Hybridization of a sigma-delta-based CMOS hybrid detector

1
Hybridization of a sigma-delta-based CMOS hybrid
detector

• Kolb, K.E.a Stoffel, N.C.c, Douglas, B.c
  Maloney, C.W.a
• Raisanen, A.D.b Ashe, B.a Figer, D.F.a
  Tamagawa, T.d Halpern, B.d
• aRochester Imaging Detector Lab, 54 Lomb Memorial
  Dr., 76-A230, Rochester, USAbRochester
  Institute of Technology IT Collaboratory, 74 Lomb
  Memorial Dr., Rochester, USAcInfotonics
  Technology Center, 5450 Campus Dr., Canandaigua,
  USAdJet Process Corporation, 57B Dodge Avenue,
  North Haven, USA

2
Outline

• Introduction
• Process development
• Initial results
• Future work

2
3
IntroductionDetector

• Back-illuminated Silicon PIN diode, 250 µm thick
• 15 µm pixels, 256x128 array
• 50 V reverse bias during integration
• 0.1 pA/cm2 at 200 K (1 e- per second per pixel)
  dark current expected by design
• Multi-temperature operation

4
IntroductionReadOut Integrated Circuit (ROIC)

• CMOS readout circuitry (3T)
• Sigma-delta sampling (very low read noise)
• For more information, see Zeljko Ignjatovics
  talk Tuesday at 410pm
• Fully digital image sensor employing sigma-delta
  indirect feedback ADC with high-sensitivity to
  low-light illuminations for astronomical imaging
  applications

5
IntroductionHybridization

• 100 fill factor of illuminated surface
• Individual pixel addressing

h?
5
6
IntroductionRoadmap for Hybridization Process
Develop Process for Polymer Structure Fabrication
Design Structure for Metal Bump Formation
Evaluate Polymer Structure Efficacy
1
2
3
Polymer Structure Development
Test Sacrificial Structure (LOR) Removal
Test Metal Deposition Process
Metal Deposition
Bump Bonding Process
4
5
Evaluate Alignment Process and Tolerance
Test Reflow of Bump Metal
Evaluate Bond Integrity
6
7
8
6
7
Process DevelopmentDesign Structure for Metal
Bump Formation
Bump Metal
Bump Metal
Bump Metal
Photosensitive Resist
Photosensitive Resist
Photosensitive Resist
LOR
LOR
LOR
Bump Metal
Aluminum Contact
Silicon Substrate ROIC / Detector
7
8
Process DevelopmentPolymer Structure Fabrication
A spin speed of 2000 rpm for 45 seconds gives a
thickness of 4 µm to the LOR layer.
Key Dimensions

• Opening in the photosensitive
• resist over the contact
• pads (7µm).

7 µm

• Base opening of the LOR
• layer around the contact
• pads (8-9µm).

4 µm

• Height of the LOR layer in the polymer structure
  (4µm).

8-9 µm
All points received a bake at 170 C for 5
minutes. 170 C was chosen based on etch rate of
the LOR (a separate experiment) and the bake time
of 5 minutes is based on the manufacturers
recommendation.
8
9
Process DevelopmentPolymer Structure Fabrication
Before fabricating the sacrificial polymer
structure on working die, the structure was first
built on bare silicon wafers to prove the
viability of the fabrication process.
7 µm
7 µm
4 µm
4.1 µm
9 µm
8-9 µm
9
10
Process DevelopmentMetal Deposition Process

• Collimated jet of gas (Argon/Helium)

• Bump metal is vaporized and enters gas jet

• Evaporated metal atoms bond together into
  nanoclusters

• Highly perpendicular trajectory at impact

Vaporization Chamber
Collimated Jet V103 cm/s
Substrate
Inert Gas Into the System
Choked Flow
Localized Deposition
Vapor Source
10
11
Process DevelopmentEvaluate Alignment Process

• Mirrors show alignment marks on both die and an
  operator can align them using the tools relative
  positioning

Alignment Mark

• Verification of alignment after hybridization is
  accomplished with a real-time infrared optical
  system.

11
12
Initial ResultsPolymer Structure Fabrication
Polymer structure before metal deposition.
Photosensitive Resist Opening is 7.0 µm in
diameter, the LOR Base Opening is 8 µm, and the
LOR Height is 4.0 µm.
12
13
Initial ResultsSacrificial Polymer Structure
Efficacy
AuSn deposited on top of the sacrificial polymer
structure
Indium deposited on top of the sacrificial
polymer structure
Missing Overhangs in Photosensitive Resist Layer
13
14
Initial ResultsPolymer Structure Removal and
Reflow
AuSn before reflow
AuSn after reflow
Indium after deposition
14
15
Initial ResultsBond Integrity
Image with focus set at bump plane to highlight
the material left after pull test (demonstrating
that failure occurred inside the bumps rather
than at the interface between the bumps and the
bonding pads). 5 kg bond force was used to bond
and 2.2 kg pull (separation) force was required
to separate the die.
15
16
Initial ResultsBond Integrity Evaluation

• Pull test gives strength of bond between die
• Best failure mode is ductile
• taffy pull

Examples of taffy pull failures
16
17
Future WorkPlanned

• Increase strength of photosensitive resist
  overhangs to withstand JVD metal deposition
• change thickness / material for photosensitive
  resist
• Once bump size is correct, hybridize and test
  bond integrity
• Electrical testing / characterization of the
  hybridized device

17
18
Future WorkAlternative Materials for Polymer
Structure
Sacrificial polymer structure (new photosensitive
resist on top of LOR) on a bare Silicon wafer.
All dimensions are small, but this can be
addressed in future process development.
19
References

• (1) Gorski, Mike Halpern, Bret. "Jet Vapor
  Deposition for AuSn Solder Applications".
  Advanced Packaging, February, 2003.
• (2) Komarenko, Paul et. al. "Jet Vapor
  Deposition". Chapter 18 in Handbook of Deposition
  Technologies for Films and Coatings, 3rd Edition,
  edited by Peter Martin, Elsevier, 2009.
• (3) Ingjatovic, Zeljko et. al. Fully digital
  image sensor employing sigma-delta indirect
  feedback ADC with high-sensitivity to low-light
  illuminations for astronomical imaging
  applications. SPIE Conference Proceedings,
  Astronomical Telescopes and Instrumentation.
  June, 2010.

19
20
Acknowledgements
This material is based upon work supported by the
National Aeronautics and Space Administration
under Grant NNX07AG99G, issued through the
Astronomy and Physics Research and Analysis
Program of the Science Mission Directorate. Part
of this research was performed in the Rochester
Imaging Detector Laboratory and was supported by
a NYSTAR Faculty Development Program grant for D.
Figer.

• The author would also like to thank
• BAE Systems, Inc.
• Dr. Karl Hirschman (Rochester Institute of
  Technology)
• Christopher Shea (Rochester Institute of
  Technology)
• Almon Fisher (Infotonics Technology Center)

20