Replacing platinum with tungsten carbide for decalin dehydrogenation

Article abstract of DOI:10.1007/s10562-014-1284-7

Pt was replaced with tungsten carbide (WC) in the Ni-Pt structure for decalin dehydrogenation on base of their similar properties. The catalytic performance of Ni-WC catalyst was predicted by density functional theory (DFT) calculations and further confirmed by dehydrogenation reaction. Replacing Pt with WC significantly reduces the side reactions such as ring-opening and coking. The highest conversion of decalin is about 93 % and the selectivity to naphthalene is nearly 100 % over Ni-WC/AC catalyst. The improved catalytic activity of Ni-WC/AC demonstrates the feasibility to replace Pt with WC as an active and less expensive dehydrogenation catalyst. Graphical Abstract: [Figure not available: see fulltext.].

Full text of DOI:10.1007/s10562-014-1284-7

Catal Lett
DOI 10.1007/s10562-014-1284-7
Replacing Platinum with Tungsten Carbide for Decalin
Dehydrogenation
•
Suitao Qi Jiaqi Yue YingYing Li
•
•
•
Jun Huang Chunhai Yi BoLun Yang
•
Received: 25 March 2014 / Accepted: 20 May 2014
Ó Springer Science+Business Media New York 2014
Abstract Pt was replaced with tungsten carbide (WC) in
the Ni–Pt structure for decalin dehydrogenation on base of
their similar properties. The catalytic performance of Ni–
WC catalyst was predicted by density functional theory
(DFT) calculations and further confirmed by dehydroge-
nation reaction. Replacing Pt with WC significantly redu-
ces the side reactions such as ring-opening and coking. The
highest conversion of decalin is about 93 % and the
selectivity to naphthalene is nearly 100 % over Ni–WC/AC
catalyst. The improved catalytic activity of Ni–WC/AC
demonstrates the feasibility to replace Pt with WC as an
active and less expensive dehydrogenation catalyst.
cleavage of C–H bond. It is well known that high temperature
is in favor of dehydrogenation of liquid organic hydride.
However, the side reaction, such as cracking, ring open, and
carbon deposition, would appear at high temperature, which
will result in the interruption of reversible hydrogen storage
cycle and catalyst coking. How to design a rational catalyst to
improve the selectivity to unsaturated aromatic and to realize
low-temperature dehydrogenation is the key for reversible
hydrogen storage by liquid organic hydride. Although Pt [11,
12], Rh [13] has been known as a good catalyst for dehy-
drogenation, its high price restricted its application on a large
scale. Levy and Boudart [14] first reported that the transition
metal carbide WC is similar to Pt in the electronic structure
and catalytic properties. In addition, WC has a strong acid
resistance, good electrical conductivity, CO and hydrogen
sulfide poisoning tolerance. These properties reveal that it is
feasible to use WC replacing precious metals to reduce the
catalyst cost for dehydrogenation reaction.
Keywords Decalin Á Dehydrogenation Á Ni–WC Á
Density functional theory
1 Introduction
In this work, Ni–WC supported on active carbon are
prepared by incipient wetness impregnation and tested by a
probe reaction, decalin dehydrogenation in a flow reactor.
The dehydrogenation performances of different catalyst
surfaces are predicted by density functional theory (DFT)
calculation. The effects of different process conditions such
as temperature and feed rate on the catalytic activity of
decalin dehydrogenation in the reaction have been discussed.
Hydrogen storage and transportation is one of the bottle-
necks for hydrogen consumption. In the numerous hydrogen
storage methods, hydrogen storage in liquid organic hydrides
is considered as a safe, efficient method which can be used to
achieve hydrogen storage and transportation on a large scale
and at a distance [1]. Using a reversible hydrogenation–
dehydrogenation catalytic reaction pair between unsaturated
aromatic hydrocarbons such as benzene, toluene and naph-
thalene and its corresponding liquid organic hydrides such as
cyclohexane [2–4], methylcyclohexane [5–9] and decalin
[10], hydrogen is stored or released through formation and
2 DFT Calculations and Experimental
2.1 DFT Calculations
S. Qi (&) Á J. Yue Á Y. Li Á J. Huang Á C. Yi Á B. Yang
Xi’an Jiaotong University, Xi’an 710049, Shaanxi,
People’s Republic of China
The catalytic performance of metal catalysts greatly
depends on the electronic structure of a metal active sur-
face, which affects the binding energy between reactant
e-mail: suitaoqi@mail.xjtu.edu.cn; suitaoqi@gmail.com
123

S. Qi et al.
and the catalyst surface. Therefore, the active components
promoting C–H bond were selected by DFT simulations
using the Vienna Ab-initio Simulation Package (VASP)
version 5.2 [15–18]. The spin-polarized Perdew–Wang
(PW91) generalized gradient approximation (GGA) [19,
20] functional was employed for the exchange–correlation
energy. A cutoff energy of 400 eV was employed for the
plane-wave basis set. As shown in Fig. 1, the hydrogen
binding energies remain stable when the k points was more
than 3 9 3 9 1, which indicated that 3 9 3 9 1 k points
for sampling the Brillouin zone was enough. All slab
calculations were performed using a 3 9 3 unit cell with a
3 9 3 9 1 Monkhorst-Pack k-point grid. Our previous
work on hydrogen adsorption on Pt and Pt–Ni surfaces [21]
found that the hydrogen binding energy (HBE) on Pt (111)
surface was the lowest among Pt (100), (110) and (111)
surfaces, thus Pt (111) surface was selected again in this
work. HBE were calculated on five kinds of surfaces, such
as Pt (111), Ni (111), the tungsten-terminated WC (0001),
nickel-terminated Ni–WC (0001), and Ni–Pt–Pt (111)
surface. The surfaces are modeled with a three-layer slab
˚
with a vacuum thickness of 10 A [21]. Each layer of WC
(0001) was composed of alternating tungsten and carbon.
The Ni–WC (0001) and Ni–Pt–Pt (111) model was
obtained by substituting the top layer with Ni atoms. The
top two layers were allowed to relax, and the bottom layer
was kept frozen.
20
15
10
5
The catalytic dehydrogenation activity could be reflec-
ted from the hydrogen adsorption on different metal sur-
faces. Figure 2 shows the four hydrogen adsorption sites
such as top site, bridge site, fcc site and hcp site on Pt (111)
surface. The other metal surfaces are similar to the Pt (111)
surface.
0
2.2 Experimental
Granular powders of a commercial activated carbon were
used as the catalyst support. (NH4)6W7O24Á6H2O aqueous
solution was added to the active carbon by incipient wet-
ness impregnation. And then the mixtures were dried at
120 °C, followed by the temperature-programming
-5
1×1×1 2×2×1 3×3×1 4×4×1 5×5×1 6×6×1 7×7×1 8×8×1 9×9×1
K-points
Fig. 1 Hydrogen adsorption energies under the condition of different
K-point densities
Fig. 2 Different adsorption sites on Pt(111) surface: a top, b bridge, c fcc, d hcp. White hydrogen, Blue platinum
Table 1 HBE (eV) of four sites
on five different surface
Surface
Top
Bridge
PW91
Hcp
Fcc
PW91
PBE
PBE
PW91
PBE
PW91
PBE
Pt (111)
-2.976
-3.029
-2.936
-3.685
-3.601
-2.87
-2.28
-2.63
-2.31
-1.87
-3.326
-3.422
-3.985
-3.801
-3.913
-2.83
-2.71
-3.07
-2.74
-2.80
23.547
23.592
23.919
23.952
23.838
22.89
22.79
22.99
22.81
22.91
23.788
23.863
24.269
24.023
24.013
22.90
22.82
23.31
22.95
23.09
Ni (111)
WC (0001)
Ni–Pt–Pt (111)
Ni–WC (0001)
123

Replacing Platinum with Tungsten Carbide
Fig. 3 Local density of states
of hydrogen on different surface
2.5
2.0
1.5
1.0
0.5
0.4
0.2
(a)
(b)
H
Ni-WC
Ni-Pt
(d)
(e)
0.0
0.4
0.0
0.4
Pt
0.2
0.2
0.0
0.4
0.0
0.4
WC
Ni
(c)
(f)
0.2
0.0
0.2
0.0
-10
-5
0
5
10
-10
-5
0
5
10
Energy/eV
Energy/eV
reduction from 293 to 943 K at the ramping rate of 3.25 K/
min in a stream of 60 mL/min N2 for 1 h, and carbonized
from 943 to 1,243 K at the ramping rate of 1.5 K/min in a
stream of 80 mL/min H2 and 20 mL/min CH4 for 2 h. And
then, WC/AC was obtained. Ni/AC and Ni–WC/AC were
obtained by the same method as WC/AC. There are some
differences in the TPR process. The dried mixture was
temperature reduced from 293 to 723 K at the ramping
rate of 9 K/min in a stream of 20 mL/min N2 and 40 mL/
min H2 for 2 h.
3 Results and Discussion
3.1 DFT Results
Hydrogen bonding energy (HBE) of four adsorption sites
on five metals surfaces are shown in Table 1. Compared
with other adsorption sites, HBE at fcc site on the five
metal surfaces was the largest, revealing that the corre-
sponding adsorption state was the most stable. HBE trend
was WC [ Ni–Pt–Pt & Ni–WC [ Ni [ Pt, suggesting
Ni–WC would have the similar catalytic dehydrogenation
activity with Ni–Pt–Pt surface.
The catalysts structure were investigated by an X-ray
diffractometer (XRD) (Holland PANalytica X’Pert ProX)
using a Cu Ka radiation. The catalyst morphology were
characterized using field emission scanning electron
microscopy (SEM) coupled with energy dispersive spec-
trometer (EDS) (sirion200, U.S.) and transmission electron
microscopy (TEM) images with selected area electron
diffraction (JEM-2100, Japan).
Hydrogen binds stronger to W-terminated WC (0001)
surface than to Pt(111) surface for 0.5 eV and the prefer-
ential sites on W-terminated WC(0001) and Pt(111) sur-
faces are the fcc site. All of these results are agree with the
DFT-PBE and the DFT-PW91 calculation results by Vasic
et al. [22] and Gaston et al. [23], respectively. In addition,
we also calculated the HBE using three layers and four
layers for our slab models by PBE with the top two layers
relaxed. The results showed that the difference in HBE was
less than 0.02 eV by PBE method between these two
models, indicating that the slab model with three layers
was enough for HBE calculation. Although there was about
1.0 eV difference in HBE between PW91 and PBE, the
trend of HBE was similar.
The catalytic activity of Ni/AC, WC/AC and Ni–WC/
AC were evaluated in a flow reactor. The catalytic dehy-
drogenation of decalin was performed with a certain
amount of Ni–WC/AC catalyst at 673 K in a tube furnace
(with a core diameter of 6 mm and a length of 60 cm, 0.5 g
Ni–WC/AC cat.). Decalin was supplied continuously to the
catalyst bed at a constant feed rate (0.01 mL/min). All of
the gas mixture was removed from the reactor through a
thermally protected part connecting to a condenser cooled
at 298 K. The accumulated condensates were analyzed gas-
chromatographically every 15 min.
Local density of states of Hydrogen with or without
adsorption on different surface are shown in Fig. 3. Fig-
ure 3a shows the density of state (DOS) of hydrogen
123

S. Qi et al.
Fig. 4 a XRD patterns of AC,
20 wt%WC/AC, 30 wt%Ni/
AC, 30 wt%Ni–20 wt%WC/
AC; b, c SEM images of
30 wt%Ni–20 wt%WC/AC at
different magnifications;
d TEM image of 30 wt%Ni–
20 wt%WC/AC; e HRTEM
image and SAED pattern
(inset) of 30 wt%Ni–
20 wt%WC/AC. g SEM image
of 35 wt%Ni/AC; h TEM
image of 35 wt%Ni/AC
without adsorption on metal surface, from which it can be
seen that there is a strong peak at the Fermi level. The DOS
of hydrogen exhibits delocalizability after adsorbed on
these catalysts surface as shown in Fig. 3b–f, demonstrat-
ing the bond formation between catalysts surface and
Hydrogen. The degree of delocalizability can imply the
123

Replacing Platinum with Tungsten Carbide
100
90
3.2 Catalysts Characterization
b
The XRD patterns of AC, Ni/AC, WC/AC and Ni–WC/AC
were shown in Fig. 4a. Compared with AC patterns, three
strong diffraction peaks of Ni could be seen in Ni/AC
patterns, corresponding to Ni (111), (200), (220). Similarly,
seven WC diffraction peaks could be seen, corresponding
to WC (001), (100), (101), (002), (111), (200), (102). A
few W2C diffraction peaks could been seen in WC/AC
diffraction patterns, corresponding to W2C (100), (002),
(101), (102), (110), which indicated that the formed car-
bides were a mixture of WC and W2C. The intensity of WC
diffraction peaks were reduced in Ni–WC/AC patterns,
which may be due to WC surface partly covered by Ni.
The smooth small Ni nanoparticles about 20 nm could
be clearly observed in Fig. 4b, c. A coral-like structure was
formed by Ni–WC/AC nanoparticles connected end to end
because of the magnetic property of Ni. The corresponding
energy dispersive spectroscopy (EDS) results were basi-
cally consistent with that of the bulk phase, revealing that
elements dispersed evenly in the prepared catalysts.
80
a
70
Conversion
Selectivity of naphthalene
60
50
40
30
20
10
0
0.8
0.7
0.6
0.5
0.4
0.3
0.2
b
0
5
10 15 20 25 30 35
WC content/wt%
a
5
10
15
20
25
30
35
Ni or WC content/wt%
Fig. 5 Correlation between metal content and decalin dehydrogena-
tion activity over different catalysts. a WC/AC; b Ni/AC the inset is
the correlation between WC content and conversion of decalin over
WC/AC catalyst
Seen from TEM images, the small grains about 10 nm
were evenly dispersed in Fig. 4e. There were two different
sizes of interplanar spacing which were corresponding to
Ni (111) (0.196 nm) and WC (100) (0.252 nm) respec-
tively in Fig. 4f. The junction of the two lattice fringes
revealed that Ni was on the top of WC in Fig. 4f, which
consisted with the Ni–WC structure in DFT caculations.
SAED patterns exhibited the well-ordered rings, which
indicated that the Ni–WC/AC was polycrystalline. The
congregated catalyst surface with the higher Ni content
could be seen in Fig. 4g, h which may result in the decrease
of the dehydrogenation conversion and the selectivity of
naphthalene over 35 wt% Ni/AC (Fig. 5).
intensity of H–M bond and it was found that the bonds
between Hydrogen and Ni–WC, Ni–Pt, WC surface are
stronger than H–Pt.
As shown in Fig. 3, the degree of delocalizability on
WC (0001) surface was too strong while doping Ni maked
the delocalizability of Ni–WC (0001) surface weaker, and
the DOS of Ni–WC (0001) surface was similar with that of
Ni–Pt–Pt (111) surface. It could be deduced that the Ni–
WC (0001) surface probably had the good dehydrogenation
activity similar to Ni–Pt–Pt (111) surface, which was in
good agreement with the HBE results. The correlation
between HBE and the catalytic activity abided by the
‘‘Balanlin’’ volcano curve. If hydrogen binds too weakly, it
will desorb molecularly before dehydrogenation can take
place. On the other hand, if hydrogen binds too strongly, it
will lead to the complete decomposition of hydrocarbons or
hamper the evolution of hydrogen. Thus, the dehydroge-
nation activity appears to be maximized when there is an
intermediate intensity of hydrogen bond. The HBE on WC
surface was much too strong to show high conversion of
dehydrogenation, which may hamper the hydrogen evolu-
tion. Doping WC with Ni decreased HBE, which was close
to the top of volcano and almost equal to that of Pt-Ni. It is
important to point out that the surfaces used in the exper-
iment contained multiple facets and step edges. The main
purpose of DFT results on the closed-packed facet such as
Pt (111) or WC (0001) is to provide a general trend of
catalytic activity to the experimental work for selecting
active metals.
3.3 The Catalytic Activities of Ni/AC, WC/AC,
Ni–WC/AC for Decalin Dehydrogenation
The effect of Ni content on decalin conversion and selec-
tivity of naphthalene is shown in Fig. 5. The very low
dehydrogenation activity on WC catalyst (curve a) agreed
with DFT results. The HBE of WC and delocalizability of
DOS is much too strong to show high conversion, which
implies the strong intensity of H–WC bond. The strong H–
WC bond hampers the evolution of Hydrogen. It would be
likely that there was a high activation barrier for decalin
dehydrogenation over WC/AC catalyst.
Interestingly, the conversion of decalin and selectivity to
naphthalene increases with the content of Ni (curve b). How-
ever, the conversion of decalin and selectivity to naphthalene
has a downward trend when Ni content is more than 30 wt%. It
indicates that the amount of the active sites in catalyst increases
with the Ni content when the Ni content is low. However, the
123

S. Qi et al.
100
90
80
70
60
50
40
30
20
100
90
80
70
60
50
40
30
20
100
90
80
70
60
50
40
30
20
Conversion
Selectivity of naphthalene
(a)
(b)
(d)
(f)
(c)
(e)
40
60
80
100
120
140
160
40
60
80
100
120
140
160
Reaction time/min
Reaction time/min
Fig. 6 The decalin dehydrogenation over different Ni–WC/AC
catalysts at 673 K. a 10 wt%Ni–10 wt%WC/AC; b 10 wt%Ni–
wt%WC/AC; e 30 wt%Ni–10 wt%WC/AC; f 30 wt%Ni–20 wt%WC/
AC reaction conditions: catalyst 0.500 g, feed rate 0.01 ml/min, 101
kpa, 673 K
20 wt%WC/AC;
c
20 wt%Ni–10 wt%WC/AC; d 20 wt%Ni–20
100
superfluous Ni leads to the low activity due to the poor dis-
persion of Ni when the Ni content has exceeds a certain extent.
These findings suggest that the Ni/AC catalysts with highly
dispersed Ni ensembles containing appropriate Ni content
should exhibit high performance for decalin dehydrogenation.
With the increasement of the WC or Ni content of Ni–
WC/AC catalysts, the catalytic conversion of decalin
gradually improved, and the selectivity to naphthalene is
still high, as shown in Fig. 6. Both the conversion and
selectivity to naphthalene over 30 wt%Ni–20 wt%WC/AC
catalyst are highest among the selected catalysts. Com-
pared with Ni/AC and WC/AC catalysts, the activity of Ni–
WC/AC greatly increased, which may benefit from the
synergy between the Ni and WC.
90
80
70
60
Conversion
Selectivity of naphthalin
0
2
4
6
8
10 12 14 16 18 20 22 24
Reaction time/h
As discussed previously on Ni–Pt supported active
carbon bimetallic catalyst [24], the side-products, tertian,
appeared at the beginning of the reaction, and then tetralin
Fig. 7 Stablity of decalin dehydrogenation over 30 wt%Ni–
20 wt%WC/AC catalyst. Reaction conditions: catalyst 0.500 g, feed
rate 0.01 ml/min,101 kPa, 673 K, LHSV of 0.8 h-1
123

Replacing Platinum with Tungsten Carbide
2. Escobar J, De Los Reyes JA, Viveros T, Barrera MC (2006) Ind.
Eng. Chem. Res. 45:5693
3. Biniwale RB, Kariya N, Yamashiro H, Ichikawa M (2006) J Phys
Chem B 110:3189
further converted to naphthalene. In addition, the corre-
sponding dehydrogenation conversion of tertalin and the
selectivity of naphthalene was 95 and 99 %, respectively.
4. Akamatsu K, Ohta Y, Sugawara T, Hattori T, Nakao SI (2008)
Ind Eng Chem Res 47:9842
3.4 The Stablity of Ni–WC/AC Catalyst
5. Alhumaidan F, Cresswell D, Garforth A (2011) Energy Fuels
25:4217
6. Hodoshima S, Shono A, Saito Y (2008) Energy Fuels 22:2559
7. Ali JK, Rippint DWT (1996) Ind Eng Chem Res 34:722
8. Alhumaidan F, Cresswell D, Garforth A (2011) Ind Eng Chem
Res 50:2509
9. Oda K, Akamatsu K, Sugawara T, Kikuchi R, Segawa A, Nakao
SI (2010) Ind Eng Chem Res 49:11287
10. Wang Y, Shah N, Huggins FE, Huffman GP (2006) Energy Fuels
20:2612
The 30 wt%Ni–20 wt%WC/AC catalyst stability was
evaluated by a continuous dehydrogen test at 400 °C, with
liquid hourly space velocity (LHSV) of 0.8 h-1. The result
was shown in Fig. 7. After 22 h, the conversion and
selectivity rate were maintained at about 93 and 99.7 %.
The result suggested that the catalyst was stable, which can
be explained by the stability of Ni and WC.
11. Alhumaidan F, Cresswell D, Garforth A (2010) Ind Eng Chem
Res 49:9764
12. Usman M, Cresswell D, Garforth A (2012) Ind Eng Chem Res
51:158
4 Conclusion
13. Mendes FMT, Schmal M (1997) Appl Catal A 151:393
14. Levy R, Boudart M (1973) J Sci 181:547
15. Kresse G, Hafner J (1993) Phys Rev B 47:558
16. Kresse G, Hafner J (1994) Phys Rev B 49:14251
17. Kresse G, Furthmu¨ller J (1996) Comput Mat Sci 6:15
18. Kresse G, Furthmu¨ller J (1996) Phys Rev B 54:11169
19. Perdew JP, Chevary JA, Vosko SH, Jackson KA, Pederson MR,
Singh DJ, Fiolhais C (1992) Phys Rev B 46:6671
20. Perdew JP, Chevary JA, Vosko SH, Jackson KA, Pederson MR,
Singh DJ, Fiolhais C (1993) Phys Rev B 48:4978
21. Gao ZF, Chen H, Qi ST, Yi CH, Yang BL (2013) Acta Phys-
Chim Sin 29:1990
(1) HBE of DFT calculation on five metal surfaces (Ni, Pt,
WC, Ni–Pt–Pt, Ni–WC) is: WC [ Ni–Pt–Pt & Ni–
WC [ Ni [ Pt, suggesting Ni–WC would have the similar
catalytic dehydrogenation activity with Ni–Pt–Pt surface.
(2) Decalin dehydrogenation was selected as the probe
reaction for the activities of Ni–WC/AC catalysts. The
results suggest that the dehydrogenation performance of
Ni–WC/AC catalyst is better than that of Ni/AC and WC/
AC catalysts. The decalin conversion and selectivity to
naphthaline on 30 wt%Ni–20 wt%WC/AC are the highest
among all the Ni–WC/AC catalysts. In addition, it shows
the good stability for decalin dehydrogenation. Although
there are more details need to do in catalyst synthesis, it is
feasible to replace Pt with WC as the active and less
expensive catalysts for dehydrogenation reaction.
´
ˇ
22. Vasic DD, Pasti IA, Mentus SV (2013) DFT study of platinum
and palladium overlayers on tungsten carbide: Structure and
electrocatalytic activity toward hydrogen oxidation/evolution
reaction. Int J Hydrog Energy 38:5009–5018
23. Gaston N, Hendy S (2009) Hydrogen adsorption on model
tungsten carbide surfaces. Catal Today 146:223–229
24. L LingHu, W Wu, S Qi, C Yi, G Zhu, B Yang (2011) Decalin
dehydrogenation on active carbon supported Pt–Ni bimetallic
Catalysts. 2011 2nd International Conference on Mechanic
Automation and Control Engineering, MACE 2011 –Proceedings,
July 15–17, 6403–6, Inner Mongolia
Acknowledgments This work was supported by the National Nat-
ural Science Foundation of China (21006076), Specialized Research
Fund for the Doctoral Program of Higher Education of China
(20110201130002).
References
1. Shukla A, Karmakar S, Biniwale RB (2012) Int J Hydrog Energy
37:3719
123