Table 11. Analysis of Alkaline Stannate Tin-plating Baths for Tina Bath No. Sn ( g / l J b Sn ( g / P Sn (g/lJd 1 2 3 4 5
112.0 113.2 112.8 108.0 101.6
44.3 44.4 43.0 43.6 44.4
42.4
... 42.2 42.8
...
Nominally: 37.5-45 g/l. Sn as KiSn(0H)B and 37.5-60 g/l. KOH. Used method of standard additions. Used calibration plot with K added to standards. Iodimetric titration,
The results in Table I1 also indicate that by adding potassium to the standard tin solutions in order to account for the enhanced absorption, the AAS analysis for tin compares
favorably with wet chemistry. The AAS results are within rt 5 of the titration results and were obtained in about one fourth the time. One can also use this enhanced absorption to advantage. When working with small quantities of tin by AAS, the addition of potassium to the standards and samples (provided no other problems are caused by the potassium) will enhance the normally weak signal obtained for tin in the air-acetylene flame. ACKNOWLEDGMENT The authors thank George E. Calley who performed the titrations and prepared the tin-plating baths.
RECEIVED for review August 22, 1969. Accepted December 22,1969.
Homogenized Fission Track Determination of Uranium in Whole Rock Geologic Samples David E. Fisher Rosenstiel School oj' Marine and Atmospheric Sciences, Unicersity of' Miami, 10 Rickenbacker Causeway, Miami, Fla. 33149
DATAON THE uranium contents of many different geologic materials show considerable discrepancies. Activation analysis of the same ultrabasic rocks have shown discrepancies of up t o a factor of three between different investigators ( I , 2). One lherzolite nodule was measured by both activation analysis and flame photometry and showed a variation in abundance of nearly a factor of two (3). Analyses by activation analysis and isotope dilution on three olivine nodules show discordancies of factors of two and three ( 4 ) . Activation analysis of different samples of the same achondritic meteorhes varies by a factor of two from one investigator to another (5-8). Determinations by both activation analysis and delayed neutron counting of chondritic meteorites generally show good agreement, although again variations of a factor of two have been noted (8,9). The U abundances in several standard rocks have been determined by many investigators using a variety of techniques. For some of these rocks the agreement is quite good but in others there are variations of up to an order of magnitude (IO).
(1) H. Wakita, H. Nagasawa, S . Uyeda, and H. Kuno, Earth Plunet. Sci. Lett., 2, 377--81 (1967). (2) V. Becker, J. H. Bennett, and 0. K. Manuel, ibid., 4, 357-62 ( 1968). ( 3 j D. H. Green, J. W. Morgan, and K. S . Keier, ibid., pp 155-66. (4) G. R. Tilton and G. W. Reed, Earth Sci. Meteorit., 31-42 (1963). ( 5 ) H. Von Konig and H. Wanke, Z . Naturforsch. 14 (1959). (6) J. F. Nix and P. K. Kuroda, Nature, 221, 726 (1969). (7) R. S . Clark, M. W. Rowe, R. Ganapathy, and P. K. Kuroda, Geochim. Cosmochim. Acta, 31, 1605-14 (1967). (8) J. W. Morgan and J. F. Lovering, Talanta, 15, 1079-95 (1968). (9) S. Amiel, J. Gilat, and D. Heymann, Geochim. Cosmochim. Acta, 31, 1499-1504 (1967). (10) F. J. Flanagan, ibid., 33, 81-120 (1969). 414
ANALYTICAL CHEMISTRY, VOL. 42, NO. 3, MARCH 1970
While a t least some of these differences are undoubtedly due to heterogeneous U distributions in these rocks, it is possible that there might be some deficiency in any of the several analytic techniques used. I have therefore thought it worthwhile to modify the technique of fission track analysis to make it suitable for whole-rock U determinations. The method as presented here is simple, direct, inexpensive, and rapid. Further, the results clearly indicate, though in a qualitative manner, the degree of heterogeneity of the U distributions within the total rock. In previous fission track studies (11, 12) a plastic detector was placed next to a polished section of rock and recoil fission fragments induced by a neutron irradiation were counted in the plastic. These studies showed that U is heterogeneously distributed within many geologic materials. The technique is useful for studying such distributions, but it is difficult t o analyze the data in terms of average U concentrations for the whole rock. Also, the necessary polishing may remove U from water-soluble U-rich minerals, or introduce contaminant U. The present technique uses a well-powdered homogenized rock surface to avoid this difficulty. The rocks are thoroughly crushed in an agate mortar and pestle and passed through a 100-mesh sieve. The gig used to mount the samples is shown in Figure 1. Milligram amounts of the powder are poured through B, leaving a -0.5-cm disk on A . B is removed and methyl cellulose powder is poured in through C. A pressure of -2000 1 b / h 2 produces a -0.1-cm thick disk with smooth, cohesive surfaces. Lexan, with a n inscribed circle of diameter 20.5 cm is then taped over the sample. A series of -50 disks, together with (11) R. L. Fleischer, Geochim. Cosmochlm. Acta, 32,989-98 (1968). (12) R. L. Fleischer, C . W. Naeser, P. B. Price, R. M. Walker, and U. B. Marvin, Science, 148,629-32 (1965).
Table I. U Contents of USGS Standard Rocks
u(
P a
Sample G-2 granite
Source Flanagan
Previous results (Ref.) 1.3-3.4 (10)
DTS-1 dunite
Flanagan
0.003-O.004 (10)
BCR-1 basalt
Hamaguchi Flanagan
0.0040; 0.0041 (15) 1.2-2.2 (10)
Tatsumoto
1.726 (la)
PCC-1 peridotite
Flanagan
0.0044.005 (10)
AGV-1 andesite
Hamaguchi Flanagan
0.0046; 0.0052 (15) 1.4-2.6 (10)
Tatsumoto
1.962 (16)
0.0042 f 0.0005 1.90 f 0.2 1.92 i 0.2 1.77 f 0 . 2
Flanagan
0.19-2.7 (10)
2.0 * 0 . 2 overwhelming clusters
GSP-1 grandiorite
a glass standard (13, 1 4 , fits into a specially made aluminum can which is then sent for reactor irradiation. N o difference was observed in early runs between powdered and unaltered glass standards; therefore, for the sake of simplicity and reuse of the standards, they were not powdered in subsequent runs. The irradiation intensity is determined by the range of U values expected in the samples. For example, a neutron dose of -1016 cm-2 is appropriate for U concentrations in the range 10-1000 ppb. After irradiation the density of induced tracks is counted in the covering plastic. Etching conditions are 30 minutes in 6N NaOH at 70 OC. About 15 minutes are needed to prepare each sample and about 15 minutes are needed to count the tracks in each sample. The cost of neutrons is about $1 .OO/disk. U concentrations within particles which are smaller than -100 microns will pass undisturbed into the powdered rock surface and will show up in the lexan detector as clusters of tracks. Occasional materials may show such large clustering effects that it is difficult to decide on an average U determination by this method. Of the rock types studied so far, only granite falls into this class. Some other rocks do show occasional clusters (see Table I). When the number of tracks involved in the clusters is significant compared to the total number of tracks, then a suitable error above and beyond that attributed to the purely statistical nature of the total track count must be included. In general, such clusters are rare. They may be considered an indication of the heterogeneity of the U distribution within the sample. Thus, GSP-1, which we note to be overwhelmed with clusters, shows a variation in previous results of greater than an order of magnitude. G-2, with large clustering, shows variations in U abundances of 1.3-3.4 ppm. In general, rocks with less clustering show better agreement among different workers. The problem of contamination introduced by the analytic procedure is always a serious consideration when concentra(13) J. W. H. Schreurs,A. M. Friedman, D. J. Rokop, M. W. Hair, and R. W. Walker, Trans. Am. Geophys. Union (abstract) 50, 336 (1969). (14) R. L. Fleischer, General Electric, Schenectady, N. Y . , personal communication, 1969. (1 5 ) H. Wakita, H. Nagasawa, and N. Onuma, University of Tokyo, private communication, 1967. (16) M. Tatsumoto, U. S. Geological Survey, Denver, Colo., private communication, 1969.
This work 1 .O without clusters 3.5 with clusters 0.0028 f 0.0005 0.0024 f 0.0005 0.0044 f 0.0008 1.75 f 0 . 2 1.87 f 0 . 2 2.00 f 0 . 2 1.70 f 0.2 1.62 f 0.2 0.0046 f 0.0005 0.0041 f 0.0006 0.0058 f 0.0006
-
Figure 1. Gig for preparation of disks 0 5cn 3 c
tion levels of ppm and lower are to be reached. This technique allows any such contamination to be clearly recognized, since tracks associated with U within the rock vanish abruptly at the boundary of the rock circle within the sample disk. The methyl cellulose shell surrounding the rock sample will give rise to tracks only if contamination U has been introduced (the U content of the methyl cellulose is low enough that it gives essentially zero tracks under the irradiation intensities used). In no cases were such tracks (due to contamination) observed. A blank sample (silica sand) showed