Selective Oxidation of H2S over CuO/Al2O3: Identification and Role of the Sulfurated Species formed on the Catalyst during the Reaction

Article abstract of DOI:10.1006/jcat.1999.2691

A CuO/Al₂O₃ catalyst was used for selective oxidation of hydrogen sulfide into sulfur at low temperature (383 K). H₂S is totally consumed during the 28 first hours of the reaction, without detection of SO₂ nor S_n. The decrease of H₂S conversion is concomitant with the detection of sulfur at the bottom of the reactor. Analysis of the catalyst evidenced the presence of several sulfurated species, such as sulfate, sulfur, and especially sulfide species. The sulfide/copper molar ratio varies versus reaction time and reaches values greater than 1, which suggests the formation of polysulfide species. Its maximum value (3.5) is concomitant with the conversion decrease and the formation of S_n. Complementary reactivity experiments show that the active phase cannot be a simple copper sulfide CuS and underline the peculiar role of oxygen in the network. It finally seems that the oxidation step can be divided into three parts: adsorption of H₂S to create the active copper oxysulfide phase, insertion of S⁰ from oxidation of H₂S leading to the creation of polysulfide species, desorption of S_n sulfur when the polysulfide chain becomes too long. Deactivation of the catalyst can be due to the storage of sulfur in its porosity or to the impoverishment of the network oxygen. Regeneration treatment at 603 K under water leads to the removal of S_n and sulfate species.

Full text of DOI:10.1006/jcat.1999.2691

Journal of Catalysis 189, 63–69 (2000)
Article ID jcat.1999.2691, available online at http://www.idealibrary.com on
Selective Oxidation of H S over CuO/Al O : Identification and Role of the
2
2
3
Sulfurated Species formed on the Catalyst during the Reaction
�
�
,1
�
�
E. Laperdrix, G. Costentin, O. Saur, J. C. Lavalley, C. N e´ dez,† S. Savin-Poncet,‡ and J. Nougayr e` de‡
�
al
Laboratoire Catalyse et Spectrochimie, UMR 6506, Ismra, 6, boulevard du M Juin, 14050 Caen C e´ dex, France; †Procatalyse,
Usine de Salindres, 30340 Salindres, France; and ‡Elf Aquitaine Production, G. R. Lacq, BP 34, 64170 Lacq, France
Received May 14, 1999; revised September 10, 1999; accepted September 13, 1999
reaction of H2S into sulfur below the dewpoint of sulfur,
A CuO/Al2O3 catalyst was used for selective oxidation of hy- with an overall recovery of 99.9% . This process involves
drogen sulfide into sulfur at low temperature (383 K). H2S is totally two catalytic beds, the first for performing the Claus re-
consumed duringthe28first hoursofthereaction, without detection
of SO2 nor Sn. The decrease of H2S conversion is concomitant with
the detection of sulfur at the bottom of the reactor. Analysis of the
catalyst evidenced the presence of several sulfurated species, such
as sulfate, sulfur, and especially sulfide species. The sulfide/copper
sulfur (i.e., between 363 and 413 K). The sulfur produced
molar ratio varies versus reaction time and reaches values greater
than 1, which suggests the formation of polysulfide species. Its max-
imum value (3.5) is concomitant with the conversion decrease and
the formation of Sn. Complementary reactivity experiments show
action and the second for the transformation of the H2S
remaining downstream by direct oxidation. Both of these
beds operate in a cyclic mode, with the adsorption reac-
tion conducted between the dew point of water and that of
is trapped in the optimized porous network of the catalyst
carrier. Regeneration is carried out at approximately 573 K,
leading to the sulfur removal from the catalyst.
that the active phase cannot be a simple copper sulfide CuS and un-
New catalysts have been developed by Procatalyse for
derline the peculiar role of oxygen in the network. It finally seems this oxidation step, one of them being CuO/Al2O3. To op-
that the oxidation step can be divided into three parts: adsorption timise its performances in the industrial process, better un-
0
of H2S to create the active copper oxysulfide phase, insertion of S
derstanding of the involved oxidation mechanism and of
the efficiency of the regeneration step has to be achieved.
The aim of this work was to identify and study the role
of the species formed on the catalyst during the oxidative
from oxidation of H2S leading to the creation of polysulfide species,
desorption of Sn sulfur when the polysulfide chain becomes too long.
Deactivation of the catalyst can be due to the storage of sulfur in its
porosity or to the impoverishment of the network oxygen. Regen-
step of H2S into sulfur. Among them, sulfate, sulfide, and
eration treatment at 603 K under water leads to the removal of Sn
sulfur species were expected. It was therefore necessary
and sulfate species.
�c 2000 Academic Press
first to define the best analytical methods for detecting and
to quantifying them. Capillary electrophoresis was used to
analyze sulfate, thiosulfate, and sulfite species. Attempts
to use IR spectroscopy to differentiate copper, alumina,
INTRODUCTION
The Claus process (1, 2) is the most commonly employed and mixed sulfate species (5) were unsuccessful, due to
technique for removing sulfur from natural gas, crude oil, or the lack of transparency of the IR beam. S was quanti-
n
industrial and combustion streams. However, the catalytic fied by UV spectrometry, after extraction in chloroform.
reaction SO2 + 2 H2S (+ 3/n Sn + 2 H2O being an equilib- The total amount of sulfide species was determined by
rium, the best sulfur recoveryefficiencyofClausplantsunits complexometry method (6). Attempts to characterize poly-
does not exceed 97–98% (3). The remaining H2S has to be sulfide species by Raman technique, according to Perrin
eliminated and one of the most promising possibilities is et al. (7), failed because of the instability of the samples
the direct H2S oxidation to elemental sulfur:
under the laser beam. Yet, their presence was evidenced by
calculation of sulfide/copper molar ratio. Sulfur balances
were obtained by comparison between total sulfur esti-
mated from addition of these analysis results and total sul-
fur quantified by elemental microanalysis. The variation of
these several species amount has been followed with time
on stream, in order to establish reaction mechanisms. The
role of sulfide and sulfate species has been particularly
studied.
H2S + 1/2 O2 → 1/n Sn + H2O.
Recently, Elf and Lurgi together developed a tail gas pro-
cess, called Doxosulfreen (4), based on the direct oxidation
1
To whom correspondence should be addressed at Laboratoire de
R e´ activit e´ de Surface, Universit e´ Paris VI, 4 pl. Jussieu, Tour 54/55, 75252
Paris C e´ dex 05. E-mail: costenti@ccr.jussieu.fr.
63
0021-9517/00 $35.00
Copyright �c 2000 by Academic Press
All rights of reproduction in any form reserved.

6
4
LAPERDRIX ET AL.
EXPERIMENTAL
before each analysis. It was prepared from a mere solution
of sodium chromate (0.1 M), obtained from dissolution of
Na2CrO4, 4H2O into ultrapurified water; 5 mL of this solu-
tion was added to 2.5 mL of a commercial solution of OFM
The CuO/Al2O3 catalyst was provided by Procatalyse:
its copper content was 17.3 (wt% ) and its specific area
2
� 1
1
36 m g .
(
Osmotic Flow Modifier) which is specific for anionic anal-
ysis. The pH of the so-obtained solution was higher than
. After adding ultrapurified water until 100 mL, the elec-
The catalytic tests were carried out in a fixed bed contin-
uous flow quartz microreactor under atmospheric pressure,
at 383 K. The reactor was heated externally with a furnace
and the temperature of the catalyst bed was controlled by a
thermocouple being localized close to the wall of the quartz
tube. For each experiment, 4 g of catalyst was heated up to
9
trolyte was evacuated for 15 min. The detection mode was
an inverted UV spectrometer, which was equipped with a
mercury lamp, at a wavenumber of 285 nm. The samples
were diluted in a NaOH solution (pH 9), stirred at 323 K
for 10 min, and filtered. It was checked by a subsequent
analysis that all of the sulfur has passed into the filtrate.
3
83 K in the stream of reactants. The feed was a mixture
of H2S/O2/H2O = 0.9/2/30 diluted in N2. The total flow rate
�
1
was 53.5 mL/min (W/F = 28 g � h mol ). For regeneration,
the CuO/Al2O3 catalyst was heated up to 603 K under a
mixture of nitrogen and water vapor (� 25% ).
Sulfide species titration. The method used is related in
ISO 4284 (6), for spaths fluor sulfide analysis. Its principle
is first to dissociate the M–S bonds by an acidic attack: a
mixture of boric and chlorhydric acids (respectively 500 mg
At the end of each experiment, the catalyst was cooled
under nitrogen flow and was then kept in a glove box under
dry nitrogen for subsequent analysis.
�
1
and 50 ml of a 400 g L solution) with tin chloride (10 ml of
�
1
a 200 g L solution) was added to 500 mg of sample previ-
ously finely ground. The H2S evolved being, under nitrogen
flow, trapped in a cadmium acetate solution, leading to CdS
cadmium sulfide formation. The reaction of this latter so-
lution with 10 mL of an acidic iodine solution (0.0086 N)
liberates H2S, which reacts with I2 stoichiometrically. Titra-
tion of the iodine excess by a sodium thiosulfate solution
The quartz or Teflon tubes of the catalytic test were
heated to avoid water or sulfur condensation. Water and
sulfur produced by the reaction were trapped at room tem-
perature; remaining traces of water were condensed by
magnesium perchlorate. Gaseous products were analyzed
by gas chromatography (Varian 3400 CPG equipped with a
FPD detector using a 25-m Porapak Q column). The con-
version of hydrogen sulfide and the sulfur selectivity were
calculated using the following equations:
(
0.01 N) allows one access to the H2S evolved and so to the
amount of sulfide species initially present.
Sn species titration. Samples 500 mg were ground in an
agate mortar and then added to 50 mL of HCCl3 (Prolabo).
The obtained mixture was stirred and heated to 323 K for
conversion ꢀ% )
=
(mol H S reacted)/(mol H S fed) � 100%
2
2
5
h. After two successive washings, the mixture was fil-
sulfur selectivity ꢀ% )
tered, transformed into methylene blue, and analyzed using
a Beckman UV spectrometer, equipped with a deuterium
lamp, the maximum Sn absorbance being above 280 nm.
=
(mol H S reacted � mol SO2 formed)/
2
(
mol H S reacted) � 100% .
2
XRD characterization. X-ray diffractograms were reg-
Analytical Methods
�
�
istered in the 5–70 2� range on a Philips PW 1800 diffrac-
Total amount of sulfur. After each experiment, the sul- tometer (scan 0.02 , accumulation time 2 s) equipped with
fur content of all the sulfurated species forms was system- a Cu anticathod (� = 1.5406 A˚ ).
atically quantified by the Vernaison CNRS microanalysis
center. Burning the samples at 2073 K under oxygen flow
transformed the whole sulfurated species into SO2, which
was further titrated using IR spectroscopy.
RESULTS AND DISCUSSION
CuO/Al2O3 Catalyst, Adsorption Step
Sulfate, thiosulfate, and sulfite species titration. Sulfate,
A first experiment has been performed for 50 h on the
thiosulfate, and sulfite species analysis was performed using CuO/Al2O3 catalyst according to the above-described ox-
a lone technique, the capillary electrophoresis, which made idation conditions. The evolution of the H2S consumption
it possible to separate ionized compounds in relation with is reported in Fig. 1. During the first 28 h, H2S is totally
their own mobility under an electric potential (4). The sys- consumed; then the conversion progressively decreases to
tem used (Waters) was equipped with a capillary (Supelco, reach 64% after 50 h. In any case, no SO2 has been detected.
7
5 �m � 60 cm) made of fused silica. The applied potential
Moreover, sulfur formation is observed after 28 h: it ap-
was 20 kV, with a current intensity around 18 �A. The injec- pears at the cool bottom of the reactor, since the water
tion was performed by hydrostatic mode, with an injection trapped is no longer clear. The decrease of the H2S conver-
time of 30 s. Analyses were realized at room temperature. sion into sulfur via the selective oxidation reaction is con-
Due to its instability, the electrolyte had to be prepared just comitant with the formation of sulfur. It could be noticed

SELECTIVE OXIDATION OF H2S OVER CuO/Al2O3
65
also reported in Fig. 2. One can actually define a second
sulfur balance, taking into account in one hand the H2S
inlet starting from the beginning of the experiment and in
the other hand the amount of total sulfur and especially
the quantity of sulfide species formed. This balance is very
good up to 28 h (see Fig. 2). The subsequent deficit (after
2
8 h) is related to the Sn trapped at the bottom of the
reactor, whose amount can be estimated to 14% (grams of
S per gram of catalyst) after 50 h of experiment.
Finally, the main result is relative to the evidence of sul-
fide species formation. In order to get information about
their structure, the molar ratio between sulfide and cop-
per has been calculated and reported in Fig. 2. It shows
that sulfide species are progressively enriched in sulfur, to
reach a maximum of 3.5 mol of sulfide per mol of copper
at 28 h of adsorption time. This value clearly indicates that
there is formation of polysulfide species. Once a maximum
of polysulfide condensation is reached, elemental sulfur Sn
appears and so the S/Cu ratio decreases. One must thus
conclude that the complexometry method used to analyze
sulfide species also titrates polysulfide species. Since metal
polysulfides are not commercially available, it was not pos-
sible to check this point. However, it is well known that cop-
per is likely to form polysulfide chains (9). X-ray diffraction
patterns (see Fig. 3) have been performed on the fresh cata-
lyst, after 28 h of experiment, and at the end of the reaction
FIG. 1. H2S conversion (% ) versus time on stream (h).
that its coloration has passed from gray before the experi-
ment to black after the reaction. This observation suggests
that sulfidation phenomena have occurred during the re-
action. Analysis results show that the total amount of sul-
fate, thiosulfate, and sulfite species remains low (� 3.4% ).
Content of Sn is also very low (6% ), since the major part of
sulfurated species (24.4% , i.e., more than 70% of the whole
sulfurated species) are sulfide ones. Thus, the sulfur balance,
defined as the ratio between the addition of these results
and the elemental sulfur amount (34.5% ), leads to a ratio
of 0.98. It can thus be concluded that the analytical meth-
ods used are convenient for the titration of the sulfurated
species on CuO/Al2O3 after reaction.
(
50 h on stream). The pattern of fresh catalyst evidences the
presence ofthe two main characteristiclinesofcopper oxide
(
10). Their intensity decreases with time on flow and only a
In order to follow the variation of the amount of sulfided
species formed with time on stream, several experiments
were performed and stopped at different times on stream.
Corresponding analyses were performed. As shown in
Fig. 2, the amount of sulfide species increases during the
first hours of the reaction to reach a maximum value near
2
8 h of reaction (30 wt% of S per 100 g of catalyst). It is
further stabilized around 25% . Moreover, no Sn is detected
in the catalyst before 28 h of reaction, showing that the
sulfur appearance visualized at the bottom of the reactor
or in water corresponds to the beginning of its formation.
After 28 h of time on stream, the difference between the
amount of total sulfur and that of sulfide species is due to
the formation of a low amount of sulfate species and to
the formation of Sn remaining on the catalyst, as shown
by results after 50 h on stream. We can therefore suspect
a link between the sulfur storage and the beginning of
deactivation, as already mentioned in the literature (8).
Actually, the sulfate amount is too low to provoke catalyst
deactivation and we have checked that the H2S selective
oxidation effectively occurs on a presulfated catalyst. The
FIG. 2. Total sulfur and sulfide amount (wt% ) on the CuO/Al2O3
catalyst compared to S stream (corresponding to total H2S passed over
total amount of sulfur (as H2S) passed over the catalyst is the catalyst). S/Cu, sulfide/copper molar ratio.

6
6
LAPERDRIX ET AL.
FIG. 3. XRD spectra of CuO/Al2O3: (a) fresh, (b) after 28 h of time on stream, (c) after 50 h of time on stream.
very small amount of copper oxide still remains on the cata- viously described. H2S is first totally consumed up to 45 min,
lyst at the end of the experiment. The middle-reaction cata- before being progressively released. The total equivalent
�
1
lyst pattern presents several lines, which can be attributed consumption of H2S corresponds to 0.003 mol g of sulfur,
to different sulfided phases such as CuS, Cu2� xS, �-Cu2S, which is 10 times lower than that needed to reach the max-
and CuS2. However, two unknown lines absent from the imum polysulfide condensation in the previous experiment
�
� 1
two other diffractograms arise (2� = 16.97 and 20.39 ). This (0.034 mol g of sulfur after 28 h). The obtained S/Cu ratio
corresponding transient phase could be relative to copper (0.28) and the absence of oxygen in the flow suggest that
polysulfide not referenced in literature, since it was already formation of a sulfide phase such as CuSxOy.
noticed that the amount of sulfided species reaches a max-
Starting from this presulfided catalyst, the oxidation re-
imum value precisely at 28 h of reaction. At the end of the action was then performed under the same oxidative con-
experiment, the copper oxide phase remains very minor. ditions (0.9% H2S and 2% O2) as for the first experiment.
Several sulfided phases still remain, since lines relative to H2S is first totally consumed, without any formation of el-
Sn have grown, which is in agreement with the analytical emental sulfur. After 22 h, sulfur appears and H2S is par-
results obtained on the catalyst and previously detailed.
tially released: this corresponds to a new consumption of
�
1
The sulfur formation mode could consist first of sulfida- 0.026 mol g of H2S. The H2S consumption and the equiv-
0
tion ofsome copper sitesand then S formation through H2S alent S/Cu ratio are reported in Table 1. It clearly appears
0
oxidation, followed by S incorporation, creating a polysul- that the molar ratio S/Cu is above 1 in both experiments,
fide chain. When this chain reaches a critical length, Sn is
desorbed. In this feature, a copper sulfide phase should be
efficient to transform H2S into Sn. Two kinds of experiments
have been performed, starting from a CuO/Al2O3 presul-
fided sample or from a CuS/Al2O3 sample.
TABLE 1
S/Cu Molar Ratio at H2S Release Time
� 1
H2S consumption (mol g
)
S/Cu
3.15
CuO/Al2O3, first experiment
Comparison with Oxidation Performed on a CuO/Al2O3
Presulfided Sample
Oxidative treatment
CuO/Al2O3, second experiment
Presulfidation step
0.034
0.003
0.026
0.28
2.41
The CuO/Al2O3 catalyst was first pretreated under H2S
flow (2.2% in N2) without oxygen, under the conditions pre-
Oxidative treatment

SELECTIVE OXIDATION OF H2S OVER CuO/Al2O3
67
reaching 3.15 in the first one and 2.69 in the second. These
results confirm that a high S/Cu ratio has to be reached
before the catalyst produces sulfur. Moreover, the second
experiment clearly shows that the presence of oxygen in the
flow is necessary for the consumption of a large amount of
H2S without formation of Sn. Indeed, oxygen is consumed
0
to oxidize the sulfur present in H2S into sulfur S which
will be incorporated into the polysulfide chain. This exper-
iment confirms the formation of polysulfides as intermedi-
ate toward sulfur Sn formation. The mechanism could thus
involve, first, an adsorption step leading to a partial sulfi-
dation of the catalyst and, second, an oxidation step which
leads to polysulfide formation. Sn production could finally
result from the breaking of the (–S–)x chains.
A second experiment has been performed starting from
a commercial CuS/Al2O3 catalyst (S/Cu = 1, wt% of Cu =
1
0.4) in order to determine the role of the oxygen atoms ini-
FIG. 4. Influence of the O2/H2S molar ratio on H2S conversion (% )
tially present in CuO/Al2O3. According to the above pro- versus time (h). The arrow corresponds to the changing of the O2/H2S
ratio.
posed mechanism, we expect that this CuS/Al2O3 catalyst
presents a good activity toward H2S oxidation and behaves
exactly in the same way as CuO/Al2O3, without of course fore changing the ratio to 5% after). But it is important to
the first adsorption step. The H2S release time has been note that, as observed in the case of a pre-sulfated catalyst,
compared for the both catalysts. In the case of CuO/Al2O3, the increase of sulfate amount does not affect the perfor-
�
1
H2S release time is reached once 8500 �mol g of H2S mance of the catalyst. This ability to maintain good catalytic
has been consumed. In the case of CuS/Al2O3, the very performances in spite of the presence of sulfates, in partic-
�
1
short H2S release time corresponds to only 420 �mol g of ular an excellent selectivity, constitutes the uniqueness of
�
1
H2S (equivalent to 700 �mol g for a CuS/Al2O3 sample the catalyst, contrary to many other catalysts previously
containing 17.4 wt% of copper). The comparison between studied, especially NiO/Al2O3 (11, 12). This suggests that
�
1
the two values (700 and 8500 �mol g ) clearly shows that the mechanism of H2S selective oxidation and particularly
CuS/Al2O3 leads to the formation of fewer condensed poly- the reactive species might be thus different in the case of
sulfides than CuO/Al2O3 and is therefore less effective in CuO/Al2O3 relative to that of NiO/Al2O3.
H2S transformation. This experiment emphasizes the pecu-
The promoting role of oxygen could be explained involv-
liar importance of the network oxygen of the CuO phase. ing polysulfide reactivity, as reported in the literature (13).
The active phase is thus not simply a single copper sulfide, Actually, if one considers that the reaction of H2S oxidation
but a copper oxy-sulfide. H2S release time could thus de- is ruled by a radical mechanism (14), oxygen would pref-
pend on the oxygen storage available in the network.
erentially react with the radicals present at the end of the
�
Besides, we have shown that the ratio O2/H2S = 2.2 is sulfur chain, to create Sx–S–O–O species extremely reac-
not a limiting factor since its increase promotes the recov- tive with H2S present in the flow. When the O2/H2S ratio
ery of the activity after 40 h of reaction, as shown by the is increased, the excess of oxygen could weaken the –S–S–
following experiment: the experiment was first performed bond, resulting in a cleavage of this latter, as previously
under the conditions previously described, till a significant explained by Steijns (15) and Dudzik (16). This cleavage
deactivation was observed (84% of conversion after 38 h could induce Sn and SO2 formation and liberate an active
of experiment). Then the reactor was bypassed and some copper site, boosting then the H2S conversion. Neverthe-
of the catalyst was taken for analysis. The O2/H2S ratio was less, the latter interpretation is not convenient for two main
then increased from 2.2 to 3, and the stream led once again reasons: (i) SO2 is not observed even when the gaseous oxy-
to the reactor. As shown in Fig. 4, the increase of oxygen gen amount is increased, and (ii) it does not explain the
rate in the inlet flow leads to an increase of H2S conversion importance of oxygen contained in the oxy-sulfide active
(
around +6% ). The analysis of the samples at 38 h of ex- phase. One could also consider, as suggested by Zhenglu
periment and at the end of the process show an increase of et al. (17), that a Mars and Van Krevelen mechanism (18) is
% of sulfide and polysulfide species. It therefore appears involved in the H2S-selective oxidation reaction. The oxy-
5
that O2 increase enhances the reaction rate. The conversion gen of the active phase CuSxOy network would allow sulfur
finally decreases because of the loss of active phase due to species oxidation (and H2O formation) and the subsequent
Sn present in the pores. As expected, the increase of O2/H2S polysulfide formation; it would be regenerated by gaseous
ratio enhances the sulfate species formation (from 3% be- oxygen. Consequently, the observed deactivation could be

6
8
LAPERDRIX ET AL.
TABLE 2
2 3
promoted alumina such as CuO/Al O . This study showed
that these species are partially reducible by H2S as low as
73 K (19) but that water alone has no effect on their stabil-
ity. Since the regeneration flows contain only N2 and H2O,
Concentrations of Different Sulfurated Species (wt%) in the
Catalyst Pretreated in Different Ways
5
Sulfide Sulfate Elemental sulfur the decrease of sulfate species amount is certainly due to
their reduction by H2S formed by a reverse Claus reac-
After oxidation reaction
After 3 h of regeneration treatment
After 14 h of regeneration treatment
24.5
4.8
3.5
1.8
0.0
6.0
0.0
0.0
�
1
tion. Indeed, emissions of SO2 (920 �mol g ) and H2S
�
1
(
400 �mol g ) during the regeneration step were observed
3.9
by CPG, as described in Fig. 5. The experimental H2S/SO2
ratio is different from the stoichiometry of a reverse Claus
reaction (0.4 versus 2). This discrepancy could be attributed
to H2S consumption due to sulfate reduction.
partly due to the progressive decrease of oxygen content in
the oxy-polysulfide active phase, as shown by the decreas-
ing intensities of CuO diffraction lines versus time (Fig. 3).
This hypothesis could also explain the low performances
of CuS/Al2O3, which contains much fewer oxygen atoms.
Increasing O2/H2S ratio may enhance gaseous oxygen in-
corporation in the oxy-polysulfide phase and thus partly
inhibit the deactivation.
CONCLUSION
CuO/Al2O3 catalyst was shown to be very selective for
the H2S selective oxidation reaction, since SO2 has never
been detected. Results imply that the formation of sulfide
and even polysulfide species on the surface of the catalyst
is a result of simple sulfur adsorption and the oxidation
Regeneration Step
�
II
0
of S to S , respectively. Elemental sulfur Sn appears af-
ter several hours, resulting from the cleavage of too-long
polysulfides chains. Concomitantly, H2S is no longer totally
consumed, because of the loss of active surface due to sul-
fur accumulation inside the alumina pores. Oxygen plays
a fundamental role in the formation of polysulfide species
and in the enhancement of the reaction.
Because of the unavoidable decrease of conversion,
mainly due to sulfur Sn storage in the pores, the catalyst has
to be regenerated regularly in order to eliminate elemen-
tal sulfur or even sulfate species. Table 2 reports analysis
results gathered from a CuO/Al2O3 catalyst after an oxida-
tion step and duringthe regeneration step. The totalamount
of sulfur present on the regenerated samples is much lower
than that on the nonregenerated one. In particular, no more
Sn elemental sulfur remains on samples, which is expected
since the regeneration temperature is much higher than the
sulfur dewpoint. Most of sulfide species are removed, since
only 3.9% remains after regeneration (versus 24.4% after
the oxidation step). Sulfate species are also eliminated. We
have already studied the reduction of sulfate species on
The remarkable characteristics of CuO/Al2O3 presently
studied are its poor ability to form sulfate species and its ex-
cellent selectivity in sulfur, even with an excess of oxygen.
Moreover, sulfates are not poisons of the selective H2S ox-
idation. This constitutes, with the formation of polysulfide
chains, a second uniqueness of this catalyst.
ACKNOWLEDGMENTS
The authors are grateful to Mr. Nwal-Amang (IUT Caen) and Mr.
Belleil (Rhodia) for technical support. E. Laperdrix thanks ADEME for
financial support.
REFERENCES
1
2
3
4
5
. Claus, C. F., British patent 5958 (1883).
. Behr, H., Refin. Nat. Gasoline Manuf. 17, 237 (1938).
. Patents WO 97 17 283 and WO 97 17 284.
. Kuhr, W. G., Anal. Chem. 62, 403R-414R (1991).
. Waqif, M., Saur, O., Lavalley, J. C., Perathoner, S., and Centi, G.,
J. Phys. Chem. 95, 4051 (1991).
6. “Spaths Fluor Pour la Fabrication d’Acide Fluorhydrique—Dosage
desSulfures—M e´ thode Iodom e´ trique,” InternationalNorm ISO 4284,
1st ed., 1978.
7. Perrin, C., Perrin, A., and Prigent, J., Bull. Soc. Chem. 8 (1972).
8. Schoofs, G. R., “Hydrocarbon Processing,” 71, February 1985.
9. Isabeg, J., and Pascal, P., in “Nouveau Trait e´ de Chimie Min e´ rale”
(
Masson and Cie, Eds.), Vol. III, p. 183. Paris, 1957.
FIG. 5. Emissions of SO2 and H2S during the regeneration by
N2 + H2O and corresponding temperature profile.
10. Strohmeier, B. R., Leyden, D. E., Scott Field, R., and Hercules, D. M.,
J. Catal. 94, 514 (1985).

SELECTIVE OXIDATION OF H2S OVER CuO/Al2O3
69
1
1
1. Laperdrix, E., DEA, Caen, France, 1996.
2. Legendre, O., “H2S Selective Oxidation on a Ni Promoted Cata-
lyst,” Internal Report. Rhodia-ISMRa-E.lf Aquitaine Production,
15. Steijns, M., Koopman, P., Nieuwehuijse, N., and Mars, P., J. Catal. 42,
96 (1976).
16. Dudzik, Z., and Bilska-Ziolek, M., Bull. Acad. Polon. Sci. Ser. Sci.
Chim. 23, 699 (1975).
1
994.
1
1
3. Van den Brink, P. J., “The Selective Oxidation of Hydrogen Sul- 17. Zhenglu, P., Weng, H. S., Han-Yu, F., and Smith, J. M., AIChE J. 30,
fide to Elemental Sulfur on Supported Iron-Based Catalysts.” Thesis,
Utrecht, 1992.
1021 (1984).
18. Mars, P., and Van Krevelen, D. W., Chem. Eng. Sci. 3, 41 (1954).
4. Steijns, M., Koopman, P., Nieuwehuijse, N., and Mars, P., J. Catal. 42, 19. Laperdrix, E., Sahibed-Dine, A., Costentin, G., Bensitel, M., Saur, O.,
7 (1976). and Lavalley, J. C., submitted for publication.
8