Tokyo Electron Limited

Megan Wooley

Jun Shinagawa

Toshihiro Kitao

Alok Ranjan

Carlos A Fonseca

Nathan Ip

Atsushi Suzuki

Aspects of the disclosure provide a method for wafer result prediction. The method includes determining predictor parameters of a semiconductor process using domain knowledge that includes knowledge of the semiconductor process, a processing tool associated with the semiconductor process, a metrology tool, and/or the wafer. The method also includes removing collinearity among the predictor parameters to obtain key predictor parameters, and selecting a subset of the key predictor parameters based on metrology data of the wafer obtained from the metrology tool. The method further includes building a virtual metrology (VM) model on the subset of the key predictor parameters and may include predicting wafer results using the VM model.

Virtual metrology for wafer result prediction

Sensitivity calculations are provided of a process model through the rate of change of a model fingerprint with respect to process variables and defects. A fingerprint sensitivity table is generated, where process variables are associated with a set of fingerprint sensitivities. The fingerprint of incoming substrates is monitored through a production process by applying the same fingerprint method that is used in the process model. Calculations are made of the difference between the incoming substrate fingerprint and the process model predicted fingerprint. This difference fingerprint is compared against the table of fingerprint sensitivities to find the process variable most likely to be responsible for the difference. Spatial relationships between process variables and actual measurements on the substrate may be obtained. Correlation through fingerprint sensitivity improves the ability to pinpoint faulty process tools. The difference fingerprint may also identify the formation of defects on a substrate.

Method and process using fingerprint based semiconductor manufacturing process fault detection

A method of optimizing a recipe for a plasma process includes (a) building a virtual metrology (VM) model that predicts a wafer characteristic resulting from the plasma process based on a plasma parameter and (b) building a control model that describes a relationship between the plasma parameter and a recipe parameter. (c) The wafer characteristic is measured after performing the plasma process according to the recipe. (d) Whether the wafer characteristic is within a predetermined range is determined. (e) The VM model and the control model are calibrated based on the wafer characteristic. (f) The recipe is optimized by updating the plasma parameter based on the wafer characteristic using the VM model and updating the recipe parameter based on the plasma parameter using the control model. (c), (d), (e) and (f) are repeated until the wafer characteristic is within the predetermined range.

Growing community of inventors

Austin, TX, United States of America

Megan Wooley