Hydrocarbon From Methanol

CATAL . REV.-SCI. ENG., 25(1).
1-118 (1983)
Hydrocarbons from Methanol

Downloaded By: [2007-2008-2009 Kyunghee University - Suwon Campus] At: 10:27 12 February 2009
CLARENCE D CHANG .
Mobil R e s e a r c h and Development C o r p o r a t i o n
C e n t r a l R e s e a r c h Division
P r i n c e t o n . New Jersey 08540
I . INTRODUCTION .................................... 2
I1 . GENERAL CONSIDERATIONS ....................... 2
A . Stoichiometric C o n s t r a i n t s ...................... 2
B . Thermochemistry ............................... 3
I11 . CATALYSTS ....................................... 4
A . E a r l y O b s e r v a t i o n s (pre-1960) .................. 4
B . Zeolite C a t a l y s t s ............................... 8
C . Nonzeolitic C a t a l y s t s ........................... 35
IV . REACTION KINETICS .............................. 36
V. REACTION MECHANISM ............................ 44
A . E t h e r Formation ................................ 44
B . Hydrocarbon Formation ......................... 48
VI . THE MOBIL METHANOL-TO-GASOLINE
(MTG) PROCESS ................................... 75
.
A Fixed-Bed Process ............................. 75
.
B Fluid-Bed P r o c e s s ............................. 84
........................
C . Product C h a r a c t e r i s t i c s 99
D . Olefin P r o d u c t i o n .............................. 100
VII . DUAL-FUNCTIONAL CATALYSIS ................... 103
VIII . POSTSCRIPT ...................................... 110
REFERENCES ...................................... 111
Copyright 0 1983 by Marcel Dekker. Inc . 0022-2348/83/250 1-0001%3.50/0

2 CHANG
I. INTRODUCTION
The conversion of methanol to hydrocarbons is a remarkable

reaction. The mechanism involves C-C bond formation from C,
fragments generated in the presence of certain acidic catalysts
and reagents. The precise nature of these reactive C1 species
is unknown at present and is the subject of lively debate. The
considerable diversity of current opinion will become apparent

from the following account.
Occasional reports of hydrocarbon formation from methanol,
often as a minor side-reaction accompanying ether formation, can
be found in the literature dating back as far a s 1880. However,
it was not until the early 1970s that the industrial potential of
the reaction emerged, due mainly to the confluence of two events:
the discovery by workers at Mobil of the selective catalytic con-
version of methanol to high octane gasoline over zeolite catalysts
[l, 21, and the 1973 Arab oil embargo. The embargo instigated
a worldwide reassessment of alternatives to petroleum. Since
methanol can be made from coal (or virtually any carbon source)
by proven technology, the Mobil development offered a new al-
ternative to classical synfuel processes such as the Fischer-
Tropsch and Bergius.
A sizable body of literature dealing with the theoretical and
applied aspects of hydrocarbon formation from methanol has
grown over the last decade and continues to expand.
An attempt is made in this survey to cover the journal litera-
ture comprehensively, while the patent literature is, of necessity,
treated with greater selectivity.
11. GENERAL CONSIDERATIONS
A. Stoichiometric Constraints
Hydrocarbon formation from methanol requires the elimination

of oxygen, which can occur via three modes: elimination a s H,O,
CO, or CO, (elimination as 0,is prohibited by thermodynamics).
Reactions involving the coproduction of oxygenates such as for-
mates can also be written. However, with the possible exception
of methane formation, as discussed later, no clear-cut experimen-
tal examples of this reaction are known. For the present, there-
fore, the discussion will be limited to hydrocarbons as the sole
organic product.
Depending on the nature of hydrocarbons formed, H 2 or coke
may also be produced. In the simplest case, where H 2 is the only
coproduct , the stoichiometry can be written a s follows:
HYDROCARBONS FROM METHANOL 3
CHSOH * [CHZI + H2O ( 1)

Here [CH,] represents mono-olefins, cycloparaffins, or the
average composition of a mixture of paraffins and aromatics.
Acetylenic compounds o r polyenes have not been observed in
significant amounts in studies reported to date.
When paraffins a r e the sole hydrocarbon products, several
reactions can be written. The principal reactions a r e
( n + l ) C H 3 0 H * CnH2n+2 + C + ( n + 1)HZO ( 2)
(2n + 1 ) C H 3 0 H + 2CnHzn+z iCO + H20 (3)
n = 1 , , 2 , 3, ... (4)
Equation ( 4 ) is derivable from ( 3 ) by allowing the water-gas

shift reaction to occur and go to completion.
In forming aromatics from methanol, 3 mol of H 2 (exclusive of
H 2 in HzO) are necessarily eliminated or transferred for each aro-
matic ring generated. The general equation for mononuclear aro-
matics formation is
nCHJOH + C n H z n - 6 + 3H, + nH,O, n = 6, 7 , . .. ( 5)
Although pure examples of (5) have not been observed to date,

the enhancement of aromatics in the presence of certain metal/zeo-
lite combinations, with concomitant H2 production, is known [ 31 .
B. Thermochemistry
Hydrocarbon formation from methanol is an exothermic reaction.

B y way of illustration. reaction enthalpies 141 associated with E q s .
(11, ( 2 1 , and (5) are plotted in Fig. 1 for n-paraffin. linear 1-
olefin, and methylbenzenes formation at 316OC. The olefins curve,
which may also be considered representative of a balanced paraf-
fin-aromatic distribution ( i . e . , with no H, or coke make) is seen
to level off at -AHR = 300-400 cal/g. This is in accord with the
enthalpy data of Chang and Silvestri [21 for hydrocarbon forma-
tion from methanol at 3 7 l o C , which is shown in Fig. 2 a s a func-
tion of conversion.
It is also apparent from Fig. 1 that the degree of exothermicity
will vary depending on product distribution.
From a practical standpoint, the control and removal of this
large reaction heat is a major constraint in reactor design.
4 CHANG
- 400
r PARAFFINS + C + H20
- 600 bl 2 4 6
CARBON NUMBER
8 10 12
FLG. 1. Hydrocarbon formation from methanol. Heats of re-

action for selected hydrocarbon products at 600 K .
111. CATALYSTS
A. Early Observations (pre- 1960)
The earliest report of hydrocarbon formation from methanol

is that of LeBel and Greene [ 5 ] , who described in 1880 the
lo3 I I 1 1 1 1 1 1 I I I l r l l l I I I I 1 1 1 1 1 1 I 1 I l l
2
10
1 AT 100% CONVERSION: OHR -400 cal I g ( - I674 k J / kg 1
I I I I I I 1 I l l I I I I I I l l 1 I I I I 1 l I l I I 1 1 1 1 1 1
lo-' 10' 3 10-2 10- I 1

LHSV-'
FIG. 2. Methanol conversion to hydrocarbons. Heat of reac-

tion vs space-time at 371% [ 2 ] .
decomposition of methanol in molten ZnC1,. Hexamethylbenzene

(HMB) and light gases, mostly CH,,, were identified as the main
products, A possible equation for this extraordinary reaction
is
C
15 CH30H FUSED ZnC& '@

C
+ 3CH4+ I5yO
AG~"" = -265.2 kcal/rnol
A s written, the reaction has AG = -261 kcalfmol of HMB at 283OC.

the melting point of ZnC1,.
6 CHANG
This large driving force for HMB formation has mechanistic a s

well as process implications which are later discussed.
LeBel and Creene interpreted their result, perhaps propheti-
cally, in terms of V H , " condensing to benzene, followed by ex-
haustive ring methylation by a Friedel-Crafts reaction with CH,Cl
generated in situ.
In 1914 Sernagiotto [ 61 reported the decomposition of methanol
by P,O, in a paper titled "On the Chemistry of Nascent Methylene."

The vigorous reaction yielded a mixture of hydrocarbons with the
general formula (CH,)n. Propene and butene were identified via
bromination. Hexene and hexane, along with unidentified solid
products, were also produced. For this example, P,O, would prob-
ably be more accurately described as a reactant rather than a cat-
alyst in the true sense.
A patent was later granted to Grosse and Snyder 171 for the
conversion of methanol and DME to hydrocarbons over ZnC1, at
375-675OC and superatmospheric pressures. Table 1 contains
typical results. The light hydrocarbons are mostly isobutane.
The heavier fractions contained significant HMB.
Adkins and Perkins [8] studied the reactions of methanol over
Al,O,. At 300-350°C methanol was quantitatively converted to
DME, and above 350°C, CH,, CzH4 as well as C O , CO,, and H,
were detected.
Topchieva and Ballod [ 91 compared the activity of silica gel,
Al$, and silica-alumina (Al,O, 30% Sio,70%), for methanol con-
.
version to DME The aluminosilicate catalyst, after adsorbing
methanol (0.58 mmol/g) at 20°C was heated to 4 O O O C and gave
CO,, C2H4, CO. and C,H, in addition to DME, C , and unreacted
methanol,
Cullinane et al. [ 101 found that methanol contacted with acti-
vated Al,O, or Type-C Al,O, at 45OOC yielded small quantities of
hexamethylbenzene.
Gorin [ 111 reacted a mixture of dimethyl ether (18.2 mol%)and
isobutane over amorphous silica-alumina at 37OoC, 150 psig, and
6.8 h-l GHSV. The conversion of dimethyl ether to hydrocar-
bons was 94.58, while no net change in isobutane occurred. The
yield distribution on an isobutane-free basis is shown in Table 2.
In the absence of isobutane, at higher temperature, lower pres-
sure, and higher space velocity, considerable methane, CO , and
carbon were produced, while the yield of C,+ hydrocarbon was
less than 15%. Although no explanation of the role of isobutane
was offered, one can surmise that it served to moderate the large
.
reaction exotherm According to Gorin, methanol reacts more
slowly and produces less liquid hydrocarbons, although no data
were given.
TABLE 1
Conversion of Methanol to Hydrocarbons over Fused ZnC1, [ 71
(425OC, 166 atm)
Mole percent MeOH

converted to
CH4 1.0
C2H6 0.4
1.3
4.0
1.0
19.6
1.5
C4H8 0.7
C, to 85OC (1.5 tom) 33.1
Higher boiling products 30. Oa
Other -
7.4
100.0
a
18%hexarnethylben zene based on MeOH converted.
A patent issued to Fawcett and Howk [12] claims the direct

conversion of methanol to hydrocarbon wax in the presence of
a metal rnolybdite catalyst, e . g . , CoMoO,, modified with nickel
chromite at 100-350°C at 70-1000 atm. Here the question arises
a s to whether methanol is not simply a source of synthesis gas
via dissociation, and therefore whether the actual hydrocar-
bon-forming reaction is a Fischer-Tropsch reaction. Such is
the case in an earlier study by Eidus [13] where methanol was
reacted over a cobalt-thoria-kieselguhr catalyst under Fischer-
Tropsch conditions and found to give the s a m e hydrocarbon
products a s with CO + 2H2, albeit in considerably lower
yield.
8 CHANG
TABLE 2
Dimethyl Ether Conversion over Si02-A1,03 [ 111
(37OoC, 10 atm, 6 . 8 h-l QHSV)
Products %
CHI4 5.7
C2H4 10.2
C3H6 + C3H8 8.1
n-C4H,, + C,H8 5.2
i-C 26.4
i-Cs 11.0
c7 3.0
c 8+ 21.6
Carbon 8.6
99.8
B. Zeolite Catalysts
1. Introduction
Zeolites are porous, crystalline aluminosilicates composed of
A104 and S O 4 tetrahedra, interconnected through shared oxy-
gen atoms, forming a three-dimensional framework. Since every
oxygen in such a structure (viewed a s an infinite lattice) is
shared by two tetrahedra, the framework will possess a net
negative charge. This is balanced by (exchangeable) cations
Mn+ in the structure, leading to the general representation
where n is the charge on the cation and z is the water of hydra-

tion. In general, y / x ? 1, and in some exceptional cases, e . g . ,
the synthetic zeolites ZSM-5 114, 151 and ZSM-11 [16, 171, 5 <
y /x 0 0 . At ylx = w , the framework will, of course, be elec-
t rically neutral .
Erionite 8 Ring ZSM-5 10 Ring Y Zeolite 12 Ring

(Stroight Channel1
FIG. 3. Typical zeolite pore geometries.
When M is a proton, the zeolite becomes a strong Bronsted

acid. A s catalysts, zeolites are unique in their ability to dis-
criminate between reactant molecules and to control product se-
lectivity, depending on molecular size and shape, This phenom-
enon, known as "shape-selective catalysis [ 18-22] , is a con-
sequence of the well-defined geometry of zeolite pores, chan-
nels, and cages. These are of molecular dimensions.
The catalytically most significant zeolites are those having
pore openings characterized by 8-, l o - , and 12-rings of oxy-
gen atoms. Some typical pore geometries are shown in Fig. 3.
For comprehensive treatments of zeolite structure and chem-
istry, the reader is referred to the classic monographs of Breck
1231 and Rabo 1241.
2. ZSM-5 and Related Zeolites

The most extensive work on hydrocarbons from methanol has
been done with ZSM-5 zeolite catalysts. There is little question
that these are the most effective catalysts yet discovered for
this reaction.
a. Structure of ZSM-5 and ZSM-11

The unique catalytic properties of ZSM- 5 zeolites have been
attributed to their crystal structure [ 1, 1621.
Zeolite ZSM-5 [ 14, 15, 251 has orthorhombic symmetry Pnma
with cell parameters a = 20.07, b = 19.92, c = 13.42 A. The
10 CHANG
FIG. 4. ZSM-Kchannel system.
channel system, represented in F i g . 4, consists of straight chan-

nels running parallel to 10101 and intersecting sinusoidal chan-
nels parallel to [ 1001. The channels are ellipsoidal with 10-ring
openings, having the approximate free dimensions 5 . 4 X 5.6 w
(straight channels) and 5 . 1 X 5 . 4 1 (sinusoidal channels) based
on oxygen radii of 1 . 3 5 A. Figure 5 shows a stereo-pair drawing
of the ZSM-5 framework viewed along [ 0101.
Closely related to ZSM-5 in structure and properties is ZSM-11
116, 17, 251. ZSM-11 has tetragonal symmetry I h 2 , with cell
parameters a = 2 0 . 1 , c = 1 3 . 4 1. The channel structure, shown
in Fig. 6, consists of intersecting straight channels, with 5 . 1 X
5 . 5 free dimension. The framework is shown in Fig. 7.
b.General Reaction Path

The conversion of methanol over ZSM-5 zeolites was investi-
gated by Chang and Silvestri [ 21 who found the reaction to pro-
ceed according to the following general reaction path:
paraffins
-H20 -H20
2CH3OH rCH,OCHS -C2'--C5'-- aromatics
cycloparaffins
+H20 C,+ olefins
Scheme A
This was established by monitoring changes in product distribu-

tion as a function of varying contact time, shown in Fig. 8. An
essentially identical reaction path was obtained with DME as feed,
Fig. 9, confirming its intermediacy in the reaction sequence.
HYDROCARBONS FROMMETHANOL 11
ZSM-5 Pnma 8 T,-,*P 1
viewed along (010J
Secondary buildlng units: complex 5-1
Framework density:
Channels:
1 7 . 9 T/1000 A3
{ [OIO] 5. 4 x 5.6 - [loo] -

10 5.1 x 5.5r.t-
Fault planes: (100)

Type species: Mobil synthetic zeolite ZSM-5
Na,Al,Si,~,O,,, - 16 H,O with n c 27 and
typically about 3
orthorhombic, Pnma, a=20.1 b=19.9 c = 1 3 . 4 A

FIGURE 5 1251.
FIG. 6 . ZSM-11channel s y s t e m .
12 CHANG
viewed along [loo]
Secondary building units: complex 5-1
Framework density: 17.7 T/1000 A3
Channels: < 100, 5.1 x 5 . 5 * * *
Fault planes: (100)

Type species: Mobil synthetic zeolite ZSM-11
Na,Al,Si,.,O,, - 16 H,O with n < 16 and
typically about 3
tetragonal, Iam2, a=20.1 c=13.4 A
FIGURE 7 [25].
This reaction path has been confirmed in general by many work-

ers [26-291
A variation has been proposed by Voltz and Wise [301 based on
results from a fluid-bed kinetic study. The modified scheme,
Scheme B , contains a direct path to olefins from methanol:
2CHjOH CH,OCH, + H2O
\ J
olefins
Scheme B
70t
/
\a DIMETHYL ETHER 1’
KY2 lo-’ I I0
I
SPACEITME (m)
FIG. 8. Reaction path for methanol conversion to hydrocarbons over HZSM-5 (371OC) [ 21.
14 CHANG
- -
-d DIMETHYL ETHER
a\
5
m ,,----4-
+
K
Y
5
30
20
10
0
10' lo-' 1
1
SPACE TIME 1-(
FLG. 9. Reaction path for dimethyl ether conversion to hydro-

carbons over HZSM-5 (371OC) 121.
Anderson et al. [27] and Ceckiewicz I431 also recognized this

possibility, This is formally identical to the mechanism of
Balandin et al. [31] for ethanol dehydration over alumina, which
was elucidated by the kinetic isotope method. The relative ki-
netic significance of this auxiliary step is unknown.
c. Product and Distribution

Table 3 [22] contains typical hydrocarbon selectivities at 371OC
for ZSM-5 and ZSM-11, showing the general similarity between the
two zeolites. The hydrocarbons span a relative narrow range of
molecular weights, terminating at about C l 0 . At this stage of re-
action, conversion of the intermediate olefins is essentially com-
plete and the main products are isoparaffins and aromatics.
The aromatics are mostly methyl-substituted. as seen in Table
4. Alm shown i s a comparison of the observed isomer distribu-
tion with their thermodynamic equilibrium values. It is evident
that the xylene distribution is close to equilibrium, while among
TABLE 3 [22]
Methanol Conversion to Hydrocarbons over Various Zeolites
(37OoC, 1 atm, 1 LHSV)
Hydrocarbon distribution (wt%) in

Erionite ZSM-5 ZSM-11 ZSM-4 Mordenite
c, 5.5 1.0 0.1 8.5 4.5

c2 0.4 0.6 0.1 1.8 0.3
c22- 36.3 0.5 0.4 11.2 11.0
c3 1.8 16.2 6.0 19.1 5.9
c,*- 39.1 1.0 2.4 8.7 15.7
c, 5.7 24.2 25.0 8.8 13.8
c,2- 9.0 1.3 5.0 3.2 9.8
C 5 + aliphatic 2.2 14.0 32.7 4.8 18.6
A, 1.7 0.8 0.1 0.4
4 7 10.5 5.3 0.5 0.9
A8 18.0 12.4 1.3 1.0
A3 7.5 8.4 2.2 1.0
A10 3.3 1.5 3.2 2.0
A 11+ 0.2 26.6 15.1
the higher aromatics, 1,3,5-trimethylbenzene, 1,2,3,4-, and

1,2,3,5-tetramethylbenzenes fall significantly short of their
equilibrium values. This was attributed to zeolite "shape-selec-
tivity," a s was t h e sharp cut-off in product molecular weight.
d. Effect of Temperature
The effect of temperature on ZSM-5 selectivity a t low w a c e
velocity (LHSV = 0 . 6 - 0 . 7 h - l ) is shown in Fig. i0 121. Below
3OOOC the main reaction is dehydration t o DME. With increasing
temperature t h e sequence of Scheme A is followed, while above
45OOC light olefins and CHI, become significant a s a result of
secondary cracking. Above about 5OOOC t h e dissociation of
methanol to H2 and CO becomes measurable.
16 CHANG
TABLE 4 [2]
Aromatics Distribution from Methanol
Conversion over HZSM-5
Normalized Normalized Equilibrium

distribution isomer distributions
(wt %) distributions (371OC)
Ben zene 4.1

Toluene 25.6
Ethylben zene 1.9
Xylenes:
0
m
P
Trimethylbenzenes :
9.0
22.8
10.0
I::[ [":I
23.9 23.5
[:3
123 0.9
124
135
Ethyltoluenes :
11.1
2.1 [
0 0.7
m+P 4.1
Isopropylben zene 0.2
Tetramethylbenzenes:
1234
1235
1245
0.4
1.9
2.0
[ [: I
33.4
Other Al0a 2.7
All+ 0.4
a
Diethylbenzenes + dimethylethylben zenes .
I00
90
80
WATER
I-
70
W
V
a
W
60
c
I
'3
y 50
w
1
2
J
40
3
I C2 - C 4 HYDROCARBONS
3
0 30
20
10 AROMATICS -
0
300 400 500
T E M P E R A T U R E 'C
FIG. 10. Effect of temperature on methanol conversion over

HZSM-5 (0.6-0.7 h-' LHSV, 1 atm) [ 2 ] .
e. Effect of Pressure
The main effect of varying reactant partial pressure on ZSM-5
selectivity is to change the relative rates of the olefin-forming and
.
aromatization steps [ 321 Decreasing pressure tends to decouple
the two reactions, while increasing pressure enhances the overlap
of the two reactions. This is illustrated in Figs. 11 and 1 2 for
PMeOH = 0.04 and 50 atm. These plots may be compared with
Fig. 8 depicting the normal atmospheric reaction path. The sig-
nificance of the shaded regions will b e discussed later.
18 CHANG
'"I
60
- - -0- WATER
8 50-
/
5
z .O'
0
40-
2
30
- DIMETHYLETHER o
/
5
3
0
0
g 20-
,,
10 -
0 - v
- AROMATICS
FIG. 11. Methanol reaction path at low pressure (HZSM-5 cat-

alyst, 371OC) [321.
Thus at subatmospheric partial pressures, high selectivity to

olefins is possible at complete conversion of methanol (Table 5).
The most striking effect of raising pressure above atmospheric
is an increase in the degree of aromatic substitution, especially
in durene selectivity. A s shown in Table 4, durene is not the
thermodynamically favored tetramethylbenzene isomer. Its pref -
erential formation is a consequence of zeolite shape-selectivity.
Among the tetramethylbenzenes , durene possesses a critical di-
ameter most compatible with the zeolite channel dimensions, i.e. ,
it is the bulkiest isomer able to move with relative ease through
the zeolite channel system, and is therefore more easily formed
than the other Clo aromatics within the zeolite. It is shown later
that in large pore zeolites, Clo equilibrium is attained.
f . Modifications and Isotypes

Kaeding and Butter [331 reported that ZSM-5 modified by treat-
ment with phosphorus compounds showed high (70%) selectivity for
E 501
/
50 /
/
2- /
0
5 40 I /
I
,
/
I
m
K
I-
5!?
0 30
I-
V
V 0- PARAFFINS
3
0
0
K 20-
20
n
0- AROMATICS
10-
10
0 L
10-2 10-1 1 10
LHSV-' , hr.
FIG. 12. Methanol reaction path at high pressure (HZSM-5 cat-

alyst, 371OC) [32].
C2-CI, olefin formation from methanol. The following was proposed

as to the nature of trimethyl phosphite modification :
H
I
O\ /o\*l/o\si/o
0 /si\o o/+,\o o\
20 CHANG
TABLE 5
Methanol Conversion at Low Pressures [ 1611
(427OC, 1 LHSV, 1 atm total pressurea)
Methanol partial pressure, atm: 1.00 0.25 0.17 0.07

Conversion, 8: 99+ 99+ 99+ 99+
Hydrocarbon distribution, wt %:
Ethylene 3.2 12.4 17.4 21.0
Propylene 4.8 18.2 26.5 38.7
Butenes 2.2 6.3 7.6 18.5
Pent ene s 0.4 0.3 0.7 2.4
(Total C,-C, olefins) (10.6) (37.2) (52.2) (80.6)
Methane 1.5 0.8 0.6 0.5
C2-C5paraffins 43.0 39.4 24.2 15.6
C,+ nonaromatic 3.9 2.3 2.7 1.3
Aromatics 41.0 20.3 20.3 2.0
a
Helium diluent at subatmospheric MeOH partial pressures.
CH30\ I
P/OCH
+CH ,OH
Other than to reduce the activity of the ZSM-5 acid sites, the
function of P-modification is unknown.
The conversion of ethanol over "Ultrasil" zeolite, which is
ZSM-8 [ 3 4 , 351, has been described by Russian workers 128, 3 6 ,
371. Representative data in Table 6 show that this catalyst be-
haves similarly to ZSM-5.
TABLE 6
Methanol Conversion over "Ultrasil" Zeolite [ 361
Reaction conditions:
Temperature, O C 300 300 450 4 50 4 50
Pressure, torr 30 80 30 80 80
Time on stream, min 60 60 60 3 60
Hydrocarbons, wt %:
CH4 7.99 5.52 11.94 9.48 1.74
C2H4 + C2H6 4.97 21.84 15.74 8.10 6.38
C,He 3.03 2.09 5.24 1.47 10.35
C3H6 5.52 15.39 31.83 29.88 14.98
i-C 4H lo 22.25 10.12 7.69 4.98 23.06
n-C4H10 1.57 0.33 1.22 0.51 4.86
C4Hs 7.08 3.98 14.49 20.98 10.66
i-C,H,, 16.78 3.29 2.62 3.54 10.08
n-C,H,, and higher 30.81 37.44 9.23 21.06 17.80
Spencer and Whittam 1551 have described a high silica zeolite

of unknown structure called Nu-1, which behaves similarly to
ZSM- 5 in methanol conversion, but deactivates quickly.
3. Other Zeolites
a. Small Pore Zeolites
Zeolites which sorb linear hydrocarbons and exclude branched
ones have received attention as catalysts for methanol conversion
to light hydrocarbons such as ethene and propene. Structures of
some of these small pore zeolites are illustrated in Figs. 13-16.
The erionite structure, Fig. 13, is hexagonal, containing
"supercages" supported by columns of cancrinite units linked
through double-6-rings (D6R). Access to, and between, the
supercages is gained through 8-rings.
Offretite is closely related to erionite except that the sequence
of 6-rings is AABAAB [231 as compared to AABAAC in erionite.
22 CHANG
Secondary building units: single 6-rings in AABMC sequence

(single 4-rings)
Framework density: 15.6 T/l000 A3
Channels: -I [OOl] -8 3.6 x 5.2 ***

Fault planes: (001)
Type species: Erionite (Na2,Ca...),,, Al,Si,,O,, 27 H 2 0

hexagonal, P63/mmc, a=13.3 c=15.1 A
FIG. 13. Erionite.
Intergrowths of erionite and offretite, e. g., zeolite T , occur fre-

quently.
In the chabazite framework, Fig. 14. the D6R layer sequence
is ABCABC, and the D6R units are linked together through tilted
CHABAZITE Rim 36 T[1]
viewed along [OOl] hex
Secondary building units: single 6-rings in AABBCC sequence

(double 6- o r single 4-rings)
Cham els: -
1 [OOl] -
8 3.6 x 3.7**+
Fault planes: (0011
Type species: Chabazite Ca,Al,,Si,,O,, a 40 HzO

trigonal, Rjm, a=13.2 c=15.1 A
(True symmetry i s lower, possibly triclinic)
FIGURE 14 1251.
4-rings. The framework contains large ellipsoidal cavities (Fig.

15), each entered through six 8-rings. These cavities a r e joined
together via their 8-rings, forming a 3-dimensional channel sys-
tem.
Zeolite ZK-5, Fig. 16, consists of truncated cuboctahedra
joined through D6R units in a body-centered structure. The
main channel system is defined by 8-rings.
24 CHANG
FIG. 15. Chabazite cavity,
Methanol conversion over a variety of small pore zeolites has

been reported by Chang et 81. 1381. Representative data are
shown in Table 7. At 339-538OC the products, restricted largely
to the C2-C, range, were mainly olefins. The presence of signi-
ficant amounts of methane in some experiments is symptomatic of
coke laydown.
The conversion of methanol to light olefins over various cation-
exchanged chabazites has been investigated by Cobb et al. [391
and Singh et al. (40, 411. Both teams of investigators observed
short-term catalyst deactivation due to coking, but found the cat-
alyst to be regenerable. Cobb et al. observed long-term irrevers-
ible deactivation, which was attributed to structural degradation,
based on drastic changes observed in the x-ray pattern. This
was disputed by Singh et al., who saw no such deactivation in a
three-month study comprising 21 regenerative cycles. It was
speculated that the discrepancy may have been due to differ-
ences in ion-exchange procedures. Singh et al. contended that
their method of preparation led to an ffultrastableform of cha-
.
bazite If
Ceckiewicz [42, 431 studied methanol over the hydrogen form
of zeolite T , Coking was rapid and was attributed to the forma-
tion of nondesorbable cyclic hydrocarbons in the zeolite super-
cages.
ZK-5 lm3m 96 T[1]
viewed along [loo]
Secondary building units: double 6-rings

(single 4-, 6- or 8-rings)
Channels: < l o o > -8 3.9***1 (100, S 3.9*+*
Fault planes: ----

Type species: synthetic zeolite ZK-5 Na3,A1,Si,,O,,, * 98 H,O
cubic, ImJm, a=18.7 A
FIGURE 16 [25].
Wunder and Leupold [44] report that selectivity to ethene and

propene is significantly enhanced over a mixture of chabazite and
erionite when these zeolites are Mn-exchanged. A methanol/water
mixture (30 ~ 0 1 H
%2 0 ) was reacted at 4OO0C, and gave the result
shown in Table 8. The product was 66.3% C2-C4olefins.
Whittam and Spencer [45] report that "zeolite MCH," which has
an x-ray pattern similar to herschelite (isostructural with cha-
bazite) but with line broadening due to small crystal size, is ac-
tive for converting methanol to C,-C hydrocarbons at 30O-55O0C.
Typical results are shown in Table 9. The product consists
mainly of ethene (16.6 vol%),propene (41.2%), and propene
(22.4%).
26 CHANG
TABLE 7
Methanol Conversion over Small Pore Zeolites [ 381
Erionitea Zeolite T Chabazite ZK-5
Reaction conditions:
Temperature, OC 3 70 341-378 538 538
Pressure, atm 1 1 1 1
LHSV (WHSV) , h-’ 1 (3.8)
b b
Conversion, % 9.6 11.1 100 100
Hydrocarbons, wt %:
CH 4 5.5 3.6 3.3 3.2
CH2H6 0.4 0.7 4.4 0
CzH4 36.3 45.7 25.4 21.4
C3Ha 1.8 0 33,3 31.8
C3H6 39.1 30.0 21.2 13.5
C4H10 5.7
10.4 22.6
C4HB 9.0 10.0
c 5+ 2.2 3.1 2.0 7.5
a
De-aluminized ; Si02/A1,03 = 16.
bPulse microreactor, 1 pL MeOH in He, 500 h - l GHSV.
Inui et al. [ 4 6 , 471 synthesized an erionite-offretite catalyst by

rapid crystallization in the presence of tetramethylammonium hy-
droxide. Methanol (128 in N1) was reacted over the calcined
(54OOC) material at 4OO0C, 1000 h - l GHSV and gave (m18) ethane
( 2 5 ) , propene ( 3 2 . 7 ) , butene ( 1 9 . 2 ) , methane ( 4 . 6 ) , propane
( 6 . 4 1 , butane ( 4 . 7 1 , and C,+ ( 7 . 5 ) at complete conversion.
b.Large Pore Zeolites

Faujasite-type zeolites, e . g . , X and Y . mordenite, and mazzite
(ZSM-4), are large pore zeolites which have been found to catalyze
methanol conversion to hydrocarbons.
The faujasite structure, Fig. 17, is built up of truncated octa-
hedra interconnected via D6R units. Faujasites contain extremely
large supercages ( ~ 1 1
3 diameter) entered into through 12-rings.
TABLE 8
Methanol Conversion over Mn-Exchanged Chabazite-Erionite I441
(40O0C. 90%conversion)
Products Vol %
C2H4
3H6
C4HB
CH4
C2H6
C3H8
C4H10
C sf nonaromatic <o. 1
Aromatics <o. 1
co 0.5
H2 5.2
H20 s2.0
Dimethyl ether 4.7
Mordenite, Fig. 18, is characterized by a one-dimensional sys-

tem of parallel elliptical channels, defined by 12-rings.
Mazzite (ZSM-4), shown in Fig. 19, also has a one-dimensional
system of parallel channels defined by 12-rings.
Schwart z and Ciric [ 481 were the first to achieve significant
methanol conversion to hydrocarbons over zeolite catalysts. They
reacted methanol over REX and ZnX at 330-390°C, and reported
the data shown in Table 10. Coke formation was rapid, which
may account for the high methane selectivity.
Topchieva et al. [49] studied methanol dehydration to dimethyl
ether over various alkali metal-exchanged X and Y faujasites. In-
terestingly, the only hydrocarbon they reported (426OC) was 1-
butene.
Attempts have been made to decrease the coking tendency of
faujasites by cation exchange with numerous metal cations having
hydrogenation activity [ 50, 511. These efforts have met with
28 CHANG
TABLE 9
Methanol Conversion over Zeolite H -MCH [ 451
(45OoC, pulse microreactor)
Products Vol %
~~~ ~
CH4 3.6
C2H6 0.2
C2H4 16.6
c 3H8 22.4
C3H6 41.2
i-C4Hlo 0
n-C4H10 4.9
l-C4H8 2.7
i-C4H8 2.0
2-C4H8 4.9
little success. The effect of increasing hydrogenation activity

of the metal component was mainly to enhance methanol dissocia-
tion [51].
The influence of acid strength of the protonic sites in zeolite
Y on dimethyl ether conversion a t 35OOC was investigated by
Cormerais et al. [ 521. A series of zeolites containing varying
concentrations of Na and K ions was prepared. The acid strengths
were determined by pyridine adsorption-desorption at different
temperatures. I t was found that dimethyl ether reacted only when
the total number of Na and K ions per unit cell was less than 16.4,
and that the active sites must be of sufficient strength to retain
pyridine at temperatures at least as high as 450OC.
The catalytic dehydration of methanol by synthetic H-mor-
denite was studied by Swabb and Gates [ 531 a t 99.5-240OC.
Olefins were detected a t 24OOC ; however, deactivation was rapid.
Natural mordenite, exchanged with various cations, was found
by Zatorski and Krzyzanowski [ 541 to be highly active for meth-
anol conversion to C1-C hydrocarbons at 350-500°C, but to have
a short life-time.
Mordenite and ZSM-4 were investigated by Chang et al. [ 221
for methanol conversion. Table 3, a s shown previously, sum-
marizes the reaction conditions and hydrocarbon distributions
FAUJASITE Fd3m 192 T[1]
viewed along (1111
Secondary building units: double &rings

(single 4- or 6-rings)
Framework density: 12.7 T/1000 A 3
Channels: <111> g 7.4***
Fault planes: (1111
Type species: Faujasite (Na,, Ca, Mg),,Al,,Si,,,O,,, 240 H,O

cubic, Fd3m, a=24.7 A
FIGURE 17 [25].
obtained. Included for comparison are results from small pore

erionite and intermediate pore ZSM-5 and ZSM-11.
4. Zeolite Shape-Selectivity
Table 3 provides a concise illustration of zeolite shape-selec;
tivity in methanol conversion to hydrocarbons, Small pore (4.3 A )
erionite, as already shown, produces only low molecular weight hy-
drocarbons. It does not sorb benzene and therefore cannot pro-
duce aromatics. The intermediate pore ( < 6 (A> ZSM-5 and ZSM-11
truncate the hydrocarbon distribution at C,, , while the large
30 CHANG
MORDENITE Cmcm
viewed along

n
Framework density: 17.2 T/1000 A’
Channels: [OOlI 12 6.7 x 7.0* tj I0101 -

8 2.9 x 5. 7*
Type species: Mordenite Na,Al,Si,O,, * 24 H,O

orthorhombic, Cmcm, a=18.1 b=20. 5 c=7. 5 A
FLGURE 18 [ 2 5 ] .
pore ( < 8 8 ) ZSM-4 and mordenite, which can accommodate mole-

cules as bulky a s hexamethylbenzene, produce this compound
along with other aromatics.
The normalized aromatics distribution is shown in Fig. 20 and
serves to highlight the influence of pore size. These results are
consistent with the critical dimensions of the methyl-substituted
benzenes, some of which are listed in Table 11.
It should be pointed out, however, that crystallographic pore
dimensions cannot generally be used to gauge precisely what
MAZZITE PGJmmc 24 T,[1] . 12 T,[m]
viewed along [OOl]

(or single 4-rings)
Framework density: 16.1 T/1000 A 3
Channels: [OOl] 12 7.4*
Fault planes: (100)
Type species: Mazzite Mg,(Ca, K2)3Al,oSiz60,2 28 H,O

hexagonal, P63/mmc, a=18.4 c=7.6 A
FIGURE 19 [ 251.
molecules will be sorbed and what molecules excluded. Molecules

larger than the calculated aperture size of zeolites are sometimes
freely sorbed. It has been found that for certain sorbate mole-
cules, the kinetic diameter from the Lennard-Jones potential pro-
vides a more satisfactory, though still not an absolute, match
with zeolite pore dimensions [ 231.
32 CHANG
TABLE 10 1481
Methanol Decomposition over Metal
Cat ion -Exchanged Paujasitesa
Temperature (OC) 330-380 360-390

% conversion of CH,OH to
gaseous products 51 30
Gaseous product composition
( m l %) :C
H2 Trace 10.3
CH4 22.5 43.0
C2H4 24.5 17.0
C2H6 2.7 2.9
C3H6 10.8 9.2
C3Hs 4.4
C4Hs 8.0 7.8

i-C4Hlo 13.9 4.0
i-CsHlz 6.2
Hexanes 4.3 2.0
Other
-
2.7 3.8
-
100.0 100.0
a
Continuous-flow; 1 atm; 1.5 LHSV; samples analyzed at 2-3 h
on stream.
b9.46 w t % Zn.
CLiquid product mainly unreacted CH,OH , H,O, and (CH,) *O.
The nonequilibrium distribution of tetramethylbenzenes from

ZSM-5 has already been noted. Comparison is here made be-
tween 1,2,4,5-tetramethylbenzene(durene) distributions in the
Al, fractions among the intermediate and large pore zeolites,
Table 12. This demonstrates that the large pore zeolites per-
m i t the attainment of equilibrium. It will be recalled that durene
loo
vo:L
ZSM-5 -
ISM I I
75 -
50 -
25
0 0
A6 A? A6 A7 A8 A9 AIO All
100
I S M-4 MORDENITE
75 75
YO
50 - 50
25 - 25
FIG. 20. Aromatics distribution from methanol conversion over

various zeolites [ 221.
is the bulkiest molecule that can diffuse readily through ZSM-5

and ZSM-11.
We have shown earlier that there is a large driving force for
methanol to form hexamethylbenzene in the presence of acid cat-
alysts, Thermodynamic calculations [ 1721 reveal, however, that
34 CHANG
TABLE 11
Critical Diameters of Selected Polymethylbenzenes [ 221
6.1
6.4
6.7
6.9
7.1
a
Equilibrium diameters (rmin) estimated from Courtauld models.
Kinetic diameter % 2 - 1 ’ 6 r,h.
TABLE 12 [22]
A,, Shape- Selectivity of Various Zeolites
%1245TMB in %1245TMB in
Catalyst hydrocarbon A10
ZSM - 5 1.6 50.1

ZSM- 11 1.2 84.0
ZSM-4 1.1 34.6

Mor denite 0.7 37.2
Equilibrium (37OOC) 33.4

TABLE 13 [ 1651
Coke Formation and Deactivation in Dimethyl
Ether Conversion, 35OOC
Residual Coke lhydro-

Catalyst activity (8) Wt%coke carbons X l o 3
HZSM-5 90 1.5 0.3

HY 8 9.3 40
H -erionite lchabazite 0.15 8.6 90
H-mordenite 0.15 6.8 200
polymethylbenzenes formation must be a kinetically controlled phe-

nomenon. Shape-selective zeolites such as ZSM-5 prevent the com-
pletion of this reaction as a consequence of steric hindrance.
Catalyst deactivation due to coke deposition is fairly rapid in
small pore and large pore zeolites. The intermediate pore zeo-
lites such as ZSM-5, on the other hand, have a high coke toler-
ance. Rollmann and Walsh [ 1641 have presented evidence that
carbon formation in zeolites is a shape-selective reaction con-
trolled by pore geometry. This has been confirmed by Cormerais
et al. [ 1651. Table 13 summarizes the results obtained by these
workers on the conversion of dimethyl ether over H-erionitel
chabazite, HZSM-5, H Y , and H-mordenite. An inverse correla-
tion is evident between deactivation rate and coke selectivity.
C. Nonzeolitic Catalysts
According to Dolgov [ 561, heating methanol in the presence

of 86%H,SO, at 135-140°C gives 2 - 4 % ethene, with the balance
dimethyl ether.
Matyushenskii and Freidlin [ 571 converted dimethyl ether to
ethene by decomposition over a catalyst composed of 39.1-43%
A120, + T i 0 2 , 0.83-0.93% Fe203, 0.15-0.3% Mg, and a trace of
CuO deposited on clay pellets and calcined at 700-750OC. A t
100-15O0C, 1-1.5 atm, and ~ 0 . 0 3h-’ WHSV, conversion was 85%
and only ethene was detected in the reactor effluent. The cat -
alyst required air regeneration after 15-20 h .
Pearson [58] reported that when methanol or trimethyl phos-
phate was heated with phosphorus pentoxide [ 61 andlor polyphos-
phoric acid at 190°C, a 36-39%yield of hydrocarbons was obtained
36 CHANG
with the average formula CH2. over a range CnH2n+z to CnHZn-10.

Apart from dimethyl ether, the products contained alkanes, cyclo-
alkanes, alkenes, substituted monoaromatics (maximal at C 1 2 ) , and
other compounds,
Bell and Chang [ 591 found that coke formation during methanol/
dimethyl ether conversion to hydrocarbons in fused ZnC1, (283OC)
is suppressed by cofeeding hydrogen. Methyl chloride was s i m i -
larly converted to hydrocarbons in the fused salt.

K i m et al. [SO] converted methanol (>99%conversion) over bulk
zinc iodide at 2OOOC to hydrocarbons. The gasoline range ( C 5 -
C15) fraction consisted mainly of branched paraffins, and con-
tained a high percentage (49.7%) of triptane. Only small amounts
of C,- hydrocarbons were formed and about 56%(based on C feed)
gas oil range (230-270OC) hydrocarbons were produced. Heavier
fractions amounted to 2-3% based on C feed, with little solid resi-
due.
Supported aluminum sulfate was found by Hargis and Kehoe
[611 to catalyze methanol and dimethyl ether conversion to hy-
drocarbons.
Kikkawa et al. [ 621 used supported aluminum dihydrogenphos-
phate as catalyst to convert methanol into a predominantly olefinic
product (20-608 yield of C,-C, olefins at 375-425OC).
Heteropolyacids and salts of heteropolyacids have been used
by On0 et al. 163, 641 to convert methanol. Typical data, ob-
tained at 300°C, 2.12 X mol/h feedrate, and atmospheric
pressure are shown in Table 14. The hydrocarbons were mostly
in the C2-C5 range. Very little aromatics were formed.
The results of Cormerais et al. [ 651 using silica-alumina cat-
alyst will be discussed later in connection with reaction mechan-
isms.
IV. REACTION KINETICS
Relatively little has been published to date on the kinetics of

hydrocarbon formation from methanol. Global kinetic treatments
are mainly found in the accessible literature. However, in view
of the great complexity of the reaction, rigorous kinetic treat-
ments may neither be practicable nor have much practical justi-
fication. Further, the large heat effects require special consid-
eration. This has usually been neglected in the published ex-
perimental work,
The kinetic scheme is simplified somewhat by the observation
[30, 66, 671 that over a wide range of conversions the initial
step of ether formation is much more rapid than the subsequent
TABLE 14 [63]
Product Distribution of the Conversion of
Methanol into Hydrocarbons
Catalyst
HTPa CuTP AgTP HTSa CUTS AgTS
Product distribution, %:b

MeOH 1.3 2.1 0 3.9 4.7 0
MeOMe 38.6 37.3 2.0 57.5 35.5 20.1
Hydrocarbons 60.1 61.5 98.0 38.6 50.8 79.9
Hydrocarbon distribution, %:b

CH4 3.7 5.2 9.0 1.6 7.3 3.2
C2H4 11.3 9.5 9.0 10.3 11.2 10.3
C2H6 0.9 0.8 5.2 0.5 0.5 1.2
8 6 8.3 8.5 3.8 8.3 8.7 8.3
C3H8 16.1 13.4 34.0 9.8 14.5 21.9
c4 39.3 39.2 26.1 41.8 35.1 36.4
c5 12.5 13.7 7.2 15.8 15.1 13.3
c6 7.9 9.7 5.7 11.9 7.6 5.4
a
TP and TS indicate dodecatungstophosphate and dodecatung-
stosilicate, respectively.
bCalculated on a carbon-number basis.
olefin-forming step, and is essentially at equilibrium. Data illus-

trating this fact appear in Fig. 21. Thus the equilibrium oxy-
genate mixture can be conveniently treated as a single kinetic
species or "lump" [ 681 .
Data from a pilot plant study of gasoline synthesis from meth-
anol over ZSM-5 [SO] indicate that the rate of oxygenate dis-
appearance is first-order in oxygenates. This work will be de-
scribed in some detail later.
0 METHANOL FEED, CONTfNUCUS FLOW

$ METHANOL FEED, PULSE REACTOR
DME FEED, CONTINUOUS FLOW
MOM-DME EQUILIBRIA
---- }
DM E nc
Q
F I G . 21. Conversion of methanol and dimethyl ether to hydrocarbons over HZSM-5. 371OC
normalized [CH,] basis.
Chen and Reagan [ 691 discovered that the reaction is autocat -

alytic over ZSM-5. They proposed the following scheme:
where A = oxygenates, B = olefins, and C = aromatics + paraffins.

Neglecting k, at low conversions, experimental data were fitted
according to the expression
giving characteristic sigmoidal curves.

Figure 22 shows the fit with data obtained using various cat-
alyst loadings and reactor configurations, confirming reactor iso-
thermicity, and demonstrating that the results were not due to
autothermal effects. Figure 23 shows data obtained with catalysts
of different activity. A linear correlation between k, and intrinsic
acid activity is indicated.
Autocatalysis was confirmed by Ono et al. [70, 711. These
workers, however, proposed a sequence of two bimolecular re-
actions :
where A and B are oxygenates and olefins, respectively. With

x = conversion of A and w = weight of catalysts, these result in
where c1 = k1/k2 and f3 = [Bl,/[A],. Figure 24 shows the fit ob-

tained at three temperatures.
The kinetic scheme of Chen and Reagan was modified by Chang
[ 681 to include an additional bimolecular term due to homologation
of olefins :
A + B - Ck2
40 CHAN G
1.o 1
kl lk2=.02 k2-55
0.8 - 0 CATALYST (a) -

A CATALYST (b)
0 CATALYST (c)
VCATALYST (d)
-
0.6 -
04- -
-
0.2 -
0
0 I I
(a) Reactor: &in.-o.d. copper; catalyst:

10% zeolite diluted with 7-alumina.
(b) Reactor: Q-in.-0.d. copper; catalyst :
1.4% zeolite diluted with M/50 copper
powder and y-alumina.
(c) Reactor : A-in.-0.d. copper ; catalyst :
3% zeolite diluted with 50/50 copper
powder and y-alumina.
(d) Reactor: +in.-0.d. stainless steel ;
catalyst : 10% zeolite diluted with
?-alumina.
FIG, 22. Autocatalysis in methanol conversion (HZSM-5cat-

alyst, 37OOC) 1691.
B+C k3 w C
C A D
where A = oxygenate, B = (:CH,), C = olefins, and D = paraffins +

aromatics.
The species B was assumed to have carbene-like character,
attacking both oxygenates and olefins. Invoking the steady-
7 . - I 7 I 1 1
0 CATALYST A
A CATALYST B
0 CATALYST C
FIG. 23. Correlation between methanol rate constant k, and

intrinsic catalyst acid activity [691 . For catalyst identifications,
see Fig. 22.
state assumption on B and eliminating time, the following ex-

pression resulted:
where u = C I A , K1 = k , / k z , and k, = k+/k,. Data a t three pres-

sures were correlated with this model and are shown in Fig. 25.
42 CHANG
reaction timeh
FIG. 24. Hydrocarbon yield vs reaction time. Solid lines are

theoretical [ 7 11 .
Doelle and co-workers 1721 studied both sorption and reaction

kinetics of methanol and dimethyl ether conversion over ZSM-5.
A t 115-2OO0Cthe kinetics of methanol conversion obeyed the rate
law
k l PH~ ,OH
r =
1 + k2P
H2 0
with
k , = 6 X l o 5 exp (-80 y p )’ mol

gesebar
For dimethyl ether, no effect of crystal size on product distribu-

tion was observed, and the intracrystalline Thiele modulus was
less than unity for 0 . 5 pm crystals.
oxygenate
F I G . 25. Methanol reaction path. Solid lines are theoretical

[681.
The conversion of methanol over chabazite was investigated by

Anthony et al. [ 731. Using curve-fitting techniques for model
discrimination, it was found that the data could be represented
by
kloa3P
A
r =
A (1 + k,oclPA)3
where
1
a=
1 + bt/SL
44 CHANG
accounts for catalyst deactivation and Sk is a selectivity function

defined by
S.. = a. + b.x. + c x.2
11 1 1 1 i l
where
moles of i produced for the j-th data point
s.. =
11 moles of methanol reacted for the j-th data point

xj = conversion of methanol for the j-th data point
The kinetics of dimethyl ether conversion over HZSM-5 was
studied over the temperature range 227-300°C by van den Berg
1741. These results are discussed in the following section.
V. REACTION MECHANISM
The reaction path of acid-catalyzed hydrocarbon formation from
methanol may be viewed essentially as composed of three key steps:
ether formation, initial C-C bond formation, and aromatization with
ti-transfer.
The mechanism of the first two steps is discussed in this sec-
tion. The final stages, comprising olefin condensation, cyclization ,
and H-transfer over acidic catalysts, have been well studied and
proceed via classical carbenium mechanisms. This topic will there-
fore not be covered in detail here since it has been comprehensively
treated in a number of excellent reviews [ 75-77]. In addition,
papers by Galwey [781 on carbenium pathways in reactions of C2+
alcohols on montmorillonite, and Dejaifve et al. [ 791 on methanol
and olefins conversion over ZSM-5 are pertinent.
The mechanism of polymethylbenzenes formation will, however,
be discussed since these compounds are peculiar to the methanol
transformation reaction.
A. Ether Formation
The mechanism of ether formation from alcohols over oxide cat-

alysts, particularly Al2O3,has been extensively investigated. The
bulk of work has concentrated on alcohols having B-hydrogens.
Several comprehensive surveys have been published [ 80, 81, 841.
In contrast, the literature on methanol dehydration is relatively
sparse. Methanol etherification is similar in many respects to that
of the higher alcohols; however, since methanol lacks a parent
olefin, sufficient differences may be found a s to warrant its dis-
cussion here.
HYDROCARBONS FROMMETHANOL ~~ ~~ ~~
45
The dehydration of methanol on alumina and amorphous silica-

alumina was studied by Parera and Figoli 1821. The effect of a
series of nitrogenous poisons (various amines and N -heterocycles)
or dimethyl ether formation was monitored. Silica-alumina was ir-
reversibly poisoned, while alumina was reversibly poisoned. This
was taken as evidence that the reaction over alumina involved s u r -
face methoxy groups:
/A1\
0
/A1\
0
/A1\
0
P'\
This interpretation was based on the assumption that bases such
as diethylamine dissociatively but reversibly chemisorb on alu-
mina :
Et Et H
I I I
Because of the greater nucleophilicity of N vs 0 , the nitrogenous

base competes with methanol for the surface oxygen, thereby in-
hibiting methoxylation. On the other hand, the reaction over
silica-alumina involves Bronsted sites, which are strongly poi-
soned by nitrogen base, forming stable quaternary ions. Ac-
cording to these investigators ether formation over silica-alu -
mina may be represented by the following scheme:
CH,OH + H+ + CH,OH,+
CH30H; + CH30H -+ (CH,),O + H 2 0 + H+
Figueras et al. [83] studied the kinetics of methanol dehydra-

tion over silica-alumina at 160-200°C and derived the rate expres-
sion
46 CHANG
where Pa is the alcohol partial pressure. It was concluded that

the rate-limiting step cannot be the interaction of two surface
alcoholate groups, which would lead to a rate equation of the
form
r = ke2 = k a P / ( l + aPI2
At low pressures the reaction order would be unity, which is not

in agreement with the observed half-order. Figueras et al. pro-
posed two modes of chemisorption, both of which assumed the
concerted action of acidic and basic centers [85-881. Nucleo-
philic attack on carbon by the basic site would generate CH,+,
which then interacts with an acid-generated surface methoxy
group [89] to give ether. The observed half-order was ration-
alized on the ground that one of the species, CH,O or CH,, must
be reversibly adsorbed. Primary carbonium ions are much less
stable than alkoxide structures, and therefore equilibrium with
the undissociated species would b e readily established :
H H
H
C-0-H
/ \ ..
H ! * H I
A- -
///?/-, I //// /////?-/$/,
The thermal decomposition of methanol adsorbed on alumina
was investigated by Matsushima and White [go] using deuterium
labeling. Desorption of methanol started at 77OC. At T > 77OC,
significant concentrations of CH,O, H20, and CH,OCH, were ob-
served, and above 427OC, CO was predominant while CH, was
significant. Coadsorption of CD,OD and CH,OH was carried out.
Under desorption at T < 237OC, the ether contained only CH,OCH,,
CH,OCD,, and CD,OCD,, while at higher temperatures deuterium
distribution became random. Thermal decomposition of CH,OHad
in the presence of gaseous CD,OCD, gave only CH,OCH,, indi-
cating that ether is formed by a bimolecular reaction between ad-
jacent surface methoxides.
Schmit z [ 911 studied the dehydration of methanol over silica-
alumina at 289-418OC and found that the reaction becomes first-
order in methanol at the higher temperatures. Interestingly,
Schmitz observed an induction period, possibly the result of
the initial formation of surface carboxylate groups [ 8 4 ] , which,
though not ether intermediates, could nevertheless influence the
initial rate of dehydration by occupying active sites. Another
explanation might involve competitive sorption of product water

prior to establishment of steady-state with respect to surface hy-
droxyl concentration.
Detrekoy and Kallo [ 921 investigated the dehydration of meth-
anol over clinoptilolite by infrared spectroscopy. Methanol ( 6
torr) was adsorbed on H -clinoptilolite at various temperatures.
At 25OC the 3620 cm-' band (acidic OH on clinoptilolite) disap-
pears and two bands at 2950 and 2840 cm-' (CH stretching) ap-
pear. Upon evacuation, the CH bande decreased significantly ;
however, the OH band did not reappear. This indicated that
only weakly adsorbed methanol is removed. Absorption at 16OOC
followed by evacuation caused only a partial disappearance of the
3620 cm-l band and a lower intensity of the 2950 and 2840 cm-I
bands, At 4OO0C, methanol can be completely removed and the
CH bands are shifted to 2860 and 2960 cm-l, indicating the pres-
ence of surface methoxyls.
Detrekoy and Kallo found that methanol dehydration also oc-
curs on dehydroxylated (at T > 400OC) clinoptilolite. They a t -
tribute this to the formation of Bronsted sites from Lewis sites
by hydration during reaction with methanol. The following mech-
anism was proposed :
/Si, /O\ A
, ,l /O\ ,iS/ ,A\
/ CH OH
3
Further reaction would be catalyzed by the acidic OH groups thus

generated.
The dehydration kinetics was also investigated in this study.
An Rideal-Eley mechanism was proposed.
In a study of methanol reaction on synthetic germanic near-
faujasite using I3C-NMR, Derouane et al. [931 observed partial
methoxylation of the surface at 3OO0C, with ether formation.
48 CHANG
Surface methoxyls could be hydrolyzed with water back to meth-

anol at 25OC. Additional small amounts of dimethyl ether and s u r -
face formate were also found after hydrolysis.
In summary, the weight of evidence favors the intermediacy of
surface alkoxyls in ether formation from alcohols. Beranek and
Kraus [ 941 conclude that the mechanism involves essentially a
nucleophilic substitution whereby the surface alkoxide is attacked
by another molecule, either from the gas phase or from a weakly

adsorbed state. Arguments in support of this mechanism were
summarized as follows:
1. "Similar products are obtained by the decomposition of metal

alkoxides containing no B-hydrogens and by the reaction of
corresponding alcohols on alumina at lower temperatures" [ 951.
2. "Acetic acid and pyridine are poisons for the formation of
ethers" [ 961.
3. "The different degrees of water inhibition on the ether and
olefin formation from ethanol on alumina, and the agreement
of ether /ethylene selectivity ratios found experimentally with
those calculated by the Monte Carlo simulation of the hydrated
surface of alumina" [ 971 .
4. ttCorrelation between the rate of ether formation from ethanol
and the surface concentration of ethoxide species. determined
by IR spectroscopy" [ 1601.
5. '%Thepositive value of the Taft reaction parameter for the for-
mation of ether in contrast to negative values for the olefin
formation on the same catalyst" [ 943 .
The, second part of Statement 5 applies only to C2+ alcohols.
B . Hydrocarbon Formation
The mechanism of initial C-C bond formation from methanol is

an unresolved question at present. Hypothetical mechanisms
abound in the literature, and run the gamut from carbene to free
radical schemes. A s of this writing, however, little supporting
experimental evidence has appeared. It seems appropriate, never-
theless, to survey and discuss the diverse entries in this t'mech-
anism sweepstakes.
The speculation of the earliest workers [5, 61 has already been
acknowledged.
1. Via Surface Alkoxyls

Among contemporary investigators, the first to consider this
question were Topchieva and collaborators [ 91 in connection with
TABLE 15
Decomposition of Methyl Derivatives [ 951
C H 3 0 H over (CH3)zO over

Al(OCH3)s at Al,03 at 45OoC, A1203at 45OoC,
Product 38 5OC (mol8) 0.5 LHSV 0.5 LHSV
CH4 22.5 Major Major

H, 35.2 Major Major
co 31.1 Major Major
C2H4 1.3 Minor Minor
C3H6 2.5 Minor Minor
(cH 3 ) 2 0 7.1 Major Major
Other 0.3 Minor Minor
a study of the adsorption of methanol vapor on SiO,, SiO2-Al2O3,

and Al,03 surfaces. It was found that a portion of the methanol
was irreversibly adsorbed on SiO,-Al,O, and A1,0,, which upon
heating and pumping to 4 O O O C evolved C,H,, C,H,, CO. and CO,.
The formation of surface methoxy groups was regarded as the pri-
mary step. Hydrocarbon formation was considered to occur by
condensation of methoxy groups, accompanied by dehydration and
H-transfer. The mechanistic details of this condensation were not
specified. Subsequently, Heiba and Landis [951 showed that the
thermolysis products of aluminum alkoxides are virtually identical
to the products of alumina-cataly zed decomposition of alcohols or
ethers, as shown in Table 15. It is seen that the main products
are CH,, H,, CO, (CH,),O. and smaller amounts of C2H4 and
C,H,. Based on the observation that Al(OCH,Q), decomposed
more readily than Al(OCH,),, it was concluded that cleavage of
the C - 0 bond is a heterolytic process, with the flow of electrons
in the direction of Al, leaving a positively charged carbon moiety
a s the reactive intermediate. A negative activation entropy was
observed, suggesting a cyclic transition state. A free radical
mechanism was rejected on the basis of negligible reactivity over
nonacidic solids, e.g., quartz, at temperatures up to 45OOC.
In a later study the thermal decomposition of methoxides of N a ,
Mg, and A1 and the compound Na[Al(OMe),] was reported by
Pfeifer and Flora [981. Ethene was the only hydrocarbon prod-
u ct ob served.
50 CHANG
2. Carbenes and Carbenoids

Venuto and Landis [ 751 proposed an a-elimination mechanism
to account for olefins formed during methanol dehydration to di-
methyl ether over NaX at 26OoC, a s reported by Mattox 1991, and
from methanol reaction over Rex and ZnX at 330-390°C, as ob-
served by Schwart z and Ciric [48]. According to this view,
methanol adsorbed on the zeolite surface loses water to form a
divalent carbenoid species, which then polymerize to form ole-

fins :
/3
H-CH,-OH -+ H20 + :CH,
u
n:CH, + (CH,),, n = 2 , 3, 4, 5
Swabb and Gates [ 531 studied the dehydration of methanol over

H-mordenite at 155-240OC. Traces of olefin were detected at
240OC. It was speculated that the olefins were formed by an
a-elimination mechanism, where bond scission is facilitated by
cooperative action of acidic and basic sites in the zeolite lattice:
Basic Bronsted
site \
Salvador and Kladnig [ 1001 investigated the surface reactions

of methanol on HY and NaY at 20-350°C using IR, GLC , adsorp-
tion isotherm, and TGA techniques. A t room temperature, physi-
cal adsorption occurs on both zeolites. With HY, methoxylation of
surface hydroxyl groups begins a t 2OoC, reaching a maximum a t
Q
, 130OC. A t 12OoC, dimethyl ether formation begins and reaches a
maximum at 21OoC, and around 25OOC secondary cracking reactions
occur forming predominantly butane and propene. This was ac-
companied by darkening of the catalyst due to coking which was
enhanced with a further temperature increase. Salvador and
Kladnig favored an a-elimination mechanism to explain olefin for-
mation. They differed with the acid-base mechanism of Swabb
and Gates 1531, proposing that the generation of carbenoid spe-
cies occurs by decomposition of the methoxylated surface :
\
JSiOCH, -+ -SOH + :CH,
/ /
HYDROCARBONS FROMMETHANOL 51
Condensation of the carbene would give olefin. Alkanes were a s -

sumed to arise via H -transfer reactions.
An a-elimination mechanism involving a carbenoid intermediate
was also proposed by Chang and Silvestri 121 for methanol reac-
tion over ZSM-5. However, it was considered unlikely the olefins
were formed by polymerization of the diradical intermediate. In
view of the high reactivity of carbenes, the probability of such an
event would be low, as demonstrated in studies on ketene photoly-

sis [ 1013. By the same token, the presence of free carbenes would
b e unlikely. Rather, a concerted reaction between methylene donor
and acceptor was proposed involving simultaneous a-elimination and
sp3 insertion into methanol or dimethyl ether as the primary step.
CHz -+ CH,CH,OR' + R O H , R , R ' = H or alkyl
An ionic mechanism involving methyl cations was rejected since

these species would be expected to form methane readily via hy-
dride abstraction from methanol, dimethyl ether, or hydrocar-
bons. The reaction
CH,+ + CH,OH + &H,OH + CH,,
is extremely rapid in the gas phase [102]. However, methane

normally accou t s for <l% of the hydrocarbons formed over ZSM-5.
2
Note also that H,OH could deprotonate to formaldehyde, which is
not observed over ZSM-5.
Chang and Silvestri suggested that bond dissociation can be
induced b y , or facilitated in the presence of the strong electro-
static field in the zeolite interior [ 103-1051 , the zeolite acting as
a potent ionizing ''solvent .I1 This model prompted a theoretical
study by Beran and Jiru [ 1061 of the influence of electrostatic
field strength on the reactivity of methanol in the zeolite cavity.
The small size of the methanol molecule permitted approximation
of the zeolite field by a homogenous field. Computations were
done by the INDO method. Orientation of the methanol molecule
with respect to the external field is shown below:
52 CHANG
Upon indreasing the field strength from 0-8 V/& the 0 - H , bond
length increases from 1.01 to 1.31 A, the C-0 bond is weakened
by stretching from 1.33 to 1.38-1.40 8 , while one of the hydro-
1).
gens moves toward 0 ( e . g . , 0 - H , decreases from 2.04 to 1.79
Further, with increasing field strength H, becomes more nega-
tively charged while the methyl hydrogens become more positive.
A similar calculation was performed on a pair of colliding methanol
molecules to simulate ether formation. Beran and Jiru concluded

that at 2-4 V / 8 the formation of dimethyl ether becomes possible
while at 4-8 V/8 species such as CH,O and CH, can exist.
The electrostatic field gradient has been estimated to be 1-3
V I A at 2.5-3.0 1 from surface cations, and in zeolites such as
ZSM-5 can be much higher because of structural influences 1741.
Pickert et al. [lo71 calculated a field strength of 6.3 VIA for
points 2 1 from an occupied S,, site in Cay, with SiIA1 = 2.
The presence of small amounts of methyl ethyl ether in the
products of methanol conversion over ZSM-5 was reported by
Chang and Silvestri [21. Cormerais et al. [65] found MeOEt
among the products of dimethyl ether decomposition over silica-
alumina at 423 K . This compound could either be a key reaction
intermediate or simply a secondary product of the methanol-to-
ethene reaction. From kinetic evidence, Cormerais et al. deduced
that this compound was not formed via reaction of Me,O with
ethene. It was determined that in the presence of excess ethene,
amounting to 30X that of the products from Me,O, the rate of
MeOEt formation from Me,O was increased by only a factor of 5.
More significantly, the formation of propene from MeOEt was more
than 10 times faster than from Me,O. In view of the observed un-
reactivity of ethene, it was concluded that propene is formed by
two successive CH, insertions into Me,O, leading to MeOPr, which
cleaves to propene. Another set of experiments gave analogous re-
sults for C, hydrocarbons. Cormerais et al. [65, 1281 proposed a
chain growth "rake mechanism," Fig. 26, to explain their results.
In this scheme, chain growth occurs by carbene insertion into
surface alkoxy species, which are transformed into olefins via
carbenium ions.
Chang and Chu 11081 reported that when methanol is reacted
over ZSM-5 in the presence of propane, the usual high iso-to-nor-
ma1 ratio of product butanes [ 21 is significantly lowered. This is
shown in Table 16 where the butane iln is seen to be 3.8 for the
control experiment and 1.1 when propane is added. The thermo-
dynamic equilibrium i/n is 0 . 7 5 at the reaction conditions. However,
under the same conditions propane and isobutane were virtually in-
ert in the absence of methanol. In the presence of l3CH3OH (90%
l 3 C . 10% " C ) , it was found that the selectivity to singly-labeled
butanes was 30-45 times higher than that expected from random
Me8O
, MeOEt C2H4 MeOPr C,H, MeOBu C4H,...
4 1 4 1 A I
PHASE
PHASE
..It - It It - It I t It
Me2
FIC. 26. "Rake" mechanism for dimethyl ether conversion to

hydrocarbons [ 651.
TABLE 16
Effect of Propane on Methanol Conversion over HZSM-5a [ 1081
(37OoC, 1 atm, 0.4 h-' LHSV (CH,OH))
Hydrocarbon product
(wt %P CH ,OH /He C H ,OH / C ,H8
CH, 0.86 0.90

C2O 0.93 0.90
cz2- 2.35 2.36
(C,O) -c -d
C,2- 3.63 4.53
i-Cbo 35.17 23.14
n-C,O 9.20 20.24
C,2- 1.88 2.34
C5+ aliphatic 21.32 20.98
Aromatics 24.66 24.62
100.00 100.00
i/n c,' 3.8 1.1
'Three moles CH30H/1 mole diluent.

bNormalized on a propane-free basis.
CActual C: = 25.36% of hydrocarbon.
dNet conversion of C 3° % 5%.
54 CHANG
3t
i /n
0I
n=O I 2 3 4
I3C"
FIG. 27. Isolnormal ratio of -labeled butanes [ 1081 .
distribution. It was concluded that propane methylation had oc-

curred. Further, the butane i l n increased with increasing I3C
substitution, as shown in Fig. 27. The iln of the 13C,H, is %2.8,
and reflects the fact that they arise primarily from self-reaction
of I 3 C H 3 0 H . The singly-substituted butanes, mostly the product
of methylation, have iln % l ,close to the equilibrium value. From
these considerations it was deduced that the reactive C , interme-
diate is carbene-like, and the mode of attack is insertion into an
sp3 C-H bond. Such carbene insertions are indiscriminate in liq-
uid phase, all C-H bonds of the substrate being subject to attack
with equal probability [ 1091 :
(75%) (25%)
although for a homogeneous gas-phase reaction there is some evi-

dence of discrimination in favor of secondary over primary C-H
bonds [ 1101. In the CH,OH /C ,He reaction, carbene insertion will
lead statistically to higher concentrations of n-butane relative to
i-butane. Attack by a cationic species such a s C H 3 + o r methyl
oxonium ion according to an Olah-type mechanism (vide infra) , on
the other hand, will yield high i/n butane ratios due to stability
of tertiary carbenium ions in the transition state. In the methyla-
tion of propene with the CH,F-SbF, complex in S0,ClF [ 1111, for
example, CH, is formed via H -transfer, and i-butane is the major
alkylation product. The following pathway has been proposed to
account for i-butane formation :
""'\ H
II
f H - ?, SbF 6
CH3 CH,
The alternative to sp3 C-H insertion, namely C, addition to the

double bond of propene generated from propane via dehydrogena-
tion, also received consideration, Although propane itself is
largely unreactive, it was believed possible that the unfavorable
dehydrogenation equilibrium could be displaced by a "drain-off"
reaction involving addition of a highly reactive species across the
double bond of propene. Carbene addition would yield, classically,
methylcyclopropane [ 1091 . However, subsequent protonation and
rearrangement gives preferentially the tertiary butyl cation, which
will either yield 2-methylpropene via loss of a proton, or i-butane
via H-transfer. If the attacking ~ p e c i e swere cationic, branched
products would similarly result :
56 CHANG
C
/ \
-
C
RH I
C-C-C +R+
+
Another pathway involving propylium ion a s an electrophile can

also be written:
but leads to the same conclusion. This alternative was therefore

rejected as an explanation of the observed results. However, in
the case of reaction of methanol alone, these mechanisms, with
propene as an intermediate, can be invoked to explain the char-
acteristic high i/n butane selectivity.
In another attempt at "trapping" the reactive C, intermediate,
Chang and Lang [1151 reacted methanol over HZSM-5 in the pres-
ence of acetonitrile. It was reasoned that a carbene would insert
mainly into the C-H bonds and to a lesser extent, add to the
C S N group. A cationic intermediate, on the other hand, would
attack the C Z N electron system exclusively, forming a nitrilium
ion [ 173-1751 :
+ +
CH3 + CH,C=N -t CH3C=N-CH,
Upon hydrolysis, N-substituted acetamides would result. This

.
is a Ritter-type sequence [ 1761 Carbene addition to the C G N
group would result in a highly reactive azirine intermediate t 1771,
which would yield the same nitrilium ion in the presence of acidic
catalysts. Thus the sole presence in the final product of N-sub-
stituted acetamides would be evidence of a cationic intermediacy
(although the alternative of acetamide formation via reaction of
amines with acetic acid from acetonitrile hydrolysis cannot be
ruled out). On the other hand, the presence of higher nitriles ,
with or without the acetamides, must be taken a s a strong indi-
cation of carbene involvement, It was also found that acetonitrile,
itself stable over the temperature range of interest, served to

moderate the reaction by competing strongly for the acid sites.
At 454OC an equimolar mixture of methanol and acetonitrile
gave mostly acetic acid and methylamines (36%methanol conver-
sion). This is the result of acetonitrile hydrolysis and methanol
amination. At 496OC ( 79% methanol conversion), hydrocarbons,
CH,CH,CN, and N-substituted acetamides were in evidence, as
well as CH,COOH + CH,COOCH,, CH,CH,COOH, and methylamines.

Upon raising the temperature to 538OC, the reactor effluent con-
tained 16.3%hydrocarbons and 7.0% 0- and N-compounds (exclu-
sive of unreacted acetonitrile). These 0- and N-compounds con-
sisted of 12.8% CH,COOH + CH,COOCH,, 46.6% CH,CH,CN, 37.2%
N-methyl-w-aminonitriles, and 3 . 4 % higher esters, pyrroles, etc.
The CH,CH,CN is taken as evidence of carbene insertion,
while the N -methyl-w-aminonitries may result from an analogous
nitrenium insertion [ 1781 where the nitrenium ion is generated
from a methylamine precursor:
+
R,R,R,N * R1R2N: + :R,
H
+ PI
RIRZN: + CH3CN * R,R,N :CH,CN --H+
RlR,NCH*CN
The reaction of methanol in the presence of zinc iodide at 2 O O O C

was reported by Kim et al. [60] to yield highly branched hydro-
carbons. Particularly high selectivity to triptane was observed.
They postulated that the intermediate is a carbene complexed with
the salt, similar to the Simmons-Smith reagent (CH,I, + Zn(Cu)) ,
which reacts with alkenes to form cyclopropane derivatives via
carbene addition to the double bond.
3. Oxonium Ions and Ylids
In their mechanism Chang and Silvestri left open the question

of stabilization of the intermediate carbene in the transition state.
A plausible resolution of this question may be found in the mech-
anism of van den Berg et al. [ 1121. In their view, dimethyl ether
from methanol dehydration reacts with a Bronsted acid site to form
a dimethyloxonium ion I , which reacts with another molecule of di-
methyl ether to form, via 11. and after elimination of methanol, a
trimethyloxonium ion 111:
58 CHANG
H+
/O\ /O\-,O\
si
I1
A1
--CH30H
I11
The critical step in their proposal is a Stevens-type intramolecu-

lar rearrangement of the trimethyloxonium ion 111 to a methyl
ethyloxonium ion V :
4 0
111 IV V
N
VI VII
~ ~~~ ~- ~
Structure V I , which is equivalent to the oxonium ylid (CH,),O+CH,-,

is seen to contain the stabilized C, carbenoid species. Cis-inser-
tion of the carbenoid into the adjacent C - 0 leads to C -C bond for -
mation. The formation of ylid depends on the assumption that the
conjugate basic sites in ZSM-5 a r e sufficiently strong to induce po-
larization of a C-H bond on a methyl group. A s indicated pre-
viously, the ab initio field calculations of Beran and Jiru [ 1061
lend support to this assumption.

An alternate cationic mechanism involving carboxonium ions was
considered by van den Berg "741. In this variation t h e carbox-
onium ion is generated by hydride abstraction. Van den Berg
presented the following scheme comparing the steps in the alkox-
onium Route ( A ) and the carboxonium Route ( B ) .
CH,-O-CH,
+ 1
Hf
"i
CH3-0-CH,
+
+ H,
A2/
B21 +
CH ,CHOH
B3 i-CH,OCH,
+ 4 +
C H 3 0 C H z C H 3 + CHBOH, CH,CH,OH + CHSOCH,
~~l
CH,OH,+ + C,H, + CH,OH
B41
CZH, + HZO + CH,OCH,
+
Route A Route B
The overall gas phase enthalpies of reaction of the various steps

were calculated and are shown in Table 17.
A major difference is found between Steps A 2 and B2. How-
e v e r , ab initio calculations [ 1131 indicate that although the 1-
hydroxyethyl cation is more stable than the methoxymethyl cat -
ion, the energy barrier separating the two is on the order of
60 CHANG
TABLE 17 [32]
The Conversion of Methanol and Ethanol to Hydrocarbons
over ZSM-5 Class Zeolitesa
Methanol Ethanol
Percent age conversion 99+ 99

Hydrocarbon distribution (wt 8) :

C1 1.2 0.1
c2 0.8 1.8
6.0 7.6
12.4 17.3
C 5+ nonaromatic 44.8 40.2
Aromatics 34.8 33.0
Wt%durene in aromatics 58.8 0.9
Wt%ethylbenzenesb in aromatics 0.6 4.1
a
T = 370 k l0C ; P = 50 atrn; LHSV = 1 . 2 h - l .
bEt benzene, MeEt benzenes, diMeEt benzenes, and diEt ben-
zenes.
260 kJJmo1,- assuming that the rearrangement proceeds through the

0-protonated oxirane :
This is comparable to the expected energy barrier of Step A 2 .

Step A 3 is highly exothermic and, in view of the behavior of the
N-analogue [ 1141, is expected to have a low activation energy.
Figure 28 is an energy diagram comparing Routes A and B. It
was concluded that Route A, involving the trimethyloxonium ion,
is favored.
According to van den Berg [ 741, kinetics of dimethyl ether
reaction over HZSM-5 is zero-order at 227-300OC. Chang and
HYDROCARBONS FROM METHANOL

61
62 CHANG
Lang [ 1151 found the reaction order of methanol decomposition

over HZSM-5 at 371OC to be zero-order up to about 60%conver-
sion, and increasing in order at higher conversions, suggestive
of Langmuir-Hinshelwood behavior. This is consistent with van
den Berg's proposal that the formation of adsorbed alkoxyoxonium
species is favored, with the C-C bond formation as the most de-
manding step. However, van den Berg observed an anomalous
temperature effect, illustrated in the Arrhenius plot of Fig. 29.

To explain this phenomenon it was proposed that at 227-26OoC,
C -C bond formation proceeds via the intramolecular rearrange-
ment of trimethyloxonium ions while at higher temperatures a con-
certed reaction between dimethyl ether and alkyl cations, with
the oxonium ion as the transition state, becomes significant.
It should be noted that ylid mechanisms have been earlier pro-
posed to explain C-C bond formation from organic N and S com-
pounds in the presence of zeolites. One such example is the for-
mation of stilbene from benzyl mercaptan over Nal3X reported by
Venuto and Landis [ 1161. Although these workers preferred an
a-elimination mechanism for this reaction, they recognized that a
sulfonium ylid mechanism is also possible. Wu et al. [ 1171 stud-
ied the t her ma1 decomposition of met hylammonium cation exchanged
Y-type faujasites. At 275-45OoC, ethene was observed among the
products of decomposition. To account for this result, a mechan-
i s m involving a Stevens rearrangement followed by Hofmann elim-
ination was proposed
In another example (not involving zeolites) reported by Lepley

and Giumanini [ 1181, the thermolysis of N ,N-dimethylpyrrolidinium
bromide gives N -ethylpyrrolidine by carbene insertion :
A m o +H'
/ \ I
CH, CH3 CH~CH~
The Stevens rearrangement of oxonium compounds has not been

reported up to this time. However, the insertion of CH,, generated
F I G . 29. Arrhenius plot of the dimethyl ether conversion on

zeolite HZSM-5 [74]. P, = 5 0 . 7 kPa; WHSV = 0.72 h-'.
by ketene photolysis, into C-H bonds of alkyl ethers is believed

to occur via an ylid intermediate [ 1091. N o insertion into alkyl
ether is observed when the rnethylene is generated by mercury
photosensitization, indicating that singlet methylene is involved
in the attack [ 1091. Methylene generated from diazomethane b y
cuprous halide -catalyzed decomposition was similarly ineffectual
[ 1091. However, when the substrate contains an electron-with-
drawing group, reaction with diazomethane occurs, a s in the fol-
lowing reaction reported by Meerwein et al. [ 1681 :
Methanol conversion over phosphorus-modified ZSM- 5 was stud-

ied by Kaeding and Butter [33], who proposed the following mech-
anism based also on oxonium intermediates :
H
@/ 8
CH,OCH, + HZeol S CH3-0, f Zeol
64 CHANG
CH3-0,
i
' -
'\
'CH,
CH,OH + CH,CH2-OCH, + HZeol
Zeol CH,-0-CH,
0
H
I
CH,OCH,CH, + HZeol * CH3-0-CH2CH3Zeol 8
H
I
CH,-0-CH,CH,-HZeol
t/ .u
m e
- CH,OH t CH2=CH2 t HZeol
The key step in this mechanism is the labilization of a C-H bond,

assisted by anionic sites in the zeolite. The resultant nucleophilic
species then attacks the Yncipient methyl carbonium ion" from pro-
tonated methyl ether (or methanol).
Recently, Perot et al. [ 1271 carried out a 13C tracer study of
dimethyl ether conversion over silica-alumina and HZSM-5. A mix-
ture of ( 12CH3)20 and ("CH,),O was reacted at 35OOC. The iso-
tope distribution in the product ethene (at ~ 6 0 conversion)
% was
found to be random. From this it was inferred that olefin forma-
tion was intermolecular. In agreement with van den Berg et al.
[ 1121 , it was proposed that the mechanism involves rearrange-
ment of a trimethyloxonium ion which is formed by a bimolecular
reaction between dimethyl ether and methanol or with another
molecule of dimethyl ether. However, data at low conversion
were not given to confirm this interpretation.
4. Carbenium Ions and Superacid Chemistry

Mechanisms based on electrophilic attack by methyl cations at
saturated carbon are favored by some workers. Such reactions
were first observed in superacid media by Olah and co-workers
[ 119- 1211 and are exemplified by the remarkable reaction of
methane :
The transition state is believed to be a pentacoordinate carbonium

ion, typified by the CH 5+ methanonium ion shown above, and in
the following substitution reaction :
R3C-H + CH3+ [R3C-+:l

H
CH 3
1+
R3C-CH3 + HS
Pearson [58] proposed this as a possible mechanism of C-C

bond formation in methanol decomposition by phosphorus pent -
oxide [6].
Based on their observation that methanol converts to hydro-
r
carbons over heteropolyacids and Nafion-H (perfluorinated sul-
fonic acid resin), which are Bronsted acids, On0 and Mori [ 711
concluded that the mechanism involves methyl cations:
+
CH, + + CH30R + \l>-cH30R] + CH,CH,OR + H+
CH:
Generation of the methyl cation was considered to occur by po-
larization of surface methoxyl species:
YH 3 CH,’
0 0 0
\ \ /O \ /O\ /O
0’ A1
\o 0 ’
Si
‘
0 0
’
A1
‘00’
Si
‘
0
66 CHANG
Autocatalysis [ 691 involving condensation of methanol and olefins

was cited as the reason why little methane is produced via hydride
abstraction.
Ono et al. [70] rejected the posshility of carbene involvement.
They found that HC1 did not poison the methanol reaction over
HZSM-5. They reasoned that since HC1 poisons base-catalyzed
reactions, basic sites do not participate in the reaction. They
assert that these results "plead against the carbene mechanism,

in which the abstraction of a proton from a methyl group by basic
sites is essential. I'
The Olah superacid mechanism was also favored by Kagi [ 1221,
who proposed a series of oxonium intermediates, e . g . ,
However, Chang [ 1231 considered the bulky transition states

arising from such reactions improbable in ZSM-5 on account of
limitations due t o channel size. +
I t should be noted that neither CH,' nor C H , have thus far
been directly observed in superacid solutions [ 124, 1251. It has
also been pointed out [ 1261 that SbF, and SO, a r e strong oxidizing
species and could generate carbenium ions from alkanes by oxida-
tion, although Olah points out that similar carbocation transforma-
tions are observed in systems such a s HF-TaF, and HF-BF, [ 1251,
which have high redox potentials, and therefore would not be po-
tent oxidizing media,
Salem [ 1691 has reported the conversion of methanol and di-
methyl ether to hydrocarbons over TaF, and NbF,. At 3OOOC in
an autoclave, a twofold excess of methanol was converted to a
mixture of paraffins and aromatics. The NbF,-catalyzed reac-
tion gave more light hydrocarbons (>81% C1-C 5> than the TaF,
reaction ( 5 3 %C,-C,). It was not clear whether the TaF, and
NbF, were stable to hydrolysis under the reaction conditions
or deactivated during the course of reaction.
The conversion of methanol over the alkylammonium zeolite

Nu- 1 was investigated by Spencer and Whittam [ 551. They con-
cluded that strong acidity is needed for t h e fast initial step and
that cationic intermediates are involved. However, neither CH ,+
nor oxonium ions were considered to be plausible.
5. Chain Mechanisms
Anderson et al. [67], using deuterium labeling, found that in

the methylation of benzene with methanol over HZSM-5 a t 207OC,
there is no exchange involving methyl hydrogens. From this it
was inferred that in the methanol-to-hydrocarbon reaction, the
initial step does not involve lability of the C-H bond prior to
C -C bond formation, and that therefore the reactive interme-
diates a r e not carbene, carbenoid, o r oxyme hylene species. In-
b
stead, a active C, entity of the type (CH,)R H (R = H or CH,)
was proposed, which participates in a chain propagation mech-
anism
active C , + reactive Cn -+ reactive Cn+,
where Cn is an olefin. This is a highly attractive scheme in

that it is consistent with the observed autocatalysis of the r e -
action [ 6 9 ] . In such a scheme, moreover, the precise nature
of the C, species becomes somewhat of a moot question. How-
ever, a s has been noted [ 1231, the temperature a t which these
D-labeling experiments were carried out was below the threshold
for hydrocarbon formation from methanol in the presence of
ZSM-5. Thus their bearing on the mechanistic question may
be somewhat tenuous. The results of Matsushima and White
[90] on deuterium exchange between CH30H and CD30D over
alumina have already been mentioned. It will be recalled that
these authors found little D-exchange below 237OC but complete
scrambling at higher temperature (T 2 397OC).
The "rake mechanism" of Ahn et al. [128] and Cormerais et
al. [ 6 5 ] has already been mentioned (Fig. 2 6 ) .
6. Other Schemes
The conversion of methanol to C ,-C hydrocarbons over nat-
ural mordenite was reported b y Zatorski and Krzyzanowski [ 541,
who proposed a free radical mechanism:
CH,OH -+ 6H, + OH
68 CHANG
CH, + C,H, + CH, + CH,CH,
CH, + CH,CH, + C,H,
2CH,CH, -+ C,Hlo
CH3CH2 -+ C2H4 + H
N o supporting evidence was given. In this context it may be rele-

vant that the Stevens rearrangement has b e n known to give rise
to C I D N P spectra, indicating some degree of radical involvement.
A possible mechanism is dissociation into a radical pair within a
solvent cage [ 1661 :
R' R
+
\..N - C I /
\ \ / \
Lin et al. [ 1291 studied methanol conversion over tungsten ox-
ide on various supports including zeolite A , rnordenite, and cha-
bazite. A mechanism involving the formation of a "HPC-O site"
species as the rate-controlling step was proposed. "These spe-
cies then migrate to sites where the external bonding between
the sites and the oxygen is stronger than the internal bonding
between C+ and 0-. Divalent specie C H , is then formed and sub-
sequently polymerized to produce the olefins." (It was not speci-
fied how charge conservation is to be preserved.) The tungsten
apparently has a dehydrogenation function.
Langner [ 1701 has proposed the following imaginative mechan-
>C=C< + -
CH30H
(H+)
H<C-(-H + CH20
0 + CH20 -0(H*) CH20H
uCH3+ CH3OH __c(H+) CH20

+ OCH3
CH30H CH3 +2CH30H

.....
-HzO * a C H 3 -2H20)
-
(H+l ( H*l CH3
(6) ISOBUTENE, ISOPENTENE, PROPYLENE

+ CYCLOOLEFIN
-0
CH3
3 PROPYLENE (H*) + PROPANE
The critical steps are H-transfer from methanol to double bonds,

producing formaldehyde (Steps 1, 4, 5) ; the Prins condensation
of formaldehyde arid cycloolefin (Steps 2 , 5) ; dehydration of the
hydroxymethyl substituent , followed by double-bond isomeriza-
tion to the methyl derivative (Steps 3, 5); and generation of ole-
fins from the polymethyl cyclic intermediate via the "paring re-
action" [ 1 7 1 ] (Step 6). The initial cycloolefin is generated ac-
cording to Step 7.
The formation of methane deserves comment.
According to Wu et al. [ 1171, the formation of methane, C O ,
and H, from zeolite surface methoxy species may proceed by a
"methyl disproportionation" reaction which produces formalde-
hyde as an intermediate. An intermolecular hydride transfer is
envisioned :
Zeol
This reaction may also lead to coke formation:
HCHO -f C + H20
Venuto and Landis [ 751 considered methane to be formed from
methanol via radical pathways.
Finally, hydride abstraction by the reactive C, species can-
not be discounted as a source of at least some of the methane,
usually present in minor amounts. A s pointed out by Ono et al.
[ 701 , autocatalysis will minimize this reaction. However, hydride
70 CHANG
abstraction could be significant during the induction period of

such autocatalysis and account for methane a t low conversions.
7. I s Ethene the Initial Olefin?

Associated with the problem of C-C bond formation is the
question of the nature of the "firstttolefin formed. This ques-
tion has generated some controversy. Chang and Silvestri [ 2 1 ,
Derouane et al. [ 261, and van den Berg [ 741 considered ethene
to be an active intermediate in the reaction sequence. Consis-
tent with this view, Jiru et al. [ 1301 found that ethene-H,O mix-
tures gave products identical with methanol over ZSM-5. Simi-
larly, Bragin et al. [ 3 8 ] compared t h e reactions of ethene and
methanol over ZSM-5 under various conditions and concluded,
on the basis of similarity of product distributions, that ethene
is a primary intermediate in methanol transformation.
According to Kaeding and Butter [ 331 , ethene is the initial
olefin a t low methanol conversion in the presence of phosphorus-
modified ZSM-5. They assumed that the phosphorus treatment
did not alter t h e catalytic mechanism.
The view that ethene is the initial olefin was challenged b y
Ahn et al. [ 1281 and Anderson et al. [ 271. They found ethene
t o be substantially less reactive than propene over a variety of
catalysts. This was regarded as evidence that ethene could not
be a primary intermediate.
The catalysts investigated b y Ahn et al. [ 1281 were silica,
alumina, silica-alumina, HY, and H-mordenite. At 35OoC, over
silica-alumina, although the contact time of ethene was 10 times
higher than methanol, less than 0 . 5 %ethene was reacted, while
methanol gave 20-25%hydrocarbons. Further, the rate of pro-
pene formation from ethene was 100 times slower than from meth-
anol.
Anderson et al. [27] utilized various ZSM-5 preparations in
their study. A t 4OOOC propene was readily converted to aro-
matic products, while ethene generally exhibited low reactivity.
However, it was observed that certain catalyst preparations
showed considerable activity in promoting ethene conversion.
It was speculated that these particular catalysts may possess
highly active sites which are easily poisoned.
Ceckiewicz [ 431 , in FTIR studies of methanol sorption on
zeolite H-T a t 25-450°C, observed the propene band pattern
but failed to detect ethene. Upon temperature-programmed de-
sorption, however, C,-C, olefins were identified. The follow -
ing reaction network with propene a s the initial olefin was
proposed :
[methanol 1
cycli zat ion cycloparaf fins
ads
ads ’ aromatics
ads
cracking
4
C,-C, olefinsg carbonaceous
propyleneg and paraffinsg deposit
ads
In this scheme, Region A represents static conditions (400°C,

~3 t o r r ) , and Region B dynamic conditions (flowing He, 25-
4OO0C, ~ 3 tom) .
0
Contrasting results were reported b y Novakova et al. [1311
who investigated the sorption reaction of ethene on HZSM-5
and HNaY zeolites by IR and MS, Oligomerization was observed
on both catalysts a t temperatures a s low a s 37OC. Upon heat-
ing to 97-197O. C3-C7aliphatic hydrocarbons were evolved; at
197- 2 9 7 O , alicyclic hydrocarbons; and at 327-427O, methyl-sub-
stituted aromatics were evolved. With HNaY , saturated hydro-
carbons predominated, while HZSM-5 gave mainly unsaturates.
Water vapor did not affect the ethene reaction with NaHY; how-
ever, with HZSM-5, water vapor facilitated the desorption of
products, especially aromatics. Also in the case of HZSM-5,
the presence of water altered the ratio of saturated to unsat-
urated hydrocarbons in favor of saturates. The different b e -
havior of HZSM-5 was attributed to the following possible fac-
tors: (a) different electrostatic field, ( b ) high Si:A1 ratio, (c)
crystal structure and pore geometry.
Bolis et al. [ 1321 studied the sorption activation of ethene by
HZSM-5 using TGA, IR, and 13C-NMR. They found that ethene
is protonated (rate-determining step) on the Bronsted acid sites
even near room temperature. Depending on zeolite pretreatment
conditions, competitive sorption between water and ethene was
72 CHANG
observed. Physisorbed water inhibited ethene sorption. This

could explain the apparent low ethene reactivity reported by
others.
According to Dejaifve e t al. [ 791, the reactions of C,H,+(ads)
are faster than i t s desorption as ethene, and desorption is faster
than chemisorption of ethene from the gas phase.
Van den Berg [ 741 carried out detailed investigations on the
sorption of light olefins in HZSM-5 using TGA and 13C-NMR. At
room temperature an average of 4 . 3 molecules was adsorbed per

Al-atom in the zeolite. Assuming that each Al-atom represents
an active site, and based on the shape of the sorption isotherms,
it was concluded that the chemisorption of olefins on HZSM-5 is
an oligomerization process occurring readily even a t room tem -
perature. This was confirmed by 13C-NMR which revealed that
the ethene oligomer had an average chain length of 25 carbon
atoms. In agreement with Novakova et al. [ 1311, van den Berg
observed an inhibitory effect of water vapor on ethene reactiv-
ity. However, at temperatures used in methanol conversion to
gasoline (227-377OC) the sorption equilibrium between water and
the Bronsted acid sites will be shifted toward desorption, allow-
ing ethene chemisorption and reaction. These considerations sup-
port the view that ethene can act a s a primary intermediate.
In conclusion, the results of Novakova [ 1311 and van den Berg
[ 741 leave little doubt that ethene is extremely reactive in the
presence of zeolites. Water may inhibit or retard the reaction
a t low temperatures, especially in ZSM-5 zeolites. With most un-
modified zeolites, methanol conversion gives product ethene to
propene ratios less than unity. The most satisfactory explana-
tion for this, in light of the above-mentioned ethene studies, is
that of Dejaifve et al. [ 791, based on desorption-adsorption of
ethene as slow steps relative to reaction of the adsorbed species,
C2H5+.
8. Origin of the Polymethylbenzenes

It has already been noted that polymethylbenzenes are distinc -
tive products of methanol conversion to hydrocarbons over a vari-
ety of catalysts. Their origin is discussed in the following pages.
A series of experiments at different flow rates and pressures
[66] ruled out the possibility that the degree of ring substitution
was simply a consequence of increased contact time brought about
by increased pressure. This is illustrated in Fig. 30 which plots
aromatic chain-to-ring ratio, defined a s (alkyl C)/(aryl C ) ,
against contact time. The plot demonstrates that chain-to-ring
ratio is parametric in pressure.
700 OF
0.61 I I I I I I ~ I I I I I ~
I I I I ~ I I ~ I I I I I I I I I I I I I ~ ~ I ~ II 1 1 1 1 1 1 1 1 1 1
2 -
2a 0.5- -
u -
I
[ r -
353 psig
169 psig
0
!-
I - + 59 psig -
$
V
0.4-
0
- 0 -
0 n
w 0 psig -
- 0 -
0.3- -
- -
I I11111111111 I I11111111111 I I 1 1 l 1 1 1 1 1 1 1 1 I I I I I I I I I L I I .
Subsequent experiments indicated that the increase in poly-

methylbenzenes with pressure is correlatable with the degree of
overlap of the methanol/DME reaction curve and the aromatiza-
tion curve in the reaction path, Figs. 8 and 12. The shaded por-
tion in these plots denotes this region of overlap, wherein ring
methylation by unconverted methanol /DME was assumed to occur
[ 321. The overlap region is seen to increase with pressure.
Evidence in support of this mechanism was twofold. First, it
was shown [32] that when ethanol was reacted under pressure,
the durene selectivity was exceedingly low a s compared with
methanol feed at the same conditions (Table 17). Second, that
the alkylation mechanism was kinetically feasible was shown in
a study 11151 comparing plug-flow and backmixed reactor per-
formance for methanol conversion over HZSM-5. With 708 back-
mixing, ring methylation increased as illustrated in Fig. 31,
which indicates a shift in distribution toward the higher aro-
matics. This is the anticipated result of the influence of back-
mixing on the selectivity of a multistep sequential reaction.
74 CHANG
"1 PLUG FLOW

70% BACKMIXED FLOW
CARBON NUMBER
FIG. 31. Normalized aromatics distribution [ 1151 .

VI. THE MOBIL METHANOL-TO-GASOLINE

(MTG) PROCESS
The conversion of methanol to aromatic hydrocarbons over

ZSM-5 zeolites is the basis of Mobil's MTG (methanol-to-gaso-
.
line) process [ l ] Two versions of this process, the fixed-bed
and fluid-bed, have been extensively investigated in bench-
scale and pilot -plant studies.

The fixed-bed process has been selected by the government
of N e w Zealand to convert natural gas from their extensive Maui
field to gasoline via methanol [ 1331 . The plant will produce
13,000 barrels of gasoline per day. The fluid-bed process is
scheduled for demonstration a s a lOO-bbl/d unit in West G e r -
many 1133, 1451.
A s noted previously, hydrocarbon formation f r o m methanol
is a highly exothermic process. The heat of reaction varies
with product distribution. At 4OOOC it is 1510-1740 k J / k g of
methanol converted, which translates to an adiabatic tempera-
t u r e rise of about 650OC [ 661. Control and dissipation of this
large reaction heat is a major constraint in reactor design. Al-
though discussion of reactor design is outside the scope of this
survey, it may be noted that the Mobil team considered a variety
of configurations, including tubular heat exchanger reactors
11343, staged fixed-bed reactors 1135, 1361, and fluid-bed r e -
actors [ 137- 1401. A fixed-bed process with light gas recycle
and the fluid-bed routes were selected for the initial process
development studies.
A. Fixed-Bed Process
The fixed-bed unit [30, 66, 1411, shown in Fig. 32, was a
two-stage system consisting of a dehydration reactor, where a
near equilibrium MeOH /DME /H20 mixture was produced, followed
by the hydrocarbon-forming reactor containing H ZSM- 5. The
second-stage product was cooled and flashed in a high-pressure
separator. Light gases were recycled to the second-stage reac-
tor to control the temperature rise. In this scheme, 208of the
total reaction heat is released in the first -stage and the balance
in the second. Depending on the recycle ratio, a AT < 95OC
could be maintained in the second stage.
Initial operating conditions for a bench-scale version of the
dual reactor system are shown in Table 18 [66]. The feed was
methanol containing 17%water to simulate crude product from a
methanol plant.
Results of a 167-h test a r e shown in Tables 19 and 20. A s
the cycle progresses, C,-C, paraffins decrease, C 7 - C 2 olefins
76 CHANG
~~
Fixed Bed
H20'-h
Crude
Methanol
"'' ---
Dehydration
Reactor
Compressor
Gas Product
+ 334%
HydrOca rbon
Recycle Liquid
Conversion Gas Product
Reactor
I
Water
408'C
FIG. 32. Mobil methanol-to-gasoline process [661.
TABLE 18
Operating Conditions for Dual Reactor Fixed-Bed Process 1661
First Reactor
Temperature, inlet 316OC
Temperature, outlet Q399OC
Pressure, inlet 13.6-23,8 atm
Space velocity 20 h - l , WHSV
Second Reactor
Temperature, inlet 343oc
Temperature, outlet Q454OC
Pressure, inlet 1 3 . 6 - 2 3 . 1 atrn
Space velocity 1.5-5.0 h - l . WHSV
Molar recycle ratio 3-4 (based on total fresh feed)
SeDarator
Temperature 37.8OC
TABLE 19 [66]
A Fixed Bed Aging Study. Run Conditions, Material Balance,
and Gasoline Propertiesa
Time on stream ( h )
8 47 95 142 167
Run conditions:
Reactor 1 inlet tern- 626 625 624 625 625
perature, OF
Reactor 1 outlet tem- 77 1 768 769 768 735
perature, OF
Reactor 2 inlet tem- 648 648 663 665 660
perature, OF
Reactor 2 outlet tem- 862 8 57 888 891 8 19
perature, OF
WHSV (MeOH over 3.3 3.3 3.3 3.3 3.3
catalyst
Molar recycle ratio 3.6 3.8 3.4 3.6 3.6
Pressure, psig 26 5 26 5 265 26 5 265
Product, wt% of total
product :
Hydrochrbons 36.28 35.75 37.65 36.11 34.44
H2 0 63.54 64.13 62.18 63.82 65.20
co 0.09 0.04 0.08 0.02 0.04
CO2 0.09 0.08 0.09 0.05 0.05
Methanol 0.00 0.00 0.00 0.00 0.27
Dimethyl ether 0.00 0.00
- - 0.00 -0.00 -0.00
100.00 100.00 100.00 100.00 100.00
Product, wt % of feed 99.72 100.45 101.69 99.76 100.22
Gasoline (including
alkylate) :
Reid vapor pressure, 9.0 9.0 9.0 9.0 9.0
p sia
Research octane, un- 96.2 95.0 94.3 94.6 94.2
leaded
Research octane, 3 g 101.0 100.3 100.2 100.3 100.3
lead
Specific gravity 0.746 0.738 0.736 0.729 0.726
Molecular weight 94.3 94.4 94.2 93.8 93.8
a
Feed crude methanol ( 17 wt%H,O) ; unit ; process development
unit.
78 C HAN G
TABLE 20 [66]
A Fixed Bed Aging Study. Hydrocarbon Product Distribution.
Basis : Weight Percent
Time on stream ( h )
8 47 95 142 167
Methane, ethane, ethylene 2.4 2.0 2.1 1.5 2.0

Propane 11.0 7.7 7.6 5.6 4.9
n -Butane 5.2 4.0 3.9 2.9 2.6
I sobu t ane 10.2 9.5 9.5 8.7 8.6
C 3 - C 4 Olefins 2.5 2.7 3.4 4.2 3.7
C 5+ Nonaromatics 35.0 42.1 42.4 48.3 50.8
Aromatic s 33.7 -
- 32.0 -
31.1 -
28.8 -
27.4
100.0 100.0 100.0 100.0 100.0
c 5+ 68.7 74.1 73.5 77.1 78.2
9 RVP gasoline including 75.9 81.9 82.8 88.2 88.4
alkylate
Durene 2.5 4.2 4.3 4.2 4.5
+
increase, C nonaromatic hydrocarbons increase significantly,
and aromatics remain fairly constant.
For the final gasoline product, alkylate could be produced
from the isobutane and C , - C I , olefins, so that the total of R V P
gasoline selectivity was 75.9-88.4 wt%. This product had an un-
leaded research octane of 94.2-96.2.
In a subsequent test [ 1421 the space velocity was lowered from
3 . 3 to 1.6 and the recycle ratio increased from 3 . 4 - 3 . 8 to 9. A
total of 208 d on stream was achieved, during which 8000 lb meth-
anol/lb catalyst was processed. The total test consisted of nine
cycles, varying in length from 1 3 . 1 d (first cycle) to 24.9 d
(Table 21), with oxidative regenerations between cycles to re-
move accumulated coke on catalyst. Typical data, shown in
Tables 22 and 23, are similar to the data in the preceding tables.
X
TABLE 21 11421
Cycle Lengths of Long-Term Aging Test in Fixed Bed Unit
Cycle length to MeOH

breakthrou gha Total cycle length Cumulative time on stream
MeOH proc., MeOH proc., MeOH proc.,
lb of MeOH/ lb of MeOHl lb of MeOHI
Run no. Cycle lb of cat. 2 Time ( d ) lb of cat 2. Time (d) Ib of cat. 2 Time (d)
212-48 1 50 1 13.1 595 15.5 595 15.5

212-49 2 982 24.9 1105 28.0 1700 43.5
212-50 3 643 16.9 795 20.8 2495 64.3
212-51 4 695 18.0 808 21.0 3303 85.3
212-52 5 607 15.8 708 18.5 4011 103.8
212- 53 6 96 3 24.9 1046 27.0 5057 130.8
212-54 7 810 21.1 895 23.3 5952 154.1
212- 55 8 9 18 23.9 1048 27.3 7000 181.4
212- 56 9 920 24.2 1006 26.2 8006 207.6
a
Methanol breakthrough taken as first material balance during which MeOH was detected in product
water phase.
TABLE 22 11421
Operating Conditions and Products for Selected Cycles in Long-Term A g i n g Test
Cycle 1 Cycle 6 Cycle 9
Conditions
T h e on stream in cycle h 73 314 94 333 77 315
Cumulative charge, w t of MeOH/ 117 501 4162 4545 7123 7504
w t conv. cat.
Dehydration reactor inlet temp., OF 601 604 583 586 576 58 7
Dehydration reactor outlet temp., OF 779 776 773 771 771 774
Conversion reactor inlet temp., OF 648 652 649 652 650 651
Conversion reactor outlet temp., OF 774 777 766 774 778 779
High pressure separator temp., OF 128 129 125 135 126 124
WHSV, MeOH over conv. cat. 1.6 1.6 1.6 1.6 1.6 1.6
Recycle ratio, mol/mol of charge 9.2 8.9 9.7 8.8 9.2 8.3
Pressure, psig 300 300 300 245 283 250
Products, wt 0 of charge :
Hydrocarbons 35.5 35.3 36.2 35.2 36.9 35.9
HZO 64.1 64.2 63.8 64.7 63.0 64.0
HZ 0.01 0.01 0.0 0.0 0.0 0.0
co 0.04 0.03 0.01 0.01 0.01 0.0
CO, 0.4 0.3 0.0 0.05 0.08 0.07
Methanol 0.0 0.2 0.0 0.0 0.0 0.0
Dimethyl ether 0.0 0.0 0.0 0.0 0.0 0.0
Material balance, wt 90 98.4 97.8 98.2 98.6 100.2 99.0
9 RVP gaso1ine:a
Research octane (R + 0) 95.1 93.3 93.4 93.2 93.2 93.0
Leaded octane ( R + 3) 100.8 99.4 99.9 99.3 99.9 99.1
Specific gravity 0.756 0.734 0.736 0.728 0.731 0.723
Molecular weight 95 95 95 95 94 94
Yield, w t % of hydrocarbons 73.3 86.2 82.6 92.0 85.5 89.3
a
Gasoline includes alkylate ; properties are calculated from compositions.
03
N
TABLE 23 11421
Hydrocarbon Product Distribution in Long-Term Aging Test
Cycle 1 Cycle 6 Cycle 9
Time on stream in cycle, h 73 3 14 94 333 77 3 16

Cumulative charge, wt of MeOH/wt of conv. cat. 117 50 1 4162 4545 7123 7504
Hydrocarbon composition, wt %:
Methane 2.1 1.8 0.9 0.7 0.9 0.8
Ethane 1.3 0.4 0.6 0.2 0.4 0.3
Ethene 0.0 0.0 0.0 0.0 0.0 0.0
Prop m e 11.6 4.6 6.6 3.0 5.3 3.9
Propene 0.1 0.2 0.2 0.2 0.2 0.3
n -Butane 5.9 2.6 3.7 2.0 3.1 2.3
Isobutane 8.5 8.5 9.3 6.4 8.9 8.0
Butenes 0.5 1.1 1.0 1.1 1.2 1.4
Total C k 30.0 19.2 22.3 13.6 20.0 17.0
n -Pentane 2.7 1.3 1.7 1.2 1.6 1.2
Isopentane 10.1 11.3 11.4 10.1 11.2 10.6
Pentenes 0.7 2.4 1.9 2.8 2.3 2.9
C yclopentane 0.3 0.2 0.3 0.3 0.3 0.3

C6+ Nonaromatic 20.0 38.i 31.5 43.6 35.3 41.1
Benzene 0.4 0.2 0.2 0.2 0.2 0.2
Toluene 4.4 1.7 2.0 1.8 2.1 1.8
Ethylbenzene 0.9 0.5 0.6 0.6 0.6 0.6
p- and m - X y l e n e s 9.5 6.7 7.4 7.1 7.2 6.2
o -Xylene 2.7 1.7 2.0 1.8 2.0 1. 7
Trimethylbenzenes 8.4 7.3 8.2 7.3 7.3 6.5
Methylethylben zenes 2.3 2.5 2.8 2.8 2.8 2.6
Propylben zenes 0.1 0.1 0.1 0.2 0.2 0.3
1,2.4,5-Tetramethylbenzene 3.6 4.2 4.2 3.6 3.2 3.2
1 , 2 , 3 , 5 - T e t r a m e t hylben zene 1.1 0.2 0.4 0.2 0.4 0.1
1,2,3,4-Tetramethylbenzene 0.4 0.1 0.2 0.1 0.1 0.1
O t h e r C l o benzenes 1.5 1.6 1.9 1.9 1.9 1.8
Cll A l k y l b e n z e n e s 0.8 0.6 0.8 0.7 0.7 0.7
Naphthalenes 0.1 0.1 0.1 0.1 0.2 0.1
Unknowns 0.0 0.0 0.0 0.0 0.4 0.7
Total C,+ 70.0 80.8 77.7 86.4 80.0 83.0
Total aromatics 36.1 27.5 31.0 28.5 29.4 26.7
84 CHANG
1.0-
0.8 -
0.6
0.2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Fractional Bed Length
FIG. 33. Band aging of methanol-to-gasoline catalyst in the

first cycle of the long-term aging test [ 1421.
A characteristic of this type of operation is catalyst band-aging,

exemplified in Fig. 33, due to progressive coking. The end of
cycle occurs when the reaction front arrives at the reactor out-
let. The corresponding change in product distribution through
a cycle is plotted in Fig. 34 (fifth cycle).
The fixed-bed process was subsequently scaled up to a 4 bblld
pilot plant [ 1431. Data summarized in Tables 24 and 25 indicate
that the bench-scale results were duplicated on the larger scale.
The reactor temperature profiles in Fig. 35 indicate identical
aging patterns in the bench and pilot-plant tests.
B . Fluid-Bed Process
Development of the fluid-bed process followed a course similar

to the fixed-bed. Initial studies were carried out in the bench-
scale reactor illustrated in Fig. 36 [ 1441.
50 I I I 1
- 40 - _h
V v
lsoparaffins
r
- V Y
m
V
> \r
-
-
7
I
.I-
? 20-
s
0
o / 0 Olefins
10-
0 I a L
Time On Stream (Hours)
FLG. 34. Changes in hydrocarbon composition during the fifth

cycle of the long-term aging test [ 1421.
TABLE 24 [1431
Four Barrels-per -Day Fixed -Bed MTG Demonstration
Plant Operating Conditions
DME reactor inlet temperature, OC 3 15
MTG reactor inlet temperature, OC 360
MTG reactor exit temperature, OC 4 12
Gas recycle ratio, molar 9
Catalyst WHSV (pure methanol basis) 1.6
Reactor pressure, kPa 2170
Separator temperature, OC 52
86 CHANG
TABLE 25 [143]
MTG Scale-Up : Yield Comparison
(average of first cycle)
Product, w t % MeOH 4 b b l l d unit Bench unit
Hydrocarbons 43.75 43.73

Water 56.19 56.17

C o t CO2, H, 0.06 0.10
Methanol 0.00 0.00
Dimethylet her 0.00 0.00
100.00 100.00
Hydrocarbon product, wt8:
Methane + ethane 2.1 2.1
Ethylene 0.03 0.02
Propane 8.6 8.5
Propylene 0.2 0.2
n-B utane 4:2 4.1
i-Butane 8.4 8.5
But ene s 0.7 0.7
C 5+ Gasoline 75.8 75.9
- -
100.0 100.0
9 RVP gasoline
Yield, w t % of HC 80. 2 80.2
R + O 95 95
Process conditions scoped were: 316-482OC; 1.0-4.4 atm;

WHSV, 0.50-3.5 h - l . Typical data a t 371°C, 1 atm, 1 WHSV a r e
shown in Tables 26 and 27 and plotted in Figs. 37-40.
The changes in selectivity through a cycle a r e directionally
similar to the fixed-bed. In practice, however, a fraction of
HY DROC ARB 0 N S FROM METHANOL 87
Fractional Bed Length
FIG. 35. Fixed-bed MTG. Movement of temperature profiles

[ 1431.
the catalyst would be removed continuously, regenerated. and

returned so that the selectivity would b e neld at a constant
level.
In Figs. 41 and 4 2 the effects of temperature and pressure
on methanol in the effluent a r e plotted against SC-’, defined
by
where Sc = corrected space velocity, S = experimental space ve-

locity, r = dilution factor (moles reactant + moles diluent)l(moles
reactant), and q = relative catalyst activity expressed a s a func-
tion of initial activity q o . An exponential deactivation was as-
sumed.
Conversion increases with increasing temperature and llSc.
The pressure effect appeared not to b e first-order. This was
assumed to be due to adsorption inhibition effects, To account
for both inhibition and change in gas residence time due to
88 CHANG
FIG. 36. Schematic of fluidized dense bed pilot plant [ 1441.

All dimensions are in inches.
different pressures, the data were correlated with the group

defined by
(~+KAPA,)
-@= --
00 ( ~ + K A P A )PO Sc
TABLE 26 f1441
Typical Material Balances in Fluid-Bed Pilot Planta
Material balance no. 26-01 25-04 25-07 25-10

MeOH charged (Ib of MeOH/ 6 71 13 5 203
lb of cat.)
Product (wt%of charge) :

Methanol 0.32 0.84 1.13 1.82
Dimethyl ether 0.02 0.07 0.15 0.43
Hydrocarbons 42.40 42.20 42.80 43.20
Water 56.77 55.81 55.54 55.08
C O , CO,, H, 0.15
- -
0.28 0.21
- 0.23
-
99.7 99.2 99.8 100.8
Hydrocarbons (wt % of
hydrocarbons) :
Methane 0.93 0.90 0.81 0.77
Ethane 0.25 0.17 0.17 0.14
Ethylene 4.84 5.64 6.26 6.54
Propane 3.65 3.08 2.83 2.37
Propene 5.41 5.98 7.21 7.97
I sobut ane 15.71 13.14 12.30 10.78
n -Butane 1.48 1.25 1.47 0.92
But enes 4.64 5.22 5.64 6.24
c 5+ 63.11 64.62 63.30 64.27
where PA^ = sum of the partial pressures of methanol, DME, and

H,O at a set of arbitrary reference conditions (1 atm, 4OOOC) ;
po = gas density at reference conditions; while PA and p = ex-
perimental values.
Figure 43 shows a plot of effluent methanol concentration
vs
90 CHANG
TABLE 27 [144]
Composition of Hydrocarbon Products from
Fluid-Bed Pilot Planta
Material balance no, 26-01 25-04 25-07 25-10

MeOH charged (Ib of MeOH/ 6 71 13 5 203
lb of cat.)
Isopentane 12.36 11.45 11.06 10.07
n-Pentane 0.49 0.43 0.42 0.33
Pentenes 2.19 2.60 3.04 3.12
Cyclopentane 0.44 0.05 0.10 0.07
Methylcyclopent m e 0.71 0.67 0.63 0.58
n - Hexane 4.06 5.02 5. 14 5.42
Met hylpentanes 5.46 6.04 6.03 5.40
Dimethylbutanes 0.99 1.17 1.18 1.00
Hexenes 0.26 0.39 0.43 0.47
Cyclohexane 0.03 0.23 0.09 0.25
C ,-PON 3.31 3.87 4.00 4.29
Ce-PON 2.89 3.62 3.46 3.91
C,-PON 1.64 2.43 2.15 2.20
CIO-PON 0.11 0.22 0.43 0.27
Benzene 0.00 0.00 0.00 0.00
Toluene 1.38 2.12 2.38 4.03
E t hylben zene 0.16 0.18 0.17 0.17
X ylenes 6.05 6.30 5.80 5.50
Trimet hylbenzenes 8.39 8.30 7.30 7.20
Methylethylben zenes 1.34 1.44 1.30 1.31
Propylben zenes 0.02 0.03 0.02 0.03
1,2,4,5-Tetramethylbenzene 1.94 3.38 2.87 3.93
lI2,3,5-Tetramethylbenzene 1.52 1.80 1.74 1.02
1,2,3,4-Tetramethylbenzene 0.56 0.50 0.41 0.45
Other C ,,,-benzenes 1.36 1.23 0.97 1.07
C ,,-Benzenes 1.72 0.48 0.43 0.33
Naphthalenes 1.14 0.02 0.00 0.01
Other aromatics 2.58 0.67 1.71 1.86
'700°F. 0 psig, 1 WHSV; concentrations are w t % of total hydro-

carbons.
HYDROCARBONS
-
a
I
V 30t
AROMATICS -0.
AROMATICS
O D
n o
I DIMETHYLETHER 7
1 DIMETHYLETHER- 1
METHANOL
METHANOL PROCESSED METHANOL PROCESSED METHANOL PROCESSED

(Lb MeOH/Lb Cat.) (Lb MeOH/Lb Cat.) (Lb Me OH/Lb Cat.)
371OC 371 "C 399O c

I atm. 2.7 atm. 2.7 atm.
I WHSV 2 WHSV 2 WHSV
FIG. 37. Fluid bed reactor (total hydrocarbons and aromatics) [ 1441.
>
r
Y
0
c5+ x-
X
X
.. ..
I X
100 200
I
I00 i I00 2(
METHANOL PROCESSED
(LbMeOH/Lb Cot.)
METHANOL PROCESSED
(Lb MeOH/Lb Cat.)
- METHANOL PROCESSED
(Lb Me OH/Lb Cat. I
371OC 371OC 399OC

FLG. 38. Fluid bed reactor ((2,' and CSi hydrocarbons) 11441.
r I I $ 1 I
PROPANE x
P
w n - BUTANE
-6-r
METHANE +ETHANE
0
0 I00 2 20 I00 200
(Lb MeOH/Lb Cat. 1 (Lb McOH/Lb Cot.) (Lb MeOH/Lb Cot.)
371 OC 371 OC 399°C

FIG. 39. Fluid bed reactor (light paraffins) [ 1441.
W
W
Gj I0 -
w
>
E3t7
I - 2 20 I00 200
0 100 200 0 I00 200
(Lb McOH/Lb Cat.) (Lb MeOH/Lb Cat.) (Lb McOH/Lb Cat.)
371 OC 371 O C 399OC

FIG. 40. Fluid bed reactor (light olefins and durene) 11441.
1 ISC
FIG. 41. Effect of temperature on methanol conversion at 25

psig [ 1441.
25
20 -
-
8
-5- 15-
25 psig
0
E
5
s 10-
50 Psig\
I 0 psig
5-
K O
0 1 1 I I I l e
FIG. 42. Effect of pressure on methanol conversion at 700°F

[ 1441.
96 CHANG
FIG. 43. Plot of methanol concentration versus @/@,11441.
Observed
-
c
Rediaed From
Equilibrium
-8 1.25
5
y 1.00 -
0
5
-
0.75 -
.-E
n
0.50 -
0.25 -
I
0 1 2 3
Methanol (Wt %)
FIG. 44. Comparison of calculated equilibrium and experimen-

tal concentrations of methanol and dimethyl ether [ 1441.
-+ To Product Recovery
4
Disengager
0 -
16" x 3'
Flue gas
Regenerator
lo" x 2%'
Reactor
Air + N2
4" x 25'
-
Hopper
lo" x 2%'
Scale: 1/4" = 1'
Liquid Feed -
Vapor Feed -
FIG. 45. Unit 231 reactor-regenerator system [ 1461.
Figure 44 plots DME vs methanol at the reactor exit. Compari-

son with the theoretical curve confirms that the initial etherifica-
tion step is close to equilibrium.
The fluid-bed process was scaled up to a 4-bbl/d pilot plant
[ 1461. The reactor-regenerator system is shown in Fig. 45 and
a schematic flow diagram of the pilot plant is shown in Fig. 46.
Long-term demonstration runs of 34 and 75 d were conducted
in the 4-bblJd pilot plant. Table 28 gives the range of operating
conditions.
A typical reactor temperature profile is displayed in Fig. 47,
indicating the uniform temperature achieved along the consider-
able length of the reactor for this highly exothermic reaction.
98
t
313A33U SVZ) 1HE)Il
CHANG
TABLE 28
Operating Conditions for Fluid-Bed Pilot Plant 1 1461
-~
Temperature 388-427OC
Pressure 2.4-3.7 atm

WHSV, MeOH 0.5-1.3 h-'
Gas superficial velocity 0.7-1.8 f t l s
Feed composition 83- 100 wt%MeOH
In Table 29 [ 1631 representative product yields are shown and

compared with typical data obtained from the 4-bblld fixed-bed
pilot plant. It can be seen that the fluid-bed produces more light
olefins and less C5' gasoline than the fixed-bed. However, the
light olefins, when alkylated with isobutane and blended back,
give final gasoline yields higher than that of the fixed-bed.
C. Product Characteristics
Extensive product quality testing has been conducted on the

gasoline from the Mobil MTG process [ 1471. In general the syn-
thetic gasoline is identical in all major aspects to commercial high
octane gasolines. Table 30 presents a summary of the properties
of MTG gasoline.
It is seen that the octane and volatility characteristics are ex-
cellent. Unlike commercial gasolines, S and N contaminants are
totally absent. The one hydrocarbon in MTG gasoline not found
in significant amounts in conventional gasoline is durene. The
mechanism of durene formation has already been discussed. Be-
cause of its high melting point (79.3OC), the durene content of
the gasoline must be controlled at suitable levels, depending on
the ambient temperature, so a s not to crystallize out of solution.
Concentrations as high as 3% of durene were found to give no
adverse engine effects.
Fleet testing confirmed that the MTG gasoline was comparable
to commercial premium gasolines in road octane, cold-start, drive-
ability, and emissions.
100 CHANG
25
20
-
c
-
C
15
B
z
c
U
c'
.-
c
-B
10
t
K
0
500 700 900 1 loo
Temperature, OF
FIG. 47. Fluid-bed reactor temperature profile [ 1461 I
D . Olefin Production
A s indicated earlier, several techniques may be applied to maxi-

mize olefin selectivity in methanol conversion over ZSM-5 zeolites.
Reduced partial pressure [ 321 and catalyst modification with phos-
phorus compounds [ 331 are effective. A third technique utilizes
TABLE 29
Comparison of Fixed- and Fluid-Bed Yields [ 1631
~~
Yields from methanol in Yields from methanol in 4 bbl/d

fixed-bed reactor system fluid-bed pilot unit
Temperature, OC : Average bed tempera-

ture, OC 413
Inlet 360
Outlet
415 Pressure, kPa 275
Pressure, kPa 2170
Recycle ratio (mol) 9.0
2.0 Space velocity (WHSV) 1.0
Space velocity (WHSV)
Yields, wt % of methanol Yields, w t % of methanol

charged : charged :
Methanol + ether 0.0 Methanol + ether 0.2
Hydrocarbons 43.4 Hydrocarbons 43.5
Water 56.0 Water 56.0
c o , coz 0.4 C O , COZ 0.1
Coke, other -0.2 Coke, other 0.2
100.0 100.0
Hydrocarbon product, Hydrocarbon product,

wt%: wt %:
Light gas 1.4 Light gas 5.6
Propane 5.5 Propane 5.9
Propylene 0.2 Propylene 5.0
i-Butane 8.6 i-Butane 14.5
n -Butane 3.3 n-Butane 1.7
B u tene s 1.1 Butenes 7.3
C,+ Gasoline 79.9 C,+ Gasoline 60.0
- -
100.0 100.0
Gasoline (including alka- Gasoline (including alka-

late) (93 R + 0. RVP = late) (96 R + 0, RVP =
9 psi [62 kPa]) 85.0 9 psi [ 6 2 kPa]) 88.0
LP G 13.6 LPG 6.4
Fuel gas 1.4 Fuel gas

- 5.6
100.0 100.0
102 CHANG
TABLE 30 [ 1471
Properties of Gasoline f r o m Methanol (yield proportions of liquid
product from 4 bbl Id unit, alkylates and light hydrocarbons)
API gravity, (ASTM 0 2 8 7 ) 62.1

Reid vapor pressure, lb (ASTM D323) 11.1
Distillation, OF, % evap. (ASTM D86) :

lBP 86
5% 102
10% 116
20 % 136
30 % 159
40 % 186
50 % 2 16
60 % 240
70 % 263
80 % 297
90 % 33 6
95% 3 58
End point 422
% Residue 1.2
Research octane number (ASTM D2699) :
Unleaded 96.8
+3.17 g Pb/USG a s TEL 102.8
+3.17 g Pb/USG as TML 103.2
Motor octane number (ASTM D2700) :
Unleaded 87.4
+3.17 g Pg/USG a s TEL 95.1
+3.17 g Pb/USG a s TML 96.1
Distribution octane number (ASTM D2886) :
Unleaded 95.9
+3.17 g PblUSG as TEL 103.4
+3.17 g PblUSC as TML 104.9
Front end (IBP-212OF fraction) octane number:
Unleaded 94.7
+3.17 g Pb/USG as TML in distillation charge 96.7
Existent gum, m g / 1 0 0 mL (ASTM D381) 1.3
Potential gum, mg/100 mL (ASTM D873; 5 h) 2.5
(continued)
TABLE 30 (continued)
Hydrocarbon type, vol % (ASTM D1319):

Aromaticsa 27
Olefins 6
Saturates 67
Freeze point, OF (ASTM D2386) - 35
Sulfur, wt % (Mobil method) Nil

Nitrogen, w t % (Mobil method) Nil
a
Includes 3 wt% durene
the fact that olefins are intermediates in the MTG reaction path
[ 2 , 1481 . Conditions a r e chosen to operate at partial conversion.
Olefins are separated and unconverted feed recycled.
A fluid-bed study was carried out a t partial conversion of
methanol [ 1491. The methanol was diluted with various amounts
of water. Results are shown in Table 31.
Using methanol containing 17%water, 48%C 2 - C , olefins selec-
tivity was obtained at 52%conversion. Increasing water dilution
was found to improve ethene selectivity. This was attributed to
a kinetic effect, since the olefin distribution was not a s predicted
b y thermodynamics under reaction conditions.
VII. DUAL-FUNCTIONAL CATALYSIS
It is natural to inquire whether the two reactions of methanol

synthesis and subsequent hydrocarbon formation could b e done
in a single step over a dual-functional catalyst. This question
was investigated by Chang et al. [ 150, 1511. The overall ki-
netics may be represented by
The catalyst is composed of a metal oxide (methanol synthesis)

component and an acidic ( zeolite) component.
104 CHANG
TABLE 31 11491
Conversion of Methanol to Light Olefins
83% CH,OH/ 16% CH,OH/

Reactant composition 17% H 2 0 84% H 2 0
Conditions : 1 2 3
Temperature, P F 570 650 6 15
Pressure, psig 6 6 27
Methanol WHSV 0.4 0.4 0.2
Product yield, wt % of methanol :
Hydrocarbons 22.7 36.6 27.2
Methanol 15.9 8.9 26.8
Dimet hylet her 23.0 5.1 7.9
Water 38.3 49.2 38.0
Coke, C O , C 0 2 , other 0.1 0.2 0.1
- - -
100.0 100.0 100.0
Per pass conversion, % 52 84 62
Hydrocarbon product, wt 8:
Ethylene 21.3 18.8 27.6
P ropylene 17.2 11.7 17.5
Butenes 7.1 6.5 6.1
Pentenes 2.1 2.9 1.3
Paraffins (C,-C,) 18.3 20.8 24.7
Aroma tics 14.5 16.3 6.6
Nonaromatics 19.5 23.0 16.2
-
100.0 100.0 100.0
C2-C5 Olefins 47.7 39.9 52.5
The above kinetic scheme is a special case of a general class of

polystep reactions 1521 represented by
kl k2
A B -C; K = kl/kl' << 1
kl
The transformations A -+ B and B -+ C are effected by different

catalysts. A s indicated, the first reaction is subject to thermo-
dynamic limitation. However, the second reaction provides a
"drain-off" equilibrium displacement, and it can be shown [ 1521

that the attainable conversion of the coupled system can exceed
the product of the conversions of the constituent reactions. In
addition to activity considerations, certain diffusional criteria
must be met before the maximum polystep kinetics are realized.
Weisz t 1521 has given the following general criterion :
dN RT r2
@ =(=)(pi)(o)
<
where D = diffusivity of the intermediate (cm2/s)

r = mean diffusing distance, or particle radius (cm)
Pi = intermediate species partial pressure (atm)
(dN/dt) = reaction rate (mol sec-' cm-')
R = gas constant (cm' atm mole-' OK-')
T = reaction temperature (OK)
When Q is <1, diffusive transport will not significantly influence

the polystep reaction rate.
Chang et al. [ 1511 found that the polystep reaction of inter-
est contained an undesirable side-reaction : hydrogenation of the
intermediate olefins. In some cases, only light paraffins were
produced via this competitive pathway.
By varying the activity of each catalyst component, it was
determined that the unwanted side-reaction became significant
whenever the rates of methanol formation and aromatization were
not pmperly matched. It was found that the interception of the
olefins could be controlled simply by changing the activity of
the metal component. This is illustrated for the ZnCrZSM-5
catalyst in Fig. 48. At ZnICr = 4 , which is optimum for meth-
anol synthesis 11531, only light paraffins are observed. A s the
ZnJCr is lowered, the syngas conversion activity drops; how-
ever, aromatics formation becomes evident. Figure 49 shows a
strong correlation between the specific hydrogenation activity
(for pmpene) of the ZnCr component and the aromatization se-
lectivity of the composite.
The effect of increasing zeolite activity is shown in Fig. 50
which plots a conversion and selectivity vs the activity pa-
rameter a, which is a measure of the relative n-hexane crack-
ing acid activity [1541. .It can be seen that with increasing
a, the efficiency of the "drain-off" mechanism increases, a s re-
flected in the conversion behavior, and the selectivity changes
in a pattern similar to the methanol-to-hydrocarbon pathway in
Fig. 8.
106 CHANG
t itTYPI
I
I
CAL HIGH PRESSURE
METHANOL SYNTHESIS
CATALYST
ATOM PERCENT
FIG. 48. Syngas aromatization over ZnCrZSM-5. Effect of

metals composition, 800°F, 1200 psig. H,/CO = 1, 1800 h-’ GHSV
[ 1511.
An interesting result, as seen in Fig. 50, is the suppression

of methane with increasing a. Since methane is thermodynamically
stable, its suppression was believed to be achieved by intercep-
tion of a precursor, which was common also to methanol. The fol-
lqwing mechanism was proposed :
7c
6C
0
50
z
5
c-
-
40
i,
A
W
30
v)
0
-I
4
5
g 20
4
10
0
SPECIFIC PROPENE HYDROGENATION RATE, k, hr-'
FIG. 49. Syngas aromatization over ZnCrZSM-5. Correlation

of catalyst hydrogenation activity with aromatics selectivity [ 1511.
In support of this proposal, the result of Vlasenko et al. [ 1551

showing that methanation and methanol formation a r e parallel r e -
actions on methanol synthesis catalysts was cited. Furthermore,
108 CHANG
30
z
$20
I
10
60
50
40
8
5
>-
I- 5
230
l-
0
!
I
In
W
20 ?
z
c
3
10 I
0 3
1 10 IUU
a
FIG. 50. Syngas aromatization over ZnCrZSM-5. Influence of

zeolite cracking activity. Zn/Cr = 0.34. H,/CO = 1, 427OC, 83
.
atm, GHSV = 1740- 1780 h - l [ 1511
TABLE 32
Conversion of Syngas and Methanol over Various
Catalysts [ 1571 (350-380OC)
~~
Feed CO i2H2 CO + 2H, CH,OH

Catalyst ZnCra ZnCr + ZVK ZVK
Pressure, at m 20 20 1
Hydrocarbons, wt 8:
CH4 Traces 8.0 Traces
c2-c4 100 9.0 42.6
5.7 5.23
c6 14.9 3.1
c, 1.71 2.0
C8 2.1 1.0
B en zene 2.9 2.79
Toluene 11.7 14.9
Xylene 3.06 4.0
Et hylben zene 11.8 15.3
c 9+ 19.1 8.8
a
14.25% conversion (to CH,OH and mainly propane) at 35OOC.
Aharoni and Starer [ 1561 found evidence for a common precursor

for methane and methanol.
Similar results have been reported by Ione et al. [ 1571 in their
study of syngas conversion over dual-functional catalysts. A zinc-
chromium methanol synthesis component was combined with a high
silica zeolite ZVK, which is reported to be analogous to ZSM-5
11581. Syngas (H,/CO = 2) was reacted over this composite at
20 atm, 35O-38O0C, and 200-1000 h - I GHSV. Typical results are
shown in Table 32, which also contains data from methanol con-
version over zeolite ZVK for comparison. A close similarity is
evident.
The conversion of syngas to light hydrocarbons over a dual-
functional ZnO-CuO/H-erionite catalyst has been reported [ 1671 .*
At 322-458OC and 27-68 atm the product was entirely Cl-C4 hy-
drocarbons. This selectivity was attributed to the small pore
size of erionite. Paraffins were predominant, indicating inter-
ception and hydrogenation of olefinic intermediates. Addition of
water suppressed double bond hydrogenation and increased
ethene selectivity.
110 CHANG
V I I I . POSTSCRIPT
This survey has provided a documentary account of the history

and nature of a catalytic reaction which h a s become the object of
widespread attention. Aside from its obvious practical implica-
tions, the formation of hydrocarbons from methanol raises a host
of intriguing theoretical questions.
There is clearly a diversity of opinion concerning the mechan-

i s m of initial C-C bond formation, and a wide gap separating
theory from experiment. Nevertheless, there seems to b e a
majority consensus that the reactive C, intermediate is cationic
in character, at least during some stage of its existence.
Regardless of the nature of the electron-donor substrate, the
reaction "bottom line," a s it were, is the transfer and incorpora-
tion of a six-electron CH, unit, initially embedded in a stabilizing
cationic environment, into the substrate. It would seem there-
fore that the central question is really concerned with timing,
i . e . , at what point during its career does this C , intermediate
lose a proton (or H, in the case of CH,')? Before, during, o r
after C-C bond formation?
It is extremely tempting to draw analogies with gas-phase
ionic reactions, recognizing, however, that ion solvation could
(and probably will) drastically alter the picture in catalytic en-
vironments. Nevertheless, such exercises may be useful heuris-
tically. For example, reactions such a s
have been studied by ion cyclotron resonance spectroscopy [ 1591.

Such gas-phase reactions a r e extremely fast and apparently p m -
ceed without activation energy. I s it possible that, even under
the influence of high solvation energies, completely analogous r e -
actions with similar (net) energetics prevail in the zeolitic e n -
vironment? Could this explain, for example, the anomalous tem-
perature dependency observed by van den Berg [ 741 during the
early stages of dimethyl ether reaction?
Because of the high exothermicity and fast kinetics of the
methanol transformation, classical methods of mechanism elucida-
tion, e.g. , trapping of intermediates, tracer, and kinetic stud-
ies, have failed to provide unequivocal answers. Direct obser-
vation of the transition state seems beyond our capabilities at
present. Indeed, it is even conceivable that the C-C bond for-

mation may involve a continuous spectrum of transition states, a
situation that is not too appealing to mechanist.
The solution to the C-C bond problem remains an open chal-
lenge at present.
The literature on this question has grown apace. It seems
certain that significant progress will b e made even a s this s u r -
vey goes to press.
REFERENCES
S. L . Meisel, J . P . McCullough, C. H . Lechthaler, and

P. B . Weisz, Chemtech, 6, 86 (1976).
C . D. Chang and A . J . Silvestri, J. Catal.. 47, 249
(1977).
C . D. Chang, W . H . Lang, and A . J . Silvestri, U.S.
Patent 3,894,104 ( 1975).
D . R . Stull, E. F . Westrum, Jr., and G . C . Sinke, The
Chemical Thermodynamics of Organic Compounds, Wiley ,
New York, 1969.
J . A . LeBel and W . H . Greene, A m . Chem. J . , 2, 20
( 1880).
E. Sernagiotto, Gazz. Chim. Ital., 44(I), 587 (1914).
A . V. Grosse and J. C. Snyder, U s . Patent 2,492,984
(1950).
H . Adkins and P. D . Perkins, J . Phys. Chem., 32, 2 2 1
(1928).
K . V. Topchieva and A . V . Ballod, Dokl. Akad. Nauk
E R , 75, 247 (1950).
N . M . Cullinane, S. J. Chart, and R . Meatyard, J. SOC.
Chem. Ind., 67, 142 (1948).
E . Gorin and%. H . Gorin, U. S. Patent 2,456,584
(1948).
F. S. Fawcett and B. W . Howk, U.S. Patent 2,744,151
(1956).
Ya. T. Eidus, Akad. Nauk SSSR Izv. Otd. K h i m . Nauk,
1, 65 (1943).
E . J. Argauer and G . R . Landolt, U.S. Patent 3,702,886
(1972).
G . T . Kokotailo, S. L . Lawton, D. H . Olson, and W . M .
Meier, Nature, 272, 437 (1978).
P. Chu, U.S. Patent 3,709,979 (1973).
G . T. Kokotailo, P. Chu, S. L . Lawton, and W . M .
Meier, -Nature,
- 275, 119 (1978).
112 CHANG
P . B . Weisz and V . J . Frilette, J . Phys. Chem., 6- 4 , 382

(1960).
P. B . Weisz, V . J . Frilette, R . W . Maatman and E . B .
Mower, J . Catal., 1 , 307 ( 1 9 6 2 ) .
S. M . Csicsery, inACS Monograph 171 (J. A . Rabo,
ed. ) , American Chemical Society, Washington, D. C , .
1976, p. 680.
J . Shabtai, Chim. Ind. (Milan), 6 1 ( 1 0 ) , 734 ( 1 9 7 9 ) .

C. D. Chang, W . H . Lang, and K . Bell, Catalysis
of Organic Reactions ( W . R . Moser, e d . ) , Dekker,
New York, 1981, p. 73.
D. W . Breck, Zeolite Molecular Sieves, Wiley, New York,
1974.
J . A . Rabo ( e d . ) , ACS Monograph 171, American Chem-
.
ical Society, Washington, D .C , 1976.
W. M. Meier and D. H . Olson, Atlas of Zeolite Structure
Types, International Zeolite Association, Juris-Druck ,
Zurich, Switzerland, 1978.
E . G . Derouane. J. B . Nagy, P. Dejaifve, J . H . C . van
Hooff, B . P . Spekman. J . C . Vedrine, and C . Naccache,
J. Catal., 53, 40 (1978).
J. R . Anderson, K . E'oger, R . Mole, R . A . Rajadhyaksha,
and J . V . Sanders, Ibid., 58, 114 ( 1 9 7 9 ) .
0 . V . Bragin, B . K.Nefedov, T. V . Vasina, T. V .
Alekseeva, V. N . Lutovinova, and K h . M . Minachev,
Izv. Akad. Nauk SSSR, Ser. Khim., g,2639 ( 1 9 7 9 ) .
F . S - S . Hwu, PhD Dissertation, Rice University, 1981.
S. E . Voltz and J . J . Wise, Final Report, November
1976, Contract No. E(49-18)-1773, U.S. Energy Re-
search and Development Administration.
A . A . Balandin, G. V . Isagulyants, E . I . Popov, Yu. I .
Derbentsev, and S. L . Vinogradov, Izv. Akad. Nauk
SSSR, Ser. Khirn., p. 218 ( 1 9 5 8 ) .
C . D. Chang, W . H . Lang. and R . L . Smith, J. Catal.,
56, 169 ( 1979).
-
W . Kaeding and S. A . Butter, Ibid., 6 1 , 155 ( 1 9 8 0 ) .
Mob2 Oil Corporation, Netherlands Patent 7 , 0 1 4 , 8 0 7
(1971).
T . V. Lirnova and I . E . Neimark, Kolloidn. Zh., - 41(2),
354 ( 1 9 7 9 ) .
S . N . Khadziev, M. 1. Levinbuk, Yu. V . Shumovskii,
and K . V . Topchieva, Kinet. Katal., z o < 6 ) , 1596 ( 1 9 7 9 ) .
N . S. Khashgul'yova, S. N . Khadziev, and A. A .
Kubasov, Vestn. Mosk. Univ. Ser. 2 , Khim., g ( 2 ) .
156 ( 1 9 8 1 ) .
C . D . Chang, W . H . Lang, and A . J . Silvestri, U.S.

Patent 4,062,905 ( 1977).
J . T . Cobb, J r . , V . T. Coon, and P . Tipnis, F B
Report, September 1978, Contract No. EW-78-S-02-4691,
U.S. Department of Energy.
B . B . Singh and R . G. Anthony, Prepr. Can. Symp.
Catal., 6, 113 (1979).
E.S h g h , F. N . Lin, and R . G . Anthony, Chem.

Eng. Commun., 4, 749 (1980).
~
S. Ceckiewicz, React.
~- Kinet. Catal. Lett., 16(1), 11
(1981).
S. Ceckiewicz, J. Chem. Soc., Faraday Trans., 1, 77,
269 (1981).
F. A . Wunder and E . I. Leupold, Angew. Chem., Int.
Ed. Engl., 19(2), 126 (1980).
T. V . Whittam and M . S. Spencer, U . K . Patent Appl.
GB 2,061,999A.
T. Inui, T. Ishihara, and Y. Takegami, J . Chem. Soc.,
Chem. Commun., p. 936 (1981).
T. Inui, H. Matsuda, N . Morinaga, G . Takeuchi,
T . Ishihara, and Y. Takegami, Japan-Soviet Catal.
Seminar, 6th Osaka, p. 196 (1981).
A. B . Schwartz and J . Ciric, in Ref. 75.
K . V . Topchieva, A . A . Kubasov, and T . V . Dao,
Vestn. Mosk. Univ., Khim.,- - 27(6), 628 (1972).
N . Davidova, D. Shopov, N . Jaeger, and G . Schulz-
Ekloff, React. Kinet. Catal. Lett., 12(3), 229 (1979).
S. Hochevar, G . V . Echevskii, B . Szaj, and K . G .
Ione, Ibid., 13(4), 425 (1980).
F. X. Cormerais, Y. S. Chen, M . Kern, N . S . Gnep,
G . Perot, and M. Guisnet, J . Chem. Res., ( 5 ) , 290
(1981).
F. A . Swabb and B . C . Gates, Ind. Eng. Chem., Fun-
dam., 11, 540 (1972). .
K Z a t G s k i and S. Krzyzanowski, Acta Phys. Chem.,
29, 347 (1978).
S. Spencer and T . V . Whittam, Ibid., 29, 307 (1978).
B . N . Dolgov, Die Katalyse in der Oyanischen Chemie,
DVW, Berlin, 1963, p. 439.
B . V . Matyushenskii and G . N . Freidlin, Zh. Vses. Khim.
Obshchestv., 12(6), 710 (1967).
D. E. PearsonrJ. Chem. Soc., Chem. Commun., p. 397
(1974).
W. K . Bell and C. D. Chang, U.S. Patent 3,969,427
( 1976) .
114 CHANG
L . Kim, M. M. Wald, and S. G . Brandenberger, J . Org.

C a . , %(17), 3433 (1978).
D. C . Hargis and L. J . Kehoe, U.S. Patent 4,072,733
(1978).
S. Kikkawa, Y . Shimizu, and S. Higuchi, Chem. Lett.,
p. 848 (1979).
Y . Ono, T . Mori, and T . Keii, Prepr. 7th Int. Congr.
Catal., Tokyo, 1980, Paper C12.

Y. Ono, T . Baba, J . Sakai, and T . Keii, J . Chem. SOC.,
Chem. Commun., p . 400 (1981).
F. X. Cormerais, G . Perot, F . Chevalier, and M. Guisnet,
J . Chem. Res. (S), p . 362 (1980).
C . D. Chang, J . C . W . Kuo, W . H . Lang, S. M . Jacob,
J . J. Wise, and A . J . Silvestri, Ind. Eng. Chem.,
Process Des. Dev., lJ, 255 (1978).
J . R . Anderson, T . Mole, and V. Christov, J. Catal.,
61, 477 (1980).
-
D. Chang, Chem. Eng. Sci., 5,619 (1980).
N . Y . Chen and W . J . Reagan, J. Catal., 59, 123 (1979).
Y . Ono, E . Imai, and T . Mori, 2. Phys. Chem., N . F . ,
115, 99 (1979).
-
Y.-Ono and T. Mori, J. Chem. Soc., Faraday Trans. 1,
77, 2209 (1981).
H . - J . Doelle, J. Heering, and L . Riekert, J . Catal., 71,
-
27 (1981).
R . G . Anthony and B . B . Singh, Chem. Eng. Commun.,
6, 215 (1980).
-
F. P. van den Berg, Doctoral Dissertation, Eindhoven,

1981.
P. B . Venuto and P. S . L . Landis, Adv. Catal., Is, 259
( 1968).
M . L. Poutsma. in Ref. 24.
H . Pines, The Chemistry of Catalytic Hydrocarbon Con-
versions, Academic, New York, 1981.
~
A . K . Galwey, J. Catal., 19, 330 (1970).

P. Dejaifve, J. C . Vedrine, V. Bolis, and E . G .
Derouane. m d . , 63, 331 (1980).
M. E. Winfield, in Catalysis, Vol. 7 (P. H . Emmett, e d . ) ,
Reinhold, New York, 1960.
H . Pines and J . Manassen. Adv. Catal... - 16.. 49 (1966).
J . M. Parera and N . S. Figoli, J . Catal., 14, 303 (1969).
F. Figueras, A . Nohl, L . de Mourgues, a n d Y . Trambouze,
Trans-. Faraday Soc., 67, 1155 (1971).
H . Knozinger, Angew. Chem., Int. E d . , - 7, 791 (1968).
1851 F. Figueras, P. Renard, and L. de Mourgues, J . Chim.

- -65, 1393 (1968).
Phys.,
F. Figueras, L . de Mourgues, and Y . Trambouze,
1861
J . Catal., 14, 107 (1969).
1871 C . Naccache. Y . Kodratoff, R . C . Pink, and B . Imelik,
J . Chim. Phys., 63,542 (1966).
t 881 B . D. Flokhart, I T R . Leith, and R . C . Pink, Trans.
Faraday SOC. , 5, 542 (1969).
891 B . Notari, Chim. Ind. (Milan), 51, 1200 (1969).

901 T. Matsushima and J . M . White,?. Catal., 44, 183
(1976).
r911 G . Schmitz, J. Chim. Phys., 75(6), 650 (1978).
_I_
[921 E . Detrekoy and D . Kallo, Acta Chim. Acad. Sci. (Hungary),

95, 201 (1977).
931 G . Derouane, P . Dejaifve, and J. B . Nagy, J. Mol.
Catal., 3, 453 (1977/78).
t 941 m e r a n e k and M . Kraus, in Comprehensive Chemical
Kinetics, Vol. 29 (C. H . Banford and C . F. H . Tipper,
e d s . ) , Elsevier, New York, 1978, Chap. 3.
[ 951 E . I . Heiba and P. S. Landis, J. Catal., 3, 4 7 1 (1964).
[ 961 J . R. Jain and C . N . Pillai. I&. , 9, 32271964).
[ 971 J . E. Dambrowski, J. B. Butt, and H. Bliss, g.. Is,
297 (1970).
[ 981 G. Pfeifer and T. Flora, Magy. Kern. Foly., c ( 8 ) , 343
( 1965).
[ 991 W. J . Mattox, U.S. Patent 3,036,134 (1962).
1001 P. Salvador and W. Kladnig, J. Chem. SOC., Faraday
Trans. 1, p. 1153 (1977).
r 1011 G . B . Kistiakowski and K . Saner, J . Am. Chem. Soc., 2,
5699 (19560; so, 1066 (1958).
[ 1021 -
R . D. Smith and J . H . Futrell, Chem. Phys. Lett., 41(1),
64 (1976).
[ 1031 J . V. Smith, in Ref. 24.
1041 . .
J A . Rabo and P H . Kasai, in Progress in Solid State
Chemistry, Vol. 9 (J. 0. McCaldin and G . Somorjai, e d s . ) ,
Pergamon, Oxford, 1975.
[ 1051 H. Noller and W. Kladnig, Catal. Rev., 13,150 (1976).
[ 1061 S. Beran and P . Jiru, React. Kinet. Catal. Lett., 2, 4 0 1
(1978).
[ 1071 P. E . Pickert, J . A . Rabo, E . Dempsey, and V . Shomaler,
Proc. Third Int. Congr. Catal., Amsterdam, 1964.
1081 C . D . Chang and C. T-W. Chu, J. Catal., 7 4 , 203 (1982).
[ 1091 W. Kirmse, Carbene Chemistry, 2nd ed. , Academic, N e w
York, 1971.
116 CHANG
[ 1101 H . M. Frey and G . B . Kistiakowski, J. Am. Chem. Soc.,

79, 6373 (1957).
6-,A . Olah, J . R . DeMember, and J. Shen, -Ibid.,-95,
4952 (1973).
11121 J. P. van den Berg, J . P . Wolthuizen. and J. H. C .
van Hooff, Proc. Vth Conf. Zeolites, Naples, Italy,
1980, p. 649.
11131 R . H . Nobes, W. R . Rodwell, W . J. Bouma, and

L . Radom, J . Am. Chem. Soc., 103, 1913 (1981).
t 1141 M. J. S. Dewar and C . A . Ramsden, J. Chem. Soc.,
Perkin Trans. 1, p. 1839 (1974).
[ 1151 C . D. Chang and W. H . Lang, Unpublished Data.
[ 1161 P. B . Venuto and P. S. Landis, J. Catal., 2,330 (1971).
1171 E. L. Wu, G. H . Kuhl, T. E. Whyte, J r . , and P. B .
Venuto, Adv. Chem. Ser., 101, 490 (1971).
1181 .
A. R Lepley and A . G. Giumanini, in Mechanisms of
.
Molecular Migrations, Vol. 3 (B J . Thyagarajan, ed. ) ,
Wiley-Interscience, New York, 1971, p. 297.
t 1191 G. A. Olah and R . H . Schlosberg, J . Am. Chem. Soc.,
-
90, 2726 (1968).
G. A . Olah, G . Klopman, and R . H . Schlosberg, - Ibid.,
91, 3261 (1969).
r. A . Olah, Y. K . Mo, and J . A . Olah, W d . , 95, 4939
( 1973).
D. Kagi, J . Catal., 2, 242 (1981).
C. D . Chang, K d . , s, 244 (1981).
G. A . Olah, J. Sommer, and E. Namanworth, J. Am.
Chem. Soc., g , 3576 (1967).
1251 G. A . Olah, in Chemical Reactivity and Reaction Paths
( G . Klopman, ed.), Wiley, New York, 1974, Chap. 6.
M. Herlem, Pure Appl. Chem., 49, 107 (1977).
G. Perot, F. X. Cormerais, and M. Guisnet, J. Chem.
R e s . (S), p . 58 (1982).
1281 B. J . Ahn. J. Armando. G. Perot, and M. Guisnet,
C. R. Acad. Sci., Ser. C , 288, 245 (1979).
[ 1291 F. N. Lin, J. C . Chao, and R . G. Anthony, & C
Process Technol. , 4, 72 (1978).
1301 P. Jiru, L. Kubelkova, J. Novakova, Z . Tvaruzkova,
and B . Wichterlova, All-Union Conference on Mechanism
of Catalytic Reactions, Nauka, Moscow, 1978, p. 25.
J. Novakova, L. Kobelkova, Z . Dolejsek, and P. J k u ,
Collect. Czech. Chem. Commun., 44, 3341 (1979).
V . B o b , J. C . Vedrine, J. P . van den Berg, J. P.
Wolthuizen, and E. G. Derouane, J . Chem. Soc.,
Faraday Trans. - 1, 76, 1606 (1980).
[ 1331 S. L. Meisel, Philos. Trans. R . SOC. London, A, 157

(1981).
1341 C . D . Chang and S. S . Grover, U . S . Patent 4 , 0 5 8 , 5 7 6
(1977).
[ 1351 C. D. Chang, A . J . Silvestri, and R . L. Smith, U . S .
Patent 3 , 9 2 8 , 4 8 3 ( 1 9 7 5 ) .
1361 J. C . W. Kuo, U . S . Patent 3 , 9 3 1 , 3 4 9 ( 1 9 7 6 ) .
11371 C . D . Chang, S . M. Jacob, A . J. Silvestri, and J . C .

.
Zahner, U. S Patent 4 , 1 3 8 , 4 4 0 ( 1979).
[ 1381 W . Lee and S. Yurchak. U.S. Patent 4 , 1 9 7 , 4 1 8 ( 1 9 8 0 ) .
r 1391 H. Owen and P. B . Venuto, U . S . Patent 4 , 0 4 6 , 8 2 5 ( 1 9 7 7 ) .
[ 1401 N . Daviduk and J . H . Haddad, U . S . Patent 4 , 2 5 1 , 4 8 4
(1981).
r 1411 J . J . Wise and A . J . Silvestri, Oil Gas J . (November 2 2 ,
1976).
[ 1421 S. Yurchak, S . E . Voltz, and J . P . Warner, Ind. Eng.
Chem., Process Des. Dev., lo, 527 ( 1 9 7 9 ) .
[ 1431 D. Liederman, S . Yurchak, J . C . W . Kuo. and W. Lee,
Papers, 15th Intersociety Energy Conversion Eng. Conf. ,
-1980, p . 1573.
[ 1441 D. Liederman, S . M . Jacob, S. E . Voltz, and J. J . Wise,
Ind. Eng. Chem., Process Des. Dev., g ( 3 ) , 340 ( 1 9 7 8 ) .
[ 1451 D . F . Palazm and W . Y . Lee, Prepr., Coal Technol.
-
Conf., Vol. 5, Houston, Texas, November 1980, p . 319.
t 1461 A . Y . Kam and W. Lee, Final Report, Contract No.
EX-76-C-01-2490, U.S. Department of Energy, April
1978.
1471 C . R . Morgan, J. P. Warner, and S. Yurchak. Ind. Eng.
Chem., Prod. Res. Dev., g ( 1 ) . 185 ( 1 9 8 1 ) .
I1481 C. D . Chang, W. H . Lang, and A . J. Silvestri, U . S .
Patent 4 , 0 5 2 , 4 7 9 ( 1977) .
11491 J . A . Brennan, W . E . Garwood, S. Yurchak, and W . Lee,
Proc. Int. Seminar on Alternate Fuels ( A . Germain, e d . ) ,
Liege, Belgium, 1981, p . 19.
I1501 C . D. Chang, W . H . Lang, and A . J . Silvestri,
J . Catal., 56, 169 ( 1 9 7 9 ) .
[ 1511 C. D. Chang, W . H . Lang, and W . K . Bell, Chemistry
in Energy Production (R. G . Weiner, e d . ) , In Press.
[ 1521 P . B . Weisz, Adv. Catal., 9, 137 ( 1 9 6 2 ) .
1531 G. Natta. in Catalysis, Vol. 3 (P. H . Emmett, e d . ) , 1955.
1541 P. B . Weisz and J. N . Miale, J. Catal., 4. 527 ( 1 9 6 5 ) .
1551 V . L. Vlasenko, V. L . Chernobrivets. V. L. Lunev, and
N . K . Melchevskii, React. Kinet. Lett., 6, 195 ( 1 9 7 7 ) .
1561 C. Aharoni and H. Starer, Can. J. Chem., 52, 4044
(1974).
118 CHANG
[ 1571 K . G . Ione, L . S . Yegorova, and L. B . Orlova, P e .

6th Japan-Soviet Catal. Seminar, Osaka, Japan, 1981,
p. 180.
[ 1581 S. S. Shepelev and K . G . Ione, React. Kinet. Catal.
-
Lett., g ( 2 - 3 ) , 233 (1981).
J . L . Beauchamp and R . C . Dunbar, J . A m . Chem.
SOC., 92, 1477 (1970).
-
H . Knozinger and H . Stolz, Ber. Bunsenges. Phys.

Chem., 75, 1056 (1970).
G.Chang and W . H . Lang, U.S. Patent 4,025,575
( 1977).
D. H . Olson, G . T. Kokotailo, S. L . Lawton, and W . M .
Meier, J. Phys. Chem., 85, 2238 (1981).
C. H . Lechthaler, J . J. %e, P. B . Weisz, and A. J .
Silvestri, Proc. 13th Intersociety Energy Conversion
Eng. Conf., San Diego, California, 1978, p . 1.
11641 L . D. Rollmann and D . E . Walsh, J. Catal., 56, 139
( 1979).
1651 F . X . Cormerais, G . Perot, and M. Guisnet, Zeolites,
1(1), 1 4 1 (1981).
[ 1661 A.R . Lepley, in Chemically Induced Magnetic Polariza-
tion ( A . R . Lepley and G . L . Closs, e d s . ) , Wiley,
N e w York, 1973, p . 323.
[ 1671 T. Y . Chan, Report 1981, DOE/PC/04717-6, NTIS.
1681 H . Meerwein, H . Disselnkotter, R . Rappen, H . V.
Rintelen, and H . van de Vloed, Justus Liebigs. Ann.
Chem.-
- , 604, 1 5 1 (1956).
11691 .
G F. Salem, Doctoral Dissertation, University of
Southern California, 1980.
[1701 B . E. Langner, Appl. Catal., 2, 289 (1982).
1711 R . F . Sullivan, C . J . Egan, G . E . Langlois, and R . P.
Sieg, J. Am. Chem. SOC., 83, 1156 (1961).
.
R . F Socha, Personal Communication.
L . Miller and A . K . Hoffmann, J . Am. Chem. SOC., 7_1,
4128 (1949).
1741 G . A . Olah, B . G . B . Gupta, and S. C . Narang, Syn-
thesis, p . 274 (1979).
I1751 F. Cacace, G . Ciranni, and P. Giacomello, d . Am. Chem.
-
SOC., 104, 2258 (1982).
[ 1761 - 213
L. I. Krirnen and D. J . Cota, Org. React., 17,
(1969).
F. W . Fowler, Adv. Heterocycl. Chem., 13, 45 (1971).
P. G. Gassman, Acc. Chem. Res., 3, 26 (1970).

Hydrocarbon From Methanol

Enviado por

Dados do documento

Descrição original:

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Hydrocarbon From Methanol

Enviado por

Direitos autorais:

Formatos disponíveis

CATAL . REV.-SCI. ENG., 25(1).

Hydrocarbons from Methanol

Copyright 0 1983 by Marcel Dekker. Inc . 0022-2348/83/250 1-0001%3.50/0

The conversion of methanol to hydrocarbons is a remarkable

considerable diversity of current opinion will become apparent

11. GENERAL CONSIDERATIONS

Hydrocarbon formation from methanol requires the elimination

CHSOH * [CHZI + H2O ( 1)

(2n + 1 ) C H 3 0 H + 2CnHzn+z iCO + H20 (3)

Equation ( 4 ) is derivable from ( 3 ) by allowing the water-gas

nCHJOH + C n H z n - 6 + 3H, + nH,O, n = 6, 7 , . .. ( 5)

Although pure examples of (5) have not been observed to date,

Hydrocarbon formation from methanol is an exothermic reaction.

FLG. 1. Hydrocarbon formation from methanol. Heats of re-

A. Early Observations (pre- 1960)

The earliest report of hydrocarbon formation from methanol

1 AT 100% CONVERSION: OHR -400 cal I g ( - I674 k J / kg 1

lo-' 10' 3 10-2 10- I 1

FIG. 2. Methanol conversion to hydrocarbons. Heat of reac-

decomposition of methanol in molten ZnC1,. Hexamethylbenzene

15 CH30H FUSED ZnC& '@

AG~"" = -265.2 kcal/rnol

A s written, the reaction has AG = -261 kcalfmol of HMB at 283OC.

This large driving force for HMB formation has mechanistic a s

by P,O, in a paper titled "On the Chemistry of Nascent Methylene."

Mole percent MeOH

A patent issued to Fawcett and Howk [12] claims the direct

where n is the charge on the cation and z is the water of hydra-

Erionite 8 Ring ZSM-5 10 Ring Y Zeolite 12 Ring

FIG. 3. Typical zeolite pore geometries.

When M is a proton, the zeolite becomes a strong Bronsted

2. ZSM-5 and Related Zeolites

a. Structure of ZSM-5 and ZSM-11

FIG. 4. ZSM-Kchannel system.

channel system, represented in F i g . 4, consists of straight chan-

b.General Reaction Path

This was established by monitoring changes in product distribu-

ZSM-5 Pnma 8 T,-,*P 1

viewed along (010J

Secondary buildlng units: complex 5-1

{ [OIO] 5. 4 x 5.6 - [loo] -

Fault planes: (100)

orthorhombic, Pnma, a=20.1 b=19.9 c = 1 3 . 4 A

viewed along [loo]

Secondary building units: complex 5-1

Framework density: 17.7 T/1000 A3

Channels: < 100, 5.1 x 5 . 5 * * *

Fault planes: (100)

This reaction path has been confirmed in general by many work-

FLG. 9. Reaction path for dimethyl ether conversion to hydro-

Anderson et al. [27] and Ceckiewicz I431 also recognized this

c. Product and Distribution

Hydrocarbon distribution (wt%) in

c, 5.5 1.0 0.1 8.5 4.5

the higher aromatics, 1,3,5-trimethylbenzene, 1,2,3,4-, and

Normalized Normalized Equilibrium

Ben zene 4.1

Isopropylben zene 0.2

Other Al0a 2.7

FIG. 10. Effect of temperature on methanol conversion over

FIG. 11. Methanol reaction path at low pressure (HZSM-5 cat-

Thus at subatmospheric partial pressures, high selectivity to

f . Modifications and Isotypes

Channels: < l o o > -8 3.9***1 (100, S 3.9+