Effects of Pad Structure on the Scrub Performance of MEMS Probes for Memory Devices Aaron Woodard

Transcription

1 Effects of Pad Structure on the Scrub Performance of MEMS Probes for Memory Devices Aaron Micron Technology, Inc

2 Outline Background / Trends / Test Requirements Experiment Design Results What Ifs Summary 2

3 Background 3

4 Diversified Memory Market Compute & Networking Embedded Storage Mobile 4

5 Industry Trends 5

6 Industry Trends Implications for wafer test Smaller pad size Thinner pad metal Increased test insertions Higher die-per-wafer / increased parallelism 6

7 Generic Wafer Test Requirements Scrub Depth < Pad Thickness All Probe Marks Inside the Pad Boundary 100% Electrical Continuity Dual Temperature Capability 7

8 The Challenge Trend Device Performance Applications Increased Test Insertions Physical Variation Time-to-Market Small Pad Area Thin Pad Metal Time Balance scrub depth vs. scrub size to meet pad dimension reductions for devices with a breadth of physical properties 8

9 Experiment 9

10 Advanced Probe Cards for Memory Devices Micro-Cantilever MEMS Full 300mm Probe Arrays >100,000 Probes per Array 10

11 MEMS Probe Key Characteristics MEMS Scrub Vertical + Lateral Pad Contact Scrub Ratio (Length vs. Overdrive) Probe Force After Overdrive Tip Initial Contact Area Lateral Travel via Overdrive Initial Contact Pad Scrub Depth Scrub Length 11

12 Evaluation Items Micron and MJC evaluated the following: Evaluation Item Variables Measured Value Scrub Characteristics Single Contact Scrub Characteristics Multi-Contact Overdrive Tip Size Overdrive Tip Size Contact Count Length Width Depth Length Width Depth 12

13 Conditions and Factors Four different probe tips were evaluated on three different devices Item Wafers Prober Temperature Probe Tip Type Value Device A, B, C 300mm Wafer Prober Room Temperature 1: Baseline 2: +70% Tip Area 3: +13% Probe Force 4: +13% Probe Force, +70% Tip Area 13

14 Device Probe Pad Characteristics Pad Thickness Comparison Other properties (Pad Material, Surface Oxide, Hardness, Modulus) do not differ significantly Device A Pad Thickness 1.0x Pad Underlayer Device B Pad Thickness 6.3x Device C Pad Thickness 1.0x 14

15 Device Probe Pad Microstructure Comparison Device A Grain Size: 1.0x %GB > 15 : 50.0x Device B Grain Size: 4.6x %GB > 15 : 29.0x Device C Grain Size: 1.3x %GB > 15 : 1.0x 15

16 Results 16

17 Char Results Scrub Length, Single Contact Device A Device B Device C Operating OD Baseline +70% Tip Area +13% Probe Force +13% Force, +70% Area Scrub Ratio characterization to determine AOT/POT is device-dependent Unpredictable effect on scrub length of changing tip characteristics 17

18 Char Results Scrub Depth, Single Contact Device A Device B Device C Operating OD Baseline +70% Tip Area +13% Probe Force +13% Force, +70% Area Scrub depth response of a given probe tip varies by pad thickness Tip size is the dominant factor in determining scrub depth 18

19 Scrub Mark Operating Overdrive Baseline +70% Tip Area +13% Force +13% F, +70% A Device A Device B The Bulldozer Device C 19

20 Char Results Punch Through, Multiple Contacts Device A Device B Device C Baseline +70% Tip Area +13% Probe Force +13% Force, +70% Area Scrub depth approaches a limit at ~80-90% of pad depth Force characteristic has little to no impact on scrub depth. 20

21 Punch Through Observed on Device A Punch Maximum Overdrive Baseline +70% Tip Area +13% Probe Force +13% Force, +70% Area 21

22 Scrub Depth Images MJC Confocal Scope Device B, Baseline Probe Tip Device A, Baseline Probe Tip Scrub direction Single Contact Scrub direction Punch Through Multi-Contact Multi-Contact Multi-Contact + 22

23 Additional Considerations 23

24 What If Pad Metal Becomes Thinner? Device Operating OD Pad thickness reduction of 30% Baseline +70% Tip Area +13% Probe Force +13% Force, +70% Area Potentially punch through on the first contact with small tips Large tips may not punch through (assuming pad hardening occurs); marginal at best 24

25 What If Testing Temperature Increases? Device A, Probe Tip: +13% Force, Baseline Area Ambient High Temp Ambient High Temp Operating OD Scrub depth shows high temperature dependence; length does not Scrub depth rate of increase slows at high overdrive 25

26 What If.Pad Size Shrinks? Edge Margin Fails Larger Probe Tip Smaller Probe Tip Edge Margin Pad Boundary Overlay of Probe Mark Top, Bottom, Left, Right Edges Larger tip area = reduced scrub depth, but edge margin is strained 26

27 Summary 27

28 Summary Scrub characterization of probe types is critical to understanding advanced probing processes Depth and length can vary significantly across a breadth of similar devices To meet the challenge of reduced pad dimensions, an array of probe tip offerings from probe card suppliers is necessary Varying tip area for optimized scrub depth Varying tip size for optimized pad edge margin The right balance of tip area vs. tip size 28

29 Innovation Required: The Gap Can the probe card industry develop a one-size-fits-all probe tip for tomorrow s device requirements? Or, can probe card suppliers cost-effectively produce an array of probes tailored for specific device properties? 29

30 Final Thoughts Cooperative partnerships are critical in developing solutions for next-generation probing processes Thank You from Micron to MJC and MEC for partnering in this study Acknowledgements Micron Technology: Kurt Guthzeit, Alistair Laing MJC: Miho Kitayama, Yuma Tanaka MEC: Dick Duncan 30