standards; the others served as controls. The tests were limited and the results (Table I) are intended only as a guide. Phosphorus, in large amounts, was the only element that caused significant interference. The important point is that platinum had no significant effecta n advantage in comparison with the titration methods.
RESULTS
Seven bauxite samples containing 5 to 32yo ferric oxide were prepared and analyzed (Table 11). The average difference between the titration and colorimetric values was only 0.10%. The standard deviation of the colorimetric procedure was 0.085yO. Determinations also were made on samples containing smaller amounts of iron, in which the final volume was 100 ml. instead of 500 ml., with a corresponding decrease in the amounts of reagents (Table 111). With either the 500- or 100-ml. final volume, the colorimetric method appears accurate and precise.
Comparison of Iron Results by 100-MI. Procedure with Those by Dichromate Titration
Table 111.
Fe.,O.. Sample KO.
NBS 69A RRC-1 SRC-6 SRC-2 SRC-3 SRC-5 Av. Std. dev.
Type Bauxite Bauxite Bauxite Bayer mud Bayer mud sinter Sinter mud
ACKNOWLEDGMENT
The author gratefully acknowledges the helpful suggestions of D. W. Utley, M. L. Moss, H. B. Hartman, and the analysts of Alcoa Research Laboratories. LITERATURE CITED
(1) Fortune, NT. B., Mellon, M. G., IND.ENG. CHEM.,ANAL. ED. 10, 60
(1938). ( 2 j Hillebrand, W. F., Lundell, G. E. F., Bright, H. A., Hoffman, J. I., “Applied Inorganic Analysis,” 2nd ed., p. 392, Wiley, New York, 1953. (3) Saywell, L. G., Cunningham, B. B.,
CT,
Colorimetric Range of Av. duplicates 5.93 0.02 5.47 0.03 20.52 0.17 4.27 0.02 2.86 0.05 5.79 0.09 7.47 0.06 k0.06
Titration av. 5.82 (assigned) 5.43 20.62 4.42 2.93 5.78 7.50
IND.EXG. CHEM., ANAL. ED. 9, 67 (1937). ( 4 ) Smith, G. F., “Cerate Oxidimetry,” p. 21, G. Frederick Smith Chemical co., Columbus~Ohio, 1942. (5) Smith, G. F., Richter, F. P., “Phenanthroline and Substituted Phenanthroline Indicators,” pp. 3, 13, G. Frederick Smith Chemical Co., Columbus, Ohio, 1944. (6) Walden, G. H . l Hammett, L. p . l Chapman, R. P., J. Am. Chem. Sac. 5 5 , 2649 (1933).
RECEIVED for review November 23, 1959. Resubmitted August 26, 1963. Accepted September 30, 1963. Southwest Regional Meeting, ACS, Baton Rouge, La., December 4, 1959.
New Fire Assay Method for Iridium G. H. FAYE, W. R. INMAN, and P. E. MOLOUGHNEY Mineral Sciences Division, Mines Branch, Departmenf o f Mines and Technical Surveys, Ottawa, Canada
b
The tin-collection scheme for the determination of the precious metals has been extended to include iridium. During the fusion process, the collection of both microgram and milligram quantities of iridium from synthetic charges is essentially complete. Tin powder is used to recover iridium from the parting-acid solution. Ion exchange and solvent extraction procedures are used to isolate iridium prior to its gravimetric or spectrophotometric estimation. The proposed method has been successfully applied to the analysis of a flotation concentrate, a copper-nickel matte, and a specimen of osmiridium from New Guinea.
A
analytical scheme for the determination of the platinum metals in ores and rocks as well as mineralogical and metallurgical products has been under development in the Mines Branch laboratories. This scheme is based on the collection of the precious metals in tin during the crucible-fusion process of fire assaying, NEW
366
ANALYTICAL CHEMISTRY
and has proved successful for platinum and palladium (6), gold ( 7 ) , and rhodium ( 8 ) . The work described in this paper was undertaken to determine the efficiency of tin as a collector for iridium and, also, to extend the analytical scheme so that iridium can be separated from interfering elements and determined by wet chemical methods subsequent to the crucible-fusion process. Results show that the collection of 5 pg. to 4.5 mg. of iridium from synthetic charges is essentially complete and that assay buttons containing base metals and other precious metals can be analyzed readily for iridium. Slag and crucible wall losses and the behavior of iridium during the parting operation were determined by radiochemical experiments using iridium-192 as tracer. Reproducible results were obtained on the application of the new method to the determination of iridium in a flotation concentrate, a copper-nickel matte, and a sample of osmiridium from Kew Guinea.
EXPERIMENTAL
Apparatus and Reagents. The apparatus and flux used for crucible fusion have been described (7, 8 ) . STANDARD IRIDIUM SOLUTION. This was prepared by dissolving Johnson and Matthey Specpure ammonium chloroiridate in 200 ml. of 1N hydrochloric acid. The iridium content of this solution was determined with 2-mercaptobenzothiazole (2) and was 1.06 mg. of iridium per ml. More dilute solutions were prepared from the stock solution by appropriate dilution with liV hydrochloric acid. Procedure.
PREPARATION OF TIN
ASSAY BUTTONS. The procedure for preparing the assay buttons was identical t o t h a t described in earlier papers on gold (7) and rhodium (8). ANALYSIS OF ASSAYBUTTONS. Each assay button was granulated and parted in hydrochloric acid as described previously (7, 8). The parting acid was treated with tin powder (8) to recover any iridium that dissolved during the parting operation and the resulting precipitate plus excess tin powder was recovered by decantation and filtration and dissolved as in the
method for rhodium (8). Tin was volatilized with mixed hydrochloric and hydrobromic acids ( 6 ) , and the sample was then taken to (dryness. The resulting salts were treated cautiously with 20 ml. of aqua regia (bromine may be evolved vigorously when large quantities of base-metal bromides are present) to ensure thrtt the iridium was in a soluble form. Nitrogen compounds were destroyed by evaporating the sample to dryness after repeated treatments with hydrochloric acid. The sample was taken up in dilute hydrochloric acid and passed through a cation exchange column t o remove base metals ( 6 , 8 ) , When necessary, the sample was passed through a second smaller column t o remove traces of iron, copper, and nickel and, except where otherwise noted, the effluent solution was treated according to the solvent extraction scheme of Faye and Inman (9) for the separation of platinum, palladium, rhodium, and iridium. The iridium fraction (in 1 to 9 hydrobromic acid) was treated with approximately 1 ml. of Concentrated sulfuric acid and the solution was evaporated to fumes of sulfur trioxide. To volatilize ruthenium, approxirrately 1 ml. of concentrated perchloric acid was added dropwise, with swirling, to the fuming solution. After the excess perchloric acid had evaporated, the solution was cooled, taken up in bydrobromic acid, and then analyzed for iridium by a modification of the method of Berman and McUryde (3). RESULTS AND DISCUSSION
Radiochemical Experiments with Iridium-192. Because of the con-
venience afforded by radiochemical techniques, certain preliminary experiments were performed with iridium-192 as tracer. The preparation of the standard iridium-192 solution and the counting techniques were similar to those described in previous publications (6, 10). RECOVERY OF IRIDIIJM FROM SALTED CHARGES. To determme the efficiency of tin as a collector for iridium during the crucible-fusion prccess, a series of experiments was performed in which 115-gram portions of the flux were salted with a similar aliquot of the standard iridium-192 solution, microgram amounts of inactive iridium carrier, and, in most experiments, 0.25 gram each of copper and nickel. These charges were fused to produce buttons weighing approximately 20 grams each. Each button was granulated and dissolved in a mixture of hydrochloric acid and hydrogen pensxide (6, 8) and the resulting solution was subsequently analyzed radiometrically for iridium192. The results of these experiments indicated that more than 98% of the iridium can be "collecsed" during the fusion process. Loss OF IRIDIUM TO SLAG AND CRUCIBLE WALL. The slags from, and the crucibles used in, the fusions of three of the salted charges (100 fig. of iridium
iridium-192 showed that this fraction of the iridium can be completely recovered by stirring the parting acid for approximately 20 minutes with 6 grams Iridium of powdered tin. Thus, in practice, both Taken, Found, Taken, Found, iridium and rhodium can be recovered mg. mg. fig. rg. simultaneously from the parting acid. 10.6 12 4.57 4.64 I n experiments with buttons t h a t did 21.2 22 4.09 4.08 5.85 5.85 42.4 41 not contain base metals, iridium (100 6.03 101" 5.96 106.0 fig.) remained insoluble during the 209" 6.11 212.0 6.28 parting operation; in this respect, Aliquots taken for analysis. iridium is similar to rhodium (8). Chemical Analysis of Synthetic Assay Buttons. Radiochemical experiments indicated t h a t the collection of iridium was essentially complete carrier) described above were analyzed radiometrically for iridium-192. The during the crucible-fusion process; slag losses in these tests were 0.6, 0.1, however, it was necessary to have a n and O.l%, respectively, and the loss entirely chemical procedure for isolatto the crucible walls was approximately ing and determining iridium sub0.2% in each test. sequent to fire assay. It was conRECOVERYOF IRIDIUM FROM PART- sidered t h a t the necessary features for ING-ACIDSOLUTION WITH TIN POWDER. such a procedure were available in existing cation exchange ( 4 ) , solvent I n the previously reported methods extraction ( 9 ) , and spectrophotometric based on the tin-collection scheme, the (3) methods. assay buttons were parted in conTo evaluate these processes, and to centrated hydrochloric acid and determine the recovery of iridium from platinum and palladium (6) and gold synthetic samples, the following experi(7) remained completely in the parting ments were performed. residue. A substantial proportion of any rhodium in the button dissolved in A series of assay buttons was prepared the parting acid when base metals were from charges containing from 10 pg. t o also present; however, this fraction was 4.5 mg. of iridium and 0.25 gram each easily recovered by treating the parting of copper and nickel. With the excepacid with tin powder (8). tion of the tests involving milligram Radiochemical experiments showed amounts of iridium, each button was that approximately 60$70 of the iridium analyzed as described under "Prodissolves during the parting of assay cedure" (Table I). buttons containing iridium and comWhen milligram amounts of iridium paratively small amounts of copper were to be determined, the effluent and/or nickel. Subsequent tests with solutions from the cation exchange
Table 1.
Recovery of Iridium from Salted Charges
(I
Table II.
Nature of sample Copper-nickel matte
Determination of Iridium in Various Materials
Sample wt., assay ton 1 1 1
Test no. 1" 2" 3" 4" 55 6b 7b
Flotation concentrate
9 10 11 12 13
Mg 8.68 16.65 11.65
0.025 0.027 0.026 0.026 0.024 0.026 0.023 0.024 Av. 0.025 0.003 0.005 0.004 0.003 0.003 Av. 0.004
0.024 (4)"
-
.
Osmiridium
Found, Troy ounce/ton Proposed Independent method laboratory
% 14d 15d 16d
34.0 33.3 34.8 -
Ave. 34.0 Samples leached in 12N hydrochloric acid before fusion. * Samples fused directly after roasting. c Figure in parenthesis indicates number of replicate determinations. Aliquots of final sample solution analyzed spectrophotometrically.
VOL. 36, NO. 2, FEBRUARY 1964
367
columns were analyzed gravimetrically using. 2 - mercaptobenzothiazole ( 2 ) (Table I). Preliminary experiments of a similar kind had shown the necessity of leaching the ignited iridium precipitate in boiling 1 to 1 hydrochloric acid, prior to reduction in hydrogen, t o eliminate contaminants that follow iridium through the cation exchange columns. This problem has been noted by Tertipis and Beamish (11). Table I shows t h a t tin is a n efficient collector for both milligram and microgram amounts of iridium during the crucible.fusion process; also, that the chemical procedures selected for button analysis are satisfactory. I n certain early experiments of the kind described above, low and variable results were occasionally obtained. This difficulty was traced to the incomplete dissolution of iridium in the residue obtained after the button-parting operation. However, when the sample solutions were treated with aqua regia after the tin volatilization step, satisfactory results were obtained consistently. Application of Method. The proposed method was applied t o t h e determination of iridium in samples of a copper-nickel matte, a flotation concentrate, and osmiridium from h'ew Guinea (Table 11). The copper-nickel matte and flotation concentrate were derived from an ore of the Sudbury district of Ontario; their base metals, platinum, palladium, rhodium, and gold contents have been given (6, 7 , 8). The matte was also analyzed for iridium by a n independent laboratory using a method in which the sample was leached with an acid mix-
ture, the leach residue was decomposed by Wichers bomb technique (12), and the determination was completed spectrophotometrically ( I ) . The results obtained by this method are also given in Table 11. The osmiridium was obtained from the donor in the form of flattened grains (average weight approximately 0.3 mg.) that had been leached with aqua regia to remove soluble impurities. To prepare the charges for the analysis of osmiridium, the weighed samples were mixed with approximately one third of the flux (40 grams) and placed in the crucible in such a way that the mixture was surrounded by the remainder of the flux (approximately 80 grams) and approximately centered in the crucible prior to fusion. By using the solvent extraction scheme of separation (9) on solutions obtained from the samples of osmiridium subsequent to fusion, this material was found to contain 2.33% platinum, 0.9% rhodium, and less than 0.3% palladium. Osmium and ruthenium were not determined. Table I1 shows that the results obtained by the proposed method for the copper-nickel matte are in agreement with those obtained by an independent laboratory using a n alternative wet chemicaI method. It is also shown that, for convenience, the copper-nickel matte may be given a preliminary acid leach t o remove the bulk of the base metals without losing iridium to the leach liquor. Platinum, palladium ( 6 ) , rhoand gold (7) are also insoluble dium (8), during this treatment.
The similarity of the results obtained for the separate samples of osmiridium suggests that the mineral was completeIy decomposed during the fusion process and that collection of iridium was essentially complete. ACKNOWLEDGMENT
The authors are grateful to Falconbridge Nickel Mines, Ltd., for supplying samples of flotation concentrate and copper-nickel matte, and also for the analyses used for comparison in this paper. They also express their thanks t o Johnson, Matthey and Co., Ltd. (Toronto), for supplying the osmiridium used in the present work. LITERATURE CITED
(1) Ayres, G. H., Bolleter, W. T., ANAL.
CHEM.29. 7. -2 (19.57'). ~~~~
~
~
I
\ - - - . I -
(2) Barefoot, R. R., McDonnel, W. J., Beamish, F. E., Ibid., 23, 514 (1951). (3) . . Berman, S. S., McBrvde, " , W. A. E., Analyst 81, 566 (1956). (4) Coburn, H. G., Beamish, F. E., Lewis, C. L., ANAL.CHEM. 28, 1297
(7) Ibid.; p. '1914.' ' (8) Ibid., 34,972 (1962). (9) Ibid., 35, 985 (1963). (10) Faye, G. H., Inman, W. R., Talanta 3, 277 (1960). ( 1 1 ) Tertipis, G. G., Beamish, F. E., ANAL.CHEM.34, 108, 521 (1962). (12) Wichers, E., Schlect, W. G., Gordon, C. L., J. Res. Natl. Bur. Std. 33, 568 (1944) (R.P. 1614).
RECEIVED for review September 20, 1963. Accepted October 27, 1963. Crown copyright reserved.
Fluorescent Compounds for Calibration of Excitation and Emission Units of Spectrofluorometer ROBERT J. ARGAUER' and CHARLES E. WHITE Department of Chemistry, University of Maryland, College Park, Md.
b An ethanol solution of the aluminum chelate of 2,2'dihydroxy-l,l '-azonaphthalene-4-sulfonic acid serves as a fluorescence standard to calibrate the excitation source of a spectrofluorometer. The emission standards quinine sulfate, 3-aminophihalimide, m-nitromethylaniline, and 4-dimethyl4'-nitrostilbine in dilute solution were found applicable to the AmincoBowman instrument. The aluminum chelate listed above is an addition to this group.
M
for the calibration of the Aminco-Bowman Spectrophotofluorometer have been described by White, Ho, and Weimer (12) and
368
ETHODS
ANALYTICAL CHEMISTRY
Melhuish (5, 6) and for other spectrofluorometers by Parker et a!. (8-10). Mechanical methods of calibration are often expensive and not available. The object of this paper is to describe methods for the calibration of both the excitation source and emission unit by means of easily available fluorescent standards. Parker (8) has described a fluorescent screen method which serves nicely for the corrections of the emission spectra in the ultraviolet region. Lippert (4) has proposed several compounds as standards for the calibration of the emission unit of a prism spectrofluorometer when the emitted light is viewed from the front surface of a fluorescent sample. These compounds are now
shown to be applicable as standards when the fluorescence is measured at 90' angle to the exciting light. EXPERIMENTAL
Apparatus. Aminco-Bowman recording spectrophotofluorometer, a n infrared power supply and detector attachment (American Instrument Co., Cat. No. 4-8170), Kipp thermopile No. E20-3734 attached t o a Kintel electronic galvanometer Model 204.4, Beckman DK1 spectrophotometer.
1 Present address, Socony Mobil Oil Co., Inc., Paulsboro, N'. J.
