Method of determining model parameters for a MOSFET compact model using a stochastic search algorithm

• US 6,314,390 B1
• Filed: 11/30/1998
• Issued: 11/06/2001
• Est. Priority Date: 11/30/1998
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Watts, Josef S., Bittner, Calvin J., Hoffmann, James P.
• Primary Examiner(s): Powell, Mark R.
• Assistant Examiner(s): JONES, HUGH M
• Application Number:
  US09/201,206
• Time in Patent Office: 1,072 Days
• Field of Search: 703/14, 703/4
• US Class Current: 703/14
• CPC Class Codes: G06F 30/367 Design verification, e.g. using simulation, simulation program with integrated circuit emphasis [SPICE], direct methods or relaxation methods

Abstract

A method of determining a set of parameters for modeling an active semiconductor device in which current flow through a channel or other area is regulated by voltage applied to the device terminals, for example, MOSFETs. The method comprises first providing a plurality of measured values for current as a function of voltage for a plurality of active semiconductor devices of differing geometries. There is then determined an initial population of vectors comprising individual values representing a plurality of desired active semiconductor device model parameters. Fitness is then evaluated for each of the vectors by comparing calculated values for current as a function of voltage from the population to the plurality of measured values for current as a function of voltage of the vectors, converting any current differences to voltage errors and adding any such voltage errors together to arrive at a fitness value for each vector. Vectors of best fitness are selected and at least one genetic operator is applied thereto to create a new population of the vectors. Vectors of best fitness are then selected. The steps of evaluating fitness and selecting vectors of best fitness are optionally repeated for such vectors of best fitness until a desired fitness is achieved to determine the desired active semiconductor device model parameters.

Claims

1. A method of determining a set of parameters for modeling an active semiconductor device in which current flow through a channel or other area is regulated by voltage applied to the device terminals comprising the steps of:

a) measuring a plurality of values for current as a function of voltage for a plurality of active semiconductor devices of differing geometries;

b) determining an initial population of vectors comprising individual values representing a plurality of desired active semiconductor device model parameters;

c) evacuating fitness for each of said vectors by;

i) comparing calculated values for current as a function of voltage from said population to the plurality of measured values for current as a function of voltage of said vectors,

ii) converting any current differences to voltage errors and

iii) adding any such voltage errors together to arrive at a fitness value for each vector;

d) selecting vectors of best fitness and applying at least one genetic operator thereto to create a new population of said vectors;

e) adding meta evolution parameters to each vector from the new population of said vectors having a sufficiently low population of diversity to control genetic operators and permit the rate and direction of change to evolve with each vector from generation to generation;

f) selecting vectors of best fitness and applying at least one genetic operator to the vectors having added meta evolution parameters to create a new population of said vectors; and

g) selecting vectors of best fitness and optionally repeating steps (c) through (f) for such vectors of best fitness until a desired fitness is achieved to determine said desired active semiconductor device model parameters.

2. The method of claim 1 wherein the measured values for current as a function of voltage for a plurality of active semiconductor devices comprise values for drain current as a function of gate voltage, drain current as a function of drain voltage, and self gain.

3. The method of claim 1 wherein step (a) comprises measuring a plurality of values for current as a function of voltage for a plurality of MOSFETs of differing geometries.

4. A method of determining fitness of a vector comprising a set of parameters for modeling an active semiconductor device in which current flow through a channel or other area is regulated by voltage applied to the device terminals, said set of parameters of said vectors including added meta evolution parameters for controlling genetic operators and permitting the rate and direction of change to evolve with each vector from generation to generation, comprising the steps of:

a) measuring a plurality of values for current as a function of voltage for a plurality of active semiconductor devices of differing geometries, the measured values for current as a function of voltage for a plurality of active semiconductor devices comprising values for drain current as a function of gate voltage, drain current as a function of drain voltage, and self gain;

b) assigning an error to each current value calculated from said vector which differs from the measured values of drain current as a function of gate voltage and drain current as a function of drain voltage by converting any current differences to voltage errors;

c) assigning an error to each self gain value calculated from said vectors which differs from the measured values of self gain by converting any self gain differences to voltage errors;

d) assigning a penalty based on drain conductance and transconductance model parameters;

e) assigning a penalty based or threshold voltage calculated from model parameters for a plurality of active semiconductor device channel width and length combinations; and

f) combining said errors and penalties to determine fitness for said vector having said added meta evolution parameters.

5. The method of claim 4 wherein step (a) comprises measuring a plurality of values for current as a function of voltage for a plurality of MOSFETs of differing geometries.

6. A method of determining a set of parameters for modeling an active semiconductor device in which current flow through a channel or other area is regulated by voltage applied to the device terminals comprising the steps of:

a) measuring a plurality of values for current as a function of voltage for a plurality of active semiconductor devices of differing geometries;

b) determining an initial population of vectors comprising individual values representing a plurality of desired active semiconductor device model parameters;

c) calculating values for a subset of parameters comprising fewer than all of said desired model parameters from each of said vectors;

d) evaluating fitness for each of said vectors by comparing vector individual values to the measured values using a fitness function;

e) selecting vectors of best fitness and applying at least one genetic operator thereto to create a new population of said vectors;

f) testing diversity of said population of vectors and optionally repeating steps (c) through (e) until a sufficiently low vector population diversity is achieved;

g) adding meta evolution parameters to each vector from the vectors having sufficiently low population diversity of step (f) to control genetic operators and permit the rate and direction of change to evolve with each vector from generation to generation;

h) applying at least one genetic operator to the vectors having added meta evolution parameters to create a new population of said vectors;

i) calculating values for said subset of parameters from each of said new population of vectors created in step (h);

j) evaluating fitness for each of said vectors by comparing vector individual values to the measured values using a fitness function; and

k) selecting vectors of best fitness and optionally repeating steps (h) through (j) for such vectors of best fitness until a desired fitness is achieved to determine said desired active semiconductor device model parameters.

7. The method of claim 6 wherein the measured values for current as a function of voltage for a plurality of active semiconductor devices comprise values for drain current as a function of gate voltage, drain current as a function of drain voltage, and self gain.

8. The method of claim 6 wherein said subset of parameters is selected from the group consisting of threshold voltage, mobility factor, subthreshold swing factor and offset voltage in the subthreshold region.

9. The method of claim 6 wherein in step (c) calculating said values for a subset of parameters from each of said vectors is by applying a Newton-Raphson algorithm.

10. The method of claim 6 wherein in step (d) using said fitness function comprises assigning an error to calculated values for current as a function of voltage from said population as compared to the plurality of measured values for current as a function of voltage of said vectors by converting any current differences to voltage errors and adding any such voltage errors together.

11. The method of claim 6 wherein the measured values for current as a function of voltage for a plurality of active semiconductor devices comprise values for drain current as a function of gate voltage, drain current as a function of drain voltage, and self gain, and wherein in step (d) using said fitness function comprises the steps of:

i) assigning an error to each current value calculated from said vectors which differs from the measured values of drain current as a function of gate voltage and drain current as a function of drain voltage by converting any current differences to voltage errors;

ii) assigning an error to each self gain value calculated from said vectors which differs from the measured values of self gain by converting any self gain differences to voltage errors;

iii) assigning a penalty based on drain conductance and transconductance model parameters;

iv) assigning a penalty based on threshold voltage calculated from model parameters for a plurality of active semiconductor device channel width and length combinations; and

v) combining said errors and penalties to determine fitness for each of said vectors.

12. The method of claim 6 wherein in step (e) genetic operators of crossover and mutation are applied to the selected individuals of best fitness.

13. The method of claim 6 wherein in step (h) said genetic operator is selected from the group consisting of crossover and mutation.

14. The method of claim 6 wherein steps (c) through (e) are repeated until a sufficiently low vector population diversity is achieved and wherein steps (h) through (j) are repeated until a desired fitness is achieved to determine said desired active semiconductor device model parameters.

15. The method of claim 6 wherein step (a) comprises measuring a plurality of values for current as a function of voltage for a plurality of MOSFETs of differing geometries.

16. A program storage device readable by a machine, tangibly embodying a database for containing i) a plurality of measured values for current as a function of voltage for a plurality of active semiconductor devices of differing geometries in which current flow through a channel or other area is regulated by voltage applied to the device terminals and ii) an initial population of vectors comprising individual values representing a plurality of desired active semiconductor device model parameters, said device further having a program of instructions executable by the machine to perform a method of determining a set of parameters for modeling an active semiconductor device, said method comprising the steps of:

a) evaluating fitness for each of said vectors by;

i) comparing calculated values for current as a function of voltage from said population to said plurality of measured values for current as a function of voltage of said vectors,

ii) converting any current differences to voltage errors, and

iii) adding any such voltage errors together to arrive at a fitness value for each vector;

b) selecting vectors of best fitness and applying at least one genetic operator thereto to create a new population of said vectors;

c) adding meta evolution parameters to each vector from the new population of said vectors having a sufficiently low population of diversity to control genetic operators and permit the rate and direction of change to evolve with each vector from generation to generation;

d) selecting vectors of best fitness and applying at least one genetic operator to the vectors having added meta evolution parameters to create a new population of said vectors; and

e) selecting vectors of best fitness and optionally repeating steps (a) through (d) for such vectors of best fitness until a desired fitness is achieved to determine said desired active semiconductor device model parameters.

17. A program storage device readable by a machine, tangibly embodying a database for containing i) a plurality of measured values for current as a function of voltage for a plurality of active semiconductor devices of differing geometries in which current flow through a channel or other area is regulated by voltage applied to the device terminals and ii) an initial population of vectors comprising individual values representing a plurality of desired active semiconductor device model parameters, said device further having a program of instructions executable by the machine to perform a method of determining a set of parameters for modeling an active semiconductor device, said method comprising the steps of:

a) calculating values for a subset of parameters comprising fewer than all of said desired model parameters from each of said vectors;

b) evaluating fitness for each of said vectors by comparing vector individual values to said measured values using a fitness function;

c) selecting vectors of best fitness and applying at least one genetic operator thereto to create a new population of said vectors;

d) testing diversity of said population of vectors and optionally repeating steps (a) through (c) until a sufficiently low vector population diversity is achieved;

e) adding meta evolution parameters to each vector from the vectors having sufficiently low population diversity of step (d) to control genetic operators and permit the rate and direction of change to evolve with each vector from generation to generation;

f) applying at least one genetic operator to the vectors having added meta evolution parameters to create a new population of said vectors;

g) calculating values for said subset of parameters from each of said new population of vectors created in step (f);

h) evaluating fitness for each of said vectors by comparing vector individual values to said measured values using a fitness function; and

i) selecting vectors of best fitness and optionally repeating steps (f) through (h) for such vectors of best fitness until a desired fitness is achieved to determine said desired active semiconductor device model parameters.

18. The program storage device of claim 17 wherein said subset of parameters is selected from the group consisting of threshold voltage, mobility factor, subthreshold swing factor and offset voltage in the subthreshold region.

19. The program storage device of claim 17 wherein in step (a) calculating said values for a subset of parameters from each of said vectors is by applying a Newton-Raphson algorithm.

20. The program storage device of claim 17 wherein in step (b) using said fitness function comprises assigning an error to calculated values for current as a function of voltage from said population as compared to said plurality of measured values for current as a function of voltage of said vectors by converting any current differences to voltage errors and adding any such voltage errors together.

21. A program storage device readable by a machine, tangibly embodying a database for containing a plurality of measured values for current as a function of voltage for a plurality of active semiconductor devices of differing geometries, said measured values for current as a function of voltage for a plurality of active semiconductor devices comprising values for drain current as a function of gate voltage, drain current as a function of drain voltage, and self gain, said device further having a program of instructions executable by the machine to for modeling an active semiconductor device in which current flow through a channel or other area is regulated by voltage applied to the device terminals, said set of parameters of said vectors including added meta evolution parameters for controlling genetic operators and permitting the rate and direction of change to evolve with each vector from generation to generation, said method comprising the steps of:

a) assigning an error to each current value calculated from said vector which differs from the measured values of drain current as a function of gate voltage and drain current as a function of drain voltage by converting any current differences to voltage errors;

b) assigning an error to each self gain value calculated from said vectors which differs from the measured values of self gain by converting any self gain differences to voltage errors;

c) assigning a penalty based on drain conductance and transconductance model parameters;

d) assigning a penalty based on threshold voltage calculated from model parameters for a plurality of active semiconductor device channel width and length combinations; and

e) combining said errors and penalties to determine fitness for said vector having said added meta evolution parameters.