Académique Documents
Professionnel Documents
Culture Documents
Article
Mechanisms of CO and COS Formation in the Claus Furnace
Peter D. Clark, Norman I. Dowling, M. Huang, William Y. Svrcek, and Wayne D. Monnery
Ind. Eng. Chem. Res., 2001, 40 (2), 497-508• DOI: 10.1021/ie990871l • Publication Date (Web): 21 December 2000
Downloaded from http://pubs.acs.org on April 28, 2009
Additional resources and features associated with this article are available within the HTML version:
• Supporting Information
• Access to high resolution figures
• Links to articles and content related to this article
• Copyright permission to reproduce figures and/or text from this article
APPLIED CHEMISTRY
The aim of this study was to determine the major pathways leading to COS and CO formation
and consumption during the processing of H2S and CO2 in the partially oxidizing conditions of
the Claus furnace. Both species were found to be produced by a multitude of pathways, which
include the direct reaction of H2S with CO2 to form COS and H2O and the reaction of CO2 with
S2, one of the major primary products in a Claus furnace. This last reaction produced SO2 and
CO as the major products, with COS being formed in lesser quantities. The dissociation of H2S
to H2 and S2 at high temperatures (>1000 °C) was shown to promote a further cascade of reactions
stemming from the reduction of COS and CO2, both of which lead to CO. Because of the known
formation of CS2 from hydrocarbon carry-over into the furnace, the reactions of CS2 with CO2,
H2O, and SO2 were also studied as potential CO- and COS-forming reactions. Reaction with
CO2 was slow at <1200 °C, but reaction with either H2O or SO2 was fast above 900 °C. Conversion
of CS2 by H2O led to CO, H2, H2S, S2, and CO2, whereas reaction with SO2 resulted in CO2 and
S2 as the major products. Similar observations were made for the reactions of COS with H2O
and SO2. The summary of pathways presented in Scheme 5 shows a complex interlinkage among
many reactions involving H2S, CO2, CO, COS, SO2, and S2, leading to the conclusion that previous
explanations of the production of COS from CO + S2 and CO from incomplete combustion
dramatically oversimplifies the formation/consumption for these compounds. It also shows that
modeling of individual kinetic rate expressions is somewhat impractical.
to account for field COS trends, indicating that other It is well-known that hydrocarbons (methane, aro-
mechanisms must be involved in COS formation. Behie matics) contaminate the acid gas fed to a Claus furnace
et al.11,12 comment that no systematic study has been because of their solubility in the amines used to separate
published on mechanisms involving H2, CO, COS, and the acid gases from hydrocarbon mixtures. Earlier work
CS2 formation/consumption in the Claus furnace. Con- done in our laboratories20,21 has resolved the question
sequently, we report the results of our research on this of CS2 formation in the Claus furnace, as we have shown
topic, which shows that COS can be formed by several that it is the hydrocarbon impurities that react with S2,
pathways, requiring that an intrinsic kinetic model for or other sulfur species, to form the CS2. An extensive
COS formation and destruction in a commercial furnace series of experiments20 showed that CO2 does not react
include much more than simply the reaction of CO with with H2S to give CS2; only COS is produced in this
S2. reaction.21 However, CS2 concentrations in the gases
Potential Pathways to COS. The Claus furnace can leaving the reaction furnace are known to rapidly
be divided into two distinct chemical regions: a partially increase and then decrease depending on the temper-
oxidizing flame zone with the highest concentrations of ature and residence time in that unit. Because COS
O2 near the burner tip and temperatures as high as could be formed by decomposition of CS2 under reaction
2500 °C and an anoxic region with temperatures in the furnace conditions, it is important to consider such
range of 900-1200 °C depending on the H2S content of reactions in the modeling of COS. Possible CS2 conver-
the acid gas and whether O2 enrichment is being sion reactions are as follows:
applied. In the case of a mixed acid gas (H2S and CO2),
a series of reactions can ensue from the thermal CS2 + H2O H COS + H2S (7)
decomposition of H2S and the reaction of H2S with CO2
in the anoxic region. CS2 + CO2 H 2COS (8)
Table 1. Equilibrium Constants for Possible Overall Reactions within the Anoxic Region of a Claus Furnace
Kp
reaction 900 °C 1000 °C 1100 °C 1200 °C
2H2S H 2H2 + S2 (1) 1.25 × 10-3 5.36 × 10-3 1.85 × 10-2 5.33 × 10-2
H2S + CO2 H H2O + COS (2) 4.59 × 10-2 6.18 × 10-2 7.98 × 10-2 9.98 × 10-2
H2 + COS H CO + H2S (3) 30.1 29.6 29.0 28.4
2COS H 2CO + S2 (4) 1.14 4.69 15.5 43.1
CO + H2O H CO2 + H2 (5) 0.72 0.55 0.43 0.35
2CO2 + S2 H COS + CO + SO2 (6) 1.57 × 10-4 3.95 × 10-4 8.84 × 10-4 1.81 × 10-3
CS2 + H2O H COS + H2S (7) 109 83.3 66.8 55.4
CS2 + CO2 H 2COS (8) 4.98 5.15 5.33 5.53
2CS2 + SO2 H 2COS + 3/2 S2 (9 + 10) 1.68 × 105 1.45 × 105 1.26 × 105 1.11 × 105
CS2 + 4H2 H CH4 + 2H2S (11) 2.70 0.34 5.74 × 10-2 1.28 × 10-2
COS + H2O H CO2 + H2S (12) 21.8 16.2 12.5 10.0
2COS + SO2 H 2CO2 + 3/2 S2 (13 + 14) 6.78 × 103 5.48 × 103 4.45 × 103 3.63 × 103
2CO + SO2 H 2CO2 + 1/2 S2 (15 + 16 + 17) 5.96 × 103 1.17 × 103 2.87 × 102 84.2
that the cooler anoxic zone of the furnace makes up the partial oxidation and steam reforming of methane or
largest region in the furnace. Thus, molecular reactions other hydrocarbons entering the furnace. The objective
might have a dominating effect in establishing the final of this work was to determine the dominant mechanistic
product composition, as radical intermediates transform pathways to COS and CO under Claus furnace condi-
to molecular species on departure from the flame zone tions to allow for a logical approach to the kinetic
and react to establish the equilibrium composition at modeling of these species.
the cooler conditions of the anoxic zone.
Because a large number of reactions can potentially Experimental Section
lead to or assist in the formation/destruction of COS and
CO within the Claus furnace, equilibrium constants for Gases and Materials. All gases were obtained from
these reactions were calculated in order to determine standard suppliers and had >99.9% purity according to
the most likely pathways. These values are given in the gas chromatographic procedures described below.
Table 1 for reactions 1-17, in some cases as overall Water was distilled before use.
reactions. Although this type of analysis can be used to Experimental System. A schematic of the system
examine the relative likelihood of different reactions to used in the experiments is given in Figure 1. Two types
act as a given pathway, it should be remembered that of reactor were used: one constructed of fully dense
it does not provide any information about the relative alumina (supplier, Vesuvius McDanel) with dimensions
rates of these processes. Nonetheless, the range of Kp of 19 mm o.d × 12.7 mm i.d. × 74 cm length and the
values listed, for the reactions as written, are suf- other fabricated from quartz with the same dimensions.
ficiently large that all of the listed reactions should be The purpose of using two reactor types was to establish
considered. The large values of the equilibrium con- the catalytic nature of the surface on some of the
stants for the reactions of CS2, COS, and CO with SO2, reactions suggested as possible COS formation path-
in particular, suggest that these reactions might rep- ways. Temperature profiles for these externally heated
resent major pathways for conversion of these species reactors were measured under a nonreactive flow of
in the Claus furnace, barring any kinetic limitations. argon using an alumina sheathed S-type thermocouple.
The smaller Kp values for reactions 1, 2, and 6 also do These profiles presented an initial ramp-up section,
not exclude these reactions from the anoxic chemistry, (∼400 °C to reaction temperature) corresponding to 20%
as these processes represent the starting points or of the length of the reactor, followed by an isothermal
primary reactions for this chemistry. hot-zone section, and finally, a cool-down section. Noted
Clearly, the number of reactions that can lead to COS residence times for the experiments given in the data
and CO could be quite large and, for CO, would include tables refer to the hot-zone section of the furnace and
500 Ind. Eng. Chem. Res., Vol. 40, No. 2, 2001
Table 2. Effect of Reactor Type on Product Yield for the Reaction of a 50% H2S/50% CO2 Feed Mixture
equilibrium, quartz, and R- alumina product yields(mol of product per 100 mol of reactant)
T °C tres (ms) H2 CO COS SO2 S2 H2O
600 3200 0.6 0.18 0.14 1.8 0.11 0.63 4.1 0 3.2 0 0 0 1.3 0.25 0.5 6.0 0.32 4.1
700 2870 1.3 0.22 1.1 4.2 0.12 1.5 4.2 0.2 5.8 0 0 0 2.7 0.25 1.0 8.5 0.48 6.7
800 2600 2.4 2.4 3.3 7.9 0.2 6.0 3.8 0.4 5.3 0 0 0 5.0 1.4 3.4 11.7 0.8 8.7
900 2400 4.0 5.9 4.3 12.3 1.3 8.6 3.1 0.7 5.5 0 0 0 8.0 4.4 5.2 15.4 3.3 11.5
1000 2230 5.8 7.6 5.5 16.9 3.7 11.5 2.5 1.0 5.4 0.09 0 0 11.1 6.2 8.6 19.2 5.8 17.0
1100 2070 7.5 6.7 6.7 21.0 14.7 14.8 1.9 3.3 6.1 0.2 0 0 13.9 11.2 10.7 22.6 19.0 20.8
1200 1930 9.3 7.4 6.74 24.9 20.5 16.9 1.4 3.7 7.1 0.4 0 0 16.3 13.9 11.4 25.4 24.1 23.1
1300 1800 10.9 7.9 6.0 28.1 23.1 17.9 1.1 3.4 8.0 0.7 0 0 18.3 15.6 12.1 27.7 26.7 26.2
1400 1700 12.2 8.2 6.3 30.8 25.6 20.2 0.9 3.3 8.0 1.3 0 0 19.5 17.0 13.2 29.1 29.2 28.1
can therefore be used for relative comparisons between the system comprising reactants and all potential
different reactions. products. The list of species included consisted of N2,
Although some reaction is unavoidable within the COS, CS2, CO, CO2, SO2, H2O, H2S, H2, S2, S3, S4, S5,
initial ramp-up zone, use of a reactor design to imple- S6, S7, and S8. Thermodynamic data in the form of the
ment a separate reactant entry and preheat was con- Gibbs energy function, [G° - H°(298.15)]/T, and the heat
sidered but not implemented. While this approach of formation at 298.15 K, ∆H°f, for these species were
appears to have merit, it is limited by the ability of obtained from the JANAF Thermochemical Tables,24
several of the reactants (notably H2S and COS) to and the program was constructed using the approach
undergo unimolecular decomposition. Subsequent reac- described by Holub and Vanka25 for the White-Johnson-
tion within the cool-down section of the reactor was Dantzig method.
minimized by employing a quench system with a quench Equilibrium constants, Kp, for the reactions listed in
time on the order of 20-100 ms from reaction temper- Table 1 were also determined from the results of these
ature to ca. 400 °C. Quenching of the gases was achieved free energy minimization calculations.
by withdrawal of gas samples through a quartz capillary
connected to a pumping system. Reactants were me- Results and Discussion
tered using mass flow controllers with reactant compo-
sitions being checked by gas chromatography (see The H2S/CO2 Reaction. The basic data for reaction
below). In experiments using sulfur as a reactant, the of H2S and CO2 in both quartz and fully dense alumina
reactant gas mixture was passed through liquid sulfur reactors are given in Table 2, along with the equilibrium
at 330 °C with the quantity being taken up ascertained values predicted by free energy calculations. It is clear
by cooling the hot gas to 20 °C to condense and trap that H2S does react with CO2 under Claus furnace
the sulfur carried by the gas. conditions, although surface effects might have a con-
Experimental Procedures. A typical experiment siderable effect on many of the reactions involved in this
consisted of stabilizing the reactor system at the desired system. In the quartz reactor, formation of COS was
temperature, passing the feed mixture through the very slow below 1000 °C such that equilibrium conver-
reactor, and analyzing the dried (P2O5) product gas. sions were not obtained (Figure 2). However, above 1050
Several samples were collected for each condition, with °C, COS formation appeared to exceed equilibrium,
average data being presented in the tables (absolute suggesting, perhaps, that decomposition of COS to CO
error is estimated to be <2% of the quoted values). and S2 is a rate-limited process. Thus, immediately it
Gas Chromatographic Methods. To avoid any is seen that the final product distribution from reaction
unwanted reactions due to the presence of condensed of H2S with CO2 depends not just on the initial reaction
water during sampling of the product gases, a P2O5 drier but on consecutive reactions of the initial products. The
was always placed in front of the syringe. The temper- apparent levels of COS above those predicted by equi-
ature of the sampling position was also maintained at librium at 1050 °C and beyond could also be explained
approximately 150 °C. In this way, both water and by recombination of CO and sulfur during quenching
sulfur were removed from the gas stream prior to of the products from these higher temperatures. As will
collection of the sample, while all other components be seen later, however, the reaction of S2, produced by
were collected and analyzed by dual gas chromato- H2S dissociation, with CO2 (reaction 6) could also
graphs. Analysis of H2, CO, and N2 was performed on a contribute to COS formation above predicted levels,
Haysep A column (Ar carrier, 15 mL/min), while all of provided that COS decomposition was slow. Interest-
the other components were analyzed on a Chromosorb ingly, in the alumina reactor, COS formation appeared
108 column with He as carrier (28 mL/min). to exceed equilibrium predictions at all temperatures
In all of the systems examined, N2 or argon was above 600 °C, indicating that the reaction of CO2 with
always maintained as the balance component in the feed H2S (reaction 1) is catalyzed quite efficiently at an
composition. This allowed these inert species to be used R-alumina surface and again suggesting that COS
as an internal standard in the GC analysis. In this way, decomposition or indeed, other COS-consuming reac-
the product yield of each component after reaction in tions are rate-limited.
units of moles per 100 moles of reactant was obtained As discussed as part of the Introduction, the inde-
from the GC data. Because water and sulfur were pendent dissociation of H2S and the subsequent reaction
removed prior to collection of the sample, the yields of of H2 with other species could have a significant effect
these components were calculated from H and S mass on the overall product composition in the H2S/CO2
balance equations, respectively. system. Because H2 formation is close to the predicted
Free Energy Calculations. Equilibrium composi- equilibrium value in both quartz and alumina reactors
tions and yields for the various reactions examined were at temperatures up to 1100 °C, dissociation of H2S must
calculated using Gibbs free energy minimization23 for either be catalyzed equally by the two reactor surfaces
Ind. Eng. Chem. Res., Vol. 40, No. 2, 2001 501
Figure 4. Comparison of equilibrium and experimental results Figure 5. Comparison of equilibrium and experimental conver-
for the CO2 + S2 reaction (tr ) 1 s). sions for COS decomposition (tr ) 0.5 s).
is thermodynamically favored even at these lower Scheme 1. Pathways for Reaction of CO2 with S2
temperatures (see also Figure 5). The comparison of the
experimental and equilibrium conversions for this sys-
tem in Figure 4 further shows good agreement only at
the high-temperature extreme of this plot, attesting to
the slow rate of reaction at lower temperatures.
Interestingly, the reaction of CO2 with either S or S2
gives another pathway for the shuttling of oxygen from
CO2 to sulfur, forming both COS and SO2, species, which
can then undergo further chemistry leading back to
sulfur and H2O. The actual mechanism of reaction 6 is
open to debate, as at least two potential pathways can
be written (Scheme 1, A and B). Both pathways involve as work done on the combustion of sulfur compounds
unstable sulfur oxides (SO and S2O), which cannot be has suggested that this species would be a less likely
observed by standard analytical methods, and each intermediate at high temperatures compared to SO.26
correctly predicts the reaction products. It is likely, The data in Table 3 also show that the COS postulated
however, that the scheme involving S2O is less probable, to be formed must undergo dissociation to S2 and CO
Ind. Eng. Chem. Res., Vol. 40, No. 2, 2001 503
(reaction 4), as COS is found in lower, and CO in higher, The absence of any significant amount of COS also
molar quantities than predicted by eq 6. In addition, suggests that the reaction might proceed by the follow-
the two pathways in Scheme 1 indicate that the CO2/S2 ing sequence:
system is reversible by reaction of SO2 with either CO
or COS. These reactions were tested by passing these SO2 + CO f SO + CO2 (15)
reagents through the fully dense alumina reactor at
>1000 °C. SO + CO f CO2 + [S] (16)
The CO/SO2 Reaction. Although this reaction has
been examined at low temperatures (<600 °C) over [S] + [S] f S2 (17)
various catalysts in connection with SO2 removal from
flue gases,27-29 only limited work has been done at This would further imply that COS might not form from
temperatures over 1000 °C. Most recently, Russian the reaction of CO with [S], if, indeed, reaction 16 is an
workers have published the results of static experiments important pathway and might only form via reaction
in quartz at 827-1077 °C, finding sulfur, CO2, and COS with S2 (the reverse of eq 4) when CO is present in large
to be formed as the major products.30 Our experiments, quantities.
carried out in a flowing regime (Table 4), yielded sulfur The COS/SO2 Reaction. The feed ratios chosen for
and CO2 as the major products, with COS being formed these experiments were based on the rationale that
only at high CO/SO2 feed ratios and only in small there is always much more SO2 in the Claus furnace
amounts relative to the other products. This type of than COS. In particular, we wished to determine
product distribution was also found from the results of whether the relative kinetics of COS decomposition and
equilibrium calculations carried out on this system and its reaction in the presence of excess SO2 would permit
also reported in Table 4. This too showed that COS was decomposition to also occur. The data obtained from
expected to be formed only at higher CO/SO2 feed ratios, these experiments (Table 5) show that the reaction with
and even then only in relatively small amounts. This SO2 could be a major pathway for the destruction of COS
result can be taken as an indication of the thermody- in the furnace, as quantitative conversion was found at
namic preference for the 2CO + S2 H 2COS equilibrium 1055 °C for residence times of >1 s, similar to those
to sit largely on the side of CO and S2 at the reaction found in commercial furnaces. The major products of
temperatures studied. the reaction were CO2 and sulfur, with only around 25%
504 Ind. Eng. Chem. Res., Vol. 40, No. 2, 2001
Table 6. Reaction Products from the Reaction of CS2 O2 incorporated for combustion of the hydrocarbons,
with CO2 allowing for formation of CS2 by reaction of the uncom-
feed mole products busted hydrocarbon with S2 or other sulfur species. As
T tr ratio (mol per 100 mol of reactant) % CS2 pointed out in the Introduction, the possible decomposi-
°C (s) CO2 CS2 CO2 CO COS CS2 S2 conversion tion reactions of CS2 are of special interest with respect
effect of temperaturea to COS formation in the Claus furnace and WHB, as
1055 2.0 5.75 1.9 5.63 0.47 0.04 1.37 0.5 27.9 each can lead to CO and COS production. As in some of
1155 2.0 5.75 1.9 5.33 0.90 0.07 1.23 0.6 31.5 the previous work, we have examined the reaction of
1255 2.0 5.75 1.9 4.36 2.14 0.13 0.73 1.1 61.3 CS2 with an excess of the reagent in order to ap-
effect of residence time proximate conditions in Claus furnaces and so, perhaps,
1055 0.5 6.22 1.6 6.06 0.09 0.13 1.48 0.1 7.5 deduce the more important mechanisms of a commercial
1055 1.0 6.78 1.9 6.50 0.39 0.25 1.58 0.3 18.1 system.
1055 1.5 5.76 2 5.57 0.62 0.13 1.55 0.4 21.3 Interestingly, CO2 was found to react with CS2 under
effect of CO2/CS2 feed ratio Claus furnace conditions although relatively more slowly
1055 2.0 20 1.9 19.6 0.91 0.05 1.33 0.5 29.8 (Table 6) when compared to some of the other reactions
1055 2.0 11.1 2 10.7 0.80 0.07 1.52 0.5 24.6 included in this study. Thus, equilibrium calculations
a Equilibrium calculations predict quantitative conversion of predict quantitative conversion of CS2 under the condi-
CS2 under all conditions investigated. tions studied, because pf the excess of CO2 in the feed.
Nevertheless, over 60% CS2 conversion was observed at
of the COS appearing as CO in the experiments showing 1255 °C using a 2-s residence time, giving CO and sulfur
complete COS conversion. This suggests that reaction as major products. Because it is difficult to envisage how
of COS with SO2 is relatively facile and is in reasonable CO2 might be converted directly to CO, it is probable
agreement with equilibrium predictions for this system. that COS is an intermediate product of this reaction
When the data with variable residence time obtained that then undergoes further decomposition to CO in a
at 1055 °C (Table 5) are compared to the model for COS combination of reactions 8 and 4.
kinetics published by Behie et al.,11 it is clear that
reaction of COS with sulfur dioxide is considerably CO2 + CS2 H [2COS] H 2CO + S2
faster than COS decomposition. This conclusion is
reinforced by the data shown in Table 10 for COS The observation of small concentrations of COS in all
decomposition in the reactor system of this study. Thus, experiments and more COS than CO at a 0.5-s residence
overall, the following reaction sequence, which is reac- time at 1055 °C supports the formation of COS as an
tions 13 and 14 given in the Introduction, can be written initial product. The last set of data in Table 6 shows
to describe the major pathway in the SO2/COS system: that the conversion of CS2 by CO2 appears to be
independent of the CO2 concentration (within this
COS + SO2 f CO2 + SO + 1/2S2 (13) limited data set), although a more extensive study of
this reaction would be required to confirm this tentative
COS + SO f CO2 + S2 (14) conclusion.
Most workers have assumed that conversion of CS2
In addition to these processes, direct decomposition of in the Claus furnace occurs by reaction with H2O, which
COS (eq 4) seems to account for approximately 25% of is present in large concentrations. The data in Table 7
the total COS conversion. bear out this supposition, as 100% conversion was
In summary, it can be concluded that the CO2/S2 observed over 2 s even at the relatively low temperature
reaction is an important reaction in the Claus furnace, of 1055 °C. In addition, conversions >90% were obtained
particularly above 1200 °C (i.e., in the flame zone), and over 0.5 s at 1055 °C. Both CO2 and CO were formed as
that it is reversible by reaction of either CO or COS with major products, along with lesser amounts of COS.
SO2. Interestingly, CO2 was the major oxide product at 1055
Reactions of CS2 with H2O, SO2, and CO2. We °C, whereas CO was the predominant oxide at 1255 °C.
have reported data on CS2 formation in earlier publica- These observations, together with the increased CO2/
tions20,21 and shown that it arises whenever hydrocar- CO ratios at higher H2O/CS2 feed ratios, suggest that
bons are present in the feed acid gas. Our laboratory COS is an intermediate product that can undergo either
data indicated that H2S consumed much or all of the decomposition to CO and sulfur or hydrolysis to CO2
Table 7. Reaction Products from the Reaction of CS2 with H2O
feed mole
ratio products (mol per 100 mol of reactant) % CS2
T °C tr (s) H2O CS2 H2S CO2 H2 CO COS SO2 CS2 S2 H2O conversion
effect of temperaturea
1055 2.0 5 2 1.53 1.09 1.71 0.80 0.11 0 0 1.2 1.8 100
1155 2.0 5 2 1.49 0.90 2.00 0.95 0.15 0 0 1.2 1.5 100
1255 2.0 5 2 1.55 0.74 1.88 1.09 0.17 0 0 1.1 1.6 100
effect of residence time
1055 0.5 5 2 1.78 0.95 1.86 0.82 0.08 0 0.15 0.9 1.4 92.5
1055 1.0 5 2 1.81 1.00 1.96 0.81 0.10 0 0.10 1.0 1.3 95.0
1055 1.5 5 2 1.61 0.96 1.85 0.88 0.12 0 0.05 1.1 1.5 97.5
effect of H2O/CS2 feed ratio
1055 2.0 20 2 2.29 1.49 3.22 0.48 0.03 0.16 0 0.8 14.5 100
1055 2.0 10 2 1.90 1.34 2.52 0.62 0.04 0.08 0 1.0 5.6 100
a Equilibrium calculations predict quantitative conversion of CS2 under all conditions investigated.
Ind. Eng. Chem. Res., Vol. 40, No. 2, 2001 505
Scheme 2. Conversion of CS2 by H2O at Claus Scheme 3. Pathways for Reaction of CS2 with SO2
Furnace Conditions
equations are shared to simplify the scheme. Also, some (4) Hyne, J. B.; Goar, B. G. The Claus Revisited-Part 4.
reactions are repeated within the scheme to highlight Sulphur 1997, 248, 43-48.
the formation of certain species (CO, COS) by more than (5) Kerr, R. K.; Paskall, H. G.; Ballash, N. Claus Process:
Catalytic Kinetics: Part 1. Modified Claus Reaction. Energy
one pathway. Species highlighted in squares can be Process. Can. 1976, 5, 66-71.
considered to represent feed species in the Claus (6) Kerr, R. K. and Paskall, H. G. Claus Process: Catalytic
furnace, although SO2, of course, is derived from the Kinetics: Part 2. COS and CS2 Hydrolysis. Energy Process. Can.
primary reaction of H2S with O2. 1976, 6, 38-44.
One of the most interesting aspects of this chemistry (7) Paskall, H. G. Capability of the Modified Claus Process. A
is the high reactivity of SO2 in COS, CS2, and CO Final Report to the Department of Energy and Natural Resources
of the Province of Alberta; Western Research and Development:
formation and destruction. Indeed, SO2 can actually be Calgary, Alberta, Canada, 1979; Chapter IV.
formed by reaction of CO2 with S2 although this reaction (8) Paskall, H. G. Reaction Furnace Chemistry and Operational
would only proceed quickly in the flame zone. Also, it Modes. Sulphur Recovery; Western Research: Calgary, Alberta,
is evident that the amount of hydrocarbon impurity 1981 (republished 1990).
being fed with the acid gas will have a marked influence (9) Sames, S. A.; Paskall, H. G. So You Don’t Have a COS/CS2
on the CO, COS, and CS2 levels of the furnace/WHB Problem, Eh?. Sulphur Recovery; Western Research: Calgary,
Alberta, 1984 (republished 1990).
process gas as these species derive directly from hydro-
(10) Monnery, W. D.; Svrcek, W. Y.; Behie, L. A. Modeling of
carbon reactions and the myriad of CS2 conversion the Modified Claus Process Reaction Furnace and the Implications
pathways. It should be noted, however, that hydrocar- on Plant Design and Recovery. Can. J. Chem. Eng. 1993, 71, 711.
bons entering the furnace could undergo the desired (11) Karan, K.; Mehrotra, A. K.; Behie, L. A. COS-Forming
oxidation to CO2, although partial oxidation or steam Reaction Between CO and Sulfur: A High-Temperature Intrinsic
reforming would lead to CO, among other products (see Kinetics Study. Ind. Eng. Chem. Res. 1998, 37, 4609.
ref 21 for further discussion of these points). As can be (12) Karan, K.; Mehrotra, A. K.; Behie, L. A. Use of New
Kinetics for COS Formation to Achieve Reduced Sulfur Emissions
noted from Scheme 5, multiple pathways give rise to from Claus Plants. Can. J. Chem. Eng. 1999, 77, 392.
CO. Thus, all of these reactions will affect the CO/S2/ (13) Sames, J. A.; Paskall. H. G.; Brown, D. M.; Chen, M. S.
COS chemistry in the WHB by virtue of controlling the K.; Sulkowski, D. Field Measurements of Hydrogen Production
CO concentration entering that unit. in an Oxygen-Enriched Claus Furnace. Proceedings of the Sulfur
Overall, it would appear that a vast amount of kinetic 1990 International Conference, Cancun, Mexico, 1990; pp 89-105.
data on individual reactions would be needed to model (14) Towler, G. P.; Lynn. S. Development of a Zero-Emissions
Sulfur Recovery Process: 1. Thermochemistry and Reaction Kinet-
CO and COS species within the furnace/WHB system. ics of Mixtures of H2S and CO2 at High Temperature. Ind. Eng.
Although this task is not impossible, we are taking the Chem. Res. 1993, 32, 2800.
more direct approach of producing data sets under (15) Faraji, F.; Safarik, I.; Strausz, O. P.; Torres, M. E.;
relevant combustion conditions with the goal of produc- Yildirim, E. CO-Catalyzed Conversion of H2S to H2S + S. 1.
ing a global kinetic model from these data. Reaction Between CO and H2S. Ind. Eng. Chem. Res. 1996, 35,
3854.
(16) Dowling, N. I.; Clark, P. D. Kinetic Modeling of the
Acknowledgment Reaction Between Hydrogen and Sulfur and Opposing H2S De-
composition at High Temperatures. Ind. Eng. Chem. Res. 1999,
We thank the member companies of Alberta Sulphur 38, 1369.
Research Ltd. for supporting these studies: AEC West; (17) Graven, W. M.; Long, F. J. Kinetics and Mechanisms of
Agrium Inc.; Air Liquide Canada Inc.; Air Products & the Two Opposing Reactions of the Equilibrium CO + H2O f CO2
Chemicals Inc.; ALCOA Industrial Chemicals; Ametek + H2. J. Am. Chem. Soc. 1954, 76, 2602.
Process & Analytical Instruments Division; Amoco (18) Sames, J. A.; Paskall, H. G. So You Don't Have a COS/
CS2 Problem, Eh?. Sulphur 1984, 172, 47-50.
Canada Petroleum Company Ltd.; Apollo Environmen- (19) Svrcek, W. Y. Kinetic Modeling of the Modified Claus
tal Systems Corp.; Bechtel Corporation; BG Technology; Process Reaction Furnace; Contract #509-4913-2795; Progress
Black & Veatch Pritchard Inc.; BOC Group; Brimstone Report, July-December 1998; Gas Research Institute: Chicago,
Engineering Services Inc.; Burlington Resources; Ca- IL, 1998.
nadian 88 Energy Corporation; Canwell Enviro-Indus- (20) Clark, P. D.; Dowling, N. I.; Huang, M. Mechanisms of CS2
tries Ltd.; Chevron Canada Resources; Crescent Tech- Formation in the Claus Front End Reaction Furnace. Proceedings
of the 47th Annual Laurance Reid Gas Conditioning Conference;
nology Inc.; Delta Hudson Engineering Ltd.; Dow
College of Continuing Education, The University of Oklahoma,
Chemicals U.S.A.; Enersul Inc.; ENI S.p.A.; ESSECO Norman, OK, March 2-5, 1997; p 321.
S.p.A.; Exxon Production Research Company; Goar, (21) Clark, P. D., Dowling, N. I.; Huang, M. Understanding
Allison & Associates Inc.; Husky Oil Operations Ltd.; Claus Furnace Chemistry: Development of a “Modified” Claus for
Kvaerner Process Systems; Parsons Energy & Chemi- Low H2S-Content Acid Gases. Proceedings of the 48th Annual
cals Group Inc.; Petro-Canada Oil and Gas; Porocel Laurance Reid Gas Conditioning Conference; College of Continuing
Adsorbents and Catalysts; Prism Sulfur Corporation; Education, The University of Oklahoma, Norman, OK, March 1-4,
1998; pp 241-263.
Procatalyse; Raisio Chemicals Canada Inc.; Sandvik (22) Ray, J.-L.; Nedez, C.; Legendre, O. Improving Sulphur
Process Systems Inc.; Shell Canada Limited; SiiRTEC Recovery in Claus Plant Operations by the Use of Promoted
NIGI S.p.A.; Stork Engineers & Contractors BV; Sultran Aluminas. Proceedings of the 1995 British Sulphur Conference,
Ltd.; Suncor Energy Inc.; Syncrude Canada Ltd.; Talis- Abu Dhabi, UAE, October 15-18, 1995; p 111.
man Energy Inc.; and TPA, Inc. (23) White, W. B.; Johnson, S. M.; Dantzig, G. B. Chemical
Equilibrium in Complex Mixtures. J. Chem. Phys. 1958, 28, 751.
(24) Chase, M. W.; Davies, C. A.; Downey, J. R.; Frurip, D. J.;
Literature Cited McDonald, R. A.; Syverud, A. N. JANAF Thermochemical Tables,
3rd ed.; American Chemical Society and the American Institute
(1) Hyne, J. B.; Goar, B. G. The Claus Revisited-Part 1. of Physics for the National Bureau of Standards: Washington,
Sulphur 1996 a, 245, 39-44. D.C., 1985.
(2) Hyne, J. B.; Goar, B. G. The Claus Revisited-Part 2. (25) Holub, R.; Vonka, P. The Chemical Equilibrium of Gaseous
Sulphur 1996 b, 246, 73-80. Systems; D. Reidel Publishing Co.: Boston, MA, 1976.
(3) Hyne, J. B.; Goar, B. G. The Claus Revisited-Part 3. (26) Cullis, C. F.; Mulcahy, M. F. R. The Kinetics of Combustion
Sulphur 1996 c, 247, 40-45. of Gaseous Sulphur Compounds. Combust. Flame 1972, 18, 225.
508 Ind. Eng. Chem. Res., Vol. 40, No. 2, 2001
(27) Kim, H.; Park, D. W.; Woo, H. C.; Chung, J. S. Reduction (33) Woiki, D.; Roth, P. Shock Tube Measurements on the
of SO2 by CO to Elemental Sulfur Over Co3O4-TiO2 Catalysts. Thermal Decomposition of COS. Ber. Bunsen-Ges. Phys. Chem.
Appl. Catal. B 1998, 19, 233. 1992, 96 (1), 1347.
(28) Hibbert, D. B. Reduction of Sulfur Dioxide on Perovskite
Oxides. Catal. Rev.-Sci. Eng. 1992, 34 (4), 391. (34) Woiki, D.; Markus M. W.; Roth, P. A Shock Tube-Laser
(29) Khalafalla, S. E.; Foerster, F.; Haas, L. A. Catalytic Flash Photolysis Study of COS + S f CO + S2. J. Phys. Chem.
Reduction of Sulfur Dioxide on Iron-Alumina Bifunctional Cata- 1993, 97, 9682.
lysts. Ind. Eng. Chem. Prod. Res. Dev. 1971, 10 (2), 133. (35) Schecker, H. G.; Wagner, H. Gg. On the Thermal Decom-
(30) Arutyunov, V. S.; Baserich, V. Ya.; Vedeneev, V. I.; position of COS. Int. J. Chem. Kinet. 1969, 1, 541.
Ushakov, V. A.; Chernysheva, A. V. Kinetics of Sulfur Dioxide
Reduction. II. Reduction by Carbon Monoxide. Kinet. Catal. 1991,
31, 919.
(31) Fern, R. J. The Chemistry of Carbonyl Sulfide. Chem. Rev. Received for review December 2, 1999
1957, 57, 621. Revised manuscript received October 24, 2000
(32) Nemeth, L.; Gati, G.; Gervasini, A.; Aaroux, A.; Mink, G.; Accepted October 26, 2000
Pap, I. S.; Szekely, T. Kinetic Study of the Carbonyl Sulphide
Synthesis from Carbon Dioxide and Carbon Disulphide on Alu-
mina Catalysts. Appl. Catal. 1990, 64, 143. IE990871L