[1]United States Patient Office 2,971,987 2,971,987 PROCESS FOR PRODUCING CRYSTALLINE TRL%IETRYLOLPHENOL Cal Y. Meyers, Princeton, NJ., assignor to Union Carbide Corporation, a corporation of New York No Drawing. Filed July 29, 1957, Ser. No. 674,627 13 Claims. (Cl. 260-621) T-his invention relates to a procm for producin- 2, 4,6trimethyloylphenol. More particularly, this invention relates to a process for producing crystalline trimethylolphenol from an alkali or alkaline earth metal salt of trimethylolphenol under carefuhy controlled -conditions. As used herein, the term "trimethylolphenol" signifies f-he compound 2,4,6- triinethylolphenol, whic-h compound may be represented by the graphic formula OH I HOH2C- CH20H CH20H T,rimethylolphenol is a useful water-soluble, resinforming compound -having a melting point of about 84-86' C., which is suitable for use in many applications, particularly as a component of phenolic resin forming compositions. It may be readily homopolymetized or. reacted with polyhydroxy cgm-pounds which systems are generally water-soluble or at least water dispersible so as to be readily appli-ed from aqueous solutions. After application, the resins can be cured with heat, or wit@h acidic cure accelerators, or with both to a hard, infusible, water-insoluble thermoset resin. One of the principal difficulties heretofore of obta'Ming trimethylolphenol in pure form has been its extreme ease of self-condensing and polymerization. Under even mildly acidic or alkaline conditions the trimethylolphenol will self-condense and otherwise be. unusable. -In addition, its great @@ity for water has made recovery of the crystalline product extremely difficult. Heretofore preparation of the pure trimethylolphenol has been accompeshed only throu.-h extended and expensive methods of preparation. Carpenter and Hunter, J. Appl. Chem., 1, 217 (1951), report that they were able to secure trimethylolphenol by the @reduction of acetoxytrimesic acid triethyl ester with litbium aluminum hydride. Martin, J. Am. Chem. Soc., 74, 3024 (1952), was able to secure a crystalline product believed to be nearly puretrimethylolphenol (melting point 84' C.) starting with phenol and formaldehyde. A mixture of polymethylolphenols formed in the reaction of the phenol and formaldehyde was converted to the trimethylsilyl derivatives with trimethylchlorosilane, the.respective sflyl derivatives were separated by fractionation and each fraction then hydrolyzed. From one fraction Martin was able to isolate crystalline trimethylolphenol in an unspecified yield having a melting point of 84' C. after a recrystallization from ethyl acetate. Tbis material is believed to be nearly pure trimethylolphenol. Freeman, J. Am. Chem. Soc., 74, 6257 (1952), was likewise able to secure trimethylolphenol by the neutralization of sodium trimethylolphenate with acetic acid in a dilute suspension in acetone. in this process, Frec@man reported a yield -of trimethylolPatented Feb. 14, 1961 2 phenol of 71 percent but of an obviously impure product (M.P. 74-75' C.), which he found impossible to further purify. -In addition, Freeman notes that in his process the trimethylolpbenol was not always recover.5 able in crystalline form, but was quite often secured only as a viscous oily product. The oily product secured by Freeman undoubtedly contains sow-e mono- and dimethylolphenols and dinuclear products which may be caused by the pres--nce of acetic acid and sodium acetate, 10 since Martin, J. Am. Chem. Soc., 73, 3954 (1951), showed that the sodium trimethylolphenate in the presence of acetic acid is not stable and wiH lose formaldehyde. An undesirable bufferin.- action also exists with the usD of acetic acid in this process which prevents the 15 recovery of nearly pure trimethylolphenol in high yields. Acetic acid and sodium acetate form a buffer so that the neutral end point is difficult to detect. It is obvious that aH of these processcs are either entirely too cum-bersome and expensive for prepaling 20 trimethylolphenol suitable for commercial use or tooinefficient in securin.- pure material to justify conimercial exploitation. It is therefore an object of the present invention to provide a process for the direct production of crystalline 25 triinethylolphenol in high yields and excellent purity in an inexpe-Tisive and simplified process. According to the present invention, I have no-,v found a process whereby ciystauine trimethylolphenol can be secured in high yields and purity from an alkali or alka30 line earth metal salt of trimethylolphenol. Basically this proci--ss includes the steps of forming a mixture of the alkali or alkaline earth metal salt of trimethylolphe-.iol . with an organic, solvent, and neutralizing the said mixture with an acidic material which forms a 35 neutral an,d substantially insoluble salt with the cation of t-he alkah OT alkaline earth metal salt of trimethylolphenol, . removiDg the said insoluble salt thus formed and recovermg the trimethylolphenol from the solvent. Careful control over the reaction and reaction condi40 tions must be maintained in order to successfury operate this process to secure the high yield and purity of trirrethylolphenol. Ibis process is appheable t(> the use of any of the alkaline metal salts of 2,4,6- trimethylolphenol, i.e., the 45 alkali or alkaline ear-th metal salts of trimethylolphenol. I particularly prefer the use of the calelum trimethylolphenate for reasons hereinafter discussed, but other salts such as the sodium trimethylolphenate and the barium triinethylolphenate can also give highly desirable results 50 in this process. These can be prepared in a manner as described in U.S. Patent 2,579,329 issued December 18, 1951, to R. W. Martin. It is highly desirable that the alkaene metal 9alt of trimethylolphenol employed in this invention be substantially free of mono- and di-methylol 55 products and from di- and polynuclear phenols. The absence of such compounds makes it possible to obtain crystals of substantially pure trimethylolphenol melting at 94-86' C. In the first step of this process, the alkali or alkaline 60 earth metal salt of trimethylolphenol is admixed with an organic solvent inert to both reactants and products and in which the trimethylolphenol is substantially completely soluble. It is not critical, however, that the trimethylolphenol salt be soluble in the organic solvent for 65 either solutions of the salt or suspensions can be used as desired. I have found, however, that trimethylolphenol salt solutions -are preferable in that the salt is immediately available for reaction and that less time is consijmed in the reaction and a better purity product is @also achieved. 70 As a result it is preferred that the trimethylolphenol salt be at least partially soluble in the solvent selected. Preferr@d solvents are the lower alkyl alcohols sui;h as'
[2]3 methanol, ethanol, isopropanol, and th-- like, in which all the trimethylolphenol salts are soluble to some degree. Other solvents stich as the lower aliphatic ketones, esters, and other polar solvents can also be used if dtsired, althou,-'a Lie trimethylolphenol salts are Icss soluble in these tllan -in the alcohols. -It is highly desirable in this proce@s that the niixture of the trimethylolphenol salt and solvent be as hi.-hly concentrated with salt as possible. If the solvent is present in aniounts of greater than abolit five ml. per gr,am of salt, ithe recovery of trimethylolphenol become's troublesome and time consuming. I p;articularly prefer the use of solvent in amounts of between about 3/4 to 3 rnl. per -ram of salt, 4although even less can be used if there is a sufficient amount to facilitate the reaction and take up in solution all of the trimethylolphenol formed. Duriiig the neutralization of this mixture, usually to a pH of about 5 to 7, the temperature of the mixture preferably sho-,ild not be permitted to exceed about 65' C. At temperatures much above this, the trimethylolphenol and/ or the salts tend to enter into undesirable side reactions. I particularly prefer to keep the temperature between about 15' and 45' C., althou.-h lower temperatures can be used if desired. It is highly critical in the opera-tion of this process that the neutralization be carried out with an acid which forms a neutral and substantially insoluble salt with the cation of the trimethylolphenol salt. The term "neutral and substantially insoluble" salt as used herein denotes a salt which is substantially neutral and so little soluble in the selected neutralization rnedium as to precipitate substantially completely therefrom, or a salt which is so insoluble and so little ionized in said medium as to have no signlqcant effect on the pH of the overall reaction mixt,,ire. It is this feattire of insolubility and neutrality that makes it possible for the crystalline tr@methylolphenol to be recovered free of inorganic salts and any mono- and di-methylolphenol compo7,inds and of di- and polynuclear phenols suell as are forrned by acid or base catalyzed side reactions. If the neutralizing acidic material forms a substantially acidic or alkaline by-product salt havirig an appreciable solubility in the neutralization medium, such a salt will promote undesirable side reactions a.-,id polymeric material formation a-id pure trimethylolphenol cannot be reco@,,ered. Also, it is aifficult to remove the dissolved portion of said salt from the trimethylolphenol atid the prodiiet vill be contaminated and impure. The acidic Peutralizing agents must therefore be judiciously selected so as to form a salt which is neutral and substantially insoluble in the organic solve-@it selected. It is for this reason, among others, that I prefer the use of the calcium trimethylolphenate. This salt is readily prepared in the anhydrous crystalline form and can be neutralized with acids such as carbonic acid, sulfuric acid, (>xalic,acid, and the like, the calcium salts of which are neutral and substantially insoluble in most or.-anic Sol_ ve,its. Likewise when water is present, acid anhvdrides such as carbon dioxide, sulflr trioxide, and the lfke can be used. The barium trimethyjolphenate acts in about the same manner apd ih-, same acids can be employed for theneutralization step. Sodium trimethylolphenate presents a somewhat different problem because of the high degree of solubility of most sodium salts in even trace amounts of water and becatise so many of these by-product salts are not neutral. In addition, sodium trimethylolphenate itself generally is isolated as the monohydrated crystalline forrn, yielding one mole oj' water per mole of sodium trimethylolphenate, this waterbeingdifficidt to remove. Thus when employing this salt, the most desi.-able fluid acids whi:ch form the necessary neutral and substantially insoluble salts are acids such as HCI and H2SO,, (and its anhydride, SO,). H<)wever other aci@ds formin-g neutral and substantially insoluble sodium salts can be used. VVhile the neutralization reaction need not necessarily @e conducted under anhydrous coiiditions, theabnnce of 2,971,987 4 water duri@ng the reaction facilitates high yields and good purities of the trimethylolphenol. I have found that up to about one mole of excess water per equivalent of trimelhylolphenate ion can be tolerated but exce-ss water mtich greater than this amount prevents crystallization of the trimethylolphenol and cannot be tolerated. The telm "excess water" denotes water over and above the arnount required by the stoi--hlometry of the neutralization reaction. For instance, if an acid such -as HCI is 10 used as the neutralizing agent, no water is required; and any water present in either the organic solvent or the trimethylol salt is "excess water" and can be tolerated only in such lirr.Lited amounts. Altematively, if the neutralizing agent is added in the fotm of an acid anhydride such 1,5 as C02, then an equimolar quantity of water is required to fol-m the corresponding acid; and in such cases, the water over and above said required quantity would be .,excess water." Trimethylolphenol appears to retain moisture tenaciously; and my experience has shown that 20 when the trimethylolphenol-forming reaction mixture contains a concentration of excess water greater than about I mole per mole trimethylolphenate ion, the reaction prodlict is an oily-appearing liquid which resists crystallization-even after it has been subjected to strong25 ly desiccative treatn,-ent. Therefore control over the water content ;of reactants and solvents should be maintained. With the calcium and barium trimethylolphenate salts, which exist in the anhydr(yus crystalline state, it is possi30 ble to use solvents containing up to one mole of water per trimethylolph@-nol equivalent, or to use a solvent containing even more water with an acid anhydride neutralizing agent which reacts with said additional water to form the corresponding acid. 35 However, again the sodium salt of trimethylolphenol presents difficulties in that it has been found to exist as the monohydrated crystals and thus one mole of water is present in the salt-solvent mixture. In this case it is necessar-y to use substantially anhydrous solvents and 40 acids. It is also frequently advantageous, in this regard, to conduct the neutralization with an acid anhydride so as to take up an equivalent quantity of water from the reaction mixture.These acid anhydrides, such as carbon dioxide and s,,ilfur trioxide, are desirable in many cases, 45 and particularly carbon dioxide when used to neutralize calcium or barium trimethylolphenate' -It not only serves to consume water and thereby to reduce the exccss water content, but also with this agent, i.e., C02, it is virtually impossible to overacidify the reaction mixture, 50 thereby avoiding any possibility of causing polymerization or side reactions. The calcium or barium carbonate is neutral and insoluble in mos,t solvents, including water. Thus, no problems are created by its presence even,when some water is present in the mixture being neutralized. 5 5 The neutralization reaction, while critical in the abovementioned features, can readily be. cbnducted by followin@@ the pH of the reaction mixture with an indic,ator such as bromphenol blue or meth-yl oran,@e or with a pH meter. It is also possible to pre-determine, the amount of 60 acid needed to reach the neutral end pbint by gravime'ric or volumetric means. With strongly acidic agents such as HCII S03 H2SO4, etc., it is desirable to add the soliible neutralizing acid or acid rnaterial slowly qo as to avoid localized overacidity. Generally, no such pre65 caution need be taken with the cation exchange resins or with weakly acidic soluble agents such as C02- Where tne neutral end point is overshot, it is possible in this process to reduce excess acidity with a suitable acid ac70 ceptor, such as K2CO3 Or the like, but more preferably it can be accomplisbed by addin.- a corresponding additional amount of alkal,i or alkaline earth metal salt of trimethylolphenol@ A particularly desirable embodirrent of this invention 75 resiot! ift the @ise of an insoluble, solid acid As the neu-
[3]5 tralizing agent for all of the trimethylolphenol salts. An insoluble, solid acid, i.e., a cation exchange resin, can be advantageously employed with an ' of the alkali and alkaline earth metal salts -to secure the trimethylolphenol in hi.-h yield and purity without dangere of acid catalyzed side reactions or polymerization. Par-ticularly desirable of the cation resins which can be used are the polysulfonated cationic exchange resins and the polycarboxylic cationic exchange resins, for instance the polystyrene acid resins, such as the polystyrene sulfonic acid and polystyrene carboxylic acid resins as the Rohm and Haas Co. "Amberlite IR-120" resin and the "AmberliteIRC50." However other cation exchange resins having shnilar ion exchangin.- properties can also be employed. Neutralization with such solid acids follows the same course as with the soluble acids, i.e., the by-products salt formed is neutral and @ubstantially insoluble. Thus a solution or suspension or the trimethylolphenate salt in the organic solvent is contacted with the ion exchange resin. The trimethylolphenol released is immediately picked up by the solvent and after removal of the exchan.-e resin, the trimethylolphenol is recovered from the solvent. A frequently convenient technique is to continuously pass a solvent solution of the trtimethylolphenol salt through a packed column thereby combining the neutralization and by-product salt separation Simultaneously. Very high yields and purity of the trimethylolphenol are secured by this method. It also has the distinct advantage of bein.- able to handle any salt of trimethylolphenol with relatively little danger of over acidifying the mixture or forming other than a neutral and substantially insoluble by-product salt. During neutralization of the triniethylolphenol salt, the trtimethylolphenol remains dissolved in the solvent and the neutral and substantially insoluble by-product salt precipitates from the solution. After the neutralization is completed, the precipitated by-product salt can then be removed from the solvent solution by Mtration, centrifugation, decantation, or other suitable means and the solvent solution concentrated by distilling off the solvent under reduced pressure and at temperatures preferbly not exceedin- about 50' C. Increased yields are secured if the salts after removal from the solvent solution are washed with portions of fresh solvent or with another suitable solvent for trtimethylolphenol but a nonsolvent for the salt. By taking the solvent-trimethylolphenol solution to dryness by solvent evaporation, the trimethylolphenol crystallizes. It can then be gro-und to a uniform size and stored under moisture-proof conditions. As an altemative method f6r recovery, the trimethyl,olphenol can be precipitated from the solvent solution by the addition of appropriate amounts of a solovent-miscible liquid in which the trimethylolphenol is not soluble. Some suitable organic compounds for causing the crystallization of trimethololphenol from solvent solutions in this manner include benzene, chloroform, methyl ethyl ketone, and CC14. If desired, further purification of the trimethylolphenol can be secured by recrystallizing the product from an anhydrous solvent such as ethyl acetate. Again, water in amounts greater than one mole per rnole of trirriethylolphenol in the recrystallization solvents must not be present durin.- the recrystallization step. In most cases, however, recrystallizing the trimethyloiphenol is not necessary as the product can be recovered having a sharp meltin- point at 84' to 86' C. Melting points in this ran.-e signify a pure or nearly pure product. Yields of 75 percent o,' theoretical or better are possible by this process, and quite often yields of better than 90 percent have been secured where care was exercised in the process inmaintaining nearly anhydrous conditions. The following examples are illustrative of this invention. 2,971,987 6 Exainple I A 3-Eter, 3-neck, round-bottomed flask equipped with an agitator, drying tube, thermometer and gas-inlet tube was charged with 206 g. of dried sodium trimethylolphenate, 150 ml. of anhydrous methanol (prepared by distillation from magnesium) -and 3 drqds of 0.5% bromphenol blue indicator and placed in an ice bath. The mixture was agitated vigorously until the temperature of the suspension was below 20' C. Anhydrous hyd.-ogen 10 chloride, dried by passage through concentrated sulfuric acid, was then adniitted as rapidly as possible without having the reaction mixture's temperature exceed 25' C. at any time. Within 1/2 hour the solution suddenly became yellow, indicating @an acid medium. The HCI flow was 15 discontinued and the reaction mixture was immediately neutralized with anhydrous potassium carbonate. The sodium and potassium chlorides were removed by filtration and the methanolic filtrate was concentrated in vacuo (pressure about 120 mm.) tat room temperature until it 20 attained a viscous, oil-like consistency. About 100 ml. of acetone which had been dried over anhy&ous calcium sulfate was added and the rather viscous, turbid mixture was refrigerated overnight. The large crop of fine white crystals which settled out was isolated. The mother liquor 25 was concentrated in vacuo, about 40 ml. of dry ethyl acetate added, the mixture refrigerated, and the crystalline material which settled out was isolated and combined with the first crop. There was so obtained 170 g. (92.4% of theory) of a fine white crystalline product which had a 30 melting point of 84-85' C. and a methylol (i.e., -CH20H) content of 49.6%. (Calculated for trimethylolphenol: 50.5%.) The methylol content is determined by reacting a sample with excess m-cresol containing 1-2% p-toluene35 sulfonic acid as catalyst in a sealed ar@ipoule for 16 hrs. at 105' C., then m6asuring, by Karl Fischer Titration, the moisture evolved during the reaction. wt. H20 evolved 31 40 Percent methylol= sariaple wt. x Tsxloo The method is, in essence, a modified version of that described by Martin, Anal. Chem., 23, 883 (1951). 45 Example 11 A 12-liter, 3-neeked, round-bottomed flask equipped as @in Example I was charged with 2200 g. (10 moles) of partially hydrous sodium trimethylolphenate containing 140 g. moisture as determined by Karl Fischer Titration, -50 and 4400 ml. of 99% isopropanol @and placed in an icesalt bath. (Total moisture content of suspension was therefore 184 g. or 1.02 moles water per mole trimethylolphenate.) When the vigorously agitated suspension had been cooled t<) 10-15' C., anhydrous HC1 (dried by 55 passage through cone. H2SO4) was bubbled in at such a rate that the reaction mass temperature did not exceed 25' C., until neutralization was completed, as signalled by the reaction mixture's changing color from tan to white. (The pH of the reaction mixture at this point was about 60 5.) 'Me reaction mixture was then ffltered, the sodium chloride cake washed with dry isopropanol -and said washings added to the filtrate which was then concentrated in vacuo (pressure about 5 mm.) at room temperature overnight. The solids that separated out were collected on a 65 Buchner funnel and the liquor expressed therefrom by means of a rubber membrane over the crystal pack. The stals then were vacuum desiccated at room temperature Lry over sodium hydroxide pellets and paraffin. The 1020 g. (55% of theory) of light grey solids so obtained had a 70 melting point of 75-85' C. The mother liquors upon I successive concentrations, fLItrations, etc.,as above, yielded an,additional 480 g. of product. The total mixture (1500 g. corresponding to 81% of theory) was thoroughly mixed and pulvprized. It had a melting point ran.-e of 65-84' 75 C.. Recrvstallization by dissolution in methanol, fo.Ilowed
[4]7 by acetone addition, vacuum. concentration, etc. as above yielded a fine, white crystalline product melting at 8485' C. Example III A mixture consisting of 188 g. (2 moles) of pbenol, 324 g. (4 moles) of 37% aqueous formahn and 90 g. (3 moles) of paraformaldehyde was cooled in an ice bath to 15' C. and 56 g. (I mole) of reagent grade lime, taken from a newly opened bottle, was added to the mixture with vigorous agitation at such a rate that the reaction mass temperature remained below 35' C. through out. After 2 hrs. of agitation, 200 ml. of water was added, the ice bath was remoxed and agitation was continuqd for an additiona,l 14 hrs. during which period the reaction tempera@ture remained at 25-30' C. without extemal cooling. The Teaction mixture was pou@red slowly @into 2.5 liters of a vigorously agitated 2:1 (by volume) acetone-isopropanol solution. The ivhite precipitate whicb formed was allowed to settle for Z hours, then collected on a Buchner funnel and th-- solvents completely expresed therefrom by means of a rubber dam. The filter cake was desiccated for several hour in a vacuum desiccator over sodium hydroxide pellets and paraffin at room temperature, then comminuted. There was so obtained 362 g. (89% yield) of a fine, a white powder which was readily soluble in water, less soluble in methanol and ethanol, slightly soluble in isopro-panol and insolube in acetone. It had a neutralization equivalent of 204 (avera.-e of 4 determinations) as detennined by titration with hydrochloric acid to a bromphenol blue endpoint. (Calculated for calcium trimethylolphenate: 203.) Example IV A I-liter, 3-neeked, round bottomed flask out-fitted with ,an agitator, dryin.- tube, thermometer, and gas inlet -tube was charged with 102 g. (0.25 mole) of calcium trimethylolphenate (the product of Example III), 4.5 g. (0.25 mole) of water @and 300 ml. of anhydrous methanol. C02 was then bubbled into the viborously agitated suspension for about 3 holirs. The pH was 7. Fur-ther addition Of C02 did not effect -any further change in pH. About 150 ml. of dry acetone was then added, with agitation, to the viscous mixture, and the voluminous precipitate of CaCO3 which settledon standin.- was removed by filtration and washed several times with dry acetone. The combined washings and filtrate was vacuum concentrated (at about 5 mri. pressure at room temperature) to thick yellow syrup which changed, on refrigeration, to crystalline mass. The entire mass was washed with cold (-5' C.) dry acetone and desiccated. The crystal,line product had a mell'ing point of 65-70' C. at this point. It Nvas re-dissolved in the minimum amount of dry methanol, the methanol solution diluted with an approximately equal volume of dry acetone and vacuum concentrated (at 5 mm.) at room temperature witil occasional additions of small por-tions of dry acetone as needed until al@l the methanol was removed. The t@-imethylolphenol was precipitated out of this acetone mixtlire by additig thereto a dry 1: I (by vol.) mixture Of CC]4-acetone, collected by filtration and dried. The wi'iite cr-,,stalline product so obtained melted sharply at 82-83' C. and was identified as t-rimetbylolphenol. Example V Seventeen -rams of a cation exchange resin (Amberlite IRC-50, Rehm and Haas Co.), which had previo-asly been thorouglily washed with methar-ol and dried, was added to a rnixture consisting cf 10 g. of dry calcium trirr .ethylolphenate hiving a neutralization equivalent of 204 and 40 m@l ' dry methanol. The mixture was agitated at 30' C. until it attained a pH of 6.5-7.0 (35-40 minutes) then filtered. The exchan-e resin was washed with two 10 ml. portions of dry met-hanol, the methanolic filtrates combined and concentrated in vacuo to a total volume of about 15 ml. A mixture consisting of 25 ml. 2,971,987 8 dry acetone and 25 ml. dry jcarbon tetrachloride was added, th,e resuitant slurry concentrated in vacuo to an overall volume of about 20 ml. and an additional 25 ml. of the 1:1 dry aceto'ne-dry CC14 n-iixture added. The mixture was refrigeraled overnight, then ffltered and the sol@ld's vacuum dried at room temppr@ture. There was so ob-tained 4 g. of a white, crystalli-.ie product which melted sharply at 80-81' C. and was.identified ascrystalline tr,'methylolphenol by @aiing a mixed melting point 10 with an authentic sample.
