PENTAERYTHRITOL DERIVATIVES I. MECHANISM OF FORMATION OF DIPENTAERYTHRITOL1 L. W. TREVOY AND M. E. MYERS Research Department, Canadian Chemical Company, Limited, Edmonton, Alberta Received October 30, 1962 ABSTRACT Dipentaerythritol, (CHZOH)~CCH~OCH~C(CH~OH)~, is a product of the base-catalyzed Tollens condensation of acetaldehyde with formaldehyde. The mechanism for formation of dipentaerythritol has not been hitherto established. Three types of experimental evidence support the conclusion that monopentaerythritol, C(CH*OH)d, is an intermediate in the sequence of reactions leading to dipentaerythritol: (I) the isolation of radioactive dipentaerythritol from a Tollens condensation in which monopentaerythritol-c14 was present, (2) comparison of the radioactivity levels of mono- and di-pentaerythritol prepared by reaction of formaldehyde-c14 with inactive monopentaerythritol and inactive acetaldehyde, and (3) the observed dependence of the yield of dipentaerythritol on the concentration of monopentaerythritol present in the condensation reaction mixture. Dipentaerythritol is a by-product (1) of the base-catalyzed condensation of acetaldehyde with formaldehyde. A number of different reaction mechanisms have been postulated to explain the formation of dipentaerythritol (11) from smaller building blocks in the Tollens hydroxymethylation reaction. Wawzonek and Rees (2) suggested a mechanism in which P-hydroxypropionaldehyde, from the condensation of formaldehyde with acetaldehyde, was considered to react with acrolein to form a dialdehyde (I). Further - reaction of I with formaldehyde by consecutive condensation and Cannizzaro reactions was considered to lead to dipentaerythritol (111. Wawzonek and Rees reached the conclusion that monopentaerythritol was ndt necessary for the formation of 11. A rather different reaction sequence was favored by Barth et al. (3). The hydrated dinler of formaldehyde (111) was thought to be capable of condensing with 2 moles of acetaldehyde, thus resulting in the generation of I, the precursor of the dimeric ether of monopentaerythritol (11). CHzOH I An alternate mechanism not heretofore considered seriously is the one in which monopentaerythritol is postulated as an intermediate which is converted by reaction with acetaldehyde and formaldehyde into the pentaerythrityl ether of P-hydroxypropionaldehyde (IV), which could readily be transformed into dipentaerythritol by further reaction with formaldehyde. 1P~esented at the Annual Meeting of the Chemical Institute of Canada in Montreal, August, 1961. Canadian Journal of Chemistry. Volume 41 (1963) 770
TREVOY AND MYERS: DIPENTAERYTHRITOL 771 Formation of ethers in the pentaerythritol reaction has also been observed when the reaction was conducted with alcohols as solvents (2,4). Thus Wawzonek and Rees isolated both the monomethyl and dimethyl pentaerythrityl ethers as products of reaction in methanolic solution. They explained this ether formation by the addition of the alcohol to the olefinic linkage of acrolein; the base-catalyzed addition of alcohols to a,p-unsaturated aldehydes is a well-established reaction. Wawzonek and Rees considered that acrolein may be an intermediate in both the formation of methyl pentaerythrityl ethers and in the formation of dipentaerythritol. By using a tracer technique it should be possible to distinguish between the mechanisms of Wawzonek and Barth on the one hand, in which dipentaerythritol is built up directly from acetaldehyde and formaldehyde, and the distinctly different mechanism in which a 5-carbon unit such as monopentaerythritol or pentaerythrose is first synthesized and reacts in further steps with acetaldehyde and formaldehyde to form dipentaerythritol. Thus, by introducing monopentaerythritol-c14 in the reaction medium and by examination of the dipentaerythritol produced for radioactivity, it should be possible to determine whether or not monopentaerythritol is a precursor of dipentaerythritol. EXPERIMENTAL Radioactivity Measurenzents Relative activities of mono- and di-pentaerythritol preparations were determined by counting solid samples under standard conditions. "Infinitely thick" samples were prepared by pelletizing 0.5 g of material into disks approximately 1.2 cm in diameter. The counting chamber was a CFlS-2 gas-flow counter purchased from Atomic Energy of Canada, Chalk River, Ontario, and Tracerlab G-1 Geiger Gas was used. Samples were counted for 10,000 counts in order to reduce the counting error to the 17, level. Correction was made for background count and the stability of the counter was checked periodically using a standard sample. (1) Synthesis of Monopentaerythritol-C14 Monopentaerythritol-C14 was synthesized by the sodium hydroxide catalyzed reaction of formaldehyde-cl4 with inactive acetaldehyde. The preparation was conducted on a 0.4-mole scale, using formaldehyde (20% aqueous solution, sodium hydroxide (20% aqueous solution), and acetaldehyde in the mole ratio of 4.6:l.l:l.O. The initial reaction temperature was 25' C and the Cannizzaro reaction was completed at 50' C. Total radioactivity used was 0.1 mc and the monopentaerythritol which was isolated possessedan activity of approximately 0.12 pc/mmole. After neutralizing the reactian mixture with acetic acid, excess formaldehyde was removed by distillation and the crude pentaerythritol crystallized upon cooling the residual solution. The monopentaerythritol-c14 was purified by con~ersi~n to the tetraacetate ester with acetic anhydride followed by fractional distillation to remove the dipentaerythritol hexaacetate. Methanolysis of the monopentaerythritol tetraacetate, b.p. 145-165' C at 0.4 mm, with methanolic hydrogen chloride regenerated the monopentaerythritol-c14, which was further purified by recrystallization kom water. (2) Synthesis of Pentaerythritol in the Presence of Monopentaerythritol-CX4 A pentaerythritol synthesis was conducted in aqueous medium, in which, in addition to the acetaldehyde, formaldehyde, and sodium hydroxide which are required for the condensation reaction, monopentaerythritol-c14 was also present. The monopentaerythritol-cl4 (0.5 mole, radioactivity 817 c.p.m.) was dissolved in water and placed in a stirred reactor. After cooling the mixture to 16' C, a formaldehyde solution (75 g of 20% aqueous solution) and a sodium hydroxide solution (2 g in 8 g water) were added. Acetaldehyde (0.5 mole) was then added to the reactor and the temperature was held at 16' C for an additional 20 minutes. More formaldehyde (262 g of 2070 aqueous solution) and sodium hydroxide (100 g of 20% aqueous solution) were added in succession. The reactor temperature was subsequently raised to 50' C for 90 minutes in order to complete the Cannizzaro
772 CANADIAN JOURNAL OF CHEMISTRY. VOL. 41, 1963 reaction. After concentration of the solution and removal of the excess formaldehyde by distillation, the crude product crystallized upon cooling the solution and was collected by filtration. The tnonopentaerythritol and dipentaerythritol were separated by conversioll to the acetate esters, fractional distillation at reduced pressure (1 mm), and alcoholysis of the individual ester fractions. The recovered monopentaerythritol (68 g) and dipentaerythritol (7.7 g) were purified by recr>stallization from water; purity was established by hydroxyl value determination (5), and the radioactivity levels of both monopentaerythritol and dipentaerythritol were determined. The final purified monopentaerythritol possessed an activity of 392 c.p.m. and hydroxyl \slue of 49.4 (calcd. for monopentaerythritol, 50.0); the purified dipentaerythritol had an activity of 316 c.p.m. and hydroxyl >slue of 39.0 (calcd. for dipentaerythritol, 40.0). (3) Synthesis of Pentaevythvitol in the Presence of Formaldehyde-C14 Formaldehyde-C14 (0.1 mc) was diluted with inactive formaldehyde and was used with inactive monopentaerythritol (0.25 mole) and inactive acetaldehyde (0.5 mole) in the synthesis of pentaerythritol. The procedure was similar to that used in Section (2). Based on pentaerythritol recovered and the pentaerythritol present in the mother liquor, the conversion of acetaldehyde to pentaerythritols was calculated to be 70%. An efficient separation of the acetates was achieved by distillation at to mm Hg pressure, and monopentaerythritol tetraacetate, dipentaerythritol hexaacetate, and tripentaerythritol octaacetate were recovered. Methanolysis of the esters yielded monopentaerythritol (31 g), dipentaerythritol (9.5 g), and tripentaerythritol (1 g). Radioactivities of the purified monopentaerythritol and dipentaerythritol, respectively, were found to be 1070 c.p.m. and 1255 c.p.m. (4) Effect of Mole Ratio of Monopentaerythritol to Acetaldehyde on Product Composition Using constant mole ratios of formaldehyde, acetaldehyde, and sodium hydroxide (4.5:l.O:l.l) the quantity of monopentaerythritol added prior to the condensation reaction was varied from 0 to 2.5 moles per mole of acetaldehyde. The product from each reaction was isolated, was recrystallized from aqueous solution after boiling the solution with dilute sulphuric acid in order to hydrolyze formals, and the hydroxyl value of the product was determined. The experimental data are presented in Table I and in Fig. 3A. The effect of monopentaerythritol conce~~tration formation of dibentaerythritol Mole ratio of initial monopenta- Pentaerythritol Hydroxyl Experiment erythritol to recovered* value Dipentaerythritol No. acetaldehyde (8) (%I (mole) *Based on the reaction of 1 mole of acetaldehyde. DISCUSSION OF RESULTS In the initial experiment in which monopentaerythritol-c14 was used, the recovered rnonopentaerythritol had approximately one-half the original activity because of dilution with a similar quantity of inactive rnonopentaerythritol which had been synthesized in the reaction from acetaldehyde and formaldehyde. The dipentaerythritol was found to be radioactive, which was a definite indication that rnonopentaerythritol had been a building block in the formation of dipentaerythritol. By analysis of the data obtained from this tracer experiment it is possible to conclude that all of the dipentaerythritol formed was built up from monopentaerythritol as starting material. The reactions by which mono- and di-pentaerythritol are fornled are coillpeting reactions and occur simultaneously. For the reaction mono-pe-c14 + CHICHO + 4CHz0 + OH- -+ di-pe-c14 + HCOO-
TREVOY AND MYERS: DIPENTAERYTHRITOL 773 with an initial inonopentaerythritol activity of 817 c.p.m. the initial dipentaerythritol activity would be 817)<(136/254) = 437 c.p.m. Similarly for a final measured monopentaerythritol activity of 392 c.p.in. the final dipentaerythritol formed would have an activity of 392X(136/254) = 210 c.p.in. Since dipentaerythritol is formed in parallel with rnonopentaerythritol via a competing reaction then the average monopentaerythritol activity of 604 c.p.m. should lead to an average dipentaerythritol activity of 323 c.p.111. Figure 1 presents the data for the reaction in which monopentaerythritol and 0 DIPENTAERYTHRITOL, g CONVERSION, '10 FIG. 1. Synthesis of pentaerythritol using monopentaerythritol-c'1. A. Dilution of monopentaerythrito1 during the condensation reaction. B. Dependence of dipentaerythritol activity on monopentaerythritol activity. dipentaerythritol were formed in the presence of pentaerythritol-c14. The experimentally determined activity of 316 c.p.m. for dipentaerythritol is the same as that-calculated within experimental error and indicates that all of the dipentaerythritol synthesized was formed via the monopentaerythritol route. In the tracer experiment in which C14-formaldehyde was used in the synthesis of monoand di-pentaerythritol, the data obtained are also in accord with a mechanism involving monopentaerythritol as an intermediate in dipentaerythritol formation. From the final observed activity of monopentaerythritol, and taking into account the 0.25 mole of inactive monopentaerythritol initially present, the activity of rnonopentaerythritol synthesized in the reaction may be calculated to be 1070X[0.60/(0.60-0.25)] = 1835 c.p.m. If all of the dipentaerythritol were formed directly from acetaldehyde and formaldehyde-c1"hen the dipentaerythritol activity would be 1835 X (272/254) = 1964 c.p.m. But since the dipentaerythritol activity was found to be 1255 c.p.m. then this is evidence against direct formation of dipentaerythritol from acetaldehyde and formaldehyde, and it is apparent that inactive monopentaerythritol initially present must also have been incorporated into the dipentaerythritol during synthesis. By comparing the calculated activity of 1964 c.p.m. with the observed activity of 1255 c.p.m. we may conclude that 64% of the monopentaerythritol units incorporated in the final dipentaerythritol were active and 36y0 were inactive. It follows that 72% of the dipentaerythritol was of the mixed variety, synthesized by reaction of inactive monopentaerythritol with inactive acetaldehyde and radioactive formaldehyde, while 28% was fully labelled dipentaerythritol, synthesized from radioactive monopentaery thritol or pentaerythrose formed
774 CANADIAN JOURNAL OF CHEMISTRY. VOL. 41, 1963 during the reaction which had entered into further reaction with radioactive formaldehyde. Assuming that the dipentaerythritol is formed in parallel with the monopentaerythritol, we may again calculate the dipentaerythritol activity based on an average rnonopentaerythritol activity during the reaction. By reacting monopentaerythritol, average activity 535 c.p.m., with inactive acetaldehyde and labelled formaldehyde-c14, the calculated activity of dipentaerythritol formed = [(535+ 1835)/2] X [272/254] = 1268 c.p.m., which is in good agreement with the observed activity of 1255 c.p.m. (see Fig. 2). INACTIVE MONOPENTAERYTHRITOL >. 0 r DIPENTAERYTHRITOL, g 9.5 CONVERSION, '10 FIG. 2. Synthesis of pentaerythritol using formaldehyde-c14. A. Change in rnonopentaerythritol activity during the synthesis. B. Dependence of dipentaerythritol activity on monopentaerythritol activity. It may be concluded that some inactive monopentaerythritol which was originally present in the reaction mixture was used for synthesis of dipentaerythritol. The experimental data presented in Fig. 3A show the effect of increasing monopentaerythritol concentration in the reaction mixture on the quantity of dipentaerythritol formed. The mass action effect of the increasing nionopentaerythritol concentration is responsible for the substantial reduction of the hydroxyl value of the isolated reaction product, indicating an increase in the quantity of dipentaerythritol formed. With monopentaerythritol to acetaldehyde ratios greater than 1 the effect in Fig. 3A is obscured by the large quantities of rnonopentaerythritol present. From the experimentally determined hydroxyl values the quantity of dipentaerythritol produced in each experiment may be calculated and in Fig. 3B the dipentaerythritol formed is observed to increase as the concentration of monopentaerythritol in the original reaction was increased from 0 to 2.4 moles per mole of acetaldehyde. It may be noted that when the ratio of monopentaerythritol to acetaldehyde was increased to 2.4 then approximately 40y0 of the acetaldehyde was converted to dipentaerythritol in the course of the condensation reaction. The fact that the addition of monopentaerythritol to the pentaerythritol reaction mixture resulted in the formation of increased amounts of dipentaerythritol provides additional evidence in support of the conclusion that monopentaerythritol is an intermediate in the reaction sequence leading to dipentaerythritol. The evidence presented here supports a reaction mechanism for dipentaerythritol formation in which a &carbon unit, monopentaerythritol, reacts with a 3-carbon unit (formed from acetaldehyde and formaldehyde) to generate an 8-carbon unit which in
TREVOY AND MYERS: DIPENTAERYTHRITOL 775 INITIAL MONOPEKTAERYTHRITOL ( MOLE RATIO) ACETALDEHYDE FIG. 3. The effect of monopentaerythritol concentration on the formation of dipentaerythritol in the condensation reaction. A. The effect of mole ratio of monopentaerythritol to acetaldehyde on the hydroxyl value of the isolated product. B. The relationship between dipentaerythritol formed and monopentaerythritol present in the condensation reaction. turn reacts in the final stage with two single-carbon formaldehyde units. The results obtained in this investigation are contrary to the previously suggested mechanisms in which either a dimeric formaldehyde unit or 4-oxa-1,7-heptanedial was assumed to provide the ether linkage which was eventually converted to dipentaerythritol. The present work has established that monopentaerythritol is an intermediate in the formation of dipentaerythritol and this conclusion is based on three different types of experimental evidence: (1) the isolation of labelled dipentaerythritol from a pentaerythritol reaction in which monopentaerythritol-c14 was used, (2) a comparison of the
776 CANADIAN JOURNAL OF CHEMISTRY. VOL. 41, 1963 radioactivity levels of monopentaerythritol and dipentaerythritol produced from formaldehyde-c14 and inactive rnonopentaerythritol and acetaldehyde, and (3) the dependence of the yield of dipentaerythritol upon the concentration of monopentaerythritol in the reaction mixture. ACKNOIYLEDGMENT The authors wish to acknowledge helpful discussions with Dr. J. W. T. Spinks, University of Saskatchewan. REFERENCES 1. 117. FRIEDERICK and W. BRUN. Ber. 63, 2681 (1930). 2. S. I17~wzo~~~ and D. A. REES. J. Am. Chem. Soc. 70, 2433 (1948). 3. R. H. BARTH, J. E. Show, and E. H. WOOD. Paper presented before North Jersey Section, -American Chemical Society, January, 1951. E. BERLOW, R. H. RARTH, and J. E. SYOW. The pentaerythritols. Reinhold Publishing Corp., New York. 1958. pp. 5-6. 4. R. H. BARTH. U.S. Patent No. 2,644,013 (June 30, 1953). 5. H. S. m71~so~ and IV. C. HUGHES. J. SOC. Chem. Ind. 58, 74 (1939).