in-situ production process detection and chamber matching … · • process deviations identified...
TRANSCRIPT
InIn--SituSitu Production Production Process Detection and Process Detection and
Chamber Matching Chamber Matching Using Temperature Using Temperature
Metrology Sensors and Metrology Sensors and DiagnosticsDiagnostics
NCCAVSNCCAVSPlasma Etch Users GroupPlasma Etch Users Group
February 12, 2004February 12, 2004
Paul MacDonaldPaul MacDonaldOnWafer TechnologiesOnWafer Technologies
PEUG February 12, 2004PEUG February 12, 2004
Abstract
• Concrete Methodologies and Technology Enabling In-SituDetection of Process Excursions– Wafer-Based Metrology System Overview
• Multiple Production examples that illustrate in-line process excursion detection and chamber matching. – Process monitoring
• Process deviations identified before any device wafers were scrapped• Thermal signatures shown to be early indicators of process instability
and drift – Chamber Matching
• Chambers at sites on different continents are easily matched to ensure process transferability
• Thermal signatures shown to quickly and easily match etch reactors
PEUG February 12, 2004PEUG February 12, 2004
Monitoring Wafer Temperature in a Plasma System to Diagnose Chamber Problems
• Etching is Impacted by Various Interacting Mechanisms
– Direct Chemical Reaction, Reactive Etching, Deposition, Mask Erosion
• Same Etch Mechanisms are Extremely Sensitive to Temperature
– Temperature is a first order indicator of etching progress or “Health”
• Wafer-Based Metrology of Spatially Resolved Temperature Profiles for all Operating Conditions
• Utilization of Wafer-Based Metrology is Vital for Realizing Real-Time Fault Detection and Isolation
Typical Plasma Etch EnvironmentTypical Plasma Etch Environment
CoolingCoolingFluidFluid
Plasma
HeliumHelium
PEUG February 12, 2004PEUG February 12, 2004
– Data Analysis System
– Data Transfer Unit
– PlasmaTemp® SensorWafer®
• System Components
PlasmaTemp® In Situ Metrology Hardware OverviewHardware Overview
- Base material is SEMI standard wafer with thermal oxide layer
- Sensor network isolated using a high purity polyimide coating- “System on a Wafer” electronics module for high speed, autonomous
data acquisition & wireless data transfer
- Thermistors located for high resolution thermal profiling
PEUG February 12, 2004PEUG February 12, 2004
Typical Usage In High Volume Production In High Volume Production FabsFabs and R&D Environmentsand R&D Environments
• PlasmaTemp SenorWafers are Used like any Other Production Monitor– Data Collection Sequence
• Load PlasmaTemp® Wafer in Standard Cassette / Place FOUP in Load Port
• Launch Tool’s Automatic Transfer Sequence
• Run the Standard Production Plasma Etch Recipe– Standard Power, Pressure,
Temperature, etc.• Remove Cassette / FOUP• Download Data
– Immediately Bring Tool Back to Production• No Dummy Wafer• No Particle Test
OnWafer built all the required services for the SensorWafer metrology system
right into a SEMI standard 300mm FOUP. Product is shown with its
convenient “docking station”.
PEUG February 12, 2004PEUG February 12, 2004
BT Main-Etch
In Situ MetrologySample Mission DataSample Mission Data
MERIE Chamber- 1500W with 30 sec O2 ash- 0oC chuck temperature
Ash
1 Line, 1 Sensor
Typical 3-D View
PEUG February 12, 2004PEUG February 12, 2004
P lasm aT em p Process Reg im e
0
50
100
150
200
250
0 2000 4000 6000 8000
Total P ow er (W )
Pres
sure
(mT)
200mm 300mm
PlasmaTemp Process RegimeSample Installed Base DataSample Installed Base Data
High Power Si EtchStrip
Silicon Etch& Metal Etch Oxide Etch
PlasmaTemp PlasmaTemp SensorWaferSensorWafer Operates in Operates in All Plasma Process RegimesAll Plasma Process Regimes
PEUG February 12, 2004PEUG February 12, 2004
0
50
100
150
200
250
Num
ber
of W
afer
s
0 - 0.050.05 - 0.10.1 - 0.150.15 - 0.20.2 - 0.250.25 - 0.30.3 - 0.35
3 Sigma Spread (C)
PlasmaTemp Isothermal 3 Sigma Spread 361 Total Wafers
PlasmaTemp PrecisionFactory CalibrationFactory Calibration
PlasmaTemp Isothermal Absolute Accuracy361 Total Wafers
0
50
100
150
200
250
0 - 0.05 0.05 - 0 .1 0.1 - 0 .15 0.15 - 0 .2 0.2 - 0 .25 0.25 - 0 .3 0.3 - 0 .35
Accuracy (C)
Num
ber o
f Waf
ers • PlasmaTemp Calibration Completed in Stable Thermal Environment– NIST-Traceable Calibration
Standard • Why Calibrate in Isothermal
Environment?– Small Variations in Plasma
Processing are Detected by Sensitive, Accurate SensorWafers
• Absolute Accuracy and 3σ Spread Better than 0.3°C– Data Sample of Last 361 Wafers
Built as of 1/1/04
PlasmaTemp SensorWafersDemonstrate Superior
Calibration Repeatability
Absolute Accuracy to Absolute Accuracy to NISTNIST--Traceable Traceable
Standard Better Than Standard Better Than 0.30.3°°CC
SensorSensor--toto--Sensor Sensor 33σσ SpreadSpread Better Better
Than 0.3Than 0.3°°CC
PEUG February 12, 2004PEUG February 12, 2004
W a fe r - to -W a fe r R e p e a ta b ilityin P la s m a P r o c e s s e s
8 5
9 0
9 5
1 0 0
1 0 5
1 1 0
1 1 5
Recipe ARecipe BRecipe CRecipe DRecipe ERecipe FRecipe GRecipe HRecipe I
R e c ip e C o n d it io n
Tem
pera
ture
(C)
W a f e r 1
W a f e r 2
W a f e r 3
PlasmaTemp PrecisionWaferWafer--toto--Wafer RepeatabilityWafer Repeatability
• Wafer-to-Wafer Repeatability– Wafer-to-Wafer Repeatability is
Limited by Chamber Repeatability – Typically 2-3%
• Plasma Reactor Variation of Only 1% Equates to as Much as 1.09°C Thermal Variation
– Wafer-to-Wafer Repeatability Better than 1.30°C
• Within 0.21°C of Plasma Reactor Stability Threshold
– Each Data Sample Includes Multiple Runs under Same Conditions
• Different SensorWafers• Same Recipe, Same Chamber
1.291.29°°CC
0.710.71°°CC0.250.25°°CC
0.610.61°°CC
0.060.06°°CC
0.040.04°°CC
0.400.40°°CC 0.830.83°°CC
1.071.07°°CC
PEUG February 12, 2004PEUG February 12, 2004
Plasma Reactor Optimization and Hardware TroubleshootingTypical ApplicationsTypical Applications
25Torr He BSC
6Torr He BSC
He Cooling Influence
Faulty Source Match
Chucking Voltage Selection
Positive V
Negative V
Process Kit Selection
SiO2 Process Kit
Si Process Kit
Backside Particle DetectionMis-Adjusted Lift Pins
Har
dwar
eTr
oubl
esho
otin
gR
eact
orO
ptim
izat
ion
PEUG February 12, 2004PEUG February 12, 2004
Data Analysis & CorrelationExample DataExample Data
• Thermal profiles and CD bias are closely related• Data indicates issue with the CD at the Notch (circled)
Spatially-Resolved CD Data can be Easily Analyzed for Correlation to Thermal Profiles
CD Bias and Thermal Profile
40RFH
59RFH
Temperature (°C) CD Bias (nm)38 pts.
PEUG February 12, 2004PEUG February 12, 2004
• Regular monitoring after known good wafer lots defines baseline tool performance (“fingerprinting”)
• Detect tool problems– Apply Diagnostics Package
• Determine appropriate time for PM and/or WC• Help bring tool back online after PM and/or WC
xx x
xxx x x
x
baseline
x x
Prob
lem
x
x x x x xxx
WC / PMT
time
Routine Temperature MonitoringInIn--Line Fault DetectionLine Fault Detection
x
Chamber Ready
3σ
Tavg
PEUG February 12, 2004PEUG February 12, 2004
Spatial Modeling Methodology Improving SignalImproving Signal--toto--Noise RatioNoise Ratio
• Profile A and Profile B Show the Same Mean and Range but VERYDifferent Thermal Profiles
• OnWafer has Developed a Modeling Approach to Determine Spatial Chamber Health Signal and Minimize Normal Noise Due to Chamber Variations
• Metric Must be Able to Capture Change in the “Shape” of the Thermal Profile while Filtering Standard Chamber Fluctuations
• Shape-Fitting Metric is Denoted in the Next Slide by “O-Factor”
22--D Temperature Profile BD Temperature Profile B
Mean T: 85°C
Ran
ge 1
0°C
22--D Temperature Profile AD Temperature Profile A
Ran
ge 1
0°C Mean T: 85°C
PEUG February 12, 2004PEUG February 12, 2004
T h e r m a l a n d F a b D a ta C o r r e la tio n
0 .0 0
0 .2 0
0 .4 0
0 .6 0
0 .8 0
1 .0 0
1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103
109
115
121
127
133
139
145
151
157
D a y
O-F
acto
r
IS S U E O -F ac to r
• Thermal Data and Fab Data Collected over 140 days • Temperature is Used as a Proxy for Chamber Health • In-Situ Thermal Measurements Identified 2 out of 3 Fab
Issues Prior to Any Other In-Line Measurements– Increasing Data Collection Frequency Improves O-Factor Resolution
Process Deviation IdentificationExample 1Example 1
Issue
Issue
Missed
NormalNormal
Normal
Normal
Normal
Normal
Normal Normal
PEUG February 12, 2004PEUG February 12, 2004
Process DriftWet Clean CycleWet Clean Cycle
• Temperature Mean and CD Bias Mean Show Steady Increase over Wet Clean Cycle
• Wet Clean Trigger Point Determined to 90RFh– CD Within Spec @ <.85nm (normalized)– Thermal Range within Spec @ <0.05°C (normalized)
Defining Wet Clean TriggerMERIE Reactor
0.8
0.9
1
33RFh 55RFh 67RFh 48RFh 100RFh
Norm
aliz
ed T
empe
ratu
re (
o C)
0
0.6
1.2
Norm
aliz
ed C
D (n
m)
Mean Temperature CD BIAS
PEUG February 12, 2004PEUG February 12, 2004
Chamber Matching
0.82
0.84
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
1.02
1A 1C 2B 2A 1B 1D
Chamber
Nor
mal
ized
Mea
n Te
mpe
ratu
re (o C
)
0
0.2
0.4
0.6
0.8
1
1.2
Nor
mal
ized
Tem
pera
ture
Ran
ge
(o C)
Mean Temperature Temperature Range
Chamber Matching6 Chamber Population 6 Chamber Population –– Single SiteSingle Site
• 6 Chambers Examined– One Chamber Clearly
Out of Distribution• Custom Process Kit
– Two Pairs of Chambers Well Matched
– Last Chamber at Distribution Median
• Hardware Changes Implemented to better Match Chambers– Chiller Adjustment
Resolved Tool 1 and Tool 2 Thermal Disparity
ChambersMatched
Cha
mbe
r Dev
iatio
n
ChambersMatched
PEUG February 12, 2004PEUG February 12, 2004
Chamber Matching Results
70 75 80 85 90 95 100 105
1A
2C
4A
3B
3C
3D
3A
Cham
ber
Temperature (C)
Step 1 Step 2 Step 3 Step 4
• 7 Chambers, Multiple Steps Examined
• Tool 3 Shows Excellent Mean Matching– Within 1.29°C for all Steps
• Chambers 4A and 2C Did Not Respond Normally to Process Steps
• Chamber 1A was Running 2°C Warmer than Tool 3
• All Deviations Required Hardware Repair
Chamber Matching7 Chamber Population 7 Chamber Population –– US and Europe SitesUS and Europe Sites
WellWell--MatchedMatchedChambersChambers
PEUG February 12, 2004PEUG February 12, 2004
HeatFlux™Advanced Analysis ToolAdvanced Analysis Tool
Plasma Application Module• Analyzes the data from a
single PlasmaTemp run• Determines:
– Cooling efficiency across the area of the chuck
– Heating uniformity of the plasma• This enables:
– Differential diagnosis of chuck or plasma-related heating problems in production
– Analysis of chamber/chuck geometries during equipment design
– Identify the Root Cause of Reactor Issues
HeatFlux
Temporal ProfileMain Etch Step from a 200mm
Poly Etch System
Chuck’s Conductance Profile
PEUG February 12, 2004PEUG February 12, 2004
Chamber MatchingActing on Thermal DataActing on Thermal Data
• Week 1
• Week 2 - Deviant
• Week 3 - Deviant
• Week 4 - Deviant
• Week 5 - Resolved
Conductance AnalysisConductance Analysis• Thermal Profile Shift in Thermal Profile In Chamber • Action is Required to Resolve Discrepancy• Correlated to Edge-Specific Yield Issue•HeatFlux™ is Applied to Separate Plasma and Chuck Effects on the Wafer
• Conductance Analysis Clearly Show Chuck Thermal Transfer Efficiency Change in Weeks 2 – 4• Process Kit Change Resolves Issue
PEUG February 12, 2004PEUG February 12, 2004
Summary
• Concrete methodologies for in-situ detection of process excursions and rapid chamber matching demonstrated
• Multiple examples that illustrate in-line process excursion detection and chamber matching. – Process monitoring
• Process deviations are identified in-line • Temporal/spatial temperature signatures show early indicators of
process instability and drift – Chamber Matching
• Chambers at one site, and chambers at sites on different continents are matched
• Temporal/spatial temperature signatures quickly and easily matchetch reactors
• Advanced modeling allows component-level troubleshooting