Quantitative In-Situ Measurement of Asperity Compression Under
the Wafer During PolishingCaprice Gray
MRS Spring MeetingSan Francisco, CA
March 30, 2005
Overview
• Introduction• Experimental Setup
– Polisher– Dual Emission Laser Induced Fluorescence
• Pad Compression Experiments and Results– Static Compression– Dynamic Compression
• Conclusion
Introduction
• DELIF is used to collect grayscale images that represent properties under the wafer during polishing– pH, temperature, fluid layer thickness
• Previous thickness measurement– time averaged– low spatial resolution
• New DELIF Setup– Laser enables instantaneous imaging– Larger distance between lens and CCD – 4 m/pixel
• In these experiments we use fluid thickness measurements to infer what is happening to pad asperities during polishing
Polishing Setup
Motor
Platen
RotoPol-31
Wafer
Force Table
Steel Table
DELIF
0
0.2
0.4
0.6
0.8
1
300 400 500 600
Wavelength (nm)
Inte
nsi
ty (
au)
Camera B Filter
Camera A Filter
Pad Em.392 nm
Calcein Abs492 nm
Calcein Em.530 nm
Pad Abs. 346nm
Laser Em.355 nm
Slurry Abs.
Slurry Em.
Why Examine 2 Emissions?
Pad
Slurry
=
Ratio
Division of 2 images cancels variationsin image source intensity
Experiments
• Pads: flat Fruedenberg FX9– Global vs. Local topography
• Static: flat glass wafer• Dynamic: etched glass wafer
Static Pad Compression• 2 Images were taken in a single
section of the pad.– 0 kPa– 70 kPa
• 17 total sections were imaged and analyzed.
Histogram Analysis
• The asperity size distribution is the same shape as the fluid layer thickness distribution.
• When force is applied to the wafer, the distribution changes both shape and location.
• Standard deviation comparison → pad compression
• Peak location → fluid layer thickness.
• Compression factor: 0
0
Static Compression Results
= 133.9 - 22.2R2 = 0.7
0
10
20
30
40
50
60
0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Compression Factor
(S
lurr
y T
hic
knes
s) (
um
)
Dynamic Pad Compression
• Two wafers were etched and contain square wells– 14.5 m deep– 27 m deep
• Asperities can expand under these wells.
• Histograms from regions 1 and 2 are compared
• Run Conditions:– Pad/wafer rotation: 30 RPM– Continuous conditioning– Slurry flow: 50 cc/min– Force applied to wafer: 10 kPa
Dynamic Results
(7m)
Conclusions
• DELIF can be used to measure slurry film thickness with asperity scale resolution.
• Over the 6 mm2 regions imaged, Different amounts of fluid can be displaced due to the same applied global forces.
• Pad topography deformation can be inferred from fluid layer thickness measurements.– During a static application of force, local pad compression is
roughly linearly related to the amount of fluid displaced by the compression
– In the dynamic case, asperity expansion into 27 m deep wells was greater than expansion into 14.5 m deep wells.
Acknowledgements
• Tufts– Professor Chris Rogers– Professor Vincent P. Manno– Dan Apone – TUFTL
• Intel Corporation– Chris Barns– Mansour Moinpour
• Cabot Microelectronics– Sriram Anjur
• University of Arizona– Ara Philipossian