Andreas Steinbach Sven Barth Memory Development Center - Infineon Technologies Dresden 4th European AEC/APC Conference 2003 Grenoble, France Monitoring of process stability and chamber matching by plasma parameter measurement using High speed- SEERS Andreas Steinbach Sven Barth Memory Development Center Unit Process Development Coordination Infineon Technologies Dresden [email protected]4 th European AEC/APC 2003
24
Embed
Monitoring of process stability and chamber matching by ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 1
Monitoring of process stability and chamber matching
by plasma parameter measurementusing High speed- SEERS
Andreas SteinbachSven Barth
Memory Development CenterUnit Process Development Coordination
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 2
Acknowledgement
We greatly appreciate our colleagues support:Lars Christoph Lincoln O´RiainAndreas Thamm Jörg RadeckerGary Skinner Mark- Stefan Scheftner
Acknow-ledgement
The results presented in this paper were supported:by the EFRE fund of the European Communityand by funding of the State Saxonyof the Federal Republic of Germany, project 7700 / 1220.
The authors are responsible for the content of the paper.
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 3
Outline
High speed SEERS featuresStrategy of plasma parameter application within AEC/APCHigh speed SEERS application examples
Process stabilityChamber matchingFDC - Fault detection and classification
Summary
Outline
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 4
High speed SEERS features
High speedSEERS
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 5
Self Excited Electron Plasma Resonance Spectroscopy -How it does work
RF up to 30th harmonicsPassiv, no impact on process
Easy current sensorqualification
SEERS model limits:One RF frequency higherthan ion plasma frequency(4 … 6 MHz)Pressure max. 300mTorr
Electron Collision Rate [s-1] Mean electron collisions with gas moleculesElectron Density [cm-3] = Plasma Density
Bulk Power [mW/cm²] Power dissipated by electrons in plasma
DC Bias Voltage [V]it`s well known
MeasurementRF Current
RF Peak Voltage
CalculationFFT and
Model SEERS
Resultsin real time
Plasma Parameters
High speedSEERS
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 6
Plasma parameter vs. Process time
1,E+08
3,E+08
5,E+08
7,E+08
9,E+08
1,E+09
200 205 210 215 220time [s]
el. C
oll.
rate
[s-1
]
2,2E+09
2,4E+09
2,6E+09
2,8E+09
3,0E+09
3,2E+09
3,4E+09
3,6E+09
el. d
ensi
ty [c
m-³
]
High speedSEERS
B- field sine wave mode 0,5 Hz:
B- field square wave mode 0,25 Hz:
Dynamic effects in plasma: B- field rotation mode effecton process conditions - Results
Electron DensityElectron Collision Rate
Plasma parameter vs. Process tim e
1,E+08
2,E+08
3,E+08
4,E+08
5,E+08
6,E+08
200 205 210 215 220time [s]
el..
coll.
rate
[s-1
]
2,E+08
4,E+08
6,E+08
8,E+08
1,E+09
el. d
ensi
ty [c
m-³]
Electron DensityElectron Collision Rate
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 7
Dynamic effects in plasma: B- field rotation mode effecton process conditions - Discussion
High speed SEERS modus: up to 15 measurements per secondElectron collision rate and electron density show B- field mode (square or sinus) impact on process conditions.
Oscillation indicate electrically asymmetric chamber geometryElectron density (plasma density) with sine wave B- field mode recipe indicates plasma instabilitiesPhase shift between electron density (= plasma density) and electron collision rate (see B-field square wave mode 0,25Hz)Hypothesis of phase shift root cause: Retardation between plasmaexcitation (electron density) and resulting gas composition
High speedSEERS
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 8
Dynamic effects in plasma: Plasma stabilisation at process start
Electron density vs. Process time
2.E+08
4.E+08
6.E+08
8.E+08
1.E+09
0 5 10 15 20 25 30 35time [s]
elec
tron
den
sity
[cm
-3]
Plasma density oscillationcaused by rotating B- field
Process stability reachedafter about 20 s
High speedSEERS
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 9
Strategy of plasma parameter application
within AEC/APC
ApplicationStrategy
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 10
The Plasma is your Tool
Some parameters can bemeasured, for example:
MFC gas flow, pressure …Other parameters cannot bemeasured, for example:
Gas adsorption and desorption at chamber wall = conditioning
Processes on wafer surfaceare difficult to monitor.
The chamber is the tool box only.The plasma in the chamber – that´s your Tool really.
Cathode, chamber kit,RF Power installation
E- chuck
Plasma
Wafer
ApplicationStrategy
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 11
Plasma parameters = key indicators of process conditions
How electron collision rate νe and electron density neindicate process conditions in plasma (simplified equations):
RF power input Gas PressureB- Field Gas Temperature
Gas Composition
Plasma parameters summarise process impacts, which are:known and measured easilyknown and difficult to measureunknown and cannot be measured at all
Plasma parameters = key indicators of process conditions
ne ~ B02 Upeak f(pGas)
)veσ1(ve-1)
ε02νe =
e2 ne2
B2
pgas
kBTgas( p1
p + ...
ApplicationStrategy
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 12
Plasma parameter application as a process health indicator in AEC/APC
Detection of statistical significant process drift:Caused by measurable effects with very high sensitivityCaused by indirectly measurable effectsCaused by recipe parameters, chamber, wafer
Data compression:By the measured parameter, „by Physics“ itself
Plasma parameters give youdirect information about the plasma - your Tool.
Plasma parameters arereal time process health indicators.
ApplicationStrategy
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 13
High speed SEERS applications within AEC/APC
Process stabilityChamber matching
FDC
ApplicationExamples
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 14
Process stability & chamber matching of an etch process
Stability and Matching
Electron Collision Rate Electron Density
Mainframe 14 chambers
Mainframe 24 chambers
Mainframe 34 chambers
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 15
Process stability and chamber matching of mainframe 2
Superimposition of long term drift and short term processcondition drift in plasmaWet clean impacton processconditions is clearlyseen by bothparameters
Stability and Matching
Mainframe 24 chambers
Mainframe 24 chamber
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 16
Process stability and chamber matchingof mainframe 2, chamber B and C
Electron collision rate and electron density indicate different effects on processconditions(see page 11)
Stability and Matching
Mainframe 2chamber B,C
Mainframe 2chamber B,C
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 17
Wet clean monitoring with plasma parametersof an etch chamber
Electron collision rate shows normal and abnormal wet clean clearly
Wet clean abnormal Wet clean normal
Stabilityand FDC
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 18
FDC: E- chuck fault detection in an etch chamber
Electron collision rate and electron density show e-chuck faults clearly
Stabilityand FDC
Plasma unstableafter wet clean
Plasma stable afterESC exchange
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 19
FDC: Detection of unstable plasma and arcing in an etch chamber with rotating B- field
Detection of unstable plasma and arcingby electron density(= plasma density) measurement
FDC
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 20
FDC: Conditioning monitoring and leak detection of CVD process
Leakdetected
Stabilityand FDC
Tool down
Abnormal firstwafer effectafter dry clean
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 21
Conditioning monitoring –First wafer effect after dry clean of CVD process
is normal First wafer effect is abnormal
First product waferafter dry clean
one point - one wafer one point - one wafer
one curve - one wafer one curve - one wafer
First product wafer after dry clean
Stabilityand FDC
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 22
FDC: First wafer effect and leak detection - discussion
Conditioning:Dry clean every 5th wafer Impact on chamber wall surfaceChamber re- conditioning takes about 20 .. 25s on followingproduct waferConditioning effects are not detected by tool parameters
Electron collision rate = „Conditioning indicator“Leak detection:
Leak was detected by increase of chamber pressure and electroncollision rateBut electron collision rate rise was about ten times higher thanpressure increase !
Electron collision rate detects very small leaks, which cannotbe found using tool parameters, e.g. pressure
Stabilityand FDC
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 23
Chamber matching: CVD chamber redesign effect on process stability
Same recipe afterchamber redesignElectron collision rate less noisy, butsignificant drift (gas composition)
Increase of electrondensity indicatesincreased plasmadensity (effectivepower input)
Stability and Matching
before redesignafter redesign
before redesignafter redesign
Andreas SteinbachSven BarthMemory Development Center - Infineon Technologies Dresden
4th European AEC/APC Conference 2003Grenoble, France
Page 24
Summary
Plasma parameters, like electron collision rate and electron density, characterise process conditions in plasma directly.High speed SEERS enables real time plasma parameter measurement with fast response time Short dynamic processes in plasma can be monitored.Applications of High speed SEERS were demonstrated with etch and CVD processes:
Process stability monitoring Chamber matching FDC
Plasma parameters can be used as real time process health indicators.