Data processing: measuring – calibrating – or testing – Measurement system – Remote supervisory monitoring
Reexamination Certificate
2002-01-25
2004-03-09
Nghiem, Michael (Department: 2863)
Data processing: measuring, calibrating, or testing
Measurement system
Remote supervisory monitoring
C438S710000, C702S085000, C702S121000, C702S182000
Reexamination Certificate
active
06704691
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to in-line monitoring of semiconductor quality, and more particularly to a method and system for in-line monitoring semiconductor quality using measurable equipment signals.
BACKGROUND OF THE INVENTION
Process performance in semiconductor fabrication can be monitored by in-line analysis. In-line analysis is usually performed with analytical tools after certain manufacturing steps or processes have been completed to determine the quality of the wafer. However, in-line monitoring can become a bottleneck in the production flow because the so many wafers are produced in one lot. There are always more fabrication tools than analytical tools in semiconductor plants; thus, examining every wafer in the entire lot will take a long time and delay the production schedule. The shorten the in-line examining time, conventionally, only a few wafers are checked by in-line tools and the examining results are considered sufficient to represent the entire lot. The in-line analysis tools can be destructive, such as Secondary Ion Mass Spectrosmetry (SIMS) for depth profiling or non-destructive, such as In-line CD-SEM for measuring the size of critical dimension, profiler for measuring etch depth or KLA scan for checking particle defect on the wafers.
One disadvantage of the conventional in-line monitoring is that the quality of the whole lot is statistically estimated from the test results of the samples, when actual results can be diverse. There is uncertainty of the exact quality of the un-sampled wafers. For high-priced IC products, the low quality of the un-sampled products will cause a great loss.
Another disadvantage is time delay. Although the in-line analytical instruments are designed for rapidly analysis, it still takes time to examine the samples and the lot is held back. Otherwise, the lot may continue in the production flow while the samples are being analyzed, but if the outcome are unsatisfactory, it is usually too late to stop the lot and to make proper remedies. The lot may be discarded, causing a great loss for the semiconductor manufacturer and a time delay to reproduce the lot.
Therefore, there is a need for a method to in-line monitoring the process performance of semiconductor fabrication in a quick and efficient way.
SUMMARY OF THE INVENTION
To solve the problems mentioned above, one object of the present invention is to provide a method and system achieving rapid in-line monitoring of process performance during semiconductor fabrication.
Another object of the present invention is to provide a method and system for in-line monitoring of process performance during semiconductor fabrication wherein the quality of every wafer can be tested.
An additional object of the present invention is to provide a method and system for in-line monitoring of process performance during semiconductor fabrication using fabrication equipment signals.
According to the present invention, a method and system for in-line monitoring process performance during wafer fabrication is provided. First, first signals generated by a fabrication tool are collected and filtered by a filtering unit to exclude abnormal signals while a model wafer processes. Then, the filtered model signals are regulated and normalized by a normalizing unit to generate model wafer data. After the fabrication process is completed, the model wafer is measured with an analytical tool to generate a measured value representing the process quality of the model wafer. The model wafer data and the measured value of the model wafer are used by a correlation unit to build a correlation model. After the correlation model is built, the process performance can be predicted as follows. Second signals generated by the fabrication tool are collected when a run wafer is processed, and then filtered to exclude abnormal signals by the filtering unit. Then, the filtered second signals are regulated and normalized by the normalizing unit to generate run wafer data. A predicted value representing the process quality of the run wafer is generated by inputting the run wafer data into the correlation model.
The correlation unit in the present invention comprises a switch for alternately selecting a modeling function or a predictive function. When the modeling function is selected, the correlation model is built by inputting the model wafer data and measured values of wafer processing quality, and when the predictive function is selected, a predicted value of process quality of a run wafer is generated by inputting wafer data of the run wafer into the correlation model.
The wafer data and the measured values of a plurality of model wafers can be inputted into the correlation unit to generate a best fit correlation model for process performance prediction. The data can be accumulated and stored in a database. Further, when the measured value of a run wafer is outside a predetermined value, the measured value of said wafer data can be inputted into the correlation unit to update the correlation model.
REFERENCES:
patent: 5864773 (1999-01-01), Barna et al.
patent: 6446022 (2002-09-01), Coss, Jr. et al.
patent: 6455437 (2002-09-01), Davidow et al.
patent: 6556949 (2003-04-01), Lyon
Le John H
Nghiem Michael
Promos Technologies Inc.
LandOfFree
Method and system for in-line monitoring process performance... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for in-line monitoring process performance..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for in-line monitoring process performance... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3258232