I

H. N. MADONl and R. F. STRICKLAND-CONSTABLE

Department of Chemical Engineering, Imperial College of Science and Technology,
London S.W. 7, England

Production of Carbon Disulfide

A process based on the system carbon monoxide plus sulfur at 500" C.

THE traditional method of manu-

facturing carbon disulfide involves re-
tice, only reaction I V is left for further
investigation.
it is seen that carbonyl sulfide decom-
poses to carbon monoxide and sulfur
action between charcoal and sulfur at Recent years have seen concentrated on the one hand and to carbon dioxide
about 850' C.; charcoal is used because efforts to prepare carbonized briquets and carbon disulfide on the other.
the reaction of sulfur with coke is pro- from low grade coals which are highly Reaction 2 is endothermic and in the
hibitively slow at all reasonable tem- reactive to sulfur and are promising absence of 3' an increase in temperature
peratures. In England both raw ma- substitutes for natural beech chars in should therefore progressively increase
terials have to be imported. Efforts carbon disulfide manufacture. As there carbon disulfide and carbon dioxide
to produce carbon disulfide by other are very few published data on the kinet- concentrations; but above 550" C.
means may be conveniently divided into ics of the charcoal-sulfur reaction, it was carbonyl sulfide begins to decompose
two groups: decided to study the behavior of these according to Reaction 3' and this
chars under various conditions. reduces the amount of carbonyl sulfide
Chemical reactions other than the and hence carbon disulfide and carbon
straightforward C + SZ = CS2, involving dioxide which can be formed. There-
sulfur dioxide or hydrogen sulfide in- Reaction between Carbon fore yield of carbon disulfide is optimum
stead of ,elemental sulfur and carbon Monoxide and Sulfur
monoxide or methane instead of ele- in the neighborhood of 550" C., as
mental carbon. This reaction has been studied inten- illustrated by the following values of
Replacing natural beech chars by sively by Stock ( 7 7 , 72) and others equilibrium gas compositions.
artificially prepared carbonized briquets (8, 73) and shown to occur in two
prepared from cheap low grade in- consecutive steps: Equilibrium Gas Composition,
digenous coals, and using the chemical
reaction C + S:! = CS2. 2co+ sz = 2cos (1)
Temp.,
c. co SE
%
cos cot CSZ

By far the most widely used method is 2cos = cog + csz (2)
400
600
0.3
3.2
0.26
1.6
61.0
54.4
19.2
20.4
19.2
20.4
the reaction between charcoal and 700 14.8 7.4 39.8 19.0 19.0
sulfur, which gives high equilibrium The published values of Stock for the
conversions around 850' C. (Table I). equilibrium constants for Reaction 2
in the temperature range 300' to 550' C. Before the war attempts were made in
All the other reactions have been tried as Germany (7) to manufacture carbon
possible alternatives, but none has (which is of interest in the present
investigation) were found to agree within disulfide by interaction of carbon mon-
found appreciable application in industry oxide with sulfur on a pilot plant scale,
in Great Britain. The catalytic reaction limits of experimental accuracy. The
rate of decomposition of carbonyl sulfide but the tests were given up in 1942
of natural gas and sulfur (reaction V)
to carbon monoxide and sulfur is very because of shortage of manpower. No
is used in the United States, where
much faster than its rate of conversion results have been reported.
ample supplies of cheap natural gas are I t was thought that if the carbon
available. This reaction has been to carbon disulfide and carbon dioxide,
thus limiting the over-all yield of carbon monoxide-sulfur reaction were to be
studied in considerable detail by Folkins, made the basis of a carbon disulfide
Miller, and Hennig (2) and it is estimated disulfide that can be produced.
Gaseous sulfur is known to exist (3, 9 ) manufacturing process in preference to
that the process is responsible for around
as a mixture of three species: Sz, S6, the charcoal-sulfur process, a catalyst
40% of U. S. carbon disulfide.
and S S ; a t low temperatures (in the must be found to give a very fast re-
I n the case of reactions I1 and 111,
range 300' to 600" C . ) s6 and SS pre- action at low temperatures, to offset the
temperatures well above 1000' C.
dominate, whereas at high tempera- disadvantage of low equilibrium con-
would have to be attained to obtain
tures (above 700' C.) SZ predominates. version to carbon disulfide.
favorable equilibrium conversion to
carbon disulfide. This would be a dis- Therefore, strictly speaking Equation 1
advantage from the standpoint of ma- should be written as:
Kinetic Studies
terials of construction; however, opera-
tions a t these temperatures are practical 2c0 + EsE
2
= 2c0s (3) Experimental Procedure. -4 mix-
with modern refractories (Carbofrax), ture of carbon monoxide and sulfur
as shown by laboratory and pilot scale where '8 represents an vapors (in varying proportions) was
experiments (70). mixture of s?, s6, and SS. Rewriting passed over the catalyst mass contained
If in any alternative process, the Equations and as: in a 1-inch glass tube of 25-cc. capacity
maintained at the required temperature,
operating temperature must be less than
and preferably. very. muchless than 850°C.. 2COS = 2CO +2
Sz ( C O reaction)
- between 300 ' and 600 ' C. I n a second
the temperature of industrial prac- (3') set of tests carbon monoxide plus sulfur
was replaced by pure carbonyl sulfide
1 Present address, Glaxo Laboratories, and 2cos = coz -k csz reac- gas.
Ltd., Ulverston, Lancashire, England. tion) (2) After excess sulfur had been condensed

VOL. 50, NO. 8 AUGUST 1958 1 1 89

 passed over these surfaces for 1 to G

"I C O A L CHARCOAL+
-1,- 7
minutes; conversions to carbon disulfide
were very poor. These are referred to
as gas phase tests (to differentiate them
from liquid phase tests).
\
4
2.4 t
I \
\ Figure 1. Ar-
Interesting results were obtained when
rhenius equa-
fine silica powder was suspended in liquid
tion plots show
low values of
sulfur and carbon monoxide bubbled
through it. The liquid sulfur reaction
1 2,000 to 14,-
-\,. 000 cat. for
beech and coal
was maintained at 440' C.-just below
the boiling point of sulfur (445" C.).
These tests are designated as liquid phase

\
*COKE cha rcoa Is
tests.
The results obtained are expressed in
C O + S E REACTION' \ Table I1 as a space velocity, which may
\ be defined as the gaseous volume ol'
1
\\
carbon disulfide (measured at N.T.P.)
formed per hour per unit volume of the
0.6
l0OO0C , s00"c.
reactor. The only available value for
the carbon disulfide production rate
in industrial practice has also been
out, carbon monoxide, carbon disulfide, the determination of the equilibrium included, expressed as a space velocity.
carbon dioxide, and carbonyl sulfide constant for the reaction, no efforts were Under laboratory investigations, opti-
were determined. Carbon monoxide, made to investigate the reaction rates mum values of the space velocities are
the only noncondensable gas, can be systematically. As Reaction 1 is very given. An average value is given at
removed by the use of liquid air ; carbon fast, more or less independent of the each temperature.
disulfide can be separated by distillation, nature of the surface, an extensive search From the liquid phase tests a value
using a bath of solidified ethyl alcohol was undertaken to find a catalyst surface for the activation energy, E , of the
at -114' C. However, it is not easy to speed up the slower Reaction 2. reaction 2COS = C O Z + CS2 may be
to separate carbonyl sulfide from carbon obtained from the Arrhenius equation
dioxide by a similar method because of METALLICSURFACES.The choice of d In K j d T = E / R T 2 , ivhere K is the spe-
the proximity of their boiling points; catalyst was guided by reference to those cific reaction rate at temperature T and E
therefore advantage was taken of the used in a similar reaction between the activation energy expressed in
thermal instability of carbonyl sulfide methane and sulfur to produce carbon calories. K could be replaced by the
gas at high temperature ( I I ) , whereby disulfide (74). These included silver carbon monoxide flow rates for equal
sulfide, molybdenum sulfide, vanadium
it is decomposed into carbon monoxide trioxide, zinc sulfide, copper promoted (below equilibrium) values of carbon
and sulfur; the carbon monoxide can with chromium oxide, nickel promoted disulfide concentrations. -4value of E
then be removed from carbon dioxide with alumina, and alumina promoted equal to about 50,000 cal. was obtained.
by the use of liquid air. Carbonyl with manganese sulfide. The catalyst Its magnitude indicated that a small
sulfide was decomposed over a glowing masses were prepared by precipitation in increase in the operating temperature
platinum filament fused in a glass bulb; place of pumice or alumina. All should substantially increase the space
the filament was maintained a t about attempts proved unsuccessful; in some velocity. Thus on plotting the log-
1000' C. The carbonyl sulfide could be cases they appeared to act as poisons, arithm of space ve1ocit)- ts. the reciprocal
quantitatively converted to carbon mon- retarding the approach to equilibrium. of absolute temperature at 400' and
SILICA AND GLASS WOOL. I n the
oxide and sulfur within a minute. original work of Stock on equilibrium 440" C. and extrapolating to a slightly
Discussion. Previous workers had determinations, silica and glass ~ 7 0 0 were
1 higher temperature (say 500" C.) a space
used silica and glass wool to catalyze the extensively used as catalyst surfaces. velocity of 500 hr.-I is obtained (see
carbon dioxide reaction. As all such These were re-examined. A mixture of Figure l ) - i e . , a 50" C. rise should
studies were principally concerned with carbon monoxide and sulfur vapors was increase the reaction rate about fifteen-

Table I. Carbon Disulfide M a y Be Made in Several Ways

CSZPro-
duction Rate,
Lb./
Re- Temp., Hr./Cu. Ft.
action Basic Chemical Equation Equilibrium, Mole %
' Lit. Ref. Type Size, Inch O C. of Reactor
I c + sz = csz Sz = 9.4; CSz = 90.6 (1) Charcoal Lumps 850 1.5
Sz = 10.6; CSz 89.4 (4) Coal charcoal 0.07 650 18
SI = 10.6; CSz = 89.4 (4) Beechwood 0.07 650 13
charcoal
I1 C + 2HzS CSz + 2Hz HzS = 39.5; Ht 40.3 (6,6 ) Charcoal Small particles 1125 Not detd.
CSI = 17.8; Sz = 2.4
I11 5C + 2SOz = CSz + 4CO so2 = 0; sz = 2.2 (10) Anthracite 0.05 (approx.) 1250 1.27
CSz = 16.5; COS 3.3; CO =
78
IV 2co + Sa = 2 c o s = coz + csz CO = 0.4; COz = 19.5
Sa 0.3; CSz = 19.5; COS
(119 12) Bubbling CO through liquid
sulfur containing suspension
440 7"

= 60.3 of silica dust

V CHI + 2Sz = CSz + 2H2S CH4
Sz
=0 ; CSz = 33.3
0 ; HzS 67.7
(14) ... ... 700 57

a Present investigations.

1 190 INDUSTRIAL AND ENGINEERING CHEMISTRY

 CARBON DISULFIDE PRODUCTION

fold. This would mean operation under

pressure (boiling point of sulfur, 500' C.
a t 35 p.s.i.g.).
Space velocities of this high order
might be required to offset the very b
large amounts of unconverted sulfur
as carbonyl sulfide in the product gases. Figure 2. Kinetic
A practical method must be found for studies show the
removing carbon dioxide from carbonyl reaction is of a
sulfide and recycling the purified car- fractional order
bonyl sulfide. This is purely an eco-
nomic consideration and may be solved
only by large scale trials and comparing
the costs of the two processes-i.e., carbon
monoxide plus sulfur a t 500' C. us.
carbon plus sulfur a t 850' C.

Reaction between
Charcoals and Sulfur
In industrial practice 1- to 2-inchlumps
of charcoal react with sulfur; however,
P A R T I A L PRESSURE O F SULFUR VAPORS, ATMOSPHERE
for investigations on a laboratory scale
smaller particles, about */8 inch, had to through the sulfur boiler and a mixture TEMPERATURE VARIATION. Prelimi-
be used. This meant that some of the of nitrogen and sulfur vapors was passed nary tests had brought out two charac-
results obtained in the laboratory may over charcoal. At the exit end excess teristics of the charcoal-sulfur reaction :
not be applicable to plant practice, sulfur was condensed cut in a bath Below about 625' C. there was no per-
because the size of the charcoal particles maintained a t 140' C. and the crude ceptible formation of carbon disulfide;
might have a considerable influence on carbon disulfide was condensed in a if the charcoals were heated appreciably
the rate of formation of carbon disulfide. solid carbon dioxide-methanol bath. above 550" C. in absence of sulfur vapors
The reaction rate of a solid-gas reac- The system was always operated under a but in a n inert atmosphere of nitrogen,
tion like the one under consideration is pressure of 1 atm. the charcoal surface underwent partial
governed by two factors: Charcoals Used in Tests. Tests were deactivation. Thus a charcoal sample
Chemical resistance, determined by carried out on the following samples : treated in this way gave a space velocity
the function exp ( E / R T ) , where E is Natural beechwood charcoal supplied only about half to a third of that in an
the activation energy and T is the as 1-inch lumps. Analysis: Moisture = atmosphere of sulfur above 550' C.
absolute temperature, is greatly affected 6%, ash = 1.5%, and volatiles = 13%, No explanations for this effect can be
by temperature; in the case of the com- Coal charcoal prepared from low given. No such effect was noticed in
bustion of carbon, a 15O rise in temper- grade coals by briquetting and sub- the case of coke.
ature approximately doubles the com- sequent carbonization a t 500" C. in the
bustion rate but is independent of the Experiments with coke showed that it
laboratory of the British Coal Utilization
size of the solid particles. Research Association a t Leatherhead. was reactive only above 1000' C. and
Diffusional resistance, only slightly Analysis: Moisture = 4.3%, ash = even a t that temperature did not compare
affected by temperature but greatly 3.4%, and volatiles = 13.5%. The favorably with the charcoals. I t would
affected by the diameter of the particle sample was supplied as cubes about '/8 be interesting to obtain the reaction rates
and the gas velocity. Diffusional resist- inch in size. with pulverized coke; the reaction be-
ance, in the case of a coal, is governed Coke lumps from boiler house. tween coke and sulfur vapors might
by the nature of the volatiles contained well be predominantly diffusiqn con-
in it. This has to be remembered in Variables Studied. The rates of
comparing reaction rates of different reaction, expressed as carbon disulfide trolled and therefore very much de-
varieties of charcoal. space velocities, were determined under pendent on the particle size. Limita-
the following conditions (Table 111) : tions of the apparatus make such studies
Depending on the temperature, one or impossible.
the other of the two resistances would 1. At varying temperatures but fixed I n the tests given in Table 111 the
predominate. For the combustion of partial pressure of sulfur vapor
2. At varying partial pressures of charcoals were always kept in contact
carbon such a temperature is about sulfur vapor but fixed temperature. with sulfur above 550' C. The log-
1050' K. (7, 75). 3. At varying total flow rates but arithms of the carbon disulfide space
I t is not possible to translate the fixed partial pressure of sulfur vapor and velocities are plotted in Figure 1 against
reactivity results obtained on a laboratory fixed temperature. the reciprocal of the absolute tempera-
scale to a full scale plant. However,
they are useful for comparing different
types of coal and showing trends. Table [I. Comparison of Space Velocities
Liquid phase tests gave high space velocities
Experimental Procedure Laboratory Investigations
.~
About 25 cc. of charcoal as l/s-inch Metallic */ls-Inoh Silica Beads
pieces was contained in a "4 inch silica Industrial surfaces,
Practice Cu, Ni, etc., Silicic Acid (3 : 1),
tube which was inserted horizontally in a (1) on p-lce Gas Phase Tests Liquid Phase Tests
Gallenkamp laboratory furnace. Liquid Temp., a. C. a50 500 450 550 400 440
sulfur was contained in a glass ,tube Space velocity,
2 inches in diameter carrying an exter- hr.-l CSZ 7 2 3 3.3 5.6 34
nal winding. Nitrogen was bubbled ~~

VOL. 50, NO. 8 AUGUST 1958 1 19 1

 Table 111. Coal Charcoal Was as Reactive as Beech Charcoal, Coke Was Relatively Unreactive
Set 1 Set 2 Set 3
Constant Variable Constant Variable
mean
partial Space
mean
partial Space
Mean
partial
total cas
flow at
- Space
Variable, pressure velocity, Constant, pressure velocity, pressure outlet, velority,
To,C. sulfur, atm. hr.-1 CS2 T o , C. sulfur, atin. br,-l CS2 To,C. sulfur, atin. cc./rnin. lrr-1 (:S2
Coal charcoal 675 0.813 200 625 0.107 35 675 0.882 1366 195
750 0.770 341 625 0.212 8 9 675 0.923 2110 178
825 0.734 437 625 0.386 70 675 0.921 2738 198
625 0.662 97
625 0.935 12 1
Beech charcoal 675 0.743 211 625 0.084 24 675 0.877 1456 219
750 0.622 444 625 0.187 37 675 0,925 2201 184
825 0.590 592 625 0.541 81 675 0.944 28b8 118
625 0.831 89 675 0.945 2749 109
625 0.954 85
Coke 825 0.991 ... ... .. ... ... ... ... ....
950 0.986 9 ... .. ... ... ... ... ...
1050 0.966 44 .. ... ... ... ... ...
1150 0.956 62 .. ... ... ... ... ...
e Stray result.

tures. The slopes of straight lines with a slope of about 0 6 (Figure 2 ) ; ever: it might be possibie to increase
drawn as nearly as possible through these the reaction is therefore of a fractional the reactivity of coke (or cheap loiv
points correspond, according to the order. grade coals) by employing it in pulver-
Arrhenius equation, to activation ener- VARYIUGFLOWRATESO F SLLFUR. ized form.
gies of 12,000 and 14,000 cal., respec- The reaction rate Mould not be expected The reaction C +
S2 = CS2 \vas
tively, for the beech and coal charcoals. to be affected by mass flow rates, if the found to be of a fractional order and its
I t is necessary to consider whether the partial pressure of sulfur was constant energy of activation was determined as
observed rates of reaction correspond This expectation was fulfilled for the coal about 13>000 cal. 4 possible explana-
to true surface reaction rates, or whether charcoal but not for the beech charcoal; tion for this rather low value is suggested.
the over-all rates are partly determined in the latter case there was a progressive
by rates of diffusion. I t seems unlikely falling off in the space velocity with
Acknowledgment
that at the flow rates and reaction rates increased flow rates, although the sulfur
used in these experiments diffusion in concentration was never less than 0 88 The authors are indebted to Sir
the gas phase could affect the over-all atm. Because of this rather unexpected Alfred Egerton for many helpful sugges-
rate. O n the other hand the charcoals result the experiments were repeated tions and to D. M. Newitt for his interest
are highly porous, and diffusion in the under identical conditions; results were in the work. especially concerning the
pores may be partly controlling and may reproducible to within 6%. No ex- reactivities of charcoal.
affect the over-all rate. As the diffu- planation has so far offered itself.
sional retardation increases as the rate Literature Cited
of reaction increases, such an effect Conclusion
would reduce the apparent activation ( I ) FIAT Final Rept. 723 (1946).
energies; this may account for the Various methods of manufacturing (2) Folkins, H. O., Miller, E., Hennig. H.,
IND. ENG.CHEM. 42. 2202 (1950'1.
rather specially low values of 12,000 carbon disulfide have been reviewed (3) Kelley, K. K., U. S. B'ur. Mines, B d l .
to 14,000 cal. quoted above. If this and compared with the traditional 406 119371.
is the case, the true activation energies method involving the use of beechwood (4) , J . Appl. C h m .
may be much higher. charcoal. The reaction between carbon (London)
VARYING PARTIAL PRESSURES O F SUL- monoxide and sulfur was singled out for (5)
FUR. The rate of production of carbon investigation. Rates of reaction were 419 (195
disulfide increased with the partial high when carbon monoxide was bubbled Owen,
- _ A. _ J.,
..
Syke3, K. W., Thomas.
pressure of sulfur vapor, as expected; through liquid sulfur at its boiling point U. J . U., Mhite,. P.,. Ibid., 49, 1198
(1953).
reaction rate increased three- to four- containing a suspension of fine silica (5) Parker, A. S.: Hottel, H. C., 1x1).
fold for a tenfold increase in the partial powder. An approximate value for the ENC. CHEM. 28, 1334 (1936).
pressure of sulfur. I n the case of activation energy, E , equal to 50.000 (8) Partington, J. R., Neville, H. H.,
beech charcoal, but not coal charcoal, cal. was obtained. This suggests that an J . Chem. SOG.1951, p. 1230.
(9) Preuner, G., Schupp, W., Z . phyrik.
the reaction tended to become a zero- increase in operating temperature (pres- Chem. (Lezpri?) 68, 129 (1909).
order reaction as the partial pressure of sure reaction) should substantially in- (IO) Siller, C. W., I N D .ENG. CHEM. 40,
sulfur vapor approached unity. In crease the reaction rate. This might 1227 (1948).
industrial practice economic considera- perhaps offset the disadvantage, for (11) Stock, .4,, Seelig, P., Ber. 5 2 , 681
f1919).
tions would tend to limit the sulfur industrial application, of low carbon Stock, A,,Siecke, W., Pohland, E.,
throughput, so that sulfur partial pres- disulfide equilibrium yield. Ibid.,57, 719 (1924).
sures approaching unity are seldom ob- The behavior of wood charcoal, Terres, E., Wesemann, H., Angezu.
tained and are more likely to be in the artificially prepared reactive charcoal, Chem. 45, 795 (1932).
Thacker, C. M.,Miller, E., IND.
region 0.2 to 0.4 atm. (total pressure and coke with sulfur was studied under ENG.CHEM.36, 182 (1944).
unity). various experimental conditions. Coal Tir, C. M,, Davis, H., Hottel, H. C.,
A plot of the logarithm of space charcoalwas as reactive as a natural beech Ibid., 26,749 (1934).
velocity us. logarithm of the partial charcoal. Coke was relatively very RECEIVED
for review August 2 , 1957
pressure of sulfur gives a straight line unreactive, even a t 1000° C. How- ACCEPTED March 1, 1958

1 192 INDUSTRIAL AND ENGINEERING CHEMISTRY