A New Multiplexable Superconducting Detector

Transcription

1 A New Multiplexable Superconducting Detector Jonas Zmuidzinas California Institute of Technology Supported by: NASA Code R, A. Lidow Caltech Trustee, Caltech President s Fund, JPL DRDF Caltech Anastasios Vayonakis Ben Mazin (at this workshop) Peter Mason Jonas Zmuidzinas Jet Propulsion Laboratory Peter Day Rick LeDuc

2 Why Superconducting Detectors? Astrophysics Millimeter-wavelength arrays CMB polarization (CMBPOL) Submm/far-IR arrays imaging and spectroscopy of dusty high-z galaxies (SAFIR) Energy resolved photon counting Extract more information from every photon! UV/optical (NHST?) X-rays: imaging spectroscopy of galaxy clusters, AGN, (Con-X ) Dark matter searches, neutrino mass experiments X-ray microanalysis SAFIR Con-X

3 Pair-breaking Detectors (e.g. STJ) Analogous to photoconductors, with mev gap (tunable) Photon-counting with energy resolution in optical/uv/x-ray How to measure quasiparticles? Must separate from Cooper pairs Can use tunnel junction as a filter (STJ) Can trap quasiparticles into TES (zero gap energy) 2 energy quasiparticles (N ~ hν/ ) Cooper pair (2 ~mev) photon hν

4 Advantages of pair-breaking detectors Operate well below T c, not at phase transition physics should be simpler (n.b. HEB/TES not fully understood) Finite gap energy Thermal quasiparticle density scales as n qp ~ exp(- /kt) Heat capacity ~ n qp can use much larger detector volume Quasiparticle lifetime τ qp ~ 1/n qp Fundamental sensitivity set by quasiparticle generation-recombination noise Sensitivity scales as (n qp / τ qp ) 1/2 ~ exp(- /kt) NEP = Nqp τ qp ( 2 η E FWHM = ) 1 4 NEP 2 (2πν) dν 2

5 Energy resolution: Fano limit Ultimate resolution set by quasiparticle creation statistics Energy resolution: E = 2.35 [F ε E] 1/2 F = 0.2 is the Fano factor Photon energy per quasiparticle: ε = 1.7 E = 0.04 ev [E/1eV] 1/2 for Ta (R = 28 at 1 ev 1.2 µm) E = 0.02 ev [E/1eV] 1/2 for Al (R = 56 at 1 ev) STJ s have extra tunnel noise

6 S-Cam (ESTEC - Rando et al., RSI, 2000)

7 High quantum efficiency over a broad band (Peacock et al., ESTEC, A&A Suppl., 1998) 100 Magnesium Fluoride Substrate + Tantalum Film 80 Reflectivity Efficiency (%) Quantum Efficiency Wavelength (nm)

8 Problems with STJ quasiparticle readout Junction fabrication is very challenging! Need ultra-low leakage current Only certain materials combinations have been successful e.g. Nb/AlOx/Nb, Al/AlOx/Al, Ta/Nb/Al-AlOx-Al/Nb/T STJ detectors need uniform magnetic field Fiske modes Each STJ needs separate low-noise amplifier Fairly high impedance devices Use JFET amplifiers Noise margin is small Efficient multiplexing not possible

9 New concept: microwave readout of quasiparticles quasiparticles change the kinetic inductance (surface reactance) of superconductor use a thin-film microwave resonant circuit kinetic inductance influences the resonant frequency quasiparticles can change resonant frequency Measure microwave transmission amplitude and phase Use low-noise HEMT amplifier Frequency domain multiplexing!

10 Surface Impedance of Superconductors Surface resistance drops exponentially as temperature is lowered Surface reactance (kinetic inductance) increases near T c

11 Measure variations in kinetic inductance δx s, R s, n qp all decrease exponentially with temperature δx s, R s have nearly constant response to changes in n qp

12 CPW Resonator Measurements

13 CPW Resonator Measurements , 85dbm, 60mK 1.0 S21 Data and Model Fit for D S 12 log mag S21 2 (db) Fit to Resonator Data - Magnitude GHz f 0 = e+009, Q = f (Hz) x 10 9 Phase (degrees) Fit to Resonator Data - Phase GHz Q = 2 x 10 6! (Al on sapphire)

14 Analysis of Resonance Data derive properties of the superconductors and resonators use Mattis-Bardeen surface impedance

15 Quarter-wavelength resonator: Al on sapphire Note: resonator has positiondependent response, ~ cos 2 (πx/2l)

16 Resonance vs. Temperature (Q c ~ 50,000) mk 2 Transmission (db) 4 6 Q 260 mk 8 I 120 mk Frequency (GHz)

17 IQ readout of amplitude and phase = V cos φ = V sin φ V cos(ωt - φ) = V cos φ cos ωt + V sin φ sin ωt

18 Responsivity: phase shift vs. number of thermal quasiparticles 150 Phase (degrees) Number of thermal quasiparticles (millions)

19 It works!!! Rise time: resonator bandwidth Fall time: quasiparticle decay Nyquist sampled readout

20 5.9 kev X-ray produces 2.5η x10 7 thermal qp mk Phase (degrees) mk Time (µs)

21 Pulse Fitting

22 Pulse tail decay time 1000 Decay time (µs) Temperature (K)

23 Phase Noise Can measure I,Q noise Calculate phase noise power spectrum S θ (ω) Gives NEP and expected energy resolution: NEP 2 (ω) =S θ (ω) ( ητ0 dθ dn qp ) 2 (1 + ω 2 τ 2 0 ), E FWHM =2.355 ( 0 ) 1 4 NEP 2 (2πν) dν 2 ocontributions to phase noise: ohemt amplifier omicrowave synthesizer oreference frequency odevice noise (ultimate limit: GR noise)

24 Noise-Equivalent Power: E ~ 10 ev NEP (W/Hz 1/2 ) Total Amplifier Synthesizer Frequency (Hz)

25 Noise, continued

26 What is currently limiting Q? X-ray pulse device is limited by coupling Test resonator Q increases as width of CPW is decreased inconsistent with ohmic loss or dielectric loss consistent with radiation loss Calculated and measured Q reasonably consistent, within factor of 2 but radiation Q calculation highly idealized Q radiation =3.4(L/s) 2 Significant increases in Q are likely Q of 10 7 or 10 8?

27 What sensitivity can we expect? Understand and eliminate apparent excess noise Improve amplifier noise from 50 K to 5 K NEP ~ W Hz -1/2, E ~ 0.3 ev Responsivity given by dθ/dn qp = α center γqv 1 [µm 3 rad/qp] o Agrees quite well with measured responsivity o Already demonstrated Q up to o Decrease volume (film thickness) o Obtain NEP below W Hz -1/2? Fano-limited E?

28 Frequency-domain multiplexing Lithographically tune each detector to a slightly different frequency Use a single HEMT amplifier to simultaneously read out many ( ) detectors Two microwave (coax) cables to sub-k stage eliminates wiring problem No complex readout electronics inside cryostat! Phase noise of HEMT amplifier, frequency reference are common to all detectors (can reduce or eliminate) RF signal processing electronics can take advantage of rapidly advancing semiconductor technology for wireless communications

29 Frequency-domain multiplexing Transmission (db) Frequency (GHz)

30 RF Prototype Board Block Diagram Phillips SA8028 PLL Honeywell HRF-AT4521 Digital Attenuator Cryostat & HEMT Analog AD8347 IQ Mixer SPI Multiplexer And Logic Level Converter Analog DAC AD8347 Amplifier Gain control I Q Programming from PC Small, low power RFICs readily available (used in cell phones)

31 Prototype RF Board ) Phillips SA8028 PLL: GHz, 101 dbc/hz phase noise 2) Analog Device AD8247 IQ Mixer: GHz, Includes 69.5 db of Gain 3) Voltage Controlled Oscillator 4) Honeywell HRF-AT4521 RF Digital Attenuator (1-31 db of programmable attenuation)

32 32-channel, 2 Msa/s, Σ ADC VME board

33 X-ray absorber with KID readout Similar in concept to X-ray STJs developed at Yale Simultaneous low-noise pulse readout of both CPW resonators Absorber, resonator design need optimization

34 New Mask Design 32x32 Optical Array, 50 µm square pixels, Q=2х10 6, f 0 = 6 GHz Optical and X-ray Devices 3 Layer Design (Absorber, Sensor, and Protect) Resonant Frequencies of 1.8, 6, and 10 GHz Devices with up to 1024 pixels Design Q s Range from 100,000 to 20,000,000 Various trapping geometries and test structures are included.

35 UV-Optical KID Array

36 Future Prospects Straightforward to increase format to Nx32 (N > 32) Larger formats also possible Need to reduce area used by resonator. E-beam lithography? Use resonator as absorber? Already position sensitive. Use microstrip resonator on top of absorber? Use phonon coupling to resonator underneath absorber? Readout electronics with 10 4 channels conceivable. Gain factor of with position-sensitive readouts. Array formats of spatial pixels?

37 SUMMARY Invented and demonstrated a new superconducting detector concept: very simple to fabricate compatible with a wide variety of materials simplifies instrument design leverages wireless communications technology high SNR single photon X-ray detection demonstrated 2-way multiplexing demonstrated A wide range of NASA SEU/Origins applications: Millimeter-wave detectors (CMBPOL) Submillimeter & far-ir arrays (128 x 128; SOFIA, SAFIR) Optical/UV energy resolving arrays (NHST?) X-ray spectroscopy Not commercial needs NASA support