Note: Descriptions are shown in the official language in which they were submitted.
--1--
IMPROVED CATALYTIC SYSTEM FOR EPOXIDATION OF
ALR~NES EMPLOYING LOW SODIUM_CATA~YST SUPPORTS
Technical Field:
The present invention i5 directed to an impro~ed
system ~or the preparation of alkene oxide from al-
i5 kene and an oxygen-containin~ gas employing a sup-
ported silver catalyst. More particularly, the pre-
sent inven~ion relates to a direct oxidation of al-
kene to the corresponding epoxide in which both effi-
ciency of the catalytic system and stability o~ the
catalyst are enhanced by combination of a gaseous
member of a redox-half reaction pair present in the
gaseous mixture of oxygen and alkene, a salt of a
member of a redox-half reaction pair in combination
with the catalyst, and use of a stability-enhancing
5upport.
; ~ Background Ar~:
The production of alkene oxides, or epoxides,
particularly ethylene~oxide by the direct:oxidation
of the corresponding alkene in the presence of a
silver containing catalyst has been known for many
years. For example, the basic process was described
by Lefort in U. S. Patent 1,998,878 and by Van Peski
in U. S. Patent 2,040,782. The basic reaction pro-
,
13457
~4
.
3~
ceeds, as illustrated for ethylene, according to the
equation:
/o\
2 C~2 CH2 + 2 ~ 2 CH2-CH2 (I)
and production of an unwanted by~product according to
the reaction:
CH2=CH2 + 3 2 > 2 CO2 + 2 H20 (II)
or by further oxidation of the epoxide.
In the years between the Van Peski patent and
the present inventions, research efforts have been
lS directed to improving both the activity and longevity
or useful life of the catalyst and the efficiency of
the overall catalytic reaction. As is indicated by
reactions I and II, the oxidation of an alkene may
produce either the alkene oxide (I) sought in the
process or the by-products CO2 and H2O.
Several terms are commonly used to describe some
of the parameters of the catalytic system. For in-
stance, "conversion" has been defined as the percent-
age of alk~ fed to the reactor which undergoes
reaction. The "efficiency" or, as it is sometimes
called, the "selectivity" of the overall process is
an indication of the proportion, usually represented
by a percentage, of the converted material or product
which is alkene oxide. The commercial success of a
30 reaction system depends in large measure on the effi-
ciency of the system. At present, maximum efficien-
cies in commercial production of ethylene oxide by
epoxidation are in the low 80s, e.g., 80 or 81 per-
cent. Even a very small increase in efficiency will
- 35 provide substantial cost benefits in large-scale
13457
-3- ~293~9~
operation. For example, taking 100,000 metric tons
as a typical yearly yield for a conventional ethylene
oxide plant and assuming 80 percent conversion, an
increase in efficiency of from 80 to 84 percent, all
other things being equal, would result in a savings
of 3790 metric tons of ethylene per year. In addi-
tion, the heat of reaction for reaction II ~formation
of carbon dioxide) is much greater than that of reac-
tion I (formation of ethylene oxide) so heat-removal
problems are more burdensome as the efficiency de-
creases. Furthermore, as the efficiency decreases,
there is the potential for a greater amount of impur-
ities to be present in the reactor effluent which can
complicate separation of the desired alkene oxide
product. It would be desirable, therefore, to de-
velop a process for the epoxidation of alkene in
which the efficiency is greater than that obtained in
conventional commercial processes, e.g., with ethy-
lene, efficiencies of 84 percent or greater, while
maintaining other performance characteristics, parti-
cularly the activity, as described below, in a satis-
factory range.
The product of the efficiency and the conversion
is equal to the yield, or the percentage of the al-
kene fed that is converted into the correspondingoxide.
The "activity" of the catalyst is a term used to
indicate the amount of alkene oxide contained in the
outlet stream of the reactor relative to tha~ in the
inlet stream. Activity i5 generally expressed in
terms of pounds of alkene oxide produced per cubic
foot o~ catalyst per hour at specified reaction con-
ditions and rate of feed. The activity may also be
stated in terms of the amount of ethylene oxide in
35 the outlet stream or the difference between the ethy-
13457
-4- 1293~9~
lene oxide content of the inlet and outlet streams.
If the activity of a reaction system is low,
then, all other things being equal, the commercial
value of that system will be low. The lower the
activity of a reaction system, the less product pro-
duced in a unit time for a given feed rate, reactor
temperature, catalyst, surface area, etcetera. A low
activity can render even a high efficiency process
com~ercially impractical. For production of ethylene
oxide, an activity below 4 pounds of ethylene oxide
per hour per cubic foot of catalyst is unacceptable
for commercial practice. The activity is preferably
greater than 8 pounds, and in some instances an acti-
vity greater than 11 pounds of alkene oxide per hour
lS per cubic foot of catalyst is desired.
In some instances, activity is measured over a
period of time in terms of the amount of alkene oxide
produced at a specified constant temperature. Alter-
natively, activity may be measured as a function of
the temperature required to sustain production of a
specified constant amount of alkene oxide. Plots of
such measurements yield "aging rates" which reflect
the stability or useful life of the catalyst. The
useful life of a reaction system is the length of
time that reactants can be passed through the reac-
tion system during which acceptable activity is ob-
served. The area under a plot of activity versus
time is equal to the number of pounds of alkene oxide
produced during the useful life of the catalyst per
cubic foot of catalyst. The greater the area under
such a plot, the more valuable the process is since
regeneration or replacement of the catalyst involves
a number of expenses, sometimes referred to as turn-
around costs. The rate at which activity decreases,
i.e., the rate of deactivation at a given point in
13457
~ _5~ 3~
time, can be represented by the slope of the activity
plot, i.e., the derivative of activity with respect
to time:
s deactivation = d[activity~/dt.
The average rate of deactivation over a period
of time can be represented then by the change in
activity divided by the time period:
average deactivation = a activity/ a t.
At some point, the activity decreases to an
unacceptable level, for example, the temperature
required to maintain the activity of the system be-
comes unacceptably hi~h or the rate of production
becomes unacceptably low. At this point, the cata-
lyst must either be regenerated or replaced. Some of
these definitions may be represented as set out be-
~0 low:
Conversion = moles alkene reacted x 1~0moles alkene fed
~ Efficiency = moles alkene oxide produced x 100
moles alkene reacted
Typically, in commercial production, since the
outlet or~effluent stream emanating from the reactor
may contain substantial amounts of unreacted alkene,
the effluent stream is recycled and combined with the
feedstream after removal of at least a portion of the
alkene oxide. Generally, as the activity of a cata-
lyst decreases with time, in order to obtain the same
ultimate yield of epoxide product, the effluent
13457
-6~ 9~
stream must either be recycled a greater number of
times or the temperature within the reactor must be
raised to increase the activity of the catalyst. The
former approach to increasing the yield of product
requires additional energy expenditures and the lat-
ter, which is most frequently used~ causes faster
catalyst deterioration.
As used herein, an activity-reaucing compound
refers to a compound which, when present in an acti-
vity-reducing amoun~, causes a reduction in activity,
some or all of which activity may subsequen~ly be
regained by returning to a situation in which the
concentration of the compound is below the minimum
actLvity-reducing amount. The minimum activity-
reducing amount varies depending on the particularsystem, the feedstream and the activity-reducing
compound.
Conversely, deactivation, as used herein, refers
to a permanent loss of activity, i~e., a decrease in
activity which cannot be recovered. As noted above,
activity can be increased by raising the temperature,
but the need to operate at a higher temperature to
maintain a particular activity is representative of
deactivation. Furthermore, catalysts tend to deacti-
vate more rapidly when reaction is carried out athigher temperatures.
In contrast to problems associated with low or
decreasing catalyst activities, less than satisfac-
tory efficiencies result in loss of starting mater-
ial, the alkene, as the unwanted product CO2. Ulti-
mately, this also increases product costs.
To be considered satisfactory, a catalyst must
not only have a sufficient activity and the catalytic
system provide an acceptable efficiency, but the
catalyst must also demonstrate a minimum useful life
13457
-7~
or stability. When a catalyst is spent, typically
the reactor must be shut down and partially disman-
tled to remove the spent catalyst. This results in
losses in time and productivity. In addition, the
S catalyst must be replaced and the si`,ver salvaged or,
where possible, regenerated. Even when a catalyst is
capable of regeneration in situ, generally production
must be halted for some period of time. ~t best,
replacement or regeneration of catalyst requires
additional los~es in time to treat the spent catal~st
and, at worst, requires replacement of the catalyst
with the associated costs.
Since even small improvements in activity, ef-
ficiency or useful life may have significance in
large scale commercial production, such improvements
have been the object of a great deal of research in
the direct epoxidation of alkenesO The focus Oe
attempts to improve performance, such as the activity
and useful life of the catalyst and the efficiency of
the system, has included such areas as feedstream
additives or removal of components therefrom; methods
o~ preparation of the catalyst; deposition or impreg-
nation of a particular type or form of silver; com-
position, rormation, physical properties and morpho-
logy of the support; additives deposited on or im-
pregnated in the support; shape of support aggregates
used in the reactor; and various types o~ reactors
and bed designs, such as stationary and fluidized
beds.
Early work on the silver-catalyzed direct oxida-
tion of alkenes to alkene oxides in many instances
resulted in improvements in activity and particularly
the selectively of the system, in many cases the
efficiency increasing by several percent. However,
recent modifications in such systems have resulted in
13~57
~ -8- ~3496
only small incremental improvements in efficiency.
In terms of operating cos~s, even fractions of a
percent improvement in efficiency can translate into
large savings in production. Accordingly, current
research is still being directed to improvements in
the activity and useful life of the catalyst and
selectivi~y of the system.
Although a vast number o~ elements and compounds
are known to have effective catalytic properties in
1~ various reactions, many have at least one shortcom-
ing, such as very high cost and/or limited availabil-
ity, thermal instability in the temperature range in
which the reaction is to be conducted, low mechanical
strength, small surface area per unit of volume,
susceptibility to poisoning, short useful lifetime,
etcetera. Such undesirable characteristics make such
substances of limited utility as catalys~s. Some of
these shortcomings, however, may be overcome and in
some instances the effectiveness of the catalyst may
be improved by applying the substance to a carrier or
support.
New support materials are continuously being
tried. ~owever, many of those which were employed in
the early development of the silver-bearing catalysts
are, with some modifications, still being used.
Materials which have found most widespread use are
typically inorganic and generally are of a mineral
nature. Such materials commonly include alumina,
fire brick, clay, bauxite~ bentoniter kieselguhr,
carbon, silicates, silica, silicon carbide, zirconia,
diatomaceous earth, and pumice.
In addition to the physical strength of the
support materials, other physical properties, such as
surface area, pore volume, pore dimensions, and par-
ticle size have drawn considerable attention. These
13~57
. ~ ~
3~ 3~
properties have been examined with great scrutinywhen evidence indicated that there was a correlation
between the size o~ silver particles and the effici-
ency of the overall system or useful life of the
catalyst. Some materials are also preferred for
their chemical properties, i.e., their "inertness" or
"promoting" properties.
The support serves a number o~ functions in a
heterogeneous catalytic system. Ease of handling is
facilitated by a support which generally takes the
form of discrete particles or aggrega~es of varying
shape or size which, depending on usage, have a major
dimension of about 1 millimeter to about 20 milli-
meters~ Thus it is not necessary for the catalyst to
form a permanent or semi-permanent part of the reac-
tor.
The support, however, serves primarily to in-
crease the surface area of the "active" component of
the catalyst, silver, which is important in that most
epoxidation occurs at the silver surface-fluid inter-
face. Many of the substances commonly employed as
catalyst supports not only have the usual external
surface, which provides a varying surface area, de-
pending on the shape of the support bodies and the
packing of the bodies, but are also of a porous na-
ture and, therefore, have a large internal surface
which contributes to the overall surface area of the
supported catalyst. Such support materials provide a
greater capacity for sorbing not only the catalyst
material during catalyst preparation, when the sup-
- port is impregnated with a solution containing the
catalyst component(s) in soluble form, but also a
greater capacity for the flow of the fluid reactants
within the catalyst during the reaction for which the
catalyst is intended. The support also improves
13457
"` lo 1~3~
performance by lo~ering the pressure drop through the
reactor and by facilitating heat and mass transfer.
Among the large variety of substances employed
in the past as supports for catalytic materials,
alumina has exhibited superiority in many respects as
a catalyst support material. In addition to the low
cost of the material, alumina has good thermal sta-
bility and some forms have a relatively large surface
area.
Alumina, in its various forms, particularly
alpha-alumina, has been preferred as a support mater-
ial for silver-containing catalysts in the prepara-
tion of alkene oxides. Numerous variations of sur-
face area, pore dimensions, pore volume and particle
size have been suggested as providing the ideal phys-
ical property or combination of properties for im-
proving efficiency, activity or useful life of the
catalyst.
Holler (U. S. Paten~ 3,908,002) discloses an al-
pha-alumina, useful as a catalyst support for reac-
tions conducted at temperatures below 800 degrees C,
such as oxidation reactions of hydrocarbons to oxyhy-
drocarbons. The support, having a surface area re-
ported to be at least about 4~ m2/g, is produced by
thermally decomposing a porous aluminum ion chain-
bridged, polymeric carboxylate. Indicating that a
large surface area in a carrier may be detrimental to
its efficient operation and catalyst activity, Belon
(U. S. Patent 3,172,866) describes a method of pro-
ducing a macroporous catalyst carrier which may beused in the catalytic production of ethylene oxide
having pore diameters of between 0.1 and 8.0 microns
and a specific surface area between a few square
meters and one square decimeter per gram. The sup-
port is prepared by heating a mixture of active and
.
~ 13457
3~
calcined aluminum oxides and a small amount of boronoxide at temperatures of between about 1,600 and
1,800 degrees C. Waterman ~U. SO Patent 2,901,441)
describes a process for preparing highly active and
selective catalysts for the oxidation of olefins to
olefin oxides on a support having an average porosity
of at least 35 percent. The method involves washing
an alpha-alumina or silicon carbide support having an
average porosity of between 35 and 65 percent with an
aqueous solution of lactic acid, washing with water
until neutral, and then impregnating the support with
an aqueous solution of silver lactate. The impreg-
nated support is thereafter heat-treated to deposit
elemental sllver. A silver-supported catalyst for
the vapor phase oxidation of ethylene to ethylene
oxide, exhibiting improved production of ethylene
oxide and catalyst longevity, is described by Brown
et al (U. S. Patent 3,725,307). The catalyst is
disclosed as being formed from support particles
having an average pore diameter of at least 10 mi-
crons up to, preferably, 70 microns and a surface
area of less than about l m2/g. The selectivities
reported do not range above about 73 percent. The
support is preferably composed of silica-alumina. A
silver-supported catalyst which includes a support of
alpha-alumina, silicon carbide, fused aluminum oxide,
or mixtures of alumina and silica was asserted by
DeMaio (U. S. Patent 3,664,970) to eliminate the need
for halogenated inhibitors in the oxidation of ethy-
lene to ethylene oxide. The support is composed ofparticles having a minimum apparent porosity of about
30 percent and wherein at least 90 percent o the
pores have diameters in the range of l to 30 microns,
the average of the diameters being in the range of 4
to 10 microns. Wattimena ~U. S. Patent 3,563,914)
13457
-12~
discloses silver catalysts using aluminum oxide sup-
ports having pore volumes between 15 and 30 ml/g and
surface areas below about 10 m2/g.
~ayden et al tU. K. Patent Application
2,014,133) disclose a silver catalyst employing a
support having a specific surface area in the range
of 0.05 to 10 m2/g, an apparent porosity of at least
20 percent, and mean pore diameters of 0.1 to 20
microns, the pore size distribution being bimodal, in
which the smaller pores preferably account for at
least 70 percent of the total pore volume. Alpha-
alumina supports are described by Rashkin (U. K.
Patent Application 2,122,913A) having a "relatively
low surface area" of less than 30 m2/g. Mitsuhata et
al (Japanese Published Patent Application 56-089843)
and ~itsuhata et al ~U. S. Patent 4,368,144) describe
supported silver catalysts in which the support is
formed from alpha-alumina having a specific surface
area of 0.5 to 5 m2/g. Watanabe et al (Japanese
Published Patent Application 56-105750) employ a
similar catalyst suppor having a surface area of 1
to 5 m2/g. Hayden et al (U. S. Patent 4,007,135)
describe silver-containing catalysts in which the
porous neat-resisting support has a specific surface
area in the range of 0.04 to 10 m2/g, an apparent
porosity of at least 20 percent, and a median pore
diameter of 0.3 to lS microns. Mitsuhata et al (U.
S. Patent 4,248,740) describe the use of high alpha-
alumina content supports having a specific surface
area of not more than 10 m2/g, an apparent porosity
of 40 to 60 percent by volume, and a pore volume of
0.1 to 0.5 cc~g. Armstrong et al (U. S. Patent
4,342,667) disclose a supported silver catalyst,
useful in the oxidation of ethylene to ethylene ox-
3s ide, in which the support has a surface area of 0.02
13457
-13- ~ ~3~
to 2 m~/g, an average pore diameter of 0.5 to 50
microns and an average pore volume of 0.2 to 0 5
cc/g.
There has also been some interes~ in the purity
of supports employed, both as to composition and
phase. Examples of high purity alumina include U. S.
Patent 2,901,441 which uses alpha-alumina having a
purity of about g9.5 percent as a support for cata-
lysts used to oxidize olefins to olefin oxides. An
ethylene oxidation catalyst is disclosed in German
Patent Publication DE 2,933,950 which attains a long
catalyst life without a loss in activity or selec-
tivity by using an alpha-alumina support having less
than 0.001 weight percent of alkali-soluble silicon
compounds. The catalyst is prepared by boiling com-
mercial quality alpha-alumina with l weight percent
sodium hydroxide solution and washing to a pH value
of 8. If desired, the silicon compound concentration
may be reduced below l part per million (ppm) by
washing further with l weight percent HF. U. K.
Patent Application 2,122,913A describes supported
silver catalysts in which the support is composed of
silica, alumina or mixtures thereof, one example of
which is an alumina having a purity of 9~.3 percent
by weight. The silver-supported catalyst described
in Japanese Published Patent Application 56-089843
employs an alpha-alumina carrier having a sodium
content of less than 0.07 weight percent. Japanese
Published Patent Application 56-105750 describes the
use of an alpha-alumina support in conjunction with a
silver catalyst for producing ethylene oxide, which
support has a sodium content less than 0.07 weight
percent. A silver catalyst including an alpha-
alumina carrier having a sodium content of not more
~ 35 than 0.07 percent is described by Mitsuhata et al
':
13457
~ " ~
-14- ~ 3~
~U. S. Patent 4,368,144). The support also has a
surface area within the range of 0.5 to 5 m2/g, an
apparent porosity of 25 to 60 percen~, a specific
pore volume of 0.2 to 0.5 ccjg, and a particle dia-
meter within the range of 3 to 20 mm. An alpha-
alumina support having a purity of 98~ weight per-
cent, for use with silver in the catalytic oxidation
of ethylene, is described by Warner et al in U. 5.
Paten~ 4,455,392. The patent additionally discloses
that the carrier is generally a conventional micro-
porous support with surface areas of less than 10
m2~g, pore volumes ranging from about 0.15 to 0.8
cc/g, and pore diameters of about 0.1 to 100 microns.
In addition to compositional purity, both phase
lS purity and morphology of the support have been areas
in which improvements in eEficiency, selectivity or
stability of the catalyst have been sought. Examples
include U. S. Patent 2,901,441 in which aluminum
oxide is substantially completely converted to the
alpha form of alumina by heating aluminum oxide to a
temperature of about 1,500 to 2,050 degrees C. Weiss
(U. S. Patent 2,209,908) and Carter tU. S. Patent
2,294,383) d-escribe the use of "Tabular Corundum" as
: . a catalyst support for metallic oxides, such as those
oxides of metals selected from the fifth and sixth
group of the periodic system, for example, vanadium,
molybdenum, uranium, etcetera, in the oxidation of
various organic materials to maleic acid and maleic
anhydride and silver for the catalytic oxidation of
~` 30 ethylene to ethylene oxide, respectively. Weiss
indicates that Tabular Corundum, which is almost
entirely aluminum oxide and has the alpha-corundum
: crystalline form of aluminum oxide, may be formed by
mixing aluminum oxide with one or more of several
compounds, such as sodium oxide and chromic oxide,
: 13457
and heating the mixture to a temperature in the range
of about 800 to about 1,800 degrees C. Tabular Co-
rundum is further described as having impurities
present in only small quantities, the material also
includes "readily bonded surfaces and consisting
essentially of interlocked corundum crystals in tabu-
lar form, having the contained impurities disseminat-
ed in minute globules throughout the crystalline
aluminan. Brengle et al (U. S. Patent 2,709,173)
also employ Tabular Corundum as a support in one of
their examples.
U. S. Patents 4,039,481 and 4,136,063 to Kimura
et al disclose a catalyst carrier and a method for
making same, the catalyst being the type used in
catalytic converters in automobile exhaust systems.
Specifically, the catalysts have a surface layer
containing alpha-phase alumina and an inner portion
consisting essentially of alumina of a phase other
than that of the alpha phase. The pores in the al-
pha-alumina surface layer are larger than those in
the inner portion of the catalyst body. A method of
preparing the phase gradient support particles is
described which provides for treating the surface of
the alumina to a depth o about 400 microns with a
transition element, particularly iron, and thereafter
firing the carrier particles.
Weber et al ~U. S~ Patent 4,379,134) describe
high purity alpha-alumina bodies, at least 85 percent
of the pore volume of the bodies having pores with a
diameter of from 10,000 to 200,000 Angstroms. The
high purity alpha-alumina bodies are prepared by
peptizing boehmite in an acidic aqueous, fluoride
anion-containing mixture. An extrudable mixture is
formed thereby which is extruded and shaped into
formed bodies which are thereafter dried at 100 to
13457
-16~ 36
300 degrees C, calcined at a temperature of from 400
to 700 degrees C to convert the alumina to the gamma
phase, and subsequently calcined further at a temper-
ature of from 1,200 to 1,700 degrees C to convert the
gamma phase to alpha-alumina phase.
A method of producing granulated porous corundum
having a homogeneous porous structure with a total
pore volume of 0.3 to 1.0 cm3/g and a predominant
pore size of 5,000 to 30,000 A is described by Bores-
kov et al (U. S. Patent 3,950,507). The method opreparing the alpha-alumina includes treating active
alumina or aluminum hydroxide having a porous struc-
ture to a first heat treatment in which the tempera-
ture is increased from 20 to 700 degrees C, a second
heat treatment in the range of from 700 to 1,000
degrees C, and a third treatment in the range of from
l,OU0 to 1,400 degrees C. Each of the heat treat-
ments is for a period of at least one-half hour, the
first heat treatment being conducted in an atmosphere
of hydrogen 1uoride in which the alumina absorbs the
hydrogen fluoride and the second heat treatment de-
sorbs the hydrogen fluoride. The patent also de-
scribes a similar procedure employing stationary
thermal conditions in which the granules of alumina
or aluminum hydroxide are impregnated with other
fluorine-containing compounds prior to the first
thermal treatment. The recommended starting mater-
ials used to form alpha-alumina include granulated
pseudo-boehmite, boehmite or bayerite as the granu-
lated aluminum hydroxide and granulated alpha-, eta-,
or theta-alumina as the ac~ive alumina.
Although alpha-alumina has been considered by
most to be the preferred alumina support material,
; Smith e~ al (U. S. Patent 2,422,172) have suggested
that beta-aluminas are more desirable than the alpha
13457
-17- ~ ?3~
phase as a support material for catalysts, particu-
larly those used in catalytic conversion processes
such as dehydrogenation and hydroforming.
In seeking the ideal support material, there has
S been some departure from the commonly employed sub-
stances. For example, some use has been made of
alkali metal and alkaline earth metal carbonates,
both as the sole support material and in combination
with other materials as the carrier for processes
such as direct oxidation of alkenes to epoxides.
A number of supported silver-containing cata-
lysts have been employed for epoxidation of alkenes
in which the carrier includes, sometimes labelled as
a promoter, a carbonate of a metal, generally an
alkali metal or alkaline earth metal. Some examples
of the use of one or more al~ali and/or alkaline
earth carbonates may be found in U. S. Patents
2,424,084, 2,424,086, 2,615,900, 2,713,586,
3,121,099, 3,258,433, 3,563,913, 3,563,914,
3,585,217, 4,007,135, 4,033,9~3, 4,039,561,
- 4,~66,575, 4,094,889, 4,123,385, 4,125,480,
4,168,247, 4,186,1~6, 4,226,782, 4,229,321,
4,324,699, European Patent Publications 0,003,542 and
0,011 .' " Japanese Patents 41-11847 and 57-107242,
U. K. Patents 590,479, 1,571,123 and 2,014,133A, and
Murray, "A Study Of The Oxidation Of Ethylene To
Ethylene Oxide On A Silver Catalyst~, Australian
3Ournal of Scientific Research, Volume 3A, Pages 433-
; 449 (1950). In additionl U. S. Patent 3,332,837
; 30 employs zinc and/or cadmium carbonates, Gelbstein,
(DS 2,352,608) aiscloses the use of the latter car-
bonate and European Patent Publication 0,003,642
mentions the use of molybdenum carbonate.
Several patents have described the use of fluo-
rine-con~aining co~pounds to treat support materials,
13457
-18-
in some cases to provide a compositionally pure sup-
port, and in other cases as a fluxing agent to im-
prove the phase purity of the support. Thus, U. R.
Published Patent Specification 590,479 and U. S.
Patent 2,424,086 indicate that a more active catalyst
is formed if the support material ha~ undergone a
preliminary treatment with a dilute solution of hy-
drofluoric acid prior to impregnation with silver.
U. S. Patent 4,379,134 teaches the preparation of
high purity alpha-alumina bodies by peptizing boeh-
mite alumina in an aqueous acidic mixture containing
fluoride anions and water. German Patent 2,933,950
teaches the reduction of silicon content by treatment
with HF. U. S. Patent 3,950,507 teaches the prepara-
tion of granulated porous corundum by a multiple stepheat treatment in which initial steps may be carried
out in an atmosphere of hydrogen fluoride. Hosoda et
al (U. S. Patent 3,144,416) suggest that a small
amount o~ a halogen compound, sulfur compound, nitro-
gen compound, or phosphorous compound may be addedeither to the reaction gas or the catalyst to improve
the selectivity of the catalyst.
The nature of the silver itself has also been
examined and modified in attempts to impro~e the
ef f iciency and stability of the catalyst. Cavitt (U.
S. Patent 4,229,321) teaches that a supported silver
catalyst of improved selectivity and activity may be
prepared by mechanically removing the outer surface
or skin of the catalyst after the impregnated cata-
lyst has been heated to évaporate volatile materialand reduction of the silver salt to silver metal,
thereby activating the catalyst.
Since the early work on the direct catalytic
oxidation of ethylene to ethylene oxide, workers in
the field have suggested that the addition of certain
13457
-19- ~Z~3~
compounds to the gaseous feedstream or direct incor-
poration of metals or compounds in the catalyst could
enhance or promote the production of ethylene ox-
ide. Such metals or compounds have been known vari-
ously as "anti-catalysts", "promoters" and "inhi-
bitorsn. These substances, which are not themselves
considered catalysts, have been proposed by prior
workers to contribute to the efficiency of the pro-
cess by inhibiting the formation of carbon dioxide or
promoting the production of ethylene oxide. The
scientific literature is replete with examples o the
use of alkali metals and alkaline earth metals and
their cations to promote the efficiency of silver
catalysts used in epoxidation reactions. For exam-
ple, ~odium, potassium and calcium were disclosed asbeing suitable promoters in U. S. Patent 2,177,361.
Numerous examples may be found in literature of pref-
erence for one or several metals or cations and ex-
clusion of one or more metals or cations as promoters
in silver catalysts.
~; Among those anions associated with the cation
used in preparing silver-con~aining catalysts em-
ployed in direct epoxidation reactions that have been
suggested as being suitable include carboxylates, for
example, formate, acetate, malonate, oxalate, lac-
tate, tartrate, and/or citrate, and inorganic salts,
such as carbonates, bicarbonates, phosphates, ni-
trates, and/or nitrites, chlorides, iodides, bro-
mates, and isopropoxides. However, although many
examples may be found in the literature indicating
that such compounds are suitable, numerous patents,
such as U. S. Patents 3,962,136; 4,012,425;
4,066,575; 4,207,210; and 4,471,071, suggest that no
unusual effectiveness, particularly with regard to
catalytic activity, is observed with any particular
13457
~3~
anion of an alkali metal promoter. U. S. Pa~ents
4,007,135; 4,094,889; 4,125,480; 4,226,782;
4,235,757; 4,324,699; 4,342,667; 4,356,312;
4,368,144; and 4,455,392 disclose that potassium
nitrate may be added to the catalyst as a suitable
promoting ma~erial. Potassium nitrat:e may also be
formed in situ when a carrier material is treated
with certain amines in the presence of potassium ions
as, for instance! when silver is introduced to a
1~ carrier material in a silver-impregnating solution
containing an amine and potassium ions, followed by
roasting.
A number of compounds have been proposed in the
literature as additives to the feedstream or
reactants to improve the efficiency of the direct,
silver-catalyzed oxidation of alkenes to alkene
oxides. For example, Law and Chitwood ~U. S. Patene
2,194,602) disclose the use of a "repressantr, i.e.,
anti-catalyst, such as ethylene dichloride, ch~orine,
2~ sulfur chloride, sulfur trioxide, nitrogen dioxide,
or other halogen-containing or acid-forming mater-
ials. Numerous additional anti-ca~alysts are pre-
sented by the same patentees in U. S. Patent
2,279,469. The anti-catalysts, broadly listed in
categories such as halogens and compounds containing
halogen, hydrocarbons, compounds containing carbon,
hydrogen and oxygen, compounds containing sulfur, and
compounds containing nitrogen are represented and, in
addition to those compounds already mentioned above,
additional representative compounds include, as ni-
trogen-containing compounds, nitric oxide, ammonia,
amines such as ethylenediamine, diphenylamine and
analine, nitro compounds such as o-nitroanisole and
o-nitrotoluene as organic oxygen-containing organic
compounds, alcohols such as methyl, ethyl and iso-
propyl alcohols~ ethers such as isopropyl and dibutyl
ethers, as well as glycol ekhers, ketones such as
methyl ethyl ketone and acetone, as hydrocarbons such
as benzene, and N-hexane; sulfur compounds such as
sulfur dioxide, hydrogen sulfide and diethylsulfide;
chlorine-containing compounds such as carbon tetra-
chloride, chlorobenzene and dichloroethyl ether.
Berl (U. S. Patent 2,270,780) lists a number of com-
pounds as anti-detonating or anti-knock materials to
control the oxidation of ethylene and propylene to
their oxides. Oisclosures of other feedstream addi-
tives used in the production of alkene oxides, parti-
cularly halogen compounds, may be found in U. S.
Patents 2,279,470; 2,799,687; 3,144,416; 4,007,135;
4,206,128; and 4,368,144. In addition, EPO Patent
0,003,642 and U. R. Patent Application 2,014,133A
disclose processes for the production of olefin ox-
ides employing silver-containing catalysts in which a
chlorine-containing reaction modiier and a nitrite
or nitrite-forming substance are described. Rumanian
Patent 53,012, published December 2, 1971, discloses
a direct, silver-catalyzed direct epoxidation proce-
dure which employs oxides of nitrogen in the feed-
stream. U. K. Patent 524,007 includes ethylene di-
chloride or nitrogen dioxide in the feedstream of a
silver-catalyzed epoxidation procedure.
Although much of the art discussed above has
resulted in improvements in the efficiency, activity
or stability of the catalytic system, many of the
13457
-22- lZ~3~
improvements have individually been rather slight.
In some of the catalytic systems, gains in one of
these performance parameters have been frequently
offset by losses in another; that is, enhancement of
one index of performance has been accompanied by a
deleterious effect on another of the indices. For
example, if a reaction system is designed which has a
very short useful life, the system may be commer-
cially impractical even though the efficiency and
initial activity o the catalyst are outstanding.
Accordingly, a system that provides an increase in
the efficiency of the overall catalytic reaction
system, while only minimally affecting the activity
and useful life of the catalyst, or perhaps increas-
ing one o these performance indices, would be parti-
cularly bene~icial.
The presence of leachable sodium in a silver
catalyst employed for epoxidation tends, in some
instances, to improve the efficiency of the system
~20 under epoxidation conditions generally used. In the
;presence of CO2 and certain efficiency-enhancing
compounds, however, sodium exhibits deactivating and
effective life-shortening effects on epoxidation
ca~,ysts and systems. The loss of stability or
life-shortening effect may be so marked, particularly
when the sodium is added as a promoter, that any
gains in efEiciency provided by the sodium are lost
several days after the catalyst is placed in ser-
vice. In many commercially used epoxidation reactors
the effluent stream contains relatively high propor-
tions of unreacted alkene. To diminish losses of the
alkene, the effluent stream is recycled and i~tro-
duced to the reactor with the feedstream. Since
efficiency in such reactions never reaches 100 per-
cent, the effluent stream al~ays contains some carbon
13457
~ -23~ 3~r~
dioxide. In a reactor in which the effluent stream
is recycled to the reactor, therefore, the feedstream
always contains some carbon dioxide which, in combin-
ation with sodium and certain efficiency-enhancing
compounds, results in losses of activity and stabil-
i~y of the catalyst and system. Commonly, the carbon
dioxide is removed by a scrubbing device, such as a
Benfield scrubber, placed in the effluent stream
between the effluent outtet and the reactor inlet.
Such devices require, however, a substantial capital
expenditure and additional plant space. The catalyst
and process of the present invention diminish the
deactivating and life-shortening effects of CO2.
Disclosure Of The Invention:
The present invention is directed to a high
~; performance catalytic process for epoxidation of
alkene in the presence of an oxygen-containin~ gas
which comprises contacting the alkene and the oxygen-
containing gas under epoxidation conditions in the
; presence of at least one efficiency-enhancing gaseous
member of a redox-half reaction pair and a solld
catalyst. The catalyst contemplated by ~ present
; 25 invention comprises a catalytically-effective amount
of silver on a solid support and an efficiency-
enhancing amount of at least one efficiency-enhancing
salt of a member of a redox-half reaction pair. The
support has less than about 50 and most frequently
less than about 20 parts per million, by weightj of
leachable sodium.
The invention, described in greater detail be-
low, is also directed to a catalyst suitable for use
in a process in which an epoxide is formed from an
alkene. The catalyst of the present invention, as
13457
-24- ~ 3~9~
suggested above, comprises a catalytically-effective
amount of silver on 3 substantially inert support and
an efficiency-enhancing amount of at least one effi-
ciency-enhancing salt of a member of a redox-half
reaction pair. The support contains less than about
50 and most frequen~ly less than about 20 parts per
million, by weight, of leachable sodium.
The present invention provides a catalytic sys-
tem which includes a stable catalyst even when used
with recycled effluent streams containing carbon
dioxide. In some aspects, the present invention
allows a collateral benefit in that it obviates the
need for a costly scrubbing unit in the effluent
stream while still maintaining desired efficiencies
and activities ~or useful periods of time.
In additlon to providing an enhanced stability,
the present invention also furnishes an additional
benefit of enhanced efficiency.
2~ Brief Descri~on Of The Drawings-
Figure 1 is a graphical comparison of aging
rates or plo~s of activity (percent ethy~ene oxide in
the outlet) of a catalyst according to the present
invention having low sodium and one having 80 ppm
sodium;
Figure 2 is a graphical comparison similar to
Figure 1 comparing a catalyst having 40 and 100 ppm
of sodium, respectively, and a feedstream having 3
percent CO2; and
Figure 3 is a graphicaI comparison of the cata-
lysts and conditions o~ Figure 2 in which effi~ien-
cies are compared.
:,
13457
-25- ~Z~3~
Detailed DescriPtion Of The Invention:
The present invention is directed to a process
for the vapor phase oxida~ion of alkenes to alkene
5 oxides, i.e., an epoxidation process, in the presence
of an oxygen-containing gas and to the silver cata-
lysts employed therein.
The process and catalyst of the present inven-
tion are useful in the epoxidation of the alkenes
10 ethylene and propylene, the epoxides of which are in
great demand for use as intermediates in producing
such materials as polymers, surfactants, synthetic
fibers and antifreeze. However, the present inven-
tion is not limited to these compounds but may be
used to oxidize cyclic and acyclic alkenes which are
in the gaseous state or have signiicant vapor pres-
sures-under epoxidation conditions. Typically these
compounds are characterized as having on the order of
12 carbon atoms or less which are gaseous under epox-
idation conditions. In addition to ethylene andpropylene, examples of alkenes which may be used in
the present invention include such compounds as bu-
tene, dodecene, cyclohexene, 4-vinylcyclohexene, sty-
rene and norbornene.
Supeort:
The support material used in the present inven-
tion may be any solid! porous, reractory material
which can withstand the roasting temperatures, if
that is the method employed to reduce the silver to
its free metallic state, and which can also withstand
the temperatures employed within the reactor during
the epoxidation process. Regardless of the method
used to reduce the silver, the support should also be
13457
~ -26-
~3~316
able to withstand the temperatures employed within
the reactor under epoxidation process conditions.
The support should not have any undue deleterious
effect on the performance of the system. Examples of
- S suitable materials include magnesia~ zirconia, sili-
ca, silicon carbide, and alumina, preferably alpha-
alumina.
The carrier materials of the present invention
may generally be described as porous or microporous,
10 having median pore diameters of about 0.01 to about
100 microns, preferably about 0.5 to about 50 mi-
crons, and most preferably about 1 to about 5 mi-
crons. Generally, they have pore volumes of about
0.6 to about 1.4 cc/g, preferably about 0.8 to about
15 1.2 cc/g. Pore vol~lmes may be measured by any con-
ventional technique, such as conventional mercury
porosity or water absorption technique.
Generally improved results have been demon-
strated when the support material is compositionally
pure and also phase pure. By "compositionally pure"
is meant a material which is substantially a single
substance, such as alumina, with only trace impuri-
ties being present. "Phase purity" or like terms
refer to the homogeneity of the support with respect
to its phase. In the present invention, alumina,
having a high or exclusive alpha-phase purity (i.e.,
alpha-alumina), is preferred. Most preferred is a
material composed of at least 98 percent, by weight,
of alpha-alumina. Under some conditions, as when
used in conjunction with the salt and gaseous members
of redox-half reaction pairs, even small amounts of
sodium can adversely affect the activity and useful
life of the catalyst, i.e., have a deactivating ef-
fect on the catalyst. Improved results have been
observed when the support contains leachable sodium
13457
-27~
levels less than about 50 parts per million (ppm) by
weight, preferably less than about 40 ppm, based on
the weight of the to~al support. Most preferred are
supports having a leachable sodium content of less
5 than about 20 ppm, by weight. Supports which are
also useful in the present invention are those having
less than a catalyst deactivating amount of ~otal
sodium. This corresponds in some instances to a
concentration of about 200 ppm. Suitable alpha-
10 aluminas having concentrations of leachable sodiumbelow 50 ppm may be obtained commercially from sup-
pliers, such as the Norton Company.
The low sodium carrier materials of the present
invention may be prepared by any conventional method
L5 of removing sodium from a solid, particularly mineral
or mineral-type material suitable in other respects
as a support material. Such treatment should not,
however, substantially adversely affect the mechani-
cal O structural characteristics of the support
material nor chemically alter the support material in
a manner which adversely affects the catalytic per-
formance indices of efficiency, activity, or catalyst
stability. Typically, the techniques involve extrac-
~L~n and/or volatilization of the sodium present. A
suitable extraction procedure may involve conversion
of the sodium present to a more easily extractabIe
material either in the same step in which extraction
takes place or in separate conversion and extraction
steps. A suitable volatilization procedure typically
3~ includes an initial step in which the sodium present
in the support is converted to a material which is
volatile upon heating. In some instances, it ~ay be
preferable to initially extract as much of the sodium
present as possible, followed by a volatilization
procedure to remove-residuaI sodium. Exemplary of
.~ .
~ 13457
-28-
~2~3~L9~
extrac~ion or leaching procedures is treatment of the
support material with a mineral acid, particularly
nitric acid in a concentration of about 10 percent,
by volume, at a temperature of about 90 degrees C,
5 for a period of about l hour and thereafter washing
the support with water. The rinsed support material
i5 then dried at a temperature of from abou~ lO0 to
1,000 degrees C for a period of from about 1 to about
3 hours. The leaching procedure described above also
10 forms the basis of a method of analysis of ~leachable
sodium~ and the definition of this term.
Alternatively, suitable alpha-alumina support
materials may be prepared so as to obtain sodium
concentrations below 50 ppm by the method described
15 by Weber et al in U. S. Patent 4,379,134.
A preferred procedure for preparing a low-sodium
support involves treatment of a support material,
particularly gamma-alumina, with an organic or inor-
ganic fluorine-co~taining substance, preferably in
aqueous solution, and thereafter firing the treated
support material at a suitable temperature. In the
present invention, the support material may either be
extruded by conventional techniques known to the art
and formed into pellets after fluori~.~ reatment and
before firing or, alternatively, formed, i.eO, ex-
truded, pellets may be fluorine-treated and then
fired. The fluorine-containing substance i9, prefer-
ably, a volatile material or one which can be readily
volatilized under firing conditions. Examples of
~ suitable fluorine-containing materials include alumi-
num trifluoride, ammonium fluoride, hydrofluoric
acid, and dichlorodifluoromethane.
The fluorine compound is used in an amount suf-
ficient to remove a major portion of the sodium
; 35 present in the sample. This amount will, o~ course,
13457
-29-
33~
vary with the amount of sodium present in the sample
but will also depend on other factors, such as the
condition under which the support material is
treated, such as the firing temperature and heating
S rate, as well as the depth of ~he b~d of material
being treated, the amount of gamma-alumina being
treated, the level of contamination of the gamma-
alumina, and how well the firing chamber is sealed.
Typically, a suitable amount of fluorine compound is
10 not more than about 3 percent, by weight, based on
the weight of the support material being treated.
Preferably, the fluorine compound is present in an
amount of about 0.8 to about 2 percent, by weight. A
suitable firing temperature for fluorine-treated
alumina is generally less than about 1,200 degrees C,
preferably from a temperature over 750 to about 1,100
degrees C. The rate of heating depends in part on
the amount of fluorine compound used. Thus, with
lower levels of fluorine, support materials having
desirable properties are generally obtained with
rapid heating. As used herein, "rapid heating"
; refers to heating from room temperature to the
- desired temperature in about 1 hour. However, with
lower concentrations of fluorine compound, slower
heating rates are yenerally preferred to achieve the
same type of product. The "slow heating" treatments
generally consist of heating from room temperature to
about 750 degrees C in about 0.5 to 1 hour and from
750 degrees C to the final temperature at a rate of
about 100 degrees C per hour.
The treatment of support materials with fluo-
rine-containing substances may provide a collateral
benefit in converting the support material to one
having a preferred "platelet" morphology. This sup-
port is described by Notermann in an application
.
13457
~ 30 ~3~
entitled "Improved Catalytic System for Epoxidationof Alkenes", in Canadian Application No. 515,865,
filed August 13, 1986, in the name of Thomas M.
Notermann.
The support particles are preferably formed
into aggregates or "pills" of a size and
configuration to be usable in commercially operated
ethylene dioxide tubular reactors. These pills may
be formed by conventional extrusion and firing, as
discussed above. When gamma-alumina is the precursor
to an alpha-alumina support material, treatment with
the fluorine recrystallizing agent may be preformed
before or after extrusion. The pills generally range
in size from about 2 mm to about 15 mm, preferably
about 3 mm to about 12 mm. The size is chosen to be
consistent with the type of reactor employed. In
general, in fixed bed reactor applicationsJ sizes
ranging from about 3 mm to about 10 mm have been
found to be most suitable in the typical tubular
reactors used in commerce. The shapes of the carrier
aggregates useful for purposes of the present
invention can vary widely. Common shapes include
spheres and cylinders, especially hollow cylinders.
Other shapes include amphora (such as defined in U.S.
Patents 3,848,033, 3,966,639 and 4,170,569),
amorphous, Raschig rings, saddles, cross-partitioned
hollow cylinders (e.g.~ having at least one partition
extending between walls), cylinders having gas
channels from side wall to side wall, cylinders
having two or more gas channels, and ribbed or finned
structures. While the cylinders are often circular,
other cross-sections, such as oval, hexagonal,
quadrilateral, trilateral, etcetera, may be useful.
.~,
~ 13457-C
,A.
-31- ~3~6
Ca~alysts:
The catalysts of the present invention are con-
ventional to the extent that silver is coated or
deposited on and/or within a solid porous carrier.
Any known method of introducing the silver to the
catalyst support may be employed. Numerous examples
and procedures are given in the patents discussed
above. In brief, in a coating or suspension process
a slurry, preferably aqueous, of ~he active catlytic
material, such as silver or its oxide, is applied to
the support to form a coherent silver layer on the
support. In the preferred impregnation process, a
solution of a soluble salt or complex of silYer in an
amount sufficient to deposit the desired weight of
silver upon the carrier is dissolved in a suitable
solvent or "complexing/solubilizing~ agent. This
solution may be used to impregnate a porous catalyst
support or carrier by immersing the carrier in the
silver-containing impregnating solution. Alterna-
tively, the support may be sprayed or sprinkled with
the impregnating solution. The excess solution may
then be allowed to drain off or the solvent may be
removed by evaporation under reduced pressure at a
suitable temperature. The silver salt or compound
used to form the silver-containing impregna~ing solu-
tion in a solvent or a complexing/solubilizing agent
is not particularly critical and any silver salt or
compound generally known to the art which is both
soluble in and does not react with the solvent or
complexing/solubilizing agent to form an unwanted
product may be employed. Thus, the silver may be
introduced to the solvent or complexing/solubilizing
agent as an oxide or a salt, such as nitrate or car-
boxylate, for example, an acetate, propionate, buty-
13457
,9~i
rate, oxalate, malonate, malate, maleate, lactate,citrate, phthalate, generally the silver salts of
higher fatty acids, and the like.
The chemical practitioner may choose from a
5 large number of suitable solvents or complexing/solu-
bilizing agents to form the silver-containing impreg-
nating solution. Besides aaequately dissolving the
silver or converting it to a soluble form, a suitable
solvent or complexing/solubilizing agent should be
10 capable of being readily removed in subsequent steps,
either by a washing, volatilizing or oxidation proce-
dure, or the like. The complexing/solubilizing
agent, preferably, should also permit solution to
provide silver in the finished catalyst to the extent
of about ~ to about 60 percent silver or higher,
based on the total weight of the catalyst. It is
also generally preferred that the solvents or com-
plexing/solubilizing agents be readily miscible with
water since aqueous solutions may be conveniently em-
ployed. Among the materials found suitable as sol-
vents or complexing/solubilizing agents for the pre-
paration the silver-containing solutions are alco-
hols, including g]ycols, such as ethylene glycol
~U. S. Patents 2,825,701 to Endler et al and
3,563,914 to Wattimena), ammonia ~U. S. Patent
2,463,228 to West et al), amines and aqueous mixtures
of amines (U. S. Patents 2,459,896 to Schwartz,
3,563,914 to Wattimena, 3,702,259 to Nielsen, and
4,097,414 to Cavitt, and carboxylic acids, such as
lactic acid (U. S Patents 2,477,435 to ~ries and
3,501,417 to DeMaio)
Typically, a silver-containing solution is pre-
pared by dissolviny silver in a suitable solvent or
complexing/solubilizing agent as, for example, a
mixture of water, ethyienediamine, oxalic acid, sil-
13457
-33
ver oxide, and monoethanolamine. The solution is
then mixed with support particles and drained.
Thereafter the particles are suitably dried.
After impregnation, the silver-impregnated car-
S rier particles are treated to convert the silver salt
or complex to silver metal and thereby effect deposi-
tion of silver on the surface of the support. As
used herein, the term "surface", as applied to the
support, includes not only the external surfaces of
10 the carrier but also the internal surfaces, that is,
the surfaces defining the pores or internal portion
of the support particles. This may be done by treat-
ing the impregnated particles with a reducing agent,
such as oxalic acid or alkanolamine and/or by roast-
ing, at an elevated temperature on the order oE about100 to about 900 degrees C, preferably about 200 to
about 650 degrees C to decompose the silver co~pound
and reduce the silver to its free metallic state.
The duration of roasting is generally for a period of
about 1 to about 10 minutes for temperatures above
200 degrees C, depending on the temperature used, and
is commonly effectuated in a hot-air belt roaster.
The concentration of silver in the finished
catalyst may vary from about 2 per~ to 60 percent,
by weight, based on the total weight of the catalyst,
more preferably the silver concentration ranges from
about 8 percent to about 5~ percent, by weight. When
a ~high silver" content catalyst is preferred, the
silver ranges from about 30 to about 60 percent, by
weight. The preferred concentration for "low silver"
content catalysts ranges from about 2 to about 20
weight percent. When a catalyst having a silver
concentration in the preferred high silver range is
prepared, the silver is preferably introduced in a
series of at least two sequential impregnation and
13457
~` -34- ~3~
roasting cycles, as discussed in greater detail be-
low. Lower silver concentrations are preferred from
a capital expense standpoin~. Xowever, the optimum
silver concentration for a particular catalyst should
also take into consideration increased productivity
resulting from performance characteristics, such as
catalyst activity, system efficiency and the rate of
catalyst aging. In many instances higher concentra-
tions of silver are preferred since they demonstrate
levels of enhanced performance, particularly catalyst
stability, which compensates for the greater capital
expenditure.
Efficiency-Enhancinq Compound:
A preferred aspect of the present invention
includes an efficiency-enhancing amount of at least
one efficiency-enhancing salt of a member of a redox-
half reaction pair. The term "redox-half reaction"
~`20 is defined herein to mean half-reactions like those
found in equations presented in tables of standard
reduction or oxidation potentials, also known as
standard or single electrode potentials, of the type
;found in, for instance, "~andbook of Chemistry", N.
A. Lange, Editor, McGraw-Hill Book Company, Inc.,
pages 1213-1218 (1961) or "CRC Handbook of Chemistry
and Physics", 65th Edition, CRC Press, Inc., Boca
Raton, Florida, pages D155-162 (1984). The term
nredox-half reaction pair" refers to the pairs of
atoms, molecules or ions or mixtures thereof which
undergo oxidation or reduction in such half-reaction
~equations. Such terms as redox-half reaction pairs
-are used herein to include those members of the class
of substance which provide the desired performance
~35 enhancement, rather than a meshanism of the chemistry
:
13457
-35-
occurring. Preferably, such compounds, when associ-
ated with the catalyst as salts of members of a half
reaction pair, are salts in which the anions are
oxyanions, preferably an oxyanion of a polyvalent
5 atom, that is, the atom of the anion to which oxygen
is bonded is capable of existing, when bonded to a
dissimilar atom, in different valence states. Potas-
sium is the preferred cation and the preferred anions
are nitrate, nitrite and other anions capable of
lO undergoing displacement or other chemical reaction
and forming nitrate anions under epoxidation condi-
tions. Preferred salts include KN03 and KN02, with
KN03 being most preferred.
Introduction 0~ Ef~iciency-
Enhancing Salt To The Carrier:
The efficiency-enhancing salt of a member of a
redox-half reaction pair may be introduced to the
catalyst in any known manner. Various sequences of
impregnating or depositing silver and efficiency-
enhancing salt on the surfaces of the carrier may be
employed. Thus, impregnation and deposition of sil-
ver and an efficl~ncy-enhancing salt of a member of a
redox-half reaction pair may be effected coinciden-
tally or sequentially, i.e., the salt or salts may be
deposited prior to, during or subsequent to silver
addition to the carrier. When more than one salt of
a member of a redox-half reaction pair is employed,
they may be deposited together or sequentially.
Typical, and in many cases preferred, of such~me~hods
include concurrent, or coincidental, impregnation in
which the solution which is used to impregnate the
support with silver also contains the dissolved effi-
ciency-enhancing salt member of a redox-half reaction
L3457
-36-
pair. This procedure permits introduction of both
the silver compound and the efficiency-enhancing salt
to the support in a single step and solution. The
other commonly employed method is the sequential
impregnation of the support in which initial intro-
duction of the silver-containing solution or effi-
ciency-enhancing salt solution (depending upon the
sequence employed) is followed by drying o the sil-
ver-containing support (and heating and/or chemical
reduction of the silver if this is the first added
substance). This support is then impregnated with a
solution of the second substance, that is, the effi-
ciency-enhancing salt (if the silver was the first
added substance). In order to perform the former
L5 procedure, i.e., coincidental impregnation, the effi-
ciency-enhancing salt must be soluble in the same
solvent or complexing/solubilizing liquid used with
the silver-impregnating solution. With the sequen-
tial procedure in which the silver is added firstl
any solvent capable of dissolving the salt which will
neither react with the slIver nor leach it from the
support is suitable. Aqueous solutions are generally
preferred, but organic liquids, such as alcohols, may
also be employed. Suitable procedures for effectin~
introduction of the efficiency-enhancing salt to the
solid support may be found in many of the patents
listed above.
In some instances, the coincidental method of
preparation of the catalyst is decidedly less prefer-
red or may not be used as, ~or examptel where a bari-
um, calcium or magnesium salt is intended to be solu-
bilized and the solution contains materials which may
precipitate the cation as, for instance, a carboxylic
acid or dicarboxylic acid such as oxalic acid.
; 35
13457
~37~ 1 Z ~3 ~
The salt of a member of a redox-half reaction
pair is added in an amount sufficient to enhance the
efficiency of the epoxidation reaction. The precise
amount will vary depending u~on such variables as the
5 gaseous efficiency-enhancing member of a redox-half
reaction used and concentration thereof, the concen-
tration of other components in the gas phase, the
amount of silver contained in the catalyst, the sur-
face area of the support, the process conditions,
10 e.g., space velocity and temperature, and morphology
of support. Generally, however, a suita~le range of
concentration of the added efficiency-enhancing salt,
calculated as cation, is about 0.01 to about 5 per-
cent, preferably about 0.02 to about 3 percent, by
weight, based on the total weight of the catalyst.
Most preferably the salt is added in an amount of
about 0.03 to about 2 weight percent.
It has been noted that when conventional
analyses have been conducted with catalysts prepared
by co-impregnation with silver and efficiency-
enhancing salt, not all the anion associated with the
cation has been accounted for. For example, cata-
lysts prepared by co-impregnation with a potassium
nitrate solution have been analyzed by conventional
techniques and about 3 moles of the nitrate anion
have been observed for every 4 moles of the potassium
cation. This is believed to be due to limitations in
the conventional analytical techniques and does not
~ necessarily mean that the unaccounted for anions are
not nitrate. For this reason, the amount of the
efficiency-enhancing salt in the catalyst is given,
in some instances, in terms of the weight percentage
of the cation of the efficiency-enhancing salt (based
on the weight of the entire catalyst), with the un-
derstanding that the anion associated with the cation
~ 13457
:~
,
~ -38- ~93~9~
is also present in the catalyst in an amount roughly
proportional (on a molar basis) to the ca~ion.
As indicated above, when the catalyst is in-
tended to contain higher concentrations of silver, it
S is generally preferred to use multip}e cycles of
impregnation and reduction to elemental silver. This
seems to result in a more uniform distribution of
silver throughout the catalyst pill. As with the
catalyst containing lower amounts of silver intro-
1~ duced in a single impregnation, the introduction ofthe efficiency-enhancing salt of a member of a redox-
half reaction pair may also be introduced in a se-
quential or coincidental procedure. Regardless of
the method employed, a first cycle includes impregna-
5 tion only with silver in an appropriate complex-
ing/solubilizing agent followed by reduction to
metallic silver, such as by roasting. (This is known
as a "silver only" procedure.) If the sequential
procedure is to be followed, the first cycle of im-
20 pregnation and reduction is repeated one or more
times. The final step includes impregnation with a
solution of the efficiency-enhancing salt of a member
of a redox-half reaction pair, draining and drying,
as discussed above.
When the coincidental or concurrent procedure is
employed in preparing catalysts containing high con-
centrations of silver, the first cycle is the same as
in the sequential technique, i.e., a silver only
procedure~ This is then followed either directly by
3~ a conventional coincidental impregnation (i.e., im-
pregnation with a solution containing both silver in
a soluble form and the efficiency-enhancing salt of a
member of a redox-half reaction pair, followed by a
reduction of silver to its elemental form, as by
roasting) or by one or more repetitions of the im-
13457
~ _39- .~ 3~
pregnation cycle with a silver only procedure inter-
posed between the silver only cycle and the coinci-
dental impregnation and reduction cycle with a solu-
tion of silver and at least one efficiency-enhancing
salt. Suitable results have been obtained with both
the sequential and coincidental procedures with two
silver impregnation cycles, although there are indi-
cations that greater amounts of silver with more
uniform distribution of silver throughout the pill
can be obtained by three or possibly more silver-
impregnation cycles. High silver-containing cata-
lysts prepared by the coincidental impregnation tech-
nique provide somewhat better performance than those
prepared by the sequential technique.
EpoxidatLon Procedure:
As in conventional processes of this type, an
alkene and an oxygen-containing gas are brought to-
gether in a reactor in the presence of a suitableepoxidation catalyst under epoxidation conditions.
Typical epoxidation conditions include temperatures
within the reaction zone of the reactor on the order
of about 180 to 300 degrees C and pressures from
about 1 to about 30 atmospheres.
The gaseous eficiency-enhancing member of a
redox-half reac~ion pair may generally be supplied to
the reaction zone within the reactor by introducing
the component to the feedstream containing alkene and
oxygen Under commercial epoxidation conditions,
such as those used in the present invention, the
feedstream also contains a gas phase halogen com-
pound, such as an alkyl halide, a hydrocarbon, and,
when the effluent stream from the reactor is recy-
cled, unreacted alkene. When recycle of the eefluent
13457
-~ _40~ 3~
stream is used, carbon dioxide may also be present.
The presence and amount of carbon dioxide depends, at
least in part, on whether a scrubbing device is em-
ployed.
The terms "gaseous member of a redox-half reac-
tion pair~, "gaseous efficiency-enhancing member o~ a
redox-half reaction pair", or like terms referred to
herein have a meaning similar to that for the "sal~
of a member of a redox-half reaction pair" or like
terms, defined above. That is, these terms refer to
members of half-reactions, represented in standard or
single electrode potential tables in standard refer-
ence texts or handbooks which are in a gaseous state
and are substances which~ in the reaction equations
lS represented in the texts, are either oxidized or re-
duced~ The preferred gaseous efficiency enhancing
materials are compounds containing an element capable
o~ existing in more than two valence states, prefer-
ably nitrogen and another element which is, prefer-
ably, oxygen. Examples of preferred gaseous effi-
ciency-enhancing members of redox-half reaction pairs
include at least one of NO, NO2, N2O4, N2O3 or any
gaseous substance capable of forming one of the
aforementione~ gases, particularly NO and NO2, under
~5 epoxidation conditions, and mixtures of one of the
foregoing, particuIarly ~O, with one or more of PH3,
CO, SO3, SO2, P2O5, and P2O3. NO is most preferred
as the gaseous efficiency-enhancing compound.
Although in some cases it is preferred to employ
members of the same half-reaction pair in the reac-
tion system, i.e., both the efficiency-enhancing salt
member associated with the catalyst and the gaseous
member in the feedstream, as, for example, with a
preferred combination of potassium nitrate and nitric
oxide, this is not necessary in all cases to achieve
13457
;3L2~3~;
satisfactory results. Other combinations, such as
KNO3/N2O3, KNO3/NO2~ KN3/N24~ KNO3/S2' KN2/N'
KNO2/NO2, and KNO3/a mixture of SO2 and NO, may also
be employed in the same system. In some instances,
S the salt and gaseous members may be found in differ-
ent half-reactions which represent the first and last
reactions in a series of half-reaction equations of
an overall reaction.
The gaseous efficiency-enhancing member of a
redox-half reaction pair is also present in an amount
sufficient to enhance the performance, such as the
activity of the catalyst, and, particularly, the
efficiency of the epoxidation reaction. The precise
amount is determined, in part, by the particular
efficiency-enhancing salt of a member of a redox-half
reaction pair used and the concentration thereof, the
particular alkene undergoing oxidation, and by other
factors noted above which influence the amount of
efficiency-enhancing salt of a member of a redox-half
reaction pair. Typically a suitable concentration of
the gaseous member of a redox-half reaction pair for
epoxidation of most alkenes, including propylene, is
about 0.1 to about 2,000 ppm, by volume, of the gas-
eous feedstream when N2 is used as ballast. When a
preferred gaseous member of a redox-half reaction
pair, such as NO, is used in the epoxidation of pro-
pylene, the preferred concentration is about 2,000
ppm, by volume, with an N2 ballast. However, when
ethylene is being oxidized, a suitable concentration
for ethylene is from about 0.1 to about 100 ppm, by
volume, of the gaseous feedstream components. Pref-
erably, the gaseous efficiency-enhancing member of a
redox-half reaction pair is present in an amount of
about 1 to about 80 ppm when about 3 percent, by
~ .
13457
-42~ 9~
volume, CO2 is present. When nitric oxide is em-
ployed as the gaseous efficiency enhancing compound
in an ethylene epoxidation system, it is present in
an amount of about 0.1 to about 60 ppm, preferably
about l to about 40 ppm, when CO2 is present.
The oxygen employed in the reaction may be in-
troduced to the reactor either as air, commercially
pure oxygen or other substance which under epoxida-
tion conditions both exists in a gaseous state and
forms molecular oxygen. The gaseous components which
are supplied to the reaction zone, or that region of
the reactor where reactants and catalyst are brought
together under epoxidation conditions, are generally
combined before being introduced to the reactor. The
reactors in which the process and catalyst of the
present invention are employed may be of any type
known to the art.
In addition to an alkene, oxygen, and the gas-
eous efficiency-enhancing member, the feedstream also
contains a halogen-containing compound, preferably an
organic halide, including both saturated and unsatur-
ated halides, such as ethylene dichloride, ethyl
chloride, vinyl chloride, methyl chloride and methy-
lene chloride. Preferably, ethyl chloride is em-
ployed as the halogen-containing compound. The
amount of halide employed will vary depending upon a
variety of factors, including the particular alkene
being oxidized and the concentration thereof, the
particular efficiency-enhancing gaseous component and
salt ar.d the concentrations thereof, as well as other
~ factors noted above as influencing the amount of
; efficiency-enhancing gaseous compound and salt.
suitable range, however, of concentration for the
halogen-containing compound is typically about 0.1 to
about 60 ppm, by volumie, of the gaseous makeup feed-
' .
~ 13457
_ ~ Z~ 3 ~3~
stream. In addition, a hydrocarbon, such as ethane,can be included in the feedstream. The feedstream
may also contain a diluent, such as nitrogen, or
other inert gas, particularly when air is used as the
5 oxygen containing gas. Varying amounts of carbon
dioxide and water vapor may also be present, depend-
ing upon whether means have been provided to remove
such substances from the effluent stream prior to
combination of at least a portion of the effluent
stream with the inlet stream. Other than the gaseous
efficiency-enhancing member of a redox-half reaction
pair, the other components are present in convention-
ally used amounts, as shown in the following table.
Volume Percent ~or ppm)
Component for Propylene Oxidation
propylene about 2 to about 50
oxygen about 2 to about 10
20 alkyl halide about 5 to about 2,000 ppm
hydrocarbon 0 to about 5
carbon dioxide up to about 15
nitrogen remainder
or other ballast
gas, e.g., methane
13457
-44- .~Z~3~
Volume Percent (or ppm3
Component for Ethylene Oxidation
ethylene at least about 2,
often about 5 to about 50
oxygen about 2 to about 8
alkyl halide about 0.1 to about 60 ppm
hydrocarbon about 0 ~o about 5
carbon dioxide up to about 7
10 nitrogen remainder
or other ballast
gas, e.g., methane
When higher alkenes, such as those previously
discussed, are epoxidized, conditions and concentra-
tions typically used for the epoxidation of propylene
may be employed.
Standard Alkene Oxide Process Test Conditions-
The successful commercial production of alkene
oxides, particularly ethylene oxide, by the silver-
catalyzed oxidation of alkene, particularly ethylene,
depends upon a variety of factors. Many of these
factors influence, either directly or indirectly, the
performance indices, such as the efficiency of the
catalytic system, the activity or the aging rate,
i.e., stability, of the catalyst. The manner in
which catalysts and catalytic systems are evaluated
in the laboratory strongly influences the values
obtained for these parameters. Techniques and exper-
iments designed to assess such catalysts and cata-
lytic systems commonly employ microreactors (i.e.,
tiny tubular reactors for testing crushed catalyst
particles) or back-mixed autoclaves of the Berty type
. .
13457
_~5- ~Z93~
ti~e.~ larger reactors which test full-sized catalyst
pellets and generally employ full gas recycle) as
described in Figure 2 of the article by J~ M. 8erty,
"Reactor For Vapor Phase-Catalytic Studies", Chemical
5 ~n~ineerinq Progress, 70, Number 5, pages 78-84
(1974), and particularly Figure 2. Microreactors are
capable of yielding, in most test situations, the
highest efficiency numbers, typically approximately
the same as or somewhat greater than those obtainable
in commercial tubular reactor operations employing
the same catalysts in non-crushed condition. Back-
mixed autoclaves commonly provide lower efficiency
values because, although conditions can be varied,
generally the entLre catalyst is exposed to the out-
let gas which has the lowest concentration of re-
actants and the highest concentration of products.
Values obtained using one type of reactor are seldom
directly comparable to those obtained in another
reactor system. As a result, claims of superior
results or the desirabi~ity of one catalyst over
another are preferably based on tests conducted under
controlled and comparable conditions.
Although the conditions set forth supra may be
employed boti~ for reactors employed in commercial
production as well as those employed in a laboratory,
as a basis of comparison, the catalysts and catalytic
systems for epoxidation of ethylene employed in the
examples set forth below have been tested u*der com-
parable conditions known as Standard Alkene Oxide
Process Test Conditions, or Standard Test Conditions
(referred to hereinafter as STC). The STC employed
for testing and characterizing the catalysts and the
catalytic systems of the present invention involve
the use of a standard back-mixed bottom-agitated
nMagnedrive" autoclave or Berty autoclave, as de-
13457
46~ 3
scribed above.
In discussing the enhancement of efficiencyprovided by the present invention, it may be noted
that, when an efficiency-enhancing amount of a salt
of a member redox-half reaction pair is employed, an
efficiency of at least about 84 percen~ i5 obtained
under Standard Test Conditions. "S~andard Test Con-
ditions" may be deined as comprising the following:
by volume, 30 percent C2H4, 8 percent 2~ 5 ppm ethyl
chloride, 5 ppm, by weight, NO, no added C2H6 or CO2,
N2, ballast, 240 degrees C, 275 psig, gas hourly
space velocity (GHSV) = 8,000 hr 1.
Although the present invention can be used with
any size and type of alkene oxide reactor, including
both fixed bed and ~luidized bed reactors known to
the art, it ls contemplated that the present inven-
tion will ind most widespread application in
standard fixed bed, multi-tubular reactors. These
generally include wall-cooled as well as adiabatic or
non-wall-cooled reactors. Tube lengths may typically
range from about 5 to about 60 feet but will fre-
quently be in the range of from about 15 to about 40
feet. The tubes may have internal diameters from
about 0.5 to about 2 inches and are expected to u~
typically from about 0.8 to about 1.5 inches. G~SV
range from about 16,000 to about l,000 hr~l.
Typically GHSV values range from about 2,000 to about
8,000 hours~l at pressures from about 1 to about 30
atmospheres, commonly about 10 to about 25 atmos-
pheres.
While the invention is susceptible to variousmodifications and alternative forms, certain specific
embodiments thereof are described in the examples set
forth below. It should be understood, however, that
these examples are not intended to limit the inven-
13457
` ~~7~ lZ~3~6
tion to the particular forms disclos~d but, on thecontrary, the invention is to cover all modifica-
tions, equivalents and alternatives falling within
the spirit and scope of the invention.
s
Example 1 - Preparation Of
Alpha-Alumina Support Havin~o~ Sodium:
Approximately 1,200 grams of 1/4 inch hollow
gamma-alumina pills or rings (obtained from ~orton
Corporation as Alumina 6573) were soaked for one-half
hour in 1 molar NH4F with occasional shaking to li-
berate bubbles that formed durinq impregnation.
Excess solution was drained from the rings, which
were then dried overnight at 120 degrees C. The
dried rings were distributed between four 3 inch
diameter x 6 inch deep cylindrical alumina crucibles
and were fired in a P. D. H. high-temperature fur-
nace. Lids for the crucibles were placed askew so
that crucible openings were covered approximately 70
percent to maintain a fluoride atmosphere in the cru-
cible but at the same time permit escape of excess
fluoride. The firing schedule was one hour to raise
the temperature to 700 degrees C, four hours to raise
2S the temperature from 700 to 1,100 degrees C, and a
hold temperature at 1,100 degrees C for one hour.
The product had a surface area of 1.56 m~/g and a
sodium content of less than 50 ppm.
Example 2 - Sequential Preparation Of A Potassium
Nitrate-Containing Suppor~ed Silver Catalyst:
.
Formation of a potassium nitrate-containing
supported silver ca~alyst resulted from a multi-step
procedure in which a support was initially impreg-
13457
~` -48- ~ 2~
nated with a silver-containing solution, the impreg-
nated support was roasted, and the silver-inpregnated
catalyst was thereafter impregnated with a potassium
nitrate solution and then dried.
S The silver-containing solution used in the first
step was prepared by dissolving 51.49 grams ethylene-
diamine in 51.03 grams of distilled water and stir-
ring the mix~ure for a period of 10 minutes. To the
stirred solution was slowly added 51 56 grams of
10 oxalic acid dihydrate. The resulting solution was
stirred for 10 minutes. To this solution was added,
in portions, ga.34 grams silver oxide. The resulting
silver-containing solution was thereafter stirred for
an additional hour and 18.07 grams of monoethanol-
amine was then added to the stirred silver-containing
solution. Stirring was continued for an additional
10 minutes. This solution was then diluted to a
total volume of 437 ml by addition of distilled
~ater.
2~ In a catalyst-impregnation tube were placed
177.85 grams of an alpha-alumina carrier, similar to
that prepare~ in Example 1. A catalyst-impregnation
tube is an elongated tube, which in use is arranged
vertically. The upper end of the tube is provided
with an inlet to supply impregnation solution to the
carrier contained in the tube and the lower end with
an outlet from which solution may be drained. The
- tube is also provided with a means to connect the
tube to a vacuum source. Prior to introduction o~
impregnation solution the tube was evacuated. The
solution described immediately above was slowly
poured into the tube to totally immerse the
support. The carrier was allowed ~o remain in the
impregnating solution for about l hour to achieve
saturation of the support. The unabsorbed solution
13457
:~293~96
was thereafter drained from the support with drainlng
continuing for about 30 minutes. The wet,
impregnated carrier was then roasted in a hot-air
belt roaster for 2.5 minutes at 500 degrees C ~66
SCFH air flow) to produce a catalyst containing 11.39
percent silver.
~ pon cooling, 44.57 grams of the roasted, im-
pregnated carrier was placed in an impregnation tube
and covered with a solution formed by dissolving 2.08
grams potassium nitrate in lO0 ml distilled water.
After covering the roasted silver-containing carrier
with the KNO3-containing solution and allowing to
stand for 15 minutes, the catalyst pellets were
drained on a funnel for 5 minutes. The resulting
drained pellets were dried in an oven for 2 hou~s at
120 degrees C, producing a catalyst containing 0.46
percent K, calculated as 1.20 percent KNO3, by
weight.
Example 3 - Sequential Preparation Of
Sodium- And Potassium Nitrate-Containing
Su~orted Silver CatalYst: ~
An impregnated and roasted silver-containing
25` catalyst was prepared as described in Example 2.
Into an impregnation tube was placed 44.57 grams of
the silver-containing catalyst (11.39 percent, by
weight, Ag) and a solution prepared by dissolving
2.08 grams potassium nitrate and 0.04 grams sodium
nitrate in lO0 ml distilled water was placed in the
tube after evacuation of the tube to totally cover
the silver-containing. The pellets were allowed to
stand in the solution for 15 minutes, the excess
solution was drained from the pellets on a funnel for
5 minutes, and the impregnated catalyst was oven-
13457
-50-
~33~9~
dried for 2 hours at 100 degrees C. The catalyst
pellets contained, by weight, 0.44 percent K, calcu-
lated as 1.15 percent KNO3 , and 0.008 percent ~80
ppm) Na.
Example 4 - Preparation Of A Potassium
Nitrate-Containing Silver Catalyst
BY A Co-Impre~nation Technique: _
A co-impregnation solution was prepared by plac-
ing 18.77 grams of ethylenediamine into a 400 ml
beaker and mixing therewith 25 grams of distilled
water to form a solution. To the stirred solution
was slowly added 18.7 grams oxalic acid and, with
continuous stirring, 3$.2 grams of silver oxide was
slowly added. When solution was complete, 6.55 grams
of monoethanola~ine was added directly to the solu-
tion. To the silver-containing solution was added
7.14 gra~s of potassium nitrate solution, 10 percent
with respect to potassium (0.258 grams KNO3/g solu-
tion). Sufficient water was added to the resulting
solution to dilute the solution to 100 ml.
i To make 100 ml of finished catalyst, 53.5 grams
of an alpha-alumina support material of the type
prepared in Example 1 but fired at a temperature of
1,000 degrees C in the form of quarter-inch ring
` extrudate and having a pore volume of ~.62 g H2O/g
support was placed in a suitable container. A source
of vacuum was attached to the container and evacuated
and thereafter-the silver/potassium-containing im-
pregnating solution was added until the carrier was
completely covered. After about one hour the excess
solution was drained from the catalyst.
Roasting of the impregnated catalyst was con-
ducted as indicated above in Example 2. The result-
13457
~` sl
~93~
ing catalyst in~luded! by weight, 17.5 percent silverand, as calculated, about 1 percent potassium nitrate.
Example ~ - Sequ~rltial Preparation
Of A Potassium Nitrate-Containing
Sup~orted Hi~h Sllver Concentr_tion Catalyst
A solution containing 17.5 g
ethylenediamine, 17.3 g distilled water, 17.5 g
oxalic acid, 30.7 g silver oxide, 6.1 g
monoethanolamine, diluted to 125 ml, was prepared in
the manner described in Example 2.
Into a catalyst impregnation tuhe were
placed 43.9 g of an alpha-alumina carriee, similar to
that prepared in Example 1, and having a surface area
of 1.12 m2/g, a pore volume of 0.8 cc/g, less than
50 ppm, by weight, leachable sodium, and a platelet
morphology of the type disclosed in Canadian
Application ~o. 515,865. The support pellets were
initially vacuum-impregnated at 28 inches mercury
vacuum, releasing the vacuum when the pellets were
completely covered with the aqueous silver amine
solution, permitting them to stand in the amine
solution for an additional hour at 1 atmosphere
before draining. The impregnated pellets were then
belt-roasted at 500 degrees C in an oven having an
air flow of 66 SCFH for 2.5 minutes. The material
contained 14.9 percent silver as determined by weight
gain. This material was then impregnated a second
and third time with fresh silver/amine solution
having the composition indicated above. The
impregnations and belt-roastings were conducted in
the same manner as described above. The catalyst
material was
13457
~ .
-
-52~ 3~
then impregnated with KNO3 by placing 67.4 9 of the
silver-impregnated pellets into a solution containing
2.1 g KNO3 dissolved in 100 ml distilled water. The
KNO3-impregnated silver catalyst material was then
5 dried at 120 degrees C for 1 hour to yield a catalyst
containing, by weight, 34.7 percent silver and 0.33
percent K (calculated as 0.86 percent K~03), as de-
termined by weight gain.
10 Example 6 - Coincidental Preparation
Of A Potassium Nitrate-Containing
Supported High Silver Concentration Catalyst:
A silver impregnation solut;on was prepared in a
15 manner similar to that described in Example 2 by
mixing 196.8 g ethylenediamine, 195.0 g distilled
water, 197.1 g oxalic acid, 345.2 g silver oxide,
69.1 g monoethanolamine and diluting to a total vol-
ume of 780 ml with distilled water. After evacuation
of an impregnation tube containing pellets of an
alpha-alumina support material ~293.1 g) similar to
those described in Example 1 and having a B. E. T.
surface area of 1.16 m2/g by a Quantasorb apparatus,
the pellets were impregnated by allowing them to
stand in the silver impregnation solution for a per-
iod of 1 hour and then draining. The impregnated
pellets were then belt-roasted at 500 degrees C in a
66 SCFH air flow for 2.5 minutes. As determined by
gain in weigh~t, the material contained 24.6 percent
silver
A co-impregnation solution was prepared in a
manner similar to that described in Example 3 by
mixing 37.5 g ethylenediamine, 37.2 g distilled
~water, 37.6 g oxalic acid, 65.9 g silver oxide, 13.2
;;35 g monoethanolamine, and 0.80 g potassium nitrate and
13457
-53-
diluting to 150 ml total solution with distilled
water. The silver-impregnated catalyst pellets (77.8
g) were impregnated by immersing them in the sil-
ver/potassium nitrate-impregnating solution for 1
hour and then draining. The resulting pellets were
then belt-roasted as described above. The pellets
resulting from the two impregnations yielded a mater-
ial, as determined by an analysis, containing, by
weight, 39.8 percent silver and 0.098 percent K (cal-
culated as 0.25 percent KNO3).
Examples 7 to 12 Production OfEthylene Oxide With Potassium
Nitrate-Containing Supported Silver Catalyst:
The examples set forth below were conducted
employing catalyts using carriers of the type pre-
pared in Example 1, the catalysts themselves being
prepared in the manner described in Examples 2 to 6
and identified as being prepared by either a sequen-
tial (S) procedure or a coincidental impregnation (C)
procedureO The epoxidation studies for which data
are presented below were conducted in a continuously
stirred tank reactor, alsG ~.~own as a back-mixed
autoclave of the type described above. The procedure
involved charging approximately 80 ml of catalyst to
the autoclave. The volume of catalyst was measured
in a one inch I. D. graduated cylinder after tapping
the cylinder several times to thoroughly pack the
catalyst. The weight of the catalyst was noted and
is indicated in the table appearing below. The back-
mixed autoclave was heated to about reaction tempera-
ture in a nitrogen flow of 11.3 SCFH with the fan
~ operating at about 1,500 rpm. The nitrogen flow was
; 35 then discontinued and the feedstream was introduced
13457
-54
to the reactor.
All ethylene epoxidation reactions, except where
indicated otherwise, were examined under Standard
Test Conditions comprising, by volume, 30 percent
C~H4~ 8 percent 2' 5 ppm ethyl chloride, 5 ppm, by
weight, NO, no added C2H6 or CO2, N2 ballast, 240
degrees C, 275 psig, GHSV = 8,000 hr 1. The data
summarized below includes surface areas of the cata-
lysts, Ag and calculated KNO3 contents, maximum ob-
served outlet ethylene oxide, maximum observed effi-
ciency, and aging rate.
13457
-55-
T~BLE
Pre-
para- Calcu-
tion lated
Example Proce- Na ~ Ag % RN03 Activity Effi-
dure ~m ~EØ) ciency
7 C 12 19.7 2.0 1.18 91~0
108 C 12 20.0 2.0 1.41 90.3
9 C 12 19.7 2.0 1.23 90.0
C 12 20.4 2.0 1.35 ~0.3
11 S 12 15.6 1.0 1.12 86.5
12* C 12 39.8 0.1 2.03 8~.0
1513** S 12 11.4 1.2 See 88.2
Figure
14** S 80 11.4 1.2 See 87.8
Figure
15*** C 40 39.8 0.1 See See
Figure Figure
2 3
16*~ C 100 39.8 0.1 See See
Figure Figure
2 3
*Test conditions, by volume, 30 percent C2H4, 8 per-
cent 2~ 20 ppm ethyl chloride, 15 ppm N0, 270 de-
grèes C, 270 psig, GHSV = 8,000 hour 1. The catalyst
Of Example 6 is used. The catalyst of Example 3 is
used.
**GHSV = 4,000 hr~l, all other conditions the same as
for Examples 7-11.
***Conditions the same as Example 12 but additionally
includes 3 percent C02 in the feedstream. The cata-
13457
-56
~39~
lyst of Example 6 is used.
Example 13 - Production of Propylene
Oxide With Potassium Nitrate-
1 Supported Silver Ca a~st:
Into a back-mixed autoclave, having a
slightly smaller volume than the autoclave employed
in the previous examples, were placed 5.0 grams (10.0
ml) of a catalyst having a support prepared as in
Example 1, The support was an alpha-alumina having a
surface area of 1.12 m2/g, a pore volume of 0.8
cc/g, less than 50 ppm, by weight, of leachable
sodium, a residual fluorine concentration o-f 0.7
percent, and a platelet morp~ology of the type
disclosed in Canadian Application No. 515,865. The
catalyst included, by weight, a silver con~entration
of 17.7 percent and a potassium concentration of 0.38
percent (calculated as 1 percent potassium nitrate).
A mixture, by volume, of 9.9 percent propylene, 7.75
percent oxygen, 200 ppm ethyl chloride, 200 ppm
nitric oxide, 2 percent methane, and the remainder
being nitrogen ballast gas, was fed to the reactor at
a temperature of 252 degrees C, a pressure of 25 psi
and a gas hourly space velocity of 832 hr 1 After
22.5 hours, a selectivity of 47.3 percent was
observed with an activity of 0.6 pounds of propylene
oxide/cubic foot of catalyst/hour (corresponding to
an outlet percent of propylene oxide of 0.465
percent).
13457
., .~