Cryogenic Particle Detectors Instrumentation Frontier Community Meeting (CPAD) Argonne National Laboratory - Jan 9-11, 2013
Total Page:16
File Type:pdf, Size:1020Kb
Cryogenic Particle Detectors Instrumentation Frontier Community Meeting (CPAD) Argonne National Laboratory - Jan 9-11, 2013 Blas Cabrera - Spokesperson SuperCDMS Physics Department Stanford University (KIPAC) and SLAC National Accelerator Center (KIPAC) References: LTD13 (2009) - SLAC http://ltd-13.stanford.edu LTD14 (2011) - Heidelberg http://ltd-14.uni-hd.de LTD15 (2013) - Caltech June 24-28, 2013 1 TES sensors invented by DOE HEP program Original motivation was search for Dark Matter funded by DOE HEP at Stanford + NIST SQUIDs Breakthourgh in 1995 when Kent Irwin suggested voltage bias negative feedback Rapidly implemented in CDMS Dark Matter Rapidly implemented in x-ray astrophysics and materials studies at NIST, Goddard and beyond Implemented in CMB by Adrian Lee in 1996 Implemented in IR-optical-UV sensors in 1998 DOE HEP should take credit for these spin offs CPDA - Cryogenic Particles Detectors Page 2 Blas Cabrera - Stanford University Original Motivation: Neutrinos & Dark Matter ! - wanted better resolution for large mass ! - now for CMB & IR-Optical-Xray ! - for Cosmic Frontier & Intensity Frontier Thermal detectors with NTD Ge sensors (T) Thermal sensors with doped semiconductors (T) Superconducting Tunnel Junctions (E) Superconducting Transition Edge Sensors (T) Superconducting Kinetic Inductance Devices (E) CPDA - Cryogenic Particles Detectors Page 3 Blas Cabrera - Stanford University WIMP Search Sensitivity past future x10 5 yrs courtesy of Vuk Mandic CPDA - Cryogenic Particles Detectors Page 4 Blas Cabrera - Stanford University Discrimination strategies ~10% ~1% 100% CPDA - Cryogenic Particles Detectors Page 5 Blas Cabrera - Stanford University Intrinsic Resolution Phonon and Ions (CDMS, EDELWEISS) Scintillation and Ions/Phonons (XENON, CRESST) Quanta E0 % Equanta Ne/keV NN/keV Noiseamp ΔEe-FWHM ΔEN-FWHM Phonons 100% 1 meV 1000000 1000000 0.1 keV 0.1 keV 0.1 keV Ions 10% 1 eV 300 100 1 keV 1 keV 3 keV Photons 1% 10 eV 10 1 0.01 keV 1 keV 10 keV CPDA - Cryogenic Particles Detectors Page 6 Blas Cabrera - Stanford University NaI -> HPGe -> µcalorimeters Cryogenic Sensors For High-precision Safeguards Measurements weblink CPDA - Cryogenic Particles Detectors Page 7 Blas Cabrera - Stanford University Scope of Cryogenic Detectors Sub-Kelvin Sensor types Transition Edge Sensors (T) Kinetic Inductance Detectors (E) Metallic Magnetic Calorimeters (T) Novel detection techniques Te c h n o lo g i e s Micro-fabrication with superconducting materials Cryogenics using dilution refrigerators and ADRs Multiplexing, SQUIDs, readout, & data analysis Particle absorbers & antennas: physics & design considerations CPDA - Cryogenic Particles Detectors Page 8 Blas Cabrera - Stanford University Large TES arrays progressing 1,280-pixel SQUID TDM multiplexer for the SCUBA-2 weblink MUX chip has 32 columns each with 40 multiplexed SQUIDs 50 X 50 mm2 CPDA - Cryogenic Particles Detectors Page 9 Blas Cabrera - Stanford University Detectors and Physics Detector Physics 3 Insulators - Debye heat capacity ~T at low temperature Conductors - Fermi liquid theory ~T at low temperature Semiconductors - electrons & holes ~ 1eV excitation Superconductors - quasiparticles ~ 1meV excitation Magnetism - paramagnetism and diamagnetism Science applications Neutrino mass experiments (IF) Dark matter searches (CF) Alpha & beta spectroscopy, mass spectroscopy, heavy ions, and neutrons X-ray & gamma spectroscopy in atomic, nuclear, astrophysics & other fields UV-optical-IR single photon detection Bolometers in mm / sub-mm wave for astrophysics, THz applications (CF) CPDA - Cryogenic Particles Detectors Page 10 Blas Cabrera - Stanford University Calorimeter Principle particle Thermal relaxation time: thermometer Thermal conductance t absorber weak thermal link : phonons electrons spins thermal bath tunneling states quasi particles CPDA - Cryogenic Particles Detectors Page 11 Blas Cabrera - Stanford University Semiconducting Thermistors R T Si – ion-implanted (P,B) Ge NTD ( Neutron-Transmutation-Doped) High impedance device particle Ibias Thermistor + JFET - 100 K CPDA - Cryogenic Particles Detectors Page 12 Blas Cabrera - Stanford University Superconducting Transition Edge Sensor (TES) R self regulated working point T Materials Mo/Cu Electro-thermal feedback K. D. Irwin, Appl. Phys. Ir/Au Lett. 66, 1945 (1995) V W X-ray heat input: TES RTES goes up shunt I joule heating decreases fast response time SQUID t CPDA - Cryogenic Particles Detectors Page 13 Blas Cabrera - Stanford University Metallic Magnetic Calorimeter (MMC) M H dc SQUID T Au:Er Au:Yb Ag:Er Bi2Te3:Er main differences to resistive calorimeters: PbTe:Er non-contact readout no dissipation due to readout current LaB6:Er CPDA - Cryogenic Particles Detectors Page 14 Blas Cabrera - Stanford University So what is intrinsic resolution for thermal detectors ? Physical Principles of Low Temperature Detectors Heat capacity C at temperatureRandom T tra nshasport of enenergyergy between he aCTt sink and detector over thermal link G produces fluctuations in the energy The average energy per carriercontent o~f CkT. The magnitude of these can easily be calculated from the fundamental assumption and definitions of statistical So there are N ~ CT / kT carriersmechanics: 2 2 So statistical thermal noise (ΔE )!rmEs rms k=T NkT=C k T C But we can detectFi gsmaller. 2. Fundament alsignal thermodyna miasc fl uctshownuation noise. Physical Principles of Low Temperature Detectors Random transport of energy between heat sink and detector over thermal link G produces fluctuations in the energy content of C. The magnitude of these can easily be calculated from the fundamental assumption and definitions of statistical mechanics: 2 !Erms = kT C Fig. 2. Fundamental thermodynamic fluctuation noise. Fig. 3. Signal can still be measured to high accuracy in presence of CPDA - Cryogenic Particles DetectorsthermodynamicPage fluct u15ation noise Blasby lCabreraooking at- Stanfordthe “corn Universityers”. Here the signal amplitude is equal to the r.m.s. value of the fluctuations. effective fmax that can be substituted for Δf. Since the resolution depends only on the ratio fc//fmax, something can be gained by making the detector as slow as the application can withstand. Detector design usually ends up a complicated tradeoff of internal time constants, thermometer noise, heat Fig. 3. Signal can still be measured capto haciight yaccu, andracy op eiratn pinresg entempce oeratf ure. thermodynamic fluctuation noise by looking at the “corners”. Here the signal amplitude is equal to the r.m.s. valu e of the flTuhcteu atpioronbs.l em of how to filter the signal to get the best possible estimate of the its amplitude in the presence of an arbitrary noise spectrum effective fmax that can be substituted for canΔf. bSien scoe ltvhede res quoliuteti ognen deerpenaldlys for the linear small-signal case if we assume only on the ratio fc//fmax, something can bteh gatai nthede bsyig maknali npgu lthse dsethapecteo ran asd the noise power spectrum are both fixed and slow as the application can withstand. Detector design usually ends up a complicated tradeoff of internal time coknnsotanwtns., tTherhemo nmoietsere ino disie,fferen heat t time bins is correlated, except for the special capacity, and operating temperature. case of white noise, but is uncorrelated in frequency space so long as it is The problem of how to filter th"es tatsigionalnary to ,"g etw thhiech b esmeant posss ibthleat its statistical properties do not change with estimate of the its amplitude in the presence of an arbitrary noise spectrum can be solved quite generally for the linearthe s malsignl-alsig. n alT hcasise miif nwiemal ass ureqme uirement should be met by almost any linear that the signal pulse shape and the noise spyowsterem. sp ect Wruem canare btohthen fi xreedad anildy construct the optimal filter in the frequency known. The noise in different time bins diso comairrelnat, edgi,v exencep tht atfo ri nth et hsep eciithal f requency bin the expected noise is n and the case of white noise, but is uncorrelated in frequency space so long as it is i "stationary," which means that its statistical properties do not change with the signal. This minimal requirement should be met by almost any linear system. We can then readily construct the optimal filter in the frequency th domain, given that in the i frequency bin the expected noise is ni and the Signal and Noise EnergyEnergy Fluctuation fluctuation is not Energy is Resolutionnot energy resolution 1 2 1 ⎛ 22π⇡ffc ⎞ 2 2 GG ΔE = c k T 2C; f = ∆E = ⎜ ⎟ kBT C; fc =c ⎝ ∆Δf ⎠ 2⇡2CπC ✓ ◆ p McCammon CPDA - Cryogenic Particles Detectors Page 16 Blas Cabrera - Stanford University National Institute of Standards and Technology • U.S. Department of Commerce Slide 14/48 Signal & Noise But finite thermalization time and amp noise Then add some noise, and some more time constants so Δ f limited 112 2⇡f 2 G ⎛ 2π fcc ⎞ 2 G ∆ΔEE == kBTT CC; fc f== ⎝⎜ Δ∆ff ⎠⎟ B c 2⇡2CπC ✓ ◆ p CPDA - Cryogenic Particles Detectors Page 17 Blas Cabrera -McCammon Stanford University National Institute of Standards and Technology • U.S. Department of Commerce Slide 15/48 JLowTempPhys(2008)151:5–24 19 Fig. 12 (Color on line) Different designs of TES detectors for different applications Many ways to measure temperature (Kα1 / Kα2 for 55Fe at 6 keV) Transition Edge Sensors (TES) Doped semiconductors - Si and NTD Ge Doped paramagnetism - MMC 1 4 4 1 4 2 (40) 2 2 ⎛ τ 0 ⎞ Fig. 13 E(Color onk line)T C Some of the best results(ΔE) for the= MnkBT KCα line at 5.89 keVΔ (l-r:E