University of Southern California Dissertations and Theses

Fundamental studies of methanol to olefin (MTO) catalysis

(USC Thesis Other)

Fundamental studies of methanol to olefin (MTO) catalysis

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content FUNDAMENTAL STUDIES OF METHANOL TO
OLEFIN (MTO) CATALYSIS
by
Weiguo Song
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Chemistry)
May 2001
Copyright 2001 Weiguo Song
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3027781
UMI'
UMI Microform 3027781
Copyright 2002 by Bell & Howell Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
Bell & Howell Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES. CALIFORNIA 90007
This dissertation, written by
under the direction of hXs Dissertation
Committee, and approved by all its members,
has been presented to and accepted by The
Graduate School, in partial fulfillment of re
quirements for the degree of
DOCTOR OF PHILOSOPHY
D ean of Graduate Studies
D a te ..... M a * . j j .,.. 2001
DISSERTATION COMMITTEE
Chairperson
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Weiguo Song James F. Haw
ABSTRACT
FUNDAMENTAL STUDIES IN METHANOL TO OLEFIN CATALYSIS
Methanol to olefin (MTO) reaction mechanisms on zeolite HZSM-5 and
HSAPO-34 were studied experimentally and theoretically. Experimental studies
were performed with solid state NMR and gas chromatograph (GC) using a pulse-
quench reactor. On both catalysts “Carbon Pool” type mechanisms, which require a
kinetic induction period, were confirmed. On HZSM-5, 1, 3-dimethylcyclopentenyl
carbenium ion is an intermediate in the MTO reactions and is a precursor to
aromatic compounds. The presence of 1, 3 -dimethyIcyclopentenyl cation on HZSM-
5 reduced the induction period of the MTO reactions. On HSAPO-34 catalyst, the
organic reaction centers were methylbenzenes which were synthesized through “ship
in a bottle” mechanisms from methanol. The average number of methyl groups per
benzene ring governs the selectivity of ethylene and propene. Ethylene was favored
with lower average methyl number per ring. The SAPO-34 was modified by adding
phosphorous compounds into the SAPO-34 cages. A phosphorous treated catalyst
showed higher selectivity towards ethylene.
The chemistry of carbenium cations on zeolites was studied. Pentamethyl
benzenium ion was synthesized on HZSM-5 from methanol and benzene with a
pulse quench reactor. The Heptamethylcyclopentenyl cation was synthesized on
HSAPO-34 from acetone with CAVERN techniques. The presence of both cations
was verified by theoretical calculations. The reactions betw een 1, 3-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
dimethy Icy clop entenyl cation and several bases were also studied. Theoretical
calculations predicted that a base with gas phase proton affinities higher than 215
kcal/mol, such as trimethyl phosphine and pyridine, will deprotonate the cation to
form a neutral cyclic diene and a protonated base; while weak bases can only form
hydrogen bonds with the cation. Some bases can also react with the cation to form
onium ion by nucleophilic addition. All theoretical predictions were verified by
experimental results.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Dedication
To my Family
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Acknowledgements
I want to thank Dr. Haw for several things. As my academic advisor, he has
always been resourceful and many times been ingenious in guiding my research
throughout the past five years. When his research group moved from Texas A&M
to USC in 1998, Dr. Haw and I went through all the difficulties to put things on
track again. That was a great experience and a great move.
Dr. John Nicholas did almost all the theoretical work in my dissertation and
my papers. Without his work, much of my research would be incomplete.
Dr. Teng Xu was my mentor when I first joined Haw’s group in 1996. He
taught me the research techniques and helped me to be an independent researcher.
My fellow group members in either Texas A&M or USC have been very
helpful and kind to me. I thank all of them, especially Katy who helped me a lot
with dissertation preparation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table of Content
Preface Dedication II
Acknowledgments III
List of Tables VII
List of Figures VIII
Chapter 1 Introduction 1
1.1 Methanol to Olefin 2
1.2 Zeolites 10
1.3 Solid State NMR 19
1.4 Combined Solid State NMR and GC Study of 26
Heterogeneous Catalysis
1.5 Reference and Notes 34
Chapter 2 SAPO-34 Synthesis and Characterization 38
2.1 SAPO-34 Synthesis 39
2.2 SAPO-34 Characterization 41
2.3 Reference and Notes 48
Chapter 3 Roles for Cyclopentenyl Cations in the Synthesis of 49
Hydrocarbons from Methanol on Zeolite
Catalyst HZSM-5
3.1 Introduction 50
3.2 Experimental Section 53
3.3 Theoretical Methods 56
3.4 Experimental Results 58
3.5 Theoretical Results. 76
3.6 Discussion 93
3.7 Conclusion 101
3.8 References and Notes 104
Chapter 4 Acid-Base Chemistry of a Carbenium Ion in a 109
Zeolite under Equilibrium Conditions: Verification
of a Theoretical Explanation of Carbenium ion
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
stability.
4.1 Introduction 110
4.2 Experimental Section 114
4.3 Results 117
4.3.1 Theoretical Results 117
4.3.2 Experimental Results 130
4.4 Discussion 142
4.5 Conclusion 146
4.6 References and Notes 148
Chapter 5 Synthesis of a Benzenium Ion in a Zeolite 150
with Use of a Catalytic Flow Reactor
5.1 Results and Discussion 151
5.2 References and Notes 158
Chapter 6 A Persistent Carbenium Ion on the 160
Methanol-to-Olefin Catalyst HSAPO-34:
Acetone Shows the Way
6 .1 Introduction 161
6.2 Experimental Section 164
6.3 Results 167
6.4 Discussion 179
6.5 References and Notes 191
Chapter 7 Methylbenzenes are the Organic Reaction 193
Centers for Methanol
to Olefin Catalysis on HSAPO-34
7.1 Results and Discussion 194
7.2 References and Notes 201
Chapter 8 Supramolecular Origins of Product Selectivity 202
For Methanol-to-Olefin Catalysis on HSAPO-34
8.1 Introduction 203
8.2 Experimental Section 205
8.3 Results 207
8.4 Discussion 222
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8.5 References and Notes
Chapter 9 Improved Methanol-to-Olefin Catalyst with
Nanocages Functionalized through
Ship-in-a-Bottle Synthesis from PH3
9.1 Results and discussion
9.2 References and Notes
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of tables
Chapter 1
Table 1.1
Table 1.2
Chapter 2
Table 2.1
Chapter 3
Table 3.1
Table 3.2
Chapter 4
Table 4.1
Representative zeolites with their pore geometries
Chemical shifts of mesityl oxide and acetone on solid acids
Elemental analysis results
Theoretical and experimental l^C isotropic chemical
shifts for cations 1 and 4.
Theoretical (B3LYP/6-311G*) barriers for various
methylation reactions in the gas phase
Experimental and Theoretical (B3LYP/6-311G*) Proton
Affinities (AH), Theoretical Methyl Cation Affinities
(AH), and Theoretical Energies of the Three Possible
Adsorption Complexes Relative to the Sum of the
Total Energies of Isolated C7H n+ and the Co Adsorbate
(AE).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of Figures
Chapter 1
Figure 1.1 Global oil production for resources o f 1800, 2200 3
2600 billion barrels.
Figure 1.2 General mechanism of methanol to olefin Catalysis. 6
Figure 1.3 Representative structure of zeolite ZSM-5 and SAPO-34. 12
Figure 1.4 General steps in zeolite synthesis 16
Figure 1.5 Cross polarization pulse sequence 24
Figure 1.6 CAVERN device 27
Figure 1.7 Scheme of pulse-quench reactor 30
Figure 1.8 Three kinds of quench experiments 32
Chapter 2
Figure 2.1 SAPO-34 synthesis procedure 40
Figure 2.2 XRD spectrum of SAPO-34 45
Figure 2.3 !H, 2 7A1 and 31P spectra of SAPO-34 46
13
Figure 2.4 75.4 MHz C Spectra of as-synthesized and 47
calcined SAPO-34
Chapter 3
Figure 3.1 75.4 ^ C CP/MAS NMR spectra of the reaction products 59
viii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3.2
Figure 3.3
Figure 3.4
Figure 3.5
Figure 3.6
Figure 3.7
Figure 3.8
Figure 3.9
of dimethyl ether-C 2 retained on zeolite HZSM-5
following various times at 573 K.
13c CP/MAS NMR spectra of the reaction products of 61
ethylene- 13C2 retained on zeolite HZSM-5 following
various times at 623 K in a pulse-quench reactor.
13 c CP/MAS NMR spectra o f the reaction products of 63
ethylene-13C2 retained on zeolite HZSM-5 following
4 s of reaction at various temperatures (indicated)
in a pulse-quench reactor.
GC traces (thermal conductivity) characterizing the 65
Volatile products from the reactions of ethylene-13C2
on zeolite HZSM-5 (4 s reaction time) at various
temperatures in a pulse-quench reactor.
13 c CP/MAS NMR studies of the reactions of DME- 13C2 6 8
with cyclopentenyl cations at 548 K on zeolite HZSM-5
in a pulse quench reactor.
13c CP/MAS NMR studies of the reactions of DME-13C2 71
with cyclopentenyl cations at 573 K on zeolite HZSM-5
in a pulse quench reactor.
Variable temperature 13c MAS NMR study of 1 on 73
zeolite HZSM-5. A single sample was prepared by
pulsing ethylene-13 C2 onto a catalyst sample at 623 K
in a pulse-quench reactor and allowing it to react for 0.5
seconds.
B3LYP/6-311G** optimized geometry of the 1,3- 79
dimethylyclopentadienyl cation 1 coordinated to the
zeolite anion (ion-pair structure).
B3LYP/6-311G** optimized geometry of the n complex 82
formed by adsorbing neutral cyclic diene 2 to the zeolite
acid site model used to obtain the ion-pair structure
in Figure 8 .
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3.10
Figure 3.11
Figure 3.12
Figure 3.13
Chapter 4
Figure 4.1
Figure 4.2
Figure 4.3
Figure 4.4
Figure 4.5
Figure 4.6
Figure 4.7
B3LYP/6-311G** optimized geometry of the 84
framework-bound alkoxy species that could form as an
alternative to cation 1 .
B3LYP/6-311G* optimized geometries of the three stable 8 6
points along the reaction pathway for the conversion of
trimethyloxonium cation and cyclic diene 2 to DME and
cation 3.
Proposed reaction mechanism for the conversion of 1 to 91
toluene.
Potential energy surfaces (kcal/mol) for removal of ethylene 99
or propene from cations 11 or 13, respectively.
B3LYP/6-311G* optimized geometries for possible 119
complexes formed by C7H n + and NH 3 .
B3LYP/6-311G* optimized geometries for 122
C7H 1/ — :B hydrogen bonded complexes.
B3LYP/6-311G* optimized geometries for C7H 10—H:B+ 124
hydrogen bonded complexes.
B3LYP/6-311G* optimized geometries for onium ions. 126
13C CP/MAS NMR spectra of samples prepared by 131
forming small amounts (less than 0 . 1 equiv.) of the
C7H n + cation in zeolite HZSM-5 and then absorbing
an excess (typically 2 equiv.) of the indicated base.
Experimental and theoretical (GIAO-MP2/tzp/dz//B3L- 133
YHP/6-311G*, in parenthesis) 13C chemical shifts (ppm)
for: (a) C7H 10; and (b) C7H 11*
1 3 C CP/MAS NMR spectra of samples prepared by 134
forming small amounts (less than 0.05 equiv.) of the
C7H n + cation in zeolite HZSM-5 and then absorbing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
0, 1, or 2 equiv. of dimethylacetamide (DMA).
Figure 4.8
Figure 4.9
Figure 4.10
Chapter 5
Figure 5.1
Figure 5.2
Chapter 6
Figure 6.1
Figure 6.2
Figure 6.3
Figure 6.4
CP/MAS NMR spectra of samples prepared by forming
small amounts (less than 0.05 equiv.) of the C7H n +
cation in zeolite HZSM-5 and then absorbing the indicated
amounts of pyridine (Py)-15N.
13
C CP/MAS NMR spectra of samples prepared by
forming small amounts (less than 0.05 equiv.) o f the
C7H n + cation in zeolite HZSM-5 and then absorbing an
excess (typically 2 equiv.) of the indicated base.
CP/MAS NMR spectra of samples prepared by forming
small amounts (less than 0.05 equiv.) of the C7H n +
cation in zeolite HZSM-5 and then absorbing the indicated
amounts of phosphine.
13
Selected 75.4 MHz C MAS spectra of benzenium ion
3 on zeolite HZSM-5.
B3LYP/6-311G** geometry of 3 optimized in Cs symmetry.
75.4 MHz 13C CP/MAS spectra from a CAVERN study
of acetone-2-1 3 C (1 equiv.) on HSAPO-34.
75.4 MHz 1 3 C CP/MAS spectra showing the effect of
dipolar dephasing (50 p,s) on the specrtum from figure 6.1
of acetone-2-1 3 C (1 equiv.) on HSAPO-34 after heating to
maximum of 513 K.
75.4 MHz 1 3 C CP/MAS spectra from a CAVERN study
of acetone-1,3-13C2 (1 equiv.) on HSAPO-34.
75.4 MHz 13C CP/MAS spectra from pulse-quench
137
139
141
154
155
168
170
171
173
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 6.5
Figure 6 . 6
Figure 6.7
Chapter 7
Figure 7.1
Figure 7.2
Chapter 8
Figure 8.1
Figure 8.2
Figure 8.3(a)
Figure 8.3(b)
Figure 8.4
studies of acetone-2-13C(l equiv.) on HZSO-34.
Gas chromatographs (FID detection) sampling the 175
product streams of reactions carried out on HSAPO-34
catalyst bed at 673 K.
B3LYP/6-311G* optimized geometries for cations la-3a. 177
B3LYP/6-311G* optimized geometries for olefins lb-3b. 183
75 MHz 1 3 C CP/MAS NMR spectra of samples from a 195
pulse-quench study of methanol conversion on HSAPO-34
at 673 K.
GC (flame ionization) analyses of the gases exiting the 197
HSAPO-34 catalyst bed operated as in the previous figure,
except as a “double-pulse” experiment.
75 MHz 1 3 C CP/MAS NMR spectra showing the loss o f 209
methyl groups as a function o f time from methylbenzenes
trapped in the HSAPO-34 nanocages at 400 °C.
Gas chromatography (flame-ionization detection) analyses 211
of the volatile products captured immediately prior to the
thermal quench from the experiments used to prepare some
of the samples for Fig. 8.1
The rates of formation of ethylene, propene, and 2-butene 213
as a function of time from a single experiment similar to
those used for Figs. 8.1 and 8.2.
Ethylene and propene selectivity as a function of the average 214
number of methyl groups per ring, Meave-
75 MHz 1 3 C CP/MAS NMR spectra from experiments 215
probing the steady-state average number of methyl groups
xii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 8.5 (a)
Figure 8.5 (b)
Figure 8 . 6
Figure 8.7
Chapter 9
Figure 9.1
Figure 9.2
per ring at various methanol space velocities.
Ethylene selectivity at 400 °C vs. the logarithm of 216
methanol weight-hourly space velocity.
The ethylene selectivity data in a plotted against Meave. 217
75 MHz 1 3 C CP/MAS NMR spectra probing the steady-state 219
average number of methyl groups per ring at fixed methanol
space velocity ( 8 h '1 ) at various temperatures.
75 MHz 1 3 C CP/MAS NMR spectra from experiments 221
probing the effect of co-feeding water on the average
number of methyl groups per ring at a 400 °C reactor
temperature.
31P MAS NMR spectra showing functionalization of 233
HSAPO-34.
GC (flame ionization) analyses of the product gases in MTO 236
conversion on catalysts functionalized by introducing
phosphorous species into the nanocages.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 1.
Introduction
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.1 Methanol to olefin (MTO)
l.l.l.O verview
Ethylene and propene are among the highest volume chemicals. With
increased demand for polyethylene, the ethylene market continues to increase.
World production of ethylene reached 209 billion lbs. in 1999, and is projected to
reach 248 billion lb. in 20031. Most light olefins are manufactured through catalytic
cracking of crude oil, a limited resource. It is difficult to estimate when we will run
out o f crude oil, the more pessimistic projections estimate that production will
decline in a matter of only 10 to 20 years . Figure 1.1, from one such projection,
shows that global oil production will peak by 2020, then decline. Even if petroleum
production somehow remains constant, the strong and growing olefin demand will
create incentives for alternative routes to olefins from abundant sources.
In 1977, Clarence Chang and coworkers at Mobil discovered that gasoline-
range hydrocarbons were produced when they passed methanol through a catalyst
bases on the aluminosilicate zeolite HZSM-5. This discovery was accidental in that
their intention was to make oxygenates with higher carbon numbers. This discovery
was then quickly developed into what was called methanol-to-gasoline (MTG)
technology, and was commercialized in New Zealand in 1979.4 MTG did not
became a world-wide industry, as originally envisioned, because of what proved to
be a relatively stable and cheap gasoline supply. The MTG process produces a high
yield of aromatics and these are undesirable in modem gasoline. Furthermore,
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
u
e s
( U
> *
u
V
a
ja
< U
es
X
o
o
3 5
• • • •
1800 billion barrels
30
• • •
25
20
2200 billion barrels
15
10
2600 billion barrels
5
0
1950 1960 1970 1980 1990 2000 2010 2020 2030
Year
Figure 1.1 Projected global oil production. Adapted from
http://www.igc.org/wri/climate/jm_oil_ 0 0 1 .html
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
olefins, rather than aromatics and alkanes have emerged as valuable hydrocarbon
feedstocks as a result of growing polyolefin demand. Thus, the object of methanol
conversion has changed from MTG to methanol-to-olefin (MTO). In order to
convert methanol to olefins such as ethylene and propene, one must use a catalyst
and process conditions that limit secondary reactions leading to aromatics and other
products.
Once economical MTO technology is feasible, it will be possible to convert
natural gas to polyolefins in integrated chemical complexes. The advantages o f this
strategy include economic incentives for natural gas utilization and the remoteness
of many gas reserves. It is cheaper to ship polyolefin pellets by train than build and
operate pipelines across vast, politically unstable regions. Natural gas, primarily
methane, can be steamed using mature technology to form synthesis gas, a mixture
of CO, CO2 and H 2 . Other mature processes, using CuO/ZnO/A^Os catalyst, takes
synthesis gas to methanol with extraordinary selectivity. 5 O f course there are many
mature processes for polyolefins, and several new ones based on metallocene
catalysts are coming on-line.
The missing step, MTO, has been around in some forms since shortly after
the development of MTG processes. Some of the earliest MTO processes were
based on the aluminosilicate zeolite HZSM-5, a material with channels ca. 0.55 nm
in diameter. Aromatics such as toluene and -xylene readily diffuse through the
ZSM-5 topology, and HZSM-5 was used for several years in New Zealand in a
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
process that made high-aromatic gasolines to meet 30 % o f that nation’s needs.
Aromatics are no longer desirable in US gasoline, and to serve as an MTO catalyst,
HZSM-5 must be modified to greatly reduce aromatic formation. This has been
attempted, with some degree of success, by further restricting the channel size,
2+
usually by partial exchange with Mg or by adsorbing an organic phosphorous
compound and calcining to leave some kind of phosphate debris in the channel.
More recently, HSAPO-34 and closely related materials (e.g., those with small
amounts of metals in the framework) emerged as the materials of choice for MTO
catalysis, in part because aromatics can not diffuse through the ca. 0.38 nm windows
in these materials, and commercialization of SAPO-type MTO catalysts seems
certain.
1.1.2. M TO mechanisms
MTO is not only important to the chemical industry and the economy, but is
also of great interest to academic chemists. The mechanisms o f MTO have been
studied by many research groups and hundreds of papers have been published. The
most recent review cited over 300 papers published within the last two decades.6
Figure 1.2 shows a general mechanism of the methanol to olefin reaction.
Methanol reaches fast equilibrium with dimethyl ether and water over a solid acid
catalyst, such that the methanol/dimethyl ether mixture is usually considered as
thereactant. Through a slow step, compounds with initial carbon-carbon bonds are
5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
" H 2 0
Cjj30H ^ ---------- -► CH3OCH3
+H20
Slow
c h 2=c h 2
C H 3 0 H /
C H 3 0 C H 3
Reaction Center
Fast
'CH3OH/
C H 3 O C H 3
C2, C3, and higher olefins
J
Aromatics, paraffins
Figure 1.2 General mechanisms o f MTO reactions.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
formed, presumably as ethylene. More light olefins are formed following the
formation of the initial C-C bonds, through two possible consecutive and parallel
mechanisms. Aromatics and paraffins are the secondary reaction products from
olefins.
The mechanism of fast equilibration among methanol, dimethyl ether and
water over a solid acid catalyst is simple.7 Methanol is protonated by a Brpnsted
acid site to form an electrophile. Another methanol, which acts as a nucleophile,
attacks the protonated methanol. W ater is then eliminated and dimethyl ether is
formed.
The mechanisms of initial C-C bond formation have been debated since the
discovery of the MTG reaction. No general consensus has been reached yet, though
over 2 0 possible mechanisms have been proposed including the oxonium ylide
8-9
mechanism by van de Berg and Olah, ’ the carbene mechanism by Chang and
Venuto, 10-11 the carbocationic mechanism by Ono and M ori, 12" 13 and the free
radical mechanism by Zatorski. 14 All these mechanisms have some direct or
indirect experimental evidence. This dissertation will not focus on the mechanism of
initial C-C bond formation.
After the initial C-C bonds are formed there are two possible pathways to
large-scale formation of light olefins: the consecutive pathway and the parallel
pathway.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In the consecutive pathway, the mechanism is that after ethylene is formed,
the ethylene either leaves the catalyst as the product or adds to another methanol to
form a propene. Propene then reacts with another methanol to make C4 olefins or
diffuses out directly. This is shown in scheme l . 15
2C\ H m C2 H4 + H 2 O
C2 H 4 + C | — — — ----- ► C3 H6
C 3 H 6 + C 1 —----------------------► C 4 H 8
Scheme 1
The parallel pathway, also called the “hydrocarbon pool” mechanism, was
1 6 1 7
proposed by Kolboe et al. " They studied the isotopic distribution o f propene in
MTO reactions on SAPO-34. By flowing 13C-methanol (as MTO reactant) and 12C-
ethanol (as an ethylene source) together over SAPO-34 at 400 °C, they found that
propene and C4 olefins almost exclusively came from methanol with large excess of
13 13 12
C atoms, while most ethylene was either all C or all C. They also found that
ethylene alone was unreactive under the reaction condition employed. These
findings suggested that propene is not formed by methylation of ethylene, as
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
proposed in the consecutive pathway. Kolboe proposed a “carbon pool” mechanism
to account for their findings, in which a phenomenological Carbon Pool was the
intermediate. The Carbon Pool is continuously methylated by methanol during the
reaction; at same time, it decomposes to make ethylene, propene, and possibly
butenes.
The nature of the carbon pool was not revealed till recently. Haw et al. found
out that dimethyl cyclopentenyl cation was an important intermediate in MTO
reactions over HZSM-5. The dimethyl cyclopentenyl cation, it’s derivatives and
HZSM-5 framework form a so-called “reaction center”, which reacts with methanol
18
and eliminate ethylene, propene and other products. They also found out that on
SAPO-34, methylbenzenes were the organic part o f the reaction center. 19 The
identification of the reaction center on HZSM-5 and SAPO-34 will be discussed in
chapter 3 and 7 of this dissertation.
1.1.3. Product selectivity
Product selectivity has always been the focus in MTO. An understanding of
how one product is favored over another product will certainly help to develop
technology with enhanced product selectivity. The goal in MTO is to have higher
light olefin selectivity, especially towards ethylene, and lower aromatics and
alkanes. Naturally, shape selectivity of the catalyst is vital in enabling us to meet
these goals. There are three types of shape selectivity: reactant shape selectivity,
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
product shape selectivity and transition state shape selectivity. Since aromatic
compounds are much larger than light olefins, we can improve light olefm
selectivity by using a catalyst with better product shape selectivity; for example,
using SAPO-34 instead of HZSM-5. So far, SAPO-34 has been the best catalyst
available for MTO as far as olefin selectivity is concerned. Ethylene and propene are
the major products of the MTO reaction using SAPO-34. However, reactant and
product shape selectivity cannot improve ethylene selectivity over propene in that
ethylene and propene come from the same reactant, and their kinetic diameters are
similar. Transition state shape selectivity may improve the ethylene yield, if we can
modify the catalyst so that the transition state that leads to ethylene is favored.
In Chapter 8 of this dissertation, the molecular origin o f the product
selectivity o f MTO on SAPO-34 will be discussed. The average number of methyl
20
groups per benzene ring controls whether ethylene or propene is favored. Chapter
9 of this dissertation will discuss an effort to modify the SAPO-34 framework with
21
PH 3 so that transition-state shape selectivity will favor ethylene formation.
1.2 Zeolites
1.2.1. Introduction
22
Zeolites are possibly the most important heterogeneous catalysts in the
23
chemical industry. Representative applications of zeolites are catalytic cracking
(ultrastable Y), hydroisomerization (Pt-modemite), benzene alkylation (ZSM-5),
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
methanol to olefins (SAPO-34), partial oxidation (titanium silicate), etc. . 24 Four
25
characteristics of zeolites are vital to their activities and selectivities.
1 Well defined, uniform framework structure.
2 Strong acid sites, (either Bronsted or Lewis acid sites)
3 High surface area.
4 High thermal stability.
1.2.2 Zeolite structures
More than 50 different kinds o f naturally occuring zeolites and over 160
kinds of synthetic zeolites have been discovered. The International Zeolite
26
Association (IZA) has published "The Atlas o f Zeolite Structure Types".
Interested readers can get detailed structural information for each zeolite from the
IZA homepage.26
The primary building units of any zeolites are TO 4 tetrahedra (T-Si, Al,
etc.). These tetrahedra are connected to each other through oxygen bridges. Different
connection patterns result in different zeolite structures. Figure 1.3 shows the
structure of two representative zeolites, HZSM-5 and SAPO-34. These two zeolites
are also the two catalysts used in MTO, and they are studied in this dissertation. As
mentioned earlier, zeolites control shape selectivity in three ways: reactant shape
selectivity, transition state shape selectivity and product shape selectivity. Figure 1.4
27
illustrates these three types o f shape selectivity. One of the deciding factors for
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Elliptical straight
channels (5.1 x 5.6 A)
Q
a
Near circular channels
(5.4 x 5.6 A)
C
Figure 1.3. Structures of zeolite ZSM-5 and SAPO-34. a. Schematic representation
of the channel arrangements of ZSM-5. b. Structure of ZSM-5 showing 10 member
oxygen windows, c. SAPO-34 structure showing a cage and 8 member oxygen
windows.
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
shape selectivity in zeolites is their cyclic oxygen bridge windows. Oxygen
windows are the links between the internal surface of the framework and the outside
media. They are also the entrance for the reactant and exit for the product. The size
o f the oxygen window determines the size of the molecule that can diffuse through
the zeolites. Zeolites are usually classified according to the number of oxygen
bridges in their oxygen windows. Table 1.1 lists some zeolites with their oxygen
window geometry. HZSM-5, which has 10-member ring oxygen windows, is a
medium pore zeolite. SAPO-34 has an 8 -member ring oxygen window. Its 3.8 A by
3.7 A oxygen windows allow small molecules, such as methanol, ethylene, propene
and linear C4 hydrocarbons to freely diffuse through, while larger molecules, such as
branched C4 hydrocarbons and aromatics are restricted. These provide outstanding
product shape selectivity in the MTO reaction. Aromatics are a substantial
constituent of the products when using HZSM-5 in MTO, while when SAPO-34 is
used, aromatics are restricted inside the SAPO-34. Over 90% olefin selectivity can
7 0
be achieved on SAPO-34.
Another deciding factor for shape selectivity is the zeolite pore structure.
ZSM-5 and SAPO-34 have different kinds of pores. The ZSM-5 framework has a
channel structure that is composed o f one sinuous and one straight channel. These
channels are interconnected to each other. The intersections of the channels can hold
very large molecules. The SAPO-34 framework has cage structure about 6.7 A by
1 0 A. This is large enough to hold an aromatic molecule, but the 8 -member oxygen
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 1.1 Representative zeolites with their pore geometry
Zeolite
structural
code
Representative
material
Number o f oxygen atoms
in oxygen window
Sizes of the
oxygen windows
(A )
AFI
AIPO4 - 5
1 2 7.3 x 7.3
FAU Y, SAPO-37 12 7.4 x 7.4
MOR Mordenite 12 6.5 x 6.5
MFI ZSM-5, TS-1 1 0 5.3 x 5.6
MEL ZSM-11, TS-2 1 0 5.3 x 5.4
FER Ferrierite 1 0 4.2 x 5.4
CHA Chabazite,
SAPO-34
8 3.8 x 3.8
ERI Erionite 8 3.6 x 5.1
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
windows are only 3.8 A by 3.7 A , and only small molecules may diffuse through
them. If the reaction center in SAPO-34 is a large species that cannot pass through
the 8 -member oxygen window, then this species must be synthesized inside the cage
from small molecules to make a "ship in a bottle" catalyst.
1.2.3. Zeolite synthesis
Zeolites are usually synthesized through hydrothermal routes. Figure 1.4
29
shows the general steps in the hydrothermal route. Al(OH )3 and Si(OH) 4 gel are
formed when silica and aluminum sources are stirred with water. The gel is mixed
with template agent and the mixture is then sealed in an autoclave at high
temperature for an extended time. During this period, aluminum and silica sources
form secondary building units (SBU) of the zeolites. Under the influence of the
template agent (structural directing agent), SBUs are connected to each other in a
specific way to form a specific zeolite crystal.
1.2.4. Zeolite acidity
The frameworks of silicates and aluminophosphates are neutral. If some of
the silicon atoms are substituted by aluminum atoms in silicates (to form
aluminosilicates), or some of the aluminum atoms are substituted by silicon atoms in
aluminophosphates (to form silicoalum inophosphates), negative charges are
generated on the framework. These negative charges are balanced by cations. If the
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
h 2 o Gelling
Dissolution M ineralizing agent
Structure directing agent Nucleation growth
Zeolite crystal
Structural building units
Amorphous hydrogel
Reagents
Si, Al, P sources
Figure 1.4 General steps in hydrothermal synthesis of zeolites
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
cations are protons, these protons form Bronsted acid sites. Usually a Bronsted acid
site on a zeolite is illustrated as a cluster.
H
The acidity o f acid sites on zeolites is a much debated topic.
Aluminosilicates, such as HZSM-5 and HY, were believed to be solid super acids.
Carbenium ions, which are persistent in solution super acid, were proposed as
intermediates in reactions such as isomerization, alkylation and cracking on
30
zeolites. In liquid solution, the acidity o f an acid can be conveniently measured by
31
the pH value or by the Hammett scale. However, these two measurements cannot
be applied to solid acids. Farcasiu proposed an acidity scale based on the chemical
32
shift of carbonyl carbon of mesityl oxide in acid solution. The same scale can also
be applied to solid acids. Haw et al. studied the carbenium ions on solid acids with
33
solid state NMR. They found that active carbenium ions, such as the isopropyl ion,
were persistent on solid super acids, such as frozen SbFs, but not on zeolites. They
proposed an acidity scale for solid acids. The acidity is determined by whether
certain carbenium ion was persistent on that solid acid. By this standard, zeolites are
strong acids, but not superacids. HZSM-5 has the strongest acidity of the material
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
studied, equals to about 80% H2SO4. They also found out that C chemical shift of
13
acetone-2- C can be used to measure solid acid acidity, as mesityl oxide does.
Table 1.3 lists the chemical shift of mesityl oxide and acetone on some solid acids.
The chemical shift of the carbonyl carbon of acetone decreases with the acidity of
the solid acid. On the solid super acid AICI3, it is 240 ppm; on HZSM-5, it is 222
ppm; on HSAPO-34, which is less acidic than HZSM-5, it is 217 ppm.
Table 1.2 Chemical shift of mesityl oxide and acetone on solid acids
Solid acid Chemical shift of carbonyl
carbon on acetone (ppm)
Chemical shift of carbonyl
carbon on mesityl oxide
(ppm)
SbF5
250 205
H 2SO4
244 203
HZSM-5 223 190
HY 2 2 0 188
HSAPO-34 217 N/A
CDCI3
205 155
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.2.5. Carbenium ion chemistry on zeolites
Only a handful of carbenium ions have been proved by solid state NMR to
be persistent on the zeolites. These cations are cyclopentenyl cations, 3 4 indanyl
35 36 37
cations, trityl cation and benzenium ions . From theoretical calculations, Haw
and Nicholas proposed that for a carbenium ion to be persistent on the zeolites, its
38
parent olefin's proton affinity must be 209 kcal/mol or higher. Finding any new
persistent carbenium ions on any zeolites has always been a combination of hard
work and good luck. This dissertation will report the discovery o f the pentamethyl
17
benzenium ion on HZSM-5 in chapter 5 and the heptamethyl cyclopentenyl cation
on SAPO-34 in chapter 6 .
Carbenium ions are electron deficient species. They can act as electrophiles
to react with a nucleophile or they can act as acids to react with bases. This
dissertation will discuss the chemistry between cyclopentenyl cations and various
40
bases on HZSM-5 in chapter 4.
1.3 High resolution solid state NMR.
Nuclear magnetic resonance (NMR) spectroscopy has been a highly
informative technique for chemists. They have used NMR to determine molecular
structure and to monitor chemical reactions. Solid state NMR is one of the best tools
to study chemistry on solid materials.
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Every nuclear with non-zero spin number is NMR active. When an NMR
active spin I is put in an external magnetic field, Bo, it interacts with not only Bo, but
also the environment surrounding itself. From quantum mechanics, the total
Hamiltonian, H, which describes the interaction of spin I experiences can be given
i 41
by:
H zeem an H chem ical Shift ^D ipolar + Hj Coupling HQuadrapolar
Hzeeman, is the interaction between I and external magnetic field , Bq. It is the
strongest interaction among all NM R interactions. The nuclear energy states of I are
degenerate in the absence of Bo- When B o interacts with I, the nuclear energy states
split into different states according to the following equation.
E = -YmhBo/27i. (m = -I, -I + 1 ,... I - 1 ,1)
The selection rule is Am = ±1. This lead to the resonance frequency (Larmor
precession frequency) (Do of spin I in Bo:
(O0 = Y B0.
As in all spectroscopic techniques, when a radiation with frequency (Do is applied to
the system, the system will selectively adsorb energy with frequency (Do from the
radiation, the system will then release the adsorbed energy through relaxation. So by
monitoring either the changes of the radiation (CW NMR) or the released energy
(pulse FT NMR), we can get information about the system.
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
H chem icai shift comes from the electron clouds that surround the nuclei. In an
external Bo field, electron clouds generate a small loop current. This loop current
subsequently generates a small magnetic filed. The magnitude of this secondary
field is proportional to the magnitude of B o, and more importantly, also depends on
the electron cloud density, which is determined by the nuclei’s chemical
environment. The effective magnetic filed is then B o ( I - c t) for spin I, where a is the
shielding tensor that describes the effect of electron clouds. The resonance
frequency of I is then given by:
C00 = Y B o(1 - c t)
The chemical shift arises when spins in different chemical environments have
different shielding tensors. Chemical shift is usually reported in the 5 scale:
Pobserved - Preference ,
8 ----------------------------------------- x 1 0 b
Preference
The chemical shift tensor, cr, is a second rank tensor, and the chemical shift
is orientation dependent. Chemical shift anisotropy (CSA) describes the orientation
dependence of the chemical shift tensor. The chemical shift tensor is usually
characterized by three principal components, 5 n , 8 2 2 and 833 ( 8 1 1 > 8 2 2 > 8 3 3 )- In
solution NMR, molecules are tumbling fast enough to average the principal
components into an isotropic value, Siso, which is the value normally reported for
chemical shift.
21
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5iso = l/3((5i i + 6 2 2 +833)
In solid materials, chemical shift anisotropy causes line broadening of the
NMR spectrum. It also provides information about symmetry in the solid as well.
With magic angle spinning (MAS), 8 jso can be obtained along with spinning
sidebands. 8n, 8 2 2 and 833 values can also be calculated from 8iso and sidebands.
HDipoiar describes direct m agnetic coupling betw een nuclei. These
interactions include homonuclear and heteronuclear coupling. For ordinary organic
species, the homonuclear interactions are H-H and C-C interactions and the
heteronuclear interaction is the H-C interaction. It is averaged to zero in solution by
molecular tumbling. Dipolar-dipolar interactions are the major source for line
broadening in solid state NMR. Strong proton decoupling is an effective way to
eliminate the C-H interaction. The C-C interaction is much weaker than the C-H
interaction and can be reduced by magic angle spinning, thus high resolution solid
1 3
state C NMR is possible. The H-H interaction is very strong. Though some
progress has been made to reduce the H-H interaction by special pulse sequences,
high resolution solid state JH NMR of organic species is still a challenge.
H j coupling is a through-bond interaction. It is usually a weak interaction. In
solution NMR, where other interactions such as chemical shift anisotropy and
dipolar-dipolar interaction are averaged to zero, J coupling can be observed.
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
HQuadrapoiar applies to spins with spin number 1 or higher. HQuadrupoiar usually
results in broad lines, which makes it difficult to obtain useful information from
solid state NMR.
Two techniques frequently used in high resolution solid state NMR are
magic angle spinning (MAS) 42 and cross polarization (CP) 43
Interactions, such as chemical shift anisotropy and dipolar-dipolar
interactions can all be averaged to zero in a solid if the sample can be spun at high
enough speed, but mechanical difficulties limit the spinning speed o f the NMR
rotor. Quantum mechanical results show that many terms in Hchemical Shift and
Hoipoiar contain the factor (3cos 0 - 1). If the sample is spun at a specific angle with
respect to the external magnetic field, so that (3cos 0 - 1) = 0, then all these terms
become zero. This specific angle is called magic angle, 54.7° relative to Bo- In
practice, MAS can greatly reduce the line broadening from dipolar-dipolar
interactions and chemical shift anisotropy.
43
Cross polarization (CP) was developed by Pines. It is the most frequently
13
used pulse sequence in solid state NMR. Taking C NMR as an example, it utilizes
1 13
the dipolar-dipolar interaction between H (abundant nuclei) and C (rare nuclei)
13
to improve C NMR sensitivity. Figure 1.5 shows the event sequence o f cross
2 2
polarization experiments. First there is a H 90° pulse that excites the H spins. Then
the * 1 1 channel is switched to the Y direction in the rotating frame to have a spin-
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
90x
i contact
H
Polarization
transfers
Acquisition
c
signal signal
grows tracks'h
Time
Figure 1.5 Pulse sequences of cross polarization
50 qs
dephasing
contact
1 Polarization
| transfers
Acquisition
c
signal signal
grows tracks
Time
and dipolar dephasing (interrupted decoupling, ID).
24
13
lock. At the same time, the C channel is switched on. This period is called the CP
contact time. The contact time lasts from 1 ms to several ms. During the contact
time, the Hartmann-Hahn match condition:
Y hB jh ~ YcBic
13 1
is met, where Y h and Yc is the magnetogyric ratio of C and H; B in and B ic are
13 1
the local magnetic fields generated by the C and H channels, respectively. A
rapid magnetization transfer from *H to 13C occurs during this contact. 13C spins are
thus polarized. Following the contact period there is normal acquisition o f FID as in
any other pulse NMR experiment. Theoretically, cross polarization can enhance the
13
C signal by a factor of 4 (Y h^ Y c ~ 4) in each scan. CP experiments also take less
time for the same number of scans than normal NMR experiments.
A very useful variation of normal cross polarization is called cross
polarization with interrupted decoupling (ID), also known as dipolar dephasing.
After the contact period, *H channel (this becomes the decoupler channel after the
contact period, so that the C-H dipolar-dipolar interaction is eliminated) is turned off
for about 50 (is then is turned on again when acquisition begins. During this 50 (is
dephasing period, strong C-H dipolar-dipolar interactions resume, resulting in very
1 3
broad lines for C signals on CH and CH2 carbons. In contrast, signals for
quaternary carbon or methyl carbons are not broadened by C-H dipolar-dipolar
interactions, because quaternary carbons have no *H attached to them, so that C-H
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
dipolar-dipolar interactions are weak, while C-H dipolar-dipolar interactions on
methyl carbons are greatly reduced by the rapid rotation of methyl group.
Compared to normal CP spectra, CH and CH 2 signals seem to be dramatically
reduced in a dipolar dephasing spectrum. This is very useful for peak assignment in
solid state NMR.
All of the NMR spectra in this dissertation were acquired with magic angle
spinning, and most with cross polarization.
1.4 Combined solid state NMR and GC study of heterogeneous catalysis
In situ solid state NMR has been widely used to study reaction
mechanisms.4 4 The catalyst is usually activated on a vacuum line at high
temperature, then a controlled amount of reagent is adsorped on the catalyst. If the
solid state NMR system is equipped with a variable temperature probe, the catalyst
is directly transferred into NMR rotor and heated to desired temperature. NMR
spectra can be acquired over a broad temperature range.45 The catalyst can also be
heated in a sealed glass ampoule in an oven for a measured time, and then the glass
ampoule is transferred into a NMR rotor to acquire NMR spectra.4 6 Many studies
have been done on MTO using in situ NMR .4 4 , 4 7 ,4 8 Figure 1.6 shows a device that
4 9
was developed in the Haw group for in-situ NM R studies, the CAVERN
(Cryogenic Adsorption Vessel Enabling Rotor Nestling) apparatus. This device can
conveniently transfer the activated catalyst into a NMR rotor and seal the rotor
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Vacuum
Line
Trap door.
Joint
Seal
M A S rotoi
e
M echanism for opening
trap door and driving
seal into rotor
Thermocouple insert
Figure 1.6 A CAVERN device for in situ NMR study
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
without exposing the catalyst to the atmosphere. One outstanding feature of the
CAVERN is that the adsorbates can be loaded on the catalyst at cryogenic
temperature. This feature is critical when the adsorbates are very reactive at ambient
temperature. This dissertation includes in situ NM R studies of the chemistry
carbenium ions on HZSM-5 and HSAPO-34 using CAVERN experiments in chapter
4 and chapter 6 . 3 9 ' 40
While CAVERN experiments have been successful, their shortcomings in
heterogeneous catalysis studies are also apparent. In CAVERN studies, the reactions
take place in a closed vessel, a sealed NMR rotor, so that the reactants and products
are immobile on the catalyst. The contact time o f reagents on the catalyst is
relatively long. An NMR spectrum usually takes several minutes or hours to
complete acquisition, and the spectrum acquired is in fact a time average of the
sample during the acquisition. Several NMR spectra with different pulse sequences
are usually acquired to get more information about the sample, so it takes a long
time to analyze a sample. A typical heterogeneous catalytic reaction takes place at
high temperature. Reagents are feed into the catalyst bed continuously with the
carrier gas, and products diffuse out continuously also. Product analysis is by on
line GC or GC/MS. The average contact time o f regents is only a few seconds or
shorter. In many heterogeneous catalytic reactions, the first several seconds o f the
reactions are o f much interest for mechanistic studies. An ideal NMR approach to
study heterogeneous catalytic mechanisms should be able to study the reaction under
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
standard reaction conditions. The product should be analyzed by GC or GC/MS at
various time during the reaction.
50 49 52
One approach to this goal is by Hunger, Haw and Munson. They each
developed a solid state NMR probes that can do flow reactions at high temperature.
In Hunger’s design, on-line GC is also employed. With GC analysis of the product
and solid state NMR detection o f the intermediates on the catalyst while the reaction
is on-going, this technology provides real time information on the reaction. On the
other hand, this approach compromises the application o f both solid state NMR and
the standard GC/flow reactor. For instance, to have better signal to noise ratio it is
preferable to acquire NMR spectra at low temperature, but here NMR spectra are
acquired at reaction temperature. In order to use magic angle spinning, the catalyst
bed has to be packed in a tubular shape with a hollow channel inside, thus the flow
patterns of the reactants and products are very different from normal reactor.
52
Haw et al. approached this problem in another way. Figure 1.7 shows a
diagram of their "pulse quench reactor". They used a standard bench-top flow
reactor, as is common used in traditional heterogeneous catalysis. The reaction takes
place under standard conditions. The carrier gas, normally He, flows through the
catalyst bed at controlled flow rate. The reactants are introduced into the catalyst bed
as a pulse, by means of a Valeo injector, to do pulse reactions. Reactants can also be
introduced by syringe pump or mass flow meter to do flow reactions. What makes
this reactor different from a conventional reactor are two additions. One is the
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Mass flow
controller
Injector
valve j
s
I Reacto
V 7
Helium gas
LLkJ w a r
n m
Quench
vent
Heat
exchanger
GC/GCMS
Nitrogen for
cooling
Figure 1.7 Schematics of the Pulse Quench Reactor
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
computer controlled valves that control the timing of each event during the reaction,
the other is a thermal quench unit to thermally quench the reaction in a very short
time. A typical experiment is as follows:
The catalyst is activated in the reactor with carrier gas. After catalyst
activation, the catalyst bed temperature is adjusted to the desired reaction
temperature. When it is time to introduce the reactants, for instance, by a Valeo
injector, the injector is switched on by the command from the computer, and the
reactants that are pre-stored in the sample loop are injected onto the catalyst bed by
the carrier gas. After a period o f reaction time, a GC sample is taken by the GC
sampling valve, which is also controlled by the computer. At the planned time to
quench the reaction, nitrogen gas pre-cooled to near liquid nitrogen temperature is
blown through the catalyst bed. The catalyst bed temperature decreases by 100 °C
during the first one tenth second o f quenching, and to ambient temperature in
seconds. The reactor is then sealed and put into a glove box. The catalyst sample is
transferred to a NMR rotor inside the glove box, and the catalyst sample is never
exposed to the air. NMR spectra are then acquired at room temperature.
Two or more reactants could be pulsed on the catalyst in a desired order.
These multi-pulse experiments enable researchers to pre-treat the catalyst with some
reagents, and to study the effect o f this pretreatment on the latter reagents' chemical
behavior. Besides pulse-quench type reactions, flow-quench reaction can also be
done; the reactants are introduced in as a flow, and the flow reaction is quenched in
steady state. Figure 1.8 illustrates the time events during the reactions.
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Single pulse-quench experiment
Pulse
Quench
Double pulse experiment
Pulse 1 Pulse 2
Quench
Flow quench experiment
Flow Quench
Figure 1.8 Three kinds of quench experiments
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The reaction can be quenched at any time. If we do a series of experiments
under identical conditions, except the time to quench, we will effectively have a
series of "snap shots" of the reaction, giving us a detailed picture of the reaction.
In this dissertation, MTO studies on HZSM-5 and SAPO-34 were all
carried out using pulse quench reactor.
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.5 References and notes
(1) Thayer, A. M. Chem & En. News 2000, 78, 19-24.
(2) MacKenzie, J. J. World Resource Institute home page.
http://www.igc.org/wri/climate/im oil_001.html (Acessed Jan. 2001)
(3) Chang, C. D.; Silvestri, A. J. J. Catal. 1977, 47, 249.
(4) Meisel, S. L. Chemtech. 1988, 1, 32.
(5) Thomas, J. M.; Thomas, W. J. Principles and Practice o f Heterogeneous
Catalysis', VCH: New York, 1996; p 515.
(6) Stocker, M. Microporous Mater. 1999, 29, 3-48.
(7) Chang, C. D.; Bibby, D. M. Methane conversion', Elsever: Amsterdam,
1988; p 127.
(8) Van der Berg, G. H.; Wolthuizen, J. P.; van Hoff, J. H. C. Proceedings 5th
International Zeolite Conference', Heyden: London, 1980; p 649.
(9) Olah, G. A. Pure Appl. Chem. 1981, 53, 201.
(10) Chang, C. D.; Silvestri, A. J. J. Catal. 1977, 47, 249.
(11) Venuto, P. B.; Landis, P. S. Adv. Catal. 1968, 18, 259.
(12) Ono, Y.; Mori, T. J. C. S. Faraday Trans. 1981,177, 2209.
(13) Nagy, T. B.; Gilson, J. P. Derouane, E. G. J. Mol. Catal. 1979, 5, 393.
(14) Chang, C. D.; Hellring, S. D.; Pearson, J. A. J. Catal. 1989, 115, 282.
(15) Sanchez del del Campo A. Ee.; Gayubo, A. G.; Aguayo, A. T.; Tattio, A.;
Bilbao, H. Ind. Eng. Chem. Res. 1998, 37, 2336-2340.
(16) Dahl, I. M.; Kolboe, S. Catal. Lett. 1993, 20, 149.
(17) Dahl, I. M.; Kolboe, S. J. Catal. 1994, 149, 458.
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(18) Haw, J. F.; Nicolas, J. B.; Song, W.; Deng, F.; Wang, Z.; Xu, T.;
Heneghan, C. S. J. Am. Chem. Soc. 2000, 122, 4763-4775.
(19) Song, W.; Haw, J. F.; Nicholas, J. B.; Heneghan, C. S. J. Am. Chem. Soc.
2000, 122, 10726-10727.
(20) Song, W.; Fu, H.; Haw, J. F. J. Am. Chem. Soc. 2000, in press.
(21) Song, W.; Nicholes, J. B.; Haw, J. F. J. Am. Chem. Soc. 2000, submitted.
(22) Strictly, zeolites are aluminosilicates, either natural or synthetic. All
other materials with the same topology as one of the zeolites are called
zeotypes. In this dissertation, to simplify discussion, zeotypes are called
zeolites too.
(23) Weitcamp, J. In Proceedings of the Ninth International Zeolite Conference,
Ballmoos, R. V.; Higgins, J. B.; Treacy, M. M., Eds.; Butterworth-
Heinemann: Montreal, 1992; Vol. 1, pp 13-45.
(24) Thomas, J. M.; Thomas, W. J. Principles and Practice of Heterogeneous
Catalysis, VCH: New York; 1996, p 362.
(25) Zhang, J. Ph. D. Thesis, Texas A&M University, 1999.
(26) IZA homepage, http://www.iza.org. (acessed Jan, 2001)
(27) Csicsery, S. M. Zeolites 1984, 4, 202-213.
(28) Kaiser, S. W. Arab. J. Sci. Engng, 1985, 10, 361.
(29) Dyer, A. in Encyclopedia o f Inorganic Chemistry, King, R. B. Eds;
John Wiley&Sons: New Yory, 1990, p4381.
(30) Froment, G. F.; Dehertog, W. J. H.; Marchi, A. J. Catalysis, 1992, 9, 1.
(31) Hammett, L. P.; Deyrup, A. J. J. Am. Chem. Soc. 1932, 54, 2721.
(32) Farcasiu, D.; Ghenciu, A. J. Am. Chem. Soc. 1993, 115, 10901.
(33) Haw, J. F.; Nicholas, J. B.; Teng, X.; Beck, L. W.; Ferguson, D. A. Acc.
Chem. Res. 1996, 29, 259.
(34) Oliver, F. G.; Munson, E. J.; Haw, J. F. J. Phys. Chem. 1992, 96, 8106.
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(35) Xu, T.; Haw, J. F. J. Am. Chem. Soc. 1994, 10188.
(36) Maciel, G. E. in Heterogeneous catalysis, Shapiro B. L. Eds; Texas A&M
University press: College station, 1984, pp 349-381.
(37) Xu, T.; Barich, D. H.; Goguen, P. W.; Song, W.; Wang, Z. Nicholas, J. B.;
Haw, J. F. J. Am. Chem. Soc. 1998, 120, 4025.
(38) Haw, J. F.; Nicholas, J. B. J. Am. Chem. Soc. 1998, 120, 11804.
(39) Song, W.; Nicholas, J. B.; Haw, J. F. J. Phys. Chem. 2000, in press.
(40) Song, W.; Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 2001,123, 121-
129.
(41) Wind, R. A. Solid State NMR of Spin-1/2 Nuclei. In M odem NMR
Techniques and Their Application in Chemistry; Popov, A. I., Hallenga, K.,
Eds.; Practical Spectroscopy Series Volume 11; Marcel Dekker: New
York, 1991, p 125.
(42) Andrew, E. R.; Bradbury, A.; Eades, R. G. Nature, 1959, 183, 1802.
(43) Pines, A.; Gibby. M. G.; Waugh, J. S. J. Chem. Phys. 1973, 59, 569.
(44) Xu. T.; Haw, J. F. Topics in Catal. 1997, 4, 109-118.
(45) Haw, J. F.; Richardson, B. R.; Oshiro, I. S.; Lazo, N. D.; Speed, J. A. J.
Am. Chem. Soc. 1989, 111, 2052.
(46) Anderson, M. W.; Klinowski, J. Nature, 1989, 339, 200.
(47) Anderson, M. W.; Klinowski, J. J. Am. Chem. Soc. 1990, 112, 10.
Bosacek, V. J. Phy. Chem. 1993, 97, 10732.
(48) Munson, E. J.; Kheir, A. A.; Lazo, N. D.; Haw, J. F. J. Phys. Chem. 1992,
96, 7740.
(49) Haw, J. F. Topics In Catal. 1999, 8 , 81-86.
(50) Hunger, M.; Seiler, M.; Horvath. T. Catal. Lett. 1999, 57, 199-204.
(51) Carlson, L. K.; Isbester, P. K.; Munson, E. J. Solid State NuclearMagnetic
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Resonance, 2000, 16, 93-102.
(52) Haw, J. F.; Goguen, P. W.; Xu, T.; Skloss, T. W.; Song, W.; Wang, Z.
Angew. Chem. 1998, 37, 948.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 2
SAPO-34 synthesis and characterization
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.1 SAPO-34 syntfae
The SAPO-34 catalyst used in this dissertation was synthesized according to
the hydrothermal route in the original patent. 1 For each batch o f SAPO-34, 30.5 g
aluminum isopropoxide (98% Al(0 -i-C3Hv)3, Aldrich), 17.5 g phosphoric acid (85%
H3PO4 , Aldrich), 2 g fumed silica (95% Si0 2 , Aldrich, 100 mL tetraethylammonium
hydroxide (TEAOH) solution (35% TEAOH in water, Aldrich), and 50 mL distilled
water were used as starting materials. Figure 2.1 illustrates the procedure of
synthesis.
The first step was alum inum isopropoxide hydrolysis. Aluminum
isopropoxide was mixed with water in a 500 mL single neck flask equipped with a
magnetic stirrer and a condenser. The mixture was heated to 50 °C to accelerate the
hydrolysis. The hydrolysis was completed in two hours with strong stirring,
resulting in a homogeneous white gel. Isopropyl alcohol, the product of the
hydrolysis, was left in the gel without further treatment. The second step was to mix
phosphoric acid with the white gel. Phosphoric acid was added to aluminum gel in 5
minutes with stirring. The resulting white homogeneous gel was part A of the final
mixture. While part A was stirred, the second step was to prepare part B of the final
mixture. In a 200 mL beaker, TEAOH solution was mixed with fumed silica. The
mixture was stirred till homogeneous. In the fourth step, part B was added to part A
with stirring. The final mixture was stirred for additional two hours, resulting in a
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A 1 ( 0-i-C3H7)3,30.2 g
add 50 mL H2O
Stir for 3 hrs.
Homogeneoues mixture
(^TEAOH (35%), lM m L ^ )
add 17.5 g H3PO4 (85%),
Stir for 1 hr.
Part A
As-synthesized SAPO-34
Wash, dry
add 2 g fumed silica
Stir 30 min.
partB
final mixture
stir for 1 hr.
Transfer to autoclave,
keep at 200 °C for
48hrs.
Figure 2.1 SAPO-34 synthesis procedure
40
semi-transparent homogeneous white gel. The composition of the final mixture is
(SiO2)3.3(AlO2)i8(PQ2)i5(H2O)60 With TEAOH.
The final step was hydrothermal crystallization using a Parr 4086 high
pressure reaction system. The final mixture was transferred into an autoclave
equipped with a 200 mL Teflon liner and pressure gauge. The autoclave was sealed
and maintained at 200 °C for 48 hours without stirring. The pressure inside the
autoclave was the autogeneous pressure of water at 200 °C.
The autoclave was slowly cooled down after 48 hours. As-synthesized
SAPO-34 was recovered from the autoclave, washed three times with 200 mL water
and dried at 120 °C for 24 hours. Usually 16 g white powder was recovered after
drying, which represents about 80 wt% yield from starting material. As-synthesized
SAPO-34 was calcined in a furnace at 600 °C for 10 hours.
2.2 SAPO-34 characterization
2.2.1 Characterization method. SAPO-34 samples were characterized by the
following methods:
Elem ental analysis. Elemental analysis was performed by Galbraith
Laboratories, Inc. to determine silicon, aluminum and phosphorous contents.
X-ray diffraction. XRD patterns were acquired on a Rigaku DMAX/B
diffactomter with Cu Koc radiation. The X-ray source was an anode operating at 50
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
kV and 180 mA with a Cu target (no filter). XRD data were collected between 5°
and 55° in terms of 20 with a 0.04° step size and l°/min scan rate.
NMR. NMR spectra were acquired on a modified Chemagnetics CMX-300
spectrometer operating at 299.7 MHz for ^ 78.1 MHz for 2 7Al, 121.3 MHz for 31P
and 75.4 MHz for 13C. Acetone (2.11 ppm), A 1(NC> 3 )3 solution (0.1 M, 0 ppm),
H3PO 4 (85%, 0 ppm) and solid hexamethylbenzene (17.4 ppm), respectively, were
1 13
used as external chemical shift standards,. H and C chemical shifts are reported
27 31
relative to TMS. Al chemical shifts are reported relative to A 1(NC> 3 )3 and P
chemical shifts are reported relative to H 3PO4 .
Results
Elemental analysis. Table 2.1 lists the elemental analysis results for a
representative sample. From this result, the mole ratios of silicon, aluminum and
phosphorous were calculated as: Si : Al : P = 1: 5.32 : 4.33. Further calculations
show that oxygen accounts for 44.3 wt % of the framework and the framework
accounts for 81.2 wt % of the sample. Crystalline water accounts for the rest of the
sample weight at 18.8 wt %. Thus, the experimental SAPO-34 unit cell composition
is determined to be: H 3.3gSi3.38Ali7.98Pi4 .6 4 0 7 2 . 3 4 H2O. This is very close to the
composition of the starting material. The acid site concentration is calculated from
this composition to be 1.17 mmol/g sample.
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 2.1. Elemental analysis result of SAPO-34
Element Weight Content %
Si 3.76
Al 18.67
P 17.45
N 0.15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
X-ray diffraction. Figure 2.2 shows the XRD pattern o f a representative
sample. It is consistent with that o f a standard chabazite sample. X-ray diffraction
proves that the SAPO-34 sample is purely crystalline with CHA topology.
Solid State NMR. Figure 2.3 shows 1H, 2 7 A1, and 31P spectra o f a
representative calcined SAPO-34 sample. A single peak at 3.9 ppm in the
spectrum shows that Bronsted acid sites on the SAPO-34 framework are uniformly
distributed. A *H signal from silanol groups is not present in the spectrum of
27 27
SAPO-34. The -30 ppm peak in the Al spectrum is typical for Al in tetrahedral
27
environment. No signal due to extra-framework Al (usually at -80 ppm) is present.
3 1
The -28 ppm peak in the P spectrum shows that all phosphorous atoms are in a
uniform environment. The NMR results further suggest that the SAPO-34 sample
13
has a uniform framework structure. Figure 2.3 reports C spectra o f the as-
synthesized SAPO-34 sample and the calcined SAPO-34 sample. Two sharp peaks
at 9 and 55 ppm for as-synthesized SAPO-34 are from the template agent,
1 3
tetraethylammonium hydroxide (TEAOH), used in SAPO-34 synthesis. No C
signal was detected on calcined SAPO-34, showing that all template was removed
from SAPO-34 after calcination.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
*n
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 2 .2 X-ray pattern o f SAPO-34
I I I i T | l l l 'r'l 1 11 ! |............"|.iT T T jlT T T 'l H 1 S[ I i i I (
4 0 3 0 2 0 10 0 -1 0 -2 0 -30 -4 0
ppm
p m
200 120 4 0 -4 0 -200
-28
ppm
3 1 p
* s is
ITT
2 0 10 0 -1 0 -20 -3 0 -4 0 -50 -60 -70 -8 0
ppm
Figure 2.3 *14, 27Al and J1P spectra of SAPO-34
31t
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Calcined SAPO-34
As-synthesized
SAPO-34
70 60 50 40 30 20 io
0 -10 -20 80
ppm
Figure 2.4 75.4 MHz 1 3 C Spectra of as-synthesized and calcined SAPO-34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
References and notes
(1) Lok, B. M.; Messina, C. A.; Patton, R. L.; Gajek, R. T.; Cannan, T. R.;
Flanigen, E. M. U.S. Patent 4, 440, 871, 1984.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 3
Roles for Cyclopentenyl Cations in the Synthesis of
Hydrocarbons from Methanol on Zeolite Catalyst HZSM-5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.1 Introduction
1 2
Methanol-to-gasoline (MTG) chemistry ’ on zeolite HZSM-5 has
a
motivated more fundamental study than any other mechanistic problem in
heterogeneous catalysis. There is also much current interest in the closely related
process of methanol-to-olefm (MTO) chemistry, which is also catalyzed by HZSM-
5 using different process conditions4 to minimize aromatic formation. Dehydration
of methanol to dimethyl ether (DME) is rapid, and identical hydrocarbon
distributions are obtained with either oxygenate. From a fundamental perspective,
MTG and MTO processes on HZSM-5 pose the same mechanistic questions: On a
working catalyst, how is methanol (or DME) converted into olefins, why is propene
formed with higher initial selectivity than other olefins, and, how are aromatics
synthesized under MTO/MTG conditions?
MTG/MTO chemistry has been intensely studied using theoretical
methods, 5" 11 infrared spectroscopy12 , 13 and N M R . 1 4 "21 At least 20 distinct
mechanisms have been proposed for the first carbon-carbon bond-forming step, and
most of these predict that ethylene is the first hydrocarbon product. It is usually
assumed that propene and other higher olefins arise by homologation of ethylene
with methanol or dimethyl ether (DME). However, a key feature of the reaction is a
kinetic induction period that precedes large-scale hydrocarbon synthesis, and the
reaction occurring during the induction period need not be the same as that on a
working catalyst. Several workers have applied classical flow-reactor methods to the
study of the induction period. In a previous series of papers that is very pertinent to
the present investigation, Kolboe studied the effect of prior introduction of olefin
22-25
precursors on the subsequent conversion of methanol. He found that treated
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
catalysts were more active than untreated catalysts and proposed a
phenomenological "carbon-pool" model to account for reactions on working
catalysts. Kolboe did not attempt to characterize the structure of the carbon pool
species, but he speculated that it might be a carbenium ion. In another series of
investigations, Mole and coworkers demonstrated that simple aromatic hydrocarbons
26 27
functioned as “co-catalysts” for methanol conversion. ’ They proposed a
mechanism by which exo-methylene species, in equilibrium with protonated methyl
aromatics (e.g., benzenium cations), undergo electrophilic addition to form
ethylbenzenes which eliminate ethylene in acidic media. An alternative mechanism
involving well-established six- to five-membered ring interconversions was also
28
proposed to account for the findings of Mole et al.
Here we report extensive experimental work using a pulse-quench catalytic
21 29 31
reactor ’ ’ to probe the transition between induction reactions and hydrocarbon
synthesis on a working catalyst. NM R characterization of quenched catalyst
32
samples, in combination with theoretical calculations, reveal that formation o f the
1,3-dimethylcyclopentenyl carbenium ion (1) on the zeolite ends the induction
period for MTO chemistry. The active site for hydrocarbon synthesis on a working
catalyst is a composite of cyclic organic species, on which carbon-carbon bonds are
made and broken in a catalytic cycle, and one or more Bronsted acid sites. We find
that 1 is synthesized in less than 0.5 s with high selectivity when ethylene is pulsed
onto the zeolite at 623 K. Variable temperature MAS NMR experiments reveal that
I is present on a sealed catalyst sample at 523 K. We measured the half-life of 1 in
the catalyst bed in the presence of carrier gas flow and obtained values of 1 0 min. at
548 K and 6 s at ca. 623 K. In the absence of 1, olefin synthesis from DME exhibits
a clear kinetic induction period. If 1 is first synthesized in the zeolite from natural
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
abundance ethylene, olefin synthesis from DME- C2 proceeds without an induction
period, and labels are incorporated uniformly into 1. The induction period
returns if the time interval between injection of ethylene and DME is greater than
several times the half-life of 1. We also show that 1 is an intermediate in the
synthesis of toluene, an important product in MTG chemistry.
1
Theoretical chemistry was used to further clarify several aspects of the
experimental work. We explored the mechanism by which isolated cation 1 is
converted to toluene by calculating the relative energies (MP2/6-311+G*) of various
possible species that might exist on the reaction pathway. We optimized the gas
13
phase geometry of 1 and calculated the C chemical shift tensors using the gauge-
including atomic orbital (GIAO) method at the MP2 level. The predicted isotopic
chemical shifts are in good agreement with those measured for 1 in the zeolite.
Density functional theory (DFT) optimizations at the B3LYP/6-311G** level
confirmed that 1 forms a stable ion pair complex with a large cluster model o f the
HZSM-5 conjugate base site. We also found a stable neutral n complex of the cyclic
diene 2 formed by deprotonation of 1 and the Bronsted site of the catalyst that was
only 2.2 kcal/mol higher in energy than the ion pair. A third species, a framework
alkoxy produced by forming a covalent bond between 1 and an oxygen on the
conjugate base site of the catalyst, is also a stable point on the potential energy
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
surface, but the relative energy of this species is so high as to suggest no role in the
observed chemistry.
Finally, theoretical chemistry was also applied to understand how formation of
1 , 2 and related species could end the kinetic induction period and catalyze
hydrocarbon synthesis. We calculated the reaction pathways for a number of
methylation reactions in the gas phase. For example, methylation of ethylene or
propene with trimethyloxonium to yield protonated DME and propene or 1-butene,
respectively, has predicted barriers greater than 45 kcal/mol. In contrast, methylation
of cyclic diene 2 with trimethyloxonium formed DME and carbenium ion 3 with a
barrier o f only 33 kcal/mole. The aggregate experimental and theoretical evidence led
to a proposed catalytic cycle for MTO chemistry on HZSM-5 which also accounts
for propene selectivity. Similar schemes can also be constructed for alkylation and
olefin elimination from aromatics as a parallel or competing pathway once these are
formed from cyclopentenyl cations.
3.2 Experimental Section
M aterials and Reagents. All results reported were obtained on a zeolite
HZSM-5 catalyst with a Si/Al ratio of 14 that was pressed into pellets (10 to 20
mesh) without binder. Some of the experiments were repeated using a zeolite with a
Si/Al of 19 as pellets with 30% by weight alumina binder. No obvious differences
were observed for the two materials studied. Ethylene- 1 3C2 (99% 1 3C), propene-1-
2 3
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13C (99% 13C), propene-2-13C (99% 1 3C) and 2 -methyl-propene-2 - 13C (99% 1 3 C)
were obtained from Cambridge Isotopes. Dimethyl ether (DME)-1 3C2 , (99% 13C)
was obtained from Isotec.
Catalysis Experim ents. We used a pulse-quench reactor to study
hydrocarbon synthesis reactions. For each experiment we loaded the reactor with a
cylindrical bed (7.5 mm diameter by 8 mm in length, typical weight of 0.3 g) of fresh
catalyst pellets. In every case, the catalyst bed was activated in place immediately
prior to use by heating at 573 K for 1 hr in flowing helium. DME pulses were 0.46
moles per mole o f acid site, (typically 0.15 mmol), and ethylene pulses were 1.9
moles per mole of acid site (typically 0.63 mmol). The pulse-quench reactor is
essentially a fixed bed, micro flow reactor. Helium (600 seem) was used as the carrier
gas, and reactants were introduced in pulses using injection valves as in references 2 1 ,
29 and 30. Schematics of a pulse-quench reactor in a single-pulse configuration can
be found in Figure 1 of ref. 29. The double-pulse experiments were of the following
type: Pulse 1, Xi, Pulse 2, X 2> Quench. Pulse 1 was the first compound delivered to
the catalyst bed, usually ethylene, and the treated catalyst was allowed to age for
time Xi prior to delivery of a second reagent, frequently DME or methanol, in Pulse
2. Further aging of the catalyst occurs during the time interval % 2- Following X2 the
catalyst bed temperature is rapidly decreased to ambient using a thermal quench.
Previous studies have shown that the temperature of the catalyst pellets decreases
150 K in the first 170 ms of a quench. In some cases we used N 2 as one of the
reagents for control experiments that precisely matched other double-pulse
experiments. Single pulse experiments were of the type: Pulse 1, Xb Quench. In
both double-pulse and single-pulse experiments the gas stream exiting the reactor was
sampled for gas chromatographic or GC-MS analysis. After quenching each reacted
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
catalyst sample, the reactor was sealed off and transferred into a glove box filled with
nitrogen (0.3 ppm H2 O). The catalyst pellets were ground and transferred to a 7.5-
mm MAS rotor which was sealed with a Kel-F end-cap. This procedure affords a
sealed sample for MAS NMR study that has not been exposed to atmospheric
moisture.
NMR Spectroscopy. 13C solid state NMR experiments were performed
with magic angle spinning (MAS) on a modified Chemagnetics CMX-300 M Hz
spectrometer operating at 75.4 MHz for 1 3C. Hexamethylbenzene (17.4 ppm) was
used as an external chemical shift standard, and all 13C chemical shifts are reported
relative to TMS. Chemagnetics-style pencil probes spun 7.5 mm zirconia rotors at
typically 6.5 kHz with active spin speed control (± 3 Hz).
Typical 13C experiments included: cross polarization (CP, contact time = 2
ms, pulse delay = 1 s, 2 , 0 0 0 to 16,000 transients); cross polarization with
interrupted decoupling (contact time = 2 ms, pulse delay = 1 s, 2 , 0 0 0 to 16,000
transients, dipolar dephasing time o f 50 ps); single pulse excitation with proton
decoupling (Bloch decay, pulse delay = 10 s, 400 to 4000 transients). All the spectra
shown were obtained with cross polarization, except where otherwise stated.
Quantitation of Cation 1 in Catalyst Samples. The absolute yields of cation
1 in several catalyst samples prepared using the pulse-quench reactor were
determined by both spin counting and elemental analysis. A calibration of 13C NM R
signal intensity vs. carbon content was obtained for measurements o f five samples of
hexamethylbenzene (natural abundance) in sulfur. All 13C NMR spectra for spin
counting experiments were measured using Bloch decays with 10 s pulse delays. The
number of 13C spins corresponding to one or more of the prominent signals of 1 in
the catalyst samples was deduced by comparison to the standard curve. This
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
measurement was then used to calculate the number of 1 cations per unit measure of
zeolite. For several samples, confirmatory evidence was obtained by carbon analysis
(Galbraith). Spin counting and elemental analysis results were in good agreement.
3.3 Theoretical Methods
Several different types of theoretical methods were employed in this
investigation. We first needed to verify the identity o f the cyclopentenyl species 1.
To this end we optimized the geometry of 1 using second order Moller-Plesset
33
theory (MP2) and the 6-311+G* basis set. The core electrons were not included in
the correlation treatment. Chemical shieldings for 1 were then calculated using the
GIAO method at the MP2 le v e l.3 4 We used Ahlrich's tzp basis set35 {51111/311/1}
(with 6 Cartesian d orbitals) on the carbon and dz {31} on the hydrogens in the
chemical shift calculations. We term this level of theory GIAO-MP2/tzp/dz.
The calculated chemical shielding tensors were symmetrized and then
diagonalized in order to yield principal components. These were then referenced to
the 13c isotropic chemical shift of carbon in tetramethylsilane (TMS) calculated at
the same level of theory (for both the shielding and geometry) such that Scale ~
S tM S - 8 C alc- The absolute shieldings of in TMS is (in ppm) 199.0 at
MP2/tzp/dz//MP2/6-311+G*. The isotropic chemical shift is the average o f the
principal components, which are defined such that Si \ > 8 2 2 > 8 3 3 . Thus,
8 iso ~ 1/3 (8 1 1 + 8 2 2 + 8 3 3 )
The asymmetry factor (r|) and chemical shift anisotropy (CSA) are defined as
previously in ref. 36.
It is also important to theoretically verify if 1 can exist as a stable cation in
the presence of the conjugate base of the zeolite acid site. To this end we optimized 1
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
complexed to a model of a deprotonated zeolite acid site in HZSM-5. Specifically,
HZSM-5 was modeled as a cluster of stoichiometry [(H3 SiO)3 SiOAl(OSiH3 )3 ]~.
Due to the large size of the system we were unable to use the MP2 method for the
geometry optimization. We thus used DFT, which provides geometries and energies
comparable to MP2 calculations at a considerably reduced computational cost. For
37
the DFT optimizations we used the hybrid B3LYP exchange-correlation functional
and the 6-311G** basis set. In all the optimizations the terminal silyl (H3Si-) groups
were held fixed in crystallographic positions corresponding to the T(12)-0(24)-T(12)
38
site in HZSM-5. The positions of all other atoms were allowed complete freedom.
The constraints imposed in the optimizations impart errors in the calculated
thermodynamic properties, we thus report only changes in energy (AE) rather than
enthalpy. We also identified two additional adsorption complexes.
We used theoretical methods to explore the mechanism for conversion o f 1 to
toluene. Each of the possible reaction intermediates was first optimized at the
B3LYP/6-311G* level o f theory. Frequency calculations were done to verify that
each optimized geometry was a true energy minimum (no negative eigenvalues). The
B3LYP/6-311G* geometries and force constants were then used as input to MP2/6-
311+G* optimizations.
Finally we calculated at the B3LYP/6-311G* level the reaction pathways for
the methylation of various species as well as for elimination reactions leading to
olefin products. All the optimizations and frequency calculations were done with
Gaussian 94.39 Chemical shift tensors were calculated with the program ACES II.40
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.4 Experim ental Results
In the pulse-quench studies reported here, the NMR spectra were measured
at room temperature, and each spectrum corresponds to a unique catalyst sample.
Catalyst samples were discarded after a single use, and all experiments began with
freshly activated catalyst. In this investigation, catalyst samples have "histories"
only within a given double-pulse experiment, when Pulse 1 could influence the
outcome of reactions following Pulse 2.
Single-pulse experiments. We previously used a pulse-quench reactor to
observe the induction period in the conversion of methanol to hydrocarbons on
21
zeolite HZSM-5. The first step in that reaction, equilibration of methanol with
DME and water, is exothermic and not relevant to the mechanism of hydrocarbon
synthesis. For this investigation we used DME as the starting material to avoid this
pre-equilibrium. Figure 3.1 presents 13C MAS NMR spectra from a pulse-quench
study of the reactions of DM E-1 3C2 on HZSM-5 catalyst pellets. DME (60 ppm)
does not react at all on fresh catalyst for the first 2 s at 573 K; between 4 and 8 s
there is a very modest conversion to form small amounts of cation 1 (peaks assigned
below), after which large scale hydrocarbon synthesis commences. GC or GC-MS
analysis of the hydrocarbon products exiting the reactor invariably show a
preponderance of propene relative to other olefins and, in particular, high propene to
41
ethylene ratios. High selectivity for propene has previously been observed. The
results in Figure 3.1 are consistent with the kinetic induction period well-known to
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
147
♦
A x
I — ...........
48
249
24
16 s
48
v Mh *m « W V m iA s w
^ iTTir i~*~......................
48
2 s
60
51
-50 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0
ppm
Figure 3.1. 75.4 MHz 13C CP/MAS NMR spectra of the reaction products of
dimethyl ether-13C2 (DME) retained on zeolite HZSM-5 following for various times
at 573 K in a pulse-quench reactor. In each case, 7.2 mg of DME, corresponding to
ca. 0.46 molecules DME per acid site, was injected onto a freshly activated, previously
unused catalyst bed. Signals due to DME (60 ppm), a trace of methanol (51 ppm) and
cation 1 are highlighted. All spectra were measured at 298 K. The asterisks denote
spinning sidebands.
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
MTG chemistry. By similar experiments, we determined that the induction period
for DME on HZSM-5 was > 16 s at 548 K and ca. 0.5 s at 623 K.
We used the pulse-quench reactor to study the reactions o f ethylene-13 C2 on
the catalyst. The heats of adsorption and reaction of ethylene are significant and
must be considered in evaluating the reaction temperature. The pulse-quench reactor
was equipped with a high-speed (5 ms) thermocouple that was placed in contact
with the catalyst pellets 4 mm from the front of the catalyst bed. Under steady-state
reaction conditions the temperature of the catalyst bed is well regulated; this is not
possible during high concentration pulses of reactive species. Briefly, the arrival of
the ethylene pulse onto the catalyst pellets is heralded by a ca. 100 K temperature
increase which persists for several seconds. We also found that 0.15 mmol DME
pulses resulted in ca. 20 K temperature transients. In the following we distinguish
between the temperature of the catalyst bed prior to (or several seconds after) an
ethylene pulse, and a transient temperature ca. 100 K higher immediately following
an ethylene pulse, whenever the distinction may be important.
Figure 3.2 reports 13C MAS NMR results for single-pulse experiments in
which ethylene was exposed to the catalyst at an initial temperature of 623 K (the
temperature during the first few seconds of reaction rose to ca. 723 K) and allowed
to react between 0.5 s and 16 s prior to quenching. The most prominent peaks in the
spectrum obtained for 0.5 s of reaction are all due to 1. These are the isotropic
resonances at 250 ppm, C l and C3 of the allyl cationic moiety; 148 ppm, C2 of the
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
129
A
19
129
48
48
250 ) 24
xS X i'
148
250
0 .5 s A
148
5 0 -5 0
ppm
Figure 3.2. 13C CP/MAS NMR spectra of the reaction products of ethylene-13 C2
retained on zeolite HZSM-5 following various times at 623 K in a pulse-quench
reactor. Signals from cyclopentenyl cations (250, 148, 48 and 24) and toluene (129
and 19) are indicated in the spectra. All spectra were measured at 298 K. The
asterisks denote spinning sidebands.
61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
allyl cationic moiety; 48 ppm, CH2 carbons; and 24 ppm, CH 3 carbons. These
assignments were verified by interrupted decoupling (also called dipolar dephasing)
experiments. They are also consistent with previous studies of cation 1 in acidic
solutions and our theoretical studies (vide infra). As the catalyst ages for 2 to 4 s,
signals due to 1 decrease with a commensurate increase in signals due to toluene at
129 and 19 ppm. With further aging in the flow reactor, signals due to toluene and
other organics diminish, and after 16 s only a modest amount o f 1 remains in the
catalyst bed. A semi-log fit of the decrease of 1 over time yielded an approximate
half-life of 6 s. We carried out an analogous experiment at a nominal temperature of
548 K and found that the half-life of cation 1 in the catalyst bed was ca. 10 minutes
at the lower temperature.
Figure 3.3 reports a series of ethylene single-pulse experiments in which the
reaction time was fixed at 4 s, but the initial reaction temperature was varied between
323 K and 773 K. A remarkable reaction occurred at the lowest temperature studied,
323 K. The 13C NMR spectrum of this sample is reminiscent of those of long-chain
alkanes. Interrupted decoupling experiments established that the 12 ppm signal was
the only one due to methyl groups in this sample. The 33 ppm signal is consistent
with CH2 groups well removed from chain ends and the 24 ppm signal is
characteristic o f methylenes immediately adjacent to chain ends. This spectrum is
due to unbranched oligomers of ethylene, and, if no cyclic species are present,
integration suggests an average chain length of up to 40 carbons. When similar
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
773 K
ion- ____
673 K
129
573 K
148
250
255
155
245
473 K
48
24
48
19
24 250
129
48
148
148
373 K
250
24
323 K
-5 0
ppm
Figure 3.3. 13C CP/MAS NMR spectra of the reaction products of ethylene-13 C2
retained on zeolite HZSM-5 following 4 s of reaction at various temperatures
(indicated) in a pulse-quench reactor. Toluene (129 and 19 ppm) is one o f the
products formed from 1 at higher temperatures. All spectra were measured at 298 K.
The asterisks denote spinning sidebands.
63
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
experiments were conducted at higher temperatures, the unbranched oligomers were
replaced with more complex products including cation 1 and closely related species.
For example, the peak at 155 ppm has previously been assigned to the C2 carbons in
cations with substitution patterns similar to 4 .4 2 We also verified this assignment by
theoretical calculations (vide infra). At least three isotropic peaks are partly resolved
in the 245-255 ppm range for samples prepared at initial temperatures of 473 K or
573 K; this is further evidence of cyclopentenyl cations with diverse substitution
patterns. Accumulation of toluene in the catalyst bed is apparent for the reaction
carried out at an initial temperature o f 573 K. Essentially no organic species were
retained on the catalyst bed after 4 s of reaction at 773 K.
% 2 ,
4
GC traces from the experiments which also provided the NMR spectra
shown in Figure 3.3 are collected in Figure 3.4. Some of the ethylene passes through
the reactor without conversion to other products. For samples prepared at
intermediate temperatures (473 - 573 K) for which 1 and related cations were
apparent in the NMR spectra, the corresponding GC traces are dominated by C4 and
C5 products in approximately equal amounts. The high yield of C5 compounds is
clearly inconsistent with formation of primary products by a simple oligomerization
of ethylene. At higher temperatures (e.g., 773 K) the products are dominated by
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ethylene
c 3 benzene toluene
C4 C5 A y \ xylenes 7 7 3 K
673 K
573 K
473 K
I 1 373 K
. « ■ n* H—» .......... ......
j i8|Tp i i8j sgli p u j f | , 11111, |H | iJIT11 [• Ir|1 1 T T |
0 5 10 15 20 25 30 35 40 45 50
Retention Time (min)
Figure 3.4. GC traces (thermal conductivity) characterizing the volatile products
from the reactions of ethylene-1 3C2 on zeolite HZSM-5 (4 s reaction time) at various
temperatures in a pulse-quench reactor. Toluene is a major product at higher
temperatures.
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
propene and aromatics, most prominently toluene, with smaller amounts of benzene
and xylenes. Again, the high yield of an odd-carbon product, toluene, is significant.
We discovered that the formation of 1 and similar cations in flow reactors is a
general characteristic of simple olefins. Single-pulse experiments were performed in
which propene (two isotopomers) and 2-methylpropene were reacted for 4 s at 623
K. In every case aromatics were also formed and label scrambling in the cations
appeared to be quantitative. 13C NMR spectra from these studies are included as
Figure S3 in the supporting materials; they are similar in appearance to those from
analogous experiments using ethylene-1 3C2 -
We used NMR spin counting and elemental analysis to measure the absolute
yields of cation 1 in experiments representative of those reported here. In particular,
we studied samples prepared by pulsing ethylene-1 3C2 onto catalyst samples at 623
K followed by reaction between 0.5 and 60 s. Spin counting and elemental analysis
results were consistent; the yields of 1 were only 1 - 4 % of all acid sites,
corresponding to absolute yields o f up to one cation per ca. four unit cells or one
cation per ca. 16 channel intersections. This yield is low enough that we might
expect to see mass transport in the catalyst little affected by formation of 1. Using
standard vacuum line techniques, we measured benzene uptakes for both fresh
catalysts and catalysts reacted to form 1 and found them indistinguishable.
Double-pulse experiments. Double pulse experiments established that the
formation o f 1 by prior injection of ethylene had a profound effect on the subsequent
reaction of DME. Figure 3.5 reports experiments carried out at 548 K, a temperature
at which the induction period is much greater than 8 s and the half-life of 1 in the
catalyst is 10 min. Without prior injection of ethylene (Figure 3.5a) there is
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3.5. 13C CP/MAS NMR studies of the reactions of DME-1 3C2 with
cyclopentenyl cations at 548 K on zeolite HZSM-5 in a pulse quench reactor.
Spectrum a corresponds to the reaction of DM E-1 3C2 for 8 s on zeolite HZSM-5,
showing almost no hydrocarbon formation. Spectrum b corresponds to a double
pulse experiment; ethylene (natural abundance) was first injected and reacted for 60 s
on zeolite HZSM-5 to form cyclopentenyl cations, then DME-1 3C2 was injected and
reacted for 8 s. The yields of hydrocarbons, including cation 1, are much higher in
the second experiment. Spectrum c corresponds to a double pulse control experiment;
ethylene-13 C2 was first injected and reacted for 60 s, then N 2 was injected and the
catalyst was quenched after a further 8 s. The control experiment demonstrates that
cation 1 , formed from the injection of ethylene, is present on the catalyst long enough
to influence the conversion of DME 6 8 s later.
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60
a. (13CH3 )2 0
w i t 1 # 1 n.in ,iM -1„ n -^ *
51
b. c 2h 4 + (1 3c h 3 )2 o
A
c. 1 3C2H4
350 300 250 200 150 100 50 0 -50
ppm
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
negligible conversion of labeled DME after 8 s. Figure 3.5b reports a result obtained
by injecting a pulse of natural abundance ethylene, waiting 60 s, injecting labeled
DME, allowing a further 8 s of reaction followed, then quenching. In this case there
was significant conversion of DME as measured by hydrocarbon products in the GC
trace, the diminished DME peak in the NMR spectrum and the appearance o f 13C
labels in peaks for cation 1. Figure 3.5c is a control experiment which establishes
that 1 derived from ethylene in the first pulse is still present in the catalyst bed 6 8 s
later. Ethylene-1 3C2 was injected followed 60 s later by a pulse o f N 2 ; 8 s later the
sample was quenched. Signals due to 1, related cations and some toluene are readily
apparent in the spectrum. We carried out analogous experiments in which the
catalyst was allowed to age up to 1 0 minutes after injection of ethylene and prior to
injection of DME, and still observed that the induction period for hydrocarbon
synthesis from DME was eliminated by the presence of 1. Very significantly, when
we allowed the catalyst bed to age for 180 min. (many times the half-life of 1 ) the
induction period for reaction of DME returned as if ethylene had never been injected.
This experiment strongly supports the role of 1 and related species in hydrocarbon
synthesis on working catalysts in that removal o f the proposed cause eliminates the
observed effect.
Figure 3.6 reports an analogous set of experiments at a higher reaction
temperature, 573 K. After 8 s of reaction in a single-pulse experiment, DME was still
within the induction period on fresh catalyst, although traces of 1 had begun to form
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3.6. 13C CP/MAS NMR studies of the reactions of DME-1 3C2 with
cyclopentenyl cations at 573 K on zeolite HZSM-5 in a pulse quench reactor.
Spectrum a corresponds to the reaction of DM E-i 3 C2 for 8 s on zeolite HZSM-5,
showing almost no hydrocarbon formation. Spectrum b corresponds to a double
pulse experiment; ethylene (natural abundance) was first injected and reacted for 60 s
on zeolite HZSM-5 to form cyclopentenyl cations, then DME-1 3C2 was injected and
reacted for 8 s. The yields of hydrocarbons, including cation 1, are much higher in
the second experiment. Spectrum c corresponds to a double pulse control experiment;
ethylene-13 C2 was first injected and reacted for 60 s, then N 2 was injected and the
catalyst was quenched after a further 8 s. The control experiment demonstrates that
cation 1 , formed from the injection of ethylene, is present on the catalyst long enough
to influence the conversion o f DME 6 8 s later.
70
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a. (1 3 CH3 )2 0
b. C2 H4+ (1 3 CH3 )2 0
«... .
C . 1 3 C 2 H 4
p TTTp 111p rrrTp T | r -|,r r | ! I 1 1 I I I I 1 I I | I 1 I I I
350 300 250 200 150 100 50 0 -50
ppm
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(Figure 3.6a). When a larger amount of 1 was pre-formed by injecting natural
abundance ethylene and allowing the catalyst to age for 60 s, the subsequent
conversion of labeled DME after 8 s of reaction was greatly enhanced (Figure 3.6b),
13
and C labels from DME were incorporated into 1. In similar double-pulse
experiments we observed some conversion of DME with X j as short as 0.5 s. Figure
3.6c reports a control experiment demonstrating that 1 is indeed present in the
13
catalyst 6 8 s after injection of ethylene- Cj.
Stability of cation 1 in the catalyst. Most of the in situ NMR studies
carried out in our laboratory in the past decade have reacted adsorbates on catalysts
in sealed rotors heated or cooled in variable temperature MAS NMR probes. 4 3 While
this experimental approach does not simulate flow reactors or permit studies to be
carried out at short reaction time scales, it has one advantage in that the NM R
measurement can be made at higher temperatures, sometimes approaching those in
a.catalytic reactor. Since the pulse-quench reactor prepares samples at high
temperatures for NMR measurements at room temperature, we considered the
possibility that 1 is not present at high temperature but is the reversible reaction
product o f other species (e.g., olefins) trapped in the catalyst by the thermal quench.
We used the pulse-quench reactor to prepare catalyst samples loaded primarily with
1 (formed from ethylene-13 C2), sealed these in MAS rotors and then carried out
conventional variable temperature in situ studies. A representative result is reported
in Figure 3.7. These studies show that 1 is stable on the zeolite in a sealed rotor up
to at least 523 K; we observed some conversion to toluene, but most of the 1 was
72
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
background
473 K
3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 -5 0
ppm
Figure 3.7. Variable temperature 13C MAS NMR study of 1 on zeolite HZSM-5. A
single sample was prepared by pulsing ethylene-13C2 onto a catalyst sample at 623
K in a pulse-quench reactor and allowing it to react for 0.5 seconds. All spectra
(Bloch decays) were measured at the temperatures indicated on the figure. 13C
signals from Kel-F end caps sealing the MAS rotor are seen in the spectrum
measured at 523 K. Cation 1 is present on the catalyst at temperatures approaching
those used in the pulse-quench experiments.
73
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
unreacted after prolonged heating. The resonances of 1 were slightly broader at 523
K compared to room temperature, but we did not clearly observe scrambling or
exchange that was rapid on the NMR (chemical shift) time scale. This shows that 1 is
stable on the zeolite at a temperature only 25 - 50 K below those used in the double
pulse experiments. Cation 1 is not the reversible oligomerization product of
unrelated species, trapped or condensed in the zeolite pores during the thermal
quench.
We also designed a more direct experiment to control against the possibility
that 1 is formed by quenching other active species of unrelated structure (e.g., simple
olefins). We carried out a double-pulse experiment similar to that in Figure 3.5b with
the following difference. After pulsing ethylene onto the catalyst and allowing it to
react at a nominal temperature of 548 K for 60 s, we rapidly quenched the catalyst to
298 K and held it near room temperature in the flow reactor for 300 s. Had we
measured the NM R spectrum at this point we would have observed that 1 and
related cations accounted for nearly all o f the organic species in the catalyst bed.
Instead, we rapidly raised the reactor temperature back to 548 K, held it there for 60
s and then injected a pulse of dimethyl ether. As in the experiment described in
Figure 5b, the GC trace showed hydrocarbon synthesis without an induction period.
The thermal quench does not eliminate the catalytic effect.
Trapping studies were performed to identify the volatile products that exit
the catalyst bed during the decomposition of 1 in flow reactors. We achieved this by
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
adding a second zeolite bed downstream of the catalyst bed. The trapping bed
contained inactive NaY zeolite pellets, and it was cooled externally with liquid
nitrogen to promote condensation of organic species. 1 was generated by pulsing
ethylene onto HZSM-5 pellets at 548 K; the gases that immediately exited the
reactor were vented to avoid collection of products unrelated to the decomposition of
1 , but between 1 0 min. and 1 hr. following injection the effluent from the reactor was
flowed through the trapping bed. 13C NMR spectra of the NaY pellets were then
measured at room temperature; representative results are presented in the supporting
Figure S5. Those experiments showed that the decomposition products of 1 that exit
the reactor are propene, toluene and a smaller quantity o f ethylene. Note that this
implies that some fraction of the olefins that form cation 1 do so reversibly, although
stoichiometry precludes complete reversibility.
Aromatic Compounds as MTO Co-catalysts. Other workers have
previously reported that co-feeding toluene or other aromatic compounds enhances
26 27
the conversion of methanol to hydrocarbons on zeolite HZSM-5. ’ We considered
the possibility that other cyclic species (clearly distinct from 1 ) derived from
aromatic compounds could also function as organic reaction centers in M TO
chemistry. This proved to be the case. Experiments similar to those described here
confirmed that co-injection of large amounts of toluene accelerated the conversion of
methanol to hydrocarbons and also formed the pentamethylbenzenium cation 5.
Calculations suggest (vide infra) that methylation of toluene has a higher barrier than
for some of the species related to 1 , and we defer detailed discussion of the
conditions under which aromatic species might act to a later report (but see reference
30 for related experiments). In brief, we suspect that a pattern of alkylation and
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
olefin elimination reactions, very similar to those described below for cyclopentenyl
species, also operates on aromatic species, once they form at higher temperatures.
+ .*
5
3.5 Theoretical Results.
13C NMR Chemical Shifts of Cations 1 and 4. We optimized the
geometries of cations 1 and 4 at the MP2/6-311+G* level; cartesian coordinates for
these are given in the supporting materials. The C2v symmetry and bond lengths
obtained for cation 1 reflect the electron delocalization and partial double bond
character expected for this allylic cation. Cation 4 exhibits similar characteristics.
The lowest energy geometry of 4 has Ct symmetry due to methyl group rotations. In
order to decrease the computational cost of the GIAO-MP2 chemical calculations,
we obtained an optimized geometry for 4 in which the molecule was constrained to
Cs symmetry. The energy difference between the two conformers is negligible.
We report the GIAO-MP2/tzp/dz values of the isotropic 13C chemical shifts
for cations 1 and 4 in Table 1. The predicted chemical shift for C l and C3 in 1 is
255.3 ppm, which compares to the experimental value of 250 ppm. For C2, theory
predicts a chemical shift o f 152.8 ppm, whereas the experimental value is 148 ppm.
The CH2 carbons are predicted to have isotropic chemical shifts o f 50.0 ppm which
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
compare to the measured values of 48 ppm. Finally, the methyl carbons have
theoretical values of 28.9 ppm, whereas the experimental chemical shifts are 24 ppm.
In all cases the theoretical values are downfield of the experimental chemical shifts, in
the worst case by ~5 ppm. It is likely that the presence of the conjugate base of the
zeolite and motional averaging has some effect on the chemical shifts. Indeed, a
notable change in the geometry of the cation is observed theoretically (see below)
when it is complexed with a model of the zeolite anion. Although the agreement is
less exact than we would like, it is sufficient to verify the presence of 1 within the
zeolite.
Similar agreement between theory and experiment was obtained for cation 4.
13
The distinctive experimental C shift for C2 of cation 4, 155 ppm (see Figure 3.3),
compares with the theoretical value o f 163 ppm. The theoretical values for C l and
C3, 249.3 ppm, are also downfield o f the 245 ppm experimental value assigned from
Figure 3.3. Other signals in the experimental spectra are good matches for the
remaining theoretical shifts if the theoretical values are assumed to be ca. 5 ppm
downfield o f experiment, as for cation 1 .
Stability of 1 and Related Species on a Zeolite Cluster Model. We have
reported that only in rare cases do carbenium ions exist as persistent species in
zeolites (under typical conditions o f temperature and loading) . 4 4 Whereas we are
claiming in this report that 1 can persist in the zeolite catalyst HZSM-5, it is also
important to see if a reasonable theoretical treatment o f the system also predicts that
1 is stable in the presence o f the zeolite conjugate base. In Figure 3.8 we show the
B3LYP/6-311G** optimized geometry of 1 on a cluster model of HZSM-5. The
most acidic proton of 1 is in close proximity (2.04 A ) of the zeolite lattice, and could
be considered to be involved in a hydrogen bond with the negatively-charged zeolite
77
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 3.1. Theoretical and experimental 13C isotropic chemical shifts for cations 1
and 4. Theoretical values calculated at the GIAO-MP2/tzp/dz level of theory using
MP2/6-311+G* geometries and referenced to the chemical shift in TMS (199.0
ppm) calculated at the same level of theory.
Cation 1 Cation 4
Exp. GIAO-MP2 Exp. GIAO-MP2
C1-C3 250 255.3 245 249.3
C2 148 152.8 155 163.1
C4-C5 48 53.0 43 50.4
Methyl (C1-C3) 24 28.9 2 2 26.7
Methyl (C2) 1 0 12.4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3.8. B3LYP/6-311G** optimized geometry of the 1,3-
dimethylyclopentadienyl cation 1 coordinated to the zeolite anion (ion-
pair structure). The binding energy for the complex is -80.7 kcal/mol.
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
oxygen. The O-H-C involved in a hydrogen bond with the negatively-charged zeolite
oxygen. The O-H-C angle is 161.2°. The C-H distance for the proton interacting with
the oxygen is 0.008 A longer than the other C-H methylene bonds. Note that the
molecule exhibits distortions from that of the free cation. In particular, the C1-C2
(1.375 A) and C2-C3 (1.405 A) bond distances, which are equal in the free cation, are
different in the complex. Similarly the C1-C5 and C3-C4 distances differ by 0.013 A
and the bonds to the methyl groups differ by 0.027 A. The total Mulliken charge on
the cation is 0.84 |e|. Although there are certainly many other possible ways the
cation could orientate itself with regard to the zeolite, this optimization demonstrates
that cation 1 is stable over the zeolite anion in the most likely geometry that would
promote proton transfer back to the zeolite. The geometry of the complex also
supports the indication from the NMR calculation that the lattice has a measurable
effect on the cation. Considering the asymmetry imposed on the cation in the
complex, we must assume that motional averaging is sufficient to retain the
appearance of C2v symmetry in the experimental NMR spectra.
It is important to consider the energetics involved in the formation of 1 in the
zeolite (all values reported are AE's). Cation 1 could be formed in principle by the
protonation of the parent olefin 2. The protonation energy o f 2 is -229.4 kcal/mol at
the B3LYP/6-311G** level. This large exothermic contribution to the reaction
energy is countered by the energy needed to remove the proton from the zeolite. A t
the B3LYP/6-311G* level of theory the deprotonation energy of our zeolite model is
303.2 kcal/mol. If we just consider these two aspects of the reaction, the overall
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
reaction would be endothermic by ~74 kcal/mol, and thus would be very unlikely to
occur. However, upon protonation of the parent olefin and formation of 1, we also
gain the energy of the Coulomb attraction between 1 and the conjugate base of the
zeolite. The interaction energy of the ion pair is not insignificant; the Coulomb
attraction for two point charges separated by 3 A is ca. 1 0 0 kcal/mol. Indeed, the
calculated difference in energy between the ion pair complex and the sum of the
energies of the ions at infinite separation is -80.7 kcal/mol.45 Summing the energy
contributions (303.2 - 229.4 - 80.7) we arrive at an overall reaction energy of -6.9
kcal/mol. For comparison, the energy difference between the ion pair complex and the
isolated parent olefin and protonated zeolite model is also -6.9 kcal/mol. This result
further validates the decomposition of the overall reaction energy into the three
contributions cited above. Thus, the large endothermic energy contribution due to
deprotonation o f the zeolite is balanced by the large exothermic contributions from
the basicity of the parent olefin and the ion-pair interaction, leading to a reaction that
is overall slightly exothermic.
We have previously found the basicity of parent olefins to be a good
predictor of the existence of stable carbenium ions in zeolites.4 4 In the prior work,
theoretical calculations (MP4(sdtq)/6-311+G*) suggested that a basicity of 209
kcal/mol or more was required in order for a stable carbenium ion to form. Those
results are consistent with the basicity of the parent olefin of 1 and the observation
of 1 as a stable cation in the zeolite.
81
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3.9. B3LYP/6-311G** optimized geometry of the n complex formed by
adsorbing neutral cyclic diene 2 to the zeolite acid site model used to obtain the
ion-pair structure in Figure 3.8.
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In previous studies of propene on acidic zeolites, we did not find any
evidence for a stable isopropyl carbenium ion on the zeolite. However, experiment
and theory were in agreement4 6 about the existence of two other states: a complex of
47
the olefin on the acidic hydrogen and a framework bound isopropoxy species of a
48
type originally proposed by Kazansky. We considered the possibility that a
complex and a framework alkoxy might also be stable states. This proved to be the
case. Figure 9 is the B3LYP/6-311G** optimized geometry of the complex of cyclic
diene 2 on a cluster model of HZSM-5. The distance between the acidic proton and
the nearest carbon is 2.20 A and the O-H-C angle is 167.9°. The total Mulliken
charge on the olefin is 0.02 |e|. The binding energy for the complex is -4.7 kcal/mol.
The electronic energy of the complex at this level of theory is only 2.2 kcal/mol
higher than that of the ion-pair structure in Figure 3.8. As such, the complex could
plausibly exist in low concentrations on the zeolite.
Figure 3.10 reports the B3LYP/6-311G** optimized geometry of a
framework alkoxy species that was obtained as a stable state when a structure similar
to the ion pair was optimized from an initial geometry with a close contact between
one of the charged carbons of the cation and a basic oxygen on the zeolite cluster. The
energy of the alkoxy complex is 28.4 kcal/mol higher than that of the ion pair. Thus,
even if the kinetics were such that the alkoxy could form (we did not make any
attempt at determining the barrier to the formation of the alkoxy from the ion pair or
from the parent olefin and the protonated zeolite) at equilibrium, the concentration of
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3.10. B3LYP/6-311G** optimized geometry of the framework-bound
alkoxy species that could form as an alternative to cation 1 (c/., the ion-pair
structure in Figure 3.8).
84
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the alkoxy species would be effectively zero, even at the highest temperatures used
in MTG chemistry. We can thus theoretically rule out any involvement of this
alkoxy species in the observed chemistry.
Modeling the Roles of 1 in Hydrocarbon Synthesis. The experimental
evidence reported above suggests three pathways from 1 to volatile hydrocarbon
products. 1. Formation of 1 and related species, including aromatics, from ethylene
ends the induction period and opens lower barrier pathways from DME to propene
and other olefins. 2. At higher temperatures and while hydride acceptors are present
on the catalyst, 1 is an intermediate in the formation of toluene. At longer reaction
times, after other products have exited the reactor, cation 1 slowly decomposes on
the catalyst to yield toluene and propene. We modeled parts of the first two
pathways using theoretical chemistry.
In the traditional view of MTG/MTO chemistry, ethylene forms from DME
by some means and then propene is formed by methylation of ethylene. Further
methylation steps lead to successively higher olefins. As a step to understanding
how the formation of cation 1 in the zeolite accelerates DME conversion, we used
theoretical methods to calculate the reaction pathways for methylation of various
organic species present or assumed to be present under reaction conditions.
Theoretical (B3LYP/6-311G*) calculations of the reactants, products and transition
states were carried out for each o f the reactions in Table 3.2. In each case we
49
calculated the reaction path (using the intrinsic reaction coordinate procedure in
85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission
2.65
2.60
a) Reactants
2.19
3.02
b) TS
2.34
1.10
c) Products
Figure 3.11. B3LYP/6-311G* optimized geometries of the three stable points along the reaction pathway for
the conversion of trimethyloxonium cation and cyclic diene 2 to DME and cation 3. a) Reactants (0 kcal/mol);
b) Transition state (+32.9 kcal/mol); c) Products (-35.1 kcal/mol). 8 6
Gaussian98) to ensure that that the transition state linked the reactant and product
states. The hydrocarbon reactants included ethylene, propene, toluene, the cyclic
diene 2 and other olefins related to cation 1. Consideration of 2 is justified by the
low energy difference between the complex of 2 on the zeolite and the ion-pair
complex of 1 and the anion site (cf., Figures 3.8 and 3.9).
In modeling the methylation of neutral hydrocarbon reactants, we had several
choices for the methylating agent. CH 3+ is not a plausible intermediate by itself,
although CH3+ groups bound to the conjugate base sites (so-called "surface
methoxys") might be able to act as methylating agents. We do not observe any such
groups under our experimental conditions. The low proton affinity of methanol (exp.
180.3 kcal/mol) argues against CH 3 0 H 2 + existing in the zeolite as a stable species.
Dimethyl ether is more basic (exp. 189 kcal/mol), and thus, protonated dimethyl
ether, (CH3)2 0 H+ (DMEH+), would be a more likely to have an appreciable lifetime
in the zeolite. An even more stable species is the trimethyloxonium cation, TM O+.
At room temperature and slightly above, this cation has been observed in the zeolite
by NMR in the presence of an excess of ether. 5 0 "52 Furthermore, several of the
classical mechanisms of MTG chemistry are formulated using TMO . Running
counter to the indications of methylation agent stability is the energy required for
methyl group uphill by 99.3 kcal/mol. For the reaction of TM O+ and ethylene, the
initial products obtained are CH 3 -CH 2-CH2+ and DME. However, the relative
proton affinities o f these molecules are such that a proton is further transferred
during the optimization, giving as final products propene and DMEH+. The overall
reaction is exothermic by 15.1 kcal/mol and the barrier is 48.9 kcal/mol. Similarly, the
- f-
reaction of TMO and propene passes through a barrier of 45.3 kcal/mol to initially
87
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
form CH3-CH+-CH 2-CH3 and DME, followed by proton transfer to form 1-butene
and DMEH+. The overall reaction is exothermic by 10.7 kcal/mol. The barrier for
methylation of toluene to form />-xylene, 41.6 kcal/mol, is also fairly high. Figure
3.11 reports the structures of the stable states on the pathway for the reaction of
TM O + with cyclic diene 2. In contrast to the reactions of the simple olefins or
toluene, the products obtained with 2 were DME and a carbenium ion, 3. This
reaction is also exothermic, by 35.1 kcal/mol, but the reaction barrier is significantly
lower than for the simple olefins, 32.9 kcal/mol. Methylation of 2 by either
CH 3 0 H2+ or DM EH+ involves much higher energy barriers (62.9 and 63.5
kcal/mol). Thus, the lower energy requirements for CH 3+ dissociation by these
species (versus TMO+) does not result in more favorable energy barriers for methyl
donation to 2. A low barrier was also obtained for an isomer of 2 with an exo-cyclic
double bond (34.6 kcal/mol). In the gas phase, the calculated energy of exo-2 is only
1.1 kcal/mol above that of the endo isomer. We also considered the possibility that
the cyclic dienes could ring-open to open-chain species, and these could be active
species in methylation reactions. Table 3.2 reports results that the barrier for
methylating an open-chain triene related to cation 3 is 40.3 kcal/mol. However,
formation of the triene from the cyclic diene is endothermic by 17.7 kcal/mol. Thus,
the effective barrier for this route is also high.
We also considered the possibility that cations 1 or 3 undergo direct
methylation (i.e., without prior deprotonation) by either DME or methanol. The
transition states for the reactions reported in Table 2 involve a concerted exchange of
H for CH3 . In effect, in the transition state H must be substantially removed
from the cations prior to methylation. Not surprisingly, the barriers for these direct
methylation reactions are quite high. In summary, the calculations reported in Table
88
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 3.2. Theoretical (B3LYP/6-311G*) barriers for various methylation reactions
in the gas phase
reactants products E* (kcal/mol)
Ethylene TMO+ Propene DMEH+
48.9
Propene TMO+ 1-Butane DMEH+
45.3
Toluene TMO+ ^-Xylene DMEH+
41.6
TMO+ DME
32 9
DME+ CH 3 0 H
63.5
c h 3 o h 2+
H 2 0
62.9
DME
34.6
DME CH3 OH 63.5
CH 3 OH
H 2 O
62.9
TMO+
40.3
89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 suggest that cyclic dienes, in equilibrium with cyclopentenyl cations under reaction
conditions, provide a pathway for carbon-carbon bond formation with a lower barrier
than for other species likely to be present in the catalyst.
We studied the mechanism by which cation 1 can form toluene under reaction
conditions. Skeletal rearrangement reactions are invariably complex and involve a
number of bond making and breaking steps; nevertheless, they are characteristic of
organic chemistry in liquid and solid acids. Here we simply consider reactions in the
gas phase and do not attempt to model the hydride accepting species or the zeolite.
The reaction stoichiometry for producing a C7H 8 species from C 7Hjj+ implies that
there must be a means by which H2 and H+ are lost. There is no apparent mechanism
by which H 2 can be lost directly, therefore the reaction must involve the lost of H-
and two H+ in three separate steps. The following is a speculative mechanism, based
entirely on theoretical predictions. In several cases the proposed mechanism involves
molecules that can exist as two or three different isomers. In each case we assume
that only the lowest energy isomer is involved.
Each species in Figure 3.12 was optimized at MP2/6-311+G*, but we focus
only on the relative electronic energy of each species and do not detail the geometries
obtained. Cation I can lose H+ to form the parent olefin 2, as discussed above.
There are two other configurational isomers of this species that vary by the positions
o f the methyl groups on the ring. The lowest energy isomer is 2a. Isomer 2b is only
0.5 kcal/mol higher in energy, and thus is very likely to be present in the zeolite
during the course of the reaction. Protonation of either of these two isomers gives 1.
The other isomer (2c) is 3.6 kcal/mol higher in energy, and is thus less likely to be
formed. Isomer 2c also does not give 1 upon protonation. We will only consider 2a
in subsequent steps of the reaction. We have earlier shown that the protonation
90
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
3a
2a
-H -
~ H
4a
5 6a
Alternate Isomers:
-H +
T oluene
2b 2c
4b
3b
4c
Figure 3.12. Proposed reaction mechanism for the conversion of 1 to toluene.
While this is shown in schematic form, each structure shown (as well as all
reasonable isomers thereof) were optimized at the MP2/6-311+G* level.
91
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
energy of 2a is 228.6 kcal/mol at the B3LYP/6-311G* level of theory. At the
MP2/6-311+G* level of theory used for the calculation of the reaction path, the
protonation energy is reduced to 217.9 kcal/mol. Isomer 2a can lose H~ to any cation
equivalent present (e.g., an alkoxy species derived from propene) to form the C7H9+
species 6a. Loss of H~ is highly endothermic (270.1 kcal/mol at MP2/6-311+G*)
and thus must be countered by a similarly large exothermic contribution from the
binding of H~ to some other cation or incipient cation. Another isomer (6b) is 7.6
kcal/mol higher in energy.
Cation 6 a can then undergo a hydride shift, to form a species with an
exocyclic double bond (Figure 12). There are three such C7 H9+ isomers. The lowest
energy isomer predicted by the MP2/6-311+G* level of theory is 7a. Isomer 7b is
4.3 kcal/mol higher in energy. The third isomer (7c) is much less stable, being 21.1
kcal/mol higher in energy than 1. Formation of 7a from 6a is predicted to be
exothermic by 22.0 kcal/mol. We again assume that the reaction involves only isomer
7a.
Isomer 7a can then undergo an additional hydride shift, which results in the
formation of 8 (Figure 12). Molecule 8 is 13.6 kcal/mol higher in energy than 7a. This
destabilization is due in large part to the change from a tertiary carbenium ion (7a) to
a secondary carbenium ion (8 ), although 8 is an allyl as well as a cyclopropylcarbinyl
cation. The formation of the three-membered ring also imparts strain, and thus a
higher energy state.
However, from 8 the ring can open to form the toluenium cation. The ring
opening of 8 naturally produces the para isomer of the toluenium cation (9a). At the
MP2/6-311+G* level, this step is predicted to be 20.9 kcal/mol exothermic. There are
92
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
obviously three isomers of the toluenium cation. As expected, the para isomer is
lowest in energy. The ortho isomer is 1.3 kcal/mol higher in energy, whereas the meta
isomer is 4.3 kcal/mol less stable than the para cation. We assume that the reaction
proceeds via the para isomer.
In the final step o f the reaction 9a would lose H+, possibly by donation to a
deprotonated acid site, and generate toluene (Figure 12). The calculated protonation
energy of toluene is 185.8 kcal/mol. We previously stated that a basicity of -209
kcal/mol is required o f the parent olefin or aromatic in order for a long-lived
44
carbenium ion to form. Thus, we would assume that the toluenium cation is not
stable in the zeolite, and readily forms toluene.
3.6 Discussion
Formation and Decomposition of Cyclopentenyl Cations. Signals due to
carbenium ion 4 were observed in an early NMR study of products formed when
53
propene was exposed to zeolite HY at room temperature, and this assignment was
42
established in 1989. Cations 1, 4, and similar cations with other substituents have
since been observed by NMR in studies of various olefins and alcohols on acidic
54, 55
zeolites. ’ Alkyl-substituted cyclopentenyl cations are indefinitely stable in
zeolites at room temperature; as such they join the trity l,^ in d an y l^ and
30
pentamethylbenzenium cations in a small group of persistent cation classes
characterized in zeolites by NMR. Cyclopentenyl cations have also been detected
co
on zeolites and other solid acids using UV spectroscopy. The observation here
93
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
that theory correctly predicts the stability of the ion-pair structure of 1 on the
zeolite anion site (Figure 3.8) is a satisfying result.
We have not fully identified the mechanism by which ethylene forms cation 1
in the zeolite, but Figure 3 shows that oligomerization occurs at low temperatures
and cation 1 is present even at 373 K. Stoichiometry dictates that the formation of 1
from ethylene also requires formation of alkanes. For example, consider a reaction
where the only products are 1 and propane:
13 C2 H4 + 2 ZeoOH -» 2 C7 H n + + 2 ZeoO' + 4 C3H 8
where ZeoOH denotes a zeolite Bransted site and ZeoO the conjugate base of such a
site. Analogous schemes may be written for other alkane co-products, and indeed we
observe light alkanes in the volatile products following the injection of ethylene. In
the double-pulse experiments described above, the ethylene that formed cation 1 was
delivered in the first pulse. Without an olefin pre-pulse, the conversion of methanol
or DME exhibits an induction period (cf., Figure 3.1). An unknown, but inefficient
mechanism of olefin synthesis operates during the induction period and accumulation
of oligomeric species in the zeolite leads to 1. We have not here attempted to
identify the mechanism of the induction reaction, but we sometimes detect small
quantities of ethylene during the induction period, which is consistent with most
mechanisms for the formation of the "first" carbon-carbon bond.
Balanced chemical reactions (albeit with large coefficients) can be written that
are consistent with the observed decomposition products o f 1 at long reaction times.
For example:
9 C 7 H n + + 9 ZeoO” -» 6 C7H 8 + 7 C3H 6 + 9 ZeoOH
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
This balanced reaction predicts that 1 decomposes to yield roughly equal amounts of
propene and toluene, as is observed. Similar balanced reactions can account
stoichiometrically for the ethylene that is observed. Unimolecular decomposition of 1
would form products that we do not observe, such as acetylene and cylopentene or
ethylene (observed) and cyclopentadiene. A more plausible mechanism for the
decomposition of 1 is deprotonation of 1 to form 2 (calculated to be uphill by only
ca. 2 kcal/mol) which then diffuses in the zeolite until it encounters a second 1 .
Intramolecular hydride transfer (disproportionation) between 1 and 2 would then
form toluene and a cyclic olefin that would further react on the catalyst. An
inefficient decomposition mechanism is consistent with the long half-life of cation 1
in the absence of other reactants which can act as hydride acceptors and open a more
direct pathway to toluene (Figure 3.12).
Earlier work has speculated on roles for cyclopentenyl cations in the
synthesis of aromatics. 5 4 ,5 8 ’ 59 The experimental demonstration of the intermediacy
of 1 in the synthesis of toluene is an important feature of the present study. In the
design of MTO processes, it is desirable to form olefins while minimizing aromatics.
Ironically, on zeolite HZSM-5 cation 1 promotes the formation of olefins, but it is
also an intermediate in the pathway to aromatics. A C7 carbenium ion is the
precursor to the C7 aromatic product. We also observe cations with substitution
patterns similar to 4 and speculate that some of the xylene products form directly
from this more highly substituted Cs carbenium ion.
How Could Cyclopentenyl Cations Catalyze Propene Synthesis?
Immediately following the induction period, we see a very high selectivity for
propene, 80+% of the total olefin yield. In particular, we see far more propene than
ethylene. The results in Table 3.2 indicate that methylation o f propene to form
95
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
butene should be even easier than methylation of ethylene to form propene. If, on a
working catalyst, the propene is formed from ethylene, why would chain growth
stop after a single methylation reaction? Pulsing ethylene onto zeolite HZSM-5
catalyst rapidly synthesizes cation 1 (and related species such as cation 4) with very
high selectivity. Cation 1 also forms naturally from the ethylene obtained when
methanol or DME are slowly converted during the induction period. A number of
observations in the literature point to roles for cyclopentenyl cations or closely
related species in olefin synthesis on acidic zeolites. For example, Langner observed
an induction period in the synthesis of hydrocarbons from methanol on NaHY
catalyst and found that this was shortened by co-feeding small amounts of other
60
alcohols. Co-feeding ethanol shortened the induction period by a factor of ca. 2,
but co-feeding cyclohexanol reduced the induction period by a factor of nearly 20. In
21
our recent communication, we used the pulse-quench reactor to asses the effect of
pulsing various compounds on the subsequent conversion o f methanol. Pulsing
isopropanol was effective in eliminating the induction period, but an even greater
conversion was obtained with a smaller quantity of the cyclic diene 10, a more direct
precursor to cation 1. Formation of 1 in the catalyst, by any means, is correlated
with a dramatic increase in hydrocarbon synthesis. When cation 1 is removed from
the zeolite bed by decomposition over time, the catalytic effect is removed and the
induction period returns.
10
96
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The barriers to direct methylation of cations 1 or 4, reported in Table 2, are
quite high. However, if the cations first deprotonate, the methylation barriers for the
resulting olefins are significantly lower than for other species in the Table. Recall
that the theoretical energy for the complex of olefin 2 on a zeolite is only 2 . 2 kcal/mol
higher than that o f cation 1. Thus, the total barrier for 1 to deprotonate followed by
methylation o f 2 to form cation 3 is still 10 kcal/mol lower than the barrier to convert
ethylene to propene. Methylation of the exocyclic isomer of 2 directly produces a
cyclopentenyl cation with an ethyl substituent, 11.
11
Cation 11 could also form by rearrangement o f cation 3. Like many skeletal
rearrangements, the detailed mechanism for converting 3 to 11 requires a number of
elementary steps, but the essence of this rearrangement is a 5 ^ 6 ring expansion-
contraction, Scheme 1.
3 12 11
scheme 1
Alkyl-substituted cyclohexenyl cations such as 12 are slightly less stable than
cyclopentenyl cations with similar substitution patterns. By similar routes, cation
11 can undergo an additional deprotonation, methylation step to form cation 13, with
97
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a propyl sidechain. Skeletal rearrangements as in Scheme 1 are consistent with our
earlier observation that propene molecules exiting the reactor contain carbon atoms
derived from both the olefin pulse that synthesized 1 and a subsequent methanol
pulse.^ *
13
Cations 11 and 13 can eliminate ethylene and propene, respectively, by
hydride migration followed by elimination, as shown for cation 11 in Scheme 2.
scheme 2
In Scheme 2, a primary carbenium ion (or a closely related protonated cyclopropane)
is formally required as an intermediate in the elimination of ethylene. The analogous
form al route from 13 to propene requires only a secondary cation intermediate. On
this basis alone, a qualitative argument for the propene selectivity could be made
98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
4 2 .2
H3C
19 2 H
( W . C H 3
H3C
CH3
H -aC
H2Ci I CH2
Figure 3.13. Potential energy surfaces (kcal/mol) for removal of ethylene or propene from cations 11 or 13,
respectively. All minima and transition states were optimized at B3LYP/6-311G*. Schematic structures are shown only
for the reactions of 1 1 to form ethylene; very similar structures were obtained for the analogous pathway to propene.
Note that that the pathway avoids primary and secondary carbenium ion intermediates (c f the formal mechanisms in
Scheme 2). For each state the relative energy along the ethylene pathway was higher in energy and than that of propene.
99
based on the relative stabilities of primary and secondary cations. Better yet, we
calculated (B3LYP/6-311G*) both reaction pathways, and these are reported in
Figure 3.13. The structures of intermediates and transition states as well as their
relative electronic energies are shown for elimination of ethylene, and energies only
are shown for the analogous elimination of propene (the structures are very similar).
Figure 3.13 shows that these elimination reactions do not in fact have either primary
or secondary carbenium ions as intermediates. The first hydride-transfer step leads to
cyclopropane ring closure to form a tertiary cation intermediate, and structures
similar to primary or secondary cations are seen only in the transition states, the
second of which forms ethylene or propene. Cyclopropylcarbinyl cations, seen in
Figures 3.12 and 3.13, are well characterized intermediates in a variety of acid-
catalyzed rearrangements in solution. 6 1 , 62
Comparing the energies in Figure 3.13, one sees that elimination of ethylene
by cation 11 involves an overall barrier of 58.3 kcal/mol, but the corresponding
barrier is 16.1 kcal/mol lower for elimination of propene by cation 13. This
difference easily accounts for the selectivity of propene over ethylene. Cyclic
cations with C4 or larger alkyl chains would eliminate C4 or higher olefins as primary
products, but if the elimination of propene is rapid at reaction temperature, longer
chains would have little opportunity to form. In the zeolite, the intermediates and
transition states corresponding to those in Figure 3.13 would be stabilized by
interaction with the framework, but this will not greatly affect the relative energetics
for producing propene vs. ethylene. Note that cyclopentenyl cations are not the
100
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
only precursors to elimination reactions that account for the olefin selectivity; we
would expect to see analogous mechanisms and the same selectivity prediction for
eliminating olefins from, for example, aromatic rings.
Olefin limination from cations 11 and 13 also forms a monomethyl
cyclopentenyl cation. A closed catalytic cycle is obtained by deprotonation and
methylation of this species to reform cation 1. B3LYP/6-311G** optimizations of
the monomethylcation and olefin in contact with the zeolite model indicate that this
is possible. In this case, the carbenium ion-zeolite complex is only 0.9 kcal/mol
lower in energy than the corresponding olefin complex. Thus the overall reaction, on
a working catalyst, converts three methanol molecules to propene and water, and the
carbon-carbon bond making and breaking steps occur on cyclic dienes to form
cyclopentenyl cations. The proposal advanced here that penta-cyclic species are
intermediates in the synthesis of olefins suggests parallels to early proposals of the
63 68
mechanisms of catalytic reforming and bi-functional ring opening. ’ The
elimination of olefins from polymethylated cyclopentyl cations echoes another early
observation that hexamethylbenzene reacts under hydrocracking conditions to form
isobutane and xylenes.69
3.7 Conclusions
The mechanism of MTO/MTG chemistry on zeolite HZSM-5 has several
distinct roles for cyclopentenyl cations. These cations form during the induction
101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
period from small amounts of olefin formed in an "induction reaction". On zeolite
HZSM-5 the principal role of the induction reaction is to synthesize the
cyclopentenyl cations that characterize a working catalyst. Once these cations are
present, a more efficient mechanism for hydrocarbon synthesis opens up, and the
induction reaction takes on a subordinate role. Cyclopentenyl cations are stable
species on zeolite acid sites, but the cyclic dienes obtained by their deprotonation are
also stable at very slightly higher energies (cf., 1 and 2). Thus, the cations act as
reservoirs of cyclic dienes at reaction temperature and these are much more easily
methylated than ethylene or propene. Sidechain methylation and skeletal
isomerizations lead to cations with alkyl substituents, and these eliminate propene
more readily than ethylene. At higher temperatures and conversions, many olefmic
and aromatic compounds are present in the catalyst, and some of these may also be
methylated and eliminate olefins.
In the presence of hydride acceptors such as olefins, cation 1 is an
intermediate in the synthesis of toluene. Thus, a species with a catalytic role in
MTO processes is also an intermediate in MTG chemistry, and suppressing
aromatic products in MTO on HZSM-5 is necessarily challenging. In the absence of
feed, cation 1 slowly decomposes to toluene, propene and ethylene.
Olefins, aromatics and other hydrocarbons are, o f course, reactive on solid
acids, and secondary products are readily observed in MTO/MTG chemistry,
especially at higher temperatures or with longer contact times. For example, toluene
102
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
is methylated to />-xylene (Table 3.2) and propene oligomerizes and cracks to an
equilibrium distribution of olefins. These processes are also very interesting and
important. The present investigation has focussed on primary reactions leading to
olefinic (MTO) and aromatic (MTG) products.
103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.8 References and notes
(1) Chang, C. D; Silvestri, A. J. J. Catal. 1977, 47, 249-259.
(2) Chang, C. D. Catal. Rev. - Sci. Eng. 1983, 25, 1-118.
(3) Stocker, M. Microporous Mesoporous Mater. 1999, 29, 3-48.
(4) Kei, F. J. Microporous Mesoporous Mater. 1999, 29, 49-66.
(5) Haase, F.; Sauer, J. J. Am. Chem. Soc. 1995,117, 3780-3789.
(6 ) Blaszkowski, S. R.; van Santen, R. A. J. Am. Chem. Soc. 1996,118, 5152-
5153.
(7) Blaszkowski, S. R.; van Santen, R. A. J. Am. Chem. Soc. 1997,119, 5020-\
5027.
(8 ) Blaszkowski, S. R.; van Santen, R. A. J. Phys. Chem. B 1997,101, 2292-
2305.
(9) Tajima, N.; Tsunda, T.; Toyama, F.; Hirao, K. J. Am. Chem. Soc. 1998,120,
8222-8229.
(10) Stich, I.; Gale, J. D.; Terakura, K.; Payne, M. C. J. Am. Chem. Soc. 1999,
121, 3292-3302.
(11) Hutchings, G. J.; Watson, G. W.; Willock, D. J. Microporous Mesoporous
Mater. 1999, 29, 67-77.
(12) Forester, T. R.; Howe, R. F. J. Am. Chem. Soc. 1987,109, 5076-5082.
(13) Campbell, S. M.; Jiang, X.-Z., Howe, R. F. Microporous Mesoporous Mater.
1999, 29,91-108.
(14) Anderson, M. W.; Klinowski, J. Nature 1989, 339, 200-203.
(15) Anderson, M. W.; Sulikowski, B.; Barrie, P. J.; Klinowski. J. Phys. Chem.
1990, 94, 2730-2734.
(16) Munson, E. J.; Lazo, N. D.; Moellenhoff, M. E.; Haw, J. F. J. Am. Chem.
Soc. 1991,113, 2783-2786.
104
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(17) Munson, E. J.; Haw, J. F. J. Am. Chem. Soc. 1991,113, 6303-6305.
(18) Munson, E. J.; Khier, A. A.; Lazo, N. D.; Haw, J. F. J. Phys. Chem. 1992,
96, 7740-7746.
(19) Hunger, M.; Horvath, T. J. Am. Chem. Soc. 1996,118, 12302-12308.
(20) Ernst, H.; Freude, D.; Mildner, T. Chem. Phys. Lett. 1994,229, 291-296.
(21) Goguen, P. W.; Xu, T.; Barich, D. H.; Skloss, T. W.; Song, W.; Wang, Z.;
Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 1998,120, 2651-2652.
(22) Kolboe, S. Acta. Chem. Scand. 1986, A-40, 711-713.
(23) Dahl, I. M.; Kolboe, S. J. Catal. 1994,149, 458-464.
(24) Dahl, I. M.; Kolboe, S. J. Catal. 1996,161, 304-309.
(25) Kolboe, S. In Methane Conversion; Bibby, D. M., Chang, C. D., Howe, R. F.
and Yurchak, S., Eds.; Elsevier Science Publishers: Amsterdam, 1988, pp
189-193.
(26) Mole, T.; Whiteside, J. A.; Seddon, D. J. Catal. 1983, 82, 261-266.
(27) Mole, T.; Bett, G.; Seddon, D. J. Catal. 1983, 8 4 ,435-445A
(28) Pines, H. J. Catal. 1985, 93, 205-206.
(29) Haw, J. F.; Goguen, P. W.; Xu, T.; Skloss, T. W.; Song, W.; Wang, Z. Angew.
Chem. 1998, 37, 948-949.
(30) Xu, T.; Barich, D. H.; Goguen, P. W.; Song, W.; Wang, Z.; Nicholas, J. B.;
Haw, J. F. J. Am. Chem. Soc. 1998,120, 4025-4026.
(31) Barich, D. H.; Xu, T.; Song, W.; Wang, Z.; Deng, F.; Haw, J. F. J. Phys.
Chem. B 1998,102, 7163-7168.
(32) Haw, J. F.; Nicholas, J. B.; Xu, T.; Beck, L. W.; Ferguson, D. B. Acc. Chem.
Res. 1996, 29, 259-267.
(33) Hehre, W. J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. Ab Initio Molecular
105
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Orbital Theory, John Wiley & Sons: New York, 1986.
(34) Gauss, J. Chem. Phys. Lett. 1992, 191, 614-620.
(35) Schafer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571-2577.
(36) Barich, D. H.; Nicholas, J. B.; Xu, T.; Haw, J. F. J. Am. Chem. Soc. 1998,
120, 12342-12350.
(37) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652.
(38) van Koningsveld, H.; van Bekkum, H.; Jansen, J. C. Acta Crystallogr. 1987,
43, 127-132.
(39) Frisch, M. J.; Trucks, G. W.; Schlegal, H. B.; Gill, P. M. W.; Johnson, B. G.;
Robb, M. A.; Cheeseman, J. R.; T. Keith; Petersson, G. A.; Montgomery, J.
A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J. V.;
Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.; Nanayakkara, A.;
Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.;
Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J.;
Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-Gordon, M.;
Gonzalez, C.; Pople, J. A. Gaussian 94, Revision B.2.; Gaussian, Inc.:
Pittsburgh, PA, 1995.
(40) ACES II, an ab initio quantum chemical program system. Stanton, J. F.;
Gauss, J.; Watts, J. D.; Lauderdale, W. J.; Bartlett, R. J.
(41) Chu, C.T-W.; Chang, C. D. J. Catal. 1984, 86, 297-300.
(42) Haw, J. F.; Richardson, B. R.; Oshiro, I. S.; Lazo, N. L.; Speed, J. A. J. Am.
Chem. Soc. 1989, 111, 2052-2058.
(43) Xu, T.; Haw, J. F. Topics in Catalysis 1997, 4, 109-118.
(44) Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 1998,120, 11804-11805.
(45) In this case we are considering the difference in energy between the
optimized ion-pair complex, and the sum of the energies of 1 and the zeolite
conjugate base, both of which were optimized in isolation at the B3LYP/6-
311G** level of theory.
(46) Nicholas, J. B.; Xu, T.; Haw, J. F. Topics in Catalysis 1998, 6, 141-149.
106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(47) White, J. L.; Beck, L. W.; Haw, J. F. J. Am. Chem. Soc. 1992,114, 6182-
6189.
(48) Kazansky, V. B. Acc. Chem. Res. 1991, 24, 379-382.
(49) Gonzalez, C.; Schlegel, H. B. J. Chem. Phys., 1989, 90, 2154-2161.
(50) Haw, J. F.; Xu, T.; Nicholas, J. B.; Goguen, P. W. Nature 1997, 398, 832-
835.
(51) Munson, E. J.; Kheir, A. A.; Haw, J. F. J. Phys. Chem.. 1993, 97, 7321-
7327.
(52) Munson, E. J.; Haw, J. F., J. Am. Chem. Soc., 1991,113, 6303-6305.
(52) Zardkoohi, M.; Haw, J. F.; Lunsford, J. H. J. Am. Chem. Soc. 1987,109,
5278-5280.
(54) Xu, I.; Haw, J. F. J. Am. Chem. Soc. 1994,116, 7753-7759.
(55) Stepanov, A. G.; Vladimir, N. S.; Zamarev, K. I. Chem. Eur. J. 1996, 2, 157-
167.
(56) Tao, T; Maciel, G. E. J. Am. Chem. Soc. 1995,117, 12889-12890.
(57) Xu, T.; Haw, J. F. J. Am. Chem. Soc. 1994,116, 10188-10195.
(58) Sommer, J.; Sassi, A.; Hachoumy; M., Jost, R.; Karlsson, A.; Ahlberg, P. J.
Catal. 1997,171, 391-397.
(59) Shulz, H .; Wei, M. Microporous Mesoporous Mater. 1999, 29, 205-218.
(60) Langner, B. E. Appl. Catal. 1982,2 ,289-302.
(61) Olah, G. A.; Reddy, V. P.; Prakash, G. K. S. Chem. Rev. 1992, 92, 69-95.
(62) Mayr, H.; Olah, G. A. J. Am. Chem. Soc. 1977, 99, 510-513.
(63) Sinfelt, J. H.; Rohrer, J. C. J. Phys. Chem. 1961, 65, 978-981.
107
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(64) Carter, J. L.; Cusumano, J. A.; Sinfelt, J. H J. Catal. 1971, 20, 223-229.
(65) Cristoffel, E.; Fetting, F.; Vierrath, H. J. Catal. 1975, 40, 349-355.
(6 6 ) Cristoffel, E.; Robschlager, K.-H. Ind. Eng. Chem. Prod. Res. Dev. 1978,17,
331-334.
(67) Weitkamp, J.; Schulz, H. J. Catal. 1973, 29, 361-366.
(6 8 ) Dautzenberg, F. M.; Platteeuw, J. C. J. Catal. 1970,19,41-48.
(69) Sullivan, R. F.; Egan, C. J.; Langlois, G. E.; Sieg, R. P. J. Am. Chem. Soc.
1961, 83, 1156-1160.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 4
Acid-Base Chemistry of a Carbenium Ion in a Zeolite
Under Equilibrium Conditions: Verification of a
Theoretical Explanation of Carbenium Ion Stability
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.1 Introduction
Acid-base chemistry in the solution and gas phases is well understood in
terms of quantitative thermodynamic properties such as acid dissociation constants
and deprotonation energies. In the gas or solution phases, there are three general
reactions between an acid AH and a base B: (1) formation of a hydrogen-bonded1
complex without proton transfer, AH—B; (2) formation of a zwitterionic A '—HB+
complex that is stabilized by Coulombic interactions and hydrogen bonding; or (3)
form ation o f a covalently-bound onium ion complex. The application of
thermodynamic principles to acid-base reactions on solid surfaces— most
importantly those used in heterogeneous catalysis— is far less clear than for the
solution and gas phases. One concern is the possibility that adsorption of acids or
bases onto a solid surface may not be sufficiently reversible for thermodynamic
equilibrium to be established.
The most important and most studied solid acids are aluminosilicate zeolites,
which are micro-crystalline and have well-defined pore systems. Of particular
interest are the acid-base reactions of hydrocarbons on zeolites. Ethylene, propene
and other simple olefins have been shown to form neutral, weakly hydrogen-bonded
complexes with Bronsted sites in zeolites, e.g., 1. A similar complex is formed with
benzene. For the case of propene, it is well established that proton transfer can occur
in concert with a nucleophilic attack by an oxygen o f the conjugate base o f the
zeolite Bronsted site. This reaction produces a kind of oxonium ion that is called a
110
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
framework-bound alkoxy species in the zeolite literature, e.g., 2 4 The protonated
forms of unsaturated hydrocarbons are carbenium ions.
c h 3
H 2C = c
H
H3C \ / c h 3
CH
H
\ / ° \ / ^ \ / 0\ A |/
A l\ A h
H
\ / \ /
A h
C7 H,o
3
a
Sb fi\
c 7Hh+
4
Over the last several years we have developed energetic criteria for the
formation of stable carbenium species in zeolites. 5 Simply put, we find that in order
111
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
for a stable carbenium ion to form on a given zeolite the proton affinity of the parent
olefin must be greater than a certain threshold. In our previous work we used a
combination o f MP4(SDTQ)/6-311G* calculations and experimental data and to
establish the threshold at approximately 209.0 kcal/mol for olefins and aromatics in
HZSM-5. 5 It is important to understand this threshold in context: the proton affinities
of propene, hexamethylbenzene and pyridine are 179.6, 205.7 and 222 kcal/mol,
respectively. Clearly, only exceptionally basic hydrocarbons w ill react
exothermically with a zeolite Bronsted site in an elementary reaction to form a
persistent carbenium ion. One of the three persistent carbenium ions identified in our
earlier work4,6’7 was the 1,3-dimethylcylclopentenyl cation (4, C7H n +), which could
in principle be formed by protonation o f 1 -3-dimethylcyclopenta-1,3-diene (3,
C 7 H 1 0 , PA - 215.6 cal/mol). Note that the C7H n + cation does not form an oxonium
(alkoxy) species with the framework of the zeolite— analogous to 2. Our theoretical
work suggests this is due to steric constraints imposed by the topology of the zeolite
channel.8
We were led to the present work by a prediction developed from our previous
theoretical study5 — that once formed, the C7H n + cation could act as an acid and
deprotonate to form the parent diene when presented with a coadsorbed base
molecule having a proton affinity greater than -2 1 6 kcal/mol, the deprotonation
enthalpy of C7H n +. By analogy to acid-base reactions in other media, the reaction of
the C7H 1/ cation with a coadsorbed base could have three outcomes: ( 1) no reaction
112
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
beyond hydrogen-bonding of the base to the carbenium ion, giving C7H n +— :B; (2)
deprotonation of the carbenium ion by the base to form a hydrogen bonded complex
between the neutral hydrocarbon and protonated base, C7 H 11—H:B+; or (3)
formation o f an onium ion complex between the carbenium ion and added base
(C7H n B)+. Implicit in this prediction is the requirement that the acid-base reactions
o f carbenium ions in zeolites are sufficiently reversible for the application of
thermodynamics. Furthermore, since our predictions are based on gas phase
thermodynamic quantities, quantitative agreement with experiment would require
that these not be strongly altered by the zeolite environment.
To test out theoretical predictions, we reproducibly prepared samples of the
4-
C7H 11 cation in zeolite HZSM-5 (2 s of reaction at 573 K) using a pulse-quench
7-11 13 13
reactor and ethylene- C2 as a convenient starting material. C solid state NMR
was used to observe the reaction products obtained upon adsorption (usually at 298
K) of a number of base molecules. These coadsdsorbed bases covered a range of
12
proton affinities between 142.0 kcal/mol (carbon monoxide) to 229.2 kcal/mol
(trimethylphosphine), allowing us to effectively "titrate" the C7H n +acid. When the
formation of C7H n +— H:B+ complexes was theoretically predicted to be energetically
favored over the initial state, C7 H n +— :B, we observed deprotonation o f the
carbenium ion to form the neutral parent olefin in the zeolite. Deprotonation of
C7H n +did not occur when the C7H 10—H:B+ complex was predicted to be higher in
113
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
energy than the corresponding C7H n +— :B complex. In addition, C7H n + formed an
ammonium ion or phosphonium ion when the weak bases NH 3 or PH 3 were co
adsorbed. In these cases theory predicted the onium ion to be the most stable of the
three possible states. In several cases some of the three possible states were not
found to be theoretically stable. Experimental observations verified these
predictions. For example, oxonium ion formation did not occur with water or
dimethylether. Neither did CO bind to C7f i n + in the zeolite to form an acylium ion.
Theory also predicted that pyridine and trimethylphosphine would preferentially
form onium ions. Indeed, these were observed experimentally as minority products.
However, steric constraints in the zeolite channel work against onium ion formation
in the cases of these bulky bases, and deprotonation is dominant reaction observed.
This work provides an important extension and verification o f our previously
proposed theory5 of carbenium ion stability in zeolites.
4.2 Experimental Section
M aterials and Regents. All results reported were obtained on a pelletized
catalyst composed o f 70% by weight zeolite HZSM-5 with a Si/Al ratio of 19 and
30% alumina binder. E th y len e-^C 2 (99% ^ C ) , acetone-2-1 (99% l^C ) and
p y r i d i n e - (99% l^N ) were purchased from Isotec. Phosphine gas and all other
reagents were purchased from Aldrich. Safety Note: Phosphine gas is toxic and
spontaneously flammable in air.
114
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Preparation of the C 7H n + Cation in Zeolite HZSM -5. We used a pulse-
quench reactor8' 11 to synthesize the cation on the zeolite. For each sample, 0.63
mmol of ethylene-1 ^C 2 was pulsed onto the catalyst at 573 K and allowed to react
for 2 seconds prior to a thermal quench to room temperature. The major products of
the reaction o f ethylene are propene, butenes, isobutane and isopentane; these
volatile products exit the catalyst before the quench. A small, but reproducible
amount of the C7H n + cation remains on the catalyst as the major adsorbed organic
species. NMR spin counting indicates that for the catalyst and conditions used, the
yield o f the C7H n +cation corresponds to 1-4 % of the acid sites. 8
After quenching each sample, the reactor was sealed off and transferred into
a glove box filled with nitrogen. Typically, the catalyst pellets were then transferred
into a 7.5 mm MAS rotor which was sealed with a Kel-F end-cap and MAS
NMR spectra were acquired then to establish the presence o f the C7H 11 cation prior
to co-adsorption of a base. The MAS rotor was then returned to the glove box and
13-15
opened. The catalyst pellets were transferred to a shallow bed CAVERN device
which was attached to a vacuum line and evacuated for 30 minutes. The base was
then adsorbed at room temperature, with the exception o f acetone, which was
introduced at 193 K to prevent oligomerization. The amount o f base used is
expressed in units o f equivalents (equiv.) relative to the number of Bronsted acid
sites in the sample. Typical loadings were ca. 2 equiv. to ensure that the zeolite acid
115
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
sites were completely titrated and excess base was available for reaction with the
C7H n+cation.
NM R Spectroscopy, 1 5 ^ ancj 3 Ip solid state NMR experiments were
performed with magic angle spinning (MAS) on a modified Chemagnetics CMX-300
MHz spectrometer operating at 75.4 MHz for ^ C , 30.4 MHz for and 121.3
MHz for 31p. Hexamethylbenzene (17.4 ppm), g l y c i n e - (-347.6 ppm) and 85
% H 3 PO4 (0 ppm) were used as external chemical shift standards for ^ c , 1 5 ^ an(j
31?, respectively. chemical shifts are reported relative to TMS, chemical
shifts relative to nitromethane, and 3 Ip chemical shifts relative to 85 % H3 PO4 .
Chemagnetics-style pencil probe spun 7.5 mm zirconia rotors at 6.5 kHz with active
spin speed control (± 3 Hz). Typical NMR experiments included: cross polarization
(CP, contact time = 2 ms, pulse delay = 1 s, 4000 scans); cross polarization with
interrupted decoupling (contact time = 2 ms, pulse delay = 1 s, 4000 scans, dipolar
dephasing time = 50 ps); single pulse excitation with proton decoupling (Bloch
decay, pulse delay = 10 s, 4000 scans). All spectra shown were obtained at room
temperature using cross polarization, except where otherwise stated.
Theoretical M ethods. We optimized the geometries of all the coadsorbate
base molecules, their protonated and methylated analogs, and all possible complexes
formed between the bases and CyHn*. We calculated the vibrational frequencies for
the bases, protonated bases, and methylated bases, in order to obtain estimates of the
enthalpies o f protonation (proton affinities) and methylation (methyl cation
116
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4-
affinities). The deprotonation enthalpy o f C7H 11 was also calculated. We did not
obtain frequencies for the larger complexes, as the relative energies were sufficient
to predict the behavior observed experimentally. All the optimizations and frequency
calculations were done using density functional theory (DFT) at the B3LYP/6-
311G* level16, 17 as implemented in Gaussian98.1 8 We obtained the isotropic
+ 1 9
chemical shifts of C7H 11 and its parent olefin C7H 10 from GIAO-MP2 calculations
with a tzp basis set on C and a dz basis set on H .20 We used the B3LYP/6-311G*
optimized geometries for the GIAO calculations. The reported chemical shifts are
relative to those obtained for TMS at the same level o f theory. The GIAO-MP2
21
calculations were done with ACES-II.
4.3 Results
4.3.1 Theoretical
The 11 basic coadsorbates we used are listed in Table 4.1. The geometries of
the isolated molecules and their protonated analogs are unremarkable and will not be
presented. However, the complexes formed between the bases and C7H n + are
significantly more interesting. We show in Figure 4.1, as a representative example,
optimized geometries for the three complexes involving NH 3 . In Figure la we show
the hydrogen bonded complex between NH 3 and C7H 1/ . Nitrogen coordinates to
117
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4.1. Experimental and Theoretical (B3LYP/6-311G*) Proton Affinities (AH),
Theoretical Methyl Cation Affinities (AH), and Theoretical Energies of the Three
Possible Adsorption Complexes Relative to the Sum of the Total Energies of Isolated
C 7H n + and the Co-Adsorbate (AE). All Values in kcal/mol. Deprotonation Enthalpy
o f C7H h + is 222.4 kcal/mol at B3LYP/6-311G*.
Proton Affinity
c h 3+
A ffinity
R elative Energies
Co-Adsorbate Exp. Theory Theory C7H „ + — :B
C7 H 1 0 — :HB+ (C7H „ B ) +
10.2 CO 142 135.0 78.6 -2.2
h 2 o 165 169.8 73.9 -12.6
Nitromethane 180.4 176.8 82.8 -13.2
p h 3 188 185.1 102.2 -4.0 21.0 -4.3
DM E 189 187.7 88.2 -10.3
A cetone 194 193.1 93.0 -13.1
n h 3 204.0 207.8 109.1 -12.2 -6.6 -19.3
DM FA 212.1 208.4 106.2 -17.6 -6.3 -17.0
D M A 217.0 215.6 109.8 -16.6 -8.0 -13.3
Pyridine 222 223.4 123.6 -11.1 -15.0 -22.9
TMP 229.2 225.2 137.7 -8.1 -10.0 -31.1
118
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.04
3.16
176.3
1.12
1.08
168.9
1.94
2.22
C)
Figure 4.1. B3LYP/6-311G* optimized geometries for possible complexes
formed by C7H11 + and NH3: (a) the C7H11 +—NH3 hydrogen bonded
complex; (b) the C7H10— HNH3+ hydrogen bonded complex; (c) the
covalently bound ammonium ion, C7H11NH3+.
119
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the most acidic hydrogen of C7H 11A The C-H distance is 1.12 A , 0.02 A longer than
it is in the isolated carbenium ion calculated at the same level o f theory. The H-N
distance is 2.04 A , the C-N distance is 3.16 A , and the C-H-N angle is 176.3°. All
these parameters are consistent with a strong hydrogen bond, as is the hydrogen
bonding energy of -12.2 kcal/mol. The C-H-N angle is 176.3°, very close to linear.
Figure lb shows the optimized geometry for the C7 H 11—NH 4 + complex. In this
case, the acidic ammonium proton interacts with both of the carbons of the double
bond in an asymmetric fashion; the C-H distances are 1.94 A and 2.22 A . The N-H
bond is 1.08 A , 0.06 A longer than calculated for isolated NH 4 + , and the C-N
distance is 3.01 A , 0.15 A shorter than in the C7H n +—NH 3 complex. The C-H-N
angle is 168.9°, still close to linear. The geometry suggests the strong tendency for
ammonium to donate the proton back to the olefin, regenerating the complex shown
in Figure 4.1a. The hydrogen bonding energy in this case is -6 . 6 kcal/mol. Finally,
Figure 4.1c shows the corresponding covalent ammonium ion complex, in which
nitrogen is bonded to carbon. The optimized N-C bond length of 1.59 A compares to
a value of 1.52 calculated for CH 3 -NH3 +. This is the lowest energy of the three
possible states, being -19.3 kcal/mol below the sum of the total energies of isolated
NH3 and C7H n +. In contrast to NH 3, stable geometries for all three types of
complexes involving some of the other bases could not be obtained.
120
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4.2. B3LYP/6-311G* optimized geometries for C7H 114— :B hydrogen
bonded complexes involving: (a) CO; (b) H 20; (c) pyridine (Py); and (d) P(CH3)3.
acetone complexes are within -0.05 A of the H - 0 distances in the C7H n +—H 2 O
complex. The more basic DMF and DMA complexes tend to bond more strongly to
the acidic proton, resulting in H- 0 distances o f 2.14 A and 2.32 A (DMF) and 2.14 A
and 2.26 A (DMA). Pyridine forms a hydrogen bonded complex (Figure 2c) very
similar to that formed by NH 3 . The C-H, H-N, and C-N distances are all within 0.01
A o f those in C7H n +—NH 3 , and the C-H-N angle (174.5°) is only -2° farther from
linear. Lastly, both PH3 and trimethylphosphine (P(CH3)3) form hydrogen bonds
with the acidic proton. Figure 4.2d shows the optimized geometry of the C7H 11 —
P(CH3 )3 complex. The P-H distance in the P(CH 3)3 complex is 2.61 A , whereas it is
2.90 A for C7H n +—PH 3 , reflecting the fact that PH 3 is much less basic than
P(CH3)3. Similarly, the C-H bond shows some lengthening in the P(CH3)3 complex
(1.12 A ), whereas it is 1 . 1 0 A in the PH 3 complex. In both cases the C-H-P bond
angles are close to linear; 176.0° for P(CH3)3, and 177.5° for PH3 .
121
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
178.5°
174.5°
122
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4.2 shows the geometries of several of the other C7H n+—:B hydrogen
bonded complexes. Figure 2a shows the complex formed between CO and C?Flii+.
The C-H bond distance is 1.10 A, and H-C distance is 2.70 A, and the C -0 triple
bond length is 1.12 A. The bonding arrangement is almost linear, with C-H-C and H-
C -0 angles o f 178.5° and 173.8°. The corresponding C-C distance is 3.80 A. In
contrast to CO, all of the complexes in which the base molecule interacts with the
carbenium ion through an oxygen give geometries with bifurcated hydrogen bonds,
involving both the oxygen lone pairs. The complex formed between C7H n + and H2O
is given as an example in Figure 4.2b. The bonding arrangement is somewhat
asymmetric, with H -0 hydrogen bonding distances o f 2.23 A and 2.31 A. The
bifurcated nature of the interaction results in C-H-0 angles that are far from linear;
144.3° and 142.4°. The closest C -0 distance is 3.19 A. It is somewhat surprising that
the oxygen prefers to interact with one of the acidic and one of the methyl hydrogens
of the carbenium ion, rather than interacting with the two, symmetry equivalent,
acidic hydrogens. Whereas we can optimize a stable complex with that hydrogen
bonding geometry (not shown), it is 1 .1 kcal/mol higher in energy than that of the
geom etry in Figure 4.2b. N itrom ethane, dim ethylether (DM E), acetone,
dimethylformamide (DMF), and dimethylacetamide (DMA) also all form bifurcated
hydrogen bonded complexes (not shown) similar to that for H2 O (Figure 4.2b). The
H-O distances in the DME and the optimized geometries of some of the other
123
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
171.2
167.4
170.9° ~ 3.83
2.42
Figure 4.3. B3LYP/6-311G* optimized geometries for C7H10—H:B+ hydrogen
bonded complexes involving: (a) PH3; (b) DMF; (c) Py; and (d) P(CH3)3.
124
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
C7H 11—H:B+ complexes are shown in Figure 4.3. These complexes exhibit
_-cation interactions between the acidic hydrogen of the protonated bases and the
olefin double bond. Figure 3a shows that the acidic hydrogen o f PH4 + coordinates
primarily with one carbon of the olefm, giving an optimized H-C distance of 1.92 A.
The other H-C distance is 2.21 A. The P-H bond length of 1.49 A is considerably
longer than the value of 1.40 in isolated PH 4+ calculated at the same level of theory.
The primary P-H-C angle of 161.7° results in a P-C distance of 3.37 A. Overall the
geometry suggests the pending transfer of the proton back to the olefm. Indeed, the
energy of this state is much higher than that of the isolated carbenium ion and neutral
base (+21.0 kcal/mol, Table 4.1). There is evidently some energetic barrier that
prevents spontaneous transfer of the proton to the olefm. In contrast, formation of the
complex between the olefin and P(CH 3)3H+ (Figure 4.3d) is exothermic, due to the
much greater basicity o f P(CH3)3 . Thus the P-H bond length o f 1.41 A is much
closer to that in isolated P(CH3)3H+ (1.40 A). P(CH3)3H+ is also much farther from
the olefin, with a H-C distances of 2.42 A and 2.50 A, and a P-C distance of 3.83 A.
The primary P-H-C angle is 170.9°. Both DMF (Figure 4.4b) and DMA (not shown)
form stable complexes in which the base is protonated. These complexes have the
shortest H-C distances, with values of 1.79 A and 2.19 A for DMF and 1.97 A and
2.32 A for DMA. These distances reflect the fact that DMA is more basic than DMF.
The O-H bond distance o f 1.04 A (DMF) and 1.01 A (DMA) also indicate the
DMAH+ cation is more stable and less prone to giving the proton back to the olefin.
125
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.87
1.57
Figure 4.4. B3LYP/6-311G* optimized geometries for onium ions formed by : (a)
PH3; (b) DMA; (c) Py; and (d) P(CH3)3.
126
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The resulting 0-C distances are 2.82 A (DMF) and 2.92 A (DMA), with
corresponding O-H-C angles o f 171.2° and 155.2°. Finally, formation of a C7H 10—
H;Py+ complex (Figure 4.4c) is also exothermic. The H-C distances of 2.12 A and
2.28 A are longer than the corresponding distances in the NH 3 complex (1.94 A and
2.22 A ), consistent with the greater basicity o f pyridine. Similarly, the H-N bond
distance of 1.04 A is shorter than the 1.08 A obtained for NH 3 . The C-N distance is
3.15 A , 0.14 A longer than calculated the NH 3 complex. In Figure 4.4 we present the
optimized geometries for some other onium ions. Although formation of the
C7HiiP(CH 3)3+ onium ion (Figure 4.4d) is much more exothermic than that o f the
C 7H uPH 3 + ion (Figure 4.4a), both have C-P bond lengths of 1.87 A . These bond
lengths compare to calculated values of 1.81 A in isolated CH 3 PH 3 + and P(CH 3 )4 +.
This is in agreement with the much weaker attraction of these bases for C7H n + (AE
= -4.3 and -31.1 kcal/mol) versus CH 3 + (AH = -102.2 and -137.7 kcal/mol). The C-O
bond distance in the C7HnDM A+ ion (Figure 4.4b) is 1.62 A , whereas it is 1.64 A in
+
the C7H11DMF ion (a very similar geometry which is not shown). These bond
+ * f
distances compare to values of 1.46 A in isolated CH3 -DMF and CH 3 -DMA .
Finally, the C7H nPy onium ion (Figure 4.4c) gives an optimized C-N bond length
of 1.57 A , which corresponds to 1.59 A in the C7HnNH 3 + ion (Figure 4.1c). For
comparison, the C-N bonds in isolated CH 3 -NH3 + and CH3 -Py+ are shorter at 1.52
127
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A and 1.48 A , again reflecting the higher affinity o f the bases for CH 3 + (AH = -
109.1 and -137.7 kcal/mol) versus C7H n+ (AE = -19.3 and -31.1 kcal/mol).
Table 4.1 gives the experimental and theoretical proton affinities for all of the
coadsorbed bases as well as the calculated deprotonation enthalpy for CyHn* (for
which no experimental data is available). The B3LYP/6-311+G* level of theory does
a reasonable job o f predicting the proton affinities o f the coadsorbed bases.
Generally the theoretical values are within 4 kcal/mol of experiment, although the
proton affinity of CO is significantly underestimated. The deprotonation enthalpy of
CyHii+ is probably considerably overestimated at the B3LYP/6-311G* level, which
we will discuss below. Table 1 also gives the energy of complex formation between
the coadsorbates and C7H n +. If we first consider the theoretical values, we find that
only pyridine and P(CH 3 )3 have proton affinities greater than the deprotonation
enthalpy of C7H n +. As expected, for only these two coadsorbates is the C7H 11—
H : B + complex lower in energy than the C 7 H n +— :B complex. Thus, we predict that
these are the only two coadsorbates for which it might be possible to experimentally
observe deprotonation o f the carbenium ion and formation o f the olefm. For all other
coadsorbates the C 7 H n +— :B complex is energetically preferred. For several of the
coadsorbates (CO, H 2 O, DME, and acetone) no C7H 11—H:B+ complex can even be
obtained theoretically. Attempts at optimization o f these C7H 11— H:B+ complexes
all revert to C7H n +— :B complexes during the course o f the calculation.
128
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
If we instead consider the experimental proton affinity data, there appears to
be a small discrepancy— the experimental proton affinity of pyridine ( 2 2 2 kcal/mol)
is lower, rather than higher, than the B3LYP/6-311G* deprotonation enthalpy of
C7H n + (222.6 kcal/mol). However, the problem here is simply that the B3LYP/6-
311G* level o f theory does not do a good job o f estimating the deprotonation
enthalpy of C7H n +. We have found this deficiency with the B3LYP functional in
predicting the proton affinities of olefins in general. Calculation of the deprotonation
enthalpy at the more accurate MP4(sdtq)/6-311G* level gives a value of 215.6
5
kcal/mol, which eliminates the discrepancy and is entirely in line with our
prediction of a stable C7H 11—H:Py+ complex.
In addition to the equilibria discussed above, there is also the possibility that
the coadsorbate will react with C7H n + to form a stable onium ion. From Table 4.1
we see that the theoretical results predict onium ions can form for all coadsorbates
except H 2 O, DME, and acetone. For CO an onium ion can be optimized, but its
formation is highly endothermic, and can thus be ignored. In four cases (PH3 , NH 3 ,
Py, and P(CH3)3), the onium ion is the energetically preferred state. Whereas the
difference in energy between the C7H n +— :B complex and the onium ion is small for
PH 3 , it is possible that both species will be observed. However, for NH 3 , Py, and
P(CH3 )3 the onium ion is clearly the energetically preferred state. Refinement of
129
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
these predictions involves steric difficulties in fitting the larger onium ions in the
zeolite channels, which we will discuss below.
In Table 4.1 we also present the calculated methyl cation affinities for all of
the base molecules. There is little experimental data with which to judge the
accuracy o f the theoretical numbers. However, it is possible, using heats of
formation and proton affinities, 1 2 to derive experimental methyl cation affinities
(AH) for H20 (6 6 . 6 kcal/mol), CO (78.3 kcal/mol), and NH 3 (105.3 kcal/mol).
These compare to calculated values o f 73.9, 78.3, and 109.1 kcal/mol, which we
consider reasonable agreement. Onium ion formation is predicted to be exothermic
for every adsorbate with a CH3+ affinity greater than or equal to that of PH3, while
no adsorbate with a CH3+ affinity less than or equal to that of acetone is predicted to
form an onium ion exothermically with C7H n+. There is a fair correlation (R.2 = 0.8)
between the CH3+ affinities and the binding energies of the onium complexes,
although the bases bind to CH3+ approximately five times stronger than they bind to
C7H n +.
4.3.2 Experimental result.
The key theoretical prediction was that coadsorption of a base with a proton affinity
greater than 217 kcal/mol would deprotonate the C7H 11+ carbenium ion in the zeolite
13
to form the neutral C 7 H 1 0 cyclic diene. Figure 4.5 reports C CP/MAS spectra of
130
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19'
5.9
40
P ( C H 3 ) 3 ( 229 )
147
125
86
Pyridine-d5 ( 222 )
DMF ( 212 )
- / ■ 'N
Acetone-dg ( 194 )
yv
Nitromethane ( 180 )
No added base
f i r r r | 1"-|TXp T T r i11 i t i | n i 111111 f i 11 r | n n |
3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 -5 0
ppm
13
Figure 4.5. C CP/MAS NMR spectra o f samples prepared by forming small
amounts (less than 0.1 equiv.) o f the C7H n + cation in zeolite HZSM-5 and then
absorbing an excess (typically 2 equiv.) of the indicated base. Numbers in
parentheses denote the experimental gas-phase proton affinities o f the bases in units
o f kcal/mol. Adsorption of the weaker bases has no obvious effect on the C7H 1 1
cation, but the NMR spectra show that the stronger bases deprotonate the cation to
form the neutral C7H 11 diene. All spectra were measured at 298 K, with the
exception o f acetone-d 6 which was m easured at 193 K to prevent aldol
condensation.
131
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
zeolite HZSM-5 samples containing the C7H 1 \+ cation and various adsorbed bases.
The lower-most spectrum is a control experiment with no added base. This spectrum
reveals the signature isotropic chemical shifts of this cation, 247, 148, 48, and 24
ppm; the full assignment of this spectrum is presented in Figure 4.6, which also
reports GIAO-MP2 theoretical chemical shifts for this cation (modeled as a gas
phase species) as well as for 1, 3 -dimethylcyclopenta-1,3 -diene, the most stable
olefin obtained by deprotonation o f C7H 11+. As shown previously, 8 this level of
13 +
theory consistently overestimates the C isotropic shifts of C7H 1 1 by approximately
5 ppm.
13
For the five examples of co-adsorbed bases shown in Figure 4.5, C signals
from the bases were minimized by the use of compounds with 1 3 C levels found in
nature, and in some cases also with deuterium substitution to further reduce the cross
polarization signal intensity. It is clear from Figure 4.5 that nitromethane, acetone,
and DMF, which all have proton affinities well below the deprotonation energy of
C7H n +, do not deprotonate the cation. However, coadsorption o f the more basic
pyridine has a profound effect: the C7H n + cation is deprotonated. The most intense
13
C signals observed following adsorption of pyridine are consistent with those of
for 1, 3-dimethylcyclopenta-1, 3-diene (see Figure 4.6), and these assignments are
13
also supported by dipolar dephasing experiments. C spectra obtained with co
adsorbed pyridine invariably show a small signal at 8 6 ppm due to a minority
132
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(134)
19 (19)
19(18)
ca. 147 (145)
ca. 4 0 (48)
ca. 147 ( 144)
125 (121)
148 (153)
24 (29)
b)
Figure 4.6. Experimental and theoretical (GIAO-MP2/tzp/dz//B3LYHP/6-
311G*, in parenthesis) 13C chemical shifts (ppm) for: (a) C7H10; and (b)
C7H11+. Only symmetry distinct atoms are annotated for C7H11+. The
resonance predicted by theory to be at 134 ppm in the spectrum of C7Hl 1
can not be resolved in our experimental spectra as a result of the broad
lines for this species in the zeolite and consequent spectral overlap.
133
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 equiv,
1 equiv.
24
No added base
148
247
T T H ' T T ffl m 1 1 I I I 1! I I f i l l I'l l I 1 111 1 1 1 ' T T T T T
350 300 250 200 150 100 50 0 -50
ppm
13
Figure 4.7. C CP/MAS NMR spectra o f samples prepared by forming small
amounts (less than 0.05 equiv.) o f the C7H n +cation in zeolite HZSM-5 and then
absorbing 0, 1, or 2 equiv. of dimethylacetamide (DMA). Deprotonation o f the
C7H n + cation occurs with 2, but not 1 equiv. DMA. All spectra were measured at
298 K.
134
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
product (vide infra). Adsorption of P(CH3 )3 almost quantitatively deprotonated the
C7H 1 1 +cation to 1,3 -dimethylcyclopenta-1, 3-diene, there was a very small amount
of a second product (vide infra). The 1 3 C signal at 5.9 ppm is due to P(CH3)3; this is
one of the more intense and better resolved signals from co-adsorbed bases in Figure
4.5.
The results in Figure 4.5 experim entally bracket the enthalpy o f
deprotonation of the C7H U+ cation between 212 kcal and 222 kcal/mol. Coadsorption
o f DMA (Exp. PA = 217 kcal/mol) helps to narrow these limits. Figure 4.7 shows
that co-adsorption of 1 equiv. of DMA has little if any effect on the C7H n +cation,
but a higher loading, 2 equiv., deprotonates some o f the cation to form the diene.
From this result, we conclude that the deprotonation enthalpy of the C7H U+cation is
slightly greater than 217 kcal/mol, very close to the best theoretical value of 215.6
kcal/mol (MP4(sdtq)/6-311G).5
Figure 4.8 reveals that some, but not all, o f the CyHn* cations are
deprotonated by the first 0.5 equiv. of pyridine. Since the yield of the C7H n + cation
prepared by the method used here typically corresponds to 0.01 - 0.04 equiv. of all
Bronsted sites o f the zeolite, this suggests that not all Bronsted sites are equally
preferred as adsorption sites. No C7H n + cation remains after adsorption of a full
13
equivalent of pyridine. The C signal at 8 6 ppm, which reproducibly forms when
135
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4.8. CP/MAS NMR spectra of samples prepared by forming small amounts
(less than 0.05 equiv.) o f the C7H n +cation in zeolite HZSM-5 and then absorbing
15 13
the indicated amounts o f pyridine (Py)- N. C spectra are shown for several
loadings and an 15N spectrum is shown for the 2-equiv. sample. The 8 6 ppm
resonance due to a small amount of the pyridinium complex o f the C7H n + cation,
c 7 H n py+>is indicated in the 13C spectra. This species is not resolved in the 1 5 N
spectrum which shows a signal at — 174 ppm due to pyridine protonated by Br0 nsted
sites in the zeolite and a signal at -99 ppm due to pyridine on Lewis sites in the
13 15
catalyst binder. All C spectra were measured at 298 K, and the N spectrum was
measured at 123 K.
136
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
-174
-99
0 -1 0 0 -2 0 0 -3 0 0 -40C -5 0 0 200 100
125
146
86
2 equiv.
1 equiv.
No added base
-5 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0
ppm
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4.9. I3C CP/MAS NMR spectra of samples prepared by forming small
amounts (less than 0.05 equiv.) of the C7H U+ cation in zeolite HZSM-5 and then
absorbing an excess (typically 2 equiv.) of the indicated base. Numbers in
parentheses denote the experimental gas-phase proton affinities of the bases in units
of kcal/mol. The 72 ppm signal from the ammonium complex of C7H n +,
C7H u NH3 +, is indicated for the spectrum of the sample prepared by co-adsorption of
ammonia. As a result of the use of cross polarization, physisorbed CO is not seen in
the spectrum of the sample prepared with this co-adsorbate. All spectra were
measured at 298 K.
138
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
119
149
NH3 (204)
I .
H 2O (1 6 5 )
CO (1 4 2 )
No added base
3 5 0 3 0 0 2 5 0 2 0 0 150 1 00 50 0 -5 0
ppm
139
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
cryopumped onto the zeolite (containing the cation) at 77 K and 40 Torr. In every
case, a modest amount of CO was physisorbed by the zeolite to yield a peak at 190
ppm in Bloch decay spectra (not shown), but no reaction products were observed.
Binding of CO to the C7H 11+ cation was also not observed when the sample was
cooled to ca. 123 K in the NMR probe and acylium ions also did not form upon
sample heating in the probe. We consider the experimental finding that the C7H n +
cation does not bind CO in zeolite HZSM-5 to form an acylium ion to be
unequivocal, completely inline with our theoretical predictions.
In contrast, the C7H 11+ cation reacted quantitatively with NH 3 (Exp. PA =
204.0 kcal/mol) to form a new species, C7H n N H 3 + , with a distinctive 1 3 C
resonance at 72 ppm, exactly where one would expect to find a tertiary carbon bound
to an ammonium (-NH3+) group. This spectrum is assigned to species shown in
Figure 4.1c. Recall that adsorption of pyridine converted a minor amount of the
+ 13
C7H 11 cation into a species with a C shift of 8 6 ppm. By analogy to ammonium
species, the minority reaction product with pyridine is assigned to the pyridinium ion
+ +
C7H 11 Py shown in Figure 4.4c.
PH3, with an experimental proton affinity of 188 kcal/mol, is much less basic
than NH 3, and Figure 4.10 shows that the C7H n +cation is not deprotonated by this
13
base. The C resonance for the carbon bonded to phosphorous is at 50 ppm, and is
140
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31p -219
-67
-5 0 -1 0 0 -1 5 0 -2 0 0 -2 5 0 -3 0 0 1 0 0 5 0 0
13c
2 equiv, PH3, dipolar dephasing
ppm
2 equiv. PH3, cross polarization
No added base, dipolar dephasing
P P m
No added base, cross polarization
m i I iiif^ ln ,, iiiyjn^jj^ inir^lni \ m iw iiift
| r i 111......1 ' " | " i i " ' i T [ 1 1 1 1 1 11 n r j T T T q - r r
3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0
Figure 4.10. CP/MAS NMR spectra of samples prepared by forming small amounts
(less than 0.05 equiv.) of the C7H 11 cation in zeolite HZSM-5 and then absorbing
1 3
the indicated amounts o f phosphine, PH3 . C spectra are shown for 0 and 2 equiv.
3 1
and a P spectrum is shown for the 2-equiv. sample. The 44-ppm resonance due to
the phosphorous-substituted quaternary carbon of the phosphonium complex of
C7H n + , C7 H h +PH 3+, is clearly resolved in a 13C spectrum measured with dipolar
3 1
dephasing. This species is also observed at -67 ppm in the P spectrum which also
shows a signal at — 219 ppm due to phosphine protonated by Br0 nsted sites in the
zeolite. All spectra were measured at 298 K.
141
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
poorly resolved from signals due to the other carbons in the complex. We resolved
this signal using a dipolar dephasing (interrupted decoupling) pulse sequence at
+
reduced temperature (Figure 4.10). Further evidence for assignment to C7H 11 PH 3
31 * 3 1
was a P resonance at -67 ppm. The dominant signal in the P spectrum, at -219
ppm, is due to unreacted PH3 hydrogen bonded to Bransted sites on the catalyst. By
a similar procedure we were able to determine that P(CH3 )3 does indeed form a
+ . +
phosphonium complex, C7HnP(CH 3 )3 with the C7H 11 cation as a minority product
(not shown). In this case, steric constraints prevent complete reaction, and
deprotonation to form the neutral diene is the preferred pathway in the zeolite.
4.4 Discussion
Excellent agreement between experimental results and theoretical predictions
was obtained for a wide variety o f coadsorbed bases. The B3LYP/6-311G*
calculations correctly predicted the equilibrium state of the carbenium ion and each
coadsorbed base. The only complication is the known overestimation o f the
deprotonation enthalpy of C7H n + at the B3LYP/6-311G* level of theory.
Weaker bases such as CO, H 2 O, nitromethane, DME, acetone, and DMF
neither deprotonated C7H u +nor reacted with it to form an onium ion. Rather, they
formed hydrogen bonded complexes. In contrast, coadsorption of the strong bases
pyridine and P(CH3)3 into the zeolite deprotonated the C7H n + cation to form the
neutral cyclic diene, 1, 3-dimethylcyclopenta-l,3-diene. In a recent study o f the
142
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
mechanism of methanol to hydrocarbon catalysis, we showed that (in the absence of
a coadsorbate) the adsorption complex of this neutral diene on a zeolite Bronsted site
was only 2 kcal/mol less stable than the ion pair o f the C7H n + cation and the
conjugate base of the Bronsted site. We proposed that the higher energy state was
thermally accessible under reaction conditions and that the neutral diene was
methylated as one of a sequence of steps in a catalytic cycle by which methanol was
converted to olefins. Neutral 1, 3-dimethylcyclopenta-1,3-diene was not observed
experim entally in that investigation. The observation here, that 1, 3-
dimethylcyclopenta-1,3-diene is liberated on the zeolite by a stronger base, does not
prove that cyclic dienes are involved in the mechanism of hydrocarbon synthesis
from methanol. However, had the opposite result been obtained, that the C7H n +
cation formed irreversibly on the zeolite and could not be deprotonated, the proposed
mechanism would have been undermined.
Coadsorption o f DMA is a particularly interesting case. Using a combination
of NMR and theoretical methods, we recently showed that the degree of proton
transfer from a zeolite Bronsted site to acetone was increased by the coadsorption of
23 24
nitromethane or similar weak bases. Our theoretical calculations indicate that the
interaction between coadsorbates and acetone increases acetone's proton affinity,
thus promoting proton transfer. In this study, coadsorption of 1 equiv. of DMA did
not appear to deprotonate C7H n + , while 2 equiv. did result in some deprotonation.
As expected, at higher loading there is also an effective increase in the basicity of
143
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DMA due to the interaction w ith neighboring molecules, which promotes
deprotonation of the carbenium ion, an independent demonstration of the previously
reported solvent effect in zeolite catalysis.
In this study we theoretically obtained geometries for 11 C7H 114— :B
complexes with C-H—X hydrogen bonds (Figure 4.1a and Figure 4.2). Most
previously observed C-H—O hydrogen bonds are quite weak, giving (absolute)
1
binding energies o f less than 4 kcal/mol. In this study we obtain much stronger
binding energies that range from -10.3 to -17.6 kcal/mol. The very strong nature of
these hydrogen bonds is clearly related to the very high acidity of the carbenium
proton. Vargas, et al.2 5 proposed that the relationship between the C-H— O
hydrogen bond strength and the deprotonation enthalpy of the C-H proton is linear;
De = 0.0551(AH(jgpro|) -23.6 (values in kcal/mol). From this relationship and our
-f-
calculated deprotonation enthalpy for C7H 11 , we can predict a hydrogen bond
enthalpy o f -11.7 kcal/mol. This value is similar to our calculated binding energy for
the C7H i1+- H 20 complex (-12.6 kcal/mol). There does not appear to be a well-
defined trend in the binding energies for the C-H—O bonds formed by all o f the
bases, although as the difference between the proton affinity of the base and the
deprotonation enthalpy of C7H n +becomes less, the binding energy tends to increase.
Less commonly observed are C-H—N hydrogen bonds 1, as exhibited in the NH 3
and pyridine complexes. These also are generally assumed to be very weak
interactions (< 4 kcal/mol), whereas our complexes are strongly bound (-12.2
144
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
kcal/mol for NH3 and -11.1 kcal/mol for pyridine). Even more unusual are the C-H—
-C hydrogen bond in the CO complex, and the two C-H—P hydrogen bonds in the
PH 3 and P(CH3 )3 complexes. Not surprisingly, the C-H—C hydrogen bond is quite
weak (-2.2 kcal/mol), as the carbon in CO is very weakly basic. The C-H—P
hydrogen bonds are relatively weak in comparison to those involving oxygen as an
acceptor, despite the relatively high proton affinities of the bases. This may be due to
the high s character of the lone pair on phosphorous, and thus its unavailability for
hydrogen bonding.26 We are not aware of any previous experimental reports of C-
H—C or C-H—P hydrogen bonds.
We were also able to theoretically obtain stable geometries for six complexes
which have 71-cation interactions between the olefin double bond and the acidic
proton of the protonated bases. The weak basicity of PH 3 results in a high energy
C7H 10—PH4 + complex that certainly does not exist under experimental conditions.
However, formation of the other five complexes is exothermic, with values ranging
from -6.3 to -15.0 kcal/mol below the sum o f the total energies of the isolated
reactants. These binding energies are quite reasonable. For example, the u-cation
binding energy between Na+ and ethylene is ~ -13.6 kcal/mol.27
Onium ions were observed experimentally in every case for which theory
predicted these to be the most stable states. For reaction with NH 3 and PH3
experiment verified that the onium ions were the major products. Onium ions were
minority products with pyridine and P(CH3 )3 . However, these are very large ions.
145
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Given that the channels of HZSM-5 are only -5.5 A in diameter, accommodation of
bulky onium ions such as shown in Figures 4c and d will be difficult, even in the
more spacious channel intersections. We did not seek to model the effect of zeolite
framework topology on the accessibility of the three states.
A fundamental concern in efforts to quantify the acid or base strengths of
surfaces is that if adsorption of probe molecules occurs irreversibly, thermodynamic
equilibrium will not be established. Here we saw that the reactions between the
carbenium ion and coadsorbed bases went to equilibrium in the zeolite, with the
position of equilibrium being that predicted by gas phase thermochemical properties.
Implicit in the notion of reversibility is the time scale of the measurement, which
was quite slow in this case, but not slower than other spectroscopy-based methods
for evaluating the strengths of solid acids.
4.5 Conclusions
Our recent theoretical approach for predicting which carbenium ions will be
persistent species in acidic zeolites has been validated experimentally by showing
that C7H u+is deprotonated to form the C7H 10 cyclic diene by a base with gas phase
proton affinity greater than the calculated deprotonation enthalpy (215.6 kcal/mol) of
C7H 11 Extensions of this theoretical approach also made a number o f specific
predictions about onium ion formation with a number of bases. These predictions
were verified experimentally with the caveat that bulky ions such as those formed
between C7H n + and pyridine and P(CH3 )3 (Figures 4.4c and d) were minority
146
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
products due to steric constraints in the zeolite. Although phosphonium and
ammonium ion formation was quantitative for PH 3 and NH 3 , respectively, no
acylium cation was obtained with CO and no oxonium ion was obtained with
H 2 O— all in agreement with theoretical predictions. We also obtained theoretical
geometries for several complexes that exhibit unusual C-H—X hydrogen bonds.
147
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.6 References and notes
(1) Jeffrey, G. A. An Introduction to Hydrogen Bonding; Oxford University
Press: New York, 1997.
(2) Olah, G. A.; Laali, K., Wang, Q.; Prakash, G. K. S. Onium Ions; John Wiley
and Sons, Inc.: New York, 1998.
(3) White, J. L.; Beck, L. W.; Haw, J. F. J. Am. Chem. Soc. 1992,114, 6182-
6189.
(4) Haw, J. F.; Richardson, B. R.; Oshiro, I. S.; Lazo, N. L.; Speed, J. A. J. Am.
Chem. Soc. 1989, 111, 2052-2058.
(5) Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 1998,120, 11804-11805.
(6 ) Xu, T.; Haw, J. F. J. Am. Chem. Soc. 1994,116, 7753-7759.
(7) Xu, T.; Barich, D. H.; Goguen, P. W.; Song, W.; Wang, Z.; Nicholas, J. B.;
Haw, J. F. J. Am. Chem. Soc. 1998,120,4025-4026.
(8) Haw, J. F.; Nicholas, J. B.; Song, W.; Deng, F.; Wang, Z.; Heneghan, C. S.
J. Am. Chem. Soc. 2000,122, 4763-4775.
(9) Haw, J. F.; Goguen, P. W.; Xu, T.; Skloss, T. W.; Song, W.; Wang, Z. Angew.
Chem. 1998, 37, 948-949.
(10) Goguen, P. W.; Xu, T.; Barich, D. H.; Skloss, T. W.; Song, W.; Wang, Z.;
Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 1998,120, 2651-2652.
(11) Barich, D. H.; Xu, T.; Song, W.; Wang, Z.; Deng, F.; Haw, J. F. J. Phys.
Chem. B 1998,102, 7163-7168.
(12) The experimental proton affinity values reported here are the most recent
values or the values with the smallest uncertainties listed in the NIST
Chemistry WebBook (http://webbook.nist.gov/chemistry/).
(13) Xu, T.; Haw, J. F. Topics in Catalysis 1997, 4, 109-118.
(14) Munson, E. J.; Murray, D. K.; Haw, J. F. J. Catal. 1993,141, 733-736.
(15) Haw, J. F.; Nicholas, J. B.; Xu, T.; Beck, L. W.; Ferguson, D. B. Acc. Chem.
Res. 1996, 29, 259-267.
(16) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652.
148
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(17) Hehre, W. J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. Ab Initio Molecular
Orbital Theory', John Wiley & Sons: New York, 1986.
(18) Frisch, M. J. et. al, Gaussian 98, Revision A.4. Gaussian, Inc.: Pittsburgh,
PA, 1998.
(19) Gauss, J. Chem. Phys. Lett. 1992,191, 614-620.
(20) Schafer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 9 7 ,2571-2577.
(21) Stanton, J. F.; Gauss, J.; Watts, J. D.; Lauderdale, W. J.; Bartlett, R. J.,
ACES II, an ab initio quantum chemical program system.
(22) Xu, T.; Barich, D. H.; Torres, P. D.; Nicholas, J. B.; Haw, J. F. J. Am. Chem.
Soc. 1997,119,396-405.
(23) Haw, J. F.; Xu, T.; Nicholas, J. B.; Goguen, P. W. Nature 1997, 389, 832.
(24) Nicholas, J. B. Topics in Catalysis 1999, 9(3-4), 181-189.
(25) Vargas, R.; Garza, J.; Dixon, D.A.; Hay, B.P. J. Am. Chem. Soc. 2000,104,
5115-5121.
(26) a. Dixon, D.A.; Dunning, T. H., Jr.; Eades, R. A.; Gassman, P. G. J. Am.
Chem. Soc. 1983,105, 7011; b. Cherry, W.; Epiotis, N. D. J. Am. Chem. Soc.
1976, 98, 1135.
(27) Feller, D. Chem. Phys. Letters 2000, 322, 543-548.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 5
Synthesis of a Benzenium Ion in a Zeolite with Use of
Catalytic Flow Reactor
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.1 Results and discussions
The observation and mechanistic significance of carbenium ions in acidic
1-3
zeolites is one o f the central problems in heterogeneous catalysis. Early
speculation included suggestions of superacidic zeolites and carbenium intermediates
of all types. It is now generally recognized that zeolite acid strength is more in line
with conventional strong acids, and the characterization o f a long-lived, free
carbenium ion within a zeolite is a rare achievement. Previously, only three types of
long-lived carbenium ions have been unambiguously identified in zeolites, mostly by
13 4 9
C solid-state NM R with magic angle spinning. In 1989 Haw et al. identified
alkyl-substituted cyclopentenyl cations similar to 1 which formed in the low-
temperature reactions of propene on zeolite HY. Cyclopentenyl cations have since
been shown to form on several zeolites from a variety o f olefins and their
precursors.5,6,9 In 1994 Xu and Haw7 identified several indanyl cations, including 2,
g
which formed from styrene on zeolite HZSM-5. In 1995 Tao and M aciel
synthesized the trityl cation, (C 6H 5)3C+, in zeolite HY from the Friedel-Crafts
reaction of CCI4 and benzene. Cano et al. recently reported the synthesis of several
trityl cation derivatives inside zeolites Y and. 10 Previous NMR studies of carbenium
ions in zeolites (and many similar, but unsuccessful, attempts) used sample
preparation methods that differ in significant ways from the conditions used in
151
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
catalytic flow reactors. In particular, the organic precursor(s) and the zeolite were
invariably sealed such that nothing could escape the zeolite.
3
We prepared the benzenium cation 3 in zeolite HZSM-5 by a novel
experimental procedure. The key to the preparation o f 3 was the use of flow reactor
conditions for sample preparation; 1 1 we pulsed an aromatic hydrocarbon and an
excess of methanol onto a zeolite bed at 573 K with continuous He carrier gas flow,
allowed them to react for 4 s, and then rapidly quenched the sample temperature to
ambient. In a flow reactor, the coproduct (water) diffuses out of the zeolite
crystallites and is swept out o f the sample by carrier gas, leaving cation 3 trapped in
the zeolite. Our various attempts to observe 3 in sealed MAS NMR rotors were
uniformly unsuccessful; in the presence o f water the equilibrium concentration o f 3
in the zeolite is negligible.
We synthesized 3 in zeolite HZSM-5 a number o f times, pulsing either
13
benzene or toluene into the flow reactor with methanol, and we permuted the C
13
label in the reactants as an aid to spectral assignment. Representative C MAS
152
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.1 Selected 75.4 MHz 13C MAS spectra of benzenium ion 3 (206, 190,139,
58, and incompletely resolved signals between 23 and 26 ppm) on zeolite HZSM-5
(Si/Al = 19). Signals near 20 ppm are due to the methyl carbons of neutral aromatic
compounds, e.g., toluene and xylenes. Benzene or toluene and methanol were
injected into the flow reactor as a pulse and allowed to react for 4 s at 573 K before
13 12
quench: (a) 0.5 equiv. of benzene- Ce and 3 equiv. of methanol- C; (b) 0.5 equiv.
13 12
of toluene-ring- C$ and 2.5 equiv. of methanol- C; (c) 0.5 equiv. of benzene and 3
13 13
equiv. of methanol- C; (d) 0.5 equiv. of toluene and 2.5 equiv. o f methanol- C; and
13 12
(e) 0.5 equiv. of toluene— C and 2.5 equiv. of methanol- C. 1 equiv. on this
catalyst corresponds to 0.58 mmol reactant/g of zeolite. An asterisk denotes spinning
sideband.
153
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
p i i T p n q i i'i 1 1» i » 1 1 r i i T ' p m p r i i ] i m \
350 300 250 200 150 100 50 0 -50
ppm
Dimethyl ethei Methanol
20
50 100
ppm
154
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.2 B3LYP/6-311G** geometry o f 3 optimized in Cs symmetry. Selected
internal coordinates are shown (in A). The angle C-7 to C-l to C-4 is 113.9, C- 8 to
C -l to C-4 is 137.4, and C- 8 to C -l to C-7 is 108.7. Theoretical 13C isotropic
chemical shifts are calculated at GIAO-MP2/tzp/dz and referenced to TMS at the
same level of theory (absolute shielding = 198.8 ppm): C -l, 65 ppm; C-2 and C-6 ,
209 ppm; C-3 and C-5, 139 ppm; C-4, 191 ppm; C-7, 23 ppm; C-8 , 35 ppm; C-9 and
C-10, 28 ppm; C-l 1, 29 ppm.
155
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
spectra showing 3 prepared from various precursors are summarized in Figure 5.1.
Signals due to the cation were also enhanced by cross polarization (not shown).
Verification o f the identity of 3 was obtained with theoretical methods and by
13C
comparison with shifts of model carbenium ions such as heptamethylbenzenium
ion. 12 We optimized the geometry of 3 at the B3LYP13 level using the 6-311G**
basis set14 and the program NWChem . 15 GIAO-M P216 13C chemical shifts were
17 18
determined with the tzp (C) and dz (H) basis sets by using ACES II. The Cs form
of 3 is shown in Figure 5.2. The Cs conformer is less than 0.1 kcal/mol above
that of the Ci global minimum; these structures are interconverted by rotation of
methyl group C-7. Differences in other internal coordinates between the two
conformers are negligible. The GIAO-MP2 calculations were tractable only for the
Cs conformer. The geminal methyl groups C-7 and C- 8 are nonequivalent according
to theory, but this is not observed experimentally due to dynamical averaging of
these geminal methyl groups at 298 K. The agreement between the observed (Figure
1) and predicted shifts for 3 (caption to Figure 5.2) is at least as good as seen for
19
other benzenium cations, and given the exotic chemical shifts of benzenium
20
cations, this agreement is powerful evidence for the proposed assignment. The
observed shifts are also in line with experimental values for heptaalkylbenzenium
cations in solution. 12,21
156
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
It is noteworthy that the formation of cation 3 occurs not by protonation, but
rather by alkylation. This is consistent with our recent observation that the
benzenium ion is not a stable state for benzene H/D exchange on acidic
22
zeolites. In the absence of water or another nucleophilic adsorbate, decomposition
of 3 in the zeolite would require methylation of the conjugate base site on the zeolite
framework. Therefore, the thermodynamic reference state for evaluating the stability
o f gem-dialkylbenzenium cations in zeolites is not a neutral complex of the proton
form of the zeolite but rather of the methyl (methoxyl) form of the zeolite.
A recent mechanistic study of the economically important process of toluene
23
disproportionation in HZSM-5 proposed key roles for benzenium cations; those
cations were somewhat different from the cation reported here, but the observation
of 3 suggests that theoretical and additional experimental studies of benzenium ion
mechanisms in zeolite catalysis might be important. We believe that 3 forms in the
channel intersections o f HZSM-5, which are more spacious than the channels
themselves, and that steric constraints account for the substitution pattern in which
C-4 and C - 6 are not alkylated.
157
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.2 References and notes
(1) Haw, J. F.; Nicholas, J. B.; Xu, T.; Beck, L. W.; Ferguson, D. B. Acc. Chem.
Res. 1996, 29, 259-267.
(2) van Santen, R. A.; Kramer, G. J. Chem. Rev. 1995, 95, 637-660.
(3) Corma, A. Chem. Rev. 1995, 95, 559-614.
(4) Haw, J. F.; Richardson, B. R.; Oshiro, I. S.; Lazo, N. L.; Speed, J. A. J. Am.
Chem. Soc. 1989, 111, 2052-2058.
(5) Xu, T.; Haw, J. F. J. Am. Chem. Soc. 1994, 116, 7753-7759.
(6 ) Stepanov, A. G.; Vladimir N. S.; Zamaraev, K. I. Chem. Eur. J. 1996, 2,
157-167.
(7) Xu, I.; Haw, J. F. J. Am. Chem. Soc. 1994, 116, 10188-10195.
(8 ) Tao, T.; Maciel, G. E. J. Am. Chem. Soc. 1995, 117, 12889-12890.
(9) Oliver, F. G.; Munson, E. J.; Haw, J. F. J. Phys. Chem. 1992, 96, 8106-8111.
(10) Cano, M. L.; Corma, A.; Fomes, V.; Garcia, H.; Miranda, M. A.; Baerlocher,
C ; Lengauer, C. J. Am. Chem. Soc. 1996, 118, 11006-11013.
(11) In brief, samples were prepared by using a device we call a "pulse-quench
catalytic reactor" that will be described in detail elsewhere. It is much like a
standard benchtop flow reactor equipped for pulsed introduction o f reagents,
and differs primarily in the extensive use of computer-controlled valves to
rapidly reduce the temperature of the gas flowing over the catalyst. A very
rapid thermal quench is probably not essential for the observation of 3, and
we anticipate that replication of this result could be achieved by the diligent
application o f standard flow reactors. Quenched samples are sealed in the
reactor, and transfer o f the catalyst to MAS rotors is carried out in a
glovebox; thus, once prepared, samples of zeolite containing 3 were not
exposed to moisture.
(12) Rezvukhin, A. I.; Mamatyuk, V. I.; Koptyug, V. A. Z. Org. Khim. 1972, 8 ,
2443.
158
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(13) Becke, A. D. Phys. Rev. A 1988, 38, 3098-3100.
(14) Hehre, W. J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. Ab Initio Molecular
Orbital Theory; John Wiley & Sons: New York, 1986.
(15) Guest, M. F.; Apra, E.; Bemholdt, D. E.; Friichtl, H. A.; Harrison, R. J.;
Kendall, R. A.; Kutteh, R. A.; Long, X.; Nicholas, J. B.; Nichols, J. A.;
Taylor, H. L.; Wong, A. T.; Fann, G. I.; Littlefield, R. J.; Nieplocha, J.
Future Generations Comput. Syst. 1996, 12, 273-289.
(16) Gauss, J. Chem. Phys. Lett. 1992, 191, 614-620.
(17) Schafer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571-2577.
(18) ACES II, an ab initio quantum chemical program system. Stanton, J. F.;
Gauss, J.; Watts, J. D.; Lauderdale, W. J.; Bartlett, R. J.
(19) Xu, T.; Barich, D. H.; Torres, P. D.; Haw, J. F. J. Am. Chem. Soc. 1997,
119, 406-414.
(20) Other evidence for the assignment is the result of dipolar dephasing
experiments conducted at both ambient and low temperatures (not shown)
which confirm that C-2, C-4, and C- 6 are substituted. This result requires
two substituents for sp3-hybridized carbon C-2 (58 ppm).
(21) Olah, G. A.; Spear, R. J.; Messina, G.; Westerman, P. W. J. Am. Chem. Soc.
1975, 97, 4051-4055.
(22) Beck, L. W.; Xu, T.; Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 1995,
117,11594-11595.
(23) Xiong, Y.; Rodewald, P. G.; Chang, C. D. J. Am. Chem. Soc. 1995, 117,
9427-9431.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 6
A Persistent Carbenium Ion on the Methanol-to-Olefin
Catalyst HSAPO-34: Acetone Shows the Way
160
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6.1 Introduction
The catalytic conversion o f methanol to olefins (MTO) 1 is a major emerging
chemical technology and the key step in the conversion of natural gas to polyolefms
using otherwise mature chemistry. The most promising catalysts for commercial
trials are solid acids based on silico-aluminophosphates o f the chabazite (CHA)
structure, most simply HSAPO-34. The topology o f this catalyst is characterized by
cages of ca. 1.0 nm by 0.7 nm diameter that are interconnected through 8 -ring
windows o f ca. 0.38 nm in diameter. Thus, reactants such as methanol and
dimethylether and products such as ethylene, propene, and 1 -butene may freely
diffuse through active catalysts, but larger molecules, even those with kinetic
diameters comparable to that of isobutane, are not adsorbed.
Applied research in MTO catalysis is directed toward increasing both
ethylene selectivity and catalyst lifetime, and this has motivated basic research into
the mechanisms of olefin synthesis and catalyst deactivation. In a recent
communication, we used pulse-quench reactors to study MTO chemistry on
HSAPO-34. GC and GC-MS analysis of the volatile products confirmed the
existence of a kinetic induction period on this catalyst. Only a modest conversion of
methanol to hydrocarbons was observed following the first pulse of methanol onto a
fresh catalyst, but essentially 1 0 0 % conversion was obtained for a second, identical
pulse applied a few minutes later. Furthermore, GC-MS showed that when the first
13
pulse was methanol- C and the second pulse was methanol- X , the predominant
161
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
isotopomer of ethylene in the product stream had one carbon of each isotope. We
13
used C MAS NMR analysis o f quenched samples from these reactor studies to
understand the changes occurring on HSAPO-34 following first exposure of
m ethanol to create a working catalyst. Those m easurements showed that
methylbenzenes self-assemble in the cages of HSAPO-34 in the first few seconds
following methanol exposure, and that these “organic reaction centers” are part of
the active site of a working catalyst.
Closely related work on the aluminosilicate zeolite HZSM-5 showed that
methylbenzenes are synthesized in a mechanism that passes through cyclopentenyl
4
carbenium ions, most typically the 1, 3-dimethylcyclopentenyl cation la.
Cyclopentenyl cations were proposed to be organic reaction centers for MTO
chemistry through a mechanism involving deprotonation to neutral, cyclic dienes.
On HSAPO-34 we im agine an analogous m echanism in which neutral
methylbenzenes are in equilibrium with benzenium cations similar to a species
recently identified on HZSM-5.5 Cation l a forms on HZSM-5 with very high
selectivity using either ethylene, propene or methanol as a starting material,4 ’ 6’ 7 and
162
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
it is indefinitely persistent at room temperature. In contrast, in the study of HSAPO-
34 we did not observe any NM R signals that we could confidently assign to
carbenium ions, even though methylaromatics readily formed.
Our failure to observe cyclopentenyl cations on HSAPO-34 raised a number
of questions about the relative acid strengths of HSAPO-34 and HZSM-5 as well as
the generality of our claim that methylaromatics form through cyclopentenyl cation
intermediates on solid acids. We therefore explored other synthetic routes to
persistent carbenium ions in this catalyst. Procedures that previously formed robust
signals due to cation l a on HZSM-5 did indeed form very low intensity signals on
HSAPO-34, but small signals make for weak evidence and suggest that only a
fraction of the acid sites are sufficiently strong to sustain these cations. We were
13
successful when we turned to acetone as a precursor. The isotropic C chemical
13
shifts of acetone-2- C on HSAPO-34 do indeed show that most acid sites on this
g
material are weaker than those on HZSM-5, and that a few percent are stronger.
When samples of acetone on HSAPO-34 were heated in sealed MAS NMR rotors,
we observed robust signals unmistakably due to a cyclopentenyl carbenium ion with
an unusual substitution pattern, probably the heptamethyIcyclopenty 1 cation 2 a.
9
Theoretical chemical shift calculations at the GIAO-MP2/tzp/dz level supported this
assignment for most of the signals, but there was an appreciable discrepancy for C4 ,
C5 , where the C4 -C5 bond distance is predicted (B3LYP) to be lengthened by steric
163
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
repulsion. While the analogous carbons were in better agreement with the
predictions for a less-crowded heptamethylcyclohexenyl cation 3a, the characteristic,
more up-field C i, C3 shift for cyclohexenyl cations, reproduced theoretically, ruled
out this assignment.
We outline the mechanism by which cation 2 a could form from acetone in
HSAPO-34. We also present theoretical calculations of proton affinities of parent
olefins that suggest why 2 a could form to a greater extent than l a on this catalyst.
On HSAPO-34, cyclopentenyl cations are indeed intermediates in the formation of
methylbenzenes.
6.2 Experimental Section
Materials and Reagents. HSAPO-34 was prepared according to reference
10. XRD showed a pure crystalline phase with the CHA structure. The product was
calcined at 873 K for 10 hours to remove the template agent and pressed into 10-20
mesh pellets. The Bronsted site concentration was determined to be 1.1 mmol/g. In
typical experiments 0.3 g of catalyst was activated at 673 K under 600 seem He flow
13 13
for 2 h immediately prior to use in a pulse-quench reactor. Acetone-2- C, (% C)
and acetone-1, 3 - 13C2 were obtained from Cambridge Isotopes, Inc.
Catalysis Experiments. CAVERN experiments were perform ed as
gH12
described previously. ’ ’ For each experiment we activated ca. 300 mg of
164
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
calcined HSAPO-34 in the shallow-bed CAVERN at 723 K for 1 hr to remove water.
0.38 mmol of acetone was adsorbed at room temperature. The rotor was then sealed
in the CAVERN and transferred to the NM R probe where it was heated to
progressively higher temperatures followed by spectral acquisition at room
temperature.
7 13
Samples were also prepared using a pulse-quench reactor ’ in a single
pulse configuration. For each experiment we loaded the reactor with a cylindrical
bed (7.5 mm diameter by 8 mm in length, typical weight o f 300 mg) o f fresh
HSAPO-34 pellets. In every case, the catalyst bed was activated in place
immediately prior to use by heating 1 hr in flowing helium and then equilibrating at
the desired reaction temperature. Previous studies have shown that the temperature
o f the catalyst pellets decreases 150 K in the first 170 ms of a quench. After
quenching each reacted catalyst sample, the reactor was sealed off and transferred
into a glove box filled with nitrogen. The catalyst pellets were ground and
transferred to a 7.5-mm MAS rotor which was sealed with a Kel-F end-cap similar to
that used in the CAVERN studies.
Gas Chromatography. A Hewlett-Packard Model 6890 gas chromatograph
with flame-ionization detector was used to analyze gases sampled from the reactor
product streams using a Valeo valve. The column was 150 m dhl50 (Supelco)
operated isothermally at 323 K to permit sampling of the gas stream more frequently
than the total analysis time for any given sample. Acetic acid did not elute from the
column using these conditions.
13
NMR Spectroscopy. C solid state NMR experiments were performed with
magic angle spinning (MAS) on a modified Chemagnetics CMX-300 MHz
165
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
spectrometer operating at 75.4 MHz for C. Hexamethylbenzene (17.4 ppm) was
used as an external chemical shift standard, and all 13C chemical shifts are reported
relative to TMS. Chemagnetics-style pencil probes spun 7.5 mm zirconia rotors at
typically 6.5 kHz with active spin speed control (± 3 Hz).
13
Typical C experiments included: cross polarization (CP, contact time = 2
ms, pulse delay = 1 s, 4000 transients); cross polarization with interrupted
decoupling (contact time = 2 ms, pulse delay = 1 s, 4000 transients, dipolar
dephasing time of 50 ps); single pulse excitation with proton decoupling (Bloch
decay, pulse delay = 10 s, 400 transients).
Theoretical Details. The geometries of the three carbenium ions (la-3a) and
their parent olefins were studied using density functional theory (DFT). In each case
the molecules were fully optimized using the hybrid B3LYP exchange-correlation
14
functional and the 6-311G* basis set. Analytic frequency calculations were done to
verify that all species were minima on the potential energy surface, and to obtain the
13
thermodynamic data needed for the determination of enthalpies. The C NMR
9
chemical shifts were calculated at the MP2 level using the GIAO method. For the
NMR calculations we used a polarized triple-zeta basis set on the carbons, and a
double-zeta basis set on hydrogens. 1 5 We will refer to this basis set as tzp/dz. The
13
reported C chemical shifts are relative to those of the chemical shift standard TMS,
16
obtained by the same computational procedures. We used Gaussian98 for all of the
calculations.
166
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6.3 Results
CAVERN Experim ents. Figure 6.1 reports 13C CP/MAS spectra from a
CAVERN study of 1 equivalent of acetone-2-13C on HSAPO-34. This loading
corresponds to 1.1 mmol/g, and approximately one molecule per cage. In each case
the NMR spectrum was measured near room temperature following heating o f the
sample rotor in the probe for ca. 2 0 min at progressively higher temperatures, as
indicated in the Figure. The spectrum acquired prior to heating (293 K) shows that
most of the acetone was shifted downfield to 217 ppm as a result o f hydrogen
8 1 7
bonding to the Bronsted sites in the catalyst. ’ In expanded views (not shown) a
small shoulder can be discerned at ca. 226 ppm. The resonance at 30 ppm is due to
other isotopomers of acetone. After heating to 433 K, small signals were observed
due to dimeric diacetone alcohol 2 (75 ppm) and its dehydration product mesityl
oxide 3 (186 ppm and 205 ppm). A small resonance due to acetic acid is also
apparent at 181 ppm. Significantly greater conversion of acetone to these products
was achieved after heating to 473 K. In addition, small signals were also observed
at 244-253, 155 and 56 ppm; their assignment to cation 2 and the significance of this
species is central to this paper. Conversion of acetone was complete after heating to
513 K. This final spectrum shows the formation o f a significant amount of
methylaromatics. The resonance at 135 ppm is due to substituted aromatic carbons
167
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
135 I
4 / “ I
^ 181
75
155 i
25
253 244
513 K
30
56
254
473 K
205
186
75
181
X 4
433 K
217
30
293 K
350 300 250 200 150 100 50 0
-50
ppm
Figure 6.1. 75.4 MHz 13C CP/MAS NMR spectra from a CAVERN study of
13
acetone-2- C (1 eqiv.) on HSAPO-34. The sample was heated in the MAS rotor to
progressively higher temperatures, as indicated, and then spectra of the reaction
products in the catalyst were measured at 298 K. * denotes spinning sideband.
168
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
while the shoulder at 130 ppm is assigned to unsubstituted aromatic carbons.
Integration by spectral simulation suggests an average o f four substituents per ring.
Methyl groups on the benzene rings are seen at 19 ppm. The above assignments are
also consistent with Bloch decay spectra (not shown).
13
Figure 6.2 compares the CP/MAS spectrum of acetone-2- C after heating at
513 K with the dipolar dephasing (interrupted decoupling) spectrum o f the same
sample. O f particular interest are the signals at 244-253, 155, and 56 ppm, which all
survive dipolar dephasing and hence are emphasized in this spectrum. For rigid
molecules, neither CH nor CH2 carbon resonances are seen in dipolar dephasing
spectra. Figure 6.2 also supports the finding that the aromatics formed from acetone
in the HSAPO-34 cages are highly substituted as ca. 2/3 of the aromatic signal
survives dipolar dephasing.
13
A CAVERN study of acetone-1, 3- C is reported in Figure 6.3. Since many
13
of the C labels from the methyl groups in acetone ended up as methyl groups in
various products with similar chemical shifts, Figure 6.3 is generally less informative
than Figure 6.1. After heating to 433 K, a small signal was seen at 126 ppm due to
C3 of mesityl oxide. Some of the lable also found its way into the aromatic carbons
as well as the siganl at 244-253, 155 and 56 ppm.
169
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
56
19
155
244
253
dipolar
dephasing
135
130
217
cross
polarization
300 250 200
350 150 100 50 0 -50
ppm
Figure 6.2. 75.4 MHz 13C CP/MAS NMR spectra showing the effects o f dipolar
1 3
dephasing (50 ms) on the spectrum from Figure 1 o f acetone-2- C on HSAPO-34
after heating to a maximum o f 513 K. Spectra were measured at 298 K. The
essential observation is that the signals at 244-253 ppm, 155 ppm, and especially 56
ppm survive dipolar dephasing, supporting the assignment to cation 2a. Dipolar
dephasing also shows that most o f the aromatic carbons are substituted and that the
reaction products are rich in methyl groups. * denotes spinning sideband.
170
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
135
. . ■ s u a O i T
1 9
253 244
155
473 K
5 6
*
K
126
X 8 \
433 K
J
30
273 K
1
111111i i n 1 111111111111111h 1111111111111
350 300 250 200 150 100 50 0 -50
ppm
Figure 6.3. 75.4 MHz 13C CP/MAS NMR spectra from a CAVERN study o f
acetone-1, 3 -13C2 (1 eqiv.) on HSAPO-34. The sample was heated in the MAS rotor
to progressively higher temperatures, as indicated, and then spectra of the reaction
products in the catalyst were measured near 298 K.
171
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Pulse-Quench Studies. Figure 6.4 reports 13C CP/MAS spectra of samples
7 13
prepared in a pulse-quench reactor ’ at various temperatures as indicated on the
Figure. In each case ca. 300 mg of fresh catalyst was activated in the reactor at 723
K and then equilibrated at the desired reaction temperature prior to injection of 16 uL
1 3
o f acetone-2- C. In each case acetone reacted on the catalyst for 4 s with
continuous product removal by He flow, then the catalyst temperature was reduced
13
to near ambient by a rapid thermal quench. C CP/MAS spectra were measured at
room temperature. GC analysis showed that most o f the acetone passed through the
catalyst without conversion within the temperature range studied; however, some
acetone was retained, unreacted, on the catalyst and gave rise to the peak at 217 ppm
that was also seen in the CAVERN study. Remarkably, we also saw a peak at 226
ppm with a large relative intensity, indicating that acetone is adsorbed more strongly
to this site than to that responsible for the 217 ppm resonance. Aromatics were also
trapped in the HSAPO-34 cages in the samples prepared under flow conditions, but
these were substituted to a lesser degree at the highest temperature studied, 2 . 0
methyl groups per ring on average after reaction at 673 K. Figure 6.4 shows no
evidence of the signals at 244-253, 155, and 56 ppm.
Flow R eacto r Studies. A pulse-quench reactor was configured for
continuous injection o f acetone by replacing the injection valve with a
programmable syringe pump. Gas samples were taken periodically and analyzed by
172
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13C
135,
20
673 K
623 K
181
573 K
217
226
513 K
350 300 250 200 150 100 0 50 -50
ppm
Figure 6.4. 75.4 MHz 13C CP/MAS NMR spectra from pulse-quench studies of
acetone-2-13C on HSAPO-34. In each case a fresh catalyst bed (ca. 300 mg) was
activated in place and then equilibrated at one of the temperatures shown on the
13
figure. 16 uL of acetone-2- C was injected onto the catalyst bed and allowed to
react for 4 s in flowing He prior to a rapid thermal quench. All spectra were
measured near 298 K. The essential observation is the resolution of the 226 ppm
signal from acetone on a stronger adsorption site from that at 226 ppm for acetone on
a weaker adsorption site. Aromatics form at higher temperatures. * denotes spinning
sideband.
173
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Acetone
1-Butene
1 -----; -----j — -j---- 1 ---- j -----j---- 1 ---- j -----1 ---- 1 -----j---- 1 -----1 ---- 1 -----1 -----1 -----r
14 16 18 20 22
Retention Time
(min.)
Figure 6.5 Gas chromatographs (FID detection) sampling the product
streams o f reactions carried out on HSAPO-34 catalyst beds operated at
673 K: a. A control experiment showing activity and product selectivity
using pure methanol as feed (WHSV = 8 h r 1 ). Conversion is
quantitative, propene selectivity exceeds that of ethylene, and C4
selectivity is low.; b. Pure acetone (WHSV = 8 hr'1 ), sample taken 16
min after initiating feed. Conversion is much lower than with methanol,
and the C4 (but not C5 ) selectivity is clearly elevated.; c. Pure acetone
(WHSV = 8 h r'1 ), sample taken 28 min after initiating feed. The catalyst
is almost completely deactivated, and selectivity for both ethylene and 1 -
butene are elevated.
174
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
GC. Not surprisingly, the conversion of acetone on HSAPO-34 was lower than for
identical experiments using methanol as a feed, and the catalyst deactivated far more
rapidly, even when the acetone was diluted to 50 % v/v with water. Nevertheless,
the product selectivity from acetone was insightful. Figure 6.5a shows a
representative GC trace from the conversion o f methanol on HSAPO-34 at 673 K
(WHSV = 8 hr ). This measurement was taken on a working catalyst, after a very
short induction period, but well before deactivation. Methanol (and dimethylether)
were almost entirely converted to hydrocarbons in this “control experiment”— the
yield o f propene was greater than that of ethylene, and the selectivity for C4 products
was low.
Figures 6.5b and 6.5c are from an otherwise identical study of the conversion
of pure acetone (WHSV = 8 hr *). The sample for chromatogram 6.5b was taken
only 16 min after beginning the acetone flow. The ethylene/propene ratio was
similar to that obtained during methanol conversion on a working catalyst, but with
acetone a much higher C4 selectivity was seen early in the life of the catalyst. The
catalyst rapidly deactivated; Figure 5c is the analysis of a gas sample taken 28 min
after commencing the acetone flow. Total conversion was decreased by a factor of
ca. 10 in Figure 6.5c compared to Figure 6.5b. In Figure 6.5c, all C4 products were
greatly reduced on the partially deactivated catalyst with the exception o f 1 -butene,
175
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
which was only slightly reduced. As deactivation reduced the conversion of acetone,
the ethylene/propene ratio increased.
Theoretical Chemical Shifts for the Heptamethylcyclopentenyl Cation
2a. Three carbenium ions that could potentially be responsible for the observed
NM R data were theoretically considered. The B3LYP/6-311G* optimized
geometries of these are given in Figure 6 .6 . Symmetry-distinct C-C bond distances
13
are shown, as are the C chemical shifts, which are discussed below. We have
p resen ted the optim ized geom etry and NM R data for the 1,3-
dimethylcyclopentadienyl cation (la ) in previous work. We confirmed the presence
of this cation during MTO and MTG chemistry on HZSM-5. Cation la has
symmetry. The delocalized nature o f the positive charge is evident by the C 1-C2 and
C2-C3 bond lengths of 1.39 A, intermediate between those o f pure single and double
bonds. On HSAPO-34, the dipolar dephasing experiment im plies that the
cyclopentadienyl framework becomes completely methylated. Thus, the optimized
geometry o f the heptamethylcyclopentadienyl cation 2a is also shown in Figure 6 .
The steric repulsion between the eclipsed methyl groups in the 4 and 5 positions
causes significant distortion o f the ring, which is essentially planar in cation la.
Although cation 2a has C[ symmetry, this is only due to the hydrogens on the methyl
group in the 2 position. If these are ignored, the carbon skeleton has C2 symmetry.
176
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
la (C2v)
^ 1.5
C are
3a (Cs)
2a (Ci)
All other methyl
Figure 6 . 6 B3LYP/6-311G* optimizied geometries for cations la-3a. Selected symmetry-distinct C-C bond
lengths (A) are show in smaller print. Predicted MP2-GIAO/tzp/dz//B3LYP/6-311G* 13C NMR chemical shifts
for symmetry-distinct atoms are shown in larger, underlined print.
177
To simplify the graphics, we assume C2 symmetry in the presentation on the bond
length and NMR data in Figure 6 .6 . Comparing l a and 2a, the only significant
geometric difference is in the C4 -C5 bond length. This bond is 0.04 A longer in 2 a
than in la , consistent with the steric crowding of the eclipsed methyl groups. For
reasons that will be apparent when we discuss the theoretical shifts we also
considered the 1, 2, 3, 4, 4, 6 , 6 -heptamethylcyclohexadienyl cation 3a. This
carbenium ion contains the same methyl-substituted alyllic substructure as cations la
and 2a . Steric clashes are reduced in 3a by the additional methylene in the 5
position. The NMR signal from this methylene could potentially be lost in the peaks
from the various methyl groups. The calculated bond lengths in 3a are very similar
to those in la.
GIAO-MP2/tzp/dz theoretical 13C isotropic shifts for cations la , 2a, and 3a
are also reported in Figure 6 . The shifts of cation la have been commented on
4
previously, we focus here on the differences between the three cations. Predicted
13
C NMR chemical shifts for symmetry-distinct atoms are shown in larger,
underlined print in Figure 6 .6 . The NMR shift of C2 is relatively constant, varying
between 151 and 155 ppm for the three cations. The chemical shift of the Ci, C 3
carbons is similar for la and 2a (253 and 255 ppm, respectively), but much different
for 3a (237 ppm). Finally, the C4 , C5 carbons in la and 3a have the same shift, (50
ppm) whereas it is shifted downfield to 73 ppm for 2a.
178
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6.4 Discussion
The R elative Acid S trength of HSAPO-34. The acid strength of
19
aluminosilicate zeolties has been a contentious issue in the literature. While most
workers would agree that zeolites are not superacids, well-founded differences of
opinion remain about the merits of various method for measuring acid strengths and
13 13
the relative strengths of diverse solid acids. The C isotropic shift of acetone-2- C
g
is commonly applied as one operational measure of acid strength, and this effect has
17
been explored theoretically. The acetone shift is influenced by adsorption on either
Lewis or Bronsted sites. Figures 6.1 and 6.2 show that most adsorption sites in our
samples of HSAPO-34 induce an acetone shift of 219 ppm. Our interpretation is that
most Bransted acid sites in HSAPO-34 are effectively weaker than those in HZSM-
5. Preliminary calculations indicate that meaningful theoretical modeling of either
the deprotonation energy of an acid site in HSAPO-34 or the adsorption of acetone in
this material would require a periodic treatment of an entire cage. We have not yet
attempted such calculations. The acetone probe would measure a lower acid strength
if the deprotonation energy o f HSAPO-34 were indeed lower than for HZSM-5, but
it is also possible that the small size o f the SAPO-34 nanocage could prevent the
most favorable hydrogen-bonding geometry and hence the largest acetone shift.
The small peak at 226 ppm, which corresponds to a few percent o f the total acid sites
does imply an acid strength for a minority site that is stronger than even those in
HZSM-5. We searched for spectroscopic evidence o f Lewis sites without success;
179
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
for example, the A1 spectrum showed no extra-framework aluminum. Future
studies o f HSAPO-34 materials of diverse origin and composition will be needed
before the 226 ppm can be assigned to a specific structural feature.
A ssignm ent o f the 244-253, 155, and 56 ppm Signals to the
H eptam ethylcyclop en tenyl C ation 2a. The 1, 3-dimethyl and 1, 2, 3-
trimethylcyclopentenyl cations have been identified in numerous NMR studies of
microporous solid acids dating back to 1989.6 The most easily seen resonance for
these cations is a signal near 48 ppm that does not survive dipolar dephasing; this is
due to C4 , C5 which were invariably unsubstituted (methylene) carbons in all
previous studies. For the previously observed di- and trimethyl cations, C j, C3 are in
the vicinity of 247-250 ppm. These signals are sometimes broad or show evidence
o f structure, as here, and this has been interpreted as evidence o f interaction with the
zeolite framework. For the dimethyl cation C2 is near 147 ppm and does not
survive dipolar dephasing; it is near 155 ppm for the trimethyl cation and does
survive dephasing.
Like the analogous resonances for the other cylopentenyl cations, the 244-
253, 155, and 56 ppm signals here always appear as a constellation, and must be
from the same species. This com bination o f chemical shifts is uniquely
characteristic of a cyclopentenyl cation. The signal at 56 ppm is quite far downfield
o f the 48 ppm resonance previously seen for less-highly substituted cations. The
observation that this signal also survives dipolar dephasing can be explained, for a
180
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
cyclopentenyl cation, only by disubstitution on carbons C4 and C5 . Substitution
would also tend to shift this resonance downfield, as observed. If the substituents on
these carbons were diverse, the 56 ppm resonance would be broad; it is instead very
sharp, suggesting that the substituting groups on all cations are identical. The
dipolar dephasing spectrum in Figure 6.2 shows that nearly all o f the upfield signal
intensity is from methyl groups, and the substituents are so assigned.
In 2 a the theoretical chemical shifts for Q,C3, and C2 are in excellent
agreement with experimental values, whereas the predicted shift for C^Cs is not.
Although theory predicts that C4 and C5 in 2a will resonate downfield o f the
analogous carbons in la , consistent with the experimental result, the predicted value
overshoots this downfield shift by 17 ppm, a large difference. In previous studies we
13
have compared theoretical GLAO-MP2 C shifts with experimental results for many
4 5 20 22
carbenium, oxonium, and related cations, ’ ’ " and in no case did we observe a
difference this large. Because of this discrepancy, we also considered an alternative
assignment for the NMR signals— the heptamethylcyclohexenyl cation 3a. In this
alternative structure steric repulsion between the two pairs o f methyl groups is
negligible, and this is reflected in normal C4 -C5 and C5-C6 .single bond lengths.
The theoretical shift o f C4 and C$ of cation 3a is 50 ppm, the same as predicted for
analogous carbons in la. Thus, although the theoretical shift is in closer agreement
with the experimental value of 56 ppm, the theoretical methods do not predict the
181
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
downfield shift observed experimentally. In addition, whereas the value for C2 (151
ppm) is in reasonable agreement, the value for Ci and C3 (237 ppm) deviate
significantly from experiment. Indeed, the theoretical shifts for cation 3 a are in
excellent agreement with experimental values for cyclohexenyl cations, and we must
look elsewhere to explain the discrepancy between the experimental and theoretical
shift o f C4 , C5 of cation 2a. It is possible that the large, heavily substituted cation is
sterically hindered by the SAPO-34 cage, which alters its geometry from that which
we calculate free from the zeolite. Furthermore, the cation may interact closely with
the zeolite conjugate base site, which could also result in differences in chemical
shift. Although the theoretical results are less telling than we have come to expect,
they still support the assignment of the cyclopentenyl cation signals to species 2a.
Significance of the Heptamethylcyclopentenyl Cation 2a. This is the first
reported observation o f any persistent carbenium ion in a SAPO material. In
unpublished experiments we have observed very low concentrations of other
carbenium ions in HSAPO-34. Low intensity signals from the 1, 3-
dimethylcyclopentenyl cation were observed in some pulse-quench work with
propene, but these results were not compelling. Here, the 244-253, 155, and 56 ppm
signals are at least comparable to those in the first reports o f various cations in
aluminosilicate zeolites, and the claim of a persistent cation in HSAPO-34 can be
made with confidence. The cation observed here is also the first cyclopentenyl cation
with subtituents on C4 , C 5 characterized on any microporous solid acid.
182
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Figure 6.7 B3LYP/6-311G* optimizied geometry for olefins lb-3b. Selected symmetry-distinct C-C bond lengths (A)
are show in smaller print.
183
In previous work we have shown that the presence of persistent carbenium
ions in aluminosilicate zeolites requires that the parent aromatic or olefin be highly
23
basic. Specifically, we found that only those aromatics or olefins with proton
affinities (enthalpies) greater that ~209 kcal/mol would form persistent carbenium
ions on aluminosilicate zeolites. This concept is illustrated in Scheme 1 which
depicts protonation of the exocyclic diene 2b to form carbenium ion 2a. Note that
2b is the only olefin that affords cation 2a in an elementary protonation step without
oligomerization or skeletal isomerization, and no olefin can be formulated with two
endocyclic double bounds as a result of the substitution pattern on C4 and C5. Figure
6.7 reports the theoretical geometries o f the neutral dienes lb , 2 b, and 3b that give
rise to the corresponding cations by protonation. Olefin 3b also has an exocyclic
double bond, conforming to the same restrictions as 2b. All three structures show
C-C bond lengths consistent with expected bond orders and conjugation.
2b
Schem e 1
As mentioned above persistent carbenium ions are expected to form on
aluminosilicate zeolites if the proton affinities o f the parent olefins are greater than
209 kcal/mol. At the B3LYP/6-311G* level of theory, the proton affinities of lb , 2b,
184
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and 3b are 222.4, 227.6, and 226.6 kcal/mol, well above our previously reported
threshold. However, while we know that DFT usually gives a reasonable prediction
o f the geometries of the cations and their parent olefins, DFT does not give a good
23
prediction of the proton affinities. In previous work we found that we need higher
order correlation, specifically MP4(sdtq)/6-311+G*, in order to obtain reasonable
values. At the MP4(sdtq)/6-311+G*//B3LYP/TZVP level of theory the proton
affinity of lb is 215.6 kcal/mol, about 7 kcal/mol lower than the B3LYP/6-311G*
value. While we did not attempt the much more computationally expensive MP4
proton affinity calculations for 2b and 3b; we expect the difference between the two
levels of theory to be similar. Thus, even after adjustment, all three olefins are
sufficiently basic to form persistent carbenium ions in aluminosilicate zeolites.
However, when we attempted to form cation la in HSAPO-34 using methods
that produce robust NMR signals in zeolite HZSM-5, we saw either no signals at all
or weak signals of slight regard. Given that acetone reports that most acid sites in
HSAPO-34 are weaker than those on HZSM-5, it is likely that the proton affinity
threshold for carbenium ion persistence is higher in HSAPO-34. Olefin 2b is over 5
kcal/mol more basic than olefin lb which likely explanis the ready persistence of 2a
but not la on HSAPO-34.
Reaction M echanism for Conversion of Acetone to Aromatics on
g
HSAPO-34. Earlier CAVERN studies of acetone on aluminosilicate zeolites
elucidated several reaction pathways for acetone on those materials. Acetone was
185
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
shown to cyclotrimerize to the unsaturated ketone isophorone, which at higher
tem peratures would loose w ater and hydrogen to form trimethylbenzenes.
Alternatively, dimeric mesityl oxide would crack in the presence of water to make
acetic acid and isobutylene (which would in turn react to make other products).
Here, on HSAPO-34, we saw neither the cyclic trimer isophorone nor the
linear trimer phorone. We also did not observe any CO2 , which commonly formed
on the zeolites and was attributed to the cracking of phorone. Instead, we saw the
formation of much more acetic acid than on any zeolite studied other than HZSM-5.
These observations show that the aldol condensation o f acetone on HSAPO-34
proceeds no further than mesityl oxide, all of which cracks with water to form acetic
acid (which reacts no further) and isobutylene (which forms all o f the hydrocarbon
products).
In the CAVERN experiments we saw three major hydrocarbon products, the
heptamethylcyclopentyl cation 2a, alkanes, and heavily methylated benzenes. We
believe that the very high selectivity for cation 2 a reflects the fact that this C 12
species is made in the SAPO cages by a fairly clean trimerization o f C4 olefins, the
latter produced by cracking o f mesityl oxide. This contrasts with most previous
routes to cylopentenyl cations in zeolites. For example, in reference 4, the 1, 3-
dimethylcyclopentenyl cation (a C7 species) formed from ethylene through a
complex and ill-characterized route, but here a C 12 species apparently forms through
a more-direct trimerization of C4 olefins.
186
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Methyl migrations and skeletal isomerizations are facile for hydrocarbons on
solid acids at high temperature, and the identification of a unique mechanism for the
formation of 2a is not practical, but Scheme 2 is offered as a representative pathway.
In HSAPO-34, acetone is converted to acetic acid and isobutylene. The latter is not
capable o f leaving (or entering) the cages in HSAPO-34. We verified this by
flowing pure isobutylene over a bed o f HSAPO-34 at 723 K. There was no
conversion whatsoever, and even after prolong exposure the catalyst contained no
detectable carbon. However, isobutylene can isomerize in the cages to linear butenes
which, as Figure 5 demonstrates, exit the catalyst in appreciable amounts when
acetone is flowed over fresh catalyst. Scheme 2 suggests trimerization of two
isbutylenes and one molecule of 2 -butene followed by a straightforward methyl
migration to set up the correct carbon skeleton for ring closure. The loss of two
equivalents of H2 in the passage from C4 olefins to cation 1 requires that two other
olefin molecules be reduced to alkanes by proton transfer and hydride abstraction
~ c h 3
Scheme 2
187
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
steps. Indeed, much of the upfield signal intensity in Figures 6.1 through 6.3 is
accounted for by alkanes (for a well-resolved example, note the peak at 9 ppm,
which is characteristic of one carbon type in a number of highly branched alkanes).
Conversion o f cation 2a (C]2H2 i+) to hexamethylbenzene (C ^H is) requires
loss of a final equivalent of H 2 (by reduction of another olefin molecule) and transfer
23
of a proton back to the acid site. As pointed out previously, the proton affinity of
hexamethylbenzene (205.7 kcal/mol) is below the threshold for formation of a
persistent carbenium ion on aluminosilicate zeolites. Reference 4 detailed the steps
and theoretical energetics for the conversion o f the 1 , 3-dimethyIcyclopenteny 1
cation to toluene. A key step in that pathway was hydride abstraction from a ring
carbon o f C7H 1 0 to form a tertiary carbenium ion. We note that the
heptamethylcyclopentenyl cation has no hydrogens on ring carbons, and an
analogous hydride abstraction from olefin 2b would necessarily occur on a methyl
group to yield (formally) a primary carbenium ion. The energetic difficulty of this
hydride abstraction step may contribute to the high yield o f cation 2a in Figures 1
through 3.
While the present study is further evidence for the intermediacy of
cyclopentenyl cations in the formation of methylaromatics, it may be that cation 2 a
will more readily loose hydrogen after transferring one methyl group to either
another organic species or to the catalyst framework. Thus, we might expect to see
188
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the formation of benzenes with fewer than six methyl groups, and an average number
o f 4 was indeed estimated from Figures 1 and 2.
Relevance of This Study to Metfaanol-to-Olefm Catalysis. We recently
proposed that methylbenzenes are part of the active site for olefin synthesis on
HSAPO-34. These form during the kinetic induction period, and undergo side-
chain methylation to yield ethylbenzene and cumene derivatives. Alternatively,
highly methylated benzenes can rearrange to extend alkyl chains. Either way these
species generate the primary olefmic products. Here we have shed some light on the
formation of methylbenzenes and demonstrated that cyclopentenyl cations can
indeed form on HSAPO-34 under conditions that also form methylbenzenes.
While acetone does react to form methylbenzenes on HSAPO-34, and these
apparently eliminate ethylene and propene (Figure 6.5b), acetone cannot re-
methylate the aromatics to form a complete catalytic cycle. Instead, isobutylene
from cracking of mesityl oxide forms additional aromatics until all pores are full and
the catalyst is deactivated. This occurs rapidly. Note that as the catalyst becomes
almost completely deactivated, the ethylene selectivity increases (Figure 6.5c), by
itself a desirable property. We speculate that the last olefmic products formed on the
deactivating catalyst are eliminated from benzene rings that have already lost most of
their substituents, and rings capable of eliminating ethylene outnumber those that can
eliminate propene.
189
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Acetone conversion also differs from methanol conversion in the high initial
selectivity for C4 olefins obtained with the former. In Figure 6.5b, the total C4
selectivity exceeds that of C2 and C4 , while it does not using methanol as a feed (c/,
Figure 6.5a). Yet, the C5 selectivity with methanol exceeds that with acetone. This
is a reflection o f the primacy o f isobutylene as a hydrocarbon product of acetone and
its isomerization to linear products, which escape the HSAPO-34 cages.
190
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6.5 Refewrences and notes
(1) Stocker, M. Microporous Mesoporous Mater. 1999, 29, 3-48.
(2) Wilson , S; Barger, P Microporous Mesoporous Mater. 1999, 29, 117-126.
(3) Song, W,; Haw, J. F.; Nicholas, J. B.; Heneghan, K. J. Am. Chem. Soc. 2000,
122, Web Release Date October 13, 2000.
(4) Haw, J. F.; Nicholas, I. B.; Song, W. G.; Deng, F.; Wang, Z. K.; Xu, T.;
Heneghan, C. S. J. Am. Chem. Soc. 2000,122, 4763-4775.
(5) Xu, T.; Barich, D. H.; Goguen, P . W.; Song, W.; Wang, Z.; Nicholas, J.
B.; Haw, J. F. J. Am. Chem. Soc. 1998,120, 4025-4026.
(6 ) Haw, J. F.; Richardson, B. R.; Oshiro, I. S.; Lazo, N. L.; Speed, J. A. J. Am.
Chem. Soc. 1989, 111, 2052-2058.
(7) Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 1998,120, 2651-2652
(8 ) Xu, T.; Munson, E. J.; Haw, J. F. J. Am. Chem. Soc. 1994,116, 1962-1972.
(9) Gauss, J. Chem. Phys. Lett. 1992,191, 614-620.
(10) Lok, B. M.; Messina, C. A.; Patton, R. L.; Gajek, R. T.; Cannan, T. R.;
Flanigen, E. M. U.S. Patent 4, 440, 871,1984.
(11) Munson, E. J.; Murray, D. K.; Haw, J. F. J. Catal. 1993,141, 733-736.
(12) Xu, T.; Haw, J. F. Topics in Catalysis 1997, 4, 109-118.
(13) Haw, J. F.; Goguen, P. W.; Xu, T.; Skloss, T. W.; Song, W.; Wang, Z.
Angew. Chem. 1998, 37, 948-949.
(14) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652.
(15) Schafer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571-2577.
(16) Frisch, M. J.; Trucks, G. W.; Schlegal, H. B.; Gill, P. M. W.; Johnson, B. G.;
Robb, M. A.; Cheeseman, J. R.; T. Keith; Petersson, G. A.; Montgomery, J.
A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J. V.;
Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.; Nanayakkara, A.;
Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.;
191
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J.;
Binkley, J. S.; Defrees, D, J.; Baker, J.; Stewart, J. P.; Head-Gordon, M.;
Gonzalez, C.; Pople, J. A. Gaussian 94, Revision B.2.\ Gaussian, Inc.:
Pittsburgh, PA, 1995.
(17) Barich, D. H.; Nicholas, J. B.; Xu, T.; Haw, J. F. J. Am. Chem. Soc. 1998,
120, 12342-12350.
(18) Weight hourly space velocity (WHSV) formally has units of g reactant (g
catalyst hr) .
(19) Haw, J. F.; Nicholas, J. B.; Xu, T.; Beck, L. W.; Ferguson, D. B. Acc. Chem.
Res. 1996, 29, 259-267.
(20) Nicholas, J. B.; Xu, T.; Barich, D. H.; Torres, P. D.; Haw, J. F. J. Am. Chem.
Soc. 1996,118, 4202-4203.
(21) Xu, T.; Barich, D. H.; Torres, P. D.; Haw, J. F. J. Am. Chem. Soc. 1997,119,
406-414.
(22) Xu, T.; Barich, D. H.; Torres, P. D.; Nicholas, J. B.; Haw, J. F. J. Am. Chem.
Soc. 1997,119, 396-405.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 7
Methylbenzenes are the Organic Reaction Centers
for Methanol to Olefin Catalysis on HSAPO-34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The catalytic conversion o f methanol to olefins1 is a major emerging
chemical technology and the key step in the conversion of natural gas to polyolefins
using otherwise mature chemistry. The most promising catalysts for commercial
trials are solid acids based on silico-aluminophosphates of the chabazite (CHA)
2
structure, most simply HSAPO-34. The acid function in this material 1 is similar to
that in aluminosilicate zeolites, but it is imbedded in an aluminophosphate lattice.
The CHA topology features cages of ca. 1.0 nm by 0.67 nm diameter that are
interconnected through windows o f ca. 0.38 nm diameter. Thus, reactants such as
methanol and dimethylether and products such as ethylene and propene may freely
diffuse through active catalysts, but products with larger kinetic diameters, even
isobutylene, are trapped within the cages.
H
O
o o
1 0 2 °
13 3
Figure 7.1 reports C MAS NMR spectra of HSAPO-34 samples that were exposed
13
to identical pulses of methanol- C for between 2 s and 7200 s in a flowing He
194
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
5 6
7 2 0 0
3 6 0
120
6 0 s
3 0 s
1 6 s
130
134
0
150 50 150 10 50 100
p p m w p p m
13
Figure 7.1. 75 MHz C CP/MAS NMR spectra o f samples from a pulse-quench
study of methanol conversion on HSAPO-34 at 673 K. Each sample was prepared
13
by injecting 20 pL o f methanol- C onto a freshly activated catalyst bed (0.3 g)
while He was flowed at 600 mL min , and reaction occurred for the times shown
followed by a rapid thermal quench. All spectra (4000 scans) were measured at 298
K using a 2-ms contact time.
195
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
stream at 673 K prior to a rapid thermal quench. In this experiment, the product
gases were also analyzed by GC. At the shortest reaction times, the conversion of
13
methanol and dimethylether was low, and the C NMR spectra reveal both adsorbed
methanol (50 ppm) and a shoulder at 56 ppm that we assign to the chemisorbed
methoxy (methoxonium) species 2 in agreement with previous NMR studies of
catalysts, 5 and in particular, to account for the time-course of this resonance in
Figure 7.1.
After 4 s of reaction there is a dramatic reduction in the amount of methanol
on the catalyst and methyl-substituted aromatics form as signaled by the aromatic
resonance at 129 ppm (with a shoulder at 134 ppm due to substituted ring carbons)
and methyl groups at 20 ppm. Other upfield signals are accounted for by alkane
products trapped in the cages. The average number o f methyl groups per ring reaches
a maximum o f ca. 4 between 30 and 120 s of reaction, but this decreases to ca. 1.4
after the catalyst ages for 7200 s at 673 K without injection of additional feed. GC
analysis confirms that traces of olefin products exit the catalyst bed even 7200 s after
methanol injection. The 56 ppm resonance persists in catalyst samples for at least
several minutes, and a small amount is present after even two hours at 673 K. This
evidence of strong adsorption supports the assignment o f this resonance to the
framework methoxonium species 2 .
Figure 2 reports GC traces that compare the activity of a completely fresh
catalyst exposed to single pulse of methanol with that of a second catalyst that was
196
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ilJ L l
L . 1
MeOH
DME
J L
c 2 h 4
c 3 h 6
c 3 h 8
c4
c5 -c6
A m J L .
liaAiiiii 1
x 2 0
n l i | i f t t ] i i i i I n
20 25
Retention Time (min.)
30
Figure 7.2 GC (flame ionization) analyses of the gases exiting the HSAPO-
34 catalyst bed operated as in the previous figure, except as a "double
pulse" experiment. Identical, 20 pL methanol pulses were applied at 0 s
and at 360 s: (a) 4 s after a first pulse the total conversion of methanol and
dimethylether (DME) to hydrocarbons was only ca. 14 %. (b) 364 s after
the first pulse and 4 s after the second pulse, the conversion was essentially
100 %. (c) This control experiment shows that only traces of products exit
the reactor 358 s after the first methanol pulse; hence, the products
observed at 364 s reflect conversion o f the second methanol pulse.
197
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
first treated with a methanol pulse 360 s prior to a second, identical pulse. With fresh
catalyst the conversion was only 14 %, but the catalyst pretreated to form
methylbenzenes achieved nearly 100% conversion of the second methanol pulse. A
control experiment in Figure 2 shows that the products treated with a methanol pulse
360 s prior to a second, identical pulse. With fresh catalyst the conversion was only
14 %, but the catalyst pretreated to form methylbenzenes achieved nearly 100%
conversion of the second methanol pulse. A control experiment in Figure 7.2 shows
that the products from the second pulse vastly overwhelm those from the first pulse
after 360 s of reaction. Experiments such as those in Figures 7.1 and 7.2 provided
very strong evidence that in order for HSAPO-34 to be active for MTO chemistry, it
must have methylbenzenes trapped in some of its cages.
13
We performed a number o f experiments using methanol- C in a first pulse
to synthesize methylbenzenes followed by a smaller pulse o f natural abundance
methanol. In a typical result, we found that the ethylene sampled 3 s after the second
(natural abundance) m ethanol pulse had the following carbon isotopom er
distribution: 43 % 1 3Q), 47 % 1 3 C i, and 10 % 1 3 C 2 - The formation o f ethylene-
13
Ci is consistent with side-chain alkylation followed by elimination. Competing
reactions including ring m ethylation and de-methylation and exchange o f
methoxonium species 2 account for the formation of other isotopomers. Propene
was also synthesized using carbon from both the labeled methylbenzenes and
unlabeled methanol.
198
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figures 7.1 and 7.2 show that MTO chemistry on HSAPO-34 exhibits a
kinetic induction period and that the ship-in-a-bottle synthesis o f methylbenzenes in
the cages produces a working catalyst. These aromatic rings are trapped by the
catalyst’s topology, but the methyl groups are free to exchange or react with
methanol. In recent pulse-quench NMR and theoretical studies of hydrocarbon
synthesis on the aluminosilicate zeolite HZSM-5, we found that methyl-substituted
cyclopentenyl carbenium ions, in equilibrium with neutral cyclic dienes, served as
organic reaction centers for MTO chemistry.6 Theoretical calculations outlined a
catalytic cycle that led to elimination of ethyl and propyl groups to form the primary
olefmic products.6 HSAPO-34 is a weaker acid than HZSM-5 and it is used at a
higher temperature. Also, the topology of HZSM-5 permits methylbenzenes, which
are synthesized through cyclopentenyl cation intermediates, 6 to readily desorb from
the catalyst and enter the product stream. Indirect evidence of the “co-catalytic” role
7
o f toluene in methanol conversion on HZSM-5 includes the work of Mole. Dahl
and Kolboe have previously suggested that methanol conversion on solid acids
g
passes through a phenomenological “carbon-pool” mechanism.
We also discovered that the deactivation o f HSAPO-34 is the result of the
conversion o f m ethylbenzenes to polycyclic aromatics. For example, with
continuous injection o f methanol at a WHSV o f 33 h , catalyst deactivation
occurred after ca. 40 min., at which time the catalyst was 16.2 % C by weight. This
199
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
shows that as the catalyst deactivates, there is not only an increase in carbon content,
but also a structural change in the organic component to polycyclic aromatics
typified by methylnapthalene. Indeed, the observed carbon content o f the
deactivated catalyst is consistent with, on average, a C jo species in every cage.
In conclusion, the active site for olefin synthesis on HSAPO-34 is a
composite of a well-defined organic species and one or more inorganic acid sites,
which can activate methanol and hold methyl cation equivalents. All of this takes
place in a nanocage that preserves the organic component and regulates selectivity
through steric constraints. The HSAPO-34 active site then is no less elegant than
those praised in enzyme chemistry, and if it operates at far higher temperatures it
also self-assembles with little to work with. The detailed mechanism by which this
site operates is under intensive investigation.
2 0 0
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7.2 References and Notes
(1) Stocker, M. Microporous Mesoporous Mater. 1999, 29, 3-48.
(2) Wilson , S; Barger, P Microporous Mesoporous Mater. 1999, 29, 117-126.
(3) HSAPO-34 was prepared according to: Lok, B. M.; Messina, C. A.; Patton,
R. L.; Gajek, R. T.; Cannan, T. R.; Flanigen, E. M. U.S. Patent 4, 440, 871,
1984. XRD showed a pure crystalline phase with the CHA structure. The
product was calcined at 873 K for 10 hours to remove the template agent and
pressed into 10-20 mesh pellets. The Bronsted site concentration was
determined to be 1.1 mmol/g. In typical experiments 0.3 g of catalyst was
activated at 673 K under 600 seem He flow for 2 h immediately prior to use
in a pulse-quench reactor.
(4) For discussion of the pulse-quench in situ NMR experiment see: (a) Haw, J.
F.; Goguen, P. W.; Xu, T.; Skloss, T. W.; Song, W.; Wang, Z. Angew. Chem.
1998, 37, 948-949. (b) Goguen, P. W.; Xu, T.; Barich, D. H.; Skloss, T. W.;
Song, W.; Wang, Z.; Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 1998,120,
2651-2652.
(c) (30) Xu, T.; Barich, D. H.; Goguen, P. W.; Song, W.; Wang, Z.;
Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 1998,120,4025-4026.
(5) Anderson, M. W.; Sulikowski, B.; Barrie, P. J.; Klinowski. J. Phys. Chem.
1990, 94, 2730-2734.
(6) Haw, J. F.; Nicholas, J. B.; Song, W.; Deng, F.; Wang, Z.; Heneghan, C. S.
J. Am. Chem. Soc. 2000,122, 4763-4775.
(7) (a) Mole, T.; Whiteside, J. A.; Seddon, D. J. Catal. 1983, 82, 261-266. (b)
Mole, T.; Bett, G.; Seddon, D. J. Catal. 1983, 8 4 ,435-445.
(8 ) Dahl, I. M.; Kolboe, S. J. Catal. 1996,161, 304-309. (b) Dahl, I. M.;
Kolboe,S. J. Catal. 1994,149, 458-464.
201
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 8
Supramolecular Origins of Product Selectivity for
Methanol-to-Olefin Catalysis on HSAPO-34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8.1 Introduction
Control of selectivity is a central problem in catalysis, and this is particularly
1-3
challenging with the microporous solid acid catalysts used in most industrial
catalytic processes. We recently demonstrated that for some reactions on solid acids,
4 5
the active site is a hybrid of inorganic and cyclic organic components. ’ The latter
function as scaffolds for bond making and breaking steps that would otherwise be
energetically unfavorable. In the case of methanol-to-olefm (MTO) conversion6 on
7
the preferred catalyst HSAPO-34 (a silico-aluminophosphate of CHA topology),
g
the organic components are methylbenzenes that self-assemble in the nanometer-
size cages during a kinetic induction period, where they must remain until burned
out, as even benzene is too large to pass through the 0.38 nm windows connecting
adjacent cages. An outstanding problem in MTO catalysis is improving ethylene
selectivity at the expense o f propene to better match the demand for the
corresponding polyolefins.
Here we show that ethylene selectivity is governed by the number of methyl
substituents on the aromatic rings, and the presence or absence of other molecules in
the nanocages. At higher methanol space velocities (partial pressures), the average
number of methyl groups per ring reaches a steady-state value of greater than 5 at
400 °C, and the ethylene selectivity is only ca. 25 %. At far lower space velocities,
the average number of methyl groups per ring reaches a steady-state value below 2 ,
and the catalyst then yields over 60 % ethylene. Similarly, the average number of
203
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
methyl groups per ring reaches lower steady-state values at higher temperatures, and
this contributes to higher ethylene selectivities.
As a step toward a kinetic analysis of MTO chemistry in HSAPO-34, we
treat each nanocage containing a methylbenzene as an independent supramolecule
capable of unimolecular decomposition to either ethylene, {n}— > {n-2, e}, or
propene, {n}— »{n-3, p}, where n implies a methylbenzene molecule with n methyl
groups, and {} designates the nanocage and its associated Bransted acid site. We
performed a series of experiments where we allowed a working catalyst to reach
steady-state during continuous introduction of methanol before abruptly cutting off
reagent introduction. We then followed the time evolutions of the rates o f product
formation and the average number of methyl groups on benzenes trapped in the
catalyst’s nanocages. The correlation of these observables permitted us to compare
the relative magnitudes of some o f the unimolecular rate constants for the reactions
for supramolecules composed o f cages containing a methylbenzene molecule, but no
other adsorbate. A rate equation is derived to account for the time evolution of the
average number of methyl groups per benzene ring in such experiments.
Co-adsorption of a water molecule into a nanocage with a methylbenzene
produces unique supramolecules, {n, H2O}, and we find that rate constants for their
decomposition to olefmic products, for example, {n, H 2O}— >{n-2 , H 2O, e} are
slower than in the absence o f water, leading to higher steady-state averages for the
number of methyl groups per benzene rings. Furthermore, the presence of water in
204
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the nanocage significantly increases ethylene selectivity at the expense of propene,
and we attribute this to enhanced transition state shape selectivity.
8.2 Experimental Section
Materials and Reagents. HSAPO-34 was prepared according to a patent
procedure. XRD showed a pure crystalline phase with the CHA structure. The
product was calcined at 873 K for 10 hours to remove the template agent and
pressed into 10-20 mesh pellets. The Bronsted site concentration was determined to
13
be 1.1 mmol/g. Methanol- C was obtained from Isotech, Inc. In general, we used
13
methanol- C for experiments in which an NMR measurement was to be made (with
or without GC) and natural abundance methanol if GC was to be the only analytical
method.
Catalysis. Experiments were performed using the pulse quench reactor
described elsewhere10 with the exception that methanol was delivered using a
motor-driven syringe pump (Harvard Apparatus model PHD 2000). For each
experiment a bed consisting of 0.3 g of catalyst was activated at 673 K in the reactor
under 200 seem He flow for 2 h immediately prior to use. This carrier gas feed rate
was also used during methanol introduction in all experiments. For NMR sample
preparation, methanol flow was abruptly ceased a predetermined time (usually 0 to
60 min) prior to quench. Previous studies have shown that the temperature o f the
catalyst pellets decreases 150 K in the first 170 ms of a quench. After quenching
205
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
each reacted catalyst sample, the reactor was sealed off and transferred into a glove
box filled with nitrogen. The catalyst pellets were ground and transferred to a 7.5-
mm MAS rotor which was sealed with a Kel-F end-cap.
Gas C hrom atography. A Hewlett-Packard Model 6890 gas chromatograph
with flame-ionization detector was used to analyze gases sampled from the reactor
product streams using a Valeo valve. The column was 150 m dhl50 (Supelco)
operated isothermally at 323 K to permit sampling of the gas stream more frequently
than the total analysis time for any given sample.
NM R Spectroscopy. 1 3 C solid state NMR experiments were performed with
magic angle spinning (MAS) on a modified Chemagnetics CMX-300 MHz
spectrometer operating at 75.4 MHz for 13C. Hexamethylbenzene (17.4 ppm) was
used as an external chemical shift standard, and all 1 3 C chemical shifts are reported
relative to TMS. Chemagnetics-style pencil probes spun 7.5 mm zirconia rotors at
typically 6.5 kHz with active spin speed control (± 3 Hz).
Typical 1 3 C experiments included: cross polarization (CP, contact time = 2
ms, pulse delay = I s, 2 0 0 0 transients); cross polarization with interrupted
decoupling (contact time = 2 ms, pulse delay = 1 s, 2 0 0 0 transients, dipolar
dephasing time of 50 ps); single pulse excitation with proton decoupling (Bloch
decay, pulse delay - IQs, 400 transients). CP and Bloch decay spectra gave very
similar values for the average number o f methyl groups per aromatic ring. Similar
to other NMR work on chemically dilute carbonaceous materials, the CP spectra
here were generally “cleaner” than the Bloch decay spectra. All spectra reported
206
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
here were measured using CP, but the average methyl group numbers reported were
determined from the Bloch decay spectra since these are more commonly regarded
as quantitative. Our conclusions would not change if we reported integrations from
the CP spectra.
8.3 Results
Decomposition of Methylbenzenes Following Methanol Cut-Off. We
carried out a series of experiments in which we first flowed 0.1 mL of methanol or
methanol-13C onto a 300 mg bed o f pelletized HSAPO-34 at 400 °C at a weight
hourly space velocity (WHSV) o f 8 h" 1 (for a 300 mg catalyst bed this corresponds
to 50 pL/min) before abruptly terminating methanol flow and waiting a variable
time before quenching the catalyst temperature to ambient. Figure 8.1 reports solid
13
state C MAS NMR spectra o f catalysts prepared using delays from 0 to 60 min.
These spectra show an aromatic carbon signal between 129 and 134 ppm, and a
methyl group resonance at 2 0 ppm that drops with increasing delay between
methanol cut-off and thermal quench. A small, sharp resonance at 25 ppm is due to
isobutane, which like the methylbenzenes, is too large to exit the nanocages.
13
The C NMR spectra permit direct measurement of the average number of
methyl groups per benzene ring in the catalyst, Meave (Eq. 1).
207
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 8.1. 75 MHz 13C CP/MAS NMR spectra showing the loss of methyl groups
as a function o f time from methylbenzenes trapped in the HSAPO-34 nanocages at
400 °C. For each case, a fresh catalyst bed was used to convert 0.1 mL of methanol-
1 3 C at a WHSV of 8 h’1 , and then methanol flow was abruptly cut off. The catalyst
bed was maintained at temperature with He flow (200 seem) for the time indicated,
and then the reactor temperature was rapidly quenched to ambient. Entire catalyst
beds were loaded into MAS rotors to avoid sampling errors, and cross polarization
spectra were measured at room temperature. The average numbers of methyl groups
per ring Meave, were calculated from Bloch decay spectra very similar to the cross
polarization spectra shown.
208
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
129
134
5 . | M eave
60 min.
2.7
3.0
3 min.
3.7
1 m in.
4.8
0 m in.
5.6
50 100 250 200 150 -50
ppm
209
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
Meave = [methyls]/[rings] = n f n (1 )
n = 0
fn denotes the fractions of rings with n methyl groups. As shown in Figure 8.1,
Meave decreased from a nearly full complement o f 5.6 immediately after methanol
cut-off to 1.9 (e.g., xylenes on average) after 60 min. The last methyl group could
be removed only with great difficulty. At 450 °C, Meave was 1.1 after 60 min and
0.2 after 14 h.
Figure 8.2 reports gas chromatographic analyses of samples taken from the
product streams immediately prior to thermal quench for several of the experiments
used to prepare the samples for Figure 8.1. One minute after methanol cut-off, the
catalyst was still producing olefinic products with an ethylene selectivity only
slightly higher than that immediately prior to cut-off; in particular, one notes a much
higher selectivity for propene than for ethylene. As the catalyst continued to age at
400 °C with He flow but no further addition o f methanol, the total rate o f olefin
production necessarily fell, but the ethylene selectivity increased dramatically,
surpassing that of propene between 6 and 12 minutes after cut-off. We measured the
rates of formation o f ethylene, propene, and a representative C4 product from a
single experiment in which we made a number of gas chromatographic injections as
a function o f time, and these data are plotted in Figure 8.3a; ethylene dominated
after 8 min. C4 products may form in part from secondary reactions and are
210
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12 min.
X 2
6 min.
3 min.
A — A j JL a.
ethylene
propene
c5
1 min.
i — i — r — |— i — i — i — i — |— ! — i — i — i —
15 20 25
Retention time (min.)
Figure 8.2 Gas chromatography (flame-ionization detection) analyses of the
volatile products captured immediately prior to thermal quench from the
experiments used to prepare some o f the samples for Fig. 1. Note the significant
increase in ethylene selectivity at longer times.
211
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
hereafter neglected. Figure 8.3b compares the ethylene and propene selectivities,
computed from Figure 3a, as a function of the time evolution of the average number
o f methyl groups per ring— a mapping made possible by a smooth fit of the data in
Figure 8.1. The relationship between molecular structure and selectivity is clear, a
change-over in selectivity to ethylene occurred when the number of methyl groups
per ring decreased to 2 .8 , and further decreases in propene selectivity accompanied
further decreases in methyl substitution.
Effects of Space Velocity or Tem perature. We explored rational means of
limiting the average number of methyl groups per ring to see if these conditions did
13
indeed increase ethylene selectivity. Figure 8.4 reports C NM R spectra that
demonstrate that as the methanol space velocity (and hence partial pressure) was
reduced at 400 °C, the benzene rings reached steady state with fewer methyl groups
per ring. The figure shows that as the space velocity was decreased (at 400 °C) from
8 h *1 to 0.0008 h’1 , Meave decreased progressively from 5.6 to 1 .2 .
We varied the methanol space velocity by four orders o f magnitude (while
keeping the He flow constant) and measured the steady-state olefin selectivities (all
at 400 °C and 100 % conversion) reported in Figure 8.5a. The ethylene selectivity,
which was only 25 % at the highest space velocities, increased dramatically as the
space velocity was reduced, and we measured an extraordinary steady-state
selectivity o f 63 % at the lowest space velocity studied. Using the Meave
measurements in Figure 8.4, Figure 8.5b maps the selectivity data in Figure 8.5a
212
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
G
©
©
fa
©
©
©
©
t £ 3
©
"©
16
14
12
10
8
6
4
2
0
10 15 20 25 5
Delay Time (min.)
Figure 8.3a. The rates of formation o f ethylene, propene, and 2-butene
as a function o f time from a single experiment similar to those used for
Figs. 8.1 and 8.2. For these measurements, multiple gas samples were
analyzed from a single catalyst bed as it evolved over time after cessation
of methanol flow. Ethylene and propene are primary products ofM TO
chemistry. It can be shown that C4 products are due to secondary
reactions o f propylene.
213
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Selectivity (%)
50 i
48
46 - I
44
42 - I
40
38
36
34
32
30
2.5
—♦ — ethylene
— i r -propene
3.5 4.5
Me
ave
Figure 8.3b. Ethylene and propene selectivity as a function of the
average number o f methyl groups per ring, Meave. The data in b are re
plotted using an abscissa derived by empirically fitting the time
evolution of Meave in Fig. 8.1 to a smooth curve generated by Excel.
This mapping shows that ethylene is favored by methylbenzenes with 2
or 3 methyl groups, while propene is favored with four or more methyl
groups.
214
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.2
1.7
2.1
WHSV=0.8 h"1
3.6
130
134
WHSV=8 fa " 1
5.6
T T nT l'l' f I F T 1 S II I I T T Vi i i i I i i j | g i i | i
300 250 200 150 100 50 0 -50
ppm
13
Figure 8.4. 75 MHz C CP/MAS NMR spectra from experiments probing the
steady-state average number of methyl groups per ring at various methanol space
velocities. In each case we converted methanol to reach steady-state selectivities,
and then abruptly ceased methanol flow and thermally quenched the catalyst bed
without further delay. Lower space velocities reduce the average number of methyl
groups.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
65
60
55
o
ju
%
m
40
35
30
25
20
♦ — ethylene
A f -propene
■
m
m
m
m
-3.5 -2.5 -1.5 -0.5 0.5 1.5
log (WHSV)
Figure 8.5a. Ethylene selectivity at 400 °C vs. the logarithm of
methanol weight-hourly space velocity. The catalyst bed, 300 mg, was
first treated with a total o f 0.1 mL o f methanol at WHSV of 8 h ' 1 at 400
°C to create methylbenzenes, and then was held at a given space velocity
until selectivity reached a steady-state value. Note the very high
ethylene selectivities at very low space velocities. ^ ,
216
R e p r o d u c e d with p e r m is s io n o f th e c o p y rig h t o w n e r. F u r th e r r e p ro d u c tio n p ro h ib ite d w ith o u t p e r m is s io n .
Selectivity (%)
■ •— ethylene
it -propene
45-
3 4 5 6 1 2
Figure 8.5 (b).The ethylene selectivity data in a plotted against Meave.
217
R e p r o d u c e d with p e r m is s io n o f t h e c o p y rig h t o w n e r. F u r th e r r e p ro d u c tio n p ro h ib ite d w ith o u t p e r m is s io n .
onto an Meave coordinate. Under steady-sate conditions, in a catalyst that also
contains some methanol, dimethylether, and water in the nanocages, the change-over
in selectivity for dominance of ethylene or propene occurred at Meave - 2 . 1 in
Figure 8.5b compared to 2.8 without methanol, dimethylether, and water, Figure
8.3b.
We also found that for a fixed space velocity o f 8 h"1 , Meave decreased with
increasing temperature, and again this correlated with higher ethylene selectivity.
Figure 8 . 6 shows that at 450 °C Meave was 4.3 (cf. 5.6 at 400 °C ) and further
dropped to 1.8 at 550 °C. The ethylene selectivities were 25 % at 400 °C, 34 % at
450 °C, and 50 % at 550 °C.
Effect of Added W ater. As the first step toward studying the
decomposition of methylbenzenes in HSAPO-34 cages in the presence o f other
molecules, we studied the effect of added water at 400 °C. Water is a co-product of
MTO chemistry that must be present in at least some nanocages during methanol
conversion, especially at higher space velocities. Figure 8.7 a and b show that if the
methanol space velocity is kept fixed but water is also introduced at a large space
velocity, Meave will reach a larger steady-state value, suggesting that olefin
elimination is retarded by the presence of water in the cages. GC analysis o f the
volatile products revealed an ethylene selectivity o f 35 % with pure methanol
delivered at 5 pL/min, but this increased to 42 % when 10 % (v/v) methanol in
218
R e p r o d u c e d with p e r m is s io n o f th e c o p y rig h t o w n e r. F u r th e r r e p ro d u c tio n p ro h ib ite d w ith o u t p e r m is s io n .
600 °C
Me^ve
0.9
550 °C
500 °C
A A j:
4.3
450 °C
130
134
5.6
250 150 100 50 -50 300 200 0
1 3
Figure 8 .6 . 75 MHz C CP/MAS NMR spectra probing the steady-state average
number of methyl groups per ring at fixed methanol space velocity ( 8 h"1 ) at various
temperatures. In each case, the thermal quench occurred immediately upon cessation
of methanol flow. The average number of methyl groups decreased with increasing
temperature.
ppm
219
R e p r o d u c e d with p e r m is s io n o f th e c o p y rig h t o w n e r. F u r th e r r e p ro d u c tio n p ro h ib ite d w ith o u t p e r m is s io n .
Figure 8.7. 75 MHz 13C CP/MAS NMR spectra from experiments probing the
effect of co-feeding water on the average number of methyl groups per ring at a 400
°C reactor temperature. In every case 100 |iL of pure methanol was first delivered
at WHSV = 8 h ' 1 (50 \iL/ min) to form methylbenzenes, and then conditions were
varied: (a) Control experiment, pure methanol at 5 p,L/ min to reach steady-state
and then thermal quench immediately upon cessation of methanol flow; (b) 1 0 %
(v/v) methanol in water at 50 jxL/ min to reach steady-state and then thermal quench
immediately upon cessation of flow; (c) Control experiment, wait 60 min with no
flow (other than He gas) after forming methylbenzenes, then thermal quench, (d)
W ait 60 min w hile flowing 10 pL/min. w ater with He gas after forming
methylbenzenes, then thermal quench. The effect of added water is to decrease the
rate o f methylbenzene decomposition to olefins, hence an increase in Meave relative
to the control experiments. Water also increases ethylene selectivity (see text).
2 2 0
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
130
134
2.5
3.6
200 150 100 300 250 -50
ppm
221
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
water was delivered at 50 pL/min. We also investigated the time evolution of Meave
following methanol cut-off in the presence o f water. Figure 8.7c shows that 60 m in
after cuff-off Meave drops to 1.9 without water, but with 1 0 pL/m in of water
following methanol cut-off (Figure 8.7d) Meave drops to only 2.5.
8.5 Discussion
The overall kinetics o f olefin synthesis on HSAPO-34 are no doubt
complicated, and will vary with co-adsorption of methanol, water, or other species
into the nano-cages containing the methylaromatics. It is convenient to consider
each nanocage with a methylbenzene molecule to be a supramolecular entity, each
including and delimited by the cage boundaries and possessing one acid site, but
distinguished by the number of methyl groups on the benzene ring and the presence
or absence o f other adsorbed molecules. We will assume that equilibration of
isomers is rapid compared to olefin elimination, and that we can treat, for example, a
cage containing only xylene and an acid site as a single supramolecule, and not
distinguish the three isomeric forms.
Decomposition of Methylbenzenes in HSAPO-34. The easiest case to
discuss is the cut-off experiments o f Figures 1 to 3, where there are no water or
methanol molecules in the cages (at least at longer times). We can treat the
decomposition o f methylbenzenes to ethylene or propene and less-substituted
methylbenzenes as unimolecular dissociation steps. Plots o f Meave versus time from
experiments like Figure 1 follow no simple integrated rate law, because of
222
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
contributions from rings with different numbers o f methyl groups and the time
evolution of the fractions of rings, fn, with n methyl groups. Consider the possible
reactions o f a trimethylbenzene molecule isolated in a nanocage. Trimethylbenzene
can either loose ethylene and leave toluene with a rate constant k({3}— >{1, e}) = kj6
or it can instead lose propene and leave benzene with a rate constant k({3}— >{0, p})
= k3P. If all methylbenzenes function independently, and diffusion of olefin
products out o f the cages is rapid, then the rate of change o f the average number of
methyl groups per ring in an experiment like Figure 1 is described by Eq. 2,
( 2 )
d t n = 0 d t B = 0
where koe = koP = k ie = kiP = k 2P = 0 , and fj = 0 when i > 6 .
Figures 1 through 3 allow us to infer the relative values o f some of the rate
constants. While xylene cannot eliminate propene, k 2 6 is not necessarily zero.
Figure 3 shows that the change-over in selectivity at 400 °C occurs with three
methyl groups per ring, implying that k 3C > k3P. We conclude that retention o f at
least one methyl group on the ring stabilizes transition states leading to olefin
elimination. Also, k6P is greater than kg6, and propene is favored at higher space
velocities and lower temperatures. Unfortunately, we also have the following
inequalities, k^p > k3P and k 6S > k3C , and the conditions for the most rapid methanol
223
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
conversion on this catalyst (large Meave) are not those that provide the highest
ethylene selectivity (small Meave).
Tem perature Effect. At higher temperatures the catalyst equilibrated to
lower values of Meave and produced higher ethylene selectivities. O f course, this
correlation does not rule the possibility of other contributions to the temperature
dependence of olefin selectivity. In a recent theoretical analysis of the elimination
of ethylene and propene from another cyclic molecule, we found a higher barrier for
4
the former relative to the latter. If (as we believe) ethylbenzenes and propyl or
isopropylbenzenes are precursors to olefins on HSAPO-34, elem entary
consideration of the stabilities of reactive intermediates would predict a higher
barrier to ethylene than to propene. A higher barrier to ethylene would also
contribute to improved ethylene selectivity at higher temperatures, as we observed.
Thus, we suspect that both the decreased Meave, and the greater rate of barrier
crossing at higher temperatures contribute the increase in ethylene selectivity with
temperature.
Effect of W ater. Under realistic operating conditions o f appreciable
methanol space velocity, some of the cages will also contain methanol, dimethyl
ether, and/or water molecules. A supramolecule consisting of a nanocage, an acid
site, trimethylbenzene, and a single water molecule could still eliminate an olefin,
but we would not expect, for example, k({3, H 2 O}— >{1, H 2 O, e}) to be identical to
224
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
k ({3}— >{1, e}) = k3C , since the adsorbate would impose steric constraints, especially
on the transition state, and solvate every point on the potential energy surface.
(Note that a cage with trimethylbenzene and two water molecules would also be a
distinct supramolecule with unique rate constants for decomposition to ethylene or
propene). Another possible reaction for a nanocage containing a methylbenzene and
water is demethylation to form methanol, but this can be ruled out based on the
endothermicity of such reactions in the gas phase. For example, the reaction of
toluene and water to form benzene and methanol has a AH o f + 17.7 kcal/mol,
implying a negligible equilibrium constant. 11
The results in Figure 8.7 can most easily be interpreted in the context of
k({n, H 2 O}— »{n-2, H2 O, e} and k({n, H 2 O}— »{n-3, H 2 O, p}) relative to the
corresponding rate constants in the absence o f water. Figure 8.7 implies that the
rates o f both ethylene and propene production are lower with water in the
nanocages. Furthermore, the additional steric constraints provided by adsorbed
water provide the most likely explanation for the increased ethylene selectivity in
the water co-feed experiments.
In the more general case o f a catalyst operating at appreciable methanol
apace velocity we also have reactions such as {n, CH 3 OH}— »{n+l, H2 O}. The
analogous reaction with dimethylether would produce a methanol molecule.
Furthermore, methylation could occur on the sidechains to more directly give
precursors to ethylene or propene, but we do not consider the detailed mechanism of
225
R e p r o d u c e d with p e r m is s io n o f th e c o p y rig h t o w n e r. F u r th e r r e p ro d u c tio n p ro h ib ite d w ith o u t p e r m is s io n .
olefin formation in this contribution. The interplay between these various reactions,
equilibration of methanol, dimethyl ether, and water, and mass transport of reactants
and products through the catalyst are reflected in observables such as Meave and
ethylene selectivity.
8.4 Conclusions
This investigation has found the origins of product selectivity in methanol-
to-olefm chemistry on HSAPO-34. At 400 °C nanocages containing only
methylbenzenes preferentially yield ethylene when the aromatic ring has two or
three methyl groups, while four to six methyl groups is more likely to lead to
propene as a product. This insight directly led us to significant improvements in
ethylene selectivity by either reducing the methanol space velocity or increasing the
temperature—we showed that either change reduces the average number o f methyl
groups per ring at steady-state. We partially interpreted the kinetics of this process
by focussing on the unimolecular decomposition of methylbenzenes in the
nanocages viewed as the reaction o f one supramolecule forming another; for
example, {n}-»{n-2 , e}.
Co-adsorption o f water produces distinct supramolecular species, for
example {n, H 2 O}, and these form olefins more slowly than the corresponding
species without water, {n}. Co-feeding of water during methanol conversion results
in an increase in the average value o f n (Meave) in the catalyst when steady state is
reached between methanol introduction and olefin formation. The presence o f H 2 O
226
R e p r o d u c e d with p e r m is s io n o f th e c o p y rig h t o w n e r. F u r th e r r e p ro d u c tio n p ro h ib ite d w ith o u t p e r m is s io n .
in the nanocage shifts the selectivity in favor of ethylene— the product formed
through the smallest transition state.
We believe that the present study is a major step forward in the
understanding of an important problem in heterogeneous catalysis. Also, the level
o f integration o f volatile product analysis and NMR data on the solid catalyst
structure (Figures 8.3b and 8.5b) represents a m ethodological advance.
Supplementary analysis by GC has become commonplace in NMR studies of
catalysis, but these data are not typically combined to the extent or advantage
demonstrated here.
8.5 References and Notes
(1) VanSanten, R. A.; Kramer, G. J. Chemical Reviews 1995, 95, 637-660.
(2) Corma, A. Chemical Reviews 1995, 95, 559-614.
(3) Corma, A. Chemical Reviews 1997, 97, 2373-2419.
(4) Haw, J. F.; Nicholas, J. B.; Song, W.; Deng, F.; Wang, Z.; Heneghan, C. S.
J. Am. Chem. Soc. 2000,122,4763-4775.
(5) Song, W.; Haw, J. F; Nicholas, J. B.; Heneghan, K. J. Am. Chem. Soc.
2000,122, 10726-10727.
(6 ) Stocker, M. Microporous Mesoporous Mater. 1999, 29, 3-48.
(7) Wilson, S; Barger, P Microporous Mesoporous Mater. 1999, 29, 117-126.
(8 ) Toluene has also been implicated as a “co-catalyst” for methanol conversion
on aluminosilicate zeolites: (a) Mole, T.; Whiteside, J. A.; Seddon, D. J.
Catal. 1983, 82, 261-266; (b) Mole, T.; Bett, G.; Seddon, D. J. Catal. 1983,
84, 435-445. Other workers have proposed a phenomenological carbon-pool
mechanism for MTO chemistry on HSAPO-34 without explicitly identifying
227
R e p r o d u c e d with p e r m is s io n o f th e c o p y rig h t o w n e r. F u r th e r r e p ro d u c tio n p ro h ib ite d w ith o u t p e r m is s io n .
the organic species responsible for the catalytic effect: (c) Dahl, I. M.;
Kolboe, S. J. Catal. 1996,161, 304-309; (d) Dahl, I. M.; Kolboe, S. J. Catal.
1994, 149, 458-464.
(9) Lok, B. M.; Messina, C. A.; Patton, R. L.; Gajek, R. T.; Cannan, T. R.;
Flanigen, E. M. U.S. Patent 4, 440, 871, 1984.
(10) (a) Haw, J. F.; Goguen, P. W.; Xu, T.; Skloss, T. W.; Song, W.; Wang, Z.
Angew. Chem. 1998, 37, 948-949. (b) Goguen, P. W.; Xu, T.; Barich, D. H.;
Skloss, T. W.; Song, W.; Wang, Z.; Nicholas, J. B.; Haw, J. F. J. Am. Chem.
Soc. 1998, 120, 2651-2652. (c) Xu, T.; Barich, D. H.; Goguen, P. W.; Song,
W.; Wang, Z.; Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 1998, 120,
4025-4026.
(11) NIST Chemistry WebBook (http ://webbook.nist.gov/chemistry/).
R e p r o d u c e d with p e r m is s io n o f th e c o p y rig h t o w n e r. F u r th e r r e p ro d u c tio n p ro h ib ite d w ith o u t p e r m is s io n .
Chapter 9
Improved Methanol-to-Olefin Catalyst with Nanocages
Functionalized through Ship-in-a-Bottle Synthesis from
p h 3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The catalytic conversion of methanol to olefins1 is the remaining link in a
chain o f catalytic processes that convert natural gas to polyolefins. The most
prom ising catalysts for com m ercial use are solid acids based on silico-
2
aluminophosphates o f the chabazite (CHA) structure, specifically HSAPO-34,
which as we prepare it has one acid site per cage. We recently demonstrated that
the active site for MTO chemistry on HSAPO-34 is a composite o f an inorganic acid
site and a methylbenzene molecule. The latter self-assembles during a kinetic
induction period and can not pass through the 0.38 nm windows interconnecting the
1.0 nm by 0.67 nm cages. The organic component is a scaffold on which carbon-
carbon bonds can be made and broken without recourse to very high energy
intermediates and transition states.5 Here we build on this prior work to rationally
design HSAPO-34 catalysts with improved selectivity, a very desirable property.
Specifically, transition states leading to propene are necessarily bulkier than those
leading to ethylene. Whereas we now know that MTO chemistry occurs in a
nanocage already partially filled by a large organic molecule, the reaction should be
very sensitive to further steric constraints.
For our first attempt at functionalization we chose phosphorous chemistry,
because phosphate species have proven to be hydrothermally-stable modifiers of
aluminosilicate zeolites, 6 and also because we could use 31P solid-state NMR to
correlate structure with activity and selectivity. Few phosphorous reagents are
230
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
c h 3o h
adsorbed into HSAPO-34; P(CH3 )3 is sterically excluded, and neither PH(CH3 )2 nor
PH 2CH3 are commercially available. We thus chose to work with P H 3 , a toxic and
pyrophoric gas that nevertheless proved to be easily handled using lecture bottles
and standard bench-top reactors assembled in a fume hood. Scheme 1 outlines our
strategy for modification of
P(CH3)3, P(CH3)4+ — * ♦ Phosphate
o2
1'
OP(CH3)3
Scheme 1
HSAPO-34 through ship-in-a-bottle routes starting with PH 3 . In every case we held
ca. 300 mg of HSAPO-34 in a flow reactor at 250 °C and then introduced PH 3 gas
into the He flow stream while simultaneously delivering methanol using a syringe
7 31 8
pump. We realized each step in the scheme, and report representative P spectra
3 1
o f thermally quenched catalyst beds in Figure 9.1. In each spectrum the P
resonance at -28 ppm is due to phosphorous in the framework. The initial products
+ 9
are P(CH3 )3 (-4 ppm) and P(CH3 )4 (23 ppm). The relative yields of these species
can be controlled by varying the ratio of PH 3 to CH3OH. P(CH3 )3 is commonly
10
used as a spectroscopic probe of acidity. The observation that all of the P(CH 3 )3
231
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
in HSAPO-34 is protonated to HP(CH 3)3+ and none is complexed to Lewis sites
(Figure 9.1a), is by itself an important finding. We were able to synthesize one
P(CH3)4+ cation in every nanocage by flowing larger amounts o f PH 3 and methanol
31 11
(Figure 9.1c). In this case the P spectrum showed a perfect 5:1 ratio of
3 1
framework to cation signals. Figure 9 .Id shows a P spectrum o f SAPO-34
containing both P(CH3)4+ and methylbenzenes. Note that with aromatics present
there are two P(CH3)4+ resonances, one as before at 23 ppm, and one at 20 ppm.
Although we believe this 20-ppm peak is almost certainly from P(CH3)4+ in cages
12
with an aromatic ring, the exact cause of the upheld shift is unknown.
When P(CH3)4+ in SAPO-34 was heated to 500 °C, it decomposed to an
inorganic phosphate species with a chemical shift almost identical to framework
phosphorous. This species is visible in Figure 9.1e as a broad feature at the base of
the -28 ppm resonance. Evidence indicates that the presence of this species
improves ethylene selectivity.
1 3
Figure 9.2 reports GC traces from the product streams o f reactors
converting methanol under steady-state conditions (WHSV = 8 h’1 ) at 450 °C using
either standard HSAPO-34 or functionalized materials. For HSAPO-34 without PH 3
treatment, methanol conversion was 100% and ethylene selectivity was 37 %.
Conversely, with one P(CH3)4+ cation per cage and all acid sites consumed by
phosphonium formation, conversion was near zero. Calcination o f this sample in
232
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 9.1. 31P MAS NMR spectra showing functionalization o f the HSAPO-34
nanocages according to the Scheme. The chemical shift of phosphorous in the
HSAPO-34 framework is -26 ppm. In all cases PH 3 was flowed over the catalyst
bed along with methanol for 10 min. The PH 3 flow rate was 14 seem except for the
sample in c where 32 seem was used. Methanol/PH3 mole ratios are shown on the
figure where important, (a) With a limiting amount of methanol the major product
in the nanocages was P(CH 3)3 , which was protonated by the Bronsted acid sites (-4
ppm), (b) P(CH3)4+ (23 ppm) formed almost quantitatively in most nanocages when
methanol was in excess, (c) With forcing conditions, P(CH 3)4+ formed in every
cage, (d) P(CH3)4+ yields signals at both 23 ppm and 20 ppm when ca. one-half of
the nanocages also contain aromatics synthesized from methanol, (e) If P(CH 3)4+ is
formed in the nanocages and the catalyst is heated to high temperatures in air, a
phsophate species forms, as indicated by the broad signal near that o f the
framework. * denotes spinning sidebands.
233
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
P(CH3)4 + oxidized
at 500 °C
P(CH3 )4+ and aromatics 20
MeOH/PH3 = 1.8
450 °C
MeOH/PH3 = 1.2
250 °C
-28
MeOH/PH3 = 0.4
250 °C
0 50 -50 -100
ppm
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 9.2. GC (flame ionization) analyses o f the product gases in MTO conversion
on catalysts functionalized by introducing phosphorous species into the nanocages.
All experiments at 450°C and WHSV = 8 h’1. (a) Control experiment on standard
HSAPO-34; 100 % conversion, ethylene selectivity = 37 %. (b) Modified catalysts
with P(CH 3)4+ in all nanocages; 0 % conversion; (c) Modified catalyst with a
phosphate species formed in every nanocage by air oxidation (500 °C) o f P(CH3)4+;
50 % conversion, ethylene selectivity = 46 %. (d) Modified catalyst with phosphate
formed in 50 % of all nanocages by air oxidation (500 °C) of P(CH3)4+; 8 8 %
conversion, ethylene selectivity = 44 %. Note that the catalyst in this case has
nearly the same activity as the standard catalyst, but significantly higher ethylene
selectivity. DME denotes dimethylether.
235
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ethylene Propene
P(C H 3 )4 + 50%
Cages, Oxidized
at 5 0 0 °C
P(CH3 >4 + Every Cage,
Oxidized at 500 °C
P(CH3 ) / Every Cage
No Oxidation
Control Experiment
Methanol
DME
12
T "
14
g I I I . I m - . y , . . ..... . . . . . . . . . . . . .
16 18 20
Retention time (min.)
236
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the reactor to form a material with a phosphate species in every cage produced a
catalyst with greatly improved ethylene selectivity (46 %) but reduced activity.
When we prepared a material with phosphate in ca. one-half o f all cages, we
achieved 95 % conversion and retained an impressive ethylene selectivity of 44 %.
The results in Figures 9.1 and 9.2 are representative and reproducible.
13
Functionalized catalysts deactivated at the same rate as standard catalysts. C
NM R showed, as before, that deactivation occurred through the formation of
aromatic species in nanocages well beyond the degree required for an active
31
catalyst. P NM R o f deactivated catalyst showed no loss of phosphate.
Furthermore, we found that phosphated catalyst reactivated by burning the
hydrocarbons out o f the cages regained the activity and selectivity o f freshly
prepared catalyst.
Shape selectivity in zeolites and related materials is logically divided into
reactant, product, and transition state selectivity. Reactant selectivity does not apply
here. When we flow propene over standard catalyst we observe formation of
aromatics and C4 , C5 olefins, but less ethylene. Thus, product shape selectivity is
also excluded, and we conclude that phosphated HSAPO-34 achieves higher
ethylene selectivity through transition-state selectivity.
237
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9.2 References and Notes
(1) Stocker, M. Microporous Mesoporous Mater. 1999, 29, 3-48.
(2) W ilson, S; Barger, P. Microporous Mesoporous Mater. 1999, 29, 117-126.
HSAPO-34 was prepared according to: Lok, B. M.; Messina, C. A.; Patton,
R. L.; Gajek, R. T.; Cannan, T. R.; Flanigen, E. M. U.S. Patent 4,440,871,
1984. XRD showed a pure crystalline phase with the CHA structure. The
product was calcined at 600 °C for 10 hours to remove the template agent and
pressed into 1 0 - 2 0 mesh pellets.
(3) Song, W.; Haw, J. F; Nicholas, J. B.; Heneghan, K. J. Am. Chem. Soc. 2000,
122, 10726-10727.
(4) (a) Haw, J. F.; Nicholas, J. B.; Song, W. G.; Deng, F.; Wang, Z. K.; Xu, T.;
Heneghan, C. S. J. Am. Chem. Soc. 2000,122,4763-4775. (b) Dahl, I. M.;
Kolboe, S. J. Catal. 1996,161, 304-309. (c) Dahl, I. M .; Kolboe, S. J. Catal.
1994,149, 458-464. (d) Mole, T.; Whiteside, J. A.; Seddon, D. J. Catal.
1983, 8 2 ,261-266.
(5) (a) Kaeding, W. W.; Butter, S. A. U.S. Patent 3,911,041, 1975. (b) Sun, H.
U.S. Patent 5,925,586,1999.
(6 ) Air oxidation o f catalysts containing P(CH3 )3 gave partial conversion to
OP(CH3 )3, which had a 31P chemical shift o f 39 ppm in SAPO-34 (spectra
not shown).
31
(7) 121 MHz P Bloch decay MAS NMR spectra were acquired on a
Chemagnetics CMX-300 spectrometer (400 scans, 90° flips, 10 s delays).
For discussion of the pulse-quench in situ NMR experiment see: (a) Haw, J.
F.; Goguen, P. W.; Xu, T.; Skloss, T. W.; Song, W.; Wang, Z. Angew.
Chem. 1998, 37, 948-949. (b) Goguen, P. W.; Xu, T.; Barich, D. H.; Skloss,
T. W.; Song, W.; Wang, Z.; Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc.
1998,120, 2651-2652.
(8 ) P(CH3)4+ has been synthesized on the surface of sulfated zirconia catalyst by
the disproportionation o f P(CH3 )3. Haw, J. F.; Zhang, J.; Shimizu, K.;
Venkatraman, T. N.; Luigi, D.-P.; Song, W.; Barich, D. H.; Nicholas, J. B.
J. Am. Chem. Soc. 2000.1 2 0 ,12561-12570.
(9) Rothwell, W. P.; Shen, W.; Lunsford, J. H. J. Am. Chem. Soc. 1984,106,
2452-2453.
238
R e p r o d u c e d with p e r m is s io n o f th e c o p y rig h t o w n e r. F u r th e r r e p ro d u c tio n p ro h ib ite d w ith o u t p e r m is s io n .
(10) Each unit cell of CHA contains 36 T sites and three cages. Our HSAPO-34
has one Si, five P, and six A1 per cage.
239
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Fundamental investigation of the interactions of zeolite and zeotype acid catalysts with hydrocarbons and oxygenates

PDF

I. Metal nanoparticles on high surface area latex supports: Preparation and applications, and, II. Modified semiconductor quantum dots as luminescent label for in situ hybridizations

PDF

I. Oxidation of diazo phosphonates and diazo bisphosphonates. II. NMR and crystallographic studies of bioactive semicarbazides and Schiff bases of hydroxyguanidine

PDF

Investigation into the syntheses and the chemistry of hard -ligated organoiridium complexes

PDF

Chemisorption studies of hydrocarbons on clean and hydrogen precovered platinum(111) and tin/platinum(111) surface alloys

PDF

Electron detachment of singly charged anions

PDF

Characterization of ordered tin-platinum surface alloys and tin-oxide overlayers on platinum(111) using ultra-high vacuum-scanning tunneling microscopy

PDF

Coupling reactions of polystyrene anions with bifunctional electrophiles

PDF

Crotonaldehyde hydrogenation on platinum(111)-based model catalysts

PDF

Imaginary time quantum Monte Carlo simulations of many fermion systems

PDF

Electrochemical studies of membranes, catalysts, and fuels for direct oxidation fuel cells

PDF

Factors affecting child survival in Pakistan

PDF

Computer simulation studies of charge transfer through biological and artificial membrane channels

PDF

Estrogen effects on excitability and plasticity in hippocampus

PDF

Design and synthesis of novel eicosanoids

PDF

Carbon-hydrogen bond activation: Investigation into the dynamics and energetics of alkane complexes within a hard-ligated rhodium system

PDF

Design considerations in the application of membrane reactors for use in steam-carbon dioxide reforming environments

PDF

Contribution of Lifshitz-van der Waals interaction to surface free energy of perfluorocarbons

PDF

At what age are gait characteristics mature? Evaluation of gait kinematics, kinetics, and intersegmental dynamics in 7 year-old children

PDF

History written with lightning: Film, television and experimental historiography

Asset Metadata

Creator Song, Weiguo (author)

Core Title Fundamental studies of methanol to olefin (MTO) catalysis

School Graduate School

Degree Doctor of Philosophy

Degree Program Chemistry

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag chemistry (physical chemistry),chemistry, analytical,chemistry, inorganic,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Haw, James F. (committee chair), [illegible] (committee member), Sahimi, Muhammad (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-110212

Unique identifier UC11338184

Identifier 3027781.pdf (filename),usctheses-c16-110212 (legacy record id)

Legacy Identifier 3027781.pdf

Dmrecord 110212

Document Type Dissertation

Rights Song, Weiguo

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA