INDUSTRIAL AND ENGINEERING CHEMISTRY
April 1955 Table 111.
Low Temperature Flexibility and Plasticizing Efficiency Related to Compatibility
Britile point, ‘C. C. brittle point Quantitative exudation test, ’X loss Shore A hardness Modulus (lOO~o), lb./sq. inch
-’;
(Alkyl pinate esters) OctylDi-nDi(2-ethyloctyl hexyl) dedyl -46 -48 -40 -65 -58 -46.5 19 10 -0.5
-
-
Di-butoxyethyl -47 -46
Di-nhexyl -54 -50.5 f3.5
+I
4.3 78
3.4 72
2.2 70
1.9 66
1.6 65
1310
1300
1300
1190
1110
Table IV. Predicted versus Observed Behavior in Polyvinyl Chloride of 1 :1 Octyldecyl Pinate-Dioctyl Phthalate Britile point,
’ C.
gkore? hardness Modulus (loo%), lb./sq. inch
Octyldecyl Pinate - 46 65 78
-
1310
DOP -31 -33.5 68
Calcd. Mean -38.5 -49.3 73
1190
1250
1 : 1 Octyldecyl Pinate-DOP -42 -47.5 69 1220
Gross differences in plasticizer permanence properties are in accord with expectations-Le., ( a ) increased volatility with decreasing molecular weight and with branching in the alkyl group; (6) increased water and soapy water sensitivity with decreasing molecular weight and with the presence of an ether linkage in the alkyl group; (c) increased gasoline extraction with increaeing length of the alkyl group and decreased gasoline extraction as a result of the presence of an ether linkage in the alkyl group. The observed apparent decrease in oil extraction with increasing chain length of the alkyl group is probably caused by ( a ) a
855
tendency toward higher oil absorption by the films plasticized with the higher alkyl pinates and ( 6 ) compensation for the probable inherently greater oil sensitivity of the higher alkyl pinates by their lower rates of diffusion. Minor deviations from the predictable order of influence of this series of pinates on permanence properties can be interpreted in terms of the extent of deviation from complete compatibility. Thus, the abnormally high volatility and soapy water extraction losses of the octyldecyl pinate plasticized films are undoubtedly composites of loss to the indicated hazard plus loss through exudation. This likelihood is borne out by the fact that both the volatility and soapy water extraction values shown by 1: 1 octyldecyl pinate-dioctyl phthalate are much lower than the corresponding values for each of these two esters when present as the sole plasticizer. CONCLUSION
The permanence, stability, and low temperature properties of the n-octyl, octyldecyl, and 2-ethylhexyl diesters of pinic acid make these esters useful secondary plasticizers for polyvinyl chloride. Pinic acid diesters derived from lower alcohols are excessively volatile while the di-n-decyl ester is for most applications inadequately compatible as a secondary plasticizer. LITERATURE CITED
(1) A.S.T.M. Standards, 1949, Part 6, p. 546. (2) Ibid., p. 574. (3) Murphy, C. hl., O’Rear, J. G., and Zisman, W. A,, IND.ENG. CHEM., 45, 119 (1953). (4) Rider, D. K., and Sumner, J. K., IND.ENG.CHEiw., A N . ~ LED., . 17,730 (1945). RECEIVED for review August 20, 1954.
ACCEPTED Sovernber 19, 1954.
Terpene-Derived Plasticizers PREPARATION OF PINIC ACID AND ITS ESTERS VIRGINIA & LOEBLICH ‘I. Naval Stores Research Station, Olustee, Flu.
FRANK C. MAGNE AND ROBERT R. MOD Southern Regional Research Laboratory, New Orleans, La.
I
Pi’THE past few years there has been increasing demand for a
domestic supply of dibasic acids, such as sebacic acid, that could be used in the preparation of synthetic lubricants, low temperature plasticizers, polymers, resins, and fibers. a-Pinene, the main constituent of turpentine, will, by stepwise oxidation, yield a series of dibasic acids; three of these are structurally identified as shown in Figure 1. The structural similarity of these acids t o the more common dicarboxylic acids suggests their potential application in the synthesis of plasticizers and low temperature lubricants. While the presence of cyclic groups, such as phenyl or cyclohexyl, in diesters is generally considered unfavorable to their performance as satisfactory low temperature lubricants by virtue of the large temperature coefficients of viscosity, high freezing temperature, or pour points imparted, Murphy, O’Rear, and Zisman ( 6 )have shown that the presence of the cyclobutane ring in the pink acid (I) diesters does not cause such adverse effects. Therefore, the esters of pinic acid should be potentially good low temperature plasticizers and those of sym-homopinic acid (11) should be better ones because of the more centered position of the
cyclobutane ring. Although several isomers of each acid (I, 11) are indicated from structural considerations, this study covers only esters of what is reported as the d-trans isomer of pinic (6) and symhomopinic acids ( 2 , 9 ) . The octyl P-(hydroxyisopropy1)pimelate y-lactone (111) (S), on the other hand, with its oxygen-containing ring was hoped to have an enhanced compatibility as well as the middle-range low temperature characteristics of an alkyl ester somewhere between a phthalate and an adipate.
VINYL PLASTICIZERS based on domestic turpentine co nstituents
. . . are promising plasticizers for polyvinyl chloride and PVC-PVA copolymers
. . . in some cases rival sebacic acid esters in physical properties and performa nce
INDUSTRIAL AND ENGINEERING CHEMISTRY
856
Vol. 47, No. 4
flask fitted with a reflux condenser and a Barrett distilling receiver. The mixture was heated a t reflux until no more water accumulated in the distillation trap. The solution was cooled, transferred t o a 5-liter separatory funnel, washed with water, then extracted with 850 ml. of 5% sodium hydroxide, and finally washed eight times with water to neutrality. The toluene and excess alcohol were removed by distillation. The crude ester was then distilled with fractionation through a Vigreux column 33 cm. long and 35 mm. in diameter provided with a heating element. A vacuum jacketed 75" angle distilling head, condenser, and fraction cutter completed the distillation apparatus. The heart cut of the ester distilled over a temperature range of 153' to 175' C. and a pressure range of 0.15 to 0.40 mm. The fractions were combined on the basis of refractive index-a difference not exceeding f0.0002 unit of the standard value used in this work, 1.4520 a t 20" C.-and acid number not greater than one. A yield of 790 grams of acceptable dihexyl pinate was obtained. This represents a 48Y0 yield of acceptable ester based on the weight of distilled pinic acid used.
a-pinene
CH3
H2
Figure 1
PLASTICIZER SCREENING
Accordingly, the method of preparation of pinic acid and several of its diesters, the plasticizing characteristics of these esters, and a typical ester of the two other terpene-derived acids are described here.
These esters have all been screened as plasticizers for 95% polyvinyl chloride-5yo polyvinyl acetate copolymer (VYDR), and those with the more promising characteristics were given an additional screening with Geon 101 (polyvinyl chloride). All esters were compounded for testing purposes according to the following basic formulation:
PREPARATION AND PURIFICATION OF PINIC ACID Vinylite VYDR or Geon 101 Plasticizer Stearic acid Basic lead carbonate
The oxidation of a-pinene to pinic acid was carried out on small pilot plant scale following the general procedure of Delepine ( 1 ) : Thirty-four pounds of commercial a-pinene and 80 pounds of potassium permanganate were allowed to react in the presence of 18 pounds of ammonium sulfate buffer and 76 pounds of water. The reaction temperature was kept below 10' C., utilizing about 700 pounds of ice. After completion of the oxidation, the reaction mixture was centrifuged t o remove the manganese dioxide, concentrated to one third the volume, and then acidified; 20% pounds of crude crystalline pinonic acid (44.5% yield) was obtained. The crude pinonic acid was oxidized in water solution with 38 pounds of calcium hypochlorite. The pinic acid was recovered from t h e water solution by continuous ether extraction. 11 pounds (53.6% yield based on weight of pinonic acid used) of crude pinic acid, neutral equivalent = 105 115, was obtained. The over-all yield based on or-pinene was 25%.
-
Wt. % 63.5 35
0.5 1.0
I n those instances where more than 35y0 plasticizer was used the resin content was varied in accordance with the formula, per cent resin = (98.5 - z), where z is the per cent plasticizer. These formulations were milled and molded at 310' F. The detailed procedures followed in these operations and also the preparation of the test specimens have been previously reported (6). The plasticized stocks were tested for tensile strength, 100% modulus, and elongation using an IP-4 Scott Tester at a loading rate of 200 pounds per minute a t 70' F. and 65 f 2% relative humidity. The brittle point was basically that described in the ASTM test D746-44T. All compatibility observations were based on examination of the 0.07-inch molaed panels for exudation after a period of aging at room temperature. Vapor pressures at various temperatures were determined by means of a tensimeter still ( 7 ) . Volatilities were determined in duplicate on specimens 1 inch in diameter and 0.07 inch thick, by both the oven and activated carbon methods. The oven method (4)involved a 96-hour exposure a t 100 i. 5' F. in a forced draft oven while
The crude pinic acid contained some chlorinated acids which were difficult to remove after esterification. Therefore, the crude acid was first "dehydrochlorinated" and then vacuum distilled, The dehydrochlorination step consisted in heating the acid at 160' to 170" C. under 3 to 5 mm. pressure for 2 hours. The acid was then fractionally distilled att2 to 3 mm. pressure, and the heart cuts were combined on the basis of neutral equivalent. The acceptable neutral equivaTable I. Characteristics of Copolymer-Plasticizer Compositions lent range was 93-100 (theory = 93). ESTERIFICATION OF PINIC ACID
The preparation of dihexyl pinate illustrates the procedure used in the synthesis of all the diesters of pinic acid. A mixture of 869 grams (4.67 moles) of distilled pinic acid, 870 ml. of toluene, 1397 ml. ( 2 : l mole ratio 20% excess) of normal I-hexanol, and 24 grams of 1: 1 sulfuric acid were placed in a 5-liter, three-necked
+
Plasticizer Bis(butoxyethy1)pinate Diamyl pinate Dihexyl pinate Dioctyl pinate Dihexyl sun-homopinate
P!astioizer,
%
35 35 35 35 40 35 40
Tensile Strength, Lb./Sq. Inch 2790 2760 2740 2670 2280 2810
...
100% Modulus, Lb./Sq. Ultimate Inch Elongation, % 1250 320 1260 330 1330 300 1490 320 1180 350 1260 3!0 980
Octyl P-(hydroxyisopropyl)pimelate 30 3260 1590 3tO ... 1140 y-lactone 35 Control (dioctyl 35 3040 1550 330 phthalate) Dioctyl sebacate 35 2520 1410 310 (monoplex DOS) 40 2180 870 350 a Activated carbon test 24 hr. a t 70' C. Oven methqd, 96 hr. at 100 =k 5' C. (4). c Beyond limit measurable on IP-4 tester, on I-inch standard test length.
Brittle Point, C . 45 - 44 - 49 54 -71 - 52 - 58
-
-
Volatility Loss, % Activated carbon Oven methoda methodb ... 6.9
... ...
0.40 . . I
...
14.0
...
...
0.72
10.5
- 32 - 33
0.36 0.50
4.6 5.3
- 58
... ...
- 23
- 69
... ...
... ... ...
...
April 1955
INDUSTRIAL AND ENGINEERING CHEMISTRY
857
The dihexyl esters of pink and sym-homopinic acids are best Tensile 100% Volatility Loss, % in imparting low temperature P!asti- Stren th, Modulus, Activated characteristics t o p o 1y v i n y 1 mer, Lb./gq. Lb./Sq. Ultimate Brittle carbon Oven Plasticizer % Inch Inch Elongation, % Point, C. methoda methodb chloride stock but even the Dihexyl pinate 35 300 -48 ... 19.6 poorest are comparable to Diootyl pinate 35 170 ... ... -32 dioctyl phthalate. 37 200 - 28 ... ... 40 200 40 ... ... The vapor pressures for these 50 -49 to -53 ... ... esters, reported in Table 111, Dihexyl sum-homopinate 35 300 1 2 . 8 -51 were read from log of vapor Octyl p- (hydroxyisopropyl) pimelate 7-lactone 30 3610 2180 250 -27 ... ... pressure versus 1/T plots of the 35 2970 1370 320 ... 4.6 - 35 40 ... 940 ... ... -41 experimental data. E x c e p t ' Control (diootyl phthalate) 35 3110 1820 260 4.7 0.35 -36 for the diamyl and dihexyl 40 2520 1200 310 ... ... - 37 pinate these esters all exhibit a Activated carbon test, 24 hr. a t 70° C. Oven method, 96 hr. a t 100 zt 5O C. (4). vapor pressures below 4 mm. 0 Beyond limit measurable on IP-4 tester, on I-inch standard test length. a t 204' C.-an empirical value which in general should not be exceeded by potential plasthe activated carbon test was that described in the ASTM test ticizers (4). As an approximate measure of plasticizer volatility, D1203-52T. Losses are reported on the basis of the original this indicates that these esters, excluding the two mentioned, weight of the plasticized samples. might have a satisfactory degree of permanence in the resin. Table 11.
Characteristics of Polyvinyl Chloride-Plasticizer Compositions
O
I . .
POLYVINYL
a 3
CHLORIDE-POLYVINYL STOCKS
ACETATE
COPOLYMER
The milling and molding characteristics of all these esters were quite satisfactory with the copolymer; the dihexyl sym-homopinate and the octyl p-(hydroxyisopropy1)pimelate 7-lactone were outstanding in this respect. All molded slabs were opaque white or cream color depending on the water whiteness of the plasticizing ester. None of the esters has shown any signs of incompatibility with the copolymer during observation periods over 1 year for the pinates, 6 months for the homopinate, and 3 months for the ester of the @-(hydroxyisopropyl)pimelic acid 7-lactone. The characteristics of these esters as plasticizers for the copolymer are shown in Table I. All are very efficient plasticizers as indicated by the 100% modulus, which is below that observed for dioctyl phthalate in all instances and above that of dioctyl sebacate in only one. With but one exception they produce copolymer stocks having excellent low temperature characteristics; in fact dioctyl pinate seems t o be about the equal of dioctyl sebacate in this respect. POLYVINYL CHLORIDE STOCKS
The results of the tests on some selected esters as plasticizers for polyvinyl chloride are shown in Table 11. These data show that the plasticizers perform somewhat differently in polyvinyl chloride stock and in most respects not quite so satisfactorily as in the copolymer stock. Modulus and elongation are adversely affected, in a varying degree, and in some instances tensile strength and brittle point. Dioctyl pinate in particular is much less effective in Geon 101, showing greatly impaired milling characteristics bordering on incompatibility at concentrations in the 40 t o 50y0range. Polyvinyl chloride stocks containing this ester have a much poorer modulus, elongation, tensile strength, and brittle point than correspondingly plasticized copolymer stocks. Dioctyl pinate does not offer plasticizing characteristics comparable to dioctyl sebacate in polyvinyl chloride ( 8 ) . While it is true that no incompatibilities have been observed for any of these plasticizers a t or below the 40% level aft,er 3 months of aging, the fact that dioctyl pinate mills badly and does exude a t the 5Oy0 level during this time leaves some doubt as to the finality of these observations, since it is quite possible that additional failures might develop over a more prolonged period. Dihexyl sumhomopinate and the octyl @-(hydroxyisopropyl) pimelate 7-lactone stand out as the most efficient of these plasticizers with polyvinyl chloride as evidenced by their low moduli. However, all are superior to dioctyl phthalate in this respect.
Table 111. Plasticizer Vapor Pressures Ester Bis(butoxyethy1) pinate Diamyl pinate Dihexyl pinate Diootyl pinate Dihexyl sym-homopinate Octyl p- (hydroxyisopropyl) pimelate 7-lactone Control (dioctyl phthalate)
Boiling Point, ' C., a t Specified Pressure 0.04 0 . 1 0 . 4 0 . 5 1.0 4 . 0 10.0 mm. mm. mm. mm. mm. mm. mm. 135 149 173 177 191 219 241 104 117 139 142 156 182 202 116 131 158 169 173 202 224 151 166 191 194 208 237 257 130 145 168 172 185 213 234 141 141
156 154
178
.
.
182 181
198 193
223 220
244 239
Volatility determinations by two different methods (Tables I, 11), however, indicate that the percentage of loss from stocks plasticized with either dihexyl pinate or sym-homopinate is considerably higher than from the control-plasticized stock. These results indicate that vapor pressures of 1.8mm. or less a t 204' C. are essential in this group of plasticizers to achieve volatility losses below the control. The darkening and stiffening which developed in some of the plasticized stocks during the oven exposure correlated very closely to the volatility losses observed. Those specimens, containing dihexyl pinate, were considerably affected, those with dihexyl sym-homopinate and bis(butoxyethyl)pinate, much less, and the others not to any noticeable extent. SUMMARY
The preparation of pinic acid from a naturally occurring terpene and of its bis(butoxyethyl), diamyl, dihexyl, and dioctyl ester has been reported. Vapor pressures have been determined on these esters and two other terpene-derived esters, dihexyl sym-homopinate, and the octyl p-( hydroxyisopropy1)pimelate y-lactone. All have been screened for their plasticizing characteristics with 95yo polyvinyl chloride-5 yo polyvinyl acetate copolymer and in some instances with polyvinyl chloride. The results show that all are satisfactory plasticizers for the copolymer, and although they are somewhat less compatible and not quite so effective with polyvinyl chloride, they will, except for the dioctyl pinate, perform adequately as plasticizers. It appears that the presence and location of the cyclobutane ring in the plasticizer are of greater import with polyvinyl chloride than with the copolymer. Reduced compatibility results when it is unsymmetrically located, as in the pinates, whereas in the sym-homopinates the effect is not so apparent: Furthermore, the more symmetrical configuration of the sym-homopinate results in better heat stability, greater permanence, and improved
INDUSTRIAL AND ENGINEERING CHEMISTRY
858
low temperature characteristics in the resin. Although the presence of the oxygen-containing ring in the pimelic acid derivative results in a sacrifice of low temperature characteristics, there is less difference in its performance in the two resins, both as to modulus and compatibility. I n addition, it exhibits a better permanence and heat stability than either the same alkyl pinates or sym-homopinates. Where compatibility is borderline] the incorporation of primary plasticizers-e.g., dioctyl phthalate-with such esters might be desirable. The terpene-derived acids described, two of which yield esters comparable to those of sebacic acid in phyEica1 properties and performance as lubricants and plasticizers, can be made from constituents of domestic turpentine. LITERATURE CITED
( 1 ) Delepine, AT., Bull. SOC. chim., France, 3 (5), 1369-82 (1936).
(2) Guha, P. C., and Ganapathi, K., Current Sci. ( I n d i a ) , 5, 244 (1936).
Vol. 41, No. 4
(3) Halbrook,
K.J., and Lawrence, R. V., private communication. (4) Kent, D. L., and Weaver, P. J., India Rubber World, 115, 813-16 (1 947). (5) Magne, F. C., and Mod, R. R., IND. ENG.CHEM.,45, 1546-7 (1953). (6) Murphy, C. M., O'Rear, J. G., and Zisman, W. A., Ibid., 45, 11930 (1953). (7) Ratchford, W. P., and Rehberg, C. E., Anal. Chem., 21, 1417-19 (1949). (8) Rider, D. K., Sumner, J. K., and Myers, R. J., IND.ENG.CHEM., 41,709-15 (1949). (9) Stinson, J. S., and Lawrence, R. V., J . 070. Chem., 19, 1047-53 (1954). RECEIVED for review July 15, 1954. ACCEPTED November 4, 1984. Southern Regional Research Laboratory and the Naval Stores Research Section, laboratories of the Southern Utilization Research Branch, Agricultural Research Service, U. S. Department of Agriculture. Mention of the names of firms or trade products does not imply that they are endorsed or recommended by the U. 9. Department of Agriculture over other firms or similar products not mentioned.
Lactic Acid Purification by Extraction J
ROBERT B. WEISER' AND CHRISTIE J. GEANKOPLIS T h e Ohio State University, Columbus 10, Ohio
ISOAMYL ALCOHOL
.. .is
a good selective solvent for liquid-liquid extraction of crude lactic acid
. . .with steam distillation for solvent recovery offers some industrial possibilities for obtaining high purity lactic acid
A
CCORDING to Needle and Aries (14) the potential annual market for lactic acid is about 200,000,000 pounds if the price of the purer grades can be reduced significantly. Principal new markets would be for the manufacture of alkyd, acrylic, and other resins. Considerable recent work has been carried out to study methods of purification, which represent a considerable portion of lactic acid production costs. One method that has received attention and is currently used is the esterification of the acid in the crude solution and consequent separation of the ester (2,6-7). Liquid-liquid extraction has also been investigated as a possible purification method (11-13, 17-19). I n a patent, Waite (20) has discussed the extraction of the acid with amyl alcohol. Background information on the various purification methods and physical properties of lactic acid are also available ( 4 , 16, 21). Bass (1)reports that usual impurities in crude lactic acid solutions are alkaline earth metal salts, sugars, and volatile impurities such as acetic and butyric acids. The separation of lactic acid from these impurities is complicated by the formation of a self-polymer of lactic acid when the acid concentration rises above about 20%. Because liquid-liquid extraction seemed promising as a . purification method, this research investigation was 1 Present address, Polychemicals Department, E. I . du Pont de Nemours and Co., Inc., Wilmington, Del.
undertaken to determine the distribution of lactic acid and the various impurities between water and a wide variety of organic solvents under various operating conditions. Then, the best selective solvent was chosen, and the remaining processing problems were investigated to determine the potentialities of the extraction process. SOLVENT SEARCH
Materials and Analytical Methods. The lactic acid used for the majority of the tests was C.P. grade. A few tests were also run with the 22y0 technical grade of lactic acid. All the chemicals employed for analysis met ACS standards of purity. The organic solvents were used as obtained for the preliminary solvent search. Wherever possible, acid solutions were analyzed by colorimetric titration with standard base. When the concentration of lactic acid in water solution is above about 20% analysis, it is complicated by the presence of a self-polymer of lactic acid (3, 10). I n this case, excess base was added t o the solution and the mixture was heated for about 10 minutes in a boiling water bath. During this time the flask was stoppered, but a fine drawn-out tube cut down carbon dioxide absorption from the air. The excess base was then immediately back-titrated with standard sulfuric acid. This method was accurate within f0.2%. The presence of inorganic salts had no effect on the analysis of the acid, For analysis of acid in the solvent layer, water was added so that the colored end point could easily be detected. I n every analysis of all types, blanks were determined. A semiautomatic titrimeter was used for acid-base analysis in colored solutions. The end point setting was between 7.0 and 7.8 pH for lactic acid solutions. The method of Friedemann and Graeser (8) was used when other acids were present in the lactic acid solution. I n this method lactic acid is oxidized to acetaldehyde, and the aldehyde is collected in bisulfite. The bound bisulfite is determined with standard iodine, and a measure of the lactic acid content is obtained. This method was accurate within &0.2y0when compared with knowns. This method will yield erroneous values in the presence of sugars. I n some cases the distribution coefficient of sugars alone between water and organic solvents was determined. Sugars were analyzed
