KR-102963118-B1 - Semiconductor inspection process accident prevention system using AI technology

Abstract

A semiconductor inspection process accident prevention system utilizing AI technology comprises: a data receiving module that receives normal inspection data and defective inspection data for multiple wafers inspected from inspection equipment; an image acquisition module that acquires surface images of the normal wafer in the received normal inspection data and the defective wafer in the defective inspection data, respectively; a condition range calculation module that calculates a normal image condition range and a defective image condition range using an artificial intelligence model on the acquired surface images; a wafer defect judgment module that determines the normality and defect status of multiple wafers using the calculated normal image condition range and defective image condition range; a data comparison analysis module that compares the normal data and defective data of multiple wafers determined by the wafer defect judgment module with the normal inspection data and defective inspection data for multiple wafers inspected from the inspection equipment received by the data receiving module; and a condition range recalculation module that recalculates the normal image condition range and defective image condition range calculated by excluding the data of the wafers determined by the wafer defect judgment module that do not match the result of comparison by the data comparison analysis module.

Inventors

• 심정도

Dates

Publication Date

20260508

Application Date

20250612

Claims (5)

1. In a semiconductor inspection process accident prevention system utilizing AI technology, A data receiving module that receives normal inspection data and defective inspection data for multiple wafers inspected from inspection equipment; An image acquisition module that acquires surface images of a normal wafer of received normal inspection data and a defective wafer of received defective inspection data from multiple directions using multiple scanning electron microscopes; A condition range calculation module that calculates normal image condition ranges and defective image condition ranges using an artificial intelligence model on acquired surface images; A wafer defect determination module that determines the normality and defects of a plurality of wafers using the calculated normal image condition range and the defect image condition range; A data comparison analysis module that compares normal data and defective data of a plurality of wafers determined by the wafer defect determination module with normal inspection data and defective inspection data for a plurality of wafers inspected by inspection equipment received by the data receiving module; A condition range recalculation module that recalculates the normal image condition range and the defective image condition range calculated by excluding the wafer data determined by the wafer defect judgment module that does not match the result of comparison in the data comparison analysis module; A version management storage module that enables real-time storage, real-time updating, and real-time security of the normal image condition range and the defective image condition range recalculated by the above condition range recalculation module; and It includes a monitoring module that monitors conditions including temperature, pressure, vibration, and current in real time for a wafer manufacturing process, an oxidation process, a photolithography process, an etching process, a deposition process, and a metal wiring process for manufacturing the plurality of wafers, and It further includes a control module that controls the data receiving module, the image acquisition module, the condition range calculation module, the wafer defect judgment module, the data comparison analysis module, the condition range recalculation module, the version management storage module, and the monitoring module. The above control module is, It stores the normal image quality range and normal illumination range for the surface images and pattern images of the normal wafers in the normal inspection data and the defective wafers in the defective inspection data, respectively; if it is determined that the image quality and illumination of the surface images and pattern images of the normal wafers in the normal inspection data and the defective wafers in the defective inspection data, respectively, captured by a scanning electron microscope, deviate from the normal image quality range and normal illumination range, it calculates the movement value of the zoom drive unit and the illumination output intensity to ensure that the image quality and illumination of the surface images and pattern images of the normal wafers in the normal inspection data and the defective wafers in the defective inspection data, respectively, become within the normal image quality range and normal illumination range, and controls the zoom drive unit and the illumination unit to achieve the calculated movement value and illumination output intensity of the zoom drive unit; it records multiple conditions including temperature, pressure, vibration, and current for the wafer manufacturing process, oxidation process, photolithography process, etching process, deposition process, and metal wiring process during the manufacturing of each of the multiple wafers, and uses an artificial intelligence model based on the multiple conditions of the normal wafers determined to be normal to [manufacture] the wafer manufacturing process, oxidation process, photolithography process, etching process, deposition process, and metal wiring process of A semiconductor inspection process accident prevention system utilizing AI technology, characterized by calculating a normal condition range including temperature, pressure, vibration, and current, storing surface images and pattern images corresponding to multiple conditions including temperature, pressure, vibration, and current during the manufacturing of each of a plurality of wafers, including a wafer manufacturing process, an oxidation process, a photolithography process, an etching process, a deposition process, and a metal wiring process, analyzing the surface images and pattern images of a wafer determined to be defective to calculate at least one abnormal condition among the wafer manufacturing process, the oxidation process, the photolithography process, the etching process, the deposition process, and the metal wiring process, and transmitting the calculation result to a manufacturing central control panel.
2. delete
3. delete
4. delete
5. delete

Description

Semiconductor inspection process accident prevention system using AI technology The present invention relates to a semiconductor inspection process accident prevention system utilizing AI technology, and more specifically, to a semiconductor inspection process accident prevention system utilizing AI technology to prevent accidents caused by human error during the inspection stage of a semiconductor manufacturing process. The semiconductor manufacturing process involves various steps, such as wafer fabrication, etching, inspection, and packaging. Among these, the process of verifying whether the manufactured semiconductor wafer is defective is called the inspection process. To perform semiconductor inspection, a reference information file called a "recipe" is required. This file is loaded into the inspection equipment to perform the inspection. However, because these recipes are manually created and managed by humans, inspection accidents caused by human error occur frequently, resulting in significant losses for the company. Recipes are files containing inspection standard information that are manually written and entered into inspection equipment. A problem arises when inspections are performed using errors among these manually entered recipes, leading to incorrect inspection results and situations where wafers must be discarded. Along with this, other recurring issues include difficulty in tracking change history, inadequate version control, input of missing conditions, insufficient standard settings, compatibility problems, insufficient reflection of equipment status changes, and inadequate review and approval processes where recipes are used without error detection. FIG. 1 is a control block diagram of a semiconductor inspection process accident prevention system utilizing AI technology according to the present invention. Figure 2 is a control block diagram of an image acquisition module. Figure 3 is a control block diagram of the monitoring module. FIG. 4 is a flowchart of a first embodiment of a semiconductor inspection process using a semiconductor inspection process accident prevention system utilizing AI technology according to the present invention. Figure 5 is a flowchart of a second embodiment of a semiconductor inspection process using a semiconductor inspection process accident prevention system utilizing AI technology. Hereinafter, a semiconductor inspection process accident prevention system (1) utilizing AI technology according to a preferred embodiment of the present invention will be described in detail with reference to the attached drawings. FIG. 1 is a control block diagram of a semiconductor inspection process accident prevention system (1) utilizing AI technology according to the present invention, FIG. 2 is a control block diagram of an image acquisition module (20), and FIG. 3 is a control block diagram of a monitoring module (80). Referring to FIGS. 1 to 3, the configuration of a semiconductor inspection process accident prevention system (1) utilizing AI technology is described. A semiconductor inspection process accident prevention system (1) utilizing AI technology includes a data receiving module (10), an image acquisition module (20), a condition range calculation module (30), a wafer defect judgment module (40), a data comparison analysis module (50), a condition range recalculation module (60), a version management storage module (70), and a monitoring module (80). The data receiving module (10) receives normal inspection data and defective inspection data for a plurality of wafers inspected by inspection equipment. The data receiving module (10) may be composed of a communication unit, and the data receiving module (10) communicates with a plurality of devices. The data receiving module (10) may perform wireless communication, and the wireless communication includes at least one of infrared communication, RF, Zigbee, and Bluetooth. The data receiving module (10) receives a signal and transmits it to the wafer defect judgment module (40) to be described later, and may be implemented in various ways corresponding to the specifications of the received signal and the implementation form of the user terminal. For example, the data receiving module (10) may wirelessly receive an RF (radio frequency) signal transmitted from a broadcasting station (not shown), or may receive a video signal according to composite video, component video, super video, SCART, HDMI (high definition multimedia interface) specifications, etc. via a wired connection. The image acquisition module (20) acquires surface images of the normal wafer of the received normal inspection data and the defective wafer of the received defective inspection data, respectively. The image acquisition module (20) can acquire surface images of the normal wafer of the received normal inspection data and the defective wafer of the received defective inspection data, respectively, from multiple directions. The image acquisition