JCMT - Stanford University ACT ACT SZ sim. James Beall Sherry Cho ... base temperature SQUID readout...

SLAC Advanced Instrumentation Seminar

Kent IrwinNIST

Superconducting detector arrays: from cosmology to nuclear non-proliferation

ACTJCMT

ACT SZ sim

James BeallSherry ChoRandy DorieseWilliam DuncanLisa FerreiraGene HiltonKent IrwinRob HoranskyBen Mates

Nathan MillerGalen O’NeilCarl ReintsemaDan SchmidtJoel UllomLeila ValeYizi XuKi Won YoonBarry Zink

The Quantum Sensors Project at NIST

• Breakthroughs in superconducting and cryogenic technology are enabling practical large superconducting cameras and spectrometers

Superconducting transition-edge sensorsMicromachiningSQUIDs multiplexersCryocoolers

• Astronomical and particle physics applications have driven the development of superconducting detector arrays, but a broader variety of fields are being impacted

CosmologyX-ray and submm astronomy Quantum informationElectron probe materials analysisSecurityNuclear materials Particle physics

Outline

Applications of TES arrays

WIMPs, neutrino mass

CDMS

Dark matter, neutrinosParticle physics

Nuclear non-proliferation, forensics

γ-ray, alpha, neutronNuclear materials

Concealed weapons, Suicide bombers, nuclear smuggling

THz, γ, neutronSecurity

Electron-probe materials analysis, synchrotron

x-rayMaterials analysis

Quantum key distribution, computing

opticalQuantum information

Planet/star/galaxy formation, WHIM, black holes…

submm, x-ray, γ-rayAstronomy

Dark energy, inflation, neutrino mass…

CMB, submm, x-rayCosmology

G ThermalConductance

E

CHeat

X-ray Thermometer

Capacity

Tem

pera

ture

Time

C

EC

Gτ =

Photon → Heat

Thermal detection

Thermodynamic power noise: NEP2=4kbT2G (W/√Hz)2

Energy fluctuations: ΔErms2 = kbT2C (J)2

Operate at low temperatures (T~ 0.1K to 0.3K) where C, G and thermodynamic fluctuations are small.

Thermal Conductance, G

Incident Radiation, P , E Thermometer, ΔT

Absorber(Heat Capacity C)

Bolometers and calorimeters

P=G×ΔT

E=C×ΔT

Bolometer

Calorimeter

Response time: τ ∝ C/G (s)

Si(Li) EDS

Cou

nts

Energy (eV)4300 4600 4900 5200 55000

500

1000

1500

2000

Ba Lγ1

Ba Lβ2

Ba Lβ3

Ti Kβ1,3

Ba Lβ1

Ti Kα1,2

Ba Lα1 BaTiO3

μcal EDS

X-ray energy resolution

• At 6 keV, conventional semiconductor detectors have energy resolution as good as about 120 eV

• TES x-ray calorimeters can have energy resolution ~ 2 eV – about 60 times better than SiLi.

rms BE k TCΔ =

Bk Boltzmann’s constant

T Temperature of operation

C Heat capacity

Thermodynamic energy fluctuations

• The energy resolution can be a bit better than the thermodynamic energy fluctuations.

Superconducting Transition-Edge Thermometer

Photon → Heat → Resistance

0.02

0.04

0.06

095.8 96 96.2

Temperature (mK)

Res

ista

nce

(Ω)

Transition-Edge Thermometer (TES)

membraneTES thermometer

• A bilayer of a thin superconducting film and a thin normal metal acts as asingle superconductor with atunable Tc - the “proximity effect”

• One example system is Mo-CuRobust and temperature stableMolybdenum Tc ~ .92 KCopper normal

• Sharp• Reproducible <~ 5 mK• Tunable• Robust

0.02

0.04

0.06

095.8 96 96.2

Temperature (mK)R

esis

tanc

e (Ω

)

Copper Molybdenum

Substrate

Proximity bilayers

Other systems: Ti / Al, Mo / Au, W, Ti, etc.

FIBRE 8-pixel bolometer array

Three 1,280-pixel TES bolometer subarrays for SCUBA-2; with UK ATC; SMC

TES arrays2001 – FIBRE 1

2006 – SCUBA-2 subarrays

Stable equilibrium

As the film cools, , and Pjoule increases.

As the film cools, , and Pjoule increases.

R 0→

Thermalconductance

TESThermometer

P sink

Heat Sink

2

JOULEVPR

=2

SINKV PR

=

V

I

If you cool the heat sink to well below the transition temperature of a voltage-biased TES, the TES will self-heat into its transition (electrothermal feedback).

Temperature self-biasing enables for large TES array applications

Temperature self-biasing

SpiderwebsJPL / Caltech / UCB

Popup bolometersGSFC

Silicon “Bricks”NIST /SMC / ATC

Micromachining

Mushroom absorbersNIST / GSFC

SQUID power: ~ nW per ‘on’channel –operate at base temperature

SQUID readout

NIST

SQUID response

Input flux

time

SQ

UID

Out

put

Δ Temperature

Same idea, different orthogonal basis set

SQ

UID

Out

put

Multiplexing basis sets

Frequency division (FDM): different pixels at different frequencies

100-SQUID series array for

~MHz frequency-domain readout with

Berkeley/LBNL

Dissipationless rfSQUID multiplexer

~4-8 GHz

Time division (TDM): different pixels at different times

~MHz TDM SQUID switches

• 2 x 2 array is shown as example of N-row by M-column array

•TDM operation:– each TES coupled to its

own SQ1– TESs stay on all the

time– rows of SQ1s turned on

and off sequentially– SQUIDs are nonlinear

amplifiers, so use digital FB

Each colored block is 1 pixel

Analog time-division multiplexing



20 mm

• Electronics– MCE from UBC – 8

columns per mux card, fully supported system, 10 person years of FPGA firmware development

– NIST/UBC MUX used for SCUBA-2, ACT, SPIDER, BICEP-2, SPUD, ClOVER, ZEUS…

Canadian SCUBA-2Consortium

UBC Room-temperature electronics

One crate controls 1,280 pixels

4K - 100 mK cryogenics

2-stage adiabaticdemagnetization refrigerator(ADR)

• Cryocooler: no liquid cryogens• Push button operation from room temperature to 50 mK• Each magnet cycle gets you to ~ 45 mK and ~ 7 days < 100 mK• NIST-designed cryostat and insert; now commercially available

1 mm

Absorbers for the electromagnetic spectrum + beyond

submm/mm-wave: resonant cavity

absorbers

submm/mm-wave: antennas(Berkeley)

Near IR & optical: TES with antireflection

coating(Nam / NIST / Stanford)

Soft x-ray: thick Bi “mushroom” absorber

Gamma-ray: thick metal foil

Dark matter: thick Si or Ge crystals

Submm astronomy: SCUBA-1

Detection of a gas giant around Fomalhaut

Survey of the galactic centerSCUBA-1

Artist concept

• A collaboration of the UK, Canada, Raytheon, and NIST• SCUBA-2 will consist of 10,240 TES bolometer pixels (half at 450 μm, half at 850 μm) on the James Clerk Maxwell Telescope real soon.

James Clerk Maxwell Telescope


1,280-pixel TES bolometer

1,280-pixel SQUID Multiplexer

bump-bonded subarray (TES+MUX)

Deep-etched

trench (10 μm)

Indium bump

bonds (80 per pixel)

TES

Nitride membrane(0.5 μm)

Quarter-wave Si Brick

1.135 mm

SQUID MUX Wafer

Detector Wafer


THz/submm astronomy: SCUBA-2

CMB

Optical

X-Ray

Cosmic structure

CosmologicalParameters

ν Mass

Dark energyEqn. of state

Cosmology by the coordinated study of cosmic structure

Most of the normal, baryonic matter in clusters lies in the hot X-ray emitting gas (106 - 108 K)

Probe of total mass of galaxy cluster – constrains cosmological parameters

X-rays and & WHIM cosmology

Clusters

SNIa

CMB

ChandraCon-X

Constellation-X

Allen et al., 2004

X-ray MicrocalorimeterSpectrometer (XMS) for Contellation-X, GSFC, NIST, etc.

CMB: SZ SignatureHot electron gas imposes a unique spectral signature

NO SZ Contribution in Central Band

145 GHzdecrement

220 GHznull

270 GHzincrement

1.4°x 1.4°Devlin, ACT collaboration

Z-independent cluster surveys

Columbia Haverford NASA/GSFCPenn Rutgers Univ of TorontoUniv. de CatolicaCardiff

UMASSCUNY

Princeton

Collaboration:NIST

The Atacama Cosmology Telescope - SZ

3,000 TES pixels(256 shown)

South Pole Telescope and APEX-SZAPEX-SZ, 320 pixels

SPT, 960 pixels

Berkeley, Chicago, etc.

(NIST SQUIDs)

• Signature of primordial gravitational waves

• CMB polarimetry microlensing: “cosmic shear”– Probe of expansion history of universe with different

systematics

Future: TES CMB polarimeters for cosmology

Polarization-sensitive TES provide excellent sensivity – need good systematic control

Balloons: SPIDER, EBEX

Ground-based: BICEP-2, SPUD, ClOVER, SPT, ACT, …

WMAP EE mode (HEMTs)

380,000 yratoms form

epn

He

The universe is opaque to photons before 380,000 years: but not to gravitational waves and neutrinos

• Detection of the cosmic gravity wave (CGB) background would probe the inflationary era (10-35 s?)

• Direct probe of inflation vs. cyclic / ekpyrotic models

reionizationrecombination

400 million yrsStars form

Surfa

ce o

f Las

t Sca

tterin

g

CMB polarimetry and primordial gravity waves

Two different types of anisotropies: 1. Decompose the temperature and

gravitational wave distribution at the surface of last scattering into a superposition of plane waves

2. The two types of waves will produce quadropole temperature anisotropies with different spin

3. Integrate the light from all directions scattering off towards the observer

4. The characteristic anisotropies of gravitational waves result in polarization with a curl-like symmetry (BB)

Density-wave anisotropy

Gravitational-wave anisotropy

Density-Wave vs. Gravity-Wave Anisotropies

Credit: Wayne Hu, Chicago

WMAP: EE polarization of the CMB; large angular scale

DASI: EE polarization of the CMB; small angular scale

CMB polarization so far (no gravity waves)

TES polarimeter arrays

Feedhorn array

PolarimeterBackshort

cuta

way

GSFC: metal platelet feedhorns Berkeley: planar antennas

JPL/Caltech: beam-synthesized antennasNIST: silicon micromachined corrugated feeds

γ spectroscopyGross neutronCalorimetryComputer modelsOperator’s reactor historyLimited destructive analysisContainment and surveillance

Current IAEA Methods

About 96 % (>600 tons) of the all the Pu under International safeguards is in spent fuel.

Is it all where it is supposed to be? Is any of it missing?

Gamma-ray: nuclear non-proliferation

Spent fuel in North Korean cooling pond

• Nuclear materials (various isotopes of U, Pu, etc.) emit characteristic γ rays and X rays around 100 keV.

• Industry-standard 100 keV spectrometer is HPGe:– best ΔE ~ 500 eV (can’t resolve some important lines)– active collecting area ~ a few cm2

– max count rate ~ 50 kHz

• Arrays of TES microcalorimeters promise much better energy resolution, comparable collecting area and count rate.

Nuclear materials assay

Efforts at LANL / NIST and LLNL

TES Pu spectrum measured at LANL

1 mm

Mo/Cu TES

Sn absorber

Key measurements:

• Pu isotopes239Pu : 240Pu γ ray

(239Pu = weapon)



Key measurements:


(239Pu = weapon)

• spent reactor fuel(different spectrum):Pu Ka1 : U Ka1x-ray fluorescence ratio(learn total Pu content

without knowing reactor history)


Key measurements:


(239Pu = weapon)

• spent reactor fuel(different spectrum):Pu Ka1 : U Ka1x-ray fluorescence ratio(learn total Pu content

without knowing reactor history)

Need TES arrays:HPGe: t = 15 minutesucal: t = 7 hours

1

2

3

4

5

std. Mo/Cu TES process;add lithographed epoxy posts

spin-coat anotherSi wafer w/ glue

stamp-print glueonto epoxy posts

grey = TEScyan = epoxy post

micro-machined “egg crate”containing Sn absorbers

yellow = Sn absorbersdark blue = glue

glue on absorbers to complete array

Batch fabrication of TES γ-ray arrays

6.25 mm

6.25 mm

MUXed spectrum of 153Gd γ-ray source

600

500

400

300

200

100

0

coun

ts /

5 eV

bin

103.4103.2103.0

energy (keV)

13 coadded pixels51 eV FWHM at 103 keV

gamma-ray arrays

70

60

50

40

30

20

10

0

coun

ts /

4 eV

bin

103.30103.20103.10

energy (keV)

NIST/LANL 1st generation γ-ray arrays

13 coadded pixels

Best pixel

51 eV

27 eV

14-pixel γ-ray array system delivered to LANL

2nd-generation γ-ray arrays

1.5 mm

absorber removedNIST/LANL

Four 66-pixel array chips

Pixel from underneath

Discrimination of 235U from 226Ra“The Kitty Litter Problem”

HEU Detection

simulation

Alpha spectroscopy is a powerful tool for low level actinide measurements (μg – pg)

Approximately 3,000 samples analyzed per year at LANL for the IAEA and others

Poor resolution of conventional alpha spectroscopy (10-30 keV) is a problem:

Can’t split 239Pu/240Pu → drives use of slow and expensive mass spec

Elemental overlaps → slow and expensive wet chemistry to separate

data239Pu240Pu

typical alpha spectrum

typical alpha samples Energy (keV)

4800 4900 5000 5100 5200 5300 5400

Cou

nts

0

500

1000

1500

2000

2500Typical Pu Alpha Spectrum239Pu Alpha Particles240Pu Alpha Particles

High resolution α spectroscopy

Pu sources and mounting hardware

electroplated mixtures of Pu isotopesbelow 2 nCi (Boulder limit)

Completed sensors

4 mm

thermometer andposts for absorber

0.4 mm

absorber removed

High resolution α spectroscopy

300

250

200

150

100

50

0

counts

/ 2

50 e

V b

in

489548904885488048754870energy (keV)

unknown high Eshoulder

209Po spectrum

300

250

200

150

100

50

0

counts

/ 2

50 e

V b

in

531553105305530052955290energy (keV)

2.46 keV resolution

210Po spectrum

in-source straggling (expected)

world-record resolution: 2.46 keV at 5.3 MeV

satellite line -never beforeobserved

Alpha spectroscopy of 209Po/210Po

First splitting of 239Pu/240Pu α lines



CDMS














detector filter SQUID

pixel variety #1

pixel feedline

resonator

coupling T

resonator

Microwave SQUID multiplexer

rf-SQUIDs (flux variable inductors)• unshunted + non-hysteretic• dissipationless• CPW resonators similar to MKIDs, P. Day et al, Nature (2003)• Junctions tune microwave resonances

Chip with one feedline; 32 dissipationlessSQUIDs

HEMT

Measured resonance at 3 flux values

ResultsTransmitted phase as a function of fluxInitial prototypes work well

• 32 resonators on feedline

• Q = 4,000-20,000 at 5 GHz

• Resonator accuracy compatible with > 1,000 SQUIDs per HEMT

• Parasitic power loss (in dielectric) ~ 5 pW

• Stable operation

• Noise dominated by dielectric:

at 100 kHz

• Input low-pass filters block microwaves from detector

• Now working coupling to TESs

00.17n HzμΦ = Φ00.17n HzμΦ = Φ

100 kHz calibration tone

Flux-ramp modulationMicrowave MUX SQUIDsoperate open loop (no feedback, but flux modulated)

• One common “flux ramp” wire for all SQUIDs

• 10 kHz sawtooth current on flux ramp wire with 10 Φ0 amplitude → SQUID response modulated at 100 kHz, above dielectric noise

• Detector signal recovered from phase of modulation

Should be possible to operate >1,000 SQUIDs with two coax cables (input & output), one HEMT, and one flux-ramp wire

Data: recovery of exponential input signal with flux-ramp modulation

Red: input flux

Green: reference baseline

White: signal measured from phase

time

phas

e

Time (ms)0 10

0

0.5

Input and Recovered Flux (

Φ

0

)



CDMS














Electron-probe materials analysis128-pixel x-ray array

X-ray TES on an SEM

Microcalorimeter

Energy (eV)

W MγW Mζ1,2

Si KβW Mβ

W Mα1,2

WSi2 on Si O2Si Kα1,2

EDS spectrometer

1400 1600 1800 220020000

500

1000

1500

2000

1700 1725 1750 17750

400

800

1200

1600

Si Kα3,4Ta Mβ

Si Kα1,2Ta Mα

3 keV beam energy~500 s-1 input rate

400 s live time~0.6 nA beam current

TaSiN thin films on Si (SUNY-Albany)

NIST μcal EDS 74 nm 32 nm 12 nm 7.5 nm 3.5 nmCo

unts

Energy

Thin-Film Analysis

Ref. Geer et. al.

The high energy resolution of the microcalorimeter also provides a high peak-to-background ratio, allowing better thin film & trace element analysis

• Contaminant particles in semiconductor processing lower yield• Low beam energies to localize x-ray production in ~60 nm particle • High energy resolution to resolve overlapping x-ray lines

Low beam energy; x-rays from particle

High beam energy; x-rays from substrate

Nanoscale Particle Analysis

Si(Li) EDS

Cou

nts

Energy (eV)4300 4600 4900 5200 55000

500

1000

1500

2000

Ba Lγ1

Ba Lβ2

Ba Lβ3

Ti Kβ1,3

Ba Lβ1

Ti Kα1,2

Ba Lα1 BaTiO3

μcal EDS

360 400 440 480 5200

200

400

600

μcal EDSTiN

Ti Lβ3,4Ti Lβ1

Ti Lα1,2Ti Lη)

(Ti L

N Kα

Cou

nts

Energy (eV)

SiLi EDS

Resolving peak overlapsThe need for low-energy analysis of small particles means that many peaks

are overlapping when measured with SiLi

0.7 wt. % Cu 0.7 wt. % Cu measured in WSimeasured in WSi22

thin filmthin film

500 1000 1500 20000

500

1000

1500

2000

2500

0

50

100

850 900 950 1000

Cu-LO-K W-M

W-M

W-M

Si-K

μcal

Cou

nts (

0.16

eV

/cha

nnel

)

Energy (eV)

Cu Lβ1

NIST μcal EDS3 keV, 1.3 nA

500 s-1 input rate1000 s live time

Cu Lα1,2

Trace-Element Analysis

Collaboration with Vartuli and Stevie (Lucent)

The high energy resolution of the microcalorimeter also provides a high peak-to-background ratio, allowing better thin film & trace element analysis

Al Ka1,2

Al oxide particle1460 1480 1500 1520

0

20

40

60

80

100

120

0

100

200

300

Cou

nts

Cou

nts

Energy (eV)

Heatpulse

Al particle

• Chemical bonding state causes small(< 1 eV) shifts in x-ray line position

• Industrially important problem: Alparticles on oxide substrates.

Particle samples provided by Alain Diebold (SEMATECH)

Chemical shift analysis

0

0.2

0.4

0.6

Δ eV

Al oxide Al

Chemical shift map

JCMT - Stanford University ACT ACT SZ sim. James Beall Sherry Cho ... base temperature SQUID readout...

Documents

Transcript of JCMT - Stanford University ACT ACT SZ sim. James Beall Sherry Cho ... base temperature SQUID readout...