CA1293240C - Catalyst and process for manufacture of alkene oxide - Google Patents

Abstract

Abstract:
A process for the epoxidation of alkene to form alkene oxide comprising contacting alkene and oxygen-containing gas under epoxidation conditions in the presence of at least one gaseous efficiency-enhancing member of a redox-half reaction pair, e.g., nitric oxide, carbon dioxide and a supported silver catalyst comprising a catalytically effective amount of silver and at least one efficiency-enhancing salt of a mem-ber of a redox-half reaction pair, e.g., potassium nitrate, with efficiency-enhancing salt present in the catalyst in an amount sufficient to provide an efficiency of at least about 84 percent under speci-fied test conditions but below the amount which under a second set of specified test conditions would re-duce the activity to less than four pounds of ethylene oxide per cubic foot of catalyst per hour, thereby reducing the activity-reducing effect of carbon dioxide in carrying out the process.

Description

CATALYST AND PROCESS FOR
MANUFACTURE OF ALKENE OXIDE
Technical Field:

The present invention is directed to improved processes for the preparation of alkene oxide from alk~ne and oxygen-containing gas employing a sup-ported silver catalyst. Particular aspects of the present invention relate to processes for epoxidizing alkene in the vapor phase to produce the correspond-ing alkene oxide at high efficiencies.
Back~round And Background Art:

The production of alkene oxides, or epoxides, particularly ethylene oxide, by the direct epoxida-tion of the corresponding alkene in the presence of a silver-containing catalyst has been known for many years. One of the earliest disclosures of a process : : for the direct epoxidation of ethylene was that of ~ Lefort~ U. S. Patent i,998,878, issued in 1935 (re-:: 30 issued in 1942~as Re. 20,370). Lefort discloses ~hat : e:~hylene oxide can be formed by reacting ethylene and ~ oxygen according to the followlng equation:
: ~ , ~3~2'~

2cH2=cH2 + 02~ 2CH2-CH2 ~I) Lefort recognized, however, that some of the ethylene, when reacted with oxygen, is comDletely oxidized to carbon dioxide according to the following equation:

CzH4 + 32 ~2H2O +2C02 (II) Reaction II, as well as other reactions in which alkene is converted to products other than alkene oxide, is undesirable since the alkene reactant is consumed in the formation of undesired products.
lS Further, the undesired products, e.g., carbon di-oxide, may adversely affect the reaction systesn. In general, the overall effectiveness of an alkene oxide production system is gauged by the performance char-acteristics of the system. The most important per-formance characteristics are the efficiency, theactivity, and the useul life of the catalyst, all of which are defined and more completely described be-low.
Percent conversion is defined as the percentage of the alkene introduced to the reaction system that undergoes reaction. Of the alkene that reacts, the percenta~e that is converted into the corresponding alkene oxide is referred to as the selectivity or efficiency of the process. The commercial success of a reaction system depends in large measure on the efficiency of the system. At present, maximum effi-ciencies in commercial production of ethylene oxide by epoxidation are in the low 80s, e.g., 80 or 81 percent. Even a very small increase in efficiency will provide substantial cost benefits in large-scale lS015 ~3~

operation. For example, taking 100,000 metric tons as a typical yearly yield for a conventional ethylene oxide plant, an increase in efficiency of from 80 to 84 percent, all other things being equal, would re-sult in a savings of 3790 metric tons of ethylene peryear. In addition, the heat of reaction for Reaction II (formation of carbon dioxide) is much greater than that of Reaction I (formation of ethylene oxide) so heat removal problems are more burdensome as the e'ficiency decreases. Furthermore, as the efficiency decreases, there is the potential for a greater amount of impurities to be present in the reactor effluent which can complicate separation of the de-sired alkene oxide product. It would be desirable, therefore, to develop a process for the epoxidation of alkene in which the efficiency is greater than that obtained in conventional commercial processes, e.g~ t with ethylene, efficiences of 84 percent or greater, while maintaining other performance charac-teristics, particularly the activity, as de~cribedbelow, in a satisfactory 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 corresponding oxide. The definitions of conversion, efficiency and yield may be represented as follows:

1~3;~ ~

% Conversion = moles alkene-reacted x 100 moles alkene fed % Eff iciency a mol~es alkene oxide produced x 100 S moles alkene reacted % Yield _ moles a1kene oxide produced x 100 moles alkene fed :
1 0 :
Generally, in a process for the epoxidation of alkene, a reaction inlet stream containing reactants and perhaps other additional materials enters the reactor or reaction zone in which catalytic material is provided and in which favorable reaction condi-tions (e.g., temperature and pressure) are maintain-ed. The reactor effluent is withdrawn or collected from the reactor. The reactor effluent contains reaction products, together with unreacted components ~rom the reaction inlet stream.
Since at least some alkene is generally not converted during its initial pass through the reac-tor, alkene is generally present in the reactor ef-fluent. To increase the overall yield of the pro-cess, at least a portion of the alkene in the reactoreffluent is returned to the reactor via a recycle stream. Means are provided for removing and recover-ing at least a portion of the alkene oxide from the reactor effluent, prior to recycling, to form the product stream. At least a~portion oE ~he remaining reactor effIuent (after the product stream has been withdrawn) becomes the recy~le stream. The recycle stream preferably contains substantially all of the alkene that was contained in the reactor effluent.
Reactants, iOe. r oxygen-containing gas and al-kene, need to be continuously replaced and are pro-3~

vided to the reactor by means of a makeup feed-stream. In general, the makeup feedst:ream and the recycle stream are combined to form the reaction inlet stream which is sent into the reactor or reac-tion zone. Alternatively, the makeup feedstream andthe recycle stream can be introduced into the reactor separately.
The epoxidation of alkene is generally carried out in the presence of a supported silver catalyst located within the reactor or reaction zone. The silver may be supported by a conventional support material, for example, alpha-alumina. The performance of the catalyst may be affected by the presence of solid, liquid or gaseous compounds which may, for example, be incorporated in the catalyst or provided via the makeup feedstream.
The activity of a reaction system is a measure o the rate o~ production of the desired product, e~g., ethylene oxide, for a particular reaction sys-tem at a particular temperature. In order to providea meanin~ful comparison of the effectiveness of two or more reaction systems or of a single reaction system at different times, factors, such as feed rate, feed composition, temperature, and pressure, that affect the rate of production of the desired alkene oxide must be normalized or accounted for, preferably by using a standard or fixed set of oper-ating conditions. Since the rate of production of alkene oxide is proportional to the volume of cata-lyst in the reaction system, the activity is usuallyexpressed in terms oE pounds of alkene oxide produced per hour per cubic foot of catalyst. It should be noted that this method of measuring activity does not take into account variations in the densities of the catalysts since the controlling factor is the volume of the reaction system available, not the weight of catalyst which will fit into a given volume. Other factors that have an effect on the rate of production of the desired compound include the following:
(1) the composition of the reaction stream' (2) the gas hourly space velocity of the reac-tion stream;

(3) the temperature and pressure within the reactor or reaction zone.
In order to compare the efectiveness of two or more reaction systems or of a single reaction system at different times, differences in factors 1 through 3 above should be minimized and/or factored into the evalua~ion of relative effectiveness.
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 commercially impractical. In general, an activity below 4 pounds of ethylene oxide per hour per cubic foot of catalyst is unacceptable for commercial prac-tice. The activity is preferably greater than 8pounds, and in some instances an activity greater than Il pounds of ethylene oxide per hour per cubic foot of catalyst is desired.
Reaction systems generally deactivate over time, i.e., the activity of the catalyst begins to decrease as~the process is carried out. Activity may be plot-ted~as a function of time to generate a graph showing the aging behavior of the catalyst. Experimentation for the purpose of developing an activity plot is usually conducted at a set temperature slnce, in 3;~

general, activity can be increased by raising the reaction temperature. Alternatively, an activity plot can be a graph of the temperature required to maintain a given activity versus time. The rate at which activity decreases, i.e., the rate of deactiva-tion at a givçn point in time, can be represented by the slope of the activity plot, i.e., the derivative of activity with respect to time:

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 = Q activity/ ~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 high or the rate of production becomes unacceptably low. At this point, the cata-lyst must either be regenerated or replaced. 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 useul 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. More specifically, the replacement of the catalyst generally requires that the reactor be ~015 -8- ~32 ~

shut down for an extended period of time, e.g., two weeks or more, to discharge the catalyst, clean the reactor tubes, etcetera. This operation requires extra manpower and the use of special equipment. The costs involved, which may include replacement cata-lyst t can mount into the millions of dollars.
As used herein, an activity-reducing compound refers to a compound which, when present in an acti-vity reducing amount, causes a reduction in activity, some or all of which activity may subsequently be regained by returning to a situation in which ~he concentration of the compound is below the minimum activity reducing amount. The minimum activity-reducing amount varies depending on the particular system, 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 de~activation. Catalysts tend to deactivate more rapidly when reaction is carried out at higher tem-peratures.
As previously noted, since the~work of Lefort(U. S. Patent 1,998,878)~ research efforts have been directed toward improving the performance character-istics of reaction systems, i.eO, improving the acti-`
vity, efficiency and useful life. Research has beencon~ducted in~areas such as feedstream additives~
removal of materials in the recycle stream and methods of catalyst preparation, including the deposition or impregnation of a particular type or form of silver. Additionally, research efforts have 3 ~ ~ ~
g been directed toward the composition and formation of the support, as well as toward additives deposited on or impregnated in the support.
One of the difficulties in carrying out research is the necessity of considering the interrelationship of the various variables. The improvement or en-hancement of one performance characteristic must not b~ at the expense of, or have ~oo great an adverse effect on~ one of the other performance characteris-tics. For example, if a reaction system is designedwhich has a very short u~eful life, the system may be commercially impractical even though the efficiency and initial activity of the catalyst are outstand-ing. 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, would be particular-ly beneficial.
Diluents have generally been included in the gaseous mixture to reduce the likelihood of explo-sion. Diluents are generally supplied via the makeup feedstream. Such diluent materials have generally been b~lieved to be inert, i.e , their function is primarily to act as a heat sink and to dilute the gaseous mixture. Nitrogen has been found to be a suitable diluent material. It is well known to use air to supply both oxygen and nitrogen to the reac-tion zone. Another ma~erial that has been used as a diluent is carbon dioxide. EPO Patent 3642 discloses that a diluent, for example, heliumt nitrogen, argon, carbon dioxide, and/or a lower paraffin, for example, ethane and/or methane, may be present in proportions of 10-80 percent and preferably 40-70 percent by volume in total. Similarly, U~ K. Patent Application GB 2 014 133A mentions carbon dioxide as a possible ~2932 ~C~

diluent. Other patents, e.g., U. S. Patents 3,043,854, 4,007,135/ and 4,206,128, Japanese Patent 53-39404, and U. K. Patents 676,358 and 1,571,123, also mention that carbon dioxide is suitable for use as an additiveO
Lefort, in U. S. Patent 1,998,878 (Re. 20,370), states ~hat carbon dioxide may be introduced into the reactor to limit the rate of complete oxidation of ethylene to carbon dioxide. Similar disclosure is found in U. S. Patent 2 f 270,780. U. S. Patent 2,615,900 discloses a process for producing ethylene oxide in which carbon dioxide gas may be added to the feed gases to act as a "depressant" or "anti-cata-lytic material". U. S. Patent 4,007,135 discloses a process in which, according to the patent, carbon dioxide may be used to raise the selectivity of the xeaction. According to Chem. Abstracts, Vol. 80, Issue 11, ~ection 22, Abstract 059195, the presence of carbon dioxide tends to retard the deactivation of ~he silver catalyst.
As mentioned above, any alkene contained in the reactor effluent stream is preferably returned to the ; reaction zone via a recycle stream. It is sometimes preerred to remove some of the gas contained in the reactor effluent stream via a purge stream prior to introducing the recycle stream into the reaction zone. The purge stream may comprise a straight purge, i~e., the purge stream can merely draw off a percentage of the recycle stream. Since a straight purge stream generally has a composition substantial-ly similar to that of the stream fro~ which it is ~removed, some alkene will generally be purged when a straight purge is employed. FGr this reason, means are sometimes provided to ensure that purge streams have relatively high concentrations of materials .

~5015 `~' Z~

other than alkene, such as nitrogen and carbon diox-ide~.
U. S. Patent 2,241,019 discloses a process in which the purge gas is carried through and in contact with an adsorptive agent which is adapted to adsorb selectively the ethylene content of the purge gas, while the nitrogen and much of the carbon dioxide present in the purge gas pass through the adsorption agent and are discharged to the atmosphere.
L0 U. S~ Patent 2,376,987 discloses a process for the two-stage preparation of butadiene in which, in the first stage, ethylene is oxidized in a converter to form ~thylene oxide. The converter contains an oxidizing catalyst which is preferably finely divided silver on a carrier such as alumina. According to the patent, if concentrated oxygen is used as the oxygen source, the ethylerle in the stream containing the oxidation products from the converter may be con-centrated and recycled to the process by scrubbing to remove carbon dioxide, etcetera.
U. S. Patent 2,653,952 discloses a process for the manu~acture of ethylene oxide in which the pro-ducts from the reactor, consis~ing essentially of ethylene oxide, ethylene, oxygen, nitrogen, helium and carbon dioxide, are dellvered to an ethylene oxide absorber. The gases are then passed in contact with a solvent for carbon dioxide. Ordinarily, ethanolamine is used as a ~solvent in this process.
The gas discharged from ~he;carbon dioxide absorber contains ethylene and is r~ecycled to ~e again passed in contact with the catalyst in the reactor. This patent recognizes that nitrogen tends to build up to higb concentrations~when the oxygen is supplied by air. The process of this paten~ therefore employs relatively nitrogen-free oxygen in the feedstream and ~5015 .

dilutes the gaseous mixture with helium.
U. S. Patent 2,799,687 discloses a preferred embodiment for the oxidation of olefins in which the reactor effluent may be passed into an ethylene oxide absorber after which about 70-90 percent of the ef-fluent from the ethylene oxide absorber is recycled and the remainder plus additional oxygen is diverted to a second reactor. According to the patent, by the di~ersion of a portion of the effluent from the first reactor to the second reactor, the buildup of carbon dioxide above certain limits, such as above about 5-7 percent, can be prevented. Similarly, U. S. Patent 4,206,128, Netherlands Patent Ap~lication 6,414,284, and U. K. Patent 1 1~1 983 all disclose processes in which some carbon dioxide is removed from the recycle stream.
According to U. K. Patent 1,055,147, one must remove carbon dioxide from the ethylene oxide produc-tion system to keep the carbon dioxide concentration in an acceptable range since, according to the pat-ent, carbon dioxide acts as an inhibitor and suppres-ses the reaction of ethylene to form both ethylene oxide and carbon dioxide.
U. S. Patent 1,998,878, U. S. Patent 3,904,656, ~he Manufacture Of Ethylene Oxide And Its Deriva-tives", The Industrial Chemist, February, 1963r Kirk Othmer, "Ethylene Oxide", Volume 8, pages 5~4,545, "Ethylene Oxide By~Direct Oxidation Of Ethylene", Petroleum Processin~, November, 1955r all include, as a process step, the removal of carbon dioxide from the alkene oxide for the purification of the alkene oxide product.
Since the early work on the direct catalytic oxidation of ethylene to ethylene oxide, it has been suggested that the addition of certain compounds to , .

13~ 32~

the gaseous feedstream or direct incorporation of metals or compounds in the catalyst could enhance or promote the production of ethylene oxide. Such metals or compounds have been known variously as 'lanti-catalystsn, "promoters" and "inhibitors".
These substances, which are not considered catalysts, are believed to contribute to the overall utility of the process by inhibiting the formation of carbon dioxide or by promoting the production of ethylene oxide, Various compounds have been found to provide some beneficial effects when contained within the gaseous mixture supplied ~o the reactor. It is well known that chlorine-containing compounds, when sup-plied to an ethylene oxide production process, helpto improve the overall effectiveness of the pro-cess. For example, Law and Chitwood, in U. S. Patent 2,194,602, disclose that higher yields of olefin oxide are ob~ained by retarding the complete oxida-tion of the olefin through the addition of very smallamounts of deactivating materials (also reEerred to by Law and Chitwood as repressants or anti-catalytic materials) such as ethylene dichloride, chlorine, sulfur chloride, sulfur trioxide, nitrogen dioxide, or other halogen-containing or acid-forming materials. U. S. Patents 2,270,780j 2,279,469, 2,279,470, 2,799,~87, 3,144,416, ~,007,135, 4,206,128, 4,368,144, EPO Patent 11 355, U. K. Pat-ents 675,358, I,055,147 and 1,571,123 also discuss the addition of halide compounds, such as ethyl chloride, ethylene dichloride, potassium chloride, vinyl chloride and alkyl chloride.
U. S. Patent 2,194,602 discloses a method for the activation of silver catalysts in which the acti-vation is accomplished by bringing the catalyst in contact with an aqueous solution of barium, strontium or lithium hydroxide after the catalyst has first been treated with a "repressant" such as ethylene di-chloride, nitrogen dioxide, or other halogen-contain-ing or acid-forming materialO
U. S. Patents 2,279,469 and 2,279,470 disclose processes of making olefin oxides in which very small amounts, i.e., less than 0.1 percent of the total volume, of "anti-catalysts" are incorpora~ed with the reactants. Halogens and compounds containing halo-gens, e.gO, ethylene dichloride, and compounds con-taining nitrogen, e.g., nitric oxide, can be used as the anti-catalysts. According to the patents, it is possible to employ mixtures of the individual anti-lS catalyst substances.
U. S. Patent 3,144,416 discloses a method ofmanufacturing silver catalysts to be used for the oxidation of oIefins. According to the patent, in order to increase the selectivity of the catalyst, a small quantity of halogen compound or nitrogen com-pound may be added to the reaction gas or catalyst.
EPO Patent 3642 and U. K. Patent Application GB
2 014 133A disclose processes of producing an olefin oxide by contacting an olefin with oxygen in the presence o~ a silver-containing catalyst and a chlorine-containing reaction modifier, for example, dichloroethane, methyl chloride, or vinyl chloride.
According to these references, the catalyst per-formance ~is improved, for example, the selectivity is increased, by contacting~the catalyst with a nitrate-or nitrite-forming substance, for example, a gas containing dinitrogen tetroxide, nitrogen dioxide and/or a nitrogen-containing compound, together with an oxidizing agent, such as nitric oxide and oxy-gen. The catalyst preferably comprises 3 to 50 per-~5015 ~2~

cent, more preferably 3 to 30 percent, by weight silver. According to the references, it is preferred that the catalyst should contain cations, for exam-ple, alkali and~or alkaline earth metal cations, as the corresponding nitrate or nitrite, particularly if the catalyst is treated with the nitrate- or nitrite-forming substance intermittently. According to the patents, suitable concentrations of the cations may be, for example, 5 x 10-5 to 2, preferably 5 x 10 4 to 2, more preferably 5 x ]0~4 to 0.5, gram equiva-lents per kilogram of catalyst. Suitably, Mo, K, Sr, Ca and/or Ba are present in amounts of 2 to 20,000, preferably 2 to 10,000, more preferably 10 to 3,000, microgram equivalents per gram of silver. In the processes of these two references, a diluent, for example, carbon dioxide, may be present and uncon-verted olefin may be recycled, suitably after removal of carbon dioxide.
Rumanian Pa~ent No. 53012, published December 2, 1971, discloses a process in which the catalyst is brought in direct contact w~th a gas ~ixture composed of 5-15 percent oxygen, 8-20 percent carbon dioxide, 60-80 percent nitrogen, completed by 1-5 percent nitrogen oxides.
U. ~. Patent 524,007 discloses a method for activating catalysts which may be accomplished by contacting the catalyst with an aqueous solution of a hydroxide of lithium, after the catalyst has first been treated with an "anti catalyst", such as ethylene dichloride or nitrogen dioxide. Accordin~
to the patent, the treatment may most advantageously be conducted simultaneously with the oxidation reaction of the olefins, inasmuch as the presence of very small amounts of anti-catalyst (less than 0~1 percent) increases the efficiency by limiting the ~ 2 9 3 formation of carbon dioxide.
Scientific literature is replete with examples of the use of alkali metals and alkaline earth metals and their cations to promote the efficiency of silver catalysts used in epoxidation reactions. Numerous examples may be found in literature regarding prefer-ence for the inclusion or exclusion oE one or several metals or cations in silver catalysts. Although many reports have indicated that no particular effective-ness is observed with one alkali metal or alkalineearth metal cation vis-a-vis another, several have suggested clear preferences for particular metal cations.
Potassium is well known as a catalyst promoter for the epoxidation o~ alkenes. One of the first patents to recognize potassium as a suitable promoter was U. S. Patent 2,177~361. According to this pat-ent, the catalyst may be promoted by the presence of very small proportions of alkali or alkaline earth metals, U. K. Patent Application 2,122,913A discloses a catalyst and a process for oxldation of ethylene in which an amount of alkali metal is deposited on the catalyst which removes substantially all acti~vity from the silver catalyst and then activity and selec-tivity are recovered by heating the catalyst in a nitrogen atmosphere.
When potassium is employed in the catalyst, it ; is generally introduced in conjunction with an an-ion. The choice of the anion has not always been regarded as significant. For example, U. S. Patents 3,962,136~, 4,010,115, 4,012j425, and 4,356,312 state that no unusual effectiveness is observed ~ith the ~use of any particular anion in the alkali metal salts used to prepare the catalysts and suggests that ni-~Z~3~
-17~

trates, nitrites, chlorides; iodides, bromates, et-cetera, may be used. Potassium nitrate was employed in the silver salt solution of Example 1 in each patent. According to the patents, from about 4.0 x 10 5 to about 8.0 x 10 3 gram equivalent weights of ionic higher alkali metal, e.g., rubidium, cesium or potassium or mixtures thereof, per kilogram of cata-lyst is deposited on the catalyst support simultan-eously with the deposit of silver. The amount of higher alkali metal preferably ranges from about 2.0 x 10 4 to about 6.5 x 10-3 gram equivalent weights per kilogram of finished catalyst. According to the patents, the amount of the higher alkali me~al (or metals) present on the catalyst surface is critical and is a function of the surface area of catalyst.
According to the patents, the alkali metal is present in final form on the support in the form of its ox-ide. U. 9. Patents 3,962,136, 4,010,115 and 4,01~,425 note tha~ the highest level of selectivity obtainable when potassium is employed typically is lower than that obtainable when rubidium or cesium is employed while U. S. Patent 4,356,312 notes that particularly good results are obtained with potas-sium.
U. S. Patent 4,066,575 notes that alkali metal nitrate is suitable for supplying an alkali metal promoter, but it notes that the anion associated with the promoter metal is not critical. U. S~ Patent 4,207,210 discloses a process for preparing an ethy-lene oxide catalyst in which higher alkali metals, such as potassium, rubidium and cesium, are deposited on a catalyst support prior to the deposition of silver. According to the patent, the amount of high-er alkali is a critical function of the surface area of the support. This patent also notes that no un-]~015 usual effectiveness is observed with the use of any particular anion in preparing the catalysts and lists nitrates as one type of salt that may be used. Car-bon dioxide and steam are listed as diluent ; 5 materials.
The use of potassium nitrate, however, to impart a promotin~ effect on the catalyst has been widely described. ~or example, U. S. Patent 4,007,135 lists a number of materials, including potassium, which can be used as promoters. According to the patent, in yeneral, 1 to 5,000, preferably 1 to 1,000, more preferably 40 to 500, and particularly 20 to 200, atoms of potassium are present per 1,000 atoms of silver. Suitably an aqueous solution of a compound, such as a chloride, sulfate, nitrate, nitrite, et-cetera, of the promoter is used for impregnation.
U. S. Patent 4,094,889 discloses a process for re-storing the selectivity of silver catalysts in which alkali metal may be introduced as a nitrate and in which the preferred content o~ potassium is in the range of 2 x 10-2 to 3 x 10-5 grams/square meter of surface area of support. U. S. Patent 4,125,480 discloses a process for reactivating used silver catalyst comprising (a) washing the used catalyst, and (b) depositiny from 0.00004 to 0.008~ preferably from 0.0001 to 0.002 gram equivalent weights per kilogram of catalyst of ions of one or more of the alkali metals, su~h as sodium, potassium, rubidium, or cesium. The ions of, e.g., potassium are deposited on the ca~alyst by impregnating it with a solution of one or~more compounds, such as potassium nitrate. U. S. Patents 4,~26,782, 4,235,757, 4,324,699, 4,342,667, 4,368,144, 4,455,3g2, Japanese Patent 56/89843, and U. K. Patent 1,571,123 suggest the use of potassium nitrate in various amounts 12~3~

Potassium nitrate may also be formed in situ when a carrier material is treated with certain amines in the presence of potassium ions, for example, when silver is introduced to a carrier material in a sil-ver-impregnating solution containing an amine and potassium ions, followed by roasting.
There has been some disclosure directed to cata-lysts for use in an ethylene oxide production system in ~hich silver is present in relatively large pro-portionst e.g., 35 percent or more. For example,U. S. Patents 3,565,828 and 3,654,318 disclose cata-lysts for the synthesis of ethylene oxide from oxygen and ethylene. According to the patents, the cata-lysts contain from 60 percent to 70 percent by weight of silver.
U. S. Patent 2,593,099 discloses a magnesium oxide-barium oxide silver catalyst support. Accord-ing to the patent, the conventional amount of silver is deposited on the support, namely, 2 to 50 percent, with the best results being obtained between 4 and 20 percent.
U. S. Patent 2,713,586 discloses a process for the oxidation of ethylene to ethylene oxide in which, according to the patent, the conventional amount of silver is deposited on the support, namely, 5 to 50 percent, with the best results being obtained between 4 and 20 percent.
U. S. Patent 3,793,231 discloses a process for the preparation of silver cataly~ts for the produc-tion of ethylene oxide in which the silver content ofthe catalysts generally range between lS to 30 per-cent by weight, preferably 19 to 27 percent by weight.
A large body of art directed to various aspects of alkene oxide production has been developed over ` -20~ 3~

the years since Lefort (U. S. Patent 1,998,~78).
Much of it is contradictory and incapable o~ re~on-ciliation. None of the art is believed to disclose or suggest a process for the high-efficiency epoxiaa-tion of alkene in which the catalyst includes anef~iciency-enhancing salt of a member of a redox-half reaction pair and which recogni~es the relationship between ~he amount of efficiency-e~hancing salt pro-vided in the catalyst and the concentration of carbon dioxide in th~ reaction inle~ stream and their e~fect on the activity and ef~iciency of th~ reaction sys-tem.

Disclosure of the Invention-Carbon dioxide, when present in a large enough quantity, can act as an activity-reducing compound in high eficiency p~ocesses for the epoxidation of alkene in which alkene and oxygen-containing gas are contacted in the presence of at least one gaseou~
efficiency-enhancing member of a redox-half reaction pair, e~g. r nitric oxide and/or nitrogen dioxide, carbon dioxide, and a supported silver catalyst which includes a salt of a member o~ a redox-half reaction 2~ pair, e.g., potassium nltrate. Carbon dioxide is generally continuously produced as a by-product of the alkene epoxidation reaction. ~5 a result, carbon : dioxide is generally contained in the reactor ef-~luent and at least some portion of it is normally returned to the reaction zone via the recycle streamc One way to avoid or limit the deleterious effects o~ carbon dioxide on the catalyst is to re-move the carbon dioxide from the recycle stream ~y, `: for example, use o a Benfield scrubber. Such an additional separa~ion step entails an additional 1~?3Z ~CI

capital expenditure, particularly with existing equipment designed for operation with lower efficien-cy catalysts where carbon dioxide is not a problem.
In commercial operations, removal of carbon dioxide S to a level at which the adverse effects of carbon dioxide are minimized can require a ~;ignificant capi-tal outlay which may not be cost-~ustifiable. The present invention is directed to a method of minimiz-ing the adverse effects of carbon dioxide on the catalyst by providing the efficiency-enhancing salt in an appropriate amount. The method can be carried out in conjunction with the removal of some carbon dioxide to control its concentration in the reaction zone.
The present invention provides a process then for the epoxidation of alkene comprising contacting alkene and oxygen-containing gas under epoxidation conditions in the presence of at least one gaseous efficiency-enhancing member of a redox-half reaction pair, carbon dioxidel and a supported silver cata-lyst. The catalyst comprises a catalytically-effec-tive amount of silver on a support and at least one efficiency-enhancing salt o a member of a redox-half reaction pair. ~he efficiency-enhancing salt is Z5 present in an amoun~ sufficient to provide an effi-ciency of at least about 84 percent under Standard Test Conditions I (as defined below) but below the amount which under Standard Test Conditions II (as defined below) would reduce the activity to less than 30 ~ 4 pounds of ethylene oxide per hour per cubic foot of catalyst, thereby reducing the activity~reducing effect of carbon dioxide in carrying out the process of the invention .

~;

-22~

Brief D~scription Of The Drawings:

Figure 1 is a flow chart of a process in accord-ance with the invention; and Figure 2 is a graph of (1) the activity and (2) the eficiency of the catalysts of Examples 1-4 ver-sus the concentration o~ the efficiency-enhancing salt contained in the catalyst when (a) no carbon dioxide is present in the reaction inlet stream, and (b) 3 volume percent of carbon dioxide is present in the reaction inlet stream.
.

Detailed Description Of The Invention:

The present invention is directed to processes for the epoxidation of alkene to form alkene oxide by contacting alkene and oxygen-containing gas under epoxidation conditions in the presence of a gaseous eficiency-enhancing member of a redox-half reaction pair, carbon dioxide, and a supported silver cata-lyst. The silver catalyst generally comprises a catalytically-effective amount of silver and an effi-ciency-enhancing salt o a member of a redox-half reaction pair on a porous support. The efficiency-enhancing salt is present in the catalyst in anamount sufficient to provide an eEiciency of at least about 84 percent under Standard Test Co~ditions ~I but below that amount which, under Standard Test Conditions II would reduce the activity to less than 4 pounds of ethylene oxide per cubic foot of catalyst per hour. Preferably, the amount of efficiency-enhancing salt is~below that amount which, under Standard Test Conditions II would reduce the activity to less than 6 pounds of ethylene oxide per cubic foot of catalyst per hour.

-23~ 3~

The following description of the preferred sys-tem for epoxidation of alkene in accordance with the present invention may be better understood by refer-ence to the flow chart in Figure 1.
In steady-state operation of the preferred sys-tem, a reaction inlet stream containing reactants, together with other gaseous materials as discussed below, is fed to a reactor at a controlled gas hourly space velocity (GHSV). The reactor may take a vari-ety of forms, but is preferably a collection of ver-tical tubes containing a supported silver catalyst.
The reaction inlet stream enters the reactor, passes through the catalyst, and exi~s the reactor as the reactor effluent. The desired product, e.g., ethy-lene oxide, is separated from the other components inthe reactor e~fluent, preferably by a scrubbing oper-ation. The remainder of the reactor effluent becomes the recycle stream. It is sometimes preferred to remove a portion of the material in the recycle stream in order to, for example, prevent buildup of certain materials in the system. The removal of material from the recycle stream may be selective, i.e., certain compounds may be removed from the re-cycle stream in greater proportions than other com~
pounds. The remainder of the recycle stream is us-ually combined with a makeup feedstream to form the reaction inlet stream. The reaction system will be ; discussed in greater detail below.
Although the present invention can be used with any size and type of alkene oxide reactor, including both fixed bed and fluidized bed reactors~ it is contemplated that the present invention will find most widespread application in standard fixed bed, multi-tubular reactors. These generally include wall-cooled as well as adiabatic reactors. Tube ~s~3;~
-24~

lengths typically range from about 5 to about 60 feet (1.52 to 18.3 meters), frequently from about 15 to about 40 feet (1.52 to 12.2 meters). The tubes gen-erally have internal diameters rom about 0.5 to about 2 inches (1.27 to 5.08 centimeters), typically from ab~ut 0.8 to about 1.5 inches (2.03 to 3.81 CentimeteES).
The catalyst generally comprises a support hav-ing catalyst material or a mixture of catalyst mater-ial and an efficiency-enhancing material impre~nated or coated on the support. The support can be gener-ally described as a porous, inorganic substrate which is not unduly deleterious to the performance of the system and .is preferably substantially inert toward the other materials in the system, i.e., the catalyst material, any other components present in the cata-lyst, e.g., eficiency-enhancing salt, and components in the reaction inlet stream. In addition, the sup-port should be able to withstand the temperatures employed within the reactor, as well as, of course, the temperatures employed in manufacturing the cata-lyst, e.g., if the catalyst material is reduced to its free metallic state by roasting. Suitable sup-ports for use in accordance with the present inven-tion include silica, magnesia, silicon carbide, zir-conia, and alumina, preferably alpha-alumina. The support preferably has a surface area of at least about 0.7 m2/g~ preferably in the range of from about 0.7 to about 16 m2/g, more preferably about 0.7 to about 7 m2/gO The surface area is measured by the B. E. T. nitrogen method described by Brunauer, Emmet and Teller in J. Am. Chem. Soc. 60, 309-316 (1938).
The support may be composed of a particulate matrix. In a preferred support, at least about 50 percent of the total~number of support particles ~ 3 having a particle size greater than about 0.1 micro-meter have at least one substantially flat major surface. The support particles are preferably formed into aggregates or "pills" of such a size and shape S that they are readily usable in commercially operated tubular reactors. These aggregates or pills general-ly range in size from about 2 millimeters to about 15 millimeters, preferably about 3 millimeters to about 12 millimeters. The size is chosen to be consistent with the type of reactor employed. In general, in fixed bed reactor applications, particle siæes rang-ing from about 3 millimeters to about 10 millimeters 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, especi-ally hollow cylinders.
The preferred support particles in accordance with the present invention have at least one substan-tially flat major surface and may be characterized as having a lamellate or platelet-type morphology. Some of the particles have two, or sometimes more, flat surfaces. The major dimension of a substantial por-tion of the particles having platelet-type morphology is less than about 50 micr~ns, preferably less than about 20 microns. When alpha-alumina is employed as the support material, the platelet-type particles frequently have a morphology which approximates the shape of hexagonal plates.
The carrier materials of the present invention may generally be described as porous or microporous and they generally have median pore diameters of from about 0.01 to about 100 microns, preferably about 0.5 to about 50 microns, and most preferably about 1 to ~L293Z ~

about 5 microns. Generally, they have pore volumes of about 0.6 to about 1.4 cc/g, preferably about 0.8 to about 1~2 cc/g. Pore volumes may be measured by any conventional technique, such as conventional mercury porosity or water absorption techniques.
Generally improved results have been demon-strated when the support material is composition~pure and also phase-pure. By "composition-pure" is meant a material which is substantially a single substance, such as alumina, with only trace impurities being present. The term "phase-pure" refers to the homo-geneity 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, alpha-alumina.
Under some conditions even small amounts of leachable sodium can adversely affect the service life of the catalyst. Notably improved results have been observed when the support contains less than about 50 parts per million (ppm) by weight, prefer-ably less than 40 ppm, based on the weight of the total catalyst. The term leachable sodium, as used herein, refers to sodium which can be removed from the support by immersing the support in a 13 percent by volume nitric acid solution at 90 degrees C for one hour. Suitable alpha-aluminas~having concentra-tions of sodium below 50 ppm may be obtained commer-cially from suppliers such as the ~orton Company.
Alternatively, suitable alpha-alumina support ma`ter-ials may be prepared so as to obtain leachable sodium concentrations below 5~ ppm by the method described by Weber et al in U. S~ Patent 4,379,134.
A particularly preferred support is a high-purity alpha-alumina support, having platelet 1~3Z~3 morpht~lo~y, ;.)f the type disclosed in Canadian Applicat iOII No 515,865~ filed August 13, 1986, in ttl~ nan-e of TAom~s M. NG~ermann, entitled "ImprGved C~t~lytic System for Ep(xi~ation of Alkenes".
The present invenl iOIl includes in the catalyst at least one efficiency-enhancing salt of a member of a redox-half reaction pair. The term "redox-half reaction" is defined herein to mean half~reactions such as those found in equations presented in tables of standard reduction or oxidation potentials, also known as standard or single electrode potentials. These equations are found in, for instance, "Handbook 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 D 155-162 (1984). The term "redox-half reaction pair" refers to the pairs of atoms, molecules or ions, or mixtures thereof, which undergo oxidation of reduction in such half-reaction equations, A member of a redox-half reaction pair is, therefore, one of the atoms, molecules or ions that appear in a particular redox-half reaction equation. The term redox-half reaction pair is used herein to include those members of the class of substances which provide the desired performance enhancement rather than a mechanism of the chemistry occurring. Preferably, such compounds, when associated with the catalyst as salts of members of a redox-half reaction pair, are salts in which the anions are oxyanions, preferably an oxyanion of polyvalent atom, i.e., the atom of the anion to which oxygen is bonded is capable of existing, when bonded to a dissimilar atom, in the different valence states.

-28- 12~3~

The preferred efficiency-enhancing salts are potas-sium nitrate and potassium nitrite.
The catalysts of the present invention are preferably prepared by depositing catalyst material and at least one efficiency enhancing salt, sequentially or simultaneously, on and/or within a solid porous ~upport. The preferred catalyst material in accordance with the present invention comprises silver, preferably of a particle size less than about 0.5 micron. Any known method of introducing the catalyst material and efficiency-enhancing salt into the catalyst support may be employed, but it is preferred that the support is either impregnated or coated. The more preferred of these is impregnation wherein, in general, a solution of a soluble salt or complex of silver and/or one or more efficiency-enhancing salt 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~ and/or efficiency-enhancing salt~containing impregnation solution.
Sequential impregnation means that silver is first deposited within the carrier in one or more impregnation steps, and then salt is deposited in a separate impregnation step.
One aspect of the present invention involves the beneficial effects observed when the catalyst con-tains high concentrations of silver. In order to provide such a catalyst by impregnation, it has been found that it is preferable to deposit the silver via several impregnation steps. Thus, if a high silver-content catalyst, e.g., a catalyst containing 30 or more percent silver, is desired and a sequential impregnation procedure is to be used, a four-step -29- ~3~ ~

process may be employed. Such a process would in-volve three silver-only impregnation steps ~ollowed by one salt-only impregnation step.
In general, a silver-only impregnation step is carried out by first immersing the support in a sil-ver-containing impregnation solution, preferably by placing the support particles in a vessel, evacuating the vessel and then adding the impregnation solu-tion. The excess solution may then be allowed to drain off or the solvent may be removed by evapora-tion under reduced pressure at a suitable tempera-ture. Typically, a silver-containins solution is prepared by dissolving silver oxide in a sui~able solvent or complexing/solubilizing agent as, for example, a mixture of water, ethylenediamine, oxalic acid, silver oxide, and monoe~hanolamine.
After impregnation, the silver-impregnated car-rier particles are treated to convert silver salt to silver metal to effect deposition of silver on the surace o~ the support. This may be done by treating the impregnated particles with a reducing agent, such as oxalic acid, alkanolamine or by roasting at an elevated temperature on the order of about 100 to about 900 degrees C, preferably about 200 to about 650 degrees C~ to decompose the silver compound and reduce the silver to its free metallic state. ~he duration of roasting is generally for a period of from about 1 to about 10 minutes, with longer times for lower temperatures, depending on the temperature used. As used herein, the term "surface~, as applied to the support, not only includes the external sur-faces of the carrier but also the internal surfaces, that is, the surfaces defining the pores or internal portion of the support particles.

-30~ 3~

The efficiency-enhancing salt may be introduced into the catalyst in any suitable manner. In gener-al, the pr~ferred amount of efficiency-enhancing salt can be deposited in one impregnation step. After immersion of the silver loaded support in the effi-ciency-enhancing salt impregnation solution, the excess solution is generally drained and the silver-and efficienoy enhancing salt-containing support is dried, for example by heating to from 80 to 200 de-grees C. When more than one salt of a member of aredox-half reaction pair is employed, the salts may be deposited together or sequentially.
Concurrent or coincidental impregnation means that generally the final (perhaps the only) impregna-tion step involves immersion of the support in animpregnation solution which contains silver as well as one or more efficiency-enhancing salts. Such an impregnation step may or may not be preceded by one or more silver-only impregnation steps. Thus, to make a high silver-content catalyst by coincidental impregnation, several silver-only impregnation steps might be carried out, ollowed by a silver- and effi-ciency-enhancing salt-impregnation step. A low sil-ver-content catalyst, e.g., from about 2 to about 20 weight percent silver, may be made by a single sil-ver- and effîciency-enhancing salt-impregnation step. For the purposes oE this invention, these two sequences are both referred to as concurrent or coin-cidental impregnation.
The three types of impregnation solutions, name-ly, silver-containing, efficiency-enhancing salt-containing, and silver- and efficiency-enhancing salt-containing, are discussed in more detail below.
There are a large number of suitable solvents or complexing/solubilizing agents which may be used to -31- ~2~3Z~

form the silver-containing impregnating solution. A
suitable solvent or complexing/solubilizing agent, besides adequately dissolving the silver or convert-ing it to a soluble form, should be capable of being readily removed in subsequent stepsl either by a washing, volatili2ing or oxidizing procedure, or the like. It is aIso generally preferred that the sol-vents or complexing/solubilizing agents be readily ;~ miscible with water since aqueous solutions may fre-quen~ly be employed.
~ mong the materials found suitable as solvents or complexing/solubilizing agents for the preparation of the silver-containing solutions are alcohols, including glycols, such as ethylene glycol, ammonia~
amines and aqueous mixtures of amines, such as ethy-lenediamine and monoethanolamine, and carboxylic acids, such as oxalic acid and lactic acid.
The particular silver salt or compound used to form the silver-containing impregnating solution in a solvent or a complexing/solubilizing agent is not particularly critical and any known silver salt or compound generally known to the art which ls soluble in and does not react with the solvent or complex-~` ing/solubilizing agent may be employed. Thus, the silver may be introduced into the solvent or complex-ing/solubilizing agent as an oxide, a salt, such as a nitrate or carboxylate, for example, an acetate, propionate, butyrate, oxalate, lactate, citrate, phthalake, generally the silver salts of higher fatty acids, and the like.
Materials which may be employed in the efficien-cy-enhancing salt-containing impregnation solution to act as a solvent for the efficiency-enhancing salt include generally any solvent capable of dissolving the salt, which solvent will neither react with the 12~2 ~3 silver nor leach silver from the support. Aqueous solutions are generally preferred but organic li-quids, such as alcohols, may also be employed.
In order to perform coincidental impregnation, S the efficiency-enhancing salt and the silver catalyst material must both be soluble in the solvent or com-plexing/solubilizing liquid used.
Suitable results have been obtained with both the sequential and coincidental procedures. Some results have indicated that greater amounts of silver with more uniform distribution of silver throughout the pill can be obtained by three or more silver-impregnation cycles. High silver-containing cata-lysts prepared by a coincidental impregnation techni-que generally provide better initial performance thanthose prepared by a sequential technique.
If the catalyst material is to be coated on the catalyst support rather than impregnated in the sup-port, the catalyst material, e.g., silver, is pre-formed or precipitated into a slurry, preferably an aqueous slurry, such that the silver particles are deposited on the support and adhere to the support surface when the carrier or support is heated to removed the liquids present.
Z5 The concentration of silver in the finished catalyst may vary from about 2 percent to 60 percent or higher, by weight, based on the total weight of the catalyst, more preferably from about 8 percent to about 50 percent, by weight. When a high silver content catalyst is employed, a silver concentrationrange of from about 30 to about 60 percent, by weight, is preferred. When a lower silver content catalyst is used, a preferred range is from about 2 to about 20 weight percent. The silver is preferably distributed relatively evenly over the support sur -33- ~3~

faces. The optimum silver concentration for a par-ticular catalyst must take into consideration per-formance characteristics, such as catalyst activity, system efficiency and rate of catalyst aging, as well as the increased cost associated with greater concen-trations of silver in the catalyst material. The approximate concentration of silver in the finished catalyst can be controlled by appropriate selection of the number of silver-impregnation steps and of the concentr~tion of silver in the impregnation solution or solutions.
When more than one salt of a member of a redox-half reaction pair is employed, the salts may be deposited together or sequentiallyO I~ is preferred, however, to in~roduce the salts to the support in a single solutlon, rather than to use sequential treat-ments using more than one solution and a drying step between impregnation steps, since the latter tech-nique may result in leaching the first introduced salt by the solution containing the second salt.
Concurrent or coincidental impregnation may be ac-complished by forming an impregnating solution which contains the dissolved efficiency-enhancing salt of a member of a redox-half reaction pair as well as sil-ver catalyst material. Silver-first impregnation can be accomplished by impregnating the support with the silver-~ontaining solution, drying the silver-con-taining support, reducing the silver, and impreg-nating the support with the efficiency-enhancing salt solution.
Reaction conditions maintained in the reactor during operation of the process are those typically used in carrying out epoxidation reactions. Tempera-tures within the reaction zone of the reactor gener-ally range from about 180 to about 300 degrees C and15015 3~

pressures generally range from about 1 to about 30 atmospheres, typically from about 10 to about 25 atmospheres. The gas hourly space velocity (GHSV) may vary, but it will ~enerally range from about 1,000 to about 16,000 hr 1.
'9Standard Test Conditions I~ (STC-I3 and "Stand-ard TPst Conditions II~ (STC II), as used herein, refer to two sets o~ conditions under which catalysts may be tested to determine if they contain the 10 requisite amount of efficiency-enhancing salt. Note that STC-I and STC-II are carried out using ethylene as the alkene in determining the operable level of the efficiency-enhancing salt. ~owever, the catalyst may be used in systems for the epoxidation of other alkenes, e.g., propylene, as described herein. The reactor used in both STC-I and STC-II is a micro-reactor, approximately 14 cm in leng~h and having an inside diameter of approximately 7.8 mm. The micro-reactor is charged with sufficient fine particulate catalyst comprisin~ particles of a size Oe from about 0.5 to about 1.5 mm, i.e., from about 35 to about 12 U. S. sieve to provide a catalyst bed height of 5 centimeters~
No recycle stream is used under STC-I and STC-II, so the makeup feedstream is also the reactioninlet stream. Measurements of efficiency (STC-I) and activity ~STC-II1 are made 24 hours af~er the respec-tive test has begun~ i.e., 24 hours a~ter the respec-tive microreactor has been brought on stream under the respective conditions described below.
The conditions under STC-I comprise a tempera-ture of 240 degrees C, a pressure of 150 psigl. a GHSV

_35~ 3~

of 8,000 hr l, a reaction inlet stream containing 30 volume per-cent ethylene, 8 volume percent oxygen, 1 volume percent ethane, 26 parts per milliont by vol-ume, ethyl chloride, 10 parts per million, by volume, nitric oxide, and the balance nitrogen.
The conditions under STC-II comprise the same set of conditions recited above for STC-I except that, under STC-II, the temperature is 260 degrees C, the reaction inlet stream contains about 3 volume percent carbon dioxide and 33 parts per million, by volume, nitric oxide with the amount of nitrogen adjusted accordingly.
The product, for example, ethylene oxide, is recovered from the reactor effluent, e.g., by an absorption process. One such method comprises sup-plying the reactor effluent stream to the bottom of an absorption column while adding a solvent, for example, water, to the top of the absorption col-umn. The solvent preferably absorbs the ethylene ~ oxide and carries it out of the bottom of the ab-sorption column, while the remainder of the reaction effluent passes out of the top of the absorption column to form the recycle stream. The desired pro-duct is thereafter recovered, for example, by passing the solvent and absorbed product through a stripper.
As noted above, it may be preferable to remove a portion of the recycle stream prior to returning the recycle stream to the reaction zone. It is generally preferable to selectively remove certain compounds.
An absorption column or other types of separation means can be used to provide a selective purge.
The recycle stream generally contains the dilu-ents and inhibitors fed to the system, unreacted alkene and oxygen, together with by-products of the reaction, such as carbon dioxide and water, and any ~3~

minor amount of alkene oxide which is not recovered as product. After removal of the purge stream, the recycle stream is returned to the reaction zone, preferably being mixed with the makeup feedstream - 5 prior to or as it enters the reaction zone.
The makeup feedstream replaces reactants, i.e., alkene and oxygen-containing gas, as well as other materials not contained in the recycle stream in sufficient amounts. Alkene, as used herein, refers to cyclic and acyclic alkenes which are in a gaseous state or have significant vapor pressures under epox-idation conditions. Typically these compounds are characterized as having on the order of 12 carbon atoms or less and which are gaseous under epoxidation conditions. In addition to ethylene and propylene, examples of alkenes which may be used in the present invention include such compounds as butene, dodecene, cyclohexene, 4-vinylcyclohexene, styrene, and norbor-nene.
The oxygen-containing gas employed in the reac-tion may be defined as including pure molecular oxy-gen, atomic oxygen, any transient radical species derived from atomic or molecular oxygen capable of existence under epoxidation conditions, mixtures of anoth~r gaseous substance with at least one of the foregoing, and substances capable of forming one of the foregoing under epoxidation conditions. ~uch oxygen-containing gas is typically oxygen introduced to the reactor either as air, commercially pure oxy-gen or any other gaseous substance which forms oxygenunder epoxidation conditions.
The makeup feedstream may also contain one or more additives, for example, a performance-enhancing gaseous halide, preferably an organic halide, includ-ing saturated and unsaturated halides, such as 1,2-~37~ ~2~

dichloroethane, ethyl chloride, vinyl chloride, methyl chloride, and methylene chloride, as well as aromatic halides. The performance-enhancing gaseous halide preferably comprises l,2-dichloroethane and~or ethyl chloride. In addition, a hydrocarbon, such as ethane, can ~e included in the makeup feedstream.
~he makeup feedstream may also contain a diluent or ballast, such as nitrogen, as is the case when air is used as the oxygen-containin~ gas.
The makeup feedstream generally also includes at least one gaseous efficiency-enhancing member of a redox-half reaction pair. The phrase "gaseous effi-ciency-enhancing compoundi', as used herein, is an alternative expression for the expression "at least one gaseous efficiency-enhancing member of a redox-half reaction pair." Both phrases are therefore meant to include single gaseous efficiency-enhancing members oE redox-half reaction pairs as well as mix-tures thereo. The term "redox-half reaction pair"
has essentially the same meaning as defined in con-nection with efficiency-enhancing salts, above. The preferred gaseous efficiency-enhancing materials are, preferably, compounds containing oxygen and an ele-ment capable of existing in more than two valence s~.ates. Examples of preferred gaseous efficiency-enhancing members of redox-half reaction pairs in-clude NO, NO2, N2O4, N2O3, any substance capable of forming gaseous NO and/or NO2 under e oxidation con-di~ions, or mixtures thereof. In addition, mixtures of one of the compounds listed above, particularly NO, with one or more of PH3, CO, SO3, and SO2 are suitable. Nitric oxide is particularly preferred.
In some cases it is preferable to employ two members of a particular half-reaction pair, one in the efficiency-enhancing salt and the other in the -38- ~Z~ J'~

gaseous efficiency-enhancing compound employed in the feedstream, as, for example, with a preferred combin-ation of KNO3 and NO. Other combinations, su~h as KNO3/N2O3, KNO3/NO2, and KNO~/N2O4 may also be em-ployed in the same system. In some instances, thesalt and the gaseous members may be found in half-reactions which represent the first and last reac-tions in a series of half-reaction equations of an overall reaction.
The gaseous efficiency-enhancing member of a redox-half reaction pair is preferably present in an amount that favorably affects the efficiency and/or the activity. The precise amount is determined, in part, by the particular efficiency-enhancing salt employed and the concentration thereo , as well as the other factors noted above which influence the amount of efficiency-enhancing salt. Suitable ranges of concentration for the gaseous efficiency-enhanciny compound are generally dependent upon the particular alkene which is being epoxidized, larger amounts of the gaseous efficiency-enhancing compound generally being preferable with higher alkenes. For example, in an ethylene epoxidation system, a suitable range of concentration for the gaseous eEficiency-enhancing member of a redox-half reaction pair i5 typically about 0.1 to about 100 ppm by volume of the reaction inlet stream. Preferably, the gaseous efficiency-enhancing compound is present in the reaction inlet stream in an amount within the range of from 0.1 to 80 ppm, by volume, when about 3 percent, by volume, carbon dioxide is present in the reaction inlet stream. When nitric oxide is employed as the gaseous efficiency-enhancing compound in an ethylene epoxida-tion system, it is preferably present in an amount of from about 0.1 to about ~0 ppm by volume. When about 3Z~53 3 percent, by volume, carbon dioxide is present in the reaction inlet stream, nitric oxide, if used as the gaseous efficiency~enhancing compound, is prefer-ably present in an amount of from about 1 to about 40 ppm. On the other hand, in a propylene or higher alkene epoxidation system a suitable concentration of the gaseous efficiency-enhancing compound is typical-ly higher, e.g., from about 5 to about 2,000 ppm by volume of the reaction inlet stream when using nitro-gen ballast.
Similarly, the concentration of the performance-enhancing gaseous halide, if one is used, is depen-dent, inter alia, upon the particular alkene which is being oxidized. A suitable range of concentration for gaseous halide in an ethylene epoxidation system is typically from about 0.1 to about 60 pp~ by volume of the ~eaction inlet stream. A suitable concentra-tion for gaseous halide in the reaction inlet stream in a prop~lene epoxidation system is typically from about 5 to about ~,000 ppm by volume when using ni-trogen ballast. The preEerred concentration of gas-eous halids, if one is used, varies depending on the particular compounds used as the e~ficiency-enhancing salt an~ the gaseous efficiency-enhancing compound ` 25 and the concentrations thereof, as well as the other factors noted above which influence the preferred amount of efficiency-enhancing salt.
The ranges for the concentration of alkene, oxygen, hydrocarbon, carbon dioxide and nitrogen or other ballast gas such as methane, in the reaction inlet stream, are dependent upon the alkene being epoxidized. The tables below show typical ranges for the materials~(other than the gaseous halide and the efficiency-enhancing compound) in the reaction inlet stream for the epoxidation of ethylene (Table A) and ~ -40- lZ~32'~

.
propylene (Table B).
'.~
; Table A

5 ComPOnent Concentration ~, .
Ethylene at least about 2, o~ten about 5 ~ to about 50, :~ 10 : volume percent Oxygen about 2 to about 8 volume percent 15 Hydrocarbon about 0 to about 5 ` volume percent . ,~
Carbon Dioxide up to about 7 volume percent Nitrogen or other remainder ballast gas, e.g., methane `

::: : ~ :
:: : 30 ~:
:

:: : :
~: : 35 --41~ 3~

Table B

Component Concentration -5 Propylene about 2 to about 50 volume percent Oxygen about 2 to about lO
volume percent Hydrocarbon about 0 to about S
volume percent Carbon ~ioxide up to about 15 lS volume percent Nitrogen or other remainder ballast gas, e.g~, methane The ranges set out in Table B for the concentra-tion of materials in the reaction inlet stream may be useful fo.r epoxidation of higher alkenes, e.g., al-ken~s having from 4 to 12 carbon atoms.
The amount of the efficiency-enhancing salt of a member of a redox-half reaction pair present in the catalyst directly affects the activity and/or effi-ciency of the epoxidation reaction. The most pre~er-able amount of the salt of a member of a .redox-half reaction pair varies depending upon the alkene being ~: epoxidized, the compound used as the gaseous effi-: ciency-enhancing member of a redox-half reaction pair, the concentration of components in the reaction inlet stream, particularly the gaseous efficiency-enhancing compound and carbon dioxide, the amount of lSOlS

-42~ 3Z

:`:
., silver contained in the catalyst, the surface area, morphology and type of the support, and the process conditions, e~g., gas hourly space velocity, temper-ature, and pressure. The approximate concentration of efficiency-enhancing salt in the finished catalyst can be controlled by appropriate selection of the concentration of efficiency-enhancing salt in the salt impregnation solution. Operable amounts of the efficiency-enhancing salt can be deter~ined by carry-ing out tests on similar catalysts containinq varyingamounts of the salt, i.e., by traversing across a range of salt concentrations from relatively too small to relatively too high an amount.
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 wi~h 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. ~or this reason, the amount of the ef~iciency-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 is also present in the catalyst in an amount roughly proportional (on a molar basis) to the cation.
The efficiency-enhancing salt is preferably provided in such an amount, in an ethylene epoxida-"

-43~

tion system, that the finished catalyst contains from ~bout 0.01 to about 0.7 percent, by weight, of ~he cation of the salt, based on the total weight of the catalyst, more preferably from about 0.02 to about ~-5 0.3 weight percent, most preferably from about 0.05 ¦to about 0.1 weight precent. The preferred salt is potassium nitrateO The approximate eoncentration of efficiency-enhancing salt in the finished catalys~
can be controlled by appropriate selection of the concen~ration o efficiency-enhancing salt in the salt-impregnation solution~

E mples 1 through 4:

Exam~les 1 throuyh 4 were carried out in a 14 ` centimeter long tubular, s~ainless steel micro-reactor, having a 7~8 millimeter inside diameter and a 5 centimeter long catalyst bed. The microreactor temperature was controlled in each case by tempera-ture controllers and a rectangular oven with air circulation. Approximately 1.1 grams of catalyst in the form of crushed pellets O.S to 1.5 mm in size was introduced into the microreactors in the amounts specified in Table 1. Prior to initiating the reac-tion, the reactor with the catalyst in place washeated to 200 degrees Centigrade in a flowing nitro-gen atmosphere. The ~ests were then carried out with a process gas mi~ture con~aining 30 vol~me ~ercent ethylene, 8 volume percent oxygen, 1 volume percent eth~ne, 26 parts per million by volume ethyl chloride, 0 or 2.99 volume percent carbon dioxide as specified in Table II,10 par~s per million by volume nitric oxide when no carbon dioxide is present in the ~rocess gas mixture, 33 parts per million by volume nitric oxide when 2.99 volume percent car~on dioxide `5015 2~0 is present, and the balance nitrogen. The tests were carried out at a pressure of 150 psig and a gas hour-ly space velocity (GHS~) of 8,000 hr 1.
Examples 1 through 4 each correspond to experi-mentation with a separate catalyst. The catalysts were prepared by coincidental impregnation. An im-; pregnating solution was prepared by dissolving 23.0 grams ethylenediamine with 60.0 grams of distilled water and stirring for a period of about 10 min-utes. ro the stirred solution was slowly added 23.0 grams of oxalic acid dihydrate. The resulting solu-tion was stirred for 10 minutes. To this solution were added, in portions, 43.3 grams of silver ox-ide. The resulting solution was therea~ter stirred for one hour, completely dissolving the silver ox-ide. To the resulting solution were added 8.1 grams of monoethanolamine, followed by stirring for an additional 10 minutes. The resulting solution was divided into 4 equal parts, each part to be used for one of the four examples. To each part was added the following amount of a potassium nitrate in water solution having a potassium concentration of 0.05 grams of potassium per gram of solution:

Example 1: 0.56 grams KNO3 solution ~xample 20 1.13 grams KNO3 solution Example 3: 2.26 grams KNO3 solution Example 4: 6.77 grams RNO3 solution Each soLution was then diluted with distilled water to 31.25 cubic centimeters.

" .

-4S~ Z ~0 High-purity alpha-alumina support pellets (13.8 grams), having a surface area of 1.12 m2/g, a porosity of 0.78 cc/g and a platelet morphology of the type disclosed in Canadian Application No.
515,865, were placed in a tube which was then evacuated, following which the support pellets were impregnated by immersing them in impregnating solution formed as described above for one hour. The excess impregnation solution was then drained. The resulting pellets were then belt-roasted at 500 degrees C in a 66 SCFH air flow for 2.5 minutes. The amount of potassium and silver in each of the four catalysts are set out in Table I.
In testi.ng each catalyst, the activity and eficiency were determined at specified reaction temperatures and with specified amounts of carbon dioxide in the reaction inlet stream, i.e., either 0 volume percent or 2.99 volume percent carbon dioxide. The results are set out in Table II below and are depicted in Figure 2.
_ABLE I
Example Catalyst Amount of Amount of Reactor Silver In Potassium Charge Catalyst In Catalyst (grams) (weight (weight percent) (Percent) 1 1.1 19.8 0.05 2 1.1 19.2 0.11 3 1.1 18.6 0.21 4 1.1 19.7 0.~6 15~15 , ~.

-46- 1Z~3~

TABLE II

Volume Parts Per Temper- Activity Efficiency ; ; Percent Million, ature (Pounds (%) Of CO2- By Vol- Within Of Ethy In Re- ume, Of Reac- lene ; action Nitric tion Oxide Pro-Inlet Oxide In Zone duced Per Stream Reactlon (C) Hour Per Inlet Cubic Foot ; Stream Of Cata-. lyst) .... , ._. . _ , ._ . ___ Example 1 0.0 10 * 15.7 85.5 ~.99 33 258.87.3 87.3 Example 2 0.0 10 239.6L7.0 90.7 2~9g 33 25g.62.9 88.0 Example 3 0.0 10 238.115.3 90.8 2.99 33 259.41.7 86.7 Example 4 0.0 10 237.516.0 91.1 2.99 33 259.2 ;1.7 ~ 85.7 : 30 : *Activity and efficiency corrected to a temperature of 240 degrees C from data at 225.8 degrees C using the standard Ar~henius temperature dependence as determined from experimentation.

~2~3;~

As can be seen from Table 1 and Figure 2, foL catalysts of the type described in Examples 1-4, th~ level of potassium which provides an efficiency of at least 84 percent is a minimum of about 0.05 weight percent while the maximum amount which can be present when the reaction inlet stream contains 3 volume percent carbon dioxide without reducing the activity to less than 4 pounds of ethylene oxide percubic foot of catalyst per hour was about 0.1 weight percent.
Example 5 - Coinciden-tal or Coimpregnation Method of Preparation Of A Potassium Nitrate-Containing ~pE~ d High Silver Concentration Catalyst:
A first silver-containing impregnation solution was prepared by dissolving 1,292.8 grams ethylene-diamine with 1.281.6 grams of distilled water and stirring for a period of 10 minutes. To the stirred solution were slowly added 1,294.8 grams of oxalic acid dlhydrate. The resulting soluti.on was stirred for 10 minutes. To this solution were added, in portions, 2,268.4 grams of Ag2O. The resulting silver-containing solution was thereafter stirred for an additional hour and 453.6 grams of monoethanolamine were 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 5,000 ml by addition of distilled water.
High-purity alpha-alumina support pellets (1,925.4 grams) having platelet morphology of the type disclosed in Canadian Application No. 515,865, having a surface area of ~Z~Z ~0 about 1.2 m2/g and a porosity of about 0.8 cc/g were ` placed in a tube which was then evacuated, following which a first impregnation was conducted by immersing the support particles in the first silver-containing impregnation solution formed as described above, for one hour. Excess impregnation solution was then , drained and the resulting pellets were ~hen belt-- roasted at 500 degrees C in a 66 SCF~ air flow for 2.5 minutes. The resulting material contained 24.9 percent silver by weight.
A co-impregnation solution was prepared by plac-ing 1,260.5 grams of ethylenediamine into a 5,000 ml beaker and mixing therewith 1,249.6 grams of dis-tilled water to form a solution. To the stirred solution were slowly added 1,262.4 grams of oxalic acid dihydrate and, with continuous stirring, 2,211.7 grams of silver oxide were slowly added. When dis-solution was complete, 442.3 yrams of monoethanol-amine were added direc~ly to the solution. To the `i 20 silver-containing solution were added 26.4 grams o~
potassium nitr~te dissolved in 50 milliliters of distilled water. To the resulting solution was added sufficient water to dilute the solution to 4,875 ml. The silver-impregnated catalyst pellets (2,495.6 grams) were impregnated in a manner similar to the first impregnation described above. The resulting catalyst contained 39.8 weight percent silver and 0.098 weight percent potassium.
: This: catalyst may be used in carrying out the process of this invention.
~`

Claims (20)

1. A process for the epoxidation of alkene to form alkene oxide comprising contacting said alkene and oxygen-containing gas under epoxidation condi-tions in the presence of at least one gaseous effici-ency-enhancing member of a redox-half reaction pair, carbon dioxide, and a supported silver catalyst, said catalyst comprising a catalytically effective amount of silver and at least one efficiency-enhancing salt of a member of a redox-half reaction pair on a sup-port, wherein said efficiency-enhancing salt is present in an amount sufficient to provide an effi-ciency of at least about 84 percent under Standard Test Conditions I but below the amount which under Standard Test Conditions II would reduce the activity to less than 4 pounds of ethylene oxide per cubic foot of catalyst per hour, thereby reducing the acti-vity-reducing effect of carbon dioxide in carrying out said process.

2. A process for the epoxidation of ethylene to form ethylene oxide comprising contacting said ethy-lene and oxygen-containing gas under epoxidation conditions in the presence of at least one gaseous efficiency-enhancing member of a redox-half reaction pair, carbon doioxide, and a supported silver cata-lyst, said catalyst comprising a catalytically effec-tive amount of silver and at least one efficiency-enhancing salt of a member of a redox-half reaction pair on a support, wherein said efficiency-enhancing salt is present in an amount sufficient to provide an efficiency of at least about 84 percent under Standard Test Conditions I but below the amount which under Standard Test Conditions II would reduce the activity to less than 4 pounds of ethylene oxide per cubic foot of catalyst per hour, thereby reducing the activity-reducing effect of carbon dioxide in carry-ing out said process.

3. A process for the epoxidation of propylene to form propylene oxide comprising contacting propy-lene and oxygen-containing gas under epoxidation conditions in the presence of at least one gaseous efficiency-enhancing member of a redox-half reaction pair, carbon dioxide, and a supported silver cata-lyst, said catalyst comprising a catalytically effec-tive amount of silver and at least one efficiency-enhancing salt of a member of a redox-half reaction pair on a support, wherein said efficiency-enhancing salt is present in an amount sufficient to provide an efficiency of at least about 84 percent under Standard Test Conditions I but below the amount which under Standard Test Conditions II would reduce the activity to less than 4 pounds of ethylene oxide per cubic foot of catalyst per hour, thereby reducing the activity-reducing effect of carbon dioxide in carry-ing out said process.

4. The process of claim 1, 2 or 3 wherein said at least one gaseous efficiency-enhancing member of a redox-half reaction pair comprises nitric oxide, nitrogen dioxide, N2O3, N2O4, a gas capable of generating nitric oxide and/or nitrogen dioxide under epoxidation conditions, or mixtures thereof.

5. The process of claim 1, 2 or 3 wherein said at least one gaseous efficiency-enhancing member of a redox-half reaction pair comprises nitric oxide, nitrogen dioxide, N2O3, N2O4, a gas capable of generating nitric oxide and/or nitrogen dioxide under epoxidation conditions, one or more of phosphine, carbon monoxide, sulfur dioxide, sulfur trioxide, or mixtures thereof.

6. The process of claim 1, 2 or 3 wherein said at least one gaseous efficiency-enhancing member of a redox-half reaction pair comprises nitric oxide.

7. The process of claim 1, 2 or 3 wherein said at least one efficiency-enhancing salt of a member of a redox-half reaction pair comprises potassium nitrate.

8. The process of claim 1, 2 or 3 wherein said at least one efficiency-enhancing salt of a member of a redox-half reaction pair comprises potassium nitrate and potassium is present in said catalyst in an amount of from about 0.03 to about 0.3 percent, by weight, based on the weight of said catalyst.

9. The process of claim 2 wherein said contacting occurs in a reaction zone, the gaseous compounds are fed to said reaction zone in a reaction inlet stream, and the amount of carbon dioxide present in said reaction inlet stream comprises up to 7 volume percent.

10. The process of claim 3 wherein said contacting occurs in a reaction zone, the gaseous compounds are fed to said reaction zone in a reaction inlet stream, and the amount of carbon dioxide present in said reaction inlet stream comprises up to 15 volume percent.

11. The process of claim 1, 2 or 3 wherein said contacting occurs in the presence of a performance-enhancing gaseous halide.

12. The process of claim 1, 2 or 3 wherein said contacting occurs in the presence of a performance enhancing gaseous halide comprising ethyl chloride, 1,2-dichloroethane or mixtures thereof.

13. The process of claim 12 wherein said catalyst comprises from about 8 to about 50 weight percent silver.

14. The process of claim 1, 2, or 3 wherein said gaseous efficiency-enhancing member of a redox-half reaction pair and said efficiency-enhancing salt of a member of a redox-half reaction pair comprise members of the same redox-half reaction pair.

15. The process of claim 1, 2, or 3 wherein said catalyst is provided in a fluidized bed.

16. The process of claim 1, 2, or 3 wherein said catalyst is provided in a fixed bed.

17. A supported silver catalyst for the epoxi-dation of alkene to form alkene oxide comprising at least 30 weight percent silver and an efficiency-enhancing amount of a salt of a member of a redox-half reaction pair on a support.

18. The supported silver catalyst of claim 17 wherein said silver is present in an amount of from about 30 to about 60 weight percent.

19. The supported silver catalyst of claim 18 wherein said salt is present in an amount of from about 0.01 to about 0.7 percent by weight.

20. The supported silver catalyst of claim 19 wherein said salt is potassium nitrate.