Reduction of SO2 by CO and COS over La2O2S - A mechanistic study

Article abstract of DOI:10.1016/S1381-1169(03)00369-8

The mechanism of the catalytic reduction of SO₂ by CO and COS over La₂O₂S was studied using a combination of TPR coupled with MS and feed-perturbation step-analysis methods. COS interacted more readily and favorably with La₂O₂S compared to CO. The interaction of COS with the lanthanum catalyst differed from that of CO. COS preferentially reacted with SO₂ to produce CO₂ and sulfur instead of following the disproportionation (to form CO₂ and COS) and decomposition (to form CO and sulfur) reactions. The redox mechanism was not the primary reaction route for the reduction of SO₂ by CO and COS over La₂O₂S. SO₂ was strongly adsorbed by the oxysulfide and remained in the oxysulfide even when heated in an inert gas stream, while COS was very reactive and did not remain long in the oxysulfide. COS had access to certain specific active sites in the oxysulfide, and the catalytic reduction of SO₂ by COS possibly occurred via adsorbed COS reacting with pools of SO₂ adspecies to form CO₂ and sulfur.

Full text of DOI:10.1016/S1381-1169(03)00369-8

Journal of Molecular Catalysis A: Chemical 203 (2003) 221–229
Reduction of SO2 by CO and COS over
La2O2S—a mechanistic study
Ngai T. Laua,∗, Ming Fanga, Chak K. Chanb
Institute for Environment & Sustainable Development, Hong Kong University of Science & Technology,
a
Clear Water Bay Road, Hong Kong, China
Department of Chemical Engineering, Hong Kong University of Science & Technology, Clear Water Bay Road, Hong Kong, China
b
Received 28 January 2003; received in revised form 7 April 2003; accepted 7 April 2003
Abstract
The mechanism of the catalytic reduction of SO2 by CO and COS over lanthanum oxysulfide was studied using a combi-
nation of temperature-programmed reaction coupled with mass spectrometry (TPR/MS) and feed-perturbation step-analysis
techniques. The results showed that COS interacts more readily with lanthanum oxysulfide than CO does. The interaction of
COS with the lanthanum catalyst was different from that of CO. In addition, COS preferentially reacted with SO2 to form CO2
and sulfur instead of following the disproportionation (to form CO2 and COS) and decomposition (to form CO and sulfur)
reactions. The redox mechanism is not the main reaction route for the reduction of SO2 by CO and COS over lanthanum
oxysulfide. SO2 was strongly adsorbed by the oxysulfide and was retained in the oxysulfide even when heated in an inert gas
stream, while COS was very reactive and did not remain long in the oxysulfide. These results suggest that COS has access to
certain specific active sites in the oxysulfide and the catalytic reduction of SO2 by COS possibly proceeds via adsorbed COS
reacting with pools of SO2 adspecies to form CO2 and sulfur.
© 2003 Elsevier Science B.V. All rights reserved.
Keywords: Lanthanum oxysulfide; Sulfur dioxide; Carbonyl sulfide; Carbon monoxide; Reduction mechanism
1. Introduction
and the COS-intermediate mechanism, shown in Eqs.
(5)–(7), have been proposed for the reaction. The
redox mechanism in which oxygen vacancies are cre-
ated by reducing the oxide catalysts seems to describe
the oxide catalysts such as ceria and titania well. The
oxygen anion vacancies are the active sites for the
reduction of SO2 and they extract the oxygen atoms
from the sulfur dioxide molecules.
The catalytic reduction of SO2 by CO (Eq. (1))
has been extensively studied using various catalyst
systems such as lanthanum oxysulfide [1], lanthanum
oxysulfide–cobalt disulfide [2,3], titanium dioxide [4],
titanium dioxide–cobalt disulfide [4], ceria [5] and oth-
ers [4].
2CO + SO2 = S + 2CO2
Two generic types of reduction mechanisms,
namely, the redox mechanism, shown in Eqs. (2)–(4),
(1)
cat-O + CO = cat-ᮀ + CO2
cat-ᮀ + SO2 = cat-O + SO
cat-ᮀ + SO = cat-O + S
where ᮀ denotes the oxygen anion vacancy. COS has
been observed to reduce SO2 faster than CO does and it
(2)
(3)
(4)
∗
Corresponding author. Tel.: +86-852-2358-6935;
fax: +86-852-2719-6710.
E-mail address: ntlau@ust.hk (N.T. Lau).
1381-1169/03/$ – see front matter © 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S1381-1169(03)00369-8

222
N.T. Lau et al. / Journal of Molecular Catalysis A: Chemical 203 (2003) 221–229
has been argued that COS is a stronger reducing agent
than CO to produce the active vacancy in Eq. (2) [4].
This redox mechanism suggests that a pre-reduced cat-
alyst can and will more promptly reduce SO2 at the an-
ion vacancies created. In other words, if a pre-reduced
catalyst does not reduce the SO2 which the catalyst
comes into contact immediately after the catalyst treat-
ment, the SO2 reduction reaction will not follow the
redox route.
For catalysts containing lanthanum oxysulfide
and/or cobalt disulfide, the COS mechanism seems
to describe the process better since the intermediate
COS is readily produced by the reaction between CO
and cobalt disulfide.
in this study. The TPR system contains an electrical
furnace with a programmable temperature controller
and a quartz micro-reactor. It has been described else-
where [6]. The lanthanum oxysulfide catalyst used in
this study was prepared by the sulfidization method
described in [7]. Briefly, 0.1 g lanthanum oxysulfide
was heated in a feed of 98 ml min−1 from room tem-
perature (∼20 ◦C) to 600 ◦C at 10 ◦C min−1. The feed
stream for the COS study comprised 820 ppmv COS
in He while that for the CO study contained 820 ppmv
CO in He. The oxysulfide sample was purged with
the feed until the effluent composition was steady be-
fore the heating started. A split stream of the effluent
from the reactor was sent to a high-speed process
mass spectrometer (MS-250, ABB EXTREL) for
composition measurement.
CoS2 + CO = CoS + COS
2COS + SO2 = 3S + 2CO2
CoS + S = CoS2
(5)
(6)
(7)
2.2. Catalyst pre-treatment effect
COS is produced by Eq. (5) and by lanthanum oxysul-
fide and CO [6]. Once formed, COS is more reactive
than CO in reducing SO2 and the reaction is catalyzed
by lanthanum oxysulfide [1,2]. Thus, the COS reduc-
tion of SO2 (Eq. (6)) is favored once COS is formed.
However, in the reported COS-intermediate mecha-
nism studies involving lanthanum oxysulfide, only the
formation of COS via the reaction of CO with cobalt
disulfide and lanthanum oxysulfide has been eluci-
dated in detail [2,3,6]. The reaction between COS and
SO2 (Eq. (6)) has not received much attention and the
mechanism involved is still not clear. It is possible that
the COS reduction of SO2 can proceed via an accel-
erated COS-redox mechanism over the catalyst. Other
possible reaction routes such as the surface reaction
between co-adsorbed COS and SO2 molecules cannot
be ruled out. This paper focuses on the reaction of
COS with SO2 over lanthanum oxysulfide and deter-
mines if the reduction of SO2 over lanthanum oxysul-
fide follows the redox mechanism. The role of COS
as a reaction intermediate in the reduction of SO2 by
CO will also be examined.
This experiment was designed to determine the role
of the redox mechanism in the reduction of SO2 over
lanthanum oxysulfide and to demonstrate the impor-
tance of the simultaneous presence of COS and SO2 to
make the reduction reaction occur. Two 0.1 g catalyst
samples were used.
The first lanthanum oxysulfide sample was main-
tained at 600 ◦C in a 90 ml min−1 He carrier stream
for 1 h before a reaction stream containing 820 ppmv
CO and 410 ppmv SO2 in He was fed to the reactor
at 98 ml min−1. The reactor was kept at 600 ◦C until
the reduction of SO2 started again. The catalyst sam-
ple was then reduced by a reduction stream contain-
ing 820 ppmv CO in He at 98 ml min−1 at 600 ◦C for
2 h before switching back to the reaction stream. The
reduced catalyst was then kept at 600 ◦C for 2 h.
The second lanthanum oxysulfide sample was
heated in the reduction stream at 600 ◦C for 2 h.
It was then cooled to room temperature in the He
carrier stream before the COS reaction stream con-
taining 820 ppmv COS and 410 ppmv SO2 in He at
98 ml min−1 was switched into the reactor. The sam-
ple was finally heated to 600 ◦C at 10 ◦C min−1 after
the composition of the effluent became steady.
2. Experimental
2.3. Step changes in feed composition
2.1. Interaction of COS and CO with La2O2S
The temperature-programmed reaction technique
coupled with mass spectrometry (TPR/MS) was used
About 0.05 g lanthanum oxysulfide was heated in
the reactor to the reaction temperatures with Ar flow.

N.T. Lau et al. / Journal of Molecular Catalysis A: Chemical 203 (2003) 221–229
223
Three temperature steps at 335 ◦C, 400 ◦C and 600 ◦C
were used to give low, moderate and almost complete
conversion of COS for the reduction of SO2. Three
feed streams at a flow rate of 100 ml min−1 with the
following compositions were used:
1. 2200 ppmv SO2, 3600 ppmv COS and 4.6% He in
Ar;
2. 3500 ppmv SO2 and 6.7% He in Ar; and
3. 5000 ppmv COS and 4.4% He in Ar.
The feed sequence was (1)→(2)→(3)→(2)→(1)→
(3)→(1). This sequence was chosen to examine the
effect of the SO2 and COS streams on each other.
Sufficient time was allowed for the composition of the
effluent to reach a steady state after the feed switch.
The oxysulfide sample was then purged with Ar until
the SO2 and COS in the effluent was zeroed out before
a new feed gas stream was switched to the reactor.
Fig. 1. TPR of lanthanum oxysulfide using CO as the reducing
agent. The relative concentration curves were obtained by adjusting
the raw MS intensity data for zero, the fragmentation effect and
instrument sensitivity. Note that the CO curve is a consumption
curve, which was computed from the difference of the CO in the
feed and the effluent. All the other curves are production curves.
3. Results and discussion
3.1. Interaction of COS and CO with lanthanum
oxysulfide
3.1.1. Carbon monoxide
∗
∗
where S is a lattice sulfide ion and is an anion va-
cancy with two negative charges created in lanthanum
oxysulfide. Two types of CO oxidation processes are
thought to be associated with the CO2 formation [6].
They include a low-temperature oxidation attributable
to the loosely bound oxygen species in the surface of
the oxysulfide, and a high-temperature oxidation as-
sociated with the more tightly bound lattice oxygen,
which becomes available when the temperatures are
high enough.
The TPR results for the CO reduction of lanthanum
oxysulfide from room temperature to 600 ◦C are
shown in Fig. 1. There were no significant changes in
the relative concentrations of CO, CO2, COS and CS2
at temperatures below 300 ◦C. CO2 began to appear
in the effluent at just above 300 ◦C, well before the
CO was consumed. This was probably due to the des-
orption of the CO2 adspecies formed by the reaction
of the pre-adsorbed CO on the lanthanum oxysulfide
[6]. The desorption of the CO2 adspecies made room
for the further adsorption of CO, which could in turn
react to form more CO2. COS was produced later at
higher temperatures. The consumption of CO and the
production of CO2 peaked at about 370 ◦C and rose
abruptly at temperatures above 500 ◦C. On the other
hand, the production of COS peaked at 380 ◦C and
then fell back to a lower level. The mechanism of the
production of COS has been discussed in detail else-
where [6] and can be summarized in the following
two reaction steps:
3.1.2. Carbonyl sulfide
The TPR results for the reaction between COS and
lanthanum oxysulfide are shown in Fig. 2. The reac-
tion started at about 220 ◦C with a moderate produc-
tion of CO2 from the oxidation of COS. The reaction
became more vigorous at temperatures above 400 ◦C
and produced significantly more CO2, COS and CO.
The disproportionation of COS to CO2 and CS2 shown
in Eq. (10) occurred at temperatures around 400 ◦C
and the production of CS2 peaked at around 550 ◦C.
The decomposition of COS to CO and sulfur shown
in Eq. (11) became increasingly important as the
∗S =∗ +S
CO + S = COS
(8)
(9)

224
N.T. Lau et al. / Journal of Molecular Catalysis A: Chemical 203 (2003) 221–229
Fig. 2. TPR of lanthanum oxysulfide using COS as the reducing
agent. The relative concentration curves were obtained by adjusting
the raw MS intensity data for zero, the fragmentation effect and
instrument sensitivity. Note that the COS curve is a consumption
curve, which was computed from the difference of the COS in the
feed and the effluent. All the other curves are production curves.
temperature rose.
2COS = CO2 + CS2
COS = CO + S
(10)
(11)
The oxidation of COS to form CO2 at temperatures
below 300 ◦C was probably due to the oxidation of
the loosely bound oxygen species in the surface of the
oxysulfide catalyst.
A comparison between Figs. 1 and 2 indicates
clearly that both CO and COS reduced lanthanum
oxysulfide but COS interacted more readily with lan-
thanum oxysulfide and in a way different from CO.
A possible reason for the latter is that COS may have
access to certain active sites in lanthanum oxysulfide
that are not available to CO. These active sites could
be the key to unlocking the mechanism of the reduc-
tion of SO2 over lanthanum oxysulfide and thus the
role of the COS intermediate in this reaction.
Fig. 3. Recovery of SO2 reduction reaction by CO over lanthanum
oxysulfide at 600 ◦C after the oxysulfide was heated in a stream
of (a) He and (b) CO. COS curves are scaled up by 25 times. All
these intensity curves have been adjusted for zero but not for the
fragmentation effect and instrument sensitivity.
heated in He and (b) being reduced by CO. When the
SO2 and CO reaction stream was first introduced to the
oxysulfide catalyst, no immediate catalytic reaction
took place. Instead, SO2 was preferentially adsorbed
by the catalyst and displaced both CO and CO2 from
the catalyst. It took more than 1 h for the He-heated
catalyst sample to achieve a significant level of
3.2. Catalyst pre-treatment
Fig. 3 shows the reactivation of the oxysulfide cat-
alyst in the reduction of SO2 by CO after (a) being

N.T. Lau et al. / Journal of Molecular Catalysis A: Chemical 203 (2003) 221–229
225
reaction between SO2 and CO to produce CO2 and
COS (Fig. 3a). The effect of heat-treating the catalyst
in inert gas has been discussed elsewhere [6]. This
treatment could drive off most of the loosely bound
and reactive sulfur from the catalyst causing difficulty
in the production of the reaction intermediate COS,
thus prohibiting the reduction of SO2. Although heat-
ing alone can deactivate the oxysulfide for the SO2
reduction reaction, this deactivation is temporary and
the reaction can resume eventually. Analogous to the
formation of oxygen anion vacancies on lanthanum
oxide [8], Lau and Fang [6] proposed the formation of
anion vacancies on lanthanum oxysulfide upon heat-
ing. Should this reaction follow the redox mechanism,
these vacancies would reduce the SO2 molecules
when they came into contact and thus would reactivate
the oxysulfide catalyst immediately. This immediate
reactivation, however, was not observed.
This phenomenon was even more pronounced for
the sample pre-reduced by CO (Fig. 3b). The sample
could not be reactivated in 2 h. On the other hand,
the as synthesized La2O2S catalyzed the reduction
of SO2 by CO immediately once the SO2/CO mix-
ture was fed to the material at the same reaction
temperature (∼600 ◦C). Thus, the He-treatment and
pre-reduction processes deactivate the catalytic ma-
terial. This strongly supports the argument that the
CO-redox mechanism described by Eqs. (1)–(3) is not
the main reaction path for the reduction of SO2 by
CO over lanthanum oxysulfide. Otherwise, the redox
mechanism would lead to the reactivation of the cat-
alyst as soon as the reaction stream was introduced.
It should also be pointed out that since the interac-
tion between CO and lanthanum oxysulfide produces
COS, the CO-reduced oxysulfide should have been
reduced by COS. Thus, the COS-redox mechanism
could not be a reaction route for the reduction of
SO2. Furthermore, there was no COS produced in the
2 h after the SO2 and CO feed was introduced to the
CO-reduced sample. This observation together with
the simultaneous production of COS during the reac-
tivation of the heat-treated oxysulfide sample, shows
that the catalytic reduction of SO2 over the oxysulfide
catalyst was strongly associated with COS and the si-
multaneous presence of COS and SO2 was vital to the
reaction. This argument is further supported by the
rapid reduction of SO2 by COS over the CO-reduced
catalyst (see Fig. 4).
Fig. 4. Recovery of SO2 reaction by COS in TPR after the oxy-
sulfide was heated in carbon monoxide. All the intensity curves
have been adjusted for zero but not for the fragmentation effect
and instrument sensitivity.
The COS–SO2 reaction started at temperatures as
low as 300 ◦C. A large amount of SO2 and COS was
consumed when the temperature was greater than
400 ◦C. This suggests that COS was more reactive than
CO for the reduction of SO2 over lanthanum oxysul-
fide because the ignition temperature of the CO–SO2
reaction was greater than 360 ◦C [1]. These results
suggest that the active sites for the COS–SO2 reduc-
tion reaction were readily present on the CO-reduced
catalyst at temperatures as low as 300 ◦C. In addition,
it is critical for COS and SO2 to be simultaneously
present over the oxysulfide catalyst for the reduc-
tion of SO2 to occur. The reaction mechanism for
the COS and SO2 reaction could involve the direct
interaction of these two species over the oxysulfide
catalyst.
In the COS reduction of SO2, there was no de-
tectable CS2 and the CO2 peak at ∼200 ◦C, which was
observed in the COS-TPR of lanthanum oxysulfide,
was absent. This suggests that COS was preferentially
scavenged by SO2 before disproportionation could oc-
cur, and the oxidation of COS below 300 ◦C (Fig. 2)
was suppressed in the presence of SO2. Since SO2 was
preferentially adsorbed onto the oxygen ions of the
catalyst [6], the adsorption could have stabilized the

226
N.T. Lau et al. / Journal of Molecular Catalysis A: Chemical 203 (2003) 221–229
loosely bound oxygen as adsorbed sulfite and sulfate
and made the oxygen unavailable to oxidize COS.
3.3. Step changes in feed concentration
In Stream 1, the COS to SO2 ratio was less than
stoichiometric and a portion of SO2 always remained
while COS was completely consumed at 600 ◦C. The
conversion increased with temperature from 335 ◦C to
600 ◦C as illustrated in Figs. 5 and 6a. The steady-state
conversion of COS was only about 10% at 335 ◦C and
70% at 400 ◦C. The responses to the perturbation se-
quence in the feed were the same at all three tempera-
tures and in the following discussion we focus on the
results for 400 ◦C (Fig. 6).
In general, as soon as the COS and SO2 stream
(Stream 1) was introduced to the oxysulfide catalyst,
COS reacted immediately with the SO2 to produce an
equivalent amount of CO2. Only a negligible amount
of COS, if any, was adsorbed to the oxysulfide catalyst.
There was no ratable amount of CS2 found in the
effluent gas. COS is very reactive over the oxysulfide
catalyst and tends to react with SO2 rather than staying
adsorbed on the catalyst or disproportionating to CS2.
When SO2 (Stream 2) was fed to the catalyst, a sig-
nificant amount of SO2 was consumed at the begin-
ning with no COS, CO, CO2 and CS2 found in the
effluent gas suggesting that the SO2 was adsorbed onto
the catalyst. If there were any COS adspecies left on
the catalyst from the previous experiment, CO2 would
have been found in the effluent. Thus, it can be de-
duced that the COS adspecies was too reactive to re-
main on the oxysulfide at 400 ◦C.
When COS (Stream 3) was sequenced into the re-
actor, COS was rapidly consumed and a surge in CO2
was observed. This surge was probably due to the re-
action of COS with the SO2 adspecies left behind from
the previous SO2 containing feed streams. CS2 was
produced at a later stage, as a result of the dispropor-
tionation of COS when the COS–SO2 reaction slowed
down due to the depletion of the SO2 adspecies. No
SO2 was detected throughout this step. The SO2 ad-
species, if present, would be strongly adsorbed by the
oxysulfide.
Fig. 5. Step perturbation with COS + SO2 at (a) 335 ◦C and
(b) 600 ◦C. The relative concentration curves were obtained by
adjusting the raw MS intensity data for zero, the fragmentation
effect and instrument sensitivity. The COS feed, COS consumption
and CO2 curves are nearly coincident indicating almost complete
conversion of COS in the SO2 reduction reaction.
the subsequent reaction over the lanthanum oxysul-
fide catalyst are summarized in Table 1. It is appar-
ent that pre-treating the oxysulfide catalyst with COS
increased the subsequent adsorption of SO2 but SO2
pre-treatment showed no apparent effect on the sub-
sequent reaction with COS. However, the adsorption
The feed-perturbation experiment was carried out
in the following order (a) COS + SO2, (b) SO2, (c)
COS, (d) SO2, (e) COS+SO2, (f) COS and (g) COS+
SO2. The effects of COS and SO2 pre-treatments on

N.T. Lau et al. / Journal of Molecular Catalysis A: Chemical 203 (2003) 221–229
227
of SO2 was reduced by the SO2 pre-treatment. A pos-
sible explanation is that SO2 is strongly adsorbed to
the oxysulfide. The SO2 adspecies is so stable that the
adsorption sites are only freed for further adsorption
when the adspecies reacts with COS to form sulfur
and CO2. Thus, SO2 pre-treatment does not alter the
population of the SO2 adspecies to affect the subse-
quent COS reaction.
Table 1
Effects of COS and SO2 pre-treatment on the subsequent reaction
catalyzed by lanthanum oxysulfide
Pretreatment
Subsequent
reaction
Effects
COS
COS
SO2
SO2
SO2
SO2 adsorption was increased
SO2 adsorption was increased
No significant effect
SO2 + COS
COS
SO2 + COS
SO2 adsorption was reduced
Fig. 6. Step change experiment at 400 ◦C: (a) COS + SO2, (b) SO2, (c) COS, (d) SO2, (e) COS + SO2, (f) COS, and (g) COS + SO2.
The relative concentration curves were obtained by adjusting the raw MS intensity data for zero, the fragmentation effect and instrument
sensitivity.

228
N.T. Lau et al. / Journal of Molecular Catalysis A: Chemical 203 (2003) 221–229
Fig. 6. (Continued ).
In the COS (Stream 3) experiment, the amount of
may not be stable long enough to survive the feed
change and the Ar purge.
CO2 produced at steady state was larger than CS2,
suggesting that in addition to the disproportionation of
COS, the oxysulfide catalyst was also reduced by COS.
Furthermore, more COS was consumed than the total
CO2 and CS2 produced. This indicates the presence
of at least another reaction, i.e. the decomposition of
COS. It is also possible that COS was adsorbed to the
oxysulfide, probably at the adsorption sites opened up
by COS reduction, even though the COS adspecies
4. Conclusions
In summary, we can conclude that the redox mech-
anism route is not the main reaction path for the
reduction of SO2 by CO and COS over lanthanum
oxysulfide. The reduction probably proceeds through a

N.T. Lau et al. / Journal of Molecular Catalysis A: Chemical 203 (2003) 221–229
229
surface reaction path involving both the SO2 and COS
adspecies. Undoubtedly SO2 is readily and more fa-
vorably adsorbed on the lanthanum oxysulfide and
adheres to the oxygen ions with the existence of other
adsorption sites for SO2. Nonetheless, we can imagine
that there are always pools of SO2 molecules adsorbed
on the oxysulfide when SO2 is present in the feed gas.
On the other hand, COS is very reactive and may
have access to certain active sites in the lanthanum
oxysulfide, which are not available to other species
such as CO. The reduction of the oxysulfide by COS
possibly produces more adsorption sites for the reac-
tants. Analogous to lanthanum oxide and other rare
earth oxides [8,9], oxygen vacancies are expected to
form in the oxysulfide, and the COS reduction to the
oxysulfide is expected to promote this vacancy forma-
tion process through the scavenging of the loose oxy-
gen species and possibly via other more direct routes.
That is,
sulfur. The reduction of SO2 by CO follows the same
route but with a step to form the COS intermediate.
Although some of the present results suggest that the
reduction reaction is associated with the anion va-
cancies, understanding how the reduction reaction is
related to these vacancies and how the reactants are
accommodated in these sites needs more research.
Acknowledgements
The authors thank Professor MA Jianxin for his
valuable advice and critical comments and the Re-
search Grants Council, Hong Kong, for the financial
support.
References
[1] J.X. Ma, M. Fang, N.T. Lau, Appl. Catal. A: Gen. 150 (1997)
253.
[2] J.A. Baglio, Ind. Eng. Chem. Prod. Res. Dev. 21 (1982) 38.
[3] J.X. Ma, M. Fang, N.T. Lau, J. Catal. 158 (1996) 251.
[4] H. Kim, D.W. Park, H.C. Woo, J.S. Chung, Appl. Catal. B:
Environ. 19 (1998) 233.
∗O =∗ +O
(12)
∗
∗
where O is a lattice oxide ion and is an anion va-
cancy.
But the COS adspecies is very unstable in the oxy-
sulfide and tends to react immediately with the SO2
coming into contact, to disproportionate to CS2 and
CO2, or to decompose to CO. Thus, the catalytic re-
duction of SO2 by COS probably takes place when
newly adsorbed COS comes into contact with pools
of adsorbed SO2 to react to form carbon dioxide and
[5] W. Liu, A.F. Sarofim, M. Flytzani-Stephanopoulos, Appl. Catal.
B: Environ. 4 (1994) 167.
[6] N.T. Lau, M. Fang, J. Catal. 179 (1998) 343.
[7] J.X. Ma, M. Fang, N.T. Lau, J. Catal. 163 (1996) 271.
[8] E.R.S. Winter, J. Catal. 22 (1971) 158.
[9] F.P. Netzer, E. Bertel, in: K.A. Gschneidner Jr., L. Eyring
(Eds.), Handbook on the Physics and Chemistry of Rare Earths,
vol. 5, North-Holland, Amsterdam, 1982, Chapter 43, p. 217.