Process Optimization

Process Optimization

Process Flow for non-critical layer optimization START Find the swing curve for the desired resist thickness. Determine the resist thickness (spin speed) from the swing curve and find the corresponding Eo value (on silicon). The Eop can be determined as below:- i) 2 times the Eo(si). END

START Critical Layer Optimization Find the swing curve for the desired resist thickness. Swing Curve Determine the resist thickness (spin speed) from the swing curve and find the corresponding Eo value (on silicon). Find the smile curve for the desired line or space CD or hole size. Smile Curve Eth for underneath layer Determine the Eop from the smile curve. Determine the Eth for the layer underneath.

Critical Layer Optimization Cnt d The estimated Eop for the layer underneath can be determined as below:- i) 2 times the Eo(layer underneath) for line and space ii) 3 time the Eo (layer underneath) for hole. iii) Ratio of Eop/Eo (si) multiplies Eo (layer underneath). This method will be used when we want a more accurate estimation. Take the Eop for the layer underneath as the center exposure energy. Set the exposure time step to be about 5% of the Eop. Set the Center focus to be 0.0um and step = 0.25um. Coat the wafer with layer underneath (flat or topography) with desired resist thickness. Expose the wafer with the selected exposure condition (Test-2 Mode). Develop the wafer with the desired development program.

Critical Layer Optimization Cnt d END Find and decide on the location of measurement and measure the same point for every measurement chip. From the results, determine the best focus position that gives the best profile and profile margin. Measure the CD at bottom or cross-section (if needed) of the line or space or hole with FESEM or CD SEM. Take photographs for every measurement. Measure the linewidth or hole size dependence on focus position at the exposure time that gives the desired linewidth or hole size. Measure the linewidth or hole size dependence on exposure time at center focus position. From the results, determine the exposure time that gives the desired linewidth or hole size.

Resist thickness Optimization Firstly, the thickness vs. spin speed graph should be created to find out the spin speed for i.e 1.0 um resist thickness. Usually the thickness depends on the technology, layer to be patterned. I.e. 1.0 um for patterning gate and 1.80 um for patterning metal line. Thickness (A) 35000 30000 25000 20000 15000 10000 5000 0 2800 3000 3200 3400 3600 3800 4000 4200 4400 Spin speed (rpm)

Swing curve Swing curve is the graph of resist sensitivity versus thickness. This graph will show oscillating behavior as shown. The swing between maximum and minimum sensitivity occurs with a thickness change of l/4n where l is the wavelength and n is the resist's index of refraction. 140 clearing dose E [mj/cm²] 0 130 120 110 100 90 80 top of bottom 1.5 line 1.6 of line 1.7 1.8 resist thickness [µm]

Swing Curve Once the thickness is selected, the swing curve experiment for resist thickness was carried out to determine the optimum resist thickness (spin speed) and the threshold exposure time (Eo). I.e spin speed is A rpm For the swing curve experiment, 10 wafers can been coated with different spin speeds ranging from A - 100 rpm to A + 100 rpm. The current program spin speed of A krpm is chose as the center of the graph. The spin speed was distributed evenly by a factor of 100-rpm reduction/increment.

Swing curve At some points of the curve, slight variations in resist thickness shows serious variation in line width. Best control, resist thickness at +/- 100 A of optimum thickness. Variations in chip surface topography makes it impossible to achieve this level of control across the wafer.

Resist response to Exposure and Focus (Smile curve) line-width vs focus position 1.4 Line-width (um) 1.2 1 0.8 0.6 0.4 200 250 300 350 400 450 500 550 0.2-0.75-0.5-0.25 0 0.25 0.5 0.75 1 Focus position (um)

Smile Curve Coat wafer with desired resist thickness Expose the wafer with below condition For Line and space, Exposure time: 2xEo(si) Exposure Step: 0.1xEo(si) Focus: 0um, step:0.25um For hole, Exposure Time: 3xEo(si) Exposure Step: 0.25xEo(si) Focus: 0um, Step: 0.25um Develop the wafer with the desired development recipe (same as we use for swing curve experiment). Measure the line and space CD or hole size of the desired resolution for every chip. Plot the smile curve (linewidth/hole size vs focus position).

Metrology Film thickness Measurement resist thickness measurement Overlay System Alignment accuracy CD SEM Critical dimension Microscope Inspection of pattern and defects

Microscope

Overlay Structure

Defects

Defects

CD SEM

Process Control

Motivation for Photolithography Process Control Factory Cost Leading Edge Exposure Tool Cost 1400 1200 1000 800 600 400 200 0 1970 1975 1980 1985 1990 1995 2000 Factory capitol cost (USDM) versus Year 12 10 8 6 4 2 0 1970 1975 1980 1985 1990 1995 2000 2005 Cost (USDM) versus Year

CD Bias CD bias is the value of CD loss or gain in related to litho patterning and etching process. E.g CD litho = 0.55 um CD after etch = 0.53 um CD bias = -0.02 um CD bias will determine the CD for litho patterning CD. E.g. Usually specification for CD is for after etch. The litho CD is determined after finding the CD bias.

CD Control In high imaging resolution in IC fabrication, the printed CD must be controlled to ever higher precision. CD control can be affected by many parameters, such as illumination, resist transmission and contrast, etc,

Photolithography Yield Pyramid Total Yield Process Yield Electrical Yield Defects Overlay Litho Process critical Stability Etch dimension Process Mask Image Contrast Resist Contrast Resist masking Substrate

Source of CD Variation Process Apply / Softbake Expose Expose PEB Delay latitude Post Exposure Bake Develop Control Parameter Thickness latitude Dose / Focus latitude PEB Delay Latitude PEB Temp latitude Develop latitude CD Variability ACLV, Xwafer, Xlot, Lotlot ACLV, Xlot, lot-lot ACLV, Xwafer, Xlot, Lotlot ACLV, Xwafer, Xlot, Lotlot Xwafer, Lot-lot

Across Chip Linewidth Variation Wavelength 436/365 365 365/248 248 248 248 Groundrule (nm) 800 500 350 250 180 150 Within Chip Photofield uniformity 50 30 25 20 18 15 Reticle CD uniformity 70 45 35 20 9 7 Proximity effect 65 40 25 20 18 15 Reflective notching / TFI 50 30 20 15 12 10 Total ACLV RSS 119 74 54 38 30 24 RSS = root sum square analysis Across-field CD variation can be regarded as largely systematic and can be represent at least 15% (3sigma) of nominal CD.

Scanner vs. Stepper

Scanner System

Scanner vs Stepper Scanner Stepper Resolution * Lens Aberration * Overlay Accuracy = = Mix & Match * Throughput * * Stability * * Price * *: Good

Scanner Advantage Smaller Projection Lens Diameter Averaged Effect by Scanning Flexibility for Intra- Field compensation Higher NA Better Lens Aberration Larger DOF Better matching Accuracy

Smaller Lens

Glossary Reticle - photomask: a master template of IC circuit used for IC fabrication Photomask - same as reticle Pellicle - a membrane mounted on the reticle / photomask to protect the IC designs from contamination and damage. CD -critical dimension is the size of critical pattern fabricated on the wafer. SEM - Scanning Electron Microscope Overlay - the measure of alignment accuracy from one pattern layer to the next pattern layer. Resist - photosensitive chemical used in lithography that changes dissolution properties when exposed to light energy. UV - ultra violet Excimer laser - Short wavelength light source for advance photolithography, 248 nm, 193 nm and 157 nm

Glossary I-line - light at wavelength 365 nm g-line - light at wavelength 436 nm swing curve - the plot of exposure dose versus thickness or spin speed smile curve - the plot of dimension versus focus at varying exposure dose CA - chemically amplified resist, the resist for advance photolithography (DUV)