Integrating Out Astrophysical Uncertainties

TL;DR

This paper develops a method to compare dark matter detection experiments by removing astrophysical uncertainties, focusing on integral quantities in velocity space, and applies it to analyze signals from CoGeNT, DAMA, CRESST, XENON10, and CDMS-Si.

Contribution

It introduces a model-independent approach to relate experimental signals by mapping observed data into velocity space integrals, reducing astrophysical uncertainties.

Findings

Expected CoGeNT rate in XENON10 is higher than observed unless scintillation is low.

S2-only analyses constrain charge yield Q_y<2.4 electrons/keV.

XENON10 and CDMS-Si would detect signals if CRESST excess is due to WIMPs.

Abstract

Underground searches for dark matter involve a complicated interplay of particle physics, nuclear physics, atomic physics and astrophysics. We attempt to remove the uncertainties associated with astrophysics by developing the means to map the observed signal in one experiment directly into a predicted rate at another. We argue that it is possible to make experimental comparisons that are completely free of astrophysical uncertainties by focusing on {\em integral} quantities, such as $g(v_{min})=\int_{v_{min}} dv\, f(v)/v $ and $\int_{v_{thresh}} dv\, v g(v)$. Direct comparisons are possible when the $v_{min}$ space probed by different experiments overlap. As examples, we consider the possible dark matter signals at CoGeNT, DAMA and CRESST-Oxygen. We find that expected rate from CoGeNT in the XENON10 experiment is higher than observed, unless scintillation light output is low. Moreover, we determine that S2-only analyses are constraining, unless the charge yield $Q_y< 2.4 {\, \rm electrons/keV}$. For DAMA to be consistent with XENON10, we find for $q_{Na}=0.3$ that the modulation rate must be extremely high ($\gsim 70%$ for $m_\chi = 7\, \gev$), while for higher quenching factors, it makes an explicit prediction (0.8 - 0.9 cpd/kg) for the modulation to be observed at CoGeNT. Finally, we find CDMS-Si, even with a 10 keV threshold, as well as XENON10, even with low scintillation, would have seen significant rates if the excess events at CRESST arise from elastic WIMP scattering, making it very unlikely to be the explanation of this anomaly.