Semiconductor device manufacturing: process – Including control responsive to sensed condition – Optical characteristic sensed
Reexamination Certificate
2001-02-07
2003-09-30
Kunemund, Robert (Department: 1765)
Semiconductor device manufacturing: process
Including control responsive to sensed condition
Optical characteristic sensed
C438S009000, C438S710000, C703S002000, C703S010000
Reexamination Certificate
active
06627464
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to semiconductor plasma etching. More particularly, the invention relates to a method and system for characterizing a semiconductor plasma process using fuzzy logic and neural networks.
BACKGROUND AND SUMMARY OF THE INVENTION
In the semiconductor industry, plasma etching has become an integral part of the manufacture of semiconductor circuits. For example, plasma etchers are frequently used in semiconductor processing when a relatively straight vertical edge is needed. For instance, when etching the polysilicon gate of a MOS transistor, undercutting the polysilicon can adversely affect the operation of the transistor. Undercutting is frequently encountered when etching is performed using a liquid etching method. Plasma etching, which uses ions accelerated by an electric field, tends to etch only horizontal exposed surfaces and therefore avoids undercutting.
An important aspect of all plasma processes is stopping the plasma process after the layer being etched has been removed but before the next layer down is destroyed. This is often called “end point” detection—for detecting the completion of etching of a particular layer. Determining the etch rate is often critical in end point detection. Another important aspect of plasma processes is chamber maintenance. Chamber maintenance involves tracking the condition of various equipment used in the plasma process to determine when cleaning, repair, or other changes need to be made.
Conventional approaches to end point detection and etch rate determination involve the use of devices such as residual gas analyzers, crystal peering scopes (visual inspection), and scanning spectrophotometers. In the case of scanning spectrophotometers, the process engineer can visually inspect the plasma for color changes which provide an indication of the etch rate. When the derivation of the etch rate becomes zero, the end point can be inferred. Residual gas analyzers operate similarly by monitoring the off gas content of the plasma process. All of the above approaches allow process engineers to monitor “outputs” of the plasma process in order to control the process.
In recent years, the monitoring of the radio frequency (RF) power (i.e., the inputs) delivered to the plasma chamber has become a valuable technique in end point detection and etch rate determination. For example, by monitoring the variables associated with the electrical power used to control the plasma chamber, a number of process-related determinations can be made.
Although a number of data collection techniques using probes such as impedance analyzers have been developed, room for improvement in the characterization still remains. For example, while process engineers typically characterize the plasma process based on the above variables, the result is both labor intensive and inaccurate. With regard to end point determination, engineers typically use the fact that certain frequencies of light transmission can be used in determining whether the etching end point has been reached. The selection of these frequencies, however, is done on a trial and error basis. The result is a relatively expensive process characterization approach that is latent with inaccuracies. Once the process has been characterized, output parameter values are predicted based on subsequent input values. Thus, if the RF input suggests a certain output value based on collected training data, that output parameter value is somewhat predictable under conventional approaches.
While the above-described conventional process characterization techniques are somewhat useful in certain instances (despite their labor intensive aspect), it is important to note that other difficulties still remain. One particular difficulty relates to the fact that the input parameters to a plasma process are not “cut and dry” in relation to the output of the plasma process. Thus, while the process engineer may be able to glean from the plasma process a few of the specifics necessary to truly control the process, the conventional approach fails to address the subtle nuances of the process. It is therefore desirable to provide a method for characterizing a semiconductor plasma process that provides “membership functions” that enable the estimation of output parameter values and are not subject to the above-described shortcomings.
The above and other objectives are provided by a method for characterizing a semiconductor plasma process in accordance with the present invention. The method includes the step of collecting training data, where the training data is based on variables associated with electrical power used to control a plasma etching chamber and resulting from execution of the plasma process. The method further includes the step of generating fuzzy logic-based input and output membership functions based on the training data. The membership functions enable estimation of an output parameter value of the plasma process such the membership functions characterize the plasma process with regard to the output parameter. Using fuzzy logic accounts for the inherent inaccuracies in the training data. The result is a characterization of the plasma process that is more reliable than conventional approaches.
Further in accordance with the present embodiment, a method for estimating an output parameter value of a semiconductor plasma process is provided. The method includes the step of collecting input data, where the input data is based on variables associated with the electrical power used to control a plasma etching chamber and corresponding to execution of the plasma process. The method further provides for estimating the output parameter value based on fuzzy logic-based input and output membership functions. The method also provides for modifying the membership functions based on a neural network learning algorithm and output data. The output data defines an actual parameter value resulting from execution of the plasma process.
In another aspect of the invention, an adaptive plasma characterization system has an impedance analyzer, a fuzzy inference system, and a neural network. The impedance analyzer is connected to a plasma etching chamber, wherein the analyzer collects data resulting from execution of a plasma process. The data is based on variables associated with electrical power used to control the plasma etching chamber. The algorithm generates the fuzzy-logic based input and output membership functions and associated fuzzy rules set of the fuzzy inference system based on the data, where the fuzzy inference system enables autonomous estimation of the output parameters of the plasma process. A neural network modifies the membership functions based on a neural network learning algorithm and output data. The output data defines an actual parameter value resulting from execution of the plasma process.
It is to be understood that both the foregoing general description and the following detailed description are merely exemplary of the invention, and are intended to provide an overview or framework for understanding the nature and character of the invention as it is claimed. The accompanying drawings are included to provide a further understanding of the invention, and are incorporated in and constitute part of this specification. The drawings illustrate various features and embodiments of the invention, and together with the description serve to explain the principles and operation of the invention.
Further areas of applicability of the present invention will become apparent from the detailed description provided hereinafter. It should be understood that the detailed description and specific examples, while indicating the preferred embodiment of the invention, are intended for purposes of illustration only and are not intended to limited the scope of the invention.
REFERENCES:
patent: 4846920 (1989-07-01), Keller et al.
patent: 4954212 (1990-09-01), Gabriel et al.
patent: 5135604 (1992-08-01), Kumar et al.
patent: 5175472 (1992-12-01), Johnson et al.
patent: 5576629 (1996-11-01), Turner et al
ENI Technology Inc.
Harness & Dickey & Pierce P.L.C.
Kunemund Robert
LandOfFree
Adaptive plasma characterization system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Adaptive plasma characterization system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Adaptive plasma characterization system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3097629