Determination of the decay-constant of 87Rb by laboratory accumulation of 87Sr

The decay-constant of the geochronometer Rb-87 (lambda(87)) has been measured numerous times over the course of nearly a century, yet consensus over an accurate determination has remained elusive. This has resulted in persistent uncertainty, with different laboratories using different values. We have re-determined lambda(87) by laboratory accumulation of Sr-87 in a batch of purified RbClO4 over a ca. 30 year time span. Daughter Sr-87 was extracted by RbClO4 precipitation and Sr-selective extraction chromatography and measured by isotope dilution thermal ionization mass spectrometry. Sr isotopic compositions are extremely radiogenic, making fractionation correction using a single spike highly imprecise. A double-spike enriched in Sr-84 and Sr-86 was prepared in conjunction with this experiment, and used to measure Sr-87 as well as to correct for instrumental fractionation during mass spectrometer measurements. The isotopic ratios of the double-spike were calibrated against Sr standard SRM-987 using the NIST certified isotopic ratios of that standard: Sr-84/Sr-86 = 0.056 55 +/- 0.000 14; Sr-87/Sr-86 = 0.710 34 +/- 0.000 26; Sr-88/Sr-86 = 8.378 61 +/- 0.003 25 (Sr-86/Sr-88 similar to 0.119 352). Rb-87/Rb-85 of SRM-984 was measured with the Sr double-spike by ICP-MS and determined to be = 0.386 353 +/- 0.000 004, and this value was used to correct Sr measurements for Rb-87 interference. Individual measurements of the decay-constant with the double-spike are very precise, but the measurements scatter asymmetrically outside of estimated analytical errors toward high values of lambda(87). Our best estimate for lambda(87) is 1.396 8(-0.001 8)(+0.002 7) x 10(-11)a(-1)(T-1/2: 49 624(-95)(+65) Ma), based on the median and absolute deviation of the median after rejection of two outliers (normalizing Sr ratios to Sr-86/Sr-88 = 0.119 4 yields a value for lambda(87) of 1.397 1 x 10(-11) a(-1)). The value measured here is more precise than other recent estimates based on radiometry and comparison with U-Pb ages and it is consistent with those values within the uncertainties of the measurements. We propose that the value reported here be used for Rb-Sr age determinations in place of all previous values. (C) 2012 Elsevier Ltd. All rights reserved.