Dark Matter Search Results from a One Ton-Year Exposure of XENON1T

We report on a search for weakly interacting massive particles (WIMPs) using 278.8 days of data collected with the XENON1T experiment at LNGS. XENON1T utilizes a liquid xenon time projection chamber with a fiducial mass of $(1.30\pm 0.01)\mathrm{t}\mathrm{o}\mathrm{n}$, resulting in a 1.0 ton yr exposure. The energy region of interest, $[1.4,10.6]{\mathrm{keV}}_{\mathrm{ee}}$ ($[4.9,40.9]{\mathrm{keV}}_{\mathrm{nr}}$), exhibits an ultralow electron recoil background rate of $[{82}_{-3}^{+5}(\mathrm{syst})\pm 3(\mathrm{stat})]\mathrm{events}/(\mathrm{t}\mathrm{o}\mathrm{n}\mathrm{yr}{\mathrm{keV}}_{\mathrm{ee}})$. No significant excess over background is found, and a profile likelihood analysis parametrized in spatial and energy dimensions excludes new parameter space for the WIMP-nucleon spin-independent elastic scatter cross section for WIMP masses above $6\mathrm{GeV}/c^2$, with a minimum of $4.1\times {10}^{-47}{\mathrm{cm}}^2$ at $30\mathrm{GeV}/c^2$ and a 90% confidence level.

Figure 1

Best-fit total efficiencies (black), including the energy ROI selection, for SR0 (dashed) and SR1 (solid) as a function of true NR energy (${\mathrm{keV}}_{\mathrm{nr}}$). The efficiency of $S1$ detection (green) and that of both $S1$ detection and selection (blue) are shown. The shaded bands show the 68% credible regions for SR1. The expected spectral shapes (purple) of $10\mathrm{GeV}/c^2$ (dashed), $50\mathrm{GeV}/c^2$ (dotted), and $200\mathrm{GeV}/c^2$ (dash-dotted) WIMPs are overlaid for reference.

Figure 2

Spatial distributions of DM search data. Events that pass all selection criteria and are within the fiducial mass are drawn as pie charts representing the relative probabilities of the background and signal components for each event under the best-fit model (assuming a $200\mathrm{GeV}/c^2$ WIMP and a resulting best-fit ${\sigma }_{\mathrm{SI}}=4.7\times {10}^{-47}{\mathrm{cm}}^2$) with the color code given in the legend. Small charts (mainly single-colored) correspond to unambiguously background-like events, while events with larger WIMP probability are drawn progressively larger. Gray points are events reconstructed outside the fiducial mass. The TPC boundary (black line), the 1.3 tons fiducial mass (magenta), the maximum radius of the reference 0.9 ton mass (blue dashed), and the 0.65 ton core mass (green dashed) are shown. Yellow shaded regions display the $1\sigma$ (dark) and $2\sigma$ (light) probability density percentiles of the radiogenic neutron background component for SR1.

Figure 3

DM search data in the 1.3 tons fiducial mass distributed in the ($\mathrm{c}S1$, $\mathrm{c}S{2}_b$) (left) and ($R^2$, $\mathrm{c}S{2}_b$) (right) parameter spaces with the same marker descriptions as in Fig. 2. Shaded regions are similar to Fig. 2, showing the projections in each space of the surface (blue) and ER (gray) background components for SR1. The $1\sigma$ (purple dashed) and $2\sigma$ (purple solid) percentiles of a $200\mathrm{GeV}/c^2$ WIMP signal are overlaid for reference. Vertical shaded regions are outside the ROI. The NR signal reference region (left, between the two red dotted lines) and the maximum radii (right) of the 0.9 ton (blue dashed) and 1.3 tons (magenta solid) masses are shown. Gray lines show isoenergy contours in NR energy.

Figure 4

Background and $200\mathrm{GeV}/c^2$ WIMP signal best-fit predictions, assuming ${\sigma }_{SI}=4.7\times {10}^{-47}{\mathrm{cm}}^2$, compared to DM search data in the 0.9 ton (solid lines and markers) and 1.3 tons (dotted lines and hollow markers) masses. The horizontal axis is the projection along the ER mean (${\mu }_{\mathrm{ER}}$), shown in Fig. 3, normalized to the ER $1\sigma$ quantile (${\sigma }_{\mathrm{ER}}$). Shaded bands indicate the 68% Poisson probability region for the total BG expectations.

Figure 5

90% confidence level upper limit on ${\sigma }_{\mathrm{SI}}$ from this work (thick black line) with the $1\sigma$ (green) and $2\sigma$ (yellow) sensitivity bands. Previous results from LUX [6] and PandaX-II [7] are shown for comparison. The inset shows these limits and corresponding $\pm 1\sigma$ bands normalized to the median of this work’s sensitivity band. The normalized median of the PandaX-II sensitivity band is shown as a dotted line.