Determination of Sulfur by Peroxide Bomb Decomposition and Am peromet ric Titration BENJAMIN WARSHOWSKY, THOMAS E. SHOOK, and E. J. SCHANTZ
Md.
Cml C Biological Laboratories, Camp Detrick, Frederick,
A
LTHOUGH numerous procedures have been described for the determination of micro quantities of sulfur in organic compounds, no detailed study has been reported on the use of amperometric techniques following perovide combustion. The standard gravimetric micromethod is time-consuming and the turbidimetric procedure is subject t o considerable error. For this reason, a study was undertaken to develop a n amperometric titration of the sulfate resulting from a peroxide bomb combustion. Majer (4)was the first to report an amperometric titration of sulfate in which lead nitrate was used as the titrating reagent. Since the solubility of lead sulfate varies greatly with the ionic strength of the medium, his proceaure was not satisfactory for samples containing unknown quantities of salt. Spalenka ( 5 ) improved the method by specifying that the titration be performed in a medium containing 30 to 40% ethyl alcohol t o repress the solubility of the precipitate. Kolthoff and P a n ( 3 )described the optimum conditions for the amperometric titration of sulfate with lead. For a 0,OOlM sulfate solution, it was recommended that the medium contain a t least 30% of ethyl alcohol and the potassium nitrate concentration be less than 0.1s. When this titration was applied by the authors t o a system containing a high salt concentration, such as would be encountered in a peroxide fusion, the precipitation of lead sulfate was almost completely inhibited and, therefore, the method could not be adapted t o the peroxide combustion. Another amperometric determination of sulfate was carried out by Heyrovskq and Berezicky ( I ) , who measured the decrease in the diffusion current of barium upon titration with sulfate. Apparently their results were inconclusive. Later Kolthoff and Gregor ( 2 ) made studies of the polarographic behavior of barium and established conditions for its determination by amperometric titration with potassium chromate. The procedure described in this paper involves combustion of the sample in a Parr peroxide bomb, followed by precipitation of the resulting sulfate with excess barium, and an amperometric titration of the excess barium with a standard solution of potassium chromate. As the solubility product of barium sulfate is less than that of barium chromate, the precipitated barium sulfate does not redissolve in presence of small excess of chromate ions. APPARATUS
A Heyrovskg polarograph, Model X I I , manufactured by E. H. Sargent and Co., was used for the amperometric measurements. T h e drop time for the capillary in a 3M sodium chloride solution was 4 seconds a t zero applied voltage. The cell was a t room temperature. A silver-silver chloride electrode was used as the reference electrode. This was made by electrolyzing a saturated sodium chloride solution for approximately 3 minutes using a heavy gage silver wire as the anode. T h e large titration cell consisted of a 100-ml. wide-mouthed bottle fitted with a rubber stopper which contained openings for the silver-silver chloride electrode, a dropping mercury electrode, a nitrogen inlet and outlet tube, and the tip of a 5-ml. buret. T h e buret was graduated in 0.01 ml. For the microgram-scale titrations, the cell was a 50ml. centrifuge tube, which used the same components as the larger unit except that they were reduced t o a smaller scale. A Parr electric ignition bomb (22-ml. capacity) was used for all combustions. REAGENTS
Merck’s sodium peroxide and benzoic acid, and Fisher’s potassium perchlorate were used for the combustion. Stock solu-
tions of 0.01M barium chloride, 0.01fM potassium chromate, and 3M sodium chloride were of Merck’s reagent grade materials. T h e potassium chromate was standardized iodometrically with standard sodium thiosulfate solution. The other reagents, 10% sodium hydroxide solution, concentrated hydrochloric acid, sodium hypochlorite, and ethyl alcohol, also were reagent grade. T h e sodium hypochlorite solution was prepared by passing chlorine through a 5 % solution of sodium hydroxide. Commercial sodium hypochlorite was not suitable, as it appeared t o contain significant quantities of sulfate. Sational Bureau of Standards cystine, Merck’s sulfanilamide, and Baker’s sodium sulfate were used as standards. PROCEDURE
Combustion. T o an 8- to 12-mg. sample, containing preferably 1 to 5 mg. of sulfur, are added 50 mg. of benzoic acid, 100 mg. of sodium perchlorate, and 1.5 grams of sodium peroxide. These materials are well mixed and then placed in a small paper cup which is placed in the nickel cup of a Parr macro electric bomb. The paper cup is constructed simply from a small piece of quantitative filter paper which weighs approximately 70 mg. It is used to hold the charge for combustion in the Parr electric bomb. The electric fuse wire is embedded in the charge inside the paper cup and ignited in the usual manner. After combustion, the bomb is cooled. T h e metal cup is placed in a small beaker, and the melt is dissolved in approximately 20 ml. of hot water. T h e cup is removed and washed with warm water, and the solution is transferred t o a 250-ml. Erlenmeyer flask. A few glass beads are added and the solution is boiled on a hot plate for 5 minutes in order to decompose most of the peroxides. T h e remaining peroxides, which interfere with the amperometric titration, are decomposed by the addition of 5 ml. of sodium hypochlorite. The solution is boiled for an additional 2 minutes and then acidified by the dropwise addition of concentrated hydrochloric acid. When sufficient acid has been added, the vigorous evolution of gases ceases and the color of the solution changes from dark gray t o a clear amber. Approximately i ml. of concentrated hydrochloric acid are usually sufficient to decompose the excess sodium hypochlorite and to neutralize the alkaline digest. The solution is allowed to cool and is then transferred to a 50-ml. volumetric flask and diluted t o the mark with distilled water. Precipitation of Sulfate and Preparation for Titration. A 20ml. aliquot of the combustion solution is placed in the large titration cell and warmed for approximately 5 minutes in a boiling water bath. Ten milliliters of the 0.01M barium chloride solution are added slowly to the aliquot, and the precipitate of barium sulfate is allowed to digest for 20 minutes. The solution is cooled, and the p H is carefully adjusted to the orange color of methyl red (ea. p H 6). In order to keep the volume to a minimum, this first is done approximately with a 10% solution of sodium hydroxide and the final adjustment is made with more dilute alkali. Thirty milliliters of ethyl alcohol are added t o the solution, making a total volume of about 60 ml. For the micro scale titration, all quantities are reduced by a factor of 10-e.g., a 2-ml. aliquot sample, 1 ml. of 0.01M barium chloride and 3 ml. of ethyl alcohol. Amperometric Titration. The cell is connected t o the titration assembly and deaerated with nitrogen for 5 minutes. T h e initial residual current is noted on the polarograph, which is a l justed to a sensitivity of 5 X and an applied voltage of -1.35 volts. The sample is titrated amperometrically with 0.01M potassium chromate. After each increment has been added, nitrogen is again bubbled through the solution for 2 minutes. Following the first sharp rise in the current, the increments of potassium chromate are limited t o 1 ml. Four or five points past the first rise in current will usually be sufficient t o obtain a straight line. T h e end point is determined graphically by plotting the galvanometer deflections (corrected for dilution effect) against the volume of titrant added. A blank determination is made t o correct for any impurities in the reagents and to eliminate any effect due t o the high salt concentration or other factors resulting
1051
ANALYTICAL CHEMISTRY
1052 from conditions of the determination. The same procedure is followed for the microtitration, b u t smaller increments (0.1 ml.) of titrant are added. The sulfur content of the sample is calculated as fol1ow.p: Sulfur, wt.
yo =
V
X M X 32.06 X 100
w
where V = volume in milliliters of standard potassium chromate used by the sample, which equals the difference in volumes required by the sample and blank t,itrations; M = molarity of the standard potassium chromate; and TY = weight in milligrams of the sample in the aliquot tit'rated.
Table 11. Effect of Sodium Chloride on Titration of Barium Chloride Solutiona with Potassium Chromate 0.01.w Sodium Chloride .4dded, J I g 20 100 200 400 600
Potassium Chromate Consumed AI1 9.99 9.96 9.80 9.74 9.66
Difference, 1\11 -0.01 -0.04 -0.20 -0.26 - 0 34
a Solutions contained 10 ml. of 0 01.11 BaC12, 20 rnl. of H20, and 30 ml. of ethyl alcohol.
EXPERIMENTAL AND DISCUSSION
The paper cup arrangement for combustion in the electric ignition bomb was found to be more satisfactory than that in which the microbomb involving ignition by flame was used. During the course of these studies much more difficulty was encountered in repeatedly obtaining complete combustion with t,he micro flame bomb than with the electrically ignited bomb, even a h e n micro samples were used. For that reason only the electric ignition met,hod outlined in the procedure was followed. Size of Sample. Using the procedure described, Sational Bureau of Standards cystine and Merck's sulfanilamide were analyzed. Samples ranging in sizes from 9 to 14 mg. were oxidized, and aliquots containing 0.1 to 1.5 mg. of sulfur were titrated. The results of these tests on known materials are tabulated in Table I and show an accuracy in the order of 9891, of the theoretical value. I n Figure 1 are recorded the data of a
I
I
f
i
I
0 IO 20 30 40 50 60 VOLUME OF 0.01 M POTASSIUM CHROMATE,ML.
Figure 2. Effect of Alcohol Concentration on .-imperometric Titration of Barium
K
w
80-
I-
W
z
0
3
$ 1 60-
40
0 0.2 0.4 06 OB 1.0 1.2 VOLUME OF 0.01 M POTASSIUM CHROMATE, ML.
Figure 1. Amperometric Microtitration of Sulfate
Table I. Sample 1
2 3 4
5
1
2 3 a
Determination of Sulfur in Cystine and Sulfanilamide Wt. of Sulfur in Aliquot, M g . Present Foundn Cystine (26.7% Sulfur) 1 . 4 7 , l .51 1.50 1.36 1.33 1.02 0.133
Error, $2 -0.8 -2.2
1.33 1,32,1.31 1 .00 0.131
-1.1 -1 .!4
Av.
-1.5 - 1. .5
A\.
-2.6 -1.1 -1.6 - 1. 8
Sulfanilamide (18.67,Sulfur) 1.05,l.07 1.09 0.91 0.89,0.90 0.173.0.173 0.176
Except for cystine sample 5 , all results were obtained by macrotitration
typical experimental titration of the sulfate resulting from a cystine combustion. Curve a represents t,he microtitration of a 2ml. aliquot containing 102 y of sulfate sulfur from a peroxide combustion, 1 ml. of 0.01111 barium chloride and 3 ml. of ethyl alcohol; curve b is the corresponding blank. Salt Effect. T o determine the effect of various concentrations of salts on the accuracy of the titration of barium ivith chromate, standard barium chloride solutions containing increasing concentrations of sodium chloride xere used as the media. It was found that as the concentration of salt increased, the amperometric end point appeared earlier. The extent of this salt effect is seen in the data presented in Table 11. ;ilthough not systemat'ically investigated, an even more pronounced effect \vas caused by the presence of nitrate ion. For this reason, the presence of nitrate ions v a s avoided hy the use of h?-drochloric acid for the neutralization of the combustion products rather than nit,ric acid. The effect of salt on the end point of the titration is presumed to be due either to a change in the solubility of barium chromate or to a measurable amount of coprecipitation of barium ions. I n any event, the blank determination, which is performed in the same manner as the unknown and with the same amounts of reagents, will compensate for these variations. Effect of pH. Preliminary studies indicated that unless the proper p H was maintained, erratic and nonreproducible results would be obt,ained. This factor !vas studied by performing determinations on synthetic solutions which were adjusted to a selected pH value prior to the chromate titration. This synthetic solution consisted of 5 ml. of 0,0131 sodium sulfate, 5 ml. of 3.11 sodium chloride, 10 ml. of 0.01JI barium chloride, 10 ml. of water, and 30 ml. of ethyl alcohol. The p H was adjusted to the desired value, as measured with a Beckman Model G p H meter, by the careful addition of acid or alkali to the unbuffered solution. Below p H 3, no precipitate of barium chromate formed owing
V O L U M E 2 6 , N O . 6, J U N E 1 9 5 4
1053
to its solubility in mineral acids. At a p H between 3 to 5.5, precipitation n-as elow, and irregular results R ere obtained. I n the p H range betrveen 5.5 and 6.2, accurate and reproducible values were found. Above pH 6.5 values again became irregular and high, presumablv owing to the presence of carbon dioxide from the air, resulting in the precipitation of barium carbonat_. hfter several determinations the solutions with high p H values had a poisoning effect on the silver-silver chloride electrode Several buffei systems were investigated for the purpose of maintaining the p H within the designated range of 5.5 to 6.2. These included acetic acid-sodium acetate, malonic acid-sodium malonate, and succinic acid-sodium succinate. -411 these buffers, even in very small amounts, completelv inhibited the reaction tietween barium and chromate ions. This was apparent from the fact that in addition to the absence of an observable precipitate of barium chromate, the current increased after the first increment of chromate had been added. I n view of this unpredicted phenomenon, the pH thereafter was adjusted to the proper value without the use of buffers bv simplv adding acid or alkali until the solution was orange to methyl red. Effect of Alcohol Concentration. Because the solubility of barium rhromate is noticeably affected by the presence of in-
different salts, Ilolthoff and Gregor ( 2 ) carried out the titration in the preience of alcohol. I n order to ascertain the minimum concentration of ethyl alcohol that would be required to yield reliable results under the conditions of the determination, this factor was studied. The results obtained by titrating synthetic media similar to those used for the p H study but containing increasing concentrations of ethyl alcohol are illustrated in Figure 2. As seen in the figure, the solubility of barium chromate is appreciable when the alcohol concentration is below 50%, making the base line, and therefore the end point, uncertain. Consequently, i t is recommended that the ethyl alcohol concentration be at least 5070. LITERATURE CITED
(1)
Hegrorsk$, J., and Berezicky, S.,Collection Czechosloc. Chem.
Communs., 1, 19 (1929). ( 2 ) Kolthoff, I. AI., and Gregor, H. P., - ~ N . A L .CHEM., 20, 541 (1948). (3) Kolthoff, I. AI., and P a n , Y. D., J . .4m. Chem. Soc., 62, 3332 (1940).
Majer, V., 2. Elektrochent., 42, 120, 123 (1936). ( 5 ) Spalenka. A f . , Collection Czechosloc. Chem. Commms., 11, 146
(4)
(1939). R E C E I V Efor D review October 1 , 1953.
Accepted February 5 , 1954.
Determination of Formic Acid by Oxidation with lead Tetraacetate A. S. PERLIN Division o f Applied Biology, National Research Laboratories, Ottawa, Canada
S
LOIT oxidation of formic acid to carbon dioxide by lead tetraacetate in aqueous acetic acid was originally reported by Grosheintz (4). Alore recently, Mosher and Kehr ( 5 ) have described the reaction with glacial acetic acid as solvent. The oxidation is greatly accelerated by potassium acetate, and the consequent enhanced rate of evolution of carbon dioxide makes the reaction suitable for manometric measurement of formic acid production during lead tetraacetate oxidations of carbohydrates (6). T h e present paper proposes t h a t the accelerated reaction may have M ider application for the determination of formic acid. I n some respects, it offers advantages over the usual analysis by oxidation with mercuric chloride (1, 2 ) . For example, the lead tetraacetate oxidation provides a simple, accurate, volumetric analysis (Table I, procedure 1) which is much more rapid and less tedious than the gravimetric procedure of the latter method. Further, 4hlBn and Samuelson ( 1 ) have shon-n t h a t
Table I.
Determination of Formic Acid
formaldehydp interferes with the mercuric chloride method and must first be separated from the formic acid-e.g., in the analysis of sulfite waste liquors. I n contrast, formaldehyde has no effect on the determination of formic acid by oxidation with lead tetraacetate (Table I). The lead tetraacetate procedure can also be used for determinations in the presence of other acids such as acetic, propionic, and succinic (Table I), and may therefore find application in the analysis of fermentation liquors. Formic acid may be determined not only by titration of the amount of reduced lead tetraacetate but also by measuring t,he amount of carbon dioxide evolved. T h e latter procedure is especially useful in the presence of substances-e.g., glycolswhich consume lead tetraacetate but do not yield carbon dioxide or formic acid. The Warburg respiromet'er is suitable for the micro range (6) (Table I, procedure 2.i), or a gas adsorption train for larger quantities of formic acid (Table I, procedure 2B). Procedures 2 d and 2B are suited also to the determination of compounds, such as glycerol and osalic acid, which yield formic acid or carbon dioxide vihen oxidized with lead tetraacetate. EXPERIXIENTA L ~
Procedure 1
2A
2B
Other Compounds .4dded, Added. ME. hlg. ... 1 05 2 14 2.1.4 2 40 (Formaldehyde) 2.11 5.0 (Propionic acid) 0.5 (Succinic acid) 4.22
HCOOHa
0.080 0.16 0.33a 21.5
0:k3b (Formaldehyde)
1.06 2.12 2 13
Reaction Time, Min. 20 20 20
2.11
20
4 18
20
0 077 0 10 0 33
30 30 30
21.3 21.4 32 1 32.7 BY alkali titration of a stock aqueous solution of formic acid. Produced i n siiu by oxidation of 0.66 mg. of glycerol. 32.3
a
...
HCOOH Found. I l g .
75
90
(io
7.5
~~
Reagents. Formic acid, reagent grade. Glacial acetic acid, reagent grade. Lead tetraacetate. A satisfactory method of preparation is given by Vogel (8). Anhydrous potassium acetate, reagent grade. Stopping solution: 10 grams of potassium iodide and 50 grams of sodium acetate dissolved in 100 ml. of water (3). ~, Standard 0 . 0 2 5 sodium thiosulfate. Procedure 1. Five milliliters of a solution of formic acid (1 to 5 mg.) in 80% acetic acid are added to 5 ml. of oxidizing solution, consisting of 100 mg. each of lead tetraacetate and potassium acetate in glacial acetic acid, in a glass-stoppered flask. X blank is prepared from 5 ml. of 807, acetic acid and 5 ml. of oxidizing solution. When the reaction has proceeded at room temperature (25" to 27" C.) for 20 to 30 minute$, 10 ml. of stopping solution is added. The yellow precipitate of lead iodide which appears may be dissolved b v addition of water and the iodine is titrated with standard t h i o d f a t e to a starch end point. The titration is equally satisfactory with the lead iodide present, but a pea-green color is given by the starch-iodine complex instead of the normal
