Direction-sensitive dark matter search with a low-background gaseous detector NEWAGE-0.3b''
Tomonori Ikeda, Kiseki Nakamura, Takuya Shimada, Ryota Yakabe, Takashi Hashimoto, Hirohisa Ishiura, Takuma Nakamura, Hiroshi Ito, Koichi Ichimura, Ko Abe, Kazuyoshi Kobayashi, Toru Tanimori, Hidetoshi Kubo, Atsushi Takada, Hiroyuki Sekiya, Atsushi Takeda, Kentaro Miuchi
PProg. Theor. Exp. Phys. , 00000 (14 pages)DOI: 10.1093 / ptep/0000000000 Direction-sensitive dark matter search with alow-background gaseous detector
NEWAGE-0.3b”
Tomonori Ikeda , Kiseki Nakamura , Takuya Shimada , Ryota Yakabe , TakashiHashimoto , Hirohisa Ishiura , Takuma Nakamura , Hiroshi Ito , Koichi Ichimura ,Ko Abe , Kazuyoshi Kobayashi , Toru Tanimori , Hidetoshi Kubo , AtsushiTakada , Hiroyuki Sekiya , Atsushi Takeda , and Kentaro Miuchi Department of Physics, Graduate School of Science, Kobe University, Rokkodai-cho,Nada-ku, Kobe-shi, Hyogo, 657-8501, Japan ∗ E-mail: [email protected] Kamioka Observatory, Institute for Cosmic Ray Research, the University of Tokyo,Higashi-Mozumi, Kamioka-cho, Hida-shi, Gifu, 506-1205, Japan Research Center for Neutrino Science, Tohoku University, Sendai 980-8578, Japan Kavli Institute for the Physics and Mathematics of the Universe (WPI), theUniversity of Tokyo, 5-1-5 Kashiwanoha, Kashiwa-shi, Chiba, 277-8582, Japan Waseda Research Institute for Science and Engineering, Waseda University, 3-4-1Okubo, Shinjuku, Tokyo 169-8555, Japan Division of Physics and Astronomy, Graduate School of Science, Kyoto University,Kitashirakawaoiwake-cho, Sakyo-ku, Kyoto-shi, Kyoto, 606-8502, Japan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
NEWAGE is a direction-sensitive dark matter search using a low-pressure gaseous timeprojection chamber. A low alpha-ray emission rate micro pixel chamber had been devel-oped in order to reduce background for dark matter search. We conducted the darkmatter search at the Kamioka Observatory in 2018. The total live time was 107.6 dayscorresponding to an exposure of 1.1 kg · days. Two events remained in the energy region of50-60 keV which was consistent with 2.5 events of the expected background. A directionalanalysis was carried out and no significant forward-backward asymmetry derived fromthe WIMP-nucleus elastic scatterings was found. Thus a 90% confidence level upper limiton Spin-Dependent WIMP-proton cross section of 50 pb for a WIMP mass of 100 GeV/ c was derived. This limit is the most stringent yet obtained from direction-sensitive darkmatter search experiments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Subject Index Dark matter, MPGD, µ TPC
1. Introduction
Dark matter is one of the biggest puzzles of the modern cosmology and the particle physics.A number of experimental efforts aiming to find the Weakly Interacting Massive Particle(WIMP) dark matter through direct searches which observe the scatterings of the WIMPand nuclei have been carried out [1–8]. However, the dark matter has not been discoveredyet. In the direct search, the annual modulation and the directional signature would be twopossible signals among the characteristic signals of the dark matter. The annual modulation © The Author(s) 2012. Published by Oxford University Press on behalf of the Physical Society of Japan.This is an Open Access article distributed under the terms of the Creative Commons Attribution License(http://creativecommons.org/licenses/by-nc/3.0), which permits unrestricted use,distribution, and reproduction in any medium, provided the original work is properly cited. a r X i v : . [ h e p - e x ] J a n s caused by the orbital motion of the Earth around the Sun. The modulation amplitude isexpected to be a few % [9]. On the other hand, the directional signature is due to the circularmotion of the solar system around the galaxy center. The forward-backward ratio in nuclearrecoil angular distribution derived from the WIMP-nucleus elastic scatterings could be anorder of magnitude [10]. In addition, the directional method could discover the WIMP darkmatter beyond the neutrino floor, which represents the ultimate background by the coherentneutrino-nucleus scatterings [11], and reveal the astrophysical and particle properties of thedark matter [12–14].NEWAGE (NEw generation WIMP search with an Advanced Gaseous tracker Experiment)is a direction-sensitive WIMP dark matter search experiment using a low pressure gas microtime projection chamber ( µ TPC) [15]. A direction-sensitive experiment needs to detect thedirection of the recoil nuclei. Hence a TPC at low pressure gas and a readout device, µ -PIC [16], which is one variation of micro pattern gaseous detectors, are used. In 2015, anunderground measurement was performed and the best directional constraint testing theforward-backward asymmetry in the nuclear recoil angular distribution was achieved [17].Then we increased the statistics by a factor of more than ten and the first 3d-vector direc-tional dark matter search was performed [18]. However, a certain amount of radioactivity,which potentially contributed to the background, was later found inside the µ -PIC. There-fore the surface material of the µ -PIC, which was the dominant background source, wasreplaced with less radioactive material. This newly developed low-background µ -PIC wascalled a low alpha-ray emission rate micro pixel chamber (LA µ -PIC) [19]. In this paper, thefirst results of the direction-sensitive dark matter search using the LA µ -PIC are reported.
2. NEWAGE-0.3b” detector
The NEWAGE-0.3b’ detector was upgraded to NEWAGE-0.3b” by replacing the readoutdevice from the standard µ -PIC to LA µ -PIC. Most part of the detector system is unchangedand we briefly summarize the structure of the NEWAGE-0.3b” detector and its performancein this section. The NEWAGE-0.3b” detector is a low-pressure gaseous µ TPC that is comprised of a LA µ -PIC [19], a gas electron multiplier (GEM [20]) and a TPC cage. The schematic drawing isshown in Fig. 1. The LA µ -PIC was manufactured by Dai Nippon Printing Co. Lt.. It has768 ×
768 pixels with a pitch of 400 µ m forming the detection area of 30.7 × .These electrodes are connected by 768 anode strips and 768 cathode strips. The anode andcathode strips of LA µ -PIC are orthogonally formed and thus a two-dimensional position ofa hit pixel can be known. The structure of the pixel electrode is same as that of a standard µ -PIC [16], while the material facing the detection volume was changed. The new materialis a compound of epoxy and polyimide without the glass fiber, which is a factor of onehundred less contaminated by isotopes of U and
Th. A performance comparable tothe standard µ -PIC was confirmed [21]. A GEM with an effective area of 32 ×
31 cm isused as a first stage amplifier in order to obtain a sufficient gas gain while keeping a stableoperation. The substrate of the GEM is a 100 µ m thick liquid crystal polymer. Cylindricalholes are formed using the laser etching technique. A hole size and pitch are 70 µ m and140 µ m, respectively [22]. The TPC field cage, which is made of four plates of polyether ther ketone plastic plates, was installed in order to make a uniform electric field. Copperwires with a spacing of 1 cm are placed on the side walls and chained by resistors. The TPCcage has a length of 41 cm. The vessel was filled with CF at 76 Torr. CF is chosen becauseof its small diffusion and a large cross section for the Spin-Dependent (SD) interaction offluorine. A gas circulation system with a cooled charcoal (TSURUMICOAL 2GS) of 100 gwas installed in order to remove radons. c m Gas circulationCupper wires41 cmLAμ-PIC(470 V) GEM(-280 V,-530 V) Drift plane(-4.10 kV)CF
76 Torr ZY 𝜃 " B glass plateMicroscope picture of LAμ-PIC
400 μm 𝜙 %&’ XYCross section view of LAμ-PIC anode cathodePI (w/o GC) 5 μmEpoxy 75 μm
Fig. 1
Schematic drawings of the NEWAGE-0.3b” detector. The left-top microscope pic-ture shows the pixel electrode structure of the LA µ -PIC. Left-bottom drawings is the crosssection of the LA µ -PIC. Surface material is combination of epoxy and polyimide withoutthe glass-cloth-sheet (GC). NEWAGE-0.3b” chamber is shown in the right panel. LA µ -PIChas a detection area of 31 ×
31 cm and the drift length is 41 cm. The gap region is definedas a volume between the LA µ -PIC and the GEM. The B glass plate is set inside the TPCfield cage for the energy calibration.The charge signals are read out from the anode and cathode electrodes with Amplifier-Shaper-Discriminator chips (SONY CXA3653Q [23]). These signals are divided into digitizedand analog signals. The digitized signals are send to the FPGA-based encoding system andthe hit-patterns are recorded with a clock of 100 MHz. In addition, the time-over-threshold(TOT) of each strip are recorded. The track-length and direction are reconstructed usingthese information (see Ref. [18] for details). Analog signals of the 768 cathode strips aregrouped into four channels and each channel is then divided into two. One of the dividedsignals is directly connected to a waveform digitizer (REPIC RPV160, 100 MHz), while theother is attenuated by a factor of three and read by the waveform digitizer. The waveformsare mainly used to determine the energy deposit of each event. he energy calibration is performed with alpha-rays generated in a B( n, α ) Li reaction.A glass plate with a thin B layer is installed inside the µ TPC. The detector is irradiatedwith neutrons from a
Cf fission source placed outside of the vessel and the neutronsare thermalized by polyethylene blocks. A continuous spectrum with an maximum edge at1.5 MeV because of the thickness of the B layer is obtained. The measured spectrumis fitted with simulated ones and the calibration factor converting the charge to energy isdetermined.
The following four analysis cuts were applied in order to select nuclear recoil events. ◦ Fiducial cut: The fiducial area of XY plane was defined as 28 ×
24 cm . The wholepart of the track is required to be in the fiducial volume. This cut removes the chargedparticles from the wall and the B glass plate. ◦ Length-Energy cut: The stopping powers of nuclei are larger than those of electrons.Hence the track-length vs. energy distribution can be used to identify the nuclear recoiltracks. Figure 2 (a) shows the track-length vs. energy distributions for the
Cf calibra-tion (black points) and the
Cs calibration (blue points). From the
Cf calibrationdata, we determined nuclear band by fitting with Gaussian function for every energybin with a width of 10 keV. This cut removes electron and alpha-rays. ◦ TOT-Energy cut: Energy deposition on a single strip is stored as TOT. Since nuclearrecoil events have larger energy losses than electron events, sum of TOT (TOT-sum)for a given energy tends to be large. In addition, since TOT-sum was expected to belinear with respect to the total energy deposit, a parameter defined by dividing theTOT-sum by energy was suitable for this purpose. Figure 3 (a) shows TOT-sum/energyvs. energy distributions for the
Cf calibration (black points) and the
Cs calibration(blue points). From the
Cf calibration data, we determined the nuclear band by fittingwith gaussian function every energy bin with a width of 10 keV. This cut rejects electronevents. ◦ Roundness cut: The roundness is defined as a reduced chi-square value of a linear fitto the track. The roundness tends to be small for an event traveling short drift length,since the effect of electron diffusion is small and the track is fitted with a straight linewell. Hence this parameter is known to have a correlation to the absolute z position.Especially, this parameter is useful to remove “gap events”. The gap events are definedas events depositing their all energy in the gap region (Fig. 1). These gap events arenot amplified by the GEM and thus the measured charge is smaller by a factor of theGEM gain than those in the detection volume. In order to reproduce the gap events, weirradiated the detector with neutrons from a Cf fission source without a drift electricfield. Figure 4 (a) shows the roundness vs. energy distribution for the gap events. Theroundness of gap events (red points) were found to be small. The events with roundness > µ TPC with neutrons froma
Cf fission source. In order to cancel the position dependence of µ TPC and have isotropicrecoils, neutrons from a
Cf fission source were irradiated from several directions and thecombined spectrum was used. The detection efficiency was calculated by comparing the r a ck Leng t h ( c m ) T r a ck Leng t h ( c m ) Energy (keV)Energy (keV)
Fig. 2
Track-length vs. energy distributions. (a) The gradation and red points representthe
Cf neutron calibration data and the
Cs electron calibration data, respectively. Thered solid and dotted lines indicate the median and ± σ quantiles of the neutron calibration.(b) Scientific RUN data (DM RUN) after the fiducial cut. T O T - s u m / E ne r g y T O T - s u m / E ne r g y Energy (keV)Energy (keV)
Fig. 3
TOT-sum/energy vs. energy distributions. (a) The gradation and red points rep-resent
Cf neutron calibration data and
Cs electron calibration data, respectively. Thered solid and dotted lines indicate the median and ± σ quantiles of neutron calibration.(b) Scientific RUN data (DM RUN) after the fiducial cut.measured energy and the simulated one. The measured detection efficiencies of the nuclearrecoil events are shown in Fig. 5. The black and red lines are fitted lines of experimentaldata and indicate the detection efficiency when we applied only the fiducial cut and all eventselections. The efficiency with all event selections was 14% at 50 keV. oundne ss R oundne ss Energy (keV)Energy (keV)
Fig. 4
Roundness vs. energy distribution. (a) The gradation and red points representthe
Cf neutron calibration with a drift electric field and without a drift electric field,respectively. The cyan dotted line (roundness= 0 .
05) indicate the nuclear selection line.(b) Scientific RUN data (DM RUN) after the fiducial cut.The µ TPC has non-isotropic response with regard to the nuclear recoil track direc-tion because of the track-reconstruction algorithm. Thus we need to measure the relativedirection-dependent efficiency of nuclear recoil in the energy range of 50-100 keV. Figure 6shows the measured distribution of the elevation angle θ ele and the azimuth angle φ azi inthe detector coordinate. The efficiency is low around XY plane, XZ plane and YZ planein any energy region because of the poor track reconstruction along anode/cathode strips.In addition, there are higher efficiency areas along the diagonal lines from the direction ofthe anode and the cathode strips. This is because the current tracking algorithm tends torecognize the diffused tracks to such directions. We evaluated the gamma rejection power,or the detection efficiency of electrons, by irradiating the detector with gamma-rays from a Cs source. The gamma rejection power or the electron detection efficiency for the energyrange of 50-60 keV was 1 . +3 . − . × − .The angular resolution was evaluated using the neutron-nuclei elastic scatterings of samemethod in previous study [24]. The angular θ is defined as the angle between the directionof the scattered nuclei and the neutron source. We evaluated the angular resolution by thecomparison of measured and simulated distributions of the recoil angle cos θ . The obtainedangular resolution was 48.0 +6 . − . degree in the energy range of 50-100 keV. The expectedforward-backward ratio of | cos θ cygnus | for the 100 GeV/c WIMP is 30%. Here θ cygnus isdefined as the angle between the WIMP-wind direction and the measured direction of therecoil nucleus.
00 200 300 400Energy (keVe.e)00.20.40.60.8 E ff i c i en cy DetectionSelection
Fig. 5
Typical detection efficiency for the nuclear recoil events. The gray and black pointswith errors are experimental data after the fiducial cut and the roundness cut, respectively.The fitted black solid line is the detection efficiency in the fiducial volume. The fitted redsolid line is the detection efficiency after all event selections. (a) 50-60 keV (b) 60-70 keV (c) 70-80 keV(d) 80-90 keV (e) 90-100 keV (f) 50-100 keV
Fig. 6
Relative direction-dependent efficiencies in each energy region. The axis with whiteand black label are the azimuth angle φ azi and the elevation angle θ ele in the detectorcoordinate, respectively. . Experiment A directional dark matter search (RUN22) was carried in Laboratory B of the KamiokaObservatory (36.25’N, 137.18’E) located at 2700 m water equivalent underground. The LA µ -PIC plane was placed vertically and the z -axis is aligned to the direction of S30 ◦ E. The firstsub RUN was performed from Jun. 6th 2018 to 24th Aug. 2018 and the second sub RUNwas carried out from 20th Sep. 2018 to 14th Nov. 2018. The target gas is CF at 76 Torr(0.1 bar) and the target mass in the fiducial volume of 28 × ×
41 cm (28 L) is 10 g. Thetotal live time is 107.6 days corresponding to an exposure of 1.1 kg · days.The energy calibration and the detection efficiency measurement were carried out everytwo weeks. The gas gain at the beginnings of the sub RUNs was about 1100 and time-dependent variation was observed due to the gas deterioration. The energy scale of the datawas corrected considering the time-dependence of the gas gain. The energy resolution were13 . ± .
3% above 50 keV. The measured drift velocity at the beginnings of the sub RUNswas 9.6 cm/ µ s and the gas deterioration gave +3 .
3% and −
12% of the maximum uncertainty.Length-Energy, the TOT-Energy and the Roundness-Energy distributions after the fiducialcut for the dark matter search data are shown in Fig. 2 (b), 3 (b) and 4 (b), respectively.A large fraction of the events have long track lengths and small TOTs, which indicates thatmost of the measured events are electrons. These events are effectively reduced by analysiscuts introduced in Sec. 2.2. Energy spectra at each cut stage are shown in Fig. 7 (a). Thefinal event sample was reduced to 17 events in energy region of 50-400 keV owing to analysiscuts using track information without any shields like Pb. Since the directionality are lostin the energy below 50 keV because of the short tracks and the diffusion effect, the lowerenergy bound is set at 50 keV. Energy spectrum for the final sample unfolded by the nucleardetection efficiency is shown in Fig. 7 (b) together with the one of the previous resultsRUN14 [17] using the standard µ -PIC. The main background of RUN14 in 50-100 keV arealpha-rays radiated from the surface material of the standard µ -PIC. These backgroundswere reduced in RUN22 by about factor 10 thanks to the LA µ -PIC whose surface materialis less contaminated with U and
Th. Here it is demonstrated that the LA µ -PIC worksas expected to reduce the alpha-ray backgrounds. Figure 8 shows the skymap of the finalsample on the detector coordinate. We calculated the nuclear recoil distribution | cos θ cygnus | for the energy region of 50-100 keV in order to evaluate the forward-backward asymmetry.Figure. 9 shows the measured and expected | cos θ cygnus | distribution binned into two foreach energy range. The systematic uncertainties of the expected rate for the WIMP weresummarized in Table 1. The angular resolution give the dominant systematic uncertaintywhich impact on the shape of the nuclear recoil distribution and considered in the followingstatistic test.
4. Results
In order to obtain a possible anisotropic | cos θ cygnus | distribution, a binned likelihood-ratiomethod was used [25]. The minimized statistic value χ was defined as, χ = 2 n (cid:88) i =0 (cid:20) ( N exp i − N data i ) + N data i ln (cid:18) N data i N exp i (cid:19)(cid:21) + α , (1)where the subscript i is the bin number of | cos θ cygnus | distribution, N data i is the measurednumber of events and N exp i is the expected number of events. A nuisance parameter α
100 200 300 400Energy (keV) −
10 110 R a t e ( c oun t s / k e V / k g / da ys ) -PIC) µ RUN22 (LA -PIC) µ RUN14 (conventional = 20 pb χ σ WIMP at
100 200 300 400Energy (keV) −
10 110 C oun t s / k e V No cutAfter Fiducial cut After Length-Energy cut After TOT-Energy cut After Roundness cut (a) (b) (Standard μ-PIC)(LAμ-PIC)No cutAfter Fiducial cutAfter Length-Energy cutAfter TOT-Energy cutAfter Roundness cut
Fig. 7 (a) Energy spectra of the dark matter search at each selection step. The black,red, blue and green line are no-cut energy spectrum, the one after the Fiducial cut, Length-Energy cut and TOT-Energy cut, respectively. The black points with error bars are the finalevent sample after the Roundness cut. (b) Final energy spectrum considering the detectionefficiency. The black and gray points with error bars represent RUN22 (using the LA µ -PIC)and RUN14 (using the standard µ -PIC), respectively. Error bars indicate statistic poissonerrors. The blue dotted line shows the expected spectrum of the WIMP-nucleus scatteringwith the WIMP mass of 100 GeV/c , the WIMP-proton cross section of σ χ = 20 pb and theenergy resolution of 13.2%.
0° 90°30° 60°-30° -60° -90°0°-30°-60° 30° 60°
Fig. 8
Skymap of the final sample on the detector coordinate. The x axis is ( φ azi , θ ele ) =(0 ,
0) and the y axis is ( φ azi , θ ele ) = (90 , ξ/σ κ ) was introduced to consider the systematic uncertainty of the angular resolution σ κ . A possible angular-resolution shift is ξ . The expected event number N exp i was obtainedby a signal Monte Carlo simulation. It depends on the WIMP mass m χ , the WIMP-protoncross section σ χ − p and the astrophysical parameters. In addition, the nuclear quenchingfactor and the detector responses were considered. Here the nuclear quenching factor was Data =20 pb χ σ WIMP at N u m be r o f e v en t s |cosθ cygnus | Fig. 9
Measured and expected | cos θ cygnus | distribution for each energy range. The blackpoints with errors are measured data. The cyan lines show the WIMP expected signal of theWIMP-nucleus scattering with the WIMP mass of 100 GeV/c and the WIMP-proton crosssection of σ χ = 20 pb. Table 1
Systematic uncertainties of the expected rate for the WIMP mass m χ = 100GeV/c . Source cos θ range Relative uncertainty (%)Energy resolution [ 0, 1 ] < . < .
2[ 0.5, 1 ] < . − .
2[ 0.5, 1 ] +1.7 − . | cos θ cygnus | were binnedinto 2-bin.The measurement data was fitted by minimizing χ . The WIMP-proton cross section σ χ − p and the nuisance parameter α were treated as fitting parameters. The minimum χ value forthe 50-60 keV bin was 3.3 where σ χ − p and α were 18.5 pb and 0.12, respectively. Figure 10shows the measured | cos θ cygnus | distribution in the energy region of 50-60 keV along withthe expected one using best-fit values. In order to calculate the p-value, we made a χ distribution of an isotropic background (BG) model and an anisotropic WIMP model fromdummy samples. One thousand dummy samples were produced by Monte Carlo simulations nd χ value of each dummy sample was calculated. P-values for both of the WIMP and BGmodel were 3.3%. Hence we cannot claim the detection of WIMP dark matter with sufficientsignificance from the observed data. This is a natural result because of the large statisticerror and the small expected anisotropic ratio. Since no significant amplitude was found, a90% confidence level (C.L.) upper limit was set on the SD cross section. The 90% C.L. upperlimit on the SD cross section was obtained as 50 pb for 100 GeV WIMPs. Figure 11 shows the90% C.L. upper limit on the SD WIMP-proton cross section as a function of the WIMP mass.This result marked a new best sensitivity record of the SD WIMP search with the direction-sensitive method. This result improved the constraint by about 15 times compared to theprevious result of RUN14-18. This improvement owes to the surface background reductionof the µ -PIC detector. cygnus θ |cos0123456 N u m be r o f e v en t s DataBest fit (DM signal)90% C.L. upper value
Fig. 10
The | cos θ cygnus | distribution in the energy range of 50-60 keV. The black pointsshow observed data. The solid and dotted red lines are the simulated distribution usingbest-fit values and excluded values of 90% C.L., respectively.
5. Discussion
One of the milestones of the NEWAGE is to search an allowed region from DAMA exper-iment [29] with the directional method. More than two orders magnitude of improvementsis required in order to cover the entire DAMA region. Since the sensitivity is limited by theremaining backgrounds, we investigated the origin of the these backgrounds in RUN22 usingGeant4 [30] simulation.The main internal background candidates are alpha-rays coming from decay of radons andthe LA µ -PIC surface. The alpha-rays cannot be discriminated from nuclear recoil eventsaround 50 keV in NEWAGE-0.3b” detector by analysis cuts. Hence screening the mate-rial of the TPC and the LA µ -PIC is needed. A broad peak is observed in the energyspectrum around 6 MeV corresponding to the Rn and
Rn decay during the darkmatter search [31]. Estimated contamination of
Rn and
Rn are 5.7 ± WIMP mass (GeV/c -
10 110 ( pb ) - p cs S D W I M P - p r o t on c r o ss s e c t i on R U N ( T h i s w o r k ) RUN RUN - D M T P C ( F ) D R I F T ( F ) DAMA allowed
Fig. 11
90% C.L. upper limits on the SD WIMP-proton cross section as a function of theWIMP mass. The red thick solid line is the result of the directional method in this work.The blue solid lines of RUN14 and RUN14-18 are our previous results [17, 18]. The light blueand black solid lines show the results from the directional analysis of DMTPC [27] and theconventional analysis without the directional sensitivity of DRIFT [28], respectively. Allowedregion from DAMA/LIBRA experiment [29] is shown by the green area.and (5.3 ± × − mBq/m , respectively. In Geant4 simulation, we generated alpha-rays from decay chains of Rn and
Rn inside the TPC according to the branchingratio and estimated the number of events or rate with the analysis cut same to RUN22.The expected number of events (rate) due to
Rn and
Rn contamination in the lowenergy region of 50-60 keV were (6 . ± . × − (5.5 × − dru ) and (5 . ± . × − (4.7 × − dru), respectively. The remaining alpha-ray emission rate of the LA µ -PIC was(2.1 ± × − alpha/cm /hr [21] and the dominant component was 5.3 MeV alpha-rays from Po decay. Expected number of events (rate) in 50-60 keV region from thisbackground was less than 1 . × − (1.1 × − dru).Ambient gamma-rays and neutrons from rocks in the mine are two of the main componentsof the external background. Contributions of the cosmic-ray muons is negligible comparedwith ambient gamma-rays and neutrons. Measured ambient gamma-rays flux in the Labo-ratory B [32] was used for the estimation. The expected counts (rate) in the energy rangeof 50-60 keV was 1 . ± . × − dru). The ambient neutrons flux was measuredusing a He proportional counter and an energy spectrum of ambient neutrons produced by( α, n ) reactions and spontaneous fission was predicted [33]. The expected counts (rate) inthe energy range of 50-60 keV is (3 . ± . × − (3.1 × − dru). The expected numberof background events in the energy region of 50-60 keV is summarized in Table 2. The resultwithout roundness-cut is also shown in order to confirm the alpha-ray backgrounds fromthe LA µ -PIC surface, directly. The measured number of events are in good agreement withpredicted one within errors in both with and without roundness-cut. dru=counts/sec/keV/kg 12/14 he Rn backgrounds was reduced by the gas circulation system with cooled charcoal andtheir contribution was found to be negligible. Ambient gamma-rays and ambient neutronscontribute some part of the backgrounds. These backgrounds can be reduced by the externalshields comprising of materials like lead (Pb) and water (H O). On the other hand, internalbackgrounds of
Rn and the LA µ -PIC surface would remain with the external shields andthey would be dominant backgrounds. A straightforward way of reducing these backgroundis to replace the detector components with radiopure materials. A µ -PIC with a furtherbackground reduction is being developed. Another approach to reduce the background isto detect the absolute z position of the events. Majority of the remaining backgrounds areknown to locate at low z (around the LA µ -PIC and the GEM) and high z position (aroundthe drift plate). A discovery of minority carriers in CS + O gas mixtures by the DRIFTgroup opened the potential of an absolute z measurement in self-triggering TPCs [28]. Wehave recently demonstrated a three-dimensional tracking with a spatial resolution of 130 µ musing the µ -PIC in a negative ion gas SF . Simultaneously, the absolute z coordinates wasdetermined with a location accuracy of 16 mm [34]. The negative ion gas TPC enables us toreduce these backgrounds using the z fiducialization, effectively. Thus the use of the negativeion gas TPC is another promising approach to reduce the surface alpha-ray backgrounds.It should be noted that the surface background will be one of the ultimate backgroundsources even after significant efforts of material selection since radioactive isotopes can beembedded by the decays of radons in a normal atmosphere even after the production. Thus itis important to take both possible ways to reduce the background so as to start investigatingthe DAMA region and further searches. Table 2
Summary of the expected numbers of background events and measured numbersin the energy region of 50-60 keV.Source w/ roundness w/o roundnessAmbient gamma-rays 1 . ± . . ± . . ± . × − (4 . ± . × − Rn (5 . ± . × − (8 . ± . × − Rn (6 . ± . × − . ± . µ -PIC surface < . × − . ± . . ± . ± . . ± . ± .
6. Conclusion
We developed a low background µ TPC detector, NEWAGE-0.3b”, for the directional darkmatter search with an LA µ -PIC which is a two-dimensional tracking gaseous detector madeof low radioactive materials. A directional dark matter search in the Kamioka Observatorywas carried out in 2018. Total exposure was 1.1 kg · days and number of the observed eventsin the energy region of 50-60 keV was two which was consistent with 2.5 events of theexpected background. No significant forward-backward asymmetry of a WIMP signal wasfound, therefore we derived a 90% confidence level upper limits on the SD WIMP-proton ross section of 50 pb for 100 GeV/c WIMPs. We improved the constraint of the previousresult [17] by a factor of 15 and marked the best direction-sensitive limit.
Acknowledgment
This work was partly supported by JSPS (Japan Society for the promotion of Science) KAK-ENHI (Grant-in-Aids for Scientific Research) (grant nos. 16H02189, 19684005, 23684014 and26104005, 19H05806), JSPS Bilateral Collaborations (Joint Research Projects and Semi-nars) program, ICRR Joint-Usage, Program for Advancing Strategic International Networksto Accelerate the Circulation of Talented Researches (R2607), and JSPS Research Fellow (grant nos. 17J03537).
References [1] E. Aprile et al., Phys. Rev. Lett., , 111302 (2018).[2] D. S. Akerib et al., Phys. Rev. Lett., , 021303 (2017).[3] R. Agnese et al., Phys. Rev. Lett., , 051301 (2018).[4] P. Agnes et al., Phys. Rev. Lett., , 081307 (2018).[5] R. Ajaj et al., Phys. Rev. D, , 022004 (2019).[6] E. Armengaud et al., Journal of Cosmology and Astroparticle Physics, , 019 (2016).[7] X. Cui et al., Phys. Rev. Lett., , 181302 (2017).[8] K. Abe et al., Physics Letters B, , 272–276 (2016).[9] S. Baum et al., Physics Letters B, , 262–269 (2019).[10] D. Spergel, Phys. Rev. D, , 1353–1355 (1988).[11] C. A. J. O’Hare et al., Phys. Rev. D, , 063518 (2015).[12] J. Billard et al., Phys. Rev. D, , 075002 (2011).[13] S.K. Lee and A.H.G Peter, Journal of Cosmology and Astroparticle Physics, , 029–029 (2012).[14] K. I. Nagao et al., Physics of the Dark Universe, , 100426 (2020).[15] K. Miuchi et al., IEEE Transactions on Nuclear Science, , 825–830 (2003).[16] A Takada et al., Nuclear Instruments and Methods in Physics Research Section A: Accelerators,Spectrometers, Detectors and Associated Equipment, (1), 195–199 (2007).[17] K. Nakamura et al., Progress of Theoretical and Experimental Physics, (2015).[18] R. Yakabe et al., Progress of Theoretical and Experimental Physics, ptaa147 (2020).[19] T. Hashimoto et al., AIP Conference Proceedings, (1), 070001 (2018).[20] F. Sauli, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers,Detectors and Associated Equipment, , 2–24 (2016).[21] T. Hashimoto et al., Nuclear Instruments and Methods in Physics Research Section A: Accelerators,Spectrometers, Detectors and Associated Equipment, , 164285 (2020).[22] T. Tamagawa et al., Nuclear Instruments and Methods in Physics Research Section A: Accelerators,Spectrometers, Detectors and Associated Equipment, (3), 390–396 (2009).[23] R. Orito et al., IEEE Transactions on Nuclear Science, , 1337–1342 (2004).[24] K. Nakamura et al., Journal of Instrumentation, (02), C02023 (2012).[25] J. Conrad and otehrs, Phys. Rev. D, , 012002 (2003).[26] J. F. Ziegler et al., Nuclear Instruments and Methods in Physics Research Section B: Beam Interactionswith Materials and Atoms, (11), 1818–1823 (2010).[27] S. Ahlen et al., Physics Letters B, (1), 124–129 (2011).[28] J.B.R. Battat et al., Physics of the Dark Universe, , 1–7 (2015).[29] C. Savage et al., Phys. Rev. D, , 123513 (2004).[30] S. Agostinelli et al., Nuclear Instruments and Methods in Physics Research Section A: Accelerators,Spectrometers, Detectors and Associated Equipment, (3), 250–303 (2003).[31] T. Ikeda, Doctor Thesis, Kobe University (2020).[32] H. Nishimura, Doctor Thesis, Kyoto University (2008).[33] K. Mizukoshi et al., Progress of Theoretical and Experimental Physics, , 123C01 (2018).[34] T. Ikeda et al., Journal of Instrumentation, (07), P07015 (2020).(07), P07015 (2020).