Device monitoring method during semiconductor process

Abstract

The invention discloses a device monitoring method during a semiconductor process. The method comprises the following steps that: step a, a sampling scheme with the fixed sampling number is formulated before starting of a semiconductor process; step b, before starting of the technology implementation, it is determined which wafers need to be sampled and which wafers do not need to be sampled according to the sampling scheme; and the wafers that need to be sampled are equally distributed to each technological device; step c, the technological processes are executed; step d, sampling is carried out according to the sampling scheme; according to the result of the sampling, on-line detection is carried out on the sampled wafers; step e, the steps from b to d are repeated until all the technological processes are completed; and step f, electrical performance detection is carried out on all the wafers. According to the invention, the beneficial effect of the method is as follows: a sampling scheme and a dynamic risk sign are cooperated, so that a potential risk is minimized during the semiconductor process.

Description

technical field [0001] The invention designs a monitoring method in the field of semiconductor manufacturing, in particular, an equipment monitoring method in a semiconductor manufacturing process that can improve product yield. Background technique [0002] The manufacture of semiconductor integrated circuits involves hundreds of process steps, and any one of these steps out of control may cause the final product to fail even if the other steps are well controlled. Therefore, monitoring each process step and having early warning is one of the key issues in the semiconductor integration process. [0003] Traditional semiconductor factories use two forms of monitoring, offline monitoring and online monitoring, to check the performance of equipment during operation. However, both of the above monitoring forms have limitations and risk periods, which bring uncertainty to production. Offline monitoring can only be carried out when the equipment is not in production, so that th...

AI Technical Summary

Problems solved by technology

However, both of the above forms of monitoring have limitations and risk periods, thus creating uncertainty in production

Offline monitoring can only be carried out when the equipment is not in production, so that the monitoring frequency is low, and the uncertainty in production cannot be monitored in time

Online monitoring is a part of the whole process, but it is usually impossible to inspect all wafers. After the products arrive at the online inspection station, they are selected according to the defined sampling rules. Products that do not meet the sampling rules are automatically skipped to the next station without making defects. Inspection, and the selection of running equipment after the product arrives at the main process site is random, so it may cause that the products released by a certain main process equipment within a period of time have not been inspected for defects. Time out of monitoring and monitoring, so that the risk expands

The actual practice of traditional semiconductor factories is the above-mentioned sampling inspection, which leads them to bear the risk of a large number of wafers being skipped in production

Images