stress in flip-chip solder bumps due to package warpage matt pharr es-240 project 12/9/08

Stress in Flip-Chip Solder Bumps due to Package

Warpage

Matt PharrES-240 Project

12/9/08

Flip Chip

Applied Load

Circuit Board

e e eSi

Circuit Board

e e eSi

W

MTTF = 183 hrs

MTTF = 880 hrs

Origin of Applied LoadSi – rigid,

Small CTE

Solder in Molten State

Substrate, large CTE

Cooling

Finite Element Model

200

400

400

5000

Silicon

Bismaleimide Triazene (BT) Substrate

200200

150

Underfill Solder

Material Properties

Material Young’s Modulus, E

(GPa)

Poisson’s Ratio ()

Thermal Exp. Coefficient (10-6/K)

Sn-3.5Ag Solder 50 0.3 23

Underfill 6 0.35 30

Silicon chip 131 0.3 2.8

Bismaleimide Triazene (BT)

26 0.39 15

Mesh

4-node linear coupled temperature-displacement quadrilateral

Fairly fine – why not? Refined near regions of interest

Edges and solder

Loading Conditions Step 1: 221°C – melting point of solder Step 2: 23°C

Coupled temp-disp steady state x-Symmetry Condition on Right End

Step 3 (Attempted): 1A current through solder Coupled thermal-electric Inputted thermal properties of materials Did not converge

Not sure why

Loading Conditions (cont.) Step 3: Solder and underfill at 100°C;

linear variation in substrate and Si to ambient temp of 70°C Used subroutine to define this temp field

Study 2: Ran same procedure except that it was assumed that the Si was very rigid and hence could not deform in the vertical direction

Mises Stress

Curvature agrees with intuition Slight variation (few MPa)

σ22

Stress is ~20 MPa in Solder Bumps Slight variation (~5 MPa)

Mises Stress Rigid Si

More variation in stress among solders

σ22 Rigid Si

Variation in stress in solders: ~20 MPa on right-side to ~35-40 MPa near left-side

Discussion Curvature seems physically intuitive Variation in solder location seems to have

minimal effect on stress Only ~5 MPa for σ22

I guessed it would be larger but that was assuming Si is perfectly rigid

If we make Si completely rigid, we get larger variation in stress among solders

Lessons Learned about FEA

FEA has advantages (over experiments): Relatively easy Easy to change material parameters

Do not assume FEA can handle everything Model could be wrong Solution may not converge