Study of the isomerization of 13C labelled methylpentanes on oxygen modified bulk tungsten carbides

Article abstract of DOI:10.1039/b000635l

The reactions of labelled 2- and 3-methylpentanes were carried out on three different oxygen modified bulk tungsten carbides varying by their oxygen treatment temperature. In all cases, isomerization took place via bond shift mechanisms; no cyclic mechanism was involved in contrast to metals like Pt, Pd or Ir where such mechanism occurs. Alkoxy and σ-alkyl intermediates were put forward as adsorbed species responsible for isomerization for bulk tungsten carbides treated by O₂ at moderate temperature (350 °C) over a short period of time (5 min) and at high temperature (700 °C) for 4 h, respectively. These intermediate species are respectively correlated to an acidic and a metallic behavior of the catalytic surfaces. The increase of the O₂ treatment time at 350 °C results in the presence of both kinds of these mechanisms for isomerization, a bifunctional behavior takes place where metallic and acidic characters are present. To explain the presence of terminal-¹³C on acidic catalysts, a 'protonated cyclopropane alkoxy species' intermediate responsible for isomerization was proposed. The driving force for the formation of such species being the presence of stable alkoxy species, general reaction pathways are discussed for the different kinds of catalysts.

Full text of DOI:10.1039/b000635l

View Article Online / Journal Homepage / Table of Contents for this issue
Study of the isomerization of 13C labelled methylpentanes on oxygen
modiÐed bulk tungsten carbides
V. Keller,* F. Garin and G. Maire
L aboratoire dÏEtudes de la Reactivite Catalytique, des Surfaces et Interfaces, U.M.R. 7515,
CNRS-UL P-ECPM, 25, Rue Becquerel, BP 08, 67087 Strasbourg Cedex 2, France.
E-mail: vkeller=chimie.u-strasbg.fr
Received 24th January 2000, Accepted 10th April 2000
Published on the Web 9th June 2000
The reactions of labelled 2- and 3-methylpentanes were carried out on three di†erent oxygen modiÐed bulk
tungsten carbides varying by their oxygen treatment temperature. In all cases, isomerization took place via
bond shift mechanisms; no cyclic mechanism was involved in contrast to metals like Pt, Pd or Ir where such
mechanism occurs. Alkoxy and r-alkyl intermediates were put forward as adsorbed species responsible for
isomerization for bulk tungsten carbides treated by O at moderate temperature (350 ¡C) over a short period of
2
time (5 min) and at high temperature (700 ¡C) for 4 h, respectively. These intermediate species are respectively
correlated to an acidic and a metallic behavior of the catalytic surfaces. The increase of the O treatment time
2
at 350 ¡C results in the presence of both kinds of these mechanisms for isomerization, a bifunctional behavior
takes place where metallic and acidic characters are present. To explain the presence of terminal-13C on acidic
catalysts, a ““protonated cyclopropane alkoxy speciesÏÏ intermediate responsible for isomerization was
proposed. The driving force for the formation of such species being the presence of stable alkoxy species,
general reaction pathways are discussed for the di†erent kinds of catalysts.
2% O was introduced in Ñowing H after activation at dif-
ferent temperatures and for di†erent periods of time: (i) at
1. Introduction
2
2
Catalytic properties of bulk tungsten carbides for hydrocar-
bon conversion have been widely studied. Tungsten carbides
show interesting catalytic properties for many chemical
reactions1h3 and some of these properties have been compared
to those of platinum.4 Recently, it has been shown that the
presence of oxygen plays a determinant role for activity and
selectivity of tungsten carbides,3,5,6 as well as for their chemi-
sorptive properties.7 Previous works8 have clearly demon-
strated that fresh bulk tungsten carbides without oxygen show
a metallic behavior leading mainly to extensive cracking of
hydrocarbons. Their treatment by oxygen leads to a decrease
of this important cracking activity and the isomerization
process is enhanced.6,9 Moreover, the temperature at which
350 ¡C for 5 min; this sample will be denoted WC
at 700 ¡C for 4 h; this sample will be denoted WC
(iii) at 350 ¡C for 3 h; this sample will be denoted WC
, (ii)
and
.
(Ov350)fm
(Ov700)4 h
(Ov350)3 h
2.2. Labelled hexanes
The experiments were performed by replacing commercial 2-
and 3-methylpentanes by labelled ones in the classical Ñow
system previously described in ref. 11. Labelled hydrocarbons
were synthetized by catalytic dehydration of the correspond-
ing labelled alcohols, obtained by Grignard reaction. The
resulting oleÐns were then hydrogenated leading to the
labelled saturated hydrocarbon.12 The procedure, the calcu-
lation of the isomer isotopic species distributions, as well as
the location of the 13C atom in the molecules have already
been described elsewhere.13 The di†erent isotopic species were
characterized by GS-MS coupling.
the bulk tungsten carbide powders are treated by O , has a
2
determinant inÑuence on their catalytic properties10 and on
the adsorbed species involved in cracking and isomerization.
The aim of this work consists in determining the di†erent
adsorbed species involved in isomerization as a function of
3. Catalytic activity of labelled reactants
temperature of treatment by O using 13C tracer studies.
2
Isomerization
of
three
labelled
hydrocarbons
(2-
3-
Thus, isomerization processes of 2-methylpentane(2-13C), of 2-
methylpentane(2-13C),
2-methylpentane(4-13C) and
methylpentane(4-13C) and of 3-methylpentane(3-13C) have
methylpentane(3-13C)) was studied on three kinds of oxygen
been studied on bulk tungsten carbides treated by O at 350
2
modiÐed bulk tungsten carbides. The partial pressures of
and 700 ¡C.
hydrocarbon and H are approximately 5 and 755 Torr,
2
2. Experimental
2.1. Catalyst samples
respectively. The reaction temperature is 350 ¡C and the
amount of catalyst (200 mg) was chosen so that the conversion
around 10% could be considered as very close to the initial
distribution.
The isotopomer distributions are summarized in Table 1,
where the started labelled reactants are reported in the Ðrst
column and the di†erent isotopic species obtained are noted
in the other columns. It must be noted that the following iso-
topic species: 2-methylpentane(2-13C) and 2-methylpentane(3-
13C), 2-methylpentane(4-13C) and 2-methylpentane(5-13C), as
well as 3-methylpentane(2-13C) and 3-methylpentane(1-13C)
The catalyst samples we used were prepared in the Catalysis
Laboratory of Lille (G. and L. Leclercq) in cooperation under
a European Community Contract (C.E.E. ST2J04607-CCTT).
They were prepared from WO by reduction to metallic W,
3
followed by carburization with a mixture of 20% CH ÈH at
4
2
800 ¡C and passivated by a previously described method.11
Prior to catalytic tests, the samples were activated by a hydro-
gen Ñux at 700 ¡C for 5 h.
DOI: 10.1039/b000635l
Phys. Chem. Chem. Phys., 2000, 2, 2893È2902
This journal is ( The Owner Societies 2000
2893

View Article Online
Table 1 Isotopic species distributions of the isomers from labelled 2-methylpentanes and 3-methylpentane at T \ 350 ¡C obtained with the
three di†erent oxidized WC bulk samples
WC
(Ov700)4 h
1.4
1.6
1.4
3.5
3.6
2.1
0.5
8.5
3.5
33.5
È
È
0.7
2.4
2.9
7.3
8.5
24.0
28.9
32.8
0.7
2.4
0.2
20.1
16.1
5.5
8.3
7.9
4.9
20.0
12.7
33.9
È
WC
WC
(Ov350)fm
(Ov350)3 h
0.4
0.9
0.7
1.5
21.7
34.2
0.7
È
È
1.2
1.9
6.5
69.2
53.1
1.1
0.0
2.6
2.4
0.7
1.2
0.0
0.09
5.5
1.0
1.3
3.0
7.8
0.8
È
È
2.5
6.0
24.2
4.0
22.4
30.2
1.7
21.6
8.8
12.2
10.3
25.7
0.0
10.9
Scheme 1 (a) Example of r-alkyl adsorbed species for hexane isomerization. (b) Di†erent routes for methyl shift isomerization of 2-
methylpentane(4-13C) via r-alkyl adsorbed intermediates.
2894
Phys. Chem. Chem. Phys., 2000, 2, 2893È2902

View Article Online
are not distinguishable because their mass spectra are very
similar in the C(5) (demethylation fragmentation) and C(4)
regions (deethylation fragmentation).12 Furthermore, the dif-
ferent isotopic species of 2,2-dimethylbutane can not be quan-
tiÐed because of the absence of the parent peak group; this is
why they have not been reported in Table 1.
\ 3.4 and (iii) the relative amount of
\ 28.9%.
(c) With 3-methylpentane(3-13C) we note that the relative
contribution of the self isomerization is equal to 35.7%.
Moreover, we have, with the two 2-methylpentanes, almost
the same relative contributions for chain lengthening, n-
hexane formation, for chain shortening, 2,3-dimethylbutane
formation and for 3-methylpentane formation, which are in
the ratio 40 : 7 and 53%. Starting from 3-methylpentane, we
have 16, 5 and 78% for n-hexane, 2,3-dimethylbutane and 2-
methylpentane formation, respectively.
3.1. On WC
(Ov700)4 h
Isomerization of the three labelled methylpentanes, 2-
methylpentane(2-13C),
2-methylpentane(4-13C)
and
3-
methylpentane(3-13C) has been performed on this catalyst
sample.
3.1.1. Isomerization of 2-methylpentane(4-13C). Isomer-
ization of 2-methylpentane (4-13C) mainly leads to di†erent
isotopic species of 3-methylpentane and n-hexane, with a 3-
methylpentane/n-hexane ratio of about 1. As already shown,6
this value of 1 is characteristic of a metallic behavior of the
catalyst sample. The most abundant labelled species are those
obtained from self-isomerization, i.e., 2-methylpentane(2-13C)
(33.5%). Isomerization into 3-methylpentane(2-13C) and n-
hexane(2-13C) represents 24.0 and 20.1%, respectively.
With the same intermediate species, via only bond shifts,
from both 2-methylpentanes (2-13C) and (4-13C), we should
obtain the same isomer distributions, and this is in fact the
case:
Let us see if these adsorbed intermediates implying r-alkyl
adsorbed species and mainly methyl, ethyl and propyl shifts
are consistent with the di†erent isotopic distributions
observed here. Starting from 2-methylpentane(4-13C), Scheme
1b shows the di†erent methyl shifts issued from r-alkyl inter-
mediates. It must be noted that the possibilities of methyl and
ethyl shifts leading to 2,3-dimethylbutane and 2,2-dimethyl-
butane are not represented because their isomers are formed
in a small amount. Scheme 1b clearly explains the important
amounts of 3-methylpentane(2-13C) and n-hexane (2-13C). At
this stage, if a propyl shift in chain lengthening takes place, it
is impossible to detect it due to the position of the 13C at C(2)
or C(5).
Starting from 2-methylpentane(2-13C), methyl and propyl
shifts explain well the formation of 28.9% 3-methylpentane(2-
13C), and of 16.1% n-hexane(2-13C) plus 7.9% n-hexane(3-
13C) (Schemes 2a and b). The propyl shift may also exist when
starting from 2-methylpentane(4-13C) but it will not bring any
isotopomer other than n-hexane(2-13C), already obtained via
methyl shifts (Scheme 1b).
Concerning isomerization of 3-methylpentane(3-13C), the
di†erent methyl and ethyl shifts (except for the formation of
2,2- and 2,3-dimethylbutanes) are shown in Scheme 3 and
account for the same amount in the formation of 33.9% 2-
methylpentane(3-13C) and 32.8% 3-methylpentane(2-13C).
Nevertheless, starting from 2-methylpentane(4-13C), the
most abundant isotopic species observed, the 2-
methylpentane(2-13C) cannot be formed directly from a simple
methyl or ethyl shift. The same observation applies when it
comes to the formation of 20% 2-methylpentane(4-13C) start-
ing from 2-methylpentane(2-13C). However, it could also be
supposed that, whatever the starting molecule is, 3-
methylpentane(2-13C) is a common adsorbed intermediate,
which isomerizes by methyl shift giving large amounts of 2-
methylpentane(2-13C) and 2-methylpentane(4-13C), respec-
tively from 2-methylpentane(4-13C) and 2-methylpentane(2-
13C), as shown in Scheme 4a.
3.1.2. Isomerization of 2-methylpentane(2-13C). As in isom-
erization of 2-methylpentane(4-13C), the value of the ratio
between 3-methylpentane and n-hexane is around 1. The most
important species are 3-methylpentane(2-13C) (28.9%), self-
isomer (2-methylpentane(4-13C) (20.0%)) and n-hexane(2-13C)
(16.1%).
3.1.3. Isomerization of 3-methylpentane(3-13C). Isomer-
ization of 3-methylpentane(3-13C) mainly leads to 2-
methylpentane(2-13C) (33.9%), to self isomerization (3-
methylpentane(2-13C) (32.8%)), and to 2-methylpentane(4-13C)
(12.7%).
Several points have to be emphasized: (i) branched isomers
with lower carbon number on the main chain are formed; (ii)
self isomerization is an important process on such catalysts
and its relative contribution is more or less similar whatever
the starting molecules; and (iii) no oleÐns are formed, and the
relative contribution of the cracking is low. Taking into
account such remarks, we can start to explain these results,
keeping in mind that the reactions are performed under a
large excess of hydrogen.
3.1.4. InterpretationÈmechanisms. As we have already
described for the catalyst sample treated by O at high tem-
2
perature, WC
,10 the isomer distribution can be
(Ov700)4 h
explained by a metallic behavior using r-alkyl adsorbed inter-
mediates. These r-alkyl adsorbed species were put forward by
Rooney et al.14 to explain the isomerization of neopentane on
noble metals without any cracking process. This mechanism
involves three-center orbitals with simultaneous p-bonding to
the metal (Scheme 1a) and accounts for the increase of methyl
shift relative to alkyl shift by increasing the chain length of the
reacting hydrocarbon.
Can such species explain the following results obtained?
(a) With 2-methylpentane(4-13C) we observe that: (i) the
relative contribution of self-isomerization is equal to 34.0%;
(ii) the value of the ratio between
Following on in the same way, isomerization of 3-
methylpentane(2-13C) by ethyl shift can explain the presence
of n-hexane(2-13C) and 3-methylpentane(3-13C) in the isomer
species distribution (Scheme 4b). The formation of these
species is of course less favored because methyl shift is pre-
dominant over ethyl shift. This point can also be explained by
the presence of steric hindrance which decreases ethyl shift
processes. It can also be added that the intermediate species
are probably not very electron deÐcient which explains the
importance of methyl shift compared to ethyl or to other alkyl
shifts.
is equal to 3.3 and (iii) the relative amount of
is equal to 24.0%.
(b) With 2-methylpentane(2-13C), observations similar to
the above mentioned ones can be made: (i) the relative contri-
bution of the self isomerization is equal to 28.5%; (ii) the value
of the ratio
Phys. Chem. Chem. Phys., 2000, 2, 2893È2902
2895

View Article Online
Scheme 2 (a) Di†erent routes for methyl shift isomerization and (b) propyl shift isomerization of 2-methylpentane(2-13C) via r-alkyl adsorbed
intermediates.
3.2. On WC
(Ov350)fm
ratio 3-methylpentane/n-hexane around 16.5, characteristic of
an acidic behavior of the catalyst, in contrast to the
On this catalyst sample modiÐed by O at 350 ¡C, isomer-
WC
sample, which exhibited rather metallic isomer-
2
(Ov700)4 h
ization of 2-methylpentane(2-13C) and 2-methylpentane(4-13C)
ization properties. Among all the isotopic species, the 3-
methylpentane(2-13C) is the most important isomer (69.2%).
Furthermore self-isomerization giving 2-methyl(1-13C)
pentane, for which the labelled carbon is located on the
methyl group, represents 21.7% of the isotopic species. All the
other isotopomers are present in minority.
has been studied and the distributions of the di†erent isotopic
species have been analyzed. We can observe that, on this cata-
lyst, the isomers distributions are quite di†erent from those
obtained on the former system. It indicates that the surface
intermediates are certainly di†erent.
3.2.1. Isomerization of 2-methylpentane(4-13C). The isom-
erization of 2-methylpentane(4-13C) leads to the di†erent iso-
3.2.2. Isomerization of 2-methylpentane(2-13C). As to
isomerization of 2-methylpentane(4-13C), the main isomer
observed is 3-methylpentane(2-13C) (53.1%), with a ratio
between methyl-3-pentane and n-hexane species of 16.5, fol-
lowed by self-isomerization into 2-methyl(-13C)pentane. The
formation of 6.5% of 3-methylpentane(3-13C) is also detected.
topic
species
of
3-methylpentane,
n-hexane,
2,3-
dimethybutane, as well as self-isomerization yielding the iso-
topic species of 2-methylpentane. As already observed on this
sample,6 3-methylpentane is the main isomer produced, with a
2896
Phys. Chem. Chem. Phys., 2000, 2, 2893È2902

View Article Online
Scheme 3 Di†erent routes for methyl and ethyl shift isomerization of 3-methylpentane(3-13C) via r-alkyl adsorbed intermediates.
It must be noted that 2-methyl(-13C)pentane is an unusual
isotopic species observed in our case, starting either from 2-
methylpentane(4-13C) or from 2-methylpentane(2-13C). This
Although, in both cases, the reaction implies the transform-
ation of a tertiary carbenium-like ion into a secondary carbe-
nium ion, the greater proportion of 3-methylpentane(3-13C) is
explained by the stability induced by the inductive e†ect19) of
the isobutyl group compared to the inductive e†ect of the
butyl one, and also by the possible 1-2 hydride migration
giving a tertiary carbenium ion.
A high amount of 2-methyl(1-13C)pentane, with the labelled
carbon on the methyl group, is observed starting from 2-
methylpentane(4-13C) or 2-methylpentane(2-13C). So, its for-
mation may result from successive isomerization in the
adsorbed phase of 3-methylpentane(2-13C) (Scheme 7), into 3-
methylpentane(1-13C), through the formation of a protonated
cyclopropane like-ring, the driving force of which is the forma-
tion of a tertiary carbenium ion [7-1]. Then, this intermediate
can either desorb or be transformed into 2-methyl(1-13C)
pentane. It must be noted, that as 3-methylpentane(2-13C) and
3-methylpentane(1-13C) cannot be distinguished, these reac-
tion pathways represented in Scheme 7 also account for the
high proportion of 3-methylpentane(2-13C) (Table 1).
speciÐcity was not observed on the WC
sample.
catalyst
(Ov700)4 h
3.2.3. InterpretationÈmechanisms. In previous work per-
formed on the catalyst sample, WC
, treated by O at
(Ov350)fm
2
350 ¡C,6 it has already been supposed that the adsorbed inter-
mediates responsible for cracking and isomerization were
alkoxy species leading to a carbenium ion-like mechanism.
This mechanism takes the interaction with the surface into
account, which is not necessarily the case with a ““classicalÏÏ
acidic mechanism implying carbenium ions. The isomerization
mechanism has also been interpreted in such a way that the
presence of OÈH surface groups yields alkoxy species as inter-
mediates [5-1]10,15 when a hydrocarbon molecule approaches
the surface (Scheme 5). Then a 1È2 hydride shift may take
place and gives the two species [5-1] and [5-2]. It has already
been shown15h18 that the covalent CÈO bond of this alkoxy
species can be strongly polarized, up to its ionic form, and
thus, the [5-2] intermediate may lead to a protonated cyclo-
propane [5-3] intermediate and gives the isomer by methyl
shift. Now we will see if this reaction pathway can be con-
Ðrmed using 13C tracer studies.
3.3. On WC
(Ov350)3 h
On this sample, only the isomerization of 2-methylpentanes
(4-13C) and (2-13C) has been studied. The di†erent isotopic
species
of
2,3-dimethylbutane,
2-methylpentane,
3-
Starting either from 2-methylpentane(4-13C) (Scheme 6a) or
from 2-methylpentane(2-13C) (Scheme 6b), the formation of
the corresponding tertiary-like alkoxy species, denoted respec-
tively [A1] and [A2], should be favored due to the most basic
CÈH bond.19 Then a hydride shift occurs and the secondary
alkoxy species, less stable, will react to give the isomer. It is
also shown in Scheme 6a and b that isomerization of the
resulting alkoxy species, [B1] or [B2], can lead to 3-
methylpentane(2-13C) or n-hexane(2-13C) when opening of the
cyclopropane ring occurs in the a or b position respectively.
methypentane and n-hexane issued from these two reactants
are summarized in Table 1.
3.3.1. Isomerization of 2-methylpentane(4-13C). Isomer-
ization of 2-methylpentane(4-13C) mainly gives di†erent iso-
topic species of 3-methylpentane and n-hexane with
a
3-methylpentane/n-hexane ratio around 1.4. This metallic
behavior is rather similar to the one already observed on the
WC
sample (Section 3.1.). Nevertheless, the isotopic
(Ov700)4 h
Phys. Chem. Chem. Phys., 2000, 2, 2893È2902
2897

View Article Online
Scheme 4 (a) Isomerization of 3-methylpentane(2-13C) into 2-methylpentane(4-13C) and 2-methylpentane(2-13C) by methyl shifts via r-alkyl
adsorbed intermediates. (b) Isomerization of 3-methylpentane(2-13C) into n-hexane(2-13C) and 3-methylpentane(3-13C) by ethyl shifts via r-alkyl
adsorbed intermediates.
2898
Phys. Chem. Chem. Phys., 2000, 2, 2893È2902

View Article Online
Scheme 5 Formation of a protonated cyclopropane-like ring from alkoxy species.
species distribution is quite di†erent from this latter one.
Indeed, only a negligible amount (0.8%) of 2-methylpentane(2-
have some more detailed information about the di†erent kinds
of isomerization mechanisms and adsorbed species, depending
on the oxygen treatment.
13C) has been detected, as on the WC
sample, whereas
sample. The
(Ov350)fm
it is the main isotopic species on the WC
(Ov700)4 h
very abundant species are 3-methylpentanes (2-13C) and (3-
13C), followed by n-hexane(2-13C) and n-hexane(3-13C). On
the other hand, we notice the presence of a relatively high
amount of 2-methyl(1-13C)pentane, with the labelled carbon
on the methyl group, which was already detected on the
4.1. WC
(Ov700)4 h
WC
, treated by O at 700 ¡C for 4 h, is supposed to
(Ov700)4 h
2
show metallic character both for cracking and isomerization
processes, taking place through classical metallacyclobutane
and r-alkyl species, respectively;10 in this case a very high
selectivity for isomerization has been observed.10 A general
reaction pathway (Scheme 8a) can be proposed, where it is
shown that almost only methyl shifts are involved (except for
the isomerization of 3-methylpentane(3-13C) into 3-
methylpentane(2-13C), where it is necessary to consider an
ethyl shift and for the isomerization of 2-methylpentane(2-
13C) into n-hexane(2-13C), where a propyl shift may occur).
This is in line with the assumption that r-alkyl adsorbed
intermediates may be responsible for metallic isomerization,
because they account for the predominance of methyl shift
relative to other alkyl shifts.14
WC
catalyst sample.
(Ov350)fm
3.3.2. Isomerization of 2-methylpentane(2-13C). The same
general features as those observed from 2-methylpentane(4-
13C) can be stressed here: (i) a methyl-3-pentane/n-hexane
ratio around 1.3, which is characteristic of metallic properties
for isomerization; (ii) the main isotopic species is 3-
methylpentane(2-13C) and (iii) a very high proportion of 2-
methyl(1-13C)pentane (25.7%), with the labelled carbon on the
methyl group, has been detected, which is rather characteristic
of acidic isomerization, as observed on the WC
sample.
(Ov350)fm
Furthermore, all the di†erent isotopic species produced are
obtained from reactions that take place in one or two steps,
supposing that there is no strong adsorption of the surface
species and that the desorption rate is rather important (not
rate determining). This is in accordance with the predomi-
nance of metallic properties of the sample.
Indeed, looking at the n-hexane isotopic species distribu-
tion, it must be mentioned that it is close to a statistical dis-
tribution.
3.3.3. InterpretationÈmechanisms. From Table 1, it can
been shown that, starting from 2-methylpentane(4-13C) or
from 2-methylpentane(2-13C), the isotopic species distribu-
tions are intermediates between those obtained on
4.2. WC
(Ov350)fm
WC
is supposed to develop few acidic sites,
WC
(Ov700)4 h
postulated than both types of adsorbed species, r-alkyl and
and WC
catalyst samples. Then, it can be
(Ov350)fm
(Ov350)fm
responsible for isomerization, in addition to the majority of
the remaining cracking metallic sites, responsible for 80È85%
of cracking processes.10 This results in the appearance of a
classical bifunctional isomerization mechanism with metallic
sites responsible for dehydrogenation/hydrogenation and
acidic ones on which skeletal isomerization occurs.
alkoxy ones, are present on the surface of WC . Such
(Ov350)3 h
behavior looks like a bifunctional isomerization, where two
kinds of sites are present and responsible for isomerization,
i.e., metallic and acidic.
The corresponding general reaction pathway, deduced from
the present 13C tracer studies is summarized in Scheme 8b,
and takes into account the formation of cyclopropane-like
intermediates resulting from the presence of OH surface
groups leading to alkoxy species as intermediates. In this case
all the isotopic species that are detected are issued from a
common intermediate, 3-methylpentane(2-13C), present in a
large majority (between 55È60%). This catalyst sample is
much more selective than the previous one, because only two
4. Discussion and conclusions
It has already been demonstrated in previous work,6,10 that
the conditions under which bulk tungsten carbides are treated
with oxygen have a determinant inÑuence on the catalytic
properties for alkane cracking and isomerization. The use of
13C tracer experiments conÐrms these results and allows us to
Phys. Chem. Chem. Phys., 2000, 2, 2893È2902
2899

View Article Online
Scheme 6 Isomerization of (a) 2-methylpentane(4-13C) and (b) 2-methylpentane(2-13C) via a protonated cyclopropane-like ring from alkoxy
species.
2900
Phys. Chem. Chem. Phys., 2000, 2, 2893È2902

View Article Online
Scheme 7 Isomerization of 3-methylpentane(2-13C) into 2-methyl(1-13C)pentane via a protonated cyclopropane-like ring from alkoxy species.
main isotopic species (3-methylpentane(2-13C) and 3-
methylpentane(1-13C) (which cannot be distinguished alone)
and 2-methyl(1-13C)pentane) can be observed, the other ones
being present in very low amounts. In this case, the driving
force of the reaction is the stability of the resulting alkoxy
intermediate. All these observations can be well demonstrated
by a cyclopropane ring from alkoxy species (Section 3.2.3.),
especially the formation of the exo-labelled 2-methyl(1-13C)
pentane, which cannot be explained by a ““classicalÏÏ carbe-
nium ion mechanism without the formation of a primary
carbenium ion.
acidity strength is attenuated in favor of an increase of metal-
lic sites for isomerization. The diminution of the acidity
strength, in comparison with WC
is revealed by the
(Ov350)fm
3-methylpentane/n-hexane ratio close to 1 and is conÐrmed by
a decrease in the formation 2-methyl(1-13C) pentane and by a
relative decrease of 3-methylpentane(2-13C). On the other
hand, the distribution of n-hexane isotopic species is much
more like that observed on WC
.
(Ov700)4 h
In conclusion, the use of 13C tracer experiments, in addition
to catalytic reactivity studies, allowed us to point out di†erent
observations: (i) two kinds of isomerization mechanisms can
be distinguished, a metallic and an acidic one, implying
r-alkyl intermediates and OH surface groups generating
alkoxy species, respectively; (ii) the metallic isomerization
mechanism, characteristic by a predominance of methyl shifts,
It has already been shown that protonated cyclopropane
rings account for the exchange of the position of two carbon
atoms in n-butane.20
is detected under strong O treatment; and (iii) the acidic
isomerization mechanism is observed on bulk tungsten
4.3. WC
2
(Ov350)3h
As already shown on this sample,6 looking at the catalytic
results and distribution of all the products formed, 13C tracer
studies conÐrm the coexistence of both kinds of surface
species, r-alkyl and alkoxy ones, resulting in metallic and
carbide modiÐed by oxygen under mild conditions. The for-
mation, in this case, of 2-methyl(1-13C)pentane starting either
from 2-methylpentane(4-13C) or 2-methylpentane(2-13C) is
directly correlated with the presence of a protonated cyclo-
propane ring as the adsorbed species. This latter species is
issued from the alkoxy species and its presence depends on the
polarity of the CÈO bond in the alkoxy species.
acidic isomerization, respectively. Indeed, as the O treatment
2
at 350 ¡C becomes longer, the amount of metallic cracking
sites decreases, leading to a decrease of total activity, the
Phys. Chem. Chem. Phys., 2000, 2, 2893È2902
2901

View Article Online
Scheme 8 General reaction pathway on (a) the WC
(Ov700)4 h
sample and (b) the WC
(Ov350)fm
sample.
11 A. Frennet, G. Leclercq, L. Leclercq and G. Maire, in Proceedings
of the 10th International Congress of Catalysis, Budapest, July
19È24, 1992, (ed. L. Guczi, F. Solimosi and P. Tetenyi, Elsevier,
Amsterdam, 1993, p. 927.
12 F. Garin, PhD Thesis, Strasbourg, France, 1978.
13 F. Garin and F. G. Gault, J. Am. Chem. Soc., 1975, 97, 4466.
14 M. A. McKervey, J. J. Rooney and N. G. Samman, J. Catal.,
1973, 30, 330.
15 Z. X. Cheng and V. Ponec, Catal. L ett., 1994, 25, 337.
16 J. Szabo, J. Perrotey, G. Szabo, J. C. Duchet and D. Cornet, J.
Mol. Catal., 1991, 67, 79.
17 V. B. Kazansky, Acc. Chem. Res., 1991, 24, 379.
18 V. B. Kazansky, M. V. Frash and R. A. Van Santen, Appl. Catal.,
1996, 146, 223.
19 N. L. Allinger, in Chimie Organique, McGraw-Hill, Paris, 1975, p.
282.
References
1
2
3
J. M. Muller and F. G. Gault, Bull. Soc. Chim. Fr., 1970, 2, 416.
R. B. Levy and M. Boudart, Science, 1973, 81, 547.
E. Iglesia, F. H. Ribeiro, M. Boudart and J. Baumgartner, Catal.
T oday, 1997, 15, 307.
4
5
6
7
8
9
R. J. Colton, J. T. Huang and J. W. Rabalais, Chem. Phys. L ett.,
1975, 34, 337.
F. H. Ribeiro, R. A. Dalla-Betta, M. Boudart, J. Baumgartner
and E. Iglesia, J. Catal., 1991, 130, 86.
V. Keller, P. Wehrer, F. Garin, R. Ducros and G. Maire, J.
Catal., 1997, 166, 128.
P. Wehrer, P. Vennegues, J. M. Bastin and G. Maire, Ann. Chim.
Fr., 1993, 18, 128.
V. Keller, P. Wehrer, F. Garin, R. Ducros and G. Maire, J.
Catal., 1995, 53, 9.
20 J. Houzvicka and V. Ponec, Catal. Rev. Sci. Eng., 1997, 39(4), 319.
E. Iglesia, J. E. Baumgartner, F. H. Ribeiro and M. Boudart, J.
Catal., 1991, 134, 523.
10 F. Garin, V. Keller, R. Ducros, A. Muller and G. Maire, J. Catal.,
1997, 166, 136.
2902
Phys. Chem. Chem. Phys., 2000, 2, 2893È2902