INSTANTANEOUS IN-SITU IMAGING OFSLURRY FILM THICKNESS DURING CMP
Caprice Gray, Daniel Apone, Chris Rogers, Vincent P. Manno, Chris Barns, Mansour Moinpour, Sriram Anjur, Ara Philipossian
Motivation
Microelectronic devices continue to decrease in size; current features are routinely smaller than 100nm
The semiconductor industry requires a deeper understanding of the physical processes involved in CMP to help attain smoother surfaces
Using Dual Emission Laser Induced Fluorescence (DELIF) we can measure instantaneous fluid film thicknesses (and temperatures) during a polishing run
Here we look at how the pad conforms to features on a wafer
Polishing Setup
Struers RotoPol-31 table top polisher
Polisher sits atop a force transducer table capable of measuring down and shear forces during a polish
Optical Setup
Evolution VF 12 bit digital cameras Region of
Interrogation: 2mm by 3mm on the pad
355 nm Nd-YAG Laser provides excitation light Laser Pulse Length: 6ns
Dual Emission Laser Induced Fluorescence
Calcein, Coumarin in slurry solution UV light excites Coumarin Coumarin emission excites Calcein Each emission is captured by a camera Taking the ratio of the two emissions normalizes the image
by initial excitation intensity Images taken are 3 second temporal averages Note: pads must be dyed black to mute any fluorescence
DELIF with One Dye Natural pad fluorescence replaces
Coumarin; Laser replaces UV lamps Allows for non dyed pads Laser now excites pad Pad emission then excites Calcein in
slurry Since Laser is much more powerful than
UV lamps, we can now take instantaneous images, not 3 second averages as before.
Experimental Parameters Freudenberg FX9 Pad Wafer & Platen Rotation: 30 rpm
Relative Velocity: 0.34 m/s Downforce: 1.8 PSI Slurry
Flow Rate: 50 cc/min9:1 dilution0.5 g/L Calcein
Previous Work
Film thickness increases as pad speed increases
Inverse relationship for downforce and thickness
Film thickness are measured from the wafer surface down to some mean height within the pad
Calibration
These known heights allow for a calibration of intensity to fluid film thickness
The 27 micron deep well (b) is brighter than the 14 micron deep well (a), indicating more fluid
(b)(a)
2mm
Results Surface roughness calculations compare
single points in an image to a mean thickness value
Indicates the wafer is compressing the pad
Roughness of:Red square = 3.40.3Blue square = 4.20.3mm FX9 Pad (from profilometer) =
4.30.3mm
Results
22% of images of the 27 micron deep well show air bubbles
The roughness in the air bubble is between the roughness inside and outside of the well
(b)(a)
ResultsAverage Surface Roughness with Variable
Down-Force
4.24.8
4.4
0
1
2
3
4
5
6
0.4 1.8 3.2
Pressure (psi)
Su
rfa
ce
Ro
ug
hn
es
s (
um
)
Previous modeling research has shown that pressure varies locally beneath a wafer suggesting that we must interrogate many other regions before we can draw any significant conclusion about roughness variation with applied global down force.
Conclusion
This work supports the notion that the CMP polishing regime is in the partial lubrication regime Wafer is partly supported by asperities, partly by fluid
pressure If it were true hydrodynamic lubrication, the
roughness under the wells would be the same as the rest of the pad
The asperities are free to expand under the etched wells and do so
Future Work
Investigate larger region/multiple regions under the waferCorrelate downforce with surface roughness
Refine calibration method to determine absolute thicknessRoughness reported here is a relative
measurement