Applied Catalysis A: General 397 (2011) 218–224

Contents lists available at ScienceDirect

Applied Catalysis A: General

journal homepage: www.elsevier.com/locate/apcata

Methyl mercaptan and carbonyl sulfide traces removal through adsorption

and catalysis on zeolites and layered double hydroxides
Andrey Ryzhikov a , Vasile Hulea a , Didier Tichit a , Catherine Leroi b , Didier Anglerot b ,
Bernard Coq a , Philippe Trens a,∗
a
Institut Charles Gerhard de Montpellier, UMR 5253 CNRS/ENSCM/UM2/UM1, “Matériaux Avancés pour la Catalyse et la Santé”,
8 rue de l’Ecole Normale, 34296 Montpellier Cedex 5, France
b
TOTAL SA, Pôle Recherche et Développement Mont/Lacq, BP 47 – 64170, Lacq, France

a r t i c l e i n f o a b s t r a c t

Article history: In this paper, the adsorption of methyl mercaptan and carbonyl sulfide by metal exchanged zeolites and
Received 25 January 2011 layered double hydroxides (LDH) is studied. The affinity of CH3 SH for metal-exchanged zeolites was found
Received in revised form 25 February 2011 to increase with the basicity and the valence of the cation. Yoon–Nelson modelling gave good predictions
Accepted 2 March 2011
of half time of breakthrough for each zeolite tested. Thermogravimetry was used in the case of the system
Available online 10 March 2011
CH3 SH/NaX for elucidating the desorption process of CH3 SH from NaX. In contrast, the affinity of COS
was found clearly lower for Me-zeolite in comparison with CH3 SH. Catalytic hydrolysis of COS was tested
Keywords:
with the same metal exchanged zeolites and LDH materials and the basicity of the catalysts was found
Methyl mercaptan
Carbonyl sulfide
to strongly influence the hydrolysis.
Zeolites © 2011 Elsevier B.V. All rights reserved.
Layered double Hydroxides
Adsorption
Hydrolysis

1. Introduction and desorption at 200–350 ◦ C. Active carbons exhibit larger adsorp-

tion capacities than zeolites, but they are much more sensitive to
Natural gas always contains contaminants or other unaccept- the regeneration process, which is more efficiently withstood by
able substances including carbon dioxide, hydrogen sulfide, water, zeolites. On the other hand, most of the adsorption studies on zeo-
carbonyl sulfide, mercaptans, heavy hydrocarbons and mercury lites deal with ethylmercaptans [3–45]. There is no dedicated study
[1]. Conditioning natural gas for marketing generally requires the on the more volatile methylmercaptan capture in a methane feed.
removal of these contaminants, which comprises several main Finally, upon desorption operation, the mercaptan concentrated
operations including deacidizing (CO2 and H2 S removal, partial flue gas needs then to be processed by additional desulfurization
extraction of methyl-, ethyl- and propylmercaptan), dehydration, treatment (vide infra).
gasoline extraction. The deacidized gas usually contains 2% CO2 , Deep mercaptans removal through chemical processes may
4 ppm H2 S and 20–50 ppm total sulfur (mercaptans and COS). In include oxidation [10], alkaline treatment [11], reaction with
view of both increasingly stringent environmental regulations and olefins [12] and hydrodesulfurization [13]. These processes need
technical requirements for further catalytic processes (e.g., syn- additional reactants (O2 , H2 O2 , H2 , olefins, . . .) and some of them
gas fabrication), deeper desulfurization becomes in high demand. produce dialkyl sulfides or disulfides. MEROX process is the most
Moreover, it is desirable for domestic users of natural gas to have popular of these processes: mercaptans are transformed by caustic
concentrations of mercaptans lowered to 2–5 ppmv. Modern pro- washing into mercaptides, which in turn are oxidized to dialkyld-
cesses for the deep removal of mercaptans and COS are generally isulfides with soda regeneration. The presence of COS reacting
based on fixed bed adsorption and chemical reactions. irreversibly with soda leads to a high base consumption; the
The most efficient adsorbents for mercaptans capture are metal- MEROX process was initially designed for fuel sweetening, but it
promoted zeolites [2–6] and active carbon [7–9]. The method used may also be applied for natural gas processing [14]. Obviously, these
is the thermal swing adsorption (TSA), with adsorption at 20–60 ◦ C chemical treatments can be coupled with the treatment of mercap-
tans concentrated flue gas issued from the regeneration stream in
the TSA process (vide supra).
∗ Corresponding author. Regarding COS removal the main purifying treatment is the
E-mail address: [email protected] (P. Trens). catalytic hydrolysis leading to CO2 and H2 S, which are in turn elim-

0926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2011.03.002
 A. Ryzhikov et al. / Applied Catalysis A: General 397 (2011) 218–224 219

inated through classical deacidizing processes. Usually the COS D8 diffractometer with CuK␣ radiation. The specific surface areas
hydrolysis is performed at temperature higher than 100 ◦ C, but were evaluated by the BET method, using conventional N2 adsorp-
recent studies showed that the reaction can be effectively achieved tion/desorption procedure at −196 ◦ C (Micromeritics ASAP 2010).
under milder reaction conditions [15–17]. To date, a number of dif- The samples were outgassed at 150 ◦ C for 8 h before the N2 adsorp-
ferent types of catalysts have been described in the patent and open tion measurement. Thermogravimetric experiments (SETARAM
literature for the COS hydrolysis. Most of them are based on alumina TG-DSC-111) were carried out on the metal-exchanged zeolites in
and titania [16–23]. In order to improve the catalytic activity, alu- order to determine both the water weight loss in the course of the
mina was doped with different promoters: alkali metals [16,24], activation step (5 K min−1 ) and the amount of sulfur compounds
alkaline earth metals, transition metal oxides (Fe, Zn, Ni, Cu, Co) evolved during their desorption.
[16,15,25] and rare earth metal oxides [26]. The adsorption of COS
is much less documented. The low polarity of COS makes its diffi- 2.3. Experimental set up
cult to adsorb on most adsorbents. In the open literature, only few
materials, including supported rhodium [27], activated carbon [28] The adsorption and catalytic experiments were carried out in
and layered double hydroxides [29] have been reported as possible a fixed-bed steel reactor, at atmospheric pressure, using 100 mg
sorbent for COS. of adsorbent/catalyst (particles of 0.15–0.35 mm). Before their use
On that account, there is the need for basic knowledge about the the materials were treated into the reactor at 300 ◦ C for 30 min
capture and elimination of CH3 SH and COS traces in CH4 , in relation under a flow of nitrogen. Adsorption experiments were carried
with adsorption and heterogeneous catalytic processes. out at 25 ◦ C using a feed stream with a composition of 0.05%
With this objective in view, we have started a research project on CH3 SH or COS in CH4 . The total flow rate of the gas stream was
a finishing treatment of natural gas containing methyl mercaptan 60 mL min−1 (36,000 mL h−1 g−1 ) for CH3 SH and 20 mL min−1 (12,
and COS. In a first approach we have: (1) revisited the behavior 000 mL h−1 g−1 ) for COS. COS hydrolysis was performed in the tem-
of some metal-exchanged zeolites, mixed oxides elaborated from perature range of 50–200 ◦ C, using a reaction mixture COS/H2 O/CH4
layered double hydroxides (LDH) as adsorbents; (2) evaluated the (0.05/2.3/97.5, v/v/v), obtained by passing a COS/CH4 gas stream
catalytic properties of basic metal-zeolite and mixed oxides for COS (20 mL min−1 ) through a water saturation evaporator kept at 20 ◦ C.
hydrolysis. The contact time defined as the ratio of catalyst weight to total
flow rate was 0.3 g s mL−1 . For both adsorption and catalytic mea-
surements, the composition of the reactor effluent was monitored
2. Experimental
on-line by a gas chromatograph (Shimadzu GC 2014) equipped with
a flame photometric detector.
2.1. Materials
2.4. Breakthrough modelling
Exchanged zeolites: Metal exchanged zeolites (Me-zeolites) NaX,
CsNaX, CaX, MgNaX, BaX, ZnNaX, NiNaX, NaY, CsY and NiY
Many models have been used for the prediction of breakthrough
were prepared on the base of commercial NaX (Si/Al ratio = 1.2,
time. Some of these can be found in references [30,31]. These have
BET surface = 770 m2 g−1 ) and NH4 Y (Si/Al ratio = 2.6, BET sur-
been developed for particular situations, such as the determina-
face = 780 m2 g−1 ) provided by Sigma–Aldrich and Zeolyst, respec-
tion of the initial part of the breakthrough curves [32]. Some others
tively. The ion exchange of NaX and NH4 Y was realized in 0.5 M
can be used with the requirement that particular assumptions have
solution of the corresponding nitrate (Ca(NO3 )2 ·4H2 O, Ba(NO3 )2 ,
to be fulfilled, for instance negligible diffusion limitations [33,34].
Mg(NO3 )2 ·6H2 O, Zn(NO3 )2 ·6H2 O, Ni(NO3 )2 ·6H2 O, CsNO3 ) with a
Another useful model has been proposed by Yoon and Nelson [35]
liquid-solid ratio of 50 mL g−1 . After stirring at 80 ◦ C for 3 h, the
taking into account the adsorption/desorption equilibrium rate as
exchanged zeolites were recovered and dried at 80 ◦ C for 12 h. This
well as the sorbate breakthrough which is very convenient in view
exchange procedure was repeated twice.
of quantifying the dynamic sorption properties of our adsorbents.
Layered double hydroxides: NO3 -MII /Al LDHs (MII = Mg, Ni or
The modelling equation they developed can be written as follows:
Zn) were synthesized by a conventional co-precipitation method
 C

under ambient atmosphere. 0.06 mol MII nitrate salt and 0.03 mol
ln = kYN t − t0.5 kYN (1)
Al(NO3 )3 ·9H2 O (MII /Al3+ molar ratio of 2) were dissolved in 200 mL Co − C
deionized water. This solution was delivered by a peristaltic pump
where kYN represents the Yoon–Nelson rate constant in min−1 and
(0.8 mL min−1 ) into a beaker at room temperature, under vigor-
t0.5 the time for 50% adsorbate breakthrough in min. This model
ous magnetic stirring. Simultaneously, a 2 M NaOH solution was
can therefore yield the time for 50% sorbate breakthrough which
added at a controlled rate to maintain the pH close to 8 using a pH-
is very convenient to evaluate the efficiency of various adsorbents
STAT Titrino (Metrohm) apparatus. After complete precipitation,
as well as their interaction distribution using the first derivative of
the slurry was refluxed at 80 ◦ C for 12 h. The solid was recovered
the breakthrough curve at t1/2 . The fit between experimental data
by centrifugation, washed several times with water and finally
and the linearized form of Eq. (1) was found by determining the
dried overnight at 80 ◦ C. The CO3 -MII /Al LDHs were obtained by ion
best linear regression coefficients (r2 ) obtained with different cou-
exchange of the previous NO3 -MII /Al LDHs precursors with excess
ples (kYN , t1/2 ). Calculated t1/2 and kYN were obtained from this fit.
Na2 CO3 solution, at room temperature for 2 h.
Experimental t1/2 were determined by interpolation of the different
Mixed oxides: CO3 -MII /Al LDHs and NO3 -MII /Al LDHs samples
breakthrough curves at C/Co = 0.5.
were calcinated under air flow at 450 ◦ C for 4 h in order to obtain
MII -Al mixed oxides, which were used as catalysts for the COS
3. Results and discussion
hydrolysis.
3.1. Adsorption on Me-zeolites
2.2. Characterization
Adsorption of CH3 SH: The choice of the metals in Me-zeolite was
The composition of the materials was determined by chemi- motivated by modulation of basicity (Na, Cs, Ca, Ba, Mg) and because
cal analysis (Service Central du CNRS, Vernaison, France). Powder Zn and Ni exhibit high affinity to sulfur compounds [36]. The metal
X-ray diffraction (PXRD) patterns were collected on a Bruker AXS content of the metal-exchanged zeolite is given in Table 1.
220 A. Ryzhikov et al. / Applied Catalysis A: General 397 (2011) 218–224

Table 1
Cation composition, CH3 SH or COS uptakes in adsorption experiments at 298 K.

Sample Cation (wt%) Cation (mol g−1 ) × 103 CH3 SH uptake Molar ratio COS uptake (mg g−1 )
−1 3 −1
Na Me Na Me (mg g ) 10 (mol g ) CH3 SH/Na CH3 SH/Me

NaX 14.7 – 6.4 – 183 3.80 0.59 – 30.0

CsNaX 5.1 31.9 2.2 2.4 66 1.38 0.20 0.10 11.0
MgNaX 4.5 5.8 2.0 2.4 132 2.75 0.18 0.46 4.5
CaX – 12.8 – 3.2 140 2.92 – 0.91 55.0
BaX – 33.5 – 2.5 111 2.31 – 0.94 55.0
NiNaX 3.0 13.9 1.3 2.4 88 1.83 0.12 0.38 7.5
ZnNaX 1.9 16.7 0.8 2.6 115 2.40 0.08 0.64 4.5
NaY 9.3 – 4.2 – 78 1.62 0.38 – –
CsY – 23.9 – 1.8 69 1.43 – 0.79 –
NiY – 12.1 – 2.1 62 1.29 – 0.63 –

Note that there was no significant change either in the specific Figs. 1 and 2 show the breakthrough curves and the mass
surface area or the XRD pattern upon cation exchange. Thus, for uptake plots of CH3 SH for the various adsorbents along with
all Me-zeolites the specific surface area was 770 ± 25 m2 g−1 . XRD the Yoon–Nelson modelling for each breakthrough curve.
patterns confirm that both parent and exchanged zeolites con- These exhibit different shapes that may be explained by dif-
sist of crystalline phase with X- or Y-FAU topology [37]. Prior to ferent affinity of CH3 SH molecules for the various cations.
adsorption experiments, the materials were activated in situ at The saturation adsorption capacities of the metal-exchanged
300 ◦ C in N2 flow for 30 min. The water weight loss during the zeolites for CH3 SH are also presented in Table 1 (with
activation process was 28%, 22%, 23%, 23%, 19%, 21%, 25%, 18%, respect to the dry materials), they decrease in the sequence:
15% and 16%, for NaX, CsNaX, CaX, MgNaX, BaX, ZnNaX, NiNaX, NaX > CaX > MgNaX > ZnNaX > BaX > NiNaX > NaY > CsY > CsNaX > NiY.
NaY, CsY and NiY, respectively. After cooling at 25 ◦ C, the gas The Yoon–Nelson modelling of the breakthrough curves is pre-
mixture CH3 SH/CH4 (0.05/99.95, v/v) was fed to the reactor at sented in Table 2. A good fit was found for any of the zeolites
60 mL min−1 . studied.
The maximum CH3 SH uptake of 183 mg g−1 is observed on NaX
corresponding to 0.203 mL g−1 (CH3 SH density of 0.9). Such loading

A
1
1
A
CCH3SH / Co CH3SH

0.75 0.75
CCH3SH / CoCH3SH

0.5
0.5

0.25
0.25

0
0 50 100 150 200 250 300 350
Time / min 0
0 20 40 60 80 100 120 140 160
Time / min
0.18

0.16
B 0.09

0.08
B
Mass uptake, gCH3SH / gads

0.14
Mass uptake, gCH3SH / gads

0.07
0.12
0.06
0.1
0.05
0.08
0.04
0.06
0.03
0.04
0.02
0.02
0.01
0
0 50 100 150 200 250 300 350 0
Time / min 0 20 40 60 80 100 120 140 160
Time / min
Fig. 1. Adsorption of CH3 SH on Me-zeolites; conditions: feed CH3 SH/CH4
(0.05/99.95, v/v), gas flow 50 mL min−1 , mass 100 mg, T = 298 K; (A) breakthrough Fig. 2. Adsorption of CH3 SH on Me-zeolites; conditions: feed CH3 SH/CH4
curves; (B) uptake profiles. () NaX, (♦) CaX, () NiNaX, () BaX, () ZnNaX, (×) (0.05/99.95, v/v), gas flow 50 mL min−1 , mass 100 mg, T = 298 K; (A) breakthrough
MgNaX, () CsNaX. curves; (B) uptake profiles. () NaY, (♦) CsY, () NiY.
 A. Ryzhikov et al. / Applied Catalysis A: General 397 (2011) 218–224 221

Table 2
Yoon–Nelson equation modelling.

Material NaX CaX BaX MgNaX ZnNaX NiNaX CsNaX

−1
kYN (min ) 0.05 0.06 0.09 0.05 0.06 0.045 0.25
Exp t1/2 181 148 125 123 118 83 71
Calc t1/2 180 140 134 100 95 83 65
r2   0.999 0.985 0.997 0.993 0.997 0.992 0.999
dC/Co
103 dt
10.2 12.7 22.7 9.5 11.1 7.8 54.3
t1/2

corresponds to 2/3 filling of the ␣-cages of the zeolite-X or -Y. It is is then 106 and 79 mg g−1 for NaX and BaX, respectively, the most
lower than that observed with adsorption of C2 H5 SH on NaX for efficient adsorbents. The larger decrease of uptake, with respect to
which a complete filling of the ␣-cages was reported [4]. This is the saturation uptake capacity, observed for NaX (−42%) in compar-
likely due both to the higher volatility of CH3 SH and to the stronger ison with BaX (−28%) comes from the less steep rise breakthrough
specific interaction of CH3 SH with Na+ cation. profile for NaX. The Yoon–Nelson rate constant for BaX is indeed
However, the molar concentration of each cation in the Me- twice as for Na–X (Table 2). This smoother adsorption profile is
zeolite is different. Therefore, for a better comparison of the affinity correlated with a weaker bonding strength between CH3 SH and
of individual cation for CH3 SH, the CH3 SH uptake was normalized Na+ than with Ba2+ . Various reasons may explain this behavior: (i)
in moleCH3 SH /molecation . Because, full exchange of Na by Cs, Mg, Ni the wider distribution of accessible cationic sites for adsorption in
or Zn cannot be achieved, the normalized uptake for Cs, Mg, Ni and NaX (sites III and II), (ii) an increase of lateral interaction between
Zn was determined after subtraction of CH3 SH adsorbed on residual adsorbed CH3 SH in the more populated ␣-cage of NaX than BaX
Na. It was assumed that the CH3 SH/Na stoichiometry on residual which will decrease the bonding strength between CH3 SH and the
Na was the same as in the parent NaX, i.e., 0.59. cation [41], and (iii) a Ba–S bond stronger than Na–S due to the
It is interesting to comment about the values of greater charge at the Ba2+ cation. Anyway, from a practical point
moleCH3 SH /moleNa in NaX and NaY of 0.59 and 0.38, respec- of view, NaX presents the larger uptake even at low breakthrough
tively. From molecular simulation studies, Abrioux et al. [38] point of CH3 SH.
concluded that at full number of Na per unit cell in NaX (Si/Al = 1), Because in the TSA process the trap needs to be regenerated, we
about 2/3 of Na+ cations occupies sites III and II of the ␣-cage, 1/3 have studied the thermal desorption of NaX after full uptake with
sitting at site I in the hexagonal prisms. Regarding NaY (Si/Al = 2.4), CH3 SH by TG-DSC experiment.
2/3 Na+ cations located at sites II, and 1/3 at sites I within the Fig. 3 presents the weight loss and the associated thermal
sodalite cavity. In both cases, 66% about of Na+ cations are acces- events as a function of temperature. First of all, the weight loss
sible for molecules within the ␣-cage. Because CH3 SH cannot (14.5%) closely corresponds to the full uptake of CH3 SH in NaX
access to the sodalites cavities, one may conclude that in NaX the (0.183/1.183 = 0.155). The thermogram can be decomposed in four
stoichiometry of adsorption, i.e., 0.59, merely corresponds to one zones of temperatures. Up to 100 ◦ C (zone A), this first zone corre-
CH3 SH molecule adsorbed per Na atom in the ␣-cage. In contrast, sponds to the thermal equilibrium of the experimental set up and
in NaY only 60% of Na cations are saturated by CH3 SH. One can no weight loss is observed. In the zone B (100–240 ◦ C) there is a
therefore conclude that the adsorption of CH3 SH is stronger in NaX weight loss at nearly constant rate associated with an endotherm,
as compared to NaY. Following the CH3 SH adsorption on Na-ZSM5 which corresponds to the desorption of CH3 SH. The phenomena
by IR spectroscopy, Garcia and Lercher [3] concluded that CH3 SH occurring between ca 240 and 280 ◦ C (zone C) need more com-
is coordinatively bound to Na+ through the lone electron pair of ments. We propose that CH3 SH which remains adsorbed reacts
the sulfur atom with a stoechiometry of one CH3 SH per Na+ site according to: 2CH3 SH → CH3 SCH3 + H2 S. Basic thermodynamic cal-
at low equilibrium pressure (10−1 Pa). Ziolek and Sugioka reached culations show that this reaction is exothermic (−12.6 kJ mol−1 )
similar conclusions [39] for adsorption of CH3 SH on NaY, and in agreement with the thermal event observed. Moreover, the
concluded that the adsorption is stronger on the more basic NaX, catalytic conversion of CH3 SH to CH3 SCH3 on Na-ZSM5 showed
with a possible dissociation at the S–H bond. This is in line with that the onset of this reaction took place around 220–250 ◦ C
the stronger affinity we observed of CH3 SH for NaX as compared [42]. It is worthy to note that the condensation of methanol
to NaY. It is more difficult to discuss about CsNaX because the to dimethylether demonstrates a similar exothermicity (exper-
respective locations of Na and Cs are not known. In contrast, this imental value: H = −11.9 kJ mol−1 ) [43]. In addition, the small
uncertainty does not exist for NaY and CsY, which means that the endothermic peak at ca 260 ◦ C would probably be associated with
moleCH3 SH /moleNa,Cs values of 0.38 and 0.79 obtained for NaY and the fast desorption of H2 S and/or CH3 SCH3 product. Finally, the
CsY respectively, would provide additional evidence of a higher slow desorption process at above 300 ◦ C (zone D) may be associ-
affinity of CH3 SH upon increase of basicity. ated with H2 S desorption in strong interaction with cations within
Regarding CaX and BaX, we can reasonably propose that more the sodalite cavities (sites II and I ).
than one CH3 SH species is bonded per divalent cation. The value of
moleCH3 SH /moleNa,Ba is close to unity, but Ca2+ and Ba2+ exhibit a 3.2. Adsorption of COS
higher affinity for hidden sites within the sodalite cavity [40,41].
This means that the fraction of Ca2+ or Ba2+ located within the The uptake of COS in LDH is very low (1–2%). For Mg/Al LDH of
␣-cage does bond more than one CH3 SH molecule. As conclud- various compositions, Sparks et al. [29] reported a COS uptake of
ing comments, for a given zeolite structure, it comes out that the 2–4%.
affinity of CH3 SH for metal-exchanged zeolites increases with the More significant COS uptakes are observed on CaX and BaX
basicity and the valence of the cation. This is well reflected by the materials, with values around 55 mg g−1 (Fig. 4, Table 1), which
ranking of the Yoon–Nelson rate constants (Table 2). corresponds to 0.36 molCOS /molBa . In contrast, NaX which exhibits
The maximum CH3 SH uptake of the Me-zeolite is certainly an the largest CH3 SH uptake does not behave in a similar fash-
important parameter, but the actual uptake at a given CH3 SH leak ion with respect to COS, and only 30 mg g−1 COS were taken up
(breakthrough point) is more relevant for a clean-up process. If we (0.08 molCOS /molNa ). If we set the breakthrough point at 5 ppm COS
set the breakthrough point at 5 ppm CH3 SH leak, the CH3 SH uptake leak, the COS uptake becomes 33 mg g−1 for BaX. It comes out that
222 A. Ryzhikov et al. / Applied Catalysis A: General 397 (2011) 218–224

Fig. 3. Thermodesorption of CH3 SH from NaX. Initial mass of NaX = 10.568 mg.

0.06 Table 3
COS conversion (%) over Me-containing X zeolite; GHSV = 7000 h−1 .

0.05 Temperature (◦ C) 50 100 150 200

Mass uptake, g COS/g ads

CsNaX 10 85 100
0.04 NaX 2 6 100
BaX 1 3 12 100
CaX 1 2 4 5
0.03 MgNaX 1 2 4 5

0.02
3.3.1. COS hydrolysis over metal-exchanged X zeolite
0.01 The catalytic performances of a series of metal-exchanged X
zeolite are summarized in Table 3.
0.00 It can be observed that there are considerable differences of
0 50 100 150 200 250 activity among the catalysts, indicating that the nature of the
Time / min cations plays a major role. Thus, CsNaX and NaX are excellent and
good catalysts, respectively, for COS hydrolysis, whereas CaX and
Fig. 4. Adsorption of COS on Me-exchanged X zeolite; conditions: feed COS/CH4 MgNaX show insignificant activity. BaX shows measurable activity
(0.05/99.95, v/v), flow 50 mL min−1 , mass 100 mg, T = 298 K. () NaX, (♦) CaX, ()
NiNaX, () BaX, () ZnNaX, (×) MgNaX, () CsNaX, ( ) NaY, (+) CsY, (−) NiY.
only at temperature higher than 150 ◦ C.
Based on these results, it is reasonable to consider that the cat-
alytic activity of these materials can be mainly related to their
the affinity of COS is clearly lower for Me-zeolite in comparison surface basicity. Indeed, the more basic NaX and CsNaX zeolites are
with CH3 SH, and this decrease is larger for NaX than it is for BaX. by far the best catalysts. This finding is broadly consistent with the
literature results obtained on alumina-type catalysts. For instance
it was reported that the catalytic activity in COS hydrolysis can be
3.3. COS hydrolysis on Me-zeolites and LDHs
increased by the presence of the basic sites [44,45]. Fiedorow et al.
reported that the basic sites are essential for the COS hydrolysis
Among the methods developed to remove COS from different
[46]. As detailed by Hoggan et al. for alumina, the first step of the
gas streams, the catalytic hydrolysis (Eq. (2)) is considered as the
hydrolysis mechanism involves a nucleophilic attack of the basic
most promising.
hydroxyl groups on the COS carbon atom [19]. As shown in Table 3,
COS + H2 O → H2 S + CO2 (2) the Me-zeolites exhibit insignificant conversion at low tempera-
ture, particularly at 50 ◦ C. This finding could be interpreted as a
A major advantage of this process is that the H2 S formed can be strong adsorption of the acidic products (H2 S and CO2 ) on the basic
easily removed using conventional deacidizing processes. sites at low temperature.
In the present study we report for the first time about the use
of metal-exchanged zeolites and mixed oxides derived from LDH 3.3.2. COS hydrolysis over LDH-based materials
as catalysts for the hydrolysis of COS. In order to avoid significant LDHs with different composition activated in situ by calcination
deactivation of catalyst, the catalytic tests were achieved for an under air at 300 ◦ C for 30 min, as well as the mixed oxides, derived
overall period of about 60 min on stream (15 min at 50, 100, 150 and from the LDHs by calcinations at 450 ◦ C (see experimental part),
200 ◦ C). As previously reported [21,26], the catalyst deactivation were screened for their ability to catalyze the COS hydrolysis under
during the COS hydrolysis can be related mainly to the decrease the conditions described in the experimental part. The as-prepared
in the number of basic sites due either to the strong adsorption of samples exhibited the typical XRD pattern of the LDH structure and
acidic molecules (CO2 , H2 S, COS) or blocking by sulfur and sulfate specific surface areas between 10 and 25 m2 g−1 . It is well known
species. that the XRD pattern of the LDH samples activated at 300 ◦ C still
 A. Ryzhikov et al. / Applied Catalysis A: General 397 (2011) 218–224 223

Table 4 a blocking of the stronger O2− basic sites in the former case due to
COS conversions (%) on LDH-type materials; GHSV = 7000 h−1 .
the presence of remaining nitrates [60,61].‘
Temperature (◦ C) 50 100 150 200 The results obtained with the LDHs based catalysts confirm that
Mg-Al(CO3 ) 7 100 COS hydrolysis involves basic sites of high strength. It is worth not-
Mg-Al(CO3 , calcined) 58 100 ing that these basic sites can be of different nature such as CO3 2−
Mg-Al(NO3 ) 3 6 100 8 present in the dehydroxylated LDHs or OH− and O2− sites present
Mg-Al(NO3 , calcined) 3 9 32 36 in the mixed oxides are found active in the reaction. Interestingly,
Zn-Al(CO3 ) 8 92 0
Zn-Al(CO3 , calcined) 0 24 48
some LDH based materials are more active catalysts for COS hydrol-
Zn-Al(NO3 ) 0 0 100 0 ysis than the reference ␥-alumina.
Zn-Al(NO3 , calcined) 0 12 41
Ni-Al(CO3 ) 0 83 0
Ni-Al(CO3 , calcined) 44 100 0 4. Conclusion
Ni-Al(NO3 ) 0 0
Ni-Al(NO3 , calcined) 12 100 The adsorption of CH3 SH and COS, and the hydrolysis of COS
␥-Al2 O3 26 100
in the presence of the metal-exchanged zeolites and LDHs were
investigated in this study. The affinity of both sulfur compounds
for Me-zeolites was higher than that for LDHs. On the other
exhibited typical lamellar structure with a decrease of crystallinity.
hand, the affinity of COS was found lower in comparison with
As compared to the as-prepared samples, the main change observed
CH3 SH. Generally, for a given zeolite structure, the saturation
is a decrease of the basal spacing due to a collapse of the structure
adsorption capacity of CH3 SH for metal-exchanged zeolites was
[52]. The LDHs calcined at 450 ◦ C exhibited the characteristic XRD
found to increase with the basicity and the valence of the cation.
patterns of the poorly crystallized M-aluminum (M = Mg, Ni or Zn)
Thus, the adsorption capacities of the metal-exchanged zeolites
mixed oxide phases. Their specific surface areas range from 60 to
for CH3 SH decreased in the sequence: NaX > CaX > MgNaX > ZnNaX
125 m2 g−1 .
> BaX > NiNaX > NaY > CsY > CsNaX > NiY. The amount of CH3 SH
The COS conversions obtained at various temperature are listed
adsorbed on NaX was 183 mg g−1 . Yoon–Nelson modelling gave
in Table 4. Results obtained under similar conditions over ␥-
good predictions of half time of breakthrough for each zeolite
alumina (288 m2 g−1 ) as catalyst were also included in this Table 4.
tested. Using the TG analysis it was found that the Me-zeolite can
Based on these results, some interesting remarks could be
be regenerated by thermal desorption up to 350 ◦ C.
drawn:
Hydrolysis of COS was tested with the same metal exchanged
The three types of CO3 2− -containing LDHs activated at 300 ◦ C
zeolites and LDH materials and the basicity of the catalysts was
perform COS hydrolysis and their reactivity ranges as follows:
found to control the hydrolysis. The LDH-based materials are more
Mg-Al(CO3 ) > Zn-Al(CO3 ) > Ni-Al(CO3 ). As previously reported, the
active catalysts than the Me-zeolites. The CO3 2− -containing LDHs
uncalcined LDHs are weakly active or totally inactive in most of the
activated at 300 ◦ C and the mixed oxides obtained after calcina-
catalytic reactions, particularly those requiring strong basic sites
tions at 450 ◦ C exhibited outstanding catalytic activity for the COS
[47–53]. In contrast, activation at 300–350 ◦ C of Mg-Al(CO3 ) LDH
hydrolysis performed at temperatures lower than 100 ◦ C.
which lead to partially decarbonated and dehydroxylated layered
material, gives rise to an highly reactive catalyst in a typical base-
catalyzed reaction such as the conversion of 2-methyl-3-butyn-2-ol References
(MBOH) into acetone and acetylene [54]. It was suggested that
[1] A. Rojey, B. Durand, Le gaz naturel: production, traitement, transport, Publica-
the basic sites are CO3 2− ions at external surface [55]. This type tions de l’institut Français du Pétrole, Editions Technip, 1994, pp. 17–19 and
of active surface sites can actually account for the reactivity of references therein.
the LDH samples activated at 300 ◦ C in COS hydrolysis. The nature [2] M. Ziolek, J. Czyzniewska, J. Lamotte, J.C. Lavalley, React. Kinet. Catal. Lett. 53
(1994) 339–346.
of the divalent cation also influences the activity, Mg2+ , as gen-
[3] C.L. Garcia, J.A. Lercher, J. Phys. Chem. 95 (1991) 10729–10736.
erally observed, leading to the most basic sample [49]. Moreover, [4] G. Weber, F. Benoit, J.P. Bellat, C. Paulin, P. Mougin, M. Thomas, Microporous
exchange of NO3 − for CO3 2− greatly decreases the basicity of the Mesopororous Mater. 109 (2008) 184–192.
[5] Z. Sarbak, Appl. Catal. A 147 (1996) 47–54.
LDHs which become almost totally inactive. This confirms that
[6] J.M. Dysard, G.F. Stuntz, T.R. Halbert, A. Malek, US Patent 20050252831 (2005).
CO3 2− are the active sites and this is in agreement with the results [7] S. Bashkova, A. Bagreev, T. Bandosz, J. Environ. Sci. Technol. 36 (2002)
reported for the reaction between formaldehyde and acetone with 2777–2782.
NO3 − - and CO3 2− -containing Mg/Al LDHs activated at 350 ◦ C [56]. [8] H. Cui, S.Q. Turn, M.A. Reese, Catal. Today 139 (2009) 274–279.
[9] T. Hisashi, H. Nagoya, T. Shiono, J. Colloid Interface Sci. 300 (2006) 814–817.
The Mg-containing mixed oxide obtained after calcination at [10] J.J. Alcaraz, B.J. Arena, R.D. Gillespie, J.S. Holmgren, Catal. Today 43 (1998) 89–99.
450 ◦ C of the Mg-Al(CO3 ) LDH is slightly more active than that [11] K.M. Brown, W.K.T. Gleim, P. Urban, Oil Gas J. 57 (1959) 73–78.
obtained from the Ni-Al(CO3 ) LDH. Besides, both materials are more [12] P. Briot, R. Cadours, S. Drozdz, F. Lecomte, US Patent 0193925 (2007).
[13] A. Carlsson, G.J. van Heeringen, US Patent 00447201 (2009).
active than the parent LDH precursors activated at 300 ◦ C as well as [14] http://www.uop.com/refining/.
alumina ␥-Al2 O3 . A substantial drop in activity is observed for the [15] J. West, B.P. Williams, N. Young, C. Rhodes, G.J. Hutchings, Catal. Commun. 2
Zn-containing mixed oxide. Such influence of the divalent cation (2001) 135–138.
[16] B. Thomas, B.P. Williams, N. Young, C. Rhodes, G.J. Hutchings, Catal. Lett. 86
in the mixed oxides can be attributed to their respective basic- (2003) 201–205.
ity already determined either by ␤-isophorone isomerisation [57], [17] J. West, B.P. Williams, N. Young, C. Rhodes, G.J. Hutchings, Catal. Lett. 74 (2001)
adsorption of CO2 followed by calorimetry [49,55] or ethanoly- 111–114.
[18] J. Bachelier, A. Aboulayt, J.C. Lavalley, O. Legendre, F. Luck, Catal. Today 17 (1993)
sis of propylene oxide [49]; this results in the following basicity
55–62.
scale: Mg(Al)O > Ni(Al)O > Zn(Al)O. Accordingly, higher activity of [19] P.E. Hoggan, A. Aboulayt, A. Pieplu, P. Nortier, J.C. Lavalley, J. Catal. 149 (1994)
Mg- than Ni-containing mixed oxide was reported in the conden- 300–306.
[20] B.P. Williams, N.C. Young, J. West, C. Rhodes, G.J. Hutchings, Catal. Today 49
sation of acetone [47,58] and acetaldehyde [59].
(1999) 99–104.
As previously observed for the dehydroxylated lamellar struc- [21] C. Rhodes, S.A. Riddel, J. West, B.P. Williams, G.J. Hutchings, Catal. Today 59
tures, the mixed oxides obtained from the NO3 − -LDHs precursors (2000) 443–464.
are less active than those obtained from the CO3 2− -containing pre- [22] H. Huang, N. Young, B.P. Williams, S.H. Taylor, G. Hutchings, Catal. Lett. 110
(2006) 243–246.
cursors. We can suggest that the slightly higher thermal stability of [23] J. Liu, Y. Liu, L. Xue, Y. Yu, H. He, Acta Phys. -Chim. Sin. 23 (2007) 997–1002.
the nitrates as compared to that of the carbonates species induces [24] S. Tan, C. Li, S. Liang, H. Guo, Catal. Lett 8 (1991) 155–168.
224 A. Ryzhikov et al. / Applied Catalysis A: General 397 (2011) 218–224

[25] H. Huang, N. Young, B.P. Williams, S.H. Taylor, G. Hutchings, Catal. Lett. 104 [42] E. Huguet, V. Hulea, B. Coq, unpublished data.
(2005) 17. [43] R.V. Ermakov, V.A. Plakhotnik, Pet. Chem. 48 (2008) 1.
[26] H. Huang, N. Young, B.P. Williams, S.H. Taylor, G. Hutchings, Green Chem. 10 [44] Z.M. George, J. Catal. 35 (1974) 218–224.
(2008) 571–577. [45] V.A. Ivanov, A. Pieplu, J.C. Lavalley, P. Nortier, Appl. Catal. A: Gen. 131 (1995)
[27] S.D. Jackson, D. Winstanley, J. Catal. 121 (1990) 312–317. 323–334.
[28] M.L. Sattler, R.S. Rosenberk, J. Air Waste Manage. Assoc. 56 (2006) 219– [46] R. Fiedorow, R. Leaute, I.G. Dalla Lana, J. Catal. 85 (1984) 339.
224. [47] A. Guida, M.H. Lhouty, D. Tichit, F. Figueras, P. Geneste, Appl. Catal. A: Gen. 164
[29] D.E. Sparks, T. Morgan, P.M. Patterson, S.A. Tackett, E. Morris, M. Crocker, Appl. (1997) 251.
Catal. B: Environ. 82 (2008) 190–198. [48] S. Abello, F. Medina, D. Tichit, J. Pérez-Ramirez, J.C. Groen, J.E. Sueiras, P. Salagre,
[30] V.C. Srivastava, B. Prasad, I.M. Mishra, I.D. Mall, M.M. Swamy, Ind. Eng. Chem. Y. Cesteros, Chem. Eur. J. 11 (2005) 728.
Res. 47 (2008) 1603–1613. [49] F. Prinetto, D. Tichit, R. Tessier, B. Coq, Catal. Today 55 (2000) 103.
[31] S. Singh, V.C. Srivastava, I.D. Mall, Colloids. Surf. A. 332 (2009) 50–56. [50] J. Palomeque, J. Lopez, F. Figueras, J. Catal. 211 (2002) 150.
[32] G.S. Bohart, E.Q. Adams, J. Am. Chem. Soc. 42 (1920) 523–544. [51] W. Kagunya, Z. Hassan, W. Jones, Inorg. Chem. 35 (1996) 5970.
[33] H.C. Thomas, J. Am. Chem. Soc. 66 (1944) 1664–1666. [52] F. Winter, X. Xia, B.P.C. Hereijgers, J.H. Bitter, A.J. van Dillen, M. Muhler, K.P. de
[34] Z. Aksu, F. Gonen, Process Biochem. 39 (2004) 599–613. Jong, J. Phys. Chem. B 110 (2006) 9211.
[35] Y.H. Yoon, J.H. Nelson, Am. Ind. Hyg. Assoc. J. 45 (1984) 509–516. [53] J.C.A.A. Roelofs, A.J. van Dillen, K.P. de Jong, Catal. Today 60 (2000) 297.
[36] H.H. Kung, Catal. Today 11 (1992) 443–453. [54] V.R.L. Constantino, T.J. Pinnavaia, Inorg. Chem. 34 (1995) 883.
[37] M.M.J. Treacy, J.B. Higgins, Collection of Simulated XRD Powder Patterns for [55] V.R.L. Constantino, T.J. Pinnavaia, Catal. Lett. 23 (1994) 361.
Zeolites, Elsevier, 2007. [56] E. Suzuki, Y. Ono, Bull. Chem. Soc. Jpn. 61 (1988) 1008.
[38] C. Abrioux, B. Coasne, G. Maurin, F. Henn, A. Boutin, A. Di Lella, C. Nieto-Draghi, [57] J. Sanchez Valente, F. Figueras, M. Gravelle, P. Kumbhar, J. Lopez, J.-P. Besse, J.
A.H. Fuchs, Adsorption 14 (2008) 743–754. Catal. 189 (2000) 370.
[39] M. Ziolek, M. Sugioka, Res. Chem. Intermed. 26 (2000) 385–412. [58] W.T. Reichle, J. Catal. 94 (1985) 54.
[40] P.P. Lai, L.V.C. Rees, J. Szilard-Chalmers, J. Chem. Soc. Faraday Trans. 1 (72) [59] E. Dumitriu, V. Hulea, C. Chelaru, C. Catrinescu, D. Tichit, R. Durand, Appl. Catal.,
(1976) 1809–1817. A 178 (1999) 145–157.
[41] M.L. Costenoble, W.J. Mortier, J.B. Uytterhoeven, J. Chem. Soc. Faraday Trans. 1 [60] O. Lebedeva, D. Tichit, B. Coq, Appl. Catal. A: Gen. 183 (1999) 61–71.
(74) (1978) 466–476. [61] M. Lal, A.T. Howe, Solid State Chem. 39 (1981) 368–376.