Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

~ APPLIED

CATALYSSI
AG
: ENERAL
ELSEVIER Applied Catalysis A: General 137 (1996) 167-177

Alkane isomerization over solid acid catalysts


Effects of one-dimensional micropores
H. Liu, G.D. Lei, W . M . H . Sachtler *
V.N. lpatieff Laboratory, Center fi~r Catalysis and Surface Science, Department of Chemistry, Northwestern
University, Evanston, IL 60208, USA
Received 25 September 1995; revised 20 November 1995; accepted 20 November 1995

Abstract

Isomerization of small alkanes is compared over two bifunctional catalysts with similar
concentrations of metal and acid sites and similar strengths of Brcnsted acid sites. The catalytic
performance and the kinetic parameters are significantly different for P t / H - M o r and P t / Z r O 2 - S O a
catalysts. Under conditions where n-butane isomerization is rapid over P t / Z r O z - S O 4 , it is
undetectable over P t / H - M o r . The apparent activation energy of n-pentane isomerization is
significantly higher over the zeolite based catalyst. These differences can be rationalized in terms
of different pore geometries: in the one-dimensional pore system of mordenite the rates of
monomolecular reactions are single file diffusion controlled; as a result, bimolecular processes
such as isomerization via intermediate dimer formation are all but impossible. Likewise, hydride
ion transfer, which is the chain propagation step in alkane reactions via carbenium ions, is
sterically inhibited in micropores and replaced by a process involving Pt-proton adducts.

Keywords: zirconia (sulphated); Platinum promotion; Mordenite; Micropores; Kinetic parameters; Single file
diffusion; Alkane isomerization; Bifunctional catalysis; Collapsed bifunctional sites; Platinum-proton adducts

1. Introduction

Since the invention of 'catalytic reforming' and 'platforming', isomerization


of n-alkanes to branched alkanes has become a large-scale industrial process.
Lower reaction temperatures yield a greater proportion of branched alkanes at
thermodynamic equilibrium. Chlorided platinum/alumina catalysts, which can
be considered to be superacids, are used at low temperatures (400-460 K).
However, these catalysts are very sensitive to poisons such as water or aromat-

" Corresponding author. Tel. ( + 1-708) 4915263, fax. ( + 1-708) 4671018, e-mail
wmhs@casbah.acns.nwu.edu.

0926-860X/96/$15.00 © 1996 Elsevier Science B.V. All rights reserved


SSDI 0 9 2 6 - 8 6 0 X ( 9 5 ) 0 0 3 2 1 - 5
168 H. Liu et al./Applied Catalysis A: General 137 (1996) 167-177

ics. The environmental problems with the regeneration of theses catalysts have
also revived interest in halogen-free isomerization catalysts. Bifunctional cata-
lysts, in particular Pt/H-Mordenite, are halogen-free alternatives currently
employed though they are less active, requiring a temperature above 520 K.
The mechanism for skeletal isomerization of small alkanes has been studied
extensively in liquid superacids and is found to proceed via a protonated
cyclopropane intermediate [1,2]. For C5+ hydrocarbons, this path avoids the
energetically prohibitive stage of a primary carbenium ion. In the case of butane,
however, opening of the protonated methylcyclopropane ring to a carbocation
with i s o - C 4 s t r u c t u r e inevitably leads to a primary carbenium ion. Therefore,
skeletal isomerization of butane was reported to be negligible under conditions
where C5+ hydrocarbons swiftly isomerized in strong liquid acids. Likewise,
over the solid catalyst Pt/A1203-SiO 2 the isomerization rate of n-butane at 573
K was found to be much slower than that of n-pentane [3].
Hsu et al. reported that ZrO2-SO4-based catalysts, and in particular Fe,
M n - Z r O 2 - S O 4, were superactive for n-butane isomerization [4]. Adeeva et al.
showed that the acid sites of these catalysts are not stronger than the Br~nsted
sites in acid zeolites or the Lewis sites in y-alumina [5]. By using isotopically
labelled n-butane and kinetic analysis it was further shown that over unpromoted
ZrO2-SO4, Fe, Mn promoted ZrO2-SO 4, and P t / Z r O 2 - S O 4 (Pt-SZ) catalysts
the isomerization of n-butane to i-butane is a bimolecular process, involving
intermediate formation of butene and an adsorbed C 8 entity [6,7]. In contrast,
the isomerization of n-pentane over the same catalyst appears to mainly use a
monomolecular pathway with only a small contribution of a parallel bimolecular
mechanism [8]. These findings reveal a strong similarity of the basic chemistry
that determines acid catalyzed alkane reactions in liquid acids and over solid
acid catalysts. An unsolved problem is, however, the striking difference in
catalytic performance between the bifunctional solid catalyst Pt-SZ and zeolite
based catalysts, such as Pt/H-mordenite (PtHMor) with roughly the same
numbers of acid and exposed Pt sites and similar acid strength of the Brc~nsted
acid sites.
In the present study, the catalytic isomerization of n-pentane and n-butane
over Pt/H-Mordenite (PtHMor) is studied under the same conditions and in the
same reactor that was used for similar work with Pt-SZ, with the objective to
identify and rationalize the differences in the kinetics and the reaction mecha-
nism over these solid acid catalysts.

2. Experimental

The preparation of the Pt-SZ catalysts was as previously reported [7]. PtHMor
samples were prepared by ion exchange of ammonium mordenite (LZ-M-8,
UOP, lot# 13744-24, Si/A1 = 8) with aqueous solution of Pt(NH3)4CI e. ICP
H. Liu et a l . / Applied Catalysis A: General 137 (1996) 167-177 169

analysis showed that the Pt-SZ catalyst contained 0.38 wt.-% Pt and 1.8 wt.-%
S. The PtHMor catalyst contained 0.45 wt.-% Pt. The catalytic data were
collected in a microflow reactor at 11 bar total pressure as described in Ref. [7].
The partial pressures of H 2 and the alkanes were adjusted by controlling the
flow rates of the feed components. In one series of experiments over Pt-SZ, the
partial pressure of one feed component was varied at a constant overall flow rate
while the partial pressures of the other reactants were kept constant by adapting
the helium make-up. The reaction orders in hydrogen and n-pentane or n-butane
were then determined from the resultant changes in the conversion and produc-
tion rates. The effluent was analyzed by on-line gas chromatography with a 50
m × 0.2 mm capillary column of PONA (Hewlett-Packard) attached to a FID
detector. The production rates of hydrocarbons were normalized to mole num-
bers of n-pentane or n-butane per gram catalyst per second (mol (gcat)-! S-1).
The reactant and products of methane, ethane, propane, butane, i-pentane,
n-pentane, and hexane were assigned as C l, C 2, C 3, C 4, i-Cs, n-C 5, and C 6,
respectively.

3. Results

3.1. Catalytic tests of n-pentane and n-butane isomerization over P t / H -


mordenite
Fig. 1 shows the reaction rates for the products of n-pentane isomerization
over PtHMor at 479 K. At a hydrogen pressure of 2.2 bar, activity and
selectivity remain constant with time on stream for the duration of the 15 h test.
The i-pentane selectivity is 97%. No hexanes were observed.
For n-butane reaction over the PtHMor catalyst, no activity was observed
below 500 K; conversion and selectivities at 544 K are shown in Fig. 2. The
selectivity to i-butane is low, viz. 9.5%; the predominant products are C 2 and
C 3. Only a very small amount of pentane was observed. This behavior is
strongly in contrast to that of Pt-SZ where n-butane conversion at 479 K is
3.45%, with a selectivity to i-butane of 94.8%.
For n-pentane isomerization over PtHMor and Pt-SZ the relevant parameters
are compiled in Table 1. The isomerization rate of n-pentane is considerable
both over Pt-SZ and PtHMor, with the latter catalyst being more active. The
cracking rate is roughly five times higher over Pt-SZ than over PtHMor. No C 6
product was observed over the PtHMor catalyst.

3.2. Temperature effects


An Arrhenius plot for n-pentane isomerization over PtHMor over a wide
range of reaction temperatures is found to significantly deviate from linearity.
Over Pt-SZ, however, perfectly linear Arrhenius plots are obtained, both for
n-pentane and for n-butane; the corresponding activation energies are much
170 H. Liu et al. / Applied Catalysis A: General 137 (1996) 167-177

n-Cs c o n v e r s i o n
i. . . . . . . . . . . . . . . . . . . . . . . . . . • - - . . . . . . . ~'''''l''vrv'" t~ * 1. . . . . . . r'l*"l*r . . . . . . . . . . . . . . . . . . . . . . .

i-C~ f o r m a t i o n

lO -7

2
10-8 C 1"-C2+C3 f o r m a t i o n
J
----~---2~22-2_2--- ~_'''_ -~222:22222- -- -~---

C~ f o r m a t i o n
| ! t ' ! : I ' I , I , I ! I ~ I ' I ~ | !
10-9

0 I 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Time on Stream hr

Fig. 1. n-Pentane conversion with time on stream over PtHMor. Total flow rate = 10 cm 3 g - J s -~ . Total
pressure = 11 bar: p(n-C 5) = 0.054 bar, p(H 2) = 2.2 bar, He balance. T = 479 K.

lower, viz. 92 and 72 kJ/mol, respectively. The values of the kinetic parameters
are compiled in Table 2.
It is possible to describe the Arrhenius data for n-pentane isomerization over

50
"
C.o sel.
i A A & & Ik & & & A A A & & & & ik & A A & A & &

40
©

i C~ sel.
- + + + - + + + + + + -= - + + + + + -t- + + + +
>
30

7- n-C 4 conv.

20

C~ sel.
10
I i-C 4 sel.
C 5 sel.
t-
O ~ ' v'v v v ~ ' v v v v v v v ' r ' l r YT T Y T T ' I r YY !

0 i 2 3 4 5 6 7 8 9 I0 1 1 12 13

T i m e o n S t r e a m / hr
Fig. 2. n-Butane conversion with time on stream over PtHMor. Total flow rate = 10 cm 3 g - ~ s - ~. Total
pressure = 11 bar: p(n-C s) = 0.054 bar, p(H 2) = 2.2 bar, He balance. T = 544 K.
H. Liu et al./Applied Catalysis A: General 137 (1996) 167-177 17t

Table 1
Reaction o f n-pentane and n - b u t a n e over solid acid catalysts at 4 7 9 K
Catalyst PtHMor Pt-SZ
Reactant n-Pentane n-Pentane n-Butane

n-Alkane pressure/bar 0.072 0.072 0.076


Hydrogen Pressure/bar 2.23 2.23 2.22
n - A l k a n e flow r a t e / 1 0 - 7 mol ( g c a t ) - 1 s - I 27.0 27.0 28.8
H 2/n-alkane 31.0 31.0 29.3

n-Alkane conversion/% 16.9 24.8 3.45


Total c o n v e r s i o n r a t e / 1 0 - 7 mol ( g c a t ) - t s - l 3.36 6.69 0.995
Isomerization r a t e / 1 0 - 7 mol ( g c a t ) - 1 s - 1 3.26 6.13 0.944
C r a c k i n g r a t e / 1 0 - 7 mol ( g c a t ) - 1 s - la 0.10 0.56 0.051
( C r a c k i n g / i s o m e r i z a t i o n ) ratio 0.03 0.08 0.05
(C 6 / C 4) ratio 0 0.96 -
(C 5 / C 3) ratio - - 0.12

Selectivity / %
CH 4 +C2H 6 - 3.04
CH 4 +C2H 6 +Call 8 2.28 1.76 -
C3H 8 - - 1.78
i-C4Hj0 - - 94.8
i-C4Hlo +n-C4Hi0 0.85 2.71 -
i ' C s H l0 96.9 91.6 0.22
n-CsHio - - 0.13
Hexanes 0 3.9 -

P(total) = 11 bar.
a C r a c k i n g rate = total c o n v e r s i o n rate - isomerization rate.

PtHMor with two linear segments which intersect near 487 K, as shown in Fig.
3. The apparent activation energies thus determined are 111.8 kJ mol-1 in the
range of 4 6 2 - 4 8 7 K and 54.7 kJ m o l - I in the range of 4 9 6 - 5 2 3 K. This
behavior possibly indicates single file diffusion in the one-dimensional pores of
mordenite.
3.3. Reaction orders in hydrogen and n-alkane
The variations in the rates of n-pentane conversion and product formation
were measured while changing the hydrogen pressure in the range of 2.2-6.1

Table 2
Kinetic p a r a m e t e r s o f n-alkane isomerization over solid acid catalysts
Alkane n-Butane n-Pentane

Catalyst Pt-SZ Pt-SZ PtHMor

Parameter
nil2 -- 1.19 --0.15 --0.89
nalkane h- 1.58 + 1.06 + 0.53
Eapparent/kJ mol - i 92 72 145
Temp. r e g i o n / k 479-503 443-479 450-479

The reaction conditions are the s a m e as in Fig. 4.


172 H. Liu et al. / Applied Catalysis A: General 137 (1996) 167-177

-12

-13

-14 \
\
m

-15

-16 i J

1.90 2.00 2.10 2.20

1000 / T

Fig. 3. n-Pentane conversion over PtHMor as a function of temperature. Total flow rate = 20 cm 3 g - ' s - ~.
Total pressure = 15 bar: p(n-C 5) = 0.15 bar, p(H 2) = 1.5 bar, He balance. T = 462-523 K.

bar at a constant n-pentane pressure and while changing the n-pentane pressure
in the range of 0.036-0.073 bar at a constant P(H2). The reaction temperature
for all these data is constant, viz. 479 K. The results for the PtHMor catalyst are
depicted in Figs. 4 and 5; (results for the Pt-SZ catalyst were published
previously [7].) The reaction orders are listed in Table 2.

-14.50
~ conversion
-15.50

-16.50
Z"
-17.50
=
O
-18.50

-19.50

-20.50 C4 f o r m a t i o n

-21.50

l ~ E f r I
-22.50
0.50 1.00 1.50 2.00

In (p(H2))
Fig. 4. n-Penmne reaction rate over PtHMor as a function of hydrogen pressure. Total flow rate = 10 cm 3 g -
s - ~. Total pressure = 11 bar: p(n-C 5) = 0.054 bar, p(H 2) = 2.2-6.1 bar, He balance. T = 479 K.
H. Liu et al. / Applied Catalysis A: General 137 (1996) 167-177 173

-14.50
n-C 5 conversion
• w
-15.00

-15.50 i-C 5 formation

-16.00

-16.50
e-.
o -17.00

-17.50
2
e-
-18.00
C 1+C2 +C3 formation
-18.50

-19.00
C Af o r m a t i o n
-19.50
I I I ', 1
-20.00
-3.50 -3.00 -2.50

In ( p ( n - C s ) )

Fig. 5. n-Pentane reaction rate over PtHMor as a function of n-pentane pressure. Total flow rate = 10 c m 3 g - '
s - i. Total pressure = 11 bar: p ( n - C 5) = 0 . 0 3 6 - 0 . 0 7 3 bar, p ( H 2) = 2.2 bar, He balance. T = 479 K.

All lines in the In rate vs. In p(H 2) graph, taken at a constant n-pentane
pressure of 0.054 bar, are linear and have negative slopes. The reaction orders in
hydrogen of n-pentane conversion and the formations of i-pentane, butane, and
(C l + C 2 + C 3) are - 0 . 8 9 , - 0 . 8 8 , - 2 . 3 , and - 1.2, respectively. The orders
in hydrogen are generally more negative at higher H2/n-C 5 ratios.
The rates of n-pentane conversion and product formation increase with an
increase in the n-pentane pressure. All lines in the In rate vs. In p(n-C 5) graph
are linear and have positive slopes, except for the formation of butane, which is
independent of the n-pentane pressure within experimental error. The rate of
n-pentane conversion has an order in n-pentane of + 0.53; the formation rates of
i-pentane, butane, and (C~ + C 2 + C3), have orders in n-pentane of +0.54, 0,
and + 0.23, respectively. The orders in n-pentane change little with the H J n - C 5
ratio; only the order in butane formation becomes - 0 . 6 1 at high H2/n-C 5.

4. Discussion

4.1. Single-file diffusion over Pt / H-mordenite

The most striking result of this work is the huge difference of these catalysts
in the isomerization of butane. The Pt-SZ catalyst is very active, PtHMor is
virtually inactive under the same conditions. In contrast, the isomerization rates
of n-pentane over these catalysts are quite comparable; the conversions differ by
174 H. Liu et al./Applied Catalysis A: General 137 (1996) 167-177

less than a factor of two. As the isomerization of n-butane includes a bimolecu-


lar step, whereas n-pentane isomerizes via a monomolecular mechanism, it
appears that a decisive difference between the two catalysts is their different
propensity to catalyze b/molecular reactions. Both catalysts contain comparable
numbers of acid and metal sites; also the strength of the acid sites is comparable
[5]. T h e results, therefore, suggest that the different pore geometry is a key
factor. Mordenite has a system of one-dimensional micropores, whereas SZ has
a wide distribution of pore sizes. Under conditions of pressure and temperature
where physisorption is significant, single file diffusion is expected for PtHMor,
but not for Pt-SZ.
For a mathematical analysis of the effects of single file diffusion on the rate
of catalytic reactions we refer to papers by K~irger et al. [9,10]. Their recently
published model calculations also include the effect of this diffusion type on the
apparent activation energy [10]. The effect of single-file diffusion on the kinetic
parameters of catalytic processes is strikingly different from that of ordinary
diffusion control, as described by the Thiele modulus and by Wheeler's formula.
At low temperatures, when physisorption in such pores is high, only the few
sites near the pore mouths contribute effectively to the observed reaction rate. At
higher temperatures, when physisorption is less important, the much more
numerous sites in the interior of the pores fully contribute to the reaction rate
registered outside the pores. Consequently, there is a temperature regime in
which the observed reaction rate increases more steeply with temperature than it
would if the number of contributing sites were constant. In other words, the
apparent activation energy is higher than the intrinsic activation energy of the
chemical process. In previous work we found this to be true of neopentane
conversion on P d / K L catalysts [11]. In this case, the one-dimensional micro-
pores of the L zeolite led to single-file diffusion, with an apparent activation
energy substantially higher than the intrinsic activation energy [ 11 ]. In contrast,
the same reaction over P d / N a Y with a three-dimensional pore system displayed
a lower than intrinsic activation energy, in agreement with the classical pore
diffusion model. Related work on Pt/mordenite has shown that the transforma-
tion, by local destruction of the zeolite framework, of a one-dimensional pore
system into a three-dimensional pore system results in a marked increase of the
observed reaction rate [12,13].
Physisorption of alkanes on zeolites, including H-mordenite has been mea-
sured by Eberly [ 14]. This author reports for n-hexane in H-mordenite a value of
- 65.7 k J / m o l for the enthalpy of adsorption. Carvill calculated from Eberly's
data that at 240°C and 40 Torr partial pressure of n-hexane, 385 molecules are
adsorbed p e r / x m of channel length [ 15]. In the present work n-pentane has been
used, with a slightly lower heat of adsorption, at a higher pressure (55 Torr) and
a lower temperature (typically 205°C) than in the example used by Carvill. We
estimate that crowding of physisorbed molecules in the channels will be slightly
higher, i.e. the average distance between physisorbed molecules in H-mordenite
H. Liu et al./Applied Catalysis A: General 137 (1996) 167-177 175

will be roughly 2.5 nm. This matches the conditions for which single file
diffusion applies.
The opening of the main channel in mordenite is 6.5 × 7.0 ,~2 wide, whereas
the minimum kinetic diameters for n-butane and i-butane are 4.3 and 5.0
respectively [16]. In such one-dimensional pores two-way traffic is difficult or
impossible; transport of molecules through the channels resembles the 'string of
pearls' situation described by KS_rger et al. [9]. For this geometry a higher than
intrinsic activation energy is predicted at low temperatures; at higher tempera-
tures, when the physisorption is small, the activation energy will approach the
intrinsic value. It therefore seems plausible to attribute the exceptionally high
apparent activation energy over PtHMor in the low temperature regime to the
limitation by single file diffusion in the one-dimensional mordenite channels.
The higher rate of n-pentane isomerization over Pt-SZ is thus caused, at least
partially, by the more favorable geometry of this catalysts system.

4.2. Effect of pore geometry on isomerization mechanism of n-butane


For n-butane isomerization over Pt-SZ a bimolecular pathway is used,
involving an adsorbed C 8 intermediate. The huge difference in catalytic activity
between Pt-SZ and PtHMor indicates that PtHMor is unable to catalyze this
bimolecular reaction. In previous research, Iglesia and Baumgartner argued that
the high performance and stability of P t / K L catalysts for C 6 aromatization is
due to the very low chance for bimolecular reactions in one-dimensional
micropores [17]. This is a consequence of the general principle that bimolecular
encounters are extremely unlikely in one-dimensional pores [18]. Since butanes
and pentanes have the same minimum kinetic diameters, the geometric con-
straints which force pentane isomerization to be single file diffusion controlled
will also prevent butane molecules from passing each other in one-dimensional
mordenite channels. As the inter-molecular route requires that a butene molecule
interacts with an adjacent C 4 carbenium ion, this process would suffer greatly
from the geometrical constraints in PtHMor. It is interesting to note that in the
case of n-pentane isomerization over Pt-SZ the main reaction is monomolecular,
but that the minor contribution of a bimolecular mechanism is indicated by the
formation of hexane as a byproduct. This suggests that with Pt-SZ a pathway
exists in which C 5 is dimerized to C10 intermediates. In contrast, this reaction
pathway appears to be impossible with PtHMor under the same conditions: no
hexane formation was observed. This lends further credence to the view that
under the conditions studied the peculiar geometry of the one-dimensional pores
in mordenite suppresses the bimolecular pathway.

4.3. Isomerization on Pt-proton adducts


Although the rates per gram catalyst of n-pentane isomerization differ less
than a factor of two between PtHMor and Pt-SZ, their kinetic parameters in
176 H. Liu et al.//Applied Catalysis A: General 137 (1996) 167-177

Table 2 reveal a rather dramatic difference between these catalysts. Acid


catalyzed reactions of alkanes are known to be chain reactions in which hydride
ion transfer is an important step. In the case of Pt-SZ it is likely that hydride ion
transfer occurs when an alkane molecule impinges to an adsorbed carbenium
ion. This bi-molecular process of the Eley-Rideal type enables every alkane
molecule to be temporarily in the state of an adsorbed carbenium ion. Such a
process will, however, be very difficult inside the micropores of mordenite. It
thus appears plausible that on PtHMor adsorbed carbenium ions are formed in a
different way, viz. via dissociative adsorption of n-pentane on Pt metal clusters,
or as suggested in other work of the present group, dissociative adsorption on
Pt-proton adducts. In view of those results, which we shall not repeat here, it is
most likely that isomerization will take place on Pt-proton adducts which thus
act as 'collapsed bifunctional sites' [19,20]:
n - CsHI2 + H ÷ - Pto ~ n - C s H ~ ] - Pt n + H 2 (1)
n - C s H ~ ] - Pt, ---> i - C s H ~ ] - Pt n (2)
i - CsH~-1 - Pt, + H 2 ~ i - C5H12 + H + - Pt, (3)
The concentration of these CsH~]-ions will be proportional to the ratio
( P c n,2/PH ); accordingly, the reaction orders over PtHMor are negative (0 >
n >~ - 1) i~a H 2 and positive (0 < n < 1) in pentane, but over Pt-SZ they are
near zero in H 2 and > 1 in pentane. The predictions of this model agree with
the observed kinetic parameters.

5. Conclusions

The remarkable differences in isomerization rate of n-butane between PtH-


Mor and Pt-SZ and the significant differences in kinetic parameters for the
isomerization of n-pentane indicate that the reaction mechanisms are fundamen-
tally different over these catalysts. Whereas the numbers of metal and acid sites
are comparable and the strength of the acid sites is similar in these bifunctional
catalysts, the microgeometry is significantly different. It has been shown that the
observed differences in catalysis and kinetic parameters can be rationalized in
terms of geometric factors. The high apparent activation energy and the pre-
dicted break in the Arrhenius plot at a temperature where physisorption of
n-pentane becomes small, indicate that over PtHMor t single-file diffusion
controls n-pentane isomerization. Additional kinetic data suggest that the pre-
vailing pentane isomerization mechanism is intra-molecular and that a Pt-pro-
ton adduct is the catalytic site for this bifunctional reaction. For butane
isomerization, the geometrical constraints of one-dimensional channels in mor-
denite suppress the dimerization step. Therefore, PtHMor is virtually inactive for
n-butane isomerization, whereas Pt-SZ with acid sites of comparable strength is
a highly active and stable catalyst.
H. Liu et al./Applied Catalysis A: General 137 (1996) 167-177 177

Acknowledgements

We gratefully acknowledge financial support from the Director of the Chem-


istry Division, Basic Energy Sciences, U.S. Department of Energy, Grant
DE-FG02-87ER13654. A start-up fund from the Center for Catalysis and
Surface Science of Northwestern University and a grant in aid from Shell
Development Company are gratefully acknowledged. We thank Magnesium
Elektron Inc. for kindly providing us with samples of sulfated zirconia.

References

[1] D.M. Brouwer and H. Hoogeveen, Progr. Phys. Org. Chem., 9 (1972) 179.
[2] D.M. Brouwer, in R. Prins and G.C.A. Schuit (Editors), Chemistry and Chemical Engineering of
Catalytic Processes, Sijthoff & Noordhof Publ., Alphen a.d. Rijn, 1980, p. 173.
[3] F. Chevalier, M. Guisnet and R. Maurel, in G.C. Bond, P.B. Wells and F.C. Tompkins (Editors),
Proceedings of the 6th International Conference on Catalysis, London, The Chemical Society, 1977, p.
478.
[4] C.-Y. Hsu, C.R. Heimbruch, C.T. Armes and B.C. Gates, J. Chem Soc., Chem. Commun., (1992) 1645.
[5] V. Adeeva, T.W. de Haan, J. J~inchen, G.D. Lei, V. Schiinemann, L.J.M. van de Ven, W.M.H. Sachtler
and R.A. van Santen, J. Catal., 151 (1995) 364.
[6] V. Adeeva, G.D. Lei, W.M.H. Sachtler; Appl. Catal. A, 118 (1994) LI 1-LI5.
[7] H. Liu, V. Adeeva, G.D. Lei and W.M.H. Sachtler, J. Mol. Catal., 100 (1995) 35.
[8] H. Liu, G.D. Lei and W.M.H. Sachtler, Appl. Catal. A, in press.
[9] J. K~irger, M. Petzold, H. Pfeifer, S. Ernst and J. Weitkamp, J. Catal., 136 (1992) 283.
[10] Chr. Rodenbeck, J. K~irger and K. Hahn, J. Catal., 157 (1995) 656.
[11] Z. Karpifiski, S. Gandhi and W.M.H. Sachtler, J. Catal., 141 (1993) 337.
[12] B.T. Carvill, B.A. Lemer, B.J. Adelman, D.C. Tomczak and W.M.H. Sachtler, J. Catal., 144 (1993) I-8.
[13] B.A. Lerner, B.T. Carvill and W.M.H. Sachtler, Catal. Today, 21 (1994) 23.
[14] P.E. Eberly, J. Phys. Chem., 67 (1963) 2404.
[15] B.T. Carvill, PhD Thesis Northwestern University, Evanston, IL, USA, December 1993, p. 143.
[16] D.W. Breck, Zeolite Molecular Sieves, Structure, Chemistry, and Use, Wiley, New York, 1974, p. 636.
[17] E. Iglesia and J.E. Baumgartner, Stud. Surf. Sci. Catal., 75 (1993) 993-1006.
[18] R. Kopelman, Science (Washington), 241 (1988) 1620.
[19] W.M. Sachtler and Z.-C. Zhang, Adv. Catal., 30 (1993) 129.
[20] T.J. McCarthy, G.-D. Lei and W.M.H. Sachtler, J. Catal., in press.

You might also like