ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd 1 Prolith Modeling • Prolith Simulation INPUTS Database of input files Film Stack Photo resist Add a film to stack Resist thickness and Softbake mask Exp tool Focus expos ure PEB Develop CD metrology and Process latitude window Specifications
31
Embed
Prolith Modeling - Frontier Homepage Powered by Yahoomyplace.frontier.com/~stevebrainerd1...ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd 1 Prolith Modeling • Prolith Simulation
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
ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd
1
Prolith Modeling
• Prolith Simulation INPUTS
Database of input
files
Film Stack
Photo resist
Add a film to stack
Resist thickness
and Softbake
mask
Exp tool
Focus exposure
PEB
Develop
CD metrology and Process latitude window Specifications
ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd
2
Prolith Modeling
• Prolith Simulation OUTPUTS:
Aerial Image
Simulation Sets
Single run simulation photoresist profile
Develop time contours
Image in reisist
PAC conc Pre-PEB
PAC conc Post-PEB
ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd
3
Prolith Modeling Resist file .res• [Version]
• 7.2 <<<Prolith Version• [Parameters]• IX500EL :Resist Name• JSR ;Resist Vendor• 1 ;Read Only• 1 ;Resist Type (0=Negative, 1=Positive)• 0 ;Resist Type (0=Conventional, 1=Chemically Amplified)• 1 ;Number of Developers
• 1 ;Dev model (1=Mack, 2=Enhanced, 3=Notch) << Mack Model for Develop<< Mack Model for Develop• PD523AD ;Developer Used << TMAH Developer• 85.000 ;Development Rmax (nm/s) < Max Develop Rate 100% exposed photoresist• 0.009 ;Development Rmin (nm/s) < Min Develop Rate unexposed photoresist• 0.060 ;Development Mth << PAC threshold value in for Mack Model• 5.800 ;Development n << Development selectivity for models. High value higher contrast
• 0.050 ;Surface Development Rate << Develop rate relative to bulk <1.0 less rate at surface
• 10.000 ;Inhibition Depth (nm) <<,the depth for transition between surface rate and bulk rate• 34.320 ;Thermal Decomp. Ea(kcal/mole) << Activation energy for PAC decomp rate• 36.800 ;Thermal Decomp. ln(Ar) (1/s) << logarithm of the Arrhenius coefficient for PAC decomp• 35.000 ;PEB Diffusivity Ea (kcal/mole) << Activation energy for PEB diffusion• 49.350 ;PEB Diffusivity Ln(Ar) (nm2/s) << logarithm of the Arrhenius coefficient for PEB
1 ;Read Only1 ;Resist Tone (0=Negative, 1=Positive)0 ;Resist Type (0=Conventional, 1=Chemically Amplified)1 ;Number of Developers1 ;Dev model (1=Mack, 2=Enhanced, 3=Notch)PD523AD, 23C ;Developer Used
183.000 ;Development Rmax (nm/s)0.006 ;Development Rmin (nm/s)0.450 ;Development Mth15.000 ;Development n0.300 ;Surface Development Rate200.000 ;Inhibition Depth (nm)34.320 ;Thermal Decomp. Ea(kcal/mole)36.800 ;Thermal Decomp. ln(Ar) (1/s)35.000 ;PEB Diffusivity Ea (kcal/mole)49.350 ;PEB Diffusivity Ln(Ar) (nm2/s);ABC data is in the following format:;wavelength A B C Unexposed n Completely Exposed n ; (nm) (1/um) (1/um) (cm2/mJ)[ABC Data]
• 1.00um dimension Dose 197 mj/cm2� dose to achieve the 0.5um width dimension
ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd
17
Prolith Simulation Example: JSR IX965G 8500A1.00um Dimension FEM
ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd
18
Prolith Simulation Example: JSR IX965G 8500A1.0um length SEM Array Process Window
ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd
19
Prolith Simulation Example: JSR IX965G 8500A1.0um length Process Window
ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd
20
Prolith Modeling Calibration of Swing Curve : Making Simulation match actual data Procedure
• PROLITH MODELING METHODOLOGY: The following is a step by step instructions for simulation calibration of a specific photoresist using Prolith as used on a specified process layer.
•
• 1. Generate an actual swing curve on the process layer being characterized by varying thephotoresist thickness plus and minus approximately 1000A in 100A increments. Vary the thickness by varying the COATER spin speed RPMs. Prior to running this test, it is assumed that one has generated a plot of photoresist thickness Vs COATER RPM. Assure that the correct Cauchy values for this photoresist ( obtain these dispersion coefficient values from the vendor) are used in the Prometrix tool when the thickness are measured.
•
• 2. Vary the doses on each wafer ( execute an FEM routine but keep the focus constant) to assure that the nominal CD is obtained at the CD max. or CD min. point on the curve. Also runEo exposes on the wafers to obtain additional information.
•
• 3. Measure the CD feature of concern using the KLA 8100 CD -SEM on each wafer at 5 different exposure doses.
ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd
21
Prolith Modeling Calibration of Swing Curve : Making Simulation match actual data Procedure
• 4. Plot CD at a fixed dose Vs Photoresist thickness in an Excel spreadsheet. These curves are labeled ACTUAL DATA @ XX mj/cm2 dose.
•
• 5. Obtain the actual film thickness and optical constants (n real refractive index and extinction coefficient k) at the actinic radiation wavelength ( i-line 365nm) for all films under thephotoresist.
•
• 6. Setup the Parameter tables in Prolith 6.03 or higher for the substrate ( use actual film thickness and optical constants at i-line 365nm for all films under the photoresist ; film 1 isphotoresist on top), photoresist ( obtain ABC and develop parameters from vendor), Softbaketime/ temperature and photoresist thickness, critical dimension size and pitch, stepper exposure tool (define NA, sigma, and wavelength), nominal dose and focus setting that one used to achieve the desired CD at the desired photoresist thickness (CD max for linewidth and CD min. for spacewidth ), PEB time and temperature, develop time, and feature measurement metrology. For feature measurement metrology one can set the process specification CD limits and measurement tool threshold %.
ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd
22
Prolith Modeling Calibration of Swing Curve : Making Simulation match actual data Procedure
• Note: Based on my experience over the years using Prolith the AB and Development parameters Rmax and Rmin supplied by the vendor seem to be good enough to obtain very good results. Very good results mean a 5% to 8% error between simulator and actual data.
•
• 7. Run a Prolith simulated swing curve using the first revision setting defined in 6 above and the same thickness range and increment you used in generating the actual swing curve.
•
• 8. Import the Prolith simulated swing curve in the Excel spreadsheet sheet containing the actual swing curve. This is done by Holding down the CONTROL key and clicking and dragging the graph into the spreadsheet. One will need to transpose the data and change the thickness from nanometers to Angstroms and the CDs from nanometers to microns.
•
• 9. Add this simulated data to the actual swing curve plot.
ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd
23
Prolith Modeling Calibration of Swing Curve : Making Simulation match actual data Procedure
• 10. If the phase of the actual and simulated is off adjust the refractive index of the photoresistin Prolith and re-run the simulation. If the simulated curve is to the left of the actual curve for a linewidth decrease the unexposed and exposed n refractive index values. It means that the simulated photoresist thin films interference effect is incorrect.
•
• 11. If the amplitude of the actual and simulated is off adjust the Dill C parameter of thephotoresist in Prolith and re-run the simulation. If the simulated curve is lower than the actual curve for a linewidth decrease the C value. It means that the simulated photoresist photospeedis too fast ( assuming a linewidth).
•
• 12. Repeat steps 10 and 11 until the simulated curve matches the actual curve for that substrate. Once this is done you have “calibrated” the simulator for that photoresist. One can now simulate with good confidence the effects of changing the exposure tool, the thin films under the resist, and the thin film thickness under the photoresist for example.
ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd
24
Prolith Modeling Effect of Swing Curve Calibration EXAMPLE: JSR IX405 on Polysilicon
IX965G ;Resist Name JSR ;Resist Vendor1 ;Read Only1 ;Resist Tone (0=Negative, 1=Positive)0 ;Resist Type (0=Conventional, 1=Chemically Amplified)1 ;Number of Developers1 ;Dev model (1=Mack, 2=Enhanced, 3=Notch)PD523AD, 23C ;Developer Used183.000 ;Development Rmax (nm/s)0.006 ;Development Rmin (nm/s)0.450 ;Development Mth15.000 ;Development n0.300 ;Surface Development Rate200.000 ;Inhibition Depth (nm)34.320 ;Thermal Decomp. Ea(kcal/mole)36.800 ;Thermal Decomp. ln(Ar) (1/s)35.000 ;PEB Diffusivity Ea (kcal/mole)49.350 ;PEB Diffusivity Ln(Ar) (nm2/s);ABC data is in the following format:
;wavelength A B C Unexposedn Completely Exposed n ; (nm) (1/um) (1/um) (cm2/mJ)
[ABC Data]365.0000.790 0.050 0.0120 1.700 1.700
ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd
26
Prolith Modeling Effect of Dills B, C, and n on swing Curve
PROLITH Swing CURVES: n , Dill B and C effects IX965G
750770790810830850870890910930950970990
101010301050
800 820 840 860 880 900 920 940 960 980 1000
IX965G Thickness nm
CD
sp
ace
nm
CD nm n =1.70 b = 0.05c =0.012 cm2/mjCD nm n =1.60 b = 0.05 c =0.012 cm2/mjCD nm n =1.60 b = 0.05 c =0.010 cm2/mjCD nm n =1.60 b = 0.45 c =0.010 cm2/mj
CCCC
BBBB
n
ECE-580-MPE Prolith_Modelling.ppt Steve Brainerd
27
Prolith Modeling Effect of Dills B, C, and n on Swing Curve
• ;ABC data is in the following format: • ;wavelength A B C Unexposed n Completely Exposed n• ; (nm) (1/um) (1/um) (cm2/mJ)• [ABC Data]• 365.000 0.973 0.104 0.0161 1.665 1.665
• ;ABC data is in the following format: • ;wavelength A B C Unexposed n Completely Exposed n• ; (nm) (1/um) (1/um) (cm2/mJ)• [ABC Data]• 365.000 1.010 0.102 0.0154 1.680 1.680
• ;ABC data is in the following format: • ;wavelength A B C Unexposed n Completely Exposed n• ; (nm) (1/um) (1/um) (cm2/mJ)• [ABC Data]• 365.000 0.860 0.089 0.0100 1.700 1.700
• ;ABC data is in the following format: • ;wavelength A B C Unexposed n Completely Exposed n• ; (nm) (1/um) (1/um) (cm2/mJ)• [ABC Data]• 365.000 0.918 0.152 0.0153 1.655 1.655