Phosphine synthesis via cathodic reduction of molten white phosphorus

Article abstract of DOI:10.1007/s10789-005-0303-x

Cathodic reduction of molten phosphorus in an acid electrolyte with the formation of phosphine is studied. The results demonstrate that the process can be accelerated by sonication, which increases the current density corresponding to phosphine synthesis by a factor of 5, up to 1 A/cm². It is shown that the electrochemical synthesis of phosphine is due to charge transport across the phosphorus film on the cathode. The optimized electrochemical synthesis conditions are described.

Full text of DOI:10.1007/s10789-005-0303-x

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Inorganic Materials, Vol. 41, No. 12, 2005, pp. 1289–1293. Translated from Neorganicheskie Materialy, Vol. 41, No. 12, 2005, pp. 1463–1467.
Original Russian Text Copyright © 2005 by Vorotyntsev, Mochalov, Balabanov.
Phosphine Synthesis via Cathodic Reduction
of Molten White Phosphorus
V. M. Vorotyntsev, G. M. Mochalov, and S. S. Balabanov
Nizhni Novgorod State Technical University, ul. Minina 24, Nizhni Novgorod, 603600 Russia
e-mail: vlad@vorotyn.nnov.ru
Received April 8, 2005
Abstract—Cathodic reduction of molten phosphorus in an acid electrolyte with the formation of phosphine is
studied. The results demonstrate that the process can be accelerated by sonication, which increases the current
density corresponding to phosphine synthesis by a factor of 5, up to 1 A/cm2. It is shown that the electrochem-
ical synthesis of phosphine is due to charge transport across the phosphorus film on the cathode. The optimized
electrochemical synthesis conditions are described.
INTRODUCTION
was mounted in a thermostated ultrasonic bath (7). Dur-
ing experiments, the electrolyte temperature was main-
tained at 75–80°C. The cathode (4) was of nickel, and
its surface area was 1 cm2. When a potential was
applied, a phosphorus film appeared on the cathode
surface owing to the electrocapillary effect [3]. The
anode (6) was a Pb rod with a surface area of 1 cm2. The
anode compartment was separated from the cathode
compartment by a glass-fiber membrane. The cathode
potential was measured by a P-5848 potentiostat using
a Luggin–Haber capillary (3) and silver/silver chloride
reference electrode (1) situated in a beaker (2) filled
with a saturated aqueous potassium chloride solution.
High-purity phosphine is widely used as a starting
material in microelectronics and semiconductor tech-
nology. It is used in vapor-phase epitaxy processes for
the preparation of III–V semiconductors (gallium and
indium phosphides) and for doping Group IV semicon-
ductors.
On an industrial scale, phosphine is produced
through acid hydrolysis of aluminum or calcium phos-
phide, followed by fine purification to remove water,
sulfuric acid, hydrogen chloride, polyphosphines,
hydrocarbons, and other impurities [1].
As shown earlier [2–12], electrochemical synthesis
of phosphine is cheaper and easier to implement than
chemical processes and is commercially viable.
The phosphine concentration in the cathode gas was
determined on a Tswett-500 gas chromatograph
equipped with a vacuum sample introduction system.
Cathodic reduction of white and red phosphorus
allotropes has been the subject of intense attention [1–
28]. The current efficiency of this process in terms of
phosphine is notably higher in the case of white phos-
phorus. Moreover, the molecular structure of white
phosphorus facilitates the study of its hydrogenation.
The mechanism of the cathodic reduction of molten
white phosphorus is, however, not yet fully understood.
It is, therefore, of interest to study this process experi-
mentally and theoretically in order to elucidate its
mechanism.
Chromatographic separation was performed at 90°C
in a glass column 1.5 m in length and 3 mm in diameter,
packed with PS-1 sorbent. The flow rate of the carrier
gas (H2) was 30 ml/min. The phosphine peak time was
1 min 57 s.
The components were detected at 130°C using a
katharometer. The phosphine concentration was deter-
mined by an absolute calibration technique. As a refer-
ence, we used ultrapure (99.995 mol %) phosphine.
Calibration mixtures were prepared by an exponential
dilution technique, as described in the Tswett-500
Series Chromatographs user’s guide. The relative error
of each dilution was within 0.4%. The impurity detec-
tion limit was evaluated using the noise band width: an
analytical signal was considered detected if its height was
equal to or greater than the noise band width. The phos-
phine detection limit thus evaluated was 0.026 vol % at a
sample pressure of 101.325 kPa.
EXPERIMENTAL AND RESULTS
The kinetics of phosphine formation were followed
using potentiodynamic, potentiostatic, and galvano-
static techniques.
Experiments were performed in the apparatus
schematized in Fig. 1, which included a quartz elec-
trolyzer (8) with an emulsion of molten white phos-
phorus in 40% aqueous orthophosphoric acid as elec-
Electrolysis was conducted under constant sonica-
trolyte (5) [3]. To prepare the emulsion, the electrolyzer tion or without it. In the latter instance, sonication was
0020-1685/05/4112-1289 © 2005 Pleiades Publishing, Inc.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
1290
VOROTYNTSEV et al.
To analysis
+
–
To potentiostat
3
6
1
4
2
7
5
8
Fig. 1. Schematic of the apparatus for electrochemical synthesis of phosphine: (1) silver/silver chloride electrode, (2) beaker con-
taining a saturated aqueous KCl solution, (3) Luggin–Haber capillary, (4) cathode, (5) electrolyte, (6) anode, (7) ultrasonic bath,
(8) electrolyzer.
only used in preparing the emulsion of molten phos- that the cathode gas contained an insignificant amount
phorus in the electrolyte.
of diphosphine, a gas which always forms via consecu-
tive attachment of hydrogens to P4 molecules, suggests
that one phosphorus atom is hydrogenated predomi-
nantly, presumably the atom near which a metal atom is
oriented. Since we used a (soluble) Pb anode, and Shan-
drinov and Tomilov [13] used a Pb cathode, there are
grounds to believe that, in the initial stages of electrol-
ysis (before gas evolution), intermediate lead phos-
phides are formed, rather than P4H4.
The results obtained by the potentiodynamic tech-
nique (without sonication) are presented in Fig. 2. At
potentials from –0.1 to –0.7 V, the current density
increases, and no gas release occurs. Between –0.7 and
–0.9 V, the current density drops to 0.02 A/cm2. At
potentials from –1 to –1.4 V, it rises again, and a gas
containing 92 vol % PH3 is released.
The absence of gaseous products at the cathode in
the initial stages of electrolysis was also reported by
Shandrinov and Tomilov [13]. In their experiments, gas
evolution was detected after the quantity of electricity
passed had attained Q ꢀ 3 A h. A similar result was
obtained from the area of the peak in Fig. 2, reduced to
unit area of the cathode. Shandrinov and Tomilov [13]
pointed out that, at a phosphine solubility in an aqueous
electrolyte of 0.15 g/l, the solution would be saturated
with PH3 at Q as small as 0.2 A h. Therefore, no phos-
phine was formed in the initial stages of electrolysis.
The observed Q = 3 A h was attributed to preliminary
reduction of phosphorus and the formation of a hypo-
thetical phosphorus hydride of composition P4H4.
Indeed, our results, as well as those reported in [13],
The formation of a metal phosphide is also evi-
denced by the polarogram of phosphorus reduction at a
j, A/cm2
0.8
0.7
0.6
0.5
0.4
0.3
attest to the formation of a phosphorus compound on 0.2
the cathode. There is, however, no solid evidence that
this compound has the above composition.
0.1
As shown earlier [3, 14–17], if the electrolyte con-
0
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
tains Pb, Hg, Cd, Bi, Sb, Co, or Zn ions phosphine syn-
thesis proceeds through the formation of intermediate
phosphides of unknown composition. Moreover, the
finding, also reported by Osadchenko and Tomilov [14],
–E, V
Fig. 2. Current density as a function of the cathode potential
without sonication (potentiodynamic technique).
INORGANIC MATERIALS Vol. 41 No. 12 2005

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
PHOSPHINE SYNTHESIS VIA CATHODIC REDUCTION
1291
j, A/cm2
1.8
mercury dripping electrode in ethanolic solutions [29].
The polarogram also shows two phosphorus reduction
peaks, but the first peak is shifted by 0.9 V compared to
our results and occurs in the range –1 to –1.7 V, which
may be related to the difference in formation potential
between mercury and lead phosphides and also to the
influence of the electrolyte composition.
The data reported in [13, 29] suggest that the
assumed phosphide is partially hydrogenated: the
decrease in the amount of free phosphorus near the first
phosphorus reduction peak corresponds to a one-elec-
tron process; near the second peak, where gas release
begins at the cathode, the current efficiency in terms of
PH3 exceeds 100% (based on a three-electron process).
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0
1.5
In Fig. 2, one can see a narrow plateau around
−1.4 eV, which corresponds to the maximum rate of
phosphine release. The current density at this potential
is 0.18 A/cm2. Further increase in potential leads to a
steep rise in current density, but under these conditions
the cathode gas contains only hydrogen, with no phos-
phine.
The data in Fig. 3 demonstrate that sonication shifts
the onset of synthesis to –1.6 eV. In the range –2.2 to
−2.6V, there is a poorly defined saturation plateau, with
a current density of 0.8–1.2 A/cm2. Analysis of the
cathode gas released under these conditions showed a
phosphine content of 95–98 vol %. Further increase in
potential leads to a gradual increase in hydrogen con-
tent, which is the only component of the cathode gas at
–3.0 V.
2.0
2.5
3.0
3.5
–E, V
Fig. 3. Current density as a function of the cathode potential
under sonication (potentiodynamic technique).
The conductivity of the phosphorus melt was mea-
sured by an M4100/4 megohmmeter in a beaker, using
platinum electrodes 1.3 × 10–5 m2 in area, separated by
1.1 cm.
The measured conductivity of the phosphorus melt
used in PH3 synthesis was three orders of magnitude
higher than the value reported in the literature: 2 ×
10−2 S/m. Since the conductivity of liquid dielectrics is
determined primarily by impurity ions [31], this dis-
crepancy is attributable to the presence of impurities in
the film. The likely impurity ions in the system under
consideration are protons from the aqueous acid solu-
tion and Pb ions from the soluble anode.
The termination of phosphine formation at
increased potentials is the result of the destabilization
of the phosphorus film. With increasing current density,
the consumption of phosphorus increases. As a conse-
quence, the film becomes thinner and experiences elec-
The resistivity of the phosphorus thus determined
tric breakdown (we observed blue glow of the cathode). was used to evaluate the thickness of the cathode film,
When the cathode comes into contact with the electro- which turned out to be close to that inferred from opti-
lyte, vigorous hydrogen evolution begins, and the cal measurements. Thus, the conductivity of the phos-
hydrogen bubbles destabilize the phosphorus film on phorus film is high enough to ensure the observed cur-
the cathode.
rent density.
The above observations lead us to assume that the
The fivefold increase in current density under soni-
reduction of phosphorus takes place on the phosphorus
electrode, on the film surface, and that bringing the
electrolyte into contact with the cathode stops phos-
phine formation. As is well known, white phosphorus is
a dielectric: its electrical conductivity is 4 × 10–5 S/m
[30]. It follows from Eq. (1) that, to ensure the observed
current density at this conductivity of the film, its thick-
ness must be on the order of 10–8 m. According to opti-
cal measurements, however, film thickness is about
10−5 m:
cation can be explained as follows: Since the electrical
properties of liquid dielectrics are dominated by impu-
rity ions, charge transport at the metal–dielectric inter-
face is also mainly due to impurities. If a liquid dielec-
tric is immobile, the ions forming in the contact layer
are only removed via migration, whose velocity is lim-
ited by the low mobility of the ions. The excess charge
appearing in the electrode zone is of the same sign as
the electrode, which reduces the electrode overpotential
and, hence, the electrode current [31]. Sonication sets
the liquid dielectric in motion, sharply accelerating
charge removal from the electrode zone owing to con-
vection.
SgI
--------
l =
,
(1)
E
Using the galvanostatic technique, we measured the
voltage as a function of time at a fixed current density
from 0.95 to 1.25 A/cm2 under sonication (Fig. 4). At
where I is the current, g is the conductivity of phospho-
rus, S is the cathode area, E is the cathode potential, and
l is film thickness.
INORGANIC MATERIALS Vol. 41 No. 12 2005

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
1292
VOROTYNTSEV et al.
Current oscillations at a fixed voltage may be due to
film thickness and resistivity fluctuations. The current
in lower resistivity regions is then higher. As a result,
the film thickness in such regions decreases, and their
resistance drops further. The process develops to the
point of complete phosphorus depletion in such
regions. When the electrolyte comes into contact with
the cathode metal, the vigorous hydrogen evolution
destabilizes the film over the entire cathode surface,
leading to well-observed disruption of the phosphorus
film.
–E, V
2.3
7
6
2.2
2.1
2.0
1.9
1.8
1.7
1.6
5
4
3
CONCLUSIONS
The present results in conjunction with earlier data
indicate that phosphine synthesis takes place on the
phosphorus cathode. The experimentally determined
conductivity of phosphorus and thickness of the phos-
phorus film on the cathode are consistent with the
observed current density.
2
1
1.5
0
1
2
3
4
5
6
7
8
9
10 11
Time, min
The presence of Pb, Hg, Cd, Bi, Sb, Co, or Zn ions
in the electrolyte increases the current efficiency in
terms of PH3 by an order of magnitude, which is due to
the formation of intermediate, partially hydrogenated
metal phosphides. It seems likely that those phosphorus
atoms oriented on the metal experience predominant
hydrogenation.
Fig. 4. Cathode potential as a function of time at j = (1) 1.25,
(2) 1.20, (3) 1.15, (4) 1.10, (5) 1.05, (6) 1.00, and
(7) 0.95 A/cm (galvanostatic technique).
2
current densities from 1.05 to 1.25 A/cm2, the voltage
first increased with time and then saturated.
Sonication increases the current density correspond-
ing to phosphine synthesis by a factor of 5, up to
1 A/cm2, which may be due to a change in charge trans-
port mechanism in the phosphorus film—from migra-
tion to convection.
We believe that this behavior is due to the penetra-
tion of protons into the phosphorus film and subsequent
reduction of the protons. As pointed out by Frumkin
[32], the rate of this process depends on the energy
spent for the transfer of a hydrated ion from the solution
bulk to the adlayer (phosphorus film in our case).
Therefore, increasing the voltage must increase the
amount of reduced hydrogen, as confirmed by analyses
of the cathode gas.
REFERENCES
1. Devyatykh, G.G. and Zorin, L.D., Letuchie neorgan-
icheskie gidridy osoboi chistoty (Ultrapure Volatile Inor-
ganic Hydrides), Moscow: Nauka, 1974.
At a current density of 1.65 A/cm2, electric break-
down of the film occurred, bringing the electrolyte into
contact with the cathode surface and giving rise to dis-
integration of the film.
2. Gordon, I., US Patent 3 109 785.
3. Miller, G. and Steingart, J., US Patent 3109786.
4. Thornton, D. and Gordon, I., US Patent 3109787.
5. Miller, G. and Steingart, J., US Patent 3109788.
6. Miller, G. and Steingart, J., US Patent 3109789.
7. Miller, G., US Patent 3109790.
In contrast, at current densities below 1.00 A/cm2
the voltage first decreased with time and than stabilized
at –2.2 V. Analysis showed that the cathode gas con-
tained 98 vol % phosphine. Consequently, the electric-
field energy was insufficient for the penetration of pro-
tons into the film, and the only reaction in the system
was phosphine evolution.
8. Gordon, I., US Patent 3109791.
9. Gordon, I., US Patent 3109792.
10. Gordon, I., US Patent 3109793.
Using the potentiostatic technique, we fixed the
potential at a level of –1.6 to –2.2 V, where phosphine
evolution had been observed earlier. We, however,
revealed current oscillations with a gradual increase in
amplitude, which eventually led to disintegration of the
phosphorus film on the cathode. As a result, phosphine
synthesis ceased, and only hydrogen was released.
These findings confirm that the process takes place at
the phosphorus electrode.
11. Gordon, I., US Patent 3109794.
12. Gordon, I., US Patent 3109795.
13. Shandrinov, N.Ya. and Tomilov, A.P., Electrochemical
Reduction of Phosphorus at a Pb Cathode, Elek-
trokhimiya, 1967, vol. 3, no. 4, pp. 237–239.
14. Osadchenko, I.M. and Tomilov, A.P., Electrochemical
Synthesis of Phosphine, Zh. Prikl. Khim. (Leningrad),
1970, vol. 43, no. 6, pp. 1255–1261.
INORGANIC MATERIALS Vol. 41 No. 12 2005