SuperCDMS Public Documentation

HVeV Run 3

Abstract This Public Documentation accompanies the electron-recoil Dark Matter search, as well as the dark photon and axion-like particle absorption search, with gram-scale HVeV detectors operated 107 meters underground at the Northwestern Experimental Underground Site (Batavia, IL). 12.22 gram-days (8.48 gram-days after the live-time selection) of exposure was acquired at 100V bias. The resulting limits for DM-electron scattering and DM absorption are given below, along with plots approved for public release.
These results are published in (add reference as soon as available).

Num.	Topic	Last Updated	Download	Comments
All results assume a relic DM density of ρ_DM = 0.3 GeV/cm³. All limits are at 90% confidence level. $\Bar{\sigma}_e$ is the DM-electron scattering cross section. F_DM is the DM form factor where F_DM = 1 (F_DM ∝ 1/q²) corresponds to a DM model with a heavy (ultra light) dark photon mediator. q is the momentum transfer. m_DM is the DM mass. More detailed information on the assumed DM models can be found in the US Cosmic Visions Community Report. Astrophysical parameters can be found in Recommended conventions for reporting results from direct dark matter searches.
1	Experimental Setup	Dec 13^th, 2022	png, pdf	Detector setup diplayed using CAD A CAD image of the run 3 detector payload. The four detectors are split between the two vertical boards with one on the front and back side of each board. This design improves the likelihood of coincidence signals from the environment between each pair on board as well as across boards by their alignment in height. The chips are labeled by their detector mask name. The crossed out detector on the left is the NF-F detector mask that was not used in the run due to a short in detector.
2	Triggering	Mar 30^th, 2023	png, pdf	Gaussian filter trigger example A comparison of Gaussian Filter (GF) and Matching Filter (MF) triggering capabilities on a pile-up event in NFC detector.
3	Laser Data	Apr 3^rd, 2023	png, pdf	Laser OFL amplitude spectrum before and after laser live-time cuts Laser OFL amplitude spectrum after various live-time cuts. Used for calibration and chi-square cut development.
4	Detector Response	May 3^rd, 2023	png, pdf	Resolution-optimized channel weighting calculation (Top) Fitted 1-eh resolution (with uncertainty) for a range of weighting factors. This example uses science data taken on Jan 11, 2021 with the NF-C detector. Fit to model is shown in black. The weighting factor which provides minimum resolution is marked by a dotted line. (Bottom) Residuals with respect to the model fit, using the same legend.
5	Detector Response	Mar 28^th, 2023	png, pdf	First electron-hole pair peak amplitude variation. Fitted position of the first eh peak in each science dataset for the NF-C detector. The black line represents the fitted position of the first peak in the fully summed background data. The summed value is also used to calculate the percent variation.
6	Detector Response	Mar 28^th, 2023	png, pdf	Example gaussian fits to laser calibration data Example laser spectrum with cuts, from NF-C data taken on Feb 9, 2021. Black fits are gaussian fit results. Grey regions are initial fitting windows. The x-axis is roughly calibrated to energy values using a simple multiplicative factor. This is not the same as the final energy calibration.
7	Detector Response	Mar 28^th, 2023	png, pdf	Example gaussian fits to laser calibration data with background estimate Example laser spectrum with cuts, from NF-C data taken on Feb 9, 2021. Black fits are gaussian fit results. Grey regions are initial fitting windows. Blue points and fits were used to approximate the backgrounds and calculate the systematic uncertainties. The x-axis is roughly calibrated to energy values using a simple multiplicative factor. This is not the same as the final energy calibration.
8	Detector Response	Mar 28^th, 2023	png, pdf	NFC energy calibration fit Calibration fit results for the 3 laser datasets together.
9	Detector Response	Mar 28^th, 2023	png, pdf	NFC energy calibration fit residuals Calibration fit results for the 3 laser datasets together. y-direction residuals with the uncertainties on offset shown. Uncertainty bands for the fit, daily gain variation (sys), and both (combined in quadrature) are shown. Uses legend from "NFC energy calibration fit".
10	Detector Response	Mar 28^th, 2023	png, pdf	NFC energy calibration fit, zoomed in to each peak Calibration fit results for the 3 laser datasets together. Zoom-in's to see the fit around each peak, with all uncertainties and bands displayed. Uses legends from "NFC energy calibration fit" and "NFC energy calibration fit residuals".
11	Live-time cuts	Mar 30^th, 2023	png, pdf	Mean baseline histogram The distribution of mean-baseline of MIDAS traces in one of the DM search series. The 3-sigma cut is shown as a green shaded region.
12	Live-time cuts	Mar 30^th, 2023	png, pdf	Mean baseline example An example of MIDAS-trace mean baseline as a function of time. MIDAS traces with elevated baseline removed by the cut are shown in orange.
13	Live-time cuts	Mar 30^th, 2023	png, pdf	Time between NFC events and nearest triggers Distributions of time intervals between triggers in NFC and the closest triggers in the detector, specified in the legend. Positive (negative) values mean that the event under consideration arrived after (before) the closest event in the veto detector. A +-20 ms window, excluded by the live-time cut, is shown in red.
14	Data-quality cuts	Dec 13^th, 2022	png, pdf	Frequency-domain Χ² cut The frequency-domain Χ² cut belongs to the data-quality cuts and was studied with the laser data. This plot shows the frequency-domain Χ² as a function of the energy. The magenta dots represent 3σ position of the chi2 distribution around each peak. The magenta line corresponds to the position of the cut.
15	Data-quality cuts	Feb 28^th, 2023	png, pdf	Simulated comparison of chi-square and delta-chi-square cuts The histogram of real timing separations for each signal in the simulated data sets. The remaining distribution after the chi-square cut is shown in orange and after the delta chi-square cut is shown in light blue. The peak at zero is all true non-pileup events. The peak at 30 samples is a binning artifact.
16	Data-quality cuts	Dec 13^th, 2022	png, pdf	Frequency-domain delta Χ² cut The frequency-domain delta Χ² cut belongs to the data-quality cuts and was studied with the laser data. This plot shows the frequency-domain delta Χ² as a function of the energy. The magenta dots represent 3σ position of the chi2 distribution around each peak. The magenta line corresponds to the position of the cut.
17	Data-quality cuts	February 21^st, 2020	png, pdf	Total quality-cut efficiency Efficiency as a function of the energy for the science data. The efficiency is fit with a flat curve. The fit curve and its uncertainty are used to set the limit.
18	Data	Apr 3^rd, 2023	png, pdf	Dark-matter-search data energy spectrum Dark-matter-search data energy spectrum before and after various cuts with the live-time reduction and efficiency taken into account. This uses 100% of the science data (rather than being partitioned like the limit setting).
19	Data	Mar 30^th, 2023	png, pdf	Dark-matter-search data energy spectrum Dark-matter-search data energy spectrum before and after cuts. The data are plotted in three cases: (1) after the calibration before the cuts; (2) after the live-time cuts; (3) after the live-time and data-quality cuts, corrected by the efficiency. This uses 100% of the science data (rather than being partitioned like the limit setting).
20	Laser Data	Apr 3^rd, 2023	png, pdf	Laser data energy spectrum Laser data energy spectrum before and after various cuts.
21	Detector Response	Jun 1^st, 2023	png, pdf	Energy resolution measurements Peak widths in laser (orange) and DM-search datasets (blue) and the energy-independent resolution model (black) with systematic uncertainty (gray) used in the limit-setting. The nominal resolution value is 4.26 eV, the uncertainty boudnries are 3.03 - 5.49 eV.
22	Detector Response	Mar 28^th, 2023	png, pdf	Energy resolution stability in 1 electron-hole pair peak Series-by-series trend of the width of the first electron-hole pair peak in the DM-search data. The width for all science datasets combined is shown as the black line, with uncertainty shown in grey.
23	Data	Jun 19^th, 2023	png, pdf	HVeV Run 3 data with cuts and energy calibration Calibrated 10% and 90% data spectra, post all live-time and data-quality cuts. The 10% data corresponds to an exposure of 0.85 gramdays. The 90% data corresponds to an exposure of 7.63 gramdays.
24	Data	Jun 19^th, 2023	png, pdf	HVeV Run 3 data compared to Run 2 data Calibrated 10% and 90% data spectra, post all live-time and data-quality cuts, converted to DRU for comparison to the spectrum used for the R2 limit setting.
25	Modeling	May 31^st, 2023	png, pdf	Example of the dark matter signal model. Example signal model for DMe (form factor = 1) with a DM mass of 1 GeV and resolution of 4.26 eV. The crystal bias is 100 V. Trapping and impact ionization contribute events to the between-peak regions. Note that "ct" refers to the charge trapping and "if" refers to the impact-ionization fraction for both electron and hole production (0.8% probability of producing an electron, and a separate 0.8% probability of producing a hole). The mass, cross-section, x-scale, and y-scale were chosen to match those in the Run 2 documentation's Figure 6.1.2.
26	Limits	May 31^st, 2023	png, pdf	Example of the peak-selection process for dark-photon absorption (top) The 10%-data dark-photon absorption limits used to determine the peak selections. Each eh limit is filled between its under- and over-fluctuation results. (bottom) The peak selection based on the above limits. The colors correspond to the legend in the top plots. If the bottom of a limit's envelope is lower than the top of all other envelopes, that limit is included in the selection for that mass.
27	Limits	Jun 20^th, 2023	png, pdf	Dark-matter-electron scattering limits with form factor = 1 DM-electron scattering limits with momentum-transfer form factor = 1, produced using the R3 peak selection and poisson limit setting methods. These limits use the newer R3 halo parameters. Systematic-uncertainty sensitivity bands are generated using 5000 variations on the input parameters. The bands are difficult to see in the limit plots, and so are also plotted as residuals (in percentage of the best-fit limit). Comparison limits are taken from the CDMS limit plotter (13 Oct 2022).
28	Limits	Jun 20^th, 2023	png, pdf	Dark-matter-electron scattering limits with form factor = 1/q² DM-electron scattering limits with momentum-transfer form factor = 1/q², produced using the R3 peak selection and poisson limit setting methods. These limits use the newer R3 halo parameters. Systematic-uncertainty sensitivity bands are generated using 5000 variations on the input parameters. The bands are difficult to see in the limit plots, and so are also plotted as residuals (in percentage of the best-fit limit). Comparison limits are taken from the CDMS limit plotter (13 Oct 2022).
29	Limits	Jun 20^th, 2023	png, pdf	Dark-photon absorption limits Dark-photon absorption limits, produced using the R3 peak selection and poisson limit setting methods. Systematic-uncertainty sensitivity bands are generated using 5000 variations on the input parameters. The bands are difficult to see in the limit plots, and so are also plotted as residuals (in percentage of the best-fit limit). Comparison limits are taken from the CDMS limit plotter (13 Oct 2022).
30	Limits	Jun 20^th, 2023	png, pdf	Axion-like particle absorption limits Axion-like particle absorption limits, produced using the R3 peak selection and poisson limit setting methods. Systematic-uncertainty sensitivity bands are generated using 5000 variations on the input parameters. The bands are difficult to see in the limit plots, and so are also plotted as residuals (in percentage of the best-fit limit). Comparison limits are taken from the CDMS limit plotter (13 Oct 2022). We modified the R2 limit by the square-root of the silicon index of refraction, to make it a valid comparison to our model.
31	Limits	Jun 5^th, 2023	png, pdf	Dark-matter-electron scattering exclusion region with form factor = 1 DM-electron scattering limits with momentum-transfer form factor = 1, produced using the R3 peak selection and poisson limit setting methods. The upper boundary of the exclusion region is calculated by taking into account the Earth shielding effect using DaMaSCUS-CRUST.
32	Limits	Jun 5^th, 2023	png, pdf	Dark-matter-electron scattering exclusion region with form factor = 1/q² DM-electron scattering limits with momentum-transfer form factor = 1/q², produced using the R3 peak selection and poisson limit setting methods. The upper boundary of the exclusion region is calculated by taking into account the Earth shielding effect using DaMaSCUS-CRUST.
33	Limits	Jul 20^th, 2023	png, pdf	Dark-photon absorption limit with overburden comparison Dark-photon absorption limit and the parameter-space where we estimate Earth-shielding to have a significant effect. The hashed region is where 107 meters (the NEXUS depth) of silicon overburden would attenuate the DM flux by the factor e (or more). Silicon is used to approximate the overburden because its complex conductivity is known down to the required energy.
34	Limits	Jul 20^th, 2023	png, pdf	Axion-like particle absorption limit with overburden comparison Axion-like particle absorption limit and the parameter-space where we estimate Earth-shielding to have a significant effect. The hashed region is where 107 meters (the NEXUS depth) of silicon overburden would attenuate the DM flux by the factor e (or more). Silicon is used to approximate the overburden because its complex conductivity is known down to the required energy.