Aueust. 1924
INDUSTRIAL A N D ENGINEERING CHEMISTRY
823
The Equilibrium between Lactic Acid and Its Anhydrides' By G.I. Thurmond and Graham Edgar UNIVERSITY OF VIRGINIA,UNIVERSITY, VA.
OMMERCIAL lactic that this might be the case, A method of analysis has been devised for determining all four acid of the British but no experimental data constituents of mixtures of lactic acid, lactic anhydride, lactide, and and U. S. Pharmaon the subject occur in the water. This method has been applied to the analysis of mixtures of literature. copeia is supposed to conthese constituents heated for some time at 155" and 100" C. The present paper pretain about 75 per cent of The results indicate that a definite and calculable equilibrium consents the results of an inacid together with about 8 dition exists between the lactic acid, lactide, and water. Temperature vestigation of the reaction per cent of anhydrides. has been shown to haoe no appreciable effect upon the equilibrium. in which the conditions of The acid is usually sold on The anhydride is always present in relatively small amounts. There equilibrium have been dethe basis of its free acid conare indications that it is present in the colloidal state. termined. tent. as defined below, and in its manufacture the effort is made, therefore, to avoid formation of anhydrides as far as EXPERIMEKTAL possible. Considerable losses from this cause invariably Preliminary experiments soon demonstrated the existence of occur, however, and the conditions governing them seem somewhat obscure. The present study grew out of some a condition of equilibrium, for upon heating lactic acid solutions of different concentrations in sealed tubes a t 140" to considerations of the technical production of lactic acid. Several substances are known which may be derived from 150°C. the concentration of free acid came rapidly to a lactic acid by loss of water. The two to which most atten- definite value, which was unaltered by further heating. Attion has been paid are the anhydride (I),and the lactide (11), tempts were then made to devise a method of analysis of the as they are apparently the only ones normally occurring in solutions by which the concentrations of the four constituents, lactic acid, anhydride, lactide, and water, could be lactic acid of any concentration. determined. This proved to be a very difficult problem. ICHICHOHC=O CHgCH-C=O The accepted method of analysis of commercial lactic acid6 I I 1 consists in (1) titrating a sample with 0.1 N sodium hydroxide, 0 0 0 using phenolphthalein as indicator, ( 2 ) adding an excess of I 1 I 0 C-CH-CHI CHICHCOOH standard alkali and heating on a steam bath for 15 minutes, (1) (11) (3) adding a slight excess of 0.1 N sulfuric acid and titrating The literature contains many conflicting statements con- the excess of acid. The first titration neutralizes the lactic cerning these substances. Wislicenus2 states that the an- acid present as such and also the free acid hydrogen of the hydride is obtained by heating lactic acid in a current of anhydride (if present). The alkali used up in thd second dry air at 130' C. and describes it as a light yellow, amorphous titration corresponds to hydrolysis of the anhydride salt and body, only slightly soluble in water but very soluble in alcohol of lactide (if present). From these two titrations the amount and ether. On the other hand, Monin3 says that the anhy- of all four constituents cannot be obtained. Ordinarily, dride does not exist in commercial acid, and is purely a the alkali consumed in the first titration is said to measure the synthetic product. Neither gives experimental data upon free acid, and lactic acid is sold on this basis. If the amounts which his opinion is based. The lactide is described as a of lactic acid, anhydride, and lactide in a given sample are white, crystalline body, of definite melting and boiling point^,^ represented by x, y, and z, respectively, and if a and b reprealmost insoluble in alcohol or ether. Occasionally the two sent the number of cubic centimeters of 0.1 N sodium hysubstances are confused. Statements vary as to the condi- droxide consumed in the first and second titrations, respections under which they are formed from lactic acid. Pure tively, we have the following relations [1.0 cc. of 0.1 N (100 per cent) lactic acid has been shown to remain un- sodium hydroxide is equivalent to 0.009 gram lactic acid, changed and crystalline for months a t temperatures below 0.0162 gram anhydride, 0.0072 gram (hydrolyzed) lactide]: 15" C.4 Partial conversion to anhydrides takes place when Y a =.- X the acid is allowed to stand a t a higher temperature. The 0,009 + o m 2 commercial acid is said to form anhydride when heated to 130'-140' C., and lactide a t 200" C. Both anhydride and lactide are said to be formed when commercial acid is allowed to stand over sulfuric acid a t room temperature.2 I n nearly If we assume either y or x to be zero, as is usually done, or all cases partial formation of anhydrides occurs when lactic determine either x, y, or x separately, these two equations suffice to permit the calculation of the concentration of each acid solutions are concentrated by evaporation. Hydrolysis of both anhydride and lactide with formation constituent. METHODOF ANALYsIs-After searching the literature no of lactic acid occurs when either substance is heated with previous work on methods of distinguishing between the water or with alkali and acid. All the statements outlined above seem to indicate that the anhydride and lactide was found to be available. By using reactionh by which anhydride and lactide are formed are re- different indicators, titrations by the electrical conductivity versible, and that a condition of equilibrium may exist be- method, the formation of metallic salts supposed to be intween the three substances and water. TiTislicenus2suggested soluble in various organic solvents, and the distribution of the constituents between two immiscible solvents, unsuccessReceived March 14, 1924. ful efforts were made to determine by a separate method
C
1
Ann , 164, 181 (1872); 167,302 (1873). Rev. gen. mat. color., 14,279 (1910). 4 Krafft and Dyes, B e y . , 28, 2589 (1895).
2
8
6 Allen's "Commercial Organic Analysis;" J . Am. Leafher Chem. Assoc., 10 (1915).
I N D UXTRIAL A N D ENGINEERING CHEMI8TRY
'824
some one of the four Constituents. After experimenting with the acetates and carbonates of several metals, however, it was found that zinc carbonate reacted readily with lactic acid in alcoholic solution, and apparently with the anhydride, forming insoluble salts and leaving the lactide in solution. Evidence of the precipitation of the anhydride salt was given by the fact that an examination of the precipitate showed the presence of material using up alkali on hydrolysis. (The zinc present was first removed by the method given below.) This reaction was adapted to the quantitative determination as follows: (1) The sample for analysis was dissolved in absolute alcohol and a portion corresponding to about 1 gram of acid was a t once titrated with 0.3 N sodium hydrox?de, using phenolphthalein as indicator (Titratiop a). An excess of alkali was then added, the solution heated on the steam bath for 15 minutes, and the excess alkali determined by titration with standard acid (alkali used, Titration b). (2) A second portion of the alcoholic solution, equal to that taken above, was then treated with 2 grams of zinc carbonate and was warmed, with stirring, until flocculent zinc lactate began to separate (usually about 5 minutes). The precipitate was filtered off under suction, and washed with absolute alcohol. The filtrate was evaporated to about 10 cc. and again filtered if further precipitation occurred. Twenty cubic centimeters of water were added, and the solution was heated with an excess of standard alkali to hydrolyze the lactide. An excess of standard acid was then added, together with a little potassium ferrocyanide to precipitate zinc, (see note 2, below), and the excess of acid was determined by titration with standard alkali (Titration c ) . The following notes explain the procedure: Titrations a and b establish the relations given in Equations 1 and 2. The net alkali consumed in Titration c measures the lactide present. From the three titrations the calculation of the concentrations of all the constituents of the system is evidently possible. (1) Treatment of the alcoholic solution with zinc carbonate, as indicated above, gives a filtrate that is neutral, after the small amount of zinc in solution has been precipitated as ferrocyanide, showing that both the free acid and anhydride have been neutralized. Since the lactide i s quite soluble in alcohol, it must all be present in the filtrate, unless i t has been hydrolyzed by the zinc carbonate. Such hydrolysis can hardly have occurred to any appreciable extent, since duplicate determinations which had been heated for varying lengths of time gave closely agreeing results, which could hardly be expected if partial hydrolysis were occurring. Determinations allowed to stand a t room temperature for 30 minutes also gave closely agreeing results, which would seem t o indicate the absence of hydrolysis. (2) A small amount of zinc remains in solution after the zinc carbonate treatment. Most of this separates out on evaporation of the solution, The zinc is apparently present as lactate, as the precipitate is found to be neutral and not to use up alkali on heating with an excess of sodium hydroxide and then titrating back with acid. I t also separates almost completely on evaporation of the solution in the cold by a stream of dry air, without causing the remaining solution to become acid, as would he the case if it were present as the anhydride salt, which might be decomposing as follows: (CHrCr
