Adaptive virtual metrology for semiconductor chemical mechanical planarization process using GMDH-type polynomial neural networks

Jia, Xiaodong; Di, Yuan; Feng, Jianshe; Yang, Qibo; Dai, Honghao; Lee, Jay

doi:10.1016/j.jprocont.2017.12.004

Cited by 72 publications

(24 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Experimental results have shown that an average root-meansquare error (RMSE) of 2.7 can be achieved. Jia et al [18] proposed an adaptive method based on polynomial neural networks to predict the MRR for CMP. Experimental results have shown that the proposed method outperforms the k-nearest neighbors (KNN), logistic regression (LR), support vector regression (SVR), and random forests (RF) in terms of mean squared error (MSE) and coefficient of determination (R 2 ).…”

Section: Related Workmentioning

confidence: 99%

“…The validation and test datasets include 144,148 and 156,262 trajectories, respectively. It should be noted that four outliers in the training dataset under stage A were removed before processing the data [18,30].…”

Section: Data Descriptionmentioning

confidence: 99%

“…Prediction accuracy is measured using R-square (R 2 ), RMSE, relative error (RE), and score function (S-score) (see Eqs. (15)- (18) Pressure applied to the center air bag…”

Section: Data Descriptionmentioning

confidence: 99%

See 2 more Smart Citations

Prediction of Material Removal Rate for Chemical Mechanical Planarization Using Decision Tree-Based Ensemble Learning

2019

Journal of Manufacturing Science and Engineering

View full text Add to dashboard Cite

Chemical mechanical planarization (CMP) has been widely used in the semiconductor industry to create planar surfaces with a combination of chemical and mechanical forces. A CMP process is very complex because several chemical and mechanical phenomena (e.g., surface kinetics, electrochemical interfaces, contact mechanics, stress mechanics, hydrodynamics, and tribochemistry) are involved. Predicting the material removal rate (MRR) in a CMP process with sufficient accuracy is essential to achieving uniform surface finish. While physics-based methods have been introduced to predict MRRs, little research has been reported on monitoring and predictive modeling of the MRR in CMP. This paper presents a novel decision tree-based ensemble learning algorithm that can train the predictive model of the MRR. The stacking technique is used to combine three decision tree-based learning algorithms, including the random forests (RF), gradient boosting trees (GBT), and extremely randomized trees (ERT), via a meta-regressor. The proposed method is demonstrated on the data collected from a CMP tool that removes material from the surface of wafers. Experimental results have shown that the decision tree-based ensemble learning algorithm using stacking can predict the MRR in the CMP process with very high accuracy.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Data Descriptionmentioning

confidence: 99%

See 1 more Smart Citation

Prediction of Material Removal Rate for Chemical Mechanical Planarization Using Decision Tree-Based Ensemble Learning

2019

Journal of Manufacturing Science and Engineering

View full text Add to dashboard Cite

show abstract

“…Jia et al utilized the GMDH, a training algorithm for Polynomial Neural Network based model development for feature selection and model complexity to estimate virtual metrology in semiconductors. GMDH was also used for training the neural network model for estimation of the impact of climate change on financial analysis of a small hydro power project etc.…”

Section: Literature Reviewmentioning

confidence: 99%

Real time reliability monitoring of hydro‐power plant by combined cognitive decision‐making technique

Majumder

Saha

et al. 2019

Int J Energy Res

View full text Add to dashboard Cite

Summary Hydro power plants are one of the most cost‐effective renewable technologies which have the potential to replace the conventional fuels. However, reliability of the plant varies with time due to irregularity in flow pattern of the river which depends on the pattern of climatic parameters like precipitation and evapo‐transpiration. The reliability of the plant performance also varies with the efficiency of the power equipment and conveyance coefficient of penstock, both of which are time dependent. When reliability reduces, performance efficiency also gets degraded. That is why, monitoring of reliability in hydro power plants is essential and required to be conducted periodically. But at present, such activities are expensive due to the requirement of extensive surveys, data collection and performance audits of the equipment and structures involved in the power production. There is a lack of single indicator based real time monitoring processes which can be utilized to continuously regulate the reliability of the power plant. Such system can reduce both cost and time required for conducting reliability assessment procedures. That is why, in the present study, an objective and cognitive technique was utilized to develop a single index for real time monitoring of the reliability of a power plant. The results indicate that efficiency of generator, levelized cost of electricity and rate of selling per unit electricity are the most significant indicator which have the capability to represent reliability. A model developed for the present study by Polynomial Neural Network also ensures that regular monitoring of reliability of an HPP is possible.

show abstract

“…Data-driven SS are common in industrial environments such as semiconductor manufacturing [7,8], chemical [9][10][11], and automotive [12,13]. The methodologies employed in the literature usually vary from simple regression/classification techniques such as linear regression [14] and Bayesian Networks [11] to more complex neural-network-based algorithms [7].…”

Section: Introductionmentioning

confidence: 99%

Laundry Fabric Classification in Vertical Axis Washing Machines Using Data-Driven Soft Sensors

et al. 2019

View full text Add to dashboard Cite

Embedding household appliances with smart capabilities is becoming common practice among major fabric-care producers that seek competitiveness on the market by providing more efficient and easy-to-use products. In Vertical Axis Washing Machines (VA-WM), knowing the laundry composition is fundamental to setting the washing cycle properly with positive impact both on energy/water consumption and on washing performance. An indication of the load typology composition (cotton, silk, etc.) is typically provided by the user through a physical selector that, unfortunately, is often placed by the user on the most general setting due to the discomfort of manually changing configurations. An automated mechanism to determine such key information would thus provide increased user experience, better washing performance, and reduced consumption; for this reason, we present here a data-driven soft sensor that exploits physical measurements already available on board a commercial VA-WM to provide an estimate of the load typology through a machine-learning-based statistical model of the process. The proposed method is able to work in a resource-constrained environment such as the firmware of a VA-WM.

show abstract

Adaptive virtual metrology for semiconductor chemical mechanical planarization process using GMDH-type polynomial neural networks

Cited by 72 publications

References 38 publications

Prediction of Material Removal Rate for Chemical Mechanical Planarization Using Decision Tree-Based Ensemble Learning

Prediction of Material Removal Rate for Chemical Mechanical Planarization Using Decision Tree-Based Ensemble Learning

Real time reliability monitoring of hydro‐power plant by combined cognitive decision‐making technique

Laundry Fabric Classification in Vertical Axis Washing Machines Using Data-Driven Soft Sensors

Contact Info

Product

Resources

About