Development of FD-SOI cryogenic amplifier for

application to STJ readout in COBAND experiment

Sep. 10-13, 2018 / Sorent, Italy

Yuji Takeuchi (TCHoU, Univ. of Tsukuba)

S.H.Kim, T.Iida, K.Takemasa, C.Asano, R.Wakasa, A. Kasajima, K. Takahashi, Y. Tsuji,

Y. Terada (U of Tsukuba), H.Ikeda, T.Wada, K.Nagase (ISAS/JAXA), S.Matsuura (Kwansei

gakuin U), Y.Arai, I.Kurachi, M.Hazumi (KEK), T.Yoshida, T.Nakamura, M.Sakai,

W.Nishimura (U of Fukui), S.Mima, K.Kiuchi (RIKEN), H.Ishino, A.Kibayashi (Okayama U),

Y.Kato (Kindai U), G.Fujii, S.Shiki, M.Ukibe, M.Ohkubo (AIST), S.Kawahito (Shizuoka U),

E.Ramberg, P.Rubinov, D.Sergatskov (FNAL), S.B.Kim (Seoul National U)

COBAND collaboration1

COBAND(Cosmic Background Neutrino Decay)

Search for Neutrino decay in Cosmic background neutrino

➔To be observed as FIR photons around ~50m in the

overwhelming zodiacal emission

𝜈2

𝜈3𝜈1

𝜈3

𝜈1

𝜈3

𝜈2𝜈3

𝜸

𝜈1

𝜸𝜈3

𝜈2

2

Neutrino Decay signal and backgrounds

3

CMB

ISD

SLDGL CB decay

wavelength [m]

Su

rfac

e b

rig

htn

ess

I [M

Jy/s

r]

Zodiacal Emission

Zodiacal Light

Integrated flux from galaxy counts

CIB summary from Matsuura et al.(2011)

10001001010.0001

0.001

0.01

0.1

1

10

100Zodiacal

Emission

Expected Neutrino

Decay signal

We can identify decay signal by highly precise measurement

of photon energy spectrum around ~50m

𝜏 = 1 × 1014yr𝑠

𝒎𝟑 = 𝟓𝟎𝐦𝐞𝐕

Current experimental

lower limit is 1012 yrs

➔NEP = 𝜖𝛾 2𝑓𝛾 ~ 110−19 Τ𝑊 𝐻𝑧

Requirements for the photo-detector for sensitivity to neutrino

lifetime of 𝜏=1014 yr𝑠 in COBAND experiment

4

• Sensitive area of 100m100m for each pixel

• High detection efficiency for a far-infrared single-photon in

=40m～80m

• Dark count rate less than 300Hz (expected real photon rate)

We are trying to achieve NEP ~10−19 Τ𝑊 𝐻𝑧 by using

• Superconducting Tunnel Junction (STJ) sensor

• Cryogenic amplifier readout

Superconducting Tunnel Junction (STJ) Sensor

A constant bias voltage (|V|<2Δ) is applied across the junction.

A photon absorbed in the superconductor breaks Cooper pairs and creates

tunneling current of quasi-particles proportional to the deposited photon energy.

• Superconductor / Insulator /Superconductor

Josephson junction device

Δ: Superconducting gap energy

2

E

Ns(E)

Superconductor Superconductor

Insulator

Insulator

Superconductor300nm

• Much lower gap energy (Δ) than FIR photon ➔ Can detect FIR photon

• Faster response (~s) ➔ Suitable for single-photon counting 5

Temperature(K)0.3 0.4 0.5 0.6 0.7

Le

akage

100pA

1nA

10nA

100nA

Nb/Al-STJ development at CRAVITY

Ileak~200pA for 50m sq. STJ, and achieved 50pA for 20m sq.

➔ This satisfies our requirement!

0.1nA

50m sq. Nb/Al-STJ fabricated at CRAVITY

6

50

0p

A/D

IV

0.2mV/DIV

I

V

T~300mK

w/ B fieldLeakage

Far-infrared single photon detection is feasible with this

Nb/Al-STJ sensor and a cryogenic amplifier which can

be deployed in close proximity to the STJ.

FD-SOI-MOSFET at cryogenic temperatureFD-SOI : Fully Depleted – Silicon On Insulator

7

Vgs (V)

Ids

0 0.5 1 1.5 21pA

1nA

1A

1mA

ROOM

3K

n-MOSp-MOS

ROOM

3K

-Ids

1nA

1A

1mA

Vgs (V)

0-0.5-1-1.5-2

Id-Vg curve of W/L=10m/0.4m at |Vds|=1.8V

Both p-MOS and n-MOS show excellent performance at 3K and below.

~50nm

Channel Length : L

Channel

Width : W

Very thin channel layer in MOSFET on SiO2

No floating body effect caused by charge

accumulation in the body

FD-SOI-MOSFET is reported to work at 4KJAXA/ISIS AIPC 1185,286-289(2009)

J Low Temp Phys 167, 602 (2012)

Increasement

of drain

current

Shift in VTH

8m

V1

50

mV

10m

V

100s

SOI prototype amplifier for demonstration test

Ｖ

Amplifier stage

Buffer stage

T=350mK

Test pulse input through C=1nF at

T=3K and 350mK

• Power consumption: ~100μW

• Output load: 1M and ~0.5nF at

this test

INPUT

OUTPUT1nF

Test pulse

8

SOI prototype amplifier for demonstration test:Common source stage

9

● Room Temp.

(V3=1.10V)

● 3K

(V3=1.50V)

Ga

in0

10

20

30

40

50

60

0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

Current source bias (V2) [V]

V3

V2

In

Out

We can compensate the effect of shifts in the thresholds by

adjusting bias voltages to archive similar gain!

Demonstration Test for Amplification of STJ response to laser

pulse on cold stage

10

We connect 20m sq. Nb/Al-STJ and SOI amplifier on the

cold stage through a capacitance and illuminate visible

laser pulses onto the STJ.

STJ

10M 3He sorption

cold stage

T~350mK

465nm laser

pulse through

optical fiber

GND

Vss

Vd

d4.7F

Cold amp.

input monitor

Cold amp.

output

STJ SOI

4.7F

Demonstration Test for Amplification of STJ response to laser

pulse on cold stage

11

Demonstrated to show amplification of Nb/Al-STJ response to laser pulse

by SOI amplifier situated close to STJ at T=350mK

Voltage [μ

V]

70

60

50

40

18μV

Voltage [

μV

]

time [μs]

1600

1200

800

400

0-100 -80 0-40 -40 -20 20 40 60 80 100

1.2mV

Input to SOI amp. from STJ

Output from SOI amp.

Laser pulse(=465nm)

Development of SOI cryogenic amplifier for STJ signal readout

is now moving to the stage of design for practical use !

T=350mK

Gain:

~70

CC operationBiasing STJ with large R keeps constant current on STJ

CV operationSTJ connection to virtual short through C keeps constant voltage on STJ

Read voltage by high

input impedance amp.

STJ readout byConstant Current and Constant Voltage operation

12

~𝟏𝐦𝐕 →

𝑽𝟎

𝑪

𝑪𝐒𝐓𝐉

𝑹𝐢𝐧

𝑹

STJ

I

V

STJ I-V curve w/light w/o light

V0=V+RI

~1mV

Read charge by

low input

impedance

amp.

Large fraction of signal is lost through CSTJ

I

V1mV

𝑹𝐅𝐁

𝑽𝟎

𝑪

𝑪𝐒𝐓𝐉

𝑹

+-𝑪𝐅𝐁

STJ

~𝟏𝐦𝐕 →

Better readout for STJ

Charge sensitive amplifier

20μA

10μA 2μA

20μA

Buffer

stage

40μA

Telescopic

cascode current

mirror

Feed

back

13

Differential amp.• Negative feedback

with C=3pF and

R=5M

• GBWP~2MHz

Differential amplifier with negative feedback of CR

Power consumption𝑉𝑑𝑑 − 𝑉𝑠𝑠 = 3𝑉𝐼𝑡𝑜𝑡𝑎𝑙 = 40𝜇𝐴

➔~120W

Response to charge injection

Input charge ~400fC

Observed gain: ~45mV/400fC~1/9pF

• Due to stray capacitance in wire

bonding used for feedback CR?

Show almost same performance in 3K

to one in room temp.

Now we are working on STJ signal

readout with this amp.

We are developing amp. with smaller C

• Target: C=200e/0.5mV ~ 60fF

200ns

0.4

A

400fC

Input

output Room Temp.

50s10m

V

50s10

mV

3Koutput

CH2

CH1

𝑉ss

×10

101pF 1k

𝑉dd𝐼ref

DC,1MΩ

Refrig.

F.G.3pF

5MΩ

-+

14

FET Threshold voltage control in Double-SOI at cryogenic temp.

We found Vth control is also possible at cryogenic temperature

• Can compensate shift of Vth in cryogenic temp.

• Can realize a FET with steep rising of Id around Vgs=0 ➔ Potential to ultra-low power consumption device

VDSOI

Middle silicon layer

BOX

pMOS-FET Id-Vgs curve

-1.5 -1 -0.5 0 0.5 Vgs(V)

1A

1nA

1pAI d

Vth control by VDSOI at room

temperature is well established. 3K

15

16

Potential to large pixel STJ with readout

viaSTJ

capacitor

FET

70

0 u

m

640 um

Development of Direct Intergradation of STJ on SOI (SOI-STJ)

• STJ layers are fabricated directly on a

SOI pre-amplifier board and cooled down

together with the STJ

FD-SOI on which STJ is fabricated

17

• Both nMOS and pMOS-FET in FD-SOI wafer on which a

STJ is fabricated work fine at temperature down to ~100mK

• We are also developing SOI-STJ fabrication at CRAVITY

B~150Gauss

2mV/DIV1m

A/D

IV

I-V curve of a STJ fabricated

at KEK on a FD-SOI wafernMOS-FET in FD-SOI wafer on which a STJ is fabricated at KEK

gate-source voltage (V)

dra

in-s

ou

rce

cu

rre

nt

0 0.2 0.4 0.6 0.8-0.2

1pA

1nA

1A

1mA

Summary• In COBAND project, we require a photo-detector with NEP~10-19

W/Hz to detect single FIR photon (~50m).

• We are developing Nb/Al-STJ with cryogenic amplifier readout.

– Nb/Al-STJs fabricated at CRAVITY satisfy our requirements.

– Cryogenic FD-SOI amplifiers are under development and we

demonstrated STJ signal amplification by a prototype SOI

amplifier at T~350mK.

• We are designing and testing more practical readout for STJ

signal with low input impedance charge integration amplifier.

– Using SOI technology, a wealth of knowledge in circuit at room temperature is

available even at cryogenic temperature almost as it is.

18

* SOI design in this work is supported by VDEC, the U. Tokyo in collaboration with Synopsys, Inc., Cadence Design

Systems, Inc., and Mentor Graphics, Inc.

Backup

19

COBAND (COsmic BAckground

Neutrino Decay)

Heavier neutrinos in mass-eigenstate (2, 3) are not stable

─ Neutrino can decay through the loop diagrams

✓ However, the lifetime is expected to be much longer than

the age of the universe

➔ We search for neutrino decay using Cosmic Background

Neutrino (CB) as the neutrino source

𝑊

𝜈3

γ

e, 𝜇, 𝜏 𝜈1,2

20

𝝆(𝝂𝟑 + ത𝝂𝟑)~ Τ𝟏𝟏𝟎 𝐜𝐦𝟑

Cosmic background neutrino (C𝜈B)

CMB(=Photon decoupling)

𝒏𝜸 = 𝟒𝟏𝟏/𝐜𝐦𝟑

𝑻𝜸 = 𝟐. 𝟕𝟑 𝑲

~380,000yrs after the

Big Bang

CB (=neutrino decoupling)

~1sec after the big bang21

Expected photon wavelength spectrum from CB decays

50𝜇𝑚(25meV)

dN/

d(a

.u.)

[m]100 50010

Red Shift effect

Sharp Edge with

1.9K smearing

𝒎𝟑 = 𝟓𝟎𝐦𝐞𝐕

E [meV]100 50 20 10 5

distribution in ν3 → 𝜈2 + 𝛾

No other source has such a sharp edge structure!!

If assume

𝑚1 ≪ 𝑚2 < 𝑚3,

𝑚3~50𝑚𝑒𝑉 from

neutrino oscillation

measurements

22

Motivation of -decay search in CB

If we observed the neutrino radiative decay at the lifetime

much shorter than the SM expectation, it would be

• Physics beyond the Standard Model

• Direct detection of C𝜈B

• Determination of the neutrino mass

– 𝑚3 = ൗ𝑚32 −𝑚1,2

2 2𝐸𝛾

➔ Aiming at a sensitivity to 3 lifetime in Ο 1013 − 1017 yr𝑠

• Standard Model expectation: 𝜏 = Ο(1043) yr𝑠

• Experimental lower limit: 𝜏 = Ο(1012) yr𝑠

• Left-Right symmetric model predicts 𝜏 = Ο(1017) yr𝑠 for

𝑊𝐿-𝑊𝑅 mixing angle 𝜁 ~0.02

3 Lifetime

23

Existing FIR photo-detectors

Detectors (μm)Operation

Temp.

NEP

(W/Hz1/2)

Monolithic

Ge:Ga50-110 2.2K ～10-17 Akari-FIS

Stressed

Ge:Ga60-210 0.3K ~0.9×10-17 Herschel-

PACS

Need more than 2 orders improvement from

existing photoconductor-based detectors

24

STJ energy resolution

Signal = Number of quasi-particles

𝑁𝑞.𝑝. = 𝐺𝐸𝛾1.7Δ

Resolution = Statistical fluctuation in number of quasi-particles

Τ𝜎𝐸 𝐸 = Τ1.7Δ 𝐹 𝐸

➔Smaller superconducting gap energy Δ yields better energy

resolution Δ: Superconducting gap energy

F: fano factor (~0.2 for Nb)

E: Photon energy

G: Back-tunneling gain

Tc :SC critical temperature

Need ~1/10Tc for practical

operation

25

Si Nb Al Hf

Tc[K] 9.23 1.20 0.165

Δ[meV] 1100 1.550 0.172 0.020

STJ current-voltage curve

Optical signal readout➔ Apply a constant bias voltage (|V|<2Δ)

across the junction and collect tunneling current of quasi particles created by photons

✓ Leak current causes background noise

Tunnel current of Cooper pairs (Josephson current) is suppressed by applying magnetic field

2Δ

Leak current

B field

I-V curve with light illumination

Voltage

Current

Signal current

26

STJ back-tunneling effect

Photon

• Bi-layer fabricated with superconductors of different gaps

Nb>Al to enhance quasi-particle density near the barrier

– Quasi-particle near the barrier can mediate multiple Cooper pairs

• Nb/Al-STJ Nb(200nm)/Al(70nm)/AlOx/Al(70nm)/Nb(200nm)

• Gain: ～10

Nb Al NbAl

Si waferNb

NbAl/AlOx/Al

27

Proposal for COBAND Rocket Experiment

JAXA sounding rocket S-520

Telescope with 15cm diameter and 1m focal length

At the focal point, a diffraction grating covering =40-80m and an array of photo-detector pixels of 50() x 8() are placed.

Each pixel has 100mx100m sensitive area.

Aiming at a sensitivity to 𝜈 lifetime for 𝜏 𝜈3 = Ο 1014 yr𝑠

28

𝜃

Drain Ａｖａｒａｎｃｈｅ

In the region of 𝑉gs < 𝑉ds , drain current causes drastic increasement. We need to be care not to use this region in design.

0 0.5 1.0 1.5 2.0 Vds[V]

0.6

0.5

0.4

0.3

0.2

0.1

0I ds[m

A]

-2.0 -1.5 -1.0 -0.5 0

0.3

0.2

0.1

0

−I ds[m

A]

Vds[V]

𝐼ds − 𝑉ds curve (pmos)＠3K

W/L=10μm/5μm

𝐼ds − 𝑉ds curve (nmos)＠3K

29