The ATLAS Upgrade Introduction Planar Pixel Sensor R&D Project - - PowerPoint PPT Presentation

The ATLAS Upgrade Planar Pixel Sensor R&D Project

• Recent Progress -

Reiner Klingenberg TU Dortmund University for the PPS R&D Collaboration 8th International Hiroshima Symposium of the Development and Application of Semiconductor Tracking Detectors HSTD-8, Taipei, Taiwan, 5-8 December 2011

Introduction LHC-ATLAS-Pixel and its upgrades The ongoing Research and Development Examples of some recent R&D aspects

1

Large Hadron Collider and the ATLAS Experiment

so far: LHC collides p+p at 3.5 + 3.5 TeV some 1300 bunches with 2·1014 protons instantaneous luminosity ~ 3·1033 cm-2s-1 or ~ 3 kHz/µb

2

ATLAS - Pixel Detector

innermost detector component 3 barrel layers + 2·3 endcap disks covers pseudorapidity range |η| < 2.5 Performance requirements transverse impact resolution < 15 µm minimal material to reduce multiple scattering and conversions high efficiency radiation hard: 500 kGy, 1015 neqcm-2 Module concept detector consists of 1744 modules of planar silicon n+-in-n sensors + frontend electronics FE-I3+ flex hybrid ~80 millions read-out channels

3

ATLAS - Pixel Detector

Pixel Sensor Design DOFZ Si n-substrate, 250 µm thick planar n+-in-n pixels, 400x50 µm2 16 guard rings on p side to reduce HV in steps 1.1 mm inactive edge incl. safety margin Read-out electronics FE-I3 chip DC coupled and bump bonded shaper + amplifier + discriminator ⇒ ToT ⇒ charge Q

4

LHC: goal to extend physics reach with increased peak luminosity to increase sensitivity for physics beyond the standard model like Higgs, SUSY, extra dimensions, heavy bosons Detectors increased occupancy and high event rate O(100) events / bunch crossing event pile-up ⇒ challenging reconstruction higher pixel granularity

Challenges at High Luminosity

Radiation damage ⇒ charge collection, depletion fluences of 1016 neqcm-2 in inner layer radiation hard material, design of sensor layout, thinness of bulk, higher bias voltage, lower threshold in read-out electronics

5

Envisaged LHC and ATLAS Upgrade Phases

now

first step: IBL, the Insertable B-Layer phase 0: IBL Insertable b-Layer phase I: NewPix under evaluation phase II: New Inner Detector increasing luminosity ➥ radiation hardness increasing occupancy ➥ better granularity required

6

Goals

evaluate & improve efficiencies & charge performances @ high & highest fluences ⇒ radiation hardness? operation conditions geometry optimization ⇒ pixel size; pixel implant; bias grid slim/active edges; dicing cost reduction at large areas ⇒ bulk options & bump bonding

ATLAS Planar Pixel Sensor Project

Tools

different sensors: CiS / HLL-MPI / MICRON / HPK and read-out FEI3/FEI4 n&p irradiations (Ljubljana reactor n, CERN 24 GeV p, Karlsruhe 26 MeV p, LANL 800 MeV p) up to 2·1016 neqcm-2 characterisation in lab and test beams CERN 120 GeV π DESY 4 GeV e TCAD simulations and dopant profile measurements for optimization

n+-pixels n-substrat, DOFz, 3-16 Vbias> 50x250 75-250 n+-pixels p-substrat, Fz, 3-21 Vbias> 50x250 Vbia 75-250

FE Sens 27 µm Cu3Sn Cu6Sn

7

Sensors measured in lab and test beam

sensors bonded to FE-I3 or FE-I4 read-out chips are tested in the laboratory with the help of radioactive sources and in a test beam set-up equipped with a beam telescope which allows tracking

source scans: hit maps & energy deposition test beam: hit maps & energy deposition & efficiencies & spatial resolution

EUDET Reports 2010-01 & 2008-04

single chip adapter card w/ sensor+FE

• ne or

several DUTs cooling

8

n+-in-n sensors: collected charge after neutron irradiations

are currently used in the ATLAS pixel detector for a fluence range 1015 neqcm-2 tests of the radiation hardness of up to 2⋅1016 neqcm-2 are ongoing neutron irradiation (Ljubljana reactor) up to 2⋅1016 neqcm-2, FE-I3 read-out measurements with e- from 90Sr source and π&e test beam charges seen are higher than model predictions which include trapping

5·1015 neqcm-2 (n) 2·1016 neqcm-2 (p)

9

n+-in-n sensors: collected charge after proton irradiations

proton irradiation (KIT, 26MeV) up to 1.4⋅1016 neqcm-2, FE-I3 read-out unirradiated FE with indium bumps flip-chipped to irradiated sensor measurements with e- from 90Sr source and π test beam

threshold FE-I3 3200e- cluster charge

1·1016 neqcm-2 (p) 1.4·1016 neqcm-2 (p)

10

n+-in-n sensors: hit efficiencies after neutron & proton irradiations

hit efficiency can be fully recovered by increasing bias voltage charge distributions @ 1.4·1016 neqcm-2 (p)

CiS 250 µm + 285 µm (2-14)·1015 neqcm-2 (p) (5-20)·1015 neqcm-2 (n)

%

O(1015) neqcm-2 O(1016) neqcm-2

11

n+-in-n sensors: hit efficiencies after neutron & proton irradiations

sub-pixel resolved hit efficiency for a 250 µm sensor @ 4⋅1015 neqcm-2 (n) mean efficiency > 97% @ 600 V

LUB2

sub-pixel resolved hit efficiency for a 250 µm sensor @ 1.4⋅1016neqcm-2 (p) mean efficiency > 97% @ 1800 V

1300V 1800V

12

n-in-p sensors

investigation of charge multiplication

parallel ¡ column parallel ¡ row

wire ¡bondable ¡ pixels

6’’ ¡300µm ¡wafers, ¡double ¡metall, ¡n-­‑in-­‑p ¡FZ

NIM ¡A ¡636 ¡(2011) ¡56

300 µm - 140 µm has been investigated effect more significant in thin sensors

strips pixels

n+-pixels p-substrat, Fz, MCz p+-implant 3-21 GR 0 V Vbias> 500V 50x250µ Vbias 75-250 µm

13

n-in-p sensors

assemblies were irradiated up to 5⋅1015 neqcm-2, neutrons and mixed significant charge is collected above threshold, Vbias > 600V

5·1015 neqcm-2 Vbias=1kV FE-I3 threshold lab data

14

n-in-p sensors

300 µm, BCB insulation on pixel side, HV stability proven n-irradiation @ 5⋅1015neqcm-2: significant charge is collected above threshold

99.8% 98.6%

lab and test beam measurements deliver comparable results Vbias = 600V MPV(charge)=6.4ke

• verall efficiency 98.6%

losses around bias dot higher eff in the center of the pixels

lab + test beam

15

Thin bulk sensors

first HPK n-in-p sensors, 150 µm, 6’’ wafer with different pixel biasing and isolations bias structure aims to reduce less-efficient area Punch-thru PTLA (a bias-dot in the 4-corner) PolySilicon (no bias dot, encircling pixel implant, ~2 MΩ ) isolation structure p-stop ~4·1012 ions cm-2, Common & Individual p-spray ~2·1012 ions cm -2 150 µm

PTLA PolySi Bias ¡rail

Biasing Scheme

IsolaCon ¡scheme

HPK ¡n-­‑in-­‑p ¡6-­‑in. ¡wafer

FE-­‑I4 ¡2-­‑chip ¡pixels FE-­‑I4 ¡1-­‑chip ¡pixels

FE-­‑I3 ¡1-­‑chip ¡pixels

FE-­‑I3 ¡4-­‑chip ¡pixels

have been investigated in test beams

16

Thin bulk sensors: test beam results

150 µm

black ¡ ¡ ¡: ¡avg. ¡non-­‑irradiated red ¡ ¡ ¡ ¡: ¡avg. ¡irradiated

Bias ¡voltage ¡[V] Q ¡[e] ¡(Arbitrary ¡unit)

Bias ¡voltage ¡dependence ¡of ¡collected ¡charges

Note: ¡independent ¡charge ¡calibraCon ¡for ¡non-­‑irradiated ¡and ¡irradiated ¡samples. ¡ ¡

“ToT charge” in arbitrry units

p-­‑stop ¡common, ¡Poly-­‑Si p-­‑stop ¡individual, ¡PTLA

#5 #6

8 6 4 2 8 6 4 2

p-­‑stop ¡individual, ¡Poly-­‑Si p-­‑stop ¡common, ¡Poly-­‑Si

#3

#4

[μm] [μm]

1

2D ¡efficiency ¡map ¡of ¡a ¡pixel, ¡non-­‑irrad, ¡at ¡150 ¡V ¡bias ¡voltage

non-irrad: FDV~40 V, saturated > 40 V

• perated at 150 V bias voltage in test beam

shown are sub-pixel resolved efficiency

• f non-irradiated FE-I4 samples

p-stop common on Poly Silicon shows best efficiency map slight inefficiency beneath the bias rail

4 FE-I4 samples also investigated after irradiation 2·1015 neqcm-2 (FE-I4) 2x PolySi-common p-stop: KEK4,5 PolySi-individual p-stop: KEK3 PTLA-individual p-stop: KEK6

17

Slim edges: DRIE

control potential drop along edge cut and reduce inactive area DRIE (Deap Reactive Ion Etching, CNM/IFAE, FBK/LPNHE, VTT/MPI&LAL) needs deep trenches, aspect ratio (1:20) polysilicon filling width of 8-12µm encouraging results on FBK test diodes electrical characteristics rely on well defined trenches; ongoing improvements

doi:10.1016/j.nima.2011.04.050

18

Slim edges: SCP

Laser scribe & cleave & edge passivation keeps edge as a clean crystal boundary passivation defines edge state and prevents conductive channels by SiO2 oxide growth for n-type atomic layer deposition or Al2O3 for p-type encouraging results; in collaboration with US Navel Research Lab

also see presentation by Marc Christophersen

100-wafers in production

19

Slim edges

Edge efficiency

CiS n+-in-in with slim edges, 250 µm irradiated to 4·1015 neqcm-2 data from pion test beam @ CERN significant part of edge pixel sensitive depends on HV and thickness inactive edge reduced down to ~ 200 µm design chosen for the planar IBL sensor

LUB2 4·1015 neqcm-2 (n)

minimize inactive area, avoids shingling along beam axis shifting pixel beneath guard rings

20

Planar IBL sensors

here, the edge design has been choosen during protoyping FE-I4 assemblies have been investigated including irradiations up to 5·1015neqcm-2

(IBL fluence)

tuning with low threshold is working well proved perfomance in test beam threshold “charge” hit map

21

Planar IBL sensors

IBL sensor production is ongoing first batches received and cross checked presently Under Bump Metallisation and dicing ongoing high yield

• n 2x1 FE-I4 Double Chip Sensors

including dicing step

22

Device simulations

TCAD simulations are done for electrical fields potential drops in the edge region leakage currents & breakdown behaviour charge generation & transport

unirradiated irradiated V[V] V[V] p n 5·1015neqcm-2 leakage current I[A]

from irradiated devices for input & cross check for simulation

23

Dopant profiles

SIMS: Secondary Ion Mass Spectrometry to determinde dopant density profile SRP: Spreading Resistance Method for carrier density profile analyzing test structures on n+-in-n and n-in-p sensors

n+ pixel implant n+-in-n pixel structure

comparison with measurements

to calibrate simulations and improve sensor design for the HL-LHC

24

SLID: Solid Liquid Inter-Diffusion

less process steps compared to bump bonding flexible in geometries and pitches however needs high planarity, no rework possible needs good opening in the isolating BCB

25

SLID: Solid Liquid Inter-Diffusion

irradiations 2·1015 neqcm-2 (n) reactor neutrons 6·1014 neqcm-2 (p) 26MeV protons charge fully recovered @ both conditions test beam with non-irradiated n-in-p device show high efficiency in center of pixel: 99.3%

Qirrad/Qnon-irrad

300 µm

26

Conclusion: Planar Pixel Sensors

now

are suited for coming LHC detector upgrades radiation hardness general understanding & performance lower costs

27

Acknowledgement

This presentation includes contributions from the whole Planar Pixel Sensor R&D Collaboration CERN, AS CR, Prague (Czech Rep.), LAL Orsay (France), LPNHE / Paris VI (France), University of Bonn (Germany), HU Berlin (Germany), DESY (Germany), TU Dortmund (Germany), University of Goettingen (Germany), MPP und HLL Munich (Germany), Università degli Studi di Udine – INFN (Italy), KEK (Japan), IFAE-CNM, Barcelona (Spain), University of Liverpool (UK), UC Berkeley/LBNL (USA), UNM, Albuquerque (USA), UCSC, Santa Cruz (USA) and especially from the PPS test beam group.

28