1318
ANALYTICAL CHEMISTRY
The p-diketone is a good nonspecific spot test reagent capable of detecting from 2 to 5 y of some metal ions in 0.1 ml. of solution. Because of the various colors exhibited with different metal ions, this reagent has found useful applications in this laboratory as a spot test reagent. Specificity is not always desired in a reagent. Considerable interest has been sholvn recently in the chromatographic separation of metals as chelates. Small amounts of metals may be concentrated for chromatography or other uses, after chelation, in two ways. The chelate may be iemoved by filtration, dried, and then dissolved in an organic solvent, or it may be extracted directly into an organic solvent. In either case, the organic extract is used to spot the paper for chromatography. Table I1 gives the Rj values for the paper chromatographic separation of mixtures of the iron(III), copper(II), nickel(II), cobalt( 11), and manganese( 11) chelates of 2-furoyltrifluoroacetone when the composition of the developing solvent v a s varied. 9 total of 1 to 2 y of each metal was present in the mixtures. The method may also be useful for other systems. Investigation of the absorption spectra for several metal chelates revealed the curves shown in Figure 1. It is probable that spectrophotometric procedures, based on the light absorption of the metal chelates in organic solvents, may be devised for the quantitative determination of several metals. A pro-
cedure for the determination of copper appears most promising. Flnorophotometric procediires for some metals may also he possible. 2-Furoyltrifluoroacetone compared favorably with dimethylglyoxinie in the determination of palladium (Table 111). Xiinierical comparison of clieniicnl factors for palladium in the 2furoyltrifluoroacetonate and diniethylglyosiniate, 0.2064 and 0.3167, respectively, favors the new reagent greatly. The effect of the presence of other platinum group metals m s investigated. Platinuni, rhodium, and iridium int,erfered slightly because their hydro123 osides precipitate a t pH 7. A reprecipitation technique or R masking reagent map remove these interferences and make this a useful technique for the determination of palladium in ores or alloys. The reagent is excellent for the gravimetric standardization of stock palladium solutions. LITER.ATURE CITED
(1) Berg, E. W., \IcIiitg-re, R.T., -\SAL. (?HEX. 27, 195 (1955). (2) Carlton, J. K., I h i d . , 22, 1072 (1950). (3) Feigl, F., “Qualitative .Indysis by Spot Tests.” Elsevier, Sen-
Tork, 1947.
(4)
Hillebrand, W. F..Lundell. G. E. F., “.lpplied Inorganic .inaly$is.” p. 292, JTiley, S e w T o r k , 1929.
R E C E I V Efor D review January 3 , 19.56. .Iccepted April ,5, :95C,.
Determination of Carboxylic Acids, Acid Chlorides, and Anhydrides by Chlorine-36-Isotope Dilution Method POUL SORENSEN Central Laboratory, Sadolin & Holmblad, Ltd., Copenhagen, Denmark
The use of chlorine-36 derivatives as a means of determining acj-latable compounds by isotope dilution w-as found to offer so many advantages that the method has been extended to systems characterized by other functional groups. The principle has been applied to the determination of carboxylic acids, acid chlorides, and anhydrides. The compound to be analyzed is quantitatively converted to the p-chloroanilide, which is then determined by an ordinary isotope dilution method, using the chlorine-36-tagged p-chloroanilide of the acid in question.
T
HE principle of a chlorine-36-isotope dilution method has
been presented recently for the determination of acylatable compounds ( I S ) . The compound to be determined is quantitatively converted to a chlorine-containing derivative, which is determined by an ordinary isotope dilution method. The method has the follovhg advantages: The determination IS usually absolutely specific for the acid in question; with one radioactive compound [p-chloro(36)-aniline in this case] different carboxylic acids may be analyzed; the reference compound for the determination is a p-chloroanilide, which is easy to prepare in a pure state; radioactivity measurements are done Kith chlorine-36, and, therefore, good precision is obtainable. Compared with an ordinary isotope dilution analysis, the disadvantage of the method is that the compound to be determined must be quantitatively converted to a p-chloroanilide. T h e quantitative conversion of acid chlorides and anhydrides t o anilides is well known (6, 8-10). By preparing an anilide from a carboxylic acid. the acid is usually first converted to the acid chloride, and lf the acid chloride is nonvolatile, this two-
step reaction may be carried out quantitatively (1 ). -4one-step reaction, by which the caiboqlic acid is treated with a phosphazo compound, has been suggested by Grimmel, Guenther, and Morgan ( 5 ) , and has further heen investigated by GoldSchmidt and coworkers ( 2 - 4 ) . When p-chlorophenylphosphazo-p-chloroanilideis used under the proper reaction conditions, carboxylic acids may be quantitatively converted to p-chloroanilides. p-Chloroanilides are easily purified and have sharp melting points. METHOD
Reagents. ~-CHLOROASILIKE. ‘4technical product xvas distilled under vacuum and a middle fraction was crystallized twice from 60% alcohol; the melting point was 70-71’ C. p-CHLOROPHEh.YLPHOSPH.LZO-p~~HLORO.4~lLlDE.The compound mas prepared according to Grimmel’s directions ( 5 ) . To a thiee-necked flask equipped x i t h an efficient stirrer mere added 130 grams of p-chloroaniline (melting point, 70-71’ C.) and 1000 nil. of toluene (reagent grade, dried over sodium); 100 ml. of toluene was distilled t o remove the last traces of moisture from the flask, The solution was stirred at refluxing temperature and treated dropwise with a solution of 25 grams of phosphorus trichloride in 50 ml. of toluene. The reaction mixture was further heated for 1 hour. The hot reaction mixture y a s filtered to remove p-chloroaniline hydrochloride, then the filtrate was evaporated nearly to dryness a t reduced pressure. The residue was treated with 200 nil. of absolute alcohol a t room temperature and allowed to cool to 0 ” C.: the precipitate was filtered and washed with 50 ml. of alcohol Finally the product was crystallized by dissolving it in 200 nil of hot toluene and adding 400 ml. of alcohol to the clear solution The crystals were cooled overnight at 0’ C , removed hj*filtiation, ITashed with a mixture of alcohol and toluene, then viith petroleum ether, and finally dried at room temperature. The yield was 13 grams. The product decomposes a t about 160” C. The elementary analysis indicates a content of half a mole of alcohol per mole of the monomer.
1319
V O L U M E 28, N O . 8, A U G U S T 1 9 5 6 Analysis of Cl-CsH~-N=P--NH-C~H~-CI. /2C2HjOH Calcd., % Found, P; Carbon 50.8 51.4 Hydrogen 3.91 3.; Nitrogen 9.14 9.3 Chlorine 23.1 23.0 P-CHLOROANILIDES, These compounds were prepared by conventional methods from the acid chloride or anhydride, and were varefully crystallized to a constant melting point. R A D I O A C T I V E Co>rPocsDs. p-Chloro(36,1-aniline. A solution of 1.40 grams of acetanilide in 10 nil. of acetic acid was chlorinated a t room temperature with the chlorine produced from 2.9:3 grams of active silver chloride ( I d ) . The reaction misture n-as diluted with 50 ml. of water, heated to give a clear solution, and crystallized by cooling. The product \$-as recrystallized from 50 nil. of 20% acetic acid and yielded 0.02 gram of acet-p-chloro(36 :inilide with a melting point of 1T8.5-179.0° C. An amount of 0.80 gram of acet-p-chloro(36)-anilide ITas hydrolyzed by 20 nil. of concentrated hydrochloric acid by heating on a vapor bath f u r 1 hours, followed by evaporation to dryness a t reduced pressure. The residue was then dissolved in 10 ml. of water and 5 nil. of 4 S sodium hydroside as added. The solution was cooled to 0" C. and the crystals ivere removed by filtration, washed with icae water, and dried a t 40" C. a t 1 mm. Finally the product was distilled on a cold finger at, 1 mm., the distillation flask being slowly heated in a water bath to 100" C. The distilled p-chloro(86)-aniline (0.50 gram) had a melting point of 70.5-T1.0' C:. The product was prepared with an activity of 1 pc. per mmole. Some chlorine-36 was regenerated as silver chloride. Yield of :icet-p-chloro(3G)-nnilicle on chlorine consumed n-as about 65c;b, p-Chloro(36)-anilides. p-Chloro(36)-aniline was treated with xi excess of acid chloride (in pyridine) or acid anhydride. The crude products tvere crystallized twice from alcohol or from alcohol diluted with water. The melting points differed not more than a ie\y tenths of a degree from that of the purc compound, I n addition to these reagents, the follon-ing are also needed: Tolueiie, reagent grade dried over sodium. Acetone, reagent grade dried over Drierite. Dioxane, reagent grade. Alcohol, reagent grade, 99 to 99.55;. Thionyl chloride, purified. Measurements. DETERMISATIOS OF PL-RITY. The puritroi a p-chloroanilide was determined from the melting point ( l a ) . The molar melting point depression must be determined for. each compound. RADIOACTIVITY MEASUREXESTS. The technique has been bed ( 1 4 ) . d ratio of two activities as determined with a ical error corresponding to a standard deviation of 0.6 t o paration of Active Solution. Dissolye a n amount of the p-chloro(3G)-anilicle corresponding to 0.7 pc. in 100 nil. of diosane. Preparation of Standard Sample. IT-eigh accurately about 200 mg. of the p-chloronnilide (inactive) and 1.3 nil. of active solution. Aidd dioxane t o the mixture until a tained by boiling. Pour into water and recr Procedure I. ACIDS. Keigh accurately estimated to contain 1 to 2 meq. of the acid to be determined. I d d 10 ml. of toluene. Add about 0.4 gram of p-ch1orophen)-1phosphazo-p-chloroanilide and 0.8 gram of p-chloroaniline for each milliequivalent of acid. Reflux for 11 '2 hours. Cool, then add 1.5 nil. of active solution, determining the amount accurately by weighing before and after addition. 1 - o ~add 10 nil. of alcohol and boil until a practically clear solution is obtained. Remove the organic solvents with water r a p o r , add 10 nil. of 4-1hydrochloric acid. cool, and remove the precipitate by filtration. .ift,er washing with water, crystallize the precipitate from absolute or diluted alcohol or from water! until the melting point differs 11y less than l o C. from that of the pure p-chloroanilide. Determine the purity from the melting point. RIeasure the specific activity (activity per unit Ti-eightj of this final sample ratio of the specific activity of the st:rnclurd sample (correct for background arid self-absorption). Procedure 11. .ICIDS (acid chloride iion~olatilej. Keigh accurately in the reaction flask (Figure 1 ) ahout 1 nieq. of the acid t o be determined. *kdd 3 nil. of thion:-l chloride and heat t h e mixture under reflux for half an hour n-ith stopcocks B clo.sed and 1) open. Reniove excess thion>-l chloride into the freezing trap hy distillation first a t water pump-vacuum and then a t 0.1 n i m . for 1 5 minutes. The reaction flask should not be heated during distillation. but, because of the flexible connection a t E, it mal- be shaken gently by hand. Then close st'opcock U and: n-ithout breaking the vacuum, add from the dropping funnel 2 grams of p-chloronniline dissolved in 10 nil. of acetone. Heat gentlj- until a c.le:tr solution appears. .Idd 1.5 ml.of active solu-
tion, determining the amount accurately as in Procedure I. Add water containing 10 ml. of 4-4- hydrochloric acid, crystallize the precipitate, and continue as described by Procedure I. Procedure ACID CHLORIDES AXD ANHYDRIDES. The procedure is essentially the same as Procedure 11, except that t h e first treatment with thionyl chloride is omitted. Calculation.
Iu.
and mg. of compound =
molecular weight of compound X B molecular weight of p-chloroanilide
where
r = A
=
specific activity of final sample specific activity of standard sample
milligrams of active solution added
B = milligrams of p-chloroanilide present after conversion of
compound t o p-chloroanilide milligrams of active solution used in preparing standard sample h = milligra-ms of p-chloroanilide used in preparing standard sample g = milligrams of p-chloroanilide per milligram of active solution P = per cent purity of final sample
a
=
DETERMINATIOS OF ACIDS
The method has been applied to the assay of the four conipounds listed.
Figure 1. Apparatus for quantitative conversion of carboxylic acid to p-chloroanilide via acid chloride A . 1 0 0 - i d reaction flask 6, D . Stopcocks C . Dropping funnel E . Flexible rubber connectioii
Acetic acid, analytical grade. Titration with standard sodium hydroxide showed a purity of 100.0 =k 0.2%. Procedure I was folloFed, except that a mixture of 5 ml. of benzene and 5 ml. of toluene was substituted for the toluene. The p-chloroanilide wae
ANALYTICAL CHEMISTRY
1320 Table I.
Taken, hlg. 126.6 127.8 125. Oa 125.56 122.2b 123.5C 128.2d 128.6d 128.O d
10 25 50 d 10 a 25 I50 a b
'
Recovered ME. Acetic Acid 126.8 127.3 126.8 127.2 121.4 122,7 127.0 130.4 127.2
% 108.2 99.6 101.4 101.3 99.3 99.3 99.0 101.4 99.4
Acetic Anhydride 114.3 118.2
100.1 100.8
131.2 134.0 134.7 132.0 134.9 126.04 132.4b 130.9b
Benzoic Acid 131.1 134.3 133.5 132.3 135.3 127.1 131.9 128.3
99.9 100.2 99.1 100.2 100.3 100.9 99.6 98.0
251.9 255.0 239 3 248.1 246 30 231 4 6 250.6/ 254 71
Stearic Acid, Procedure I 247 250 254 247 245 244 248 233
98.1 98,O 98.0 99.6 99.5 97.1 99 0 99.3
114.2 117.3
and the melting points of the prepared p-chloroanilides were as follows: acetic acid, 179.2-179.5' C.; benzoic acid, 193.2193.5' C.; and stearic acid, 103.8-104.2' C. I n the determinations of acetic acid, acetic anhydride, and benzoic acid the samples for radioactivity measurements were purified until the melting point differed by less than 0.2' C. from that of the pure substance. The correction due to the impurity is then very small; an estimated melting point depression of 0.5" C. per per cent of impurity was used. I n the determination of stearic acid the difference in melting point varied from 0.1" to 0.6' C.; the melting point depression of stearo-p-chloroanilide was determined to be 0.2" C. per per cent content of palmito-p-chloroanilide.
Determination of Carboxylic Acids and An hydrides
DISCUSSION
Stearic Acid, Procedure I1 280.1 248 99,2 255.4 258 101.0 244.5 244 99.8 264.5 264 99.8 mg. of water added t o sample. mg. of water added t o sample. mg. of water added t o sample. mg. of formic acid and 10 mg. of propionic acid added t o sample. mg. of palmitic acid added t o sample. mg. of palmitic acid added t o sample. ~~
Table 11. Effect of Vary-ing Excess of p-chlorophenylphosphazo-p-chloroanilide on Determination of lcetic and Benzoic Acids Acid p-chlorophenylphosTaken, M g . phazo-p-chloroanilide, Mg.a dcetic Acid 125.2 400 400 126.6 GOO 124.5
Recovered Rlg. i20,5 125.9 125.0
70 96.2 99,s 100.4
Benzoic Acid 130.9 200 131.8 100.7 134.0 300 132.5 98.9 136.3 500 135.3 99.3 134.2 600 134.6 100.3 a Amount of p-chloroanilide was twice t h a t of p-chlorophenylphosphazo-pchloroanilide.
crystallized from 80 ml. of water and twice from 10 ml. of 30% alcohol. Acetic anhydride, analytical grade. Determination by the method of Kappelmeier (6) showed an acetic anhydride content of 100.0 It 0.47,. Procedure I11 was followed and the p chloroanilide was crystallized twice from 10 ml. of 30y0 alcohol. Benzoic acid. This was an analytical grade product that was crystallized from toluene and then sublimed a t 0.1 mm. Titration with standard sodium hydroxide showed a purity of 100.1 =t 0.2%. Procedure I was followed for this determination. The p-chloroanilide was crystallized from 10 ml. of absolute alcohol and twice from 10 ml. of 60% alcohol. Stearic acid, melting point, 69-70" C. A laboratory sample with an estimated purity of 99% was furnished by The Danish Soyacake Factory, Ltd. Procedures I and I1 were both used. The p-chloroanilide was crystallized a t least three times from 10 ml. of absolute alcohol. The p-chloroanilides of the acids are all known (7, 11, 15),
The results given in Tables I and I1 show that the acids are quantitatively converted to p-chloroanilides by treatment with a mixture of p-chlorophenyl-phosphazo-p-chloranilide and pchloroaniline. Even moderate amounts of water are permissible. The reaction may also occur quantitatively with p-chlorophenylphosphazo-p-chloroanilide alone, but experiments showed that a purer reaction mixture was obtained when p-chloroaniline was added. Grimmel suggests that the reaction should be carried out in toluene, and this solvent has also been recommended in Procedure I. However, in the analysis of acetic acid satisfactory results were obtained only with a mixture of benzene and toluene. The same mixture could also be used in t'he analysis of benzoic acid and stearic acid, but experiments indicated that the determinations were considerably more sensitive to water and to varying excess of p-chlorophenylphosphazo-p-chloroanilide. Only a few experiments were performed to confirm the quantitative reaction of acid chlorides and anhydrides with p-chloroaniline, because other investigators have shown this type of reaction to be quantitative. In the analysis of stearic acid via stearoyl chloride (Procedure 11) the p-chloroanilide was a little difficult to purify. However, it was possible to obtain melting point differences as low as 0.2"C., but it is supposed that the correction for impurities does not hold in this special case. This may explain the slightly higher results obtained by Procedure 11. Dicarboxylic acids cannot usiually be determined by the method proposed, because of the formation of half-anilides or imides. Experiments with succinic acid using Procedure I gave converPions of only about 60%. ACKSOW LEDGhIENT
The author wishej to express his thanks to Sadolin &- Holmblad, Ltd., for permission to publish this article. The assistance of Jytte JBrn-Jensen in performing the radioactivity measurements is gratefully acknowledged. Ernst Boss kindly analyzed the phosphazo compound. LITERATURE CITED
(1) Escher, H. H., Hela. Chim. Acta 12,27 (1929). (2) Goldschmidt, S., Iirauss, H. L., Angew. Chem. 67,471 (1955). (3) Goldschmidt, S..Lautenschlager, H., A n n . Chem. Justus Liebigs 580, 68 (1953). (4) Goldschmidt, S., Obermeyer, F., Ihid., 588, 24 (1954). ( 5 ) Grimmel, H. IT.,Guenther, A., IIorgan, J. F., J . Am. Chem. SOC.68,539 (1916). (6) Kappelmeier, C. P. .I.,Anal. Chim. Acta 2, 146 (1948). (7) King, H., Orton, K. J. P., J . Chem. Soc. 99, 1377 (1911). (8) Menschutkin, S., Wassilief, M., J . Buss. Phys.-Chem. SOC.21, 192 (1889). (9) Xicolas, L., Burel. R., Chim. anal. 33, 341 (1951). (10) Orton, K. J. P., Bradfield, A. E., J . Chem. SOC.1927,983. (11) Owen, G., Ibid., 123, 3392 (1923). (12) Sorensen, P., ASAL.CHEM.26, 1581 (1954). (13) Ibid., 27, 388 (1965). (14) I b i d . , 391. (15) Walther, R. yon, J . prakt. Chem. ( 2 ) , 67, 445 (1903). RECEIVED for review February 4, 1956.
Accepted April 20, 1956.
