Highly efficient metal sulfide catalysts for selective dehydrogenation of isobutane to isobutene

Article abstract of DOI:10.1021/cs5000944

Metal sulfide catalysts were highly efficient in the activation of C-H bond for isobutane dehydrogenation, and the dehydrogenation performance was better than that of the commercial catalysts Cr₂O₃/Al ₂O₃ and Pt-Sn/Al₂O₃, providing a class of environmentally friendly and economical alternative catalysts for industrial application.

Full text of DOI:10.1021/cs5000944

Letter
pubs.acs.org/acscatalysis
Highly Efficient Metal Sulfide Catalysts for Selective
Dehydrogenation of Isobutane to Isobutene
Guowei Wang, Chunyi Li,* and Honghong Shan
State Key Laboratory of Heavy Oil Processing, China University of Petroleum, Qingdao 266580, People’s R China
S Supporting Information
*
ABSTRACT: Metal sulfide catalysts were highly efficient in the activation of C−H
bond for isobutane dehydrogenation, and the dehydrogenation performance was
better than that of the commercial catalysts Cr O /Al O and Pt−Sn/Al O ,
2
3
2
3
2
3
providing a class of environmentally friendly and economical alternative catalysts
for industrial application.
KEYWORDS: isobutane, dehydrogenation, isobutene, metal sulfide, sulfidation
2
0
21
ith increasing demand for isobutene, a widely utilized
acceptor as early as 1966. In 1994, Resasco et al. treated a
Ni/Al O catalyst with dimethyl sulfoxide and observed that an
1
W
raw material for the synthesis of polymers and octane-
2
3
2
boosting additives in gasoline, isobutane dehydrogenation has
increased selectivity toward isobutene and a decreased coke
formation rate, as well as an isobutene yield of 21%, could be
achieved. Although the secondary reactions were effectively
inhibited after the modification, the catalyst activity still needed
to be improved. Moreover, Kobayashi and Shimizu also
reported similar promoting effect of sulfur on isobutane
dehydrogenation over a Pt-alkali metal/Al O catalyst.
3
−5
drawn extensive attention in recent years. CrO
and Pt-
x
6
−8
based catalysts are known to be two typical catalysts applied
in the catalytic dehydrogenation of isobutane on a commercial
scale for decades. Although the catalytic performance of these
catalysts is relatively satisfactory, the harmful impacts of Cr and
the high cost of Pt have limited their application to some
22
2
3
9
−12
23
extent. Moreover, even with extensive research
focused on
Supported metal sulfide catalysts, for example, MoS ,
2
24
25
alternative routes, that is, oxidative dehydrogenation, it is still
hard to control the extent of oxidation, which inevitably leads
CoS, and NiS, have been widely applied in hydrogenation
processes for years. Hydrogenation and dehydrogenation are
reverse reactions; therefore, the questions to be answered are
why sulfide catalysts have not been applied in dehydrogenation
processes, and whether metal sulfides are able to catalyze
dehydrogenation reactions. Recently, we have found that
MgAl O -supported NiS catalysts exhibited excellent perform-
to an increase in the formation of CO and decrease in the
x
1
3,14
selectivity toward olefins.
Therefore, the major challenge in
dehydrogenation technology is to develop a new type of
catalyst with environmentally friendly and cost-effective
characteristics as well as excellent performance in the absence
1
5
2
4
of oxidants.
26
ance in isobutane dehydrogenation
and applied for
To achieve this goal, many attempts have been made.
However, when Co and Ni were evaluated as active
components, a great amount of methane was generated.
27,28
patents.
Then, how about other metal sulfides?
In this paper, to investigate and compare the performance of
various metal sulfides, inert silica was chosen as the support. A
variety of metal oxides, including ZnO, CuO, MnO , MoO ,
Fe O , Co O , and NiO (some even with inferior hydro-
genation activity, e.g., ZnO, CuO and MnO ), all environ-
mentally friendly and inexpensive, with 13 wt % loading
supported on silica were prepared and pretreated in H S/H
flow. Afterward, all these sulfided catalysts were evaluated and
16−18
According to the literature,
the dissociation energy of
2
3
C−C bonds is smaller than that of C−H bonds; therefore, the
2
3
3
4
active species preferentially insert into C−C bonds, and
19
2
cracking reactions occur as a result. Georgiadis et al. also
held the point that, for both Co and Ni, demethanation is
preferred over dehydrogenation on energetic grounds. Con-
sequently, certain methods should be carried out to suppress
the undesired C−C bond cleavage and selectively activate the
C−H bond of the alkane molecules.
2
2
Received: November 18, 2013
Revised: March 1, 2014
Published: March 5, 2014
Molybdenum sulfide-alumina catalyst was found to be
efficient in paraffin dehydrogenation with CO as a hydrogen
©
2014 American Chemical Society
1139
dx.doi.org/10.1021/cs5000944 | ACS Catal. 2014, 4, 1139−1143

ACS Catalysis
Letter
Figure 1. XRD patterns of sulfided SiO -supported oxide catalysts before (a) and after (b) reaction: (A) ZnO/SiO , (B) CuO/SiO , (C) MnO /
2
2
2
2
SiO , (D) MoO /SiO , (E) Fe O /SiO , (F) Co O /SiO , and (G) NiO/SiO .
2
3
2
2
3
2
3
4
2
2
Table 1. Physicochemical Properties and Catalytic Performance of SiO -Supported Metal Oxide and Corresponding Metal
2
a
Sulfide Catalysts for Isobutane Dehydrogenation
oxide catalysts
sulfide catalysts
2
3
b
c
2
3
b
c
SBET, m /g
V , cm /g
Cisobutane , wt %
Sisobutene , wt %
SBET, m /g
V , cm /g
Cisobutane , wt %
Sisobutene , wt %
p
p
ZnO/SiO2
CuO/SiO2
252
248
245
234
239
225
236
1.01
0.97
0.91
0.87
0.88
0.86
0.88
4.2
3.7
66.9
68.6
56.6
66.4
43.5
25.3
7.6
249
244
240
229
236
211
234
0.98
0.93
0.87
0.81
0.84
0.71
0.74
28.4
64.9
62.8
65.2
69.2
71.1
67.0
80.6
84.7
84.5
79.8
86.6
87.0
87.6
MnO /SiO2
19.1
6.4
2
MoO /SiO2
3
Fe O /SiO
2
13.3
14.1
91.1
2
3
Co O /SiO
2
3
4
NiO/SiO2
a
−1
b
Reaction conditions: temp, 560 °C; 4 g of catalyst loaded; 14.3 vol % i-C H in nitrogen at a total flow rate of 14 mL min . Conversion of
4
10
c
isobutane. Selectivity to isobutene, and all the data were obtained at the very beginning of the reaction.
characterized. Moreover, a detailed catalyst preparation
procedure is provided in the Supporting Information.
Isobutane conversion and isobutene selectivity of the oxide
catalysts behaved in opposite ways, as reflected by the results
XRD patterns of these sulfided SiO -supported oxide
that NiO/SiO , with the highest isobutane conversion,
2
2
catalysts are shown in Figure 1a. Diffraction peaks marked
with certain symbols represent corresponding metal sulfides. As
presented the lowest selectivity to isobutene, whereas CuO/
SiO2 with the lowest conversion, exhibited the highest
selectivity. As for the Co and Ni-based catalysts, the generation
of a large amount of methane was indicative of high activity for
for the sulfided ZnO/SiO catalyst, ZnS was detected on the
2
catalyst surface in the form of both wurtzite and sphalerite.
Moreover, two forms of Cu S existing on the surface of the
2
C−C bond breaking. Surprisingly, after sulfidation with H S/
2
sulfided CuO/SiO catalyst are chalcocite and digenite. In brief,
2
H , all the catalysts but NiO/SiO2 exhibited significantly
2
all the oxides on the investigated catalysts were effectively
converted to the corresponding sulfides after sulfidation
treatment. From A to G, the corresponding sulfides were
ZnS, Cu S, MnS, MoS , FeS, CoS, and NiS.
improved isobutane conversion. Meanwhile, the selectivity
toward isobutene was basically higher than 80 wt % for all
catalysts, demonstrating the highly efficient activation ability of
C−H bond for the sulfided catalysts. Most importantly, in
addition to Mo, Ni, and Co, widely applied in hydrogenation
processes, the other metals rarely used in hydrogenation
reaction also exhibited improved dehydrogenation performance
2
2
Table 1 compares the physicochemical properties and
dehydrogenation performance of SiO -supported metal oxide
2
and corresponding sulfide catalysts. The surface area of all the
2
oxide catalysts was larger than 220 m /g, and the pore volume
3
after sulfidation with H S/H . Furthermore, detailed product
was in a range between 0.86 and 1.01 cm /g. Although both
2
2
surface area and pore volume of the catalysts decreased slightly
after sulfidation, the sulfidation treatment has no obvious
negative effect on the pore structure of the catalyst in general.
distributions of isobutane dehydrogenation over NiO/SiO
2
catalyst before and after sulfidation are listed in Table S1 in
the Supporting Information.
1
140
dx.doi.org/10.1021/cs5000944 | ACS Catal. 2014, 4, 1139−1143

ACS Catalysis
Letter
For comparison, the industrially relevant catalysts Cr O /
2
3
Al O3 and Pt−Sn/Al O were prepared according to the
2
2
3
2
9,30
literature,
and their dehydrogenation performance was
evaluated under the same reaction conditions. As listed in Table
, both catalysts were less active in dehydrogenation than most
2
Table 2. Catalytic Performance of Commercial Relevant
Cr O and Pt-Based Catalysts for Isobutane
2
3
a
Dehydrogenation
b
c
Cisobutane , wt %
Sisobutene , wt %
d
Cr O /Al O
3
54.5
49.7
48.4
42.7
82.4
86.2
84.9
78.1
2
3
2
e
3
Pt−Sn/Al O
2
f
Cr O /SiO
2
3
2
g
Pt/SiO2
a
Reaction conditions: temp, 560 °C; 4 g of catalyst loaded; 14.3 vol %
−1
b
i-C H in nitrogen at a total flow rate of 14 mL min . Conversion of
isobutane. Selectivity to isobutene, and all the data were obtained at
4
10
c
Figure 2. Catalytic activity of metal sulfide catalysts and commercial
relevant catalysts for isobutane dehydrogenation.
d
very beginning of the reaction. Containing 15 wt % Cr O .
2
3
e
f
Containing 0.5 wt % Pt and 1.5 wt % Sn. Containing 13 wt %
g
Cr O . Containing 13 wt % Pt.
2
3
exhibited relatively low catalytic activity (see SI Table S2). For
ZnS/SiO , MnS/SiO , MoS /SiO , and FeS/SiO catalysts after
reaction, the intensity of the diffraction peaks corresponding to
2
2
2
2
2
of the sulfide catalysts. The isobutene yield decreased in the
following order: CoS/SiO > FeS/SiO > NiS/SiO > Cu S/
metal sulfides decreased significantly. In addition, MoO was
2
2
2
2
2
SiO > MnS/SiO > MoS /SiO > Cr O /Al O > Pt−Sn/
also detected on MoS /SiO catalyst after reaction, indicating
2
2
2
2
2
3
2
3
2
2
Al O > ZnS/SiO . The isobutene yield and isobutene
the loss of sulfur species. As for the Cu S/SiO catalyst, low
2
3
2
2 2
selectivity versus conversion curves for all tested catalysts are
compared in Figure S1 in the Supporting Information.
To exclude the effects of metal content and support, 13Pt/
SiO and 13Cr O /SiO catalysts were prepared and evaluated
chalcocite, another form of Cu S, appeared in much lower
2
intensity after reaction. Moreover, accompanied by the
disappearance of CoS and NiS, Co S and Ni S emerged on
9
8
3 2
the CoS/SiO2 and NiS/SiO2 catalysts after reaction,
respectively. In summary, although the pathways of sulfur loss
were different for these sulfide catalysts, realized by either the
decrease in the amount of original metal sulfides or the
formation of new compounds with less sulfur content, the loss
of sulfur was probably the primary reason for the deactivation
of the catalysts. To quantify the sulfur content on sulfided
catalysts, elemental analysis measurement has been carried out.
2
2
3
2
in isobutane dehydrogenation, with the results integrated in
Table 2. The dehydrogenation performance was not better than
that of Cr O /Al O and Pt−Sn/Al O catalysts, which
2
3
2
3
2
3
probably resulted from the weak interaction between the silica
support and active component. Moreover, the aggregation of Pt
on 13Pt/SiO catalyst could also contribute to the unsat-
2
isfactory dehydrogenation performance.
Moreover, when Al O and MgAl O were taken as supports
Sulfur contents of the fresh sulfided NiO/SiO catalyst and the
2
3
2
4
2
of metal sulfide catalysts, similarly high levels of isobutane
conversion and isobutene selectivity could be achieved (see
Supporting Information (SI) Figure S2), demonstrating that
the excellent performance of metal sulfides has little to do with
the support.
catalysts after 1 and 8 h of reaction were 5.38, 5.11, and 3.43 wt
%, respectively, directly reflecting the sulfur loss from the
catalyst during the reaction.
To further determine the variations of surface sulfur species
during the reaction, XPS characterization was performed over
On the basis of the above discussions, the metal sulfides
detected by XRD) were probably responsible for the
sulfided NiO/SiO catalysts before and after reaction. The S 2p
2
(
XPS spectra are illustrated in SI Figure S4. For the sulfided
catalysts before reaction, observation of the doublet character-
istic of S2 species directly supported that the metal oxides on
the catalysts were converted to the corresponding sulfides after
sulfidation treatment. After 8 h of reaction, the intensity of the
significantly improved dehydrogenation performance. More-
over, according to the literature, sulfur adsorbed at the kinked
edges of the Pt catalysts substantially deactivated cracking
−
3
1
32
sites and inhibited C−C bond breaking. Therefore, one
possible explanation is that, with the introduction of sulfur, the
steric hindrance of the metal−sulfur bond hindered the
interaction between the carbon atoms of isobutane molecule
and the metal sites active for C−C bond cleavage and,
consequently, led to limited cracking reactions and improved
dehydrogenation performance.
2−
S
doublet decreased greatly. For the Ni 2p XPS spectra
shown in SI Figure S5, a characteristic peak at 852.4 eV
represents the presence of NiS, and the intensity of this peak
decreased after reaction, which is consistent with the results of
S 2p XPS spectra. Such observation further proved the
consumption of metal sulfide during the reaction.
Although the dehydrogenation performance was improved
remarkably after sulfidation, the catalytic activity of the metal
sulfide catalysts was much less stable than that of the
commercial catalysts (see Figure 2). Taking the sulfided
Furthermore, an online mass spectrometer was applied to
detect the released gases during isobutane dehydrogenation
over the sulfided catalysts. For the sulfided NiO/SiO catalyst,
2
in addition to isobutane and the main products (including
hydrogen and isobutene), hydrogen sulfide was also detected
(see SI Figure S6), indicating that sulfur was lost mainly in the
form of hydrogen sulfide. Complete mass balance calculation of
the reaction has also been conducted, with the results listed in
NiO/SiO catalyst as an example (see SI Figure S3a), the
2
decreased dehydrogenation activity was probably caused by the
change in the active sites. Figure 1b illustrates the XRD patterns
of SiO -supported metal sulfide catalysts reacted for 8 h, which
2
1
141
dx.doi.org/10.1021/cs5000944 | ACS Catal. 2014, 4, 1139−1143

ACS Catalysis
Letter
Table S3 in the Supporting Information. As listed in the table,
the sulfur mass balance and overall mass balance reached up to
catalysts. Nevertheless, possible sulfur loss occurs during the
application of these catalysts, and therefore, measures for sulfur
replenishment should be implemented.
9
8 and 97 wt %, respectively. Moreover, the mass balance data
indicated that the reaction was not a stoichiometric reaction. If
the reaction occurred in a stoichiometric way, that is, , NiS + i-
ASSOCIATED CONTENT
* Supporting Information
Further details are given in Figures S1 to S8 and Tables S1 to
S3. This material is available free of charge via the Internet at
http://pubs.acs.org.
■
C H → i-C H + H S + Ni, only 0.019 g of isobutane could
S
4
10
4
8
2
be converted according to the sulfur loss weight of the catalyst
0.0108 g). However, the reacted isobutane (0.181 g) was
(
much more than that amount, so it could be concluded that the
reaction was not stoichiometric, and the catalyst served more
like a catalyst rather than the reactant.
AUTHOR INFORMATION
Corresponding Author
E-mail: chyli@upc.edu.cn, chyli_upc@126.com.
Notes
■
*
To recover the catalyst activity, the spent NiS/SiO catalyst
2
was sulfided by H S/H for another 3 h after 8 h reaction, and
2
2
in total, five sulfidation−reaction cycles were conducted. Within
the five cycles, the initial isobutane conversion was maintained
at a relatively steady level (see SI Figure S3b), that is, the
catalyst activity could be fully recovered after every sulfidation
treatment. In practice, to reduce cost, coproduced sulfur-
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
This work was financially supported by the National Natural
Science Foundation of China (No. U1362201), the National
73 Program of China (No. 2012CB215006), and the
Fundamental Research Funds for the Central Universities
No. 12CX06040A).
■
containing dry gas can be used to replace H S to treat the
2
catalyst in the sulfidation process.
To deepen the understanding of catalyst deactivation,
9
isobutane dehydrogenation over sulfided NiO/SiO catalyst
2
(
under a cofeed of H S (2 mL/min) has been further carried
2
out, and the results are illustrated in Figure S7 in the
Supporting Information. It can be observed that both isobutane
conversion and isobutene selectivity decreased to some extent
in the initial 2 h. The catalytic performance remained relatively
steady afterward, with isobutene yield up to 50 wt %, which was
indicative of improved catalyst stability under the cofeed of
REFERENCES
■
(1) Xie, H.; Wu, Z.; Overbury, S. H.; Liang, C.; Schwartz, V. J. Catal.
009, 267, 158−166.
2) Centeno, M. A.; Debois, M.; Grange, P. J. Catal. 2000, 192, 296−
06.
3) Zimmermann, H.; Versluis, F. U.S. Patent 5378350, 1995.
4) Buonomo, F.; Jezzi, R.; Notari, B.; Kotelnikov, G. R.; Michailov,
K. R.; Patanov, V. A. U.S. Patent 4746643, 1988.
5) Fritz, P. M.; Boelt, H. V.; Hackner, H.; Van, D. G. J. EP. Patent
0947247, 1999.
2
(
3
(
(
H S. Meanwhile, a small quantity of sulfided hydrocarbons,
2
including methanethiol, ethanethiol, and thiophen, were
generated during the reaction.
It should also be mentioned that a small amount of coke
(
(
0.15 wt % for the spent NiS/SiO catalyst after 8 h reaction)
(6) Iezzi, R.; Buonomo, F.; Sanfilippo, D. U.S. Patent 5143886, 1992.
(7) Herber, R. R.; Thompson, G. J. U.S. Patent 4806624, 1989.
2
was accumulated on the catalyst during the continuous reaction
process, which also contributed to the catalyst deactivation to
some extent. However, it was fortunate that the treatment with
H S/H could eliminate most of the deposited coke at the
reaction temperature of 560 °C, as evidenced by the H -TPR-
MS profile of the spent NiS/SiO catalyst illustrated in SI
(
8) Olbrich, M. E.; McKay, D. L.; Montgomery, D. P.; Jean, B. U.S.
Patent 4926005, 1990.
9) Li, J.; Wang, C.; Huang, C.; Sun, Y.; Weng, W.; Wan, H. Appl.
Catal., A 2010, 382, 99−105.
10) Schimmoeller, B.; Jiang, Y.; Pratsinis, S. E.; Baiker, A. J. Catal.
010, 274, 64−75.
11) Trotus, I.-T.; Teodorescu, C. M.; Par
ChemCatChem 2013, 5, 757−765.
(
2
2
2
(
2
2
Figure S8, which demonstrated the removal of coke in the form
of methane under hydrogen atmosphere. Given that the sulfur
loss and coke deposition occurred simultaneously during the
reaction, to achieve continuous operation, a circulating fluidized
bed reactor equipped with a dehydrogenation reactor, it is
suggested that a regenerator for coke burning and a sulfidation
section be employed in future industrial application. However,
the coke deposited on the catalyst during isobutane
dehydrogenation was low, and the heat released from coke
burning is not sufficient for the reaction. Therefore, an
appropriate fuel (e.g., coproduced dry gas) can be introduced
and burned in the regenerator to supply extra heat and achieve
heat balance.
In conclusion, metal sulfide catalysts, a novel class of
nonnoble metal catalysts for alkane dehydrogenation, exhibited
relatively satisfactory dehydrogenation performance with
improved activation ability of the C−H bond over the C−C
bond. In addition to the active components, such as Mo, Ni,
and Co widely applied in hydrogenation processes, the sulfides
of Cu, Mn, and Fe, barely reported in hydrogenation reactions,
also showed excellent performance. Under the same operating
conditions, the performance of these catalysts was even better
than that of the industrial Cr O /Al O and Pt−Sn/Al O
(
̧
̂
vulescu, V. I.; Marcu, I.-C.
(12) Siahvashi, A.; Chesterfield, D.; Adesina, A. A. Ind. Eng. Chem.
Res. 2013, 52, 4017−4026.
(13) Zhang, Q.; Cao, C.; Xu, T.; Sun, M.; Zhang, J.; Wang, Y.; Wan,
H. Chem. Commun. 2009, 45, 2376−2378.
(14) Wang, H.; Cong, Y.; Yang, W. Chem. Commun. 2002, 38, 1468−
1
469.
15) Wang, G.; Li, C.; Shan, H.; Wu, W. Ind. Eng. Chem. Res. 2013,
2, 13297−13304.
16) Halle, L.; Houriet, R.; Kappes, M. M.; Staley, R. H.; Beauchamp,
J. J. Am. Chem. Soc. 1982, 104, 6293−6297.
17) Blomberg, M. R. A.; Siegbahn, P. E. M.; Nagashima, U.;
(
5
(
(
Wennerberg, J. J. Am. Chem. Soc. 1991, 113, 424−433.
(18) Jacobson, D.; Freiser, B. J. Am. Chem. Soc. 1983, 105, 5197−
5
206.
(19) Georgiadis, R.; Fisher, E. R.; Armentrout, P. J. Am. Chem. Soc.
1
(
989, 111, 4251−4262.
20) Hepp, H. J.; Johnson, M. M. U.S. Patent 3280210, 1966.
(21) Resasco, D. E.; Marcus, B. K.; Huang, C. S.; Durante, V. A. J.
Catal. 1994, 146, 40−55.
(
63.
22) Kobayashi, J.; Shimizu, T. Bull. Chem. Soc. Jpn. 2000, 73, 759−
7
(23) Skrabalak, S. E.; Suslick, K. S. J. Am. Chem. Soc. 2005, 127,
9990−9991.
2
3
2
3
2
3
1
142
dx.doi.org/10.1021/cs5000944 | ACS Catal. 2014, 4, 1139−1143

ACS Catalysis
Letter
(
24) Dhas, N. A.; Ekhtiarzadeh, A.; Suslick, K. S. J. Am. Chem. Soc.
001, 123, 8310−8316.
25) Zhang, L.; Yu, J. C.; Mo, M.; Wu, L.; Li, Q.; Kwong, K. W. J.
Am. Chem. Soc. 2004, 126, 8116−8117.
26) Wang, G.; Meng, Z.; Liu, J.; Li, C.; Shan, H. ACS Catal. 2013, 3,
992−3001.
27) Li, C.; Wang, G.; Sun, N.; Sun, Y. C.N. Patent 201210536414.7,
012.
28) Li, C.; Wang, G.; Sun, N.; Sun, Y. C.N. Patent 201310014789.1,
013.
29) Weckhuysen, B. M.; Schoonheydt, R. A. Catal. Today 1999, 51,
23−232.
30) Iglesias-Juez, A.; Beale, A. M.; Maaijen, K.; Weng, T. C.; Glatzel,
P.; Weckhuysen, B. M. J. Catal. 2010, 276, 268−279.
31) Coughlin, R. W.; Hasan, A.; Kawakami, K. J. Catal. 1984, 88,
63−176.
32) Somorjai, G. A.; Li, Y. Introduction to Surface Chemistry and
Catalysis; Wiley: Hoboken, NJ, 2010.
2
(
(
2
(
2
(
2
(
2
(
(
1
(
1
143
dx.doi.org/10.1021/cs5000944 | ACS Catal. 2014, 4, 1139−1143