Applied Catalysis B: Environmental xxx (xxxx) xxxx

Contents lists available at ScienceDirect

Applied Catalysis B: Environmental

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

Efficient recovery of hydrogen and sulfur resources over non-sulfide based

LaFexAl12-xO19 hexaaluminate catalysts by H2S catalytic decomposition
Guoxia Jianga,b, Xin Zhanga,c,*, Fenglian Zhanga,b, Zizhong Liue, Zhuo Wangf, Zhengping Haoa,b,c,
Caihong Lind
a
National Engineering Laboratory for VOCs Pollution Control Material & Technology, University of Chinese Academy of Sciences, Beijing, 101408, PR China
b
Key Laboratory of Environmental Nanotechnology and Health Effects, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing, 100085,
PR China
c
Research Center for Environmental Material and Pollution Control Technology, University of Chinese Academy of Sciences, Beijing, 101408, PR China
d
Center of Research & Development, Shandong Sunway Petrochemical Engineering Share Co., Ltd, Beijing, 100015, PR China
e
Chemistry and Environment Science College, Inner Mongolia Normal University, Inner Mongolia Key Laboratory of Green Catalysis, Hohhot, Inner Mongolia 010022, PR
China
f
Beijing Institute of Nanoenergy and Nanosystems, Chinese Academy of Sciences, Beijing, 100083, PR China

A R T I C LE I N FO A B S T R A C T

Keywords: Simultaneous recovery of hydrogen and sulfur resources from acidic gas hydrogen sulfide (H2S), whose content
Hydrogen resource may reach 20–100 % after hydrodesulfurization in oil refineries, is increasingly attractive. Nowadays, the
Sulfur recover widespread utilization of H2S decomposition technology is mainly restricted by the low hydrogen and sulfur
H2S yield, i.e. 20 % at 800 °C. In this work, for the first time a novel non-sulfide based LaFexAl12-xO19 hexaaluminate
Active phase
catalysts was used for high (50 %) hydrogen yield at 800 °C. At the same time, the active phases and reaction
Catalytic mechanism
mechanism were confirmed by various characterization techniques. The XRD, Raman, XPS, Mössbauer and XAFS
results verified that the Fe species in the hexaaluminate structure, especially octahedral-coordinated Fe3+,
served as the main active phases. The reaction routes/mechanism for H2S decomposition were also verified by
density functional theory (DFT) method. H2S was firstly adsorbed on the active sites and then completed the
decomposition by direct dehydrogenation mechanism. Briefly, current study has minimized the challenges for
H2S decomposition at industrial scale. It is anticipated that our findings will offer potential application of H2S
decomposing technology.

1. Introduction H2S decomposition [5]:

2H2S ↔ 2H2 + 1/4S8, ΔH =79.5 kJ/mol (1)

Hydrogen sulfide (H2S) is a flammable, corrosive and very poiso-
nous gas, which could be harmful to living organisms even at low
concentrations. However, except for relatively small amounts, i.e.
0.001–16 %, produced by natural sources, H2S is widely generated in
chemical industries [1]. From hydrodesulfurization in oil refineries, the
content of H2S may reach 20–100 % [2]. The famous Claus process is
extensively applied to H2S treatment, during which H2S is converted
into sulfur and water [3]. Meanwhile, it is worthy pointing out that
hydrogen (H2), as alternative energy source and important chemical
feedstock, is needed due to its increasing demand [4]. Therefore, taking
seriously for both the sustainability of resource and the fulfillment of
stringent environmental requirements, an attractive alternative could
be the recovery of hydrogen and elemental sulfur simultaneously by
In recent years, various methods have been explored and tested on
H2S decomposition, including thermolysis, electrolysis, photolysis,
plasmolysis and many their variants [6–10]. Among these, the most
promising process is thermal decomposition. However, the H2S de-
composition reaction is highly endothermic and the equilibrium con-
version is low even at high temperatures (only 20 % H2 yield at
1000 °C) [11]. Therefore, certain amounts of studies have been ongoing
to overcome and tackle the drawback by employing suitable catalysts to
improve the H2S conversion and hydrogen yield.
Various catalysts for H2S decomposition based on metal sulfides,
disulfides and oxides have been investigated. For metal sulfides, mainly
focused on FeS, CuS and NiS, however, maximum H2 yield attained was

⁎
Corresponding author at: National Engineering Laboratory for VOCs Pollution Control Material & Technology, University of Chinese Academy of Sciences, Beijing,
101408, PR China.
E-mail address: xzhang@ucas.ac.cn (X. Zhang).

https://doi.org/10.1016/j.apcatb.2019.118354
Received 25 August 2019; Received in revised form 21 October 2019; Accepted 27 October 2019
Available online 04 November 2019
0926-3373/ © 2019 Elsevier B.V. All rights reserved.

Please cite this article as: Guoxia Jiang, et al., Applied Catalysis B: Environmental, https://doi.org/10.1016/j.apcatb.2019.118354
G. Jiang, et al.
only 20 % at 770 °C [12–14]. Notably, metal disulfides (such as MoS2,
WS2, FeS2, CoS2 and NiS2) and metal oxides (mainly focused on per-
ovskites [15–17]) presented analogous results, maximally receiving 20
% H2 yield at 800 °C [16,18]. Thus, the decomposition activity of H2S
was still low and unsatisfactory. And the relatively low H2 yield ap-
peared to be the mainly restricted parameter for the widespread utili-
zation of the decomposition catalysts and the H2S decomposition pro-
cess.
More importantly, the reaction processes for metal oxides catalysts
were partly based on “two-step process” [10]. The key point of this
mechanism was that metal oxides were sulfided to sulfides, which were
the actual catalysts:
MxOy + yH2S → MxSy + yH2O, (2)
And then, the “two-step process” associated with metal sulfides
could happen. A “lowest” metal sulfide MxSy has affinity with sulfur,
forming the “highest” sulfide MxSy+z and hydrogen. In their turn, the
“highest” sulfides can give back the corresponding “lowest” sulfide by
thermal decomposition:
MxSy + zH2S → MxSy+z + zH2, (3)
MxSy+z → MxSy + zS0, (4)
H2S → H2 + S0. (5)
While the main drawbacks of the two-step processes were the
handling of large mass of sulfides, and the need to cool and re-heat
them because of the differences in operating temperatures of the two
steps. Meanwhile, due to the phases change, the real catalysts were the
formed intermediate sulfides but not oxides. Thus, metal oxides could
not be recovered after reaction.
Therefore, it is extremely desired to explore new type H2S decom-
position catalysts with new reaction routes, excellent catalytic activity
and high thermal stability by taking into account the fact that the de-
composition reaction is high temperature favorable thermodynamically
(almost higher than 800 °C).
Hexaaluminates consisting of alternatively stacked spinel blocks
and mirror planes have been regarded probably as the most promising
catalytic materials for high-temperature applications due to their ex-
ceptional thermal stability in combination with extraordinary sub-
stituting capability [19,20]. The general formula can be presented as
ABxAl12−xO19, A represents a mono-, di-, or trivalent large cation re-
siding in the mirror plane, and the component B could be a transition-
metal (Mn, Fe, Co, Cu, Ni, etc.) or a noble metal ion (Ir, Ru, Pd, Rh),
which can partially or even completely substitute Al crystallographic
sites [21]. Particularly, the stable spinel structure made sulfidation
reactions difficult, just like the results on perovskites [16,22]. There-
fore, the new reaction route possibility for H2S catalytic decomposition
was provided.
In this study, the active Fe species and the remarkably outstanding
thermostability of hexaaluminates were combined. And Fe substituted
hexaaluminates catalysts LaFexAl12-xO19 (LaFex, x = 2, 4, 6, 8, 10 and
12) were synthesized and applied for H2S decomposition reaction. The
results shown that as high as 50% hydrogen yield was achieved, which
has never been reported before. Subsequently, the main active phases
were considered to be the Fe species presented in the hexaaluminate
structure, especially octahedral-coordinated Fe3+. In addition, it was
verified that, in contrast to traditional two-step process, H2S was firstly
adsorbed on the active sites and then completed the decomposition by
direct dehydrogenation processes.
2. Experimental
2.1. Catalyst preparation and activity test
Samples with nominal composition LaFexAl12-xO19 (LaFex, x = 2, 4,
2
Applied Catalysis B: Environmental xxx (xxxx) xxxx
6, 8, 10 and 12) were prepared using the carbonate route. Typically, to
prepare LaFe2Al10O19 (LaFe2, x = 2), La(NO3)3, Fe(NO3)3·9H2O, and Al
(NO3)3·9H2O were dissolved individually in 60 °C deionized water in
proportion. Except aluminum nitrate, metal nitrates were mixed to-
gether firstly to form clear solutions, followed by adjusting the pH = 1
with nitric acid. Then, the aluminum nitrate solution was added into
the metal nitrates mixture. This mixed solution was poured under
vigorous stirring into a saturated (NH4)2CO3 solution heated at 60 °C to
form precursor precipitate. During the precipitation process, the pH
value of the solution was maintained between 7.5 and 8.0. After con-
tinuous stirring at 60 °C for 4 h and aging for 3 h, the precipitate was
filtered, washed, and then dried at 120 °C overnight. The obtained
precursor materials were calcined by step-wise in the temperature
range 500–1200 °C for 6 h under air condition. All prepared samples
were named as LaFex, such as LaFe6 (LaFe6Al6O19).
Catalytic decomposition of H2S was tested in a vertical fixed bed
quartz reactor placed in an external heating furnace at atmospheric
pressure. In a typical experimental run, 0.5 g of catalysts (20–40 mesh)
was loaded between quartz wool in the center of the quartz tube. The
typical compositions of feed gas were 1000 ppm H2S/N2, and the total
flow rate of the reaction mixture was 200 mL/min with a gas hourly
space velocity (GHSV) of 24000 h−1. The outlet gas stream from fur-
nace was first passed through a condenser to trap the sulfur gas in ef-
fluent stream, and then were analyzed via an online gas chromato-
graphy (GC) equipped with a thermal conductivity detector (TCD, with
Plot/Q columns) and a flame photometric detector (FPD, with GasPro
columns). And the carrier gas of GC was nitrogen (N2). Before entering
TCD detector, the outlet gas stream was passed through a gas bubbler
containing saturated aqueous Cu (II) acetate solution to remove un-
converted H2S.
The conversion of hydrogen sulfide and the hydrogen yield were
calculated in accordance with equation:
(H2 S)in − (H2 S)out
H2 S conversion = × 100%
(H2 S)in (6)
(H2 )out
H2 yield = × 100%
(H2 S)in (7)
2.2. Catalyst characterization
The structure of the samples were investigated using powder X-ray
diffraction (XRD, PAN analytical X’Pert-Pro, Cu Kα radiation, λ
=1.5418 Å) operated at 40 kV and 40 mA in the 2θ range of 10-90° with
5°/min scanning rate. The morphology and chemical composition of
prepared samples were examined using field emission scanning electron
microscopy with energy dispersive X-ray spectroscopy (FESEM/EDX,
JSM-6700 F), transmission electron microscopy and high-resolution
transmission electron microscopy (TEM/HRTEM, JEM-2100 F, 200 kV).
Raman spectra were performed at room temperature using a micro-
Raman spectrometer (HORIBA, France) with a 532 nm Ar+ laser as the
excitation source in a backscattering geometry. The laser power in-
cident on the catalysts was less than 10 mW. The time of acquisition
was varied according to the intensity of the Raman scattering.
Surface compositions of the samples were investigated with X-ray
photoelectron spectroscopy (XPS, Thermo ESCALAB 250, USA) using a
monochromatic Al Kα exciting radiation (1486.71 eV). The shift of the
binding energy due to relative surface charging was corrected using the
C1 s level at 284.8 eV as an internal standard. Near-surface atomic
concentrations of the fresh and spent materials were calculated ac-
cording to peak areas and sensitivity factors.
The 57Fe Mössbauer spectra of the as-prepared materials were re-
corded using a Topologic 500A spectrometer and a proportional
counter at 80 K with a spectrometer working at constant acceleration
using a Rh57Co source. 57Fe Mössbauer spectral parameters such as the
isomer shift (IS), the electric quadrupole splitting (QS) and the relative
G. Jiang, et al.
spectral areas (A) of the different components of the absorption patterns
were evaluated. All IS values were reported with respect to metallic
iron (α-Fe) at the measurement temperature. Data analysis was per-
formed using a nonlinear least square fitting routine that models the
spectra as a combination of singlets, quadruple doublets and magnetic
sextuplets based on a Lorentzian line shape profile.
The powder samples were prepared for X-ray Absorption Fine
Structure (XAFS) specimens with thickness appropriate for Fe K-edge
measurements. The XAFS measurements were carried out at the Beijing
Synchrotron Radiation Facility (BSRF). Fe foil was used as a reference
sample, and all spectra were measured in the transmission mode at
room temperature.
2.3. Theoretical calculation
The DFT calculations were conducted using CASTEP with a plane
wave basis set. The Perdew-Burke-Ernzerhof (PBE) functional with the
generalized gradient approximation (GGA) was applied to calculate the
exchange-correlation energy. The ion core and valence electron inter-
action was described by Vanderbilt-type ultrasoft pseudopotential. A
plane wave cut-off energy of 380 eV was used in the calculation.
Brillouin zone sampling was calculated using a Monkhorst-Pack grid
with respect to the symmetry of the system. The convergence criteria
for configuration optimization were set to the tolerance for SCF, energy,
maximum force, with a maximum displacement of 1 × 10−6 eV/atom,
2 × 10−5 eV/atom, 0.05 eV/Å and 2 × 10−3Å, respectively.
3. Results and discussion
3.1. Catalytic performance of LaFex catalysts
The catalytic performance of LaFex catalysts for H2S decomposition
were displayed in Fig. 1a and b. As shown, the LaFe6 samples were the
most effective catalysts, which could present as high as 50 % H2 yield at
800 °C. To the best of our knowledge, it was remarkably higher than the
previous results reported in literatures (Table S1). Especially, as an
oxide catalyst, Fe-substituted hexaaluminate seems to be more active
than perovskites with Cr and/or Co occupying B-site, which has the best
known catalystic activity (20 % H2S conversion at 800 ℃) [16,17].
Besides, the two-cycle durable experiments (Fig. 1c) with 10 h for
per cycle were performed over LaFe6 catalysts at 800 °C. And the results
revealed that the LaFe6 catalysts exhibited excellent stability, which
could proceed consecutively for two cycles (20 h) without notable de-
activation. It seemed not appear to be a situation where the resulting
produced sulfur covered the active sites and reduced activity.
Particularly, some interesting phenomena were also observed. As
shown in Fig. 1a, the H2 yield increased gradually with the rise of Fe
substitution from 2 to 6. Subsequently, however, the H2 yield drasti-
cally decreased upon the growth of Fe substitution from 8 to 12. At the
same time, unexpected behaviors were obtained in the case of H2S
conversion in Fig. 1b. For LaFe2, LaFe4 and LaFe6 catalysts, the H2S
conversion gradually increased with the rise of Fe substitution and
temperature. Whereas, the H2S conversion for LaFe8, LaFe10 and LaFe12
catalysts increased greatly (almost higher than 50 %), expressing an
opposite trend compared with H2 yield. Apart from these, in Fig. 1c, an
initially increased stage of H2 yield (almost 4 h) was observed for the
first cycle. However, the increasing stage vanished in the second cycle
and the H2 yield was maintained at 50% throughout the second cycle.
Overall, according to the catalytic performance, several critical is-
sues need to be explored. Such as why the LaFex catalysts with lower Fe
substitution (x = 2, 4 and 6) could achieve higher H2 yield than those
with higher Fe substitution (x = 8, 10 and 12)? Why the H2S conversion
of LaFex catalysts were always higher than the corresponded H2 yield at
each specific temperature, especially for the LaFe8, LaFe10 and LaFe12
catalysts (almost lower than 30 % H2 yield with respect to higher than
50% H2S conversion)? Why an initially increased stage of H2 yield on
3
Applied Catalysis B: Environmental xxx (xxxx) xxxx
durable experiments emerged in the first cycle and then vanished in the
second cycle? Therefore, it is strongly desired to elucidate the inter-
correlation between the catalysts structure, properties, active phase and
the extraordinarily excellent catalytic performance as well as the cat-
alytic mechanism.
3.2. The chemical state of Fe species in LaFex catalysts
The X-ray powder diffraction (XRD) patterns of LaFex catalysts were
provided in Fig. 2a and b. As illustrated in Fig. 2a, typical diffraction
peaks corresponding to a well crystallized magnetoplumbite (MP,
JCPDS 36–1314, marked with β) phase were the dominant features of
LaFe2, LaFe4 and LaFe6 catalysts. In the case of LaFe8 catalysts, several
weak reflections associated with the impurities of α-Fe2O3 (JCPDS 84-
0310) and LaFeO3 (JCPDS 75-0439) were simultaneously detected.
What’s more, the neat patterns were even predominated by the reflec-
tions related to α-Fe2O3 and LaFeO3 with Fe substitution increased
further to 10 and 12. Especially for LaFe12 catalysts, monophasic MP
peaks were not detected.
The phase evolution of LaFex catalysts due to Fe substitution could
be clearly revealed from the 30–40° (2θ) range of the XRD patterns. As
shown in Fig. 2b, the peak positions corresponding to MP phase
(marked with β) continuously shifted to lower diffraction angles with
the x increased from 2 to 10, which mainly resulted from the enlarge-
ment of the cell parameters of MP hexaaluminate. In other words, the
iron content incorporated into the hexaaluminate lattice increased with
the Fe substitution. Furthermore, the intensity of peaks related to MP
phase gradually weakened as Fe substitution increased, indicating that
excessive substitution would essentially hindered the formation of
monophasic MP phase [23].
Raman spectroscopy is a powerful means for the conformation of
structural phase transitions, moreover, to determine phases and struc-
ture of multioxide systems. As shown in Fig. 2c, the LaFex catalysts with
x below 8 presented similar features. In the case of LaFe2 catalysts, the
dominated broad band around 800 cm−1 was mainly originated from
the motions of the tetrahedral and bipyramidal group in hexaaluminate
crystal structure [24]. However, the characteristic band gradually
shifted to lower wavenumbers with the rise of Fe substitution to 8. The
shifted modes in the 600–700 cm−1 region could ascribe to the
stretching modes of octahedral units in hexaaluminate structure, which
became rather obvious along with the substitution increased [24].
Hence, it indicated that the hexaaluminate structure with the Fe sub-
stitution up to 8 could be maintained. Furthermore, it was evident that
the LaFe10 and LaFe12 catalysts exhibited distinctly different char-
acteristics with hexaaluminate structure. The band at 217, 241, 286,
398 and 489 cm−1 were related to the presence of α-Fe2O3 [25].
Moreover, the emerging band at 432 and 608 cm−1 for LaFe12 catalysts
belonged to the orthorhombic structure of LaFeO3 [26]. These results
were well aligned with the above XRD analysis.
Fig. 3 displayed the transmission electron microscope (TEM) images
of the LaFe6 samples. And the figures (Fig. 3a) verified that the LaFe6
samples had a typical hexxalluminate sheet structure, which was more
evident in high-resolution image (HRTEM) and selected-area electron
diffraction (SAED) characterizations. The lattice fringes in the HRTEM
image (Fig. 3b) and discrete electron diffraction points (insert image in
Fig. 3c) clearly proved that the samples were highly crystallized. The
inter-planar spacing derived from SAED spectra was calculated to be
0.56 nm, corresponding to the crystal face of (004) (JCPDS 36–1314).
57
Fe Mössbauer spectroscopy was employed at 80 K to characterize
the iron state in the as-prepared materials. The fitted hyperfine para-
meters, namely, isomer shift (IS), quadrupole splitting (QS) and mag-
netic hyperfine field (H) were collected in Table 1. And the spectra
fitting results were displayed in Fig. 4.
For LaFe2, LaFe4 and LaFe6 catalysts, it was clear that the H values
in Table 1 were all bellowing 520 KOe, testifying that monophasic MP
structure was presented. Namely, Fe species exist in the MP crystal
G. Jiang, et al.
Fig. 1. Catalytic performance of LaFex catalysts. a H2 yield and b H2S conversion of the LaFex catalysts as a function of temperature in the feed gas of 1000 ppm H2S/
N2 (24000 h−1). c Durability experiment results with two cycles on LaFe6 catalysts at 800 °C in the feed gas of 1000 ppm H2S/N2 (24000 h−1).
lattice by replacing the Al sites [27–29]. In addition, the IS values for
LaFe2, LaFe4 and LaFe6 catalysts were 0.18–0.5 mm/s, indicating the
unique presence of Fe3+ species [30–33]. It was verified by the Fe 2p X-
ray photoelectron spectroscopy (XPS) analysis (Fig. S1a) as well.
On the other hand, it has been widely accepted that there are three
different kinds of crystallographic sites in the MP-type La-hex-
aaluminate structure, which are tetrahedral (Al2), octahedral (Al1, Al4,
Al3) and bipyramidal (Al5) sites [34]. As for the fitting results in Table 1
related to LaFe2, LaFe4 and LaFe6 catalysts, the IS < 0.3 mm/s illu-
strated the existence of Fe3+ ions in tetrahedral (Th) coordination
[35,36]. More detailed, when IS < 0.3 mm/s, the highly distorted tri-
gonal bipyramidal (Tr) Al5 sites were expected to show higher QS value
than that observed for the symmetric tetrahedral Al2 sites in the spinel
block. Subsequently, the 0.3 mm/s < IS < 0.5 mm/s revealed the
presence of Fe3+ ions in octahedral (Oh) coordination [37–39]. In ad-
dition, the remaining sublattices of LaFe6 catalysts associated with Fe3+
ions in octahedral sites (Al1, Al3, and Al4) can be further separated by
the value of H. The H with the largest value (515.6 kOe) was compar-
able to that of the octahedral Al3 sites in MP-hexaaluminate. Mean-
while, the sextets with the lowest hyperfine fields (377.1 kOe) were
attributed to the octahedral Al4 sites [28]. Moreover, it is significant to
emphasize that the Fe3+ (Oh) content increased progressively with the
4
Applied Catalysis B: Environmental xxx (xxxx) xxxx
rise of Fe substitution to 6 (4.2 wt%, 9.7 wt% and 14.9 wt% for LaFe2,
LaFe4 and LaFe6 catalysts, respectively. Calculated based on the relative
area and actual Fe content listed in Table 2), demonstrating that more
Fe ions entered into the hexaaluminate lattice by occupying the octa-
hedral sites of MP structure. It was extremely consistent with the
Raman results that the band related to stretching modes of octahedral
units was gradually enhanced with the rise of Fe substitution.
In the LaFe8 and LaFe10 catalysts cases, two new sextets with H
values higher than 520 KOe emerged, which were corresponded to
Fe3+ in α-Fe2O3 and LaFeO3 structure, respectively. The higher hy-
perfine field values (533–536 KOe) was regarded as the presence of
Fe3+ in α-Fe2O3, while Fe3+ in LaFeO3 resulted in H values locating at
524–525 KOe, taking into account the fact that the H values of Fe3+ in
α-Fe2O3 were greater than that of Fe3+ in LaFeO3 [27–29]. However,
the results suggested that the MP structure vanished in LaFe12 catalysts,
which was well aligned with the XRD and Raman analysis. Importantly,
states with IS > 0.8 mm/s are assigned to Fe2+ ions (0.7 wt%),which
also proved by the Fe 2p XPS analysis in Fig. S1a.
3.3. Active phases of the reaction
Although the above analysis revealed the phase compositions of the
G. Jiang, et al. Applied Catalysis B: Environmental xxx (xxxx) xxxx

Fig. 2. Structural characterization of LaFex catalysts. a X-ray diffraction patterns of LaFex catalysts. b The 30–40 degree (2 theta) range of the X-ray diffraction
patterns LaFex catalysts. c Raman spectra of LaFex catalysts.

catalysts, however, the actual active phases and reaction mechanism
were still uncertain.
Therefore, in order to get insight to the feasibility of traditional two-
step process over the LaFex catalysts, two cycle H2 yield experiments
versus reaction temperature were carried out over the LaFe6 catalysts
under the employed reaction conditions. And the catalytic perfor-
mances were exhibited in Fig. 5a.
From the standpoint of traditional two-step process, it seemed lo-
gical to expect that the sulfurized LaFe6 active phases should be formed
during the first cycle. Consequently, the catalytic performance should
be improved significantly for the second cycle at each or specific re-
action temperature. However, in terms of H2S conversion and H2 yields,
the used catalysts exhibited very similar results compared with those of
fresh catalysts (Fig. 5a). Importantly, the used catalysts under-
performed the fresh catalysts, which was contrary to the expectation. It
was suspected that different reaction mechanisms occurred among the
LaFex catalysts.
To clarify these phenomena, various characterizations were

Fig. 3. Morphology characterization of LaFe6 catalysts. a TEM images, b and c HRTEM images of LaFe6 catalysts. The insert image in c was SAED pattern of LaFe6
catalysts.

5
G. Jiang, et al. Applied Catalysis B: Environmental xxx (xxxx) xxxx

Table 1
57
Fe Mössbauer spectra results of LaFex catalysts at 80 K.
Catalyst IS (mm/s)a QS (mm/s)b H (KOe)c A (%)d Mass ratio (%)e Site assignment

LaFe2 0.215 0.677 50.0 6.9 doublet MP, Fe3+(Thf), Al2

0.271 2.328 20.1 2.8 doublet MP, Fe3+(Trg), Al5
0.490 0.640 30.0 4.2 doublet MP, Fe3+(Ohh)
LaFe4 0.303 0.390 doublet PARA
0.220 0.636 426.6 37.2 8.0 sextet MP, Fe3+(Th), Al2
0.270 2.430 324.0 17.7 3.8 sextet MP, Fe3+(Tr), Al5
0.429 0.620 505.0 45.1 9.8 sextet MP, Fe3+(Oh)
LaFe6 0.360 0 doublet PARA
0.214 0.634 463.2 27.0 6.5 sextet MP, Fe3+(Th), Al2
0.280 2.480 396.0 11.2 2.7 sextet MP, Fe3+(Tr), Al5
0.473 0.372 377.1 6.4 1.5 sextet MP, Fe3+(Oh), Al4
0.482 0.410 496.1 25.0 6.0 sextet MP, Fe3+(Oh), Al1
0.505 0.212 515.6 30.4 7.4 sextet MP, Fe3+(Oh), Al3
LaFe8 0.300 −0.100 536.0 6.3 1.7 sextet α-Fe2O3
0.390 0.180 525.0 12.8 3.4 sextet LaFeO3
0.210 0.630 470.0 29.0 7.8 sextet MP, Fe3+(Th), Al2
0.270 2.660 405.0 42.0 11.3 sextet MP, Fe3+(Tr), Al5
0.465 0.494 490.6 9.6 2.6 sextet MP, Fe3+(Oh)
LaFe10 0.369 0.180 533.0 16.0 4.8 sextet α-Fe2O3
0.352 0.148 524.0 30.0 8.9 sextet LaFeO3
0.261 0.324 517.0 29.0 8.7 sextet MP, Fe3+(Th), Al2
0.280 2.364 512.1 18.0 5.4 sextet MP, Fe3+(Tr), Al5
0.460 0.510 444.0 7.0 2.1 sextet MP, Fe3+(Oh)
LaFe12 0.304 0.155 534.0 38.0 13.6 sextet α-Fe2O3
0.347 0.123 524.1 60.0 21.4 sextet LaFeO3
1.020 −0.260 530.8 2.0 0.7 sextet Fe2+

a
Isomer shift.
b
Quadrupole splitting.
c
Hyperfine field.
d
Relative area.
e
The actual content of each specific Fe species.
f
Tetrahedral.
g
Trigonal bipyramidal.
h
Octahedral.

conducted on fresh, used and reused LaFe6 catalysts. The XRD analysis
(Fig. S2a) revealed that the MP structure of LaFe6 catalysts were well
maintained even after two-cycle reactions. More importantly, no peaks
related to sulfides and sulfates appeared. Whereas, the sulfur species
were detected by XPS and energy dispersive X-ray spectroscopy (EDX),
and the actual surface elemental contents were listed in Table 2. The
EDX mapping (Fig. S3) shown that sulfur species distributed uniformly
in all the catalysts. Besides, it was considered that the peaks located at
160.88 eV, 162.24 eV and 168 eV in S 2p XPS (Fig. S2d) were ascribed
to S2-x2− species, S2- species and SO42− species, respectively [40–42].
Moreover, further reaction could only cause a slight increase in sulfur
content (Table 2), which increased from 2.5 wt% for the used catalyst to
4.3 wt% for the reused catalyst.
X-ray absorption fine structure spectroscopy (XAFS) was further
applied to explore the phase compositions of the fresh used and reused
catalysts. The normalized Fe K-edge X-ray absorption spectra, X-ray
absorption near edge structure (XANES) region and radial distribution
functions were displayed in Fig. 5b–d. The Fe-K absorbing edge shifted
to a lower energy position (Fig. 5c), indicating the decrease of valent of
Fe ions [43,44]. These results were in line with Fe 2p XPS analysis in
Fig. S2b, in which the relative amount of Fe2+ species increased slightly
after two cycle reaction (area percentage 15% in used catalyst and 20 %
in reused catalyst).
It was astonishing to be observed that the spectra profiles in Fig. 5d
for all the catalysts were almost consistent with each other, revealing
that the structure were stable during the reaction and no more phases
emerged. There were two strong amplitude peaks between 1.0 Å and
4.0 Å. The first peak at 1.50 Å was related to the first coordinate level of
ferrite, which was corresponding to Fe-O coordinate peak caused by the
scattering of nearby O2− ions [45]. The second peak appeared at 2.6 Å
were ascribed to second coordinate level, which was caused by the
scattering of second-rate nearby metal ions [45]. As known, the radial
distribution function may indicate the distance between atom level and
central absorbing atom. The first coordinate level in catalyst samples
implied that the coordination of Fe-O was changed after reaction. It was
shown that the distance between the scattering center (Fe ions) and the
O2− ions was shorten after used, due mainly to the reduction of a few
Fe3+ to Fe2+. Besides, as stated in literature, the Fe-S bond distance
was generally considered to be located at 2.23 Å [46]. While, there was
no discernible peak in this position.
Hence, the traditional two-step mechanism may not favorable to
explain the results in this work by considering that the MP structure of
LaFe6 catalysts were well maintained even proceeded two cycles
(Figs. 5d and S2a) and only a tiny proportion of Fe was sulfurized to
sulfide and sulfate species (Table 2). As from the perspective of tradi-
tional two-step mechanism, the catalysts should be sulfurized easily and
abundantly, and more sulfur species were beneficial to the reaction.
Therefore, the sulfides in traditional two-step mechanism seemed not
play the main role of actual active phases.
Consequently, it was likely acceptable to conclude that the Fe spe-
cies presented in the stable hexaaluminate structure rather than sulfides
played the role of actual active phases in the reaction. Particularly, it
was extremely important to observe that the H2 yield in Fig. 1a over
fresh LaFex catalysts at 800 °C were positively correlated with the
content variation trend (Table 1) of Fe3+ ions in octahedral (Oh) co-
ordination within hexaaluminate structure. Hence, the Fe3+ (Oh)
maybe among the major active site species contributing to the excellent
catalytic performance.
3.4. Reaction mechanism over hexaaluminate catalysts
Based on theradical reaction mechanism [47] and the analysis of

6
G. Jiang, et al. Applied Catalysis B: Environmental xxx (xxxx) xxxx

Fig. 4. 57Fe Mössbauer spectra fitting results of LaFex catalysts. The solid lines show the curve fits to the raw data. Average uncertainties in Mössbauer para-
meters: ± 0.03, Fe fraction; ± 2 KOe, hyperfine magnetic field; ± 0.03 mm/s, isomer shift; ± 0.05 mm/s quadrupole splitting.

Table 2 well known that the crystal structure of hexaaluminates is very com-
The chemical composition of surface elements in LaFex catalysts. plicated, consisting of alternatively stacked spinel blocks and mirror
Catalyst Atomic ratio (%) Oβ/Oα Mass ratio (%)
planes. The analysis in Table 1 proved that Fe mainly entered the spinel
blocks substituting the lattice sites of Al in Fe-substituted hex-
La Fe O Al S Fe S aaluminate samples. Thus, the structure of FeAl2O4 spinel was adopted
to simulate hexaaluminates and simplify the calculation. Because hy-
LaFe2 3.1 6.4 57.1 33.2 0.51 13.9
drogen generation was reduced with LaFe8, LaFe10 and LaFe12 catalysts,
LaFe4 3.9 10.9 57.6 27.4 0.56 21.6
LaFe6 4.7 12.7 59.5 22.9 0.61 24.2 related calculation on α-Fe2O3 and LaFeO3 were also conducted for
LaFe8 6.1 15.2 60.7 17.8 0.77 26.9 reference and comparison. The related slab models were displayed in
LaFe10 7.5 17.9 64.4 10.0 1.46 29.9 Fig. S4.
LaFe12 10.3 24.6 65.1 1.61 35.7
To explore the reaction pathways, the stable co-adsorption config-
LaFe6 used 5.2 10.4 55.1 26.8 2.4 0.56 19.4 2.5
LaFe6 reused 5.0 8.1 53.6 29.2 4.0 0.54 15.4 4.3 urations of SH and H or S and 2H served as the reactants and inter-
mediates. All structure and bond length information for the transition
states (TS) and products were displayed in Figs. S5 and S6. And sche-
active phases, a probable direct dehydrogenation reaction pathway was matic potential energy profiles were presented in Fig. 6 to illustrate the
tentatively proposed as following: energy changes for the whole process of H2S interacting with the sur-
face. As the research about H2S interacting with the α-Fe2O3 (001)
-Fe- + H-SH → -Fe-SH + H, (8) surface has been already clear, the results of potential energy profiles
-Fe-SH → -Fe-S + H, (9) were used for comparison [48,49].
In Figs. S5 and S6, H2S firstly approached and adsorbed on the Fe
-Fe-S → -Fe- + S, (10) top site, parallel to the surface of LaFeO3 (010) and FeAl2O4 (100),
H + H → H2. (11) eventually, analogous to the direct dehydrogenation process on the α-
Fe2O3 (001) surface [49]. After the fracture of H–S bond in absorbed
H2S firstly adsorbed on the catalysts surface of Fe (Oh) sites, and
3+
H2S and OeH bond formation, the first dehydrogenation process (H2S
then complete the decomposition by direct dehydrogenation processes. → SH + H) occurred, with the energy barriers of 25.9 and 7.7 kJ/mol
The density functional theory (DFT) method within a periodic slab on p (2 × 2) super-cells of LaFeO3 (010) and FeAl2O4 (100) surface,
model was employed to validate the new proposed mechanism. It is respectively. These values were much smaller than the corresponding

7
G. Jiang, et al. Applied Catalysis B: Environmental xxx (xxxx) xxxx

Fig. 5. Catalytic performance and characterization of LaFe6 catalysts. a H2 yield (line and symbol) and H2S conversion (column) of the fresh and used LaFe6 catalysts
as a function of temperature in the feed gas of 1000 ppm H2S/N2 (24000 h−1). b Normalized Fe K-edge X-ray absorption spectra of the fresh and used LaFe6 catalysts.
c XANES region of the normalized Fe K-edge X-ray absorption spectra of the fresh and used LaFe6 catalysts. d Radial distribution functions around Fe atom of samples
of the fresh and used LaFe6 catalysts.

hypothetical route was feasible and the reaction could occur on the
FeAl2O4 (100) surface easiest.
Differences of the catalytic activities were perhaps mainly derived
from the electronic and structural properties of the catalysts. The Fe-O
atoms on the catalysts surface were used to cleave H2S electronically,
and the O atom acted as a proton acceptor, which meant that the more
negative charge the O atom had, the higher ability to attract protons
[50]. Calculation of Mulliken charge (Table S2) confirmed that the
absolute charge number of O on the three surfaces followed the order:
FeAl2O4 > LaFeO3 > Fe2O3, consistent with the yield performance.
Structurally, differences between the H–S-H angle in H2S (the angle of
H-S-H in the free H2S is 92.1°) and the O-Fe-O angle on the surface for
adsorption and cleavage significantly have an effect on the energy
barrier [51]. Greater differences in angle lead to higher energy barrier
to be overcome [52]. The angle of O-Fe-O on the FeAl2O4 surface had a
moderate difference with the angle of H-S-H in the free H2S, resulting in
the decomposition happen possibly.
Fig. 6. Energy profile of the H2S decomposition process. Finally, the catalytic performance can be explained reasonably by
the direct dehydrogenation mechanism. The catalysts with lower Fe
substitution (LaFe2, LaFe4 and LaFe6) could produce higher H2 yield
α-Fe2O3 (001) (72.0 kJ/mol) [49]. Another dehydrogenation process
(Fig. 1a) were due mainly to the higher content of Fe3+ (Oh) sites. And
(SH → S + H) proceeded, with the energy barriers of 52.8 and 17.3 kJ/
the slightly decreased catalytic performance in Fig. 5a for the second
mol of LaFeO3 (010) and FeAl2O4 (100) surface, respectively. These
cycle might resulted from the formation of little amount sulfur species.
values were also much lower than the corresponding α-Fe2O3 (001)
Besides, it was worth mentioning that the O 1s XPS analysis re-
(130 kJ/mol) [49]. Thus, the FeAl2O4 (100) surface presented the
vealed that oxygen species were presented on the catalyst surface and
lowest energy barriers in the whole process, indicating that the
consumed during the reaction (Fig. S2c). As known, the peak at ∼531.5

8
G. Jiang, et al.
eV was related to surface adsorbed oxygen species (donated as Oα), and
the band at ∼529.5 eV was ascribed to lattice oxygen species (donated
as Oβ) [53]. And the Oβ/Oα in Fig. S2c decreased from 0.61 to 0.54 after
two cycle reaction, indicating that the lattice oxygen was involved in
the reaction.
The O 1s XPS spectra for all the LaFex catalysts were also provided
in Fig. S1b. It indicated that lattice oxygen species went up with the
increase of Fe substitution, and the Oβ/Oα rise sharply especially when
iron oxides appeared (LaFe8, LaFe10 and LaFe12 catalysts). As known,
lattice oxygen were nucleophilic reagents and were usually responsible
for selective oxidation reactions. In conjunction with the phenomenon
that Fe2+ emerged after the reaction, H2S selective catalytic oxidation
perhaps occurred simultaneously [54], as a Fe3+ redox cycle was in-
volved in the H2S selective oxidation process. Moreover, the sequence
of H2S conversion were extremely consistent with that of Oβ/Oα. Thus,
the occurrence of H2S selective catalytic oxidation associated with the
sulfidation are responsible for the higher H2S conversion for LaFe8,
LaFe10 and LaFe12 catalysts (Fig. 1b) and the initially increased stage of
H2 yield in the first time experiment of catalysts durability (Fig. 1c).
4. Conclusion
Synergistic recovery of hydrogen and sulfur resources were con-
ducted over catalytic decomposition of acidic gas H2S on non-sulfide
based LaFex hexaaluminate catalysts. The LaFe6 catalysts could present
more than 50% H2 yield and proceed consecutively for two cycles
(20 h) without notable deactivation. The outstanding catalytic perfor-
mance in this work provided possibility for the technological innova-
tion in chemical industries involved acidic gas H2S produced.
Compared with conventional Claus process, hydrogen in H2S is no
longer wasted as H2O, but can be converted into valuable hydrogen
resource and energy.
Additionally, it was considered that the Fe species presented in the
stable hexaaluminate structure, especially the Fe3+ (Oh), rather than
the sulfides were the actual active phases in the reaction. And the tra-
ditional two-step mechanism was not involved in the reaction.
Significantly, a probable reaction pathway was tentatively proposed
and verified by DFT method. H2S firstly adsorbed on the catalysts
surface of Fe3+ (Oh) sites, and then completed the decomposition by
direct dehydrogenation processes. It was revealed that the synthesized
hexaaluminate catalysts presented the lowest energy barriers in the
whole process. Moreover, the H2S selective catalytic oxidation reaction
could occur simultaneously, and the lattice oxygen participated in the
reaction. The generated little sulfide and sulfate species were re-
sponsible for the decrease of catalytic activities.
Declaration of Competing Interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
Acknowledgements
This work was financially supported by the National Natural Science
Foundation (21976177, 21976176, 21577158, 21507148), the
National Key R&D Program of China (2018YFA0209302) and the
University of Chinese Academy of Sciences.
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.apcatb.2019.118354.
9
Applied Catalysis B: Environmental xxx (xxxx) xxxx
References
[1] M.S. Shah, M. Tsapatsis, J.I. Siepmann, Chem. Rev. 117 (2017) 9755–9803.
[2] A.K. Gupta, S. Ibrahim, A. Al Shoaibi, Prog. Energy Combust. Sci. 54 (2016) 65–92.
[3] A. Shaver, M. ElKhateeb, A.M. Lebuis, Angew. Chemie Int. Ed. English 35 (1996)
2362–2363.
[4] C. Duan, R.J. Kee, H. Zhu, C. Karakaya, Y. Chen, S. Ricote, A. Jarry, E.J. Crumlin,
D. Hook, R. Braun, Nature 557 (2018) 217.
[5] X. Zong, J.F. Han, B. Seger, H.J. Chen, G.Q. Lu, C. Li, L.Z. Wang, Angew. Chemie Int.
Ed. English 53 (2014) 4399–4403.
[6] L. Zhao, Y. Wang, L. Jin, M.L. Qin, X. Li, A.J. Wang, C.S. Song, Y.K. Hu, Green
Chem. 15 (2013) 1509–1513.
[7] K.M. Kwok, S.W.D. Ong, L. Chen, H.C. Zeng, ACS Catal. 8 (2017) 714–724.
[8] D. Ipsakis, T. Kraia, G.E. Marnellos, M. Ouzounidou, S. Voutetakis, R. Dittmeyer,
A. Dubbe, K. Haas-Santo, M. Konsolakis, H.E. Figen, N.O. Güldal, S.Z. Baykara, Int.
J. Hydrogen Energ. 40 (2015) 7530–7538.
[9] T.Y. Cong, A. Raj, J. Chanaphet, S. Mohammed, S. Ibrahim, A. Al Shoaibi, Int. J.
Hydrogen Energ. 41 (2016) 6662–6675.
[10] A.P. Reverberi, J.J. Klemeš, P.S. Varbanov, B. Fabiano, J. Clean. Prod. 136 (2016)
72–80.
[11] F. Faraji, I. Safarik, O.P. Strausz, E. Yildirim, M.E. Torres, Int. J. Hydrogen Energ. 23
(1998) 451–456.
[12] A. Bishara, O.A. Salman, N. Khraishi, A. Marafi, Int. J. Hydrogen Energ. 12 (1987)
679–685.
[13] O.A. Salman, A. Bishara, A. Marafi, Energy 12 (1987) 1227–1232.
[14] V.E. Kaloidas, N.G. Papayannakos, Int. J. Hydrogen Energ. 12 (1987) 403–409.
[15] N.O. Guldal, H.E. Figen, S.Z. Baykara, Int. J. Hydrogen Energ. 40 (2015)
7452–7458.
[16] N.O. Guldal, H.E. Figen, S.Z. Baykara, Chem. Eng. J. 313 (2017) 1354–1363.
[17] N.O. Guldal, H.E. Figen, S.Z. Baykara, Int. J. Hydrogen Energ. 43 (2018)
1038–1046.
[18] T. Chivers, J.B. Hyne, C. Lau, Int. J. Hydrogen Energ. 5 (1980) 499–506.
[19] A.J. Zarur, J.Y. Ying, Nature 403 (2000) 65–67.
[20] J.G. McCarty, Nature 403 (2000) 35–36.
[21] M. Tian, X.D. Wang, T. Zhang, Catal. Sci. Technol. 6 (2016) 1984–2004.
[22] L.Z. Zhang, J.-M.M. Millet, U.S. Ozkan, J. Mol. Catal. A-Chem. 309 (2009) 63–70.
[23] G. Groppi, C. Cristiani, P. Forzatti, J. Catal. 168 (1997) 95–103.
[24] J. Kreisel, G. Lucazeau, H. Vincent, J. Solid State Chem. 137 (1998) 127–137.
[25] D.L.A. De Faria, S. Venâncio Silva, M.T. De Oliveira, J. Raman Spectrosc. 28 (1997)
873–878.
[26] E. Traversa, P. Nunziante, L. Sangaletti, B. Allieri, L.E. Depero, H. Aono, Y. Sadaoka,
J. Am. Ceram. Soc. 83 (2000) 1087–1092.
[27] Y. Zhang, X.D. Wang, Y.Y. Zhu, X. Liu, T. Zhang, J. Phys. Chem. C 118 (2014)
10792–10804.
[28] Y.Y. Zhu, X.D. Wang, A.Q. Wang, G.T. Wu, J.H. Wang, T. Zhang, J. Catal. 283
(2011) 149–160.
[29] Y.Y. Zhu, X.D. Wang, G.T. Wu, Y.Q. Huang, Y. Zhang, J.H. Wang, T. Zhang, J. Phys.
Chem. C 116 (2011) 671–680.
[30] P. Sazama, B. Wichterlová, E. Tábor, P. Šťastný, N.K. Sathu, Z. Sobalík, J. Dědeček,
Š. Sklenák, P. Klein, A. Vondrová, J. Catal. 312 (2014) 123–138.
[31] G. Fierro, G. Moretti, G. Ferraris, G.B. Andreozzi, Appl. Catal. B: Environ. 102
(2011) 215–223.
[32] P. Boroń, L. Chmielarz, J. Gurgul, K. Łątka, T. Shishido, J.-M. Krafft, S. Dzwigaj,
Appl. Catal. B: Environ. 138–139 (2013) 434–445.
[33] C.H. Choi, W.S. Choi, O. Kasian, A.K. Mechler, M.T. Sougrati, S. Bruller,
K. Strickland, Q.Y. Jia, S. Mukerjee, K.J.J. Mayrhofer, F. Jaouen, Angew. Chemie
Int. Ed. English 56 (2017) 8809–8812.
[34] E. Tronc, F. Laville, M. Gasperin, A.M. Lejus, D. Vivien, J. Solid State Chem. 81
(1989) 192–202.
[35] G. Ung, J. Rittle, M. Soleilhavoup, G. Bertrand, J.C. Peters, Angew. Chemie Int. Ed.
English 53 (2014) 8427–8431.
[36] M.M. Rodriguez, B.D. Stubbert, C.C. Scarborough, W.W. Brennessel, E. Bill,
P.L. Holland, Angew. Chemie Int. Ed. English 51 (2012) 8247–8250.
[37] P. Sazama, N.K. Sathu, E. Tabor, B. Wichterlová, Š. Sklenák, Z. Sobalík, J. Catal. 299
(2013) 188–203.
[38] M.M. Khusniyarov, T. Weyhermuller, E. Bill, K. Wieghardt, Angew. Chemie Int. Ed.
English 47 (2008) 1228–1231.
[39] I. Yamada, K. Tsuchida, K. Ohgushi, N. Hayashi, J. Kim, N. Tsuji, R. Takahashi,
M. Matsushita, N. Nishiyama, T. Inoue, T. Irifune, K. Kato, M. Takata, M. Takano,
Angew. Chemie Int. Ed. English 50 (2011) 6579–6582.
[40] Z. Guo, X.W. Wang, Angew. Chemie Int. Ed. English 57 (2018) 5898–5902.
[41] T. Jurca, M.J. Moody, A. Henning, J.D. Emery, B.H. Wang, J.M. Tan, T.L. Lohr,
L.J. Lauhon, T.J. Marks, Angew. Chemie Int. Ed. English 56 (2017) 4991–4995.
[42] Y.D. Shao, Z. Guo, H. Li, Y.T. Su, X.W. Wang, Angew. Chemie Int. Ed. English 56
(2017) 3226–3231.
[43] W.C.M. Gomes, D.M.A. Melo, P.M. Pimentel, E.P. Marinho, M.A.F. Melo, R.S. Nasar,
Chem. Phys. 490 (2017) 67–74.
[44] M. Soloviov, A.K. Das, M. Meuwly, Angew. Chemie Int. Ed. English 55 (2016)
10126–10130.
[45] L.X. Wang, J. Zhang, Q.T. Zhang, N.C. Xu, J. Song, J. Magn. Magn. Mater. 377
(2015) 362–367.
[46] A. Matamoros-Veloza, O. Cespedes, B.R.G. Johnson, T.M. Stawski, U. Terranova,
N.H. de Leeuw, L.G. Benning, Nat. Commun. 9 (2018).
[47] A.M. El-Melih, L. Iovine, A. Al Shoaibi, A.K. Gupta, Int. J. Hydrogen Energ. 42
(2017) 4764–4773.
G. Jiang, et al.
[48] J.J. Song, X.Q. Niu, L.X. Ling, B.J. Wang, Fuel Process. Technol. 115 (2013) 26–33.
[49] L.X. Ling, J.J. Song, S.P. Zhao, R.G. Zhang, B.J. Wang, RSC Adv. 4 (2014)
22411–22418.
[50] M. Haranczyk, R. Bachorz, J. Rak, M. Gutowski, D. Radisic, S.T. Stokes, J.M. Nilles,
K.H. Bowen, J. Phys. Chem. B 107 (2003) 7889–7895.
10
Applied Catalysis B: Environmental xxx (xxxx) xxxx
[51] D.C. Look, I.J. Lowe, J.A. Northby, J. Chem. Phys. 44 (1966) 3441–3452.
[52] R.E. Johnson Jr., R.H. Dettre, J. Phys. Chem. 68 (1964) 1744–1750.
[53] C.N.R. Rao, V. Vijayakrishnan, A.K. Santra, M.W.J. Prins, Angew. Chemie Int. Ed.
English 31 (1992) 1062–1064.
[54] X.R. Ren, L.P. Chang, F. Li, K.C. Xie, Fuel 89 (2010) 883–887.