CRES Control Using CDA as a Shielding Gas

Transcription

1 ti CRES Control Using CDA as a Shielding Gas Steve Austin, Gary Grayson and Al Wegleitner Texas Instruments, Dallas, TX June 11 th

2 Outline Background the journey Problem Statement Measurement/Methodology Key Findings Solution's Acknowledgments Page 2

3 Background Factory Expansion Phase 4 just completed Phase 4 automotive factory was experiencing significant CRes yield loss issues between testers and devices Phase 5 to ramping ~ 150 systems migrating to automotive 3 insertion test methodology Page 3

4 Problem Statement Factory was experiencing significant CRes yield loss issues between testers and devices during 3 insertion probing sequence at different temps CRes fails were causing significant yield loss and requiring significant reprobe time losses Traditional CRes solutions were reaching the limits of effectiveness Page 4

5 Objective Show the effects of CDA on CRes Review the potential impact of CDA that effect die temperature Review design/hardware procedures needed to manage CDA Share key learning's the journey Page 5

6 CRes Sources Probe Bond Pad Surface Test Program Diode Structure PC Planarity Cleaning Frequency Debris Field Test Head Level Z-Axis Travel Chuck Speed Muffin Fans Air Handlers Raised Floor Equipment Density H 2 0? Conditions that effect CRes M a x C R E S i n O h m s Typical CRES performance curve in SCT and EBT. VLB105 ~ 7 ohms ~ 7 minutes Page 6

7 Cres Causes by Category Nearly all are out of the control of operations Page 7

8 Cleaning Frequency change most common approach to control CRES at probe. Page 8

9 Cleaning Effectiveness A Paradigm Shift Improving Contact Resistance without Increasing on-line Cleaning? Injecting Compressed Dry Air (CDA) on the Die Under Test has improved the stability of CRES and reduces the need for additional on-line cleaning Page 9

10 Discovery Experiments Blanket Aluminum Wafer Cantilever Dual-Site Probe Card VLCLT Tester CDA Nitrogen Maximum Contact Resistance (MAXCRES) used to plot effects Page 10

11 Single Power Pin Maximum Contact Resistance (MAXCRES) 10 Ω 20 Ω Probe Card Needles 13 Ω Die Under Test 12 Ω 11 Ω 12 Ω Page 11

12 100 Max Contact Resistance wo/w CDA No CDA CDA Ohms Contact Resistance Page Die 600 Die

13 Effectiveness of CDA vs. Nitrogen Purge Induced vibration Normal CDA Nitrogen Page 13

14 Oxidation Theory Consider the Mechanical Contact to Al Pads A native oxide layer of Al2O3 exists on the Al pads to be tested at probe Some of this native oxide is removed by the mechanical probing & good contact is made by mechanical abrasion However, once micro-contacts are made, local heating can lead to oxide growth in the presence of air, leading to the formation of Al2O3 ( GT ~ - 400kcal, therefore spontaneous) In the presence of moisture, hydrated aluminum oxides (Al2O3+H2O) are formed which can be thicker & more porous than the native Al2O3 The hydrated oxides are colorless to white and can lead to resistive or semiconducting contacts 1 CDA Purge Effects on Contact Process Elimination of moisture and therefore, hydration of the oxide 1 Broz & Rincon, Probe Contact Resistance During Elevated Temperature Wafer Test, ITC Proceedings, p. 396, Page 14

15 CDA Contamination Within Factory Specs. Moisture readings pre and post filter (Spec ~ 100 ppm) Pre filter was at 9.02 ppm after 14 hours of analysis Post filter was at 10.8 ppm and dropping after only 2 hours of analysis Hydrocarbon (THC) readings from pre and post filter (Spec ~ 5 ppm) Both Samples were analyzed using a GC/FID for total hydrocarbons (THC) as CH4 per the site specification for CDA The pre filter sample averaged 3.26 ppm THC as CH4 The post filter sample averaged 3.46 ppm THC as CH4 Page 15

16 Page 16

17 CDA vs. Non - CDA Results MP3 Yield Percentage With and Without CDA BIN % 95.0% Count of DIE_NUM % 85.0% % 75.0% % 65.0% % 55.0% % 45.0% % 35.0% % C CDA V3D620B09CDA 125C CDA Off V3D620B PROBE_CNT PROG CDA_ON TEMP WAFER 11 1 Page 17

18 CDA vs. Non - CDA Results Contact Sensitive Fallout (Bins 41, 55, 74, and 201) With and Without CDA 180 Count of DIE_NUM BIN C CDA V3D620B09CDA 125C CDA Off V3D620B09 0 PROBE_CNT PROG CDA_ON TEMP WAFER Page 18

19 CDA vs. Non - CDA Results Contact Sensiive Fallout (Bins 41, 55, 74, and 201) With and Without CDA Page Count of DIE_NUM VLB160 VLC45 VLC93 129C 125C CDA No CDA V3D944X03CDA V3D944X03 0 PROBE_CNT PROG CDA ON TEMP TESTER WAFER BIN

20 CDA influence on die temperature Response of ICCQ to chuck temperature was characterized with no CDA. CDA was injected at room temp (30 C) and hot chuck(125 C). Changes in ICCQ were recorded The equation (CDA Temp Chuck Temp) / 45 C closely defines the effect of CDA on die temperature. CDA temperature is 20 C typical Die temp is effected by <1 C at 30 C chuck temperature Die temp is effected by 2-4 C at 125 C chuck temperature Page 20

21 Effect of chuck temperature on ICCQ current ua ICCQ Current Dev A Dev B Chuck Temperature Page 21

22 CDA Hardware Example Page 22

23 No Measurable ESD Buildup Measured ESD on probe needles ACL Model 300B Electrostatic Locator Two testers CDA on needles Measure every 15 minutes No Measurable static build up <1v Page 23

24 Contact Stability - By Week Opens Before/After Immediate Reprobe - First Probe in EBT/SCT 70% 60% 50% 40% 30% 20% 10% 0% Page % 0.80% 0.70% 0.60% 0.50% 0.40% 0.30% 0.20% 0.10% 0.00% % Opens Changed May-Wk1 May-Wk3 Jun-Wk1 Jun-Wk3 Jun-Wk5 Ju1-Wk2 Ju1-Wk4 Aug-Wk2 Aug-Wk4 Sep-Wk2 Sep-Wk4 Oct-Wk1 Oct-Wk3 Nov-Wk1 Nov-Wk3 Dec-Wk1 Dec-Wk3 Dec-Wk5 Jan-Wk2 Jan-Wk4 Feb-Wk2 18-Feb 20-Feb % Opens Impact on First Pass Opens CDA Turned On % Opens Before Reprobe (right axis) % Opens After Reprobe (right axis) Goal for % Changed

25 Conclusion An in-situ method of stabilizing contact resistance using CDA (compressed dry air) Reducing oxidation and contaminate build up Maximizing yield and reducing reprobe on contact sensitive BINs Reduces the need to increase cleaning intervals. This results in longer probe card life and improved throughput. Page 25

26 Acknowledgements Steve Austin Kelly Daughtry Gary Grayson Frank Mesa Mark Gillette Robert Davis PE Techs Page 26

27 Thanks For Listening Enjoy the Conference Page 27