High-temperature sulfur removal from biomass-derived synthesis gas over bifunctional molybdenum catalysts

Article abstract of DOI:10.1002/cctc.201300185

The removal of sulfur species from the biomass-derived producer gas after gasification is required to protect downstream catalysts. Finding suitable materials for high-temperature desulfurization is one of the main challenges for the improvement in the ef

Full text of DOI:10.1002/cctc.201300185

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201300185
High-Temperature Sulfur Removal from Biomass-Derived
Synthesis Gas over Bifunctional Molybdenum Catalysts
Christian F. J. Kçnig,[a] Patrick Schuh,[a, b] Tilman J. Schildhauer,[a] and Maarten Nachtegaal*[a]
The removal of sulfur species from the biomass-derived pro-
ducer gas after gasification is required to protect downstream
catalysts. Finding suitable materials for high-temperature de-
sulfurization is one of the main challenges for the improve-
ment in the efficiency of catalytic biomass conversion. The bio-
mass-derived producer gas usually contains not only H2S but
also organic sulfur species such as thiophene (C4H4S), which
cannot be removed by most sorbent materials. Herein, we ex-
plored Al2O3-supported molybdenum catalysts as bifunctional
materials for the removal of H2S and catalytic conversion of
C4H4S at high temperatures. By using X-ray absorption spec-
troscopy under reaction conditions, we show that H2S is re-
moved through the sulfidation of MoO3. C4H4S is catalytically
converted over MoO2 to MoS2 and hydrocarbon species. The
subsequent oxidation of MoS2 to MoO3 and SO2 regenerates
the material and allows the sequestration of sulfur from the
gas stream. Furthermore, the negative effect of steam on
sulfur removal is shown to be caused by competitive adsorp-
tion with sulfur species. These findings show the possibility of
the high-temperature desulfurization of biomass-derived gas
for catalytic conversion, for instance, to synthetic natural gas.
Introduction
In recent years, biomass has received increased interest as
a chemical feedstock or as fuel. Technologies for the thermo-
chemical conversion of biomass to fuels, for example, through
gasification, have the potential to be cost competitive with
gasoline.[1]
sary re-evaporation of H2O for the adjustment of the H2 to CO
ratio via steam reforming, and regeneration of scrubbing liq-
uids make cold gas cleaning thermally inefficient and expen-
sive. Technoeconomic analysis showed that processes such as
tar reforming and acid gas and sulfur removal comprise up to
31% of the minimum selling price for ethanol produced from
biomass.[12] This finding suggests that efficient sulfur and tar re-
moval can lower costs significantly.
During gasification of biomass feedstocks, a gas mixture
mainly of CO, H2, CO2, CH4, and H2O is generated.[2] In addition,
depending on gasification parameters, several undesired by-
products and contaminants with different concentrations are
produced, which include olefins, tars, sulfur, and nitrogen het-
eroatom species (thiophene [C4H4S] and pyridine); inorganic
constituents such as sulfur, chlorine, and nitrogen (H2S, COS,
HCl, NH3, and HCN); and alkali metals.[3–5] Sulfur-containing
compounds are corrosive to pipelines and downstream instal-
lations and thus limit plant lifetime.[6,7] Moreover, sulfur species
poison downstream catalysts, which negatively affects the cat-
alyst performance. For example, small amounts of H2S poison
the nickel or ruthenium catalysts used for methanation,[8,9] hot
gas cleaning,[10] and Fischer–Tropsch synthesis.[11]
Hot gas desulfurization operating between 300 and 8508C is
preferred to improve the efficiency of biomass-to-fuel conver-
sion.[13,14] Desulfurization over sorbent materials and/or cata-
lysts leads to the accumulation of sulfur on the sorbent over
time on stream. To maintain operation, the sorbent needs to
be replaced or regenerated: the latter either periodically (e.g.,
in a swing reactor) or continuously. For the continuous remov-
al of sulfur from the gas stream, a chemical looping process is
used.[15] For regeneration, the sorbent is exposed to an oxidiz-
ing atmosphere, which leads to the production of SO2, which
in turn can be sequestrated from the cleaned fuel gas stream.
The sequestrated SO2 can then be captured in scrubbers by
a slurry or solid sorption materials.[16,17]
In coal or biomass gasification process chains, sulfur compo-
nents are often removed by using low-temperature processes
(e.g., scrubbing), such as cooling the producer gas from 8508C
(gasification temperature) to below the dew point of H2O. Re-
heating up to 250–4008C for downstream fuel catalysis, neces-
One major challenge for hot gas desulfurization is the iden-
tification of materials that remove different sulfur species from
the gas in the presence of steam, and that are stable over
many cycles of sulfur uptake and regeneration. Zinc titanate
sorbents were proposed as regenerable, attrition-resistant H2S
sorbents.[18] The literature focuses largely on the removal of
H2S from coal gas without investigating organic sulfur com-
pounds.[19,20] Such organic sulfur compounds are usually pres-
ent in gasified biomass owing to the low gasification tempera-
ture (compared to the temperature of coal gasification) and
are not usually removed by sorbents.[5]
[a] C. F. J. Kçnig, P. Schuh, Dr. T. J. Schildhauer, Dr. M. Nachtegaal
Paul Scherrer Institut
5232 Villigen PSI (Switzerland)
Fax: (+41)56-310-2199
E-mail: maarten.nachtegaal@psi.ch
[b] P. Schuh
Technische Universitꢀt Mꢁnchen
05747 Garching (Germany)
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3700 – 3711 3700

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Molybdenum, along with Ni, Co, or other promoters, re-
moves sulfur-containing components in the hydrodesulfuriza-
tion (HDS) process in petroleum refining.[21] In contrast to the
HDS process, water vapor is always present in the producer
gas from the gasified biomass, either from the moist feedstock
or from steam used as a gasification medium.[22] The competi-
tive adsorption of H2O on the catalyst has been reported to
reduce the desulfurization of C4H4S over CoꢀMo catalysts,[23] in
hydrodeoxygenation (HDO) over NiꢀMo[24] and CoꢀMo,[25] and
in H2S removal over ZnO.[26] Furthermore, the partial pressure
of H2 in the producer gas from atmospheric gasifiers is of the
order of several 100 mbar whereas HDS reactors usually oper-
ate at a H2 partial pressure of greater than 10 bar (1 bar=
0.1 MPa).[21] As the biomass-derived synthesis gas has low par-
tial pressures of H2 and high amounts of steam, the develop-
ment of specific materials for the high-temperature desulfuriza-
tion of biomass-derived synthesis gas is required. Sulfide cata-
lysts, as they are usually used in the HDS process under high
H2 partial pressures, cannot be used for the hot gas desulfuri-
zation of biomass with continuous regeneration, because the
sulfur on the catalyst would be
significantly lower steam content. To study the state of Mo
under reaction conditions, in situ X-ray absorption spectrosco-
py (XAS) was used. Recent advances in synchrotron instrumen-
tation are able to perform in situ studies on time scales of sec-
onds or subseconds, which are relevant for dynamic processes,
such as chemical looping.[30] The mechanism of the removal of
H2S was found to be the concurrent sulfidation of MoO3 to
MoS2 and reduction to MoO2. The removal of C4H4S proceeds
via the desulfurization of the C4H4S ring, which leads to the
sulfidation of MoO2.
Results
Characterization of the fresh catalyst
To characterize the local structure of the fresh, calcined cata-
lyst, the sample was investigated by using XAS. The X-ray ab-
sorption near-edge spectroscopy (XANES) spectra of the fresh
Mo/Al2O3 catalyst along with the spectra of reference com-
pounds are shown in Figure 1a. The XANES spectra of the
immediately oxidized to SO2
upon regeneration.
The sulfidation of Mo catalysts
and their HDS activity were stud-
ied in the past with various tech-
niques, and the findings are
summarized in reviews.[27,28] The
Mo-based catalysts for HDS are
usually regarded as slabs of
MoS2, which are terminated by
sulfur and eventually doped
with Co on the edges of the
slab.[29] The active sites are as-
sumed to be vacancies on the
edges of MoS2. Furthermore, in
HDS processes, C4H4S is continu-
ously converted to H2S, which is
eliminated downstream in a ZnO
bed,[21] whereas in hot gas desul-
furization processes, all sulfur is
ideally converted to SO2 and se-
questrated from the synthesis
Figure 1. a) Mo K-edge XANES spectra and b) Fourier transforms of the EXAFS spectra of the fresh Mo/Al2O3 cata-
lyst sample versus Mo reference compounds for MoO3, MoO2, MoS2, and Mo foil.
gas stream.
Herein, we investigate mecha-
nisms of sulfur removal in
a cyclic process from gasified biomass at high temperatures.
The goal is to develop a high-temperature gas cleaning pro-
cess and to understand the mechanisms of removal not only
of H2S but also of organic sulfur species, which are rarely con-
sidered in high-temperature gas cleaning studies. As Mo-based
catalysts are often used for the conversion of organic sulfur
species, Mo is a promising material to accomplish sulfur re-
moval. However, the cyclic operation of sulfur removal and re-
generation leads to requirements for the material that are dif-
ferent from those for HDS, which is usually performed at lower
temperatures, at much higher H2 partial pressures, and with
fresh catalyst demonstrate features that are similar to those of
the reference MoO3 spectrum, such as the shoulder in the
spectrum at 20005 eV and the position of the absorption
edge. This qualitative comparison suggests that the fresh Mo
catalyst is oxidized to MoO3 after calcination. The correspond-
ing Fourier transforms of the extended X-ray absorption fine
structure (EXAFS) spectra are shown in Figure 1b. In the Fouri-
er transforms, the fresh catalyst shows one peak at approxi-
mately 1.2 ꢁ but no significant peaks at higher distances.
Based on the comparison to the Fourier transforms of the ref-
erences, this peak is a MoꢀO scattering shell at a distance simi-
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3700 – 3711 3701

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
lar to that in the MoO3 reference spectra. According to the
XANES and EXAFS spectra, the spectrum of the fresh catalyst is
identified as MoO3. In contrast to the reference spectra, the
fresh, calcined MoO3/Al2O3 sample does not show scattering
shells at higher distances (such as the MoꢀMo scattering shell
at ꢁ1.8 ꢁ), which suggests a small particle size.
the catalyst was oxidized in 4% O2 for 180 s to regenerate the
sample. The cycles were repeated 10 times. The selected MS
traces for H2 (detected at m/z 2), H2S and O2 (both measured
at m/z 33), and SO2 (m/z 64) of one representative cycle of re-
duction, sulfidation, and regeneration at 6008C are shown in
Figure 3. Although H2S is not detected at the reactor outlet
Fitting the EXAFS spectrum of the fresh catalyst with
a MoO3 reference structure[31] yields a coordination number of
3.7ꢂ0.5 for the MoꢀO shell at a distance of (1.71ꢂ0.01) ꢁ,
with a Debye–Waller factor of (0.005ꢂ0.001) ꢁ2. No scattering
shells at higher distances could be fitted. The scanning TEM
(STEM) analysis of the fresh MoO3/Al2O3 catalyst (Figure 2) re-
Figure 2. STEM micrograph of the fresh Mo/Al2O3 catalyst as measured in
the Z-contrast mode.
veals a particle size of approximately 1 nm, which supports the
findings of the EXAFS analysis.
Figure 3. Cyclic exposure of the Mo/Al2O3 catalyst to reduction in H2, sulfida-
tion in H2 +H2S, and oxidation in O2 at 6008C. The selected MS traces of one
cycle for the reactor outlet are shown along with the sample composition,
which is derived from the linear combination fit of the Mo K-edge XANES
spectra.
To investigate the effect of Mo impregnation on the surface
area and pore volume of the catalyst particles, N2 physisorp-
tion measurements were performed. The blank Al2O3 support
showed a surface area of 170 m2 gꢀ1 (measured by using the
BET method[32]) and a pore volume of 0.46 cm3 gꢀ1 (measured
by using the BJH method[33]), whereas for the fresh MoO3/Al2O3
catalyst (4 times impregnated), a BET surface area of 180 m2 gꢀ1
and a pore volume of 0.45 cm3 gꢀ1 were measured. According
to these results, impregnation and calcination of the support
change the porous structure only slightly. The impregnation re-
sulted in 9 wt% loading of Mo on the Al2O3 support, as mea-
sured by using inductively coupled plasma optical emission
spectroscopy (ICP–OES).
during exposure to 2000 ppm H2S within 900 s, SO2 is detected
upon oxidation with 4% O2. One measure for the stability of
this process is the amount of SO2 that is generated during
each cycle upon oxidation, which is shown in Figure 4 as the
integrated signal at m/z 64 over 10 cycles. These data suggest
that H2S is effectively removed from the producer gas stream
by the Mo/Al2O3 catalyst over multiple cycles and oxidized to
H2S removal under dry conditions
In the producer gas from biomass or coal gasification, H2S is
usually the most abundant sulfur species. Therefore, we first in-
vestigated H2S removal under dry conditions. The fresh catalyst
was heated to 6008C and exposed to periodically changing at-
mosphere, which simulated the conditions in a chemical loop-
ing desulfurization reactor. After reduction in 1.25% H2 for
180 s, 2000 ppm H2S was added to the total feed of
25 mLminꢀ1 for 900 s, which corresponded to a total sulfur
amount of 33.5 mmol. These conditions represent the reducing
conditions for the sulfur-containing producer gas, which usual-
ly also contains CO, CO2, and CH4. After a purge in He for 60 s,
Figure 4. Integrated SO2 signal (detected at m/z 64) over 10 cycles of H2S
exposure to the Mo/Al2O3 catalyst at 6008C.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3700 – 3711 3702

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
SO2. The resulting SO2 stream can be separated by performing
oxidation in a separate reactor, for instance, in a chemical loop-
ing reactor or a swing reactor.
While the XANES region (up to 100 eV above the absorption
edge) probes the electronic structure of the sample, the local
geometric environment is probed by the EXAFS part of the
spectrum. The Fourier transforms of the selected EXAFS spec-
tra during the sulfidation and regeneration cycle are shown in
Figure 5b. After oxidation, a single peak at approximately 1.2 ꢁ
(not phase corrected) is visible at a position similar to that in
the fresh, calcined sample (Figure 1), which indicates full oxida-
tion to MoO3. Upon exposure to H2, the intensity of this peak
decreases whereas a second peak at approximately 1.7 ꢁ ap-
pears. During sulfidation, the peak at 1.2 ꢁ decreases in inten-
sity whereas the peak at 1.7 ꢁ increases in intensity and shifts
to higher distances. The MoꢀO scattering shell of MoO3 has
the peak at 1.2 ꢁ, whereas MoꢀO and MoꢀS scattering shells
of MoO2 and MoS2 have peaks between 1.5 and 2 ꢁ. This result
suggests that the fraction of MoO3 in the sample decreases
over time whereas the fractions of MoO2 and MoS2 increase.
Fitting of the full EXAFS spectra of such mixtures, with many
scattering paths overlapping at
To gain insight into the mechanism of sulfur removal by the
Mo-based catalyst, time-resolved in situ XAS at the Mo K-edge
was used during the cycles of reduction, sulfidation, and oxida-
tion. The XAS spectra were averaged over 10 cycles. Linear
combination fitting (LCF) can be used to fit the percentage of
known reference states of the catalyst. Applying LCF to time-
resolved XAS spectra enables the detection of trends in the re-
action. The spectra were fitted with three components: MoO3,
MoO2, and MoS2. Here, we used the spectrum of the fresh cata-
lyst (Figure 1) as reference for MoO3, because the MoO3 nano-
particles differ slightly in their XANES signals from the MoO3
reference sample. Linear combination fits of the selected
XANES spectra at the beginning of sulfidation in H2S, during
sulfidation, at the end of sulfidation, and after oxidation are
shown in Figure 5a. The resulting time-resolved composition
similar distances, is not possible
with the data quality obtained
herein. Nevertheless, the qualita-
tive interpretation of the EXAFS
spectra supports the findings of
LCF of the XANES spectra in
Figure 3.
H2S removal in the presence of
steam
All woody biomass contains
moisture, with content varying
between 6.5% (for fir) and 60%
(for willow).[22] In addition, steam
is often used as a gasification
medium.[34] Thus, the producer
gas after biomass gasification
contains rather large amounts of
steam. Therefore, it is imperative
that any high-temperature gas
cleaning process operates relia-
bly in the presence of steam, be-
cause condensation for steam
removal would require cooling,
Figure 5. a) XANES spectra (bold lines) with linear combination fits (thin lines) and b) Fourier transforms of the se-
lected spectra after reduction, during sulfidation, after sulfidation, and upon oxidation.
of the Mo catalyst over 1 cycle of sulfidation and oxidation is
shown in Figure 3. Exposure to 1.25% H2 does not substantially
reduce the catalyst, and even a small fraction of MoS2 is deter-
mined from linear combination fits. However, upon exposure
to 2000 ppm H2S, MoO3 is consumed whereas MoO2 and MoS2
are formed at a similar rate. At the end of the sulfidation
period, approximately 50% of the sample in the X-ray beam is
MoO3, 30% is sulfided to MoS2, and 20% is MoO2. The LCF re-
sults indicate that the Mo catalyst is fully oxidized to MoO3
upon exposure to 4% O2 within a few seconds. The difference
between 1 and the sum of the three fractions, that is, the non-
fitted fraction, is represented by the black line. The nonfitted
fraction never exceeded 2%.
which is to be avoided. The experiment of reduction, H2S re-
moval, and oxidation was repeated with additional steam, and
the mass of the catalyst in the reactor was varied. The feed
rate of H2O was 0.1 mLhꢀ1, which corresponds to approximate-
ly 7.7% steam content at a dry gas flow rate of 25 mLminꢀ1
.
The integrated H2S signals at the reactor outlet are shown in
Figure 6, which are normalized to the He signal and averaged
over 10–15 cycles. For comparison, the integrated signal of the
dry experiment is also shown in the figure. Although the pres-
ence of steam reduces the H2S uptake of 25 mg of the catalyst
by approximately 40% compared to that under dry conditions,
higher amounts of the catalyst lead to complete H2S removal
even in the presence of steam, which indicates that the pres-
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3700 – 3711 3703

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 6. Integrated H2S signal at the reactor outlet (measured at m/z 33) for
different catalyst masses under identical experimental conditions at 6008C
with 7.7% steam. The data for a mass of 0 g represent the measurement
through an empty tube. For comparison, the data from the dry experiment
for a mass of 25 mg are denoted by the blank triangle.
ence of steam reduces H2S removal but can be compensated
for by using a larger amount of the catalyst.
C4H4S removal under dry conditions
In the producer gas from gasified biomass, sulfur is present
not only as H2S but also as organic sulfur species, which also
need to be removed from the gas stream. As a model com-
pound for organic sulfur species, C4H4S was chosen here. Simi-
lar to the cyclic sulfidation under H2S (see above), the catalyst
was exposed to periodic reduction, exposure to H2 +C4H4S,
and oxidation at a constant temperature of 6008C. In Figure 7,
a comparison of the MS data for the reactor inlet (dashed
lines) and the reactor outlet (solid lines) shows that all
200 ppm C4H4S (m/z 84) that the Mo catalyst is exposed to is
removed in the presence of 40% H2. Upon oxidation, SO2 is
generated, which indicates that sulfur was bound to Mo/Al2O3
and oxidized in 4% O2. A control experiment with a blank
Al2O3 support under identical conditions did not show detecta-
ble interaction of Al2O3 with C4H4S (data not shown). The re-
moval of C4H4S from the gas stream can therefore be attribut-
ed to the Mo particles.
Figure 7. Cyclic exposure of the Mo/Al2O3 catalyst to reduction in H2, reduc-
tion in H2 +C4H4S, and oxidation in O2 at 6008C. The selected MS traces for
the reactor inlet (dashed line) and the reactor outlet (solid lines) are shown
along with the sample composition (bottom panel), which is derived from
the linear combination fit of the Mo K-edge XANES spectra.
3.3 mmol of C4H4S is expected to result in approximately 3% of
Mo to be present in the MoS2 phase, which is below the detec-
tion limit of the LCF of the XAS spectra (typically 5–10%).
Results from the Fourier-transformed EXAFS spectra, shown
in Figure 8, are qualitatively similar to those from LCF. Upon re-
duction, the intensity of the scattering shell at approximately
1.2 ꢁ decreases whereas a scattering shell at approximately
1.8 ꢁ appears. The first peak at 1.2 ꢁ can be attributed to the
first MoꢀO coordination shell from MoO3, and the scattering
shell at 1.8 ꢁ can be assigned to the first MoꢀO coordination
shell from MoO2 (see the reference spectra in Figure 1).
LCF was performed on the time-resolved XANES data of the
averaged cycle of reduction, exposure to H2 +C4H4S, and sub-
sequent oxidation. The LCF results illustrated in Figure 7 indi-
cate that MoO3 readily reduces to approximately 30% MoO2
and 70% MoO3 after exposure to 40% H2. Compared to the re-
sults of the experiment on H2S uptake (Figure 3), the amount
of MoO2 is higher owing to the higher H2 concentration. Over
the time on stream during exposure to 40% H2 +200 ppm
C4H4S, the Mo catalyst fraction slowly reduces further. Upon ox-
idation in 4% O2, Mo is completely oxidized to MoO3.
Modulated excitation spectroscopy was shown to enhance
the fitting precision of the XAS spectra by filtering out all
those parts in the spectrum that do not follow the excitation
frequency, which is here the switching of the gas feed.[35–37]
The remaining spectra, called demodulated or phase-detected
spectra, are essentially difference spectra between the different
states of the catalyst during one excitation cycle but with
a better signal-to-noise ratio and only those contributions to
the spectra that change upon excitation. Although the XANES
or EXAFS data in Figures 7 and 8 provide no conclusive evi-
dence for the formation of MoS2, the demodulation of the
spectra, which are modulated with the switching of gases,
allows sensitive detection of minute changes in the structure
of the sample.[37] The demodulation of the spectra transforms
them from the time domain to a phase domain. The selected
In contrast to the results of the experiment on H2S uptake,
no significant fraction of MoS2 is found. Within 15 min, the cat-
alyst is exposed to approximately 3.3 mmol of C4H4S. In the ex-
periment with sulfidation by H2S (see above), approximately
30% of the Mo catalyst in the X-ray beam sulfides to MoS2
after exposure to 33.5 mmol of H2S. Therefore, exposure to
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3700 – 3711 3704

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 8. Selected Fourier transforms of the XAS spectra during the cyclic
exposure of the Mo/Al2O3 catalyst to H2, H2 +C4H4S, and O2 at 6008C.
demodulated spectra are shown in Figure 9 in k space and R
space for different phase angles Fk. The demodulated spectra
in k space show that at least two contributions to the spectra
exist, with the major contribution at Fk =3108 and the minor
contribution at Fk =2238. The two spectra demonstrate a dis-
tinctly different shape with strong contributions at different k
values. The Fourier transformation of the demodulated spectra
at Fk =310 and 2238 in R space shows those scattering shells
that change upon excitation, that is, the switching of the gas
atmosphere from H2 to H2 +C4H4S and to O2. For qualitative
comparison, they are plotted along with the difference spectra
of reference compounds of MoO3, MoO2, and MoS2. This com-
parison indicates that the demodulated spectrum at 3108 is
similar to the difference spectra of MoO3 and MoO2 between
1 and 2 ꢁ, which indicates that this spectrum corresponds to
the reduction of MoO3 to MoO2 of a fraction of the sample.
The peak of the demodulated spectrum at 2238 at Rꢁ1.7 ꢁ is
in phase with the difference spectra of MoO2 and MoS2, which
is indicative of the sulfidation of MoO2 to MoS2. The amplitude
of the demodulated spectrum at 2238 is approximately
10 times smaller than that of the spectrum at 3108, which indi-
cates that a much smaller fraction of Mo atoms is sulfided.
Thus, the demodulated spectra suggest that the sample trans-
forms from MoO3 to MoO2, which is confirmed by the linear
combination fit of the XANES spectra. Furthermore, a fraction
of the sample is transformed from MoO2 to MoS2, which is not
readily visible in the XANES or EXAFS spectra.
Figure 9. a) Selected demodulated XAS spectra during the cyclic exposure of
the Mo/Al2O3 catalyst to H2, H2 +C4H4S, and O2 at 6008C. The spectrum at
2238 is shown as a thick red line. b) Fourier transforms of the selected
demodulated XAS spectra and of the difference spectra of reference
compounds.
Effect of prereduction on C4H4S removal
In a cyclic process in which the catalyst is exposed to an alter-
nating atmosphere of the sulfur-containing producer gas,
which is usually reducing, and oxidizing gas for regeneration,
the oxidation state of most of the Mo atoms is changing
(Figure 3). To study the effect of the oxidation state of Mo on
C4H4S removal, the duration of reduction before the addition
of C4H4S was varied. The C4H4S signal detected at the reactor
outlet is shown in Figure 10 for different durations of reduc-
tion. The removal of C4H4S from the gas stream depends on
the time of previous reduction. The amount of C4H4S that is
detected at the reactor outlet decreases over time on stream,
which coincides with the progressing reduction of MoO3 to
MoO2 (Figure 7). For longer times of prereduction, the amount
of C4H4S that is detected upon the addition of C4H4S to the
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3700 – 3711 3705

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 10. C4H4S signal (measured by m/z 84) over time during cyclic reduc-
tion, exposure to H2 +C4H4S, and oxidation for different times of prereduc-
tion.
Figure 11. Integrated C4H4S signal (measured at m/z 84) at the reactor inlet
(cycles 1 and 2) and at the reactor outlet in the absence of H2 (cycles 3–7)
and in the presence of H2 (cycles 8–10).
stream decreases, which suggests that MoO2 is more favorable
for C4H4S removal than is MoO3.
catalyst (Figure 9). As the sulfidation of Mo by H2S (from either
the feed or the hydrogenation of C4H4S) and C4H4S removal
occur in parallel, it is important to understand the effect of the
sulfidation of the catalyst on C4H4S removal. To test this, 50 mg
of the Mo/Al2O3 catalyst was presulfided before exposure to
C4H4S. After approximately 220 min of sulfidation at 6008C in
the presence of 40% H2 +1000 ppm H2S, the catalyst was ex-
posed to 40% H2 +100 ppm C4H4S at the same temperature.
The selected MS traces obtained during sulfidation and subse-
quent C4H4S exposure are shown in Figure 12. After sulfidation,
C4H4S was removed completely and small amounts of H2S and
C4H6 were generated. This finding confirms that a sulfided Mo
catalyst is also active for the removal of C4H4S from the gas
stream as expected because of its use in HDS applications.
Effect of H2 on C4H4S removal
Although the removal of C4H4S under dry conditions is accom-
plished in the cyclic process (see above), the aforementioned
results do not clarify whether the removal of C4H4S is due to
pure sorption of C4H4S on MoO2 or whether it is a catalytic pro-
cess. To study this question, the cycles of reduction, exposure
to C4H4S, and oxidation were modified and the simultaneous
feed of C4H4S with H2 was compared with the feed of C4H4S
without H2. In both cases, the catalyst was reduced in 40% H2
before the addition of 200 ppm C4H4S because, as shown
above, a previously reduced catalyst is required for C4H4S con-
version. If the presence of H2 affects the removal of C4H4S, H2
must participate in a catalytic process. Otherwise, the removal
of C4H4S can be described as pure sorption.
Breakthrough of C4H4S
For a better understanding of the mechanisms of C4H4S remov-
al, a breakthrough experiment was performed under steady-
state conditions, and the selected MS traces obtained through
that experiment are shown in Figure 13. After reduction in
40% H2 for 6 min, 25 mg of the catalyst was exposed to 40%
H2 and 100 ppm C4H4S at 6008C at a total gas flow rate of
25 mLminꢀ1. At the beginning of the exposure, no H2S (m/z
33) or C4H4S (m/z 84) is detected at the reactor outlet, as ob-
served in the previous dry experiments. Approximately 2700 s
(45 min) after the start of C4H4S exposure, C4H4S is detected at
the reactor outlet, along with H2S. During the next 5 h, the sig-
nals for C4H4S and H2S rise slowly and reach a steady state
eventually. In addition, various C1ꢀC4 species are detected.
Before breakthrough of C4H4S, the signals for CH4 (m/z 15),
C2H4 (m/z 26 and 27), and C2H6 (m/z 27 and 30) are detected,
along with those for C4H8 (m/z 27, 28, and 41) and C4H10 (m/z
27, 28, 41, 42, and 43). At the time when C4H4S breaks through,
the signals for CH4, C2H4, C2H6, and C4H10 reach a maximum.
After breakthrough, the signals for CH4 and C2H6 decrease to
The integrated C4H4S signal (detected at m/z 84) for the re-
actor inlet, the reactor outlet with only C4H4S in the feed, and
simultaneous feed of H2 and C4H4S is shown in Figure 11. As
before, all experiments were performed at 6008C. In the ab-
sence of H2, approximately 40% of the C4H4S fed into the reac-
tor is measured at the outlet; hence, 60% sorbs on the
sample. On feeding H2 simultaneously, no C4H4S is detected at
the reactor outlet. This result suggests that the removal of
C4H4S over a prereduced Mo catalyst proceeds at least partially
catalytically, most likely owing to the hydrogenation of C4H4S
to H2S as well as C4H6 and other hydrocarbons. The experiment
in the absence of H2 shows that a fraction of C4H4S is removed
from the gas stream through sorption on the catalyst.
Effect of sulfidation on C4H4S removal
The decomposition of C4H4S in the presence of H2 (see the pre-
vious section) will most likely lead to the sulfidation of the Mo
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3700 – 3711 3706

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
C4H4S removal in the presence
of steam
To assess the effect of steam on
the removal of C4H4S, the cyclic
experiments of reduction, expo-
sure to C4H4S, and oxidation
(Figure 7) were repeated in the
presence of 7.7% steam in the
feed. The selected MS traces of
one representative cycle are
shown in Figure 14. In contrast
to the dry experiment, the MS
results show only partial conver-
sion of C4H4S. Directly after the
addition to the gas feed, the
C4H4S signal detected at the re-
actor outlet is approximately
60% of the signal detected at
the reactor inlet.
Figure 12. Selected MS traces obtained during the sulfidation and subsequent exposure of the Mo/Al2O3 catalyst
to H2 +C4H4S at 6008C.
The sample composition, determined by the linear combina-
tion fit of the time-resolved XAS spectra, suggests no qualita-
tive difference from that in the dry experiment (Figure 7).
Upon addition of H2, approximately 40% of the sample re-
duces from MoO3 to MoO2, which oxidizes quickly to MoO3
Figure 13. Selected MS traces obtained during the continuous exposure of
the Mo/Al2O3 catalyst to H2 +C4H4S at 6008C.
zero, the signals for C2H4 and C4H10 return to near initial levels,
and a signal for C4H6 (m/z 54) is detected.
These data indicate that over the time on stream of continu-
ous exposure to H2 +C4H4S, the selectivity toward carbon con-
version changes. In the beginning, some of the carbon rings of
C4H4S are cracked to C1 and C2 species. After C4H4S break-
through, the selectivity shifts toward C3 and C4 species. Al-
though the detailed mechanism for the selectivity toward dif-
ferent carbon species is beyond the scope of this study, the re-
sults indicate a shift in selectivity that coincides with that in
C4H4S breakthrough.
Figure 14. Cyclic exposure of the Mo/Al2O3 catalyst to reduction in H2, re-
duction in H2 +C4H4S, and oxidation in O2 at 6008C in the presence of 7.7%
steam. The selected MS traces (top panel) for the reactor inlet (dashed line)
and the reactor outlet (solid lines) are shown along with the sample compo-
sition, which are derived from the linear combination fit of the Mo K-edge
XANES spectra.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3700 – 3711 3707

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
upon oxidation in O2. The degree and speed of reduction and
oxidation are not changed significantly upon addition of
steam.
Similar to the experiment on H2S uptake (see above), the
amount of catalyst in the reactor was varied whereas all other
experimental parameters were kept constant during reduction,
exposure to C4H4S, and oxidation in the presence of steam.
The integrated signal for C4H4S at the reactor outlet is shown
in Figure 15. The MS data detected at m/z 84 (C4H4S) was nor-
Scheme 1. Representation of the mechanisms during H2S and C4H4S remov-
&
&
&
al. : MoO3, : MoO2, : MoS2.
occurs through the sulfidation of MoO3 to MoS2 under the con-
current formation of MoO2. Under oxidizing atmosphere, MoS2
is oxidized to MoO3 along with the release of SO2 [Eqs. (1) and
(2)].
Sulfidation :
ð1Þ
2 MoO3 þ 2 H2 þ 2 H2S ! MoO2 þ MoS2 þ 4 H2O
Figure 15. Integrated C4H4S signal at the reactor outlet (measured by m/z
84) for different catalyst masses under identical experimental conditions at
6008C in the presence of 7.7% steam. The data for a mass of 0 g represent
the measurement through an empty tube. For comparison, the data from
the dry experiment for a mass of 25 mg are denoted by a blank triangle.
Regeneration :
ð2Þ
2 MoS2 þ 7 O2 ! 2 MoO3 þ 4 SO2
This mechanism is in line with earlier findings, which indicat-
ed that MoO3 is sulfided to MoS2 via OꢀS exchange followed
by reduction.[38–41] Temperature-programmed sulfidation experi-
ments showed that MoO3 supported on Al2O3 can be sulfided
at 500 K via OꢀS exchange on Mo6+ ions followed by reduc-
malized to the He signal (m/z 4) and integrated over single
cycles, and the mean values and standard deviations were cal-
culated for data from 13 to 15 cycles. Although the resulting
standard deviations are relatively large, the trend shows that
with the increasing catalyst mass a lower C4H4S signal is de-
tected at the reactor outlet, which indicates that the detrimen-
tal effect of steam is due to the competitive adsorption of H2O
on the MoO2 surface, blocking sites for C4H4S to adsorb. This
effect can be compensated for by using more catalyst material.
For comparison, the data for 25 mg of the catalyst material
from the dry experiment are shown, in which no C4H4S is de-
tected at the reactor outlet.
tion to Mo6+ [38]
This finding was later supported by X-ray pho-
.
toelectron spectroscopy (XPS) and IR emission spectroscopy
studies on crystalline MoO3,[39] by XPS studies on SiO2-support-
ed MoO3 model catalysts,[40] and by XAS and TEM studies on
MoO3/SiO2 catalysts.[41]
For the removal of C4H4S, a MoO2 surface is preferred to
a MoO3 surface (Figure 10). The experiments of C4H4S removal
reveal several plausible mechanisms of the removal:
1) C4H4S can adsorb on the MoO2 surface, on which it does
not react further, until all possible sites are occupied
(sorption).
Discussion
Mechanism of sulfur removal
2) C4H4S can adsorb on the MoO2 surface, on which it is hy-
drogenated by H2 to H2S and hydrocarbons, mostly C4 and
only small amounts of C1ꢀC3 species (hydrogenolysis).[42]
Subsequently, the generated H2S reacts with MoO2 and
form MoS2 and H2O.
The results for the reactivity (measured by using MS) and struc-
ture of the Mo catalyst (measured by using XAS) enable us to
propose mechanisms of sulfur removal. These proposed mech-
anisms of the removal of H2S and C4H4S over Mo are illustrated
in Scheme 1 and discussed below.
3) The adsorbed C4H4S can directly sulfide the Mo site on
which it adsorbs and forms C1ꢀC4 hydrocarbons (direct
desulfurization).
From the experiments on H2S uptake (Figures 2–4), it can be
concluded that the removal of H2S from the producer gas
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3700 – 3711 3708

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
In mechanism 1, the only measured species after all Mo sites
are occupied would be C4H4S. Mechanism 2 would lead to the
detection of H2S and hydrocarbons after breakthrough, where-
as mechanism 3 would result in the breakthrough of C4H4S
and hydrocarbons.
Effect of steam
Steam, which is always present in the producer gas from gasi-
fied biomass, has a detrimental effect on the desulfurization
over Mo/Al2O3 catalysts. However, the reduction/oxidation of
Mo is not significantly affected by the presence of steam (Fig-
ures 6 and 7), which is in line with the findings from ex situ ex-
periments using XPS.[23] In contrast, it seems that H2O adsorbs
competitively on the surface to H2S and C4H4S, which in effect
slows down the desulfurization reactions. Although the de-
creased activity can be compensated for by using more cata-
lyst material, the activity in a real system will depend on the
process parameters (temperature, steam content, etc.) and the
catalyst formulation (Mo loading, dispersion, addition of
promoters, etc.).
The presented data indicate that the mechanism of C4H4S re-
moval is not pure sorption (mechanism 1) of the C4H4S mole-
cule on the MoO2 surface, because H2 is required to completely
remove C4H4S from the gas (Figure 11), indicating a catalytic
process. The demodulation of the XAS spectra (Figure 9) sug-
gests that a fraction of the catalyst undergoes transition from
MoO2 to MoS2 during exposure to H2 and C4H4S.
The breakthrough of C4H4S with some production of H2S in-
dicates that hydrogenolysis is not the prevailing mechanism of
C4H4S removal over MoO2. If that was the case, as it is for pre-
sulfided Mo (Figure 12), only H2S would be observed in the
breakthrough experiment, but no C4H4S.
Suitability of Mo catalysts for hot gas desulfurization
Thus, C4H4S is removed from MoO2 mainly through the
direct desulfurization of C4H4S. Sulfur-free surfaces are covered
initially with C4H4S, which completely removes C4H4S from the
gas. In the presence of H2, the adsorbed C4H4S then sulfides
MoO2 whereas hydrocarbons are released. The observation of
CH4 in this initial phase indicates that the carbon structure is
completely cracked to C1 species, possibly owing to the pres-
ence of acid sites on MoO2.[43] This process occurs at a rate
slower than the feed rate of C4H4S, which leads to the ob-
served breakthrough of C4H4S [Eqs. (3) and (4)].
The presented results indicate that Mo/Al2O3 is a suitable cata-
lyst for hot gas desulfurization. At 6008C, H2S and C4H4S, which
are used here as model compounds for sulfur species in the
producer gas, are removed under dry conditions. Over the 10–
15 cycles under a reducing and oxidizing atmosphere, the cata-
lyst is stable, as indicated by the stable levels of SO2, which is
removed upon oxidation (see Figure 4), enabling long-term
use for desulfurization. The adjustment of the residence time
of the catalyst in the sulfur-containing producer gas will avoid
the breakthrough of C4H4S. In a real-size plant, the desulfuriza-
tion over the Mo catalyst is likely to be implemented after
a high-temperature filtration step to avoid interaction of the
particulate matter with the catalyst.
Direct desulfurization :
ð3Þ
H2 þ C4H4S þ MoO2 ! MoS2 þ CH4 þ C2ꢀC4 species
The promotion of the catalyst with other metals, such as Co,
can be an option to minimize the negative effect of steam. A
comparison of the unpromoted Mo/Al2O3 catalyst with CoMo/
Al2O3 catalysts showed that the former was more prone to the
inhibition of hydrodeoxygenation by steam than the latter.[25]
In contrast, promotion with Ni was found to enhance the neg-
ative effect of steam owing to the oxidation of nickel sul-
fides.[24] Therefore, the development of the dedicated materials
for the improved desulfurization of the producer gas must
take the effect of H2O into account.
Hydrogenolysis :
ð4Þ
H2 þ MoS2 þ C4H4S ! MoS2 þ H2S þ C4 species
The MoS2 phase formed through sulfidation in H2S is active
for the removal of C4H4S in the presence of H2 (Figure 12),
which forms C4H6 and H2S.
As the Mo catalyst removes H2S through sorption and subse-
quent sulfidation and C4H4S through a catalyzed reaction, we
consider it a bifunctional material. In this context, bifunctionali-
ty is defined as the capability of the material to remove sulfur
from the gas stream through two mechanisms: sorption and
catalysis.
Conclusions
The suitability of Mo catalysts supported on Al2O3 for a hot-
gas removal of sulfur in the absence and presence of steam
was investigated. Although the activity for sulfur removal was
studied by using MS, the state of Mo was probed by using
time-resolved in situ X-ray absorption spectroscopy at the Mo
K-edge. Experiments at 6008C showed that under dry condi-
tions, H2S, which is the most abundant sulfur species in the
producer gas, and C4H4S, which is used as a model compound
for organic sulfur species, are completely removed by Mo. The
presence of steam is detrimental for the removal of H2S and
C4H4S from the gas owing to the competitive adsorption of
steam on the Mo surface, which blocks sites for sulfur
adsorption.
The direct mechanistic insight that is presented in this con-
tribution is possible only owing to the use of time-resolved
in situ XAS. Although in situ studies using other techniques
will suffer from small particle size (not possible using XRD),
large amounts of steam (not possible using IR spectroscopy),
or ambient pressure (not possible using XPS), XAS is a tool
that is complementary to the routinely used gas analysis. Al-
though gas analysis provides direct information only on the
catalyst activity, its structure can be directly observed by using
XAS.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3700 – 3711 3709

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Swiss Light Source using a QuickXAS monochromator with a chan-
nel-cut Si (111) crystal that was rotated at a frequency of 0.2 Hz,
which resulted in 5 s time resolution per spectrum. The X-ray
energy was calibrated with use of a Mo metal foil placed down-
stream of the sample. The beam intensity was monitored with ion
chambers filled with N2 (I0, before the sample) and Ar (I1 and I2,
after the sample and the reference foil, respectively). The X-ray
beam was collimated by a Pt-coated mirror upstream of the mono-
chromator and was focused to approximately 100 mm2 with a Pt-
coated toroidal mirror placed downstream of the monochromator.
The ion chamber signals were amplified, the angular encoder
signal from the monochromator was converted to analog voltage,
and all signals were sampled by an analog-to-digital converter at
1 kHz. The resulting data were averaged over multiple cycles, and
the encoder signal was converted into an energy scale with
custom MATLAB scripts. Normalization, energy calibration, LCF, and
EXAFS fitting were performed with IFEFFIT.[44] The EXAFS fitting of
the fresh catalyst sample was performed over a k range of 3–
14 ꢁꢀ1 in k2 weighting. The value for the amplitude reduction
The mechanisms of sulfur removal were proposed on the
basis of the measurements of reactivity and structure of the
Mo catalyst; H2S is removed through the sulfidation of MoO3
to MoS2 with the concurrent reduction of MoO3 to MoO2. The
removal of C4H4S occurs over MoO2 through the hydrogenoly-
sis of C4H4S, which forms MoS2 and C1ꢀC4 species. This process
is, however, slower than the catalytic hydrogenation of C4H4S
over MoS2. Subsequent oxidation to SO2 enables the sequestra-
tion of sulfur from the producer gas stream and regeneration
of the catalyst.
Herein, we showed that Mo-based catalysts can be used for
sulfur removal from biomass-derived gas at high temperatures,
which removes H2S and organic sulfur compounds. This pro-
cess improves the efficiency of catalytic biomass conversion.
Experimental Section
2
factor (S0 ), 0.87, was derived by fitting a spectrum of the Mo foil,
Materials
which was acquired under identical experimental conditions. LCF
was performed over a range of ꢀ20 to +50 eV, which was relative
to the absorption edge. All components used the same energy
shift E0 and were forced to have a weight between 0 and 1.
Summing of the components to 1 was not enforced.
The materials were prepared through repeated wetness impregna-
tion. Ammonium heptamolybdate tetrahydrate [(NH4)6Mo7O24
·4H2O, Sigma–Aldrich] was dissolved in deionized water and im-
pregnated on the g-Al2O3 support (Sasol). Excess water was evapo-
rated slowly under vacuum, and the impregnated support was
washed with deionized water. Finally, the catalyst was dried at
110 8C for 1 h and calcined at 6008C for 6 h. This process was re-
peated 4 times to increase the Mo loading. The sample
(ꢁ25 mg) was used for the in situ experiments with H2S and C4H4S.
Reference compounds MoO2, MoO3, and MoS2 were purchased
from Sigma–Aldrich, diluted in cellulose, and pressed into pellets.
Samples were characterized by N2 physisorption (Micromeritics
TriStar II), XRD (Brucker D8), ICP–OES (Varian Vista Pro AX), and
STEM. STEM was performed on a Hitachi HD2700C STEM micro-
scope.
Acknowledgements
We thank F. Krumeich and D. Fodor from ETH Zꢁrich for STEM
measurements, S. Kçchli for ICP–OES analysis, M. Steib for assis-
tance during beamtime, E. de Boni for technical assistance, the
Swiss Light Source for providing beamtime. C.F.J.K. acknowledges
the Competence Center for Energy and Mobility for funding.
Keywords: biomass · desulfurization · in situ spectroscopy ·
molybdenum · X-ray absorption spectroscopy
Experimental setup
The reactor for the in situ experiments consisted of a quartz capil-
lary (3 mm outer diameter and 0.1 mm wall thickness; Hilgenberg)
that was heated by a hot air blower (Leister). The samples were
fitted in the reactor between two plugs of quartz wool. The tem-
perature inside the capillary was measured with a thermocouple
placed downstream of the bed and was kept at 6008C. The pres-
sure inside the reactor was kept constant at 0.5 bar overpressure
with pressure regulating valves (Bronkhorst). A computer-con-
trolled setup was used to mix gases with mass flow controllers
(Bronkhorst). All gas lines were coated with amorphous SiO2 to
minimize sulfur adsorption (Sulfinert, Restek). Gas analysis was per-
formed with a quadrupole mass spectrometer (Max 300-LG, Extrel).
Deionized water was injected into the gas stream with a syringe
pump (Harvard apparatus) through a fused silica capillary into
a crosspiece in a heated box. As this method has limited dynamics
of starting and stopping the water feed, steam flow was constant
during the experiments with steam. The transfer lines between the
crosspiece, the reactor, and the mass spectrometer were heated to
avoid condensation.
[1] A. V. Bridgwater, Appl. Catal. A 1994, 116, 5–47.
[2] H. de Lasa, E. Salaices, J. Mazumder, R. Lucky, Chem. Rev. 2011, 111,
5404–5433.
[3] A. van der Drift, J. van Doorn, J. W. Vermeulen, Biomass Bioenergy 2001,
20, 45–56.
[4] W. Torres, S. S. Pansare, J. G. Goodwin, Catal. Rev. 2007, 49, 407–456.
[5] H. Cui, S. Q. Turn, V. Keffer, D. Evans, T. Tran, M. Foley, Energy Fuels 2010,
24, 1222–1233.
[6] K. Natesan, Corrosion 1985, 41, 646–655.
[7] S. Mrowec, Oxid. Met. 1995, 44, 177–209.
[8] P. K. Agrawal, J. R. Katzer, W. H. Manogue, J. Catal. 1982, 74, 332–342.
[9] W. D. Fitzharris, J. R. Katzer, W. H. Manogue, J. Catal. 1982, 76, 369–384.
[10] J. Hepola, P. Simell, Appl. Catal. B 1997, 14, 305–321.
[11] K. Okabe, K. Murata, M. Nakanishi, T. Ogi, M. Nurunnabi, Y. Liu, Catal.
Lett. 2009, 128, 171–176.
[12] S. D. Phillips, Ind. Eng. Chem. Res. 2007, 46, 8887–8897.
[13] M. M. Yung, W. S. Jablonski, K. A. Magrini-Bair, Energy Fuels 2009, 23,
1874–1887.
[14] K. Engvall, H. Kusar, K. Sjoestroemm, L. J. Pettersson, Top. Catal. 2011,
54, 949–959.
[15] R. D. Solunke, G. Veser, Fuel 2011, 90, 608–617.
[16] R. K. Srivastava, W. Jozewicz, C. Singer, Environ. Prog. 2001, 20, 219–
228.
[17] Y. Mathieu, L. Tzanis, M. Soulard, J. Patarin, M. Vierling, M. Moliꢂre, Fuel
Process. Technol. 2013, 114, 81–100.
Setup for XAS
The in situ XAS experiments at the Mo K-edge (20.0 keV) were per-
formed in the transmission mode at the SuperXAS beamline at the
[18] J. H. Swisher, J. Yang, R. Gupta, Ind. Eng. Chem. Res. 1995, 34, 4463–
4471.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3700 – 3711 3710

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
[19] S. Cheah, D. L. Carpenter, K. A. Magrini-Bair, Energy Fuels 2009, 23,
5291–5307.
[33] E. P. Barrett, L. G. Joyner, P. P. Halenda, J. Am. Chem. Soc. 1951, 73, 373–
380.
[20] X. Meng, W. de Jong, R. Pal, A. H. M. Verkooijen, Fuel Process. Technol.
2010, 91, 964–981.
[21] H. Topsøe, B. S. Clausen, F. E. Massoth Catalysis, Science and Technology,
Springer, Berlin, 1996, p. 1–269.
[22] P. McKendry, Bioresour. Technol. 2002, 83, 37–46.
[23] T. A. Patterson, J. C. Carver, D. E. Leyden, D. M. Hercules, J. Phys. Chem.
1976, 80, 1700–1708.
[24] E. Laurent, B. Delmon, J. Catal. 1994, 146, 281–291.
[25] M. Badawi, J. F. Paul, S. Cristol, E. Payen, Y. Romero, F. Richard, S. Brunet,
D. Lambert, X. Portier, A. Popov, E. Kondratieva, J. M. Goupil, J. El Fallah,
J. P. Gilson, L. Mariey, A. Travert, F. Mauge, J. Catal. 2011, 282, 155–164.
[26] I. I. Novochinskii, C. Song, X. Ma, X. Liu, L. Shore, J. Lampert, R. J. Farrau-
to, Energy Fuels 2004, 18, 576–583.
[27] R. Prins, V. H. J. de Beer, G. A. Somorjai, Catal. Rev. Sci. Eng. 1989, 31, 1–
41.
[28] P. T. Vasudevan, L. G. Fierro, Catal. Rev. Sci. Eng. 1996, 38, 161–188.
[29] H. Topsøe, B. S. Clausen, Catal. Rev. Sci. Eng. 1984, 26, 395–420.
[30] T. Ressler, J. Wienold, R. E. Jentoft, T. Neisius, M. M. Guenter, Top. Catal.
2002, 18, 45–52.
[34] T. Grçbl, H. Walter, M. Haider, Appl. Energy 2012, 97, 451–461.
[35] D. Ferri, M. S. Kumar, A. Eyssler, O. Korsak, P. Hug, A. Weidenkaff, M. A.
Newton, Phys. Chem. Chem. Phys. 2010, 12, 5634–5646.
[36] D. Ferri, M. A. Newton, M. Nachtegaal, Top. Catal. 2011, 54, 1070–1078.
[37] C. F. J. Kçnig, J. A. van Bokhoven, T. J. Schildhauer, M. Nachtegaal, J.
Phys. Chem. C 2012, 116, 19857–19866.
[38] P. Arnoldy, J. A. M. van der Heijkant, G. D. de Bok, J. A. Moulijn, J. Catal.
1985, 92, 35–55.
[39] Th. Weber, J. C. Muijsers, J. H. M. C. van Wolput, C. P. J. Verhagen, J. W.
Niemantsverdriet, J. Phys. Chem. 1996, 100, 14144–14150.
[40] J. C. Muijsers, Th. Weber, R. M. van Hardeveld, H. W. Zandbergen, J. W.
Niemantsverdriet, J. Catal. 1995, 157, 698–705.
[41] A. J. de Boer, van Dillen, D. C. Koningsberger, J. W. Geus, J. Phys. Chem.
1994, 98, 7862–7870.
[42] D. L. Sullivan, J. G. Ekerdt, J. Catal. 1997, 172, 64–75.
[43] N. Sotani, M. Hasegawa, Bull. Chem. Soc. Jpn. 1973, 46, 25–29.
[44] M. Newville, J. Synchrotron Radiat. 2001, 8, 322–324.
Received: March 14, 2013
Published online on September 9, 2013
[31] E. M. McCarron, J. C. Calabrese, J. Solid State Chem. 1991, 91, 121–125.
[32] S. Brunauer, P. H. Emmett, E. Teller, J. Am. Chem. Soc. 1938, 60, 309–319.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3700 – 3711 3711

Copyright of ChemCatChem is the property of Wiley-Blackwell and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.