Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN Ch. Brönnimann, SLS Detector Group Bump Bonding of PILATUS Multichip Modules Christian Brönnimann Swiss Light Source SLS 5232 Villigen-PSI Pixel Apparatus for the SLS PILATUS Outline of the talk: 1. Project and component description 2. PSI bump bonding process 3. What can go wrong
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Bump Bonding of PILATUS Multichip ModulesChristian Brönnimann
Swiss Light Source SLS5232 Villigen-PSI
Pixel Apparatus for the SLS
PILATUS Outline of the talk:
1. Project and component description
2. PSI bump bonding process
3. What can go wrong
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Swiss Light Source: The Beamlines
10m
LINAC (100 MeV)
Booster (2.4 GeV)
Storage Ring (2.4 GeV, 288 m)
Undulators
DiagnosticsXA
S
1L
2S
3M
4S
5L
6S7M
8S
9L
10S
11M
12S
IRVUV
MYTHENDetector
PILATUSDetector
Surfaces/Interfaces Spectroscopy
Surfaces/Interfaces Microscopy
ProteinCrystallography
Material Science
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
PILATUS DetectorFinal Size: 40 x 40 cm2 (0.16m2)
2000 x 2000 pixels
Modular detector -> dead area ~6%Frame rate: >10Hz
High duty cycle: <6% (Tro~6ms)
6x3 modules (2003)
Sensor
RO-Chips
Module Control Board(RO-electronics) High Density
Interconnect (HDI)
Al-Support
1 Module: (35 x 80 mm2)
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Module Geometry
+
(0.7,0.7)(-0.2,-1.285)
(1.85,-1.18)
(0,0)
+(79.422,34.069)
(80.122,34.769)(79.620,35.355)
18.2
6m
m0.
60m
m
35.4
8m
m36
.64
mm
37.6
5m
m38
.60
mm
6.24 mm
9.90 mm
79.77 mm
80.80 mm81.20 mm
STEP: 9.982
Active Area: X (8*44+7*2)*0.217=79.422 Y(2*78+1)*0.217=34.069
Ch. Brönnimann, SLS Detector Group
SLS06 Chip: Specification
Radiation tolerant DMILL technologyAvailable since Mai 20016”-wafersChip size : 20 x 10 mm2, (44 x 78 pixels)64 chips/ wafer Yield: 30% Low noise analog block (ENC tot = 70 e-)Shaping time tsh = 100 ns -> 1MHzIndividual threshold adjustment Low overall threshold variation (σ < 100e-)15 bit pseudo random counter (Semi static)chip size 44 x 78 pixels (20 x 10 mm2)Each pixel is XY-addressableRead-out time: 6.7ms (at 10 MHz)
Ch. Brönnimann, SLS Detector Group
Sensors
• Colibrys SA, Neuchatel, Switzerland• 4” wafer, single sided process• Si-N-Material, 6-10kΩcm, d=0.3mm• Backside: n+ implant, aluminized• Vdep ~+30 V, Vbias ~+90V• UBM deposition Ti/W, Au
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Indium Bump Bonding (PSI)Developed for CMS by R. Horisberger, S. Streuli, F. Glaus
Aluminum - Chip Pads
Under Bump Metal
Foundry or LNS (PSI)
Photoresist for lift off
LMN (PSI)
Indium
Reflow Bump diameter 18 µm
Lift-off
Single Chips
10mm Wafers
15mm Wafers
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Bump bonding (cont)Sensors:- Wafers with Indium- Dicing- Reflow
Chips:- Wafers with UBM+Indium- Dicing- Testing of the chips
-Bump pressing-Reflow of bump bonded assembly
PILATUS Module
Ch. Brönnimann, SLS Detector Group
PILATUS ModuleModule Data• Active Area: 79.6 x 35.3 mm2
(continuously sensitive)
• 157 x 366 = 57462 pixels
• 16 chips (radiation hard)
• Pixel size 0.217x0.217 mm2
• Readout-time: 5 ms
• Energy Range:Eγ >6 keV
• Minimum Threshold 3 keV
• Threshold adjust per pixel
• Threshold RMS 6% of threshold
• Rate: 1MHz/pixel
• 15-bit counter/pixel
• single photon counting, no readout noise
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Indium Evaporation
• 4,6,8” wafers
• single wafer evaporation
• 4 wafers/day
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Sensor Liftoff
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Sensor after Reflow
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Sensor after Reflow
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Sensor after Reflow
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
SLS06 Chip: Testing
• Manual Probestation• Chips are tested after dicing
Probecard
6” CMOS Wafer (diced)
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Bump bonder
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Geometrical verification
Deviation from x=341,y=0
-4.0
-3.0
-2.0
-1.0
0.0
1.0
2.0
3.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14
gap nr
[um
] dxdy
(341,0)
Sensor
Chips
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Pull tester, removing of chips
Pull Tests:
- Good chips hold > 300 g (~0.1 g / bump)
- Pull each chip with 75 g
- Miss-positioned chips hold 0 g
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
• Module with 54912 bumps
•55Fe (6keV) illumination
• Result corrected by a map of digital+analog defects+ high threshold: 3293 pixels (5.9%)
• 217 pixels give no response for Fe55 (0.39%)
• If test is done with 90Sr, < 50 bad pixels
Bad bond map (Mod 9)
Eric F. EikenberrySLS Detector Group
PILATUS Detector at Spring-8 BL38B1
Lysozyme1° rotation10 s exposure
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Modul Nr Date Chips Remarks Modul Nr Date Chips Remarks
Wafno Electrically Defect Wafno Defects
Module 1 Aug 01 4 2 chips D1 Jan 03 15,17 0 chips 1 bad chip W15
Module 2 Aug 01 1 2 chips D2 Jan 03 15,17 0 chips 7 bad chips W15
Module 3 Oct 01 3 4 chips D3 Jun 03 18 ? ?
HDI 1 Oct 01 2 2 chips D4
HDI 2 Oct 01 5 2 chips D5
Module 4 Nov 01 6 1 chip Bank D6
Module 5 Nov 01 3,5,10 1 chip Bank D7
Module 6 Feb 02 10,11 0 chips Bank D8
Module 7 Mar 02 10 Unusable Sensor overetched D9
Module 8 Jun 02 12 0 chips Overheated D10
Module 9 Jan 03 13,14 0 chips good D11
Module 10 Jan 03 14,15 0 chips 3 bad chips W15 D12
Module 11 Jan 03 17,18,20,21,22 0 chips good D13
Module 12 Jun 03 18 0 chips good D14
Module construction history
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
What can go wrong:Module 7: Scanning electron microscope pictures
• Passivation topology
• Passivation attacked by UBM etching
• lift-off with MAR 404 etched Al away
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
What can go wrong:Modules 10,D1,D2: Problems with Chips from wafer 15
Scanning electron microscope analysis with EDX, Auger electron spectroscopy (surface sensitive)
0 1 2 3 4 5 6
-200
20406080
100120140160180200220240260280
3.3/4.40.6/3.4
24/14
TiSi
Al
C20A_intens: Pad2 C20B_intens: Pad2 (im Kratzer)
#X-R
ays
Energy [keV]0 1 2 3 4 5 6
0
50
100
150
2.7/2.60.6/1.1
0.9/1.78.5/0.9
0.09/0.070.17/0.21
8/40.5/13
1.7/1.360.6/3.6
AuTi
Al
Si
In
C17BUMB1: Pad C15BUMP1: Pad C15BUMB2: Indium C15BUMB3: Ring
#X-R
ays
Energy [keV]
C15bumb1C17bumb1
C15bump2
C15bump3
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN
Ch. Brönnimann, SLS Detector Group
Summary and Outlook
• PSI Bump-Bonding Process based on a lift-off process with Indium
• Processing of the 4 and 6”-wafers is done professionally in the clean-room of the LMN at PSI
• Bump-bonding of big multi-chip modules is working
• Problems at the processing level are solved: Tools like SEMs, EDX, AES are very helpful for debugging
• Visual and mechanical tests are very efficient and do not harm the modules (as opposed to electrical tests)
• Rework on mis-positioned chips is routinely done
• Plan for this year: Fabricate 20 more modules of this type
• Plan for the future: Upgrade clean-room for 8”-processing,
• Ev. move bump-bonding equipment to clean room.
Workshop on Bonding and Die Attach Technologies, 11-12 June 2003 CERN