1
(bl 2100
2300
0.7
2000
2100 2450
0.3.
0.1.
TIME 1 ng lead in Aqueous Solution
Figure 7. Comparison of 1 ng Pb signals at different atomizing temperatures using the unmodified tube ( a )and design I (b)
Table I. Comparison of Temperatures Using t h e P - E Design a n d t h e New Design Dial set a t
P-E design,
OC
OC
100 200 3 00 400 500 600 750
104 190 290 400 510 580 671
2.5-mm i d . , makes it easier for the analyst to introduce a syringe or micropipet through the hole into the tube without touching the sides of the hole. This ensures that a portion of the sample is not lost to the hole surface in the sample introduction phase. Furthermore, the sensitivities obtained with the new tube for the analyses of P b and Hg in aqueous solutions are better than those with the old tube (Figures 5-7). In this laboratory, trace metal analysis in microsamples of biological and clinical nature was improved by use of new designs of a hollow graphite tube. Preliminary results indicate that the new design is superior to the commercial design for such analyses. I t is felt that the new tube designs should provide an improved utility for direct analyses of clinical and biological samples with little or no sample preparation. A study on the use of these tube designs for the determination of volatile metals in blood, serum, and tissues is under way and will be reported later. ACKNOWLEDGMENT
New design,
The authors express their sincere thanks and appreciation to L. Carignan for providing the blood and serum samples, and to R. Wiley for cutting the windows in the graphite tube.
OC
225 426 660 855 926 1054 1164
LITERATURE CITED
~
tion step due largely to incomplete tube clearance of decomposed organic matter in the charring step, diminishes and disappears a t higher charring temperatures. The absence of this peak when using the modified tube designs is hypothesized as being due to a) the minimization of the internal surface area on which mobile particles may deposit during the charring step; b) the greater removal of such particles through the windows by the inert gas flow; and c) the provision of a more uniform distribution of thermal energy due to the lowering of the thermal mass of the tube. In the case of the unmodified tube, the temperature a t the center of the tube appears higher than a t the ends. I t is readily observed t h a t in the modified tube, the center area between the windows, 2 cm is glowing red a t 650 OC; while the standard design has a glowing band of less than 1 cm a t the same temperature. The presence of a temperature profile in the unmodified tube enhances the deposition of particles in the cooler regions of the tube. Table I shows t h a t the temperature of the new design is higher than that of the unmodified tube when the same instrument setting is used. T h e enlargement of the center hole from 1.5-mm t o
0
E. J. Underwood, "Trace Elements in Human and Animal Nutrition", 3rd ed., Academic Press, New York, N.Y., 1971. W. Mertz and W. E. Cornatzer, Ed.. "Newer Trace Elements in Nutrition", Marcel Dekker, New York, N.Y., 1971. E. L. Kothy, Ed., "Trace Elements in the Environment", American Chemical Society, Washington, D.C., 1973. A. A. Kbg. Astrophys. J., 75, 379 (1932). 9. V. L'Vov, Spectrochim. Acta,'l7, 761 (1961). B.V. L'vov, Spectrochim. Acta, Part 6, 24, 53 (1969). R. Woodriff and G. Ramelow, Spectrochim. Acta, Part 8, 23, 665 (1968). H. Massman. Spectrochim Acta, Part 6,24, 53 (1969). M. D. Amos, P. A. Bennet. K. G. Brodie. P. W. Y. Lung, and J. P. Matousek, Anal. Chem., 43, 21 1 (197 1). T. S. West and X. K. Williams, Anal. Chim. Acta, 45, 27 (1969). F. Dollmsek and J. Stupar, Analyst(London), 98,841 (1973). J. F. Lech, D. D. Siemer, and R. Woodriff. Spectrochim. Acta, Part 6, 28, 435 (1973). H. M. Donega and T. E. Burgers, Anal. Chem., 42, 1521 (1970). J. Y. Hwang, P. A. Ullucci, and S. B. Smith, Jr., Amer. Lab., 3 (8),41 (1971). J. P. Matousek. Amer. Lab., 3, (6), 45 (1971). H. L. Kahn and D. C. Manning, Amer. Lab., 4 (8),51 (1972). H. J. lssaq and W. L. Zielinski, Jr., Anal. Chem., 46, 1328 (1974). H. J. lssaq and W. L. Zielinski, Jr., Anal. Chem.. 46, 1436 (1974).
RECEIVEDfor review July 23, 1974. Accepted August 7, 1975. Presented in part a t the 1975 Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, Cleveland, Ohio, March 1975. Research sponsored by the National Cancer Institute under Contract No. NO;-CO-25423 with Litton Bionetics, Inc.
Relative Molar Response of a Flame Ionization Detector to Some Organic Compounds Containing Sulfur W. G. Filby, K. Gunther, and R.-D. Penzhorn lnstitut fur Radiochemie, Kernforschungszentrum, Karlsruhe, West Germany
The relative molar response (RMR) to flame ionization detectors (FID) of many oxygen-containing organic compounds has been determined in recent years (1,2).Ackman
showed that, for these compounds, the RMR could be predicted reasonably well when a specific response is assigned t o certain polyatomic groups (3).During the course of our
ANALYTICAL CHEMISTRY, VOL. 47, NO. 13, NOVEMBER 1975
2283
Table I. Relative M o l a r Response (RMR)” of Some (S-0)-Containing Compounds Carbon Substance (molecular weight)
Dimethyl sulfone (94) Methyl e s t e r of methanesulfonic acid (110) Methyl ester of ethanesulfonic acid (124) Methyl e s t e r of n-propanesulfonic acid (138) Methyl e s t e r of isopropanesulfonic acid (138) Methyl e s t e r of n-butanesulfonic acid (152) Methyl e s t e r of fert-butanesulfonic acid (152) Methyl e s t e r of tert-butanesulfinic acid (136) n-Butyl sulfite (194) n-Butyl sulfone (178) Di-n-butyl thiosulfonate (210) Diphenyl ( 154) Diphenyl disulfide (218)
RMR
number
ECK
126
2
1.3
176
2
1.8
326
3
3.3
397
4
4 .O
444
4
4.4
483
5
4.8
368
5
3.7
428 573 610 850 1050 1040
5 8 8 8 12 12
4.3 5.7 6.1 8.5 10.5 10.4
=The RMR values were placed on the heptane = 700 scale by employing the experimentally determined laboratory factor and the Ackman increments for diethyleneglycol monoethyl ether (RMR = 355). * The effective carbon number (ECN) refers to the response of the FID to a compound, relative to its “parent” hydrocarbon which is assumed fully ionized. recent work on the photochemistry of sulfur dioxide in the presence of alkanes ( 4 , 5 ) ,we observed t h a t for most of the sulfur containing compounds (RS02H, RSOsH, RS02SR, RSO2R) which are formed in these systems, scarcely any information on the RMR was available in the literature. In fact, most of the aliphatic sulfinic acids had not been synthesized u p to very recently (6). Hence, we determined the FID response of a number of compounds containing a variety of S,O, groups in order t o investigate the correlation between the degree of S,O, substitution, chain length, and branching with the RMR. EXPERIMENTAL Materials. While alkyl sulfinic acids were prepared according t o a method developed in our laboratory ( 6 ) , the alkyl sulfonic acids were obtained via concentrated nitric acid oxidation of the corresponding thiols (7-9). The n-butyl thiosulfonate was prepared by hydrogen peroxide (30%) oxidation of n-butyl disulfide according to the method of Hinsberg (IO). All other materials were of the purest quality commercially available. Prior to RMR determination, all substances were purified by double recrystallization or vacuum distillation. Gas Chromatographic Operating Conditions. All determinations were performed on a Hewlett-Packard 5710 A Gas Chromatograph coupled to a Hewlett-Packard 3380 A Integrating Recorder. The following operating conditions were employed: Column: 10% diethylene glycol succinate on chromosorb; carrier gas flow rate: 30 ml/min; starting temperature: 80 “C; temperature programming: 2 OC/min up to 190 OC; air flow rate: 240 ml/min; hydrogen flow rate: 30 ml/min. Procedure. To an ethereal solution of the substance to be investigated, 1 ml of a standard solution of m-chlorobenzoic acid methyl ester (10-2M) was added and the whole made up t o 25 ml with ether. The latter compound served as an internal standard. In order t o obtain accurate molar response values, at least three solutions of different concentrations were prepared for each of the investigated compounds. Usually three replicate 2-pl samples of each solution were injected. To place the determined values on a RMR basis, the procedure was repeated for solutions of diethylene glycol monoethyl ether. The RMR value of the latter was calculated to be 355, using the increments provided by Ackman (3).In general, if 2284
the laboratory factor (LF) of a compound A, whose RMR value is known (2, 3), has been determined, it is possible to obtain the LF (in the same particular arbitrary units chosen for A) of any compound B listed in Table I by employing the relationship RMRA LFB =i -X LFA RMRB The data for the heteroatom-containing compounds investigated in this work have been summarized in Table I. The effect of S,O, incorporation, percent carbon, and degree of branching was examined in the light of the following classification: 1)two sulfones, dimethyl and di-n-butyl, 2) a series of variously branched aliphatic methyl esters of sulfonic acid (the methyl ester of tert-butane sulfink acid may also be incorporated here, 3) symmetrically (di-nbutyl sulfite and di-n-butyl sulfone) and unsymmetrically (di-nbutyl thiosulfonate) substituted sulfone (>SO11 groups and, 4) diphenyl and diphenyl disulfide.
.RESULTS A N D DISCUSSION Bearing in mind the classification given above, the following features can be derived from the results: 1) For sulfones, characterized by two C-S bonds, the effective carbon number (ECN) appears to be lower than that of the corresponding hydrocarbon by approximately 1-2 units, depending on the chain length. This may be a reflection of the stability of the alkyl sulfonyl radicals formed during the primary chemi-ionization ( 1 1 ) . 2) In the case of the methyl esters of sulfonic acids, two trends become evident: a) a n approximately linear increase in the RMR value with increasing chain length for straight chain compounds (ARMR = 100-150) and no effective carbon deficiency. This contrasts interestingly with the data of Ackman and Sipos (12) who found carbon deficiencies of 1.5 for some analogous carboxylic acid esters. Only a minor change is observed on “reducing” the tert-butane sulfonic to tertbutane sulfinic acid. b) An E C N value for the tert-butyl member one unit smaller than the nominal. A similar conclusion was reached by Clementi e t al. (13) for tert-butyl groups attached t o heteroaromatic rings. It is probably associated with the reluctance of a quaternary carbon atom t o chemi-ionize. 3) Here a striking difference becomes visible between symmetrically and unsymmetrically substituted sulfone groups. Whereas in t h e case of the thiolsulfonate no carbon deficiency is evident, both of the symmetrically substituted compounds show carbon deficiencies of approximately 2. T h e cause of this effect is not clear, though it may be associated with the known stability of sulfones and the ease of fission of S-S bonds. 4) Unfortunately our preestablished program (extremely short retention times rendered peak integration inaccurate) did not enable us t o investigate the effect of sulfur incorporation in the aliphatic series. T h e results for diphenyl and diphenyl disulfide are therefore presented. In both cases the E C N values are two units smaller than expected and the RMR values identical within experimental error. It is attractive t o propose that the bridge carbons behave similarly t o the quaternary carbon atom of a tert-butyl moiety (see above), Le., each contributes a carbon deficiency of 1. T h e introduction of sulfur a t the bridge position apparently has a negligible effect. CONCLUSION T h e gas chromatographic response factors for 13 S,O, compounds have been determined. With the sole exception of compounds containing quaternary carbon, the RMR values are, within groups of compounds of similar structure, roughly proportional to the number of available carbons in the molecule. T h e incorporation of sulfur has little effect on the RMR when they are situated between two “non-effective” carbon atoms. T h e degree of oxidation of a
ANALYTICAL CHEMISTRY, VOL. 47, NO. 13. NOVEMBER 1975
sulfur atom in a SO ,, containing hydrocarbon has only a minor effect on the RMR.
LITERATURE CITED (1) G. Perkins, R. E. Laramy, and L. D. Lively, Anal. Chem., 35, 360 (1963). (2) R. Kaiser, "Chromatographie in der Gas Phase", Hochschultaschenbucher-Verlag, MannheimEurich, 1969, Part 2. (3) R. G. Ackman. J. Gas Chromatogr.,2, 173 (1964). (4) R.-D. Penzhorn, L. Stieglltz, W. G. Fiiby, and K. Gunther. Chemosphere, 3, 111 (1973). (5)R.-D. Penzhorn, W. G. Filby, K. Gunther, and L. Stieglitz, Int. J. Chem. Klnet., in press.
(6) W. G. Filby, K. Guhther, and R.-D. Penzhorn, J. Org. Chem., 38, 4070 (1973). D. L. Vivian and E. E. Reid, J Am. Chem. SOC.,57, 2559 (1935). C. R. Nolier and J. J. Gordon, J. Am. Chem. Soc., 55, 1090 (1933). E. C. Wagner and E. E. Reid, J. Am. Chem. SOC.,53, 3447 (1931). G. Hinsberg, Ber. Deut. Chem. Gas.,111, 4294 (1908). J. A. Good and J. C. J. Thynne, Trans. Faraday SOC.,63, 2708, 2720 (1967). (12) R. 0. Ackman and J. C. Sipos, J. Chromatogr., 16, 298 (1964). (13) S. Clementi, G. Savelli, and M. Vergoni, Chromatographla, 5 , 413 (1972).
(7) (8) (9) (10) (11)
RECEIVEDfor review June 23, 1975. Accepted August 18, 1975.
Determination of Silver in Rocks by Neutron Activation Analysis R. A. Nadkarni and G. H. Morrison Department of Chemistry, Cornell University, Ithaca, N.Y. 14853
Silver is a major industrial metal in which a major worldwide imbalance of more than 100 million ounces per year exists between production and consumption (1).This is because silver is comparatively rare in the Earth's crust, being the 67th element in order of natural abundance. The crustal abundance is estimated as 0.07 ppm ( 2 ) . Because of the low abundance levels of silver in rocks or minerals, simple and rapid methods of analysis are required for its determination. Neutron activation analysis (NAA), because of its inherent sensitivity, is one of the preferred methods for the determination of silver. Purely instrumental NAA is not feasible for determining silver in common rocks because the presence of high activity of major and other minor-trace elements will swamp the small amounts of silver activity produced after irradiation. Hence, radiochemical separation has to be employed. There are several methods published for the determination of silver, only a few of which allow one-step separation of silver ( 3 ) .Most others depend on tedious and elaborate radiochemical procedures involving solvent extraction, ionexchange, and precipitation, usually combined together (4-7). Recently, we had proposed a method for the determination of gold and five platinum metals in geological materials using NAA (8). The key aspect of the method was the use of an ion-exchange resin which was specific for noble metals. T h e details of this resin-Srafion NMRR are described in that paper (8).In continued work with this resin, we found that silver(1) is also quantitatively adsorbed on this resin. Since then Muzzarelli and Rocchetti (9) have also studied this resin with 14 transition metal ions and found that, in addition, only Cr(VI), Mo(VI), Sb(tartrate), and Hg(I1) approached quantitative adsorption. In this communication, we present our use of this resin for the determination of silver in rocks after neutron irradiation.
EXPERIMENTAL Reagents. Samples. The rock samples analyzed included the USGS standards andesite AGV-1 and granodiorite GSP-1, South African platinum ore standard PTO-1, and Canadian standard reference materials Syenite rock Sy-1 and sulfide ore SU-1. These were all dried for 24 hours at 90 "C before weighing out for irradiation. For tracer work IosAg (1.38 X lo2 years half-life) obtained from Oak Ridge National Laboratory, Oak Ridge, Tenn., was used after suitable dilution. Standards. Irradiation standards were prepared by dissolving 5-9s pure silver metal in "03 and diluting to the desired strength. Aliquots were weighed out on about 50 mg of "specpure"
Si02 in high purity quartz vials and dried under a heat lamp, after which they were sealed. The standards usually contained about 5 Fg of silver. The silver carrier solution was prepared from AR grade AgNO3 and contained 1 mg Ag/ml. Ion-Exchange Resin. Srafion-NMRR resin was obtained from the Ayalon Water Conditioning Co., Ltd., Haifa, Israel. The resin was packed in glass columns 10 cm in length and 1.5-cm inner diameter. To remove any chloride on the resin, the column was washed with 100 ml of 0.05N "03 until the effluent did not give a positive test for chloride. Irradiations. Geological samples (0.5-0.9 g) sealed in high-purity quartz vials along with (-5 fig) Ag standards were irradiated either for 100 hours at a thermal neutron flux of 2 X lOI3n cm-2 sec-I in the GTRR reactor of the Georgia Institute of Technology, Atlanta, Ga., or for eight hours a t a flux of 3.5 X 10% cm-2 sec-I in the central thimble facility of the Cornell TRIGA Mark I1 reactor. The samples were allowed to "cool" for several days before processing. Radiochemical Procedure. This is practically identical with the method for other noble metals published by us earlier (8) with the following differences: 1) "03 is used in place of HC1; 2) 1 mg of Ag(1) carrier is used instead of noble metals carrier solution; 3) Srafion NMRR ion-exchange column is washed till C1- free before using; 4) a 56-cm3 coaxial Ge(Li) detector is used instead of a 30cm3 detector.
RESULTS AND DISCUSSION Since it is possible that Ag+ may be adsorbed on the resin column as AgC1, the resin was eluted with diluted "03 until free of chloride ions. However, it was still found that lo8Ag+ was 99.98% absorbed (six determinations) on the resin column a t p H 1.5-2. This is the same p H range a t which Hg2+ and noble metals are adsorbed on this column. Once adsorbed, Ag+ cannot be eluted with either 1M KCNS or 3N NH40H. In the case of NH40H, the column turned brownish black because of silver oxide formation. The 1 M KCN could elute up to 15% of the adsorbed Ag. A 5% thiourea solution could essentially quantitatively (90%) elute the adsorbed silver; however, in view of the quantitative adsorption of silver on the resin, elution and subsequent chemical yield determination steps were considered unnecessary. This adsorption and elution behavior is again similar to that observed in the case of noble metals (8).
In addition to the quantitative ion-exchange behavior of silver, we also carried out the entire radiochemical procedure with lo8Ag. The yield of losAg through the complete procedure was 99.6% (average of triplicate determination). In view of this quantitative adsorption, chemical yield determination was considered unnecessary.
ANALYTICAL CHEMISTRY, VOL. 47, NO. 13, NOVEMBER 1975
2285
