Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
1
Sensors and Metrology
A Survey
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
2
Outline
• General Issues & the SIA Roadmap
• Post-Process Sensing (SEM/AFM, placement)
• In-Process (or potential in-process) Sensors
– temperature (pyrometry, thermocouples, acoustic waves)
– pressure and flow (manometers, momentum devices)
– composition (OES, LIF, RGA, mass Spectroscopy Actinometry)
– thickness (reflectometry, ellipsometry, scatterometry)
– smart-dummy wafers and smart substrates
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
3
Introduction
• Sensors (and actuators) are key limiting factors in
application of control techniques to semiconductor
manufacturing
• sources of difficulty
– implementation environment (vacuum, clean facilities, etc.)
– perception that in-situ sensors affect process
– ex-situ sensors can reduce throughput
– cost of ownership
– traditional resistance in industry
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
4
General Remarks on Sensors
• modeling is often key part of sensing
– physical quantity of interest may not be directly measured (ex:
OES indirectly contains info about etch process state)
– thus, sensors are based on a model of the underlying physical
process
sensors = data + model
• signal processing
– needed to reduce noise, improve bandwidth
– difference between data and information
• problems
– sensors require calibration
– must account for drift
• other issues
-
sensor fusion
- data compression
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
5
Key Issues in Sensors
• some key tradeoffs
– non-invasive vs. invasive
– non-destructive vs. destructive
– in-situ vs. ex-situ
– speed vs. accuracy
– noise
• bias (accuracy) vs. precision (repeatability + reproducibility)
– a sensor could be inaccurate, (thermocouple readings are off by 4
o
K)
– but the sensor might have good precision, (it is consistently off)
– precision is often more important for process control
• modern filtering and estimation methods can be of great use
in improved sensing software.
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
6
Accuracy vs. Precision
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
7
SIA Road Map Challenges above 45nm (through 2009)
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
8
SIA Road Map Challenges below 45nm (beyond 2009)
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
9
The 2004 update Metrology Road Map – near term
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
10
The 2004 update Metrology Road Map – long term
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
11
The 2004 update Lithography Metrology Road Map –near term
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
12
The 2004 update Lithography Metrology Road Map – long term
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
13
CD Metrology
• CD-SEM is today’s pre-eminent technique…
• Electron yield in interaction volume is a function of
surface topography (Secant effect) and atomic number.
• Extracting CD is not so simple...
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
14
Various Edge Detection Algorithms are in use...
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
15
CD-SEM is often calibrated with AFM
Atomic force Microscopy uses either
repulsive forces (sub nm proximity), or
weaker attractive forces (a few nm
away).
Tip tracks surfaces using feedback
control.
Shape and size of tip is the critical
source of errors.
Standard features are use to calibrate
and “de-convolve” the tip profile from the
measurements.
AFM is 100-1000 times slower than a
CD-SEM.
AFM is sensitive to line-edge roughness.
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
16
CD-SEM vs. AFM standards
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
17
CD-SEM cannot “see” the actual profile, so AFM is
used for this purpose...
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
18
AFM/SEM reading of Dense vs. Isolated Lines
dense
iso
AFM
SEM
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
19
Comparison for Contact holes
AFM
SEM
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
20
More CD-SEM vs. AFM comparisons
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
21
Optical CD Measurement
I
• Not very repeatable
• Limited spatial resolution
• Relatively inexpensive
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
22
Electrical CD Measurement
Measurement is very repeatable and fast.
Can only be used in conductive layers.
Need at least one conductive layer and one insulator.
Can be extended for misalignment measurements.
SEM
Lines
ID: XXN
Grating
Horz Iso
W/S= 180/
1000
Grating
Vert Iso
W/S= 180/
1000
Grating
Horz Iso
W/S= 180/
360
Grating
Vert Iso
W/S= 180/
360
Grating
Horz
Medium
W/S = 180/
240
Grating
Vert Medium
W/S = 180/
240
Grating
Horz Dense
W/S = 180/
180
Grating
Vert Dense
W/S = 180/
180
DUT1
Vert
DUT2
Vert
DUT3
Horz
DUT4
Horz
VDP
DUT5
Horz
DUT6
Horz
DUT7
Vert
DUT8
Vert
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
32
SEM
Lines
ID: XXO
DUT1
Vert
DUT2
Vert
DUT3
Horz
DUT4
Horz
VDP
DUT5
Horz
DUT6
Horz
DUT7
Vert
DUT8
Vert
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
32
SEM
Lines
CF_LENS coma/flare
CD_LIN linearity
MEFH
MEF
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
23
Basic SEM Structure
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
24
The Many Modes of SEM
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
25
The Issue of Spatial Resolution
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
26
Chrome on Silicon Example
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
27
Signal Depends Strongly on Material
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
28
Voltage Contrast SEM
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
29
CD-SEM Resolution
Scanning Resolution shown to 1-5nm.
CD metrology on resist has 5-10nm precision.
Other solutions:
ATF, has (theoretically) atomic resolution.
Problem:
What is CD??
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
30
CD-SEM Conclusions
• Accuracy is several (many) nm
• Precision (1-
σ repeatability + reproducibility) is
1-2nm today
• CD-SEM is stand-alone (i.e. expensive)
• CD-SEM measurements are available only after
patterning, and data integration with control
systems is difficult at best.
• AFM-based calibration will not be possible for
trenches less than ~100nm wide.
Lecture 17: Sensors and Metrology I
Spanos
EE290H F05
31
State of the Art CD/Imaging-SEM