CoGeNT Interpretations

TL;DR

The paper investigates whether the CoGeNT low-energy excess can be explained by dark matter by examining a broad set of interaction hypotheses, including spin-independent and spin-dependent scattering, generalized proton–neutron couplings, momentum-dependent operators, and inelastic transitions. It couples a detailed methodology for rate calculations and background treatment with comprehensive scans against constraints from CDMS-Si, XENON10, and SIMPLE, while exploring the DAMA modulation data under channeling uncertainties. The key finding is that standard SI elastic scattering with equal couplings is strongly disfavored, but generalized couplings plus momentum dependence (and, in some cases, a light mediator or inelasticity) can reconcile CoGeNT with DAMA and other limits, offering concrete model-building directions. The work emphasizes that the interpretation hinges on experimental backgrounds and velocity-distribution assumptions, and points to future low-threshold experiments as crucial tests for these scenarios.

Abstract

Recently, the CoGeNT experiment has reported events in excess of expected background. We analyze dark matter scenarios which can potentially explain this signal. Under the standard case of spin independent scattering with equal couplings to protons and neutrons, we find significant tensions with existing constraints. Consistency with these limits is possible if a large fraction of the putative signal events is coming from an additional source of experimental background. In this case, dark matter recoils cannot be said to explain the excess, but are consistent with it. We also investigate modifications to dark matter scattering that can evade the null experiments. In particular, we explore generalized spin independent couplings to protons and neutrons, spin dependent couplings, momentum dependent scattering, and inelastic interactions. We find that some of these generalizations can explain most of the CoGeNT events without violation of other constraints. Generalized couplings with some momentum dependence, allows further consistency with the DAMA modulation signal, realizing a scenario where both CoGeNT and DAMA signals are coming from dark matter. A model with dark matter interacting and annihilating into a new light boson can realize most of the scenarios considered.

Figures (12)

• Figure 1: SI scattering with equal proton and neutron couplings, $f_p=f_n$. Plotted are the confidence interval for CoGeNT data shown in blue (90%) and cyan (99%), taking the constant and L-shell background Eq. (\ref{['eqn:simplebg']}). Also plotted are the 90% exclusion limits coming from CDMS-Si (red) and XENON10 (orange) and SIMPLE (brown). Also plotted are the DAMA confidence intervals shown in magenta (90%) and green (99%), where the filled (unfilled) contours assume 100% (no) channeling.
• Figure 2: Best fit (red-solid) for the low energy CoGeNT data with a WIMP signal component (green light-dashed). The WIMP best-fit parameters are $m_{ \chi} = 9.4~\mathrm{GeV}$ and $\sigma_{\rm n} = 0.84\times 10^{-40}~\mathrm{cm}^2$ and the velocity distribution was taken to be Maxwellian with $v_0 = 230~\mathrm{km}/\sec$, and $v_{\rm esc} = 500~\mathrm{km}/\sec$. Notice that the Gaussian peaks from the L-shell background component (blue-dashed) do not contribute to the first 5 energy bins from the left. This part of the spectrum is entirely explained by the WIMP contribution.
• Figure 3: Confidence interval for CoGeNT data for $0\%$, $10\%$, $30\%$, and $50\%$ exponential background contribution, solid, dashed, dotted, dot-dashed, respectively. Below the green light-dashed line DM fails to explain 50% of the signal in at least one bin ( see text). Other curves and contours are labeled as in figure \ref{['fig:nobgcontours']}.
• Figure 4: On the left we plot a fit (red-solid) for the low energy CoGeNT data with a WIMP signal component (green-dashed) and a $50\%$ exponential background (blue-dotted). The WIMP parameters are $m_{ \chi} = 7~\mathrm{GeV}$ and $\sigma_{\rm n} = 0.64\times 10^{-40}~\mathrm{cm}^2$ and the velocity distribution was taken to be Maxwellian with $v_0 = 230~\mathrm{km}/\sec$, and $v_{\rm esc} = 500~\mathrm{km}/\sec$. On the right, the fit's WIMP component (dashed-green) in each bin is stacked on top of the residual in that bin (solid-blue). The residual is simply the WIMP component subtracted from the data at every bin. It is evident that a very large background (i.e. non-WIMP) component is needed at low energy in order to accommodate such a light WIMP.
• Figure 5: Spin independent scattering with $f_n = -f_p$. Contours and lines are labeled as in figure \ref{['fig:nobgcontours']}, with the addition of the XENON10 limit from varying $\mathcal{L}_{eff}$ at 1$\sigma$ (orange thin).