Physics-based MOSFET model for variational modeling

1. A method of optimizing MOSFET device production comprising:

defining, using a computer, key independent parameters;

formulating key independent parameters into a canonical variational form, wherein the canonical variational form comprises a key independent parameter extracted at nominal, or target, device specifications plus a product of magnitude of a variation of the key independent parameter, relative to a skew of a distribution on device characteristics, and a selected distribution;

calculating theoretical extracted parameters using at least one of key independent parameters in canonical variational form, physics-based analytical models, and corner models;

calculating simulated characteristics of a device using at least one of key independent parameters in canonical form or theoretical extracted parameters using a variational model;

extracting target data parameters based on at least one of measured data and predicted data;

comparing the simulated characteristics to the target data parameters;

modifying the theoretical extracted parameters based on the comparison to the target data parameters;

modifying the key independent parameters in canonical form based on the modified theoretical extracted parameters;

calculating and outputting the simulated characteristics based on the modified theoretical extracted parameters and the modified key independent parameters in canonical form; and

modifying the key independent parameters in the device production using the simulated characteristics.

2. The method of claim 1 , wherein key independent parameters comprise at least one of:

a flatband voltage, for both n-type and p-type devices;

a channel doping concentration at depletion edge of zero body bias, for both n-type and p-type devices;

an electrical oxide thickness;

a zero biased lightly doped drain resistance per unit width, for both n-type and p-type devices;

a channel length offset correction due to production techniques; and

a channel width offset production due to production techniques.

3. The method of claim 1 , wherein the variational model uses system parameters comprising BSIM4 physical model parameters.

4. The method of claim 1 , wherein the method of extracting the theoretical statistical parameters based on the target data parameters comprises at least one of:

identifying and measuring data parameters and distribution of data parameters in process control monitoring;

extracting statistical model parameters from measured data parameters; and

determining a distribution of model parameters from the measured data parameters and the distribution of the measured data parameters.

5. The method of claim 4 wherein the statistical model parameters comprise at least one of:

an electrical oxide thickness;

a standard deviation of the electrical oxide thickness;

a flatband voltage, for both n-type and p-type devices;

a standard deviation of the flatband voltage, for both n-type and p-type devices;

a channel doping concentration at depletion edge of zero body bias, for both n-type and p-type devices;

a standard deviation of the channel doping concentration at depletion edge of zero body bias, for both n-type and p-type devices;

a zero biased lightly doped drain resistance per unit width, for both n-type and p-type devices; and

a standard deviation of the zero biased lightly doped drain resistance per unit width, for both n-type and p-type devices.

6. The method of claim 4 , wherein the distribution of measured data parameters in process control monitoring comprises at least one of global variations, local variations, gradient variations, or combinations thereof.

7. The method of claim 1 , wherein the variational model is implemented on BSIM4.

8. The method of claim 1 , further comprising:

defining and nesting statistical model cards;

adding SPICE elements to a .subckt to model extrinsic characteristics;

implementing multiple spice simulations for different seed parameters based on the theoretical extracted parameters;

flattening a circuit netlist to transistor level;

perturbing at least one of the key independent parameters and the theoretical extracted parameters;

storing results in an accessible format;

comparing results to manufactured circuits; and

modifying at least one of manufacturing characteristics, key independent parameters, or theoretical extracted parameters based on the comparison.

9. The method of claim 1 , further comprising:

defining earliest and latest arrival times;

defining delays and output slews based on the key independent parameters and the theoretical extracted parameters;

using a statistical characterization tool to simulate delay arcs at random process points;

generating a response surface model to describe deviations in global and local variations;

storing results in an accessible format;

identifying and correcting circuits violating statistical requirements.

10. The method of claim 1 , wherein process control monitors are defined, and global inter-die, gradient intra-die, and local intra-die variations are extracted in the domains of chip mean, across chip, and mismatch.

11. A non-transitory computer readable medium comprising software instructions to optimize MOSFET device production, wherein the software instructions comprise instructions for causing a computer to perform:

calculating theoretical extracted parameters using at least one of key independent parameters in canonical variational form, physics-based analytical models, and corner models;

calculating simulated characteristics of a device using at least one of key independent parameters in canonical form or theoretical extracted parameters using a variational model;

extracting target data parameters based on at least one of measured data and predicted data;

comparing the simulated characteristics to the target data parameters;

modifying the theoretical extracted parameters based on the comparison to the target data parameters;

modifying the key independent parameters in canonical form based on the modified theoretical extracted parameters;

calculating and outputting the simulated characteristics based on the modified theoretical extracted parameters and the modified key independent parameters in canonical form; and

modifying the key independent parameters in the device production using the simulated characteristics,

wherein the canonical variational form comprises a key independent parameter extracted at nominal, or target, device specifications plus a product of magnitude of a variation of the key independent parameter, relative to a skew of a distribution on device characteristics, and a selected distribution.

12. The computer readable medium of claim 11 , wherein key independent parameters comprise at least one of:

a flatband voltage, for both n-type and p-type devices;

a channel doping concentration at depletion edge of zero body bias, for both n-type and p-type devices;

an electrical oxide thickness;

a zero biased lightly doped drain resistance per unit width, for both n-type and p-type devices;

a channel length offset correction due to production techniques; and a channel width offset production due to production techniques.

13. The computer readable medium of claim 11 , wherein the variational model uses parameters comprising BSIM4 physical model parameters.

14. The computer readable medium of claim 11 , wherein the method of extracting the theoretical statistical parameters based on the target data parameters comprises at least one of:

identifying and measuring data parameters and distribution of data parameters in process control monitoring;

extracting statistical model parameters from measured data parameters; and

determining a distribution of model parameters from the measured data parameters and distribution of the measured data parameters.

15. The computer readable medium of claim 14 wherein the statistical model parameters comprise at least one of:

an electrical oxide thickness;

a standard deviation of the electrical oxide thickness;

a flatband voltage, for both n-type and p-type devices;

a standard deviation of the flatband voltage, for both n-type and p-type devices;

a channel doping concentration at depletion edge of zero body bias, for both n-type and p-type devices;

a standard deviation of the channel doping concentration at depletion edge of zero body bias, for both n-type and p-type devices;

a zero biased lightly doped drain resistance per unit width, for both n-type and p-type devices; and

a standard deviation of the zero biased lightly doped drain resistance per unit width, for both n-type and p-type devices.

16. The computer readable medium of claim 14 , wherein the distribution of measured data parameters in process control monitoring comprises at least one of global variations, local variations, gradient variations, or combinations thereof.

17. The computer readable medium of claim 11 , further comprising:

defining and nesting statistical model cards;

adding SPICE elements to a .subckt to model extrinsic characteristics;

implementing multiple spice simulations for different seed parameters based on the theoretical extracted parameters;

flattening a circuit netlist to transistor level;

perturbing at least one of the key independent parameters and the theoretical extracted parameters;

storing results in an accessible format;

comparing results to manufactured circuits; and

modifying at least one of manufacturing characteristics, key independent parameters, or theoretical extracted parameters based on the comparison.

18. The computer readable medium of claim 11 , further comprising:

defining earliest and latest arrival times;

defining delays and output slews based on the key independent parameters and the theoretical extracted parameters;

using a statistical characterization tool to simulate delay arcs at random process points;

generating a response surface model to describe deviations in local variations;

storing results in an accessible format; and

identifying and correcting circuits violating statistical requirements.