Явища на межах поділу біля поверхні індивідуальних та складних пірогенних нанооксидів
The aim of this review paper was to analyze investigation results on interfacial and temperature behaviors of nonpolar and polar adsorbates interacting with individual and complex fumed metal or metalloid oxides (FMO), initial and subjected to various treatments or chemical functionalization and com...
Gespeichert in:
| Datum: | 2019 |
|---|---|
| Hauptverfasser: | , , , , |
| Format: | Artikel |
| Sprache: | Englisch |
| Veröffentlicht: |
Chuiko Institute of Surface Chemistry National Academy of Sciences of Ukraine
2019
|
| Online Zugang: | https://surfacezbir.com.ua/index.php/surface/article/view/675 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Назва журналу: | Surface |
| Завантажити файл: |  |
Institution
Surface| _version_ | 1869291838020517888 |
|---|---|
| author | Гунько, В. М. Туров, В. В. Гончарук, О. В. Пахлов, Є. М. Матковський, О. К. |
| author_facet | Гунько, В. М. Туров, В. В. Гончарук, О. В. Пахлов, Є. М. Матковський, О. К. |
| author_institution_txt_mv | [
{
"author": "В. М. Гунько",
"institution": "Інститут хімії поверхні ім. О.О. Чуйка Національної академії наук України"
},
{
"author": "В. В. Туров",
"institution": "Інститут хімії поверхні ім. О.О. Чуйка Національної академії наук України"
},
{
"author": "О. В. Гончарук",
"institution": "Інститут хімії поверхні ім. О.О. Чуйка Національної академії наук України"
},
{
"author": "Є. М. Пахлов",
"institution": "Інститут хімії поверхні ім. О.О. Чуйка Національної академії наук України"
},
{
"author": "О. К. Матковський",
"institution": "Інститут хімії поверхні ім. О.О. Чуйка Національної академії наук України"
}
] |
| author_sort | Гунько, В. М. |
| baseUrl_str | |
| collection | OJS |
| datestamp_date | 2020-01-28T14:32:15Z |
| description | The aim of this review paper was to analyze investigation results on interfacial and temperature behaviors of nonpolar and polar adsorbates interacting with individual and complex fumed metal or metalloid oxides (FMO), initial and subjected to various treatments or chemical functionalization and compared to such porous adsorbents as silica gels, precipitated silica, mesoporous ordered silicas, and polymeric composites. Note that the particulate morphology of FMO depends strongly not only the flame reaction conditions but also on the types and amounts of reagents, as well their distribution in the flame. Therefore, complex nanooxides can include core-shell nanoparticles, (CSNP) of 50-200 nm in size with titania or alumina cores and silica or alumina shells in contrast to simple and smaller nanoparticles of individual FMO. CSNP could be destroyed under high-pressure cryogelation (HPCG) or mechanochemical activation (MCA). These treatments as well simple hydrocompaction (controlled wetting-drying) affect the structure of aggregates of nanoparticles and agglomerates of aggregates, resulting in their becoming more compacted. The analysis shows that complex FMO could be more sensitive to external actions than simple nanooxides such as fumed silica. Any treatment of ‘soft’ FMO affects the interfacial and temperature behaviors of polar and nonpolar adsorbates. Rearrangement of secondary particles and surface functionalization affect the freezing-melting point depression of adsorbates. For some adsorbates, open hysteresis loops became readily apparent in adsorption-desorption isotherms. Clustering of adsorbates bound in pores causes reduced changes in enthalpy during phase transitions (freezing, fusion). Freezing point depression and melting point elevation cause significant hysteresis freezing-melting effects for adsorbates bound to initial and treated FMO in textural pores (voids between nanoparticles in secondary structures). Relaxation phenomena for both low- and high-molecular weight adsorbates or of filled polymeric composites are affected by the morphology of primary particles, structural organization of secondary particles of differently treated or functionalized FMO, content of adsorbates, co-adsorption order, and temperature. |
| doi_str_mv | 10.15407/Surface.2019.11.003 |
| first_indexed | 2025-07-22T19:34:56Z |
| format | Article |
| fulltext |
Поверхность. 2019. Вып. 11(26). С. 3–269 3
ТЕОРИЯ ХИМИЧЕСКОГО СТРОЕНИЯ И РЕАКЦИОННОЙ
СПОСОБНОСТИ ПОВЕРХНОСТИ.
МОДЕЛИРОВАНИЕ ПРОЦЕССОВ НА ПОВЕРХНОСТИ
________________________________________________________________________________________________________________
PACS: 31.15.A-; 31.15.bu; 31.15.E-; 31.70.Dk; 36.40.-c; 61.43.Gt; 68.03.-g; 68.43.-h; 81.40.Ef 10. doi: 15407/Surface.2019.11.003
INTERFACIAL PHENOMENA AT A SURFACE OF INDIVIDUAL
AND COMPLEX FUMED NANOOXIDES
V.M. Gun’ko, V.V. Turov, O.V. Goncharuk, E.M. Pakhlov, O.K. Matkovsky
Chuiko Institute of Surface Chemistry, 17 General Naumov Street, 03164 Kyiv, Ukraine,
e-mail: vlad_gunko@ukr.net
The aim of this review paper was to analyze investigation results on interfacial and
temperature behaviors of nonpolar and polar adsorbates interacting with individual and complex
fumed metal or metalloid oxides (FMO), initial and subjected to various treatments or chemical
functionalization and compared to such porous adsorbents as silica gels, precipitated silica,
mesoporous ordered silicas, and polymeric composites. Note that the particulate morphology of
FMO depends strongly not only the flame reaction conditions but also on the types and amounts of
reagents, as well their distribution in the flame. Therefore, complex nanooxides can include core-
shell nanoparticles, (CSNP) of 50-200 nm in size with titania or alumina cores and silica or
alumina shells in contrast to simple and smaller nanoparticles of individual FMO. CSNP could be
destroyed under high-pressure cryogelation (HPCG) or mechanochemical activation (MCA).
These treatments as well simple hydrocompaction (controlled wetting-drying) affect the structure
of aggregates of nanoparticles and agglomerates of aggregates, resulting in their becoming more
compacted. The analysis shows that complex FMO could be more sensitive to external actions
than simple nanooxides such as fumed silica. Any treatment of ‘soft’ FMO affects the interfacial
and temperature behaviors of polar and nonpolar adsorbates. Rearrangement of secondary
particles and surface functionalization affect the freezing-melting point depression of adsorbates.
For some adsorbates, open hysteresis loops became readily apparent in adsorption-desorption
isotherms. Clustering of adsorbates bound in pores causes reduced changes in enthalpy during
phase transitions (freezing, fusion). Freezing point depression and melting point elevation cause
significant hysteresis freezing-melting effects for adsorbates bound to initial and treated FMO in
textural pores (voids between nanoparticles in secondary structures). Relaxation phenomena for
both low- and high-molecular weight adsorbates or of filled polymeric composites are affected by
the morphology of primary particles, structural organization of secondary particles of differently
treated or functionalized FMO, content of adsorbates, co-adsorption order, and temperature.
Keywords: nanosilica, complex nanooxides, interfacial phenomena, adsorption, evaporation,
confined space effects.
Introduction
Particulate morphology of FMO
Fumed (or pyrogenic) metal and metalloid oxides, FMO (alumina, silica, titania, zirconia,
germania, etc.) are widely used in industry, medicine, and agriculture as individual powder
materials or components of composite materials such as filled polymers, powder or monolith
solids, suspensions, etc. [1-12]. Flame (H2/O2/N2) synthesis at 1000-1500 oC using MClx or MRx
or MClx1Rx2 (M is a metal, and R is an organic functionality, e.g., OCH3, CH3, etc.) as precursors
leads to certain general characteristics of FMO much of it related to their nano-particulate
morphology corresponding to the formation of nonporous nanoparticles (NPNP), which form
4
aggregates (< 1 m in size) and agglomerates of aggregates (> 1 m) forming a loose powder with
low bulk density (b 0.05-0.15 g/cm3 dependent on composition) [13-30]. There is also silica
fume (microsilica) analogous to fumed silica that is a by-product of the silicon and ferrosilicon
alloy production, and it consists of spherical-like particles with an average particle diameter of 150
nm. Silica fume is mainly used as pozzolanic material for high performance concrete (ASTM
C1240, standard specification for silica fume used in cementitious mixtures, http://astm.or).
Typically, FMO are composed of roughly spherical-like nonporous nanoparticles (NPNP) sized in
the 5-100 nm range (Figs. 1-11) that determines their relatively large values of the specific surface
area (Table 1, SBET). For silica fume, the primary particle size distribution (PPSD) is typically
broader than that of FMO studied (Fig. 1).
Fig. 1. Particle size distributions of FMO estimated from nitrogen adsorption data using the V/SCR method.
High resolution transmission electron microscopy (HRTEM) images recorded for
individual, binary and ternary FMO initial and differently treated [25-33] show (Figs. 1-11) that
the shape of nanoparticles could be varied from nearly ideal spherical one to irregular one, and
with very weak contacts between adjacent nanoparticles (especially of spherical shape) or more
strongly adherent nanoparticles (especially of irregular shape).
Fig. 2. (a) SEM image of initial fumed silica A-300 and (b) its TEM image after cryogelation at high pressure (~1000
atm); (c) AFM image of zirconium-containing mesoporous MCM-41 type silica [34]; (d) TEM image of a
mechanical mixture (1:1) of fumed silica A-300 and fumed alumina; and (e) SEM image of fumed silica A-300
wetted and dried.
5
Note that nanoparticles in silica gels or ordered mesoporous silicas (see, e.g., Fig. 2c) are
very strongly adherent in secondary structures (such as hard microglobules or beads) in contrast to
FMO composed of ‘soft’ agglomerates of aggregates of primary nanoparticles. This difference in
the nano- and micro-particulate morphology and hierarchy leads to differences in various
characteristics and properties of highly disperse FMO and granulated materials that will be
analyzed below.
The FMO materials are frequently called nanooxides due to the morphology of primary
nanosized particles. FMO with primary nanoparticles forming aggregates and less dense
agglomerates of aggregates are typically characterized by a small portion of the solid fraction in
the powder ~2-5% (Fig. 2). The nano-particulate morphology of FMO provides these materials
with characteristics useful for fillers of polymers, drug carriers, pigments, and thickeners due to
easy distribution of nanoparticles in polymeric or other matrices [3-5,35-45]. Additionally, these
morphological features enhance a role of the interfacial phenomena in the characteristics of the
materials applied in different media.
The properties of FMO depend not only on their composition but also on treatment
conditions and the history of the materials because of facile rearrangement of the secondary
structures with nanoparticles bonding by electrostatic and van der Waals forces with one another
in ‘soft’ powder. Note that chemical bonds between nanoparticles are practically absent in fresh
FMO powders. In contrast to individual FMO, complex (binary and ternary) nanooxides could
include not only simple uniform primary nanoparticles but also core-shell nanoparticles (CSNP)
(Figs. 3, 4, 7-11) [17,46,47]. Strong external actions (e.g., high-pressure cryogelation, HPCG, in
cryo-bombs at approximately 1000 atm) could lead to decomposition of CSNP [46,47] (Fig. 4) that
caused significant changes in the textural and other characteristics of strongly treated FMO.
Fig. 3. TEM image of CSNP with alumina/silica/titania
(AST1).
Fig. 4. HRTEM images of decomposed CSNP of AST1.
The formation of CSNP can occur when one component of a binary or ternary FMO tends to
be in a crystalline form with practically pure individual oxide (e.g., titania or alumina). However,
conditions during the flame synthesis (high temperature, flame turbulence, high velocity of
molecules, radicals, ions, protoparticles, and nanoparticles [1-23]) may prevent the formation of
large crystallites [1-8]. Small crystallites (protoparticles) could form dense aggregates
6
(polycrystalline cores), which are then covered by some oxide layers in the flame to form CSNP.
Another reason for the formation of complex CSNP is the difference in reactivity of the precursors
of different oxides, and the difference in the conditions of their appearance in the flame. For
example, SiCl4 or TiCl4 (liquids at standard conditions) are volatile and reactive compounds, but
AlCl3 dimerizes to Al2Cl6 forming a solid compound, which should be sublimated to be added to
the flame. Al2Cl6 is less reactive than SiCl4 or TiCl4. Similar differences are characteristic for other
precursors containing various metals. Therefore, distributions of different precursors in the flame
(a relatively long jet used is characterized by certain gradients of temperature and concentration of
reactants and active particles) could be nonuniform. This leads to different rates of the formation
of nuclei, protoparticles, and layers with different oxides. Additionally, the formation of M1-O-M1
or M2-O-M2 bonds (where M1 M2) could be preferable to the formation of M1-O-M2 bonds at the
surface of nuclei of some phases, especially if they tend to be crystalline [27-49]. This leads to the
formation of protoparticles with a nearly pure single phase (especially nanocrystallites such as
titania or alumina in which other oxides, e.g., silica, could be present as low-content impurities)
present in complex FMO. Note that the formation of asymmetric bridges M1-O-M2 could affect the
catalytic properties of materials [50-52], as well as aggregation of nanoparticles and agglomeration
of the aggregates [53] that play an important role in applications of these materials [1-12,54-56].
Fig. 5. HRTEM images of individual fumed oxides (a) A-300 and (b) alumina and electron diffraction pattern (insert).
For FMO, the toxicity (an important consideration in applications of these materials) of
nanomaterials depends on the surface chemistry and PPSD, as well as on the behavior of
aggregates and agglomerates in aqueous media or in the gas phase [37-40,57,58]. Some treatments
and surface modifications could reduce the toxicity of FMO [39,40,53], as well as their dusting.
Applications of FMO could give significant advantages in comparison with materials composed of
much larger particles due to high activity of mobile NPNP characterized by a significant value of
the specific surface area (up to 500 m2/g). FMO applications generate additional problems [57-60],
7
which, however, could be solved using treatment or modification of nanomaterials. Some
properties of FMO are sensitive to relatively gentle external actions (e.g., treatment using a
grinder, microbreaker, ball-mill, or even hand pressing) [48,49,53]. For example, several minute
treatments of fumed alumina or silica/alumina in a grinder (using powder matter under vibrations
of a ball (0.5-1 cm in diameter) in a cylinder, both with stainless steel) could result in changes in
the phase composition of alumina nanocrystallites [48,49] and reduce dusting of the material.
Clearly, these effects changing the FMO properties could alter their behavior in certain
applications in liquid, gaseous, or polymeric media.
Fig. 6. HRTEM images of fumed complex oxides (a) SA23, (b, d) ST20, (c) SA96, (e) ST63, and (f) AST50.
8
Fig. 7. HRTEM images of (a) MCA A-300, (b, c) A-300 cryogel, and (d) initial fumed alumina (scale bar 20, 50, 100,
and 10 nm, respectively).
Fig. 8. HRTEM images of initial AST50 at different magnifications at scale bar of (a) 20 nm, (b) 5 nm, and (c) 2 nm,
and (d) electron diffraction pattern showing the presence of titania crystallites.
9
Fig. 9. HRTEM images of (a-e) AST1 (a, b) initial (scale bar 50 and 5 nm) and (c, d, e) cryogel (scale bar 200 and 100
nm), and (f) initial AST71 (scale bar 5 nm) and electron diffraction pattern (insert) showing the presence of
both crystalline and amorphous phases.
10
Fig. 10. HRTEM images of initial (a) ST20 (scale bar 10 nm) and (b-d) ST63 at different magnifications at scale bar
of (b) 500 nm, (c) 10 nm, and (d) 5 nm.
Fig. 11. HRTEM images of initial (a) SA23 and (b) SA75 (scale bar 20 and 5 nm, respectively).
11
Table 1. Composition, specific surface area (SBET) and pore volume (Vp) of fresh fumed oxides (SCV/SCR method for
individual FMO and VCV/SCR for complex FMO) and silica gels (cylindrical pore model).
Sample
2SiOC
(wt%)
2TiOC
(wt%)
32OAlC
(wt%)
SBET
(m2/g)
Vp
(cm3/g)
A-50 99.8 52 0.126
A-100 99.8 92 0.155
A-200 99.8 206 0.463
A-300 99.8 294 0.524
A-400 99.8 409 0.859
A-500 99.8 492 0.874
SA1 98.7 1.3 203 0.416
SA3 97 3 185 0.405
SA8 92 8 303 0.688
SA23 77 23 347 0.788
SA30 70 30 238 0.643
SA75 25 75 118 0.320
Al2O3 100 125 0.262
ST2 98 2 77 0.263
ST9 91 9 235 0.580
ST14 86 14 156 0.386
ST20 80 20 84 0.174
ST40 60 40 148 0.333
ST63 33 63 84 0.215
ST65 35 65 34 0.080
ST94 6 94 30 0.100
TiO2 100 42 0.117
AST50 28 50 22 37 0.095
AST71 8 71 21 74 0.127
AST82 6 82 12 39 0.150
AST87 4 87 9 42 0.148
AST88 8 88 4 39 0.123
Si-40 99.9 742 0.636
Si-60 99.9 456 0.822
Si-100 99.9 349 1.225
The presence of CSNP in complex FMO could lead to non-monotonic changes in the surface
content of one of the components, e.g., alumina in silica/alumina and alumina/silica/titania [26,27].
Changes in the surface content of an active component (e.g., alumina or titania in binary or ternary
FMO based on silica) in complex FMO could strongly affect the interfacial behavior of adsorbates
of low and high-molecular weights as well as other important properties [53]. The non-monotonic
changes in the surface composition could influence the properties of the final materials (FMO
filled polymers, nanocomposites, aqueous suspensions, etc.). Therefore, deeper insight into such
relationships as properties vs. bulk/surface composition of complex FMO (silica/alumina,
silica/titania, alumina/silica/titania and other compositions) vs. interfacial phenomena, in
comparison with related individual FMO affected by certain additional material treatments, is of
interest from both theoretical and practical points of view. Many of publications analyzed here
were based on investigations of individual fumed oxides (silica and alumina), binary FMO
silica/titania (e.g., ST20 and ST63) and silica/alumina (SA23, SA75, and SA96), and ternary
alumina/silica/titania AST (AST1, AST50, and AST71) (Table 1) used as the initial powder
materials [25-30]. Binary and ternary fumed oxides were prepared by simultaneous high-
temperature hydrolysis of the corresponding metal chlorides (SiCl4, AlCl3, TiCl4) [25-30,46-
49,53]. Cab-O-Sil HS-5 (Cabot Corporation) was used as the initial material to prepare cryogel
and suspended-dried samples.
To prepare high-pressure cryogels with FMO, aqueous suspensions of nanooxides with
doubly distilled water or NaCl solution sonicated were frozen at 260 K (for 12 or 24 h) or 208 K
(for 12 h) or 77.4 K (for 4 h) in thick-walled stainless-steel reactors at pressures of up to 1050 atm
12
[46,47]. This high pressure was caused by ice crystallites formed in the frozen suspensions (~10–
15 mL) placed in the strongly restricted volume of cryo-bombs. The pressure of ice (~1000 atm) in
the cryo-bombs was estimated [46,47] according to [61,62]. Cryonanooxides (CNO) were dried in
air at room temperature to air-dry state. The same suspensions prepared at standard conditions
were kept at room temperature or 255-260 K for 24 h and 1 atm to prepare gelled or cryogelled
samples and then dried to air-dry state. All the final materials studied were prepared in the powder
state [46,47].
Fig. 12. HRTEM images of fumed complex oxides (a) AST1 and (b) AST71.
Features of nano-particulate morphology of various FMO (Figs. 1-16) depend specifically
on the nature, composition and amounts of different oxide components, which could be both
crystalline and amorphous (e.g., titania, alumina) or only amorphous (silica). The main feature of
FMO containing alumina and/or titania is the formation of core-shell nanoparticles, which clearly
manifest as dense alumina aggregates (mono- or polycrystalline cores) with thin silica or alumina
shells in SA (Figs. 6a and 11a) and AST (Figs. 3, 4, 9, and 12a) or titania cores (frequently
monocrystalline) with a relatively thick silica shell (Figs. 6c,d and 10). Additionally, a decrease in
the specific surface area (i.e. an increase in the average diameter (d) of primary nanoparticles since
SBET ~ 1/d) of FMO leads to a broadening of the size distribution of primary nanoparticles (Figs. 1-
16) [25-30,46-49,53,63]. For SA materials, relatively large CSNP (50 nm in diameter (d) or larger)
were observed even for SA23 (Figs. 6a and 11a), which exhibits a large SBET value (Table 1) that
corresponded to a small average nanoparticle size (for SA23, dav = 6.9 nm).
13
Fig. 13. HRTEM images of A-300 (a) initial and (b) cryogel, and AST1 (c) initial and cryogel.
For AST materials with lower SBET values, CSNP could be 50-200 nm in size, while
smaller simple nanoparticles could be also found in these FMO (Figs. 6-13). FMO nanoparticles
with titania cores and silica shells are also clearly visible in AST (Figs. 2c,d, 8 and 10). However,
complex ST nanoparticles (or AST at high TiO2 amounts) frequently include only one titania
crystallite (anatase/rutile) core with a silica shell (Figs. 14-16). Titania polycrystalline cores were
less frequently observed than polycrystalline alumina cores. This difference could be explained by
features of the pyrogenic synthesis of complex FMO prepared using such precursors as TiCl4,
SiCl4, and Al2Cl6 with different properties and reactivity [25-28,64-67].
In our works analyzed here, X-ray diffraction (XRD) patterns (Figs. 14-16) were recorded
over 2θ = 10-70° range using a DRON-4-07 (Burevestnik, St. Petersburg) diffractometer with Cu
K (λ = 0.15418 nm) radiation and a Ni filter. Analysis of the crystalline structure of alumina was
carried out using the JCPDS Database (International Center for Diffraction Data, PA, 2001) for -
Al2O3 (JCPDS#29-0063) and -Al2O3 (JCPDS#46-1212).
The crystallinity of complex FMO (Figs. 14-16) increased with decreasing silica content
and increasing temperature during the synthesis or treatment [25-30,53,68]. Additionally, silica in
complex FMO could delay the phase transition of anatase to rutile or low-temperature aluminas to
corundum [53]. The crystallinity of titania in FMO is greater than that of alumina, and silica is
14
totally amorphous (Figs. 14 and 16); however, preheating of A-300 at 1400 oC resulted in
crystallization of silica to -cristobalite (Fig. 14d). The mentioned difference in titania and
alumina could be explained by the higher reactivity of TiCl4 compared with Al2Cl6. The higher
reactivity resulted in titania nuclei, protoparticles and nanoparticles growing faster in the flame
than alumina nanostructures. These differences appeared in the 29Si and 27Al NMR spectra (Figs.
17 and 18, Table 3), as well as in the surface content of alumina and titania vs. their total content in
FMO (Fig. 19). The latter is nearly monotonic for titania and non-monotonic for alumina. At low
silica content (~5 wt.%) in complex FMO (e.g., ST94, SA96, AST82), silica formed a solid
solution in the main phase (titania or alumina). This was indicated by 29Si NMR signals of silica
structures with Si-O-Si bonds absent in the spectra.
Fig. 14. XRD patterns of FMO: (a) alumina, SA, AST1, (b) ST and A-300, (c) AST and pure anatase and rutile; and
(d) initial and heated A-300.
Fig. 15. Spacing distribution function for fumed silica (amorphous) and titania (crystalline) calculated from HRTEM
images of A-200 and fumed titania using Fiji software (http://fiji.sc/Fiji).
15
Fig. 16. XRD patterns of initial and MCA silica A-300 and alumina and their blend (1:3) initial and MCA.
Table 2. Textural characteristics of FMO initial and differently treated [46,47,53].
Sample SBET
(m2/g)
Snano
(m2/g)
Smeso
(m2/g)
Smacro
(m2/g)
Vp
(cm3/g)
Vnano
(cm3/g)
Vmeso
(cm3/g)
Vmacro
(cm3/g)
Initial A-300 330 29 288 13 0.826 0.012 0.621 0.193
MCA 6 h A-300 328 85 223 20 1.325 0.019 0.978 0.327
A-300 (II) 303 94 200 10 0.735 0.035 0.523 0.176
A-300 (II) HPCG 297 38 258 2 0.826 0.012 0.798 0.017
Alumina 70 16 51 3 0.181 0.007 0.118 0.056
SA23 347 93 238 15 0.815 0.024 0.526 0.265
SA75 118 13 99 6 0.320 0.002 0.200 0.017
ST20 84 5 76 3 0.179 0.001 0.128 0.050
ST63 84 11 68 5 0.215 0.002 0.123 0.090
Initial AST1 78 12 62 4 0.234 0.006 0.156 0.072
MCA 30 min AST1 60 6 50 4 0.308 0.003 0.237 0.068
ATS1 HPCG 143 49 92 2 0.594 0.016 0.544 0.034
AST50 37 7 29 1 0.084 0.001 0.054 0.029
AST71 74 12 60 2 0.127 0.002 0.076 0.049
Initial A-300/AST1 (1/1) 184 19 158 7 0.481 0.005 0.366 0.110
MCA 30 min A-300/AST1 142 17 118 7 0.456 0.004 0.324 0.128
A-300/AST1 HPCG 145 41 53 51 1.077 0.010 0.190 0.877
Note. The Vnano and Snano values were calculated by integration of the fV(R) and fS(R) function, respectively, at 0.35 nm
< R < 1 nm, Vmeso and Smeso at 1 nm < R < 25 nm, and Vmacro and Smacro at 25 nm < R < 100 nm. Cryogels with A-300
and A-300/AST1 were prepared (20 wt.% and 14 wt.% suspensions, respectively) at 208 K and ~1000 atm for 12 h.
AST1 cryogel was prepared (sonicated 20 wt.% suspension) in a cryo-bomb at 260 K for 24 h and 77.4 K for 4 h at
~1000 atm.
The surface chemistry of the silica powders was further characterized via solid state 29Si
NMR analysis using a 400 MHz Bruker Avance FT-NMR spectrometer (Bruker Corporation,
Billerica, MA, USA). Cross polarization and magic angle spinning (CP/MAS) were utilized to
increase 29Si signal intensity and to overcome signal broadening associated with solid state NMR.
A 4 mm probe was used, the number of scans was 16,384 (a run time of about 14 hours), the
sweep width was 30 kHz (380 ppm), the delay time between scans was 3.00 seconds and the CP
contact time was 6 ms. The transmitter frequency (SFO1) for the Si nucleus was 79.4886563 MHz,
while the transmitter frequency used for the 1H nucleus was 400.1314005 MHz. Solid-state 1H
MAS NMR spectra were recorded at a spin speed of 10 kHz and a recycle delay of 1 s, and
adamantane was used as a reference. For complex FMO, the 29Si (resonance frequency 59.595
MHz) cross-polarization spinning at the magic angle (CP/MAS) NMR spectra and the 27Al
(resonance frequency 78.172 MHz) MAS NMR spectra (program ZG) of complex oxides were
recorded by a Bruker Avance 300 NMR spectrometer (magnetic field of 7.046 Tesla, a spinning
rate of 8 kHz of 4 mm zirconia rotor). Chemical shifts of 29Si and 27Al were referenced to
16
tetramethylsilane and an Al(NO3)3 aqueous solution respectively; i.e., the resonance of Si(CH3)4
and Al(H2O)6
3+ was set to 0 ppm (Fig. 18, Table 3).
Fig. 17. (a) 1H MAS and (b, c) 29Si CP/MAS NMR spectra of (a, b) unmodified silicas and (c) initial Cab-O-Sil HS-5,
suspended-dried and suspended in NAOH solution (pH 10) and dried. =Si(OH)2, SiOH, and Si(OSi)4
correspond to Q2
Si, Q
3
Si and Q4
Si sites at 91, 100, and 110 ppm, respectively. Silicas: fumed silica A-300
(Pilot plant of Chuiko Institute of Surface Chemistry, Kalush, Ukraine) preheated at 450 °C for 2 h; preheated
A-300 hydro-compacted with water (1:2 w/w), stirred for 10 min, and then heated at 105 °C for 8 h (cA-300);
fumed silica A300 (Evonik) preheated at 450 °C for 2 h; silica TS 100 (Evonik Ind.); Syloid® 244
(precipitated silica, Grace Davidson), and Cab-O-Sil HS-5 (Cabot Corporation).
Table 3. Contributions (in %) of different centers with Si and Al in SA, ST and AST samples determined as relative
integral intensity of the bands obtained on deconvolution of the 29Si CP/MAS and 27Al MAS NMR spectra.
Sample Q4
Si Q3
Si Q2
Si QAl(VI) QAl(V) QAl(IV)
SA1 45.7 33.3 21.0 60.3 39.7
SA3 31.1 52.6 14.2 77.9 1 21.1
SA8 48.0 45.0 7.0 73.4 26.6
SA23 45.5 33.4 21.1 55.9 3.5 40.6
SA30 28.8 50.7 20.5 51.0 4.2 44.8
ST9 33.6 44.7 21.7
ST14 44.0 31.8 24.2
ST20 29.1 54.6 16.3
ST63 57.3 16.6 26.1
ST65 64.9 19.5 15.6
AST50 68.1 23.6 8.3 86.2 13.8
AST82 94.1 3.4 2.5
17
Fig. 18. (a, b) 29Si and (c, d) 27Al MAS NMR spectra of fumed oxides (a) ST and A-200, (b, d) SA and A-200, and (c)
AST and Al2O3 samples.
The properties of FMO could be easily changed by modification of a nanoparticle surface
using functionalization by various silanes or other reagents [69-73], treatment in plasma [74],
mechano-chemical activation [75,76] or by grafting of another oxide (e.g., titania, zirconia, iron
oxide) onto silica matrices [53,77-82]. Note that similar nanomaterials could be synthesized at low
temperature in liquid media [83], e.g., precipitated silica or complex oxides. All these
modifications of FMO change morphological, structural, and textural characteristics of the
materials, and, therefore, the interfacial behavior of adsorbates [53].
Total chemical composition of complex oxides (Fig. 19) was analyzed by XRF (Canberra,
USA) spectrophotometer with a 55Fe (or 109Cd) radioactive source and an amplitude analyzer
(Canberra) coupled with a computer with the AXIL program. The titania phase in titania-
containing samples (at
2TiOC 50 wt.%) consists of a blend of anatase and rutile at Canatase/Crutile
between 7.3 (AST50) and 0.84 (ST65). In the case of low amounts of titania in ST, the titania
phase consists of a major portion of anatase with a minor contribution of amorphous oxide
distributed in the silica matrix. All silica samples as well as silica and alumina in complex oxides
are amorphous. Initial fumed alumina is mainly amorphous with approximately 20% of a
crystalline (-Al2O3) phase (impurities 0.5 wt.% of metal oxides, mainly Fe2O3 (< 0.25 wt.%) and
TiO2 (< 0.16%)).
18
Fig. 19. Relationships between total CX and surface content S
XC of (a) alumina and titania, (c) silica in SA, ST and
AST samples; (b) relationship between ratio S
XC /CX and the total content of the second phases, and (d)
relationship between the surface content of alumina or titania and the peak temperature of TPD MS
thermograms of desorbed water.
Surface content of aluminum (at%) in SA and AST and titanium in ST and AST (Fig. 19)
was determined by Auger electron spectra (AES) recorded by a JAMP-10S (JEOL) spectrometer.
The nanooxide powders were prepared with an indium/carbon matrix to avoid the charge buildup
on the dielectric samples. The spectra with minimum intensity of the indium and carbon lines were
selected for subsequent analysis. Sample regions most characteristic and optimum for the AES
studies were determined by the electron scanning microscopy (energy of electron beam 5 keV,
beam current 2×1010 A, beam diameter 0.05-0.1 m). The differential Auger electron spectra
E×dN(E)/dE were recorded with an energy analyzer such as a “cylindrical mirror” with the energy
resolution of E/E = 0.007, a step of 1 eV, modulation amplitude of voltage on the energy
analyzer 4 V, circuit voltage 2.5 kV, and the time constant of the amplification circuit of 1 s. The
surface Si and Al contents were determined by the analysis of the LVV lines, and the surface
titanium content was estimated using the LMM line.
The NMR spectra show that the state of the Si atoms depends more strongly on the
treatment of silica (Fig. 17c) or the presence of another phase (Fig. 18) than on a type of initial
silica (Fig. 17b). Additionally, nonuniform distribution of the various phases in FMO NPNP (Fig.
19) can strongly affect the properties of the materials (e.g. Fig. 19d).
Textural characteristics of FMO
To analyze the textural characteristics of nanooxides (typically degassed at 373-473 K for
several hours, Table 1), low-temperature (77.4 K) nitrogen adsorption–desorption isotherms were
recorded using Micromeritics ASAP 2405N or ASAP 2420 adsorption analyzers, Quantachrome
Autosorb adsorption analyzers or other apparatuses [25-30,46-49,53,63-83]. The specific surface
19
area (SBET) was calculated according to the standard BET method [84]. The total pore volume Vp
was evaluated from the nitrogen adsorption at p/p0 0.99, where p and p0 denote the equilibrium
and saturation pressure of nitrogen at 77.4 K, respectively [85]. The nitrogen desorption data were
used to compute the pore size distributions (PSD, differential fV(R) ~ dVp/dR and fS(R) ~ dS/dR)
using DFT method or a self-consistent regularization (SCR) procedure [86] under non-negativity
condition (fV(R) 0 at any pore radius R) at a fixed regularization parameter = 0.01 using a
complex pore model applied with slit-shaped (S) and cylindrical (C) pores and voids (V) between
spherical nonporous nanoparticles packed in ordered or random aggregates (SCV/SCR for
individual FMO or VCV/SCR for complex FMO) [86]
ppp
pr
r
r
pr p
pp dRRfRpt
R
w
dRRfA
k
k
)(),()(
)(
)(min
max
(1)
where rmin and rmax are the minimal and maximal half-width or pore radius, respectively; w = 1
(slit-like pores), 2 (cylindrical pores) and 1.36 (voids between spherical particles packed in a cubic
lattice);
)/ln(
cos2
),()(
0 ppTR
Rptpr
g
m
pk
(2)
])2/2/()2/2/()1(1[
])2/2/()1()2/2/(1[
)1(
),(
1
11
nn
nnn
BET
m
p zccbzccbzc
znnbznbznnb
z
cz
S
a
Rpt (3)
b = exp(/RgT); is the excess of the evaporation heat due to the interference of the layering on
the opposite wall of pores (determined as a varied parameter using local isotherm approximation,
LIA); t(p,Rp) is the statistical thickness of adsorbed layer; am is the BET monolayer capacity;
)/)exp(( TRQQcc gsps ; cs is the BET coefficient for adsorption on flat surface (calculated
using LIA); Qs and Qp are the adsorption heat on flat surface and in pores, respectively; z = p/p0; n
= Rp/tm is the number (noninteger) of statistical monolayers of adsorbate molecules. Desorption
data were utilized to compute the f(Rp) distribution with Eq. (1) using a regularization procedure
under non-negativity condition (fV(Rp) 0 at any Rp), at a fixed regularization parameter = 0.01.
A model of a pore mixture with slitshaped and cylindrical pores and voids between spherical
nanoparticles packed in random aggregates (SCV model, MND method) was developed [86]
because the standard pore models with slitshaped or cylindrical pores are inappropriate to describe
the textural porosity of nanosilicas. The self-consistent regularization (SCR) procedure was used
to calculate the PSDs and to determine contributions of different types of pores to the pore
volume. The fV(Rp) functions linked to pore volume can be transformed to the fS(Rp) distribution
functions with respect to surface area using the corresponding pore models
))(()(
p
p
pV
p
pS R
V
Rf
R
w
Rf (4)
where w = 1, 2, 3, and 1.36 for slitlike, cylindrical, spherical pores and gaps between spherical
particles packed in the cubic lattice. The fV(Rp) and fS(Rp) functions were used to calculate
contributions of nanopores (Vnano and Snano at the pore radius Rp < 1 nm), mesopores (Vmeso and
Smeso at 1 nm Rp 25 nm), and macropores (Vmacro and Smacro at Rp > 25 nm) to the total pore
volume and the specific surface area. The values of Snano, Smeso, and Smacro were corrected that Snano
+ Smeso + Smacro = SBET. For estimation of deviation of the pore shape from the model, a criterion
w = SBET/Ssum – 1 where
j
R
R
p
p
jp
pjV
p
j
j
j
R
R
ppjSjsum dR
R
V
Rf
R
w
cdRRfcS
max
min
max
min
))(()( ,
,, , (5)
cj is the weight coefficient, Rmin = 0.3 nm, Rmax = 100 nm, j denotes a pore model, fS,j(Rp) is the
differential distribution function with respect to the specific surface area of pores of the j-th type.
The integral adsorption equation (1) may be re-written as follows
20
( ) ( )j j
j
A p c A p =
, max,
min ,
( )
, ,
( )
( ) ( , , ( )) ( )
/ 2
k j j
k j
r p r
j
j V j j V j
j sfr r p
w
c f R dR t p R a f R dR
R
, (6)
where fV,j(R) is the PSD for the j-th type of pores, wj is the pore type formfactor (1 for slit, 2 for
cylindrical, 1.36 for voids between spherical particles packed in a cubic lattice), cj = cslit, ccyl, and
cvoid are the weight constants (cslit + ccyl + cvoid = 1) determining contributions of slitshaped and
cylindrical pores and voids between spherical particles to the volume filled by adsorbate at
pressure p (A(p)); rk,j is the half-width of slit pores, the radius of cylindrical pores or meniscus
radius for voids determined through modified Kelvin equation including the thickness of the
adsorbate layer tj. In the case of the presence of nanoparticles (cj = cvoid) characterized by the size
distribution function (a), Eq. (6) can be re-written with additional integral in the second term
(V/SCR method)
max, max
, min
,
( )
( ) ( , , ) ( )
/ 2
j
k j
r a
j
V j j
sfr p a
w
f R t p R a a da
R
(7)
which is due to a certain distribution function of nanoparticle sizes becoming broader with
decreasing SBET value of fumed oxides. The cj values were determined for the best fitting of the
experimental isotherms using self-consisting regularization procedure with several steps in
solution of the sum of integral equations. sf = (s + f)/2 is the average collision diameter of
surface and fluid atoms; rk,j is determined through modified Kelvin equation including the
thickness of the adsorbate layer tj. Lennard–Jones potential was used for slitshaped and cylindrical
pores and gaps between spherical particles. This method gives better results than that used
previously for complex adsorbents with one distribution function for all the pore shape models
because of the use of independent fV,j(Rp) functions for each pore type. For a pictorial presentation
of the pore size distributions, the fV(Rp) functions were re-calculated to incremental PSDs (IPSDs)
)))(()((5.0)( 1,,1,,, ipipipVipVipV RRRfRfR . (8)
Additionally, fS(R) was used to estimate the deviation (w) of the pore shape from the model using
a self-consisting (to better fit the nitrogen adsorption isotherms) regularization in the case of the
use of a complex model of the pore shape
1
)(
max
min
R
R
S
BET
dRRf
S
w (9)
where Rmax and Rmin are the maximal and minimal pore radii respectively. The S*
nano, S
*
meso and
S*
macro values were corrected by multiplication by (w+1) that gives S*(w+ 1) = Ssum = Snano +
Smeso + Smacro = SBET. The effective w value (wef) can be estimated with equation
max
min
max
min
)(
)(
R
R
V
R
R
V
p
BET
ef
dRRf
dRRRf
V
S
w (10)
The specific surface area (S) of materials composed of spherical nanoparticles
characterized by the particle size distribution (a) (calculated using the self-consisting
regularization for fV(R) and (a) with the model of voids between spherical particles) can be
calculated with equation
daat
A
ar
taNaA
A
a
Nrta
a
S
a
a
m
m )())((arcsin)(2
2
3max
min
222
3
(11)
where mrtaA , a is the particle radius, the density of material, N the average coordination
number of nanoparticles in aggregates, t the thickness of an adsorbed nitrogen layer, and rm is the
21
meniscus radius determined at 0.05 < p/p0 < 0.2 corresponding to the effective radius R of voids
between spherical particles. Condition S = SBET can be used to estimate the N value. An additional
criterion |<S> SBET| < 1 m2/g was used to determine the amin and amax values for the (a)
distributions calculated at p/p0 < 0.5 (i.e. before capillary condensation starts) with
m
mm
dtdr
dtdrtrS
S
),(
. (12)
The values of <RV> and <RS> as the average pore radii calculated as a ratio of the first moment of
fV(R) or fS(R) to the zero moment
<R> = f(R)RdR/f(R)dR (13)
The fV(R) and fS(R) functions were also used to calculate contributions of nanopores (Vnano and
Snano at 0.35 nm < R < 1 nm), mesopores (Vmeso and Smeso at 1 nm < R < 25 nm), and macropores
(Vmacro and Smacro at 25 nm < R < 100 nm) [86]. The morphology of primary nanoparticles as well
as of particles of higher hierarchic levels is typical and nearly the same for all individual and
complex fumed oxides [53,63-83]. However, the difference in the structure of active surface sites
of acidic or basic characters could affect the interfacial behavior of adsorbates, even nonpolar such
as nitrogen. For example, the interaction energy is higher on the adsorption of nitrogen molecules
on titania than on silica or alumina due to stronger dispersive, electrostatic (13.3, 9.5, and 5.4
kJ/mol) and charge transfer (5.2, 2.9, and 1.6 kJ/mol) interactions in complexes –O–HNN
with TiO(H)Ti, SiO(H)Al and SiOH groups, respectively [53], according to the Kitaura-
Morokuma analysis of these complexes using the GAMESS program suit with the 6-31G(d,p)
basis set [87,88]. Both structural and energetic factors caused a small difference in the nitrogen
adsorption at very low pressures p/p0 < 104 (Fig. 20a) [88]. This resulted in the same shape of
normalized adsorption isotherms (p/p0) (Fig. 20a) despite the difference in the particle size
distributions (PaSD) as well as in the values of SBET of FMO. This similarity is better seen for
derivatives d/d(p/p0) (Fig. 20b). However, there is some difference in the nitrogen adsorption
energy distributions for these FMO (Fig. 20c) because of the effects of the chemical structure of
silica, alumina, titania and related binary and ternary systems possessing different types and
amounts of terminal and bridging hydroxyls of different acidity. Notice that these factors could
more strongly affect the adsorption/desorption of water because of its more specific interactions
with FMO [53].
Heating of nanosilica at different temperatures (over the 473-1173 K range) for different
time (10 min – 5 h) weakly affected the morphology of particles (Fig. 21a,b) [88]. However,
heating of nanosilica to 1673 K led to formation of -cristobalite (Fig. 14d). In contrast to the
nitrogen adsorption on various FMO when their difference appeared at low pressures (Fig. 20)
heating of nanosilicas led to a larger difference in the adsorption at higher pressures. This was due
to changes in the characteristics of both primary and secondary particles that are well seen for PSD
(Fig. 21c) more sensitive to heating than the normalized isotherms because these changes were
over the total range of the pore size caused by associative desorption of water from both surface
and bulk of primary particles. Therefore, the specific surface area depended on treatment
conditions (Fig. 22) since the |SBET/SBET| values change over a wide range up to ~0.3 on heating
at 1173 K for 2 h [88]. It should be noted that both (p/p0) and d/d(p/p0) functions are sensitive
to changes in the particle morphology. This is well seen for a set of silicas: fumed silicas – silica
gels – mesoporous ordered silicas (Fig. 23) [88]. However, these differences appear mainly at p/p0
> 0.4 when capillary condensation starts in mesopores of various shapes, and this effect depend on
PSD and topology of pores. Additionally, the d/d(p/p0) curve shapes differ at p/p0 < 0.02 for
porous and fumed silicas (Fig. 23c) due to the presence of narrow pores in the first one in contrast
to nanosilicas loosely composed with nonporous primary nanoparticles practically without the
formation of nanopores (R < 1 nm) between them.
22
Fig. 20. (a) Relative adsorption of nitrogen ( = a/am)
and (b) derivatives d/d(p/p0) as function of
relative pressure; and (c) adsorption energy
distributions f(E) for set of fumed oxides (A-
300, Al2O3, TiO2, SA1, SA3, SA23, SA30,
ST9, ST14, ST20, ST29, AST50).
Fig. 21. (a) Relative adsorption of nitrogen ( = a/am)
and (b) derivatives d/d(p/p0) as function of
relative pressure; and (c) DFT IPSD (model of
voids between spherical particles) for two
samples of A-300 (weight loss of 3.2 and 1.3
wt.% on heating to 1173 K and bulk density of
0.024 and 0.028 g/dm3, respectively) heated at
different temperatures (473, 573, 873, and 1173
K) for different time (from 10 min to 5 h).
Fig. 22. Changes in the specific surface area of two samples A-300 heated at different temperatures.
23
Fig. 23. (a, b) Relative adsorption of nitrogen ( = a/am) and (c, d) derivatives d/d(p/p0) as function of relative
pressure for (a, c) fumed silicas (A-50, A-100, A-200, A-300, A-400, and A-500), silica gels (Si-40, Si-60,
and Si-100) and mesoporous ordered silicas (MCM-41, MCM-48 and SBA-15); and (b, d) A-300 initial (1),
wetted-dried (2), and dried after hydrothermal treatment at different temperatures (curves 3-5).
Fig. 24. Relationship between the specific surface area (SBET) and the pore volume (Vp) for (a) initial various
individual (silica, alumina, and titania), binary (silica/alumina, SA, and silica/titania, ST) and ternary
(alumina/silica/titania, AST) fumed oxides and MCA A-300; and (b) porous silica gels (Si-40, Si-60, and Si-
100) and mesoporous ordered silicas (MCM-41, MCM-48, and SBA-15) [53].
24
The presence of pores with a narrow PSD in ordered silicas or silica gels causes appearance of a
maximum in the d/d(p/p0) curves at 0.2 < p/p0 < 0.85 depending on the size of these pores, since
the narrower the pores, the lower the pressure of this maximum [88]. Fumed oxides do not have
ordered structure of pores; therefore, a similar d/d(p/p0) maximum is absent. A similar maximum
was absent after hydrothermal treatments of fumed silica A-300 at different temperatures (Fig.
23d) despite significant re-arrangement of primary particles and changes in their size [53,89].
Additionally, a similar maximum was absent on suspending of nanosilica (CA-300 = 1-20 wt.%),
sonication, drying on air, and heating at 473 K for 2 h [53]. Consequently, fumed oxides are
morphologically relatively stable with respect to primary nanoparticles, but secondary structures
are ‘soft’ and could be easily rearranged.
Fig. 25. Pore size distributions (SCV/SCR) of initial and differently treated FMO (a) A-300 initial, ball-milled for 6 h
and cryogel prepared in a cryo-bomb (using sonicated 20 wt.% aqueous suspension) at 208 K and ~1000 atm
for 12 h, (b) AST1 initial, MCA in a microbreaker for 30 min and cryogel using sonicated 20 wt.% aqueous
suspension frozen in a cryo-bomb at 260 K for 24 h and 77.4 K for 4 h at ~1000 atm, (c) initial SA and
alumina, (d) initial ST, (e) initial AST, and (f) A-300/AST1 (1:1) initial, MCA in a microbreaker for 30 min
and cryogel (using sonicated 14 wt.% aqueous suspension) prepared in a cryo-bomb at 208 K and ~1000 atm
for 12 h.
The adsorption isotherms of nitrogen could be used to determine not only the values of SBET
and Vp and the PSD functions but also the primary particle size distributions (PPSD) using pair
integral equations related to PSD and PPSD and solved with the self-consistent regularization
procedure [86,88]. Heating of nanooxides causes several processes (such as dehydration of the
surface and the bulk of primary particles, chemical binding of adjacent nanoparticles due to
25
condensation of hydroxyls, rearrangement of secondary and ternary particles) that result in
structural changes of nanooxides depending on the heating temperature. As a whole the
morphology of primary nanoparticles of fumed oxides is stable on different treatments in contrast
to the structure of secondary particles and especially higher level particles [53,88,89].
Fig. 26. NLDFT pore size distributions of (a) fumed
silica A-380 initial and after hydrothermal
treatment at 150 oC (A-380HTT) and silica gel
200DF Gasil; (b) silica gels Si-40, Si-60, and
Si-100; and (c) precipitated silica Syloid 244
and mesoporous ordered silicas MCM-41,
MCM-48 and SBA-15.
Fig. 27. (a) Bulk density of the silica powders (dried in
air for 24 h) treated in different media (air and
water, ethanol or water/ethanol ~0.5 g per
gram of dry silica) as a function of MCA time;
(b) desorption (in air at room temperature) of
water or ethanol from MCA-treated silicas,
and (c) corresponding desorption rate as a
function of time at the initial fast stage of the
desorption.
Typically, FMO with increasing SBET value (i.e., with decreasing size of primary particles)
are characterized by increased pore volume (Vp) due to enhanced agglomeration of nanoparticles
of smaller sizes (Fig. 24). The empty volume in the FMO powders can be about 95% [53]. For
example, b = 0.045 and 0.13 g/cm3 for A-300 and A-50, respectively, that correspond to Vem =
21.8 and 7.2 cm3/g; however, the Vp values are much lower (0.524 and 0.126 cm3/g, Table 1).
26
Voids between nonporous nanoparticles in secondary structures determine the textural porosity of
the powders.
These are several types of voids between nanoparticles in FMO. Voids between adjacent
nanoparticles contribute nanopores at R < 1 nm, and narrow mesopores at 1 nm < R < 3-5 nm.
Neighboring nanoparticles in the same aggregate give narrow mesopores, while distant
nanoparticles in the same aggregate, or from neighboring aggregates, contribute to broader
mesopores or macropores. Larger voids between nanoparticles from neighboring aggregates in
agglomerates give rise to macropores. Voids between agglomerates correspond to large
macropores, which remained practically empty during adsorption of nitrogen even at p/p0 close to
1. All these voids give 4-6 peaks in the PSD (Fig. 14) [90]. Practically any treatment led to an
increase in the bulk density of the FMO powders [53]. The secondary structures (aggregates of
primary nanoparticles and agglomerates of aggregates) became denser and the PSD profiles
changed (Figs. 25a,b,f and 26a). This led to an increase in Vp despite a decrease in Vem. All of
these effects are clearly visible in changes in the PSD calculated using the SCV/SCR or VCV/SCR
methods [86], taking into account the presence of silica, alumina and titania in complex FMO (Fig.
25). Large changes in the PSD were observed for high-pressure cryogels of A-300 (Fig. 25a),
AST1 (Fig. 25b) and a mixture of A-300 and AST1 (Fig. 25f). MCA of FMO in a ball mill (Fig.
25a) or a microbreaker (Fig. 25b,f) have smaller effects than HPCG [46,47,63,90]. MCA shifts the
main PSD peaks toward smaller pore size with a reduction in SBET (mainly due to decrease in
Smeso), and could either increase Vp (A-300, AST1) or decrease Vp (A-300/AST1 mixture). HPCG
increased the value of SBET of AST1 due to decomposition of core-shell nanoparticles [46,47].
However, it did not affect SBET of A-300, and decrease of SBET for A-300/AST1, since the
aggregates became denser (i.e. accessibility of nanopores decreased for nitrogen molecules). The
Vp value increased for all samples due to cryogelation. Note that normal-pressure cryogelation is
frequently used for structuring metal oxide materials [92-99]; however, HPCG could cause much
larger changes in the texture of these materials [46,47,63]. Hydrothermal treatment (HTT) of
fumed oxides [89] caused strong changes in PSD (Fig. 26a), which became similar to those of
porous silicas (Fig. 26b,c).
Note that intensive MCA could strongly change the properties of treated materials. For
example, changes of gibbsite and fumed silica mixture during MCA investigated using XRD,
MAS-NMR and XPS [100] were observed for crystallinity and other intrinsic structural
characteristics. Formation of new Al–O–Si chemical bonds was detected by the changes in Al
coordination number of gibbsite in the mixture during milling, and the appearance of new
resonance in 29Si NMR spectrum of the ground mixture. The presence of Al–O units in silica
framework was demonstrated by the increase of Al 2p binding energy and the decrease of Si 2p
binding energy [100].
Mesopores (1 nm < R < 25 nm) are the main contributors to porosity and specific surface
area of various FMO [53]. Wetting, suspending-drying, MCA, and HPCG of FMO resulted in an
increase in mesoporosity and sometimes macroporosity of treated materials [46,47,53,63,90].
These effects could be explained by the origin of the textural porosity of FMO that is caused by
voids between nanoparticles in secondary structures. Note that the empty volume Vem = 10-25
cm3/g in the FMO powders is much greater than the value of Vp (0.1 – 1.3 cm3/g, Table 1)
determined from nitrogen adsorption because the adsorption of nitrogen in large macropores is too
small (the molecules do not sense the oxide surface far from pore walls).
Solvents (e.g., water, alcohol) in the amounts of 30-50 wt.% in respect to the FMO amount
could form a multilayer coverage of all primary particles in the wetted powders. The water or
ethanol molecules could penetrate in contact area between adjacent nanoparticles in aggregates.
The water molecules (half smaller in volume than ethanol molecules) could penetrate in narrower
voids and strongly decompose the particle-particle linkages in the aggregates. Contrariwise, water
clusters could be bridges between nanoparticles from neighboring aggregates that caused
diminishing voids between these aggregates in the wetted-dried powders. Primary particles in the
initial silica practically do not have the siloxane bonds with neighboring particles. They were
27
attached one to others by the hydrogen bonds and electrostatic and dispersion interactions [53].
The solvent molecules (especially water) could break these inter-particle bonds and increase the
mobility of nanoparticles during MCA treatment or sonication. The solvents in the used small
amounts provide essential plasticizing effects on the wetted silica powders. The MCA treatment of
water-wetted silica gave the powder with a larger bulk density (~0.4 g/cm3, Fig. 27a) due to better
packing of primary particles than that after the MCA in the dry or ethanol atmosphere [75].
However, this maximal b value corresponded to approximately 35% of dense and ordered-packed
silica spheres of the same size (d = 6/(0SBET) = 8.26 nm at 0 = 2.2 g/cm3 and SBET = 330 m2/g)
that in a cubic lattice corresponded to b 1.15 g/cm3. This difference could be explained by, at
least, two factors: (i) primary particles of nanosilica have different sizes; and (ii) certain amounts
of non-decomposed aggregates could remained after the MCA. Both factors, as well as a random
structure of initial aggregates, prevent the ordered packing of silica nanoparticles in secondary
particles upon MCA treatment. However, compaction of primary particles became stronger with
increasing MCA time (tMCA) under the plasticizing effect of a solvent because of much deeper
rearrangement of agglomerates and aggregates [75].
Fig. 28. Thermogravimetric (TG) and differential TG (DTG) curves of weight loss on heating of initial silica (curves
1) and water-wetted MCA-sample treated for 6 h (curves 2).
The particle compaction can affect the desorption rate of water or ethanol on evaporation at
room temperature in air (Fig. 27b). However, this dependence is complex, as well as changes in
the PSD of FMO affecting this process. The pore (voids between spherical nanoparticles could be
considered as textural pores) size distributions depended on tMCA [75].
On heating, desorption of both intact water (T < 150 oC) and associatively desorbed water
(or water desorbed from the inner volume of primary particles at T > 150 oC) increased after the
MCA treatment (Figs. 27b and 28) [75]. The difference in the amounts of water desorbed from
initial and MCA-treated silica slightly increased with heating temperature to 900 oC. This could be
explained by larger amounts of water molecules and OH groups inside silica nanoparticles treated
with 0.5 g of water per gram of silica (added into the ball-mill) for 6 h, as well as retardation of
water desorption from narrower pores in denser secondary structures. Thus, a variety of treatments
could be used to change the characteristics of FMO as relatively soft materials. The changes in the
structure – property relationships could be analyzed using XRD, Raman, UV-vis, FTIR, NMR,
DSC, and other methods [25-30,46-49,53,71-91]. Note that FMO are more sensitive to various
treatments in aqueous media than in organics and the effects depend on the content of nanooxide
in the suspension. The adsorption and interfacial and temperature behaviors of nonpolar and polar
compounds interacting with FMO at different temperatures are very sensitive to the
morphological, structural and textural changes of the materials caused by the mentioned treatments
[53].
28
Structural features of FMO
The crystallinity of simple and complex nanooxides depend on the crystallinity of titania or
alumina since silica is always amorphous in initial FMO (Figs. 14-16 and 29) [30,53]. Individual
titania includes both anatase and rutile (Figs. 29b,d). For other titania-containing FMO, the titania
phase is composed mainly of anatase. This is of importance because the crystallinity and type of
titania phases could influence the adsorption, catalytic, and other important characteristics of the
materials [53].
The UV-vis DRS spectra (Fig. 29) recorded at room temperature in the wavelength range of
200–900 nm (using ZnO as a reference material) show the MCA or HPCG effects on the structure
of treated FMO [90]. The differences in the nature and morphology of FMO caused differences in
adsorption and other properties of the materials [53]. Note that many properties of fine powder of
FMO depend on material history because their ‘soft’ nature due to easy rearrangement of
secondary particles upon pressing, MCA, wetting or suspending and drying, and high temperature
treatments [25-30,46-49,53,63-68,101-103].
Fig. 29. (a, b) XRD patterns of nanooxides; (c) changes in the SBET,X/
2BET,NS value as function of relative crystallinity
of nanooxides determined from the XRD data; and (d) Raman spectra of some oxides containing.
29
Fig. 30. UV/visible spectra (Jasco UV/vis spectrometer) of samples containing AST1 or A-300/AST1 initial, MCA 30
min in a microbreaker (label ‘MCA’), or high-pressure cryogels (label ‘cryo’).
In addition to rearrangement of the secondary particles during MCA or high-pressure cryogelation,
changes in the chemical structure of nanoparticles could also occur. Such chemical changes could
be seen in UV-vis spectra of treated samples in the 300-600 nm range (Fig. 30). High-pressure
cryogels with AST1 and MCA-treated AST1 or A-300/AST1 change color from white to beige of
different tints. The color centers correspond to defects formed in the lattices of titania and alumina
(e.g., oxygen vacancies) during MCA or high-pressure gelation. Note that MCA and HPCG gave
different effects in the 300-400 nm (or 4.1-3.1 eV) range. For example, A-300 damped the high-
pressure effects on AST1 during HPCG, but this influence of A-300 decreased during MCA (Fig.
30) [90]. There were patterns in how FMO properties, characteristics and behavior were affected
by different media, depending on their composition and history [53]. First, typically FMO
containing titania and alumina were characterized by smaller SBET values (i.e. larger nanoparticles)
and pore volumes, (Vp, Table S1) because aggregation of nanoparticles decreased with increasing
size. Contributions of nanopores at R < 1 nm (i.e. voids between adjacent nanoparticles in their
aggregates) decreased with decreasing value of SBET [53].
The infrared (IR) spectra of FMO with bound adsorbates give useful information on the
structure of the interfacial layer, especially for silica because of appropriate sensitivity of the O-H
stretching vibrations of surface silanols toward types and amounts of adsorbates [37,38,53]. It was
of interest to compare the FTIR spectra of FMO (nanosilica) and silica gel (e.g., 200DF with SBET
= 540 m2/g, Vp = 0.34 cm3/g) [104]. The FTIR spectra of nanosilicas A-300 and A-380 with
adsorbed water (from air) and water/methane co-adsorbed in the spectral cell (Fig. 31a,b) were
characterized by a broad band at 2500-3700 cm1 which could be assigned to the OH-stretching
vibrations of adsorbed water and disturbed surface silanols [38,53]. A narrow band at 3750-3747
cm1 linked to free silanols [38,105] was observed for nanosilicas (more readily observed for A-
300 than A-380) but not for silica 200DF (Fig. 31a). This result could be explained by clustering
of water adsorbed on nanosilicas (i.e., a continuous layer of adsorbed water was absent), in
contrast to porous silica 200DF whose narrow pores (Fig. 26a) were practically entirely filled by
water adsorbed from air.
Gaseous methane alone and adsorbed onto nanosilica was characterized by an intensive
narrow band at CH = 3016 cm1 and a set of low-intensive satellite bands split due to vibrational-
rotational motions of methane molecules in the vapor phase in the FTIR cell. A small quantity of
methane adsorbed on nanosilica was observed after purging the cell with air (Fig. 31b, curve 2)
[104]. The adsorption of methane did not lead to significant changes in the band intensity of free
silanols because of the weak interaction of nonpolar CH4 molecules with the silica surface [104].
However, calculations of the free surface area (SIR) on the basis of the integral intensity of free
30
silanols at OH = 3747 cm1 (as a surface characteristic) and a band at 1870 cm1 (as a bulk Si–O
combination mode used as an internal standard related to the nanoparticle volume), according to a
method described elsewhere [106], give SIR = 263.3 m2/g for A-300 alone (with adsorbed water
from air) and 266.9 and 265.9 m2/g for silica with methane adsorbed for 1 and 7 min, respectively.
An increase in the SIR value was due to partial desorption of water resulting from methane flow in
the cell. It should be noted that the value of SIR could be considered as a part of the surface area
free from adsorbates, since it was determined from the band of free silanols at 3747 cm1, and it
was close the value of SBET for dehydrated silicas [53,104,106]. According to the values of SIR and
SBET, adsorbed water (h 0.1 g/g) disturbed about 22% of surface silanols. According to the values
of SIR after exposure to methane for 1 and 7 min, methane disturbed about 0.4 % of surface
silanols free from adsorbed water. Comparing these results with the changes caused by known
amounts of water it is possible to estimate the amount of adsorbed methane that give 1.8 % CH4
with respect to adsorbed water (i.e. ~0.2 wt.% with respect to silica since h 0.1 g/g) at 288 K.
This value was smaller than that obtained from 1H NMR spectral measurements (discussed below),
giving a greater ratio of adsorbed methane and water as 0.116 : 1, but at a slightly lower
temperature (280 K) [104].
Fig. 31. FTIR spectra over the OH-stretching vibrations range for (a) silica A-300 with adsorbed water (from air)
(curve 1), after blowing by methane (2), methane alone (3), and silica 200DF with water adsorbed from air
(4); and (b) A-380 with co-adsorbed water and methane (1) and after blowing by air (2); (c) FTIR spectra (in
a transmission mode) of nanosilica with adsorbed DON (1,5-dioxinaphthalene) at CDON = 0 (1), 0.1 (2), 0.2
(3), 0.4 (4), 0.6 (5), 0.8 (6) 1.0 (7) and 1.5 mmol/g (8); (d) degree (X) of distortion of surface silanols (OH =
3748 cm1) as a function of the adsorbed amount of DON.
31
Nanosilica A-380 is more aggregated and has a larger textural porosity than A-300, since
the pore volume determined from the nitrogen adsorption isotherm is Vp = 0.943 and 0.714 cm3/g,
respectively [104]. Therefore, nanosilica A-380 could adsorb larger amounts of water from air.
This results in a much lower intensity of the band at 3747 cm1 (Fig. 31) and much smaller free
surface area SIR = 186 m2/g at SBET = 378 m2/g (50.8 % reduction) than that for A-300. Purging
this sample with methane gave SIR = 194 m2/g because of partial desorption of water, as seen with
A-300. Subsequent purging with air led to diminution of the SIR value to 183 m2/g, because of co-
adsorption of water and residual methane. Assuming that water displaced adsorbed methane, one
could estimate the value of free surface area occupied by methane as 1.8 %, which is higher than
that observed for A-300, due to stronger aggregation of silica nanoparticles of A-380. Notice that
these estimations were not exact due to features of the used calculation methods but clearly
showed tendency of the influence of pre-adsorbed water on the adsorption of methane and vice
versa. Silica 200DF included both micro- and narrow mesopores in contrast to nanosilicas
characterized by broader textural pores (Fig. 26). Therefore, silica 200DF could adsorb larger
amounts of water from air than nanosilicas. Thus, the confined space effects in pores of different
types, and the influence of pre-adsorbed water on the adsorption of methane, significantly differ
for nanosilicas and 200DF [104].
The FTIR spectra of nanosilica A-300 impregnated by 0.2-1.0 mmol/g of 1,5-
dioxinaphthalene (DON) from water-ethanol solution and dried (using tablets with a mixture with
KBr (1:5)) showed interaction of DON with a surface of nanosilica. It was estimated by changes in
the FTIR spectra related to the OH-stretching vibrations of free surface silanols (OH at 3748 cm1)
depending on the DON amounts (Fig. 31c). The IR bands of stretching and deformation vibrations
of bonds in adsorbed DON molecules as well as the OH vibrations of surface silanols could be
identified in the spectra. The intensity of the OH band of free silanols decreased with increasing
amounts of adsorbed DON (CDON) because of the hydrogen bonding of these molecules to the
silanols. A sharp increase in the degree of distortion of free silanols by the DON molecules was
observed at CDON < 0.5 mmol/g (Fig. 31d). Notice that the values of SIR for DON/A-300 were
relatively high (300-338 m2/g) that were in agreement with a relatively low degree of distortion of
surface silanols (Fig. 31d). This result could be explained by the adsorption of DON in the form of
clusters but not in the form of individual molecules. This is characteristic for polyaromatic
compounds because of great energy of dispersion interactions upon their clustered adsorption [53].
Suspending-drying of FMO as ‘soft’ materials led to rearrangement of secondary structures
[53,107]. AFM images of a dried solid residual of treated silica A-300 alone showed (Fig. 32) that
repulsive particle-particle interactions caused by negative charging of the silica surface at pH 12
resulted in a reduction of the size of aggregates of primary nanoparticles. However, changes in
silica concentration in aqueous suspension sonicated at native pH value (close to 5 with only slight
negative charging of the silica surface [53]) weakly affected the aggregate structure in the dried
solid residual (Fig. 32a,c). This result was confirmed by small variations in the structural
characteristics (pore volume Vp = 1.42-1.58 cm3/g, specific surface area SBET = 220-239 m2/g) and
pore size distributions (Fig. 33) of these systems. However, the Vp values differ quite a bit from the
initial A-300 powder, in contrast to the SBET values (Fig. 34), because primary particles were not
changed but their packing was changed. Similar results with respect to the morphology of
secondary FMO particles were observed for the dry silica powder and dried solid residual of
treated A-50, A-300, and A-380 [53]. Primary particles (30-100 nm in diameter) and their
relatively small aggregates (< 300 nm) were observed for silica A-50. Mainly aggregates (< 200
nm) were observed for A-300 (primary particles were 6-18 nm in diameter) and A-380 (5-15 nm
primary particles).
Rearrangement of aggregates of primary particles, due to sonication of aqueous
suspensions and drying of the solid residual, led to significant changes in the shape of PSD in
comparison with the initial dry A-300 powder (Fig. 33) [107]. The contributions of pores of
different sizes to the total porosity and specific surface area demonstrated a concentration-
dependent character (Fig. 34). The interfacial Gibbs free energy of sonicated or ball-milled
32
aqueous suspensions of A-300 alone has a maximum at CA-300 12 wt.% [53], as well as the
S/S(CA-300) and V/V(CA-300) functions for the dried residual (Fig. 34). This content of silica
corresponded to a critical value because the sonicated aqueous suspension could transform into a
gel-like state at this or higher values of CA-300 after 1-3 days (depending on CA-300). A continuous
cluster of silica particles formed at these concentrations. However, at lower concentrations (CA-300
< 10 wt.%) the aqueous suspension did not transform into the gel-like state even over a period of
several years. Changes in the PSD shape of the solid residual slightly depended on silica
concentration in the aqueous suspension (Fig. 33b), because particle packing rearrangement
depended on many-particle interactions which is a function of concentration in the suspension.
However, the position of the main one-two meso- and macropore peaks in the PSD did not depend
on silica content, but it depended on the silica type. This suggested that the rearrangement of
secondary particles resulting from the applied treatments was a function of PaSD. The process of
suspending-sonication-drying led to a significant increase in the bulk density of treated silica (0.41
g/cm3 for dried solid residual of the suspension at CA-300 = 16.7 wt.%) in comparison with the
initial powder (0.05-0.06 g/cm3), or the dry powder ball-milled for 24 h (0.3 g/cm3), or after
hydrothermal treatment of fumed silica A-300 (0.2 g/cm3) [53,89,107].
33
(a)
(b)
(c)
Fig. 32. AFM images of dried solid residual of aqueous suspension with A-300 (a) CA-300 = 1 wt.% (without NaOH),
(b) 1 wt.%, pH 12 (adjusted by NaOH), and (c) 20 wt.% (without NaOH) [53].
34
Drying of the solid residual from the aqueous suspension (16.7 wt.%) led to an
approximately 3-fold reduction of the size of the continuous cluster. Despite this effect, there was
no loss in the specific surface area (SBET was close to zero) for suspended-sonicated-dried
nanosilica (Fig. 34), in contrast to the dried solid residual of the ball-milled suspensions or ball-
milled dry powders (SBET was between 50 and 160 m2/g for these samples) [53,107].
Fig. 33. IPSDV of (a) initial dry A-300 powder (curve 1) and dried residual of aqueous suspensions of A-300 (1 wt.%)
(2), with addition of 0.9 wt.% of NaCl (3), pH 12 (4), and 20 wt.% + 0.9 wt.% of NaCl (5) with a model of
pores as voids between spherical particles; and (b) initial dry A-300 powder (curve 1), dried residual of
aqueous suspensions of A-300 at CA-300 = 7.4 (2), 9.9 (3), 12.3 (4), 14.5 (5) and 16.7 (6) wt.% with a model of
cylindrical pores.
Fig. 34. Relative changes in the specific surface area and porosity: (a) total, (b) nanopores, (c) mesopores, and (d)
macropores for suspended-sonicated A-300 and then dried as a function of concentration.
(Dimethylamino)azobenzene (DMAAB, pKa = 3.3) was chosen as a color indicator to study
the active site distributions on the nanooxide surfaces using the optical spectroscopy. The diffuse
reflection spectra of adsorbed DMAAB (Fig. 35) were recorded using a SF-18 (LOMO, St.-
Petersburg) spectrophotometer, and then converted to absorption spectra according to the Kubelka-
Munk formula.
35
Fig. 35. Optical spectra of DMAAB adsorbed onto different fumed oxides (sample weight was selected to provide
close surface area of all samples).
The DMAAB adsorption from the gas phase onto samples previously evacuated to 104 Torr, then
heated in special optical glass vessels at defined temperatures for 1 h and hermetically sealed, was
carried out at 3385 K for 2-4 h. The doses by weight of samples were different to maintain a
constant surface area in all measurements. The assignment of the DMAAB absorption bands was
done by analogy to the spectra for this substance in neutral and acidic solutions. Four absorption
bands of DMAAB adsorbed on the complex oxide surface can be detected: (a) = 430-470 nm,
physisorbed DMAAB through dispersion interaction, (b) 470-490 nm, hydrogen-bonded DMAAB,
(c) 490-545 nm, complexes with H+ transferring from the Brønsted (B) sites to DMAAB, and (d)
550-560 nm, DMAAB complexes with Lewis (L) acid sites.
Wetting-drying and cryogelation effects depend on FMO content in the treated suspensions
that is shown using fumed silica Cab-O-Sil HS-5 (Tables 4 and 5, Figs. 36-38). All these processes
result mainly in compaction of the secondary structures (NPNP aggregates and agglomerates of
aggregates) that affect the PSD. However, the specific surface area insignificantly changes with
one exception of a sample treated at pH 10 with NaOH.
Fig. 36. Pore size distributions calculated using NLDFT (cylindrical pores), MND (cylindrical pores) and SCV/SCR
methods for (a) initial silica Sab-O-Sil HS-5, and treated samples: (b) 301, (c) 301A, and (d) 302 (see Tables
4 and 5).
36
Table 4. Samples based on fumed silica Cab-O-Sil HS-5.
Sample
number
Processing parameter Sample
number Processing parameter
301 Normal 3 wt% slurry 406 stirred 48 hours
301 A Frozen 72 hours 407
Silica heated 650 C 12 hrs prior to slurry
making, freezing and freeze drying
302 3 wt% frozen 11 days 501 normal 5 wt %
401 Normal 4 wt% slurry 502 not frozen or freeze dried
402
sonicated 5 mins (~30 kHz in a
bath sonicator)
502
crushed
crushed up chunks before TGA
403 sonicated 25 mins 503 briefly frozen and freeze dried
404 pH 4.3, not adjusted 504 frozen 48 hrs and air dried
405 Adjusted pH 10 with NaOH 1001 10 wt%
Bare plain fumed silica
Table 5. Textural characteristics of silicas based on Cab-O-Sil HS-5.
Sample b
(g/cm3)
SBET
(m2/g)
Vp
(cm3/g)
Snano
(m2/g)
Smeso
(m2/g)
Smacro
(m2/g)
Vnano
(cm3/g)
Vmeso
(cm3/g)
Vmacro
(cm3/g)
Initial 0.045 272 0.689 66 198 8 0.031 0.535 0.123
301 0.294 284 1.677 77 183 24 0.036 1.291 0.350
301A 0.256 265 1.449 70 184 11 0.033 1.241 0.175
302 0.179 281 1.329 78 184 19 0.035 0.979 0.315
401 0.228 288 1.530 78 176 33 0.034 1.011 0.485
402 0.197 286 1.611 76 156 54 0.033 0.649 0.929
403 0.191 281 1.152 13 244 24 0.003 0.778 0.371
404 0.176 282 0.913 69 195 17 0.033 0.643 0.237
405 0.227 171 1.153 43 90 39 0.019 0.501 0.633
406 0.211 255 1.240 61 156 38 0.027 0.635 0.577
407 0.120 241 0.914 54 167 20 0.024 0.605 0.284
501 0.241 276 1.486 66 168 42 0.026 0.857 0.602
502 0.344 267 1.510 60 183 24 0.022 1.186 0.301
503 0.176 268 1.319 33 192 44 0.010 0.627 0.682
504 0.328 277 1.442 66 198 12 0.026 1.271 0.145
37
Fig. 37. Pore size distributions calculated using NLDFT (cylindrical pores), MND (cylindrical pores) and SCV/SCR
methods for treated samples: (a) 401, (b) 402, (c) 403, (d) 404, (e) 405, (f) 406, and (g) 407 (see Tables 4 and
5).
38
Fig. 38. Pore size distributions calculated using NLDFT (cylindrical pores), MND (cylindrical pores) and SCV/SCR
methods for treated samples: (a) 501, (b) 502, (c) 503, and (d) 504 (see Tables 4 and 5).
Described above structural features of FMO can affect adsorption and other interfacial
phenomena occurring in both gaseous and liquid dispersion media [53] that will be analyzed
below.
Interfacial phenomena
Individual adsorption of low-molecular weight compounds
To analyze the adsorptive characteristics of a variety of initial and treated FMO (Tables 1 and
6), the adsorption of water, n-hexane and other low-molecular weight adsorbates was studied using
an adsorption apparatus with a McBain–Bark quartz scale at 293 K [28,30,53]. Samples were
evacuated at 103 Torr and 473 K for several hours to a constant weight, then cooled to 2930.2 K,
and the adsorption of water, hexane and other adsorbates was studied at varied relative pressures
p/ps [28,30]. For all studied FMO [30], the nitrogen adsorption-desorption isotherms were
characterized by narrow hysteresis loops (Figs. 39-41) and could be assigned to the type II of
IUPAC classification [84,85]. This regularity was due to a spherical-like (but non-ideal) shape of
nonporous primary particles (average diameter dav = 6.9-52.5 nm for various FMO, Tables 1 and 6,
and Figs. 2-10) forming aggregates (secondary particles of 50-1000 nm in size) and looser
agglomerates of aggregates (ternary particles > 1 m in size). Voids in secondary and ternary
particles provided the textural porosity of the materials but with virtually no contribution from
nanopores (Fig. 42). Notice that the adsorption isotherms of various adsorbates (Figs. 39-41) could
also be assigned to the type II [84,85] because of textural features of FMO. The empty volume in
loose nanooxide powders was very great Vem = 1/b 1/0 = 10-30 cm3/g (where b = 0.03-0.10
g/cm3 is the bulk density of the dry powders).
Table 6. Structural characteristics of nanooxide powders [28,30,53].
Sample
2SiOC
(wt.%)
2TiOC
(wt.%)
2 3Al OC
(wt.%)
2
s
SiOC
(wt.%)
2
s
TiOC
(wt.%)
2 3
s
Al OC
(wt.%)
dav
(nm)
Δw
2BET,NS
(m2/g)
2p,NV
(m2/g)
SBET,ac
(m2/g)
SBET,hex
(m2/g)
SBET,DEA
(m2/g)
SBET,TEA
(m2/g)
SA8 92 8 29.5 70.5 8.6 0.33 303 0.68 223 207 271 265
SA23 77 23 73.6 26.4 6.9 0.23 347 0.82 362 247 434 384
ST20 80 20 75.6 24.4 27.9 0.40 84 0.18 116 80 97 115
AST03 2.7 0.3 97.0 13.9 0.30 125 0.31 162 101 114 126
AST1 10.0 1.0 89.0 18.0 0.33 99 0.25 180 99 113 117
AST50 28 50 22 58.7 40.0 1.3 47.9 0.31 37 0.08 44 31 44 42
AST71 8 71 21 5.0 80.4 4.6 21.6 0.43 74 0.13 83 63 60 70
AST82 6 82 12 0 91.2 8.8 40.2 0.27 39 0.11 43 48 46 58
A-50 >99.8 >99.8 52.5 0.38 52 0.13 42 77
A-200 >99.8 >99.8 11.9 0.31 230 0.48 232 187
A-300 >99.8 0 >99.8 9.2 0.17 295 0.52 261
Al2O3 >99.8 >99.8 12.9 0.16 133 0.45 246 187 205 238
Al2O3 >99.8 >99.8 19.3 0.35 89 0.13 95 90
TiO2 >99.8 >99.8 35.7 0.32 42 0.09 58 43
Note. Cs is the surface content of oxides, and Δw is relative error of the model of voids between spherical particles packed in random aggregates.
39
40
Fig. 39. Adsorption isotherms of (a, b) nitrogen, (c, d) hexane, (e, f) acetonitrile, (g, h) diethylamine, DEA, and (i, j)
trimethylamine, TEA (left - initial and right - normalized by SBET,N2) onto different oxides (1) AST03, (2)
AST1, (3) AST50, (4) AST71, (5) AST82, (6) SA8, (7) SA23 and (8) ST20 degassed at 200 oC.
41
Fig. 40. Comparison of adsorption of different adsorbates (1) nitrogen, (2) hexane, (3) acetonitrile, (4) DEA, (5) TEA,
and (6) water onto (a) AST03, (b) AST1, (c) AST50, (d) AST71, (e) AST82, (f) SA8, (g) SA23 and (h) ST20
degassed at 200 oC.
42
Fig. 41. Adsorption of different adsorbates onto (a) A-300 degassed at 200 oC; (b) fumed alumina (
2BET,NS = 133
m2/g) degassed at * 200, ** 600 and *** 900 oC; and adsorption of (c) hexane and (d) acetonitrile onto A-50,
A-200, alumina (
2BET,NS = 89 m2/g) and titania.
A major contribution of broad voids (macropores) (Fig. 42) in secondary (aggregates) and ternary
(agglomerates) particles, and thermodynamically unfavorable condensation of gases or vapors in
macropores [84,85] resulted in ineffectively filling of them by any gaseous or vapor adsorbates
even at p/p0 = 0.98-0.99. For instance, the maximal
2p,NV value among studied oxides was 0.82
cm3/g for SA23 (Tables 1 and 6) (possessing the maximal
2BET,NS value) which corresponded to
about 3% of the Vem value (b = 0.036 g/cm3). Another effect appearing due to the textural porosity
was that the nitrogen adsorption was proportional to the
2BET,NS value at p/p0 up to 0.5-0.6 (Fig.
39a,b) because of weak capillary condensation and insignificant volume filling of broad voids. The
2BET,NS values were calculated using the standard range of adsorption at 0.05 < p/p0 < 0.3 [84,85].
This did not give large errors in estimation of the
2BET,NS values because of similar results of
adsorption linearization in the BET coordinates for narrow and broad p/p0 ranges. For instance,
2BET,NS = 318 and 35 m2/g for SA23 and AST50 (with the largest and smallest
2BET,NS values) at
0.05 < p/p0 < 0.5, respectively, that were close to the
2BET,NS values calculated at 0.05 < p/p0 < 0.3
(Table 6). Additionally, the average area occupied by a N2 molecule could be varied (e.g., 0.162
nm2 for carbons and nitrogen molecules located parallel to the surface or 0.137 nm2 for silica due
to changes in the orientation of the N2 molecules around surface hydroxyls) [53].
43
Fig. 42. Pore size distributions for studied complex oxides calculated on the basis of (a) nitrogen desorption, and (b)
hexane adsorption with the model of voids between spherical particles.
Thus, despite a certain difference in the normalized (divided by
2BET,NS ) nitrogen adsorption
at p/p0 > 0.5 (Fig. 39b) caused by variations in the PSD and packing of primary particles in
secondary and ternary structures [30,53], the nanooxides morphology could be assumed as roughly
the same for all FMO studied. This was due to the flame synthesis of them at high temperatures T
> 1200 oC; i.e., primary particles could be in state close to liquid one in the flame [2-4] to form
nano-beads. However, the studied oxide powders strongly differed in the specific surface area
(Tables 1 and 6, SBET), PSD and the aggregation degree of particles (Vp) that, as well as the
difference in the nature of oxide surfaces, could affect the adsorption of any adsorbates. The
analysis of these aspects could be undertaken by comparison of the structural characteristics (S, V)
calculated from the adsorption isotherms of different adsorbates [30].
Greater differences in the adsorption on the same oxide were observed for more complex
and polar adsorbates (diethylamine (DEA), trimethylamine (TEA), CH3CN, H2O) than for nitrogen
(Figs. 39 and 40) [30]. These effects were caused by stronger bonding of stronger electron-donor
and polar adsorbates to Brønsted (bridging hydroxyls) and Lewis (incompletely O-coordinated Al
or Ti atoms appeared due to degassing at relatively high temperature) acid sites and other active
sites (e.g., terminal hydroxyls). However, there was a clear tendency toward a decrease in the
adsorption of all adsorbates with decreasing BETS value, despite the difference in their polarity,
electron-donor properties, and molecular size. Clearly, the specific surface area and textural
features of adsorbents could play an important role on the adsorption of various adsorbates. The
strong influence of the value of SBET on the adsorption of any adsorbates onto FMO could be
explained by the fact that a significant contribution to the adsorbed amounts was due to the first
monolayer [30,53] with insignificant capillary condensation of adsorbates in broad but short voids
between neighboring nanoparticles in their aggregates and agglomerates. Note that some isotherms
were recorded up to p/p0 = 0.62-0.75 due to certain difficulties in the measurements caused by the
behavior of the adsorbates studied at higher p/p0 values [30]. Additional effects were due to the
difference in the nature and content of active surface sites (terminal (MOH) and bridging
(M'O(H)M") hydroxyls and others) responsible for the formation of strong adsorption
complexes with electron-donor and proton-donor molecules [30,53].
The degassing temperature of FMO can play an important role on measurements of the
adsorption and, therefore, estimation of the structural characteristics from the adsorption
isotherms. However, morphological changes in fumed alumina or titania were much smaller upon
heating than that of silica [30,53]. This effect was due to the difference in the types of hydroxyls
44
(only terminal for silica and terminal and bridging for alumina and titania) and the O-coordinating
numbers of Si (only fourfold O-coordinated) and Al (from fourfold to sixfold) or Ti (five- and
sixfold O-coordinated) atoms. Therefore, associative desorption of water from a silica surface led
to stronger changes in the lattice structure than for other FMO [53].
Alumina (
2BET,NS = 133 m2/g) degassed at 600 or 900 oC and cooled to room temperature
(without contact to air) adsorbed greater amounts of acetonitrile and especially TEA than alumina
degassed at 200 oC (Fig. 41b). This result as well as the maximal SBET,X/
2BET,NS values (Fig. 43,
Table 6) could be caused by enhancement of donor-acceptor interactions between the N atoms of
adsorbates and surface Lewis acid sites (effectively formed upon strong dehydration of alumina at
high temperatures) and certain diminution of the size of primary particles on thermoevacuation
[30,53]. Additionally, calcination at 900 oC could increase the crystallinity degree of alumina that
could also influence the adsorption of electron-donor compounds. Notice that the first portion of
adsorbates (especially water) could dissociatively adsorb onto strained bonds appearing at the
FMO surfaces degassed at high temperatures [30,53].
Fig. 43. Relationships between the
2BET,NS and (a) SBET,X or (b) SBET,X/
2BET,NS values.
The degree of aggregation of the primary particles (affecting Vp), which depended on
treatment conditions, can play a certain role in the adsorption and changes in the
SBET,X/
2BET,NS values [30]. For example, for two samples of alumina, the ratio SBET,X/
2BET,NS was
greater for a sample with a larger value of Vp (Table 6) because the adsorption was greater for
FMO with more aggregated nanoparticles. Additionally, variations in orientation of adsorbed
molecules at a surface affected the surface area occupied by each molecule () and, therefore, the
SBET,X/
2BET,NS ratio [30,53].
A minimal adsorption onto FMO degassed at 200 oC was observed for water (Figs. 40 and
41), despite the fact that it could form strong hydrogen bonds with surface hydroxyls [30]. This
low adsorption was due to several reasons. First, saturated vapor of water has low pressure (17.5
mmHg at 293 K). Second, water adsorption was clustered rather than monolayered, e.g., the
surface was incompletely covered by water molecules even at Cw = 15-20 wt.% corresponding to
several statistical monolayers [30,53,84,85]. Third, nanopores were practically absent in FMO and
the contribution of narrow mesopores was low in the powders (Fig. 42). Among the adsorbates
studied [30] only water adsorbed in the form of clusters. Even during single molecule adsorption,
it tends to form two hydrogen bonds with neighboring surface hydroxyls. Therefore, nanopores
and narrow mesopores were more appropriate for effective clustered or nanodomain adsorption of
water than broad mesopores and macropores. In the latter, the formation of 3D structures totally
45
filling voids needed very large amounts of adsorbed water [53]. However, the adsorption of great
amounts of water in broad pores at p/p0 < 0.98 was thermodynamically unfavorable as well as for
other studied adsorbates [30,53]. All the factors mentioned could play a certain role on the
adsorption of organic adsorbates. However, thermodynamic conditions of their adsorption differ
from that of water. This led to greater adsorption of organic adsorbates onto FMO in comparison
with water (Figs. 40 and 41) at the same temperature (~20 oC).
Nonpolar hexane and weakly polar acetonitrile adsorbed onto FMO better than water but
worse than polar DEA and TEA [30]. This was due to strong hydrogen bonding of the amines to
the surface hydroxyls and Lewis acid sites. The adsorption of TEA was typically greater or close
to that of nitrogen (at the same p/p0 values), despite a large difference in the adsorption
temperatures (293-297 K and 77.4 K, respectively). Temperature determines the average kinetic
energy of molecules. Elevating temperature leads to diminution of the adsorption in voids between
nonporous spherical nanoparticles where the steric effects are absent. For AST71, the nitrogen
adsorption was much higher than that of TEA. This could be caused by dense packing of primary
AST71 particles since the N2 hysteresis loop was long. This loop started at p/p0 = 0.4 for AST71
and AST1, but for other samples, it began at p/p0 0.7. Consequently, capillary condensation of
nitrogen began on AST71 (or AST1) at lower pressures than for other FMO that enhanced the
adsorption [30].
In contrast to dispersion interactions of hexane and nitrogen, specific interactions of
acetonitrile and amines with surface active sites (mainly various hydroxyls) could cause an
overestimation of the SBET,X value in comparison with the values of
2BET,NS or SBET,hex (Fig. 43,
Table 6) [30]. This was due to several effects linked to features of orientation and packing of
molecules in the first adsorbed layer. Orientation of adsorbed molecules interacting with surface
hydroxyls, especially bridging ones, could corresponded to their non-maximal projection onto the
surface. Therefore, the effective value could be smaller than the maximal value estimated from
the molecular geometry of the adsorbates. This overestimation was maximal for acetonitrile. Its
adsorption complexes mainly corresponded to non-lengthwise orientation of the molecules that
caused a significant diminution of the effective value. However, the value for acetonitrile used
for estimation of SBET,ac (Table 6) corresponded to the average projection area of a molecule to a
surface [30].
The SBET,X overestimation for TEA was greater than that for DEA for FMO at
2BET,NS 125
m2/g, significant crystallinity and surface roughness (Table 6, ∆w) [30]. However, for SA8 and
SA23 with large values of
2BET,NS and amorphous silica and crystalline alumina phases, this effect
was greater for DEA (Fig. 43). These results could be caused by different changes in the packing
of adsorbed molecules of DEA and TEA in the first monolayer at the surface of smaller
nanoparticles more strongly aggregated in secondary ones because the steric factor was lower for
smaller DEA molecules than for TEA. This factor was greater for FMO with larger values of SBET
because they were characterized by enhanced aggregation of nanoparticles. For hexane, the
overestimation of SBET,hex was observed for A-50, AST82, and alumina degassed at 200 and 900oC,
respectively, that could be explained by, at least, two factors [30]. First, a low SBET value for both
AST82 and A-50 with large primary nanoparticles (Table 6, dav) weakly aggregated (Fig. 40g,
loop beginning at p/p0 > 0.8) provided better accessibility of the surface for relatively large C6H14
molecules in comparison with other oxides with larger SBET values and more strongly aggregated
nanoparticles (Fig. 42). Second, a significant surface content of titania in AST82 (Table 6)
enhanced dispersive interactions of hexane molecules with nanoparticles. For alumina degassed at
900 oC, there were additional factors such as diminution of sizes of nanoparticles (due to
dehydration and crystallization without sintering), increased crystallinity and dissociative
adsorption of a portion of the molecules [30].
An increase in the values of w (Table 6), i.e. enhancement of the surface roughness of
nanoparticles and deviation of their shape from spherical, corresponded to increasing
46
SBET,X/
2BET,NS ratio for, e.g., X = acetonitrile. This correlation for acetonitrile was clear for ten
FMO samples from fourteen ones studied. In other words, it was not strong correlation because
many structural and other effects overlap here [30].
The hydrophilicity of FMO was analyzed using calorimetry (oxides were degassed at 473
K at low pressures for several hours) and 1H NMR spectroscopy (oxides were equilibrated in air)
methods applied to samples after different pre-treatments [30,53]. This characteristic was linked to
the possibility of the formation of strong hydrogen and donor-acceptor bonds or/and dissociative
adsorption of water. The treatments before the calorimetric measurements resulted in desorption of
intact water and a portion of dissociatively adsorbed water (MOH, M'O(H)M", where M = Si,
Al or Ti) from both surface and volume of oxide nanoparticles [53]. However, for the NMR
measurements, surface and volume water was re-adsorbed from air. Therefore, one could expect
that the heat effects on the adsorption of water on the calorimetric measurements should be
stronger than that on the NMR measurements. This was typically observed for the samples studied
with the exception of SA8 and ST20 [30,53]. There was a tendency for an increase in the
difference between the Qw and S values with increasing size of primary particles. It was minimal
(and the hydrophilicity coefficient K = Qw/Qd < 1) for amorphous SA23 (dav = 6.9 nm, Δw = 0.23)
and maximal for AST50 (dav = 49.7 nm, Δw = 0.31, i.e., roughness AST50 > SA23) with
crystallinity > 90% (K = 1.83). However, this ratio depended on the particle composition, e.g.,
AST50 (crystalline) and A-50 (amorphous, Δw = 0.38) had close dav values but the reverse
relationships between the Qw and S values; however, their S values were close. These results
could be caused by stronger heat effects on the water adsorption onto degassed samples composed
of larger primary particles but depending on silica content and surface structure. Larger particles
could adsorb greater amounts of water in the volume because the V/S ratio increased (especially
for silica) as well as the ratio of volume and surface amounts of adsorbed water. Similar effects
observed for different silicas resulted in a significant overestimation of the
2BET,H OS value for
heated and degassed FMO and
2BET,H OS
2BET,NS for samples equilibrated in air [30,53]. Notice
that FMO studied with maximal
2BET,NS values were characterized by minimal hydrophilicity. The
Kh value, determined as the ratio of the heats of immersion in water and decane Kh = Qw/Qd, was
low (< 1) for SA23 [53], and this nanooxide had the maximal
2BET,NS value (i.e., smallest
nanoparticles) among the SA samples studied. Titania-containing samples (typically with small
2BET,NS values) had larger K values than samples with alumina/silica, alumina or silica [30,53].
To analyze the effects of specific (hydrogen bonding) and nonspecific (dispersive)
interactions of adsorbates with different surface sites, a variety of adsorption complexes was
calculated by the ab initio method with the 6-31G(d,p) basis set appropriate to study the hydrogen
bonding. For the adsorption of nonpolar hexane and nitrogen molecules, the interaction energy
(EHF) was much smaller than that for weakly polar acetonitrile and polar DEA, TEA, and water
molecules. This was due to the electrostatic interactions giving the main contribution to the EHF
value for the hydrogen bonds (Table 7, Eel), and this contribution was much larger than the
dispersive interactions. Therefore, for weakly polar acetonitrile, the EHF value was smaller than
that for more polar adsorbates (TEA, DEA, H2O). The formation of two intermolecular (hydrogen)
bonds between an adsorbed molecule and neighboring hydroxyls in the oxide clusters led to
greater EHF values up to 65-93 kJ/mol (Table 7). Notice that the DFT calculations (with the
B3LYP/6-31G(d,p), the HF/6-31G(d,p) geometry and consideration for the BSSE [30]) of the two-
hydrogen-bonding complexes of H2O and NH3 with the silica cluster give slightly smaller values
EDFT,BSSE = 41.6 and 52.2 kJ/mol, respectively, than EHF. On the formation of two bonds,
each of them could be weaker than an individual strong hydrogen bond. Diminution of the
wavenumber of the OH stretching vibrations (OH) (strongly linked to E) upon the hydrogen
bonding for individual bond OHN(H)(CH2CH3)3 for SiO(H)Al was larger (1444 cm1) by two
times than that for similar complex with TiO(H)Al (731 cm1). However, the EHF values were
47
56 and 93 kJ/mol, respectively. Notice that the calculated OH values could be underestimated
due to the cluster approach and harmonic approximation used [30]. According to the experimental
data OH = 45, 305 and 975 cm1 for hexane, acetonitrile and TEA, respectively, adsorbed onto
silica, but the calculations gave 4, 130, and 496 cm1, respectively. Therefore, the OH values
calculated could be used only for qualitative analysis of certain tendencies in interactions between
polar and nonpolar adsorbates with different surface sites of FMO [30,53].
Table 7. Interaction energy between Brønsted acid site =(HO)SiO(H)Al(OH)= and an adsorbate
molecule calculated using the Kitaura-Morokuma method (6-31G(d,p) basis set) [30,53].
Adsorbate EHF (kJ/mol) EBSSE (kJ/mol) Eel (kJ/mol)
H2O 69.6 56.5 108.9
NH3 75.5 65.2 121.0
CH3CN 40.5 36.0 52.5
N2 7.9 5.1 9.2
C6H14 4.7 1.7 1.7
Note. E is the changes in the total energy on the bonding, EBSSE is the E value with
consideration for the basis set superposition error (BSSE) and EHF > EBSSE, and Eel is the
electrostatic component of E.
Quantum chemical calculations of amorphous (cluster approach) and crystalline (periodic
boundary conditions) SA and ST oxides showed that crystalline systems have stronger Brønsted
acid sites [53], which could form stronger hydrogen bonds. There was a tendency, however, for a
non-monotonic increase in the ratio SBET,X/
2BET,NS with increasing crystallinity of FMO (Fig. 29c)
[30].
Thus, according to the calculation results, the interaction energy depended not only on the
type of surface sites but also on the structure of neighboring sites, especially for adsorbates tending
to form several intermolecular bonds (per a molecule) with surface sites [30]. Additionally, the
electron-donor properties of adsorbates can play an important role. Therefore, the maximal OH
values observed for TEA and DEA interacting with the Brønsted acid sites are in agreement with
higher interaction energy for these amines or NH3 (as their simple model) with the acid sites in
comparison with other adsorbates. This was also due to the maximal contribution of the
electrostatic component to the EHF values (Table 4) and higher polarizability of amines than other
studied molecules. The latter was due to location of the highest occupied molecular orbital
(HOMO) on the N atoms of amines (DEA, TEA) at 2-3 eV higher than the level of the HOMO
localized on the O atoms in water, the N atoms in the N2 or CH3CN molecules or the C and H
atoms in hexane [30,53].
Large values of EHF could provide enough energy for certain deformations (required
additional energy) of the adsorbed molecules interacting with neighboring active surface sites
(depending on the pore sizes, especially in nanopores, and surface coverage). This condition obeys
if the attractive interactions with the surface sites are stronger than the repulsive lateral interactions
between neighboring adsorbed molecules, which could be more compacted than in the free fluid or
vapor. Therefore, adsorbed molecules could occupy surface area smaller than average one
calculated for the free molecules. Additionally, orientation of adsorbed molecules, which are
strongly non-spherical, could be varied depending on structure of surface sites that leads to
different projection area of them onto the surface. Both mentioned factors lead to overestimation
of the SBET,X values in comparison with the
2BET,NS values for the same FMO [30,53].
Investigations of the adsorption of polar, weakly polar and nonpolar adsorbates (X) onto
individual and complex FMO showed that the ratio SBET,X/
2BET,NS = 0.68-1.85 depended on the
48
structural characteristics of adsorbents, the nature of active surface sites, conditions of degassing,
the polarity of adsorbates and their ability to form strong hydrogen (with Brønsted acid sites) or
donor-acceptor (with Lewis acid sites) bonds with surface sites [30,53]. A stronger overestimation
of the specific surface area SBET,X was found for complex oxides at
2BET,NS = 100-133 m2/g with
maximal crystallinity and high roughness of nanoparticles than that at
2BET,NS = 30-50 (with
weakly aggregated nanoparticles) or 200-300 m2/g with lower crystallinity (i.e. with lower
Brønsted acidity) but strongly aggregated. Additionally, changes in the SBET,X/
2BET,NS values
correlated to changes in the oxide hydrophilicity (i.e., strength and acidity of surface sites)
estimated from the heat of immersion of nanooxides in water and decane [53]. The amounts of
water bound to oxide nanoparticles in the aqueous suspensions were much larger than the pore
volume
2p,NV of the initial powders or dried solid residua of the suspension of the same materials.
This was due to rearrangements of secondary and ternary particles in the strongly wetted powders
or aqueous suspensions of FMO. Therefore, the rearrangement of secondary structures of FMO
under action of adsorbates or/and solvents could affect their adsorption values and energetic
characteristics as well as interactions of nanoparticles with their surroundings in complex liquid
and polymer media. The results obtained showed that not only the specific surface area, pore
volume, pore size distributions, and particle sizes but also surface composition and crystallinity of
the surface phases and the surface roughness and rearrangement of FMO nanoparticles affected the
adsorption of different adsorbates, as well as the adsorbed layer structure and the behavior of the
interfacial layer. This influence was maximal for adsorbates forming the strong hydrogen and
donor-acceptor bonds with different sites of FMO surfaces. The effects described are of
importance for the optimization of nanooxide materials for a range of applications including
adsorbents, fillers, and additives [53].
Treatment of FMO in aqueous media (and subsequent drying) more strongly affected the
adsorption of both polar (water, DEA) and nonpolar (hexane) adsorbates than MCA of the
powders (Figs. 44 and 45). This could be explained by a larger rearrangement of nanoparticles in
sonicated aqueous suspensions compared to the rearrangement of them in organic solvents
[47,53,90].
Fig. 44. Adsorption isotherms of diethylamine onto AST1 initial, cryogel (208 K for 12 h) and gel (24 h) dried and
degassed.
The adsorption of water (Fig. 45a) and benzene (Fig. 45b) was similar for MCA-treated mixture
with A-300/AST1. For relatively large molecules of adsorbates such as diethylamine (Fig. 45c),
lower adsorption occurred after HPCG. The presence of NaCl (during HPCG and then in treated
and dried samples) could reduce the compaction of aggregates, as well as silica can do in A-
300/AST1 [46,47]. After HPCG, the adsorption of diethylamine onto A-300/AST1/NaCl was
similar to that for suspended-dried A-300/AST1; however, the adsorption onto A-300/AST1
cryogel was much lower (Fig. 45c) [90]. The adsorption of nonpolar hexane (Fig. 45d)
49
demonstrated other features of the process on differently treated A-300/AST1 samples. The
adsorption of hexane was largely independent on treatment type at p/ps < 0.4, corresponding to
mainly monolayer coverage of the FMO surface. This was in contrast to diethylamine at p/ps < 0.4
where large adsorption differences were seen with treatment type. It was higher than that for initial
AST1. The open hysteresis loops for the hexane isotherms, especially for the suspended-dried
mixture with A-300/AST1 (Fig. 45d), could be explained by the formation of long curved pores
with narrow throats in compacted secondary structures of treated FMO. With this pore geometry
the desorption could be activated and, therefore, it was slow [90].
Fig. 45. Adsorption isotherms (at 20±1 oC) of (a) water, (b) benzene, (c) diethylamine, and (d) n-hexane onto AST1 or
A-300/AST1 differently treated.
Surface modification of ST could lead to substitution of not only surface silanols but also
Ti-OH groups [52,106]. It was believed that Ti-O-Si surface bonds were relatively easy to
hydrolyze, thus the surface modified ST20 was slurred in water for 30 minutes, filtered and dried
(Figs. 46 and 47). The concentration of surface C≡N was quantitatively analyzed before and after
exposure to water by monitoring the CN stretching band area ratioed to the 800 cm1 SiO2 band as
an internal standard. Prior to water exposure the band area ratio was 0.372 ± 0.012, and after water
exposure = 0.343 ± 0.020. A paired t-test indicates that these values were different at the 95%
confidence level, suggesting that some surface functionalized Ti-OH species were hydrolyzed
upon water exposure. After this water treatment, the material was exposed to UV irradiation for 1
hour and the infrared spectrum of the dried ST20 was reacquired, with no apparent changes. Thus,
the remaining surface SCN groups were stable to both hydrolysis and photocatalysis [108].
Although the differences in adsorption of methylene blue between the unmodified and
surface-modified ST20 samples was evident, the differences were not as large as that of pure silica
since unmodified ST20 has some strong adsorption sites that silica did not (Fig. 48) [108].
50
Fig. 46. Isotherms of methylene blue adsorption on
unmodified and modified FMO (amine - Cab-
O-Sil HS5 modified by 3-
aminopropyltrimethoxysilane, CN - 3-
cyanopropyltrimethoxysilane, SCN - 3-
thiocyanatopropyltriethoxysilane, thiol – 3-
mercaptopropyltrimethoxy silane, and HMDS -
hexamethyldisilazane).
Fig. 47. Distribution functions of Gibbs free energy of
methylene blue adsorption on unmodified and
modified FMO.
Fig. 48. Distribution functions of Gibbs free energy of methylene blue adsorption on unmodified and modified ST20
(error bar of the calculations is shown).
Complex silica/titania (ST20) more strongly affected bound water structure (Fig. 49), and
was characterized by a small activation energy of dc relaxation (Table 8, Ea,dc) [109]. Residual
sulfate groups affected the number and mobility of ions and, therefore, small Ea,dc values were
found for suspensions with PC-titanias (Table 8, Ea,dc), and the larger the content of sulfate groups,
the smaller was the Ea,dc value. Fumed ST20 has stronger Brønsted acid sites than titania alone;
therefore, Ea,dc,TiO2 > Ea,dc,ST20. The number of acid/base sites at a surface should not strongly
affected such redox reactions as photodecomposition of methylene blue, but could affect the
adsorption of MB and its electronic state, e.g., MB protonation. The MB molecules adsorbed on
the PC-500 surface reduced the interaction of the titania surface, with interfacial water [53,109]
displaced from the surface, since the TSDC of the frozen suspension of PC-500/MB was lower
than that of PC-500 at T < 170 K (Fig. 49c) [108,109].
51
Fig. 49. TSDC thermograms for frozen 3 wt.% aqueous suspensions of (a) PC titania (and water alone) and (b) fumed
oxides (thermograms were normalized to Fp = 1 MV/m).
Table 8. Structural and adsorption characteristics of nanooxides [109].
Sample SBET
(m2/g)
Vp
(cm3/g)
Him,w
(mJ/m2)
Him,d
(mJ/m2)
dXRD
(nm)
Ea,dc
(kJ/mol)
PC-100 89 0.358 453 157 18 (A)a 34.9
PC-105 78 0.356 456 147 20 (A)a 61.7
PC-500 207 0.375 423 137 9 (A)a 76.2
71.9 (MB)c
Fumed
TiO2
60 0.194 184 24 (A)a
30 (R)b
146.7
ST20 64 0.148 469 12 (A)a 45.4
AST1 99 0.253 548 - 126.4
Note. Crystallite size of a anatase and b rutile. The heat of immersion in water (Him,w) and decane (Him,d);
c 0.3 wt.%
MB was added to 3 wt.% PC-500 suspension.
However, polar MB molecules (both adsorbed on the titania surface and dissolved in the bulk
water) strongly contribute to dipolar relaxation at T > 170 K. Additionally, the activation energy of
dc relaxation decreased for PC-500/MB (Table 8, Ea,dc) because of the influence of MB changing
the water characteristics (Fig. 49). All the mentioned effects could play an important role on
photodecomposition rate since adsorbed MB molecules were most likely to be decomposed [109].
The enthalpy of phase transition of bound adsorbates (Tables 9 and 10, H) is smaller by
the modulus than that of bulk liquids (freezing) or solids (fusion) with one exception of n-decane
(40% longer than n-hexane) bound to initial A-300/AST1. The decrease in the |H| occurs due to
52
(i) small sizes of bound structures (clusters, nanodomains) located in voids between nanoparticles,
and (ii) certain disorder of bound liquids (being in amorphous state after freezing) which are
frozen at temperatures lower than the freezing point of bulk liquids [53]. Both factors cause
smaller values of energy which should be spent (for melting) or realized (upon freezing) per a
bound molecule of the adsorbate.
The values of the exotherms (cooling freezing) and endotherm (heating melting) on
the DSC thermograms depend on the amounts of adsorbates and adsorbents [53]. The adsorbate
amounts more strongly affect the exotherms/endotherms related to strongly (SBA) or weakly
bound adsorbates (WBA). For the majority of the studied samples, the amounts of adsorbates were
larger than that of adsorbents. Therefore, a significant portion of adsorbates was weakly bound that
gives the sharp exotherms of freezing and the main endotherms at temperatures close to melting
point of the bulk frozen compound. For better view of the effects of the types of adsorbates and
adsorbents, the DSC thermograms were normalized per mg of adsorbate with subtraction of the
baseline (Figs. 50-52). Relatively large amounts of adsorbates interacting with FMO lead to low
intensity of melting endotherms of SBA, which are located at T < Tm at T = T Tm > 10 oC, in
comparison with the melting endotherms of weakly bound adsorbates that is located at T close to
melting point at T < 10 oC [53].
For aromatics (benzene, toluene), the freezing point depression is stronger than that for n-
decane (Figs. 50-52). Melting delay (i.e. melting of frozen adsorbates at T > Tm) is observed for
both nonpolar and polar adsorbates. However, it is absent for DMSO bound to initial or MCA A-
300 (Fig. 50f). Note that DMSO possesses maximal donor number DN = 125 kJ/mol among
studied adsorbates (e.g. for water DN = 75 kJ/mol) [53]. DN is a quantitative measure of Lewis
basicity, i.e. the possibility to form strong hydrogen bonds with surface hydroxyls (Table 11).
Therefore, the DMSO molecules form the hydrogen bonds with surface hydroxyls stronger than
other studied adsorbates that strongly change the structure of interfacial DMSO layers in
comparison with bulk liquid (and frozen DMSO). However, the acceptor number (AN) is higher
for water than that for DMSO. Therefore, in contrast to other studied adsorbates, water tends to
form adsorbed clusters, in which water molecules act as both proton-acceptors and proton-donors,
and even at low content, water forms cyclic clusters around surface hydroxyls. For other studied
adsorbates, the values of DN are much lower than those for DMSO or water, e.g. DN = 0 for
chloroform. However, AN of chloroform is larger (by 11 kJ/mol) than that of DMSO. These
features of the electronic and molecular structures of adsorbates differently affect their behavior in
pores (voids between nanoparticles causing confiened space effects) such as freezing-melting point
depression (or delay) vs. pore sizes and structures of adsorbates and pore surfaces. The observed
depression is stronger for DMSO and water bound to MCA A-300 than that for initial A-300 (Figs.
50-52), because voids become narrower after MCA. This is due to the pure confined space effect,
because MCA does not practically change the surface structure of nanosilica in contrast to AST1
more strongly affected by MCA (due to destroy of large core-shell nanoparticles) till to change the
sample color. The HPCG more strongly affects the melting thermograms of adsorbates bound to
AST1 than the MCA of AST1 does (Fig. 51). In the A-300/AST1 blend, nanosilica plays a role of
a damper during HPCG. Therefore, the difference in the melting endotherms in comparison with
the initial powder can be smaller for this blend (Fig. 52) than that for AST1 initial and treated
alone (Fig. 51).
For certain samples, the freezing point depression corresponds to relatively small shifts of
the phase transition (melting) temperature at |T| < 10 oC (Tables 9 and 10, Figs. 50-52). This
range of the shifts of the melting points corresponds to WBA, since condition |T| > 10 oC
corresponds to SBA. Typically, larger structures of adsorbates (domains) located in broader voids
(pores) mainly correspond to WBA, but structures located in narrow voids (i.e. relatively small
clusters of adsorbates) or adsorption layers located close to the pore walls frequently represent
SBA. The position of very narrow exotherms related to freezing of WBA (due to excess of
adsorbates much larger than the pore volume Vp, Table 1) depends on the cooling rate (β), and the
greater the value of β, the lower is the freezing point. MCA results in the shifts of the PSD peaks
53
toward smaller pore sizes, and high-pressure or normal-pressure cryogelation leads to compacting
of the secondary particles. Therefore, these treatments can enhance the contributions of SBA.
These strong effects of the treatments on the texture of the powders can result in the appearance of
open hysteresis loops for even non-polar compounds adsorbed onto treated FMO (Fig. 45) [53].
This open shape of the isotherms can correspond to activated desorption or swelling of adsorbents
in liquid nitrogen.
Fig. 50. DSC thermograms (cooling (a, c, e) and heating (b, d, f)) of nonpolar adsorbates bound to initial or MCA
FMO: (a, b) benzene (melting point Tm = 5.53 oC, boiling point Tb = 80.1 oC), (c, d) toluene (Tm = −95 oC, Tb
= 111 oC), and (e, f) n-decane (Tm from −30.5 oC to −29.2 oC, Tb from 173.8 oC to 174.4 oC). Cooling/heating
rate is β = 10 oC/min for all samples with exception of toluene at β = 20 oC/min.
54
Fig. 51. DSC thermograms (cooling (a, c) and heating (b, c)) of polar adsorbates bound to initial or MCA FMO: (a, b)
water (melting point Tm = 0 oC, boiling point Tb = 100 oC), (c) chloroform (Tm = −63.5 oC, Tb = 61.15 oC), and
(d) DMSO (Tm = 18.5 oC, Tb = 189 oC). Cooling/heating rate is β = 10 oC/min for all samples with exception of
chloroform at β = 20 oC/min.
55
Fig. 52. DSC thermograms (cooling (a, c, e, f) and heating (b, d, e, f)) of (a, b, c) nonpolar and (d) polar adsorbates
bound to FMO high-pressure cryogels: (a, b) benzene, (c, d) toluene, (e) decane, and (f) water (curves for
initial FMO are also shown in (a)-(c)). Cooling/heating rate is β = 10 oC/min for all samples with exception of
toluene at β = 20 oC/min.
56
Table 9. Shifts T = T Tm (K) in peak temperatures and corresponding enthalpy H (J/g) in DSC thermograms.
Adsorbate
Water Benzene Toluene n-Decane
Sample Process
T H T H T H T H T H T H T H T H
A-300
initial
Cooling
Heating
6.2
4.3
244.3
233.6
4.4
204.0
-15.0
0.3
88.1
89.4
2.4
90.1
30.7
3.0
37.6
47.1
12.0
4.3
5.7
1.5
194.9
200.6
9.7
19.0
13.1
12.3
A-300 MCA
30 min
Cooling
Heating
11.7
4.3
151.0
155.7
6.5 108.7 17.2
4.6
11.8
10.0
13.0
10.6
2.0
14.9
12.6
137.4
2.4
15.0
88.5
12.4
Alumina
initial
Cooling
Heating
11.0
12.0
72.5
1.6
3.2
1.5
57.9
81.9
-16.7
0.3
6.4
2.7
6.1
139.0 7.6 3.9
AST1
initial
Cooling
Heating
14.5
1.5
149.9
140.0
3.3
262.5
15.7
22.3
27.1
0.9
1.9
19.1
22.0
1.9
29.4
35.1
13.5
7.3
5.0
1.4
165.6
163.4
10.4
15.8
18.9
18.8
8.1
3.4
AST1 MCA
30 min
Cooling
Heating
7.3
1.5
121.3
177.3
2.4
22.3
11.9
1.0
-32.0
33.4
24.0
1.9
24.1
26.8
11.5
14.1
8.0
2.1
103.5
124.8
13.8
16.1
15.5
15.7
AST1
cryo
Cooling
Heating
12.5
2.8
166.7
285.3
5.3
205.0
11.1
2.5
31.7
34.6
9.0
4.0
9.8
0.8
29.8
3.8
28.9
34.8
11.7
5.4
11.4
12.7
1.3
17.7
44.6
37.2
20.2
36.0
2.9
3.2
117.5
154.0
A-300/AST1
initial
Cooling
Heating
15.5
1.6
167.1
207.0
3.6 3.1 17.0
1.4
67.9
45.6
18.7
0.9
3.6
<1.0
12.9
<1.0
5.5
1.4
250.1
240.7
11.3
16.7
93.7
82.3
A-300/AST1
cryo
Cooling
Heating
9.9
1.3
249.9
275.9
3.3 249.0 12.3
1.5
70.9
94.9
6.0
1.5
36.3
-0.5
12.4
45.0
13.6
4.1
7.3
11.0
0.3
16.7
-48.1
38.1
41.6
34.9
3.5
3.2
75.6
59.0
Note. Enthalpy of fusion at freezing point is 333.5 (water), 126.3 (benzene), 71.8 (toluene), 73.7 (chloroform), 183.9 (DMSO), and 201.82 (n-decane) J/g. Values of H < 0 and H > 0 correspond to
exotherms and endotherms, respectively.
57
Table 10. Shifts T = T Tm (K) in peak temperatures and corresponding enthalpy H (J/g) in DSC thermograms.
Adsorbate
DMSO Chloroform
Sample Process
T H T H T H
A-300
initial
Cooling
Heating
16.3
9.9
74.1
110.5
9.4
25.1
-1.6
26.2
3.5
5.6
17.0
35.0
20.5
17.2
44.9
26.3
A-300 MCA
30 min
Cooling
Heating
19.8
13.0
51.6
82.2
31.6
16.3
6.0
8.8
22.5
44.3
Table 11. Energy of molecules of adsorbates in the gaseous and liquid state (methods B97XD/cc-pVDZ (gas phase)
and SMD/B97XD/cc-pVDZ (liquid phase)).
Compound Et,g (Ha) Et,l (Ha) Es
(kJ/mol)
ECDS
(kJ/mol)
Benzene 232.180344228 232.188434255 21.2 -15.1
Toluene 271.487909798 271.496489581 22.5 -15.8
n-Decane 394.239040751 394.248132818 23.9 -21.9
n-Hexane 237.020340834 237.026503668 16.2 -14.8
Chloroform 1419.28566524 1419.29329148 20.0 -12.6
DMSO 553.134547788 553.145496094 28.7 0.0
Water 76.3996333120 76.4109660870 29.8 6.1
Note. Et,g and Et,l are total energy of a molecule in the gas and liquid states, respectively. Es is the solvation energy.
ECDS is the non-electrostatic components of the solvation energy.
Evaporation of liquids and condensation of vapors or gases vs. temperature are important
interfacial phenomena occurring in nature, agriculture, industry, and medicine. These processes
can be affected by a type of liquid or vapor/gas, structure of liquids, surroundings (free bulk,
droplets or pore-confined liquids, characteristics of adsorbents, i.e. confined space effects),
presence of solutes (co-adsorbates), gas and liquid flow parameters, pressure, and they depend
exponentially on temperature as activated ones [53,101,110-124]. Confined space effects lead to
deceleration of evaporation, especially from long and narrow pores, despite changes in the surface
tension (as well as some other characteristics such as density, critical temperature, etc.) of liquids
located in pores. The stronger the intermolecular interactions between molecules of a liquid and
solutes or between them and active surface sites of pore walls, the slower is the evaporation,
especially from the depth of narrow pores [110-124]. These effects could enhance condensation of
vapors or gases upon adsorption in porous media [84]. Adsorbates could form layers [53,125], e.g.,
upon adsorption of plane molecules (e.g., benzene) in slit-shaped pores of activated carbons,
exfoliated graphite, graphenes, clays, etc. that lead to changes in their properties in comparison
with bulk liquids free of layered structures. Free bulk and bound waters form clustered structures,
which depend on confined space effects [53]. The properties of liquids (surface tension, boiling
and critical temperatures, density, diffusivity, mobility, etc., see Table S8 in SI) in strongly
confined space could significantly differ from those of bulk liquids [53,126,127]. These effects
become stronger in narrower pores depending on the pore wall nature (chemical structure) and
shape. However, in mesopores at radius Rp > 10t (where t is the adsorbate monolayer thickness),
the mentioned above changes in the properties of adsorbates could be relatively small [53,127].
Experimental techniques based on evaporation (such as thermogravimetry (TG),
temperature-programmed desorption (TPD) of adsorbates) [128-130] or condensation (used in
recording of adsorption/desorption isotherms of gases or vapors) [84,85] give information on
features of interactions between adsorbates and adsorbents depending on the nature of both ones,
the texture and morphology of adsorbents [53]. Polarity, features of intermolecular interactions,
molecular weight and size of adsorbates could strongly affect their temperature and interfacial
behaviors [53,84]. The evaporation rate and evaporation / accommodation coefficients give
58
important information for deep insight into the phenomena affected by confined space effects
[53,84,110-115,122-124].
Theoretical and experimental investigations of spreading of liquid droplets on solid
surfaces and their evaporation were discussed for pure liquids and solutions [131]. Evaporation of
complete wetting and partial wetting liquids into a nonsaturated vapor atmosphere were considered
in this work. The main attention was focused on partial wetting characterized by the hysteresis of
static contact angle. For complete wetting, spreading/evaporation proceeded in two stages that was
theoretically described with a good agreement with experimental data. For partial wetting, the
spreading/evaporation of a sessile droplet of pure liquid could occur through four subsequent
stages [131]. The initial one, spreading, could be relatively short (1–2 min) and, therefore,
evaporation could be neglected during this stage. During the initial stage, the contact angle reaches
the value of advancing contact angle and the radius of the droplet base reaches its maximum value.
The first stage of evaporation is characterized by the constant value of the radius of the droplet
base. The value of the contact angle during this stage decreases from static advancing angle to
static receding contact angle. During the second stage of evaporation, the contact angle remained
constant and equal to its receding value, while the radius of the droplet base decreased. At the third
stage of evaporation both the contact angle and the radius of the droplet base decrease until the
drop completely disappears. It has been shown theoretically and confirmed experimentally that
during the first and second stages of evaporation the volume of droplet to power 2/3 decreased
linearly with time [131]. The universal dependence of the contact angle during the first stage and
of the radius of the droplet base during the second stage on the reduced time has been derived
theoretically and confirmed experimentally. The theory developed for pure liquids is applicable to
nanofluids, where a good agreement with the available experimental data could be found.
However, for evaporation of surfactant solutions the process deviates from the theoretical
predictions for pure liquids at concentration below critical wetting concentration and it is in
agreement with the theoretical predictions at concentrations above it. Thus, the theoretical models
of evaporation of free and sessile droplets have been well developed [115,123,131].
For similar processes occurred for droplets in the gaseous phase or upon location of liquids
in pores [101], quantum chemical (QC) calculations related to evaporation of polar and nonpolar
liquids were carried out with density functional theory (DFT) methods using a hybrid functional
ωB97X-D [132-134] (labelled as wB97XD in Gaussian 09) and various basis sets (aug-cc-pVQZ,
aug-cc-pVTZ, cc-pVTZ, and cc-pVDZ) with the Gaussian 09 program suit [132]. The solvation
effects (for molecule alone or in adsorbed state on a silica cluster in a medium with this compound
or in the gas phase) were analyzed using the solvation model SMD [135] implemented in Gaussian
09. To compute the Gibbs free energy of solvation (subscript s), it was assumed that ΔGs = Gl
Gg, where Gl and Gg were the Gibbs free energies of a molecule free or bound to silica cluster in
the liquid (subscript l) and gas (g) media, respectively. The calculations were performed taking
into account zero-point and thermal corrections to the Gibbs free energy in the gas phase and for
solved molecules and silica clusters using the geometry optimized using HF/cc-pVDZ, ωB97X-
D/cc-pVDZ, ωB97X-D/cc-pVTZ or ωB97X-D/aug-cc-pVQZ (last basis was used only for water)
[132-136]. The solvation effects were analyzed for the silica clusters with 8 and 22 tetrahedrons
alone and with an adsorbed molecule (see SI). Note that functional ωB97X-D introduces empirical
damped atom-pairwise dispersion terms into the functional containing range-separated Hartree-
Fock exchange for better description of van-der-Waals (vdW) interactions [132-136]. Therefore,
this functional was selected to obtain more adequate results for the gas and liquid (SMD) phases.
For linear molecules of 1-butanol or n-hexane, four conformers free or bound to a silica cluster
were taken into account. The effect of conformerization of n-hexane and 1-butanol molecules,
resulting in conformation-dependent changes in the Gibbs free energy of the molecules in both
phases, and the evaporation/condensation processes depended on the population of conformer
states [138,139].
Silica surface was modelled by clusters with eight SiO4/2 tetrahedrons and eight hydroxyls
or 22 tetrahedrons with 16 OH. Interactions of molecules H2O, C6H6, n-C6H14, H3C(CH2)3OH,
59
(CH3)2CO, and CCl4 with the silica cluster were analyzed using ωB97X-D/cc-pVDZ with
complete optimization of the geometry using ωB97X-D/cc-pVDZ (smaller silica cluster) and
HF/cc-pVDZ (larger cluster) [101].
The effect of conformerization of molecules, resulting in conformation-dependent changes
in the Gibbs free energy of the molecules in the gas and liquid phases, was studied for 1-butanol
and n-hexane considering linear and strongly bent conformers (see SI). The geometry of free
conformers and their complexes with the small silica clusters was completely optimized using
ωB97X-D/cc-pVDZ. In complexes with the larger cluster, only linear conformers were used [101].
The Gibbs free energy of an ensemble of conformers could be determined by formula [137]
B
1
ln exp( / )
N
N j
j
G RT G k T
, (14)
N is the number of conformers, R is the universal gas constant, kB is the Boltzmann constant, and T
is the temperature in the gas or liquid phase. To estimate the average changes in the Gibbs free
energy upon evaporation of a molecule, a formula could be used [101,138,139]
l g ,g ,l( ) /N NG G G N , (15)
where subscripts l and g corresponded to molecules in the liquid and gas phase, respectively.
Besides the conformerization of n-hexane and 1-butanol, this approach was used to study
interaction of the molecules with the silica clusters. The solvation effects were computed for the
corresponding complexes of the molecules with the silica clusters. In this case Eq. (15) was
corrected
l g ,g ,l cl,g cl,l( ) ( ) / 2N NG G G G G N , (16)
where Gcl,g and Gcl,l were the Gibbs free energies of the pure silica cluster in the gas and liquid
phases, respectively. Thus, several processes should be considered such as evaporation of
molecules (i) adsorbed onto pore walls at the gas/solid interface, (ii) located at the gas/liquid
interface, (iii) adsorbed onto pore wall in liquid/solid interface and transferred toward gas/liquid of
gas/solid interfaces [101,138-144].
For simplicity, one could assume that the Gibbs free energy of evaporation Gev ΔGlg
[139,140]. The evaporation enthalpy (Qev(T) > 0) contributing to Gev(T) could be estimated for
liquids as [143]
ev c( ) (1 / )mQ T q T T , (17)
where q and m are the constants, and Tc is the critical temperature, to estimate changes in Gev as
[139,140]
ev l g 0 ev ev 0( ) ( ) ( ) / ( )G T G T Q T Q T , (18)
where ΔGlg = ΔGs = Gg Gl, T0 = 298.15 K. The estimation of Gev(T) based on Eq. (18) is
more reliable than direct QC calculation of Gev(T) at high temperatures [101,144] using ωB97X-
D/cc-pVTZ and SMD/ωB97X-D/cc-pVTZ. Eq. (18) could overestimate a decrease in Gev with
temperature due to different changes in the entropy with temperature for different conformers in
the gas and liquid phases.
The equation for the averaged (by states of N conformers) evaporation rate ( )i i jk [101,144]
of the ith-molecule (which could be in a different conformer state) from the jth cluster or nano-,
micro-droplet (hereafter referred to as droplets), taking into account conformerization effects (for
n-hexane and 1-butanol), could be re-written in the form [138,139]:
( )
B 0
exp
g l
i i j ij
Gp
k b
k Tn RT
, (19)
where bij is the collision rate of the ith molecule with the jth droplet calculated using the kinetic
gas theory [101,138-144], n0 is the initial number of molecules in a droplet (here 4 or 20 nm in
diameter), p is the reference pressure (which could be estimated according to experimental
60
conditions). The values of ΔGgl and ΔGlg for adsorbed liquids (Eq. (20)) were corrected taking
into account changes in the interactions of molecule-molecule and molecule-silica surface. For
simplicity, the Gibbs free energy of a microdroplet with a very large n0 was assumed to be not
affected by a removal of a single molecule. The conformerization was assumed to take place not
only in a certain phase but also during possible changes in conformer states when molecules
transfer from liquid to gas phase and vice versa. This effect depends on the population of states of
various conformers in both phases and could affect the values of and k vs. temperature and
pressure. Clearly, the evaporation rate for liquids from pores differs from the rate of free liquids.
To account the effect of pores, modified equation described in detail elsewhere [101,142] was used
2
0 2
4 2 ( ) ( )cos
( ) ( ) expp p m
p g p
R T V T
k T k T
L R TR
, (20)
where k0 is the evaporation rate for free liquid with corrections of the Gibbs free energy in respect
to interactions of adsorbates with a silica surface, Vm is the molar volume, Rp is the pore size, Lp is
the pore length, is the contact angle (assuming cos = 1), and p(T) is the surface tension for
liquid located in pores determined using equation for bulk liquid [143]
b ( ) (1 / )mT A T B , (21)
where A, B, and m were constants, and in pores p(T) = b(T)f(Rp), where f(Rp) is the correction
function caused by confined space effects (analyzed only for water). The form-factor
2
2
p
p
R
L
in Eq.
(20) corresponded to the ratio of the area of pore section and the whole bead surface. To determine
k0 in Eq. (20), Eq. (19) could be used for evaporation of free droplets [101,144] of sizes
corresponding to the average diameter of pores with corrected Gibbs free energy due to the
interactions of a molecule with the silica surface (modelled by a cluster). For simplicity, one could
suggest that Lp = d/4 = 75 m, where d = 0.3 mm is the average diameter of silica gel beads. The
value of Rp in Eq. (20) could be calculated as the ratio of the first and zero moments of the
distribution function of pore sizes [53,145].
The confined space effects are very strong in narrow pores (nanopores at Rp < 1 nm and
narrow mesopores at Rp < 2 nm) [46,47,53,125-127,146-149]. These effects could be ignored for
mesoporous silica gels Si-60 and Si-100 and fumed silica A-300 characterized by small
contribution of pores in the mentioned range. A portion of pores in Si-40 is narrower than Rp = 2
nm (Fig. 26); however, contribution of nanopores at Rp < 1 nm is small. Consequently, the errors
of developed approach could be significant for the evaporation from narrow pores of Si-40 in
contrast to other silicas studied. This aspect was analyzed using appropriate experimental data
[101].
Large models of narrow mesopore of silica gel filled by water or n-hexane or two
nanoparticles surrounded by a shell with water molecules were calculated using PM7 method
(MOPAC 2016) [150-152]. For visualization of the fields around molecules and small silica
clusters, the TorchLite program was used [153,154]. Visualization of molecular structures was
performed with the help of the GaussView [155], ChemCraft [156] or UCSF Chimera [157]
programs.
The evaporation of all studied compounds from narrower pores of Si-40 was slower under
ambient conditions than from Si-60, Si-100 or A-300 (Fig. 53) [101]. However, a smaller volume
of pores of Si-40 could provide shorter time of evaporation of adsorbed compounds under heating.
A larger value of SBET of Si-40 could give the opposite result due to higher interaction energy of
adsorbates with a silica surface than that of molecule-molecule interactions [53,101].
61
Fig. 53. Evaporation as a ratio of content of residual liquid (mt) to initial one (m0) of (a) hexane, (c) water, and (e) 1-
butanol, difference between evaporation of liquid alone and in the presence of silicas for (b) hexane (0.1 g per
0.1 g of dry silica), and (d) water (1 g per 0.3 g of dry silica), (f) relative amounts of evaporated liquids alone or
in the presence of Si-40 as a function of time.
However, in narrow pores, the surface tension of pore-confined liquid drops down that could
positively affect the evaporation. In the case of an increase in the boiling temperature of adsorbate,
the difference in the evaporation (at room temperature) from silicas with various pore sizes
decreased from hexane (Fig. 53a,b), water (Fig. 53c,d) to butanol (Fig. 53e) that was also
accompanied by increasing time of evaporation for this series (Fig. 53f). Note that the evaporation
of water from Si-100 could be faster (especially from the layers located close to the surface of
silica beads) than that from textural pores (voids) of A-300 (Fig. 53c,d). However, for hexane, the
opposite effect was observed (Fig. 53a,b). This difference in the behavior of hexane (weakly
interacting with a silica surface) and water (strongly interacting with a silica surface) in respect to
Si-100 and A-300 could be explained by the difference in the particulate morphology and texture
of these silicas. In contrast to total values of the interaction energies, a relative difference in the
interaction energies of water-water molecules and water-silica was smaller than those for hexane
[101].
The TG study of the evaporation of liquids as a function of temperature showed a role of
structural, textural and morphological characteristics of adsorbents, as well as the adsorbate
properties [101]. The effects of narrow pores (i.e. confined space effects) were stronger than that
caused by the differences in the interaction energies of adsorbates with a silica surface because the
evaporation of nonpolar (hexane, benzene, CCl4) and polar (acetone, butanol, water) compounds
occurred from Si-40 (possessing narrower pores) at higher temperatures than from Si-100
62
(possessing broader pores) at the same fraction sizes of beads (0.2-0.4 mm). This difference tends
to be larger for liquids at lower boiling point. In contrast to water and 1-butanol, acetone molecule
did not have proton-donor groups. Therefore, intermolecular bonds (e.g., >C=OHC) in liquid
acetone were very weak in comparison with the bonds >C=OHOSi between acetone
molecules and the silica surface. This effect and confined space effects caused the maximal
difference in the evaporation vs. T for acetone desorbed from Si-40 and Si-100 (Figs. 37 and S56).
Note that the effect of diminution of the value of the surface tension () in narrow pores gives a
small increase in the rate of evaporation from depth of long pores of Si-40 and it could be ignored
for other silicas. The effect of an increase in the evaporation rate could be much stronger for
liquids located around the entrances of narrow pores close to the surface of silica beads (Figs. 53
and 54), because the value of Lp (used in Eq. (20)) strongly decreased there [101].
Fig. 54. (a, b) Evaporation rate vs. temperature for various compounds evaporated (a) in the gas phase (droplet of 20
nm in diameter) and from the depth of pores of Si-40 and (b) Si-100. (c, d) Evaporation time vs. temperature
for different compounds evaporated from pores of (c) Si-40 (assuming dp = 4 nm, Lp = 75 m) and (d) Si-100
(assuming dp = 20 nm, Lp = 75 m) and from a droplet of 20 nm in diameter located in the gas phase.
Excess of liquids (i.e. the amounts of liquids were larger than the pore volume of the
samples used) in the TG measurements caused a significant step in the thermograms at the boiling
point of the liquids studied [101]. A part of the TG curves related to the evaporation of liquids at T
> Tb after evaporation of bulk liquids (located out of pores) characterizes the evaporation from
pore volume and pore walls that corresponded to different G values. The confined space effects
for all liquids studied were more clearly observed at T > Tb in the TG curves drawn vs. ΔT = T
Tb. First, the evaporation from narrower pores of Si-40 was completed at higher temperatures but
in a much narrower temperature range (429-467 K) than those for Si-100 having broader pores
(Figs. 26). For the latter, the temperature of completed evaporation for liquids studied was from
347.4 K (acetone) to 468.1 K (butanol). The gap between the point of completed evaporation (Tev)
and the boiling point (which depended on both molecular weight of a compound and strength of
intermolecular forces) ΔTev = Tev Tb was much greater for Si-40 (up to ΔTev = 118 K for CCl4).
This gap decreased for liquids having a higher value of Tb and evaporated from Si-40. It was
minimal for butanol ΔTev 55 K that has maximal Tb = 390.81 K among studied liquids. For Si-
100, minimal ΔTev 18 K was for acetone, and maximal ΔTev 77 K was for butanol [101].
63
The order of the values of Tb and Tc differs for studied liquids: Tb is higher for butanol than
that of water but the value of Tc is higher for water. Therefore, the evaporation rate curves vs. T for
water and butanol intersect at temperatures between the points of Tb and Tc for both free droplets
and liquids evaporated from silica gels (Fig. 54). This effect could be explained by the fact that all
atoms in water molecules could take part in the formation of strong hydrogen bonds in contrast to
butanol molecules, in which only two atoms (from the OH group) from 15 ones could do that
[101].
According to the TG data and theoretical calculations (Fig. 54) [101], the temperature
ranges of evaporation of liquids from silica gels and the value orders of k and correlated to the
values of Tb. This is due to the fact that the value of Tb (as well as evaporation features) depends
on (i) the strength of intermolecular bonds that depends on the type of intermolecular bonds
(hydrogen, polar, nonpolar) and confined space effects for adsorbed compounds, polarity and
polarizability of molecules, that determine in total, e.g., the surface tension; (ii) molecular weight
and intermolecular bonds affecting molecular mobility, diffusivity, and viscosity of liquids. Of
course, for the evaporation from porous media, the textural characteristics of the adsorbents play
an important role. Therefore, there is a difference in the evaporation of liquids bound to silica gels
and nanosilica, which are characterized by very different PSD (Fig. 26). The evaporation of a
droplet with size corresponding to size of pores of Si-100 occurred much faster at the rate greater
by several orders of magnitude than that observed for Si-100 upon the evaporation from the depth
of pores (Fig. 54a,b). The evaporation time vs. temperature τ(T) ~ 1/k(T) (Fig. 54c,d) is
characterized by the opposite shape and orders than values of k vs. T (Fig. 54a,b) for liquids
evaporated from silica gels and free droplets. However, the function τ(T) demonstrates the curve
shapes similar to those of the evaporation rate (Fig. 54) [101].
The porosity type of silica gel Si-100 with long cylinder-like mesopores of 4-20 nm in
radius (see Fig. 26) located in millimeter-sized beads (assumed Lp = 75 m) and fumed silica with
textural pores as short voids between spherical nonporous nanoparticles (~ 8.3 nm in diameter)
forming aggregates < 1 m in size and agglomerates of aggregates > 1 m (assumed Lp = 1 m)
strongly affected the evaporation rate (Fig. 55a) [101]. Fumed silica A-300 has a close value of
SBET (330 m2/g) to that of Si-100 (314 m2/g). However, these silicas have very different PSD
affected by suspending-drying for A-300 only (Fig. 26). The curve of the water evaporation rate
from A-300 (from the depth of textural pores) is located between the curves for water evaporated
from Si-100 (from the depth of long pores) and a free drop (from open surface of drop) (Fig. 55a).
Note that the structures of a water layer bound to nonporous silica nanoparticles or a water cluster
located in narrow silica mesopore, as well as the structures of n-hexane clusters free or bound in
narrow mesopore, significantly differ [101]. The structure of the water layer bound to the outer
surface of silica nanoparticles could be less dense than the water cluster located in narrow
mesopore. This is in agreement with 1H NMR spectra of water bound in silica gel and fumed silica
[53]. However, the n-hexane cluster in narrow mesopore is less dense than the free n-hexane
cluster. These differences in the organization of interfacial water and n-hexane are due to the
difference in intermolecular interactions between molecules per se and with a silica surface [101],
as well as molecular and pore sizes. Similar differences in confined space effects were observed
using 1H NMR and calorimetry methods for water, decanol and decane bound to Si-40, Si-60 and
Si-100 [158] and nanosilica [53].
64
Fig. 55. Evaporation rate vs. temperature for water evaporated from (a) a droplet (20 nm in diameter) in the gas phase
and bound in the depth of textural pores (voids between nanoparticles in their aggregates) of fumed silica A-
300 (pore diameter dp = 10 nm, pore length Lp = 1 µm in aggregates of nanoparticles) or in the depth of long
mesopores of silica gel Si-100 (dp = 20 nm, Lp = 75 µm), and (b) the effect of a decrease in the pore length
and surface tension in pores on evaporation of water from short pores (Lp = 4 nm, p(T)) and depth of long
pores (Lp = 4 nm, b(T)). Inserts showed a small portion (over 370-375 K) of evaporation curves calculated
using the surface tensions of free bulk (b) and pore-confined (p) waters.
The confined space effects could lead to changes in the surface tension of liquids [53]. This
effect was analyzed for water bound in pores of Si-40 (having the narrowest pores among silica
gels studied, Fig. 26) at dp = 4 nm (31% of water from the first monolayer affected by interactions
with pore walls and 69% of water distant from the walls and having Gs close to that of bulk
water) [101]. An increase in the evaporation of water from the depth of long pores due to
diminution of the value of was small (Fig. 55b, bottom insert) and could be ignored. The
evaporation of water located around the entrance of the same pore at dp = 4 nm and Lp = 4 nm was
much faster than from long pore at Lp = 75 µm (Fig. 55b). The use of p(T) instead of b(T) gives a
certain (relatively small) increase in the evaporation (Fig. 55b, upper insert) [101].
Thus, to calculate the average value of the evaporation rate vs. temperature several averaging
procedures should be applied [101]: (i) by the pore length (since the evaporation of adsorbate from
the entrance of pore is much faster than from the depth of the same pore); (ii) by changes in the
surface tension p(T) in pores of different sizes (however, it could be ignored at Rp > 10t = 3.2 nm
(for water) and in the depth of long narrow pores, and this effect grows in absolute values (Fig.
55b, inserts) at the entrance of narrow pores but its relative contribution remained small); (iii) by
changes in the density of liquid layered in narrow pores at the presence of specific adsorption sites
[159] (changes in the values of liq and gas vs. temperature were taken in to account); (iv) by
effects of conformerization of long linear molecules; and (v) by changes in the interaction energy
with neighboring molecules (in bulk and in pores) and pore walls. Note that the first, fourth, and
fifth effects are much stronger than others. All these effects result in very fast evaporation of
liquids such as hexane from the outer surface of silica gel beads and the entrance of pores (in
shallow depth). However, in full depth of narrow pores, liquids could remain during a long period
at relatively low temperatures (Figs. 53-55). The evaporation rate is nearly exponential function of
temperature (Fig. 54, linear portions in log-scale of the Y axis). Therefore, even narrow pores
become empty with increasing temperature, especially at T > Tb [101]. However, this process is
not very rapid even at T > Tb. For instance, acetone was evaporated from Si-100 for approximately
4 min at T > Tb [101], but it was evaporated from Si-40 for 34 min. Butanol was evaporated from
both silicas for 25-30 min at T > Tb similar to water. If the results shown in Fig. 38b are compared
to the experimental data on the evaporation of water from Si-40 that one could conclude that the
experimental time of evaporation falls approximately midway between the theoretical curves
related to the evaporation from the entrance and full depth of pores. This effect could be explained
by the diffusion of liquid from the depth toward the entrance of pores during a whole evaporation
process. Therefore, the process could occur faster than for evaporation only from the depth of
65
pores, since the liquid can continually move toward the outer surface of the beads upon heating
[101].
It should be noted that the values of evaporation/condensation coefficients V and g
(assuming quasi-equilibrium state for gas and liquid states for free or pore-confined adsorbates at a
certain temperature) [138,139,144] were typically greater for liquids evaporated from silica gels
than that from free droplets with one exception for water. The values of V and g (related to the
probability of condensation of molecules into droplets from the gas phase) relatively increase (at
high temperatures) for compounds with a greater value of Tc in a series from acetone to water
[101].
Quantum chemical calculations of interactions of adsorbate molecules with a small silica
cluster with counterpoise corrections accounting the basis set superposition error (BSSE), as well
as the solvation effects for the molecules (ΔGs), cluster (ΔGs,c) and cluster with adsorbed molecule
(ΔGs,m+c), showed that the values of the mentioned energies were smaller for nonpolar adsorbates
(hexane, carbon tetrachloride, benzene) than those for polar adsorbates (water, acetone, butanol)
[101]. This was due to the formation of strong hydrogen bonds by the latter, i.e. due to strong
electrostatic interactions between polar molecules and silica surface. Features of interactions of
water (dense coverage of the surface) and hexane (distant location of molecules from the surface)
with silica were well seen from the structure of liquids in pores. The effect of stronger interactions
of polar compounds with the silica surface in parallel to the effect of location of adsorbates in
narrow pores of silica gels led to stronger deceleration of desorption of the polar molecules in
comparison with nonpolar ones (Figs. 53-55). However, for narrow pores of Si-40, the effect of
location of molecules in the depth of narrow pores was much stronger than the effect of stronger
interactions of polar molecules with the silica surface at the entrance of pores (Figs. 53-55). This
influences about 85-95 wt.% of adsorbates located in pores. For Si-100 possessing broader pores
than Si-40 (Fig. 26), the influence of the silica surface on deceleration of desorption deals with
only 15-20 wt.% of adsorbates located in pores [101]. These changes in the Gibbs free energy due
to changes in the molecule location were taken into account on calculations of the evaporation rate
from the depth of pores (Fig. 54). These differences could be affected by the differences in the
values of both average pore diameter and specific surface area of silicas (i.e. the amounts of
molecules located in the layer contacted to the silica surface), as well as the evaporation from
depth of pores. For a small content of polar active adsorption sites (e.g., carbons and silicalite),
density of liquids in pores could be lower than that for bulk liquid, and air/vapor bubbles could be
located in pores. This could inhibit the adsorption and accelerate the evaporation from pores, but
this process was not analyzed in [101].
Note that the conformerization of molecules in the liquid and gas phases results in the
appearance of additional channels of transfer of molecules evaporated from liquid to gas phase
with decreased positive changes in the Gibbs free energy [101,144]. The values of k(T) calculated
for hexane without taking into account the conformerization could be smaller than that calculated
with consideration of conformerization in bulk and bound liquids and in the gas phase (a smaller
difference was at higher values of T) upon evaporation from silica gel [101]. In general, there were
channels of the evaporation of different conformers both with smaller and larger values of Gev.
However, the effects of the former could be greater if transfers through the channels with high
values of Gev were not practically occurred. Therefore, the full effect of the conformerization
could result in increasing evaporation, which, however, depend on populations of states of
different conformers [101,144]. If the population of more stable conformers is greater in the liquid
phase than in the gas phase that the conformerization could diminish the evaporation [101].
The results of the SMD calculations for molecules confined in narrow pores could be
questionable due to some reasons [101,138,139]. For example, the structure of a solvation shell
differs from that for bulk liquids, and these differences increase with decreasing size of pores. The
interaction energy of studied molecules with a silica surface was typically larger (i.e. ΔG was
lower) than that with neighboring molecules. This was taking into account (with the corresponding
weighting coefficients) upon calculation of the evaporation rate and condensation/evaporation
66
coefficients by dividing of adsorbed compounds into three types: free bulk, bulk in pores and first
monolayer of molecules bound to a silica surface using the values of SBET (amounts of bound
liquids in the first monolayer were estimated as tSBET, where t was the monolayer thickness), and
pore volume Vp (amounts of bulk liquids in pores estimated from the difference Vp tSBET). As a
whole, for similar corrections, the experimental adsorption isotherms of studied compounds or
NMR data could be used to determine the Gibbs free energy of adsorption [46,47,53,101,149].
Fig. 56. Evaporation of (a, b) hexane (at room temperature) from (a) wetted samples at amounts of hexane of 0.06-
0.07 g per 0.02 g of FMO, and (b) suspensions containing FMO initial or treated (1 g of hexane per 0.05 g of
FMO). To reduce the errors, only a final portion of evaporation of (a) 0.02 g or (b) 0.05 g of hexane is shown
using the point at residual 0.02 or 0.05 g of hexane as t – tx = 0 min at x = 0.02 g (t0.02 ~ 70 min) or 0.05 g
(t0.05 ~ 450 min), respectively. Data shown in (a) were normalized by dividing by 0.02 g. (c) Evaporation of
n-decane (1 g per 0.05 g of FMO).
For lower amounts of hexane (Fig. 56a), the rearrangement of secondary structures during
wetting and evaporation was much smaller than for suspension (Fig. 56b) [90,101]. The difference
in the evaporation of residual amounts (0.02 g) of hexane from different FMO was much smaller
for lower amounts of hexane than evaporation of residual amounts (0.05 g) of hexane from
suspended FMO. A relatively weak interaction of hexane with FMO results in slightly faster
67
evaporation of hexane from FMO than from pure hexane; however, this difference was very small
(Fig. 56a). For suspended FMO, the evaporation of bound hexane (Fig. 56, curves 1-7) was slower
than that of individual hexane (Fig. 56b, curve 8) [90,101].
Water molecules could strongly interact with oxide surfaces and other water molecules
[53,90,101]. Therefore, the differences between evaporation or water alone and water bound oxide
surfaces were relatively small (Fig. 57). The rearrangement of nanoparticles in aggregates or
decomposition of larger core-shell nanoparticles affected the desorption of water (Fig. 57b)
differently in comparison with evaporation of hexane (Fig. 56b). However, there were similar
regularities for both adsorbates. For example, evaporation of them was slower from initial AST1
(Figs. 56b and 57b, curve 3) [90,101].
Despite decomposition of large CSNP and an increase in SBET for AST1 cryogel,
evaporation of both adsorbates was slower than that from initial AST1 [90]. However, water could
be more slowly evaporated from AST1 cryogel (Fig. 57b, curve 5) than from MCA AST1 (curve
4), but for hexane, there were the opposite results (Fig. 56b). All these results could be explained
by several reasons. First, hexane much weakly affected the texture of the nanooxide powder than
water; i.e., hexane could not change packing of nanoparticles in the secondary structures. Hexane
molecules interact with oxide surfaces and other hexane molecules much weaker than water
molecules. Hexane molecules were larger than water molecules. Therefore, contact surface area
was smaller for hexane molecules, which could not penetrate into narrow voids between adjacent
nanoparticles in contrast to water molecules [90,101].
Fig. 57. Evaporation of water (at 13±1 oC) from (a) suspended samples at amounts of water of 1 g per 0.05 g of FMO,
and (b) final portion of water evaporation starting at residual amounts of water of 0.05 g (i.e. after
evaporation of 0.95 g of water).
Evaporation of n-decane occurred very slowly (Fig. 56c), approximately four times slower
than water (Fig. 57) or two orders of magnitude than hexane (Fig. 56a). Decane alone is
evaporated slower than decane disturbed in the suspensions. However, decane evaporation from
the suspension of AST1/A-300 cryogel (composed of compacted particles) was similar to that for
decane alone. Note that decane evaporation from AST1 cryogel occurred faster than from
AST1/A-300 cryogel due to narrower pores in the latter [90].
Thus, experimental and theoretical investigations of evaporation of various liquids (polar
water, acetone and 1-butanol and nonpolar n-hexane, benzene and carbon tetrachloride) from
various silicas showed [90,101] that the evaporation rate depended on molecular properties of
adsorbates (molecular weight, intermolecular bond strength, surface tension, density, boiling and
critical temperatures), surface chemistry and texture of adsorbents, as well as location of molecules
in depth of pores. Different orders of values of certain characteristics (e.g., Tb and Tc for water and
butanol) could lead to intersection of curves of evaporation rate (or time) vs. temperature for
68
different compounds. However, in general, the order (position) of curves of the rate and time of
evaporation vs. T well correlated to the boiling point of bulk liquids and the pore diameter of
adsorbents at the same values of other their characteristics and upon the evaporation from the
depth of pores [90,101].
The decelerating of evaporation from the depth of pores was much stronger than that from
the entrance of pores due to slower motion of the molecules toward the outer surface of adsorbent
beads with decreasing size of pores [90,101]. The difference in the evaporation rate from various
silica samples at a fixed temperature decreased with increasing boiling point (from hexane, water,
to butanol). The effect of diminution of the surface tension in narrow pores on the evaporation of a
certain adsorbate was small in the depth of pores in comparison with the effects of increasing pore
length, decreasing pore diameter and increasing boiling point of adsorbates. Additionally, to
estimate the time of complete evaporation of liquids from pores, one should take into account the
continuous movement of adsorbate molecules along pores toward the outer surface of beads. The
rate of evaporation from a surface of beads and the entrance of mesopores was close to the rate of
evaporation of free droplets of similar size. However, the rate of evaporation from depth of long
and narrow pores was much lower (by several orders of magnitude) than that from the surface of
beads. The conformerization effects could enhance evaporation, especially if the populations of
bent conformers at higher energy was significant in the liquid phase in pores (e.g., due to
increasing temperature) [90,101].
Developed theoretical approach used to calculate the evaporation rate and time as well as
condensation/evaporation coefficients for various liquids from various porous media could be
useful for deep analysis of the experimental data obtained with such techniques as
thermogravimetry, temperature-programmed desorption, gas chromatography, adsorption-
desorption at a fixed temperature, calorimetry, etc. [90,101]. Note that the averaging of the
theoretical results should be performed for such characteristics as the pore diameter and length of
adsorbent particles, degree of pore filling by adsorbates, surface tension p(T) of adsorbates in
narrow pores (however, changes in p(T) could be ignored at Rp > 10t 3.2 nm for water, 3.4 nm
(benzene), 4.9 nm (CCl4), etc.), interactions of molecule-molecule and molecule-surface, and
conformerization of adsorbate molecules, especially long linear ones [90,101].
Note that the interfacial behavior of water plays a very important role in building industry
(especially in respect to the properties of cement and concrete). Cement hydration led to the
dissolution of alite (C3S, 3CaO·SiO2) and belite (C2S, 2CaO·SiO2) with the formation of calcium
silicate hydrates (C–S–H) and calcium hydroxide (C-H) [160-166]. While calcium hydroxide is a
very well-known crystalline phase [160,161]; however, the C–S–H structure, chemical
composition, density and morphology remained uncertain [162]. C–S–H is a highly disordered
nanoscale material containing a significant amount of bound water (Fig. 58). There were a few
techniques (e.g., NMR spectroscopy (Fig. 59), relaxometry, and cryoporometry) able to
characterize C–S–H without removing the bound water (Figs. 58 and 59). Note that dehydration
could damage the nanoscale structures of the material. The C–S–H precipitate characteristics
depended on the chemical and physical conditions in which the hydration takes place [163–166].
Note that silica fume (an industrial solid waste formed as a byproduct during the production of
alusil alloy or metallurgy grade silicon, and having pozzolanic properties and phase minerals
comprising major silica components (95.1%) in the amorphous form, with traces of alumina,
alkaline metal oxides, and transition metal oxides [167]) was used in preparation of cement and
concrete. The particle size distribution of silica fume is very wide, and primary particles were from
several nanometers up to 200-300 nm in diameter [167].
69
Fig. 58. (a) Evolution of the different water populations as a function of hydration time for white cement with 10%
silica fume (colors), all at w/b = 0.4 (diamonds were solid signals, Isolid, squares were C–S–H interlayer water,
ICSH, triangles were C–S–H gel pore water, Igel, and solid circles were “free water”, Icap, becoming
interhydrate water. The solid signal for the plain white cement mix was not displayed. (b) Evolution of the
associated T2 relaxation times for the different water populations. The grey lines were the equivalent of the
mobile fractions and T2 relaxation times for the plain white cement.
A surface of FMO nanoparticles is practically open because FMO nanoparticles are
nonporous [1-12,53]. Therefore, low-molecular weight polar or charged adsorbates or metal ions
were poorly adsorbed onto unmodified surface of FMO [53,168-173] because the desolvation
energy for bound adsorbate could be greater than the solvation energy of this adsorbate in the bulk
solution. These properties could be changed due to adsorption of polymers [174-183] or chemical
functionalization of the FMO surface [184,185]. Complex FMO or FMO matrices modified by
deposion of catalytically active phase (e.g., anatase) could be used as effective catalysts, especially
photocatalysts because nanoparticulate structure of FMO is appropriate for excitation of electrons
under UV emission and interactions with bound compounds [108,109,186].
A variety of experimental and computational techniques including macro- to meso- and
microscale approaches, pivotal applications, defines the most important interfacial quantities and
their experimental investigations, providing theoretical background and detailed solutions [187]. In
this work, vital techniques used in interfacial flow problems, such as modern meshless numerical
methods and conventional computational fluid dynamics methods were described. The
technicalities of correctly using the computational methods developed for interfacial flows, as well
as the simulation of interesting interfacial flow physics were discussed. This work offers an
authoritative and state-of-the-art overview of computational methodologies and simulation
techniques for the quantification of interfacial quantities [187].
70
Fig. 59. 29Si MAS NMR spectra (9.4 T, νR = 6.0 kHz) of the anhydrous white cement with silica fume and of the blend
after hydration for 3, 7, 14, and 28 days.
Catalytic effects
It is difficult to expect that FMO could be good selective catalysts for fine organic
synthesis because flame (pyrogenic) synthesis leads to poorly controlled structure of complex
oxides. However, individual fumed titania (such as Degussa P25, Fig. 60) or titania doped by a
small amount of another oxide (not more than several percent) giving a solid solution in the main
phase could be effective photocatalysts for deep decomposition of organics under UV or visible
light. Besides titania some other fumed oxides (e.g., containing ceria, zirconia, ZnO, etc.) can be
used as catalysts and photocatalysts. Titania (anatase) initial or doped by various metal oxides as
well as titania/silica composites possess significant catalytic activity [188-193]. Catalytic activity
for the photodecomposition of methylene blue, MB (selected as representative aromatics for
photocatalytic decomposition) was found as greatest (per gram of TiO2) for non-treated ultrafine
titania PC-500, which has the largest SBET value and smallest particle size of the titania-containing
materials studied [186]. However, this activity calculated per m2 is higher for PC-105, possessing
a much smaller value of SBET than PC-500. The activity per unit surface area of titania is greatest
for the fumed silica-titania complex oxide ST20, possibly due to enhanced adsorption properties of
this material [53]. Calcination of PC-500 at 650 oC led to enhancement of anatase content and
catalytic activity, but heating at 800 and 900 oC lowers the anatase content (since rutile appears)
and diminishes catalytic activity, as well as the specific surface area because of nanoparticle
sintering [186].
Nano-sized particle TiO2-doped SiO2 gels demonstrated the dependence of the structural
and other characteristics on content of silica and a method of material preparation (Fig. 60) [194].
The concentration of TiO2 loading, the particle size, and the surface characteristics were shown to
relate to the degree of UV absorption and the measured energy bandgap (Eg) of the composites
(Fig. 60a). Note that the value of Eg is one of the main characteristics of photocatalysts since the
lower this value, the effective is the catalysts and it can operate even under visible light if this
value is appropriately small.
71
Fig. 60. (a) The energy bandgap (Eg) of TiO2 (synthesized using titanium (IV) isopropoxide, TIP) – SiO2 gel, TiO2
(TIP) – fumed SiO2, TiO2 (Degussa P25)–SiO2 gel and TiO2 (P25) – fumed SiO2 versus wt.% TiO2 present;
HRTEM image of (b) TiO2 (P25) and mixture with TiO2 (P25) (10 wt.%) and fumed silica Aerosil 200
(Degussa); and (c) SEM image of TiO2 (P25).
The reaction rate constant of catalytic photodecomposition of MB could be considered as a
distribution function due to the effects of surface heterogeneity, titania particle size distribution
(i.e. band gap distribution), and other structural and energetic factors. Therefore, the concentration
of MB in solution (with catalyst suspension) as a function of time of UV emission could be written
as follows using a simple equation [186]
max
min
0
0
( ) ( )
1
k
e
k
C
C t f k dk
kC t
(22)
Calculations of the f(k) function were carried out using a regularization procedure with unfixed
regularization parameter and nonnegativity condition [195]. More complex experimental Ce(t)
graphs were observed for some systems. Therefore, the kernel in the integral equation was
modified to include additional terms related to diffusion and decomposition of MB
[108,109,186,193].
max
min
0 3
1 2
1
( ) [ exp( )] ( )
1
k
e
k
C t C a z f k dk
a z a z
(23)
where z = kC0t; a1, a2, a3, and were the equation parameters determined by minimization of the
functional
max
2
0 3
1 20
1
( , , , ) ( ( )) [ exp( )] min
1
t
i ek a spline C t C a z dt
a z a z
(24)
The effectiveness of photodecomposition at a titania surface (Figs. 44 and 45) depended on
many factors such as (i) anatase/rutile ratio and amorphous phase content (since rutile and
amorphous phases generally have lower catalytic activity than anatase); (ii) specific surface area
(i.e. contact area for adsorbed reactant); (iii) PSD, particle morphology (e.g., PaSD), and particle
aggregation (accessibility of the titania surface to reactant because hindered diffusion could be in
72
very narrow pores, and reduced UV radiation transmitted into narrow, deep and long pores); (iv)
the presence of surface functionalities (e.g., sulfate) which could act as a catalyst poison, electron
trap, mediator of nanoparticle aggregation, etc. [10,15,53,188-192]. To elucidate some of these
effects the catalytic activity of PC titanias, fumed titania and ST20 was compared for the
photodecomposition of MB (Fig. 61) [109]. The main results were: (i) PC-500 composed of the
smallest anatase particles and with few sulfate groups possesses the largest catalytic activity per
gram of oxide. PC-100 (with more sulfate groups than PC-105) is less active than PC-105, despite
a larger SBET value. Fumed titania, composed of anatase and rutile and having a lower SBET value
than PC-100 or PC-105, as well as reduced hydrophilicity potentially resulting in nanoparticle
aggregation, exhibited a smaller photocatalytic effect. Finally, ST20, with 80 wt.% silica, not
surprisingly was the least catalytically active as measured per gram of oxide [109,186].
Re-calculation of the photocatalytic activity as A = (CMB,0 - CMB)/CMB,0/SBET (i.e.
normalized to surface area) (Fig. 61b) showed that PC-500, PC-105 and fumed titania exhibited
high activity at short times t < 6 min [109]. With time t > 6 min, PC-105 and fumed titania,
possessing low SBET values, exhibited the greatest activity per m2. Taking into account the surface
content of titania in ST20 (24 wt.% TiO2) one could assume that surface anatase in this sample is
the most active (per m2) because of stronger and/or a greater number of adsorption complexes
formed with MB [109], and due to the Langmuir-Hinshelwood reaction mechanism [196]. The
latter was supported by the fact that the initial MB concentration (~20 ppm), and the equilibrium
MB concentration after 20 minutes of stirring in the absence of UV light (sufficient time for
equilibrium to be achieved), was nearly identical for PC titanias and fumed titania. However, the
MB solution concentration exhibits a significant reduction with ST20 to 13 ppm, indicating MB
adsorption was much greater for ST20 than the other materials [109].
Calculations of the distribution functions of effective MB decomposition rate were
performed with Eqs. (22)-(24) using the data shown in Fig. 44a. Both integral equations give
similar results with respect to f(k) peak position (Fig. 62). However, the model equation with a
more complex kernel in Eq. (10) gives narrower distributions, and splits f(k) into two peaks for
ST20. PC-500 exhibits a broader distribution than the other titania samples (Fig. 62b). This was
possibly due to the more complex textural properties of PC-500 [109], and the difference in the
reaction rate for MB adsorbed at the outer surface of aggregates of primary particles and in voids
between primary particles in aggregates (i.e. in pores).
Heating of titania strongly affected its catalytic activity because of changes in the anatase-
to-rutile ratio and amorphous-to-crystalline phase transformation, as well as sintering of
nanoparticles resulting in significant reduction of the value of SBET. The results of this influence
were readily observed for PC-500 [109]. The greatest catalytic activity for MB
photodecomposition was observed for titania treated at 650 oC, when the anatase-to-rutile
transformation was not yet occurred, but the amorphous-to-crystalline phase transformation had
[109]. Taking into account the 5-fold reduction of the SBET value for PC-500 resulting from 650 oC
heat treatment, the enhancement of the activity (per m2) was substantial. This sample was more
active (per m2) than titania Degussa P25 (frequently used as a reference photocatalyst) at t > 20
min (Fig. 61b). Calcination at higher temperatures (800 and 900 oC) results in loss of titania
catalytic activity due to at least two factors: i) anatase-to-rutile transformation and, ii) sintering of
nanoparticles since the dXRD value grows for all calcined PC titanias [109]. Additionally, it may be
that certain treatments of sulfate-containing PC titanias (produced by liquid phase synthesis) could
provide reduction of the band gap, because of embedding of S atoms in the lattice instead of O
atoms and, therefore, the photoexcitation threshold [197,198].
73
Fig. 61. (a) Relative diminution of residual concentration
of MB in solution as a function of UV emission
with 0.2 g of oxides, and (b) relative activity of
catalysts per m2 of their surface area; PC-500*
(4) after heating at 650 oC; fumed titania
Degussa P25 (5); and ST20** (8) with
consideration for surface titania content.
Fig. 62. (a) Normalized distribution functions of the
reaction rate constant of catalytic
photodecomposition of MB determined with
Eq. (22); (b) distribution functions of the
reaction rate constant determined using Eqs.
(23) and (24).
Thus, investigations of the morphological, structural, adsorption, and catalytic
characteristics of ultrafine titanias and fumed titania-containing complex nanooxides showed [109]
that There were several key parameters that affected the catalytic properties: (i) accessible particle
surface area and porosity of primary and secondary particles; (ii) titania nanoparticle phase
composition; (iii) surface content of titania (anatase) in complex nanooxides; (iv) sintering of
nanoparticles affected by calcination conditions; (v) content and type of surface hydroxyls as not
only participants in photocatalysis but also as potential adsorption sites on complex nanooxides.
As a whole the titania composed of the smallest anatase nanoparticles (PC-500) possesses the
largest catalytic activity per gram of oxide in the photodecomposition of MB. However, estimation
of the catalytic activity per m2 of titania surface suggests that PC-105 and fumed titania, composed
of larger nanoparticles than PC-500, possess a higher activity than non-heated PC-500 at reaction
time t > 6 min. PC-500 calcined at 650 oC exhibits a striking increase of activity in MB
photodecomposition (per m2), at least comparable to fumed titania Degussa P25. The complex
oxide ST20 possesses the highest catalytic activity per m2 of titania, possibly because of enhanced
adsorption of the reactant MB to the surface of this material, and the Langmuir-Hinshelwood
reaction mechanism [109].
Confined space effects and interfacial phenomena
Confined space effects for various adsorbates could be analyzed using 1H NMR, DSC,
TSDC, DRS, TG and some other methods [53,199-217]. To avoid contribution of signals of
water molecules from ice or other solid fractions or macromolecules, low-temperature 1H
74
NMR spectra could be recorded for static samples using a narrow bandwidth (20 kHz) [53].
Changes in the Gibbs free energy (G) of bound water and free surface energy (S), estimated
as the modulus of integrated changes of the G values in the bound water layers, were
determined from the temperature dependences of the amounts of unfrozen water (Cuw in mg of
water per gram of silica) at T = 200-273 K < Tf = 273.15 K [53]. The amount of water (h) in
samples was dependent on the type of the studied materials. The temperature dependence of
the Gibbs free energy of ice [53,199] could be approximated as follows:
Gice = 0.0295 0.0413Т +6.64369105(Т)2 + 2.27708108(Т)3 (kJ/mol), (25)
where Т = 273.16 T at 160 K < T 273.15 K. The fact that bound water may remained
unfrozen at T < 273 K suggests that its structure is disturbed and due to stronger interactions with
an adsorbent surface than with neighboring water molecules that results in inequality ice
i
w GG .
Further lowering of temperature reduces this inequality until the point Tf,bw, at which a certain
amount of bound water became frozen. At Tf,bw the relationship Gw = Gice pertains,
where 0)( w
i
w
i
w GTGG ( 0
wG denotes the Gibbs free energy of unperturbed bulk water at 273.15 K
and superscript i stands for interfaces). It was assumed that neither Gice, nor Gice were influenced
by the solid surface. The area under the G(Cuw) curve (temperature dependences G(T) and
amounts of unfrozen pore water Cuw(T) could be simply transformed into the relationship
G(Cuw)) determines overall changes in the Gibbs free energy of interfacial water [53]
max
0
uwC
uwS GdCA , (26)
where max
uwC is the total amount of unfrozen water at T 273 K, and A is a constant dependent
on the type of units used in this equation.
Water or other liquids could be frozen in narrower pores at lower temperatures, which
could be described by the Gibbs-Thomson (GT) relation for the freezing point depression
[53,200-205]
,
,
2
( ) sl m GT
m m m
f
T k
T T T R
H R R
, (27)
where Tm(R) is the melting temperature of a frozen liquid in cylindrical pores of radius R, Tm,
is the bulk melting temperature, the density of the solid, sl is the energy of solid-liquid
interaction, Hf is the bulk enthalpy of fusion, and kGT is a constant (e.g., kGT = 60 K nm for
water adsorbed on silicas, however, it can be varied in a wide range of 30-200 K nm depending
on the type of adsorbents [53]). On the basis of this equation and the methods sensitive to
transition of phase, different versions of cryoporometry, relaxometry, and thermoporometry
were developed to study the interfacial phenomena, behavior of adsorbates, and structural
characteristics of a variety of solid and soft materials and bioobjects [53,200-212].
Typical 1H NMR spectra were observed for unfrozen water bound in wetted powder (total
amount of water Cw = 11 wt.%) and aqueous suspension (Cw = 90 wt.%) of fumed silica/alumina
SA8 (used as a representative sample of complex FMO) recorded at different temperatures (Fig.
63) [30]. Relative amounts of bound unfrozen water (Cuw,270K/Cw) in the suspension were much
smaller (< 30 %) than that in the hydrated powder (100 %) in which practically all water (Cuw,270K)
was unfrozen at 270 K. In the suspensions of AST82 and SA8, the Cuw,270K value corresponded to
17 and 27 % of Cw, respectively (Figs. 63 and 64). The
2BET,NS value of SA8 was seven and four
times larger than that of AST82 and AST71, respectively (Tables 1 and 6). Therefore, the amounts
of interfacial bound unfrozen water were larger for both hydrated powder and aqueous suspension
of SA8 with lowering temperature than that for AST71 or AST82 (Fig. 64a). This corresponded to
stronger binding of interfacial water to smaller nanoparticles of SA8 (as FMO with larger specific
75
surface area) than to larger particles of AST materials, and the Gibbs free energy was lower (at the
same Cuw values) for the systems with SA8 (Fig. 64b).
Fig. 63. 1H NMR spectra of unfrozen water in (a) hydrated powder of SA8 (Cw = 11 wt.%) and (b) aqueous
suspension of SA8 (Cw = 90 wt.%).
Fig. 64. (a) Temperature dependence of unfrozen water content (Cuw) and (b) relationships between Cuw and changes
in the Gibbs free energy of bound water for hydrated powders (Cw = 11 wt.%) and aqueous suspensions (Cw =
90 wt.%) of different oxides.
The secondary (aggregates) and ternary (agglomerates) particles of FMO have relatively
soft structure, which could be rearranged on interactions with adsorbates, especially water, other
small polar molecules or polar polymers such as poly(vinyl alcohol), poly(ethylene glycol), etc.
[53]. Similar rearrangements were revealed on comparison of the PSD (Fig. 65) calculated on the
basis of the nitrogen adsorption onto dry samples and using NMR-cryoporometry (solving integral
Gibbs-Thomson equation) applied to the Cuw(T) functions for wetted samples or aqueous
suspensions [30,53]. Contribution of narrow voids at R < 3 nm filled by structured water in wetted
powder at Cw = 11 wt.% was greater than that of voids filled by nitrogen fluid in dry powder. The
changes in the wetted powder could be due to several effects. Water molecules penetrating
between adjacent nanoparticles (nanopore contribution grows) plasticize their contacts and
contract neighboring particles in denser aggregated structures (contribution of narrow mesopores
grows). Notice that these effects clearly appeared on wetting and drying of the nanooxide powders
[53].
76
Fig. 65. PSD for SA8 calculated from the nitrogen desorption data (curves 1) and the Cuw(T) function (NMR) for
hydrated powder (2) and aqueous suspension (3) (a) with normalized curves 2 and 3, and (b) initial range of
the PSD at R < 3 nm without the normalization of the curves.
On suspending, the rearrangements of secondary and ternary particles could be enhanced. This
caused stronger textural changes in FMO after suspending-drying than after wetting-drying or
interaction of the powders with organic solvents (e.g., ethanol) [53]. In the suspension of SA8
(Fig. 64) Cuw = 2.5 g/g at 270 K, i.e., the volume of bound water (V = 2.5 cm3/g) was
approximately two and four times larger than the pore volume (
2p,NV ) of suspended-dried SA8
(1.33 cm3/g) and the initial powder of SA8 (0.68 cm3/g), respectively (Tables 1 and 6). These
effects were due to loosing of contacts between nanoparticles on wetting or suspending and then
strong binding (up to formation of siloxane bonds) of nanoparticles one to several others on
drying. These aspects of the behavior of nanooxides were of importance on treatment or use of
these materials in liquid and polymer/solvents media [53].
Competitive adsorption and confined space effects
Co-adsorption of water and organic solvents onto nanooxides can lead to diminution of the
water interaction with the oxide surfaces. This effect was clearly observed on suspending of wetted
powder of AST71 in the CDCl3 medium since the G value of bound water became smaller as
well as Cuw (Fig. 64b, curves 1 and 2) [30,53].
Table 12. Parameters in correlation equation h (ppm) = a + bqh used to estimate the h values for different compounds (based on
GIAO/B3LYP/6-31G(d,p)//HF/6-31G(d,p) and semiempirical calculations) [53,209,210].
Compound Method a b Probes for calibration
Water PM3 18.81135 92.83742 8H2O-24H2O
Water PM6 27.97889 87.56668 16H2O
Water PM7 44.757685 131.473652 43H2O
Methane PM3 2.33089 26.24548 16CH4
Decane PM6 3.199315 16.632049 C10H22
Decane PM7 9.318837 61.145893 C10H22
PEG (CH) PM7 10.769365 55.335375 C6H14O4
PVA (CH) PM7 6.929922 50.712501 (-CH2CH(OH)-)4
PDMS (CH) PM7 5.044178 30.462862 H(-O-Si(CH3)2-)7OH
H(-OP(OH)(O)-)nOH PM7 27.526046 95.608505 H(-OP(OH)(O)-)3OH
For theoretical analysis of the influence of the confined space effects and surface structure
of adsorbents, quantum chemical calculations were performed for relatively large models (up to
15000 atoms) using semiempirical PM3, PM6, and PM7 methods (MOPAC 2012 & 2016). To
compute the values of the chemical shift of the proton resonance (H), a linear correlation function
77
H = aqH + b linking the values of H with the atomic charge (qH) of atoms H in bound water
molecules (Table 12) [53,209,210].
Fig. 66. (a) Chemical shift H (experimental data) as a function of temperature for bulk water and hydrated silica gels
(Si-40 (specific surface area SBET = 732 m2/g, pore volume Vp = 0.54 cm3/g), Si-60 (SBET = 369 m2/g, Vp =
0.75 cm3/g) and Si-100 (SBET = 349 m2/g, Vp = 1.23 cm3/g)) and fumed silica A-380 (SBET = 378 m2/g, Vp =
0.94 cm3/g), and (b) incremental pore size distributions (MND method) for these silicas [53].
The interfacial behavior of liquids depended on the type of a solid surface, e.g., contents and
types of hydrophilic or hydrophobic functionalities, number and density of active surface sites,
particle topology and porosity of the materials [53]. The chemical shift of water interacting with a
silica surface depended on the type of silica porosity (i.e., confined space effect, see Fig. 66),
surface modification (e.g., hydrophobization) degree of silica and presence of co-adsorbates (Fig.
67), and distribution of water in pores (silica gels) or onto the outer surface (fumed silica) of silica
particles (Figs. 66-68). There were two main tendencies for the properties of water interacting with
such hydrophilic adsorbents as silica gels. First, a decrease in the size of pores enhanced
clusterization of water; i.e. average H value should decrease. Second, an increase in a fraction of
water directly interacting with the silica surface with increasing specific surface area (SBET)
enhanced the polarization of water molecules leading to a decrease in magnetic shielding of the
protons; i.e. downfield shift should be observed [53].
Thus, these tendencies were opposite. Therefore, the H(T) curve for water bound to Si-40
(possessing narrowest pores among silica gels studied, Fig. 66b) at partial filling of pores was
located between the curves for water bound to Si-100 (broadest pores) and Si-60 (middle sizes of
pores). Changes in the water organization at a surface of silicas depended on both temperature and
pore size distribution. Thus, for silica gels with broader pores (Si-100, Si-60), the H(T) curves at T
< 250 K (when only mobile strongly bound water, SBW, was observed in the 1H NMR spectra)
demonstrate a weak dependence on T in contrast to water bound to Si-40 possessing much
narrower pores than Si-100 or Si-60 that caused the strongest confined space effects
[53,158,209,210].
78
Fig. 67. 1H NMR spectra of water calculated using PM6 and calibration function for 2000H2O (curve 1) and small
water clusters (mainly weakly associated water) interacting with hydrophobic silica (with trimethylsilyl
(TMS) groups, curve 2) or with the presence of methane molecules (3), silica was modeled by porous particle
(pore of 1.1 nm in radius), and experimental 1H NMR spectrum (266 K) of water (0.24 g/g) bound to silica
gel Si-60 partially modified by trimethylsilyl groups (curve 4) [209,210].
Fig. 68. 1H NMR spectra of water (517H2O) bound mainly onto outer surface or mainly in pore of silica particle (1620
atoms) calculated using PM7 method and a calibration function for water (Table 12) [209,210].
The textural features of nanosilica A-400 resulted in the formation of relatively large
domains of adsorbed water even at a low hydration degree (h = 0.02 g/g) [158]. This water could
be assigned to strongly associated water, SAW characterized by the chemical shift (δH) of
approximately 5 ppm (Fig. 67a) close to that of bulk liquid water [53]. In SAW, the average
number of the hydrogen bonds per a molecule was 3.5-4.0 that caused δH 5 ppm; however, δH
7.4 ppm for ice Ih. Notice that in liquid water, a part of molecules was interstitial; i.e. these
molecules were out of the hydrogen bond network. Therefore, the average coordination number of
water molecules could be about 4.5 (i.e. larger than that in ice having, therefore, a lower density
than liquid water) but δH,lw < δH,ice due to thermal distortion of the hydrogen bonds (length O-HO
and angle OHO were non-optimal) and the presence of interstitial water molecules [53].
79
Fig. 69. 1H NMR spectra (for static samples) of adsorbates bound to nanosilica A-400 (a-d) initial (b = 0.07 g/cm3)
and (e, f) compacted (b = 0.15 g/cm3): (a) water (h = 0.02 g per gram of dry A-400); and water (h = 0.05 g/g)
and n-decane of content Cd = (b) 1.5 and (c, d) 7.3 g/g, (d) with addition of polyphosphoric acid (PPA) at
CPPA = 0.05 g/g, (e) 0.73 and (f) 3.7 g/g; (c) A-400 initial (solid lines) and heated (dot-dashed lines) at 723 K
for 10 min, loaded into the ampoule, heated at 473 K, slightly cooled, decane (7.3 g/g) was added to hot silica
and boiled for 1 min.
A fraction of water bound to A-400 was unfrozen at T < 260 K, i.e., it was strongly bound
water, SBW since changes in the Gibbs free energy G < 0.5 kJ/mol [158]. The δН value (i.e.
magnetic shielding of protons) of water depended on the average number of the hydrogen bonds
per a molecule (<nHB>) and the length O-HO and angle OHO of these bonds. It was equal to
7.4 ppm for single ice Ih crystals and ca. 1 ppm for free water molecules [53].
Despite water and decane are immiscible liquids, addition of decane (Cd = 1.5 g/g) to
nanosilica A-400 weakly hydrated (h = 0.05 g/g) resulted in the upfield shift of 1H NMR signal of
water, since the δH value changed in the 3.2-3.5 ppm range (Fig. 69b) instead of 4.5-5 ppm for
water adsorbed alone (Fig. 69b) with a weak dependence of the δH values on temperature [158].
This decrease in the δH value could be explained by enhanced clusterization of bound water under
action of hydrophobic decane. Decane could displace water into both narrower pores (inaccessible
for larger decane molecules) and larger pores (to decrease contact area between two immiscible
liquids). Both displacements result in decrease in the Gibbs free energy of the system. However,
80
the second type of the displacements was more characteristic for the systems with greater amounts
of water than h = 0.05 g/g [53,158]. Despite a relatively great amount of decane (1.5 g/g), the
sample remained in the powder state; i.e. all decane was in bound state [158].
The 1H NMR spectra (Fig. 69b,d) were recorded for both decreasing (from 290 K to 210 K)
and increasing (from 210 K to 300-310 K) temperatures [158]. Adsorbed decane demonstrated two
signals at 0.9 and 1.25 ppm (Fig. 69b), corresponding to methyl and methylene groups,
respectively. A small shift of decane signals with temperature could be due to the temperature
dependence of the magnetic susceptibility of samples and/or changes in the parameters of a
resonant circuit of a sensor without stabilization of resonance conditions by using deuterons.
During cooling of a sample, signal intensity of both water and n-decane strongly decreased at T =
220 K (Fig. 69b). During heating, intensive signals of n-decane (and water) appear at T > Tf (Fig.
53) [158].
A sample placed in the n-decane dispersion medium (Cd = 7.3 g/g), stirred at 293 K, was
characterized by an additional upfield shift of bound water signal to approximately 1.5 ppm (Fig.
69c), i.e. the water clusterization increased in comparison with the sample at Cd = 1.5 g/g [158].
Water with similarly low values of δH could be assigned to weakly associated water, WAW, which
represents individual molecules, 1D or strongly branched 2D or 3D clusters at the boundaries of
polar (e.g., silica) and nonpolar (e.g., n-decane) media or mosaic polar-nonpolar structures
[53,158]. The δH values for H atoms in the OH groups participating and non-participating in the
hydrogen bonds as a proton-donor strongly differ because O or other electron-acceptor atoms
displace the electron density from the H atoms that leads to a decrease in the magnetic shielding of
the protons. Therefore, WAW is characterized by a low (< 1) average number of the hydrogen
bonds per a molecule as a proton-donor. However, this WAW bound to nanosilica A-400
corresponded to strongly bound water (SBW) since its signal already appears at T > 240 K (i.e. G
< 1 kJ/mol) when decane signal appears too. This may suggest that water (i.e. SBW) locates in
narrow voids (pores) of A-400 corresponding to the first peak of PSD at R = 0.6-1.5 nm [158].
To analyze the influence of bound water on the behavior of bound n-decane,
polyphosphoric acid (PPA) was added as a dehydrating agent in the amount corresponding to the
amount of water (Fig. 69d) [158]. Really, water signal was not observed after addition of PPA, i.e.
all water was bound to PPA. Signal of water/PPA could be observed at ca. 10 ppm as a very broad
signal. This signal was not observed (Fig. 69d) due to features of the used measurement technique
and a narrow bandwidth (20 kHz). The PPA addition led to a decrease in splitting of decane
signals (compare the spectra in Fig. 69b,c,d). Thus, strong bonding of water by PPA led to
enhancement of interaction of bound decane with the nanosilica surface [158].
Pretreatment of nanosilica A-400: heating at 723 K for 10 min, loading a hot sample into a
NMR ampoule, heating at 473 K, slight cooling, and then addition of decane to hot silica sample
and boiling in decane for 1 min, was used to reduce the silica surface hydrophilicity [158]. This
treatment resulted in diminution of melting delay (Figs. 69c and 70) since signal intensity of a
fraction of mobile decane increased much stronger at T > Tf, especially at T > 260 K.
This result could be explained by a better distribution of silica nanoparticles (in comparison with
the initial silica) in the decane medium and stronger interaction of decane with the silica surface,
which was characterized by a decreased number of surface silanols and the amount of residual
water due to heating under the mentioned conditions. The stronger interaction of decane with the
silica surface appears in broadening of the 1H NMR bands (Fig. 69c, dot-dashed lines) similar to
samples at a low content of decane located only in pores of silica [158].
81
Fig. 70. Temperature behavior of signal intensity (normalized to intensity I0 at 290-310 K) of n-decane bound to initial
(bulk density 0.07 g/cm3) and compacted (0.15 g/cm3) A-400 depending on Cd and presence of PPA; curve
labeled as ‘hot’ corresponded to A-400 heated at 723 K for 10 min, loaded into the ampoule, heated at 473 K,
slightly cooled, and addition of decane to hot silica and boiling in decane for 1 min.
Freezing/melting curves, showed as relative intensity of 1H NMR signals of mobile decane
molecules confined in textural pores of A-400, were characterized by a broad hysteresis loop (Fig.
70) [158]. This was due to melting of solid alkane structures with a certain delay in comparison
with their freezing at the same temperatures. During heating of the frozen A400/water/decane
system at Cd = 7.3 g/g, melting of a significant fraction of decane occurred at temperatures higher
than the freezing point (Tf = 243.5 K) of bulk decane (Fig. 70). An increase in signal intensity was
observed up to 280-310 K. Addition of PPA led to certain changes in the freezing and melting
temperatures of decane. During freezing, a strong decrease in the signal intensity of decane was
observed at 220-230 K, i.e. at T < Tf (Fig. 70). This effect could be due to relatively weak
interaction of a significant fraction of decane (frozen at T > 230 K) with the pore walls. This was
clearly seen from the difference in the freezing curves at different content of decane. Since at Cd =
7.3 g/g (only ~20% decane was unfrozen at 230 K), a fraction of decane weakly bound or non-
interacting with the silica surface was much greater than that at Cd = 1.5 g/g (~80% decane was
unfrozen at 230 K), because decane located in pores was unfrozen at 230 K [158].
It is known that compacting of nanosilicas (e.g., during MCA) could result in an increase in
contribution of mesopores due to rearrangement of secondary structures with increasing bulk
density from initial b = 0.04-0.07 g/cm3 to 0.2-0.3 g/cm3 for treated samples [53]. Therefore,
confined space effects for n-decane could be enhanced (in comparison with initial A-400 at b =
0.07 g/cm3) if it was bound to compacted A-400 at b = 0.15 g/cm3 (Figs. 69 and 70) [158]. Really,
at least, two effects were observed. First, in contrast to the initial nanosilica, the 1H NMR spectra
include broader lines without fine structures related to the CH3 and CH2 groups of n-decane (Fig.
69e,f). However, at Cd = 3.7 g/g, a certain line structure was observed (Fig. 69f) but less clearly
than for n-decane bound to the initial silica (Fig. 69b-d). Second, the integral intensity
demonstrated a smaller decrease during heating at T > 260 K than even at a larger Cd value for the
initial A-400 sample (Fig. 70). Thus, enhanced mesoporosity of compacted A-400 led to the
stronger confined space effects for bound n-decane and delay in melting of frozen structures
diminished [158].
82
Fig. 71. 1H NMR spectra (for static samples during heating) of adsorbates bound to silica gel Si-60 at decane amount
Cd = (a) 3.7 g/g and (b) 0.44 g/g and residual water at h 0.005 g/g.
As it was mentioned above, nanosilicas could be assigned to adsorbents with a ‘soft’
texture because adsorbates (as well as thermal, mechanical, mechanochemical, and hydrothermal
treatments) could easily rearrange aggregates of primary nanoparticles and especially
agglomerates of aggregates [53]. To elucidate the effects of the textural changes in A-400 due to
interaction with n-decane (and water), silica gel Si-60 (composed of much more rigid beads than
agglomerates of A-400) with the SBET and Vp values similar to those of A-400 (Table 1) was used
in similar 1H NMR studies [158]. Notice that alkanes adsorbed onto Si-60 could not affected its
textural characteristics since silica gels were rigid adsorbents in contrast to ‘soft’ FMO.
Samples of silica gel Si-60 were prepared with different amounts of n-decane (Cd = 0.44
and 3.7 g/g). Si-60 was preheated at 445 K, and then decane added to cooled silica filled a portion
of pores (Cd = 0.44 g/g) or completely filled of pores and formed the dispersion medium (Cd = 3.7
g/g) [158]. The last Si-60/alkane sample was heated at 445 K for 1 min to remove air bubbles, and
it was transparent after this heating. It included ca. 0.005 g/g of residual water (Fig. 71a). The
amount 0.44 g/g of decane (liquid density 0.73 g/ml) provided filling of a major portion (0.6
cm3/g) of pores of Si-60 (Table 1, Vp). At decane excess, the 1H NMR spectra include two
relatively broad signals at 1.25 and 0.9 ppm at T > Tf 243 K (Fig. 71a). Signal of residual water
appears as a shoulder at 1.5 ppm. Additionally, methylene groups linked to CH3 groups in the
decane molecule could contribute this signal. Water bound in this sample corresponded to WAW,
as well as for alkane/A-400 system (Fig. 69) [158]. At T 240 K, the signal width decreased and
the upfield shift was observed due to changes in the magnetic susceptibility of the sample with a
frozen fraction of decane. At Cd = 0.44 g/g when alkane molecules were located only in pores of
Si-60, signals of methyl and methylene groups were broadened, overlapped and characterized by a
small downfield shift with decreasing temperature (Fig. 71b). This effect was much stronger than
that for decane bound to the treated nanosilica A-400 with a greater content of alkane [158]. This
83
difference could be explained by stronger confined space effects for decane located in narrow
mesopores of rigid silica gel beads in comparison with the ‘soft’ A-400 powder [158].
Fig. 72. Temperature behavior of signal intensity (normalized by dividing by the intensity I0 at maximal temperature
used, 290-300 K) of decane bound to Si-60 at Cd = 3.7 and 0.44 g/g or A-400 at Cd = 7.3 g/g during heating
regime.
At T < Tf, the temperature dependences of signal intensity of decane were similar for both
samples with Si-60 at Cd = 3.7 or 0.44 g/g (Fig. 72) but they were located above that of A-400 at
Cd = 7.3 g/g due to a greater excess of bulk decane in the last sample [158]. This caused signal
appearance only at T > 240 K during thawing of frozen decane/A-400 with significant melting
delay (Fig. 72). Consequently, a greater amount of decane bound to Si-60 was unfrozen in
comparison with A-400 at the same temperatures. This difference could be explained not only by
the difference in the amounts of decane but also by the difference in the texture of these silicas. At
Cd = 3.7 g/g, the total amount of alkane was unfrozen at T > Tf during heating (Fig. 72) in contrast
to the sample at a low amount of decane (0.44 g/g). This effect was due to a very small outer
surface area of Si-60 beads, which could affect the temperature behavior of decane located out of
pores [158]. Therefore, the effect of the silica gel on bulk alkane (melting delay) was much smaller
than that for nanosilica having a much greater outer surface of aggregates of nanoparticles which
affected a greater amount of nearly bulk decane and enhanced the melting delay [158]. Notice that
similar structurizing effects were observed for aqueous suspensions of nanosilicas, since 10-12
wt.% of nanosilica was enough for total structurization of all water in the suspension [53]. If all
decane was bound in pores of Si-60 at Cd = 0.44 g/g, its fraction was immobilized at T > Tf, and an
increase in intensity was observed up to 280-310 K similar to that for decane/A-400 (Fig. 72); i.e.
solid or quasi-solid decane structures did not melt due to the effect of silicas and this effect was
stronger for nanosilica. In other words, the same amount of A-400 could structure a much greater
amount of decane than silica gel. Thus, different morphological characteristics of A-400 and Si-60
at close SBET values led to very different effects on decane located in pores and out of pores [158].
The H(T) function depends on the number of possible configurations of the water
molecules in the hydrogen bonds network. Considering that this number is inversely proportional
to the average number of the hydrogen bonds <nHB>, according to the entropy definition S
kBlnnHB [211]. Therefore, the temperature derivative of the measured fractional chemical shift
[211]
lnln ( ) HB
p pp
nT S
T T T
(28)
should be proportional to the constant pressure specific heat CP(T) (CP = T(S/T)P). This aspect
was analyzed in detail in [211]. The T(ln(T)/T)P functions were compared for water and decane
bound to A-400 and Si-60 (Fig. 73) [158].
84
Heating of Si-60/decane (Cd = 0.44 g/g) frozen at 210 K led to an increase in the entropy
(Fig. 73, curve 1) at T < 230 K. Then the entropy decreased (minimum is at 243 K Tf,decane) due
to ordering of mobile molecules at T < Tf,decane. At T > Tf,decane, the entropy increased due to
increasing disorder in a melting fraction of the adsorbate. The next minimum at ca. 273 K (i.e. at
Tf,water) is due to the effects of ordering (separating) mobile water and decane (Fig. 72). Excess of
decane (Cd = 3.7 g/g), when the dispersion medium for Si-60 beads was with decane, resulted in
much lower entropy at Tf,decane (Fig. 73, curve 2) due to freezing of the major portion of decane at
T < Tf,decane (Fig. 72). For adsorbed water (Fig. 73, curves 3 and 4), the entropy decreased at T
close to Tf,water, and then it increased due to increasing disorder of liquid water with increasing
temperature [158].
Fig. 73. Function T(ln(T)/T)P vs. temperature for n-decane (C10) and water bound to Si-60 or A-400 or bulk water.
Theoretical calculations of the 1H NMR spectra of n-decane and water molecules free and
bound in silica mesopore (1.1 nm in radius) or bound to a silica nanoparticle (~2.1 nm in diameter)
showed (Fig. 74a) that interactions with silica resulted in downfield shift of the band maximum of
decane and broadening of the spectrum (Fig. 74a, curves 3 and 4) [158]. This occurred despite
weak interactions between a silica surface and nonpolar alkane molecules. The effect was stronger
for decane molecules located in the narrow pore (curve 4) than that for the molecules bound to the
silica nanoparticle (curve 3). Water molecules co-adsorbed with decane without formation of
clusters were characterized by a band at 1-2 ppm (Fig. 74a, curve 5), i.e. this was WAW. These
results were in agreement with the experimental data showing the WAW appearing due to co-
adsorption with decane [158]. Notice that decane-decane interaction is relatively weak. The free
energy of solvation of n-decane in the n-decane medium calculated using the SMD method (HF/6-
31G(d,p) basis set) gives 16 kJ/mol. This was approximately 2.5 times smaller (by the modulus)
that the solvation energy of silica in water. Therefore, the 1H NMR spectra calculated for decane
molecules free and in the decane dispersion medium were practically the same in contrast to the
molecules free and interacting with the silica surface because the interaction of a decane molecule
with the silica surface is stronger than that with neighboring decane molecules [158].
Interaction of water molecules with the polyphosphoric acid (PPA) fragments caused
significant downfield shifts for both water molecules (compare curve 1 in Fig. 74b and curve 5 in
Fig. 74a, H = 3-5 ppm) and OH groups in the PPA (compare curves 2 and 3 in Fig. 74b, H =
5-7 ppm) [158]. This effect is a typical downfield shift for acids interacting with water molecules
due to decreasing of the electron density on the H atoms (i.e. decreasing magnetic shielding of
protons) which tend to be de-attached as H+ to form Н3О
+(Н2О)n (i.e. hydronium ion H3O
+, Zundel
(H5O2
+) or Eigen (H9O4
+) cations) and acidic residua both strongly affecting the water structure
[53]. Notice that for the strongest hydrogen bonds of basic compounds with strongly acidic OH
85
groups, the chemical shift of proton resonance could reach much higher values up to 16-20 ppm
[53,158].
Fig. 74. Theoretically calculated 1H NMR spectra of (a) n-decane and water molecules free (curves 1 and 2) and
bound in mesopore of silica (curve 4) or to silica nanoparticle (curves 3, 5); and (b) water (curve 1, PM7) and
OH groups in PPA fragments hydrated (2, PM7) and non-hydrated (3, GIAO/B3LYP/6-31G(d,p)).
The interaction energy of water or decane with silicas, measured as the heat of immersion
of powder samples in liquids, increased with increasing SBET of silicas (Fig. 75) [158]. The curve
course is practically the same for nanosilicas and silica gels interacting with water because of the
identity of the nature of their surfaces composed of the siloxane bonds and silanol groups. The
latter (in contrast to the siloxane bonds) were the main adsorption sites for water [53]. In the
alkane medium, the increase in the interaction energy with increasing SBET was smaller than that
for the aqueous medium. For silica gels, the ratio Kh = Qwater/Qdecane was large (2.5-3.3), because
silica gels are highly hydrophilic materials, and it increases with SBET value. For nanosilicas, the
Kh values were smaller (1.2-1.5) than that of silica gels due to a smaller number of surface silanols
(~2OH/nm2) than that on the silica gel surface (~5OH/nm2). This difference is a consequence of
the flame synthesis of nanosilicas at high temperatures (>1300 K) in contrast to silica gels
synthesized in the liquid reaction media. This led to greater values of Qdecane for nanosilicas caused
by stronger dispersion interactions with more open surface of nanosilicas than for silica gels with
similar values of SBET. Thus, nanosilicas were less hydrophilic (it adsorbs only 1-3 wt.% of water
from air) than silica gels. Additionally, the surface area of nanosilicas (possessing the textural
porosity caused by aggregates < 1 m in size of nonporous nanoparticles < 10 nm in size) is more
accessible for relatively large molecules of n-decane (Fig. 74) than that of silica gels representing
rigid mesoporous particles of sizes in the millimeter range, i.e. possessing narrow and long
mesopores.
86
Fig. 75. The relationship between the heat of immersion of nanosilicas and silica gels in water or decane vs. the
specific surface area (SBET) of silicas.
Dispersion interactions of decane molecules with the silica surface were weaker than the hydrogen
bonding of water molecules to the surface silanols (per mole of silanols), but were greater per mole
of adsorbates due to a much larger size of C10H22 than H2O [158]. Additionally, the narrowest
pores of silica gels are less accessible for decane molecules larger by approximately ten times than
water molecules. Adsorbed water molecules form clusters around surface hydroxyls that could
reduce the effects of different numbers of surface hydroxyls per nm2 of surface of nanosilicas and
silica gels [158]. Therefore, the curves of Qwater vs. SBET (Fig. 75) have similar courses for
nanosilicas and silica gels. The difference in the values of Kh for these silicas is mainly due to the
difference in the values of Qdecane. For example, Qdecane = 34.4 and 17.4 J/g for A-300 and Si-100,
respectively, having similar values of SBET (Table 1). Stronger interactions of n-decane with less
hydrophilic nanosilicas than with more hydrophilic silica gels could explain the difference in the
temperature behavior of 1H NMR signal intensity of n-decane bound to nanosilica A-400 and silica
gel Si-60 (Figs. 70 and 72) [158].
DSC thermograms of cooling of n-decane, adsorbed onto different silica gels in the
amounts greater than the pore volume of silicas, demonstrate the presence of two crystallization
peaks (Fig. 76) [158]. A low-temperature crystallization peak shifts toward higher temperature
with increasing pore size (Fig. 76). The shift of a low-temperature melting peak with respect to a
low-temperature crystallization peak depended on the pore size distribution of silica gels because
melting of larger frozen structures bound in larger pores occurred at higher temperatures (Fig.
76b,c) since the heating rate of 5 K/min used was not too small to prevent certain delay observed
in the melting process. Therefore, the hysteresis loop could be broader for decane located in
broader pores due to the formation of larger frozen structures (quasi-crystallites). Similar effects
were observed (Figs. 50-52, Tables 9 and 10) for different amounts of decane adsorbed onto the
same silica because it filled only narrower pores at a lower adsorbed content. A narrow peak at T
close to Tf (Fig. 76) was due to freezing of nearly bulk decane weakly bound or unbound to silica
gels because decane was adsorbed in the amounts greater than the pore volume. Consequently,
there was a fraction of decane weakly bound or unbound to the silica surface. The second peak,
which is located at lower temperatures and broader than the first one but it depends on the pore
size of silica gels, appears due to freezing of decane bound in mesopores. Melting curves of
decane at T < Tf,d, i.e. for alkane located in pores, and water crystallites at T < Tf,w confined in
narrow pores of Si-40 were similar in shape as they demonstrate broad peaks but different in
position (Fig. 76a) because these adsorbates were of very different nature and characterized by
different types of intermolecular bonding affecting Tf. Water did not show the melting peak at T
Tf,w (Fig. 76a, curve 4) because all water (h = 0.35 g/g) was bound in narrow pores of Si-40.
During freezing of the sample (Fig. 76a, curve 3), a narrow crystallization peak was observed at
ca. 273 K (i.e. there was strongly associated but weakly bound water) in addition to a broad
crystallization peak at 230-250 K for water strongly bound in pores. During melting of ice bound
87
in pores of Si-40, the heterogeneity of adsorbent and step-by-step melting of ice in broader pores
resulted in the absence of a sharp melting endotherm at temperatures close to Tf,w (Fig. 76a, curve
4). Residual water bound to Si-60 was characterized by a small crystallization peak at 256 K (Fig.
76b, curve 1) [158].
Fig. 76. DSC thermograms of cooling-heating of (a-c) n-decane (curves 1 and 2), (a) water (curves 3 and 4), and (b, c)
1-decanol (curves 3 and 4) bound to silica gels for (a) Si-40 (Cd = 0.61 g/g, h = 0.35 g/g), (b) Si-60 (Cd = 1.15
g/g, 1-decanol Cdl = 0.45 g/g), and (c) Si-100 (Cd = 1.11 g/g, Cdl = 1.11 g/g) at the cooling-heating rate of 5
K/min, silica sample mass was 12.8-25.6 mg.
1-Decanol (Tf,dl = 279.5 K, less polar than water but more polar than decane) bound to Si-
60 or Si-100 (Fig. 76b,c) demonstrated two freezing peaks and two melting peaks [158]. Despite a
relatively low amount of decanol bound to Si-60 (0.45 g/g), a sharp freezing exotherm was
observed close to its freezing point Tf,dl (Fig. 76b). Decanol bound to Si-100 (Cdl = 1.11 g/g)
demonstrated broader peaks than that bound to Si-60. This difference is similar to that for decane
adsorbed onto these silicas [158].
Thus, there are several factors governing the temperature behavior of polar (water), weakly
polar (1-decanol), and nonpolar (n-decane) compounds bound in pores or located out of pores of
different silica gels Si-40, Si-60, and Si-100 that differs for nanosilicas. The first one is the pore
size, since the temperature changes in the structure of bound liquids are more homogeneous for
adsorbates located in narrower pores with a narrow PSD. The second one is the amount of
adsorbate, which could be strongly or weakly bound to the surface of adsorbent or unbound (bulk),
located out of pores. Additionally, the heating-cooling rate and the temperature gradient could
88
affect the interfacial and temperature behavior of adsorbates, especially if adsorbents have broad
PSD (such as FMO) that caused the formation of inhomogeneous structures (liquid – solid) of
adsorbates at different temperatures in pores of different sizes. The temperature dependence of the
behavior of liquids confined in pores is the basis of several porosimetry methods, including DSC
thermoporometry [53,158,212-214].
The Gibbs-Thomson equation applied to DSC melting thermograms allows one to calculate
the PSD for silica gels (with the model of cylindrical pores since the errors of this model are
relatively small for silica gels [158]) using melting curves at T < Tf, i.e. using the low-temperature
DSC peaks. Melting curves of decane at T < 243 K (Fig. 76) were used to calculate the PSD with
the equation R = kGT/(T0,m – Tm) for cylindrical pore radius at kGT = 64.6 K nm, where T0,m and Tm
were the melting temperatures for pure bulk crystallites and confined in pores of radius R,
respectively. For the freezing point depression of decane confined in pores, these calculations
resulted in the PSD similar to the NLDFT PSD (Fig. 77) [158]. Notice that the DSC PSD have a
simple contour than the NLDFT PSD curves because the latter were calculated using a small value
of the regularization parameter fixed in the Quantachrome software. For Si-40, DSC PSD (Fig. 77,
curve 1) was zero at R = 1.0-1.5 nm in contrast to NLDFT PSD (curve 2). This difference could be
explained by two factors: (i) decane was poorly adsorbed in narrow pores at R < 1.5 nm due to
presence of residual water and air bubbles; and (ii) the used temperature was not enough low to
obtain a fraction of decane unfrozen in the narrowest pores (melting curve starts from nonzero
values, Fig. 76a, curve 2). Hence, the appropriate correspondence between the NLDFT and DSC
PSD for silica gels was evidence that the melting low-temperature peaks of decane were due to
alkane molecules located in mesopores of silica gels [158].
Fig. 77. Differential PSD of silica gels calculated using the NLDFT method (applied to nitrogen adsorption-desorption
isotherms with the model of cylindrical pores) and the Gibbs-Thomson equation for the freezing point
depression for n-decane confined in pores using DSC melting thermograms.
Maximal excess in the decane amount was for Si-60 (Vd = 1.58 cm3/g at Vp = 0.82 cm3/g)
[158]. Therefore, a marked high-temperature peak at T 263 K was observed for Si-60. However,
Si-60 has narrower pores than Si-100 (Fig. 77). Therefore, decane structures confined in
mesopores have maximal sizes for Si-100 but the frozen decane structures out of pores could be
larger for Si-60. A minimal Cd value (slightly higher than the pore volume) was for Si-40,
possessing the narrowest pores (Fig. 77) with the highest adsorption potential [158].
The differences in the PSD of silicas and the decane excess led to different hysteresis loops
for integrated heat flow ((T)) [158]
min
( ) | ( ) |
T
T
T F T dT , (29)
during cooling (exo-effects due to crystallization) and heating (endo-effects due to melting) of
decane bound to silicas (Fig. 78).
89
Fig. 78. Integral heat as a function of temperature (T) (with subtraction of a nonlinear baseline and normalization to
unit) for exo (cooling, crystallization) and endo (heating, melting) effects for decane confined in pores of (a)
Si-40, (b) Si-60, and (c) Si-100.
For easier comparison of the (T) graphs, the integration in Eq. (29) of both freezing and melting
curves (after subtraction of a nonlinear baseline) for the modulus of the heat flow |F(T)| was
carried out from a minimal temperature (~170 K) toward higher temperatures (~290 K). The (T)
curves for decane bound to Si-60 and Si-100 were similar (Fig. 78b,c) but strongly differ from that
for alkane bound to Si-40 (Fig. 78a). This difference could be explained not only by the difference
in the PSD of these silicas but also by the smallest amount (0.61 g/g) of decane adsorbed on Si-40.
A nearly vertical step on the (T) curve of freezing of bulk (or nearly bulk) decane at ca. 240-243
K was greater for Si-60 (with maximal excess of adsorbed alkane) than for Si-100 or Si-40 (Fig.
78) due to larger decane structures formed out of pores of Si-60, despite pore sizes of Si-100 were
greater than that of Si-60 (Fig. 77). Larger structures of nearly bulk decane gave a greater
exothermic effect during freezing, since a greater endothermic effect was observed during melting
(Figs. 76-78). Notice that a certain inclination of a step around Tf was observed on the melting
curves, especially for Si-60 and Si-100 (Fig. 78b,c) due to certain delay in melting of large frozen
structures during heating at T > Tf [158].
The DSC results allowed one to explain broad hysteresis loops for the integral 1H NMR
intensity for decane, both bulk and confined between silica nanoparticles of A-400 (Fig. 70), as
well as for decane bound to Si-60 (Fig. 72) depending on the excess of the alkane [158]. The effect
of decane excess for freezing-thawing during 1H NMR measurements could result in the formation
of large quasi-solid structures (quasi-crystallites) which were not observed in the 1H NMR spectra
90
recorded at a narrow bandwidth (20 kHz) because of a relatively short time of transverse relaxation
of an immovable fraction (despite appearance of a fraction of unfrozen bulk decane at T > Tf). For
DSC measurements (Fig. 76), a relative significance of this effect was smaller (since only a small
peak is at T 263 K) due to a smaller decane excess and a certain physical difference in the
phenomena studied by the NMR (changes in the molecular mobility with temperature) and DSC
(release or absorption of heat with decreasing or increasing temperature) methods [53,158].
Thus, combined NMR/DSC/thermoporometry/nitrogen adsorption/quantum chemistry
analysis [53,158] of the temperature and interfacial behavior of n-decane, water and 1-decanol
bound to nanosilica and silica gels showed that a portion of decane or decanol remained frozen
(quasi-solid) at temperature higher than the freezing point of bulk liquids during heating of
samples. Additionally, a fraction of adsorbates remained unfrozen at temperatures below the
freezing point during cooling due to the confined space effects. However, freezing and melting
curves did not coincide and the hysteresis loop width depended on the NMR (or DSC) experiment
conditions [158].
For n-decane adsorbed onto silica gels Si-40, Si-60 and Si-100, two-four freezing or melting
peaks were observed on the DSC thermograms over the 170-300 K range during cooling or
heating of samples that depended on the pore size distribution of silicas and the amount of decane
[158]. Integrated heat flow curves vs. temperature demonstrated the hysteresis loops similar to the
effects observed by the NMR method; however, the shapes of the hysteresis loops differed because
of the difference in the observed effects: changes in molecular mobility (NMR) or absorption
(melting) and release (freezing) of heat (DSC). According to 1H NMR data, a portion of n-decane,
which is in quasi-crystalline state characterized by fast molecular exchange (short transverse
relaxation time) and not observed in the NMR spectra, was greater during heating of samples than
a portion of decane frozen during cooling at temperature close to the freezing point of bulk liquid
(Tf). A similar effect appearing in the DSC endotherms for samples heated at the same heating rate
was smaller than that in the NMR measurements due to different excess in the decane amounts and
the difference in the phenomena studied by the low-temperature 1H NMR (changes in molecular
mobility with temperature) and DSC (release or absorption of heat with decreasing or increasing
temperature) methods. The enhanced confined space effects for nanosilica A-400 compacted
toward twice greater bulk density (0.15 g/cm3 instead of 0.07 g/cm3 for the initial A-400) resulted
in an increase in a portion of unfrozen n-decane at T > Tf. For a small content of water co-adsorbed
with n-decane (of a large content > Vp), water displaced by decane into narrower pores where it
could not form large structures. Therefore, this water is weakly associated water. The studied
regularities in the temperature and interfacial behavior of water and n-decane or 1-decanol co-
adsorbed onto different silicas could be used to explain features of the work and the efficiency of
adsorbents applications at different temperatures and concentrations of adsorbates [53,158].
Treated FMO
Water adsorbed on the initial silica A-300 was characterized by two 1H NMR signals (Fig.
79a) at δН = 2.8-4.8 (SAW) and 1.5 ppm (WAW) [215]. The presence of two signals suggests the
difficulties in the exchange processes between the corresponding water structures. This could be
explained by the formation of small clusters and nanodomains of water bound in narrow pores
between adjacent particles (WAW) and broader voids between several neighboring particles in
aggregates (Figs. 26 and 80), respectively, as it was described above. Similar values of δН were
observed for water adsorbed onto porous silicas and assigned to the hydrogen-bound and non-
bound protons in the water molecules [53,158]. However, it should be noted that the topology of
pores in silica gels or ordered mesoporous silicas and voids in the nanosilica powders were
strongly different. Therefore, conditions of water clusterization were much better for nanosilica (at
low amounts of water) than for silica gels [53,213-215].
91
Fig. 79. 1Н NMR spectra recorded at different
temperatures of water (5 wt.%) adsorbed on
А-300 (a, b) initial and (c) MCA in (a)
CDCl3 and (b, c) ССl4 media.
Fig. 80. (a) Temperature dependence of content of
unfrozen water, Cuw (total water content 5
wt.%) adsorbed to initial A-300 and MCA A-
300; (b) cluster size distributions of this water
(CCl4 medium); and (c) PSD (MND method)
for initial and MCA treated A-300.
Signals at δН = 0 and 7.16 ppm were due to tetramethylsilane and CHCl3 (admixture in
CDCl3), respectively. According to the 1H NMR signal classification [53,213], observed water
signals (Fig. 79a) could be assigned to strongly and weakly associated waters, respectively. The
SAW signal shifted toward the strong magnetic field with increasing temperature. This could be
explained by breakage of a portion of the hydrogen bonds, decomposition of large polyassociates
of unfrozen water and stabilization of smaller structures of unfrozen water in narrow pores [215].
At T < 250 K the SAW signal intensity decreased due to freezing of this water with
lowering temperature [215]. This process was accompanied by signal broadening because of
decreased mobility of adsorbed water at lower temperatures. A certain broadening of the SAW
signal at T > 250 K was due to such factors as accelerating exchange processes between SAW and
WAW and an increase in nonuniform broadening due to appearance of solvent vapor babbles
formed with elevating temperature. The latter could lead to a decrease in the SAW signal intensity.
The WAW signal decreasing with lowering temperature was narrower than the SAW signal
(Fig. 79a) [215]. Two split signals of different intensity were observed for WAW because of
nonuniformity of this water. At T > 250 K the WAW signal intensity increased due to decreased
content of SAW. Two signals of WAW (Fig. 79a) could be attributed to water dissolved in
92
chloroform (the right signal of lower intensity) and water bound to silica nanoparticles (the left
signal of greater intensity).
NMR-cryoporometry results showed the difference in the organization of water clusters
adsorbed onto the initial and MCA A-300 (Fig. 80) because of the difference in the PSD of these
silicas [215]. Despite an increase in contribution of smaller water clusters bound to MCA A-300,
the contact surface area between water and silica (determined by integration of the fS(R)
distribution functions) was larger for the initial silica (78 m2/g) than for MCA A300 (56 m2/g).
Notice that 5wt.% H2O was too low content to form a continuous water layer at the silica surface
and to fill all textural pores. Therefore, water was strongly bound (freezing starts at T < 250 K,
Fig. 63a) and could cover only a portion of the silica surface and filled only a small portion of
pores, mainly narrow mesopores and nanopores, forming clusters of radius R < 3 nm (Fig. 80b)
[215]. The behavior of water bound by silica initial and after the MCA of the water-wetted powder
for 1 and 6 h was studied by the 1H NMR spectroscopy with layer-by-layer freezing-out of bulk
and bound water at 200 < T < 273 K in different media (air, water, chloroform) (Fig. 81) [75,215].
It was shown that the longer the MCA causing narrowing voids between silica
nanoparticles, the stronger bound the interfacial water and the stronger the CDCl3 effects on this
water [215]. For low hydration h = 0.11 g/g, the displacement of bound water by chloroform into
larger voids became more visible at T < 260 K since the amounts of strongly bound water [53]
frozen at T < 260 K decreased differently. For strongly wetted powders the 1H NMR spectra of
unfrozen water were observed at lower temperatures to 200 K with the single signal at chemical
shift of the proton resonance H = 4.5-5.5 ppm close to that of bulk water [53,215]. For the
aqueous suspension of the initial silica, the 1H NMR signal was observed at T 220 K [215].
Water in both cases was strongly associated [53] since the H value was close to that of bulk water.
A portion of adsorbed water frozen close to 273 K could be assigned to weakly bound water
(WBW) at |G| < 0.5 kJ/mol. At h = 0.11 g/g, the spectra shape depended on the tMCA value and
medium (air or CDCl3). Water adsorbed on all samples being in the air medium was strongly
bound since the signal decreased only at T < 250 K. For the chloroform medium, a portion of
adsorbed water was weakly bound since the signal strongly decreased at T < 265 K. The signal
shifted toward the weak magnetic field with lowering temperature because of a stronger influence
of the silica surface on closer located water layers frozen at lower temperatures. Additional signals
at H 7 and 0 ppm were due to H in chloroform (as an admixture in CDCl3) and trimethylsilane
(added to chloroform as a reference compound), respectively. A signal of weakly associated water,
WAW appeared at H = 1.0-1.5 ppm [53,213] for the sample in CDCl3 medium. This signal could
be assigned to water dissolved in chloroform or forming small clusters in contact with both
chloroform and the silica surface in narrow voids. The amounts of this water decreased for MCA-
treated samples because of changes in the textural porosity and associativity of water molecules
adsorbed in different voids. At temperatures close to 273 K the WAW signal was split into two
signals due to structural nonuniformity [53] of the silica-water-chloroform system (Fig. 81) [75].
SBW (frozen at T < 260 K, Fig. 81) at h =0.11 g/g was located in narrow voids at R < 5 nm
between neighboring primary particles [75]. The pores filled by unfrozen water were narrower in
MCA-treated samples. Chloroform could displace bound water into larger voids since this
provided a smaller boundary area between larger water structures and weakly polar chloroform
located in voids (note that water and chloroform are immiscible liquids in the bulk). This
displacement depended on tMCA since the larger the bulk density of the powder (longer MCA
time), the smaller the displacement of the water. For weakly hydrated initial silica at a low
hydration degree h = 0.11 g/g, the specific surface area determined with NMR-cryoporometry Scryo
< SBET (Table 7) because water could cover only a portion of the particle surface at h = 0.11 g/g
[53]. This portion decreased in the CDCl3 medium because of the displacement of bound water
and the Scryo value decreased (Table 7). For MCA-treated powders at h = 1.2 and 1 g/g, Scryo < SBET
was due to the same reason and possible remaining of air babbles in narrow voids. For the initial
silica at h = 40.7 g/g, Sc,mes > Sc,mic but at h = 3 g/g, Sc,mes < Sc,nano (Table 13) because of attractive
93
interactions between neighboring nanoparticles through water bridges in the wetted powder and
repulsive interactions between charged particles in the suspension. Therefore, DLS showed mainly
individual nanoparticles for the initial silica. The Sc,meso value was small at h = 0.11 g/g since water
in the form of clusters located only in narrow voids and did not form coverage of the total surface
of particles. In the suspension or strongly hydrated powder, the Sc,mes value strongly increased,
especially for the initial silica [75].
Fig. 81. Differential changes in the amounts of unfrozen water (dCuw/dT) as a function of temperature (bottom x-axis)
and changes in the Gibbs free energy (G, top x-axis) in (a) weakly wetted samples (h = 0.11 g/g) initial (b =
0.045 g/cm3) and treated for 1 h (b = 0.215 g/cm3) and 6 h (b = 0.394 g/cm3) without and with CDCl3
medium and (b) strongly hydrated (h = 3, 40.7, 1.2 and 1 g/g, respectively) silicas; (c, d) differential functions
with respect to the sizes of unfrozen water structures for (c) weakly and (d) strongly hydrated silicas.
Table 13. Characteristics of water unfrozen at T < 273 K and bound to initial and wet-MCA-treated A-300 nanosilica
estimated by NMR-cryoporometry [75].
tMCA
h
b
g/cm3
h
g/g
Scryo
m2/g
Sc,nano
m2/g
Sc,meso
m2/g
Vc,nano
cm3/g
Vc,mes
cm3/g
S
J/g
G
kJ/mol
Medium
0 0.045 0.11 159 111 48 0.045 0.055 10.3 3.03 Air
0 0.045 0.11 69 56 13 0.022 0.077 6.3 2.96 CDCl3
0 0.045 3.0 322 207 115 0.085 1.367 32.2 3.08 H2O
0 0.045 40.7 281 80 201 0.031 2.136 26.6 3.12 H2O
1 0.215 0.11 188 131 57 0.053 0.076 10.8 2.87 Air
1 0.215 0.11 142 101 41 0.024 0.076 9.4 3.00 CDCl3
1 0.215 1.2 123 52 71 0.022 0.772 15.6 3.21 H2O
6 0.394 0.11 332 332 0 0.100 0 16.7 3.21 Air
6 0.394 0.11 100 94 6 0.040 0.058 8.3 2.71 CDCl3
6 0.394 1.0 124 65 59 0.028 0.676 17.9 3.09 H2O
Note. G corresponds to the changes in the Gibbs free energy of a water layer closely located to the silica surface; S
is the modulus of overall changes in the Gibbs free energy of all bound water.
94
The s values varied between 32.2 J/g (98 mJ/m2) and 6.3 J/g (19 mJ/m2) for initial silica at
h = 3 and 0.11 g/g (in CDCl3) were relatively small [75]. This was due to relatively small amounts
of the surface silanols on fumed silica synthesized at high temperature [1-23]. Therefore, this silica
can adsorb ~1.6 wt.% of water from air (desorbed at T < 150 oC), and only 7.9 % of the surface
area was occupied by the adsorbed water molecules. More hydrophilic fumed silicas could adsorb
5-7 wt.% of water from air and have s = 150-200 mJ/m2 [53]. The s values for MCA-treated
silica were smaller than that of suspended or strongly hydrated initial silica (Table 13). This result
suggested that the number of surface silanols could insignificantly change on MCA but particles
were strongly aggregated that reduce total amounts of bound water since Cuw < h (1.2 or 1 g/g) and
the number of silanols accessible for adsorbed water [75].
The low-temperature 1H NMR spectroscopy used dealt with static samples to avoid signals
of solid adsorbates and silanols of silica [53] that contribute the spectra recorded using the magic
angle spinning (MAS) technique [216-225]. However, the difference in the MAS and non-MAS
1H NMR spectra of water bound to fumed silica could be small [104].
Addition of n-decane to wetted precipitated silica Syloid 244 at h = 0.3 g/g resulted in the
displacement of a portion of water into larger pores (Table 14 and Fig. 82) [91]. However, the
associativity of water in larger pores increased that resulted in a downfield shift of the water signal
[91]. Besides the water associativity effect on the value of the chemical shift of proton resonance,
H, stronger interactions of water with silica were observed in narrower pores [53,91,158]. For
example, MCA (6 h) A-300 was characterized by increased contribution of nanopores. Therefore,
the values of H of water (h = 0.1 g/g) bound to MCA A-300 were greater than those of water
bound to Syloid (h = 0.3 g/g) [91].
Fig. 82. (a, b) Amounts of unfrozen water vs. temperature (a) initial and (b) normalized by dividing by the total water
content; (c, d) pore size distributions calculated using NMR cryoporometry (c) differential and (d)
incremental PSD.
Note that partial freezing of water in pores changed the confined space effects for another
fraction of water remained in unfrozen state [91]. This could lead to greater contribution of
95
nanopores estimated from the Cuw(T) function than from nitrogen adsorption. This difference was
well observed for Si-100 at h = 0.5 and 0.59 g/g (Table 14, Fig. 83c). The PSD from the NMR
cryoporometry has peaks absent in the PSD from nitrogen adsorption (Fig. 83c). This effect was
less visible for Syloid (Fig. 83a) or A-300 (Fig. 83b) having less ordered porosity with broader
PSD than silica gel Si-100 has. Additionally, the secondary structures of A-300 or Syloid 244
could be more easily rearranged during interaction with water/ice than globules of Si-100 [91].
Table 14. Characteristics of water unfrozen at T < 273 K and bound to silicas studied estimated by NMR-
cryoporometry [91].
Sample h
g/g
Suw
m2/g
Suw,nano
m2/g
Suw,meso
m2/g
Vuw,nano
cm3/g
Vuw,meso
cm3/g
S
J/g
G
kJ/mol
<Tm>
K
Syloid 0.3 69 47 22 0.020 0.197 9.2 3.05 252.8
Syloid (+decane) 0.3 48 19 29 0.008 0.253 7.3 2.77 259.7
Syloid 0.9 24 0.1 20 0.0 0.298 3.2 2.27 268.2
Si-100 0.5 104 58 45 0.024 0.448 17.1 3.10 255.6
Si-100 0.59 275 239 34 0.104 0.401 30.4 3.15 244.0
A-300 0.1 159 111 48 0.045 0.055 10.3 3.03 223.0
A-300 3.0 322 207 115 0.085 1.367 32.2 3.08 263.3
A-300 MCA 1.0 124 65 59 0.028 0.676 17.9 3.09 261.3
Note. h is the hydration degree of silicas (amounts of water in gram added per gram of dry silica); ΔG is the changes in
the Gibbs free energy of water layer closely located to a surface of intracellular structures; γS is the modulus of the
total changes in the Gibbs energy of bound water unfrozen at T < 273.15 K; Suw is the specific surface area in contact
with unfrozen water, Suw,nano and Vuw,nano, Suw,meso and Vuw,meso were the specific surface area and pore volume of
nanopores at R < 1 nm and mesopores at 1 nm < R < 25 nm, respectively, filled by unfrozen water; and
0,
min
0,
min
)(
/
)( mm T
T
uw
T
T
uw
m dT
dT
TdC
TdT
dT
TdC
T is the average melting temperature, Tm,0 is the melting temperature
of individual bulk ice, Tmin correspond to the temperature at which Cuw = 0 [91].
However, at h = 1 g/g, contribution of water located in nanopores decreased (Table 14, Suw,nano and
Vuw,nano). A similar difference was observed for 1H NMR spectra of water bound to silica gel
(narrow PSD, greater downfield shift) and Syloid (broad PSD, smaller downfield shift) (Fig. 26)
[91].
96
Fig. 83. Comparison of the incremental PSD calculated using the amounts of unfrozen water vs. temperature (NMR
cryoporometry) and the amounts of adsorbed nitrogen vs. pressure using the SCV/SCR method for (a) Syloid
244, (b) A-300, and (c) Si-100.
Comparison of the DSC thermograms of nonpolar (benzene, toluene, decane), weakly polar
(chloroform) and polar (water, DMSO) adsorbates bound to nanosilica A-300 initial and MCA 6 h
and Syloid 244 showed (Fig. 84) the hysteresis effects of freezing-melting [91]. Significant
differences in the temperature positions of freezing exotherms and melting endotherms were
observed for both nonpolar and polar adsorbates, and the maximal values were mainly observed
for Syloid (up to 70 oC for toluene, 28 oC (decane), 24 oC (benzene), 21 oC (water)). For
chloroform (~50 oC) and DMSO (~6 oC), the gaps were similar for all adsorbents studied. The heat
effects (Tables S15 and S16, H) for both exotherms and endotherms were smaller than that for
bulk liquids/solids, and the difference depended on the amounts of adsorbates. This difference
increased with decreasing amounts of adsorbates due to preventing of the crystallization of liquids
upon freezing in the interfacial layers and narrow pores, in which, maximal disturbance of the
structure of adsorbates was provided [53]. Among the studied adsorbates, the enthalpy of fusion at
the freezing point was maximal for water (333.5 J/g) and minimal for toluene (71.8 J/g), and the
hysteresis effect was much greater for toluene than that for water for all FMO studied. Note that
the hysteresis effect was smaller for polar DMSO and water than that for nonpolar adsorbates. This
result could be explained by stronger interactions between an adsorbent surface and nonpolar
adsorbates than the molecule-molecule interactions that results in stronger changes in the structure
(organization) of organic liquids bound to silicas [91].
97
Fig. 84. DSC thermograms for (a) benzene, (b) toluene, (c) n-decane, (d) chloroform, (e) water, and (f) DMSO
interacting with Syloid 244 and A-300 during cooling and heating runs at a cooling-heating rate of 20 oC/min
for toluene and chloroform and 10 oC/min for other adsorbates.
Typically, the freezing exotherms were narrower than the melting endotherms (Fig. 84) [91].
However, if the amount of adsorbate was relatively small (e.g., 3.878 mg of toluene per 1.834 mg
of Syloid) that the freezing exotherm could be absent (Fig. 84b). For A-300, the amount of toluene
was greater (8.127 mg per 1.578 mg of A-300) that resulted in a large and sharp freezing exotherm
(Fig. 84b). For some adsorbates such as water (Fig. 84e), benzene (Fig. 84a), and decane (Fig.
84c), there were exotherms at T > Tm that could be explained by changes in the structure of
unfrozen liquids bound to silicas with temperature before the formation of a solid phase [53,91].
Rearrangement of silica nanoparticles in the secondary structures can occur due to
interaction with water [53]. One of the sequences of this effect could be changes in the PSD
calculated using the DSC melting thermograms for water (ice) bound to silicas in comparison with
the PSD for the silica powder estimated from the nitrogen adsorption/desorption isotherms (Fig.
85) [91]. The maximal difference was observed for the initial A-300 (Fig. 85a). It was smaller for
MCA (6 h) A-300 (Fig. 85b). It was minimal for precipitated silica (Fig. 85c) because of features
of its synthesis. Therefore, adsorption of water could produce minimal changes in the textural
98
characteristics of precipitated silica. However, wetting-drying-wetting could affect the structure of
Syloid too (Fig. 85c) due to swelling of particles, decomposition of a part of inter-particle bonds,
and formation of new inter-particle bonds [91].
Fig. 85. Comparison of the IPSD calculated
from nitrogen adsorption-desorption
isotherms and the DSC melting
thermograms of water (ice) bound to
(a) initial A-300, (b) MCA 6 h A-300,
and (c) Syloid 244.
Thus, investigations of fumed silica A-300 initial and mechanochemically activated,
precipitated silica Syloid 244 and silica gel Si-100, which have a similar specific surface area
(~330-350 m2/g), using infrared spectroscopy, DSC, NMR, and adsorption methods, showed some
similarities and differences in their properties appearing in different media under different
conditions [91]. Mechanochemical activation of A-300 in a ball-mill or a microbreaker affected
the structure of aggregates of nanoparticles and agglomerates of aggregates, which became more
compacted (since the bulk density increased from 0.045 to 0.394 g/cm3), and slightly changed the
morphology of nanoparticles. The latter appeared as changes in the shape of the IR band of the Si-
O stretching vibrations at 1100 cm1. Any treatment of ‘soft’ nanosilica affected the interfacial
behavior of polar and nonpolar adsorbates because significant rearrangement of secondary
particles changes the freezing-melting point depression of bound adsorbates. Precipitated silica
Syloid 244 and silica gel Si-100 demonstrated higher stability of the properties than fumed silica
upon interaction with water and other adsorbates in a broad range of temperatures [91].
Clusterization of adsorbates bound in pores caused diminution of the heat effects during
phase transition (freezing and fusion). The freezing point depression and increasing melting point
caused significant hysteresis freezing-melting effects for adsorbates bound in pores, that was larger
for nonpolar compounds than for polar ones due to the differences in such interactions as
molecule-molecule and molecule-silica surface affected by stronger confinement in narrower pores
[91].
99
The study [91] showed that fumed and precipitated silicas could be more sensitive than silica
gel to external actions such as wetting-drying and MCA. These effects should be considered upon
applications of the materials in different media.
Cryogelation and confined space effects
For high-pressure cryogelation in cryo-bombs, fumed silica A-300 (SBET = 330 m2/g) was
used as the starting material [63]. Aqueous suspensions of A-300 (5, 10, 15 and 20 wt.%) were
prepared using doubly distilled water and sonicated (22 kHz) for 5 min. The resulting dispersion
was then frozen at 259, 208, or 77 K in thick-walled-stainless-steel reactors at pressures from 1 to
1050 atm, caused by ice formed in the frozen suspension (~10-15 ml) for 24 h. The pressure
(estimated according to [61,62]) was controlled using partial (450 atm) or complete (~1000 atm)
filling of a freezing bomb with stainless steel. Then cryosilica (CS) samples were placed in a glass
dish and dried in air at room temperature for 2-5 days to an air-dry state. The final materials were
in the powder state [46,47,63].
Cryosilicas CS20 (CA-300 = 20 wt.% in the aqueous suspension frozen where ice produced a
pressure of 450 atm) and CS20m (prepared at ~1000 atm) were studied at different degrees of
hydration (h = 0.08, 0.75, 1.1 and 5.67 g of water per gram of dry silica) [46,47,63]. Hydration
levels were controlled by addition of a known amount of water to a powder sample placed into an
ampoule. Samples were equilibrated at 293 K for 1 h before NMR measurements. Experiments
with methane were performed at a slightly increased pressure because an ampoule was linked to a
rubber vessel with methane at 1.1 atm. Liquids (n-decane, chloroform, dimethylsulfoxide
(DMSO), acetonitrile, water, or concentrated aqueous solutions of H2O2, HCl or CF3COOH) were
added to fill the total volume of a NMR ampoule with hydrated silica (the weight of dry silica was
~100 mg). After addition of hydrophobic decane, the sample was heated to 320 K for 1-2 min to
remove air bubbles and to completely fill free pores of hydrated silica powder [46,63].
The freezing of aqueous suspensions of nanosilica in high-pressure reactors resulted in dried
powdered materials that were very different from silicas treated in vapor phase water at high
pressures and temperatures [37,38,53,91,226-228], at high temperatures and normal pressure in air
[90,91], or at room temperature and normal pressure followed by air drying [68,229-231]. The air-
dried high-pressure cryosilicas were in the powder state with agglomerates of 1-20 m in size (Fig.
86), which was far more compacted than in the fumed silica starting material [46,63].
Fig. 86. Microphotograph (Primo Star optical microscope, Carl Zeiss) of CS20m powder prepared at 1000 atm (scale
bar 10 m).
100
Fig. 87. HRTEM images of (a, b, c) initial A-300 and (d, e, f) CS20m at different magnifications (scale bars (d) 100
nm, (a, e) 50 nm, (b) 20 nm, (f) 10 nm, and (c) 5 nm).
The bulk density of the cryosilica powder, as well as of any suspended and dried nanosilicas,
increased approximately five-fold (Table 15, b), since b = 0.045 g/cm3 for the initial A-300
powder [63]. This was due to re-arrangement of nanoparticles in aggregates and agglomerates of
aggregates during suspension and subsequent drying. Primary particles became mobile in
sonicated aqueous suspensions including a large proportion of individual nanoparticles [53].
During drying of the suspensions, nanoparticles tended to form more compacted aggregates and
agglomerates than those formed in the flame (flow velocity > 20 m/s) during pyrogenic synthesis.
Typically, the pore volume (Table 15, Vp) determined from the adsorption of nitrogen at p/p0 =
0.98-0.99 was much smaller than the empty volume, Vem of the powder. For example, for initial
fumed silica (b = 0.045 g/cm3) Vem = 21.8 cm3/g, but Vp = 0.65 cm3/g was much smaller because
nitrogen could not effectively fill large pores, especially at R > 100 nm. For suspended-frozen-
dried silica CS20m b = 0.31 g/cm3 and Vem = 2.8 cm3/g but Vp = 1.28 cm3/g; i.e. the difference
between the Vem and Vp values strongly decreased after cryogelation. These textural changes were
readily observed in the HRTEM images (Fig. 87), IPSD (Fig. 88a-e) and nitrogen adsorption
isotherms (Fig. 88f) [63].
According to HRTEM images (Fig. 87), CS20m was composed of amorphous primary
particles more strongly compacted in aggregates (comp. Fig. 87a,e). These aggregates remained
101
porous with the textural porosity as voids between nonporous primary particles (Fig. 87d). The
shape of primary particles was not ideally spherical for both initial A-300 and CS20m.
Table 15. Textural (specific surface area SAr and SBET), bulk density (b) and adsorption (desorption of water in TG)
characteristics of silicas vs. silica content in the frozen suspension (CA-300), pressure and treatment
temperature (Ttr) [63].
Sample CA-300
(wt.%)
SAr
a
(m2/g)
SBET
(m2/g)
Vp
(cm3/g)
Snano
(m2/g)
Smeso
(m2/g)
Smacro
(m2/g)
Vnano
(cm3/g)
Vmeso
(cm3/g)
Vmacro
(cm3/g)
w
Initial 330 0.826 51 270 9 0.026 0.653 0.147 0.025
CS5 5 317 280 1.474 14 258 8 0.009 1.358 0.107 0.392
CS10 10 303 293 1.341 23 268 1 0.015 1.294 0.032 0.431
CS15 15 295 292 1.347 29 262 1 0.019 1.289 0.039 0.454
CS20 20 297 1.279 33 263 1 0.022 1.231 0.026 0.463
CS20m 20 285 1.277 31 253 1 0.021 1.220 0.037 0.458
CS5* 5 332 286 1.393 10 275 1 0.006 1.364 0.023 0.363
CS10* 10 326 289 1.298 15 272 2 0.010 1.247 0.041 0.459
CS15* 15 323 304 1.470 27 276 1 0.017 1.423 0.029 0.401
CS20* 20 306 303 1.443 20 281 2 0.013 1.402 0.028 0.404
CS5** 5 302 300 1.449 24 273 2 0.015 1.390 0.044 0.389
CS10** 10 318 303 1.445 23 278 2 0.014 1.390 0.041 0.405
CS15** 15 304 1.455 19 281 3 0.012 1.390 0.053 0.397
CS20** 20 317 296 1.408 36 257 3 0.023 1.328 0.056 0.451
Sample b
(g/cm3)
Intact waterb
(T < 200oC)
(mmol/g)
Total waterc
(T < 900oC)
(mmol/g)
Wd
(W200–W600)
(mmol/g)
cvoid ccyl cslit Pressure
(atm)
Ttr
(oC)
Initial 0.045 0.889 1.556 0.383 0.308 0.368 0.324 - -
CS5 0.235 0.619 3.417 1.780 0.159 0.611 0.230 450 -14
CS10 0.208 1.238 2.600 0.744 0.197 0.387 0.416 450 -14
CS15 0.228 0.745 1.397 0.537 0.194 0.268 0.538 450 -14
CS20 0.294 5.177 6.225 0.798 0.203 0.203 0.622 450 -14
CS20m 0.309 1.556 3.103 0.962 0.209 0.171 0.620 1000 -14
CS5* 0.285 12.742 13.545 0.708 0.182 0.695 0.123 1 -14
CS10* 0.294 2.596 4.035 1.005 0.227 0.380 0.393 1 -14
CS15* 0.349 1.977 2.924 0.731 0.163 0.458 0.379 1 -14
CS20* 0.282 2.290 3.047 0.609 0.151 0.555 0.294 1 -14
S5** 0.277 1.711 3.202 1.170 0.180 0.522 0.298 1 25
S10** 0.311 1.630 3.204 1.143 0.174 0.522 0.304 1 25
S15** 0.216 2.054 3.692 1.011 0.168 0.565 0.268 1 25
S20** 0.202 1.576 5.894 1.548 0.185 0.223 0.592 1 25
Notes. aSamples were heated at 100 oC for 2 h before measurements of the specific surface area. The amounts of water
desorbed during heating from b20 oC to 200 oC, c20 oC to 900 oC, and d200 oC to 600 oC [63].
However, the average value ( )d / ( )dd df d d f d d (i.e. the first moment of the
distribution function) was smaller for CS20m (<d> = 0.3766 nm) than for initial A-300 (<d> =
0.3979 nm). Thus, high-pressure cryogelation resulted in not only stronger aggregation of primary
particles (the bulk density b (Table 9) increased) but also in certain deformation of nanoparticles
(the <d> value decreased). Changes in the textural porosity of treated nanosilica reflected in
changes in the b value could be analyzed using the pore size distributions (Fig. 88) [63].
102
Fig. 88. Incremental PSD (SCV/SCR) for initial powder of A-300 and silicas prepared at CA-300 = (a) 5 wt.%, (b) 10
wt.%, (c) 15 wt.% and (d) 20 wt.% in the aqueous suspensions under different conditions: frozen at high
pressure (450 atm) or at maximum high pressure (1000 atm), frozen at normal pressure (*), suspended under
normal conditions (**), as well as initial A-300 powder, (e) NLDFT for cryosilicas prepared from 5-20 wt.%
suspensions and CS20m; (f) nitrogen adsorption-desorption isotherms for differently treated samples at CA-300
= 20 wt.%.
For treated silica powders, contribution of nanopores and narrow mesopores decreased,
while the contribution of mesopores at R = 10-30 nm increased compared with the initial fumed
silica A-300 (Fig. 88, Table 15) [63]. The primary particles were not strongly changed during
suspending-cryogelation-drying (Figs. 86 and 87). Therefore, the nitrogen adsorption isotherms
were practically the same at p/p0 < 0.8 (Fig. 88f, insert) but they characterized by different
hysteresis loops at p/p0 > 0.8 and different maximal adsorption at p/p0 0.99 (Fig. 88f).
Table 16. Textural characteristics of initial and differently treated nanooxides.
Oxide Medium Tt
(K)
Pt (atm) Note SBET
(m2/g)
Snano
(m2/g)
Smeso
(m2/g)
Smacro
(m2/g)
Vp
(cm3/g)
Vnano
(cm3/g)
Vmeso
(cm3/g)
Vmacro
(cm3/g)
w
gsPS300* 0.1M NaCl 293 1 Gel/NaCl 264 20 242 2 1.257 0.010 1.207 0.040 0.325
gswPS300* 0.1M NaCl 293 1 Gel/NaCl washed 276 24 251 1 1.313 0.012 1.263 0.038 0.331
csPS300* 0.1M NaCl 260 1000 Cryogel/NaCl 262 27 234 1 1.194 0.015 1.147 0.032 0.362
cswPS300* 0.1M NaCl 260 1000 Cryogel/NaCl
washed
279 29 248 1 1.291 0.016 1.244 0.031 0.376
PS300+AST(gSAST) water 293 1 Gel 147 57 55 36 1.029 0.017 0.374 0.638 0.042
PS300+AST(cSAST) water 208 1000 Cryogel 147 51 62 34 1.098 0.017 0.542 0.539 0.114
PS300+AST(gsSAST) 0.1M NaCl 293 1 Gel/NaCl 155 58 62 35 1.064 0.018 0.454 0.592 0.066
PS300+AST(csSAST) 0.1M NaCl 208 1000 Cryogel/NaCl 155 58 66 31 1.142 0.018 0.608 0.515 0.123
Note. *Gelation during 24 h, gelation of all other samples was during 12 h. Labels in dried samples correspond to cryogel (c), cryogel with NaCl (cs), gelation at room temperature
(gel, g) with NaCl (gs), and washed samples gelatinized with NaCl (csw or gsw).
103
104
Fig. 89. NLDFT PSD for (a) silica and (b) alumina: initial powder and differently treated nanooxide cryogelation in
0.1 M NaCl solution at 208 K, suspended in water or 0.1 M NaCl solution and dried at room temperature and
standard pressure and then washed (sample labels correspond to them in Table 15).
These results showed significant changes in packing of primary particles in aggregates. This
packing increased with increasing pressure during cryogelation. Additionally, the effect became
stronger with increasing silica content in the suspensions (comp. Fig. 88a-e) [63]. An increase in
the silica content in the suspensions resulted in an increase in mesopore contribution at R 10 nm,
and in a decreased contribution of mesopores at R 20 nm for high-pressure cryosilicas in contrast
to samples prepared at standard pressure.
The main PSD peak (both IPSD and dV/dR, Fig. 88) for HP-cryosilicas shifted toward
smaller R values with increasing silica content in the suspensions. Additionally, HP-cryosilicas, in
contrast to silicas frozen or non-frozen at normal pressure, the contribution of pores modeled by
slit-shaped pores (Table 15, cslit) increased due to the high-pressure effects. This effect was due to
certain changes in the shape of primary particles (Fig. 87), which were discussed above [63].
Contributions of nanopores (Figs. 88 and 89, Tables 15 and 16, Snano and Vnano) and
macropores (Smacro and Vmacro) were much smaller than that of mesopores (Smeso and Vmeso) for all
silicas treated in the aqueous media under different conditions [63]. However, in contrast to Vmeso,
the Snano, Smacro, Vnano, and Vmacro values for treated silicas decreased in comparison with the initial
silica. The Smeso values for high-pressure cryosilicas were smaller than that for the initial silica but
the Smeso values for normal-pressure cryosilicas (Table 15, label *) and non-frozen suspended-dried
silicas (label **) were larger than that of the initial silica. All these textural changes were due to
different rearrangement of aggregates of primary nanoparticles depending on treatment conditions
[53].
Fig. 90. IR spectra of PS100, cPS100, and gPS100 (1:100 mixture with KBr).
105
Fig. 91. IR spectra of (a) alumina samples: initial air-dry
(curve 1), cAl2O3 (2), gAl2O3 (3), and gsAl2O3
(4); and (b) ST samples: initial air-dry (curve 1),
cryogel prepared at 208 K and ~1000 atm pure
(2) and with 0.1 M NaCl (3), gelatinized at
standard conditions (4) and with addition of
NaCl (5) (1:100 mixture with KBr). All spectra
of ST samples were normalized to the
asymmetric mode of Si-OSi at 1100 cm-1.
Fig. 92. 1H NMR spectra of water (0.1 g/g) and n-decane
(0.15 g/g) bound to AST (dashed lines), gAST
(dot-dashed lines) and cAST (solid lines) in
media (a) air and (b) CDCl3 with addition of
TFAA (6 : 1).
Cryogeletion affects the textural and structural characteristics of FMO. The IR spectra of
the Si-O and Al-O stretching vibrations change (Figs. 90 and 91) due to particle compaction.
These changes affect the behavior of bound water (Figs. 92 and 93) that is also influenced by the
dispersion media (air, water, chloroform + TFAA). These effects appear in the pore size
distributions those filled by unfrozen water (Figs. 94 and 95).
106
Fig. 93. Amounts of unfrozen water as a function of
temperature for (a) cAST, (b) gAST, and (c)
AST located in different media.
Fig. 94. Incremental PSD calculated from 1H NMR
spectra of water (0.1 g/g) bound to AST,
gAST and cAST with addition of n-decane in
media (a) CDCl3 and (b) CDCl3 with addition
of TFAA (6:1).
Fig. 95. Differential PSD calculated from 1H NMR spectra of water (0.1 g/g) bound to AST, gAST and cAST (b-d)
with addition of n-decane in media (a, b) air, (c) CDCl3, and (d) CDCl3 with addition of TFAA (6:1).
DFT B3LYP/6-31G(d,p) calculations of the chemical shifts of 1H were performed for pure
water alone (Fig. 96a, curves 1 and 3) or with NaCl (curves 2 and 4). Larger structures were
calculated using the PM7 method (curves 5 and 6). These calculations show that the difference in
107
the f(H) shapes for pure water and water with dissolved Na+ and Cl ions decreases with
decreasing salt content. Additionally, the presence of dissolved Na+ and Cl- ions results in a
downfield shift (Fig. 96b, curves 4 and 5) in comparison with the system with NaCl
nanocrystallites (curves 2 and 3) bound to a silica surface. The H values for the SiOH groups
interacting with water molecules are greater than that for water molecules (comp. curves 2 and 3 or
4 and 5, Fig. 96b). These results explain a certain difference in the spectra of water bound to
PS300 without (Fig. 96a) and with NaCl (Fig. 96b).
Fig. 96. DFT B3LYP/6-31G(d,p) and semiempirical PM7 calculations of the 1H chemical shifts in pure water, water
with NaCl (a) alone or (b) bound to silica surface (silica pore with 1620 atoms).
Fig. 97. IR spectra of PS300 unmodified (curve 1) and modified at TMS = 27.2 % (2) and 37.2 % (3); the spectra of
modified silicas were normalized to the intensity of the band at (a) 1867 cm1 or (b) 1100 cm1 of unmodified
silica.
For high-pressure cryogels, there were two tendencies: (i) a certain decrease in the specific
surface area (Table 15, SAr and SBET) with increasing silica content in the suspension, and (ii)
nonlinear changes in the amounts of desorbed water (total water desorbed at T < 900 oC) with a
maximum for the material prepared at CA-300 = 5 wt.% (Table 15) [63]. Both tendencies remained
108
for cryogels prepared at normal pressure (Table 15, CS-X*). Thus, the desorbed water amounts
(intact and total water) change with increasing silica content in the suspension. This could be
explained by the formation of more compacted secondary structures with increasing CA-300 value,
changes in the sizes of pores filled by water, and the number of hydroxyls. Note that changes in
the specific surface area depended on the preparation conditions (C, P, T) much less than the pore
volume (Vp) and the content of desorbed water (Table 15). These results could be explained by
greater changes in the organization of primary silica nanoparticles in secondary structures (Table
15, Fig. 88), with less change in nanoparticle size (Fig. 87), e.g., observed during hydrothermal
treatment of silicas [89].
Fig. 98. Temperature dependences of the amounts of unfrozen water (Cuw(T)) and changes in the Gibbs free energy
(G(T)) of bound water as well as relationships between Cuw and G for CS20 in different dispersion media
at (a) h = 0.08 g/g (in methane, chloroform, or decane) and 5.67 g/g (concentrated 15% suspension); (b) h =
0.75 g/g and (c) h = 1.1 g/g (in air or decane).
Since CS20 and CS20m exhibited the largest structural changes among high-pressure
cryosilicas, low-temperature 1H NMR spectroscopy studies focused on these materials. Water was
either the sole probe compound or it was used as a co-adsorbate with different organics in various
dispersion media (Figs. 98-100, Table 17) for deeper insight into the interfacial behavior of
different adsorbates and co-adsorbates such as nonpolar methane and n-decane, weakly polar
chloroform, and polar water, DMSO, acetonitrile, H2O2, HCl, and CF3COOH [63]. CS20
(cryogelation at 450 atm) studied by NMR contained 0.08 g of water per gram of dry silica. To
analyze the interfacial behavior of bound water, the sample was studied in different dispersion
109
media [63]. In the methane dispersion medium, the spectra of bound water included a single signal
exhibiting a chemical shift (with respect to methane of H = 0 ppm), H 4-5 ppm [63]. This was
similar to the H value of bulk water. Consequently, this water could be assigned to strongly
associated water, SAW, with an average number of hydrogen bonds per molecule of 3-4, similar to
that in bulk water [53,63].
Fig. 99. Pore size distributions (a, c, d) differential and (b, d) incremental calculated using NMR cryoporometry
(calculated as size distributions of water structures unfrozen in pores) (IPSD for initial A-300 was calculated
using nitrogen adsorption-desorption isotherm, dot-dashed line), (c) NMR cryoporometry and TG
thermoporometry curves; (d) h = 0.75 g/g and 1.1 g/g (air or decane dispersion media) for CS20 in different
dispersion media; (a-d) CS20 and (e) CS20m samples.
Freezing of bound water occurred at low temperatures at T < 260 K (Fig. 98) because of the
freezing point depression for liquids confined in pores [53,63]. Since freezing of bound water
occurred at T < 260 K, all of this water could be assigned to strongly bound water, SBW. The 1H
NMR signal of methane at H 0 ppm had a weak dependence on temperature [63]. From the
intensity ratio of water (h = 0.08 g/g) and methane, the amount of adsorbed methane could be
estimated as small as ~8 mg/g at 280 K. This low adsorption of CH4 was due to filling of
nanopores (voids between adjacent nanoparticles) by water (Fig. 99, Table 17), which methane
could not displace. This was because methane has much weaker interactions with the silica surface
compared to water [53,63].
In the CDCl3 dispersion medium, bound water (h = 0.08 g/g) corresponded to SAW
characterized by H 5 ppm [63]. A fraction of this water became weakly bound, WBW frozen at
T > 260 K. Additionally, at higher temperatures, a weak signal was observed at H 1.8 ppm
corresponding to weakly associated water, WAW [53]. The signal at H 0 ppm corresponded to
110
tetramethylsilane used as a chemical shift standard. In contrast to methane, weakly polar
chloroform diminished the interaction of water with silica: the average melting temperature (Table
17, <Tm>), filling of mesopores by unfrozen water (Vmeso, Smeso), and changes in the Gibbs free
energy of the first layer (Gs) of bound water increased; while the free surface energy (S), filling
of nanopores (Vnano, Snano), and the SBW content decreased. Similar effects were observed for a
variety of adsorbents due to the displacement of water by weakly polar chloroform from narrow to
broader pores or out of pores altogether [53]. The main driving force of these effects led to a
reduction of the contact area between immiscible polar (water) and weakly polar (chloroform) or
nonpolar (decane) liquids. This reduction resulted in a free energy decrease because water-water
interactions are similar to water-silica interactions, but interactions of organic co-adsorbates with
silica are stronger than that of organics-organics or organics-water interactions [53]. The character
of re-organization of bound water under the action of chloroform was seen in changes in the PSD,
and in particular the IPSD, (Fig. 99b) when compared with water bound to CS20 in the methane
dispersion medium (Fig. 99a) [63].
Nonpolar n-decane more effectively displaced bound water than weakly polar chloroform
(Figs. 98 and 99, Table 17) [63].
Table 17. Characteristics of water bound to cryosilica powder CS20 and CS20m in different media [63].
h (g/g) Medium Сuw
s
(g/g)
Сuw
w
(g/g)
ΔGs
(kJ/mol)
ΔGw
(kJ/mol)
γS
(J/g)
0.08 Methane 0.08 - 3.50 - 9.1
0.08 CDCl3 0.06 0.02 2.97 0.47 5.2
0.08 Decane 0.08 - 2.23 - 5.7
0.75 Air 0.30 0.45 2.91 0.50 18.9
0.75 Decane 0.14 0.61 1.88 0.35 38.3
5.67 Water 0.80 3.27 2.34 0.30 67.4
0.1#a Air 0.08 0.02 2.77 0.39 7.8
0.1#b Air 0.08 0.02 2.78 0.44 8.0
1.1# Air 0.21 0.890 3.21 0.35 15.0
1.1# Decane 0.34 0.760 1.35 0.5 24.7
h (g/g) <Tm>
(K)
Suw
(m2/g)
Snano
(m2/g)
Smeso
(m2/g)
Vnano
(cm3/g)
Vmeso
(cm3/g)
0.08 216.2 89 71 18 0.03 0.05
0.08 241.8 45 21 24 0.01 0.07
0.08 240.0 65 0 65 0 0.08
0.75 259.7 115 32 83 0.01 0.67
0.75 248.5 110 0 110 0 0.74
5.67 260.9 301 0 301 0 4.10
0.1#a 229.6 41 36 5 0.02 0.07
0.1#b 229.0 42 36 6 0.02 0.07
1.1# 261.0 103 23 80 0.01 0.52
1.1# 258.4 84 0 78 0 0.67
Note. #CS20m, and other samples were based on CS20; CS20m stored at a 298 K and b255 K for a week [63].
This result could be explained by much greater dispersion interactions of larger decane
molecules with the silica surface than small methane or chloroform molecules. Notice that similar
effects of the water displacement were observed for small but polar molecules such as acetonitrile.
Adsorbed decane demonstrated two 1H NMR signals at 0.9 and 1.25 ppm [63] corresponding to
methyl and methylene groups, respectively. The water signal was weak due to the relatively small
amount compared to the decane content, and observed at H 1.8 – 2.5 ppm [63]. However, at h =
0.75 g/g, the water signal was much more intense but remained lower than the decane signal. The
water observed was a WAW-SAW mixture, with a major contribution of SBW since it was frozen
at T < 260 K. However, this water was more weakly bound to silica than that in either the methane
or chloroform media (Fig. 98). The WBW content increased with increasing water content (Table
17, Cuw
w), this led to an increase in the <Tm> value. This was because the relative content of SBW
111
decreased significantly from 75-100% (h = 0.08 g/g), 20-40% (h = 0.75 g/g) to 14% at (h = 5.67
g/g). Both highly hydrated powder CS20 at h = 0.75 g/g, and the concentrated (15%) aqueous
suspension at h = 5.67 g/g, exhibit strongly and weakly bound water (Table 17).
Water adsorbed onto CS20 at h = 0.75 g/g could be divided into SBW and WBW, frozen at
lower and higher temperatures, respectively. SAW (Table 17) was also found in the suspension
[63]. This effect could be explained by the formation of clusters and domains of water (Fig. 99d)
which were not spatially separated. When this sample at h = 0.75 g/g was in the decane dispersion
medium, the characteristics of bound water change quite a bit (Figs. 98b and 99d, Table 17).
WAW and two types of SAW appeared, and the PSD became bimodal.
Notice that even at a low adsorbed water content (h = 0.08 g/g), There was a broad PSD
(Fig. 99a,b) including clusters at 1 < R < 2 nm and domains at 2 < R < 20 nm [63]. These structural
features exhibited traces of the IPSD (determined from the nitrogen adsorption isotherm)
characteristic of the initial fumed silica (Fig. 98b, dot-dashed line) which contained mesopores at 2
< R < 10 nm. However, smaller and larger pores (voids between nonporous nanoparticles) differed
because of the rearrangement of primary nanoparticles in the secondary structures due to
suspending, cryogelation at high pressure (CS20 was prepared at 450 bar), and drying in air
[53,63,232].
Results thus far have addressed systems using co-adsorbates that were either non- or much
less polar than H2O. It was of interest to study the systems with CS20 containing water co-
adsorbed with more polar compounds such as H2O2, HCl, and CF3COOH [63].
A 30% aqueous solution of H2O2 (0.05 g per gram of dry CS20) was added to CS20
preheated at 470 K for 15 min (to desorb residual water) and then equilibrated for 1 h [63]. For the
sample in air, two signals of H2O2 (H = 12 ppm) and H2O (H = 5 ppm) were observed at T < 240
K. From the intensity ratio of these signals, and the temperature behavior of the 5 ppm signal
(compared with pure water), one could assume that water included a small portion of dissolved
H2O2 (since bound water is a poor solvent [53]). Similarly, the H2O2 structures responsible for the
signal at 12 ppm were practically free of water. In other words, not only was bound water a poor,
inactive solvent, but hydrogen peroxide was also [63].
Thus, partial freezing of the solution and its cryoconcentration with decreasing temperature
could cause the formation of spatially separated clusters [63] since proton exchange between H2O
and H2O2 was slow on the NMR timescale [53]. At T > 240 K, the exchange became faster and the
signal width of H2O2 became broader and, therefore, not detected (since the bandwidth was 20
kHz). The water resonance exhibited a slightly higher chemical shift due to the interaction of water
with H2O2. Besides the signal of SAW at 5 ppm, a weak signal of WAW was observed at 1.8 ppm,
the intensity of which increased with increasing temperature due to melting. The WAW did not
interact with hydrogen peroxide because of space confined effects, and the formation of spatially
separated structures with WAW and hydrogen peroxide occurred [63].
The behavior of water bound to CS20m (h 100 mg/g) was also analyzed depending on
storage of the sample under different conditions (Fig. 100) [63]. The sample CS20m stored at 255
K for a week was loaded into cooled NMR ampoule, and after the NMR spectra recording, this
sample was stored at 293 K for a week in air, and the spectra were recorded again. A small
difference in the spectra (Fig. 74a) and the Cuw(T) curves (Fig. 100b) could be caused by
adsorption of an additional portion of water from air onto the cooled sample. This resulted in an
increase in the signal intensity. Adsorbed water was located in narrow voids mainly at R < 3 nm
(Fig. 100c). Therefore, the surface area of silicas in contact with bound water was about 40 m2/g
(Table 17, Suw), and Snano > Smeso; however, Vnano < Vmeso. Therefore, the major portion of water
(~80%) was strongly bound water (Table 17, Cuw
s). This led to the value of S
* = S/h recalculated
per gram of water (~80 J/g) greater than that for the aqueous suspension but lower than that for a
sample in the methane atmosphere. As a whole, the observed changes in the behavior of water
bound to CS20m stored under different conditions were small due to the stability of cryosilica
properties [63].
112
Fig. 100. (a) 1HMR spectra of CS20m stored at 255 K (solid lines) for a week and then stored at 298 K (dashed lines)
for a week; and the corresponding (b) Cuw(T) dependences and (c) PSD and IPSD.
This aspect is of importance because silica samples were studied using a set of methods
under very different conditions including TSDC measurements of strongly cooled samples (Fig.
101) [63]. TSDC thermograms showed dipolar relaxations at T < 220-220 K and direct current (dc)
relaxation at T > 210-220 K (Fig. 101). In contrast to 1H NMR showing nonzero intensity due to
local mobility of molecules, the dc relaxation was due to through conductivity when There were
conditions for percolation of protons or Н3О
+(Н2О)n (i.e., Zundel (H5O2
+) and Eigen (H9O4
+)
cations) between two electrodes [53]. Therefore, the dc relaxation could start at slightly higher
temperatures than the NMR spectra appear for mobile molecules of adsorbates. The activation
energy (assuming that the dc relaxation obeys the Arrhenius law) was relatively low for the system
studied (54.8, 52.7, 41.3 and 55.2 kJ/mol for curves 1-4, Fig. 101, respectively). The Ea value for
the CS5 suspension was smaller than that for initial 5 wt.% aqueous suspension of nanosilica
[53,63].
The TSDC was higher for wetted powder of CS20m (h = 1.6 g/g) than for other samples
(Fig. 101) because the relative amount of water contacted with the silica surface was greater than
in other samples [63]. The high-pressure cryogelation effects appear since the TSDC intensity was
much lower for 20 wt.% suspension of A-300 (nontreated at low temperature and only sonicated)
than that of not only for CS20m but also for CS5m. The TSDC intensity for dipolar relaxations
increased with increasing silica content from 5 wt.% to 20 wt.% for cryosilicas because water
molecules were strongly polarized by the electrostatic fields near the silica surface. However, this
effect depended also on the organization of silica nanoparticles in secondary structures [63].
113
Fig. 101. TSDC thermograms (normalized to Fp = 100 V/mm) of cryosilicas CS5 and CS20m dried and hydrated at h
= 1.6 g/g, thawed aqueous suspension of CS5m (1000 atm) and gel-like concentrated suspension (S20**, 20
wt.% of A-300) prepared at standard condition without freezing.
Thus, high-pressure cryogelation affected the properties of cryosilica powders and
suspensions depending on the silica content in initial aqueous suspensions (5-20 wt.%) frozen-
dried [46,47,53,62,63]. The largest changes were observed for CS20 and CS20m (20 wt.%
suspension frozen at 450 or 1000 atm, respectively). The adsorption of water onto CS20 from air
increased several-fold in comparison with CS5, CS10 or CS15. The specific surface area decreased
slightly during high-pressure cryogelation. This was due to rearrangement of primary
nanoparticles in the secondary structures, which were denser than the initial silica powder.
However, the structure of primary particles per se changes only slightly according to HRTEM
images. This rearrangement resulted in changes of the PSD shape and contributions of pores of
different shapes (cylindrical, slit-shaped, and voids between nanoparticles) and sizes (nano-, meso-
, and macropores) [63].
The morphological and textural characteristics are of importance for applications of
nanooxides; therefore, changes in these characteristics of CNO were analyzed in detail (Tables 15-
18, and Figs. 101 and 102) [46,47,63]. Maximal change (double increase) in the specific surface
area of CNO was observed for cAST in comparison with the initial AST powder. This result could
be explained by lower stability of relatively large AST particles at 50-200 nm in diameter, which
represented polycrystalline particles covered by a shell, in comparison with smaller single alumina
crystallites or silica nanoparticles (Fig. 102). Ternary AST particles were less uniform than binary
ST (with much smaller increase in SBET approximately by 3%) or pure silica and alumina with
decreased SBET for CNO during high-pressure cryogelation at low temperature.
114
Table 18. Textural characteristics of initial and differently treated nanooxides [46,47,63].
Oxide Medium Tt (K) Pt (atm) Note SBET
(m2/g)
Snano
(m2/g)
Smeso
(m2/g)
Smacro
(m2/g)
Vp
(cm3/g)
Vnano
(cm3/g)
Vmeso
(cm3/g)
Vmacro
(cm3/g)
w
PS300 – – – Initial powder 302 94 200 9 0.734 0.035 0.523 0.176 0.021
cPS300 water 208 1000 cryogel 297 38 258 2 0.827 0.012 0.798 0.017 0.749
gsPS300 0.1M NaCl 293 1 Gel/NaCl 292 27 265 1 0.826 0.009 0.807 0.010 0.741
csPS300 0.1M NaCl 208 1000 Cryogel/NaCl 283 16 266 1 1.212 0.007 1.193 0.012 0.281
PS100 – – – Initial powder 84.4 32 49 3 0.206 0.012 0.124 0.070 0.290
gPS100 water 293 1 Gel 83.7 8 53 23 0.427 0.003 0.059 0.365 –0.108
cPS100 water 208 1000 Cryogel 83.8 28 43 13 0.474 0.014 0.207 0.253 –0.094
Al2O3 – – – Initial powder 89 12 75 2 0.167 0.006 0.132 0.029 –0.227
gAl2O3 water 293 1 Gel 76 13 60 2 0.545 0.007 0.504 0.034 0.450
gsAl2O 0.1M NaCl 293 1 Gel/NaCl 72 4 43 26 0.521 0.002 0.150 0.369 0.209
csAl2O 0.1M NaCl 208 1000 Cryogel/NaCl 72 7 44 22 0.553 0.004 0.238 0.311 0.437
ST – – – Initial powder 87 27 56 4 0.228 0.012 0.144 0.072 –0.291
gST water 293 1 Gel 90 31 49 11 0.365 0.010 0.161 0.195 –0.168
cST water 208 1000 Cryogel 89 30 49 10 0.366 0.010 0.168 0.188 –0.182
gsST 0.1M NaCl 293 1 Gel/NaCl 86 33 39 14 0.415 0.010 0.121 0.284 –0.105
csST 0.1M NaCl 208 1000 Cryogel/NaCl 88 30 44 13 0.402 0.009 0.138 0.255 –0.100
AST – – – Initial powder 83 12 68 3 0.217 0.005 0.166 0.047 –0.164
gAST water 293 1 Gel 74 6 46 22 0.532 0.003 0.231 0.298 0.363
cAST water 208 1000 Cryogel 160 15 136 9 0.595 0.008 0.466 0.121 0.125
gsAST 0.1M NaCl 293 1 Gel/NaCl 120 7 100 13 0.506 0.004 0.294 0.208 –0.060
csAST 0.1M NaCl 208 1000 Cryogel/NaCl 133 9 112 13 0.508 0.005 0.305 0.199 –0.090
Note. Gelation of all other samples was during 12 h. Labels in dried samples correspond to cryogel (c), cryogel with NaCl (cs), gelation at room temperature (gel, g) with NaCl (gs).
115
Fig. 102. HRTEM images of (a) AST, (b) cAST, and (c, d) csAST.
Among primary AST nanoparticles there near pure alumina, silica, and titania
nanoparticles. In the presence of a small amount of NaCl (2.8 wt.% in dried samples due to 0.1 M
NaCl in the 20 wt.% suspensions), an increase in the SBET value for csAST (by 64%) and gsAST
(by 45%) was smaller than that for cAST (by 93%) [45,46].
Salt crystallites formed in rigid pores could damage the materials that depended on pore
sizes, as well as on content of salts and water [233]. However, occurrence of this process during
cryogelation of the aqueous suspensions was less probable because oxide nanoparticles were
nonporous but pores were the voids between nanoparticles in their “soft” loose aggregates. During
drying of samples, salt crystallites could form in these voids; i.e. they fill a portion of the volume
of textural pores and block a portion of the oxide surface (Fig. 102c,d). However, this effect was
small since the NaCl amount was only 2.8 wt.% and NaCl crystallites were relatively large (Fig.
102d). Additionally, salt could increase the amounts of residual liquid water during cryogelation;
i.e., this could diminish pressure of ice crystallites onto oxide nanoparticles. Perhaps this effect can
play the most important role in diminution of the SBET value for csAST and gsAST in comparison
with cAST, and decomposition of complex AST nanoparticles decreased (Fig. 102). For a
mechanical mixture with AST and PS300 (1 : 1 w/w), after gelation or HPCG and drying the SBET
value decreased for mixtures without (gSAST, cSAST by 23%) and with addition of NaCl
(gsSAST, csSAST, by 19%) in comparison with a non-treated blend powder having SBET = 192
m2/g. Thus, small nanoparticles of PS300 (dav 9 nm) could play a role of a damper decreasing
high pressure effects of ice crystallites on larger AST particles (dav 30 nm). A small increase (2-
3%) in the SBET value was observed for binary oxide ST after cryogelation and drying. These
116
results for AST and ST could be due to the difference in the temperature/pressure behavior of
silica, alumina and titania during extreme treatment of binary and ternary fumed oxides consisting
of complex nanoparticles, since similar effects were absent for treated individual fumed silica and
alumina. The SBET value diminution was greater for cPS300 (by 1.7%) than that for cPS100 (by
0.7%). It also decreased for CNO with pure alumina (by ~5%). Thus, for CNO with complex
ternary and binary fumed oxides, an increase in the specific surface area could be observed due to
partial decomposition of complex nanoparticles characterized by much greater residual stress than
such uniform fumed oxides as amorphous silica and crystalline alumina (mainly with -Al2O3
which was stable up to 1150 oC and then transforms into -Al2O3) [46,47,63]. Therefore, the
difference in the temperature behavior of silica, alumina and titania phases could lead to cracking
of complex nanoparticles during HPCG. For CNO with individual oxides, SBET decreased but
nanoparticle sizes do not practically change after HPCG and drying. Thus, the SBET value could
decrease due to rearrangement (compaction) of aggregates and agglomerates of pure silica or
alumina.
The pore volume (Tables 15-18, Vp), increased for all the CNO or systems suspended-dried
at standard conditions [46,47,63]. This was due to the morphological features and related textural
characteristics caused by the formation of loose aggregates with nonporous nanoparticles and very
loose agglomerates of aggregates in the initial powders of nanooxides strongly changed during
HPCG or gelation. Compaction of nanoparticles in aggregates during HPCG (or even gelation at
standard conditions) and drying led to the opposite effects for Vem (which strongly decreased) and
Vp (which increased, Tables 15-18). The gelation/drying led to an increase in the b value from
0.067 g/cm3 (PS300) to 0.31 g/cm3 (cPS300). Therefore, Vem decreased from 14.5 cm3/g (PS300)
to 2.8 cm3/g (cPS300) due to diminution of the volume of macropores, which, however, were
poorly filled by such low-molecular weight adsorbates as nitrogen, water, etc., and, therefore, Vem
>> Vmacro. However, after HPCG or gelation (and compaction of aggregates) despite a decrease in
Vmacro, the Vp value increased due to a significant increase in Vmeso (Tables 15-18). Similar effects
were observed for other studied nanooxides that are, however, smaller due to their larger
nanoparticles, an increase in their bulk density, and a decrease in their pore volumes over the total
pore size range (Fig. 103, Tables 15-18) [46,47,63].
According to HRTEM images (Fig. S66), during HPCG and drying of cPS300, the
nanoparticle morphology of nanosilica did not practically change [46,47,63]. However, the textural
characteristics (Tables 15-18, Fig. 103) could significantly change because of rearrangement of
CNO nanoparticles in their aggregates and agglomerates. Contribution of nanopores at R < 1 nm
and narrow mesopores at 1 nm < R < 5–10 nm into the PSD decreased after gelation/drying due to
compaction of the aggregates/agglomerates of nanoparticles. In other words, each nanoparticle
could be closer to neighboring nanoparticles in CNO than in the initial powder (Fig. 103).
Contribution of large mesopores at 5–10 nm < R < 25 nm and narrow macropores at 25 nm
< R < 60-100 nm into the PSD strongly increased after gelation/drying (Fig. 103) [46]. However,
contribution of large macropores R > 100 nm decreased due to compaction of nanooxides after
gelation/drying since the bulk density increased by several times. Notice that primary
nanoparticles of PS300 have non-ideal spherical shape. Therefore, the SCV/SCR model errors
could be relatively large (Tables 15-18, w) and the w value could increase for CNO because of
certain deformation of nanoparticles and changes in the shape of the contact zones between them
in aggregates. Therefore, w = 0.021 for the initial PS300 (i.e. about 2 % error in the SCV/SCR
model) but it increased for cryogels and suspended/dried samples (Tables 15-18, w). For PS100,
w was greater (0.29) than for PS300 (0.02) since the deviation in the particle shape from the
spherical one could grow with increasing size of primary particles of fumed oxides [1-
23,46,47,53].
117
Fig. 103. Incremental SCV/SCR PSD for silicas (a) PS300 and (b) PS100, (c) alumina, (d) silica/titania, (e)
alumina/silica/titania, and (f) blend PS300 + AST (1:1): initial powder and differently treated nanooxides:
cryogelation in pure water at 208 K, in 0.1 M NaCl solution at 260 K or 208 K, suspended in water or 0.1
M NaCl solution and dried at room temperature and standard pressure (sample labels correspond to them in
Tables 15-18).
During HPCG or gelation and drying NaCl particles could affect the PSD (Fig. 103) and
other textural characteristics (Tables 15-18) [46,47,63]. This was due to several reasons. First,
NaCl dissolved in water affected oxide particle-particle interactions [46,47,53,63,170] and caused
an increase in the amounts of unfrozen water during HPCG due to enhanced freezing point
depression also caused by confined space effects that led to a certain diminution of the pressure
and reduce the nanoparticles compaction. Second, NaCl in the shape of nanocrystallites could
negatively affected the interactions between adjacent oxide nanoparticles that result in decreasing
compaction (including valence bonding of nanoparticles) of nanoparticles in the rearranged
aggregates during drying. Third, NaCl crystallites per se (NaCl density of 2.16 g/cm3 was close to
the density of amorphous silica of 2.2 g/cm3) could give a certain contribution into the textural
characteristics of the materials depending on the NaCl crystallite sizes. Additionally, NaCl
nanoparticles could fill voids between oxide nanoparticles that influence the S and V
characteristics of the materials. However, these direct effects of NaCl crystallites on the textural
118
characteristics were small since the NaCl amount was low (2.8 wt.%) and crystallites were
relatively large [46,47,63].
For csPS300, the presence of NaCl in the dried CNO powder caused diminution of the
specific surface area (Table 18, SBET, Snano) but an increase in the pore volume (Vp, Vmeso) in
comparison with pure cPS300 [46]. For alumina, AST and ST, the NaCl effects differed (in
general, it was smaller) from those for PS300 because of the difference in the textural
characteristics, density, primary particle sizes, and the aggregation degree of these nanooxides.
Minimal effects were observed for ST (Table 18, Fig. 103). This could be due to relatively
uniform distribution of titania in the volume and surface of complex ST nanoparticles
characterized by a maximal number of the SiOTi and SiO(H)Ti bridges among silica/titania
nanooxides [53]. This prevents cracking ST nanoparticles during HPCG.
Sample cPS300 with strongly increased bulk density (0.31 g/cm3) and decreased
contribution of nanopores and macropores (Table 18) was characterized by increased total Vp
because of a significant increase in contributions of pores over the 10–30 nm range (Fig. 103a)
[46,47,63]. The rearrangement of silica nanoparticles in 0.1 M NaCl suspensions frozen at 260 K
and 208 K was similar (Tables 15-18). However, their PSD differ (Figs. 103) since the PSD of
csPS300 shifted toward larger pore sizes. Washing of csPS300 (giving cswPS300) resulted in a
certain increase in all the textural parameters (Tables 15-18) because of removal of NaCl particles
filling voids between neighboring silica nanoparticles and changes in particle-particle interactions
after washing-drying. Therefore, the PSD intensity of cswPS300 increased around R 10 nm (Fig.
103a, curves 2 and 3). Similar results were observed for non-washed and washed suspended-dried
samples gsPS300 and gswPS300 (Tables 15-18). However, their PSD intensity around R 10 nm
was lower than that for cryogels (Figs. 103a, curves 6 & 7 vs. 2 & 3) due to weaker compaction of
suspended/dried but non-frozen samples.
For a mixture PS300/AST, the effects of NaCl differ from those for the corresponding
individual systems PS300 and AST (Tables 15-18, Fig. 103f) [46,47,63]. For individual PS300,
addition of NaCl into the aqueous media led to decreasing textural parameters of HPCG samples
but for AST, the opposite results were observed. For csSAST, a smaller increase (in comparison
with pair of AST and csAST) in the parameters was observed in comparison with cSAST (Tables
15-18). For alumina, NaCl enhanced contribution of macropores but the opposite effect was for
mesopores and nanopores. This feature was well seen in the PSD (Figs. 103).
Not only textural and morphological characteristics of nanooxides changed due to HPCG but also
the crystalline structure of samples containing alumina [46]. An intensive line at 2 31.7o could
be assigned to NaCl crystallites (2 31.1o for individual NaCl) bound to alumina in csAl2O3.
This band was low-intensive in Al2O3 and gAl2O3 (Fig. 104). However, this line could be rather
attributed to -Al2O3 having a doublet at 31.3 and 31.7o. In other words, an increase in the
crystallinity degree of alumina after HPCG could lead to an increase in the intensity of -Al2O3
lines. However, intensity of a line at 32.85o which was higher in both pure -Al2O3 and initial
fumed alumina or gAl2O3 was lower than that at 31.7o in csAl2O3. The latter could be due to the
effect of NaCl. Additionally, in both initial Al2O3 and gAl2O3, There was a line at 32.0o which
could be assigned to -Al2O3 and which was not observed in csAl2O3. Thus, certain but not strong
changes in the crystalline structure were observed in csAl2O3 (Fig. 104). However, the hydroxide
Al(OH)3 lines were very weak in csAl2O3 in contrast to treated AST (Fig. 105). These insignificant
changes in the crystalline structure of csAl2O3 were in agreement with relatively small changes in
the SBET value ( 19 %) of treated alumina samples in contrast to AST ( 93%) (Table 18)
[46,47,63].
119
Fig. 104. XRD patterns for initial alumina (Al2O3),
suspended/dried at standard conditions
(gAl2O3), and cryogel prepared in 0.1 M
NaCl solution at 208 K (csAl2O3) (sample
labels correspond to them in Table 18),
and pure crystalline NaCl and –Al2O3.
Fig. 105. XRD patterns for initial AST, and cryogel prepared
in water (cAST) or 0.1 M NaCl solution at 208 K
(csAST) (sample labels correspond to them in
Table 18), and pure crystalline NaCl, anatase, and
different crystalline alumina phases: corundum (–
Al2O3), –Al2O3, and –Al2O3.
Greater changes in the XRD patterns, especially at 20–40o, were observed for AST after
HPCG in comparison with individual alumina (Fig. 104) [46]. From these changes, the HPCG
effect on the XRD pattern of AST was stronger than the presence of NaCl (due to its low content
of 2.8 wt.%). The lines of such several phases as aluminum hydroxide Al(OH)3, corundum –
Al2O3, –Al2O3 and –Al2O3 were better observed without NaCl (cAST) than in the presence of
NaCl (csAST) or in initial AST. Stronger XRD changes in treated AST (Fig. 105) than that of
alumina (Fig. 104) were in agreement with greater changes in the textural characteristics of CNO
with AST than alumina (Tables 15-18). The crystallinity degree of AST, csAST and cAST was 32
%, 37 and 43 %, respectively (estimated from the area of crystalline peaks and residual amorphous
halo). The size of alumina crystallites was about 10 nm that was in agreement with the particle size
of cAST (12 nm) estimated from the SBET value but much smaller than that for the initial AST
(~25 nm). The increase in the crystallinity degree of cAST and csAST was mainly due to
appearance of hydroxide Al(OH)3 (lines at 18.2 and 20.5o) and a certain increase in intensity of
lines over the 25-70o range of several alumina phases in comparison with the initial AST (Fig.
105) [46,47,63].
The changes in the structural characteristics of treated samples studied were accompanied
by changes in the surface structure. Treated silica samples studied using the IR spectroscopy were
more hydrated than initial nanosilica since a broad band over 2500–3800 cm–1 was more intensive
for them (Fig. 97). Therefore, the surface area free of water (estimated according to [106]) was 262
(PS300), 90 (cPS300), 67 (csPS300), 71 (gPS300), and 41 (gsPS300) m2/g. This was due to
compaction of silica nanoparticles in aggregates during treatment in the aqueous media and drying.
Additionally, NaCl bound to silica was highly disperse since NaCl particles and adsorbed water
(bound to silica and NaCl particles) strongly disturb a larger portion of the surface hydroxyls than
in pure initial silica and CNO (compare the IR spectra for cPS300, gPS300 and csPS300,
gsPS300) [46,47,63].
The IR spectra of AST change due to pre-heating (Fig. 106, curves 1 and 2) and gelation
under different conditions (curves 3-6). There was no narrowing of the cooperative Si–O–Si
asymmetric modes at 1100 cm1, as well as of the bands related to Ti–OSi, AlOAl, and AlOSi.
This result was in agreement with diminution of the sizes of AST particles after HPCG and some
changes in the phase composition, e.g., appearance of Al(OH)3 (Fig. 105) [46]. Changes in
intensity of the IR bands over the 900-500 cm1 range could be due to dehydration of the surface
on heating at 723 K and appearance of four- or fivefold O-coordinated Al atoms instead of sixfold
120
O-coordinated ones. This led to an increase in intensity of the bands at 800-900 cm1 and decrease
in intensity of the bands at 500-600 cm1. HPCG of the aqueous suspensions results in the opposite
effects due to enhanced hydration (formation of the Al(OH)3 phase) and O-coordination number of
Al atoms [46].
Fig. 106. IR spectra of AST (curve 1) initial and (2) preheated at 723 K for several hours, (3) cAST, (4) csAST, (5)
gAST, (6) gsAST, (7) fumed ST94 with 6 wt.% SiO2 and 94 wt.% TiO2, (8) pure fumed alumina, corundum
-Al2O3 (9) IR and (10) Raman spectrum, and (11) pure anatase, (1:100 mixture with KBr). All spectra of
AST samples were normalized to the asymmetric mode of Si-OSi at 1100 cm-1.
Adsorption energy of water or other compounds depended strongly on the content of
already adsorbed amounts (Fig. 107) [53,58]. This effect caused different changes in the structure
of bound water depending on the distance from the adsorbent surface. This is also reflected in the
relationship of G vs. Cuw vs. T [53].
Fig. 107. Adsorption enthalpy vs. uptake for H2O vapor adsorption (T = 303 K) on (i) silica-coated anatase (A-S: )
and silica-coated rutile (R-S: , ◊) in comparison with uncoated anatase (A8/A72, dot-dashed curves),
uncoated rutile (R, dashed curves) and a plain (fumed) silica specimen (S: ); (ii) alumina/dimethicone-
coated rutile (R-Alu-D: ) and a plain (fumed) alumina specimen (Alu: ). Volumetric isotherms in the
inset. Solid symbols 1st run, open symbols 2nd run of adsorption. All samples were preliminary outgassed at
T = 303 K. Latent enthalpy of liquefaction of water: LH(H2O) = 44 kJ/mol (dotted line).
The changes in the texture and surface properties of treated samples led to significant
changes in adsorption-desorption of polar water and nonpolar hexane (Fig. 108) [46,47,63]. A
maximal increase in the water (Fig. 108a) and hexane (Fig. 108b) adsorption was observed for
pure cryogels cPS300 and cAST, and for the latter this effect was stronger, since for the latter this
121
increase for water adsorption corresponded to a factor 2.5 but for the former it was 2.2. For
nonpolar hexane, this effect was similar. This difference could be explained by stronger changes in
the morphology of AST particles than for PS300 during cryogelation at high pressure (Table 18).
Practically the same changes in the adsorption of polar water and nonpolar hexane could be
explained by the effects of the textural changes in the materials but not by changes in the content
of surface hydroxyls due to particle decomposition.
The presence of NaCl resulted in diminution of changes in both structural (Table 18) and
adsorption (Fig. 108) properties of gAST and gsAST in comparison with cAST and csAST
[46,47,63]. Changes in the texture (Fig. 103) and structure (Fig. 105) of differently treated AST
samples resulted in an increase in water desorption mainly in the 350-500 K for cAST and gAST
in comparison with the initial AST [46]. Compaction of the particles with their simultaneous
decomposition (Fig. 75) and formation of the aluminum hydroxide in cAST caused a shift of the
main peak toward higher temperatures. For gAST, desorption of water increased by several times
in comparison with AST but the peak position was practically the same [46].
Fig. 108. Isotherms of (a) water and (b) n-hexane
adsorption-desorption onto initial PS300,
cryogels cPS300 and cAST (prepared at 208
K and ~1000 atm), and AST initial and
gelatinized at standard conditions.
Fig. 109. Size distributions (a) differential or (b)
incremental for structures of unfrozen water
bound to cPS300 in different dispersion
media at hydration degree h = 0.1 g/g in air or
chloroform-d with addition of 0.15 g of n-
decane (at cryoporometry constant 67 K nm).
To analyze the NaCl influence on the HPCG results, 1H NMR study of water bound to
cPS300 and csPS300 was carried out for samples being in air or chloroform-d media also affected
by added non-polar n-decane (Fig. 109) [46]. The NaCl effects were stronger than the effects of
CDCl3 or n-decane [63]. Typically, hydrophobic media (CDCl3 here) displace a portion of water
from pores to form larger water structures in larger pores or out of pores [53]. This resulted in
faster decrease in signal intensity with decreasing temperature [46]. However, addition of a small
amount of n-decane (0.15 g/g) caused re-arrangement of water adsorbed in amounts close to the
122
equilibrium adsorption (0.1 g/g) at p/ps 0.9 for initial PS300 or at p/ps 0.9 for cPS300 in air
medium (Fig. 108). Therefore, addition of CDCl3 to n-decane/water/cPS300 resulted in the
opposite effect, since water structures became smaller and PSD intensity increased at R < 2 nm
(Fig. 109). For csPS300, a similar effect was weaker due to influence of NaCl crystallites on the
organization of adsorbed water. In other words, strong interactions of water molecules with NaCl
nanocrystallites and dissolved ions Na+ and Cl led to diminution of the influence of changes in the
dispersion media [46,47,63].
Significant changes in the texture and structure of cAST in comparison with AST and
gAST (Tables 15-18, Figs. 102, 103e, 105, 106, and 108) led to a difference in the interfacial
behavior of water alone or in the presence of decane (Figs. 110 and 111) [46,47]. Small amounts
of water (0.1 g/g) and decane (0.15 g/g) were used to fill only a portion of pores of the materials
(approximately Vp half of cAST and gAST, Table 18). The greatest difference in the 1H spectra
was observed for AST samples with water/decane in chloroform-d with addition of trifluoride
acetic acid (TFAA) (Fig. 110b). This was due to the influence of the texture of AST (differently
changed for gAST and cAST as shown above), i.e. confined space effects, on organization of both
water and decane and dissolution of TFAA in this water. A maximal amount of TFAA was
dissolved in water bound to AST since the 1H NMR peak was at H = 9-10 ppm. The TFAA
amounts dissolved in water decrease for gAST and it was minimal for cAST [46,47,63].
For treated AST (gAST and cAST), decane displaces water into narrower voids (comp. Fig.
111a,b). Chloroform medium enhanced this trend. A minimal effect of the chloroform medium
was for water bound to AST, i.e. water structures bound to AST were larger than those bound to
gAST or cAST. Therefore, TFAA could be better dissolved in water bound to initial AST. There
was apparent contradiction between this effect and the shapes of water structures. However, it was
due to a fixed constant kGT = 67 K nm in the Gibbs-Thomson equation [53] for all calculations
with NMR-cryoporometry (Figs. 109 and 111). Clearly, this constant for pure water and water
with dissolved TFAA could be different [46,53].
Weakly associated water (WAW) characterized by H = 1.2-2.0 ppm appeared after
addition of decane, especially for gAST [46,47,63]. There was certain content of WAW in the
nanooxide systems in the chloroform media (Fig. 110). The main reason of the WAW appearing
was the formation of branched 3D clusters of water located in narrow voids. Weakly polar
(chloroform) or nonpolar (decane) co-adsorbates were immiscible with water and their molecular
sizes were larger than that of water. Therefore, they could displace water into very narrow voids
where they could not be located. The number of such narrow voids increased for cAST due to
nanoparticle decomposition. Therefore, displacement effect of decane onto water was stronger for
cAST than that for gAST or AST (Fig. 111b). Some theoretical calculations of the 1H NMR
spectra of water alone and bound to silica, pure and with NaCl analyzed in the ESI confirm this
interpretation of the experimental 1H NMR spectra [46,47].
High-pressure cryogelation of 20 wt.% aqueous suspensions of nanooxides (silicas PS300
and PS100, alumina, silica/titania, alumina/silica/titania) pure or with addition of 0.1 M NaCl at
208 K or 260 K led to significant changes in the textural characteristics of dried cryonanooxides
and changes in the crystallinity of certain oxides [46,47,63]. Maximal changes in the specific
surface area and crystalline structure were observed for high-pressure cryogels with ternary
nanooxide AST. Large ternary nanoparticles (composed of nanocrystallites covered by a shell)
could be decomposed during high-pressure cryogelation, the crystallinity degree increased by
approximately 10 % and a new phase of aluminum hydroxide appears. Addition of PS300 to AST
(mechanical blend at 1:1 w/w) led to diminution of the decomposition of AST particles during
HPCG. For individual nanooxides (silica, alumina), high-pressure cryogelation caused a decrease
in the specific surface area, but for ternary AST and binary ST, the specific surface area increased
(much greater for ternary AST). For all treated samples, a significant increase in the volume of
large mesopores (at pore radius 5–10 nm < R < 25 nm) and macropores (25 nm < R < 60–100 nm)
was observed. However, contribution of nanopores (R < 1 nm) and narrow mesopores (1 nm < R <
5–10 nm) decreased after HPCG due to compaction of nanoparticles in aggregates. Thus, such
123
individual nanooxides as amorphous nanosilica and crystalline nanoalumina were morphologically
more stable during HPCG than complex nanooxides such as ternary nanooxide with alumina
(crystalline)/silica (amorphous)/titania (solid solution in alumina and silica), especially composed
of nanocrystallites covered by a shell, and binary nanooxide with silica (amorphous)/titania
(partially crystalline). Nanosilica PS300 with the highest SBET value among studied oxides were
characterized by maximal rearrangement of aggregates and agglomerates during HPCG because of
the most minimal sizes of primary nanoparticles that provide maximal flexibility of secondary
structures of the material [46,47].
Fig. 110. 1H NMR spectra of water (0.1 g/g) and (b) n-
decane (0.15 g/g) bound to AST (dashed
lines), gAST (dot-dashed lines) and cAST
(solid lines) in media (a) air and (b) CDCl3
with addition of TFAA (6 : 1).
Fig. 111. Incremental PSD calculated from 1H NMR
spectra of water (0.1 g/g) bound to AST, gAST
and cAST (b) with addition of n-decane in air.
Addition of NaCl (0.1M) into the aqueous suspensions of nanooxides sonicated before
HPCG affected the textural and structural characteristics on the finished cryogels due to several
reasons [46,47]. First, ice was much pure than water with respect to dissolved compounds.
Therefore, concentrating NaCl solution occurred during HPCG that results in additional depression
(besides the depression due to confined space effect) of the freezing point of interfacial water with
dissolved NaCl. This could slightly reduce pressure caused by ice crystallites formed in the cryo-
bombs. Second, interfacial water was characterized by lower activity as a solvent. Therefore, the
formation of NaCl nanocrystallites could start during HPCG (and finish during drying of cryogels)
and they could affect oxide inter-particle interactions. This could reduce the transformation of
complex oxide nanoparticles, e.g., cAST was composed of smaller nanoparticles than csAST due
to stronger decomposition complex nanoparticles. Third, NaCl crystallites per se could affect the
124
textural characteristics of dried cryogels. However, this influence could not be strong since the
amounts of NaCl were small (~2.8 wt.% in dried powders) and NaCl crystallites were relatively
large up to 500 nm.
The difference between IPSD of initial powders and treated samples was smaller at R < 2
nm (contact zones between adjacent nanoparticles were only slightly compacted) than at R > 10
nm since the whole texture of aggregates and agglomerates undergoes significant rearrangement
during HPCG, simple gelation or MCA in contrast to nanoparticles per se (Figs. 97, 103, and 112)
[46,47]. MCA for 5 or 30 min gave small changes in the IR spectra in the range of the MO (M =
Si or Al) stretching vibrations in the SA13 blends (Fig. 113). These small changes of nanoparticles
affected by MCA slightly touch crystalline structure of alumina (silica was amorphous) whose
crystallinity increased, sizes of nanoparticles (decreased) and their electronic structure [47]. A
decrease in the size of primary particles should result in increasing SBET; however, enhanced
contacts between adjacent nanoparticles could compensate this effect and the value of SBET for
MCA SA samples was smaller than that for HPCG or gelled samples [47].
Fig. 112. Incremental pore size distributions (VCV/SCR method for silica and titania) (a) CVT-ST20 initial (curve 1),
HPC (2) and gel (3), (b) ST63 initial (curve 1), HPC (2) and gel (3), (c) ST80 initial (curve 1), HPC (2) and
gel (3), and (d) ST94 initial (curve 1), HPC (2) and gel (3).
Fig. 113. Infrared spectra of initial PS300 (curve 1) and SA13 blend initial (2) and after MCA for 5 min (3) or 30 min
(4).
125
During HPCG freezing temperature plays an important role because it could affect pressure
in the freezing bombs [46,47,63]. AST sample was the most interesting among all nanooxides
studied under HPCG [46,47] because it included a fraction of core-shell nanoparticles of a
relatively large size (50-200 nm), which could be destroyed under HPCG. To analyze the
temperature behavior of AST during HPCG, three regimes of freezing were used (Fig. 114) [47].
Two-step HPCG (sample c26077AST frozen at 260 K and then at 77.4 K) resulted in maximal
compaction of AST since the hysteresis loop shifted toward smaller pressures (Fig. 114a) and the
main IPSD peak shifted toward smaller pore sizes (Fig. 114b). A similar but slightly smaller effect
was observed for c77AST after HPCG at 77.4 K. HPCG at 208 K (c208AST) resulted in weaker
compaction of aggregates and agglomerates. However, the IPSD at R < 5 nm for all HPC samples
were very similar (Fig. 114b). Thus, the destroying degree of the core-shell large nanoparticles
was practically the same at these three temperature regimes, since the value of SBET was the same
for all HPCG AST samples and approximately twice larger than that of the initial AST powder
[47]. However, rearrangement (compaction) of the secondary structures depended on the freezing
conditions (Fig. 114) [46,47].
Fig. 114. (a) Nitrogen adsorption-desorption isotherms and (b) incremental pore size distributions (VCV/SCR method
for alumina and silica) for initial dry powder AST (curve 1), cryogels prepared at (2) 260 K for 24 h and then
77.4 K for 4 h, (3) 77.4 K for 4 h, and 208 K for 12 h.
Unmodified and modified silicas and interfacial behavior of adsorbates
Fumed silica could be modified using a variety of reagents (e.g., a set of silanes) with
different functionalities including simple hydrophobic groups (CH3 and longer), groups with
reactive bonds (C=C, N-H, etc.), polar groups (NH, COOH, CO, COH, SH, etc.) [1,11,37,38].
Typically, hydrophobicity of silica increased with increasing TMS content on the silica surface
(Table 19) [53,72]. A decrease in the hydrophilicity (accompanied by an increase in intensity of
the CH stretching vibration bands of the TMS groups at 2970 and 2907 cm1) of modified silicas
was accompanied by a decrease in the adsorption of water from air. This led to a reduction of
intensity of the broad band of adsorbed water at 3300 cm1 [72].
Table 19. Heat of immersion of unmodified and modified silicas in water and n-decane [72].
TMS
(%)
Qw
(J/g)
Qd
(J/g)
= Qw/Qd
0 48 22.0 2.0
27.2 39 24.5 1.6
37.2 33 24.0 1.4
Deconvolution of a broad band of the OH stretching vibrations (into four bands for PS300
and six bands for partially hydrophobized nanosilica HS1 (TMS = 27.2 % substitution of SiOH by
TMS groups) and HS2 (TMS = 37.2 %) due to the presence of two CH bands [53]) gave integral
126
intensity of a band of strongly bound water or strongly disturbed silanols at 3290-3280 cm1 as
68.2, 56.0, and 50.8 % for PS300, HS1, and HS2, respectively [72]. Relative contribution of
slightly disturbed silanols at 3670-3680 cm1 (which were poorly accessible for adsorbates)
increased for modified silicas, since integral intensity of this band was 5.2, 7.8, and 8.1 % for
PS300, HS1, and HS2, respectively. Notice that There was a slight broadening of a band of
asymmetric SiO stretching vibrations at 11501250 cm1 of HS2 in comparison with PS300 (Fig.
S75). This could be due to surface Si–O–Si(CH3)3 groups (the spectra were normalized to the
intensity of the band at 1867 cm1 (which was the bulk silica overtone mode proportional to the
total mass of silica probed by the IR beam and used as the inner standard) or 1100 cm1 (Si–O
stretching vibrations) of unmodified silica) [72]. Typically, a decrease in separation between the
cooperative Si–O–Si asymmetric modes at 1100 cm1 (i.e. band narrowing) in the IR spectra of
silica was usually associated with an increase in the polymerization degree, i.e. an increase in the
number of Si–O–Si bonds in larger particles [37,38,105,235]. However, it was difficult to suppose
that silica nanoparticles’ size decreased during silylation since the specific surface area of
modified silica decreased. Therefore, the observed broadening could be assigned to the surface
groups because their symmetric and asymmetric stretching vibrations could occur at higher
frequency [72].
Calculations of the specific surface area according to the IR data [106] (using the
normalized ratio of integral intensity of the bands at 1867 and 3750 cm1) give SIR = 328, 296, and
250 m2/g at TMS = 0 (PS300), 27.2 % (HS1), and 37.2 % (HS2), respectively. This reduction of
the SIR value could be due to two effects. First, real diminution of the specific surface area was due
to attaching TMS groups. Second, diminution of the content surface silanols caused apparent
decrease in the specific surface area since this content was used to estimate the SIR value [106].
Thus, real diminution of the specific surface area was slightly smaller than the SIR value showed
[53]. The reduction of the specific surface area of modified silica with increasing modification
degree was a typical effect [53,236,237]. Besides this surface area reduction, the adsorption
potential also decreased with increasing value with respect to both polar and nonpolar
compounds and polymers [29]. Thus, the interfacial behavior of such co-adsorbates as methane or
hydrogen and water could depend on several factors: (i) the textural characteristics vs. surface
composition; (ii) changes in the adsorption potential; (iii) changes in clustered adsorption of water;
(iv) freezing and melting points of co-adsorbates; (v) current temperature, and (vi) concentration
of co-adsorbates and dispersion medium type. Some of these aspects were analyzed with
lowtemperature 1H NMR measurements [72].
Partial silylation of PS300 (TMS = 27.2 and 37.2%) caused an increase in nanoparticle
aggregation observed in the suspension with increasing TMS value (Fig. 115a-c). This could be
explained by rearrangement of modified silica nanoparticles in aggregates to reduce the contact
area between hydrophobic TMS groups and water molecules. Notice that the aggregation was
higher for HS1 than for HS2 (Fig. 115). This could be explained by a greater nonuniformity of the
surface at a lower degree of the silylation. Similar effects were observed for other samples
[53,237,238]. However, the observed changes in aggregation were similar to typical changes in
aggregation of different unmodified nanosilicas [238]; i.e., they were not extensive since the silica
surface was only partially hydrophobized. Changes in the zetapotential of partially modified
nanosilicas were relatively small compared to the initial silica (Fig. 115d). Only a portion of
surface silanols was deprotonated in the used range of pH. The modified silica surface had enough
number of silanols to provide a close value of the surface charges. Additionally, the zetapotential
was linked to average charging under slipping plane in the electrical double layer, EDL, around
primary particles and their aggregates. Thus, partial silylation of the nanosilica surface led also to
changes in a dense part of EDL [72].
A self-consistent electrostatic theory was used to predict disjoining pressure isotherms of
aqueous thin-liquid films stabilized by non-ionic surfactants and air-water surface tensions and
zeta potentials of electrolyte solutions with and without non-ionic surfactant. This model
127
combined specific adsorption of hydroxide ions at the interface with image charge and dispersion
forces on ions in the diffuse EDL [239].
Fig. 115. Particle size distributions with respect to (a) scattered light intensity, (b) particle volume, and (c) particle
numbers, and (d) zetapotentials of PS300, HS1, and HS2 samples in the diluted (0.02 wt.%) aqueous
suspensions.
128
Fig. 116. 1H NMR spectra recorded at different temperatures of water bound to modified PS300 (HS2, = 37.2 %)
for (a, b) wetted powder at h = 1.2 g/g and (c, d) aqueous suspension at h = 13 g/g, (e) amounts of unfrozen
water Cuw vs. temperature and vs. changes in the Gibbs free energy of bound water (errors < 10%), and (f)
size distributions of unfrozen water clusters bound to silica in wetted powder and aqueous suspension (Figs.
b and d correspond to subsets of Figs. a and c at low temperatures).
The mechanisms of the protonation of solid metal oxides and hydroxides in aqueous media
were analyzed using theoretical approaches and experimental results [240]. The apparent
acidity/basicity of each kind of surface sites of metal oxides and hydroxides in aqueous
suspensions was strongly influenced by the overall surface charge of the materials and thus by the
electrical potential smeared out at the interfacial region. Depending on its sign this increased or
decreased the hydrogen ion concentration on the surface, thus promoting or hindering protonation.
This was manifested by the shifts of the protonation peaks of the various kinds of sites with
respect to the −pK values of the corresponding intrinsic protonation constants and the appearance
of an extra peak in the d[H+
cons,surf]/dpH vs. pH curves. Potentiometric titrations experiments
performed for four technologically important oxides showed that the proposed protonation
mechanism describes indeed the protonation of polycrystalline oxides and hydroxides in aqueous
media [240].
129
For modified nanosilicas in more concentrated aqueous suspensions, stronger aggregation
could be expected than observed in [72] for the more diluted suspensions or for unmodified silicas
[53,237,238]. Therefore, one could expect that the aggregation of modified silica nanoparticles
could affect the interfacial behavior of water bound to a surface.
The forces acting in colloidal suspensions and affecting their stability and aggregation
kinetics have been analyzed [241] and the approximations used for these forces in numerical
simulations and the importance of the balanced account for both colloidal forces and
hydrodynamic interactions have been discussed. Simulations were based on the Langevin
equations with pairwise interaction between particles and take into account Brownian,
hydrodynamic and colloidal forces. It was confirmed that the neglecting of hydrodynamic
interaction results in an accelerated growth of aggregates. The results of numerical simulations of
aggregation kinetics were compared with well-known analytical solutions [241].
Water adsorbed onto a surface of any adsorbent could exist in clusters due to nonuniformity
of the surface (e.g., mosaic coverage with surface hydroxyls) if the water amounts were relatively
small (< 20 wt.% for nanosilica [53]). Additionally, bulk or bound water tends to form a maximum
number of hydrogen bonds per molecule possible [53,242]. The appearance of hydrophobic TMS
groups at a silica surface (with mosaic SiOH and SiOSi(CH3)3 coverage) could affect the
structure of bound water (Fig. 116, Table 20) [72]. At a relatively high hydration degree h = 1.2
g/g, water could form a continuous layer at the modified silica (HS2) surface. Therefore, the only
signal observed was from strongly associated water (SAW) [53] in the 1H NMR spectra (mobile at
T < 273 K), identified by the chemical shift of the proton resonance H = 5.05.5 ppm. These H
values were typical for hydrated silicas [53,218] and close to that of bulk liquid water. At 265 K <
T < 273 K, a main fraction of this water was frozen. Therefore, it could be assigned to weakly
bound water [53]. At low temperatures, two relatively weak signals were observed at 5 ppm
(signal 1) and 7 ppm (signal 2) (Fig. 116b). On the basis of freezing temperatures and H values,
this water could be assigned to strongly bound water [53] of the SAW type. The H value of signal
2 was close to that of ice. This ordering of interfacial water (similar to ice) could be caused by the
effects of hydrophobic TMS groups on the water [53,242]. In the concentrated aqueous suspension
at h = 13 g/g, besides the signal of SAW at H = 4.56.0 ppm (Fig. 116c,d), a signal of weakly
associated water was also observed at H 1 ppm. This could be explained by water filling of
adsorption sites at the surface in a confined space between neighboring TMS groups [72]. For
wetted powder (h = 1.2 g/g), the degree of filling of adsorption sites by water is likely much lower
(i.e., the hydration shells of nanoparticles were nonuniform) than in the suspension. WAW
corresponds to 1D, 2D or branched 3D clusters, in which a relative number of the H atoms in the
hydrogen bonds was smaller than 50% [53]. The complex signal shape seen at 272.5 K and 280 K
(Fig. 116c) may be due to nonuniformity of the modified silica [72].
For wetted powders, the greater the amount of adsorbed water, the smaller the fraction of
strongly bound water [53]. The WBW fraction could increase with increasing water content
because the amount of water distant from the adsorbent surface increased. For aqueous
suspensions, three main types of water could be present: SBW, WBW and bulk (unbound water,
UW) water, and one could expect that all of the water was SAW. There were two sections in the
Cuw(T) curves (Fig. 116e) corresponding to SBW (at lower Gibbs free energy, ΔG) and WBW (at
higher Gibbs free energy) [72]. The amounts of SBW were 0.13 and 0.45 g/g for the wetted
powder and the suspension, respectively, while the total volume of bound water was 0.965 and
4.325 cm3/g, respectively (Table 20, Vuw). Contact area between bound water and the HS2 surface
increased nearly five-fold in the suspension (Table 20, Suw).
130
Table 20. Characteristics of unfrozen water bound to modified silica (HS2, = 37.2%) at different conditions [72].
Sample h (g/g) Vuw
(cm3/g)
Suw
(m2/g)
Suw,nano
(m2/g)
Suw,meso
(m2/g)
Suw,macro
(m2/g)
Vuw,nano
(cm3/g)
Vuw,meso
(cm3/g)
Vuw,macro
(cm3/g)
S
(J/g)
<T>
(K)
ΔGs
(kJ/mol)
Wetted powder 1.2 0.965 62 3 52 7 0.001 0.860 0.103 9.911.0 267.45 2.310.2
Suspension 13 4.325 287 47 143 97 0.022 2.628 1.675 34.863.5 267.96 2.620.3
Weakly wetted powder +CH4 0.05 0.050 15 12 3 0 0.005 0.045 0 3.360.34 230.92 2.630.3
Nearly dry powder +H2 0.005 0.005 0.3 0 0.3 0 0 0.005 0 0.120.01 258.33 1.630.2
Powder in CDCl3+TFAA 0.05 0.038 19 15 4 0 0.007 0.031 0 2.490.25 242.27 2.630.3
Note. Vuw is the volume of bound water unfrozen at T < 273 K; Suw is the specific area of silica in contact with unfrozen water; nano-, meso- and macro-components of Suw and Vuw
were determined by integration of the distribution functions at radius values of 0.31.0 nm (nano), 125 nm (meso), and 25100 nm (macro); S is the surface free energy (in J per
gram of dry oxide); <T> is the average melting temperature; and ΔGs is changes in the Gibbs free energy of strongly bound water.
131
Fig. 117. Theoretically calculated 1H NMR spectra (by PM7) of water (~200H2O) bound to unmodified (curve 1, 828
atoms) or partially modified (curve 2, 1272 atoms) silica particle, and pure water with 2000H2O (curve 3).
Therefore, contributions of structures of three types of waterfilled nanopores (at radius R
< 1 nm, Suw,nano, Vuw,nano), mesopores (1 nm < R < 25 nm, Suw,meso, Vuw,meso) and macropores (25 nm
< R < 100 nm, Suw,macro, Vuw,macro) all increased significantly. Changes in contacts between bound
water and the HS2 surface caused an increase in the interaction energy. Therefore, the surface
energy increased by several times (Table 13, S), and the Gibbs free energy of the first adsorption
layer (corresponding to SBW) decreased (ΔGs). The S values were calculated in J per gram of dry
silica. Therefore, at a small content of adsorbed water these values were small (Table 20, h = 0.05
or 0.005 g/g). However, for wetted powder, the S value was small despite h = 1.2 g/g that
suggested weak interactions between water and HS2 [72].
The average melting temperature for water bound in the suspension was close to that for
the wetted powder (Table 20, <T>) [72]. The <T> values for wetted powder and concentrated
suspension suggest that WBW was predominant in both systems, in agreement with the
dependences of Cuw vs. ΔG and T (Fig. 116e) showing that the major portion of bound water was
WBW. This was also in agreement with the bound water cluster sizes (Fig. 116f), since water in
larger clusters should be more weakly bound to the surface as their distance from the surface
increased. There was a tendency of a decrease in the <T> value with decreasing hydration degree
(Table 20). However, this also depended on other conditions (e.g., compare systems at h = 0.05
and 0.005 g/g, Table 20).
According to theoretically calculated 1H NMR spectra, water bound to nanosilica was
characterized by a broader band exhibiting a larger contribution of water with downfield shift (Fig.
117, curve 1) in comparison with pure water (curve 3) [72]. Water (SAW) bound to silylated silica
surface was characterized by an upfield shift with appearance of WAW at 12 ppm (Fig. 117,
comp. curves 2 and 1) similar to experimentally observed spectra. These effects were due to
enhanced clustering of water bound to TMSsilica with mosaic hydrophilichydrophobic patches,
in comparison with water bound to an unmodified silica surface [72].
Methane molecules interacting with adsorbents due to weak van-der-Waals bonds could be
effectively adsorbed only into nanopores or narrow mesopores. Therefore, methane adsorption
onto nanosilica was relatively low, particularly on unmodified nanosilicas [53,72,104]. It was also
shown that co-adsorbed water could increase the adsorption of methane. To increase methane
adsorption, partially silylated nanosilica HS2 was used with a small amount of pre-adsorbed water
(h = 0.05 g/g). This water, partially freezing at T < 273 K, could change the topography of the
silica surface and the topology of voids between adjacent nanoparticles. Under these conditions,
clathrate formation was impossible (clathrates only form at high pressures) and the low content of
132
water forming strongly clustered structures at residual silanols located between attached TMS
groups [72].
HS2 with pre-adsorbed water (h = 0.05 g/g led to a larger content of SAW than WAW or
SBW than WBW), various portions of which could be frozen at different temperatures, was a
nonuniform adsorbent for methane (Fig. 118) [72].
Fig. 118. (a, b) 1H NMR spectra recorded at different temperatures of water and methane bound to HS2, (c)
temperature dependences of concentrations bound unfrozen water and adsorbed methane (errors < 10%),
and (d) size distribution of voids filled by unfrozen water.
This nonuniformity was also observed in the distribution of unfrozen water structures filling voids
between silica particles (Fig. 118d) and structural characteristics of voids filled by bound water
(Table 20). Therefore, two signals of methane were observed in the 1H NMR spectra at H = 0.1
0.4 ppm (Fig. 118a,b). Signal intensity of methane increased with increasing temperature (Fig.
118a). The amounts of unfrozen water and adsorbed methane vs. temperature (Fig. 118c)
demonstrated concerted changes at T > 240 K. This could be caused by changes in the topology of
voids (pores) where melting of water occurred resulting in increased mobility of water molecules
with increasing temperature. The observed downfield shift of methane signal with lowering
temperature could be caused by the temperature dependence of the sample’s magnetic
susceptibility [243]. For weakly hydrated unmodified nanosilica (h = 0.005 g/g), methane
adsorption decreased from 0.034 g/g at 200 K to 0.01 g/g at 280 K, opposite to the trend seen with
the partially silylated nanosilica. With more highly hydrated unmodified nanosilicas when h =
0.0450.15 g/g, methane adsorption was much lower (0.0010.002 g/g) over this temperature
range, again exhibiting decreasing adsorption with increasing temperature [104]. Thus, partial
silylation of nanosilica caused changes in the adsorption trend of methane with temperature in
comparison with the unmodified silica [72,104]. Water was adsorbed as SAW (45 ppm) and
WAW (0.81.5 ppm) (Fig. 118a). The downfield shift was observed for both water types, and this
shift was stronger for SAW possessing larger sizes. This occurred due to ordering of water clusters
on freezing (e.g., mobile water amorphous ice crystalline ice). Signal intensity decreased
with lowering temperature due to water fraction freezing.
133
To study changes of confined space effects, additional water was added to the HS2 sample
to form a wetted powder (h 1 g/g), then stirred and heated at 400 K for 10 min (bulk density ~0.3
g/cm3). The resulting sample contained a residual amount of bound water (h = 0.005 g/g). This
water gave a broad 1H NMR signal whose intensity could not be determined (Fig. 119a); therefore,
the amount of adsorbed methane was given in arbitrary units (Fig. 119b). Estimation of the CH4
signal noise level in the spectra of initial HS2 (Fig. 118a) compared to the additionally treated HS2
(Fig. 119a), suggests that the adsorption of methane onto treated HS2 was lower than on initial
HS2. This may be a result of changes in the confined space effects and the amounts of co-adsorbed
water located in narrow pores mostly able to bind adsorbed methane molecules. However, for the
treated HS2, there was a larger temperature dependence of the adsorbed amount of methane than
that for the initial HS2. This result could be caused by enhancement of confined space effects in
compacted nanosilica [72].
Fig. 119. (a) 1H NMR spectra recorded at different temperatures of water (h = 0.005 g/g) and methane bound to
compacted HS2, and (b) changes in the methane adsorption with temperature (errors < 10%).
Theoretically calculated 1H NMR spectra of water and methane in different structures (Fig.
120) showed that for the H atoms of methane molecules interacting with neighboring water
molecules in 10CH4@8H2O (curves 46), There was a downfield shift of ~1 ppm. This led to a
broadening of the methane band at 01 ppm (curves 46). The peaks at 3 and 5.5 ppm were linked
to the water molecules from the same clusters. This splitting resulted from calculation features of
clusters of restricted sizes, since experimentally there should be only one band due to fast
molecular exchange. Changes in the DFT functionals (in GIAO calculations) and optimization
type (DFT or HF) resulted in small differences in the spectra. Use of the PM7 method with the
calibration function gave the spectrum of 10CH4 (Fig. 120, curve 3), which was close to that
calculated using the GIAO method with B3LYP/631G(d,p)//HF/631G(d,p) (curve 2). Therefore,
this approach was applied to a large system (Fig. 120, curve 7). A downfield shift for methane
bound to hydrated TMSsilica nanoparticles was similar to that for 10CH4@8H2O. WAW appears
as a shoulder at 12 ppm (curve 7). SAW peak of maximum intensity was at 4 ppm, similar to the
experimental data (Fig. 91a). Additionally, there was a shoulder at 67 ppm (Fig. 120, curve 7)
similar to the experimental signal seen at low temperatures (Fig. 116b). These theoretical results
confirmed interpretation of the experimental 1H NMR spectra described above [72].
134
Fig. 120. 1H NMR spectra of water and methane in different structures: CH4 (curve 1), 10CH4 (2 and 3), and
10CH4@8H2O (curves 46) calculated with GIAO using B3LYP/631G(d,p)//HF631G(d,p) (curves 1, 2
(line), and 4) wB97XD//HF631G(d,p) (curve 5) or wB97XD (curve 6), and PM7 with a correlation
function (10CH4, 3 (symbols) and water molecules bound to TMSsilica nanoparticle with co-adsorbed
methane molecules, curve 7).
Hydrogen is poorly adsorbed onto any adsorbent at normal pressure because of very weak
van-der-Waals interactions. Nearly dry HS2 at h = 0.005 g/g was used here to study the interfacial
behavior of hydrogen co-adsorbed with a small amount of pre-adsorbed water [72]. Adsorbed
hydrogen was observed in the low-temperature 1H NMR spectra as a broadened signal at H = 4.5
ppm (Fig. 121a). This broadening suggested adsorption, since the resonance for H2 gas would be
much narrower at 4 ppm. The small upfield shift for adsorbed hydrogen was due to location of the
electron density in the H2 molecules mainly between the two protons. This led to a decrease in
their magnetic shielding upon hydrogen adsorption. Residual water was observed as both SAW at
H = 4.04.5 ppm and WAW at H = 1.01.5 ppm (Fig. 121a) [72]. Water was in a strongly
clustered state, since its contact area with the HS2 surface was very small (Table 20, S).
Additionally, there was a methane signal (added to hydrogen as a standard) at H = 0 ppm. In
contrast to hydrogen, water was frozen with lowering temperature that led to signal reduction at T
< 260 K and its disappearance at T 220 K.
The adsorption of hydrogen onto HS2 was very low ( 0.2 mg/g) and smaller by an order
of magnitude than that of methane. Hydrogen adsorption (Fig. 121b) increased with increasing
temperature, similar to the behavior of adsorbed methane (Figs. 118 and 119). Notice that despite
the low content of pre-adsorbed water (Table 20), it formed both smaller (~12 nm) and larger
(1030 nm) structures (Fig. 121c). This corresponded to known clustered adsorption of water [53]
enhanced by partial silylation of the silica surface. Melting of ice nanocrystallites located in
nanopores and narrow mesopores with increasing temperature, and enhanced mobility of water
molecules, could free up a portion of these smaller pores. Such pores were better able to adsorb
hydrogen molecules than larger mesopores. Therefore, the adsorption of hydrogen increased with
temperature [72].
135
Fig. 121. (a) 1H NMR spectra recorded at different temperatures of water (h = 0.005 g/g), hydrogen and methane
bound to HS2, (b) amounts of WAW, SAW and hydrogen vs. temperature (average errors were smaller than
10%), and (c) size distribution of pores filled by unfrozen water bound to HS2.
Nonpolar or weakly polar media could slow down molecular exchange between water
molecules from different clusters and domains located at an adsorbent’s surface. This could result
in narrowing or even splitting of bands in the 1H NMR spectra [53]. Addition of acids to the
dispersion media allows one to differentiate several types of SAW, which could dissolve acids or
salts differently [46,53,244]. Typically, interfacial WAW could not dissolve the acids. Here
trifluoroacetic acid (TFAA) was added to chloroform-d medium (CTFAA = 15 wt.% in CDCl3).
Before addition of TFAA, bound water was observed as SAW at 3.5 ppm at 280 K and 5 ppm at
210 K, as well as WAW at H = 1.01.5 ppm (Fig. 122a,c). The intensity of the SAW signal
decreased more rapidly than the WAW signal (Fig. 122c). Besides these signals of interfacial
water, signals of CHCl3 (as an admixture in CDCl3) and tetramethylsilane (as a chemical shift
standard added to CDCl3) were also observed. From comparison of the 1H NMR spectra of water
bound to HS2 without [72] and with CDCl3 (Fig. 122a), a disordering of SAW was observed due
to the weakly polar liquid dispersion medium. Evidence of disordering came from H values being
136
smaller in the presence of CDCl3, and the SBW amount decreased from 0.035 g/g to 0.023 g/g.
This was due to the displacement of a portion of bound water from the silica surface by CDCl3
(Table 20, S, V) leading to a decrease in the S value, and an increase in <T>. Despite the low
water content (h = 0.05 g/g), several types of water structures including nanoclusters at R < 1 nm,
larger clusters (R = 110 nm) and domains (R > 10 nm) were observed (Fig. 122d, Table 20) [72].
Addition of TFAA resulted in significant downfield shifts of SAW signals (H = 1213
ppm) due to fast proton exchange between TFAA and water molecules (Fig. 122b) [72]. At T >
240 K, two SAW signals (signals 1 and 2, Fig. 122b) corresponding to different amounts of
dissolved TFAA (more dissolved TFAA for water with signal 1) were observed. Besides SAW
with TFAA, a weak signal of SAW without dissolved TFAA (signal 3) was observed at 4 ppm. At
273250 K, signal 2 decreased more than signal 1, and a downfield shift was observed for signals
1 and 2. But at T < 250 K, an upfield shift was observed in parallel to a strong decrease in signal
intensity due to freezing of both TFAA and water. Besides signals of SAW, signals from WAW at
1.01.5 ppm were observed (Fig. 122b). The chemical shift of WAW was the same with and
without TFAA suggesting that WAW did not dissolve TFAA [72].
Theoretical calculations of the 1H NMR spectra (GIAO/DFT) of hydrated TFAA in
molecular and ionized states (Fig. 123) suggested that signals 1 and 2 (Fig. 122b) may be caused
by hydrated TFAA in similar states [72].
Fig. 122. 1H NMR spectra recorded at different temperatures of water (h = 0.05 g/g) bound to HS2 in the CDCl3
dispersion medium (a) without and (b) with addition of TFAA (15 wt.% with respect to CDCl3 amount), (c)
amounts of SAW and WAW vs. temperature without TFAA, and (d) size distribution of water structures
without TFAA.
137
Fig. 123. 1H NMR spectra of hydrated TFAA in molecular (curve 1) or ionized (curve 2) state calculated with
GIAO/B3LYP/631G(d,p)//HF/631G(d,p).
Despite partial hydrophobization (heterogenization resulting in a mosaic surface with
hydrophilic and hydrophobic patches) of the PS300 surface by attachment of TMS groups (TMS =
37.2%), modified silica could bind significant amounts of water (up to ~5 g/g) in an aqueous
suspension. However, only approximately 0.5 g/g of this water was strongly bound while the
major fraction of water was weakly bound. The presence of surface TMS groups caused the
appearance of weakly associated water (at chemical shift H = 12 ppm) at the interfaces, even in
the aqueous suspension of TMSsilica. The adsorption of methane onto partially silylated
nanosilica with pre-adsorbed water (0.05 g/g) increased with temperature in contrast to the
adsorption of methane onto unmodified silica where a reduction of methane adsorption was seen
with increasing temperature. It was believed that changes in both confined space effects and the
temperature dependent organization of interfacial water may explain these results [72].
Changes in the organization of interfacial water were observed upon changes in the types
of dispersion media and co-adsorbates [72]. In weakly polar CDCl3 medium, interfacial water
existed in states that were strongly (chemical shift H = 45 ppm) and weakly (H = 12 ppm)
associated, as well as strongly (changes in the Gibbs free energy ΔG > 0.50.8 kJ/mol) and
weakly (ΔG < 0.50.8 kJ/mol) bound [53]. Water in these different forms resulted in differences
in their activity as solvents. WAW did not dissolve trifluoroacetic acid, but some SAW could
dissolve TFAA. In addition, the SAW activity differs for different fraction of SAW (e.g., located
in narrower and broader pores). This appeared as downfield and upfield shifts of water/TFAA
clusters and domains with temperature. Therefore, the increase in the adsorption of methane onto
partially silylated nanosilica with increasing temperature could be explained by enhanced mobility
of unfrozen water which was displaced from some narrow pores, and these pores could be
occupied by methane [72].
Typically, water confined in narrower pores was characterized by a smaller average
number of the hydrogen bonds per a molecule [53,245]. Therefore, the value of H of confined
water decreased in comparison with bulk water (Fig. 124). Notice that H 1 ppm was
observed for individual water molecules in the gas phase or dissolved in nonpolar or weakly
polar organic solvents, H 4-5 ppm was for bulk water and H 7 ppm was for Ih ice [53].
The hydrogen bond network structure of bound water differed from that of bulk water or ice
and depended on the topology of pores and chemical composition of the pore walls affecting
the energy of the hydrogen bonds (EH) [53]. The hydrogen bonds between water molecules
and active surface sites (hydroxyls) of silica or other oxides were stronger (EH = 40-50
kJ/mol) than the hydrogen bonds between water molecules (EH = 25-28 kJ/mol) [53].
Therefore, the heat of water adsorption on hydrophilic adsorbents was greater than the latent
heat of bulk water condensation (Q = 45 kJ/mol). This led to reduction of the Gibbs free
energy (G < 0) of interfacial water in comparison with the bulk depending on the pore size
138
and chemical structure of the pore walls [245].
Fig. 124. Temperature dependence of the H value for bulk water (line without symbols) and water adsorbed on (a)
nanosilicas and silica gels at h = 0.21 (A-150), 0.163 (A-200), 0.24 (A-380), 0.12 (Si-40), 0.115 (Si-60),
and 0.114 (Si-100) g/g, and (b) fumed alumina and SA at h = 0.2 (Al2O3), 0.37 (SA3) and 0.223 (SA23) g/g.
The H values of bound water characterize both topological and surface nature effects
on this water (Fig. 124) [53,245]. The values of H were larger for water adsorbed on
mesoporous silica gels than on fumed silica, alumina and SA composed of nonporous primary
nanoparticles forming aggregates and agglomerates of aggregates [245].
Features of the primary particle size distribution ((a)) and their aggregation in
secondary particles (i.e. textural porosity characteristics) influence the diffusion of water
molecules in concentrated aqueous suspensions of nanosilicas [53]. Therefore transverse
relaxation time (T2) was two-three times longer (i.e. the molecular mobility was higher) for
water in the aqueous suspensions of A-50 (average diameter of primary particles d 52 nm)
than that for A-300 (d 9.3 nm) at the same silica concentrations. Additionally, the
temperature effects were stronger for the A-50 suspension. In other words, at the same weight
concentration of nanosilicas, water was more strongly bound in the suspension with smaller
particles (A-300) because of larger amounts of the interfacial water in the suspension of A-300
(by several times) because this silica had much larger SBET value and more strongly aggregated
primary particles than A-50. The 1H NMR spectroscopy investigations of a variety of
nanosilicas and other oxides showed that concentrations of surface hydroxyls and their acidity
could strongly influence the amounts and characteristics of bound water in the aqueous
suspensions [53]. Therefore, not only the morphology, the (a) distributions, and the SBET
values but also the surface structure of nanoparticles could affect the molecular mobility, the
dipolar and direct current (dc) relaxations and other dynamic processes in interfacial water
[245].
139
Fig. 125. Temperature dependences of the amounts of unfrozen water (Cuw) and TSD current (dc) on the adsorption of
water on (a) ordered mesoporous silicas MCM-48 (hydration h = 0.25 (NMR) and 5 g/g (TSDC)) and SBA-
15 (0.19 (NMR) and 5 g/g (TSDC), (b) silica gels Si-40, Si-60 and Si-100 (h = 19 (NMR) and 5 g/g (TSDC)),
(c) activated carbon C-47 (h = 4.71 (NMR) and 13.3 g/g (TSDC)); and (d) nanosilica A-300 at CA-300 = 3.0
(TSDC, h 32.3 g/g), 3.7 (NMR, h 26.0), 7.4 (h 12.5), 12.3 (h 7.1) and 16.7 wt.% (h 5.0 g/g).
The 1H NMR spectra were linked to mobile water molecules [53] but the dc relaxation
(TSDC) [214] was caused by mobile protons and other ions when condition of throughout
percolation of ions (between two electrodes in a TSDC cell) was achieved on heating of a
frozen system [53,245]. Both dynamic phenomena were temperature dependent (Fig. 125). For
all samples, a linear dependence of ln(ITSDC) versus 1/T was observed for the dc relaxation; i.e.
it obeys the Arrhenius law. The mobility of water molecules (NMR) demonstrated more
complex character caused by non-Arrhenius or several Arrhenius processes since two-three
linear portions were observed for the curves of ln(Cuw(T)) versus 1/T [245]. This effect could
be explained by the presence of several types of interfacial water (i.e., SAW, WAW, SBW and
WBW) located in different pores at different distances from the oxide surfaces [53] and
responsible for different local molecular mobility of interfacial water observed at different
temperatures (due to layer-by-layer freezing-out of confined water with lowering temperature).
Additionally, certain cooperative effects could be observed in water bound in nanopores or
interacting with surface functionalities because of a clustered structure of bound water
characterized by a dense hydrogen bond network that led to cooperative relaxations of water
clusters or nanodomains [53,214]. Typically, a more complex and broader PSD led to a more
complex lnCuw(1/T) function (Fig. 125) [245].
The complexity of the behavior of interfacial water reflecting in complex temperature
dependences of the molecular mobility and related characteristics reflected in a broad
distribution of the activation energy of the local molecular mobility (LMM) at ELMM = 5-83
kJ/mol (NMR) and the dc relaxation at Edc = 46-98 kJ/mol (TSDC) calculated assuming that
the processes obey the Arrhenius law, and typically ELMM < Edc for the same systems [245].
The ELMM values were greater for the suspensions with A-300 (maximal values ELMM = 46 and
52 kJ/mol at h = 7.1 and 5 g/g, respectively) than ELMM for porous silicas: 40 (MCM-48), 25
140
(SBA-15), 34 (Si-100), 18 (Si-60) and 21 kJ/mol (Si-40) [245]. These results could be due to
larger SBET values of porous silicas than that of nanosilicas and stronger confined space effects
in narrower pores. These effects led to diminution of the average number of the hydrogen
bonds per a molecule that diminish the ELMM values. Calculations of the distribution functions
of activation energy (f(E)) of the local mobility (NMR) [53] and the dipolar and dc relaxations
(TSDC) [6] in the interfacial water at a surface of polymer adsorbent LiChrolut EN [214] (Fig.
126) and nanosilica A-300 (as representative samples with high nanoporosity (C-47) or
textural porosity (A-300)) revealed the presence of several dynamic processes in bound water
studied by both methods. The dipolar relaxation (TSDC) [214] was observed at temperatures
(90 K < T < 210-220 K) lower than that characteristic for appearance of the local molecular
mobility (NMR) [245].
The throughout molecular mobility causing the dc relaxation (TSDC) appeared at
higher temperatures (T > 210-230 K) than local molecular mobility (NMR) and both processes
depended (but differently) on the confined space effects [245]. These results were due to the
influence of surface electrostatic fields on the structure of the hydrogen bond network in
interfacial water (i.e., the average number of these bonds per a molecule and their strength)
and due to additional conditions necessary for the dc relaxation (ion percolation) in
comparison with the local molecular mobility. An increase in the content of an adsorbent (and
therefore contribution of interfacial water) led to diminution of the low-temperature (LT) peak
of the dipolar relaxation at 120 K characteristic for bulk water and related to the dipolar
relaxations of OH groups in small clusters (without the cooperative effects) and the relaxation
of interstitial water molecules (without the hydrogen bonds and H 1 ppm) [53,245].
Fig. 126. Distribution functions of the activation energy of the mobility of water molecules at a surface of polymer
LiChrolut EN adsorbent calculated on the basis of the 1H NMR data (curve 1, hydration h = 2 g/g) and
dipolar and dc relaxation (TSDC, curves 2 (regularization) and 3 (piecewise linearization), h = 5.4 g/g).
Complex nanooxides (silica/alumina and silica/titania) have strong Brønsted acid sites
(SiO(H)M where M = Al or Ti) and a set of weaker sites (bridging (MO(H)M) and
terminal (MOH, SiOH) hydroxyls, Lewis acid sites, etc.) that affected the behavior of
interfacial water (Fig. 97) [53,245]. The chemical nonuniformity of the SA, ST or AST
surfaces could cause a stronger non-Arrhenius character of the dc relaxation and the molecular
mobility in the aqueous suspension (Fig. 127) in comparison with nanosilica [245]. The
molecular mobility of interfacial water appears in the aqueous suspensions of SA at lower
temperatures than that for nanosilica (Fig. 100). However, the dc relaxation began at
temperatures close to that for nanosilica. Consequently, changes in the local structures in
interfacial water and the temperature behavior of this water (NMR) more strongly depended on
the oxide surface structure than the ion percolation effects (TSDC) depended on the
composition of oxide particles and their aggregation in the suspensions [245].
141
Fig. 127. Amounts of unfrozen water (NMR, curves 1-4) and dc relaxation (TSDC, curves 5-8) as functions of
reciprocal temperature for fumed Al2O3 (curves 1 and 5), SA1 (2 and 6), SA3 (3 and 7) and SA23 (4 and 8)
at Cox = 5 wt.% (1-4) and 3 wt.% (5-8).
Comparative investigations of water confined in pores of solid materials and bioobjects
using the 1H NMR, TSDC and microcalorimetry methods showed that the difference in
temperature of transition from the local molecular mobility of water bound at the pore walls or
intracellular functionalities to ion percolation effects (throughout conductivity) could be 10-40
K [245]. This value depended on several factors such as the morphology of particles, textural
and structural characteristics of adsorbents, chemical structure of the surfaces, concentrations
of components, the presence of dissolved salts (e.g., NaCl) and other compounds. There were
four types of water confined in solid adsorbents and bioobjects such as strongly (G < 1
kJ/mol) and weakly (G > 1 kJ/mol) bound and strongly (H = 4-5 ppm) and weakly (H = 1-
2 ppm) associated waters. At low amounts of adsorbed water (much smaller than the pore
volume) typically all water was strongly bound. A portion of this water could be weakly
associated if the pore walls were mosaic and composed with hydrophilic and hydrophobic
patches, e.g., for partial silylation of silica surface. The local molecular mobility of bound
water depicted a non-Arrhenius character or included several Arrhenius-type processes
characterized by different activation energies because the corresponding water was differently
clustered and bound to different structures at the interfaces. Therefore, the dynamic
characteristics of these types of bound water could significantly differ. The temperature
dependence of ion percolation in bound water has the Arrhenius character for relatively
uniform systems such as the aqueous suspensions of nanosilica or mesoporous silicas.
However, for complex materials (e.g., complex nanooxides) or adsorbents possessing both
broad and very narrow pores (e.g., activated carbons) certain deviations from the Arrhenius
behavior were observed on the dc relaxation. The behavior of interfacial water depended
strongly on the presence of nonpolar or polar organic solvents which could displace a portion
of water from narrow pores into larger ones, change the associativity of bound water and its
interaction energy with solid surfaces or intracellular functionalities. Water and polar organics
confined in narrow pores were much poorly complex than in the bulk that led to enhanced
clusterization of water bound to the mosaic surfaces [53,245].
The 1H NMR spectra of mobile water and water/methane at T < 273 K were recorded
without the use of the MAS technique [53,104]. The spectral bandwidth (non-MAS) of surface
silanols and OH groups from ice was much broader than mobile water. However, immobile OH
structures could affect the baseline of the spectra of mobile water bound to silicas characterized by
different PSD (Fig. 128). The values of H of free unbound silanols was 1.5-2.0 ppm (Fig. 129,
rotor frequency fr = 9 kHz, curve 2) [218-222]. This band was absent for A-300 at h = 0.05 (Fig.
142
129, curve 1, fr = 0 kHz) and 0.08 g/g [104]. However, the O-H stretching vibrations of free
silanols were observed at 3747 cm1 in the FTIR spectra of A-300 at a similar h value, because
free silanols could be observed at the nanosilica surface at h < 0.2 g/g due to clustered adsorption
of water [53]. The bands at H = 1-2 ppm (i.e. in the range of free silanols observed in the 1H MAS
NMR spectra of silica (Fig. 129) [218-222]) were observed for silica 200DF at h 0.005 g/g
[104]. However, intensity of these bands decreased with lowering temperature and they were not
observed at 210 K. This behavior was more characteristic of weakly associated but strongly bound
water than for free silanols [53]. Additionally, the O-H stretching vibrations of free silanols were
not observed for silica 200DF; i.e. all the SiOH groups were hydrogen bonded [104]. Notice that a
small portion of water could be dissolved in chloroform and contribute to the spectra at H 1.2
ppm. The band at H = 3-4 ppm, which was not observed at T < 240 K, could be assigned to
strongly associated water with a major fraction of weakly bound water (frozen at T > 260 K). As a
whole clustering of water adsorbed on nanosilica diminishes with increasing amounts of water; i.e.
it became more strongly associated [53]. Therefore 1H NMR signal intensity decreased faster with
lowering temperature for A-300 than 200DF with a greater amount of adsorbed water, which was
readily seen at T 240 K. Thus, this NMR technique allowed one to observe the 1H NMR spectra
of mobile water, in contrast to the 1H MAS NMR technique giving the spectra of all OH structures
(silanols, bulk and adsorbed liquid and frozen water). This technique, with layer-by-layer freezing-
out of bound water at 200-280 K, could give detailed information on co-adsorption of methane and
water on the studied silicas [104].
Fig. 128. Incremental pore size distributions of silicas A-300 (1), A-380 (2) and 200DF (3) calculated from the
nitrogen adsorption/desorption isotherms (77.4 K) using the DFT method. A model of cylindrical pores was
used for 200DF, and a model of pores as voids between spherical nonporous primary particles forming
aggregates of random structures for nanosilicas was employed.
143
Fig. 129. 1H NMR spectra of initial A-300 (h 50 mg/g) recorded at 293 K using the MAS technique (rotor frequency
fr = 0 (1) and 9 (2) kHz).
The 1H NMR spectra of water and methane co-adsorbed on nanosilica A-300 at different
hydration levels were recorded over the 200-280 K range [104]. These spectra include a single or
weakly-split signal of methane at H = 0.0-0.3 ppm. This signal slightly increased with lower
temperature due to increasing adsorption of methane, and an enhanced interaction with the silica
surface. Water was responsible for two signals at H ≈ 1 and 4-5 ppm corresponding to weakly
associated water (WAW) with a significant portion of unbound H atoms, and strongly associated
water (SAW), with dipolar coupled molecules due to OHO hydrogen bonds, respectively. The
presence of several 1H NMR signals of an adsorbate suggests slow (on the NMR timescale)
molecular exchange between the corresponding states [53,243]. The signal intensity of SAW at H
≈ 4-5 ppm decreased, but the H value increased, with lower temperature because of layer-by-layer
freezing-out of fractions of more weakly bound SAW and WAW. However, complete freezing of
WAW was observed only at h = 0.005 g/g and T < 220 K, and at h > 0.03 g/g only a certain
portion of WAW was frozen at 200 K. Equilibration of the water/A-300 system for a week led to
several important effects [104]: (i) the adsorption of methane decreased for the equilibrated
compared with the non-equilibrated system at the same h values; (ii) water became more strongly
associated due to equilibration, i.e. it became less clustered with a greater number of hydrogen
bonds per molecule; (iii) this latter effect was stronger with increasing hydration of nanosilica as
well as silica 200DF (Fig. 130). For the latter clustered adsorption of water was observed only at
minimal hydration h = 0.005 g/g (shoulder at H = 1-2 ppm). At greater amounts of water adsorbed
onto silica gel 200DF only SAW was observed at H 4-6 ppm. This led to strong diminution of the
adsorption of methane with increasing amounts of pre-adsorbed water [104].
When h < 0.3 g/g all water was strongly bound to the nanosilica surface because it freezes
at T < 260 K, and changes in the Gibbs free energy G < 0.5 kJ/mol (Figs. 130 - 132) [104]. Four
of six non-equilibrated A-300/H2O/CH4 samples (with the exception of samples at h = 0.005 and
0.1 g/g), showed an increase of WAW with lower temperature (Fig. 130b). The most methane
adsorption was observed at h = 0.1 g/g (Fig. 130c) when the amount of WAW was largest and the
system was non-equilibrated (Fig. 130b). Equilibration of this system for 7 days caused significant
diminution of the adsorption of methane. However, this effect was much smaller at lower
hydration (Fig. 130c) because adsorbed water was more clustered at smaller amounts despite the
long equilibration time. Notice that changes in the G and S values after equilibration were
relatively small (Fig. 131); however, changes in contributions of SAW (Fig. 130a) and WAW (Fig.
131b) were significant. Consequently, changes in the adsorption of methane after pre-equilibration
of the A-300/water system were more influenced by structural changes in adsorbed water (e.g.,
144
increase in the associativity and the size of adsorbed water structures) than because of changes in
its energetic characteristics [104].
Fig. 130. Temperature dependences of concentration of (a) SAW, (b) WAW and (c) methane co-adsorbed on
nanosilica A-300 and (d) relationship between concentrations of methane and WAW.
There was a correlation between the amount of adsorbed methane and WAW (Fig. 130d).
Lower amounts of adsorbed methane were observed for samples at minimal (h = 0.005 g/g) and
maximal (1 g/g) hydration of silica [104]. The former included too little water to form effective
nanoporosity (between the surface of adjacent silica nanoparticles, clusters and domains of
unfrozen water and ice nanocrystallites), and the latter included too much water which could form
a nearly continuous surface SAW film (Fig. 130a), and bound water was less clustered. Strong
clustering of bound water was a necessary condition for the maximal adsorption of methane onto
nanosilica [104].
145
Fig. 131. (a) Relationship between changes in the Gibbs free energy and the SAW amounts on co-adsorption of water
and methane onto nanosilica A-300, and (b) normalized interfacial energy (s/h or s,SAW/CSAW) as a function
of water content (SAW* corresponds to samples equilibrated for a week) for A-300 and 200DF.
Fig. 132. Temperature dependence of the concentration of (a) SAW, (b) WAW and (c) methane co-adsorbed on silica
200DF and (d) relationship between concentration of SAW and changes in Gibbs free energy of bound
water.
The value of interfacial energy (s), determined as the modulus of the total decrease in the
Gibbs free energy of the adsorbent-adsorbate system [53], increased with increasing water content
[104]. However, normalized interfacial energy (s,SAW/CSAW for SAW and s/h for all bound water)
decreased at h = 0.3 and 1.0 g/g (Fig. 131b), because the influence of the silica surface decreased
for distant layers of bound water. For a sample at h = 0.1 g/g the s/h and s,SAW/CSAW curves have
a maximum. This was due to the equilibrium state of intact sample (with water adsorbed from air
and equilibrated for long periods) compared to the incomplete equilibration of heated (at 150 oC to
h = 0.005 g/g) and subsequently wetted samples. Complete equilibration of treated and wetted
nanosilica required several days due to the slow penetration of water molecules into the particle
146
volume [37,38,53]. Notice that the s/h and s,SAW/CSAW maxima at h = 0.1 g/g corresponded to the
largest contribution of WAW and greatest adsorption of methane [104].
The relative contribution of WAW of total bound water was maximal (up to 20%) at the
smallest hydration (0.005 g/g) of silica, and minimal (< 1 %) at the largest hydration (1.0 g/g)
[104]. Consequently, the characteristics of bound water strongly corresponded to the
characteristics of SAW, the major contributor of bound water. There were four-three types of
SAW structures (Figs. 133 and 134); e.g., clusters, nanodomains and microdomains. These SAW
structures became frozen at different temperatures depending on the strength of their interactions
with the silica surface, reflecting changes in the relationship between G and CSAW (Fig. 131a). A
displacement of the dCSAW/dT maxima toward higher temperatures with increasing h value (Fig.
133) was seen because of the reduced interaction of SAW fractions with the silica surface with
growing hydration. This results in the appearance of weakly bound water (characterized by G >
0.5 kJ/mol and frozen at T > 260 K) at h 0.3 g/g (Figs. 130 and 131), due to the formation of
relatively large water structures. These water structure sizes were determined by the Gibbs-
Thomson (GT) relation for the freezing point depression of confined liquids [53], were relatively
large (Fig. 134), and frozen at T > 260 K (Fig. 133). Notice that the relative contribution of
nanopores was largest at h = 0.1 g/g; however, absolute values were larger at h = 0.3 g/g. At h = 1
g/g the Snano and Vnano values strongly decreased but the Smeso and Vmeso values strongly increased.
These results corresponded to significant enhancement of water associativity with increasing h
value [104].
For weakly hydrated samples at h 0.1 g/g, the maximal contributions of dC/dT and f(R)
(Figs. 134 and 135) were for structures with strongly bound water frozen at T < 230 K with sizes R
< 3 nm [104]. These water structures, attributed to clusters at R < 1 nm and nanodomains at R = 1-
3 nm, corresponded to nanopores and narrow mesopores, respectively. Water partially filling
narrow pores at R < 10 nm between primary silica nanoparticles provided enhancement of
effective nanoporosity at R < 1 nm. This results in an increase in the adsorption of methane,
especially at h = 0.1 g/g, when the contribution of small water structures at R 1 nm was maximal
as the relative contribution of nanopores filled by bound unfrozen water [104].
The molecular mobility of water could appear in different structures where water
molecules have different numbers of hydrogen bonds (per molecule) of different strength.
Therefore, the distribution functions of the activation energy (f(E)) of the molecular mobility
(calculated using the temperature dependences of dCSAW(T)/dT assuming that this process obeys
the Arrhenius law) [53,245] were relatively broad over the 15-85 kJ/mol range (Fig. 136a) [104].
This range of E values corresponded to the motion of water molecules having from one to four
hydrogen bonds per molecule.
The activation energy of the molecular motion of methane (determined from the
temperature dependences of dCCH4/dT) was 1-10 kJ/mol for a minimally hydrated sample (Fig.
136b). However, for more strongly hydrated silica (h = 0.1 and 1.0 g/g) the values of E increased
by several times. This could be explained by several reasons. First, enhanced nanoporosity formed
by water clusters, nanodomains and ice nanocrystallites, partly or completely frozen with lowering
temperature, enhanced the barriers for methane motion. Second, complex water/methane structures
could demonstrate cooperative motion; therefore, the activation energy of the methane motion
depended on the water motion. Thus, co-adsorbed water and methane were not individual phases
independent from one another. This was also confirmed by the correlation between the amounts of
adsorbed methane and WAW (Fig. 131d) [104].
147
Fig. 133. Temperature derivatives of the SAW content for (a) weakly and (b) strongly hydrated samples; dCWAW/dT
was shown for a sample at h = 0.005 g/g (a, curve 2).
Fig. 134. Distribution functions of sizes of unfrozen bound water structures at (a) low and (b) high hydration and the
pore size distribution of A-300 calculated using modified Nguyen-Do equation with the model of cylindrical
pores (a, curve 5) and DFT method with the model of voids between spherical particles (b, curve 4).
The adsorption of methane onto nanosilica A-300, composed of nonporous primary
nanoparticles (average diameter 8.1 nm) at standard pressure, was a function of temperature and
silica hydration. The silica hydration dependence was nonlinear, and maximal adsorption of
methane (1.9-1.2 wt.% at 200-280 K) was observed at hydration h = 0.1 g/g for intact silica. A
decrease (on heating) and increase (on wetting) of the silica hydration both led to a reduction of
methane adsorption. Co-adsorption of methane and water led to the appearance of a 1H NMR
signal from weakly associated water at H 1 ppm. The amount of this water correlated to
concentration of adsorbed methane, because weakly associated bound water was most clustered at
the surface of nanosilica composed of nonporous primary nanoparticles. The adsorption of
methane on micro/mesoporous silica 200DF decreased with increasing amount of pre-adsorbed
water characterized by significant associativity (H ≈ 5 ppm) at h 0.005 g/g [104].
148
Fig. 135. Distribution function of sizes of unfrozen water structures (curve 1, h = 0.15 g/g) and pore size distribution
of silica 200DF (2).
Fig. 136. Activation energy of molecular motion of unfrozen water at T < 273 K and different hydration of A-300 and
200DF with co-adsorbed methane.
Thus, to enhance the adsorption of methane onto adsorbents with pre-adsorbed water at
standard pressure and T < 280 K, this water should exhibit maximal clustering. The amount of
clustered water should be equal to a value characteristic for a given adsorbent, to maximize the
contribution of narrow nanopores appropriate for the adsorption of methane [104].
Dry powder of A-380 could be characterized by certain structural hierarchy of particles
starting from primary particles (5-15 nm in size, average diameter 7.2 nm), aggregates of primary
particles and agglomerates of aggregates [53,246]. Adsorbed water at low amounts (< 20 wt.%)
localized mainly in the zone of contacts between adjacent nanoparticles. At higher amounts,
adsorbed water could form continuous layers covering whole nanoparticles [53]. Consequently, for
two samples at h = 0.5 and 1 g/g continuous water layers should be formed but for other samples at
h = 0.07, 0.08, and 0.15 g/g island adsorption of water could be to form only partial coverage of
the surface. Despite this structural difference, water bound to nanosilica gives a single 1H NMR
signal at H = 4-5 ppm, i.e. it was SAW, whose intensity decreased with lowering temperature at T
< 250 K due to partial freezing of water. The H value decreased with elevating temperature
because of the influence of thermal motion of water molecules on the hydrogen bond network
structure. There were fractions of SBW and WBW but WAW was absent. Addition of organic
solvents to A-380/water slightly affected the shape and the position of the water signal. A weak
signal of the CHD2 groups of admixtures in CD3CN and DMSO was observed at H = 2-2.5 ppm.
149
Co-adsorbed DMSO (in a mixture with CDCl3) more strongly affected freezing-out of water since
even at 210 K an intensive signal of unfrozen water was observed [246].
Table 21. Characteristics of bound water in hydrated powders of nanosilica A-380 with the presence of organic solvents [53,246].
h
(g/g)
Organics
(g/g)
CSBW
(mg/g)
CWBW
(mg/g)
�-GS
(kJ/mol)
S
(J/g)
Suw
(m2/g)
Snano
(m2/g)
Smeso
(m2/g)
Smacro
(m2/g)
Vnano
(cm3/g)
Vmeso
(cm3/g)
Vmacro
(cm3/g)
- 250 750 2.52 30.2 116 14 102 0.4 0.007 0.976 0.007
1 СDCl3 150 850 2.87 23.0 150 35 115 0.5 0.016 0.966 0.008
2 СDCl3 170 830 2.37 27.0 130 0 130 0.1 0 0.996 0.002
1
6 СDCl3 30 970 2.57 21.6 121 6 115 0.3 0.003 0.985 0.005
- 125 375 2.84 16.5 87 43 43 0.5 0.020 0.451 0.007
2 СDCl3 100 400 2.59 15.8 56 23 32 0.8 0.011 0.449 0.014
0.5
8 СDCl3 75 425 3.16 15.2 53 19 34 0.7 0.009 0.458 0.011
- 100 50 2.78 11.1 78 67 11 0.1 0.029 0.117 0.001 0.15
3 СDCl3 100 50 2.43 9.3 40 28 12 0.1 0.013 0.134 0.001
6 СDCl3 100 50 2.70 8.8 10 0 10 0.1 0 0.145 0.002
0.5 DMSO 100 50 3.37 14.7 181 177 4 0.4 0.073 0.055 0.006
0.5DMSO+1.5CDCl3 145 5 3.35 21.4 227 221 6 0 0.094 0.056 0
- 60 20 3.48 7.8 75 71 4 0.2 0.029 0.043 0.002
2.0 CDCl3 55 25 2.47 4.6 5 1 4 0.1 0 0.074 0.002
1.0 DMSO 70 10 3.92 12.3 178 177 1 0 0.065 0.013 0
0.08
1 DMSO+2 CDCl3 70 10 3.76 10.7 116 114 2 0 0.046 0.029 0
1 CD3CN+0.5 CDCl3 35 35 2.32 3.1 15 0 15 0 0 0.054 0
1 CD3CN+1 CDCl3 35 35 2.42 3.3 17 0 17 0 0 0.059 0
0.07
1 CD3CN+2 CDCl3 30 40 2.14 2.9 15 0 15 0 0 0.056 0
Note. Suw, Snano, Smeso and Smacro were the surface area total, nano- (radius R < 1 nm), meso- (1 < R < 25 nm) and macropores (R > 25
nm), respectively, of silica in contact with unfrozen water; Vnano, Vmeso, and Vmacro were the pore volume of nano-, meso-, and
macropores, respectively [53,246].
To study features of the influence of organic solvents on the characteristics of bound water
localizing in voids between adjacent silica nanoparticles in their aggregates on the basis of the
temperature dependences of changes in the integral intensities of the signals of unfrozen water, the
relationships between changes in Gibbs free energy on water adsorption and the amounts of
unfrozen bound water were determined (Fig. 137). At relatively high level of hydration the
influence of organic solvents was weaker than that at low hydration. For the former, There were
near vertical (related to SBW) and horizontal (WBW) sections in the G(Cuw) graphs. The
maximal amount of SBW was 250 mg/g which corresponded to approximately third volume of
pores (voids between adjacent primary particles in aggregates) filled by nitrogen at 77.4 K.
Addition of chloroform alone or in a mixture with DMSO led to stronger changes in the
characteristics of bound water at lower hydration (Table 21, Fig. 137). The effects of chloroform
in a mixture with acetonitrile were smaller than that of CDCl3+DMSO. These effects could be
explained by the difference in the solubility of water in these organics individual and in mixtures.
Co-adsorbed chloroform decreased the interfacial energy (S) of bound water at different hydration
(Table 21) [246].
150
Fig. 137. Relationships between content of unfrozen water and changes in the Gibbs free energy of unfrozen water in
differently hydrated powders of nanosilica A-380 at h = (a) 1, (b) 0.15, (c) 0.07 and different amounts of
organic solvents [246].
The interfacial energy normalized to h = 1 g/g as S*= S/h increased with decreasing h value
because of a decrease in the amounts of WBW. There was tendency of a decrease in the S* value
with increasing content of chloroform. This could be explained by the displacement of a portion of
water from narrow voids to broader ones, and this effect was stronger at lower hydration (Fig. 137)
when bound water did not form continuous adsorption layers and co-adsorbed organics could
easily contact the silica surface. Such electron-donor solvents as CD3CN and DMSO were well
dissolved in water and chloroform. However, chloroform-DMSO (CD3CN) mixtures could stratify
on addition of a small quantity of water to form layers enriched by chloroform and an electron-
donor. To avoid this effect such quantities of mixtures were added to the silica powder to prevent
the formation of a liquid phase. CD3CN and DMSO differently affected the interaction of water
with the silica surface (Table 21, Fig. 138) since CD3CN enhanced but DMSO diminishes the
interfacial energy value of bound water [246].
151
Fig. 138. Size distribution of unfrozen water structures for hydrated A-380 powder at h = (a) 1, (b) 0.15, (c) 0.07 and
different amounts of organic solvents [246].
152
Fig. 139. 1H NMR spectra of aqueous 30% solution of H2O2 (a) bulk and adsorbed (10 wt.%) onto A-400 being in (b)
CCl4 and (c) CDCl3 media.
On the co-adsorption of water and studied organics, there were several factors affecting
changes in the structure (clusterization) of bound water. If only weakly polar chloroform was co-
adsorbed with water that the phase boundary tends to be minimal. Therefore, if pre-adsorbed water
formed a continuous coverage of the silica surface that the effects of chloroform were minimal.
These effects increased if the island coverage with pre-adsorbed water was realized (Fig. 137).
Chloroform in a mixture with polar organics could be mixed with bound water because DMSO
effectively interacting with both polar water and nonpolar chloroform could separate water and
chloroform clusters. There was no tendency to decrease the contact area between water and
chloroform at the silica surface because they contact only with DMSO and the silica surface. This
was well seen from the structural characteristics of bound water (Table 21) since maximal specific
surface area (total and of nanopores) of unfrozen water structures was observed for the
DMSO+CDCl3 (or DMSO alone). However, even for these cases these values were much smaller
than the SBET value of silica A-380. Notice that S < SBET on the adsorption of water alone at h =1
g/g. This could be explained by structural nonuniformity of bound water (i.e. its clusterization,
formation of nanodomains (Fig. 138) and absence uniform continuous water layers covering silica
nanoparticles) and the absence of contacts between total surface area of silica particles and water
[246].
153
Water bound to nanosilica did not form uniform layers even at relatively high hydration (1
g of water per 1 g of dry silica). Co-adsorption of water with weakly polar chloroform alone or in a
mixture with polar DMSO or acetonitrile more strongly affected the structure of water bound to
nanosilica A-380 at lower hydration when significant portion of silica surface remained uncovered
by water. Maximal clusterization of bound water was observed on the co-adsorption with
chloroform/DMSO mixture because DMSO could effectively interact with both nonpolar (through
dispersion interactions of chloroform with CH3 groups) and polar (water interaction with S=O)
structures [246].
The aqueous solution of hydrogen peroxide was characterized by two 1H NMR signals at
H 6 (H2O) and 12 (H2O2) ppm (Fig. 139a) [247]. The signal width increased with increasing
temperature but stronger for H2O2 due to enhancement of the proton exchange H2O ↔ H2O2. A
weak signal at H 2 ppm could be assigned to methyl groups of acetone added as a standard for
determination of the chemical shift. The chemical shift of water interacting with H2O2 was 1-1.5
ppm larger than that for bulk liquid water. Symbols at the peaks (Fig. 139a) showed the integral
intensity of these peaks. At T < 230 K the intensity drops down because of water and hydrogen
peroxide freezing. Complete melting of the frozen components occurred at Tm = 250 K. The ratio
of the integral intensity of the signals was 4:1 that was close to that of their molar fractions. A
decrease in the intensity at T > Tm was due to changes in nuclear level population (Curie law)
[243]. Additionally, H2O2 could decompose with elimination of oxygen, which solving in water
could provide additional channel (paramagnetic centers) of relaxation of the nuclear
magnetization. It was possible that the presence of dissolved oxygen in the H2O2 solution could be
a reason of a stronger manifestation of the Curie law in comparison with other similar
heterogeneous systems [53]. The intensity of the H2O2 signal (Fig. 139a) decreased with
increasing temperature much stronger than that of water. This could be explained by changes in
the component clusterization (due to cryoconcentration) which decreased with increasing
temperature. There were two states of H2O2 characterized by slow and fast proton exchange (in
NMR time scale) and the contribution of the latter increased with increasing temperature.
According to clustered structure of water, liquid water is composed of clusters and
nanodomains, state of which could be changed due to external actions, solutes, co-solvents,
confined space effects, etc. [53,242]. For such miscible liquids as water and hydrogen peroxide,
more clustered structure could be formed at lower temperatures (Fig. 140, right picture) [247].
This clustered structure could be destroyed with increasing temperature due to enhanced motion of
the molecules forming random structure (Fig. 140, left picture). The latter was characterized by
broadened 1H NMR spectrum theoretically calculated (Fig. 140). For the solution, the water peak
shifted toward weak magnetic field (curves 2w and 3w) due to interaction of water molecules with
hydrogen peroxide molecules [247].
At T < 250 K water (SAW) mixed with H2O2 was partially unfrozen but characterized by
lower activity as a solvent worse dissolving H2O2. Therefore, the latter tends to be more clustered
(as shown in Fig. 140, right picture). Greater H values for water than that for pure liquid water
was due to the hydrogen bonding of water molecules with H2O2 molecules. With increasing
temperature, the solubility of H2O2 in SAW increased that resulted in enhancement of the proton
exchange between water and hydrogen peroxide. One could expect the appearance of broad but
non-split signal of the H2O-H2O2 mixture shifted toward weak magnetic field in comparison with
water. However, this effect was absent since the separated signals of water and hydrogen peroxide
were observed at T > 250 K (Fig. 139a). This effect could be explained by the existence of the
clustered H2O2 structure even at 280-290 K [247].
154
Fig. 140. 1H NMR spectra of clusters calculated using PM6 method and the correlation functions for water and
hydrogen peroxide based on the GIAO/DFT and PM6 calculations of the same systems (a, curves 1-3, and
b), GIAO/B3LYP/6-31G(d,p) (a, curves 4-6) and GIAO/IEFPCM/B3LYP/6-31G(d,p) (a, curve 7) with (a)
hydrogen peroxide alone or with water molecules; two types of the mixtures were shown with pure solution
(curves 3) and clustered (curves 2) structures; and (b) 317H2O*32H2O2 adsorbed onto silica particles (199
tetrahedrons) and experimental 1H NMR spectrum of H2O2/H2O/A-400 (10 wt.% of 30 wt.% H2O2 solution
in water) recorded at 200 K (curve 4).
For interaction of the aqueous solution of H2O2 with nanosilica (dehydrated before mixing
with the solution) being in nonpolar (CCl4) or weakly polar (CDCl3) solvents, the temperature
behavior of the 1H NMR signals differed from that for the pure solution (Fig. 139) [247]. Several
signals were in the spectra: WAW (δН ≈ 1 ppm), two signals of SAW with different amounts of
dissolved H2O2 (δН ≈ 5-6.5 ppm). Additionally, in the CCl4 medium, separated signal of hydrogen
peroxide was observed at δН ≈ 11 ppm. Freezing temperature of CCl4 was 240 K. The dispersion
medium freezing (Fig. 139b) resulted in broadening of TMS signal at δН = 0 ppm. Near freezing
temperature of CCl4 transition of SAW from state 1 at δН ≈ 5 ppm into state 2 at δН ≈ 6 ppm was
observed. Both signals were observed at 210-240 K. The H2O2 signal decreased with temperature
and it disappeared at T > 240 K [247].
In weakly polar CDCl3 medium (Fig. 139b) the δН value of SAW increased [247]. Splitting
of the SAW peak into 1 and 2 states occurred at T > 250 K. Signal 2 shifts toward strong magnetic
field with increasing temperature because of decreasing clusterization of the solution. This effect
for signal 1 was much smaller. Silica A-400 as a nanostructured material affected the clusterization
of the solution located in narrow nano- and mesopores (voids between silica nanoparticles). This
process was also dependent on cryoconcentration, i.e. relative concentration of H2O2 in certain
structures increased due to partial freezing of water [247].
155
Fig. 141. 1H NMR spectra of aqueous 30% solution of H2O2 adsorbed (10 wt.%) onto A-400 being in (a) CCl4 + 8 %
CD3CN, (b) CDCl3 + 8 % CD3CN, (c) CDCl3 + 15 % CD3CN, (d) CDCl3 + 22 % CD3CN, (e) CDCl3 + 8 %
DMSO media.
The 1H NMR spectra of the H2O/H2O2 mixture adsorbed onto nanosilica being in different
mixtures of nonpolar (CCl4) or weakly (CDCl3) solvent with addition of polar solvents (CD3CN,
(CD3)2SO) (Fig. 141) demonstrated complex temperature behavior [247]. Besides signals of water,
H2O2 and TMS, signals of methyl groups of CH3CN (as admixture in CD3CN) at δН = 2.2 ppm and
(CH3)2SO (admixture in (CD3)2SO) at δН = 2.5 ppm were observed. In frozen CCl4 medium (T <
240 K) signals of SAW and H2O2 were observed separately (Fig. 141a). The SAW signal shifted
toward strong magnetic field with increasing temperature and minimal value δН = 3.5 ppm was
observed at 230 K and then it shifted toward weak magnetic field. At 240 K two signals of H2O2 in
the adsorbed solution were observed at δН = 10 and 8.5 ppm. However, at higher temperatures
only one weak signal was observed and its δН value decreased with temperature. Since the δН
value of this signal was lower than that of pure H2O2 one could assume that the H2O2 molecules
form clusters with water and acetonitrile molecules [247]. After melting of CCl4 (T > 240 K) a
major portion of H2O2 dissolved in the SAW nanodomains (that resulted in increasing δН value of
SAW) or formed the adsorption complexes with surface silanols groups. Their signal could be
156
absent in the spectra due to the proton exchange SiOH↔HOH causing significant shortening of
the relaxation time [247].
In the CDCl3 with addition of CD3CN (Fig. 141b), a weak signal of H2O2 was observed at
δН ≈ 9 ppm. It could be due to the formation of clusters with the participation of H2O2, H2O and
CDCl3 or CD3CN molecules (as well as in the system shown in Fig. 141a). SAW was observed as
one (T < 240 K) or two signals (signals 1 and 2) which shift toward strong magnetic field with
increasing temperature. Signal 1 shifts stronger than signal 2. The maximal δН value for SAW was
7 ppm at 220 K. At lower temperatures, the signal intensity and its δН value decreased due to
partial freezing of water and H2O2. Weak signal of water (signal 3) was also observed at δН = 2.2-
2.4 ppm (to the left from the signal of methyl groups of acetonitrile). This δН value was
intermediate between the values of WAW (1-1.5 ppm) and water bound to acetonitrile molecules
(3 ppm). This signal intensity increased with increasing concentration of acetonitrile in the mixture
with CD3Cl (Fig. 141b-d). It was possible that it could be caused by clusters with CD3Cl-CD3CN-
H2O bound to the A-400 surface in narrow voids. The H2O2 signal intensity at δН ≈ 9 ppm
increased with increasing CD3CN concentration (Fig. 141b-d). The minimal δН value of SAW
decreased to 4 ppm. Signals 1 and 2 were observed separately only at T > 270 K. Signal 3 was the
most intensive at 22% CD3CN in mixture with CD3Cl. It intensity increased with increasing
temperature but intensity of signal 1 decreased. For a CDCl3 mixture with DMSO (Fig. 141e), the
δН value of hydrogen peroxide increased to 11 ppm. Significant changes in the intensity of signals
3 and 1 (similar to that at 22% CD3CN) were already observed at 5% DMSO added to CD3Cl
[247].
Observed patterns of relationship of the 1H NMR signals of aqueous 30% solution of
hydrogen peroxide bound to nanosilica A-400 being in solvents of different polarity and
hydrophobicity could be interpreted in terms of the formation of clusters and nanodomains in
voids between spherical silica nanoparticles, and these clusters contain different amounts of H2O2
[247]. At low temperatures the confined space effects increased due to partial freezing of
components forming crystallites in voids between silica nanoparticles. This led to
cryoconcentration of certain components. In the hydrophobic nonpolar medium, the solution of
H2O2 became inhomogeneous and formed two types of structures with strongly associated water
with different content of hydrogen peroxide. The content of weakly associated water was much
smaller than for similar samples with water without H2O2. For the H2O/H2O2 mixtures in weakly
polar CD3Cl with addition of polar CD3CN or strongly polar DMSO, three types of clusters and
domains were formed at the silica surface: (i) clusters with concentrated H2O2 at H = 9-11 ppm;
(ii) clusters with concentrated water and dissolved certain amounts of co-solvents; and (iii) clusters
with organic solvents containing certain amount of water and H2O2. The activity of solvents
increased with increasing temperature and the cryoconcentration effects decreased that result in the
formation of more homogeneous interfacial solutions including organic solvents. Hydrogen
peroxide formed clusters with water molecules which could slowly exchange by protons with the
clustered SAW and WAW. Significant amounts of WAW were observed only for CD3Cl/DMSO
medium [247].
Effects of salts in free media and confined space
The presence of dissolved salts in the aqueous media can affect the value of the chemical
shift H (Fig. 142) [53,210]. These effects depended on the salt type and content (Figs. 142 and
143). Frequently, a diminution of the water structure order due to dissolved salts results in a
decrease in the values of H. However, for some salts (e.g., KCN), the opposite effects were
observed. These results could be explained by kosmotropic and chaotropic properties of ions
[242].
157
Fig. 142. Temperature dependence of the 1H NMR chemical shift H for pure water or concentrated solutions of
different salts (experimental data) [53,210].
During grinding salt (KCl was selected as a color representative) with strongly hydrated
PMS at h = 9 g/g (Fig. 144a,b) or more weakly hydrated nanosilica at h = 0.1 g/g (Fig. 144c,d)
KCl microcrystallites were observed. A low content of water could not allow the dissolution of
salts. Their sizes were smaller in KCl/PS300 (d 5 m) than KCl/PMS (d 20 m) because of an
abrasive effect of nanosilica which was much harder than the polymer used. Additionally, the
KCl/PS300 has uniform color that suggests a uniform distribution of KCl particles in the mixture
[53,210].
Fig. 143. 1H NMR spectra of (a) small clusters of water (8-43 molecules) and water/DMSO calculated using
GIAO/B3LYP/6-31G(d,p) used for calculation of calibration functions for (b) water clusters with 275 and
2000 molecules (pure and dissolved ions 4Na+ + 4Cl) calculated by the PM6 method, and experimental 1H
NMR spectrum of water in 3.6 % solution (0.44 g/g) of KCl bound to nanosilica A-380. [53,210].
Water bound to NaCl with the presence of n-decane showed a broad signal at H 4.5 ppm
at 280 K (Fig. 145a, dash-dotted lines) [248]. At 270 K, its intensity became much lower and it
158
disappears at 265 K due to freezing of water. This suggests that all water was weakly bound
because it was frozen close to the freezing point of bulk water. This water was strongly associated
since H was close to that of bulk water. Decane gave two signals at H = 0.95 ppm (CH3 groups)
and 1.25 ppm (CH2) (Fig. 145a). At low temperatures, a weak signal at 1.5-2.0 ppm could be
assigned to weakly associated water. Addition of a certain amount of TFAA to chloroform (1 : 9)
led to a downfield shift of signal of TFAA/water toward 7-8 ppm. It was broad at high
temperatures and shifts toward 10 ppm with decreasing temperature close to that for bulk TFAA
solution (~11 ppm) [248].
Fig. 144. Microphotographs (Primo Star optical microscope, Carl Zeiss) of grinded samples (during 10 min) (a, b)
hydrated PMS (h = 9 g/g) with KCl (1 : 3 w/w) and (c, d) KCl/PS300 (1:6) at h = 0.1 g/g in (c) air and (d)
silicone oil (scale bar 20 μm (a), and 10 μm (b, c, d)) [248].
159
Fig. 145. 1H NMR spectra recorded at different
temperatures of water and n-decane bound
to the NaCl powder in (a) CDCl3 (dash-
dotted lines) and CDCl3/TFAA (9:1) (solid
lines) at 0.1 g/g of water and 0.1 g/g of
decane; (b) a mixture of hydrated PMS (h =
9 g/g) with decane (1 : 1) (dash-dotted lines)
and a mixture with hydrated PMS/KCl (1 :
3) at h = 9 g/g and decane (4 : 1) (solid
lines) [248].
Fig. 146. Size distributions of unfrozen water structures
in (a) NaCl powder with 0.1 g/g of water and
0.1 g/g decane in CDCl3 medium; (b)
hydrated PMS (90 wt.% water) with decane (1
: 1); and (c) a mixture with hydrated
PMS/KCl (1 : 3) and decane (4 : 1) [248].
A decrease in the content of decane and addition of KCl to hydrated PMS resulted in
unexpected upfield shift of water signal from 5-6 ppm (dash-dotted lines) toward 4.5-5.0 ppm
(solid lines, Fig. 145b) [248]. This result could be explained by diminution of sizes of bound water
structures after addition of KCl (Fig. 146b,c). In contrast to the systems with NaCl (Fig. 145a),
signals of water and decane were not observed at T < Tf,d = 243.5 K (freezing point of bulk n-
decane) (Fig. 145b). However, in the presence of KCl signal of water was observed at 250-260 K.
Thus, a small fraction of SAW was strongly bound, and signal of WAW disappears at low
temperatures even at T < Tf,d. Certain broadening of decane signal could be due to nonuniformity
of magnetic susceptibility of samples including relatively large solid particles of salts [248].
160
Fig. 147. 1H NMR spectra recorded at different temperatures of water and n-decane bound to (a) PS300/KCl (6 : 1) at
h = 0.1 g/g and Cd = 0.2 g/g in air; (b) PS300/KCl (6 : 1) at h = 0.1 g/g in a mixture with CDCl3/C10H22 (7 :
1); (c) PS300/NaCl (6 : 1) at h = 0.1 g/g and Cd = 0.15 g/g in a mixture CDCl3/CD3CN (4:1) medium; (d)
PS300/NaCl (6 : 1) at h = 0.1 g/g and Cd = 0.15 g/g in a mixture CDCl3/CD3CN/TFAA (4:1:1) medium (e)
PS300/KCl (6 : 1) at h = 0.1 g/g and Cd = 0.2 g/g in CDCl3 medium; and (f) PS300/LiCl (1 : 1) at h = 0.1 g/g
and Cd = 0.15 g/g in CDCl3 medium [248].
Despite a relatively small amount of water (insufficient to dissolve NaCl in the studied
samples) bound to NaCl particles, water formed relatively large structures at R > 3 nm in radius
(Fig. 146a). They corresponded to SAW but WBW, since surface area of relatively large NaCl
crystallites was small [248]. Water bound to PMS, which was characterized by developed surface
area ~200 m2/g [53,69], formed much smaller structures (Fig. 146b,c), which corresponded to both
SBW and WBW, as well as SAW and WAW (Fig. 118b). Addition of KCl led to diminution of
sizes of these structures (Fig. 146c) [248].
In a mixture of PS300 nanoparticles with small KCl particles being in air, the freezing
point depression was observed for both bound water and decane (Fig. 147a) [248]. Unfrozen water
was observed even at 210 K and decane at 220 K. Water formed relatively small structures (Fig.
148a) in narrow voids between silica and salt particles. In CDCl3 medium, lines became narrower
(due to diminution of molecular exchange effects) (Fig. 147b). A portion of water was displaced
from pores. This led to differentiation of signals, since two signals of SAW were observed at H =
5 ppm (SAW1) and 3.5 ppm (SAW2), as well as WAW at H = 1.7-2.0 ppm. Addition of CD3CN
to this mixture (Fig. 147c) caused an upfield shift of water signals and a relative increase in the
amounts of SAW2 and a decrease in the content of SAW1. This suggested decreasing associativity
of bound water molecules. Addition of TFAA (Fig. 147d) resulted in simplification and averaging
of the spectra due to fast proton exchange between molecules of water and TFAA. Intensive signal
of the aqueous solution of TFAA was observed at 10 ppm. Replace of NaCl by KCl (Fig. 147e) or
161
LiCl (Fig. 147f) led to enhancement of SAW1, including both SBW and WBW. Appearance of
these waters corresponded to the formation of relatively small structures (Fig. 148c, d) [248].
Fig. 148. Size distributions of unfrozen water structures in (a) PS300/KCl (6:1) in air; (b) PS300/NaCl (6:1) in CDCl3;
(c) PS300/KCl (6:1) in CDCl3; and (d) PS300/LiCl (1:1) in CDCl3 [248].
The interfacial behavior of bound decane depended on the type of adsorbents (silica PS300
and HPS300, PMS, salts), the presence of water, and the type of dispersion medium [248]. All
these features reflect in the changes in the amounts of unfrozen decane vs. temperature recorded at
increasing temperature from 210 K to 290 K (Fig. 149). At I/I0 = 0, all decane was in solid-like
state (freezing point Tf,d = 243.5 K), and at I/I0 = 1, all decane was liquid. The I/I0 curves for
decane bound to PMS (Fig. 149a, curves 3 and 4) showed a weak influence of hydrated PMS on
the temperature behavior of decane that was similar to bulk decane. Consequently, decane could
not displace water from the PMS surface, since the confined space effects for decane were
practically absent. For all other systems, the confined space effects were observed because a
fraction of unfrozen decane was observed at T < Tf,d. There was another effect of inhibition of
melting a fraction of solid-like decane at T > Tf,d. This effect was maximal for decane bound to
PS300/NaCl located in the CDCl3 medium (Fig. 149b, curve 1). This immobilization of bound
decane could be maximal effective in mesopores of appropriate sizes. PS300 possesses textural
mesopores and macropores. During any treatment (grinding, suspending/drying, gelation,
mechanochemical activation) of PS300, it became more compacted and its bulk density strongly
increased (up to 0.3 g/cm3), and this led to increase in contributions of textural mesopores
[53,248]. A fraction of solid-like structures of decane formed in mesopores could remained during
heating at T > Tf,d because the heating rate was relatively high. This led to delay in melting of
bound structures. Notice that the I/I0 curves for the systems with CDCl3 dispersion medium were
located lower than that for the systems in air (Fig. 149). In both air and CDCl3 medium at T Tf,d,
approximately 70 % of decane was in a liquid state. This showed a strong colligative effect of
solvents with lower freezing point than that of solutes [53,248].
162
Table 22. Sample characteristics and integrated changes in the temperature range of phase transition of n-decane
[248].
Solids h
(g/g)
Cd
(g/g)
Medium
(K)
+
(K)
I/I0
T = Tf
NaCl 0.15 0.1 CDCl3 17.8 5.1 0.73
NaCl 0.15 0.1 5CDCl3+1TFAA 17.8 9.0 0.71
PMS 9.0 1.0 Air 1.7 4.2 0.40
1PMS/3KCl 9.0 0.23 Air 1.5 3.4 0.35
6PS300/1KCl 0.1 0.20 Air 7.9 6.5 0.73
6PS300/1NaCl 0.1 0.15 CDCl3 3.7 20.8 0.11
6PS300/1NaCl 0.1 0.15 4CDCl3+1CD3CN+1TFAA 12.7 8.7 0.77
6PS300/1KCl 0.1 0.20 CDCl3 21.3 7.3 0.68
PS300/LiCl 0.1 0.15 CDCl3 14.8 5.6 0.65
HPS300/KCl 0.1 0.10 CCl4 18.9 2.9 0.83
Note.
f
min
0( ) /
T
T
I T I dT and
max
f
0
0
( )
T
T
I I T
dT
I
.
Fig. 149. Relative intensity in the 1H NMR spectra of decane (I0 corresponds to intensity at 290 K) bound to (a) NaCl
at Cd =0.1 g/g and h = 0.1 g/g in CDCl3 (curve 1), with added TFAA as CDCl3/TFAA = 5 : 1 (2); PMS/KCl
(1 : 3) at Cd =0.225 g/g and h = 9 g/g (3); PMS at Cd =1 g/g and h = 9 g/g (4); and PS300/KCl (6:1) at Cd
=0.2 g/g and h = 0.1 g/g in air (5); (b) PS300/NaCl (6:1) at Cd =0.15 g/g and h = 0.1 g/g in CDCl3 (curve 1);
HPS300/KCl (1:1) at Cd =0.1 g/g and h = 0.1 g/g in CCl4 (2); PS300/NaCl (6:1) at Cd =0.15 g/g and h = 0.1
g/g in CDCl3/CD3CN/TFAA (4:1:1) (3); PS300/LiCl (1:1) at Cd =0.15 g/g and h = 0.1 g/g in CDCl3 (4); and
PS300/KCl (6:1) at Cd =0.2 g/g and h = 0.1 g/g in CDCl3 (5) [248].
Minimal integrated effects on melting of decane were observed for the
poly(methylsiloxane) (PMS) systems (Table 22, ) [248]. This due to filling of pores by water (h
= 9 g/g), relative hydrophilicity of PMS due to the presence of a number of residual silanols [39],
and relatively large sizes of KCl crystallites. Notice that the maximal difference in changes in the
163
Gibbs free energy of solvation of non-dissociated was for LiCl. Calculations with SMD/B3LYP/6-
31G(d,p) clusters M10Cl10 (M = Li, Na) and K8Cl8 with the geometry optimized using the HF/6-
31G(d,p) method gave for solvation in water and n-decane, respectively: 41.3 and 34.9 kJ/mol
(Li10Cl10), 28.2 and 24.9 kJ/mol (Na10Cl10), and 22.5 and 18.2 kJ/mol (K8Cl8). Thus, the
difference for KCl particles in solvation by water and decane added a reason to form of large
decane structures (bound to PMS/KCl) with the properties close to that of bulk decane. A
significant decrease in the amounts of water from h = 9 g/g bound to PMS to 0.1-0.15 g/g bound to
silica with salts or individual NaCl led to enhancement of the interfacial effects on decane (Table
22, ). In many cases, the freezing point depression effects (i.e. ) were stronger than the melting
delay effects (i.e. +). Both effects depended more strongly on the textural characteristics and the
amounts of water (i.e. filling of pores by water) than the types of silica and salts. The latter could
be explained by relatively weak effects of salt crystallites or dissolved ions on the 1H NMR spectra
of water without of adsorbents (Fig. 150) or bound to silica (Fig. 151) because all the spectra were
characterized by a signal at 4-5 ppm typically observed for bound SAW or bulk water [53]. The
spectrum of hydrated LiCl nanoparticles was slightly different due to partial dissolution of LiCl
and formation of charged particles including H3O
+ and OH. Anions Cl were chaotropes but
cations Li+ were kosmotropes, which differently affected the water structure, especially at a
surface of polar nanoparticles. The dissolution effects were much weaker for NaCl nanoparticles
(Fig. 150, curve 2). However, some separated fragments were formed.
Fig. 150. Semiempirical PM7 calculations of the 1H chemical shifts of water with salt crystallites (curves 1-3) or
dissolved ions (curves 4-6) [210].
The interfacial and temperature behavior of mixed water and n-decane depended on such
several factors as the texture and surface nature of adsorbents, the presence and content of
microparticles or nanoparticles of salts, the amounts of components [53,210,248]. If water
completely fills pores of poly(methylsiloxane), which contains both non-polar Si-CH3 and polar
Si-OH groups, that decane co-adsorbed onto already hydrated PMS practically did not sense the
confined space effects and its freezing/melting occurred near the freezing point of bulk decane. If
co-adsorbed water and decane fills only a portion of pores that their interfacial and temperature
behavior depended on the texture of the adsorbents and amounts of salt crystallites. This led to
broadening of the temperature range of melting of decane in both sides from the freezing point of
bulk decane. Typically, the freezing point depression due to confined space effects resulted in
stronger changes in the temperature behavior of adsorbed decane than the effects of melting delay
164
due to both kinetic suppression and immobilization of solid-like structures of decane in mesopores
of adsorbents. The dispersion media (nonpolar CCl4, weakly polar CDCl3, polar CD3CN and
TFAA) influence the interfacial and temperature behavior of co-adsorbed water and n-decane
because decane could be easily dissolved in non-polar or weakly polar solvents but water could
strongly interact with polar solvents. This caused additional differentiation of the interfacial
structures with water (including its four types SAW & WAW, SBW & WBW) and decane
[53,210,248].
Fig. 151. (a) DFT B3LYP/6-31G(d,p) and (b) semiempirical PM7 calculations of the 1H chemical shifts in pure water,
water with NaCl (a) alone or (b) bound to silica surface [210].
Acidic solutions in confined space
Interfacial water at a small content could form large clusters interacting with silica with the
presence of hydrophobic or weakly polar surroundings [53]; however, practically all water
remained strongly bound [53,215] and did not form the phase separated from silica. Similar
regularities in the temperature dependences were observed for adsorbed aqueous solution of HCl
(Fig. 152). Pure liquid hydrochloric acid (36 wt.%) has the signal at H = 8.8-9.4 ppm increasing
with lowering temperature (Fig. 152a). Hydrated complexes of HCl with a dissolved fraction (Fig.
152b, signal 1) and SAW (signal 2) bound by nanosilica differ in their temperature dependence.
The H value and intensity of signal 1 depended weakly on temperature in contrast to signal 2
which increased and shifts toward the strong magnetic field with increasing temperature. Signal 2
was not observed at T < 240 K. Addition of hydrochloric acid (Fig. 152b) led to diminution of the
signal of WAW interacting with silica (H = 1.5-2 ppm) in comparison with the A-300-water
system (Fig. 152a). However the signal of water dissolved in chloroform (H = 1 ppm at very low
amount of this water) did not depend on the presence of hydrochloric acid [215].
The spectra shape of SAW bound in samples placed in the CCl4 medium changed weakly
(Fig. 152c); however, the difference in the chemical shift of two signals of SAW decreased [215].
The difference between signals 1 and 2 decreased with increasing HCl concentration since relative
intensity of signal 1 increased. A decrease in the intensity of signal 2 was observed at T < 240 K
(Fig. 152).
165
Fig. 152. 1H NMR spectra at different temperatures of (a) 36% hydrochloric acid, and hydrochloric acid and water
adsorbed to (b)-(d) initial and (e) MCA A-300 in (b) CDCl3 or (c)-(e) CCl4; (b, c, e) 38 mg/g H2O and 44
mg/g HCl, (d) 140 mg/g H2O and 120 mg/g HCl.
In 36% hydrochloric acid used six water molecules were per HCl molecule. This solution
was frozen at T < 200 K (Fig. 152a) with simultaneous freezing of water and HCl. Freezing of the
HCl solutions was possible at higher temperatures if the HCl concentration decreased.
Consequently, the results shown in Fig. 152b-d could be interpreted as the formation of two types
of HCl/water structures with different concentrations of HCl (CHCl was smaller in structures
corresponding to signal 2 at a lower H value and these structures include mainly non-dissociated
HCl molecules) [215]. The chemical shift of signal 1 was close to that of individual HCl solution,
i.e. the structures corresponding to signal 1 have the properties similar to that of the bulk HCl
solution with significant contribution of dissociated HCl molecules [215].
For water/HCl adsorbed onto MCA A-300 the 1H NMR spectra (Fig. 152e) were similar to
that of water adsorbed on MCA A-300. Freezing of the adsorbates in these systems occurred at
similar temperatures. Water in confined space of smaller voids in MCA A-300 could have a lower
activity as a solvent. Therefore, water clusters adsorbed on MCA A-300 include a lower amount of
dissolved HCl molecules. Quantum chemical calculations of the An anhydrous cluster with 16HCl
but calculated with consideration of the solvation effects (GIAO/IEFPCM/B3LYP/6-31G(d,p))
was characterized by a narrow f(H) distribution function (Fig. 153) because all the HCl molecules
remained non-dissociated.
166
Fig. 153. The distribution function of the H values for clusters with 16H2O, 16HCl and dissociated HCl interacting
with 16H2O (geometry was optimized using IEFPCM/B3LYP/6-31G(d,p) and the NMR spectra were
calculated using GIAO with consideration of the solvation effects), and experimental 1H NMR spectrum of
water/HCl/A-300 (concentrated solution of HCl) recorded at 280 K (curve 4) [210,215].
The 1H NMR spectra of the anhydrous cluster with 16HCl and of the water cluster with
dissociated HCl strongly differ (Fig. 153). The latter had characterized by greater H values up to
17 ppm. These calculations showed that the position of the observed resonance (signal 2, Fig.
152b) could be affected by the presence of non-dissociated HCl. Zundel and Eigen cations located
far from Cl ions [249] provide signals at higher H values (signal 1, Fig. 152b) [215].
Adsorption of a homogeneous HCl/water mixture onto nanosilica A-300 initial or
mechanochemically activated resulted in diminution of HCl dissolution in water (at a low content
of water in the system), especially for MCA silica with narrower voids between nonporous
nanoparticles [215]. This effect appeared as changes in the 1H NMR spectra of the HCl solution
characterized by signals at the H values corresponding to both dissociated and non-dissociated
HCl molecules. Obtained results showed that the use of nanosilica as an oral sorbent could change
the characteristics of gastric juice with the HCl solution. The concentration of dissociated HCl
molecules could be smaller in voids between silica nanoparticles in their aggregates. These
changes could depend on the type of nanosilica, e.g., initial or MCA silica. For the latter the
mentioned effect was stronger because of diminution of the size of voids between silica
nanoparticles [53,215].
The 1H NMR spectra of fresh 30% solution of phosphonic acid in D2O demonstrate the
main signal of the protons in the PO-H groups (and H2O or HDO due to H-D exchange reactions
between PO-H and D2O) with H = 6 ppm (290 K) increased to 9.8 ppm (200 K) with decreasing
temperature [244]. The signal intensity decreased with lowering temperature and its width
increased because of freezing of a portion of water and acid as well as concentrating of the acid
solution with decreasing temperature. Notice that for a similarly concentrated HCl solution, H =
9.0-9.4 ppm at 290-200 K [215] because the dissociation degree of HCl is much higher than that of
POA. Two signals of lower intensity were observed at weaker and stronger fields than the main
signal. The difference between their H values was approximately 2.1 ppm. These H values
depended weakly on temperature; however, the intensity sharply increased at T 260 K but the
width increased at T < 260 K. This occurred due to freezing of a portion of POA. After boiling of
the sample and its storage at 290 K for a week, the intensity of the HP doublet signals decreased
because of the HP-DD2O exchange reaction. The 1H NMR spectra of solid phosphonic acid with a
low content of water (50 mg per gram of solid POA) in CCl4 medium include three signals at H =
6 (signal 1), 5 (weak signal as a shoulder of signal 1) and 3.5 (signal 2) ppm at 290 K (Fig. 154a).
The shape of signal 1 was asymmetrical. Its intensity decreased with decreasing temperature (Fig.
127a), as well as the H value decreased to 5 ppm at 200 K (in contrast to the behavior of the
signal in the concentrated POA solution). The H values for other two signals depended weaker on
167
temperature than signal 1. Addition of water to 70 mg/g results in an increase in the signal
intensity and width (Fig. 154b). However, the temperature behavior of the spectra was similar to
that at CH2O = 50 mg/g [244].
Fig. 154. 1H NMR spectra recorded at different temperatures of weakly hydrated solid phosphonic acid (a, b) alone
and (a, dashed lines) adsorbed onto nanosilica at A-300/POA = 4/1 at h = (a) 50, (b) 70 and (a, dashed lines)
100 mg/g in CCl4 medium.
The displacement of signal 1 toward the weak magnetic field with increasing temperature suggests
that contribution of protons in water with dissolved acid increased with temperature (Fig. 154)
because increased solubility of H3PO3. Contribution of hydrated protons (i.e. dissociated POA)
was small at small content of water (Fig. 154a) because the H value decreased with decreasing
temperature in contrast to that for the concentrated POA solution. The spectra shape showed that
structures with water/dissolved H3PO3 were nonuniform since several signals could be found.
Signal 2 at 3.5 ppm could be attributed to water molecules with one hydrogen bond as a proton-
donor and PO-H without the hydrogen bonding (see quantum chemical calculation results) [244].
Dried suspension with 16 wt.% of silica and 4 wt.% H3PO3 as a powder with
approximately 10 wt.% of water was characterized by the 1H NMR spectra with only one
relatively broad signal at 4-4.5 ppm (Fig. 154) [244]. Similarly hydrated nanosilica A-300
modified by phosphonic acid obtained by hydrolysis of adsorbed PCl3 was characterized by larger
values H = 6.5-10.5 ppm [53]. During drying of the concentrated suspension with silica and POA,
solution supersaturation results in formation of acid crystallites. Additionally, residual water
bound by nanosilica has low activity as a solvent. Therefore, according to the 1H NMR spectra
(Fig. 154), bound water could dissolve a low amount of H3PO3 since the H value depended
weakly on temperature as well as the signal shape [244].
168
Fig. 155. Temperature dependences of concentration of unfrozen water responsible for signal (a) 1 and (b) 2 and (c)
chemical shift of signal 1 for weakly hydrated solid H3PO3 (CH2O = 50 or 70 mg/g) and A-300/H3PO3 = 4/1
(CH2O = 100 mg/g) in CCl4 medium.
Water unfrozen at T < 273 K due to bonding to solid POA or dried A-300/POA powder
(Fig. 155a) could be attributed to strongly (SBW unfrozen at T < 250 K) and weakly (WBW
frozen at T < 250 K) bound water. However, there was not a clear boundary between SBW and
WBW because of changes in the concentration of dissolved POA. A major fraction of water
responsible for signal 1 (Cuw1(T)) (Fig. 155a) corresponded to SBW but a major fraction of water
giving signal 2 corresponded to WBW (Fig. 155b) [244].
The H value of sample with 50 mg H2O per gram of acid depended weakly on temperature
(Fig. 155c) [244]. It slightly increased at T > 230 K. An increase in the water content to 70 mg/g
gives an increase in the H value with a minimum at 240 K. For the dried A-300/H3PO3 powder,
the H value decreased with increasing temperature (Fig. 155c). These results were caused by two
competitive processes such as diminution of associativity of water molecules (H decreased) and
increase of concentration of dissolved acid (H increased) with temperature [244]. During freezing
of the solution, pure ice and acid crystallites (which do not contribute the 1H NMR spectra here)
form separately. However, interfacial water could be frozen at lower temperature than acid that
affected the concentration of acid dissolved in the interfacial water and the chemical shift H
decreased with lowering temperature. However, signal of the concentrated acid solution was
absent in weakly hydrated A-300/H3PO3 powder because water bound to the silica surface did not
169
dissolve the acid which formed crystallites weakly bound to silica surface, according to the FTIR
spectra [244].
Fig. 156. Changes in the Gibbs free energy of bound water in weakly hydrated solid H3PO3 in CCl4 medium at CH2O =
(a, e) 50 or (b, f) 70 mg/g and (c) A-300/POA at CH2O = 100 mg/g and (d, g) water cluster size distributions
(WCSD) with respect to signals (d) 1 and (g) 2.
Water bound to solid POA or silica/POA powders was energetically and structurally
nonuniform (Fig. 156) because of spatial restrictions dependent on void size distribution between
nanoparticles and the difference in the surface effects on water layers nearest to the surface and
located in the next layers [53,244]. As a whole, at low content of adsorbed water (50-100 mg/g) it
tended to form small clusters (Fig. 156d) characterized by relatively large changes in the Gibbs
free energy (Fig. 156a-c) of bound water with the cluster size distributions (WCSD) over 0.6-10
nm in radius (Fig. 156d). Water responsible for signal 2 was nonuniform too but it did not form
nanoclusters at R < 1 nm (Fig. 156g). Its nanodomains weaker interacted with solid POA (Fig.
156e,f) than water responsible for signal 1 [244].
Melting of anhydrous H3PO4 occurred at 290 K; however, the aqueous solution of the acid
could be frozen at much lower temperature because of colligative properties of the aqueous
solutions [53,242,244]. Initial sample OX-50/H3PO4 at CH2O = 5 mg/g has only a weak 1H NMR
signal at H = 4 ppm. Its intensity decreased with lowering temperature because of freezing of
water which was weakly bound. Signal at 0 ppm was due to tetramethylsilane (0.2 wt.%) used as a
reference compound. For dried OX-50/POA powder at CH2O = 45 mg/g, the values of all the
structural characteristics were much lower than those of A-300/POA at CH2O = 50 mg/g because
OX-50 has the SBET value six times smaller than that of A-300. However, the difference in the Suw
values was smaller than that in the SBET values because of contributions of POA nanocrystallites.
Very large values of the structural characteristics of unfrozen water were observed for the
concentrated suspension of OX-50/POA. In other words, water in this suspension was strongly
clustered (vide infra) due to interactions with POA molecules and oligomers [244].
There was a significant difference in the solubility of POA in bulk and bound water
because of decreased activity of the interfacial water [244]. Low-temperature 1Н NMR
spectroscopy investigations of water bound by phosphoric and phosphonic acids solid alone or
adsorbed onto nanosilicas OX-50 or A-300 showed that concentrated solutions or weakly hydrated
solid POA or dried silica/POA powders being in CCl4 medium were characterized by different
170
temperature dependences of the H values because of only partial dissociation of the PO-H bonds.
The H values depended strongly on water amounts, silica type and temperature. NMR
cryoporometry showed that small water clusters (< 1 nm) and nanodomains (up to 20 nm in size)
were present at the interfaces of hydrated solid POA and silica/POA powders. Quantum chemical
calculations of the 1H NMR spectra demonstrate the influence of POA/water cluster structure and
dissociation of the PO-H bonds on the H values [244].
Hydrophilic-hydrophobic characteristics of FMO and confined space effects
Interesting information on the interfacial and temperature behaviors of adsorbates bound to
FMO can be obtained using the DSC method [53]. Comparison of the results of low-temperature
1H NMR spectroscopy and DSC (Table 23) can allow one to attain a deeper insight into these
phenomena. The differences in the values of the heat of immersion in water (Qw) and free surface
energy (S) (Table 23) are due to the differences in the amounts of adsorbates and the approaches
used to calculate these values. Note that Qw deals with all water bound to FMO at T 293 K, but
S deals with only a fraction of unfrozen water at T < 273 K [53].
The interaction of the silica surface with polar and non-polar molecules upon immersion in
liquids is primarily determined by the nature and concentration of surface functional groups. The
initial fumed silica surface contains hydrophilic silanol groups and siloxane bridges, which have
rather hydrophobic properties [37,38], and the concentration of free silanol groups on the silica
surface is 2.25–2.5 μmol/m2. Therefore, fumed silica exhibits pronounced hydrophilic properties.
Substitution of polar silanol groups for trimethylsilyl (TMS) groups leads to a change in the nature
of the silica surface and the interaction with the polar and nonpolar molecules [53]. In the case of
immersion of modified silica surface in polar liquids there is a significant reduction of the heat of
immersion Q, normalized to 1 g of the sample, with increasing modification degree (Table 24).
Immersion in nonpolar decane shows a decrease in Q with increasing degree of surface
modification, but it is insignificant. Since the heat of immersion is proportional to changes in the
Gibbs free energy of surface, the thermal effect is determined, respectively, not only by the nature
of the surface, but also by the specific surface area. The modification of silica causes a decrease in
the specific surface area of the samples (Table 24). Therefore, the heats of immersion reflect these
changes. The average Q values normalized to 1 m2 of the sample surface (Fig. 157) tend to
decrease Qw with increasing degree of surface modification. However, the values of Qd
(normalized to 1 m2 of surface) are independent on modification degree.
Table 23. Adsorption characteristics determined from calorimetrical (Qw, Qd, K) and 1H NMR spectroscopy (S)
measurements.
Sample Qw
(J/m2)
Qd
(J/m2)
K=Qw/Qd
S
(J/m2)
SA8 0.119 0.087 1.37 0.288
SA23 0.132 0.151 0.87 0.122
ST20 0.328 0.082 4.0 0.435
AST03 0.345 0.240
AST1 0.548 0.136 4.03 0.163
AST50 0.459 0.251 1.83 0.077
AST71 0.591 0.161 3.67
AST82 0.493 0.148 3.33 0.479
Al2O3
a 0.411 0.295 1.39 0.231
Al2O3
b 0.400 0.160 2.50
A-200c 0.183 0.149 1.23 0.124
A-50 0.231 0.177
TiO2 0.262 0.083 3.16
Note.
2BET,NS = a133, b86, and c230 m2/g; Qd is the heat of immersion of FMO in n-decane.
Relatively high values of the heat of immersion in polar liquids (triethylamine, TEA,
acetonitrile, 2-propanol and water) and small values of Qd are caused by the differences in
171
intermolecular interactions between the liquids and the polar and nonpolar groups of the modified
silica surface. Interactions of nonpolar liquid with surface functionalities (polar OH or nonpolar
CH3) are determined only by van der Waals forces. Therefore, the heats of immersion of silica in
nonpolar decane are relatively small (due to low energy of the interactions) and they practically do
not depend on the concentration of OH or CH3 groups on the silica surface. The interaction of
polar liquids (TEA, acetonitrile, water, 2-propanol) with the solid surface is more complex. There
are additive effects of various components of polar interactions, whose energy is much higher than
that of the dispersion forces. The heat of immersion for all polar liquids depends on the
modification degree of the silica surface, because polar interactions occur only with the polar
centers at the surface (OH groups). The heat of immersion in TEA, acetonitrile, and 2-propanol
depends almost linearly on the concentration of silanol groups. The decrease in the heat of
immersion of modified silica in water at a small degree of modification is insignificant, while for
samples with high modification degree small values of the heat of immersion (0.23 – 0.24 J/m2)
were observed. Such dependency for water is probably explained by the high energy of
interactions between the molecules of water (surface tension of water is in 3 – 4 times higher than
for the rest of investigated liquids) and low ability to interact with nonpolar CH3 groups.
Table 24. Heats of immersion modified silicas in polar and non-polar liquids.
Sample �a Qwater
(J/g)
QTEA
(J/g)
Q2-propanol
(J/g)
Qacetonitrile
(J/g)
Qdecane
(J/g)
SBET
(m2/g)
Кh
A-300 (initial) 0 41.0 63.8 49.4 48.4 17.9 291 2.41
Si1 0.07 40.4 64.0 44.5 44.3 16.8 282 2.41
Si2 0.14 36.1 60.2 41.6 44.9 15.6 275 2.35
Si3 0.30 31.8 53.0 39.5 40.5 15.6 270 2.20
Si4 0.43 30.1 50.3 22.1 38.1 13.7 274 2.12
Si5 0.53 26.0 44.2 33.3 33.6 13.5 250 1.88
Si6 0.62 22.0 45.5 29.8 29.4 15.2 245 1.61
Si7 0.89 4.9 33.4 23.3 23.6 15.3 250 0.43
Si8 1.0 5.5 29.5 23.0 24.6 12.2 246 0.42
Note. a Degree of OH group substitution with TMS groups determined by integral intensity of the band at 2966 cm-1.
Fig. 157. Dependence of the heat of immersion, normalized to 1 m2, on the degree of substitution of the silanol groups
of the silica surface for TMS groups.
Adsorption of polar polymers onto silica surface occurs due to formation of the hydrogen
bonds between polar groups of polymers and surface silanols. Therefore, if the amount of adsorbed
polymer corresponds to monolayer, the surface properties are mainly determined by functional
groups of the polymers. If the adsorbed polymer amount is less than the monolayer, the
interactions of liquid occur with both silanols and functional groups of the polymers. Therefore,
the heat of immersion in polar and nonpolar liquids is determined by the total surface area of the
172
composite, the accessibility of silanols and functional groups of the polymer for interactions with
solvent molecules.
Figure 158 shows the heat of immersion in water (normalized to 1 g and 1 m2) for oxide-
polymer composites with various concentrations of PVP (12 000 Da), PVA (43 000 Da), and PEG
(35 000 Da). The hydrophilicity of composites may decrease (PEG) or increase (PVP and PVA)
depending on the polymer nature and the degree of surface coating. For PVP/silica composites
(Fig. 158a) the heat of immersion in water decreases in the initial section due to reducing specific
surface area of composites (Fig. 159a). However, at high content of PVP, the heat of immersion in
water increases again almost to the initial level. Since the heat of immersion in decane decreases
proportionally to the specific surface area, the hydrophilicity index (Kh=Qw/Qd) increases (Fig.
159b). Such behavior indicates higher hydrophilicity of surface covered with PVP compared to the
initial silica, that confirmed by the values of the heat of immersion normalized per 1 m2 (Fig.
158b). These data correlate with a high heat of dissolving PVP (J/g) (Table 25), which
demonstrates the high solvation energy. For silicas modified with PVA, an increase in the heat of
immersion in water normalized per 1 m2 is also observed with increasing polymer concentration
(Fig. 158b), but composites with PEG show a significant decrease in the heat of immersion in
water with increasing polymer content (Fig. 158a,b). The heat of dissolution of PEG in water is
negative (Table 25) that can argue that PEG reduces the surface hydrophilicity compared with the
initial silica. The increase in the surface hydrophilicity of PVA/silica composites in comparison
with the initial silica may be caused by the hydrophilic properties of PVA and silanols.
Table 25. The heat of dissolution of polymers in water.
PVP PEG PVA
Heat of dissolution
(J/g)
50 –10 soluble only at 80 ºС (n/m)
Fig. 158. The heat of immersion in water, normalized to
1 g (a) and 1 m2 (b), for oxide-polymer
composites with different concentrations of
PVP, PVA and PEG.
Fig. 159. Specific surface area S (a) and hydrophilicity
index Kh (b) of oxide-polymer composites
with different concentrations of PVP, PVA
and PEG.
173
Changes in the surface composition for complex oxides are accompanied by changes in the
amounts and types of surface sites that can affect the hydrophilicity of these oxides. Therefore, a
question arises about regularities in the relationships between the surface composition and
hydration properties of complex FMO. The heat of immersion of ST, SA and AST (Tables 26-28)
calculated per 1 g of oxide differs over a wide range because of the large differences in the values
of SBET.
To avoid the influence of the SBET value, the heat of immersion was also calculated per m2
of oxide surface area. To estimate the hydrophilicity of oxides, the heat of immersion was
measured for both hydrophilic (water) and hydrophobic (decane) liquids, and the hydrophilicity
index Kh was calculated. According to the data (Tables 26 and 28), at small content of the second
oxide (alumina in SA and titania in ST) the heat of immersion (per m2) increases. However, a
monotonic increase in this value is not observed since the hydrophilicity decreases and then
increases at high content of the second phase. This difference can be caused by non-monotonic
changes in the surface content of the second phase at the surface of the materials [53].
Additionally, the most acidic sites with bridging OH groups actively interacting with water are
formed at the interfaces of SiO2/Al2O3 and SiO2/TiO2. Contribution of the interfacial region
depends on the surface content of the second phase, its crystallite sizes, and the amounts of sites
with isomorphous substitution of Si for Ti or Al or vice versa. The use of the Kh values
independent of the SBET value but dependent on the ratio between polar and nonpolar sites allows
one to observe certain correlations between the surface content of the Al or Ti atoms in the binary
and ternary oxides, their surface content and the Kh values.
Notice that the hydrophilicity of fumed titania is much higher than that of silica and
alumina, and, accordingly, the availability of surface structures with TiO2 in binary and ternary
oxides can promote the hydrophilicity of the surface in comparison with SiO2/Al2O3. For ternary
oxides, there is a synergetic effect since the values of the hydrophilicity index of ternary oxides is
higher than that of individual SiO2, Al2O3, and TiO2.
Table 26. Heat of immersion of FMO with SiO2/Al2O3 at different content of Al2O3.
Sample SBET (m2/g) СAl (аt.%) Liquid Q, J/g Q, J/m2 Kh
A300 230 0 water 42.2 0.18 1.2
decane 34.4 0.15
Al2O3 140 39 water 55. 3 0.39 1.41
decane 39.5 0.28
SA1 207 6.2 water 55.2 0.27 1.44
decane 38.3 0.18
SA3 188 27.7 water 62.8 0.31 1.42
decane 44.2 0.23
SA6 76 water 18.7 0.24
SA8 308 29.5 water 36.7 0.12 1.36
decane 26.9 0.09
SA23 353 11.9 water 46.5 0.13 0.87
decane 53.6 0.15
SA75 85 34.1 water 39.7 0.47 1.38
decane 28.7 0.34
SA96 81 water 44.0 0.54 1.51
decane 29.1 0.36
Therefore, one can assume that the presence of TiO2 (having high hydrophilicity) at the
oxide surface (especially at high surface content of TiO2) and the formation of bridge OH groups
(having high Brønsted acidity at low surface content of Ti and Al) can result in increase of the
AST hydrophilicity. The influence of bridge OH groups is especially noticeable for ternary oxides
with high amounts and different nature of bridge OH groups leading to the increase in the
hydrophilicity of the surface in comparison with SiO2, Al2O3 and TiO2 (Tables 26-28).
174
Table 27. Heat of immersion of FMO with SiO2/TiO2 at different content of TiO2.
Sample
SBET (m2/g) CTi (at. %) Liquid Q (J/g) Q (J/m2) Kh
TiO2 50 31.5 water 11.05 0.22 3.2
ST2 77 4.3 water 3.47 0.07
decane 27.97 0.36 2.5
ST9 188 7.1 water 11.3 0.15
decane 40.56 0.22 2.1
ST14 137 7.8 water 19.7 0.10
decane 35.85 0.26 1.6
ST20 86 6.5 water 22.3 0.16
decane 28.18 0.33 1.7
ST40 109 7.4 water 16.42 0.19
decane 26.54 0.24 1.4
ST65 34 14.4 water 18.9 0.17
decane 13.53 0.40 2.8
ST94 30 30.9 water 4.91 0.14
The enthalpy of phase transitions of bound adsorbates was smaller by the modulus than that
of bulk liquids (freezing) or solids (fusion) with the exception of n-decane (40% longer than n-
hexane) bound to initial A-300/AST1 [90]. The decrease in the |H| occurred due to (i) the small
size of bound structures (clusters, nanodomains) located in voids between nanoparticles, and (ii)
the disorder of bound liquids (being in amorphous state after freezing) which were frozen at
temperatures lower than the freezing point of bulk liquids. Both factors caused smaller endotherms
for melting, or exotherms for freezing per bound molecule of adsorbate [90].
Table 28. Heat of immersion of SiO2/Al2O3/TiO2 of different composition.
FMO
SBET
(m2/g)
СSi
(at. %)
СAl
(at. %)
СTi
(at. %)
Liquid
Q
(J/g)
Q
(J/m2)
Kh
ATS0.3 125 water 43.2 0.35
ATS0.6 97 water 49.3 0.51 3.7
decane 13.2 0.14
ATS1 99 water 54.3 0.55 3.0
decane 17.94 0.18
ATS50 38 32.1 0.7 11.1 water 17.4 0.46 1.8
decane 9.52 0.25
ATS71 74 10.2 2.7 24.9 water 43.7 0.59 3.7
decane 11.93 0.16
ATS82 39 0.3 5.8 31.9 water 19.2 0.49 3.3
decane 5.77 0.15
ATS87 42 3.8 5.4 30.8 water 30.3 0.72 5.0
decane 6.08 0.14
ATS88 39 5.3 4.8 27.7 water 24.8 0.63 2.1
11.81 0.30
175
Fig. 160. Normalized (per mg of adsorbates with subtracted baseline) DSC thermograms of (a-c) nonpolar and (d-f)
polar adsorbates bound to initial A-300 and alumina or MCA A-300: (a) benzene (melting point Tm = 5.53 oC,
boiling point Tb = 80.1 oC), (b) toluene (Tm = −95 oC, Tb = 111 oC), (c) n-decane (Tm from −30.5 oC to −29.2
oC, Tb from 173.8 oC to 174.4 oC), (d) chloroform (Tm = −63.5 oC, Tb = 61.15 oC), (e) water (melting point Tm
= 0 oC, boiling point Tb = 100 oC), and (f) DMSO (Tm = 18.5 oC, Tb = 189 oC). Cooling/heating rate was β =
10 oC/min for benzene, decane, water, and DMSO, and β = 20 oC/min for chloroform and toluene.
176
Fig. 161. Normalized (per mg of adsorbates with subtracted baseline) DSC thermograms of (a-c) nonpolar and (d)
polar adsorbates bound to initial, MCA, and high-pressure cryogel of AST1: (a) benzene, (b) toluene, (c)
decane, and (d) water. Cooling/heating rate was β = 10 oC/min for all samples.
The values of the exotherms (cooling freezing) and endotherms (heating melting) on
the DSC thermograms depended on the amounts of adsorbates and adsorbents [90]. The adsorbate
amounts have a greater influence on the exotherms/endotherms related to strongly (SBA) or
weakly bound adsorbates (WBA). For the majority of the studied samples, the amounts of
adsorbates were larger than that of adsorbents. Therefore, a significant portion of adsorbates was
weakly bound. This gives rise to the sharp exotherms of freezing, and the main endotherms at
temperatures close to the melting point of the bulk frozen compound. For a better view of the role
the types of adsorbates and adsorbents play, DSC thermograms were normalized per mg of
adsorbate with subtraction of the baseline (Figs. 160-162). Relatively large amounts of adsorbates
interacting with FMO led to a low intensity of melting endotherms of SBA located at T < Tm at T
= T Tm > 10 oC, compared with the melting endotherms of weakly bound adsorbates located
close to the melting point at T < 10 oC [90].
For aromatics (benzene, toluene), the freezing point depression was greater than for n-
decane (Figs. 160-162) [90]. Melting delay (i.e. melting of frozen adsorbates at T > Tm) was
observed for both nonpolar and polar adsorbates. However, it was absent for DMSO bound to
initial or MCA A-300 (Fig. 160f). Note that DMSO possesses the largest donor number, DN = 125
kJ/mol, among the studied adsorbates (e.g., for water DN = 75 kJ/mol) [208]. DN was a
quantitative measure of Lewis basicity, i.e. the tendency to form strong hydrogen bonds with
surface hydroxyls. Therefore, DMSO formed stronger hydrogen bonds with surface hydroxyls
than the other adsorbates studied. This has a strong influence on the structure of interfacial DMSO
layers in comparison with bulk liquid (and frozen DMSO) [90].
177
Fig. 162. Normalized (per mg of adsorbates with subtracted baseline) DSC thermograms of (a-c) nonpolar and (d)
polar adsorbates bound to initial and high-pressure cryogel of A-300/AST1: (a) benzene, (b) toluene, (c)
decane, and (d) water. Cooling/heating rate was β = 10 oC/min for all samples.
On the other hand, the acceptor number (AN) was higher for water than for DMSO.
Therefore, in contrast to the other studied adsorbates, water tends to form adsorbed clusters in
which water molecules act as both proton-acceptors and proton-donors [90]. Even at low water
content, water formed cyclic clusters around surface hydroxyls. For the other studied adsorbates,
the values of DN were much less than DMSO or water, e.g., DN = 0 for chloroform. However, the
AN of chloroform was larger (by 11 kJ/mol) than DMSO. These features of the electronic and
molecular structures of adsorbates affected their behavior in pores (voids between nanoparticles)
in different ways. For example, the freezing-melting point depression (or delay) vs. pore sizes and
structures of adsorbates and pore surfaces could be a result of adsorbate electronic and molecular
structure. The observed freezing point depression was greater for DMSO and water bound to MCA
A-300 than that for initial A-300 (Figs. 160-162), because voids became narrower after MCA.
This was due to the pure confined space effect. MCA did not appreciably change the surface
structure of nanosilica, in contrast to AST1 which was more strongly affected by MCA (due to the
destruction of large core-shell nanoparticles) evidenced by the change in sample color. The HPCG
results in much larger changes to the melting thermograms of adsorbates bound to AST1 than
MCA of AST1 (Fig. 161). In the A-300/AST1 blend, nanosilica plays an inhibiting role during
HPCG. Therefore, the difference in the melting endotherms in comparison with the initial powder
was smaller for this blend (Fig. 162) than for both initial and treated AST1 (Fig. 161) [90].
For some samples, the freezing point depression corresponded to relatively small shifts of
the phase transition (melting) temperature at |T| < 10 oC. This melting point shift range
corresponded to WBA, since |T| > 10 oC corresponded to SBA [53,90]. Typically, larger
structures of adsorbates (domains) located in broader voids (pores) mainly corresponded to WBA.
Structures located in narrow voids (i.e. relatively small clusters of adsorbates), or adsorption layers
located close to the pore walls, frequently represent SBA. The position of very narrow exotherms
was related to freezing of WBA (due to excess of adsorbates much larger than the pore volume
178
Vp), and was dependent on the cooling rate (β); the greater the value of β the lower the freezing
point. MCA results in shifts of the PSD peaks toward smaller pore sizes, and high-pressure or
normal-pressure cryogelation led to compacting of the secondary particles. Therefore, these
treatments could enhance the contributions of SBA. These treatment effects the texture of these
powders could result in the appearance of open hysteresis loops for even non-polar compounds
adsorbed onto treated FMO. The open shape of these isotherms could corresponded to activated
desorption.
DFT quantum chemical calculations of the adsorption complexes of water and hexane
showed that water molecules could form two strong hydrogen bonds with neighboring active
surface sites or cyclic clusters [53,90,101,209,210]. This caused the adsorption energy values
(Table 29), calculated taking into account the solvation effects for the free clusters and with
adsorbed molecule, to be larger for water than for adsorbed hexane. Non-electrostatic components
of the solvation energy (Table 29, ECDS corresponding to cavitation (C), dispersion (D), and
changes in the otherwise homogeneous solvent structure (S) induced by the solute) were
destabilizing for the adsorbed water molecule (ECDS > 0), but stabilizing for hexane (ECDS < 0). For
complex oxides, the interaction energy is greater for bound water, but it could be lower for a
bound hexane molecule. The difference in the interaction energy with and without consideration of
solvation effects was much smaller for bound hexane than for bound water (comp. Et,l,cl+m and
Et,g, Table 29). The values of Et,l,cl+m for water interacting with oxide clusters were much
larger than for hexane. These calculation results showed why hexane could be evaporated much
faster than water from any FMO [101]. Clearly, the evaporation rate for liquids from pores differs
from the rate of free liquids [254-256].
Table 29. Interaction energy of water and n-hexane (hex) molecules with silica, silica/alumina (SA), silica/titania (ST) and alumina/silica/titania (AST) clusters (DFT B97X-D
calculations with the cc-pVDZ basis set) [90,101].
Parameter H2O/silica H2O/SA H2O/ST H2O/AST hex/silica hex/SA hex/ST hex/AST
Et,g (Ha) 3902.69920591 3856.27283154 4539.17465345 4416.33144339 4063.30390686 4016.86762615 4699.77637712 4576.92397921
Et,l (Ha) 3902.75652436 3856.33414420 4539.25412599 4416.42150981 4063.32657790 4016.88976195 4699.79978938 4576.94681724
Et,g (kJ/mol) 66.9 (2H-bonds)
56.1 (cor. BSSE)
97.0 (2H-bonds) 75.6 (2H-bonds) 96.3 (2H-bonds) 28.3
25.3 (cor. BSSE)
32.3 29.2 25.7
Et,l,cl+m (kJ/mol) 41.7 66.0 51.6 63.3 24.8 28.0 25.6 22.3
ECDS (kJ/mol) 23.9 25.9 28.7 26.4 35.4 32.6 36.4 34.3
silica SA ST AST silica SA ST AST
Et,g (Ha) 3826.27502792 3779.83720394 4462.74716547 4339.89607929 3826.27502792 3779.83720394 4462.74716547 4339.89607929
Et,l (Ha) 3826.33089333 3779.89924168 4462.82470557 4339.98762667 3826.29289925 3779.85486709 4462.76580134 4339.91407967
Es,cl (kJ/mol) 146.7 162.9 203.6 240.4 46.9 46.4 48.9 47.3
Note. The total energy of water molecule Et,g = 76.3986999307 Ha and Et,l = 76.4097574717 Ha (solvation energy Es = 29.0 kJ/mol); and hexane molecule Et,g =
237.018101921 Ha and Et,l = 237.024235496 Ha (Es = 16.1 kJ/mol). ECDS corresponds to non-electrostatic components of the solvation energy. Subscripts ‘g’ and ‘l’
correspond to calculations in the gas or liquid phase, respectively. Es,cl is the energy of solvation of a pure cluster; Et,l,cl+m = Et,l,cl+m Et,l,cl Et,l,m is the solvation energy of an
adsorbed molecule (it takes into account the solvation effects and interaction of a molecules with a cluster). Et,g is the interaction energy of a molecule with a cluster in the gas
phase. Structure of the clusters is shown in Figs. S53 (water) and S54 (hexane). Et,g = 71.4, 93.5, and 85.1 kJ/mol and Et,l,cl+m = 46.5, 67.2, and 58.2 kJ/mol per a water
molecule for 9H2O, 3H2O, and 6H2O interacting with the silica and SA clusters.
179
180
High-molecular weight adsorbates
Interactions of polymers with FMO
FMO are appropriate materials to be used in composites with polymers [32-45,257-266]
because nanoparticles (especially modified silicas) could be easily distributed in the polymer
matrices. The polymer-nanoparticle, polymer-polymer, and nanoparticle-nanoparticle interactions
in the composites play a very important role in respect to the properties of whole materials.
Composites could be nanoporous or macroporous [267].
The PSD functions and AFM images of the treated FMO and FMO/polymer systems
revealed that the structure of secondary particles depended on oxide type, adsorbed polymers, and
treatment conditions [53,107]. Monolayer coverage could cause a small loss of the specific surface
area 1.0
BET
BET
S
S
if nanoparticles do not stick tightly together in aggregates, because of
interactions with polymer molecules. More than monolayer coverage, or stronger particle-polymer
interactions, could lead to a greater reduction of the specific surface area. For instance, for A-
380/PEG 3.0
BET
BET
S
S
at CPEG = 180 mg/g (monolayer coverage), a significant reduction of pore
volume (from 1.56 to 0.59 cm3/g) was observed. PEG is an adhesive for strongly aggregated
nanoparticles of A-380, since this silica has a Vp value much greater than dry initial A-300 powder
[107]. The increase in aggregate size of A-380/PEG was readily observed in the AFM images
[53,107]. Smaller effects of monolayer adsorbed PEG (with respect to reduction of the SBET and VP
values) were observed for A-300/PEG (CPEG = 125 mg/g, 3.0
BET
BET
S
S
and 2.0
p
p
V
V
) and A-
50/PEG (CPEG = 25 mg/g, 17.0
BET
BET
S
S
and Vp > 0) because of the weakening of primary
particle aggregates with increasing sizes of nanoparticles for a variety of FMO [53].
Despite the scatter in changes in the SBET and Vp values for a large set of the A-
300/polymer or A-300/protein systems [107], due to variations in the type of macromolecules and
treatment techniques, there were several clearly visible trends. The adsorption capacity Vp for
treated silica or silica/polymer powder samples was typically greater than that for the initial dry
powder if the polymer content Cpol < 30 wt.%. However, this value decreased with increasing
polymer concentration and the maximal Vp value (1.8 cm3/g) was observed for suspended-
sonicated-dried A-300 alone. Generally, the adsorption of polymers led to a reduction of the
specific surface area; however, these changes could be relatively small if the polymer content Cpol
< 20 wt.%; i.e., at monolayer or lower coverage. Minimal loss of the SBET value for
polymer/nanooxide powders occurred under the following conditions: (i) relatively weak polymer-
polymer intermolecular interactions (e.g., dispersive without hydrogen bonding as in PEO, PEG,
PVP, PDMS); (ii) maximal decomposition of secondary particles upon their coating by linear
polymer molecules and minimal aggregation of coated particles (i.e. maximal number of segments
should interact with the silica surface); (iii) decreased possibility for the interactions of adsorbed
polymers with many primary particles simultaneously (e.g., short polymer molecules); (iv) good
correlation of the distances between active sites on particle surfaces and polar groups of polymer
molecules; (v) relatively short distances between polar groups of polymer molecules; and (vi) high
rotational mobility of side polar groups (e.g., in PVP) forming hydrogen bonds with the silica
surface.
The PSD functions for A-300/PEG (Fig. 163b) were narrower at pore radii R > 10 nm and
have lower intensity at 1 < R < 10 nm than A-50/PEG (Fig. 163a) [107]. Additionally, the
difference in the PSD functions for the initial A-300 powder (Fig. 163b, curve 1) and treated
samples with A-300/PEG was much larger than that for A-50 and A-50/PEG (Fig. 163a).
181
However, differences in the effects of various ball-milling and other treatments were larger for A-
50/PEG (Fig. 136a) and A-380/PEG (Fig. 164) than A-300/PEG (Fig. 163b).
Fig. 163. IPSDV for dried (a) initial A-50 (curve 1), residual of aqueous suspension with A-50 (2), A-50/PEG (25 mg
per gram of silica) impregnated (3), ball-milled for 5 min (4) then treated in (5) saturated water vapor or (6)
ethanol for 24 h, and ball-milled A-50/PEG with addition 30 wt.% of water for 5 min (7); (b) initial А-300
(curve 1); dried A-300/PEG (125 mg per gram of silica) impregnated (2), ball-milled (3), then treated in
saturated (4) water vapor or (5) ethanol for 24 h, ball-milled A-300/PEG with addition of 30 wt.% of water
(6) (DFT with the model of voids between spherical particles in aggregates).
Significant changes in the amount of adsorbed PVA (≤ 1 monolayer) have a weak influence
on the IPSD shape for the A-300/PVA and SA/PVA systems (Fig. 165). Similar results were
observed for A-300/PEO (Fig. 166). However, the IPSD functions with respect to the specific
surface area (Fig. 166b,d) demonstrate a reduced contribution of both narrow and broad pores.
This could be explained by a relatively uniform distribution of the linear and long (molecular
weight ~600 kDa) PEO molecules on the silica surface and the interaction of a PEO with many
particles simultaneously. This affected the topology of voids in both aggregates of primary
particles, and agglomerates of aggregates. The changes in total porosity as a function of PEO
content was Vp < 0 at CPEO > 15 wt.% (Fig. 167a). On the other hand SBET monotonically
decreased with increased CPEO, while macroporosity increased and mesoporosity decreased (Fig.
167b) [107].
Fig. 164. IPSDV of initial dry A-380 powder (curve 1) and dried residual of aqueous suspension of A-380 (1 wt.%)
(2), A-380/PEG (180 mg/g) (3) impregnated, (4) ball-milled dry mixture A-380/PEG (180 mg/g) and then
treated in saturated (5) water or (6) ethanol vapor for 24 h, (7) ball-milled with addition of 30 wt.% of water
(DFT with the model of voids between spherical particles in aggregates).
182
Fig. 165. IPSDV of (a) initial A-300 (1), dried solid residual of A-300 (2), A-300/PVA at CPVA = 15 (3), 75 (4) and 150
(5) mg/g; (b) initial A-300 (1), dried solid residual of A-300 (2), and A-300/PVP at CPVP = (3) 25, (4) 75, (5)
87.5 and (6) 175 mg/g; (c) initial SA8 (1), dried solid residual of SA8 (2), and SA8/PVA at CPVA = (3) 10, (4)
50 and (5) 100 mg/g; (d) initial ST9 (1), dried solid residual of ST9 (2), and ST9/PVA at CPVA = (3) 9, (4) 45
and (5) 90 mg/g (DFT calculations with the model of cylindrical pores).
Different effects of adsorbed PVA on the Vp and SBET values were observed for A-
300/PVA (both parameters decreased with CPVA) and SA/PVA (Vp grows or remained nearly
constant while SBET slightly decreased or remained nearly constant) [107]. These differences,
resulting from the structure of the PVA coating on the oxide nanoparticles, may be caused by
differences in the acid-base properties of amphoteric silica and silica/alumina, the latter possessing
a set of acid-base sites which could strongly interact with the PVA molecules.
Strong aggregation of ossein molecules with silica nanoparticles led to PSDV functions
[107] similar to A-300/PEO samples at pore radii R < 10 nm. However, for A-300/ossein samples
the contribution of broad mesopores at R > 15 nm and macropores at R > 25 nm was much smaller
than that for other nanooxide/polymer systems. This was due to stronger aggregation of particles
and protein macromolecules indicated by a loss in SBET and Vp with increasing Cossein (SBET = 221,
151 and 112 m2/g and Vp = 0.72, 0.54 and 0.41 at Cossein = 31.6, 158 and 316 (monolayer coverage)
mg/g, respectively) [107]. Similar IPSDV functions were observed for A-300/gelatin (SBET = 145
and 125 m2/g and Vp = 1.0 and 0.88 cm3/g, respectively) and A-300/BSA (SBET = 178, 160 and 141
m2/g and Vp = 1.17, 1.08 and 0.95 cm3/g, respectively). However, the IPSDV peaks for A-300/BSA
and A-300/gelatin were characterized by larger contributions of broad pores compared to A-
300/ossein samples [107].
183
Fig. 166. (a, c) IPSDV and (b, d) IPSD for dried solid residual of A-300/PEO (adsorption from aqueous solution) at
different content of PEO (DFT with the models of (a, b) voids between spherical particles in aggregates and
(c, d) cylindrical pores).
Fig. 167. Relative changes in the specific surface area and the porosity (a) total and (b) meso- and macropores as a
function of adsorbed PEO.
184
The adsorption of several types of polymers onto nanosilica A-300 gives different
interfacial layer structures. PEG and PEO having the same structure of the segments were
characterized by a close (Cpol) shape (Fig. 168a) [107]. Despite weak polymer-polymer
interactions, PDMS gives the lowest (Cpol) values (as a function of Cpol in mg/g) because of (i) a
large segment (mseg) weight (CH3)2SiOSi(CH3)2, (ii) its weak interaction with silanols due to
relatively poor electron-donor properties of oxygen atoms in the siloxane bonds, (iii) steric effects
of the CH3 groups, and (iv) the helical structure of the PDMS chain. PVA molecules could form
strong hydrogen bonds with both silanols and OH groups of neighboring molecules, leading to low
values as a function of Cpol and the lowest values as a function of Cpol/mseg (Fig. 168c). PVP
has a larger segment weight than PEG, PEO, and PVA by a factor of 2.4. Therefore, (Cpol) for
PVP was lower than PEG or PEO, but close to that of PVA (Fig. 168a) [107].
Fig. 168. Perturbation degree () of free surface silanols as a function of the polymer loading (a) Cpol in mg per gram
of silica, (b) = Cpol/Cpol,monolayer and (c) Cpol normalized by dividing by the molecular weight of a segment
(mseg) for PEG (35 kDa), POE (600 kDa), PVP (12.7 kDa), PVA (43 kDa), PDMS (8 kDa) and BSA (67
kDa) onto A-300.
185
PVP molecules could more effectively interact with surface silanols than other linear
polymer molecules because of the rotational mobility of the side groups responsible for the
formation of hydrogen bonds (Fig. 168c). BSA molecules demonstrate relatively low (Cpol)
values because of the high average mseg value, and the globular protein shape. Therefore, a
significant portion of the molecules could not be in contact with the silica surface. However,
protein molecules have several types of polar side groups in addition to the polypeptide chain
which could form strong hydrogen bonds with silanols. That could partially compensate for the
effect of the globular shape of the molecules on the (Cpol) values. Normalization of Cpol by
dividing by the segment mseg molecular weight (Fig. 168c) results in similar (Cpol/mseg) graphs
for all A-300/polymer systems except A-300/PVA, probably because of the effects of strong
hydrogen bonds between PVA molecules. The globular structure of BSA molecules prevents
strong lateral interactions, therefore (Cpol/mseg) for BSA was similar to other systems
characterized by relatively weak polymer-polymer interactions [107].
Fig. 169. TPD MS thermograms at m/z (a, b) 18 (water) and (c, d) 28 (CO) formed on decomposition of PVA (a, c)
and PEG (b, d) adsorbed on silica A-300, silica/alumina SA1, SA3, SA8, silica/titania ST2, ST9, ST20 and
alumina/silica/titania AST50.
The effects of residual interfacial water, which could disturb silanols and form hydrogen
bond bridges between polar groups of polymer molecules and surface silanols, could provide a
small difference in the (Cpol/mseg) graphs for different polymers (especially at Cpol/mseg > 1) (Fig.
168c). Thus a minimal loss of the specific surface area for the nanooxide/polymer powders results
from a monolayer coating of nanoparticles by PEO or PEG, because of the formation of strong
hydrogen bonds with silanols and weak polymer-polymer interactions, result in a dense coverage
with more intensive interaction between polymer molecules and surface silanols. Notice that the
shape of the () graphs (Fig. 168b) slightly differs from that of (Cpol) (Fig. 168a) due to certain
differences in the monolayer capacity for the studied polymers: 125 (PEG), 150 (PVA), 170
(PVP), 190 (PEO), 200 (PDMS) and 400 mg/g (BSA) [107].
The strong interaction of adsorbed polymers with oxide surface caused changes during
their thermal decomposition. Additionally, the catalytic effect of active surface sites on complex
oxides on thermolysis of adsorbed PVA and PEG molecules (Fig. 169) caused displacement of the
186
TPD peaks toward lower temperatures in comparison with thermolysis on silica. There was a
difference in the catalytic effects for PVA and PEO because of their structural differences and the
type of surface interactions. Dehydration of PVA/silica was complete at 650 K, but the
dehydration of fumed silica alone occurred at much higher temperatures (up to 1000 K and higher)
[37-40,53]. The strong hydrogen bonding of OH groups of PVA and silica was the main reason for
this difference [107].
The sample mass used in all of the TPD MS measurements was almost identical [107].
Therefore, differences in the TPD peak intensities of eliminated water and CO on decomposition
of PVA and PEG (Fig. 169), could be caused by not only the differences in the catalytic effects of
SA and ST samples or the reaction mechanisms, but also the differences in the specific surface
area (by order of the value) of nanooxides and the amounts of polymers (monolayer coverage was
used here).
Investigations of the morphological and structural characteristics of a set of powder
nanocomposites with fumed oxides and different linear polar polymers (PEO, PEG, PVA, PVP,
and PDMS) and proteins (ossein, gelatin, and BSA) showed that monolayer or lower coverage of
nanoparticles by the macromolecules results in relatively small reductions of specific surface area
and adsorption capacity in comparison with similarly treated fumed oxides alone [107]. Proteins
result in larger reductions than linear polar polymers. Oxide/polymer powders have a more ordered
pore structure than the initial oxide powders, evidenced by the narrower pore size distribution for
the oxide/polymer systems. This structure and secondary particle morphology depended only
slightly on the content and the type of polymers at coverage less than a monolayer. If polymer-
polymer interactions were weaker than polymer-oxide interactions (PEO, PEG, PVP, and PDMS),
the amounts of hydrogen bonds between silanols and polymers per segment depended weakly on
the type and size of polymers [107].
Interfacial and temperature behaviors of low- and high-molecular weight compounds
Studied silica gels Si-60 and Si-100 and nanosilica A-400 could be assigned to mainly
mesoporous materials since contributions of mesopores to the specific surface area and pore
volume were predominant [268]. Silica gel Si-60 did not practically have nanopores and
macropores with PSD in the range of R = 2-7 nm (Fig. 26). Silica gel Si-100 has mesopores at R =
3.5-15 nm at small contributions of nanopores and macropores [268]. Nanosilica A-400 was
characterized by a broader PSD than silica gels with significant contributions of macropores. This
was due to its texture and particle morphology, since nanoparticles (average diameter ~ 6.7 nm)
form relatively rigid aggregates (< 1 µm in size) and soft agglomerates (> 1 µm) with empty
volume in them Vem 16 cm3/g. The pore volume determined from low-temperature nitrogen
adsorption was much lower than Vem. However, this volume could be filled by polymers. The
adsorption of polymers onto A-400 could be accompanied by re-organization of the texture of
aggregates and agglomerates [53] in contrast to silica gels with rigid microparticles. The absence
of macropores in silica gels could cause slow diffusion of macromolecules in mesopores that could
result in partial filling of pores by PDMS during the NMR measurements. A portion of narrow
pores accessible for nitrogen molecules was inaccessible for much larger PDMS molecules. A
helix shape of PDMS with six Si–O bonds in a cycle restricts the number of segments which could
directly interact with a matrix surface to form the hydrogen bonds SiO–H···O(Si(CH3)2–)2.
However, the PDMS conformation could be changed in the adsorption layer depending on the
PDMS content [53,268].
At a low content of water (hydration h = 0.005 g/g) in dried silicas with adsorbed PDMS,
single signal of water observed at 2.5-5 ppm depended weakly on temperature (Fig. 170) [268].
187
Fig. 170. Temperature dependence of the chemical shift H of water at different hydration (h = 0.005 or 0.1 g/g) bound
to A-400 or Si-60 with adsorbed Oxane 5700.
Assuming that H = 0 ppm for PDMS, the H values could be estimated for adsorbed water
(Fig. 170). Minimal H values were observed for residual water (h = 0.005 g/g) bound to Si-60
with adsorbed 0.3 g/g of Oxane 5700 corresponding to approximately one third of the pore
volume. However, for a higher hydration at h 0.1 g/g and at 0.5 g/g of adsorbed Oxane 5700, the
H values were close to that for bulk water; i.e. adsorbed water became strongly associated water
(SAW). However, at the residual amounts of water (h 0.005 g/g), an increase in the amounts of
Oxane 5700 to 0.5 g/g, corresponding to 62-65 % of Vp for Si-60 and A-400, results in an increase
in the H values (3.2-3.5 ppm). The H values for water adsorbed to A-400/Oxane 5700 were lower
than that for Si-60/Oxane 5700 because of the difference in the texture of these silicas [268]. The
increase in the H values could be explained by an increase in the sizes of the water clusters and
nanodomains since the associativity of bound water increased under action of PDMS. However, it
was lower than that at h = 0.1 g/g. PDMS displaces water from pores or from the silica surface,
and to decrease the surface area of contact of water with hydrophobic silicone oil, water formed
larger structures [269]. However, water at low hydration at h = 0.005 g/g was not typical SAW
characterized by H = 4-5ppm [53] because of too low amounts of adsorbed water. However, 10-
15% filling of pores by water (h = 0.1 g/g) results in the appearance of SAW [268].
The temperature dependences of 1HNMR signals of the CH3 groups of PDMS (Oxane 1000
or 5700) bound to silicas or bulk oils alone (Fig. 171a) were obtained for samples during their
thawing [268]. The intensity was normalized to that of silicone oils before freezing. The PDMS
adsorption onto silicas results in an increase in the melting temperature for a main fraction of
adsorbed polymers (Fig. 171a). However, a certain fraction of adsorbed polymers could be melted
at lower temperatures than the bulk polymer. This heterogenization of adsorbed polymers was due
to their interactions with the silica surface and confined space effects affecting both crystallization
and melting processes. These phenomena were typical for different polymers, which could be in
crystalline, semicrystalline or amorphous states, under confined space effects [53].
Relative changes in the Gibbs free energy of adsorbate-adsorbent systems (ΔG) were
proportional to changes in temperature (ΔT) of the corresponding phase transition (e.g., freezing or
melting of adsorbate crystallites in pores and in the bulk) and the difference in content of liquid
phase (ΔCPDMS,liq) vs. temperature [268] PDMS,liqG k T C , where k was a constant. The data
(Fig. 171a) could be transformed into the temperature dependences of relative changes in the
Gibbs free energy of PDMS. The ΔCPDMS,liq values were determined as the differences in the
amounts of liquid fraction in bulk and adsorbed oils (Fig. 171b,c). These dependences have ranges
with positive and negative values. Adsorption interactions should be accompanied by diminution
of the Gibbs free energy of adsorbates. However, confined space effects could reduce density of
188
polymers adsorbed in relative narrow pores. Additionally, crystalline structures in pores could be
more stable because of stronger interactions with the pore walls than with neighboring
macromolecules of residual water molecules. Therefore, the segmental mobility of a fraction of
PDMS could appear at higher temperatures than that for bulk polymers but for another fraction of
the polymer loosely packed in pores the opposite effect could be observed [268].
Fig. 171. Temperature dependences of (a) relative intensity of 1H NMR signals of silicone oils alone (dashed lines) or
bound to silicas (solid lines), and (b, c) relative changes in the Gibbs free energy of polymers bound to
silicas.
A minimal stabilization of solid Oxane 1000 at T < 260 K was observed for Si-100 because
it has broader pores than Si-60 and did not have narrow pores as A-400 [268]. Notice that the PSD
calculated with the cryoporometry method based on the Gibbs-Thomson equation for freezing
point depression for adsorbates confined in pores for Si-100/Oxane 1000 has a shape broader than
PSD calculated from the nitrogen adsorption isotherm. This could be explained by location of a
fraction of PDMS out of pores because the surface area of silica in contact with PDMS was equal
to 47 m2/g (~13% of SBET of Si-100) but the PDMS amount corresponded to 37% of the pore
volume of Si-100 [268].
A maximal stabilization of solid Oxane 1000 and Oxane 5700 was observed for Si-60
possessing narrow and rigid mesopores. A-400 was composed of soft agglomerates which could be
re-organized during the adsorption of polymers [53,107,268,270]. Temperature of melting was
189
lower for Oxane 5700 than for Oxane 1000 (Fig. 144a). The stability of solid Oxane 5700
decreased (at T < 260 K) with increasing amounts of polymer and water (Fig. 171c).
The effects of both decreasing and increasing of the melting temperature were observed for
silicone oils confined in pores of silica gel particles and voids between nanoparticles in nanosilica
powder [268]. The broadening of the range of melting of PDMS depended on the pore size
distribution and it was minimal for silica gel Si-100 with broader pores. An increase in the
amounts of polymer and adsorbed water results in decreasing effects because of decreasing
interactions of a fraction of the macromolecules with the silica surface and location of PDMS out
of pores. The cryoporometry results showed that the surface area in contact with macromolecules
(Oxane 1000) was about 13% of the SBET value of silica gel Si-100. Consequently, PDMS fills
only a portion of pores despite its volume corresponded to 37% of the pore volume of Si-100
[268].
Nanosilica dispersed in PDMS represents mainly agglomerates of 1-10 m in size (Fig.
172) [271]. Stronger agglomeration of nanosilica in silicone oil in comparison with nanosilica
suspended in aqueous media was due to a hydrophilic character of silica. Confined space effects
could be expected for the PDMS molecules located between silica nanoparticles in agglomerates
that differ for PDMS bound to silica gel.
Fig. 172. Microphotograph (Primo Star optical microscope, Carl Zeiss) of PS400 (30 wt.%) in PDMS (scale bar 10
m).
The temperature dependences of relative intensity of 1H NMR signal differ for individual
PDMS initial and preheated at 420 K for 20 min (Fig. 173b). The I/I0 value corresponded to a
weight fraction of liquid PDMS. The mentioned difference could be explained by additional
disordering of PDMS supramolecular structures [267-271] during heating at 420 K that slows
down the PDMS freezing during subsequent cooling; i.e. this disorder was partially irreversible.
Fig. 173. (a) 1H NMR spectra of individual Oxane 1000 recorded at different temperatures during cooling and
subsequent heating (initial sample); and (b) relative intensity (normalized by dividing of current intensity by
I0 of signal recorded at 290 K) of 1H NMR signals for PDMS initial and preheated at 420 K for 20 min.
190
The 1H NMR signal width was narrow (Fig. 173a) that allows us to observe a small upfield
shift of δH up to 0.1 ppm upon cooling from 280 K to 210 K. Subsequent heating gives a
downfield shift toward practically the same δH values at the same temperatures. There was a
hysteresis loop for signal intensity (Fig. 173b). A significant decrease in intensity of PDMS signal
occurred during cooling from 240 K to 220 K (Fig. 173b), corresponding to the melting point of
Oxane 1000 (231 K). However, complete freezing of PDMS was not observed even at 210 K.
During heating, melting of a part of a frozen PDMS fraction occurred at T > 250 K. However
according to DSC data, melting endotherm was at ~233 K. The hysteresis loop was narrower for
preheated PDMS than for initial one due to disordering of supramolecular PDMS structures [271].
The temperature behavior of PDMS depended on the surroundings. Confined space effects
could be analyzed for PDMS bound to silica gel Si-100 and nanosilica PS400. One could assume
that both mesopores and macropores of silicas could be accessible of PDMS molecules [271].
In contrast to the upfield shift for PDMS alone with decreasing temperature (Fig. 173a),
PDMS adsorbed onto silica gel demonstrates a downfield shift with decreasing temperature (Fig.
174a). The spectra of PDMS bound to nanosilica PS400 have a similar shape (not shown here).
The 1H NMR spectra were recorded for PDMS/silica samples equilibrated at 293 K for a week
(labeled as “initial”) and then these samples were heated at 420 K for 1 h (labeled as “preheated”)
and measured after preheating [271].
Fig. 174. (a) 1H NMR spectra of PDMS (33.3 wt.%) bound to silica gel (preheated sample), and changes in relative
intensity of 1H NMR signal of CH3 groups of PDMS adsorbed onto (b) silica gel (0.5 g PDMS per gram of
Si-100), (c) 8 wt.% PS400 in PDMS, and (d) PS400 : PDMS or PS300 : PDMS = 1 : 1 for initial and
preheated samples.
Comparison of relative intensity curves for PDMS alone (Fig. 173b) and adsorbed onto
silicas (Fig. 174b-d) suggests that PDMS molecules became solid-like at temperatures higher than
Tf (freezing point) for bulk PDMS [271]. Typically, the freezing point depression was observed for
low-molecular weight adsorbates located in pores, and this effect increased with decreasing pore
size [53]. For adsorbed polymers, this picture was more complex, since a polymer fraction
remained liquid at T < Tf but another fraction remained frozen at T > Tm (melting point). It was
191
possible that freezing occurred for an interfacial layer of adsorbed polymer at T > Tm. The next
layers of the polymer located in pores could be in unfrozen (mobile) state at T < Tf. Notice that the
hysteresis loops of 1H NMR signal intensity were observed for smaller adsorbates such as
hydrophobic n-decane bound to silica gel and nanosilica [158,248,271].
To compare the effects of different sizes of nanoparticles and time of equilibration of
PDMS/nanosilica, PDMS/PS300 was also studied (Fig. 174d) [271]. Nanosilica PS300 was
composed of larger nanoparticles (average d = 9.2 nm) than PS400 (d = 6.7 nm). Therefore, the
surface area of contact of the PDMS molecules with silica nanoparticles in PDMS/PS300 should
be smaller than that in PDMS/PS400. This led to a curve which was closer to that for initial PDMS
alone (Fig. 173b) than the curve for PDMS/PS400. Consequently, the relaxation of PDMS was
deeper due to a smaller effect of PS300 than that of PS400. The heating rate was the same (5
K/min) for PDMS/PS300, but the equilibration time at a certain temperature was longer: 10 min
(instead of 7 min used for PDMS/PS400) at 210 K < T < 250 K for each temperature point, then 30
min (an intensity step observed) and 10 min for subsequent points. Changes in the heating regime
led to certain changes in the intensity curve due to deeper relaxation of the PDMS structure [271].
Theoretical calculations of the 1H NMR spectra of PDMS fragments (7, 18 or 36 units) free
and bound in a pore or out of a pore of silica nanoparticle (Fig. 175) showed the effects of both a
silica surface (in pore or out of pore) and co-adsorbed water [271].
Fig. 175. Experimental 1H NMR spectra of PDMS alone (curve 1) and adsorbed onto Si-100 (0.5 g/g PDMS) (curve 2)
at 280 K, and theoretically calculated spectra of PDMS alone (7 units) by GIAO/B3LYP/6-31G(d,p) (curve
3), and PM7 (with full optimization of geometry) with the correlation function H = 5.96214 + 33.7732qH
(qH is the atom H charge) for PDMS alone (36 units) (curve 4), two PDMS molecules (each of 18 units)
bound into pore and out of dry pore (curve 5), out of (curve 6) and in (curve 7) hydrated (517H2O) pore of
silica particle (1624 atoms), one PDMS molecule (18 units) in dry pore (curve 8) or hydrated pore (curve 9).
For PDMS located in the pore, the downfield shift was greater than that for PDMS placed
out of the pore (Fig. 175, curves 6 and 7) [271]. For PDMS located in a dry pore, there was the
upfield shift (curves 5 and 8) in comparison with PDMS located in a hydrated pore (curves 6, 7,
and 9) due to enhanced van-der-Waals and electrostatic interactions with surroundings (both silica
and water). PDMS in a liquid state (curve 1) was also characterized by the downfield shift in
comparison with individual molecule without neighbors (curves 3 and 4) due to the effects of the
surroundings. Notice that the interaction energy between two PDMS fragments (each with three
units) calculated using the DFT wB97XD/631G(d,p) method (with HF/6-31G(d,p) geometry)
was Et = 9.4 kJ/mol. However, according to calculation of these two fragments using the
Morokuma method by HF/631G(d,p), contribution of electrostatic energy was 46.9% due to
interactions of the H atoms of the CH3 group of a molecule with the O atom in the Si-O-Si bridge
of another molecule. This interaction caused a certain polarization of the bonds (H charge
192
increased in the complex from 0.132 a.u. to 0.145 a.u. and qO = 0.634 and 0.636 a.u.,
respectively) and the shielding of protons decreased. This explains the downfield shift of PDMS in
the liquid state in comparison with individual PDMS molecules [271].
Investigations performed by spectroscopic (NMR), relaxation (DRC, TSDC), scattering
(SAXS, SANS), thermal (DSC) methods showed importance of the confined space effects on the
interfacial behavior of polymers interacting with a solid surface in pores or at open surface of
disperse but nonporous particles [53]. To estimate the effects of open silica surface and confined
space in silica pores on the phase transition of adsorbed PDMS, the temperature dependences of
the differences between the І/І0 values for adsorbed ((І/І0)a) and individual ((І/І0)i) polymers as
Δ(І/І0) = (І/І0)a − (І/І0)i for initial and preheated samples were calculated (Fig. 176) [271].
The negative values of Δ(І/І0); i.e. diminution of the relative intensity І/І0 for adsorbed
PDMS, corresponded to a relative increase in freezing temperature of a fraction of adsorbed
PDMS [271]. For all samples, there was a significant difference in the temperature behavior of 1H
NMR signals of initial and preheated samples. The hysteresis loops of the Δ(І/І0)(T) curves for
preheated samples were much narrower than that for the initial samples. This difference could be
interpreted as diminution of the surface and confined space effects on the temperature behavior of
PDMS bound to silica due to enhanced disordering of the supramolecular structure of preheated
PDMS and removal of adsorbed water. Notice that the DSC data showed the difference in the
temperature behavior of PDMS adsorbed onto silicas in the first and second scans analogous to
initial and preheated samples studied by the NMR method [53,271].
Fig. 176. Changes in relative intensity of 1H NMR signals of bulk PDMS and PDMS bound to silicas for initial and
preheated samples: (a) 33.3 wt.% PDMS bound to silica gel, (b) 8 wt.% PS400 in PDMS, (c) PS400 :
PDMS = 1 : 1; and (d) temperature dependence of chemical shift of proton resonance for PDMS alone and
bound (33.3 wt.%) to Si-100.
The positive values of Δ(І/І0) (i.e. a decrease in the freezing temperature) was more
characteristic for initial samples than for preheated ones (Fig. 176) [271]. For preheated samples,
this effect was well observed for PDMS adsorbed onto silica gel (Fig. 176a) that was due to the
confined space effects. During freezing of PDMS alone the upfield shift was observed with
decreasing temperature in contrast to PDMS bound to silica gel (Fig. 176d). Thus, the PDMS-
PDMS interactions result in an increase in proton shielding with lowering temperature. However,
PDMS-silica surface interactions caused certain diminution in proton shielding with decreasing
193
temperature. This could be explained by stronger polarity of adsorption sites of the silica surface
interacting with the macromolecules than that of PDMS molecules [271].
The low-temperature 1H NMR spectroscopy used here gives information of the behavior of
PDMS with respect to phase transition between liquid and solid-like fractions [271]. Additional
information on the temperature behavior of PDMS interacting with different silicas could be
obtained using the DSC method. According to the DSC data, the crystallization temperature (Tc)
for the sample with 33.3 wt.% PDMS/silica gel was about 177 K during the first scan (Fig. 177a)
that corresponded to published results [272]. During the second scan, Tc was about 188 K but this
feature was much weaker than that during the first scan. This effect could be explained by
disordering supramolecular structure [272,273] of PDMS macromolecules bound to silica that
remained during the second scan [267-271].
Fig. 177. Results of the first (1) second (2) DSC scans for PDMS (33.3 wt.%) bound to (a) silica gel Si-100 and (b)
nanosilica PS400 at cooling-heating rate of 10 K/min; and (c) integral heat as a function of temperature (T)
(with subtraction of a nonlinear baseline and normalization to unit) for endo-effects (heating, melting) for
PDMS (1) confined in pores of Si-100 or (2) bound to non-porous PS400 nanoparticles.
194
For PDMS adsorbed in pores of silica gel with the average diameter dp 14 nm similar to
the size of typical lamellar structures of PDMS (~10 nm), the melting endotherm was well seen at
227 K for both scans (Fig. 177a) in contrast to that for the PDMS/PS400 system with a very small
endotherm observed only upon the first scan (Fig. 177b) [271]. This effect could be due to much
stronger disordering effect of small nanoparticles of PS400 onto PDMS supramolecular structure
than that of silica gel with relatively broad mesopores, where the PDMS structure could be close to
that in the bulk PDMS. These DSC results were in agreement with the NMR data showing the
stronger effects of nanosilica than silica gel on the temperature behavior of PDMS. During the first
cooling-heating scan, there was a broad endotherm at 273-393 K (Fig. 177) which could be due to
several processes [53,274-279]. Water evaporation could contribute this endotherm but this effect
could be small, since the water content in both systems was small < 0.5 wt.% [271].
For PDMS (33.3 wt.%) bound to nanosilica surface, the melting endotherm peak at 233 K
was very small during the first scan (Fig. 177b), and it was practically absent during the second
heating run [271]. Additionally, the crystallization exotherm (around 183 K) was absent. However,
a small exotherm was observed at 233 K during the first cooling run. This could be due to freezing
a small portion of PDMS and/or water bound in narrow nanopores of PS400 which were poorly
accessible for PDMS macromolecules. Notice that the broad endotherm at 300-400 K observed
during the first scan was larger for PDMS/PS400 (H = 47 J/g) than that for PDMS/Si-100 (H =
39 J/g). This difference could be caused by stronger interactions of the PDMS molecules with the
open PS400 surface than that with the silica surface in pores of Si-100 [271]. Despite the PDMS
amount was much lower (~0.35 cm3 per gram of silica) than the pore volume of Si-100 (Vp = 1.23
cm3/g), a film of PDMS was observed at a surface of Si-100 particles. The difference in the
morphology of silicas caused a significant difference in the curves of integrated heat flow [134,
213, 271]
min
( ) | ( ) |
T
T
T F T dT during heating (endo-effects due to melting) of PDMS bound to Si-
100 and PS400 (Fig. 177c), since the PDMS/Si-100 curve has a significant step at 220-230 K in
contrast to a smooth curve for PDMS/PS400 (Fig. 177c). This step for PDMS/Si-100 in not
vertical because of certain nonuniformity of PDMS with respect to both molecular weight and
supramolecular structure of PDMS located in pores of different sizes and out of pores.
The temperature behavior of PDMS (Oxane 1000) differs for polymers alone or bound to
silica gel Si-100 and nanosilica PS400 [271]. The low-temperature 1H NMR spectroscopy showed
that two fractions of PDMS (liquid and solid-like, frozen) co-exist over a broad temperature range
spreading toward both sides from the Oxane 1000 melting point. The transverse relaxation times
of protons in methyl groups strongly differ for liquid and solid-like PDMS that allows quantitative
estimation of contributions of both fractions vs. temperature.
The cooling-heating of PDMS and PDMS/silica samples gives the hysteresis loop for
intensity of 1H NMR signals. The loop width depended on the surrounding (bulk PDMS or PDMS
adsorbed on a silica surface) and confined space effects (silica gel or nanosilica) [271]. It also
strongly differs for initial and samples preheated at 420 K. The observed differences could be
explained by different order or disorder of supramolecular structures of PDMS alone, adsorbed
and preheated.
DSC measurements of polymers (33.3 wt.%) bound to silica gel and nanosilica showed that
bound PDMS was characterized by different temperature behavior on the first and second cooling-
heating scans and this behavior differs for silica gel and nanosilica. For PDMS/silica gel,
supramolecular structures could be more ordered than that for PDMS/nanosilica. Both melting
endotherm and crystallization exotherm were observed for PDMS/silica gel. However, both
thermal features were much weaker for PDMS/nanosilica and observed only during the first
temperature scan. This difference could be explained by stronger interactions of polymers with
silica nanoparticles than with a surface in pores of silica gel because penetrating of
macromolecules into relatively narrow and long pores (since silica gel beads have 0.2-0.5 mm in
diameter) could be more difficult than mixing nanoparticles with PDMS molecules [271].
195
Oxide composite with SiO2/ZrO2 (labeled as AZr) at 2.4 wt.% of ZrO2 was prepared by a
wet impregnation method using zirconium acetylacetonate dissolved (0.4 wt.%) in CCl4, and
nanosilica A-300 as a substrate, calcined up to 823 K [149]. The amounts of unfrozen water (Cuw)
at T > 240 K for AZr (h = 50 mg/g) in air was equal to the h value (Fig. 178a), i.e. all water was
unfrozen at 240 < T < 273 K and G < 1.2 kJ/mol; i.e. it was SBW [53,149]. The behavior of
bound water changes in the chloroform medium or/and after the adsorption of PDMS (Fig. 178a).
The amount of SBW (Table 30, Сuw
s) decreased and WBW appears (Сuw
w). This led to strong
decrease in the γS value (Table 30); i.e. water more weakly interacts with the oxide surface. These
changes could be caused by two factors. First, water and chloroform were immiscible liquids.
Therefore, bound water tends to reduce the boundary area with weakly polar (hydrophobic)
dispersion medium. This could occur by the formation of larger water domains (Fig. 179) [149].
Fig. 178. Relationships between the amounts of unfrozen water (Cuw) and changes in the Gibbs free energy of bound
water and temperature for AZr and AZr/PDMS in (a) air or CDCl3 and (b, c) in mixture with CDCl3/CD3CN
at 9:1 or 9:2 for three types of bound water: SAW, WAW and ASW.
196
Table 30. Characteristics of water bound to AZr or AZr/PDMS samples in air or chloroform medium [149].
Sample Medium Water
type
Сuw
s
(mg/g)
Сuw
w
(mg/g)
ΔGs
(kJ/mol)
ΔGw
(kJ/mol)
γS
(J/g)
<T>
(K)
Suw
(m2/g)
Snano
(m2/g)
Smeso
(m2/g)
Vnano
(cm3/g)
Vmeso
(cm3/g)
Air SAW 50 - 3.6 - 6.9 205.6 63.3 55.8 7.5 0.024 0.026
CDCl3 SAW 14 36 4.0 0.8 2.1 250.4 12.4 5.0 7.4 0.002 0.040
AZr
9CDCl3/1CD3CN SAW 10 40 3.2 0.8 1.9 254.0 8.2 1.3 6.9 0.001 0.045
Air SAW 10 40 3.9 0.6 1.7 254.4 10.5 3.5 7.1 0.001 0.038
SAW1 10 10 2.7 0.6 0.7 255.0 2.6 0.8 1.8 0 0.017
SAW2 0 26 - 0.4 0.4 266.9 2.9 0 2.8 0 0.021
CDCl3
CDCl3
CDCl3 WAW 1 3 2.6 0.5 0.1 257.4 0.7 0.03 0.6 0 0.003
SAW1 5 17 2.3 0.5 0.5 260.6 1.7 0 1.6 0 0.020 9CDCl3/1CD3CN
9CDCl3/1CD3CN SAW2 0 21 - 0.5 0.3 267.0 2.8 0 2.7 0 0.020
SAW1 4 13 1.6 0.5 0.3 263.3 1.1 0 1.0 0 0.016
AZr/PDMS
9CDCl3/2CD3CN
9CDCl3/2CD3CN SAW2 0 17 - 0.5 0.2 266.7 2.0 0 2.0 0 0.015
197
Fig. 179. Size distributions of pores filled by unfrozen water for AZr and AZr/PDMS in (a, c) air (*) or CDCl3 (**)
and (b, d) in mixture with CDCl3/CD3CN at 9:1 (*) or 9:2 (**) shown as (a, b) incremental and (c, d)
differential distributions. The PSD based on the N2 adsorption isotherm [34] was multiplied by 0.06 since
the water content corresponded to approximately 6% of the pore volume calculated from the nitrogen
adsorption.
However, an increase in the water domain sizes led to diminution of the interaction of water with
the oxide surface (Table 30, S). Second, PDMS was hydrophobic polymer, which displaces a
water portion from the AZr surface. Therefore, water bound to AZr/PDMS in air or chloroform
tends to reduce the contact area with both hydrophobic PDMS and CDCl3. This results in
diminution of water interaction with the oxide surface and the γS value drops down and <T>
increased (Table 30). The effects of PDMS and organic solvents on bound water structure were
well seen in the PSD (Figs. 179 and 180). The water domain size increased in comparison with
water bound to AZr in air. The Suw, Snano and Vnano values strongly decreased (Table 30). Water
structures in the mesopore range give the main contributions (Table 30, Smeso, Vmeso). Therefore,
<T> (Table 30) increased by 45-60 K because the main fraction of water became WBW but it
remained SAW. Contribution of WAW was small in all samples (Fig. 181, Table 30). The
cryoporometry and thermoporometry give similar PSD for AZr in air (Fig. 180), despite
significant difference in the used temperature ranges and related phenomena (melting and
evaporation of water). The difference in the N2, NMR and TG PSD shapes (Figs. 179 and 180),
especially at R < 2 nm, could be caused by a strong influence of water on the organization of
aggregates of nanoparticles. In contrast to nitrogen molecules, water molecules could penetrate
between adjacent oxide nanoparticles bonded in aggregates by the hydrogen (electrostatic) and
van-der-Waals bonds, therefore, individual silica nanoparticles could be observed in concentrated
aqueous suspensions [53,149,238].
Water bound in weakly hydrated powders (h = 50 mg/g) with silica/zirconia (2.4 wt.%)
alone or with immobilized PDMS was mainly strongly associated (δН = 4-5 ppm) but weakly
bound to the oxide surface in the chloroform medium. Both PDMS and CDCl3 stimulate an
increase in size of water structures to reduce the contact area (Suw) between water and hydrophobic
polymer or organic solvent. Adsorbed water did not completely cover the oxide surface in air since
Suw = 63 m2/g at the specific surface area of 290 m2/g. The Suw value decreased by 5-10 times for
samples with or without PDMS in the chloroform medium. Thus, the used techniques allow
198
quantitative description of structures of water bound to complex surfaces in different dispersion
media [149].
Fig. 180. Comparison of the size distributions of pores filled by unfrozen water at h = 50 mg/g (NMR) or evaporated
water at h 60 mg/g (TG) for AZr in air.
Fig. 181. Changes in (a, c, e) the amounts of unfrozen water dCuw/dT as a function of changes in the Gibbs free
energy, and (b, d, f) size of unfrozen water structures in starch at h = (a, b) 0.3 g/g, (c, d) 1 (curve 1), 1.5 (2,
3) and 2.3 g/g (4), (e, f) 0.1 (1), 0.2 (2-5) g/g; on addition of CDCl3 at CCDCl3 = (a, b) 2 g/g (2, 3), (e, f) 2 g/g
(3-5), and CDMSO = (a, b) 1 g/g (3), (e, f) 0.5 (4) and 1 (5) g/g, CQc = 4.8 wt.% (c, d, curves 2-4) and (e, f, 1-
5), CA-300 = 4.8 wt.% (c, d, curve 4) and (e, f curves 1-5); PSD for A-300 calculated on the basis of the
nitrogen adsorption was shown (d, f).
199
Comparison of changes in the dCuw/dT and fV(R) functions for unfrozen water (Fig. 181)
showed that SBW (G < (0.50.8) kJ/mol) corresponded to mainly water structures at R < 2-3
nm, and WBW ((G > (0.50.8) kJ/mol) was linked to structures at R > 2-3 nm [280]. The
boundary G and R values between SBW and WBW depended on system composition. For more
complex systems with starch/Qc/A-300/water this boundary was nonabrupt (Fig. 181e) in contrast
to starch/water or starch/water/CDCl3 (Fig. 181a) because There were the differences in the
freezing point depression (i.e. the corresponding size (R) of pores where water was unfrozen at this
temperature) and the G values for water interacting with organics and silica nanoparticles
[53,280]. Addition of DMSO to starch/water/CDCl3 caused significant changes in the dCuw/dT
(Fig. 181a, curve 3) and fV(R) (Fig. 181b, curve 3) functions. Addition of Qc and A-300 (Fig. 181)
to hydrated starch led to more complicated shapes of both dependences because of different
tendencies in the DMSO, CDCl3, Qc, starch, and A-300 effects on structure of interfacial water.
For instance, the bound water displacement to both broader and narrower pores (in starch and/or
silica) occurred due to its interactions with organic solvents [53,280].
The interfacial behavior, energetic characteristics and structure of both SBW and WBW
depended strongly on the amounts of water and the type and amounts of co-adsorbates (organic
solvents) in hydrated starch with addition of quercetin, nanosilica, and weakly polar (CDCl3) and
polar (DMSO) organic solvents [53,280]. The addition of these solvents to hydrated
quercetin/starch/nanosilica caused the appearance of three additional water signals at δH = 1.3, 3
and 4 ppm differently dependent on temperature. Thus, the composite systems with hydrated
starch/quercetin/nanosilica were characterized by structural and energetic nonuniformities which
could be varied due to changes in the amounts of water and treatment temperature.
Fig. 182. Theoretically calculated 1H NMR spectra for (a) water cluster with 12H2O (1) and distribution function f(H)
for 12Н2О*6((СН3)2SO) (only for water) (2) (method GIAO/B3LYP/6-31G(d,p)//6-31G(d,p)), (b)
distribution functions f(H) for water nanodomain (1), hydrated two starch fragments (2), quercetin molecule
in water nanodrop (3), and hydrated starch/Qc (4) calculated using correlation equation (geometry was
optimized using PM6/MOZYME), curve 5 – experimental 1H NMR spectrum of water (0.3 g/g) bound to
starch recorded at 280 K.
In the presence of organics, the hydrogen bond network in water was disturbed that led to
upfield shift in the 1H NMR spectra [53,280]. This effect depended on the type of organics. The
calculated 1H NMR spectra were broadened for water interacting with small organic molecules,
e.g., DMSO (Fig. 182a). The presence of the quercetin molecule or two starch (amylose)
200
fragments in water nanodomain results in the upfield shift in the spectra (Fig. 182b). This suggests
weakening of the hydrogen bonding or changes in orientation of the OH bonds of water
molecules with respect to lone-electron pairs of the O atoms of neighboring molecules.
For hyaluronic acid (HA), which could form strong hydrogen bonds, the upfield shift for
bound water was slightly smaller (Fig. 183) than for water disturbed by quercetin or starch (Fig.
182). Appearance of a small number of chloroform molecules (ratio CHCl3/H2O = 17 : 624)
around hydrated HA (Fig. 183, curve 3) weakly affected the spectrum. Much stronger upfield shift
of the 1H NMR spectra was observed for water bound to DNA or DNA intercalated by
doxorubicin (anticancer medicine) (Fig. 184) due to chemical structure of DNA. A water shell of
DNA was more strongly clustered and tends to be weakly associated water (WAW) than water
around polysaccharides (Figs. 182 and 183) [53,210].
Fig. 183. PM6 calculations (using the calibration function) of 1H NMR spectra of pure water (curve 1) and water
bound to a fragment (773 atoms) of hyaluronic acid HA (curve 2) and with addition of CHCl3
(17CHCl3/624H2O) (curve 3), experimental 1H NMR spectrum of water/HA/A-300 (4 wt.% H2O and 10
wt.% HA) recorded at 280 K (curve 4). [210]
Fig. 184. PM6 calculations (using the calibration function) of 1H NMR spectra of water bound to a DNA fragment
without (curve 1) and with (2) one or (3) two molecules of doxorubicin (Dox) intercalating DNA,
nanodomain of pure water with 2000H2O (curve 4), experimental 1H NMR spectrum of water/DNA/A-300
(50 wt.% H2O and 6 wt.% DNA) recorded at 280 K (curve 5) [210].
To characterize interfacial interactions, the peculiarities of molecular dynamics and
conformational state in PVP nanoshells compared with those in bulky PVP, far-IR spectra at 20-
500 cm1 were registered [277]. These spectra obtained for neat PVP and PVP shells with different
thicknesses including monomolecular layers in the 20PVP/80SiO2 and 20PVP/80Al2O3 samples
(Fig. 185). Seven partly overlapping absorption bands with the maxima at 42, 83, 100, 169, 247,
355 and 446 cm1 were present in the far-IR spectrum of neat PVP. The band assignment was
given in Table 31. The relative values of absorption coefficients were calculated for the bands
characterizing interfacial interactions and shell thicknesses (Table 32).
201
(a) (b)
Fig. 185. Far-IR spectra of neat PVP and PVP shells in (a) PVP/SiO2 and (b) PVP/Al2O3 at different content CPVP and
shell thickness. Monomolecular PVP coverage was at 20 wt.% of PVP. The spectra of PVP shells were
shifted along the ordinate axis relative to the spectra of neat PVP. Shaded rectangle between ~260 and 350
cm1 characterizes the region of Lewis/Brønsted interactions at polymer/oxide interfaces (tentatively, see
text).
Table 31. Absorption bands in the far-IR spectra of PVP in bulky state and nanoshells [277].
Absorption band, cm1 Assignment
40-46 Hydrogen bonds, torsional vibrations a
78-104 Small-angle pyrrolidone ring vibrations (librations,
Poley-type absorption)
167-172 Hydrogen bonds, stretching vibrations a
245-248 Torsional skeletal vibrations
~260 to 350 PVP/oxide Lewis/Brønsted interactions b
350-357 Deformation pyrrolidone ring vibrations
442-447 Ditto
Note. a The hydrogen bonds in neat PVP were assigned to PVP-residual tightly-bound water complexes. b Tentative
assignment (see text).
The PVP far-IR spectrum substantially changes when PVP was transformed into nano-
dispersive adsorbed state, i.e., for monomolecular and thicker PVP nanoshells [277]. These
changes relating to interactions, dynamics and conformational state of chains at interfaces were
most pronounced for PVP monolayers (CPVP = 20 wt.%). Additionally, somewhat different spectra
of PVP nanolayers adsorbed on silica and alumina surfaces were also observed.
Generally, four peculiarities of the nanoshell spectra have to be noted [277]. First, the
doublet 83/100 cm1 librational band transforms into the single band at ~80 cm1 for the
monomolecular shell. Splitting of the librational bands into two bands with the difference in the
maxima values of 10-20 cm1 has earlier been observed for polymers and explained by the
complicated conformation of molecules with the existence of two isomeric states. The
disappearance of the doublet may indicate, therefore, “simplifying” conformational state of the
majority of the PVP segments (flattened conformations, extending chains) in the adsorbed
monolayer strongly interacting with the oxide substrate. Secondly, the absorption band at 245-247
cm1 of torsional skeletal vibrations disappears at CPVP = 20 wt.% indicating some suppression of
202
the chain dynamics in the PVP monolayer. At last, the changes in the spectra in the regions of
~170 and 260-350 cm1 directly characterize strong physical interactions at the PVP-oxide
interfaces (Table 32) [277].
Table 32. Far-IR characterization of PVP-oxide interfacial interactions [277].
Hydrogen
bonding
PVP/oxide Lewis/Brønsted interactions
(~ 260 to 350 cm1 region)
Sample Nanoshell
thickness,
Nm k170/k80 k318/k80 k350/k80 k350/k444
Neat PVP 1.05 0.85 1.15 0.72
40PVP/60SiO2 2-5 1.14 1.20 1.23 0.82
20PVP/80SiO2 1-2 1.25 1.53 1.35 0.96
90PVP/10Al2O3 20-90 1.05 0.89 1.23 0.76
80PVP/20Al2O3 15-70 1.04 0.94 1.24 0.78
40PVP/60Al2O3 5-20 1.02 1.26 1.40 0.88
20PVP/80Al2O3 2-9 1.02 1.56 1.72 1.02
Note. The coefficients k80 and k444 and their ratio remained practically unchangeable and were used as internal
standards.
Thus, if interfacial interactions between PVP and silica core were controlled basically by
the hydrogen bonds and to some extent by the Lewis/Brønsted (LB) interactions, the latter mainly
control the PVP-alumina interactions. Strong PVP/oxide interactions and peculiar conformational
state of adsorbed polymer at the interfaces crucially affected the glass transition dynamics,
especially for the monomolecular layer shells. This was confirmed by the DSC experiments
presented below [277].
a b
Fig. 186. DSC curves obtained in the temperature region of glass transition for neat PVP and PVP shells in PVP/oxide
nanoparticles with (a) silica and (b) alumina. The second scans were taken after heating samples to 217oC at
V = 20oC/min and subsequent cooling down to 20oC at V = 320oC/min. The curve obtained for
20PVP/80SiO2 after annealing at 150oC for 1 h was also shown.
203
The DSC thermograms (Fig. 186) characterized PVP glass transition in waterless neat
polymer and PVP/oxide nanoparticles [277]. Glass transition temperature Tg at the half-height of
heat capacity step, and the transition breadth Tg = Tg"
Tg', where Tg'
and Tg"
are the temperatures
of the glass transition onset and completion, respectively, were shown. One could see a peculiar
glass transition manifestation in core-shell samples depending on their composition (shell
thickness) and oxide type. The main point here was that no simple dependence of Tg on the shell
thickness was observed. The strongly broadened in both directions, at the limit from 63 to 230oC,
and two-stage glass transition may be seen for the PVP shells unlike the transition range Tg =
18oC for neat PVP. The glass transition width Tg =100-150oC was typical of the monomolecular
PVP shells at Cpol = 20 wt.% in the core-shell particles. For instance, Figure 186 showed that
instead of Tg = 143oC for neat PVP one could see Tgs at 104 and 158oC for the 90PVP/10SiO2
composition (hpol=7-25 nm); at 109 and 157oC for the 40PVP/60SiO2 composition (hpol=2-5 nm)
but at 127oC for the 20PVP/80SiO2 composition (hpol = 1-2 nm) or at 129 and 216oC for the latter
shell after its annealing at 150oC for 1 h [277].
The shell transition characteristics differed for fillers with silica or alumina [277]. The
basic difference was higher Tg'
and Tg" temperatures for 20PVP/80Al2O3 and 40PVP/60Al2O3. At
the close PVP nanolayer thickness in the 40PVP/60SiO2 and 20PVP/80Al2O3 (Table 32) stronger
LB interactions in the latter result in higher Tg'
and Tg" values: 87 and 213oC instead of 76 and
169oC for the silica-containing particles, respectively (Fig. 186) [277].
It was found that “anomalous” dynamic modes contribute to the broad, complicated glass
transition in the monomolecular PVP nanoshells [277]. This was corroborated in the experiments
performed at different heating rates V, from 5 to 40oC/min, with the determination of apparent
activation energy Q of segmental dynamics in transition as a function of temperature.
Combined far-IR/DSC study allowed us also to reveal the cardinal changes in the dynamic
and structural characteristics of PEG nanolayers adsorbed on alumina and silica surfaces compared
to those for neat bulky PEG [277]. DSC data were obtained for melting of neat PEG and PEG
nanoshells (Figs. 187 and 188 and Table 33). One could see that the characteristics of this
transition depended both on the type of oxide cores and the nanoshell thickness (i.e. PEG content
in core-shell particles). Despite a high PEG ability to crystallization, the total amorphization of the
PEG monolayer adsorbed on amorphous silica surface was observed. Some crystalline phase
appeared for PEG-silica pairs only in the thicker PEG nanolayers (not shown therein). Vice versa,
the PEG monolayer adsorbed on the alumina surface remained highly-crystalline (Table 33,
20PEG/80Al2O3). It was characterized with the anomalous, doublet melting endotherm consisting
of the lower-temperature (I) and “normal” (II) peaks (Fig. 187). Crystallinity of the PEG nanoshell
somewhat decreased with its thinning from = 93-95% for neat PEG and PEG shells at the
thickness hpol = 20-90 nm to = 70% for the PEG shells at hpol = 2-9 nm (20PEG/80Al2O3) [277].
Table 33. Melting transition characteristics, crystallinity and crystallite sizes of neat PEG and PEG nanolayers in core-
shell nanoparticles [277].
True melting points,
oC
Sample Melting point
Tm at V = 20
oC/min,
oC 'tr
mT
tr
mT "tr
mT
Melting
enthalpy
Hm,
J/g PEG
Crystallinity
%
Crystallite
thickness
lc,
nm
PEG 75a 59 65 65 185±10 93±5 9-22
46b 58b 80±3b 70±5 3-8b 20PEG/80Al2O3
66 59 59 58±3 9
40PEG/60Al2O3 68 155±7 79±5
80PEG/20Al2O3 71a 162±8 82±5
90PEG/10Al2O3 71a 187±10 95±5
20PEG/80SiO2 0 0
Note. a These experimental melting points were higher than the equilibrium melting temperature of a perfect PEG
crystal T0
m = 69oC for methodical reasons (due to thermal delay). b These true melting temperatures, melting
enthalpy and crystallite thickness values relate to anomalous low-temperature part I of PEG monolayer melting peak
(see Fig. 187).
204
Fig. 187. DSC curves obtained at a heating rate V = 0.62oC/min in the temperature region of PEG melting for
20PEG/80SiO2 nanoparticles (1, monolayer), neat PEG (2), 90PEG/10Al2O3 (3), 40PEG/60Al2O3 (4), and
20PVP/80Al2O3 nanoparticles (5, “monolayer”).
True melting points were determined for neat PEG and PEG monolayers adsorbed on
alumina cores, in particular for the correct calculation of lamellar crystallite thickness [277]. The
characteristic melting temperatures vs. V1/2 were obtained at the heating rate ranging from 0.62 to
20oC/min; temperature Tm
anom relates to the maximum of anomalous peak I (Fig. 188).
Additionally, the equilibrium melting point Tm
o for PEG crystal was also indicated. One could see
that only two plots for Tm and Tm" in neat PEG were linear over the whole range of heating rates,
whereas the deviations from linearity toward higher temperatures were observed at low heating
rates in other cases, especially considerable for PEG nanoshells. As known, this effect was
associated with metastability and reorganization of lamellae during the experiment; however,
extrapolation of the linear sections of these plots to V = 0 gives true melting points typical of the
initial lamellae [277].
Fig. 188. Melting points vs. (heating rate V)1/2 plots obtained for neat PEG and PEG nanolayer in 20PVP/80Al2O3.
Temperatures of melting peak maxima (Tm, Tm
anom), onset (Tm’) and completion (Tm”), and the equilibrium
melting point Tm
o for PEG crystallites were indicated. True melting temperatures of these samples were
obtained by extrapolating linear parts of these curves to V = 0.
205
Onset of true melting range (T tr
m)' was at 59oC for neat PEG but at 46oC for the thinnest
PEG shells (Fig. 188 and Table 33) [277]. The calculation of the crystalline lamellae thicknesses lc
showed that the anomalously thin (lc = 3-8 nm) lamellae prevail in the monomolecular PEG shells
adsorbed on the alumina surface. Of significance, this value coincides in fact with the shell
thickness estimated (Table 32). It means that alumina surface was mainly covered with single
lamellae forming these nanolayers.
PEG depicts a high affinity to the fumed silica surface in aqueous medium, and the
hydrogen bonding SiOHO(RR) between ether oxygen atoms and surface silanols was
responsible for strong adsorption of PEG. The IR spectra showed that PEG molecules disturbed
accessible surface silanols at the monolayer coverage (20PEG/80SiO2). This resulted in
disappearance of the majority of free silanols; however, only 20% of oxygen atoms of PEG
molecules took part in the hydrogen bonding with the surface silanols, obviously, in the train
segments [277].
Combined AFM/FTIR/fluorescence spectroscopy study of PEO films with 15-500 nm in
thickness adsorbed on oxidized silicon substrate showed that lamellar growth rates, morphology,
and the degree of crystallinity were functions of film thickness [277]. The degree of crystallinity
decreased steadily when the film thickness became smaller than 200 nm; the lamellar growth rates
decreased in the thinnest (15 nm) films to below 1% of their bulk value. In the experiments,
correlation between polymer crystallization in thin films and the surface free energy of several
substrates was observed. It was concluded, in particular, that the crystallization was strongly
influenced by the interfacial hydrogen bonding (as in our PEG-silica pair); and it was suppressed
on quartz surface due to constraining effect that influenced orientational and conformational
changes of polymer chains at the interfaces. Especially strong suppression of the crystallization
was observed for PEO chains (the same of PEG) after their grafting onto fumed silica particles that
inhibited the chain mobility necessary for this process. As it follows from the NMR spectra, some
train PEO segments with suppressed mobility were trapped near the silica surface in this case,
alongside the presence of segments in the molecular loops and tails with anomalously high
mobility. Therefore, totally amorphous PEG monolayer of ~1-2 nm in thickness adsorbed on
amorphous fumed silica in our experiments was in fact the limiting case of structural disordering
of this polymer caused by local hydrogen bonding of chains to the substrate hydroxyls [277].
Monte Carlo simulations [281] of polymer crystallization in thin films of thickness
comparable to the polymer−coil sizes (radius of gyration Rg of a few nanometers) suppose
nucleation on “sticky walls” (due to increased affinity to a substrate) with formation of
preferentially flat-on lamellar crystals. When the PEO film of hundreds of nanometers in thickness
anneals at Au substrate it changes from isotropic film to one with some degree of preferred chain
orientation. The effect was more pronounced in the thinner films, suggesting that this phenomenon
was caused by the presence of polymer-substrate interface. The PEO chains within the crystallites
were oriented preferentially with their helical axes perpendicular to the substrate surface [282].
Really, the PEO chains crystallized in the layer of ~10 nm in thickness on the Au surface into the
lamellae where chains were oriented perpendicular to the film plane [283]. Analogously, the PEO
films of 300-400 nm in thickness were formed by spin coating onto aluminized glass substrate,
i.e., in fact on an alumina surface; on in-situ recrystallization, macromolecules oriented themselves
to be normal to the substrate. Such recrystallization occurred in our work during preparing the
samples for spectroscopic measurements.
Layer-multiplying co-extrusion of alternating polystyrene (or poly(ethylene-co-acrylic
acid)) and PEO layers was used to study crystallization of PEO in a confined, two-dimensional
space [284,285]. When the PEO layer thickness was equal to 20-25 nm these layers crystallized as
single, high-aspect-ratio flat-on lamellae that resembled single crystals. The fold surfaces of
lamellae were in the plane of the layer, and the chains were perpendicular to the film plane.
The alumina surface contains much less sites for the hydrogen bonding, such as free
hydroxyls, than silica surface [53,277]. Nevertheless, strong polymer-alumina interfacial
interactions were expected. For instance, a highly constrained relaxation, associated with these
206
interactions, was observed for a polymer in pores of ~20 nm in size [286]. Therefore, it may be
supposed that the formation of the highly-crystalline PEG shells of 2-9 nm in thickness
corresponding approximately to the main lamellar thickness on crystalline alumina substrate may
be caused by both LB interfacial interactions and surface-induced heterogeneous nucleation
process similar to heteroepitaxial crystallization.
Thus, as the thickness of the adsorbed PEO layer decreased from micro- to nanoscale the
morphology changes from complicated 3D spherulite structure to lamellar stacks and then
basically to single lamellae with a preferred flat-on orientation in films when the PEO helices
were mainly perpendicular to the substrate surface. It was obvious that such structural changes
have to result in the relevant changes in molecular dynamics, most sharply pronounced in the
monomolecular crystalline PEG films when the substrate−polymer interface and free polymer
surface dominate the dynamics of ultrathin films [287].
Such an assumption was confirmed by the far-IR spectra. PEO (or PEG) molecule has a
7/2 helix structure with the helix length (identity period) of 1.93 nm. In the frequency range
between 80 and 300 cm1 three spectral bands with the maxima around 110, 165-170, and 210-
220 cm1 have earlier been observed in the PEO spectra [277]. These bands were assigned to the
symmetric and antisymmetric torsional and deformation vibrational modes in chains, viz.,
torsional vibrations around C-O bonds (COC, ~110 cm1), or around C-O and C-C bonds (COC +
CC, 165-170 cm1), and the complex motions in chains (210-220 cm1) including deformation
(bending COC and CC modes), and torsional vibrations (CC) (Table 34). The analysis allowed us
also to assign the 110 cm1 band to the limited torsional vibrations (libration) on the scale of a
monomer unit –CH2–CH2–O– in this polymer. The polarized spectra of oriented POE showed that
(COC) torsion dipole transition moment (band at 107 cm1) was parallel to the PEO helix axis,
whereas two other far-IR bands were associated with the vibrations perpendicular to
macromolecular axis [277].
Table 34. Absorption bands in the far-IR spectra of PEG in bulky state and nanoshells [277].
Nanoshell
thickness,
nm
Absorption bands (cm1) Sample
Torsional vibrations
COC, or libration of a
monomer unit53
Torsional vibrations
CC + COC
Torsional-deformation
vibrations (СС, COC, CC)
PEG 103, 110 169 218
80PEG/20Al2O3 15-70 109 166 214
40PEG/60Al2O3 5-20 109 165 213
20PEG/80Al2O3 2-9 100 160, 182 210, 223
20PEG/80SiO2 1-2 110 170 217
Note. Enhancing absorption in the region of ~ 260-400 cm1 with shell thinning in PEG/Al2O3 was observed that may
be tentatively assigned to increasing contribution of LB interactions at the interfaces (see text).
The far-IR spectra were obtained in the region of 20-400 cm-1 for neat PEG and PEG
shells with ~1-2 nm (monomolecular PEG layer) or 3-9 nm (thin PEG lamellae thickness, see
Table 34) for 20PEG/80SiO2 and 20PEG/80Al2O3, respectively, as well as for thicker PEG
nanoshells (Fig. 189) [277]. These spectra allowed us to reveal peculiarities of the chain dynamics
in the PEG nanoshells from 1-2 to 70 nm in thickness (Table 34) compared with that in neat PEG.
Additionally, to lesser extent, information on interfacial interactions may be obtained from the
spectra. The spectra manifest three features including the presence of three mentioned absorption
bands; however, a considerable difference in the spectral contours of different samples, and
arising of an additional absorption over the range between 300 and 400 cm1 (for PEG nanolayers
in PEG/Al2O3) were observed. Modifying of the spectrum of neat PEG in the nanolayers
depended on the layer thickness and the type of oxide. The effects became more pronounced with
thinning of films suggesting that This was caused by the presence of polymer-oxide interface and
207
free surface (more strictly, contacting with nonpolar LDPE matrix in our samples). The most
cardinal changes in the PEG spectra were observed for the thinnest PEG layer adsorbed on
alumina (spectrum 5) [277].
Far-IR spectrum 1 of neat PEG (Fig. 189) consists of broad intense bands at maxima ~110,
169 and 218 cm1 [277]. A slight tendency to splitting into two modes (103 and 110 cm1) was
observed for the first of these peaks. According to XRD analysis of PEO crystals, this effect was
identified as a result of distortion of the normal helical structure of PEO chains. Nevertheless,
splitting of this band could not be resolved in the spectra of PEG nanolayers. Spectrum 2 of the
amorphous PEG monomolecular layer adsorbed on the silica surface differs from spectrum 1
mainly only by decreased band intensities and widths. This may be associated, apparently, with
the impact of interfacial interactions resulting in some constraining effect. Unfortunately, the
stretching vibrations of hydrogen bonds may manifest at the same frequencies as dynamic mode
(COC + CC, 165-170 cm1) of PEG. Therefore, it was impossible to separate the compensating
effects of decreasing the relevant band intensity and possible some increasing of the hydrogen
bond band [277].
It was assumed [277] that substantial LB interactions between the ether oxygen atoms of
PEG molecules and the silica surface were absent. Such interactions manifest themselves over the
~260-400 cm1 range (see above). Therefore, the coincidence of spectra 1 and 2 in this region
(Fig. 189) confirms such assumption. Far-IR spectra 3 and 4 of thicker PEG nanolayers in
PEG/Al2O3 composites (Fig. 189) were also characterized by some decreasing intensities of three
absorption bands. However, the absorption at ~260-400 cm1 increased in this case, obviously, as
a consequence of arising substantial LB interactions between PEG and alumina surface. Finally,
the most pronounced changes in the far-IR spectrum were observed for 2-9 nm layer of PEG in
the 20PEG/80Al2O3 composite (spectrum 5). This highly-crystalline layer consisted mainly from
thin PEG flat-on lamellae. One could see three peculiarities in spectrum 5 compared to spectrum 1
of neat PEG (Fig. 189, Table 34). First, it was large decreasing and narrowing of 109 cm1 band.
Then, it was some tendency to splitting of 169 cm1 band into the bands at 160 and 182 cm1 and
the distinct splitting of 218 cm1 band into the bands at 210 and 223 cm1. At last, the maximal
effect of absorption at ~ 260-400 cm1, with the maxima at 320 and 360 cm1, was observed for
this PEG nanolayer adsorbed on alumina [277].
Fig. 189. Far-IR spectra of neat PEG (1) and PEG shells in the 20PEG/80SiO2 (2), 80PEG/20Al2O3 (3),
40PEG/60Al2O3 (4), and 20PEG/80Al2O3 (5) nanoparticles with different shell thicknesses.
Thus, very specific chain dynamics was registered in the far-IR spectrum of the thinnest
(2-9 nm) layer of PEG adsorbed on alumina. By analogy with the data obtained for adsorbed PVP
monolayers, one could assume that arising of the absorption at ~ 260-400 cm1 characterizes,
really, LB interactions between PEG and alumina [277]. The prevailing feature of PEG nanolayer
208
structure in this case was the presence of thin (3-8 nm) lamellae located in the plane of the layer
when the lamellar folds from one side were strongly interact with alumina substrate (interfacial
forces) but remained undisturbed by surroundings (nonpolar PE matrix or air in our samples) from
the opposite side of lamellae. The PEG chains between lamellar fold include only about 2-4
identity periods (or Kuhn segments). Therefore, different dynamics in chain segments, close to
substrate or to the opposite lamellar surface, may be reflected in the observed splitting of two
absorption bands in far-IR spectrum of this nanolayer. At last, a sharp decreasing of bandwidth at
109 cm1 for nanoshell in the 20PEG/80Al2O3 composite may be due to narrowing a dispersion of
potential barriers to the smallest-scale torsional COC motion within lamellae with chains oriented
perpendicular to alumina surface.
The combined study of four series of oxide core - polymer shell nanoparticles (prepared in
the pseudo-liquid state or using polymer adsorption onto oxides from the aqueous solution then
dried) with different shell thicknesses (monomolecular and thicker nanolayers) and
comprehensively characterized was performed. Fumed oxide cores, amorphous silica (5-18 nm in
size) or crystalline alumina (12-47 nm in size) and water-soluble PVP and PEG shells were used
[277]. IR spectroscopy in the far (20-500 cm1) and middle (400-4000 cm1) spectral regions was
used to estimate the interfacial interactions and peculiarities of molecular dynamics in the polymer
nanoshells. The peculiarities of the glass transition in the PVP nanolayers and melting
transition/crystallites in the PEG nanolayers were characterized using DSC method. The obtained
calorimetric data were in an accord with the spectral results.
The contributions of strong hydrogen bonding and, tentatively, LB forces to PVP/silica
interfacial interactions but only of the enhanced LB ones to PVP/alumina interfaces were revealed,
and the distinct impact of interfaces (interfacial interactions) on dynamics was observed. Some
suppression of dynamics in chains and “simplifying” their conformational state in nanolayers were
shown. All the effects observed were the functions of the nanoshell thickness and the type of
substrate. The most pronounced peculiarities were found for the monomolecular PVP and PEG
layers when the substrate-polymer interface and “free” polymer surface dominate the film
dynamics [277].
Unlike bulky polymer, 1-2 nm thick PVP monomolecular layers were characterized with
the dramatically high dynamic heterogeneity within the glass transition extending from 70 to
230oC, while varying the activation energy from 80 to 560 kJ/mol. Three “abnormal” dynamic
modes only constituted this transition including two decelerated, constrained motion modes and
ultra-fast one caused by a collapse of intermolecular dynamic cooperativity [53,277].
Despite a high PEG ability to crystallization, total amorphization of PEG monolayer
adsorbed on silica surface was observed that was due apparently to local hydrogen bonding of
chains with substrate silanols. On the contrary, highly-crystalline PEG monolayer was formed on
alumina surface. The single anomalously thin “flat-on” lamellae, of 3-8 nm thick, prevailed in this
layer: this value coincided with the shell thickness estimated. As presumed, the latter effect was
caused by the quasi-heteroepitaxial crystallization process of PEG on crystalline alumina surface.
Such structure of the thinnest PEG layer manifested itself in the characteristic changes of the far-
IR spectra [277].
Features of the interactions between water and solutes with silicas (Fig. 190) or
macromolecules (such as starch gel, fibrinogen, etc.) and A-300 (Fig. 191) reflected in the
dielectric spectra and TSDC thermograms [53,214,245,288]. Enhanced clusterization of water
bound to silicas led to diminution of the value of permittivity (Fig. 190a). An increase in the
degree of hydration or formation of larger structures in pores of silica gel resulted in increased
values of ’. Treatment in organic solvents for 24 h, drying at 473 K and wetting gave smaller
clusters of water than that for initial A-300 or treated in CCl4 (Fig. 190a). However, if the amount
of water bound to treated A-300 was greater that this effect could be masked (Fig. 190b) because
of stronger increase in the size of clusters.
209
Fig. 190. Relaxational characteristics of water free, pure or with solutes or bound to silicas differently affected: (a)
permittivity (at frequency 3 MHz) vs. hydration degree (h) of silicas initial or suspended in organic solvents
at 298 K for 24 h, dried at 473 K (24 h) and wetted using a certain amount of water; (b) dielectric loss (at 3
MHz) vs. temperature for A-300 initial (h = 20 wt.%) and pretreated in CCl4 (h = 40 wt.%). TSDC
thermograms (sample thickness ~1 mm, polarization at 300 V) of (c) 1 wt.% suspension of A-300 in solution
at 0.15 M NaCl, solutions of NaCl and NaClO4 and pure water (polarized at Fp = 0.1 MV/m); and (d) wetted
A-300 initial (h = 66.6 wt.%) and after MCA (6 h) at h = 50 and 40 wt.%, silica gel Si-60 at h = 83.3 wt.%
and pure water [53,214].
Even relatively small content of A-300 in the solution of NaCl caused enhanced
polarization of various structures because intensity of both dipolar relaxations at 90-190 K and dc
relaxation at T > 190 K were greater in the A-300 suspension than that in pure NaCl solution (Fig.
190c). Note that the salinity led to enhanced relaxations in comparison with pure water. Changes
in the content of water or MCA treatment of A-300 or changes in the texture of silicas (compare
A-300 and Si-60) led to significant changes in the dipolar and dc relaxations of frozen and thawed
structures of wetted silicas (Fig. 190d).
A low-temperature (LT) band (T < 170 K) shits toward higher temperatures and its
intensity for water/starch/A-300 was smaller than that for water alone or A-300 suspension (Fig.
191a). A high-temperature (HT) band (T > 170 K) shifted toward lower temperatures and its
intensity increased. The first effect was due to stronger hydrogen bonding of interfacial water
molecules in composite than in starch. The second effect was due to enhancement of plasticizing
of starch by water in the presence of A-300 because of transition of adsorbate (starch
macromolecules) into nanostate (i.e. diminution of supramolecular interactions between
macromolecules), and an increase in the amounts of more mobile relaxing dipoles in glycoside
structures characterized by greater polarity due to interaction with neighboring water molecules or
surface silanols of silica nanoparticles [53,288]. Note that relaxation of dipolar structures could
210
start at much lower temperatures than the relaxations appeared in the low-temperature 1H NMR
spectra and caused by local molecular mobility (Fig. 191b) [23,245].
(a)
(b)
Fig. 191. (a) TSDC thermograms of hydrogel of gelatinized starch (14.2 wt.%) (1), aqueous suspension of A-300 (5
wt.%) (2), gelatinized starch/nanosilica (7.7: 1) at h = 16.9 g/g (3), and water (polarized at Fp = 0.35 MV/m)
(4); arrows showed directions of the displacement of LT and HT bands on addition of silica and starch to
water. (b) TSDC thermograms (curves 1 and 2) and Cuw (NMR, curves 3, 4 and 5) as functions of
temperature for A-300 alone (1 and 5), solution of fibrinogen (2.5 wt.% HPF and 0.15 М NaCl, curve 3)
and after addition of nanosilica A-300 (3.5 wt.%) (2 and 4).
The behavior of polymers or proteins interacting with FMO in the aqueous media depended
on their concentration and salinity and pH of the solution [53,287,289,290]. An increase in
concentration of PEG led to diminution of the PEG/A-300 aggregate sizes (Fig. 192a, SPSDI and
SPSDV and Fig. 192b); however, an increase in the amount of PEG/A-300 in the suspension gives
the opposite result (Fig. 192b). The broadest SPSD was observed at minimal content of PEG (Fig.
192a). The main factor determining these regularities was stronger PEG-A-300 interactions than
PEG-PEG interactions. Increasing PEG content led to an increase in the number of PEG-PEG
contacts which were less effective than PEG-silica ones. At CPEG = 20 wt.% (SBET/SBET 0.3)
the amount of the macromolecules was not enough to bond all particles in large aggregates;
therefore, the SPSDN peak shifts toward smaller D values (Fig. 192a). An increase in CPEG led to
tendencies: diminution of the sizes of large aggregates (SPSDV = 100-1000 nm), and increase
(CPEG = 40 wt.%) and then diminution (CPEG = 80 and 90 wt.%) of the smallest aggregates sizes
(SPSDN < 100 nm). Notice that the SPSD for PEG/AST samples were not measured because of
fast sedimentation of particles [287].
The potential for PEG/A-300 depended weakly on the presence and content of PEG (Fig.
192c). This could be explained by several factors: (i) sizes of aggregates with A-300 [53,287] and
PEG/A-300 change in the same range; (ii) PEG macromolecules were uncharged; (iii) the surface
charge density of silica nanoparticles was not practically changed on interactions with PEG; and
(iv) charge distributions in the electrical double layer (EDL) and the shear plane [238,289,290]
were not strongly changed. However, for PEG/AST the potential dependence on CPEG was
stronger than that for PEG/A-300 (Fig. 192c). This difference could be explained by stronger
interactions of PEG molecules with Brønsted acid sites (SiO(H)Al, SiO(H)Ti, AlO(H)Ti) and an
increase in the amounts of PEG per surface square unit for PEG/AST in comparison with PEG/A-
300 due to large difference (approximately by order in magnitude) in the SBET values of these
oxides [287].
211
Fig. 192. Characteristics of the aqueous suspensions of PEG-oxides (1 wt.%): (a) particle size distributions for
PEG/A-300 with respect to the light scattering (SPSDI, curves 1a-4a), particle volume (SPSDV, 1b-4b) and
particle numbers (SPSDN, 1c-4c); (b) effective diameter for PEG/A-300 (0.2 and 1 wt.%); and (c)
potential for the diluted suspensions (0.1 and 0.025 wt.% for PEG/A-300 and PEG/AST, respectively) as a
function of pH.
In the first DSC scan, glass transition steps were observed only for two samples and no
crystallization or melting peaks were observed. In order to delete any previous thermal history of
the samples, as usually done in DSC [291-293], the measurements were performed in the second
DSC scan. The results presented below (Fig. 193 and Table 35) refer to these second-scan
measurements. The temperature dependences of structural changes (melting, crystallization, glass
transition) of PEG in the PEG/nanooxide systems change with PEG content (Table 35, Fig. 193).
The glass transition temperature (Tg) for PEG (53 oC, Table 35) was in agreement with previous
measurements of PEG and PEO at molecular weight ~ 0.1-1 MDa [294-296]. For all PEG/oxide
systems, except for PEG20/AST, the Tg value was higher than that for individual PEG (Table 35)
[287].
212
Fig. 193. Results of the second DSC scans for individual PEG (PEG100, curve 1) and (a) PEG-A-300 and (b) PEG-
AST at heating rate =10 oC/min and cooling rate = 30 oC/min.
Table 35. Characteristics of PEG/oxide systems from the DSC measurements [287].
Sample T
g
(oC)
T
cr
(oC)
T
m
(oC)
Δc
p/am
(J/g oC)
ΔΗ
cr/pol
(J/g)
ΔΗ
m/pol
(J/g)
X
cr
X
im
h
(wt.%)
PEG 53 36 67 0.70 157 159 0.81 0 1.8
PEG90/A-300 49 37 67 0.32 146 142 0.72 0.15 1.0
PEG40/A-300 - 17 54 0 85 90 0.46 0.54 3.2
PEG20/A-300 46 - - 0.66 0 0 0 0.06
PEG90/AST 46 39 68 0.34 154 151 0.77 0.12 0.3
PEG40/AST 47 39 64 0.59 143 145 0.74 0.07 0.4
PEG20/AST 66 35 59 0.52 125 145 0.74 0.07
Note that the melting enthalpy of the 100% crystalline PEG corresponds to ΔΗ0 =197 J/g [297].
Somewhat higher Tg differences were between PEG - PEG/AST than PEG - PEG/A-300,
due to stronger local interactions between PEG - AST than PEG - A-300 as observed in the FTIR
and PCS results [287]. Glass transition was not observed for PEG40/A-300 (signal at Tg was weak,
Fig. 193), whereas glass transition was observed for PEG20/A-300 (fully amorphous) with a
smaller CPEG value and for all PEG/AST samples. It was well established that in semicrystalline
polymers glass transition could not be observed if a fraction of amorphous phase was smaller than
a critical value because of the presence of a rigid amorphous phase [298]. However, glass
transition was observed for PEG40/AST due to the interactions of a larger fraction of the polymer
213
with silica than with AST. In general, we may expect an increase in the Tg value due to constraints
for PEG due to bonding to the oxide surfaces and a decrease in the Tg value because of an increase
in the free volume due to loosened molecular packing of the chains confined in voids between
adjacent primary particles [299-302]. It seems that the latter dominates only for PEG20/AST. The
lower Tg values at lower polymer content for PEG/AST (Table 35) were reasonable in terms of
constrains imposed to packing of the polymeric chains. Notice that in polyimide/silica hybrids
glass transition was not observed at Cpol < 40 wt.% because of strong polymer-silica interactions
[300]. Similar results were reported also for poly(2-hydroxyethyl acrylate)/silica [301] and
poly(methyl methacrylate)/silica nanocomposites [302]. Water as an additional factor affected the
Tg value, so that for PEG/A-300, where hydration levels (Table 35) were higher than for PEG/AST
(Table 35), the Tg values were lower because of stronger plasticization. It should be mentioned that
water content (Table 35) was calculated gravimetrically from the masses of hydrated and dried
samples, mamb. and mdried respectively, using simple equation [287]
amb. dried water
amb. amb.
( 100%) ( 100%)
m m m
h
m m
. (30)
In addition to the effects on the glass transition temperature discussed above, filler content
strongly affected the heat capacity step on the glass transition. The heat capacity step (Table 35)
has been normalized to the same amorphous polymer fraction (Δcp,am). Compared with pure PEG
the Δcp,am value decreased strongly at low filler contents and increased again on further particle
addition (Table 35). Generally, filler addition reduces the polymer mobility, while a thin polymer
layer (~nm) close to the nanoparticles (surface, porous) was almost immobilized [299-302].
Despite some uncertainty of the Δcp calculation. This was a useful tool for estimation of a fraction
of immobilized polymer, Xim. The latter may be defined as a fraction of amorphous polymer which
did not contribute the glass transition [287]
p,am
im cr
p,am(PEG)
(1 )(1 )
c
X X
c
, (31)
where Δcp,am refers to the nanocomposites
p,DSC
p,am
cr PEG(1 )
c
c
X C
, (32)
Δcp,am(PEG) is the corresponding quantity for the neat polymer sample, CPEG is the weight fraction
of the polymer in composite, and Χcr is the degree of crystallinity calculated through the melting
enthalpy, ΔΗDSC [287]
DSC
cr
0 PEG
H
X
H C
. (33)
Between 0.06 and 0.54 (Table 35) Xim is a non-monotonous function of CPEG. This differs from
polymer nanocomposites with dispersed oxide nanoparticles synthesized by the sol-gel method in
the polymer matrix, where an approximately linear increase of Xim was observed at low filler
content increased [299]. For fumed oxides the opposite behavior was observed. The higher the
content of polymer in the systems, the less restricted mobility was observed in general. This
phenomenon indicates layer-by-layer polymer shell growth on the oxide surface during material
preparation, leading to diminution of both polymer-oxide interactions for subsequent polymer
layers and the restriction of the polymer mobility. The PEG immobilization was more intensive for
A-300 than AST50 (with smaller SBET). Larger Xim values for PEG/silica were due to a larger
fraction of PEG being in the interfacial layer at the silica surface (larger SBET value). Nonlinear
changes in Xim(CPEG) (Table 35) could be due to nonlinear changes in particle aggregation, which
differs for PEG/A-300 and PEG/AST in addition to the effects of different SBET values [287].
There are FMO effects on crystallization and melting of PEG [287]. It is obvious that at
low oxide content the crystallization temperature (Tcr) somewhat increased because the oxide
particles could act as crystallization nuclei (facilitating crystallization) (Table 35 and Fig. 193).
214
However, at higher oxide content Tcr strongly decreased (crystallization was suppressed) probably
due to spatial constrains imposed to the diffusion of polymer chains. The melting temperature Tm
was not affected at low oxide contents but decreased at higher contents corresponding to lower
quality or smaller polymer crystallites which melt (deform) easier with temperature [303]. Finally,
the degree of crystallinity Xcr decreased in the nanocomposites (Table 35); i.e. oxide nanoparticles
reduce the crystallization ability of PEG there. This was in agreement with results obtained for
nanocomposites with particles dispersed in a polymeric matrix (CPEG > Cox), where, in general,
reduction of polymer crystallization was observed with decreasing Tcr, Tm, and Xcr values with
increasing filler content [304]. The effects were stronger for PEG/A-300 (smaller oxide
nanoparticles), and no crystallinity was observed in PEG20/A-300. Moreover, a stronger Tm
reduction for that series indicates that the quality of PEG crystallites was lower comparing to the
PEG/AST series [287].
It should be mentioned that in semicrystalline polymers There was also a fraction of
immobilized amorphous polymer at the interfaces with crystallites (i.e. rigid amorphous phase)
[239]. Measured values depended on combined effects of both constrains caused by nanoparticles
and polymer crystallites. Thus the differences in the DSC results were due to the differences in the
SBET values (i.e. primary particle sizes), relative amounts of PEG per surface area unit (at the same
CPEG values), and the numbers of different active surface sites at silica and AST surfaces strongly
interacting with the macromolecules (Table 36) [287].
Table 36. Temperature of different relaxations maxima (Trel) [287].
T
rel
(oC) Sample
β β΄ β
JG
Α MWS CR
PEG - - 57 42 27 7
Dried PEG 118 87 54 22 1 -
Melted PEG 115 77 60 36 20 7
PEG90/A-300 127 92 57 39 27 3
PEG40/A-300 124 91 76 58 48 34
PEG90/AST 124 89 64 44 28 14
PEG40/AST 122 82 - 43 36 18
Dried PEG90/A-300 123 85 - 15 8 -
Dried PEG40/A-300 107 - - 27 20 -
Dried PEG90/AST 112 85 58 16 2 -
Dried PEG40/AST 118 79 - 40 20 7
The TSDC measurements were performed for both initial composites (containing some
amounts of water as they were hydrophilic) and dried samples to clarify the origin of the relaxation
processes observed (Figs. 194 and 195) [287]. At temperatures below Tg (i.e. in the glassy state,
where secondary relaxations were observed) three observed weak peaks (shoulders) could be
attributed to the β, β΄ and βJG processes discussed below. The main process, α-relaxation
(associated to the glass transition), follows as a shoulder on a stronger Maxwell-Wagner-Sillars
(MWS) peak, which was related to the interfacial relaxation phenomena. At higher temperatures,
the conductivity relaxation (CR) [305] was observed, followed by a strong current uptake, related
to ionic dc-conductivity, which was characteristic for hydrophilic systems (Table 36). The γ-
relaxation of PEG at about 160oC could not be fully seen in this particular temperature range. The
identification and discussion of that and other secondary relaxations will be based on the DRS
measurements [287].
215
Fig. 194. TSDC thermograms for (a) PEG/A-300 and (b) PEG/AST initial and dried; inserts showed low-temperature
ranges related to -relaxations.
Fig. 195. TSDC thermograms for dried PEG, PEG/A-300 and PEG/AST.
Three relaxation processes at low temperatures (sub-glass relaxations) were characterized
by relatively low values of current maxima (Figs. 194 and 195) corresponding to relatively small
values of dielectric strength typical for dipolar processes in the glassy state [287]. Additionally,
these systems include a small quantity of water (Table 35) which could affect relaxations at low
temperatures because of plasticization of the polymer matrix. In particular, the following
relaxations were observed: β-relaxation (between 127 and 107oC) affected (plasticized) by
water; β΄-relaxation (from 92 to 79oC); βJG-relaxation (from 76 to 54oC) (Johari-Goldstein
relaxation) (Table 36), which could be observed as a shoulder on the low-temperature side of the
α-relaxation peak and located at 15-30 oC lower than the latter (Fig. 195 and Table 36) [287].
The PEG relaxation processes at higher temperatures exhibit higher values of current
maxima (by about two orders in magnitude) and dielectric strength indicating significant
contribution of charge carrier motion [287,306-309]. The α-relaxation (from 58 to 42oC) was
216
associated with the glass transition of PEG plasticized by water (Table 36). A small displacement
of the -relaxation toward higher temperatures was observed in PEG90/A-300 with respect to pure
PEG, and a larger displacement but toward lower temperatures was observed for samples at 40
wt.% PEG, especially for PEG40/A-300 (Fig. 194). This led to the corresponding displacement of
the distribution function of activation energy of depolarization (Fig. 196). This effect was more
clearly observed for dried samples (Fig. 192). The results could be explained by two opposite
effects: constrains due to PEG-oxide interactions and an increase in free volume in more loosely
packed macromolecules in voids between adjacent primary oxide particles. Both effects increased
with increasing CPEG value. Relatively small amounts of bound water were in composites (Table
35); nevertheless, it caused essential plasticization effect resulting in the displacement of the
TSDC response toward lower temperatures in comparison with dry samples. This displacement
was larger for PEG/A-300 than PEG/AST because of higher water content in PEG/A-300 samples
(Table 35) [287].
The MWS relaxation (from 48 to +2oC) (Table 36), typically located by 8-20 oC higher
than the α-relaxation, was associated with interfacial polarization and space charges [287]. This
relaxation could be also assigned to the αcr-relaxation associated with the glass transition of some
parts of polymers strongly restricted between condensed crystalline regions, and observed for PEG
and PEO. However, this suggestion comes in contradiction with the high strength and the shape of
the relaxation. Judging from the fact that the dielectric strength of the relaxation has a clear
dependence on the degree of crystallinity, the most probable explanation of its origin could be
based on release of charges trapped during the polarization step at the interfaces between
crystalline and amorphous regions. It was affected by water and oxide nanoparticles similar to the
α-relaxation. The CR relaxation observed at higher temperatures following the main relaxation by
15-20 oC higher refers to changes in conductivity related to dc (direct current in isothermal
measurements at low effective frequencies) and ac (alternating current at higher frequencies). This
relaxation was not observed for some samples that could be explained by deplasticization (water
drying) of the main dielectric relaxation and the fixed polarization temperature (20oC). It was
affiliated to space charges (conductivity effects), and also observed for PEO. Compared with the
results of the DSC measurements, the Tα and TMWS values showed similar trends to the Tg value
(Table 28), providing further support that in the high polymer content samples polymer-filler
interactions dominate over free volume effects, whereas the contrary was true for the low polymer
content samples. We come back to these relaxations later in relation to the DRS data. It should be
mentioned that the two peaks which follow the α-relaxation at higher temperatures (MWS and CR
peaks, respectively) could be also MWS relaxation indicating the variety in phase separation
surfaces (polymer-filler, amorphous-crystalline polymer, etc.) [287].
Fig. 196. Distribution functions of activation energy of dipolar relaxations for individual PEG and PEG (40 wt.%) -
oxide systems (calculated from the TSDC data).
217
Typical DRS spectra (dielectric loss against frequency at different temperatures) were
presented in Fig. 197. At low temperatures the secondary γ and β relaxations were well observed.
As temperature increased, the response gets dominated by the segmental dynamics, space charge
polarization/relaxation and conductivity effects, giving rise to high values of ε΄΄ and ε΄ (Fig. 198),
orders of magnitude higher than at low temperatures. A main DRS peak was well observed for all
samples at higher temperatures (~30 oC) (Fig. 198). Judging from the TSDC and the DRS
isochronal (constant frequency) plots [287], this peak was probably composed from more than one
relaxations (at least MWS and CR).
An overall increase in the response was observed for nanocomposites with both oxides in
comparison with individual PEG; however, There was no systematic variation with composition,
but only in combination with the dehydration of the samples, which decreased the response and the
-relaxation was better resolved for the dry samples (Fig. 197c,d). The response value increased
for series pure PEG < 90 wt.% < 40 wt.% PEG in the region of the -relaxation [287].
Fig. 197. DRS for (a) PEG and (b) dried PEG/A-300 at CPEG = 40 wt.% over the range between 150 and +30 oC.
Secondary relaxations at 90 oC for (c) PEG/A-300 and (d) PEG/AST.
The resolving power of the dielectric techniques gets higher as frequency decreased. That
was the reason why in the TSDC measurements (104–102 Hz) the α, MWS, and CR relaxations
were better resolved in comparison with DRS, where one broad relaxation was observed. The data
for PEG were included that was melted between the plates of the capacitor (melted PEG). We may
expect better resolution in that case because of the better contact of the sample with the electrodes
and a smaller sample thickness (~50 μm) [287].
Based on the isothermal (Fig. 197) and the isochronal plots and by recording the frequency
and temperature maxima, respectively, for all the relaxations observed, the Arrhenius plots were
constructed (Figs. 199 and 200) [287]. The TSDC data (peak temperatures) at the equivalent
218
frequency of 1.6 mHz corresponding to a relaxation time of ~100 s and the DSC data (Tg) at the
equivalent frequency of 10 mHz have been included in the plots to facilitate comparison of the
results obtained using three techniques. The Arrhenius equation
max 0 exp aE
f f
kT
(34)
was well fitted to the data for the secondary relaxations and the activation energy (Ea) values were
determined.
Fig. 198. The dielectric loss " (a, c) and permittivity ' (b, d) as a function of frequency at 30 oC for individual PEG
(initial and dried), PEG/A-300 (a, b) and PEG/AST (c, d) at CPEG = 90 and 40 wt.%.
The -relaxation was followed only by DRS since extrapolation to low frequencies showed
that the corresponding TSDC peak (~ 160 oC) was out of the temperature range of the TSDC
measurements. The activation energy of the -relaxation for individual PEG was equal to 38
kJ/mol which was in agreement with the literature data. This relaxation was attributed to
‘crankshaft motions’ of methylene sequences [287].
The situation was more complex with the -relaxation, e.g., for individual PEG two
strained lines could be found in the Arrhenius plot (trends (1) and (2) in Fig. 199) [287]. A change
in slope was observed at about 45 oC for pure PEG to higher activation energy from Ea = 63 to
146 kJ/mol. This behavior may suggest that the relaxation was of the Goldstein-Johari type
[310,311], and for this sample the change in slope occurred close to Tg determined with DSC and
TSDC. However, the single behavior was observed for several composites (Fig. 200) with straight
lines (by Arrhenius equation). For PEG40/AST a similar behavior was observed at low activation
energy (56 kJ/mol). Close activation energy (60-63 kJ/mol) was observed for initial PEG90/oxide
and PEG40/AST samples. High activation energy (183 kJ/mol) was observed for dried
PEG40/AST sample which was higher than that for dried PEG40/A-300 (95 kJ/mol). Trend (1)
(Figs. 199 and 200) for βDRS relaxation observed for initial samples (hydration 0.3-3.2 wt.%) was
alike to a water relaxation [53,214], on mixtures of ethylene glycol oligomers and PEG600 with 35
wt.% of water. The activation energies of the trends (1), (3) and (4) were close. The observed
horizontal displacement was possibly related to the larger amounts of respective groups (water
219
molecules). Notice that the activation energies calculated from the DRS data and shown above
were in the same range as the Ea values based on the TSDC measurements (Fig. 199) because the
mechanisms of the relaxations observed here by both methods were linked to the mobility of polar
structures in PEG and bound water molecules in composites [287].
Fig 199. Comparative Arrhenius plots for pure PEG: points from isothermal and isochronal plots; and respective
TSDC and DSC (1st and 2nd scans) data were added.
At higher temperatures (mainly at T > 20 oC, Figs. 197-200) a broad relaxation could be
identified as a superposition of the α-relaxation, MWS and CR by comparing with each other
results for ε΄, tanδ, ΄΄ and TSD current. The dielectric Tg value was estimated as temperature,
where ε΄΄ values start to increase significantly, by almost one order of magnitude, at the lowest
frequency. This procedure provided the data for the α-relaxation included in the Arrhenius plots.
Figs. 172 and 173 showed the VTF behavior of the α-relaxation and plasticization by water. As a
whole the overall complex trends of the -relaxations in relation to the Johari-Goldstein relaxation
and the ‘α-β’ temperature range were of interest and need additional investigations [287].
It should be stressed that within this work the glass transition temperature values were
recorded for the various compositions by using three different methods, DSC, TSDC and DRS
[53,214,245,287]. The results by these three techniques showed in general the same trends with
changes in PEG/oxide composition and close Tg values for the systems with similar thermal
treatments and equivalent frequencies.
In the aqueous media PEG macromolecules form stronger hydrogen bonds with surface
hydroxyls of oxide nanoparticles (silica, alumina/silica/titania) than with macromolecules or water
molecules that result in the formation of hydride aggregates (50-1000 nm in size) with well
distributed PEG molecules and allow to form relatively uniform composites on drying of the
systems [53,287]. The weakly hydrated (0.3-3 wt.% of water) PEG (20-90 wt.%)/oxide powders
were characterized by full coverage of nanoparticles by macromolecules. Therefore, relaxational
characteristics of individual PEG and PEG in composites differ and depended on polymer content,
temperature and treatments (e.g., drying or melting). Addition of nanoparticles suppresses
hydrophilicity, crystallinity and mobility of the amorphous polymer fraction, the effects becoming
in general stronger with increasing filler fraction. However, some of the characteristics do not
demonstrate a linear dependence on the CPEG value because changes in the interfacial layer
structure, confined effects, number of nanoparticles interacting with a macromolecule (which was
several times longer than the size of nanoparticles, especially for A-300 at 8 nm in diameter) with
CPEG value. Drying and melting of PEG in the PEG/oxide systems, despite low amounts of residual
water (h 3.2 wt.%), results in significant inhibition of polymer motion and the activation energy
of the DRS-relaxation increased in 1.5-3 times. Consequently, appropriate changes in hydration of
hydrophilic PEG/nanooxide composites allow one to strongly change their characteristics even at
220
the same content of polymer. The influence of the nanooxides on the thermal transitions and the
dielectric relaxation mechanisms was stronger for silica than AST, although the local interaction
between polymer and AST was found to be stronger. Combining with the fact that silica particles
were much smaller, the overall conclusion was that the polymer was better distributed for silica
rather than AST [287].
Fig. 200. Comparative Arrhenius plots for PEG and PEG/oxide composites: points from isothermal and isochronal
plots; respective TSDC and DSC (1st and 2nd scans) data were added. Two relaxation lines were also added
for mixture of ethylene glycol oligomers and PEG600 with 35 wt.% of water (trends (3) and (4)
respectively).
The interfacial behavior of water located in narrow pores [312-315] could affect
interactions of macromolecules with FMO [39,40,53,69,107,174-183,214,230,248, 268,271,280,
287, 288,290, 316-322] and other materials [103,216,217,245,323-326]. These effects appear as
changes in the relaxation characteristics of adsorbates. During controlled preparation of
nanomaterials [327] and filled polymeric composites [328] these effects should be considered, as
well as in the cases of catalytically active FMO such as titania [329-331] or to control the toxicity
of nanomaterials [332]. Note that many of aspects of the interfacial and temperature behaviors of
various adsorbates bound to FMO were analyzed in several monographs [53,333,334] and recent
research papers [335-368]. It should be noted several additional aspects in changes in the
characteristics and properties of FMO due to hydrocompaction [369-372].
Hydrocompartion, surface modification and confined space effects
SEM (Fig. 201a,b) and TEM (Fig. 201c,d) images of initial (a, c) and hydro-compacted (b,
d) nanosilicas (Table 38) demonstrates that the particulate morphology does not practically change
with respect to primary NPNP [369].
The secondary structures (NPNP aggregates), as well ternary structures with agglomerates,
become denser after hydro-compaction of nanosilica [369].
The differential pore size distribution functions f(r) based on the small-angle X-ray
scattering (SAXS) data were calculated using Fredholm integral equation of the first kind for
scattering intensity I(q) [373]
max
min
2
2
sin cos
( ) ( ) ( )
( )
r
r
qr qr qr
I q C V r f r dr
qr
, (35)
where C is a constant, q = 4sin()/ the scattering vector value, 2 is the scattering angle, is the
wavelength of incident X-ray, V(r) is the volume of a pore with radius r (proportional to r3), and
f(r)dr represents the probability of having pores with radius r to r + dr. The values of rmin (=
/qmax) and rmax (= /qmin) correspond to lower and upper limits of the resolvable real space due to
instrument resolution. This equation was solved using the CONTIN algorithm [195]. The f(r)
221
function could be converted into incremental PSD (IPSD) (ri) = (f(ri+1) + fV(ri))(ri+1 ri)/2 for
better view of the PSD at larger r values.
Table 38. Textural characteristics of unmodified and hydro-compacted A-300 (SCV/SCR method)
Sample hcp
(g/g)
SBET
(m2/g)
Vp
(cm3/g)
<RV>
(nm)
<RS>
(nm)
Snano
(m2/g)
Smeso
(m2/g)
Smacro
(m2/g)
Vnano
(cm3/g)
Vmeso
(cm3/g)
Vmacro
(cm3/g)
A-300 0 294 0.850 21.13 6.22 45 233 17 0.023 0.560 0.266
cA-300 0.5 305 0.918 34.62 6.86 57 224 24 0.027 0.388 0.503
cA-300 1.0 303 1.109 31.49 8.10 52 219 32 0.025 0.484 0.600
cA-300 1.5 228 1.116 41.98 11.25 34 156 39 0.017 0.297 0.802
cA-300 2.0 227 1.447 26.90 13.79 23 164 40 0.012 0.819 0.616
cA-300 2.5 232 1.540 23.25 13.12 31 163 38 0.018 0.977 0.545
cA-300 3.0 228 1.492 19.82 12.14 34 163 31 0.020 1.082 0.389
cA-300 4.5 234 1.488 18.13 11.00 47 167 21 0.027 1.177 0.284
cA-300 5.0 231 1.515 21.64 12.34 41 157 34 0.023 1.024 0.468
Note. The values of Vnano and Snano were calculated by integration of the fV(R) and fS(R) functions, respectively, at 0.35
nm < R < 1 nm, Vmeso and Smeso at 1 nm < R < 25 nm, and Vmacro and Smacro at 25 nm < R < 100 nm. The values of <RV>
and <RS> as the average pore radii were calculated as a ratio of the first moment of fV(R) or fS(R) to the zero moment
(integration over the 0.35-100 nm range) <R> = f(R)RdR/f(R)dR.
Fig. 201. (a, b) SEM and (c, d) TEM images of A-300 (a, c) initial and hydro-compacted at water content hcp = (b) 3
g/g and (d) 4 g/g (scale bars are (a) 200 nm, (b) 100 nm, and (c, d) 20 nm).
To calculate the particle size distribution (PaSD) functions, several models of particles (e.g.
spherical, cylindrical, lamellar ones and various blends of them) could be used. For spherical
particles, integral equation similar to Eq. (35) could be written as follows
222
max
min
( ) ( , ) ( )
R
R
I q C P q R f R dR , (36)
where C is a constant, R is the radius of particles, f(R) is the distribution function (differential
PaSD), and P(R) is the form factor for spherical particles [374] (the kernel of the integral equation
S7): P(q,R) = (4πR3/3)2[Φ(q)]2 and Φ(q,R) = (3/(qR)3)[sin(qR) qRcos(qR)].
The PaSD with respect to the volume of particles (as abundance in vol%) could be
calculated as follows
abundance(vol%) = 3 3( ) / ( )R f R R f R dR . (37)
For cylindrical particles, there are two variable parameters, such as the radius (R) and
length (H) of cylinders
max max
min min
( ) ( ) ( ) ( , , )
H R
H R
I q C f H f R P q H R dHdR , (38)
where
/2
1
0
sin 0.5 cos2 ( sin )
( , , ) sin
sin 0.5 cos
qHJ qR
P q H R CV d
qR qH
, J1(x) is the first-order Bessel
function, V = πR2H is the cylinder volume, and C is a constant [374].
For lamellar particles [374]
max
min
( ) ( , ) ( )
L
L
I q C P q L f L dL , (39)
where
2
2 sin( / 2)
( , ) ( / )
( / 2)
qL
P q L L q
qL
, L is the lamellar thickness, and the prefactor (1/q2) is the
so-called Lorentz factor required to randomize the orientation of the lamellar particle [374]. In the
case of complex systems, several models with various blends of spherical, cylindrical and lamellar
particles could be used with certain weight coefficients.
For a complex model of particles, the integral equation includes three terms
max
min
max
min
max max
min min
223
3
2
2
/22
1
0
4 3
( ) (sin( ) cos( ) ( )
3 ( )
sin( / 2)
( / ) ( )
( / 2)
sin 0.5 cosθ2 ( sinθ)
( ) ( ) sinθ
sinθ 0.5 cosθ
R
sph
R
L
lam
L
H R
cyl
H R
R
I q c qR qr qR f R dR
qR
qL
c L q f L dL
qL
qHJ qRR
c f H f R d
H qR qH
dRdH
(40)
where I(q) is the X-ray scattering intensity, q = 4�sin(�)/� is the scattering vector value, 2� is
the scattering angle, � is the wavelength of incident X-ray, R is the radius of particles, H and R are
the length and radius of cylinders, L is the lamellar thickness, f(R), f(L), and f(H) are the
distribution functions, J1(x) is the first-order Bessel function, cx are the weight coefficients
calculated, as well f(x) functions, using a self-consistent regularization procedure. Equation (40)
could be solved using a self-consistent regularization procedure [86].
The chord size distribution, G(r) as a geometrical statistic description of a multiphase
medium, was calculated from the SAXS data [375,376]
2
4
2
0
sin
( ) ( ) 4
d qr
G r C K q I q dq
dr qr
, (41)
where K is the Porod constant (corresponding to scattering intensity I(q) ~ Kq4 in the Porod
range).
The specific surface area from the SAXS data was calculated (in m2/g) using equation
223
4
SAXS 10 (1 )
a
K
S
Q
, (42)
where = a/0 is the solid fraction of adsorbent, and Q is the invariant
2
0
( )Q q I q dq
. (43)
The Q value is sensitive to the range used on integration of Eq. (43) (since experimental q values
are measured between the qmin and qmax values different from 0 and ). Therefore, the invariant
value Q was calculated using equation [377]
max
min
2
max( ( ) ) /
q
i i i
q
Q I q b q q K q
(44)
where b is a constant determined using equation
I(q)q4 = K + bq4 (45)
valid in the Porod range.
Fig. 202. Pore size distributions (a, c) differential (PSD) and (b, d) incremental (IPSD) calculated using (a, b) SAXS
and (c, d) nitrogen adsorption-desorption isotherms with (c) NLDFT (dV/dR) and (d) SCV/SCR (IPSD)
methods for initial and hydro-compacted A-300 samples prepared at different content of water (hcp = 0.5 –
5.0 g/g).
The analyses of the SAXS and nitrogen adsorption-desorption isotherms (Table 38, Figs.
202 and 203) are in agreement with microscopic images. However, there are some additional
factors, which should be considered upon the analysis of the textural characteristics of nanosilicas.
First, the empty volume in the nanosilica powder (Vem = 1/b1/0, where b and 0 are the bulk
and true densities of samples) is much greater than the Vp value of adsorbed nitrogen even at p/p0 =
0.999. For example, Vem = 21.8 cm3/g for initial A-300 at b = 0.045 g/cm3, but the value of Vp is
much smaller (Table 38). This underestimation of the value of Vp calculated from the nitrogen
224
adsorption is due to very weak influence of the pore walls (nanoparticle surface) on the nitrogen
molecules located in macropores far from the solid surface in loose silica agglomerates. After
hydro-compaction of A-300, the difference between the values of Vem and Vp strongly decreases,
e.g. Vem = 3.45 cm3/g and Vp = 1.49 cm3/g for cA-300 treated at hcp = 4.5 g/g. Therefore, the
various PSD (Fig. 2) describe only a part of voids between NPNP in secondary/ternary structures,
and these parts differ for PSD based on the data of the SAXS, NLDFT, and SCV/SCR methods
due to features of these methods.
Fig. 203. Particle size distributions (PaSD) for initial and hydro-compacted nanosilica calculated using the SAXS
patterns analyzed as described in ESM file.
Fig. 204. The specific surface area estimated from nitrogen adsorption (SBET), SAXS data for maximum of PaSD at dm
(SSAXS,d = 6/(dm0), where 0 = 2.2 g/cm3 is the true density of fumed silica), IR spectra from integral
intensity of surface TO modes (Si-O) at 1200 cm-1 (SIR,SiO), and a ratio of the integral intensity of bands of
silanols at 3745 cm-1 and Si-O overtone at 1870-1860 cm-1 (SIR,OH) and bulk density (b) vs. the water
amount used on silica compaction.
According to SAXS PSD and IPSD (Fig. 202a,b), contributions of nanopores at radius R <
1 nm increases with increasing hcp value. However, contribution of narrow mesopores at 1 nm < R
< 10 nm decreases upon nanosilica compaction. Note that SAXS can give information on pores
inaccessible for nitrogen molecules. Therefore, contribution of narrow nanopores in the SAXS
PSD is much greater than that in the NLDFT PSD (Fig. 202c) or SCV/SCR IPSD (Fig. 202d).
However, the tendency in a decrease in contribution of narrow mesopores with increasing h value
is similar to that in the SAXS PSD. Both NLDFT PSD and SCV/SCR IPSD demonstrate an
increased contribution of broad mesopores at 10 nm < R < 25 nm with increasing hcp value. A
more complex picture is observed for the SAXS PSD in the range of broad mesopores, as well for
macropores at R > 25 nm, vs. the hcp value. The SCV/SCR method shows that the maximal
contribution of macropores (Table 38, Vmacro, Smacro) is observed in a middle range of hcp = 1-2 g/g
225
with a maximum of the SCV/SCR IPSD for macropores at hcp = 1.5 g/g. Sample cA-300 at hcp
=1.5 g/g corresponds to the first sample with the water amount greater than Vp and this sample is
the first one with partially collapsed secondary structure since a significant decrease in the value of
SBET (Table 38, Fig. 204) is observed. However, the value of b is middle at hcp = 1.5 g/g and it
continue grows with increasing hcp to 4.5 g/g [369].
Fig. 205. IR spectra of nanosilicas in the range of (a) the Si-O stretching vibrations, (b) 4000-1350 cm-1, and (c, d) the
SiO-H stretching vibrations depending on the amount of water used upon silica compaction.
Fig. 206. 1H NMR spectra (recorded at different temperatures upon heating from minimal one used) of water (1.1 g/g)
adsorbed on a surface of nanosilica A-300 (a) initial (dot-dashed lines) and hydro-compacted at hcp = (a) 1.5
g/g (solid lines) and (b) 2.0 g/g (dot-dot-dashed lines) and 3.0 g/g (solid lines).
The estimations of the specific surface area (Fig. 204) using a peak value in the SAXS
PaSD (Fig. 203) or the use of the IR spectra (Fig. 205) confirm that the value of hcp = 1.5 g/g is
critical for reorganization of the NPNP in secondary and ternary structures. Note that the
estimation of the specific surface area gives SSAXS = 332 and 251 m2/g for initial A-300 and cA-
300 treated at hcp = 5 g/g, respectively. The SSAXS values are larger than SSAXS,d or SBET because the
former corresponds to total surface area, a part of which is inaccessible for N2 molecules, and upon
calculation of the SSAXS,d value, only the peak PaSD value (Fig. 203) is used. Certain structural
226
changes in cA-300 are already observed at lower hcp values (Fig. 205c,d), even at hcp = 0.3 g/g. It
is possible that these changes at small hcp values (0.3-0.5 g/g) are due to complex treatment with
wetting-stirring-aging-drying. These textural changes features can be explained by the clustered
adsorption of water, i.e. it can be located in contact zones between adjacent NPNP. In this case, the
amount of water at hcp = 0.3 g/g is greater than contribution of nanopores (Table 38, Vnano).
Therefore, it can stimulate reorganization of contacts between neighboring NPNP in their
aggregates. This reorganization leads to an increase in the accessibility of the NPNP surface for
nitrogen molecules and the SBET value increases at hcp = 0.5 and 1.0 g/g (Table 38), but this effect
is stronger at hcp =0.5 g/g [369].
Additional information on the reorganization of NPNP in secondary/ternary structures of
the cA-300 samples in comparison with A-300 could be obtained using low-temperature 1H NMR
spectroscopy and NMR cryoporometry. The 1Н NMR spectra of water (h = 1.1 g/g) bound to A-
300 and cA-300 (Fig. 6) demonstrate single, slightly asymmetrical signal at the chemical shift of
proton resonance H 5 ppm close to that of liquid bulk water. Intensity of this signal decreases
with decreasing temperature due to freezing of a part of bound water. A fraction of water frozen at
T < 260 K can be attributed to strongly bound water (SBW) (Fig. 207). Water frozen at 260 K < T
< 273 K is weakly bound water (WBW). The amounts of unfrozen water (Cuw) vs. T (Fig. 207a)
could be used to estimate the size distribution of pores filled by unfrozen water (Fig. 208) using
the Gibbs-Thomson relation for the freezing/melting point depression of water (or other liquids
confined in pores) depending on the pore sizes. These distributions could be used to estimate a part
of pores (in different size ranges) containing unfrozen water (Table 39).
Fig. 207. (a) Temperature dependences of the content of unfrozen water (Cuw) estimated from integral intensity of the
1H NMR spectra recorded at different temperatures, and (b) changes in the Gibbs free energy G vs. Cuw (for
A-300 and cA-300 samples with constant hydration upon NMR measurements h = 1.1 g/g).
The amount of adsorbed water in samples studied using the 1H NMR spectroscopy was
constant at h = 1.1 g/g that is smaller than Vp of cA-300 but it is greater than Vp of A-300 and close
to the critical value of water strongly affecting the reorganization of NPNP in the
secondary/ternary structures (Table 38). The amount of SBW increases with increasing hcp (Table
39, Cuw
s); therefore, the average melting temperature <Tm> decreases. There is a tendency in
increasing of the values of Gs (related to SBW) and S (related to all bound water) with hcp
(Table 39). This is due to an increase in contribution of nanopores (Table 39, Snano,uw, Vnano,uw)
filled by unfrozen water with increasing hcp used on the silica compaction treatment [369].
Table 39. Characteristics of water bound to unmodified and hydro-compacted A-300 (h = 1.1 g/g)
Sample hcp
(g/g)
Сuw
s
(mg/g)
Сuw
w
(mg/g)
ΔGs
(kJ/mol)
γS
(J/g)
γS
(mJ/m2)
<Tm>
(K)
Snano,uw
(m2/g)
Smeso,uw
(m2/g)
Smacro,uw
(m2/g)
Vnano,uw
(cm3/g)
Vmeso,uw
(cm3/g)
Vmacro,uw
(cm3/g)
A-300 0 50 1050 2.58 11.55 114.5 266.1 32 63 6 0.014 0.790 0.085
cA-300 1.5 75 1025 2.74 14.17 115.5 264.8 48 68 6 0.020 0.796 0.084
cA-300 2.0 110 1090 2.88 15.39 121.2 264.4 48 71 8 0.020 0.819 0.105
cA-300 3.0 170 1030 2.66 14.10 102.6 261.9 82 53 2 0.035 0.578 0.030
Fig. 208. Size distributions of unfrozen water structures filling voids between NPNP (a) differential (PSD) and (b) incremental (IPSD) calculated using NMR cryoporometry
method (at constant h = 1.1 g/g) applied to initial A-300 and cA-300 at different hcp values.
227
228
This is matching with TEM images (Fig. 209) of A-300 and related modified fumed silicas
with no strong compaction of the secondary structures (note that stronger compaction results in
greater b values).
Fig. 209. TEM images of A-300 (a) unmodified and (b-e) modified (b) AP1, (c) AP2D, (d) AP4D, and (e) AP5D (scale
bar 50 nm).
229
Fig. 210. TEM images of (a) cA-300, (b) cAP3D, (c) A300, (d) AEP1, (e) AEP2D, and (f) AEP4D.
For other silicas, the modification does not result in great changes in the compaction of the
secondary structures (Fig. 210), despite a certain increase in the r values (Table 40). These results are
of importance for the application of these composites as fillers for various nonpolar media (e.g.
230
polymers), since strong compaction of ANPNP can lead to a worsened and nonuniform distribution
of filler particles in the polymer matrices. This can lead to impairment of the mechanical and other
characteristics of the final materials. Note that the secondary structures of Syloid 244 (precipitated
silica) look like more strongly compacted than for other silicas such as cA-300 or TS 100 undergoing
compaction treatments at low and high temperatures. However, Syloid 244 has a maximal SBET value
and a minimal r value (Table 40) among the silicas studied. Therefore, the contents of attached
hydrophobic functionalities (Table 40, CC+CH) for samples SP3D and cAP3D are similar. It is worth
mentioning that nanopores (voids between adjacent NPNP in ANPNP at R < 1 nm in pore radius) are
practically absent (Table 40, Vnano, Snano, Fig. 212) in the unmodified and modified silicas. This is of
importance for effective silica modification as nanopores are poorly accessible for relatively large
PMDS molecules or their fragments [345].
The values of <RV> > 25 nm and Vmacro > Vmeso (Table 40), as well the PSD (Fig. 212, PSD
maxima at R > 25 nm), suggest that the organization of ANPNP corresponds to meso/macroporous
materials rather than to mesoporous ones. This could be of importance for SiO2 surface modification
by the polymer fragments (formed as a result of interactions with DMC), since the surfaces of NPNP
are better accessible in broad mesopores and macropores than in narrow mesopores or nanopores
(voids between NPNP in ANPNP), while for TS 100 and Syloid 244, Vmacro < Vmeso, but <RV> > 25
nm. Thus, both these compacted silicas are characterized by significant macroporosity [345].
The modification of NAS by short PDMS (PMS5 at an average degree of polymerization of ca.
11); long PDMS (but with DMC addition used to cleave the Si-O bonds in PDMS); or HMDS
typically leads to reduction of the values of Vp and SBET (Table 40); and an increase in the average
radius (r) of nanoparticles. However, changes in the values of the textural characteristics of
mesopores and macropores are different for modified A-300 (A and AE series) and other silicas (cA,
T, and S series) due to the differences in the PSD of the unmodified silicas reflecting the
organization of ANPNP. However, changes in the NPNP per se after additional treatment of
nanosilicas (e.g. hydro-compaction A-300 cA-300 and A-300(AE) TS 100) are not significant.
For example, SAXS analysis of the particle size distributions (PaSD) for A-300 and cA-300 shows
similar curves with the same position for the PaSD maximum (Fig. 211), which shifts toward larger
values due to cA-300 functionalization (cAP3D). This is in agreement with the changes in the r
values (Table 40) computed from the nitrogen adsorption isotherms. It should be noted that hydro-
compaction (cA-300) or PDMS/DMC modification (cAP3D) of A-300 results in a diminishing
ANPNP contribution (Fig. 211, r > 10 nm), in comparison to unmodified A-300. In other words,
both processes cause a certain amount of decomposition of ANPNP (see e.g. Fig. 210a,b) [345].
Fig. 211. Nanoparticle size distributions (SAXS data treated assuming that particles are spherical) for unmodified A-300,
hydro-compacted cA-300, and cA-300 modified by P3D.
231
Fig. 212. Incremental pore size distribution calculated using the nitrogen ads-des isotherms treated with the DFT method
(with a model of voids between spherical NPNP in silica) for (a) A-300 unmodified and modified, (b) cA-300
unmodified and modified, (c) A300 (AE) unmodified and modified, (d) TS 100 and Syloid 244 unmodified
and modified samples.
The difference in the organization of secondary structures ANPNP causes certain changes in
the characteristics of silicas (Figs. 209-215), e.g. in the adsorption of water from air (Fig. 215, a
broad IR band at 3500-3250 cm1). Adsorbed water partially remains after pre-heating. This effect
depends also on the structure of the silica surface hydroxyl layer. This appears in variations in
intensity of 1H MAS and 29Si CP/MAS (Q3 – SiOH, Q2 – Si(OH)2) NMR spectra of unmodified
silicas (Figs. 213 and 214), and also in the values of zeta potential (Table 40, ), which becomes less
negative due to surface hydrophobization upon substitution of silanols, responsible for the surface
charging vs. pH, e.g. negative charging at pH > pH at point of zero charge, by nonpolar
functionalities. For example, samples A-300 (hydrophilic) and A-300/PMS5 (hydrophobic) are
characterized by the potential of 4.2 mV and 0.9 mV, respectively. However, there is no a linear
dependence of the potential on the CC value as a certain measure of the hydrophobicity of modified
surfaces (see Table 40) due to the structural features of a modifier layer depending on a type of
PDMS and reaction conditions, as well structural features of silicas studied [345].
Thus, features in the organization of NPNP in ANPNP (Figs. 209-211) and changes in the
porosity and specific surface area in the ranges of mesopores and macropores (Table 40, Fig. 212), as
well changes in the concentrations of surface silanols (single SiOH and twin =Si(OH)2) can affect
the hydrophobization of NAS by various silicones [345]. This appears first in the different content of
the attached hydrophobic functionalities (Table 40, CC and CH). Second, the length of PDMS plays a
very important role in the degree of hydrophobization (Table 40). As a whole, the effects of PDMS
length on the degree of hydrophobization can be explained by several factors such as: (i) longer
molecules are characterized by lower molecular mobility due to stronger molecule-molecule or
molecule-nanoparticle interactions (the value increases for a polymer alone); (ii) penetration of
longer molecules into narrower pores is stronger restricted because longer linear molecules tend to
form clews; (iii) longer molecules generate greater steric barriers for interaction of DMC with a
silica surface and neighboring PDMS; (iv) surface-attached longer PDMS fragments produce greater
negative effects (umbrella screening) on the possibility of other fragments becoming attached to
neighboring active surface sites.
232
The barriers generated by longer PDMS molecules or their fragments are well seen in the
incomplete substitution of surface silanols (the values decreased by 14-50% in comparison to the
reaction of silica with short PMS5 or HMDS) [345]. In addition, these factors result in a diminution
of intensity of the D2 (Si(CH3)2) line in the 29Si CP/MAS NMR spectra, and also the incomplete
disappearance of the Q2 (Si(OH)2) line at 91 ppm (Fig. 213) or 1H MAS NMR at H = 4-5 ppm
(related to silanols and bound water) (Fig. 214). For AP1 and AEP1, the intensity of the D2 line is
maximal and the Q2 line at 91 ppm is not observed (Fig. 213), as well the line of silanols/water at
H = 4-5 ppm (Fig. 214). The IR and NMR spectral features correspond to complete substitution of
surface silanols by hydrophobic functionalities generated by PMS5 reacted with a silica surface with
no DMC (but it is optimal at slightly higher temperatures). These results are close to the ones
obtained upon modification of cA-300 by HMDS (Figs. 213 and 214) at lower temperature.
However, the total weight of the attached trimethylsilyl (TMS) groups (upon reaction of HMDS with
cA-300) is much smaller than that after the reaction of PMS5 with A-300 (Table 40, CC+CH).
Therefore, from a practical point of view, PMS5 could be preferable to HMDS for nanosilica
hydrophobization in spite of the higher temperatures of the reaction [345].
Fig. 213. 29Si CP/MAS NMR spectra of unmodified and modified silicas based on (a) A-300, (b) cA-300, (c) A300 (AE),
and (d) TS 100 and Syloid 244 (lines Qn correspond to Si(OH)4-n(OSi)n at n = 2, 3, and 4; line D2 corresponds
to (-O)2Si(CH3)2; line T3 corresponds to OSi(CH3)3).
233
Fig. 214. 1H MAS NMR spectra of unmodified and modified silicas based on (a) A-300, (b) cA-300, (c) A300 (AE), and
(d) TS 100 and Syloid 244.
Fig. 215. The IR spectra (in the 3800-1350 cm-1 range) of cA-300 unmodified (preheated at 105 oC and 450 oC) and
modified by P3D (cAP3D) and HMDS (cAH).
There is a general tendency for an increasing value (corresponding to better
hydrophobization [345]) with an increase in the reaction temperature from 100 °C to 250 °C.
However, this effect could be lessened with increasing length of PDMS, e.g., for AP2D, = 0.67 and
0.86, and for AP5D, = 0.60 and 0.63 at the reaction temperature of 100 °C and 200 °C,
respectively.
Thus, various nanostructured silicas (fumed silicas such as initial (A-300, A300) and hydro-
compacted (cA-300, TS 100), and precipitated silica Syloid 244) were modified by different PDMS
using DMC as an initiator of the Si-O bond splitting. HMDS was used to modify cA-300.
Investigation of the materials using microscopy, infrared spectroscopy, thermo-desorption, nitrogen
adsorption-desorption, solid state NMR, SAXS, and zeta-potential methods show that the
morphological, textural, and structural characteristics of modified silicas depend greatly on the type
and content of the modifiers employed and on the organization of nonporous nanoparticles in
secondary structures, as well on the reaction temperature. The results reveal that functionalization
with PMS5 alone (Tr = 250 °C) leads to a higher degree ( > 0.95) of silanol substitution [345]. The
SBET reduction is larger; and silanol groups, that remained at the silica surface after its modification,
are mainly absent as compared with the results for longer PDMS/DMC reactions occurring at 200-
220 °C, giving = 0.50-0.86, and a larger SBET value. For nonporous particles of fumed silica, better
modification by shorter PDMS is not trivial because the material is rather macroporous with only the
234
textural porosity caused by voids between nonporous nanoparticles in soft secondary structures. The
texture of crude NAS studied is favorable for effective surface modification both by short
organosiloxane (e.g. PMS5 giving CC = 8.2 wt.% (A300) or 9.1 wt.% (A-300) at 250 °C) alone or for
longer polymers with the presence of DMC. Silicas TS 100 and Syloid 244 modified by
PMS200/DMC demonstrate similar CC values (~5 wt.%) but smaller than that (5.4 wt.%) for
hydrocompacted cA-300 modified by PMS200/DMC. Overall, the PDMS/DMC modified
nanostructured silicas could be of interest from a practical point of view, as they remain in a
dispersed state with no strong compaction of the secondary structures after the functionalization that
is appropriate for better distribution of the modified nanoparticles in various polymer matrices or
other nonpolar media [345].
Interactions of polymethylsiloxane (PMS) with nanosilica was studied under forced
hydrocompaction to elucidate changes in confined space effects upon mechanical treatments [344].
PMS hydrogel, synthesized using methyltrichlorosilane as a precursor, at ~7-8 wt.% of PMS and 93-
92 wt.% of water (Enterosgel, Kreoma-Pharm, Ukraine) was used as the initial material. After drying
at room temperature for a week, the amount of water bound in PMS was small (0.7 wt.%). Dried and
stirred PMS was rehydrated (h = 1 g/g) and stirred again. The bulk density of hydro-compacted
wetted PMS powder is b 0.5 g/cm3 at Vem 1.5 cm3/g; i.e., it remains as a disperse material, as
well the blends with nanosilica [344]. Dry PMS and dry fumed silica (nanosilica) A-300 were mixed
in a porcelain mortar, and then distilled water (h = 1 g/g) was added. If the system was stirred
without any strong mechanical loading (simple mixing) that b 0.5 g/cm3 (PMS/A-300). If the
system was stirred under strong mechanical loading (careful grinding in a porcelain mortar with
strong hand-loading giving ~20 atm, estimated from the geometry of the mortar and a pestle used
and a loading weight, for 15 min) that b 0.6 g/cm3. This is a hydro-compacted sample labeled as
cPMS/A-300 [344].
The IR spectrum of cPMS/A-300 (Fig. 216a, a broad band of the O-H stretching vibrations at
3700-2700 cm1 and a band of the bending vibrations of water molecules at 1630 cm1) shows that
water is strongly bound in the cPMS/A-300 blend (dried at room temperature) despite the presence
of a hydrophobic component (PMS). In contrast to cPMS/A-300, hydrophobic PMS, as well PDMS,
is practically free of water (Fig. 216a, curves 2 and 3, 1630 cm1 is absent for both samples, as well a
band at 3700-2700 cm1 for PDMS). A band at 3700-3200 cm1 in the PMS spectrum (Fig. 216a,
curve 2) is due to silanols, both residual and formed during grinding of the dried PMS sample. The
bending vibrations at 1630 cm1 characteristic for water molecules are practically absent for PMS
similarly to PDMS in contrast to cPMS/A-300 (Fig. 216a) [344]. Note that a band of free silanols of
nanosilica (as well residual SiOH in PMS) is observed as a shoulder at 3740 cm1. Silanols disturbed
due to interactions with adsorbed water molecules or adjacent nanoparticles contribute the broad
band at 3700-2700 cm1 also characteristic for water. PMS has intensive bands of the CH (Fig.
216a, at 2972 and 2916 cm1), SiC (Fig. 216b, asymmetrical at 1272 cm1 and symmetrical at 780
cm1), and SiO (asymmetrical at 1128 cm1 and symmetrical at 1032 cm1) stretching vibrations. In
contrast to PDMS (linear polymer), intensity of two main asymmetrical and symmetrical bands of
the SiO stretching vibrations in PMS (cross-linked 2D-3D polymer) differs, as well the bands of the
SiC stretching vibrations because the amount of CH3 groups is twice larger in PDMS than in PMS.
The additional cross-linking bonds in PMS change the ratio of the mentioned bands in comparison to
that for linear PDMS.
235
Fig. 216. Infrared spectra of air-dried cPMS/A-300, PMS, and PDMS alone in the range of (a) 4000-1300 cm1 (pure
samples) and (b) 1800-300 cm1 (samples with KBr at 1 : 400 w/w ratio).
If all intact water is desorbed from the initial PMS hydrogel upon heating at T < 200 oC (Fig.
217, curve 1) that the amount of this water is ca. 92.7 wt.%. Note that this value corresponds to firm
information on Enterosgel. The DTG minimum at 107.4 oC (curve 4) corresponds to 51.2% of
removed water. Thus, 48.8% of water (evaporated at T > 107.4 oC) could be assigned to bound water,
which really interacts with PMS. The main portion of water in Enterosgel corresponds to weakly
bound water (WBW). In the case of rehydrated dried PMS [344] at a much smaller amount of water
(h = 1 g/g), all water is bound, and 72% of this water is weakly associated water (WAW). Thus, PMS
rather weakly interacts with water due to the abundance of hydrophobic CH3 groups and a very small
contribution of nanopores (Table 41, Snano, Vnano).
Fig. 217. (a) Thermogravimetry (TG) curves (1-3), differential TG (DTG, curves 4-6) and (b) differential thermal
analysis (DTA, curves 1-3) of initial PMS hydrogel (Enterosgel) (a1, a4, b1), PMS dried at room temperature
(a2, a5, b2) and hydro-compacted wetted PMS (a3, a6, b3).
236
Fig. 218. TEM images of degassed (a) initial PMS (Enterosgel); (b) PMS dried at room temperature and stirred; (c)
PMS/A-300 (1:1); and (d) cPMS/A-300 (1:1).
Desorption of water from the PMS hydrogel is an endothermic process, and a DTA minimum
is at 123.6 oC (Fig. 217b, curve 1). For dried PMS (Fig. 218b), the residual amount of water is small
as 0.7 wt.% (Fig. 217a, curve 2, T = 200 oC), and it is hydrophobic as it remains at the top of water
[344]. For wetted (h = 1 g/g), stirred, and hydro-compacted PMS (which could be considered as
rather non-hydrophobic since it forms a suspension), the amount of water is ca. 47.5 wt.% (curve 3)
that is slightly smaller than the added amount of water due to its certain loss during sample
preparation for the measurements [344].
The second weight loss for PMS at T > 400 oC (Fig. 217a) is due to oxidizing of CH3 groups.
Note that the main DTA maximum related to this exothermic (oxidizing) peak for hydro-compacted
PMS shifts toward higher temperatures (Fig. 217b, curve 3, ~521 oC) than that for hydrogel (Fig.
217b, curve 1, ~502 oC) or dried PMS (Fig. 217b, curve 2, ~516.5 oC). However, for the latter, there
is a DTA shoulder at 590-627 oC. Thus, both drying of the hydrogel and hydro-compaction of PMS
affect not only the amounts of water, intact and differently bound to PMS, but also the temperature
course of the desorption of water (both intact and formed upon condensation of silanols) and
oxidizing of the CH3 groups. These effects could be caused by changes in the morphology and
texture of the materials, e.g. in the organization of nanoparticles in their aggregates (Figs. 218-221),
with decomposition of a part of interparticle bonds and reduced sizes of interparticle voids. The
reaction rate of oxidation kinetics of CH3 functionalities in pores depends on the PSD because there
are flows with oxygen (air) into and reaction products out of pores. Additionally, the samples were
heated at a heating rate of 10 oC/min that also affect reaction features in pores. Thus, the compaction
of the secondary structures can affect the gas flows and, therefore, the reaction rate of oxidation (or
associative desorption of water with condensation of hydroxyls) [344].
237
Fig. 219. Nanoparticle (i.e. with no secondary structures) size distributions (PaSD, curves 1-4) of calculated from TEM
images shown in Fig. 218 (using ImageJ, version 1.52w with granulometry plugin,
https://imagej.nih.gov/ij/download.html, accessed on November 13, 2018.), and (curve 5) PaSD for aqueous
suspension of Enterosgel (0.0078 wt.%, pH 3.1) measured using photon correlation spectroscopy (Zetasizer
3000, Malvern Ins., = 633 nm, scattering angle 90o).
Microscopic images (TEM, Fig. 218, and SEM, Fig. 219) show that PMS nanoparticles are
mainly of 2-4 nm in radius (Fig. 220) form secondary structures of various sizes. These small sizes
of PMS nanoparticles correspond to relatively high value of SBET (Table 41). The secondary particles
are responsible for the textural porosity of the PMS, A-300, and PMS/A-300 powders (Fig. 221,
Table 41). The pores determined from the nitrogen adsorption-desorption isotherms are mainly in the
range of mesopores (Fig. 221). However, the PSD correspond only to a part of pores because
nitrogen can only partially fill broad macropores. Therefore, the PSD (Fig. 221, Table 41) cannot
give a complete picture on textural features in the range of macropores. Additionally, practically
always there is an inequality Vem > Vp for highly disperse materials such as fumed oxides or dried
PMS that is caused by a larger contribution of macropores to Vem than to Vp. However, features in the
organization of bound water (or other liquids) depend more strongly on the textural characteristics of
nanopores (R < 1 nm) and mesopores (1 nm < R < 25 nm) than on macropores. In the latter, the water
characteristics are similar to those of bulk water.
Table 41. Textural characteristics of PMS alone and PMS/A-300 [344].
Sample SBET
(m2/g)
SDFT
(m2/g)
Snano
(m2/g)
Smeso
(m2/g)
Smacro
(m2/g)
Vp
(cm3/g)
Vnano
(cm3/g)
Vmeso
(cm3/g)
Vmacro
(cm3/g)
<RV>
(nm)
<RS>
(nm)
PMS 507 471 2 504 1 1.320 0.002 1.304 0.014 6.08 5.28
Stirred PMS 572 581 1 558 13 2.604 0.001 2.248 0.355 16.86 9.42
PMS/A-300 354 322 35 306 13 1.265 0.019 1.084 0.163 15.25 7.64
cPMS/A-300 407 357 8 399 1 1.021 0.006 1.005 0.011 6.56 5.17
A-300 294 289 44 229 16 0.850 0.023 0.567 0.259 20.41 6.14
Note. The values of Vnano and Snano, Vmeso and Smeso, and Vmacro and Smacro were calculated by integration of the fV(R) and
fS(R) functions at 0.35 nm < R < 1 nm, 1 nm < R < 25 nm, and 25 nm < R < 100 nm, respectively. The values of <RV>
and <RS> as the average pore radii were calculated as a ratio of the first moment of fV(R) or fS(R) to the zero moment
(integration over the 0.35-100 nm range) <R> = f(R)RdR/f(R)dR.
238
Fig. 220. SEM images (giving secondary structures with aggregates and agglomerates) of degassed (a) Enterosgel; (b)
PMS dried at room temperature; (c) PMS/A-300 (1:1); and (d) cPMS/A-300 (1:1) (scale bar 500 nm).
Dried (and degassed) PMS hydrogel (with no stirring) possesses a smaller (by 12.8%) surface
area than dried, stirred, and degassed PMS (Table 41, SBET) due to certain decomposition of the
secondary particles in the latter. The drying/stirring treatment results in an increased pore volume of
PMS (Table 41, Vp), however, the Vem value decreases because the value of b increases. There are
also changes in the PSD (Fig. 221) and other textural characteristics (Table 41, Snano, Vnano, Smeso,
Vmeso, Smacro, Vmacro, <RV> and <RS>) of PMS, A-300, and PMS/A-300 differently treated [344].
These results can be explained by decomposition of the bonds (both chemical and physical cross-
linkages) between adjacent polymer and silica nanoparticles, reorganization of their aggregates and
agglomerates. This is accompanied by changes in voids between nanoparticles, and contribution of
mesopores can increase, but contribution of macropores can decrease. The character of these changes
for treated PMS differs from that of the PMS/A-300 blend or nanosilica alone (Figs. 216-221, Table
41) because rigid silica nanoparticles can provide abrasive effects on the soft polymer structures.
Thus, for wetted PMS/A-300 stirred without and with strong mechanical loading and then dried, the
textural and morphological characteristics differ due to the reorganization of the secondary particles
formed by nanoparticles with practically the same sizes and structure [344].
The value of SBET increases (similar to PMS alone) only by 15% (Table 41) upon strong
stirring, but the value of Vp decreases in contrast to PMS alone due stronger compaction of the
secondary structures. Therefore, the average value of <R> (Table 41) decreases upon hydro-
compaction of cPMS/A-300, but it increases during stirring of PMS alone. Note that the maximal
<RV> value is observed for the initial loose A-300 powder characterized by minimal b = 0.05 g/cm3.
Thus, A-300 addition to PMS provides enhanced compaction of macro/meso-structures and
additional decomposition of the secondary particles both of PMS and A-300. This is due to the
abrasive effects (appearing upon mechanical treatments of the slurry) of A-300 on PMS
nanoparticles and their secondary structures, and vice versa due to the polymer effects on the A-300
aggregates and agglomerates. Therefore, nanopores and narrow mesopores practically disappear in
the composite (Fig. 221, curves 4 and 5 at R < 2 nm, Table 41, S and V). These pores could be
assigned to nanosilica aggregates because the PSD of weakly treated PMS/A-300 (Fig. 221, curve 4)
and A-300 (curve 5) are similar at R < 2 nm. During strong stirring of wetted PMS/A-300 (h = 1 g/g)
239
with subsequent drying, the nanosilica aggregates were strongly destroyed (Fig. 218), and A-300
nanoparticles were covered by nonuniform shells with smaller PMS nanoparticles. The changes in
the organization of PMS/A-300 aggregates and agglomerates upon weak and strong stirring of the
slurry are well seen in SEM images (Fig. 220c,d) since for the latter, they are much denser. Note that
the systems were dehydrated and degassed before the nitrogen adsorption, TEM and SEM studies.
These treatments can affect the organization of the nanoparticles in the aggregates and agglomerates
of aggregates. However, low-temperature 1H NMR spectroscopy and NMR cryoporometry allow one
to study hydrated PMS and PMS/A-300 for a deeper insight into the structure of the secondary
particles and the behavior of bound water. This gives information on the organization of hydrated
nanoparticles depending on pretreatment conditions and dispersion media [344].
Fig. 221. (a) Differential (dV/dR) NLDFT PSD and (b) incremental SCV/SCR PSD (V(Ri) = (fV(Ri+1) + fV(Ri))(Ri+1
Ri)/2 at V(Ri) = Vp) for A-300 (curve 1), dried PMS hydrogel (2), dried, rehydrated, hydro-compacted, and
dried PMS (3), stirred wetted PMS/A-300 without (curve 4) and with (curve 5) strong mechanical loading (the
difference in the NLDFT and SCV/SCR PSD shapes is due to the difference in the PSD and IPSD and a smaller
value of the regularization parameter upon the NLDFT calculations).
To avoid the effects of different amounts of water in various hydrated systems studied using
low-temperature 1H NMR spectroscopy and cryoporometry, all of them were investigated at a
constant hydration degree at 1 g of water per gram of dry solid matters [344].
Water bound to dried, rehydrated, and stirred PMS (h = 1 g/g) located in air gives 1H NMR
signal at the chemical shift of the proton resonance at H = 4.0-6.5 ppm (similar to that of bulk water)
dependent on temperature (Fig. 222a). Its intensity decreases with decreasing temperature due to
partial freezing of bound water (Fig. 222) [344]. The H value increases with lowering temperature
due to several effects such as (i) ordering water structure with decreasing contribution of interstitial
water; (ii) decreasing mobility of the molecules tending to spatial positions corresponding to them in
ice (for ice Ih H 7 ppm); and (iii) decreasing vibrational and rotational mobility of the bonds and
functional groups [344]. A relatively large width of signals is due to the presence of various water
clusters in interparticle voids of different sizes and decreased molecular mobility of bound water.
The 1H NMR spectra of water bound in the PMS hydrogel [344], similar to those of dried and
rehydrated PMS (Fig. 222a, h = 1 g/g), but slightly narrower, show that a significant fraction of
water is unfrozen at T < 273 K. However, it is practically all frozen at T 257 K in contrast to that in
hydrated PMS at a smaller h value because the relative content of strongly bound water (SBW)
frozen at T < 260 K increases with decreasing total amount water in the system [344].
240
Fig. 222. 1H NMR spectra recorded at different temperatures of hydrated (h = 1 g/g) samples of dried, rehydrated, stirred
PMS located in (a) air and (b, c) chloroform dispersion medium and (d) with addition of TFAA
(6CDCl3+1TFAA); (c) a part of the spectra (b) at low temperatures shown SBW.
If the hydrated PMS is located in the chloroform-d medium (Fig. 222b,d) that signals become
narrower. This, as well changes in the temperature behavior of bound water, is due to changes in the
water organization. At low temperatures (Fig. 222d), several signals are observed. They can be
attributed to structures with different associativity of the water molecules in various clusters located
in different voids (pores). Note that frozen water (ice) structures can effectively change the pore
(void) size distribution that can also affect the behavior of remained unfrozen water located in these
narrower pores. Water with H = 4.5-5.0 ppm (Fig. 222d, signal 1) can be assigned to strongly
associated water (SAW) similar to bulk water with the molecules having 3-4 hydrogen bonds per a
molecule. Water with H = 1-2 ppm (Fig. 222d, signal 3) could be considered as weakly associated
water (WAW) forming 1D or 2D clusters of LDW (with smaller average numbers of the hydrogen
bonds per a molecule than that in bulk water) with hydrophobic surroundings affecting the
organization of water clusters. Water at H 3 ppm (Fig. 222d, signal 2) is intermediate between
SAW and WAW with respect to the sizes and shapes of the clusters [53,344].
Addition of trifluoroacetic acid (TFAA) to chloroform leads to certain broadening of signals
and to a downfield shift of signals of unfrozen water. Intensity of signal demonstrates a smaller
decrease with lowering temperature due to dissolution of the acid in bound water and the colligative
properties of the solution (Fig. 222c) that enhances the SBW amount. As a whole, interfacial bound
water is rather a poor solvent even for TFAA. Therefore, only a weak shoulder is observed at high H
values (8-10 ppm) corresponding to the acidic solution, and there is a signal (at 280 K) of water with
no dissolved acid [344].
241
Fig. 223. 1H NMR spectra recorded at different temperatures of hydrated (h = 1 g/g) PMS/A-300 in air (solid lines) and
chloroform (dotted-dashed lines) stirred without (a, b) and with (c, d) strong mechanical loading; (c, d)
correspond to a part of the spectra recoded at low temperatures.
For hydrated PMS/A-300 (Fig. 223a,b, solid lines) and cPMS/A-300 (Fig. 223c,d, solid lines)
located in air, the spectra shapes are similar to those of hydrated PMS (Fig. 222a). In the chloroform
dispersion media, the spectra shapes depend on the mechanical treatment of the samples (Fig. 223,
dotted-dashed lines). Upon weak mechanical loading, signals of SAW and WAW are observed (Fig.
223a,b). WAW signal is greater in the total temperature range. For the blend after strong mechanical
loading, practically only SAW signal is observed (Fig. 223c,d). These results suggest that the
organization of bound water depends on the textural and morphological rearrangement of the
secondary structures of hydrated PMS/A-300 (h = 1 g/g) under the strong mechanical loading. For
example, contribution of nanopores and narrow mesopores decreases (Table 41, Fig. 221). Therefore,
chloroform can displace water only into broad mesopores and macropores, in which, all bound water
corresponds to SAW only [344]. Besides SAW and WAW, it is possible to differentiate unfrozen
(at T < 273 K) water into strongly (SBW) and weakly (WBW) bound waters [53]. It is assumed that
SBW corresponds to a fraction frozen at T < 260 K (changes in Gibbs free energy G < 0.5 kJ/mol)
and WBW is frozen at 260 K < T < 273 K (G > 0.5 kJ/mol) (Fig. 224a). This assumption is based
on the difference in the temperature behavior of unfrozen SBW and WBW. As a whole, freezing of
SBW occurs in broad temperature and G value ranges, but in a narrow Cuw value range in contrast
to WBW. This is due to a fast decrease of the effects of solid surface fields on distant interfacial
water layers. This decay of the surface fields is stronger for hydrophobic (i.e. nonpolar or weakly
polar) surfaces than for polar ones due to different dependences of the electrostatic fields (caused by
charges and dipoles) and van der Waals forces on distance. However, there is an additional factor
caused by the confined space because the opposite walls in pores can enhance the influence of the
surface fields on the temperature behavior of bound water. Additionally, the hydrophobic media (e.g.
chloroform) can change the organization of bound water. Water tends to locate in narrow nanopores
(inaccessible for larger chloroform molecules) or in broad mesopores and macropores to reduce the
surface area of contacts between immiscible liquids. Both in PMS and PMS/A-300, nanopores at R <
1 nm are practically absent (Fig. 221). Therefore, water is mainly displaced by chloroform into larger
242
mesopores. SBW transforms into WBW (mainly SAW) for all systems (Fig. 224a and Table 42,
compare the systems located in air and chloroform), but with one exception with the
chloroform/TFAA medium.
Fig. 224. (a) Relationships between the amounts of unfrozen water (Cuw) and changes in the Gibbs free energy (G)
depending on temperature; (b, c) size distributions of pores filled by unfrozen water (PSDuw) for PMS alone in
(1) air, (2) chloroform and (3) 6CDCl3+1TFAA; and PMS/A-300 stirred without (4-6) and with (7, 8) strong
mechanical loading in air (4, 7) and chloroform (5, 6, 8); (b) differential PSDuw and (c) incremental IPSDuw.
The latter is due to the colligative properties of the acidic solution that result in the freezing
temperature depression. Note that a fraction of water located in macropores (R > 25 nm) could be
assigned to non-bound water with the properties similar to those of bulk water. Therefore, NMR
cryoporometry with water as a probe cannot appropriately describe total contribution of macropores
(Table 42, Fig. 224b,c). The effects of chloroform on bound water are well seen in a decrease in the
S values giving the sum of total changes in the modulus of G for all interfacial water (due to
diminution of interaction of water with solid surfaces) and an increase in the average melting
temperature <Tm> (due to transformation SBW WBW) (Table 42). This is due to a decrease in the
confined space effects on bound water displaced into larger pores.
There is a certain difference in contribution of nanopores to the total porosity and specific
surface area calculated on the basis of nitrogen adsorption data (Table 41, Fig. 221) and NMR
cryoporometry (Table 42, Fig. 224b,c). This difference could be explained by the effects of frozen
water (ice) located in mesopores that can result in transformation of mesopores into nanopores upon
the use of the NMR cryoporometry. Additionally, dried (nitrogen adsorption) and wetted (NMR
study) PMS samples could have different structures of nanoparticles, as well different secondary
structures, due to a certain swelling effect, and the amount of water is smaller than Vem in the
powders.
243
Table 42. Characteristics of water bound to non-compacted and compacted PMS alone and with A-300 (1 : 1) in air,
chloroform medium alone or with addition of TFAA (h = 1.0 g/g) [344].
Sample Medium
Сuw
s
(mg/g)
Сuw
w
(mg/g)
ΔGs
(kJ/mol)
γS
(J/g)
<Tm>
(K)
Snano,uw
(m2/g)
Smeso,uw
(m2/g)
Vnano,uw
(cm3/g)
Vmeso,uw
(cm3/g)
PMS Air 280 720 2.48 25.13 258.63 138 98 0.056 0.758
PMS CDCl3 10 990 2.44 0.69 262.36 3 1 0.001 0.007
PMS CDCl3/TFAA 475 525 2.45 27.73 257.39 15 226 0.006 0.796
PMS/A-300 Air 115 885 2.85 12.06 257.04 64 86 0.026 0.297
PMS/A-300
(SAW)
CDCl3 95 25 3.04 2.88 251.26 20 14 0.008 0.041
PMS/A-300
(WAW)
CDCl3 795 95 2.77 9.91 257.88 52 54 0.021 0.202
cPMS/A-300 Air 225 775 2.68 17.45 261.78 110 70 0.044 0.675
cPMS/A-300 CDCl3 30 970 2.90 5.01 263.69 31 20 0.012 0.207
Note. Cuw
s and Cuw
w are the amounts of weakly and strongly bound waters; ΔGs is the changes in the Gibbs free energy of water layer
closely located to a surface; γS is the modulus of the total changes in the Gibbs energy of bound water unfrozen at T < 273.15 K; <Tm>
is the average melting temperature; Snano,uw and Vnano,uw, Smeso,uw and Vmeso,uw are the specific surface area and pore volume of nanopores
at R < 1 nm and mesopores at 1 nm < R < 25 nm, respectively, in contact with unfrozen water.
The PSDuw curves (Fig. 224b,c) well show the effects of chloroform (i.e., the displacement of
water into larger pores) and hydro-compaction (i.e., the diminution of a contribution of narrow pores
and an increase in contribution of broader mesopores) because both ones result in the shifts of the
PSDuw toward larger pores. These effects appear in changes in the H values and entropy of bound
water vs. temperature (Fig. 225).
The secondary structures of PMS and PMS/A-300, treatment features, and a medium type
affect the temperature dependence of the H values of bound water (Fig. 225a,b). The mentioned
effects influence the changes in the entropy of bound unfrozen (melted ice) water vs. temperature
(Fig. 225c,d). The maximal peak of the s(T) function for PMS alone (Fig. 225c) corresponds to
increased entropy both of SBW and WBW that appears due to melting of ice (with decreasing
structural order) with increasing temperature to 255-265 K. However, for intermediate clusters of
water (Fig. 222d, signal 2) bound to PMS located in chloroform, this peak becomes larger (Fig. 225c,
curve 3) than that in the air medium (curve 1) or for SAW (curve 2, signal 1 in Fig. 222d). For the
chloroform/TFAA medium (curve 4), the s(T) curve strongly differs from others due to the
colligative properties of the acidic solution with strongly shifted freezing/melting points toward
lower temperatures [53,344].
For both PMS/A-300 samples located in air (Fig. 225d, curves 1 and 4), the s(T) peaks at
263.5 K slightly shift toward higher temperature in comparison to that for PMS alone (Fig. 225c).
This is rather unexpected result because addition of hydrophilic nanosilica results in the effect, which
could be expected upon addition of a hydrophobic component. However, the appearance of the
hydrophobic surroundings (chloroform) results in the shift of the s(T) maximum toward lower
temperature both for SAW (Fig. 225d, curve 2) and WAW (curve 3). For compacted cPMS/A-300
(Fig. 225d, curve 5), a strong decrease in the s(T) values is observed at temperatures corresponding
to melting of WBW (which is mainly SAW, Fig. 223c,d). This effect could be explained by
diminution of the porosity of the compacted blend that leads to increased effects of hydrophobic
functionalities on unfrozen water, which, therefore, becomes more ordered [53].
Interaction energy (Et) in the system with dehydrated PMS particle (119 structural units
with residual 9 OH groups) surrounded by a water shell clustered is Et = 19.1 kJ/mol per a water
molecule [344]. The interaction between the water shell and PMS particle (per a water molecule)
gives Et = 7.1 kJ/mol, but in the water shell per se, it is stronger Et = 12.0 kJ/mol. This result
corresponds to a tendency of the formation of larger water structures (clusters, domains) upon
interaction with hydrophobic surroundings. However, for the water shell around a non-dehydrated
PMS particle (119 structural units with 60 OH groups), the 1H NMR spectrum is similar to that for
244
the shell around the dehydrated PMS particle (with residual 9 OH groups). This effect is due to the
clusterization of the water shells for both particles [344].
Fig. 225. Chemical shifts H of unfrozen water vs. temperature in the 1H NMR spectra of (a) PMS and (b) PMS/A-300
differently treated and in different media (air and chloroform or CDCl3/TFAA), and (c, d) corresponding
functions s(T) = T((ln(T))/T)P vs. T.
Thus, the powders of dried disperse hydrophobic PMS alone and in the blends with hydrophilic
nanosilica can be easily rehydrated upon stirring with water [344]. The properties of the slurry
depend on the mechanical treatment due to stronger compaction of the PMS/A-300 secondary
structures with increasing mechanical loading. Note that a similar behavior of the blends with
hydrophobic and hydrophilic nanostructured materials was observed for several compositions that
reflects the general regularities appearing at appropriate amounts of added water and certain
mechanical loading onto the blends of hydrophilic and hydrophobic nanostructured materials. These
conditions result in the reorganization of the secondary structures and removal of micro-scaled air
bubbles bound to hydrophobic components in the initial mixtures. Note that after drying of the
slurry, the powder can demonstrate the hydrophobic properties since it can remain at the top of liquid
water [344].
A fraction of water (h = 1 g/g) bound to PMS or PMS/A-300 located in air corresponds to
strongly bound water frozen at T < 260 K, but practically all water is strongly associated and
characterized by the chemical shift at H = 4.5-5.0 ppm. If these systems are located in the
hydrophobic chloroform dispersion medium that the organization of bound water strongly changes to
reduce the contact area between immiscible liquids. Typically, chloroform can displace adsorbed
water into narrow nanopores (inaccessible for larger chloroform molecules) or/and into larger pores
(to reduce the contact area between them). In the system studied, narrow nanopores are practically
absent; therefore, chloroform displaces water mainly into larger pores. This results in a decrease in
the amounts of SBW. A fraction of weakly associated water (H = 1-2 ppm), i.e., strongly clustered
water with 1D and 2D structures, appears in the systems undergoing low mechanical loading. After
245
stronger mechanical loading, only strongly associated water is observed due to changes in the
confined space effects in more strongly compacted secondary particles [344]. If trifluoroacetic acid is
added to the chloroform medium that the colligative and confined space effects overlap and the
amount of SBW increases. Note that in this case, the changes in the porosity caused by the
appearance of the ice clusters in interparticle voids can be smaller than that in the systems with a
main fraction of SAW/WBW [344]. The PMS/nanosilica blends can be of interest from a practical
point of view for applications as medical sorbents. The blends are with improved or better controlled
textural and adsorption characteristics, which can be easily varied in a broader range due to changes
in composition and pretreatment conditions. Note that tests of these systems in various applications
were started, and some promised results were obtained [344].
Conclusion
Complex FMO such as silica/alumina, silica/titania, and alumina/silica/titania could consist of
core-sell nanoparticles including crystalline or polycrystalline cores with an amorphous shell. In
contrast to simple and small silica or titania nanoparticles in individual FMO, complex core-shell
nanoparticles (50-200 nm in size) with titania or alumina cores and silica or alumina shells in binary
and ternary FMO could be destroyed under mechanochemical activation in a ball mill or a
microbreaker. Both can affect the structure of aggregates of nanoparticles and agglomerates of
aggregates, resulting in more compaction. This could be accompanied by changes in color from
white to beige of different tints, with resulting changes in the UV-vis spectra in the 300-600 nm, as
well as changes in the crystalline structure of alumina. This study showed that complex nanooxides
could be more sensitive to external actions than simple nanooxides such as fumed silica. This aspect
is of importance for practical applications of FMO.
The results of the hydrocompaction of nanosilica A-300 depend strongly on the amounts of
water (in the range of 0.3-5.0 g/g) used in the process. A selection of a certain amount of water
allows the appropriate reorganization of the secondary and ternary structures of nanoparticles that is
accompanied by changes in the textural characteristics (S, V, PSD) of the powder. However,
according to TEM and SAXS data, the nanoparticles per se are practically not affected by this
treatment, i.e., only the rearrangement of their secondary formations occurs. At a low amount of
water (hcp 1 g/g), the reorganization of secondary/ternary structures does not lead to reduction of
the specific surface area (it even slightly increases). However, at hcp 1.5 g/g, the specific surface
area decreases, but the pore volume (nitrogen adsorption at p/p0 0.99) increases despite the empty
volume of the powder decreases from 21.8 cm3/g for initial A-300 (b = 0.045 g/cm3) to 3.45 cm3/g
after compaction at hcp = 4.5 g/g (b = 0.256 g/cm3). Note that the structural reorganization of the
hydro-compacted powders is possible due to addition of new amounts of water. This suggests that
the chemical bonds between adjacent NPNP do not practically form upon the hydro-compaction
under soft conditions. Thus, hydro-compacted nanosilica loses the dust-forming property but remains
active with respect to the NPNP activity and mobility in the secondary structures. This is of
importance for practical applications of nanosilicas, especially in aqueous suspensions, drug delivery
systems, and upon filling of polymers.
Any treatment of nanooxides affected the interfacial and temperature behaviors of polar and
nonpolar adsorbates those depend on concentrations, treatment history and other conditions. For
some adsorbates, open hysteresis loops result in adsorption-desorption isotherms. Rearrangement of
secondary FMO particles can strongly affect the freezing-melting point depression of bound
adsorbates such as water and low-molecular weight organics. Clustering of adsorbates in pores
caused reduced enthalpy changes during phase transition (freezing, fusion). Freezing point
depression and increasing melting point caused significant hysteresis effects for some adsorbates,
246
both low- and high-molecular weight compounds, bound to oxide nanoparticles in textural pores as
voids between nanoparticles in secondary structures of FMO.
Confined space effects for various adsorbates bound to initial or treated FMO lead to reduction
of the activity of solvents in the interfacial layers depending on the pore sizes and temperature. This
results in differentiation of aqueous solutions of acids, salts and other compounds in different
structures located in pores of different sizes that are characterized by decreased amounts of solutes in
pores of smaller sizes. This differentiation enhances with decreasing temperature and partial freezing
of adsorbates because of, at least, two effects related to freezing a portion of adsorbates in pores (that
results in changes in the effective porosity and pore sizes) and cryoconcentrating of the solution
because frozen fractions tend to be with much lower content of solutes than that dissolved in the
liquid solvent.
Glossary of abbreviations Units
A constant
a adsorption cm3/g
am BET monolayer adsorption cm3/g
bij collision rate s-1
Cd concentration of decane (g/g)
CSiO2 concentration of silica wt%
Cuw amount of unfrozen water (g) per gram of adsorbent g/g
Cs
uw concentration of unfrozen water strongly bound to surface wt%
Cw
uw concentration of unfrozen water weakly bound to surface wt%
dCuw/dR derivative g/g/nm
dCuw/dT derivative g/g/K
dp diameter of pores nm
dVuw(R)/dR derivative cm3/g/nm
E adsorption energy kJ/mol
Ek kinetics energy kJ/mol
Ep potential energy kJ/mol
Fp intensity of the electrostatic field kV/m
f(k) distribution function of rate constant arb.un.
f(Rp) differential pore size distribution arb.un.
fS(R) distribution function of pore size with respect to surface area arb.un.
fV(R) distribution function of pore size with respect to pore volume arb.un.
h hydration g/g
I intensity of 1H NMR signal -
I0,i intensity of the temperature distribution curve of phase i arb.un.
Ic intensity of 1H NMR signal of water adsorbed from the gas phase -
Iuw intensity of 1H NMR signal of unfrozen water at T < 273 K -
k rate constant (s-1)
kB Boltzmann constant (1.3806488×10−23 J/K)
Kh coefficient of hydrophilicity -
kGT constant in Gibbs-Thomson equation K nm
Lp pore length nm
p equilibrium pressure Pa
p0 saturation pressure Pa
Q heat of adsorption kJ/mol
Qd heat of immersion in n-decane kJ/mol
247
Qw heat of immersion in water kJ/mol
qH charge on H-atom a.u.
R radius of pores nm
Rg gas constant kJ/K/mol
Rp pore radius nm
Rmax maximal pore radius on integration nm
Rmin minimal pore radius on integration nm
SBET specific surface area by the Brunauer-Emmett-Teller method m2/g
SBET,X specific surface area determined from adsorption of compound X m2/g
SIR specific surface area determined using IR spectra m2/g
Smacro specific surface area of macropores m2/g
Smeso specific surface area of mesopores m2/g
Snano specific surface area of nanopores m2/g
Ssum corrected total specific surface area m2/g
Suw specific surface area determined using NMR cryoporometry m2/g
T absolute temperature K
T1 longitudinal relaxation time s
T2 transverse relaxation time s
Tb boiling temperature K
Tc critical temperature K
Tcr crystallization temperature K
Tg glass-transition temperature K
Tm melting temperature K
Tm, bulk melting temperature K
Tm(R) melting temperature of a frozen liquid in pores of radius R K
<Tm> average melting temperature K
t time s
t thickness of an adsorbed nitrogen layer nm
t(p,Rp) statistical thickness of an adsorbed layer nm
tm statistical thickness of a monolayer nm
tMCA time of MCA hour
tT heating time min
Vem empty volume cm3/g
Vp pore volume cm3/g
Vmacro volume of macropores cm3/g
Vmeso volume of mesopores cm3/g
Vnano volume of nanopores cm3/g
Vuw(R) volume of unfrozen water in pores of radius R cm3/g
w parameter in Kelvin equation arb.un.
wef effective parameter in Kelvin equation arb.un.
X normalized inverse transition temperature 1/K
Xci normalized inverse transition temperature of phase i, 1/K
regularization parameter -
heating rate K/s
Gibbs adsorption
surface tension N/m
p surface tension in pores N/m
248
S module of total changes in the Gibbs free energy of interfacial water mJ/m2
EH energy of the hydrogen bonds kJ/mol
G changes in Gibbs free energy of the interfacial water kJ/mol
Gs changes in Gibbs free energy of the strongly bound water kJ/mol
Gw changes in Gibbs free energy of the weakly bound water kJ/mol
Hf bulk enthalpy of fusion kJ/mol
Him immersion enthalpy kJ/mol
Tm melting point depression degr.
changes in the surface tension mJ/m2
w relative deviation from the pore model arb.un.
δ chemical shift ppm
H chemical shift of protons ppm
(T) integrated heat flow in DSC arb.un.
density g/cm3
b bulk density g/cm3
(a) primary particle size distribution arb.un.
reduced (a/am) adsorption -
-potential mV
200DF silica gel;
A-50 fumed silica;
A-100 fumed silica;
A-150 fumed silica;
A-200 fumed silica;
A-300 fumed silica;
A-380 fumed silica;
A-400 fumed silica;
A-500 fumed silica;
AFM Atom Force Microscopy;
APMS AminoPropylMethylSilyl groups;
ASTxx Alumina/Silica/Titania nanooxide; xx = titania content (wt.%)
ASW water in associates HO-H…A (A– electron donor center);
B3LYP Exchange-correlation functional in DFT
BW Bound Water;
Cab-O-Sil HS-5 fumed silica;
CNO CryoNanoOxides;
CSNP Core-Shell NanoParticles;
DEA DiEthylAmine;
DFT Density Functional Theory;
DLS Dynamic Light Scattering;
DMAAB (DiMethylAmino)AzoBenzene;
DMSO DiMethylSulfOxide;
DON DiOxiNaphthalene;
DRS Dielectric Relaxation Spectroscopy;
DRS Diffuse Reflectance Spectra;
DSC Differential Scanning Calorimetry;
DTG Differential TG;
249
EDL Electric Double Layer;
FMO Fumed Metal or Metalloid Oxides;
FT Fourier Transformation;
FTIR Fourier Transform Infrared Spectroscopy;
HMDS HexaMethylDiSilazane;
HOMO Highest Occupied Molecular Orbital;
HPCG High-Pressure CryoGelation;
HRTEM High Resolution Transmission Electron Microscopy;
HTT HydroThermal Treatment;
IEP IsoElectric Point;
IGT Integral Gibbs-Thomson equation;
Ih hexagonal ice;
IPSD Incremental Pore Size Distribution;
LTNA Low-Temperature Nitrogen Adsorption;
MAS Magic Angle Spinning;
MB Methylene Blue
MCA MechanoChemical Activation;
MCM-41 ordered mesoporous silica;
MCM-48 ordered mesoporous silica;
MND Modified Nguyen-Do method;
MS Methylated Silica;
MW Molecular Weight;
NMR Nuclear Magnetic Resonance;
OX-50 fumed silica;
PaSD Particle Size Distributions;
PC-100 photocatalyst with titania;
PC-105 photocatalyst with titania;
PC-500 photocatalyst with titania;
PDMS Poly(DiMethyl Siloxane);
PEG Poly(Ethylene Glycol);
PM6 and PM7 Semiempirical methods
PMS Poly(Methyl Siloxane);
POA Phosphorus OxyAcids;
POE Poly(Oxy Ethylene);
PPA PolyPhosphoric Acid;
PPSD Primary Particle Size Distribution;
PS300 pyrogenic silica analogues of A-300;
PSD Pore Size Distribution;
PSDuw Pore Size Distribution determined from the Cuw values;
PVA Poly(Vinyl Alcohol);
PVP Poly(Vinyl Pyrrolidone);
QC Quantum Chemistry;
Qc quercetin;
SAxx Silica/Alumina nanooxides; xx = alumina content (wt.%)
SAW Strongly Associated Water;
SAXS Small-Angle X-ray Scattering;
SBA-15 ordered mesoporous silica;
SBA Strongly Bound Adsorbate;
250
SBW Strongly Bound Water;
SCR Self-Consistent Regularization;
SCV model of a pore mixture with Slitshaped and Cylindrical pores and
Voids between spherical nanoparticles packed in random aggregates;
SEM Scanning Electron Microscopy;
SI Supplementary Information;
Si-40 silica gel;
Si-60 silica gel;
Si-100 silica gel;
SMD Solvation Model
STxx Silica/Titania nanooxides; xx = titania content (wt.%)
TEA TriEthylAmine;
TEM Transmission Electron Microscopy;
TEOS TetraEthOxySilane; TetraEthyl OrthoSilicate
TG ThermoGravimetry;
TMS TetraMethylSilane;
TMS TriMethylSilyl groups
TPD-MS Temperature-Programmed Desorption with Mass-Spectrometry control;
TSDC Thermally Stimulated Depolarization Current;
UV–Vis Ultra-Violet–Visible (spectroscopy);
WAW Weakly Associated Water;
WBA Weakly Bound Adsorbate;
WBW Weakly Bound Water;
wB97XD DFT functional;
XPS X-ray Photoelectron Spectroscopy;
XRD X-Ray Diffraction
References
[1] Bergna H.E., Roberts W.O. (editors). Colloidal Silica: Fundamentals and Applications. (CRC Press,
Boca Raton, 2006).
[2] Hastie J.W. (editor). Materials Chemistry at High Temperatures. Vol. 1, Characterization. Vol. 2,
Processing and Performance. (NJ: Clifton, Humana Press, 1990).
[3] Ullmann's Encyclopedia of Industrial Chemistry. Wiley-VCH, Weinheim. 2008.
[4] Basic Characteristics of Aerosil. Technical Bulletin Pigments, No 11, Degussa AG, Hanau. 1997.
[5] Shchukin E.D., Zelenev A.S. Physical-Chemical Mechanics of Disperse Systems and Materials. (CRC
Press, Boca Raton, 2015).
[6] Cabot Corporation. http://www.cabotcorp.com/solutions/products-plus/fumed-metal-oxides/.
[7] DuPont. http://www.dupont.com/.
[8] Evonik Ind. http://corporate.evonik.com/en/Pages/default.aspx;
http://www.aerosil.com/product/aerosil/en/services/downloads/Pages/test-methods.aspx.
[9] Kulkarni P., Baron P.A., Willeke K. (editors). Aerosol Measurement: Principles, Techniques, and
Applications. Third Edition. (New York : John Wiley & Sons, 2011).
[10] Büchel K.H., Moretto H.-H., Woditsch P. Industrial inorganic chemistry. (Weinheim: Wiley-VCH
Verlag GmbH, 2000).
[11] Auner N., Weis J. (editors). Oganosilicon Chemistry VI. (Weinheim: Wiley-VCH Verlag GmbH, 2005).
[12] Piemonte V., De Falco M., Basile A. (editors). Sustainable Development in Chemical Engineering –
Innovative Technologies. First Edition. (UK: Chichester , John Wiley & Sons, 2013).
[13] Kriechbaum G.W., Kleinschmit P. Superfine oxide powders- flame hydrolysis and hydrothermal
synthesis. Adv. Mater. 1989.10:330.
251
[14] Gutsch A., Mühlenweg H., Krämer M. Tailor-made nanoparticles via gas-phase synthesis. Small .
2005.1:30.
[15] Jacobsen H., Kleinschmit P. Flame Hydrolysis. In: Ertl G, Knozinger H, Weitkamp J (editors),
Preparation of Solid Catalysts. (Weinheim: Wiley-VCH Verlag GmbH, 1999, pp. 99-109).
[16] Afyon S., Hagemann M., Somer M., Isfort C.S. Thermal and hydrothermal stability of flame
hydrolytically synthesized SiO2/TiO2 mixed oxides. Solid State Sci. 2013.18: 91.
[17] Albers P., Maier M., Reisinger M., Hannebauer B., Weinand R. Physical boundaries within aggregates
– differences between amorphous, para-crystalline, and crystalline structures. Cryst. Res. Technol.
2015.1: 20. DOI: 10.1002/crat.201500040
[18] Schaefer D.W., Hurd A.J. Growth and structure of combustion aerosols: fumed silica. Aerosol Sci.
Technol. 1990. 12(4): 876.
[19] Roth P. Particle synthesis in flames. Proceedings of the Combustion Institute. 2007. 31: 1773.
[20] Pratsinis S.E. Flame aerosol synthesis of ceramic powders. Prog. Energy Combust. Sci. 1998. 24: 197.
[21] Mueller R., Madler L., Pratsinis S.E. Nanoparticle synthesis at high production rates by flame spray
pyrolysis. Chem. Eng. Sci. 2003. 58: 1969.
[22] Camenzind A., Caseri W.R., Pratsinis S.E. Flame-made nanoparticles for nanocomposites. Nano Today.
2010. 5: 48.
[23] Teoh W.Y., Lutz M. Flame spray pyrolysis: An enabling technology for nanoparticles design and
fabrication. Nanoscale. 2010.2: 1324.
[24] Weston J.S., Venkataramani D., Aichele C.P., Grady B.P., Harwell J., Resasco D. Pseudosolid, shear-
thinning gel formation in binary dispersions of metal oxide nanoparticles at low volume fractions.
Langmuir. 2014. 30: 14982.
[25] Gun’ko V.M., Mironyuk I.F., Zarko V.I., Voronin E.F., Turov V.V., Pakhlov E.M., Goncharuk E.V.,
Nychiporuk Yu.M., Kulik T.V., Palyanytsya B.B., Pakhovchishin S.V., Vlasova N.N., Gorbik P.P.,
Mishchuk O.A., Chuiko A.A., Skubiszewska-Zięba J., Janusz W., Turov A.V., Leboda R. Morphology
and surface properties of fumed silicas. J. Colloid. Interface Sci. 2005. 289: 427.
[26] Gun’ko V.M., Nychiporuk Yu.M., Zarko V.I., Goncharuk E.V., Mishchuk O.A., Leboda R.,
Skubiszewska-Zięba J., Skwarek E., Janusz W., Yurchenko G.R., Osovskii V.D., Ptushinskii Y.G.,
Turov V.V., Gorbik P.P., Blitz J.P., Gude K. Relationships between surface compositions and
properties of surfaces of mixed fumed oxides. Appl. Surf. Sci. 2007. 253: 3215.
[27] Gun’ko V.M., Blitz J.P., Gude K., Zarko V.I., Goncharuk E.V., Nychiporuk Yu.M., Leboda R.,
Skubiszewska-Zięba J., Osovskii V.D., Ptushinskii Y.G., Mishchuk O.A., Pakhovchishin S.V., Gorbik
P.P. Surface structure and properties of mixed fumed oxides. J. Colloid. Interface. Sci. 2007. 314: 119.
[28] Gun’ko V.M., Zarko V.I., Turov V.V., Oranska O.I., Goncharuk E.V., Nychiporuk Yu.M., Pakhlov
E.M., Yurchenko G.R., Leboda R., Skubiszewska-Zięba J., Osovskii V.D., Ptushinskii Y.G.,
Derzhypolskyi A.G., Melenevsky D.A., Blitz J.P. Morphological and structural features of individual
and composite nanooxides with alumina, silica, and titania in powders and aqueous suspensions.
Powder Technology. 2009. 195: 245.
[29] Gun’ko V.M., Bogatyrev V.M., Borysenko M.V., Galaburda M.V., Sulim I.Y., Petrus L.V., Korduban
O.M., Polshin E.V., Zaulychnyy Ya.V., Karpets M.V., Foya O.O., Myronyuk I.F., Chelyadyn V.L.,
Dzhura U.Ya., Leboda R., Skubiszewska-Zięba J., Blitz J.P. Morphological, structural and adsorptional
features of oxide composites of different origin. Applied Surface Science. 2010. 256: 5263.
[30] Gun’ko V.M., Yurchenko G.R., Turov V.V., Goncharuk E.V., Zarko V.I., Zabuga A.G., Matkovsky
A.K., Leboda R., Skubiszewska-Zięba J., Janusz W., Phillips G.J., Mikhalovsky S.V. Adsorption of
polar and nonpolar compounds onto complex nanooxides with silica, alumina, and titania. J. Colloid
Interface Sci. 2010. 348: 546.
[31] Park H.K., Park K.Y., Jung K.Y. Alumina-precursor nanoparticles prepared by partial hydrolysis of
AlCl3 vapor in tubular flow reactor: Effect of hydrolysis conditions on particle size distribution. Ind.
Eng. Chem. Res. 2014. 53: 10372.
[32] Dorigato A., Pegoretti A. The role of alumina nanoparticles in epoxy adhesives. J. Nanopart. Res. 2011.
13: 2429.
252
[33] Sabzi M., Mirabedini S.M., Zohuriaan-Mehr J., Atai M. Surface modification of TiO2 nano-particles
with silane coupling agent and investigation of its effect on the properties of polyurethane composite
coating. Prog. Org. Coating. 2009. 65: 222.
[34] Rodríguez-Castellón E., Jiménez-López A., Maireles-Torres P., Jones D.J., Rozière J., Trombetta M., et
al. Textural and structural properties and surface acidity characterization of mesoporous silica-zirconia
molecular sieves. J. Solid. State Chem. 2003. 175: 159.
[35] Liang S., Neisius N.M., Gaan S. Progress in organic coatings recent developments in flame retardant
polymeric coatings. Prog. Org. Coating. 2013. 76: 1642.
[36] Allahverdi A., Ehsani M., Janpour H., Ahmadi Sh. The effect of nanosilica on mechanical, thermal and
morphological properties of epoxy coating. Prog. Org. Coating. 2012. 75: 543.
[37] Iler R.K. The Chemistry of Silica. (Chichester: Wiley, 1979).
[38] Legrand A.P. (editor). The Surface Properties of Silicas. (New York: Wiley, 1998).
[39] Blitz J.P., Gun’ko V.M. (editors). Surface Chemistry in Biomedical and Environmental Science. NATO
Science Series II: Mathematics, Physics and Chemistry. Springer, Vol. 228. (Dordrecht: Springer,
2006).
[40] Chuiko A.A. (editor). Chemistry of Silica Surface. (Kiev: UkrINTEI, 2001). [in Russian].
[41] Nicolais L., Borzacchiello A., Lee S.M. (editors). Wiley Encyclopedia of Composite. Materials, 5-
Volume set, 2nd ed. (NJ: Wiley, Hoboken, 2012).
[42] Dodiuk H., Goodman S. (editors). Handbook of Thermoset Plastics. Third Edition. (UK: Oxford,
Elsevier, 2014).
[43] Wicks Z.W., Jr, Jones F.N., Pappas S.P., Wicks D.A. Organic Coatings: Science and Technology.
Third Edition. (New York: John Wiley & Sons, 2007).
[44] Ghosh S.K. (editor). Functional Coatings. (Weinheim: Wiley-VCH Verlag GmbH, 2006).
[45] Wu L., Baghdachi J. (editors). Functional Polymer Coatings: Principles, Methods, and Applications.
First Edition. (New York: John Wiley & Sons, 2015).
[46] Gun’ko V.M., Turov V.V., Zarko V.I., Pakhlov E.M., Matkovsky A.K., Oranska O.I., Palyanytsya B.B.,
Remez O.S., Nychiporuk Yu.M., Ptushinskii Y.G., Leboda R., Skubiszewska-Zięba J. Cryogelation of
individual and complex nanooxides under different conditions. Colloids Surf. A: Physicochem. Eng.
Aspects. 2014. 456: 261.
[47] Gun’ko V.M., Zarko V.I., Pakhlov E.M., Matkovsky A.K., Remez O.S., Charmas B., Skubiszewska-
Zięba J. Low-temperature high-pressure cryogelation of nanooxides. Journal of Sol-Gel Science and
Technology. 2015. 74: 45.
[48] Gun’ko V.M., Ilkiv V.Ya., Zaulychnyy Ya.V., Zarko V.I., Pakhlov E.M., Karpetz МV. Structural
features of fumed silica and alumina alone, blend powders and fumed binary systems. J. Non-Crystal
Solid. 2014: 403: 30.
[49] Gun’ko V.M., Zaulychnyy Ya.V., Ilkiv BI, Zarko V.I., Nychiporuk Yu.M., Ptushinskii Yu.G., Pakhlov
E.M., Leboda R., Skubiszewska-Zięba J. Textural and electronic characteristics of mechanochemically
activated composites with nanosilica and activated carbon. Appl. Surf. Sci. 2011: 258: 1115.
[50] Ertl G., Knozinger H., Weitkamp J. (editors). Handbook of Heterogeneous Catalysis. (Weinheim: VCH
Verlagsgesellschaft mbH, 1997).
[51] Busca G. Heterogeneous Catalytic Materials: Solid State Chemistry, Surface Chemistry and Catalytic
Behaviour. (Amsterdam: Elsevier, 2014).
[52] Regalbuto J. (editor). Catalyst Preparation: Science and Engineering. (Boca Raton: CRC Press, 2007).
[53] Gun’ko V.M., Turov V.V. Nuclear Magnetic Resonance Studies of Interfacial Phenomena. (Boca
Raton: CRC Press, 2013).
[54] Pietsch W. Agglomeration in Industry. (Weinheim: Wiley-VCH Verlag GmbH, 2005).
[55] Sandkühler P., Lattuada M., Wu H., Sefcik J., Morbidelli M. Further insights into the universality of
colloidal aggregation. Adv. Colloid Interface Sci. 2005. 113: 65.
[56] Schießl K., Babick F., Stintz M. Calculation of double layer interaction between colloidal aggregates.
Advanced Powder Technology. 2012. 23: 139.
[57] Karlsson H.L., Toprak M.S., Fadeel B. Toxicity of metal and metal oxide nanoparticles. In: Nordberg
G.F., Fowler B.A., Nordberg M. (editors). Handbook on the Toxicology of Metals. Fourth Edition.
(Amsterdam: Elsevier, 2015, pp. 75-112).
253
[58] Bolis V., Busco C., Ciarletta M., Distasi C., Erriquez J., Fenoglio I., Livraghi S., Morel S.
Hydrophilic/hydrophobic features of TiO2 nanoparticles as a function of crystal phase, surface area and
coating, in relation to their potential toxicity in peripheral nervous system. J. Colloid Interface Sci.
2012. 369: 28.
[59] Theodore L., Kunz R.G. Nanotechnology: Environmental Implications and Solutions. (NJ: Hoboken,
John Wiley & Sons, 2005).
[60] Theodore L. Nanotechnology: Basic Calculations for Engineers and Scientists. (NJ: Hoboken, John
Wiley & Sons, 2006).
[61] Mills A. The freezing bomb. Phys. Education. 2010. 45: 153.
[62] Vidovskii A.L. Experimental determination of pressure during ice expansion. Hydrotechnical
Construction. 1972. 6: 791.
[63] Gun’ko V.M., Turov V.V., Zarko V.I., Pakhlov E.M., Prykhod’ko G.P., Remez O.S., Leboda R.,
Skubiszewska-Zięba J., Blitz J.P. High-pressure cryogelation of nanosilica and surface properties of
cryosilicas. Colloids Surf. A: Physicochem. Eng. Aspects. 2013. 436: 618.
[64] Gun’ko V.M., Zarko V.I., Turov V.V., Leboda R., Chibowski E. The effect of second phase distribution
in disperse X/silica (X= Al2O3, TiO2, and GeO2) on its surface properties. Langmuir. 1999. 15: 5694.
[65] Gun’ko V.M., Turov V.V., Zarko V.I., Voronin E.F., Tischenko V.A., Dudnik V.V., Pakhlov E.M.,
Chuiko A.A. Active site nature of pyrogenic alumina/silica and water bound to surfaces. Langmuir.
1997. 13: 1529.
[66] Gun’ko V.M., Zarko V.I., Turov V.V., Voronin E.F., Tischenko V.A., Chuiko A.A. Dielectric
properties and dynamic simulation of water bound to titania/silica surfaces. Langmuir. 1995. 11: 2115.
[67] Gun’ko V.M., Zarko V.I., Chibowski E., Dudnik V.V., Leboda R., Zaets V.A. Structure of fumed
titania and silica/titania and influence of the nature of surface sites on interaction with water. J. Colloid
Interface Sci. 1997. 188: 39.
[68] Gun’ko V.M., Mironyuk I.F., Zarko V.I., Turov V.V., Voronin E.F., Pakhlov E.M., Goncharuk E.V.,
Leboda R., Skubiszewska-Zięba J., Janusz W., Chibowski S., Levchuk Yu.N., Klyueva A.V. Fumed
silicas possessing different morphology and hydrophilicity. J. Colloid Interface Sci. 2001. 242: 90.
[69] Gun’ko V.M., Turov V.V., Zarko V.I., Goncharuk E.V., Gerashchenko I.I., Turova A.A., Mironyuk
I.F., Leboda R., Skubiszewska-Zięba J., Janusz W. Comparative characterization of polymethylsiloxane
hydrogel and silylated fumed silica and silica gel. J Colloid Interface Sci. 2007. 308: 142.
[70] Blitz J.P., Christensen J.M., Gun’ko V.M. Silica surface modification reactions with aluminum and
boron alkyls and (alkyl) chlorides: reactivities and surface nanostructures. J. Nanoscience &
Nanotechnology. 2008. 8: 660.
[71] Frei R., Blitz J.P., Gun’ko V.M., Frost B.E., Sergeev V.S. Kinetics and computational studies of an
aminosilane reaction with a silsesquioxane silanol. J. Phys. Chem. A. 2009. 113: 6612.
[72] Gun’ko V.M., Turov V.V., Myronyuk I.F., Goncharuk O.V., Pakhlov E.M., Bezruka N.A., , Skwarek
E., Janusz W., Blitz J.P. Interfacial phenomena at a surface of partially silylated nanosilica. J. Colloid
Interface Sci. 2014. 434(15): 28.
[73] Blitz J.P., Gun’ko V.M., Samala R., Lawrence B.A. Mixed bifunctional surface-modified silicas using
tethered aminofunctional silane catalysts. Colloids Surf. A: Physicochem. Eng. Aspects. 2014. 462: 1.
[74] Terpilowski K., Rymuszka D., Goncharuk O.V., Sulym I.Y., Gun’ko V.M. Wettability of modified
silica layers deposited on glass support activated by plasma. Appl. Surf. Sci. 2015. 353: 843.
[75] Gun’ko V.M., Voronin E.F., Nosach L.V., Turov V.V., Wang Z., Vasilenko A.P., Leboda R.,
Skubiszewska-Zięba J., Janusz W., Mikhalovsky S.V. Structural, textural and adsorption characteristics
of nanosilica mechanochemically activated in different media. J. Colloid. Interface Sci. 2011. 355: 300.
[76] Zaulychnyy Ya.V., Gun’ko V.M., Yavorskyi Y.V., Zarko V.I., Petrovska S.S., Mischenko V.N. Effect
of mechanical activation of highly disperse SiO2/α-Fe2O3 mixtures on distribution of valence electrons.
Metallofizika i Noveishie Tekhnologii. 2015. 37: 1063.
[77] Sulim I.Y., Borysenko M.V., Korduban O.M., Gun’ko V.M. Influence of silica matrix morphology on
characteristics of grafted nanozirconia. Appl. Surf. Sci. 2009. 255: 7818.
[78] Boratyrev V.M., Gun’ko V.M., Galaburda M.V., Borysenko M.V., Pokrovsky V.A., Oranska O.I.,
Polshin E.V., Korduban O.M., Leboda R., Skubiszewska-Zięba J. Synthesis and characterization of
Fe2O3/SiO2 nanocomposites. J. Colloid Interface Sci. 2009. 338: 376.
254
[79] Gun’ko V.M., Bogatyrev V.M., Leboda R., Skubiszewska-Zięba J., Petrus L.V., Nychiporuk Yu.M.,
Oranska O.I., Dudarko O.A., Osovskii V.D., Ptushinskii Y.G. Titania deposits on nanosilicas. Annales
Universitatis Marie Curie-Sklodowska Sectio Chemia. 2009. 64: 21.
[80] Gun’ko V.M., Bogatyrov V.M., Oranska O.I., Borysenko LI, Skubiszewska-Zięba J., Książek A.,
Leboda R. Structural features of ZnxOy/nanosilica composites. Appl. Surf. Sci. 2013. 276: 802.
[81] Nazarkovsky M.A., Gun’ko V.M., Zarko V.I., Skwarek E., Skubiszewska-Zięba J., Leboda R., Janusz
W. Textural characteristics of SnO2-doped titania/nanosilica and transition of phase of bound water.
Journal of Chemical Technology and Metallurgy (Sofia, Bulgaria). 2013. 48: 373.
[82] Sulym I., Goncharuk O., Skwarek E., Sternik D., Borysenko M.V., Derylo-Marczewska A., Janusz W.,
Gun’ko V.M. Silica-supported ceria–zirconia and titania–zirconia nanocomposites: Structural
characteristics and electrosurface properties. Colloids Surf. A: Physicochem. Eng. Aspects. 2015. 482:
631.
[83] Myronyuk L.I., Myronyuk I.F., Chelyadyn V.L., Sachko V.M., Nazarkovsky M.A., Leboda R.,
Skubiszewska-Zięba J., Gun’ko V.M. Structural and morphological features of crystalline nanotitania
synthesized in different aqueous media. Chem. Phys. Lett. 2013. 583: 103
[84] Adamson A.W., Gast A.P. Physical Chemistry of Surface. Sixth edition. (New York: Wiley, 1997).
[85] Gregg S.J., Sing K.S.W. Adsorption, Surface Area and Porosity. (London: Academic Press, 1982).
[86] Gun’ko V.M. Composite materials: textural characteristics. Applied Surface Sci. 2014. 307: 444.
[87] Schmidt M.W., Baldridge K.K., Boatz J.A., Elbert S.T., Gordon M.S., Jensen J.H., Koseki S.,
Matsunaga N., Nguyen K.A., Su S.J., Windus T.L., Dupuis M., Montgomery J.A. General atomic and
molecular electronic structure system. J. Comput. Chem. 1993. 14: 1347.
[88] Gun’ko V.M., Leboda R., Skubiszewska-Zięba J. Heating effects on morphological and textural
characteristics of individual and composite nanooxides. Adsorption. 2009. 15: 89.
[89] Gun’ko V.M., Skubiszewska-Zięba J., Leboda R., Khomenko K.N., Kazakova O.A., Povazhnyak M.O.,
Mironyuk I.F. Influence of morphology and composition of fumed oxides on changes in their structural
and adsorptive characteristics on hydrothermal treatment at different temperatures. J. Colloid Interface
Sci. 2004. 269: 403.
[90] Gun’ko V.M., Zarko V.I., Goncharuk O.V., Matkovsky A.K., Remez O.S., Skubiszewska-Zięba J.,
Wojcik G., Walusiak B., Blitz J.P. Nature and morphology of fumed oxides and features of interfacial
phenomena. Appl. Surf. Sci. 2016. 366: 410.
[91] Gun’ko V.M., Turov V.V., Zarko V.I., Pakhlov E.M., Charmas B., Skubiszewska-Zięba J. Influence of
structural organization of silicas on interfacial phenomena. Colloids Surf. A: Physicochem. Eng.
Aspects. 2016. 492: 230.
[92] Nishihara H., Mukai S.R., Fujii Y., Tago T., Masuda T., Tamon H. Preparation of monolithic SiO2–
Al2O3 cryogels with inter-connected macropores through ice template. J. Mater. Chem. 2006. 16: 3231.
[93] Nishihara H., Iwamura S., Kyotani T. Synthesis of silica-based porous monoliths with straight
nanochannels using an ice-rod nanoarray as a template. J. Mater. Chem. 2008. 18: 3662.
[94] Mukai S.R., Nishihara H., Tamon H. Porous microfibers and microhoneycombs synthesized by ice
templating. Catal. Surv. Asia. 2006. 10: 161.
[95] Nishihara H., Mukai S.R., Shichi S., Tamon H. Preparation of titania–silica cryogels with controlled
shapes and photocatalysis through unidirectional freezing. Materials Letters. 2010. 64: 959.
[96] Mukai S.R., Nishihara H., Shichi S., Tamon H. Preparation of porous TiO2 cryogel fibers through
unidirectional freezing of hydrogel followed by freeze-drying. Chem. Mater. 2004. 16: 4987.
[97] Shlyakhtin OA, Oh Y-J. Inorganic cryogels for energy saving and conversion. J. Electroceram. 2009.
23: 452.
[98] Pons A., Casas Ll., Estop E., Molins E., Harris K.D.M., Xu M. A new route to aerogels: Monolithic
silica cryogels. J. Non-Crystal Solid. 2012. 358: 461.
[99] Tamon H., Akatsuka T., Mori H., Sano N. Synthesis of zeolite monolith with hierarchical
micro/macropores by ice-templating and steam-assisted crystallization. Chem. Eng. Trans. 2013. 32:
2059.
[100] Chen L., Ye G., Xu D., Zhu L., Lu Z., Dong L., Liu Y. Chemical bond change of gibbsite and fumed
silica mixture during mechanic al activation. Materials Letters. 2012. 85: 91.
255
[101] Gun’ko V.M., Goncharuk O.V., Goworek J. Evaporation of polar and nonpolar liquids from silica gels
and fumed silica. Colloids Surf. A: Physicochem. Eng. Aspects. 2015. 474: 52.
[102] Voronin E.F., Gun’ko V.M., Guzenko N.V., Pakhlov E.M., Nosach L.V., Malysheva M.L.,
Skubiszewska-Zięba J., Leboda R., Borysenko M.V., Chuiko A.A. Interaction of poly(ethylene oxide)
with fumed silica. J. Colloid Interface Sci. 2004. 279: 326.
[103] Gun’ko V.M., Turov V.V., Leboda R., Zarko V.I., Skubiszewska-Zięba J., Charmas B. Adsorption,
NMR and thermally stimulated depolarization current methods for comparative analysis of
heterogeneous solid and soft materials. Langmuir. 2007. 23: 3184.
[104] Gun’ko V.M., Turov V.V., Bogatyrev V.M., Petin A.Y., Turov A.V., Trachevskyi V.V., Blitz J.P. The
influence of pre-adsorbed water on adsorption of methane on fumed and nanoporous silicas. Appl. Surf.
Sci. 2011. 258: 1306.
[105] Kiselev A.V., Lygin V.I. Infrared Spectra of Surface Compounds. (New York: Wiley, 1975).
[106] McCool B., Murphy L., Tripp C.P. A simple FTIR technique for estimating the surface area of silica
powders and films. J. Colloid Interface Sci. 2006. 295: 294.
[107] Gun’ko V.M., Leboda R., Skubiszewska-Zięba J., Goncharuk E.V., Nychiporuk Yu.M., Zarko V.I.,
Blitz J.P. Influence of different treatments on characteristics of nanooxide powders alone or with
adsorbed polar polymers or proteins. Powder Technology. 2008. 187: 146.
[108] Gude K., Gun’ko V.M., Blitz J.P. Adsorption and photocatalytic decomposition of methylene blue on
surface modified silica and silica-titania. Colloids Surf. A: Physicochem. Eng. Aspects. 2008. 325: 17.
[109] Gun’ko V.M., Blitz J.P., Zarko V.I., Turov V.V., Pakhlov E.M., Oranska O.I., Goncharuk E.V.,
Gornikov Y.I., Sergeev V.S., Kulik T.V., Palyanytsya B.B., Samala R.K. Structural and adsorption
characteristics and catalytic activity of titania and titania-containing nanomaterials. J. Colloid Interface
Sci. 2009. 330: 125.
[110] Silberberg M.A. Chemistry – The Molecular Nature of Matter and Change. Fourth edition. (New York:
McGraw-Hill, 2006).
[111] Ashgriz V. (editor). Handbook of Atomization and Sprays. (Heidelberg: Springer, 2011).
[112] Fujikawa S., Yano T., Watanabe M. Vapor-Liquid Interfaces, Bubbles and Droplets. (Heidelberg:
Springer, 2011).
[113] McElroy M.B. The Atmospheric Environment. (Princeton: Princeton University Press, 2002).
[114] Ahsan A. (editor), Evaporation, Condensation and Heat Transfer. (Rijeka: InTech, Croatia, 2010).
[115] Sazhin S.S. Droplets and Sprays. (London: Springer, 2014).
[116] Chapman S., Cowling T.G. The Mathematical Theory of Nonuniform Gases. (Cambridge: Cambridge
University Press, 1970).
[117] Tamim J., Hallett W.L.H. A continuous thermodynamics model for multicomponent droplet
vaporization. Chem. Eng. Sci. 1995. 50: 2933.
[118] Lippert A.M., Reitz R.D. Modeling of multicomponent fuels using continuous distributions with
application to droplet evaporation and sprays. SAE Technical Paper 1997, 972882.
[119] Hallett W.L.H. A simple model for the vaporization of droplets with large numbers of components.
Combustion and Flame. 2000. 121: 334.
[120] Zhu G.-S., Reitz R..D. A model for high-pressure vaporization of droplets of complex liquid mixtures
using continuous thermodynamics. Int. J. Heat Mass Transf. 2002. 45: 495.
[121] Derkachov G., Kolwas K., Jakubczyk D., Zientara M., Kolwas M. Drying of a microdroplet of water
suspension of nanoparticles: from surface aggregates to microcrystal. J. Phys. Chem. C. 2008. 112:
16919.
[122] Sefiane K., Ward C.A. Recent advances on thermocapillary flows and interfacial conditions during the
evaporation of liquids. Adv. Colloid Interface Sci. 2007. 134–135: 201.
[123] Erbil H.Y. Evaporation of pure liquid sessile and spherical suspended drops: a review. Adv. Colloid
Interface Sci. 2012. 170: 67.
[124] Erbil H.Y. Control of stain geometry by drop evaporation of surfactant containing dispersions. Adv.
Colloid Interface Sci. 2015. 222: 275.
[125] Sokolowski S., Fischer J. Liquid-vapour’ density profiles for fluids in pores from density functional
theory. J. Chem. Soc. Faraday Trans. 1993. 89: 789.
256
[126] Bucior K., Yelash L., Binder K. Molecular-dynamics simulation of evaporation processes of fluid
bridges confined in slitlike pores. Phys. Rev. E. 2009. 79: 031604 (1-12).
[127] Morishige K., Shikimi M. Adsorption hysteresis and pore critical temperature in a single cylindrical
pore. J. Chem. Phys. 1998. 108: 7821.
[128] Paulik F., Paulik J. Investigations under quasi-isothermal and quasi-isobaric conditions by means of the
Derivatograph. J. Thermal Anal. 1973. 5: 253.
[129] Meyers R.A. (editor). Encyclopedia of Analytical Chemistry. (Chichester: John Wiley & Sons Ltd.,
2000).
[130] Haines P.J. (editor). Principles of Thermal Analysis & Calorimetry. Ch. 4. (Cambridge: Royal Society
of Chemistry, 2002).
[131] Semenov S., Trybala A., Rubio R.G., Kovalchuk N., Starov V., Velarde M.G. Simultaneous spreading
and evaporation: recent developments. Adv. Colloid Interface Sci. 2014. 206: 382.
[132] Frisch M.J., Trucks G.W., Schlegel H.B., Scuseria G.E., Robb M.A., Cheeseman J.R, Scalmani G,
Barone V, Mennucci B, Petersson GA, Nakatsuji H, Caricato M, Li X, Hratchian HP, Izmaylov AF,
Bloino J, Zheng G, Sonnenberg JL, Hada M, Ehara M, Toyota K, Fukuda R, Hasegawa J, Ishida M,
Nakajima T, Honda Y, Kitao O, Nakai H, Vreven T, Montgomery JA Jr, Peralta JE, Ogliaro F,
Bearpark M, Heyd JJ, Brothers E, Kudin KN, Staroverov VN, Kobayashi R, Normand J, Raghavachari
K, Rendell A, Burant J C, Iyengar SS, Tomasi J, Cossi M, Rega N, Millam JM, Klene M, Knox JE,
Cross JB, Bakken V, Adamo C, Jaramillo J, Gomperts R, Stratmann RE, Yazyev O, Austin AJ, Cammi
R, Pomelli C, Ochterski JW, Martin RL, Morokuma K, Zakrzewski VG, Voth GA, Salvador P,
Dannenberg JJ, Dapprich S, Daniels AD, Farkas Ö, Foresman JB, Ortiz JV, Cioslowski J, Fox DJ.
Gaussian, Inc., Wallingford CT, 2009., Gaussian 09, Revision D.01, Gaussian, Inc., Wallingford CT;
2013.
[133] Chai J.-D., Head-Gordon M. Long-range corrected hybrid density functionals with damped atom-atom
dispersion corrections. Phys. Chem. Chem. Phys. 2008. 10: 6615.
[134] Yang K., Zheng J., Zhao Y., Truhlar D.G. Tests of the RPBE, revPBE, τ-HCTHhyb, ωB97X-D, and
MOHLYP density functional approximations and 29 others against representative databases for diverse
bond energies and barrier heights in catalysis. J. Chem. Phys. 2010. 132: 164117.
[135] Marenich A.V., Cramer C.J., Truhlar D.G. Universal solvation model based on solute electron density
and on a continuum model of the solvent defined by the bulk dielectric constant and atomic surface
tensions. J. Phys. Chem. B. 2009. 113: 6378.
[136] Becke A.D. Perspective: Fifty years of density-functional theory in chemical physics. J. Chem. Phys.
2014. 140: 18A301.
[137] Ho J., Coote M.L., Cramer C.J., Truhlar D.G. Theoretical calculations of reduction potentials. In :
Hammerich O., Speiser B (editors). Organic Electrochemistry. Fifth Edition. (Boca Raton: CRC Press,
2016, pp. 229-59).
[138] Gun’ko V.M., Nasiri R., Sazhin S.S., Lemoine F., Grisch F. A quantum chemical study of the processes
during the evaporation of real-life Diesel fuel droplets. Fluid Phase Equilibria. 2013. 356: 146.
[139] Gun’ko V.M., Nasiri R., Sazhin S.S. A study of the evaporation and condensation of n-alkane clusters
and nanodroplets using quantum chemical methods. Fluid Phase Equilibria. 2014. 366: 99.
[140] Ortega I.K., Kupiainen O., Kurtén T., Olenius T., Wilkman O., McGrath M.J., Loukonen V.,
Vehkamäki H. From quantum chemical formation free energies to evaporation rates. Atmos. Chem.
Phys. 2012. 12: 225.
[141] Kupiainen O., Ortega I.K., Kurtén T., Vehkamäki H. Amine substitution into sulfuric acid – ammonia
clusters. Atmos. Chem. Phys. 2012. 12: 3591.
[142] Beverley K.J., Clint J.H., Fletcher P.D.I., Thubron S. Evaporation rates of water contained within
porous silica particles. Phys. Chem. Chem. Phys. 1999. 1: 909.
[143] Yaws C.L. (editor). Thermophysical Properties of Chemicals and Hydrocarbons. (NY: William
Andrew Inc., Norwich, 2008).
[144] Gun’ko V.M., Nasiri R., Sazhin S.S. Effects of the surroundings and conformerisation of n-dodecane
molecules on evaporation/condensation processes. J. Chem. Phys. 2015. 142: 034502.
257
[145] Goworek J., Stefaniak W., Prudaczuk M. The influence of polarity of liquids on the parameters
characterizing the porosity of silica gels estimated by thermogravimetric analysis. Thermochimica Acta.
2001. 379: 117.
[146] Singh M.P., Singh R.K., Chandra S. Ionic liquids confined in porous matrices: Physicochemical
properties and applications. Prog. Mater. Sci. 2014. 64: 73.
[147] Gubbins K.E., Long Y., Sliwinska-Bartkowiak M. Thermodynamics of confined nano-phases. J. Chem.
Thermodynamics. 2014. 74: 169.
[148] Wang L., Yu Q. Methane adsorption on porous nano-silica in the presence of water: An experimental
and ab initio study. J. Colloid Interface Sci. 2016. 467: 60.
[149] Gun’ko V.M., Sulym I.Y., Borysenko M.V., Turov V.V. Interfacial behavior of water bound to
zirconia/nanosilica with adsorbed poly(dimethylsiloxane). Colloids Surf. A: Physicochem. Eng. Aspects.
2013. 426: 47.
[150] Stewart J.J.P. MOPAC 2016, Stewart Computational Chemistry, Colorado Springs, CO, USA,
http://openmopac.net/, 2019.
[151] Stewart J.J.P. Optimization of parameters for semiempirical methods VI: more modifications to the
NDDO approximations and re-optimization of parameters. J. Mol. Mod. 2013. 19: 1.
[152] Maia J.D.C., Carvalho G.A.U., Mangueira C.P. Jr., Santana S.R., Cabral L.A.F., Rocha G.B. GPU
linear algebra libraries and GPGPU programming for accelerating MOPAC semiempirical quantum
chemistry calculations. J. Chem. Theory Comput. 2012. 8: 3072.
[153] Cheeseright T., Mackey M., Rose S., Vinter J.G. Molecular field technology applied to virtual
screening and finding the bioactive conformation. Expert. Opin. Drug. Discov. 2007. 2: 131.
[154] Cheeseright T., Mackey M., Rose S., Vinter J.G. Molecular field extrema as descriptors of biological
activity: definition and validation. J. Chem. Inf. Model. 2006. 46: 665.
[155] Dennington R., Keith T., Millam J. GaussView, Version 5.0.9, Semichem Inc., Shawnee Mission KS;
2009.
[156] Zhurko G.A, Zhurko D.A. Chemcraft (version 1.8), 2019, http://www.chemcraftprog.com.
[157] Pettersen E.F., Goddard T.D., Huang C.C., Couch G.S., Greenblatt D.M., Meng E.C., Ferrin T.E. UCSF
Chimera--a visualization system for exploratory research and analysis. J. Comput. Chem. 2004. 25:
1605.
[158] Turov V.V., Gun’ko V.M., Zarko V.I., Goncharuk O.V., Krupska T.V., Turov A.V., Leboda R.,
Skubiszewska-Zięba J. Interfacial behavior of n-decane bound to weakly hydrated silica gel and
nanosilica over a broad temperature range. Langmuir. 2013. 29: 4303.
[159] Takei T., Mukasa K., Kofuji M., Fuji M., Watanabe T., Chikazawa M., Kanazawa T. Changes in
density and surface tension of water in silica pores. Colloid Polym. Sci. 2000. 278: 475.
[160] Petch H.E. The hydrogen positions in portlandite, Ca(OH)2, as indicated by the electron distribution.
Acta Crystallogr. 1961. 14: 950.
[161] Desgranges L., Grebille D., Calvarin G., Chevrier G., Floquet N., Niepce J.C. Hydrogen thermal motion
in calcium hydroxide: Ca(OH)2. Acta Crystallogr. Sect. B: Struct. Sci. 1993. 49: 812.
[162] Jennings H. Refinements to colloidal model of C–S–H in cement: CM-II. Cem. Concr. Res. 2008. 38:
275.
[163] Famy C., Scrivener K.L., Atkinson A. Effects of an early or a late heat treatment on the microstructure
and composition of inner C–S–H products of Portland cement mortars. Cem. Concr. Res. 2002. 32: 269.
[164] Nonat A. The structure and stoichiometry of C–S–H. Cem. Concr. Res. 2004. 34: 1521.
[165] Gallucci E., Zhang X., Scrivener K.L. Effect of temperature on the microstructure of calcium silicate
hydrate (C–S–H). Cem. Concr. Res. 2013. 53: 185.
[166] Muller A.C.A., Scrivener K.L., J. Skibsted J., Gajewicz A.M., McDonald P.J. Influence of silica fume
on the microstructure of cement pastes: New insights from 1H NMR relaxometry. Cement and Concrete
Research. 2015. 74: 116.
[167] Zhu W., Zhou Y., Ma W., Li M., Yu J., Xie K. Using silica fume as silica source for synthesizing
spherical ordered mesoporous silica. Materials Letters. 2013. 92129.
[168] Gun’ko V.M., Zarko V.I., Mironyuk I.F., Goncharuk E.V., Guzenko N.V., Borysenko M.V., Gorbik
P.P., Mishchuk O.A., Janusz W., Leboda R., Skubiszewska-Zięba J., Grzegorczyk W., Matysek M.,
258
Chibowski S. Surface electric and titration behaviour of fumed oxides. Colloids Surf. A: Physicochem.
Eng. Aspects. 2004. 240: 9.
[169] Gun’ko V.M., Zarko V.I., Turov V.V., Goncharuk E.V., Nychiporuk Yu.M., Turova A.A., Gorbik P.P.,
Leboda R., Skubiszewska-Zięba J., Pissis P., Blitz J.P. Regularities in the behaviour of nanooxides in
different media affected by surface structure and morphology of particles. In: Shpak A.P., Gorbyk P.P.
(editors). Nanomaterials and Supramolecular Structures. (Dordrecht: Springer, 2010, pp. 93-118).
[170] Gun’ko V.M., Andriyko L.S., Zarko V.I., Marynin A.I., Olishevskiy V.V., Janusz W. Effects of
chlorides of alkaline metals on the behavior of nanosilica in aqueous media. Central Eur. J. Chem.
2014. 12(4): 480.
[171] Andriyko L.S., Zarko V.I., Gun’ko V.M., Marynin A.I., Olishevskiy V.V., Skwarek E., Janusz W.
Electrical and physical characteristics of silica nanoparticles in aqueous media affected by cations Na+,
Ba2+ and Al3+. Adsorption Science & Technology. 2015. 33: 601.
[172] Janusz W., Skwarek E., Zarko V.I., Gun’ko VM. Structure of electrical double layer at the Al2O3-
SiO2/electrolyte solution interface. Physicochemical Problems of Mineral Processing. 2007. 41: 215.
[173] Skwarek E., Matysek-Nawrocka M., Janusz W., Zarko V.I., Gun’ko V.M. Adsorption of heavy metal
ions at the Al2O3-SiO2/NaClO4 electrolyte interface. Physicochemical Problems of Mineral
Processing.2008. 42: 153.
[174] Wiśniewska M., Terpiłowski K., Chibowski S., Chibowski E., Urban T., Zarko V.I., Gun’ko V.M.
Effect of polyacrylic acid (PAA) adsorption on stability of mixed alumina‐silica oxide suspension.
Powder Technology. 2013. 233: 190.
[175] Wiśniewska M., Terpiłowski K., Chibowski S., Urban T., Zarko V.I., Gun’ko V.M. Effect of solution
pH on the stability of mixed silica -alumina suspension in the presence of polyacrylic acid (PAA) with
different molecular weights. Central Eur. J. Chem. 2013. 11: 101.
[176] Wiśniewska M., Terpiłowski K., Chibowski S., Chibowski E., Urban T., Zarko V.I., Gun’ko V.M.
Stability of colloidal silica modified by macromolecular polyacrylic acid (PAA) – application of
turbidymetry method. J. Macromol. Sci. Part A: Pure Appl. Chem. 2013. 50: 639.
[177] Wiśniewska M., Urban T., Grządka E., Zarko V.I., Gun’ko V.M. Comparison of adsorption affinity of
polyacrylic acid for surfaces of mixed silica–alumina. Colloid Polym. Sci. 2014. 292: 699.
[178] Wiśniewska M., Urban T., Nosal-Wiercińska A., Zarko V.I., Gun’ko V.M. Comparison of stability
properties of poly(acrylic acid) adsorbed on the surface of silica, alumina and mixed silica-alumina
nanoparticles - Application of turbidimetry method. Central Eur. J. Chem. 2014. 12: 476.
[179] Wiśniewska M., Terpiłowski K., Chibowski S., Urban T., Zarko V.I., Gun’ko V.M. Investigation of
stabilization and destabilization possibilities of water alumina suspension in polyelectrolyte presence.
International Journal of Mineral Processing. 2014. 132: 34.
[180] Wiśniewska M., Ostolska I., Szewczuk-Karpisz K., Chibowski S., Terpiłowski K., Gun’ko V.M., Zarko
V.I. Investigation of the polyvinyl alcohol stabilization mechanism and adsorption properties on the
surface of ternary mixed nanooxide AST 50 (Al2O3–SiO2–TiO2). Journal of Nanoparticle Research.
2015. 17: 14.
[181] Wiśniewska M., Szewczuk-Karpisz K., Ostolska I., Urban T., Terpiłowski K., Zarko V.I., Gun’ko
V.M. Effect of polyvinyl alcohol adsorption on the mixed alumina–silica–titania suspension stability.
Journal of Industrial and Engineering Chemistry. 2015. 23: 265.
http://dx.doi.org/10.1016/j.jiec.2014.08.027.
[182] Wawrzkiewicz M., Wiśniewska M., Gun’ko V., Zarko V. Adsorptive removal of acid, reactive and
direct dyes from aqueous solutions and wastewater using mixed silica-alumina oxide. Powder
Technology. 2015. 278: 306.
[183] Grządka E., Wiśniewska M., Gun’ko V.M., Zarko V.I. Adsorption, electrokinetic and stabilizing
properties of the system: guar gum/surfactant/alumina. Journal of Sufractants and Detergents. 2015.
18: 445.
[184] Blitz I.P., Blitz J.P., Gun’ko V.M., Sheeran D.J. Functionalized silicas: structural characteristics and
adsorption of Cu(II) and Pb(II). Colloids Surf. A: Physicochem Eng Aspects. 2007. 307: 83.
[185] Kothalawala N., Blitz J.P., Gun’ko V.M., Jaroniec M., Grabicka B., Semeniuc R.F. Post-synthesis
surface modified silicas as adsorbents for heavy metal ion contaminants Cd(II), Cu(II), Cr(III), and
Sr(II) in aqueous solutions. J. Colloid Interface Sci. 2013. 392: 57.
259
[186] Gun’ko V.M., Blitz J.P., Bandaranayake B., Pakhlov E.M., Zarko V.I., Sulym I.Ya., Kulyk K.S.,
Galaburda M.V., Bogatyrev V.M., Oranska O.I., Borysenko M.V., Leboda R., Skubiszewska-Zięba J.,
Janusz W. Structural characteristics of mixed oxides MOx/SiO2 affecting photocatalytic decomposition
of methylene blue. Appl. Surf. Sci. 2012. 258: 6288.
[187] Rahni M.T., Karbaschi M., Miller R. (editors). Computational Methods for Complex Liquid-Fluid
Interfaces. (CRC Press, Boca Raton; 2015).
[188] Zhan C., Chen F., Yang J., Dai D., Cao X., Zhong M. Visible light responsive sulfated rare earth doped
TiO2/fumed SiO2 composites with mesoporosity: Enhanced photocatalytic activity for methyl orange
degradation. J. Hazard Mater. 2014. 267: 88–97.
[189] Yu Y., Wang J., Parr J.F. Preparation and properties of TiO2/fumed silica composite photocatalytic
materials. Procedia Eng. 2012. 27: 448.
[190] Diamanti M.V., Ormellese M., Marin E., Lanzutti A., Mele A., Pedeferri M.P. Anodic titanium oxide as
immobilized photocatalyst in UV or visible light devices. J. Hazard Mater. 2011. 186: 2103.
[191] Bokare A., Pai M., Athawale A.A. Surface modified Nd doped TiO2 nanoparticles as photocatalysts in
UV and solar light irradiation. Solar Energy. 2013. 91: 111.
[192] Zhan C.C., Chen F., Dai H.H., Yang J.T., Zhong M.Q. Photocatalytic activity of sulfated Mo-doped
TiO2/fumed SiO2 composite: a mesoporous structure for methyl orange degradation. Chem. Eng. J.
2013. 225: 695.
[193] Goncharuk E.V., Mishchenko V.N., Zarko V.I., Gun’ko V.M. Effect of the composition and structure
of titanosilicas on their photocatalytic activity in the decomposition of methylene blue. Theor. Experim.
Chem. 2006. 42: 26.
[194] Jaroenworaluck A., Pijarn N., Kosachan N., Stevens R. Nanocomposite TiO2–SiO2 gel for UV
absorption. Chem. Eng. J. 2012. 181–182: 45.
[195] Provencher SW. A constrained regularization method for inverting data represented by linear algebraic
or integral equations. Comp. Phys. Comm. 1982. 27: 213.
[196] Laidler K.J., Meiser J.H. Physical Chemistry. (Menlo Park CA: Benjamin/Cummings, 1982).
[197] Fujishima A., Hashimoto K., Wanable H. TiO2 Photocatalysis: Fundamentals and Applications.
(Tokyo, Japan: BKC Inc., 1997).
[198] Thompson T.L., Yates J.T. Jr. Surface Science studies of the photoactivation of TiO2 - new
photochemical processes. Chem. Rev. 2006. 106: 4428.
[199] Glushko V.P. (editor). Handbook of thermodynamic properties of individual substances. (Moscow:
Nauka, 1978). [in Russian].
[200] Mitchell J., Webber J.B.W., Strange J.H. Nuclear magnetic resonance cryoporometry. Phys. Rep. 2008.
461(1): 1.
[201] Petrov O.V., Furó I. NMR cryoporometry: Principles, application and potential. Progr. NMR
Spectroscopy. 2009. 54: 97.
[202] Aksnes D.W., Forl K., Kimtys L. Pore size distribution in mesoporous materials as studied by 1H NMR.
Phys. Chem. Chem. Phys. 2001. 3: 3203.
[203] Webber J.B.W., Anderson R., Strange J.H., Tohidi B. Clathrate formation and dissociation in
vapor/water/ice/hydrate systems in SBA-15, sol–gel and CPG porous media, as probed by NMR
relaxation, novel protocol NMR cryoporometry, neutron scattering and ab initio quantum–mechanical
molecular dynamics simulation. Magnet Reson. Imag. 2007. 25: 533.
[204] Gun’ko V.M., Savina I.N., Mikhalovsky S.V. Cryogels: Morphological, structural and adsorption
characterization. Adv. Colloid Interface Sci. 2013. 187–188: 1.
[205] Petrov O.V., Furó I. A joint use of melting and freezing data in NMR cryoporometry. Micropor.
Mesopor. Mater. 2010. 136(1-3): 83.
[206] Schulz P.S. Ionic liquids as solvent probes for NMR cryoporometry. Chem.Phys.Chem. 2010. 11: 87.
doi:10.1002/cphc.200900804.
[207] Shiko E., Edler K.J., Lowe J.P., Rigby S.P. Probing hysteresis during sorption of cyclohexane within
mesoporous silica using NMR cryoporometry and relaxometry. J. Colloid Interface Sci. 2013. 398: 168.
[208] Mitchell J., Gladden L.F., Chandrasekera T.C., Fordham E.J. Low-field permanent magnets for
industrial process and quality control. Prog. Nucl. Magn. Reson. Spectrosc. 2014. 76: 1.
260
[209] Gun’ko V.M. Interfacial phenomena: effects of confined space and structure of adsorbents on the
behavior of polar and nonpolar adsorbates at low temperatures. Current Physical Chemistry. 2015. 5:
137.
[210] Gun’ko V.M. Modeling of interfacial behavior of water and organics. J. Theor. Comp. Chem. 2013.
12(7):1350059.
[211] Mallamace F., Corsaro C., Broccio M., Branca C., González-Segredo N., Spooren J., Chen S.-H.,
Stanley H.E. NMR evidence of a sharp change in a measure of local order in deeply supercooled
confined water. Proc. Natl. Acad. Sci. USA. 2008. 105(35): 12725.
[212] Landry M.R. Thermoporometry by differential scanning calorimetry: experimental considerations and
applications. Thermochim. Acta 2005. 433(1-2): 27.
[213] Gun’ko V.M., Turov V.V., Bogatyrev V.M., Zarko V.I., Leboda R., Goncharuk E.V., Novza A.A.,
Turov A.V., Chuiko A.A. Unusual properties of water at hydrophilic/hydrophobic interfaces. Adv.
Colloid Interface Sci. 2005. 118(1-3): 125.
[214] Gun’ko V.M., Zarko V.I., Goncharuk E.V., Andriyko L.S., Turov V.V., Nychiporuk Y.M, Leboda R.,
Skubiszewska-Zieba J., Gabchak A.L., Osovskii V.D., Ptushinskii Y.G., Yurchenko G.R., Mishchuk
O.A., Gorbik P.P., Pissis P., Blitz J.P. TSDC spectroscopy of relaxational and interfacial phenomena.
Adv. Colloid Interface Sci. 2007. 131(1-2): 1.
[215] Turov V.V., Gun’ko V.M., Turova A.A., Morozova L.P., Voronin E.F. Interfacial behavior of
concentrated HCl solution and water clustered at a surface of nanosilica in weakly polar solvents media.
Colloids Surf. A: Physicochem. Eng. Aspects 2011. 390(1-3): 48. doi: 10.1016/j.colsurfa.2011.08.053.
[216] Gun’ko V.M., Turov V.V., Skubiszewska-Zięba J., Leboda R., Tsapko M.D., Palijczuk D. Structural
characteristics of a carbon adsorbent and influence of organic solvents on interfacial water. Appl. Surf.
Sci. 2003. 214(1-4): 178.
[217] Turov V.V., Gun’ko V.M., Zarko V.I., Leboda R., Jabłoński M., Gorzelak M., Jagiello-Wojtowicz E.
Weakly and strongly associated nonfreezable water bound in bones. Colloids Surf. B: Biointerfaces
2006. 48(2): 167.
[218] Kinney D.R., Chuang I.-S., Maciel G.E. Water and the silica surface as studied by variable temperature
high resolution 1H NMR. J. Am. Chem. Soc. 1993. 115(15): 6786.
[219] Chuang I.-S., Maciel G.E. Probing hydrogen bonding and the local environment of silanols on silica
surfaces via nuclear spin cross polarization dynamics. J. Am. Chem. Soc. 1996. 118(2): 401.
[220] Liu C.C., Maciel G.E. The fumed silica surface: a study by NMR. J. Am. Chem. Soc. 1996. 118(21):
5103.
[221] Humbert H. Estimation of hydroxyl density at the surface of pyrogenic silicas by complementary NMR
and Raman experiments. J. Non-Crystal Solid 1995. 191(1-2): 29.
[222] Hu J.Z., Kwak J.H., Herrera J.E., Wang Y., Peden Ch.H.F. Line narrowing in 1H MAS spectrum of
mesoporous silica by removing adsorbed H2O using N2. Solid State Nuclear Magnetic Resonance 2005.
27(3): 200.
[223] Fonseca I., Matos J., Gonçalves M.C., Carvalho A., Sebasti P.J. Silica and silica organically modified
nanoparticles: Water dynamics in complex systems. Microporous and Mesoporous Materials 2015.
217: 102.
[224] Jimura K., Hayashi S. Proton diffusion in hybrid materials of CsHSO4 and silica nanoparticles as
studied by 1H solid-state NMR. Solid State Sci. 2012. 14(1): 171.
[225] Walia J., Crone J., Liang J., Niknam M., Lemaire C., Thompson R.T., Peemoeller H. Temperature and
hydration dependence of proton MAS NMR spectra in MCM-41: Model based on motion induced
chemical shift averaging. Solid State Nucl. Magn. Reson. 2013. 49-50: 26.
[226] Brinker C.J., Scherer G.W. Sol-Gel Science: The Physics and Chemistry of Sol-Gel Processing. (NY:
Academic Press, 1990).
[227] Pierre A.C. Introduction to Sol-Gel Processing. (Boston: Kluwer, 1998).
[228] Vansant E.F., Van Der Voort P., Vrancken K.C. Characterization and Chemical Modification of the
Silica Surface. Studies in Surface Science and Catalysis. Vol. 93 (Amsterdam: Elsevier, 1995).
[229] Gun’ko V.M., Zarko V.I., Voronin E.F., Turov V.V., Mironyuk I.F., Gerashchenko I.I., Goncharuk
E.V., Pakhlov E.M., Guzenko N.V., Leboda R., Skubiszewska-Ziȩba J., Janusz W., Chibowski S.,
261
Levchuk Yu.N., Klyueva A.V. Impact of some organics on structural and adsorptive characteristics of
fumed silica in different media. Langmuir. 2002. 18(3): 581.
[230] Mironyuk I.F., Gun’ko V.M., Turov V.V., Zarko V.I., Leboda R., Skubiszewska-Zięba J.
Characterization of fumed silicas and their interaction with water and dissolved proteins. Colloids Surf.
A: Physicochem. Eng. Aspects 2001. 180(1-2): 87.
[231] Gun’ko V.M., Voronin E.F., Mironyuk I.F., Leboda R., Skubiszewska-Zięba J., Pakhlov E.M.,
Guzenko N.V., Chuiko A.A. The effect of heat, adsorption and mechanochemical treatments on stuck
structure and adsorption properties of fumed silicas. Colloids Surf. A: Physicochem. Eng. Aspects 2003.
218(1-3): 125.
[232] Turov V.V., Voronin E.F., Morozova L.P., Gun’ko V.M., Nosach L.V. The effect of mechanical
activation on the hydration properties of nanodispersed silica. Russian Journal of Applied Chemistry
2011. 84: 1304.
[233] Gupta S., Pel L., Kopinga K. Crystallization behavior of NaCl droplet during repeated crystallization
and dissolution cycles: An NMR study. Journal of Crystal Growth 2014. 391: 64.
[234] Xu J.-A., Huang E., Lin J.-F., Xu L.Y. Raman study at high pressure and the thermodynamic properties
of corundum: Application of Kieffer's model. Am. Mineral. 1995. 80(11-12): 1157.
[235] Hair M.L. Infrared Spectroscopy in Surface Chemistry. (NY: Dekker, 1967).
[236] Gun’ko V.M., Turov V.V., Bogatyrev V.M., Charmas B., Skubiszewska-Zięba J., Leboda R.,
Pakhovchishin S.V., Zarko V.I., Petrus L.V., Stebelska O.V., Tsapko M.D. Influence of partial
hydrophobization of fumed silica by hexamethyldisilazane on interaction with water. Langmuir 2003.
19(26): 10816.
[237] Gun’ko V.M., Sheeran D.J., Augustine S.M., Blitz J.P. Structural and energetic characteristics of silicas
modified by organosilicon compounds. J. Colloid Interface Sci. 2002. 249(1): 123.
[238] Gun’ko V.M., Zarko V.I., Leboda R., Chibowski E. Aqueous suspensions of fumed oxides: particle size
distribution and zeta potential. Adv. Colloid Interface Sci. 2001. 91: 1.
[239] Karraker K.A., Radke C.J. Disjoining pressures, zeta potentials and surface tensions of aqueous non-
ionic surfactant/electrolyte solutions: theory and comparison to experiment. Adv. Colloid Interface Sci.
2002. 96: 231.
[240] Bourikas K., Kordulis C., Lycourghiotis A. The mechanism of the protonation of metal (hydr)oxides in
aqueous solutions studied for various interfacial/surface ionization models and physicochemical
parameters: A critical review and a novel approach. Adv. Colloid Interface Sci. 2006. 121: 111. doi:
10.1016/j.cis.2006.06.002.
[241] Kovalchuk N.M., Starov V.M. Aggregation in colloidal suspensions: Effect of colloidal forces and
hydrodynamic interactions. Adv. Colloid Interface Sci. 2012. 179-182: 99. doi:
10.1016/j.cis.2011.05.009.
[242] Chaplin M. Water Structure and Science. http://www.lsbu.ac.uk/water/, 23 October, 2015.
[243] Pople J.A., Schneider W.G., Bernstein H.J. High-Resolution Nuclear Magnetic Resonance. (NY:
McGraw-Hill Book Company, 1959).
[244] Gun’ko V.M., Morozova L.P., Turova A.A., Turov A.V., Gaishun V.E., Bogatyrev V.M., Turov V.V.
Hydrated phosphorus oxyacids alone and adsorbed on nanosilica. J. Colloid Interface Sci. 2012. 368(1):
263.
[245] Gun’ko V.M., Turov V.V., Zarko V.I., Goncharuk E.V., Turova A.A. Regularities in the behaviour of
water confined in adsorbents and bioobjects studied by 1H NMR spectroscopy and TSDC methods at
low temperatures. Colloids Surf. A: Physicochem. Eng. Aspects. 2009. 336(1-3): 147.
[246] Petin A.Yu., Gun’ko V.M., Turov A.V., Turov V.V., Leboda R. Clusterization of water at a surface of
nanosilica A-380. Annales Universitatis Marie Curie-Sklodowska Sectio Chemia 2009. 64(1): 184.
[247] Gun’ko V.M., Turov V.V., Turov A.V. Hydrogen peroxide – water mixture bound to nanostructured
silica. Chem. Phys. Lett. 2012. 531: 132.
[248] Turov V.V., Krupska T.V., Tsapko M.D., Gun’ko V.M. Temperature behavior of water and n-decane
bound to nanosilica or poly(methylsiloxane). Chemistry, Physics and Technology of Surface 2015. 6(2):
244.
[249] Murakhtina T., Heuft J., Meijer E.J, Sebastiani D. First principles and experimental 1H NMR signatures
of solvated ions: the case of HCl(aq). Chem. Phys. Chem. 2006. 7(12): 2578.
262
[250] Wypych G. (editor), Handbook of Solvents. (Toronto: ChemTec Publishing, 2001).
[251] Gun’ko V.M. Modelling of evaporation of clusters and nanodroplets of organic molecules using
quantum chemical and the kinetic gas theory methods. Chemistry, Physics and Technology of Surface
2015. 6(1): 5.
[252] Liboff R.L. Kinetic Theory. (NY: Prentice-Hall, Englewood Cliffs, 1990).
[253] Laurendeau N.M. Statistical Thermodynamics: Fundamentals and Applications. (Cambridge:
Cambridge University Press, 2005).
[254] Sefiane K. Patterns from drying drops. Adv. Colloid Interface Sci. 2014. 206: 372. doi:
10.1016/j.cis.2013.05.002.
[255] Peng C.-C., Cerretani C., Braun R.J., Radke C.J. Evaporation-driven instability of the precorneal tear
film. Adv. Colloid Interface Sci. 2014. 206: 250. doi: 10.1016/j.cis.2013.06.001.
[256] Askounis A., Sefiane K., Koutsos V., Shanahan M.E.R. Effect of particle geometry on triple line
motion of nanofluid drops and deposit nano-structuring. Adv. Colloid Interface Sci. 2015. 222: 44. doi:
10.1016/j.cis.2014.05.003.
[257] Thomas S., Shanks R., Chandrasekharakurup S. (editors), Design and Applications of Nanostructured
Polymer Blends and Nanocomposite Systems. (Amsterdam: Elsevier, 2016).
[258] Ahmed W., Jackson M.J. (editors) Emerging Nanotechnologies for Manufacturing. Second Edition,
(Amsterdam: Elsevier, 2015).
[259] Piemonte V., De Falco M., Basile A. (editors), Sustainable Development in Chemical Engineering –
Innovative Technologies. First Edition. (Hoboken, NJ: John Wiley & Sons, 2013).
[260] Demir M.M., Wegner G. Challenges in the preparation of optical polymer composites with nanosized
pigment particles: a review on recent efforts. Macromol. Mater. Eng. 2012. 297(9): 838.
[261] Kim D.J., Jo M.J., Nam S.Y. A review of polymer-nanocomposite electrolyte membranes for fuel cell
application. J. Ind. Eng. Chem. 2015. 21: 36.
[262] Wang Y., He J., Liu C., Chong W.H., Chen H. Thermodynamics versus kinetics in nanosynthesis.
Angew. Chem. Int. Ed. 2015. 54(7): 2022.
[263] Martinez L.J., Sanchez M.L., Kikot P., Candal R., Grasselli M. Preparation of functional currant-bun-
like fumed silica/polymethacrylate nanoparticles by radiation-induced polymerization. Colloids Surf. A:
Physicochem. Eng. Asp. 2014. 463: 110.
[264] Shadjou N., Hasanzadeh M. Bone tissue engineering using silica-based mesoporous nanobiomaterials:
Recent progress. Mater. Sci. Eng. C 2015. 55: 401.
[265] Yoosuk B., Wongsanga T., Prasassarakich P. CO2 and H2S binary sorption on polyamine modified
fumed silica. Fuel. 2016. 168: 47.
[266] Ramesh S., Liew C.-W. Exploration on nano-composite fumed silica-based composite polymer
electrolytes with doping of ionic liquid. J. Non-Crystal Solid 2012. 358(2): 931.
[267] Damouny C.W., Silverstein M.S. Hydrogel-filled, semi-crystalline, nanoparticle crosslinked, porous
polymers from emulsion templating: Structure, properties, and shape memory. Polymer. 2016. 82:262.
[268] Gun’ko V.M., Turov V.V., Krupska T.V., Ruban AN, Kazanets A.I., Leboda R., Skubiszewska-Zięba J.
Interfacial behavior of silicone oils interacting with nanosilica and silica gels. J. Colloid Interface Sci.
2013. 394: 467.
[269] Kawaguchi M. Silicone oil emulsions stabilized by polymers and solid particles. Adv. Colloid Interface
Sci. 2016. 233: 186.
[270] Klonos P., Kyritsis A., Pissis P. Interfacial dynamics of polydimethylsiloxane adsorbed on fumed metal
oxide particles of a wide range of specific surface area. Polymer. 2015. 77: 10.
[271] Gun’ko V.M., Turov V.V., Turova A.A., Krupska T.V., Pissis P., Leboda R., Skubiszewska-Zięba J.
Interactions of poly(dimethylsiloxane) with nanosilica and silica gel upon cooling-heating. J. Colloid
Interface Sci. 2014. 426: 48.
[272] Dollase T., Wilhelm M., Spiess H.W., Yagen Y., Yerushalmi-Rozen R., Gottlieb M. Effect of interfaces
on the crystallization behavior of PDMS. Interface Sci. 2003. 11(2): 199.
[273] Lehn J.-M., Supramolecular Chemistry. (Weinheim: VCH Verlagsgessellschaft mbH, 1995).
[274] Bershtein V.A., Egorov V.M. Differential Scanning Calorimetry of Polymers. Physics, Chemistry,
Analysis, Technology. (NY: Ellis Horwood, 1994).
263
[275] Bershtein V., Gun’ko V., Egorova L., Guzenko N., Pakhlov E., Ryzhov V., Zarko V. I. Well-defined
silica core – poly(vinyl pyrrolidone) shell nanoparticles: interactions and multi-modal glass transition
dynamics at interfaces. Polymer. 2009. 50(3): 860.
[276] Bershtein V.A., Gun’ko V.M., Karabanova L.V., Sukhanova T.E., Yakushev P.N., Egorova L.M.,
Glievyy O.B., Lutsyk E.D., Pakhlov E.M., Turova A.A., Zarko V.I., Vylegzhanina M.E. Hybrid
polyurethane-poly(2-hydroxyethyl methacrylate) semi-ipn-silica nanocomposites: interfacial
interactions and glass transition dynamics. J. Macromol. Sc.i Part B: Phys. 2010. 49(1): 18.
[277] Bershtein V.A., Gun’ko V.M., Egorova L.M., Guzenko N.V., Pakhlov E.M., Ryzhov V.A., Zarko V. I.
Well-defined oxide core-polymer shell nanoparticles: interfacial interactions, peculiar dynamics and
transitions in polymer nanolayers. Langmuir 2010. 26(13): 10968.
[278] Bershtein V.A., Gun’ko V.M., Egorova L.M., Wang Z., Illsley M., Voronin E.F., Prikhod'ko G.P.,
Yakushev P.N., Leboda R., Skubiszewska-Zięba J., Mikhalovsky S.V. Dynamics, thermal behaviour
and elastic properties of thin films of poly(vinyl alcohol) nanocomposites. RSC Advances 2012. 2(4):
1424.
[279] Bershtein V.A., Gun’ko V.M., Karabanova L.V., Sukhanova T.E., Yakushev P.N., Egorova L.M.,
Turova A., Zarko V.I., Pakhlov E.M., Vylegzhanina M.E., Mikhalovsky S.V. Polyurethane–poly(2-
hydroxyethyl methacrylate) semi-IPN–nanooxide composites. RSC Advances 2013. 3(34): 14560.
[280] Gun’ko V.M., Turov V.V., Barvinchenko V.N., Turova A.A., Rugal A.A., Zarko V.I., Leboda R.
Nonuniformity of starch/nanosilica composites and interfacial behaviour of water and organic
compounds. Appl. Surf. Sci. 2010. 256(17): 5275.
[281] Ma Y., Hu W., Reiter G. Lamellar crystal orientations biased by crystallization kinetics in polymer thin
films. Macromolecules 2006. 39(15): 5159.
[282] Hoffman, C. L. Rabolt, J. F. Self-assembled thin-film blends by polymer codeposition - poly(ethylene
oxide) and poly(methyl methacrylate). Macromolecules 1996. 29(7): 2543.
[283] Zhu J., Wang M. Phase ordering and crystallization kinetics in ultrathin poly(ethylene
oxide)/poly(methyl methacrylate) blend films. J. Macromol. Sci. Part B: Phys. 2008. 47(5): 1008.
[284] Wang H., Keum J.K., Hiltner A., Baer E. Confined crystallization of PEO in nanolayered films
impacting structure and oxygen permeability. Macromolecules 2009. 42(18): 7055.
[285] Wang H., Keum J.K., Hiltner A., Baer E., Freeman B., Rozanski A., Galeski A. Confined
crystallization of polyethylene oxide in nanolayer assemblies. Science 2009. 323(5915): 757.
[286] Martin J., Mijangos C.,Sanz A., Ezquerra T.A., Nogales A. Segmental dynamics of semicrystalline
poly(vinylidene fluoride) nanorods. Macromolecules 2009. 42: 5395.
[287] Klonos P., Pissis P., Gun’ko V.M., Kyritsis A., Guzenko N.V., Pakhlov E.M., Zarko V.I., Janusz W.,
Skubiszewska-Zięba J., Leboda R. Interaction of poly(ethylene glycol) with fumed silica and
alumina/silica/titania. Colloids Surf. A: Physicochem. Eng. Aspects 2010. 360(1-3): 220.
[288] Gun’ko V.M., Pissis P., Spanoudaki A., Turova A., Turov V.V., Zarko V.I., Goncharuk E.V. Interfacial
phenomena in starch/fumed silicaat varied hydration levels. Colloids Surf. A: Physicochem. Eng.
Aspects 2008. 320(1-3): 247.
[289] Hunter R.J. Zeta Potential in Colloid Sciences (London: Academic Press, 1981).
[290] Gun’ko V.M., Turov V.V., Zarko V.I., Dudnik V.V., Tischenko V.A., Voronin E.F., Siltchenko S.S.,
Barvinchenko V.N., Chuiko A.A. Aqueous suspensions of fumed silica and adsorption of proteins. J.
Colloid Interface Sci. 1997. 192(1): 166.
[291] Sorai M (editor), Comprehensive Handbook of Calorimetry and Thermal Analysis (West Sussex,
England: Wiley, 2004).
[292] Beaumont R.H., Clegg B., Gee G., Herbert J.B.M., Marks D.J., Roberts R.C., et al. Heat capacities of
propylene oxide and of some polymers of ethylene and propylene oxides. Polymer. 1966. 7(8): 401.
[293] Donth E. (editor), The Glass Transition: Relaxation Dynamics in Liquids and Disordered Materials.
Springer Series in Materials Science. vol 48 (Berlin: Springer, 2001).
[294] Fontanella J.J., Wintersgill M.C., Calame J.P., Andeen C.G. Electrical relaxation in pure and alkali
metal thiocyanate complexed poly(ethylene oxide). Solid State Ionics 1983. 8(4): 333.
[295] Kyritsis A., Pissis P., Tsonos C., Laudat J., Ren J. Dielectric and conductivity relaxation in dry and
humid solid PEO electrolytes. J. Non-Crystal Solid 1994. 172-174(2): 1431.
264
[296] Kripotou S., Pissis P., Sysel P., Sindelar V., Bershtein V.A. Structure-property relationships in novel
poly(imide-amide)-poly(ethylene glycol)hybrid networks. Polymer. 2006. 47(1): 357.
[297] Jia Z., Zhang K., Tan J., Han C., Dong L., Yang Y. Crystallization behavior and mechanical properties
of crosslinked plasticized poly(L-lactic acid). J. Appl. Polymer. Sci. 2009. 111(3): 1530.
[298] Dobbertin J., Hensel A., Schick C. Dielectric spectroscopy and calorimetry in the glass transition region
of semi-crystalline poly(ethylene terephthalate). J. Thermal Analys Calorimetry 1996. 47(4): 1027.
[299] Fragiadakis D., Pissis P. Glass transition and segmental dynamics in poly (dimethylsiloxane)/silica
nanocomposites studied by various techniques. J. Non-Crystal Solid 2007. 353(47-51): 4344.
[300] Bershtein V.A., Egorova L.M., Yakushev P.N., Pissis P., Sysel P., Brozova L. Molecular dynamics in
nanostructured polyimide–silica hybrid materials and their thermal stability. J. Polymer. Sci. B:
Polymer. Phys. 2002. 40(10): 1056.
[301] Gómez Tejedor J.A., Rodriguez Hernández J.C., Gómez Ribelles J.L., Monleón Pradas M. Dynamic
mechanical relaxation of poly(2- hydroxyethyl acrylate)-silica nanocomposites obtained by the sol– gel
method. J. Macromol. Sci. B: Phys. 2007. 46(1): 43.
[302] Sargsyan A., Tonoyan A., Davtyan S., Schick Ch. The amount of immobilized polymer in PMMA SiO2
nanocomposites determined from calorimetric data. Europ. Polymer. J. 2007. 43(8): 3113.
[303] Aranguren M.I. Crystallization of polydimethylsiloxane: effect of silica filler and curing. Polymer.
1998. 39(20): 4897.
[304] Fragiadakis D., Pissis P., Bokobza L. Glass transition and molecular dynamics in poly
(dimethylsiloxane)/silica nanocomposites. Polymer. 2005. 46(16): 6001.
[305] Kalogeras I.M., Roussos M., Christakis I., Spanoudaki A., Pietkiewicz D., Brostow W., Vassilikou-
Dova A. Dielectric properties of cured epoxy resin + poly(ethylene oxide) blends. J. Non-Crystal Solid
2005. 351: 2728.
[306] Hedvig P. Dielectric Spectroscopy of polymers. (Bristol: Adam Hilger, 1977).
[307] Brauenlich P. (editor), Thermally stimulated currents in solids. Topics in applied physics. vol. 37
(Berlin: Springer, 1979).
[308] Kremer F., Schönhals A. (eds), Broadband Dielectric Spectroscopy. (Berlin: Springer-Verlag, 2003).
[309] Donth E. (editor), The Glass Transition: Relaxation Dynamics in Liquids and Disordered Materials.
Springer Series in Materials Science. Vol 48 (Berlin: Springer, 2001).
[310] Capaccioli S., Ngai K.L., Shinyashiki N. The Johari-Goldstein-relaxation of water. J. Phys. Chem. B
2007. 111(28): 8197.
[311] Shinyashiki N., Sudo S., Yagihara S., Spanoudaki A., Kyritsis A., Pissis P. Relaxation processes of
water in the liquid to glassy states of water mixtures studied by broadband dielectric spectroscopy. J.
Phys.: Condens. Matter. 2007. 19(20): 205113.
[312] Osti N.C., Coté A., Mamontov E., Ramirez-cuesta A., Wesolowski D.J., Diallo S.O. Characteristic
features of water dynamics in restricted geometries investigated with quasi-elastic neutron scattering.
Chem. Phys. 2016. 465-466: 1.
[313] Tan S.P., Piri M. Equation-of-state modeling of confined fluid phase equilibria in nanopores. Fluid
Phase Equilibria. 2015. 393: 48.
[314] Nouri-Khorasani A., Malek K., Malek A., Mashio T., Wilkinson D.P., Eikerling M.H. Molecular
modeling of the proton density distribution in a water-filled slab-like nanopore bounded by Pt oxide and
ionomer. Catalysis Today 2016. 262: 133.
[315] Weiland E., Springuel-Huet M., Nossov A., Gédéon A. Xenon NMR: Review of recent insights into
porous materials. Micropor. Mesopor. Mater. 2016. 225: 41.
[316] Klonos P., Kaprinis S., Zarko V.I., Peoglos V., Pakhlov E.M., Pissis P., Gun'ko V.M. Thermal and
dielectric studies of PEG/C/AST nanocomposites. J. Appl. Polymer. Sci. 2013. 128(3): 1601.
[317] Stamatopoulou C., Klonos P., Bondaruk O., Koutsoumpis S., Gun’ko V., Pissis P., Karabanova L.
Hydrophilic nanocomposites based on polyurethane/poly(2-hydroxyethyl methacrylate) semi-IPNs and
modified/unmodified nanosilica for biomedical applications. J. Polymer Sci. Part B: Polymer Phys.
2014. 52(5): 397.
[318] Galaburda M.V., Klonos P., Gun’ko V.M., Bogatyrov V.M., Borysenko M.V., Pissis P. Dielectric
properties and thermal destruction of poly(dimethylsiloxane)/Fe2O3/SiO2 nanocomposites. Appl. Surf.
Sci. 2014. 305: 67.
265
[319] Sulym I.Ya., Klonos P., Borysenko M.V., Pissis P., Gun’ko V.M. Dielectric and thermal studies of
segmental dynamic in silica/PDMS and silica/titania/PDMS nanocomposites. J. Appl. Polymer. Sci.
2014. 131(23): 41154.
[320] Klonos P., Sulym I.Ya., Borysenko M.V., Gun’ko V.M., Kripotou S., Kyritsis A., Pissis P. Interfacial
interactions and complex segmental dynamics in systems based on silica-polydimethylsiloxane core-
shell nanoparticles: Dielectric and thermal study. Polymer. 2015. 58: 9.
[321] Klonos P., Sulym I.Ya., Kyriakos K., Vangelidis I., Zidropoulos S., Sternik D., Borysenko M.V.,
Kyritsis A., Deryło-Marczewska A., Gun'ko V.M., Pissis P. Interfacial phenomena in core–shell
nanocomposites of PDMS adsorbed onto low specific surface area fumed silica nanooxides: Effects of
surface modifications. Polymer. 2015. 68: 158.
[322] Klonos P., Dapei G., Sulym I.Ya., Zidropoulos S., Sternik D., Deryło-Marczewska A., Borysenko
M.V., Gun’ko V.M., Kyritsis A., Pissis P. Morphology and molecular dynamics investigation of PDMS
adsorbed on titania nanoparticles: effects of polymer molecular weight. Eur. Polymer. J. 2016. 74: 64.
[323] Leboda R., Turov V.V., Charmas B., Skubiszewska-Zięba J., Gun’ko V.M. Surface properties of
mesoporous carbon-silica gel adsorbents. J. Colloid Interface Sci. 2000. 223(1): 112.
[324] Gun’ko V.M., Leboda R., Skubiszewska-Zięba J., Turov V.V., Kowalczyk P. Structure of silica gel Si-
60 and pyrocarbon/silica gel adsorbents thermally and hydrothermally treated. Langmuir. 2001. 17(11):
3148.
[325] Melillo M., Gun’ko V.M., Mikhalovska L.I, Phillips G.J., Davies J.D., Lloyd A.W., Kozynchenko O.P.,
Malik D.J., Streat M., Mikhalovsky S.V. Structural characteristics of activated carbons and ibuprofen
adsorption affected by bovine serum albumin. Langmuir. 2004. 20(7): 2837.
[326] Tomaszewski W., Gun’ko V.M., Leboda R., Skubiszewska-Zięba J. Structural characteristics of
modified activated carbons and adsorption of explosives. J. Colloid Interface Sci. 2003. 266: 388.
[327] Ahmed W., Jackson M.J. (eds) Emerging Nanotechnologies for Manufacturing. Second Edition
(Amsterdam: Elsevier, 2015).
[328] Thomas S., Shanks R., Chandrasekharakurup S. (eds) Design and Applications of Nanostructured
Polymer Blends and Nanocomposite Systems. (Amsterdam: Elsevier, 2016).
[329] Zaleska A. Doped-TiO2: a review. Recent Patents on Engineering. 2008. 2(3): 157.
[330] Bagheri S., Shameli K., Hamid S.B.A. Synthesis and characterization of anatase titanium dioxide
nanoparticles using egg white solution via sol-gel method. J. Chem. 2013. 2013: 848205.
[331] Safaei-Naeini Y., Aminzare M., Golestani-Fard F., Khorasanizadeh F., Salahi E. Suspension stability of
titania nanoparticles studied by UV-vis spectroscopy method. Iranian Journal of Materials Science &
Engineering. 2012. 9(1): 62.
[332] Buzea C., Blandino I.I.P., Robbie K. Nanomaterials and nanoparticles: sources and toxicity. J.
Biointerphases. 2007. 2(4): 17.
[333] Gun'ko V.M., Turov V.V., Gorbik P.P. Water at the Interfaces. (Kiev: Naukova Dumka, 2009). [in
Russian].
[334] Turov V.V., Gun'ko V.M. Clustered Water and Ways of Its Applications. (Kiev: Naukova Dumka,
2011). [in Russian].
[335] Gun’ko V.M., Turov V.V., Zarko V.I., Goncharuk O.V., Pakhlov E.M., Skubiszewska-Zięba J.,
Blitz J.P. Interfacial phenomena at a surface of individual and complex fumed nanooxides. Adv. Colloid
Interface Sci. 2016. 235: 108.
[336] Gun’ko V.M., Turov V.V., Zarko V.I., Pakhlov E.M., Charmas B., Skubiszewska-Zięba J. Influence of
structural organization of silicas on interfacial phenomena. Colloids and Surfaces A: Physicochem. Eng.
Aspects. 2016. 492: 230.
[337] Gun’ko V.M., Zarko V.I., Goncharuk O.V., Matkovsky A.K., Remez O.S., Skubiszewska-Zięba J.,
Wojcik G., Walusiak B., Blitz J.P. Nature and morphology of fumed oxides and features of interfacial
phenomena. Appl. Surf. Sci. 2016. 366: 410.
[338] Gun’ko V.M., Turov V.V., Zarko V.I., Goncharuk O.V., Matkovsky A.K., Prykhod’ko G.P.,
Nychiporuk Yu.M., Pakhlov E.M., Krupska T.V., Balakin D.Yu., Charmas B., Andriyko L.S.,
Skubiszewska-Zięba J., Marynin A.I., Ukrainets A.I., Kartel M.T. Multi-layer graphene oxide alone and
in a composite with nanosilica: preparation and interactions with polar and nonpolar adsorbates. Appl.
Surf. Sci. 2016. 387: 736.
266
[339] Gun’ko V.M., Turov V.V., Pakhlov E.M., Matkovsky A.K., Krupska T.V., Kartel M.T., Charmas B.
Blends of amorphous/crystalline nanoalumina and hydrophobic amorphous nanosilica. Journal of Non-
Crystalline Solids. 2018. 500: 351.
[340] Gun’ko V.M., Pakhlov E.M., Skubiszewska-Zięba J., Blitz J.P. Infrared spectroscopy as a tool for
textural and structural characterization of individual and complex fumed oxides. Vibrational
Spectroscopy. 2017. 88: 56.
[341] Zaulychnyy Ya.V., Gun’ko V.M., Yavorskyi Y.V., Gasyuk I.M., Wanderka N., Dudka O.I. Effect of
mechanical treatment on the distribution of valence electrons and characteristics of nanocomposite
(SiO2)x(Al2O3)1-x (x = 0.8, x= 0.7) electrodes in lithium power sources. Applied Surface Sci. 2019. 494:
1013.https://doi.org/10.1016/j.apsusc.2019.07.206
[342] Gun’ko V.M., Pakhlov E.M., Goncharuk O.V., Andriyko L.S., Marynin A.I., Ukrainets A.I., Charmas
B., Skubiszewska-Zięba J., Blitz J.P. Influence of hydrophobization of fumed oxides on interactions
with polar and nonpolar adsorbates. Appl. Surf. Sci. 2017. 423: 855.
[343] Mironyuk I.F., Gun’ko V.M., Vasylyeva H.V., Goncharuk O.V., Tatarchuk H.V., Mandzyuk V.I.,
Bezruka N.A., Dmytrotsa T.V. Effects of enhanced clusterization of water at a surface of partially
silylated nanosilica on adsorption of cations and anions from aqueous media. Microporous and
Mesoporous Materials. 2019. 277: 95.https://doi.org/10.1016/j.micromeso.2018.10.016
[344] Gun’ko V.M., Turov V.V., Krupska T.V., Protsak I.S., Borysenko M.V., Pakhlov E.M.
Polymethylsiloxane alone and in composition with nanosilica under various conditions, J. Colloid
Interface Sci. 2019. 541: 213.
[345] Protsak I., Gun’ko V.M., Henderson I.M., Pakhlov E.M., Sternik D., Le Z. Nanostructured amorphous
silicas hydrophobized by various pathways. ACS Omega. 2019. 4(9): 13863.
[346] Skwarek E., Janusz W., Gun’ko V.M., Pakhlov E.M., Zarko V.I., Gdula K. Characteristics of surface
and electrochemical properties of composites with fumed metal oxides and hydroxyapatite. Adsorption.
2016. 22(4-6): 725.
[347] Skwarek E., Goncharuk O.V., Janusz W., Pakhlov E.M., Gun’ko V.M. Heats of immersion of
hydroxyapatite and hydroxyapatite/fumed oxides composites in water and n-decane. Materials
Chemistry and Physics 2018. 215: 99.
[348] Wawrzkiewicz M., Wiśniewska M., Gun’ko V.M. Application of silica–alumina oxides of different
compositions for removal of C.I. Reactive Black 5 dye from wastewaters. Adsorption Science &
Technology. 2017. 35(5–6): 448.
[349] Wawrzkiewicz M., Wiśniewska M., WołowiczA., Gun’ko V.M., Zarko V.I. Mixed silica-alumina oxide
as sorbent for dyes and metal ions removal from aqueous solutions and wastewaters. Microporous and
Mesoporous Materials. 2017. 250: 128.
[350] Klonos P., Dapei G., Sulym I.Ya., Zidropoulos S., Sternik D., Deryło-Marczewska A., Borysenko
M.V., Gun’ko V.M., Kyritsis A., Pissis P. Morphology and molecular dynamics investigation of PDMS
adsorbed on titania nanoparticles: effects of polymer molecular weight. European Polymer Journal.
2016. 74: 64.
[351] Klonos P., Kulyk K., Borysenko M., Gun'ko V., Kyritsis A., Pissis P. Effects of molecular weight
below the entanglement threshold on interfacial nanoparticles/polymer dynamics. Macromolecules.
2016. 49(24): 9457.
[352] Klonos P., KyritsisA., BokobzaL., Gun’ko V.M., Pissis P. Interfacial effects in PDMS/titania
nanocomposites studied by thermal and dielectric techniques. Colloids and Surfaces A: Physicochem.
Eng. Aspects. 2017. 519: 212.
[353] Klonos P., Goncharuk O., PakhlovE., Sternik D., Derylo-Marczewska A., Kyritsis A., Gun'ko V.,
Pissis P. Morphology, Molecular Dynamics, and Interfacial Phenomena in Systems Based on Silica
Modified by Grafting Polydimethylsiloxane Chains and Physically Adsorbed Polydimethylsiloxane.
Macromolecules. 2019. 52(7): 2863.
[354] Gun’ko V.M., Lupascu T., Krupska T.V., Golovan A.P., Pakhlov E.M., Turov V.V. Influence of tannin
on aqueous layers at a surface of hydrophilic and hydrophobic nanosilicas. Colloids and Surfaces A:
Physicochem. Eng. Aspects. 2017. 531: 9.
[355] Nosach L.V., Voronin E.F., Pakhlov E.M., Charmas B., Skubiszewska-Zięba J., Skwarek E., Janusz W.,
Gun’ko V.M. Nano-particulate structures with glucose-derived char and compacted fumed silica in
267
gaseous and aqueous media. 2017. 195: 729. In: Nanophysics, Nanomaterials, Interface Studies, and
Applications. NANO 2016. Springer Proceedings in Physics. Fesenko O., Yatsenko L. (eds) (Springer,
Cham).
[356] Gun’ko V.M., Turov V.V., Krupska T.V., Pakhlov E.M. Behavior of water and methane bound to
hydrophilic and hydrophobic nanosilicas and their mixture. Chem. Phys. Lett. 2017. 690: 25.
[357] Goncharuk O.V., Andriyko L.S., Malysheva M.L., Korotych O.I., Marynin A.I., Zarko V.I.,
Ukrainets A.I., Gun’ko V.M. Influence of indifferent electrolytes on the coagulative structure formation
in aqueous silica dispersions. French-Ukrainian Journal of Chemistry. 2017. 5(2): 40.
[358] Goncharuk O.V., Gun’ko V.M., Ugnivenko A.P., Terpilowski K., Skwarek E., Janusz W. Effect of
ethonium adsorption on structure formation in nanosilica dispersions. Nano Research & Applications.
2017. 29(3): 1.
[359] Voronina O.E., Malysheva M.L., Nosach L.V., Voronin E.F., Gun’ko V.M., Charmas B.,
Skubiszewska-Zięba J. A role of free silanol groups of nanosilica surface in interaction with poly(vinyl
pyrrolidone). Annales Universitatis Mariae Curie-Sklodowska, Sectio AA Chemia 2017. LXXII(2): 51.
[360] Gun’ko V.M., Krupska T.V., Andriyko L.S., Klymenko N.Yu., Siora I.V., Novikova O.A., Marynin
A.I., Ukrainets A.I., Charmas B., Shekhunova S.B., Turov V.V. Bonding of doxorubicin to nanosilica
and human serum albumin in various media. J. Colloid Interface Sci. 2018. 513: 809.
[361] Gun’ko V.M., Turov V.V., Krupska T.V. Interfacial behavior of methane and organic solvents with low
freezing points upon interaction with hydrophilic and hydrophobic nanosilicas. Chemistry, Physics and
Technology of Surface 2018. 9(2): 107.
[362] Turov V.V., Gun’ko V.M., Pakhlov E.M., Krupska T.V., Tsapko M.D., Charmas B., Kartel M.T.
Influence of hydrophobic nanosilica and hydrophobic medium on water bound in hydrophilic
components of complex systems. Colloids and Surfaces A: Physicochem. Eng. Aspects. 2018. 552: 39.
[363] Gun’ko V.M., Pakhlov E.M., Goncharuk O.V., Andriyko L.S., Nychiporuk Yu.M., Balakin D.Yu.,
Sternik D., Derylo-Marczewska A. Nanosilica modified by polydimethylsiloxane depolymerized and
chemically bound to nanoparticles or physically bound to unmodified or modified surfaces: Structure
and interfacial phenomena. J. Colloid Interface Sci. 2018. 529: 273.
[364] Goncharuk O., Bogatyrov V., Kazakova O., Galaburda M., Oranska O., Skwarek E., Waniak-Nowicka
H., Janusz W., Gun’ko V. Silica-supported NixOy, ZnxOy and MnxOy nanocomposites: physicochemical
characteristics and interactions with water and n-decane. Bulletin of Materials Science. 2019. 42: 243.
[365] Nosach L.V., Voronin E.F., Pakhlov E.M., Guzenko N.V., Gun’ko V.M. Polymer modified nanosilica
as a sorbent for medical applications, chapter 7. In: Biocompatible Hybrid Oxide Nanoparticles for
Human Health, From Synthesis to Applications, Micro and Nano Technologies. Melnyk I.V.,
Vaclavikova M., Seisenbaeva G.A., Kessler V.G. (Eds.) (Amsterdam: Elsevier, 2019)
[366] Gun’ko V.M., Polymer composites with functionalized silica. In: Polymer Composites with
Functionalized Nanoparticles. Synthesis, Interactions, Properties and Applications. Majka T.,
Pielichowski K. (Eds.) (Amsterdam: Elsevier, 2019).
[367] Protsak I., Gun’ko V.M., Turov V.V., Krupska T.V., Pakhlov E.M., Zhang D., Dong W., Le Z.
Nanostructured polymethylsilosane/fumed silica blends. Materials 2019. 2409(12): 1.
[368] Turov V.V., Gun’ko V.M., Krupska T.V., Protsak I.S., Pakhlov E.M., Structural and adsorption features
of amorphous nanosilica modified by various addition of polymethylsiloxane. Chemistry, Physics and
Technology of Surface. 2019. 10(3): 203.
[369] Gun’ko V.M., Turov V.V., Pakhlov E.M., Krupska T.V., Charmas B. Effect of water content on the
characteristics of hydro-compacted nanosilica. Appl. Surf. Sci. 2018. 459: 171.
[370] Gun’ko V.M., Turov V.V., Pakhlov E.M., Krupska T.V., Borysenko M.V., Kartel M.T., Charmas B.
Water interactions with hydrophobic versus hydrophilic nanosilica. Langmuir. 2018. 34(40): 12145.
[371] Turov V.V., Gun’ko V.M., Krupska T.V., Kartel M.T. Influence of solid and liquid hydrophobic
compounds on characteristics of water located in an adsorption layer of a hydrophilic component of the
system. Chemistry, Physics and Technology of Surface. 2018. 9(4): 341.
[372] Gun’ko V.M., Turov V.V., Protsak I.S., Krupska T.V., Pakhlov E.M., Tsapko M.D. Effects of pre-
adsorbed water on methane adsorption onto blends with hydrophobic and hydrophilic nanosilicas.
Colloids and Surfaces A: Physicochem. Eng. Aspects. 2019. 570: 471.
268
[373] Pujari, P.K., Sen, D., Amarendra, G., Abhaya S., Pandey A.K., Dutta D., Mazumder S. Study of pore
structure in grafted polymer membranes using slow positron beam and small-angle X-ray scattering
techniques. Nuclear. Instr. Method Phys. Res. B. 2007. 254: 278.
[374] Sakurai S., SAXS evaluation of size distribution for nanoparticles. Chapter 5. In: A.E. Ares (ed.), X-ray
Scattering, InTech, Croatia, 2017, pp. 107-134.
[375] Brumberger H. (Ed.) Small Angle X-ray Scattering. Gordon & Breach, New York, Syracuse, 1965.
[376] Dieudonné Ph., Hafidi A.A., Delord P., Phalippou J. Transformation of nanostructure of silica gels
during drying. J. Non-Crystal. Solid. 2000. 262: 155.
[377] Fairén-Jiménez D., Carrasco-Marín F., Djurado D., Bley F., Ehrburger-Dolle F., Moreno-Castilla C.
Surface area and microporosity of carbon aerogels from gas adsorption and small- and wide-angle X-ray
scattering measurements. J. Phys. Chem. B 2006. 110: 8681.
ЯВИЩА НА МЕЖАХ ПОДІЛУ БІЛЯ ПОВЕРХНІ
ІНДИВІДУАЛЬНИХ ТА СКЛАДНИХ ПІРОГЕННИХ
НАНООКСИДІВ
В.М. Гунько, В.В. Туров, О.В. Гончарук, Є.М. Пахлов, О.К. Матковський
Інститут хімії поверхні ім. О.О. Чуйка Національної академії наук України
вул. Генерала Наумова, 17, Київ, 03164, Україна, e-mail:vlad_gunko@ukr.net
Мета цього огляду проаналізовати дослідження поведінки на границях поділу та в
залежності від температури неполярних та полярних адсорбатів, що взаємодіють з
індивідуальними та складними пірогенними оксидами металів та металоїдів (ПОМ),
вихідними та тренованими чи хімічно модифікованими, у порівнянні з пористими
силікагелями, осадженими кремнеземами, композитами тощо. Комплексні ПОМ можуть
представляти собою частинки ядро-оболонка (ЧЯО, розміром 50-200 нм) з ядром з TiO2 чи
Al2O3 та оболонкою з SiO2 чи Al2O3 на відміну від простих та менших наночастинок
індивідуальних ПОМ. ЧЯО можуть бути зруйнованими при кріожелюванні при високому
тиску чи при механохімічній обробці. Ці тренування, як і гідроущільнення (контрольоване
змочування та сушка) впливають на будову агрегатів наночастинок та агломератів з
агрегатів, які стають більш компактними. Аналіз вказує на те, що складні ПОМ можуть
бути більш чутливими до різних зовнішніх впливів, ніж прості ПОМ, як нанокремнезем. Любе
тренування «м’яких» ПОМ впливає на міжфазну та температурну поведінку полярних та
неполярних адсорбатів. Перебудова вторинних частинок та поверхнева функціоналізація
впливають на зсув точки замерзання-розморожування адсорбатів, локалізованих у порах. Для
деяких адсорбатів спостерігається відкрита петля гістерезису адсорбції-десорбції.
Кластеризація адсорбатів, локалізовиних у порах, призводить до зменшення змін ентальпії
при фазових переходах (замерзання, плавлення). Зсув точки замерзання та плавлення
призводить до суттєвих гістерезисних ефектів при замерзанні-розмерзанні адсорбатів, що
локалізовани у текстурних порах вихідних та тренованих ПОМ. Релаксаційні явища як для
низькомолекулярних, так і високомолекулярних адсорбатів чи полімерних композитів
залежать від морфології первинних частинок, структурної організації вторинних частинок
ПОМ, тренованих чи модифікованих різним чином, вмісту адсорбатів, порядку ко-адсорбції,
температури тощо.
Ключові слова: нанокремнезем, складні нанооксиди, міжфазні явища, адсорбція,
випаровування, ефекти замкнутого простору.
269
ЯВИЛЕНИЯ НА ГРАНИЦАХ РАЗДЕЛА У ПОВЕРХНОСТИ
ИНДИВИДУАЛЬНЫХ И СЛОЖНЫХ ПИРОГЕННЫХ
НАНООКСИДОВ
В.М. Гунько, В.В. Туров, Е.В. Гончарук, Е.М. Пахлов, А.К. Матковский
Институт химии поверхности им. А.А. Чуйко Национальной академии наук Украины
ул. Генерала Наумова, 17, Киев, 03164, Украина, e-mail: vlad_gunko@ukr.net
Цель этого обзора проанализировать исследования поведения на границах раздела и в
зависимости от температуры неполярных и полярных адсорбатов, которые
взаимодействуют с индивидуальными и сложными пирогенными оксидами металлов и
металлоидов (ПОМ), исходными и тренированными или химически модифицированными, по
сравнению с пористыми силикагелями, осажденными кремнеземами, композитами и др.
Комплексные ПОМ могут представлять собою частицы ядро-оболочка (ЧЯО, размерами 50-
200 нм) с ядром из TiO2 или Al2O3 и оболочкой из SiO2 или Al2O3 в отличии от простых и
меньших наночастиц индивидуальных ПОМ. ЧЯО могут быть разрушены при
криожелировании при высоком давлении или при механохимической обработке. Такая
подготовка, как и гидроуплотнение (контролируемое смачивание и сушка) влияют на
строение агрегатов наночастиц и агломератов из агрегатов, которые становятся более
компактными. Анализ показывает, что сложные ПОМ могут быть более чувствительными
к различным внешним воздействия, чем простые ПОМ, как нанокремнезем. Любая
подготовка «мягких» ПОМ влияет на межфазное и температурное поведение полярных и
неполярных адсорбатов. Перестройка вторичных частиц и поверхностное модифицирование
влияют на сдвиг точки замерзания-плавления адсорбатов, локализованных в порах. Для
некоторых адсорбатов наблюдается открытая петля гистерезиса адсорбции-десорбции.
Кластеризация адсорбатов, локализованных в порах, приводит к уменьшению изменений
энтальпии при фазовых переходах (замерзание, плавление). Сдвиг точки замерзания и
плавления приводит к существенным гистерезисным эффектам при замерзании-плавлении
адсорбатов, локализованных в текстурных порах исходных и тренированных ПОМ.
Релаксационные явления как для низкомолекулярных, так и высокомолекулярных адсорбатов
или полимерных композитов зависят от морфологии первичных частиц, структурной
организации вторичных частиц ПОМ, тренированных или модифицированных различным
образом, концентрации адсорбатов, порядка ко-адсорбции, температуры и т.д.
Ключевые слова: нанокремнезем, сложные нанооксиды, межфазные явления, адсорбция,
испарение, эффекты замкнутого пространства.
3-38
39
40_55
56
57-102
103
104-113
114
115_129
130
131-178
179
180_195
196
197-226
227
228_269
|
| id | oai:ojs.pkp.sfu.ca:article-675 |
| institution | Surface |
| keywords_txt_mv | keywords |
| language | English |
| last_indexed | 2026-03-12T17:17:04Z |
| publishDate | 2019 |
| publisher | Chuiko Institute of Surface Chemistry National Academy of Sciences of Ukraine |
| record_format | ojs |
| resource_txt_mv | surfacezbircomua/34/a959b9f1c2d78de81bd3545c8f54fb34.pdf |
| spelling | oai:ojs.pkp.sfu.ca:article-6752020-01-28T14:32:15Z Interfacial phenomena at a surface of individual and complex fumed nanooxides Явиления на границах раздела у поверхности индивидуальных и сложных пирогенных нанооксидов Явища на межах поділу біля поверхні індивідуальних та складних пірогенних нанооксидів Гунько, В. М. Туров, В. В. Гончарук, О. В. Пахлов, Є. М. Матковський, О. К. Nanosilica Complex nanooxides Interfacial phenomena Adsorption Evaporation Confined space effects The aim of this review paper was to analyze investigation results on interfacial and temperature behaviors of nonpolar and polar adsorbates interacting with individual and complex fumed metal or metalloid oxides (FMO), initial and subjected to various treatments or chemical functionalization and compared to such porous adsorbents as silica gels, precipitated silica, mesoporous ordered silicas, and polymeric composites. Note that the particulate morphology of FMO depends strongly not only the flame reaction conditions but also on the types and amounts of reagents, as well their distribution in the flame. Therefore, complex nanooxides can include core-shell nanoparticles, (CSNP) of 50-200 nm in size with titania or alumina cores and silica or alumina shells in contrast to simple and smaller nanoparticles of individual FMO. CSNP could be destroyed under high-pressure cryogelation (HPCG) or mechanochemical activation (MCA). These treatments as well simple hydrocompaction (controlled wetting-drying) affect the structure of aggregates of nanoparticles and agglomerates of aggregates, resulting in their becoming more compacted. The analysis shows that complex FMO could be more sensitive to external actions than simple nanooxides such as fumed silica. Any treatment of ‘soft’ FMO affects the interfacial and temperature behaviors of polar and nonpolar adsorbates. Rearrangement of secondary particles and surface functionalization affect the freezing-melting point depression of adsorbates. For some adsorbates, open hysteresis loops became readily apparent in adsorption-desorption isotherms. Clustering of adsorbates bound in pores causes reduced changes in enthalpy during phase transitions (freezing, fusion). Freezing point depression and melting point elevation cause significant hysteresis freezing-melting effects for adsorbates bound to initial and treated FMO in textural pores (voids between nanoparticles in secondary structures). Relaxation phenomena for both low- and high-molecular weight adsorbates or of filled polymeric composites are affected by the morphology of primary particles, structural organization of secondary particles of differently treated or functionalized FMO, content of adsorbates, co-adsorption order, and temperature. Цель этого обзора проанализировать исследования поведения на границах раздела и в зависимости от температуры неполярных и полярных адсорбатов, которые взаимодействуют с индивидуальными и сложными пирогенными оксидами металлов и металлоидов (ПОМ), исходными и тренированными или химически модифицированными, по сравнению с пористыми силикагелями, осажденными кремнеземами, композитами и др. Комплексные ПОМ могут представлять собою частицы ядро-оболочка (ЧЯО, размерами 50-200 нм) с ядром из TiO2 или Al2O3 и оболочкой из SiO2 или Al2O3 в отличии от простых и меньших наночастиц индивидуальных ПОМ. ЧЯО могут быть разрушены при криожелировании при высоком давлении или при механохимической обработке. Такая подготовка, как и гидроуплотнение (контролируемое смачивание и сушка) влияют на строение агрегатов наночастиц и агломератов из агрегатов, которые становятся более компактными. Анализ показывает, что сложные ПОМ могут быть более чувствительными к различным внешним воздействия, чем простые ПОМ, как нанокремнезем. Любая подготовка «мягких» ПОМ влияет на межфазное и температурное поведение полярных и неполярных адсорбатов. Перестройка вторичных частиц и поверхностное модифицирование влияют на сдвиг точки замерзания-плавления адсорбатов, локализованных в порах. Для некоторых адсорбатов наблюдается открытая петля гистерезиса адсорбции-десорбции. Кластеризация адсорбатов, локализованных в порах, приводит к уменьшению изменений энтальпии при фазовых переходах (замерзание, плавление). Сдвиг точки замерзания и плавления приводит к существенным гистерезисным эффектам при замерзании-плавлении адсорбатов, локализованных в текстурных порах исходных и тренированных ПОМ. Релаксационные явления как для низкомолекулярных, так и высокомолекулярных адсорбатов или полимерных композитов зависят от морфологии первичных частиц, структурной организации вторичных частиц ПОМ, тренированных или модифицированных различным образом, концентрации адсорбатов, порядка ко-адсорбции, температуры и т.д. Мета цього огляду проаналізовати дослідження поведінки на границях поділу та в залежності від температури неполярних та полярних адсорбатів, що взаємодіють з індивідуальними та складними пірогенними оксидами металів та металоїдів (ПОМ), вихідними та тренованими чи хімічно модифікованими, у порівнянні з пористими силікагелями, осадженими кремнеземами, композитами тощо. Комплексні ПОМ можуть представляти собою частинки ядро-оболонка (ЧЯО, розміром 50-200 нм) з ядром з TiO2 чи Al2O3 та оболонкою з SiO2 чи Al2O3 на відміну від простих та менших наночастинок індивідуальних ПОМ. ЧЯО можуть бути зруйнованими при кріожелюванні при високому тиску чи при механохімічній обробці. Ці тренування, як і гідроущільнення (контрольоване змочування та сушка) впливають на будову агрегатів наночастинок та агломератів з агрегатів, які стають більш компактними. Аналіз вказує на те, що складні ПОМ можуть бути більш чутливими до різних зовнішніх впливів, ніж прості ПОМ, як нанокремнезем. Любе тренування «м’яких» ПОМ впливає на міжфазну та температурну поведінку полярних та неполярних адсорбатів. Перебудова вторинних частинок та поверхнева функціоналізація впливають на зсув точки замерзання-розморожування адсорбатів, локалізованих у порах. Для деяких адсорбатів спостерігається відкрита петля гістерезису адсорбції-десорбції. Кластеризація адсорбатів, локалізовиних у порах, призводить до зменшення змін ентальпії при фазових переходах (замерзання, плавлення). Зсув точки замерзання та плавлення призводить до суттєвих гістерезисних ефектів при замерзанні-розмерзанні адсорбатів, що локалізовани у текстурних порах вихідних та тренованих ПОМ. Релаксаційні явища як для низькомолекулярних, так і високомолекулярних адсорбатів чи полімерних композитів залежать від морфології первинних частинок, структурної організації вторинних частинок ПОМ, тренованих чи модифікованих різним чином, вмісту адсорбатів, порядку ко-адсорбції, температури тощо.&nbsp; &nbsp; Chuiko Institute of Surface Chemistry National Academy of Sciences of Ukraine 2019-10-30 Article Article application/pdf https://surfacezbir.com.ua/index.php/surface/article/view/675 10.15407/Surface.2019.11.003 Surface; No. 11(26) (2019): Surface; 3-269 Поверхность; № 11(26) (2019): Поверхность; 3-269 Поверхня; № 11(26) (2019): Поверхня; 3-269 3154-8091 3154-8083 10.15407/Surface.2019.11 en https://surfacezbir.com.ua/index.php/surface/article/view/675/677 Авторське право (c) 2019 В.М. Гунько, В.В. Туров, О.В. Гончарук, Є.М. Пахлов, О.К. Матковський |
| spellingShingle | Гунько, В. М. Туров, В. В. Гончарук, О. В. Пахлов, Є. М. Матковський, О. К. Явища на межах поділу біля поверхні індивідуальних та складних пірогенних нанооксидів |
| title | Явища на межах поділу біля поверхні індивідуальних та складних пірогенних нанооксидів |
| title_alt | Interfacial phenomena at a surface of individual and complex fumed nanooxides Явиления на границах раздела у поверхности индивидуальных и сложных пирогенных нанооксидов |
| title_full | Явища на межах поділу біля поверхні індивідуальних та складних пірогенних нанооксидів |
| title_fullStr | Явища на межах поділу біля поверхні індивідуальних та складних пірогенних нанооксидів |
| title_full_unstemmed | Явища на межах поділу біля поверхні індивідуальних та складних пірогенних нанооксидів |
| title_short | Явища на межах поділу біля поверхні індивідуальних та складних пірогенних нанооксидів |
| title_sort | явища на межах поділу біля поверхні індивідуальних та складних пірогенних нанооксидів |
| topic_facet | Nanosilica Complex nanooxides Interfacial phenomena Adsorption Evaporation Confined space effects |
| url | https://surfacezbir.com.ua/index.php/surface/article/view/675 |
| work_keys_str_mv | AT gunʹkovm interfacialphenomenaatasurfaceofindividualandcomplexfumednanooxides AT turovvv interfacialphenomenaatasurfaceofindividualandcomplexfumednanooxides AT gončarukov interfacialphenomenaatasurfaceofindividualandcomplexfumednanooxides AT pahlovêm interfacialphenomenaatasurfaceofindividualandcomplexfumednanooxides AT matkovsʹkijok interfacialphenomenaatasurfaceofindividualandcomplexfumednanooxides AT gunʹkovm âvileniânagranicahrazdelaupoverhnostiindividualʹnyhisložnyhpirogennyhnanooksidov AT turovvv âvileniânagranicahrazdelaupoverhnostiindividualʹnyhisložnyhpirogennyhnanooksidov AT gončarukov âvileniânagranicahrazdelaupoverhnostiindividualʹnyhisložnyhpirogennyhnanooksidov AT pahlovêm âvileniânagranicahrazdelaupoverhnostiindividualʹnyhisložnyhpirogennyhnanooksidov AT matkovsʹkijok âvileniânagranicahrazdelaupoverhnostiindividualʹnyhisložnyhpirogennyhnanooksidov AT gunʹkovm âviŝanamežahpodílubílâpoverhnííndivídualʹnihtaskladnihpírogennihnanooksidív AT turovvv âviŝanamežahpodílubílâpoverhnííndivídualʹnihtaskladnihpírogennihnanooksidív AT gončarukov âviŝanamežahpodílubílâpoverhnííndivídualʹnihtaskladnihpírogennihnanooksidív AT pahlovêm âviŝanamežahpodílubílâpoverhnííndivídualʹnihtaskladnihpírogennihnanooksidív AT matkovsʹkijok âviŝanamežahpodílubílâpoverhnííndivídualʹnihtaskladnihpírogennihnanooksidív |