Method for data processing in application to ohmic contacts
The method of processing the data of electrophysical investigations of ohmic contacts has been developed. It allows obtaining more accurate results of measuring the contact resistance and additional information by analyzing the statistical and spatial distribution of input data. To test the method,...
Gespeichert in:
| Veröffentlicht in: | Semiconductor Physics Quantum Electronics & Optoelectronics |
|---|---|
| Datum: | 2019 |
| Hauptverfasser: | , , , , , , , , |
| Format: | Artikel |
| Sprache: | Englisch |
| Veröffentlicht: |
Інститут фізики напівпровідників імені В.Є. Лашкарьова НАН України
2019
|
| Schlagworte: | |
| Online Zugang: | https://nasplib.isofts.kiev.ua/handle/123456789/215433 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Zitieren: | Method for data processing in application to ohmic contacts / A.E. Belyaev, N.S. Boltovets, R.V. Konakova, V.M. Kovtonjuk, Ya.Ya. Kudryk, V.V. Shynkarenko, M.M. Dub, P.O. Saj, S.V. Novitskii // Semiconductor Physics Quantum Electronics & Optoelectronics. — 2019. — Т. 22, № 1. — С. 11-18. — Бібліогр.: 11 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| _version_ | 1860480584861614080 |
|---|---|
| author | Belyaev, A.E. Boltovets, N.S. Konakova, R.V. Kovtonjuk, V.M. Kudryk, Ya.Ya. Shynkarenko, V.V. Dub, M.M. Saj, P.O. Novitskii, S.V. |
| author_facet | Belyaev, A.E. Boltovets, N.S. Konakova, R.V. Kovtonjuk, V.M. Kudryk, Ya.Ya. Shynkarenko, V.V. Dub, M.M. Saj, P.O. Novitskii, S.V. |
| citation_txt | Method for data processing in application to ohmic contacts / A.E. Belyaev, N.S. Boltovets, R.V. Konakova, V.M. Kovtonjuk, Ya.Ya. Kudryk, V.V. Shynkarenko, M.M. Dub, P.O. Saj, S.V. Novitskii // Semiconductor Physics Quantum Electronics & Optoelectronics. — 2019. — Т. 22, № 1. — С. 11-18. — Бібліогр.: 11 назв. — англ. |
| collection | DSpace DC |
| container_title | Semiconductor Physics Quantum Electronics & Optoelectronics |
| description | The method of processing the data of electrophysical investigations of ohmic contacts has been developed. It allows obtaining more accurate results of measuring the contact resistance and additional information by analyzing the statistical and spatial distribution of input data. To test the method, the Au–Ge–TiB₂–Au contact to n-n⁺-GaAs was used. The analysis of frequency distribution for the total resistance, specific contact resistance, and surface resistance of the semiconductor has been carried out. The spatial distribution of these parameters has been analyzed. By taking the linear gradient of specific resistivity into account, the value of the contact resistance has been clarified. We have achieved a reduction of half-width of the distribution by 14%, that is, a reduction of the error in determining the contact resistance. The method has been developed for correctly analyzing the impacts of technological treatments and degradation processes and has been oriented towards research purposes. Evaluation of the gradient distributions of the contact resistance and the resistance of the semiconductor can be used to identify the defects in the technological processes of manufacturing devices.
|
| first_indexed | 2026-03-23T19:02:29Z |
| format | Article |
| fulltext |
ISSN 1560-8034, 1605-6582 (On-line), SPQEO, 2019. V. 22, N 1. P. 11-18.
© 2019, V. Lashkaryov Institute of Semiconductor Physics, National Academy of Sciences of Ukraine
11
Semiconductor physics
Method for data processing in application to ohmic contacts
A.E. Belyaev
1
, N.S. Boltovets
2
, R.V. Konakova
1
, V.M. Kovtonjuk
1,2
, Ya.Ya. Kudryk
1
*, V.V. Shynkarenko
1
,
M.M. Dub
1
, P.O. Saj
1
, S.V. Novitskii
3
1
V. Lashkaryov Institute of Semiconductor Physics, NAS of Ukraine,
41, prospect Nauky, 03680 Kyiv, Ukraine
2
State Enterprise “Research Institute “Orion”,
8а, Eugene Pottier str., 03057 Kyiv, Ukraine
3
Zhytomyr Ivan Franko State University,
40, Velyka Berdychivska str., 10008 Zhytomyr, Ukraine
*E-mail: kudryk@isp.kiev.ua
Abstract. The method of processing the data of electrophysical investigations of ohmic
contacts has been developed. It allows obtaining more accurate results of measuring the
contact resistance and additional information by analyzing the statistical and spatial
distribution of input data. To test the method, the Au–Ge–TiB2–Au contact to n-n
+-GaAs
was used. The analysis of frequency distribution for the total resistance, specific contact
resistance and surface resistance of semiconductor has been carried out. The spatial
distribution of these parameters has been analyzed. With taking the linear gradient of
specific resistivity into account, the value of the contact resistance has been clarified. We
have achieved reduction of half-width of the distribution by 14%, that is, reduction of the
error in determining the contact resistance. The method has been developed for correct
analyzing the impacts of technological treatments and degradation processes and has been
oriented on research purposes. Evaluation of the gradient distributions of the contact
resistance and the resistance of semiconductor can be used to identify the defects in the
technological processes of manufacturing devices.
Keywords: specific contact resistance, electrophysical diagnostic, ohmic contacts.
doi: https://doi.org/10.15407/spqeo22.01.11
PACS 73.40.Ns, 85.30.Kk
Manuscript received 08.01.19; revised version received 29.01.19; accepted for publication
20.02.19; published online 30.03.19.
1. Introduction
The development and improvement of ohmic contacts is
one of the first steps in the developing technology of
microelectronic devices based on a new material, also a
step is often needed to optimize the existing technology.
Therefore, the problem of correct study of parameters
inherent to ohmic contacts in order to obtain the
maximum information and trustworthiness is actual and
timely. The existing set of methods for measuring the
specific contact resistance, each of which has its limits of
applicability, accuracy and specificity of use, was
considered, for example, in [1-4], but in these works
statistical aspect of measurements was not take into
account. And in general, the experimenters, as a rule, do
not enough take into account this aspect in processing the
results of measurements.
2. Method of data processing of ohmic contacts
At the same time, the statistical measurements along the
wafer may give not only a more accurate value of the
contact resistance, but additional data related to the
geometry of distributions of the calculated parameters,
nature of errors and parameter gradients. In this work, the
method of processing the data of measurements for
ohmic contacts oriented on research purposes is
proposed. In connection with the research direction, the
visual control of processes that can occur in different
ways depending on the technological conditions of
formation of contacts, characteristics of the
semiconductor surface and post-processing is widely
used (see Fig. 1).
SPQEO, 2019. V. 22, N 1. P. 11-18.
Belyaev A.E., Boltovets N.S., Konakova R.V. et al. Method for data processing in application to ohmic contacts
12
3. Approbation of method
To test the method, we used the contact of Au–Ge–TiB2–
Au to n-n
+ with the concentration of carriers in the layer
n ~ 5·1015 cm–3. The metallization was deposited using
the magnetron method with the thicknesses of layers: Au
– 1800 Å; Ge – 300 Å; TiB2 – 1000 Å; Au – 2000 Å.
After deposition of metal by using photolithography we
form contact structures. Measuring the specific contact
resistance was carried out using the method CTLM
(circular transmission line model) (Fig. 2, insert). The
temperature of rapid thermal annealing was 480 °С.
3.1. Analysis of small sample size
The first step in studying the contact resistance, which
should be made before photolithography, is to choose a
measurement method or several alternative methods. The
critical point of many numerical methods is checking on
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
-0.06
-0.04
-0.02
0.00
0.02
0.04
0.06
4
3
2
I,
A
V, V
1
4
3
2
1
Fig. 2. Current-voltage characteristics of the contacts of one
template. In the insert – a template of radial transmission line
model (CTLM).
the verge of applicability. Since the measurement of
contacts of a large sample size is rather resource-
intensive, the second step is the study of a small sample
(up to 10 samples) that can be processed under a
simplified scheme and is necessary for timely detection
of ohmic contact parameters exceeding bounds of the
study using this method and the need for a change in the
measurement method. It is important to study the current-
voltage characteristics (I–V) of contacts of the small
sample size (Fig. 2). Formally, only the linearity and
symmetry of I–V give us the right to use the resistance to
characterize the contact. Usually, the research of the
contact resistance contains the dependence of the
resistance on the geometric factor [1-4].
One of the criteria for the applicability of methods
for calculating the contact resistance is good
approximation of the experimentally obtained points with
the theoretical curve (Fig. 3). In addition, for TLM
methods the relationship between contact length and
transfer length is often critical, for the vertical – the
relationship between the diameter and thickness of the
wafer.
0 1 2 3 4 5
2
4
6
8
10
12
R
,
Ω
1/πr
2
, 10
4
cm
-2
Fig. 3. Dependence of the resistance on the inverse contact
area.
Fig. 1. Block-scheme of the method foк statistical processing the measurement results to calculate the contact resistance.
SPQEO, 2019. V. 22, N 1. P. 11-18.
Belyaev A.E., Boltovets N.S., Konakova R.V. et al. Method for data processing in application to ohmic contacts
13
3.2. Analysis of large sample size
If the results of a small sample size show the
applicability of the method of measuring the contact
resistance, then the next step is to study a large sample
size of contact resistance. The distribution of resistance
parameters will not necessarily be normal. It is even
more likely that a normal distribution will be obtained
only for the logarithm of resistance (lognormal
distribution) [4], which should be taken into account
when 3-sigma filtering the input data is used.
Accordingly, the next step is to check the distributions
for normality.
3.2.1. Verification of distributions for normality and
3-sigma filtering
The Shapiro–Wilk method is used to verify normality,
which shows better results than other similar tests [5].
Fig. 4 shows the distribution histograms of the total
resistance of the contacts with the diameters 50, 80, 100
and 150 µm. It is evident that in the half-logarithmic
coordinates, the distribution is closer to normal, which is
confirmed by the Shapiro–Wilk test. It is especially
noticeable when comparing Figs. 4g and 4h, from this it
follows that the distribution should be described by the
lognormal dependence. Some deviation of this
distribution from the lognormal dependence can be
caused by the presence of two peaks of the distribution of
resistance (see Fig. 4b, 4d), one of which may be related
with the distribution of the contact resistance, another –
with the influence of the scatter of the semiconductor
parameters and other factors unrelated with the contact
resistance. That is why the first peak that is significant at
small contact diameters (50 and 80 µm) becomes
practically invisible on the contacts of larger diameter, in
accordance with the reduction of contribution of the
contact resistance in the total resistance of the contacts.
Only the knowledge that the distribution is close to
lognormal gives grounds to carry out 3-sigma filtering
the input data, that is, to discard several points that are
not part of the general distribution and are probably
artifacts associated with the defectness of the
semiconductor, setting before deposition of contact,
lithography and more.
After filtering, we improve the values of median,
mode and average values and a second 3-sigma filtration
is performed until a point is rejected. Only after filtering
the input data, we can go to the calculation of the specific
contact resistance.
3.2.2. Calculation of contact resistance and associated
parameters
In our case, in the method of calculating the contact
resistance we used a template containing a set of contacts
with a constant ratio of internal and external radii r1/r2
(see Fig. 2, the insert). For a single contact, the total
resistancence can be written as in [6]:
( )
( )11
10
11
2
α
α
α
1
π2
ln
π2 rJ
rJ
r
R
r
rR
R ss +
= , (1)
where r1 and r2 are internal and external radii of the
corresponding contacts; J0, J1 – modified Bessel
functions of the first kind; α ≡ 1/Lt is the inverse length
of the transfer,
S
c
t
R
L
ρ
≡ , RS is the specific surface
resistance of the surface layer of the semiconductor; ρc –
specific contact resistance.
In the case where Lt >> r1 , the equation (1) can be
simplified as follows:
2
1π
ρ
r
CRR c
s += , (2)
where
≡
1
2ln
π2
1
r
r
C is a constant. In other words, the
dependence of the total resistance on the inverse contact
area is linear with the slope coefficient equal to the
specific contact resistance, as we observe in Fig. 3 for
one of the templates. From each template on the plate,
the specific contact resistance was calculated and a
histogram of the frequency distribution of the specific
contact resistance was plotted in linear and semi-
logarithmic coordinates (Fig. 5). It should be noted that
only in semi-logarithmic coordinates the distribution
passes the test Shapiro–Wilk for normality (Fig. 5b),
which gives us the right to continue consideration of the
distribution as lognormal.
Mathematical expectation calculated on the basis
of approximation of distribution by the Gauss curve
in Fig. 5b is 1.52·10–4 Ohm·cm2. The average
value is 1.60·10–4 Ohm·cm2, the median equals
1.52·10–4 Ohm·cm2, the mode is 1.60·10–4 Ohm·cm2.
3.2.3. Correlation analysis
Due to the scatter of the ohmic contact parameters, an
error in the determination of the contact resistance may
occur, which, in accordance with (2), when the contact
resistance is exceeded, simultaneously leads to lowering
the specific surface resistance of the semiconductor, and
vice versa. That is, the error of this kind will lead to a
negative correlation between ρc and Rs. On the other
hand, the correlation between ρc and Rs can be of
physical nature, that is, changes in both parameters may
be related to the variations in the concentration of the
doping admixture in the film, which will lead to a
positive correlation between the parameters. It is known
that with a negative correlation, the replacement of Rs,
determined from a specific template as the average value
on the plate, can reduce the error in determination
of ρc [7].
In our case, a highly blurred negative correlation is
observed (Fig. 6). Let re-calculate ρc, replacing RsС
SPQEO, 2019. V. 22, N 1. P. 11-18.
Belyaev A.E., Boltovets N.S., Konakova R.V. et al. Method for data processing in application to ohmic contacts
14
6 7 8 9 10 11 12 13 14
0
5
10
15
20
Normality Test (Shapiro-Wilk)
W P Value Decision
---------------------------------------------------------------------
0.92992 1.46026E-4 Not Normal at 0.05 level
----------------------------------------------------------------------
Model: Gauss
y=y0 + (A/(w*sqrt(PI/2)))*exp(-2*((x-xc)/w)^2)
Chî 2/DoF = 26.57652
R^2 = 0.78823
y0 0.8435 ±4.91793
xc 8.87918 ±0.34594
w 3.13014 ±1.14554
A 79.72857 ±41.15566
C
o
u
n
t
R
50
, Ω
a)
0.85 0.90 0.95 1.00 1.05 1.10 1.15
0
5
10
15
20
Normality Test (Shapiro-Wilk)
W P Value Decision
-------------------------------------------------------------------
0.94922 0.00624 Not Normal at 0.05 level
-------------------------------------------------------------------
Model: Gauss
y=y0+(A/(w*sqrt(PI/2)))*exp(-2*((x-xc)/w)^2)
Chi^2/DoF = 6.05598
R^2 = 0.94069
y0 0 ±0
xc1 1.02542 ±0.00579
w1 0.05541 ±0.01153
A1 1.0921 ±0.20286
xc2 0.9154 ±0.00482
w2 0.06664 ±0.01066
A2 1.70556 ±0.22344
C
o
u
n
t
lg(R
50
)
b)
3 4 5 6 7 8 9
0
5
10
15
20
Normality Test (Shapiro-Wilk)
W P Value Decision
-----------------------------------------------------------------------
0.82193 3.70039E-15 Not Normal at 0.05 level
-------------------------------------------------------------------------
Data: Count7_Count
Model: Gauss
Equation: y=y0 + (A/(w*sqrt(PI/2)))*exp(-2*((x-xc)/w)^2)
Weighting:
y No weighting
Chi^2/DoF = 6.76152
R^2 = 0.92084
y0 0.88819 ±1.05029
xc 4.07466 ±0.07839
w 1.43681 ±0.18281
A 38.20977 ±5.28043
C
o
u
n
t
R
80
, Ω
c)
0.5 0.6 0.7 0.8 0.9
0
5
10
15
20
25
Normality Test (Shapiro-Wilk)
W P Value Decision
------------------------------------------------------------------------
0.90587 5.47523E-7 Not Normal at 0.05 level
-------------------------------------------------------------------------
Model: Gauss
Equation:
y=y0 + (A/(w*sqrt(PI/2)))*exp(-2*((x-xc)/w)^2)
Chi^2/DoF = 2.27357
R^2 = 0.98335
y0 0 ±0
xc1 0.54558 ±0.01268
w1 0.04489 ±0.0235
A1 0.99008 ±0.6856
xc2 0.64434 ±0.02039
w2 0.15224 ±0.02746
A2 3.18128 ±0.71365
C
o
u
n
t
lg(R
80
)
d)
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
0
5
10
15
20
25
Normality Test (Shapiro-Wilk)
W P Value Decision
----------------------------------------------------------------------
0.87719 6.57794E-10 Not Normal at 0.05 level
-----------------------------------------------------------------------
Model: Gauss
y=y0 + (A/(w*sqrt(PI/2)))*exp(-2*((x-xc)/w)^2)
Chi^2/DoF = 3.66837
R^2 = 0.97351
y0 2.83857 ±1.13526
xc 2.93187 ±0.05619
w 1.26543 ±0.14694
A 34.92113 ±4.52296
C
o
u
n
t
R
100
, Ω
e)
0.3 0.4 0.5 0.6 0.7 0.8
0
5
10
15
20
Model: Gauss
y=y0 + (A/(w*sqrt(PI/2)))*exp(-2*((x-xc)/w)^2)
Chi^2/DoF = 9.31806
R^2 = 0.92934
y0 1.46791 ±3.87503
xc1 0.51877 ±0.79742
w1 0.17415 ±0.60903
A1 2.12076 ±21.77167
xc2 0.43458 ±0.06789
w2 0.11523 ±0.30872
A2 1.56651 ±20.23672
Normality Test (Shapiro-Wilk)
W P Value Decision
----------------------------------------------------------------------
0.93688 3.94975E-4 Not Normal at 0.05 level
-----------------------------------------------------------------------
C
o
u
n
t
lg(R
100
)
f)
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
0
5
10
15
20
25
30
Normality Test (Shapiro-Wilk)
W P Value Decision
--------------------------------------------------------------------
0.85297 2.56739E-12 Not Normal at 0.05 level
C
o
u
n
t
R
150
, Ω
g)
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
0
5
10
15
20
25
30
Normality Test (Shapiro-Wilk)
W P Value Decision
---------------------------------------------------------------------
0.91613 4.4257E-6 Not Normal at 0.05 level
----------------------------------------------------------------------
Data: Count14_Count
Model: Gauss
Equation: y=y0 + (A/(w*sqrt(PI/2)))*exp(-2*((x-xc)/w)^2)
Weighting:
y No weighting
Chi^2/DoF = 8.1999
R^2 = 0.94998
y0 0 ±0
xc 0.29985 ±0.01284
w 0.25213 ±0.0282
A 8.83063 ±0.8054
C
o
u
n
t
lg(R
150
)
h)
Fig. 4. Total contact resistance histograms for contacts d = 50 µm (a, b), d = 80 µm (c, d), d = 100 µm (e, f), and
d = 150 µm (g, h) in linear (a, c, e, g) and semi-logarithmic (b, d, f, h) coordinates.
SPQEO, 2019. V. 22, N 1. P. 11-18.
Belyaev A.E., Boltovets N.S., Konakova R.V. et al. Method for data processing in application to ohmic contacts
15
0 1 2 3
1.0x10
-4
2.0x10
-4
ρ
c
,
Ω
·c
m
2
R
s
, arb. units
Fig. 6. Relationship between the calculated parameters ρc
and Rs.
Fig. 7. Histogram of the frequency distribution of the specific
contact resistance with the replacement RsС by CRs
.
Fig. 8. Error of approximation of the dependence inherent
to the specific contact resistance on the coordinate projection:
1 – radial distribution with the center defined from Fig. 9, and
2 – linear distribution depending on the angle of rotation.
determined for the specific template, on the mathematical
expectation of a sample across the whole plate CRs ,
found on the basis of the frequency distribution of the
specific surface resistance. We find the contact resistance
for each template as the average weighted value *
cρ
found from the formula (2) for each of the template
contact with weighted coefficients
( ) 42 1ρπ
−
+≡ csi rCRϑ , proportional to the contribution
of error of the total resistance to the contact resistance:
∑
∑
=
==
n
i
i
n
i
cii
c
n
1
1*
ρ
ρ
ϑ
ϑ
. (3)
1.0x10
-4
2.0x10
-4
3.0x10
-4
4.0x10
-4
0
2
4
6
8
10
12
14
16
18
Model: Gauss
Equation:
y=y0+(A/(w*sqrt(PI/2)))*exp(-2*((x-xc)/w)^2)
Weighting:
y No weighting
Chi^2/DoF = 5.26493
R^2 = 0.78801
y0 1.97823 ±1.27423
xc 0.00015 ±3.5561E-6
w 0.00005 ±0.00001
A 0.0006 ±0.00017
C
o
u
n
t
ρ
c
, Ω·cm
2
Normality Test (Shapiro-Wilk)
W P Value Decision
-------------------------------------------------------------------
0.95967 0.03493 Not Normal at 0.05 level
-------------------------------------------------------------------
a)
1E-4 2E-4 1E-3
0
2
4
6
8
10
12
14
16
18
Normality Test (Shapiro-Wilk)
W P Value Decision
-----------------------------------------------
0.97419 0.30844 Normal at 0.05 level
---------------------------------------------------------
Model: Gauss
Equation:
y=y0+(A/(w*sqrt(PI/2)))*exp(-2*((x-xc)/w)^2)
Weighting:
y No weighting
Chi^2/DoF = 3.60825
R^2 = 0.9238
y0 -2.00401 ±4.68736
xc -3.81766 ±0.00751
w 0.20218 ±0.05454
A 4.50608 ±2.23332
C
o
u
n
t
ρ
c
, Ω·cm
2
b)
Fig. 5. Histograms of frequency distributions of specific contact resistance in linear (a) and semi-logarithmic (b) coordinates.
1E-4 2E-4 1E-3
0
2
4
6
8
10
12
14
16 Count
Gauss aprox
from fig. 5b
C
o
u
n
t
ρ
c
, Ω·cm
2
0 60 120 180 240 300 360
8.20E-011
8.25E-011
8.30E-011
8.35E-011
8.40E-011
8.45E-011
8.50E-011
8.55E-011
8.60E-011
2
Angle
1
152°
SPQEO, 2019. V. 22, N 1. P. 11-18.
Belyaev A.E., Boltovets N.S., Konakova R.V. et al. Method for data processing in application to ohmic contacts
16
Fig. 9. Distribution of the contact resistance on the area of the
plate (lighter colour – worse). The center of the radial
distribution is marked with a dot. The red arrow indicates the
optimal angle of the linear gradient from Fig. 8.
Fig. 10. Error of approximation of the dependence inherent to
the specific resistivity of semiconductor on the coordinate
projection: 1 – radial distribution with the center defined from
Fig. 9 and 2 – linear distribution depending on the angle of
rotation.
1 2 3 4 5 6 7 8 9 10 11 12 13
1
2
3
4
5
6
7
X Position
Y
P
o
s
it
io
n
0
0.5000
1.000
1.500
2.000
2.500
3.000
3.500
4.000
Rs, arb. units
Fig. 11. Distribution of the specific resistivity of semiconductor
on the area of the plate (lighter colour – worse). The center of
the radial distribution is marked with a dot. The red arrow
indicates the optimal angle of the linear gradient from Fig. 10.
In this case, this replacement does not improve the
scatter of values of the calculated specific resistance
(Fig. 7), which can testify to the simultaneous action of
several factors of errors, in particular, the influence of the
spatial distribution of semiconductor specific resistivity,
which we will examine in the next paragraph.
3.2.4. Gradient analysis
Spatial heterogeneity can also contribute to the value of
both the specific contact resistance and specific
resistivity of the semiconductor. To study the presence of
spatial distribution, two conditions were tested, namely:
the presence of a linear gradient of the distribution of the
contact resistance and presence of a radial distribution of
the contact resistance with the center located in the center
of the plate. The physical preconditions for the
appearance of linear gradients are related with the
technological processing of a part of the sample, in
particular, the chemical processing of the surface and the
rapid thermal annealing. The radial distribution can occur
as a result of the transfer of growth inhomogeneities of
the plate to the epitaxial film, which will affect the
properties of the ohmic contact.
From Fig. 8, we see that the linear gradient of the
distribution inherent to the specific resistance is more
likely to be realized. Visually, we may observe it in
Fig. 9b, as the color gradient from the lighter in the upper
left corner to the darker in the lower right corner of the
matrix. Similar graphs can be constructed for the specific
resistance of the semiconductor (Figs. 10 and 11). The
angle to the gradient of the semiconductor specific
resistivity does not coincide with that for the contact
resistance, which is understandable, since otherwise there
would be a clear correlation between ρc and Rs with a
connection through the geometric factor.
0 1 2 3 4
1.0x10
-4
1.5x10
-4
2.0x10
-4
2.5x10
-4
ρ
c
,
Ω
·c
m
2
R
s
, arb. units
Fig. 12. Dependence between the calculated parameters ρc and
Rs is separated into three identical groups of contacts by
placement on the plate with the optimal angle of the linear
gradient.
0 60 120 180 240 300 360
0.0070
0.0071
0.0072
0.0073
0.0074
0.0075
0.0076
2
E
rr
o
r
Angle
1
51°
SPQEO, 2019. V. 22, N 1. P. 11-18.
Belyaev A.E., Boltovets N.S., Konakova R.V. et al. Method for data processing in application to ohmic contacts
17
1E-4 2E-4 1E-3
0
5
10
15
20
Count
Gauss fit
from fig. 5b
C
o
u
n
t
ρ
c
, Ω·cm
2
Fig. 13. Refined histogram of the frequency distribution
inherent to the specific resistance of the contact.
More obviously, the same can be seen in Fig. 12,
where the set of experimental points is separated by three
identical subsets with their placement along the axis
corresponding to the optimal gradient (Fig. 9), the
influence of the geometric factor on the blurring of
dependence between the parameters ρc and Rs is clearly
seen. Accordingly, to improve the value of the contact
resistance, we will replace Rs with the linear gradient of
resistivity and re-calculate the value of the contact
resistance. As can be seen from Fig. 13, this substitution
results in a significant reduction in the half-width of the
distribution by 14%, that is, enables to reduce the error in
determining the contact resistance.
In our case, the ohmic contacts are formed at a
sufficiently high annealing temperature [2, 8-11], so the
scatter of the values of the contact resistance is minimal,
but even in this situation, the method allows to obtain
additional information on the sources of the scatter in the
contact resistance parameters, which may be useful in
improvement the technological processes of forming
contacts or quality control of experimental production.
4. Conclusions
The method of processing data of electrophysical
investigations of ohmic contacts has been developed,
which allows obtaining more accurate results of
measuring the contact resistance and obtaining additional
information by analyzing the statistical and spatial
distributions of input data. It has been shown that the
distribution of the specific contact resistance, at least in
this case, is described by the lognormal dependence,
which is confirmed by the Shapiro–Wilk test. The
method has been developed for the correct analysis of the
impacts of technological treatments and degradation
processes and has been oriented on research purposes.
Evaluation of the gradient distributions of the contact
resistance and resistance of semiconductor can be used to
identify limitation in the technological processes of
production of devices.
References
1. Sidhant Grover. Effect of Transmission Line
Measurement (TLM) Geometry on Specific Contact
Resistivity Determination. A thesis for the degree of
Master of Science in Materials Science and
Engineering in the School of Chemistry and
Materials Science, College of Science Rochester
Institute of Technology, December 2016.
2. Holland A.S., Reeves G.K. New challenges to the
modelling and electrical characterization of ohmic
contacts for ULSI devices. Microelectronics
Reliability. 2000. 40, No 6. P. 965−971; doi:
10.1109/ICMEL.2000.838732.
3. Berger H.H., Holland A.S., Reeves G.K. Contact
resistance and contact resistivity. J. Electrochem.
Soc. 1972. 119, No 4. P. 507−514. doi:
10.1149/1.2404240.
4. Sheremet V.N. Metrological aspects of measuring
resistance of ohmic contacts. Radioelectronics and
Communications Systems. 2010. 53, Issue 3. P.
119–128;
https://doi.org/10.3103/S0735272710030015.
5. Razali N.M., Wah Y.B. Power comparisons of
Shapiro−Wilk, Kolmogorov−Smirnov, Lilliefors
and Anderson−Darling tests. Journal of Statistical
Modeling and Analytics. 2011. 2, No 1, Р. 21−33.
6. Reeves G.K. Specific contact resistance using a
circular transmission line model. Solid-State
Electronics. 1980. 23. No 5. P. 487−490.
https://doi.org/10.1016/0038-1101(80)90086-6.
7. Basanets V.V., Slepokurov V.S., Shynkaren-
ko V.V., Kudryk Ya.Ya, Kudryk R.Ya, Konako-
va R.V., Kovtonuk V.M. Studying the resistivity of
ohmic contacts Au−Ti−Pd−n-Si for avalanche
transit-time diodes. Tekhnologiya i konstruirovaniye
v elektronnoy apparature. 2015. №1. C. 33−37 (in
Russian); doi: 10.15222/TKEA2015.1.33.
8. Belyaev A.E., Boltovets N.S., Kapitanchuk L.M.
et al. The features of temperature dependence of
contact resistivity of Au−Ti−Pd2Si−p
+-Si ohmic
contacts. Semiconductor Physics, Quantum
Electronics & Optoelectronics. 2010. 13, No 1.
P. 8−11.
9. Sachenko A.V., Belyaev A.E., Boltovets N.S. et al.
On a feature of temperature dependence of contact
resistivity for ohmic contacts to n-Si with an n
+-n
doping step. Semiconductor Physics, Quantum
Electronics & Optoelectronics. 2014. 17, No 1. P.
1−6.
10. Belyaev A.E., Pilipenko V.A., Anischik V.M. et al.
Role of dislocations in formation of ohmic contacts
to heavily doped n-Si. Semiconductor Physics,
Quantum Electronics & Optoelectronics. 2013. 16,
No 2. P. 99−110.
11. Shepela A. The specific contact resistance of Pd2Si
contacts on n- and p-Si. Solid-State Electronics.
1973. 16, Issue 4. P. 477−481.
https://doi.org/10.1016/0038-1101(73)90185-8.
SPQEO, 2019. V. 22, N 1. P. 11-18.
Belyaev A.E., Boltovets N.S., Konakova R.V. et al. Method for data processing in application to ohmic contacts
18
Authors and CV
A.E. Belyaev, Director of V. Lashka-
ryov Institute of Semiconductor
Physics, Асаdemician of NAS of
Ukraine, Professor, Doctor of
Sciences. The area of his scientific
interests includes electrical and
galvanomagnetic properties of
semiconductors.
M.S. Boltovets, Head of Department
at the State Enterprise “Research
Institute “Orion”. The area of his
scientific interests includes IMPATT
and Gunn diode technology,
reliability of semiconductor devices.
R.V. Konakova, Professor, Doctor of
Sciences, Head of Laboratory for
physical and technological problems
of solid-state microwave electronics
at the V. Lashkaryov Institute of
Semiconductor Physics. The area of
her scientific interests includes solid state physics,
transport properties in metal-semiconductor contacts,
reliability of semiconductor devices.
V.M. Kovtonjuk, Researcher at the
State Enterprise “Research Institute
“Orion”. The area of his scientific
interests includes IMPATT and Gunn
diode technology, generator
efficiency.
Ya.Ya. Kudryk, Senior researcher at
the V. Lashkaryov Institute of
Semiconductor Physics. The area of
his scientific interests includes solid
state physics, transport properties in
metal-semiconductor contacts to SiC,
GaN, GaP, InP.
V.V. Shynkarenko, Senior researcher
of V. Lashkaryov Institute of
Semiconductor Physics. The area of
his scientific interests includes solid
state physics, thin film physics, high-
frequency devices.
M.M. Dub The area of his scientific
interests includes solid state physics,
transport properties in metal-
semiconductor contacts to diamond.
P.O. Saj The area of his scientific
interests includes solid state physics,
transport properties in metal-
semiconductor contacts to InN.
S.V. Novitskii, Senior Lecturer of
Zhytomyr Ivan Franko State
University. The area of his scientific
interests includes solid state physics,
transport properties in metal-
semiconductor contacts to InP.
|
| id | nasplib_isofts_kiev_ua-123456789-215433 |
| institution | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| issn | 1560-8034 |
| language | English |
| last_indexed | 2026-03-23T19:02:29Z |
| publishDate | 2019 |
| publisher | Інститут фізики напівпровідників імені В.Є. Лашкарьова НАН України |
| record_format | dspace |
| spelling | Belyaev, A.E. Boltovets, N.S. Konakova, R.V. Kovtonjuk, V.M. Kudryk, Ya.Ya. Shynkarenko, V.V. Dub, M.M. Saj, P.O. Novitskii, S.V. 2026-03-16T11:01:50Z 2019 Method for data processing in application to ohmic contacts / A.E. Belyaev, N.S. Boltovets, R.V. Konakova, V.M. Kovtonjuk, Ya.Ya. Kudryk, V.V. Shynkarenko, M.M. Dub, P.O. Saj, S.V. Novitskii // Semiconductor Physics Quantum Electronics & Optoelectronics. — 2019. — Т. 22, № 1. — С. 11-18. — Бібліогр.: 11 назв. — англ. 1560-8034 PACS: 73.40.Ns, 85.30.Kk https://nasplib.isofts.kiev.ua/handle/123456789/215433 https://doi.org/10.15407/spqeo22.01.11 The method of processing the data of electrophysical investigations of ohmic contacts has been developed. It allows obtaining more accurate results of measuring the contact resistance and additional information by analyzing the statistical and spatial distribution of input data. To test the method, the Au–Ge–TiB₂–Au contact to n-n⁺-GaAs was used. The analysis of frequency distribution for the total resistance, specific contact resistance, and surface resistance of the semiconductor has been carried out. The spatial distribution of these parameters has been analyzed. By taking the linear gradient of specific resistivity into account, the value of the contact resistance has been clarified. We have achieved a reduction of half-width of the distribution by 14%, that is, a reduction of the error in determining the contact resistance. The method has been developed for correctly analyzing the impacts of technological treatments and degradation processes and has been oriented towards research purposes. Evaluation of the gradient distributions of the contact resistance and the resistance of the semiconductor can be used to identify the defects in the technological processes of manufacturing devices. en Інститут фізики напівпровідників імені В.Є. Лашкарьова НАН України Semiconductor Physics Quantum Electronics & Optoelectronics Semiconductor physics Method for data processing in application to ohmic contacts Article published earlier |
| spellingShingle | Method for data processing in application to ohmic contacts Belyaev, A.E. Boltovets, N.S. Konakova, R.V. Kovtonjuk, V.M. Kudryk, Ya.Ya. Shynkarenko, V.V. Dub, M.M. Saj, P.O. Novitskii, S.V. Semiconductor physics |
| title | Method for data processing in application to ohmic contacts |
| title_full | Method for data processing in application to ohmic contacts |
| title_fullStr | Method for data processing in application to ohmic contacts |
| title_full_unstemmed | Method for data processing in application to ohmic contacts |
| title_short | Method for data processing in application to ohmic contacts |
| title_sort | method for data processing in application to ohmic contacts |
| topic | Semiconductor physics |
| topic_facet | Semiconductor physics |
| url | https://nasplib.isofts.kiev.ua/handle/123456789/215433 |
| work_keys_str_mv | AT belyaevae methodfordataprocessinginapplicationtoohmiccontacts AT boltovetsns methodfordataprocessinginapplicationtoohmiccontacts AT konakovarv methodfordataprocessinginapplicationtoohmiccontacts AT kovtonjukvm methodfordataprocessinginapplicationtoohmiccontacts AT kudrykyaya methodfordataprocessinginapplicationtoohmiccontacts AT shynkarenkovv methodfordataprocessinginapplicationtoohmiccontacts AT dubmm methodfordataprocessinginapplicationtoohmiccontacts AT sajpo methodfordataprocessinginapplicationtoohmiccontacts AT novitskiisv methodfordataprocessinginapplicationtoohmiccontacts |