Removal of radon progeny from delicate surfaces

TL;DR

The paper addresses the radon-progeny background problem in rare-event searches caused by surface $^{210}$Pb, which decays through $^{210}$Bi to $^{210}$Po and can drive neutron backgrounds via $(\alpha,n)$ reactions. It tests a minimally invasive cleaning method—acetone wiping—on copper and silicon to remove surface contamination and extend allowable construction exposure times. Using alpha spectroscopy and a time-dependent decay model of the chain $^{210}$Pb $\rightarrow$ $^{210}$Bi $\rightarrow$ $^{210}$Po, the authors extract removal fractions, finding about half of the surface activity removed after the first wipe and comparable removal efficiencies for $^{210}$Pb and $^{210}$Po; a second wipe yields little additional benefit for copper, while silicon results have larger uncertainties. This simple cleaning step provides a practical, low-cost mitigation to relax air-exposure constraints in next-generation experiments, potentially reducing assembly time and costs while accommodating delicate components.

Abstract

$^{210}Po$ $α$-decay driven neutron background is a concern for many rare event search experiments. It is a difficult to control background because its radiogenic component depends on the air exposure history of parts. In this study, we demonstrate that about half of the radon progeny $^{210}Po$ can be removed from copper and silicon surfaces relatively easily by wiping a copper sample with acetone wetted tissue and a silicon detector with acetone soaked cotton balls. For a copper sample we demonstrate that long-lived $^{210}Pb$ is removed with similar effectiveness. For copper, allocated the longest counting time, additional wiping was found to be largely ineffective. For silicon, the removal effectiveness has large uncertainties. Additional cleaning showed a small but statistically significant effect. Capitalizing on this trivial cleaning step will allow experiments to relax their requirements on the allowable air exposure time during construction, leading to cost and time savings.

Figures (4)

• Figure 1: Photo of copper sample 1 and the Si detector used to study surface activities.
• Figure 2: Energy spectra obtained with uncleaned (red, 12 days), cleaned (green, 100 days) copper sample 1 and without the sample (black, 83 days). The inset shows the energy range of the $\rm ^{210}Po$ peak. A peak is visible at the expected energy. The shaded area corresponds to the chosen integration range.
• Figure 3: Time dependence of the $\rm ^{210}Po$$\alpha$-peak counting rate obtained with copper sample 1. Each point corresponds to the rate observed during one day of counting, errors are statistical. The red points were obtained with the uncleaned sample, the blue points after the initial cleaning and the green points after repeated cleaning. The black points show rates observed during background runs. The widths of the hatched bands indicate the statistical standard errors of the averages. The x-axis gives the date in format year, month, day.
• Figure 4: Time dependence of the background subtracted $\rm ^{210}Po$ cleaned over uncleaned counting rate ratio for copper sample 1. The point-wise errors account for the subtraction. The red line and numbers in the fit box show the resulting activity ratios for floating $\rm ^{210}Pb$, $\rm ^{210}Bi$ and $\rm ^{210}Po$ rate ratios. The blue fit line was obtained fixing the $\rm ^{210}Pb$ and $\rm ^{210}Bi$ activity ratios to 1 (assume no cleaning effect) and floating only the $\rm ^{210}Po$ ratio.