Ea;

Elm-b} ‘

:31 A , '
241*f-

'23:: - m}

2&3

:3

$.24: '

 

x:
«‘4
4

 

 

 

 

3 g ,“3 ‘
53.1.41.“ A

‘4 1i:
A" r.

1'
'o

a”
3.

Ni? '42;
u no,
39 I

F!

f
A ‘ .

‘ ’;::":‘2'p"‘“

. . -

'-..‘.‘. .;
{Jr}.

|‘ ‘1 fl '

‘ O
543:3
vu‘v.

, ,

 

A 7
¢ Wty’z
‘ ’41". .

\J an.
\, 1»
I?" .‘w

m
:1,
w

I
y-

'7‘
2‘»,
,. ‘

Q

g"; "'1'
w- -

‘19s

r;

' t

w; 1',
5’5

 

llllllllllllllllllllllllllllllllllllll

31293 01409 9232 LIIRARY
“93 Michigan State
University

                                 

 

 

 

This is to certify that the

thesis entitled

 

Catalytic Conversion of Methanol to l
Hydrocarbons by Polyphosphoric Acid: 5
A Mechanistic Study 1 ;

presented by

Mohamed Al-Azab

has been accepted towards fulfillment
of the requirements for

M o S 0 degree in Chemi StY‘y

(02..

fl Major profesg

[kneDecember 10, 1994

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

PLACE ll RETURN BOX to rornovo this checkout from your record.
TO AVOID FINES rotum on or boron duo duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Afflrmofivo ActionlEmol Opportunity Intuition
W1

 

 

Catalytic Conversion of Methanol to Hydrocarbons by
Polyphosphoric Acid: A Mechanistic Study

By

Mohamed A. Al-Azab

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Chemistry
1995

ABSTRACT

Catalytic Conversion of Methanol to Hydrocarbons by
Polyphosphoric Acid: A Mechanistic Study

by
Mohamed Al-Azab

This thesis describes mechanistic studies concerning the formation of
the initial C-C bond in polyphosphoric acid (PPA) catalyzed methanol to
hydrocarbons conversion ("The Pearson Reaction"). We believe that the
Pearson reaction, in which the methanol is converted to hydrocarbons at
200°C by PPA, operates by a mechanism similar to the zeolite-catalyzed
"Mobil Process". A system was designed to run time course studies to
examine the identities and appearance times of hydrocarbon products
obtained from reacting trimethyl phosphate (TMP) over PPA at various
temperatures and with a number of additives.

Isobutane and isopentane were detected as the first products at 180°-
240°C. Ethylene, which most research groups reported as the first product
in zeolite systems, appeared at higher temperatures (>200°C) in the present
system. The observed products are saturated hydrocarbons (CnH2n+2)
rather than alkenes (CnHzn), which would be expected on grounds of the
simple dehydration stoichiometry nCH3OH to give CnHzn + nH20. The
possible role of carbon monoxide (CO) as a catalyst in the initial C-C bond
formation via the intermediacy of ketene was investigated. A modified CO
mechanism is proposed in order to explain the new results. Additive
studies were done to examine the proposed mechanisms both for the CC

bond formation and for the production of saturated hydrocarbons.

To my family
To my parents

ACKNOWLEDGMENTS

I would like to express my great appreciation to Dr. Jackson for his
guidance, assistance, encouragement, and friendship during the course of
this work. Without his support and ideas this work would never have been
accomplished. I would also like to thank those on my committee Dr.
Karabatsos, Dr. Miller, and Dr. Smith. Special thanks to those individuals
who helped me carry this work forward: Doug Hoover for his studies with
me in part of this project, Ahmad Madkuor for his discussion and
friendship, and Larry Szajek for his editing. I am grateful to Dr.
Karabatsos for his support and wisdom. I would also like to thank Dr.
Jackson's family; Evy and Kelvin for their friendship.

I would like to extend my deepest thanks to my wife Hessa for her
understanding and sacrifice, my daughter Nouf, and my sons Khaled and
Majed.

I would especially like to thank my parents for their spiritual support
and prayers, my brothers, sisters, and The United Arab Emirates
University for their financial support.

Finally, I would like to thank the members of my research group for
their friendship and help.

iv

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Page
List of Tables vii
List of Figures viii
Chapter 1: Historical and Overview of Methanol to Gasoline 1
1.1. Introduction 1
1.2. The Oxonium Ylide Mechnisem 2
1.3. The Carbene Mechanism 7
1.4. The Free Radical Mechanism 9
1.5. Nonzeolite-Catalyzed MTG Conversion "The Pearson
Reaction" 1 3
1.6. The First Products From Methanol to Hydrocarbon
Conversion 1 4
1.7. The Role of Carbon Monoxide in Methanol to Hydrocarbon
Processes 1 8
References 22
Chapter 2: Method Development and Experimental Results 24
2.1. Focus of Research 24
2.2. Method Development 26
2.2.1. Standard Materials 26
2.2.2. Nuclear Magnetic Resonance 26
2.2.3. The Reactor Design 26
2.2.4. Method Used 27
2.2.5. Reproducibility of The Reaction 29
2.3. Experimental Results 30
2.3.1. Products of Reaction of PPA with TMP 30
2.3.2. Proton Decoupling Experiment 31
2.3.3. Effects of Changing TMP Concentration 35
2.3.4. Effect of Additives 35
2.3.4.1 Addition of Methyl Formate 35

V

2.3.4.2. Addition of Isobutylene 36

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.3.4.3. Addition of t-Butanol 36
2.3.4.4. Addition of 2-Propanol 38
2.3.4.5. Addition of Pivalic Acid (2,2-Dimethyl

Propanoic Acid) 38
2.3.4.6. Addition of Acetic Acid 41

2.3.4.7. Addition of Formic Acid, Propanoic Acid and
Acetic Anhydride 41
2.3.4.8. Addition of CO 44
Chapter 3: Analysis and Discussion 50
3.1. Introduction 50
3.2. Key Observations 51
3.3. Modified CO-Catalyzed Mechanism 55
3.4. Ethylene is not The First Product 57
3.5. Hydride Source 60
3.6. Effect of Additives 60
3.6.1. Addition of Isobutylene 61
3.6.2. Addition of Pivalic Acid and t-Butanol 62
3.6.3. Addition of 2-propanol 72
3.6.4. Addition of Acetic Acid 72
3.6.5. Addition of CO 73
3.7. Future Work 74
3.8. Conclusion 76
References 78

 

List of Tables
Table Page

Table 2.1. Qualitative summary of products distributions over time

 

course at 240°C ' 32
Table 2.2. The conversion of acetic acid to methyl acetate at 180°C
(2.055 x 10'4 moles of TMP was used) 43

 

Table 2.3. Volumes of CO added 44

 

vii

List of Figures

Figure Page

Figure 2.1. The reactor design 28

 

Figure 2.2. 1H NMR for the primary products (isobutane and isopentane)
from the reaction of PPA with TMP after 30 minutes at 200°C
(80:20 PPA to M ratio) compared to the pure materials.
(A) Isobutane, (B) Isopentane, (C) Product mixture ......................... 33

Figure 2.3. 1H NMR spectrum for the reaction products. At 200°C after
30 minutes (80:20 PPAzTMP ratio) 34

 

Figure 2.4. The effects of addition of isobutylene (A) and methyl formate
(B) on the product onset (isobutane + isopentane) at 180°C
using 200:1 TMP to the additive. Curve (C) represents the
behavior of TMP alone 37

 

Figure 2.5. The effect of adding t-butanol to the reaction mixture using
(100:1) TMP to t-butanol ratio at 180°C 39

 

Figure 2.6. The effect of adding 2-propanol to the reaction mixture
compared to the reference reaction. Ratio of 200:1 of TMP to
2-propanol at 180°C 40

 

Figure 2.7. The effect of addition of pivalic acid to the reaction mixture
at 180°C using 200:1 TMP to pivalic acid ratio 42

 

Figure 2.8. Number of moles of product and moles of carbon (C)
trapped from reaction of PPA with TMP with CO added in a
ratio of 600:1 TMP:CO at 180°C 45

 

Figure 2.9. Number of moles of product and of carbon (C) trapped from
reaction of PPA with TMP with CO added in a ratio of 600:1
TMP:CO at 180°C 46

 

Figure 2.10. Number of moles of product and of carbon (C) trapped
from reaction of PPA with TMP with CO added in a ratio of
100:1 TMP:CO at 180°C 47

 

viii

Figure 2.11 Number of moles of product and of carbon (C) trapped from

reaction of PPA with TMP with CO added in a ratio of 800:1
TMPzCO at 180°C 48

 

Figure 2.12. Number of moles of product and of carbon (C) trapped

Figure 3.1.

Figure 3.2.

Figure 3.3.

Figure 3.4.

Figure 3.5.

from reaction of PPA with TMP with CO added in a ratio of
600:1 TMP:CO at 200°C 49

 

Addition of isobutylene to the reaction of PPA with TMP
compared to the reference reaction at 180°C using 100:1(TMP
to isobutylene) ratio (6.7 x 10'5 moles of isobutylene)...................63

Addition of pivalic acid to the reaction of PPA with TMP
compared to the reference reaction at 180°C 200:1 TMP :
pivalic acid ratio (3.42 x 10'5 moles of pivalic acid) ......................... 66

Addition of t-butanol to the reaction of PPA with TMP
compared to the reference reaction at 180°C 100:1 TMP :
t-butanol ratio (6.8 x 10‘5 moles of t-butanol) 68

 

Addition of pivalic acid to the reaction of PPA with TMP
compared to the reaction of pivalic with PPA without TMP
using 100:1 TMPzpivalic acid ratio (6.85 x 105 moles of
pivalic acid) at 180°C 69

 

Addition of t-butanol to the reaction of PPA with TMP
compared to the reaction of t—butanol with PPA without TMP
using 100:1 TMPzt-butanol ratio (6.8 x 10-5 moles of
t-butanol) at 180°C 70

 

ix

Chapter 1

Historical and Overview of Methanol to Gasoline

1.1. Introduction

The conversion of methanol to hydrocarbons has been intensively
studied for the two decades since its discovery.1 Numerous research
groups have extensively investigated this conversion using a variety of
catalysts.2 These investigations have led to the use of zeolite H-ZSM-S,
developed by Chang at Mobil“, as well as other zeolites in the petroleum
refining industry. In conjunction with existing technology for converting
natural gas, coal, and biomass to methanol, the methanol to gasoline (MTG)
conversion process offers a non-petroleum-based pathway to fuels and
organic intermediates. This technology saw its first commercial
application in 1985 in New Zealand where a "Mobil process" plant was
designed and built to supply one-third of the country's gasoline demand.5
Although the continuing low price of imported oil ultimately led to the
failure of these experiments, this episode illustrated the practical feasibility
of large-scale MTG conversion as a source of gasoline hydrocarbons.

H-ZSM-S zeolite has become the standard catalyst for the "Mobil
Process." This zeolite is a crystalline aluminosilicate composed of A104
and SiO4 tetrahedra linked through shared oxygen atoms to form a unique
three-dimensional crystal framework. The structure is porous with a

network of channels that allow small molecules to traverse the lattice. The

2
aluminate sites are anionic and therefore require charge-balancing cations

in the open cavities or pores of the zeolite. When these cations are protons
the zeolite (H-ZSM-S in this acid form) is an extremely strong acid. The
numbers of acidic sites in the zeolite combined with the shape selectivity of
the pore structure have been noted to be important factors which control
product selectivities depending on the molecular size and shape of the
diffusing molecules.2 The acid catalyst WO3/A103 has also been used in
the MTG conversions.t5

Many mechanisms have been proposed for the formation of the
initial C-C bond in the MTG conversion, but this process is mechanistically
not obvious. It remains a challenge for scientists in this field to understand
this reaction in detail. The most widely discussed and examined mechanistic
classes are:

1) Intermolecular alkylations of oxonium ylides;

2) OH insertion by carbenes formed via a-elimination;

3) Free radical pathways.

Brief surrunaries of these proposed mechanisms will be presented in

the sections that follow.

1.2. The Oxonium Ylide Mechnisem

The oxonium ylide mechanism has been suggested by two research
groups using two different catalysts. It has also become the most widely
invoked pathway by other researchers in this area. Van den Berg and co-
workers’, using H-ZSM-S, and Olah8 using WO3/Ale3, have proposed a

3
mechanism that invokes a methyleneoxonium ylide as intermediate.

Scheme 1 shows the mechanism outlined by Olah for the reaction catalyzed
by WO3/A1203, He proposed that the first step in this conversion is the
bimolecular dehydration of methanol to dimethyl ether. The latter is then
alkylated via acid-catalyzed trans-methylation to give trimethyloxonium ion
I-l. Deprotonation of [-1 then gives dimethyloxonium methylide 1-2,
which undergoes intermolecular methylation to ethyl dimethyl oxonium ion
1-3 instead of intramolecular Stevens rearrangement to methyl ethyl ether
[-4. Upon deprotonation by methanol, dimethyl ether, or basic sites in the
catalyst, I-3 would yield ethylene (Scheme 1). A parallel pathway would
proceed by Lewis acid coordination, deprotonation, and alkylation of
dimethyl ether on the catalyst surface as shown by [-5 through I-7 of

Scheme 1.

Sodium hydride and lithium tetramethyl piperidide were used as
bases to demonstrate the required deprotonation of trimethyloxonium ion
(TMO).6 Reaction of trimethyloxonium salts with sodium hydride (NaH)
produces a considerable amount of urethane and dimethyl ether (Scheme 2).
These results were interpreted as hydride methylation by trimethyloxonium
ion to give methane and dimethyl ether.

The formation of ethylene and ethane could be explained by proton
abstraction to form methylene dimethyloxonium ion which is then
methylated to form ethyl dimethyl oxonium ylide, leading ultimately to
formation of methane and ethane (see Scheme 2). NaH is normally
considered a strong base and a poor donor of nucleophilic hydride.

However, NaH was diverted to follow path B instead of following path A

 

 

 

 

 

 

 

 

Scheme 1
-H¢O H3C-O'CH3 H3C~ ’CH
2 013—011 u,c-o-cu3 t 3
Cat. [-1 CH3 + CH3O-Cat.
cat.
—H+

H3C\ +

1-5 ,O-cat. -
H3C H3C ~B’CH2
- H+ 1.2 (EH3 + CH3O-cat.

H C

H3C 3 \ +

1'6 - :d-cat. ,O-CH3
H2C 53¢

+
H3C\6. CH H3C ~93CH2CH3
H 3C’ 3 1-3 CH3

H3C\

1'7 ,d-cat.

H3CH2C

 

crrza=cu2 + H3C-O-CH3

 

H
CHZ-CH2 + ,O+-cat.
Hzc

H;,C.-1O»,(':H2
éHa _+_. cu,ocu,crr3

Stevens
[-2 rearrangement 1.4

5
(forming methane and dimethyl ether as major products instead of forming

ethylene) as shown in Scheme 2. This experimental result suggests that the
formation of an oxonium ylide by deprotonation of a TMO salt is a
difficult process in general and does not support the involvement of TMO

as an ylide source in MTG conversion.

 

Scheme 2
H C.+.CH NaH
(A) 3 9 3 s CH4 +crr3ocu3
CH3
NaH ’Hz
ll

 

H3C~5'CH3 (CH3)30+ _ H3C~6£H2CH3

a -1 .
( ) CH2 CH3 \w
NaH A
(CH3)20 + H2C=CH2

CH3OCH2CH3 + CH, (cu,)zo + H3C-CH3

 

In 1991 Haw9 observed trimethyloxonium ion for the first time in a
MTG process by 13C magic-angle spinning (MAS) NMR. Heating dimethyl
ether-1-13C to room temperature and then adsorbing it on H-ZSM-S gave a
signal at 80 ppm which was assigned to the trimethyloxonium ion.

"Is the conjugate base form of the acidic zeolite sufficiently basic to

deprotonate a non bonded intermediate to generate the ylide

6
intermediate?"10 This question was raised in regards to the involvement of

methyloxonium ylide in the MTG conversion.

Hunter and Hutchingsll carried out experiments to show whether an
Al-O moiety in a zeolite framework is basic or nucleophilic in its reaction
toward the TMO ion. In their work, lithium aluminum tetraisopropoxide
(LAT) was used as a model for the zeolite basic site in which isopropoxide
groups replace the silyloxy groups in the zeolite. When LAT was reacted
with TMO+SbC16', no significant products resulting from ylide formation
were observed; products were only consistent with either methylation of
isopropoxide or nucleophilic chloride attack. Furthermore, reaction of
LAT with trimethyl sulfonium (TMS) iodide (a methylating agent, but not
as strong as TMO) gave no reaction indicating that the Al-O moiety is not
sufficiently basic to deprotonate a TMS salt (ylide precursor) and not
sufficiently nucleophilic to demethylate TMS. From these results, Hunter
and Hutchings concluded that the Al-O moiety in the zeolite should be more
nucleophilic than basic, so it would not be likely to form an ylide

intermediate.

Additional evidence against the involvement of the trimethyloxonium
(TMO) ion as an ylide source was obtained by comparing the relative
reactivity of sulfur and oxygen containing reagents. Hunter and Hutchings
used (CH3)28 instead of (CH3)2O as the feed stock.12 They assumed that the
deprotonating step would be the rate limiting step, and that the sulfur
would enhance the reactivity of the reagent, but they did not observe any
enhancement. They concluded that TMO is the methylating agent but not

the ylide source.

1.3. The Carbene Mechanism

This mechanism involves the formation of carbene (:CH2) by 0.-
elimination of water from methanol by the concerted action of acidic and
basic sites of H-ZSM-S. Olefins are formed either by insertion of :CH2
into methanol or dimethyl ether and their subsequent reactions or by
polymerization. Chang and Silvestril do not agree with those13 who
suggest the polymerization of carbene to alkenes as shown in Scheme 3,

because the reactivity of the free carbene would make its presence short
lived.

Scheme 3

A 1:1 .. (1
cat-0‘ +H-C-9H + H-O-cat —-- H'O-cat 4-:CH2

H + 1120 + cat-0'

H3C-OH + :CHz —’ C2H5-OH

 

n :CH2 (CH2) n "Polymerization"

Diazomethane (CH2N2) has been used as a carbene source to provide
evidence for the involvement of carbene as an intermediate. 14 Lee and Wu
noted that the production of ethylene and propylene was enhanced when
diazomethane was reacted over H-ZSM-S. They suggested that the

methylene which is generated from diazomethane reacts over H-ZSM-S at

8
200°C leading to the formation of surface carbene (carbenoid) via a-

elimination of methylene from diazomethane. Lee and Wu concluded that
the singlet methylene, initially formed by decomposition of diazomethane,
can be stabilized by electron donation from the oxygen lone pairs of the
catalyst surface in the vacant p-orbital of the lowest unoccupied molecular

orbital to form the carbenoid or ylide species as shown bellow:

 

 

Olah” criticized the carbene mechanism suggested by Lee and Wu. He
pointed out that the bimolecular dehydration of methanol to dimethyl ether
is exothermic by ca. -24.5 kcal/mol. On the other hand the a-elimination
of water from methanol to form methylene is thermodynamically
endothermic by some 83.5 kcal/mole (in the paper the value reported is
349.3 kcal/mole, but this number appears to be the correct value in
kJ/mole. Furthermore, Olah pointed out that diazomethane is more likely
to be protonated by the acidic sites of the zeolite. Then the protonated
diazomethane (IV-1) would act as a methylating agent towards another
diazomethane molecule, yielding IV-2, and ultimately give ethylene as
shown in Scheme 4.

The carbene mechanism was also ruled out due to the results of

hydrogen and oxygen co-feeding experiments. Co-feeding of hydrogen gas

9
with methanol over H-ZSM-S and WOg/Ale3 catalysts under a broad

range of conditions offered no increase of methane productionfléol" It

would be expected that :CHz would react with Hz to give methane.

Scheme 4

cnzu2 3+ cum; __z_z.CH N CI-IgCH2N2" + N2
rv-r rv-z

_. H"
C112=CH2 + N2

 

Subsequent studies18 demonstrated that co-feeding of a low
concentration of oxygen with methanol and dimethyl ether over H-ZSM-S
caused rapid and irreversible deactivation of the catalyst, rather than the
expected formation of formic acid from the reaction of 02 with :CH2.
From these results Hunter and Hutchings concluded that a gas-phase

carbene species is not a possible intermediate.

1.4. The Free Radical Mechanism

It was noted that dimethyl other can thermally decompose over acidic
catalysts at elevated temperatures to form methyl radicals”. This
possibility was studied in the context of the MTG process over H-ZSM—S.
Using Electron Spin Resonance (E.S.R.) spectroscopy, Clark and co-

workersl9 observed the formation of free radicals during a reaction of
methanol over H-ZSM-S. They proposed that the free radical is -CH20CH3

10
and that it could be produced by either pathways V-A, V-B or V-C as

shown in Scheme 5.

Scheme 5

 

 

 

 

.S+crr,ocn3 > S-H + ~crr2-o-CH3
Mech. V-A ~cuz-ocrr3 0,11,0- WQHSOH
M001]. V-B 2 ° CHz-O-CH3 ; CH30-(CHflz-OCH3
MOCh. V'C 'CHz’O’CH3 : :CH2 '1" 'O'CH3

I CH2 + CH3OCH3 ——" CH3CH2-OCH3

-S = surface radicals

Examination of the generation of carbene via the decomposition of
the methoxymethylene radical (Scheme 5) provided evidence against
mechanisms V-A and V-C leaving some possibility for mechanism V-B.2°
Choukroun examined mechanism V-B, which is the dimerization of the
radical of dimethyl ether. He reported that radical dimerization of
dimethyl ether by (FSO3)2 in FSO3H gave dimethoxy ethane, offering
supporting evidence for mechanism B. The second step in mechanism C
could happen, but the involvement of carbene was already ruled out.

Subsequent experiments”.16 contrasting the relative rates of
methanol conversion over H-ZSM-S and WO3/A1203 upon introduction of

the two stable 1: radicals 02 and NO, showed interesting evidence

concerning the role of radical reactions. When a low level of oxygen was

11
co—fed with methanol, a rapid and irreversible deactivation of H-ZSM-S

was observed, whereas there was no deactivation of WO3/A1203 even with
high levels of co-fed 02. In the case of NO, the opposite was observed,
i.e., deactivation of WO3/A1203 and no effect on the rate with H-ZSM-S.
From this observation the conclusion was drawn that H-ZSM-S and

WO3/A1203 differ mechanistically from each other in the MTG

conversion.

After the publication of the results with NO and Oz, Chang and co-
workers21 investigated in detail the role of NO on the MTG conversion
over H-ZSM-S. They found that the catalyst was deactivated by the
addition of NO, but the deactivation was proportional to the partial
pressure of NO and on the concentration of aluminum in the catalyst. This
result supported the proposed mechanism V-B of Clark as was shown in
Scheme 5. The effect of NO was complicated due to the formation of NH3,
a catalyst poison, which could arise via a Beckman rearrangement as shown

in Scheme 6. In the previous work by Hunter and Hutchings10 the time

scale had not been long enough to observe the effects of NH3,

 

 

 

 

Scheme 6
NO
C1130CH3 ——> -CH3 ——> CH,No : > CHEN-OH
Beckrnann
rearrangement
‘1‘
C0 + NH3 ¢ O=C~NHZ 4 H$=N-H

OH

12

Chang offered an alternative mechanism for the involvement of
radicals in the MTG process.20 The radical (Ro) may be alkyl, alkoxy,
alkoxyalkyl, etc. and generated either homogeneously via gas-phase
pyrolysis or by homolytic scission of surface alkoxyls. The surface-bound
carbene VII-1, essentially a methylene oxonium ylide, would then be the
reactive carbon that undergoes alkylation by methyloxonium ion
intermediates (see Scheme 7).

 

 

 

 

Scheme 7

Z’OH + CH3 OH AV CH30'Z 4' H20

CH3O-Z + R. : (31120.2 4. R-H
:CH2 CH
,0. .. I 2

CH2 0-2 = /0\ <-—> /9\ + '02
VII-l
R-H + '02 : Ro + Z-OH
Z: zeolite

The formation of methyleneoxonium ylide as an intermediate
represents a point of convergence between the free radical mechanism
proposed by Chang (Scheme 7) and the oxonium ylide pathway (Scheme 2).
The difference between these mechanisms is whether the oxonium ylide is

formed via two one-electron processes (Scheme 7) or via one two-electron

13
process (Scheme 2). The surface ethoxy group could be formed by these

two mechanisms (Scheme 8).

 

 

Scheme 8
CH3
H\’Oj
it Elm. ‘fH’CH’
\
o _ -H O O
>Si’ \A1’9\ 2 >A1’+\Si<
| |\ l |
OR _
CH2 (EH3 (szCH3
0 - O -
I |\ I \ | |\

1.5. Nonzeolite-Catalyzed MTG Conversion "The Pearson

Reaction"

Another type of MTG process using a nonzeolite catalyst was
discovered by Pearson.22 Pearson found that dimethyl ether and
hydrocarbons are also produced from methanol in polyphosphoric acid
(PPA).” In the "Pearson reaction", heating PPA and trimethyl phosphate
(TMP) or equivalent amounts of methanol and phosphorus pentoxide, to

temperatures above 195°C produces a mixture of saturated hydrocarbons,

l4
monoaromatics and traces of alkenes with the average formula of CnH2n+2

t0 CnHZn-l range.

Considering the original proposal of ethylene as the first product
along with the observation of carbon monoxide in the H-ZSM-S catalyzed
MTG process“, Jackson and Bertsch proposed a new mechanism in which
CO was the effective catalyst in the PPA-catalyzed MTG conversion.25
This suggestion also accounted for the induction periods before the onset of
hydrocarbon production in both H-ZSM-S and PPA catalyzed processes,
assuming the two reactions operate by similar mechanisms. The
Jackson/Bertsch studies used a closed system in which mixtures of methanol
or trimethyl phosphate (TMP) and PPA were heated at 200°C for a period
of 2 hours. Ethylene was one of the products observed from this reaction,
as detected by gas phase 1H NMR. In the Jackson mechanism shown in
Scheme 10, CO activates the methyl group originating from CH30H to
form an acetyl cation (X-l). This species loses a proton to form ketene
(X-2), which then functions as a carbon nucleophile, undergoing successive
methylations, to give X-3 to X-S and leading ultimately to the formation
of olefins. This mechanism was examined in detail by ab initio

calculations.

1.6. The First Products From Methanol to Hydrocarbon

Conversion

Is ethylene (CH2=CH2) the first product in PPA-catalyzed MTG
chemistry? Many experiments have been designed to investigate and

differentiate the various mechanisms which have been pr0posed in both

15
systems (using H-ZSM-S or PPA catalyst). In many of these cases, the

apparent primary product formed from methanol is ethylene. It has been
proposed that after traces of ethylene are formed from methanol, it

undergoes other transformations yielding secondary products.

 

 

Scheme 10

H 'CH + 9’ +
0+ H3C-OXY 9* XYO= g 2C' 2 g...“ XYOH

if- _’ < . ' .C. .1

" , cu3 xvou CH2 ' XYO’

+
x-rx-r cngoxv 0+
8

V CH2CH3
+ x-3
0* CHgOXY 0 XYOH

+ "I II
0 XYO C

fl
Ill

..- éHtCHa). circa,
+ X-S X-4
H3 CHCg CHg
+ XYbH

x, Y = H, CH3, Si(OR)3. A1'(OR)3, P(OR)3(O)

Based on labeling studies using 13C NMR and deuterium-labeled
feeds in MTG conversion over H-ZSM-S zeolite, Molezm" proposed that
part of the ethylene is formed "directly" from methanol and the remainder

is formed "indirectly. Directly formed ethylene was derived only from
the carbon of methanol and did not incorporate the carbon of any added
olefin or alcohol. However, the ethylene formed indirectly incorporated
some carbons from added higher alcohol feeds (i.e. not methanol). The

"direct" formation of ethylene from methanol appeared consistent with the

l6
oxonium-ylide mechanism. However Dessau argued that most of the

ethylene observed comes from alkylation and cracking processes”, i.e. the

"indirect" mechanism.

Ahn, et a1.29 and Anderson, et al.30 found ethylene to be substantially
less reactive than propene over many catalysts. Ethylene at 400°C
exhibited low reactivity, while propene was readily converted to aromatic
products. This implied that ethylene could not be an active intermediate in
the H-ZSM-S catalyzed MTG conversion.

Although ethylene has been identified as a primary product under a
wide range of reaction conditions, Sulikowski and Klinowskiii1 found, based
on G.C. studies, that trans-2-butene is the first hydrocarbon product from
H-ZSM-S zeolite at a temperature below 200°C and ethylene was not
detected among the products under these conditions. They rationalized this
result by the formation of a methoxy group on the acidic site which reacts
with another CH3OH molecule to yield ethoxy groups on the catalyst

surface as shown in Scheme 11.

Schemell
It"?
If?" H CH
H '0' '0. 3
Sc-H ' H’o‘
H 1 9‘2 ‘ H
+ + t
\ .’O\ /O\ / \ ’0‘ /O\ ./
/Sl\o A1\0,s \ 22> /Sl‘o A1\O, Sr\

17
Instead of yielding ethylene, the ethoxy groups reacts with methanol to give

propoxy and then butoxy groups which then undergo dehydration to give
trans-2-butene. The details of this mechanism are murky.

In 1988, the novel technique of Differential Scanning Calorimetry
was used by Kolboe32 to study the methanol to hydrocarbons reaction over
H-ZSM-S and non-zeolite catalysts at low temperatures (185°-240°C).
Dimethyl ether, isobutane and apparently isopentane were observed as the
first products after adsorption of dimethyl ether on the catalysts at 185°C.
This study clearly defined the induction period, after which a sharp onset

of hydrocarbon production accompanied by a heat release, was observed.

More recently, Hutchings and co-workers33 investigated the
conversion of methanol to hydrocarbons by using a pulsed reaction method
over zeolite H-ZSM-S similar to that used by Kolboe32 and by Klinowski31.
Pulses of methanol were injected into a gas chromatograph injector
modified to function as a micro-reactor. Products from this reaction were
analyzed by packed and capillary columns. Dimethyl ether and traces of
butanes were detected at 112°C, but no C4 hydrocarbons were detected at

higher temperatures. Zeolite SAPO-34 was used under the same reaction
conditions, but again no formation of C4 hydrocarbons was observed.
However, very recently, Koboe and Dahl34 observed propene and butenes
from the reaction of 13C-methanol and 12C-ethylene (fed as ethanol) over
SAPO-34 in a flow system at 400°C using argon as a carrier gas. Isotopic
labeling studies indicated that products (propene and butenes) were mostly

formed from methanol and contained a large excess of 13C atoms.

18
Based on their results, Hutchings suggested that the first step in C4

hydrocarbon formation is the generation of a surface methoxy group
(XIII-l) via methylation of the active sites of zeolite by methanol as
shown in Scheme 13. The surface methoxy group then reacts with a
hydrogen-bonded molecule at the same O-Al active center yielding a
surface ethoxy group (XIII-2). The ethoxy group has two possible
reaction pathways. One is to undergo a B-hydrogen elimination to give
ethylene. The other is to react with another methanol molecule to give a
surface isopropoxy group (XIII-3). A further addition of another
methanol molecule yields a surface t-butoxy group which could give
isobutane by hydride abstraction from methanol (or other suitable hydride

source) or isobutylene from B-proton elimination.

When ethanol pulses were used in the same technique no C4

formation was recorded; however, a surface ethoxy group would have been
expected from the mechanism as well as formation of ethylene, propene,
and isobutylene. These species were not observed experimentally.
However, very recently Hutchings and co-workers found that isobutane is

the preferred product from the MTG conversion using zeolite [3.35

1.7. The Role of Carbon Monoxide in Methanol to Hydrocarbon

Processes

The role of carbon monoxide (CO) has been investigated previously
in methanol to hydrocarbon processes.”42 In 1979 Derouane37 reported

that CO had no effect in the MTG conversion, although some CO was
incorporated in the products. However, Jackson and Bertch noted from the

19
data displayed in their paper, in the absence of adding CO no hydrocarbons

were observed until 8 min. at 350°C. When CO was added, the

hydrocarbons appeared faster (at 4 min.) and increased linearly.

 

Scheme 13
Ethylene
11 B-Elimination
(EH3 CH3
9H3 ,0 H3613 cux'r‘g:
- + I — s
’O"'°"!'I (H 1!] 6""‘H 5 o:
H r I 2' : ——-> H/ ,: HC-CH3 [-120
"o .(1) /0\ ,0\ "'0 Q o’ o
/ \lil + slim, / \jT/+\S|i{ \
Methoxy Ethoxy XIII-2
v
cry/.3
CH3 Q)
, = '1:

1120 H3C—Cli—CH3 1120 H, H30”? CH3 H20
0, oH20 o
OVK1’9\31’0\ ‘— /0\"°' ’9\Si’ \
| 1 1 1"»

t-butoxy Isopropoxy XIII-3

H Transfer / \B-Elimination 11 B-Elimination

Isobutane Isobutylene Propcne

20
Anderson and Klinowskil-i-ii8 observed CO in the MTG reactions

over zeolite H-ZSM-S using magic-angle spinning NMR (MAS). The
initial resonances that appeared at 250°C were for methanol (50.5 ppm),
DME (60.5 ppm), and carbon monoxide (184 ppm). No higher molecular
weight hydrocarbons were observed at this temperature. After heating the
sample at 300°C for 10 minutes, other resonances for aliphatics and
aromatics were recorded. The assignment of the CO peak was based on the
following: First, the resonance was stable and increased in intensity at
increasing temperatures. Second, no proton coupling was observed for that
signal. Third, the resonance disappeared when the spectrum was measured
with cross-polarization which means that this peak belonged to a mobile

non-attached species.

Haw 35 however, reported that CO does not catalyze or get
incorporated in the MTG conversion over zeolite H-ZSM-S, based on his
l3C MAS NMR studies. Reaction of 13CH3OH in the presence of H13C02H
(formic acid) as a source of 13C0 in the reaction resulted in no increase of
hydrocarbon formation compared to the same reaction with no CO source.
The same experiments were repeated but with unlabeled methanol and no
l3C labeled hydrocarbons were recorded in the 13C MAS NMR spectra.
According to these results he concluded that CO is neither an intermediate

nor a catalyst in the MTG conversion.

Hutchings et. al.33 also ruled out the role of CO in the formation of
the initial C-C bond in the MTG reaction over zeolite H-ZSM-S at
temperatures between ZOO-300°C. They reported that CO has no important
effect on the induction period for this process either as a catalyst or as a

reactive intermediate. These results were supported by labeling and model

21
experiments. In model experiments in which ketene was exposed to

MezSO4 and Me3O+SbCl5’ they found no methylation of the ketene by
either of these methylating agents. Hutchings concluded that the ketene acts
only as an acylating agent in the normal way, and that methylation of
ketene, if it occurred, would react at the oxygen and not at either carbon

atom.

22

References

flaw-5138)

10
11

12
13
14
15

l6

17
18

19

20

21

22
23

Meisel, S. L.;Mc01110ugh, J. P.; Lechthaler, C. H.; Weisz, P. B.
Chemtech, 1976, 6, 86.

Chang, C. D.; Silvestn' A. J. J. Catal. 1977, 47, 249.

Chang, C. D. Catal. Rev. Sci. Eng. 1983, 25, 1.

Chang, C. D. Stud. Surf. Sci. Catal. 1988, 36, 127.

Maiden, C. J. Stud. Surf Sci. Catal. 1988, 36, l.

Olah, G. A. Pure Appl. Chem. 1981, 201.

Van den Berg, J. P.; Wolthuizen, J. P.; Van Hoof, J. H. C. In
"Proceedings, 5th International Conference on Zeolite; Rees, L.V.
C., Ed.; Heyden: London, England,l980; p 649.

Olah, G. A.; Doggweiler, H.; Felberg, J. D. J. Org. Chem. 1984, 49,

21 12.
Munson, E. J.; Haw, J. F. J. Am. Chem. Soc. 1991, 113, 6303.
Hutchings, G. J.; Hunter, R. Stud. Surf. Sci. Catal. 1990. 38, 279.

Hunter, R.; Hutchings, G. J. J. Chem. Soc., Chem. Commun. 1985,

886.

Hutchings, G. J.; Gottschalk, F. M.; Hall,V. M.; Hunter, R. J. Chem.
Soc., Faraday Trans. I 1987, 83, 571.

Venuto, P. B.: Landis, P. S. Adv. Catal. 1968, 18, 259.

Lee, C. 8.; Wu; M. J. Chem. Soc., Chem. Commun. 1985, 250.

Olah, G. A.; Surya Prakash, G. K.; Ellis, R. W.; Olah, J. A. J. Chem.
Soc., Chem. Commun. 1986, 9.

Hunter, R.; Hutchings, G. J .; Pickle, W. J. Chem. Soc., Chem.
Commun. 1987, 1369.

Hutchings, G. J .; Jansen Van Rensburg, L.; Pickle, W.; Hunter, R. J.
Chem. Soc., Faraday Trans. 1 1988, 84, 1311.

Hunter, R.; Hutchings, G. J. J. Chem. Soc., Chem. Commun. 1987,
377.

Clark, J. K. A.; Darcy, R.; Hetgarty, B. F. ; O'Donoghue E.; Arnir
Ebrahimi, V.; Rooney, J. J. Chem. Soc., Chem. Commun. 1986, 425.
Choukroun, H.; Brunel, D.; Germain, A. J. Chem. Soc., Chem.
Commun. 1986, 6.

Chang, C. D.; Hellring, S. D.; Pearson, J. A. J. Catal. 1989, 115,
282.

Pearson, D. E. J. Chem. Soc., Chem. Commun. 1974, 397.

Pearson, D. E. Chem. Eng. News 1983, 50.

24
25
26
27
28
29

30

31
32

33

35

36
37

38
39

41
42

23

Anderson, M. W.; Klinowski J. Nature 1989, 339, 200.
Jackson, J. E.; Bertsch, F. M. J. Am. Chem. Soc. 1990, 112, 9085.
Mole, T. J. Catal., 1983, 84, 423.

Mole, T.; Bett, G. J. Catal., 1983, 84, 435.

Dessau, R. M. J. Catal. 1986, 99, 111.

Ahn, B., J.; Aramando, J .; Perot, G.; Guisnet, M., C. R. Acad. Sci,
Ser. C 1979, 245, 288.

Anderson, J. R.; Foger, K.; Mole, R.; Rajadhyaksha, R. A.; Sanders,
J. V. Ibid. 1979, 58.

Sulikowski, B.; Klinowski, J. Appl. Catal. 1992, 89, 69.

Kolboe, S. Acta Chem. Scand. 1988, A 42, 185.

Hutchings, G. J .; Lee, D. F; Lynch, M. Appl. Catal. A: General,
1993, 106, 115.

Dahl, I. M.; Kolboe, S. J. Catal. 1994, I49, 458.

Hutchings, G. J.; Johnston, R; Lee, D. F.; Warwick, A.; Williams,
C. D.; Wilkinson, M. J. Catal. 1994, I47, 177.

Munson, E. J.; Haw, J. F. J. Am. Chem. Soc. 1991, 113, 6303.
Nagy, J. B.; Gilson, J. P.; Derouane, E. G. J. Mol. Catal. 1979, 5,

393.
Anderson, M. W.; Klinowski, J. J. Am. Chem. Soc. 1990, 112, 10.

Dwyer, J. Nature 1989, 339, 174.

Chu, C. T. W.; Chang, C. D. J. Catal. 1984, 86, 297.

Chang, C. D.; Chu, C. T. W.; Socha, R. F. J. Catal. 1984, 86, 289.
Kaeding, W. W.; Butter, S. A. J. Catal. 1980, 61, 155.

Chapter 2
Method Development and Experimental Results
2.1. Focus of Research

Our interests in the mechanistic chemistry of the Pearson reaction
arose from the possibility of transforming methanol to hydrocarbons in a
solution-phase process which is analogous to the MTG conversion in zeolite
but occurs at a lower temperature, i.e.< 200°C. The typical temperature of
the "Mobil process" in which H-ZSM-S catalyzes the dehydration of
methanol or dimethyl ether into hydrocarbons is BOO-400°C. The low
methane production and the overall similarity of the product mixtures
between the Pearson and Mobil processes are key to our assessment of the
Pearson reaction as a homogeneous model for H-ZSM-S zeolite-catalyzed

methanol conversion.

The original Jackson26 approach to the mechanistic investigation of
the Pearson reaction employed a closed system, in which a mixture of
methanol or trimethyl phosphate (TMP) with P205 or PPA was placed in a

thick-walled NMR tube, degassed, and sealed under vacuum. The NMR
tube was heated at 200°C for up to two hours. 1H NMR spectra of the

24

25
gaseous products were recorded in situ. 1H NMR spectra for the solution

were also obtained by gently warming the mixture. DME and ethylene

were detected in the gas phase, but no methane formation was recorded.

The ground work for this investigation was done in the Jackson
group and resulted in the proposed major role for CO in MTG chemistry,
as outlined above. In the present study, we have investigated the role of
CO in the PPA-catalyzed conversion of MTG. The reaction was studied in
depth focusing on the microscopic mechanism of formation of initial C-C
bond. A new technique was developed to study this reaction focusing in the
time evolution of the products and the effect of additives on the onset of
reaction. This investigation was designed to firmly establish or eliminate

the role of CO in the methanol to hydrocarbons process.

Ab initio and semiempirical molecular orbital calculations have been
carried out to evaluate segments of the proposed mechanism, especially the
role of CO and some of the suggested ylide intermediates. The
thermochemistry of the intermediates along the proposed CO-catalyzed
pathway from methanol to hydrocarbons has been examined by ab initio
molecular orbital methods, which model gas-phase chemistry. effects of
salvation have been explored at the semiempin'cal PM3-SM2 level and the
resulting corrections applied to the ab initio calculated heats of formation.
(unpublished results from studies by Jackson, J. E.; Garrett, S.; and
Bowman, J.). The overall conclusion from these efforts is that the
oxonium ylide scheme is energetically inaccessible, both in the gas phase

and in the polar PPA medium

26
2.2. Method Development

2.2.]. Standard Materials

Polyphosphoric acid (PPA), trimethyl phosphate (TMP), 1,2,4,5-
tetrachlorobenzene, methyl formate, methyl acetate, methyl prepionate,
carbon monoxide, propene, n-butane, isobutane, and isopentane were
purchased from Aldrich Chemical Company. Acetic acid, pivalic acid
(2,2-dimethyl propanoic acid), propanoic acid, 2-propanol, and t-butanol
(2-methyl-2-propanol) were obtained from Mallinckrodt Chemical
Company. Ethylene was obtained from Matheson Gas Products, Inc.
Acetic anhydride was obtained from Fisher Scientific. Acetone-d6 was
purchased from Cambridge Isotopes. All chemicals were used without

further purification.

2.2.2. Nuclear Magnetic Resonance

1H NMR spectra were recorded at 300 MHz using a Varian Gemini
300 fourier transform spectrometer and referenced to the residual proton
resonance in deuterated acetone at 8 2.04 ppm. As an internal quantitation

standard, 0.01 M 1,2,4,5-tetrachlorobenzene was included in all samples.

2.2.3. The Reactor Design

The reactor was designed to allow a time course study of
methanol to hydrocarbon conversion using PPA, a strong liquid acid, as the

catalyst. The reactor (Figure 2.1) has a 6 inch long tube (reaction cell)

27
with an inlet for N2 placed so that gas flow passes directly through the

reaction mixture, which is also agitated with a magnetic stirring star. On
the top of the reaction tube is a water condenser connected to a glass joint
capped with a rubber septum through which TMP and other liquid
additives are injected via a long needle syringe. The other outlet of this
joint is a tube that goes to a syringe-needle to allow injection to a septum-
capped, vented NMR tube. The NMR tube is cooled in a dry ice/acetone
bath in order to trap low boiling volatile products. The reaction tube is
heated to the appropriate reaction temperatures in a magnetically stirred
silicon oil bath thermostatically controlled by a temperature regulator
(temperature accuracy is i 2°C).

2.2.4. Method Used

PPA is a highly viscous, moisture sensitive and reactive material. In
order to allow for easy handling of PPA and to minimize contamination
with water, the following procedure was adopted and used throughout the

course of this investigation, unless otherwise noted.

The apparatus used for each catalytic run consisted of a reaction cell
and gas capture device, i.e., NMR tube, and is shown in Figure 2.1. Two
identical apparatuses were used simultaneously. One contained PPA, TMP
and a gas or liquid additive. The second was used as a control and
contained only PPA and TMP (the reference reaction). Each catalytic run
was repeated a total of two to three times with a parallel reference reaction

unless otherwise mentioned.

28

.0 Rubber Septum

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

,x' :’ """""""""""""""" Stopcock
Condenser E J \l ----- Fl ow Meter
§§~~‘~{ ‘
K —II. ....... N OCdIC
H20 1 11:.
\» _l _| ............ NM R Tube
.1 J 1].“
N29. _ J| ...... Dry ice / Acetone
\‘Acetone-dg
I
1 a
j “s, ITemp. Controller I
on Bath Magnetic *
Stirbar

Figure 2.1. The reactor design.

29
The reaction apparatus was dried at 100°C in an oven overnight and

set up and used immediately. PPA (3.2 ml; 6.5 :1: 0.1 g) was quickly
transferred to the reaction cell containing a magnetic stir star. The
reaction cell was placed in a constant temperature oil bath set at the desired
temperature and flushed with N2 gas, followed by setting the N2 flow rate
to approximately 7.3 to 7.8 ml per minute. TMP (0.80 ml; 6.8 x 10'3
moles) and other liquid additives were added all at once via a gas tight
syringe giving a total volume of approximately 4 ml in the reaction cell.
The start of each run was recorded at the addition of TMP I (liquid
additive if any). Liquid additives were introduced by along needle syringe
to the top of the reaction mixture, from the top of the water cooled
condenser. Gas phase additives were introduced by injection of a measured

volume into the N2 gas flow which passes directly through the hot reaction

mixture.

For each run gases evolved were trapped in a 5 mm NMR tube
containing acetone-d6 (0.8 ml). The NMR tube was removed, capped, and
replaced by another NMR tube at 5 minutes, 10 minutes, 15 minutes, 20
minutes, 40 minutes, and 60 minutes elapsed reaction time. The
temperature of the reaction mixture was kept constant over the course of
each experiment at the reaction temperature studied. Specific details
including reaction temperatures, additives introduced and products detected

will be given in the descriptions of individual experiments.
2.2.5. Reproducibility of The Reaction

The reproducibility of the reaction using the above methods was

checked by running a sequence of experiments at the same conditions for

30
the same time. Three experiments were done in a series at 240°C with

80:20 PPA:TMP ratio). All experiments gave approximately similar
results with little variation in the product concentrations. The 180°C
temperature was chosen in the studies of the additives because the products
appeared after 20 minutes, thus allowing us to follow the onset of products
with time. The reference reaction was the unmodified reaction of TMP
with PPA in 20:80 ratio. For each reaction presented in this work a

reference reaction was run for comparison.

2.3. Experimental Results

2.3.1. Products of Reaction of PPA with TMP

The distribution of the products from the reaction of PPA and TMP
(80:20 vzv) ratio is shown in Table 2.1. The 1H NMR spectra of trapped
products from the reaction shows that isobutane (i-C4) and isopentane (i-
C5) are the first products observed after dimethyl ether over the time
studied (0-60 minutes) at various temperatures. Isobutylene and ethylene
were then recorded at least 10 minutes after the appearance of isobutane
and is0pentane at higher temperatures (>200°C) (Table 2.2). Figure 2.2
shows a sample 1H NMR spectrum of the initial product mixture (isobutane
and isopentane) in d6-acetone along with spectra of the authentic pure
compounds. Figure 2.3 shows the 1H NMR spectrum for the reaction
products at 200°C after 30 minutes using an 80:20 PPA:TMP ratio. The

length of time between initial formation of DME and formation of

31
isobutane and isopentane was dependent upon reaction temperature, i.e.,

higher temperatures resulted in shorter times.

2.3.2. Proton Decoupling Experiment

Although the 1H NMR spectra for all products trapped in our
reaction were compared with spectra for the pure materials in d-acetone,
the identification of the products was also confirmed by proton decoupling
experiments at 500 MHz on the product mixture (10-30 min.samp1e) from
a reaction of TMP with PPA (20:80 ratio) at 240°C.

a) Irradiation at 0.8 - 0.87 ppm region results in collapse of a group
of peaks at 1.2 ppm (which is assigned for the methylene group in
isopentane) and a change of the multiplicity at 1.6 ppm (which is assigned
to the methine proton in isobutane). The region at 1.6 ppm showed
overlapping resonances with singlet and triplet multiplicities as expected
from the above assignment. This means that there are two overlapping
proton resonances at 1.6 ppm; these are attributed to the methine proton of
isobutane and the methyl groups of isobutylene. The latter compound's

terminal methylene protons are also seen at 4.6 ppm.

b) Irradiation at 1.6 ppm results in collapse of the group of peaks at
0.8 ppm (which is assigned to the methyl groups in isobutane and
isopentane) to became a singlet and a doublet and no other peaks were
affected. This indicates that there is coupling between these groups of
protons. The above proton decoupling results confirm the assignments of
the 1H NMR peaks for the products trapped.

32

 

 

 

time I min.
temp.
°C 5 10 15 20 40 60
180 DME DME DME DME DME DME
1 'C4H10 1 'C4H10
1 ' C51112 1 ’ C51'112
200 DME DME DME DME DME DME
1'C4Hm 1 'C4H10 14341110 1'C4H10 1 ’C4Hlo
1 -C5Hrz 1- C5H12 l - Cerz 1- (351112 1 ’ C51112
1 - C4H8 i ' C4118
QH4 C2H4
220 DME DME DME DME DME DME
1 “(341110 .1‘C4H10 1 'C4H10 .1 'C4H10 1 'Cal‘lro
i'(351112 1'(351‘112 1'C5Hiz 1’C5H12 l-C51‘112
l ' C4118 l " C4H8
C2H4 C2H4
240 DME DME DME DME DME DME
1'C4H10 1 'C4H10 1'C4H10 ,1'C4H10 1'C4H10 1 ’C4Hlo
1 “C4112 1‘ C5111: 1 ' C351112 1' CSHIZ 1 ' Can 1 ' C5111;
i-C4Hs i-C4Hs 1'C4Hs i-C4H8 1'C4H8
C2H4 C2H4 C2H4 C2H4 C2H4
DME : dimethyl ether
1 'C4H10 : Isobutane
1 ‘CSH12 : Isopentanc
i - C4113 : Isobutylene
QH4 : Ethylene

Table 2.1. Qualitative summary of products distributions over time

course at 240°C.

4:3

 

 

 

 

 

 

 

P9:

rh'ei'r'a'rb'ei '15'1'0'1': '1': '1'e'e'euu'o'.

‘l'hefirstprodueuhurnruetioe
ofPPAnthrqpedinM
(Isobutane-Idiom).

 

 

1.

Figure 2.2. 1H NMR for the primary products (isobutane and isopentane)
from the reaction of PPA with TMP after 30 minutes at 200°C (80:20 PPA
to TMP ratio) compared to a mixture of their pure material.

.32: divas—m 85% «SEE
coca Decca .< .8039:— eouoaa 05 .8. Baboon... :22 =— .n.« 8.50..“

 

 

 

 

. . 9.. . . . . . .
s... .1.. ... E w... a. c .F .._ ._....
. .1
bI—«rbbrphpppm.bLPhbb>bpr|Pl>bbnhpbbhhppbhhppbbLbbPPlePhth>>r>lpbbblhI-N
3 C
. U in— M I
m. // ..s
on C ./ ..n w
W: 1 111. w
a .— .nmmrummmm . .
h
0600 P-
. mm.” m 823... u.
o 82:33 :36» n:
85035055 "0
3-2.303 5 ON: u..—
o—Yucofiot. "E
89:38. 5 «3.8. a... 6.. e. :0. 5
«0.2.10. 6
no-2: «ID- "n
«0.2: «3:9. ‘5 on: c. 29.5? ”<

35
2.3.3. Effects of Changing TMP Concentration

Varying quantities of methylating agent were used to examine the
effect of TMP concentration on the reaction products. Different ratios of
TMP to PPA were used; (2:98), (5:95), (10:90) [or 2%, 5%, and 10% by
volume]. At 180°C, neither ethylene nor hydrocarbons (i-C4 and i-C5),
appeared although DME and TMP were observed. When those
experiments were repeated at higher temperature (220°C), i-C4 and i-C5
along with DME and TMP were observed after 15 minutes and increased in
concentration over time. But importantly, no ethylene was observed at
these low concentrations (2%, 5%, and 10%) of TMP. At a temperature of

220°C and TMPzPPA ratio of 20:80, ethylene was observed after 15
minutes, well after i-C4 and i-C5 had appeared (5 minutes).

2.3.4. Effect of Additives

In the course of additive studies with isobutylene, pivalic acid, and t-
butanol, no reference reaction was used because the parallel run was the
reaction of the PPA with additive but without TMP. Therefore, an average
of four reference reactions for other experiments was calculated and used

as a reference for those experiments unless otherwise noted.

2.3.4.1. Addition of Methyl Formate

The Jackson mechanism of Scheme 10 in Chapter 1 suggests that
addition of CO should accelerate the onset of hydrocarbon formation.
Methyl formate (MP) is known to decompose under acidic conditions to

carbon monoxide and methanol; thus the addition of MP to the reaction
mixture should result in the earlier formation of i-C4 and i-C5. When a

36
(200:1) ratio of (TMPzMF) was added, the overall reaction rate at 180°C

increased dramatically (Figure 2.4). The usual i-C4 and i-C5 hydrocarbons
were observed after 20 minutes, but with the addition of MP i-C4 and i-C5

appeared immediately after five minutes. Their concentrations increased

rapidly with time as well.

2.3.4.2. Addition of Isobutylene

Ratios (100:1); 1% (1.54 :1: 0.05 ml gas volume) and (200:1); 0.5%
(0.77 :l: 0.05 ml gas volume) of isobutylene to TMP were added to the
reaction mixture at 180°C. Isobutane and isopentane appeared immediately
after 5 minutes and continued rising with time (0-1h). Figure 2.4 shows
the effect of the addition of 1% (6.7 x 10-5 moles) isobutylene in the time
of i-C4+i-C5 formation compared to the reference reaction at 180°C. For
this reaction, the total amount of isobutane formed after 60 minutes is 1.7 x
10'5 moles, while the total amount of iSOpentane formed is ca. 8.8 x 10‘6
moles. The ratio of i-C4li-C5 is equal to 2.02. Total number of carbons
converted to i-C4 + i-Cs and isobutylene compared to the total number of
carbons added (isobutylene and TMP) is equal to 1.86 x 104.

2.3.4.3. Addition of t-Butanol

The effects of alcohol addition on PPA-catalyzed methanol to
hydrocarbon conversion at 180°C were examined. A ratio of (100:1); 1%
(5.00 i 0.02 mg; 6.8 x 10-5 moles) TMPzt-butanol was used. Results from
the 100:1 ratio are shown in Figure 2.5. It shows that t-butanol accelerate
the conversion of TMP to isobutane, isopentane, and isobutylene within 1

hour. Isobutane appeared at the first five minute of on-stream reaction

37

Effect of Additives on Product Onset Times

1.000-5

 

8.0096 '

Moles Trapped

 

 

 

80

 

Figure 2.4. The effects of addition of isobutylene (A) and methyl formate
(B) on the product onset (isobutane + is0pentane) at 180°C using 200:1
TMP to the additive. Curve (C) represents the behavior of TMP alone.

38
time. The total amount of i-C4+i-C5 formed after 60 min. is 2.187 x 10-5

moles. The ratio of i-C4/i-C5 is vary from 3.3 to 2.2.

2.3.4.4. Addition of 2-Propanol

A ratio of (200:1) TMPz2-propanol (2.05 i 0.02 mg of 2-propanol,
3.42 x 10'5 moles) was used at 180°C. The first products observed in 1H
NMR were propene and isobutane (4.52 x 10‘6 moles from 0-15 minutes
compared to the reference reaction run parallel to it at the same time).
The propene yield reached a maximum (1.16 x 10:5 moles, ca. 33%) at the
first five minutes, fell after that, and disappeared after 15 min. However,
the isobutane yield increased with time. Isopentane was not among the first
products formed, but we observed it in the period of 15-20 minutes, with a
ratio 2.2:] i-C4zi-C5. Its concentration increased with time (Figure 2.6).

2.3.4.5. Addition of Pivalic Acid (2,2-Dimethyl Propanoic
Acid)

Ratios of (100:1); 1% (6.94 :l: 0.02 mg; 6.8 x 10‘5 moles ) and
(200:1); 0.5% (3.49 i 0.02 mg; 3.4 x 10-5 moles) TMPzpivalic acid
(wt.:wt.) were used at 180°C. Isobutane, isopentane, and isobutylene were
observed at the first minutes of the reaction and their concentrations
increased with time (0-60 minutes) as will be shown in the discussion
chapter. Figure 2.7 shows a plot of the formation of i-C4 + i-Cs from
addition of pivalic acid. The 1% ratio produced more products than did
the 0.5%.

39

 

 

4.009—5
—D— sum 04405 from addition of t-butanol
'8 1 —0— sum 044-05 from rat.
& 3.000-5 -
E-t
In
D
3. 2.00e-5 -
D
e-
9
ca
.2
g 1.009-5 -1
0.00e+O r . 1

 

 

 

o 20 40 so so
Time l min.

Figure 2.5. The effect of adding t-butanol to the reaction of PPA with
TMP using (100:1) TMP to t-butanol ratio at 180°C.

 

5.009-5
—0— sum propene
—I-— cum O4+CS Ref
4.0095 . —I— cum 044-05
.5 -—-A— sum 04
0
D.
3'
1.. 3009-5 ~
[-1
m i
.2
E 2.009-5 -

 

 

 

 

 

80

Time I min.

Figure 2.6. The effect of adding 2-propanol to the reaction PPA with
TMP compared to the reference reaction. Ratio of 200:1 of TMP to 2-
propanol at 180°C. 33% of 2-propanol was converted to propene.

41

2.3.4.6. Addition of Acetic Acid

Ratios of trimethyl phosphatezacetic acid were used, (100:1; 12.00 i
0.02 mg acetic acid), and (200:1; 6.10 :l: 0.02 mg acetic acid) at 180°C.

Similar results were obtained from all these reactions, without formation
of i-C4 and i-C5 or other hydrocarbons. The only demonstrated product

was methyl acetate, suggesting that all of the "+CH3" groups are consumed
in the methylation of acetic acid at a rate faster than the decomposition of
acetic acid to give CH3CO+ or ketene. Where 2.055 x 10'4 moles of acetic
acid was used (100:1 ratio), 2.059 x 10'4 moles of methyl acetate were
formed (100% conversion) (Table 2.2). From these results, all the acetic
acid was methylated yielding methyl acetate.

2.3.4.7. Addition of Formic Acid, Propanoic Acid and
Acetic Anhydride

Formic acid (FA) was added in a ratio of 200:1 (TMPzFA) at 180°
and 200°C. It was expected that the FA would turn on the reaction faster,
but it did not. The "+CH3" groups were consumed with no formation of
hydrocarbons. Addition of propanoic acid and acetic anhydride under the
same reaction condition (180 and 200°C) gave results similar to those of the
additions of acetic and formic acid. Propanoic acid gave methyl propanoate

and acetic anhydride gave methyl acetate.

42

 

4.000-5
—I— sum 04405 from pivalic acid
—9— sum C4+CS from ref.

 

Moles Trapped

 

 

 

 

80

Time I min.

Figure 2.7. The effect of addition of pivalic acid to the reaction of PPA
with T'MP at 180°C using 200:1 TMP to pivalic acid ratio.

 

0 — 5
5 — 10
10 — 15
15 —20
20 —40
4O —60
Total number of moles

% conversion

 

 

43

0

1.13 x 10‘5
2.78 x 10‘5

5.3 x 10‘5

1.3 x 10'4
2.059 x 10‘4
2.059 x 104

100

 

Time(Min.) Cumulative number Number of moles of
of moles of methyl isobutane +
acetate iso . entane

OOOOOOO

 

Table 2.2. The conversion of acetic acid to methyl acetate at 180°C
(2.055 x 10“1 moles of acetic acid was used).

 

44

2.3.4.8. Addition of CO

A series of experiments were run using different concentrations of
carbon monoxide. Carbon monoxide is proposed to be the active catalyst
in the early phases of the reaction. So, one would expect that adding it to
the reaction mixture would enhance the reaction rate and/or increase the
reaction product yields. Various ratios of carbon monoxide to TMP
concentration were used; they are summarized in table 2.3. The results
from these experiments are shown in figures 2.8 - 2.12). Figures 2.8 and
2.9 show the difference in the results from two similar reaction (addition

of CO using 600:] TMP:CO ratio to the reaction mixture.

Table 2.3. Volumes of CO added.

 

 

TMP:CO Volumes of added
Ratio COIml
Continuos flow -
100:1 1.53 :t 0.02
200:1 0.76 i 0.02
400:1 0.38 i 0.02
600:1 0.19 i 0.02

 

 

 

800:1 0.09 :1: 0.02

 

45

 

 

 

5.000-5
+ sum 044-05 600:1
‘ —0— sum (0) 600:1
4909.5 ., —'l— sum C4+CS Rel.(no 00)
'8 --0— sum (0) Ref. (no 00)
a d
a.
a 3 000-5 -
[.4 .
m
.2
Q
2 2009-5 -
l
1.000-5 -
0.00e+O - . . . -
0 2 0 6 0 8 0

 

4o
Time/Min.

Figure 2.8. Number of moles of product and moles of carbon (C)
trapped from reaction of PPA with TMP with CO added in a ratio of 600:1
TMPzCO at 180°C.

 

 

 

 

 

 

coco-s
—I— sum C4+CS
—o-— sum(C)
5-009-5‘ —e— sumC4+CS Ref
—-0— sum(C) Hot
'1:
8,, coca-5-
a.
a
h
[-1
m 3009-5-
0
1'5 d
2 2009-5-
1.009-54
0.009+0 , ~ . - '. .
o 20 40 so so

Time/Min.

Figure 2.9. Number of moles of product and of carbon (C) trapped from
reaction of PPA with TMP with CO added in a ratio of 600:1 TMP:CO at
180°C.

47

 

6.000-5
‘ —I— sumC4+C$100z1
—o— sum(C) 100:1
5'°°°'5" —'I- cum 64405 Ref
1 —-0— sum (C) rot

Moles Trapped

 

 

 

Time/Min.

Figure 2.10. Number of moles of product and of carbon (C) trapped
from reaction of PPA with TMP with CO added in a ratio of 100:1
TMPzCO at 180°C.

 

48

 

 

 

 

 

4.000-5
—I— sumC4+C$ 800:1
—'O— sum(C) 800:1
—I— sumC4+05 Rot
3.009-5 - —+— sum(C) Rot
E
2'
s.
1" 2.0095 4
m
.2
g .
1.009-5 -
q
0.00e+O —~ ~ n I -
O 20 40 6 0 80
Time/Min.

Figure 2.11. Number of moles of product and of carbon (C) trapped
from reaction of PPA with TMP with CO added in a ratio of 800:1
TMP:CO at 180°C.

49

 

8.009-4
—I- sumC4+CS 600:1 at 200°C

. —O'— sum(C)600:1at200°C
—l— sum04+05 Rot at 200°C
6.009-4 - —°— sum(C) 1391912001

 

 

 

 

'6
0
a d
a.
a
h
1" 4.009-4 -
U}
.2
:3
2.0094 -
A
0.009+0 , r 1' . , - , T
0 20 40 so

Time/Min.

Figure 2.12. Number of moles of product and of carbon (C) trapped
from reaction of PPA with M with CO added in a ratio of 600:1

TMPzCO at 200°C.

Chapter 3

Analysis and Discussion

3.1. Introduction

Although many groups have observed CO as a product in the MTG
process112’394’596 Jackson and Bertsch7 were first to postulate a
mechanistically important role for CO in the overall process when PPA is
used. Their postulated mechanism addressed the initial C-C bond
formation in the MTG conversion in zeolites and in PPA. The similarity
of the product mixtures observed in both processes and the absence of
methane were the keys to choosing the homogeneous and more easily
investigated reaction. The relatively low temperatures used in the Pearson
reaction (using PPA) were easily accessible. In such strong acidic media
the formation of the CC bond is thought to involve an electrophilic
methylating agent and a nucleophilic carbon species. The electrophilic
methyl groups are available in protonated methanol, protonated dimethyl
ether (DME) or methyl phosphates. However, the question that arose was
the nature of the carbon nucleophile involved. In Olah's mechanismB, the
nucle0phile is the methylene oxonium ylide while in the Jackson postulate it
is ketene. The choice of oxonium ylide as nucleophile became the
predominant mechanistic candidate when the free radical mechanism was

ruled out 9’10.

50

5 1
3.2. Key Observations

There are certain key observations pertaining to MTG conversion
catalyzed by H-ZSM-S or PPA that support the mechanisms proposed to
date:

1) Most of the research groups using mass spectromeu'y11’12’13,
infrared spectroscopyl4v15, and NMR studieleé, established that methanol
is first dehydrated to DME and that the equilibrium mixture of methanol
and DME is then converted to hydrocarbons.

2) Historically, ethylene has been identified as a primary product as
discussed in chapter one. Isobutane and isopentane were also observed as
the primary products by Kolboe.” Another C4 species, trans-2-butene,
was reported by Klinowski and Sulikowski to be the first product.18

3) Carbon monoxide had been noted among the first products in the
MTG process over H-ZSM-S by many research groups. The CO was
observed subsequent to the formation of DME but prior to the onset of
hydrocarbon formation. Furthermore, at high temperatures the
concentrations of CO began to decrease as hydrocarbons were formed.
The initial CO generation in methanol to hydrocarbon process thus still
remained a question. Chang and Silvestri3 had earlier established that in
the Mobil process zeolites oxidize formaldehyde or its equivalents to
produce CO. Thus, CO presumably forms by oxidation of methanol
during an initial induction period, perhaps by protons or methyloxonium
ions which function as methyl cation equivalents.2

4 ) Adsorption of t-butanol (CH3)3COH over H-ZSM-S zeolite was
studied by Aronson19 using 13C NMR. Aronson reported that the adsorbed

52
species could be bound as a silyl ether, with the alkyl group covalently

bonded to the zeolite framework oxygen as shown below.

The reaction is assumed to proceed by transfer of a proton from the zeolite
to the alcohol to form an oxonium ion. The oxonium ion in turn
dehydrates to form a silyl ether intermediate [-1. Aronson further

suggested that isobutylene is formed as shown in equation (1).

 

 

CH3 CH3 CH3
13' 13' _ 113
H3C-(f-CH3 H3C—g— CH3 H2C""C"'CH3 (1)
02 “OZ +
1.2 1.3 1102
Z= Zeolite

No peaks were observed in the 13C NMR spectra that could be assigned to
carbenium ions I-3. However, their results are consistent with an
equilibrium between a carbenium ion and a silyl ether. According to their

energy diagram, the alkylsilyl ether L2 is lower in energy than carbenium
ion [-3 by 4 kcal/mol.

53
5) Adsorption of propene over H-ZSM-S zeolite was studied by Haw

and co-workers.20 Haw used 13C CP—MAS and assigned a peak at 250 ppm
to the isopropyl cation [-4 formed on the acidic sites of the zeolite
immediately after adsorption of propene. The iSOpropyl cation further
reacts with additional propene molecules to yield long chain carbocations

[-5 as shown in equation (2).

CH2=CH-CH3 CH¢=CH-CH3

+ +

H CH3-(fH-CH3 CH3-$H-CH2-CH(CH3)2

. ———» ——» 1

O O O /

>Si’+\A1/ >si’+\A1/ >Si’+\A1 (2)
| | \ l l \ | l \
L4 [-5

A rapid exchange of free alkyl cation could follow resulting in an
equilibrium between the ethyl isopropyl carbenium ion [-6 and the methyl
isobutyl carbenium ion [-7 as shown in equation (3). It is expected that I-
6 would rearrange to form the tertiary cation 1-8.

 

 

CH3 H CH3H

I I I I «1-
Hsc-t-i-t-Cns ~ A Wet-ems

H H H H H H

[-6 [-7

54
The Aronson mechanism involves a silyl ether as an intermediate and the

Haw mechanism involves a carbenium ion as an intermediate. Which

intermediate is consistent with the results of this investigation?

6) According to the original Jackson work, thermodynamic data
support the CO- catalyzed pathway. The methylation and protonation steps
in this mechanism are well within thermodynamic range. In contrast, the
oxonium ylide mechanism requires a deprotonation step which would be

extremely unlikely1 due to its very high endothermicity.

7) The observations of this investigation can be summed up as

follows:

a) The experimental data of this investigation point to one major
result: At temperatures between 180°C and 240°C isobutane
and isopentane were the hydrocarbon products recorded after
DME (see Figure 2.2 in Chapter 2). Isobutylene and ethylene
were then recorded at least 10 minutes after the appearance of
isobutane and isopentane at higher temperatures (>200°C) (see
Table 2.1 in Chapter 2). A small amount of propene is also

seen at large reaction times.

b) The addition of a CO precursor (Methyl formate), "turned on"
the reaction earlier and isobutane C4H10 (i-C4) and isopentane
C5H12 (i-Cs) hydrocarbons were observed after 10 minutes.

c) The addition of isobutylene turned the reaction on earlier or in

a shorter period of time than before, (i.e., after 10 minutes).

55
d) The addition of pivalic acid and t-butanol also accelerated the

reaction rate and products were observed within the first five
minutes.
e) The addition of 2-propanol yielded propene and i-C4 and i-C5

as well.

3.3. Modified CO-Catalyzed Mechanism

According to our results: i-C4 and i-C5 are the first products and the
addition of methyl formate accelerates the reactions onset compared to the
reference reaction. Furthermore, the additions of t-butanol and pivalic
acid also accelerate the reactions onset (see later) when compared to the
reference reaction. With these new results in mind, it appears that the
Jackson mechanism, as originally proposed, does not account for the
specific generation of i-C4 and i-Cs alkanes. In their work, Jackson and
Bertsch investigated the same reaction using the same catalyst and
methylating agent, but they did not undertake a time course study as we did
in this work. They heated the mixture at temperatures up to 200°C and
observed ethylene among the gaseous products. However, in the present
work the time evolution for this process was recorded. Ethylene was
detected, but not as the first product. The Jackson mechanism has
therefore been modified to rationalize the new results, and is shown in
Scheme (1).

56

Scheme 1

 

1rxo+—CH3 + co : H3c—CEo"

nH+ 11

7

H2C=C=O

”CH3+”

7

~ +
H”. CO . H2C=CH2 S 2:. CH3CH2 —C-'..'_-'0

Ethylene
“and"!
CH3CH=C=O
rnCHjO-n
_ +
11*. co, Hzc=CH2CH3 1:; ‘——. (CflshCH—C=O
Propene
«Him
(113C)2C=C=0

"CH3+n

T

+
(113C)2C: CH2: C0 + (CH3)3C+ S (CH3)3C—CEO
Isobutylene
H” + CO

x. Y =11, CH3

«CH34-n
.. .. } CH3OH, CH30CH3, Si(OR)3, AI’(OR)3. P(OR)3(O)

57
3.4. Ethylene is not The First Product

Ethylene is much less reactive than propene21; if ethylene was the
first product, it would have shown up under our trapping conditions
because it is more volatile than either i-C4 or i-C5. Trapping efficiency
for ethylene and isobutylene was found to be 80-82%. Surprisingly, we
got CnHzMz products (saturated) rather than the CnHzn (unsaturated)
products expected, based on the simple dehydration stochiometry of

methanol.

So in our case, if C2H4 is formed and acts as the active intermediate
(regardless of how it is formed), C3 or more saturated and unsaturated
hydrocarbons would be expected among the first products. Propene,
propane, trans-2-butene or l-butene could be formed and C4 hydrocarbons
as well (Scheme 2) and would be detected in our NMR samples of trapped
products. As mentioned in Chapter one, Dessau”,23 found a similarity in
the product distribution (C3+ olefins and ethylene) between the methanol
conversion to hydrocarbons and from heptene cracking (Scheme 3) under
comparable conditions (400°C and 1-2.5 Torr in N2) using zeolite H-ZSM-
5. He observed significant ethylene formation from heptene cracking,
along with C3+ olefins, when operated at low pressure and long contact
time. He reported that the ethylene is not the initial observable olefin
product from methanol, but is formed by secondary reequilibration of the
primary olefins propene and butene. So, we think that the ethylene and
propene in our reaction are formed via cracking of the larger

hydrocarbons produced as shown in Scheme 4.

58

Scheme 2
H
" VH3" + H2C"-=CHz —’ H3C’S‘CH3 —_’ H3(:/'§CH2
II-l
Q’Hggc + CH3

A")"*c113 + H3CQCH ——-> H3C\/\CH3-D->

“9:01,

H2C\
\ACH 3 H3C % CH3
l-butene trans-Z-butene 3isobutylene
"-2 “-3 [1.4

Scheme 3

/> \

“3‘5
/W\
ch + CH; ch’Jr‘cng + H >=CH2

1-. 1+1“

59

Scheme 4

H3C CH3 Hac H3C CH
>+ \ H2C# _’ _> H3CWCH3

I'13C CH3 CH3 4-

H3, 0,1 _

\ + H C=CH

CH3 H 3C CH3

CH3
+
+ ___>
H3C 4|>013 Hzc H3§MCH3

C113
rv-1

H3C H3C CH3 CH3

CH3
IV-I + Hzc=< * W
CH3 H3C + CH3

H3C
rv-z

 

H3C
1v-1 —» _._. Non, + H3c¢\CH3
H3C

H3C CH3

IV-Z —> —-> -—'> H3CMCH3 + H3C¢\CH3

H3C

3.5. Hydride Source

Methanol or DME are good sources of hydride, as shown below,
because oxygen strongly stabilizes the neighboring carbocation by
resonance (Scheme 5). As appears from the Scheme below, CO could then
be formed by further oxidation of formaldehyde, a pathway already
established in the zeolite context by Chang and Silvestrill.

 

 

 

Scheme 5
CH3 H CH3 H
W* X ‘“ Bx "' (I? X
-. .° 3 3 8
x69 H
- : BIC: —> -—-> C0 + H2

X: H, CH3

3.6. Effect of Additives

Below are detailed further the effects of additives on the timing and
composition of the PPA-catalyzed product mixture, and how the new
mechanism explains these results. The original mechanism postulated that
ethylene or propene were the reactive products. But in the course of this
study it became clear that neither ethylene nor propene is among the first

products. The detection of isobutane and isopentane led us to revise the

61
original mechanism to continue methylation of the acetyl cation (CH3-CO+)

all the way to the pivaloyl((CH3)3-C-CO+) cation and ultimately to the ‘
formation of the t-butyl cation, isobutylene, and isobutane. In fact, in the
strongly acidic medium, the t-butyl cation and isobutylene should be in
equilibrium as shown in equation (4).

 

 

G) = >
“3‘3 CH3 H3C cu, + Hi (4)
t-Butyl cation Isobutylene

As mentioned in Chapter two the reference reactions of the additive
studies are an average of four different reference reactions for other
experiments. The parallel reactions in the course of the additive studies

were the reaction of PPA with the additive in the absence of TMP.
3.6.1. Addition of Isobutylene

As noted earlier, the addition of isobutylene to the reaction mixture
accelerated the reaction when compared to the reference reaction (Figure
3.1). Protonation of isobutylene gives the t-butyl cation which can yield
isobutane via hydride abstraction as shown in path A of Scheme 6.
Methylation of isobutylene gives the t-pentyl cation. The latter could give
three products: 2-methyl-l-butene VI-l and 2-methyl-2-butene VI-2 via
proton loss, and isopentane VI-2 via hydride abstraction, as shown in path
B of Scheme 3. We did not detect either VI-l or VI-2 but did detect
isopentane VI-3. From these results, it can be assumed that the t-pentyl

62
cation (the protonated form of VI-l and VI-2) is strongly favored at

 

equilibrium.
Scheme 6
+ CH3
CH C n u u _"
., x5 1.. )3 Ci 1" H
H3C CH3 H3C CH3 H3C CH3

 

CH2 u +00 1 -
CH3 VI-l
B) G)
/U\ H3 C

 

H3C CH3 CH3 l-I+
_\ CH3
+ H‘ " l
H3C CH 3

CH3 v1.2

£8

H3C CH3
Isopentane
VI-3

3.6.2. Addition of Pivalic Acid and t-Butanol

One could propose that adding the tertiary carbocation (CH3)3C+
would also accelerate the reaction. t-Butanol and pivalic acid are well
known as sources of t-butyl cation in strongly acidic media.24 The
difference between them is that when pivalic acid is dehydrated to the t-
butyl cation, a molecule of CO is also produced.

63

"mo-~- (C) from PPA + TMP + Isobutylene
—O— C4+C$ from PPA + TMP + Isobutylene
"W0"- (C) from Ref. reaction (NO add.)
+ C4+C§ from Rat. reaction (NO add.)

 

 

 

1 .200-4
!
1.000-4 -
1 I."
8.0095 1 ’0'.
an -I f.
O. .'
2
[" 6.009-5 - '3'
U) .0.
.2 .
e
2 4.009-5 - “.0

 

 

80

Time/Min.

Figure 3.1. Addition of isobutylene to the reaction of PPA with TMP
compared to the reference reaction at 180°C using 100:1 TMP to

isobutylene ratio (6.7 x 10‘5 moles of isobutylene).

64
When pivalic acid was added to the PPA reaction mixture at 180°C,

i-C4, i-Cs and isobutylene were detected in the first minutes (Figure 3.2).
Now in this process, the formation of the tertiary carbocation or
isobutylene does not require the presence of a methylating nucleophile
because the formation of (CH3)3C+ can easily occur by dehydration of
pivalic acid to form (CH3)3C-CO+ as shown in Scheme 7.

Scheme7
o
H3C .. H3C + - H3C 4" CH3
OH "HZO c=o CO
H u —> H3C — CH
3C CH3 K. . CH3 3

H t-butyl cation

1

Addition of t-butanol (2-methyl-2-propanol) to the PPA reaction

H3C CH,

mixture in the presence of TMP gave results like those with pivalic acid
(Figure 3.3); isobutane was first produced (in the first five minutes) then
isopentane showed up and their concentrations increased with time. Once

the t-butyl carbocation is formed it can follow one of two pathways (as

65
shown in Scheme 6). It can abstract a hydride from TMP or from an

isobutylene molecule in the absence of TMP.

As proposed in our modified CO mechanism the pivaloyl ((CH3)3C-

CO+) ion readily loses CO to give (CH3)3C+ and form products by two
distinct and separate pathways: .
1) Hydride abstraction to form isobutane or 2) Proton loss to give
isobutylene. The (CH3)3C-CO+ species is proposed in the revised CO
mechanism. So, would the same results and same products be expected
from pivalic acid in PPA in the absence of the methylating agent (TMP)?
The omission of TMP in the reaction mixture yielded results qualitatively
similar to those found with TMP present. Isobutane, isopentane and
isobutylene were again the first detected. Isobutylene decreased with time.
The ratio of isobutane to isopentane (C4/C5) is approximately equal to 2.

Addition of the t-butyl cation resulted in approximately double the
amount of i-C4 than i-C5. Thus the methylation of isobutylene is seemingly
slower than protonation followed by hydride abstraction and the

equilibrium is shifted toward the t-butyl cation as shown in A in Scheme 6.

Another important point that can be drawn from the experiments
without TMP in the reaction mixture that included pivalic acid (1%) (6.8 x
10‘5 moles of pivalic acid) is that the formation of i-C4 and i-Cs decreased
dramatically compared to when a similar number of moles (6.8 x 105) of
t-butanol (1%) is added to the reaction mixture (Figure 3.4 for pivalic acid
and Figure 3.5 for t-butanol). The ratio of total number of carbon atoms
(C) produced by the reaction (in the presence of TMP) to the total number

 

 

 

 

.......... (C) from PPA + TMP + Pivalic acid
—o-— 04405 from PPA + TMP + Pivalic acid
----D--- (C) from Roi. reaction (NO add.)
—I-— 044-05 from Ref. reaction (NO add.)
2.000-4
'5
0 "
a.
:-
1.00e-4 - ......
a o"'
.2 .....
i O
0.00e+0

 

 

 

Time/Min.

80

Figure 3.2. Addition of pivalic acid to the reaction of PPA with TMP
compared to the reference reaction at 180°C 200:1 TMP : pivalic acid ratio
(3.42 x 10'5 moles of pivalic acid).

67
of carbon atoms added (not counting TMP) is equal to 0.49 in the case of

pivalic acid and 0.71 in the case of t-butanol. However, in the absence of
TMP, these ratios were found to be 0.45 in the case of pivalic acid and 0.61
in the case of t-butanol. I think that these results support the formation of
(CH3)3C+ or isobutylene during the PPA-catalyzed methanol conversion.
The ratio C4/C5 in the case of pivalic acid is 1.1-2.5 in the reaction with
TMP, and 21-32 in the reaction without TMP. In the case of t-butanol
this ratio is 2.2-3.6 with WP and 1.4-1.8 without TMP.

The source of hydride in the absence of TMP could be isobutylene,
formed via proton loss from the t-butyl cation. Gas phase thermochemical
data show that the t-butyl cation can abstract a hydride from isobutylene to
yield isobutane and 2-methyl propenyl cation. Isobutylene can also be
readily methylated by trimethyl oxonium ions or other CH3+ donors to
yield a t-pentyl carbocation which can subsequently abstract a hydride to
give isopentane. ISOpentene was not detected as a product. Rapid hydride
transfers can occur between stabilized carbocations (especially tertiary

carbocations) as shown in equation (5).

CH3 H3C CH3 HBC

Ji“ >—\= A WSW

H3C CH3 H3C CH3 CH3 H3C

 

 

>-\CH<

68

-----o----- sum(C) from PPA + TMP + t-butanol
—O—- C4+CS from PPA + TMP + t-butanol

""D'" (C) from Ref. reaction (NO add.)
—I— 04405 from Ref. reaction (NO add.)

 

1.009-4
,0
8.000-5 - V
a. 6.009-5 -
a ,
r-
E-' q
U) ’0'.
2 a". D
g 4.009-5 - ....

 

 

 

 

o 20 40 so so
Time/Min.

Figure 3.3 Addition of t-butanol to the reaction of PPA with TMP
compared to the reference reaction at 180°C 100:1 TMP : t-butanol ratio
(6.8 x 10‘5 moles of t-butanol).

69

sum G4+C§ irom add. of 1% pivalic acid
sum (C) from add. at 1% pivalic acid

isobutylene from add. oi 1% pivalic acid
"mo-~- sum C4465 without TMP

.....,.... sum (C) without TMP

"mom-- isobutylene without TMP

Hi

2.0004

 

 

Moles Trapped

 

 

 

 

 

 

80

Time/Min.

Figure 3.4. Addition of pivalic acid to the reaction of PPA with TMP
compared to the reaction of pivalic with PPA without TMP using 100:1
TMPzpivalic acid ratio (6.8 x 10‘5 moles of pivalic acid) at 180°C.

7O

aumC4+Cs from add. of 1% tobutanol
sum(C) from add. at t-butanol
Isobutylene from add. at 1% t-butanol
.......... sum 044-05 without TMP

.....¢..... sum(C) without TMP

-----n----- Isobutylene without TMP

Hi

 

1.000-4

Moles Trapped

 

 

 

Time/Min.

Figure 3.5. Addition of t-butanol to the reaction of PPA with TMP
compared to the reaction of t-butanol with PPA without TMP using 100:1
TMPzt-butanol ratio (6.8 x 10'5 moles of t-butanol) at 180°C.

 

71
Oxygen also stabilizes a neighboring carbocation by charge

delocalization as shown in Scheme (8). CO generation could occur from

either of two pathways (A) or (B).

 

 

 

Scheme8
Hx
C20 + XY
/
H\ + :r//’ H (I)
,c=ox \ Y "opo H
H - ox = 3
Y H‘C: X: CH3
’ Y
H\
(A) ’C:() ——> ——> C0 +H2
H
H\ ’Ox -H+ 4.. IOX 06X
(B) ,C\ -——-> -C\ <—> H-C\
H Y "R'hi
OX: good L.G.
Y: good L.G.
+ + _
’lox ‘X 00 'Y +
H-C\Y H-C\Y -—> H-CO —> co + H+

72

3.6.3 Addition of 2-propanol

According to the results from the addition of 2-propanol (Figure 2.6
in Chapter 2), propene was formed together with a small amount of
isobutane during the first five minutes. Dehydration of 2-propanol gives
propene under acidic conditions25v26 but nobody to our knowledge has
observed isobutane from the addition of 2-propanol in the methanol to
hydrocarbon process. Dehydration of 2-propanol could follow one of the
pathways shown in Scheme 2 in our reaction. Propene and isobutane but
no isopentane were observed in the product 1H NMR spectra in the first
five minutes. The latter (iSOpentane) appeared after 15 minutes. It might
be expected that the trans-2-butene (II-5) and l-butene (II-4) in Scheme 2

would be formed; these products, however, were not observed.

3.6.4. Addition of Acetic Acid

When acetic acid was added to the standard PPA/TMP reaction, the
formation of volatile methyl acetate (CH3COOCH3) was observed to be

faster than the formation of hydrocarbons. There was no formation of i-
C4 and i-C5. The formation of methyl acetate suggests that many of the
active "CH3“ groups are consumed via methylation of acetic acid at a rate

much faster than the decomposition of acetic acid to give CH3CO+ (see
Table 2.2 of Chapter 2).

73

3.6.5. Addition of CO

Carbon monoxide was added to the reaction mixture in different
ratios ranging from 800:1 to 100:1 as shown in the Figures (2.8 - 2.12 in
Chapter 2 ), and even as a continuous flow in the canier gas. Surprisingly,
its addition did not accelerate the rate of hydrocarbon formation to the
extent that we expected. As shown in Scheme 2, CO is the key species
proposed for the formation of the ketene which carries on the reaction.
Addition of methyl formate which is a good source of CO enhanced both
the reaction rate and product formation. At this moment, we do not have

an explanation as to why CO did not accelerate the reaction.

74

3.7. Future Work

The conversion of methanol to hydrocarbons was studied over
various catalysts by numerous research group as previously stated, but still
the mechanism of the crucial initial C-C bond formation remains
unresolved. As mentioned previously, in many studies ethylene is
considered to be the primary intermediate leading to higher hydrocarbons,
whereas some research groups, including ours, questioned this possibility.
In the present work we intended to investigate the Jackson proposed

mechanism. However the question was not fully answered.

In the new CO-modified Scheme, there are many points that need to
be examined. Explicitly looking for CO and C02 in our system needs to be
done by using other probes such as, IR., G.C., or G.C.M.S. along with
NMR.

For each intermediate in the new proposed mechanism, there is a
way to check its validity by directly adding either the corresponding
species (or immediate precursors) or the products for/from each step as

shown in the following Scheme.

Labeling studies are needed to address the question of the hydride
source. CD3OX (X=H or D) or deuterated trimethyl phosphate should be
used with the addition of either t-butanol or pivalic acid (t-butyl cation
source). *CH3COOX (X=H or R) could be used to see whether the methyl
group is incorporated in the isobutane to form (CH3)3*C-H. Longer-chain
acids R*CH2COOX could be used for the same goal.

75

 

V

Yxo+—CH, + co H3C—Cso“

I co, HCOOH, HCOOCH3 I I «3+»

Hzc=C=O

. «CH +00 7
I Ethylene I 3 I CH3CH2COOH I
+ +

H2C=CH2 .—_—_> 33¢...an 2:.» CH3CI12—CEO
H“ + C0 II

I CH3CH20H I
CH3CH =C=0

Pr H31“. ,
LUBE r |(CH3)2CHCOOH|
+

H+ "' C0 7 n H hfiunfi

CH HOHCH
M (H 3C)2C=C =0
l Isobutylene
. CH3 WI(CH:)3CCOOH I

(1‘13C)2C=CH2 :Hfl "' CH'.i‘---- (CHs)3C—C=0

H" + co
(CH3)3COH *
(CH ) CCOOH

CH3COOH
CH3COOCH3
(CH3CO)2O

   
     

 

 

 

nH-hr

 

 

 

 

 

 

 

76

3.8. Conclusion

Several significant results have come from this work:

1) The previous assumption that ethylene is the first hydrocarbon
product from the Pearson reaction was proven wrong; instead, it is the
saturated hydrocarbons isobutane and is0pentane that are the first volatiles
isolated.

2) The alkenes ethylene, propene (traces), and isobutylene are
produced, but at substantially later times than the first products; they are
most easily interpreted as products from cracking of higher molecular
hydrocarbons as interpreted earlier by Dessau for the Mobil process.

3) The PPA-catalyzed methanol to hydrocarbons reaction has been
routinely observed at the remarkably low temperature of 180°C; in
addition, the existence of a substantial induction period before the
appearance of hydrocarbon products has been verified. The combination
of an induction period followed by a sharp onset of isobutane and
isopentane production strikingly resembles Kolboe's finding in the H-ZSM-
5 catalyzed conversion of dimethyl ether to hydrocarbons.

4) The earlier proposal of Jackson and Bertsch, that CO is the key
catalytic species in the initial C-C bond formation, has now received its
first rigorous tests in the Pearson reaction. It did not fare well. Despite
the promising finding that methyl formate (HCOOCH3) accelerates the
onset of hydrocarbon production, formic acid itself —a CO source in
strong acid— did not appear to accelerate the process. Furthermore, direct
addition of CO at several concentration levels had essentially no effect on

the reaction.

77
5) Finally the observation of saturated hydrocarbons implies that

some substrate must be getting oxidized to provide the extra hydrogen.
The implicit byproducts would be CO and/or C02, Explicit tests for these

products are actively being pursued.

78

 

References

1 Nagy, J., B.; Gilson, J. P.; Derouane, E. G. J. Mol. Catal. 1979, 5,
393.

2 Anderson, M. W.; Klinowski, J. J. Am. Chem. Soc. 1990, 112, 10.

3 Dwyer, J. Nature 1989, 339, 174.

4 Chu, C. T. W.; Chang, C. D. J. Catal. 1984, 86, 297.

5 Chang, C., D.; Chu, C. T. W.; Socha, R. F. J. Catal. 1984, 86, 289.

6 Kaeding, W. W.; Butter, S. A. J. Catal. 1980, 61, 155.

7 Jackson, J. E.; Bertsch, F. M. J. Am. Chem. Soc. 1990, 112, 9085.

8 Olah, G. A.; Surya Prakash, G. K.; Ellis, R. W.; Olah, J. A. J.
Chem. Soc., Chem. Commun. 1986, 9. .

9 Hunter, R.; Hutchings, G. J.; Pickle, W. J. Chem. Soc., Chem.
Commun. 1987, 1369.

10 Hutchings, G. J .; Jansen Van Rensburg, L.; Pickle, W.; Hunter, R. J.
Chem. Soc., Farady Trans. 1 1988, 84, 1311.

11 Chang, C. D.; Silvestri A. J. J. Catal. 1977, 47, 249.

12 Van den Berg, J. P.; Wolthuizen, J. P.; van Hooff, J. H. C. In
Proceedings, 5th International Conference on Zeolites, Rees,
L.V.C., Ed.;Heyden: London, England, 1980 p 649.

13 Dessau, R. M.; LaPierre, R. B. J. Catal. 1982, 78, 136.

14 Salvador, P.; Fripiat, J. J. J. Phys. Chem. 1975, 79, 1842.

15 Salvador, P.; Kladnig, W. J. Chem. Soc., Faraday Trans. 1 1977,
73, 1153.

16 Derouane, E. G.; Nagy, J. B.; Dejaifve, P.; van Hooff, H. C.;
Spekman, B. P.; Vedrine, J. C.; Naccache, C. J. Catal. 1978, 53, 40.

17 Kolboe, S. Acta Chem. Scand. 1982, A 42, 711.

18 Sulikowski, B.; Klinowski, J. Appl. Catal. 1992, 89, 69.

19 Aronson, M. T.; Gorte, R. J.; Farneth, W. E.; David White. J. Am.
Chem. Soc. 1989, 111, 840.

20 Zardkoohi, M.; Haw, J. F.; Lunsford, J. H. J. Am. Chem. Soc. 1987,
109, 5278.

21 Derouane, E. G. J. Catal. 1973, 53, 331.

22 Dessau, R. M. J. Catal. 1986, 99, 111.

23 Dessau, R. M. J. Catal. 1987, 103, 526.

24 Knozinger, H.; Kohne, R. J. Catal. 1966, 5, 264.

25 Mole, T. J. Catal. 1983, 84, 423.

26 Mole, T.; Bett, G. J. Catal. 1983, 84, 435.

HICHIGAN STATE UNIV. LIBRRRIES
lli‘IHillWMW”WNWWINHIIWHIIWNIIIHI
31293014099232