Problems Related to Interplanetary Matter (1961)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

RADIUM-URANIUM AGE DETERMINATIONS ON MARINE SHELLS Wallace S. Broecker Lament Geological Observatory The unexplained deviations in C14 ages, discussed by Suess, make it even more important to develop an independent check of the C14 method for ages greater than 5000 years, beyond which point historical dating and dendrochronology fail. A method originally conceived by Potratz based on inequilibrium in the U series may be suitable for this purpose. This method depends upon the observation that the ratio of A 230^ 238 in recently formed marine carbonates generally seems to be quite low (<0. 05). With time, secular equilibrium will be reestablished in accord- ance with the 80, 000 year half-life of Th230. A similar systematic relationship exists for Pa231 (3. 5 x 104 year tj/2), a member of the U235 series. In order to investigate and, if possible, exploit these methods of determining ages, a program has been begun at Lamont Observatory, measuring C14, U235 and U238, Ra226, Pa23l, Th232 and total Ba in samples of marine carbonates. Ra226 (1600 year t1/2) is expected to come into equilibrium with its parent, Th230, in times the order of a few. thousand years, and can be measured with high sensitivity by looking at the decay rate of its gaseous daughter Rn222. Th232 and Ba are measured in order to estimate the initial levels of Th230 and Ra226 respectively. This allows the reliability of the assumption that Th230 and Ra226 are initially absent to be checked in each sample. The calculation of an age from these measurements involves two basic assumptions: that the concentration of Th23^, Ra226 or Pa231 at the time of formation is known, and that the shells have been a closed system since their formation. It seems well established, from the work of Sackett and Potratz, that Th230 and Pa231 are essentially absent in many recently formed marine carbonates. Ra226 is present initially and, although it decays to the level at which it is supported by Th230 within five to ten thousand years, spuriously long ages could result from excess Ra226 in very young shells. This ambiguity can be resolved by measuring the C14 age of the sample. Since Ra is known to migrate in marine clays, measurements of Th230 as well as Ra226 should be made whenever pos- sible. The cross-check provided by the Pa23l-U235 age will also be 96

helpful, although the low isotopic abundance of U23!> and the relatively short half-life of Pa231 indicate that this method has a limited range of applicability. Measurements have been made on three types of marine carbonates: contemporary shells; shells and corals of known radiocarbon age; carbon- ates sufficiently old so that radioactive equilibrium should have been attained. A check of the analytical methods on uranium and radium standard samples indicates that standard deviations are <10 percent (U) and<5 percent (Ra226), with sensitivities of 10 ppb U and 2 x 10-6 ppb Ra226 for analyses of 30 gram carbonate samples. Table 1 displays the results on contemporary gastropods. Note that the uranium concentration is very low; if this level of U had proved to be characteristic of all shells measured, dating by this method would be very difficult. The shells of non-zero age all have U contents >0. 3 ppm. There are two plausible explanations for this discrepancy. That a species effect may be operating is suggested by data of Totsomoto and Goldberg and of Sackett on U in marine carbonates. Shells of other types will be analyzed to check on this. Alternatively, the uranium may be secondary, and TABLE 1 Results on Contemporary Gastropods Ra226 10-14 g/g U238 Ra226 10-6 g/g Ra226 equil. Jamaica 1884 0. 7±0. 2 <0.01 >2.0 1930 0. 6±0. 2 <0.02 >1.0 Bahamas 1880 2. 0±0. 2 <0. 01 >3. 0 1950 1. 1±0. 2 <0.01 >6.0 Iceland 1840 <0. 2 1910 0. 2±0. 1 1946 0. 2±0. 1 Hawaii 1840 0. 5±0. 2 0. 04±. 02 0. 36±. 20 1936(?) 1. 0±0. 2 0. 13±. 03 0. 21±. 06 Tahiti 1880 0. 3±0. 1 1957 0. 4±0. 2 97

might build up with increasing age. A series of identical shells of varying ages is being analyzed to check this hypothesis. From the levels of Ra226 given in Table 1, it seems that ~1 x 10-14 g/g is a typical value for the initial Ra226. It is unlikely that more than ~3 x 10- 14 g/g could have been present in any of the shells studied, since processes of shell-formation almost certainly result in a lower Ra/Ca ratio in shells than in sea-water, judging from the behavior of Sr and Ba in this system. Assuming ~1 x 10-14 g/g Ra226 initially present, errors in the U-Ra ages due to initial Ra should be negligible in shells greater than 5000 years in age. Results of measurements on very old samples are given in Table 2. A progressive acid leach was used on the Eniwetok coral. Clearly, more data is needed to decide whether the positive deviations from equilibrium ratios are significant. TABLE 2 U-Ra Measurements on Samples of Very Great Age Fraction _. _ __ Sample Material % l(T14 g/g 10-* g/g Ra226equil. Potratz Mississip- Ls. Standard pian Lime- 0-100 36±1 1. 00±. 03* 1.00±. 05 stone 584-B Miocene Coral, 0-50 73±3 1.6±. 1 1.26±. 11 Eniwetok 50-100 58±2 1.4±. 1 1. 18±. 11 *Results by other laboratories using four different techniques average 1. 05. The data presented in Tables 3, 4, and 5 are quite encouraging, al- though some of the second-order effects are not understood. The agree- ment between Ra-U and C14 ages for Eniwetok Aragonitic Corals (Table 3) indicates that errors due to initial Ra226 are probably not important. The data on British Columbia and Spitsbergen shells generally show reasonably good agreement, when the first 10-30 percent of the sample is neglected. A puzzling feature is the systematic decrease, often by large factors, in the Ra226 content during the first stages of the progressive acid leach. 98

TABLE 3 Aragonitic Coral from Eniwetok Atoll Sample # Fraction % Ra226 10-14g/g 1 U238 0-6 g/g Ra226 Ra-U Age (103 y) C14 Age 482-A Ra226 , eqml. (103 y) 0-100 3. 5±. 3 2.4±. 2 0. 040±. 004 4. 8±0. 5 3.8±0. 2 482-B 0-100 6. 4±. 8 2. 8±. 2 0. 062±. 008 7. 3±0. 9 5. 6±0. 2 482-C 0-100 5. 0±. 3 3. 0±. 2 0. 047±. 004 5. 7±0. 5 5.9±0. 2 TABLE 4 Results on British Columbia Shells Sample # Fraction % Ra226 U238 Ra-U Age C" Ra226 Age (103y) 391-C io-14g/g i o-6 g/g Ra226 equil. (io3 y) 0-30 8.7±.4 0. 33±. 03 0. 74±. 08 155±35 30-43 1.5±. 2 0. 43±. 04 0. 10±. 01 12±2 12-0.3 43-100 0.47±. 10 0. 18±. 03 0. 072±. 014 8. 6±1.6 475-A 0-30 42±2 1. 15±. 11 1. 01±. 12 >200 >39 based 0-30 40±3 1. 14±. 10 0. 97±. 1 1 >200 on other C14 dates 30-43 10±. 5 1.07±. 11 0. 27±. 03 36±4 and strati 30-43 8. 4±1.0 0. 92±. 10 0. 25±. 03 32±4 graphic evidence; 43-100 3. 5±. 3 0. 29±. 02 0. 34±. 04 48±5 should be 43-100 2. 5±. 3 0. 20±. 02 0. 35±. 04 49±6 <50 475-B 0-30 7.6±. 6 0. 28±. 03 0. 75±. 09 160±50 >39 30-42 4. 7±.4 0. 46±. 04 0. 28±. 03 38±5 same age 42-100 4. 0±. 3 0. 36±. 04 0. 32±. 03 44±5 as 475-A 99

TABLE 5 Results on Spitsbergen Shells Sample # Ra226 U238 Ra-U C14 Fraction % io-14g/g 10- g/g t Ra226 .y .> / Age Age* 572-A Ra226 equil. (io3y) (io3 y) 0-10 7. 1±. 5 0. 62±. 05 0. 32±. 03 44±5 — 10-55 4. 9±. 2 1.29±. 10 0. 11±. 01 13±1 9. 4±. 2 55-100 3.7±. 2 1.64±. 13 0. 064±. 006 7.8±.8 9.7±. 2 572-B 0-10 6. 8±. 3 0.44±. 03 0.43±. 03 65±6 10-55 3. 3±. 2 0. 69±. 04 0. 13±. 01 16±1 9. 1±.2 55-100 2. 1±. 2 0. 54±. 04 0. 10±. 01 11±1 9. 5±.2 572-C 0-10 68±3 1. 2±. 1 1. 54±. 13 — 10-45 65±3 1. 2±. 1 1.44±. 13 — 34±1 45-100 50±3 1.7±. 1 0. 83±. 07 208±60 34±1 * Measurements by Ingrid Olsson of the Uppsala, Sweden, Radiocarbon Laboratory. This pattern is very common and does not seem to be reflected in the uranium contents. However, simple surficial contamination by Ra22° may not be the answer, since in only one case (572-C) is there more Ra226 than would be expected at equilibrium, so that some correlation with U2.^ is apparent. The origin of this Ra226 distribution must be considered an open, and perhaps important, question. The discrepancy between Ra-U and C14 ages for sample 572-C (Table 5) is probably due to inseparable contamination with modern Stratigraphically, 572-C is certainly considerably older than 572-A or 572-B and may be from a previous interglacial period, which would agree with the Ra-U ages. Table 6 gives data on miscellaneous shell samples in which the two ages do not agree at all. The C14 ages are well-established. The most likely cause for this disagreement seems to be initial Th230; comparisons of Th232 contents are in progress to determine whether this could be the case. Another interesting feature of this data is that acid leaches of 100

TABLE 6 Results on Miscellaneous Shells Sample # Ra226 u238 Ra-U C14 Fraction % Ri £60 Age Age 521 (Borneo) io-14 g/g io-6g/g Ra226 , rld equil. (103y) (103y) 0-10 2. 7±. 3 0. 39±. 03 0. 19±. 02 23±2 10-55 1.7±. 2 0. 34±. 03 0. 14±. 02 17±2 5.6±0. 2 55-100 l.3±. 1 0. 41±. 03 0. 087± . 01 11±1 248A (Ellsmere Island) 0-10 8.3±. 4 0. 71±. 05 0. 32±. 03 45±5 10-55 2.8±. 2 0. 54±. 04 0. 15±. 02 18±2 7.2±0. 2 55-100 2. 8±. 2 0. 57±. 04 0. 14±. 01 17±2 581 (Argentine Shelf) 0-15 22±1 1. 8±. 1 5 0. 33±. 03 43±5 [ 15-100 9.6±5 0. 68±. 05 0. 38±. 03 56±6 \ 16±3 sample 581 do not evidence a systematic, uncompensated decrease in Ra226. jf this can be correlated with the fact that this sample has prob- ably never been above sea level, it may indicate that deposition of Ra226 from ground-water is crucial in developing the pattern of Ra22" concen- tration noted in the other samples. The Ra-U age method has also been applied to materials believed, on various grounds, to be of last interglacial age. Ages obtained range from 100, 000 years to 134, 000 years, which is in reasonable agreement with dates for the last interglacial age estimated from extrapolations of sedimentation rates for ocean cores. In conclusion, although the Ra-U method may not be capable of pro- viding a precise independent check of the Cl4 ages, it is promising as a means of extending the range of age measurements in marine carbonates to perhaps 200, 000 years. Much accessory data (e.g., Th230, Th232, pa231 contents) is needed before the method can be applied with confidence. 101