Determination of Halogen in Organic Compounds by Automatic

---

idcntical with that based upon the liberation of hydrogen and reduction by the released hydrogen. The precision of the method niight be improved by using a more dilute hydride solution and a smaller excess, as well as more refined methods of measurcnient, drlivery, and titration, including more precise hyringes or niicroburets. Titration by weight instead of by volume might be desirable but would complicate the technique. However, for determining the number

of double bonds in a compound, the reported precision is satisfactory, particularly because of the ease with which replicate values may be obtained. ACKNOWLEDGMENT

The author nishes to thank H. C. Craig for carrying out the experimental work and H. C. Craig and 2. F. Smith for carrying out some of the preliminary experiments on which the method is based.

LITERATURE CITED

(1) Chaney, A. L., IND.EKG. CHEK, AXAL.ED. 10, 326 (1938). ( 2 ) I.evy, G. B., Murtaugh, J. J., Rosenblatt, hl., Zbzd., 17, 193 (1945). (3) Smith, I). M.,Mitchell, J., Jr., Billmeier, A. XI., ANAL.CHEY. 24, 1847 (1952). RECEIVED for review February 26, 1958. Accepted June 11, 1958. Division of Analytical Chemistry, 133rd Meeting, ACS; San Francisco, Calif., April 1958. Meeting-in-Xiniature, Analytical Group, S o r t h Jersey Section, ACS, South Orange, T.J., January 27, 1958.

Determination of Halogen in Organic Compounds by Automatic CouIo metric Titration 0. E. SUNDBERG, H. C. CRAIG, and J. S. PARSONS Research Division, American Cyanamid Co., Bound Brook,

b An automatic titrator, utilizing the coulometric principle, i s used for the determination of halides, following sodium peroxide Parr bomb fusion. The method involves precipitation of the halide b y generation of silver ions from a silver anode. The electricity required i s measured b y an integrating motor incorporated in the instrument, and the titration is automatically terminated at the end point b y the Beckman titrator control unit. The aulomatic coulometric titration technique affords several advantages over the generally used Volhard titration method for halogen analysis; it i s direct, obiedive, and timesaving. It minimizes the human error and yields greater precision.

N.J.

nitrate reagcnt and facilitates automatic operation of the rquipment. The Lingane nicthod m s applied to the automatic titration of halides in solut'ions obtainrd from Parr bomb fusions. The method n-as d s o adapted with some success to thc titration of mixtures of bromidr and chloride. During thr coursc of this work, an automatic coulometric titrator was built. T h r instrumentation was simplified over previous models by incorpomting a low-inertia integrating motor into the circuitry. This motor eliminates the need for a precise constant current supply and a precisr timing device. Details of t,hc instrumentation havr been described (,4). EXPERIMENTAL

D

halogen in organic compounds is iniportant in a number of laboiatoriee. Dwomposition of the organic sample by a sodium peroxide Parr bomb fusion, follon-ed by :t T'olhard titration, has been the most satisfactory technique for routine analys w However, colored iiiipuritirs mag ohscwe the end point. or a fading end point niay result if tht, precipitated si1vc.r halide is not filtered off. A method involving c1irrc.t electrometric tit,ration n.ith silver ion is advantageous. c,liminating thc need for filtrat'ion of the silver halide. The automatizing of the titration also saves tinir. Recently. Lingane ( I ) shon-rd that automatic coulometric titration of small amounts of halide with electrolytically generated silver ion was feasible. Coulometric generation of silver ion eliminates the need for standard silver ETERVISlTIOS Of

1842

ANALYTICAL CHEMISTRY

Apparatus. il photograph of the simplified automatic coulometric titrator used is shown (Figure 1). From left to right. ai(' the electrolysis eel1 assembly, the integrating motorcounter unit. t h e Beckman Automatic Titrator Rlodel K. and the ~ o supply unit. -4 block diagram of tlie automatic coulometric titrator is presented in Figure 2 . A simple unst,abilized current of 250 ma. is provided from a direct current sonrce (selenium rectifier bridge). The quant,ity of electricity is measured in terms of motor counts by the integrating motor 11-hich integrates current and time. Silver ion is generated in the electrolysis cell from a silver wire anodc. The potential of the solution is controlled by a Beckman titrator unit equipped with a silver-calomel electrode system. An agar-agar bridge with 0.1-Ysodium nitrate is used brtween thc calomel electrod(, and the halide solution. The antiripstor con-

trol circuit of the Beckman titrator, automatically regulates the addition of proportionally smaller amounts of silver ion a s the vicinit,y of the end point is reached. As the approach t o the end point potential is signaled by the silver indicator electrode, t'he anticipator circuit initiates a series of momentary off and on operations, stopping the titration a t the exact equivalence potential setting and thereby avoiding possible overshooting of the end point. Materials. The silver anode consists of a 3-nini. diameter wire of pure (99.9 %) silver (Baker R- Co., Inc.) of which a 6-inch, spiral-shaped length is exposed t o the electrolysis solution. The silver wire leading out of the solution is enclosed in a glass tube and is sealed at, the lower end n-ith dpiezon wax (IT-100, James C. Biddle Co.). A coarsr frit'ted glass tube comprises t h r cathode cwmpartment. The silver indicator consists of a I/*-inch length (exposed) of 3-mni. diameter silver iT-iresealed in a glass tube n i t h Apiezon wax. A new silver anode should be installed in the cell usually after 200 to 250 determinations. ~ r The U-shaped agar bridge connecting the saturated calomel electrodi. vessel with the electrolysis cell is filled with 47, agar-agar in 0.1S sodium nitrate. I t is usable for several months, provided the ends of the tube are kept immersed in solution. il solution of approximately 0.11- sodium nitrate is used in the vessel containing the saturated calomel electrode. The solution for testing tlie perforniance of the titrator consists of 140 ml. of specially prepared nitrattl-nitric- acid reagent to which is added 10 nil. of standard 0.1-Y hydrorhloric acid and 300 nil. of 3A alcohol (5 gallons commercially pure methanol added to 100 gallons of 190 proof ethyl alcohol). The

+

\

I Figure 1.

Simplified automotic coulometric titrator

nitratenitric acid reagent is prepared by dissolving 20 grams of C.P. sodium nitrate and 3 ml. of concentrated nitric acid per 100 ml. of solution. About 150 ml. of the nitratenitric acid solution is mixed with 300 ml. of 3A alcohol for use in the anode compartment, an additional volume of the nitratenitric acid solution is kept in a reservoir for subsequent addition to the cathode compartment. The titrating medium consists of a solution obtained from the sodium peroxide Parr bomb combustion; it is subsequently neutralized with excess nitric acid and diluted t o 150 ml. If the halide under analysis is chloride, 300 ml. of alcohol are then added to decrease the solubility of the silver chloride so that the rate of potential change at the equivalence point will be high enough for a precise titration. Alcohol is not required for determination of bromide or iodide. If bromide and iodide are present, hydrazine sulfate must he added to the alkaline solution t o reduce a n y bromate or iodate which may have formed during the fusion. The automatic titration step requires approximately 6 minutes per meq. of halide. Sample Preparation. T h e sample is first subjected t o t h e usual P a r r bomb fusion technique, using a threecomponent mixture of sample, sucrose, and sodium peroxide. T h e quantity of sodium peroxide used should he such t h a t t h e amount of nitric acid required for acidification will not exceed 25 ml. Sample sizes ranging from 0.05 t o 0.25 gram were used in t h e determinations. The contents of the bomb are dissolved in a minimum amount of water in a 600-ml. beaker, acidified with 25 ml. of concentrated nitric acid, and diluted t o 150 ml. If chloride is present, 300 ml. of 3A alcohol are added to decrease the solubility of the silver chloride. If bromide and iodide are

Figure 2. Black diagram of autamatic caulometric titrator present, 350 mg. of hydrazine sulfate are added to the dissolved sample, prior to acidification and dilution t o 150 ml. The solution is then subjected to automatic coulometric titration. A blank is run on the reagents. Automatic Coulometric Titration. The Beckman titrator is allowed a 5minute warm-up period. After the zero check is made, t h e current from the power supply is adjusted t o about 250 ma. and t h e pH-millivolt dial is set t o t h e previously determined end point potential for t h e system under analysis, according t o the experimentally determined values described below. The anticipator setting is adjusted, the counter reading taken, and upon completion of titration, the end point reading recorded. CALIBRATION

Adjustment of Anticipator. T h e anticipator on t h e Beckman should be adjusted so t h a t t h e current switches off and on about 50 times near the end point. An automatic counter on the instrument indicates the number of starts and stops. A minimum of 50 counts is necessary t o ensure t h a t the titrator will not overshoot t h e end point. A count of 300 or above is excessive. Testing Performance of Titrator.

T h e solution for testing the performance of t h e titrator is prepared from t h e n i t r a t e n i t r i c acid reagent, standard 0.1N hydrochloric acid and 3A alcohol and titrated. The increase in t h e counter reading should he 37.0 + 0.2. A blank is run on the reagents, and t h e correction in counts applied t o the calculations. Calibration of Low-Inertia Integrating Motor. The calibration factor of the low-inertia integrating motor, expressed as milliequivalents per count was obtained by using Faraday's law in titrations of known amounts of halide. I n t h e fundamental method, a constant current is passed through a fixed resistor for a n accurately measure d time and t h e motor counts reciirded. The calibration factor is obt ained from the equation Meq. per count =

A more practical method of motor calibration involves the automatic coulometric titration of a solution containing 10 ml. of 0.1N hydrochloric acid in 140 ml. of a specifically prepared sodium nitratenitric acid reagent to which is added 300 ml. of 3A alcohol. The calibration factor is then calculated as follows: Meq. per count = ml. 0.W HCl

x

0.1000N

No. of motor counts

The calibration factors obtained by these two methods agreed within three parts in 1o00, thus serving as a check on the current efficiency during the electrolysis. CALCULATIONS

Calibration factor, milliequivalents per count Meq. per count = ml. 0.1N HC1 x 0.1000N motor counts lo

o'looo

37.05

=

= 0.02699

37.05 equals the experimentally determined number of motor counts required to electrolyze 1 meq. of an ion, according to Faraday's law. Calculation for chlorine % c1 = (N,

35.46 - N d x 0.02699 x wt. of sample ( g . )

lono x loo

N L = number of motor counts at start of titration.

N z = number of motor counts at end of titration.

Calculation for bromine VOL. 30, NO. 11, NOVEMBER 1958

1843

Table 1.

Determination of Chlorine, Bromine, and Iodine

Per Cent Found, av. Theory CHLORIXE 17.46 17 53 37.89 37.91 58.48 58.62

KO. of

Compound

Detns. 13 13

2,4-Dinitrochlorobenzene

a-Chloroacetamide 1,3,5-Trichlorohenzene

7

BROMINE 44,89 39,62 39.75 25.77 25.84 72.32 72.46

n-Bromosuccinimide p-Bromobenzoic acid Bromobenxanthrone 2,4,6-Tribromophenol

12 11 11 4

44.88

4Iodobiphenyl Tetraiodophthalic anhydride

10 7

45.18 77.57

II.

Table

yo C1 Found

-

Coulometric 40. if3, 40.72

40.74 Tetrachloro phthaloni53.14, 53.00 trile Ar. 53.07 Research sample A 11.02, 11.24, 11.15 Av. 11.14 15,48 Research sample B Chlorinated copper phthalocyanine A chlorinated copper phthalocyanine B Chlorinated copper phthalocyanine C Chlorinated copper phthalocyanine D

47.07, 47.16, 47.21 Av. 47.15 0.42, 0.33 Av. 0.38 2.12, 2.03 Av. 2.07 0.73, 0.82 .4v. 0.78

=

- Ni) X 0.02699

X

wt, of sample (g.)

79.92

-

loo0 x 100

Calculation for iodine % I = ( N 2

-

126.91

A',) X 0.02699 X __ IOoo wt. of sample (g.)

x

100

DEVELOPMENT OF AUTOMATIC COULOMETRIC TITRATION PROCEDURE

Titration Curves of Simple Halide Systems. The determination of t h e

equivalence point potentials under exact experimental conditions must first be ascertained-either by manual or autoniatic titration of known amounts of halide ion. T h e manual titration yields a characteristic titration curve, whereas t h e automatic titration method serves as a n over-all calibration of the apparatus. Curves were obtained by manual titration of 0.5 meq. of chloride, bromide, and iodide, respectively, in a sodium nitrate-nitric acid media (140 ml. of specially prepaied nitrate-nitric acid

1844

ANALYTICAL CHEMISTRY

\-olhard

40.58, 41 04, 40 GO,

Av.

(-4'2

fO.ll f0.11 10.05

f0.13 fO.18

f O .18 szo.09 f O .19

*0.08

Comparison of Chlorine Values Obtained b y Coulometric vs. Volhard Titration

Sample Tetrachloroanthraquinone

% Br

IODIXE 45.14 77.89

Std. dev.

40 52

40.69 53 34, 53 39, 53 15 53 30 11 37 Av.

15 32, 15 18 15 25 46 59 0 18, 0 16 0 17 2 00, 2 17 2 08 0 53, 0 61 0 57

Average, Dif, $0 05

-0 23 -0 23 $0 23 +O

56

+o

21

-0 c1

-to

21

solution). Exact concentrations of halide ion \?-ere obtained by adding 10 ml. of standard hydrochloric acid, sodium bromide, and potassium iodide. I n the chloride titration, 300 ml. of alcohol were added, giving a total volume of 450 nil. The bromide and iodide titrations were made in 300-ml. volumes in the absence of alcohol. The electrolysis of each halide was carried out a t approximately 250 ma., using the silvercalomel electrode system. The equivalence point potential for chloride was $255 mv. The corresponding motor count reading n-as 18.7,which agrees well with the 18.5 counts theoretically required for 0.5 meq. of chloride. The rate of potential change near the end point was about 6 mv. per tenth of a count. The equivalence point potential for bromide was +I75 mv., with a corresponding motor count reading of 18.3. The rate of potential change near the end point was about 16 mv. per tenth of a count, which is considerably better than that obtained for chloride. The equivalence point potential for iodide was about +20 mv., and the rate of potential change near the end point was 40 mv. per tenth of a count. Of the three halides, iodide gives the best

sensitivity. I n each case, the number of observed motor counts agreed well with the theoretical value for 0.5 meq. of the halide. Symmetrical titration curves were obtained for chloride, bromide, and iodide in the presence of the supporting electrolyte described. The addition of 350 mg. of hydrazine sulfate was found to have no effect on the bromide and iodide titration curves. However, excessive use causes interference by reduction of Ag+ ion. Per cent recovery of chloride by automatic coulometric titration was investigated to eliminate any error due to incomplete combustion during Parr bomb treatment. Known concentrations of chloride were added to solutions obtained by sodium peroxide Parr bomb fusions of sucrose; the solutions were acidified, diluted with water and 3A alcohol, and titrated at +240 mv. (z's. S.C.E.). An average recovery of 99.8% chloride was obtained. The slope of the titration curve at the end point may he increased by increasing the amount of halide being titrated, or by decreasing the volume of solution. However, in the Pari bomb procedure, an aqueous volume of a t least 150 ml. is desirable. Experimental Results. Table I summarizes the results for chlorine, bromine, and iodine obtained from a series of purified halogen compounds by automatic coulometric titration. In the titration of chloride (alcoholwater solution) the equivalence point potential was set a t +a40 mv. (2s. S.C.E.). For the first three compounds listed in Table I, the standard deviation for a single determination ranged from 10.05 to f O . l l % chlorine. The electrolysis of bromine was performed in an aqueous solution to an equivalence point potential of f170 niv. Table I indicates good agreement between the average found values and theory in four purified bromine compounds. The precision is indicated by a standard deviation for a single determination of 1 0 . 0 9 to =k0.18% bromine. The data listed in Table I also show the accuracy and precision of the automatic coulometric titration technique when applied to the determination of iodine in two representative compounds. The electrolysis was carried out in aqueous solution to an equivalence point potential of 0.0 mv. Table I shows a standard deviation for a single determination of 10.08 to k 0 . 1 9 ~ oiodine. Comparative values for chlorine obtained in several organic compounds by automatic coulometric titration us. the Volhard procedure are shown in Table 11. The determinations were made on solutions obtained from a Parr bomb fusion. The average difference between the two methods was +O.l%

absolute for the coulometric method. Copper phthalocyanine green samples gave higher chlorine values by the coulometric method. Because these samples contain copper, it was thought that copper might interfere with the coulometric technique. A titration of a known chloride with 80 mg. of copper present gave 1o0.4Y0 recovery of chloride, indicating that the effect of copper was not serious. Results for chlorine obtained by automatic coulometric titration in a series of research samples are listed in Table 111. The per cent difference between theoretical and experimental values for chlorine ranges from -0.2 t o +0.3%. Results and Discussion of Halide Mixture Systems. The possibility of analyzing mixtures of halides by automatic coulometric titration was investigated. Lingane ( 2 ) discusses the theoretical aspects involved in the potentiometric titration of mixtures of chloride, bromide, and iodide. The analysis of mixtures of these three halides is based on the differences in the solubility products of the respective silver halides. According to Lingane, the analysis of mixtures of chloride and bromide is more difficult than the other binary halide mixtures because of the relative closeness of the silver chloridesilver bromide solubility products. Hon-ever, this halide mixture appears to be the most important from a practical viewpoint. The equivalence potentials for a halide mixture containing chloride and bromide ions in a 1 to 1 ratio were established by manual titration of 0.5 meq. of bromide ion and 0.5 meq. of chloride ion in an aqueous-alcohol supporting electrolyte. The equivalence point potential for bromide in this mixture is +60 mv., and that for chloride is +240 rnv. The effect of the presence of chloride ion on the bromide titration curve was clearly demonstrated. In the absence of chloride ion, the bromide equivalence point potential is +170 mv. The rate of potential change for the first break, in the vicinity of the bromide end point, is 1.4 mv. per tenth of :t count. This gradual slope places a limitation on the precision, because the potential setting of the Beckman titrator cannot be reproduced better than about &2 mv. Table Is’ shows the results from titrating two different mixtures of bromide and chloride ions. In each case, halide mixtures equal to 1 meq. total were titrated coulometrically. The supporting electrolyte consisted of a water-alcohol media. The automatic titratoi was preset a t a potential of +60 mv., which is the indicated equivalence point potential for bromide ion in a 1 to 1 bromide-chloride mixture. The titration is automatically terminated a t this potential, giving the number

I n the range of bromide to chloride ratios investigated (1 to 9 and 9 to I), the equivalence point potential for bromide varied from about +40 to 80 mv. The observed potential for chloride remained constant a t +240 mv. Data in Table V show that considerable error will result in the bromide determination unless the relative concentrations of bromide and chloride are known. Such information would indicate the proper equivalence point potential setting for the system under consideration, so that more reliable bromide values should be obtained. Lingane (3) points out that mixtures of halides cannot be titrated too accurately because the silver halides form mixed crystals or solid solutions which result in coprecipitation errors. The changc of end point with widely different ratios of chloride and bromide has recently been pointed out by Przybylowicz (6)’

of motor counts equivalent to bromide ion concentration. The potential is then set to +240 and the chloride ion determined automatically as before. As shown in Table IV, a 1 to 1 mixture of bromide to chloride can be titrated x i t h reasonable accuracy as indicated by per cent recovery and standard deviation. However, a 1 to 9 bromide to chloride ratio indicates high recovery and poor precision for bromide; the chloride results, however, appeared to be somewhat better. I n view of this, another method for establishing the equivalence point potentials was studied. A suitable end point potential for the bromide end point in known mixtures of chloride and bromide may be determined empirically. Such end point potentials are given in Table V, and some data on results in such mixtures are given in Table IV. Known mixtures of bromide and chloride ions were titrated with equivalent quantities of electrolytically generated silver ion, and the resulting potential was determined. I n each case, the total halide ion concentration was equal t o 1 meq. or 37.05 motor counts (by previous calibration). Results in Table V indicate that the equivalence point potential of a mixture of bromide-chloride ion depends on the relative concentration of the two ions. Table 111.

Sample

+

SUMMARY

The automatic coulometric titrator described has been used in this laboratory for over two years in halogen analyses of several thousand research samples covering a wide spectrum of organic compounds. The determinations were primarily for elemental

Determination of Chlorine in Research Samples b y Automatic Coulometric Titration

Empirical Formula

yG Dif.

13.5

11.5 7.04

7 04, 6 . 9 5 17 5 11.7 7.16 8.78 9.75 9.99

$0.14 -0.2 -0 2

2

3 4

70 c1

Found 12.8 13.7 22.5 17.4. 17.3

1

Theory 13.0 22.8 17.2 17 1 7.14 17.3

5 6

-

;I

9 10

17:O

9.58 9.70 10.0

11

12

Table IV.

+0.2 -0.2 $0.3 -0 15 +0.1

-0.12 -0.20

-0.05 +O.Ol

Titration of Mixtures of Bromide-Chloride (Total 1 meq.)

Per Cent

Ratio of Br-/Cl1:l

1:9 Table

V.

Ratio of Br/Cl 1/9

2/8 3/7 4/6 5/ 5 6/4 7/3 8/2 9/1

Detns. 6 8

Av. recovery of Br-

loa. 5

102 5

Std. dev. 0 48 6 80

rlv. recovery of C1-

100 4

100 0

Std. dev. 0 84 0 76

Dependence of Bromide Equivalence Point Potentials in BromideChloride Ratios (Total 1 rneq.)

Theoretical Yo. of Counts Br c1 3.70 33.30 7.40 29.60 11.10 25.90 14.80 22 20 18 50 18.50 22,20 14.80 25.90 11.10 29.60 7.40 3.70 33.30

Obsd. Equiv. Total No. of Point Potential Counts Found for Br, Mv. 37 02 40 36.97 52 36.95 19 37.00 51 37.02 58 37 04 60 37.14 48 37.11 70 37.26 80

VOL. 30, NO. 1 1 , NOVEMBER 1958

1845

chlorine, but bromine and iodine were also determined successfully. A standard deviation of 0.3 to 0.47, relative (iO.05 to 0.19% absolute) mas obtained for chlorine, bromine, and iodine by this procedure. Further work is in progress for determination of halogen in binary and ternary mixtures. The automatic coulometric titration technique for halogen determination is superior to the Volhard method because it provides a direct and objective titration that is unaffected by such factors as the presence of colored im-

purities which may be in the sample or result from .the Parr bomb fusion. The method has also its timesaving merit, and, finally, the coulometric technique affords greater precision than that obtainable by Volhard titration. ACKNOWLEDGMENT

The authors wish to express appreciation to Charles Maresh for encouraging this investigation. They also wish to thank Jean Haurand for assistance in the preparation of the manuscript.

LITERATURE CITED

(1) Lingane, J. J., -45.4~. CHEX. 26, 622

(1954).

(2) Lingane,

J. J., “Electroanalytical Chemistry,” p. 100, Interscience, New York, 1953. (3) Ihid., p. 104. (4)Parsons, J. S., Seaman, b7.,Amick, R. >I.,AXAL.CHEX.27,1954 (1955). (5) Przybylowicz, E. P., Ibid., 28, 2028 (1956).

RECEIVEDfor review February 21, 1958. Accepted July 7 , 1958. North Jersey Section Meeting-in-Miniature, ;1CS, January 1957.

Determination of Cadmium in Zinc Concentrates and Other Zinc-Rich Materials Anion Exchange Procedure SlLVE KALLMANN, HANS OBERTHIN, and ROBERT LIN Research Division, Ledoux & Co., leaneck, N. J.

b Cadmium forms a strong anion complex with iodides of the type Cdld--. Zinc and other elements with which cadmium is frequently associated do not form stable iodo complexes and, therefore, are not retained on Dowex 1 resin. A method is proposed which takes advantage of the retention of cadmium iodide on a strong base anion exchange resin while other elements with which cadmium is associated in concentrates and alloys are quantitatively removed. Cadmium is finally eluted with 3N nitric acid.

T

most important source of cadmium is zinc sulfide concentrate containing small quantities of the cadmium sulfide mineral, greenockite. The accurate determination of cadmium in zinc minerals, various concentration products, and alloys is, therefore, of considerable importance. It still represents one of the more difficult analytical operations because of lack of variable valency states of cadmium and because gravimetric and volumetric methods are subject to interference from many other elements, particularly zinc, with which cadmium is usually associated (3). iinion exchange appears to be the most promising approach with both cadmium and zinc forming complex anions HE

1846

ANALYTICAL CHEMISTRY

of the type Zn(Cd)C14--(Br4--)(14--). According to several investigators (1, 7 , a), the order of stability of these anions is ZnC14-->ZnBr4-->Zn14-and Cd14-->CdBr4-->CdC14--. Taking advantage of thegreater stability of cadmium iodide in a mixed sulfate-iodide medium, Baggot and Willcocks (1) separated micrograms of zinc from large quantities of cadmium. Hunter and Miller (4) first absorbed zinc and cadmium in a chloride medium on Amberlite IRA-400, chloride form, thus separating the two elements from various other elements forming no chloroanions or complexes not adsorbed in a 2 5 hydrochloric acid medium. Khile they extracted the zinc with hydriodic-nitric acid solution, their investigation was not extended t o cover the elution of the cadmium. Kallmann, Steele, and Chu ( 6 ) adsorbed zinc and cadmium on Dowex 1 in a chloride medium, extracted the zinc n-ith sodium hydroxide, then eluted the cadmium with nitric acid. K h e n the method n as internally evaluated by determining cadmiuni in a great number of production samples, several analysts observed that the 8% sodium hydroxide eluent caused six-elling of the resin, thus favoring the formation of channels and air pockets. If channels formed, they naturally prevented proper contact of resin and eluent, thus requiring excessive quantities of sodium hydroxide for the elution of the zinc and causing erratic cadmium results.

IODIDE MEDIUM

Because of the difficulties evperienced irith the sodium hydroxide medium, an attempt was made to discover a superior zinc eluent. As cadmium iodide is more stable than zinc iodide (1, 7 , 8), the iodide medium appeared most promising. It was also hoped that the separation of the cadmium from the zinc could be carried out simultaneously with the removal of various other elements found in zinc material. Subsequent tests indicated that Fe++ is not retained by Dowex 1 in a dilute sulfuric acid medium containing potassium iodide. Liken ise. manganese, aluminum, calcium, magnesium, nickel, cobalt, and most other elements n-ith 15-hich cadmium usually i- aqsociated, did not form stable iodo complexes in 0 . 7 5 s sulfuric acid, containing 50 grams per liter of potassium iodide, and therefore, 17-ere not adsorbed. Cu++ n-as reduced and precipitated by the iodide t o Cu+, u-hich was dightly soluble in potassium iodide. Xs tests shom-ed that soluble cuprous iodide is retained by Don-ex 1. prior remoi-a1 of the copper x i t h nietallic iron was indicated. This step can be advantageously combined with the remoral of the lead as sulfate and also assures the quantitatii-e remoral of silver, gold, bismuth, arsenic, and antimony which either would form insoluble iodides or nould be partially retained by Dowex 1 (Table I). As u-as expected from the