Catalytic ring-attachment isomerization and dealkylation of diethylbenzenes over halide clusters of group 5 and group 6 transition metals

Article abstract of DOI:10.1016/j.jcat.2004.01.009

A molybdenum halide cluster, (H₃O)₂[(Mo ₆Cl₈)Cl₆]·6H₂O, possessing an octahedral metal framework was used as a catalyst in a gas flow reactor under 1 atm of hydrogen. On reaction of p-diethylbenzene, dehydrogenation to ethylstyrene proceeded selectively at 300°C. At 400°C, mutual interconversion of o-, m-, and p-diethylbenzenes proceeded selectively. The ethyl group migrated by an intramolecular 1,2-shift mechanism without yielding disproportionation products. Niobium and tungsten chloride clusters with the same metal framework were also active catalysts for the isomerization of p-diethylbenzene. All the reactions resulted in appreciable yields of dealkylation products. The catalytic activity for isomerization can be ascribed to acid sites on the cluster surface, and the catalytic activity for dealkylation, to the metallic nature of the framework metal.

Full text of DOI:10.1016/j.jcat.2004.01.009

Journal of Catalysis 223 (2004) 54–63
www.elsevier.com/locate/jcat
Catalytic ring-attachment isomerization and dealkylation
of diethylbenzenes over halide clusters of group 5
and group 6 transition metals
Satoshi Kamiguchi,a Kunihiko Kondo,b Mitsuo Kodomari,b and Teiji Chihara a,∗
a
Institute of Physical and Chemical Research (RIKEN), Wako, Saitama 351-0198, Japan
b
Department of Applied Chemistry, Shibaura Institute of Technology, Shibaura, Minato-ku, Tokyo 108-8548, Japan
Received 29 September 2003; revised 6 January 2004; accepted 6 January 2004
Abstract
A molybdenum halide cluster, (H O) [(Mo Cl )Cl ] · 6H O, possessing an octahedral metal framework was used as a catalyst in a gas
3
2
6
8
6
2
◦
flow reactor under 1 atm of hydrogen. On reaction of p-diethylbenzene, dehydrogenation to ethylstyrene proceeded selectively at 300 C.
At 400 C, mutual interconversion of o-, m-, and p-diethylbenzenes proceeded selectively. The ethyl group migrated by an intramolecular
◦
1,2-shift mechanism without yielding disproportionation products. Niobium and tungsten chloride clusters with the same metal framework
were also active catalysts for the isomerization of p-diethylbenzene. All the reactions resulted in appreciable yields of dealkylation products.
The catalytic activity for isomerization can be ascribed to acid sites on the cluster surface, and the catalytic activity for dealkylation, to the
metallic nature of the framework metal.
2004 Elsevier Inc. All rights reserved.
Keywords: Isomerization; Dealkylation; Dehydrogenation; Diethylbenzene; Halide cluster; Molybdenum halide; Tungsten chloride; Niobium chloride;
Tantalum chloride
1. Introduction
alytic intermolecular transalkylation (disproportionation) of
toluene to produce p-xylene [5,6], which is used in its
oxidation to form terephthalic acid, and the subsequent re-
action with ethylene glycol to synthesize polyester resins.
Mordenite and ZSM-5 are currently the best catalysts for
those catalytic reactions [7–9]. Chemical modification with
antimony [10], chemical vapor deposition of tetraethoxysi-
lane [11], and chemical liquid deposition of tetraethoxysi-
lane [12] were reported to increase the selectivity of ZSM-5.
Trimethylbenzenes and bulkier polymethylbenzenes were
allowed to react over medium- and large-pore zeolites. Ze-
olites NU-87 that have 12-membered ring pockets on the
crystal surface disproportionated trimethylbenzenes largely
within the micropores, while the isomerization reaction oc-
curred mainly on the external surface [13]. Similarly, large-
pore zeolite Beta disproportionated trimethylbenzenes [14],
tetramethylbenzene, pentamethylbenzene, and hexamethyl-
benzene [15]. Over the mesoporous molecular sieve of
MCM-41, the disproportionation of trimethylbenzenes is
also reported [16]. On the large-pore aluminophosphate
molecular sieve of substituted AlPO-5 materials, dispropor-
tionation competed with isomerization. The disproportiona-
We have previously reported on the catalytic activity of
halide clusters in the isomerization of olefins [1], dehydro-
halogenation of halogenated alkanes [2], dehydration of al-
cohols [3], and decomposition of phenyl acetate to phenol
and ketene [4]. Although these reactions are common, and
are catalyzed mostly by solid acids, halide cluster catalysts
exhibit characteristic features in their selectivity and activity.
We have carried out many reactions to broaden our knowl-
edge on the application of these new catalysts and to shed
light on their characterization from a reactivity point of view.
Some of the attempted reactions were successful and, in this
article, a unique intramolecular ring-attachment isomeriza-
tion of p-diethylbenzene to o- and m-diethylbenzenes and
related reactions are presented.
Much literature has appeared on the catalytic ring-
attachment isomerization of o- and m-xylenes and the cat-
*
Corresponding author.
E-mail address: chihara@postman.riken.go.jp (T. Chihara).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.01.009
2004 Elsevier Inc. All rights reserved.

S. Kamiguchi et al. / Journal of Catalysis 223 (2004) 54–63
55
tion of trimethylbenzenes and toluene occurred on stronger
acid sites, while the isomerization of trimethylbenzenes
and xylenes predominated on weaker acid sites [17]. On
the other hand, very little has been reported on the reac-
tions of ethylbenzene and diethylbenzenes, even though p-
diethylbenzene is an industrial source of p-divinylbenzene.
Some strong molecular acids have been used as homoge-
neous catalysts in solution. The catalytic disproportionation
of ethylbenzene in the presence of AlCl3 yielded benzene,
m- and p-diethylbenzenes, and 1,2,4-triethylbenzene [18].
In an AlCl3–HCl catalytic system, the disproportionation of
m- and p-diethylbenzenes proceeded to yield ethylbenzene
and 1,2,4-triethylbenzene [19]. A protic acid, CF3COOH,
also catalyzed the disproportionation of m-diethylbenzeneto
yield ethylbenzene and 1,2,4-triethylbenzeneby an ethylcar-
benium rearrangement [20]. All of these catalytic reactions
were intermolecular transalkylations. On the other hand, ap-
parent interconversions of o-, m-, and p-diethylbenzenes
have been reported in the presence of an AlCl3 catalyst,
and the isomerization was attributed solely to a sequence
involving an intramolecular 1,2-shift mechanism [21]. This
conclusion is doubtful, because disproportionation products
account for as much as 40% of the products [22]. The
isomerization of p-substituted toluenes catalyzed by AlCl3
has been reported, and only xylenes were isomerized ex-
clusively via an intramolecular 1,2-shift mechanism. The
isomerization of t-butyltoluene proceeded by an intermole-
cular alkyl transfer mechanism, and the isomerization of
ethyltoluenes and isopropyltoluenes proceeded by both of
the above mechanisms [23,24]. On the basis of this result,
the isomerization of diethylbenzenes should follow both the
above mechanisms in the presence of AlCl3. In summary,
the ring-attached ethyl group undergoes mainly intermole-
cular migration by the homogeneous acid catalysts such as
AlCl3, AlCl3–HCl, and CF3COOH.
Similarly, diethylbenzenes disproportionated over solid
catalysts. When Y-zeolite mixed with SiO2–Al2O3 was
employed as a catalyst in a neat liquid of o-, m-, or p-
diethylbenzene at 170 ◦C, disproportionation occurred, af-
fording ethylbenzene and 1,2,4-triethylbenzene. Although
small yields of o-, m-, and p-diethylbenzenes were formed,
they were the secondary intermolecular transalkylation prod-
ucts from ethylbenzene and 1,2,4-triethylbenzene [22]. Dis-
proportionation of ethylbenzene to afford benzene and m-
and p-diethylbenzenes was reported over a ZSM-5 cata-
lyst without yielding o-diethylbenzene [25]. In this case,
o-diethylbenzene was too large to locate in the ZSM-
5 cavities [26]. Similarly, the exclusive formation of p-
diethylbenzene over ZSM-5 has been reported in the reaction
of ethylbenzene with ethylene on a commercial scale [7] or
with ethanol in the laboratory [27]. In the latter case, the
p-diethylbenzene produced subsequently isomerizes to form
the meta-isomer. Generally, in large-pore zeolites, such as
Y-zeolites and mordenite, side chains disproportionate by
way of bimolecular transalkylation. Conversely, in medium-
pore zeolites, such as ZSM-5, disproportionation proceeds
by way of successive dealkylation and realkylation [28].
Thus, ring-attached ethyl groups do not migrate around the
aromatic ring by an intramolecular 1,2-shift mechanism over
conventional heterogeneous solid acid catalysts.
2. Experimental
2.1. Materials and characterization
The cluster complexes [(Nb6Cl12)Cl2(H2O)4] · 4H2O
[29], (H3O)2[(Mo6Cl8)Cl6]·6H2O (1) [30], [(Mo6Br8)(OH)4
(H2O)2] [31], [(Ta6Cl12)Cl2(H2O)4]·4H2O [29], and (H3O)2
[(W6Cl8)Cl6] · 6H2O [32] were prepared according to pub-
lished procedures, followed by repeated recrystallization to
yield well-formed single crystals. Commercial Re3Cl9 (Fu-
ruya Metal) was used without further purification. The crys-
tals of the above complexes were crushed until they passed
through a 150-mesh screen, but not a 200-mesh screen.
Solid-state clusters of [Mo6Cli8]Cl2aCl4a/-a2 (2), were obtained
by heating 1 at a rate of 100 ◦C/h to 200 ◦C, followed by
a dwell at 200 ◦C for 6 h under vacuum [30]. Organic sub-
strates employed were commercial products, and were used
as received.
The Raman spectra of the cluster samples in glass re-
action tubes employed were recorded in situ on a Kaiser
Optical Systems HoloLab 5000 Spectrometer using a Nd-
YAG laser operating at λ = 532 nm, with a 7.6-mm focusing
lens. The data counts were accumulated 30 times over 1-s
intervals. Then, the cluster samples were transferred to an
airtight sample holder in a globe box, and powder X-ray dif-
fraction (XRD) was performed on 20-mg samples using a
MacScience MXP21TA-PO X-ray diffractometer employing
Cu-Kα radiation and a scan rate of 2◦/min.
The trapped reaction products were analyzed using a
Hewlett–Packard 5890 Series II gas chromatograph coupled
with a Jeol Automass System II analyzer, with a 15-m ×
0.25-mm × 0.25-µm DB-1 capillary column and helium car-
rier gas (GC/MS), and a GL Sciences 353B gas chromato-
graph (GLC) with a flame ionization detector, a 30-m ×
0.25-mm×0.25-µm DB-1 capillary column, and helium car-
rier gas. In the GLC analysis, all the products were iden-
tified by comparison with authentic commercial samples.
The catalytic activities were evaluated using an on-line Shi-
madzu 14B gas chromatograph. This was fitted with a flame
ionization detector, a 9.5-m × 3-mm 10% Bentone 34 on
Neopack 1A-packed column to study the reaction of diethyl-
benzenes, a 2-m×3-mm 5% Bentone 34 and 5% dinonyl ph-
thalate (DNP) on Shimalite-packed column for the reactions
of p-xylene, a 2-m × 3-mm 5% Bentone 34 and 5% DNP
on Shimalite-packed column connected to a 3-m × 3-mm
5% Bentone 34 and 5% di-iso-decyl phthalate on Uniport
HP-packed column to study the reactions of p-ethyltoluene.
The GLC used nitrogen as the carrier gas. The gaseous prod-
ucts were identified by using the on-line GLC employing a
3-m × 3-mm Unipaks-packed column.

56
S. Kamiguchi et al. / Journal of Catalysis 223 (2004) 54–63
2.2. Apparatus and procedures
Dehydrogenation
Et–C6H4–Et → Et–C6H4–CH=CH2 + H2.
(3)
(4)
(5)
Gas flow reactions were carried out in a conventional ap-
paratus [1]. The gas flow rate (300 mL/h) was controlled by
an orifice-type flow regulator at ambient pressure, and the
gas was introduced into a temperature-controlled stainless-
steel tube with a glass lining (length = 3 m, i.d. = 2 mm),
which was used as a preheater to heat the carrier gas to
the same temperature as the reactor. The general procedure
consisted of packing a known mass of the cluster complex
(200 mg) in an electrically heated vertical tubular borosili-
cate glass reactor (i.d. = 3 mm) with the aid of quartz glass
wool. Catalyst pretreatment was performed in situ immedi-
ately before the reaction under a stream of gas at the op-
erating temperature for 1 h. Then, the liquid reactant was
introduced at a controlled rate (0.40 mmol/h) via a motor-
driven glass syringe pump through a T-joint attached be-
fore the reactor vessel. The products were introduced into
a temperature-controlled gas sampler (1 mL) connected to
a six-way valve, and were analyzed using the on-line GLC.
The reactor effluent was frozen in a dry-ice trap for subse-
quent analysis, which was performed using the capillary col-
umn GLC. Preliminary experiments showed that the gaseous
products were composed of methane, ethane, and ethylene.
In a typical experiment, a weighed cluster sample of
(H3O)2[(Mo6Cl8)Cl6] · 6H2O (1) (200 mg) was placed in
the glass reaction tube surrounded by a close-fitting copper
tube, which was placed in the center of an electric furnace.
The catalyst sample was heated from room temperature to
400 ◦C, and maintained at that temperature for 1 h in a
stream of hydrogen (300 mL/h) before a kinetic run. Af-
ter a period of 10 min, the temperature reached the set point.
The reaction was initiated by feeding neat p-diethylbenzene
(62 µL/h, 0.40 mmol/h) into the stream of hydrogen using
a microfeeder. Analysis of the trapped products showed that
the material balance on recovery, based on aromatic com-
pounds, was 95% (3–5 h). Formation of methane, ethane,
and ethylene was detected by the gas analyses. The reaction
was monitored every 20 min by sampling the reaction gas
(1 mL) using the six-way valve, which was kept at 135 ◦C,
followed by analysis using the on-line GLC. In this article,
the conversion and selectivity are discussed on the basis of
the data from the on-line GLC analysis.
Intermolecular transalkylation (disproportionation)
2Et–C6H4–Et → Et–C6H5 + C6H3Et3.
Dealkylation in the side chain
Et–C6H4–Et + H2 → Et–C6H4–CH3 + CH4.
3.2. Activation of (H3O)2[(Mo6Cl8)Cl6] · 6H2O (1)
To assign their catalytic activity, the halide cluster com-
plexes were employed in the form of crushed granules,
which compromised their catalytic efficiency. The spe-
cific surface area of the crushed and screened crystals of
(H3O)2[(Mo6Cl8)Cl6] · 6H2O (1) was 3.95 m2/g.
When crystals of 1 in a glass reaction tube were treated
in a stream of hydrogen at 400 ◦C for 1 h, and then al-
lowed to react with p-diethylbenzene, a catalytic reaction
proceeded. The reaction profile is shown in Fig. 1. The cat-
alytic activity fluctuated with time, but leveled off after 3 h.
The main reaction was the ring-attachment isomerization
of Eq. (1), which yielded m-diethylbenzene in 74.0% se-
lectivity, with a smaller yield of o-diethylbenzene in 7.7%
selectivity. Dealkylation of the side chains (Eq. (2)) to yield
ethylbenzene and benzene and dehydrogenation (Eq. (3)) to
yield p-ethylstyrene proceeded to a small extent. However,
no intermolecular transalkylation (Eq. (4)) to yield ethylben-
zene and triethylbenzene was observed.
3. Results and discussion
3.1. Reactions
Fig. 1. Typical reaction profile of p-diethylbenzene over
(H O) [(Mo Cl )Cl ] · 6H O (1). After treatment of 1 (200 mg) in
3
2
6
8
6
2
◦
a stream of H (300 mL/h) at 400 C for 1 h, the reaction was initi-
Ring-attachment isomerization
2
ated by introducing p-diethylbenzene (0.40 mmol/h, 0.62 µL/h) to the
H
stream at the same temperature. Percentage conversion = ((reacted
p-Et–C6H4–Et → o-Et–C6H4–Et + m-Et–C6H4–Et.
Dealkylation of the side chain (dealkylation of the ring
attached alkyl)
(1)
2
material)/(reacted material + recovered material)) × 100. Percentage
selectivity = (product/(total amount of products)) × 100. Methane, ethane,
and ethylene originating from the side chain are not taken into account,
but one carbon atom of those molecules originated from the aromatic ring
equated to one-sixth of the reacted material.
Et–C6H4–Et → Et–C6H5 → C6H6.
(2)

S. Kamiguchi et al. / Journal of Catalysis 223 (2004) 54–63
57
Fig. 3. Effect of temperature on the reactivity and product distribution of
p-diethylbenzene over (H O) [(Mo Cl )Cl ] · 6H O (1) 4 h after the re-
3
2
6
8
6
2
action was initiated. No hydrogenolysis of the aromatic ring was observed
4 h after the reaction started. Methane, ethane, and ethylene originating
from the side chain are ignored in the calculation of composition. The
other reaction conditions are the same as those described in the caption
to Fig. 1. Conversion ("), m-diethylbenzene (2), o-diethylbenzene (Q),
p-ethylstyrene (!), benzene (P), ethylbenzene (1). Products constituting
less than 1% of the products have been omitted.
Fig. 2. Reaction profile of p-diethylbenzene catalyzed by
(H O) [(Mo Cl )Cl ] · 6H O (1) at different temperatures. The other
3
2
6
8
6
2
reaction conditions are the same as those described in the caption to Fig. 1.
The effect of temperature on the reactions over 1 was
examined, and the reaction profiles at various temperatures
are shown in Fig. 2. No catalytic activity was observed be-
low 175 ◦C, and reactions above 200 ◦C showed substan-
tial catalytic activity. Cluster 1 is a catalytically inactive
and coordinatively saturated stable complex without multi-
ple metal–metal bonds and, hence, was changed to an ac-
tive species by the thermal treatment. The activity varied
with time, but reached steady states after 3 h. The steady-
state activities were not proportional to the temperature. The
activity and composition of the products at various tem-
peratures are plotted in Fig. 3. At 300 ◦C, the main reac-
tion occurring was dehydrogenation (Eq. (3)) to yield p-
ethylstyrene in 90% selectivity; however, at 400–475◦C this
changed to ring-attachment isomerization (Eq. (1)) to af-
ford m-diethylbenzene.At higher temperatures, near 500 ◦C,
dealkylation of the side chain (Eq. (2)) proceeded more
to completion with increasing temperature, yielding ethyl-
benzene and benzene. Thus, selectivity depended strongly
on temperature. Different selectivity has been reported with
faujasite [33]. When faujasite was used at 500–568◦C un-
der cracking conditions, p-diethylbenzene yielded a large
range of products by radical and ionic mechanisms. In
this case, dehydrogenation yielded p-ethylstyrene as the
main reaction, with dealkylation yielding ethylbenzene and
p-ethyltoluene to a small extent, with traces of isomerization
products.
der the same reaction conditions, as did the example with no
catalyst (entries 13 and 14 in Table 1). Molybdenum pen-
tachloride cannot be applied as a catalyst under the same
reaction conditions, as it boils at 268 ◦C. From these results,
we inferred that the halide cluster 1 would serve as a cata-
lyst precursor by taking advantage of its low vapor pressure
and high melting point. The conversion was proportional to
the mass of 1 used, and the reaction rate was of zero order
with respect to the mass of p-diethylbenzene added, inso-
far as the conversion was not high and the added mass of
p-diethylbenzene was not small. The conversion was pro-
portional to the volume of carrier gas (entries 2–4), where
p-diethylbenzene was in saturated adsorption. The effect of
pretreatment temperature was examined. The catalytic activ-
ity of 1 treated at 300 ◦C throughout was about half that of
1 treated at 400 ◦C (entries 2 and 6). The former tempera-
ture afforded the dehydrogenation product, p-ethylstyrene,
selectively. When a 400 ◦C pretreatment of 1 was applied to
the reaction at 300 ◦C, the selectivity was the same as that
of 1 treated at 300 ◦C throughout (entries 5 and 6). Hence,
selectivity does not depend on pretreatment temperature, but
on reaction temperature.
When the reaction was performed in a helium stream in
the same manner, catalytic activity decreased noticeably, and
its selectivity changed (entry 7). Dehydrogenation yielding
p-ethylstyrene replaced isomerization as the main reaction,
indicating that hydrogen plays an important role in the reac-
tion. In a nitrogen stream, no reaction was observed. Diffu-
sion of the reactant onto the cluster surface may be hindered
in a viscous gas [34]. Alternatively, nitrogen could be ad-
sorbed onto the Mo metal sites of activated 1.
3.3. Activity of halide clusters
Table 1 lists the catalytic activities of various halide
clusters and related compounds treated in the same way at
400 ◦C. Molybdenum metal showed no catalytic activity un-

Table 1
a
Isomerization, dehydrogenation, and dealkylation of p-diethylbenzene over halide clusters
Entry
Cluster
Carrier Reaction Conversion
Selectivity (%)
Dehydrogenation
o-Diethylbenzene m-Diethylbenzene p-Ethylstyrene p-Divinylbenzene Benzene Toluene Ethylbenzene p-Ethyltoluene
gas
temp.
(%)
Isomerization
Dealkylation
◦
( C)
1
2
3
4
5
6
7
8
9
[(Nb Cl )Cl (H O) ] · 4H O
H
400
400
400
400
300
300
400
400
400
400
400
400
400
400
400
400
4.9
9.5
14.6
28.1
3.4
5.4
1.7
0.0
23.5
20.1
11.7
3.7
0.0
0.0
10.2
7.7
4.5
5.3
2.1
4.1
7.7
–
0.0
3.7
7.9
4.9
–
51.7
74.0
77.3
70.5
4.0
4.1
10.6
–
19.0
4.2
76.2
4.6
–
21.2
4.9
3.9
2.4
90.3
87.0
58.8
–
0.0
81.8
0.3
1.7
–
0.0
0.0
0.0
0.0
0.0
2.7
0.0
–
0.0
9.5
0.0
0.0
–
1.6
2.7
4.0
7.1
0.2
0.2
7.1
–
18.8
0.3
0.7
1.3
–
0.8
0.5
0.6
1.0
0.4
0.2
0.7
–
6.6
0.0
1.1
2.0
–
9.4
9.4
9.2
13.2
1.7
0.7
7.5
–
32.5
0.3
6.6
42.9
–
5.1
0.8
0.5
0.6
1.3
1.0
7.6
–
20.0
0.2
7.2
42.6
–
6
12
2
2
4
2
2
2
2
2
2
2
(H O) [(Mo Cl )Cl ] · 6H O (1)
H
H
H
H
H
3
2
6
8
6
2
b
c
(H O) [(Mo Cl )Cl ] · 6H O (1)
3
2
6
8
6
2
(H O) [(Mo Cl )Cl ] · 6H O (1)
3
2
6
8
6
2
d
(H O) [(Mo Cl )Cl ] · 6H O (1)
3
2
6
8
6
2
e
(H O) [(Mo Cl )Cl ] · 6H O (1)
3
2
6
8
6
2
(H O) [(Mo Cl )Cl ] · 6H O (1)
He
N
2
3
2
6
8
6
2
(H O) [(Mo Cl )Cl ] · 6H O (1)
3
2
6
8
6
2
f
[(Mo Br )(OH) (H O) ]
H
H
H
H
H
H
H
H
6
8
4
2
2
2
2
2
2
2
2
2
2
10 [(Ta Cl )Cl (H O) ] · 4H O
6
12
2
2
4
2
11 (H O) [(W Cl )Cl ] · 6H O
3
2
6
8
6
2
12 Re Cl
3
9
13 Mo metal
14 None
–
8.9
0.0
–
25.7
3.7
–
9.3
78.7
–
0.0
0.0
–
4.4
0.8
–
3.1
1.4
–
32.4
12.2
–
16.2
3.2
g
15 (H O) [(Mo Cl )Cl ] · 6H O (1)/SiO
2.3
1.1
3
2
6
8
6
2
2
16 MoO
2
a
◦
After treatment of cluster (200 mg) in a stream of carrier gas (300 mL/h) at 400 C for 1 h, the reaction was initiated by introduction of p-diethylbenzene (0.40 mmol/h, 62 µL/h) into the carrier gas stream.
Conversion = ((reacted material)/(reacted material + recovered material)) × 100 (%) 4 h after the reaction started. Selectivity = (product/(total amount of products)) × 100 (%) 4 h after the reaction started. No
hydrogenolysis of the aromatic ring was observed 4 h after the reaction started. No disproportionation to triethylbenzenes was observed in all the reactions. Methane, ethane, and ethylene originating from the
side chains are not taken into account.
b
H : 600 mL/h.
2
c
d
e
f
H : 1200 mL/h.
2
◦
Pretreated at 400 C in a stream of hydrogen for 1 h before reaction.
Pretreated at 300 C in a stream of hydrogen for 1 h before reaction.
Naphthalene (3.1%).
5.0 wt% of 1 was supported on silica gel (380 m /g).
◦
g
2

S. Kamiguchi et al. / Journal of Catalysis 223 (2004) 54–63
59
A halide cluster of tungsten, (H3O)2[(W6Cl8)Cl6]·6H2O,
a Group 6 metal, exhibited almost the same catalytic activity
and selectivity with 1 (entry 11). On the other hand, a halide
cluster of tantalum, [(Ta6Cl12)Cl2(H2O)4] · 4H2O, a Group
5 metal, selectively catalyzed the dehydrogenation to yield
p-ethylstyrene and a small quantity of p-divinylbenzene
(entry 10). A cluster of niobium, [(Nb6Cl12)Cl2(H2O)4] ·
4H2O, also a Group 5 metal, showed intermediate selec-
tivity with respect to the above clusters, forming both m-
dimethylbenzene and p-ethylstyrene in significant quantities
(entry 1). Trirhenium nonachloride, Re3Cl9, is reported to
change to metallic rhenium on treatment with hydrogen at
250–300 ◦C [35], which was confirmed by our XRD analy-
sis. This reduced Re selectively catalyzed the dealkylation
of the side chain (Eq. (2)) to yield ethylbenzene and dealky-
lation in the side chain (Eq. (5)) to yield p-ethyltoluene (en-
try 12). One of the main effects of platinum in bifunctional
Pt/Al2O3–ZSM-5 catalyst in the transformation of ethylben-
zene is to increase the rate of dealkylation [36]. A similar
effect has been reported for Pt/Al2O3–zeolite catalysts in
ethylbenzene hydroisomerization [37]. Rhenium is located
in a neighboring position to the platinum group metals in the
Periodic Table, and is similar to the platinum group metals
in some of its properties. Rhenium metal, as well as its ox-
ide, sulfide, and selenide, is a catalyst in the hydrogenation
of organic compounds [38,39].
lysts, which have no micropores, are shown to be applica-
ble to the isomerization of long-chain alkyl compounds.
p-Cymene exhibited quite high reactivity, but the reaction
pathway changed to dealkylation, yielding toluene, which
can be ascribed to the formation of a stable secondary car-
benium ion as an intermediate [22]. p-Propyltoluene yielded
p-cymene in 6.0% selectivity by a skeletal rearrangement of
the n-alkyl group to the secondary alkyl group, which indi-
cates that a carbenium ion had formed in this case [20].
As Table 3 shows, three diethylbenzene isomers reacted
over 1 to achieve similar conversions, with the main reaction
being ring-attachment isomerization, although the selectiv-
ity for the isomerization decreased in the order p- > m- > o-
isomer. Even traces of triethylbenzenes (Eq. (4)) were not
detected in all the reactions. The selectivities for mutual in-
terconversion of the three isomers are illustrated in Fig. 4.
One of the most striking features of the reactions is that o-
diethylbenzene afforded the m-isomer in a selectivity that
was three times that of the p-isomer, even though from a
thermodynamical point of view, the p-isomer is more stable
than the m-isomer [40]. p-Diethylbenzene yielded predomi-
nantly its m-isomer, which was only twice as abundant as the
equilibrium composition, compared with the o-isomer. On
the other hand, the m-isomer afforded the sterically hindered
o-isomer more abundantly than expected from the equilib-
rium composition, when compared with the formation of
the p-isomer. Consequently, the ethyl group preferentially
migrated to the neighboring position of the aromatic ring,
irrespective of the equilibrium distribution.
Compound 1 supported on silica gel was also an active
catalyst. However, its selectivity differed (entry 15). The sur-
face hydroxyl group of the silica gel is a likely participant in
the catalysis reaction.
When a mixture of p-xylene and p-diethylbenzene was
applied to the reaction, ring-attachment isomerization oc-
curred to yield the corresponding o- and m-isomers, as
shown in Table 2. However, the formation of o- and m-
ethyltoluenes, which would be expected from intermolecular
alkyl transfer, was not detected. Furthermore, triethylben-
zenes were not detected in any of the reactions listed in
Tables 1–3. These results clearly demonstrate that the iso-
3.4. Application of 1 to disubstituted benzenes
Table 2 summarizes the catalytic activity of 1 for some p-
substituted benzenes. All substrates tested, except p-cyme-
ne, yielded the corresponding m-isomer as the main product
by ring-attachment isomerization. The halide cluster cata-
Table 2
a
Isomerization and dehydrogenation of p-disubstituted benzenes
Substrate
Conversion
(%)
Selectivity (%)
Isomerization
Dehydrogenation
Dealkylation
Others
o-Isomer m-Isomer
Benzene Toluene Ethylbenzene p-Xylene p-Ethyltoluene
b
p-Xylene
13.9
11.8
9.4
76.5
9.5
5.3
6.0
3.7
0.3
7.7
1.0
71.3
75.6
52.6
10.7
74.0
5.6
–
2.9
23.2
0.1
4.9
–
1.7
1.6
0.0
0.1
2.7
13.4
11.4
11.3
86.6
0.5
–
–
–
–
1.9
0.0
0.8
–
8.3
p-Ethyltoluene
p-Propyltoluene
p-Cymene
1.7
0.0
0.0
9.4
–
0.8
1.3
0.1
0.0
–
0.0
c
6.0
d
2.1
p-Diethylbenzene
0.0

p-Xylene
+
–

5.6
6.3
11.1

e
f
p-Diethylbenzene
8.7
50.6
0.0
5.2
0.7
10.8
a
◦
Reactions were performed at 400 C using 0.40 mmol/h substrate. The other conditions are the same as described for Table 1.
Ethylbenzene (1.2%), 1,2,3-trimethylbenzene (0.6%), 1,2,4-trimethylbenzene (6.3%), and 1,3,5-trimethylbenzene (0.2%).
p-Cymene.
b
c
d
e
f
p-Propyltoluene (1.8%) and m-propyltoluene (0.3%).
Mixture of p-xylene (0.20 mmol/h) and p-diethylbenzene (0.20 mmol/h). The selectivity is based on the total mass of starting materials.
p-Divinylbenzene (10.8%). o-Ethyltoluene and m-ethyltoluene were not detected.

60
S. Kamiguchi et al. / Journal of Catalysis 223 (2004) 54–63
Table 3
a
Reaction of diethylbenzenes and ethylbenzene over (H O) [(Mo Cl )Cl ] · 6H O (1)
Substrate
3
2
6
8
6
2
Conversion
Selectivity (%)
b
(%)
Isomerization
Dehydrogenation
Dealkylation
o-Isomer
m-Isomer
p-Isomer
Benzene
Toluene
Ethylbenzene
Ethyltoluene
o-Diethylbenzene
m-Diethylbenzene
p-Diethylbenzene
6.3
7.9
9.5
1.7
0.3
2.9
–
26.8
7.7
7.7
–
31.7
–
74.0
10.6
–
10.9
47.7
–
–
–
26.2
7.6
4.9
58.8
61.5
88.0
2.2
2.7
2.7
7.1
26.7
8.1
1.4
0.8
0.5
0.7
11.8
3.9
27.5
11.1
9.4
7.5
–
0.1
3.2
0.8
7.6
–
c
d
p-Diethylbenzene
Ethylbenzene
Ethylbenzene
d
–
–
–
–
–
a
◦
Reactions were performed at 400 C using 0.40 mmol/h substrate. The other conditions are the same as listed for Table 1. Disproportionation to yield
triethylbenzenes from diethylbenzenes or diethylbenzenes from ethylbenzene was not detected in all the reactions.
b
o-, m-, and p-isomerism was retained unless otherwise noted. Xylenes were not detected.
m-Ethyltoluene (2.2%) and p-ethyltoluene (1.0%).
In He stream.
c
d
Fig.
4.
Isomerization
of
diethylbenzenes
over
(H O) [(Mo Cl )Cl ] · 6H O (1), showing the selectivity values (in %) at
3
2
6
8
6
2
◦
400 C, 4 h after the reaction was initiated. No hydrogenolysis of the aro-
matic ring was observed. Methane, ethane, and ethylene originating from
the side chain are not taken into account. The other reaction conditions are
the same as those described in the caption to Fig. 1. The thickness of each
arrow is proportional to the corresponding selectivity. The equilibrium
◦
compositions of diethylbenzenes in the vapor phase at 400 C are shown in
brackets.
Fig. 5. XRD patterns of (H O) [(Mo Cl )Cl ]·6H O (1) treated at various
3
2
6
8
6
2
merization proceeded via an intramolecular 1,2-shift of the
alkyl group around the aromatic ring.
temperatures with H (300 mL/h) for 1 h. The XRD patterns of 1 subjected
2
◦
to 4 h reaction at 400 C, Mo Cl (2), and MoO are also shown.
6
12
2
after the pretreatment.
3.5. Development of active sites
ꢀꢁ
ꢂ
ꢃ
(H3O)2 Mo6Cli8 Cl6a · 6H2O (1)
XRD patterns of the pretreated samples at various tem-
peratures in a hydrogen stream for 1 h are shown in Fig. 5,
with that of the Mo6Cl12([Mo6Cli8]Cl2aCl4a/–2a, 2) samples be-
ing synthesized independently. The sample heat-treated at
200 ◦C was assignable to [(Mo6Cl8i )Cla4(H2O)2] (3) [41,42].
Fig. 5 shows that cluster 1 changed above 100 ◦C from 1 via
3 to the extended Mo–Cl–Mo bonded solid-state cluster 2 as
heat was applied up to 400 ◦C (Eqs. (6) and (7), Scheme 1).
An increase of about 50% in the surface area was observed
ꢀꢁ
ꢂ
ꢃ
→
Mo6Cli8 Cla4(H2O)2 (3) + 2HCl + 6H2O,
(6)
(7)
ꢀꢁ
ꢂ
ꢃ
Mo6Cli8 Cl4a(H2O)2 (3)
ꢀ
ꢃ
→ Mo6Cli8 Cla2Cla4/–2a (2) + 2H2O.
The same trends were observed in Raman spectroscopy.
Raman spectra of the samples discussed above are shown
in Fig. 6, in which four assigned peaks [43] are retained up
to 200 ◦C. Above this temperature, the Raman peak orig-

S. Kamiguchi et al. / Journal of Catalysis 223 (2004) 54–63
61
Scheme 1.
coordinated water had not been completely removed by
pretreatment, and that the oxygen of the water molecules
had been captured to form MoO2 at 400 ◦C. The synthe-
sis of 2 from 1 is performed by gradual heating up to
200 ◦C, followed by prolonged heating at the same tem-
perature under vacuum. The coordinated water is removed
by this heat treatment, and is replaced with bridging Cla–a
ligands (Scheme 1). Rapid heating to higher temperatures
would give rise to the formation of imperfect crystals con-
taining defective [Mo6Cli ]ClaCla–a moieties, and exhibiting
catalytic activity, as in t8he c2ase4o/2f olefin isomerization on
(H3O)2[(W6Cl8)Cl6] · 6H2O [1].
We have reported on the catalytic isomerization of
olefins, dehydration of alcohols, dehydrohalogenation of
halogenated alkanes, and decomposition of phenyl acetate to
phenol and ketene by halide cluster catalysts, as mentioned
above. All of these reactions are catalyzed by conventional
solid acids and bases, and the evolution of HCl was observed
on pretreatment of the clusters. By taking into account the
electronegativity of the metal, we deduced that the active
sites on the cluster surface are hydroxo- (Eq. (8)) or oxo-
(Eq. (9)) species as exemplified in the case of the W clus-
ter. The ring-attachment isomerization would be catalyzed
by this acidic species.
Fig. 6. Raman spectra of (H O) [(Mo Cl )Cl ] · 6H O (1) treated at var-
3
2
6
8
6
2
ious temperatures with H (300 mL/h) for 1 h. The Raman spectra of 1
2
◦
subjected to 4 h reaction at 400 C and Mo Cl (2) are also shown. No
6
12
Raman peaks attributable to MoO were observed in this region.
2
ꢀ
ꢃ
ꢀ
ꢃ
inally at 321 cm−1 ascribed to the triple bridged Mo-Cli
breathing vibration was replaced by two peaks, suggesting a
change in the coordination mode, and a new peak appeared
at 278 cm−1. It is at this temperature that catalytic activity
emerged. No appreciable change in the XRD patterns and
the Raman spectra after a 5-h reaction time was observed
with p-diethylbenzene at the same temperatures.
(W6Cl8)Cl4(H2O)2 → (W6Cl8)Cl3(OH)(H2O)
+ HCl,
(8)
(9)
ꢀ
ꢃ
ꢀ
ꢃ
(W6Cl8)Cl3(OH)(H2O) → (W6Cl8)Cl2(O)(H2O)
+ HCl.
When a hydrogen gas stream was replaced with a helium
gas stream in the reaction of p-diethylbenzene, the main re-
action changed from isomerization to dehydrogenation (Ta-
ble 3). In both cases, dealkylation proceeded significantly.
Hydrogen favored isomerization, and was partly used for re-
ductive dealkylations and, when in low concentration, it was
taken from the substrate. Similarly, dehydrogenation of eth-
The Raman spectrum of 1 heat treated at 400 ◦C was
almost the same as that of a sample of cluster 2. How-
ever, closer examination of the XRD patterns shows that
the crystallinity of 1 heat-treated was poor, and that it con-
tained a small quantity of MoO2. This indicated that the

62
S. Kamiguchi et al. / Journal of Catalysis 223 (2004) 54–63
ylbenzene proceeded under a He stream, and, under both He
and H2 streams, dealkylation proceeded (Table 3). The lat-
ter reaction has not been reported for conventional molecular
strong acids or for solid acid catalysts, except under cracking
conditions. Dealkylation of ethylbenzene under hydrogen
has been reported exclusively on platinum group metals and
on Ni:Pt, Ni [44], and Rh [45] on SiO2, Al2O3, or TiO2 and
Ni on mordenite [46]. Dealkylation of propylbenzene over
Pt/SiO2 or Al2O3 [47] and of pentylbenzene over Pt/silica–
alumina catalysts [48] has been reported. Halide clusters of
Group 6 metals always afforded dealkylation products in se-
lectivities of 12–49% at 400 ◦C (Tables 1–3).
lation mechanism. Another characteristic feature of halide
cluster catalysis is that the halide clusters have no micro-
pore structures and, hence, they can be used to catalyze
o-diethylbenzene, which is too large to enter the pores of
zeolite Y and ZSM-5.
The isomerizations over cluster catalysts are always ac-
companied by an appreciable yield of dealkylation products
under a hydrogen stream, particularly over Group 6 metal
clusters. Dehydrogenation proceeds preferentially under a
helium stream. Thus, atmospheric hydrogen, as well as alkyl
hydrogen, is activated over these cluster catalysts, and can
be used to carry out reductive dealkylations that have previ-
ously been reported over platinum group metal catalysts.
In the case of Group 6 metals, another type of ac-
tive site can be envisaged. On heating of a more elec-
tronegative (Group 5 metal) Nb cluster containing aqua lig-
ands, [(Nb6Cl12)Cl2(H2O)4] · 4H2O, the ligand is not lost,
References
and hence a binary complex of [Nb6Cl14]([Nb6Cli10Cl2i–/a2
]
[1] S. Kamiguchi, M. Noda, Y. Miyagishi, S. Nishida, M. Kodomari, T.
Chihara, J. Mol. Catal. A 195 (2003) 159.
Cl Cla–a) is not formed, which can be synthesized by an-
a–i
2/2 4/2
other route involving the comproportionation of Nb metal
with NbCl5 in a sealed Nb metal tube [49]. In contrast, the
aqua ligands of 1 are easily removed by heating at 200 ◦C to
yield the binary complex 2, as mentioned above. Rhenium
chloride, a Group 7 metal complex, is easily and completely
reduced to metallic Re on heating in a hydrogen atmosphere
at above 250 ◦C. Hence, coordinatively unsaturated or, more
likely, hydrido ligands containing Mo or W metal could exist
in the poorly crystallized samples of 2.
[2] S. Kamiguchi, M. Watanabe, K. Kondo, M. Kodomari, T. Chihara,
J. Mol. Catal. A 203 (2003) 153.
[3] S. Kamiguchi, T. Chihara, Catal. Lett. 85 (2003) 97.
[4] T. Chihara, S. Kamiguchi, Chem. Lett. (2002) 70.
[5] L. Cejka, B. Wichterlova, Catal. Rev. Sci. Eng. 44 (2002) 375.
[6] T.-C. Tsai, S.-B. Liu, I. Wang, Appl. Catal. A 181 (1999) 355.
[7] W.W. Kaeding, G.C. Barile, M.M. Wu, Catal. Rev. Sci. Eng. 26 (1984)
597.
[8] A.B. Halgeri, J. Das, Catal. Today 73 (2002) 65.
[9] M.A.A. Kareem, S. Chand, I.M. Mishra, J. Sci. Ind. Res. 60 (2001)
319.
In a series of Y(η5-C5H5)2(alkyl), [Zr(η5-C5H5)2
(alkyl)]+, and [Nb(η5-C5H5)(η4-C4H6)]+ moieties, these
metals are isoelectronic, exhibiting similar polymeriza-
tion activities [50]. It is well known that Ru metals in
RuCl2(PPh3)3 and RuHCl(PPh3)3 are isoelectronic with Rh
metal in RhCl(PPh3)3, and that these complexes are cata-
lysts in the homogeneous hydrogenation of unsaturated C–C
bonds [51]. Thus, Group 6 metal atoms of the halide cluster
complexes could be isoelectronic with the platinum group
metals by taking two or more electrons from the halogen
and hydrido ligands. The selectivity for reductive dealky-
lation by the cluster complexes could be attributable to the
metallic nature of the clusters.
[10] S. Zheng, A. Jentys, J.A. Lercher, J. Catal. 219 (2003) 310.
[11] H. Manstein, K.P. Moller, W. Bohringer, C.T. O’Connor, Micropor.
Mesopor. Mater. 51 (2002) 35.
[12] S. Zheng, H. Heydenrych, P. Roeger, A. Jentys, J.A. Lercher, Stud.
Surf. Sci. Catal. 135 (2001) 1772.
[13] S.-H. Park, H.-K. Rhee, Catal. Today 63 (2000) 267.
[14] J. Cejka, A. Krejci, Stud. Surf. Sci. Catal. 130C (2000) 2627.
[15] A. Sassi, M.A. Wildman, H.J. Ahn, P. Prasad, J.B. Nicholas, J.F. Haw,
J. Phys. Chem. B 106 (2002) 2294.
[16] J.V. Cejka, A. Krejci, N.V. Zilkova, J.V. Dedecek, J.V. Hanika, Micro-
por. Mesopor. Mater. 44 (2001) 499.
[17] E. Dumitriu, C. Guimon, V. Hulea, D. Lutic, I. Fechete, Appl. Catal.
A 237 (2002) 211.
[18] M. Brini, R. Mislin, J.-M. Sommer, Bull. Soc. Chim. Fr. (1965) 3654.
[19] N. Nambu, K. Yamamoto, S. Hamanaka, M. Ogawa, Nippon Kagaku
Kaishi (1979) 925.
[20] H.J. Bakoss, R.M.G. Roberts, A.R. Sadri, J. Org. Chem. 47 (1982)
4053.
[21] G.A. Olah, M.W. Meyer, N.A. Overchuk, J. Org. Chem. 29 (1964)
2313.
4. Conclusions
The halide cluster (H3O)2[(Mo6Cl8)Cl6] · 6H2O (1) de-
veloped catalytic activity above 200 ◦C for p-diethylbenzene.
The activity depended on reaction temperature, with de-
hydrogenation, ring-attachment isomerization, and dealky-
lation increasing with increasing temperature. Groups 5
and 6 metal halide clusters exhibited a characteristic selec-
tive ring-attachment isomerization of diethylbenzenes by the
intramolecular 1,2-shift mechanism without yielding dis-
proportionation products at 400 ◦C. They are distinct from
conventional molecular strong acids, such as AlCl3 and
CF3COOH, or solid acids, such as zeolite Y and ZSM-5, iso-
merization of which is based on an intermolecular transalky-
[22] A.P. Bolton, M.A. Lanewala, P.E. Pickert, J. Org. Chem. 33 (1968)
1513.
[23] R.H. Allen, L.D. Yats, D.S. Erley, J. Am. Chem. Soc. 82 (1960) 4853.
[24] R.H. Allen, J. Am. Chem. Soc. 82 (1960) 4856.
[25] J. Engelhard, J. Catal. 145 (1994) 574.
[26] N. Arsenova-Härtel, H. Bludau, R. Schumacher, W.O. Haag, H.G.
Karge, E. Brunner, U. Wild, J. Catal. 191 (2000) 326.
[27] T. Yashima, J.-H. Kim, H. Ohta, S. Namba, T. Komatsu, Stud. Surf.
Sci. Catal. 90 (1994) 379.
[28] J.S. Beck, W.O. Haag, in: G. Ertl, H. Knözinger, J. Weitkamp (Eds.),
Handbook of Heterogeneous Catalysis, vol. 5, Wiley–VCH, Wein-
heim, 1997, p. 2136.
[29] F.W. Koknat, J.A. Parsons, A. Vongvusharintra, Inorg. Chem. 13
(1974) 1699.

S. Kamiguchi et al. / Journal of Catalysis 223 (2004) 54–63
63
[30] P. Nannelli, B.P. Block, Inorg. Synth. 12 (1970) 170.
[31] J.C. Sheldon, J. Chem. Soc. (1962) 410.
[32] V. Kolesnichenko, L. Messerle, Inorg. Chem. 37 (1998) 3660.
[33] L. Forni, S. Carra, J. Catal. 26 (1972) 153.
[42] L.J. Guggenberger, A.W. Sleight, Inorg. Chem. 8 (1969) 2041.
[43] J.R. Schoonover, T.C. Zietlow, D.L. Clark, J.A. Heppert, M.H.
Chisholm, H.B. Gray, A.P. Sattelberger, W.H. Woodruff, Inorg.
Chem. 35 (1996) 6606.
[34] H.G. Karge, J. Ladebeck, Z. Sarbak, K. Hatada, Zeolites 2 (1982) 94.
[35] H. Gehrke Jr., D. Bue, Inorg. Synth. 12 (1970) 193.
[36] J.M. Silva, M.F. Ribeiro, F.R. Ribeiro, E. Benazzi, M. Guisnet, Appl.
Catal. A 125 (1995) 1.
[44] D. Duprez, A. Miloudi, G. Delahay, R. Maurel, J. Catal. 101 (1986)
56.
[45] D. Duprez, A. Miloudi, G. Delahay, R. Maurel, J. Catal. 90 (1984)
292.
[37] L.D. Fernandes, J.L.F. Monteiro, E.F. Sousa-Aguiar, A. Martinez, A.
Corma, J. Catal. 177 (1998) 363.
[38] W.H. Davenport, V. Kollonitsch, C.H. Kline, Ind. Eng. Chem. 60 (11)
(1968) 10.
[46] S. Narayanan, J. Chem. Soc., Faraday Trans. 1 75 (1979) 434.
[47] S. Toppi, C. Thomas, C. Sayag, D. Brodzki, F.L. Peltier, C. Travers,
G. Djéga-Mariadassou, J. Catal. 210 (2002) 431.
[48] S.M. Csicsery, J. Catal. 15 (1969) 111.
[39] H.S. Broadbent, G.C. Campbell, W.J. Bartley, J.H. Johnson, J. Org.
Chem. 24 (1959) 1847.
[49] A. Simon, H.G. Schnering, H. Wöhrle, H. Schäfer, Z. Anorg. Allg.
Chem. 339 (1965) 155.
[40] C.L. Yaws, P.Y. Chiang, Chem. Eng. 95 (1988) 81.
[41] H. Schäfer, H. Plautz, H. Bauman, W. Beckmann, C. Brendel, U.
Lange, H.-G. Schulz, R. Siepmann, Z. Anorg. Allg. Chem. 389 (1972)
57.
[50] K. Mashima, Macromol. Symp. 159 (2000) 69.
[51] M.A. Bennett, T.W. Matheson, in: G. Wilkinson, F.G.A. Stone, E.W.
Abel (Eds.), Comprehensive Organometallic Chemistry, vol. 4, Perga-
mon Press, Oxford, 1982, p. 931.