Reliability Issues in Lead free Electronic Assemblies

1
Reliability Issues in Lead free Electronic
Assemblies

• COST MP 0602 Meeting
• IPM, Brno, CZ
• 27th August 2007
• Paresh Limaye

2
Introduction

• Reliability
• Types of Reliability issues
• Component reliability
• PCB reliability
• Solder joint reliability
• Others
• Sn whiskers
• Brittle Solder failures
• Summary

3
Reliability
Probability that the device performs as expected
for an expected duration.

• From the point of view of product seller
• Product has to perform as is promised to the
  customer AND has to last for a certain period
  (Warranty period/time till an newer version of
  the product is introduced)

• From the point of view of the end user
• Product has to perform as is promised to the
  customer AND has to last for as long as possible
  (or until the user gets tired of using the
  product)

4
Reliability Issues

• Broad categories
• Component reliability
• PCB reliability
• Solder joint reliability
• Electro-migration
• Sn whiskers
• Others.. Many more

5
Reliability Specific to Lead Free soldering

• Dealing with products, systems, specifications
  designed for tin lead backed with 50 years of
  field data
• The increased soldering temperature has a
  significant impact on product reliability.
• Degradation rate doubles with every 10oC.
• Lead-free solder has different mechanical
  properties compared to SnPb.

6
Components

• Components can be degraded/damaged during the
  reflow soldering process
• Thermal load temperature-time
• Damage to internal temperature sensitive
  structure of component electrolytes, insulating
  materials,
• Shift in electrical performance, reduced life
  span,

7
Components - Moisture

• Absorbed moisture rapidly expands during reflow
  soldering may lead to cracking of the component
  package pop-corning.
• Absorbed moisture may lead to excessive component
  warpage
• Opens/Shorts
• Poor quality solder joints

8
Printed Circuit Board laminate

• Lead-free reflow soldering / Hot Air Solder
  Leveling puts more thermal load on the board.
• Risks
• Via barrel cracking due to CTE mismatch between
  Cu barrel and epoxy laminate in Z-direction.
• Delamination
• Board sagging
• Discoloration
• Important parameters
• Glass transition temperature Tg
• Time-to-delamination T260, T288
• Decomposition temperature
• Z-expansion 50-250oC

9
Printed Circuit Board Board finish

• Lead-free HASL (Sn100C, SAC)
• High thermal load on PCB not suitable for thick
  multilayers
• Immersion Sn
• Sn whiskering in unsoldered areas
• Solderability shelf life, multiple reflow
• Immersion Ag
• Solderability sensitive to sulphur
• NiAu
• Au Embrittlement Immersion OK, electroless (?),
  electroplated NOT OK.
• Black pad (PCB manufacturer)
• Skip plating (PCB manufacturer)
• Soldering to Ni instead of Cu slower
• All cases risk of harmful chemistry residues in
  via holes.
• OSP (Organic Solderability Preservative)
• Solderability
• Invisible quality issues

10
Possible solder joint failures

• Poor quality solder joint
• Insufficient temperature-time cold joint
  (assembly)
• Excessive temperature-time brittle joint due to
  excessive intermetallics (assembly)
• Solderability issue
• Component leads (component manufacturer,
  storage)
• PCB surface (PCB manufacturer, storage)
• Incompatible metallurgy of lead-finish (design)
• Contaminated solder joint (design, assembly)
• Good quality solder joint Fracture failures
• Shock
• Fatigue
• Vibration
• Thermal cycling

11
Key aspects of SAC solders

• SnAg3-4Cu leads to increased stress levels!
• Stiffer material than SnPb significantly higher
  E-modulus. The same deformation leads to a higher
  stress level.
• Stronger than SnPb can bear higher stress levels
• Lower plasticity than SnPb.
• Higher solidification temperature leads to
  increased stress levels in the component/joint
  after solidification when thermal mismatch is
  present.
• Creep rate (deformation under constant load) is
  10 to 100 times slower than for SnPb.

12
Solder joint SAC

• SA3-4C may lead to failures elsewhere than in the
  solder!
• Intermetallic layer
• PCB pad lifting
• Component pads and body (ceramic chip)
• SA3-4C solder joints are more susceptible to
  shock.
• SA3-4C solder joints are less resistant to strong
  vibrations
• Increasing trend to move towards lower Ag content
  solders SAC 105 etc.
• Metallurgical Mess!!!

13
Thermo-mechanical Fatigue
Package has lower CTE 7-12 ppm/?C Board has
higher CTE 16-18 ppm/??C Thermo-mechanical load
is taken up by the solder
14
Thermal Cycling

• Solder joints experience creep-fatigue
• Fracture is initiated and the crack grows until
  the joint is mechanically separated

15
Microstructure Damage Zone
16
Solder joint Thermo-mechanical fatigue

• How is reliability ensured?
• Identify the loading mechanism during operation
• Identify failure mode and failure distribution
• Accelerate failure mode in testing
• Compare to qualification standards (themselves
  based on acceleration models and past
  experience/field data)

17
Solder joint Thermo-mechanical fatigue

• Accelerated thermal cycling test e.g. 0C-100C, 1
  cycle/hour
• Determination of failure distribution e.g.
  Weibull distribution
• Determination of acceleration factor with
  respect to field condition based on failure
  model e.g. Coffin-Manson, Norris-Landzberg
• Lifetime estimation under field conditions

18
Is SAC Reliable enough?Is it as reliable as SnPb?

• NO SINGLE ANSWER

Stress leveldependency(J.-P. Clech)
19
Thermal cycle experiments

• 10 mm x 10 mm x 0.68 mm WLCSP device daisy
  chained 64 I/O
• Assembled on 2.5 mm thick board (High Tg)
• Two pad sizes 250 ?m and 450 ?m
• Sn 4Ag 0.5 Cu 300 ?m and 450 ?m preformed BGA
  Spheres

20
Thermal Cycling
21
Acceleration FactorsSAC

• Norris-Landzberg equation
• N.Pan et al., HP, 2005 Salmela et. al., Nokia,
  2006 have published models for SAC
• Acceleration factor more sensitive to maximum
  test temperature as well as to the temperature
  amplitude. Increased acceleration factor compared
  to SnPb at constant dwell time.

22
Acceleration Factors Experimental vs Modeled
NfC(De)n
23
Acceleration Factors Comparison
Salmelas Model
24
Acceleration Factors

• Salmelas model accounts for the solder material
  and the component type used
• Tends to over predict the AF at higher range and
  under predicts at lower ranges
• AFs based on strain energy density show better
  correlation with the experimental observations
• We are far from understanding the real
  accelerated behaviour of SAC solders
• Creep mechanisms and their activation
• Creep behavior of the various new alloys being
  introduced through the range of temperature of
  accelerated testing

25
Solder joint Thermo-mechanical fatigue

• Reliability statements are based on accelerated
  thermal cycling tests.
• These tests have been designed for SnPb solders.
• Accelerated tests give different results
  depending on test conditions, joint
  configuration, failure criterion, stress level,..
• 10-100 times lower creep rate of lead-free
  solders reduces the acceleration factor of the
  accelerated test.
• Risk accelerated tests may overestimate fatigue
  resistance of lead-free solders!

26
Solder joint Contamination

• Solder joint contamination may have a negative
  impact on the solder joint reliability
• Pb in lead-free solder joint
• Source SnPb solderable finish, contaminated
  solder bath
• Effect tendency to form low melting phase SnPbAg
  (179oC). Weakened solder joint in last solidified
  region.
• Bi in SnPb solder joint
• Source SnBi solderable finish
• Effect tendency to form low melting phase PbBi
  (96oC). Severely weakened solder joint.
• Au in lead-free or SnPb solder joint
• Source NiAu solderable finish
• Effect formation of highly brittle SnAu
  intermetallics. Au embrittled solder joint.

27
Solder joint Contamination

• Reports from July 2007 suggest that leaded solder
  components are becoming scarce
• Assemblies relying on SnPb solders (high rel.
  applications) may end up being forced to use lead
  free components
• Long term reliability ?????
• Need to understand
• Creep Behaviour of mixed/contaminated alloys
• Accelerated and field cycling behaviour

28
Intermetallic related issues

• Other causes of potential solder joint failure
• Kirkendall voiding related to differences in
  diffusion between elements at solder/base metal
  interface.
• Effect weakened interface, cracking along
  interface
• More severe with lead-free soldering because of
  high Sn content.
• AgPd not compatible with SAC(?).
• Source of concern, not clear yet.

• Electroless Ni/Immersion Au Ni3P precipitation
• ENIG requires typically 8 P in Ni.
• Intermetallics growth leads to Ni3P precipitation
  along interface
• Brittle interface cracking
• Concern for both SnPb as well as lead-free
  soldering

29
Sn whiskering

• SnPb3-10 has been widely used as a solderable
  lead finish for components. Ban of Pb leads
  component manufacturers to go for pure Sn because
  of its low cost, availability and good
  solderability properties.
• Pure Sn whiskers!
• Tin whisker (inspection definition) A
  spontaneous columnar or cylindrical filament,
  which rarely branches, of tin emanating from the
  surface of a plating finish. (NEMI)

Kinked
Branched
30
What do we know about whiskers?

• It may create shorts under field operation
  conditions. It is NOT a production issue!
• Satellites/Cruise missiles even Nuclear plants
  affected by this
• It is not only an issue of pure Sn.
• Compressive stress in the Sn layer drives
  whisker growth.
• No quantitative view yet on impacting
  parameters.
• Several mitigation techniques no clear solution.

31
Brittleness Testing of Leadfree solders Charpy
Test
Clear ductile to brittle transition for Pb-free
solders!
32
Brittleness Testing of Leadfree solders Mini
Charpy Test
33
Brittle Solder FailuresEnergy for breaking joints
To Be published in the proceedings of EPTC 2007
34
Brittleness Testing of Leadfree solders Mini
Charpy Test
Test at -88ºC
K. Lambrinou
Potential concern for assemblies that operate in
high shock/ extreme temperature environment
Test at 23ºC
35
Summary

• Reliability issues in leadfree electronic
  assemblies various sources
• Component reliability, PCB reliability, Solder
  joint reliability, Sn whiskers, Flip Chip related
  .. MANY MORE!!!
• Higher melting temperature
• Higher solder stiffness
• Higher propensity to form intermetallics
• Various alloys and surface finishes being used
  Metallurgical Mess!!!
• Long term solder joint related effects Will
  lead free solders perform as well as tin lead?
  NO clear answer

36
Summary

• Depends on the loading condition is which the
  solder joint is expected to fail.
• Need for identifying creep mechanisms active in
  field conditions AND accelerating those in
  testing
• Need good creep and acceleration models
• Mixed leadfree-SnPb solders very little
  information
• Sn Whiskers is an issue for high reliability.
  applications for which we have no solution
• Low temperature brittle behaviour of LF solder
  can be an issue.

37
Special Thanks to

• Bart Vandevelde
• Ingrid De Wolf
• Geert Willems (www.rohsservice.be)
• IPSI/REMO Group
• Dr. Jean Paul Clech, Dr. Robert Darveaux
• ALSHIRA Partners Connectronics,TBP (Geel),
  Alcatel-Lucent (Antwepren) Multiboard, IMEC Gent,
  Interflux, Electronic Apparatus
  (http//www.imec.be/ALSHIRA)

38
(No Transcript)