358
INDUSTRIAL AND ENGINEERING CHEMISTRY
The potentiometric method can be applied in dilute solutions and in general is more accurate than the amperometric method, but is considerably slower. To obtain reliable results potentiometrically it is necessary that solubility equilibrium be reached a t each point of the curve, especially in the near vicinity of the end point. B~ titrating to an equivalence potential, the potenti,+ metric method can be made more rapid, but the attainment of equilibrium must be experimentally proved. LITERATURE CITED
(1) Fodk, C. W.,and Bawden, A. T., J . A m . Chem. sot., 48, 2045 (1926).
Vol. 18, No. 6
(2) Hume, D. N., and Harris, W. E.,IND.ENQ.CHIEM., ANAL. ED., 15, 465 (1943). (3) Kolthoff, I. M., and Harris, W. E., Ibid., 18, 161 (1946). (4) Kolthoff,I. M., and Lingane, J. J., ~ ~ p o l a r o g r a p h yChapter ~~, XXXIII, New York, Interscience Publishers, 1941. ( 5 ) Laitinen, H. A., IND. ENO.CHEM., ANAL.ED.,13, 393 (1941). (6) Laitinen, H. A., and Kolthoff, I. M., J. P h w . Chem., 45, 1079 (1941). (7) salomon, E., z. ,,hysik. Chem,, 24, 55 (1897). (8) Ibid., 25, 366 (1898). (9) Salomon, E., 2.Elektrochem., 4, 71 (1897). THIEinvestigation was carried out under the sponsorship of the O5ce of Rubber Reserve, Reconstruction Finance Corporation, in connection with the Government Synthetic Rubber Program.
Amperometric Titration of Mixtures OF Halides Using the Rotating Platinum Electrode T. D. PARKS', Noyes Chemical Lab0 atory, University of Illinois, Urbana, 111.
H. A. LAITINEN, W. P. JENNINGS,
f
Iodide, bromide, and chloride can be successively titrated in mixtures with silver nitrate, using the rotating platinum electrode as an amperometric indicator electrode. Ammonia i s added for the iodide titration, an excess of acid is added for the bromide titration, and gelatin i s added for the chloride titration. Multivalent metallic ions added as flocculating agents do not appreciably affect the titration of bromide in bromide-chloride mixtures, indicating that mixed crystal equilibrium is not reached during the rapid addition of silver nitrate. The amperometric method i s much more rapid than the potentiometric method for mixtures of halides, although not always as accurate.
C
ONSIDERABLE work has been done in the titration of mixtures of halides, especially by potentiometric titrations.
however, considered adsorption to be responsible. Flood ( 5 ) calculated the theoretical error due to ideal mixed crystal formation assuming (a) that the entire precipitate is in equilibrium with the solution a t all times and ( b ) that the precipitate coagulated upon formation, and only the part being precipitated a t each instant is in mixed crystal equilibrium. Flood and Bruun (6) applied the results assuming "ideal inhomogeneous mixed crystal formation" (case b) to chloride-bromide mixtures, and calculated correction factors to be applied to the results as a function of mole fraction of bromide in the mixture. Thus for an equimolar mixture, the bromide result should be 1.2% high and the chloride result an equal amount low. Flood and Bruun concluded that mixed crystal formation accounted largely for the experimental results but that adsorption also was a contributing factor. They suggested that increased dilution would diminish the error due to adsorption. Zintl and Betz (17) titrated mixtures of halides in very dilute solution (about 0,001N ) . Bromide could barely behitrated in the presence of 10 times aa much chloride, iodide in the presence of 60 times as much bromide, and iodide in the presence of 5000 times rn much chloride. Flood and Sletten ( 7 ) and Schutza (16) worked out techniques for detecting potentiometric end points in chloridcebromide mixtures which require empirical correctiom depending upon the composition of the mixture. The potentiometric titration method suffers from the inherent difficulty that the potential break for bromide in the presence of chloride is only about 0.1 volt (IS).
Bromide and i6dide mixtures were first titrated potentiometrically by Behrend (I),who added sufficient ammonia to keep silver bromide in solution. He waa unable to determine bromide in the resence of chloride. Dutoit and von Weisse (4) and F'inkhoP(14) titrated iodide in the presence of bromide without adding ammonia. In attem ts to titrate bromide in the presence of chloride, Pinkhof (14y added ammonium carbonate to prevent the precipitation of silver chloride. By using the proper reference electrode, the titration was carried out until the bromide concentrrttion waa 10-3 N , and a correction of this Table 1. Effect of Multivalent Cations on Bromide Titration in Absence and Presence of amount waa added to the result. Chloride Liebich (18) attempted to titrate mixNorNorNortures of halides. Titration of iodide gave mality mality mality high results in the presence of bromide, of of of BrCI&NOJ Electrolyte Present AgNOs Used Error but accurate results were obtained in the presence of barium nitrate. Similar high MI. % results were found for iodide in the pres0.01 ... 0.1 . , . . . . . . . .. 9.96,9.90 -0.4, -1.0 9.95,9.96,9.97 -0.5 -0.4, -0.7 0 . 2 M Bat+ ence of chloride, and bromide in the 0.005 M A I + + + 10.04 9.96 O.d, - 0 . 4 resence of chloride. The addition of -0.7, -0.6 0.005 M AI+++,0 . 8 N HNOJ 9.93:9.94 arium nitrate or alum waa shown to im0.02 M A l + + + ,0 . 8 N HNOa 10.00 0.0 0,002 M Th +4 10.05,9.95 0.5, -0.5 prove the results. Clark (2, 3) found about a 1%error in bromide titrations in 10.01 0.1 0.001 ... 0.01 0.002 M.............. 9.95,10.04 0 .5,O. 4 Th +4 the presence of an equal concentration of chloride in the presence of 5% barium ni0.01 0.1 0.1 0 . 8 "NO: 10 03 10.03 10.07 0.3,0.3,0.7 0.2 M Ba++ 9:93: 9.92:9.93 -0.7, - 0 . 8 , - 0 . 7 trate. He also titrated mixtures of bromide 0.002 M A I + + + 0 8 N HNOa 10 03 10.06 0.3,0.6 and iodide in the presence of barium 0 2 0 9 0.005 M AI+++: 018 N HNOi 10:OZ: 10.09 nitrate, and chlorideiodide mixtures with0 02 M A l + + + 0 8 N HNO: 10.00,10.04 0.'0,'0.'5 0.5 0.7 10.05 10.07 0'01 M Th+4 6 8 N HNOi out salt addition. 0.7: 1 . 2 , l . Z 10.07: 10.12,lO.12 0:OZ M Th+4: 0 : s N HNO: The high results obtained for the more 10.24,10.20 2.4,2.0 0.01 0.02 0.1 0.8NHNO: insoluble halide in chloride-bromide or 0.8,O. 1 10.084.10.01' bromide-iodide mixtures have been attrib0 . 2 M B a + + 0 . 8 N "01 10 07 10 06 0.7,0.6 l0:26: 10:22 2.6.2.2 uted by Liebich (18) and Miiller (13) to 0.02 M AIc4+ 0.7,l.Z 0.02 M Th+4 10.07,10.12 the formation of mixed crystals of silver 9.71 -2.9 halides as shown by Kuster (10) and 0.002 0.002 0.02 0 . 8 "NO: -1 8 -4.2 0 . 2 M Ba++ 9,82,9.58 Thiel (16) and studied in detail by 0.02 M AI+++, 0 . 8 N HNO: 9.43,9.65 -517: - 3 . 5 Kolthoff and Eggertsen (8). Clark (S,3),
.. .
E
1 Preaent ad&, Emeryville, Calif.
Shell Development Co..
0
2 minutes stirring after each ml. of AgNOi added.
ANALYTICAL EDITION
June, 1946
359
It was to compare the amperometric and poTable 111. Titration of Mixtures of Iodide, Bromide, and Chloride tentiometric end points as applied t o the titraGalvanomtion of halide mixtures that this study was Normality Normality eter SemiAgNOJ undertaken. Sample of Halide of -4gNOa Medium tivity Used Error The amperometric titration of the individual MI. % 1 0.01N I 9.97, 9.94 -0.3, -0.6 0.1 0.1 N N H I halides hss been described (11). Silver bromide 0.01N B r 0.1 0 . 8 N HNOa 1’2 10.10, 10.01 1.0,0.1 0.01N C1:$!O 9.86, 9.73 -1.4, -2.71 and silver iodide do not cathodically depolarize 2 0.18NI1/2 17.91 -0.5 t,he rotating platinum electrode. Silver chloride, 0.009NBr0.1 0.8NHNOs 1/5 8.92,8.81 -0.9,-2.1 0.003 N C10.05 1/20 6.08,6.01 1.2,0.2 however, causes a cathodic current even in the 3 0.003 N I 0.05 5.86, 5.93 -2.3, -1.2 presence of a large excess of chloride. Therefore 0.018 N B r 0.1 0.8 N HNO, 18.10, 18.12 0.3,0.6 0.009NCl0.1 1/20 8.91 -1.0 it is possible, in principle, to titrate either bromide 4 0.009 N I 0.1 1/2 8.98 -0.2 0.003 N B r 0.05 0.8 N “0, 1.2 6.07 or iodide in the presence of chloride if no gelatin 0.018NC10.1 0.1 %*I. @40 17.75 -1.4 is added to suppress the current due to silver chloride. A chloride end point could then be obtained by adding gelatin and shifting to a The results obtained in titrations of all three halides are shown lower galvanometer sensitivity. in Table 111. As indicated, a 0.1 N ammon‘ medium waa used T~ determine iodide in the presence of bromide, ammonia can ammonia for the for the iodide titration, followed by 0.8 be added to increase the solubility of silver bromide without bromide and 0.1% gelatin for the &loride. The galvanometer affecting the iodide end point. Thus all three halides can be sensitivity of about 0.01 microampre per mm. was cut by means of an Ayrton shunt to l/2 for the Iodide, for the bromide, and successively titrated by first adding ammonia for the iodide ‘/m for the chloride titration to obtain a suitable residual current titration, then adding an excess of acid and titrating the bromide, line in each titration. As a reference electrode, the mercur and finally adding gelatin for the chloride end point. mercuric iodide-potassium iodide cell of Kolthoff and Harris volt us. saturated calomel electrode) was used or the iodide titration in ammoniacal medium, followed by a EXPERIMENTAL saturated calomel reference electrode for the bromide and chloride titrations* The effect of adding multivalent metallic ions as flocculating agents on the titration of bromide in the absence and presence of X o difficultywas encountered except in the bromide titration chloride is shokn in Table 1, The apparatus and experimental of sample 4, when only about 10% of the total halide present was procedure have been described (11). bromide. I n this case, the bromide curve was flat with an inAlthough a trend toward higher results for bromide in the distinct end point. presence of increasing amounts of chloride is noticeable, the In Table Iv, data are s h o w for the analysis of three synthetic flocculating ions in general had no effect on the end point. It halide mixtures of varying composition, made by one worker i s possible that under the conditions of the amperometric titraunknow.rzs and by tion in which the end point is rapidly passed (the entire titration The amperometric method shows promise as a rapid method for takes only 3 to equilibrium is not minutes), mixed t’he determination of halides in mixtures, although it is ndt reached even though the precipitate is colloidally dispersed. always as accurate &S the Slower potentiometric method. In two experiments (Table I), the titration was delayed for 2 minutes after the addition of each milliliter of reagent. Contrary to expectation, the results were lower than for the same titration run rapidly. The presence of five or ten times as much Table IV. Analysis of Unknown Halide Mixtures chloride as bromide led t o a flat titration curve with no bromide Present Found end point. For solutions less than 0.002 N in both bromide and Sample IBrCI1Brc1M g . / 1 0 0 ml. chloride, low results were obtained for bromide. At higher coniUg./loO ml. centrations, a balancing of errors occurs, the “normal” low result 1 15.15 155.1 0.43 15.0 153.8 0.62 2 2.54 73.5 312.0 2.0 76.2 305.6 for the amperometric bromide titration being balanced by a 3 121.8 17.6 121.3 15.3 21.4 21.3 positive error caused by mixed crystal formation. In Table 11, data are given for several titrations of iodidebromide mixtures in the presence of various concentrations of LITERATURE CITED ammonia. It appears that 0.2 N ammonia is too concentrated (1) Behrend, R.,2. physik. Chem., 11, 466 (1893). for solutions of iodide more dilute than 0.01 N . In general, a (2) Clark, W.,J. A m . Chem. SOC.,48,749 (1926). concentration of 0.1 N ammonia is suitable for all but the most (3) Clark, W., J . Chem. SOC.,1926,768. dilute iodide solutions. (4) Dutoit, P.,and von Weisse, G . , J . chem. Phys., 9,578 (1911). (5) Flood, H., Z.anorg. allgem. Chem.,229,76 (1936). (6) Flood, H., and Bruun, B., Ibid., 229,85 (1936). (7) Flood, H., and Sletten, E., 2. anal. Chem., 15,30 (1938). (8) Kolthoff, I. M., and Eggertsen, F. T., J . A m . Chem. SOC.,61, Table II. Titration of Iodide in Presence of Bromide in Ammoniacal
0”:
0”:$& 8:: #f&
8:: g@k
f//”5
3
ytential -0.23
Solution Nor-
mality of 1-
Normality of Br-
Normality of
AgNOs
Normality of NHs
AgXOsVsed
Error
%
.Ul. 001
0.01
0 01 0.001
0001 0 0001
....
9.94 9.94,993,9.94 0.2 9.90,9.96,9.92 0.05 0,l 0.2 9.88,9.83, 9.82 0.001 0.01 0 . 1 9.99,9.93 0.2 9.64,9.58 0.01 0.01 0.1 9.80,9.80.988° 0.0001 0.001 0.05 10.07,9.95 0.01
0.1 0.1
0 1% gelatin added.
0:i
0.1
9 46. 9.31O
0.2
8 77Y
-0.6
-0.6,-0.7,-0.6 -1.0, -0.4, -0.8 -1.2, -1.7, - 1 . 8 - 0 . 1 , -0.7 -3.6, -4.2 -2.0, -2.0, -1.2 0.7, -0.5 -5.4, -6.9 -12.3
6-
1036 (1939). (9) Kolthoff, I. M., and Harris, W. E., IND. ENCI.CHEM.,ANAL.ED., 18,161 (1946). (10) Kuster, F. W., 2.anorg. Chem., 19,81 (1899). (11) Laitinen, H. A.,Jennings, W. P., and Parks, T. D., IND.ENG. CHEM.,AN.AL.ED.,18, 355 (1946). (12) Liebich, C.,dissertation, Dresden, 1920. (13) Maller, E., “Elektrometrische Massanalyse”, p. 115,Dresden, Steinkopf, 1932. (14) Pinkhof, J., dissertation, Amsterdam, 1919. (15) Schutza, H. S., Angew. Chem., 51,55 (1938). (16) Thiel, A.,Z.anorg. Chem., 24, 1 (1900). (17) Zintl, E., and Beta, K., Z.anal. Chem., 74,330 (1928).
Txxs investigation WBB carried out under the sponsorship of the Office of Rubber Reserve, Reconstruction Finance Corporation, in connection with the Government Synthetic Rubber Program.
