In Situ Host Principle

Introduction

The acquisition of trace data is not the problem. To be useful for process monitoring, trace data must be reduced to figures of merit that concisely describe tool parameter behavior over a process step (or interval) of interest. Statistical key numbers (maximum, mean, and standard deviation) calculated to represent the behavior of each signal during a process step are used for this purpose. A PVD chamber that was automated will be used as a case study.

The RGA before Automation

The first obstacle is usually a lack of synchronization. Over 95% of RGAs in use are not functionally automated. The user initiates the data collection, and the data is collected continuously at a steady sample rate until a certain amount of time has elapsed, or until the data file reaches a certain size. The analysis treats the data at any given point identically.

With the standard software, alarm thresholds have to be set to allow for normal excursions, such as wafer transfers.

Solutions

The first step is to isolate the analysis to a single wafer, and isolate the data collection from the excursions such as those caused by wafer transfers.

Now the data is much more stable, and reasonable analysis/limits can be applied.

The second obstacle is perspective. The system now generates a single file for every wafer processed. Reviewing the data can become incredibly tedious.

The solution is to collect summary data. In Situ Host collects the average, maximum, and standard deviation for each species and signal, for every wafer, and inputs the data into a summary spreadsheet.

Summary

In Situ Host facilitates the automation of a RGA by:

1. Synchronizes the operation of the RGA to that of the process
2. Collects summary data such that the perspective is enhanced.
3. Allows the creation of signals to further define the information.