Catalytic dehydroisomerization of n-alkanes to isoalkenes

Article abstract of DOI:10.1016/j.jcat.2008.01.021

An equilibrated mixture of pentene isomers was produced by dehydroisomerization of n-pentane on catalysts consisting of Pt clusters within [Fe]ZSM-5 channels. These catalysts showed high isomerization rates, excellent stability even without added H_2<

Full text of DOI:10.1016/j.jcat.2008.01.021

Journal of Catalysis 255 (2008) 134–137
www.elsevier.com/locate/jcat
Research Note
Catalytic dehydroisomerization of n-alkanes to isoalkenes
Xuebing Li, Enrique Iglesia ∗
Department of Chemical Engineering, University of California, Berkeley, CA 94720-1462, USA
Received 18 December 2007; revised 17 January 2008; accepted 18 January 2008
Available online 20 February 2008
Abstract
An equilibrated mixture of pentene isomers was produced by dehydroisomerization of n-pentane on catalysts consisting of Pt clusters within
[Fe]ZSM-5 channels. These catalysts showed high isomerization rates, excellent stability even without added H , and isopentene selectivities
2
above 60%. Metal sites on Pt clusters dehydrogenate n-alkanes and n-alkenes formed undergo skeletal rearrangements with high selectivity on
weak acid sites prevalent in Na-[Fe]ZSM-5 after reduction of exchanged Pt cations. Zeolite channels inhibit side reactions and sintering processes
that cause rapid deactivation on unprotected Pt clusters.
© 2008 Elsevier Inc. All rights reserved.
Keywords: n-Pentane; Dehydroisomerization; ZSM-5; Pt; Pentene
1. Introduction
ters in alkane dehydrogenation to catalyze the direct dehydro-
isomerization of n-pentane to equilibrated mixtures of pentene
isomers. These catalysts exhibit excellent reactivity, selectivity,
and stability even in the absence of H2 in the reactant stream.
Alkenes are useful intermediates in the synthesis of many
chemicals. One of these, methyl-tert-butyl ether, produced via
isobutene-methanol reactions and useful as a fuel additive, has
raised concerns about its solubility and permanence within
aquifers [1]. tert-Amyl methyl and larger ethers are less sol-
uble and degrade more rapidly; as a result, they have emerged
as attractive alternatives. The synthesis of these molecules re-
quires higher isoalkenes, which can be produced via catalytic
dehydroisomerization of n-alkanes. The selective direct dehy-
droisomerization of n-butane to isobutene has been reported
[2–9], but corresponding reactions of larger n-alkanes occur
concurrently with side reactions and remain largely unexplored
[10,11].
We have shown previously that Pt/Na-[Fe]ZSM-5 catalyzes
dehydrogenation of C2–C4 alkanes with unprecedented reactiv-
ity and stability [12,13]. The synthesis of n-alkenes from larger
n-alkanes required the rigorous titration of all residual Brønsted
acid sites in the catalysts using Cs [14]. These acid sites, how-
ever, catalyze alkene isomerization via monomolecular path-
ways [15]. We exploit here these isomerization pathways and
the remarkable reactivity and stability of encapsulated Pt clus-
2. Experimental
The synthesis of Pt/Na-[Fe]ZSM-5 has been previously re-
ported [12–14]. Na-[Fe]ZSM-5 precursors were prepared by
exchanging NH4-[Fe]ZSM-5, synthesized by known meth-
ods [15], three times with 0.1 M NaNO3 solutions (EMD
Chemicals, >99%) at 353 K for 15 h. Pt cations were then
grafted by exchange with [(NH3)4Pt](NO3)2 solutions (6.25 ×
10−5 M; Aldrich, 99.995%) for 12 h at 353 K, and subsequent
filtration, rinsing with deionized water, and treatment in flow-
ing dry air (Praxair, extra dry, 1.67 cm3 s−1) for 12 h at 723 K
(0.017 K s−1). [Fe]ZSM-5 crystals were ∼0.5 µm in diame-
ter (transmission electron microscopy). This sample contained
0.12 wt% Pt and 0.28 wt% Fe with a Si/Fe atomic ratio of 340
(from inductively-coupled plasma emission spectra). Al was
present in trace amounts (0.06 wt%).
The Pt dispersion (0.58) was measured from volumetric up-
takes of strongly chemisorbed H2 (4–40 kPa) at 298 K us-
ing a Quantasorb chemisorption analyzer (Quantachrome Corp.
Autosorb-1), after treating samples (1 g, 180–250 µm diame-
ter) in He (0.5 cm3 s−1) at 573 K for 1 h, then at 573 K in
*
Corresponding author.
E-mail address: iglesia@berkeley.edu (E. Iglesia).
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.01.021

X. Li, E. Iglesia / Journal of Catalysis 255 (2008) 134–137
135
H2 (0.5 cm3 s−1) for 2 h, and finally in dynamic vacuum at
573 K for 1 h. Strongly chemisorbed hydrogen by subtracting
from total uptakes those measured after evacuation at 298 K,
which correspond to weakly-held hydrogen, after extrapolating
both adsorption and backsorption isotherms to zero H2 pres-
sure. A H:Pt adsorption stoichiometry was used to calculate the
number of Pt surface atoms.
actions are also inhibited on small Pt clusters because of their
requirement for large Pt ensembles [17].
First-order deactivation constants on Pt/Na-[Fe]ZSM-5
(0.004 h−1) are similar to those observed for C2–C4 alkane de-
hydrogenation on these catalysts [12,13] and ∼10 times smaller
than for n-pentane dehydroisomerization on Pt/SAPO-11 even
though H2 was present to stabilize the catalysts in the latter
study [10] (Table 1, Fig. 1). We conclude that encapsulation of
small Pt clusters within constrained environments in [Fe]ZSM-
5 channels also inhibits chain growth reactions that form un-
saturated organic residues responsible for deactivation when Pt
clusters are present within less protected environments [18].
n-Pentane reaction rates and selectivities (25 kPa, Aldrich,
99%) were measured at 673 K and ambient pressure in two
types of reactors; one was a tubular reactor with plug-flow hy-
drodynamics and the other a gradientless recirculating batch
reactor (206 cm3 volume). Both reactors were certified to be
free of mass or heat transfer corruptions. Catalysts (0.1 g in flow
and 0.001 g in batch reactors, 180–250 µm aggregates) were
treated for 2 h in 40% H2 (Praxair, 99.999%, 0.20 cm3 s−1
)
in He (Praxair, 99.999%, 0.30 cm3 s−1) at 673 K. n-Pentane
(25 kPa) was added to He flow with a syringe pump (Tele-
dyne Isco, Model 500 D). Reactants and products were sep-
arated by gas chromatography (HP 5890 II; HP-1 capillary
column; 50 m × 0.32 mm) and detected by flame ionization.
The approach to equilibrium for each pentene isomer (ηi) from
n-pentane was calculated from measured rates and thermody-
namic data [16] and used to correct measured net rates and
determine forward dehydroisomerization rates [14].
3. Results and discussion
Fig. 1 shows n-pentane dehydroisomerization rates (per Pt
atom) and selectivities (on a carbon basis) on Pt/Na-[Fe]ZSM-
5 as a function of time-on-stream (673 K, 25 kPa n-pentane;
flow reactor, 140 h). Forward pentene formation rates decreased
from 5.7 mol (g-atom Pt)−1 s−1 (25% n-pentane conversion)
after 0.8 h on stream to 1.5 mol (g-atom Pt)−1 s−1 (11%
n-pentane conversion) after 140 h (Fig. 1a). Measured pentene
formation rates are compared per Pt atom, because dispersion
data were unavailable in the comparative previous report [10].
These rates are ∼30 times greater on Pt/Na-[Fe]ZSM-5 than on
Pt/SAPO-11 (0.5 wt% Pt) [10] at similar values (0.35) of the ap-
proach to equilibrium for pentenes from n-pentane even though
the latter rates were measured at higher temperature (773 vs.
673 K) [10] (Table 1). The higher rates on Pt/Na-[Fe]ZSM-5
apparently because of the higher dispersion and resistance to
sintering of encapsulated Pt clusters prepared by grafting of Pt
precursors onto Na-[Fe]ZSM-5. Undesired hydrogenolysis re-
Fig. 1. (a) Pentene formation rates during n-pentane dehydroisomerization
on (") Pt/Na-[Fe]ZSM-5 at 673 K (n-pentane partial pressure: 25 kPa; H
−1 −1
) and (F)
2
partial pressure: 0 kPa; space velocity: 12 mol (g-atom Pt)
s
Pt/SAPO-11 at 773 K (n-pentane partial pressure: 26 kPa; H partial pressure:
2
−1 −1
44 kPa; space velocity: 0.24 mol (g-atom Pt)
s
) [10]. (b) Selectivities to
(Q) pentenes, (F) cracking products and (2) oligomerization products during
n-pentane dehydrogenation on Pt/Na-[Fe]ZSM-5 at 673 K in the absence of H .
2
Table 1
Catalytic performance of Pt-based catalysts for n-pentane dehydroisomerization
Catalyst
Reaction conditions
n-Pentane
conversion
(%)
Measured
pentene
synthesis rate
Pentene
selectivity
(%)
First-order
deactivation rate
constant (h
Reactants
(kPa)
T
a
−1
)
(K)
Pt/Na-[Fe]ZSM-5
(this study)
Pt/SAPO-11 [10]
n-Pentane (25)
673
16
1.9
95
0.004
H
(0)
2
b
b
b
n-Pentane (26)
(44)
773
24
0.06
94
0.04
H
2
a
−1 −1
s .
mol (g-atom Pt)
b
Obtained from Fig. 5, Ref. [10].

136
X. Li, E. Iglesia / Journal of Catalysis 255 (2008) 134–137
Table 2
a
Double-bond and methyl-shift isomerization. Approach to equilibrium from n-pentane during n-pentane dehydroisomerization on Pt/Na-[Fe]ZSM-5 at 673 K
b
Time on
stream
(h)
η
i
n-Pentenes
Isopentenes
1-Pentene
trans-2-
Pentene
cis-2-
Pentene
3-Methyl-
1-butene
2-Methyl-
1-butene
2-Methyl-
2-butene
0.8
140
0.47
0.15
0.45
0.14
0.39
0.12
0.39
0.09
0.38
0.08
0.42
0.09
a
−1 −1
s .
Reaction conditions: n-pentane, 25 kPa; space velocity: 12 mol (g-atom Pt)
Approach to equilibrium from n-pentane:
b
[P
][P
]
1
H
pentene(i)
[P
2
η =
i
×
.
]
K
eq,i
n-pentane
Fig. 2. (a) Selectivities of (Q) pentenes, (F) cracking products and (2) oligomerization products during n-pentane dehydrogenation on Pt/Na-[Fe]ZSM-5 at 673 K
in RRU unit in the absence of H . (b) Ratio of (η /η ) as a function of n-pentane conversion: (Q) trans-2-pentene, (") cis-2-pentene, (!) 3-methyl-1-butene,
2
i
1-pentene
(P) 2-methyl-1-butene and (1) 2-methyl-2-butene. n-Pentane partial pressure: 25 kPa.
Scheme 1. Possible pathways for n-pentane dehydroisomerization on Pt/Na-[Fe]ZSM-5.

X. Li, E. Iglesia / Journal of Catalysis 255 (2008) 134–137
137
The selectivity to pentenes increased from 85 to 96% with in-
creasing time-on-stream (Fig. 1b), while selectivities to smaller
molecules, formed by cracking or hydrogenolysis side reac-
tions, and to larger hydrocarbons, formed by oligomeriza-
tion/dehydrogenation pathways, concurrently decreased. These
selectivity trends predominantly reflect weaker contributions
from secondary reactions as n-pentane conversion decreased
with time, but perhaps also the selective deactivation of stronger
acid sites, which favor cracking and oligomerization of primary
alkene products.
dehydrogenation of n-pentane at 673 K on Pt clusters followed
by selective skeletal rearrangement of n-pentenes on acid sites
present on [Fe]ZSM-5. Alkene selectivities above 95% with ex-
cellent stability (deactivation rate constants: 0.004 h−1), even
in the absence of co-fed H2, were achieved. Reaction rates
are more than ten-fold higher than the best values previously
reported and allow the use of catalysts with much lower Pt con-
tent. These catalysts and insights provide potentially practical
routes for the selective synthesis of molecules useful as precur-
sors to fuel additives and chemical intermediates.
Pentene isomers are essentially in thermodynamic equi-
librium with each other during n-pentane dehydrogenation
on Pt/Na-[Fe]ZSM-5. The approach to equilibrium values (η)
are similar for all isomers during the early stages of reac-
tion (0.8 h time-on-stream; Table 2). Thus, skeletal isomeriza-
tion of pentenes to form branched isomers (3-methyl-1-butene,
2-methyl-1-butene and 2-methyl-2-butene) is fast on acid sites
formed in Na-[Fe]ZSM-5 during reduction of grafted cationic
Pt precursors. At longer times (140 h), n-pentenes (1-pentene,
trans-2-pentene and cis-2-pentene) remain equilibrated with
each other via fast hydride shifts, but η values for isopentenes
become smaller than for n-pentenes, apparently because slower
methyl shift reactions cannot reach equilibrium as the number
of acid sites decreases with time-on-stream (Table 2). Never-
theless, isopentenes represent more than 60% of all pentene
products even after 140 h on Pt/Na-[Fe]ZSM-5 catalysts.
Primary and sequential pathways in dehydroisomerization
reactions catalyzed by Pt/Na-[Fe]ZSM-5 were probed by mea-
suring changes in selectivity with contact time in a recircu-
lating batch reactor (Fig. 2). Pentene isomers are almost ex-
clusively formed at low conversions (Fig. 2a), indicating that
cracking, oligomerization, and aromatization reactions require
secondary reactions of primary pentene products (Scheme 1).
All n-pentene isomers remain at equilibrium with each other
at all contact times, as their constant ηi/η1-pentene ratios indi-
cate; this indicates that hydride shifts are fast on acid sites.
Isopentene ηi/η1-pentene ratios were smaller than for n-pentenes
and increased with contact times, because isopentenes form via
slower skeletal isomerization of n-pentenes on Brønsted acid
sites (Fig. 2b, Scheme 1).
Acknowledgments
The authors acknowledge ExxonMobil Research and En-
gineering Co. for financial support and Dr. James C. Vartulli
(ExxonMobil) for the synthesis of the [Fe]ZSM-5 sample used
in this study.
References
[1] P.M. Morse, Chem. Eng. News 77 (1999) 26.
[2] G.D. Pirngruber, K. Seshan, J.A. Lercher, J. Catal. 186 (1999) 188.
[3] G.D. Pirngruber, K. Seshan, J.A. Lercher, J. Catal. 190 (2000) 338.
[4] G.D. Pirngruber, O.P.E. Zinck-Stagno, K. Seshan, J.A. Lercher, J. Ca-
tal. 190 (2000) 374.
[5] G.D. Pirngruber, K. Seshan, J.A. Lercher, J. Catal. 190 (2000) 396.
[6] M.R. Sad, C.A. Querini, R.A. Comelli, N.S. Fígoli, J.M. Parera, Appl.
Catal. A 146 (1996) 131.
[7] S.B. Derouane-Abd Hamid, D. Lambert, E.G. Derouane, Catal. Today 63
(2000) 237.
[8] A. Vieira, M.A. Tovar, C. Pfaff, B. Méndez, C.M. López, F.J. Machado,
J. Goldwasser, M.M. Ramírez de Agudelo, J. Catal. 177 (1998) 60.
[9] Y. Takiyama, H. Nagata, M. Kishida, K. Wakabayashi, Sekiyu Gakkai-
shi 41 (1998) 80.
[10] C.M. López, M. De Sousa, Y. Campos, L. Hernández, L. García, Appl.
Catal. A 258 (2004) 195.
[11] G.P. Herrera, D.G. Lardizábal, V.H.C. Martínez, A.A. Elguézabal, Catal.
Lett. 76 (2001) 161.
[12] T. Waku, J.A. Biscardi, E. Iglesia, Chem. Commun. (2003) 1764.
[13] T. Waku, J.A. Biscardi, E. Iglesia, J. Catal. 222 (2004) 481.
[14] X. Li, E. Iglesia, Chem. Commun. (2008) 594.
[15] J. Houžvicˇka, J.G. Nienhuis, S. Hansildaar, V. Ponec, Appl. Catal. A 165
(1997) 443.
[16] C.L. Yaws, Yaws’ Handbook of Thermodynamic and Physical Properties
of Chemical Compounds, Gulf, Houston, 2006.
4. Conclusions
[17] J.H. Sinfelt, Catal. Rev. 3 (1969) 175.
[18] F.H. Ribeiro, A.L. Bonivardi, C. Kim, G.A. Somorjai, J. Catal. 150 (1994)
186.
These data show that isopentene isomers (3-methyl-1-bu-
tene, 2-methyl-1-butene, 2-methyl-2-butene) can be formed via