CONCLUSIONS
For the instrument, sample containers, and the phosphor system used, the optimum amount of scintillator solution is 10 to 15 ml. Slight deviations from any established volume within this range will cause negligible error in final calculated radioactivities. It should be possible to measure the selected volume with a graduated cylinder rather than a calibrated pipet. For the internal standard technique of analysis, the ideal ratio of internal standard radioactivity to sample radioactivity would be infinite. Practically, a ratio of 10 to 1 will produce counting efficiencies with about the same expected error as that attributed to the measurement of sample plus internal standard. Experimentally this value was found to be *1.4%. Coincidence losses are not significant below a count rate of 300,000 c.p.m. Careful use of microliter pipets will not introduce a significant error The effect of optical differences in the counting vials described in this paper is insignificant.
Quantities of nonquenching sample of
at least 1 ml. can be used (in 15 ml. of scintillator solution) without causing significant loss in counting efficiency. The standard deviation of any single determination of a tritiated hydrocarbon is =k2.0’%’,. Some variables that might affect counting efficiency were not evaluated in this paper. Among these are the effect of dissolved oxygen in the scintillator solution and sample (6) and the effect of voltage-gate settings on the Packard liquid scintillation spectrometer. Preliminary investigations of discriminator settings indicate that counting efficiencies might be increased from 15 or 16% to about 2570 by changing the gate setting from 10 and 50 volts to 5 and 60 volts. The increase in efficiency comes principally from the larger number of pulses included by reducing the lower gate setting from 10 to 5. For reasons of reliable instrument stability, i t was felt that the 10 setting on the lower discriminator would tend t o minimize the error caused by occasional
increases in line noise or by more gradual increases in amplifier tube noise. LITERATURE CITED
(1) Bel!, C.. G., Jr., Haye:! F. S . , “Liquid Scintillation Counting, p. 166, Pergamon Press, London, 1958. (2) Bennett, C. A., Franklin, N. L., “Statistical Analysis in Chemistry and the Chemical Industry,” p. 21, Wiley,
New York, 1954.
(3) Davidson, Jack D., Feigelson, Philip, Intern. J. Appl. Radiation and Isotopes 2, NO. 1, 1-18 (1957). (4) Kallman, H., Furst, hI., Phys. Rev.
79,857 (1950). (5) Pringle, R. W., Black, L. D., Funt, B. L., Sobering, S., Ibid., 92, I582 (1953). (6) Reynolds, G. T., Harrison, F. B., Salvini, G., Ibid., 78, 488 (1950). ( 7 ) Wilson, E. B., Jr., “An Introduction to Scientific Research,” p. 273, McGraw-Hill, New York, 1952. (8) Wilzbach, K. E., J . Am. Chem. SOC. 79, 1013 (1957). (9) Zeigler, C. A., Chleck, D. J., Brinkerhoff, J., ANAL.CHEM.29, 1774 (1957). RECEIVEDfor review July 10, 1959. Accepted January 18, 1960. Work supported in part by the Department of the Army, Ordnance Project TB5-0010C.
Separation of Rhodium from Iridium by Copper Powder G. G. TERTlPlS and
F. E. BEAMISH
University o f Toronto, Toronto, Ontario, Canada
b A method for the separation of rhodium from iridium uses copper as a selective precipitant for rhodium in 1 .ON hydrochloric acid. Both milligram and microgram quantities, 10 mg. to 50 y, of the two metals can b e separated and subsequently determined. Losses of either rhodium or iridium to iridium or rhodium precipitate are negligible and in accordance with the filtrate losses of rhodium and iridium.
T
HIS RESEARCH arises from the lack of convenient precipitants for the determination of rhodium and iridium in a solution of both metals. Until recently, no satisfactory quantitative method of separation was available. Three acceptable methods for the separation of rhodium from iridium and the subsequent determination of these metals have been recorded (4, 6, 14). The titanium(II1) chloride (6) procedure is suitable for macro amounts, although the subsequent removal of excess reagent by cupferron is difficult. The Berman-McBryde method (4) is considered more effective than the antimony (14) method for separating
486
ANALYTICAL CHEMISTRY
microgram amounts of these metals, although there is some difficulty in recovering iridium from the anionic exchanger. The present report deals with the application of copper powder as a selective precipitant for rhodium in 1.ON hydrochloric acid solution of both rhodium and iridium metals in a range from 10 mg. to 50 y and the dissolution of the metals and their separation from copper by cation exchange procedure in forms convenient for gravimetric or colorimetric determination. The optimum conditions for a n acceptable separation of rhodium from iridium in a solution of both metals were studied. These included the volume of the solution, the acidity in the range from 0.1 to 2.ON, the heating time from 15 to 150 minutes, and the temperature from 60” to 9 5 O C. The amount of copper was kept in excess. APPARATUS,
REAGENTS, AND SOLUTIONS
STANDARD
Beckman Model A2 glass electrode p H meter and Beckman Model B spectrophotometer were used for all p H readings and absorption measurements, respectively.
Ion Exchange Columns. Large Column. A borosilicate glass tube 23 cm. in inside diameter was joined t o a draining tube about 5 mm. in inside diameter. The resin bed was 16 t o 17 em. in depth. Small Column. The borosilicate glass tube n.as 1 em. in inside diameter. The draining tube was 4 mni. in inside diameter. The resin bed iTas 5 to 6 cm. in depth. Exchanger. The exchanger m-as Dowex 50X cationic resin in sodium form, of 20 to 50 mesh, supported in the column on a small plug of glass ~vool. The exchanger was regenerated just before use with 3N hydrochloric acid until the eluents n-ere colorless and free of iron and copper, as shown by a spot test of potassium thiocyanate and rubeanic acid, respectively. The excess acid was removed from the exchanger by washing with water until the eluent was neutral to litmus paper. Chlorination Apparatus. A Vycor tube 19 mm. in inside diameter and 50 em. in total length tvith a n elongated end (outlet) t o enable joining to a rubber tube, Jvas heated by a n electric furnace of 9.53 em. in length. Liquid chlorine (Canadian Industries, Ltd.), Thiobarbituric acid (660-Eastman Organic Chemicals) , highest purity.
2-Mercaptobenzothiazole, practical (Eastman Kodak Co., No. P5470), was recrystallized from ethyl alcohol. Thiobarbituric acid and 2-mercaptobenzothiazole solutions were freshly prepared as needed and filtered before use. Cupric oxide powder reagent (General Chemical Go.) was reduced in hydrogen a t 400" to 420" C., and ground in an agate mortar, and the copper used fresh. Stock rhodium solution. Sodium rhodium chloride solution containing 10 ml. of concentrated hydrochloric acid per liter was standardized gravimetrically by thiobarbituric acid (5) and found to contain 1.01 mg. of rhodium per ml. Stock iridium solution. Ammonium iridium chloride solution, containing 10 ml. of concentrated hydrochloric acid per liter, was standardized gravimetrically by 2-merc:tptobenzothiazole (2) and found to contain 0.92 mg. of iridium per ml. REDUCTION OF IRIDIUM
Latimer (9) reports that the redox potentials for the couples of copper, rhodium, and iridium are, respectively:
+
-0.337volt 3eE o = -0.44 volt Ir 6 C1- = IrC16--3eE" = -0.77 volt IrCl6--- = IrC16-- eE" = -1.017 volts Cu = C u + + 2e- E" Rh 6 C1- = RhC16-'--
+
=
+
+
+
+
Although the recorded potential for the iridium system indicates that copper would serve as a precipitant of iridium metal, it has been the experience in this laboratory that the expected reduction does not take place under certain conditions, whereas it does for rhodium. Aoyama and Watanabe ( I ) , separating platinum from iridium by copper powder in 0.1N hydrochloric acid, report that spectrophotometric analysis of the obtained platinum and iridium did not yield traces of iridium in platinum or platinum in iridium, although the values of iridium recovered were low. Iridium is not erisily precipitated by copper, which is in agreement with that in the titanous chloride separation of rhodium (6) and analogous methods using vanadous ion (ti), chromous ion ( l a ) , and antimony metal ( 1 4 , which are stronger reducing agents than copper. When magnesium and zinc metal, commonly used as strong reducing agents, were applied at3 precipitants of iridium metal in 1.OAV hydrochloric acid, they precipitated only part of iridium, even though the solution was boiled up to 1 hour and subsequently warmed on a steam bath for 24 hours. Also, hydrazine sulfate precipitated an appreciable amount of iridium in
Table 1.
Iridium Reduced by Copper
Iridium Added, mg. 9.15
Reduced, 32
y
30 33 4.60
1.84 0.459 0.459 0.459 0.092 0.092 0.092 0,046) 0.046
1.8" 1.4"
All these were done under the same conditions as described in the recommended separation of rhodium from iridium. The results are listed in Table I. FILTRATE LOSSES OF RHODIUM
Aliquots of the stock rhodium solution were similarly treated to determine the recovery of rhodium following its precipitation by copper powder. The excess copper and the rhodium were removed by filtration. The copper in the filtrate was eliminated by a cationic exchange procedure and the rhodium was determined colorimetrically as described below. Average losses of rhodium for the recorded amount of precipitated metal mere as follows : Rhodium
0.3"
After reduction by copper, the three residues were combined and analyzed for iridium.
Taken, mg. 10.06 ~.
5.03 2.01 0.502 0.0503
Losses,
y
3 . 5 =t0 . 5 3.5 f 0.5 2.3 f0.7 0.8
0.25
SEPARATION OF RHODIUM FROM IRIDIUM, AND DETERMINATION OF THESE METALS
sodium hydroxide solution after heating.
It is possible that the resistance to reduction by iridium is the result of unusually stable dissolved complexes. This view is supported by the general analytical behavior of iridium-cg., the difficulty of arriving a t complete precipitation by hydrogen sulfide, etc. Peculiarly, the reduction of iridium is encouraged by the presence of palladium, The problem seems to be kinetic and its interpretation would provide a distinct contribution to the chemistry of the platinum metals. Because iridium is theoretically reduced by copper, it was necessary to learn the exact limits of iridium metal precipitated under the conditions of the recommended separation. Experiments were carried out to determine the amount of iridium reduced by copper powder by the process described below for the separation of rhodium from iridium. After the treatment of the iridium solution by copper, filtration, and n-ashing of the residue according to the procedure for the separation of rhodium from iridium, the residue n-as dissolved by aqua regia and filtered through the same filter crucible (filtrate A). The residue m-as chlorinated. The dissolved chlorination product was filtered and the filtrate was combined with filtrate A. These combined filtrates were evaporated, converted to chlorides by warming n i t h concentrated hydrochloric acid, and evaporated to dryness. The residue was dissolved with water. Copper was eliminated in the resulting solution through the large and small cationic exchangers and iridium was determined colorimetrically in the final effluent.
Separation. Aliquots of both rhodium and iridium stock solutions were placed in a 150-ml. beaker, made up t o 30 ml. in 1.ON hydrochloric acid, and heated on a hot plate (80" t o 85" C.). When micro amounts of the metals were to be separated, the total volume was 20 ml. After the addition of freshly reduced copper, the beaker was tightly covered with a watch glass and the solution was heated to gentle evolution of bubbles (91" to 93" C.) for 1 hour with swirling a t intervals of 1minute; the volume was kept constant by the addition of water. When copper was added, the solution instantaneously changed from dark red to light pink, iridium being reduced to a lower state. An additional 200 mg. of copper powder were added after heating the solution for 30 minutes to prevent any possible redissolution of metallic rhodium and also to complete the reduction of rhodium. The warm solution was filtered through an A2 filter crucible of 3-ml. capacity by suction and decanted three times; the residue was transferred into the filter crucible by a water stream and then washed 18 times lvith water into the crucible. Recovery and Determination of Rhodium. The filter crucible containing the residue was placed in the original beaker and treated with 24 ml. of aqua regia (3 to 1) on a steam bath until reaction ceased, the beaker being well covered u i t h a watch glass. The resulting solution was filtered through the original filter crucible and the beaker was washed several times by a m-ater stream to tranefer any residue into the filter crucible (filtrate A). The residue in the filter crucible was dried in a steam cabinet, covered with ground sodium VOL. 32, NO. 4, APRIL 1960
0
487
Table II.
Copper Added,
Mg.
+ 200 350 + 200 550
250
+ 200
Separation of Rhodium from Iridium
Rhodium Added,
Indium Added,
Rhodium Recovered,
Mg.
Mg.
Mg.
Mg.
10.06
9.15
5.03
4.73
10.07 10.10 5.00 5.02 4.98 4.99 4.98 5.02 2.00 1.99
9.11 9.10 4.71 4.75 4.72 4.73 ... 4.76 1.81 1.84 1.84 1.84 1.82 1.81 0,448" 0.456" 0.451" 0,455" 0,455= 0 .452a 0.092 0,089 0.088 0.089 0.088 0.089 0.045 0.045
2.01
1.84
...
200
+ 200
Iridium Recovered,
0.502
0.459
0.1006
0.092
1.99 2.02 2.01 0.499" 0.485~ 0 . 496a 0.497 0,498 0.497
...
...
0.0987 0.0981 0.0987 0.0503
0,046
d.Q474 0.0476
0.046
Analyzed colorimetrically by aliquots.
chloride, and chlorinated for 7 hours at 700' C. (7). The solution, resulting from the dissolution of the chlorination product with 0.1N hydrochloric acid, n a s filtered through a glass-wool stem and the filtrate was combined with filtrate A. These combined filtrates were evaporated almost to dryness, warmed a few times with concentrated hydrochloric acid, diluted with water, and finally treated again with concentrated hydrochloric acid, .and taken to dryness to expel any nitric oxide and convert the metals to chlorides. The salts were dissolved in 50 ml. of water, the pH of the solution was adjusted to 1.3 to 1.5, and copper was removed by passing the solution through an appropriate cation exchange column (10) at the rate of 1 drop per second. The large and small columns were washed with 550 and 170 ml., respectively, of dilute hydrochloric acid (pH 1.3 to 1.5). The rhodium in the effluent solution was determined as follows: When milligram amounts of rhodium were to be determined, 3.5 ml. of concentrated hydrochloric acid were added to each sample and rhodium was determined gravimetrically by thiobarbituric acid (6) ; when microgram amounts were to be determined, the effluent volume was reduced, transferred to a 30-ml. beaker, and evaporated to dryness in the presence of 2 ml. of 2% sodium chloride solution. Organic matter was destroyed by digestion with concentrated nitric acid and 30% hydrogen peroxide and the metal was converted to chloride by concentrated hydrochloric acid. Rhodium was determined colorimetrically by 488
ANALYTICAL CHEMISTRY
Table 111.
Detection of Rhodium and Iridium
Rhodium in Iridium Precipitate Iridium Rhodium Rhodium Added, Added, Detected,
Mg.
Mg.
Mg.
9.15
10.06 10.06 0 . 014a 10.06 4.60 5.03 5.03 0.012' 5.03 1.84 2.01 2.01 0,009" Iridium in Rhodium Precipitate Rhodium Iridium Iridium lidded, Added, Detected, Mg.
Mg.
AIg.
10.06
9.15 9.15 0,0243 9.15 5.03 4.60 4.60 0.016" 4.60 2.01 1.84 1.84 0.010" a Three iridium or rhodium precipitates of each set were combined and analyzed for rhodium or iridium, respectively. stannous chloride in 2147 hydrochloric acid ( I S ) at a wave length of 470 mp (11). For 2 to 100 y of rhodium, the volume of the solution and cell width were 25 ml. and 5 em., respectively, n-hereas for over 100 y, the volume was 50 ml. and the cell width 1 em.
I n all cases, a reagent blank TT-as carried through the entire procedure and standards Kere used directly from the stock rhodium solution. Results are listed in Table 11. Recovery and Determination of Iridium. T h e filtrate and washings from t h e rhodium precipitation, contained in a 400-ml. beaker, were evaporated t o dryness and the residue was dissolved in 50 ml. of water. Copper was eliminated in t h e resulting solution b y the cation exchange procedure described above in t h e recovery and determination of rhodium. The effluent from the small exchanger was evaporated to dryness in the presence of 2 mi. of 2% sodium chloride solution to remove mineral acids which interfere in the subsequent precipitation of iridium by 2-mercaptobenzothiazole (2). Iridium was determined as follows: When milligram amounts of iridium were t o be determined, the residue was dissolved in 30 ml. of water and iridium was determined gravimetrically in the resulting solution by 2-mercaptobenzothiazole (2) in the presence of one or two glass beads to prevent bumping. When microgram amounts were to be determined, the residue was dissolved with water and the resulting solution was transferred to a 30-ml. beaker. After the organic matter had been destroyed by digestion of concentrated nitric a peroxide evaporation to dryness, and subsequent treatment of the residue with concentrated hydrochloric acid, iridium was determined colorimetrically according to a modification (IO) of the Berman-McBryde (3) method. I n all cases, a reagent blank was carried through the entire procedure and standards xere used directly from the stock iridium solution and, for the micro amounts, analyzed simultaneously with the samples (Table 11). DETECTION OF RHODIUM I N IRIDIUM PRECIPITATE AND IRIDIUM I N RHODIUM PRECIPITATE
The final precipitates, obtained in the above experiments, were combined to form a single sample each of rhodiuni and iridium. Detection of Rhodium in Iridium Precipitate. The filter papers with t h e iridium precipitates were placed in a 50-ml. beaker and decomposed by heating on a hot plate with 2 to 3 ml. of fuming nitric acid. One t o 1.5 ml. of concentrated sulfuric acid mere added and organic matter was destroyed by the addition of portions of concentrated nitric acid and heating to mhite fumes. Thc solution was diluted with water and heated to light fumes to destroy any nitroso compounds formed. The solution was cooled and made u p to 8 ml. ~ i t water. h Rhodium was precipitated by antimony metal and subsequently determined colorimetrically as described (Table 111) (I,+).* Reagent blanks were used. Detection of Iridium in Rhodium Precipitate. The filter papers n i t h the rhodium precipitates m r e placed
in a 500-ml. Erlenmeyer flask and organic matter was destroyed by heating t o heavy fumes with 10 ml. of concentrated sulfuric acid and portions of concentrated nitric acid; the resulting solution was diluted with 20 ml. of water and heated t o light fumes. Rhodium was removed from the solution by 20y0 titanous chloride solution (6). T o eliminate contamination of the rhodium precipitate by iridium, the rhodium n a s precipitated three times and the third precipitate was filtered through an A2 filter crucible. I n the three filtrates combined, titanium was precipitated by cupferron three times and the three filtrates were again combined. Subsequent to destruction of the organic matter by concentrated sulfuric and nitric acids and evaporation of the solution t o 2 ml., iridium was determined colorimetrically according to the Berman-McBryde procedure (5) for samples previously fumed with sulfuric acid. Reagent blanks were used. Results are listed in Table 111. When more than about 5 mg. of iridium are present, the amount recovered from the rhodium (Table 111) is less than that retainecl by the rhodium as indicated in Table 11. This is due to the difficulty of isolating micro amounts of iridium from the bulky cupferron precipitate, because treatment of the final cupferron precipitate with concentrated sulfuric, nitric, and 707, perchloric acids (2) yielded a mauve color only when rhodium and iridium were present in excess of about 5 mg. Honever, in any case, the amount of iridium lost is small. REPRECIPITATION
From the results (Table 11) when both rhodium and iridium metals are present in more than .5-mg. amounts, the rhodium values are slightly high and the iridium values slightly low, by about 30 to 40 y. This is in agreement with the recorded values of iridium in Table I. When maximum accuracy is required, rhodium can be reprecipitated by copper in the effluent of the large exchanger. This reprecipitation could decrease the amount of iridium reduced 1%ithout increasing the filtrate losses of rhodium beyond 0.01 mg. A second precipitation of rhodium and iridium, separately, was applied as follow: Rhodium was precipitated and filtered as described above in the separation of rhodium from iridium. The filtrate and washings were kept (filtrate B). The residue was dissolved by aqua regia and dry chlorination treatments. Copper was eliminated in the resulting solution through the large exchanger as described, subsequent to conversion of the metals t o chlorides. The effluent was adjusted to 30 ml. in 1.ON hydrochloric acid and rhodium was reprecipitated by copper and filtered as previously recom-
Table IV.
Iridium Reduced and Filtrate Losses of Rhodium Applying Reprecipitation
Rhodium Added, mg. Unreduced, y 10.57 8
Copper, Added, Mg. 550 200
+
... ... ... ... .
Table V.
.
I
7 7
... ...
...
Iridium Added, mg. Reduced, ... ... ... ... ... ... 9.28 7.2 ... 7.0 ...
y
7.1
Separation of Rhodium from Iridium Applying Reprecipitation of Rhodium
Copper 550 200
+
Added, Mg. Rhodium 10.57
mended. The filtrate and washings were combined with filtrate B. After removal of the copper, rhodium was determined colorimetrically in these combined filtrates as described under the recovery and determination of rhodium. The same procedure was followed for iridium. The residue of the iridium reprecipitation was dissolved in aqua regia followed by dry chlorination, the metals were converted to chlorides by hydrochloric acid, copper was eliminated through the cationic exchangers, and iridium was determined colorimetrically (Table IV). The results of separation of rhodium from iridium, applying reprecipitation of rhodium in the effluent of the large exchanger and proceeding according to recommended method of separation, are recorded in Table V. SUMMARY
A method for separating rhodium from iridium in a solution of both metals is successful in a range from 5-mg. to 50-7 amounts of these two metals. Losses of either rhodium or iridium t o iridium or rhodium precipitate, respectively, are negligible and in accordance with the filtrate losses of rhodium and iridium. This procedure could be applied also for the metals present in excess of 5 mg. after a reprecipitation of rhodium in the effluent of the larger exchanger. ACKNOWLEDGMENT
The authors express their appreciation t o National Research Council of Canada for financial support in the form of a grant given to G. G. Tertipis. LITERATURE CITED
(1) Aoyama, S., Watanabe, K., J. Japan. Chem. 75, 20-3 (1954).
Iridium 9.28
Recovered, Mg. Rhodium Iridium 10.54 9.26 10.53 9.24 10.55 9.25
(2) Barefoot, R. R., McDonnell, W. J., Beamish, F. E., A N ~ LCHEM. . 23, 514 (1951). (3) Berman, S. S., McBryde, W. A. E., Analyst 81,566 (1956). (4) Berman, S. S., McBryde, W. A. E., IND.ESG. CHEM.,ANAL.ED.’ 18, 120 11946). (6) Gilihrist, R., J. Research Natl. BUT. Standards 9,547 (1932). (7) Hill, M. A., Beamish, F. E., ANAL. CHEM.22,590 (1950). (8) Icarpon, B. G., Federova, A. N., Ann. inst. platine (U.S.S.R.) 11, 135 (1933). (9) Latimer, W. M., “Oxidation Potentials,” 2nd ed., pp. 342-4, Prentice-Hall, Sew York, 1952. (10) Marks, A. G., Beamish, F. E., ANAL. CHEM.30,1464 (1958). (11) Maynes, A. D., McBryde, W. A. E., Analyst 79,230 (1954). (12) Pshenitsyn, K. K., Federov, I. A., SimanovskiI, P. V., Izvest Sectora Platiny i Drug. Blagorod. Metal. Inst. Obschei i Neorg. Khim. A k a d . N a u k . S.S.S.R. 22, 16-21 (1948). (13) Sandell, E. B., “Colorimetric Determination of Traces of Metals,” 2nd ed., pp. 523-5, Interscience, New York, 1950. (14) Westland, A. D., Beamish, F. E., Mikrochim. Acta 10, 1474 (1956).
RECEIVEDfor review October 1, 1959. Accepted December 17, 1959.
Correction Precision in X-Ray Emission Spect rogr a phy, Background Present I n this article by P. D. Zemany, H. G. Pfeiffer, and H. A. Liebhafsky [ANAL. CHEM. 31, 1776 (1959)], on page 1778, column 1, line 2, the amount should be 4 y instead of 40 y. VOL. 32, NO. 4, APRIL 1960
489
