A thermogravimetric study of the reactions of molybdenum disilicide with anhydrous hydrogen fluoride and fluorine

Article abstract of DOI:10.1016/j.jfluchem.2012.10.002

The results of a thermogravimetric study into the dry fluorination of molybdenum disilicide, MoSi₂, using hydrogen fluoride and dilute fluorine gas as fluorinating agents are reported. The reaction between molybdenum disilicide and fluorine follows the thermodynamically preferred route, viz. the formation of the volatile molybdenum hexafluoride along with gaseous silicon tetrafluoride, with the reaction starting just below 200 °C. The reaction with hydrogen fluoride yields solid molybdenum metal and gaseous silicon tetrafluoride, similarly thermodynamically predicted, above 250 °C. No reaction is observed at low temperatures where solid molybdenum trifluoride is expected to form. The results of a kinetic analysis of the data for the reaction with hydrogen fluoride are reported. In the range 250-450 °C the kinetics are chemical reaction controlled. Above this, up to 700 °C, the rate is controlled by diffusion through the stagnant gas films surrounding the solid particles. Evidence for a third, un-quantified, high-temperature mechanism is given.

Full text of DOI:10.1016/j.jfluchem.2012.10.002

Journal of Fluorine Chemistry 145 (2013) 66–69
Contents lists available at SciVerse ScienceDirect
Journal of Fluorine Chemistry
journal homepage: www.elsevier.com/locate/fluor
A thermogravimetric study of the reactions of molybdenum disilicide with
anhydrous hydrogen fluoride and fluorine
a,
a
b
J.S. Gama *, J.B. Wagener , P.L. Crouse
a
Applied Chemistry Department, The South African Nuclear Energy Corporation Ltd. (Necsa), P.O. Box 582, Pretoria 0001, South Africa
Department of Chemical Engineering, University of Pretoria, Pretoria 0002, South Africa
b
A R T I C L E I N F O
A B S T R A C T
Article history:
The results of a thermogravimetric study into the dry fluorination of molybdenum disilicide, MoSi2,
using hydrogen fluoride and dilute fluorine gas as fluorinating agents are reported. The reaction between
molybdenum disilicide and fluorine follows the thermodynamically preferred route, viz. the formation of
the volatile molybdenum hexafluoride along with gaseous silicon tetrafluoride, with the reaction
starting just below 200 8C. The reaction with hydrogen fluoride yields solid molybdenum metal and
gaseous silicon tetrafluoride, similarly thermodynamically predicted, above 250 8C. No reaction is
observed at low temperatures where solid molybdenum trifluoride is expected to form. The results of a
kinetic analysis of the data for the reaction with hydrogen fluoride are reported. In the range 250–450 8C
the kinetics are chemical reaction controlled. Above this, up to 700 8C, the rate is controlled by diffusion
through the stagnant gas films surrounding the solid particles. Evidence for a third, un-quantified, high-
temperature mechanism is given.
Received 20 September 2012
Received in revised form 3 October 2012
Accepted 4 October 2012
Available online 23 October 2012
Keywords:
Molybdenum disilicide
Anhydrous hydrogen fluoride
Thermogravimetric analysis
Kinetics
ß 2012 Elsevier B.V. All rights reserved.
1
. Introduction
such as high melting points, super high oxidation resistance,
2
formation of a thin protective surface silica layer (SiO ), and a high
Four molybdenum fluorides are known to exist, viz. MoF
3
, MoF
4
,
creep resistance [9–12]. Very little is known of the fluorination of
molybdenum silicides using anhydrous hydrogen fluoride (AHF) or
fluorine.
MoF
primarily in the electronics industry [5]. Usually the volatile,
hygroscopic molybdenum hexafluoride (MoF ) is formed by
reacting molybdenum metal with strong fluorinating agents, such
as F , BrF , and SF [6]. Exploding molybdenum metal into SF gives
a conversion of more than 70% to MoF , depending on the SF /Mo
was reported to be present in minor
5 6 6
and MoF [1–4], but only MoF has commercial applications,
6
Thus, in this article we present results on the fluorine chemistry
2
of molybdenum disilicide (MoSi ) using thermogravimetry as tool,
2
3
6
6
where MoSi was treated with both fluorine and anhydrous
2
6
6
hydrogen fluoride, without using any material as a fuse or as an
auxiliary substance.
ratio, while non-volatile MoF
quantities [1].
3
Equilibrium thermodynamic calculations predict that stoichio-
In most cases the lower molybdenum fluorides (MoF
and MoF ) are formed through the reduction of high oxidation
state metal fluorides [7].
However, it has been shown that the treatment of MoSi
3
, MoF
4
,
metric ratios of MoSi
all temperatures up to 1000 8C (Eq. (1)).
2
and F
2
should produce MoF
6
(g) and SiF
4
(g) at
5
MoSi
2
ðsÞ þ 7F
2
ðgÞ ! MoF
6
ðgÞ þ 2SiF ðgÞ
4
(1)
2
with
gaseous fluorine in the presence of tungsten (as an auxiliary
substance) and sulphur (as a fuse) achieve complete conversion of
In the case of the reaction with HF, the equilibrium
thermodynamic calculation with a stoichiometric ratio of 1:8
2 6 4
MoSi to MoF and SiF [8]. The objective of this work was to
MoSi
2
/HF predicts the formation of molybdenum metal, silicon
2
determine the enthalpy of formation of MoSi at a specific
tetrafluoride and hydrogen gas (Eq. (2)). While with excess HF
temperature and pressure.
(
(
8
1:11 MoSi
Fig. 1) predicts the same products as above 80 8C, but below
0 8C MoF should form as shown in Eq. (3).
2
/HF), the equilibrium of thermodynamic calculation
2 3 5 3
Molybdenum silicides (MoSi , Mo Si, and Mo Si ) are known
for high temperature applications, because of physical properties
3
MoSi
MoSi
2
þ 8HFðgÞ ! Mo þ 2SiF
4
ðgÞ þ 4H
2
ðgÞ
(2)
(3)
*
Corresponding author. Tel.: +27 12 305 6135; fax: +27 12 305 3197.
E-mail address: jabulani.gama@necsa.co.za (J.S. Gama).
2
þ 11HFðgÞ ! MoF ðsÞ þ 2SiF
3
4
ðgÞ þ 5:5H ðgÞ
2
0022-1139/$ – see front matter ß 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jfluchem.2012.10.002

J.S. Gama et al. / Journal of Fluorine Chemistry 145 (2013) 66–69
67
Fig. 3. Dynamic thermogravimetric curve of the reaction between MoSi
in nitrogen.
2
and 10% HF
2
Fig. 1. Equilibrium composition for the reaction between MoSi and HF (1:11 mol/
mol).
observed. The mass loss increased as the temperature was
increased, with the reactions at 600 8C and 700 8C having the
highest mass loss of 32.7%. Molybdenum metal was the major
phase by XRD analyses only at 700 8C. In the temperature range of
2
. Results and discussion
2
00–600 8C, MoSi
2
was the major phase, indicating incomplete
The dynamic thermogravimetric fluorination of MoSi
indicates that fluorine completely converts MoSi into the volatile
fluorides MoF and SiF (Fig. 2). Once the ignition temperature of
about 180 8C has been achieved, the reaction proceeds easily to
yield the volatile products and is complete at 330 8C. This result
corresponds well with the thermodynamic prediction (Eq. (1)).
2 2
with F
fluorination of MoSi
2
.
2
Scanning electron microscope (SEM) images of the product at
different temperatures are given in Fig. 5. At 200 8C and 300 8C no
major morphological changes are evident, which can be ascribed to
the incomplete fluorination of MoSi . At 700 8C, where molybde-
2
num metal was the major phase there are clear changes and
particle shrinkage observed (Fig. 5).
To understand the mechanism of the regimes observed in Figs. 3
and 4, kinetic parameters were extracted. A number of kinetic
models are possible for fluid–solid reactions, in this case the
following assumptions were made: the particles are spherical and
6
4
The dynamic thermogravimetric reaction between MoSi
0% AHF in nitrogen started with a small mass loss of about 2%
Fig. 3). This is possibly due to the removal of the silicon oxide layer
SiO ), which forms when MoSi is exposed to the atmosphere.
2
and
1
(
(
2
2
After what appears to be an induction phase, full ignition is
achieved at about 250 8C. The thermogravimetric curve suggests
three distinct regimes, roughly in the ranges 250–550 8C, 550–
2
the reaction occurs first at the outer surface of the MoSi particles,
with the zone of the reaction moving toward the center, leaving
behind completely converted material and unert solid. Therefore,
at anytime there is an unreacted core of material which shrinks in
size during the reaction. Thus for this work the three models of the
shrinking core (SCM) were considered (Table 1). The three models
with their fractional residue,
Table 1. The full derivation of these models can be found in, e.g.
Levenspiel [13]. Note that is the fractional residue, not the extent
of reaction. In all cases, , is the molar density of the solid reactant,
R is the initial particle size (assuming full sphericity), and C is the
reactant gas concentration. In addition, k is the chemical reaction
rate constant, k is the mass transport coefficient through the
6
50 8C, and 650–800 8C.
These three regimes resulted in a mass loss of about 31%, with a
total mass loss of about 33%, as compared to the calculated 37.1% of
a full conversion to Mo metal (Eq. (2)). Analysis of the final product
by X-ray diffraction (XRD) showed molybdenum metal as the
major phase. X-ray fluorescence (XRF) also showed trace amounts
of silicon.
a, as a function of time, are listed in
a
When the HF reactions were conducted isothermally (Fig. 4) in
the temperature range of 200–700 8C, the theoretical mass loss of
r
B
g
3
7.1% was also not obtained. In this case only two regimes were
00
g
Fig. 2. Dynamic thermogravimetric curve of the reaction between MoSi
2
and 10% F
2
2
Fig. 4. Isothermal thermogravimetric curves of the reaction between MoSi and 10%
in nitrogen.
HF in nitrogen.

6
8
J.S. Gama et al. / Journal of Fluorine Chemistry 145 (2013) 66–69
2
Fig. 5. SEM micrographs of MoSi , untreated and after reaction with HF at different temperatures.
stagnant gas film, and D
the product layer. The symbol b is the stoichiometric coefficient for
e
is the effective diffusion constant through
in the high-temperature region a better fit was obtained for gas-
film diffusion control (Table 2).
the general gas–solid reaction:
Overall, the ash-layer control model fitted poorly in all cases. In
the temperature range investigated, only two overlapping
mechanisms could be identified, which corresponds to the first
two regimes in Fig. 3. Between 250 8C and 450 8C, it appears that
the chemical reaction itself controls the process, while between
450 8C and 700 8C, diffusion through the stagnant gas film
surrounding the particles seems to be rate controlling. The third
regime observed in the dynamic TG run for the reaction of HF with
MoSi2 (Fig. 3), evidently takes place at a temperature above this
and could not be identified from the isothermal results. The
Arrhenius parameters (activation energies and pre-exponential
factors) are given in Table 3.
AðgÞ þ bBðsÞ ! Products
(4)
For model fitting purposes, the fractional residue,
a
, was
referenced to the amount of silicon present. That is, m
sample mass was taken as the mass of silicon present in the silicide,
was taken as the final mass of silicon present after the reaction.
The time-dependent mass was taken as the sample mass minus the
mass of metal (Mo) present. The span of is thus 1 to 0 in each case,
and since the LHS of the model equation tend to zero at t = , the
linear data plots were done for a zero y-intercept in each case.
0
, the initial
m
f
a
t
m
0
a
¼
(5)
m f
Table 2
R-Square values for better fit.
2
None of the models tested gave a good fit at 200 8C with R
Model
R
2
values of 0.9161, 0.9141 and 0.8526, suggesting that the reaction
only fully develops somewhat above 200 8C. Between 250 8C and
2
50–450 8C
500–700 8C
Chemical reaction control
0.999
0.654
0.689
4
50 8C a better fit was obtained for chemical reaction control, while
Diffusion through product layer
Diffusion through stagnant gas film
0.680
0.911
Table 1
Gas–solid reaction models evaluated.
Model
f(
a
)
t
t
Table 3
Arrhenius parameters.
Chemical reaction
control
a1=3
t
t
r R
B
1
ꢀ
¼
¼
bk C
00
g
Model
Ea
kJ/mol)
Pre-exponential
ꢀ
1
(
factor (m s or
2 ꢀ1
m s )
rB
R
t
t
Control by diffusion
1 ꢀ
a
¼
t
t
¼ 3bk
g
C
g
through stagnant gas film
2
rB
R
Chemical reaction control
29.54
12.56
0.032
0.329
Control by diffusion through product layer 1 ꢀ 3a2=3 þ 2
a
¼
t
t
¼ 6
e g
bD C
Diffusion through stagnant gas film

J.S. Gama et al. / Journal of Fluorine Chemistry 145 (2013) 66–69
69
3
. Conclusions
6 4
MoSi can be fully converted to MoF and SiF by reacting with
to the desired temperature using a rate of 10 8C/min. The diameter
of the furnace tube was 10 mm. The thermocouple was inserted
through the inlet of the reactive gas below the sample, and
positioned just underneath the TG pan to accurately measure the
temperature. The reactive gases were introduced at a flow rate of
60 mL/min.
2
dilute elemental fluorine. The reaction with gaseous HF only fully
develops above 250 8C, and yields pure molybdenum metal along
with SiF at high temperatures (700 8C). Three kinetic regimes
4
were detected in the dynamic thermogravimetric reaction with HF
in three different temperature ranges (250–450 8C, 500–700 8C and
above 700 8C). Arrhenius parameters for the first two of these
regimes could be extracted from the isothermal thermogravi-
metric results.
The products obtained from these reactions were analyzed
using Bruker A-D8 Advanced XRD for phase identification, ZEISS
Gemini Ultra Plus Field Emission Gun (FEG) SEM for image analysis
and EDX and Panalytical Axios XRF spectrometer for elemental
analysis.
4
. Experimental
Acknowledgements
4
.1. Materials
The South African Nuclear Energy Corporation (Necsa) in whose
laboratories the work was done, the Graduate-in-Training (GIT)
program at Necsa, and the financial support of the South African
National Research Foundation (NRF) are gratefully acknowledged
by the authors. Mr. B.M. Vilakazi of Necsa is acknowledged for his
assistance in obtaining the thermogravimetric data.
The molybdenum disilicide was purchased from Sigma–
Aldrich. The dilute anhydrous hydrogen fluoride and fluorine
gas (10% HF or F diluted in N ) were obtained from Pelchem (Pty)
Ltd. All reagents had a purity of >99%.
2
2
4.2. Experimental procedure
References
An adapted thermogravimetric analyzer (TGA) was used for
conducting the experiments. The instrument was modified to
handle the corrosive gases HF and fluorine. Details are given in
[
[
[
1] R.L. Johnson, B. Siegel, Journal of Inorganic and Nuclear Chemistry 31 (1968) 663–
955.
2] D.E. LaValle, R.M. Steele, M.K. Wilkinson, H.L. Yaker Jr., Journal of the American
Chemical Society 82 (1960) 2433–2434.
3] F.A. Cotton, G. Wilkinson, Advanced Inorganic Chemistry, 4th edition, John Wiley
and Sons, New York, 1962.
Gama et al. [14]. A starting mass of 40–60 mg MoSi
each set of experiments. The MoSi was stored in a glove box under
nitrogen and treated with special consideration to minimize the
oxygen present, even though MoSi is known as reasonably
2
was used for
2
[
[
4] M.L. Larson, F.W. Moore, D.A. Edwards, Inorganic Synthesis 12 (1970) 165–178.
5] K. Bjorklund, Microfabrication of Tungsten, Molybdenum ans Tungsten Carbide
Rods by Laser-assisted CVD, ACTA Universitatis upsaliensis, Uppsala, Sweden,
2001 http://uu.diva-portal.org/smash/record.jsf?pid=diva2:161180.
2
oxidation resistant. Two thermogravimetric methods were used,
viz. non-isothermal (dynamic) and isothermal. For isothermal
experiments the sample was first heated to the predetermined
temperature and then maintained at this constant temperature to
equilibrate before introducing the reactive gas. For non-isothermal
experiments the reactive gas was introduced at the beginning of
each reaction and a heating rate of 10 8C/min was used. Isothermal
experiments were only carried out with HF(g), since the fluorine
reaction was considered to be of less interest.
[6] O. Glemser, Journal of Fluorine Chemistry 33 (1986) 45–69.
[
[
7] R.B. Heslop, P.L. Robinson, Inorganic Chemistry, Elsevier, Amsterdam, 1967.
8] P.A.G. O’Hare, Journal of Chemical Thermodynamics 25 (1993) 1333–1343.
[9] Z. Yao, J.J. Stiglich, T.S. Sudarshan, Journal of Materials Engineering and Perfor-
mance 8 (1999) 291–304.
10] N.N. Greenwood, A. Earnshaw, Chemistry of the Elements, Elsevier, Oxford,
[
1
984 .
[
11] K. Natesan, S.C. Deevi, Intermetallics 8 (1999) 1147–1158.
[12] V.A. Maksimov, Inorganic Materials (Neorganicheskie Materialy) 5 (1996) 965–
66.
13] O. Levenspiel, Chemical Reaction Engineering, John Wiley, New York, 1996.
14] J.S. Gama, J.B. Wagener, P.L. Crouse, Journal of Fluorine Chemistry 139 (2012)
12–15.
9
The isothermal reactions were done at temperatures ranging
between 200 8C and 700 8C. The samples were placed and heated in
a small TG pan (inside diameter of 5.6 mm and a height of 1.6 mm)
[
[