Process-parameter prognostic system for predicting shape of semiconductor structure, semiconductor fabrication apparatus having the system, and method of using the apparatus
First Claim
Patent Images
1. A process-parameter prognostic system comprising:
- a process prediction unit configured to calculate predicted parameters using measured parameters of semiconductor structures and sensor parameters associated with plasmas in a process chamber used to form the semiconductor structures, wherein the process prediction unit is further configured to provide a process-predicting model based on selected measured parameters and selected predicted parameters of selected semiconductor structures and configured to repetitively update the process-predicting model based on the sensor parameters by adjusting a number of the selected semiconductor structures and a number of remaining semiconductor structures in an order of a semiconductor fabrication process to provide remaining predicted parameters corresponding to the process-predicting model; and
a process-change point corresponding unit configured to compare selected sensor parameters corresponding to the selected predicted parameters with the remaining sensor parameters corresponding to the remaining predicted parameters to confirm process environments in the process chamber;
wherein each of the semiconductor structures is one selected from a structure on a semiconductor substrate and a structure including a material layer on and different from the semiconductor substrate, the measured parameters comprise actual numerical figures for limiting the shape of the semiconductor structures, and the predicted parameters are expected numerical figures for estimating the shape of the semiconductor structures;
wherein the sensor parameters are classified to correspond to at least two different process items, and the at least two different process items correspond to the semiconductor structures, respectively, and correspond to independent variables of the process-predicting model, respectively;
wherein each of the sensor parameters comprises an average radio frequency of a waveform obtained between specific points in time during performance of the semiconductor fabrication process;
wherein the process-predicting model is linear between the selected measured parameters and the selected predicted parameters;
wherein the process prediction unit is configured to determine a boundary condition for the process-predicting model using the selected sensor parameters and is configured to compare the boundary condition with the remaining parameters;
wherein the boundary condition is one selected from a maximum value, a minimum value, an average value, a standard deviation value, which are obtained using the sensor parameters according to the at least two different process items, and a statistical value obtained using the maximum value, the minimum value, the average value, and/or the standard deviation value;
wherein when the remaining sensor parameters satisfy the boundary condition, the process prediction unit replaces a previously fabricated one of the selected semiconductor structures with a previously fabricated one of the remaining semiconductor structures to generate a first specific predicted parameter, replaces a subsequently fabricated one of the selected semiconductor structures with a subsequently fabricated one of the remaining semiconductor structures to generate a second specific predicted parameter, and repetitively performs the same process as the process of generating each of the first and second specific predicted parameters in order to generate the predicted parameters, andwherein the boundary condition is repetitively updated based on the process-predicting model using the selected sensor parameters and the remaining sensor parameters.
1 Assignment
0 Petitions
Accused Products
Abstract
Provided are a process-parameter prognostic system for predicting the shape of a semiconductor structure, a semiconductor fabrication apparatus having the process-parameter prognostic system, and a method of using the same. The process-parameter prognostic system may have a process prediction unit and a process-change point corresponding unit. The process prediction unit and the process-change point corresponding unit may obtain predicted parameters using measured parameters of semiconductor structures and sensor parameters of plasmas corresponding to the semiconductor structures.
26 Citations
5 Claims
-
1. A process-parameter prognostic system comprising:
-
a process prediction unit configured to calculate predicted parameters using measured parameters of semiconductor structures and sensor parameters associated with plasmas in a process chamber used to form the semiconductor structures, wherein the process prediction unit is further configured to provide a process-predicting model based on selected measured parameters and selected predicted parameters of selected semiconductor structures and configured to repetitively update the process-predicting model based on the sensor parameters by adjusting a number of the selected semiconductor structures and a number of remaining semiconductor structures in an order of a semiconductor fabrication process to provide remaining predicted parameters corresponding to the process-predicting model; and a process-change point corresponding unit configured to compare selected sensor parameters corresponding to the selected predicted parameters with the remaining sensor parameters corresponding to the remaining predicted parameters to confirm process environments in the process chamber; wherein each of the semiconductor structures is one selected from a structure on a semiconductor substrate and a structure including a material layer on and different from the semiconductor substrate, the measured parameters comprise actual numerical figures for limiting the shape of the semiconductor structures, and the predicted parameters are expected numerical figures for estimating the shape of the semiconductor structures; wherein the sensor parameters are classified to correspond to at least two different process items, and the at least two different process items correspond to the semiconductor structures, respectively, and correspond to independent variables of the process-predicting model, respectively; wherein each of the sensor parameters comprises an average radio frequency of a waveform obtained between specific points in time during performance of the semiconductor fabrication process; wherein the process-predicting model is linear between the selected measured parameters and the selected predicted parameters; wherein the process prediction unit is configured to determine a boundary condition for the process-predicting model using the selected sensor parameters and is configured to compare the boundary condition with the remaining parameters; wherein the boundary condition is one selected from a maximum value, a minimum value, an average value, a standard deviation value, which are obtained using the sensor parameters according to the at least two different process items, and a statistical value obtained using the maximum value, the minimum value, the average value, and/or the standard deviation value; wherein when the remaining sensor parameters satisfy the boundary condition, the process prediction unit replaces a previously fabricated one of the selected semiconductor structures with a previously fabricated one of the remaining semiconductor structures to generate a first specific predicted parameter, replaces a subsequently fabricated one of the selected semiconductor structures with a subsequently fabricated one of the remaining semiconductor structures to generate a second specific predicted parameter, and repetitively performs the same process as the process of generating each of the first and second specific predicted parameters in order to generate the predicted parameters, and wherein the boundary condition is repetitively updated based on the process-predicting model using the selected sensor parameters and the remaining sensor parameters.
-
-
2. A process-parameter prognostic system comprising:
-
a process prediction unit configured to calculate predicted parameters using measured parameters of semiconductor structures and sensor parameters associated with plasmas in a process chamber used to form the semiconductor structures, wherein the process prediction unit is further configured to provide a process-predicting model based on selected measured parameters and selected predicted parameters of selected semiconductor structures and configured to repetitively update the process-predicting model based on the sensor parameters by adjusting a number of the selected semiconductor structures and a number of remaining semiconductor structures in an order of a semiconductor fabrication process to provide remaining predicted parameters corresponding to the process-predicting model; a process-change point corresponding unit configured to compare selected sensor parameters corresponding to the selected predicted parameters with the remaining sensor parameters corresponding to the remaining predicted parameters to confirm process environments in the process chamber; wherein each of the semiconductor structures is one selected from a structure on a semiconductor substrate and a structure including a material layer on and different from the semiconductor substrate, the measured parameters comprise actual numerical figures for limiting the shape of the semiconductor structures, and the predicted parameters are expected numerical figures for estimating the shape of the semiconductor structures; wherein the sensor parameters are classified to correspond to at least two different process items, and the at least two different process items correspond to the semiconductor structures, respectively, and correspond to independent variables of the process-predicting model, respectively; wherein each of the sensor parameters comprises an average radio frequency of a waveform obtained between specific points in time during performance of the semiconductor fabrication process; wherein the process-predicting model is linear between the selected measured parameters and the selected predicted parameters; wherein the process prediction unit is configured to determine a boundary condition for the process-predicting model using the selected sensor parameters and is configured to compare the boundary condition with the remaining parameters; wherein the boundary condition is one selected from a maximum value, a minimum value, an average value, a standard deviation value, which are obtained using the sensor parameters according to the at least two different process items, and a statistical value obtained using the maximum value, the minimum value, the average value, and/or the standard deviation value; wherein when sensor parameters corresponding to the at least two different process items of the remaining sensor parameters do not satisfy the boundary condition, the process prediction unit transmits the sensor parameters corresponding to the at least two different process items to the process-change point corresponding unit, and the process-change point corresponding unit is configured to confirm the process environments in the process chamber based on the sensor parameters corresponding to the at least two different process items and the selected sensor parameters to select one of a management or a removal of the sensor parameters corresponding to the at least two different process items. - View Dependent Claims (3)
-
-
4. A semiconductor fabrication apparatus having a process-parameter prognostic system, the apparatus comprising:
-
a process chamber configured to generate plasmas between semiconductor structures and a process gas; a sensor disposed on a sidewall of the process chamber and configured to sample sensor parameters associated with the plasmas; a memory module configured to store measured parameters of the semiconductor structures removed from the process chamber; and the process-parameter prognostic system electrically connected to the sensor and the memory module and including a process prediction unit and a process-change point corresponding unit configured to provide predicted parameters using the process prediction unit and the process-change point corresponding unit, wherein the process prediction unit provides a process-predicting model based on selected measured parameters and selected predicted parameters of selected semiconductor structures and repetitively updates the process-predicting model based on the sensor parameters by adjusting a number of the selected semiconductor structures and a number of the remaining semiconductor structures in an order of a semiconductor fabrication process to obtain a remaining predicted parameters corresponding to the process-predicting model, and the process-change point corresponding unit is further configured to compare selected parameters corresponding to the selected predicted parameters with the remaining sensor parameters corresponding to the remaining predicted parameters and configured to confirm a process-change point in the process chamber wherein when the remaining sensor parameters satisfy a boundary condition, the process prediction unit repetitively performs the following two operations to generate the predicted parameters; (1) replaces a previously fabricated one of the selected semiconductor structures with a previously fabricated one of the remaining semiconductor structures to generate a first specific predicted parameter; (2) replaces a subsequently fabricated one of the selected semiconductor structures with a subsequently fabricated one of the remaining semiconductor structures to generate a second specific predicted parameter; and wherein the boundary condition is repetitively updated based on the process-predicting model using the selected sensor parameters and the remaining sensor parameters. - View Dependent Claims (5)
-
Specification