Alkane activation over acidic zeolites: The first step

Article abstract of DOI:10.1002/chem.200901737

The heterogeneous acid-catalyzed activation step of alkanes leading to the reaction intermediates (carbocationic or alkoxy species) was up to now the matter of a longstanding controversy. Gas chromatography and online mass spectroscopy measurements show that H2 and methane are formed over H-zeolites, whereas HD and CH₃D are formed over D-zeolites as the primary products in the reaction with isobutane. These results indicate that o-bond protolysis by strong acid sites is the first step for hydrocarbon activation on these catalysts at mild temperatures (473 K), in analogy to the activation path occurring in liquid superacid media.

Full text of DOI:10.1002/chem.200901737

FULL PAPER
DOI: 10.1002/chem.200901737
Alkane Activation over Acidic Zeolites: The First Step
Benoit Louis,[a] M. Maciel Pereira,[b] Fabiana M. Santos,[b] Pierre M. Esteves,[b] and
Jean Sommer*[c]
Abstract: The heterogeneous acid-cata-
lyzed activation step of alkanes leading
to the reaction intermediates (carbo-
cationic or alkoxy species) was up to
now the matter of a longstanding con-
troversy. Gas chromatography and
online mass spectroscopy measure-
ments show that H2 and methane are
formed over H-zeolites, whereas HD
and CH3D are formed over D-zeolites
as the primary products in the reaction
with isobutane. These results indicate
that s-bond protolysis by strong acid
sites is the first step for hydrocarbon
activation on these catalysts at mild
temperatures (473 K), in analogy to the
activation path occurring in liquid su-
peracid media.
Keywords: alkanes · heterogeneous
catalysis · hydrocarbons · zeolites
Introduction
formed above 573 K, but many authors have shown that H/
D exchange between iso-alkanes and zeolites or other solid
acids occurs readily and, regioselectively at temperatures
below 473 K, in the absence of significant cracking or iso-
merisation reactions.[7–11] These results were rationalised by
the involvement of adsorbed carbenium ions in equilibrium
with the corresponding olefin by a fast protonation/depro-
Acid-catalyzed transformation of hydrocarbons, such as
cracking and isomerisation, are of prime importance for the
petrochemical industry.[1,2] The chemical inertness of the
starting alkanes is generally overcome by the use of high
temperatures and strongly acidic catalysts, such as proton-
exchanged zeolites.[3] The activation step of alkanes is well
demonstrated in liquid superacids,[4] whereas the initial step
involved on solid acids has not, up to now, received a clear
answer.[5] It is generally believed that zeolites lack sufficient
AHCTUNGTREGtNNNU onation process before desorption by hydride transfer ac-
cording to Scheme 1. Nevertheless the questions “what is
first: alkene or carbenium ion? or which is the first step in
alkane activation?” were left open.
Herein we describe our recent results which, we believe,
clearly confirm the hypothesis of an initial protolytic cleav-
acid strength[6] to allow direct C C or C H bond protolysis
at mild temperatures (below 500 K), thus excluding the for-
mation of a transient pentacoordinated carbonium ion. Zeo-
lite-catalyzed alkane transformations are generally per-
À
À
À
À
age of C H and C C bonds in the activation of alkanes
[a] Dr. B. Louis
Laboratoire des Matꢀriaux Surfaces et Procꢀdꢀs pour la Catalyse
Member of ELCASS (European Laboratory for Catalysis
and Surface Sciences), UMR 7515 CNRS
25 rue Becquerel 67087 Strasbourg Cedex 2 (France)
[b] Prof. M. M. Pereira, F. M. Santos, Prof. P. M. Esteves
Universidade Federal do Rio de Janeiro
Instituto de Quꢁmica
Av. Athos da Silveira Ramos 149, CT Bloco A
Cidade Universitꢂria 21941-909, Rio de Janeiro (Brazil)
[c] Prof. J. Sommer
Institut de Chimie, UMR 7177
4 rue Blaise Pascal, 67070 Strasbourg Cedex (France)
Fax : (+33)390241486
E-mail: sommer@chimie.u-strasbg.fr
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200901737.
Scheme 1. H/D exchange for isobutane over acidic D-zeolite.
Chem. Eur. J. 2010, 16, 573 – 576
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
573

even at temperatures as low as 473 K. Scheme 2 presents
the two possible protonation pathways, as observed in liquid
superacids.
(Scheme 3), in which the smallest alkane cracking product
remains propane. All our experiments were carried out at
473 K to monitor the initial product distribution. As H2
Scheme 2. Protonation of saturated alkanes.
Results and Discussion
In 1946, Bloch, Pines and Schmerling[12] observed that n-
butane isomerises to isobutane, under the influence of pure
aluminium chloride only in the presence of HCl. They pro-
posed that the ionisation step takes place through initial
protolysis of the alkane as evidenced by formation of minor
amounts of hydrogen in the initial stage of the reaction.
Whereas all studies involving isomerisation, cracking and al-
kylation reactions under acidic conditions agree that carbo-
cations are key intermediates, the mode of the formation of
these reactive species from barely reactive hydrocarbons is
still under debate nowadays.
Scheme 3. Classical oligomerisation and b-scission mechanism showing
that methane formation is unfavourable by this mechanistic route.
could also be formed by dehydrogenation reaction, we have
completed our study by deuteration of the Brçnsted acid
sites of the zeolite. Hence, the formation of HD molecule
would ascertain a superacid-like alkane activation mecha-
nism.
Owing to this long-lasting controversy concerning the ini-
tial step in hydrocarbon conversion over solid acids, the fol-
lowing pathways can be found in the literature: Protolysis,[13]
hydride abstraction by a preexisting carbenium ion,[14] hy-
dride abstraction by a Lewis acid,[10,15] oxidation by means
of traces of metal impurities or protonation of traces of al-
kenes impurities.
In one of the most cited papers in hydrocarbon chemistry,
Haag and Dessau,[16] described the cracking of 3-methylpen-
tane and n-hexane in the presence of acidic zeolites between
623 and 823 K and suggested the coexistence of two compet-
ing mechanistic pathways. Whereas one mechanism involved
a bimolecular hydride transfer in the rate determining step,
the second one described a monomolecular reaction pro-
ceeding via a pentacoordinated transition state, as known in
superacidic conditions.[17] The latter mechanism involved the
active acid sites of the aluminosilicate[18] and was suggested
to predominate at high temperature (823 K).
Reaction on H-ZSM5: Isobutane was recirculated over H-
ZSM5 (acid form of MFI Zeolite) at 473 K and the follow-
ing product distribution was observed (see Figure 1). Isobu-
tane reacts readily on H-ZSM5 even at that low tempera-
ture. However, its conversion remains very low, 1–2% after
150 min of reaction. During the first 20 min on stream, only
H2 was produced, in agreement with a direct protolysis of
the alkane by a Brçnsted acid site of the zeolite.[9] Accord-
We have chosen isobutane as the probe molecule to re-
ACHTUNGTRENNUNGinvestigate the cracking mechanism. It is the smallest, most
reactive alkane due to its tertiary hydrogen. The protolysis
À
of this tertiary C H s bond should produce H2 and a terti-
ary butylcarbenium ion on the zeolite surface.
Figure 1. Product distribution during isobutane recirculation on H-ZSM5
These reaction products cannot be accounted by the
~
&
&
^
*
*
at 473 K methane, ethane, propane, n-butane, ethylene, iso-
classical oligomerisation and b-scission mechanism[5]
pentane, hydrogen.
^
574
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 573 – 576

Activation over Acidic Zeolites
FULL PAPER
ing to the s-reactivity of hydrocarbons,[19,20] the protolysis is
higher resolution of the signal in the mass spectrometer, we
have therefore performed the cracking reaction over deuter-
ated USY zeolite, D-USY.
In this sense, to overcome the doubt that hydrogen could
result from isobutane dehydrogenation, and further confirm
our working hypothesis, we have deuterated the zeolite, re-
À
favoured on the tertiary C H bond, which leads to H2 for-
mation, prior to the appearance of light alkanes. Following
the initial release of H2, the following products were detect-
ed: CH4, C2H6, C2H4 and C3H8. With the exception of meth-
ane, the formation of these light products can be best ration-
alised by the classical cracking mechanism, involving b-scis-
sion.[3] Scheme 3 shows some of the reactions and products
involved in these paths, thus further supports the absence of
methane. Hence, methane can only be formed through a
non-classical mechanism (Scheme 2b).
[21]
À
placing all Brçnsted acid sites by O D groups
(see the
Supplementary Information). Likewise liquid superacid
media, the mechanism proposed (Scheme 2, routes a) and
b)) may be valid, if HD or CH3D were detected. The partial
pressures, monitored by on-line MS, are presented in
Figure 2. Neither isobutene (m/z 56), nor H2 was observed.
Nevertheless, H2 was formed in extremely low amount,
ꢀ10À3 mmol, suggesting that at 473 K only few acid sites are
strong enough to operate a direct protolytic cleavage of iso-
butane to produce H2 and adsorbed tert-butyl carbenium
ion.
To investigate the dependence between the number of
Brçnsted acid sites of the zeolite and the production of hy-
drogen from isobutane cracking, we have varied the amount
of zeolite (hence the number of acid sites), monitoring the
quantity of H2 produced for the different experiments. By
dividing the mass of zeolite by two, the production of hydro-
gen was also divided by the same factor. These results indi-
cate that only 0.1% of zeolite acid sites reacted with isobu-
tane to form H2. Hydrogen, in this closed recirculation
system, is only produced as long as the acid sites are con-
sumed in the protolytic step forming the carbocation.
Figure 1 shows that later on, its amount remains constant, in
contrast with the secondary products resulting from classical
autocatalytic oligomerisation-cracking reaction of isobutane.
The protonation of saturated alkanes may result in the
formation of a non-classical pentacoordinated carbonium
ion, which decomposes into a trivalent carbenium ion and
hydrogen or methane (Scheme 2). Route a) is predominant
over b) according to the s-reactivity concept proposed by
Olah.[19,20]
Figure 2. Profiles of the primary products of [D1]isobutane reaction on
~
*
DUSY: HD (m/z 3) - -, CH3D
N
-&-.
Simultaneously to isobutane (m/z 58) consumption, m/z 3
and 17 ascribed respectively to HD and CH3D were ob-
served as the primary products followed by [D1]isobutane
(m/z 59). As m/z 19 and m/z 20 are not observed, m/z 3
cannot be formed by molecular-ion fragmentation of water
DHO, or D2O within the mass spectrometer. Moreover as
the profile of m/z 3 did not follow in shape that of m/z 59,
consequently m/z 3 can only be ascribed to HD formation
and m/z 17 can only be ascribed to CH3D formation. It was
also observed that the amount of isobutane m/z 58 con-
sumed approximately corresponds to the amount of
[D1]isobutane (m/z 59) formed. Moreover, ZSM5 and USY
frameworks favour the diffusion of reactants toward the
active sites.[23–25] This leads to a lowering in the energy barri-
er to activate alkane and, thus mimicking the behaviour of
superacids.[23,24] It is noteworthy that [D1]isobutane (m/z 59)
produced was neither accompanied by isobutene nor H2 for-
mation. In line with H/D ex-
Reaction on D2O-exchanged ultra stable faujasite (D-USY):
Table 1 presents a detailed characterisation of the acid sites
present on both ZSM5 and USY (ultra stable faujasite) zeo-
lites. The Brçnsted acid sites were titrated by means of a H/
D isotope exchange technique developed in our group.[21a–c]
n-Hexane cracking was used as a model reaction to evaluate
the performance of the two catalysts. The HUSY (an acidic,
ultra-stable, faujasite-type zeolite) zeolite exhibited a much
higher cracking rate when compared to H-ZSM5, 2370
versus 364 mmol n-hexane converted per g per min. To get a
periments reported in the liter-
Table 1. Characterisation of the acidity of zeolite materials.
ature,[7–10] we suggest that
Si/Al[a]
Quantity of extra
Total number of Brçnsted
Rate of n-hexane cracking
[D1]isobutane (m/z 59) is
[mmolgÀ1 min][d]
framework Al [%][b]
acid sites [mmolgÀ1
]
[c]
formed by hydride transfer be-
H-ZSM5
HUSY
12
6.2
–
40
1.17
2.26
364
2370
tween [D1]carbenium ion (m/z
58) and [D0]isobutane (m/z 58)
according to Scheme 1.
The catalytic transformation
of barely reactive alkane by mi-
[a] Determined by X-ray fluorescence; [b] Obtained from the difference of total Si/Al and framework Si/Al
(TO4 vibration by FT-IR); [c] H/D exchange technique reported in Ref. [21b,c] [d] Calculated at iso-conversion
between 5 and 8%. Experimental conditions are described in detail in Ref. [22].
Chem. Eur. J. 2010, 16, 573 – 576
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
575

J. Sommer et al.
experiment was repeated three times and a good agreement in each step
was observed.
croporous solid acids has been industrially implemented in
large-scale processes, such as catalytic cracking, reforming,
hydroisomerisation, oligomerisation and disproportionation.
Considering the economic importance of these reactions a
clear understanding of their activation mechanism is of pri-
mary interest. It may help researchers to design in a rational
manner the solid acid catalyst, through proper control of
Brçnsted acidity for instance, and thus increase the selectivi-
ty toward desired products.
Acknowledgements
The authors are grateful to Petrobras, CNPq and the Loker Hydrocarbon
Research Institute for financial support. The authors thank Prof. Octavio
Ceva Antunes for his support.
Conclusions
[1] G. A. Olah, ꢆ. Molnꢂr, Hydrocarbon Chemistry, 2nd ed., Wiley In-
terscience, New York, 2003.
[2] H. Pines, Chemistry of Catalytic Hydrocarbon Conversion, Academ-
ic Press, New York, 1981.
[3] A. Corma, Chem. Rev. 1995, 95, 559–614.
[4] G. A. Olah, A. Molnar, S. K. Prakash, J. Sommer, Superacid Chemis-
try, 2nd ed., Wiley, New York, 2009.
[5] M. Boronat, A. Corma, Appl. Catal. A 2008, 336, 2–10.
[6] B. Umansky, J. Engelhardt, W. K. Hall, J. Catal. 1991, 127, 128–140.
[7] S. G. Hindin, G. A. Mills, A. G. Oblad, J. Am. Chem. Soc. 1951, 73,
278–281.
[8] J. Engelhardt, W. K. Hall, J. Catal. 1995, 151, 1–9.
[9] J. Sommer, M. Hachoumy, F. Garin, D. Barthomeuf, J. Vedrine, J.
Am. Chem. Soc. 1995, 117, 1135–1136.
[10] B. Schoofs, J. Schuermans, R. A. Schoonheydt, Microporous Meso-
porous Mater. 2000, 35–36, 99–111.
[11] M. Haouas, S. Walspurger, J. Sommer, J. Catal. 2003, 215, 122–128.
[12] H. S. Bloch, H. Pines, L. Schmerling, J. Am. Chem. Soc. 1946, 68,
153.
The reaction of isobutane over acidic zeolites leads to the
formation of H2 as initial product. The proportionality rela-
tion between the amount of H2 produced and the amount of
catalyst indicates that H2 formation is directly related to the
number of Brçnsted acid sites involved in the activation
step. In the reaction of isobutane with deuterated zeolites,
HD and CH3D are the initial products. Thus the questions
“what is first: alkene or carbenium ion? or which is the first
step in alkane activation?” have now a definitive answer;
the carbenium ion is formed during the initial step of alkane
activation through protolysis of a s bond, similar to the pro-
cess that occurs for liquid superacids.
[13] C. Stefanadis, B. C. Gates, W. O. Haag, J. Mol. Catal. 1991, 67, 363–
367.
Experimental Section
[14] J. Huang, Y. Jiang, V. R. R. Marthala, M. Hunger, J. Am. Chem.
Soc. 2008, 130, 12642.
Isobutane cracking on H-ZSM5, protolysis experiments: The reaction
was run in an all glass grease-free recirculation flow system, as reported
earlier.[11] About 8.5 g of H-ZSM5 (AlSi-Penta Zeolite GmbH, Si/Al=
24, particle size: 20–30 mm, 1.17 mmol H+/g , Na2O <0.03 wt%) was acti-
vated at 823 K under N2 (40 mLminÀ1) for 1 h. The temperature was then
decreased to the reaction temperature (473 K) and a mixture of 60 mL of
gaseous isobutane (purified from traces of isobutene on Pt/Adams and
checked by GC: 99.99% purity), 190 mL of N2 was recirculated with a
[15] a) M. Marczewski, J. Chem. Soc. Faraday Trans. 1986, 82, 1687–
1701; b) T. F. Narbeshuber, A. Brait, K. Seshan, J. A. Lercher, J.
Catal. 1997, 172, 127—136; c) J. A. van Bokhoven, A. M. J. van der
Eerden, D. C. Koningsberger, J. Am. Chem. Soc. 2003, 125, 7435–
7442.
[16] W. O. Haag, R. M. Dessau, Proceedings of the 8th International
Congress on Catalysis, Berlin 1984, pp. 305–316.
peristaltic pump (Masterflex L/S) at
a
rate of 5 mLminÀ1. Samples
[17] J. Sommer, J. Bukala, Acc. Chem. Res. 1993, 26, 370–376.
[18] W. O. Haag, R. M. Lago, P. B. Weisz, Nature 1984, 309, 589–591.
[19] G. A Olah, Angew. Chem. Int. Ed. Engl. 1993, 32, 767–788.
[20] A. Goeppert, J. Sommer, New J. Chem. 2002, 26, 1335–1339.
[21] a) R. Olindo, A. Goeppert, D. Habermacher, J. Sommer, F. Pinna, J.
Catal. 2001, 197, 344–349; b) B. Louis, S. Walspurger, J. Sommer,
Catal. Lett. 2004, 93, 81–84; c) S. Walspurger, B. Louis, Appl. Catal.
A 2008, 336, 109–115.
[22] F. Ocampo, H. S. Yun, M. M. Pereira, J. P. Tessonnier, B. Louis,
Cryst. Growth Des. 2009, 9, 3721–3729.
[23] C. M. Zicovich-Wilson, A. Corma, P. Viruela, J. Phys. Chem. 1994,
98, 10863–10870.
(1 mL) were withdrawn at regular time intervals and analyzed by using a
GC during the initial 150 min on stream. Hydrocarbons were analyzed by
using a Hewlett Packard 5840 Series II chromatograph, using a HP-5 ca-
pillary column (crosslinked 5% diphenyl and 95% dimethylpolysiloxane,
non polar, 30 mꢄ0.32 mmx0.25 mm film thickness) and FID detector.
Hydrogen was analyzed by means of an Intersmat IGC 112m gas-chro-
matograph equipped by a 5 ꢅ molecular sieves column and TCD detec-
tor (after external calibration).
Isobutane cracking on D-USY zeolite, deuterolysis experiments: About
8.2 g Æ0.2 g of the protic form (no sodium was detected by X-Ray fluo-
rescence) of ultra-stable Y-type zeolite (HUSY) catalyst, with SiO2/Al2O3
ratio=12.4 (16.4, corrected for extra framework Al), and BET area=
620 m2g,À1 was used in all experiments. The hardware unit consists of a
saturator, followed by a fixed-bed reactor, connected on-line to a quadru-
polar mass spectrometer (MKS which is able to measure m/z from 1 to
100). The hardware is described in the supplementary information. This
[24] P. M. Esteves, B. Louis, J. Phys. Chem. B 2006, 110, 16793–16800.
[25] P. Kuhn, P. Pale, J. Sommer, B. Louis, J. Phys. Chem. C 2009, 113,
2903–2910.
Received: June 23, 2009
Published online: November 13, 2009
576
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 573 – 576