Method of predicting an optimal process condition model to improve a yield of a semiconductor fabrication process and method of controlling a semiconductor fabrication process based on an optimal process condition model

1 . A method of selecting an optimal process condition model for a semiconductor wafer fabrication process, the method comprising:

collecting a plurality of process parameter information for a unit process used in the semiconductor fabrication wafer process;

obtaining a plurality of first characteristic information of a test semiconductor wafer before performing the unit process and obtaining a plurality of second characteristic information of the test semiconductor wafer after performing the unit process;

calculating a process global uniformity (PGU) using the plurality of first characteristic information and the plurality of second characteristic information;

creating a data set corresponding to the unit process using the plurality of process parameter information and the PGU;

creating a virtual process environment function with respect to the unit process using the data set;

creating an optimal process condition model with respect to the unit process using the virtual process environment function,

setting, by a user, at least one process target of the unit process, and

performing the unit process using the at least one process target on a production semiconductor wafer,

wherein the process global uniformity is calculated with respect to changes in values of two or more from among a thickness, a size, a shape, crystallinity, a resistance, a current-voltage characteristic, a composition, and a concentration,

wherein the optimal process condition model includes a control information that satisfies a process capability target for performing the unit process, and

wherein the control information comprises at least one of a plurality of information for performing the unit process selected from a process pressure, a process temperature, a process time, a process period, a number of a process performance, a kind of a process gas, a kind of a process material, a flux of a supply gas, an applying power source, a bias voltage, an impedance and an exposing amount.

2 . The method of claim 1 , wherein the unit process comprises at least one of an oxidation process, a photo process, a diffusion process, a thin film process, a metallization process, an etch process, a CMP process and a cleaning process.

3 . The method of claim 1 , wherein calculating the PGU comprises pre- processing the PGU using a polynomial fitting method.

4 . The method of claim 1 , wherein creating the virtual process environment function comprises creating the virtual process environment function using a supervised learning based regression analysis model from the data set.

5 . The method of claim 1 , wherein creating the virtual process environment function comprises verifying a capacity of the virtual process environment function.

6 . The method of claim 5 , wherein verifying the capacity of the virtual process environment function performs using at least one of a determination coefficient (R 2 ) calculation and an F-statistic verification.

7 . The method of claim 1 , wherein creating the optimal process condition model comprises:

creating a solution set of the unit process satisfying the at least one process target.

8 . The method of claim 7 , wherein creating the solution set of the unit process comprises:

creating a plurality of the solution sets with respect to the at least one process target using a DNA algorithm; and

evaluating an adequacy of the solution sets to create an optimal solution set with respect to the at least one process target.

9 . A method of performing a unit process in a semiconductor fabrication process, the method comprising:

collecting a plurality of process parameter information for the unit process of the semiconductor fabrication process;

obtaining a plurality of first characteristic information of a test semiconductor wafer before performing the unit process and obtaining a plurality of second characteristic information of the test semiconductor wafer after performing the unit process;

calculating a process global uniformity (PGU) using the plurality of first characteristic information and the plurality of second characteristic information;

creating a data set corresponding to the unit process using the plurality of process parameter information and the PGU;

creating a virtual process environment function with respect to the unit process using the data set;

creating an optimal process condition model with respect to the unit process using the virtual process environment function;

setting, by a user, at least one process target of the unit process,

selecting, by a user, a process capability target for performing the unit process, creating a solution set with respect to the at least one process target of the unit process using the optimal process condition model;

creating a control information of the unit process using the solution set, and

performing the unit process using the at least one process target and the process capability target on a production semiconductor wafer,

wherein the process global uniformity is calculated with respect to changes in values of two or more from among a thickness, a size, a shape, crystallinity, a resistance, a current-voltage characteristic, a composition, and a concentration,

wherein the optimal process condition model creates the control information that satisfies the selected process capability target, and

wherein the control information comprises at least one of a plurality of information for performing the unit process selected from a process pressure, a process temperature, a process time, a process period, a number of a process performance, a kind of a process gas, a kind of a process material, a flux of a supply gas, an applying power source, a bias voltage, an impedance and an exposing amount.

10 . The method of claim 9 , wherein the unit process comprises at least one of an oxidation process, a photo process, a diffusion process, a thin film process, a metallization process, an etch process, a CMP process and a cleaning process.

11 . The method of claim 9 , wherein calculating the PGU comprises pre-processing the PGU using a polynomial fitting method.

12 . The method of claim 9 , wherein creating the virtual process environment function comprises creating the virtual process environment function using a supervised learning based regression analysis model from the data set.

13 . The method of claim 9 , wherein creating the virtual process environment function comprises verifying a capacity of the virtual process environment function.

14 . The method of claim 13 , wherein verifying the capacity of the virtual process environment function performs using at least one of a determination coefficient R 2 calculation and an F-statistic verification.

15 . The method of claim 9 , wherein creating the solution set of the unit process comprises:

creating a plurality of the solution sets with respect to the at least one process target using a DNA algorithm; and

evaluating an adequacy of the solution sets to create the-an optimal solution set with respect to the at least one process target.