Documente Academic
Documente Profesional
Documente Cultură
Conversions of
Synthesis Gas and
Alcohols to Chemicals
Catalytic
Conversions of
Synthesis Gas and
Alcohols to Chemicals
Edited by
Richard G. Herman
Lehigh University
Bethlehem, Pennsylvania
vii
viii PREFACE
extended to all of the authors for their pleasant and timely cooper-
ation and for their efforts in producing quality papers. I espe-
cially appreciate and thank my wife, Helen Lynn, for.her assistance
in the formatting of this volume and for typing the final copy.
Richard G. Herman
September 1, 1983
CONTENTS
INTRODUCTORY ORIENTATION
ix
x CONTENTS
APPENDICES
Richard G. Herman
INTRODUCTION
SOURCES OF ENERGY
To carry out its daily functions, the human body consumes about
3000 kcal, or approximately 12,300 Btu, of energy. Of course, we
enjoy eating and drinking to provide this energy, wherein our bodies
transform the food into energy. This quantity of energy is equiva-
lent to the energy contained in one lb of bituminous coal. However,
to maintain our standard of living in the United States today, the
per capita energy consumption is equivalent to about 12 short tons
of coal annually [1]. This 300 million or so Btu quantity of energy
is used for transportation, to run our households, to operate our
business establishments, and by our industrial sector. It has been
noted that this per capita energy usage is equivalent to each of us
having 1200 personal slaves [2].
3
4 R.G.HERMAN
From where is this energy derived? It is obtained mainly from
the storehouse of fossil fuels - coal, crude oil, and natural gas.
Figure 1 shows the apportionment of natural resources that are used
to generate energy in the United States. A similar figure could be
constructed to represent the world energy patterns [3], although the
early usage of coal would be at a larger percent of the total, while
the contribution of petroleum would occur somewhat later and would
be of a slightly smaller magnitude. It is clear from Figure 1 that
the source of U.S. energy shifted from wood to coal in the late
1800's, most noticeably in the 1880-1895 period, and shifted again
after World War II from coal to petroleum and natural gas. The con-
tribution of nuclear power has been steadily increasing, while hydro-
derived power has maintained a constant portion of the energy supply.
~
C!l
a:
w
z 75
w
...J
<
~
0 50
~
LL
0
WOOD
~
z 25
w
()
a:
w
D.. 0
1850 1875 1900 1925 1950 1975 2000
YEAR
Figure 1. The historical pattern of energy utilization in the
United States.
During the last decade, the price of coal has been increasing
in conjunction with the increased cost of crude oil, although not as
rapidly. This has led to even more favorable comparative economics
for utilizing nuclear power to generate electricity in Western Europe
and Japan. This point was initially emphasized by the very large
price increase for crude oil that was enforced by the OPECa nations
in 1973. During the following year, there were 62 nuclear power re-
actors ordered or letters of intent placed by the OECDb countries.
Of course, with a six year lead time, these plants would not begin
to come on stream until 1980. Between 1974 and 1980, however, new
governmental regulations, increased lead time and costs, and, due in
part to conservation, the average annual increase in electricity
demand by OECD nations dropped from the high demand rate of about 7%
annually in the 1960-1973 period to 3.1% [6]. This led to the defer-
ment or cancellation of 71 nuclear power projects in the U.S. during
1974-1981 [6]. This compares with the 72 operating nuclear power
reactors at the end of 1978 [7]. The trend in the U.S. for new elec-
OH a Coal a Nuclearb
trical capacity is now away from nuclear power and toward coal-fired
generation plants. For example, from mid 1983 to 1988, 50 nuclear
plants are scheduled to start up, as well as 52 new coal-fired in-
stallations [8]. Past that date, 11 nuclear power plants are still
on the drawing boards, while 72 coal-based plants are in the plan-
ning stage [8]. .
At the same time that U.S. crude oil imports were increasing,
the price was also increasing. This is shown in Figure 3, where a
barrel of petroleum contains 42 U.S. gallons. The most immediate and
personal impact of the price increase of crude oil was the accompa-
nying increase in the price of gasoline. It can be noted from Figure
3 that the % increase in the price of a gallon of gasoline was not
even near to the % increase in the cost of a barrel of gasoline.
Part of this is due to the taxation contained in the price of gaso-
line, e.g. in 1973, 38% of the real price of a gallon was added as
tax [llJ. Thus, the "real" cost of gasoline in 1973 was approximate-
ly $.27/gallon. In 1980, the % tax in gasoline prices was only 13%
in the U.S., that is a real price of $1.06 plus $.14 tax for a sell-
ing price of $1.20/gallon. In 1973, gasoline prices in Europe, and
most other countries of the world, were already over $l/gallon be-
cause of taxation. For example, in Germany 239% of the "real" cost
of gasoline was added to the price as tax. Even so, the price of
gasoline only increased about 3.5-fold in the 1973-1980 period, while
the cost of imported crude oil increased over 10-fold.
80
U. S. Consumption I' /-\_,~
70 ~/
I
'/
I
2 I
I
co 60 1
1
I
.;: 50 I
I ·
-
-:;: I
::l
0' 40 ./....,'\)
-/
./ _'-''-.1
~ 30
....
z 20
Import~"""'"
Petroleum ~
YEAR
35
~ 30
(;;
.0
Qj
Co 25
~
...i
6 20
UJ
0
:::l 2 =g
a: a:
:l._
() 15 a:
Q.
f0- 1.40 ~ ~
CI)
1.20 ~ tit
<
UJ
10 1.00
~ ~
O:J
UJ w 0
...J
0 .80 ~ ~
0 a: "
~ 5
.60
.40
.. ....
~ ~
...I ~
Figure 3. Plots showing the cost of Middle East crude oil and of
gasoline in the United States during the last 33 years.
Petroleum
drilling activity and new reserves are being found. However, much
of the world has been extensively explored, and the probability of
finding additional super-giant oil fields (initial reserves greater
than 5 billion barrels) is dwindling. Indeed, during the 1971-1979
period, only two super-giant fields have been discovered, and both
of them are assigned to Mexico (with combined recoverable reserves
of 15 billion barrels) [13]. Of the 35 known super-giant fields,
25 are in the Middle East and 10 of those are in Saudi Arabia. One
of the remaining super-giant fields is the Alaskan Prudhoe Bay field.
It can be noted that during the 1971-1979 period, 40 giant fields
(initial reserves between 0.5 and 5 billion barrels) were discovered,
but this is only 58% of the number discovered during the previous
decade [13]. It is of interest that the 35 super-giant fields and
the 245 giant fields together account for 75% of the current produc-
tion and 73% of the proved reserves. The 29,700 smaller oil fields
provide the remainder [13].
20 Proved Reserves
40
(f) (f)
...J ...J
w w
a: 16 a:
a: 30 a:
« «
m 12 m
LL LL
0 20 0
(f) 8 (f)
z z
Q 10 Q
...J 4 ...J
...J ...J
m m
0
1950 1955 1960 1965 1970 1975 1980 1985
Natural Gas
higher than coal) [16], the price increases have already been felt.
For example, the % gas cost outlay made by gas utilities to natural
gas producers more than doubled during the 1972-1982 period [17], and
this has been reflected in the frequent requests of gas rate increas-
es made by gas utility companies to state public utility commissions.
The increased price of natural gas is also being felt in the chemi-
cal industry. In fact, it has been estimated that by 1985, feedstock
natural gas will represent more than 85% of the cash cost of produc-
ing methanol [15].
With this in mind, what are the prospects for the future produc-
tion of natural gas in the U.S.? The U.S. is the largest gas produc-
ing nation in the world, but Figure 5 shows that the production
has been gradually declining since the 1970-1973 period of high pro-
duction. In light of the crude oil situation during the last decade,
this would hardly have been predicted. However, the low volume of
additions during the 1968-1978 period could perhaps have been pre-
dicted because of the lack of incentives for drilling and finding new
sources of natural gas. The NGPA of 1978 provided the basis for eco-
nomic incentive, and the finding of new gas has increased.
I-
Proved Reserves I-
UJ W
UJ W
LL
LL 30 240
~ ~
III III
~ ..-Production ~
U 20 160 U
LL
LL
a a
en
~ 10 80 z
a 2...J
...J
...J :::
a: a:
I- 1950 1955 1960 1965 1970 1975 1980 1985 I-
Coal
In 1977, the U.S. production of coal was 689 million tons, and
President Carter released an energy plan that would increase the
coal production to 1,265 million tons in 1985. Of this production,
PRODUCTION OF ENERGY AND CHEMICALS 15
800
en
c
.8 700
-
.<::
~
0
en
c 600
0
.-
E
;i 500
0
f=
()
:::>
Cl
0 400
0::
0..
-l
<t 300
0
()
-l
<t
:::> 200
z
Z
<t
100
In the late 1960s, the U.S. economy was strong, inflation and
unemployment were at low rates, and there was a pronounced movement
away from coal and toward petroleum. Industrial capacity was being
utilized at record levels. However, in the early 1970s the economy
was slowing down and commodity prices were increasing. The increase
in OPEC petroleum prices and crude oil disruption of 1973 deepened
the economic slowdown and extended it into 1974 and 1975. Histori-
cally, there is a close direct relationship between economic activity
and energy consumption. Table 5 shows that this is true during the
1976-1979 recovery period. During this four-year period, the GNP
increased at an annual rate of 4.7% and energy consumption increased
at an annual rate of 2.8%. Adjustments were gradually being made to
alleviate the heavy U.S. dependence on imported petroleum and the
Natural Gas Policy Act was finally passed in late 1978, which encour-
age the drilling for and production of natural gas. Direct and in-
direct coal liquefaction research and development programs were
strongly supported by DOE and private industry, although the growth
of this support had taken a period of time to develop.
aEstimated quantities
"
18 R. G. HERMAN
Table 7. U.S. Energy Usage and the Sources of the Energy (quadrillion Btu)
:0
G)
::z::
m
:0
s::
»
z
PRODUCTION OF ENERGY AND CHEMICALS 21
SUPPL Y r---~
DEMAND
and natural gas by the end of the century if new reserves are not
continuously discovered (Figures 4 and 5) and if better and more ef-
ficient means of extraction, e.g. surfactant and C02 injection, are
not further developed. The presently recoverable reserves of coal
in the U.S. are at least 177 billion tons (Table 4), and at the cur-
rent rate of mining, these reserves would last for about 220 years.
Regardless of the number of years the reserves of fossil fuels might
last, the fact of finite quantity must be emphasized.
At some time in the future, the supply of fossil fuels will
reach a limiting point, at which the physical supply of these fuels
will decrease due to irreversible depletion of reserves. This is
shown schematically in Figure 8 on a prolonged time scale. The
demand for energy derived from fossil fuels has increased with pop-
ulation growth and with technological and social advances. The
energy demand curve in Figure 8 could also be labelled as standard of
living. As the fossil fuels become depleted, they will initially be
both physically and economically limited for energy production. Con-
servation and enhanced usage efficiencies, as well as greater depend-
ence on current alternatives, will delay the effects of the depletion
of fossil fuels. However, in the long run, the demand for energy
will fall because there will be no energy available if new energy
forms have not been developed. Hydroelectric and current-generation
nuclear power will provide little per capita energy. Ultimately,
human society will return to basically an agrarian labor-intensive
society (curve a in Figure 8).
FOSSil
FUELS r---------__ ",-
- - - -c
/
/ .......................... ·······b
-""'========a
ENERGY
DEMAND~---------
TIME_
However, there are many factors that will inhibit the growth of
PRODUCTION OF ENERGY AND CHEMICALS 23
'"oc:
500 r---__
'0
'"oc:
400
.0
(/)
UJ
>
a: 300
UJ
(/)
UJ
a:
UJ 200
--'
In
«
a:
UJ
> 100
o
Ll
UJ
a:
o 10 20 30 40 50 60 70 80
YEARS OF PRODUCTION
about half, and would, in effect, prohibit the direct burning without
clean-up, e.g. in coal-fired electric power generating plants, of
all bulk coals in the U.S. It is often stated that western U.S.
coals are very low sulfur coals. However, most western coals have
apparently not been intensively explored nor analyzed. The western
coals that have been utilized and fully characterized are the highest
grade available, while western coals containing up to 10% sulfur are
known. Therefore, as time and production proceed, the quality of
mined coal can be expected to decline. Therefore, it follows from
the preceeding discussion that the actual future coal utilization
will likely be less than indicated in Table 8, and the growth rate
in coal production might be on the order of 3% during the next 45
years, which will lengthen the life-time of the U.S. reserves.
The total energy of 41.1 million barrels per day of oil equiv-
alent (MMBDOE) that is consumed corresponds to 87.0 Quads. This
energy consumption is broken down into the four principal U.S. sec-
tors in Table 9 and is compared with the energy utilization of 1980.
These numbers can be compared with those derived from the 1990 pro-
jections contained in Figure 11. The first observation to be made is
that the total energy projected to be consumed in 1985 is much too
high. Although the energy usage was 75.9 Quads in 1980, the reces-
sion resulted in a decline to only about 71.8 Quads being consumed
in 1982 (Table 7). This will increase by about 1.5% in 1983. Even
with 5% increases in 1984 and 1985, the total annual energy consump-
tion would only reach 80.34 Quads, which would be 1.4 Quads above
the 1979 level (Table 5).
COHVfAStON
..
A,O
!Lf:crl'll(;lrr TR.I\N5MISStOH
G£N£RAfION "0$$1$
1) t '''1
,U51DENTIAL
"
C:OMMEROAL
"
".
INOU$tAlAl
.
'IilANSPOAT" no'f
price tripled again, but the competitive point was not achieved.
This phenomenon is a case of the "receding break-even point" [39].
While the price of crude oil was increasing, the prices of other
resources were also increasing. Some of these, e.g. all of the energy
COIA'EJl$lON
AHO
rAANSMlSSK>N
(l(CTItiClf¥ lOSS E5
G{N(!V.TlOH
". IN
.. I Ql.jl
TIII,6.HSPQJlfATIQH
"
However, once the plants are established, the costs of raw mate-
rials for coal gasification and for synthesis gas conversion are
projected to be lower than comparable chemical syntheses from petro-
leum [40]. A projection of the relative costs [40] for some chemical
feedstocks in 1990, based on the price per 1b of carbon contained in
the feedstock, is given in Figure 12. On this basis, 1:1 synthesis
gas and propylene will be comparable in price, as will 2:1 synthesis
gas and ethylene. However, the synthesis gas mixtures contain oxygen,
and this results in the real cost of 1:1 synthesis gas to be only
40% of the cost based on carbon content alone. Therefore, it is ex-
pected that in 1990 the relative costs of the chemical intermediate
feedstocks will increase in the order 1:1 synthesis gas < CO < pro-
pylene < 2:1 synthesis gas < ethylene < methanol.
Para form
Formaldehyde (aq)
Methanol
Ethylene 2:1 Synthesis Gas
Propylene 1: 1 Synthesis Gas
Carbon Monoxide
Natural Gas
Crude Oil
Vacuum Residuum Coal
Ethylene Glycol
(1)
However, the first step of the process involves a series of two re-
actions, (2) and (3), that limit the maximum yield of ethylene oxide,
and subsequently ethylene glycol, to 86% [43].
Table 10. Reaction Stoichiometries for the Formation of Selected Chemicals from Synthesis Gas
and Methanol
Reactant
Product Reaction H2/ CO Usagea
Methanol
Specialty Chemicals
lwa.. r
~
SYNTHESIS GAS
1
ETHANOL
lco
-1 Cu/ZnO
METHANOL
j~
ACETIC
ANHYDRIDE
lZSM-S
Figure 13. Valuable chemicals can be produced from natural gas via
synthesis gas and methanol.
CONCLUSION
The chemical industry consumes approximately 10% of the U.s.
energy, and most of it is derived from petroleum. The fluctuations
in the price and availability of petroleum during the last decade has
provided the initiative for seeking alternative feedstocks and alter-
native sources of power. With the recent demands made for interfue1
substitution and inter-feedstock switching, a renewed focus on syn-
thesis gas and alcohols has appeared because of their capabilities
to satisfy both demands. Because of the versatility of these feed-
stocks, it is expected that the growth in production of these will
continue, as will the research and development in methods of utili-
zation.
REFERENCES
Michael S. Graboski
INTRODUCTION
37
38 M. S. GRABOSKI
gas for methanol and ammonia synthesis has been practiced commercial-
ly since 1930. It is the most widely used method of hydrogen manu-
facture in the United States. The major reformer feedstocks are
natural gas, propane, LPG, butane, naphtha and refinery off-gas [1].
L-~~====~~T~O~POWER
GAS/AIR STEAM
WATER
C02
CO 15.5 47.0
COZ 8.1 4.3
HZ 75.7 46.0
NZ+A O.Z 1.4
CH4 0.5 0.3
HZS o 1.0
CO + H20 -+ CO 2 + H2 (2)
GASIFICATION REACTIONS
CO + HZO -+ CO 2 + HZ (Z)
Depending on the fuel being processed, the actual full set of reac-
tions occurring is quite complicated since thermal pyrolysis is an
important reaction step. In gasification systems, the overall objec-
tive is to maximize the chemical energy of (CO + HZ) and minimize
methane in the final product gas. To accomplish this, the fuel is
processed at high temperature in an oxygen deficient environment,
with heat released by the oxidation used to raise the reaction mix-
ture temperature and supply energy for the endothermic reforming
reaction.
OIL GASIFICATION
SOOT
RECOVERY
STEAM
and steam/02 in the reaction chamber. The burner designs are proprie-
tary and feedstock dependent. Back radiation from the walls and flame
vaporize the oil and initiate combustion. Approximately 35 percent
of the stoichiometric oxygen and 0.4 lb steam/lb oil are fed to the
unit. In the flame region, pyrolysis of the oil and combustion take
place at very high temperatures (150QoC). Methane is the principle
hydrocarbon produced in pyrolysis, and some methane passes through
this zone intact. Homogeneous thermal steam cracking in the soaking
zone of the reactor allows methane to decompose. This reaction ceases
below about l300°C. About 3% of the oil by weight leaves as carbon
(soot) in the product gas. Pressure does not affect the reaction
kinetics or composition markedly. Kermode 12] presents data on 02
requirements for different fuels. The gas composition can be esti-
mated by assuming water-gas shift equilibrium at the exit temperature.
Gas composition can be varied by adjusting the 02/steam/fuel mix, and
cold gas efficiencies of 75% are typical.
COAL GASIFICATION
In terms of synthesis gas production, Texaco and KBW are the most
desirable modern day alternatives. While there are many years of
commercial operating experience with Lurgi at Sasol, South Africa
(Fischer-Tropsch to hydrocarbons), it should be stressed that Lurgi
PRODUCTION OF SYNTHESIS GAS 45
Steam
....--........- . ; L - _ - ,
Drying Shift
& t---.j Gasification Converter t----,
Grinding
Compression
Slag Ash
Plant
Steam
MeOH
Synthesis
MeOH-
Distillation CompresSion
Steam Steam
ting data. The Texaco gasifier has been operated at 150 tons per
day and this is equivalent to about 100 TPD of methanol which is
small by today's standards.
SUMMARY
Most synthesis gas produced today comes from natural gas consti-
tuents. Alternately, synthesis gas may be produced from coal and
biomass. Suitable commercial-scale gasifiers, including Koppers-
Totzek, Texaco and Winkler gasification systems are applicable to
both coal and biomass processes. Because of the differences in
feedstock characteristics including reactivity, ash, sulfur, nitrogen
and chlorine, the process requirements for the two feedstock types
are considerably different. Typical process flow schemes for commer-
cial processes operating on coal and wood have been described.
PRODUCTION OF SYNTHESIS GAS 49
REFERENCES
INTRODUCTION
53
54 F. N. LIN AND F. PENNELLA
EXPERIMENTAL METHODS
Catalysts
Catalytic Tests
Chemisorption Measurements
RESULTS
The BET surface areas and the crystallite sizes of the catalysts
are listed in Table 2. The surface area of the calcined catalysts
decreased from about 40 m2 jg to 25 m2 jg as the cobalt content in-
creased from 0 to 8% (Figure 3). A similar trend is evident in the
surface areas of the tested catalyst, but with a greater scatter of
the datapoints. In the low cobalt concentration range (0-0.6 wt %),
where the loss of activity reached 98%, the surface area varied be-
tween 41 and 38 m2 jg. This small variation in surface area does not
56 F. N. LIN AND F. PENNELLA
Yield, g/g/hr
Co Content Higher
Hydrocarbons Methanol
Wt % Alcohols
Co Content °
Average Crlsta11ite Size! A
Wt % Area, m2 /g C0304 (Co) CuO (Cu) ZnO (ZnO)
25.----------------------------------------,
o ALCOHOLS
o HYDROCARBONS
2.0
a:
J:
;::
« 15
'-'
I
~
ci 1.0
...J
W
;;:
0.5
0
0
\ ~
2.0 4.0
COBALT CONTENT, WT. %
6.0 8.0
10
a:
J: 0.5
~
'T
Ol
(j,
9
w
;;: 0.1
...J
o
~ 0.05
J:
ti:i
:::;
o
o
0.0 1!:------;;''=----:-'-:::-__-;:-L::--__~,-----l
o 2.0 4.0 6.0 B.O 1.0
COBALT CONTENT, WT. %
5or-----------------------------------------------~500
o SURFACE AREA
• CRYSTALLITE SIZE OF ZnO
A CRYSTALLITE SIZE OF CuO
!;I!
400
.
<I:
E u.i
<i N
300 iii
w
c:: w
<I: r
W
<.) j
~ 20 200 <I:
c:: r
(/)
::)
(/) >-
c::
<.)
100
O~------~~------~~------~~------~--------~ 0
o 2.0 4.0 6.0 8.0 10.0
COBALT CONTENT. WT. %
DISCUSSION
Cl
o
Yield, g/g-cat/hr a
Reduction Total
Catalyst Temperature (OK) Methanol Hydrocarbons
Alcohols
COO.3ZnA10.3 523 o o 0
Co O• 3 ZnA1 0 • 3 773 0.012 0.003 0.11
1% Cu/CoO.3ZnAlO.3b 523 0.014 0.006 0.09
ZnA1 0 • 3 523 o o 0
5% Cu/ZnA10.3b 523 0.18 0.18 0
:n
z
aAt 543°K, 6.5 MFa, 8000 GHSV
C
bCopper was added by impregnation of the calcined catalyst with an aqueous solution of cupric z
nitrate. »
z
0
:n
"'tJ
m
Z
Z
m
r
r
»
COBALT ON COPPER - BASED CATALYSTS 61
A: TOTAL ADSORPTION
B: WEAK ADSORPTION
C: STRONG ADSORPTION, A-B
f-
(f)
>-
...J
«
f-
«
0
I
0>
00
LL
0
(f)
UJ
...J
0 20
~
::t • • C •• ••
0
0 80 100
CuCrl.2AlO.9 0 0.37
CuCrl.2AlO.9 0.1 0.36
CuCrl.2Al O.9 0.45 0.40
CuCrl.2AlO.9 1.80 0.35
CuCrO.26 A1 0.04 0 0.26
CuCrO.26AlO.04 0.27 0.16
CuCrO.26AlO.04 0.55 0.11
CuCrO.26AlO.04 1.1 0.036
CONCLUSIONS
The purpose of this work was to determine whether copper sites
active for methanol synthesis are modified by cobalt to produce
higher alcohols. The results obtained indicate that on either cop-
per-zinc or copper-chromium-based catalysts, cobalt deactivates the
sites and does not alter their selectivity. The formation of higher
alcohols and hydrocarbons observed on CuZnAlO.3 at high concentra-
tions of cobalt can be attributed to the formation of cobalt ensem-
bles and not to modifications of the methanol synthesis sites.
REFERENCES
INTRODUCTION
65
66 G. B. ATKINSON ET AL.
Alloy Preparation
Activation
The alloys were ground to minus 25 plus 80 mesh and were indi-
vidually tested to develop surface area by treating with different
oxidizing gases at elevated temperatures. The gas mixture must have
sufficient oxidizing power to convert the reactive metal to its
oxide. Frequently used gas mixtures included air, air saturated with
water, carbon dioxide, hydrogen saturated with water, and hydrogen
and carbon monoxide. Th2Cu and ThCu2 alloy buttons disintegrated
when the thorium was oxidized to thoria in air at ambient conditions.
ThCu3.6 partially disintegrated under these conditions.
350· C
u Thermocouple
n
Catalyst-+__fm
bed
Liquid
Synthesis gas trap
and helium
Liquid
drain
Temperature
Catalyst Activation Total area CH 30H at
Treatment mZ/g mol % maximum, °c
to treat the alloy for surface area development. The total surface
area was measured after methanol testing. Results are reported as
the maximum mole percent methanol in the product gas and exclude
hydrogen. The temperature at which each maximum was obtained is
included.
The surface area developed and the methanol produced were low
for catalysts prepared with zinc as the active catalyst component.
Maximum methanol production was obtained at temperatures of 360 to
400°C. Copper-containing catalysts were more active than the zinc-
containing catalysts and maximum methanol production was obtained
at temperatures of Z80 to 350°C. Maximum production of methanol for
some of the alloy-based catalysts was greater than the commercial
catalyst and was obtained at lower temperatures. The alloy ThCu5
developed surface area by several different activation methods. Ox-
idation with air at 400°C produced the most active catalyst. A
series of thorium-copper alloys was prepared and investigated to
determine the best composition.
7or-------------r_----~------r_------------r_----------__,
Results from testing small samples of ThCux alloys (0.2 m1) for
methanol activity showed that ThCu6 was the best composition, and
THORIUM-COPPER INTERMETALLIC CATALYSTS 73
7or-------------,------,r------.--------------r-------------~
~ 30
~
o
z
~ 20
I-
W
::E
10
~~0-0------------2~5~0------------~30~0-------------3~5~0------------~400
TEMPERATURE, ·C
Figure 3. Methanol production as a function of temperature. Syn-
thesis gas: l6H2:CO, 60 atm pressure, 31,000 GHSV.
"0 30
.............................. ------- ThCU9. 15.0 m2/g
E
------,
- - - ThCuCr. 23.7 m2/g
......
N .....
:I:
0
,.,."..""" ,.,---........"
c: A--- __ " "-
'0 /"" / -~~,-
.2 20
~
u
)(
...J
,/
, /
"
,,,/
I "I ,
,
0
Z
«
:I:
/,,..,,,, /
I- 10
w
~
O~ _ _ _ _ _ _ _L-____________L-____________L -__________-...J
200 250 300 350 400
TEMPERATURE. °C
SUMMARY
Temperature Time Wt %
Catalyst
°c hr CH30H H20 C2H50H >C2H50Hb
% CO
Temperature g CH30H/ % CH30H in converted to
°c ml/hr exit gas CH30H
9B-173:
180 1.04 3.75 13.0
200 1.59 5.87 19.4
210 1.91 7.19 23.2
230 2.77 11.06 33.6
250 3.55 15.46 44.5
Cu/ZnO/Al203:
240 0.10 0.32 1.1
280 0.24 0.82 2.8
300 0.37 1.27 4.3
320 0.52 1.84 6.2
340 0.49 1.77 6.0
THORIUM-COPPER INTERMETALLIC CATALYSTS 77
15 4#-
'-
a~
CHsOH en
en
cr
cr
C)
C) 0
1&1
!: Q.
X !!:
1&1 II::
~
~ II)
...J
0 ~
z
cr
::E:
~
1&1
2
TIME, h
Figure 5. Catalytic stability of ThCu6 (9B-120). Synthesis gas:
H2:CO = 2.4, 60 atm pressure, 9,500 GHSV, and at 260°C.
20
a~
encr
C)
!:
x
1&1
~
...J
0
Z
cr Cu/ZnO/AI20S
::E:
~ 60 a~t;.;.m_ _......
1&1
2
300 340
TEMPERATURE, aC
REFERENCES
INTRODUCTION
81
82 J. F. KNIFTON ET AL.
(1)
ALCOHOL/ESTER GENERATION
Continuous Processing
aTypical operating conditions: 220°C; 430 atm constant pressure; CO/H2 (1:1); catalyst charge:
Ru, 2-8 mmole, quaternary salt, 10-25 g.
bLiquid product analysis by glc, also detected: water, alkyl formates, butanols, acetaldehyde,
2-methyl-l,3-dioxolane, dialkyl ethers, glycol monoalkyl ethers and ethylene glycol; C02 and CH4
present in the product gas samples, along with larger quantities of unreacted CO/H2.
cSmall quantities of methyl propionate and ethyl propionate present in these fractions. to.
dLiquid yield (wt %) calculated basis total catalyst charge. "T1
eOperating pressure, 272 atm. A
fRun in p-dioxane solvent, operating pressure 544 atm. Z
"T1
-t
o
Z
m
-t
»
r-
ALCOHOL/ESTER FUELS FROM SYNTHESIS GAS 85
(Z)
CO + ZHZ + CH30H + CZH50H + HZO (3)
260
240
OPERATING
TEMPERATURE
( ·C)
220
200 •
/
10 20 30 40 50 60 70
ALCOHOL- ESTER PRODUCTIVTY ( g/llr)
Various CI-C4 alcohols have been tested, including the typical prod-
ucts of Table 1. To some degree the solvent s.tructure has been found
to influence final product composition. Where, for example, CO
hydrogenation is conducted at temperatures close to, or higher than,
the critical temperature of the solvent alcohol (see Table 4), the
latter will of course be for the most part in the gas phase. In
turn this leads to changes in the molar concentration of ruthenium
"melt" catalyst in the reactor and an alteration in the effective
productivity of the system. The net increases in observed produc-
tivity, illustrated in Figure 2 for syntheses conducted at 240°C are
correlated in eq. 6:
aOperating conditions: gas feed rate, 200 STP l/hr; liquid feed
rate, 20 cc/hr; 476 atm total pressure.
bExperimental series where gas composition is 1/1: liquid feed
contains 7555 ppm ruthenium (as RU3(CO)12)' 36.8% BU4PI, and the
remainder is n-butanol.
CCatalyst decomposition is evident in this temperature range.
dExperimental series where gas composition is 1/2: liquid feed
contains 7318 ppm ruthenium (as RU3(CO)12), 37.1% BU4PI and the
remainder is n-butanol.
Catalyst Life
PROCESS DESCRIPTION
Effect of Gas and Liquid Feed Rates Upon Liquid Product Composition
aExperimental series where liquid feed contains 10,264 ppm ruthenium (as RU3(CO)12)' 47% BU4P1,
and the remainder is methanol; operating conditions are: 240°C; 476 atm pressure; CO/H2: 1/1.
bExperimental series where liquid feed contains 10,264 ppm ruthenium (as RU3(CO)12)' 47% BU4P1
and the remainder is ethanol; operating conditions are: 240°C; 476 atm pressure; CO/H2 : 1/1.
'-
"'T1
A
Z
"'T1
--I
o
Z
m
--I
»
:-
ALCOHOL/ESTER FUELS FROM SYNTHESIS GAS 89
400
SYNTHESIS GAS
FLOW
(STP IIhrl
300
200
100
20 30 40
240°C under pressure (450 atm). After a 3-hour hold period, synthe-
sis gas conversion is ca. 40%; synthesis gas is present in excess.
Productivity is 0.3 grams of liquid oxygenate cc-l hr-l.
PRODUCT PERFORMANCE
100
•
80
RUTHENIUM 60
CONCENTRATION
(ppm.102 )
40
20
O~--~-------r--------r--------r-
20 40 50
----.-_-..:._---.
110
"- (METHANOL
...............
40
•
a --a
_ _ _ _- : - a - a - - - - =
ALCOHOL a_a-a a \.
CONTENT (wt '11.1 30
~ ~ETHANOL
a
20
" , - PROPANOL
...
10
. _9----9-- 9--9-9
OL-----~--------~------~--------r_------~------
100 200 300 400 500
HEAVY
OXY8ENATES
SUMMARY
Example I II V VI
-III- -IV-
Catalyst composition . RU3 (COh2-Bu4PI RU3(CO)12-Co2(CO)8-Bu4PBr
Solvent n-BuOH n-BuOH EtOH MeOH EtOH c
Productivity:
Cl-C3 alcohols (g/hr) 8.4 39.0 31.0 29.1 36.6 8.7
Cl-C3 alkyl acetates (g/hr) 0.1 1.1 2.8 3.1 3.6 3.8
Volume (%)
Product mix "A" Product mix "B"
Component C1-C3 alcohols C1-C3 alcohol/esters
REFERENCES
A GORDIAN KNOT?
Cheryl K. Rofer-DePoorter
Geochemistry Group
Los Alamos National Laboratory
P.O. Box 1663
Los Alamos, NM 87545
INTRODUCTION
or its reverse, the water gas reaction, may occur. The water gas
shift is used to regulate HZ and CO concentrations in synthesis gas,
and it occurs as part of synthesis gas reactions such as the Fischer-
Tropsch synthesis, in which CO and HZ are starting materials, and
COZ and HZO may be products. At typical synthesis gas reaction tem-
peratures, the equilibrium constant for the water gas shift is close
enough to unity and the reaction proceeds rapidly enough that all
four species can be expected to be present. The economics of syn-
thesis gas reactions usually require that oxygen rejection be via HZO
rather than COZ; therefore, control of the oxygen-carrying product
through the water gas shift may be desirable.
97
98 C.K.ROFER-DePOORTER
system to iron's activity in the water gas shift reaction. The dif-
ferences between HZO- and COZ- producing catalysts are rather within
some subset of elementary reactions occurring under Fischer-Tropsch
conditions to give an effective water gas shift. Identification of
these elementary reactions will allow the design of improved Fischer-
Tropsch catalysts.
The overall kinetics of the water gas shift reaction have been
studied on unsupported iron [18] and platinum [19], on alumina-sup-
ported Group 7B, 8, and lB metals, some of which were also supported
on silica and carbon [20], on zeolite- and alumina-supported rhodium
[21], and on zeolite-supported ruthenium [22].
co + M ! M-CO
H20 + 2 M t M-H + M-OH
M-CO + M-OH ! adsorbed intermediate
adsorbed intermediate t M-C02 + M-H
M-C02 t M + C02
2 M-H t 2 M + H2
The water gas shift catalysts used industrially have been iron
and chromium oxides (high-temperature catalyst) and copper, zinc,
chromium, and aluminum oxides (low-temperature catalyst). Limited
WATER GAS SHIFT FROM FISCHER-TROPSCH 101
H20 + MO ~ HO-MO-H
H20 + MO ~ H20-MO
H20 + 2 M ! M-OH + M-H
CO + M ! M-CO
M-CO + H20-MO t OM-O(H)C(H)O
OM-O(H)C(H)O + M ! M-O(H)C(H)O
M-O(H)C(H)O t M + C02 + H2
Hydrogen Adsorption
HZ + M t M-HZ(physisorbed)
M-HZ(physisorbed) t H-M-H
CO + M t M-CO(physisorbed)
M-CO(physisorbed) t M-CO
HZO + M t M-HZO(physisorbed)
M-HZO(physisorbed) t M-OHZ
COZ + M t M-COZ(physisorbed)
M-COZ(physisorbed) t M-COZ
M-CO t M-~I
o
M-1IIC ++ O-M-C
o
M-OHZ t H-M-OH
M-OH t H-M-O
M-CO + M-O M-COZ +
M-CO + M-OH t M-C(O)OH
M-C(O)OH +
H-M-COZ
HZ + MO t MO-HZ(physisorbed) (4)
Water Adsorption
CO + MO t OM-CO(physisorbed) (12)
(14)
hydrogen transfer,
or oxygen insertion,
Carbon monoxide and water, as well as C02 and H2, have been co-
adsorbed on zinc and magnesium oxides and the surface intermediates
observed by infrared spectroscopy [101]. Formate ions were the pre-
dominant species observed. Formate [102,103], acetyl and acetate
[102], and carbonate and bicarbonate [103] intermediates have been
trapped during CO-H2 and C02-H2 reactions on copper-zinc catalysts.
Water and CO adsorbed on an iron-chromium catalyst interact to pro-
duce formate [41]. A possible reaction producing formate is
This is an oxygen insertion into the M-C bond (or carbon migration
to oxygen) with hydrogen transfer. It may be broken down into
Reaction Number
H2 + MO t MO-H2(physisorbed) (4)
MO-H2(physisorbed) t H-MO-H (5)
H-MO-H + 2 MO t OM-H-MO-H-OM (6)
H20 + MO t H20-MO(physisorbed) (7)
H20-MO(physisorbed) t H20- MO (8)
H20-MO(physisorbed) t MO-HOH (9)
H20-MO t HO-MO-H (10)
MO-HOH t HO-MO-H (11)
CO + MO t OM-CO(physisorbed) (12)
OM-CO(physisorbed) tOM-CO (13)
CO + MOMO t M2C03 (14)
C02 + MO t MO-C02(physisorbed) (15)
MO-C02(physisorbed) tOM-C02 (16)
MO-C02(physisorbed) t MO-C02 (17)
MO-C02(physisorbed) t OM-OCO (18)
OM-OCO + OM-OH t OM-OC(O)OH + MO (19)
MO-C02 + OM-OH t MO-C(O)OH + OM-O (20)
OM-C02 + OM-OH t OM-OC(O)OH + MO (21)
OM-CO + OM-OH t OM-C(O)OH + MO (23)
OM-C(O)H t OM-OCOH (24)
OM-DCOH t OM-OC(H)O (25)
M-OCO + M-H t M-OC(H)O + M (26)
OM-CO + OM-O tOM-C02 + MO (27)
OM-C02 + OM-O t OM-OC02 + MO (28)
Carbon dioxide can also react with metal complexes to give formates
[142]:
The equilibrium
. (33)
Reaction Reference
Reaction ~umber
co + M t M-CO (37)
The water gas shift reaction has also been photocatalyzed with a ru-
thenium complex, apparently because of photo-initiation by CO disso-
ciation from the complex [149].
(40)
Some of the products appear not to play a part in the water gas
shift reaction except to remove a small amount of material from the
system. Carbonates and the inactive form of hydrogen on oxides are
in this category.
METHANOL SYNTHESIS
Although the scope of this paper does not permit a detailed dis-
cussion of methanol synthesis, that reaction must be closely related
to the water gas shift over the low-temperature catalyst, since the
catalysts are similar. An ideal situation would be to be able to
write a list of elementary reactions encompassing both methanol syn-
thesis and the water gas shift. However, the intermediates in meth-
anol synthesis have not been unambiguously identified [184]. Clearly
the largest difference between the two processes is in the oxidizing
power of the reactants: synthesis gas for methanol is more reducing
than either side of the water gas equilibrium. However, small
amounts of C02 promote methanol synthesis. Again, as in the Fischer-
Tropsch synthesis, adsorption-desorption of HZ, HZO, CO, and COZ and
association-dissociation reactions of HZ, H20, and C02 are probably
common to both processes. The co that is hydrogenated to methanol
remains associated, as is likely for the CO that is oxidized in the
water gas shift. It is not clear whether the intermediates in co hy-
drogenation are bonded to the catalyst through oxygen (formate-meth-
oxy route) or through carbon (formyl-hydroxymethyl route), or whether
both of these routes are significant [185].
ACKNOWLEDGEMENTS
REFERENCES
INTRODUCTION
129
130 J. V. BAUER ET AL.
elements:
BACKGROUND
In principle then, for reactions (1) and (Z), where olefins are prod-
ucts, CO/HZ utilization ratios may vary from 0.5 to Z. Where paraf-
fins are produced (e.g. (4) and (5», the CO/H2 usage ratio may vary
from n/(Zn + 1) to Zn/(n + 1), depending on whether the by-product
rejected is water or carbon dioxide. For example, for a paraffin
SLURRY PHASE FISCHER-TROPSCH PROCESS 131
product with an average molecular weight of 15, the CO/H2 usage ratio
can vary from 0.48 to 1.88.
EXPERIMENTAL
r-------,
I
I
BACK
I
I PRESSURE CS tC6 GAS
GAUGE SPLITTER SAMPLES
I
FLOW
I
CONTROL
OIL
SAMPLES
FLOW SLURRY
CONTROL
RESERVOIR
The product stream from the partial reflux condenser is led via
a heated line to a pressure reduction stage, and thence to a CS/C6
splitting column, to produce a condensed liquid phase and a gaseous
phase. The gaseous stream consists of unreacted CO/H2, C02 and prod-
ucts with carbon numbers from C1-CS. The condensed liquid product
consists of hydrocarbons with carbon numbers C6 and above and an
aqueous phase containing dissolved oxygenates. This method of prod-
uct collection avoids the use of high pressure traps, is more suited
to continuous operation, and, by reducing the number of product frac-
SLURRY PHASE FISCHER-TROPSCH PROCESS 133
J C,/C,
1 SPLITTER I
I SC,
CARLE 397B
GAS CHROMATOGRAPH r- ~
1~C.
REACTOR C,-C,; CO; CO.; H.
I
I
I
FILTERED
I
I
CARLE 111 I
-~
EXCESS AQUEOUS/ORGANIC
SLURRY PHASE
SEPARATION
.... GAS CHROMATOGRAPH
C.+ VAPOR
,. I
I WAXES
0
,.
:zI
C)
I:)
c:
I
I
I1~
m I
0
c: I
III
SIGMA 2 I
I
~
MA INFRAME COMPUTER LIQUIDS ORGANICS
AUTO-SAMPLER I GAS CHROMATOGRAPH
ARCHI VE AND CALCULATIONS
CAPILLARY COLUMN
I
t C,-C ••
I
I
I
I I
I
I
I
D ATA ACQUISITION I
SIGMA 1
""'~I
COM PUTER SUBSYSTEM AQUEOUS
•
GAS CHROMATOGRAPH
ROH; RCO.H; ETC •
I
I
I
SIGMA 1 I
~
ORGANICS GAS CHROMATOGRAPH
I-
NON-CAPILLARY
WAXES C,-C •• I
,.-- - - - - - ------------; I
SIGMA 10
~----- COMPUTING INTEGRATOR
_ _ _ SAMPLE FLOW
- -- - DATA FLOW
Catalyst Preparation
RESULTS
a64.0% C18-C35 !-
<
CD
»
c
m
:D
m
--i
»
:-
C/l
r-
C
:II
:II
-<
."
::z::
»
C/l
m
Table 2. The Dependence of a and Product Type on CO/H2 Ratio; Slurry Phase, Baseline Fe203a en"
n
::z::
m
:II
Wt % of H~drocarbon Oxygenates, .:...
Slurry cone., P, T, Conv. , ;0
CO/H 2 a branched mol % of
wt% MPa °c mol % n-a1kanes 1-a1kenes o."
isomers product C/l
n
::z::
."
:II
19.3 3.34 260 0.5 30.4 0.72 46.1 34.4 19.5 0.2 o
19.3 3.31 280 0.5 43.4 0.71 41.6 46.4 12.1 0.1 n
m
14.7 3.17 248 1.4 24.7 0.78 25.4 50.1 24.5 7.1 C/l
C/l
14.7 3.14 277 1.4 50.8 0.78 24.9 49.0 26.1 1.2
14.7 3.17 251 2.8 9.0 0.70, 40.7 40.5 19.1 2.2
0.93 b
14.7 3.10 281 2.8 27.7 0.70[, 40.0 44.2 15.8 2.8
0.93
aMean material balance for tabulated data: C 98.9 ± 1.5% and H 98.6 ± 2.1%
bFor >C10 hydrocarbons
w
......
138 J. V. BAUER ET AL.
25~----------------------------------~
COIH, - 0.5
T - 280 ' C
P - 3.31 MPa
20
15
WEIGHT
PERCENT
10
5 10 15 20
CARBON NUMBER
~ N-ALKANE
_ 1-ALKENES
~ BRANCHED ISOMERS
25,-----------------------------------,
CO/H, - 1.4
T - 277' C
P - 3.14 MPa
20
15
WEIGHT
PERCENT
Figure 4. Hydrocarbon
10 weight distribution:
Fe2 03
5
5 10 15 20 25 30
CARBON NUMBER
~ N-ALKANE
_ l-ALKENES
~ BRANCHED ISOMERS
SLURRY PHASE FISCHER-TROPSCH PROCESS 139
25~----------------------------------~
COtH, - 2.8
T - 281 "C
P - 3.10 MPa
20
15
WEIGHT
PERCENT
10
5 10 15 20 25 30
25~---------------------------------'
COtH, - 0.9
T - 252°C
20 P - 2.03 MPa
15
WEIGHT
PERCENT
10
5 10 15 20
CARBON NUMBER
~ N-ALKANE
_ 1-ALKENES
mm BRANCHED ISOMERS
Figure 6. Hydrocarbon weight distribution: Catalyst Al
140 J. V. BAUER ET AL.
25r-----------------------------------~
COIH, - 1.0
T - 237 · C
20 P - 2.21 MPa
WEIGHT 15
PERCENT
10
40
25 r------------------------------------,
COIH, - 1.0
T - 241 · C
P - 1.12 MPa
20
15
WEIGHT
PERCENT
10
5 10 15 20 25 30 35 40
~ N-AlKANE CARBON NUMBER
_ 1-AlKENES
~ BRANCHED ISOMERS
25r----------------------------------,
COIH, - 1.0
T - 247"C
20 P - 2.07 MPa
15
WEIGHT
PERCENT
10
o
5 10 15 20 25 30 35 40
25 r----------------------------------,
COIH, - 2.0
T - 2SO' C
20 P - 2.08 MPa
15
WEIGHT
PERCENT
10
5 10 15 20 25 30 35 40
~ N-ALKANE CARBON NUMBER
_ l-ALKENES
~ BRANCHEDISOMEAS
DISCUSSION
(6)
80
f-
U
~
C.. +
C 60
0
a:
Q.
u.
0
'#.
40
~
20
C9- C25
Schulz-Flory Maxima
gasoline range 0.76 5.8 47.6 31.8 0.7
diesel range 0.88 1.4 31.9 S4.l 12.9
Baseline Fe203
CO/H2 O.S 0.70 17.4 37.6 17.3 0.1
CO/H2 = 1.4 0.78 6.8 41.1 36.2 2.7
o.---------------------------------~
P T
Curve CO/H. MPa a
2.8
°C
--
-1 .. (a) 3.10 281 0.70,0.93
(b) 1.4 3.14 277 0.78
(e) 0.5 3.31 280 0.71
-2 {
W· \ ......
LOG (_I)
Cj ........ , ° 0 • ---.
-4 • (e) •
-5~--~----~----~~
5 10 15
____ ____ ____
~
20
~
25
~
30
CARBON NUMBER
basis falls into two distinct regions separated at -ClO, with the
ClO-C30 hydrocarbons following a separate line with a high a value
of 0.93.
Table 4 lists some published data for (i) the commercial Fisch-
er-Tropsch units at SASOL [10,13], (ii) a pilot plant three-phase
bubble column reactor at Rheinpreussen, 1.6 m id x 8.6 m high [3],
and (iii) some bench scale slurry column experiments carried out by
KBlbel [12]. The operation at Rheinpreussen was geared to maximize
the yield of gasoline. Some doubt exists about the overall product
analysis, since the data reported in Table 4 was stated to be from
a "sample from large storage tanks which contained the products from
one of the longer operating periods" [3]. It may not, therefore,
fully represent the percentage of light products obtained from the
plant, which were removed by oil scrubbing and activated carbon ad-
sorption. With this in mind, a comparison of the product fractions
of Table 3 with the data of Table 4 indicates that the SASOL reactors
and previous slurry phase operations gave products that again were
unselective and of the Schulz-Flory limited type. Satterfield et
al. have also shown that the slurry phase data obtained by KBlbel can
be correlated by a Schulz-Flory distribution using the appropriate
value of a [6].
Table 4. A Comparison of Commercial SASOL and Prior Slurry Phase Fischer-Tropsch Data
o r-----------------------------------------~
P T
Curve COIH. MPa ~ Catalyst
-1 (a) 1.0 2.21 237 } A2
\ .... (b) 1.0 1.12 241
~ (e) 1.4 3.17 248 Fe 2 0 3
-2 .~.I\ •••••
W· , '.
LOG (i-) ~:~,
.' ....... (a)~
-3 .........-'"'_"""'::-... ' - ' - ' (b) -.
.
. • (e)·· . '.,
...... .........
-4 '""
5 10 15 20 25 30 35 40
CARBON NUMBER
Or---------~------~~~--------------__,
P T
Curve CO/H. MPa °C Catalyst
(a) 2.0 2.08 250
-1 (b) 1.0 2.09 247 J B
(e) 2.7 3.17 253 Fe.O.
-2
W·
LOG( Cjl )
-3
-4
5 10 15 20 25 30 35 40
CARBON NUMBER
CONCLUSIONS
Selectivity
ACKNOWLEDGEMENTS
REFERENCES
U. S. Department of Energy
Pittsburgh Energy Technology Center
P. O. Box 10940
Pittsburgh, PA 15236
INTRODUCTION
If light olefins are used as feed, ZSM-5 can catalyze the con-
version to a high octane liquid product via steps (b) and (c) out-
lined above. Since olefins and oxygenates are obtained as some of
the products of synthesis gas conversion by the Fischer-Tropsch
reaction, there is considerable interest in combining a typical
Fischer-Tropsch metal such as Fe, Co, or Ru with ZSM-5 to form a bi-
functional catalyst. It has been shown that such a catalyst can
convert synthesis gas to gasoline-range hydrocarbons in one step
[5-8].
151
152 v. U. S. RAO ET AL.
EXPERIMENTAL
Preparation of Catalysts
The products of the reaction were mainly C02, H20, and hydro-
carbons. A hot trap at 175°C, following the reactor, collected wax
and heavy oil. A cold trap at O°C, following the hot trap, collect-
ed water, gasoline-range products, and light oil. The remaining
product that was not in the condensed phase and the unconverted feed
gas moved past the back pressure regulator. At atmospheric pressure,
the gas mixture passed through a wet test meter that measured the
total volume of gas that had passed during a test period. Samples
of the exiting gas were collected periodically and analyzed using a
gas chromatograph for H2' CO, C02' and Cl-C7 hydrocarbons. The
liquid product collected in the cold trap was separated physically
into aqueous and hydrocarbon fractions. A mass balance was performed
to provide information such as that shown in Table 1. It is assumed
that the liquid hydrocarbon product consists of Cs+. Material re-
covery, as shown by the mass balance procedure, was usually better
than 96%.
The product collected in the hot trap was found to be very small
in quantity in comparison to that collected in the cold trap. The
hydrocarbon product from the latter was analyzed by ASTM-D-2887 GC
simulated distillation. It was also separated into aromatics, ole-
154 v. U. S. RAO ET AL.
Table 1. Conversion and product distribution from Co-ZSM-5 cata-
lysts during the initial 24 hour period. H2/CO = 1;
P = 21 bar; T = 280°C; Feed rate = 0.659 gig cat/hr
Physical admixture
Method of Aqueous with precipitated
adding cobalt nitrate solution cobalt oxide
and the reaction mixture, consisting of 30 vol% C2H4 and 70 vol% H2,
was introduced at 1 atm and a flow rate of 0.983 gig cat/hr. The
liquid product, which consists of hydrocarbons, was collected in a
trap at O°C. The remaining effluent, which was in the gas phase,
consisted of H2 and hydrocarbons. Samples of the effluent were ana-
lyzed for H2 and Cl-C7 hydrocarbons. Mass balances were performed
to obtain the information shown in Table 2. The liquid hydrocarbon
product was analyzed as described earlier.
RESULTS
A B C
Catalxst H-ZSM-5 1.7% Co-ZSM-5 1.7% Co-ZSM-5
Physically Ion-Exchanged
admixed
80r-----~----,-----_r----_r----_,
70
i
...= 60
~
in
ffi 50
>
8N
! 40 Catalysts: Cabalt-ZSM-5
o
o • Salution impregnated
• Physically admixed
30 Feed: H2/CO • 1.0
Test canditlons: Ts 280°C, P.21
bar, WHSV-0.66
Periad: 0-24 haur.
200~----~----~4----~6~----~--~10
tially higher, the C5+ yield was lower, and there was an increase
in the aromatic fraction of the liquid product. For example, the
solution-impregnated catalyst 5.9% Co-ZSM-5 of Table 1, when examined
at a reaction temperature of 300°C, showed CO conversion of 63.0%.
The methane selectivity was 42.0%, the C5+ selectivity was 38.7%,
and the fraction of aromatics in the liquid hydrocarbon product was
42.0%.
.e
Catalyst· Cabalt-Z SM-5 (physically admixed)
containing 9.2 w t. % cobalt. -
c
..
°
~
• 60 ~ 60
~
X.
z
~ A..... •••
~..J 40
~~---- ..----_.A.~--r-.=:.::::-:::..=-_
_ . - .... ..
----:I
Q
VI
It:
_ • ...-. 40~
z
o
l:i
.,'
-"'-'
Feed: Hz/CO = I • Conversion
u
'"
:I:
20 T=280°C, P=21 bar
WHSV=0.66 • C5+ 20 6u
• C,
o 14
80r-----,-----~------~-----r----~~----~----~
.e
° .,Jt.
(5
> 60
",.""
Z "".
Q
",,"" "" ""
~
in
0
Q.
:E
40
I"
0
I
.
u
e I
::::l
0 20 I
::i I • Aromatics
",II • Olefins
0 14
TIME ON STREAM,days
(A) H-ZSM-S
(B) 1.7% Co-ZSM-S (physically admixed)
(C) 1.7% Co-ZSM-S (ion-exchanged).
The results are shown in Table 2. The data have been provided for
two separate periods, 0-4 and 4-8 hr on stream, respectively.
6 0.0 0.0
7 11.6 8.4
8 37.9 31.0
9 30.8 36.2
10 14.5 19.0
11 3.8 4.4
12 1.3 1.0
in the reaction mixture may have been responsible for the slower
decline in the rate of aromatics production than would have been the
case otherwise.
Catalyst Characterization
CONCLUDING REMARKS
From the results and the tone of the discussion, it might appear
at first that metal-zeolite catalysts prepared by the physical ad-
mixture method may be superior to those prepared by solution-impreg-
nation methods, since the latter are likely to cause alterations in
the acid sites of the zeolite. The impregnation methods, however,
require further investigation, since they can provide ways to alter
the degree of metal dispersion in the catalyst. Metal carbonyls
[16,18] and metal-organic compounds are more likely to yield higher
metal dispersions on zeolite supports than solutions of metal salts.
Well-dispersed Fe-ZSM-5 [16,18] and carbon-supported iron [19] cata-
lysts with Fe particle size in the range 30 to 70 A show superior
maintenance of F-T activity with time. It has been reported [20]
that metal-exchanged (Fe 2+, NH4+)Y can be reacted with an anionic,
metal-containing coordination compound that is water-soluble, such
164 v. U. S. RAO ET AL.
as (NH4 )3[Fe(CN)6], to yield an insoluble compound, Fe3[Fe(CN)6]2,
distributed throughout the zeolite while the zeolite itself returns
to the ammonium form. The insoluble complex can later be reduced
in hydrogen to finely dispersed metal in the zeolite. It is thus
evident that impregnation and exchange techniques need to be inves-
tigated further to obtain metal-zeolite catalysts with highly dis-
persed metal components for the study of F-T and related reactions.
ACKNOWLEDGEMENTS
REFERENCES
INTRODUCTION
EXPERIMENTAL
167
168 F. G. DWYER AND W. E. GARWOOD
It is striking that the 4/1 ZSM-5/Fe-K mix gives the same re-
sults when diluted fourfold with quartz, i.e. the ZSM-5 component
interacts with the Fe-K component even when separated from it by
many molecular distances. This will be discussed in more detail
later as it relates to mechanism.
CO Conversion, wt % 95 96 94
H2 Conversion, wt % 54 53 45
Wt % C Converted to
C02 32 31 31
Hydrocarbon , 68 69 69
The C2-C6 data from Table 5 were used to calculate the mole %
distribution of the olefins for comparison with equilibria, shown
in Table 6 (equilibria data are available through C6 only) [12].
The partial pressure of the total olefins at the exit of the reac-
tor is about 6 psia, and the comparison is made at this pressure.
3000r----.-----r----~----._--_,----~
2000
•
1000
..
500
0
II:
~ 200
~
100
50
20
10~ ___ 4_ _ _ __ L_ _ _ _ ~_ _ _ _~ _ _~ _ _ _ _ ~
o 10 20 30 40 50 60
Olefin Content, wt %
Mechanistic Implications
R-CH-CH3
R-CH=CH2 + H-M M
I
R-CH T CH2-M
} + Further Growth.
172 F. G. DWYER AND W. E. GARWOOD
3000
2000
1000
500
0
ia:
q: 200
C
2 100
eo
50
20
10~ __~____~__~~__~____~__~~__~.
o 10 20 30 40 50 60 70
Olefin Content, wt %
CONCLUSIONS
3000.----.----~--~,_--_r----~--_.
50
20
10~---L----~--~-----L----~--~
o 10 20 30 40 50 60
c.+ 0lefln5, wt %
Figure 3. Effect of Si02/Al203 ratio
on C6+ olefin content.
'74 F. G. DWYER AND W. E. GARWOOD
3000
2000
1000
500
.2
-;
a:
cS 200
~
a
iii
100
50
20
10~ __~____~__~~__~____~__~
o 10 20 30 40 50 60
Co' Aromatics, wt %
CO Conversion, wt % 95 97 94
H2 Conversion, wt % 54 59 45
H~drocarbon Carbon No. Distribution, wt %
C1 17 17 18
C2 4 6 8
C3 8 4 6
C4 18 11 14
C5 11 10 13
C+ 42 52 41
6
Olefin Content b~ C No. wt %
C2 6 26 40
C3 7 45 53
C4 4 53 68
C5 3 57 74
C6+ Aromatics, wt % 50 13 13
C6+ 01efins, wt % 1 22 37
1-Cs= 3 3 2 2
2M1C4= 19 19 19 21
3M1C4= 2 2 2 3
trans-2Cs= 12 13 12 9
cis'-2Cs= 6 7 6 8
2M2C4= 58 56 59 57
2 10 40
3 7 53
4 17 18
5 16 74
6 15 62 3 1
7 12 59 8 1
8 11 32 -25 a 10
9 7 15 -50a 27
10 4 <1 89
11 1 <1
100
Equilibria
Found 6 psia
C2= 20 2
C3_= 12 12
C4- 29 35
c5= 24 36
C6= 15 15
100 100
SYNTHESIS GAS TO OLEFINS 177
ACKNOWLEDGNENTS
REFERENCES
INTRODUCTION
179
180 J. G. GOODWIN, Jr. ET AL.
EXPERIMENTAL
Approximate Redox
Metal -LlHf
Source Surface Area Potential
Oxide (mZ/g) (kcal/mol)
(mV)
RESULTS
DISCUSSION
Tauster et ale [7] have suggested that SMSI properties are re-
lated to the reducibility of the support. They found this criterion
to hold for 12 out of 13 transition metal oxides investigated; only
184 J. G. GOODWIN, Jr. ET AL.
80
w 70
z
«
~ 60
w
~
#- 50
5 40
I-
w
3: 30
20
10+-~--~~--~~--~-,--~
200 220 240 260 280 300 320 340 360
TEMPERATURE, C
Figure 1. Effect of the support on methane selectivity.
Cr203 did not fit the correlation. Table 4 gives the reducibilities
of the supports investigated in this study, as well as the metal
oxides (MnO, Nb20S' and Ti02) found by Tauster et al. [7] to have the
strongest metal support interactions. One might expect ZnO to be an
excellent candidate for SMSI. The reaction results in this study
#-
5 80
I-
iii
3:
~-
y 60
N
()
z
5 40
1=
()
«
a:
u.
~ 20
u.
W \
....J
o S!. -... MgO
o 'Bl - .....
200 220 240 260 280 300 320 340 360
TEMPERATURE, C
Figure 2. Effect of the support on olefin selectivity.
n
»
-t
»
r
-<
~
CJ)
C
"'tI
Table 4. Reducibilities of the Supports "'tI
o
~
Z
Metal Oxide Reduction to Reducibility [8] at 1000 0 K -t
SMSI Activity [Ref] ::z::
log (PH20/PH2) m
"T1
c:;;
n
::z::
Th02 Th -18 m
::0
Ce02 Ce -17 .6 .!.oj
MgO Mg -14.5 no [7] ::0
o"'tI
La203 La -13.7 CJ)
Zr02 Zr -13.6 no [7] n
Si ::z::
S102 - 9.1 no [7] CJ)
MnO Mn - 6.2 very strong [7] -<
Z
Cr203 Cr - 5 little [7] -t
ZnO Zn ::z::
- 2.9 m
CJ)
Nb205 Nb02 - 1.1 very strong [7] c:;;
Ti02 Ti305 - 1 strong [7]
Sn02 Sn + 0.2
00
(1l
186 J. G. GOODWIN, Jr. ET AL.
tend to confirm the possibility of SMSI with Ru/ZnO since the activ-
ity and CH4 selectivity were lower and the olefin selectivity greater
for Ru/ZnO relative to Ru/SiOZ. Such behavior is similar to that
found for SMSI Ru/TiOZ [9]. Our results also confirm those of
Tauster et al. [7] that, although CrZ03 would be a good candidate
for SMSI, it does not seem to occur. Ru/CrZ03 gave results in this
study very similar to those for Ru/SiOZ. Finally, although SnO is
the most reducible of all the oxides listed, it produced primarily
CH4. This may be explained by the fact that, if an oxide is reduced
too much, alloy formation may occur between the reduced metal of the
support and the catalytic metal. In the case of Ru/SnO, since Sn is
not active for F-T, formation of RuSn particles would be expected to
decrease the activity and increase CH4 formation. This is indeed
what our results show.
Recently, Chen and White [10] have suggested that SMSI behavior
may occur with those supports that have relatively high electrical
conductivities and work functions that are lower than the supported
metal. Comparing the conductivities of the inorganic oxide supports
in this study [11], it can be seen that CrZ03 and ZnO have the high-
est conductivities. However, as pointed out before, ZnO and Crz03
show significant differences in activities and selectivities of meth-
ane and olefin formation. A possible explanation for this difference
may lie in the fact that ZnO is an n type semiconductor and Crz03 is
a p type semiconductor [lZ]. However, it should be noted that Laz03
and Crz03 show significantly different properties although both are
p-type semiconductors. In addition, while Cr 203 has a higher con-
ductivity than SiOZ (SiO Z is an insulator), toe catalytic selectiv-
ities for Ru/cr Z0 3 and Rfi/Si0 2 are very similar.
CONCLUSIONS
The most interesting supports studied were ZnO and La203 since
188 J. G. GOODWIN, Jr. ET AL.
they were able to produce low CH4 yields and very high (> 80%) olefin
yields. Ru/La203 was one of the most active catalysts (on a site
basis), whereas Ru/ZnO was one of the least active. The results sug-
gest that ZnO may be an SMSI support.
While ZnO and La203 are not high surface area supports as is
normally desirable, the unique catalytic properties for F-T synthesis
of Ru in interaction with thern suggest interesting possibilities for
new catalyst designs.
ACKNOWLEDGEMENTS
REFERENCES
CARBOMETHOXYLATION OF OLEFINS
F. J. Waller
INTRODUCTION
~ RCH2CH2C02Rl
RCH=CH 2 + CO + R10H (3)
~
1S3
194 F. J. WALLER
(4)
EXPERIMENTAL
Materials
Carbomethoxylation Procedure
Catalyst Turnover
Acid (mequiv) Conv.
precursor a rate
a5% Pd/C (0.434 mmo1 Pd), ~3P/Pd = 15.4, 500 mmo1 MeOH, 333 mmo1
propene: (~3P)4Pd (0.433 mmo1 Pd), ~3P/(~3P)4Pd = 11.5, 500 mmo1
MeOH, 333 mmo1 propene.
(CF2CF2)n-CF2CF--
I
(OCF2iF----)mOCF2CF2S03H
CF3
x
Catalyst Turnover
Acid b (mequiv) Conv.
precursor a rate
a(03P)4Pd (0.433 mmol Pd), 03P/(03P)4Pd = 11.5, 500 mmol MeOH, 333
mmol propene.
bAmberlyst 15 is a macroreticular sulfonated polystyrene ion-exchange
resin, and Dowex is a sulfonated polystyrene ion-exchange resin.
are 44.1 and 2.4%, respectively. The origin of the latter compound
appears to be ethene insertion into the intermediate (L)nPdXCOCH2CH3.
1-Hexene yielded three products, methyl heptanoate, methyl 2-methy1-
hexanoate, and methyl 2-ethy1pentanoate in conversions of 13.7, 4.4,
and 0.7%, respectively (Equation 6). Cis-2-butene produced two pro-
ducts, methyl 2-methy1butyrate and methyl valerate in conversions
of 8.9 and 13.5%.
200 F. J. WALLER
Turnover
Olefin acid a (mequiv) Conv.
rate
f H3 f2 H5
CH3(CH2)3CHC02CH3 + CH3(CH2)2CHC02CH3 (6)
SUMMARY
Turnover
Olefin Acid a (mequiv) Conv.
rate
a(03P)4Pd (0.433 mmol Pd), 03P/(03P)4Pd 11.5, 500 mmol MeOH 333
mmol olefin.
REFERENCES
HYDROFORMYLATION OF DICYCLOPENTADIENE
INTRODUCTION
aWhile this work was in progress, a patent was issued on the hydro-
formylation of DCPD and similar materials using a Co2(CO)8/PR3
catalyst system [3].
203
204 C. D. WOOD AND P. E. GARROU
00
J eO/R2
y No ~2
HNO- ~O
eo~~~
H!~
H~OH "DCPDDH"
scheme for the production of diol. In the first step, the norbornyl
ring of the DCPD is hydroformylated to the aldehyde product. This
intermediate can either be reduced to the mono-alcohol or further
hydroformylated to the di-aldehyde intermediate. As the reaction
continues, both of these products are converted to diol.
! eO/H 2
~O~ (}Q,OB
Figure 2. Olefin Hydrogenation Scheme
HYDROFORMYLATION OF DICYCLOPENT ADIENE 205
[oJ
Figure 3. Cyclopentylmethanol (CPM) Formation Scheme
Dependence on Phosphine
~ OH
HCcf~
II
o
~ -CH ~ OCH
~II H6~1Io
o
~ OCR
Hc6~11
HCrOD
II
o
8 0
co + HCo(CO)3L t HCo(CO)4 + L.
The equilibrium was shown to lie further to the left for more basic
and less bulky phosphines. Since the hydroformy1ation reaction is
much faster with HCo(CO)4 than with its phosphine complexes, faster
reaction rates should be obtained with triary1phosphines than with
the more basic tria1ky1phosphines. Intimately tied to the phosphine
effect on rate is its effect on selectivity, or more precisely the
hydrogenation activity, of the catalyst. The rate of reduction of
aldehyde has been shown to increase with increasing basicity of PR3
because compounds of the type HCo(CO)3PR3 are better reducing agents
than HCo(CO)4 due to their increased hydridic character [5-7]. Using
C02(CO)S as a catalyst in the 150-1S0°C temperature range, synthesis
pressures of 1500-4500 psig are necessary to prevent catalyst decom-
position. However, when C02(CO)S has been modified by triorganophos-
phines, pressures as low as 100-300 psig and temperatures as high as
150-200°C may be used without any observable catalyst decomposition.
Catalyst stability has also been shown to increase with the basicity
of the organophosphorous ligand [3].
The PPh 3 modified system revealed the next fastest rate, good
stability, and an equivalent DCPDDM yield. In the C02(CO)S-PPh3
system, DCPD reacts with synthesis gas at 160°C. As the reaction
temperature increased to 195°C, both the dia1dehyde and mono-alcohol
intermediates were formed. Hydroformy1ation of the 5-membered ring
proceeds at the same rate as the reduction of the mono-aldehyde in-
termediate. These products are then converted to DCPDDM.
Cracking Products
Test
Number
Phosphine Condo pKa ova (Y'O? 0
1 a
2 c
3 PPP a b 2
4 PPh3 a 2.7 7
5 PPh3 c 2.7 4
6 PPh3 f 2.7 3
7 PPh3 g 2.7 1
8 P(n-Buh c 8.4 10 3
9 PMe3 a 8.6 20 5
10 PMe3 c 8.6 9 6
n P(C6Hnh a 9.7 3
12 P(C6Hn)3 c 9.7 7 1
13 P(C6 Hn)3 d 9.7 9
14 P(C6Hn)3 e 9.7 4 4
a4.0 g DCPD, 0.75 wt% Co, 195-200 o C, 2000 psig, H2/CO = 2, 3-4 hr
reaction time, P/Co = 1, 20 g solvent, 100% DCPD conversion
bUnknown although l-methylphosphole has pKa of +0.5
c5X amounts in footnote a
dSame as footnote a, except 500 psig H2/CO
eSame as footnote a, except 3500 psig CO/H2
fSame as footnote a, except 1.75 wt% Co
gSame as footnote c, except slow DCPD feed at operating conditions
(continued)
Table 1- (Continued)
H'\{}() HC~
II
0
+ +
Test
Number 00 ~OH ~OCH
II
0
DCPDDM
DCPDDM
Formates
1 6 1 65 9
2 5 1 59 13
3 8 62 7
4 15 63 7
5 9 64 4
6 13 66 5
7 11 65 6
8 1 17 8 34 20
9 1 13 3 38 15
10 2 12 6 38 22
11 1 19 1 60 7
12 1 23 1 55 7
13 21 22 18 1
14 18 1 55 8
wr.%
20
o 2
TIME(HR)
Figure 5. Reaction profile for the hydroformylation of DCPD with
C02(CO)8
low, a gray, insoluble powder was isolated from the reaction mixtures.
Elemental analysis indicated that this product contained >95% cobalt.
Formates
60
H~OH
40
WT,%
20
H~
I
o 2
TIME(HR)
WT.%
o 2
TIME(HR)
Figure 7. Reaction profile for the hydroformylation of DCPD with
C02 (CO)8-(n-Bu)3P
81 24 105
84 15 99
87 14 101
83 21 104
92 11 103
P(C6H11)3 90 7 97
n-Bu3P 93 2 95
PPP 19 20 99
78 25 103
1100 820 75
1400 1200 86
1100 1100 100
2500 2500 100
Reaction Poisons
CATALYST CHARACTERIZATION
The catalysts were generally prepared in situ by the addition
of 2 equivalents of PR3 to C02(CO)8. The PR3 ligand readily reacts
HYDROFORMYLATION OF DICYCLOPENT ADIENE 213
3500 4 18 8 60
2000 3 19 8 60
500 9 43 <1 l8 a
(8)
.I I I I I I I I
2100 2000 1900 1800 2100 ·2000 1900 1100
Z079 (m)
ZOZO (m)
1987 (s)
1963 (m)
1943 (s)
[Co(CO)3[P(C6Hll)3]Z]+ 75.9 1995 (m)
[Co(CO)4]- 1980 (m)
1890 (s)
ZO.l Z038 (m)
1953 (s)
o CO2(CO)7 PR3
A [Co(CO)3PRJ2
• [Co(CO)3(PR3)2 + J
o <> HCo(CO)3PR3
82 76 70 PPM
I I I
2100 2000 1900 1800 (em-I)
(PROTON DECOUPlED)
C02(CO)7PR3
[Co(CO)3PRJ2
(SElECT! VE DECOUPlED)
PPM
CONCLUSIONS
EXPERIMENTAL
DCPD Hydroformylation
Catalyst Preparations
REFERENCES
INTRODUCTION
synthesis gas since all of the feed appears in the product. There
is no loss of feed components to form water in this reaction, such
as occurs in the production of ethanol, olefins, or aromatics from
synthesis gas.
221
222 S. SUZUKI ET AL.
Thermodynamics
The thermodynamic data for glycolic acid formation are both in-
complete and unreliable. However, the magnitudes of enthalpy and
entropy changes are large enough to provide useful information. The
reaction, as written in equation 2, is exothermic with ~Ho298 = -27
± 3 kcal/mol.
(Glycolic)
K=------- (4)
I8
I
!
I BallI
~H" =-27 keel
.MS0 =-59 CaI.I"C
Ideal Solutlonl
Feed: One Mole H20
Per Mole CH20
Temperature, °C
The high pressure used in the Du Pont glycolic acid process ob-
viously requires use of unconventional equipment and probably causes
high energy use in the process. In addition, sulfuric acid at 220°C
is highly corrosive, and conventional corrosion-resistant materials
of construction cannot be used at these conditions.
Metal-Catalyzed Carbony1ations
-----<.~HOCH2+ + H20
2
HOCH 2+ + CO -----<.~HOCH2CO+
2 3
-----<.~HOCH2COOH + H+
4
- - - - - - t... HOCH2C02CH2COOH + H+
3 4 5
(6)
We have found [2Z] that alkoxyacetic acids and their esters are
readily formed by the carbonylation of formaldehyde or methylal with
HF catalyst in the presence of alcohols. This reaction is so favor-
able that any methanol present in the formaldehyde used for the pro-
duction of glycolic acid is converted almost quantitatively into
methoxyacetic acid. Diglycolic acid can be hydrogenated to diethyl-
ene glycol, but methoxyacetic acid is an undesirable by-product, as
it cannot readily be converted to a glycol. To minimize the pro-
duction of methoxyacetic acid, formaldehyde with a low methanol con-
~ent is required for commercial production of glycolic acid and
ethylene glycol by this process.
~
n
or
Table 1. Dig1yco1ic Acid Formation in Formaldehyde Carbony1ation "T1
:D
o
s::
products, mol %
run conditions s::m
starting of CH20 charged -I
material reaction :::I:
test mole ratio, time, glycolic dig1ycolic unreacted >
Z
temp., °c pressure, psig
no. CH20/H20/HF min acid acid CH20, % or
>
Z
o
1 1/0.5/2.5. 50 2070-1050 80 91 5 Ul
2 1/0.5/2.5 1
-<
Z
53 1980 50 77 6
-I
3 1/0.5/2.5 70 2000 35 90 10 -1 :::I:
m
4 1/0.2/2.5 50 1900 35 67 23 Ul
5 1/0.0/2.5 25 1940 45 34 45 >5 enG)
6 1/0.0/2.5 50 1920 120 28 51 8 >
7 1/0.0/1.5 50 1950 155 24 55 >3 Ul
I\)
w
232 S. SUZUKI ET AL.
100
50
20
19
20 40 60 80 100 120
Reaction Time, Min.
8 1/1.8/16.7 84 22 970-830 11 50 83
9 1/1.8/16.7 84 22 500-355 9 60 87
10 1/1.4/12.3 82 23 1000-655 10 47 97
11 1/1/10 81 0 50 147 a 73 27 b
12 1/1/10 81 50 80 80 122a 60 b 0.0
13 1/1/8.3 78 23 960-700 60 150 93
14 1/1/5 68 37 1920-1600 30 50 105 b
15 1/1/5 68 20 50 207 52 44 b
16 1/1/5 68 50 80 60 46 37 b
17 1/1/3.6 60 50 1500 38 212 91 1.0 1
18 1/1/3.6 60 50 1660-1500 25 40 95 3.5
19 1/1/3.6 60 50 340 110 104b
20 1/1/3.6 60 30 355 110 67 21b
21 1/0.5/2.5 56 50 2490-2135 35 88 2.7
~
CJ)
- ---------------------- - - - - c
aSma11 reactor with limited stirring. N
c
bAna1ysis approximate in these experiments. ~
m
-I
»
:-
ETHYLENE GLYCOL FROM METHANOL AND SYNTHESIS GAS 235
10
0
:::E:
I
C
0
~
c 6
~
I&.
~
;:;
~
Carbonylation Techniques
properties reaction
catalyst components
pelleted catalysts a conditions c
(8)
(9)
TEST NO.
feed, wt % DEGGa 59 61 39 30 28
ester composition, %
monomer 59 66 78 89 74
dimer 33 31 21 11 22
trimer 8 3 0.7 none 4
test conditions b
temp., °c 200 182 182 204 182
pressure, psig 1500 1500 1500 1500 1500
LHSV 4.8 4.0 4.8 6.4 3.2
conversion, %
initial 96.3 96.3 93.5 92.5 95.7
final 81.2 93.6 93.1 92.5 93.0
loss of activity
(% conversion)/100 hr 15 14 0.37 none 12
less than 22% of the glycolic acid in the form of polymeric esters
in order to avoid catalyst deactivation.
•
1990C 1500 I ~204°C, 1000 pslg
, psg .... ~
93
. '. ,r~~~----~.r--w".
, I
,, ,,,
I 204°C'
199°C I
;I. I
,,I
,
I
I
,
J , I
I
I
I , I
,
I
I
R••ctlon PNUUN, 1100 pslg
Ttlt Period, Hr
Figure 4. Effects of polymeric glycolate esters on catalyst
stability
(10)
ETHYLENE GlyeOl FROM METHANOL AND SYNTHESIS GAS 245
CONCLUSION
REFERENCES
Alan Peltzman
Halcon SD Technology
Two Park Avenue
New York, New York 10016
INTRODUCTION
249
250 A. PELTZMAN
The major commerCl.al uses for formic acid are in the textile
and leather industries. In Europe, formic acid has been in use
as a silage preservative for many years. The prospect for greatly
increased worldwide formic acid demand for agricultural purposes
has been raised in recent years.
the world's food supply. Those regions of the world having short
growing seasons and hence long periods when livestock are fed
stored forage would exhibit the greatest desire for direct acidi-
fication. Typical areas which would become markets for formic
acid for these purposes are Canada, parts of the u.s. and Northern
Europe.
Process Description
SYNTHESIS GAS ~l
COMPRESSION
REACTION
Ln
METHYL FORMATE HYDROLYSIS
HYDROLYSIS
PRODUCT PRODUCT
PURIFICATION
REACTION SEPARATION
METHYL FORMATE
METHANOL SEPARATION
CH OH + co + HCOOCH 3 (6)
(Met~anol) (Carbon (Methyl
Monoxide) Formate)
CO + + HCOOH (8)
(Carbon (Formic
Monoxide) Acid)
256 A. PELTZMAN
Hydrolysis/Purification
ECONOMICS
Labor
Operators per shift 2 2
Foremen per shift 1 1
Supervisor 1 1
Maintenance
% of BLCC 4 4
SYNTHESIS GAS TO FORMIC ACID 259
CONCLUDING REMARKS
REFERENCES
INTRODUCTION
261
262 W. R. PRETZER AND M. M. HABIB
A Caveat
EXPERIMENTAL
Product Analysis
The gas analysis was performed using a Carle series 500 CGC with
dual detectors. The oxygenates were analyzed by a flame detector
while H2' CO, C02' 02, N2, and light hydrocarbons were analyzed by a
thermal conductivity detector. The whole analysis was performed by
either a single manual or in-line injection using a set of seven dif-
ferent columns. Valve switching as well as temperature programming
were controlled by a microprocessor.
Data for initial gas uptake studies for different catalyst sys-
tems were collected in the unit described above. However, for better
accuracy in measuring the gas uptake, a small reservoir with a volume
of 120 cc was used. Sensotek pressure transducers fitted with dig-
ital read-outs were connected to both the reactor and the reservoir.
The temperature of the reservoir was held constant during an exper-
iment, while the temperature of the autoclave varied ±1.5°C from the
desired reaction temperature. Time and gas uptake readings were
266 W. R. PRETZER AND M. M. HABIB
recorded from the moment the unit reached desired reaction condi-
tions.
Methanol Homologation
Promotional Effects
Z
I/Co mmol products based on methanol mmol :x>
Methanol % HAc & Deriv. c r
Molar ()
Conversion (based on HAc) o
Ratio Me20 HAc EtOH MeOAc CH4 Others b
:J:
or
:J:
0.0 12 58 61 20 32 49 4 61 o
0.5 36 69 125 105 94 59 417 221 s::o
1.0 47 77 216 108 142 68 464 284 r
110 200 528 713
oc;)
2.0 69 523 6 75
4.5 64 305 375 2 177 81 408 663 :x>
-I
10.0 68 834 130 0 124 121 466 314 6
z
i'J
en
....,
268 w. R. PRETZER AND M. M. HABIB
eooo
Legend
4600 0 1- 0
/j.
~c~o~ .
4000 0 It..c°- t ...-'1
'V ~ClE.2_ •
8BOO 0 I!..CO=4.5 __ 'V""
0
0 It..Co=tO
ci
~ .000 --
u;
01( 2600
C!J
..J
~ 2000
1100
1000
BOO
0
0 2 4 II
• 10 12
nME. MIN.
14 18 11 20
Selectivity Effects
Table 2. Effect of Iodine/Cobalt Ratio on Product Selectivity for the Catalyst Coba1t/lodinel
DPPEa
1000~------~--------------------------------------------------'
Legend
IlOO 0 I/Co-<J.5 :,,10.'11 Cobl1ll ~I
1~o.~C. 4000"$(11
[,. !fPo..=..2_ p/ Co. ' .H2/CO='
41100
ci
~ 1500
'iCI 1000
2000
1100
1000
1000
~
100 I!!!.-
_IP~.·•...
w_ ........ .
«~U.Il~.... D "d(m."M:Z:~8t,\un. + 'ft cUdols
lOa n-HAc units
' .
700
.•...
b.
•
u 000
..." ..,....•......, ..
£ eoo
0
E
400
..~.............
E ......
saO ·····························.. 0
200
100
'f---..--t;>----..--T"I-6---.--~------=--
e ~
6
4 15
IICo Moler Retio
10
"
Figure 3. Effect of iodinelcobalt ratio on product selectivities
Previously, it had been reported that the rico ratio can dramat-
ically influence the yield (a function of conversion and selectivity)
of products obtained by homologating methanol [19]. In this earlier
work, two catalyst systems were studied. One system, producing eth-
anol, comprised Co(acac)2' tri-p-tolylphosphite, and iodine, showed
low ethanol yields when IICo > 0.5. The other system, giving acetal-
dehyde, was composed of Co(acac)2, Ph3As, and iodine. Significantly
lower acetaldehyde yields were seen when rico> 0.25. It is impor-
tant to note that the reaction time for these experiments was 3 hr
compared to 20 min for the work described above.
The observation that the yield of ethanol declines when rico >
0.5 is consistent with the selectivity pattern as a function of rico.
Hydrogenation ability is suppressed when rico > 2 and, while signif-
icant amounts of ethanol are produced when rico = 1, more acetalde-
hyde than ethanol is formed. With an extended reaction time, this
acetaldehyde is subject to condensation, a process more favorable
under the more acidic conditions prevailing when rico = 1 than at
rico = 0.5. To produce high yields of ethanol in minimal reaction
time, a catalyst capable of rapidly hydrogenating acetaldehyde is
required. Since cobalt cannot do this under conditions most favor-
able for rapid aldehyde synthesis, a cocatalyst is required. The so-
lution, as amply represented in the patent literature [12,20a-f], has
been the use of ruthenium compounds. Catalyst systems containing
ruthenium cocatalysts produce optimal ethanol yields when rico = 2,
which is the ratio giving the maximum rate of acetaldehyde production
using catalysts containing only cobalt.
Aldehydes Carbonylation
Alcohol + 5% Water Alcohols Hydrocarbons Ethers b Others
Hydrocarbons
aCatalyst = CoI2 = 1 mmol; NaI 6.5 nnnol; dppe 0.5 mmol; Alcohols 95 cc s:
Pressure = 4000 psig, H2 /CO = 1 s:
Temperature = 200°C I
l>
Time = 3 hr !:!:!
OJ
RECENT ADVANCES IN ALCOHOL HOMOLOGATION 277
tion reaction proceeds via a carbonium ion [3], and that an alcohol
dehydration mechanism is operative [8], at least for those alcohols
which can be dehydrated to olefins. The observed reactivity pattern
(3°>2°>1°) is consistent with an El elimination involving a proton-
ated alcohol giving a carbonium ion. Subsequent reaction of this
carbonium ion could involve nucleophilic attack by iodide or Co(CO)4,
or elimination of a proton to give an olefin. Hydroformylation of
olefin leads to product aldehyde.
Our results (vide supra) differ from this prior work. Using a
catalyst comprising CoI2' NaI, and dppe, the present work has ob-
served the order of reactivity to follow methanol > primary > second-
ary > tertiary. Further, much less rearrangement was observed, i.e.
the position of the hydroxy group in the starting alcohol was pre-
dominantly the position of addition of the -CH2- unit. These results
are consistent with an SN2 mechanism, most likely involving the for-
mation of alkyl iodide from alcohol and iodide.
In the presence of the CoI2, NaI, and dppe catalyst system, al-
cohol should still be protonated and olefin formed as described
above. However, olefin will accumulate since no effective hydrofor-
mylation catalyst is present. HCo{CO)4 is not present in any signif-
icant amount and cobalt carbonyl modified with aromatic diphosphine
is a drastically less active hydroformylation catalyst [22]. This
pathway to homologated product is effectively blocked. A second
pathway, however, is available, i.e. nucleophilic attack by 1- on the
protonated alcohol to give alkyl iodide. The mechanism for this
process will be SN2 and thus favors primary alcohols. Oxidative ad-
dition of alkyl iodide to a cobalt compound will give an alkyl-cobalt
complex and subsequent alkyl migration (or carbon monoxide insertion)
affords an acyl complex. Hydrogenation andlor reductive elimination
will give product aldehyde. This second pathway is not viable for
primary alcohols in iodine-free or -deficient systems. In the ab-
sence of iodine no strong nucleophile is present, while in those sys-
tems with low IICo ratios, the concentration of free 1- is low. Fur-
ther, for those systems promoted with 12, the iodide is probably
present as HI, a considerably weaker nucleophile than 1-.
II-)
.....
co
280 w. R. PRETZER AND M. M. HABIB
mary alcohols will not react since the SNZ path is minimized and
elimination is not favorable for primary alcohols.
-HZO -w
R+ - - - -__~ olefin co + HZ _ RCHZOH \
El pathway HCo(CO)4
ROH + ~ ~ [ROHZ]+
+1- RI CO + HZ RCH OH
----- Z
-HZO [Co]
ACKNOWLEDGEMENTS
REFERENCES
M. L. Deem
Department of Chemistry
Lehigh University
Bethlehem, Pennsylvania 18015
oII 0
C III
[.0\ ~ \ }; /0:\
O-Co-C=Co-O
\.0 'oj
1
RhIX(alkene)LZ
ll~s alkene,
RhIXL3 ~ RhIX(S)LZ +H~ HZRhIIIX(S)LZ (1)
Solvate Reference
RhC12(COC2HS) (PPh3)2·C2HSOH 2
RhCl(PPh3)2(dioxane) 3
RhC12(COC2HS) (PPh3)2·0.SCH2C12 2
[H2RhCl(PPh3)2]2·CH2C12 4
[H2RhBr(PPh3)2]2·CH2C12 4
[H2RhCl(PPh3) (CSHSN)]2 4
[H2Rh(PPh2cy) 2 (Me2CO)2]PF6a S
[H2Rh(PPh2cy) 2 (MeCN)2]PF6 S
a cy = cyclohexyl.
aFrom Reference 2.
bDownfie1d to 8S% H3P04 is positive.
-55 100/0
-30 75/25
-15 66/33
o 45/55
+15 40/60
+30 20/80
I
R,R'CHOH + X=Z ~ R,R'C=O + HX-YH
"donor" "acceptor"
2 3 4 5
H H 80 39
H H 80 39
H H CHCH:CH 2 80 39
H H CHCH:CHCH 3 80 39
H CH 3 CHCCH 3 :CH 2 80 39
CH 3 CHPh 83 29 40
CH 3 80, 83 39, 40
CH 3 norbornene 42
CH 3 norbornadiene 83 40
CH 3 1,5-cyclooctadiene 83 40, 42
CHiH3 ~H(CH2)5~H 80 42
CH 3 "
CH(CH «
2 \CH 100 38 42, 43
CH 3 ~H(CH2\~H 42
CH 2CH 3 3H(CH2)5~H 80 42
Ph ~H(CH2) 4~CH3 43
(continued)
296 M. L. DEEM
Table 50 (continued)
Alcoho17Donor! ~ Acceptor! ..:L Temper- Yield (%) Refer-
R R' X Y ature 1- ence
CH 3 CH 3 0 C(CH Z)4 45
CH 3 CH 3 0 CCH3,(CH2)ZCCH3:CHZ 83 Z005 c , Z8 b 44
CH 3 CH 3 gH(CO)(CH2)3~H 40 15 43
H Ph ~H(CO)(CHZ)3~H 40 90 43
CH 3 Ph ~H(CO)(CH2)3~CH3 70 43
CH 3 Ph HZC CHCOCH 3 40 43
CH 3 Ph CH 3CH CHCOCH 3 50 43
II II
CH 3 Ph CH(CO)(CH 2)3CH 40, 50 46, -- 43
H CH 3 PhCH CHCOCH 3 50 Z7 43
CH 3 CH 3 PhCH CHCOCH 3 50 lZ 43
CH 3CH 2 Ph ~H(CO)(CHZ)3~H 40 Z9 43
Table 5. (continued)
H Ph PhCH CHCOCH 3 50 79 43
CH 3 Ph PhCH CHCOC(CH 3 )3 50 43
CH 3 Ph ArCH CHCOPh 50 43
Ph Ph PhCH CHCOPh 50 39 43
CH 3 CH 3 PhCHl CHPh 83 91 48
CH 3 CH 3 PhN CHPh 83 85 48
aFor ketone.
bFor alcohol(s).
cFor alkane.
s::
r
om
m
s::
POLYETHERS AND ORGANORHODIUMS 299
: CH 2- -----7'"-=------.... -CH2CHCH2CH2CH2-
• -CH2CH2CH2CH2- I III
RhlLu ~h Ln
H
10 11
;;>'"
CICOCH2CH3 HOCH 21H2cH3
.. -CH 2CH=CHCH 2 - (6)
I
RhIL n
+ RhIL n 12
13
Z oxygen of an ether
o (ppm) Found
Reaction Mixture Known n-PrOH
64.413 64.522
25.895
10.266 10.185
alkenes) toward the value of the completely free alkene. The addi-
tion of D20 breaks polyether-rhodium(III) interactions and causes
two other simultaneous changes in the lH NMR spectrum. All residual
cis-RhC12(COC2H5)(PPh3)2 is converted to its trans isomer following
introduction of D20, where this transformation is monitored by the
highly characteristic lH NMR signals for the COCH2CH3 group of the
cis and trans isomers (cf. Table 2). This may imply that association
of MeO(CH2CH20)10(CH2)15CH3 with some cis-RhC12(COC2H5) (PPh3) 2 inhib-
ited the cis to trans conversion, possibly by impeding a Berry pseu-
dorotation process. The -CH20- units of MeO(CH2CH20)10(CH2)15CH3
show two characteristic signals. The dominant one is at 3.64-3.61
ppm, and an environmentally sensitive signal is present in the region
of 3.72-3.66 ppm. This latter signal shows variations in its rela-
tive intensity following successive addition to MeO(CH2CH20)10(CH2)15-
CH3 of RhCI(PPh3)3 plus CICOC2H5 and, then, D20. Finally, after the
mix is left with D20 for about one day, this minor band disappears.
More dramatic interactive effects are exhibited [30] in the IH NMR
spectra for a polyether following its admixture with R4~, which is
a less diffuse cation than are Rh(I) and Rh(III). A tight 1:1 ad-
duct, where the solid is stable after thermolysis at >65°C and 3-4
Torr for >3.75 hr, of l,2-dimethoxyethane and n-octyltriethylammonium
chloride shows lH NMR resonances of 3.23, 3.10, and 3.02 ppm (of rel-
ative intensities 4:2:3), whereas the separate components have res-
onances of 3.46, 3.53, and 3.36 ppm for (CH3CB2)4NCl, CH30C]2C]20CH3'
and C]30CH2CH20C]3, respectively.
i i i I
phinopentane complex has given [55,61] RhHC1(t-Bu2PCH2CH2CMeCH2CH2P-
Bu-t2) (in -30% yield) and RhHC1(t-Bu2PCH2CH2CHCH2CH2PBu-t2) (in 75%
yield). An X-ray crystallographic study [61] firmly establishes
these structures.
EXPERIMENTAL
to lower field). For 3lp NMR signals, 85% H3P04 was used as an ex-
ternal standard (0 = 0, with increasingly positive values to lower
field). For 3lp{lH}FT_, lH-, and l3C{lH}FT-NMR spectra, the probe
was set at 36.23, 89.55, and 22.49 MHz, respectively. Gas chromato-
graphy was done with a Hewlett-Packard 5730A chromatograph equipped
with a 5880A Series GC terminal. The single column measured 6' x 1/8"
(Ni) and was packed with 5% SE-30 on an 80/100 mesh support. Glass
injection port liners were loaded with silica gel (to prevent rhodium
from entering the column and injection port), and the injection port
and TC detector were maintained at 150 and 250°C, respectively.
Daily GC calibration curves were prepared using standard solutions
(equal to 13, 35, 55, and 78% reaction of ClCOC2H5)'
SUMMARY
ACKNOWLEDGEMENTS
1574 (1967).
14. W. Voe1ter and C. Djerassi, Chem. Ber., 101:58 (1968).
15. L. Horner, H. Blithe, and H. Siege1y Tetrahedron Lett., 4023
(1968).
16. w. Strohmeier and E. Hitze1, J. Organomet. Chem., 87:257 (1975).
17. J. P. Cand1in and A. R. Oldham, Discuss. Faraday Soc., (46):60
(1968).
18. A. Levi, G. Modena, and G. Scorrano, J. Chem. Soc., Chem.
Commun., 6 (1975).
19. W. S. Knowles, M. J. Sabacky, B. D. Vineyard, and D. J. Wein-
kauff, J. Am. Chem. Soc., 97:2567 (1975).
20. M. D. Fryzuk and P. Bosnich, J. Am. Chem. Soc., 99:6262 (1977).
21. D. G. Constable, C. R. Langrick, B. Shabanzadeh, and B. L. Shaw,
Inorg. Chim. Acta, 65:L151 (1982).
22. Y. Dror and J. Manas sen , J. Mol. Cata1., 2:219 (1977).
23. A. F. Borowski, D. J. Cole-Hamilton, and G. Wilkinson, Nouv. J.
Chim., 2:137 (1978).
24. F. Wuesthoff, E. F. von Pechmann, D. Behrens, and R. Goetz,
German Patent 2,627,354 (Dec.' 23, 1976).
25. R. Bar, Y. Sasson, and J. Blum, J. Mol. Cata1., 16:175 (1982).
26. R. G. Nuzzo, D. Feit1er, and G. M. Whitesides, J. Am. Chem. Soc.,
101:3683 (1979).
27. R. G. Nuzzo, S. L. Haynie, M. E. Wilson, and G. M. Whitesides,
J. Org. Chem., 46:2861 (1981).
28. Y. Dror and J. Manassen, Proc. 7th Int~rn. Congr. Cata1., Tokyo,
July 1-2, 1980, Abstr. B14.
29. S. Franks, F. R. Hartley, and J. R. Chipperfie1d, Inorg. Chem.,
20:3238 (1981).
30. M. L. Deem, unpublished experimental results/submitted for pub-
lication.
31. Anal. Ca1cd for C37H76011'3H20: C, 59.17; H, 11.01. Found: C,
59.43; H, 11.18.
32. Anal. Cal cd for C48H83010P,: C, 67.73; H, 9.83; P, 3.64. Found:
C, 67.91; H, 9.83; P, 4.03.
33. J. P. Collman, Acc. Chem. Res., 1:136 (1968).
34. F. H. Jardine, Prog.Inorg. Chem., 28:144 (1981).
35. J. Chatt and B. L. Shaw,Chem. Ind. (London), 931 (1960).
36. M. A. Bennett and P. A. Longstaff, Chem. Ind. (London), 846
(1965).
37. A. J. Birch and K. A. M. Walker, J. Chem. Soc., C, 1894 (1966).
38. A. Emery, A. C. Oeh1sch1ager, and A. M. Unrau, Tetrahedron Lett.,
4401 (1970).
39. A. Yamada, T. Fukuda, and M. Yamagita, Bull. Chem. Soc. Jpn.,
48:353 (1975).
40. R. Uson, L. A. Oro, R. Sariego, and M. A. Esterve1as, J. Organ-
omet. Chem., 214:399 (1981).
41. T. Nishiguchi, K. Tachi, and K. Fukuzumi, J. Am. Chem. Soc.,
94:8916 (1972).
42. H. Imai, T. Nishiguchi, and K. Fukuzumi, J. Org. Chem., 39:1622
(1974).
POLYETHERS AND ORGANORHODIUMS 305
Jack D. Chase
INTRODUCTION
307
308 J. D. CHASE
when blended into gasoline at the 10% level, increase research and
motor octane by 2-3 octane numbers.
ETHERIFICATION REACTIONS
Chemistry
Catalyst
CH3-C-CH3 + CH30H • • (1)
II
CH2
Catalyst CH3
I
CH3-CH2=C-CH3 + CH30H • • CH3-CH2-C-O-CH3 (2)
I I
CH3 CH3
The above reactions occur below 180°C in the liquid phase and
are generally carried out at about the minimum pressure to assure
liquid phase at reaction temperature. There are three isomeric iso-
amylenes: only the 2-methyl butenes react, i.e. 2-methyl-l-butene and
2-methyl-2-butene, while the 3-methyl-l-butene does not react.
The addition reactions with alcohol are very selective and only occur
with the tertiary olefins. Consequently, when the C4 fraction of an
olefin stream from an industrial installation such as an ethylene
plant is passed through an esterification reactor, the other olefins
pass through the process unreacted. Incidentally, this situation has
lead to the proposed scheme [8] for separating isobutylene from other
C4s, whose boiling points are so close as to make distillation sep-
aration difficult. In such a case, the isobutylene is first ester-
ified, then separated from other C4s, and finally decomposed back to
isobutylene and methanol.
Kinetics
first order in isobutylene and zero order in methanol when more than
stoichiometric amounts of methanol are present. This was similarly
reported by work at Englehard [10]. This was confirmed by Gulf
Canada studies where results are plotted in Figure 1. The etheri-
fication of isoamylenes have similar type kinetics with respect to
the orders of the various reactants as reported by Snamprogetti [9].
1.6
w
~ 1.2
a:
z
Q 0 ••
I-
CJ
C 0.4
w
a:
o 1 2 3
REACTANTS MOLAR RATIO
(CH 3 0H/ iC4)
MTBE PROCESS
Gulf Canada was first to develop and promote [12] process tech-
nology for integrating a MTBE plant into a petroleum refinery using
the less concentrated (10 - 18%) ·isobutylene in the C4 stream from
a refinery cat cracker unit (FCCU).. A major aspect of this technol-
Conversion Temperature
Space Ve10cityb of Reaction Rate C
% °c
Most of the unreacted methanol comes off the bottom of the dis-
tillation column leaving from 1 to 5% methanol with the MTBE product,
depending on process conversions. This product stream typically also
contains up to 1% diisobuty1ene (a good gasoline component) and
about 0.2% C4 hydrocarbons.
aBPSD = barrels per stream day. This is distinct from barrels per
(calendar) day, since there is usually a 3-week maintenance shut
down annually.
312 J. D. CHASE
1&.1
III 25
I-
:E
e 20
Z 15
0
iii 0 0 ~
II:
11.1
10
>
Z
0 5
(J
~ 0
0 20 40 60 80
MODIFIED REYNOLDS NUMBER
TAME PROCESS
General
~ 15
in
ffi 10
>
Z
0 5
(J
~
0
160 180 200 220
TEMPERATURE. of
Figure 3. Effect of temperature and catalyst particle size (0, 35-48
Mesh; 0, +16 Mesh) on the % conversion to MTBE.
HIGH OCTANE ETHERS FROM METHANOL AND ISO-OLEFINS 313
HTBE PROCESS
STRIPPING COLUMN
REACTOR Dl ST. COLUHN UNREACTABLE Cq' S TO
ALKYLATION OR POL YHER I ZA TI ON
Cq OLEFINS
HEOH RECYCLE
(TO GASO)
Process Description
Table 2. MTBE Economics with the Gulf Canada Process; CAT Cracked
Feedstock
The location of the bleed stream to purge off the lighter ends
is at the reactor exit. As shown in Figure 5, a single distillation
column can serve as initial feed depentanizer and as an ether column
for separating TAME and C5 hydrocarbons. This doubling-up of service
provides both operating and capital cost savings, in comparison with
a TAME process with depentanizer and ether column upstre~ and down-
stream of the reactor, respectively. However, this economy may be
bought at the cost of loss of some blending flexibility, since the
ether product stream is more concentrated (greater than 50% TAME) in
the two column case, compared to being less than 5% TAME (remainder
being C6+ hydrocarbons) in the single column case.
TAI1E PROCESS
LCCG
on" rather than an integrated process. The former simply are added
to the olefinic gasoline stream in a refining complex. By slightly
modifying the upstream octane splitter column, which separates the
FCC gasolines into heavy cat cracked gasolines (HCCG) and light cat
cracked gasolines (LCCG), the economics of TAME processing are sig-
nificantly improved.
TAME ECONOMICS
PRODUCT PROPERTIES
w
(0
320 J.D.CHASE
ACKNOWLEDGEMENTS
Much of the work reported in this paper was carried out between
1974 and 1982 while the author was employed as a Senior Research
Engineer and Project Leader of Oxygasoline Process Development in
the Research and Development Department of Gulf Canada Limited. The
author, who is working as a consultant in early 1983, wishes to thank
Gulf Canada for permission to publish this work.
HIGH OCTANE ETHERS FROM METHANOL AND ISO-OLEFINS 321
a) Alkylation:
isobutylene + isobutane + trimethyl pentane 94 95/92
1 1.2 1. 78 volumes
b) Polymerization (Polygas):
isobutylene + n-butene + iso octene 89 98/80
1 1.0 1.62 volumes
c) MTBE
isobutylene + methanol + MTBE 110 118/101
1 0.43 1.27 volumes
REFERENCES
INTRODUCTION
323
324 L. LlU ET AL.
Cobb et al. [14] and Tipnis [15] reported on the use of Zeolon
500, a Norton product composed of erionite and chabazite, for meth-
anol conversion to low molecular weight olefins. The extent of
lanthanum ion exchange had little effect on the product distributions.
The catalyst activity changed with time on stream with paraffins
being the first product, olefins were present at intermediate times
and dimethyl ether was the major product toward the end of the reac-
tion period. The aged catalyst sample contained approximately 10%
coke. Cobb et al. [14] report substantially higher methane and C4
concentrations than reported by others as shown in Table 1. This
may be due to the existence of significant exotherms in the tubular
reactor that was used.
{ 10.4} { 9.6}
Butene 8 10 10.3 7.4
C! 2.0 9 1.9 2 9 8.2 4.6
CAl
N
UI
Table lb. Product Distributions with Small Pore Zeolites w
C)
'"
ZSM-34 Ion exchanged
Catalyst ZSM-34 a ZK5 c Zeo1on 500 e
with Thb AW500 d
tribution was 36.9% DME, 40.3% H20 and 22.4% hydrocarbons. For the methanol/water feed the C
m
product distribution exclusive of the feed was 3.9% DME, 54.7% H20 and 41.4% HC. -I
gTubu1ar reactor. »
I
hBerty reactor (internal recycle reactor).
CONVERSION OF METHANOL TO LOW MOLECULAR OLEFINS 327
Marosi et a1. [21] claim the use of H-ZSM-5 for methanol and/or
dimethy1ether diluted with 50% (by wt) of steam for selectively pro-
ducing ethylene. Experiments with dimethy1ether diluted with water,
carbon dioxide, nitrogen or helium yielded a gas composed of 74%,
54%, 39%, and 37% (mo1e%) ethylene, respectively. Propylene content
was 20%, 27%, 33%, and 37% (mo1e%), respectively. The butenes, bu-
tanes, methane, and propane were less than 1%. Presumably, the re-
maining gas would be methanol or unreacted dimethy1ether. This ex-
ample indicates that the action of water is more than just a diluent.
For dimethy1ether the affect may be a simple mass action effect ac-
cording to the reaction
(1)
Wunder et al. [22] claim the use of the molecular sieve, l3X,
containing manganese and/or magnesium for conversion of methanol
and/or dimethylether to obtain high yields of ethylene. A hydrocar-
distribution (Ex. 2) of 46.9% by wt ethylene, 29.2% propylene, 12.3%
methane, 5.3% butene, 3.7% ethane, 0.5% propane, and 2.1% butane was
obtained for a methanol feed, a reaction temperature of 400°C, and
reaction pressure of 1 bar. The methanol conversion was 90.1%, but
the conversion to hydrocarbons was only 12.5% with 47.2% conversion
to dimethylether.
lICO lIHo
kca1/mo1 at 700 0 K
Kinetics
(17)
This expression was combined with the material balance for a plug
flow reactor, integrated and rearranged to yield
a
akCM8 (18)
k'LHSV
where
k' = kCM
PMw- ]-1
[-RTPL
LHSV liquid hourly space velocity
330 l. lIU ET Al.
(19)
for methanol conversion over AW500 ion exchanged with NH1; and rare
earth chlorides. Data obtained by using a Berty internal recycle
reactor were used to develop equation (19). The activity factor (l
had the functional form,
1
(l = -=-l-+--'-bt"";:/;"'(:--LH""""S-V"""")
By combining the material balance for the well-mixed reactor with
Equation (19) and the definition of (l and rearranging, the result is
k O p2
1 M , (20)
eM: X LHSV
(1 + bt/LHSV + ki PM)3
where
A~C (21)
C~D (23)
: CH 2 + ( CH3)2 0 ~ C3 H6 + H2 0 (29)
Mechanisms
0+ H e:+
Basic I Bronsted
O---H---CH 2---O---H---O
Site I I Acid Site
H OR
\'\#
+ ~~2 + CH3CH 20R' + HOR
.. .. ,
H - - - - ·CH20R
5. Ono [36) proposes that the active centers are Bronsted sites
and writes
- 0+
CH30H + HOZ -+ Z-Oo - CH3 + H20
0- 0+ -[H...)+
Z-O - CH3 + CH30H -+ Z-O CH3' : CH2-0H
EXPERIMENTAL
The catalyst, AW-500 purchased from Union Carbide, was ion ex-
changed by using a solution of ammonium and rare earth chlorides.
The procedure was similar to that reported by Singh [38] and is de-
scribed in detail by Garza [39]. The exception was that a heated
ion exchange column was used. This enabled preparation of 5 kg of
catalyst for use in the fluid bed unit.
The coked catalyst was regenerated at 823°K with air while con-
tinuously feeding and withdrawing catalyst.
(Air)
';; Cyclone
Re<Jctor . ~
Feedstock
Monometer
Preheote
Used Catalyst
for Regeration
'11 0
0
~30 0
0
CZ H4
0
~~
.G
'E20
VI
"
C.H,
is
c
.8... "
"---0 Q
C4He
~
,..e
810 V
~o
'0 CH4
::t
FE ~ 8 Cfu ;
50 min, and
aCondition of the Catalysis: The catalyst for Runs 323, 324, and 328
was partially regenerated. Catalyst residence time during regen-
eration was approximately 60 min. Coke content was approximately
3.9 g of coke per 100 g of coke-free catalyst after regeneration
and before a run. Run 329: The catalyst was not regenerated after
Run 328.
bSelectivity C2H4 and C3H6 =100 (2 moles C2H4 + 3 moles C3H6)/(moles
methanol reacted) and reflects methanol conversion to coke, CO and
C02.
cMaximum theoretical value is 43.5% (Values reported are on a coke-
free basis).
c
.2
~
i20
o
c
.8
5u 10
i
:I:
0L--4~30--------4~~--------~~~O~ OL--430~------4-5LO--------~~~~
TEMPERATURE "C TEMPERATURE °C
50
~
::I
.0
~40
i3
c
0
.0 312
8 30
..,e
!' Feed Composition, mol. %
HaD
-.-
.E2.0
II
c 0
WHSY
0._ ~OH :& EXP. NO.
Na',2
II 0 0.' III 71 507,108,_.310 N-',I,2,5
>- 0 0.' 4. 51 SI',III,IIS N-',I,'
~ C, 0.1 100 !l9 N-I
W 10 'J 0.8 eo eo 517.'18 N = 1,2
i
0
10 30 50 70 90 110 130 150 170
N
Accumulated Catalyst Residence Time (min.) SN" .-Z:1 8..
OL---~----~----~--~----~----~--~--~
~ 30 50 70 90 00 00 ~ ~
N
Accumulated Catalyst Residence Time (min.) SN" .-Z:1 8..
i
'E o
WHSV
0..
Feed Composition, mole %
~OH :. HaD =.:. •. t,2
5 40 DO.•
o 0..
H
41
TI
al
!01,!08.1C8,1fO N-',',2.!
511,111,111 N-',I,I
.8.
c I:> 0 .•
'\} 0..
o 0.4
100
50
100
DO
5tI
517,111
It-,
N • ',2
!2!,524,la• • • Ncl,I,I,2
~3O
l.
::a::
.5 20
ec
8.
e 10 - ......."'-- 25%
Q.
~
".
48+%
i
0
10 30 50 70 90 110 130 150 170
IOO'r---------------------------------------~
!
~c90
0.51
.t~ >Ie I'
+1: 323 V
!i 80 <>.306
eo
ij
0 31•
d:t;
!e Feed Composition, mol. %
j~ TWHSV
o D.. ~OH : . HaD !:;:. •'1,2
£.E 60 o 0.. ae TI "',101,_.510 N-',',.,I
1&1 o 0.. 41
100
al 111,.11.51....1, •••
.. ._,
I:> 0 .•
~ '\} 0.1 DO eo 517,118 It -',2
o 0.4 100 125.124.11',121 N-',I,I,2
i
50~--------------------------------------~
10 30 50 70 90 110 130 150 170
.ti 10
FBBd Composition, mole %
~
.G
"i:
D
TWHSV
0..
0.'
~OH
18
: . HrfJ
71
::;:e N'I,2
507,108.101,310 H-',',2,5
1ii 8 0
0 0.. 48 SI 511,112.115 "- •• 2.1
0 o.a
.
{> 100 !f9 N~I
c 'V 0.8 SO 50 117,118 N =1.2
11 <> 0.4 100 311.114,128,128' N-',I,',2
8 6
,..e
"0
J:
.E 4
•
c
~
:::J
GI 2
it
~
o~--~----~----~--~----~--~~
10 30 50 70 90 110 130
__ __
~
150
~
170
N
Accumulated CatalYlt Relidence Time (min.! eN"!.. 8..
10 ...
48% MIN
12%825%
10
c Feed Composition, mol. %
! WHSV
~OH ~~,:. ""',e
o
~~
0.8 :. H,.o
o 0 .• 28 75 !07,508,3OI,atO Na',1,2,1
8 o 0.. 41 81 III,SI2,all N -',.,3
C D. 0.9 100 319 Na,
'J 0.8 50 80 317,348 N =1,2
c o 0.4 100 321, 324, a2~,528 He ',1,',2
~0 6
e
u
'0
>-
J:
.E 4 /12%, 25and48% MINe
~7 305 I
-
306 309 312
to
C j38 0311 ~13
0
.c o~ ~K>
Ui 2 6319~ ..... ,
•:.e 100 % M and 50-50 MIW
~
0
10 30 50 70 90 110 130 150 170
cIO~----------------------------------------------~
~
~
la
o
c 0'" 48-52 M-N
.8 _ _-=-----~00313
~ 6
~
~
J:
.E 4 2 317 SO-50 M-W
• 6319 n 318_ Feed Composition, mole %
c WHOY ....OH ". . . . "P. NO.
~ 0 0.8 12 81 508,301
; 2 0 0.8 ae 71 !O7,50.,_,310
~ 0 0.. 48 51 ''',31e,31'
'I-
~
~
0
g: ':
0.4 100
50 ~I~,al.
!lI,524,1.8,529
~O~--~----~--~----~--~----~--~----~
50
10 30 70 90 110 130 150 170
N
Accumulated Catalyst Residence Time (min.) eN" .~t 8..
10r-------------------------------------------,
0.15
DISCUSSION
Co)
~
(11
Co)
,flo.
0)
Table 4b. The Effect of Nitrogen and Water Di1utiona on the Distribution of Product Hydrocarbons
323 5.23 30.2 2.4 32.5 17 .6 4.0 6.4 0.03 1.85 91.0
324 4.8 29.7 2.3 32.0 16.5 4.0 9.8 0.04 1.7 90.0
311 5.9 30.6 2.6 31.5 16.8 4.0 8.2 0.08 97.0
318 3.55 29.3 2.16 31.9 17 .5 3.83 9.7 0.03 98.1
319 3.44 27.2 1.9 30.0 19.2 4.84 8.88 0.09 2.19 86.0
333 9.25 39.5 3.69 34.0 4.95 1.54 7.17 0.14 97.5
334 5.89 36.7 2.04 37.5 7.09 2.05 8.18 0.10 1.14 99.3
330 7.0 35.0 1.8 37.1 9.0 1.5 8.8 98.0
r
r
C
m
-i
}>
r
()
0
z
<
m
:D
(/)
(5
z
0
"T1
Table 4c. The Effect of Nitrogen and Water Di1utiona on the Product Yields ~
m
-I
:r
WT % PRODUCT YIELD DISTRIBUTION
»
z
RUN 0
H2 O CO CO2 H2 CH4 C2 H4 C2 H6 C3 H6 C3 H8 C4 HI0 C4 H8 C4 H6 C5 DME r
-I
0
r
319 57.9 0.16 0.26 0.08 1.39 l1.00 0.80 12.14 7.79 1.96 2.78 0.94 0.89 1.06 0
~
333 36.76 1.52 0.77 0.278 1.67 7.2 0.7 5.97 0.88 0.22 1.28 0.026 42.6 ~
334 34.19 0.138 0.78 0.139 2.02 12.36 0.74 12.91 2.44 0.69 2.816 0.037 0.38 30.31 0
r
m
()
C
WT % CARBON YIELDS r
»
:D
0
319 0.02 0.15 2.6 23.4 1.6 26.0 15.9 4.0 5.8 2.0 1.8 1.0 r
m
333 1.53 0.55 3.1 15.5 1.4 12.8 1.8 0.45 2.8 0.10 56.0 ::!!
334 0.10 0.50 3.5 23.8 1.4 24.8 4.5 1.2 5.4 0.10 0.56 35.6 z(/)
w
-...J
"""
w
~
(Xl
Table 5a. Effect of C02 Dilution (CH30H/C02 = 50/50) on Hydrocarbon Yields at 430°C with MWHSV =
0.40 and TWHSV = 0.954 (Batch Catalyst Run 336).
aEquivafent conversion to- hydrocarbolls(HC1- = (moTes CH30H converted to HC) / (Moles of CH30H
reacted - 2(mo1es of DME in the product» x 100.
r
r
C
!!l.
»
r
("')
o
z
<
m
:ll
en
6
z
o"T1
:s::
m
Table 5b. Effect of C02 Dilution (CH30H/C02 = 50/50) on the Distribution of Hydrocarbons Formed -I
at 430°C with MWHSV = 0.40 and TWHSV = 0.954 (Batch Catalyst Run 336). J:
»
Z
o
r
TIME ON WT% HYDROCARBON DISTRIBUTION -I
STREAM CARBON o
BALANCE r
(min) CH4 C2 H4 C2 H6 C3 H6 C3 H8 C4 H10 C4 H8 C4 H6 C5 o
:E
:s::
or
15 8.16 19.56 1.72 17 .47 30.8 7.82 7.27 0.03 7.0 85.6 m
("')
30 4.67 22.20 1.92 22.6 27.3 6.85 7.43 0.03 6.9 91.6 C
60 4.17 25.65 2.23 32.0 18.30 3.43 11.08 0.05 2.99 96.0 ~
90 4.18 31.36 1.90 41.73 9.93 2.04 7.67 0.068 1.09 95.6 :ll
120 4.83 36.90 1.95 40.25 6.95 1.58 5.97 0.07 1.44 97.0 or
m
150 9.25 43.38 3.70 30.0 6.08 0.73 6.1 0.09 0.66 99.0 "T1
180 16.6 36.60 6.73 26.0 4.77 1.12 7.09 0.13 0.73 102 Z
en
w
""co
IN
UI
o
Table 6a. Effect of H20 Dilution (CH30H/H20 = 50/50) on Hydrocarbon Yields at 430°C with MWHSV =
0.416 and TWHSV = 0.651 (Batch Catalyst Run 337).
r
c
C
m
-I
»
r
(")
oz
<
m
::0
C/)
oz
o
'T1
s:::
m
-I
Table 6b. Effect of H20 Dilution (CH30H/H20 = 50/50) on the Distribution of Hydrocarbons Formed ::r:
at 430°C with MWHSV = 0.416 and TWHSV = 0.651 (Batch Catalyst Run 337). l>
z
or
-I
TIME ON WT% HYDROCARBON DISTRIBUTION o
CARBON r
STREAM BALANCE o
(min) CH4 C2 H4 C2 H6 C3 H6 C3 H8 C4 HlO C4 H8 C4 H6 C5 :E
s:::
or
m
30 4.08 20.3 2.16 20.0 32.2 7.36 8.46 0.01 5.29 98 (")
C
60 3.55 25.8 2.07 27.27 20.7 4.65 10.9 0.011 4.96 97 r
90 3.6 30.2 1.99 35.8 12.36 2.48 11.03 0.016 2.41 97 l>
::0
120 3.83 32.3 1.80 43.3 8.46 1.8 7.45 0.017 0.95 97 or
150 4.57 39.8 1.86 41.31 5.52 1.30 5.08 0.023 0.47 97 m
180 6.77 44.4 3.19 31.52 5.59 0.95 6.54 0.28 0.74 97 !!
Z
C/)
w
(11
352 L. LlU ET AL.
40
N
I-
:. 35
z
....
0
I-
30
~
m
....
~
25
l-
....
I/)
c 20
z
0 15
m
~
<[
0 10
0
~
c
>- 5
:J:
o •• •
3 5 7 9 II 13 15 17 19 21 22
Propane WI. %
Figure 13. Hydrocarbon distribution for 703, 723, and 753°K and
WHSV of 0.4, 0.9, and 2.0.
20
e
16 100% CH,OH
12
UJ
Z
<[
e_
Q.
o
~
o
Q.
8 50 % N, 100% CH,OH
I:. 50 % H,O
-4
MWHSV
Figure 14. Effect of space velocity and dilution on propane content
at 703°K.
CONVERSION OF METHANOL TO LOW MOLECULAR OLEFINS 353
----- .,
II)
z
iL
.--
101
.. .
..J
o
60
~
~ ~ ~ ~ ~ u u ~ ~ M
PARAFFINS WT. %
MWHSV-O.4
SO
50
40 Hydrocarbons
s.o
5.0
~
4.0
~ 3.0
2.0
Z
0 1.0
~
::J 70 MWHSV·2.0
a:
III
l-
VI
so
0
I-
(.) 50
::J
0
0 20
0::
Q.
10
~
....;=====:::jt:::~::=--..!:'HYdrOCarbons
0
5.0
4.0
3.0
2.0
1.0
"'"
RUN NO. 328 329 330 333 334 331 325 332 326
TWHSV (hr-1 ) 0.4 0.4 2.0 1. 75 1.54 1.92 0.4 1.92 0.37
DILUENT 1 N2 H2 O
MWHSV (hr- ) 0.4 0.4 2.0 0.93 0.98 1.92 0.4 1.92 0.37
MeOH CONVERSION (%) 100% 98% 35% 55% 41% 44% 100% 49% 100%
% CLOSURE OF 82.0 90.9 98.9 97.5 99.1 95.9 81.9 95.4 81.0
CARBON BALANCE
r
C
C
m
-t
»
r
CONVERSION OF METHANOL TO LOW MOLECULAR OLEFINS 355
,..
Material
80 lances
In"ludln9 Ca1l..
~
,..
'04
••
- - - ' 3'
331
t-~
33
32
329
-- r'"
32.5
- -
..'
•. 3
Coke
Oepo.ition •. 2
or.ml/ ""ln
•. 1
...
,..
Carbon
Closures
EMCll,ld l"'9 Co._ 9.
~
e.
10 20 30 40
•
Parallin (1:. C.H2 •••) Conlenl
.',
WI ,., In th . Hydroc;.rbon F • •c.llon
et ale [lOJ.
The activity of this catalyst was higher than the catalyst stud-
ied by Singh [38J even though similar procedures were used in its
preparation. By using the rate equation, equation (19), developed
by Anthony et ale [24J in the material balance equation for a plug
flow reactor approximately 95% conversion is obtained for a space
velocity of 0.4 hr -1, whereas the space velocity for the internal
recycle reactor required for this conversion was 0.1 hr- l • There-
fore, based on reactor type, space velocities of 3 to 4 times that
obtained in the internal recycle reactor should yield the same meth-
anol conversion. As shown in Table 4, Run 319, methanol conversion
of 94% was obtained at a WHSV of 0.94. The increase in apparent ac-
tivity is probably due to a reduction in catalyst particle size. One
eighth inch cylindrical pellets were used by Singh [38], whereas
356 L. LlU ET AL.
% CARBON
BALANCE
90
CLOSURE
85 .~
11 13 15 17 19 21
PROPANE WT %
average particle diameter for this study was 173 microns. This could
result in reduction in the resistance to pore diffusion, thereby in-
creasing the global reaction rate. A corresponding increase in space
velocity would be the result for the same conversion of methanol.
CONCLUSIONS
ACKNOWLEDGEMENTS
APPENDIX 1: NOMENCLATURE
Methanol Equivalent
moles of methanol converted to Hydrocarbons
Conversion to
[moles of methanol reacted - (2 moles of DME)]
Hydrocarbons
REFERENCES
361
362 o. M. KUT ET AL.
for the selective production of light olefins over modified zeolites
will be discussed later. The use of biomass as an ethylene source
can be interesting from an economic standpoint for developing coun-
tries that have scant petroleum resources and balance of payment
problems. If such a country has substantial amounts of fermentation
ethanol available (derived from agricultural products like sugar mo-
lasses), the production of ethylene by catalytic dehydration in small
production units (5000-10,000 t/yr) at low capital costs is efficient
and economic. India and Brazil are typical examples of such coun-
tries. Winter and Eng [3] have compared the economics of ethylene
production by naphtha cracking and ethanol dehydration. The naphtha
cracking process is capital cost intensive and not very selective in
ethylene, but the ethylene cost can be reduced by selling other crack-
ing products like propylene. The capital investment costs for the
ethanol dehydration are lower and there are normally no valuable side
products; the ethylene price is dictated by the price of the fermen-
tation alcohol [3,4]. To attract investment in the ethanol-based
chemistry in Brazil, the price of one liter absolute fermentation al-
cohol for use as a chemical feedstock is stipulated to be limited to
35% of the official petrochemical price of one kilogram of ethylene
[5]. In Brazil, ethanol is replacing naphtha in two of its major
applications: as raw material for ethylene production and as an auto-
motive fuel component. Luchi and Trindade [5] estimated the econom-
ics of various ways of ethanol utilization. Their analysis showed
that the highest economic value for fermentation ethanol is achieved
when aqueous alcohol is used as a feedstock for ethylene production.
In this case, the value of the hydrated alcohol is 95% of the cost of
an equal volume of naphtha in Brazil. When anhydrous alcohol is used
as a 20/80 fuel blend with gasoline, ethanol has an economic value
of 85% of the cost of an equal volume of naphtha.
hydration of ethanol dates from 1797 when Bondt, Deiman, van Troo-
stwyk, and Lauwerenburg [9] produced ethylene by passing ethanol or
diethylether vapors over a heated contact. The early literature was
reviewed by Winfield [10] in 1960 and included the main developments
in the industrial realization of the dehydration processes [3,11].
The catalysts used were alumina, silica, acid treated clays and phos-
phoric acid supported on coke granules. The operation temperatures
were ca. 360°C. The phosphoric acid system gave a better ethylene
selectivity (99.5%) compared to 97% on alumina contacts. The main
drawback of the phosphoric acid was leaching out of the active compo-
nent, which caused serious corrosion problems. Furthermore, the cat-
alyst, which was difficult to handle, required frequent discharge and
replacement [11]. The dehydration of ethanol is endothermic, requir-
ing 1590 kJ/kg ethylene produced [12]. The main products are ethyl-
ene and diethylether. Ethylene is produced by a parallel-consecutive
reaction scheme that will be discussed later. Depending on the cata-
lyst system and operation conditions, either the parallel or the con-
secutive character of the reaction can predominate:
~ Acetaldehyde _ _
Ethanol---------:, Ether - - - -
{-
Ethylene - - ~ Carbon
General Features
~Ether
Ethanol ~ 1 lk3
~ Ethanol + Ethylene
k2 'Ethylene
Little work has been done with pure silica because of its low
activity. In the temperature range of 20S-30SoC, however, attrition
ground silica shows a relatively high olefin selectivity for ethanol
dehydration, which depends on the impurities in the catalyst materi-
al [Sl]. Various silica-aluminas were studied by De Mourgues and
co-workers [40,S2,S3]. At low temperatures (ISO-200°C) with this
mixed catalyst, twice as much ether as ethylene was produced. Pyri-
dine inhibited the olefin formation more strongly than the ether
formation, and therefore olefin selectivity decreased [S3]. Poison-
ing experiments with sodium ions and perylene showed that the acidity
is correlated with Bronsted acidity and not with Lewis acidity [40].
Water inhibited both the olefin and the ether formation [S3]. The
natural and synthetic zeolites, as a special class of aluminosili-
cates, show such a distinct activity and selectivity behavior that
CATALYTIC CONVERSION OF ALCOHOLS TO OLEFINS 369
Winfield [10] has pointed out that thoria is a very active de-
hydration catalyst with a high olefin selectivity and with negligible
ether formation. Preparation and pretreatment strongly influence
the activity and the dehydration/dehydrogenation selectivity of these
catalysts [54,55]. For example, a most active and selective catalyst
for dehydration was prepared by the pyrolysis of thorium oxalate at
350-400°C [10]. Thoria prepared from thorium hydroxide, on the other
hand, gave more dehydrogenation products than dehydration products
[54]. This strong dependence of the selectivity on the pretreatment
conditions was demonstrated by Davis and Brey [55] on carefully char-
acterized thoria samples with 2-octanol dehydration as a test reac-
tion [55]. It was concluded, that the high surface area thoria was
selectively dehydrating, whereas the low surface area catalyst was
more selective for dehydrogenation.
(B/P ~ 0.8) gave ethylene as the main product (70%). On the other
hand, for boron-rich catalysts (B/P > 0.8) only small amounts of
ethylene (10%) were formed, whereas the ether selectivity was about
70%. The high ether selectivity of boron-rich catalysts can be de-
stroyed by poisoning with hydrogen chloride; in that case the ethyl-
ene yield increased from 10 to about 50% [73]. X-Ray analysis showed
that in phosphorous-rich samples (B/P ~ 0.8), segregation of P205 on
the surface can be observed and BrBnsted acid sites are dominant.
By contrast, on boron-rich catalysts most of the acidic sites are
Lewis type. Since this type of catalyst has a low olefin selectivity
and this selectivity can be increased by the introduction of protonic
acidity (HCl poisoning), Haber and Szybalska [73] came to the con-
clusion that on phosphate catalysts the olefin selectivity is corre-
lated with the BrBnsted acidity. Two calcium phosphate catalysts,
with well defined hydroxyapatite structure, were studied by Kibby
and Hall [74]. On the calcium-deficient hydroxyapatite, only dehy-
dration took place and olefins were the only products formed, but the
studied reaction temperature of 352°C was far too high for ether
formation from primary alcohols. With the stoichiometric proportion
of calcium in the hydroxyapatite, dehydrogenation also took place.
EXPERIMENTAL DATA
Apparatus
Materials
Products formed
Weight
Reaction Ethanol PPA Temperature Total gas Ethylene Ether C4- increase
mode (g) (g) (OC) (m1) (%)a (m1) (m1) (m1) PPA (g)
aYie1d in moles of gaseous product per mole of ethanol feed, with appropriate stoichiometry
o
:s:
c
"
-i
m
-i
»
r
CAT AL YTIC CONVERSION OF ALCOHOLS TO OLEFINS 381
VGtot [ml]
r-----------------------------------,
T rOC]
208
8000
171
158
130
8
t [h]
8
t [hI
containing compounds.
The total gas production showed zero order kinetics in the first
hours of the experiments, accelerating towards the end of the opera-
CAT AL YTIC CONVERSION OF ALCOHOLS TO OLEFINS 383
[mil
r-------------------------------------,
VG
~ ct • Ct.
[llIl] Ether
D Ethylene
~
LIquId
remaInIng
Reaction EA (kJ/mol)
The activation energy for the overall gas formation under semi-
batch experimental conditions is only 10.5 kJ/mol compared to 67
kJ/mol in batch experiments. This indicates that the reaction is now
under mass transfer control. Since no ethanol can be detected in the
gas phase at the beginning of the reaction, the gas-liquid mass
transfer of the reactant cannot be rate determining. In a first re-
action step, ethanol forms esters with PPA, which decompose to give
the different products [126,127]. These products have to permeate
through the catalytically active PPA layer before they reach the gas
phase. During this transport, various polymerization steps can occur
depending on the instantaneous strength of the acid. Downing and
Pearson [132] have shown that the acidity of PPA continuously decreas-
es with the decreasing P205 content. During the dehydration, the
water formed is bound to the catalyst; its acid strength, and simul-
taneously its catalytic activity for individual reaction steps,
changes in different ways.
% Im
r _II______________________________________--.
~ Ether
Ethanol -} C4 - -... C4+ - -... Hydrocarbons.
----- Ethylene ...-------
r-----------------------------------,
VG Imll
0.6
o 6 8 2 L 6 8
t [ hI
XN-lOlO and AG MP-50 (Figure 7a). Under the same experimental con-
ditions (below the boiling point of ethanol in order to have a liquid
phase reaction), Amberlyst 15 produced diethylether with a selectiv-
ity of 97.5%. Figure 7a shows that initially only ethylene and its
oligomers (C4 + C4+) are formed. With increasing conversion, the
water formed blocks the most active sites and ether is formed in an
increasing amount. Starting with aqueous ethanol (80% w/w) , no eth-
ylene was formed over Amberlyst XN-lOlO, only its oligomers formed
(Figure 7b). The activity in terms of gas produced was only about
75% of the value for the dry alcohol. In all of the experiments no
other product could be detected in the liquid phase. After 8 hr re-
fluxing, only about 1400 ml gas was produced from 40 g of absolute
alcohol used (xA = 6.5%) with all three resins.
.-_-_Ether
0.4
1.0. r . ; : : - - - - - - - - - - - - - - - - - - - - - - - - ,
x
0..8
0..5
0.4
0..2
CONCLUSIONS
REFERENCES
197 (1969).
44. H. KnHzinger and H. Stolz, Ber. Bunsenges. Phys. Chern., 74:1056
(1970) •
45. K. Jiratova and L. Beranek, App1. Cata1., 2:125 (1982).
46. H. KnHzinger and R. KHhne, J. Cata1., 5:264 (1966).
47. J. E. Stauffer and W. L. Kranich, Ind. Eng. Chern., Fundam.,
1:107 (1962).
48. H. J. Solomon, H. Bliss, and J. B. Butt, Ind. Eng. Chern.,
Fundam. 6:325 (1967).
49. P. K. Gupta and M. Ravindram, Hung. J. Ind. Chem., 6:115 (1978).
50. H. Knozinger, K. Koch1oefl, and W. Meye, J. Catal., 28:69 (1973).
5l. W. J. Hatcher, Jr. and L. Y. Sadler III, J. Cata1., 38:73 (1975).
52. L. de Mourgues, F. Peyron, Y. Trambouze, and M. Prettre,
J. Cata1., 7:117 (1966).
53. L. de Mourgues, D. Barthomeuf, F. Figueras, M. Perrin, Y. Tram-
bouze, and M. Prettre, Proc. Int. Congr. Cata1., 4th, 2:187
(1971) •
54. w. S. Brey, Jr., B. H. Davis, P. G. Schmidt, and C. G. Moreland,
J. Cata1., 3:303 (1964).
55. B. H. Davis and W. S. Brey, Jr., J. Cata1., 25:81 (1972).
56. A. J. Lundeen and R. van Hoozer, J. Am. Chern. Soc., 85:2180
(1963).
57. A. J. Lundeen and R. van Hoozer, J. Org. Chem., 32:3386 (1967).
58. P. Canesson and M. Blanchard, J. Cata1., 42:205 (1976).
59. K. Thomke, Z. phys. Chem., Neue Fo1ge, 105:75 (1977).
60. K. Thomke, Z. phys. Chem., Neue Fo1ge, 105:87 (1977).
6l. K. Arata and H. Sawamura, Bull. Chern. Soc., Japan, 48:3377
(1975).
62. B. H. Davis and P. Ganeson, Ind. Eng. Chem., Prod. Res. Dev.,
18:191 (1979).
63. T. Takeshita, R. Ohnishi, and K. Tanabe, Cata1. Rev., 8:29
(1974).
64. T. Yamagushi and K. Tanabe, Bull. Chern. Soc., Japan, 47:424
(1974).
65. I. Mochida, A. Kato, and T. Seiyama, J. Cata1., 22:23 (1971).
66. J. B. Moffat, Cata1. Rev.-Sci. Eng., 18:199 (1978).
67. L. K. Freid1in, V. Z. Sharf, and Z. T. Tukhtamuradov, Kinet.
Kata1. 5:347 (1964).
68. J. B. Moffat and A. S. Riggs, J. Cata1., 28:157 (1973).
69. J. B. Moffat and A. S. Riggs, J. Cata1., 42:388 (1976).
70. H. Knozinger, H. Buehl, and E. Ress, J. Cata1., 12:121 (1968).
7l. A. Tada and K. Mizushima, Nippon Kagakti Kaishi, 278 (1972).
72. S. S. Jewur and J. B. Moffat, J. Cata1., 57:167 (1979).
73. J. Haber and U. Szyba1ska, Discuss. Faraday Soc., 72:263 (1981).
74. C. L. Kibby and W. K. Hall, J. Cata1., 29:144 (1973).
75. A. Tada and M. Yoshida, Nippon Kagaku Kaishi, 856 (1973).
76. A. Tada, M. Yoshida, and M. Hirai, Nippon Kagaku Kaishi, 1379
(1973).
77. S. Kikkawa, Y. Shimizu, and S. Higuchi, Chern. Lett., 849 (1979).
78. Y. Ono, T. Mori, and T. Keii, Proc. Int. Congr. Cata1., 7th.,
392 O. M. KUT ET AL.
p 1414 (1981).
79. Y. Ono and T. Mori, J. Chern. Soc., Faraday Trans I, 77:2209
(1981).
80. Y. Ono, T. Baba, J. Sakai, and T. Kei1, J. Chern. Soc., Chern.
Commun., 400 (1981).
8l. T. Okuhara, A. Kasai, N. Hayakawa, M. Misono, and Y. Yoneda,
Chern. Lett., 391 (1981).
82. N. Hayakawa, T. Okuhara, M. Misono, and Y. Yonada, Nippon Kagaku
Kaishi, 356 (1982).
83. R. Rudham and A. Stockwell, in "Catalysis (Specialist Periodi-
cal Reports)," C. Kemba11, ed., The Chemical Society, London,
Vol. 1, p 87 (1977).
84. M. S. Spencer and T. V. Whittam, in "Catalysis (Specialist Pe-
riodical Report)," C. Kemba11 and D. A. Dowden, ed., The
Chemical Society, London, Vol. 3, p 189 (1980).
85. D. Barthomeuf, Stud. Surf. Sci. Cata1., 5 (Catalysis by Zeo-
lites), Elsevier, Amsterdam, p 55 (1980).
86. N. Y. Chen, J. Phys. Chern., 80:60 (1976).
87. D. H. Olson, W. O. Haag, and R. M. Lago, J. Catal., 61:390
(1980).
88. E. M. F1anigen, J. M. Bennett, R. W. Grose, J. P. Cohen, R. L.
Patton, R. M. Kirchner, and J. V. Smith, Nature (London),
271:512 (1978).
89. M. Ra1ek and O. Grubner, Proc. Int. Congr. Cata1., 3rd, 1302
(1965).
90. 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).
91. v. J. Fri1ette, W. O. Haag, and R. M. Lago, J. Cata1., 67:218
(1981).
92. S. L. Meisel, J. P. McCullough, C. H. Lechtha1er, and P. B.
Weisz, CHEMTECH, 6:86 (1976).
93. C. D. Chang and A. J. Silvestri, J. Catal., 47:249 (1977).
94. C. D. Chang, W. H. Lang, and R. Smith, J. Cata1., 56:169 (1979).
95. N. Y. Chen.and W. J. Reagan, J. Cata1., 59:123 (1979).
96. Y. Ono and T. Mori, J. Chem. Soc., Faraday Trans. I, 77:2209
(1981).
97. P. Dejaifve, J. C. Vedrine, V. Bo1is, and E. G. Derouane,
J. Cata1., 63:331 (1980).
98. J. R. Anderson, T. Mole, and V. Christov, J. Cata1., 61:477
(1980) •
99. W. W. Kaeding and S. A. Butter, J. Cata1., 61:155 (1980).
100. P. G. Rodewald, U.S. Patent 4,700,219 (July 11, 1978); assigned
to Mobil Oil Corp.
101. Y. Matsumura, K. Hashimoto, S. Watanabe, and S. Yoshida, Chern.
Lett., 121 (1981).
102. J. C. Oudejans, P. F. van den Oosterkamp, and H. van Bekkum,
App1. Cata1., 3:109 (1982).
103. R. A. Rajadhyaksha and J. R. Anderson, J. Cata1., 63:510 (1980).
104. E. N. Givens, C. J. Plank, and E. J. Rosinski, U.S. Patent
CATALYTIC CONVERSION OF ALCOHOLS TO OLEFINS 393
(1961) •
133. G. A. 01ah, A. P. Fung, and D. Meidar, Synthesis, 280 (1981).
CONVERSION OF METHANOL TO HYDROCARBONS ON HETEROPOLY COMPOUNDS
395
396 J. B. MOFFAT AND H. HAYASHI
INTRODUCTION
Heteropoly compounds have been known for over one hundred and
fifty years. The first heteropoly compound, ammonium l2-molybdophos-
phate, was prepared by Berzelius in 1826. Although considerable work
has been done on heteropoly compounds since that time, it is only
recently that interest in their catalytic properties has developed.
Since several excellent reviews of the general properties of heter-
opoly compounds have appeared, one in 1976 [1] and the other in 1978
[2], discussion of these aspects will be abbreviated here.
Both free acids and salts are found within the class of substan-
ces labelled as heteropoly compounds. The anions, of high molecular
weight and structurally complex, consist of one or more central
atoms, surrounded by a number of atoms different from the central
one, together with oxygen atoms. There are many elemental possibil-
ities for the central atom but those of special interest in the pres-
ent work are phosphorus and silicon. Equally well the choices for
the peripheral atoms are many. The discussion here will be limited
to molybdenum and tungsten.
[5,6]. This acid has a greater hydrolytic stability than the corre-
sponding molybdenum compound [1]. The l2~tungstophosphoric acid com-
pound can also be prepared by ether extraction of acidified solutions
of tungstate and phosphate [3,6]. An almost white crystalline solid,
which decomposes on standing and hydrolyzes in water, is obtained.
The dimeric acid containing phosphorus and tungsten, that is lS-tung-
stodiphosphoric acid (H6P2W1S062), may be prepared from sodium tung-
state and phosphoric acid [3].
tached to one peripheral metal atom, and which are located on the
surface of the anion [9,10].
EXPERIMENTAL
H3 PMo 12 04 0 Na3PW12040
H3 PW12040 Ca4[ PW12 040]2
H4 Si 12040 Zn3[ PW12 0 40]2
Mg3[PW12040]2
AlPW12040
BPW12040
Zr3[PW12040]4
Zr3[PW12040]4
RESULTS
With the four heteropoly acids studied, the acid strength in-
creases with calcination temperature. The highest strength measured
CONVERSION OF METHANOL TO HYDROCARBONS ON COMPOUNDS 401
650~ AIR
525,"AIR
[ I I ! I I I I I
60 55 50 45 30 25 20 15 10
28 (degrees)
/
H3PWl2040
J
No
1'. lI\ ..J,
-
1 cAlli LA
B
.1 J.. .A
Zr
A A ~ A..- .A .A
AI
Co
.A
A.
.-
M9
Zn ,,".,...
""'-
60 50 40 30 20 (0
28(0 )
Figure 6. Powder X-ray diffraction patterns of metal salts of 12-
tungstophosphoric acid calcined at 150°C in air for 2 hr
[20].
CONVERSION OF METHANOL TO HYDROCARBONS ON COMPOUNDS 403
100~---------.-=~-----r~--,14
XA
(0 1 12 ~
325°C
10 ::
w
60 8 >=
Z
6 g
cO 40 a::
"l
)(
4 u "
o
a::
20 2 ~
x
0orT--------~~~~=rr_--~14
( b1 12~
80 352°C
log
W
60 8 >=
z
40
6 g
a::
)(
"l
4 U "oa::
20
2 ~
%
10
100~------~~--~-----r----,14
(c) 12~
80 403°C
109
~
60 8 >-
Z
6g
40 a::
)(
"l
4g"
20
a::
2 ~
x
C2
OLL~--~~==~~-L--~O
10 100 W/F 1000
100 ,00, ~
I~ • .'. 1'4 I'" :it= •• I'"
xA ( b)
'zi 12 ~ 12~
80l- He 80
350.C 4 100 100
i 60~
1:(0) 103
LLJ
>=
i
... 60 8
~
..,
>=
~ 8
~
..,
>=
B
>- Z
>-'" I \.. C3 ~ Z >- '"
6 II) c!I 6 g 6 g
ct! 40 II: "0 c!I "0 II:
« II:
C « c( «
xC )( X <I U
"g " U
o o
II: a: II:
20 o 2.0 o
2 )0- 2 ~ 2 >-
% % %
The results obtained with the sodium salt showed good agreement
between the conversion of methanol and the expected yield of ether,
thus demonstrating that the dehydration reaction
XA
c
irl
)::
Z 10
II!
~~ ~~ J I~~/ T~~~ c
I
o x
~ ~ 25
)0-
r
01 -dO A: ~" - _ . -0
~
(t )
---::1 Z, 3/4 PW'2 l 'l' 9PW'2
C
...J ~
!AI
10 ~
II!
i c
!l'
X s::
l5'" 0
lS ~
~ ""»
--i
O~ »Z
30 100 300 1000 30 100 300 1000 30 100 300 1000 30 100 300 1000
W/F W/f W/f W/F 0
J:
Figure 9. Methanol conversion (XA)' yield of dimethyl ether (Y E), and hydrocarbon J:
yields at various residence times on metal salts of 12-tungstophosphoric acid »
[20]. The reaction temperature was 350°C. The catalysts were previously »-<
calcined at 400°C in He for 2 hr. en
J:
CONVERSION OF METHANOL TO HYDROCARBONS ON COMPOUNDS 407
DISCUSSION
Calcined in He H2 He H2
Methanol Converted (%) 99.0 94.2 98.0 100.0
Dimethyl ether 56.0 41.8 25.0 26.1
Produced (%)
11'1 11'1
~ ~
u u
:::> :::>
c
8
a:
1 oa:
0.. 0..
20 AI
;e
t...
.,.
u 15
u..
0
0
...J
W
~
~
::::>
~
x
ex
~
No
0.28
Figure 11. Partial charge on oxygen (00) in the metal salts of 12-
tungstophosphoric acid and the maximum yield for C4 hy-
drocarbons from the conversion of methanol [20].
410 J. B. MOFFAT AND H. HAYASHI
further experimental work. The DME may interact with Bronsted acid
sites through the negatively charged oxygen of the DME:
17////1//11111111/
Such an interaction presumably increases the electron density between
the 0 of DME and the proton, to the detriment of that between the
former oxygen and the methyl groups. This will lead to a scission
of one of the O-CH3 bonds, producing methanol and a methoxy group at-
tached to the catalyst. Since the hydrogen atoms of the catalyst
bound methoxy group will have their binding electron density reduced,
the process may be repeated, thus forming one or more C-C bonds.
ACKN9WLEDGEMENTS
REFERENCES
INTRODUCTION
SILVER PROCESS
413
414 C. J. MACHIELS ET AL.
°z Ag ZO(s) (1)
ZH(s) + HZ (5)
this stage, only methoxy groups remain on the surface, and this has
been confirmed by spectroscopic methods. On further heating, the
methoxy groups begin to react, and hydrogen is abstracted and spills
onto the metal surface. This hydrogen may either combine with an-
other hydrogen resulting in desorption of H2 , or it may combine with
a methoxy group causing methanol to desorb. The resultant formalde-
hyde formed by methoxy decomposition readily desorbs. The mechanism
of the methanol conversion to formaldehyde is similar on a silver
surface, except that it is complicated by competing reaction channels.
This mechanism explains why the limiting ratio of hydrogen-to-water
in the silver process is one:
MOLYBDATE PROCESS
REDOX CYCLE
o
II
o
/ Mo .., 0
H
HCfJ '_H
H .' IC
"o ,~'
• 07
....
Mo
0/ '0
I I
Figure 1. Schematic showing the proposed mechanism for methanol ox-
idation to formaldehyde over a molybdate catalyst.
0.12
Surface Sl!!!;ies
0.11
MeOH
HOH
0.10
-OM. ~
MeOH
-OH
0.09 -OMe
•uc:
~0
0.08
0.07
~
MeOH
HOH
CH Stntch
Anllsym .£!!..!!!l!
~'~I
MeOH
~ 0.08
-OH
1 MeOH MeOH
0.05
0.02
4000 3700 3400 3\00 2800 2500 2200 1900 1600 1300 1000
Wavenumber
3). The slow step in the reaction cycle (Figure 1) is the breaking
CONCLUSION
REFERENCES
INTRODUCTION
Uses
Methyl chloride is mainly used as raw material for other chI oro-
methanes, in the manufacture of silicone resins, and for solvent of
butyl rubber manufacturing.
419
420 S. AKIYAMA ET AL.
120.000
/
100.000
...
..
~
0
80.000
!i 80.000
8:
....
ill
g 40.000
"
-c
20.000
The choice between the two processes would depend upon raw mate-
rial availability, but more importantly upon the ability to utilize
the by-product hydrogen chloride, either through marketing as muri-
atic acid or as raw material for other processes such as oxychlorin-
ation of ethylene.
Chlorinnation
Hydrochlorination Chlorination
catalytical (250~3500C)
vapor phase thermal (400~500°C)
A1203
HYDROCHLORINATION OF METHANOL
Catalyst
~
250 270 300
Si02 17.1 52.7 47.3 20.0 56.1 43.9 21.6 48.6 51.4
8i02-AbOa
3 1.6 33.5 66.5 43.4 41.6 58.4 57.7 51.6 48.4
(AbOa 28.6%)
8i02 -AbOa
(AbOa 87%) 89 68.1 31.9
T -AbOa 93.6 98.5 1.5 92.8 98.8 1.2 93.8 99.2 0.8
Reaction Condition
Reaction Mechanism
99
98
~
r::
0
0.-1 97
r/I
I-t
CD
I-
r::
0
0
96 ~
SCD ~
::II 0.-1
95 2 :;:l
I-
0
....CDCD
fIl
~
CD
::II
1.05 1.1 1.15 1.2 0
HCf/MeOH mol ratio
98
§ 97
0.-1
r/I
I-t
CD
I-
§ 96
o ~
S ~
:; 95 . 2 o~
0.-1
b
....CDCD
fIl
~
o
j
-L____~____L __ _-L____~O
L __ _
Residence time
Figure 5. Residence time dependence on methanol conversion and
selectivity. HCl/MeOH = 1.1 mol/mol at 270°C
CATALYTIC CONVERSIONS OF METHANOL TO CHLOROMETHANES 425
Q = 25 Kcal/mol, exothermic
Catalyst
Reaction Control
Q)
H
;l A ../stable
01
H ," .. ___ ....1'.. _____ - - - Cf2Conversion
Q)
.'
..-1
+'
o
01
Q)
IX:
Time ~
;!
,,,
'iR I
I
s::i
0
•.-1
fIl
H
Q)
I> I
,
~
0
0 99.5 I
Q)
I
~
•.-1
H I
0
...... 0
.0::: I
0 I
I
99.0 L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _---l
CATALYST, PPM
Figure 7. Catalyst concentration and chlorine conversion at 110°C
as follows:
Cat R - R'·,
H H H H
I I I I
C1·+H-C-C1 - - H-C·+HC1, H - C ·+C12 H-C-C1+C1·
I I I I
H C1 C1 C1
H H H H
I I I I
C1'+H-C-Cl - - C1-C·+HC1. C1-C·+ C1 2 C1-C-C1+C1·
I I I I
C1 C1 C1 C
CATALYTIC CONVERSIONS OF METHANOL TO CHLOROMETHANES 427
Product Quality
Process Description
50
40
~
30
...
0
II
20
Temp. kl k2 k3
Liquid Phase 90°C 1.0 0.38 0.076
,
;t j
ttl +,
:~U
. r .•.
-H I!H!!~ f-.+
•• i-
.'-t-
---:-
+-
+ ·+II~:.
~I'
•.
. ++ ... ~ .. ~ ,
---"-+-
,-
. -
- --- ~
f-L - -- ~j ~-
- .-
-.
-r- -~
-
-+. - t-- 1I-=t 1-+ t-- r--
f+--i-..+- r-
,
h- f+
Process Features
C1 ~ __________ ~ ________- - J
WiSHIHG AN D OJ:HTDRJ.TION'
Figure 12. The new Tokuyama Soda Process for producing chloro-
methanes from methanol and chlorine.
ECONOMICS
COMMERCIAL APPLICATIONS
cently by LCP Chemicals & Plastics Inc., for revamping of its exist-
ing plant in }!oundsville, West Virginia, USA. Completion of the
revamping is scheduled for late 1983.
REFERENCES
Richard G. Herman
Center for Surface and Coatings Research
Sinclair Laboratory, #7
Lehigh University
Bethlehem, PA 18015
INTRODUCTION
433
434 R. G. HERMAN
Formation of Alkylamines
Ever since the early research of Sabatier and Mailhe [4], the
formation of alkylamines from ammonia and alcohols has been well-
known. This catalytic reaction is carried out in the vapor phase at
elevated temperature and pressure and results in the formation of a
mixture of primary, secondary, and tertiary amines and water, as in-
dicated by the unbalanced equation 1, where R = an alkyl group such
(1)
(2)
In both cases, the reactor system would be pressurized with H2, and
in flow systems the hydrogen would function as the carrier gas. The
catalysts that will be discussed are subdivided into the following
three classes:
(i) dehydration oxidic catalysts,
(ii) hydrogenation-type supported Group VIII metals, and
(iii) copper-based catalysts.
supports, and of alloys of copper with Group VIII metals. The re-
action conditions used to alkylate ammonia or amines with alcohols
are similar to those utilized with the supported Group VIII metals.
However, the copper-based catalysts exhibit appreciably better se-
lectivities than those found with other catalysts. The recently
described copper-based catalysts are listed in Table 2, along with
the conditions of some test reactions [19-35]. The selectivity is
measured as the efficiency with which one hydrogen atom on the react-
ing amine (ammonia) is replaced by the alkyl group of the alcohol,
relative to other substituted amines. Even with (CH3)2NH, the se-
lectivity is not 100% because of other possible reactions, e.g. dis-
proportionation and transalkylation.
ROH +
o
N
+ OR
N
+
6
N
(3)
~
~
442 R. G. HERMAN
+
o 57% 24%
(4)
where the reactant methanol/pyridine mole ratio was 15/1 [38]. Uti-
lizing ethanol as the alkylating agent at 263°C, in only a 3-fold
excess, produced 2-picoline (2-ethylpyridine) in a 61% yield. More
recently over a Ni catalyst, pyridine has been methylated by methanol
with a 50% conversion to yield 100% 2-picoline, which can then be
converted to 2,6-lutidine (2,6-dimethylpyridine) with a 65% selec-
tivity at a 50% conversion [40]. The contact time used in some of
the investigations [38,39] was a rather long 12 sec in order to
achieve high conversions. However, long contact times can result in
poor selectivities because the picolyl hydrogens of 2- and 4- alkyl-
pyridines are more acidic than the corresponding benzylic hydrogens
[41]. This can lead to alkylation and growth of the side-chains on
substituted pyridines.
(5)
(6)
Table 3. Formation of Ethylenediamine from Ethanolamine and Ammonia »
r
r
"-<»
Reactants Product Total Temp. , -I
Catalyst Reference 6
(mole ratio) Se1ectivitya Conversiona Pressure
Z
0
."
205°C 44 0
Ni-Co-A1 203 NH3' 10 3:
H2N(CH2)20H, 1 54.8% 67%b 204 atm "tJ
0
H2' 14.4 c
z
Cu-Ni-A1 203 NH3, 10 195°C 44 0
CJ)
H2N(CH2)20H, 1 58.8% 57.2%b 149 atm
H2, 14.4
Cu-Co-Ni- NH3 ~1)' 3.5 c 160°C 45
A1203 H2N( H2)20H , lC -76%d 64%d 300 atm
H2
aWt %
bBatch autoclave result
cVol ratio .j:Io.
.j:Io.
dMole % w
444 R. G. HERMAN
alumina.
OH
°
II
R-C-OH + CH30H t
I
R-C-O-CH3 t
°
II
R-C-O-CH3 + H20 (13)
I
OH
Net Reaction:
(18)
EXPERIMENTAL
Catalysts
Reactor System
.
Temperatur@
Controller • I
c=J---- , .
Pump
II Themocouple Resevoir
, I
Switch Box
--rQ,-@ •01-
Cylindrical :: I I" 1 Digital
Heaeer , I I '1": Temperature
I II I' I Readout
I 1111
,) Ihl
Stainless
Steel ,/ )111 : :
Reactor I, I \ \
Back Pressure
Vent Regulator ',', Gas ChrOma~OgraPh
/,' \ \
I
I
\
,
\
Co Cas Bubble
Sampling Flow
Wet
Cylinder Meter I Test
_.J
Meter
Vent
Cas Cylinders
in the reactor with Pyrex beads, and the reactor was extensively
purged with hydrogen under ambient conditions. The catalyst was then
reduced in flowing H2 at 300°C for 2 hr and at 500°C for 2.5 hr. The
system was cooled to ambient temperature before pressurization was
carried out.
RESULTS
Synthesis of Methylbutylamine
at STP), and the catalyst ~as heated to 190°C. Using the high pres-
sure pump, liquid n-butano1 was injected into the gas stream at the
rate of 25 ~l/min. A 57.3% butanol conversion to products occur with
a 96% selectivity to methy1buty1amine. A small amount of methy1-
dibuty1amine was formed, as indicated in Table 4, and the quantity
of water formed corresponded exactly to the number of moles of n-bu-
tanol that had reacted. No ammonia or dimethylamine was detected in
the exit stream.
The GHSV of the gas stream was decreased, and this effectively
decreased the partial pressure of the monomethy1amine in the reactor.
Accordingly, the conversion of n-butanol decreased somewhat to 53.6%
as shown in Table 4. However, the same high selectivity toward
methylbutylamine was observed, even though the GHSV was 17% lower
than in the previous test.
Table 4. Reactiona of Methylamine with n-Butanol at 190°C over 2.45 g of the Cu/2nO 30/70 mol%
Catalyst
a±0.0002 mol/hr
jJ
G)
:r
m
jJ
s::l>
z
~
r
:;0:;
-<
r
~
-t
5
z
o"TI
~
Table 5. Reactiona of Isopropylamine with Isopropanol and Disproportionation of Isopropylamine ~
over Cu/ZnO and Pd/Si02 Catalysts at 75 atm z
o
9
:::c
m
C t I t Temperature Isopropylamine Isopropanol GHSV Total Amine Selectivity to % Carbon -t
a a ys (OC) (mol/hr) (mol/hr) (hr- l ) Conversion (%) Diisopropylamine Balance m
:0
o
~
-t
Cu/ZnO 190 0.0064 0.0086 4700 70.3 100 100.7 o
~
(')
Cu/ZnO 190 0.0071 0 4700 22.5 54 101.7 o
~
Cu/ZnO 159 0.0066 0.0083 4700 54.5 100 102.0 "'0
o
C
Cu/ZnO 159 0.0071 0 4700 2.8 50 100.0 z
o
Pd/Si02 190 0.0057 0.0094 1500 56.1 61 99.3 en
a±O.OOOl mor/hr
~
(J'I
452 R.G.HERMAN
The same reactants were used with a 4 g (10 ml) portion of the
Pd/Si02 catalyst at 190°C in flowing nitrogen. The molar ratio of
the ethano1/isopropy1amine reactants was 1.45/1.00, and the liquid
injection rate was again 20 ~l/min. Table 6 shows that the isopro-
py1amine was transformed into a number of products, with a 34% selec-
tivity to the ethy1isopropy1amine alkylation product and a 34% selec-
tivity toward the disproportionation product, diisopropylamine. Al-
though the carbon balance was satisfactory, the nitrogen balance was
only 89% that of the determined inlet feed. This appears to be caused
by a low determination of the ammonia content in the exit gas and
would reduce the above reported selectivities somewhat. In any case,
it seems that the trans-a1kylation/disproportionation sequence is
also occurring to a small extent over this catalyst. However, the
disproportionation of isopropy1amine is clearly the significant side-
reaction pathway over the Pd/Si02 catalyst.
DISCUSSION
Table 6. Reactiona of Isopropylamine with Ethanol at 190°C and 75 atm over Cu/ZnO and Pd/Si02
Catalysts
a-±O-:-bOOl- moT/hr
bi-PrNHEt = ethylisopropylamine
C(i-Pr)2NH = diisopropylamine
dEt2NH = diethylamine
':JJ
Gl
:c
m
':JJ
s:
:t>
z
ALKYLATION OF N- AND O-HETEROATOM COMPOUNDS 455
ano1 from synthesis gas can also be used for the alkylation of amines
with alcohols. These include both the Cu/ZnO based catalysts [53,
54,57-60] and the Pd-based catalysts [55,57,61,62]. In addition,
the following features have been established.
(1) Unsupported Cu/ZnO can be successfully utilized as an amine
alkylation catalyst,
(2) Alkylation of a primary amine by a primary or secondary al-
cohol occurs at fast rates, high selectivities, and mild
conditions over the Cu/ZnO catalyst,
(3) Alkylation of a primary amine by a primary or secondary al-
cohol occurs at slower rates and with poorer selectivities
over the Pd/Si02 catalyst than were obtainable with the
Cu/ZnO,
(4) Disproportionation of the primary amine to the secondary
amine and ammonia occurs over these catalysts in the absence
of alcohols,
(5) Disproportionation of the primary amine occurs to a greater
extent over the Pd/Si02 catalyst than over the Cu/ZnO cata-
lyst, and, in fact, this reaction is inhibited over the
latter catalyst by the presence of alcohols,
(6) The extent to which the disproportionation reaction proceeds
can be suppressed by lowering the temperature,
(7) Hydrogen is not a necessary component of the reactant stream,
but its presence can enhance the net yield of alky1ated
amine,
(8) An appreciable excess of the reactant amine is not necessary
for high selectivity toward the monoa1ky1ated product to be
achieved, and
(9) Utilization of ethanol as the feedstock alcohol in the al-
kylation of an amine substituted with a secondary alkyl
group tends to lead to a small degree of trans-alkylation.
As can be noted from Tables 1-3, the catalysts that can be used
for the alkylation of amines are typically supported on alumina or
silica. The Cu/ZnO catalyst utilized in the present work can be
regarded as an unsupported catalyst in comparison with the Pd/Si02
catalyst. It will become evident in this discussion that the pres-
ence of catalyst supports is a predominant factor in the occurrence
of side-reactions that compete with the desired alkylation reaction.
Therefore, if a support is necessary for the physical stability of a
particular catalyst, the support should be chosen or treated so that
side-reactions are minimized. Recommendations for this will be given.
H H H R
,I I"
N.':".R N.-:-.H (2l)
7777777777777777777777
involve an ammonium ion species. Since SiD2 is a more acidic mate-
rial than is ZnD, the trend in this work is that the more acidic the
catalyst, the greater is the degree of disproportionation of primary
amines. The disproportionation reaction over Pd/SiD2 is somewhat
suppressed by the presence of alcohols in the reactant stream, while
the reaction is inhibited over the Cu/ZnD catalyst under the same
conditions. This might reflect the degree of competition between the
alcohol and the primary amine for reactive surface sites on the ma-
terials. The experimental results indicate that alcohols are more
successful competitors over the more basic Cu/ZnD catalyst than they
are over the more acidic Pd/SiD2 catalyst.
NH3 -46.11 65
CH3NH2 -22.97 65
( CH3)2 NH -18.45 65
CH3CH2NH2 -47.15 65
(CH3CH2>2NH -71.42 65
Two Cu/ZnD catalysts that have been used for the hydrogen-free
alkylation of isopropy1amine with ethanol were examined by X-ray
photoelectron spectroscopy (XPS). In neither case was nitrogen de-
tected on the surface of the catalysts. However, the previously ob-
served deactivation of copper-based catalysts in the absence of hy-
drogen has been postulated to be caused by surface nitride formation,
e.g. as represented by equation 22 [66]. The ammonia, in this case,
458 R. G. HERMAN
(22)
of the two reactants could be coupled, e.g. the alcohol could form
an aldehydic surface species and the released hydrogen atoms could
then be utilized by the other reactant. This discussion indicates
a need for further mechanistic research to be carried out. We are
currently studying the reaction sequence and intermediate species
involved in the alkylation of ammonia and amines with alcohols and
with synthesis gas.
ACKNOWLEDGEMENTS
REFERENCES
67. J. M. Pommersheim and J. Coull, Am. Inst. Chem. Eng. J., 17:1075
(1971) •
68. P. Wiseman, "An Introduction to Industrial Organic Chemistry,"
Applied Science Publ. Ltd., London, 391 (1972).
APPENDIX 1.
463
~
Ol
Appendix 1. U.S. Energy Conversion Factors ~
1 million barrels
of oil a 1.000 0.227 0.251 5.680 1.700 5.800
1 thousand barrels
of oil per day
(for one year)a 0.365 0.083 0.091 2.073 0.620 2.117
1 million metric
tons of coal b 4.398 1.000 1.102 24.974 7.474 25.51
1 million short
tons of coal b 3.991 0.907 1.000 22.664 6.783 23.15
1 billion cubic feet
of natural gas C 0.176 0.037 0.041 1.000 0.299 1.021
1 billion kilowatt
hours 0.588 0.123 0.136 3.342 1.000 3.412
1 trillion Btu
(.001 quad) 0.172 0.039 0.043 0.979 0.293 1.000
1 thousand trillion
calories 0.684 0.143 0.158 3.886 1.163 3.968
Chemical Nomenclature
465
466 APPENDIX 2
Acetaldehyde Ethana1
Acetaldehyde dimethylacetal l,l-Dimethoxyethane
Acetic acid Ethanoic acid
Acetone Propanone
Acrolein or Acrylaldehyde 2-Propen-l-al
Acrylonitrile Propenonitrile
Benzil or Diphenyldiketone 1, 2-Diphenylethan-l,2-dione
Biacetal 2,3-Butanedione
Butyraldehyde Butanal
Caprolactam or 1,6-Hexanolactam 6-Hexanelactam
Chloroform Trichloromethane
Crotonaldehyde 2-buten-l-al
Cumene Isopropylbenzene
Diacetone alcohol 4-Hydroxy-4-methylpentan-2-one
APPENDIX 2 467
Diethanolamine Bis(2-hydroxyethyl)amine
Dimethyl terephthalate Dimethyl-l,4-benzene-
dicarboxylate
Ethylbenzene Phenyl ethane
Ethylene glycol l,2-Ethanediol
Formaldehyde Methanal
Formamide Methanamide
Formic acid Methanoic acid
Glyceraldehyde 2,3-Dihydroxypropanal
Glycerine or Glycerol l,2,3-Propane-triol
Glycolaldehyde or 2-Hydroxyethanal
Glycolic aldehyde or
Hydroxyacetaldehyde
Glyoxal Ethanedial
Hemiacetal l-Methoxyethanol
Hexamethylenediamine l,6-Diaminohexane
Isobutyryl chloride or 2-Methylpropanoyl chloride
Isopropyl carbonyl chloride
Isopropyl alcohol 2-Propanol
2,5-Lutidine or 2,5-Dimethylazine
2,5-Dimethylpyridine
Maleic anhydride cis-Butendioic anhydride
Methacrylic acid 2-Methyl-2-Propenoic acid
Methyl acetate l-Methoxyethanal
Methyl acetoacetate l-Methoxybutan-l,3-dione
Methylethylketone Butanone
Pentaerythritol Tetrakis(hydroxymethyl)methane
Phthalic anhydride l,2-Benzene-dicarboxylic
anhydride
2-Picoline or a-Picoline or 2-methylazine
2-Methylpyridine
Propylene Propene
Stilbene l,2-Diphenylethene
Styrene Phenylethene
Vinyl acetate Ethenylethanoate
Vinyl chloride Chloroethene
o-Xylene l,2-Dimethylbenzene
p-Xylene l,4-Dimethylbenzene
INDEX
469
470 INDEX