3D Workshop at CPPM

Download Report

Transcript 3D Workshop at CPPM

3D Workshop at CPPM
The 3-D IT:
The new challenge
for ATLAS Pixel detectors
CPPM Marseille (France): B. Chantepie, J.-C. Clémens, R. Fei, D. Fougeron, S. Godiot, P.
Pangaud, A. Rozanov
Bonn University (Germany): D. Arutinov, M. Barbero, T. Hemperek, M. Karagounis, H.
Krüger, A. Kruth, N. Wermes.
LBNL Berkeley (USA): J. Fleury, M. Garcia-Sciveres, A. Mekkaoui
March 18-19, 2010, CPPM, Marseille, France
3D Workshop at CPPM
Hybrid Pixel Detectors
for particles trackers

An early 3-D approach!!
Choice of several sensors
for particles detection
(HR Si, diamant,3D..)

Dedicated electronic chip
AND

A bump-bonding solder
for interconnection


Sensors for ionizing particles
(e-, photon, gamma, etc ..)

Electronic pixel readout



Monolithic device
Analog detection (low noise, low power)
Digital readout
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
ATLAS Pixel Detectors
for LHC/SLHC

Hybrid Pixels Detector of ATLAS/LHC


Hybrid Pixels Detector of ATLAS/SLHC



LHC : Luminosity of 1034 cm-2.s-1
SLHC expected 10 times more !!!
Like a big camera with a 1.5 m2 area and 80
Million of Pixels with a snapshot every 25ns
More luminosity, more pixels more ionizing
particles, more … !!!
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
Atlas upgrades / SLHC :
Why 3-D approach?
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
Tezzaron-Chartered 3D MPW :
ATLAS/SLHC sub-part
ATLAS/SLHC
C and D Sub-part
LBL
G Sub-part
G1
March 18-19, 2010, CPPM, Marseille, France
G2
P.Pangaud
3D Workshop at CPPM
Fermilab 3-D Multi-Project Run :
C-Band ATLAS
SEU-3D
SEU-3D
FETC4-AE
FETC4-DS
TSV Daisy Chain + BI
Electrical Test
TSV vs Transistors
Mechanical stress
DFF + Trans + Cap
Electrical Test
TSV vs Transistors
+ capacitors
FETC4-AE (CPPM) : same than FEI4_Proto1, but in Chartered 0.13LP
 FETC4-DS (CPPM) : Shift Register + counter + readout data
and ”Drum registers“
 SEU-3D (CPPM) : SEUless memories blocks
 General test structures (CPPM) : TSV + BI Daisy chain (electrical parameters) ; TSV capacitors value with and

without BackMetal and BI ; Transistors (Linear and ELT) closed to TSV ; Mechanical stress effects of devices (Trans,
Cap, Res, DFF)
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
Fermilab 3-D Multi-Project Run :
D-Band ATLAS
FETC4-AE
Electrical Test
TSV vs Transistors
TSV, Cap and Bump
OmegaPix
Analog
OmegaPix
Digital
FETC4-DC
Electrical Test
TSV vs Transistors
FETC4-AE (CPPM) : same than FEI4_Proto1, but in Chartered 0.13LP
 FETC4-DC (Bonn-CPPM) : Digital pixels Read-out "à la FEI4“ (M.Barbero)
 OmegaPix (LAL) : see (D. Thienpont ) presentation
 General test structures (CPPM) : TSV + BI Daisy chain (electrical parameters) ; TSV capacitors value with and

without BackMetal and BI ; Transistors (Linear and ELT) closed to TSV ; Mechanical stress effects of devices (Trans,
Cap, Res, DFF)
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
Fermilab 3-D Multi-Project Run :
G-Band LBNL- Fermilab
FETC4-AH
FETC4-DS
FETC4-AH (LBL-CPPM) : same than FEI4_Proto1 but with holes collection.
 FETC4-DS (CPPM) : Shift Register + counter + readout data
and ”Drum registers“

March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
FETC4 - Atlas 3-D Read-Out chip
Fermilab proposal
3D HEP consortium
Tezzaron-Chartered MPW
Tezzaron Chartered
1st MPW Run (Summer 2009)
Chartered
1st Prototype (Feb 2009)
Chartered
2nd Prototype (Nov 2009)
ATLAS 3D Read-Out chip
Pixel size : 50 x 125 µm
Pixel array size : 160 x 336
Chip size : 19 x 20 mm
Chartered
3rd Prototype (June 2010)
Tezzaron Chartered
2nd MPW Run (?)
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
FETC4_AEDS (DC) : Atlas 3-D Read-Out proto chip
2 Tiers + Sensor
sensor

Digital part : Tier 2 (C2, D2).

Sensor : Tier 0
Back Side Metal
SuperContact
M1
M2
M3
M4
M5
M6
M6
M5
M4
M3
M2
M1
Tier 1
(thinned wafer)
Analog part : Tier 1 (C1, D1)
thinned to use SuperContact
which are connected to the
Back side metal (wire and
bump-bonding interface)
Bond Interface
Tier 2

SuperContact
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
FETC4_AEDS (DC) : Atlas 3-D Read-Out proto chip
Sensor aspect
Sensor layout :
Max-Planck-Institut für Physik, Werner
Heisenberg Institut , Munich
Bump-Bonding :
IZM , Munich
Due to the bonding constraints, the geometrical sensor area had to be shrink


Sensor : 7 columns of 48 pixels
Tier 1 and Tier 2 : 14 columns of 61 pixels
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
Tezzaron-Chartered 3D MPW (C1 and D1) :
Analog Tier FETC4-AE
Based from the FEI4-P1 chip made in IBM 0.13µm technology :
Signal from Sensor,
via TSVs
2 ways for read-out the discriminator:

By the classical way in Tier1

By the FEC4-DS or DC chip (Tier 2) via the BI.
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
Tezzaron-Chartered 3D MPW (C2) :
Digital Tier FETC4-DS

3 functions :
ANALOGUE

Read-out the analog Tier 1

Noise generator (pick-up, coupling, etc…)
in front of 11 specifics area of Tier 1
(preamplifier, feed-back capacitor,
DAC…)

Several programmable shielding
configuration
DIGITAL
SuperContact
Back Side
Metal for
bonding
Tier 1
(thinned
wafer)
M1
M2
M3
M4
M5
M6
M6
Bond Interface
M5
M4
M3
M2
M1
Tier 2
SuperContact
Analog tier and Digital tiers are face-to-face with only
~5µm of distance
FE-TC4-DS is mainly dedicated to study the
parasitic coupling between two tiers
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
ATLAS chips - historical review :
FEC4-P1 and FETC4-AEDS(DC)

FEC4-P1, run 2D (submission February 2009, test April 2009) :



2D run in 0.13LP Chartered technology. 1st approach
Layout translation of FEI4-P1 IBM 0.13µm 8LM to FEC4-P1 CHRT 0.13µm 8LM
FETC4-AEDS (DC) (submission July 2009, test expecting before summer 2010) :


3D run in 0.13LP Tezzaron-Chartered technology.
Layout translation of FEI4-P1 IBM 0.13µm 8LM to FETC4-P1 CHRT 0.13µm 5LM



Move from de 8LM to 5LM
3D structures
Two additional Digitals Tiers
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
ATLAS chips - historical review :
FEC4-P1 et FEC4-P2

Test Results of FEC4-P1 :





Chip is working
The minimum threshold is # 1100 e- ; threshold resolution# 200 e-; Noise < 80 eIrradiation : the chip is working for a X-rays dose rate up to 200MRad
Defected has been found for an protons dose-rate up to 160 MRad. The memories structures
(Latches) stacked up to 1.
FEC4-P2, run 2D (submission November 2009, chip under test) :




2D run in 0.13LP Chartered technology. 2nd prototype to corrected the first one.
Analog Part : Optimization of the transistors size dedicated to the Chartered constraints technology
Digital Part : Memories latches => modification with 2 kinds of architecture.
Layout optimization.
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
FEx4 ATLAS chips : next in 2010…

FEC4-P3, run 2D (submission June 2010) :





2D run in 0.13LP Chartered technology. Hope last prototype .
To finished the translation of the last building blocks in CHRT technology : Current Ref,
PLL, LVDS, etc …
Pixel modification : to shrink pixel size to 125µm and to reduce the metal levels to 5.
Structures ready for the 3D chip
FETC4-A, run 3D (before the end of 2010) :



Very large matrix size : 336 x 80
Small pixel size : 125µm x 50µm
Twin of FEI4-A but with an half size pixel.
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud
3D Workshop at CPPM
Feedback of 3D development

3D vision

Long understanding 3D Tezzaron process





PDK

Chartered PDK is not really a PDK but pieces of PDK





Check face to face BI connections
Check TSV density
Bad layers definitions, source of mistakes


Very poor PDK with not optimized PCELLS regarding min values of rules !!!
Chartered PDK is not stable and need to be wary all the time (no support but only
access to GlobalFoundries site!!) Ever requested, never replied !!!
Sub-versions of Calibre files are not compatible between us (bad experience with
FEC4_P1)
Home-made 3D DRC-LVS calibre deck tools


Bad layer numbers definition
2 TopMetal but 1 TM for BI !!!
Backside Metal : poor rules!!
Filling!! And floating metal !!
Developed our own StreamLayers file
Now we switched to the Cadence OA version
March 18-19, 2010, CPPM, Marseille, France
P.Pangaud