Search for Low-Mass Weakly Interacting Massive Particles Using Voltage-Assisted Calorimetric Ionization Detection in the SuperCDMS Experiment

SuperCDMS is an experiment designed to directly detect weakly interacting massive particles (WIMPs), a favored candidate for dark matter ubiquitous in the Universe. In this Letter, we present WIMP-search results using a calorimetric technique we call CDMSlite, which relies on voltage-assisted Luke-Neganov amplification of the ionization energy deposited by particle interactions. The data were collected with a single 0.6 kg germanium detector running for ten live days at the Soudan Underground Laboratory. A low energy threshold of $170{\mathrm{eV}}_{\mathrm{ee}}$ (electron equivalent) was obtained, which allows us to constrain new WIMP-nucleon spin-independent parameter space for WIMP masses below $6\mathrm{GeV}/c^2$.

Figure 1

Recoil energy spectrum of WIMP-search events, after application of event-selection cuts. Inset: Low-energy spectrum in terms of raw counts (blue, solid histogram); also shown is the analysis efficiency (red, curve). Both are expressed in ${\mathrm{keV}}_{\mathrm{ee}}$. The analysis threshold of $170{\mathrm{eV}}_{\mathrm{ee}}$ is indicated by the vertical dot-dashed line. The resolution of the 1.3 keV line is $43{\mathrm{eV}}_{\mathrm{ee}}$ ($1\sigma$).

Figure 2

The efficiency-corrected WIMP-search energy spectrum is shown in ${\mathrm{keV}}_{\mathrm{nr}}$ and compared with expected rates for WIMPs with the most likely masses and cross sections suggested by the analysis of CoGeNT [8] (green, upper dashed curve) and CDMS II Si [10] data (red, lower dashed curve). Note that the $k=0.157$ Lindhard yield model was used to convert from an electron-equivalent to a nuclear-recoil-equivalent energy scale. The $170{\mathrm{eV}}_{\mathrm{ee}}$ ionization threshold translates to $841{\mathrm{eV}}_{\mathrm{nr}}$ (amber dot-dashed line). The $1.3{\mathrm{keV}}_{\mathrm{ee}}$ activation line appears at $\sim 5.3{\mathrm{keV}}_{\mathrm{nr}}$.

Figure 3

The 90% upper confidence limit from the data presented here are shown with exclusion limits from other experiments. These are grouped as Ge bolometers in blue: CDMS II Ge regular (dot dash) [39], CDMS II Ge low threshold (solid) [40], EDELWEISS II low threshold (dash) [37]; point-contact Ge detectors in purple: TEXONO (dash) [41], CDEX (dot dash) [42]; liquid Xenon in red: XENON100 (dot dash) [43], XENON10 S2 only (dash) [44], LUX (solid) [45]; and other technologies in magenta: low threshold reanalysis of CRESST II data (dot dash) [46], PICASSO (dash) [47]. The contours are from CDMS II Si (light and dark gray correspond to 68% and 90% C.L. regions, respectively) [10], CRESST II (blue) [9], DAMA (orange) [6, 7], CoGeNT (pink) [8].

Figure 4

The effect of the choice of the yield model on the 90% confidence level upper limit is shown. For the Lindhard model, the $0.170{\mathrm{keV}}_{\mathrm{ee}}$ analysis threshold corresponds to $0.84{\mathrm{keV}}_{\mathrm{nr}}$ ($k=0.157$), $1.1{\mathrm{keV}}_{\mathrm{nr}}$ ($k=0.1$), $0.73{\mathrm{keV}}_{\mathrm{nr}}$ ($k=0.2$). For the power-law model used by CoGeNT, the analysis threshold corresponds to $0.75{\mathrm{keV}}_{\mathrm{nr}}$.