Titrations with a Photoelectric Titrimeter M. L. NICHOLS AND B. H. KINDTI, Cornell Unioeraity, I t h , N. Y. Varioun neutralizationn and a precipitation titration have been made with a photoelectrio titrimeter, which ha8 ale0 been applied to the titratios of fluorides and to the determination of the critical concentration of soap solutions.
T
However, with fairly high indicator concentrations, with the ratio of the intensitiea of the two beams near unity, and with a pH near the middle of the indicator range the error is slight. It may be preferable to use as a second filter one wbich transmits in the region where little or no change in the absorbency of the indicator occurs rather than one which transmits in the violet region, aa in the latter case one may have to compensate for large differencee in the transmittances of the filters aa well aa in the responsea of the photocells. Sodium Hydroxide-Hydrochloric Add. In these titrations 20 drops of a 0.1% isohydric solution of bromothymol blue in a final volume of 150 ml., filters with transmittance-band maxima a t 420 and 620 mfi ( 8 ) ,and a Clark and Lubs buffer solution of a pH of 6.75 were used. Titrations were made to determine the dilution limit with strong acids and bases. The solutions and water were boiled to remove carbon dioxide and the titration was completed in an atmosphere of nitrogen. The following averages of several titrations were obtained:
H E authors (f 7 ) have described a photoelectric titrimeter and its principle of operation. Miiller (16) haa shown that compensation for variations in the intensity of the light source is inherent in the circuit used. This instrument haa been applied to several titrations, some of which are very difficult by visual methods,to show ita precision and stability. NEUTRALIZATIONREACTIONS
The spectrophotometric study by Brode ( 8 )shows that an ideal hydrogen ion concentration indicator has two narrow, sharp absorption bands both in the visible and yet separated far enough SO that one does not affect the height of the other. When one absorption maximum decreases the other should increase, and there should be a relatively large change in the height of these maxima in relation to the change in hydrogen ion concentration. Holmes ( f 1 , l b ) has shown that the indophenols satisfy these conditions, but that some are relatively unstable in aqueous solution. Except for the fact that their secondary bands are in the violet region, the phthaleins and sulfonphthalein are satisfactory. The azo dyes have broad bands which affect each other's height and show a definite shift (9)toward the violet with increased pH. Holmes (11 )was the fmt to point out that the maximum degree of alteration per unit change in indicator transformation is obtained by measuring the absorbencies a t the two wave lengths a t or near the maxima of the two absorption bands. Furthermore, the ratio of the absorbencies of the two bands does not vary with indicator concentration, and dichromatism introduces no interference. For all neutralization titrations the metho! of balancing the circuit with a buffer solution and then titrating to this balanced position was wed. The precision of this method depends upon the ability to reproduce a given difference in intensities, which neceasitates a careful control of the indicator concentration. 1 Present a d d r e , General Eleotrio Company, Boheneotrdy, N. Y
'
Normality 0.1 0.01 0.002
NaOH.
HCI
24.33 24.33 24.37
25.00 25.00 25.00
MI.
MI.'
Acetic Acid-Ammonium Hydroxide. As a further indication of the precision, these two solutions were titrated aa above and the average values were compared with the average obtained visually by standardizing the acetic acid against sodium hydroxide with phenolphthalein and the ammonium hydroxide against hydrochloric acid with modified methyl orange. Visual Normslity, Normality, HGHaOr NHtOH 0.084713 0.10325 0.09493 0.10317
Ratio 1.088
Titrimeter HGHaOs, NILOH. MI. MI. 25.05 23.04 25.05 23.01
Ratio 1.088
Boric Acid-Sodium Hydroxide and AnilirrtHydrochloric Acid. The possibility of making a direct titration of very wesk acids and bases waa investigated with these solutions. winn alizarin vellow R and methyl yeliow aa ;he best a v k s b l e indicators. However, these titrations were found to be impractical. The absorption spectra of alizarin yellow R (Figure 1) showed it to be unsatisfactory and the potentiometric titration of aniline (Figure 2) showed that a variation of 40% in the salt concentration would cause an error of about 2% in the equivalent volume of the titrant, assuming the pH a t the equivalence point to be fixed. Therefore, the final volume and sample weight must be constant or volume correctione must be applied. Titrations dth Mixed Indicators. A differential instrument of this type should be particularly well adapted for certain mixed indicators. The spectral transmittance cwves ( 9 ) of equal parts of methyl red and 3;5 4bo 4 b 4 h 5& 5b & 5 6;5 G O 6$5 bromocresol green show that this mixed indiWavelength (mw) cator has the advantage that it exhibite two Figure 1. Transmittancy Curves for Alicurin Yellow R absorption band8 in the middle of the visible
4k
Sbo
7bO
6b
785
'
786
ANALYTICAL CHEMISTRY
portion of the spectrum, whose intensities change in opposite directions at the same time. Modified indicators, however, offer no improvement over the indicators themselves with an instrument of this type. The mixed indicator of 3 parts of bromocresol green and 2 parts of methyl red exhibits a sharp change from green to wine-red at a pH of 5.1. Potentiometric titrations of 0.1 N and 0.01 N sohtions of hydrazine and hydrochloric acid gave the equivalence points a t p H values of.4.9 and 5.2, respectively. When 25.00 ml. of 0.1 N and 0.01 N hydrazine were titrated with hydrochloric acid, using filters of 515 and 620 mg, the following values wwe obtained: “2, All.
Normality
NiHr, MI.
Potentiometrically, the equivalent volume of hydrochloric acid was between 27.10 and 27.20 ml. in both cases. A mixture of 3 parts of cresol red and 3 parts of thymol blue has no significant advantage over eitber with the titrimeter because the alkaline violet color is a single broad absorption band caused by the auperposition of the thymol blue band a t 596 mp upon that of the cresol red a t 572 mp. TURBIDITY TITRATIONS
The phototurbidimetric titration of barium sulfate has been made by del Campo and co-workers ( 4 , 5 )and by Ringbom (go). The single photocell method of operation, in which the galvanometer spotlight is returned to ten divisions from the null position after each addition of titrant, was used. Titration to maximum turbidity is based upon the principle that the addition of each portion of the precipitant results in the formation of some precipitate and an increase in the turbidity of the solution. An increase in the absorption of light takes place until the equivalence point is reached, and further addition of the precipitant decreases the degree of turbidity by dilution. If coagulation is prevented, the equivalence point corresponds to maximum turbidity. Ringbo? gave the optimum experimental conditions for the precipitation of barium sulfate as:
1. The concentration of the sample must be high enough for one drop of precipitant to cause immediate precipitation but low enough for the precipitate to remain in suspension. E’ulfillment of these two conditions is aided by the addition of a miscible organic solvent and a protective colloid. 2. The concentration of precipitant should be such that one drop corresponds to the accuracv of measurement. 3. The thickness of layer ihould be such that the loss in illumination at the end point is 40 to 80%. Ringbom’s procedure was followed using a 0.02994 iM solution of barium chloride and a 0.02979 M solution of sodium sulfate. A 490 mp filter was used to decrease the level of illumination. The sulfate was added to the barium ealt, although no detectable difference was noticed with the reverse procedure. The barium chloride solution was placed in a Wml. titration cell with a 32mm. light path, 35 ml. of water, 25 ml. of 95% ethyl alcohol, and 1 ml. of a freshly prepared 1% solution of gum arabic were added, and the solution was stirred for a few minutes. About 5 ml. of the sulfate solution were added at a rate sufficient to cause the formation of very small particles and then the addition was continued in 0.10 ml. increments a t first and in 0.05ml. increments near the end point. After each addition the solution was stirred for 1 to 2 minutes and the galvanometer light was adjusted, the maximum dial reading corresponding to the equivalence point. The following values were obtained: 0.02994 M
BaCh Used, MI. 5.77 8.13 0.24
NanSOr Observed, Mg. 24.59 25.94 26.49
NaSSOt Calculated,
Deviation,
24.55 20.07 26.53
0.04 0.13 0.04
Ms.
Mg.
The presence of the alcohol causes immediate precipitation of barium sulfate and the gum arabic maintains a finely dispersed precipitate but, even so, Schlieren lines were sometimes observed with 0.02 M solutions and invariably with 0.01 M solutions. Occmionally they appeared a t the start of the titration with 0.03 M solutions if the first 5 ml. were not added rapidly enough. The best results were obtained with the maximum turbidity a t a dial reading between 20 and 35. TITRATION O F FLUORIDE WITH THORIUM NITRATE
Originally this investigation vas begun on the design ( 1 6 ) and application of an instrument to the titration of fluoride with cerous nitrate, using methyl red (18). However, because of the variation of the p H a t the equivalence point and the difficulty of maintaining a constant 80 C. temperature, attention was directed to this titration. Willard and Winter ( 2 4 )introduced a procedure using a lake of zirconium and alizarin red as indicator in 50% alcohol. Armstrong ( 2 ) improved the method by omitting- the use of zirconium and using the lake with thorium. Other workers have confirmed this procedure, although the p H must be controlled carefully. Hoskins and Ferris ( 1 3 ) state that the optimum pH is approximately 3.5 to 50% ethyl alcohol and recommend an equimolar solution of monochloroacetic acid and sodium monochloroacetate in an amount sufficient for a buffer concentration of 0.02 M. They used 0.04 ml. of a 0.0570 aqueous indicator solu-0.37599 . tion in 50 ml. and titrated to the very light pink shade of a matching blank. Later Armstrong ( I ) and Rowley and Churchill (%I)both found the color change sharper in aqueous than in 50% alcoholic solution with a pH from 2.9 to 3.1. Reynolds and Hill (19) confirmed these indicator results &sand found that the concentration of indicator, the fluoride concentration, and the temperature also affected the titration. I l , ,I 1l 1 I II I I I I I I 19 IS 18 20 22 24 ~ 261N2&,130s033i0~ 28 30 , 32 ~ 34 36 ~ 30 ~40 4 2 44 46 46‘ 4 They also reported a small blank in MI.O.1 N HCI Solution aqueous solution which increased with the volume. Approximately the same Figure 2. Potentiometric Titration of Aniline with Hydrochloric Acid O
I
I
V O L U M E 2 2 , N O . 6, J U N E 1 9 5 0
187
The use of dextrin, agar-agar, gum arabic, gelatin, and starch aa a protective colloid was tried. Of these, only starch was wtisfactory and it was found that 10 ml. of a 1%starch solution with 70 ml. of the fluoride solution improved the detection of the rolor change in aqueous solution (23).
I
1
,
1
1
1
I
I
I
,
,
,
Sn 400 425 450 475 500 525 550 575 600 625 650 675 Wavelength (mu)
I
mo
Figure 3. Transniittancy Curves of Sodium Alizarin Sulfonate and Thorium Ideke
Table I. 0.05
,v
Th(NOdc, 111.
4.40 4.50
4.60 4.70 4.80 4.90 5.00 5.10 5.20 4.40
4.50 4.60 4.70
Titration of Fluoride
1-Min. Interval Dial reading Change Aqueous Solutions 99.2 2.0 97.2 3.9 93.3 6.8 86.4 8.5 77.9 8.6 69.3 6.9 62.4 4.7 57.7 3.3 54.4 Alcoholic Solutions 90.5 1.9 97.6 5.0 02.6 10 1 82.5 14.9
4.80
67.6
4.00
54 0
5.00
46.3
A . 10
.5.20
12.4 40.4
13.G 7.7 3.9 2.0
2-Min. I n t e r v a l Dial reading Change 98.8 96.0 90.7 82.1 73.4
2.8 5.3
8.6 8.7
8.0 64.5 58.3 53.6 50.4 98.6 96.5 92.8
6.2 4.7
3.2
2.1 3.7
9.0 83 8 6!I r,
aG. 1 47.5
14 3
13.4 S O
3 .!I
13.6 41.7
1.9
optimum aiialytical conditions were found by Jlatuszak and Brown ( 1 4 ) . From all these investigations it can be concluded that, although the ability to judge the end point can be acquired. it requires considerable experience, the use of a color standard, and much time and patience. For investigation of this titration a solution of pure sodium fluoride containing 0.04505 gram of fluorine per 100 ml., a 0.04845 A: solution of thorium nitrate, both 0.1 and 1.0% aqueous solutions of sodium alizarin sulfonate, and a monochloroacetic acid buffer solution 0.2 d l in both acid and salt of a pH of 3.15 were prepared. The transmittancy curves of both the yellow form of the indicator and the red thorium lake are shown in Figure 3. Accordingly, filters which had transmittance-band maxima a t 420 and 515 nip were selected.
The end point of the titration was determined by observing the maximum change in. the dial readings of one of the slidewires. A suitable uantity of sodium fluoride solution was ajded from a buret to a titration cell of 32-mm. light path. This was diluted to about 70 ml. with water, or a mixture of 40 ml. of water and 20 ml. of 95y0 ethyl alcohol was added. Fifteen dro s o f a 0.1%indicator solution wereadded a n z t h e color was discharged with 1 drop of 0.1 N hydrochloric acid. Then 3.5 ml. of buffer solution were added and the titration was made with 0.05 N thorium nitrate. A fixed time interval of l or 2 minutes was allowed t~ elapse between the ddition of the thorium nitrate and the adjustment of the rheostat.
With the titration in aqueous solution t h e gr:ttiual precipitation of the thorium fluoride in the vicinity of the end point caused a definite drift of the galvanometer. It was found that the use of ethyl alcohol not only causedrapid precipitation but also increased the magnitude of the change in the dial readings a t the equivalence point, which can bc seen from Table I and the curves of Figure 3. Under these conditions the optimum concentration of indicator was found to be 15 drops of a 0.1% solution in 70 ml. of fluoride solution. This is about twice that used by Reynolds and Hill ( 1 9 ) and Matuszak and Brown (14). The use of 10 drops necessitated the diminishing of the light intensity to prevent fluctuations of the galvanometer and 20 drops broadened the maximum in the curve of dial readings versus volume and caused the results to be farther removed from the stoichiometric values, as also found by Reynolds and Hill. On the assumption that the rheostats are wound accurately to the nearest 0.1 dial division, that the reaction will permit a precision of 0.1 division in adjustment, and that the smallest change in the vicinity of the end point is greater than 0.1 division, the precision of the determination should be better the smaller the increments in the volume of thorium nit'rate solution. All these conditions were not satisfied with this apparatus. With aqueous solutions the difference in changes in the dial readings for 0.05-ml. increments near the end point was of the same order of magnitude as the uncertainty in the d i d readings. Although this was not true with alcohol present, still in this case the interpolat.ion of the readings with 0.1-ml. increments gavc sufficient precision and increased the speed of titration. The titration can be made with 0.02 N as well as 0.1 N thorium nitrate solution, although the precision decreases with decreasing concentration. The effect of earying the amount of fluorine was studied to determine if the relationship between the amounts of fluorine and thorium nitrate was lincar over a given range and also whether the results were stoichiometrically exact. The titrations were made both by balancing the circuit with the rheostat in the 420 mfi arm, with the rheostat in the 515 mp arm fixed, and vice versa. In the first case the position of the dial in the 420 mp arm is directly proportional to the ratio of the transmittancy of the 515 mp beam to that of the 420 mp beam. Because thc change in this ratio is assumed to pass through a maximum a t the end point, the change in dial readings also passes through a maximum. I n the serond case the dial readings are inversely proportional to this same ratio, so the change in the reciprocals of the dial readings should pass through a maximum.
A N A L Y T I C A L CHEMISTRY DETERMINATION OF CRITICAL CONCENTRATION OF SOAP SOLUTIONS
Table 11. Determination of Fluoride FF-
Detn.
Added,
Found.
2.26 4.51 6.78
2.20 4.41 6.65 8.85 2.20 4.42 6.64
ME.
NO.
9,Ol 2.26 4.51 6.77
‘ 0 ‘
IlO 2:O
Figure 4.
hlg.
i.0 4fO i.0 i.0 i 0 Milligrams of Fluorine
Aqueous solutions of most soaps exhibit a more or less abrupt change in physical properties over a relatively short concentration range. This phenomenon has been attributed to the formation of oriented soap aggregates, and the concentration a t which it occurs has been termed “the critical concentration for the formation of micelles.” Corrin, Klevens, and Harkins (8) were the first to use the alteration of the absorption spectrum of dye solutions to determine the critical concentration. They employed a visual titration based upon the fact that the absorption spectrum of pinacyanol chloride in aqueous solutions of anionic soaps changes sharply to that characteristic of its solutions in organic solvents over a short range of soap concentrations. Corrin and Harkins (6) used the same technique for the determination of the critical concentrations of anionic soaps with rhodamine 6G and of cationic detergents with Sky Blue FF, eosin, fluorescein, and acidified 2,6-dichlorophenolindophenol. They ( 7 ) also measured the depression of the critical concentration by the addition of an electrolyte. The transmittancy curves for various concentrations of anionic soaps in aqueous pinacyanol chloride solutions (19) show that tlie color change a t the critical concentration is the result of a rise in the intensity of the absorption band a t 480 mp and a simultaneous fall in the intensity of two others a t about 570 and 615 mp. Because no absorption spectra of dyes for the titration of cationic detergents were available and because the use of 2,&dichlorophenolindophenol might introduce an error due to the acid required, Sky Blue FF was chosen.
Error, Mg. -0.06 -0.10 -0.13 -0.16 -0.06
-0.09
-0.13
8-0 9.0 !IO ,
Linearity of Sodium Fluoride-Thorium Nitrate Relationship
The results for varying amounts of fluorine are given in Table
I1 and the linear relationship is shown in Figure 4.
The sodium dodecyl sulfate was recrystallized twice from ethyl alcohol-ether solutions. The n-decyl and n-dodecyl trimethyl ammonium bmmides were prepared by refluxing t h e c o r r v n d m g n-alkyl bromides with alcoholic solutions of tnmethylarmne ($2). h c h soap was recrystallized three times from methanol-ether solutions, and the solvent was removed b vacuum desiccation and stored over pho horns pentoxide. TEe potassium bromide was reagent grade. %he pinacyanol chloride and Sky Blue FF were used as received from Eastman and National Aniline. A stock solution of sodium dodecyl sulfate was pre ared by die solvin 1.900 grams of the soap in 1 liter of 10-6 pinacyanol chlorilie solution. The cationic soap solutions were M individual1 for each determination and a Sky Blue J F was used.
!k
SOE%ZI
The spectral behavior of pinacyanol chloride in various conThe resulta are not stoichiometrically exact. However, this is centrations of sodium dodecyl sulfate, shown in Figure 6, a g r w to be expected because of the high indicator concentration used with that for other anionic soaps (8). Filters with transmittanceand because it was aesumed that the change in trammittancy band maxima a t 490 and 610 mp were used. passes through a maximum a t the end point, whereaa actually it is the changein absorbency or the logarithms of the transmittancy which does. The l second error was calculated and found to be only 0.03 ml., so the main error was due t o the indicator. Within the concentration 4-0-Aqueous solution range studied fluorine can be titrated with 0.05 N thorium nitrate solution with a precision of 10.01 ml. and the same accuracy could be attained by applying an indicator correction or by standardizing the thorium n i t r a t e s o l u t i o n b y t h i s met hod. The method of extrapolating two straight lines to determine the end point waa also applied to this titration. Some results are shown in Figure 5 . This method is more time-consuming than the other, the nearly horizontal line i a not very clearly defined probably because of incomplete precipitation, coagu5 lation, etc., and the results are not aa accurate because of the higher indicator concentration required.
Figure 5.
Titration Curvea for Internetting-Line Method
V O L U M E 22, N O . 6, J U N E 1 9 5 0
789
The titration was made by first placing 65 ml. of the stock solution of sodium dodecyl sulfate containing the proper amount of dye in a 32-mm. titration cell. Then the dye solution (same concentration as in sample) was added in 0.50-ml. increments and the rheostat in the 610 mp arm waa adjusted to balance after a k e d time interval of 1 o or 2 minutes. The maximum change in the position of the rheostat dial was taken as the end point. r h e results in Table I11 show that equilibrium is more nearly approached for 2-minute than for 1-minute intervals, but the titration is empirical and should always be made under the same conditions. These results agree well with the value of 6.02 X lo-* M obtained by Corrin and Harkins(6) by visual titration. 10In addition to the fact that the dye solution itself faded very slowly, the spectrum of the dye in the soap solution (Figure 6) apparently was altered on standing, probably owing to the disFigure appearance of the band a t 480 rnp. Spectra for Sky Blue FF in solutions of cationic detergents were not available (IO), so these were determined for 10-6 M aqueous solutions as shown in Figure 7. An examinstion of these spectra shows that they represent simply a cumulative effect of several close component bands. Nevertheless, it is obvious from these spectra and those determined for several other cationic soaps in solutions of this dye that at the critical concentration a sharp decrease occurs in the in-
Bye solution olme Above crit. conc.
Below crit. conc.
Wavelength (mu)
7.
600 625
650 675 700
Transmittancy Curves of Sky Blue FF in n-Dodec) I Trimethyl Ammonium Bromide
tensity of an absorption band whose maximum is about 630 mp. However, the simultaneous increase in the intensity of any other absorption band is considerably masked. Filters with transmittance-band maxima a t 515 and 750 mp were selected EB the best available. Titrations were made as with sodium dodecyl sulfate, except that the increments of dye solution were 1.0 ml. and all time intervals were 2 minutes. The results are given in Table IV. The slightly lower value for the n-decyl salt, compared to the value of 6.36 X IO-* M of Corrin and Harkins, may be attributed to the lower dye concentration or a slight impurity in the soap. Inasmuch as no determinations of the value of the n-dodecyl salt with Sky Blue FF have been reported, the value obtained must be compared with the value of 1.45 X lo-’ M from the conductometric measurements of Scott and Tartar (28). Just t u in conductometric measurements, the transition occurred over a wider concentration range for the n-decyl than for the n-dodecyl soap. SUMMARY
Above crit. conc. IO-
-*-
Below, offer standing I day
O L d 0 4’rs 560 5;5
5;O 5;5 6 0 6;5 Wovelength (mu)
660 6;
Figure 6. Transmittancy Curves of Pinacyanol Chloride in Sodium Dodecyl Sulfate
Table 111. Critical Concentration of Sodium Dodecyl Sulfate Detn. N o .
Table IV. Detergent n-Decyl n-Dodecyl
1-bfin.Interval 6.09 X 10-1 M 6.09 X 10-1 M 6.12 X lo-’ M
2-Min. Interval 6.18 X 10-8 A l 6.18 X 10-1 h l
...
Critical Concentration of Cationic Detergents Wt. of Detergent, Gram 1.3502 1.2500 1.2200 0.3100 0.3010 0.3000
Dye Soln., MI. 12.7 7.7 7.2 3.1 1.3 1.8
Critical Concn., M 6.19 x 106.19 X 10-2 6.03 X 10-1 1.48 X 10-8 1.47 X 10-9 1.46 X 10-2
As Mika (15) and others have pointed out, although photometric measurements can be used to good advantage for the rapid, routine detection of color changes difficult to find with the naked eye, they increase only the precision, not the accuracy of the result. Thus, in acid-base titrations the accuracy of the result is still susceptible to the errors associated with indicators (salt error, protein error, etc.) as well as factors affecting the pH of the solution directly-e.g., temperature. The simplicity of the method of titrating to a previously balanced position of the galvanometer and the ease and rapidity with which it may be executed make it especially suitable for routine analyses. It has at least another distinct advantage in that the end point need not be a t or near the mid-point of the transition range of the indicator. This method has been applied to various sorts of acid-base titrations with a precision of about 0.01 ml. in 25 ml. The speed and accuracy of a phototurbidimetric titration have been shown with the titration of sulfate. The average deviation of three titrations from the correct value of about 25 mg. of sodium sulfate was 0.07 mg. and the maximum deviation was 0.13 mg. Although it is somewhat empirical, the method of titrating to the maximum change in the dial readings has been suggested for those titrations t o which the other methods definitely cannot be applied or as a substitute for the longer intersecting-line methods. A study haJ been made of a number of optimum analytical con-
ANALYTICAL CHEMISTRY
190
ditions for the photoelectric titration of fluoride with thorium nitrate using sodium alizarin sulfonate. It w&sfound, in spite of previous reports to the contrary, that the end point of this titration is apparently improved by the presence of alcohol. I n the range of 2 to 10 mg. of fluoride the average deviation in the volume of thorium nitrate was t0.01 ml. Finally, the apparatus has been applied to the visual titration technique, suggested by Harkins, for the determination of the critical concentration for micelle formation in soap solutions by the spectral change of a dye. LITERATURE CITED
Armstrong, W,D., IND.ENG.CHEM.,ANAL.ED.,8, 384 (1936). Armstrong, W.D.,J . Am. Chem. SOC.,55, 1741 (1933). Brode, W.R., Ibid, 46,581 (1924). Campo, -4. del, Burriel, F., and Garcia Escolar, L., Anales aoc. esGn. fls. 2/ qutm., 34, 829 (1936). Ibid., 35,41 (1937). Corrin, M. L.,and Harkins, W.D., J . Am. Chem. Soc., 69,679 (1947). Ibid., p. 683. Corrin, M. L.,Klevens, H. B., and Harkins. W. D., J . Chem. Phy8., 14,480(1946).
(9)Fortune, W.B., and Mellon, M. G., J . Am. Chem. SOC.,60,2607 (1938). (10) Harkins, W. D.,private communication. (11) Holmes, W.C.,J . Am. Chem. Soc., 46,627 (1924). (12)Ibid., 47,2232 (1925). (13) Hoskins, W.M.,and Ferris, C. A., IND. ENG.CHEM., ANAL.ED., 8,6(1936). (14) Matuszak, M. P.,and Brown, D. R., Ibid., 17,100 (1945). (15) Mika, J., 2.anal.Chem., 128, 159 (1948). (16) Mtiller, R. H., IND.END.CHEM.,ANAL.ED., 11, 1 (1939). (17) Nichole, M. L.,and Kindt, B. H., ANAL.CHEM., 22, 781 (1950). ENG.CREM., ANAL.ED., 15, (18) Nichols, M.L.,and Olsen, J. S., IND. 342 (1943). (19) Reynolds, D.S.,and Hill, W. L., Ibid., 11, 21 (1939). (20) Ringbom, A., 2. anal. Chem., 122, 263 (1941). (21) Rowley, R. J., and Churchill, H. V.,IND.ENQ.CHEM., ANAL. ED., 9, 551 (1937). (22) Scott, A. B., and Tartar, H. V., J . Am. Chem. Soc., 65, 692 ( 1943). MetalluTgia, 36,346 (1947). (23)Stross, W., (24) Willard, H. H., and Winter, 0. B . , IND. END.CHEH.,ANAL.ED., 5, 7 (1933). RECEIVEDJanuary 3, 1950. From a thesis preaented by B. H. Kindt to the Graduate School of Cornel1 University in partial fulfillment of the requirements for the degree of doctor of philosophy.
Amperometric Titrations with Ferrocyanide EDWARD L. NIMER, RANDALL E. HAMM, AND GARTH L. LEE University of Utah, Salt Lake City, Utah Conditions for the amperometric titration of zinc and indium with ferrocyanide are described. The method for zinc depends upon titration in a supporting electrolyte of ammonium acetate that is 1.7 molar at an applied potential of -1.4 volts vs. the saturated calomel electrode. The precipitate formed approaches Zn*Fe(CN)ein composition. The method for indium depends upon titration in a supporting medium of potassium chloride that is 0.1 molar at an applied potential of -0.75 volt vs. the saturated calomel electrode. The precipitate formed has a composition of approximatelyInr[Fe(CN)&. Complex forrnation is suggested as the reason for precipitation of normal ferrocyanidesrather than double salts.
T
HE ampemmetric titration of ferrocyanide with zinc ion in 0.2 N hydrochloric acid solutions at a potential of -1.2
volts us. the saturated calomel electrode was reported by Spalenka ( 6 ) . Spalenka also reported that the reverse titration could be performed in an ammoniacal solution containing ammonium chloride, but that the results were somewhat erratic. The well-known method of titrating zinc with ferrocyanide, as adapted for internal indicators and for potentiometric titration, gives a precipitate with a composition of K*Zna[Fe(CN)s]*. The method, reported in this paper, of amperometrically titrating zinc in ammonium acetate supporting electrolyte gives a precipitate approaching ZnlFe(CN)#in composition. The titration of indium with ferrocyanide by a potentiometric method was reported by Bray and Kirschman ( 1 ) . A similar method using diphenylamine as an internal indicator in acetatebuffered solution was reported by Hope, Ross, and Skelly ( 8 ) . The composition of the precipitate given by these investigators was KIns [Fe(CN)6]4. The composition of the precipitate which was found in this investigation was In4[Fe(CN)e]p when the precipitation was performed in 0.1 molar potassium chloride solution. APPARATUS
The titration vessel waa a 150-ml. beaker which was closed with a rubber stopper through which were four holes for entrance of the dropping mercury electrode, the buret tip, the saturated potassium chloride-agar bridge to the saturated calomel electrode (S.C.E.), and the tube to furnish a stream of hydrogen or of nitrogen for removing dissolved oxygen.
The circuit used for making voltage and current measurements waa similar to the manual polarograph of Kolthoff and Laitinen (9), although some of the titrations were checked, reading the scale manually, on a Sargent Model XI1 olaro aph. Polarograms were run using the Sargent Model $11 poKrograph. The buret used for titration was of 1Gml. capacity. REAGENTS AND METHODS
Stock Solutions. Pure zinc metal was weighed out, dissolved in sulfuric acid, and diluted to volume in a volumetric flask. Pure indium metal waa weighed out, diesolved in hydrochloric acid, and diluted to volume in a volumetric flask. Recrystallized potassium ferrocyanide was dried at 105' C. for 24 hours, pulverized, again dried for 24 hours, weighed, dissolved, and diluted in a volumetric flask to prepare a 0.1 molar solution. A small amount of sodium carbonate waa added as a preservative. Further dilutions of these stock solutions were prepared by standard volumetric techniques. Supporting Electrolytes. The ammonium acetate and the potassium chloride used as supporting electrolytes were reagent grade chemicals, found by experiment to give no reduction currents in the range of potential in which they were to be used. Analytical Methods. I n addition to being prepared as accurately aa possible, the solutions also were carefully analyzed by volumetric and gravimetric methods. A standard ceric solution was prepared from primar standard grade ammonium hexanitratocerate and used to checi the concentration of the ferrocyanide solution. A thiosulfate solution waa standardized by using the ceric solution and waa then used to check the stock zinc sulfate solution by the method of Lang (6). In this way the ratio of the concentration of the zinc to
