UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Optical characterization of the DEAP 360 Dark Matter detector Litvinov, Oleksandr 2019

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


24-ubc_2019_november_litvinov_oleksandr.pdf [ 19.35MB ]
JSON: 24-1.0380749.json
JSON-LD: 24-1.0380749-ld.json
RDF/XML (Pretty): 24-1.0380749-rdf.xml
RDF/JSON: 24-1.0380749-rdf.json
Turtle: 24-1.0380749-turtle.txt
N-Triples: 24-1.0380749-rdf-ntriples.txt
Original Record: 24-1.0380749-source.json
Full Text

Full Text

OPTICAL CHARACTERIZATION OF THEDEAP-3600 DARK MATTER DETECTORbyOleksandr LitvinovA THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinThe Faculty of Graduate and Postdoctoral Studies(Physics)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)August 2019c© Oleksandr Litvinov, 2019The following individuals certify that they have read, and recommend to the Faculty of Gradu-ate and Postdoctoral Studies for acceptance, the thesis entitled:Optical characterization of the DEAP-3600 dark matter experimentsubmitted by Oleksandr Litvinov in partial fulfillment of the requirements for the degree ofMaster of Science in Physics.Examining Committee:Fabrice Retiere, Physical Sciences Division, TRIUMFSupervisorGordon Walter Semenoff, PhysicsSupervisory committee memberiiAbstractDEAP-3600 (Dark Matter Experiment using Argon for Pulse shape discrimination) is a darkmatter experiment using liquid argon as a target to detect dark matter’s Weakly InteractingMassive Particles (WIMPs). This work is a contribution to the experiment, focused on opticalprocesses inside the inner detector. In order to re-scale and calibrate complex optical systemused in the apparatus, the distribution of external light is studied. This work includes a qual-itative description of the optical model, data analysis of light reflection from liquid/gas argonboundary and an attempt of estimating the refractive index of wave-shifter material.iiiLay SummaryVery little is known about Dark Matter, despite the fact it comprises roughly 26% of the totalenergy of the Universe. DEAP-3600 (Dark Matter Experiment using Argon for Pulse shapediscrimination) is a dark matter experiment focused on the direct detection of WIMPs (WeaklyInteracting Massive Particles). It is designed to detect the scintillation light, emitted in clodliquid argon after interacting with WIMPs. Therefore, precise optical characterization of theapparatus is crucial for accurate particle detection. This work is a contribution to the opticalpart of the experiment, focused on its calibration methods.ivPrefaceThis work is based on the apparatus and data of the DEAP-3600 experiment, the subject ofa large international collaboration. The data analysis in chapter 3 is completed by author,O.Litvinov under supervision of F.Retiere and P.Giampa.vTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiLay Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Evidence For Dark Matter . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1.1 Gravitational Anomalies . . . . . . . . . . . . . . . . . . . . . . . . . 21.1.2 Weak-Gravitational Lensing . . . . . . . . . . . . . . . . . . . . . . . 31.1.3 Cosmic Microwave Background . . . . . . . . . . . . . . . . . . . . . 41.2 Dark Matter Candidates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.2.1 WIMPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61.2.2 Axions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71.3 Detection of Dark Matter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71.3.1 Dark Halo Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71.3.2 Detection Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8vi1.3.3 WIMP direct detection . . . . . . . . . . . . . . . . . . . . . . . . . . 91.3.4 Current Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 DEAP-3600 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132.1 Noble Liquids Physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.1.1 Liquid Argon Dimers . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.1.2 Pulse Shape Discrimination . . . . . . . . . . . . . . . . . . . . . . . 162.2 Detector Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172.2.1 Inner Detector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172.3 Radioactive Backgrounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212.3.1 39Ar Beta Decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222.3.2 238U and 232Th Alpha Decay Chains . . . . . . . . . . . . . . . . . . . 232.3.3 Neutron Backgrounds . . . . . . . . . . . . . . . . . . . . . . . . . . 252.4 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262.4.1 Region of Interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262.4.2 Data Acquisition System . . . . . . . . . . . . . . . . . . . . . . . . . 273 Optical model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293.1 Hamamatsu R5912 Photo-Multipliers . . . . . . . . . . . . . . . . . . . . . . 293.2 AARF Calibration System . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313.3 Argon Boundary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343.4 TPB Optical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393.4.1 Changes in Light Distribution . . . . . . . . . . . . . . . . . . . . . . 413.4.2 Pulse Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 434 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48viiList of Tables1.1 The main parameters of the Dark Halo Model [21]. . . . . . . . . . . . . . . . 82.1 The main parameters of the DEAP-3600 detector [3]. . . . . . . . . . . . . . . 203.1 Current estimation of main optical parameters for the DEAP-3600 inner detector 45viiiList of Figures1.1 Observed galactic rotation curves of NGC 6503 (left) and NGC 3198 (right)as functions of the distance to the center of mass, and modelled portion of thedisk, gas and halo structure [11, 41] . . . . . . . . . . . . . . . . . . . . . . . 21.2 Observation of the lensing effect in space [29] . . . . . . . . . . . . . . . . . . 31.3 Gravitational lensing in the Bullet Cluster [16]. Green lines identify basedon gravitational lensing predicted mass distribution, while while X-ray scan(right) predicts concentrated mass locations to be in yellow and red zones . . . 51.4 Recent results of Planck collaboration on measuring the Cosmic MicrowaveBackground [17] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61.5 Cosmic Microwave Background temperature spectrum measured by Planckcollaboration in 2015 [1]. The relative amplitudes at the bottom show the ap-proximate ratio of baryonic matter to dark matter to be 1 to 5. . . . . . . . . . . 61.6 Schematic Feynman diagram of three different detection methods based on theinitial and final particles, participating in interactions. Dark matter candidatesare labeled as χ , and P represents baryonic matter. . . . . . . . . . . . . . . . . 91.7 Different techniques using for WIMP direct detection [40] . . . . . . . . . . . 101.8 Current status of leading WIMP-search experiments. Green section is provedto be forbidden for WIMP parameter space, while current lower limit of thesearch is set to be a neutrino floor [19]. . . . . . . . . . . . . . . . . . . . . . . 122.1 Schematic diagram of the scintillation photon emission in the recombinationprocess . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15ix2.2 Schematic diagram of the scintillation photon emission in the excitation process 162.3 Cross-sectional diagram of the DEAP-3600 detector [3]. The lower sphere isfiled with approximately 3300 kg of liquid argon. Cooling and filling systemsare realized using a vertical neck cylindrical part. . . . . . . . . . . . . . . . . 182.4 Acrylic vessel (left) and copper shielding (right) in the process of installation . 192.5 Schematic cross section of the inner Hamamatsu PMT coupled to the acrylicvessel [3] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212.6 Detailed design of PMT and light guide connection [3] . . . . . . . . . . . . . 222.7 Thorium decay chain, including lifetimes of all the intermediate particles. Al-pha decays are labeled in green, beta decays are labeled in red . . . . . . . . . 232.8 Uranium decay chain, including lifetimes of all the intermediate particles. Al-pha decays are labeled in green, beta decays are labeled in red . . . . . . . . . 242.9 Schematic diagram of possible surface alpha radioactivity events [3] . . . . . . 252.10 Fprompt versus recoil energy parameter space for the DEAP-3600 [2]. Data for64 h AmBe NR events. The boundaries of the ROI are shown in black. Alldata-quality cuts are applied. . . . . . . . . . . . . . . . . . . . . . . . . . . . 272.11 Data flow architecture for DEAP-3600 [3] . . . . . . . . . . . . . . . . . . . . 283.1 Crossectional drawing of the R5912 Hamamatsu HQE photomultiplier tube.The active photocathode area is approximately 530 cm2 [26] . . . . . . . . . . 303.2 Quantum efficiency of the R5912 Hamamatsu HQE photomultiplier as a func-tion of the incoming photon’s wavelength . . . . . . . . . . . . . . . . . . . . 313.3 AARF method of light injection into an acrylic vessel [4] . . . . . . . . . . . . 323.4 Occupancy distribution among PMTs when one of the AARFs is fired [4]. Datafor approximately 2 million events. . . . . . . . . . . . . . . . . . . . . . . . . 343.5 Schematic diagram of light injection inside the AV and argon reflection pro-cesses with low and high LAr levels . . . . . . . . . . . . . . . . . . . . . . . 35x3.6 Light distribution for low LAr levels (indicated with red line)) with a lightsource at the bottom of the vessel. Statistics for 500K events . . . . . . . . . . 363.7 Light distribution for high LAr levels (indicated with red line) with a lightsource at the bottom of the vessel. Red line indicates the LAr level. Statisticsfor 500K events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363.8 Comparison of the light distribution for low LAr level (shown with dash line)and an almost empty vessel for 500K events. . . . . . . . . . . . . . . . . . . . 373.9 Ratio of the light yield for high LAr level (shown by a dashed line) and analmoslt empty vessel for 500K events. . . . . . . . . . . . . . . . . . . . . . . 383.10 Simulation data. Ratio of the light distribution for low LAr level and an emptyvessel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393.11 Simulation data. Ratio of the light distribution for high LAr level and an emptyvessel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393.12 Qualitative impact of TPB presence: more photons get trapped inside acryliclight guides. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403.13 Comparison of light distributions with a light source in the AARF ID 1. Statis-tics for 500K events. Each slot represents the ratio of detected light after andbefore the TPB was installed. . . . . . . . . . . . . . . . . . . . . . . . . . . . 413.14 Comparison of light distributions with a light source in the AARF ID 3. Statis-tics for 500K events. Each slot represents the ratio of detected light after andbefore the TPB was installed. . . . . . . . . . . . . . . . . . . . . . . . . . . . 423.15 Comparison of light distributions with a light source in the AARF ID 8. Statis-tics for 500K events. Each slot represents the ratio of detected light after andbefore the TPB was installed. . . . . . . . . . . . . . . . . . . . . . . . . . . . 42xi3.16 Sample of total amount of incoming pulses over time distribution. The timescale is the difference between PMT reading and trigger time. PMTs are re-arranged based on the distance to the light source. Statistics for 500K events.AARF light source ID 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443.17 Average time delay for first incoming photons after implementing TPB over allPMTs. Statistics for 500K events. AARF light source ID 1. . . . . . . . . . . . 453.18 Average time delay for first incoming photons after implementing TPB over allPMTs. Statistics for 500K events. AARF light source ID 3. . . . . . . . . . . . 463.19 Average time delay for first incoming photons after implementing TPB over allPMTs. Statistics for 500K events. AARF light source ID 8. . . . . . . . . . . . 46xiiAcknowledgmentsFirst and foremost, my sincere gratitude goes to my supervisor, Fabrice Retiere, who made mea part of the DEAP collaboration, where I met so many incredible people. Thank you for allyour help, support and opportunities you gave me over the last two years.I owe special thanks to Pietro Giampa for sharing his experience, knowledge and wisdom.My progress would be barely possible without his leadership. Thank you, Pietro.Thank you everyone in the DEAP collaboration. I learned a lot from each of you and I amhappy I had the opportunity to get to know you better during our travels. Especial thanks go toJoe McLaughlin, who gave me a ride on my very first day with DEAP in Sudbury and becamemy good friend during his stay in Vancouver.Apart from that, I owe a debt of gratitude to everyone who did not let me feel lonelythroughout my workdays at TRIUMF. I am also grateful to all my university friends who mademy studying and life enjoyable and fun.Lastly, I would like to thank my family for all their love, patience and care. Thank you.xiiiChapter 1IntroductionThe nature and composition of Dark Matter remains one of the most pressing open-questionsin modern physics. Growing interest in this topic can be easily explained with the fact that only5% of the total mass-energy of the universe consists of baryonic matter. The remaining 95%are predicted to be a mixture of dark matter and dark energy with the portions of 26% and 69%correspondingly. Despite growing effort in experimental physics during the past few decadesno direct observations have been made yet, although astrophysics provides multiple indirectevidences. Successful discovery of the dark matter may be the next major step in modernphysics and a considerable expansion of the Standard Model.1.1 Evidence For Dark MatterVery little is known about Dark Matter except that it makes up for the vast majority of the massof the universe and it does not interact with light. The first indirect evidence for the existenceof Dark Matter came at the beginning of the 20th century. Even though we can not see it,we do know it is there due to gravitational anomalies which astrophysicists could observe atlarge scales. Astronomical observations reported more gravitational oddities, which could notbe described without introducing the new concept of matter. Gravitational lensing and CosmicMicrowave Background stand out among others as essential indirect evidences for the existence1Figure 1.1: Observed galactic rotation curves of NGC 6503 (left) and NGC 3198(right) as functions of the distance to the center of mass, and mod-elled portion of the disk, gas and halo structure [11, 41]of Dark Matter.1.1.1 Gravitational AnomaliesFritz Zwicky first coined the term Dark Matter while studying the Coma cluster [17]. Bycalculating the velocity dispersion of the cluster, using Doppler shifts in its visible spectrumin 1933, Zwicky used the virial theorem (Eq. 1.1) to estimate the gravitational potential of thecluster:< T >=−12<V > (1.1)where <T> is the mean kinetic energy in the center of mass frame and <V> is the mean po-tential energy of the system. Zwicky’s calculations showed that the actual mass of the clusterwas much higher than any predicted one based on the total masses of individual objects (esti-mated using luminosity measurements). He assumed that the bulk of the cluster must be filledwith some new sort of matter which energy is dominated over the energy of visible cosmicobjects.The second indirect evidence of Dark Matter came from Vera Rubin in the early 1970s,2Figure 1.2: Observation of the lensing effect in space [29]when she applied studying the rotational curve of the Andromeda galaxy. On the example ofAndromeda Galaxy, Rubin first used a standard Newtonian approach:mv2r=GmMr2(1.2)where m is the mass of some particular star in the galaxy, v is linear velocity, r is thedistance between the star and the center of mass of the galaxy, M is the mass of the part ofthe galaxy inside the sphere of a given radius r, and G is the gravitational constant. As simplemechanics predicts, velocity keeps growing within the center of the disk and it falls off in thespiral arms. However, the observed distribution was consistent with flat after eighth kiloparsecs(∼26000 light years).The flat behaviour could be accounted for by introducing into account a growing darkmatter halo within the galaxy. Astronomers were able to reproduce this result with multiplegalaxies, such as NGC 6503 and NGC 3196.1.1.2 Weak-Gravitational LensingWeak-gravitational lensing studies provide another strong evidence for the existence of DarkMatter. In general, gravitational lensing occurs when light that travels large distances is dis-torted by an heavy object placed between the source and the observer.3The deflection angle of the perturbed light is proportional to the mass of the object:θ =4GMrc2(1.3)where theta is the deviation of real angle between the source of light and observer, G is theuniversal gravitational constant, M is the mass of the object and r is an impact parameter. Thisapproximation for point mass can be applied to galaxies allowing to measure mass-distributionswhen the position of an emitter is known precisely. In some cases, when an heavy object islocated on the same line between source and receiver, radial symmetry allows to observe so-called Einstein rings (Figure 1.2).However, weak-gravitational lensing can also be used to measured the effects of dark matterhalos in far objects like the Bullet cluster. As a result of the merging of two neighbor galaxyclusters, the Bullet cluster was analyzed by scientists using X-ray measurements. The X-rayitself was emitted by hot plasma in the epicenter of clusters’ collision [17]. The observationshown in Figure 1.3 revealed that zones of dense baryonic matter are not the same as calculatedlocations of mass peaks, which can only mean that clusters are dominated with non-lumiousmatter [29].1.1.3 Cosmic Microwave BackgroundA further indirect evidence of dark matter comes from the early ages of our universe and canbe found in the Cosmic Microwave background (CMB). First detected in the 20th century byArno Penzias and Robert Wilkinson, the CMB describes the conditions of the early ages ofthe Universe. CMB is a remnant electromagnetic radiation from the Big-Bang, which containscrucial information about the composition of our universe. While temperature distribution ofthe CBM seems to be flat on a large scale, more precise analysis reveals angular fluctuations oftemperature in range of 10−5K (Figure 1.4).Using standard cosmological model, the Plank collaboration [1] studied the energy densitydistribution as a factor of angular distances (Figure 1.5). Parameter l represents expansion of4Figure 1.3: Gravitational lensing in the Bullet Cluster [16]. Green lines iden-tify based on gravitational lensing predicted mass distribution, whilewhile X-ray scan (right) predicts concentrated mass locations to bein yellow and red zonesthe energy density distribution. The current estimations from the study show that Dark Mattermakes up for 25% of the entire universe.1.2 Dark Matter CandidatesDespite the list of indirect evidences discussed in the previous section, very little is knownabout Dark Matter. The mainstream theory expands Beyond the Standard Model of particlephysics (BSM), with a large number of proposed candidates. However, there are few limits andconditions, which all the dark matter candidates must satisfy:Non Baryonic, motivated by the lack of interaction with light; Non Relativistic, motivatedby large-scale structure of the universe [10]; Long-Lived, the lifetime has to be of the order ofthe life of the universe for it to be in relic density.It is impossible to mention every candidate in a huge variety, so the next sections will befocused on the most popular among current theories.5Figure 1.4: Recent results of Planck collaboration on measuring the Cosmic Mi-crowave Background [17]Figure 1.5: Cosmic Microwave Background temperature spectrum measured byPlanck collaboration in 2015 [1]. The relative amplitudes at the bot-tom show the approximate ratio of baryonic matter to dark matter tobe 1 to WIMPsThe current most favourite candidate are Weakly Interacting Massive Particles (WIMPs). Fromits name, WIMPs interact weakly and have expected masses in the GeV-TeV range [28].6During the early stages of the Universe, WIMPs were in perfect equilibrium between pro-duction and annihilation, but as the universe started expanding and therefore cooling, the over-all energy would have fallen below the threshold for WIMP production. This effect would haveleft the relic WIMP abundance that we may observe today [25].The work put forth in this thesis is focused on directly on WIMPs searches.1.2.2 AxionsCharge parity (CP) violation remains to be one of the biggest open questions in current StandardModel theory, breaking entire CP-conservation in quantum chromodynamics (QCD). In theirattempt of solving one, Peccei and Quinn suggested a new Symmetry (U(1)) [37] for the lastterm in Lagrangian Eq. 1.4 for QCD, while first two have symmetry (SO(3)):LQCD =−14FαµνFµνα −∑nψ¯nγµ [∂µ − igAαµ tα ]ψn−∑nmnψ¯nψn (1.4)where the first term describes gluon interaction with gluon field tensors, ψn is a quark fieldof Dirac four-spinor, and ta stays for 8 gluon colour matrices in the second term, describingchromodynamic part of the equation.This new symmetry requires the existence of a new particle - Axion [42]. Today, the ADMX(Axion Dark Matter eXperiment) is a current leading collaboration searching for this new par-ticle [39]. Using Primakoff effect with a supercondating microwave resonator, they have grad-ually been narrowing the mass-region of Axions in last several years.1.3 Detection of Dark Matter1.3.1 Dark Halo ModelThe characteristics of the Dark Matter halo in our galaxy are fundamental to any proposedDark Matter measurement, as they directly impact the expected rate in our frame. Spherical7dark matter halos are strongly motivated by gravitational lensing and rotational curves mea-surements. This model predicts the density of the Dark Matter halo ρx falls off proportionalto r−2, where r is the distance to the galactic center. According to Lewin and Smith, the halodensity for Milky Way galaxy varies from 0.2 [GeV/(c2 cm3)], to 0.4 [GeV/(c2 cm3)], with anaverage 0.3 [GeV/(c2 cm3)] [32]. The simplest physical distribution for this theory would be toassume Maxwellian velocity distribution for the dark matter particles in the halo [32]. Exceptthat, Standard Halo Model (SHM) assumes halo to be isothermal, meaning that its tempera-ture is stable and does not depend on the exact location [31]. Exact parameters of SHM arepresented in Table 1.1Parameter Standard valuev0 220 [km/sec]vsun 232 [km/sec]vesc 544 [km/sec]ρhalo 0.3 [GeV/c3cm3]Table 1.1: The main parameters of the Dark Halo Model [21].1.3.2 Detection MethodsThere are many techniques used to probe dark matter. In general these searches can be classifiedin three groups: direct detection, indirect detection, and production in collisions. Dependingon how one reads the Feynman-style process diagram on Fig. 1.6, all three processes can bedescribed qualitatively.Production in collision is based on generating the same conditions of early stages of theUniverse in a laboratory, by colliding particles at high Energies. Accelerated to TeV energies,baryonic particles could create dark matter directly. Traditional collider detectors can measureenergies of all the outgoing particles, imbalance of transverse energy after the explosion wouldmean a loss of energy due to the creation of dark matter, which can not be detected.8Figure 1.6: Schematic Feynman diagram of three different detection methodsbased on the initial and final particles, participating in interactions.Dark matter candidates are labeled as χ , and P represents baryonicmatter.Indirect detection is based on annihilation decay of Dark Matter particles that can lead tothe production of baryonic matter. Annihilation is proportional to dark matter density squared,and resulting particles can be caught and analyzed with detectors [18].Direct search of dark matter particles relies on the recoil signal from elastic or quasi-elasticscattering of Dark Matter particles with a target nuclei or electrons [21]. The primary challengeof this method is to separate WIMP-produced recoils from backgrounds in the detector sincenot only dark matter can produce a detectable signal.1.3.3 WIMP direct detectionElastic and quasi-elastic scattering of WIMPs with matter can lead to the production of thefollowing effects in a detector: phonon production, charge production, and photon production.Since the basic process of direct detection method is recoil kinematics, one can postulatethat recoil energy Er of a target nuclei, scattered elastically with a WIMP particle of mass mχand velocity v, can be calculated using Eq. 1.5:Er =µ2r v2mn(1−Cosθ) (1.5)9Figure 1.7: Different techniques using for WIMP direct detection [40]where mn is target nucleus mass, θ is scattering angle in the WIMP-nucleus centre of massframe and µr is reduced WIMP-nucleus mass:µr =µχmnµχ +mn(1.6)Generally, differential cross-section of the scattering Dark Matter particle can be spin-dependent (SD) or spin-independent (SI), so both terms must be taken into account. In thiscase, the differential cross-section of WIMP-nucleus can be expressed asdσχndEr=mn2µ2r v2(σSI0 F2SI +σSD0 F2SD) (1.7)where FSI and FSD are form factors for spin-independent and spin-depended term correspond-ingly, and are functions of the transferred energy. For the SD case the cross-section can becalculated for the total spin of the target nucleus J as following:σSD0 =(32pi)G2Fµ2r(J+1J)[〈Sp〉ap+ 〈Sn〉an]2 (1.8)with Fermi constant GF , ap and an to be effective coupling of the WIMP to proton andneutron correspondingly for the SD case, 〈Sp〉 and 〈Sn〉 to be the expectation values of spin for10proton and neutron:〈Sn,p〉= 〈N|sn,p|N〉 (1.9)WIMP-proton and WIMP-neutron cross-section limits for SD case can be expressed asfollowing:σSDn,p =24G2Fµ2n,pa2n,ppi(1.10)when only one of two interactions (proton or neutron) is dominated and the second one canbe reasonably neglected.For SI case, WIMP-nucleus cross-section is given by [38]:σSI0 =4µ2rpi[Z fp+(A−Z) fn]2 (1.11)where A and Z are atomic mass and atomic number of the target nucleus, fp and fn are theeffective SI coupling of a WIMP to proton and neutron correspondingly.1.3.4 Current StatusCurrently a large number of experiments cover both the SI and SD WIMP detection channel.No observation has been made in the field. Various experiments have been able to further con-strain different combinations of WIMP masses and cross-sections. A list of principle publishedWIMP exlusion cross-section curves are shown in Figure 1.8. The DEAP-3600 experiment,the focus of this thesis, achieved the best sensitivity among LAr-based experiments.11Figure 1.8: Current status of leading WIMP-search experiments. Green sectionis proved to be forbidden for WIMP parameter space, while currentlower limit of the search is set to be a neutrino floor [19].12Chapter 2DEAP-3600The Dark matter Experiment using an Argon Pulse-shape discriminator with about 3300 kgof liquid argon (DEAP-3600) is the focus of the work reported in this thesis. The detectoris located 2 kilometers underground in an active mine at SNOLAB, in Sudbury ON, Canada[20]. The DEAP-3600 is a single-phase liquid argon (LAr) detector optimized for the detectionof spin-independent WIMP-nucleus interactions. The maximum target-medium load of thedetector is 3600 kilograms, however, due to post-commissioning hardware issues it currentlyoperates at approximately 3300 kg [12]. The sensitivity of the detector to the WIMP-nucleoncross-section is 10−46 cm2 for a WIMP mass of 100 GeV/c2 [5]. The main purpose of theexperiment is to detect a WIMP directly using scintillation light, which would be producedafter a dark matter particle scatters off a target atom in the detector. A further benefit of LAris the two distinct time-constants of the generated scintillation light: a fast time-constant onthe order of a few nanoseconds and a long time-constant on the order of one microsecond,depending on a quantum state of an excited argon dimer. Depending on the nature of therecoil (nuclear or electronic), the population distribution of the quantum states is considerablydifferent, making it easy to distinguish between nuclear recoils and beta/gamma radiation inthe experiment.132.1 Noble Liquids PhysicsAlthough the scintillation method can be applied to other noble liquids, there are still someadvantages of using argon specifically. Helium should not be used as a target medium due toits extremely low boiling temperature and comparatively light mass. Other possible candidatesare neon, argon, and xenon. The atomic number (Z=18 for argon) plays a significant role inthe process of recoils. The higher the density of a medium, the larger the number of targetnucleons occupying the same volume, linearly increasing the probability of interaction withtraveling particles. Another crucial feature of a noble gas for dark matter searches is a photonyield per deposited energy. Per 1 keV of energy, argon emits 42 photons at zero electric field,which is much higher in comparison to neon’s 15 [13] and results in a more efficient detector.The deciding factor prioritizing argon over xenon is the simplicity of extracting it from theenvironment. Due to a large portion of argon in the atmosphere, it is relatively easy to getargon from liquid air by distilling. It affects the cost of an experiment drastically. To date,xenon is more than 300 times more expensive than argon, while its scientific benefit is stillunder discussion.2.1.1 Liquid Argon DimersA dimer is an unstable bound state of two noble liquid atoms in cold temperature. The dimers’decay leads to a photon emission [35], called scintillation light. Emitted light during the dimerdecay process has a peak wavelength of roughly 128 nm, being in an ultraviolet (UV) spectrum[27]. Dimers are the resulting products of argon atom excitation, which can be achieved in twoways: recombination and self-trapped excitation [23].The recombination process is schematically shown in Figure 2.1. First, an argon atom inthe ground state interacts with a traveling particle and gets ionized, emitting a free electron.A positively charged argon ion combing with another ground-state argon atom to form anunstable dimer ion. Recombination with a free electron makes the dimer electrically neutral.14Figure 2.1: Schematic diagram of the scintillation photon emission in the re-combination processThe decaying, unstable electrically neutral dimer creates two ground-state argon atoms and aphoton in the ultraviolet spectrum to compensate for the difference in energy.The more direct process of creating a spontaneous dimer is called self-excitation, schemat-ically shown in Figure 2.2. In this process, the interaction of an argon atom with a movingparticle and the energy transfer between them causes an excitation of the atom. Two argonatoms in a ground state can not bound together making a dimer, but it can be produced af-ter the interaction between one ground-state atom and one exciton [30]. Again, similar to therecombination process, the unstable dimer decays and emits an ultraviolet photon.As described before, dimers can be produced by excitation but also by the recombinationof electrons and ions especially when the density of ionization is high, which is the case fornuclei energy deposition. Therefore the recombination process will boost the scintillation yieldof nuclei compared to electrons. On the other hand, dimers can decay non-radiatively due todimer-dimer interaction or interactions with excited atoms. This process suppresses scintilla-tion when the density of dimers is high, which again is much higher for nuclei than electrons.This process suppresses scintillation light production by nuclei. Furthermore, the time scale ofnon-radiative de-excitation are comparable to the scintillation time scale and they suppress the15Figure 2.2: Schematic diagram of the scintillation photon emission in the exci-tation processlong lived triplet dimer much more than the short lived singlet dimer. Overall the results is thatelectrons traveling in LAr will yield a lot more late photons due to triplet dimer decays thannuclei.2.1.2 Pulse Shape DiscriminationA liquid argon dimer produced using any of the methods described above can only be in one ofthree states: one triplet and two singlets. The parity conservation principle does not allow oneof the singlet states to emit scintillation light in the process of a dimer decaying down to twoground-state argon atoms [6]. The lifetimes of the other two decay states are very distinct, andthis is the main principle of the Pulse-Shape Discrimination (PSD) method. The triplet statedecay is carried out over 1.6 µs while the splitting of a singlet exciton into two atoms happensduring a much shorter time of 7 ns [27]. Fortunately for the WIMP detection analysis, there isa direct correlation between the probability of creating a singlet or a triplet dimer state and thecharacteristics of traveling particles, whose energy transfer leads to an excitation [5].For this reason, one of the main parameters in the DEAP-3600 light detection process isFprompt, which identifies the ratio of the amount of a prompt light to the total amount of detected16light during one event, i.e., the interaction of LAr with one traveling particle:Fprompt =γprompt[150ns]γtotal(2.1)where γprompt[150ns] represents the amount of light recorded in the first 150 ns of a particularevent and γtotal is the total amount of detected light. This crucial feature for a new particledetection method was proven to be extremely efficient in a previously conducted DEAP-1experiment [22].2.2 Detector DesignThere are a few basic principles and challenges of building such a large detector. First of all,since the expected WIMP interaction rate is tiny, a proper calibration system must be designed,implemented and tested, in order to be able to reconstruct each event independently usingreconstruction methods and probability algorithms. When reducing the dominant backgroundsources of both the NR and ER events, radioactive shielding construction is used to isolatethe fiducial argon from external background sources, while careful material selection mitigatesinternal radiation from each detector part. It is also necessary to collect as many photons aspossible using high efficiency sensors and reflecting materials. This section briefly describesthe motivations behind and realization of the essential hardware parts in the DEAP-3600.2.2.1 Inner DetectorA schematic cross-section of the detector is shown in Figure 2.3 and a list of key parame-ters is shown in table 2.1. The inner detector consists of a spherical acrylic vessel (AV) withan 85 cm inner radius and 5 cm thickness. Acrylic material is well suited for this purposenot only because it is clean, both optically and radioactively, but also because it can shieldthe medium from external neutron background [3]. Besides AV, acrylic cryostat includes 255acrylic light guides, extending 45 cm in length and 19 mm in thickness, connecting the AV17Figure 2.3: Cross-sectional diagram of the DEAP-3600 detector [3]. The lowersphere is filed with approximately 3300 kg of liquid argon. Coolingand filling systems are realized using a vertical neck cylindrical part.with 255 Hamamatsu R5912-HQE photomultiplier tubes (PMTs) for complete light detection.The light guides shield the inner detector from PMT glass radioactivity and allow operating the18Figure 2.4: Acrylic vessel (left) and copper shielding (right) in the process ofinstallationPMT close to room temperature. The light guides cover approximately 75% of the total AVsurface area.Having light reflectors around the AV is important for the DEAP-3600 detector. Reflectorsprevent photon leakage in between light guide, hence maximizing light collection.Diffusivereflectors cover the outer surface of the AV to bounce photons back into the LAr until they gettrapped into one of LGs [3]. Specular reflectors surround the LG surface salvaging a fractionof the photons escaping from the light guides. A 150 µm aluminized mylar foil was chosenas a reflector material. Thin copper shields surround the LGs, and the space between themis filled with polyethylene and Styrofoam materials [3] providing both neutron shielding andtemperature isolation.Inside a thin spherical stainless steel vessel is where the inner detector is placed. The vesselprovides a connection to the inner detector from the outside through a vertical cylindrical neck.A cooling coil inside the neck uses liquid nitrogen for keeping the argon in a liquid state. Onits top, the neck part has a glove box construction. The whole stainless steel installation issubmerged into a 7.8-meter diameter tank. The tank is filled with ultra-pure water, shieldingthe inner detector from external gamma ray radiation and enabling the detection of cosmicmuons. For precise muon detection, 48 outside-looking Hamamatsu R1408 muon veto (MV)19Parameter SpecificationRadius 85 [cm]Total argon mass 3300 [kg]Water shielding tank volume 375 [m3]Number of inner detector PMTs 255Number of Cherenkov veto PMTs 48Light yield About 8 [Npe/keV]Sensitivity around 100 GeV/c2 10−46 [cm2]Table 2.1: The main parameters of the DEAP-3600 detector [3].PMTs are located on the outer surface of the stainless steel frame, scanning the light inside thewater shield [3].The wavelength of scintillation light in liquid argon is in a vacuum ultraviolet spectrum(VUV) with a 128 nm peak [15]. This wavelength corresponds to lower photon energy thanthe energy of the argon’s first excited atomic state. It prevents photons from being absorbed byLAr in AV. Thus, every interaction which happened in the inner detector leading to the creationof a dimer can be detected and reconstructed.For converting VUV photons into visible light and further detection by PMTs, a thin-filmwavelength shifter was deposited onto the inner surface of the AV. Organic 1,1,4,4-tetraphenyl-1,3-butadiene (TPB, C28H22) covers the total internal area of the AV with a layer thickness ofnearly 3 µm [14]. After interacting with TPB, 420 nm photons travel in the AV and then getdetected by PMTs.Since the PMT’s efficiency is dependent on its temperature, sixteen PMTs have a temper-ature sensor installed on the copper parts of the light guides. Copper shields surround everyPMT to equalize the temperature along the photodetectors. Sensors are distributed evenlyamong LGs so that the approximate temperature for each specific PMT can be accurately esti-mated based on the readings of surrounding sensors.20Figure 2.5: Schematic cross section of the inner Hamamatsu PMT coupled tothe acrylic vessel [3]The way a PMT is connected to a light guide is shown in detail in Figure 2.6. Opticalcoupling is achieved between the PMT and LG by using Sigma Aldrich material (378399) to fillthe entire intermediate space of the PMT-LG system [3] inside a cylindrical barrel. The materialwas chosen because of its glass-imitating optical parameters and because it satisfied thermalconduction parameters. A copper sleeve with a high index of thermal conductivity eliminatesthermal fluctuations along the assembly and prevents possible temperature instabilities. AFINMET [8] shield around each PMT lessens the effect of background magnetic fields.2.3 Radioactive BackgroundsVarious background source may mimic wIMP interactions. Background signals from all thesurrounding radioactive sources must be estimated, identified and removed. Therefore, reduc-ing radioactivity is vital for the experiment. Corresponding parameters must be taking intoaccount in the design, which requires a proper material selection. In this way, sensitivity to21Figure 2.6: Detailed design of PMT and light guide connection [3]WIMP cross-section can be maximized.There are some sources of background signals which might look like predicted WIMPevents. Among them are radon decays within an argon medium, material radioactivity of theacrylic in the inner detector, 39Ar beta decay, neutron background, and cosmic background.2.3.1 39Ar Beta DecayAs mentioned before, one of the reasons for using argon for the dark matter search is thesimplicity of its extraction from the atmosphere. However, it also means that stable 40Ar carriesunstable 39Ar isotopes (with a 269 year lifetime) as a result of interaction with cosmic rays [33].Stable potassium 39K is a final product of 39Ar β -decay, which also produces a free electronand electron anti-neutrino particle:39Ar→ 39K+ e−+ ν¯e (2.2)In natural argon, the frequency of isotope decay is estimated to be roughly 1 Bq/kg [9]. As-suming this number to be constant due to its long lifetime, the DEAP-3600 experiences around223300 39Ar β -decay events per second, or 285 million events per day. This rate can be dimin-ished by extracting argon from underground deposits or using various distillation techniques.Nevertheless, this dominant background can be reduced entirely using a PSDmethod describedin section 2.1.2, based on previous studies of a previously completed DEAP-1 experiment [5].2.3.2 238U and 232Th Alpha Decay ChainsFigure 2.7: Thorium decay chain, including lifetimes of all the intermediate par-ticles. Alpha decays are labeled in green, beta decays are labeled inredThe main sources of unstable particles with lifetimes comparable to the age of the earth(∼10 billion years) for the DEAP-3600 are 238U and 232Th decay chains. These chains mix23Figure 2.8: Uranium decay chain, including lifetimes of all the intermediate par-ticles. Alpha decays are labeled in green, beta decays are labeled inredboth α- and β -decay processes, emitting a significant number of electrons and alpha particles.Helium nucleus similarly interact with LAr as argon nuclei do, mimicking scintillation events.In both 238U and 232Th decay chains, starting with 220Rn and 222Rn correspondingly, daughterisotopes have a short lifetime and high energies (approximately a few MeV) of emitted par-ticles [36]. Despite the similarity of Fprompt parameters of the events with predicted WIMPinteractions, these events can still be separated because their energy deposited is much higherthan WIMP interactions. However, in the case that a radioactively emitted 4He nuclei loses asignificant part of its energy before entering the LAr volume, it can be classified as a WIMPROI event. Those events are called surface alpha events.Surface alpha events might occur in three regions: the acrylic vessel, the TPB and theLAr detector. All possible ways of losing energy by alpha-particles before interacting with24Figure 2.9: Schematic diagram of possible surface alpha radioactivity events [3]the LAr are shown in Figure 2.9. In the LAr volume, both daughter and alpha particles willgenerate too many scintillation photons be classified in the WIMP ROI. In the case of alphadecay happening inside the TPB layer, fewer photons are produced though the TPB itself mayproduce scintillation photons. Finally, decay within the acrylic radiation may generate justenough energy to enter the ROI and be classified as WIMP. Position reconstruction is the onlyway to identify and reject radon decay daughters on the surface of the AV or within the TPB.2.3.3 Neutron BackgroundsAnother type of background which might produce the same signal parameters as a WIMP isneutron radiation. It is especially important for this type of experiment to lessen the effect ofinteracting neutrons. The main sources of this flux are natural reactions in surrounding rocks,cosmic muons, and the PMT glass material. Light guides and filler blocks, implemented aroundPMT tubes and described in section 2.2.1, drastically suppress neutron background. Spallationneutrons produced by cosmic muons can be rejected by detecting the muon induced Cherenkovphotons in the water tank system. Since the DEAP-3600 setup is fully submerged into 500 tons25of ultra-pure water, incoming muons emit Cherenkov light [34], which is observed by 48 muonveto PMTs. Engineering solutions allow keeping a stable temperature and fast water flow inthe tank for more accurate light detection.2.4 Data CollectionDue to having hundreds of sensors recording the data continuously, data analysis is anotherimportant part of the project. The motivation of the whole project is to gather statistics ofinteractions in the region whereWIMP scattering can occur. Thus, it is crucial to select physicaldata properly, applying low-level data cleaning cuts, automatically rejecting all the possiblehardware issues and irrelevant signals.2.4.1 Region of InterestThere are two physical parameters, defining the small window of the region of interest (ROI)for the DEAP-3600: Fprompt and the recoil energy of a candidate WIMP-event. Setting thecriteria of these values yields a selection of potential WIMP interactions among a variety ofdetected signals. Since interactions between dark matter and LAr yield nuclear recoil events,electron recoil event can be suppressed. The study of the DEAP-1 experiment was based onthe AmBe radiation source. Radioactive 241Am emits alpha-particle, which can be capturedby 9Be nucleus. As a result of this interaction, the AmBe source emits photon and neutron.The study showed the variation of Fprompt for NR events between 0.7 and 0.9, proving that NRevents have much higher Fprompt in comparison with ER’s ∼0.3 [5].The region of interest is shown in Figure 2.10, having low recoil energy and high Fprompt.It allows to diminish the expected rate of argon background events to be 0.2. The lower PEparameter is set as 80 PE (10 keV) to reject argon beta-decay events to a fraction of 0.5, whilethe higher bound of 240 PE separates the ROI from surface alpha-decay events. Less then10% of NR events are outside the ROI after applying the framework which is a reasonable26Figure 2.10: Fprompt versus recoil energy parameter space for the DEAP-3600[2]. Data for 64 h AmBe NR events. The boundaries of the ROIare shown in black. All data-quality cuts are applied.compromise.Targeted number of a 3-tonne-year fiducial exposure is 30 neutron events in the ROI, to-gether with 150 surface alpha events and 1.6 billion events of gamma or beta 39Ar decay [3].2.4.2 Data Acquisition SystemAll signals from the PMT detectors are digitized in order to enable identifying individual PMTpulses with a timing resolution limited by the PMT transit time spread of about 1ns (σ ). Themain criteria for the DEAP-3600 Data Acquisition System (DAQ) is to detect all possible veryrare WIMP signals without being swamped with the dominant 39Ar background events. Thedata flow architecture for the DEAP-3600 is shown schematically in Figure 2.11. A WIENERMPOD crate powers the PMTs with a high voltage module. The PMTs convert a signal fromphotons into an analog pulse. The PMT response is described with more details in Section3.1. All PMTs are connected to signal conditioning boards (SCB). There are 27 SCBs with12 channels each, supporting LAr, muon veto and veto neck PMTs. A Digitizer and TriggerModule (DTM) analyzes the analog signals produced by summing the 12 PMT signals within27the SCBs connected to the LAr PMTs. The DTM produces trigger signals that are used toselect when to acquire data with the digitizers CAEN V1720s and V1740s. 250 Mega sampleper second (MS/s) V1720 boards are used for most data providing dynamic range between 1and about 60 photo-electrons. 62.5MS/s V1740 digitizers are used when the V1720s saturatemostly for alpha-decays of radon[3]. Test pulses fed into each SCBs are used to measure thetiming offsets between digitizers. The DTM generate both clock and a trigger signals ensuringsynchronization of the digitizers and calibration systems.Figure 2.11: Data flow architecture for DEAP-3600 [3]28Chapter 3Optical model3.1 Hamamatsu R5912 Photo-MultipliersThe Hamamatsu R5912 8 inch diameter, High-Quantum-Efficiency (HQE) Photo-MultiplierTubes (PMTs) are devices capable of measuring individual photons. These 255 devices are thecore of the DEAP-3600 experiment, as they measure the LAr scintillation light (converted intovisible by the wavelength shifter) from interactions that occur inside the detector. The designof these PMTs is shown in Figure 3.1. An 8-inch 700-gram bulb consists of a borosilicateglass window, bialkali photocathode, eleven dynodes and an anode. When an incoming photonis absorbed within the photocathode material on the inner part of the PMT glass, a photo-electron may be liberated. The electric field set between the photocathode and first dynodeextracts the photo-electron into the vacuum space where it is subsequently accelerated towardsthe first dynode. For DEAP-3600, a series of eleven dynodes amplify the signal by increasingthe number of electrons after each dynode stage. Each dynode is a metal plate under highvoltage covered with a material with a high electron emission coefficient, which is critical forcreating avalanches. The total charge is collected at the anode. The avalanche process enablesthe detection of single electrons by converting single photo-electron into a million of electrons.The PMT quantum efficiency is the probability that an impinging photon generates a photo-electron and consequently a readable signal. This process depends on the photon wavelength.29Figure 3.1: Crossectional drawing of the R5912 Hamamatsu HQE photomulti-plier tube. The active photocathode area is approximately 530 cm2[26]The peak of the quantum efficiency for the Hamamatsu R5912 is approximately 420 nm. It ismuch higher than the VUV argon scintillation spectrum that is peaked at 128nm. At that wave-length photons are not detected by any optical device used in the experiment. Therefore usinga wavelength shifter is essential for detecting the liquid argon scintillation light. Furthermorethe VUV light does not go through acrylic and the wavelength shifting must be deposited onthe inside of the acrylic vessel. Figure 3.2 shows the PMT’s quantum efficiency as a functionof an incoming light’s wavelength.The Hamamatsu R5912 was chosen for DEAP-3600 due to its size, relatively low intrinsicradioactivity, and high quantum efficiency.30Figure 3.2: Quantum efficiency of the R5912 Hamamatsu HQE photomultiplieras a function of the incoming photon’s wavelength3.2 AARF Calibration SystemThere are twenty-two light injection sources based on light emitting diodes (LED) implementedin DEAP-3600 for PMT and optical calibration, twenty of which are distributed equally andsymmetrically around the detector. The other two are located in the neck part. The aim ofthis system is in part the calibration of optical parameters of AV, TPB and LAr including theirvariations over the detector. The Acrylic and Aluminum Reflectors and Fibre-Optics system(AARF) is the combination of a 445 nm LEDs, LED driver system and optical fibers transport-ing photons to the edge of the light guides with minimum losses. The light from each AARFfiber is reflected onto a PMT using an aluminum mirror as shown in detail in Figure 3.3. Asmall fraction of the photons, which is approximately 20% of the incoming light, is reflectedon the PMT window and redirected towards the acrylic vessel. The so-called AARF’s PMTdetects the rest of the light. Some of the reflected light gets trapped inside the acrylic, bounc-ing in the material between the reflector system on one side and the TPB layer on the other31Figure 3.3: AARF method of light injection into an acrylic vessel [4]side. Bouncing photons are also detected by the AARF’s PMT if they bounce back into LGafter reflecting from the AV surface, or possibly with one of the neighboring PMTs. A typicaldistribution of the detected amount of light over the PMTs are shown in Figures 3.6 and 3.7.The main advantage of using optical calibration with AARFs is the ability to run the cal-ibration process at any time during the detector construction. It makes it possible to compareoptical parameters of the inner detector with different settings and properties, such as the pres-ence or absence of a TPB layer. Also, it makes it possible to measure the relative quantumefficiency of each PMT over time, which is essential for data analysis.32While working in the calibration mode, AARFs inject the light into the system in shortpulses. The DAQ system creates a trigger for the PMT’s data collection, which copy provokesLED light emission. Each light flash counts as one event, during which a signal from everyPMT is recorded and collected with the DAQ. The intensity of each AARF light source can betuned from zero to a few hundred photons per event.When working with PMT data, one crucial parameter of each PMT is occupancy. It isdefined as the ratio of a number of LED flashes during which a PMT detects at least onephoton over the total number of events during the calibration data run:Occ[PMTi] =NevdetectedNevtotal(3.1)This number identifies the percentage of the events during which PMT[i] detected at leastone photon. Usually, the AARF’s intensity is tuned in such a way that most of the PMTs havearound 5% occupancy. However, even though 95% of the events do not have photons getting toa PMT[i], there is still a small probability that two or more photons will get the PMT[i] duringsome event. This probability can be estimated with Poisson statistics:P(N,λ ) =λNe−λN!(3.2)where P(N,λ ) is the probability of detecting N photo-electrons by some PMT and λ is themean number of photo-electrons observed by the same PMT for a certain AARF intensity:λ =−ln(1−Occ[PMTi]) (3.3)Equation 3.3 represents the relationship between the mean PE number detected by a PMTand its occupancy. A standard occupancy distribution over 255 PMTs is shown in Figure 3.4,together with its distribution over rearranged PMTs based on the distance from the light source.Thus, for 5% PMT occupancy only 2.5% of detected events record correspond to multiple PEevents. After three circles of neighbor PMTs, corresponding to a 60 degree opening angle fromthe AARF, all PMTs record approximately the same occupancy. Photons detected by the PMTsaround the AARF source have most likely been reflected at the acrylic vessel (or wavelength33Figure 3.4: Occupancy distribution among PMTs when one of the AARFs isfired [4]. Data for approximately 2 million events.shifter) interface. Photons detected by the PMT further away from AARF are maximally scat-tered, as all information about the source location has been erased. Such photons may havetravelled along the acrylic vessel or traversed the inner volume of the detector.3.3 Argon BoundaryOptical properties of a liquid/gas argon boundary have not been described in detail yet. Someattempts of characterization are described in this chapter together with a comparison of the datawith simulations.In the perfect case, the best way to study exactly how the surface between liquid andgaseous argon changes along the Z-direction would be to examine an AARF light source lo-cated horizontally at the equator of the acrylic vessel. Then, comparing the optical parametersof the half-filled vessel at slightly different levels and subtracting the signal difference in pulse34Figure 3.5: Schematic diagram of light injection inside the AV and argon reflec-tion processes with low and high LAr levelsdistributions, a detailed analysis could be done. Since such an AARF light injector does notexist in the DEAP-3600, alternative solutions must be found. One of them is using bottomAARF and studying the interaction of an almost vertical photon beam with the surface.Since a significant number of PMTs are located above the LAr level in the acrylic vessel, theliquid/gaseous boundary has a significant impact on recorded signals and photon distributionsin the upper part of the detector. The purpose of this investigation is to analyze different levelsof LAr when one of the bottom AARFs is fired. The index of refraction of liquid Argon issignificantly higher than gaseous argon and Fresnel equations are expected to apply though theliquid-gas interface may not be flat. A schematic representation of the effect of the liquid-gasinterface is shown in Figure 3.5.For each of the 255 PMTs, the value of the mean charge per event was processed indepen-dently for a 40 nanosecond time window at the trigger time.On every charge distribution plot, no matter what the LAr level is, a significant amount ofthe light never gets inside the liquid. It is caught inside the acrylic and TPB layers, reflectinginternally on the walls back and forth until it is detected by one of the nearby PMTs (the green35Figure 3.6: Light distribution for low LAr levels (indicated with red line)) witha light source at the bottom of the vessel. Statistics for 500K eventsFigure 3.7: Light distribution for high LAr levels (indicated with red line) witha light source at the bottom of the vessel. Red line indicates the LArlevel. Statistics for 500K eventscircles and a few other nearby circles around the red AARFs PMT in Figures 3.6 and 3.7).The overall increase in the mean charge function at all PMTs after filling the chamber with36Figure 3.8: Comparison of the light distribution for low LAr level (shown withdash line) and an almost empty vessel for 500K events.LAr is caused by random 39Ar scintillation in liquid Argon. The actual background of the LArsignal is estimated looking at a time window between 2 to 1 microsecond before the pulsespeak. In order to adjust each PMT to itself eliminating a PMT efficiency and gain variation,the empty tank data is taken as the reference point. The ratio of two distributions: with andwithout liquid argon is shown in Figures 3.8 and 3.9 for slightly filled and nearly empty detectorrespectively.The [pq] parameter stays for the mean charge of detected photo-electrons with some partic-ular PMT per one light injection event. Since the amount of photo-electrons is proportional tothe photon number of detected light, plotting this value over all PMT’s spherical coordinates inthe detector (Cos(θ) and φ ) gives an accurate picture how photons are distributed in the vessel.The high intensity of incoming light into bottom PMTs in Figure 3.8 is caused by highreflection of photons from a LAr/GAr boundary. It is the same effect as in the Figure 3.9, butthe flashed region is more extensive due to increasing an area of the light front through crosseddistance. Another important observation from Figure3.9 is the reduction of the amount of lightseen by the first neighboring PMTs around the AARF. The higher index of refraction of LAr37Figure 3.9: Ratio of the light yield for high LAr level (shown by a dashed line)and an almoslt empty vessel for 500K events.compare to GAr[7] indeed reduces the reflections at the TPB optical interface. Its evidencealso can be seen in Figure 3.8, where three bottom PMTs of the six first neighbors around theAARF detect less light than the upper three although an angle distribution of incoming lightmust have axial symmetry.However, comparing the data with Monte-Carlo simulations, an important inconsistencyin the optical model was observed. The total amount of reflected light in the simulations wasgreater than in the real data. Correspondingly, there is a significant difference in the number ofrefracted photons. It brings up the inaccuracy in TPB refractive index used in the simulations.It is also possible that TPB is so thin that the Fresnel formalism fails due to interference. formore precise correlation between the simulations and data, a complex data analysis must beprovided, first attempts of which are presented in the next chapter.38Figure 3.10: Simulation data. Ratio of the light distribution for low LAr leveland an empty vessel.Figure 3.11: Simulation data. Ratio of the light distribution for high LAr leveland an empty vessel.3.4 TPB Optical PropertiesIn order to work in the spectrum of PMT quantum efficiency, a few micrometers layer oforganic 1,1,4,4-tetraphenyl-1,3-butadiene (TPB) wavelength shifter was placed into the internalside of the acrylic chamber. An evaporation method allowed to distribute the material equally39over the surface. However, since evaporation can’t guarantee a perfect spherical symmetryof grown crystal patterns, experimental data analysis is important for calculating the effect ofTPB for optical processes in the DEAP-3600. The refractive index and typical distance betweenscatters are two main parameters which dictate the behavior of theoretical simulations. Monte-Carlo statistics were provided simultaneously in order to compare existing TPB parametersalong with measured ones. Unfortunately, it is impossible to separate all the optical parametersfor a direct measurement from each other, which makes the data analysis very complex.The data used in this study are sets of AARF data with the same flashing condition butthe following detector configurations: before the TPB deposition, after TPB deposition withvacuum in the AV and after TPB deposition with liquid Argon in the AV. In the case of perfectspherical symmetry, this would be enough for first-order results of the TPB optical properties.However, the surface roughness of both the acrylic and TPB materials does not allow for solereliance on that data.Figure 3.12: Qualitative impact of TPB presence: more photons get trapped in-side acrylic light guides.403.4.1 Changes in Light DistributionThe impact of TPB can be explained qualitatively. The additional thin layer between the acrylicand inner vessel medium (liquid argon or gaseous nitrogen) creates more boundaries of mate-rials with different indexes of refraction. In the first case, with no TPB, photons move froma region of lower phase velocity to a higher one. As a result, a significant portion of photonsget reflected and some photons may remain confined within the AV due total internal reflectionuntil they get detected by one of the PMTs close to the AARF. Data shows that a significantfraction of photons are detected inside a 60◦ opening angle zone around the AARF (Figure3.4). By replacing gaseous nitrogen or vacuum in the chamber (refractive index 1.00) with LAr(refractive index 1.23 for 400 nm wavelength [24]), this effect is diminished, which was shownin the previous section. Even though the refractive index of TPB is still under discussion, fromthe first data analysis results it was estimated to be 1.7 for 400 nm wavelength. Thus, eventhough photons can leave acrylic and transfer into a higher phase velocity zone, they may getreflected from the TPB-medium interface.Figure 3.13: Comparison of light distributions with a light source in the AARFID 1. Statistics for 500K events. Each slot represents the ratio ofdetected light after and before the TPB was installed.41Figure 3.14: Comparison of light distributions with a light source in the AARFID 3. Statistics for 500K events. Each slot represents the ratio ofdetected light after and before the TPB was installed.Figure 3.15: Comparison of light distributions with a light source in the AARFID 8. Statistics for 500K events. Each slot represents the ratio ofdetected light after and before the TPB was installed.42In order to find the difference between photon distributions, an average number of detectedphotons was calculated for every PMT for two cases: before and after TPB deposition. Threedifferent AARF locations were used to compare the consistency and symmetry of the pattern.Working with the ratio of the distributions, one can cancel out a PMT’s efficiency, which differsslightly over PMTs. As predicted, the total amount of transmitted light into the vessel decreasedafter depositing TPB crystals. Since the power of the LED source dictates the integral of thetotal charge among the PMTs, it remains the same.3.4.2 Pulse DelayInconsistency in charge distributions over different light sources might be caused by the rough-ness of the TPB layer and its scattering distance. Comparison of the shortest times, in whichphotons can travel between PMTs, can be a useful method for estimating TPB optical parame-ters. The electronics used in the DEAP-3600 allows for detection of the photon’s arrival timewith a precision of 1 ns driven by the PMT transit time spread. Applying the pulse-countingmethod pulse-time distribution can be recorded for every PMT. Then, fitting the front partof the distribution with a Gaussian distribution, the shortest time was estimated at the half-maximum point. The sample of all the pulse distributions by PMT ID is shown in Figure 3.16.The normalization was done assuming zero time difference of arriving photons to the AARF’sPMT, though this method has a considerable uncertainty due to the saturation effect. Figures3.17-3.19 show the impact of the TPB on the travelling time of photons from one of AARF’sPMT, serving as the light injector.After analyzing data for three different light sources (Figures 3.17, 3.18, 3.19), the firstconclusion drawn is that photons not in the direct vicinity or line of sight of the AARF sourcespend more time in the AV after TPB deposition. It does not affect the first two circles ofneighbors, which detect photons reflected in acrylic. The difference for light transmitted intothe chamber is significant. Not only bouncing, but also scattering affects this time as well.When transiting into the AV with a wider spectrum of angle distribution, the amount of paths43Figure 3.16: Sample of total amount of incoming pulses over time distribution.The time scale is the difference between PMT reading and trig-ger time. PMTs are rearranged based on the distance to the lightsource. Statistics for 500K events. AARF light source ID 1.due to scattering and TPB roughness between two PMTs increases. High dependency of timefluctuations over the position of the light source proves the fact that the reflection angle of thelight beam from the AARF’s PMT varies significantly in different light guides. However, insome regions data is consistent for different light sources. In some zones of the longest timedelays the total amount of detected photons decreases. Unfortunately, it is still unclear whetherthis effect is the consequence of high dependency on the refraction angle over the TPB surfaceor the evidence of more complex effects such as thin film interference.Based on the charge and time analysis, together with the use of Monte-Carlo simulationmethods, the TPB refractive index was estimated as 1.7 for a 400 nm wavelength, which isconsistent with Reference [7]. However, studying the optical parameters of the TPB for a VUVspectrum is still in the active phase. Meanwhile, the main optical parameters estimated for theDEAP-3600 are listed in Table 3.1.44Figure 3.17: Average time delay for first incoming photons after implementingTPB over all PMTs. Statistics for 500K events. AARF light sourceID 1.Parameter Estimated valueLAr index of refraction 1.25LAr scattering length ∼10 [m]TPB average thickness 3 +/- 0.02 [um]TPB index of refraction ∼1.7TPB scattering length 3 [um]Table 3.1: Current estimation of main optical parameters for the DEAP-3600inner detector45Figure 3.18: Average time delay for first incoming photons after implementingTPB over all PMTs. Statistics for 500K events. AARF light sourceID 3.Figure 3.19: Average time delay for first incoming photons after implementingTPB over all PMTs. Statistics for 500K events. AARF light sourceID 8.46Chapter 4SummaryThe main result of this work is a summary and a qualitative description of the DEAP-3600optical calibration, along with the first attempts of estimating some physical values of usedmaterials in the inner detector. The results shown in the section 3.3 emphasize an inaccuracyof current estimations of TPB optical parameters, such as index of refraction. However, thedata justified the assumption of a glassy LAr/GAr boundary model, used in the Monte-Carlosimulations. The data also demonstrated a presence of a much higher scattering rate in TPBthan was predicted. Comparison of different LED sources proved an inconsistency in angulardistributions of incoming light. The detailed study of the TPB optical parameters is still in theactive phase. This work showed the estimation of its index of refraction to be greater than 1.5.More accurate estimation of this number requires a complex data analysis and deep usage ofthe Monte-Carlo simulations.47Bibliography[1] Peter AR Ade, N Aghanim, M Arnaud, M Ashdown, J Aumont, C Baccigalupi, AJ Ban-day, RB Barreiro, JG Bartlett, N Bartolo, et al. Planck 2015 results-xiii. cosmologicalparameters. Astronomy & Astrophysics, 594:A13, 2016.[2] P. A. Amaudruz, M. Baldwin, M. Batygov, B. Beltran, C. E. Bina, D. Bishop, J. Bonatt,G. Boorman, et al. First Results from the DEAP-3600 Dark Matter Search with Argon atSNOLAB. Physical Review Letters, 2018.[3] P.-A. Amaudruz, M. Baldwin, M. Batygov, B. Beltran, C.E. Bina, D. Bishop, J. Bonatt,G. Boorman, M.G. Boulay, B. Broerman, et al. Design and construction of the DEAP-3600 dark matter detector. Astroparticle Physics, 108:1–23, 3 2019.[4] P-A Amaudruz, M Batygov, B Beltran, CE Bina, D Bishop, J Bonatt, G Boorman,Mark Guy Boulay, B Broerman, T Bromwich, et al. In-situ characterization of the hama-matsu r5912-hqe photomultiplier tubes used in the deap-3600 experiment. Nuclear In-struments and Methods in Physics Research Section A: Accelerators, Spectrometers, De-tectors and Associated Equipment, 922:373–384, 2019.[5] P-A Amaudruz, M Batygov, B Beltran, J Bonatt, K Boudjemline, Mark Guy Boulay,B Broerman, JF Bueno, A Butcher, B Cai, et al. Measurement of the scintillation timespectra and pulse-shape discrimination of low-energy β and nuclear recoils in liquid argonwith deap-1. Astroparticle Physics, 85:1–23, 2016.48[6] Claude Amsler, V Boccone, A Büchler, R Chandrasekharan, C Regenfus, and J Rochet.Luminescence quenching of the triplet excimer state by air traces in gaseous argon. Jour-nal of Instrumentation, 3(02):P02001, 2008.[7] Ashu-Kumar Bansal and Alfons Penzkofer. Spectroscopic and travelling-wave lasingcharacterisation of tetraphenylbenzidine and di-naphtalenyl-diphenylbenzidine. AppliedPhysics B, 91(3-4):559–569, 2008.[8] Vitalia Evgenievna Baranova, Pavel Fedorovich Baranov, Sergey V Muravyov, andSergey V Uchaykin. Fluxgate magnetometer for measuring ultra low magnetic induction.In Measurement in Research and Industry: Proceedings of XXI IMEKO World Congress,pages 1–4, 2015.[9] P Benetti, F Calaprice, E Calligarich, M Cambiaghi, F Carbonara, F Cavanna, AG Cocco,F Di Pompeo, N Ferrari, G Fiorillo, et al. Measurement of the specific activity of 39ar innatural argon. Nuclear Instruments and Methods in Physics Research Section A: Accel-erators, Spectrometers, Detectors and Associated Equipment, 574(1):83–88, 2007.[10] George R Blumenthal, SM Faber, Joel R Primack, and Martin J Rees. Formation ofgalaxies and large-scale structure with cold dark matter. Nature, 311(5986):517, 1984.[11] Roelof Bottema and Jeroen PE Gerritsen. An investigation of the structure and kinemat-ics of the spiral galaxy ngc 6503. Monthly Notices of the Royal Astronomical Society,290(4):585–597, 1997.[12] M G Boulay and the Deap Collaboration. DEAP-3600 Dark Matter Search at SNOLAB.Journal of Physics: Conference Series, 375(1):12027, 2012.[13] Mark G Boulay and A Hime. Direct wimp detection using scintillation time discrimina-tion in liquid argon. arXiv preprint astro-ph/0411358, 2004.[14] B Broerman, Mark Guy Boulay, B Cai, D Cranshaw, K Dering, S Florian, R Gagnon,P Giampa, C Gilmour, C Hearns, et al. Application of the tpb wavelength shifter49to the deap-3600 spherical acrylic vessel inner surface. Journal of Instrumentation,12(04):P04017, 2017.[15] Ori Cheshnovsky, Baruch Raz, and Joshua Jortner. Emission spectra of deep impuritystates in solid and liquid rare gas alloys. The Journal of Chemical Physics, 57(11):4628–4632, 1972.[16] Douglas Clowe, Marua Bradacˇ, Anthony H. Gonzalez, MaximMarkevitch, Scott W. Ran-dall, Christine Jones, and Dennis Zaritsky. A Direct Empirical Proof of the Existence ofDark Matter. The Astrophysical Journal, 648(2):L109–L113, 9 2006.[17] Planck Collaboration, P. A. R. Ade, N. Aghanim, M. Arnaud, M. Ashdown, J. Aumont,C. Baccigalupi, et al. Planck 2015 results. XIII. Cosmological parameters. 2 2015.[18] Jan Conrad. Indirect Detection of WIMP Dark Matter: a compact review. 11 2014.[19] P Cushman, C Galbiati, DN McKinsey, H Robertson, TMP Tait, D Bauer, A Borgland,B Cabrera, F Calaprice, J Cooley, et al. Snowmass cf1 summary: Wimp dark matter directdetection. arXiv preprint arXiv:1310.8327, 2013.[20] F Duncan, AJ Noble, and D Sinclair. The construction and anticipated science of snolab.Annual Review of Nuclear and Particle Science, 60:163–180, 2010.[21] Richard J Gaitskell. Direct detection of dark matter. Annu. Rev. Nucl. Part. Sci., 54:315–359, 2004.[22] P Gorel. Search for dark matter with liquid argon and pulse shape discrimination: Resultsfrom deap-1 and status of deap-3600. arXiv preprint arXiv:1406.0462, 2014.[23] Emily Grace, Alistair Butcher, Jocelyn Monroe, and James A Nikkel. Index of refrac-tion, rayleigh scattering length, and sellmeier coefficients in solid and liquid argon andxenon. Nuclear Instruments and Methods in Physics Research Section A: Accelerators,Spectrometers, Detectors and Associated Equipment, 867:204–208, 2017.50[24] Emily Grace, Alistair Butcher, Jocelyn Monroe, and James A Nikkel. Index of refrac-tion, rayleigh scattering length, and sellmeier coefficients in solid and liquid argon andxenon. Nuclear Instruments and Methods in Physics Research Section A: Accelerators,Spectrometers, Detectors and Associated Equipment, 867:204–208, 2017.[25] Kim Griest. The Search for the Dark Matter: WIMPs and MACHOs. Annals of the NewYork Academy of Sciences, 688(1):390–407, 6 1993.[26] Hamamatsu. Photomultiplier Tubes Basics and Applications THIRD EDITION (Edition3a). Technical report, 2007.[27] T Heindl, T Dandl, M Hofmann, R Krücken, L Oberauer, W Potzel, J Wieser, and A Ul-rich. The scintillation of liquid argon. EPL (Europhysics Letters), 91(6):62002, 2010.[28] Richard J. Hill and Mikhail P. Solon. WIMP-nucleon scattering with heavy WIMP effec-tive theory. 9 2013.[29] Karl Hille. Hubble Finds an Einstein Ring. 2018.[30] Akira Hitachi, Tan Takahashi, Nobutaka Funayama, Kimiaki Masuda, Jun Kikuchi, andTadayoshi Doke. Effect of ionization density on the time dependence of luminescencefrom liquid argon and xenon. Physical Review B, 27(9):5279, 1983.[31] Bradley J Kavanagh and Anne M Green. Improved determination of the wimp mass fromdirect detection data. Physical Review D, 86(6):065027, 2012.[32] JD Lewin and PF Smith. Review of mathematics, numerical factors, and corrections fordark matter experiments based on elastic nuclear recoil. Astroparticle Physics, 6(1):87–112, 1996.[33] JG Mitchell. The argon-40argon-39 method for potassium-argon age determination.Geochimica et Cosmochimica Acta, 32(7):781–790, 1968.51[34] Kenzo Nakamura. Hyper-kamiokandea next generation water cherenkov detector. Inter-national Journal of Modern Physics A, 18(22):4053–4063, 2003.[35] A Neumeier, M Hofmann, L Oberauer, W Potzel, S Schönert, T Dandl, T Heindl, A Ul-rich, and J Wieser. Attenuation of vacuum ultraviolet light in liquid argon. The EuropeanPhysical Journal C, 72(10):2190, 2012.[36] Jon M Olley, Andrew Murray, and Richard G Roberts. The effects of disequilibria in theuranium and thorium decay chains on burial dose rates in fluvial sediments. QuaternaryScience Reviews, 15(7):751–760, 1996.[37] R. D. Peccei and Helen R. Quinn. CP Conservation in the Presence of Pseudoparticles.Physical Review Letters, 38(25):1440–1443, 6 1977.[38] Chung-Lin Shan. Determining ratios of wimp-nucleon cross sections from direct darkmatter detection data. Journal of Cosmology and Astroparticle Physics, 2011(07):005,2011.[39] Ian P. Stern, ADMX, and ADMX-HF ADMX-HF collaborations. Axion dark mattersearches. In AIP Conference Proceedings, volume 1604, pages 456–461. American Insti-tute of Physics, 2 2014.[40] Teresa Marrodan Undagoitia and Ludwig Rauch. Dark matter direct-detection experi-ments. Journal of Physics G: Nuclear and Particle Physics, 43(1):013001, 2015.[41] Tjeerd S van Albada, John N Bahcall, K Begeman, and R Sancisi. Distribution of darkmatter in the spiral galaxy ngc 3198. The Astrophysical Journal, 295:305–313, 1985.[42] Steven Weinberg. A New Light Boson? Physical Review Letters, 40(4):223–226, 1 1978.52


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items