Activation of La2O3 for the catalytic reduction of SO2 by CO

Article abstract of DOI:10.1006/jcat.1996.0328

Lanthanum oxide by itself has been reported to be inactive for the reduction of SO₂ by CO; however, we discovered that La₂O₃ can be activated in a reaction gas atmosphere if previously hydrated. The active phase of the activated lanthanum oxide was identified to be lanthanum oxysulfide by using X-ray diffraction. This showed that the activation was in essence a sulfidization of the oxide. An activation mechanism is proposed based on the surface acidic and basic properties resulting from the dehydroxylation of the lanthanum hydroxide.

Full text of DOI:10.1006/jcat.1996.0328

JOURNAL OF CATALYSIS 163, 271–278 (1996)
ARTICLE NO. 0328
Activation of La2O3 for the Catalytic Reduction of SO2 by CO
Jianxin Ma,1 Ming Fang,2 and Ngai Ting Lau
Research Centre, The Hong Kong University of Science & Technology, Clearwater Bay, Hong Kong
Received September 18, 1995; revised June 18, 1996; accepted July 15, 1996
summarizing what has been learned from the work on the
catalytic oxidative coupling of methane using lanthanum
catalysts (11), it becomes apparent that the difference in
the catalytic performance reported by the various research
groups may be due to the extreme sensitivity of lanthanum
oxide to water vapor and CO2. Indeed, the freshly high-
temperature-calcined hexagonal A-form La2O3 can suffer
bulk hydration and carbonation in lessthan 24h ofexposure
to the atmosphere (12). Furthermore, it has been found that
the lanthanum in the perovskite-type oxides is sulfidized
into sulfides, oxysulfide, and/or oxysulfate after reaction
(8, 9, 13); thus, it is the general belief that lanthanum in
the oxide form cannot be responsible for the catalytic ac-
Lanthanum oxide by itself has been reported to be inactive for
the reduction of SO2 by CO; however, we discovered that La2O3 can
be activated in a reaction gas atmosphere if previously hydrated.
The active phase of the activated lanthanum oxide was identified to
be lanthanum oxysulfide by using X-ray diffraction. This showed
that the activation was in essence a sulfidization of the oxide. An
activation mechanism is proposed based on the surface acidic
and basic properties resulting from the dehydroxylation of the
c
lanthanum hydroxide.
ꢀ 1996 Academic Press, Inc.
INTRODUCTION
Sulfur dioxide discharged into the atmosphere as a prod- tivity. In the literature, there is little detailed description on
uct ofcombustion processes, such ascoal-fired power plants, the preparation of the La2O3, and even more scarce is the
is one of the major components of acid rain and other forms discussion on whether lanthanum oxide and/or its sulfidized
of atmospheric pollution. Numerous flue gas desulfuriza- derivatives are good catalysts for the SO2 reduction reac-
tion (FGD) processes have been commercialized (1), and tion. Therefore, we decided to evaluate the performance of
most of them are wet scrubbing processes. Dry flue gas SO2 lanthanum oxide in the catalytic reduction of SO2 by CO.
reduction methods are highly desirable, especially if the
In the present paper, the findings of our study of the ac-
sulfur can be recovered in a salable form. The catalytic re- tivation of hydrated La2O3, prepared from different start-
duction of SO2 by CO has received much attention because ing materials and/or pretreatment conditions, for the reduc-
it offers a dry desulfurization method using a flue gas con- tion of SO2 by CO are reported. Catalyst samples before
stituent as the reducing agent, while producing elemental and after the reaction were characterized by X-ray powder
sulfur.
diffraction (XRD) to relate the phase composition to the
Among many of the catalysts investigated for the reac- activation behavior and catalytic performance, and a mech-
tion between SO2 and CO, lanthanum oxide has been con- anism is proposed to describe the role of hydration in the
sidered to be inactive by researchers. Palilla (2) observed activation process.
that the activity of La2O3 was much lower than that of the
LaCoO3 perovskite oxide; Happel et al. (3) reported that
pure lanthanum oxide is inactive; and Bajars (4) used a
EXPERIMENTAL
La2O3 catalyst pretreated in a mixture of 75 vol% CO and
In our earlier investigation of the catalytic reduction of
25 vol% N2 at 600ꢁC and obtained not only a low conver-
SO2 by CO using perovskite type oxides as catalyst, we
sion of SO2 but also a substantial formation of COS. De-
found that the oxides can be activated in the reaction gas
spite these reports, it is not difficult to notice that in many
stream due to a sulfidization process (9). We also learned
of these studies lanthanum is a vital component of some of
that the lanthanum, coordinated by 12 oxygen ions in a
the active catalysts developed (2–10). On the other hand,
perovskite oxide, can be sulfidized into sulfide, oxysulfide,
and/or oxysulfate, depending on the particular chemical en-
vironment. The same activation method was used in this
work.
A fixed-bed flow reactor (a 2-cm diameter by 50-cm-long
quartz tube) was used in this study. The reactor was heated
1
Present address: Institute of Industrial Catalysis, East China Univer-
sity of Science & Technology, Shanghai 200237, The People’s Republic of
China.
2
To whom correspondence should be addressed. E-mail:
rcmfang@usthk.ust.hk.
271
0021-9517/96 $18.00
c
Copyright ꢀ 1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.

272
MA, FANG, AND LAU
ilar results can be obtained by simply exposing the La2O3
TABLE 1
powder to ambient air for one week.
Physical Properties of Lanthanum Oxide Samples
LaOOH was prepared by the in situ decomposing of the
hydrated product La(OH)3 at 400ꢁC under vacuum for 4 h.
The BET surface area of the lanthanum compounds be-
fore and after activation was determined using Micromerit-
ics ASAP2010, and the bulk structure was characterized
using X-ray powder diffraction (Philips MPD-1880 I X-ray
Sample
no.
Assay Surface area Particle size
Supplier
(% )
(m2 gꢂ1
)
(mesh)
1
Yiaolong Non-Ferrous 99.99
Metals Co. Ltd.
5.4
120–400
2
3
4
5
Fluka Chemika
BDH Lab. Supplies
Merck
97.0
99.5
99.995
99.5
6.1
5.2
4.7
5.6
120–200
120–200
120–200
120–200
˚
diffractometer, CuKꢁ radiation, ꢂ = 1.542 A).
Sigma Chemical Co.
RESULTS
Activation of Pure and Pretreated La2O3
externally in an electrical furnace and the temperature of
the catalyst bed was controlled to within 1ꢁC. The feed gas
entered from the top of the reactor at a constant flow rate
and passed through the catalyst bed into a section where the
reaction products were treated and measured. The reactor
was operated at atmospheric pressure.
Sulfur dioxide concentration in the feed and effluent was
measured using an on-line nondispersive infrared analyzer
(CFA-321A, Horiba). A second on-line infrared gas ana-
lyzer (VIA-510, Horiba) was used to determine the carbon
dioxide concentration in effluent. A HP 5980 Series II Gas
Chromatograph with two columns(one molecular sieve and
one Porapak Q) and two TCD detectors was used to ana-
lyze CO, CO2, and the sulfur-containing compounds such
as SO2, COS, H2S, and CS2. The effluent stream compo-
sition was measured after the elemental sulfur has been
condensed by passing the effluent through an ice-bath trap
and a filter with a pore size of 2 ꢀm.
For safety reasons and better control, the feed gas con-
sisting of 0.5 vol% SO2, 1.0 vol% CO, and balanced with
UHP nitrogen was prepared in a gas blender. The feed flow
rate was held at 180 ml minꢂ1 measured at atmospheric
pressure for all activation experiments performed.
Five lanthanum oxide samples purchased from different
sources were used in the activation study. The specific sur-
face area and the particle size of the samples are summa-
rized in Table 1. All experiments were conducted using 0.5 g
of the lanthanum oxides without any pretreatment.
In addition to the above as-received samples, catalyst pre-
cursors prepared by different pretreatment methods were
also investigated. The preparation of the precursors are de-
scribed in the following paragraphs.
La2O3. A freshly prepared La2O3 sample was heated
in the reactor to 600ꢁC in air purified by activated carbon
and molecular sieve 5A for 30 min before the reaction gas
mixture was introduced. In our reaction system, SO2 is a re-
actant, while CO2 is a product; therefore, by following the
concentration change of these two compounds in the reac-
tor, we can interpret what is happening inside the reactor.
Curve a in Figs. 1a and 1b shows the concentration changes
of SO2 and CO2 in the reactor. The SO2 concentration de-
creased at the beginningofthe reaction and returned almost
to the feed concentration after about 5 min, while the CO2
curve showed a small peak before decreasing to zero. If the
CO2 formed wasdue to the reaction ofSO2 and CO, then the
molar ratio between the consumption and formation rates
would be 1 : 2. By integrating the peak areas under the cor-
responding curves, the amount of SO2 consumed and CO2
formed were estimated to be 0.049 and 0.019 mmol, respec-
tively, indicating there must exist another pathway for the
disappearance of SO2, or the formation of CO2 was not due
to the above mentioned reaction alone. Furthermore, the
small CO2 peak and the deep SO2 valley have also been
observed when the feed contained no SO2 or CO, thus the
CO2 could be formed by the reaction of CO with the active
oxygen species adsorbed on the surface. We can in turn say
that the consumption of SO2 was due to the adsorption on
La2O3.
After 6 h of treatment, only a minimal change in the
SO2 and CO2 concentrations was observed showing that
the A-La2O3 could not be easily activated. This result is
consistent with the reports in the literature.
La(OH)3. A calcined La2O3 powder sample was ex-
The phase pure La2O3 was prepared by first calcining the posed to an atmosphere saturated with water vapor for
oxide powder purchased from Yiaolong at 900ꢁC for 12 h. 1 week. XRD analysis revealed that the La2O3 was trans-
It was then quickly transferred to the reactor and calcined formed to La(OH)3. The hydrated lanthanum oxide was
in situ at 600ꢁC under a dry air stream for half an hour activated in the same manner described earlier, and the re-
before performing the activation experiments.
sults are shown as curve b in Figs. 1a and 1b.
The hydration and carbonation of lanthanum oxide were
It is very interesting to see that there was a second sharp
carried out by exposing the calcined La2O3, as described decrease in the SO2 concentration after about 6 min, while
above, in a sealed vessel containing water vapor or with the CO2 concentration increased synchronously; after 1 h
CO2 produced from dry ice for 1 week. Alternatively, sim- we observed the deposition of elemental sulfur in the cold

ACTIVATION OF La2O3 FOR CATALYTIC REDUCTION OF SO2 BY CO
273
FIG. 1. Change in (a) SO2 and (b) CO2 concentration during activation. Curve a: A-La2O3; Curve b: La(OH)3; Curve c: La2O3 exposed to dry ice
vapor; Curve d: La2O3 exposed to air.
trap. This clearly shows that sulfur dioxide was catalyti- (curve d) for comparison to the other samples. As can be
cally reduced to sulfur and the water-vapor-pretreated lan- seen, the pretreatment of lanthanum oxide in air is also very
thanum oxide could be and was activated. Furthermore, effective in achieving activation. The adsorption of SO2 of
the decrease in the SO2 concentration at the beginning was this catalyst was the same as the sample treated with water
much greater than that for the unexposed A-La2O3, indicat- vapor. It was found that the major phase after the exposure
ing that the adsorption of SO2 on the oxide was significantly was also La(OH)3.
promoted by the hydration process.
Comparison of untreated La2O3 samples purchased from
different sources. The activities of six as-received lan-
thanum oxide samples from different sources after acti-
vation are summarized in Table 2. The activity is defined
as the average specific SO2 conversion rate at 600ꢁC at
the termination of activation; i.e., the SO2 concentration
in the effluent does not change any more. The samples
from Yiaolong and Fluka were found to be more active
than the other three, while the sample from Merck was al-
most inactive. It should be mentioned that the containers
of the Yiaolong and Fluka samples were opened a year
ago, while the activation experiments using samples from
La2O3 exposed to dry ice vapor. A calcined La2O3 pow-
der sample was exposed to CO2 produced from dry ice in a
closed vessel for one week and it was found that this sample
could also be easily activated. See curve c in Figs. 1a and 1b.
This sample behaved virtually the same as La(OH)3, differ-
ing only in that the second sharp decrease in SO2 occurred
some 5 min later and the adsorption of SO2 was weaker
(0.073 mmol).
Indeed, XRD analysis revealed that after such a treat-
ment, the sample was almost all La(OH)3 with only a small
quantity of La2OCO3. The formation of La(OH)3 was prob-
ably due to the moisture in the atmosphere condensed by
the dry ice. This implies that the hydration of lanthanum ox-
ide is much easier than carbonation and is consistent with
the observation made by Bernal et al. (12). They observed
the lanthanum hydroxide in three different air-stabilized
lanthanum samples was only partially transformed to the
hydroxycarbonate phase, and there was no further progress
of the carbonation reaction after exposing the samples to
atmospheric CO2 for more than 3 years.
TABLE 2
Activities of Untreated Lanthanum Oxide Samples
after Activation
SO2
Surface area
Conversion
(% )
Ave. specific conv.
(mol mꢂ2 sꢂ1) ꢃ 107
Sample
(m2 gꢂ1
)
Yiaolong
Fluka
BDH
Merck
Sigma
4.0
6.0
4.9
3.1
4.9
94.6
86.5
61.2
6.1
3.1
1.9
1.7
0.3
2.1
La2O3 exposed to air. A sample of the calcined La2O3
powder was exposed to air at room conditions (relative
humidity, 85–95% ) for one week and the activation curve
of the air-exposed sample is presented in Figs. 1a and 1b
77.1

274
MA, FANG, AND LAU
Sigma, BDH, and Merck were conducted immediately after
opening the storage bottles. XRD analysis revealed that the
Yiaolong sample was completely changed to the hydroxide,
the Merck sample maintained the pure La2O3 phase, and
the other samples were found to be partially hydrated to
different extent.
Influence of the dehydroxylation of La(OH)3 on activa-
tion. According to Bernal et al. (12, 14), LaOOH is formed
in the decomposition of lanthanum hydroxide at 377ꢁC be-
fore forming the oxide at a higher temperature:
La(OH)3 → 2 LaOOH
→
La2O3
[1]
ꢄ 527ꢁC (ꢂH2O)
ꢄ 377ꢁC (ꢂ2H2O)
In this work we prepared the lanthanum oxyhydroxide
in situ in the reactor in an inert atmosphere at a decompo-
sition temperature of 400ꢁC. The activation course of the
decomposed product is shown in Fig. 2, and it can be seen
that the SO2 concentration, after an initial decrease, did not
return to as high a level as that shown in Fig. 1a, indicat-
ing that the reactivity of the dehydrated product LaOOH
increased significantly.
FIG. 3. Activation of La2O3 by means of water injection.
in the bone-dry reactive gas mixture, however, activation
was initiated as soon as 0.05 ml of water was injected into
the gas stream. Furthermore, it can be seen that the fur-
ther injection of water had no influence on the steady-state
concentrations of SO2 and CO2.
Effect of water during the activation of La2O3. So far we
have demonstrated that the prehydrated lanthanum oxide
can be easily activated to catalyze the reduction of SO2
by CO; there is, however, still another way to hydrate the
oxide: by injecting water into the reaction gas stream during
activation. This method has practical significance because
flue gas usually contains water vapor in addition to sulfur
dioxide.
Phase composition. All of the samples before and after
activation were characterized using XRD. The phase com-
position identified are summarized in Table 3. Some of the
diffractograms are presented in Figs. 4a and 4b. The results
show that La2O2S was always formed after activation, and
if the major phase in the starting material was La(OH)3,
then La2O2S existed as the major phase. In the case of the
Yiaolong sample, the only identifiable phase was La2O2S
after activation.
In this experiment, an in situ calcined La2O3 was used
as the starting material. Figure 3 shows the lanthanum ox-
ide could not be activated after about 90 min of exposure
DISCUSSION
Active Phase
From the results shown in Table 2, lanthanum oxysulfide
was found to exist in all catalysts after activation, strongly
suggesting that La2O2S was the active phase for the re-
duction reaction. Indeed, there is a relationship between
the catalytic activity and the La2O2S content in the cata-
lyst. The relationship is shown in Fig. 5, where the relative
La2O2S content was calculated from diffractograms of some
of the samples after activation according to the following
equation:
.
X
P = Ii
Ii ꢃ 100.
[2]
i
Here P is the content of a species in phase i in % and
i
Ii is the first intensive diffraction peak of the ith phase.
FIG. 2. Activation of decomposed La(OH)3 sample.
This is also consistent with the observations made in our

ACTIVATION OF La2O3 FOR CATALYTIC REDUCTION OF SO2 BY CO
275
TABLE 3
Phase Composition of Catalysts Useda
Description of sample
Before activation
After activation
As received, Yiaolong
La(OH)3
La2O2S
As received from Fluka Chemika
As received from BDH Lab. Supplies
As received from Merck
La2O3, La(OH)3
La2O3, La(OH)3
La2O3
La2O3, La(OH)3, La2O2S, (La2O2S2)
La2O3, La2O2S
La2O3, (La2O2S)
As received from Sigma
La2O3, La(OH)3
La2O3
La(OH)3
La(OH)3, (La2O2CO3)
La(OH)3, (La2O2CO3, La2O(CO3)2)
LaOOH, La2O3, La(OH)3
La2O3, La2O2S
La2O3, (La2O2S), (La(OH)3)
La2O2S, (La2O3)
La2O2S, (La2O3), (La2O2CO3)
La2O2S, (La2O(CO3)2)
La2O2S2, La2O2S, La2O3, (La2S3)
Dissolved, precipitated, and calcined at 900ꢁC for 12 h
Exposure to saturated water vapor for 1 week
Exposure to CO2 (dry ice) for 1 week
Exposure to atmospheric air for 1 week
Decomposed in situ for 4 h
a Minor phases are in parentheses.
previous investigation using perovskite oxides as catalyst, means of CO adsorption and found that both acidic and
where La2O2S was always found to exist as the major phase basic sites existed on the surface of hydroxylated samples
in the active catalysts (9).
after evacuation at various temperatures. The removal of
certain OH groups after outgassing treatment led to the
formation of coordinatively unsaturated La ions which are
Effect of Hydration
The activation results and XRD analysis showed that the strong Lewis acid sites. It was found that after outgassing
hydration of lanthanum oxide, regardless of the method at 573ꢁK, the La(OH)3 phase decomposed and the bound-
used, promotes the sulfidization of the oxide to oxysulfide. ing of the OH groups to one lanthanum atom was formed,
Using Eq. [2], the content of La(OH)3 in the sample before providing strong Lewis base (or Brønsted acid) sites. It
activation and the La2O2S concentration after activation was further found that the pretreatment of the catalyst
can be correlated and a linear relationship is obtained, as at temperatures ꢅ773ꢁK resulted in the formation of ba-
shown in Fig. 6. This shows the effect of hydration on the sic O2ꢂ sites because of the dehydroxylation. Li et al. (17)
activation of the catalyst.
reported that surface unsaturated sites on CeO2 can be
It is customary to characterize an oxide by its acid and generated by dehydroxylation at high temperatures, and
base properties which in turn are dependent on the de- these sites are vital to the activation of CO to form carbon-
gree of dehydroxylation. In addition, surface OH groups ate and inorganic carboxylate species. A severe change in
themselves can play the role of acidic or basic sites. The in- the surface properties would be experienced when a lan-
fluence of surface hydroxyl groups on the adsorption prop- thanum oxide was rehydrated and dehydroxylated, and it
erties and catalytic activity of rare-earth oxides has been would have the capability in promoting the interaction of
reported for several systems in literature and was reviewed SO2 and/or CO with the surface and thus achieving activa-
by Netzer and Bertel (15). Tsyganenko et al. (16) investi- tion. This phenomenon will be discussed in the following
gated the acid and base properties of lanthanum oxide by section.
FIG. 4. X-ray diffractograms of some samples (a) before and (b) after activation.

276
MA, FANG, AND LAU
FIG. 7. Chemisorption of CO and SO2 on dehydrated La2O3 surface.
sorption of CO on Lewis acid sites is cationic and the ad-
sorption of SO2 on the basic part of a Brønsted sites is an-
ionic; thus, the interaction between the two adsorbates on
two neighboring sites could be the transfer of one oxygen
atom from SO2 to CO, resulting in the formation of a more
reactive intermediate, sulfur monoxide, which then reacts
with the carbon monoxide and the solid to form lanthanum
oxysulfide. If there were no such Lewis–Brønsted pairs
present on the solid, such as the case with lanthanum oxide,
the formation of SO and the conversion of lanthanum oxide
to oxysulfide would be difficult, because the SO formation
step is endoenergic and the chance of interaction between
SO2 and CO is low. This is why lanthanum oxide calcined
at high temperature cannot be activated, but can be after
rehydration. Based on the results and discussion above, the
following activation mechanism can be proposed:
FIG. 5. Relationship between La2O2S content and specific activity.
Activation Mechanism
The foregoing discussion points out that (1) the rehydra-
tion of the lanthanum oxide increases its adsorption of SO2
and greatly promotes activation; (2) it is easier to activate
the LaOOH formed by the decomposition of La(OH)3 (the
first step in Reaction [1]); (3) the dehydration may lead to
the formation of a surface with both acidic and basic sites;
and (4) the activated catalyst is lanthanum oxysulfide. It is
known that CO can adsorb on coordinatively unsaturated
metal cations (Lewis acid sites) formed by the removal of
certain OH groups (16), while SO2 is a strong Lewis acid
and therefore will preferably attach to the basic part of a
Brønsted site (18). Thus we can pictorially represent the
chemisorption of CO and SO2 as shown in Fig. 7 in a way
analogous to the surface model for alumina (19). The ad-
2La(OH)3 → 2LaOOH + 2H2O
CO + SO2 → CO2 + SO
[3]
[4]
[5]
[6]
[7]
CO + SO → CO2 + S
2LaOOH + CO → La2O2 + CO2 + H2O
La2O2 + S → La2O2S
Here represents lattice oxygen vacancy. Reaction [6] sug-
gests that it may be easier for the adsorbed CO to obtain
an oxygen atom during the decomposition of LaOOH to
La2O3, because of the formation of structural defects and
the existence of edges or steps where oxygen has a lower co-
ordination number. The overall gas-solid reaction is shown
as follows:
2La(OH)3 + 3CO + SO2 → La2O2S + 3CO2 + 3H2O. [8]
It was observed that COS was formed during activation,
and Fig. 8 is the time-dependent COS concentrations mea-
sured for the lanthanum oxide exposed to air. It can be seen
that the formation of COS increased gradually and reached
a constant level of about 90 ppmv after about 300 min.
Carbonyl sulfide can be formed by the reaction of carbon
monoxide with elemental sulfur:
CO + S → COS.
[9]
FIG. 6. The effect of La(OH)3 content in the starting material on the
formation of La2O2S in the catalyst.
Thus the formation of COS is dependent on the amount of

ACTIVATION OF La2O3 FOR CATALYTIC REDUCTION OF SO2 BY CO
277
sulfur in the reaction system. Elemental sulfur can be pro-
duced by Reaction [5] as well as by the following reactions
catalyzed by lanthanum oxysulfide:
SO2 + 2CO → 2CO2 + S
SO2 + 2COS → 2CO2 + 3S.
[10]
[11]
Intuitively, sulfur plays a balancing act in Reactions [7] and
[9]–[11] during the activation and reduction reactions. As
activation progresses, less sulfur is consumed by Reaction
[7] while at the same time more La2O2S is produced, which
in turn catalyzes Reactions [10] and [11] to produce more
sulfur leading to the formation of COS by Reaction [9].
However, upon the completion of the activation process,
Reaction [7]isno longer involved and the formation ofCOS
is now determined by the competition between Reactions
[9], [10], and [11].
The role of these reactions can be unraveled by exam-
ining the stoichiometry: Reaction [8] shows that the stoi-
chiometry of CO and SO2 during activation is 3 : 1 and is
2 : 1 for the reduction reaction (Reaction [10]).
FIG. 9. Difference in SO2 and CO2 conversion during activation.
ization also took place. As time progressed, the difference
In this experiment, a feed stoichiometric ratio of 2 : 1 was decreased gradually because of the increasing contribution
used. If Reaction [8] were the dominant reaction, the con- of desulfurization. As a result more and more sulfur became
version of CO would be 50% higher than that of SO2; on available for the formation of COS, as reflected in Fig. 8.
the other hand, if Reaction [10] dominated, there would be After about 300 minutes, the difference decreased to zero
no difference between the two conversions.
upon the completion of sulfidization. This is also consistent
It can be seen in Fig. 9 that the difference between the with the constant formation of COS shown in Fig. 8. It is
CO and SO2 conversions was very large and negative at estimated that complete sulfidization requires only about
the beginning of the reaction because of the strong ad- 50 min for 0.5 g of La(OH)3 using a reactive gas containing
sorption of SO2. In this stage no or only little sulfur was 0.5 mol% SO2 and 1.0 mol% CO when no desulfurization
available to form COS; see Fig. 8. Thereafter, the conver- reaction occurred.Therefore we believe Reaction [10] is a
sion of CO became higher than SO2, but by no more than part of the activation process and longer activation time is
20% , indicating that in addition to sulfidization desulfur- needed.
CONCLUSION
Lanthanum oxide can be activated to catalyze the re-
duction of SO2 by CO to elemental sulfur. The activation
is achieved by the rehydration of lanthanum oxide to the
hydroxide prompting the creation of Lewis and Brønsted
sites in the subsequent dehydroxylation of the hydroxide.
These sites are able to chemisorb SO2 and CO simultane-
ously, promoting the formation of reactive SO species and
lattice oxygen vacancies which readily accept sulfur. The
activation is a gas/solid sulfidization reaction between the
reaction gas mixture and the pretreated oxide. The active
phase formed after the sulfudization is found to be lan-
thanum oxysulfide, La2O2S. The catalytic performance of
this material has been investigated and will be reported
elsewhere.
ACKNOWLEDGMENTS
The authors appreciate the invaluable technical support provided by
FIG. 8. Time-dependent COS concentration during activation.
Messrs. K.L. To and H.S. Tsui of the Research Centre, HKUST. We are

278
MA, FANG, AND LAU
also grateful to Mr. Aiping Chen’s contribution to the current work. The
Materials Preparation and Characterisation Centre has been more than
generousin performingXRD on our behalf, and we are indebted to itsstaff.
9. Ma, J. X., Fang, M., and Lau, N. T., in “Environmental Catalysis”
(G. Centi, and S. Perathoner, Eds.), EFCE Publications, Series 112,
p. 555. Societa’ Chemica Italiana, Rome, 1995.
10. Liu, W., and Flytzani-Stephanopoulos, M., in “Environmental Catal-
ysis” (J. N. Armor, Ed.), Am. Chem. Soc., Symp. Ser., Vol. 552, p. 375.
Am. Chem. Soc., Washington, DC, 1994.
REFERENCES
11. Taylor, R. P., and Schrader, G. L., Ind. Eng. Chem. Res. 30, 1016 (1991).
1. Noyes, R. (Ed.), “Handbook of Pollution Control Processes.” Noyes, 12. Bernal, S., Diaz, J. A., Garcia, R., and Rodriguez-Izquierdo, J. M.,
Park Ridge, NJ, 1991.
J. Mater. Sci. 20, 537 (1985).
2. Palilla, F. C., U.S. Patent 3,931,393 (1976).
3. Happel, J., Leon, M. A., Hnatow, M. A., and Bajars, L., Ind. Eng.
Chem. Prod. Res. Dev. 16, 150 (1977).
13. Baglio, J. A., Ind. Eng. Chem. Prod. Res. Dev. 21, 38 (1982).
14. Bernal, S., Botana, F. J., Garcia, R., and, Rodriguez-Izquierdo, J. M.,
Thermochim. Acta 66, 139 (1983).
4. Bajars, L., U.S. Patent 3,978,200 (1976).
5. Happel, J., Hnatow, M. A., Bajars, L., and Kundrath, M., Ind. Eng.
Chem. Prod. Res. Dev. 14, 154 (1975).
15. Netzer, F. P., and Bertel, E., in “Handbook on the Physics and Chem-
istry of Rare Earths” (K. A. Gschneidner and L. Eyring, Eds.), p. 217.
North-Holland, Amsterdam, 1982.
6. Hibbert, D. B., and Tseung, A. C. C., J. Chem. Tech. Biotechnol. 29,
713 (1979).
16. Tsganenko, A. A., Lamotte, J., Gallas, J. P., and Lavalley, J. C., J. Phys.
Chem. 93, 4179 (1989).
7. Hibbert, D. B., and Champbell, R. H., Appl. Catal. 41, 273
(1988).
17. Li, C., Sakata, Y., Arai, T., Domen, K., and Onishi, T., J. Chem. Soc.
Faraday Trans. 1 85, (1989).
8. Hibbert, D. B., and Champbell, R. H., Appl. Catal. 41, 289
(1988).
18. Khalafalla, S. E., and Haas, L. A., J. Catal. 24, 115 (1972).
19. Khalafalla, S. E., and Haas, L. A., J. Catal. 24, 121 (1972).