GILBERTIY.SMITHAND FREDERICK XELSON
471 $
molecule a t the dropping mercury electrode. The values of ~ S (Table E I X ) are certainly unreliable. However, the values of n, as calculated from the average of three experimental values of D are definitely 2 , 4, 0 and 8 for butyl nitrate, ethylene glycol dinitrate, glycerol trinitrate and pentaerythritol tetranitrate, respectively. From the polarographic data and the number of electrons involved in the reduction of one molecule (as calculated from Ilkovic equation with experimentally determined values of D) of a polynitrate ester, it appears that each ON02 group in the molecule reduces in a 2-electron step. This has been shown to give an alcohol and nitrite ion as the products
+
RON02 2eROH20-
+
+RO-
+ SO2-
+ROIl + 013-
Thus, for glycerol trinitrate the following over-all
[CONTRIBUTION FROM TIIE OAK
L’ol. 70
reduction probably takes place a t the dropping mercury electrode in neutral or slightly acidic solutions. C3Ha(OS02)3
+ Be- + 31120 +
C?H,(OH)$
+ 3\‘02- + 30FI
It may be postulated that in alkaline solution the secondary ON02 group is readily attacked by thc large concentration of hydroxyl ion and thus n more complicated process is involved. Polynitrate esters are certainly less resistant to attack by hydroxyl ion a t room temperature than a simple nitrate ester such as n-butyl nitrate. n’ork now in progress on the alkaline decomposition of glycerol trinitrate may lead to an understanding of the reactions a t the dropping mercury electrode in alkalitic solution. ClIIUA T241iC. CA1,ITORTIA
RIDGENATIONAL LAnORATORY,
CHEMISTRY D I V I S I O N ]
The Polarography of Thallium(II1) BY GILBERTW. SMITHAND FREDERICK NELSON RECEIVED APRIL 3, 1954 The polarography of thallium(II1) in chloride solutions has been investigated, Two diffusion current plateaus were found in TI(II1) polarograms; the first plateau can be interpreted to correspond t o the reduction TI(II1) to Tl(1) and the second to the reduction TI(II1) to Tl(O)(Hg). The magnitude of both diffusion current plateaus is appnrently controlled by the rate of diffusion of TI(II1). The diffusion current constants for the two-electron and three-electron reductions Tl(II1) t o TI(1) and TI(II1) t o TI(O), were found to be 3.83 f 0.03 and 5.72 f 0.08, respectively, in 0.6 M HCl corresponding t o a diffusion coefficient, D TI(II1) of 1.00 X 10-6 cm.2 sec.-I. The diffusion current constant for TI(1) in 0.6 M HCI was determined from separate experiments with Tl(1) solutions and found to be 2.65 zt 0.03, corresponding to D TI(1) = 1.91 X c m 2 sec.-l, in reasonably good agreement with previously reported values. Polarograms from solutions containing both Tl(1) and TI(II1) were found t o be suitable for analytical determination of TI(1) and TI(II1). The half-wave potential of the “Tl(I)/(O) wave” for Tl(II1) solutions was practically the same as that for Tl(1) in the same supporting electrolyte.
Although the polarography of thallium(1) has been studied e x t e n s i ~ e l y , ~apparently -~ only one note, by Hughes and discusses the polarography of thallium(II1). These authors found that for perchloric acid solutions, polarograms of Tl(II1) had a “reproducible and well defined” diffusion current plateau which merged with the anodic dissolution current of mercury. This plateau is followed by a wave which on the basis of its half-wave potential was identified with the TI(I)/(O) wave. The first diffusion current plateau was thus assumed to result from the reduction of Tl(II1) to Tl(I), and the second plateau was assumed to result from the reduction of Tl(II1) to TI(0). Since mercury dissolution occurs a t a more negative potential than the TI(III)/(I) reduction potentialfia polarographic “half-wave” for the Tl(III)/(I) reduction is not observed. No values were given for the diffusion current constants of TI(II1). The present investigation was primarily concerned with the measurement of diffusion current (1) This document is based on work performed for the Atomic Energy Commission a t the Oak Ridge National Laboratory. (2) J . J. Lingane, THISJOURKAL, 61, 2099 (1939). (3) L. Meites, ibid., 73, 4257 (1951). (4) “Bibliography of Polarographic Literature,” Leeds and Northr u p Co., Philadelphia, Pa., 1950. (5) G.K.Hughes a n d N. S. Hush, Azrsf. J . Sci., 10, 184 (1948). (6) T h e potential of t h e Tl(I)/(III) couple is -1.25 v. (us. Hz electrode); W. M. Latimer, “Oxidation Potentials,“ Prentice-Hall, Inc., New York, N. Y.,2nd Ed., 1962.
constants of Tl(II1) from the Tl(III)/(I) and the TI(III)/(O) reduction currents in chloride solution to test the possibility of utilizing polarographic methods for the determination of Tl(1) and Tl(II1) in Tl(II1)-(I) mixtures. Half wave potentials of the TI(I)/(O) wave in Tl(II1)-(I) mixtures were also measured and compared with previously reported values for Tl(1) solutions.
Experimental M TI) were prepared h y TI(1) stock solutions ( f a . dissolving recrystallized nitrate-free’ thallous chloride i n 0.1,O.6antl1 0 3 1 H C l o r 0 . 1 MKClcontaining5 X 10-3.11 HCI. Tl(II1) stock solutions were prepared from Tl(1) solutions by oxidation with chlorine gas. Thallium concentrations were determined by iodometricaJ titration after rcmoval of the excess chlorine with phenol. Polarograms were recorded in an air conditioned room a t 25 i 1’ with a Sargent Model XXI recording polarograph. A polarographic cell having a capacity of about 20 ml. was used in conjunction with a KCI-agar bridge and a saturated calomel reference electrode. Oxygen (or excess chlorine) was purged from the solutions with nitrogen purified by passing through a train of ammoniacal cuprous chloride or, in some cases, through copper turnings at 500’. When the solutions contained chlorine, the dropping mercury electrode was introduced after the chlorine had been expelled, t o (7) By brucine sulfate test-F. Feigl, “Spot Tests,” Elsevier Press, New York, N. Y., 3rd Ed., 1946. ( 8 ) I . M. Kolthoff and E. B. Sandell, “Textbook of Quantitative Inorganic Analysis,” Rev. Rd., The Macmillan Co., Xew York, N. Y., 1947. (9) C. W,Sill and H. E. Pctrrson, Anal. Chem., 21, 1208 (1019).
Sept. 20, 1954
THEPOLAROGRAPHY
avoid oxidation of the mercury. To obtain reproducible diffusion currents with TI(II1) solutions, it was necessary to begin measurements as soon as the dropping mercury electrode was introduced since it was found that Tl(II1) slowly reduces upon contact with the mercury pool in the cell. Maxima were not usually observed; however, when they did appear gelatin (0.01%) was added as suppressor. Half wave potentials of the Tl(I)/(O) waves were usually i) os. potential E (corrected taken from plots of log i / ( i d for iR drop). The current i was obtained a t several points along the wave after stopping the bridge t o obtain a constant current. The total limiting current, i d , was measured a t -0.7 v. v s . S.C.E. and both i and id were corrected for residual current. At each point on the wave, the applied potential was measured with a Leeds and Northrup potentiometer. Since, as mentioned, Tl(II1) slowly reduces on contact with mercury, measurements were made rapidly and second polarogram was usually run immediately to determine whether the bulk T1( 111) concentration had remained essentially constant.
OF
4715
THALLIUM(III)
difference indicates that the diffusion coefficient of
TI(1) is considerably larger than t h a t of Tl(II1).
+4.5
-
Results and Discussion A. Polarograms of Tl(II1) and Tl(I).-A typical polarogram of TI(II1) in 0.6M HC1 is shown in
+3.0 ?
a
E
3 .
v
+I
3b 1.5
e
0 Fig. 1 (curve A). The reduction current of TI(II1) merges without break with the anodic dissolution RESIOUAL CURRENT current of mercury near zero volts (us. S.C.E.), I rises to a constant current, ~ ( I I I ) , I ( I (corrected ) for residual current), and then rises again near -0.45 v. vs. S.C.E. to a second constant current, ~ c I I I ) , ~ o ) . 0 -0.2 -0.4 -0.6 -0.8 T h e current ~ ~ I I I ) / ~may I) be assumed to result from Volts vs. S.C.E. the reduction of Tl(II1) to Tl(1) and the current Fig. 1.-Polarograms of thallium(II1) and- (I) solutions &III)/(n) can be assumed to represent the reduction 25', 0.6 hl HC1; A, 3.33 X M TI(II1); 13, 3.33 X of Tl(II1) to TI(0) (Tl-amalgam). The current, 10-4 TI(I). ~ ~ I I I ) / ( Iwas ) , measured over the range 0.02 to 1.0 X M Tl(II1) and found to be proportional to Diffusion current constants, I(I~I)/(I) and Tl(II1) concentration. I(I)/(o) for Tl(II1) and Tl(1) solutions, respectively, The diffusion currents, &III),(T) and &III)/(o) , were calculated in the usual way, by substituting were measured a t -0.26 and -0.70 v. vs. S.C.E., values of &III)/(I) and i(1)/(0) in the equation I = respectively. It was found by averaging the results id/cm2/8t'/0 (cf. IlkoviE equation, i d = C107nD'/~. for several TI(II1) solutions in 0.6 M HC1 that the ~ m ' / ~ where t ' / ~ n, D , c, m and t have their usual ratio, ~ ( I I I ) / ( o )to ~ ( I I I ) / ( I )was , 1.51 0.01. This definition.I0 For 0.6 M HCI solution, I(III)/(I) = value is in accordance with the assumption t h a t 3.83 f 0.03, I(III)/(o) = 5.72 f 0.08, and I(I)/(o) = these currents result from three-electron and two- 2.65 0.03." Each of these values is based on electron reductions, respectively, and that the dif- measurements of currents from a t least nine sepafusion rate of Tl(II1) controls the magnitude of both rate polarograms of pure TI(II1) or Tl(1) solutions. diffusion currents. This observed value for the Diffusion coefficients corresponding to these difratio is in excellent agreement with the expected fusion current constants were calculated with the value of 1.516, which takes into account small dif- IlkoviE equation and found to be 1.00 and 1.91 X ferences in current arising from differences in charcm.2 set.-' for TI(II1) and Tl(I), respectively. acteristics of the capillary electrode a t the above B. Polarograms of Tl(II1)-Tl(1) Mixtures.potentials, ;.e., mz/atl/s= 2.161 and 2.184 m.g.'/' Polarograms were run on several solutions containsec.-'/z, a t -0.26 and -0.70 v. vs. S.C.E., respec- ing varying amounts of Tl(II1) and TI(1) in 0.6 M tively. The drop times, t , a t these potentials were HCI. The total thallium concentration (0.332 X 4.34 and 4.42 seconds, respectively. M ) was held constant while the concentration I t was shown that by varying the mercury height of TI(1) was increased from 0 to 0.315 X M. &III)/(I) was very nearly proportional to the square The results are summarized in Fig. 2. root of the mercury pressure (expressed in cm. of At a potential -0.27 v. vs. S.C.E. the current mercury) over the range cu. 25 to 79 cm. of mercury. ~(III)/(I) observed for solutions containing TI(II1) This observation and the fact that the TI(III)/(I) and Tl(1) results from the reduction of TI(II1) to diffusion current plateau was very nearly parallel to Tl(1) and as shown in Fig. 2 is proportional to the residual current may be taken as evidence that TI(II1) concentration. The total current a t -0.7 v. the reduction of Tl(II1) to Tl(1) occurs practically vs. S.C.E., however, may be considered to represent instantaneously a t the dropping mercury electrode. the sum of two reduction currents, ~ ( I , / ( o ) and For comparison a polarogram of TI(1) a t the (lo) I. M. Kolthoff and J. J. Lingane, "Polarography, Vol. I, same thallium concentration (3.33 x M is also Interscience Publishers, Inc., New York. N. Y., 1952. (11) This value is in good agreement with literature values (a) I = shown in Fig. 1 (curve B). It can be seen that 61, &I)/(o) corrected for residual current in curve B is 2.69 for 0.1 N HCI, J. J. Lingane and I. M. Kolthoff, THISJOURNAL, (1939); (b) I 2.71 in 0.1 N KCI, F. Buckley and J. K. Taylor, about 40% larger than the corresponding reduction J825 . Research Nail. Bur. Standards, 34, 106 (1945); and ( c ) I 2.693 in This large 0.613 N HCI, L. Meites, T H IJOURNAL, current ~ ( I I I ) / C O-) ~ ( I I I ) / ( I of ) curve A. ~ 13, 4257 (1951).
f
*
*
-
-
4716
ROBERTE. LANGAND CECILV. KING
Vol. 70
tures may be carried out rather easily, by utilizing the linear relationship expressed in equation I . Certain reducible substances, such as Fe(II1) which also produce a constant reduction current in the region 0 to -0.45 x 7 . ZIT. S.C.E. must of course be taken into consideration in analytical applications. C. Half-wave Potential of Tl(1).-The most precise measurements of the Tl(1) '(0) half-wave potential (LIS. S.C.E.) reported in the literature appear to be those by Lingane? who found E = -0.455 0.003v. in 0.10MKCI and v. in 1.0 M KCl. The average of three deterinmations in this Laboratory in 0.10 Ji KCl was I< = 0 0.2 0.4 0.6 0.8 1.0 0.338 st 0.003 and a single measurement in 0 1 li Fraction of Tl(1). KCI, 0.005 d i HC1 gave E 1% = -0.462, in reacoiiable agreement with Lingane's value for 0.1 31KC1 Fig. 2.--l)iffusion currents of Tl(II1) and Tl(1) in 0.6 M HC1; total TI = 0.332 X -11,25', no maximum In 1.0 31 HC1, El/? was found to be -0.191 x-., considerably more negative than Lingane s \ d u e suppressor. for 1.0 ?I1 KC1. This more negative value probably i(m)/mj. By substituting the measured current, is due, in part, to the fairly large liquid junction po~ ( I I I ) / ( I in ~ , the relationship, ~ ( I I I ) I ( O = ) cy ~ ( I I I ) / ( I ) tential which exists between l II HC1 and the satuwhere N = a proportionality constant, the Tl(1) rated calomel reference electrode. current, i , ~/,o) ) may be calculated by the expresHalf-wave potentials of the wave appearing in sion the polarograms of pure Tl(II1) solutions were found to be essentially the same'? as those for pure i m c 0 , = ;tot - a i ( r r r ) , ( r ) (1) where i t o t = total current corrected for residual cur- T1(I) solution in the same supporting electrolyte. rent a t -0.7 v. ZJS. S.C.E. Figure 2 also shows a For example, El/, = -0.403 i 0.003 and -0.493 plot of i l 1 ) / ( o ) calculated by equation 1 (using the + 0.003 in 0.10 -11KCI, 0.005 -11HC1 and 1 0 experimental value N = 1.51) vs. fraction of Tl(1). HC1, respectively. Each measurement was based I t can be seen that i c ~ , /is( o proportional ) to Tl(1) on at least three separate determinations with pure A11 waves were tested for reversiconcentration. Xt small fractions of Tl(I), i c ~ ) /Tl(II1) ~ ~ ~solutions ~ values were, of course, subject to greater errors be- bility, plotting log (i - i I I I j I t I ) ) ( ~ ( I I I ~ I ~-o ) i) cause of difficulty of measuring small differences LIS. E,,,, and the resulting slopes were found to ngrec within +O.OW .'1 with the expected Sernst dol)e, between two relatively large reduction currents. similar series of measurements with solutions con- -0.030v. a t 23". Acknowledgment.-The authors are indebted to taining a total thallium concentration of 1.00 x Dr. Kurt A. Kraus for helpful criticisnis and advicc ;21gave essentially similar results. It can be concluded from these experiments that during the course of the work. diffusion current constants for Tl(1) and Tl(II1) (12) This wave is thus associated with the Tl(1) to 'II(0) rccluclion are independent of the Tl(I)/Tl(III) ratio in Tl(1)- as pointed out previousl! h > Hughes a n d Hush (ref 5 ) Tl(II1) mixtures and hence analysis of such mix- OAK RIDGE,TEYY
*
[CONTRIBUTION FROM THE
DEPARTMEW OF
CHEhfISTRY, NEW Y O R K U N I V E R S I T Y ]
Transference Numbers in Aqueous Zinc and Cadmium Sulfates1 BY ROBERTE. LXNG.4ND CECILV. KING RECEIVED FEBRUARY 26, 1954 Transference numbers of zinc and cadmium ions in aqueous solutions of the sulfate? a t 23" have l ~ c e ntletcrmitied from the electromotive force of concentration cells with lead amalgam-lead sulfate electrode5 The results for bine ion agree n ith those found by Purser and stoke^,^ and values for cadmium ion are sitnil,ir
Transference numbers of the sulfate ion in aqueous zinc and cadmium sulfates were calculated by Wolten and King2 from the e.m.f. of concentration cells with zinc and cadmium amalgam electrodes. The measurements with zinc sulfate were repeated by Purser and stoke^,^ who pointed out that the experimental e.m.f. values of m'olten and King were not adequate for the purpose, and that the mathematical treatment and interpreta-
tion mere improper. The present authors agree with these conclusions, and have made new measurements in order to establish transference numbers in both systems with certainty. The values obtained for zinc ion agree closely with those found by Purser and Stokes, and those for cadmium ion are similar. Transference numbers of the metal ions were determined directly with the aid of the cells
(1) From a P h D thesis submitted t o the Graduate School of N e w York University b y Robert E Lang (2) G i V Wolten a n d C V King, THISJ D U R U A L , 71, 570 (1010) n) r P P,WM pnrl R H $ t r l l r P q , S M , 73, m n r i ~ s i )
two phase
Pb-Hg
I PbS04(s),MSOl(nz) ,
?rISO,(ref), PI)SOl(s) i Ph-FIg ti\ o p h n v
MSO4(rcC indicates n constant molality ior a l l
