Single Photon Counting Detectors for Submillimeter Astrophysics:

Concept and Electrical Characterization

John TeufelDepartment of Physics

Yale University

Yale:Minghao Shen

Andrew SzymkowiakKonrad LehnertDaniel Prober

Rob Schoelkopf

NASA/GSFCThomas Stevenson

Carl Stahle Ed Wollack

Harvey Moseley

Funding from NASA Explorer Tech., JPL, GSFC

Overview

• Types of detectors

•Noise and sensitivity in detectors

•What is the Submillimeter?

•The “SQPC” – a high-sensitivity sub-mm detector

•Dark currents and predicted sensitivities of SQPC

• Time scales and saturation effects

• Future Work

Types of Detectors

Coherent• Measures Amplitude & Phase• For Narrow-band Signals• Sensitivity given in Noise

Temperature [K]• Adds a 1/2 photon of noise

per mode• Minimum Noise Temperature:

TQ=hf/2k

• Example: a mixer

Incoherent• Measures only Amplitude• For Broad-band Signals• Sensitivity given by NEP

[W/rt(Hz)]• No fundamental noise limit on

detector• Ideally limited only by photon statistics of signal or

background• Example: a photomultiplier

8

6

4

2

0

<n

>

0.12 4 6 8

12 4 6 8

10hf/kT

Wien Raleigh-Jeans

Average occupancy per mode

In the Wien limit:

1/2 photon per mode of noise is unacceptable!

When to Use an Incoherent Detector

1

1bbhf kTn

e

1n

bb

Photon Counting in Optical

PMTPhotonsSignal Source

Background Radiation

Ntot=(n + ndark)• t

Ntot = Ntot

nbackground + nsource ndark

Rate of detector false counts

n =Rate of incoming photons

Photons

Direct Detection with Photoconductor

Bandpass Filter, B

P = h (n + n )incident signal background

NEP n Bbg background

Background Radiation, e.g. CMB,

Atmosphere...

Signal Source

Typical NEP ~10 W/ Hz-17bg

V

+

-

-

+• •

100 1

100 1

1

GHz THz

m mm

E h meV

Infrared

What is the Sub-Millimeter?

How Many Photons in the Sub-mm “Dark?”

3 K blackbody

10 % BW

single-mode

Photon-counting (background) limit:

see e.g. SPECS mission concept, Mather et al., astro-ph/9812454

Future NASA projects need NEP’s < 10-19 W/rt(Hz) in sub-mm !

NEP ~ h(n )1/2

The SQPC: Single Quasiparticle Photon Counter

•Antenna-coupled Superconducting Tunnel Junction (STJ)

•Photoconductor direct detector•Each Photon with

excites 2 quasiparticles

Nb antennaAl absorber

(Au)

~ 1 m

STJ detector

junction

sub-mmphoton AuNb Al Al

AlOx

2 2Al Nb

Responsivity = 2e/photon = e/ = 5000A/W

•Incident photons converted to current

Lower Idark=> Higher sensitivity

What is measured

Nb antenna

(Au)

STJ detectorjunction

sub-mmphoton

Ultimate Sensitivity

V

•Current readout should not add noise to measurement

•FET or RF-SET should have noise

•RF-SET is fast and scalable

2 falseI e n n

Photocurrent Dark current

TotalN

false darkn I

Integration of RF Circuits, SETs, and sub-mm Detectors

• 16 lithographic tank circuits on one chip

• one of four e-beam fields, with SETs and SQPC detectors, and bow-tie antenna

Sensitivity and Charge Sharing with Amplifier

Q ~ 1000 e-

CSTJ ~ 250 fF CSET ~ 1/2 fF

FET (2SK152; 1.1 nV, 20 pF) RF-SET (30 nV, ½ fF)

Either FET or SET can readout STJ @ Fano limit,

But only SET is scalable for > 50-100 readouts

0.15 e/rt(Hz) 1 x 10-4 e/rt(Hz)

Collects all charge Collects CSET/CSTJ ~ 0.2%

still ~ 3 times better

Experimental Set-up and Testing

•Small area junctions fabricated using double angle evaporation

1 µm

Bow Tie Antenna

Detector

140 µm

•Device mounted in pumped He3 cryostat (T~250mK)

Fig. 2. (a) SQPC detector strip and tunnel junctions are located between two halves of a niobium bow-tie antenna for coupling to submillimeter radiation. A gold quasiparticle trap is included here in the wiring to just one of two dual detector SQUIDs. (b) Close-up view of detector strip and tunnel junctions made by double-angle deposition of aluminum through a resist mask patterned by electron beam lithography. Pairs of junctions form dc SQUIDs, and critical currents can be suppressed with an appropriately tuned external magnetic field.

1 µm

junction

detector strip

SQUIDloop

quasiparticle trap

antennaantenna

Al/AlOx/Al Junctions: ~ 60 x 100 nm

XB

Detector Junctions form a SQUID 40

20

0

-20

-40

Cu

rren

t [n

A]

4002000-200-400Voltage [mV]

Supercurrent Suppression

3

2

1

0

Cu

rren

t [n

A]

2.01.51.00.50.0Magnetic Field [mT]

4

3

2

1

0

Cu

rren

t [p

A]

280270260250Magnetic Field [mT]

Supercurrent Contributions to Dark Current

Supercurrent

•Cooper pair tunneling affects the subgap current both at zero and finite voltages

•DC Josephson effect:

•AC Josephson effect:2 ( ) Re ( )

( )2

c Jdark

bias

BB

I ZI

V

Zen Ic sin(J t)

21Re[ ( )]

2RF C JP I Z DC biasP IV

V

Zen

SQPC

RF PowerDC Power

cos( )C oI I

( )

2C J

J

I I sin t

eV

*

*Holst et al, PRL 1994

80

60

40

20

0

Cur

rent

[pA

]

4003002001000Voltage [mV]

Magnetic Field Dependence of Sub-gap Current

2 ( ) Re ( )( )

2c J

darkbias

BB

I ZI

V

60

40

20

0C

urr

ent

[pA

]1.51.00.50.0

Magnetic Field [mT]

BCS Predictions for Dark Current

2 2( )

2 2 2BTk

dark on B B

eV eVI e eV Sinh K

eR e TkT

V k T

T=1.6 K

T=250 mK

{} eV

-8

-6

-4

-2

0

2

4

6

8

Cu

rre

nt

[nA

]

-400 -200 0 200 400Voltage [mV]

10-13

10-12

10-11

10-10

10-9

Cu

rre

nt

[A]

2 3 4 5 6 7 8 91

Temperature [K]

Rn= 13.1 kW

Rn= 9 kW

Rn= 47 kW

Thermal Dark Current Measurements

BCS Predicts:

Tc =1.4 K

I @ 50 V

( ) Bkdark

TTI e

20

10

04002000

BCS (357mK) 357 mK 256 mK

I/500

Cu

rren

t [p

A]

Voltage [µV]

1912min ~ 10I pA NEP W Hz

Recombination and Tunneling Times

absorber

lead

(large

volume)

sub-mmx-ray

Vabs

RN

tunnel

1000 m3 0.01 m3

½

2 s

50 k

2 s

Vabs

thermal recomb ~ 100 s

@ 0.26 K

tunnel ~ VabsRN

tunnel << recomb

so quantum efficiency

is high

at low power:

False count rate = Idark/e = 3 MHz for ½ pA

Saturation: Recombination vs. Tunneling

Current

Power (P)Idark

(or photon rate, N)

Noise

N~ Id/e

rec ~ tunn

Nsat ~ (th/tun) Id/e

Psat~ 0.02 pW; scales as 1/RN

Absorber gap reduced by excess q.p.’s

I ~ P

NEP ~ P1/2

NEP ~ P1/4

I ~ P1/2

Demonstration of an RF-SET Transimpedance Amplifier

Trim gate

Input gate

0.5 fF

Electrical Circuit Model and Noise

2

2

42 n

bdark

kT ee

ZII

R

Shot Noise

Johnson Noise

Amplifier Noise

:

2 dark

Sensitivity

I eI

VRb

en

SQPC

4

610

-16

2

4

610

-15

2

4

Cu

rre

nt N

oise

[A/r

t (H

z)]

100

101

102

103

104

Frequency [Hz]

610

-20

2

4

610

-19

2

4

610

-18

NE

P [W

/rt.(Hz)]

Total Noise Amplifier Noise Johnson Noise Shot Noise

•Problem: Need to couple known amount of sub-mm radiation to detector

•Solution: Use blackbody radiation from a heat source in the cryostat

Future Work: Detecting Photons

10-15

10-14

10-13

10-12

10-11

Po

we

r [W

]

12 4 6 8

10Blackbody Temperature [K]

10-13

10-12

10-11

10-10

10-9

Cu

rren

t [A]

Cryogenic Blackbody as Sub-mm Photon Source

1 cm

104

105

106

107

108

Re

sist

an

ce [W

]

1 10

Temperature [K]

~oTTe

V•Hopping conduction thermistor

•Micro-machined Si for low thermal conduction

Coming Soon: Photoresponse Measurement

T= 1-10K

T= 250 mK

Quartz Window

Si Chip with SQPC

Advantages of SQPC

• Fundamental limit on noise = shot noise of dark current

•Low dark currents imply NEP’s < 10-19 W / rt.Hz

•High quantum efficiency – absorber matched to antenna

• High speed – limited by tunneling time ~ sec

• Can read out with FET, but SET might resolve single ’s

• Small size and power (few m2 and pW/channel)

• Scalable for arrays w/ integrated readout

Summary

• When hf>kTbb, a photon counter is preferred

•In the sub-mm, no such detector exists

•The SQPC would be a sub-mm detector with unprecedented sensitivity

•Contributions to detector noise have been measured and are well-understood

•Photocurrent measurements in near future