ANALYTICAL CHEMISTRY
1036 powder mixed with the lime as suggested by Bennett et ai. (1) gave a false positive test with triphenyl phosphate. Copper powder in the calcium oxide as suggested by Emich and Schneider (S,6 ) offers no advantage, since compounds containing nitrogen linked to oxygen give positive tests \T ith lime alone. LITERATURE CITED
(1) Bennett, E. L., Gould, C. K., Jr., Swift, E. H., and Siemann, Carl, ANAL.CHEM.,19, 1035 (1947). (2) Cheronis, X . D., and Entrikiri, J. B., “Semimicro Qualitative Organic Analysis,” Sew Tork, Thomas Y . Crowell Co., 1947.
(3) Emich, Friedrich, and Schneider, Frank, “Microchemical Laboratory hlanual,” New York, John Wiley & Sons, 1932. (4) Feigl, Fritz, “Qualitative Analysis by Spot Tests,” Xew York, Elscvier Publishing Co., 1946. ( 5 ) Fletcher, M , H., White, C. E., and Sheftel, M. S., IND.ENG. CHEM.,ANAL.ED.,18, 179 (1946). (6) Schneider, Frank, “Qualitative Organic Microanalysis,” Xew Tork, John Wiley 8; Sons, 1946. (7) Velluz, L., and Pesez, >I., Ann. p h a r m . franc., 4, 10 (1946). (8) Ihid., p. 12. RECEIVED September 25, 19.50.
Colorimetric Assay of Diacyl Amides J. B. POLYA AND P. L. TARDEW Chemistry Department, University of Tasmania, Hobart, Australia IPRIANN and Tutt’le (5) have developed a procedure for assaying acyl phosphates by converting the organic acyl radicals t o hydroxamic acids, which give intense coloration with ferric chloride. Because purrly organic anhydrides or acyl chlorides are not likely t’o occur under biological conditions, and esters or amides do not react under the experimental conditions, the method is claimed t o be specific for acyl phosphates in biological materials. It is possible, however, that diacylamides of the general formula RCO.KR‘.COR” may occur in biological systems. The Lipmann-Tuttle assay was applied t o a number of representative compounds of this class. Dibenzamide ( 1), formyl benzamide ( 4 ) ,diacet’amide (IO),and other diacyl amides with R’ = H (8) were prepared by previously described methods. Other diacyl amides were prepared by the acylation of the appro. priate amines or amides with anhydrides in the presence of acyl halides or thionyl chloride in catalytic amounts ( 3 ) . Ot,her reagents were commercial products of analytical quality. The reagent solutions were those recommended by Lipmann and Tuttle. They were prepared fresh and used in the quantities and order as suggested by the originators of the method. T h e colors were compared with a Hilger Spekker absorptiometer using 1-cm. cells, Ilford 608 gelatin filters, and a mixture of the reagents only as reference. This gave convenient readings for the assay of succinic anhydride or acet’yl phosphate in concentrations ranging from 0.025 to 0.100 M with an error of +4%.
This procedure was not suitable for diacyl amides, because some of them were not sufficiently soluble and even diacetamide gave low density readings with a n error of = t 1 8 ~ o .Table I refers t o average values from four measurements each, with probable errors indicated above.
Table I. Optical Density Molar Concentration 0.100 0.075 0.050 0,025
Optical Density Succinic anhydride Diacetamide 1.18 0.13 0.11 0.89 0.59 0.30
0.09 0.06
Better results were obtained by increasing the alkalinity of the reaction mixture and limiting the reaction time t o 2 minutes. ’ I n this modified procedure 1 ml. each of 4 Jf hydroxylamine hydrochloride solution and 3.5 .IT sodium hydroxide are mixed and 1.8 ml. of the rrcommended acetate buffer are added. The solution to be analyzed (2 ml.) followed by another 0.2 ml. of the 3.5 M sodium hydroxide is added in this order and the mixture is allowed t o stand for 2 minutes a t room temperature (18’ to 20”). The subsequent operations are those recommended by Lipmann and Tuttle. Ilford 604 gelatin filters were used in the modified procedure as they were found to give better results than Ilford 608.
Dibenzamide is not sufficiently soluble in water and was used dissolved in 1% sodium hydroxide. Other relatively insoluble diacyl amides like acetyl and propionyl benzamide hydrolyze too rapidly in alkaline solutions. I n such cases the assays had to be restricted to lower concentrations in water only. Although Smethyldiacetamide is very soluble in water, such’solutions hydrolyze rapidly and the assay must be carried out without delay. against +4% in other Even then the assay has a n error of =ti% cases, n-ith the exception of acetyl chloroaeetamide, which is assayed v i t h an error of =t22’%. The method is not suitable for the assay of h’-formyl amides, as the color fades within a few seconds. KH20H+ The reaction therefore proceeds as PhCONHCHO PhCONH2 HC( .O) SHOH, in agreement with previous theories (8). Hydantoin, S-methyl-S’-acetvlurea, surcinimide, and acetamidine gave little or no color. hcetamide in a 5% (0.847M) solution gave a n intensity of 0 27 * 0.02, corresponding to that given by approuimately 0.0025 Jf solutions of diacyl amides The S-bromo and S-bromoniagnesium derivatives of diacetamide (6) gave the color test which, however, was not suitable for quantitative purposes, owing to the instability of these compounds under the experimental conditions. The same applies to a lesser extent to acetyl chloroacetanlide, but the colorimetric assay of some halogenated diacyl amides could be supplemented by fluorometric or spectrographic methods ( 7 , Q ) . Esters interfere in the presence of excess alkali ( I l ) , although this is not the case in the original method of Lipmann and Tuttle. However, the removal of esters from solutions containing acyl phosphates is difficult or impossible, whereas the separation of esters from most diacyl amides is comparatively simple. Average results of six assays (four assavs only for acetyl chloroacetamide) are shown in Table 11. The errom indicated above refer to concentrations from 0.0025 to 0.0100 M . At 0.0010 M concentration the errors are estimated a t *9 to *loyo. The curves of density versus concentration may be regarded as linear within the stated range of errors and may be used for the rapid determination of small amounts of diacyl amides. The assays may be duplicated with a n error of *2% only if succinic anhydride is used in the reference cell as in the original LipmannTuttle procedure. Diacetamide is unsatisfactory as a reference in the original but useful in the modified method. Analytically pure disilver acetyl phosphamide has been prepared by dehydrating and acylating monoammonium phosphate with acetic anhydride in excess in the presence of thionyl or acetyl chloride and precipitating from a n aqueous solution with silver nitrate ( 2 ) . Like benzoyl phosphamide ( l a ) , acetyl phosphamide appears to be relatively stable jn aqueous solution. Removal of the silver by thioacetamide resulted in solutions from Khich the original silver compound salt could be regenerated with a loss of 71 to 86%. Such silver-free solutions could not be used for the assay because of a sloa-ly fading green color due to residual thioacetamide in the solution. Fresh extracts of disilver acetyl
+
+
V O L U M E 23,
NO. 7, J U L Y
1951
1037 ACKNOWLEDGMENT
Table 11.
Optical Density Optical Density
Compound Succinic anhydride Diacetamide .V-Methyldiacetamide Diacetanilide Acetyl propionamide Dipropionamide -4oetyl benzamide Propionyl b e n z a m d e Dibenoamide Acetyl chloroaoetamide
0 0100 .M
1.18 1.33 1 07 1 33 1.29 1.39
Insol. Insol. 1.75 1.10
0 0075
M
0.89 0.99 0.73 0.97 0.95 1.02
Insol. Insol.
1.25 0.65
0 0050 M
0 0025 -M.
0.0010 M
0.59 0.65 0.50 0.65 0.64 0.69 0 69 0 72 0.82 0 52
0.30 0.33 0.26 0.30 0.32 0.34 0.36 0.35 0.41 0.27
0.14
phosphamide with a slight excess of 10% sodium chloride solution did not give the test by either the original or the modified Lipmann-Tuttle procedure. A faint color appeared occasionally after standing for a few hours. This suggests that the cleavage of acetyl phosphamide by hydroxylamine yields acetamide and phosphohydroxylamine, probably followed by the rapid hydrolysis of the latter. The original and modified test has been applied to cancer, diabetes, pregnancy, and normal human serum and t o albumin, globulin, euglobulin, and pseudoglobulin fractions of such sera. As a rule the test was negative, but a few fresh, defibrinated sera from diabetic patients gave a faint purple color which was retained in the centrifuged precipitated proteins.
0.15 0.12 01.4
0.15 0.14 0.16 0.16 0.19
...
The authors wish t o acknowledge a grant from E. J. Hallstrom and thank G. C. Bratt and Peter Dunn for pure preparations of some of the compounds investigated. LITERATURE CITED
(1) Barth, L., and Senhofer, C., Ber., 9, 978 (1876).
(2) Bratt, G. C., and Polya. J. B., material prepared for publication, cf. (6). ( 3 ) D u m , P., and Polya, J. B., unpublished material. (4) Einhorn, A,, Ann., 343, 223 (1905). (8)Lipmann, L., and Tuttle, L. C., J . B i d . Chem., 159, 21 (1945). (6) Polya, J. B., DSc. thesis, University of Twmania, 1951. ( 7 ) Polya, J. B., aiid Spotswood, T. AI., J . Australian Chem. Inst., 15,445 (1948).
( 5 ) Polya, J. B.,and Spotswood, T. >I,, Rec. Irav. chim., 67, 927 (1948). (9) Ibid., 68,573 (1949). (10) Polya, J. B., and Tardew, P. L., J . Chem. Soc., 1948, 1081. (11) Thompson, A. R., ~41i.siruZiun J . Sei. Research, A3 ( I ) , 128 (1950). (12) Titherley, A. IT., and K o r r d l , E., J . Ciiern. Soc., 95, 1143 (1909).
R E C E I V EJDu n e 7, 1930.
Determination of Total Chlorine in Inorganic Salts WILLIAM A. JAMES
AND RICHARD E. BELSER’ Midwest Research Institute, Kansas City, LMo.
MANY of the methods ( 5 ) available for the analysis of I Nmixtures containing chloride, chlorate, perchlorate, thiocyanate, and cyanide ions, the total chlorine is determined with some difficulty. I n mixtures containing chloride, thiocyanate, and cyanide ions it is necessary first to liberate the cyanide ions and destroy the thiocyanate ions or determine the amount of rach present before the “chloride chlorine” is obtained. To overcome these difficulties, a method for the direct determination of the total chlorine has been developed. Should the mixture contain chlorate or perchlorate ions, the “chlorate chlorine” and “perchlorate chlorine” would also be determined M ith no additional handling or treatment of the sample. PRINCIPLES OF METHOD
The technique employed is similar to t h a t used by LaForce ef al. (2)in determining the carbon dioxide content of inorganic carbonates. Their method involves fusion of the sample with potassium bisulfate in a micro combustion boat placed in the standard organic carbon-hydrogen train. In the chlorine determination the reaction takes place in a spiral or Friedrich tube used for the determination of chlorine in organic compounds ( 8 ) . In addition, manganese dioxide is added to the fusion mixture to ensure complete liberation of the chlorine. Preising et al. ( 4 ) have shown that heating chlorates and perchlorates with manganese dioxide releases the chlorate chlorine and the perchlorate chlorine. The reaction with potassium perchlorate is
2KCIO4
+ 2H2S01 + M n 0 2 + C ~+ Z hlnSOI + K2S04 + 2 H 2 0 + i o 2
The chlorine gas formed then reacts with a mixture of sodium carbonate and hydrogen peroxide in the spiral of the tube t o form the chloride, which is titrated with standard silver nitrate. I
Present address, Remington Arms, Inc., Independence, M o
PROCEDURE
T h e spiral tube is prepared in the following manner: One milliliter of a saturated solution of sodium carbonate and 4 ml. of 3% hydrogen peroxide are mixed in a 125-ml. Erlenmeyer flask and drawn into the tube until the spiral is covered. The mixture is drained back into the flask, and after the tube is positioned in the furnace, the flask is hung over the exit end. Any material spraying from the tube is retained by the flask. The platinum catalysts are positioned and heated to between 600’ and 700” C. The train is then ready to receive the boat containing the fusion mixture. The sample, weighing 10 to 20 mg., depending upon the amount of chlorine present, is weighed in a semimicro porcelain combuvtion boat. Sixty mg. of potassium bisulfate and 9 mg. of manganese dioxide are added without mixing with the sample. The boat is placed in the train and the oxygen is connected. The flow is regulated to a rate of 5 to 6 ml. per minute. The boat is first heated cautiously to prevent excessive foaming. After the foaming ceases, gradual heating is continued until the full heat of the burner can be applied. After completion of the fusion, which requires approximately 20 minutes, the tube is removed from the furnace and allowed to cool, the oxygen continuing t o pass through the tube. When the tube has reached room temperature, the hoat is removed and the contents of the tube are rinsed into the flask with distilled water. The carbonate solution in the flask is acidified with dilute nitric acid (1 t o 1) and concentrated to 5 or 10 ml. After cooling, a n equal volume of acetone and two drops of 0.1% dichlorofluorescein indicator (0.1 gram in 100 ml. of 70% alcohol) are added (1). T h e p H is then adjusted by addingysmall increments of solid sodium bicarbonate until the indicator is greenish and fluorescent. Acetone is added because i t gives greater dispersion of the silver chloride precipitate and therefore yields a sharper end point. The titration with 0.01 N silver nitrate is conducted in a partially darkened room using a black background. The titration flask is illuminated with a daylight lamp located to the side and slightly t o the rear of the observer. A blank is carried through all steps of the combustion. Because the chloride solution must be a t least 0.005 N in sodium chloride to obtain an end point, it is necessary t o add, accurately
