Recent observation of anomalous superconducting behavior of ultrathin YBCO films and single crystals of cluster borides

We report the number of deviations from conventional behavior in superconducting properties of ultrathin (1–3 unit-cell (UC)) YBa₀₂Cu₃O₇₋x films sandwiched between semiconducting Pr₀,₆Y₀,₄Ba₂Cu₃O₇₋x layers and for single crystals of cluster superconductor: dodecaboride ZrB₁₂. We have found a qua...

Повний опис

Збережено в:
Бібліографічні деталі
Опубліковано в: :Физика низких температур
Дата:2006
Автор: Gasparov, V.A.
Формат: Стаття
Мова:English
Опубліковано: Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України 2006
Теми:
Онлайн доступ:https://nasplib.isofts.kiev.ua/handle/123456789/120350
Теги: Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Цитувати:Recent observation of anomalous superconducting behavior of ultrathin YBCO films and single crystals of cluster borides / V.A. Gasparov // Физика низких температур. — 2006. — Т. 32, № 8-9. — С. 1105–1114. — Бібліогр.: 33 назв. — англ.

Репозитарії

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id nasplib_isofts_kiev_ua-123456789-120350
record_format dspace
spelling Gasparov, V.A.
2017-06-11T18:25:14Z
2017-06-11T18:25:14Z
2006
Recent observation of anomalous superconducting behavior of ultrathin YBCO films and single crystals of cluster borides / V.A. Gasparov // Физика низких температур. — 2006. — Т. 32, № 8-9. — С. 1105–1114. — Бібліогр.: 33 назв. — англ.
0132-6414
PACS: 74.25.Nf, 74.72.Bk, 74.70.Ad, 72.15.Gd
https://nasplib.isofts.kiev.ua/handle/123456789/120350
We report the number of deviations from conventional behavior in superconducting properties of ultrathin (1–3 unit-cell (UC)) YBa₀₂Cu₃O₇₋x films sandwiched between semiconducting Pr₀,₆Y₀,₄Ba₂Cu₃O₇₋x layers and for single crystals of cluster superconductor: dodecaboride ZrB₁₂. We have found a quadratic temperature dependence of the kinetic inductance, Lk⁻¹(Т) , at low temperatures independent of frequency, with a break in slope at TBKT dc , the maximum of real part of conductance, ωσ₁(T), and a large shift of the break temperature and the maximum position to higher temperatures with increasing frequency ω. We obtain from these data the universal ratio T /Lk⁻¹ (Tdc ) = 25, 25, and 17 nHK for 1-, 2- and 3-UC films, respectively in close agreement with theoretical prediction for vortex—antivortex unbinding transition. Superfluid density of ZrB₁₂ displays unconventional temperature dependence with pronounced shoulder at T/Tc equal to 0.65. Contrary to conventional theories we found a linear temperature dependence of Hс₂(Т) from Tc down to 0.35 K. We suggest that both λ(T) and Hс₂(Т) dependencies can be explained by a two band BCS model with different superconducting gap and Tc.
We are grateful to V.F. Gantmakher, A. Hebard, M. Chan, R. Huguenin, C. Lobb, P. Martinoli, D. van der Marel, C. Rogers, D.J. Scalapino, J.-M. Triscone, T. Venkatesan, X.X. Xi for stimulating discussions. This work was partially supported by Russian Ministry of Industry, Science and Technology (MSh-2169.2003.2) and Russian Academy of Sciences Program: New Materials and Structures.
en
Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
Физика низких температур
К 100-летию со дня рождения Б.Г. Лазарева
Recent observation of anomalous superconducting behavior of ultrathin YBCO films and single crystals of cluster borides
Article
published earlier
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
title Recent observation of anomalous superconducting behavior of ultrathin YBCO films and single crystals of cluster borides
spellingShingle Recent observation of anomalous superconducting behavior of ultrathin YBCO films and single crystals of cluster borides
Gasparov, V.A.
К 100-летию со дня рождения Б.Г. Лазарева
title_short Recent observation of anomalous superconducting behavior of ultrathin YBCO films and single crystals of cluster borides
title_full Recent observation of anomalous superconducting behavior of ultrathin YBCO films and single crystals of cluster borides
title_fullStr Recent observation of anomalous superconducting behavior of ultrathin YBCO films and single crystals of cluster borides
title_full_unstemmed Recent observation of anomalous superconducting behavior of ultrathin YBCO films and single crystals of cluster borides
title_sort recent observation of anomalous superconducting behavior of ultrathin ybco films and single crystals of cluster borides
author Gasparov, V.A.
author_facet Gasparov, V.A.
topic К 100-летию со дня рождения Б.Г. Лазарева
topic_facet К 100-летию со дня рождения Б.Г. Лазарева
publishDate 2006
language English
container_title Физика низких температур
publisher Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
format Article
description We report the number of deviations from conventional behavior in superconducting properties of ultrathin (1–3 unit-cell (UC)) YBa₀₂Cu₃O₇₋x films sandwiched between semiconducting Pr₀,₆Y₀,₄Ba₂Cu₃O₇₋x layers and for single crystals of cluster superconductor: dodecaboride ZrB₁₂. We have found a quadratic temperature dependence of the kinetic inductance, Lk⁻¹(Т) , at low temperatures independent of frequency, with a break in slope at TBKT dc , the maximum of real part of conductance, ωσ₁(T), and a large shift of the break temperature and the maximum position to higher temperatures with increasing frequency ω. We obtain from these data the universal ratio T /Lk⁻¹ (Tdc ) = 25, 25, and 17 nHK for 1-, 2- and 3-UC films, respectively in close agreement with theoretical prediction for vortex—antivortex unbinding transition. Superfluid density of ZrB₁₂ displays unconventional temperature dependence with pronounced shoulder at T/Tc equal to 0.65. Contrary to conventional theories we found a linear temperature dependence of Hс₂(Т) from Tc down to 0.35 K. We suggest that both λ(T) and Hс₂(Т) dependencies can be explained by a two band BCS model with different superconducting gap and Tc.
issn 0132-6414
url https://nasplib.isofts.kiev.ua/handle/123456789/120350
citation_txt Recent observation of anomalous superconducting behavior of ultrathin YBCO films and single crystals of cluster borides / V.A. Gasparov // Физика низких температур. — 2006. — Т. 32, № 8-9. — С. 1105–1114. — Бібліогр.: 33 назв. — англ.
work_keys_str_mv AT gasparovva recentobservationofanomaloussuperconductingbehaviorofultrathinybcofilmsandsinglecrystalsofclusterborides
first_indexed 2025-11-25T21:07:27Z
last_indexed 2025-11-25T21:07:27Z
_version_ 1850550735709667328
fulltext Fizika Nizkikh Temperatur, 2006, v. 32, Nos. 8/9, p. 1105–1114 Recent observation of anomalous superconducting behavior of ultrathin YBCO films and single crystals of cluster borides V.A. Gasparov Institute of Solid State Physics RAS, Chernogolovka 142432, Moscow Region, Russia E-mail: vgasparo@issp.ac.ru Received March 1, 2006, revised April 17, 2006 We report the number of deviations from conventional behavior in superconducting properties of ultrathin (1–3 unit-cell (UC)) YBa2Cu3O7–x films sandwiched between semiconducting Pr0.6Y0.4Ba2Cu3O7–x layers and for single crystals of cluster superconductor: dodecaboride ZrB12. We have found a quadratic temperature dependence of the kinetic inductance, L Tk �1( ), at low tem- peratures independent of frequency, with a break in slope at TBKT dc , the maximum of real part of conductance, ��1(T), and a large shift of the break temperature and the maximum position to higher temperatures with increasing frequency �. We obtain from these data the universal ratio TBKT dc /Lk �1 (TBKT dc ) = 25, 25, and 17 nHK for 1-, 2- and 3-UC films, respectively in close agreement with theoretical prediction for vortex—antivortex unbinding transition. Superfluid density of ZrB12 displays unconventional temperature dependence with pronounced shoulder at T/Tc equal to 0.65. Contrary to conventional theories we found a linear temperature dependence of H Tc2( ) from Tc down to 0.35 K. We suggest that both �(T) and H Tc2( ) dependencies can be explained by a two band BCS model with different superconducting gap and Tc. PACS: 74.25.Nf, 74.72.Bk, 74.70.Ad, 72.15.Gd Keywords: YBCO films, dodecaboride ZrB12, inductance, penetration depth, two band supercon- ductions. 1. Introduction The unusual superconducting behavior of high-Tc superconductors has stimulated efforts to test whether two-dimensional (2D) behavior is an essential aspect of these materials. The question of whether an isolated unit-cell thick layer could exhibit superconductivity and how his properties relate with bulk materials re- main controversial. From other side, it has been sug- gested from Berezinskii—Kosterlitz—Thouless (BKT) theory that there exist bound pairs of thermally ex- cited vortices and antivortices (with opposite circula- tion) in 2D layers below the TBKT and dissociated on free vortices and antivortices above [1,2]. Although many observations of the BKT transition in YBCO, BiSrCaCuO, and TlBaCaCuO compounds have been reported (see [3] and references therein), detailed comparison of the experimental data with the theory by Davis et al. [4] showed disagreements possibly due to inhomogeneity and vortex pinning. Rogers et al. reported that the usual BKT transition, i.e., all ther- mally activated vortices form vortex—antivortex pairs at temperatures below TBKT, was not observed in ultrathin Bi2Sr2Cu2O8 films from a low-frequency noise measurement due to vortex pinning [5]. Repaci et al. [6] showed from the study of dc I–V curves that free vortices exist at low temperatures even in one-unit-cell-thick YBCO films, indicating the ab- sence of the dc BKT transition. The binding energy between a vortex—antivortex pair U(r) � 1/r at r > �eff (here �eff = 2�2/d is the effective penetration depth and d is the film thickness), diverges at high distances r. Thus they had pointed out that a precon- dition [7] for the BKT transition to occur in a super- conductor, i.e., the sample size Ls< �eff, is not satis- fied even in YBCO films as thin as one unit cell. According to the BKT theory extended to finite frequencies [8–10], higher frequency currents sense © V.A. Gasparov, 2006 vortex—antivortex pairs of smaller separations. At high frequency, the electromagnetic response of a 2D superconductor is dominated by those bound pairs that have r � l�, where l� = (14D/�)1/2 is the vortex diffusion length and D is the vortex diffusion constant. Using the Bardeen—Stephen formula for free vortices [11], we estimate that l� < 1 �m at � > 10 MHz, which is much less than �eff = 40 �m for the 1-UC YBCO film [3]. This implies that it is possible to detect the response of vortex—antivortex pairs with short separation lengths at high frequencies in the samples even though the usual BKT transition is not present as shown in dc and low-frequency measurements. From other side, the recent discovery of super- conductivity in magnesium diboride [12] has initiated a substantial interest in potential «high–temperature» superconducting transition in other borides (see refe- rences in Refs. 13–16). Yet, only nonstoichiometric boride compounds (MoB2.5, NbB2.5, Mo2B, W2B, BeB2.75) demonstrate such transition. A potential clue to this contradiction may lay in the crystal structure of boron compounds, in particular in their cluster structure. Although it is widely accepted that the layered structure is crucial for high-Tc superconducti- vity, one can argue that clusters of light atoms are important for high Tc as well. In particular, there are a number of rather high-Tc superconductors among three-dimensional cluster compounds. Those are alkali metal doped C60 compounds (fullerides) Me3C60 (Me = K, Na, Rb, Cs) with the highest Tc up to 33 K for RbCs2C60 [17,18]. It is also known that boron atoms form clusters. These are octahedral B6 clusters in MeB6, icosahedral B12 clusters in �-rhombohedral boron, and cubo-octahedral B12 clusters in MeB12. The quest for superconductivity in these cluster compounds has a long history. Several superconduct- ing cubic hexaborides, MeB6, and dodecaborides, MeB12, have been discovered by Matthias et al. back in late 1960’s [19] (Me = Sc, Y, Zr, La, Lu, Th). Many other cluster borides (Me = Ce, Pr, Nd, Eu, Gd, Tb, Dy, Ho, Er, Tm) were found to be ferromagnetic or antiferromagnetic [19,20]. It was suggested that the superconductivity in YB6 and ZrB12 (Tc of 7.1 and 6.03 K, respectively) is exactly due to the effect of a cluster of light boron atoms. Clearly a systematic study of ZrB12 is needed to address the question of superconductivity in this compound. This has been the motivation for current systematic study of the temperature dependencies of the fre- quency and temperature dependences of the complex sheet conductance, �(�,T), of 1-UC to 3-UC thick YBCO films sandwiched between semiconducting Pr0.6Y0.4Ba2Cu3O7–x layers in a frequency range bet- ween 1 MHz to 30 GHz, as well as �(T) and upper critical magnetic field, H Tc2( ), in single crystals of ZrB12. We found a large increase of the transition temperature as a function of frequency for those films from 4 MHz to 30 GHz. We observe unusual super- conducting properties of ZrB12 and argue that these results can be reconciled by two-band superconducti- vity with different Tc. Currently these data were published in Refs. 3, 16, here we present most im- portant results of these studies. 2. Experimental setup Ultrathin YBCO layers sandwiched between 100 � buffer and 150 � cover layers of Pr0.6Y0.4Ba2Cu3O7–x were grown epitaxially on atomically flat and well- lattice-matched (100) LaAlO3 substrates using a multitarget pulsed-laser deposition (PLD) system [3]. For few UC thick films and the substrate used, the thicknesses are below the critical value for forming islands and the growth is in the range of layer by layer Stranski–Krastanov growth. The sample thickness is rather uniform due to the nature of the growth mode as characterized by using cross sectional transmission electron microscopy (TEM), atomic force microscope (AFM) and in situ RHEED. The three films thickness we examined were nominally 1-, 2-, and 3-UC thick and had the c axis normal to the film surface. The samples were made at different oxygen composition and therefore we will call them as S1 and S2 ones. The contacts were made at the edges of the 1�1 cm 3UC film for van der Pauw four-point resistance mea- surements. Under ambient conditions, dodecaboride ZrB12 crystallizes in the fcc structure of the UB12 type [15,16]. In this structure, the Zr atoms are located at interstitial openings among the close-packed B12 clusters. Our ZrB12 single crystals were grown using a floating-zone method [14]. The obtained single crystal ingots had a typical diameter of about 5 to 6 mm and a length of 40 mm. To assure good quality of our samples we performed metallographic and x-ray investigations of as grown ingots. We discovered that most parts of the ZrB12 ingot contained a needle like phase of nonsuperconducting ZrB2. We believe that ZrB2 needles are due to preparation of ZrB12 single crystals from a mixture of a certain amount of ZrB2 and an excess of boron [14]. We believe that unconventional properties observed from other studies may be due to these sample problems [16]. Therefore, special care has been taken to cut the samples from ZrB2 phase free parts. We used the spark erosion method to cut the single crystal ingots into rectangular <100> oriented bars of about 0.5�0.5�8 mm. The samples were lapped with 1106 Fizika Nizkikh Temperatur, 2006, v. 32, Nos. 8/9 V.A. Gasparov diamond paste and etched in hot nitrogen acid to remove any damage induced by lapping deteriorated surface layers. A well-defined geometry of the samples provided for the precise �(T) and superconducting properties measurements. Temperature was mea- sured with platinum (PT-103) and carbon glass (CGR-1-500) sensors. Magnetic measurements of �(T,H) and �(T,H) were carried out using a super- conducting coil in applied fields of up to 6 T down to 1.3 K. Additional dc and ac �(H) measurements were performed in the National High Magnetic Field La- boratory in Tallahassee, Florida (NHMFL) at tem- peratures down to 0.35 K. The dc magnetic field was applied in the direction of the current flow. For this study, two highly crystalline, supercon- ducting films of MgB2 were grown on an r-plane sapphire substrate in a two-step process [22]. De- position of boron precursor films via electron-beam evaporation was followed by ex situ postannealing at 890 �C in the presence of bulk MgB2 and Mg vapor. We investigated films of 500 and 700 nm thickness with corresponding Tc0’s of 38 K and 39.2 K, res- pectively. The details of the preparation technique are described elsewhere [22]. The �(�,T) at RF in thin 1–3 UC YBCO and MgB2 films was investigated employing a single coil mutual inductance technique. This technique, origi- nally proposed in Ref. 23 and lately improved in Ref. 24, has the advantages of the well known two-coil ge- ometry, and was extensively used for the study of the �(T) dependence for YBCO and MgB2 films [3,15,16]. In this radio frequency technique, the change of inductance L of a one-layer pancake coil located in the proximity of the film and connected in parallel with a capacitor C is measured. The LC cir- cuit is driven by the impedance meter (VM-508 TESLA) operating at 2–50 MHz, with a high fre- quency stability of 10 Hz. The film is placed at small distance (� 0.1 mm) below the coil. Both sample and coil are in a vacuum, but the coil holder is thermally connected with helium bath, while the sample holder is isolated and may be heated. During the experiment the coil was kept at 2.5 K, whereas the sample temper- ature was varied from 4.2 up to 100 K. Such design al- lows us to eliminate possible effects in temperature changes in L and C on the measurements. The complex mutual inductance M between the coil and the film can be obtained through Re ( ) ( ) M T L f f T � � � � � � � �0 0 2 2 1 , (1) Im ( ) ( ) ( ) ( ) ( ) M T f T C Z T Z T f T f � � � � � � � � � 1 2 1 1 2 0 2 0 2� . (2) Here L, Z(T), f(T), L0, Z0 and f0 are the inductance, impedance and the resonant frequency of the circuit with and without the sample, respectively. In the low frequency regime, where the coil wire diameter is much thinner than the skin depth at the working frequency, the expression of the variation of the M(T) (relative to the case where no sample is in the coil, M0), as a function of the �(T) may written as [24]: M T M q ql d l dq( ) ( ) � � � � � � � ���0 0 1 2 coth , (3) where M(q) plays the role of mutual inductance at a given wave number q in the film plane and depends on the sample-coil distance h, d is the sample thick- ness, and l is a complex length defined as l = = [1/(i��0�1+�2)]1/2, (more details can be found in Refs. 16, 24). A change in real, Re M(T), and imagi- nary, Im M(T), parts of M(T) were detected as a change of resonant frequency f(T) of the oscillating signal and impedance Z(T) of the LC circuit, and converted into Lk(T) and �1(T) by both, using Eqs. (1), (2) with inversion mathematical procedure and Eq. (3). The high frequency (100 MHz–1 GHz) measure- ments were performed using the cavity formed with a similar spiral coil with no capacity in parallel. The coil form the radio frequency resonator coupled to a two coupling loops and is driven by the radio-fre- quency signal generator/receiver. In this case the quality factor of the resonator Q and the resonance frequency were measured and converted to Re M(T) by Eq. (1) and to Im M(T) by: Im ( ) ( ) ( ) ( ) M T L f f T Q T f f T Q � � � � � � �0 0 0 0 1 1 , (4) where Q(T) and Q0(T) are the quality factors with and without the samples, respectively. The MW losses were measured using a resonant cavity technique with the gold-plated copper cylind- rical cavity operated in the TE011 mode at 29.9 GHz. The samples were mounted as a part of the bottom of the cavity through a thin gold-plated Cu-diaphragm with a small central hole so that the sample itself occupies only the holes part of the endplate through a transparent Teflon film. The resonator was operating in a transmission configuration. During the experi- mental run, we measured the amplitude of the trans- mitted signal at resonance as a function of tem- Recent observation of anomalous superconducting behavior of ultrathin YBCO films Fizika Nizkikh Temperatur, 2006, v. 32, Nos. 8/9 1107 perature. The resonator was coupled weakly to the input and output waveguides so the amplitude of the transmitted signal at resonance is proportional to the square of unloaded quality Q. We assume the MW electric field E in the film to be uniform along normal direction and equal to the E without the film. The �1(�,T) is thus proportional to Q–1(T) = Q–1(T) — – Q0 –1(T), where Q(T) and Q0(T) are the quality factors with and without the samples, respectively. We used similar radio frequency LC technique [15,16,25] to measure �(T) of ZrB12 samples. This technique employs a rectangular solenoid coil into which the sample is placed rather then use of spiral coil. The connection between parameters of the circuit and �(T) is described by following equation: � � �( ) ( ) ( ) ( ) ( ) ( ) T f T f f T fc � � � � � � � 0 0 0 2 2 2 2 . (5) Here � = 0.5(c2�/2��)1/2 is the imaginary part of a skin depth above Tc, which was determined from the �(T) measurements close to Tc, f(T) is the resonance frequency of the circuit at arbitrary T and f(Tc) and f(0) are the resonance frequency of the circuit at the superconducting transition and at zero temperature, respectively. 3. Magnetic penetration depth in YBCO films Figure 1 displays the Re M(T) and Im M(T) curves for a 2UC S2 film at three different frequencies from 3 MHz to 500 MHz as measured by different techniques: LC circuit and spiral coil resonator. The HF data are normalized to RF Re M(0) data because of different gap h values used. The most noticeable feature of these data is rather high shift of the onset point Tc0 of Re M(T) transition with frequency, not observed in such measurements on thick films. Notice also, that the inductive response, Re M(T), starts at lower tem- peratures than Im M(T), characterized by a peak close to transition, and this shift is raised with frequency. We have carried out the mutual inductance mea- surements on Pr0.6Y0.4Ba2Cu3O7–x films and observed no any features in the temperature dependences of the mutual inductance M(T). The Re M(T) and Im M(T) data are converted to L Tk �1( ) and Re �(T) using Eq. (3) and the mathemati- cal inversion procedure [3] based on the same ap- proach as in the two-coil mutual inductance method. Figure 2 shows the L Tk �1( ) curves in very low perpen- dicular magnetic fields, and zero field �Re �(T) for the 1-UC and 2-UC films (S1). We found that L Tk �1( ) fit well over a wide temperature range by a parabolic de- pendence: L Tk �1( ) = Lk �1(0)[1 – (T/Tc0)2], shown as thin solid lines in Fig. 2. We emphasize that this qua- dratics equation fit the data below characteristic tem- perature which we define as TBKT dc , and which is below the positions of the peaks in �Re �(T), which we de- fine as TBKT � . The mean field transition temperature, Tc0, determined by extrapolation of L Tk �1( ) to 0, is larger than the onset of transitions of L Tk �1( ), while is close to the onset point of �Re �(T) curves. Also, the Lk �1 0( ) fitted data are the same for H = 0 and 5 mT while have different Tc0. In Fig. 3, we plot �Re �(T) at 8 MHz and Q–1(T)/Q0 –1(4.2 K) determined from MW data (30 GHz) for 3-UC sample (S1). The dc resistive tran- sition of the same sample is also shown in the figure. According to the Coulomb gas scaling model, the re- 1108 Fizika Nizkikh Temperatur, 2006, v. 32, Nos. 8/9 V.A. Gasparov L C � , Z Fig. 1. The Re M(T) and Im M(T) curves for a 2-UC film (S2) at different frequencies, MHz: 3 (triangles), 26 (squares) and 500 (circles) calculated from raw f(T), Z(T) and Q(T) data. The solid lines describe a guide for the eye. Inset shows experimental set-up. . .. . Fig. 2. Temperature dependence of L Tk �1( ) for 1- and 2-UC films (S1) at 8 MHz and different magnetic fields, mT: 0 (circles), 2 (squares), 3 (triangles) and 4 (crosses). The solid lines curves shows �Re �(T) at zero field. The thin solid lines are quadratic fits to Lk �1 (T) below TBKT dc and for magnetic field data. Also shown is the theoretical BKT function (dashed line). sistance ratio R/Rn is proportional to the number of free vortices and should follow a universal function of an effective temperature scaling variable X = T Tc( 0 – TBKT dc )/TBKT dc (Tc0 – T) [8], which can be approximated by �/�n = C0X exp [–C1(X–1)–1/2] (here �n is the normal state resistivity, C0 = 1.7 and C1 = 4.9 are constants). We plot T/X(T) as a func- tion of T obtained by fitting �/�n data for the 3-UC film with this equation in the inset of Fig. 3. The best fit was found with Tc0 = 78.5 K (the point where T/X = 0) and TBKT dc = 56 K. Here TBKT dc represents a nomi- nal dc BKT transition temperature which is lower then TBKT � determined from RF and MW data. There are three major features observed in our RF and MW measurements: (i) the large frequency de- pendence of Tc(�), (ii) a foot jump in temperature dependence of the L Tk �1( ), which is destroyed in weak magnetic fields, and (iii) a maximum in �Re �(T) with the onset of transition at higher temperatures than that of the L Tk �1( ). Indeed, as we can see from Fig. 1, the TBKT � value, determined as the maximum position of losses, in- creases on 4 K (from 36.6 to 40.7 K) as the frequency raise from 3 to 500 MHz for 2-UC YBCO S2 film. The TBKT � value shift at 30 GHz is much larger: 74 K as compared with 61.5 K at 8 MHz (see Fig. 3). Even larger shift of TBKT � was observed in 1-UC film and no shift was detected in a 2000 � thick YBCO film. We determined the TBKT � as the peak position of �1(T) point. We also used the TBKT dc as the point where Lk �1 (T) deviates from the square fit (see Fig. 2), as was used by Hebard et al. [30]. The ratio TBKT dc /Lk �1(TBKT dc ) thus equal to 25, 25, and 17 nHK for 1-, 2- and 3-UC S1 films, respectively, is about constant, which however is larger than the theoreti- cal estimation TBKT dc �2(TBKT dc )/d = �0 2/32�2kB = = 0.98 cm·K (or TBKT dc /Lk �1(TBKT dc ) = 12.3 nHK) [31] ignoring small dynamic theory corrections of Ambe- gaokar et al. [9] to the renormalized coupling cons- tant KR. In order to see whether this assumption is correct, we plot theoretical BKT function L Tk �1( ) as dashed straight line on Fig. 2 derived from the universal relationship: K d c e k T R B BKT � � �2 2 2 216 2� (6) predicted by theory [4]. Notice however, that this theoretical dependence is valid for dc case, while the frequency dependence of TBKT � is obvious from the picture discussed above. This is why the critical temperatures determined from the intercept of dashed theoretical line with experimental L Tk �1( ) is lower then the peak position of �1(T). To see whether this description is correct, we plot the dependence of the penetration depth �–2(T) derived from L Tk �1( ), versus scaling variable — normalized temperature T/TBKT � on Fig. 4. It is obvious, that all data for three studied S1 films at 8 MHz fall on the same curve, which proof of our definition of TBKT � as the peak position of �1(T). The Abrikosov vortex lattice parameter av is the scale limiting the formation of vortex–antivortex pairs in magnetic field. We can estimate the field Hext which destroys the vortex pair unbinding from the following relation: l�� av = (�0/Hext) 1/2. This estimation gives Hext �1.5 mT at 10 MHz for 1-UC and 2-UC films in good agreement with the expe- rimental results (see Fig. 2). It is generally believed that the inhomogeneity can destroy the interaction of Recent observation of anomalous superconducting behavior of ultrathin YBCO films Fizika Nizkikh Temperatur, 2006, v. 32, Nos. 8/9 1109 � � � � m cm ; . .. . . . R e n H – 1 � Q /Q – 1 – 1 0 T dc T rf T mw BKT BKT BKT Fig. 3. Temperature dependence of dc �(T) (squares), �Re �(T) (circles) at 8 MHz and �Q–1 (T)/Q0 �1(4.2 K) at 30 GHz (solid line) for a 3-UC sample (S1). Inset shows the universal plot: T/X vs T. Arrows indicate the TBKT values determined from the dc resistivity, loss func- tion at 8 MHz and MW measurements (30 GHz). Pe ne tra tio n epd th , – – 2 2 � m Fig. 4. The penetration depth � –2(T) derived from Lk(T) = �0� 2/d vs normalized temperature T/TBKT � for: 1-UC (crosses), 2-UC (squares) and 3-UC (circles) films. vortex—antivortex pairs. However, inhomogeneity can never help to create the BKT transition. It is easy to see from above, we report several re- sults which are in qualitative agreement to the predic- tion of the dynamic theory of vortex—antivortex pairs with short separation lengths. These effects must be from intrinsic effect and can not be the result of inhomogeneity or due to the skin effect in the samples [3]. 4. Magnetic penetration depth in ZrB12 In the BCS theory the London penetration depth is identical with �(T) for specular and diffuse surface scattering and for negligible nonlocal effects. For a BCS-type superconductor with the conventional s-wave pairing form, the �(T) has an exponentially vanishing temperature dependence below Tc/2 (where �T) is almost constant): � � � ( ) ( ) ( ) exp ( ) T k T k TB B � � � �� � � �� � � � � � �0 1 0 2 0 (7) for clean limit: l > �, and � �( ) ( ) ( ) T k TB � �� � � �� 0 1 0 2 tanh (8) for dirty limit: l < � [13]. Here (0) is the energy gap and �(0) is the penetration depth at zero tem- perature. Close to Tc �(T) dependence has a BCS form [16]: � � ( ) ( ) T T Tc � � �� � � �� 0 2 1 . (9) Important problems for �(T) measurements are (i) determination of basic superconducting parameter �(0) and (ii) temperature dependence law, to see whether s-wave or d-wave pairing forms exist. Both these problems can be addressed from low-T �(T) dependence according to Eqs. (7) and (8). We used Eq. (5) to extrapolate the resonance frequency f(T) of our LC circuit down to zero temperature. To address the problem with �(0) we plot �(T) – �(0) data versus BCS reduced temperature: 1 2 1/ T/Tc( )� in the region close to Tc (see the inset to Fig. 5). The advantage of this procedure is the insensitivity of such analysis to the choice of f(0) on this temperature scale. The value of Tc = 5.992 K used in this data analysis is obtained by getting best linear fit of the �(T) – �(0) vs 1 2 1/ T/Tc( )� plot. Re- markably there is only a few millidegrees difference between Tc obtained from the fit and actual Tc0. We use the slope of �(T) – �(0) vs 1 2 1/ T/Tc( )� and Eq. (9) to obtain the value of �(0) = 143 nm. In order to investigate the temperature dependence of �(T) in the whole temperature region, in Fig. 6 we plot the superfluid density �2(0)/�2(T) versus the re- duced T/Tc for ZrB12 sample using this �(0) deter- mined from one gap fit close to Tc. One can easily no- tice an unconventional behavior of superfluid density with pronounced shoulder at T/Tc equal to 0.65. This feature can be explained by a model of two independ- ent BCS superconducting bands with different plasma frequencies, gaps and Tc’s [16]. We label these two bands as p- and d-bands in accordance to electron structure of ZrB12 [26]. It is clear from partial BCS 1110 Fizika Nizkikh Temperatur, 2006, v. 32, Nos. 8/9 V.A. Gasparov Fig. 5. Temperature variations of �(T) for ZrB12 sample. The inset shows �(T) – �(0) versus BCS reduced temperature. �2 �2 Fig. 6. Superfluid density, [�(0)/�(T)]2, of the ZrB12 single crystal sample and MgB2 thin film (crosses). The predicted behavior of [�(0)/�(T)]2 within the two band model as described in the text is shown by the solid, dashed (p-band) and dotted (d-band) lines. The solid line represent BCS fit of MgB2 data using 2�(0)/kBTc as fit parameter. curves, that single gap BCS model is unable to fit a full [�(0)/�(T)]2 unconventional dependence. For a dirty limit [27] we can write for two band superconductor in: 1 2 0 0 2 2 2� �( ) ( ) ( ) ( ) ( ) ( ) ( T T k T T kp p B p p d d � tanh tanh B d d T ) ( ) ( ) . �2 0 0 (10) Here i is the superconducting energy gap and �i(0) is residual penetration depth in the p- or d-band. Using this two gap �(T) BCS-like dependence and interpolation formula ( ) ( ) tanh [ .T �0 188 � ( – ) ]T /Tc /1 1 2 we fit the experimental data with six fitting parameters: �i, i and Tci . From this fit we obtain Tc p= 6.0 K, Tc d= 4.35 K, p(0) = 0.73 meV, d(0) = 1.21 meV, �p(0) = 170 nm and �d(0) = = 260 nm, for p- and d-bands, respectively. Dashed and dotted lines in Fig. 6 show the contributions of each p- and d-bands, respectively. Clearly low-tem- perature dependence of �2(0)/�2(T) is dominated by the d-band with the smallest Tc, whereas the high- temperature behavior results from the p-band with the larger Tc. The reduced energy gap for p-band, 2 p(0)/kBTc p = 2.81, is rather small relative to the BCS value 3.52, while d-band value, 2 d(0)/kBTc d = = 6.44, is twice as big. Thus we suggest that ZrB12 may have two superconducting bands with different Tc and order parameters. Notice that this unusual conclusion may be right for two bands in the limit of zero interband coupling in agreement with resistivity data [13]. Also, the (0) of ZrB12 may not be constant over the Fermi surface. We based our conclusion on the two-gap model for dirty-limit superconductors, suggested by Gurevich [27]. In this model, we can write 1 4 2 2 2 2� � ( ) ( ) T e c N D N Dp p p d d d � � � 1 0 1 02 2� �p d( ) ( ) , (11) where Ni, i and Di are the density of states, the energy gap and the diffusivity in p- and d-bands, respectively. It follows from electron band structure calculations that the dominant contribution to the density of states N(EF) is made by the Zr4d and B2p states, with Nd = 7.3·1021 st/eV·cm3 and Np = = 8.7·1021 st/eV·cm3, respectively [26]. The B2p bonding states are responsible for the formation of B12 intra cluster covalent bonds. In turn, Zr4d bands are due to Zr sub lattice. We use this two band approach (Eq. (10)) to obtain p-band diffusivity of Dp = 57 cm2/s and d-band diffusivity of Dd = = 17 cm2/s. Note that there is almost a six times difference between the p- and d-band diffusivity. We use this result for our discussion of H Tc2( ) data in the following paragraph. The important goal of this chapter is comparison of ZrB12 and MgB2 data. In Fig. 6 we show the tem- perature variation of a superfluid density, ��(0)/��(T), versus reduced temperature T/Tc for the best MgB2 film as determined from the one-coil technique (Eq. (1)) and inversion procedure from Eq. (3) with �(0) = 114 nm. The solid line represents BCS single gap calculations by the aid of a single term of Eq. (10) and using finite energy gap ( (0) = = 1.93 meV) as the fit parameter. According to Fig. 5, there is a very good agreement between experimental data and the single gap BCS curve over the full tem- perature range. The reduced energy gap 2 (0)/kBTc is evaluated to be 1.14. It is actually within the range of values for 3D � bands obtained by PCS on MgB2 single crystals [16]. Both (0) and �(0) are consistent with microwave measurements on similar c-axis ori- ented thin films (3.2 meV and 107 nm, respectively) [28]. Notice that we studied here the penetration depth in the ab-plane due to the samples being c-axis oriented thin films. This feature predicts that our �ab(T) is determined by the small energy gap for the � band. 5. Upper critical magnetic field We now turn to the electronic transport data acquired in magnetic field. Figure 7 displays the Recent observation of anomalous superconducting behavior of ultrathin YBCO films Fizika Nizkikh Temperatur, 2006, v. 32, Nos. 8/9 1111 �( H H ) – � (0 ), � m � Fig. 7. Magnetic field variation of �(H) of a single crystal ZrB12 sample at different temperatures, K: 5.66, 5.53, 5.02, 4.06, 3.45, 2.84, 2.15 and 1.43, from the left to the right. The solid lines are the guides for the eye. The arrow describes how Hc2 has been deduced. penetration depth magnetic field transitions at various temperatures down to 1.43 K with the fields oriented along the sample bar. Figure 7 clearly demonstrates a well defined onset of �(H) transition. We used this onset to estimate Hc2. Similar approach was used to �(T) data [16]. Figure 8 presents the Hc2 dependence defined at the onsets of the finite �(H) and �(H). Remarkable feature of this plot is an linear increase of Hc2 with decreasing temperature down to 0.35 K. To see whether one gap BCS model may work for ZrB12, we extrapolate H Tc2( ) to zero temperature by use of the derivative of dH Tc2( )/dT close to Tc and the assumption that the zero temperature Hc2 0( ) = = 0.69T dHc c2/dT�Tc. The resulting Hc2 0( ) = 114 mT is substantially lower than the low temperature onset data below 3 K. Linear extrapolation of H Tc2( ) to T = 0 gives Hc2 0( ) = 162 mT. We used this value to obtain the coherence length �(0), by employing the re- lations Hc2 0( ) = �0/2��2(0). The latter yields �(0) = = 45 nm, which is substantially larger than a few ang- stroms coherence length of high-Tc superconductors. In contrast to Ref. 21 our estimations agree well with the Ginzburg–Landau parameter � � �! Using our values of �p and �d data we obtain �p = 3.8 and �d = = 5.8. Both values of �d are larger then 1 2/ that im- plies that ZrB12 is type II superconductor at all T. Using the GL expression forH T /c1 0 24( ) ln � � �� we obtain: Hc2/Hc1 = 2�2/ln �. From the value Hc2 0( ) obtained above and the Hc1 data from [16], we find � = 6.3 and �(0) = 280 nm which is in good agreement with the value �p(0) = 260 nm obtained from two gap BCS fit for d-band. In contrast to the conventional BCS theory, H Tc2( ) dependence is linear over an extended temperature range with no evidence of saturation down to 0.35 K. Similar linear H Tc2( ) dependence has been observed in MgB2 (Refs. 29 and 30) and BaNbOx (Ref. 31) compounds. One can describe this behavior of upper critical field using the two-gap approach. According to Gurevich [27], the zero-temperature value of the Hc2 is significantly enhanced in the two gap dirty limit superconductor model: H k T D D g c B c 2 0 1 2 0 112 2 ( ) . exp � � � � � � � , (12) as compared to the one-gap dirty limit approximation Hc2 0( ) = �0kBTc/1.12�D. Here g is a rather compli- cated function of the matrix of the BCS supercon- ducting coupling constants. In the limit of D2 << D1 we can simply approxi- mate g � �ln (D2/D1)�. The large ratio of D2/D1 leads to the enhancement of Hc2 0( ) and results in the upward curvature of the H Tc2( ) close to T = 0 [27]. According to our �(T) data (see above), we found very different diffusivities for p- and d-bands: Dp/Dd � 3. Thus we can speculate that the limiting value of Hc2 0( ) is dominated by d-band with lower diffusivity Dd = 17 cm2/s, while the derivative dHc2/dT close to Tc is due to larger diffusivity band (Dp = 56 cm2/s). Indeed, simple estimation of Dp = = 4�0kB/�2 �dHc2/dT = 39 cm2/s from derivative dHc2/dT = 0.027 T/K close to Tc gives almost the same diffusivity relative to one estimated from �(T) for the p-band. Thus we believe that the two gap theo- retical model qualitatively explains the unconven- tional linear H Tc2( ) dependence, which supports our conclusion about the two gap nature of superconduc- tivity in ZrB12. The possibility of the multigap nature of the super- conducting state was predicted for a multiband super- conductor with large difference of the e–p interaction at different Fermi-surface sheets (see Ref. 32 and references therein). To date MgB2 has been the only compound with the behavior consistent with the idea of two distinct gaps with the same Tc. We believe that our data can add ZrB12 as another unconventional example of multi gap and multi-Tc superconductor. Although observed two gap behavior of �2(0)/�2(T) in ZrB12 is similar to that in high-Tc su- perconductors, observation of two different Tc in these bands is unconventional. This also relates to the linear H Tc2( ) dependence in the wide temperature range up to Tc. Striking two gap BCS behavior observed calls certainly for a new study of low-T energy gap and Hc2(T) of ZrB12 for understand the nature of super- conductivity in this cluster compounds. 1112 Fizika Nizkikh Temperatur, 2006, v. 32, Nos. 8/9 V.A. Gasparov H c2 , m T Fig. 8. Temperature variations of H Tc2( ) of ZrB12. Symbols: the onset H Tc2( ) data determined from �(H) (circles) and �(H) (squares). Dotted line is the BCS data determined from the HW formula [33]. 6. Conclusion In summary, we have compared our experimental results on ultrathin YBCO films with the extended dynamic theory for BKT transition and found that the vortex—antivortex pairs with short separation lengths are present. The unbinding of the vortex pairs were observable at high frequencies even though a true BKT transition is absent in the samples. Our results also indicate that part of the transition broadening in ultrathin YBCO films can be related to the dissociation of vortex–antivortex pairs. While the temperature dependence of �(T) in thin c-axis oriented thin film MgB2 samples is well des- cribed by an isotropic s-type order parameter, we find unconventional behavior of ZrB12 superfluid density with pronounced shoulder at T/Tc equal to 0.65. The H Tc2( ) dependences have been deduced from the �(H) and �(H) data. Both techniques reveal an un- conventional linear temperature dependence of H Tc2( ). We conclude therefore that ZrB12 presents a evidence of the unconventional two-gap superconduc- tivity with different Tc in the different bands. We are grateful to V.F. Gantmakher, A. Hebard, M. Chan, R. Huguenin, C. Lobb, P. Martinoli, D. van der Marel, C. Rogers, D.J. Scalapino, J.-M. Triscone, T. Venkatesan, X.X. Xi for stimulating discussions. This work was partially supported by Russian Ministry of Industry, Science and Technology (MSh-2169.2003.2) and Russian Academy of Sciences Program: New Materials and Structures. 1. V.L. Berezinskii, ZhETF 59, 907 (1970); ibid 61, 1144 (1971). 2. J. M. Kosterlitz and D.J. Thouless, J. Phys. C6, 1181 (1973); Prog. Low Temp. Phys. B7, 373 (1978). 3. V.A. Gasparov, G. Tsydynzhapov, I.E. Batov, and Qi Li, J. Low Temp. Phys. 139, 49 (2005). 4. L.C. Davis, M.R. Beasley, and D.J. Scalapino, Phys. Rev. B42, 99 (1990). 5. C.T. Rogers, K.E. Myers, J. N. Eckstein et al., Phys. Rev. Lett. 69, 160 (1992). 6. J.M. Repaci, C. Kwon, Qi Li et al., Phys. Rev. B54, 9674 (1996). 7. M.R. Beasley, J.E. Mooij, and T.P. Orlando, Phys. Rev. Lett. 42, 1165 (1979). 8. P. Minnhagen, B.I. Halperin, and D.R. Nelson et al., Phys. Rev. B23, 5745 (1981); Rev. Mod. Phys. 59, 1001 (1987). 9. V. Ambegaokar, B.I. Halperin, and D.R. Nelson et al., Phys. Rev. Lett. 40, 783 (1978); Phys. Rev. B21, 1806 (1980). 10. B.J. Halperin and D.R. Nelson, J. Low Temp. Phys. 36, 599 (1979). 11. J. Bardeen and M.J. Stephen, Phys. Rev. 140, A1197 (1965). 12. J. Nagamatsu, N. Nakagawa, T. Muranaka, Y. Zani- tani, and J. Akimitsu, Nature 410, 63 (2001). 13. V.A. Gasparov, N.S. Sidorov, I.I. Zver’kova, and M.P. Kulakov, Pis’ma Zh. Eksp. Teor. Fiz. 73, 601 (2001) [JETP Lett. 73, 532 (2001)]. 14. V.A. Gasparov, M.P. Kulakov, I.I. Zver’kova, N.S. Sidorov, V.B. Filipov, A.B. Lyashenko, and Yu.B. Paderno, Pis’ma Zh. Eksp. Teor. Fiz. 80, 376 (2004) [JETP 80, 330 (2004)]. 15. V.A. Gasparov, N.S. Sidorov, I.I. Zver’kova, S.S. Khassanov, and M.P. Kulakov, Zh. Eksp. Teor. Fiz. 128, 115 (2005) [JETP 101, 98 (2005)]. 16. V.A. Gasparov, N.S. Sidorov, and I.I. Zver’kova, Phys. Rev. B73, 094510 (2006). 17. K. Tanigaki, T.W. Ebbesen, S. Saito, J. Mizuki, J.S. Tsai, Y. Kubo, and S. Kuroshima, Nature (London) 352, 222 (1991). 18. O. Gunnarsson, Rev. Mod. Phys. 69, 575 (1997). 19. B.T. Matthias, T.H. Geballe, K. Andres, E. Coren- zwit, G. Hull, and J.P. Maita, Science 159, 530 (1968). 20. Y. Paderno, N. Shitsevalova, I. Batko, K. Flahbart, H. Misiorek, J. Mucha, and A. Jezowski, J. Alloys Comp. 219, 215 (1995). 21. R. Lortz, Y. Wang, S. Abe, C. Meingast, Yu.B. Paderno, V. Filippov, and A. Junod, Phys. Rev. B72, 024547 (2005). 22. M. Paranthaman, C. Cantoni, H.Y. Zhai, H.M. Chris- ten, T. Aytug, S. Sathyamurthy, E.D. Specht, J.R. Thompson, D.H. Lowndes, H.R. Kerchner, and D.K. Christen, Appl. Phys. Lett. 78, 3669 (2001). 23. V.A. Gasparov and A.P. Oganesyan, Physica C178, 445 (1991). 24. A. Gauzzi, J. Le Cochec, G. Lamura, B. J. Jonsson, V.A. Gasparov, F.R. Ladan, B. Placais, P.A. Probst, D. Pavuna, and J. Bok, Rev. Sci. Instr. 71, 2147 (2000). 25. V.A. Gasparov, M.R. Mkrtchyan, M.A. Obolensky, and A.V. Bondarenko, Physica C1, 197 (1994). 26. I.R. Shein and A.L. Ivanovskii, Fiz. Tverd. Tela (St. Petersburg) 45, 1363 (2003) [Phys. Solid State 45, 1429 (2003)]. 27. A. Gurevich, Phys. Rev. B67, 184515 (2003). 28. B.B. Jin, N. Klein, W.N. Kang, H.-J. Kim, E.-M. Choi, S.-I. Lee, T. Dahm, K. Maki, Phys. Rev. B66, 104521 (2002). 29. L. Lyard, P. Samuely, P. Szabo, T. Klein, C. Mar- cenat, L. Paulius, K.H.P. Kim, C.U. Jung, H.-S. Lee, B. Kang, S. Choi, S.-I. Lee, J. Marcus, S. Blanchard, A.G. M. Jansen, U. Welp, G. Karapetrov, and W.K. Kwok, Phys. Rev. B66, R180502 (2002). 30. A.V. Sologubenko, J. Jun, S.M. Kazakov, J. Karpinski, and H.R. Ott, Phys. Rev. B66, R180505 (2002). 31. V. A. Gasparov, S.N. Ermolov, G.K. Strukova, N.S. Sidorov, S.S. Khassanov, H.-S. Wang, M. Schneider, E. Glaser, and Wo. Richter, Phys. Rev. B63, 174512 (2001); V.A. Gasparov, S.N. Ermolov, S.S. Khasanov, G.K. Strukova, L.V. Gasparov, H.-S. Wang, Qi Li, M. Schnider, Wo. Richter, E. Glaser, F. Schmidl, Recent observation of anomalous superconducting behavior of ultrathin YBCO films Fizika Nizkikh Temperatur, 2006, v. 32, Nos. 8/9 1113 P. Seidel, and B.L. Brandt, Physica B284–288, 1119 (2000). 32. A. Brinkman, A.A. Golubov, H. Rogalla, O.V. Dol- gov, J. Kortus, Y. Kong, O. Jepsen, and O.K. Ander- sen, Phys. Rev. B65, R180517 (2001); A.A. Golubov, A. Brinkman, O.V. Dolgov, J. Kortus, and O. Jepsen, Phys. Rev. B66, 054524 (2002). 33. E. Helfand and N.R. Werthamer, Phys. Rev. Lett. 13, 686 (1964); Phys. Rev. 147, 288 (1966). 1114 Fizika Nizkikh Temperatur, 2006, v. 32, Nos. 8/9 V.A. Gasparov