CA1089877A - Ethylene glycol process - Google Patents

Abstract

GLYCOL ALDEHYDE AND ETHYLENE GLYCOL PROCESSES
ABSTRACT
A process for preparing glycol aldehyde by reacting formaldehyde, hydrogen and carbon monoxide at elevated tem-perature and superatmospheric pressure in the presence of rhodium catalyst and conversion thereof to ethylene glycol as the substantially exclusive polyol product.

Description

10~9877 This invention is concerned with processes for the preparation of glycol aldehyde, and conversion thereof to ethylene glycol, by reaction of formaldehyde, carbon monoxide and hydrogen in the presence of a rhodium catalyst.
Ethylene glycol is a very valuable commercial chemical with a wide variety of uses including use as a cool-ant and anti-freeze, monomer for polyester production, solvent and an intermediate for production of commercial chemicals.
Glycol aldehyde is a valuable intermediate in organic synthesis, including the preparation of serine, and is part-icularly usefl as an intermediate in the production of ethylene glycol by catalytic hydrogenation.
The reaction of formaldehyde with carbon monoxide and hydrogen is a known reaction and yields, inter alis, ethylene glycol, methanol, and higher polyhydroxy compounds.
For example, U.S. Patent 2,451,333 describes the reaction of formaldehyde, carbon monoxide and hydroven over a cobalt catalyst to produce mixtures of polyhydroxy compounds which include ethylene glycol, glycerol, and higher polyols..
Various metal catalysts are also disclosed including nickel, manganese, iron, chromium, copper, platinum, molybdenum, palladium, zinc, cadmium, ruthenium and compounds thereof.
U.S. Patent 3,920,753 describes the production of glycol aldehyde by reaction of formaldehyde with carbon mon-oxide and hydrogen in the presence of a cobalt catalyst f ~ ' .

..
.' ~

lU89~77 1 under controlled reaction conditions, but with low con-version yields.
Polyols are also produced by reaction of carbon monoxide and hydrogen over various metal catalysts. U.S.
Patent 3,833,634 describes this reaction catalyzed by rhodium to produce ethylene glycol, propylene glycol, glycerol, methanol, ethanol, methyl acetate, etc. Rhodium catalysts are also employed in the production of oxygenated derivatives of alkenes, alkadienes and alkenoic acid ester by reaction with carbon monoxide and hydrogen, as described, for example, in U.S. Patents 3,081,357; 3,527,809; 3,544,635;
3,577,219; and 3,917,661.
The prior art processes for production of ethylene glycol have characteristically provided mixtures of products, the principal co-products being propylene glycol and glycer-ine, along with the lower alcohols, methyl and ethyl alcohol.
Thus, these processes are encumbered by the need for expensive and time-consuming separation techniques where ethylene glycol is the desired product. In addition, the efficiency of the reaction in terms of yield of ethylene glycol is not high due to the concomitant formation of the co-products, which are usually present in significant amounts.
It has now been found that the reaction of for-maldehyde, carbon monoxide and hydrogen over rhodium cata- -lyst appears to involve a two-stage reaction, with the first stage yielding glycol aldehyde and methanol, and the second stage yielding ethylene glycol. Thus, this 1(~89877 1 reaction is analogous to that realized with cobalt catælysts as collectively disclosed in the aforementioned U.S. Paten~s

2,451,333 and 3,920,753, the surprising difference residing in the high selectivity of the present in~entive process which exclusively leads to ethylene glycol as the sole detectable polyol obtained in the second stage of the reaction. Further, the yield o~ gylcol aldehyde realized in the first stage of the present process is substantially greater than that obtained by the process described in tO U.S. Patent 3,920,953.
Thus, the present process provides glycol aldehyde in æubstantially higher yield than heretofore attainable from formaldehyde, carbon monoxide and hydrogen, and in the preferred f~rm of the 1nvention, provldes ethylene glycol as the exclusive, detectable polyol product, in improved yields when compared to slmilar processes.
The very desirable results obtained in accordance with the present process renders the process particularly amenable to commercial production of ethylene glycol, not only from ~he viewpoint of attainable high yields of ethylene gylcol, but also the ease of recovery o~ ethylene glycol from the co-produced methanol, e.g., by simple fractional distilla~ion. The ease of recovery is extreme~y important s~nce it permits separation o~ ~he ethylene glycol from the product mix~ure e~en in those process runs where methanol may be produced as the ma~or product, the glycol belng the minor product. Thus, even where the glycol 108987'7 1 is present in amounts corresponding to about lO mole-per-cent, and even less, of the reaction product mixture, tlle ease of separation will permit recovery of the glycol.
Glycol aldehyde is also produced in a high order o~ purity, in that the exclusive detectable co-pro-duct is methanol from which it can be readily separated.
Alternatively, where methanol does not present difflculty?
~he reaction product of the first stage of the present process can be used as such, without separating, as the source o~ glycol aldehyde in organic synthesis. Since the only detectable aldehyde, aside from ~ormaldehyde, in the first stage reaction mixture is glycol aldehyde, the ; reaction m~xture can be used as such in the second stage to produce ethylene glycol by reduction of glycol aldeh~de to obtain the glycol as the sole polyol product.
It is, o~ course, axiomatic that prior art pro-cedures are seriously encumbered by the ~act that the reaction product is a mixture of polyols (includ~ng ethylene glycol) which are extremely di~ficult to separate even when employing mult~ple ~ractional dist~llations.
The process o~ the present ~nvention iB accom~
plished by contact~ng formaldehyde, carbon m~noxlde and hydrogen, preferably in a suitable sol~ent, in the presence o~ a rhodi~m-containing catalyst at elevated temperature 2~ and superatmospheric pressure. The ma~or product o~ the two stage ~eaction is ethylene glycol, with the ma~or by-:~ product ~eing methanol. The manner o~ contact is not .
''' ~089877 1 critical s~nce any of the various procedures normally employed in this type of rea~tion can be used as long as efficient gas-liquid contact is provided. Thus, the process may be carried out by contacting a solution of formaldehyde together wlth the rhodium catalyst with a m~xture of carbon monoxide and hydrogen ~t the selected conditions. Alternatively, the solution of ~ormaldehyde may be passed over the catalyst in a trickle phase under a mixture of carbon monoxide and hydrogen at the selected conditions o~ temperature and pressure.
In ~i-ew o~ the two-sta~e- nature of the present process to produce ethylene glycol, the implementation can take several forms. The reaction can be accomplished by allowing both stages to proceed consecutively at suit-able temperature and pressure, or alternatively the reactioncan be stopped at the end of the flrst phase where the pro-duct ~s glycol aldehyde and the second phase can be caxried out under any applicable reduction process which t~ill result i~ convers~on of the aldehyde group o~ glycol aldehyde to a primary alcohol group resulting in ethylene glycol.
A wide ~ariety of reduction processes can ~e employed ~or the second phase reaction including the well-known chemical reduc~ng agents employed in reducing aldehydes to primary alcohols. For commercial processes, however, 2~ where possible, catalytic reductions employing hydrogen are usually pre~erred s~nce they are more practical and e~icient?
especially with ca~alysts which can be regenerated and thus 10~987 7 1 are re-usable. In the presen~ process, catalytic hydro-~enation is pre~erred ~or these same reasons, especially with catalysts which can be regenerated. Any hydrogenation catalyst can be employed.
Thus, typical hydrogenation catalysts include, ~or example, Raney Nickel, cobalt, copper chromite, rhodium, palladium, platinum, and similar such me~al catalysts, which can be used convenlently on supports such as charcoal silica, alumina, k~eselguhr and the li~e. The conditions of catalytic hydrogenation are ~Jell-~nown and, in general, the reaction can be conducted at temperatures ranging from about 30 to about 150C., usually at pressures of from about 100 to about 5000 psig. me use oi higher temperatures and pressures, though operable, prov~des no special advan-tage and usually requires special equipmenk which economicallyis disadvantageous and therefore not preferred.
Particularly pre~erred catalysts are those wkich characteristicallg require short react~on times, e.g.
palladlum and nickel, which is most desirable ~or commercial processes ~or economic reasons.
As mentioned hereinbefore, the main product of the ~irst stage react~on is glycol aldehyde, along with methanol. Glycol aldehyde tends to form acetals, a reaction typical o~ aldehydes, and ~n view of the primary alcohol group present in the molecule, th~s compound ~orms hem~-acetals and acetals with itsel~ ~n the ~orm o~, ~or example, l~near and cyclic acetals, represented by ~he formulas:

" 1089877 1 A.

C~ _ CHOH
O \O
\ CH - CH2 OH

B.

/ \ CHOH
CIH--O

C. OH ~0 HOCH2 ~OC~ C~
H
In addit~on, glycol aldehyde ~oxms acetals and hem~-acetals with methanol, and, i~ present, ethylene glycol. Acetals ~n particular are resistant to hydrogenation and should preferably be hydrolyzed to the free aldhyde so that ef~ic-ient reductlon to ethylene glycol can be effected.
~ he hydrolysis reaction can be accomplished me~ely by assuring the presence of water ~n the reaction mixture, pre~erablg ~n at least equ~valent molar ~uantit~es.
Thus, equ~molar amounts of water are required to assure com-2~ plete hydrolysis, with less than equimolar resulting ~n lessthan complete hydroly~is o~ the acetal present 1n the mlxtu~e wh~ch, ~n turn, results ~n lower y~eld o~ ethylene glycol. It is convenient to hydrolyze the acetal ~mmed~ately 1~8g877 1 prior ~o and/or concurrent ~rith the reduction stage.
Oftentimes, the a~ount o~ water required for substantial hydrolysis of the a~oremen~ioned acetals may already be present in the first stage reaction which ideally contains small amounts of water for best results, e.g. ~rom about 0.5 to about 10~ by volume. Alternatively, where insu~ficient water is present, the necessary water level can be achieved by mere addition of wa~er to the second stage reaction, either batchwise or by metering over the course of the reaction. In experience to the present tlme, optimum final levels of water are in the range o~ from about lB-30% by volume based on the hydro-gena~ion mix~ure.
To ~acilitate hydrolysis, the presence o~ an lS acid is p~rticularly desirable. Thus, s~rong mineral acids, such aæ hydrohalic acids, sulfuric, and phosphoric acids or, preferably, weak organic acids, especially lower aIkanolc acids such as acetic and propion~c ac~ds, can be employed for this purpose. Strong mineral acids should be avoided where the reaction solvent is reactive therewlth, e.g. am~de solvents which hydrolyze. As will be apparent from the following disclosure, amide solvents are usually pre~erred, particularly in the f~rst stage reaction and with these æolvents, it is prePerred to employ weak acids to catalyze the acetal hydrolysis. The amount o~ acid employed does not appear to be critical and even ~race amounts a~e e~fec~e, as shouId be obvious to those skilled in this art .
Thus, it is apparent tha~ a separate hydrolys~s ; 30 step is not always necessary since the normal water content 1(~8~87~

1 of the first stage reaction will hydrolyze at least part of the acetals produced and the hydrolyzed part will reduce to ethylene glycol. However, maximizing yield of ethylene glycol dictates the inclusion of a hydrolysis step to assure maximum hydrolysis and thus the highest realizable yield of ethylene glycol. Accordingly, the inclusion of the hydrolysis step, though not always essential, amounts to good technique, which, in view of the simplicity of adding water, with or without acid present, is readily practicable.
The combined hydrolysis - hydrogenation step can be carried out by art-recognized techniques as described, for example, in U.S. Patents 4,024,197; 2,721,223; 2,888,492 and 2,729,650.
The catalyst for the first stage reaction may be elemental rhodium, or a rhodium compound, complex or salt, or mixtures thereof, employed as such or deposited or affixed to a solid support such as molecular sieve zeolites, alumina, silica, anion exchange resin or a polymeric ligand.
In the active form, the catalyst comprises rhodium in complex combination with carbon monoxide, i.e., rhodium carbonyl, which may contain additional ligands as described, for example, in U.S. Patent 3,527,809 and the aforementioned U.S. Patent 3,833,634, each of which discloses rhodium complexes containing carbon monoxide and organic ligands as well as hydrogen as a ligand. As described in U.S. Patent 1~8g87 7 1 3,833,634, suitable organic ligands are compounds which contain at least one nitrogen and/or at least one oxygen atom, said atoms having a pair of electrons available for formation of coordinate bonds with rhodium. Illustrative organic ligands include various piperazines, dipyridyls, N-substituted acetic acids; tetrahydrofuran, dioxane, 1,2-dimethoxybenzene, alkyl ethers of alkylene glycols, alkano-lamines, iminodiacetic acid, nitrilotriacetic acid, ethylenediaminetetraacetic acid, and the like. In U.S.
Patent 3,527,809 are described phosphorus-containing ligands such as trialkyl, triaryl and tricycloalkyl phosphites and triarylphosphines, as well as the analogous antimony and arsenic compounds.
The catalyst can be employed in soluble form or in suspension in the reaction medium, or alternatively deposited on porous supports. The catalyst can be prepared by various techniques. For example, the complex with carbon monoxide can be preformed and then introduced into the reaction medium, or, alternatively, the catalyst can be formed in situ by reaction of rhodium, or rhodium compound, directly with carbon monoxide which may be effected in the presence of a selected organic ligand to form the organic ligand-carbon monoxide-rhodium complexes in the reaction medi~m.

.

-- 10~9~77 1 When glycol aldehyde is the desired product, of c~urse, only ~he ~irst stage reaction need be carried out. The product obtained is usually in ~he form of the aforementioned acetals and can be separated from the co-produced methanol and reaction solvent, if necessary, by fractional distillation. Gas chromatography and mass spectrophotometric analysis is used to identi~y the pro-duct as glycol aldehyde. ;In addition, the dimedone (5,5-dime~hylcyclohexane-1,3-dione) derivative o~ pure glycol aldehyde were prepared and compared with the dimedone derivative of the product obtained from a typical reaction according to the present process to show them to be ident-~Oal. NMR analysls of the der~vat~ve con~irmed glycol aldehyde as the product. No glyox~-l was detected by any of the a~orementioned analytical techniques. Further, the ; only aldehydes detected in the reaction product were for-maldehyde and glycol aldehyde.
The ~irst stage reaction which results in glycol aldehyde, and methanol, production is usually substantially complete in relati~ely short reaction times, usually less than about one hour, w~th sub~tantial yield o~ product realized ~n as little as 30 minutes, and e~en less time.
Usually, only small am~unts o~ ethylene glycol, 1~ any, can be detected.
As should be appare~t, the rhodlum catalyst employed in the ~irst stage reaction can also serve as the h~drogenation catalyst for the second stage reaction to 1 produce ethylene glycol. Thus, if the first phase reaction is allowed to con~inue, eventually the hydrogenation rezction will yield ethylene glycol. Particularly excellent yields are obtained by adding water, where necessary, to hydrolyze the glycol aldehyde acetals present from the first stage reaction thus realizin~ maximum yields of ethylene glycol.
In general, the rhodium catalyst of the first stage reaction is an effective catalyst ~ r the second stage hydrogen~tion, but does not provide as short reaction ~imes as can be realized with other hydrogenation catalysts, under the usual reaction conditions.
To shorten the second stage reaction time, it is possible to effect the reduction step over metal catalysts such as palladium and nickel, and it is usually preferred to ef~ect the second stage reaction in a separate reactor.
Thus, the first stage reaction can be conducted in a first reactor under selected conditions of temperature and pressure, and after completi~n the first stage product, with or with-out lsolation ~rom the reaction mixtuxe, can then be trans-ferred to a second reactor under selected cond~tions oftemperature and pressure to e~fect the hydrogenation reaction under hydrolysis conditions, l.e. in the pxesence of at least the stoichiometric amount of water to hydrolyze the glycol aldehyde acetals present.
Alternati~ely, the two stage reaction can be conducted in one reactor wlth provision ~or controll~ng the reaction parameters. At the time o* the hydrogenation ~o~9877 1 stage, the selected hydrogenation catalyst can be added, conveniently with the water required for hydrolysis, if any is needed, and the hydrogenation reaction can then pro-ceedO In this latter modification, the hydrogenation cata-lyst can be added to the ~irst phase reaction mixture withor without the ~irst phase rhodium catalyst being present.
Generally, it is preferred to remove the rhodium catalyst, particularly if th~s can be done con~en~ently so that competiti~e catalysis will not impede the hydrogenation reaction, and, more importantly, to permit more accur~te control over the reaction.
The present invention, therefore, provides a simplifled process for select~ve production of glycol aldehyde as the sole detectable aldehyde product. In 1~ addition, this invention affords a simplified process for obta~ning ethylene glycol by either allowing the initial process for glycol aldehyde to continue so that hydrogenation under hydrolytic conditions yields ethylene glycol or, alternati~ely, the glycol aldehyde product of the first stage reaction is reduced under hydrolytic conditions employing art-recognized reduction processes to ethylene glycol. In the latter process, the glycol aldehyde pro-duct of the first stage reaction can be used in the form of the reaction mixture, or the product can be isolated, 2~ as by ~ractionat~on, and used ~n puri~ied ~orm.

-- . .. _ The amount o~ catalyst emplbyed in the first stage ; reaction does not seem to be critical and can ~ary consider-ably. At least a catalytically e~fecti~e amount of catalyst ; 30 should be used, of course. In general, an amount o~ catalyst 10~987~

1 which is effective to provide a reasonable reaction rate is sufficient. As little as 0.001 gram atoms of rhodium per liter of reaction medium can suffice while amounts in excess of 0.1 gram atoms do not appear to materially affect the rate of reaction. For most purposes, the effective preferred amount of catalyst is in the range of from about 0.005 to about 0.025 gram atoms per liter.
The reaction conditions are not overly critical in that wide ranges of elevated temperature and superatmospheric pressures are operable. The practical limitations of pro-duction equipment will dictate to a great extent the selection of temperatures and pressures at which the reaction is to be effected. Thus, using available production systems, the selected elevated temperature should be at least about O O
75 C. and can range up to about 250 C. and even higher. For most purposes, the preferred operating temperature ranges from about 100 to about 175 C. The superatmospheric pressure should be at least about 10 atmospheres and can range up to almost any pressure attainable with production apparatus.
Since extremely high pressure apparatus is quite expensive, pressures to about 700 atmospheres are suggested. Most desirably, the pressure should be in the range of from about 250 to about 400 atmospheres, particularly when employing the aforesaid preferred temperature range.
The reaction is preferably carried out in a solvent which will dissolve polar materials and which preferably is aprotic in order to maximize selectivity to ethylene glycol.
Suitable solvents include a wide variety and are exemplified by N-substituted amides in which each hydrogen of the amido nitrogen is substituted by a hydrocarbon group, e.g., l-methyl-10~9~77 1 pyrrolidin-2-one, N,N-dimethylacetamide, N,N-diethylacetamide, N-methylpiperidone, 1,5-dimethylpyrrolidin-2-one, l-benzyl-pyrrolidin-2-one, N,N-dimethylpropionamide, hexamethylphos-phoric triamide and similar such liquid amides; nitriles, such as acetonitrile, benzonitrile, propionitrile and the like; cyclic ethers such as tetrahydrofuran, dioxane and tetra-hydropyran; ethers such as diethyl ether, 1,2-dimethoxy-benzene alkyl ethers of alkylene glycols and polyalkylene glycol and di-, tri- and tetraethylene glycols; ketones such as acetone, methyl isobutyl ketone, and cyclohexanone; esters, such as ethyl acetate, ethyl propionate and methyl laurate;
organic acids such as acetic acid, propionic acid and caproic acid; and alkanols, such as methanol, ethanol, propanol, 2-ethylhexanol and the like; and mixtures thereof. Many of the solvents are non-reactive in the medium whereas others are capable of functioning as ligands. The selected solvent should preferably be liquid under the reaction conditions.
When employed, solvents appear to exert varying influ-ences on the yield of ~roduct formed and the selectivity to ethylene glycol, depending on the nature of the solvent. For example, when lower alkanoic acids, e.g., acetic acid, are present for example as a co-solvent in the first state reaction the reactlon appears to proceed more rapidly but the yield of glycol decreases somewhat while that of methanol in-creases. When acetic acid was employed at a level of from about 10 to about 20 volume percent of the reaction mixture, the reaction proceeded in about one-half the time required for the same solvent containing no acetic acid but with increased methanol production (55% vs. 40%) and decreased glycol -1(~89877 l production (30~ vs. 48%). Further, basic amines such as pyridine, triethylamine and amines of comparable basicity appear to exert a negative influence on the yield of glycol aldehyde obtained and this influence becomes more pronounced as the molar ratio of amine to rhodium increases. Thus, even when the amine is present as a co-solvent, the tendency is towards reduced yield of glycol aldehyde when compared to solvent systems from which amines are excluded. Protic solvents such as water, phenols and carboxylic acids, e.g., acetic acid, in large quantities, e.g., greater than about 30-40~ by volume, exert a similar negative influence on the yield of glycol aldehyde. In most cases, the decrease in yield of glycol aldehydr is accompanied by an increase in methanol yield, while in some cases the conversion of formaldehyde is reduced so that the yield of both products is reduced. Thus, where optimum yields of glycol aldehyde and ethylene or protic sol-vents in significant amounts are usually avoided, particularly in the first stage reaction.
On the other hand, certain solvent systems favor high selectivity for glycol aldehyde and ethylene glycol production, and in many cases substantially lower yields of methanol are obtained. Solvents such as organic amides, in particular, ; favor high selectivlty for glycol aldehyde and ethylene glycol production, and in many cases substantially lower yields of - methanol are obtained. Solvents such as organic amides, in particular, favor high selectivity for glycol aldehyde and ethylene glycol production, and in many cases substantially lower yields of methanol are obtained, for which reason these solvents are preferred. Hydrocarbon solvents can be employed but apparently result in lower yiedls of glycol aldehyde and glycol than obtained with the preferred solvents.

1 The preferred solvents are aprotic organic amides. The contemplated amides include cyclic amides, i.e., in which the amido group is part of a ring structure such as in pyrrolidinones and piperidones; acylated cyclic amines, such as N-acyl piperidines, pyrroles, pyrrolidines, piperazines, morpholines, and the like, preferably in which the acyl group is derived from a lower alkanoic acid, e.g., acetic acid; as well as acyclic amides, i.e., wherein the amido group is not part of a ring structure as in acetamides, formamides, propionamides, caproamides and the like. The most preferred of the amides are those in which the amido hydrogen atoms are fully replaced by hydro-carbon groups preferably containing not more than 8 carbon atoms. Exemplary hydrocarbon groups are alkyl, preferably lower alkyl such as methyl, ethyl and butyl; aralkyl, such as benzul and phenethyl; cycloalkyl, such as cyclopentyl and cyclohexyl; and alkenyl, such as allyl and pentenyl. The preferred amido nitrogen substituents are lower alkyl, especially methyl, ethyl and propyl groups and aralkyl groups, especially benzyl. The most preferred amide solvents include l-methylpyrrolidin-2-one, 1-ethylpyrrolidin-2-one, l-benzylpyrrolidin-2~one, N,N-diethylacetamide, and N,N-diethylpropionamide.
The nitrile solvents include any organic nitrile solvent preferably containing up to about 3 carbon atoms, such as acetonitrile, benzonitrile, phenylacetonitrile, capronitrile and the like. Mixtures of solvents can be employed.

_18-~0~9877 1 As is the case when basic-nitrogen containing solvents are employed, the use of basic nitrogen-containing ligands with rhodium carbonyl complexes is generally attended by a lower selectivity for ethylene glycol production, usually accompanied by an increase in methanol production. For this reason, catalysts including basic nitrogen-containing ligands are usually avoided since the yield of the desired product, ethylene glycol, is diminished. Oxygen-containing ligands seem to provide best results in terms of the yield of ethylene glycol, for which reason they are preferred over the aforesaid basic nitrogen-containing ligands. The most pre-ferred of the rhodium carbonyl complexes are the complex formed with the acetylacetonate anion, Rh(CO)2(C5H702) and hexarhodium hexadecacarbonyl, Rh6 (CO)16, since up to the present, best results are obtained therewith and they are readily obtainable or preparable.
The reaction pressures represent the total pressure of the gases contained in the reactor, i.e., carbon mon~xide and H2, and, if present, any inert diluent gas such as nitrogen. As in any gaseous system, the total pressure is the sum of partial pressures of component gases. In the present reaction, the molar ratio of hydrogen to carbon monoxide can range from about 1/10 to about 10~1, with the preferred ratio, from about 1/5 to about 5/1, and the reaction pressure can be achieved by adjusting the pressure of these gases ir. the react~r.

~ .

., .

~089877 1 For best results, the molar ratio of carbon monoxide to hydrogen is maintained at high values in the first stage reaction where high partial pressures of carbon monoxide favor production of glycol aldehyde. In the second stage reaction, high partial pressure of hydrogen is desirable for reduction reaction. Thus, in the first stage reaction to produce glycol aldehyde, the partial pressure of carbon monoxide is usually adjusted to be about 3 to about 10 times that of hydrogen. In the second stage reaction, i.e. the hydrogenation, the partial pressure of hydrogen is adjusted to a high value to facilitate the reaction. Such adjustments of the gas feed can be readily accomplished. For example, after the first phase reaction is substantially complete, the reactor need only be bled to lower the pressure and then pressurized with hydrogen gas to achieve the desired high partial pressure of hydrogen. Carbon monoxide present in the gaseous system of the first phase reaction need -not be completely purged from the reactor prior to repressuriz-ing with hydrogen gas. Of course, carbon monoxide can reduce the efficiency of certain catalyst systems, possibly through 2~ poisoning as is known, and preferably is excluded when such systems are employed.
Where the second phase reaction is carried out in a separate reactor whether over the originally present rhodium catalyst or a different metal hydrogenation catalyst, the reaction is normally conducted under hydrogen gas without diluent gas, as is usual in catalyzed hydrogenation reactions.

, 1~89877 1 The source of formaldehyde for the present process can be any of those commonly used in this technology including paraformaldehyde, methylal, formalin solutions, and polyoxy-methylenes. Of these, paraformaldehyde is preferred since best yields are attained therewith. Solutions of formaldehyde in solvents, advantageously the reaction solvent, can be used, e.g. solutions of formaldehyde in aqueous reaction solvent, such as N-methyl pyrrolidin-2-one. The use of methylal may be attended by a reduction in yield of ethylene glycol. If trioxane is employed, because of its stability, a hydrolyzing agent should be employed to release formaldehyde.
As with any process of this kind, the present pro-cess can be conducted in batch, semi-continuous, and continuous operation. The reactor should be constructed of materials which will withstand the temperatures and pressures required, and the internal surfaces of the reactor are substantially inert.
The usual controls can be provided to permit control of the reaction such as heat-exchangers and the like. The reactor should be provided with adequate means for agitating the react-ion mixture; mixing can be induced by vibration, shaking,stirring, oscillation and like methods. Catalyst as well as reactants may be introduced into the first stage or the second stage reactor at any time during the process for replenishment.
Recovered catalyst, solvent and unreacted starting materials may be recycled.
The relative yields of ethylene glycol and methanol are not overly critical since the product mixture can be readily separated into the components by known techniques, especially by fractional distillation, regardless of the .~

1 proportions contained in the mixture. Therefore, even where ethylene glycol is 10-20% of the reaction mixture, it can be readily separated from the mixture, especially in continuous process production of ethylene glycol, with the methanol recycled as formaldehyde. Of course, the preferred processes yield mixtures in which ethylene glycol predominates as the reaction product.
In addition to the aforementioned solvent effects, other factors also affect the yields of ethylene glycol and methanol and the conversion of formaldehyde in the process.
For example, in the combined two-stage reaction, the use of low partial pressures of carbon monoxide appears to favor greater methanol production, whereas the use of high partial pressure of Co, particularly during the first stage, results in lower methanol yields without significant change in glycol yield. Thus, at a partial pressure of carbon monoxide at 1900 psig., the conversion of formaldehyde amounted to 57% with a 76%
molar selectivity for ethylene glycol whereas at 1055 psig., the conversion was 72% and molar selectivity was 56% under otherwise identical conditions. Increased partial pressure of hydrogen particularly in the combined reaction resulted in increased glycol selectivity and increased convertion of formaldehyde with little, if any, change in methanol yield.
The effect of temperature variation in the pre-ferred temperature range is not as pronounced, with higher formaldehyde conversion and ethylene glycol selectivity being .

1085~8'77 1 obtained in the 100 - 175 C. range, particularly during the first stage reaction.
The process conditions for the separate first stage reaction are essentially the same as employed in the first stage of the combined two-stage reaction. Thus, the reaction is carried out at a temperature of at least about 100 C. to obtain a reasonable reaction rate although somewhat lower temperatures can be employed with slower reaction rates being realized. For reaction times of about one hour, and even less, the temperature should be in the range of from about 100 C.
O O O
to about 175 C., preferably from about 120 C. to about 140 C.
As in the combined two stage reaction, the partial pressure of carbon monoxide is preferably high, in comparison to that of hydrogen, with the preferred ratios being from about 2:1 to about 10:1, the more preferred being from about 3:1 to about 8:1. The total pressure of gas used is generally main-tained at from about 1000 psi up to about 9000 psi, with from about 3000 to about 7000 psi being preferred. Of course, higher pressures and higher temperatures can be used but with no appreciable advantage and, since they require the use of spec-ial high pressure equipment, they are usually avoided.
The reaction conditions employed in the second stage reaction, i.e. the hydrogenation, can be any of the standard reaction temperatures and pressures employed for such reac-tions since neither temperature nor pressure are critical for this reaction. Preferably, the hydrogenation is conducted at .

10~877 1 a temperature of at least about 100 C. in order to effect a reasonable reaction rate. Of course, lower temperatures can be used if longer reaction times can be tolerated. The pressure of hydrogen gas is not excessively critical as long as suffic-ient gas is available for the hydrogenation. For convenience, the pressure will range from about 500 psi to as much as 5000 psi, although even higher pressures can be employed.
When the catalyst selected for the hydrogenation step is other than rhodium, it is preferred to remove the rhodium catalyst from the first stage reaction mixture. This prefer-ence is primarily predicated on the desirability of avoiding concomitant catalytic effects which may tend to reduce the yield of ethylene glycol, the desired product. It has been determined, for example, that the yield of ethylene glycol was considerably lessened when the hydrogenation was effected over supported nickel or palladium catalyst using the first stage reaction mixture without removing the rhodium catalyst present therein. When these hydrogenations were repeated with the addition of water to the reaction mixture, the water preferably containing at least catalytic amounts of acid, usually acetic acid, almost quantitive conversion to ethylene glycol occured, particularly when Palladium catalyst, e.g. Pd/C, is used. How-ever, after the glycol aldehyde is separated from rhodium cata-lyst, e.g. by distillation, the glycol aldehyde is reduced al-most quantitatively with catalysts such as palladium on carbon in the absence or presence of rhodium. The aforesaid reduced yields of ethylene glycol are explainable by the production of unidentified high boiling liquid product which remains after distillation of ethylene glycol from the reaction mixture.
Apparently, secondary competitive reactions proceed where 108987~

1 both the rhodium catalyst and the hydroge~ation metal catalyst are simultaneously present in the hydrogenation reaction mixture, the nature of which reactions is not understood up to the present. Surprisingly, no significant amounts of the high boiling residue were discovered in the reactions mixtures obtained with either rhodium or other metal as the sole hydrogenation catalyst. With Pd/C, glycol aldehyde is almost quantitatively reduced to ethylene glycol.
The results obtained with the present new process are surprisin~ and totally unexpected. As hereinbefore described, the prior art processes of reacting formaldehyde, carbon monoxide and hydrogen have let to mixtures of polyol products principally ethylene glycol, glycerol and higher diols from which it is extremely difficult to separate the individual components. The present process on the other hand, selectively yields ethylene glycol as the polyol product.
Analysis of the product produced by means of gas-liquid chromatography has failed to reveal any polyol other than ethylene glycol, which is readily separated from methanol, the monohydric alcohol product, as hereinbefore mentioned.
The following examples further illustrate the in-vention.

A 71 ml. stainless steel reactor fitted with a glass liner is charges with 0.5 g of commercial paraformaldehyde 0.019 g ~h(CO)2(C5H702) and 5 ml. N-methylpyrrolidinone. The reactor is pressured to 43~0 psig with H2 and CO at a ratio of 2.2/1 and then shaken by a wrist action shaker in a hot air oven at 150 C. for five hours. After cooling and venting the 1(~89877 1 gases, the reaction mixture is analyzed via gas-liquid chromatography and is found to contain 0.07 g. of methanol and 0.43 g. of ethylene glycol. No higher polyols are observed.

The reaction is carried out as in Example 1 except the reactor is pressured to 3350 psig with H2 and CO
at a ratio of 1.5/1. The reaction solution is analyzed and found to contain 0.08 g. of methanol and 0.34 g. of ethylene glycol. Identification of ethylene glycol is confirmed by mass spectrometry.
EXA~PLE 3 The reaction is carried out as in Example 1 except the reactor is pressured to 2350 psig with H2 and CO
at a ratio of 1.7/1. Analysis after the reaction shows the presence of 0.07 g. of methanol and 0.25 g. of ethylene glycol.

-The reaction is carried out as in Example 1 except that 2.5 g. of methylal is charged in place of paraformal-dehyde and the reactor is pressured to 3330 psig. with H2 andCO at a ratio of 1.5/1. Anal~sis of the solution after reaction shows the presence of 0.26 g. of methanol and 0.06 g.
of ethylene glycol.

The reaction is carried out as in Example 1 except the reactor is pressured to 3750 psig. with H2 and CO at a ratio of 4/1. Analysis of the reaction mixture shows the presence of 0.16 g. of methanol and 0.40 g. of ethylene glycol.

-2~-1 EXA2~LE 6 Th~ reaction is carried out as in Example 2 except the formaldehyde is charged as 1.28 g. of 37%
aqueous solution stabilized with methanol. Analysis 5 of the reaction solution shows the presence of 0.29 g.
of methanol (after correcting for the initial methanol) and 0.25 g. o~ evhylene glycol.
EXA~IE 7 The reaction is carried out as in Example 1 10 except the formaldehyde is charged as 0.5 g. of alkal~
precipitated ~-polyoxymethylene and the reactor is pressured to 3500 psig. with H2 and C0 at a ratio of 2.3/1. Analysis of.the reaction solution sAows the presence of 0.17 g. of methanol and 0.30 g. of ethylene 15 glycol-The reaction is carried out as in Exampie 2 except the reaction temperature is 175C. Analys~s of the reaction solution shows the presence of 0.05 g. of methanol and 0.25 g. of ethylene glycol.
EXAMPLE 9, The reaction is carried out as in Example 2 except the reaction temperature is 125C. Analysis of the reaction solution shows the presence of o.o8 g. of 2~methanol and 0.33 g. o~ ethylene glycol.

A 71 ml. stainless steel reactor equippPd w~th a glass 7iner is charged with 0.037 g Rh(C0)2(C~H702), 1.0 g, paraformaldehyde ænd 5 ml. N-methylpyrrolidinone, pressured 9~7'7 1 to 3000 psig. with H2 and CO in a ratio of 1.5/1, and shaken ten hours at 20Q C. After cooling and venting the gases analysis of the reaction solution shows the presence of 0.51 g. of methanol and 0.16 g. of ethylene glycol.

The reaction is carried out as in Example 10 except the charge is 0.037 g. Rh(CO)2(C5H7O2), 1.0 g.
paraformaldehyde and 5 ml. of hexamethylphosphoric triamide, 10 and the pressure is 3330 psig with H2 and CO in a ratio of 1.5/1. The reaction is carried out for five hours at 150 C.
Analysis of the reaction solution shows the presence of 0.58 g. of methanol and 0.20 g. of ethylene glycol.

The reaction is carried out as in Example 11 except the solvent is N,N-dimethylacetamide. Analysis of the reaction product shows the presence of 0.64 g. of methanol and 0.32 g. of ethylene glycol.

The reaction is carried out as in Example 2 except the solvent is acetonitrile. Analysis of the reaction product shows the presence of 0.10 g. of methanol and 0.14 g. of ethylene glycol.

The reaction is carried out as in Example 2 except the solvent is N-methylpiperidone. Analysis of I the reaction product shows the presence of 0.32 g. of s, methanol and 0.16 g. of ethylene glycol.

,,-.

The reaction is carried out as in Example 2 except the solvent is N-benzylpyrrolidinone. Analysis of the reaction product shows the presence of 0.28 g. of methanol and 0.09 g. of ethylene glycol.

The reaction is carried out as in Example 7 except the solvent is N,N-diethylacetamide and the formaldehyde is charged as 0.5 g. of paraformaldehyde.
Analysis of the reaction product shows the presence of 0.05 g. of methanol and 0.29 of ethylene glycol.

The reaction is carried out as in Example 16 except the solvent is 1,5-dimethyl-2pyrrolidinone.
Analysis of the reaction product shows the presence of 0.35 g. of methanol and 0.17 g. of ethylene glycol.

The reaction is carried out as in Example 7 except that the formaldehyde is charged as paraformaldehyde, the solvent is 1,4-dioxane and the H2/CO ratio is 2.0/1.
Analysis of the reaction product shows the presence of 0.07 g. of methanol and 0.16 g. of ethylene glycol.

The reaction is carried out as in Example 18 except the solvent is benzonitrile. Analysis of the reaction product shows the presence of 0.13 g. of methanol and 0.1 g. of ethylene glycol.

98'77 1 EXAMPLE 2~
A 300 ml. Magne-Stir autoclave equipped with a Disperso-Max stirrer which was operated at 1500 rpm was charged with 0.285 g. of Rh(CO)2(C5H7O2), 7.5 g. of 95~
paraformaldehyde and 75 ml of N-methylpyrrolidinone. The reactor is closed and, while the solution is stirred, pressured to 3500 psig with H2 and CO at a 3/1 ratio. The reactor is heated to 150 C. Maximum pressure of 4650 psig is reached at 138 C. When the pressure drops to 4100 psig, the reactor is repressured to5000 psig with H2 and CO at a 2/1 ratio. Total reaction time at 150 C. is 3 hours. After cooling and venting the reactor is opened and the product solution recovered. Analysis of the product ~hows the presence of 3.0 g. of methanol and 6.0 g. of ethylene glycol.

The procedure of Example 20 is repeated except that the autoclave is pressured to give 5000 psig at 125 C
with H2 and CO at a ratio of 1.86/1. The reactor is heated, with stirring at 1500 rpm, to 125 C. When the pressure drops to 4200 psig, the autoclave is repressured to 5000 psig with the same gas mixture. The total time at 125 C. is five hours. Analysis of the product solution shows the presence of 1.3 g. of methanol and 6.6g. of ethylene glycol.

The reaction is carried out as in Example 21 except the H2 and Co were at a ratio of 1/1 and at an initial pressure of 5200 psig at 125 C. Analysis of the product solution shows the presence of 4.2 g of methanol and 2.8 g.
of ethylene glycol.

lQ~9~7'7 The reaction is carried out as in Example 21 except that 0.143 g. of Rh(CO)2(C5H7O2) is charged-Analysis of the product solution shows the presence of 1.3 g. of methanol and 1.8 g. of ethylene glycol.
EXAMPLE 2~
The reaction is carried out as in Example 21 except that 0.428 g. of Rh(CO)2(C5H7O2) is charged and the reaction time is three hours. Analysis of the product solution shows the presence of 1.0 g. of methanol and 6.3 g. of ethylene glycol.

The reaction is carried out as in Example 1 except the catalyst is charged as 0.008 g. powdered elemental rhodium and the temperature is 250 C. Analysis of the product solution shows the presence of 0.32 g. of methanol and 0.17 g. of ethylene glycol.

The reaction is carried out as in Example 2 except the catalyst is charged as 0.067 g. RhH(CO)[P(C6H5)3]3.
Analysis of the product solution shows the presence of 0.16 g. of methanol and 0.11 g. of ethylene glycol.

The procedure of Example 25 is repeated using rhodium oxide and rhodium chloride, respectively, in lieu of elemental rhodium with comparable results.

The procedure of Example 1 is repeated except the solvent is N,N-diethylpropionamide.
Analysis of the product shows 0.02 g. methanol and 0.18 g. ethylene glycol.
-EXAMæLE ~9 The procedure of Example 1 is repeated except the solvent is N-ethylpyrrolidin-2-one.
Analysis of the product shows 0.17 g. methanol and 0.38 g. ethylene glycol.

The procedure of Example 1 is repeated except the solvent is N,N-diethyl-m-toluamide.
Analysis of the product shows 0.07 g. methanol and 0.17 g. ethylene glycol.
In each of the foregoing examples, the analyses were carried out using gas-liquid chromatography and in no instance was there any polyol, excepting ethylene glycol, detected.
The amide solvents used in the oregoing examples were freed of amine contaminants by distillation.
As mentioned hereinbefore, amines appear to show a negative influence on the yield of glycol and thus are preferably avoided. Accordingly, amine solven~s and amine ligands for the rhodium carbonyl complex are not employed because of the said negative influence which can result in little, if any, yield of the dèsired ethylene 1 glycol depending on the amount of amine present. Thus, when pyridine is present in the reaction mixture, the yield of ethylene glycol is appreciably diminished, the extent of diminution of yield being proportional to thç
molar ratio of pyridine to rhodium; when hydroxypyridine is used as ligand for the rhodium carbonyl complex a similar diminution of yield of ethylene glycol is observed. The following examples illustrates the neqative influence of amines on the present process.

:
The procedure of Example 1 is repeated using 0.073 millimole Rh(CO12(C5H702) and 15.8 millimoles paraformaldehyde in 5 ml. N-methylpyrrolidin-2-one which is stirred for 5 hours at 150 C. The initial pressure is 5000 psig (H2/CO=2.2).
Using this procedure, the effect of addition of various levels of pyridine is determined and the results given in Table 1.

% Yield Productivity Expt. Pyridine Added (Glycol and moles/mole Rh No. moles/mole Rh Methanol) Glycol Methanol 2 0.16 67 45 114

3 0.3 80 26 163

4 1.0 75 11 167 lQ~9877 1 From these data, it is apparent that pyridine exerts a neqative influence on the ethylene qlycol yield. Similar results are obtained with other amines such as methylamine, triethylamine and 2-hydroxvpyridine.
The effect of hydroqen and carbon monoxide partial pressures, previously discussed herein, is demonstrated by the data of Table II which is determined by repeating the procedure of Example 31 without amine present but varying the gaseous components.
TA~LE II

Initial Partial~ Yield Productivity Expt Pressures, psi(Glycol + Moles/mole Rh No. H2 COMethanol) Glycol Methanol -.
1 1410 l90Q 38 51 30 2 2850 lgO0 45 74 23 Variation in rea~tion temperature provides a some-what lesser effect, although fairly pronounced, the lower temperatures providing higher conversions of formaldehyde and higher yields of ethylene glycol as illustrated in Table III.

-3~-~089877 1 TA~I.EIII

% Yield Productivity

5 Expt. Temp. (Glycol and (moles/mole ~h) No. C Methanol) Glycol Methan~l 1 (1) 175 . 37 56 - 25 2 (1) 150 45 74 23 3 (1) 125 51 74 36 4 (2) 175 47 68 34 5 (2) 150 57 77 46

6 (3) 125 62 98 37 (1) 5 hrs in 71 ml shaken reactor -Rh(CO) (C H702), 0.073 millimole; paraformaldehyde, 15.8 m~ moles; N-methylpyrrolidinone, 5 ml;
H~:C0, 3:2 at 4760 psi (initial).
(2) 3 hrs in 300 ml stirred reactor -Rh(C0) (C H 2)~ 1.095 millimole; paraformaldehyde, 237 mi~ oIes, N-methylpyrrolidinone, 75 ml; H , 3250 psi (initial); C0, 1750 psi (initial). 2 (3) Identical to (2) except 5 hrs.

The Psllowing example illustrates the ine~iect~veness 20 of rhodi~m carbonyl catalyst in the reaction of carbon monoxide and hydrogen to produce ethylene glycol and methanol, under the same conditions of temperature and pressure as employed in the preceding examples.

2~ .

-` 1089877 1 ~XAMPLE 32 Usin~ the procedure of the foregoing examples, except that formaldehyde is omitted, a mixture of Rh(CO)2(C5H70~) (0.145 millimoles); ligand (when present~
(0.57 millimole) and solvent (5 ml. ? is heated at an initial pressure o~ 5000 psig (H2/C0=1.5) and 200C with stirring for 10 hours. A series of runs, ~Jith and without ligand (2-hydroxypyridine and pyrocatechol) using various solvents including N-methyl pyrrolidin-2-one, tetrahydro~uran, tetraglyme and mixtures thereof ~th methanol and methyl formate~ resulted in no detectable amounts of ethylene glycol and from 0 to 5 millimoles o~
methanol.
The use o~ lon~er or shorter reaction time shows 15 no appreciable change as is also the case when the catalyst is increased to flve times the a~orestated amount.
The results are summarized in Table IV

3~
-~6-108987'7 o~
.,, aJ ~ _ ~ ~ ~ ~ ~ ,, O _ ~ -- O
tJ h , .- 3 -S
.~ a O ~ Z
.k ~ 1l ~ .,~
_I k ~ Z
~ ~ ~ ~) `D N O ~ tn O 1-- 1~
.C ~1 I ,1 0 ~ ~ ,1 0 ~ U~ O O ~ ~ O
5_ O

t~` ~ ....
'I O ~ ,.
~ _~
_I /D .~ ~ .
O ~ O OO O ~ V

H ~ al h ~ O ~J 3 5~ ~I ~ ~q ~ ' '.
I U~
C~ ~: oN ~
N O -- I ~ O al ~ ..
~ 8 ~c ~N O ~ 8 .~
~ _ _ ~0 ~
~Q ~ Z O a~ -- _ _ _ t~ ~
~ ~ .
Q~ aJ ~q`
h --I

ra ~5 ~ r-l ,1 -1 U
S ~ h rl x a3 X --O~ ~ ~ I k ~
U
O
~ I
~ O
O
O _I ~ k ~1 O
. ~~ ~ ~ ~
O ~ _l ~ O ~ S 3;1 ~I Z ,-1 ~ ~ d' 117 ~9 1~ ) ~ O ~
S~ ~ U " ,1 - .
.C ~ ~ 11 O O ~ rl ~ O
,1 u~ ~ E~ E~ Z

~ _ _ _ _ ~08987 7 EX~MæLE`33 1 Using the procedure of Example 7 but using paraformaldehyde as formaldehyde source and H2/CO=2.1, a variety of N,N- disukstituted amides are evaluated as solvents with the results summarized in Table V.
TABLE V

N,N-dipropylacetamide 0.11 0.01 N,N-diburylacetamide 0.09 0.03 N-acetyl piperidine 0.19 0.31 N-propyl pyrrolidin-2-one 0.11 0.36 N-butyl " 0.~3 0.32 N-isopropyl " 0.17 0.31 N-3 butyl " 0.19 0.21 1(~89877 1 The foregoing examples are illustrative of the com-bined two stage reaction. Example 34 illustrates the pro-duction of glycol aldehyde by the first stage reaction.

EXP~PL~ 34 The following reaction mixture is charged to a pres-sure vessel as employed in the preceding examples:
2.5 mmole Rh (CO)2 (C7H502) 237 mmole paraformaldehyde (95%) 5 ml. H2Q --114 ml. N-me~hylpyrrolidinone The ~essel is pressured to 2500 psi (PCo=2000 psi and PH = 500 psi) and then heated to 130C and stirred at 1750 rpm.
Samples are remoYed at lS minute intervals and analyzed with the following results (concentrations are in mmoles): ~ ~
CONCENTRATION

Reaction Glycol Ethylene Time (min.) CH20 Aldehyde MeOH Glycol 45 89 81 2~ 0 The aldehydes present in the final reaction solution are identified as formaldehyde and glycol aldehyde with no obher aldehyde or carboxyl compound being detected. Glycol aldehyde can be separated from the reaction mixture, e.g. by distillation, or the reaction mixture can be used in the second stage reaction as in Example 35. When this procedure - ; .. ; . - . ... ~ . .. .. . .. . . .. .

10~9877 1 i5 repeated at 160C., the yield of glycol aldehyde decreased substantially after the first 30 minutes. On repeating this procedure at lower total pressure ~Pco= 2000 psi and PH =
500 psi? slightly lower yield of glycol aldehyde is obtained.
Example 35 illustrates the two stages reaction using the rhodium catalyst as the hydrogenation catalyst.

The procedure of Example 34 is repeated except that the first stage is terminated at the end of one hour and the vessel depressurized by bleeding and then repressurized with hydrogen to 5000 psi so that the hydrogen is 80 mole % of the total gas. The second stage reaction then proceeds and is sampled at 15 minute intervals for analysis with the following results (concentrations are in mmoles).

_~Q_ 10~9877 1 CON~N~RATI~NS
Experiment Ti~e H ~ A~ ~e Me~H _thy-lene Glycol Ç~ 126 26 0 3~ ~1 43 78 83 45 <1 14 73 112 - -60 Cl 6 74 117 .
2 3~133 67 1~ ~

lS 2 78 69 47 45 <2 6 65 117 60 <2 6 62 120 . _ .. ... . . . . .

~5 C2 6 82 127 60 <2 6 79 127 ~ -In ~xperimen~ 2, 10 ~1 of glacial acetic acid was added to the reaction mixture before the start of the hydro-genation. In Experiment 3,20 ml. of H20 was added at the beginning of the hydrogenation.

A pressure vessel is charged with the following:
4 ml. N-methyl pyrrolidinone

7.5~ mmole formaldehyde (as paraformaldehyde-95~) 0.7 mmole Rh ~CO)2 (C7H502) and the vessel is pressured to 4aoo psi with Co (80 mole %) and H2 ~2Q mole %) and heated at 130C for 90 minutes.

1 The product on analysis showed:
1.2 mmole MeOH
. mmole H2CO
4.7 mmole glycol aldehyde The reaction mixture is then pressurized with hydrogen to 75 mole % H2 and 25 mole % CO and heated at 150C. for five hours to obtain a product of the following composition:
2.2 mmole MeOH
3.6 mmole ethylene glycol The reduction step is repeated but with 0.5 g.
Ni on kieselguhr and 0.5 ml. each of water and acetic acid added to the reaction mixture. The product obtained has the following composition:
1.4 mmole ethylene glycol 1.0 mmole MeOH
and high boiling residue When this procedure is repeated using Pd/C (5%) in lieu of the Nickel catalyst, ~he product has the following composition: 1.8 mmole eth~lene glycol -1.7 mmole MeOH
and high boiling residue When the reduction procedure is repeated but with pure glycol aldehyde using Pd/C with N-methylpyrrolidinone as solvent at hydrogen pressure of 3000 psi for 5 hrs. at 150C. an almost quantitative yield of ethylene glycol is obtained. Using Nickel on kieselguhr in lieu of Pd/C
resulted in lower conversion to ethylene glycol.

3~

_ 1089~77 1 The following e~ample shows the results obtained ~ith pure glycol aldehyde in the hydrogenation, as contrasted with Example 36.
EXAMPL~ 37 A pressure vessel is charged with the following:

8.3 mmol. glycol aldehyde 4 ml. N-methyl pyrrolidinone and the selected catalyst system is added. The ve~sel is then pressurized to 3000 psi H2 and heated at 150C for five 10 hours to obtain the hydrogenation product with the following results:

Residual Ethylene aldehyde glycol -Run Cata~yst Additives (mmol.) (mmol.

1 Rh(Co)2 (C7H502) --- 0.9 2.4 Pd/C (5~) -2 Rh(C)2 (C7H52)50~ aq. HOAC 0.1 7.4 Pd/C (5%) 3 Pd/C (5%) ~~~ 9'9 The aqueous acetic acid (50%) is present at a 20 level of 20% by volume of the reaction mixture.
The results indicate quantitative conversion to ethylene glycol using Pd/C as the sole hydrogenation catalyst. Further, the results with the combined rhodium `
and palladium catalysts show that the hydrogenation proceeds substantially better under hydrolytic conditions.

-~ -43-987~

1 From the experimental data, it is apparent that in the present process a high catalyst efficiency is attained. In general, the optimum average yield of glycol aldehyde based on catalyst employed is about 100 moles/mole of rhod;um catalyst. In contrast, in U.S. Patent 2,451,333 the reported experimental results show that an optimum of about 14 moles of glycol aldehyde are produced per mole of cobalt catalyst employed.

_4~_

Claims (10)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A process for producing ethylene glycol which comprises reacting in a first reaction stage formaldehyde, carbon monoxide and hydrogen at superatmospheric pressure up to about 700 atmospheres and at temperatures in the range of about 75 to about 250°C. to obtain a glycol aldehyde containing reaction product, and subsequently catalytically reducing the glycol aldehyde containing reaction product to form ethylene glycol as the only detectable polyol, a catalytically effective amount of rhodium in complex combination with carbon monoxide being present at least during the first reaction stage.

2. A process according to claim 1, wherein the glycol aldehyde is catalytically hydrogenated to ethylene glycol.

3. Process according to claim 1, wherein the reactions are conducted in an aprotic organic amide.

4. Process according to claim 3, wherein the amide is an N-lower alkyl pyrrolidin-2-one.

5. Process according to claim 3, wherein the amide is N-methyl pyrrolidin-2-one.

6. Process according to claim 2, wherein the conver-sion to ethylene glycol by hydrogenating glycol aldehyde is con-ducted under hydrolytic conditions.

7. Process as in claim 6, characterized in that the hydrogenation is carried out in the presence of aqueous acid.

8. Process as in claim 7 characterized in that the acid is acetic acid.

9. Process as in claim 8 characterized in that the acid is about 50% aqueous acetic and the volume of aqueous acid is from about 10 to about 20 volume % based on the reaction mix-ture.

10. Process according to claim 1, 2 or 3, wherein the pressure in the first reaction stage is from about 250 to about 400 atmospheres and the temperature is from about 100°C. to 175°C.