Scalable, hierarchical control for complex processes

Data processing: generic control systems or specific application – Generic control system – apparatus or process – Optimization or adaptive control

Reexamination Certificate

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

C700S121000, C438S005000

Reexamination Certificate

active

06810291

ABSTRACT:

FIELD OF THE INVENTION
The invention relates to the field of data processing and process control. In particular, the invention relates to nonlinear regression prediction and/or control of complex multi-step processes.
BACKGROUND
Process prediction and control is crucial to optimizing the outcome of complex multi-step production processes. For example, the production process for integrated circuits comprises hundreds of process steps (i.e., sub-processes). In turn, each process step may have several controllable parameters, or inputs, that affect the outcome of the process step, subsequent process steps, and/or the process as a whole. In addition, the impact of the controllable parameters on outcome may vary from process run to process run, day to day, or hour to hour. The typical integrated circuit fabrication process thus has a thousand or more controllable inputs, any number of which may be cross-correlated and have a time-varying, nonlinear relationship with the process outcome. As a result, process prediction and control is crucial to optimizing process parameters and to obtaining, or maintaining, acceptable outcomes.
SUMMARY OF THE INVENTION
The present invention provides a method and system for complex process prediction and optimization utilizing sub-process metrics and optimization of the sub-process metrics with respect to a cost function for the process.
In one aspect, the invention comprises: (1) providing a map between the sub-process metrics and process metrics using a nonlinear regression model; (2) providing one or more target process metrics; (3) providing an acceptable range of values for the sub-process metrics to define a sub-process metric constraint set; (4) providing a cost function for the sub-process metrics; and (5) determining values for the sub-process metrics that are within the constraint set, and that produce at the lowest cost a process metric(s) that is as close as possible to the target process metric(s) in order to define target sub-process metrics for each sub-process.
In another aspect of the invention, steps (1) to (5) in the preceding paragraph are applied at different hierarchical process levels. For example these steps may be repeated for sub-sub-processes of one or more sub-processes. (That is, the sub-process becomes the “process” of steps (1) to (5) and the sub-sub-process becomes the “sub-process.”) Similarly, in another aspect of the invention, steps (1) to (5) in the preceding paragraph are repeated for a higher level process composed of two or more processes from step (1). (That is, the process becomes the “sub-process” of the new higher level process.) In effect, the present invention may perform a hierarchical series of steps (1) to (5) that can be scaled to any level of the overall production process.
As used herein, the term “metric” refers to any parameter used to measure the outcome or quality of a process or sub-process. Metrics include parameters determined both in situ during the running of a sub-process or process, and ex situ, at the end of a sub-process or process.
The map between sub-process metrics and process metrics is preferably determined by training a nonlinear regression model against measured sub-process and process metrics. The sub-process metrics from each of the sub-processes serve as the input to a nonlinear regression model, such as a neural network. The output for nonlinear regression model is the process metric(s). The nonlinear regression model is preferably trained by comparing a calculated process metric(s), based on measured sub-process metrics for an actual process run, with the actual process metric(s) as measured for the actual process run. The difference between calculated and measured process metric(s), or the error, is used to compute the corrections to the adjustable parameters in the regression model. If the regression model is a neural network, these adjustable parameters are the connection weights between the layers of the neurons in the network.
In one embodiment, the nonlinear regression model is a neural network. In one version, the neural-network model architecture comprises a two-layer feedforward model with cascade correlation of the single hidden-layer nodes and an adaptive gradient algorithm for back-propagation of prediction errors to adjust network weights. Hidden units are added one at a time (or in vector candidate groups) and trained to maximize the correlation between the hidden unit's outputs and the residual error at the output of the current training process metrics (i.e., the training vector). Previously hidden units are connected or “cascaded” through weights to subsequent units to reduce the residual error not explained by previous hidden nodes.
In another aspect, the present invention comprises: (1) providing a map between the sub-process metrics and process metrics using a nonlinear regression model; (2) providing one or more target process metrics; (3) providing an acceptable range of values for the sub-process metrics to define a sub-process metric constraint set; (4) providing a cost function for the sub-process metrics; (5) determining values for the sub-process metrics that are within the constraint set, and that produce at the lowest cost a process metric(s) that is as close as possible to the target process metric(s) in order to define target sub-process metrics for each sub-process; (6) determining a map between the operational variables of a sub-process and the metrics of the sub-process; (7) providing a cost function for the sub-process operational variables; and (8) determining values for the sub-process operational variables that produce at the lowest cost the sub-process metric, and that are as close as possible to the target sub-process metric values.
As used herein, the term “operational variables” includes sub-process controls that can be manipulated to vary the sub-process procedure, such as set point adjustments (referred to herein as “manipulated variables”), variables that indicate the wear, repair, or replacement status of a sub-process component(s) (referred to herein as “replacement variables”), and variables that indicate the calibration status of the sub-process controls (referred to herein as “calibration variables”). Furthermore, it should be understood that acceptable values of sub-process operational variables include, but are not limited to, continuous values, discrete values and binary values.
For example, where the process comprises plasma etching of silicon wafers, the operational variables for a plasma etch sub-process, such as performed by a LAM 4520 plasma etch tool, may be as follows: manipulated variables (“MV”) may include, e.g., RF power and process gas flow; replacement variables (“RV”) may include, e.g., time since last electrode replacement and/or a binary variable that indicates the need to replace
ot replace the electrodes; and calibration variables (“CalV”) may include, e.g., time since last machine calibration and/or the need for calibration.
In another aspect, the invention comprises: (1) providing a map between the sub-process metrics and sub-process operational variables and the process metrics using a nonlinear regression model; (2) providing one or more target process metrics; (3) providing an acceptable range of values for the sub-process metrics and sub-process operational variables to define a sub-process operational constraint set; (4) providing a cost function for the sub-process metrics and operational variables; and (5) determining values for the sub-process metrics and operational variables that are within the constraint set, and that produce at the lowest cost a process metric(s) that is as close as possible to the a target process metric(s) in order to define target sub-process metrics and target operational variables for each sub-process.
In another aspect, the invention comprises: (1) providing one or more target process metrics; (2) determining values for the sub-process metrics that are within a constraint set, and that produce at the lowest cost a process metric(s) that is as close as possible t

LandOfFree

Say what you really think

Search LandOfFree.com for the USA inventors and patents. Rate them and share your experience with other people.

Rating

Scalable, hierarchical control for complex processes does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Scalable, hierarchical control for complex processes, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Scalable, hierarchical control for complex processes will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFUS-PAI-O-3306460

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.