Heat capacities of titanium silicides Ti5Si3, TiSi and TiSi2

Article abstract of DOI:10.1016/S0925-8388(00)01223-8

Titanium silicides Ti₅Si₃, TiSi and TiSi₂ have been prepared and characterized by X-ray powder diffraction and optical metallography. The heat capacities of the silicides have been measured by means of differential scannin

Full text of DOI:10.1016/S0925-8388(00)01223-8

Journal of Alloys and Compounds 314 (2001) 99–102
L
www.elsevier.com/locate/jallcom
The heat capacities of titanium silicides Ti Si , TiSi and TiSi
5
3
2
a
a ,
b
b
b,1
*
Shilpa Agarwal , Eric J. Cotts , Sergei Zarembo , Robert Kematick , Clifford Myers
a
Department of Physics, State University of New York at Binghamton, P.O. Box 6016, Binghamton, NY 13902-6016, USA
b
Department of Chemistry and Materials Research Center, State University of New York at Binghamton, P.O. Box 6016, Binghamton,
NY 13902-6016, USA
Received 21 August 2000; accepted 28 August 2000
Abstract
Titanium silicides Ti Si , TiSi and TiSi have been prepared and characterized by X-ray powder diffraction and optical metallography.
5
3
2
The heat capacities of the silicides have been measured by means of differential scanning calorimetry (DSC) in the temperature range of
2
1
2
30 to 600 K. The heat capacities of Ti Si , TiSi and TiSi at 298 K have been found to be 18163, 47.260.6 and 66.660.9 J mol K,
5 3 2
respectively. The obtained results are compared with literature heat capacity data.
 2001 Published by Elsevier Science B .V .
Keywords: Heat capacity; Differential scanning calorimetry; Titanium silicides
1
. Introduction
sured by means of adiabatic calorimetry. For TiSi, Archer
evaluated his measurements of enthalpy increments, along
with those reported by others [3,9] at temperatures between
298.15 and 1473 K, in order to present values of heat
capacities between 4.2 and 1500 K. In this investigation
the heat capacities of Ti Si , TiSi, and TiSi have been
measured in the temperature range of 230–600 K by
means of differential scanning calorimetry. The goal of the
present investigation is to provide direct measurements of
the heat capacities of these Ti silicides in this temperature
range to compare them with previous measurements in an
effort to clarify the thermal functions for titanium silicides.
Titanium disilicide is used for Si interconnects because
of its low electrical resistivity and thermal stability [1]. As
integrated circuit and electronic package requirements
increase, greater efficiency in fabrication processes such as
chemical vapor deposition (CVD) will be required. Reli-
able thermodynamic data on titanium silicides are essential
to the continued development and optimization of these
advanced thin-film fabrication techniques [2].
Previous reports [3–9] of the thermal properties of Ti
silicides revealed some large disparities between values of
heat capacities, particularly in the temperature range
between 300 and 600 K. Golutvin [3] previously reported
heat capacities for Ti Si , TiSi and TiSi at temperatures
5
3
2
2. Experimental details
5
3
2
between 300 and 1400 K calculated from enthalpy incre-
ment measurements performed by means of drop
calorimetry. Sylla et al. [4] published heat capacity values
Titanium silicide samples were prepared by arc melting
stoichiometric mixtures of elemental titanium (less than
0.01% metal impurities) and silicon (less than 0.005%
metal impurities) on a water-cooled copper hearth under a
purified argon atmosphere. To remove traces of oxygen
and water from argon inside the arc melting chamber, a
titanium slug, used as a getter, was arc melted for about 1
min just before a Ti–Si sample. The resulting titanium
silicide ingots were turned on their side and remelted at
least once. X-ray powder diffraction patterns indicated that
for TiSi between 350 and 500 K, obtained by differential
2
scanning calorimetry. Sychev et al. [8] reported heat
capacity values for Ti Si in the temperature range of 15
5
3
to 300 K obtained by means of adiabatic calorimetry.
Archer et al. [5–7] reported heat capacity values for all
three of these silicides between 4.2 and 350 K calculated
from enthalpy increments of the titanium silicides mea-
the arc melted Ti Si and TiSi samples were single phase.
*
Corresponding author. Tel.: 11-607-777-4371.
5
3
2
In the TiSi samples, however, a significant amount of
foreign phases was detected. The arc melted TiSi samples
E-mail address: ecotts@binghamton.edu (E.J. Cotts).
Deceased.
1
0
925-8388/01/$ – see front matter
 2001 Published by Elsevier Science B .V .
PII: S0925-8388(00)01223-8

1
00
S. Agarwal et al. / Journal of Alloys and Compounds 314 (2001) 99–102
2
6
were annealed in vacuum (10 Torr) at 11508C for 18 h.
After annealing of the TiSi samples, X-ray diffraction
analysis showed that they were single phase. Lattice
parameters of Ti Si , TiSi and TiSi , calculated from
anneal at the highest temperature of a specific interval.
This completed anneal cycle was repeated three times for
each sample during 1 day. The third data set was used for
data analysis. Sapphire was used as a primary standard for
calibrating the measurements of heat flow, while copper
was utilized to check the accuracy of the measurements.
The analysis of the differential scanning calorimetry
data to determine heat capacities has been previously
described in detail [13–15]. First, the DSC signals for both
the Ti–Si and for the reference samples were obtained by
subtracting the data due to the empty Al pan from the total
DSC signal of an encapsulated sample. For each set of
runs, the ratio between the known heat capacity values for
sapphire [16] and the observed signal for a sapphire
reference sample was then used to calculate a calibration
factor as a function of temperature within a specific
temperature range. The heat capacities for Ti Si , TiSi and
5
3
2
X-ray powder diffraction data, are given in Table 1. These
values are in good agreement (within 0.5%) with previous-
ly reported measurements [4,8,10–12].
Polished cross-sections of the Ti–Si samples were
examined by means of optical metallography. The metal-
lographic images revealed that there were some foreign
phases present in the samples. The amount of impurity was
estimated by measuring the relative area of the phases
observed on the metallographic images. The TiSi sample
contained less than 1 at.% of foreign phases. The impurity
in the TiSi sample was less than 0.1 at.%, and the Ti Si
2
5
3
sample was contaminated by less than 1 at.%.
The heat capacities of the Ti Si , TiSi and TiSi
5
3
2
5
3
samples in the temperature range of approximately 250–
00 K were measured by differential scanning calorimetry
DSC) [13,14]. The equipment used was a Perkin-Elmer
TiSi samples were calculated by multiplying the observed
2
6
(
signal for the samples by the calibration factor. Variations
in the calibration factor from run to run or from day to day
were typically within 1%.
DSC2 interfaced to a microcomputer for data acquisition.
Sapphire and copper (both obtained from National Institute
of Standards and Technology [NIST]) were used as heat
capacity standards. The Ti–Si samples were ground into
fine a powder, while the sapphire and copper standard
samples were used in the form of disks weighing approxi-
mately 18 and 50 mg, respectively. The Ti Si , TiSi and
3
. Results and discussion
Figs. 1–3 show heat capacities of Ti Si , TiSi and
5
3
5
3
TiSi , respectively, in the temperature region of approxi-
TiSi samples were all approximately 30 mg each. All the
2
2
mately 250–600 K. Solid lines on the plots represent
polynomial fits to the data in the form
samples were encapsulated in commercially available
aluminum pans. Variation in the masses of the pans chosen
for different samples was less than 0.05 mg.
2
22
C 5 A 1 BT 1 CT 1 DT
(1)
p
Temperature calibration of the Perkin-Elmer DSC2 in
the temperature range of 230–700 K was performed using
high-purity Hg, In, Pb and Zn metals and cyclohexane (the
cyclohexane sample was obtained from NIST) as reference
materials. Based on the differences between observed and
actual melting points of the references, corrections were
made to compensate for the discrepancy between actual
temperature and the programmed temperature of the
instrument.
Coefficients of the polynomial fits along with heat
capacities for the three titanium silicides at 298 K are
summarized in Table 2. The majority of the data are within
Heat capacity measurements were performed by the
scanning method with temperature intervals of 75 K and
2
1
scan rates of 0.333 K s . The samples were examined in
the following sequence: empty Al pan, Cu reference,
Ti Si , TiSi and TiSi samples, sapphire reference. Data
5
3
2
were compiled during an initial 180 s isothermal anneal at
the lowest temperature, followed by the heating of the
sample during 338 s, and finally by a 360 s isothermal
Table 1
Lattice parameters for Ti Si , TiSi and TiSi
2
5
3
˚
˚
˚
a, A
b, A
c, A
Ti Si
TiSi
TiSi2
7.45460.001
6.52260.006
8.27460.005
5.15060.001
5.02260.005
8.54460.005
5
3
3.64060.003
4.79560.002
Fig. 1. Heat capacity vs. temperature for Ti Si as measured by means of
5
3
DSC. (m) – data points; (——) – polynomial fit to all of the data.

S. Agarwal et al. / Journal of Alloys and Compounds 314 (2001) 99–102
101
Fig. 4. Percent deviations of the data from this study for Ti Si from
5
3
Fig. 2. Heat capacity vs. temperature for TiSi as measured by means of
DSC. (m) – data points; (——) – polynomial fit to all of the data.
previously reported experimental data: (h) – Archer et al. [5]; (d) –
Sychev et al. [8]; (m) – Golutvin [3].
with Golutvin’s data at temperatures between 300 and 450
K is significant, reaching approximately 30% at 300 K. In
a similar fashion, Fig. 5 compares the data for TiSi from
2
this study with the data of Archer et al. [7], Sylla et al. [4],
and Golutvin [2]. The data for TiSi from this study agree
2
with those of Archer et al. [7] and Sylla et al. [4] within 1
Fig. 3. Heat capacity vs. temperature for TiSi2 as measured by means of
DSC. (m) – data points; (——) – polynomial fit to all of the data.
1
% of the fitted values, while some of the data points agree
to within 2–3%.
Fig. 4 shows the percent deviation of the heat capacity
values for Ti Si obtained in this study from the previous-
5
3
ly published data of Archer et al. [5], Sychev et al. [8], and
Golutvin [3]. Agreement with the data of Archer et al. and
Sychev et al. is within 1 to 2%, while the disagreement
Fig. 5. Percent deviations of the data from this study for TiSi2 from
previously reported experimental data: (h) – Archer et al. [7]; (d) –
Sylla et al. [4]; (m) – Golutvin [3].
Table 2
a
Heat capacities at 298 K and heat capacity parameters for Ti Si , TiSi and TiSi in the temperature range of 250–600 K as measured by DSC
5
3
2
A
B
C
D
C at 298 K,
p
21
J mol
K
2
2
2
4
5
5
6
5
5
Ti Si
236
49.8
68.9
20.101
0.0177
0.0249
1.38310
21.18310
21.38310
23.07310
25.98310
27.56310
18163
47.260.6
66.660.9
5
3
TiSi
TiSi2
a
2
22
C 5 A 1 BT 1 CT 1 DT
.
p

1
02
S. Agarwal et al. / Journal of Alloys and Compounds 314 (2001) 99–102
over comparable temperature ranges. Archer’s values of
heat capacity for TiSi for temperatures above 350 K are
based primarily on reanalysis of Golutvin’s enthalpy
increments. The improved agreement between our direct
measurements of heat capacity and Archer’s calculations
for heat capacity in the temperature range of 250–600 K
would lend credence to Archer’s fitting procedure [6].
To summarize, in the temperature range of 250–350 K,
the measured heat capacity values for the three titanium
silicides are more consistent with the data reported by
Archer et al. [5–7] and others [4,8], then with the values of
Golutvin [3]. Archer’s reanalysis [6] of Golutvin’s data
would indicate that while Golutvin’s enthalpy increment
data are of reasonable accuracy, his heat capacity functions
fail at the lower temperatures of his data ranges. A suitable
combination of Archer’s thermal functions [5–7], the data
from this study, and, for temperatures above 600 K,
Golutvin’s heat capacity functions would be recommended
for use in the thermal functions for the titanium silicides.
Fig. 6. Percent deviations of the data from this study for TiSi from
previously reported experimental data: (h) – Archer et al. [6]; (m) –
Golutvin [3].
to 2%, but disagree considerably with the values given by
Golutvin [3]. Fig. 6 displays percent deviation of the heat
capacity values for TiSi obtained in this study from the
data of Archer et al. [6] and Golutvin [3]. Also included in
Fig. 6 are data gleaned by Archer [5] from analysis of
others’ (primarily Golutvin’s) enthalpy increment data.
The data from this study for TiSi and those of Archer
disagree by between 4 and 8%. Nevertheless, it can be
seen in Fig. 6 that the agreement with the data of Archer is
much better than with those of Golutvin, where deviations
of up to 50% are observed.
Acknowledgements
We acknowledge support from NSF DMR 9902783.
References
[1] S.P. Murarka, Silicides for VLSI Applications, Academic Press,
Orlando, FL, 1983.
[2] J. Engqvist, C. Myers, J. Carlsson, J. Electrochem. Soc. 139 (1992)
3197.
General agreement of the data obtained in this study
with the calculated heat capacity values of Archer would
tend to validate to his thermal functions. Agreement of 1 or
[3] Y.M. Golutvin, Russ. J. Phys. Chem. 33 (8) (1959) 164.
[
4] W.K. Sylla, S.E. Stillman, M.S. Sabella, E.J. Cotts, J. Appl. Phys. 76
5) (1994) 2752.
5] D.G. Archer, D. Filor, E. Oakley, E.J. Cotts, J. Chem. Eng. Data 41
1996) 571.
(
[
2
%, or the accepted capability of the differential scanning
(
calorimeter, was generally observed. In the case of TiSi, it
can be speculated that the 4–8% deviation between the
data from this study and those of Archer is due to
differences in compositions of the TiSi samples used in the
two investigations. Of the three Ti–Si phases of interest,
the single phase TiSi samples were found to be the most
difficult to prepare. High temperature annealing and room
temperature grinding were required for the TiSi samples to
limit the amount of secondary phases. Contamination by
foreign phases during these preparation procedures may
have caused the observed difference in the reported heat
capacity values.
[6] D.G. Archer, R.J. Kematick, C.E. Myers, E.J. Cotts, J. Chem. Eng.
Data 44 (1999) 167.
[7] D.G. Archer, M.S. Sabella, S.E. Stillman, E.J. Cotts, J. Chem. Eng.
Data 40 (1995) 1237.
[8] N.I. Sychev, P .V . Gel’d, G.I. Kalishevich, Russ. J. Phys. Chem. 54
(
1) (1980) 125.
[9] S .V . Meschel, O.J. Kleppa, J. Alloys Comp. 267 (1998) 128.
[10] V.N. Svechnikov, Y.A. Kocherzhinski, L.N. Yupko, O.G. Kulik, E.A.
Shishkin, Doklady Akademii Nauk 193 (2) (1970) 393.
11] N .V . Ageev, V.P. Samsonov, Russ. J. Inorg. Chem. 4 (7) (1959).
12] F. Laves, H. Wallbaum, Z. Kristallogr. 101 (1939) 78.
13] Perkin Elmer DSC-2 Operator Manual.
[14] E.S. Watson, M.J. O’Neill, J. Justin, N. Brenner, Anal. Chem. 36 (7)
(1964) 1233.
15] S.C. Mraw, in: C.Y. Ho (Ed.), Specific Heat of Solids, Hemisphere,
New York, 1988, p. 395.
[
[
[
[
In the temperature region of interest, approximately
2
30–600 K, reasonably good agreement with the data of
[
16] D.G. Archer, J. Phys. Chem. Ref. Data 22 (1993) 1441.
Golutvin was observed only near 600 K. However, good
agreement with the data of the other investigators is found