Chuck Temperature Control Using Learned Pattern Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing temperature control methods for semiconductor chip inspection face challenges in achieving high accuracy and efficiency due to design limitations, wafer type variations, and manual parameter adjustments, leading to time delays, hunting in control amounts, and decreased measurement throughput and accuracy.
Innovation Solution
A temperature control device and method that utilize a learned model to automatically adjust chuck temperature control parameters based on acquired chuck temperature data, classifying patterns and deriving appropriate heating and cooling control parameters using a correspondence relationship between features and patterns of chuck temperature changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If manual temperature control parameter adjustments are used, then flexibility in handling different wafer types is improved, but time delays and hunting in control amounts occur, reducing measurement throughput and accuracy
Solution Approach 1:
The system performs preliminary classification of temperature change patterns using a learned model before actual temperature control. By pre-analyzing the chuck temperature data and categorizing it into specific patterns (e.g., rapid increase, gradual decrease), the system prepares appropriate control parameters in advance, eliminating manual adjustment delays and enabling immediate automated response to different wafer types.
Solution Approach 2:
The system continuously monitors chuck temperature changes and uses a learned model to classify the temperature change pattern in real-time. Based on the classified pattern, the temperature control parameter is automatically adjusted and fed back to the temperature adjusting unit. This closed-loop feedback mechanism eliminates hunting by providing stable, data-driven control parameters that adapt to actual temperature behavior without manual intervention.
2Adaptability or versatility
If manual temperature control parameter adjustments are used, then flexibility in handling different wafer types is improved, but time delays occur, reducing measurement accuracy
Solution Approach 1:
The system performs preliminary classification of temperature change patterns using a learned model before actual temperature control. By pre-analyzing the chuck temperature data and categorizing it into specific patterns (e.g., rapid increase, gradual decrease), the system prepares appropriate control parameters in advance, eliminating manual adjustment delays and enabling immediate automated response to different wafer types.
Solution Approach 2:
The system continuously monitors chuck temperature changes and uses a learned model to classify the temperature change pattern in real-time. Based on the classified pattern, the temperature control parameter is automatically adjusted and fed back to the temperature adjusting unit. This closed-loop feedback mechanism eliminates hunting by providing stable, data-driven control parameters that adapt to actual temperature behavior without manual intervention.
3Productivity
If learned model-based automatic parameter adjustment is used, then measurement throughput and accuracy are improved, but device complexity increases
Solution Approach 1:
The system introduces a learned model as an intermediary between temperature measurement and control parameter adjustment. This learned model acts as a intelligent mediator that automatically classifies temperature change patterns and determines appropriate control parameters, replacing complex manual adjustment procedures with a data-driven automated system that simplifies the overall control logic while improving performance.
Solution Approach 2:
The temperature control system performs self-adjustment by automatically classifying its own temperature change patterns using the learned model and selecting appropriate control parameters without external intervention. The system serves itself by monitoring its own state, making intelligent decisions based on the classified patterns, and adjusting its control parameters autonomously, thereby reducing the need for complex external control mechanisms.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables precise and efficient temperature control, reducing time delays and hunting, improving measurement throughput and accuracy, and facilitating automatic parameter adjustments without manual intervention.
Implementation Method 1
The wafer chuck of the prober includes, for example, a temperature adjusting device including a heater mechanism, a chiller mechanism, a heat pump mechanism, and the like, and heats or cools the wafer held on the wafer chuck using the temperature adjusting device.
Implementation Method 2
a chuck temperature acquiring unit configured to acquire chuck temperature indicating a temperature of a wafer chuck
Data Source
AI summary
A temperature control device, a temperature control method, a program, a prober, and a learning model generating method that can implement automatic adjustment of a chuck temperature control parameter are provided. The temperature control device includes a chuck temperature acquiring unit that acquires a chuck temperature, a classifying unit which outputs a temperature change pattern in a case where the chuck temperature is input, using a learned model generated through learning with a correspondence relationship between features of changes in the chuck temperature and temperature change patterns, and a temperature control parameter setting unit that derives a temperature control parameter corresponding to the temperature change pattern output from the classifying unit and sets the temperature control parameter, and operation of a chuck temperature adjusting unit that adjusts the chuck temperature is controlled using the temperature control parameter.