Search for low-mass WIMPs in a 0.6 kg day exposure of the DAMIC experiment at SNOLAB

We present results of a dark matter search performed with a 0.6 kg d exposure of the DAMIC experiment at the SNOLAB underground laboratory. We measure the energy spectrum of ionization events in the bulk silicon of charge-coupled devices down to a signal of 60 eV electron equivalent. The data are consistent with radiogenic backgrounds, and constraints on the spin-independent WIMP-nucleon elastic-scattering cross section are accordingly placed. A region of parameter space relevant to the potential signal from the CDMS-II Si experiment is excluded using the same target for the first time. This result obtained with a limited exposure demonstrates the potential to explore the low-mass WIMP region ($<10\mathrm{GeV}{c}^{-2}$) with the upcoming DAMIC100, a 100 g detector currently being installed in SNOLAB.

Figure 1

Copper module holding an 8 Mpix CCD being installed in the low-radioactivity copper box. Two other modules have already been installed and can be partially seen at the bottom of the box. The flex cables that carry the CCD signals are also visible.

Figure 2

(a) Cross-sectional representation of the charge produced by a pointlike ionization event (star) in the CCD bulk as it is drifted to the pixel array. (b) The $x–y$ distribution of charge on the pixel array following the ionization event. Because of diffusion, the charge is collected in multiple pixels, with the lateral spread (${\sigma }_{xy}$) being positively correlated with the depth ($z$ coordinate) of the interaction. When the CCD is read out in the $1\times 1$ configuration, this is the pattern observed in the image. (c) In the $1\times 100$ mode, the CCD is read out in column segments 100 pixels tall, collapsing the pixel contents along the $y$ axis, leading to a one-dimensional pattern with the charge spread out over fewer pixels. The one-dimensional lateral spread (${\sigma }_x$) is positively correlated to the depth of the interaction.

Figure 3

Linear constant $k$ relating the CCD output signal to the ionization generated in the substrate. Values are given relative to $k$ measured at $5.9{\mathrm{keV}}_{\mathrm{ee}}$. Calibrations at high energies were performed with x rays, while the lowest energy points were obtained using optical photons, as outlined in the text. The linearity of the CCD energy response is demonstrated down to $40{\mathrm{eV}}_{\mathrm{ee}}$.

Figure 4

A MIP observed in cosmic ray background data acquired on the surface. Only pixels whose values are above the noise in the image are colored. The large area of diffusion on the top left corner of the image is where the MIP crosses the back of the CCD. Conversely, the narrow end on the bottom right corner is where the MIP crosses the front of the device. The reconstructed track is shown by the long-dashed line. The short-dashed lines show the $3\sigma$ band of the charge distribution according to the best-fit diffusion model.

Figure 5

Example of the pixel value distribution after image processing in one 30 ks exposure (black) and its corresponding blank (blue) acquired in December 2014. The noise in the image is fitted to ${\sigma }_{\mathrm{pix}}=1.8e^{-}$.

Figure 6

$\mathrm{\Delta }LL$ distributions for all clusters in the $1\times 1$ data set. The blue line shows the distribution for clusters in the blanks, which are representative of the contribution from readout noise to the data set. The black dashed line presents the expected distribution (from simulation) of ionization events that occur uniformly in the CCD bulk, assuming a constant (flat) energy spectrum. The black line shows the distribution for all clusters in the $1\times 1$ data set. The dashed red line is the fit done to the tail of the noise distribution to determine the selection used to reject readout noise. The fit is statistically consistent with the tail of the distribution.

Figure 7

Lateral spread (${\sigma }_{xy}$) versus measured energy ($E$) of the clusters that pass the selection criteria outlined in Sec. 6. Black (red) markers correspond to candidates in the $1\times 1$ ($1\times 100$) data set. Gray markers show the simulated distribution of energy deposits near the front and back surfaces of the device. The projection on the ${\sigma }_{xy}$ axis of the identified clusters is shown on the right. The horizontal dashed lines represent the fiducial selection described in Sec. 7, while the vertical dashed line shows the upper bound of the WIMP search energy range.

Figure 8

Spectrum from ${}^{57}\mathrm{Co}$ source calibration in the laboratory after event selection to remove readout noise and surface events, as performed in the WIMP search. The event rate has been normalized to the absolute rate expected in the energy interval $0.5–1.5{\mathrm{keV}}_{\mathrm{ee}}$. The spectrum is taken as a direct measurement of the detection efficiency because the Compton scattering spectrum at these low energies is approximately constant. The simulated detection efficiency, including the fit with the functional form used for the WIMP search analysis, is shown.

Figure 9

Final detection efficiency of events as a function of reconstructed energy ($E$) for the $1\times 1$ (black) and $1\times 100$ (red) data sets after cluster selection outlined in Secs. 6 and 7. Solid lines present the acceptance of the WIMP signal, while dashed lines present the detection efficiency of background events considering both bulk and surface contributions.

Figure 10

Energy spectrum of the final candidates in the $1\times 1$ and $1\times 100$ data sets. The red line shows the best-fit model with parameters $s_{\mathrm{tot}}=0$, $b_{1\times 1}=31$, and $b_{1\times 100}=23$.

Figure 11

Upper limit (90% C.L.) on the WIMP-nucleon cross section ${\tilde{\sigma }}_{\chi -\mathrm{n}}$ derived from this analysis (red line). The expected sensitivity $\pm 1\sigma$ is shown by the red band. For comparison, we also include 90% C.L. exclusion limits from other experiments [3, 17] and the 90% C.L. contours corresponding to the potential WIMP signals of the CDMS-II Si [3] and DAMA [18] experiments.