CA2034435A1 - Method for conducting organic reactions using hexaalkylguanidinium salt as phase transfer catalyst - Google Patents

Abstract

METHOD FOR CONDUCTING ORGANIC REACTIONS
USING HEXAALKYLGUANIDINIUM SALT AS PHASE
TRANSFER CATALYST
Abstract Reactions between a solid polar and a non-polar compound, especially nucleophilic aromatic substitution reactions between an alkali metal salt of a hydroxyaromatic compound or thio analog thereof and an activated halo- or nitro-substituted aromatic compound, are conducted in a non-polar organic solvent such as toluene or xylene, in the presence of a hexaalkylguanidinium salt as a phase transfer catalyst. The method is particularly useful for the preparation of bisimides from bisphenol A or 4,4'-biphenol salts and 4-nitrophthalimides.

Description

- 1 - 2~3~3~
RD-l 97 6 IlSING HExAp~LKyIlG~ IDl~Iu~ $~T T ~S P~ASE

This invention relates to the preparation of organic compounds by the reaction of polar with non-polar compounds, and more particularly to a method of preparation thereof which employs improved phase transfer catalysts.
Various methods are known for conducting reactions between highly polar reagents, such as alkali metal salts of hydroxyaromatic compounds or thio analogs thereof, and substantially non-polar reagents such as activated halo- or nitro-substituted aromatic compounds. Typical nucleophilic aromatic substitution reactions of this type result ln replacement of the halo or nitro group with an aryloxy or 15` arylthio group.
Such nucleophilic aromatic substitution reactions are particularly useful commercially for the preparation of aromatic ether bisimides such as those of 2,2-bis[~-(dicarboxyphenoxy)phenyl]propane bisimides and 4,4'-bis(dicarboxyphenoxy)biphenyl bisimides. These bisimides maybe converted to dianhydrides, which in turn undergo reaction with diamines to produce polyetherimides. Certain bisimides also react directly with diamines to produce polyetherimides, as disclosed, for example, in U.S. Patent ~,578,470. The analogous monoimides are similarly useful as endcapping or chain-stopping agents for polyimides.
In most cases, it was formerly necessary to conduct reactions of this type (including nucleophilic displacement reactions) in polar aprotic solvents, since the alkali metal salts are typically insoluble in non-polar solvents.
Commercial preparation of a~omatic ethers was therefore inhibited by various disadvantages of polar aprotic solvents, including high cost, difficulty of recycling and toxicity.

~ .

.

- 2 - 2~
RD-197~8 More recently, it has been posslble to conduct the reaction in non-polar solvents with the employment of a phase transfer catalyst, facilitating incorporation of the salt of the hydroxyaromatic compound in the organic phase. Many S types of phase transfer catalysts are known, including quaternary ammonium and phosphonium salts as disclosed in U.S. Patent 4,273,712. More specifically, there have been used various bis-quaternary ammonium or phosphonium salts as disclosed in U.S. Patent 4,554,357, and aminopyridinium salts as disclosed in U.S. Patents 4,460,778, 4,513,141 and 4,681,949.
Despite the improvements affoxded by the use of phase transfer catalysts as described in the above-identified patents, several problems remain. In the first place, the reaction is often ~uite slow when those catalysts are employed. In the second place, decomposition of the phase transfer catalyst usually occurs during the reaction, necessitating frequent replacement thereof and resulting in the formation of by-products which cause discoloration of the produc~ and may lead to undesirable side reactions.
The results of catalyst decomposition are particularly noticeable in the preparation of 4,4'-bis(dicarboxyphenoxy)biphenyl bisimides by the reaction of nitrophthalimides with ~,4'-biphenol salts, using a bis(trialkyl~alkylenediammonium halide as catalyst. When the reaction is conducted in refluxing toluene as solvent/ high yields are obtained. However, in xylene ~which has a higher boiling point) the yield is much lower if the catalyst is exposed to re~lux temperatures prior to initiation of the reaction, in the course of drying the nitrophthalimide.
Yields increase somewhat if the catalyst is not introduced until after drying, but are still lower than desired.
The present invention is base~ on the discovery that certain hexaalkylguanidinium salts may be employed as , ,. : ,: - ; , .,, , .
. ~ . :, .

. :
'i :

- 3 - ~ 3~

phase transfer catalysts in reactions between polar and non-polar compounds. The use of these salts frequently increases the reaction rate substantially as compared with the use of previously known phase ~ransfer catalysts in comparable amounts. In addition, the hexaalkylguanidinium salts have a high degree of thermal stability and thus do not undergo substantial decomposition during the displacement reaction.
This means less color ~ormation in the product and the potential for recycling of catalyst, decreasing the cost of the process.
Accordingly, the invention is a method for effecting reaction in a non-polar organic solvent between a highly polar compound which is insoluble in said solvent and a substantially non-polar compound which is soluble therein, which comprises conducting said reaction in the presence of at least one hexaalkylguanidinium salt as a phase transfer catalyst.
The present invention is capable of use in connection with an extremely broad spectrum of reactions between organic chemicals. In general, it may advantageously be employed in any situation where reaction is to be effected be~ween one reagent which is highly polar and insoluble in the non-polar liquid to be used as solvent, and another which is substantially non-polar and is soluble therein. More particularly, it is applicable to reagents employed in nucleophilic aromatic substitution reactions, and still more particularly to the reaction between at least one alkali metal salt of a hydroxyaromatic compound or thio analog thereof and at least one activated halo- or nitro-substituted arornatic compound. For the sake of convenience~ these reagents will be the principal ones hereinafter and they will be specifically identified as "phenol salt" and "activated aroma~ic compound", respectively.

' .

.

- 4 - Z~3~3~

The phenol salts are generally compounds o~ the formula (I) Rl(ZM) a t wherein Rl is an aromatic radical containing about 6-30 carbon atoms, M is an alkali metal, Z is oxygen or sulfur and a is l or 2. The Rl radical may be a hydrocarbon radical or may contain other atoms such as oxygen or sulfur.
Illustrative monovalent radicals (i.e., those derived from compounds in which a is l) include phenyl, m-tolyl, p-tolyl, l-naphthyl, 2-naphthyl, p~chlorophenyl and 4-bromo-l-naphthyl.
Most often, Rl is a divalent aromatic radical;
i.e., a is 2. Illustrative dihydroxyaromatic compounds are resorcinol, hydroquinone, 4,4'-dihydroxybiphenyl, 4,4'-dihydroxy-3,3',5,5'-tetramethylbisphenyl, bis(4-hydroxyphenyl)methane, 3 hydroxyphenyl-4-hydroxyphenylmethane, 2,2-bis(2-hydroxyphenyl)propane, 2,2-bis(4-hydroxyphenyl)propane (hereinafter "bisphenol A"), 2-(3-hydroxyphenyl)-2-(4-hydroxyphenyl)propane, l,l-bis(4-hydroxyphenyl)ethane, 2,2-bis(4-hydroxyphenyl)pentane, l,l-bis(4-hydroxyphenyl)ethylene, 4,4'-dihydroxybenzophenone, bis(4-hydroxyphenyl)ether, bis(4-hydroxyphenyl) sul~ide, ~5 bis(4-hydroxyphenyl) sulfoxide, bis(9-hydroxyphenyl) sulfone and 3-hydroxyphenyl-4-hydroxyphenyl sulfone.
The preferred Rl radicals are usually ~2 (II) :

.

~ ' : J ; ~ ' , .

'', `

- 5 - ~ ~3~3~

(III) ~ and R2 Xl Xl R2 xl xl ~IV) ~ R -xl X

wherein each R2 ls independently hydrogen or methyl, R3 is a straight-chain or branched alkylene radical containing 1-5 carbon atoms and each xl is independently hydrogen or halogen (usually chlorine or bromine). Mixtures of the foregoing formulas are also contemplated. Especially desirable are the bisphenol A salts, having formula IV in which R3 is isopropylidene and each xl is hydrogen.
The alkali metal in the phenol salt may be any of the known alkali metals. Sodium and potassium are usually preferred by reason of availability and low cost, with sodium being especially preferred. The Z value may be oxygen or sulfur and is usually oxygen.
By "activated aromatic compound" is meant a compound having an electron-deficient aromatic ring, generally achieved by the presence of one or more electron-withdrawing substituents. Illustrative substituents of thistype are halo, nitro, acyl, cyano, carboxy, carbalkoxy, aldehydo, sulfone and perfluoroalkyl, as well as heterocyclic aromatic substituents such as pyridyl.
Most often, the activated aromatic compound is a substituked imide having the formula , .

~44L3~i o Il 4 / \

\C/
Il . .
o wherein A is an aromatic radical, R4 is hydrogen or an unsubstituted or substituted hydrocarbon radical containing about 1-13 carbon atoms and x2 is halo or nitro. The A
radical generally contains about 6 30 carbon atoms. The imide is generally derived from an o-dicarboxylic acid such as phthalic acid or 2,3-naphthalenedicarboxylic acid;
however, derivatives of acids such as 1,8-naphthalenedicarboxylic acid are also suitable. Mostpreferably, the imide is a phthalimide.
The R4 value is preferably an alkyl and especially a lower alkyl radical (i.e., one containing up to 7 carbon atGms). Most preferably, R4 is methyl or n-butyl.
According to the invention, the reaction between the phenol salt and the substituted aromatic compound is conducted in a non-polar organic solvent. Suitable solvents ; include benzene, toluene, xylene, chlorobenzene, o-dichlorobenzene, chlorotoluene7 dichlorotoluene and octane.
Aromatic solvents are preferred, with aromatic hydrocarbon solvents and ~specially toluene being particularly preferred.
An essential aspect of the invention is the emplo~nent of at least one hexaalkylguanidinium salt as a phase transfer catalyst. Suitable salts are represented by the forrnula . ~ -- 7 - ~ ~ ~4~

(VI) \ N~ C ~ ~ RB X
R6 ~
N\
Rl wherein each of R5, R6, R7, R8, R9 and R10 is a primary alkyl radical and X is an anion. -The alkyl radicals suitable as R5-10 are primary alkyl radicals, generally containing about 1-12 and preferably about 2-6 carbon atoms. The X value may be any anion and is preferably an anion of a strong acid; examples are chloride, bromide and methosulfate. Chloride and bromide ions are usually preferred.
As indicated by the dotted bonds in formula VI, the positive charge in the hexaalkylguanidinium salt is delocalized over one carbon and three nitrogen atoms. This is believed to contribute to the salt's stability under the conditions encountered according to the invention, including relatively high temperatures. As a result, decomposition of the hexaalkylguanidinium salt does not occur or occurs only to a very minor extent. The resul~s include suppression of ~- by-product formation and potential for continued use via 2Q recycle.
Hexaalkylguanidinium salts may be prepared by the reaction of a tetraalkylurea with phosgene or phosphorus oxychloride, or by the reaction of thiourea with an N,N-dialkylcarbamoyl halide, to yield a chloroformamidinIum salt, frequently r~ferred to as a "Vilsmeier salt", followed by reaction of said salt with a dialkylamine. Reference is made to Kan~lehner et al., ~iebi~ ~n. Chem_, 12~, 108-126, and ~-.

:

2~3~S

Pruszynski, an. J. ~he~n., ~, 626-629 (1987), which are incorporated by reference herein.
The preparation of hexaalkylguanidinium salts is illustrated by the following examples.

~m~

A 3-liter, 5-necked flask fitted with a mechanical stirrer, condenser, phosgene inlet tube, pEI meter and 10 addition port was charged with 182.85 grams (2.5 moles) of diethylamine, one liter of methylene chloride and 200 ml. of water. Phosgene (99 grams, 1 mole) was passed into the mixture under nitrogen, with stirring, at the rate of 3 grams per minute, with addition of aqueous sodium hydroxide 15 solution to malntain the pH in the range of 10-12. A
vigorous exothermic reaction took place during phosgene addition, causing refluxing of the methylene chloride. After phosgene addition was complete, the mixture was maintained at a pH of 12 while refluxing was continued for 2 hours. The 20 methylene chloride phase was separated, washed with water and vacuum stripped to yield the desired crude tetraethylurea in quantitative yield based on phosgene.
To a solution of 172.3 grams (1 mole) of tetraethylurea in 100 ml. of dry toluene was added under 25 nitrogen, with stirring, 170 grams (1.05 moles) of phosphorus oxychloride. The mixture was stirred and warmed at 60 C for 2 hours under nitrogen, whereupon the Vilsmeier salt separated as a second phase. Periodic analysis by nuclear magnetic resonance indicated when the reaction was complete.
30 At that point, the mixture was cooled to O C and diluted with 500 ml. of dry methylene chloride. There was then added, under nitrogen, 182 grams (2.5 moles) of diethylamine, with stirring at 0C. An exothermic reaction took place, and when it was complete the mixture was warmed to room temperature ~.' :

2~13a~s and analyzed by proton nuclear magnetic resonance.
Additional diethylamine was added until no further Vilsmeier salt was present in the mixture, after which 400 ml. of 35%
aqueous sodium hydroxide was added carefully and the mixture was extracted with methylene chloride. The organic phase was washed with saturated sodium chloride solution, dried and evaporated to afford the crude product as a yellow oil which crystallized upon addition of ethyl acetate. Upon filtration of the ethyl acetate slurry, the desired hexaethylguanidinium chloride was obtained in 87% yield. It could be recrystallized from a mixture of equal volumes of heptane and ethyl acetate, with enough chloroform added to effect solution when hot.

Exame~

Hexaethylguanidinium chloride, obtained according to Example 1, was dissolved in methylene chloride and ~he solution was washed three times with saturated aqueous sodium bromide solution. Upon workup as described in Example 1, the desired hexaethylguanidinium bromide was obtained; it had a melting point of 174-175-C.

~.m The procedure of Examples 1-2 was repeated, substituting tetra-n-butylurea and di-n-butylamine for the tetraethylurea and diethylamine, respectively. The product was the desired hexa-n-butylguanidinium bromide.
In the method of this inventiont the reaction mixture containing the ~henol salt, activated aromatic compound, hexaalkylguanidinium salt and solvent is normally heated at a temperature in the range of about 100-200-C, preferably about 125-175 C. It is ~referred to use s stoichiometric amounts of the phenol salt and activated aromatic compound, but under appropriate conditions an excess of one reagent or the other (especially the phenol salt), generally not more than about 25~, may be employed. An internal standard may be incorporated in the reaction mixture for analytical purposes. The proportion of hexaalkylguanidinium salt is a catalytically effective proportion, most often about O.S-5.0 mole percent based on activated aromatic compound. In the case of a compound containing both halo and nitro substituents, the halo substituent is normally displaced.
When isolation of the product is required, it may be achieved by conventional methods. These typically involve washing with an aqueous alkaline solution followed by drying of the organic phase and solvent stxipping.
For the preparation of 2,2-bis[4-(dicarboxyphenoxy)phenyl]propane bisimides, the preferred hexaalkylguanidinium salts are those in which the alkyl groups contain up to 3 carbon atoms, with the hexaethyl compound being most preferred. On the other hand, highest yields of 4,4'-bis(4-dicarboxyphenoxy)biphenols are generally obtained by the use of salts in which the alkyl groups contain 4-6 carbon atoms, especially the hexabutyl compound.
The reasons for this phenomenon are not presently known.
The method of this invention is illustrated by the following examples. All percentages are by weight.
"Chromatographic yield" is yield as determined by high pressure liquid chromatography.

E~=m~l~Y_i~lL

Sodium p-cresoxide was prepared by the reaction of p-cresol with sodium hydroxide in aqueous solution, followed by addition of toluene and removal of water by azeotropic - ; ' '~ ' ;" '` '~ ' , '' : ~ .
~: ~
: ' ' " `

435i distillation. A mixture of 780 mg. (6 mmol.) of sodium p-cresoxide and 800 mg. (5 mmol.) of p-chloronitrobenzene was prepared in an anhydrous nitrogen atmosphere, and there were added 22 ml. of solvent and a small amount of tetracosane as an internal standard. The solvent included at least 2 ml. of toluene, which was removed by distillation in a nitrogen atmosphere, with stirring, to effect azeotropic drying of the reactants. Hexaalkylguanidinium salt was then added in the appropriate amount, as a stock solution in o-dichlorobenzene, and heating and stirrin~ were continued with the reaction mixture being periodically removedr quenched with acetic acid, diluted with methylene chloride and analyzed by v~por phase chromatography.
The reaction parameters and yields of 4-cresyl-4'-nitrophenyl ether obtained, are listed in Table I. Solvents ~ are identified as follows:
Tol toluene;
PhCl - chlorobenzene;
DCB - o dichlorobenzene.
T~sL~_I
Example _ 4 5 6 7 8 9_ 1~
Solvcnt PhCl DCB DCB DCB PhCl Tol PhCl Hexaalkylguanidinium salt (formula VI):
R5-8 C2H5 C2Hs C2Hs C2~5 C4H9 C~3 CH3 R9-10 C2H5 C2Hs C2Hs C2H5 C4Hg CH13 CHl3 X Cl Cl Br Br 3r Br 3r-Mole ~* l.0 0.5 1.0 0.25 1.0 1.0 1.0 Solvent 35 Temperature, C 130 150 150 l~0 130 110 130 Chromato~raphic yield, %:
15 mln. --- --~ 97 ~~~ ~~~~~ -~~
30 min. 99 --- --- 65 92--- 88 1 hr. ~ - 72 --- --- 100 2 hrs. --- 98 --- --- --- 69 ---4 hrs. --- --- --- 87 --- --- ---20 hrs. --- --- ~~~ ~~~ ~~~99 ~~~
*Based on 4-chloronitroben~ene.

. .

- 12 - ~V3~

~Bm~ç~

Following substantially the procedure of Examples 1-7, various reactions between 4-substituted chlorobenzenes and sodium p-cresylate or p-thiocresylate, using hexaethylguanidinium bromide as the phase transfer catalyst.
The relative parameters and yields are given in Table II.
~
Example _ . ~ ~ 11 12 13 14 15 1~_ 17 15Alkali metal salt:
Z (formula I) S O O S S S S
Activated aromatic compound:
Substituent NO2 CN SO2CHs CN SO2CHs COC6Hs CHO
Catalyit mole % 1 5 5 Temperature, C130 180 130 180 130 130 130 Time, min. lS 480 60 120 120 180 120 Solvent PhCl DCB PhCl DCB PhCl PhCl PhCl Chromatographic94 72 97 96 95 88 a7 yield, %
E~am~L~s~

In catalyst stability tests, 308 mg. ~1 ~mol.) of hexaethylguanidinium bromide was heated under reflux, with stirring, with 260 mg. (2 mmol.) of sodium p-cresoxide in refluxing chlorobenzene (Example 18) and refluxing o-dichlorobenzene (Example 19) for 2 hours, in a nitrogen atmosphere. The mixtures were cooled and extracted with 2 aqueous sodium hydroxide, and the aqueous extracts were washed with petroleum ether and extracted with methylene chlorideO Any remaining p-cresol salt was removed by washing with 10% a~ueous potassium hydroxide solution, and the methylene chloride solutions were dried over magnesium sulfate and vacuum stripped. ~he amount of - 13 - X~3~43~

hexaethylguanidinium bromide recovered was in the range of 89-88%.

~am~

A 50-ml. round-bottomed two-necked ~lask fitted with a reflux condenser, a magnetic stir bar and a nitrogen inlet was charged with 3 grams (11 mmol.) of the disodium salt of bisphenol A, 4.54 grams (22 mmol.) of 9-nitro-N-methylphthalimide, O.S gram of 1,3,5-triphenylben~ene as an internal standard, 685 mg. (0.26 mmol.) of hexaethylguanidinium chloride and 17.1 grams of toluene. The flask was purged with nitrogen and heated under reflux for 45 minutes; samples were periodically taken, quenched with a mixture of urea, acetonitrile, methanol and acetic acid and analyzed by high pressure liquid chromatography. At the end of the reaction period, the mixture was diluted with toluene to 22% solids and 6 ml. of 0.8% aqueous sodium hydroxide solution was added at 80 C. The mixture was stirred at this temperature for 15 minutes, after which the organic layer was removed and washed twice more with aqueous base. The toluene was vacuum ~tripped to yield the desired 2,2-bis[4-(3,4-dicarboxyphenoxy)phenyl]propane bis-N-methylimide in 94.7%
yield.
A solution of exactly 500 mg. of the bisimide in exactly 10 ml. of chromatographic grade methylene chloride was stirred until all the bisimide had dissolved. The yellowness index of the solution was then measured in a Gardner Instruments colorimeter, and was found to be 5Ø
A control reaction was conducted by an iden~ical procedure except that the hexaethylguanidinium chloride was replaced by an equivalent amount of bisttri-n-butyl)-1,6-hexylenediammonium dibromide (i.e., 0.13 mmol., since the equivalent weight of this compound is half its molecular ,.
~ ` " `; ~ ` :

, ' - 14 2~

weight). The product was isolated in 92% yield and had a yellowness index value of 22.2.

Exam~les 21=~2 The procedure of Example 20 was repeated, employing half the quantities of each reagent and replacing the hexaethylguanidinium chloride with an equivalent amount of hexaethylguanidinium bromide ~Example 21) and hexa~n-butylguanidinium bromide (Example 22). The isolated yieldsand yellowness indices of the products were as follows:
Example 21 - 91%; 5.6.
Example 22 - 88%i 8.3.

The reaction vessel was a 100-ml. round bottomed flask fitted with a mechanical stirrer and a Dean-Stark trap bearing a reflux condenser and a nitrogen inlet and linked with the reaction vessel by a tube filled with calcium hydride and plugged with cheesecloth below and glass wool above the calcium hydride. The vessel was charged with 8 grams ~32.2 mmol.) of 4-nitro-N-butylphthalimide, 495 mg.
25 (0.806 mmol.) of hexa-n-butylguanidinium bxomide, 400 mg. of biphenyl as an internal standard and 16.4 ml. of toluene.
The mixture was purged with nitrogen for 5 minutes; heated under reflux for 50 minutes in a nitrogen atmosphere, with stirring; and cooled, after which 3.71 grams (16.1 mmol.) of dry 4,4'-biphenol disodium salt was added. Heating under reflux in a nitrogen atmosphere was resumed and the reaction was monitored at 30-minute intervals by high pressure liquid chromatography after the aliquots being sampled had been diluted with chloroform, treated with dimethylacetamide and filtered.

'~' : ' . ,~ :,, , ." : ~'' ';' ~3~35 After 2.5 hours of refluxing, the mixture was cooled to 80 C and washed three times with 15 ml. of 0.8%
aqueous sodium hydroxide which had been preheated to 85 C.
In each washing step, the mixture was stirred for 15 minutes before removal of the aqueous layer. The toluene was removed by vacuum stripping, to yield the desired 4,9'-bis(3,4-dicarboxyphenoxy)biphenyl bis-N-(n-butyl)imide. The chromatographic and isolated yields were 98.7% and 97.7%, respectively.
In a control similar to that described for Example 20, the chromatographic and isolated yields were 96~ and 92%, respectively.

The procedure of Example 23 was repeated, substituting xylene for the toluene and thus providing a substantially higher reaction temperature. The chromatographic and isolated yields were 99% and 96.3%, respectively.
In the control, the chromatographic yield was only 55.1% and the isolated yield was 57.5%. When the procedure was modi~ied by adding the catalyst with the biphenol disodium salt, rather than with the 4-nitro-N-butylphthalimide~ the chromatographic yield was 86.3% and theisolated yield was ~6.0%.

Claims (20)

1. A method for effecting reaction in a non-polar organic solvent between a highly polar compound which is insoluble in said solvent and a substantially non-polar compound which is soluble therein, which comprises conducting said reaction in the presence of at least one hexaalkylguanidinium salt as a phase transfer catalyst.

2. A method according to claim 1 wherein the highly polar compound is an alkali metal salt of a hydroxyaromatic compound or thio analog thereof, and the substantially non-polar compound is an activated halo- or nitro-substituted aromatic compound.

3. A method according to claim 2 wherein the alkali metal salt has the formula (I) R1(ZM)a , wherein R1 is an aromatic radical containing about 6-30 carbon atoms, M is sodium or potassium, Z is oxygen or sulfur and a is 1 or 2.

4. A method according to claim 3 wherein M is sodium and a is 2.

5. A method according to claim 4 wherein the solvent is toluene or xylene.

6. A method according to claim 5 wherein the activated aromatic compound is a substituted imide having the formula (V) , wherein A is an aromatic radical, R4 is hydrogen or an unsubstituted or substituted hydrocarbon radical containing about 1-13 carbon atoms and X2 is halo or nitro.

7. A method according to claim 6 wherein the substituted imide is a 4-nitrophthalimide.

8. A method according to claim 7 wherein the alkali metal salt is the disodium salt of bisphenol A or of 4,4'-biphenol.

9. A method according to claim 8 wherein the reaction is conducted at a temperature in the range of about 100-200°C.

10. A method according to claim 9 wherein stoichiometric amounts of the alkali metal salt and 4-nitrophthalimide are employed.

11. A method according to claim 8 wherein the proportion of hexaalkylguanidinium salt is about 0.5-5.0 mole percent based on 4-nitrophthalimide.

12. A method according to claim 8 wherein the alkali metal salt is a bisphenol A salt.

13. A method according to claim 12 wherein R4 is methyl.

14. A method according to claim 13 wherein the hexaalkylguanidinium salt is hexaethylguanidinium chloride.

15. A method according to claim 13 wherein the hexaalkylguanidinium salt is hexaethylguanidinium bromide.

16. A method according to claim 8 wherein the alkali metal salt is a 4,4'-biphenol salt.

17. A method according to claim 16 wherein R4 is n-butyl.

18. A method according to claim 17 wherein the hexaalkylguanidinium salt is hexa-n-butylguanidinium chloride.

19. A method according to claim 17 wherein the hexaalkylguanidinium salt is hexa-n-butylguanidinium bromide.

20. The invention as defined in any of the preceding claims including any further features of novelty disclosed.