Characterization of the sub-kev Germanium detector

Transcription

1 Indian J Phys (March 2018) 92(3): ORIGINAL PAPER Characterization of the sub-kev Germanium detector M K Singh 1,2 *, M K Singh 1,2, V Sharma 1,2, L Singh 1,2, V Singh 1 *, V S Subrahmanyam 2, A K Soma 1,2, G Kiran Kumar 1 and H T Wong 1 1 Institute of Physics, Academia Sinica, Taipei 11529, Taiwan 2 Department of Physics, Institute of Science, Banaras Hindu University, Varanasi , India Received: 22 November 2016 / Accepted: 26 July 2017 / Published online: 24 October 2017 Abstract: Germanium ionization detectors having sensitivities as low as 100 ev open new windows for the studies of neutrino and dark matter physics. This novel detector demands overcoming several challenges at both hardware and software levels. The amplitude of physics signals is comparable to those due to fluctuations of the pedestal electronic noise in low energy range. Therefore, it is important to study the low energy calibration properly. In this article, we focus on the optimization of the calibration scheme. Keywords: Ionization ppcge and npcge detector; Low energy calibration; Dark matter identification; Neutrino interactions PACS Nos.: Wk; ?d; Vj; sb 1. Introduction and physics motivation The Taiwan EXperiment On NeutrinO (TEXONO) Collaboration research program theme is based on the studies of low energy neutrino physics such as neutrino nucleuscoherent scattering (NNCS), neutrino millicharge, charge radius, neutrino magnetic moments etc. and dark matter search at Kuo-Sheng Reactor Neutrino Laboratory (KSNL), Taiwan [1, 2]. Germanium detectors with sub-kev sensitivities have been demonstrated as efficient means to probe Weakly Interacting Massive Particles (WIMPs) [1 11]. This novel detector technique is also adopted in the study of neutrinonucleus coherent scattering with reactor neutrinos [1, 2, 6 9]. This motivates the development of point-contact germanium detectors. The experimental signatures are the nuclear recoils, posing the challenging requirements of low background and low threshold to the detectors. With the above theme, we explore high-purity germanium (HPGe) detector technology to develop a sub-kev threshold detector for pursuing studies on low energy neutrino and dark matter physics. The TEXONO collaboration has developed *Corresponding author, singhmanoj59@gmail.com; venkaz@yahoo.com and used several HPGe detectors [1, 2]. The generic benchmark goals in terms of detector performance are: (I) modular target mass of the order of 1 kg; (II) detector sensitivities reaching the range of 100 ev; and (III) background in the range of 1 count kg -1 kev -1 day -1 (cpkkd). 2. Experimental details The Kuo-Sheng Reactor Neutrino Laboratory is located at a distance of 28 m from the core No. 1 and around 102 m from the second nuclear reactor of the Kuo-Sheng Nuclear Power Station operated by the Taiwan Power Company on the northern shore of Taiwan. A multi-purpose inner target detector space of dimension 100 cm 9 80 cm 9 75 cm is enclosed by 4p passive shielding material, which has a total weight of about 50 ton. The nominal overburden above the Kuo-Sheng Reactor Neutrino Laboratory is around 30 meter-water-equivalent (mwe). The shielding provides attenuation to the ambient neutron and gamma background, and consists of, from inside out, 5 cm of Oxygen Free High Conductivity (OFHC) copper, 25 cm of boron-loaded polyethylene, 5 cm of stainless steel to support the structure, 15 cm of lead to attenuate the gamma rays, and plastic scintillator panels to veto the cosmic rays [1, 2]. The schematic layout of the shielding structure is Ó 2017 IACS

2 402 M K Singh et al. shown in Fig. 1(a). In the inner target space, various detectors can be placed for the different scientific programs, simultaneously, on a movable platform. At the experimental site, the Germanium detector (a target detector) was enclosed by an NaI(Tl) crystal scintillator detector as an anti-compton (AC) detector and copper passive shielding inside a plastic bag purged by nitrogen gas evaporated from a liquid nitrogen dewar. This set up shown in Fig. 1(b), was housed inside the passive shielding space as shown in Fig. 1(a). Cosmic ray (CR) veto panels made of plastic scintillators read out by photomultiplier tubes (PMTs) [1 3] surrounded this whole structure. The NaI(Tl) detector named as wall detector provides three side coverage to the target detector. Fig. 1 (a) Schematic layout of the general-purpose inner target space, passive shielding and cosmic ray veto panels. (b) Schematic diagram of the experimental set up, which includes the Germanium detector and the NaI(Tl) scintillator detector 2.1. Germanium detector Working principle of semiconductor detector The Semiconductors consist of a periodic atom lattice, with the electrons in the crystal forming energy bands. In valance band, the electrons are in a bound state and the band lying above the valence band is called the conduction band. Here, in this band, the electrons are free charge carriers. The band gap between the valence and the conduction band is approximately 1 ev, for the semiconductor detectors. When the electric field is applied, the electrons in the conduction band drift towards the cathode and the resulting current can be measured. The current depends directly on the number of electrons in the conduction band. The electrons can move from the valence band to the conduction band by thermal excitation, and the current due to this process is called leakage current [12]. Electrons that jump from the valence to the conduction band leave behind a hole. The hole travels in the opposite direction of the electron due to the electric field. The leakage current can produce a noise signal, which is comparable to the physics signal in case of rare physics events. Hence it is important to keep the leakage current small as much as possible [3, 12]. The leakage current can be suppressed by cooling the crystal. In most experiments liquid nitrogen is used to control the leakage current by lowering the crystal temperature for the better performance of the detectors. In semiconductor detectors, ionizing radiation is measured by the number of charge carriers sets free in the detector material, which is placed between two electrodes. Under a reverse bias, an electric field extends across the intrinsic or depleted region. Ionizing radiation produces free electrons and holes. The high purity germanium has a net impurity level of around atoms/cm 3 so that with a moderate reverse bias, the entire volume between the electrodes is depleted, and an electric field extends across this active region. The number of electron hole pairs produced is proportional to the energy of the radiation. As a result, a number of electrons are transferred from the valence band to the conduction band, and an equal number of holes are created in the valence band. Under the influence of an electric field, electrons and holes travel to the electrodes, where they result in a pulse that can be measured in an external circuit. As the amount of energy required to create an electron hole pair is known, and is independent of the energy of the incident radiation, measuring the number of electron hole pairs allows the energy of the incident radiation to be determined.

3 Characterization of the sub-kev Germanium detector Sub-keV Germanium detector The Germanium detector is also called high purity germanium detector (HPGe) or hyper-pure germanium detector. There are several configurations of Germanium detectors with variable mass, which are available commercially such as Standard Electrode Coaxial Germanium Detector (SEGe), Broad Energy Germanium Detector (BEGe), Reverse Electrode Coaxial Germanium Detector (REGe), Ultra Low Energy Germanium Detector, Point Contact Germanium Detector (PCGe) etc. The capacitance of SEGe detector is typically 20pF, whereas the PCGe detector has capacitance less than a pf, which is the major advantage in the Germanium detector technology [1, 2]. The lower capacitance of the PCGe detector provides several benefits over commonly used HPGe detector such as: (I) the noise is reduced; (II) the energy resolution is improved because of the reduction in noise; (III) the timing resolution is improved, which separates single, and double sited events in a crystal; (IV) the energy threshold is improved. Before advanced purification techniques were developed, germanium crystals could not be produced with highpurity sufficient to enable their use as spectroscopy detectors. Impurities in the crystals trap electrons and holes, ruining the performance of the detectors. Consequently, germanium crystals were doped with lithium ions (Ge (Li)), in order to produce an intrinsic region in which the electrons and holes would be able to reach the contacts and produce a signal. The HPGe detectors commonly use lithium diffusion technique to make an n? Ohmic contact, and boron implantation technique to make a p? contact [1 3]. Coaxial detectors with a central n? contact are referred to as n-type detectors, while p-type detectors have a p? central contact. The thickness of these contacts represents a dead layer around the surface of the crystal within which energy depositions do not result in detector signals [1 3]. The central contact in these detectors is opposite in polarity to the surface contact, making the dead layer thickness in n-type detectors smaller than that in p-type detectors [1 3]. Typical dead layer thicknesses are around 0.5 mm for a Li diffusion layer, and around 0.3 lm for a boron (B) implantation layer. TEXONO collaboration has studied various customs designed and commercially made germanium detectors [1, 2]. These germanium detectors are used as ionization detectors. The outer surface electrode of the p-type point contact Germanium detector is fabricated by lithium diffusion, and has a finite thickness. Electron hole pairs produced by the radiations at the surface layer are subjected to a weaker drift field than those in the crystal bulk region. The Surface events have only a partial charge collection and slower rise-time [1 3]; while for the n-type point contact germanium detector, the outer surface is boron-implanted electrode of submicron thickness, due to which n-type detectors show no anomalous surface effect [1 3] Data readout and analysis Data readout A schematic diagram of the readout scheme for the Germanium detectors is shown in Fig. 2. Signals from germanium crystal sensors are first amplified by the front-end JFETs located near the germanium diodes. Outputs are fed to the reset preamplifiers, which are placed * 30 cm away from the detector. The typical output of the preamplifier, as observed in an oscilloscope, is displayed in Fig. 3. The saw-tooth characteristics exemplify the timing structure of the RESETs which are issued after a fixed time-period or when the charge deposition in the detector exceeds a preset value (for example, with a direct cosmic ray event). The steps in between the RESETs represent physics signals with heights proportional to their energy [1, 2]. The preamplifier signal output is distributed to a fast timing amplifier, which keeps the rise-time information. The two amplifiers, one of them have 6 ls shaping time, which provides the energy information, and the other one with a 12 ls shaping time, which is used to reduce the microphonic noise events. The discriminator output provides the trigger instant for data acquisition (DAQ). 200 MHz and 60 MHz Flash Analog-to-Digital Converters (FADC), respectively digitize pulses from timing amplifier and shaping amplifier Event selection The shaping amplifier output pulses are digitized in timing range 26 ls before and 70 ls after the trigger. The resolution of FADC is 1/60 ls, which means 1 ls has 60 data points. Figure 4 shows a typical waveform with some important parameters. The definition of some important waveform parameters are described below: (I) The minimum point (min pt ) and the corresponding time bin (min tbin ) of the pulse are searched in the entire pulse ranging from 0 to 96 ls. (II) The maximum point (max pt ) and the corresponding time bin (max tbin ) of the pulse are searched from 10 to 60 ls window. (III) Amplitude (A max ) is the maximum height of the pulse after subtraction of the pedestal. (IV) Pedestal (ped) and pedestal tail (ped tail ) are calculated for each event from each channel by the average of amplitude distribution from 0 to 10 ls and 70 to 96 ls, respectively:

4 404 M K Singh et al. Fig. 2 Schematic diagram of the DAQ system for the Germanium detector analysis is performed interactively in root framework [13]. Fig. 3 Schematic drawing of the raw preamplifier signals as recorded with oscilloscope (V) ped ¼ 1 X 10 VðtÞ; ped tail ¼ 1 X 96 VðtÞ: ð1þ t¼0 t¼70 Total area under the pulse is termed as Q. It is obtained by the integration of the whole pulse (from 0to96ls) defined as: Q ¼ X96 t¼0 VðtÞ: ð2þ (VI) The partial charge under the pulse is termed as Q 0 and is the integral of the pulse in the time window from 15 to 55 ls and defined as: Q 0 ¼ X55 t¼15 VðtÞ: ð3þ In this way, all the required information for the data analysis is stored into reduced form in a root ntuple file. We are also able to add new parameters in root ntuple file as per physics analysis requirements, from time to time. All As we know that, the neutrino and dark matter candidates are weakly interacting with the detector material. Therefore, the nuclear recoil of v N and t N events are uncorrelated with the rest of the surrounding detector components and are uniformly distributed into the point contact germanium detector volume. The background events that may be physics events must be correlated with any other detectors. It means we are looking for events having only single sited interaction and uniform distribution into the germanium detector volume. In the following, the superscript (-)/(?) denotes uncorrelated/(correlated) of cosmic ray (CR) and anti-compton veto (AC) systems with germanium detector signals. The AC? CR - represent an event having hits in anti-compton veto system and in germanium detector but no hit in cosmic ray veto system. It means that, this event belongs to the ambient gamma ray. The AC - CR? represent an event having hits in cosmic ray veto system and in germanium detector but no hit in anti-compton veto system. It means that, this event belongs to the cosmic ray induced high-energy neutron event. The AC? CR? represent events having hits in cosmic ray veto and in anti-compton veto systems along with the germanium detector, are physics events. The AC - CR - having signal only in the target germanium detector designates the desired signal, which is used for further physics study. The preamplifier RESET signal can induce noise events with definite timing structure. The timing difference

5 Characterization of the sub-kev Germanium detector 405 Fig. 5 Events timing with respect to the reset signal, where Dt? and Dt - are correspondingly the next and previous reset signal Fig. 4 (a) A typical wave form showing various waveform parameters. (b) The different integration region defines the Q and Q 0 between an event and its previous RESET is denoted by Dt - and to next RESET is denoted by Dt?. The timing distribution between Dt - and Dt? is depicted in Fig. 5. The events with constant time interval are eliminated. The corresponding signal efficiency achieved is more than 97%, given by the survival fraction of AC? CR? physics events when it is subjected to identical criteria [1, 2]. We apply several offset cuts to reduce the noise events, which are described stepwise in Ref. [2]. 3. Results and discussion 3.1. Energy calibration and spectrum A number of internal peaks arising from cosmogenic isotopes decaying via electron capture (EC) provide excellent candidates for calibration along with random trigger events, which provide zero energy definition. There are Fig. 6 Germanium detector calibration with known energy peaks as mentioned in Table 1 with random trigger events providing zero energy definition several X-ray lines below 12 kev, which help to calibrate the events occurring in point contact germanium detector. The calibration point is shown in Fig. 6. Table 1 summarizes some details about the K and L-shell X-ray lines observed in point contact germanium detector, which has originated from cosmic ray, induced radionuclide decays and the decays by electron capture. The X-ray line at kev arises due to the Germanium activation by both thermal and fast cosmogenic neutrons. In the thermal neutron case, the following reaction is possible: 70 Ge þ n! 71 Ge! EC 71 Ga þ cð10:37 kevþ; s1 2 ¼ 11:4 days: ð4þ In the fast neutron case, the possible reactions are as follow:

6 406 M K Singh et al. Table 1 The energies and lifetimes of K-shell and L-shell X-rays for different atoms Parent radionuclide Daughter nuclide Daughter K-shell energy (kev) Daughter L-shell energy (kev) Parent nuclide halflife 73,74 As Ge , 77.8 days 68,71 Ge Ga , 271 days 68 Ga Zn min 65 Zn Cu days 56 Ni Co days 56,57,58 Co Fe , 271.7, 70.9 days 55 Fe Mn years 51 Cr V days 49 V Ti days 70 Ge þ n! 68 Ge þ 3n; ð5þ 68 Ge! EC 68 Ga þ cð10:37 kevþ; 68 Ga! EC 68 Zn þ cð9:66 kevþ; s1 ¼ 271 days; ð6þ 2 s1 ¼ 67:7 mts: ð7þ 2 While the X-ray line at kev arises from fast neutron of cosmic origin. The production and decay of zinc isotopes are present through the following reactions: 70 Ge þ n! 65 Zn þ 2n þ 4 2 He; ð8þ 65 Zn! EC 65 Cu þ cð8:98 kevþ; s 1=2 ¼ 244 days: ð9þ The A max,q,q 0 and the combination of the A max and Q 0, provides four different parameters for the energy calibration. There are no physics related structures observed below * 1 kev, where energy calibrations are performed by using a test pulser. Test pulser events are produced by a precision pulse generator fed to the preamplifier. Pulser signals probe the response of electronics system independent of the electronics drift effects for physics signals. Pulser and physics events have identical profiles as depicted in Fig. 7 [1, 2]. Data were taken with decreasing pulser amplitude. The pulser measurements at zero amplitude are equivalent to random trigger events. The energy scale for the pulser data is defined by matching with the c-peaks at high energy. After performing various calibrations with the desired parameters, we have measured the resolution (sigma value) of the kev peaks as well as the noise edge as listed in Table 2. Table 2 reveals that the A max parameter gives the better resolution value, as well as lower noise edge compared to the other parameters. Therefore, we used this optimized calibration parameter to produce the energy spectrum for further study. The intrinsic pulser measurement response is linear in the range corresponding to the physics events [1, 2]. The pulser measurements for p-type, point contact germanium Fig. 7 With the same settings of the shaping amplifier, physics events (black line) and a test pulser event (red line) (color figure online) detector (ppcge) and n-type point contact germanium (npcge) detector are displayed in Fig. 8. Polynomial functions provide calibration of A max into energy unit over the entire range. The response is linear in the physics region of interest above the electronic noise-edge Results TEXONO collaboration has used Standard Electrode Coaxial Germanium Detector (SEGe) of more than 1 kg mass many years ago, and then switched to Point Contact Germanium Detector (PCGe) [3]. All the detectors used by TEXONO are procured commercially made as per custom design by the TEXONO collaboration according to their requirements [1]. This study is based on p-type pointcontact Ge detector (ppcge) and n-type point-contact Ge detector (npcge) with 500 g mass each. After applying all the optimized basic cuts [1], and energy calibration by using optimized calibration parameter (A max ), the final energy spectrum of the AC - CR - sample events for both p-type, and n-type point contact germanium detectors are shown in Fig. 9.

7 Characterization of the sub-kev Germanium detector 407 Table 2 The sigma value of kev peak and noise edge for various energy calibration parameters Calibration parameter n-type point contact Germanium detector (npcge) p-type point contact germanium detector (ppcge) kev Peak resolution r (ev) Electronic noise edge (ev) kev Peak resolution r (ev) Electronic noise edge (ev) Total area (Q) ± ± Partial area (Q 0 ) ± ± Amplitude (A max ) ± ± Amplitude (A max )? partial area (Q 0 ) ± ± Fig. 8 Response of p-type, point contact germanium detector and n-type, point contact germanium detector versus energy when the test pulser amplitude is comparable to pedestal noise fluctuations energy estimator A max The electronic noise edge for p-type, point contact germanium (ppcge) detector is 300 ev and for n-type, point contact germanium (npcge) detector is 350 ev are obtained by using the optimized calibration parameter as shown in Table 2 [1, 2]. 4. Conclusions In this article, we focus on the various parameters for the energy calibration of the data taken at KSNL with germanium detectors. From Table 2 it is clear that the energy spectrum obtained by using maximum amplitude parameter for energy calibration, shows improved energy resolution as well as better with respect to electronic noise as compared to the other three parameters. From this study, it is clear that the maximum amplitude is the best-optimized parameter for the energy calibration. In this study, we also compare the resolution of the n-type with p-type, point contact germanium detector. As we can see from Table 2 that p-type, point contact germanium detector has better energy resolution and electronic noise compared to the n-type, point contact Fig. 9 Energy spectrum for (a) p-type, point contact germanium detector, and (b) n-type point contact germanium detector right side using optimized calibration parameter A max germanium detector. Further understanding of the sub-kev background and optimization of the energy resolution of n-type, point contact germanium detector to be comparable to the p-type, point contact germanium detector constitute our ongoing research program. Acknowledgements The research programs and results presented in this article are from the efforts of the TEXONO Collaboration, consisting of groups from Taiwan (Academia Sinica, KuoSheng Nuclear Power Station), China (Tsinghua University, Institute of Atomic Energy, Nankai University, Sichuan University), India (Banaras

8 408 M K Singh et al. Hindu University) and Turkey (Middle East Technical University, Dokuz Eyll University) and being a founding partner of the CDEX Collaboration. M. K. Singh thanks the University Grants Commission (UGC), Govt. of India, for the funding through UGC D. S. Kothari Post-Doctoral Fellowship (DSKPDF) scheme. The authors are grateful to the contributions from all the collaborators. References [1] A K Soma et al. Nucl. Instru. Meth. A (2016) and references therein [2] M K Singh et al. Indian J. Phys. 91(10), (2017). DOI /s [3] H B Li et al. Astropart. Phys (2014) and references therein [4] H T Wong et al. Phys. Rev. D (2007) [5] H B Li et al. Phys. Rev. Lett (2013) [6] J W Chen et al. Phys. Lett. B (2014) [7] J W Chen et al. Phys. Rev. D (R) (2014) [8] H T Wong, THE UNIVERSE 3(4) 22 (2015) [9] S Kerman et al. Phys. Rev. D (2016) [10] H T Wong et al. J. Phys. Conf. Ser (2006) [11] H T Wong Int. J. Mod. Phys. D (2011) [12] Q Looker et al. Nucl. Instru. Meth. A (2015) [13]