Dissertations

The SuperCDMS collaboration has developed detectors to probe potential interactions between Dark Matter (DM) particles and semiconductor crystals at cryogenic temperatures. During the final phase of the Cryogenic Dark Matter Search low ionization threshold experiment (CDMSlite), vibrations originating from the cooling system were observed. These vibrations generated broadband low-frequency noise (LFN), which frequently triggered the detector. These LFN events were challenging to remove due to their variable pulse shapes, which often resembled low-energy signals produced by particle interactions. In the final analysis, a strict event selection criterion was applied to the dataset to remove the LFN events. However, this increased the energy threshold and lowered the signal efficiency, reducing the experiment's sensitivity to low-mass DM interactions. 
In this work, a new LFN selection criterion using machine learning is presented. This approach aims to improve the signal efficiency while maintaining the signal purity. Two neural network architectures are considered: a Convolutional Neural Network (CNN) and a CNN combined with Long Short-Term Memory (LSTM). The CDMSlite detector has four phonon signal channels; therefore, the LFN identification is probed using individual channels, the sum of the phonon channels, and multi-channel configurations. The LFN-background samples for training are data-driven. A novel method for signal sample generation using pulse shapes and regression is developed. Both this method and standard phonon signal templates are used for generating signal samples. Among the tested channel configurations and neural network architectures, the multi-channel configuration with a CNN architecture performs the best in terms of signal efficiency and purity. This network is used on the CDMSlite Run 3 data to remove LFN events. 
The final spectrum reveals additional events compared to the spectrum obtained with the previous LFN selection criterion. Exclusion limits on the spin-independent DM-nucleon cross-section are calculated for DM masses between 1 GeV/c² and 10 GeV/c². The new results are consistent with the previous analysis, with an improvement in sensitivity up to 17.5% in the DM mass range of 2.4 GeV/c² to 5 GeV/c². The gain in signal efficiency due to the ML-based LFN cut is limited by the application of a fiducial volume selection criterion. Therefore, case studies are conducted excluding the fiducial volume selection criterion to investigate its impact, which demonstrates the potential of the ML-based LFN cut in the low-mass DM parameter space. The sensitivities are extended down to DM masses of 0.8 GeV/c², with improvements in sensitivity of up to three orders of magnitude below 1.2 GeV/c².

 

SuperCDMS SNOLAB is a planned dark matter direct detection experiment that aims to probe dark matter with masses below 10 GeV. It will use semiconducting silicon and germanium detectors cooled to ∼15 mk by a commercial dilution refrigerator to achieve thermal noise reduction. Nonetheless, the experiment is expected to observe a non-zero background rate originating from decay daughters of radionuclides that contaminate the instrumentation. This report quantifies the increasing background rate caused by cosmogenic activation in the experiment's copper instrumentation that is stationed above ground for an extended period of time. It is found that an appreciable background rate reduction can be achieved by either re-fabricating the experiment's cryostat with new, low-exposure copper (which poses several logistical issues), or by allowing a 3-year cooldown period for these pieces. Since neither of these options are practical or cost-effective, it is recommended that the experiment allow cosmogenic activation to account for up to 15% of its background budget. 
This report secondarily confirms an issue with Geant4's production cuts mechanism. Secondary particles that are within the safety of a volume boundary are not expected to be removed from the stack by production cuts, but the contrary has been observed for electromagnetically neutral secondaries. Using TestEM5 - an example application that ships with Geant4 - it is confirmed that safety is calculated correctly for all secondaries and that neutral secondaries are not cut away. The ApplyCuts flag is investigated given its ambiguous nature in the program's documentation. A solution to this issue for neutral secondaries is not found, but it is emphasized that this issue is of moderate importance to the SuperCDMS collaboration's simulations efforts. Further investigation of each standard electromagnetic physics list is warranted.

 

The nature of dark matter remains unknown, with no experimental confirmation of existing theories. The upcoming SuperCDMS SNOLAB experiment seeks to probe dark matter particle models with masses below 10 GeV/c². To achieve this, the SuperCDMS collaboration has developed cryogenic sampling calorimeters with energy thresholds down to the sub-eV energy range. 
The response of the SuperCDMS detectors is not fully characterized theoretically, requiring empirical calibration techniques down to their energy thresholds. A key challenge for an empirical calibration technique is to achieve a uniform distribution of energy depositions throughout the detector target, as expected from dark matter particles. 
This thesis presents the development of a novel calibration technique for SuperCDMS silicon detectors utilizing energy signatures from the Compton scattering energy depositions. The simulations supporting this calibration method are discussed. Additionally, the inference model to identify the calibration signatures in the data, and the procedure to construct the calibration function is described. This thesis further demonstrates a successful calibration of experimental data from four SuperCDMS R&D detectors using this approach. Finally, a comparison is made between the new method and an alternative calibration technique using optical photons (2 eV).

 

Astrophysical observations suggest that the universe contains a substantial amount of dark matter, likely composed of particles beyond the Standard Model. The SuperCDMS (Cryogenic Dark Matter Search) SNOLAB experiment is a next-generation direct detection effort using cryogenic germanium and silicon detectors to measure phonon and ionization signals from dark-matter recoils. It aims to improve sensitivity to dark-matter particles with masses below 10 GeV by an order of magnitude. As part of this effort, HVeV detectors (high voltage eV scale), which are gram scale with single-charge sensitivity, serve as a prototype to provide insight into detector response, calibration methods and background sources. Three prior HVeV runs have yielded competitive constraints on low-mass dark matter. 
This dissertation analyzes data from the HVeV Run 4 experiment, conducted in an underground laboratory at Northwestern University. The experiment benefits significantly from the identification and elimination of the luminescence from the printed circuit boards in the detector holder used in Run 3, resulting in a lower background event rate and stronger constraints. Data from 10.80 gram-days of exposure were analyzed in a blinded study using a likelihood-based method. Limits are set on the dark matter-electron scattering cross-section in the mass range of MeV to GeV, dark photon absorbption mixing parameter and axion-like particle coupling constant in the mass range of eV to tens of eV. The results turn out to be competitive and world-leading in some of the lower mass ranges. The experiment also provides information on potential background sources in the low-energy range, where future HVeV runs are expected to reduce or model them.

 

The existence of dark matter has been inferred through many pieces of astrophysical evidence. However, much about its nature is unknown to this day. The several decades-long search for dark matter has given rise to many experiments and even more dark matter candidates. SuperCDMS is a direct detection experiment that uses cryogenic detectors to probe interactions of dark matter particles with Standard Model particles. Work done towards advancing two frontiers of the SuperCDMS SNOLAB experiment will be the main points of discussion in this thesis - (a) the characterization of the new SuperCDMS SNOLAB detectors, and, (b) the development of novel event reconstruction techniques for the experiment. 
After introducing the motivation for dark matter and the current experimental techniques used to detect it, the thesis will introduce SuperCDMS detection principles and pre-requisite knowledge to motivate and understand most of the work described within. First, preliminary results will be presented from testing and characterizing SuperCDMS detector towers at the Cryogenic Underground TEst facility, a low background test facility at SNOLAB in Sudbury, Canada. The first aspect will present a detailed discussion of advanced reconstruction algorithms to fit data sampled at non-uniform speeds to keep within the bandwidths of the readout electronics and maintain low trigger threshold at SuperCDMS SNOLAB. The second major development discussed will be a novel reconstruction technique called the N×M filter, which fits N channels with M shapes/templates simultaneously and develops a pipeline that integrates machine learning to achieve excellent resolution improvement. The key outcomes of this thesis are (a) capability demonstration of the SuperCDMS SNOLAB detectors, (b) development of a less memory-intensive algorithm to process non-uniformly sampled data, and, (c) demonstration of a two-fold improvement and nearly a four-fold improvement in energy resolution in old and new data sets using the N×M filter, respectively.

 

Rare event search experiments like Dark Matter searches, Coherent Elastic Neutrino-Nucleus Scattering (CEνNS), and Neutrino-less Double Beta Decay (NDBD) are exciting and challenging as they address unanswered physics questions and push de boundaries of current technology. Dark matter, constituting 26.8% of the Universe's mass-energy, remains an enigma. In direct detection experiments, Dark Matter particles scatter off target nuclei, producing signals such as light, phonons, charge, or a combination of them, depending on the detector material. The absence of GeV-scale Dark Matter signals compels the exploration of low-mass Dark Matter, requiring detectors with low-energy thresholds. CEνNS, which scatters neutrinos coherently with nuclei, faces similar challenges. Another challenge is the mitigation of high background rates from radiogenic and cosmogenic sources, necessitating underground laboratory with passive shielding, and active shielding for in-situ background reduction. 
In this thesis, we discuss the development and performance of an annular, cryogenic phonon-mediated active veto detector designed to significantly reduce radiogenic backgrounds in rare event search experiments. The detector consists of a germanium veto detector weighing approximately 500 g, with an outer diameter of 76 mm and an inner diameter of 28 mm. The veto detector can host a 25 mm diameter germanium inner target detector weighing around 10 g. Using inputs from a GEANT4 based simulation, the detector was optimized to be positioned between two germanium detectors, resulting in a >90% reduction in background rates dominated by gamma interactions. Operating at mK temperatures in the experimental setup, the prototype veto detector achieved a baseline resolution of 1.24 ± 0.02 keV, while maintaining a functional inner target detector with a baseline resolution of 147 ± 2 eV. Experimental results closely matched simulation predictions, affirming the efficiency of the design for aggressive background reduction necessary for neutrino and dark matter search experiments. 
This thesis also presents experimental results from a ∼ 100 g single-crystal sapphire detector, with a diameter of 76 mm and a thickness of 4 mm, equipped with transition edge sensors (TES). Sapphire, composed of aluminum oxide, emerges as a promising candidate for light dark matter search experiments due to its lower atomic mass compared to materials like germanium and silicon. This novel phonon-assisted sapphire detector exhibits a baseline recoild energy resolution of 28.4 ± 0.4 eV. We combine two low-threshold detector technologies, sapphire and ∼ 100 g Si High Voltage (HV), to develop a large-mass, low-threshold detector system. It simultaneously measures athermal phonons in a sapphire detector while an adjacent Si HV detector detects scintillation light from the sapphire detector utilizing NTL amplification. This setup allows for event-by-event discrimination between electron and nuclear events due to differences in their scintillation light yield. While previous systems with simultaneous phonon and light detection have employed smaller detectors, this system is designed to provide a large detector mass with high amplification for the limited scintillation light. 
This thesis also discusses an ongoing study to precisely measure the decay rates of ³²Si and ³²P using data from Cryogenic Dark Matter Search (CDMS) experiment, collected between 2003 and 2012 at the Soudan Underground Laboratory. The experiment employed 19 Ge and 11 Si cryogenic detectors in a five-tower configuration to detect recoil energy from particle interactions, measuring both phonon and charge energy. ³²Si, a naturally occurring isotope in Si detector material, decays to ³²O, which further decays to stable ³²S, emitting β particles contributing to background for dark matter signals. The analysis comprises three parts: (i) obtaining the main observable, the charge energy after applying all data quality cuts, (ii) modeling the beta decay spectrum of ³²Si and ³²P using Betashape software and comparing it to the Fermi theory of beta decay, and (iii) conducting GEANT4 simulations to model other relevant backgrounds present in the experimental setup. A profile likelihood analysis will be performed, utilizing the three aforementioned inputs, to determine the precise level of ³²Si contamination within Si detectors. The analysis is currently in progress, and this thesis will discuss the current status and future prospects of this investigation. This measurement is crucial not only for SuperCDMS SNOLAB, a future upgrade of the CDMS and SuperCDMS experiment, but also for all rare event search experiments utilizing Si detectors.

 

The Super Cryogenic Dark Matter Search (SuperCDMS) experiment uses cryogenic, solid-state detectors to infer the nature of particle interactions in a rare-event search for particles which may constitute dark matter. The experiment relies on the Geant4 Condensed Matter Physics (G4CMP) package to model and understand the underlying physics that compels the response of the solid-state detectors. The G4CMP simulation models solid-state excitations and their relevant processes which includes the transport and scattering of charge carriers and phonons. This report validates that G4CMP's Monte Carlo method for modeling charge carrier and phonon physics is in agreement with the theoretical description as well as evaluates the CPU cost of different processes. We conclude that further investigation is required to confirm the source of CPU consumption with the addendum that the simulation may be further optimized by migrating from a single-scatter model to a multiple-scatter model.

 

Dark matter plays an essential role in understanding modern physics and particles beyond the Standard Model. Evidence suggests that dark matter accounts for approximately 85% of the universe's matter, and 26.8% of its mass-energy composition. Key candidates for dark matter include unidentified subatomic particles like Weakly Interacting Massive Particles (WIMPs) and axions. The Super Cryogenic Dark Matter Search (SuperCDMS) employs direct detection methods to identify these elusive particles using cryogenic technologies. SuperCDMS Soudan is the latest completed CDMS experiment in Minnesota, in preparation for the next phase experiments of SuperCDMS SNOLAB in Sudbury, Canada. At SNOLAB, the Cryogenic Underground TEst (CUTE) facility is dedicated to analyze background levels prior to the full operation of SuperCDMS SNOLAB experiments. Utilizing collected data from CDMSlite Run 3 at Soudan Underground Laboratory with minimized background interference, sensitivity limits were established for solar axions within the keV energy range using the profile likelihood ration method. Our results show an axio-electric coupling constant constraints g_ae<5.91×10^-11 from the atomic recombination and de-excitation, Bremsstrahlung, and Compton channels at a 90% confidence level.

 

Cosmological observations have produced a wealth of evidence which demonstrates that the majority of the matter content in our Universe is "dark". The identity of this dark matter remains elusive, as none of the members of the standard model of particle physics accurately describe its properties. This has prompted the scientific community to launch a broad search, spanning decades in time, mass and sensitivity, with the goal of detecting and identifying this source of new physics. 
The next generation of the Super Cryogenic Dark Matter Search (SuperCDMS) is currently under construction deep underground at SNOLAB. The experiment aims to expand the search for dark matter to lower masses (≲10 GeV/c²) and greater sensitivities using silicon and germanium detectors by minimizing experimental backgrounds and operating detectors with superb energy resolution. 
SuperCDMS will accomplish its low projected background in part by developing a robust shield to protect its detectors from environmental radiation. This dissertation presents the results of simulations which demonstrate the success of the shield design at stopping radiogenic neutrons. The shield will be able to reduce these environmental sources to the point where coherent scattering from solar neutrinos are expected to dominate the nuclear recoil backgrounds. 
In order to search for such light dark matter masses, SuperCDMS uses sensitive transition edge sensors to measure small energy depositions in the detectors. The ultimate energy resolution of these devices, expected to be <1 eV, has not yet been realized. This dissertation describes the analysis of a dark matter search performed at the University of Massachusetts Amherst with a prototype detector which uses SuperCDMS style sensors to achieve a baseline energy resolution of 2.3 eV. The results of this search demonstrate sensitivity to dark matter candidates with masses as low as ∼25 MeV/c².

 

Dark matter direct detection experiments are about to hit a complex obstacle - the irreducible background of solar neutrinos. While this will complicate the search for dark matter, it will usher in the beginning of a new search for beyond Standard Model neutrino physics. However, the use of solar neutrinos as a signal of novel physics in these detectors is still in its infancy. To further explore the potential of next-generation and far-future direct detection experiments in this vein, we consider how deviations in the solar neutrino rate can be used as an indirect probe of new physics in the neutrino sector. We consider beyond Standard Model extensions that can serve as solutions to the present tension in the muon's anomalous magnetic moment, as well as the more general framework of neutrino non-standard interactions. In all cases, we find that future direct detection experiments will be able to either probe as-yet unconstrained new neutrino physics or provide us with information complementary to dedicated neutrino experiments.
We conclude that direct detection experiments are poised to become key players in the field of neutrino physics, contributing to a compelling research mission beyond their search for dark matter.

 

The SuperCDMS Collaboration is searching for dark matter via direct detection of WIMPs. We want to better understand how our detectors respond to both WIMP events and other interactions to help us analyze previously-taken data and also prepare for the next experiment. We focus on the Californium-252 calibration source used in the real experiment, as the neutrons it releases will produce signals similar to those caused by WIMPs; by comparing real Cf-252 data and data produced in our simulation framework, we can trace the physical processes that produce the features we see in the output of the detectors. We find that the initial stages of the simulation are consistent with real Cf-252 behavior and have reasonable detector responses, but we also identify potential improvements to make simulations more realistic. Further, using simulations we identify several measurement complications that might be taken into account to improve analyses for real data.

 

Abundant and diverse evidence suggests that around 85% of the matter in the Universe is non-luminous and non-baryonic. While this 'dark matter' interacts gravitationally with Standard Model matter, its nature and non-gravitational interactions are not yet understood. The SuperCDMS experiment uses cryogenic semiconductor detectors to search for interactions of dark matter with Standard Model matter. The next generation is currently under construction at SNOLAB, with plans to begin commissioning in 2024. Before their operation in the main experiment, the SuperCDMS SNOLAB detectors will be tested at the Cryogenic Underground TEst (CUTE) facility, a low background environment at SNOLAB constructed for this purpose. This thesis describes a dark matter search using data from SuperCDMS Soudan, in which constraints on dark photons and axion-like particles with masses as low as 40 eV/c² are presented. Also described are the projected sensitivities of SuperCDMS SNOLAB and upgrades to the facility for several electron-interacting dark matter candidates. Finally, a characterization study of an R&D detector at the CUTE facility is presented.

 

Many different types of astronomical evidence suggest that most of the mass in the universe is attributable to a nonluminous matter, called dark matter by physicists. Dark matter is believed to be one or more new elementary particles with currently unknown properties. So far, the prevailing method for attempting to discern the properties of the new particle(s) is to observe a direct collision between dark matter and standard model particles. 
SuperCDMS is a direct detection experiment conducted underground that measures ionization and phonon energy in cryogenic germanium crystal detectors. While taking data in the CDMSlite configuration it can amplify the ionization signal, dramatically lowering the threshold for a detectable amount of energy deposited in the detector via a collision with a hypothetical dark matter particle. 
Previous analysis of data taken in the CDMSlite mode examining the elastic scattering limit disregarded any energy loss due to the dark matter passing through the earth and atmosphere to reach the detector. The practical effect of this energy loss is that any dark matter particles which have very high interaction cross sections with normal matter (∼10^-30 cm²) will have lost too much of their energy to be detectable by the experiment once they have reached it. This work will describe a method to account for this effect and present a re-analysis of the CDMSlite run 3 data, resulting in an adjusted sensitivity band, rather than just a lower limit. The lower limit of the new band is consistent with the previously published lower limit obtained without incorporation of the overburden. The upper limit is consistent with the previous limits from other experiments, with an upper limit on the cross section of 10^-31 cm² for a 1.5 GeV WIMP.

 

Neutron-induced nuclear recoils represent a substantial background to many rare-event searches. In this thesis, neutron backgrounds are investigated in the context of searches for two types of events: coherent-elastic neutrino-nucleus scatter (CENNS) searches in the vicinity of nuclear reactors, and dark matter (DM) direct detection experiments and detector calibration in underground sites. Sensitivity of generic silicon and germanium detectors to reactor neutrino CENNS in the presence of neutron capture backgrounds is investigated by simulating detector deployment for varying energy resolutions and ambient thermal neutron fluxes. It is shown that typical reactor-adjacent fluxes must be reduced by several orders of magnitude by shielding or other means in order to make a CENNS measurement with high statistical significance. Germanium detectors are found to have higher sensitivity to the CENNS signal at finer resolutions. It is shown that unbinned and binned likelihood analyses yield similar results. The influence of an active neutron capture veto is discussed. Next, a semianalytical method for simulating neutron-induced backgrounds in SuperCDMS (Super Cryogenic Dark Matter Search) HVeV (high-voltage, electronvolt resolution) detectors is presented. A basic model of radiogenic neutron flux is presented, and the resulting recoil spectrum in silicon is calculated analytically. A new ionization model based on the models of Lindhard and Alig is presented. The resulting neutron-induced phonon spectrum is found to be separated into rounded e/h pair peaks spaced by 120 eV. Rates of neutron-induced events total 68.61±13.52 (syst.)±0.07 (stat.) events kg^-1 day^-1 below 1 keV, with around 0.2 events kg^-1 day^-1 in the first few nonzero-charge e/h pair peaks. It is found that systematic errors dominate the error budget.

 

The Super Cryogenic Dark Matter Search (SuperCDMS) SNOLAB experiment will search for interactions of dark matter particles in solid-state detectors. The dominant expected background at low energy for SuperCDMS SNOLAB comes from radon daughters. This dissertation provides improved understanding to help SuperCDMS achieve its goal of ≤50 nBq/cm² of ²¹⁰Pb contamination. 
My analysis of SuperCDMS Soudan (the predecessor to SuperCDMS SNOLAB) indicates that the ²¹⁰Pb contamination of the detector faces and side regions were 62±5 nBq/cm² and 231±8 nBq/cm² respectively. Bulk ²¹⁰Pb contamination in surrounding materials was 347±7 mBq/kg and likely produced an even larger background than the surface contamination. The angular position of these background events indicated that the detector interface boards (DIBs) mounted on opposing sides of each detector produced ∼5× the bulk event rate as compared with other locations. For SuperCDMS SNOLAB, bulk contamination is being reduced by choosing sufficiently radio-pure materials; surface contamination is being reduced by minimizing exposure to high-radon environments and reducing the radon concentration of the underground cleanroom where detectors are installed. 
In order to reduce the radon concentration of such environments, I led the construction and optimization of the SD Mines radon-reduction system (RRS), which achieved a 4000× reduction in the radon concentration to an activity of ∼25 mBq/m³ (four times lower than the SuperCDMS SNOLAB goal of 100 mBq/m³). I have also helped develop and install the SuperCDMS SNOLAB RRS. 
In order to optimize existing systems and inform development of future systems, I have written a highly configurable simulation of the RRS based on the principles of adsorption and diffusion physics. The RRS simulation has been used to guide operation of the SD Mines RRS and predict the performance of future systems for several compelling hardware changes; these predictions indicate that a radon reduction similar to that of the SD Mines RRS could be achieved at more than double the through-system flow after installing longer columns or an active heating system, or by increasing the pumping speed. 
Finally, I have demonstrated that a post-fabrication, acid-etch cleaning of the detector sidewalls can reduce ²¹⁰Pb surface contamination by >99× at 90% C.L., potentially reducing the ²¹⁰Pb contamination from ∼50 nBq/cm² to ∼20 nBq/cm².

 

Astronomical and cosmological observations suggest that roughly 85% of matter in the universe is in the form of non-luminous dark matter that interacts predominantly via gravity. The preferred explanation is that dark matter is comprised of one or more new particles, which may solve open questions in particle physics. Direct detection experiments look for interactions of galactic dark matter in sensitive detectors using a variety of technologies, including: cryogenic solid state detectors, time projection chambers with liquid noble elements, and bubble chambers with superheated fluids. The cryogenic operating temperature of solid state detectors afford these devices excellent energy resolution and low energy thresholds, providing sensitivity to low-mass dark matter particles. The Super Cryogenic Dark Matter Search (SuperCDMS) is a direct detection experiment that uses cryogenic semiconductor detectors instrumented with superconducting transition edge sensors. The first phase of SuperCDMS took place at Soudan mine in Northern Minnesota, and preparations are being made for the next phase at SNOLAB in Sudbury, ON. Prior to the completion of SuperCDMS SNOLAB, the detectors will be tested in the neighboring experimental bay at SNOLAB in the Cryogenic Underground TEst (CUTE) facility. This thesis describes contributions to the design and construction of CUTE and testing data from SuperCDMS detectors in this facility. As well, results of a dark matter search using data from SuperCDMS Soudan, and results from a prototype detector operated at the University of Massachusetts Amherst will be presented. These data sets provide constraints on axion-like particles and dark photons with masses as low as 40 eV, and thermal relics with masses down to 40 MeV. When probing such low dark matter masses with cryogenic detectors, several experimental and analytical considerations must be taken into account, which will be detailed in this thesis.

 

Numerous astronomical observations suggest that dark matter is a new type of particle and makes up the majority of the mass in the Universe. Still, its discovery and further understanding of its nature remain one of the most significant open questions in all of science. Using combined semiconductor and superconducting technologies, the SuperCDMS experiment has set world-leading limits on small-mass dark matter interactions. However, the ability to extend the sensitivity further is limited by the lack of a full understanding of how detectors respond to O(1 eV) interactions. In this thesis, we study the response of small high voltage detectors to photon interactions from a laser using simulations. We present a comparison between the full simulation results and the experimental data demonstrating that the SuperCDMS simulation successfully reproduces the main features of the data and can be used in conducting simulation-based dark matter searches.

 

In recent years, theoretical and experimental interest in dark matter (DM) candidates have shifted focus from primarily Weakly-Interacting Massive Particles (WIMPs) to an entire suite of candidates with masses from the zeV-scale to the PeV-scale to 30 solar masses. One particular recent development has been searches for light dark matter (LDM), which is typically defined as candidates with masses in the range of keV to GeV. In searches of LDM, eV-scale and below detector thresholds are needed to detect the small amount of kinetic energy that is imparted to nuclei in a recoil. One such detector technology that can be applied to LDM searches is that of Transition-Edge Sensors (TESs). Operated at cryogenic temperatures, these sensors can achieve the required thresholds, depending on the optimization of the design. 
In this thesis, I will motivate the evidence for DM and the various DM candidates beyond the WIMP. I will then detail the basics of TES characterization, expand and apply the concepts to an athermal phonon sensor-based Cryogenic PhotoDetector (CPD), and use this detector to carry out a search for LDM at the surface. The resulting exclusion analysis provides the most stringent limits in DM-nucleon scattering cross section (comparing to contemporary searches) for a cryogenic detector for masses from 93 to 140 MeV, showing the promise of athermal phonon sensors in future LDM searches. Furthermore, unknown excess background signals are observed in this LDM search, for which I rule out various possible sources and motivate stress-related microfractures as an intriguing explanation. Finally, I will shortly discuss the outlook of future searches for LDM for various detection channels beyond nuclear recoils.

 

Dark matter (DM) is one of the most outstanding problems in physics and is a promising hint for physics beyond the Standard Model. Many dark matter detection experiments have been build, with weakly-interacting massive particles (WIMP) as a popular DM candidate. There is also growing interest in light (keV-GeV) dark matter (LDM). The Super Cryogenic Dark Matter Search (SuperCDMS) experiment, which is based on transition-edge sensor (TES), is designed for low-mass WIMPs detection. In the meanwhile, SuperCDMS is using the same technology to build more sensitive detectors to probe some LDM models.  
In this thesis, I will introduce one of the SuperCDMS R&D programs called HVeV which develops high-voltage detectors with eV-scale resolution. The outstanding performance of HVeV detector enabled two topics that are important for DM research and related fields. First, ionization yield is an essential parameter to calibrate the detector response for low-mass WIMP but is not characterized in the target energy region. I used the HVeV detector to measure the ionization yield in silicon down to 100 eV. Second, I studied the backgrounds in HVeV detectors and identified one that dominates. The sensitivity of HVeV detectors can be increased by two orders of magnitude if this background source can be eliminated. I also show a DM exclusion limit with a half day of measurement from the HVeV detector.

 

Many astrophysical observations point to the abundant existence of dark matter (DM) in the universe composed of particles beyond the Standard Model. However despite numerous experiments using different techniques, DM particles have yet to be directly observed. The Super Cryogenic Dark Matter Search (SuperCDMS) SNOLAB experiment will epmply silicon and germanium crystal detectors operated at temperatures as low as 15 mK to probe DM interactions using phonon and ionization signals. Recenly, R&D facilities have developed gram-sized high-voltage eV-scale (HVeV) silicon detectors that achieve single-electron-hole-pair resolution. By probing effective DM-electron interactions, these devices can be used for searches of low-mass DM candidates. 
This dissertation presents a DM search experiment known as HVeV Run 2 that employs a second-generation HVeV detector operated in an above-ground laboratory at Northwestern University (IL, USA). Energy spectra are obtained from a blind analysis with 0.39 and 1.2 g-days of exposure with the detector biased at 60 and 100 V, respectively. The 0.93 gram detector achieves a 3 eV phonon energy resolution, corresponding to a world-leading charge resolution of 3% of a single electron-hole pair for a detector bias of 100 V. With charge carrier trapping and impact ionization effects incorporated into the DM signal models, the resulting exclusion limits are reported for inelastic DM-electron scattering for DM masess from 0.5-10⁴ MeV/c²; in the mass range from 1.2-50 eV/c² the limits for dark photon and axion-like particle absorption are reported. 
Several DM search experiments, including HVeV Run 2, are sensitive to low-mass DM candidates that rely on temperature-dependent photoelectric absorption cross section of silicon. However discrepancies in the underlying literature data result in dominating systematic uncertainties on the DM exclusion limits. In order to reduce these systematic uncertainties, this dissertation presents a novel method of making a direct, low-temperature measurement of the photoelectric absorption cross section of silicon at energies near the band gap (1.2-2.8 eV).

 

As the sensitivity of direct detection experiments improves, they will soon be subject to a new, irreducible background from the coherent elastic scattering of solar neutrinos with nuclei. The presence of new physics can modify this scattering rate, and signals of neutrino scattering may appear in direct detection experiments sooner than expected. In this thesis, we explore the effects of several simplified models of new physics on neutrino scattering at direct detection experiments. We introduce the neutrino contour, a projection of the modified coherent neutrino scattering rate on a dark matter parameter space. This contour can be used to quickly identify whether a direct detection experiment could set competitive constraints on a given model, or conversely, whether the model could produce a large enough neutrino scattering rate to hinder searches for dark matter at that experiment. We discuss the subtleties that arise while computing constraints from the results of one experiment, CDMSlite, in particular the challenges of including electron scattering in the analysis. Finally, we calculate the sensitivity of several future direct detection experiments to one model, the U(1)_Lμ-Lτ. We find that the upcoming LUX-ZEPLIN experiment will be able to test solutions to two ongoing problems in fundamental physics: the muon g-2 anomaly and the H₀ tension.

 

The Super Cryogenic Dark Matter Search (SuperCDMS) is a direct-detection dark matter search experiment that primarily aims to search for Weakly Interacting Massive Particles (WIMPs) using state-of-the-art solid-state detection technology. During its operation at the Soudan underground laboratory, germanium detectors were operated with a high bias voltage mode known as the CDMS low ionization threshold experiment (CDMSlite) to achieve below-keV thresholds. CDMSlite, for being able to measure small energy depositions in detectors, also provides sensitivity to Lightly Ionizing Particles (LIPs) with very small fractional charges. This thesis will discuss an analysis to search LIPs with the data acquired in CDMSlite mode. An important component for LIPs search is the expected energy-deposition distributions for LIPs falling on the CDMSlite detector. In this thesis, a simulation framework to calculate the energy-deposition distributions is developed. This thesis presents first direct-detection limits on the intensity of cosmogenic LIPs with electric charges smaller than e/(3×105) as well as the strongest limits for charges ≤e/160, with a minimum intensity of 1.36×10−7cm−2s−1sr−1 at charge e/160. In any rare-event search experiment, understanding background is crucial. Neutrons capable of mimicking dark matter signals are a major background for any dark matter search experiment. A simulation study to estimate the neutron background for an India-based dark matter search experiment at Jaduguda Underground Science Laboratory (JUSL) is performed. The experiment at JUSL will be the first phase of a proposed Dark matter search at India-based Neutrino Observatory (DINO). It will be a direct detection experiment with primary aims to search for WIMPs as dark matter candidates. In this thesis, we discuss the methodology of estimating neutron flux at JUSL and report the results. The total neutron flux reaching the laboratory above 1 MeV energy threshold is found to be 5.76(±0.69)×10−6cm−2s−1. The impact of neutron background on the sensitivity of the experiment to detect dark matter at JUSL is also discussed. The thesis is organized as follows. Chapter 1 provides a brief introduction to the Lightly Ionizing Particle (LIPs). The analysis to search LIPs in SuperCDMS is briefly outlined in the chapter. This chapter also discusses the importance of neutron background estimates in a dark matter search experiment, more specifically, in the context of a proposed India-based dark matter search experiment at Jaduguda Underground Science Laboratory. In Chapter 2, the SuperCDMS experiment is introduced. In Chapter 3, the framework developed to perform simulations for Lightly Ionizing Particles is presented. In Chapter 4, the LIPs search analysis with the CDMSlite data and the results are discussed. In Chapter 5, the simulation of neutron background and the feasibility of dark-matter search at JUSL is discussed. Finally, conclusions from all the results discussed in this thesis are presented in Chapter 6.

 

Dark matter (DM) makes up roughly 27% of the mass-energy budget of the Universe. Over the past two decades direct detection DM experiments have repeatedly returned null results for DM particles in the mass range of 1 GeV to 1 TeV for different DM interaction cross sections. This has shifted the focus to sub-GeV mass range for DM particles. The thesis has three main parts, (i) development of a silicon detector for low mass DM searches, (ii) measurement of the ionization yield in germanium detectors used by SuperCDMS experiment, and (iii) understanding the backgrounds from various cosmologically activated isotopes in direct dark matter experiments. The ideal requirements in semiconductor detectors used for low mass direct DM search experiments is to build large mass (100 g or more) detectors with single electron sensitivity. One of the challenges in designing such detectors has been reducing their leakage current. The thesis will have a discussion on the development of a 100 g silicon high voltage (Si HV) detector made with a contact free (CF) electrode to reduce leakage current. Its signal-to-noise (S/N) performance is studied which shows a significant improvement when compared to previous detectors of similar mass and dimensions due to the CF electrode. The thesis also discusses the single electron baseline resolution achieved by this detector and how it is an ideal candidate for low mass DM search and coherent elastic neutrino nucleus scattering (CEvNS) experiments. The detector is being used in a reactor-based experiment called Mitchell Institute Neurtino Experiment at Reactor (MINER) at Texas A & M University, USA to measure CEvNS from reactor neutrinos. Direct DM searches work on the principle of an elastic scattering process between a DM particle and the detector material. In semiconductor materials, the amount of energy transferred during this process to create electron-hole pairs in the detector is quantified by the term ionization yield (IY). A precise understanding of the IY reduces the uncertainty in measurement of recoil energy and hence the mass of the interacting particle. Super Cryogenic Dark Matter Search (SuperCDMS) is a direct dark matter search experiments that made use of germanium detectors at HV (i.e. 70 V). These detectors are called CDMS low ionization threshold energy or CDMSlite in short. In a detector like the Ge CDMSlite, where the signal gain is obtained at the cost of measuring IY on an event by event basis, an IY model has to be used. In this thesis we discuss the analysis of a dedicated calibration run using two photo-neutron sources 124Sb+9Be and 88Y+9Be to measure the IY in germanium CDMSlite detectors at ≈50 mK temperature. The analysis will make use of a likelihood method to extract the yield from data. The likelihood method takes three inputs; (i) A neutron energy distribution from GEANT4 simulations, (ii) A background probability distribution function from data, and (iii) a modified data-driven IY model inspired by Lindhard et. al. The thesis will also have a brief overview on the possible backgrounds in the Super-CDMS SNOLAB experiment. The main contributors to backgrounds will be discussed.Tritium is a major background source for direct low mass DM searches. There will be a discussion on the production rate estimates of 3H, 55Fe, 65Zn, and 68Ge from the second run of the CDMSlite. Overall, the thesis will have 5 chapters. Chapter one will be an introduction to dark matter and emphasize the need for low mass dark matter search. Chapter 2 will discuss the development and characterization of the Si HV detector. Chapter 3 will focus on the ionization yield measurement in SuperCDMS. Chapter 4 will be an overview about the major backgrounds in dark matter searches. And Chapter 5 will be conclusion and outlook.

 

The Weakly Interacting Massive Particle (WIMP) has historically been a prime candidate for dark matter due to its elegant compatibility with the Minimally Supersymmetric Standard Model (MSSM). Recent dark matter experiments have ruled out much of the GeV~TeV mass range predicted by the MSSM, however, and the newest generation of direct detection experiments, such as SuperCDMS SNOLAB, have begun to explore sub-GeV dark matter. As experiments push towards lower masses, sensitivity to eV-scale electron recoil events has become increasingly important. A variety of unexplained excesses at energy deposits of 1~100 eV have been found in many such low-mass experiments, across different detection techniques and at different excess rates. These low-energy events are not thought to be dark matter, and they must be understood and effectively removed to further improve detector energy resolution. This thesis will outline the simulation of one possible source of low-energy excess at SuperCDMS: optical photons produced from charges in uniform motion, specifically from Cherenkov radiation (CR), transition radiation (TR), and the intermediate hybrid transition-Cherenkov radiation (HR). In the latter case, the standard formulae for HR contain divergences at certain angles in transparent media, which are an artifact of their derivation and become problematic to simulate. To resolve this, we present a novel approach to normalize the divergent HR peaks, which allows HR to transition smoothly between TR and CR in a numerical simulation. We also add TR and CR as a physics process to the SuperCDMS Monte Carlo simulation package SuperSim, based on the Geant4 simulation toolkit. To verify the physics of our addition, we show some test simulations in comparison to theoretical predictions of optical radiation intensity. We also use the simulation to make some preliminary predictions on the contribution of TR and CR to the low energy background, with the expectation that more thorough analyses will be conducted in the future.

 

The mystery of Dark Matter in our universe has puzzled scientists over the past century, and a large number of experimental ventures have been conducted hoping to find an answer. The Weakly Interacting Massive Particle (WIMP) is a well motivated dark matter candidate with a number of direct detection experiments dedicated to finding WIMPs or excluding them from being the solution to the dark matter problem. SuperCDMS is a cryogenic dark matter experiment with semiconductor detectors currently being constructed in the underground low background laboratory SNOLAB, with plans to begin taking science data in 2023. The Cryogenic Underground TEst (CUTE) facility is presently operating at SNOLAB to test de detectors which will go in the SuperCDMS experiment, and perform preliminary science runs using SuperCDMS detectors in a low background, low noise environment. Calibrations of detectors is critical in order to understand and quantify the signals which they produce. The design, construction, and testing of the neutron calibration system planned to be installed in CUTE are highlighed, and an analysis of the collection efficiency of a detector was conducted by identifying the signal associated with ⁷¹Ge decay. This thesis also discusses the contributions to the improved performance of CUTE through noise studies, and cooldown optimization.

 

An encouraging avenue of investigation beyond the Standard Model (SM) is the search for particles with fractional electric charge. While the SM does not anticipate the existence of free particles with a fractional electric charge, fractionally charged particles (FCPs) have not been experimentally excluded. Free fractionally charged particles are a feature of viable extensions to the Standard Model with extra U(1) gauge symmetries. This thesis describes a FCP search using CDMSlite Run 2 Period 1 data which results in the first direct detection limits on cosmogenic FCPs with an electric charge as small as e/108 and the most stringent limit on vertical intensity for FCPs with an electric charge ≤e/160. This analysis is also the first to consider cosmogenic FCPs with a wide range of masses (5 MeV/c2 – 100 TeV/c2) and velocities (βγ=0.1−106 ). The intensity limit computed for an assumed cos2θ angular distribution is nearly three times weaker than that for an isotropic angular distribution for most assumed values of f .

 

The desire to unveil the mystery of dark matter which, according to compelling astrophysical and cosmological evidence, constitutes about 85% of the matter in the Universe, drives physicists to continuously develop their tools and search methodologies. There are several proposed candidates for new particles among which the Weakly Interacting Massive Particle (WIMP) is considered the most propitious. 
The SuperCDMS SNOLAB experiment is a direct dark matter (DM) detection experiment. It is expected to start taking science data by the beginning of the year 2023. The plan is to start the operations with a total payload of ∼30 kg of Ge and Si detectors, situated in a well-shielded cryostat with ∼15 mK at the coldest thermal stage. The experiment aims to probe DM with masses down to 0.3 GeV (0.5 MeV) through nuclear (electron) scattering. To accomplish this, detectors have to have energy thresholds of a few eV 
The Cryogenic Underground TEst facility (CUTE) is currently the underground facility equipped to operate the future SuperCDMS detectors. Since summer 2019, the facility has operated a variety of detectors to identify, quantify, and mitigate its noise sources and investigate its potential for dark matter search while the construction of the SuperCDMS experiment is in progress. In this thesis, I discuss key contributions to setting up and commissioning the CUTE facility in a timely manner. I also summarize the major tests and findings. 
The low energy reach of the new SuperCDMS detectors brings the need for new calibration methods. We propose two different methods that use LEDs operated next to the detectors. This has initiated the study of LED properties at low temperatures. One of the suggested methods has been shown to be viable, and preliminary tests for the other method have shown promising results. Finally, I show the possibility of operating LEDs in the proximity of an eV-sensitive detector without impacting its performance.

 

The dark matter (DM) problem is currently one of the most pressing issues facing physics. This thesis presents some of my work and contributions to the detectors employed by the Super Cryogenic Dark Matter Search (SuperCDMS), one of the many experiments currently investigating DM. First, we present an overview of astrophysical and cosmological observations to give insight into our current understanding of DM and its role in the universe. Some potential particle DM candidates are also discussed. Then we describe the SuperCDMS experiment and relevant detector technologies. We go into particular detail in discussing the phonon sensors, deriving the relationships that define their fundamental working principles and describing their practical deployment. We also discuss the main noise contributions to the phonon sensors which can limit experimental sensitivity.
We then present some work related to reconstruction of event position in the detectors, including new analysis techniques and a novel cryogenic calibration source mover we have developed and tested. We also present tests of improving experimental scalability by increasing individual detector masses using the largest-diameter cryogenic Si detectors yet operated. We present first experimental results from the novel use of (n,γ) processes to calibrate the low energy nuclear recoil energy scale. This work is essential to understand how DM-nuclear interactions will manifest in semiconductor detectors. We also take a short look at the newest, most sensitive detectors currently operated by SuperCDMS and discuss how they are allowing new insights into critical physics processes in all SuperCDMS detectors. Finally, we demonstrate how some of the various improvements discussed in this work can improve the experimental reach of the next generation of SuperCDMS at SNOLAB.

 

The SuperCDMS experiment uses cryogenic silicon and germanium detectors to search for dark matter candidates such as WIMPs (Weakly Interacting Massive Particles) streaming through the Earth. Collisions in the silicon and germanium crystals are expected to produce phonons whose thermal signatures can be measured.
This thesis first describes the integration of a new Signal Distribution Unit (SDU) to the SuperCDMS data acquisition system, which allows for synchronization of multiple detectors and electronic/mechanical noise characterization via accelerometer, antenna, and AC phase measurements.
From SuperCDMS detector data it is necessary to reconstruct the energies of the particle events. This thesis explores the use of Convolutional Neural Networks (CNNs) to perform this reconstruction and finds that, although they perform well, changing the noise model breaks the model and requires the neural network to be retrained. In order to mitigate this issue, a new CNN model is proposed which includes the noise Power Spectral Density (PSD) of the data as an additional input to the CNN. While it proves to be effective as a denoising algorithm, it still fails for data with a different noise model. However, including data from multiple PSDs in the neural network training sample allows it to handle data with different types of noise while still maintaining the quality of the reconstruction. Nevertheless, neural networks trained even on multiple PSDs do not robustly handle data taken with PSDs dissimilar to those in the training sample, suggesting that CNNs may need to be retrained whenever the noise environment changes in a significant way.

 

Astronomical observations lead us to believe that dark matter constitutes the majority of the mass of the Universe, yet it has never been directly observed experimentally. The Super Cryogenic Dark Matter Search Experiment (SuperCDMS) aims to discover the particle nature of dark matter, by developing and using some of the most sensitive detectors ever built. The SuperCDMS detectors are capable of rejecting known particles to search for dark matter interactions, but the discrimination is a complicated analysis task. Previous analyses with SuperCDMS set world leading sensitivity limits to dark matter interactions, but the analysis techniques have not been based on simulations. Now we have developed an efficient simulations infrastructure that will help better understand the detectors, develop a better background model, and optimize the next generation dark matter searches. In this thesis we present the fully-functional simulations infrastructure to produce high quality samples as well as results for the charge system readout simulations, and comparisons to expectations from data. We show that our simulations well-reproduce dominant features of data, bringing SuperCDMS closer to a robust simulation-based analysis method, much needed for the upcoming SuperCDMS dark matter searches.

 

Astronomical and cosmological observations indicate that large amounts of some slowly moving, unseen Dark Matter pervades the universe and outweighs normal matter by a factor of five. The Standard Model of Physics has no contender with the properties of this as-yet-undetected particle, so experimentalists build state-of-the-art radiation detectors to attempt to directly measure this Dark Matter. The Super Cryogenic Dark Matter Search (SuperCDMS) at Snolab is one such experiment, currently being constructed. This experiment uses ultra-cold, superconducting, high-purity silicon and germanium detectors to measure low-energy nuclear recoils from the elastic scattering of Dark Matter. I contributed to the calibration of the energy scale of low-energy nuclear recoils in CDMS II silicon detectors by computing an improved livetime correction for calibration data to verify the Monte Carlo rate normalization. Results indicate that the phonon collection efficiency of nuclear recoils relative to electron recoils is 95.2 +0.9 −0.7 %, and the ionization collection efficiency of low-energy nuclear recoils in silicon is lower than Lindhard prediction, consistent with other recent measurements.
Backgrounds from the progeny decay of the abundant, naturally-occurring radioactive isotope radon-222 obstruct the sensitivity of essentially every dark-matter search. Radon concentrations in the Snolab cavern would contribute prohibitively large backgrounds if the volume surrounding the detectors were not purged with a low-flow low-radon gas. By measuring the radon diffusion and emanation, we identified acceptable gasket materials for sealing this radon purge, ensuring that the radon-induced backgrounds will be significantly lower than the other experimental backgrounds. A radon emanation system with a gas handling system and low-background radon detector was commissioned and used to measure the radon emanation of the proposed gaskets. A low-cost apparatus was constructed to measure the radon diffusion of gaskets with a commercial radon detector.
The sensitivity of future generations of dark-matter detectors are expected to be dominated by long-lived low-energy beta- and alpha-emitting radon daughters such as 210 Pb on detector surfaces. I describe simulations indicating the detector could also be used to reduce background from material impurities plaguing rare-event searches, the commissioning of a prototype demonstration detector, and a gas handling system necessary to operate the detector. I demonstrated that the gas handling system reduces the otherwise dominant backgrounds by a factor of 62. This detector will therefore be able to detect 32 Si and 210 Pb 100 times better than currently available screeners.

 

Direct detection dark matter experiments are increasingly interested in the low-mass dark matter parameter space, but zero-background low-mass searches require event separation between the electron and nuclear recoil bands, which requires a proper understanding of detector energy reconstruction.

Previous simulations have shown that we do not entirely understand the ionization efficiency (yield) for electron and nuclear recoils, as the assumption that the distribution for the yield is normally distributed for a true recoil energy is violated. Since the yield distribution may directly affect dark matter low-mass limits, it is crucial we understand how the yield is distributed.

A component to understanding the yield distribution is the variance in the number of electron-hole pairs produced or ionization variance. This effect has been studied relatively infrequently as experiments have been interested in large energy deposits (10 - 100 keV) and could accurately separate electron and nuclear recoil events. For electron recoils, the ionization variance is described by a “Fano” factor. For nuclear recoils the effect can be parameterized by an “effective” Fano factor, which has similar definition but a different physical origin. The nuclear recoil “effective” Fano factor is shown to be much larger than the electron-recoil version above around 10 keV deposited energy

 

An abundance of evidence suggests that most of the Universe is composed of nonluminous matter. This "dark matter” is believed to be a new elementary particle and experiments around the world are attempting to directly detect rare collisions with terrestrial detectors.

The properties of dark matter have yet to be identified, thus efforts are ongoing to explore a range of possible masses and interaction cross-sections. For the latter, experiments can increase exposure by scaling up the detector mass and operating for a longer time. To search for dark matter with less mass than a nucleon, new technologies and analysis techniques need to be developed to be sensitive to energy deposits less than a few keV.

SuperCDMS is a direct detection experiment that measures ionization and phonon energy in cryogenic germanium crystal detectors. A special mode of operating the SuperCDMS detectors, called CDMSlite, amplifies the ionization signal via phonon creation. This amplification leads to a lower energy threshold, which provides sensitivity to smaller dark matter masses.

Typically, direct detection experiments assume dark matter scatters elastically off the nuclei in the detector. In this thesis, I will highlight the most recent CDMSlite elastic dark matter search. Then I will describe how inelastic dark matter collisions can manifest in the detector and be useful for extending experimental sensitivity to lower dark matter masses. Finally, I will present a re-analysis of CDMSlite data using a profile likelihood to search for low-mass dark matter through two inelastic scattering channels: Bremsstrahlung radiation, and the Migdal Effect.

 

SuperCDMS is a direct-detection dark matter (DM) experiment that uses cryogenically cooled germanium and silicon detectors to search for interactions between DM particles and detector nuclei, and in this thesis I describe my contributions to the experiment. I start with a brief review of DM and motivate the possibility of its detection in underground laboratories with sensitive detectors, and I review the SuperCDMS detector fundamentals. Then I focus on detector development for the future generation of the experiment, which will deploy an array of detectors at SNOLAB in Sudbury, Canada. Specifically I describe characterization of prototype detectors from surface facility testing, and discuss measurements of critical values that determine the detectors' sensitivity to DM particles, such as the baseline resolution and the phonon collection efficiency. I also describe analysis techniques developed to measure intrinsic detector noise in a high radiation environment such as a surface test facility. In the final chapters I describe a DM search analysis using four months of data from operation of SuperCDMS detectors in the Soudan Mine in northern Minnesota. I discuss how a particular detector operating mode, called CDMSlite, lowers the energy threshold of the detectors in order to improve the sensitivity to low-mass DM particles. I also present new analysis techniques that optimize the sensitivity to low-mass DM particles, including noise discrimination with multivariate classifiers, instrumental background modeling, and a profile likelihood signal and background fitting approach. In this analysis we set an upper limit on the DM-nucleon scattering cross section in germanium that is a factor of 2.5 improvement over the previous CDMSlite result for a DM mass that is five times the proton rest mass.

 

The majority of matter in the Universe is non-luminous, non-baryonic, and as of yet undiscovered. The SuperCDMS experiment searches for this "dark matter" in the form of Weakly Interacting Massive Particles (WIMPs), operating semiconducting detectors at cryogenic temperatures to measure energy deposits from dark matter interactions. These efforts have resulted in world leading limits on the interactions of dark matter with ordinary matter. The next phase of SuperCDMS is under construction at SNOLAB near Sudbury, ON.  
Interaction of cosmic rays with detector material can result in the production of radioactive isotopes, the decay of which results in background in the measured energy spectrum. Tritium is expected to be the limiting background for SuperCDMS SNOLAB. Production rates of tritium and other cosmogenic isotopes are not well known, with a wide spread in theoretical calculated rates and only a single measurement previously published for tritium. The work in this thesis details the measurement of cosmogenic production rates using low energy spectra measured by SuperCDMS.

 

Astrophysical measurements and cosmological predictions suggest the exists of a large amount of matter in the universe that does not interact via electromagnetic forces. This non-luminescent matter, known dark matter, exists in halos that encompass and are within galaxies, including the Milky Way. Therefore, dark matter particles should should be directly detectable by experiments on Earth, such as the Super Cryogenic Dark Matter Search (SuperCDMS). Dark matter is assumed to be low mass (< 100 GeV/c2) and interact via the weak force using either a spin-independent or spin-dependent coupling. However, making incorrect assumptions about dark matter interactions can lead to misleading results. Because interactions with dark matter particles are rare, direct detection experiments must be able to shield for or reject backgrounds to very low levels. Low energy neutron backgrounds that make it to the detectors are especially dangerous, because they cannot be easily distinguished from the expected dark matter signal.
Scintillator doped with a high neutron-capture cross-section material can be used to detect neutrons via their resulting gamma rays. Examples of such detectors using liquid scintillator have been successfully used in past high-energy physics (HEP) experiments. However, a liquid scintillator can leak and is not as amenable to modular or complex shapes as a solid scintillator. The light outputs and efficiencies of gadolinium-loaded polystyrene-based scintillators have been explored using a wide variety of gadolinium compounds with varying concentrations. Collection strategies using a wavelength shift- ing (WLS) fiber and silicon photomultipliers (SiPMs) were also evaluated as a possible neutron veto for an upgrade to SuperCDMS SNOLAB.
The scattering of dark matter particles off nuclei in direct detection experiments can be described in terms of a multidimensional effective field theory (EFT). A new systematic analysis technique is developed using the EFT approach and Bayesian inference methods to exploit, when possible, the energy-dependent information of the detected events, experimental efficiencies, and backgrounds. Highly dimensional likelihoods are calculated over the mass of the weakly interacting massive particle (WIMP) and multiple EFT coupling coefficients, which can then be used to set limits on these parameters and choose models (EFT operators) that best fit the direct detection data. Expanding the parameter space beyond the standard spin-independent isoscalar cross section and WIMP mass reduces tensions between previously published experiments. Combining these experiments to form a single joint likelihood leads to stronger limits than when each experiment is considered on its own. Simulations using two nonstandard operators (O3 and O8) are used to test the proposed analysis technique in up to five dimensions and demonstrate the importance of using multiple likelihood projections when determining constraints on WIMP mass and EFT coupling coefficients. In particular, this shows that an explicit momentum dependence in dark matter scattering can be identified.
CDMSlite Run 2 was a search for Weakly Interacting Massive Particles (WIMPs) with a cryogenic 600 g Germanium detector operated deep underground. It was operated in a mode optimizing sensitivity to WIMPs of relatively low mass, 2 - 20 GeV, while sacrificing background rejection. An EFT analysis of CDMSlite Run 2 data from SuperCDMS Soudan is presented here. A binned likelihood Bayesian analysis was performed on the data, optimizing over the parameters of EFT interactions and the recoil energy spectra due to the dominant Compton scattering and tritium backgrounds. Recoil energy regions within 5σ of known activation peaks were removed from the analysis. The Bayesian evidences of the resulting likelihoods show that CDMSlite Run 2 data is entirely consistent with the background models with no EFT interaction necessary. Upper limits on the WIMP mass and coupling coefficients amplitudes and phases are presented for each EFT operator.

 

Over the past eighty years, numerous complementary observations of our universe have indicated that our current description of physics is far from complete. The "ordinary matter", such as electrons, protons, photons and neutrons, that constitutes the bulk of all human physical experiences is actually only a minority (about 16%) of the total mass of the universe. The remaining 84% is very poorly understood, but has profound effects on the dynamics and evolution of our universe. Because it does not interact with light, and is not observable in telescopes on earth, this extra mass is usually referred to as "dark matter". Although the dark matter is poorly understood, Weakly Interacting Massive Particles (WIMPs) are a well-motivated candidate that can be directly detected via a non-gravitational interaction with normal matter, potentially allowing for direct terrestrial detection and characterization of this dark matter. This dissertation is focused on this direct WIMP detection and will be broken into two main parts.

The first part focuses on the blinded analysis of roughly three years of data collected from March 2012 to November 2015 by the SuperCDMS Soudan experiment. SuperCDMS Soudan consists of an array of 15, 0.6-kg, cryogenic, Ge iZIP particle detectors situated in a decommissioned iron mine in remote northern Minnesota. This analysis is optimized to be sensitive to theoretical WIMP masses above 10 GeV/c2. This result set the strongest limits for WIMP--germanium-nucleus interactions for WIMP masses greater than 12 Gev/c2.

The second part focuses on the development new kind of particle detector in the style of a SuperCDMS iZIP, designed to simplify fabrication and readout, improve phonon-based position reconstruction, and help to scale to larger target arrays. These detectors replace the TES-based phonon sensors of the iZIP with Microwave Kinetic Inductance Detectors (MKIDs).

 

Years of astrophysical observations suggest that dark matter comprises more than ~80 % of all matter in the universe. Particle physics theories favor a weakly-interacting particle that could be directly detected in terrestrial experiments. The Super Cryogenic Dark Matter Search (SuperCDMS) Collaboration operates world-leading experiments to directly detect dark matter interacting with ordinary matter. The SuperCDMS Soudan experiment searched for weakly interacting massive particles (WIMPs) via their elastic-scattering interactions with nuclei in low-temperature germanium detectors. During the operation of the SuperCDMS Soudan experiment, 210Pb sources were installed to study background rejection of the Ge detectors. Data from these sources were used to investigate energy loss associated with Frenkel defect formation in germanium crystals at mK temperatures. The spectrum of 206Pb nuclear recoils was examined near its expected 103 keV endpoint energy to extract the first experimentally determined average displacement threshold energy of 19.7 +/- 0.5 (stat) +/- 0.1 (syst) eV for germanium. This has implications for the sensitivity of future germanium-based dark matter searches including the SuperCDMS SNOLAB experiment. The SuperCDMS SNOLAB experiment will employ germanium and silicon detectors to improve current WIMP-search results by at least one order of magnitude for masses <= 10 GeV/c2. This will require substantial shielding against cosmogenic and radiogenic backgrounds. The SuperCDMS SNOLAB passive shield will be permanent for the duration of the experiment so extensive simulations were undertaken to optimize the shield design. This resulted in a design of an outer layer of 60 cm of water, a middle layer of 20 cm of lead, and 30 cm of polyethylene which limits the background rate to that required for the primary physics goals of the experiments. The experiment will begin operations in 2020 and care must be taken during the construction phase to limit exposure to the ~135$ Bq/m3 radon activity in the laboratory. The daughter products of 222Rn can attach to nearby surfaces leaving long-lived 210Pb in place for the duration of the experiment. For non-line-of-sight surfaces of the polyethylene shield, the maximum allowable 210Pb activity is 10,000 nBq/cm2. A study was conducted to experimentally determine the contamination rate of polyethylene and copper by exposing samples for 83~days at SNOLAB. From the resulting surface activities, obtained from high-sensitivity measurements of alpha emissivity using the XIA UltraLo-1800 spectrometer, the average 210Pb plate-out rate was determined to be 249 and 423 atoms/day/cm2 for polyethylene and copper, respectively. A time-dependent model of alpha activity was developed leading to a maximum exposure time of 39 days in the SNOLAB environment.

 

In this thesis, I cover three elements of the design of the CDMS HV detectors. I discuss the detector physics controlling how charges and phonons are generated in our detector crystals, com- paring theory to results of recent experiments carried out at Stanford. I move on to describe the operating principles of our phonon-mediated charge readout, as well as the design of the CDMS HV detector. I then describe the performance tests of early CDMS HV prototypes in conjunction with the SuperCDMS SNOLAB electronics, and discuss the path towards achieving single electron-hole pair resolving detectors at the kg-scale given the performance obtained thus far. As a result of these tests, we were able to refine our noise and sensor dynamics models, and develop new metrics for diagnosing non-ideal sources of noise to aid in reducing coupling of the external environment to our detectors.

 

The CDMSlite spectrum had a large contribution from electron recoil background events. From the information gained during the first two CDMSlite Runs, a background model was developed for the third and final CDMSlite Run. Analytical descriptions were identified for those backgrounds that were theoretically known, e.g. tritium β-spectrum, and Geant4 simulations were used to understand and predict the low-energy spectra from other sources, e.g. Compton scattering.

Multiple new models were developed for detectors operated in CDMSlite at Soudan. These include the analytical formula for Compton scattering, and empirical models for surface backgrounds from ²¹⁰Pb contamination of the germanium crystals and detector housing. In order to accurately describe the surface events, a new detector response function was developed that included information about the electric field and energy resolution of the detector. These models were essential to the implementation of a profile likelihood analysis of the CDMSlite Run 3 data, which improved on the sensitivity to dark matter over the Run 2 optimum interval analysis for WIMP masses above 2.5 GeV/c². This demonstrated a successful application of a likelihood analysis to the high-voltage operating mode, and the potential for these analyses in the future SuperCDMS SNOLAB experiment.

 

The detectors of the Cryogenic Dark Matter Search (CDMS) measure phonons and charge emanating from recoiling germanium nuclei. Its phonon sensors, though exquisitely sensitive, are relatively complicated to fabricate and deploy. Responding to the scientific imperative to increase dark matter sensitivity, we look for a simpler phonon sensor technology. Kinetic inductance based phonon sensors (KIPS) are superconducting thin film microwave frequency resonant circuits. Phonon energy impinging on the film alters its surface impedance, which alters the characteristics of the resonance. KIPS can be made with relatively large, millimeter sized features which makes them easier to fabricate and reduces variations between detectors.

This instrumentation focused thesis shows that KIPS can be used as a simple and sensitive phonon sensor for the CDMS detector. KIPS design aspects, competitiveness to the current transition edge phonon sensors, readout considerations and suggestions on how to instrument them in future dark matter experiments will be presented. The broader applicability of KIPS in nuclear non-proliferation and other physics investigations is also discussed.

 

The present work is a description of the search for a direct detection of super-symmetric dark matter via scattering from standard model particles. The Cryogenic Dark Matter Search (CDMS) Experiment uses ionization and athermal phonon sensor technologies to achieve event-by-event discrimination of electron and nuclear recoils in cryogenic germanium crystal detectors. At low energies, where the the ability to discriminate individ- ual nuclear recoil events from background is reduced, a periodic variation of the rate and crossover signatures in the energy spectrum can aid the identification of a WIMP signature in the presence of significant backgrounds. In general, the direct detection of dark matter is the first step toward the identification and classification of dark matter in the universe.

This work describes a background-subtracted search for annual modulation in the WIMP- search data acquired in the Cryogenic Dark Matter Search II (CDMS II) Experiment, which was the second implementation of the highly successful CDMS technology. We observe no significant modulation in the 2.7 keVnr to 11.9 keVnr (nuclear-recoil-equivalent) energy range selected for this analysis. These results are not compatible with a WIMP dark matter interpretation of the signals reported by the DAMA/LIBRA and CoGeNT experiments, and provide complementary support to earlier CDMS low-threshold germanium analyses.

 

The SuperCDMS test facility at University of Minnesota aids in the detector R&D and characterization of prototype detectors, as part of the scale-up effort for SuperCDMS SNOLAB. This thesis presents the first full ionization and phonon characterization study of a 100 mm diameter, 33 mm thick prototype Ge detector with interleaved phonon and ionization channels. Measurements include ionization collection efficiency, surface event rejection capabilities, and successful demonstration of nuclear recoil event discrimination. Results indicate that 100 mm diameter, interleaved Ge detectors show potential for use in SuperCDMS SNOLAB.

As part of detector R&D, the Minnesota test facility also looks beyond the next stage of SuperCDMS, investigating larger individual detectors as a means to easily scale up the sensitive mass of future searches. This thesis presents the design and initial testing results of a prototype 150 mm diameter, 33 mm thick silicon ionization detector, which is 5.2 times larger than those used in SuperCDMS at Soudan and 2.25 times larger than those planned for use at SuperCDMS SNOLAB. In addition, the detector was operated with contact-free ionization electrodes to minimize bias leakage currents, which can limit operation at high bias voltages. The results show promise for the operation of both large volume silicon detectors and contact-free ionization electrodes for scaling up detector mass and bias.

 

A high mass WIMP search analysis is performed on the recent collected data sets at Soudan. It aims at exploring WIMPs with masses from the order of 10 GeV/c² and above. With a raw exposure of 1657.54 kg-days, an exclusion limit is set on the spin-independent WIMP-nucleon cross section at 1.32×10⁻⁴⁴ cm² for a 75 GeV/c² WIMP at a 90% confidence level by the profile likelihood ratio technique.

Neutrons are the most dangerous background in a direct dark matter search experiment. The nuclear recoil signal that a neutron produces in a Ge detector is indistinguishable from that a WIMP produces. Protection against them is one of the key aspects for the next generation of SuperCDMS experiment at SNOLAB. An active neutron veto system was proposed to be implemented in this future experiment to make it more robust from neutrons. The feasibility of both the plastic and liquid neutron veto systems was studied.

 

SuperCDMS searches for Dark Matter in the form of WIMPs. SuperCDMS is able to reject most common background events by discriminating between electron and nuclear recoils. In the CDMS low ionization threshold experiment mode, electron recoil discrimination is sacrificed when applying a stronger electric potential across the detector. The electric field associated with this potential is not uniform, and the amplification of the phonon signal caused by drifting charges through this potential creates a radial dependence on the reconstructed energy. This thesis explores a new method of identifying and rejecting events at the edge of the detector in order to improve the ratio of signal to background by excluding background events with improperly reconstructed energy.

 

A significant evidence from galaxies and astrophysical observations, suggests that ∼ 85% of the matter in our Universe is invisible matter. The observations of the so-called "dark matter" suggest that it consists of non-relativistic, non-baryonic particles, which seldom interact with baryonic matter, or with each other. Many experiments are searching for dark matter, each of which is based on a particular dark matter candidate. Weakly Interacting Massive Particles (WIMPs) are one of the well-motivated candidates for dark matter. So far, no answers were provided by the Standard Model of particle physics to the dark matter puzzle.   

The Super Cryogenic Dark Matter Search experiment (SuperCDMS) is considered one of the pioneer experiments in the direct search for WIMPs. It is based primarily on deploying germanium and silicon detectors at cryogenic temperatures to search for direct WIMP-nucleus elastic scattering interaction through which lattice vibrations are generated and sensed in one of the coldest detectors ever built.   

The new phase of SuperCDMS experiment at SNOLAB is aiming to be sensitive to the lower WIMP mass scale. Therefore, a lower background and detector threshold energy is a necessity, and the detectors need to be calibrated and tested for the new proposed sensitivity. The tests include high bias voltages, which are required to increase the gain in signal-to-noise ration and to allow for the detection of low energy events using the phonon signal. However, the upper limit and polarity for the bias voltage need further studies in order to understand the variation of the detector's response to high voltage. Therefore, we performed the breakdown measurement (chapter 4) at Queen's Test Facility.   

Moreover, detectors have to be calibrated before being utilized in measuring low energy interactions, and that is what lead to the use of infrared photons. Once we can calibrate and understand the behavior of infrared photons in germanium detectors, they can be utilized in calibrating germanium detectors at the lower energy scale. Therefore, we performed the infrared calibration measurement which represents the bulk of the work in my thesis.

 

The detectors themselves, however, are quite complex; and a very detailed under- standing of the microscopic physics is helpful in analyzing the very rare events that occur within them. Furthermore, better understanding and modeling of the detectors can aid in the design and optimization of future iterations of the experiment.

This work describes the design and implementation of a low temperature condensed matter physics simulation library built on top of the popular Geant4 particle tracking framework. The library, named “Geant4 Condensed Matter Physics” or G4CMP, intro- duces several solid state concepts to the Geant4 framework such as crystal lattices, phonon quasiparticles, non-scalar effective masses, and implements several physics processes relevant to cryogenic temperature crystals.

In addition to the physics library, which is intended for general use, this work also describes a full Monte Carlo simulation package for the SuperCDMS iZIP detectors which utilizes G4CMP at its core and also fully simulates the detector readout sensors.

 

This paper discuses two projects that were designed around the SuperCDMS experiment. The first is an attempt to understand the SuperCDMS simulation that was designed in the Matlab environment. The tests that were conducted were designed to study how electron and hole propagation is effected by the initial event location and the charge bias within the detector. The second experiment was an attempt to systematically quantify the level of noise produced within the SuperCDMS detectors by the dilution refrigerator. The results show only initial findings and no conclusive results were obtained.

 

Lower-mass dark matter O(10 GeV/c²) has become more prominent in the past few years. The CDMS detectors can be operated in an alternative, higher-biased, mode to decrease their energy thresholds and correspondingly increase their sensitivity to low-mass WIMPs. This is the CDMS low ionization threshold experiment (CDMSlite), which has pushed the frontier at lower WIMP masses. This dissertation describes the second run of CDMSlite at Soudan: its hardware, operations, analysis, and results. The results include new WIMP mass-cross section upper limits on the spin-independent and spin-dependent WIMP-nucleon interactions. Thanks to the lower background and threshold in this run compared to the first CDMSlite run, these limits are the most sensitive in the world below WIMP masses of ~4 GeV/c². This demonstrates also the great promise and utility of the high-voltage operating mode in the SuperCDMS SNOLAB experiment.

 

Direct dark matter detection experiments usually have excellent capability to distinguish nuclear recoils, expected interactions with Weakly Interacting Massive Particle (WIMP) dark matter, and electronic recoils, so that they can efficiently reject background events such as gamma-rays and charged particles. However, both WIMPs and neutrons can induce nuclear recoils. Neutrons are then the most crucial background for direct dark matter detection. It is important to understand and account for all sources of neutron backgrounds when claiming a discovery of dark matter detection or reporting limits on the WIMP-nucleon cross section. One type of neutron background that is not well understood is the cosmogenic neutrons from muons interacting with the underground cavern rock and materials surrounding a dark matter detector. The Neutron Multiplicity Meter (NMM) is a water Cherenkov detector capable of measuring the cosmogenic neutron flux at the Soudan Underground Laboratory, which has an overburden of 2090 meters water equivalent. The NMM consists of two 2.2-tonne gadolinium-doped water tanks situated atop a 20-tonne lead target. It detects a high-energy (>∼ 50 MeV) neutron via moderation and capture of the multiple secondary neutrons released when the former interacts in the lead target. The multiplicity of secondary neutrons for the high-energy neutron provides a benchmark for comparison to the current Monte Carlo predictions. Combining with the Monte Carlo simulation, the muon-induced high-energy neutron flux above 50 MeV is measured to be (1.3 ± 0.2) × 10 −9 cm −2 s −1 , in reasonable agreement with the model prediction. The measured multiplicity spectrum agrees well with that of Monte Carlo simulation for multiplicity below 10, but shows an excess of approximately a factor of three over Monte Carlo prediction for multiplicities ∼ 10 − 20. In an effort to reduce neutron backgrounds for the dark matter experiment SuperCDMS SNO- LAB, an active neutron veto was developed. It is estimated that the current design of the neutronveto with a 40 cm thick layer of boron-doped liquid scintillator can achieve a > 90% efficiency for tagging the single-scatter neutrons. In addition, a one-quarter scale prototype detector for neutron veto has been built and tested.

 

Different pieces of evidence obtained during the last decade, from galactic to cos-mological scales, has led to the conclusion that the Universe is dominated by a non-baryonic, non-luminous and non-relativistic matter contribution – commonly knownas Dark Matter. Weakly Interacting Massive Particles (WIMPs) are one of the mostpopular candidates for the dark matter particle. In this work, a detailed description ofthe fundamentals of direct detection searches is offered, which aim to detect WIMPSvia their interaction with a nucleus target in Earth-based detectors. The differentbackgrounds that will affect these experiments are described, together with the var-ious techniques employed to reject them. SuperCDMS Soudan is a direct detectionexperiment which uses Germanium semiconductor crystals detectors operating at mKtemperatures. These detectors are equipped with phonon and charge sensors, enablingexcellent rejection of electron recoil backgrounds. However, any irreducible neutronbackground from environmental radioactivity is still present in the experiment. Theestimation of this background is presented with a detailed description of the processfollowed, in the context of the search for WIMPs with masses between 10-100 GeV/c2.

 

This dissertation focuses on ionization collection in these detectors under the sub-Kelvin, low electric field, and high crystal purity conditions unique to CDMS. The design and per- formance of a fully cryogenic HEMT-based amplifier capable of achieving the SuperCDMS SNOLAB ionization energy resolution goal of 100 eVee is presented. The experimental appa- ratus which has been used to record electron and hole properties under CDMS conditions is described. Measurements of charge transport, trapping, and impact ionization as a function of electric field in two CDMS detectors are shown, and the ionization collection efficiency is determined. The data is used to predict the error in the nuclear recoil energy scale under both CDMSlite and iZIP operating modes. A two species, two state model is developed to describe how ionization collection and space charge generation in CDMS detectors are controlled by the presence of “overcharged” D− donor and A+ acceptor impurity states. The thermal stability of these states is exclusive to sub-Kelvin operation, explaining why ioniza- tion collection in CDMS detectors differs from similar semiconductor detectors operating at higher temperature. This work represents a solid foundation for the understanding ionization collection in CDMS detectors.

 

In this dissertation, I describe a novel apparatus for studying the transport of charge in semiconductors at cryogenic temperatures. The motivation to conduct this experiment originated from an asymmetry observed between the behavior of electrons and holes in the germanium detector crystals used by the Cryogenic Dark Matter Search (CDMS).

This asymmetry is a consequence of the anisotropic propagation of electrons in germanium at cryogenic temperatures. To better model our detectors, we incorporated this effect into our Monte Carlo simulations of charge transport. The purpose of the experiment described in this dissertation is to test those models in detail.

Our measurements have allowed us to discover a shortcoming in our most recent Monte Carlo simulations of electrons in germanium. This discovery would not have been possible without the measurement of the full, two-dimensional charge distribution, which our experimental apparatus has allowed for the first time at cryogenic temperatures.

 

The SuperCDMS SNOLAB experiment will use solid state Germanium and Silicon detectors to search for Weakly Interacting Massive Particles (WIMPs), a leading candidate to explain dark matter. WIMPs are thought to exist in halos around galaxies and therefore thought to be constantly streaming through the earth. The CDMS detectors have been developed to measure the energy deposited by a WIMP-nucleon collision in terrestrial calorimeters. This thesis focusses on the Data Acquisition (DAQ) system that uses Detector Control and Readout Cards (DCRCs) and is designed to be dead- time-less. The DCRCs read in the data stream from the detector’s 12 phonon and 4 ionization energy channels. The DCRCs also control detector settings, and we develop interactive codes to allow users to easily change detector settings through the DCRC. The DAQ is designed to decide which events to write to disk in order to keep data throughput under a limit yet never miss an event that will be useful in the subsequent analysis. In this effort we develop different readout methods of the detector array for the different calibration runs and WIMP search runs. We also develop fast algorithms for rejecting events that fail a certain criteria for being usable. We also present a novel data compression method that reduces the total data volume by a factor of ∼ 16 yet retains all important information. This method involves a large covariance matrix inversion, and we show that this inversion can be consistently computed given that a sufficient amount of data has been used to build the covariance matrix. We also develop a GUI that is a critical element of the detector testing program for SuperCDMS SNOLAB. The GUI accesses the data stream as it is being written to disk, efficiently reads in the waveforms, and displays them in a user-friendly, oscilloscope-like, format. By making use of Fast Fourier Transform technology, the GUI is also capable of displaying incoming data in the frequency domain. This tool will enable a new degree of real-time analysis of detector performance, specifically noise characteristics, at the test facilities in the next stage of detector testing.

 

SuperCDMS Soudan operates specialized germanium detectors (iZIPs) that are cooled to milliKelvin temperatures deep underground in the Soudan Underground Laboratory with the hope of detecting a rare collision between dark matter and a nucleus. A search for low-mass dark matter comes with multiple unique challenges since the background discrimination abilities of these detectors becomes less powerful at the low energies needed to probe low-mass dark matter since the signal to noise ratio deteriorates. Using a sophisticated background model via a pulse rescaling technique, SuperCDMS Soudan was able to produce a world leading exclusion limit on low-mass dark matter.

Effort is to extend the analysis to higher masses require long running times during which many aspects of the detectors or the environment can change. Additional challenges are offered by the powerful background discrimination ability of the iZIP. The background distributions are well separated from the signal region, meaning most of the leakage arises from low-probability tails of the background distributions. In the absence of an enormous dataset, extrapolations from the bulk of the distribution are required. While attempting to obtain a model of gamma induced electron-recoils leaking into the signal region of the detector from high radius a curious asymmetry between the sides of the detectors was discovered potentially indicating an electronics or detector design problem.

 

As the CDMS (Cryogenic Dark Matter Search) experiment is scaled up to tackle new dark matter parameter spaces (lower masses and cross-sections), detector production efficiency and repeatability becomes ever more important. A dedicated facility has been commissioned for SuperCDMS detector fabrication at Texas A&M University (TAMU). The fabrication process has been carefully tuned using this facility and its equipment. Production of successfully tested detectors has been demonstrated. Significant improvements in detector performance have been made using new fabrication methods, equipment, and tuning of process parameters. This work has demonstrated the capability for production of next generation CDMS SNOLAB detectors.

Additionally, as the dark matter parameter space is probed further, careful calibrations of detector response to nuclear recoil interactions must be performed in order to extract useful information (in relation to dark matter particle characterizations) from experimental results. A neutron beam of tunable energy is used in conjunction with a commercial radiation detector to characterize ionization energy losses in germanium during nuclear recoil events. Data indicates agreement with values predicted by the Lindhard equation, providing a best-fit k-value of 0.146.

 

Understanding the quasiparticle diffusion process inside sputtered aluminum (Al thin films ~0.1-1 μm) is critical for the Cryogenic Dark Matter Search (CDMS experiment to further optimize its detectors to directly search for dark matter. An initial study with Al films was undertaken by our group ~20 years ago, but some important questions were not answered at the time. This thesis can be considered a continuation of that critical study.

 

Analyzing SuperCDMS Soudan data to look for low-mass dark matter comes with particular challenges because of the low signal-to-noise very near threshold. However, with a detailed background model developed by scaling high-energy events down into the low-energy signal region, SuperCDMS Soudan produced world- leading limits on the existence of low-mass dark matter.

In addition, a few SuperCDMS Soudan detectors experienced cold hardware problems that can affect the data collected. Of particular interest is one detector considered for the low-mass WIMP search that has one of its charge electrodes shorted to chassis ground. Three events were observed in this detector upon unblinding the SuperCDMS Soudan low-energy data, even though <1 event was expected based on pre-unblinding calulations. However, the data collected by the shorted detector may have been compromised since an electrode shorted to ground will modify the electric field in the detector. The SuperCDMS Detector Monte Carlo (DMC) provides an excellent way to model the effects of the modified electric field, so a new model of the expected backgrounds in the low-mass WIMP search is developed using the DMC to try to explain how the short may have affected the data collected.

 

This thesis describes the results of such a search with the SuperCDMS experiment, which uses Ge detectors cooled to 50 mK to detect ionization and phonons produced by particle interactions. We perform a blind analysis of 577 kg d of exposure on 7 detectors targeting WIMPs with masses < 30GeV/c², where anomalous results have been reported by previous experiments. No significant excess is observed and we set an upper limit on the spin-independent WIMP-nucleon cross section of 1.2×10^-42 cm² at 8 GeV/c². We also set constraints on dark matter interactions independent of the dark matter halo physics, as well as on annual modulation of a dark matter signal.

Cryogenic detectors similar to SuperCDMS also have potential applications in neutrino physics. We study several configurations in which dark matter detectors could be used with an intense neutrino source to detect an unmeasured Standard Model process called coherent neutrino scattering. This process may be useful, for example, as a calibration for next-generation dark matter detectors, and for constraining eV-scale sterile neutrinos. In addition, small cryogenic X-ray detectors on sounding rockets with large fields-of-view have the unique ability to constrain sterile neutrino dark matter. We set limits on sterile neutrino dark matter using an observation by the XQC instrument, and discuss prospects for a future observation of the galactic center using the Micro-X instrument.

 

The Cryogenic Dark Matter Search (CDMS) experiment is designed to directly detect elastic scatters of weakly-interacting massive dark matter particles (WIMPs), on target nuclei in semiconductor crystals composed of Si and Ge. These scatters would occur very rarely, in an overwhelming background composed primarily of electron recoils from photons and electrons, as well as a smaller but non-negligible background of WIMP-like nuclear recoils from neutrons. ...

Nuclear recoils have suppressed ionization signals relative to electron recoils of the same recoil energy, so the response of the detectors is calibrated differently for each recoil type. ...

I discuss systematic uncertainties affecting the reconstruction of this recoil energy, the primary analysis variable, and use several methods to constrain their magnitude. I present the resulting adjusted WIMP limits and discuss their impact in the context of current and projected constraints on the parameter space for WIMP interactions.

 

"CDMSlite" stands for CDMS - low ionization threshold experiment. Here we utilize a unique electron phonon coupling mechanism to measure ionization generated by scattering of light particles. Typically signals from such low energy recoils would be washed under instrumental noise. In CDMSlite via generation of Luke-Neganov phonons we can detect the small ionization energies, amplified in phonon modes during charge transport. This technology allows us to reach very low thresholds and reliably measure and investigate low energy recoils from light Dark Matter particles. This thesis describes the physics behind CDMSlite, the experimental design and the first science results from CDMSlite operated at the Soudan Underground Laboratory.

 

The CDMS-II phase of the Cryogenic Dark Matter Search, a dark matter direct-detection experiment, was operated at the Soudan Underground Laboratory from 2003 to 2008. The full payload consisted of 30 ZIP detectors, totaling approximately 1.1 kg of Si and 4.8 kg of Ge, operated at temperatures of ~50 mK. ...

A full re-analysis of the CDMS-II data was motivated by an improvement in the event reconstruction algorithms which improved the resolution of ionization energy and timing information. The Ge data were re-analyzed using three distinct background-rejection techniques; the Si data from runs 125 - 128 were analyzed for the first time using the most successful of the techniques from the Ge re-analysis. The results of these analyses prompted a novel “mid-threshold” analysis, wherein energy thresholds were lowered but background rejection using phonon timing information was still maintained. This technique proved to have significant discrimination power, maintaining adequate signal acceptance and minimizing background leakage.

 

This thesis documents the test facility and the work involved in its development. In the test facility, we performed the first ionization collection efficiency measurements of the ionization test devices. The test devices are fabricated with detector-grade germanium crystals that are 100 mm in diameter, which is the largest available, and 33 mm in thickness. The measured efficiencies are consistent with the earlier measurements conducted with smaller Ge crystals, demonstrating that these 100 mm crystals can be used for development of the next generation dark matter detectors.

The data taken during the last four runs of CDMS II with total raw exposure 612 kg-day were reprocessed with improved ionization pulse reconstruction algorithm. We present the classic timing analysis with the reprocessed data in this thesis. For the four runs combined, this analysis resulted in a new WIMP-nucleon cross section 4.4×10^-44 cm² for a WIMP mass of 70 GeV/c², which is a factor of 1.6 improvement compared to the original c58 classic timing analysis.

 

The Cryogenic Dark Matter Search (SuperCDMS) relies on collection of phonons and charge carriers in semiconductors held at tens of milliKelvin as handles for detection of Weakly Interacting Massive Particles (WIMPs). This thesis begins with a brief overview of the direct dark matter search (Chapter 1) and SuperCDMS detectors (Chapter 2). In Chapter 3, a 3He evaporative refrigerator facility is described. Results from experiments performed in-house at Stanford to measure carrier transport in high-purity germanium (HPGe) crystals operated at sub-Kelvin temperatures are presented in Chapter 4. Finally, in Chapter 5 a new numerical model and a time-domain optimal filtering technique are presented, both developed for use with superconducting Transition Edge Sensors (TESs), that provide excellent event reconstruction for single particle interactions in detectors read out with superconducting W-TESs coupled to energy-collecting films of Al.

 

Fundamental particles are always observed to carry charges which are integral multiples of one-third charge of electron, e/3. While this is a well established experimental fact, the theoretical understanding for the charge quantization phenomenon is lacking. On the other hand, there exist numerous theoretical models that naturally allow for existence of particles with fractional electromagnetic charge. These particles, if existing, hint towards existence of physics beyond the standard model. Multiple high energy, optical, cosmological and astrophysical considerations restrict the allowable mass-charge parameter space for these fractional charges. Still, a huge unexplored region remains.

The Cryogenic Dark Matter Search (CDMS-II), located at Soudan mines in northern Minnesota, employs germanium and silicon crystals to perform direct searches for a leading candidate to dark matter called Weakly Interacting Massive Particles (WIMPs). Alternately, the low detection threshold allows search for fractional electromagnetic-charged particles, or Lightly Ionizing Particles (LIPs), moving at relativistic speed. Background rejection is obtained by requiring that the magnitude and location of energy deposited in each detector be consistent with corresponding "signatures" resulting from the passage of a fractionally charged particle. In this dissertation, the CDMS-II data is analyzed to search for LIPs, with an expected background of 0.078±0.078 events. No candidate events are observed, allowing exclusion of new parameter space for charges between e/6 and e/200. ...

SuperCDMS (Super Cryogenic Dark Matter Search) is a leading direct dark matter search experiment which uses solid state detectors (Ge crystals) at milliKelvin temperatures to look for nuclear recoils caused by dark matter interactions in the detector. `Weakly Interacting Massive Particles' (WIMPs) are the most favoured dark matter candidate particles. SuperCDMS, like many other direct dark matter search experiments, primarily looks for WIMPs. The measurement of both the ionization and the lattice vibration (phonon) signals from an interaction in the detector allow it to discriminate against electron recoils which are the main source of background for WIMP detection.

SuperCDMS currently operates about 9 kg of Ge detectors at the Soudan underground lab in northern Minnesota. In its next phase, SuperCDMS SNOLAB plans to use 100-200 kg of target mass (Ge) which would allow it to probe more of the interesting and as of yet unexplored parameter space for WIMPs predicted by theoretical models. The SuperCDMS Queen's Test Facility is a detector test facility which is intended to serve for detector testing and detector research and development purposes for the SuperCDMS experiment.
A modied detector called the `HiZIP' (Half-iZIP), which is reduced in complexity in comparison to the currently used iZIP (interleaved Z-sensitive Ionization and Phonon mediated) detectors, is studied in this thesis. ...

 

Astrophysical and cosmological measurements on the scales of galaxies, galaxy clusters, and the universe indicate that ~85% of the matter in the universe is composed of dark matter, made up of non-baryonic particles that interact with cross-sections on the weak scale or lower. Hypothetical Weakly Interacting Massive Particles, or WIMPs, represent a potential solution to the dark matter problem, and naturally arise in certain Standard Model extensions.

The Cryogenic Dark Matter Search (CDMS) collaboration aims to detect the scattering of WIMP particles from nuclei in terrestrial detectors. Germanium and silicon particle detectors are deployed in the Soudan Underground Laboratory in Minnesota. These detectors are instrumented with phonon and ionization sensors, which allows for discrimination against electromagnetic backgrounds, which strike the detector at rates orders of magnitude higher than the expected WIMP signal.

This dissertation presents the development of numerical models of the physics of the CDMS detectors, implemented in a computational package collectively known as the CDMS Detector Monte Carlo (DMC). After substantial validation of the models against data, the DMC is used to investigate potential backgrounds to the next iteration of the CDMS experiment, known as SuperCDMS. Finally, an investigation of using the DMC in a reverse Monte Carlo analysis of WIMP search data is presented. ...

 

This dissertation describes the results of a WIMP search using CDMS II data sets accumulated at the Soudan Underground Laboratory in Minnesota. Results from the original analysis of these data were published in 2009; two events were observed in the signal region with an expected leakage of 0.9 events. Further investigation revealed an issue with the ionization-pulse reconstruction algorithm leading to a software upgrade and a subsequent reanalysis of the data. As part of the reanalysis, I performed an advanced discrimination technique to better distinguish (potential) signal events from backgrounds using a 5-dimensional chi-square method. This data analysis technique combines the event information recorded for each WIMP-search event to derive a background-discrimination parameter capable of reducing the expected background to less than one event, while maintaining high efficiency for signal events. Furthermore, optimizing the cut positions of this 5-dimensional chi-square parameter for the 14 viable germanium detectors yields an improved expected sensitivity to WIMP interactions relative to previous CDMS results. This dissertation describes my improved (and optimized) discrimination technique and the results obtained from a blind application to the reanalyzed CDMS II WIMP-search data.

This analysis achieved the best expected sensitivity of the three techniques developed for the reanalysis and so was chosen as the primary timing analysis whose limit will be quoted in a on-going publication paper which is currently in preparation. For this analysis, a total raw exposure of 612.17 kg-days are analyzed for this work. No candidate events was observed, and a corresponding upper limit on the WIMP-nucleon scattering cross section as a function of WIMP mass is defined. These data set a 90% upper limit on spin-independent WIMP-nucleon elastic-scattering cross section of 3.19×10^-44 cm² for a WIMP mass of 60 GeV/c². Combining this result with all previous CDMS II data gives an upper limit of 1.96×10^-44 cm² for a WIMP of mass 60 GeV/c² (a factor of 2 better than the original analysis). ...

 

An overwhelming proportion of the universe (83% by mass) is composed of particles we know next to nothing about. Detecting these dark matter particles directly, through hypothesized weak-force-mediated recoils with nuclear targets here on earth, could shed light on what these particles are, how they relate to the standard model, and how the standard model ts within a more fundamental understanding.

This thesis describes two such experimental efforts: CDMS II (2007-2009) and SuperCDMS Soudan (ongoing). The general abilities and sensitivities of both experiments are laid out, placing a special emphasis on the detector technology, and how this technology has evolved from the first to the second experiment. Some topics on which I spent significant efforts are described here only in overview (in particular the details of the CDMS II analysis, which has been laid out many times before), and some topics which are not described elsewhere are given a somewhat deeper treatment.

In particular, this thesis is hopefully a good reference for those interested in the annual modulation limits placed on the low-energy portion of the CDMS II exposure, the design of the detectors for SuperCDMS Soudan, and an overview of the extremely informative data these detectors produce. It is an exciting time. The technology I've had the honor to work on the past few years provides a wealth of information about each event, more so than any other direct detection experiment, and we are still learning how to optimally use all this information. Initial tests from the surface and now underground suggest this technology has the background rejection abilities necessary for a planned 200kg experiment or even ton-scale experiment, putting us on the threshold of probing parameter space orders of magnitude from where the field currently stands.

 

A wide variety of astrophysical observations indicate that approximately 85% of the matter in the universe is nonbaryonic and nonluminous. Understanding the nature of this "dark matter" is one of the most important outstanding questions in cosmology. Weakly Interacting Massive Particles (WIMPs) are a leading candidate for dark matter since they would be thermally produced in the early universe in the correct abundance to account for the observed relic density of dark matter. If WIMPs account for the dark matter, then rare interactions from relic WIMPs should be observable in terrestrial detectors. Recently, unexplained excess events in the DAMA/LIBRA, CoGeNT, and CRESST-II experiments have been interpreted as evidence of scattering from WIMPs with masses ~10 GeV and spin-independent scattering cross sections of 10^-41-10^-40 cm².

The Cryogenic Dark Matter Search (CDMS II) attempts to identify WIMP interactions using an array of cryogenic germanium and silicon particle detectors located at the Soudan Underground Laboratory in northern Minnesota. In this dissertation, data taken by CDMS II are reanalyzed using a 2 keV recoil energy threshold to increase the sensitivity to WIMPs with masses ∼10 GeV. These data disfavor an explanation for the DAMA/LIBRA, CoGeNT, and CRESST-II results in terms of spin-independent elastic scattering of WIMPs with masses ~12 GeV, under standard assumptions. At the time of publication, they provided the strongest constraints on spin-independent elastic scattering from 5-9 GeV, ruling out previously unexplored parameter space. ...

 

For the past 15 years, the Cryogenic Dark Matter Search or CDMS has searched for Weekly Interacting Massive Particle dark matter (WIMPs) using Ge and Si semiconductor crystals instrumented with both ionization and athermal phonon sensors so that the much more common electron recoil leakage caused by photons and Βs from naturally present radioactive elements can be easily distinguished from elastic WIMP nucleon interactions by looking at the fraction of total recoil energy which ends up as potential energy of e^-/h⁺ pairs.

Due to electronic carrier trapping at the surface of our semiconductor crystals, electron recoils which occur near the surface have suppressed ionization measurements and can not be distinguished from WIMP induced nuclear recoils and thus sensitivity to the WIMP nucleon interaction cross section was driven in CDMS II by our ability to define a full 3D fiducial volume in which all events had full collection. To remain background free and maximally sensitive to the WIMP-nucleus interaction cross section, we must improve our 3D fiducial volume definition at the same rate as we scale the mass of the detector, and thus proposed next generation experiments with an order of magnitude increase in active mass were unfortunately not possible with our previous CDMS II detector design, and a new design with significantly improved fiducialization performance is required.

 

The Cryogenic Dark Matter Search (CDMS) is searching for weakly-interacting massive particles (WIMPS), which could explain the dark matter problem in cosmology and particle physics.

By simultaneously measuring signals from deposited charge and the energy in nonequilibrium phonons created by particle interactions in intrinsic germanium crystals at a temperature of 40 mK, a signature response for each event is produced. This response, combined with phonon pulse-shape information, allows CDMS to actively discriminate candidate WIMP interactions with nuclei from electromagnetic radioactive background which interacts with electrons.

 

Substantial evidence from galaxies, galaxy clusters, and cosmological scales suggests that ~ 85% of the matter of our universe is invisible. The missing matter, or "dark matter", is likely composed of non-relativistic, non-baryonic particles, which have very rare interactions with baryonic matter and with one another. Among dark matter candidates, Weakly Interacting Massive Particles (WIMPs) are particularly well motivated. In the early universe, thermally produced particles with weak-scale mass and interactions would `freeze out at the correct density to be dark matter today. Extensions to the Standard Model of particle physics, such as Supersymmetry, which solve gauge hierarchy and coupling unification problems, naturally provide such particles.

Interactions of WIMPs with baryons are expected to be rare, but might be detectable in low-noise detectors. The Cryogenic Dark Matter Search (CDMS) experiment uses ionizationand phononsensitive germanium particle detectors to search for such interactions. CDMS detectors are operated at the Soudan Underground Laboratory in Minnesota, within a shielded environment to lower cosmogenic and radioactive background. The combination of phonon and ionization signatures from the detectors provides excellent residual-background rejection.

 

Although dark matter appears to constitute over 80% of the matter in the Universe, its composition is a mystery. Astrophysical observations suggest that the luminous portions of the Galaxy are embedded in a halo of darkmatter particles. Weakly Interacting Massive Particles (WIMPs) are the most studied class of dark-matter candidates and arise naturally within the context of many weak-scale supersymmetric theories. Direct-detection experiments like the Cryogenic Dark Matter Search (CDMS) strive to discern the kinetic energy of recoiling nuclei resulting from WIMP interactions with terrestrial matter. This is a considerable challenge in which the low (expected) rate of WIMP interactions must be distinguished from an overwhelming rate due to known types of radiation.

An incontrovertible positive detection has remained elusive. However, a few experiments have recorded data that appear consistent with a low-mass WIMP. This thesis describes an attempt to probe the favored parameter space. To increase sensitivity to low-mass WIMPs, a low-threshold technique with improved sensitivity to small energy depositions is applied to CDMS shallow site data. Four germanium and two silicon detectors were operated between December 2001 and June 2002, yielding 118 days of exposure. By sacrificing some of the CDMS detectors ability to discriminate signal from background, energy thresholds of ~1 and ~2 keV were achieved for three of the germanium and both silicon detectors, respectively. A large number of WIMP candidate events are observed, most of which can be accounted for by misidentification of background sources. No conclusive evidence for a low-mass WIMP signal is found. The observed event rates are used to set upper limits on the WIMP-nucleon scattering cross section as a function of WIMP mass. Interesting parameter space is excluded for WIMPs with masses below ~9GeV/c². Under standard assumptions, the parameter space favored by interpretations of other experiments data as low-mass WIMP signals is partially excluded, and new parameter space is excluded for WIMP masses between 3 and 4 GeV/c².

 

The Cryogenic Dark Matter Search (CDMS) is designed to detect Weakly-Interacting Massive Particles (WIMPs) in the Milky Way halo. The phase known as CDMS II was performed in the Soudan Underground Laboratory. The final set of CDMS II data, collected in 2007-8 and referred to as Runs 125-8, represents the largest exposure to date for the experiment.

We seek collisions between WIMPs and atomic nuclei in disk-shaped germanium and silicon detectors. A key design feature is to keep the rate of collisions from known particles producing WIMP-like signals very small. The largest category of such background is interactions with electrons in the detectors that occur very close to one of the faces of the detector. The next largest category is collisions between energetic neutrons that bypass the experimental shielding and nuclei in the detectors. Analytical efforts to discriminate these backgrounds and to estimate the rate at which such discrimination fails have been refined and improved throughout each phase of CDMS.

Next-generation detectors for future phases of CDMS require testing at cryogenic test facilities. One such facility was developed at the University of Minnesota in 2007 and has been used continuously since then to test detectors for the next phase of the experiment, known as SuperCDMS.

 

Cosmological observations in the last decades have led to a concordance model of the universe, where ~85% of matter is non-baryonic, non-luminous and non-relativistic at the time of structure formation. Theories of physics beyond the Standard Model of particle physics propose a wide array of candidates for the nature of this unseen dark matter. Weakly interacting massive particles (WIMPs) are a class of candidates which is well motivated by thermal production models for dark matter in the early universe. WIMPs (or any other dark matter candidate), distributed in a spherical isothermal halo surrounding our galaxy (the standard halo model; SHM), could be detected in terrestrial detectors.

In the standard model of disc galaxy formation, a dark matter disc forms as massive satellites are preferentially dragged into the disc-plane and dissolve. The low velocity of the dark matter particles in the dark disc with respect to the Earth enhances detection rates at low recoil energy in direct detection experiments. For WIMP masses &50GeV/c2, the detection rates increase by up to a factor of 3 in the 5-20 keV recoil energy range. Comparing this with rates at higher energies may be sensitive to the WIMP mass, providing stronger mass constraints particularly for masses ~100GeV/c². The annual modulation signal is significantly boosted and the modulation phase is shifted by ~3 weeks relative to the dark halo. The variation of the observed phase with recoil energy determines the particle's mass, once the dark disc properties are fixed by future astronomical surveys. ...

 

The Cryogenic Dark Matter Search (CDMS) is searching for Weakly Interacting Massive Particles (WIMPs) with cryogenic particle detectors. These detectors have the ability to discriminate between nuclear recoil candidate and electron recoil background events by collecting both phonon and ionization energy from recoils in the detector crystals. The CDMS-II experiment has completed analysis of the first data runs with 30 semiconductor detectors at the Soudan Underground Laboratory, resulting in a world leading WIMP-nucleon spin-independent cross section limit for WIMP masses above 44 GeV/c2.

As CDMS aims to achieve greater WIMP sensitivity, it is necessary to increase the detector mass and discrimination between signal and background events. Incomplete ionization collection results in the largest background in the CDMS detectors as this causes electron recoil background interactions to appear as false candidate events. Two primary causes of incomplete ionization collection are surface and bulk trapping.

Recent work has been focused on reducing surface trapping through the modification of fabrication methods for future detectors. Analyzing data taken with test devices has shown that hydrogen passivation of the amorphous silicon blocking layer worsens surface trapping. Additional data has shown that the iron-ion implantation used to lower the critical temperature of the tungsten transition-edge sensors causes a degradation of the ionization collection. Using selective implantation on future detectors may improve ionization collection for events near the phonon side detector surface. ...

 

In recent decades astronomers and physicists have accumulated a vast array of evidence that the bulk of the universe's matter is in some non-baryonic form that remains undetected by electromagnetic means. This "dark matter" resides in diffuse halos surrounding galaxies and other cosmic structures. Particle theorists have proposed a wide array of candidates for its nature. One particularly promising class of candidates are Weakly Interacting Massive Particles (WIMPs): quanta with masses of order 100 GeV/c2 and interactions characteristic of the weak nuclear force.

The Cryogenic Dark Matter Search (CDMS) experiment seeks to directly detect the rare elastic interactions of galactic WIMPs with terrestrial nuclei. To this end, CDMS operates an array of crystalline Ge and Si particle detectors in Soudan Underground Laboratory in northern Minnesota. These crystals are operated at millikelvin temperatures and instrumented to measure the ionization and athermal phonons generated by each particle interaction. This combination provides a powerful two-fold discrimination against the interactions of particles generated by radioactive decay and cosmogenic showers.

This dissertation describes the commissioning, analysis, and results of the first WIMP-search data runs of the CDMS experiment with its full complement of 5 "Towers" of detectors. These data represent a substantial increase in target mass and exposure over previous CDMS results. The results of this work place the most stringent limits yet set upon the WIMP-nucleon spin-independent cross section for WIMP masses above ~ 44 GeV/c2, as well as setting competitive limits on spin-dependent WIMP-nucleon interactions. This work also outlines the larger context of this and other probes of the WIMP theory of dark matter, as well as some current development efforts toward a larger cryogenic experiment.

 

Non-baryonic dark matter makes one quarter of the energy density of the Universe and is concentrated in the halos of galaxies, including the Milky Way. The Weakly Interacting Massive Particle (WIMP) is a dark matter candidate with a scattering cross section with an atomic nucleus of the order of the weak interaction and a mass comparable to that of an atomic nucleus. The Cryogenic Dark Matter Search (CDMS-II) experiment, using Ge and Si cryogenic particle detectors at the Soudan Underground Laboratory, aims to directly detect nuclear recoils from WIMP interactions.

This thesis presents the first 5 tower WIMP-search results from CDMS-II, an estimate of the cosmogenic neutron backgrounds expected at the Soudan Underground Laboratory, and a proposal for a new measurement of high-energy neutrons underground to benchmark the Monte Carlo simulations.

Based on the non-observation of WIMPs and using standard assumptions about the galactic halo [68], the 90% C.L. upper limit of the spin-independent WIMPnucleon cross section for the first 5 tower run is 6.6 x 10^-44cm2 for a 60 GeV/c2 WIMP mass.

 

Images of the Bullet Cluster of galaxies in visible light, X-rays, and through gravitational lensing confirm that most of the matter in the universe is not composed of any known form of matter. The combined evidence from the dynamics of galaxies and clusters of galaxies, the cosmic microwave background, big bang nucleosynthesis, and other observations indicates that 80% of the universe's matter is dark, nearly collisionless, and cold. The identity of the dark matter remains unknown, but weakly interacting massive particles (WIMPs) are a very good candidate. They are a natural part of many supersymmetric extensions to the standard model, and could be produced as a nonrelativistic, thermal relic in the early universe with about the right density to account for the missing mass. The dark matter of a galaxy should exist as a spherical or ellipsoidal cloud, called a "halo" because it extends well past the edge of the visible galaxy.

The Cryogenic Dark Matter Search (CDMS) seeks to directly detect interactions between WIMPs in the Milky Way's galactic dark matter halo using crystals of germanium and silicon. Our Z-sensitive ionization and phonon ("ZIP") detectors simultaneously measure both phonons and ionization produced by particle interactions. In order to find very rare, low-energy WIMP interactions, we must identify and reject background events caused by environmental radioactivity, radioactive contaminants on the detectors, and cosmic rays. In particular, sophisticated analysis of the timing of phonon signals is needed to eliminate signals caused by beta decays at the detector surfaces. ...

 

Most of the mass-energy density of the universe remains undetected and is only understood through its affects on visible, baryonic matter. The visible, baryonic matter accounts for only about half of a percent of the universe's total mass-energy budget, while the remainder of the mass-energy of the universe remains dark or undetected. About a quarter of the dark mass-energy density of the universe is comprised of massive particles that do not interact via the strong or electromagnetic forces. If these particles interact via the weak force, they are termed weakly interacting massive particles or WIMPs, and their interactions with baryonic matter could be detectable.

The CDMS II experiment attempts to detect WIMP interactions in the Soudan Underground Laboratory using germanium detectors and silicon detectors. A WIMP can interact a with detector nuclei causing the nuclei to recoil. A nuclear recoil is distinguished from background electron recoils by comparing the deposited ionization and phonon energies. Electron recoils occurring near detector surfaces are more difficult to reject.

This thesis describes the results of a x^2 analysis designed to reject events occurring near detector surfaces. Because no WIMP signal was observed, separate limits using the germanium and silicon detectors are set on the WIMP cross section under standard astrophysical assumptions.

 

The Cryogenic Dark Matter Search (CDMS) experiment is designed to search for dark matter in the form of Weakly Interacting Massive Particles (WIMPs) via their elastic scattering interactions with nuclei. This dissertation presents the CDMS detector technology and the commissioning of two towers of detectors at the deep underground site in Soudan, Minnesota. CDMS detectors comprise crystals of Ge and Si at temperatures of 20 mK which provide ~keV energy resolution and the ability to perform particle identification on an event by event basis. Event identification is performed via a two-fold interaction signature; an ionization response and an athermal phonon response. Phonons and charged particles result in electron recoils in the crystal, while neutrons and WIMPs result in nuclear recoils. Since the ionization response is quenched by a factor ~ 3(2) in Ge(Si) for nuclear recoils compared to electron recoils, the relative amplitude of the two detector responses allows discrimination between recoil types. The primary source of background events in CDMS arises from electron recoils in the outer 50 µm of the detector surface which have a reduced ionization response. We develop a quantitative model of this dead layer effect and successfully apply the model to Monte Carlo simulation of CDMS calibration data. Analysis of data from the two tower run March-August 2004 is performed, resulting in the world's most sensitive limits on the spin-independent WIMP-nucleon cross-section, with a 90% C.L. upper limit of 1.6 x 10^-43 cm2 on Ge for a 60 GeV WIMP. An approach to performing surface event discrimination using neural networks and wavelets is developed. A Bayesian methodology to classifying surface events using neural networks is found to provide an optimized method based on minimization of the expected dark matter limit. The discrete wavelet analysis of CDMS phonon pulses improves surface event discrimination in conjunction with the neural network analysis, giving a 20% improvement to the expected and final WIMP-nucleon cross-section upper limits.

 

In this thesis I discuss the development of a novel phonon-mediated distributed transition-edge-sensor X-ray detector which would be useful for astrophysical studies such as magnetic recombination in the solar corona, the warm-hot intergalactic medium and surveys of clusters and groups of galaxies.

The detector uses a large semiconductor absorber and Transition-Edge-Sensors (TESs) to readout the absorbed energy. Calorimetry is performed on individual photons and a partitioning of the energy between various TESs allows for position determination. Hence time varying astronomical sources can be spectroscopically studied and imaged.

 

The Cryogenic Dark Matter Search has completed two runs at the Soudan Underground Laboratory In the second, two towers of detectors were operated from March to August 2004. CDMS used Ge and Si ZIP (Z-sensitive, Ionization, and Phonon) detectors, operated at 50mK, to look for Weakly Interacting Massive Particles (WIMPs) which may make up most of the dark matter in our universe. These detectors are surrounded by lead and polyethylene shielding as well as an active muon veto. These shields, as well as the overburden of Soudan rock, provide a low background environment for the detectors.

The ZIP detectors record the ratio of ionization signal to phonon signal to discriminate between nuclear recoils, characteristic of WIMPs and neutrons, and electron recoils, characteristic of gamma and beta backgrounds. They also provide timing information from the four phonon channels that is used to reject surface events, for which ionization collection is poor. A blind analysis, defined using calibration data taken in situ throughout the run, provides a definition of the WIMP signal region by rejecting backgrounds. This analysis applied to the WIMP search data gives a limit on the spin independent WIMP-nucleon cross-section that is an order of magnitude lower than any other experiment has published.

 

Evidence from observational cosmology and astrophysics indicates that about one third of the universe is matter, but that the known baryonic matter only contributes to the universe at 4%. A large fraction of the universe is cold and non-baryonic matter, which has important role in the universe structure formation and its evolution. The leading candidate for the non-baryonic dark matter is Weakly Interacting Massive Particles (WIMPs), which naturally occurs in the supersymmetry theory in particle physics.

The Cryogenic Dark Matter Search (CDMS) experiment is searching for evidence of a WIMP interaction off an atomic nucleus in crystals of Ge and Si by measuring simultaneously the phonon energy and ionization energy of the interaction in the CDMS detectors. The WIMP interaction energy is from a few keV to tens of keV with a rate less than 0.1 events/kg/day. To reach the goal of WIMP detection, the CDMS experiment has been conducted in the Soudan mine with an active muon veto and multistage passive background shields.

The CDMS detectors have a low energy threshold and background rejection capabilities based on ionization yield. However, betas from contamination and other radioactive sources produce surface interactions, which have low ionization yield, comparable to that of bulk nuclear interactions. The low-ionization surface electron recoils must be removed in the WIMP search data analysis. An emphasis of this thesis is on developing the method of the surface-interaction rejection using location information of the interactions, phonon energy distributions and phonon timing parameters. The results of the CDMS Soudan run118 92.3 live day WIMP search data analysis is presented, and represents the most sensitive search yet performed.

 

The Cryogenic Dark Matter Search (CDMS) uses position-sensitive Germanium and Silicon crystals in the direct detection of Weakly Interacting Massive Particles (WIMPs) believed to constitute most of the dark matter in the Universe. WIMP interactions with matter being rare, identifying and eliminating known backgrounds is critical for detection. Event-by-event discrimination by the detectors rejects the predominant gamma and beta backgrounds while Monte Carlo simulations help estimate, and subtract, the contribution from the neutrons.

This thesis describes the effort to understand neutron backgrounds as seen in the two stages of the CDMS search for WIMPs. The first stage of the experiment was at a shallow site at the Stanford Underground Facility where the limiting background came from high-energy neutrons produced by cosmic-ray muon interactions in the rock surrounding the cavern.

Simulations of this background helped inform the analysis of data from an experimental run at this site and served as input for the background reduction techniques necessary to set new exclusion limits on the WIMP-nucleon cross-section, excluding new parameter space for WIMPs of masses 8-20 GeV/c2.

 

There is an abundance of evidence that the majority of the mass of the universe is in the form of non-baryonic non-luminous matter that was non-relativistic at the time when matter began to dominate the energy density. Weakly Interacting Massive Particles, or WIMPs, are attractive cold dark matter candidates because they would have a relic abundance today of ~0.1 which is consistent with precision cosmological measurements. WIMPs are also well motivated theoretically. Many minimal supersymmetric extensions of the Standard Model have WIMPs in the form of the lightest supersymmetric partner, typically taken to be the neutralino.

The CDMS II experiment searches for WIMPs via their elastic scattering off of nuclei. The experiment uses Ge and Si ZIP detectors, operated at <50 mK, which simultaneously measure the ionization and athermal phonons produced by the scattering of an external particle. The dominant background for the experiment comes from electromagnetic interactions taking place very close to the detector surface. Analysis of the phonon signal from these interactions makes it possible to discriminate them from interactions caused by WIMPs. This thesis presents the details of an important aspect of the phonon pulse shape analysis known as the Lookup Table Correction. The Lookup Table Correction is a position dependent calibration of the ZIP phonon response which improves the rejection of events scattering near the detector surface. ...

 

The Cryogenic Dark Matter Search (CDMS) experiment is designed to search for dark matter in the form of the Weakly Interacting Massive Particles (WIMPs). For this purpose, CDMS uses detectors based on crystals of Ge and Si, operated at the temperature of 20 mK, and providing a two-fold signature of an interaction: the ionization and the athermal phonon signals. The two signals, along with the passive and active shielding of the experimental setup, and with the underground experimental sites, allow very effective suppression and rejection of different types of backgrounds.

This dissertation presents the commissioning and the results of the first WIMP- search run performed by the CDMS collaboration at the deep underground site at the Soudan mine in Minnesota. We develop different methods of suppressing the dominant background due to the electron-recoil events taking place at the detector surface and we apply these algorithms to the data set. These results place the world's most sensitive limits on the WIMP-nucleon spin-independent elastic-scattering cross-section. Finally, we examine the compatibility of the supersymmetric WIMP-models with the direct-detection experiments (such as CDMS) and discuss the implications of the new CDMS result on these models.

 

The Cryogenic Dark Matter Search (CDMS) uses cryogenically-cooled detectors made of germanium and silicon in an attempt to detect dark matter in the form of Weakly- Interacting Massive Particles (WIMPs). The expected interaction rate of these particles is on the order of 1/kg/day, far below the 200/kg/day expected rate of background interactions after passive shielding and an active cosmic ray muon veto. Our detectors are instrumented to make a simultaneous measurement of both the ionization energy and thermal energy deposited by the interaction of a particle with the crystal substrate. A comparison of these two quantities allows for the rejection of a background of electromagnetically-interacting particles at a level of better than 99.9%. The dominant remaining background at a depth of ~ 11 m below the surface comes from fast neutrons produced by cosmic ray muons interacting in the rock surrounding the experiment.

Contamination of our detectors by a beta emitter can add an unknown source of unrejected background. In the energy range of interest for a WIMP study, electrons will have a short penetration depth and preferentially interact near the surface. Some of the ionization signal can be lost to the charge contacts there and a decreased ionization signal relative to the thermal signal will cause a background event which interacts at the surface to be misidentified as a signal event. We can use information about the shape of the thermal signal pulse to discriminate against these surface events. Using a subset of our calibration set which contains a large fraction of electron events, we can characterize the expected behavior of surface events and construct a cut to remove them from our candidate signal events. ...

 

From individual galaxies, to clusters of galaxies, to in between the cushions of your sofa, Dark Matter appears to be pervasive on every scale. With increasing accuracy, recent astrophysical measurements, from a variety of experiments, are arriving at the following cosmological model : a flat cosmology ( omega k = 0) with matter and energy densities contributing roughly 1/3 and 2/3 ( omega m = 0.35, omega lambda = 0.65). Of the matter contribution, it appears that only ~10% ( omega b ~ 0.04) is attributable to baryons. Astrophysical measurements constrain the remaining matter to be non-realtivistic, interacting primarily gravitationally. Various theoretical models for such Dark Matter exist. A leading candidate for the non-baryonic matter are Weakly Interacting Massive Particles (dubbed WIMPS). These particles, and their relic density may be naturally explained within the framework of Super-Symmetry theories. Super- Symmetry also offers predictions as to the scattering rates of WIMPs with baryonic matter allowing for the design and tailoring of experiments that search specifically for the WIMPs. The Cryogenic Dark Matter Search experiment is searching for evidence of WIMP interactions in crystals of Ge and Si. Using cryogenic detector technology to measure both the phonon and ionization response to a particle recoil the CDMS detectors are able to discriminate between electron and nuclear recoils, thus reducing the large rates of electron recoil backgrounds to levels with which a Dark Matter search is not only feasible, but far-reaching. This thesis will describe in some detail the physical principles behind the CDMS detector technology, highlighting the final step in the evolution of the detector design and characterization techniques. In addition, data from a 100 day long exposure of the current run at the Stanford Underground Facility will be presented, with focus given to detector performance as well as to the implications on allowable WIMP mass - cross-section parameter space.

 

A convincing body of evidence from observational and theoretical astrophysics suggests that matter in the universe is dominated by a non-luminous, non-baryonic, non-relativistic component. Weakly Interacting Massive Particles (WIMPs) are a proposed particle candidate that satisfy all of the above criteria. They are a front-runner among dark matter candidates because their predicted contribution to matter in the universe is cosmologically significant and because they may arise naturally from supersymmetric (SUSY) models of particle physics. The Cryogenic Dark Matter Search (CDMS) employs advanced detectors sensitive to nuclear recoils caused by WIMP scatters and capable of rejecting ionizing backgrounds.

The first phase of the experiment, conducted at a shallow site, is limited by a background of neutrons which are indistinguishable from WIMPs in terms of the acquired data. By accounting for and statistically subtracting these neutrons, CDMS I provides the best dark matter limits to date over a wide range of WIMP masses above 10 GeV/c2. These results also exclude the signal region claimed by the DAMA annual modulation search at a >71% confidence level.

The second phase of the experiment, located at a deep site, is scheduled to begin data acquisition in 2002. Due to longer exposures, larger detector mass, and low background rates at this site, data from CDMS II are expected to improve on present WIMP sensitivity by about two orders of magnitude....

 

Extensive evidence indicates that a large fraction of the matter in the universe is nonluminous, nonbaryonic, and cold: nonrelativistic at the time matter began to dominate the energy density of the universe. Weakly Interacting Massive Particles (WIMPs) are an excellent candidate for nonbaryonic, cold dark matter. Minimal supersymmetry provides a natural WIMP candidate in the form of the lightest superpartner, with a typical mass Md ~ 100 GeV c-2. WIMPs are expected to have collapsed into a roughly isothermal, spherical halo within which the visible portion of our galaxy resides. They would scatter off nuclei via the weak interaction, potentially allowingth eir direct detection.

The Cryogenic Dark Matter Search (CDMS) employs Ge and Si detectors to search for WIMPs via their elastic-scatteringin teractions with nuclei while discriminating against interactions of background particles. The former yield nuclear recoils while the latter produce electron recoils. The ionization yield (the ratio of ionization production to recoil energy in a semiconductor) of a particle interaction differs greatly for nuclear and electron recoils. CDMS detectors measure phonon and electron-hole-pair production to determine recoil energy and ionization yield for each event and thereby discriminate nuclear recoils from electron recoils.

This dissertation reports new limits on the spin-independentWIMP-nucleon elastic-scattering cross section that exclude unexplored parameter space above 10 GeV c-2 WIMP mass and, at > 75% CL, the entire 3s allowed region for the WIMP signal reported by the DAMA experiment. The experimental apparatus, detector performance, and data analysis are fully described.

 

Lots of gravitating material that doesn't emit or absorb light seems to be required in all sensible accounts of the dynamics of large-scale structures in the universe. The nature and extent of this mysterious "dark matter" has been one of the central puzzles in cosmology over the last decade. This dissertation describes an experiment that tests one possibility, that the dark matter is in the form of undiscovered Weakly Interacting Massive Particles (WIMPs) produced as a thermal relic of the big bang. In this chapter, we will review the most important observations that suggest the dark matter must exist and discuss the forms it could take.

 

Observations have shown that galaxies, including our own, are surrounded by halos of "dark matter". One possibility is that this may be an undiscovered form of matter, weakly interacting massive particls (WIMPs).

This thesis describes the development of silicon based cryogenic particle detectors designed to directly detect interactions with these WIMPs. These detectors are part of a new class of detectors which are able to reject background events by simultaneously measuring energy deposited into phonons versus electron hole pairs. By using the phonon sensors with the ionization sensors to compare the partitioning of energy between phonons and ionizations we can discriminate betweeen electron recoil events (background radiation) and nuclear recoil events (dark matter events). These detectors with built-in background rejection are a major advance in background rejection over previous searches.

Much of this thesis will describe work in scaling the detectors from 1/4 g prototype devices to a fully functional prototype 100 g dark matter detector. In particular, many sensors were fabricated and tested to understand the behavior of our phonon sensors, Quasipartice trapping assisted Electrothermal feedback Transition edge sensors (QETs). The QET sensors utilize aluminum quasiparticle traps attached to tungsten superconducting transition edge sensors patterned on a silicon substrate. The tungsten lines are voltage biased and self-regulate in the transition region. Phonons from particle interations within the silicon propogate to the surface where they are absorbed by the aluminum generating quasiparticles in the aluminum. The quasiparticles diffuse into the tungsten and couple energy into the tungsten electron system. Consequently, the tungsten increases in resistance and causes a current pulse which is measured with a high bandwidth SQUID system....

 

A substantial body of observational evidence indicates that the universe contains much more material than we observe directly via photons of any wavelength. The existence of this "missing" mass or "dark" matter is inferred by its gravitational effects on the luminous material. Accepting the existence of dark matter has profoundly shaken our understanding in most areas of cosmology. If it exists at the lowest densities measured it is hard to understand in detail the creation of the elements in the early universe. If moderate density values are correct, then we have trouble understanding how the universe came to have so much structure on large scales. If the largest densities are correct, then dark matter is not ordinary matter, but must be something exotic like a new fundamental particle.

We would like to measure the properties of the dark matter directly. Supposing that the dark matter consists of a new fundamental particle, a WIMP, that was in thermal equilibrium in the early universe, we have built an experiment to detect dark matter directly by elastic scattering with germanium or silicon nuclei. Our detectors are large (~ 200 g) calorimeters that can discriminate between interactions with the electrons, due to background photons and beta particles, and interactions with the nuclei, due to WIMPs and background neutrons. The detectors operate at low temperatures (~ 20 mK) in a specially constructed cryostat. To reduce the rate of background events to a manageable level, the detectors and cryostat have been constructed out of selected materials and properly shielded. This dissertation discusses the properties of the hypothetical WIMPs, the detectors, cryostat, and shielding system, and finally, the analysis methods.

 

A major problem currently facing astrophysics and cosmology is the question of dark matter. Although there is little doubt about the existence of dark matter, there is considerable uncertainty about the abundance and nature of this matter. One possibility is that dark matter consists of weakly interacting massive particles (WIMPs), such as the lightest stable particle in supersymmetry models.

Direct detection experiments look for nuclear recoils from WIMPs scattering in a detector. The first generation of direct detection experiments were ultimately limited by radioactive backgrounds. The Cryogenic Dark Matter Search (CDMS) is a direct detection experiment based on novel particle detectors operated at millikelvin temperatures that provide intrinsic background rejection. This capability, however, is not 100% effective. Therefore a low background environment is essential to the experiment.

To create such an environment, all possible background sources have been extensively studied both by measuring the background contribution from muons, phonons and neutrons and by performing detailed Monte Carlo simulations of the photon and neutron backgrounds. The results of this investigation, as discussed in this thesis, have influenced all aspects of the CDMS experiment....

 

This thesis describes the development of several superconducting tungsten thin film based particle detector technologies. The initial motivation for this work was the construction of detectors sensitive to dark matter and neutrino scattering events. These technologies also show promise in other applications, including high resolution x-ray spectroscopy.

The detectors described here consist of a tungsten thin film deposited on a silicon substrate. When an incident particle scatters in the silicon crystal, it deposits energy in the form of phonons which propagate to the surface of the crystal where they are absorbed in the tungsten thin film. The superconducting film is biased at or near its transition temperature. Changes in the resistance of the film are measured.

The superconducting titanium transition-edge sensors previously developed by our group exhibit a threshold phonon energy density below which no signal is detectable. This threshold density poses severe restrictions on resolution, energy threshold, and absorber mass. In order to overcome these limitations, several new technologies were developed. In each case, a superconducting film with a sharp transition well below that of titanium (~ 380 mK) is necessary. To this end superconducting W films were developed with ~ 1 mK wide transitions at 70 mK. Before this work W thin films always exhibited transition temperatures > 600 mK....

 

We have configured a double-sided Silicon Crystal Acoustic Detector (SiCAD) for simultaneous measurement of both phonons and ionization. This detector operates at ~ 370 mK and consists of a Ti Transition Edge Sensor (TES), which is the phonon detector, on one side, and a similar pattern of metal, acting as an electrode for the ionization measurement, on the other side of a 300 μm thick high-purity, monocrystalline Si wafer. The phonon sensor is also position sensitive: it measures the distance between the event location and the detecting surface and can be used to reject background events occurring near the surface. In addition, a coupling between the phonon sensor and the ionization sensor allows the determination of event position in one dimension in the plane of the detecting surface. We present the results of experiments which demonstrate the discrimination capability and position sensitivity of the detector for energy depositions above ~ 3 keV.

The physics of charge measurement, necessary for the background rejection technique, in silicon at low temperature (T < 0.5 K) and low applied electric field (E = 0.1 - 100 V/cm) has been examined in a variety of high purity, p-type silicon samples with room temperature resistivity in the range 2 - 40 kΩ-cm. The samples varied in thickness from 300 μm to nearly 5 mm. Charge loss at low electric field due to trapping during charge drift is present but the data suggest that another charge-loss mechanism is also important. We present results which indicate that a significant fraction of the total charge loss (compared to full collection) occurs in the initial charge cloud near the event location. A simple model of charge trapping both in the initial cloud and along the electric field induced drift to the electrodes is developed and satisfactory comparison to the data is found. In addition, measurements of the lateral size, transverse to the applied electric field, of the initial electron-hole cloud indicate large transverse diffusion lengths. At the lowest fields a lateral diameter on the order of 1 mm is found in samples ~ 5 mm thick.

 

One of the most important issues in astrophysics and cosmology is understanding the nature of dark matter. One possibility is that it is made of weakly interacting subatomic particles created in the big bang, such as the lightest particle in supersymmetry models. These particles should scatter elastically of nuclei in a detector on earth at a rate of ~ events/kg/week, and will deposit energies of a few keV. Current attempts to detect these interactions are limited by a radioactive background of photons and beta particles which scatter on electrons.

We have developed a novel particle detector to look for dark matter based on the simultaneous measurement of ionization and phonons in a 60 g crystal of high purity germanium at a temperature of 20 mK. Background events can be distinguished by our detector because they produce more ionization per unit phonon energy than dark matter interactions.

The phonon energy is measured as a temperature change in the detector by means of neutron transmutation doped germanium thermistors attached to they crystal. The ionization measurement is accomplished by applying a bias to implanted contacts on the faces of the disk. Charge collection differs from the normal situation at 77 K in efficiency is good with an electric field of only ~0.2 V/cm after the charged impurities in the crystal have been neutralized by free charge created by particle interactions from a radioactive source. For fields below this charge collection is poor, and affects the amount of phonon energy measured. We have modeled this in terms of charge trapping.