Nano- and microscale engineering of the molybdenum disulfide-based catalysts for syngas to ethanol conversion

Article abstract of DOI:10.1002/cctc.201402067

Nickel-promoted MoS₂, unsupported catalysts and laponite-supported alcohol synthesis catalysts were synthesised by using microemulsion (ME) and hydrothermal (HT) methods. Highly ordered sulfide slabs, consisting of up to seven layers, were visible in the TEM images of HT-based NiMoS₂ catalysts. In contrast, disordered sulfide layers were identified in ME-based NiMoS₂ catalysts. High catalytic activity was observed in ME-based supported (laponite-supported NiMoS₂) and unsupported catalysts. After the CO hydrogenation reaction, the catalysts were characterised by X-ray photoelectron spectroscopy and inductively coupled plasma-mass spectrometry elemental analyses, which detected a significant sulfur loss in ME-based NiMoS₂ catalysts and minor sulfur loss in HT-based NiMoS₂ catalysts. In addition to the large surface area (120 m ²g^-1), disordered sulfide structure, and exposed active sites, ME-based NiMoS₂ catalysts demonstrated higher alcohol selectivity (61 mol%) than HT-based NiMoS₂ catalysts (15 mol%). Correlations between the catalyst morphology, surface active components, and alcohol selectivity are discussed herein.

Full text of DOI:10.1002/cctc.201402067

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201402067
Nano- and Microscale Engineering of the Molybdenum
Disulfide-Based Catalysts for Syngas to Ethanol
Conversion
Muxina Konarova,[a] Fengqiu Tang,[a] Jiuling Chen,[b] Geoff Wang,[b] Victor Rudolph,[b] and
Jorge Beltramini*[a]
Nickel-promoted MoS2, unsupported catalysts and laponite-
supported alcohol synthesis catalysts were synthesised by
using microemulsion (ME) and hydrothermal (HT) methods.
Highly ordered sulfide slabs, consisting of up to seven layers,
were visible in the TEM images of HT-based NiMoS2 catalysts.
In contrast, disordered sulfide layers were identified in ME-
based NiMoS2 catalysts. High catalytic activity was observed in
ME-based supported (laponite-supported NiMoS2) and unsup-
ported catalysts. After the CO hydrogenation reaction, the cat-
alysts were characterised by X-ray photoelectron spectroscopy
and inductively coupled plasma–mass spectrometry elemental
analyses, which detected a significant sulfur loss in ME-based
NiMoS2 catalysts and minor sulfur loss in HT-based NiMoS2 cat-
alysts. In addition to the large surface area (120 m2 gꢀ1), disor-
dered sulfide structure, and exposed active sites, ME-based
NiMoS2 catalysts demonstrated higher alcohol selectivity
(61 mol%) than HT-based NiMoS2 catalysts (15 mol%). Correla-
tions between the catalyst morphology, surface active compo-
nents, and alcohol selectivity are discussed herein.
Introduction
Among the fuel alternatives to gasoline, ethanol has attracted
the most research interest.[1] Ethanol contains oxygen and has
a higher octane number than does gasoline. It delivers a sus-
tainable, CO2 neutral, and clean energy alternative to fuel if
sourced from renewable resources. These factors have encour-
aged the use of ethanol fuel in transport worldwide to reduce
the consumption of gasoline. For example, in the United
States and Brazil, high-ethanol blends are available as a trans-
portation fuel (e.g., E85, an 85% blend with gasoline).
um-based catalysts,[4] modified Fischer–Tropsch synthesis cata-
lysts,[5] modified methanol synthesis catalysts,[6] and MoS2-
based catalysts.[7]
Herein, we focus on the synthesis and catalytic performance
of the MoS2-based catalyst. Researchers at Dow Chemicals pa-
tented the MoS2-based catalyst in the late 1980s, as a bench-
scale catalyst testing showed promising results on conversion
of syngas to mixed alcohols.[8] The MoS2-based catalyst ach-
ieved up to 30% CO conversion and 80% ethanol selectivity
(CO2 free basis) upon mixing/doping with alkaline metals
(MAlk =K, Cs, Rb) and transition metals (MTr =Ni, Co, Fe).[9] Most
reports state that the alkaline metals promote alcohol forma-
tion by preventing reduction of MoIV to Mo0[10] whereas the
transition metals increase CO conversion by promoting carbon
chain growth.[11] To obtain the MAlk–MTr–MoS2 catalyst, the tran-
sition metals are co-precipitated with MoS2 (ammonium tetra-
thiomolybdate) precursors[12] and the alkaline metals are intro-
duced through physical mixing. Most studies aim to find opti-
mum ratios of MAlk/Mo or MTr/Mo and reaction conditions
(pressure, syngas ratio, space velocity, and CO2 or H2S addition
into feed gas).[2a,13] Meanwhile, reports on the synthesis of cata-
lysts and effects of the synthesis parameters on catalyst prop-
erties are limited.
Currently, ethanol fuel is produced from corn or sugarcane
through fermentation; however, there is an increased interest
to produce fuel from syngas (a mixture of CO and H2).[2] Nu-
merous studies have endeavoured to formulate a cost-effective
catalytic process for the production of ethanol from syngas.[3]
Nevertheless, such technology is yet to be implemented be-
cause of side reactions, which produce undesired products
(methane, methanol, and other hydrocarbons and oxygenates).
Catalysts play a major role in promoting ethanol formation
and inhibiting side reactions. Catalysts that are suitable for the
conversion of syngas to ethanol can be categorised as rhodi-
[a] Dr. M. Konarova, Dr. F. Tang, Dr. J. Beltramini
ARC Centre of Excellence for Functional Nanomaterials
The University of Queensland
The Ni(Co)MoS2 catalyst, also known as a hydrotreating cata-
lyst, is mainly used to remove sulfur, nitrogen, and oxygen
from crude oil feedstock.[14] According to previous reports, the
catalytic activity of MoS2-based hydrotreating catalysts could
be tuned by altering the synthesis parameters.[15] Devers et al.
produced the MoS2 catalyst through the thermal decomposi-
tion and hydrothermal (HT) synthesis of thiosalts; catalytic
tests of thiophene hydrodesulphurisation and tetraline hydro-
St Lucia, QLD 4072 (Australia)
Fax: (+61)7-3346-3973
E-mail: j.beltramini@uq.edu.au
[b] Dr. J. Chen, Dr. G. Wang, Prof. V. Rudolph
School of Chemical Engineering
The University of Queensland
St Lucia, QLD 4072 (Australia)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201402067.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2394 – 2402 2394

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
genation reactions revealed a significant effect of the morphol-
ogy on the catalytic activity.[16] By using a solution-based
method, Genuit et al. varied synthesis parameters (precursors
nature, surfactant types, and MeNi,Co/Mo ratios) and obtained
highly dispersed Ni(Co)–Mo–S sulfides, which led to high cata-
lytic activity in the hydrodesulfurisation of thiophene.[17]
Furthermore, the catalytic activity of MoS2 catalysts was af-
fected by the location of promoter atoms (cobalt and nickel) in
the sulfide structure.[18] High catalytic activities were observed
in the mixed sulfide (MeCo,Ni–Mo–S) phase, in which the pro-
moter atoms are located on the edge planes of the sulfide
layers.[19] In the mixed sulfide phase, a direct correlation of the
catalyst structure with the catalytic performance was con-
firmed by the Topsøe group.[20] In non-promoted MoS2 cata-
lysts, active sites were associated with sulfur vacancies (anionic
vacancies), which were formed under a hydrogen-rich environ-
ment.[21] With an atom-resolved scanning tunnelling micro-
scope, Kibsgaard[22] detected sulfur vacancies by exposing
MoS2 clusters to atomic hydrogen and concluded that the for-
mation of the vacancy (corner or edge) depends on the actual
size of MoS2 nanoclusters.
Figure 1. XRD patterns of NiMoS2-LP-HT, NiMoS2-LP-ME, and bare laponite.
The catalytic active sites of MoS2 alcohol synthesis catalysts
are expected to be the same as of hydrotreating catalysts.[23]
Most reports lack data indicating a direct correlation between
catalytic active sites and catalytic performance (CO conversion
and alcohol selectivity) of MoS2. It is unclear which morpholo-
gy would best promote the formation of catalytic active sites
(sulfur vacancies) and which synthesis route would result in
such a morphology/structure. The purpose of this study was to
find a correlation between catalytic performance and the mor-
phology/structure of MoS2. To identify a clear distinction be-
tween morphology and structure, we chose the microemulsion
(ME) and HT methods for MoS2 synthesis and compared cata-
lytic activity, alcohol selectivity, and textural properties of both
supported and unsupported NiMoS2 catalysts.
Results and Discussion
Synthesis and characterisation of laponite-supported
NiMoS2 catalysts
Figure 2. TEM images of a) NiMoS2-LP-HT, b) NiMoS2-LP-ME, and c) bare la-
ponite. Scale bars=200 (a,b) and 50 nm (c).
The XRD patterns of NiMoS2-LP-HT, prepared by using the HT
method, NiMoS2-LP-ME, prepared by using the ME method,
and the laponite support are shown in Figure 1.
The TEM images of NiMoS2-LP-HT and NiMoS2-LP-ME re-
vealed dark catalyst particles distributed on the laponite sup-
port. The TEM image of NiMoS2-LP-HT consisted of large aggre-
gates (ꢁ200 nm), which were irregularly distributed across the
laponite framework. In contrast, the TEM image of NiMoS2-LP-
ME resembled the ideal morphology of a catalyst, in which fine
particles (ꢁ20 nm) of active species were homogenously de-
posited on the support. A comparison of the TEM images of
NiMoS2-LP-HT and NiMoS2-LP-ME catalysts revealed that the ME
method results in an ideal catalyst structure, which has uni-
formly distributed NiMoS2 particles on the laponite framework.
The nitrogen physisorption isotherms of the two catalysts
corresponded to type IV isotherms, which are characteristic of
mesoporous adsorbents (Figure 3).[25] The BET surface area of
Laponite has a layered structure; the interlayer basal spacing
d(001) appeared at 2q=78 on the XRD patterns. The absence
of the basal peak (2q=78) indicated either 1) complete exfolia-
tion of laponite layers or 2) chemical decomposition of lapon-
ite via acid leaching during synthesis.[24] Laponite retained its
layered structure during syntheses (HT and ME), as XRD detect-
ed the basal (2q=78) peak in all three samples. The XRD pat-
tern of NiMoS2-LP-HT mainly consisted of laponite and MoS2
phases (Figure 1). NiMoS2-LP-ME generated weak XRD signals,
which indicated the amorphous nature of samples (Figure 1).
The TEM images of the samples are shown in Figure 2. An
image of commercial bare laponite is also provided to clarify
laponite structures.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2394 – 2402 2395

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
performed at T=3108C, P=60 bar (1 bar=100 kPa), and gas
hourly space velocity (GHSV)=1044 hꢀ1
.
The catalytic test results of the two catalysts are presented
in Figure 4 and Table 1. The NiMoS2-LP-HT catalyst maintained
stable CO conversion during the 95 hour reaction run; the
NiMoS2-LP-ME catalyst had high initial activity, which quickly
decreased with reaction time. The average CO conversion
reached up to 14.5 mol% with NiMoS2-LP-ME, whereas only
9.45 mol% of CO was converted into products with NiMoS2-LP-
HT.
Figure 3. Nitrogen physisorption isotherms and PSD curves of NiMoS2-LP-HT
and NiMoS2-LP-ME catalysts. w=Pore width
300 and 250 m2 gꢀ1 was measured for the NiMoS2-LP-ME and
NiMoS2-LP-HT catalysts, respectively. The pore size distribution
(PSD) curves of the two catalysts are compared in the inset of
Figure 3.
A bimodal PSD was observed for catalysts, which was cen-
tred at 3.5 and 5.1 nm for NiMoS2-LP-HT and at 3.5 and 7.1 nm
for NiMoS2-LP-ME. Regardless of the different synthesis meth-
ods used, both catalysts demonstrated an average pore size of
3.5 nm (the BJH method). This pore size could arise from
spaces between randomly oriented MoS2 layers. Compared
with the HT-based sample, the ME-based sample had a broader
PSD. The broad PSD was possibly due to the non-ionic surfac-
tant (Brij 30) used as an ME stabiliser; the surfactant decom-
posed during heat treatment, which left a porous structure
and amorphous carbon residue.
Figure 4. CO conversion as a function of time on stream for NiMoS2-LP-HT
and NiMoS2-LP-ME catalysts performed at 3108C, 60 bar, and a GHSV of
1044 hꢀ1
.
Table 1. CO conversion and selectivity of NiMoS2-LP-HT and NiMoS2-LP-
ME catalysts.[a]
Catalytic performances of laponite-based NiMoS2 catalysts
Variable
NiMoS2-LP-HT
9.45
NiMoS2-LP-ME
14.5
XCO[b] [mol%]
S[c] [mol%]
The catalytic performances of laponite-supported NiMoS2 cata-
lysts were studied at the laboratory scale in a fixed-bed high-
pressure reactor. By using the measured volumetric flow rate
and the concentration of component i, the molar flow rate of
component i was determined (Fi). The CO conversion was cal-
culated by using the molar flow rates of CO in the inlet and
outlet stream of the reactor. The internal standard gas (nitro-
gen) was used to calculate molar flow rates of CO in the outlet
stream [Eq. (1)].
hydrocarbons
C1ꢀOH
51.4
3.9
26.4
2.6
0.3
33.2
79.5
34.2
19.1
18.2
3.39
1.71
42.4
43.2
C2ꢀOH
C3ꢀOH
other oxygenates
alcohol
C2ꢀOH in total alcohol [mol%]
[a] Reaction conditions: P=60 bar, GHSV=1044 hꢀ1, H2/CO=2; [b] CO
conversion; [c] Selectivity based on C-containing products, including CO2.
in
CO
out
CO
F
ꢀ F
ð1Þ
XCO
¼
in
F
CO
Although NiMoS2-LP-ME resulted in high CO conversion
(14.5 mol%), methane and methanol were the major out-
stream products with 34 and 19 mol% selectivity, respectively.
To clarify the differences in the catalytic performance of the
two catalysts, X-ray photoelectron spectroscopy (XPS) analysis
was performed (Figure 5). The aim was to quantify surface ele-
ments and compare the quantity of surface elements in
NiMoS2-LP-HT and NiMoS2-LP-ME catalysts. Notably, the XPS re-
sults were obtained from a survey (general) scan without fur-
ther details on the oxidation state of the elements. As samples
The selectivity of a product was calculated by using Equa-
tion (2). Notably, CO2 content is included in the calculation of
selectivity.
out
ethanol
F
ð2Þ
Sethanol
¼
in
out
CO
F
ꢀ F
CO
To study CO conversion and the stability of the laponite-sup-
ported NiMoS2 catalyst, the CO hydrogenation reaction was
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2394 – 2402 2396

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
In the XPS of NiMoS2-LP-ME catalysts, laponite spe-
cies (magnesium, silicon, and oxygen) dominated on
the surface, which could act as a barrier between cat-
alyst species (molybdenum, nickel, and sulfur) and re-
actant molecules (carbon monoxide and hydrogen).
In the ME method, NiMoS2 particles were formed in
a reverse emulsion (water-in-oil) system and their
particle sizes were less than 5 nm (some of them
formed aggregates up to ꢁ20 nm); laponite particles
(dp ꢁ30 nm) were dispersed in water without further
size reduction. The different particle size likely caused
the presence of most laponite particles on the sur-
face layers of the catalyst; thus, laponite particles
generated stronger XPS signals than did NiMoS2 par-
ticles. This observation could explain why laponite-
supported (NiMoS2-LP-ME) catalysts failed to deliver
high catalytic activity.
The above results were based on the textural and
catalytic properties of laponite-supported NiMoS2 cat-
alysts prepared by using HT and ME methods. The
ME method produced NiMoS2-LP with more superior
catalytic performances than the HT-based catalysts. A
comparison of TEM images revealed that the ME
method resulted in homogenously distributed
NiMoS2 particles on the laponite support. According
to the XPS results, the surface of NiMoS2-LP-ME was
covered by laponite species (magnesium, silicon, and
oxygen); however, a greater exposure to the active sites (mo-
lybdenum, nickel, and sulfur) was needed to achieve high CO
conversion. For further comparisons, a new set of unsupported
NiMoS2 catalysts was prepared by using the ME and HT meth-
ods.
Figure 5. XPS results of NiMoS2-LP-HT and NiMoS2-LP-ME. High-resolution scan in Mo3d,
S2p, and Ni2p regions. CPS=Countssꢀ1
.
were exposed to air during preparation for XPS, the catalysts
were not reduced before XPS analysis. Carbon and oxygen
were the major elements on the surface of both catalysts
(Table 2). The atomic percentage of elements in NiMoS2-LP-HT
changed as follows: 1) the molybdenum content remained un-
changed, 2) the sulfur content reduced from 5.48 to 4.37 at%,
and 3) the nickel content equalled 1.31 at% before and
0.82 at% after the reaction on the surface.
Catalytic performances of unsupported NiMoS2 catalysts
The XPS results of the ME-based samples, in which the sur-
face concentration of active species (molybdenum, nickel, and
sulfur) was three times lower than that in the HT-based sam-
ples (in fresh catalysts), are also summarised in Table 2. Sulfur
was not detected in the spent NiMoS2-LP-ME catalysts, whereas
the surface concentration of nickel and molybdenum de-
creased after the reaction.
Unsupported NiMoS2-HT and NiMoS2-ME catalysts were also
tested for the CO hydrogenation reaction at T=3108C and P=
60 bar. Changes in CO conversion with time on stream are il-
lustrated in Figure 6. In the catalytic tests with NiMoS2-HT, the
CO conversion level remained stable during 92 h of reaction
whereas the activity of NiMoS2-ME decreased in the last 10 h
of reaction. High catalytic activity was observed for NiMoS2-ME
Table 2. XPS data of NiMoS2-LP prepared by using HT and ME methods,
analysed before and after reaction tests.[a]
Element
XPS data [at %]
NiMoS2-LP-MEF NiMoS2-LP-MES NiMoS2-LP-HTF NiMoS2-LP-HTS
Mo3d
Ni2p
S2p
Mg2s
Si2p
C1s
1.91
0.2
1.05
5.7
15.6
37.3
1.7
0.14
0
8.0
18.0
2.74
1.31
5.48
2.8
0.82
4.37
7.3
12.7
23.1
6.67
12.09
25.12
46.59
32.68
39.02
O1s
38.23
48.79
Figure 6. CO conversion as a function of time on stream for NiMoS2-HT and
[a] F=fresh catalysts; S=spent catalysts.
NiMoS2-ME catalysts performed at 3108C, 60 bar, and a GHSV of 1044 hꢀ1
.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2394 – 2402 2397

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
catalysts, which resulted in an average CO conversion of
33 mol%. In contrast, NiMoS2-HT catalysts converted only
23 mol% of CO into products.
The product selectivity of the catalysts is given in Table 3.
With the NiMoS2-HT catalyst, CO hydrogenation mainly yielded
methane (30.2 mol%) and carbon dioxide (43.8 mol%), which
Figure 8. SEM images of a) unsupported NiMoS2-HT and b) NiMoS2-ME cata-
lysts. Scale bars=10 mm.
Table 3. CO conversion and selectivity of NiMoS2-HT and NiMoS2-ME
catalysts.[a]
Variable
NiMoS2-HT
23
NiMoS2-ME
33
XCO[b] [mol%]
S[c] [mol%]
By using nitrogen physisorption isotherms, we measured
a BET surface area of 120 m2 gꢀ1 and a pore volume of 0.45 mL
for NiMoS2-ME catalysts. A BET surface area of 6 m2 gꢀ1 was
found for NiMoS2-HT catalysts. The TEM and SEM images of the
HT-based catalysts showed highly ordered sulfide slabs, which
was in agreement with the small BET surface area and pore
volume found for the sample. A large BET surface area of the
NiMoS2-ME catalyst was associated with the highly disordered
sulfide layers and carbon residue (26 wt%); the latter may act
as a dispersant for particles by creating a porous network.
hydrocarbons
C1ꢀOH
30.2
5.25
8.7
0.75
0.28
15
14.4
28.1
26.8
4.27
1.83
61
C2ꢀOH
C3ꢀOH
other oxygenates
alcohol
C2ꢀOH in total alcohol [mol%]
[a] Reaction conditions: P=60 bar, GHSV=1044 hꢀ1, H2/CO=2; [b] CO
58
44
conversion; [c] Selectivity based on C-containing products, including CO2.
Surface analysis of unsupported NiMoS2 catalysts
indicated a low selectivity for the formation of alcohol. The ME
-based catalysts resulted in a high percentage of alcohol
The major differences were found in the XPS data of the HT-
and ME-based catalysts; an XPS analysis was performed for the
fresh and spent catalysts. The XPS spectra and one example of
curve fittings are presented in Figure 9a and b (see the Sup-
porting Information for detailed curve fitting). An XPS survey
scan of the fresh NiMoS2-ME catalyst demonstrated strong sig-
nals of oxygen and carbon, O1s and C1s, and weak signals of
Mo3d and S2p (Table 4). The NiMoS2-ME catalyst yielded
a binding energy of 229.3 and 232.7 eV, which were character-
istic of Mo3d5/2 (Mo4+) and Mo3d5/2 (Mo6+), respectively.[15b] An
analysis of the XPS spectrum of the sulfur region revealed two
S2p3/2 doublets with a binding energy of 162 and 164 eV,
which indicated the presence of S2ꢀ ions and S22ꢀgroups.[26]
Correlating the XPS spectra of sulfur (S2p), molybdenum
(Mo3d), and a S/Mo atomic ratio of 2, we assigned a binding
energy of 229.3 eV to MoS2; however, this does not exclude
the presence of MoO2.
(61 mol%) and significantly
a low amount of methane
(14.4 mol%) and carbon dioxide (23 mol%). As the same
amount of the NiMoS2 active catalyst was loaded into the reac-
tor for both cases, such a significant difference in catalytic
properties was unexpected, particularly in product selectivity.
To clarify the major differences between the unsupported
HT- and ME -based catalysts, TEM and SEM analyses were per-
formed. As shown in TEM (Figures 7) and SEM (Figure 8)
images, the morphology of NiMoS2 depended on the synthesis
method. The TEM images of the NiMoS2-ME catalyst demon-
strated highly disordered MoS2 layers; in some areas, up to
three layers were detected. In contrast, stacks of MoS2 contain-
ing six to seven sulfide layers and well-crystallised sulfide slabs
were observed on the TEM images of the NiMoS2-HT catalyst.
These TEM images indicated that the ME method produced
highly disordered and short sulfide layers (ꢁ10 nm) whereas
the HT method produced continuous and well-crystallised mul-
tilayers of MoS2. The SEM images indicated that the NiMoS2-ME
catalyst contained small and plate-shaped particles whereas
the NiMoS2-HT catalyst consisted of large aggregates.
The analysis of the XPS spectra revealed a small peak at
229.3 eV, which was characteristic of Mo3d5/2 (Mo4+), in Mo3d
regions of the spent NiMoS2-ME catalyst. In addition, the spec-
trum demonstrated some contribution from Mo5+ with
230.9 eV and a large peak corresponding to Mo3d5/2 (Mo6+
)
with a binding energy of 232.8 eV. Measurements in the S2p
region detected sulfur in a significantly low concentration
(2.3 at%), which yielded one S2p3/2 doublet with a binding
energy of 162.2 eV. The surface atomic ratio of S/Mo corre-
sponded to 0.33. An increase in Mo3d5/2 (Mo6+) contribution
and low sulfur content indicated that a significant amount of
sulfur was lost during the reaction, which led to the oxidation
of Mo4+ species to MoV and MoVI oxides.
The XPS spectra of NiMoS2-HT catalysts are shown in Fig-
ure 9a and b. A minor difference was observed in XPS data of
Figure 7. TEM images of unsupported a) NiMoS2-HT and b) NiMoS2-ME cata-
lysts. Scale bars=20 nm (a) and 10 nm (b).
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2394 – 2402 2398

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
the fresh and spent NiMoS2-HT catalysts, particularly in the con-
centration of surface elements and in Mo3d and S2p line posi-
tions (Table 4).
The fresh NiMoS2-HT catalysts in Mo3d regions contained
three molybdenum oxidation states: 1) Mo3d5/2 (Mo4+) at
229.3 eV; 2) Mo3d5/2 (Mo5+) at 230.6 eV; 3) Mo3d5/2 (Mo6+) at
232.8 eV; these line positions remained unchanged in the
spent NiMoS2-HT catalysts. Regarding the S2p regions, a weak
and broad shoulder at 169 eV, which is consistent with S2p3/2
for SO4 groups, existed together with two S2p3/2 doublets at
162.2 and 164.5 eV, respectively. Compared with the NiMoS2-
ME catalyst, the NiMoS2-HT catalyst retained its surface sulfur
content during catalytic testing, which gave a S/Mo atomic
ratio of 2 for both fresh and spent NiMoS2-HT catalysts.
The total concentration of (molybdenum, nickel, and sulfur)
elements in fresh and spent catalysts was also analysed by
using inductively coupled plasma–mass spectrometry (ICP–
MS), and the results are summarised in Table 4. The ICP–MS re-
sults revealed that the NiMoS2-ME catalyst lost a significant
amount of sulfur during the reaction, which concurs with the
XPS results. The chemical composition of the NiMoS2-ME cata-
lyst changed from Ni0.41MoS1.93 to Ni0.48MoS0.75 after the reac-
tion; in the NiMoS2-HT catalyst, the composition changed to
Ni0.43MoS2.11 (fresh) and Ni0.47MoS1.77 (spent).
Correlation between sulfur loss and catalytic activity
During the exposure of the MoS2 catalyst to the syngas feed,
a loss of sulfur was observed. According to previous reports,
sulfur located on the edges of the sulfide layers reacts with hy-
drogen, which leaves coordinatively unsaturated molybdenum
ions and anionic vacancies.[21,27] In hydrotreating MoS2 cata-
lysts, the adsorption of molecules such as O2, CO, N2, and NH3
occurred mainly on the anionic vacancy sites of sulfide cata-
lysts.[28] A correlation between the active sites and the im-
proved catalytic activity was reported for hydrotreating cata-
lysts.[29] For other sulfide catalysts (NiWS) used in hydrocrack-
ing, hydrodesulfurisation, and hydrogenation reactions, the
anionic vacancies (sulfur-deficient metal sites) were the catalyt-
ic active centres; these could be blocked in the presence of
H2S.[30] A similar phenomenon could also occur with the MoS2
catalyst during the alcohol synthesis reaction. The catalyst, pre-
pared by using ME methods, had higher selectivity
Figure 9. a) XPS results of fresh and spent NiMoS2-HT and NiMoS2-ME cata-
lysts. High-resolution scan in Mo3d regions. b) XPS results of fresh and
spent NiMoS2-HT and NiMoS2-ME catalysts. High-resolution scan in S2p
regions.
towards alcohol than towards hydrocarbons. Based
on the XPS results, which revealed a substantial de-
Table 4. XPS data and ICP–MS results of unsupported NiMoS2 prepared by using HT
and ME methods, analysed before and after reaction tests.[a]
crease in sulfur content after the reaction [Ni0MoS1.96
(fresh); Ni0.17MoS0.33 (spent)] and catalytic tests, it is
possible to correlate the high alcohol selectivity with
the coordinatively unsaturated molybdenum sites.
Assuming that any alcohol precursors originate from
the coordinatively unsaturated molybdenum sites,
a possible mechanism for alcohol formation could be
drawn as shown in Scheme 1.
NiMoS2-MEF
NiMoS2-MES
NiMoS2-HTF
NiMoS2-HTS
XPS data [at%]
Mo3d
5.16
6.89
7.98
8.59
S2p
10.16
2.31
15.13
17.16
Ni2p
0
1.19
1.76
1.82
C1s
67.14
61.9
39.45
33.54
O1s
17.56
Ni0MoS1.96
Ni0.41MoS1.93
27.72
Ni0.17MoS0.33
Ni0.48MoS0.75
35.03
Ni0.22MoS1.89
Ni0.43MoS2.11
38.89
Ni0.21MoS1.99
Ni0.47MoS1.77
Surface composition[b]
Elemental composition[c]
This mechanism supports the reaction path in
which non-dissociated CO molecules are absorbed
into CH3-metyl species, which is also reported by Mei
and other researchers.[31] As the NiMoS2-ME catalysts
[a] F=fresh catalysts; S=spent catalysts. [b] Based on XPS analysis. [c] Based on ICP–
MS analysis.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2394 – 2402 2399

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
of data presented herein, we conclude that the ME method
produces highly active NiMoS2 catalysts with higher ethanol se-
lectivity than the HT method. The ME-based catalyst has disor-
dered sulfide structures, sulfur vacancies, and large BET surface
area. Our results emphasise that the highly disordered sulfide
structures in combination with coordinatively unsaturated mo-
lybdenum sites is the basis for the synthesis of ethanol from
syngas.
Scheme 1. Mechanism for the formation of ethanol via adsorption of CO
molecules in sulfur-vacancy molybdenum sites.
Conclusions
The aim of this study was to synthesise highly active catalysts
for ethanol synthesis from syngas. A number of NiMoS2 cata-
lysts (laponite supported and unsupported) were synthesised
by using the hydrothermal and microemulsion (ME) methods.
The unsupported NiMoS2 catalyst prepared by the using ME
method demonstrated a larger BET surface area (120 m2 gꢀ1),
higher CO hydrogenation activity (33 mol% CO conversion),
and higher selectivity toward alcohol (61 mol%) than those
prepared by using the hydrothermal-based NiMoS2 catalyst.
The X-ray photoelectron spectroscopy and inductively coupled
plasma–mass spectrometry results of the spent catalysts re-
vealed a significant loss of sulfur, particularly in NiMoS2-ME cat-
alyst. The high catalytic activity of the NiMoS2-ME catalyst is as-
sociated with its highly disordered sulfide layers, which are
readily reduced in syngas atmosphere by releasing sulfur from
its structure. The release of sulfur from the catalyst structures
creates anionic vacancies that promote the formation of alco-
hol via CH3-methyl and CO coupling.
contain layered sulfide (up to two layers), sulfur located on
other than surface planes and edge sites most likely remains in
the structure and contributes to the formation of CH3-metyl
species through the CO and H2 dissociation occurring on
charge neutral molybdenum sites or fully sulfurised molybde-
num sites. Moreover, the dissociation of hydrogen on sulfur-
deficient sites could not be ruled out.
A further question arises as to whether these coordinatively
unsaturated molybdenum sites are stable. Product selectivity
and CO conversion change over time; other factors (in addition
to sulfur loss) also lead to the change in catalytic activity, such
as pore blocking, particle sintering, and carbon deposition.
Hensley et al. claimed that the presence of a sulfur source
(H2S) in the syngas feed is crucial because the MoIV catalyst oxi-
dises over time to form MoVI oxides, which changes selectivity
from alcohol to hydrocarbons.[32] However, in their study, the
presence of H2S in the syngas feed resulted in 81% alcohol
and 17% hydrocarbon selectivity (D6-sample name, please see
Ref. [32]) whereas the catalyst (D2) tested in H2S-free syngas
yielded 77% alcohol and 19% hydrocarbon selectivity . The dif-
ferences in product selectivity are not apparent, considering
these reactions have continued for 350 (D2) and 3840 h (D6),
respectively. The study by Christensen et al.[13a] also observed
differences in product selectivity. Upon addition of H2S, they
found chain growth and increase in selectivity towards hydro-
carbons. It is most likely that H2S reacts with reaction inter-
mediates and promotes the formation of long chains of hydro-
carbons. By using H2S-free syngas feed, Gang et al.[13b] ob-
served no apparent change in the catalytic activity and alcohol
selectivity of the K–Mo/Co/C sulfide catalyst after 1000 h of re-
action. Dianis[33] observed a decrease in the alcohol selectivity
of MoS2 with feed containing 150 ppm H2S; in contrast, the al-
cohol selectivity of a Co–Mo–S catalyst was unaffected by H2S
addition. They postulated that the increased alcohol selectivity
is the result of the weakly adsorbed hydrogen and strongly ad-
sorbed CO. As numerous reactions occur simultaneously on
the surface of a catalyst, the possibility of blocking the coordi-
natively unsaturated molybdenum sites is high; however, their
presence plays a major role in alcohol synthesis. If there is
a sudden decrease in alcohol selectivity, one must perform
a treatment of the catalyst with the sulfur source (H2S). Co-
feeding with a sulfur source (H2S) during the reaction inhibits
the formation of coordinatively unsaturated molybdenum sites
and changes the product selectivity of a catalyst. On the basis
Experimental Section
Laponite-supported NiMoS2-ME catalyst preparation
The clay, Laponite RD (Laporte Industries Ltd, USA), was used as
a support material. The laponite powder (0.5 g) was dispersed in
water (50 mL). The suspension was stirred until it formed an
opaque solution. Meanwhile, the oil phase containing cyclohexane
(100 mL) and a non-ionic surfactant (6 mL; Brij-30, Sigma–Aldrich)
was stirred in a beaker at RT. Then, a sulfur source (5 mL; 21 wt%
(NH4)2S solution, Sigma–Aldrich) was added to the oil phase. After
several minutes of stirring, an aqueous solution of
(NH4)6Mo7O24·4H2O (2 mL, 25 wt%; Sigma–Aldrich; 83.0% MoO3
basis) was added dropwise to the water-in-oil (w/o) ME, which was
followed by the addition of an aqueous solution of Ni(NO3)2·6H2O
(1 mL; 37.5 wt%; Sigma–Aldrich). The black w/o ME system was
stirred for 1 h to allow sufficient mixing. Laponite (dispersed in
50 mL of water) was added to the w/o ME system. The addition of
laponite destabilised w/o ME systems, which led to the deposition
of precipitants (molybdenum, nickel, and sulfur) onto the laponite
surface. Subsequently, cyclohexane was removed with a rotary
evaporator. The black precipitate was heat-treated in N2 atmos-
phere at 3508C for 4 h to remove the surfactant.
Unsupported NiMoS2-ME catalyst preparation
The experimental method was the same as that for the supported
catalysts, except the laponite solution (laponite/water) was not
added.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2394 – 2402 2400

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
every 12 h and analysed with a gas chromatograph (GC-8A,
Shimadzu) equipped with a flame ionisation detector.
Laponite-supported NiMoS2-HT catalyst preparation
The laponite-supported NiMoS2 catalyst was prepared by using the
HT method as follows: Laponite (5 g) was dispersed in water
(250 mL). The suspension was stirred until it formed an opaque so-
lution. The surfactant (10 mL; TERGITOL 15-S-9, Sigma–Aldrich) was
added to the laponite solution. The suspension was then stirred
for 2 h to allow sufficient mixing. Meanwhile, two aqueous solu-
tions of (NH4)2MoS4 and Ni(NO3)2 were prepared. These two solu-
tions were added dropwise over a 1 h period to the stirred suspen-
sion (laponite/surfactant/H2O). The Mo/Ni molar ratio was 2:1. After
prolonged stirring for 3 h at 608C, the resulting black slurry was
transferred to an autoclave and kept at 1308C for 24 h. The black
precipitate was recovered from the mixture by centrifuging and
washing with deionised water. The wet cake was dried in air at
1008C, followed by heat-treatment at 3508C for 4 h in N2
atmosphere.
Acknowledgements
This research forms part of the Baosteel Australia Research and
Development Centre (BAJC) portfolio of projects and has received
support through the project BAJC Australian Research Council
Linkage Project (ARCLP) (project no. BALP1100). We acknowledge
the Australian Research Council (ARC) for partially funding the
project and the ARC Centre of Excellence for Functional Nanoma-
terials.
Keywords: hydrothermal synthesis
molybdenum · nanotechnology · sulfur
·
microemulsions
·
[1] a) J. J. Spivey, A. Egbebi, Chem. Soc. Rev. 2007, 36, 1514–1528; b) V. Sub-
ramani, S. K. Gangwal, Energy Fuels 2008, 22, 814–839; c) M. Gupta,
M. L. Smith, J. J. Spivey, ACS Catal. 2011, 1, 641–656.
[2] a) S. Zaman, K. J. Smith, Catal. Rev. Sci. Eng. 2012, 54, 41–132; b) R. G.
Herman, Catal. Today 2000, 55, 233–245.
Unsupported NiMoS2-HT catalyst preparation
The experimental method was the same as that for the supported
catalysts, excluding the addition of the laponite solution (laponite/
H2O).
[3] a) R. Andersson, M. Boutonnet, S. Jaeras, J. Chromatogr. A 2012, 1247,
134–145; b) V. R. Surisetty, I. Eswaramoorthi, A. K. Dalai, Fuel 2012, 96,
77–84.
[4] M. M. Bhasin, W. J. Bartley, P. C. Ellgen, T. P. Wilson, J. Catal. 1978, 54,
120–128.
Catalyst characterisation
The crystalline phase of the laponite-supported NiMoS2 catalyst
was identified with X-ray diffractometer (MiniFlex) using CoKa radi-
ation. The internal morphology of the samples was observed with
a high-resolution transmission electron microscope (JEOL-2100) op-
erating at 200 kV. The N2 adsorption/desorption isotherms of the
samples at ꢀ1968C were obtained with a TriStar II surface analyser
(Micromeritics). The XPS data were obtained with an Axis Ultra X-
ray photoelectron spectrometer (Kratos) equipped with a 165 mm
hemispherical electron energy analyser. The incident radiation was
monochromatic AlKa X-rays (1486.6 eV) at 225 W (15 kV, 15 mA).
Survey (wide) scans were taken at an analyser pass energy of
160 eV and multiplex (narrow) high-resolution scans at 20 eV.
Survey scans were taken in the binding energy range of 1200–0 eV
with 1.0 eV steps and 100 ms dwell time. Narrow high-resolution
scans were run with 0.05 eV steps and 250 ms dwell time. Base
pressure in the analysis chamber was 1.0ꢁ10ꢀ9 Torr (1 Torr=
133.3 Pa), and during sample analysis, it was 1.0ꢁ10ꢀ8 Torr. Atomic
concentrations were calculated with the CasaXPS software (version
2.3.14) and a Shirley baseline with Kratos library relative sensitivity
factors. The peak fitting of the high-resolution data was also per-
formed with the CasaXPS software. A CHNS elemental analyser
(FlashEA 1112 series, Thermo Electron Corporation) was used to de-
termine the carbon content.
[5] J. A. Dalmon, P. Chaumette, C. Mirodatos, Catal. Today 1992, 15, 101–
127.
[6] a) Y. Lu, F. Yu, J. Hu, J. Liu, Appl. Catal. A-Gen. 2012, 429–430, 48–58;
b) J. Gong, H. Yue, Y. Zhao, S. Zhao, L. Zhao, J. Lv, S. Wang, X. Ma, J. Am.
Chem. Soc. 2012, 134, 13922–13925; c) F. Pio, T. Enrico, P. Italo, Catal.
Rev. 1991, 33, 109–168.
[7] J. M. Christensen, P. A. Jensen, N. C. Schiødt, A. D. Jensen, ChemCatChem
2010, 2, 523–526.
[8] a) M. M. Conway, C. B. Murchison, R. R. Stevens, Dow Chemical Co., USA,
1986, p. 26 pp; b) C. B. Murchison, M. M. Conway, R. R. Stevens, G. J.
Quarderer, Vol. 2, Chem. Inst. Can. 1988, pp. 626–633.
[9] a) S. Venkateswara Rao, D. Ajay Kumar, K. Janusz, Appl. Catal. A-Gen.
2010, 385, 153–162; b) Z. Sharif, J. S. Kevin, Catal. Rev. 2012, 54, 41–
132.
[10] V. R. Surisetty, Y. Hu, A. K. Dalai, J. Kozinski, Appl. Catal. A-Gen. 2011, 392,
166–172.
[11] M. Xiang, D. Li, H. Xiao, J. Zhang, W. Li, B. Zhong, Y. Sun, Catal. Today
2008, 131, 489–495.
[12] a) P. Afanasiev, G.-F. Xia, G. Berhault, B. Jouguet, M. Lacroix, Chem.
Mater. 1999, 11, 3216–3219; b) P. Afanasiev, C. R. Chim. 2008, 11, 159–
182.
[13] a) J. M. Christensen, P. M. Mortensen, R. Trane, P. A. Jensen, A. D. Jensen,
Appl. Catal. A-Gen. 2009, 366, 29–43; b) L. Gang, Z. Chengfang, C. Yanq-
ing, Z. Zhibin, N. Yianhui, C. Linjun, Y. Fong, Appl. Catal. A-Gen. 1997,
150, 243–252; c) R. Andersson, M. Boutonnet, S. Jꢂrꢃs, Appl. Catal. A-
Gen. 2012, 417–418, 119–128.
The catalytic tests were performed at 3108C, 60 bar, and a GHSV of
1044 hꢀ1 in a fixed-bed high-pressure reactor. The stainless steel
fixed-bed reactor (internal diameter: 10 mm; length: 100 mm)
equipped with a thermocouple was inserted in a furnace. The cata-
lyst was reduced in H2 atmosphere at 4508C for 4 h before the re-
action. In all catalytic tests, premixed syngas with a H2/CO ratio of
2:1 and 4% of N2 (internal standard) was used as a feed. The mass
flow controller (Bronkhorst High-Tech B.V.) regulated the inlet gas
flow rate. Gaseous products leaving the reactor was passed
through the condenser to obtain the liquid fraction of products.
The outlet gas stream composition was analysed with a gas chro-
matograph (GC-2014, Shimadzu) equipped with a thermal conduc-
tivity detector. Liquid samples were discharged from the condenser
[14] a) R. R. Chianelli, G. Berhault, B. Torres, Catal. Today 2009, 147, 275–286;
b) J. Grimblot, Catal. Today 1998, 41, 111–128.
[15] a) Y. Iwata, K. Sato, T. Yoneda, Y. Miki, Y. Sugimoto, Catal. Today 1998 45,
353–359; b) J. Kibsgaard, Z. Chen, B. Reinecke, T. Jaramillo, Nat. Mater.
2012, 11, 963–969; c) Y. Boonyawan, K. Jae Hyung, S. Chunshan, N. Cha-
walit, P. Pattarapan, Catal. Today 2008, 130, 14–23; d) P. Ratnasamy, A. J.
Lꢄonard, J. Catal. 1972, 26, 352–358.
[16] E. Devers, P. Afanasiev, B. Jouguet, M. Vrinat, Catal. Lett. 2002, 82, 13–
17.
[17] D. Genuit, P. Afanasiev, M. Vrinat, J. Catal. 2005, 235, 302–317.
[18] a) S. Harris, R. R. Chianelli, J. Catal. 1986, 98, 17–31; b) J. V. Lauritsen,
M. V. Bollinger, E. Laegsgaard, K. W. Jacobsen, J. K. Norskov, B. S. Clau-
sen, H. Topsoe, F. Besenbacher, J. Catal. 2004, 221, 510–522; c) J. V.
Lauritsen, S. Helveg, E. Laegsgaard, I. Stensgaard, B. S. Clausen, H.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2394 – 2402 2401

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Topsoe, F. Besenbacher, J. Catal. 2001, 197, 1–5; d) J. V. Lauritsen, J.
Kibsgaard, G. H. Olesen, P. G. Moses, B. Hinnemann, S. Helveg, J. K. Nor-
skov, B. S. Clausen, H. Topsoe, E. Laegsgaard, F. Besenbacher, J. Catal.
2007, 249, 220–233; e) Y. Aray, J. Rodrꢅguez, ChemPhysChem 2001, 2,
599–604.
[27] F. E. Massoth, C. L. Kibby, J. Catal. 1977, 47, 300–315.
[28] a) R. R. Chianelli, A. F. Ruppert, S. K. Behal, B. H. Kear, A. Wold, R. Ker-
shaw, J. Catal. 1985, 92, 56–63; b) D. G. Kalthod, S. W. Weller, J. Catal.
1985, 95, 455–464.
[29] a) W. S. Millman, W. Keith Hall, J. Catal. 1979, 59, 311–313; b) E. A. Lom-
bardo, M. Lo Jacono, W. K. Hall, J. Catal. 1980, 64, 150–162; c) T. A. Bod-
rero, C. H. Bartholomew, J. Catal. 1983, 84, 145–155.
[19] a) R. R. Chianelli, A. F. Ruppert, M. Josꢄ-Yacamꢆn, A. Vꢆzquez-Zavala,
Catal. Today 1995, 23, 269–281; b) L. P. Hansen, Q. M. Ramasse, C. Kisie-
lowski, M. Brorson, E. Johnson, H. Topsøe, S. Helveg, Angew. Chem.
2011, 123, 10335–10338; Angew. Chem. Int. Ed. 2011, 50, 10153–10156.
[20] O. Sørensen, B. S. Clausen, R. Candia, H. Topsøe, Appl. Catal. 1985, 13,
363–372.
[21] P. Jean-FranÅois, P. Edmond, J. Phys. Chem. B 2003, 107, 4057–4064.
[22] J. Kibsgaard, Atomic-scale Investigation of MoS2-based Hydrotreating
Model Catalysts: A Scanning Tunneling Microscopy Study, PhD Thesis,
Faculty of Science University of Aarhus, 2008.
[30] a) R. J. H. Voorhoeve, J. C. M. Stuiver, J. Catal. 1971, 23, 228–235;
b) R. J. H. Voorhoeve, J. Catal. 1971, 23, 236–242.
[31] a) D. Mei, R. Rousseau, S. M. Kathmann, V.-A. Glezakou, M. H. Engelhard,
W. Jiang, C. Wang, M. A. Gerber, J. F. White, D. J. Stevens, Journal of Cat-
alysis 2010, 271, 325–342; b) T. Tatsumi, A. Muramatsu, K. Yokota, H. To-
minaga, J. Catal. 1989, 115, 388–398; c) J. M. Andrew, C. L. Adam, A.-P.
Frank, T. Burcin, C. S. Niels, K. N. Jens, S. Felix, Top. Catal. 2013, 57, 135–
142; d) M. Xu, E. Iglesia, J. Catal. 1999, 188, 125–131.
[23] G.-z. Bian, Y.-l. Fu, M. Yamada, Appl. Catal. A-Gen. 1996, 144, 79–91.
[24] Polymer Nanocomposites By Emulsion And Suspension Polymerization
(Ed.: V. Mittal), [In: RSC Nanosci. Nanotechnol. 2011, 16], Royal Society
of Chemistry, 2011.
[32] J. E. Hensley, S. Pylypenko, D. A. Ruddy, J. Catal. 2014, 309, 199–208.
[33] W. P. Dianis, Appl. Catal. 1987, 30, 99–121.
[25] S. Brunauer, L. S. Deming, W. E. Deming, E. Teller, J. Am. Chem. Soc.
1940, 62, 1723–1732.
[26] A. M. de Jong, H. J. Borg, L. J. van Ijzendoorn, V. G. F. M. Soudant, V. H. J.
de Beer, J. A. R. van Veen, J. W. Niemantsverdriet, J. Phys. Chem. 1993,
97, 6477–6483.
Received: February 21, 2014
Published online on July 15, 2014
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2394 – 2402 2402