Formfactor, Inc. · NASDAQ:FORM
Formfactor — MicroProbe — Technical Presentation (June 2011)
22 slides from Formfactor's technical presentation, June 2011
Slide 1: A Flexible Vertical MEMs Probe Card Technology for Pre-Bump and eWLP Applications
June 12 to 15, 2011 San Diego, CA A Flexible Vertical MEMs Probe Card Technology for Pre-Bump and eWLP Applications Mike Slessor Rick Marshall (MicroProbe, Inc.)
Slide 2: Vertical MEMS for Pre-Bump Probe
Vertical MEMS for Pre-Bump Probe • Introduction: eWLP and Pre-Bump Probing Requirements • Experiment Objectives & Details – Accurate Probing on Small Pads at Fine Pitch – Low Scrub Depth on Pre-Bump Pads – Wafers Probed & Measurements Taken • Observed Probe Results & Conclusions – Accuracy & Repeata
Slide 3: MicroProbe: A Leading Supplier of Logic/RF/SoC Probe Card Solutions
MicroProbe: A Leading Supplier of Logic/RF/SoC Probe Card Solutions Innovation and Growth • Technology Leadership – >1000 MEMS probe cards delivered • Market Share Growth – #1 supplier of Advanced SoC Probe Cards • Customer Collaboration – 35-year history of delivering results Breadth and Stability
Slide 4: eWLP Resurrects Pre-Bump Array Probing
eWLP Resurrects Pre-Bump Array Probing • While pre-bump probing has been largely eliminated from mature BGA Flip-Chip packaging flows, the workflow and cost considerations of eWLP are re-introducing need for effective pre-bump probing on arrays of aluminum pads. • Today’s designs challenge probing o
Slide 5: What is eWLP?
What is eWLP? • eWLP = “Embedded Wafer Level Packaging” • eWLP is an evolution of BGA-type packaging that uses molded carriers and fan-out RDLs. The original die are singulated, embedded into molded carriers, and then reconstituted onto artificial wafers. Wafer-level processes then add redistributio
Slide 6: eWLP and Pre-Bump Probe
eWLP and Pre-Bump Probe • Packaging bad die into molded carriers, and subsequently attaching them to reconstituted wafers, causes very expensive yield loss for the final eWLP wafer. • For multi-die eWLP packages, the cost impact is even worse – the problem is directly analogous to test escapes findi
Slide 7: Example eWLP Test Flow
Example eWLP Test Flow Reconstitution Into Artificial Wafer Die A Wafer Sort Dice Wafer / Singulation Redistribution Layers (RDL) Integration of Printed Die B Good Die Into Wafer Sort Molded Carrier Solder Ball / Cu Pillar Attach Die C Wafer Sort Final Test June 12 to 15, 2011 IEEE SW Test Workshop
Slide 8: Pre-Bump Probe Challenges
Pre-Bump Probe Challenges • Fine Pitch in Full Grid Arrays – RDL fan-out permits increasingly fine pitch across high pincount arrays. Today’s arrays are 130um – 180um. Next generation arrays will be < 100um. • Small Pads – Bond pad openings are getting smaller: 50um octagonal pads are migrating towa
Slide 9: MicroProbe Vertical MEMS "Mx" Probe Architecture
MicroProbe Vertical MEMS “Mx” Probe Architecture • Composite MEMS structure allows optimization of mechanical and electrical design – Multiple materials & Layers – Photolithographically Defined Shape • Resulting material & geometry flexibility provides optimal contact performance and pitch scalabili
Slide 10: Customer Experiments with Mx for Pre-Bump Probing
Customer Experiments with Mx for Pre-Bump Probing • Customer A: Focus on low force contact to minimize pressure applied to circuit under pad • Customer B: Focus on good contact with minimal scrub depth into pad material • Customer C: Focus on probe tip accuracy for contacting small pads with high pr
Slide 11: Customer A: Low Force Experiment
Customer A: Low Force Experiment • Wafer Setup for Experiment – 300mm wafer at 40nm process – eWLP pre-bump pads probing – 60um octagonal pads • Evaluation Criteria – Cres must be within acceptable range for device – Probe force over active area and low-k ILD must be minimal – Scrub mark must be sma
Slide 12: Mx Probe Contact Force at Overtravel
Mx Probe Contact Force at Overtravel Probe Contact Force 8.00 7.00 6.00 5.00 Contact Force (g) 4.00 FC-150 FP-60 3.00 FP-50 2.00 1.00 0.00 0 10 20 30 40 50 60 70 80 90 100 Overtravel (um) Recommended OT = 65um June 12 to 15, 2011 IEEE SW Test Workshop 12
Slide 13: Mx Low-Force Design & Modeling
Mx Low-Force Design & Modeling • MEMS Geometry & Metallurgy – Proprietary Mx MEMS process enables multi-layer probe design with lithographically defines shapes – Focused on low-force mechanical design while maintaining excellent Cres characteristics • Design & Modeling – Detailed FEA models are deve
Slide 14: Customer A: Low Force Findings
Customer A: Low Force Findings • Mx scrub mark ranges 8um – 15um Mx-FP Scrub Cantilever Length of Scrub Length • Maximum scrub depth of 0.55um ~10um of >30um – (After 8 touchdowns) • No ILD cracking found with 60um over-drive 12.1um 8.9um June 12 to 15, 2011 IEEE SW Test Workshop 14
Slide 15: Customer B: Small Pads Experiment
Customer B: Small Pads Experiment • Test Setup for Experiment – ≈180 Die Per Wafer – ≈ 7,500 Pads Contacted per Die – 55um Pads in 180um Array – TEL P12Ln Prober with Test Temperature of 40degC June 12 to 15, 2011 IEEE SW Test Workshop 15
Slide 16: Scrub Mark Size & Accuracy
Scrub Mark Size & Accuracy • Customer Findings on Scrub Mark Placement – Typical accuracy to pad center: +/- 9um – Worst-case accuracy observed: +20um – (Mx Typical Spec = +/-13um) 50um OD • Customer Findings on Scrub Mark Size – 20um OD 7um Scrub 30um OD 20um OD – 30um OD 10um Scrub – 50um OD
Slide 17: Accuracy & Repeatability Across Wafer
Accuracy & Repeatability Across Wafer Die = -5, 0 Die = -1, 0 Die = 0, -6 Die = 0, 5 Die = 6, 0 June 12 to 15, 2011 IEEE SW Test Workshop 17
Slide 18: Customer B: Small Pads Findings
Customer B: Small Pads Findings • Mx solution delivers highly-accurate and repeatable scrub marks that are suitable for pre-bump probing small pads – Overall planarity of a large array was very good (< 24um) – Probe mark placement was very accurate, consistently placing the scrub center within +/- 1
Slide 19: Customer C: Scrub Depth Experiment
Customer C: Scrub Depth Experiment • Wafer Setup for Experiment – 300mm qual wafer selected in Engineering lab – Entire wafer probed with 4 touchdowns on every die – Lower 2/3 probed with a 5th touchdown – Lower 1/3 probed with a 6th touchdown • Scrub Mark Review Techniques – Center and edge samples
Slide 20: Customer C: Scrub Depth Findings
Customer C: Scrub Depth Findings • Pad Scrub Findings – Starting aluminum depth of 1.20um – Worst-case image: 6 touchdowns at wafer center – Aluminum depth of 0.64um shows maximum scrub depth of 0.56um – Low scrub was very repeatable across all wafer zones Passivation (for FIB) 2.85um 1.20um 0.64um
Slide 21: Conclusions & Use Benefits
Conclusions & Use Benefits • Conclusion – The Mx MEMS vertical probe solution addresses key requirements for today’s pre-bump probing: low-force, high-accuracy contact, repeatability, and low force. These characteristics can enable effective pre-bump probe for the next generation of pads testing. •
Slide 22: Follow-On Work / Q&A
Follow-On Work / Q&A • Follow-On Work – Lifetime testing on accuracy and scrub would be beneficial to understand MEMS stability and repeatability versus legacy solutions – Copper pads testing should be conducted – these studies were all done using Aluminum pads – Additional hot-temp testing would be