Formfactor, Inc. · NASDAQ:FORM
Formfactor — SWTW 2007
21 slides from Formfactor's conference presentation, 2007
Slide 1: Contact Precision Optimization to Improve Scrub Performance
Contact Precision Optimization to Improve Scrub Performance Ying Wang and Amy Leong
Slide 2: Outline
Outline ▪ Probing Trends ▪ Design of Experiments (DOE) on Scrub Sensitivity Analysis ▫ Objectives ▫ Background ▫ DOE Setup ▫ Results ▪ Conclusions ▪ Acknowledgements SWTW 2007 - 2 - FormFactor
Slide 3: Probing Contact Precision
FFI PH150 Probing Trends Probing Contact Precision ▪ Contact Precision ▫ Precise and reproducible contact on probe pads to ensure maximum yield at wafer test and subsequent process steps −Low probing pressure (minimum pad damage) −Small scrub mark −X-y precision FFI MicroSpring Contacts −Low contact
Slide 4: Industry Trends Drive Probing Challenges
Probing Trends Industry Trends Drive Probing Challenges Industrial Trends Probing Challenges ▪ Continuous shrinkage in pad ▪ Minimize yield loss due to dimensions ▫ Unreliable wire-bond from deep ▪ Thinner pad metal layer moving scrub and large particles below 0.7um ▫ Probing damage at upper metal ▪
Slide 5: Quotes from Customer Test Floors
Probing Trends - Quotes from Customer Test Floors The deeper the Scrub marks deeper scrub, the higher than x μm could lead failure percentage of to bad contact ball-non-stick reliability Scrub Performance Metric Oxidation of exposed Build-up at underlayer metal (Cu) Too much of probing end/start of
Slide 6: What Our Customers Ask For
Probing Trends What Our Customers Ask For ▪ Customer qualification requirements ▫ “No ILD cracking” ▫ “Maximum scrub depth lower than a μm after b times consecutive probing” ▫ “Maximum prow height larger than c μm after d times consecutive probing” ▫ “No underlayer metal exposures after e number of
Slide 7: Background: Macroscopic Mechanisms
Scrub Sensitivity Analysis DOE Background: “Macroscopic” Mechanisms K OT 1 0.9 0.8 0.7 0.6 Force Pressure = P ressu re Tip Size vs. 0.5 Pressure Area 0.4 steeper 0.3 0.2 0.1 0 K vs. 5 Pressure 0.5 1 more shallow 10 Tip Size 1.5 15 2 20 K 2.5 e 3 S iz FormFactor SWTW Presentation, 2004 T ip Increasin
Slide 8: Background: Macroscopic Mechanisms
Scrub Sensitivity Analysis DOE Background: “Macroscopic” Mechanisms Tip Size ▪ Macroscopically, punch through level was found to be a direct function of Large tip pressure ▫ Tip area ▫ Spring constant ▫ Planarity ▫ Over travel Standard FormFactor SWTW Presentation, 2004 Low High K SWTW 2007 - 8 - Fo
Slide 9: Background: Microscopic Mechanisms
Scrub Sensitivity Analysis DOE Background: “Microscopic” Mechanisms ▪ “Microscopically”, punch through can be caused by tip surface roughness and/or particle Particle scratches A “good” mark A “bad” mark SWTW 2007 - 9 - FormFactor
Slide 10: DOE Setup: Procedure
Scrub Sensitivity Analysis DOE DOE Setup: Procedure Spring array design incorporate most variables Tip/spring treatments into one probe head when appropriate (Tip size and shape) Scrub mark Probing experiment on characterizations (scrub representative contact depth and height) surfaces on a prober S
Slide 11: DOE Setup: Scrub Mark Characterization
Scrub Sensitivity Analysis DOE DOE Setup: Scrub Mark Characterization Raw data Height Depth Bearing ratio analysis for scrub depth and height Both height and depth values were generated based on “Bearing Ratio” analyses SWTW 2007 - 11 - FormFactor
Slide 12: DOE Setup: Matrix
Scrub Sensitivity Analysis DOE DOE Setup: Matrix ▪ A DOE was designed to assess the sensitivities of various factors on scrub performance Tip Tip TD ▪ Output parameters size shape count ▫ Scrub depth metric Tip Treatment ▫ Prow height metric X X X Condition “A” ▪ Fixed parameters for the DOE Tip Tre
Slide 13: Results: Example scrub marks from various tip designs
Scrub Sensitivity Analysis DOE Shape Results: Example scrub marks from various tip designs Size L, Shape 1 Size S, Shape 1 Size S, Shape 2 Size L, Shape 2 Size SWTW 2007 - 13 - FormFactor
Slide 14: Results: Scrub Depth Pareto Plot
Scrub Sensitivity Analysis DOE Results: “Scrub Depth” Pareto Plot t ratio > 3.0 Pareto Plot of Transformed Estimates Term Orthog Estimate Tip treatment[B] -15.99093 Tip size[Large] -10.28125 TD count[Five] 6.95375 Tip treatment[A] -3.80436 Tip size[Large]*Tip treatment[B] -2.77981 Tip shape[1]*Tip t
Slide 15: Results: Scrub Depth Interaction Profile
Scrub Sensitivity Analysis DOE Results: “Scrub Depth” Interaction Profile Interaction Profiles Scrub depth 110 metric (%) 80 Small Small Large Tip size Small 50 Tip size Large Large 20 -10 110 Scrub depth Five metric (%) 80 Five TD count One Five One 50 TD count One 20 Macroscopic -10 Microscopic 11
Slide 16: Results: Prow Height Pareto Plot
Scrub Sensitivity Analysis DOE Results: “Prow Height” Pareto Plot t ratio > 3.0 Pareto Plot of Transformed Estimates Term Orthog Estimate Tip treatment[B] -15.99093 Tip size[Large] -10.28125 TD count[Five] 6.95375 Tip treatment[A] -3.80436 Tip size[Large]*Tip treatment[B] -2.77981 Tip shape[1]*Tip t
Slide 17: Results: Prow Height Interaction Profile
Scrub Sensitivity Analysis DOE Results: “Prow Height” Interaction Profile Interaction Profiles Prow height 110 metric (%) 80 Tip size Tip size Large Small 50 Small Small Large Large 20 -10 110 Prow height metric (%) 80 TD count Five Five TD count 50 Five One One One 20 -10 Macroscopic 110 Microscopi
Slide 18: Results: Tip Treatment Impact on Scrub
Scrub Sensitivity Analysis DOE Results: Tip Treatment Impact on Scrub Tip Treatment A Tip Treatment B MicroSpring with optimized tip treatment B demonstrated much improved scrub performance SWTW 2007 - 18 - FormFactor
Slide 19: What We Can Do to Help
What We Can Do to Help The deeper the Scrub marks deeper scrub, the higher than x mm could failure percentage lead to bad contact of ball-non-stick In-depth collaborations reliability between FFI and our customers will enable contact precision and product yield optimization Oxidation of Build-up at
Slide 20: Conclusions
Conclusions ▪ Probing contact precision is critical to improve device yield ▫ Minimize yield loss from unreliable wire bond and under-layer damage ▫ Contact precision defines multi-dimensional future probing requirement ▪ FFI MicroSpring contact scrub depth can be optimized by tip size and tip treat
Slide 21: Acknowledgements
Acknowledgements ▪ The authors acknowledge the supports and advice from following colleagues ▫ Sunil Wijeyesekera ▫ Rod Martens ▫ Ken Matsubayashi ▫ Tom Napier SWTW 2007 - 21 - FormFactor