Development of High Melting Point, Environmentally Friendly Solders, Using the CALPHAD Approach

1
TOFA 2008, Discussion meeting on Thermodynamics
of Alloys, Krakow, Poland, 22-27 June 2008.
Development of High Melting Point,
Environmentally Friendly Solders, Using the
CALPHAD Approach
Vivek Chidambaram, John Hald, Jesper Hattel
Department of Mechanical Engineering, Technical
University of Denmark, Produktionstorvet,
Building 425, 2800 Kgs.Lyngby, Denmark E-mail
vic_at_mek.dtu.dk
2
Outline/ Overview

• Introduction
• Required properties
• Ternary combinations adhering solidification
  criterion
• Scrutinization
• Analysis of Sn based systems
• Concluding Remarks
• Acknowledgements

3
High Melting Point Solders

First level Packaging
First Level Packaging Interconnection of chip to
carrier
Miniaturization
Wire bonding
Flip Chip Assembly
High Melting Point Solders
Currently used Pb-10Sn (275-302C), Pb-5Sn
(308-312C)
Applications 1.DC/DC Converter 2.Notebook
computers 3.Desktop Servers 4.High frequency
applications 5. Electronic diesel control for
Automotive.
4
Advanced Electronic Packaging
Level 1 Packaging High Melting Point Solders
Level 2 Packaging Low Melting Point Solder
SnAgCu (217C)
High Melting Point Low Melting Point plus 50C
or more
Ref. 04121279 EC exemptions, final report (ERA
Technology)
5
Required Properties in the Substitute
6
Solidus/ Liquidus (Equilibrium Solidification-CALP
HAD)
All well assessed binary and ternary systems in
COST 531 (V.2) SSOl2 databases were
extrapolated using CALPHAD approach to predict
the melting range and were optimized for a
narrow solidification range.
All Compositions are in mole-fraction
7
Non-Equilibrium Solidification (Scheil Module-
ThermoCalc)
Scrutinization of Equilibrium predictions
8
Commercial Interests Availability
Cost-Effectiveness
One of the global concerns is will the chosen
replacement be available in sufficient quantity
to supply the needs globally? and the
microelectronics industry is extremely cost
conscious.
Availabilty (31-12-07)
Cost (28-03-08)
Units 1000 metric tons
Units US/Kg
9
Wetting

• Wetting depends on a number of factors like
  wetting force, wetting
• time, surface tension and formation of IMC
  during wetting reaction.
• Low Surface tension (?) generally facilitates
  better wetting.

Source SURDAT Database
Surface tension values also influences the
natural radius of curvature
10
Miniaturization Drive
The primary reason for preferring flip-chip
assemby over wire bonding is the miniaturization
drive so it is crucial that the solder alloys
have low natural radius of curvature (R) and
easily electroplatable.
Denisty Surface tenion values- SURDAT database
Electroplating- Cheap, high productivity and
ensures miniaturization
11
Environmental Impact

• The quest for Lead-free solders is not due to
  technical reasons but
• due to environmental concerns.
• The motto would be to develop, not only
  Lead-free solders but
• Environmentally friendly solders

EPA - Environmental protective agency (drinking
water standard) OSHA - Occupational safety and
health administration
12
Scrutinization of CALPHAD Results
Ultra fine pitch bumping is currently being
enabled only by electroplating process. Ge Si
currently cannot be electroplated. Thus ternary
systems with constituents of Ge Si are not so
interesting (atleast for now).

• Sn based
• Cheap, low surface tension
• High IMC formation

• Au based
• Potential value in the solder ensures
• recyclability
• Close to Eutectic reaction
• Relatively low IMC formation

- Ternary Assessed
13
Solidification Sn-0.29Au-0.06Cu (305-322C)
Solidification Sn-0.25Au-0.11Sb (273-290C)
14
Segregation

• Elements are considered to be segregated if
  their composition in the solidified
• solid is greater than the one predicted by
  equilibrium solidification.
• Influence of segregation on thermo-mechanical
  properties depends on whether the
• phase exists in the microstructure as dispersed
  phase or matrix phase or does not exist.

Sn-0.29Au-0.06Cu
Sn-0.25Au-0.11Sb
15
Sn-0.29Au-0.08Cu Microstructure-Prediction
(Equilibrium Calculation) FCC_A1 phase is
predicted to disappear in the final
microstructure. Probably no effect of segregation
on properties.
Sn-0.25Au-0.20Sb Microstructure-Prediction (Equili
brium Calculation) SbSn phase is predicted to
exist in the final microstructure as dispersed
phase. Probably less effect of segregation on
properties.
16
Concluding Remarks
Among the proposed Sn based candidates,
Sn-0.29Au-0.06Cu is a better candidate than the
Sn-0.29Au-0.11Sb
High Sn content in the solder alloy would lead to
the formation of thick IMCs and are prone to
creep and fatigue. Then the other alternative
probably could be only Au based ones.
17
Work Ahead

• The promising solder alloy candidates would be
  prepared by
• Hot Plate Microscope.
• Solidification behaviour would be experimentally
  verified by
• Thermal Analysis.
• Intermetallic Compounds developed between Solder
  and UBM
• would be predicted by the CALPHAD approach.
• Intermetallic Compounds prediction would be
• experimentally studied and verified by
  Insitu-FIB SEM.

18
Acknowledgements

• Danish Ministry of Science, Technology and
  Development for
• financial support through the innovation
  consortium Matpack
• (project no. 07-003145).
• Dr. PeterTorben Tang, for his valuable
  suggestions and advices.

Thank You All for Your Kind Attention!!