Coulometric Titration of Aniline Use of Copper(I I ) and Bromide as Dual Intermediates RICHARD P. BUC,K AND ERNEST H. SWIFT California Znstitute of Technology, Pasadena, Calif.
I t was desired to demonstrate the application of dual intermediates to coulometric titrations with bromine of a slowly reacting compound, and to develop a method for the determination of small quantities of aniline. A procedure for the coulometric titration of aniline with bromine has been developed in which an excess of bromine is generated coulometrically and reduced with cuprous copper. Wider application of the coulometric process to the bromometric titration of organic compounds which react slowly with bromine is indicated.
P
ROCEDURES have been developed in these laboratories
whereby electrolytically generated bromine has been used aa an intermediate oxidizing agent in secondary coulometric processes (1, 4,8, 9) for the determination of certain reducing agents, and electrolyticalIy generated cuprous copper has been used as a reducing int,ermediate in secondary coulometric titrations of chromate and of vanadate (5). I n these procedures, the substance to be determined reacts with the electrolytically generated intermediate, xhich is produced in the solution by the passage of a known constant current for a measured time. The end point is obtained by observing the current flow between two platinum indicator electrodes which have a small potential difference impressed upon them. I n general, the substances titrated have reacted rapidly and stoichiometrically with the electrolytically generatcd intermediate-at least, the reaction rate has been more rapid under these conditions than the generation rate. In the course of an investigation of coulometric titrations with bromine of certain organic compounds in aqueous solutions it was found t,hat the rate oi r e a d o n of the bromine with these compounds was slow and the end-point determination difficult. In conventional volumetric procedures, such titrations are usually carried out by the addition of an excess of bromine, followed after an appropriate time h!- a back titration. In order to adapt such a bromonietric titration to the coulometric method, a mearis for. the determination of the excess bromine must be found, as direct) cathodic reduction of the bromine is not possible; t,he concentration of excess bromine is too small for reduction with 1 0 0 ~ o current efficiency. Because the cuprous-cupric copper couple has been found to possess the qualification for a coulometric intermediate (5),the application of a dual intermediate system containing cupric copper and bromide t,o tlhe titration of aniline was investigated. Preliminary experiments which indicated the usefulness of this dual int,ermediate for The deterxniiiation of so-called bromine numbers had been made bj- hIeier ( 2 ) . The aniline was titrated by generating an excess of t)roniine, permitting it, to react for the necessary time, then back-titrating the excess bromine with electrolytically generated cuprous copper.
react, for about 3 minutes in the closed vessel; 2 grams of potassium iodide were added and the iodine was titrated with standard sodium thiosulfate to a starch end point. Tlie standardizaequivalent of aniline per gram of solution tion gave 1.416 X with a n average deviation of 0.2%. The potassium bromate solution was standardized iodometrically by weight titration against standard thiosulfate. The thiosulfate solution was standardized by weight titrat.ion against’ a h n d a r d potassium iodate solution. Separate cupric copper and bromide solutions were prepared. -40.2 6’ solution of copper sulfate in 12 F hydrochloric acid was prepared from twice recrystallized copper sulfate. Previously, the 12 F hydrochloric acid had been found by electrolytic oxidation with bromine t o contain approximately 1 X 10-6 equivalent of reducing material in a IO-ml. sample; this effect was eliminated by the addition of the required amount of a saturated solution of chlorine in hydrochloric acid. One volume formal sodium liromide solutions were used. KO extraneous oxidizing or reducing materials were found in these solutions. The laboratory-distilled water was boiled t,o eliminat,e an oxidizing agent, presumed t o be chlorine. Apparatus. The apparatus used was essentially the same as that described by Meier, Myers, and Swift ( 3 ) with the modification described by Ramsey, Farrington, and Swift ( 7 ) . , Preliminary Adjustments. The current in the generator circuit, Tvas determined as described (3) and corresponded t,o 1.0375 X 10-7 equivalent per second. A potent,ial of 200 mv. was impressed upon the indicator electrodes. Khen the coulometric apparatus was not in use, the electrodes were shorted t,o the generator anode and stored in a solution approximately 2 F in hydrochloric acid. Immediately before each $et of titrations, the elect.rodes were placed in a solution 0.1 F in sodiumbromideand 1F i n hydrochloric acid contained in a titra-’ tion cell (a 40 X 80 mm. Teighing bottle), and bromine !vas generated in the solution for 30 seconds a t high rate. This procedure aided in maintaining the stability of the electrode sensitivity; treatment with cleaning solut,ion was necessary only occasionally. Procedure. In carrying out a titration, 25 ml. of a solution prepared by dilution of the stock aniline solution were pipetted into a t,itration cell; 5 ml. of the 0.2 F copper sulfate solution (12 F in hydrochloric acid), 5 ml. of the 1 F sodium bromide solution, and 10 ml. of water were added. The generator current was set a t the desired value and the initial indicator current was ohservetl,
EXI’KRIII E S T A L
Chemicals. A11 chemicals i v c w reagent grade. The stock aniline solution was prepared by dissolving frpshly distilled aniline in 0.02 F (volume formal) hydrochloric acid. Air was removed by bubbling nitrogen through the solution to prevent oxidation of the aniline. This st,ock solution was standardized by a modificat,ion of the hromometric method :is folloivs: Ten-milliliter samples of t,he stock aniline solution were 11-eighed into 500-nil. conicd flask? containing 100 ml. of boiled dist,illed water, 2 grams of potassium broinitle, and 5 ml. of 6 F hydrochloric acid. About 25 ml. of standard 0.02790 P potassium bromate solution were weighed into the solution and allowed to
499
JVhen an unhnovn aniline solution ~va6analyzed, trial titrations were made by generating bromine for a period of time, stopping the generation, and observing the change of the indicator current. with time. For titrations of greater than 100-second generation time, it was most convenient to generate bromine in the solution a t the high rate (IO-; equivalent per second) until the indicator current wis “off scale” with the intlicat,or current microammeter shunted doirn to one third of it,s normal sensitivity. -4fter stopping the generation, a pause of about 80 seconds was often nccesaary before the indicator current was again “on scale.” In the determination of smaller samples, generation of bromine in 10-second intervals, followed by a pause to observe the change of the indicator current, was found to be practical. If the equivalence point for the formation of tribromoaniline had not heen reached, the indicator current decreased continu-
so0
A N A L Y T IC A L C.HE M I S T R Y
ously with time during this pause, indicating that the bromine was still reacting with the aniline. However, as the equivalence point was approached, the rate of decrease became slower. If the equivalence point were passed, the indicator current did not decrease noticeably during the 30-second wait. When the approximate equivalence point was known, the indicator circuit was opened during the titration and generation of bromine was continued until an approximately 25-second excess was produced a t the high generation rate. This quantity was found sufficient to bring the reaction essentially to completion within 1 minute. Then the polarity of the internal generator electrode was reversed and cuprous copper was generated until most of the excess bromine had been reduced, as demonstrated by the return of the indicator current “on scale.” The generation was continued in half-second intervals until the indicator current decreased to 15 pa. Between each period of generation a 10second pause was taken while the indicator current reached a steady value. A plot of indicator current against time of generation during a titration shows an initial rise in the current, followed by an off scale period, then upon generation of cuprous copper a return to on scale, followed by a linear decrease in the indicator current during the range of 40 to 10 pa.
A correction for the impurities in the reagents used and for any loss of bromine was made by titrating six blank solutions containing 5 ml. of 0.2 F copper sulfate in 12 F hydrochloric acid, 5 ml. of 1 F sodium bromide, and 35 ml. of water. The generator current was set a t the desired value and bromine was generated for 25 seconds, the time of generation of excess bromine in an aniline titration. The solution was permitted to stand for 1 minute without stirring, then the polarity of the internal generator electrode was reversed and cuprous copper was generated until an indicator current of 15 pa. was obtained. The “correction time’’ was designated as the difference between two times: The larger time was that of anodic generation (bromine), and the shorter time was that of cathodic generation (Cu ( I ) ) back to the arbitrary value of 15 pa. The average correction time was determined from the last four blanks which were made, as these were usually the most consistent. The time of cuprous copper generation back to an indicator current of 15 pa. for the aniline titration and the correction time were subtracted from the total time of bromine generation to obtain the corrected titration time. The corresponding weights of aniline were calculated from the titration time and the rate of generation. As has been shown by Meier ( d ) , if the generation of copper(1) is continued beyond the time corresponding to an indicator current of 15 pa., the indicator current passes through a broad minimum, and again increases as cuprous copper is generated in excess. The arbitrary value of 15 pa. on the linear portion of the bromine-controlled curve was chosen as the end point because a given deviation in current measurement a t such a point causes less error in the corresponding generation time than would result with a point in the region of minimum current or on the linear portion of the more horizontal copper(1) controlled curve. The end point was found to be stable and reproducible. By controlling the polarity of the internal generator electrode, one could generate bromine or cuprous copper as desired and so pass through this arbitrary point, as well as the minimum, several times without marked error. A series of experiments showed that one could pass through the minimum five times before the time corresponding to 15 pa. of indicator current varied over 0.1 second. DISCUSSION OF BROMINE-ANILINE REACTION
Pamfilov (6)states in his review of methods for the quantitative determination of aniline that the bromometric method is the most convenient and accurate one for aniline concentrations in the range from 0.1 to 10-5 F. This method is based upon the reaction
of bromine and aniline to yield 2,4,6-tribromoaniline; however, Pamfilov and Kisselva (6) showed that the reaction is rapid only in “weakly” acid solutions. I n 0.1 P hydrochloric acid, the reaction is so slow that a direct titration with bromine is very difficult. There would be an advantage in adapting the bromometric method to solutions of higher acid concentrations because the titration would be expected to be more specific and less subject to side reactions, and because aniline is conveniently collected in such solutions. However, in studying the application of the coulometric process to the determination of aniline in solutions of strong acids, it was observed that the bromination reaction was so slow in solutions approximately 10-5 P in aniline and 1 F in hydrochloric acid that bromine accumulated in the solution during the titration; thus after 4 seconds of continuous high rate generation, the indicator current was off scale and when the end point was reached there was no characteristic increase in the indicator current,
Table I. Sample NO.
I1 a b
Confirmatory Titrations Taken
Aniline, Micrograms Found
199.5
C
d e
f
Av.
I11 a b
99.7
100.1 99.1 99.8 99.8 99.6
C
d e
w n
f
Av.
IV
a
b
:
e f
Va b C
d e
199.6 199.4 200.1 199.5 199.5 200.1 199.7
99.4 99.5 99.55
Error
0.1 -0.1
0.6 0.0 0.0 0.6 0.2 0.4 -0.6 0.1 0.1 -0.1
-n 7 -0.3 -0.2
-0.15 0.6 0.0 0.0 -0.5 -0.4
-0.3 -0.1
0.0
-0.1 0.0 0.0 -0.4 -0.1
A satisfactory procedure required the generation of an excess of bromine sufficient to complete the reaction in a reasonable time, yet the excess could not be such as to cause significant volatilization of bromine during this time. Experiments showed that the generation a t 10 pa, of a t least 25 seconds of excess bromine waa necessary for completion of the reaction in 1 minute; a 5-second excess (at 10 pa.) led to an error of several per cent. TITRATIONS IN ACETIC ACID-ACETATE SOLUTIONS
Although at p H values from 0 to 1 the rate is so slow that dual intermediates must be used, preliminary experiments have shown that in an acetate buffer with a pH value of 4.5 the rate of the reaction is sufficiently rapid that a secondary coulometric titration with bromine, and with an amperometric end point, can be performed. With p H values greater than 6 the hydrolysie of the
V O L U M E 2 4 , NO. 3, M A R C H 1 9 5 2 bromine caused late end points and correspondingly high results.
501 able the sample of redistilled aniline, and Paul S. Farrington for his cooperation during the experimental work.
CONFIRMATORY TITRATIONS
Table I contains data obtained from confirmatory titrations carried out as described. These data show a n average error of 0.5 microgram and an average deviation of 0.3 microgram for samples of from approximately 100 and 300 micrograms, and an average error of 0.2 microgram and an average deviation of 0.2 microgram for samples between approximately 10 and 50 micrograms. The factors limiting the accuracy of the above measurements are thought to be the preparations, the standardization, and the instability of the aniline solutions. I n the larger samples, there was considerably better agreement among titrations within one group than between the average titer of the group and the calculated value for the titer. The average “correction time” was 0.3 second of generation time. ACKNOWLEDGMENT
The authors wish t o thank Howard J. Lucas for making avail-
LITERATURE CITED Brown, R. A,, and Swift, E. H., J. Am. C h a . Soc., 71, 2717 (1949). Meier, D. J . , master’s thesis, California Institute of Technology, 1948. hleier, D. J., Myers, R. J., and Swift, E. H., J . Am. Chem. Soc., 71,2340 (1949). Myers, R. J., and Swift, E. H., Ibid.,70, 1047 (1948). Pamfilov, A. V., 2.anal. Chmn., 69,282-92 (1926). Pamfilov, -4. V., and Kisselva, V. E., Ibid., 72, 100-12 (1927). Ramsep. W. J.. Farrington. P. S.. and Swift. E. H.. ANAL.C m x . 22,332 (1950). (8) Sease, J. W., Niemann, C., and Swift, E. H.,Ibid., 19, 197 (1947). (9) Wooster, W. S.,Farrington, P. S., and Swift, E. H., Ibid., 21, 1457 (1949). I ~ E C E I V Efor D review August 27, 1051. Accepted December 27, 1951. Contribution 1624 from the Gates and Crellin Laboratories of Chemistry, Califprnia Institute of Technology.
Estimation of Molecular Weight of Starch Polysaccharides Determination of Their Reducing End Groups SIEGFRIED NJSSENBAURI AND F.Z. ELASSID Division of I’larit Biochemistry, College of Agriculture, University of California,Berkeley, Calif.
The methods for determining the molecular weight of amylodextrins, aniyloses, and amylopectins based on osmotic pressure or ultracentrifuge measurements are laborious and require large amounts of material. The existing colorimetric methods yield only relative molecular weights. The present method is based on the determination of the reducing group of the polysaccharide by an adaptation of the Folin and Malmros colorimetric procedure for the estimation of glucose. Alkaline ferricyanide in the presence of cyanide is used as an oxidizing agent. Comparison of the molecular weights of a number of amylodextrins, amylose starch fractions, synthetic polysaccharides, and amylopectin fractions of low molecular weight with those obtained from osmotic pressure and other measurements showed a fair agreement. The method allows rapid estimation of molecular u eights of some polysaccharides and requires little material.
T
HE methods available for the estimation of molecular weights of polysaccharides based on the determination of the aldehydic end group, yield information only about the relative molecular size (6). Generally the values for molecular weights of polysaccharides obtained by these methods have not been compared with those obtained by accepted methods, such as osmotic pressure or ultracentrifuge measurements. Lansky, Kooi, and Schoch ( 5 ) investigated a number of such procedures for determination of molecular weight of starch fractions They found that some ferricyanide ( I , 4 ) , alkaline copper (IO),and alkaline 3,5-dinitrosalicylate (6, 7) reagents, even when fairly selective towards oxidation of the terminal aldehydic group, do not give a stoichiometric relationship between glucose and maltose. It therefore appeared that, when these methods are applied to higher polysaccharides, they reflect relative sizes only. In the present investigation it has been found that Folin and Malmros’ (2, 11) method for the determination of reducing sugars could be adapted for the estimation of molecular weights of polysaccharides with a considerable degree of accuracy. It is based on oxidation of the aldehydic group by ferricyanide, the addition of ferric sulfate to form Prussian blue, which is stabilized by gum ghatti, and the determination of the color intensity of the
Prussian blue formed. The relationship of the reducing values between glucose, maltose, and heptaose (seven-glucose-unitpolysaccharide) is stoichiometric, and the molecular weights of amylodextrins, amyloses, and the smaller amylopectin molecules obtained by this method are in close agreement with values obtained by osmotic pressure measurements. The applicability of this method to the determination of molecular mights of amyloses and amylopectins was tested on a number of samples available in this laboratory. The heptaose and 23- and 42-glucose unit amylodextrins were supplied by Dexter French and J. H. Pazur of the Iowa State College. Using chromatographic analysis these workers found the heptaose to be homogeneous and virtually free of impurities (3). They also determined the molecular weights of the nonhomogeneous amylodextrins by oxidation of the aldehydic end group and titration of the resulting acidity. Examination of the results presented in Figure 1 shows that the intensity of color developed by glucose, maltose, heptaose, and 23- and 42-unit dextrins is directly proportional to the number of moles of reducing groups present, and is independent of the chain length. Table I shows a fair agreement between the molecular weight values of amyloses obtained by osmotic pressure measurmenta
