This document summarizes a presentation given at the 3rd Asia Academic Forum in Kuala Lumpur from November 13-15, 2006. The presentation introduces process control systems (PCS) and their importance to Intel's assembly and test operations. It discusses challenges faced by PCS including poor signal detection, data system limitations, and human errors. It outlines Intel's PCS strategy to improve detection and prevention, deliver automation solutions, and build a high performance PCS culture. The presentation shows how PCS has evolved and will continue to advance with new technologies. It proposes areas of joint opportunity between Intel and universities.

1. Advancing Process Control
Systems in Assembly/Test
Nik Zurin

2. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
2
• Intel Assembly/Test
• PCS 101
• Why PCS Is Important to Intel Assembly/Test
• Advancement in PCS Technology
• Shape of Things to Come
• Areas of Opportunity
Presentation Objectives:
• Introduce concept of PCS and it’s importance to Intel
Agenda

3. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
3
Arizona
Chandler ATTD
Malaysia
Penang ATTD
Assembly/Test Technology Development (TD)
Assembly/Test High Volume Manufacturing (HVM)
• 5 high volume assembly/test manufacturing sites
• 2 assembly/test technology development sites
Intel’s Global Assembly/Test Network

4. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
4
FabFab Assembly & TestAssembly & TestSortSort
Raw Silicon to
Wafer
Wafer singulation to Die, Die
Assembly to Package and
Structural and functional Testing
Electrical testing
to check silicon
integrity
Intel Assembly/Test (Flip Chip Technology)
Die PrepDie Prep
ChipChip
AttachAttach
EpoxyEpoxy
UnderfillUnderfill
IHSIHS
AttachAttach
BurnBurn--inin TestTest FinishFinish
Assembly Test

5. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
5
What is PCS?
• Process Control Systems
• Any factory system used to monitor/control overall variation in the
manufacturing process
• Provides an on-line, real time system for identifying and responding to
process and equipment variation
PCSPCS = SPC + RFC= SPC + RFC
Response Flow ChecklistResponse Flow Checklist
Documented, standardized, andDocumented, standardized, and
systematic methods forsystematic methods for
responding to unplanned events.responding to unplanned events.
Statistical Process ControlStatistical Process Control
The use of statistical techniquesThe use of statistical techniques
to analyze a process, taketo analyze a process, take
appropriate actions to achieve &appropriate actions to achieve &
maintain a stable process, &maintain a stable process, &
improve process capabilityimprove process capability

6. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
6
Why is PCS important to Manufacturing?
We need it to continuously improve:
• Customer satisfaction
• Scrap rate (Yields)
• Screening man-hours (Productivity)
• Loss of equipment availability (Productivity)
• Product and process Quality
Customers ask for it
• “Don’t just tell us that your process/product is improving, prove it by
showing data”
• Customer audit
It is an ISO Requirement
• Process Control
• The Use of Statistical Techniques

7. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
7
PCS Goal in Intel Assembly/Test
PCS is a key component of Intel’s Quality System. It provides methods,
tools and systems that ensure that the factory is able to:
• Run its process on target with minimum variation
• Rapidly detect problems
• Respond to shifts, trends, off-target
Keeps our processes stable, alerts us whenKeeps our processes stable, alerts us when
something has gone wrong, and helps ussomething has gone wrong, and helps us
identify opportunities to reduce variationidentify opportunities to reduce variation

8. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
8 Courtesy: Dev Pillai
Poor signal detection at root cause operation
• Lack of proper definition of PCS parameters resulting in quality escapees and late
detection
• Limited metrology definition early on during path finding & development
• Too many parameters to monitor causing signal with dilution
Limitations in data systems
• Cumbersome to extract and analyze data from different, unconnected databases
• Limitations in getting at a finer granularity beyond lot level
• Dependency on human to enter data by hand – source of data inaccuracies
Human induced variations
• Setup errors, wrong recipe, mis-processing, mixing, Spec non compliance
• Not paying attention to triggers, RFC
• Limited discipline to monitor data & indicators to take actions to improve process
maturity
Our A/T PCS Challenges…

9. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
9
And Escalating A/T Manufacturing Complexity…
Increased A/T complexity with every technology generation
•Complex Si – Pkg integration and New Mfg processes with steep vertical ramps
•Technology envelope extension drives reduction in historical quality & reliability margins
•Increasingly important : Faster info turns, reliable equipment with advanced PCS built in
Technology GenerationTechnology Generation
C4 / OLGAC4 / OLGA
MktMkt
SegmentationSegmentation
300mm300mm
Heat SpreaderHeat Spreader
Structural testStructural test
FCPGAFCPGA
LowLow--K ILDK ILD
CDOCDO
SiSi--PkgPkg
InteractionInteraction
FCLGAFCLGA
PbPb--freefree
BSMBSM
Hi/lowHi/low pwrpwr
Mfg flowMfg flow
OptimizationOptimization
Combined TESTCombined TEST
Dual coreDual core
Mfg ProductivityMfg Productivity
180nm 130nm 90nm 65nm 45nm
IncreaseinComplexityIncreaseinComplexity
Increase in
Increase in SiSi--PkgPkg Interaction
Interaction
More die/wafer
More data/die,
Operational complexity
More die/pkg
interaction
Process & Binning
complexity
OLGA
Technology GenerationTechnology Generation
C4 / OLGAC4 / OLGA
MktMkt
SegmentationSegmentation
300mm300mm
Heat SpreaderHeat Spreader
Structural testStructural test
FCPGAFCPGA
LowLow--K ILDK ILD
CDOCDO
SiSi--PkgPkg
InteractionInteraction
FCLGAFCLGA
PbPb--freefree
BSMBSM
Hi/lowHi/low pwrpwr
Mfg flowMfg flow
OptimizationOptimization
Combined TESTCombined TEST
Dual coreDual core
Multi Core
Mfg Productivity
90nm 65nm 45nm
IncreaseinComplexityIncreaseinComplexity
Increase in
Increase in SiSi--PkgPkg Interaction
Interaction
More die/wafer
More data/die,
Operational complexity
More die/pkg
interaction
Process & Binning
complexity
OLGA

10. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
10
… Are Driving our A/T PCS Strategy
Courtesy: Dev Pillai
Focus Area Goal
Improve Methods for
Detection,
Containment &
Prevention
ModuleModule
EngineeringEngineering
PCS BKMs/Methods
PeoplePeople
Mfg Floor PCS
Eng Data Reviews
Data SystemsData Systems
NGSC/SPC++
Eng Analysis Tools
Deliver Automation
Solutions to improve
data availability,
quality &
accessibility
Build & Sustain
High Performance
PCS Culture in HVM
Business Value to A/T
Offline PCS:
Proactive data driven actions
for excursion prevention,
reduction in process variation
& continuous improvement
Online PCS:
Real time detection,
containment & dispositioning of
process issues at root cause
Develop and execute key initiatives along the 3 focus areas to
meet demands and challenges of future A/T technologies

11. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
11
Define key and controlDefine key and control
process parametersprocess parameters
Develop & qualifyDevelop & qualify
metrology andmetrology and
measurement methodsmeasurement methods
Validate process specValidate process spec
limitslimits
Establish control chartsEstablish control charts
with valid control &with valid control &
disposition limitsdisposition limits
Develop best knownDevelop best known
methods for processmethods for process
targetingtargeting
Technology
Definition Phase
Technology
Development Phase
Establish online PCS for
manufacturing
Collect data to assess
process stability,
capability & matching
Revise RFCs for Out of
Control events
Revise or tighten control
limits
Conduct beta testing of
of PCS automation
systems
Technology
Deployment Phase
Transfer technology to
HVM – demonstrate
process matching
Data Reviews !
Proliferate all PCS tools
& systems across
multiple HVM factories
Feedback key learnings & knowledge
to future technologies (N+1, N+2…)
Module Engineering : Developing
Fundamentals of PCS

12. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
12
Engineer /
Operator
Engineer /
Q&R
Decision Support System
MARS
Reporting
Tools
Online DB
Station
Controller
Station
Controller
Station
Controller
UI UI UI
Shop Floor Control System
WorkstreamSPCCMS
SPC++ System
Adapter
1. Data Acquisition
125
150
175
200
225
250
value2
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Sample
Avg=185.42
LCL=137.08
UCL=233.76
Online Charts
3. Trigger
125
150
175
200
225
250value2
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Sample
Avg=185.42
LCL=137.08
UCL=233.76
Offline Charts
5. Offline Analysis
2. Analysis (rule based)
4. RFC
Integrated Data Systems : Automating PCS

13. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
13
People Systems : Establishing and Sustaining
a Culture of PCS in HVM
HighHigh
PerformancePerformance
PCS CulturePCS Culture
SetSet
ExpectationsExpectations
Tools &Tools &
TrainingTraining
FocalFocal
PositivePositive
ReinforcementReinforcement
TD
HVM
HVM
HVM
Engineering
• Set up and sustain PCS in HVM
• Regularly collect, analyze and present PCS
data to assess process health
• Make data driven decisions to fix process
issues, improve yield & drive reduction of
process variation
Manufacturing
• Strictly adhere to PCS procedures on the
production floor
• Accurately respond and provide clear
annotation to Out-of-control events
• Include PCS & quality as part of regular
floor communication
Management
• Set and communicate clear goals and
expectations
• Recognize and reward behaviors that uphold
quality over output
• Participate in PCS discussions and reviews

14. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
14
Evolution of PCS in Assembly/Test
Phase 1Phase 1
PCSPCS
AwarenessAwareness
Phase 2Phase 2
EnhancedEnhanced
CapabilitiesCapabilities
Phase 3Phase 3
AdvancedAdvanced
PCSPCS
90nm90nm
20032003--20042004
65nm65nm
20052005--20062006
45nm45nm
20072007--20082008
32nm32nm
20092009--20102010
Manual PCSManual PCS
Manual data analysisManual data analysis
Offline PCS analysis,Offline PCS analysis,
data reviewsdata reviews
PCS Cultural ShiftPCS Cultural Shift
50% Station Controller,50% Station Controller,
SPC, Recipe ControlSPC, Recipe Control
40% inline monitor40% inline monitor
Automated dataAutomated data
analysis, Commonalityanalysis, Commonality
tools, PCS data reviewstools, PCS data reviews
PCS People SystemsPCS People Systems
Mfg integrationMfg integration
70% Station Controller,70% Station Controller,
SPC, Recipe ControlSPC, Recipe Control
Auto RFCAuto RFC
60% inline monitor60% inline monitor
Automated suite of dataAutomated suite of data
analysis tools integratedanalysis tools integrated
into a well defined HVMinto a well defined HVM
business processbusiness process
PCS People SystemsPCS People Systems
Mfg integrationMfg integration
Sustained CultureSustained Culture
>90% inline monitor>90% inline monitor
insituinsitu metrologymetrology
Unit level PCSUnit level PCS
Fault DetectionFault Detection
50%
60%
80%
>90%
Excursion detection
& containment at root
cause
Recipe managementRecipe management
>90% Station Controller,>90% Station Controller,
SPC, Recipe ControlSPC, Recipe Control
Auto RFCAuto RFC
Our Complex Factories and rapidly evolving Demand for PCS
drive a need for Innovation and Trained Engineers…

15. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
15
Areas of Joint Opportunity
Conferences and
Industry Standards
Research
Collaboration for
software, methods,
algorithms
Qualified
students (after
graduation)
for hire
University
Curriculum
Development
• Adjust curriculum to evolving
semiconductor needs
• Areas of Interest:
Chemical Engineering
Materials Science
Mechanical Engineering
Computer Science and
Software Engineering
Statistics and Operations
Research
• Provide new and continued
opportunities to share
knowledge via conferences
• Develop pre-competitive
standards that benefit the
entire industry
• Feedback & feed forward control systems
• Factory & Equipment control systems
• Data mining
• Multivariate statistical analysis
• Predictive yield analytics
• Graduates with:
Strong applied problem
solving skills
Semiconductor or
Manufacturing knowledge
Ability to apply theory with
practical knowledge to
solve real world problems

16. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
16
Acknowledgements and Contacts
• Sridhar Rajappa, ATTD Automation, Intel Chandler
• Walter Flom, ATTD Statistics/Methods, Intel Chandler
• Wee Tatt Cheong, ATTD Statistics, Intel Penang
• Jeff Pettinato, CAS, Intel Chandler

17. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
17
Questions?

18. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
18
Backup

19. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
19
Engineering Analysis and Sustaining
Data extraction
Data sorting &
verification
Data
Analysis
Reports &
Reviews
0 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Engineering
Sustaining
Data
extraction
Data Analysis Reports & Reviews
Engineering Sustaining and Continuous
Improvement
Data ACCESSABILITY
Discovery/EPASS enhancement
Data availability-ARIES/MARS/OASYS
New PCS BKMs
Stability, capability & matching
PCS health metrics
Sustaining BKMs
PCS health-targeting, matching
Data driven CIPs-yield, MOR
Goal: Improve efficiency in engineering data analysisGoal: Improve efficiency in engineering data analysis
and sustaining (>90% of total engineerand sustaining (>90% of total engineer’’s time)s time)
In Future:
Prep time in
minutes
Today:
Prep time in
days/hours

20. 3rd Asia Academic Forum, Kuala Lumpur
Nov. 13-15, 2006
20
Manufacturing PCS Floor Reports
OOC & Compliance
Reports
WorstreamWorstream
OASYS DBOASYS DBSPC++SPC++
Station
Controllers
In-line DB Off-line DB Application
MARS DBMARS DB
Flat File
Download
Option
Features
Shopfloor
Interface
LIST OF REPORTS / GRAPHS Prototyped in TD
1. OOC/RFC compliance Report-
2. Manual reports from SPC++
3. SPC++ Charts
LIST OF REPORTS / GRAPHS Requested
1. Exceptions Report - Web based - for
supervisors
2. OOC reports
3. SPC++ charts
ARCHITECTURE
DECISION
(early to mid Q4)
Websites
Control Rooms
Crystal Reports
(existing Operations portal)
Portals
Sharepoint
PHASE 1PHASE 1
PHASE 2/CIP More inline and offline databases could be considered but
Shopfloor Interface will remain the same.
Offline viewing of
SPC++ Charts
Exception Report
OOC & Compliance
Reports
WorstreamWorstream
OASYS DBOASYS DBSPC++SPC++
Station
Controllers
In-line DB Off-line DB Application
MARS DBMARS DB
Flat File
Download
Option
Features
Shopfloor
Interface
LIST OF REPORTS / GRAPHS Prototyped in TD
1. OOC/RFC compliance Report-
2. Manual reports from SPC++
3. SPC++ Charts
LIST OF REPORTS / GRAPHS Requested
1. Exceptions Report - Web based - for
supervisors
2. OOC reports
3. SPC++ charts
ARCHITECTURE
DECISION
(early to mid Q4)
Websites
Control Rooms
Crystal Reports
(existing Operations portal)
Portals
Sharepoint
PHASE 1PHASE 1
PHASE 2/CIP More inline and offline databases could be considered but
Shopfloor Interface will remain the same.
Offline viewing of
SPC++ Charts
Exception Report