CA1051819A - Electrolytic hydrodimerization of acrylonitrile using a nitrilocarboxylic acid compound - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE

In a process for hydrodimerizing an olefinic nitrile, amide or ester by electrolyzing an aqueous solution of the olefinic starting material, a conductive salt, where the for-mation of hydrogen at the cathode can be substantially inhi-bited and the current efficiency of the process significantly increased by including in the solution a nitrilocarboxylic acid compound such as a salt of ethylenediaminetetraacetic acid, there is provided the improvement of using quaternary phosphonium cations in the aqueous solution.

Description

This invention relates to an improvement in an electrolytic hydrodimerization process.
Production of paraffinic dinitriles, dicarboxamides or dicarboxylates by electrolytic hydrodimerization of an alpha, beta-olefinic nitrile, carboxamide or carboxylate is well known, e.g. from U.S. Patents 3,193,475-79 and 3,193,481-83 issued July 6, 1965 to M. M. Baizer. Although the process has been sufficiently attractive that it has been in commercial use for over nine years, efforts to develop improvements there-on have been continued with particular emphasis on loweringelectric power costs and mitigating electrode corrosion and fouling tendencies because of which it has been heretofore commercially preferable to carry out the process with a cell-dividing membrane.
With the object of maintaining high electrolyte con-ductivity while employing an electrolysis medium containing organic salts in a proportion small enough for attractive use of a single-compartment (membraneless) cell, one approach to improvement of the process has been to use as the electrolysis medium an aqueous solution of a mixture of quaternary ammonium and alkali metal salts together with the olefinic compound to be hydrodimerized. An example of such an approach is described in Netherlands Patent Application 66,10378 laid open for public inspection January 24, 1967, and further development thereof is described in U. S. Patent 3,616,321 issued October 26, 1971, to A. Verheyden et al. and U.S. Patent 3,689,382 issued September 5, 1972 to H. N. ~ox et al. However, all known variations of the process are characterized by some degree of inefficiency in use of the electrolyzing current, and this pro-blem is typically even more significant in those process varia-tions that utilize such an undivided cell.

105~819 For example, not all of the electroreduction that occurs at the cell cathode takes the form of the desired hydro-dimerization reaction or even the generally undesired simple hydrogenation of the olefinic starting material. Instead, a minor but significant proportion normally results in generation of molecular hydrogen. This hydrogen ordinarily accumulates in the electrolysis offgas together with oxygen produced at the anode and, in fact, the proportion of hydrogen in the off-gas is a fairly accurate indicator of the proportion of con-sumed electrolysis current that was wasted on such hydrogenproduction. At relatively low concentrations of hydrogen in the offgas, the percentage by volume of hydrogen in the offgas is generally about twice the percentage of current consumed in the electrolysis by undesired production of molecular hydrogen.
More specifically, the percentage of current consumed in the electrolysis by undesired production of molecular hydrogen is normally equal to fifty times the percentage by volume of hydrogen in the offgas divided by one hundred less the percen-tage by volume of hydrogen in the offgas, i.e., 50 x %H2/(100-%H2). For example, a concentration of 10% by volume of hydrogenin an elect~ lysis offgas usually indicates that about 5.5%
of the current consumed in the electrolysis was wasted on molecular hydrogen production and, accordingly, that the current efficiency of the hydrodimerization process was not possibly any greater than about 94.5%.
Clearly, the higher the proportion of the electroly-zing current that produces molecular hydrogen rather than the desired hydrodimer, the greater the cost of production of the hydrodimer will be. Accordingly, a process improvement where-by an olefinic compound from the aforementioned class can beelectrolytically hydrodimerized with a resultingly lowered production of molecular hydrogen and a thereby increased current ~0518~9 efficiency is highly desirable, and it is an object of this invention to provide such an improvement. Additional objects of the invention will be apparent from the following descrip-tion and Examples in which all percentages are by weight except where otherwise noted.
More particularly, in accordance with the process of the present invention, in a process for hydrodimerizing an olefinic compound having the formula R2C=CR-X wherein -X is -CN, -CONR2 or -COOR', R is hydrogen or R' and R' is Cl-C4 alkyl by electrolyzing an aqueous solution having dissolved therein at least about 0.1% by weight of the olefinic com-pound, and at least about 0.1% by weight of conductive salt in contact with a cathodic surface having a cathode potential sufficient for hydrodimerization of the olefinic compound, and wherein the formation of hydrogen at the cathodic surface can be substantially inhibited and the current efficiency of the process significantly increased by including in the solution at least one nitrilocarboxylic acid compound such as, for example, a nitriloacetic or nitrilopropionic acid compound having the formula Y2N ( Z YN )n R" - COOM wherein Y is a monovalent radical such as hydrogen, - R"- COOM, ( CH2 )m+l OH or Cl-C20 alkyl; -R"- is ( CH2 )m or --t- CHR"' ) R"'is hydroxy, -COOM, - ( CH2 )m COOM or Cl-Cg alkyl, hydroxyalkyl or hydroxphenyl; Z is a divalent C2-C6 hydro-carbon radical; M is a monovalent radical such as hydrogen, alkali metal or ammonium; m is 1 or 2; n is 0-4; and at least one Y is -R"-COOM or ( CH2 3m+lOH, there is provided the improvement of including in the aqueous solution at least about 10 5 gram mol per liter of monoquaternary phosphonium cations or multivalent, multiquaternary-phosphonium cations.
Olefinic compounds that can be hydrodimerized by the improved process of this invention include those having the ~05~8~9 structural formula R2C=CR-X wherein -X is -C~, -CO~R2 or -COOR', R is hydrogen or R' and R' is Cl-C4alkyl (i.e., methyl, ethyl, n-propyl, isopropyl, n-butyl, isobutyl or tert-butyl).
Compounds having that formula are known as having alpha, beta mono-unsaturation and in each such compound, at least one R
may be R' while at least one other R is hydrogen and at least one R', if present, may be an alkyl group containing a given number of carbon atoms while at least one other R', if present, is an alkyl group containing a different number of carbon atoms. Such compounds include olefinic nitriles such as, for example, acrylonitrile, methacrylonitrile, crotononitrile, 2-methylenebutyronitrile, 2-pentenenitrile, 2-methylenevalero-nitrile, 2-methylenehexanenitrile, tiglonitrile or 2-ethyli-denehexanenitrile; olefinic carboxylates such as, for example, methyl acrylate, ethyl acrylate or ethyl crotonate; and olefinic carboxamides such as, for example, acrylamide, metha-crylamide, N-N-diethylacrylamide or ~,N-diethylcrotonamide.
Best results are generally obtained when the olefinic compound has at least one hydrogen atom directly attached to either of the two carbon atoms joined by the double bond in the afore-described structural formula. Also presently of greater util-ity in the process of this invention are those olefinic com-pounds wherein R' in that formula is methyl or ethyl, and par-ticularly acrylonitrile, methyl acrylate and alpha-methyl acrylonitrile.
Products of hydrodimerization of such compounds have the structural formula X-CHR-CR2-CR2-CHR-X wherein X and R
have the aforesaid significance, i.e., paraffinic dinitriles such as, for example, adiponitrile and 2,5-dimethyladiponitrile;
paraffinic dicarboxylates such as, for example, dimethyladi-pate and diethyl-3,4-dimethyladipate; and paraffinic dicarbox-amides such as, for example adipamide, dimethyladipamide and ~,N'-dimethyl-2,5-dimethyladipamide. Such hydrodimers can be employed as monomers or as intermediates convertible by known processes into monomers useful in the manufacture of high molecular weight polymers including polyamides and polyesters.
The dinitriles, for example, can be hydrogenated by known processes to prepare paraf~finic diamines especially useful in the production of high molecular weight polyamides. Other examples of various olefinic compounds that can be hydrodi-merized by the process of this invention and the hydrodimers thereby produced are identified in the aforecited U. S. Patent Nos. 3,193,475-79 and '481-83.
The invention is herein also described in terms of electrolyzing an aqueous solution having dissolved therein certain proportions of the olefinic compound to be hydrodimer-ized, the aforedescribed quaternary phosphonium cations and a conductive salt. Such use of the term "aqueous solution" does not imply, however, that the electrolysis medium may not also contain an undissolved organic phase. To the contrary, the process of this invention can be quite satisfactorily carried out by electrolyzing the aqueous solution in an electrolysis medium containing the recited aqueous solution and a dispersed but undissolved organic phase in any proportions at which the aqueous solution is the continuous phase of the electrolysis medium. Hence in some embodiments of the invention, the aqueous solution may be suitably electrolyzed in an electroly-sis medium containing essentially no undissolved organic phase, by which is meant either no measurable amount of undissolved organic phase or a minute proportion of undissolved organic phase having no significant effect on the hydrodimer selectivity achieved when the aqueous solution is electrolyzed in accord-ance with the process of this invention. Such a minute propor-tion, if present, would be typically less than 5% of the combined weight of the aqueous solution and the undissolved organic phase in the electrolysis rnedium. In other embodi-ments, the process of this invention can be carried out by electrolyzing the aqueous solution in an electrolysis medium consisting essentially of the recited aqueous solution and a dispersed but undissolved organic phase in a larger proportion (e.g. up to about 15%, 20% or even more of the combined weight of the aqueous solution and the undissolved organic phase in the electrolysis medium) which may or may not significantly affect the hydrodimer selectivity depending on other conditions of the process. In some continuous process embodiments in-volving recycle of unconverted olefinic compound and whether present in a minute or larger proportion, such an organic phase is normally made up mainly (most commonly at least about 65%
and even more typically at least about 75%) of the olefinic compound to be hydrodimerized and the hydrodimer product with some small amounts of organic hydrodimerization by-products, quaternary ammonium or phosphonium cations, etc. possibly also present. Typically, such an organic phase contains at least about 10%, preferably-between about 15% and about 50%, and even more desirably between about 20% and about 40% of the olefinic compound to be hydrodimerized. In any event, however, the concentrations of the constituents dissolved in the aqueous solution to be electrolyzed, as set forth herein, are with reference to the recited aqueous solution alone and not the combined contents of said aqueous solution and an undissolved organic phase which, as aforesaid, may be present but need not be present in the electrolysis medium as the invention is carried out. On the other hand, the weight percentages of undissolved organic phase described herein are based on the combined weight of the aqueous solution and the undissolved organic phase in the electrolysis medium.

~OS~819 Referring to the constituents of the aqueous phase, the olefinic compound to be hydrodimerized will be present in at least such a proportion that electrolysis of the solution, as described herein, results in a substantial amount of the desired hydrodimer being produced. That proportion is gener-ally at least about 0.1% of the aqueous solution, more typic-ally at least about 0.5% and, in some embodiments, preferably at least about 1% of the aqueous solution. Inclusion of one or more additional constituents which increase the solubility of the olefinic compound in the solution may permit carrying out the process with the solution containing relatively high proportions of the olefinic compound, e.g. at least about 5%
or even 10% or more, but in most embodiments, the aqueous solution contains less than about 5% (e.g. not more than 4.5%) of the olefinic compound and, in many of those embodiments, preferably not more than about 1.8% of the olefinic compound.
In accordance with this invention, the minimum required proportion of quaternary phosphonium cations is very small. In general, there need be only an amount sufficient to provide the desired hydrodimer selectivity (typically at least about 75%) although much higher proportions can be present if desired or convenient. In most cases, the quaternary phos-phonium cations are present in a concentration of at least about 10 5 gram mol per liter of the aqueous solution. Even more typically their concentration is at least about 10 4 gram mol per liter of the solution and, in some embodiments employing monovalent mono-quaternary phosphonium cations, pre-ferably at least about 5 x 10 4 gram mol per liter. Although higher proportions may be present in some cases, as aforesaid, the quaternary phosphonium cations are generally present in the aqueous solution in a concentration not higher than about 0.5 gram mol per liter and even more usually not higher than about 105~8~

gram mol per liter. In some preferred embodiments, the concentration of quaternary phosphonium cations in the solu-tion is between about 10 4 and about 10 2 gram mol per liter.
The quaternary phosphonium cations that are present in such concentrations are thosepositively-charged ions in which a phosphorous atom has a valence of five and is directly linked to other atoms (e.g. carbon) satisfying four fifths of tha~ va~ence. Such cations need contain only oS,~7~ofc/~
one pentavalent ~}~e~ou_ atom as in, for example, various mono-quaternary phosphonium (e.g. tetraalkylphosphonium) cations, but they may contain more than one pentavalent atoms, as in, for example, various multivalent multiquaternary phos-phonium cations such as the bis-quaternary phosphonium cations, e.g. polymethylenebis(trialkylphosphonium) cations. Mixtures of such monovalent phosphonium cations and multivalent quater-nary phosphonium cations can also be used. Suitable mono-quaternary phosphonium cations may be cyclic, but they are more generally of the type in which a pentavalent phosphorus atom is directly ;inked to a total of four monovalent organic groups preferably devoid of olefinic unsaturation and desir-ably selected from the group consisting of alkyl and aryl radicals and combinations thereof. Suitable multi-quaternary phosphonium cations may likewise be cyclic, and they are typi-cally of a type in which the pentavalent phosphorus atoms are linked to one another by at least one divalent organic (e.g.
polymethylene) radical and each further substituted by monovalent organic groups of the kind just mentioned sufficient in number (normally two or three) that four fifths of the valence of each such pentavalent atom is satisfied by such divalent and mono-valent organic radicals. As such monovalent organic radicals, suitable aryl groups contain typically from six to twelve ~5181C~
carbon atoms and preferably only one aromatic ring as in, for example, a phenyl or benzyl radical, and suitable alkyl groups can be straight-chain, branched or cyclic with each typically containing from one to twelve carbon atoms. Although quater-nary phosphonium cations containing a combination of such alkyl and aryl groups (e.g. benzyltriethylphosphonium ions) can be used, many embodiments of the invention are preferably carried out with quaternary cations having no olefinic or aromatic unsaturation. Good results are generally obtained with tetraalkylphosphonium ions containing at least three C2-C6 alkyl groups and a total of from 8 to 24 carbon atoms in the four alkyl groups, e.g. tetraethyl-, ethyltripropyl-, ethyltributyl-, ethyltriamyl-, ethyltrihexyl-, octytriethyl-, tetrapropyl-, methyltripropyl-, decyltripropyl-, methyltri-butyl-, tetrabutyl-, amyltributyl-, tetraamyl-, tetrahexyl-, ethyltrihexyl-, diethyldioctyl-phosphonium and many others referred to in the aforecited U. S. Patent Nos. 3,193,~75-79 and '481-83. Generally most practical from the economic stand-point are those C8-C20 tetraalkylphosphonium ions containing at least three C2-C5 alkyl groups, e.g. methyltributyl-, tetra-propyl-, ethyltriamyl-, octyltriethylphosphonium, etc. Parti-cularly useful are the C8-C16 tetraalkylphosphonium ions con-taining at least three C2-C4 alkyl groups.
Similarly good results are obtained by use of the divalent polymethylenebis(trialkylphosphonium) ions, particu-larly those containing a total of from 17 to 36 carbon atoms and in which each trialkylphosphonium radical contains at least two C3-C6 alkyl groups and the polymethylene radical is C3-C8, i.e., a straight chain of from three of eight methylene radi-cals.
Presently most attractive from the economic stand-point are the C18-C32 polymethylenebis(trialkylphosphonium) ions ~0518~9 in which each trialkylphosphonium radical contains at least two C3-C5 alkyl groups and the polymethylene radical is C4-C6.
In many embodiments of the invention employing such polymethyl-enebis(trialkylphosphonium) ions, the carbon atom content of such ions is preferably from 20 to 34.
Presently of specific interest for potential com-mercial use in the process of this invention are the C20-C34 hexamethylenebis(trialkylphosphonium) ions, e.g. those in which each trialkylphosphonium radical contains at least two C3-C6 alkyl groups. Also generally preferred are the hexamethylenebis (trialkylphosphonium) ions containing from 20 to 30 carbon atoms, e.g. those in which each trialkylphosphonium radical contains at least two C3-C5 alkyl groups, and especially the C24-C30 hexamethylenebis(trialkylphosphonium) ions in which each trialkylphosphonium radical contains at least one and preferably two n-butylgroups. Any of such cations can be in-corporated into the aqueous solution to be electrolyzed in any convenient manner, e.g. by dissolving the hydroxide or a salt (e.g. a Cl-C2 alkylsulfate) of the desired quaternary phosphonium cation in the solution in the amount required to provide the desired concentration of such cations.
One significant advantage of the polymethylenebis (trialkylphosphonium) ions for use in the present invention is that relative to most of the corresponding tetraalkylphosphon-ium ions of the type described hereinbefore, they tend to dis-tribute themselves in higher proportion toward the aqueous phase of a mixture of an aqueous solution of the type electroly-zed in accordance with the present invention and the undissol-ved organic phase which, as aforesaid, may be present in the 105~8~9 aqueous solution during the electrolysis. Whether or not such an organic phase is present in substantial proportion in the aqueous solution during the electrolysis, product hydrodimer is generally most conveniently removed from the electrolyzed solution by adding to the solution (either before or after the electrGlysis) an amount of the olefinic starting material in excess of its solubility therein, mixing the solution and the excess olefinic compound until they are substantially equili-brated, and then separating (e.y. decanting) from the resulting mixture a first portion thereof that is richer than said mix-ture in the olefinic compound and therefore richer than said mixture in the hydrodimer product which is normally substan-tially more soluble in the olefinic compound than in the electrolyzed aqueous solution. ~ormally, the hydrodimer pro-duct is separated from said first portion of the mixture (e.g.
by distillation) while a second portion of the mixture com-prising an aqueous solution of the type subjected to electroly-sis in accordance with the present invention is recycled and the aqueous solution comprised by said second portion is sub-jected to more of such electrolysis. In process embodimentsin which the hydrodimer product is separated from the electro-lyzed solution in the manner just described and in view of the importance of having sufficient quaternary phosphonium cations in the aqueous solution to maintain a high hydrodimer selecti-vity on further electrolysis of the solution, the use of a quaternary cation that distributes itself in relatively high proportion in the aqueous portion of a substantially equili-brated mixture of the type just described is highly attractive from the standpoint of lessening the costs of recovering such cations from the separated (e.g. decanted) organic portion of the mixture and/or loss of such cations due to incomplete recovery from said organic portion of the mixture. Surprisingly, and despite their generally higher carbon content, various bis-quaternary cations of the class definea hereinbefore have been found to distribute themselves toward the aqueous solu-tion in ratios significantly higher (e.g. up to at least 3-4 times higher) than those of the corresponding mono-quaternary cations.
The type of conductive salt employed is not usually critical to inhibition of hydrogen formation by use of a nitrilocarboxylic acid compound as described herein. Hence the conductive salt can be a quaternary phosphonium salt such as, for example, a tetraalkylphosphonium phosphate, sulfate, alkylsulfate, (e.g. ethylsulfate) or arylsulfonate (e.g. toluene sulfonate). Although organic salts of that general type can be employed as the conductive salt in a divided or single-compartment (undivided) cell, it is generally preferred to use an alkali metal conductive salt, i.e., a salt of sodium, potassium, lithium, cesium or rubidium, especially in undivided electrolytic hydrodimerization (EHD) cells, and many attractive embodiments of the invention are carried out with enough alkali metal salt dissolved in the aqueous solution to provide alkali metal cations constituting more than half of the total weight of all cations in the solution. When such alkali metal salts are used, those of lithium and especially sodium and potassium are generally preferred for economic reasons.
Also preferred for such use are the salts of inorganic and/or polyvalent acids, e.g. a tetraalkylphosphonium or alkali metal orthophosphate, borate, perchlorate, carbonate or sulfate and particularly an incompletely-substituted salt of that type, e.g., a salt in which the anion has at least one valence satisfied by hydrogen and at least one other valence satisfied by an alkali metal. Examples of such salts include disodium phosphate (Na2HP04), potassium acid phosphate (KH2P04), sodium bicarbonate (NaHC03) and dipotassium borate (K2HB03). Also useful are the alkali metal salts of condensed acids such as pyrophosphoric, metaphosphoric, metaboric, pyro-boric and the like (e.g. sodium pyrophosphate, potassium meta-borate, borax, etc.) and/or products of hydrolysis of such condensed acid salts. Depending on the acidity of the solution to be electrolyzed, the stoichiometric proportions of such anions and alkali metal cations in the solution may correspond to a mixture of two or more of such salts, e.g. a mixture of sodium acid phosphate and disodium phosphate, and accordingly, such mixtures of salts (as well as mixtures of salts of dif-ferent cations, e.g. different alkali metals, and/or different acids, e.g. phosphoric and boric) are intended to be within the scope of the expressions "conductive salt" and "alkali metal phosphate, borate, perchlorate, carbonate and sulfate" as used herein. Any of the alkali metal salts may be dissolved in the aqueous solution as such or otherwise, e.g. as the alkali metal hydroxide and the acid necessary to neutralize the hydroxide to the extent of the desired acidity of the aqueous solution.
The concentration of conductive salt in the solution should be at least sufficlent to substantially increase the electrical conductivity of the solution above its conductivity without such a salt being present. In most cases, a concentra-tion of at least about 0.1% is favored. More advantageous con-ductivity levels are achieved when the solution has dissolved therein at least about 1% of the conductive salt or, even more preferably, at least about 2% of such a salt. In many cases, optimum process conditions include the solution having dis-solved therein more than 5% (typically at least 5.5%) of theconductive salt. The maximum amount of salt in the solution is typically limited only by its solubility therein, which 105181g varies with the particular salt employed. With salts such as sodium or potassium phosphates and/or borates, it is generally most desirable that the solution contain between about 8%
and about 15% of such a salt or mixture thereof.
As aforesaid, generation of molecular hydrogen at the cathode of a process of the type discussed herein can be sub-stantially inhibited by including in the aqueous electrolysis medium at least one nitrilocarboxylic acid compound such as, for example, a nitriloacetic or nitri~lopropionic acid compound having the formula Y2N ( Z - YN ~~-n R" -COOM wherein Y is a monovalent radical such as hydrogen, -R"-COOM, + CH2 ~ lOH
or Cl-C20alkyl (preferably Cl-Cl0 alkyl such as ethyl, n-propyl, tert-butyl, n-hexyl, n-decyl, etc.); -R" is ~CH2 ~ or ~ CHR'''-~-; R''' is hydroxy, -COOM, ( CH2 ~ COOM or Cl-C8 alkyl, hydroxyalkyl (e.g. hydroxyethyl) or hydroxyphenyl (e.g.
ortho-hydroxyphenyl); Z is a divalent C2-C6 hydrocarbon (e.g.
alkylene) radical such as, for example, n-hexylene, n-butylene, iso-butylene or, generally more desirably, ethylene or n-propy-I lene; M is a monovalent radical such as hydrogen, an alkali metal (e.g. lithium or, usually more desirably, sodium or potassium)or phosphonium; m is l or 2; n represents the number of repeating --~Z - YN ) groups, if any, and may be 0, 1, 2, 3 or 4; and at least one Y in the formula is -R"-COOM or -~ CH2-~--+1OH group in addition to the -R"-COOM group on the right hand end of the formula as shown hereinbefore. At least one such additional -R"-COOM or --~CH2-~m+lOH group is usually desirably attached to the nitrogen atom at the left-hand end of the formula but when n is l, such an additional group may be attached (alternatively or otherwise) to the nitrogen atom in the ( Z - YN ) unit, and when n is 2, 3 or 4, any one or more of the nitrogen atoms in the repeating ( Z - YN ) units may have such an additional -R"-COOM or ( CH2 } +lOH group attached thereto.

1051~319 Preferably, but not necessarily, the nitrilocarboxy-lic acid compound is an aminopolycarboxylic acid compound, i.e.
one in which there are at least two -R"-COOM groups. It is also generally desirable for Y to be C2-C4 alkylene and for n to be 0, 1, 2 or 3 (even more desirably 0, 1 or 2 and most preferably 1 or 2). Representative of such compounds are nitrilotriacetic acid, diethylenetriaminepentaacetic acid, N,N-di(2-hydroxyethyl)-glycine, ethylenediaminetetrapropionic acid, N-N'-ethylenebis[2-(o-hydroxyphenyl)]glycine and, typically most favored, ethylenediaminetetraacetic acid and N-hydroxy-ethylethylenediaminetriacetic acid (hereinafter sometimes represented as EDTA and HEDTA, respectively). In the low concentrations generally employed, they may be added to the electrolysis medium as acids, or, usually more conveniently and particularly at the alkaline pH' s favored for most embodiments of the invention, as partially or fully neutralized salts there-of (e.g. the water-soluble phosphonium or alkali metal salts of such acids). In accordance with procedures known in the art, alkali metal salts of such nitrilocarboxylic acid com-pound can be prepared by reacting an appropriate amine (e.g.ethylenediamine) with an alkali metal salt of a chloracetic acid in the presence of an alkali metal hydroxide, or with hydrogen cyanide and formaldehyde and then an alkali metal hydroxide, or with ethylene glycol to provide hydroxyethyl substituents of nitrogen atom(s) of the amine and then reacting the hydroxyethyl-substituted amine with an alkali metal hydroxide in the presence of cadmium oxide to con-vert the hydroxethyl substitutents to alkali metal acetate substituents in the proportion desired, or with acrylonitrile in the present of a base (e.g. sodium hydroxide) and then hydrolyzing the cyanoethylated amine in the presence of an alkali metal hydroxide. Conveniently utilized salts of EDTA, HEDTA and other such nitrilocarboxylic acid compounds are also ~051~9 available commercially. See "Keys to Chelation", Dow Chemical Company, Midland, Michigan (1969).
The minimum concentration of the nitrilocarboxylic acid compound in the aqueous electrolysis medium is only that sufficient to inhibit formation of molecular hydrogen at the 1~
cathodic surface 4~ the process. In general, at least about 0.025 millimol of the nitrilocarboxylic acid compound per liter of the solution is desirable and at least about 0.1 millimol per liter is preferred. In most cases having a greater attrac-tion for commercial use, at least about 0.5 millimol per literis more desirable and at least about 2.5 millimols per liter usually provide even better results. Generally, not more than about 50 millimols per liter are required, although higher concentrations may be employed if desired. Even more typically, economic results are better when the concentration of the nitrilocarboxylic acid compounds in the solution is not greater than 25 millimols per liter. With reference to such concentra-tions, it should be understood that the nitrilocarboxylic acid compounds used herein may degrade under the conditions of the process, e.g. to compounds that have lower molecular weight and/or fewer -R"-COOM or ( CH2~--+-lOH groups but which never-theless provide the advantages of this invention in substantial measure, and accordingly such degradation products should be considered as equivalent to the undegraded nitrilocarboxylic acid compounds to the extent that they provide the advantages thereof, when measuring or otherwise identifying a nitrilocar-boxylic acid compound concentration with reference to the pro-cess of this invention. Mixtures of two or more of the afore-described nitrilocarboxylic acid compounds may also be used in the process of this invention and accordingly, such mixtures are meant to be within the scope of the expression "a nitrilo-carboxylic acid compound" as used in this disclosure and the appended claims.

In substantial measure when carrying out the present process in a cell divided by a cation-permeable membrane and particularly when carrying out the process in a single-compartment cell, generation of hydrogen at the cathode is even more significantly inhibited by including in the electroly-sis medium a boric acid, a condensed phosphoric acid or an alkali metal salt thereof. The boric acid or borate may be added to the solution as orthoboric acid, metaboric acid or pyroboric acid and then neutrallzed to the desired solution pH, e.g. with an alkali metal (preferably the cation of the conduc-tive salt) hydroxide or as a completely or incompletely sub-stituted alkali metal salt of such an acid (e.g. disodium or monosodium orthoborate, potassium metaborate, sodium tetra-borate or the hydrated form thereof commonly called borax).
The condensed phosphoric acid~r pho-sphate may be added as a polyphosphoric (e.g. pyrophosphoric or triphosphoric) acid and then neutralized to the desired solution pH or as a completely or incompletely substituted alkali metal salt thereof (e.g.
tetrasodium pyrophosphate or potassium hexametaphosphate or triphosphate).
In general, the condensed phosphoric acids and their alkali metal salts tend to hydrolyze in the electrolysis medium at rates dependent on their concentration, the solution pH, etc.
It is believed, however, that the products of such hydrolysis continue to inhibit the generation of hydrogen at the cathode so long as they remain condensed to at least some degree, i.e., so long as they have not been hydrolyzed to the orthophosphate form, and hence the preferred concentrations of such condensed phosphoric acid compounds are herein expressed in terms of weight percent of a condensed phosphoric acid (which may be that originally added to the solution or hydrolysis products thereof having a lower but conventionally recognizable degree ~os~9 of molecular condensation) or the molar equivalent of an alkali metal salt thereof. When such a condensed phosphoric acid is used in the process of this invention, and particularly in an undivided cell having a metallic anode (e.g. an anode compris-ing a ferrous metal such as carbon steel, alloy steel, iron or magnetite), it is generally advantageous for the solution to contain at least about 0.01%, preferably between about 0.02%
and about 3%, and often most desirably between about 0.02%
and about 2% of the condensed phosphoric acid or the molar equivalent (molecularly equivalent amount) of an alkali metal salt thereof.
The aforementioned boric acids and alkali metal salts thereof, on the other hand, tend to relatively rapidly form in the electrolysis medium a variety of boron-containing ions having relative proportions normally dependent on their concen-trations, the solution pH, etc., and generally including both uncondensed (i.e., orthoborate) and condensed (e.g. metaborate, tetraborate, polymeric ring-containing, etc.) ions, regardless of whether the acids and/or salts originally added to the electrolysis medium were in condensed or uncondensed form at that time. In other words, condensed borates (e.g. tetrabor-ates) normally convert in the electrolysis medium in part to orthoborate ions and in part to other condensed borate ions, while orthoborates added as such generally form various con-densed borate ions, depending largely on the solution pH, etc.
In any event, it appears that the boron-containing ions are effective for purposes of this invention whether they are present in condensed or uncondensed forms or a mixture thereof and accordingly, preferred concentrations of the boric acids or salts are herein expressed (on the basis of one liter of solution) in terms of gram atoms of boron which may be present in the ionic form of condensed or uncondensed borates or other boron-containing moieties provided by interaction between the electrolysis medium and the boric acids and/or salts added thereto. When such bori~ acids or salts are used in the pro-cess of this invention, and particularly in an undivided cell having a metallic anode (e.g. an anode comprising a ferrous metal such as carbon steel, alloy steel, iron or magnetite), it is generally desirable for the boron concentration in the electrolysis medium to be at least about 0.01 and preferably 0.02 gram atom of boron per liter of solution. It is gener--ally not necessary that the boron concentration in the solution be greater than about 0.9 gram atom per liter and in many cases it need not be greater than about 0.5 gram atom per liter, although higher concentrations are not necessarily detrimental and may be advantageous, e.g. if it is intended that a boric acid salt provide a substantial portion of the electrical conducitivity of the electrolysis medium.
In most cases, the pH of the bulk of the electrolysis medium is at least about two, preferably at least about five, more preferably at least about six and most conveniently at least about seven, especially when the process is carried out in an undivided cell having a metallic anode. On the other hand, the overall solution pH is generally not higher than about twelve, typically not higher than about eleven and, with the use of sodium or potassium phosphates and/or borates, generally not substantially higher than about ten.
The temperature of the solution may be at any level compatible with extence of such of the solution itself, i.e., above its freezing point but below its boiling point under the pressure employed. Good results can be achieved between about 5 and about 75C. or at even higher temperatures if pressures substantially above one atmosphere are employed. The optimum temperature range will vary with the specific olefinic compound lOS1819 and hydrodimer, among other factors, but in hydrodimerization of acrylonitrile to adiponitrile, electrolysis temperatures of at least about 25 are usually preferred and those between about 40 and about 65 C. are especially desirable.
As is well-known, electrolytic hydrodimerization of an olefinic compound having a formula as set forth hereinbefore must be carried out in contact with a cathodic surface having a cathode potential sufficient for hydrodimerization of that compound. In general, there is no minimum current density with which the present process must be carried out at such a cathodic surface but in most cases, a current 2ensity of at least about 0.01 amp per square centimeter (amp/cm2) of the cathodic surface is used and a current density of at least about O.OS amp/cm2 is usually preferred. Although hi'gher cur-rent densities may be practical in some instances, those generally employed in the present process are not higher than about 1.5 amp/cm and even more typically not higher than about 0.75 amp/cm of the aforedescribed cathodic surface.
Depending on other process variables, current densities not higher than about 0.5 amp/cm2 may be preferred in some embodi-ments of the process.
Although not necessary, a liquid-impermeable cathode is usually preferred. With the use of such a cathode, the aqueous solution to be electrolyzed is generally passed between the anode and cathode at a linear velocity with reference to the adjacent cathodic surface of at least about 0.3 meter per second, preferably at least about 0.6 meter per second and even more preferably between about 0.9 and about 2.4 meters per second although a solution velocity up to 6 meters per second or higher can be employed if desired. The gap bet~.een the anode and cathode can be very narrow, e.g. about 1 millimeter or less, or as wide as 12.5 millimeters or even wider, but is usually most conveniently of a width between about 1.5 and about 6.2 millimeters. In the process of this invention, the cathodic surface can be made of virtually any material at which the requisite cathode potential can be provided and which is not dissolved or corroded at an intolerable rate. In general, the process can be carried out with a cathode consisting essenti-ally of cadmium, mercury, thallium, lead, zinc, manganese, tin (possibly not suitable with some nitrile reactants) or graphite, by which is meant that the cathodic surface contains a high percentage (generally at least about 95% and preferably at least about 98%) of one or a combination (e.g. an alloy) of two or more of such materials, but it may contain a small amount of one or more constituents that do not alter the nature of the cathodic surface so as to prevent substantial realiza- , tion of the advantages of the present invention, particularly as described herein. Such other constituents, if present in substantial concentration, are preferably other materials having relatively high hydrogen overvoltages. Of particular preferance are cathodes consisting essentially of cadmium, lead, zinc, manganese, graphite or an alloy of one of such metals, and especiaily cathodes consisting essentially of cadmium. Best results are usually obtained with a cathodic surface having a cadmium content of at least about 99.5%, even more typically at least about 99,9% in ASTM Designation B440-66T (issued 1966).
Cathodes employed in this invention can be prepared by various techniques such as, for example, electroplating of the desired cathode material on a suitably-shaped substrate of some other material, e.g. a metal having greater structural rigidity, or by chemically, thermally and/or mechanically bonding a layer of the cathode material to a similar substrate.
Alternatively, a plate, sheet, rod or any other suitable ~05~8~9 configuration consisting essentially of the desired cathode material may be used without such a substrate, if convenient.
The process of this invention can be carried out in a divided cell having a cation-permeable membrane, diaphragm, or the like separating the anode and cathode compartments of the cell in such a way that the aqueous solution containing the olefinic compound undergoing hydrodimerization at the cathode of the cell is not in simultaneous direct contact with an anode of the cell. However, it is especially advantageously carried out in cells not divided in that manner, i.e., in cells which the solution being electrolyzed is in direct physical contact with an anode and cathode of the cell. In fact, and particularly without the presence of a boric or condensed phosphoric acid or salt thereof in the preferred concentrations described hereinbefore, it has been found that the aforemen-tioned nitrilocarboxylic acid compounds, and especially in the concentrations cited hereinbefore, generally substantially inhibit the corrosion of metallic anodes when used in such undivided cells. Anodes whose corrosion may be thereby inhi-bited include those composed of the heavy metals (i.e., metalshaving a specific gravity greater than 4.0) such as, for example, platinum, ruthenium, nickel, lead, lead dioxide and, of particular advantage when the conductive salt is a phos-phate, borate or carbonate, ferrous materials such as carbon steels, alloy steels, iron and magnetite.
In fact, an especially preferred embodiment of the invention is carried out in an undivided cell having an anode comprising a ferrous metal and with the use of an alkali metal phosphate, borate or carbonate conductive salt and an electroly-sis medium having a pH not substantially below seven. Ofpotential interest from the economic standpoint are those embodiments employing an anode consisting essentially of ~051819 carbon steel, exemplary compositions of which are listed in the 100, llU0 and 1200 series of American Iron and Steel Insitute and Society of Automotive Engineers standard steel composition numbers, many of which may be found on page 62 of Volume 1, Metals Handbook, 8th Edition (1961) published by the American Society for Metals, Metals Park, Ohio.
In general, the carbon steels that are advantageously used as anode materials in the process of this invention con-tain between about 0.02% carbon (more typically at least about 0.05% carbon) and about 2% carbon. Normally, carbon steels such as those of the AISI and SAE 1000 series of standard steel composition numbers are preferred and those containing between about 0.1% and about 1.5% carbon are typically most desirable. Regardless of the material from which it is made, each anode in the cell may be in the form of a plate, sheet, strip, rod or any other configuration suitable for the use intended. In a preferred embodiment, however, the anode is in the form of a sheet (e.g. of cold-rolled carbon steel) essen-tially parallel to and closely spaced from a cathodic surface of approximately the same dimensions.
Although the invention described and claimed herein is not to be regarded as limited to any particular mechanism proposed therefor, it is presently believed that the nitrilo-carboxylic acid compounds (and probably to a lesser extent, if present, the boric and/or condensed phosphoric acid compounds) at least partially sequester heavy metals which tend to accumu-late in the electrolysis medium (e.g. as a result of corrosion of the cathode and/or, with use of an undivided cell, corrosion of the anode) and that such sequestration inhibits the deposi-tion of those metals on the cathode of the cell. It is furtherbelieved that unsequestered heavy metals (or oxides and/or hydroxides thereof) tend to form colloidal particles in the 105181~
electrolysis medium and after such deposition, alter the nature of the cathodic surface so as to increase the generation of molecular hydrogen at the expense of process current effi-ciency. Those beliefs are mainly based on observations that increases in hydrogen production normally accompany increased deposition on the cathode of a relatively dense precipitate which has been identified as essentially completely composed of such heavy metals (principally iron in an undivided cell having a steel anode) and their oxides and hydroxides, and that deposition of the precipitate is substantially inhibited by use of the process improvement described and claimed herein.
The following specific examples of the process of this invention are included for purposes of illustration only and do not imply any limitations on the scope of the invention.
Also in these examples, acrylonitrile and adiponitrile are generally represented by AN and ADN, respectively.
EXAMPLE I
In a continuous process, a liquid electrolysis med-ium composed about 99% by (l) an aqueous solution having dis-solved therein between 1.4% and 1.6% AN, about 1.2% ADN, 10%
of a mixture of sodium orthophosphates, 0.6-1.4 x 10 3 mole per liter of methyltributylphosphonium ions, about 0.5% (14.2 millimoles per liter) of Na4EDTA and the sodium borates pro-duced by neutralizing orthoboric acid in an amount corresponding to about 2% of the solution to the solution of pH of about 8.5 and about 1% by (2) a dispersed but undissolved organic phase containing 27-29% AN, 54-58% ADN, 7-9% AN EHD by-products and 3% water was circulated at 55C. and 1.22 meters per second through an undivided electrolytic cell having an AISI 1020 carbon steel anode separated by a gap of 1.76 millimeters from a cadmium cathode composed of cadmium conforming to ASTM

Designation B 440-66T (at least 99.9% Cd), and electrolyzed as it passed through the cell with a current density of 0.185 amp/cm2 of the surface of the cathode. Organic phase containing product ADN, AN, EHD by-products and unreacted AN
was separated from the electrolyzed medium and make-up AN was added after which the medium was recirculated through the cell and electrolyzed again under the conditions just described.
For each Faraday of current passed through the medium, 0.4 millimole of Na4-EDTA was added to the circulating medium and about 12 grams of the solution were purged from the system and replaced with water containing sufficient dissolved methyltri-butylphosphonium ions and sodium orthophosphates and borates to maintain the concentrations of those constituents of the solution at the aforedescribed levels and the total volume of the medium essentially constant. After 120 hours of electroly-sis under those conditions, it was found that AN had been con-verted to ADN with average and final selectivities of 88%, the steel anode had corroded at an average rate less than 0.5 millimeter per year and the volume percent of hydrogen in the offgas had averaged below 1% with a final value of 0.8%.
EXAMPLE II
In processes essentially as described in Example I
except that the quaternary cations in the aqueous solution are any one or a mixture of those identified below, instead of hexamethylenebis(ethyldibutylphosphonium) ions, the results average and final selectivities of AN conversion to ADN and anodic surface corrosion are substantially the same as those obtained in Example I:

Hexamethylenebis(tributylphosphonium) Hexamethylenebis(amyldipropylphosphonium) Hexamethylenebis(tripropylphosphonium) Hexamethylenebis(methyldibutylphosphonium) Hexamethylenebis(ethyldihexylphosphonium) Hexamethylenebis(decyldiethylphosphonium) Pentamethylenebis(propyldibutylphosphonium) Pentamethylenebis(triamylphosphonium) Tetramethylenebis(ethyldibutylphosphonium) Tetramethylenebis(octyldipropylphosphonium) Heptamethylenebis(ethyldibutylphosphonium) EXAMPLE III
In processes essentially as described in Example I
except that the quaternary cations in the aqueous solution are any one or a mixture of those identified below instead of methyltributylphosphonium ions the results average and final selectivities of AN conversion to ADN and anodic surface corrosion are substantially the same as those obtained in Example I:
Amyltributylphosphonium Tetrapropylphosphonium Diethyldihexylphosphonium Decyltriethylphosphonium Propyltributylphosphonium Tetraamylphosphonium Ethyltributylphosphonium Octyltributylphosphonium Diethyldibutylphosphonium

Claims (30)

The embodiments of the invention in which an exclu-sive property or privilege is claimed are defined as follows:

1. In a process for hydrodimerizing an olefinic com-pound having the formula R2C=CR-X wherein -X is -CN, -CONR2 or -COOR', R is hydrogen or R', R' is C1-C4 alkyl and at least one R directly attached to either of the two carbon atoms joined by the double bond in said formula is hydrogen by electrolyzing an aqueous solution having dissolved therein at least about 0.1% by weight of said olefinic compound and at least about 0.1% by weight of conductive salt in contact with a cathodic surface having a cathode potential sufficient for hydrodimerization of said olefinic compound, and wherein there is included in the solution between about 0.1 and about 50 millimols per liter of a nitrilocarboxylic acid compound having the formula wherein Y is hydrogen, , or C1-C20 alkyl;
is or ; is hydroxy, , or C1-C8 alkyl, hydroxyalkyl or hydroxyphenyl; Z is a divalent C2-C6 hydrocarbon radical; M is hydrogen, alkali metal or ammonium; m is 1 or 2; n is an integer from 0 to 4 and at least one Y is or , the improvement comprising utilizing at least about 10-5 gram mol per liter of mono-quaternary phosphon-ium or multivalent, multi-quaternary phosphonium ions in said aqueous solution.

2. In a process for hydrodimerizing an olefinic compound having the formula R2C=CR-X wherein -X is -CN, -CONR2 or -COOR', R is hydrogen or R' and R' is C1-C4 alkyl by electrolyz-ing an aqueous solution having dissolved therein at least about 0.1% by weight of said olefinic compound and at least about 0.1%
by weight of conductive salt in contact with a cathodic surface having a cathode potential sufficient for hydrodimerization of said olefinic compound said solution also having in a concentra-tion sufficient to inhibit formation of hydrogen at the cathodic surface, 0.1 to 50 millimols per liter of solution of a nitrilocarboxylic acid compound having the formula surface, a nitrilocarboxylic acid compound having the formula wherein Y is hydrogen, , or C1-C20 alkyl, Z is a divalent C2-C6 hydrocarbon radical, M is hydrogen, alkali metal or ammonium, m is 1 or 2, n is an integer from 0 to 4 and at least one Y is or ; the improvement comprising said solution further containing at least about 10-5 gram mol per liter of quaternary phosphonium cations.

3. The process of Claim 1 wherein Y is or , Z is C2-C4 alkylene, m is 1 and n is an integer from 0 to 3.

4. The process of Claim 1, said solution having dissolved therein between about 10-5 and about 10-1 gram mol per liter of multi-quaternary phosphonium ions.

5. The process of Claim 4 wherein Y is -R''-COOM or , Z is C2-C4 alkylene, m is 1 and n is an integer from 0 to 3.

6. The process of Claim 5 wherein -R'' is , Z is ethylene and n is 0, 1 or 2.

7. The process of Claim 6 wherein the nitrilocarboxylic acid compound is selected from the group consisting of ethylenediaminetetraacetic acid, N-hydroxyethylethylene-diaminetriacetic acid, diethylenetriaminepentaacetic acid, nitrilotriacetic acid, N,N-di(2-hydroxyethyl)glycine and the alkali metal and ammonium salts of such acids.

8. The process of Claim 1 carried out in an undivided cell having a heavy metal anode in contact with said solution.

9. The process of Claim 1, wherein Y is -R''-COOM or , Z is C2-C4 alkylene, m is 1 and n is an integer from 0 to 3, said solution having dissolved therein between about 10-5 and about 10-1 gram mol per liter of quaternary phosphonium ions.

10. The process of Claim 9 wherein the solution contains an alkali metal salt selected from the group consisting of borate in a concentration corresponding to at least about 0.01 gram atom of boron per liter of solution and condensed phosphate in an amount molecularly equivalent to at least about 0.01%
by weight of the corresponding condensed phosphoric acid.

11. The process of Claim 9 wherein -R''- is , Z is ethylene and n is 0, 1 or 2, said solution having dissolved therein at least about 1% by weight of alkali metal phosphate, borate, perchlorate, carbonate or sulfate.

12. The process of Claim 9 wherein the conductive salt is an alkali metal phosphate, borate or carbonate and the anode comprises a ferrous metal.

13. The process of Claim 12 wherein the nitrilocarboxylic acid compound is selected from the group consisting of ethylenediaminetetraacetic acid, N-hydroxyethylethylene-diaminetriacetic acid, diethylenetriaminepentaacetic acid, nitrilotriacetic acid, N,N-di(2-hydroxyethyl) glycine and the alkali metal and ammonium salts of such acids.

14. The process of Claim 12 wherein the solution contains an alkali metal salt selected from the group consisting of borate in a concentration corresponding to at least about 0.01 gram atom of boron per liter of solution and condensed phosphate in an amount molecularly equivalent to at least about 0.01% by weight of the corresponding condensed phosphoric acid.

15. In a process for hydrodimerizing an olefinic compound having the formula H2C-CR-X wherein -X is -CN or-COOR', R is hydrogen or R' and R' is methyl or ethyl by electrolyzing an aqueous solution having dissolved therein at least about 0.5%
by weight of said olefinic compound, and at least about 1% by weight of sodium or potassium phosphate, borate, carbonate or sulfate in contact with a cathodic surface consisting essenti-ally of cadmium or lead with a current density of at least about 0.01 amp/cm2 of cathodic surface, said solution having a pH
between about 5 and about 11 and a temperature between about 5° and about 75°C., said solution further having between about 0.1 and about 50 millimols per liter of nitrilo-carboxylic acid compound having the formula wherein Y is hydrogen, , or C1-C20 alkyl;
-R'' is or ; R''' is hydroxy, -COOM, or C1-C8 alkyl, hydroxyalkyl or hydroxyphenyl; Z is a divalent C2-C6 hydrocarbon radical; M is hydrogen, alkali metal or ammonium;
m is 1 or 2; n is an integer from 0 to 4 and at least one Y
is or , the improvement wherein said solution contains between about 10-5 and about 0.5 gram mol per liter of multi-quaternary phosphonium ions.

16. The process of Claim 15, said solution containing between about 10-5 and about 10-1 gram mol per liter of quaternary phosphonium ions selected from the group consisting of C17-C36 polymethylenebis(trialkylphosphonium) ions in which each trialkylphosphonium radical contains at least two C3-C6 alkyl groups and the polymethylene radical is C3-C8.

17. The process of Claim 16 wherein the nitrilocarboxylic acid compound has the formula where-in Y is -CH2 COOM or -CH2CH2OH; Z is C2-C4 alkylene; M is hydrogen, alkali metal or ammonium; and n is an integer of 0 to 2.

18. The process of Claim 17 wherein the nitrilocarboxylic acid compound is selected from the group consisting of ethylene-diaminetetraacetic acid, N-hydroxyethylethylenediaminetriacetic acid, diethylenetriaminepentaacetic acid, nitrilotriacetic acid, N,N-di(2-hydroxyethyl)glycine and the alkali metal and am-monium salts of such acids.

19. The process of Claim 17 wherein the solution contains an alkali metal salt selected from the group consisting of borate in a concentration corresponding to between about 0.02 and about 0.9 gram atom of boron per liter of solution and con-densed phosphate in an amount molecularly equivalent to between about 0.02% and about 3% by weight of the corresponding condensed phosphoric acid.

20. The process of Claim 17 carried out in an undivided cell having a heavy metal anode in contact with said solution.

21. The process of Claim 20 wherein the nitrilocarboxylic acid compound is selected from the group consisting of ethylene-diaminetetraacetie acid, N-hydroxyethylethylenediaminetriacetic acid, diethylenetriaminepentaacetic acid, nitrilotriacetic acid, N,N-di(2-hydroxyethyl)glycine and the alkali metal and am-monium salts of such acids.

22. The process of Claim 16 wherein the salt is an alkali metal phosphate, borate or carbonate and the anode eomprises a ferrous metal.

23. The process of Claim 22, wherein Y is -R"-COOM or , R"' is hydroxyphenyl, Z is ethylene, m is 1 and n in an integer from 0 to 2.

24. The process of Claim 23, wherein the solution contains an alkali metal borate in a concentration corresponding to at least about 0.02 gram atom of boron per liter of solution.

25. In a process for hydrodimerizing acrylonitrile by electrolyzing an aqueous solution having dissolved therein at least about 0.5% but less than about 5% by weight of acrylo-nitrile, and at least about 1% by weight of sodium or potassium salt selected from phosphate and borate in an undivided cell having a ferrous metal anode with a current density of at least about 0.1 amp/cm2, said solution having a pH between about 7 and about 11 and a temperature between about 40° and about 65°C., and said solution having between about 0.5 and about 25 milli-mols per liter of a nitrilocarboxylic acid compound selected from the group consisting of ethylenediaminetetraacetic acid, N-hydroxyethylethylenediaminetriacetic acid and the alkali metal salts of such acids, the improvement wherein said solution includes between about 10-4 and about 10-2 gram mol per liter of quaternary phosphonium ions.

26. The process of Claim 25, said anode consisting essentially of carbon steel.

27. The process of Claim 25, wherein the solution con-tains at least about 2.5 millimols per liter of the nitrilo-carboxylic acid compound.

28. The process of Claim 27, wherein the solution con-tains sodium or potassium borates in a concentration corres-ponding to between about 0.02 and about 0.5 gram atom of boron per liter of solution.

29. The process of Claim 25, said solution having dis-solved therein between about 10-4 and about 10-2 gram mol per liter of C8-C24 tetraalkylphosphonium ions having at least three C2-C6 alkyl groups.

30. In a process for hydrodimerizing acrylonitrile by electrolyzing an electrolysis medium consisting essentially of an aqueous dispersion containing up to about 20% by weight of an undissolved organic phase, said aqueous dispersion having dissolved therein at least about 0.5% but less than about 5% by weight of acrylonitrile, and at least about 1% by weight of sodium or potassium salt selected from phosphate and borate in an undivided cell having a ferrous metal anode with a current density of at least about 0.1 amp/cm2, said dispersion having a pH between about 7 and about 11 and a temperature between about 40° and about 65°C, and said dispersion containing between about 0.5 and about 25 millimols per liter of a nitrilocarboxylic acid compound selected from the group consisting of ethylenedi-aminetetraacetic acid, N-hydroxyethylethylenediaminetriacetic acid and the alkali metal salts of such acids, the improvement wherein said dispersion includes between about 10-4 and about 10-5 gram mol per liter of quaternary phosphonium ions.