time required to make a measurement (seconds)

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 infrared section 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 this

system he demonstrates that calorific rays have the same nature as light, also demonstrating that they have polarization properties exactly like light rays.

He names the calorific rays “ultrared radiation”.

• The first astronomical observation is carried out soon after:

– IR radiation from the moon is detected by Charles

Piazzi Smyth in Tenerife, using a thermocouple. He

also 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 ¼ mile.

– The detector works because the resistance of the Pt strip changes when heated by the absorbed radiation.

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

measured.

• With his bolometer Langley is able to

measure 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 Melloni’s detector !) to measure the infrared radiation from 110 stars, as well as from planets, such as Jupiter and Saturn, and several nebulae .

– 1920’s : systematic IR observations with vacuum thermopiles (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 – Johnson’s star photometry

– First Semiconductor bolometers, slicing carbon

resistors 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.

– 1960’s and ff. bolometers and semiconductors detectors with their telescopes are carried to space using stratospheric balloons and

rockets.

• Consequence:

– First sky surveys @ λ 100 μm – 1968 First IR ground based

large area sky survey (2 μm,

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 finest details of the Cosmic Microwave Background.

• Breakthrough:

– The composite bolometer (absorber and thermistor separated and each optimized independently):

N. Coron, P. Richards …

Circa 1970

Circa 1980

Composite Bolometer

(Coron, Richards …)

monolitic bolometer (Goddard, ..)

Cryogenic Bolometers

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

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

filter

(frequency selective) Feed

Horn

(angle selective) Integrating

cavity Radiation

Absorber (ΔT)

Thermometer

(Ge thermistor (ΔR) at low T, or TES)

Incoming Photons

(ΔB)

• Fundamental noise sources are Johnson noise in the thermistor (<ΔV

²

> = 4kTRΔf), temperature fluctuations in the thermistor ((<ΔW

²

> = 4kGT

²

Δf), background radiation noise (T

_bkg⁵

) need to reduce the temperature of the detector and the radiative

background.

Load resistor

ΔV

Arno Penzias and Robert Wilson (1965):

We get microwaves isotropically from every direction of the sky. It’s the Cosmic Microwave Background.

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 microwaves have exactly a

blackbody spectrum Nobel Prize in Physics, 2006

COBE-FIRAS

• COBE-FIRAS was a Martin-Puplett

Fourier-Transform Spectrometer with

composite bolometers. It was placed in a 400 km orbit.

• A zero instrument comparing the

specific sky brightness to the brightness of a cryogenic Blackbody

• The output was nulled (within detector noise) for T

_ref

=2.725 K

• The brightness of empty sky is a

blackbody at the same temperature !

• The early universe was in thermal

equilibrium at high Temperature.

σ (cm

^-1

) wavenumber

Primeval Fireball _Additional

evidence for an early

hot phase

Srinand et al.

Nature 408 931 (2000)

COBE Molecules in

cosmic clouds (rotational levels)

) 1

( z T

T = _o +

K

T

_o

= 2 . 725

Two decades ago

• The spider-web absorber is developed

–It minimizes the heat capacity of the absorber

–It 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 micro machined as a web of metallized Si

₃

N

₄

wires, 2 μm thick, with 0.1 mm pitch.

•This is a good absorber for mm-wave photons and

features a very low cross section for cosmic rays.

Also, the heat capacity is reduced by a large factor with respect to the solid absorber.

•NEP ~ 2 10

^-17

W/Hz

^0.5

is achieved @0.3K

•150μK

_CMB

in 1 s

•Mauskopf et al. Appl.Opt.

36, 765-771, (1997)

1900 1920 1940 1960 1980 2000 2020 2040 2060

10

²

10

⁷

10

¹²

10

¹⁷

Langley'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 0.3K Spider web bolometer at 0.1K 1year

1day 1 hour

1 second

Development of thermal detectors for far IR and mm-waves

ti m e r equi red to m a k e a m eas ur em ent ( s ec onds )

year

Photon noise limit for the CMB

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

The 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)

= Photon noise limit

Multi-moded

• In steady conditions the

temperature rise of the sensor is due to the background radiative power 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

P Q

T T

G ( − )

₀

= +

G T Q

T

T T

G G T Q

T G P

+

=

−

⎥⎦ =

⎢⎣ ⎤

⎡ − +

=

0 0

0 0

'

) ' (

) (

T

₀

’

Q(pW)

0.27K 0.28K

0.26K

0 1 2

• In presence of an additional signal ΔQ e

^jωt

(from the sky)

• There is a tradeoff between high sensitivity and fast response. The heat capacity C should be

minimized to optimize both.

• Using a current biased thermistor to readout the temperature change:

Cryogenic Bolometers

Q T

dt G T

C d Δ +

_eff

Δ = Δ

G C dQ G

dT

eff

=

= + τ

ω τ

^{2 2}

1

1

2

1

2

/ ) /

( )

(

ω τ α α

α α

= +

=

= ℜ

=

=

⇒

=

G

eff

T R i

dQ dT T

i R dQ

dV

T RdT

i idR

dT dV T dR T

R T

Small sensor at low

temperature

Responsivity

• A large α is important for high

responsivity.

• Ge thermistors:

2

1

2

/

) ( )

(

ω τ α

α

= + ℜ

=

G

eff

T R i

dT T dR T

R T

10 ⁻ 1

−

≈ K

T

α

Cryogenic Bolometers

• Johnson noise in the thermistor

• Temperature noise

• Photon noise

• Total NEP (fundamental):

Cryogenic Bolometers

df kTR V

d

_J

4

2

Δ =

( )

²

2 2 2

2 4

fC G

G kT

df W d

eff T eff

π

= + Δ

( )

( ) ^{e e dx}

x h

c T k df

W d

x x Ph BG

∫ ₋ ^{− +}

Δ =

2 4

3 2

5 2 5

1 1

4 ε ε

df W d

df W d

df V

NEP d

^{J T Ph}

2 2

2 2

2

1 Δ

Δ + Δ +

= ℜ

Again, need of low

temperature and low background

Q

1900 1920 1940 1960 1980 2000 2020 2040 2060

10

²

10

⁷

10

¹²

10

¹⁷

Langley'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 0.3K Spider web bolometer at 0.1K 1year

1day 1 hour

1 second

Development of thermal detectors for far IR and mm-waves

ti m e r equi red to m a k e a m eas ur em ent ( s ec onds )

year

Photon noise limit for the CMB

Spider-Web Bolometers

Absorber

Thermistor

Built by JPL Signal wire

2 mm

•The absorber is micro machined as a web of metallized Si

₃

N

₄

wires, 2 μm thick, with 0.1 mm pitch.

•This is a good absorber for mm-wave photons and

features a very low cross section for cosmic rays.

Also, the heat capacity is reduced by a large factor with respect to the solid absorber.

•NEP ~ 2 10

^-17

W/Hz

^0.5

is achieved @0.3K

•150μK

_CMB

in 1 s

•Mauskopf et al. Appl.Opt.

36, 765-771, (1997)

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 is of the order of 10 ms @ 300mK

• Once bolometers reach BLIP

conditions (CMB BLIP), the

mapping speed can only be

increased by creating large

bolometer arrays.

Bolometer Arrays

• BOLOCAM and MAMBO are examples of large arrays

with hybrid components (Si wafer + Ge sensors)

• Techniques to build fully litographed arrays for the CMB are being developed.

• TES offer the natural

sensors. (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:

3.2 3.3

2.5 NET Focal Plane

(μK/sqrt(Hz))=

30 25

NET (μK/sqrt(Hz) ) = 15

1.6 1.9

2.4 FWHM (deg) =

83 58

37 N det =

1.4 2.1

λ (mm) 3.3

220 145

90 f (GHz)

40 25

15 N modes (geom) =

m 0.8 F =

Instrument

m 0.4 D lens =

Bolometric

0.25 eff =

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 half of the modes collected by the

Winston horn actually couple to the bolometer absorber

(in single-polarization detectors).

Simulations by L.Lamagna, G.Pisano

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 TES

Lee, UCB

3 mm

5 cm

• G=15-30 pWatt/K

• NEP = 1.4e-17 (150 GHz)

• NEQ = 156 μK*rt(sec) (150 GHz)

• τ = 3 msec,

150

150 150

150 250

250 420

Slide: Hanany

William Jones

Princeton University

for the

Spider Collaboration

The Path to CMBpol June 31, 2009

Suborbital Polarimeter for Inflation Dust and the Epoch of Reionization

Suborbital 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

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 Readout

high impedance (10 MΩ) detector

To A/D converter and data storage

B

+

i

_b

X 1000

Inside dewar Outside dewar

JFET Preamplifier for differential AC bias

• AC bias current is better

because the

amplifier is used at a frequency far from 1/f noise.

• Sine wave or triangle wave bias

• Demodulator needed.

• Differential

biasing is better because

differential amplification removes

common

interference.

• Low noise, high CNRR amplifier needed.

AC

B

-1 + 1

10MΩ 5 pF

5 pF +V_b

-V_b

180Hz

+

-

V_AC

JFETs, x100

Lock-in

^Vout

REF IN

Planck is a very ambitious

experiment.

It carries a complex CMB experiment (the state of the art, a few years ago) all the way to L2, improving the sensitivity wrt WMAP by at least a factor 10, extending the frequency

coverage towards high frequencies by a factor about 10

PLANCK

ESA’s 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 (ΔT/T∼10

^-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)

ESA : Jan Tauber

HFI 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

Scientific Laboratories

Satellite

+ subcontractors National Agencies

PI Puget PI Mandolesi

Ecliptic plane

1 ^o/day

Boresight

(85^o from spin axis)

Field of view rotates at 1 rpm

E M

L2

Observing strategy

The payload will work from L2, to

avoid 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 detected brightness is the sum of the brightness from the sky

(dominant for the solid angles directed towards the sky, in the main lobe) and the

Brightness from ground

(dominant for the solid angles directed towards ground, in the sidelobes).

RA( θ ⁾

main lobe

θ

side lobes

FWHM=

λ

/D

boresight

⎥ ⎥

⎥

⎦

⎤

⎢ ⎢

⎢

⎣

⎡

Ω +

Ω

= ^A ∫ ^{B RA d} ∫ ^{B RA d}

W

lobes side

Ground lobe

main

sky

( θ , ϕ ) ( θ , ϕ ) ( θ , ϕ ) ( θ , ϕ )

• The angular response (beam pattern) RA( θ ^, φ ⁾

is usually polarization-dependent

<<3x10^-6 7x10^-8 srad

1’

<<3x10^-4 7x10^-6 srad

10’

<<0.01 2x10^-4 srad

1^o

<<1 2x10^-2 srad

10^o

<RA

_sidelobes

>

Ω

_mainlobe FWHM

Going to L2 reduces the solid angle occupied by the Earth by a factor 2π/2x10

^-4

=31000, thus

relaxing by the same factor the required off-axis rejection.

1.5Mkm

900km L2

COBE WMAP,

Planck

No day-night changes up there … extreme stability

Planck – HFI polarization sensitive focal plane

Ponthieu et al. 2010

Scan direction

z

NEP _b = 15 aW/Hz ^1/2 -> 70 μK/Hz ^1/2

Total NET (bolo+photon) = 85 μK/Hz ^1/2

LFI

LFI

Pseudo-correlation

Differential radiometer Measures I,Q,U

30, 44, 70 GHz

t

Primary

secondary

Focal Plane

Off-axis Dragone

Telescope, wide field, good polarization

properties, 1.89mx1.50m

aperture

Thermal performance :

Planck collaboration: astro-ph/1101:2023

LFI Focal Plane Unit

Thermal performance :

Planck collaboration: astro-ph/1101:2023

Mission :

Planck collaboration: astro-ph/1101:2022

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

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 Λ-CDM)

• A large α is important for high

responsivity.

• Ge thermistors:

• Superconducting transition edge thermistors:

Cryogenic Bolometers

2

1

2

) ( )

( 1

ω τ α α

= + ℜ

=

G

eff

R i

dT T dR T

R

10

⁻1

−

≈ K

T

α

1000

⁻1

≈ K

T α

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

• Have very low R, so work better at constant voltage. Lets’

write in detail the equations:

Voltage-Biased Superconducting Bolometers

[ ]

P Ci

T G P T e

Te Ci

T G P P

Te Ci

Te G

T e T dT

dR R

T R

P V

Te Ci

Te G

dT Te dR R

dR V d

Pe

dt Te C d Te

R G Pe V

dt T dQ

Te T

R G V R

Pe V P

T T

R G P V

i i

i i

b i

t i t

i t

i b

t i

t i t

b i t

i

o t

b i t b

i

o b

ω δ δ α

δ α ω

δ

ωδ δ δ

δ

ωδ δ

δ δ

δ δ

δ δ

δ δ

δ

ϕ ϕ

ϕ ϕ

ϕ

ϕ ω ϕ

ω ϕ

ω ω

ϕ ω ϕ

ω ω

ϕ ω ω

+ +

=

⎥⎦ ⇒

⎢⎣ ⎤

⎡ + +

=

+

⎥⎦ =

⎢⎣ ⎤

− ⎡

+

⎥⎦ =

⎢⎣ ⎤ + ⎡

+

⎥ =

⎦

⎢ ⎤

⎣ + ⎡

⇒

+

− +

⎥ =

⎦

⎢ ⎤

⎣ + ⎡ +

+

−

= +

−

+ +

+

+ +

+

2

) (

) (

) (

2

) (

) 2 (

) (

2 2

2

1

) (

)

(

• The effective thermal conductivity is

• The first part is the Electro-Thermal Feedback (ETF) part.

• When δP increases (a signal arrives) T increases; this increases the resistance which in turns decreases the bias power P_b=V_b²/R. As a result the total power (P + P_b) 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 α produces a negative ETF, increasing the temperature change.

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

Voltage-Biased Superconducting Bolometers

P Ci

T G P T e

Te Ci

T G P P

i

i

δ

α ω δ

δ α ω

δ

^{ϕ ϕ}

+ +

=

⎥⎦ ⇒

⎢⎣ ⎤

⎡ + +

=

⁻

C i

T G

G

_eff

₌ P α _{+ +} ω

( )

i

o

L G

i C G T

P C

i G

T P

L ω ωτ

α ω

α

ω = +

+ + =

= 1 1

• The Responsivity is

• And using

• We get

• Defining

• We get

Voltage-Biased Superconducting Bolometers

( )

P T T

P V P

R R

R V V P

R

R V

P R V

P

i

_b

b b

b b

b b

δ δ α δ

δ δ

δ δ

δ δ

δ / 1

²

1 1

2

= − = −

−

=

=

= ℜ

) 1

(

_o

i i

i

i L G

e Ci

T G P

e P

P T Ci

T G P T e

ω ωτ δ α

δ δ α ω

δ

^{ϕ ϕ ϕ}

+

= + +

+

=

⇒ +

+

=

^{− − −}

) 1

( 1 )

1 ( 1

1

o i

b o

i b

b b

b

L i

Le V

i L G

e T

P V P

T T

P

V ωτ ωτ

α δ

δ

α

^{ϕ ϕ}

+

− + + =

− +

=

−

=

ℜ

^{− −}

+ 1

= L τ

o

τ

ωτ

ϕ

i L

L e V

b i

+

− +

=

ℜ 1

1 )

1 ( 1

For large ETF (L>>1):

• the time constant is

reduced wrt the standard one by L+1

• For slow signals (ω<<1/τ) and large ETF the

responsivity is simply

-1/V

_b

Such a high value for α (which is >0 for TES) induces a large change in the bias power when

radiation hits the detector (electrothermal feedback) This results in a large

reduction of the time constant and in

stabilization of the

responsivity.

• Are the future of this field. See recent reviews from Paul Richards, Adrian Lee, Jamie Bock, Harvey Moseley … et al.

• In Proc. of the Far-IR, sub-mm and mm detector technology workshop, Monterey 2002.

TES arrays

Cryo:

0.3K Space qual.

receiver (1pixel of 1000)

antenna

stripline filter

membrane island

load

TES

Si substrate with Si3N4 film

SQUID Readout

MUX

TES for mm waves

(Cardiff, Phil Mauskopf)

… and many others …

150 μm

PROTOTYPE SINGLE PIXEL - 150 GHz (Mauskopf) Schematic:

Waveguide

Radial probe

Nb Microstrip

Silicon nitride Absorber/

termination

TES Thermal links

Similar to JPL design, Hunt, et al., 2002 but with

waveguide coupled antenna

PROTOTYPE SINGLE PIXEL - 150 GHz (Mauskopf) Details:

Radial probe

Absorber - Ti/Au: 0.5 Ω/square - t = 20 nm Need total R = 5-10 Ω

w = 5 μm → d = 50 μm

Microstrip line: h = 0.3 μm, ε = 4.5 → Z ~ 5 Ω

TES Thermal links

TES Readout

A very low impedance, extremely low current noise amplifier 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 for 10³-10⁴ pixels. The rest of the readout is warm.

‹ Very flexible: different materials and geometries can be chosen to tune 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:

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

KIDs working principle:

In a superconductor below T_c, electrons can bind to form CPs with binding energy E=2Δ =3.5*k_bT_c .

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

CPs QPs

The values of s_s and s_n depend on the densities of QPs and CPs. By measuring them, we can get

information on n_qp .

J

_s

J

_n

-is₂n_CP -is₂n_QP

s₁n_QP

A better estimate of s_s and s_n is obtained using the Mattis Bardeen integrals:

A better theory...

Note that:

• R_s decreases exponentially

• X_s becomes constant

• X_s/ R_s grows exponentially

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

How can we measure the small variations of L

_k?

The superconductor can be inserted in a resonating circuit with extremely high Q, since:

s

s

R

X Q ∝

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

C_c

R_QP L_kin

L_mag

C_l

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

Cⁿ_c

Rⁿ_QP Lⁿ_kin

Lⁿ_mag

Cⁿ_l C¹_c

R¹_QP L¹_kin

L¹_mag

C¹_l

C²_c

R²_QP L²_kin

L²_mag

C²_l RF carrier (f¹+ f²+ ... + fⁿ )

Pixel 1, f¹ Pixel 2, f² Pixel n, fⁿ

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?

n′_CP< n_CP QPs

CPs

• Suppose a photon hits the detector

• If its energy is high enough (hν > 2ΔE) it can break CPs

• The density of CPs therefore changes

• This leads to a variation of L_kin

The same effect can be accomplished by increasing the temperature of the superconductor

The readout is accomplished by

monitoring the phase of the transmitted signal

L

f

₀

∝ 1

dq

Readout technique

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

This kind of plots can give all the information regarding resonator parameters

It is also the basis for actual measurements of radiation Remember that:

δ T δϑ

n

QP

δ

) δϑ

T

(

n

_QP

Cryogenic system overview

SCN-CN coax

VNA / IQ mixers

2xDC block 2xDC block

2x10dB atten

1xDC block 1xDC block

1x10dB 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

f_synt f_synt

PC DAC

ADC

f_synt

f_synt± f₀ , f_synt± f₁...

f₀, f₁...

f₀, f₁...

f_synt± f₀ , f_synt± f₁...

Single pixel readout system Multipixel readout system

Both systems share the core components!

0.3K 2K

Re(S₂₁)

Im(S₂₁)

PC DAC

ADC

f_synt

f_synt f₀, f₁...

f₀, f₁...

f_synt± f₀ , f_synt± f₁...

10 mm

8 mm

capacitive coupling

2.5 mm

0.7 mm

Al CPW (200nm) SiO₂ (1μm)

Silicon Substrate (0.5mm)

KID chip description

Material: Aluminium

6 resonators of varying length Substrate:

The dielectric constant is not exactly determined!

Base temperature characterization

All 6 resonances observed!

Typical resonance amplitude curve

10

5

1

≈ . 1 ⋅

Q

Effect of temperature variation - 1

Higher T Higher n_qp Higher losses

Higher T Lower n_qp Lower f₀

Amplitude

Phase

Effect of temperature variation - 2

) T ( f L

TOT

1

0 ∝

TOT kin

L

= L

α L ( )

) T ( L )

( f

) T ( f

kin kin

0 2

1

0 0

0 = − αδ

δ

Quality factor increase

Estimate of kinetic inductance fraction

α

^{≈ 0.018}

2D data analysis

A fitting procedure has been developed to estimate the parameters of 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/n_qp equivalent to 10^-9-10^-8 deg/N_qp

Optical measurements

System modified by adding a filter chain

Polyethilenewindow Fluorogold(400GHz lowpass) Fluorogold+145GHz bandpassfilter

Horn KID Gunn diode

Quasiparticle lifetime

To estimate the absorbed power that induces the signal we still need one piece of information:

QP qp abs

P n

ητ

= Δ

τ

_QP ≈ 30μs

When T decreases, the quasiparticle lifetime increases (n_QP smaller!)

A possible solution: LEKID

‹ Distributed element KIDs

‹ Lumped element KIDs

Needs some sort of antenna

Response depends on where the photon hits the sensor

C L

It is possible to tune the meanders to 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

(

^{2 2}

)

2

2 1 _QP

QP QP

S N

NEP +ω τ

⎥⎥

⎦

⎤

⎢⎢

⎣

⎡

δ δϑ Δ

= ητ

− ϑ

Theoretical limit given by GR noise is as low as 10^-20W/Hz^0.5 at 100mK

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

Conclusions

‹ The 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!