Far-Infrared Detectors and Mixers Jonas Zmuidzinas Caltech

1
Far-Infrared Detectors and MixersJonas
ZmuidzinasCaltech
2
Z-Spec wideband direct detection spectroscopy
WaFIRS Grating 62 48 3.3 cm

• Parameters
• ? 1.0 to 1.5 mm (CSO)
• 160 NTD-Ge micromesh bolometers
• ?? 900 MHz, ?v 1,000 km s-1
• Compact, stackable waveguide-coupled diffraction
  grating technology demonstration for future
  instruments (e.g., BLISS, SAFIR, )

CaltechBret Naylor, Jonas Zmuidzinas CardiffPete
r Ade CEA (France)Lionel Duband ColoradoJames
Aguirre, Lieko Earle, Jason Glenn, Phil Maloney,
Corey Wood JPLJamie Bock, Matt Bradford, Hien
Nguyen ISAS (Japan)Hideo Matsuhara
3He/4He Fridge
60 mK ADR (Salt pill courtesy Peter Timbies
group)
3
Technology 2-D Waveguide-Coupled Diffraction
Grating
Bolometers are mounted on keys with reflecting
backshorts
The parallel plate grating diffracts and focuses
incoming radiation
Waveguide bend blocks couple radiation from the
parallel plate waveguide to the bolometers
Single broadband corrugated input feed
4
Z-Spec high-redshift measurements
Cloverleaf QSO at z2.55 7.9 hours with Z-Spec at
CSO
Rest Wavelength ?m
434
372
325
289
Raw Spectrum
CO J9-gt8
CO J8-gt7
CO J6-gt5
CO J7-gt6

• Cloverleaf host galaxy
• A powerful lensed system, originally detected in
  submillimeter (redshifted dust) by Barvainis et
  al. (1992). CO 4-3 and 7-6 detected with IRAM
  30m and PdB interferometer (same group in 1994).
• Z-Spec at CSO
• 3 new lines including 2 highest-J transitions!

Line Significance
Frequency GHz
5
APM 08279 _at_ z3.9 with Zspec
6
BLISS concept

• Polarizer and dichroics couple a complete 38-433
  micron spectrum from a point source.
• Titanium suspension supports entire assembly at
  50 mK
• 4224 TES bolometers
• Fits within 50 cm x 50 cm x 50 cm volume
• -gt cold mass 30-35 kg

4 echelle modules 25cm

Could stand alone as the far-IR instrument OR
complement SAFARI.
6 WaFIRS, max36cm
7
BLISS Concept Goal Provide a complete 48-433
micron spectrum of a distant galaxy single action
5 bands, two beams, each with R700 modules (10
modules total) (Reduced from original BLISS (6
bands, R1000) to reduce mass)
Band ?min ?max L W H Ndet Spatial pitch Spectral pitch
?m ?m cm cm cm (both) ?m ?m
Cross-dispersed echelle modules (waveguide tolerances challenging, but ruled gratings easy) Cross-dispersed echelle modules (waveguide tolerances challenging, but ruled gratings easy) Cross-dispersed echelle modules (waveguide tolerances challenging, but ruled gratings easy) Cross-dispersed echelle modules (waveguide tolerances challenging, but ruled gratings easy) Cross-dispersed echelle modules (waveguide tolerances challenging, but ruled gratings easy) Cross-dispersed echelle modules (waveguide tolerances challenging, but ruled gratings easy) Cross-dispersed echelle modules (waveguide tolerances challenging, but ruled gratings easy) Cross-dispersed echelle modules (waveguide tolerances challenging, but ruled gratings easy) Cross-dispersed echelle modules (waveguide tolerances challenging, but ruled gratings easy)
1 38 67 10.3 6.6 6.0 960 306 193
2 67 116 18.0 11.5 11 960 535 337
WaFIRS modules (Compact, where conventional echelle too large massive) WaFIRS modules (Compact, where conventional echelle too large massive) WaFIRS modules (Compact, where conventional echelle too large massive) WaFIRS modules (Compact, where conventional echelle too large massive) WaFIRS modules (Compact, where conventional echelle too large massive) WaFIRS modules (Compact, where conventional echelle too large massive) WaFIRS modules (Compact, where conventional echelle too large massive) WaFIRS modules (Compact, where conventional echelle too large massive) WaFIRS modules (Compact, where conventional echelle too large massive)
3 116 180 14.9 14.9 3 768 873 140
4 180 280 23.1 23.1 4 768 1353 216
5 280 433 35.9 35.9 4 768 2097 336
4224 detectors total (TES bolometers) Fits inside
50cm x 50 cm x 50 cm envelope.
8
Low-G TES bolometers for SPICA

• Extending silicon-nitride micromesh technology
  Reducing thermal conductance with long legs--700
  ?m X 0.5 ?m X 0.5 ?m. Suitable for 1-D or 2-D
  formats
• Measured G corresponds to 4e-19 W Hz-1/2 at 220
  mK, confirmed with electrical measurements.
• When cooled to 70 mK, this G corresponds to an
  NEP of 6e-20 W Hz-1/2. (G measured, NEP
  measurement underway.)
• Close to BLISS requirement !
• (but a long way from a flight system)
• Matt Kenyon et al.

NEP(GT2)1/2
9
Bolometer Performance Recent HistoryRichards
Law
1995 Bolocam 1999 Zspec
10
The Quantum Capacitor Detector A Single Cooper
Pair Box Based Readout for Pair Breaking
Photo-detectors (Echternach et al.)
Predicted NEP x wavelength of QCD detector for
the expected loads of SPICA
Measurement of tunneling rates as a function of
qp injection rate onto ground plane of SCB via
SIN junction compared to theoretical prediction.
The only fit parameter is tqp 250ms.

• Nqp? h? /? generated by antenna coupled
  radiation in a reservoir
• Quasiparticles tunnel to and SCB island, which
  forms a trap of depth ?E?EC-EJ/2
• Capacitance of the island changes by CQ
  (4EC/EJ)(Cg2/C?) ?eo/(?eo?oe)
• Capacitance change produces phase shift
  ???2CQ/(?oZoCC2)

Real time phase measurement on SCB. Jumps of 150
degrees are due to individual qp tunneling events
11
The revolution in submm array technologyRichards
Second Law
12
UKs SCUBA 2 return of the superconducting
bolometer !

• 5000 close-packed TES pixels per band
• Voltage-biased TES Irwin 1995 (Stanford)
• SQUID TDM mux Chervenak et al 1999 (NIST)
• Credit W. Holland, UKATC

Niobium Flex Cable
Ceramic PCB
40x32 sub-array
SQUID Series Array Amplifiers
SCUBA 2 SQUID TDM mux wafer (NIST)
Ribbon Cables to Room Temperature
13
SCUBA 2 pixel design
Pixel size limited by size of SQUID mux unit cell
14
SCUBA 2 Multiplexer
Its getting complicated
15
GSFC BUG TES design

• TES leads come out the sides
• avoids bump bonding, but this limits scaling,
  mosaicing
• bump-bonded version being developed

16
Simplify ! Use kinetic inductance effect - MKID
Day et al. 2003, Nature 425, 6960 (2003)
17
Single resonator IQ readout
I A cos ? Q A sin ?
CPW coupler

• TWO outputs (I, Q)
• changes in frequency,
• dissipation (Q) are measured simultaneously

CPW
18
Frequency multiplexing
2-12 GHz HEMT
19
Digital wideband, multi-frequency readout

• Commercial board
• Dual ADC DAC
• 400 MSPS
• 14/16 bits
• Virtex-5 FPGA
• Easily handles several hundred readout channels
• Size 75 x 150 mm

Use IQ mixers for up/downconversion
20
SQUID Time Domain Multiplexing (NIST)
Multiplexing factor is set by bandwidth of series
array SQUIDs
21
Microwave SQUID TES-mux prototype (Mates et al.,
2008)

• 32 resonators (quarter-wave CPW)
• Simple and gradiometric SQUIDs
• Inductively and directly coupled

8 mm
inductively coupled
directly coupled
resonator
filter
SQUID
Niobium on Si/Si02
22
Antenna-coupled submm MKID
23
Heritage twin-slot lens
Zmuidzinas LeDuc 1992, IEEE MTT
24
4x4 antenna-coupled MKID array
Camera for CSO 24 x 24 pixels, x 4 bands
(Glenn/Golwala)
25
2007 demonstration camera for CSO
Demonstration camera for CSO 4x4x2 color array,
1.3 0.85 mm
26
April 2007 at CSO
First-light images (G34.3, Jupiter)
Under construction 24x24 4-color camera for
CSO (2300 detectors) Glenn, Golwala, Zmuidzinas,
NSF ATI
FTS response, two-color pixels (no AR coatings on
optics)
27
Frequency noise use amplitude readout ?

• amplitude noise is HEMT limited
• should be possible to approach BLIP
• requires 16x lower amplifier noise temperature

Gao et al. 2007, APL 90, 102507
28
TLS origin of MKID phase noise

• TLS dissipation in qubits Martinis et al., PRL
  95, 21053 (2005)
• Existence of low-energy two-level systems (TLS)
  in amorphous solids known for 30 years
• Anomalous low-temperature behavior of glasses
  (heat capacity conductivity, ultrasonic
  attenuation, etc.)
• Anderson 1972 Phillips 1972
• Random distribution of energy level splittings,
  with DE few K
• Caused by motion of groups of atoms
• Dipole moment couples to E-field
• Geometrical scaling of low-temperature frequency
  shift vs. temperature demonstrates existence of
  thin TLS layer on resonator surface, perhaps
  oxides (Gao et al 2008)
• Geometrical scaling of noise consistent with
  surface TLS layer (Gao et al 2008)
• Semi-empirical noise theory explains scaling,
  allows optimization

29
Two Possible Noise Mechanisms

• Two-level state switching

Phonon absorption or emission
Telegraph noise sz(t), c(t)
Upper state life time T1

• Two-level energy fluctuation

TLS-TLS interaction
Nearest neighbor j
Telegraph noise dE(t), c(t)
30
Optimizing the geometry

• Two-section CPW resonator

• Interdigitated capacitor

Sr10mm, Nb on Si
31
MKIDs for the far-IR

• NEP already approaching the range needed for BLIP
  imaging with a cold aperture (1E-18)
• Expect further rapid progress as a result of
  improved physical understanding of TLS (affects
  noise and Q)
• Radiation coupling needs to be demonstrated.
  There are several efforts now headed in this
  direction
• Cambridge/Cardiff lumped-element KID
• SRON/Delft antenna coupling
• GSFC (Wollack/Moseley) direct absorption in
  microstrip resonator
• Many other possibilities exist, utility depends
  on materials, noise, Q, etc.
• BLIP spectroscopy at NEP of 1E-20 will be a
  challenge
• true for all technologies

32
Cambridge/Cardiff
33
SRON/Delft
34
SRON/Delft
35
Quantum Well Far Infrared Detectors

• Quantum Well Infrared Photodetector (QWIP)
• Side-incidence intersubband absorption
• Background limited infrared performance (BLIP)
  demonstrated up to 93 µm by NRC, Canada
• Appl. Phy. Lett. 86, 231103 (2005)
• Requires low doping low QE
• Quantum Well Intra-Subband Photodetector (QWISP)
• Normal-incidence intrasubband absorption
• dopant impurity potential assisted absorption
• Better far IR performance than QWIP (theory)
• Ting et al, Appl. Phy. Lett. 91, 073510 (2007)
• Preliminary experimental results demonstrated
  strong normal incidence response to 25 µm
• Utilizes standard GaAs processing, CMOS mux

Far IR QWIP
36
(No Transcript)
37
(No Transcript)
38
Keck InstituteforSpace Studies
December 2007
39
Summary

• Ground-based work is giving us a small taste of
  the science possibilities for space far-IR
  spectroscopy
• ALMA, Herschel, CCAT will add to the momentum
• Good progress toward low NEPs needed for
  dispersive spectroscopy with cold aperture
• Low-G bolometers
• Several other interesting concepts being pursued
• All superconducting so far
• Systems integration will be key
• TES bolometer imaging arrays are advancing
• Microwave SQUID multiplexing should increase
  array size further
• Kinetic Inductance Detectors
• Noise mechanism now understood
• Quantitative prediction and optimization is
  becoming possible
• Physical understanding is inspiring new resonator
  designs, new materials
• Expect further rapid progress in NEP
• Far-IR radiation coupling is the next challenge
• Quantum well devices for the far-IR
• QWISP far-IR imaging arrays
• QCL THz lasers
• QWISP mixer ?

Capabilities well beyond Herschel/HIFI are
possible