Preparation of Acrylic Modified Rosin Noah J. Halbrookl and Ray V. Lawrence Naval Stores Laboratory, Southern Marketing and Nutrition Research Division, ARS, U S D A , Olustee, FL 3,9072
Gum, wood, and tall oil rosins were condensed with acrylic acid. Hard resins about one color grade darker than the starting rosin were obtained. The principal product was the dicarboxylic acid from the Diels-Alder reaction of abietic-type acids and acrylic acid. lsopimaric acid gave a monoaddition product. The composition of the unreacted rosin acids showed that other pimaric-type acids either react or isomerize to Asisopimaric acid. Most of the rosin neutrals disappeared during the condensation. About 6-1 0% of the reaction products appeared to be a mixture of rosin acid dimers and a product in which two molecules of rosin acids reacted with one molecule of acrylic acid. At temperatures between 225" and 240"C,there was no loss of carboxyl function during the reaction.
T h e reaction of rosin with acrylic acid gives a similar reaction product to t h a t obtained from propiolactone. The reaction products are a mixture of rosin acids and rosin dicarboxylic acids. The products of rosin and propiolactone have been used to prepare unsaturated polyester resins, which on copolymerization with styrene have excellent properties (Halbrook et al., 1963). T h e properties of the copolymers suggest their use as laminating, molding, casting, and coating resins. More recently, Mueller et al. (1969) have used these mixed rosin dicarboxylic acids in formulations for improved tackifiers in rubber. This paper describes the optimum conditions for reacting rosin with acrylic acid. The composition and physical properties of rosin condensed with 5-20 parts of acrylic acid per 100 parts rosin (phr) are given. The abietic-type acids of rosin (abietic, neoabietic, and palustric) react with acrylic acid to give the same dicarboxylic acids obtained when levopimaric acid is condensed with acrylic acid. This is to be expected since the abietic-type acids have all formed the same Diels-Alder adducts with numerous other dienophiles (Ruzicka et al., 1932; Fleck, 1944; Loeblich et al., 1955). The dicarboxylic acids consist of two pairs of epimers formed by an approach of the dienophile, acrylic acid, to the a face of levopimaric acid. The major products of t h e reaction have a positional assignment as indicated in Figure 1. The structures of the isomers of the adduct were reported by Halbrook et al. (1964). Gum, wood, or tall oil rosin from the southern pine usually contains 85-90% rosin acids. About 40-60% are of the abietic type, and 9-27y0 are of the pimaric type (Joye and Lawrence, 1967). We have found t h a t acrylic acid and isopimaric acid give monoaddition products. The nonacid portion is a complex mixture of alcohols, aldehydes, and other neutrals. Part of these neutrals would be expected t o react u i t h dienophiles. Experimental
Materials. Commercial grades of gum, wood, and tall oil rosin were used. T h e wood and gum rosin had a color grade of WW. T h e tall oil rosin had a color grade of X . T h e rosin acid samples were prepared by t h e heat isomerization of levopimaric acid (Loeblich et al., 1955) to give mixtures of abietic-type acids. Reagent grade arcylic acid and hydroquinone were used.
To whom correspondence should be addressed. 200 Ind. Eng. Chem. Prod. Res. Develop., Vol. 11, No. 2, 1972
General Reaction Conditions. Runs were made with 5-20 parts of acrylic acid per 100 parts of gum rosin at 200°, 2 2 5 O , 240°, 260°, and 275OC. T h e heat-isomerized levopimaric acid and the tall oil and wood rosins were condensed with 15-20 parts of acrylic acid (phr) a t 240OC. Polymerization of the acrylic acid was inhibited by addition of 150 ppm, rosin basis, of hydroquinone which was divided equally and added to the molten rosin and to the acrylic acid. The gum rosin (300 grams) was charged to a flask equipped with a stirrer, dropping funnel, inert gas inlet, thermometer, air-cooled condenser, and water trap topped with a watercooled condenser. The rosin was heated under a slow current of nitrogen to 225OC and held for 30 min. Stirring was commenced when the rosin had melted. After 30 min t h e pot temperature was adjusted to the desired reaction temperature, and the acrylic acid added slowly. The acrylic acid was added through a dropping tube below the surface of the rosin a t a low rate so t h a t little or no refluxing occurred. This required from 0.25 hr on the lower modifications to about 3 h r on the higher modifications a t the higher temperatures. The reaction temperature was held for 3 h r after all the acrylic acid was added. The heat-isomerized levopimaric acid and t h e tall oil and wood rosins were condensed with 15and 20 parts of acrylic acid (phr) a t 240°C. Pure isopimaric acid (1.0 gram) and excess acrylic acid were heated a t 240°C with stirring under nitrogen for 5 hr. Excess acrylic acid and acrylic acid polymer were removed by dissolving the product in benzene, filtering, and washing. The benzene was removed by evaporating a t 5-mm pressure and 100°C. Theisopimaric was converted into a mixture of an adduct having the same retention time as t h e abietic acid adduct, unreacted isopimaric acid, and As-isopimaric acid. Pure dehydroabietic acid, when heated for 5 h r a t 27OOC with an excess of acrylic acid, was recovered unchanged. Analytical Procedures. T h e acid numbers of the modified rosins were determined b y titrating a n alcoholic solution with 0.1N sodium hydroxide. Standardized procedures adapted for rosin were used in determining saponification numbers (ASTM, 1968a), color grades (ASTAI, 1968b), and softening points (ASTM, 1 9 6 8 ~ ) . The rosin acids in the reaction product were determined by gas chromatography (gc) of their methyl esters using procedures similar to those described by Joye and Lawrence (1967). The methyl esters were prepared by adding an ethereal s o h -
Table 1.
Preparation, Properties, and Composition of Rosin and Acrylic Acid Reaction Products
%
Prepn. no.
Reaction temp, OC
Acrylic acid, g/100 g rosin
Acid, no., mg KOH/g
Softening pt. ring & ball, 'C
Sopon. no., mg KOH/g
composition, wt Abietic Dehydrotype abietic acid acids
Acrylic acid adduct
Pimaric type acids
Rosin acids not accounted for b y gc
44.0 52.4 50.0 47.0 39.0 26.0 15.2 38.0 40.0 15.0 36.0 38.2
26.0 13.3 13.4 12.8 13.0 10.4 7.6 20.9 12.4 13.0 18.1 16.5 11.7
55.0 4.8 4.6 4.7 3.5 5.4 5.8 23.8 18.4 14.6 27.6 27.6 10.2
5.0 6.7 5.0 5.4 6.5 9.2 8.6 4.7 4.3 5.4 7.8 6.5 7.2
11.2 6.1 8.0 10.7 15.0 28.6 20.4 12.5 9.4 16.3 14.3 15.0
39.5 40.0
20.5 14.2 11.3
40.3 2.9 1.3
26.4 24.0 24.0
5.8 6.3
45.0 49.0
21.4 13.4 11.9
56.3 5.9 1.6
9.0 13.3 12.0
5.7 6.4
1.6 1.4 2.1
96.1 36.9 5.2
2.3 2.4 3.8
11.0 9.9
Gum rosin 2 3 4 5 6 7 8 9 10 11 12
200 225 240 250 260 275 225 225 225 250 250 250
20 20 20 20 20 20 10 15 5 10 15
165 252 255 252 243 238 215 184 21 1 230 186 208 227
13 14
240 240
15 20
162 236 256
15 20
162 23 1 255
10 20
186 240 276
1
5
177 273 274 273 267 262 242 20 1 222 246 198 225 243
242 268
118 122 117 124 123 121 96 111 116 102 117 124 Tall oil rosin 105 105 Wood rosin
15 16
17 18
240 240
240 240
241 267
119 119
Heat-isomerized levopimaric acid 186 241 125 47.2 127 77.4 278
tion of diazomethane to a methyl alcohol solution of the sample. For determination of t h e acrylic acid adduct, a n 8-ft X I/4-in. 0.d. copper column packed with 5% silicone rubber W-98 on 60/80 mesh Chromosorb-W was used. The trimethyl ester of fumaropimaric acid was used as am internal standard. A t a column temperature of 250°C, injector and detector temperatures a t 315' and 3OO0C, and helium flow of 80 ml per min, the trimethyl ester of fumaric acid had a retention time of 17.5 min. The ester of the acrylic acid adduct of levopimaric acid had a response of 1.20 relative t o the fumaropimaric acid ester and was separated into four peaks with retention times of 0.49, 0.62, 0.66, and 0.70 relative to the internal standard. Analysis for rosin dimer and related products was made by the method described by Sinclair et al. (1971). The adducts were also separated in acid form from the unreacted rosin acids by liquid partition chromatography (Halbrook et al., 1964). The neutral materials in rosin and t h e condensates were determined by passing a n ethereal solution through activated alumina (Moseley and Stanley, 1970). Discussion
The optimum temperature range for adduct formation under the conditions used is from 225' t o 240'C. When the inhibitor, hydroquinone, was not used, or when the reaction was carried out under reflux, acrylic acid polymer formed and yields of adduct were low. The polyacrylic acid, which could be seen as insoluble specks in t h e modified rosin, gave a flocculent precipitate on solution of the modified rosin in alchool. Some reflux was unavoidable in runs made a t 260-275'C. This reflux continued until the reaction was stopped. The gum
fl+
H3C
COOH
COOH
Figure 1. Reaction of levopimaric acid with acrylic acid
rosin used contained 55% abietic-type acids. Complete reaction of the abietic-type acids would require 13 parts of acrylic acid (phr) to give 68 parts of adduct. The addition of 1 mole of acrylic acid t o each mole of the rosin acids present would require about 20 parts of acrylic acid. Unreacted rosin acids were found in t h e products condensed with 20 grams acrylic acid (phr). Therefore, several runs were made with 25 parts of acrylic acid. There was no increase in adduct formation over t h a t obtained with 20 parts of acrylic acid. The unreacted acids were about the same. The percent composition by weight of acrylic acid adduct and unreacted rosin acids is given in Table I. These weight values do not account for all the rosin acids. Some side reactions would be expected to occur in the complex rosin mixture. The gum rosin used contained 12% of neutral materials. The product obtained when gum rosin was condensed with five parts of acrylic acid (phr) had 47, neutral material and when condensed with 20 parts of acrylic acid (phr) had 1.8% neutral material. Gas chromatography of several products by the method of of the reaction Sinclair et al. (1971) showed t h a t 6-10 wt products had retention times approximately equivalent to those of the rosin acid dimers. These high-molecular-weight Ind. Eng. Chem. Prod. Res. Develop., Vol. 1 1 , No. 2, 1972
201
products were probably a mixture of rosin acid dimers and the product formed by t h e reaction of two molecules of rosin acid with one molecule of acrylic acid. This material accounts for most of the rosin acids not accounted for either as free acid or acrylic acid adduct. The amount of all the resin acids, except dehydroabietic and A*-isopimaric, decreased during t h e reaction of the rosin with the acrylic acid. These two acids increased slightly as they usually do when rosin is heated. Experiments with pure dehydroabietic showed t h a t it did not react with acrylic acid. Pure As-isopimaric acid was not available to check its reaction with acrylic acid. Since it is present to the extent of only 201, in the starting rosin and 4 4 % in the reaction product, i t would have little effect on the reaction Abietic and isopimaric acids reacted simultaneously, but the pimaric-type acids appeared to react slightly slower. At 260°C with a slight excess of acrylic acid, the isopimaric acid reacted almost completely. Pure isopimaric acid, when heated with acrylic acid, converted to a mixture of monomeric addition product, isopimaric acid, and As-isopimaric acid. Most of the methyl ester of this addition product had the same retention time on gc analysis as the major product obtained when rosin is condensed with acrylic acid, but it is different. The abietic acid-acrylic acid adduct of rosin forms a carbontetrachloride solvate. The isopimaric acid condensation product isolated by liquid partition chromatography does not. Partition chromatography of the modified rosins gave a separation of the rosin acids and adduct. The recovery of the carboxyl placed on the column agreed with that obtained by gc analysis for adduct and unreacted rosin acids. On stripping the column with toluene, additional carboxylic material was obtained approximately equivalent to t h a t in the unaccounted for rosin acids. The greater than normal divergence between acid numbers and saponification numbers (McKelvy et al., 1957) of the acrylic acid modified gum rosin as compared with gum rosin is probably due to esterification of rosin alcohols such as elliotinol with the acrylic acid and the acrylic acid adduct. The gum rosin had 4 4 % elliotinol and a hydroxyl value of 23. No elliotinol was observed in the wood and tall oil rosins. The rapid drop in acid numbers and saponification numbers of
the preparations made above 24OOC could have been caused by decarboxylation. The modified rosins had color grades of N to WW. Conclusions
Abietic-type acids of rosin react with acrylic acid to form the expected Diels-Alder adduct. Pimaric-type acids form a 1: 1 addition product. Some isomerization of isopimaric acid occurred. Overall decrease of other pimaric-type acids suggests t h a t all the rosin acids except dehydroabietic acid and A8-isopimaric form addition products with acrylic acid. Saponification numbers show t h a t no appreciable amount of decarboxylation occurred below 240OC. About 6-10 wt % of the products have retention times by gc analysis indicating molecular weights in t h e range of 600. These were probably products involving the condensation of 2 moles of rosin acids with 1mole of acrylic acid. literature Cited
ASTM D 464-59, Part 20, 249, Philadelphia, PA, 1968a. ASTN D 509-33, Part 20, 269, Philadelphia, PA, 196813. ASTU E 28-67, Part 20, 1105, Philadelphia, PA, 1968c. Fleck, E. E., U.S. Patent 2,359,980 (October 10, 1944). Halbrook, 3.J., Lawrence, R. V., Dalluge, hl. D.. Stein, G. A., Ind. Eng. Chem. Prod. Res. Develop., 2, 182 (1963). Halbrook, N. J., Lawrence, R. V., Dressler, R. L., Blackstone, R. C., Herz, W., J . Org. Chem., 29, 1017 (1964). Joye, Jr., N. >I., Lawrence, R. V., J . Chem. Eng Data, 12, 279
11967). Loeblich, V. M.,Baldwin, D. E., O'Conner, R. T., Lawrence, R V., J . Amer. Chem. Soc., 77, 6311 (1955). McKeIvy, J. B., McConneII, N. C., Joye, Jr., N.XI., Lawrence, R. V., Paint, 021 Chem. Rev., 120 (14), 10 (1957). Moseley, P. B., Stanley, J. B., J . Amer. 021 Chem. SOC.,45, 547 iic17n~ ",. , A U #
llueller, W. J., Bennett, B., Halbrook, N. J., Schuller, W. H., Lawrence, R V., Rubber Age, 101 (7), 43 (1969). Ruzicka, L , i2nkersmit, P. J., Frank, B., Helv. Chim. Acta, 15, 1289 (1932). Sinclair, R. G., Hinnekamp, E. R., Boni, K. -4., Berry, D. A., Schuller, W. H., Lawrence, R. V., J . Chromatogr. Scz., 9 , 126 (1971). RECEIVED for review October 12, 1971 ACCEPTED January 13, 1972 hIention of firm names or commercial products does not constitute an endorsement by the U.S. Department of Agriculture.
Volatility of Dioctylsebacate on Tinplate and Electrolytic Chromium-Plated Steel Larry A. Nimonl and Glen K. Korpi Metals Research and Engineering, Continental Can Co., Inc., 7662 South Racine ilvenue, Chicago, IL 60620
D u r i n g the manufacture of tinplate (TP) and electrolytic chromium-plated steel (CCO), a thin film of di(2-ethylhexyl) sebacate, commonly known as dioctylsebacate or DOS, is applied to the surface. The DOS is applied a t a level of about 0.5 rg/cm2 and serves as a lubricant during subsequent fabrication operations. This investigation attempted to To whom correspondence should be addressed. 202
Ind. Eng. Chem. Prod. Res. Develop., Vol. 1 1 , No. 2, 1972
determine relevant differences in the volatility and decomposition characteristics of the as-received DOS-TP and DOSCCO systems. Little is known about DOS-metal systems (Yonezaki, 1970); a radioactive tracer study of DOS films on tinplate (Davis, 1970) concluded that the disappearance of DOS during storage was due to a combination of hydrolysis and decarboxylation. The authors assumed hydrolysis products