UK DM project: radioactivity test results

Introduction

At the activity levels sought for low-background experiments, separate samples of the `same' material can give significantly different results - through either genuine variability or contamination. We (and, I am sure, the various other groups whose results are included here) have made every effort to avoid contamination, but there are always stages outside our control. Samples where there is no obvious likelihood of surface contamination (e.g. the Salisbury Cathedral Pb) have been scraped, etched, or otherwise `cleaned' prior to analysis.

Where errors are quoted, these are generally those quoted by the analyst, and are one standard deviation unless otherwise indicated; in a few instances, where tests on two or three samples have given similar results, I have combined these and quoted the resulting standard deviation. Though these do not necessarily reflect experimental error, they nonetheless indicate a range of uncertainty. Reported statistical errors for samples with high K content are sometimes very small (1 in 10³ or less); these are indicated as zero. Since measurements generally involve a `background' subtraction, mean values are occasionally negative.

The analyses we have commissioned have usually been reported to us as concentrations; John Barton's counting results were converted to concentrations by him, and other counting results were converted by various members of the RAL Group. Many of the results from other groups are reported as disintegration rates; I have converted these.

On the whole, where we have more than one analysis of a material, results are reasonably consistent; however, there are clearly problems with K at the ppm level - which may be due to contamination or sample variability.

With data from other groups, it is not always clear whether what is reported is a measurement or an objective. Where there is prima facie evidence that it is the latter, I have commented to this effect.

Parent U and Th levels are measured directly in mass spectrometry and neutron activation analysis and it is often assumed, without justification, that the whole series are present in equilibrium. In gamma ray spectrometry it is customary to report the results in terms of the ²³⁸U, ²³²Th concentrations which would give the observed line intensities if the series were in equilibrium. Unfortunately the lines normally used - the stronger ones - come well down the chain in both series; the earlier g emitting isotopes only yield weak or low energy lines which are hard to detect. For U the series is usually out of equilibrium, except for unprocessed mineral samples, and it is really ²²⁶Ra and its decay products which are measured. For Th it is less easy to generalize: with Ge spectrometers it should be possible to estimate parent Th via ²²⁸Ac and compare this with `equilibrium' estimates from ²⁰⁸Tl. Note that, in the rare cases where ²³⁵U is present in significant amounts, it may also be out of equilibrium.

As well as the above caveats on the validity of the data, care is required in using them: the relationship between crude activities in target, shielding, and structural materials and observed count-rates is a subtle one. In general, Peter Smith's Monte Carlo-based chart of g and b count-rates for various U, Th, and K concentrations, adjusted for thickness as discussed in the next section, is a convenient guide. Approximate energy binnings of g, b, and a decays from U and Th chains are given in our review (Phys. Reports 187, p 266. Note the omission of 0.36 from the 2.5-3 MeV g bin for Th).

Cosmogenic radioactivity is not included in the tables below; fall-out, and other `artificial' radioactivity is noted where known.