An efficient Ni-Mo-K sulfide catalyst doped with CNTs for conversion of syngas to ethanol and higher alcohols

Article abstract of DOI:10.1016/j.apcata.2013.08.026

A type of Ni-Mo-K sulfide catalyst doped with CNTs for conversion of syngas to ethanol and higher alcohols was developed, and displayed high activity and selectivity for direct synthesis of C_1-4-alcohols, especially ethanol, from syngas. Over a Ni_0.5Mo₁K_0.5- 15%CNTs catalyst under the reaction conditions of 8.0 MPa and 593 K, the S(total oxy.) reached 64.1% (CO₂-free), with the corresponding STY(total oxy.) being 113 mg h^-1 g^-1. Ethanol was the dominant product, with S(EtOH) and STY(EtOH) reaching 33.1% (CO₂-free) and 55.6 mg h^-1 g^-1, respectively. This STY(EtOH)-value was 1.47 times that (37.9 mg h^-1 g^-1) of the CNTs-free counterpart under the same reaction conditions. Addition of a minor amount of CNTs to the sulfurized Ni_0.5Mo₁K_0.5 catalyst caused little change in the E_a for the hydrogenation-conversion of syngas. Appropriately reducing CNT's grain-size could improve its capability to adsorb hydrogen, thus increasing CO hydrogenation-conversion, yet did not influence selectivity of the products. The present work demonstrated that CNTs as promoter function through their adsorbing/activating H₂ to generate a surface micro-environment with higher stationary-state concentration of H-adspecies on the functioning catalyst. This resulted in a dramatic increase, at the surface of the functioning catalyst, of the molar percentage of catalytically active Mo⁴⁺/Mo⁵⁺ species in the total amounts of surface Mo. On the other hand, those active H-species adsorbed at the CNTs surface could be readily transferred to Ni_iMo_jK _k active sites via the CNT-assisted hydrogen spillover. The aforementioned two factors both were conducive to increasing the rate of hydrogenation conversion of syngas.

Full text of DOI:10.1016/j.apcata.2013.08.026

Applied Catalysis A: General 468 (2013) 44–51
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
An efficient Ni–Mo–K sulfide catalyst doped with CNTs for conversion
of syngas to ethanol and higher alcohols
Ji-Jie Wang, Jian-Rong Xie, Yan-Hui Huang, Bing-Hui Chen, Guo-Dong Lin,
Hong-Bin Zhang∗
College of Chemistry and Chemical Engineering, State Key Laboratory of Physical Chemistry for Solid Surfaces and National Engineering Laboratory for
Green Chemical Productions of Alcohols–Ethers–Esters, Xiamen University, Xiamen 361005, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 11 June 2013
Received in revised form 8 August 2013
Accepted 15 August 2013
Available online 26 August 2013
A type of Ni–Mo–K sulfide catalyst doped with CNTs for conversion of syngas to ethanol and higher
alcohols was developed, and displayed high activity and selectivity for direct synthesis of C1–4-alcohols,
especially ethanol, from syngas. Over a Ni0.5Mo1K0.5–15%CNTs catalyst under the reaction conditions of
8.0 MPa and 593 K, the S(total oxy.) reached 64.1% (CO2-free), with the corresponding STY(total oxy.)
being 113 mg h−1 g−1. Ethanol was the dominant product, with S(EtOH) and STY(EtOH) reaching 33.1%
(CO2-free) and 55.6 mg h−1 g−1, respectively. This STY(EtOH)-value was 1.47 times that (37.9 mg h−1 g−1
)
Keywords:
of the CNTs-free counterpart under the same reaction conditions. Addition of a minor amount of CNTs
to the sulfurized Ni0.5Mo1K0.5 catalyst caused little change in the Ea for the hydrogenation–conversion
of syngas. Appropriately reducing CNT’s grain-size could improve its capability to adsorb hydrogen, thus
increasing CO hydrogenation–conversion, yet did not influence selectivity of the products. The present
work demonstrated that CNTs as promoter function through their adsorbing/activating H2 to generate a
surface micro-environment with higher stationary-state concentration of H-adspecies on the functioning
catalyst. This resulted in a dramatic increase, at the surface of the functioning catalyst, of the molar
percentage of catalytically active Mo4+/Mo5+ species in the total amounts of surface Mo. On the other
hand, those active H-species adsorbed at the CNTs surface could be readily transferred to NiiMojKk active
sites via the CNT-assisted hydrogen spillover. The aforementioned two factors both were conducive to
increasing the rate of hydrogenation conversion of syngas.
Multiwalled carbon nanotubes
CNT-doped Ni–Mo–K sulfide catalyst
Higher alcohol synthesis
Ethanol synthesis
Syngas
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
The best results were obtained over a catalyst containing 2.5 wt%
Rh supported on SiO2 and promoted by 0.05 wt% Fe. At 573 K
Ethanol is being considered as a potential alternative synthetic
fuel or fuel-additive for use in automobiles, and as a potential
hydrogen source for fuel cells. Renewable ethanol can also serve
as a feedstock for the synthesis of a variety of industrial chemi-
cals and polymers. Ethanol can be produced from biomass or coal,
and, recently, a worldwide interest in the synthesis of ethanol
from biomass- and coal-derived syngas is growing. Significant
improvements in catalyst design need to be achieved to make this
conversion commercially attractive [1].
The direct conversion of syngas to ethanol and other C2-
oxygenates has been reported by Union Carbide Company as early
as 1975 over SiO2-supported Rh catalyst promoted by metal ions,
such as Fe, Mo, Mn, W, Th, and U, in a stirred autoclave reactor [2–5].
and 1030 psi using H2/CO, the catalyst produced 49% methane,
2.8% methanol, 31.4% ethanol, and 9.1% acetic acid. However, the
rates of methanol and ethanol production were about 50 g/(L-cat. h).
Following this early work, there are several reports on the con-
version of syngas to ethanol and other C2+-oxygenates using a
wide range of noble metals-based catalysts containing Rh, Ru, and
Re supported on various oxides, such as SiO2, Al2O3, CeO2, ZrO2
and MgO [3–5]. Although Rh-based catalysts show promise, the
commercial viability of these catalysts is questionable due to its
availability.
Development of non-noble metals-based catalysts industrially
viable for direct synthesis of ethanol and higher alcohols from syn-
gas has been the other focus of R&D efforts. The catalysts reported
in previous literatures include mainly modified methanol synthe-
sis catalysts, modified Fischer–Tropsch synthesis catalysts, and
MoS2-based catalysts [6–16]. Most of those catalysts have been
extensively studied for the synthesis of mixed higher alcohols con-
taining a small amount of ethanol from syngas, and yet few of
them are suitable for the hydrogenation-conversion of syngas with
∗
Corresponding author. Tel.: +86 592 2181888; fax: +86 592 2181888.
E-mail addresses: wangjijie1985@xmu.edu.cn (J.-J. Wang), jianronx@xmu.edu.cn
(J.-R. Xie), yhhuang@xmu.edu.cn (Y.-H. Huang), chenbh@xmu.edu.cn (B.-H. Chen),
gdlin@xmu.edu.cn (G.-D. Lin), hbzhang@xmu.edu.cn (H.-B. Zhang).
0926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.08.026

J.-J. Wang et al. / Applied Catalysis A: General 468 (2013) 44–51
45
ethanol as the major product. Recently, there are several reports
on new catalyst systems, such as K2CO3/Co–MoS2/clay catalyst
[12], K and Ni doped ␤-Mo2C catalyst [13], and Al2O3 (or AC)-
suppoted Ni(or Co, Rh)-promoted, K-modified MoS2 catalysts [16],
over which the selectivity to form ethanol and higher alcohols from
hydrogenation–conversion of syngas has improved markedly to
different extent. Nevertheless, from the practical viewpoint, the
selectivity and yield of desired alcohol products of those catalysts
still need to be further improved.
Multi-walled carbon-nanotubes (donated as “CNTs” in later
text) have been drawing increasing attention in recent years. This
kind of nanotube-C material possesses a series of unique features,
such as its highly conductive graphitized tube-wall, nanometer-
sized channel and sp2-C-constructed surface, as well as excellent
performance for adsorption and spillover of hydrogen, which make
CNTs full of promise to be a novel catalyst support or promoter.
The catalytic studies conducted so far on CNT-based systems have
shown encouraging results in terms of activity and selectivity
[17–19].
crushed, and sieved to a size of 40–80 mesh for the activity
evaluation.
2.2. Catalyst evaluation
Activity tests of the catalysts for conversion of syngas to ethanol
and higher alcohols were carried out in a fixed-bed continuous-
flow reactor-GC combination system. Catalyst (1.0 g) was mixed
with 4.0 g quartz sand (inert diluents, 40–80 mesh) in order to
maintain isothermal conditions, and placed in the reactor. Prior to
the reaction, the sample of oxide-precursor of catalyst was in situ
pre-reduced/sulfurized in a H2-carried 5% CS2 gaseous mixture
stream at 0.2 MPa and 1800 mL h−1 g−1. The sulfurization temper-
ature was programmed to rise from room temperature to 673 K
and maintain at 673 K for 6 h, and then lower to desired tempera-
ture for the catalyst test. The hydrogenation–conversion reaction
of syngas was conducted at a stationary state under reaction condi-
tions of 5.0–8.0 MPa, 573–623 K, V(H2)/V(CO)/V(N2) = 45/45/10 and
GHSV = 3000–6000 mL h−1 g−1. Exit gas from the reactor was imme-
diately transported, while maintaining at 473 K, to the sampling
valve of the GC (Model GC-950 by Shanghai Haixin GC Instruments,
Inc.), which was equipped with dual detectors (TCD and FID) and
dual columns filled with carbon molecular sieve (TDX-201, Tianjin
Chem. Reagent Co. Ltd.) and capillary column (TG-BONDQ, USA),
respectively, for online analysis. The former column (2.0 m length)
was used for the analysis of N2 (as an internal standard), CO and
CO2, and the latter (30 m × 0.32 mm × 10 ␮m) for hydrocarbons,
alcohols, and other C-containing hydrogenation products. CO con-
version (denoted as X(CO)) was determined through an internal
standard (N2), and the carbon-based selectivity for the carbon-
containing products, including alcohols, hydrocarbons (HC), and
other oxygenates (noted as Salc., SHC, etc. in later text), was calcu-
lated by an internal normalization method. The description about
calculation procedure of space–time–yield (STY) of the products
is lengthy and was placed in the Supplementary materials. All
data were taken 24 h after the reaction started (unless otherwise
specified).
In the present work,
a type of co-precipitated–deposited
Ni–Mo–K sulfide-based catalyst doped with CNTs was developed.
The catalyst displayed higher activity and selectivity for direct syn-
thesis of ethanol and higher alcohols from syngas, compared to the
CNT-free host catalyst. The catalyst was characterized by means
of TEM, XRD, XPS, H2S/H2-TPS and H2-TPD, and the nature of the
promoter action by CNTs was discussed. The results shed light on
the design of practical catalysts for direct synthesis of ethanol and
higher alcohols from syngas.
2. Experimental
2.1. Catalyst preparation
The CNTs used in the present work were prepared according
to the methods reported previously [20]. Freshly prepared CNTs
were treated with boiling nitric acid (8 mol/L, at 363 K) for 8 h, fol-
lowed by rinsing with de-ionized water twice, and then drying
at 383 K under N2-atmosphere. Open-end CNTs with hydrophilic
surface were then obtained (see Fig. S1). The dried CNTs were
crushed, and sieved to a grain-size of ≤75 ␮m for the catalyst
preparation.
2.3. Catalyst characterization
TEM measurements were performed on Technai F30 or JEM-
1400 transmission electron microscope. XRD measurements were
carried out on an X’Pert PRO X-ray diffractometer (PANalytical)
with Cu K␣ (ꢀ␣1 = 0.15406 nm, ꢀ␣2 = 0.15443 nm) radiation oper-
ating at 40 kV and 30 mA. A continuous scan mode was used to
collect 2Â data from 10◦ to 90◦. X-ray photoelectron spectroscopy
(XPS) measurements were done on a Quantum 2000 Scanning
ESCA Microprobe instrument with Al K␣ radiation (15 kV, 25 W,
hꢁ = 1486.6 eV) under ultrahigh vacuum (5 × 10−7 Pa), calibrated
internally by the carbon deposit C(1s) (Eb = 284.7 eV). Specific
surface area (SSA) was determined by N2 adsorption using a
Micromeritics ASAP 2020 system.
Tests of temperature-programmed reduction/sulfurization by
H2S/H2 (labeled as H2S/H2-TPS) of the catalysts in oxidation state
were conducted on a fixed-bed continuous-flow micro-reactor. An
anhydrous CaCl2-column and a 3A-zeolite column were installed in
sequence at the reactor-exit to remove water vapor yielded from
reduction and sulfurization of metal oxides of the catalyst sample.
Fifty mg of catalyst sample (in size of 40–80 mesh) in oxidation
state was used for each test. The sample was first flushed by an
Ar (of 99.999% purity) stream (30 mL min−1) at 573 K for 60 min to
clean its surface, and then cooled down to room temperature, fol-
lowed by switching to a H2-carried 10 vol% H2S gaseous mixture as
reducing/sulfurizing gas (30 mL min−1) to start the TPS measure-
ment from 298 K to 1073 K. The rate of temperature increase was
A
series of CNT-doped Ni–Mo–K catalysts, denoted as
NiiMojKk–x%CNTs (where x% represented mass percentage),
were prepared by combined co-precipitation and impregnation
method. Two aqueous solutions containing calculated amounts
of Ni(NO3)2·6H2O and (NH4)6Mo7O24·4H2O (all of AR grade),
respectively, were simultaneously added dropwise under ultra-
sonic agitation into a Pyrex flask containing a calculated amount
of CNTs at 353 K. After 30 min of ultrasonic agitation, an appropri-
ate amount of ammonia liquor was added into the aforementioned
solution to adjust and maintain the pH value of the solution at
∼7 so as to form precipitate. The suspension was continuously
stirred for 2 h at 353 K, followed by cooling down to room tem-
perature, aging for 4 h before filtering. The filter cake (precipitate)
was dried at 383 K for 4 h and calcined at 823 K in N2 atmo-
sphere for 4 h, followed by cooling down to room temperature.
They were next impregnated with K2CO3 aqueous solution con-
taining a calculated amount of K by the conventional incipient
wetness method, followed by aging for 10 h, drying at 383 K for
4 h and calcining at 673 K under N2 atmosphere for 4 h, finally
yielding the precursor of CNT-doped Ni–Mo–K catalysts (in oxi-
dation state). A CNT-free oxide precursor of Ni–Mo–K catalyst
(denoted as NiiMojKk), used as reference, was prepared in the
similar manner. All samples of catalyst precursor were pressed,

46
J.-J. Wang et al. / Applied Catalysis A: General 468 (2013) 44–51
10 K min−1. Change of H2S-signal was monitored by an on-line GC
(Shimadzu GC-8A) with a TCD detector.
Tests of H2-temperature-programmed desorption (H2-TPD) of
the catalysts were conducted on an adsorption/desorption system.
Two hundreds mg of sample of the pre-reduced/sulfurized cata-
lyst was used in each test. Prior to the H2-TPD test, the sample
of the pre-reduced/sulfurized catalyst was in situ treated in a H2
(of 99.999% purity) stream (1800 mL h−1) at 673 K for 3 h and then
flushed by an Ar (of 99.999% purity) stream (1800 mL h−1) at 673 K
for 60 min to clean its surface, and then cooled down to 433 K,
followed by switching to the H2 (of 99.999% purity) stream for
hydrogen adsorption for 60 min and subsequently at room temper-
ature for 2 h. Afterwards, the sample was flushed by the Ar stream
at room temperature until the stable baseline in GC appeared. TPD
measurements were then conducted from 298 K to 923 K. The rate
of temperature increase was 5 K min−1. Change of hydrogen-signal
was monitored by an on-line GC (Shimadzu GC-8A) with a TCD
detector.
Fig. 2. Reactivity of syngas conversion to ethanol and higher alcohols over the sulfu-
rized Ni0.5Mo1K0.5–15%CNTs catalysts with the CNTs of varying grain-sizes; reaction
conditions were the same as in Fig. 1.
3. Results and discussion
and 3.03%, respectively. X(CO-hydr.) of the other four catalysts with
i/1 = 0.25, 0.33, 0.75 and 1.0 was 7.7%, 8.0%, 6.6%, 5.3%, successively,
with the corresponding S(total oxy., CO2-free) being 64.0%, 64.4%,
72.0%, 74.5%, and the S(EtOH, CO2-free) being 29.0%, 31.2%, 35.0%,
30.8%, and the Y(EtOH) being 2.23%, 2.50%, 2.31%, 1.63%, succes-
sively. The Ni/Mo molar ratio of 0.5/1 was chosen to be the optimal
one (see Fig. S2).
Finally, the K/Mo molar ratio (k/1) of the Ni0.5Mo1Kk–15%CNTs
catalysts was optimized. The results showed that, over the cata-
lyst with k/1 = 0.5/1 under the aforementioned reaction conditions,
the observed X(CO-hydr.) was 8.4%, with the corresponding S(total
oxy., CO2-free), S(EtOH, CO2-free) and Y(EtOH) being 67.2%, 36.1%
and 3.03%, respectively. X(CO-hydr.) of the other four catalysts with
k/1 = 0.3, 0.4, 0.6 and 0.7 was 5.7%, 7.0%, 7.5% and 5.6%, with the cor-
responding S(total oxy., CO2-free) being 69.9%, 69.5%, 64.5%, 59.3%,
and the S(EtOH, CO2-free) being 29.8%, 33.7%, 33.0%, 29.0%, and the
Y(EtOH) being 1.70%, 2.36%, 2.48%, 1.62%, successively. The K/Mo
molar ratio of 0.5/1 was chosen to be the optimal one (see Fig. S3).
Besides the catalyst composition, the grain-size of the used
CNTs also has a certain effect on the activity of the correspond-
ing catalyst. Fig. 2 shows the reactivity of syngas conversion to
3.1. Optimization of the catalyst composition
Reactivity of conversion of syngas to C1–4-alcohols over a series
of the sulfurized Ni0.5Mo1K0.5–x%CNTs catalysts with varied CNT-
doping amount was first investigated. The results (see Fig. 1)
showed that the observed carbon-containing products involved
oxygenates (simplified as oxy., including C1∼4-alc. and dimethy
ether (DME)) and hydrocarbons (simplified as HC, including C1–3
-
alkanes), as well as CO2 yielding from the water–gas–shift (WGS)
side-reaction. The catalyst with 15% CNT-doping displayed the
best catalytic performance among those tested. Over this cat-
alyst under the reaction conditions of 5.0 MPa and 593 K, the
hydrogenation–conversion of CO (denoted as X(CO-hydr.)) and
S(total oxy., CO2-free) reached 8.4% and 67.2%, respectively. The
corresponding Y(total oxy.) (i.e., yield of the total oxy., defined as
the product of X(CO-hydr.) and S(total oxy., CO2-free)), reached
5.65%. Ethanol was the major hydrogenation product, with the
S(EtOH, CO2-free) and Y(EtOH) being 36.1% and 3.03%, respectively.
Next, the Ni/Mo molar ratio (i/1) of the NiiMo1K0.5–15%CNTs
catalysts was optimized. The results showed that, over the cata-
lyst with i/1 = 0.5/1 under the aforementioned reaction conditions,
the observed X(CO-hydr.) was 8.4%, with the corresponding S(total
oxy., CO2-free), S(EtOH, CO2-free) and Y(EtOH) being 67.2%, 36.1%
C1–4-alcohols over a series of the sulfurized Ni0.5Mo1K0.5–15%CNTs
catalysts with the CNTs of varying grain-sizes. With the grain-size of
the CNTs decreasing progressively from the range of 830–380 ␮m,
X(CO-hydr.) and X(CO-WGS) both increased monotonously before
approaching a plateau at 75–38 ␮m and below. Meanwhile, the
S(total oxy.), S(MeOH), S(EtOH) and S(PrOH), as well as S(total HC),
all maintained stable at a respective level. Thus, in order to achieve
high conversion of CO hydrogenation, the CNTs with the grain-sizes
of ≤75 ␮m was selected.
3.2. Optimization of the reaction operation conditions
3.2.1. Reaction temperature
Reaction temperature has a pronounced effect on the syn-
gas conversion and the selectivity of the formation of ethanol
and higher alcohols. Fig. 3 displays the reactivity of syngas
conversion to ethanol and higher alcohols over the sulfurized
Ni0.5Mo1K0.5–15%CNTs catalyst at varied temperatures. With the
reaction temperature rising progressively from 583 K, X(CO-hydr.),
X(CO-WGS) and Y(total HC) increased monotonously; Y(ethanol)
and Y(total oxy.) increased quickly before reaching a maximum
at 593 K, and then slowly decreased; while Y(MeOH) and Y(DME)
decreased monotonously. In order to achieve both high Y(ethanol)
and Y(total oxy.), 593 K was taken as the optimal operating tem-
perature.
Fig. 1. Reactivity of syngas conversion to ethanol and higher alcohols over the sulfu-
rized Ni0.5Mo1K0.5–x%CNTs catalysts with varying CNT-doping amounts, at reaction
conditions of 5.0 MPa, 593 K, V(H2)/V(CO)/V(N2) = 45/45/10, GHSV = 3000 mL h−1 g−1
(outlet).

J.-J. Wang et al. / Applied Catalysis A: General 468 (2013) 44–51
47
Fig. 3. Reactivity of syngas conversion to ethanol and higher alcohols over the sul-
furized Ni0.5Mo1K0.5–15%CNTs catalyst at varied temperatures; reaction conditions
were the same as in Fig. 1 except T ranging in 573–623 K.
Fig. 4. Operating stability of syngas conversion to ethanol and higher alcohols over
the sulfurized Ni0.5Mo1K0.5–15%CNTs catalyst, at reaction conditions of 8.0 MPa,
593 K, V(H2)/V(CO)/V(N2) = 45/45/10, GHSV = 4000 mL h−1 g−1 (outlet).
3.2.2. Feed-gas GHSV
In order to find out the working conditions with great extent
of reaction, the reaction of syngas converted to ethanol and
higher alcohols was conducted over the Ni0.5Mo1K0.5–15%CNTs
catalyst under reaction conditions of 8.0 MPa, 593 K,
V(H2)/V(CO)/V(N2) = 45/45/10 and varied GHSV of the feed-
gas. The results showed that, with the GHSV rising progressively
from 3000 to 6000 mL h−1 g−1, X(CO-hydr.) and X(CO-WGS) both
decreased monotonously, and S(total oxy.) increased slowly but
S(EtOH) decreased slightly (going down to 31.8% from 34.0%).
Meanwhile, STY(total oxy.) increased slightly before reaching a
maximum at GHSV = 4000 mL h−1 g−1, and then slowly decreased
(see Fig. S4). In order to simultaneously achieve relatively high
X(CO-hydr.) and S(EtOH), as well as STY(total oxy.), GHSV of the
feed-gas was set at 4000 mL h−1 g−1 (outlet).
the Ea for the hydrogenation-reaction of CO, most likely implying
that such doping did not alter the major reaction pathway of CO
hydrogenation.
3.4. Characterization of the catalysts
3.4.1. TEM characterization
Our previous studies [20–22] have demonstrated that this type
of CNTs used in the present work was a “Herringbone-type” of
multi-walled nanotubes, with the outer diameters of 10–60 nm and
the inner diameters of 3–8 nm (see Fig. S1(a)). The content of ele-
mental carbon and graphitized carbon was >99% and ≥90% (mass%),
respectively, in the purified CNTs, with the N2-BET specific surface
area (SSA) being ca. 125 m2 g−1 before crushing and sieving. Test
of H2-temperature-programmed hydrogenation (TPH) showed that
the temperature needed for initiating the hydrogenation reaction
of the CNTs with H2 was ≥773 K (see Fig. S1(b)), indicating that this
type of CNTs was stable in H2-atmosphere at the reaction temper-
atures for the hydrogenation-conversion of syngas to ethanol and
higher alcohols.
Fig. 5(a) shows the TEM image of the pre-reduced/sulfurized
Ni0.5Mo1K0.5–15%CNTs catalyst. The observed the Ni–Mo sulfide
nanoparticles were fairly uniform in size and well dispersed on the
CNTs or their scraps. Fig. 5 (b) and (c) show the TEM images of the
two catalysts, Ni0.5Mo1K0.5–15%CNTs and the CNT-free counter-
part, Ni0.5Mo1K0.5, after testing for conversion of syngas to ethanol
and higher alcohols under the aforementioned reaction conditions
of 8.0 MPa and 593 K for 24 h. Comparatively, the particle sizes of
Ni–Mo–K sulfide species were significantly smaller in the CNT-
containing catalyst than those in CNT-free counterpart (Fig. 5 (b)
vs. (c)).
3.3. Performance of Ni–Mo–K sulfide catalyst doped with CNTs
and reference system
The reaction of syngas converted to ethanol and higher
alcohols was conducted under the aforementioned optimized reac-
tion conditions: 8.0 MPa, 593 K, V(H2)/V(CO)/V(N2) = 45/45/10 and
GHSV = 4000 mL h−1 g−1 (outlet). Table 1 lists the results taken at
100 h after the reaction started. Over the Ni0.5Mo1K0.5–15%CNTs
catalyst, X(CO-hydr.) reached 8.4%, with the corresponding
STY(total oxy.) and STY(EtOH) being 113 and 55.6 mg h−1 g−1
,
respectively. The two values was 1.51 and 1.47 times those (74.7
and 37.9 mg h−1 g−1) of the CNT-free counterpart Ni0.5Mo1K0.5
respectively.
,
Fig. 4 shows the operation stability of the Ni0.5Mo1K0.5-15%CNTs
catalyst and the CNT-free counterpart Ni0.5Mo1K0.5 for the syngas
conversion to ethanol and higher alcohols lasting 112 h. Over the
Ni0.5Mo1K0.5–15%CNTs catalysts, throughout the reaction, X(CO-
hydr.), S(total oxy.), S(EtOH) and STY(total oxy.) maintained stable
at the level of ∼8.4%, ∼64%, ∼33% and 113 mg h−1 g−1, respectively,
with no obvious deactivation observed.
3.4.2. XRD characterization of the catalysts
The apparent activation energy (Ea) of the hydrogenation-
reaction of CO over the two catalysts tested was measured under
the reaction conditions of 5.0 MPa, V(H2)/V(CO)/V(N2) = 45/45/10,
GHSV = 6000 mL h−1 g−1 (outlet), under which the mass transfer
limitation was ruled out. The Ea of hydrogenation-reaction of
CO observed on the two catalysts, Ni0.5Mo1K0.5–15%CNTs and
Ni0.5Mo1K0.5, was 58.6 and 62.1 kJ mol−1, respectively (see Fig. S5).
These Ea values were fairly close to each other, indicating that
appropriate incorporation of a minor amount of CNTs into the
Ni0.5Mo1K0.5 sulfide host catalyst did not cause a marked change in
The XRD post-reaction analysis (see Fig. 6) showed that
there were few differences between the two used catalysts
in their XRD features related with the Ni–Mo–K components,
except the feature at 2Â = 26.1◦ for the CNT-containing sys-
tem (see Fig. 6(a)), which was due to the diffraction of (0 0 2)
plane of graphite-like tube-wall of the CNTs [21]. In these
used catalysts, the Ni–Mo–K components existed mainly in
the forms of MoS2, NiSx and K–Mo–S with the corresponding
XRD features appearing at 2Â = 13.6◦/31.0◦, 35.8◦/45.8◦/56.2◦ and
19.0◦/22.6◦/27.2◦/32.3◦/41.5◦, respectively [11,16,23,24]. Using the

48
J.-J. Wang et al. / Applied Catalysis A: General 468 (2013) 44–51
Table 1
Reactivity of syngas conversion to ethanol and higher alcohols over the Ni0.5Mo1K0.5–15%CNTs catalyst and the reference system.a
Catalysts
BETb (m2 g−1
)
React. condition X (CO) (%)
Selectivity of hydrogenation products (C%, CO2-free)
STY (mg h−1 g−1
)
Total Hydr. WGS Total oxy. MeOH EtOH PrOH BuOH DME HCs
Total oxy. EtOH
14.2
4.1
—
(i)
(ii)
13.4
13.3
8.4
8.4
5.0
4.9
67.3
64.1
14.5
13.8
36.1
33.1
11.1
11.7
3.2
3.8
2.4
1.7
32.7
35.9
90.3
113
45.4
55.6
Ni0.5Mo1K0.5–15%CNTs
Ni0.5Mo1K0.5
(i)
(ii)
9.1
9.9
6.2
6.4
2.9
3.5
62.6
57.7
13.9
10.2
33.8
30.1
9.4
12.0
2.7
3.9
2.8
1.5
37.4
42.3
60.9
74.7
31.0
37.9
Ni0.5Mo1K0.5–15%AC
(i)
10.6
7.1
3.5
61.6
14.8
33.1
9.0
2.4
2.3
38.3
69.2
35.0
a
Reaction conditions: (i) 5.0 MPa, 593 K, V(H2)/V(CO)/V(N2) = 45/45/10, GHSV = 3000 mL h−1 g−1 (outlet); (ii) 8.0 MPa, 593 K, V(H2)/V(CO)/V(N2) = 45/45/10,
GHSV = 4000 mL h−1 g−1 (outlet); the data were taken 100 h after the reaction started.
b
In the pre-reduced/sulfurized state.
Scherrer’s equation, the average crystallite-size of the three kinds of
crystallites, MoS2, NiS and K–Mo–S, was estimated to be 19.0 nm,
8.4 nm and 7.8 nm for the CNT-containing catalyst, and 32.0 nm,
10.5 nm and 12.3 nm for the CNT-free counterpart. This is in line
with the results of the TEM observation (see Fig. 5(b) and (c)). The
average crystallite-size of the catalytically active components of
the CNT-containing catalyst was comparatively smaller than that
of the CNT-free counterpart, indicating the positive role of CNTs in
improving the dispersion of particles of MoS2, NiSx and K–Mo–S.
3.4.4. H2S/H2-TPS test of catalysts in oxidation state
H2S/H2-TPS of the catalyst provided useful information about
the reducibility and sulphidity of catalyst. Fig. 8 shows the H2S/H2-
TPS profiles taken on the two catalysts in oxidation-state. The
H2S/H2-reduction/sulfurization of the Ni0.5Mo1K0.5 catalyst (see
Fig. 8(b)) started at ca. 453 K and the main TPS-peaks appeared
at 586 K, 648 K and 893 K. Previous literatures [32–34] reported
that reduction of MoO3 by H2S/H2 to MoO2 occurred at ca. 650 K,
and sulfiding of MoO2 to MoS2 occurred at much higher tempera-
tures, and the combination of Ni into MoO3 might bring about the
formation of a Mo–Ni–O composite phase to be easier to be sulfur-
ized, and the sulfiding appeared to be completed at ca. 1100 K in all
cases. Therefore, the observed 586-K peak may be ascribed to sul-
fiding of MoO3 to MoO2, and the 648-K peak may be due to sulfiding
of MoO2 to MoS2, while the higher temperature region (spanning
from 693 K to 893 K) likely included the contribution from the sul-
fiding of bulk compounds (MoO3, MoO2), which were difficult to
be sulfurized at the lower temperatures. It is worth noting that the
doping of CNTs into the Ni0.5Mo1K0.5 resulted in marked down-
shifting of the sulfiding peaks (from 586-K, 648-K and 893-K to
468-K, 578-K and 713-K, respectively, see Fig. 8(b) vs. (a)). Con-
ceivably, the lower reduction/sulfurization temperature would be
beneficial to inhibiting the aggregation of NiiMojKk-sulfide crystal-
lites formed by reduction/sulfurization by H2S/H2, thus favorable
to increasing the exposed area of the NiiMojKk components in
the pre-reduced/sulfurized state (see Table 1), and subsequently,
improving the catalytic activity for higher alcohol synthesis (HAS).
3.4.3. Analysis of the used catalysts by XPS
XPS measurements provide information about Ni and Mo
species at the quasi-functioning surface of the two tested catalysts
of Ni0.5Mo1K0.5–15%CNTs and Ni0.5Mo1K0.5. The measurement
results showed that the addition of CNTs to the Ni0.5Mo1K0.5 did
not affect the valence state of the Ni2+-species (see Fig. S6). There
was little difference between the two tested catalysts in the posi-
tion and shape of their Ni(2p)-XPS peaks. The XPS peaks of Ni(2p3/2
)
and Ni(2p1/2) observed on the two tested catalysts both exhibited
at 856.3 and 873.8 eV (B.E.), respectively, with their area-intensity
ratio of I(856.3) to I(873.8) being about 2. These values are charac-
teristics of the NiSx species [25].
Different from the case of Ni(2p)-XPS spectra, certain differ-
ence existed between the two used catalysts in the position and
shape as well as relative intensity of the peaks associated with
the Mo-species. With reference to Ref. [26,27] and assuming Eb(Mo
3d3/2)–Eb(Mo 3d5/2) = 3.1 eV and I(Mo 3d5/2)/I(Mo 3d3/2) (intensity
ratio of the peaks) = 1.5 for each kind of the Mon+ species in the same
valence-state, analysis and fitting of these Mo(3d)-XPS spectra were
carried out. The results (see Fig. 7) showed that, under the reductive
atmosphere containing H2, CO, etc., most of the Mon+ was reduced
to lower valence: the major amount in Mo4+, minor in Mo5+, and
few in Mo6+. This was analogous to the case of co-existence of Mo4+
(major) with Mo5+ (minor) in the related systems [28]. The molar
percentage of Mo4+, Mo5+ and Mo6+ species in the total Mo-amount
at the surface of quasi-functioning catalyst was estimated to be
40, 45 and 15 mol% for the Ni0.5Mo1K0.5–15%CNTs, and 19, 33 and
48 mol% for the Ni0.5Mo1K0.5. These results indicate that the surface
concentration of Mo4+ species of the CNT-containing catalyst was
2.1 times that of the CNT-free counterpart.
3.4.5. H2-TPD test of pre-reduced/sulfurized catalysts
The previous literature [35] reported that graphitized carbon
black surfaces were capable of rapidly equilibrating H2/D2 mix-
ture, and that the active surface area was described in terms of
atoms located at edge positions on the graphite basal plane and
was determined from the amount of oxygen able to chemisorb at
these sites. The previous H2-TPD investigation by us [36] showed
that hydrogen adsorption on the CNTs can occur at ambient tem-
perature and pressures, and that H-species adsorbed on the CNTs
may be in two forms: associative (molecular state) and dissociative
(atomic state), as evidenced in our Raman spectroscopic study of
H2/CNTs adsorption system [21].
Fig. 9 shows the H2-TPD profiles of H2 adsorption on the
pre-reduced/sulfurized Ni0.5Mo1K0.5–15%CNTs catalyst and the
CNT-free counterpart. Overall, each profile contained a lower-
temperature peak (peak I at ca. 363 K) and a higher-temperature
peak (peak II at ca. 673 K). The lower-temperature peaks resulted
from desorption of the weakly adsorbed H-species, most proba-
bly molecularly adsorbed hydrogen, and the higher-temperature
peaks were attributed to desorption of the strongly adsorbed H-
species, perhaps dissociatively chemisorbed hydrogen [21,36]. It is
conceivable that, at the reaction temperatures for HAS (573–613 K
for the present work), the concentration of H-adspecies associated
with peak-I was very low, and most of H-adspecies at the surface of
Catalysts containing Mo, Group VIII metals, and alkali have been
extensively investigated [6–11,29,30]. It is well acknowledged that
the activity of the catalysts was mainly associated with the pres-
ence of Mo, and that lower oxidation state of Mo (mainly Mo4+
)
was more active for synthesis of higher alcohols from CO hydro-
genation [31]. The higher concentration of surface Mo4+-species,
the higher the catalyst activity is. In comparison with the CNT-free
host system, higher concentration of surface Mo4+-species of the
CNT-containing catalyst was attributed to its higher reducibility,
and was also one of the important factors leading to a significant
increase of its specific activity for CO hydrogenation.

J.-J. Wang et al. / Applied Catalysis A: General 468 (2013) 44–51
49
Fig. 6. XRD patterns of the post-operation catalysts: (a) Ni0.5Mo1K0.5–15%CNTs; (b)
Ni0.5Mo1K0.5
.
Fig. 7. XPS spectra of Mo(3d) of the post-operation catalysts: (a)
Ni0.5Mo1K0.5–15%CNTs; (b) Ni0.5Mo1K0.5
.
Fig. 5. TEM images of: (a) the pre-reduced/sulfurized Ni0.5Mo1K0.5–15%CNTs; (b)
the post-operation Ni0.5Mo1K0.5–15%CNTs; (c) the post-operation Ni0.5Mo1K0.5
.
Fig. 8. H2S/H2-TPS profiles of the catalyst precursors in oxidation state (a)
Ni0.5Mo1K0.5–15%CNTs; (b) Ni0.5Mo1K0.5 (×85%).

50
J.-J. Wang et al. / Applied Catalysis A: General 468 (2013) 44–51
the sulfurized Ni0.5Mo1K0.5 catalyst did not cause a marked change
in the Ea for the hydrogenation-conversion of syngas (see Fig. S5),
implying that the doping of CNTs did not alter the composition and
structure of catalytically active site nor the major reaction pathway
of CO hydrogenation. (2) Appropriately reducing CNT’s grain size
could improve the CNT’s capability of adsorbing/activating H2 and,
thus, increase rate of hydrogenation–conversion of CO, yet did not
influence selectivity of the products (see Figs. 2 and S7).
It follows that the CNTs as promoter function through their
adsorbing/activating H2 to generate a surface micro-environment
with higher stationary-state concentration of H-adspecies on the
functioning catalyst. This resulted in a dramatic increase, at the
surface of the functioning catalyst, of the molar percentage of cat-
alytically active Mo4+/Mo5+ species in the total amounts of surface
Mo. On the other hand, those active H-species adsorbed at surface
of the CNTs could be readily transferred to NiiMojKk active sites via
the highly conductive CNT-assisted hydrogen spillover. The afore-
mentioned two factors both were conducive to increasing the rate
of a series of the surface hydrogenation reactions in the process of
syngas hydrogenation-conversion.
Fig. 9. TPD profiles of H2 adsorption on the pre-reduced/sulfurized catalysts (a)
Ni0.5Mo1K0.5–15%CNTs; (b) Ni0.5Mo1K0.5
.
4. Concluding remarks
functioning catalysts was those corresponding to peak-II. Hence, it
could be inferred that those strongly chemisorbed H-species were
closely associated with the reaction activity of CO hydrogenation.
The relative area-intensity (I) of the peak-IIs for the two catalysts
was estimated to be: I(a)/I(b) = 100/52. This was also expected to be
the sequence of increase in the stationary-state concentration of
H-adspecies at the surface of functioning catalysts, in line with the
sequence of the HAS reaction-activity observed on the correspond-
ing catalysts.
H2-TPD measurement of hydrogen adsorbed was also done
on the CNTs with varying grain-sizes: (a) 830–380, (b) 380–180,
(c) 180–120, (d) 120–75, (e) 75–38 and (f) ≤38 ␮m. The result
showed that with the grain-size of the CNTs decreasing pro-
gressively from 830–380 ␮m, the amount of desorbed hydrogen
increased quickly in the higher grain-size region of (a)–(d), and
approached a plateau when the grain-size reached 75–38 ␮m (e)
and below (f) (see Fig. S7). This result implied that the capability
of the CNTs to adsorb hydrogen was improved when appropri-
ately reducing its grain size (correspondingly increasing its specific
surface area (SSA)). The relative area-intensity (I) of those H2-
TPD profiles in the region of 323–873 K was estimated to be:
I(a)/I(b)/I(c)/I(d)/I(e)/I(f) = 45/50/65/72/95/100. This was also expected
to be the sequence of increase in the stationary-state concentration
of H-adspecies at the functioning surface of the corresponding cata-
lysts, in line with the sequence of increase in X(CO-hydr.) observed
on the corresponding catalysts (see Fig. 2).
(1) A type of highly efficient Ni–Mo–K sulfide catalyst doped with
CNTs for conversion of syngas to ethanol and higher alcohols
was developed.
(2) Appropriately reducing CNT’s grain-size can improve the pro-
moting effect of CNTs.
(3) CNTs as promoter function through their sp2-C surface to
adsorb and activate hydrogen, so as to generate a surface
micro-environment with higher stationary-state concentration
of H-adspecies on the functioning catalyst. This resulted in a
dramatic increase, at the surface of the functioning catalyst,
of the molar percentage of the catalytically active Mo4+/Mo5+
species in the total amounts of surface Mo. On the other hand,
those active H-species adsorbed at surface of the CNTs could be
readily transferred to NiiMojKk active sites via the CNT-assisted
hydrogen spillover. The aforementioned two factors both were
conducive to increasing the rate of hydrogenation conversion
of syngas.
Acknowledgements
The work is supported by “973” project (2011CBA00508), NSFC
project (20923004), PCSIRT (No. IRT1036) of China.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.apcata.
2013.08.026.
3.5. Nature of promoter action by the CNTs
It is evident that the high reaction activity of
hydrogenation–conversion of syngas to C1–4-alcohols, espe-
cially ethanol, over the CNT-doped Ni–Mo–K sulfide catalyst is
closely related to the peculiar structure and excellent perfor-
mance for adsorbing/activating H2 of the CNTs as promoter. It has
been well known that the CNTs are a type of two-dimensional
nano-carbon, which is close to hollow graphite fiber structurally.
In the case of the present work, the CNTs as promoter function
through sp2-Cx at both nanoscale and microscale, rather than
individual carbon atoms. The CNTs did not participate in the
sulfurized NiiMojKk cluster to construct new catalytically active
site. Thus, the promoter action by CNTs via their sp2-C surface to
adsorb/activate H2 may be both short range and medium range
(via hydrogen spillover), as supported by the following results in
the present work: (1) Addition of a minor amount of the CNTs to
References
[1] V. Subramani, S.K. Gangwal, Energ. Fuels 22 (2008) 814–839.
[2] G. van der Lee, B. Schuller, H. Post, T.L.F. Favre, V. Ponec, J. Catal. 98 (1986)
522–529.
[3] Y.H. Du, D.A. Chen, K.R. Tsai, Appl. Catal. 35 (1987) 77–92.
[4] H. Ehwald, H. Ewald, D. Gutschick, M. Hermann, H. Miessner, G. Ohlmann, E.
Schierhorn, Appl. Catal. 76 (1991) 153–169.
[5] R. Burch, M.J. Hayes, J. Catal. 165 (1997) 249–261.
[6] T. Tatsumi, A. Muramatsu, T. Fukunaga, H. Tominaga, Proc. 9th Intern. Congr.
Catal. 2 (1988) 618–625.
[7] J.G. Santiesteban, C.E. Bogdan, R.G. Herman, K. Klier, Proc. 9th Intern. Congr.
Catal. 2 (1988) 561–568.
[8] H.-B. Zhang, Y.-Q. Yang, H.-P. Huang, G.-D. Lin, K.R. Tsai, Stud. Surf. Sci. Catal.
75 (1993) 1495–1505.
[9] Z. Li, Y. Fu, J. Bao, M. Jiang, T. Hu, T. Liu, Y-N. Xie, Appl. Catal., A 220 (2001)
21–30.

J.-J. Wang et al. / Applied Catalysis A: General 468 (2013) 44–51
51
[10] H. Qi, D. Li, C. Yang, Y. Ma, W. Li, Y. Sun, B. Zhong, Catal. Commun. 4 (2003)
339–342.
[11] D.-B. Li, C. Yang, H.-R. Zhang, W.-H. Li, Y.-H. Sun, B. Zhong, Stud. Surf. Sci. Catal.
147 (2004) 391–396.
[25] S. Harris, R.R. Chianelli, J. Catal. 98 (1) (1986) 17–31.
[26] J. Abart, E. Delgado, G. Ertl, H. Jeziorowshi, H. Knözinger, N. Thiele, X.Z. Wang,
E. Taglauer, Appl. Catal. 2 (1982) 155–176.
[27] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray Photo-
electron Spectroscopy—A Reference Book of Standard Spectra for Identification
and Interpretation of XPS Data, Physical Electronics Inc., Eden Prairie, 1995.
[28] A.M. Venezia, Catal. Today 77 (2003) 359–370.
[12] J. Iranmahboob, H. Toghiani, D.O. Hill, Appl. Catal. A: Gen. 247 (2003) 207–218.
[13] M. Xiang, D. Li, W. Li, B. Zhong, Y. Sun, Catal. Commun. 8 (2007) 513–518.
[14] J. Bao, Y.L. Fu, G.Z. Bian, Catal. Lett. 121 (2008) 151–157.
[15] V.R. Surisetty, A. Tavasoli, A.K. Dalai, Appl. Catal. A: Gen. 365 (2009) 243–251.
[16] V.R. Surisetty, I. Eswaramoorthi, A.K. Dalai, Fuel 96 (2012) 77–84.
[17] P. Serp, M. Corrias, P. Kalck, Appl. Catal. A: Gen. 253 (2003) 337–358.
[18] H.B. Zhang, G.D. Lin, Y.Z. Yuan, Curr. Topics Catal. 4 (2005) 1–21.
[19] H.B. Zhang, X.L. Liang, X. Dong, H.Y. Li, G.D. Lin, Catal. Surv. Asia 13 (2009) 41–58.
[20] P. Chen, H.B. Zhang, G.D. Lin, Q. Hong, K.R. Tsai, Carbon 35 (1997) 1495–1501.
[21] H.B. Zhang, G.D. Lin, Z.H. Zhou, X. Dong, T. Chen, Carbon 40 (2002)
2429–2436.
[29] K. Fujimoto, T. Oba, Appl. Catal. 13 (1985) 289–293.
[30] M. Inoue, T. Miyake, Y. Takegami, T. Inui, Appl. Catal. 29 (1987) 285–294.
[31] A. Muramatsu, T. Tatsumi, H. Tominaga, J. Phys. Chem. 96 (1992) 1334–1340.
[32] P. Arnoldy, J.A.M. van den Heijkant, G.D. de Bok, J.A. Moulijn, J. Catal. 92 (1985)
35–55.
[33] J.F. Liang, R. Guo, X.Y. Yang, J. Mol. Catal. (China) 16 (2002) 15–18.
[34] C.Y. Yang, W.Y. Yang, F.X. Ling, F. Fan, Contemporary Chem. Ind. (China) 39
(2010) 120–122.
[22] J.M. Zhou, H.Y. Li, G.D. Lin, H.B. Zhang, Acta Phys. Chim. Sin. (Chinese) 26 (2010)
3080–3086.
[23] XRD Data Bank Attached to X’Pert PRO X-ray Diffractometer (PANalytical, The
Netherlands), 2003.
[35] Y. Ishikawa, L.G. Austin, D.E. Brown, P.L. Walker Jr., in: P.L. Walker Jr., P.A.
Thrower (Eds.), Chemistry and Physics of Carbon, vol. 12, American Carbon
Society, Marcel Dekker, New York, 1975, 39-39.
[36] Z.H. Zhou, X.M. Wu, Y. Wang, G.D. Lin, H.B. Zhang, Acta Phys.-Chem. Sin. (Chi-
nese) 18 (2002) 692–698.
[24] N. Rueda, R. Bacaud, M. Vrinat, J. Catal. 169 (1997) 404–406.