-1- Nicholas Vickers June 19 th, 2008 SMTA Silicon Valley Chapter, June 19 th, 2008 Board Level Failure Analysis of Chip Scale Packages Nicholas Vickers,

Post on 13-Jan-2016

215 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

Transcript

-1-Nicholas VickersJune 19th, 2008

SMTA Silicon Valley Chapter, June 19th, 2008

Board Level Failure Analysis of Chip Scale Packages

Nicholas Vickers, Kyle Rauen, Andrew Farris, Michael Krist, Ronald Sloat, Jianbaio Pan, Ph.D.

California Polytechnic State University at San Luis Obispo

-2-Nicholas VickersJune 19th, 2008

OverviewOverview

Failure analysis Methods

Failure analysis results

Mechanical testing

Conclusions

-3-Nicholas VickersJune 19th, 2008

Failure Analysis TechniquesFailure Analysis Techniques

2 techniques used:

• Dye Penetrant• Exposes cracks cause by drop testing

• Crack area, and direction

• Cross sectioning• Locates the layer in which the crack occurs

• Identifies composition of layer cracks occur

-4-Nicholas VickersJune 19th, 2008

ResultsResults

35%

**Note- Some components show more than one failure mode

-5-Nicholas VickersJune 19th, 2008

Pad Cratering and Electrical FailurePad Cratering and Electrical Failure

Not Pad Cratered

Pad Cratered

-6-Nicholas VickersJune 19th, 2008

Cross-sectioned solder joint is shown to be cracked near the board side copper pad

Solder Fracture FailureSolder Fracture Failure

-7-Nicholas VickersJune 19th, 2008

Solder Fracture FailureSolder Fracture Failure

Cross-sectioned solder joint is shown to be cracked near the board side copper pad

Copper trace failure also shown (left side)

-8-Nicholas VickersJune 19th, 2008

Cracking Under Pads (Cratering)Cracking Under Pads (Cratering)

Epoxy on the PWB board surface cracked away from the fibers within the board, allowing the copper pad to lift away from the board

-9-Nicholas VickersJune 19th, 2008

Epoxy on the PWB board surface cracked away from the fibers within the board, allowing the copper pad to lift away from the board

Cracking Under Pads (Cratering)Cracking Under Pads (Cratering)

-10-Nicholas VickersJune 19th, 2008

Input/Output Trace FailureInput/Output Trace Failure

I/O trace gets stretched when the copper pad lifts away from the PWB

If the copper pad lifts far enough away, then ductile failure occurs in the copper trace

-11-Nicholas VickersJune 19th, 2008

Input/Output (I/O) traces that connect to the daisy-chain ‘resistor’ were often broken

Many components had this broken trace and no other identifiable failure

I/O Trace FailureI/O Trace Failure

Board sideComponent side

-12-Nicholas VickersJune 19th, 2008

I/O Trace Failure LocationI/O Trace Failure Location

-13-Nicholas VickersJune 19th, 2008

I/O Trace and Daisy-chain Trace failures are both caused by pad cratering

Pad cratering was present on 88% of electrically failed components, and is directly responsible for 69% of electrical failures

Failure Mode Comparison

-14-Nicholas VickersJune 19th, 2008

Failures After 10 Drops (No EB)

Component: Green – no failure, Blue – transitional failure, Orange – full failure, Red – complete failureSolder Joints: Black – pad crater, Red – solder fracture (board side), Yellow – solder fracture (csp side)

-15-Nicholas VickersJune 19th, 2008

Failures After 14 Drops (No EB)

Component: Green – no failure, Blue – transitional failure, Orange – full failure, Red – complete failureSolder Joints: Black – pad crater, Red – solder fracture (board side), Yellow – solder fracture (csp side)

-16-Nicholas VickersJune 19th, 2008

Failures After 325 Drops (Epoxy EB)

Component: Green – no failure, Blue – transitional failure, Orange – full failure, Red – complete failureSolder Joints: Black – pad crater, Red – solder fracture (board side), Yellow – solder fracture (csp side)

-17-Nicholas VickersJune 19th, 2008

Failures After 279 Drops (Acrylic EB)

Component: Green – no failure, Blue – transitional failure, Orange – full failure, Red – complete failureSolder Joints: Black – pad crater, Red – solder fracture (board side), Yellow – solder fracture (csp side)

-18-Nicholas VickersJune 19th, 2008

SAC305 Solder MicrostructureSAC305 Solder Microstructure

-19-Nicholas VickersJune 19th, 2008

Microhardness TestingMicrohardness Testing

-20-Nicholas VickersJune 19th, 2008

Mechanical Testing of BoardsMechanical Testing of Boards

Need to know material properties of the board to simulate using FEA methods.

Tested along fiber direction using an Instron tensile tester.

-21-Nicholas VickersJune 19th, 2008

Elastic

Region

Fibers Unkinking

σ fracture

Along Fiber Results

-22-Nicholas VickersJune 19th, 2008

Along Fiber StrengthAlong Fiber Strength

0

50

100

150

200

250

300

350

10 mm/min 30 mm/min 50 mm/min

Stre

ngth

(MPa

)

-23-Nicholas VickersJune 19th, 2008

Along Fiber ModulusAlong Fiber Modulus

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

10 mm/min 30 mm/min 50 mm/min

Elas

tic

Mod

ulus

(M

Pa)

-24-Nicholas VickersJune 19th, 2008

ConclusionsConclusions

Pad cratering is the most common failure modePad cratering does not necessarily cause electrical

failure, but can cause electrical failure by introducing other failure modes

Dominance of pad cratering indicates that solder joints are not the weakest part of this lead-free assembly

-25-Nicholas VickersJune 19th, 2008

Conclusions (Cont.)Conclusions (Cont.)

Tougher board material is needed to increase reliability

Majority of failures occurred on the cable side of the board when DAQ cable is attached

First failures usually occur in the corners of the CSPsEdge-bonding is effective at reducing pad cratering

problems

-26-Nicholas VickersJune 19th, 2008

AcknowledgementsAcknowledgements

Cal Poly: Michael Krist, Kyle Rauen, Micah Denecour, Andrew Farris, Ron Sloat, and Jianbaio Pan, Ph.D.

Flextronics: Dongkai Shangguan, Ph.D., Jasbir Bath, David Geiger, Dennis Willie

Henkel: Brian Toleno, Ph.D., Dan Maslyk

-27-Nicholas VickersJune 19th, 2008

Acknowledgements

Project Sponsors:

Office of Naval Research (ONR)

Through California Central Coast Research Park (C3RP)Society of Manufacturing Engineers Education FoundationSurface Mount Technology Association Silicon Valley

-28-Nicholas VickersJune 19th, 2008

Thank You.Thank You.Any Questions?Any Questions?

-29-Nicholas VickersJune 19th, 2008

Supplementary SlidesSupplementary Slides

-30-Nicholas VickersJune 19th, 2008

Dye Stained Solder Fractures

Dye stained solder fractures were found• Partial solder fracture (left) was not completely fractured

before the component was removed

• Complete solder fracture (right) was fully fractured before the component was removed

-31-Nicholas VickersJune 19th, 2008

Failures After 10 Drops (No EB)

-32-Nicholas VickersJune 19th, 2008

Failures After 14 Drops (No EB)

-33-Nicholas VickersJune 19th, 2008

Failures After 325 Drops (Epoxy EB)

-34-Nicholas VickersJune 19th, 2008

Failures After 279 Drops (Acrylic EB)

-35-Nicholas VickersJune 19th, 2008

Test Vehicle Drop Orientation

Test vehicle is always mounted with components face down

-36-Nicholas VickersJune 19th, 2008

Component Locations

JEDEC defined component numbering• The DAQ cable attaches near component C6 (in between

components C1 and C11)

-37-Nicholas VickersJune 19th, 2008

Blank PWB – No Cable vs Cable

• Symmetry of acceleration peaks has shifted (C7 vs C9)

• Maximums greatly reduced by cable (C3, C13, C8)

1500G Input Acceleration

-38-Nicholas VickersJune 19th, 2008

Populated PWB – No Edge Bond

• Dampening due to the cable seems less significant than with blank PWB (both graphs are more similar)

1500G Input Acceleration

-39-Nicholas VickersJune 19th, 2008

Epoxy Edge Bonded CSPs

• Stiffer board with edge bonding has less symmetry disturbance

• Overall accelerations are significantly reduced vs no edge-bond

1500G Input Acceleration

-40-Nicholas VickersJune 19th, 2008

Acrylic Edge Bonded CSPs

• Stiffer board with edge bonding has less symmetry disturbance

• Overall accelerations are significantly reduced vs no edge-bond

1500G Input Acceleration

top related