1

Detectors&

Readout

History: early days The infrared range has been discovered by

astronomers! Friedrich Wilhelm Herschel, using a prism and

balckened bulb thermometers, detects the infraredsection of the solar spectrum (calorific rays, 1800)

The final demonstration that IR is also EM waves happens a bit later Macedonio Melloni in 1829 develops the

thermomultiplier, a sensitive IR detector. With thissystem he demonstrates that calorific rays have the same nature as light, also demonstrating that theyhave polarization properties exactly like light rays. He names the calorific rays ultrared radiation.

The first astronomical observation is carriedout soon after: IR radiation from the moon is detected by Charles

Piazzi Smyth in Tenerife, using a thermocouple. Healso shows that IR radiation is better detected at higher altitudes.

History: early days The first bolometers were developed for

astronomy, and allowed the first IR spectroscopy of an astronomical source Samuel Pierpoint Langley in 1878 develops the

bolometer: a thin blackened platinum strip, sensitive enough to measure the heat of a cow from a distance of mile.

The detector works because the resistance of the Pt strip changes when heated by the absorbedradiation.

The detector is differential: 4 strips are placed in a Wheatstone bridge but only one is blackenedand exposed to incoming radiation. Common-mode effects are rejected by the bridge and tinyvariations of bolometer resistance can bemeasured.

With his bolometer Langley is able tomeasure the IR spectrum of the sun, discovering atomic and molecular lines.

Old times Further developments:

1915 : William Coblentz uses thermopiles (an improved version of Macedonio Mellonis detector !) to measure the infrared radiation from 110 stars, as well as from planets, such as Jupiter and Saturn, and several nebulae.

1920s : systematic IR observations with vacuumthermopiles (Seth B. Nicholson, Edison Pettit and others): diameters of giant stars

1948: IR observations show that the moon is covered by dust.

1950s: Lead Sulphide photodetectors Johnsonsstar photometry

First Semiconductor bolometers, slicing carbonresistors to make the thermistor (W. S. Boyle and K. F. Rodgers, J . Opt . Soc . Am . 49 :66 (1959))

One generation ago The revolution :

1961: Franck J. Low develops the first cryogenic Ge bolometer, boosting the sensitivity by orders of magnitude.

1960s and ff. bolometers and semiconductorsdetectors with their telescopes are carried tospace using stratospheric balloons and rockets.

Consequence: First sky surveys @ 100 m

1968 First IR ground basedlarge area sky survey (2 m, from Mt. Wilson)

Few decades ago

mm-wave bolometers cooled at 1.5K or 0.3K operating from space

become sensitive enough to measure the finestdetails of the Cosmic Microwave Background.

Breakthrough: The composite bolometer (absorber and thermistor

separated and each optimized independently): N. Coron, P. Richards

2

Circa 1970

Circa 1980

Composite Bolometer(Coron, Richards )

monoliticbolometer(Goddard, ..)

Cryogenic Bolometers For FIR & mm-waves spectroscopy we need very wide band

detectors. Bolometers provide the optimal choice: they are senitivefrom mm-waves to the visible range.

filter(frequencyselective)

FeedHorn(angle selective)

IntegratingcavityRadiation

Absorber (T)

Thermometer(Ge thermistor (R)at low T, or TES)

IncomingPhotons (B)

Fundamental noise sources are Johnson noise in the thermistor( = 4kTRf), temperature fluctuations in the thermistor(( = 4kGT2f), background radiation noise (Tbkg5) needto reduce the temperature of the detector and the radiativebackground.

Load resistor

V

Arno Penzias and Robert Wilson (1965):We get microwavesisotropically fromevery direction of the sky. Its the Cosmic MicrowaveBackground.Nobel Prize in Physics, 1977

F. Melchiorri (high mountain, 1974), . P. Richards et al.(balloon, 1980) and then John Mather et al.(1992) with the FIRAS on the COBE satellite: these microwaveshave exactly a blackbody spectrum

Nobel Prize in Physics, 2006

COBE-FIRAS COBE-FIRAS was a Martin-Puplett

Fourier-Transform Spectrometer withcomposite bolometers. It was placed in a 400 km orbit.

A zero instrument comparing the specific sky brightness to the brightnessof a cryogenic Blackbody

The output was nulled (within detector noise) for Tref=2.725 K

The brightness of empty sky is a blackbody at the same temperature !

The early universe was in thermalequilibrium at high Temperature.

(cm-1) wavenumber

3

Primeval Fireball Additionalevidence

for an earlyhot phase

Srinand et al. Nature 408 931 (2000)

COBE Molecules in cosmic clouds(rotational levels)

)1( zTT o +=

KTo 725.2=

Two decades ago

The spider-web absorber isdevelopedIt minimizes the heat capacity of the

absorberIt minimizes the cross-section to

cosmic rays, while maintaining high cross-section for mm-waves

Spider-Web Bolometers

Absorber

Thermistor

Built by JPL Signal wire

2 mm

The absorber is micromachined as a web of metallized Si3N4 wires, 2 m thick, with 0.1 mm pitch.

This is a good absorber formm-wave photons and features a very low cross section for cosmic rays. Also, the heat capacity isreduced by a large factorwith respect to the solidabsorber.

NEP ~ 2 10-17 W/Hz0.5 isachieved @0.3K

150KCMB in 1 s

Mauskopf et al. Appl.Opt. 36, 765-771, (1997)

1900 1920 1940 1960 1980 2000 2020 2040 2060

102

107

1012

1017Langley's bolometer

Golay Cell

Golay Cell

Boyle and Rodgers bolometer

F.J.Low's cryogenic bolometer

Composite bolometer

Composite bolometer at 0.3K

Spider web bolometer at 0.3KSpider web bolometer at 0.1K

1year

1day

1 hour

1 second

Development of thermal detectors for far IR and mm-waves tim

e re

quire

d to

mak

e a

mea

sure

men

t (se

cond

s)

year

Photon noise limit for the CMB

Crill et al., 2003 BOOMERanG 1998 bolometers, 300 mKThe same kind of bolometer is used now in Planck @100mK

Measured performance of Planck HFI bolometers (0.1K)(Holmes et al., Appl. Optics, 47, 5997, 2008)

=Photonnoiselimit

Multi-moded

4

In steady conditions the temperature rise of the sensor isdue to the background radiativepower absorbed Q and to the electrical bias power P:

The effect of the background power is thus equivalent to an increase of the reference temperature:

Cryogenic Bolometers

PQTTG += )( 0

GQTT

TTGGQTTGP

+=

=

+=

00

00

'

)'()(

T0

Q(pW)0.27K

0.28K

0.26K0 1 2

In presence of an additional signalQ ejt (from the sky)

There is a tradeoff between high sensitivity and fast response. The heat capacity C should beminimized to optimize both.

Using a current biased thermistor toreadout the temperature change:

Cryogenic Bolometers

QTGdt

TdC eff =+

GC

GdQdT

eff

=

+=

2211

221/

/)()(

+===

===

effGTRi

dQdT

TRi

dQdV

TRdTiidRdVdT

TdRTR

T

Small sensorat low

temperature

Responsivity

A large isimportant forhigh responsivity.

Gethermistors:

221/

)()(

+=

=

effGTRi

dTTdR

TRT

110 KT

Cryogenic Bolometers Johnson noise in the thermistor

Temperature noise

Photon noise

Total NEP (fundamental):

Cryogenic Bolometers

kTRdf

Vd J 42

=

( )2222

24

fCGGkT

dfWd

eff

effT

+=

( )( ) dxeex

hcTk

dfWd

x

xBGPh

+=

2

4

32

552

114

dfWd

dfWd

dfVd

NEP PhTJ222

22 1 +

+

=

Again, needof low

temperatureand low

background

Q

1900 1920 1940 1960 1980 2000 2020 2040 2060

102

107

1012

1017Langley's bolometer

Golay Cell

Golay Cell

Boyle and Rodgers bolometer

F.J.Low's cryogenic bolometer

Composite bolometer

Composite bolometer at 0.3K

Spider web bolometer at 0.3KSpider web bolometer at 0.1K

1year

1day

1 hour

1 second

Development of thermal detectors for far IR and mm-waves

time

requ

ired

to m

ake

a m

easu

rem

ent (

seco

nds)

year

Photon noise limit for the CMB

Spider-Web Bolometers

Absorber

Thermistor

Built by JPL Signal wire

2 mm

The absorber is micromachined as a web of metallized Si3N4 wires, 2 m thick, with 0.1 mm pitch.

This is a good absorber formm-wave photons and features a very low cross section for cosmic rays. Also, the heat capacity isreduced by a large factorwith respect to the solidabsorber.

NEP ~ 2 10-17 W/Hz0.5 isachieved @0.3K

150KCMB in 1 s

Mauskopf et al. Appl.Opt. 36, 765-771, (1997)

5

Crill et al., 2003 BOOMERanG 1998 bolometers, 300 mK

Cryogenic Bolometers Ge thermistor bolometers have

been used in many CMB experiments: COBE-FIRAS, ARGO, MAX,

BOOMERanG, MAXIMA, ARCHEOPS

Ge thermistor bolometers are extremely sensitive, but slow: the typical time constant C/G isof the order of 10 ms @ 300mK

Once bolometers reach BLIP conditions (CMB BLIP), the mapping speed can only beincreased by creating largebolometer arrays.

Bolometer Arrays BOLOCAM and MAMBO are

examples of large arrayswith hybrid components (Si wafer + Ge sensors)

Techniques to build fullylitographed arrays for the CMB are being developed.

TES offer the naturalsensors. (A. Lee, D. Benford, A. Golding ..hear Richards..)

Bolocam Wafer (CSO)

MAMBO (MPIfR for IRAM)

SWIPE The Short Wavelength Instrument for the Polarization Explorer Uses overmoded bolometers, trading angular resolution for sensitivity Sensitivity of photon-noise limited bolometers vs # of modes:

3.23.32.5NET Focal Plane (K/sqrt(Hz))=

302515NET (K/sqrt(Hz) ) =

1.61.92.4FWHM (deg) =

835837N det =

1.42.13.3 (mm)

22014590f (GHz)

402515N modes (geom) =

m0.8F =Instrumentm0.4D lens =Bolometric

0.25eff =LSPE - SWIPE

Number of modes actually coupling to the bolometer absorber

SWIPE Overmoded detectors are obtained coupling large area bolomete

absorbers to Winston horns. Example of large-throughput spider-web bolometer (being

developed in Italy, F. Gatti)

SWIPE bolometers will be made also in Cambridge (Withington)

SWIPE Overmoded detectors are obtained coupling large area bolometer

absorbers to Winston horns.

Simulations confirm that about halfof the modes collected by the Winston horn actually couple to the bolometer absorber(in single-polarization detectors).

Simulations by L.Lamagna, G.Pisano

6

EBEX EBEX Focal Plane

Total of 1476 detectors Maintained at 0.27 K 3 frequency bands/focal plane

738 element array 141 element hexagon Single TESLee, UCB

3 mm

5 cm

G=15-30 pWatt/K NEP = 1.4e-17 (150 GHz) NEQ = 156 K*rt(sec) (150 GHz) msec, 3=

150

150 150

150250

250

420

Slide: Hanany

William JonesPrinceton University

for the

Spider Collaboration

The Path to CMBpolJune 31, 2009

Suborbital Polarimeter for Inflation Dust and the Epoch of ReionizationSuborbital Polarimeter for Inflation Dust and the Epoch of Reionization

Spider: A Balloon Borne CMB Polarimeter

Long duration (~30 day cryogenic hold time) balloon borne polarimeter

Surveys 60% of the sky each day of the flight, with ~0.5 degree resolution

Broad frequency coverage to aid in foreground separation

Will extract nearly all the information from the CMB E-modes

Will probe B-modes on scales where lensing does not dominate

Technical Pathfinder: solutions appropriate for a space mission

7

Carbon Fiber Gondola

Attitude Control flywheel magnetometer rate gyros sun sensor

Flight Computers/ACS 1 TB for turnaround 5 TB for LDB

Pointing Reconstruction 2 pointed cameras boresight camera rate gyros

Six single freq. telescopes

30 day, 1850 lb, 4K / 1.4 K cryostat

Constant current bias Very high impedance voltage follower as close as possible to the

detector (inside dewar) Very low noise OP amp amplifier (1 nV/sqrt(Hz)

Ge Bolometer Readouthigh impedance (10 M) detector

To A/D converterand data storageB

+

ib

X 1000

Inside dewar Outside dewar

JFET Preamplifier fordifferential AC bias

AC bias currentis betterbecause the amplifier is usedat a frequencyfar from 1/f noise.

Sine wave or triangle wavebias

Demodulatorneeded.

Differentialbiasing is betterbecausedifferentialamplificationremovescommon interference.

Low noise, high CNRR amplifierneeded.

BAC

-1+1

10M

5 pF

5 pF

+Vb

-Vb

180Hz +

-VACJFETs,

x100Lock-in Vout

REF

IN

Planck is a veryambitiousexperiment.

It carries a complex CMB experiment (the state of the art, a few years ago) all the way to L2,

improving the sensitivity wrtWMAP by at least a factor 10,

extending the frequencycoveragetowards high frequencies by a factor about 10

PLANCKESAs mission to map the Cosmic Microwave Background

Image of the whole sky at wavelengths near the intensity peak of the CMB radiation, with high instrument sensitivity (T/T10-6)

high resolution (5 arcmin)

wide frequency coverage (25 GHz-950 GHz)

high control of systematics

Sensitivity to polarization

Launch: 14/May/2009; payload module: 2 instruments + telescope

Low Frequency Instrument (LFI, uses HEMTs)

High Frequency Instrument (HFI, uses bolometers)

Telescope: primary (1.50x1.89 m ellipsoid)

8

ESA : Jan TauberHFI PI : Jean Loup Puget (Paris)HFI IS : Jean Michel Lamarre (Paris)LFI PI : Reno Mandolesi (Bologna)LFI IS : Marco Bersanelli (Milano)

Almost 20 years of hard work of a very large team, coordinated by:

HFI

LFI

ScientificLaboratories

Satellite

+ subcontractors

NationalAgencies

PI Puget

PI Mandolesi

Ecliptic plane1 o/day

Boresight(85o from spin axis)

Field of viewrotates at 1 rpm

E

M

L2

Observing strategyThe payload will work from L2, toavoid the emission of the Earth, of the Moon, of the Sun

Why so far ?

Good reasons to go in deep space: Atmosphere Sidelobes Stability

In the case of CMB observations, the detectedbrightness is the sum of the brightness from the sky(dominant for the solid anglesdirected towards the sky, in the main lobe) and the Brightness from ground(dominant for the solid anglesdirected towards ground, in the sidelobes).

RA()main lobe

side lobes

FWHM=/D

boresight

+= dRABdRABAW

lobesside

Ground

lobemain

sky ),(),(),(),(

The angular response (beam pattern) RA(,)is usually polarization-dependent

9

Planck HFI polarization sensitive focal plane

Ponthieu et al. 2010

Scan direction

z

NEPb = 15 aW/Hz1/2 -> 70 K/Hz1/2Total NET (bolo+photon) = 85 K/Hz1/2

LFI

LFIPseudo-correlationDifferential radiometerMeasures I,Q,U30, 44, 70 GHz

t

10

Primary

secondary

Focal Plane

Off-axis Dragone Telescope, wide field, good polarizationproperties, 1.89mx1.50m aperture

Thermal performance :Planck collaboration: astro-ph/1101:2023

LFI FocalPlane Unit

Thermal performance :Planck collaboration: astro-ph/1101:2023

Mission :Planck collaboration: astro-ph/1101:2022

Mission : Planck collaboration: astro-ph/1101:2022

11

Real data (from just 15 days of operation)

Planck 2013 CMB anisotropy map and power spectrum data (red dots with error bars)Green line is the best fit to a 6-parameters cosmology model (inflationary -CDM)

A large isimportant forhigh responsivity.

Ge thermistors: Superconducting

transition edgethermistors:

Cryogenic Bolometers

221

)()(

1

+=

=

effGRidT

TdRTR

110 KT

11000 KT

S.F. Lee et al. Appl.Opt. 37 3391 (1998)

Have very low R, so work better at constant voltage. Letswrite in detail the equations:

Voltage-Biased Superconducting Bolometers

[ ]

PCiG

TP

eTTeCiGT

PP

TeCiTeGeTT

dTdR

RT

RVP

TeCiTeGTedTdR

RdRdVPe

TedtdCTeG

RVPe

dtdQTTeTG

RV

RVPeP

TTGR

VP

ii

iiib

tititib

ti

titibti

otibbti

ob

++=

++=

+=

+=

+

+=

+

++=

+++

=+

+++

++

+

2

)()()(2

)()(2

)(22

2

1

)(

)( The effective thermal conductivity is

The first part is the Electro-Thermal Feedback (ETF) part. When P increases (a signal arrives) T increases; this increases the

resistance which in turns decreases the bias power Pb=Vb2/R. As a result the total power (P + Pb) does not decrease as much, and the temperature does not change much.

For a given incoming power, the ETF reduced the temperature change. It is the reverse of what happens in a semiconductor bolometer, where

the negative produces a negative ETF, increasing the temperature change.

But here we measure the bias current at constant voltage. The currentneeded to keep the bias more stable is increased by the ETF. So wedefine

Voltage-Biased Superconducting Bolometers

PCiG

TP

eTTeCiGT

PPi

i

++=

++=

CiGT

PGeff

++=

( )oi

L

GCi

GTP

CiGT

P

L

+

=+

=+

=11

12

The Responsivity is

And using

We get

Defining

We get

Voltage-Biased Superconducting Bolometers

( )PT

TP

VPR

RRV

VPRRV

PRV

Pi b

b

b

b

bbb

111/ 2

2 =====

)1( o

iii

iLGe

CiGT

Pe

PTP

CiGT

PeT

++=

++=

++=

)1(1

)1(11

o

i

bo

ib

b

b

b iLLe

ViLGe

TP

VPT

TP

V

++=

++==

1+=

Lo

iLL

Ve

b

i

++=

11

)1(1

For large ETF (L>>1): the time constant is

reduced wrt the standard one by L+1

For slow signals (

13

PROTOTYPE SINGLE PIXEL - 150 GHz (Mauskopf)Schematic:

Waveguide

Radial probe

Nb Microstrip

Silicon nitride

Absorber/termination

TESThermal links

Similar to JPL design, Hunt, et al., 2002 but withwaveguide coupled antenna

PROTOTYPE SINGLE PIXEL - 150 GHz (Mauskopf)Details:

Radial probe

Absorber - Ti/Au: 0.5 /square - t = 20 nmNeed total R = 5-10 w = 5 m d = 50 m Microstrip line: h = 0.3 m, = 4.5 Z ~ 5

TES

Thermal links

TES Readout

A very low impedance, extremely low current noiseamplifier is needed

SQUID (Superconducting quantum interference device)Offers the perfect solution.

Martedi.

Very resistant: materials are all suitable for satellite and space missions.

Extremely simple cold electronics: one single LNA can be used for103-104 pixels. The rest of the readout is warm.

Very flexible: different materials and geometries can be chosen totune detectors to specific needs.

order of 103-104 pixels read with a single coax

Ease of fabrication: one single layer of material is needed.

Kinetic Inductance Detectors

A possible solution:

Main characteristics:

14

The CPs have zero DC resistance, but the reactance is non-zero and has two distinctcontribution kinetic and magnetic L.

KIDs working principle:

In a superconductor below Tc , electrons can bind to form CPs withbinding energy E=2 =3.5*kbTc .

The total conductivity of the material can beestimated using the two-fluid model

CPs

QPs

The values of ss and sn depend on the densities of QPs and CPs. By measuring them, we can getinformation on nqp .

Js Jn

-is2nCP -is2nQP s1nQP

A better estimate of ss and sn is obtained using the MattisBardeen integrals:

A better theory...

Note that:

Rs decreases exponentially

Xs becomes constant

Xs/ Rs grows exponentially

D. C. Mattis and J. Bardeen, in Phys Rev 111 (1958)

How can we measure the small variations of Lk?The superconductor can be inserted in a resonating circuit with extremelyhigh Q, since:

ss RXQ

The resonator is extremely simple to do, and consists of a shorted length of superconducting line capacitevely coupled to the feedline l/4 resonator

Cc

RQPLkin

Lmag

Cl

KIDs are intrinsically multiplexable:

Unitary transmission off resonance

Q values very large (~106)

Multiplexing

Each resonator acts at the same time as detector and filter

Cnc

RnQPLnkin

Lnmag

Cnl

C1c

R1QPL1kin

L1mag

C1l

C2c

R2QPL2kin

L2mag

C2l

RF carrier (f 1 + f 2 + ... + f n )

Pixel 1, f 1 Pixel 2, f 2 Pixel n, f n

One single amplifier needed!Many potential applications

M. Calvo et al. in Conf. Proc. of 1st International LDB Workshop (2008)E.Andreotti et al. in NIMR A 572 (2008)

How do we actually measure the incoming radiation?

nCP< nCP

QPs

CPs

Suppose a photon hits the detector

If its energy is high enough (h > 2E) it can break CPs

The density of CPs therefore changes

This leads to a variation of Lkin

The same effect can be accomplishedby increasing the temperature of the superconductor

The readout is accomplished bymonitoring the phase of the transmittedsignal

Lf 10

dq

Readout techniqueUsually the phase is redefined and referred to the center of the resonant circle:

This kind of plots can give allthe information regardingresonator parameters

It is also the basis for actualmeasurements of radiation

Remember that:

T

QPn)T(nQP

15

Cryogenic system overview

SCN-CN coax

VNA / IQ mixers

2xDC block2xDC block2x10dB atten

1xDC block1xDC block1x10dB atten

KID

300K

30K

2K

300mK

SCN-CN coax

SS-SS coax

warm amplifier

cold amplifier

KIDs readout system

0.3K 2K

Re(S21)

Im(S21)

DAQ

fsynt

fsynt

PC

DAC

ADC

fsynt

fsynt f0 , fsynt f1...

f0, f1...

f0, f1...

fsynt f0 , fsynt f1...

Single pixel readout systemMultipixel readout system

Both systems share the core components!

0.3K 2K

Re(S21)

Im(S21)

PC

DAC

ADC

fsynt

fsynt

f0, f1...

f0, f1...

fsynt f0 , fsynt f1...

10 mm

8 mm

capacitive coupling

2.5 mm

0.7 mm

Al CPW (200nm)

SiO2 (1m)

Silicon Substrate(0.5mm)

KID chip descriptionMaterial: Aluminium6 resonators of varying length

Substrate:

The dielectric constant is notexactly determined!

Base temperature characterization

All 6 resonances observed!

Typical resonanceamplitude curve

51011 .Q

Effect of temperature variation - 1

Higher T Higher nqp Higher losses

Higher T Lower nqp Lower f0

Amplitude

Phase

Effect of temperature variation - 2

)T(Lf

TOT

10

TOT

kin

LL

=)(L

)T(L)(f)T(f

kin

kin

021

000 =

Quality factor increase

Estimate of kinetic inductance fraction

0.018

16

2D data analysis

A fitting procedure has been developed to estimate the parametersof the resonators and the effect of the IQ mixers

The results are in very good agreement with the data:

Temperature variation - 3

The blue points correspond to the base temperature resonant frequency

We obtain sensitivities of 10-3-10-2 deg/nqpequivalent to 10-9-10-8 deg/Nqp

Optical measurementsSystem modified by adding a filter chain

Polyethilene

window

Fluorogold(400G

Hz low

pass)

Fluorogold+

145GH

z bandpassfilter

Horn KIDGunn diode

Quasiparticle lifetimeTo estimate the absorbed power that induces the signal we stillneed one piece of information:

QP

qpabs

nP

=

QP 30s

When T decreases, the quasiparticlelifetime increases (nQP smaller!)

A possible solution: LEKID

Distributed element KIDs

Lumped element KIDs

Needs some sort of antennaResponse depends on where the photon hits the sensor

C

L

It is possible to tune the meandersto match free space impedance!

ADS simulation

Electrical NEP measurement

Dominant contribution given by the warm readout components!

Still too high for real applications, but:

High sample rate data acquired

( )222

2 1 QPQP

QP

NSNEP +

=

Theoretical limit given by GR noiseis as low as 10-20W/Hz0.5 at 100mK

P. K. Day et al. Lett. Nature 425 (2003)

17

ConclusionsThe KIDs concept has been studied and theoretical

models have been developed to analyze their reponse

The experimental testbench has been completed and characterized

The first chip has been made and thouroughly tested

The first results are very promising

Yet still some open issues

High Q factors even at 300mK multiplexing!

Good agreement with theoretical predictions

First light already seen

Develop a system to reach lower T (dilution fridge?)

Optimize optical coupling LEKID

Thanks for your attention!