Analysis of Lactic Acid-Lactate Ester Systems - ACS Publications

May 1, 2002 - The lactic acid and volatile fatty acid contents of five locally made type cheeses. A. J. G. Barnett , G. A. Tawab. Journal of Dairy Res...
0 downloads 0 Views 597KB Size
V O L U M E 2 2 , NO. 4, A P R I L 1 9 5 0

545

close to the limit of experimental error of the method. The over:ill evaporation times of the samples were not appreciably shortened nor was the amount of curve deflection due t o moisture accumulation much different from t,hat noted in tests carried out with untreated cones. The results from tests carried out with glass cloth (Fiberglass) rones show that solvents less volatile than methyl isobutyl ketone cannot be tested on thismat,erial, owing to the loss of droplets of sample as a result of drainage to the lower edge of the cone. However, the plotted data, from tests of the more volatile samples ( a port’ion of which is shown in Figures 10 and ll),closely approximated those from the method of Rent and Wik and in every case the shape of the evaporation curve was very similar to that obtained from tests on the same solvent using filter paper cones. This latter point is of particular significance in that it indicates that the chemical nature of the filter papcr does not significantly effect therelease of solvent. Theapparent increase in rate of evaporation evidenced by the use of glass cloth may be logically attributed t o its loose open texture, which does not mechanically wtard the release of solvent to as great an extent as does the relatively dense close-textured filter paper. P O S S I B L E MODIFlCATlOUS AYD IMPROVEMESTS

Since the Evaporomcxtcbr \V:IS developed to the extent dcscribed xbovc and a considerable :imouiit of experimental data obtained, several problems relative to specific uses of solvents over a wide range of temperature and Iiumidity conditions have arisen that cannot be studied with a bulk evaporation method but may be investigated with the Evaporonieter after minor modification of the instrument and improvement of the design of some of the essential elements. These altcrhons, primarily to the instrument cabinet, include the installat.ion of heating elements, proper thermal insulation, and modifiration of the air-inlet port to permit atjachnwnt to air-conditioning equipment,.

The large number of tests conducted to date by various operators show that the free draining type of manually controlled buret does not introduce serious errors due to improper drainage or to uneven distribution of sample on the cone. In working with relatively viscous resin solutions this source of error may be significant, but it may be considerably reduced by the use of a modified mercury reservoir type of buret. Automatic operation of the buret, through connection to the magnet rotating motor, may also prove feasible. ICKSOWLEDGMENT

The authors wish to express their appreciation of the valuable suggestions and technical assistance rendered by their colleagues, C. H. Klute, J. F. Fidiam, and C. J. Penther, as well as the helpful advice of other members of the staff of Shell Development Company. LITERATURE CITED

(1) Bent, P. A., and Wik, S. N., I n d . Eng. Chem., 28, 312 (1936). (2) Doolittle, A. K., Ibid., 27, 1169 (1935). (3) Gardner, H. A., and Sward, G. G., “Physical and Chemic:

Examination of Paints, Varnishes, Lacquers, and Colors, 10th ed., p. 460, Bethesda, Md., Henry A. Gardner Laboratory, 1946. (4) Hart, L. P., Sci. Sect., Educational Bur., Am. Paint Varnish Mfgrs. Assoc., Circ. 360 (1930). (5) Heen, de, J . chim. phys., 11, 205 (1913). (6) Hofmann, H. E., I d . Eng. Chem., 24, 135 (1932). (7) Lowell, J. H., ISD. ENG.CHEM.,A x . 4 ~ ED., . 7, 278 (1935). (8) Rubek, D. D., and Dahl, G. W., Ibid., 6, 421 (1934). (9) Wetlaufer, I,. .4.,and Giegor, J. B., Ibid., 7, 29&3 (1935). (10) Wilson, 31. )I., niitl IVorster, F. J., I n d . Eng. Chern., 21, 592 (1929). RECEIVED April 21, 1949. Presented before the Division of Paint, Varnish, CHEMICAL and Plastics Chemistry a t the 115th Meeting of the AMERICAX SOCIETY,i‘an Franci-ro, Calif.

Analysis of lactic Acid-lactate Ester Systems mtourw A N D KEYYETII i.KOBE L’nizersity of Texas, Austin, Texas

RALPH A.

4 method is presented for determining the composition of reaction mixtures of lactic acid with methanol, both during the reaction and at equilibrium. This is achieved by combining a determination of the free carboxyl groups by titration and a calculation of the amount of hydrolysis of the polylactic acids based on a curve showing hydrolysis ratios as a function of temperature and time. These hydrolysis ratio curve* were determined in a separate series of experiments on aqiieotis lactic acid.

T

HE ever-increasing importance of lactic acid and its deriwtives w industrial chemicals has created a need for data on lactic acid-lactate systems. Fisher (6) has pointed to the versatility of the acid as a low-cost chemical intermediate because of its two reactive functional groups. To date exact work in this field has been made difficult by the lack of suitable ailslytical methods for reaction systems of lactic acid, lactic esterq. :dcohols, and n n t r r . ANALYSIS OF THE PROBLEM

Although the constitution of lactic acid has been a controversial subject for many years, recent work (1, 3, 4, 8) indicates that the monomeric acid exists in equilibrium with polymeric condensation forms of polylactic acids. The extent of the 1 Present

address, University of Louisville, Louisville, Ky.

polymerization or iriternal esterification depends on the concentration of the acid. Thus, in a mixture of lactic acid and an alkyl lactate, the monomeric form and polylactic acids are present in addition to the alkyl lactate, which in this work was methyl lactate. In analyzing these systems, speed is as essential as accuracy in order to prevent a shift of equilibrium concentrations. The determination of free carboxyl groups by titration, followed by determination of free hydroxyl groups, was eliminated for want of a rapid, accurate method for determining hydroxyl groups. Also eliminated were gravimetric methods based on insoluble lactate salts, because the solubility of all these salts was too great for quantitative results. The Karl Fischer method was tried without sucress as a means of determining the amount of n.ater present. Physical methods were useless because of the large numbei

ANALYTICAL CHEMISTRY

546

of reactants and products to be resolved. However, these methods are of value in resolving binary mixtures and were used for lactic acid-water and methyl lactate-water systems. A method that seemed promising was one in which the free carboxyl groups are titrated followed by the addition of a base which will saponify methyl lactate but not the polylactic acids. Many bases were tried under a wide range of conditions, but none gave satisfactory results, although it was discovered that methyl lactate saponifies more readily than the polylactic acids. Solvent extraction using benzene, dichloroethane, and tetrachloroethane as solvents was attempted but was abandoned because of the failure of the system to form two phases with methanol present, unfavorable distribution coefficients, and slowness of the process. Dietz ( 2 ) successfully used dichloroethane in a simultaneous esterification-extraction process for dilute lactic acid solutions.

A check between the yield of methyl lactate a t equilibrium as determined by the proposed method and by distillation should indicate that this method may be applied to mixtures a t any time, regardless of whether or not equilibrium has been attained. Accordingly, a series of experiments was performed on mixtures a t equilibrium. For the sake of simplicitj-, the results of the distillation of the reaction mixture reported in Run B below are used. To 4.5 grams of this mixture 10 nil. of 0.1 S sodium hydroxide solution were added to neutralize the free lactic acid partially and to provide water for the azeotropic distillation of the methyl lactate. This mixture was then distilled; saponification of the distillate indicated the presence of 37.6% methyl lactate in the equilibrium mixture as compared with 37.91% by the proposed titration method. For any experimental esterification a t equilibrium a distillation gave at least 97.5% of the methyl lactate value determined by titration. EXPERLMEYTA L

v

I

Figure 1.

2

3

4 TIME

5

- HOURS

(I

7

I 0

Change in Ratio of Titratable to Total Lactic Acid with Time 85% lactic acid used

In developing a distillation method it was difficult to prevent further esterification or hydrolysis during the process. The best method found for accomplishing this consisted of partially neutralizing the excess acid with aqueous base followed by a simple distillation. The use of aqueous base not only provided a buffer (sodium lactate) but also added water which permitted an azeotropic distillation of the methyl lactate a t a lower boiling point. Because this azeotrope was only vaguely known ( 2 , 6, 7, 9 ) , numerous fractionations were performed in an effort to locate it more exactly. The boiling point was found to be 99.1" C. a t 752 mm. of mercury and the composition to be about 26.5% methyl lactate by weight It is believed that the vapor liquid equilibrium curve is flat over a wide range a t the water end. Recovery from known mixtures of 5- to 10-ml. volume by the distillation method was 97.5% or more complete. This is also the yield obtained when the same volume of methyl lactate, pure or mixed with water, is distilled. This method of analysis should prove satisfactory a t equilibrium for large quantities and where speed is not essential. The method found most suitable consisted of determining concentration ratios between monomeric and polymeric lactic acid in aqueous solution as a function of time and titratable acidity. This assumes that the polylactic acids that are present hydrolyze in a definite manner, regardless of the means by which the quantity of monomeric Figure 2. lactic acid is reduced.

Methyl lactate, furnished by National Dairy Research Laboratories, was analyzed by saponification with standard sodium hydroxide, by boiling point determination, and by refractive index and density measurement. The purity was 100.0%. Reagent grade 85% lactic acid was assayed by titration and saponification with standard base. Average of five assays was 72.55% titratable acidity and 85.67% total acidity (both reported as lactic acid). Absolute methanol which had a purity of lOO.O%, as determined by boiling point and density measurements, was employed. To study the hydrolysis of polylactic acid, approximately 50-ml. samples of 85% lactic acid were weighed out and chilled in an ice bath. To each sample standardized saturated sodium hydroxide solution was added from a calibrated buret while the sample was agitated vigorously in the ice bath. The amounts of sodium hydroxide added were calculated to divide the region of titratable acidity of the lactic acid into equal parts. For example, for the hydrolysis run a t 100" C., the region of titratable lactic acid acidity involved in the esterification experiment was divided into seven approximately equal divisions, by adding to seven 50-ml. samples of the 85% lactic acid 6.52, 9.76, 13.10, 16.28, 19.60, 22.85, and 26.00 ml. of 18.00 N sodium hydroxide solution, reqpectively. After the sodium hydroxide had been added, the mixture was agitated for 1 minute and approximately 8-ml. samples were pipetted into ampoules. These ampoules were sealed and placed in a constant temperature bath. Samples were run a t temperature? of 25", 40". 60", 80", and

PER C E N T TOTAL AClDlJY - A S L A C T I C ACID

Relation between Titratable, Monomeric, and Total Acidity of Lactic 4cid Solutions

547

V O L U M E 22, NO, 4, A P R I L 1 9 5 0

1

r CH,

1

'

T a b l e 1. Equilibrium \.-allies of T i t r a t a b l e and >lonom e r i c Lactic Acid i n Aqueous Lactic .Acid Solutions (7 Total .Iciditr,

7 Titratable Acidity.

Lactic Acid

as Lactic Acidu

H

+ ( n - 1) H20 e

11

n CH,.CHOH.COOH

(2)

nC Monomeric Acidity,

as Lactic .kcid ( 4 ) 10.0 10.0 10 29.5 29.5 30 48.67 46.67 50 53.87 57.060 64.6( 59.0 70 61,45 70.67 80 72.33 61.0 85 72.33 68.67 YO 54.67 70.5 95 66.0 47.0 100 Results of a Average of literature values and author': investigations. all invrstigations in rather close aicret'nient. 3s

'Io ~--cH-COO-

4. The monomeric lactic acid formed in the preceding steps proceeds with the esterification process as in step 1. 401

~

-~ .I

I

100' C'. .it, definite intervals of time an ampoule was removed, chilled, and broken open. .I small aliquot was removed and weighed. This was then titrated with standard base, using phenolphthalein indicat,or, folloived by the addition of more than sufficient base to neutralize the total acidity. after boiling 5 niinut,es, an excess of standard acid sufficient to render the solution colorless was added and the sample was back-titrat'ed with standard base. From the data obtained, values of the ratio, R , between t.itratable acid at any time, P , and the original total acid, I-Le., before addition of the 18.00 S base-were plot'ted against time (Figure 1).

The equilibrium values between lactic acid and its polyniers w e reported in the literature ( 1 , 4,8 ) . I n order to correlate tlie data from the several sources wit'li the present work, 85% lactic acid was diluted with different amounts of distilled water and allowed to stand a t room temperature until equilibrium was estahlished between titratable and total acidity as determined by tit,ration and saponification. Brcause in previous deterniiriations of t,his equilibrium some variation existed among the values determined by the various workers, it was considered n-orth while to present a curve based on the average of all the determinations including this current work. The data det,ermined in this work are given in Table I and Figure 2, which show monomeric lactic acid present in different strengths of solution a' presented by Filachione and Fisher (4). These data map be used in the following manner:

Run A. In following the course of the esterification of lactic acid arid methanol, ampoules containing the reaction mixture were removed from a constant temperature bath a t intervals, chilled, and opened. ..\liquets of approximately 2 ml. were pipetted out, weighed, and titrated to the neutral red end point with standard base. Sufficient sodium hydroxide to provide an excess over that needed to neutralize the total acid present was added and the mixture was boiled for 5 minutes in a flask fitted with an air condenser. Phenolphthalein was added, followed by an excess of st,andard hydrochloric acid over that needed to neutralize the solut,ion, and the mixture was then back-titrated with btandard base. This ana1ysi.QIras also performed on a sample of the original mixture. The change in the titratable acidity as the reaction proceeds is shown in Figure 3 for a typical rraction. The mole ratio of inethanol to lactic arid for this reaction was 3.88. The temperature of reaction v a s 100" C. and no catalyst was present. The niechanism by which the esterification takes place is assumed to consist of the following simultaneously occurring steps. 1. The methyl alcohol and monomeric lactic acid begin to combine to form methyl lactate.

('H3CHOH.COOH

+ CHJOH + CH,.CHOH.COOCH,

+ HZO

(1)

2. This shifts the equilibrium existing between monomeric and pol'ymeric lactic acid by removing the monomer and diluting the reaction mixture with the water formed. 3. The polymeric lactic acid begins to hydrolyze to monomer, probably in several steps, in order to re-establish equilibrium between the monomeric and polymeric forms.

-0

I

2

3

,

I

4

5

~

6

TIME-HOURS

Figure 3. R a t e of C h a n g e in T i t r a t a b l e Acidity for Typical Reaction between $lethano1 a n d 855% Lactic Acid a t 100" C. ( R u n 4)

If it were not for the hydrolysis of polymeric lactic acid (step 3), the decrease in titrahble acidity obtained by subtracting the acidity a t time e from the initial acidity would represent the carboxyl groups t,hat have been esterified. However, new carboxyl groups have been produced by this hydrolysis to replace some of those esterified. For each period of time, 0, t,hat the reaction has proceeded a definite rat.io esists DetTTeen the titratable acidity at that time and the original total acidity. Using the hydrolysis curves a t the temperature of the experiment (Figure 1)' this ratio and time are located on a curve (or as an interpolated point between curves). The curve can then be followed back to zero time and the ratio of titratable to total acidity found at zero time. The ratio a t zero t h e minus the ratio a t time 6 gives a value which is multiplied by t.he original total acidity of the reaction mixture to obtain the increase in titratable acidity due to the hydrolysis of polymeric lactic acid (step 3 ) . Sample calculations illustrate the manner in which these calculations are made. Sample Calculation. T o find the concentration of methyl lactate at the end of 0.5 hour in Run Aka t 100" C., mole ratio of methanol to lactic acid 3.88, no catalyst. Original titratable acidity = 37.00 ml. of 0.1 N base per gram of solution. Total acidity = 43.66 ml. of 0.1 N base per gram of solution = 4.366 me. per gram of solution. On Figure 3 read titratable acidity at the end of 0.5 hour a t 28.53 ml. of 0.1 N base per gram of solution = 2.853 me. per gram of solution. Disappearance of titratable acidity 37.00 - 28.53 = 8.47 ml. of 0.1 N per gram of solution = 0.847 me. per gram of solution.

A N A L Y T I C A L CHEMISTRY

548

4.366 - 0.709 = 3.657 me. per giani of solution. This must also equal the water formed, Total water a t equilibrium = 3.659 3.657 = 7.316 me. per gram of solution. Strength of acid solution in water a t eauilibrium " (0.709 x 90) (loo) = 32.60%. (7.316 X 18) (0.709 X Sq) From Figure 2 the titratable acidity 111 equilibrium with this strength acid is 32.1%. Therefore some polymer exists a t equilibrium, Total aciditv a t eauilibrium 0.709 (32.6/32.1) . . , = 0.720 me. per gram of solution. Methyl lactate formed = 4.366 - 0.720 = 3.646 me. per gram ( i f solution. Water formed = 3.646 me. per gram of solution. Total water a t equilibrium = 3.659 3.646 = 7.305 me. per gram of solution. The strength of lactic acid a t equilibrium is now calculated to IJP 33%. Because 33/32.5 is not sufficiently diflerent from 32.6/' 32.1, no further refinement will be made. Methanol a t the start (4.366) (3.88) = 16.92 me. per gram

Ratio of titratable acidity (0.5 hour) to original total acidity, P / I = 28.53/43.66 = 0.653. Locating this ratio on the hydrolysis chart, Figure 1, a t a time of 0.5 hour and following the curves back to zero time, a ratio of 0.621 is obtained at zero time. Polylactic acids reverting to monomeric acid and forming 0.621)(43.66) = 1.40 ml. of 0.1 N base methyl lactate (0.653 per gram of solution = 0.140 me. per gram of solution. Total methyl lactate formed = 0.847 0.140 = 0.987 me. per gram of solution.

+

-

+

+

Run B. To obtain the equilibrium composition for the reaction, the equilibrium curve (Figure 2) for aqueous lactic acid and its polymers is used. I t is convenient to assume for stoichiometric purposes that the lactic acid used consists only of monomeric lactic acid and water. Thus, in the sample calculation given below the solution is considered to consist of 85.67% monomeric lactic acid and 14.33% water. This value for water is used for stoichiometric purposes and it saves involved calculations of the amount of water needed to hydrolyze the polymeric lactic acid. Such assumption causes no appreciable error because the amount of polymeric form a t equilibrium is small. The initial total acidity minus the titratable acidity a t equilibrium is considered to be equivalent to the methyl lactate formed. This is also the amount of water formed in the reaction (Equation 1). The total water a t equilibrium is approximately the sum of the water introduced with the lactic acid plus that formed in the reaction. In order to use Figure 2, the lactic acid concentration must be calculated as though no methanol and methyl lactate were present. Thus the percentage of total lactic acid in this hypothetical solution is the weight of lactic acid times 100 divided by the weights of lactic acid and water. This value is used as the abscissa on Figure 2 and the corresponding percentage of titratable acidity is read from the ordinate scale. The ratio of these values of percentage total acidity t o titratable acidity is multiplied by the titratable acidity a t equilibrium to obtain an equilibrium value of total lactic acid which is more nearly correct than the value originally assumed. Nom one of the original approximations can be corrected. The initial total acidity minus the total lactic acid just calculated is a more nearly correct value for the methyl lactate or water formed. Using this new value for water formed, the above calculations may be repeated if the change in the ratio of percentage total acidity to titratable acidity is sufficient to warrant another calculation. The methanol remaining in the equilibrium mixture is equal to the methanol in the original mixture minus the methyl lactate formed. This now gives values for all components in the equilibrium mixture. Sample Calculation. T o find equilibrium composition in Run A at 100' C., mole ratio of methanol to lactic acid 3.88, no catalyst. Total acidity a t the start = 43.66 ml. of 0.1 N base per gram of solution = 4.366 me. per gram solution. (4.366) (14.33/18) Water at start of reaction (85.67/90) 3.659 me. per gram of solution, where 85.67 is the per cent total lactic acid in the lactic acid used in the experiments. Molecular weights of lactic acid and water are 90 and 18, respectively. Titratable acidity a t equilibrium 7.09 ml. of 0.1 N base Der a s m of solution = 0.709 me. Assuming &-st that no polymer is present, 0.709 represents Ithe total milliequivalent lactic acid at equilibrium. Methyl lactate formed is

+

Methanol a t equilibrium i6.92- -

of solution.

Equilibrium Composition Weight % 6.48 37.91 42.47 13.14 100.00

hIe./G. Solution 0.720 3.646 13.274 7.305

Lactic acid Methyl lactate Methanol \Tater

1I

-

l

1 //

i

i

l

I

METHYL LACTAX-HzO

I IO

I

20

1

30

I 40

50

CONCP(TRATI0N-

Figure 4.

6D

70

80

90.

Id0

WElGHT PERCENT

Refractive Indexes of Lactic Acid-Water a n d Methyl Lactate-Water Mixtures a t 25" C.

V O L U M E 22, NO. 4, A P R I L 1 9 5 0

549

Table 11. Refractive Indexes of Methyl Lactate-Wi1tc.r and Lactic Acid-Water Mixtures Refractive Index a t 25‘ ( ’ Methyl Lactate, Wt. % 0.00 15.75 33.30 50.00 67.00 71.00 79.00 92.75 95.50

1.33260 1,34710 1.36258 1.37748 1.39156 1.39467 1.40043 1.40840 1.40981 1.41205

100.00

Lactio Acid 1.33260 1.34182 1.35014 1.36416 1.37753 1.38153 1.38727 1.39324 1.39939 1.40401 1.40745 1,40863 1.41217 1.41417 1.41772 1.42132 1.42639 1.43293 1,43367

0.00 9.18 16.95 30.60 43.76 47.21 52.56 58.10 64.07 68.44 72.00 72.93 76.34 78.42 81.00 85.67 90.35 97.90 98.27

Refractive indexes of lactic acitl -water mixtures and methyl lactate-water mixtures were obtained using a Bausch & Lomb precision refractometer. These data are presented in Table I1 and Figure 4. Although no difficulty was experienced in analyzing methyl lactate-water mixtures, the lactic acid-water system values iyere difficult to reproduce because of the presence of a series of converging parallel lines rather than sharp demarcation between light and dark areas. DISCUSS1O.Y

The procedure given here is based on certain assumptions for which no proof is possible. In view of the fact that a t equilibrium the results of this method can be checked by distillation, it seems logical to assume that the results are accurate a t any stage of the reaction. The procedure was developed after a careful investigation of the factors influencing its accuracy. To establish that no saponification of polylactic acids occurs during titration, a comparison W R ? made of results obtained by

use of both aqueous and alcoholic base with those reported i n the literature for various strengths of the acid in the range 0 to 85% lactic acid. I n this range the experimental results substantially checked those of the literature. Because methyl lactate hydrolyzes readily, it wm necessary to determine how the titration would affect this ester. I t was found that if the dilute base was added in a slow steady stream with excellent agitation no saponification of the ester occurred until all the free acid had been neutralized. Neutral red was chosen as indicator because its color change a t p H 7 minimized the danger of hydrolysis. The technique of titration increased in importance as equilibrium in the esterification mixture was approached. The determination of the breakdown of the polymeric acid is a more complex problem I t was decided to assume that the mechanism is that previously discussed. To evaluate this hydrolysis it was decided to use a substance that would remove the monomeric acid and a t the same time add water. Lactic acid to which calcium lactate had been added waa heated to study the possibility of a “salt effect,” After determining that the salt effect did not influence the hydrolysis of the polylactic acids, the use of basic materials was studied. Solid bases proved too insoluble in the mixture and their distribution and heat output were too difficult to control. The use of alcoholic sodium hydroxide resulted in some esterification between the alcohol and the acid a t high temperatures. Finally, saturated aqueous sodium hydroxide solutions were utilized with satisfactory results. LITERATURE CITED

(1) Bezzi, S.,Riccoboni, L., and Sullam, C., Mem. eccad. Italia, Classe SC~.fis. mt. nut., 8,127-213 (1937). (2) Dietz, A,, Degering, E., and Schopmeyer, H., I d . Eng. Chem.. 39,82-5 (1947). (3) Eder, R.,and Kutter, F., Helv. Chim. Acta, 9,355-64 (1926). (4) Filachione, E., and Fisher, C., I d . Eng. Chem., 36,223-8 (1944). (5) Fisher, C., and Filachione, E., U. S. Dept. Agr., Bur. Agr. Ind. Chem., Bull. AIC-178 (May 1948). (6) Rehberg, C.,Faucette, W., and Fisher, C., I d . Eng. Chem., 36, 469-75 (1944). (7) Schopmeyer, H., and Arnold, C., U. S. Patent 2,350,370(June 0, 1944). (8) Watson, P.. I n d . Eng. Chem., 32,399-401 (1940). (9) Weisberg, S.,and Stimpson, E., U. S. Patent 2,290,926(July 28. 1942). RECEIVEDFebruary 10, 1949.

Determination of Sulfur Trioxide in Chlorosulfonic Acid Thermometric Methods WILLIAM SEAMAN, J. T. WOODS, AND H. N. BANK American Cyanamid Company, Calco Chemical Division, Bound Brook, A’. J .

B

ECAUSE the concentration of free sulfur trioxide in chlorosulfonic acid is of importance in some of the uses of this acid, it became necessary to develop a more precise method for its determination than that generally used, which involves a total acid value and a chloride value (2, 3, 6, 6). This method is not sufficientlyprecise because it is necessary to determine the sulfur trioxide indirectly, and the factors involved are such that an error of 0.1% each in the acid value and the chloride value causes an error of 1.0% sulfur trioxide. Considerable heat is evolved when hydrogen chloride reacts with sulfur trioxide. According to Ogier (4),the heat of formation of chlorosulfonic acid from sulfur trioxide (solid) and hgtlrogon chloride (gaseous) is 14 1 kg.-cal. and its specific heat is

0.282 calorie per gram. The method finally developed, involving measurement of the temperature rise when gaseous hydrogen chloride is passed into chlorosulfonic acid containing sulfur trioxide, is not only sufficiently precise for the determination of low concentrations of sulfur trioxide, but gives more precise values than the old method for samples containing high concentrations of sulfur trioxide. METHOD O F ANALYSIS

Apparatus. All parts of the apparatus which come into contact with the sample must be thoroughly dry. In Figure 1 are shown a cylinder or generator of dry hydrogen chloride; a gasometer for measuring t,he hydrogen chloride consisting of a