Reduction of arsenic trichloride with hydrogen

Article abstract of DOI:10.1134/S1070427207010016

The composition of an equilibrium vapor-gas mixture in the range 700-1100 K was estimated for the reaction of arsenic trichloride with hydrogen taken in excess of a = 1.0-2.5 as compared to the stoichiometry. The temperatures of the onset of arsenic condensation (dew point) were found. Kinetic parameters of the chemical reaction were determined. Optimal conditions for the reduction of arsenic trichloride with hydrogen were refined and conditions for the condensation of arsenic with the minimum fraction of the amorphous modification were found.

Full text of DOI:10.1134/S1070427207010016

ISSN 1070-4272, Russian Journal of Applied Chemistry, 2007, Vol. 80, No. 1, pp. 1 6. Pleiades Publishing, Ltd., 2007.
Original Russian Text N.A. Potolokov, A.V. Serov, V.N. Potolokov, A.V. Zhuravlev, V.P. Kolganov, E.G. Zhukov, V.A. Fedorov, V.I. Kholstov, 2007,
published in Zhurnal Prikladnoi Khimii, 2007, Vol. 80, No. 1, pp. 3 8.
INORGANIC SYNTHESIS
AND INDUSTRIAL INORGANIC CHEMISTRY
Reduction of Arsenic Trichloride with Hydrogen
N. A. Potolokov, A. V. Serov, V. N. Potolokov, A. V. Zhuravlev, V. P. Kolganov,
E. G. Zhukov, V. A. Fedorov, and V. I. Kholstov
Research Institute of Electronics Materials, Open Joint-Stock Company, Kaluga, Russia
Kurnakov Institute of General and Inorganic Chemistry, Russian Academy of Sciences, Moscow, Russia
Kaluga Branch, Timiryazev Moscow Agricultural Academy, Kaluga, Russia
Federal Industry Agency, Moscow, Russia
Received March 3, 2006
Abstract The composition of an equilibrium vapor gas mixture in the range 700 1100 K was estimated
for the reaction of arsenic trichloride with hydrogen taken in excess of a = 1.0 2.5 as compared to the
stoichiometry. The temperatures of the onset of arsenic condensation (dew point) were found. Kinetic param-
eters of the chemical reaction were determined. Optimal conditions for the reduction of arsenic trichloride
with hydrogen were refined and conditions for the condensation of arsenic with the minimum fraction of
the amorphous modification were found.
DOI: 10.1134/S1070427207010016
The reaction of gaseous arsenic trichloride with
hydrogen is one of the most important stages of the
preparation of high-purity elemental arsenic [1 4].
AsCl3g + 1.5H2g
0.25As4g + 3HClg,
(1)
(2)
Keq = exp( G0T /RT).
Systematic study of AsCl reduction with hydrogen is
3
Such representation of the AsCl transformation to
arsenic is admissible, as it was shown previously
3
required for the development of physical and chemical
principles for producing high-purity arsenic from an
unconventional raw material, products of the detoxica-
tion (utilization) of chemical weapons (lewisite).
The feasibility of organizing industrial production of
high-purity arsenic-containing compounds on the
basis of lewisite has been substantiated in [5 7].
[5, 7] that As is the main arsenic species in the vapor
4
phase at temperatures below 1100 K in the presence
of a hydrogen excess < 5 against the stoichiometry.
Equilibrium constants (2) of reaction (1) were calcu-
lated for the range 700 1100 K and an excess (
=
1 2.5) of hydrogen against the stoichiometry by two
known methods:
It was found by studying thermodynamic and ki-
netic features of the AsCl reduction with hydrogen
3
GT0
=
H0298
T S0298
[ni(S0T
H0298
+
[ni(H0T
ST298)i],
T**,
HT298)i]
[8 10] that the degree of conversion of arsenic tri-
chloride is sufficiently high even at temperatures
above 600 K and small hydrogen excesses against
a stoichiometry. The chemical equilibria for real reac-
tion conditions were simulated in [11].
T
(3)
(4)
GT0
=
T
where
** is the reduced thermodynamic potential.
T
This paper deals with an experimental study of the
reduction of AsCl with hydrogen. Appreciable atten-
3
Both methods give similar K values varying from
3.4 10 to 6.0 10 atm
to 100 kJ mol . From the resulting K values, we
estimated the composition of the equilibrium vapor
gas mixture for a total pressure of 1 atm in the reac-
tion zone. The calculated data are given in Fig. 1,
eq
4
4
0.75
0
tion is given to the process kinetics, and also to the
optimization of conditions for the production of high-
purity arsenic from unconventional raw materials,
products of lewisite detoxication and wastes from
nonferrous metallurgy.
for G varied from 60
T
1
eq
which shows that the calculated partial pressures P
eq
EXPERIMENTAL
of arsenic vapor are practically independent of the
process temperature and decrease by a factor of 1.7
with increasing from 1 to 2.5. The partial pressure
First we made a thermodynamic estimation of the
arsenic trichloride reduction with hydrogen:
1

2
POTOLOKOV et al.
becomes heterogeneous.
AsCl3g + 1.5H2g
log P [atm]
Ass + 3HClg.
(6)
Thus, we can expect that AsCl will be reduced by
3
reactions (1) and (6) simultaneously below the dew
point. In this case, the contribution of reaction (6) in-
creases as the temperature decreases and can be esti-
mated from expression (7).
785 K
D = (Peq
P)/Peq,
(7)
769 K
where P is the equilibrium pressure of arsenic vapor
eq
for reaction (1). The process can be considered as
passing homogeneously (1) in the gas phase with the
subsequent condensation of arsenic vapor:
1
103/T, K
As4g
4Ass.
(8)
Fig. 1. Temperature dependence of arsenic pressure P under
equilibrium conditions. : (1) 1.0 and (2) 2.5; (3) saturated
vapor pressure of arsenic.
The kinetics of AsCl reduction with hydrogen was
3
studied on an installation schematically shown in
Fig. 2. The main element of the installation is a quartz
reactor consisting of two sections. A high-temperature
zone is constructed as a system of three coaxial cyl-
inders ensuring a preset residence time of the reaction
mixture (AsCl + H ) at a certain temperature.
of hydrogen decreases by a factor of 1.26 with in-
creasing temperature from 700 to 1100 K at = 1, but
> 1.5 it becomes practically temperature-in-
dependent. The equilibrium concentration of AsCl
at
3
3
2
remains insignificant at all values and decreases as
A water-cooled condenser ensuring the deposition
of arsenic in the form of a compact coarsely crystal-
line ingot is arranged in the low-temperature zone of
the reactor. The reactor is placed inside a multizonal
resistance heating unit where the required temperature
of the reaction and arsenic condensation zones is
created.
increases, confirming a high degree of AsCl con-
3
version to arsenic (Table 1).
The temperature dependence (5) of the saturated
vapor pressure of arsenic over solid arsenic [8] is also
shown in Fig. 1.
logP [atm] = 8140/T + 9.244.
(5)
We used two methods for the determination of the
degree of arsenic trichloride conversion to elemental
It is seen that the straight line crosses the horizont-
al lines of P in points corresponding to 785 and
As
arsenic: (1) direct monitoring of the AsCl content
3
769 K for = 1 and 2.5, respectively. These are the
temperatures that are dew points for the initial pro-
portions of the reactants and correspond to the condi-
tions of the onset of arsenic condensation and to the
lower temperature boundary of correct thermodynamic
calculations for reaction (1). At lower temperatures,
reaction (6) occurs, and the homogeneous process
in gaseous reaction products by sampling and sub-
sequent analysis of the samples and (2) determination
of the amount of the resulting elemental arsenic from
the loss of a certain amount of the starting AsCl .
3
The first method makes it possible to reduce the
experiment duration and to make several parallel
analyses with a single portion of AsCl . The second
3
method is more labor-consuming; however, it provides
an integral characteristic of the process kinetics in
the H AsCl system where the effect of fluctuations
Table 1. Degree of arsenic trichloride conversion to
arsenic as a function of temperature and hydrogen excess
2
3
of process conditions will be leveled off. In this case,
however, an essential complication is permanent for-
mation of finely dispersed amorphous modification of
arsenic, which is partially carried away from the con-
denser by a gas stream. Therefore, it is necessary to
take into account the amount of elemental arsenic
precipitated in the absorbing system.
, %, at indicated temperature, K
700
800
900
1000
1100
1
97.32
99.97
99.99
99.99
97.55
99.97
99.99
99.99
97.74
99.98
99.99
99.99
97.78
99.98
99.99
99.99
97.86
99.98
99.99
99.99
1.5
2.0
2.5
We used a combined technique in the experiments,
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 80 No. 1 2007

REDUCTION OF ARSENIC TRICHLORIDE WITH HYDROGEN
3
Hydrogen
Nitrogen
Hydrogen
to be burnt
Fig. 2. Scheme of a pilot installation for studying kinetics of AsCl reduction with hydrogen: (1 8) control stopcocks, (9, 10) gas
3
pressure regulators for pressure adjustment, (11) installation for diffusion purification of hydrogen, (12, 13) heaters of high-tem-
perature and low-temperature zones, (14, 15) flow meters, (16) reactor for arsenic trichloride reduction with hydrogen, (17) ves-
sel with AsCl , (18 20) vessels for absorption of hydrogen chloride and residual arsenic trichloride, and (21, 22) manometers.
3
determining the total amount of unchanged AsCl for
(chemical reaction duration) from expression (10)
tr = VaT0/(Q1 + Q2)Tr, (10)
where t is the reaction duration (s); V , volume of
3
the entire experiment duration with a sufficiently large
amount of the starting substance (about 200 g). As a
result, the advantages of the second method were
realized and the error originating from incomplete col-
lection of elemental arsenic formed was decreased. To
check the correctness of the results obtained, we meas-
ured in addition the total amount of chlorine in the
system for absorption of gaseous reaction products.
r
a
the reaction zone (l); T , absolute temperature (K); Q
0
1
and Q , flow rates of vaporous AsCl and gaseous
2
3
hydrogen, respectively, reduced to the normal condi-
1
tions (l s ); and T , reaction temperature (K). Gen-
erally the average value of T was determined by the
graphical integration of the temperature variation
along the reaction zone.
r
r
To describe the kinetics of the arsenic trichloride
reduction, we used the approach developed in [8, 12,
13]. The rate of reaction (1) is determined by equa-
tion (9).
The concentrations of components in the starting
vapor gas mixture were calculated by formulas (11)
and (12).
r = dc/dt = kcXcYH,
(9)
1
1
c0 = T0/V0(1 + 1.5 )Tr,
(11)
(12)
where r is the reaction rate (mol l s ); k, the reac-
tion rate constant; c and c , concentrations of arsenic
cH(0) = 1.5 T0/V0(1 + 1.5 )Tr,
H
trichloride and hydrogen in the vapor gas mixture
(M); X and Y, partial reaction orders with respect to
arsenic trichloride and hydrogen.
where
is the ratio of the amount of hydrogen fed
(mol) to its stoichiometric amount; V , volume of
22.4 l, which is occupied by 1 mol of ideal gas under
normal conditions.
0
The reaction rate constant in a series of experiments
can be calculated from the AsCl residual concentra-
3
tion at a fixed residence time of the reaction mixture
in the reaction zone. We found the residence time
To determine the partial reaction orders and the
reaction rate constant, it is necessary to solve rate
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 80 No. 1 2007

4
POTOLOKOV et al.
Table 2. Experimental and calculated parameters of the reduction of arsenic trichloride with hydrogen
= 1
fexp
= 1
%
fcalc
= (
calc)/
exp
exp
calc
exp
Tr, K
t, s
1072
1072
1072
1072
1026
932
886
980
980
980
980
980
980
1072
1072
1.03
1.29
1.88
2.60
1.67
1.97
1.91
1.98
1.79
1.96
2.36
4.03
2.01
1.58
1.68
16.1
17.5
19.7
21.2
25.0
25.4
25.2
20.0
14.6
9.9
10.4
11.4
8.0
11.6
10.0
88.4
96.9
99.8
99.8
99.3
99.0
97.9
99.5
99.0
98.8
99.0
99.4
97.2
99.1
99.2
90.6
97.8
99.7
99.9
99.4
98.9
97.7
99.3
98.7
98.5
99.1
99.6
98.3
99.0
99.1
2.5
0.9
0.1
0.1
0.1
0.1
0.2
0.2
0.3
0.3
0.1
0.2
1.1
0.1
0.1
equation (9). Let us introduce a dimensionless vari-
and excess of hydrogen against the stoichiometric
amount = 1.02 4.03. The ranges of variation of
these variables were chosen on the basis of the exami-
nation of real process modes and the data of [8, 10].
able f, the fraction of unchanged AsCl , defined by
3
expression (13):
f = c /c0.
(13)
Typical experimental and calculated data are given
in Table 2. To determine partial reaction orders, we
used the method of finding a minimum of the depen-
dence of the standard deviation of the reaction rate
constant (and also of its logarithm) on X and Y calcu-
lated by Eq. (16) for each series of the experiments
Then, taking into account expression (14), we can
write down the rate equation of the chemical reaction
in form (15):
cH = 1.5 c,
(14)
(15)
(X+ Y
df/[f X(f + )Y] = 1.5Ykc0
1)dt,
satisfying the requirement T = const. The application
r
of such a procedure gave the following reaction orders
with respect to arsenic trichloride and hydrogen: X =
1.86 and Y = 2.78. The order of reaction (1) in the
system AsCl H is 4.6, suggesting its complicated
where
=
1. In view of the fact that at t = 0 f =
r
1, the integral form of the rate equation is
3
2
1
mechanism.
(X+ Y
df/[f X(f + )Y] = 1.5Y kc0
1) tr.
(16)
Using the X and Y values obtained, we solved
Eq. (16), calculated the reaction rate constant for the
whole array of the experimental data, and determined
its temperature dependence (17).
f
The integration of the left-hand side of Eq. (16) re-
presents a certain problem, as the form of the integral
depends on X and Y values, and only when X and Y
are equal to integer and(or) half-integer numbers the
integral can be expressed in elementary functions. In
this connection, the problem was solved in such a way
as to obtain an array of experimental concentrations of
lnk = B
A/T,
(17)
where A = 15287 and B = 36.509.
The apparent activation energy E is approximately
1
unchanged AsCl by varying independent variables
127 kJ mol . The frequency factor K , determined in
3
0
and to determine the partial reaction orders and the
temperature dependence of the reaction rate constant
on this basis.
the collision theory from the relationship lnK = B, is
7.17 10 (mol l )
rate constant at 800 K exceeds 3.6 10 and increases
to approximately 1.7 10 (mol l )
0
15
1
4.6
1
s . We note that the reaction
7
9
1
4.6
1
s
at 1000 K.
The ranges of variation of the independent param-
eters were as follows: temperature of the reaction
The degree of arsenic trichloride conversion under
these conditions exceeds 99% at the residence time of
zone T = 886 1071 K, reaction duration = 5 28 s,
r
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 80 No. 1 2007

REDUCTION OF ARSENIC TRICHLORIDE WITH HYDROGEN
5
the mixture in the reaction zone longer than 15 s and
at a hydrogen excess > 1.5.
1
A, g min
The data obtained allowed us to calculate the de-
gree of arsenic trichloride conversion (%). The
results of this calculation satisfactorily agree with the
experiment (Table 2). The relative difference between
the calculated and experimental data does not exceed
40% when estimating the fraction of unchanged AsCl
and 4% when determining the degree of its conver-
3
sion. The rate constants of AsCl reduction with
3
hydrogen were used to optimize operation conditions
for the production of high-purity arsenic on a pilot
installation.
The calculation results are plotted in Fig. 3. It is
seen that the maximal capacity of about 3 g min of
Fig. 3. Dependence of the pilot installation capacity for
1
AsCl reduction with hydrogen A on hydrogen excess for
3
various temperatures and 99.9% conversion. Temperature,
arsenic at the preset 99.9% conversion and tempera-
ture of 1100 K is reached at = 3.4. When choosing
the reaction temperature, the possibility of contamina-
tion of the desired arsenic by apparatus impurities
should be taken into account. In particular, at tem-
peratures exceeding 1100 K the resulting arsenic con-
tains an increased amount of silicon. The decrease in
temperature below 900 K results in an appreciable
decrease in the reaction rate. The increase in the hy-
drogen excess over 3.4 reduces the output capacity.
In practice, the reduction of arsenic trichloride with
hydrogen is carried out with a hydrogen excess of
1.5 2, as at > 2 the fraction of amorphous arsenic
appreciably increases because of higher rates of the
gas mixture at the reactor outlet and, as a conse-
quence, greater breakthrough of arsenic vapor into the
cold part of the reactor where arsenic condenses in
the amorphous state.
K: (1) 900, (2) 950, (3) 1000, (4) 1050, and (5) 1100.
Tc, K
Fig. 4. Dependence of the amorphous arsenic content
c
on the temperature T of the condensation zone.
As(am.)
c
condensation zone does not exceed 1.5%.
The found conditions for the reduction of arsenic
compounds with hydrogen were used as the basis for
constructing a pilot installation for the production of
high-purity arsenic from such unconventional raw
materials as products of the detoxication of chemical
weapons (lewisite) and waste from nonferrous metal-
lurgy.
We determined the optimal temperature of the con-
densation zone from the viewpoint of minimizing the
amount of amorphous arsenic. We performed a series
of experiments in which the temperature in the zone
of the maximal thickness of the product layer was
varied and the fraction of amorphous arsenic was de-
termined. The results are given in Fig. 4, which shows
that, at temperatures higher than 700 K, the amorph-
ous arsenic fraction increases considerably (from 7.2
to 22% at T > 730 K). In the range 700 670 K, this
fraction varies only slightly. Thus, the optimal tem-
perature range where the residual pressure of arsenic
The content of impurities in samples of high-purity
arsenic prepared from lewisite is shown in Table 3.
The impurity contents were monitored by spectral
analysis with excitation of the spectrum in a dc arc
and by atomic-emission spectroscopy with excitation
of the spectrum in an inductively coupled plasma dis-
charge. The high detection limit of impurities (up to
3
is lower than 5 10 atm is 670 705 K.
8
10 wt %) was reached by applying extraction pre-
To decrease arsenic removal in the form of aerosols
from the reactor, it is necessary to use a trapping sys-
tem. In our case, we used a special unit in which the
direction of the gas stream changes and residual
amorphous arsenic is trapped by a system of bubble
absorbers. The amount of arsenic removed from the
concentration of impurities in the sample preparation
stage.
It should be noted that the purity of amorphous and
crystalline arsenic modifications is practically the
same. Nevertheless, from the viewpoint of conveni-
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 80 No. 1 2007

6
POTOLOKOV et al.
Table 3. Content of impurities (A) in samples of high-
purity arsenic
(4) The degree of AsCl conversion to arsenic was
shown to be no less than 99.99%, and the minimal
amount of amorphous As, 7.2 %.
3
A
106, wt %
Impurity
ACKNOWLEDGMENTS
crystalline
arsenic
amorphous
arsenic
detection limit
This study was performed within the framework
of the Program of Presidium of the Russian Academy
of Sciences Support of Innovations and was finan-
cially supported by the Leading Scientific Schools
grant of the Russian Foundation for Basic Research
(grant no. NSh-4895.2006.3).
Al
Zn
Te
Si
<1.0
0.1
2.0
0.5 0.1
4.0
<1.0
0.1
2.0
2 0.5
2.0
2.0
<0.5
<0.1
<0.1
<0.1
<0.5
<5.0
1.0
0.03
0.6
0.06
1.0
0.1
0.01
0.01
0.07
0.1
0.1
1.0
S
Fe
Mn
Mg
Cu
Na
Cr
Pb
0.7
<0.5
<0.1
<0.1
<0.1
<0.5
<5.0
REFERENCES
1. Rtskhiladze, V.G., Mysh’yak (Arsenic), Moscow:
Metallurgiya, 1969.
2. Nisel’son, L.A., Yaroshevskii, A.G., Gasanov, A.A.,
et al., Vysokochist. Veshch., 1993, no. 4, pp. 62 64.
3. Fedorov, V.A., Pashinkin, A.S., Efremov, A.A., and
Grinberg, E.E., Vysokochist. Veshch., 1991, no. 5,
pp. 7 30.
ence in the technology of semiconductor materials,
amorphous arsenic is further sublimed to convert it to
the crystalline form. The arsenic thus obtained is of
99.9999 wt % purity (6N grade) and is highly com-
petitive with products of leading companies from the
United States (Johnson Matthey) and Japan (Furuka-
wa) produced from the traditional raw material, arsen-
ic ores.
4. Fedorov, V.A., Sovremennye problemy obshchei i
neorganicheskoi khimii. Sbornik k 70-letiyu Inst.
obshchei i neorganisheskoi khimii im. N.S. Kurnakova
Ross. Akad. Nauk (Modern Problems of General and
Inorganic Chemistry. Coll. of Works Dedicated to
70th Anniversary of Kurnakov Institute of General
and Inorganic Chemistry, Russian Acad. Sci.), Mos-
cow: Nauka, 2004, pp. 464 472.
The use of high-purity arsenic obtained from un-
conventional raw materials ensures high and repro-
ducible parameters for commercially produced single
crystals of gallium arsenide and for epitaxial struc-
tures based on it.
5. Fedorov, V.A., Efremov, A.A., Grinberg, E.E., et al.,
Ross. Khim. Zh., 1994, vol. 38, no. 2, pp. 25 33.
6. Fedorov, V.A., Efremov, A.A., Zhukov, E.G., et. al.,
Ros. Khim. Zh., 1995, vol. 39, no. 4, pp. 46 57.
7. Fedorov, V.A., Russ. J. Inorg. Chem., 2001, vol. 46,
CONCLUSIONS
suppl. 2, pp. 87 112.
8. Krapukhin, V.V., Sokolov, I.A., and Kostyuchen-
ko, M.N., Izv. Vyssh. Uchebn. Zaved., Tsvetn. Metall.,
1967, no. 4, pp. 95 103.
9. Milovanova, M.V. and Krenev, V.A., Zh. Neorg.
Khim., 1997, vol. 32, no. 6, pp. 1699 1700.
(1) The composition of an equilibrium vapor gas
mixture was estimated for the reaction of AsCl with
3
hydrogen in the range 700 1100 K. The degree of
AsCl conversion varies from 97.3 up to 97.9% at the
3
amount of hydrogen relative to the stoichiometry
=
10. Krenev, V.A. and Evdokimov, V.I., O vosstanovlenii
trikhlorida mysh’yaka vodorodom (On the Reduction
of Arsenic Trichloride with Hydrogen), Available
from VINITI, 1972, no. 4049.
1 and approaches 100% as
exceeds 1.5.
(2) The reaction order (X = 1.86, Y = 2.78), rate
constant (lnk = 36.509 15287/T), and apparent activa-
tion energy (E = 127 kJ mol ) were determined.
1
11. Potolokov, V.N., Efremov, V.A., Zhukov, E.G., et al.,
Neorg. Mater., 2003, vol. 39, pp. 20 26.
(3) The optimal conditions of the arsenic trichlor-
ide reduction with hydrogen were determined experi-
mentally: temperature in the reaction zone 1050
1100 K, temperature in the condensation zone 670
705 K, residence time of vapor gas mixture in the
reaction zone 20 25 s, and hydrogen excess = 1.5
2.0.
12. Rusin, A.D., Agamirova, L.N., and Zhukov, E.G.,
Vestn. Mosk. Gos. Univ., Khimiya, 1982, vol. 23,
no. 2, pp. 104 107.
13. Denisov, E.T., Kinetika gomogennykh khimicheskikh
reaktsii (Kinetics of Homogeneous Chemical Reac-
tions), Moscow: Vysshaya Shkola, 1978.
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 80 No. 1 2007