Exposure Optics Temperature Control With ML Focus Shift Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing exposure apparatuses face challenges in accurately predicting and correcting optical characteristic variations in optical systems due to temperature changes, which affects productivity and image quality, especially when measuring focus shifts without disrupting exposure processes.
Innovation Solution
An information processing apparatus that utilizes a learning model to predict optical characteristic variations by inputting target temperatures and incorporating temperature adjustment history, structure, and operation information, allowing for high-accuracy corrections without direct measurement during exposure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If temperature measurement elements are provided in optical elements to measure temperature distribution, then temperature change can be detected, but device complexity increases and measurement precision is insufficient when control actions are taken
Solution Approach 1:
The patent introduces a learning model as an intermediary between temperature data and optical characteristic predictions. Instead of directly measuring optical characteristics or relying on simple temperature-optical relationships, the learning model processes temperature distribution data and control action data to predict optical characteristic variations, achieving high measurement precision without adding physical measurement elements to the optical path.
Solution Approach 2:
The patent replaces physical measurement mechanisms (temperature measuring elements in optical elements) with a computational prediction system. The learning model substitutes for direct optical characteristic measurement, using temperature and control data to predict optical variations, thereby reducing device complexity while maintaining or improving measurement precision.
2Measurement precision
If direct measurement of focus shifts is performed to ensure exposure quality, then measurement precision improves, but productivity decreases due to exposure process disruption
Solution Approach 1:
The patent performs preliminary prediction of optical characteristic variations using the learning model before actual exposure occurs. By predicting focus shifts and distortion based on temperature distribution and control action history, the system can pre-compensate for expected variations, ensuring exposure quality without interrupting the exposure process for measurements.
Solution Approach 2:
The patent implements a feedback mechanism where the learning model continuously predicts optical characteristic variations based on real-time temperature data and control actions. These predictions feed back into the exposure control system, allowing dynamic adjustment of exposure parameters to maintain quality while keeping productivity high.
3Temperature
If temperature-adjusted gas is supplied to suppress temperature change, then temperature stability improves, but optical characteristic variations still occur due to control actions
Solution Approach 1:
The patent changes the approach from controlling only temperature to controlling multiple parameters including temperature distribution, control action history, and their interactions. The learning model analyzes these combined parameters to predict optical characteristic variations, recognizing that temperature stability alone is insufficient when control actions like gas supply cause refractive index changes.
Solution Approach 2:
The learning model serves as an intermediary that connects temperature control actions with their unintended optical effects. It processes data about control actions (such as gas supply) and temperature changes to predict resulting optical characteristic variations, enabling compensation for effects that direct temperature control cannot prevent.
Data Source
AI summary
In order to provide an information processing apparatus capable of obtaining a variation amount in an optical characteristic of an optical system provided in an exposure apparatus in consideration of a control for suppressing a temperature variation of the optical system, the information processing apparatus according to the present invention is configured to predict the variation amount in the optical characteristic of the optical system by inputting a target temperature of the optical system to a learning model, in which the learning model is a learning model obtained by machine learning.