US 5,859,437 AGranted

Intelligent supervision system with expert system for ion implantation process

AssigneeTaiwan Semiconductor Manufacturing Company, Ltd.

InventorsPau-Lo Hsu, Li-Cheng Shen, Chin-Shien Yang, Chi-Shun Hou

Life of this patent

Estimating the term…

Loading patent…

Abstract

An intelligent supervision system on the ion beam map during ion implantation process includes an ion implanter for introducing ions into a wafer, a feature extractor for receiving a beam current signal and a display blanking signal from the ion implanter to extract features of the beam current signal; a fuzzification subsystem for justifying confidence level on the features; and an expert system for recognizing symptoms of abnormal scanning and providing an indication thereof to a human operator. The expert system also provides an indication of appropriate corrective action, which an operator then uses to adjust the implantation process. Alternatively, the expert system may be configured to provide control signals directly to the ion implanter to adjust the implantation process without the need for an operator.

Claims18

Dependent claims

An ion implantation system with an automatic supervision system for controlling implantation of ions in a semiconductor wafer, the ion implantation system being configured to generate a beam map during ion implantation, said beam map comprising a plurality of frames each having an outmost edge and one or more minors, said outmost edge and said minors each having a width, wherein the width of said outmost edge is larger than the width of said minor, the ion implantation system comprising:

implantation means for implanting ions in the semiconductor wafer by repeated scanning of a beam of ions across the wafer;
feature extraction means for extracting information from the beam map generated by the ion implantation system, wherein for each frame of said beam map the information extracted by said feature extraction means comprises magnitudes of:
a first left segment spanning from a beginning point of said outmost edge to a first point at which said outmost edge equals a predetermined threshold;
a first center segment spanning from said first point to a second point at which said outmost edge equals the predetermined threshold;
a first right segment spanning from the second point to an ending point of said outmost edge;
a second left segment spanning from a beginning point of a minor to a third point at which said minor equals the predetermined threshold;
a second center segment spanning from the third point to a fourth point at which said minor equals the predetermined threshold;
and
a second right segment spanning from the fourth point to an ending point of said minor;
fuzzification means for determining a confidence level of the information extracted by said feature extraction means;
and
expert means for receiving the information from said feature extraction means and the determined confidence level from said fuzzification means and in response thereto recognizing abnormal ion beam scanning and providing adjustment information for adjusting the implantation means to correct the abnormal ion beam scanning.

9 more elements

an implanter configured to implant ions in the semiconductor wafer by repeated scanning of a beam of ions across the wafer;
a feature extractor coupled to the implanter, wherein the feature extractor is configured to extract information from the beam map generated by the ion implantation system, wherein said feature extractor comprises;
(a) a visual signal converter coupled to said implanter, said visual signal converter being configured to extract features of the outmost edge and the minors;
(b) a data converter coupled to said visual signal converter, said data converter being configured to generate mean, variance and combinations thereof from the features of the outmost edge and the minors;
and
(c) data examiner coupled to the implanter, wherein the data examiner is configured to separate information corresponding to the outmost edge from information corresponding to the minors, wherein for each frame of said beam map said information extracted by said feature extractor comprises magnitudes of:
a first left segment spanning from a beginning point of said outmost edge to a first point at which said outmost edge equals a predetermined threshold;
a first center segment spanning from said first point to a second point at which said outmost edge equals the predetermined threshold;
a first right segment spanning from the second point to an ending point of said outmost edge;
a second left segment spanning from a beginning point of a minor to a third point at which said minor equals the predetermined threshold;
a second center segment spanning from the third point to a fourth point at which said minor equals the predetermined threshold;
and
a second right segment spanning from the fourth point to an ending point of said minor;
a fuzzification subsystem coupled to the feature extractor, wherein the fuzzification subsystem is configured to determine a confidence level of the information extracted by said feature extractor;
and
an expert subsystem coupled to the fuzzification subsystem, wherein the expert subsystem is configured to receive the information from said feature extractor and the determined confidence level and in response thereto recognize abnormal ion beam scanning and provide adjustment information for adjusting the implanter to correct the abnormal ion beam scanning, wherein said expert subsystem comprises:
(a) an inference engine coupled to said data examining means, said inference engine being configured to provide a model for recognizing symptoms of abnormal scanning;
and
(b) a rule base coupled to said inference engine, said rule base being configured to provide a set of rules for adjusting said implantation means to correct the abnormal beam scanning.

16 more elements

Dependencies are inferred from the claim text.