Characterization of HV-CMOS pixel sensor prototypes · Ettore Zaffaroni STREAM Final Conference, 17/09/2019 Smart Sensor Technologies and Training for Radiation Enhanced Applications

Characterization of HV-CMOSpixel sensor prototypes

Ettore Zaffaroni

STREAM Final Conference, 17/09/2019

Smart Sensor Technologies and Training for Radiation Enhanced Applications and Measurements (STREAM) is a project funded by the European Commission under the Horizon2020 Framework Program under the Grant Agreement no 675587.

STREAM began in January 2016 and will run for 4 years.

Ettore Zaffaroni 2

Outline

● Introduction● The characterized sensors● TCT (Transient Current Technique)

measurements and results● Testbeam measurements and results

Ettore Zaffaroni 3

Introduction

● ATLAS will upgrade its inner tracker for HL-LHC– ITk, ~190 m2 of silicon,

~15 m2 of pixel detectors– High occupancy and radiation

damage● HV-CMOS developed as a

possible replacement for the outermost pixel layer

Technical Design Report for the ATLAS Inner Tracker Pixel Detector”,CERN-LHCC-2017-021 (2018). https://cds.cern.ch/record/2285585

Ettore Zaffaroni 4

ams development timeline

● CCPDv3, CCPDv4 (ams 180 nm) small prototypes, capacitively coupled to front-end ASICs

● H35DEMO (ams 350 nm), first full scale prototype, with monolithic parts● ATLASPix1 (ams 180 nm), fully monolithic, 3 pixel matrices● ATLASPix2 (ams/TSI 180 nm), small scale prototype, focus on periphery

and SEU tolerant memory ● ATLASPix3 (TSI 180 nm), full scale prototype, single pixel matrix

2013 2014 2015 2016 2017 2018 2019

CCPDv3~2x3 mm2

CCPDv4~2x3 mm2

H35DEMO~18x24 mm2

ATLASPix1~10x25 mm2

ATLASPix2~3x4 mm2

(ams/TSI)ATLASPix3~20x21 mm2

(TSI)

Ettore Zaffaroni 5

H35DEMO chip● H35 technology by ams

– 350 nm HV-CMOS● 4 pixel matrices and

test structures● 250x50 mμm 2 pixels● 4 different resistivities

– 20, 80, 200, 1000 Ω·cmcm

Imag

e fro

m:

http

://d

x.do

i.org

/10.

1088

/174

8-02

21/1

1/01

/C01

012

18 mm

24 mm

Ettore Zaffaroni 6

ATLASPIX1 chip

●aH18 technology by ams–180 nm HV-CMOS

●Fully monolithic●Same resistivities as H35DEMO●ATLASPIX_Simple matrix tested

–130x40 mμm 2 pixels–25x400 matrix–Column-drain readout,triggerless

ATL ASP IX_ IsoS impl e

ATL ASP IX_ Simp le

ATL ASP IX_ M2

10 mm25 m

m

Ettore Zaffaroni 7

Transient Current Technique

Ettore Zaffaroni 8

TCT (Transient Current Technique)

● Generation of carriers using a laser– Precise location

● Carriers move under electric field, generating a current

● Current signal amplified with a RF amplifier

● TCT allows to study the space charge region– Depletion depth, Neff, etc.

LASER + ++ + + + + + +

++

+

++

+

-- - - - - - - - - - - - -

- +

-

Detector bulk

Test structure

z position

Ettore Zaffaroni 9

TCT setup● Pulsed IR laser (1064 nm) with FWHM of 12 μmm● Detector at –27 °C using Peltier● 1 μmm step size in all axes

Ettore Zaffaroni 10

TCT setup● DUT mounted vertically to reduce effects of

swinging stages● PCB with controlled impedance traces and correct

termination to remove signal reflections

Ettore Zaffaroni 11

H35DEMO test structures● In the periphery of the chip● 3x3 pixel matrix (just the

sensor diodes)● Outermost pixel cathodes are

connected together– 2 channels (central and

external)● Top bias

750 um750 um

150 um150 um

LASE

RLA

SER

n+

SNTUB

DNTUB

SN

n+

SNTUB

n+

DNTUB

p+

SP

DP

HV

p+

SP

DP

HV

p+

SP

DP

HV

p+

SP

DP

HV

n+

SNTUB

DNTUB

p substrate

to amplifier

p+

SPTUB

DPTUB

GND

NOT TO

SCALE250 um250 um

zx

yx

yx

y

x

y

z

Ettore Zaffaroni 12

Irradiation campaigns

● Neutrons– TRIGA reactor in Ljubljana

● One irradiation step per sample

– Annealing● Protons (measurements at Uni. Bern and Uni. Geneva)

– BERN Inselspital cyclotron (16.7 MeV)● Multiple irradiation steps per sample

– PS IRRAD (24 GeV)● One irradiation step per sample

Ettore Zaffaroni 13

eTCT scans

● Edge TCT scans perfomed● ~150-300x 1 m steps in y, ~70x 5 μm m steps in xμm

● Several voltage steps (10 to 12 normally)– From 0 V to -100 or -165 V

● At each step the signal is averaged 40 times– To reduce noise

● All results shown for central pixel

n+

SNTUB

DNTUB

SN

n+

SNTUB

n+

DNTUB

p+

SP

DP

HV

p+

SP

DP

HV

p+

SP

DP

HV

p+

SP

DP

HV

n+

SNTUB

DNTUB

p substrate

to amplifier

p+

SPTUB

DPTUB

GND

x

y

z

DOI: 10.1088/1748-0221/13/10/p10004

Ettore Zaffaroni 14

Data analysis - depletion● Integration of current

signal to get the charge

● Selection of the region of interest

1000 Ω·cmcm, -100 V200 Ω·cmcm, -100 V

250

200

150

100

50

0

300

Ettore Zaffaroni 15

Data analysis - depletion

● Fit of the charge profiles– One fit per profile in the ROI

● Two contributions:– Smeared box function– Gaussian, to model the charge

sharing● Calculation of the FWHM

– Max of the box function considered

200 Ω·cmcm

1000 Ω·cmcm

Ettore Zaffaroni 16

Data analysis - Neff

● Neff (effective doping concentration) is calculated by fitting the depletion vs voltage data with:

● d0 and Neff free parameters– e electron charge– ε silicon dielectric constant

● d0: sensitive region depth at 0 V bias (due to built-in voltage and n-well finite depth)

d=d0+√2ϵ

e N eff

V

200 Ω·cmcmneutrons 1014 neq/cm2

Ettore Zaffaroni 17

Results - Neff

20 Ω·cmcm

80 Ω·cmcm

200 Ω·cmcm

1000 Ω·cmcm

Ettore Zaffaroni 18

Results - Neff

● Significant differences between protons and neutrons and between resistivities

● Initial increase of Neff at very low fluences (<1e14 neq/cm2, protons) for the 200 Ω·cmcm sample– Effect competing with initial acceptor removal?– Not observed in 1000 Ω·cmcm, data not available for 20 and

80 Ω·cmcm

Ettore Zaffaroni 19

Results – Neff neutrons

p 16.7 MeV p 24 GeV

● Plots combined by particle type, for different initial resisitivities

● Tend to the same Neff value

Ettore Zaffaroni 20

Results - annealing

● Measured on neutron irradiated samples

● Initial beneficial annealing, then reverse annealing

● Measurement will be performed on proton irradiated samples for comparison

Ettore Zaffaroni 21

Testbeams

Ettore Zaffaroni 22

Testbeams

● Telescope: pixel sensors used to measure tracks and generate trigger

● DUT read out at the same time

● Reconstruction and analysis

Ettore Zaffaroni 23

Reconstruction and analysis

RAW DATA

NOISE-SCANRemove noisy pixels

ALIGNMENTAlign telescope planes

and DUT

TRACK RECONSTRUCTIONFind tracks, match

with DUT hits

DATA ANALYSISEfficiency, timing, etc.

Proteus software

Ettore Zaffaroni 24

Telescope modules

● Planar modules from IBL– Innermost pixel layer in

ATLAS– Planar silicon pixel sensor

● Read out by the FE-I4 ASIC– 50x250 μmm2 pixels– 80 columns, 336 rows

FE-I4

Ettore Zaffaroni 25

DAQ: CaRIBOu system

● Read-out system for ITk and CLIC sensor prototypes

● Provides power, HV and data links

● Based on a PC, a Zynq-7000 FPGA and a custom DUT board

FPGA board(ZC706)

DUT

PC

CaR

boar

d

DOI: 10.1088/1748-0221/12/01/P01008

Ettore Zaffaroni 26

Telescope● 6 FE-I4 modules

– Spatial resolution 8x12 um

● Trigger rate up to 4 kHz● Cold DUT box (down to

-20 °C)● Successfully used in

various campaigns (SPS, FNAL) since 2014

2013 2014 2015 2016 2017 2018 2019

SPS FNAL

DOI: 10.1088/1748-0221/13/02/p02011, 10.1088/1748-0221/13/12/p12009

Ettore Zaffaroni 27

Measurement campaign

● Data acquisition at CERN SPS (06-10/2018)● Simple matrix of ATLASPIX1● Bias voltage and threshold scans● Different irradiations

– Protons (while operating the sensor)– Neutrons

Ettore Zaffaroni 28

Testbeam results

● High efficiency after irradiation● Noisy pixels and neighbors

masked● Lines masked due to issues in

row circuitry– Identified and solved in

subsequent prototype submission

Ettore Zaffaroni 29

Testbeam results

● Efficiency vs bias voltage (left) and threshold (right)

● Above 98% for a wide range of parameters

Ettore Zaffaroni 30

Testbeam results

● Cluster time: difference between trigger time (telescope) and cluster time (DUT)– 1 bin spread given by the

telescope● Cluster value: time over

threshold

Ettore Zaffaroni 31

Comments

● Testbeams show excellent performance of ATLASPix1 before and after irradiation– Efficiency up to ~99% after 1015 neq/cm2

– Excellent timing performances● Identified issues in the circuitry, feedback

provided to the designers

Ettore Zaffaroni 32

Conclusions● TCT measurements

– Initial acceptor removal observed after proton and neutron irradiation– Beneficial and reverse annealing observed for neutrons– Significant differences between neutron and protons effects

● Testbeams – High, uniform efficiency over a large pixel matrix after irradiation– Very good timing performances

● Excellent results for fully monolithic pixel sensors!

Thank you!