Efficiency & Yield Improvements with Factory-Wide Process Control Software

Transcription

1 Efficiency & Yield Improvements with Factory-Wide Process Control Software David Genova, MS, MBA PVMC c-si Metrology Workshop July, 2012

2 Outline Factory Automation Data-Collection Options Fault Detection & Classification Real-Time Tool Monitoring Factory-Wide Data Analysis Integrated Process Control Software 2

3 Factory Automation Data-Collection Options

4 Factory Automation Data Collection to link the factory together Buy Poly Poly Charge Ingot Casting Ingot Cut Brick Cut Slice Clean Inspect Sort Package Multi Mono Cell Inspect Wet Texture P/N Junction Wet Clean Edge Isolation ARC SIN Metal Line Firing Furnace Cell Sort Cell Package MOCVD Spray Direct Remote Wet Laser Module Inspect Cells String Cells Layup Connect Strings Inspect Laminate Inspect add JB & Frame Module Sort Module Package 4

5 Factory Automation Does not have to be scary Easy to start with the tools from a single line (cell line shown) Then roll-out Factory Wide With or Without MES Direct or Manual Data Entry Time, Lot, or -ID based analysis Centralized Factory-Integration Layer TEXTURE DIFFUSION PSG ETCH EDGE ISO ARC SCREEN PRINTING FIRING TEST 5

6 Factory Automation Attribute-Data for a Cell Line (Type 1/4) ATTRIBUTE DATA Desired Data Mandatory Data Relevant Data - Sample List LOT ID ID Start Start Material Material Material Supplier Ingot ID Date Time Type Size Spec. ID Brick ID 12/16-A /16/2011 1:57 AM multi 156 mm Supplier /16-A /16/2011 1:58 AM multi 156 mm Supplier /16-A /16/2011 1:58 AM multi 156 mm Supplier /16-B /16/ :00 AM multi 156 mm Supplier /16-B /16/2011 2:00 AM multi 156 mm Supplier /16-A /16/2011 1:58 AM multi 156 mm Supplier /16-A /16/2011 1:58 AM multi 156 mm Supplier /16-A /16/2011 1:59 AM multi 156 mm Supplier /16-A /16/2011 2:00 AM multi 156 mm Supplier /16-A /16/2011 2:01 AM multi 156 mm Supplier /16-A /16/2011 2:01 AM multi 156 mm Supplier /16-A /16/2011 2:01 AM multi 156 mm Supplier /16-A /16/2011 2:02 AM multi 156 mm Supplier /16-A /16/2011 2:02 AM multi 156 mm Supplier /16-A /16/2011 2:02 AM multi 156 mm Supplier /16-A /16/2011 2:02 AM multi 156 mm Supplier /16-B /16/2011 2:03 AM mono 156 mm 200mm 12/ /16-B /16/2011 2:03 AM mono 156 mm 200mm 12/ /16-B /16/2011 2:03 AM mono 156 mm 200mm 12/ /16-B /16/2011 2:04 AM mono 156 mm 200mm 12/ /16-B /16/2011 2:04 AM mono 156 mm 200mm 12/ Thickness Spec. Data is usually manually entered, or sometimes collected via an MES or factory-integration layer. More is better! 12/16-B /16/2011 2:04 AM mono 156 mm 200mm 12/ /16-B /16/2011 2:05 AM mono 156 mm 200mm 12/

7 Factory Automation WIP-Data for a Cell Line (Type 2/4) LOT ID Desired Data ID Line ID Tool Path Date Time Tool ID Date Time Tool / Tube Zone ID WIP DATA Mandatory Data - Sample List (repeat for every tool) Texturing Etch Diffusion PSG Etch SIN Printing 1-3 Firing Furnace 12/16-A Line A 5:56:02 PM Tex-A 6:28 PM POCL-A-R4-1 10:26:15 PM PSG-A 11:04 PM SINA-A 1:22 AM SP 1A 1:55:04 AM FF-A 12/16-A Line A 5:56:04 PM Tex-A 6:28 PM POCL-A-R4-1 10:26:18 PM PSG-A 10:59 PM SINA-A 1:07 AM SP-1B 1:56:06 AM FF-A 12/16-A Line A 5:56:07 PM Tex-A 6:28 PM POCL-A-R4-1 10:26:21 PM PSG-A 11:03 PM SINA-A 1:22 AM SP-1A 1:55:12 AM FF-A 12/16-B Line B SINA-A 11:22 PM SP-1B 11:57:16 PM FF-A 12/16-B Line B SINA-A 11:23 PM SP-1A 11:59:26 PM FF-A 12/16-A Line A 10:55 PM SINA-A 1:22 AM SP-1B 1:56:32 AM FF-A 12/16-A Line A 11:02 PM SINA-A 1:23 AM SP-1A 1:56:44 AM FF-A 12/16-A Line A 6:06:03 PM Tex-A 6:28 PM POCL-D-R1-2 10:33:12 PM PSG-A 11:06 PM SINA-A 1:24 AM SP-1B 1:57:16 AM FF-A 12/16-A Line A 10:36:15 PM PSG-A 11:13 PM SINA-A 1:25 AM SP-1A 1:58:18 AM FF-A 12/16-A Line A 6:05:00 PM Tex-A 6:28 PM POCL-D-R2-4 10:33:18 PM PSG-A 11:04 PM SINA-A 1:25 AM SP-1B 1:59:56 AM FF-A 12/16-A Line A 5:56:12 PM Tex-A 6:28 PM POCL-D-R2-5 10:26:43 PM PSG-A 11:03 PM SINA-A 1:25 AM SP-1A 1:59:13 AM FF-A 12/16-A Line A 10:36:45 PM PSG-A 11:03 PM SINA-A 1:25 AM SP-1B 1:59:12 AM FF-A 12/16-A Line A 6:08:03 PM Tex-A 6:36 PM POCL-C-R3-3 10:38:46 PM PSG-A 11:11 PM SINA-A 1:26 AM SP-1A 2:00:05 AM FF-A 12/16-A Line A 6:07:04 PM Tex-A 6:36 PM POCL-C-R3-3 10:38:04 PM PSG-A 11:11 PM SINA-A 1:26 AM SP-1B 2:00:08 AM FF-A 12/16-A Line A 6:08:13 PM Tex-A 6:36 PM POCL-C-R3-3 10:38:07 PM PSG-A 11:12 PM SINA-A 1:26 AM SP-1A 1:59:12 AM FF-A 12/16-A Line A 11:16 PM SINA-A 1:27 AM SP-1B 2:00:17 AM FF-A 12/16-B Line B 11:16 PM SINA-B 1:27 AM SP-2A 2:01:28 AM FF-B 12/16-B Line B 6:10:00 PM Tex-B 6:36 PM POCL-C-R3-3 10:38:00 PM PSG-B 11:10 PM SINA-B 1:27 AM SP-2B 2:01:32 AM FF-B 12/16-B Line B 11:10 PM SINA-B 1:27 AM SP-2A 2:01:35 AM FF-B 12/16-B Line B 6:04:00 PM Tex-B 6:28 PM POCL-D-R1-3 10:36:02 PM PSG-B 11:16 PM SINA-B 1:28 AM SP-2B 2:02:38 AM FF-B 12/16-B Line B 6:04:03 PM Tex-B 6:28 PM POCL-D-R1-3 10:36:05 PM PSG-B 11:16 PM SINA-B 1:29 AM SP-2A 2:02:46 AM FF-B 12/16-B Line B 6:04:06 PM Tex-B 6:28 PM POCL-D-R1-3 10:36:07 PM PSG-B 11:16 PM SINA-B 1:29 AM SP-2B 2:02:52 AM FF-B 12/16-B Line B 6:04:09 PM Tex-B 6:28 PM POCL-D-R1-3 10:33:10 PM PSG-B 11:12 PM SINA-B 1:29 AM SP-2A 2:03:03 AM FF-B Data is usually automatically collected (via an MES or factory-integration layer) Note: -ID eliminates the need for individual-tool connections and MES Date Time Tool ID Date Time Tool ID Date Time Tool ID Date Time Tool ID 7

8 Factory Automation Metrology-Data for a Cell Line (Type 3/4) Desired Data METROLOGY AND QUALITY CONTROL DATA Relevant Data - Sample List (more is better) Incoming Inspection Tex Depth Diffusion SIN Metal Line LOT ID ID Width Length Pre-etch Pre-etch Life Bulk Post Sheet Mean Std Weight Thick time Res Weight Rho 1 - Sheet Rho 5 Thick R.I. Finger Finger Width Height 12/16-A /16-A /16-A /16-B /16-B /16-A /16-A /16-A /16-A /16-A /16-A /16-A /16-A /16-A /16-A /16-A /16-B /16-B Data is manually-collected for offline tools and automatically-collected for inline tools Note: -ID makes -to- correlation possible, otherwise Lot-to-Lot correlation is the highest achievable 12/16-B /16-B /16-B /16-B /16-B

9 Factory Automation Test & Sort Data for a Cell Line (Type 4/4) CELL TEST ELECTRICAL DATA Desired Data Mandatory Data Electrical Data - Sample List Sorting Data - Sample LOT ID ID Test Date/Time Tool ID FF Rs Iap Isc Rsh Voc Pmpp Irev1 Irev2 Eff Blue Temp Class Bin 12/16-A /16/2011 1:57 AM CT-A A 6 12/16-A /16/2011 1:58 AM CT-A A 4 12/16-A /16/2011 1:58 AM CT-A A 7 12/16-B /16/ :00 AM CT-A A 4 12/16-B /16/2011 2:00 AM CT-A A 4 12/16-A /16/2011 1:58 AM CT-A A 5 12/16-A /16/2011 1:58 AM CT-A A 4 12/16-A /16/2011 1:59 AM CT-A A 5 12/16-A /16/2011 2:00 AM CT-A A 5 12/16-A /16/2011 2:01 AM CT-A A 4 12/16-A /16/2011 2:01 AM CT-A A 4 12/16-A /16/2011 2:01 AM CT-A A 5 12/16-A /16/2011 2:02 AM CT-A A 5 12/16-A /16/2011 2:02 AM CT-A A 8 12/16-A /16/2011 2:02 AM CT-A A 5 12/16-A /16/2011 2:02 AM CT-A A 5 12/16-B /16/2011 2:03 AM CT-A A 5 12/16-B /16/2011 2:03 AM CT-A A 4 12/16-B /16/2011 2:03 AM CT-B A 2 12/16-B /16/2011 2:04 AM CT-B A 2 12/16-B Absolutely the 12/16/2011 most important 2:04 AM Data-Collection CT-B Source A 3 12/16-B /16/2011 2:04 AM CT-B A 2 12/16-A /16/2011 2:05 AM CT-B A 1 Most manufacturers use Lot-ID to compare Suppliers, Tools, and Lines; although many still use 12/16-A /16/2011 2:05 AM CT-B A 1 12/16-A time-based Trend 12/16/2011 Analysis 2:05 AM CT-B A 1 Data is automatically-collected directly from the Tool. This is only a small portion of available data 9

10 Fault Detection & Classification Real-Time Tool Monitoring

11 Fault Detection & Classification Real Time Tool Monitoring Buy Poly Poly Charge Ingot Casting Ingot Cut Brick Cut Slice Clean Inspect Sort Package Multi Mono Cell Inspect Wet Texture P/N Junction Wet Clean Edge Isolation ARC SIN Metal Line Firing Furnace Cell Sort Cell Package MOCVD Spray Direct Remote Wet Laser Best fit for high-impact tools that have lots of variables line: DSS Furnace; Slicer Cell line, POCL Diffusion, SIN Deposition 11

12 Fault Detection & Classification Installation does not have to be scary Easy to connect to individual tools Easy to train IT and end-users to manage and use software Easy to centralize and roll-out factory wide, allowing for very precise tool-to-tool performance comparison. Direct Tool Connections Data centralization Real Time Alarming and Problem Notification FDC Software DSS Cut TEX DF PSG SIN PRINT Firing TEST 12

13 Fault Detection & Classification How it Works Trace Process Equip. SVID#1 MFC1 SVID#2 MFC2 SVID#3 TC1 SVID#4 TC2 SVID#5 LP1 SVID#6 LP2 SVID#7 HP1 #1 #2 #3 #4 #5... Steps Monitors and overlays tool-trace data (typically gas flow, pressure/vacuum, temperature) from individual tool Runs (Run-1 is shown) Run-2 overlays Run-3 overlays Run-4 overlays Capacity to analyze: o Traces o Steps o Runs 13

14 Fault Detection & Classification What it Makes Possible Trace Process Equip. SVID#1 MFC1 SVID#2 MFC2 SVID#3 TC1 SVID#4 TC2 SVID#5 LP1 SVID#6 LP2 SVID#7 HP1 Having FDC enables an engineer to: o Monitor critical process areas o Correlate metrology data to tool performance why is the data changing? o Optimize PM scheduling to increase tool uptime o Optimize specification limits o Optimize process recipes o Increase overall yields! #1 #2 #3 #4 #5... Steps 14

15 Factory-Wide Data Analysis Integrated Process Control Software

16 Factory-Wide Process Control Software Integrate and analyze all data sources Ingot Growth Fab Cell Fab Module Fab Metrology Data Attribute Data Tracking Data FDC data Test data Automatic root-cause analysis across every manufacturing workshop Process Control Software 16

17 Factory-Wide Process Control Software Start with Cell-Test data only What do you gain? Ingot Growth Fab 1. Data analysis for all available electrical data Efficiency, Isc, Voc, FF, Rsh, Rs, low-power rate, etc. 2. Automatic production and quality-control reports Overall efficiency, throughput, uptime, and yield, bin-classification 3. Automatic engineering analysis reports by efficiency, current, low-power rate, defect-type, supplier, material, tool, etc. Cell Fab Module Fab Cell Test 17

18 Factory-Wide Process Control Software add Metrology data What do you gain? Ingot Growth Fab Cell Fab Module Fab Inspect Tex Etch P/N Junction PSG Etch Edge Isolation SIN ARC Metal Line Firing Furnace Cell Test Cell Sort 18

19 Factory-Wide Process Control Software add Tool data Ingot Growth Fab Cell Fab Module Fab What do you gain? 1. WIP Visibility 2. LOT Tracking 3. Yield Reports 4. Tool-level SPC & FDC Analysis 5. Split-Lot Analysis 6. Operator Reports 7. Production Path Analysis Inspect Tex Etch P/N Junction PSG Etch Edge Isolation SIN ARC Metal Line Firing Furnace Cell Test 19

20 Factory-Wide Process Control Software add -ID PV ID s recommended per wafer, 4-shown 1.4mm x 1.4mm each, 5-20μm deep 130ms etch time per ID 20

21 Factory-Wide Process Control Software -ID Lasering and Tracking -scribing locations Choose 1 of 3 -reading locations The more, the better 21

22 Factory-Wide Process Control Software add -ID What do you gain? Absolute correlation of wafer properties to conversion efficiency and yield o All metrology measurements thickness, lifetime, bulk resistivity, SIN, grid lines, bow, etc. o Supplier performance (silicon/wafer) o Tool performance / FDC o Advanced recipe management; correlated directly to wafer properties Complete traceability o From wafer fab through module fab and into the field 22

23 Factory-Wide Process Control Software Summary of Benefits

24 Factory-Wide Process Control Software Summary of Benefits Factory Automation o Make data useable (attribute, WIP, metrology, test) Fault Detection & Classification (FDC) o Make metrology data more powerful (correlate to tool-trace data o Optimize tool performance (preventive maintenance scheduling, specification limits, recipes, mechanical and electrical yields) Factory-Wide Data Analysis o Correlate all data (wafer-cell-module) o Automatic Analysis: trends, tools, suppliers, lines, workshops, shifts, operators, recipes, metrology-measurements, etc. o -ID: absolute wafer-tracking and wafer-property correlation analysis 24

25 Near-Term and Long Term Challenges for PCS in PV manufacturing

26 Near-Term Challenges (1-3 Years) Challenge 1 Factory Automation o Lack of factory experience and know-how o IT-teams are often unsure of themselves Challenge 2 Corporate Strategy of Manufacturers o Economy-of-Scale has been largely successful o Operational excellence requires different thinking Challenge 3 Powerful Industry Suppliers o Bundle low-value added SPC & MES software o Displace money that could have gone to PCS o Make buyers pessimistic on software 26

27 Long-Term Challenges (4+ Years) Challenge 1 PCS Value o Quantifiable efficiency gains must be realized Challenge 2 PCS Adoption Rate o Domino affect anticipated Challenge 3 -ID Adoption Rate o Quantifiable efficiency gains must be realized o Domino affect anticipated 27

28 Thank You Questions?