Next-generation virtual metrology for semiconductor manufacturing: A feature-based framework

Suthar, Kerul; Shah, Devarshi; Wang, Jin; He, Q. Peter

doi:10.1016/j.compchemeng.2019.05.016

Cited by 21 publications

(8 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For instance, as early as in 1992, they have been explored for parameter estimation in plasma etching based on optical emission spectroscopy (OES) and mass spectroscopy 13 , 14 or global process parameters paired with measured etch characteristics 15 . Plasma virtual metrology (VM) was conducted based on multivariate sensor data, 16 with regard to real-time fault detection in reactive ion etching, 17 batch process characterization in semiconductor fabrication, 18 and a deep learning VM framework based on OES data 19 and “plasma information” descriptors 20 . Further examples have devised an inverse reconstruction of intrinsic plasma properties, such as the electron energy distribution function from OES diagnostics data 21 as well as an active learning guided scheme based on Fourier transform infrared spectroscopy data for parameter space exploration 22 …”

Section: Introductionmentioning

confidence: 99%

Review: Machine learning for advancing low-temperature plasma modeling and simulation

Trieschmann,

Vialetto,

Gergs

2023

J. Micro/Nanopattern. Mats. Metro.

View full text Add to dashboard Cite

Machine learning has had an enormous impact in many scientific disciplines. It has also attracted significant interest in the field of low-temperature plasma (LTP) modeling and simulation in past years. Its application should be carefully assessed in general, but many aspects of plasma modeling and simulation have benefited substantially from recent developments within the field of machine learning and datadriven modeling. In this survey, we approach two main objectives: (a) we review the state-of-the-art, focusing on approaches to LTP modeling and simulation. By dividing our survey into plasma physics, plasma chemistry, plasma-surface interactions, and plasma process control, we aim to extensively discuss relevant examples from literature. (b) We provide a perspective of potential advances to plasma science and technology. We specifically elaborate on advances possibly enabled by adaptation from other scientific disciplines. We argue that not only the known unknowns but also unknown unknowns may be discovered due to the inherent propensity of data-driven methods to spotlight hidden patterns in data.

show abstract

Section: Introductionmentioning

confidence: 99%

Review: Machine learning for advancing low-temperature plasma modeling and simulation

Trieschmann,

Vialetto,

Gergs

2023

J. Micro/Nanopattern. Mats. Metro.

View full text Add to dashboard Cite

show abstract

“…6,7 The soft analyzer, which is also known as "soft sensor" or "virtual sensor", virtually estimates the sulfur content of tail oil by developing a mathematical predictive model integrating process knowledge and operation data and by taking easy-to-measure variables (also called secondary variables, such as flow rate, pressure, and temperature) as inputs. 8,9 Therefore, the soft analyzer is delay-free, easy and cheap to maintain, and it has now been developed as a promising solution to the issues associated with laboratory analysis and hardware analyzer. 10 Up to now, the soft analyzer has been researched in-depth and applied extensively in the hydrocracking process.…”

Section: Introductionmentioning

confidence: 99%

“…The hardware analyzer uses spectrometers for direct measurements, which can shorten the measurement period but suffers from accuracy degradation and high investment and maintenance costs 6,7 . The soft analyzer, which is also known as “soft sensor” or “virtual sensor”, virtually estimates the sulfur content of tail oil by developing a mathematical predictive model integrating process knowledge and operation data and by taking easy‐to‐measure variables (also called secondary variables, such as flow rate, pressure, and temperature) as inputs 8,9 . Therefore, the soft analyzer is delay‐free, easy and cheap to maintain, and it has now been developed as a promising solution to the issues associated with laboratory analysis and hardware analyzer 10 .…”

Section: Introductionmentioning

confidence: 99%

Soft analyzer for sulfur content in tail oil of hydrocracking process based on adaptively regularized dynamic polynomial PLS

Yao,

Shen,

Qiao

et al. 2023

Asia-Pacific J Chem Eng

View full text Add to dashboard Cite

Hydrocracking process is an essential and widely used means to remove impurities such as sulfur, nitrogen, and oxygen from heavy oil in refineries. The sulfur content in the tail oil of the hydrogenation process is a key quality index that must be strictly monitored and controlled. However, the traditional measurement ways of the sulfur content suffer from either large delays or high investment and maintenance costs. To this end, a predictive model named “adaptively regularized dynamic polynomial partial least squares (ARDPPLS)” is proposed to develop soft analyzer for sulfur content estimation. The ARDPPLS takes complicated characteristics of the hydrogenation process into consideration, including process dynamics, nonlinearities, and transportation delays of materials and energies. Specifically, the dynamics and nonlinearities are modeled by the finite impulse response paradigm and polynomial fitting, respectively, and the delays of explanatory variables are automatically determined using differential evolution. In particular, a Bayesian inference‐based adaptive regularization scheme for the polynomial partial least squares is designed to tackle overfitting that results from high‐order variable augmentation and high‐order polynomials. The performance of the ARDPPLS is evaluated by a real‐world industrial wax oil hydrocracking process, showing the effectiveness of the ARDPPLS.

show abstract

“…The first category (see Refs. [6][7][8] for some examples of specific applications in semiconductor manufacturing) uses historical datasets, including: process parameters, sensor data, real metrology measurements, etc. (Figure 1), to infer by means of statistical methods (and, recently, machine learning techniques) the process outcomes in terms of modification of the incoming system [6][7][8][9][10].…”

Section: Introductionmentioning

confidence: 99%

“…[6][7][8] for some examples of specific applications in semiconductor manufacturing) uses historical datasets, including: process parameters, sensor data, real metrology measurements, etc. (Figure 1), to infer by means of statistical methods (and, recently, machine learning techniques) the process outcomes in terms of modification of the incoming system [6][7][8][9][10]. VM has no intrinsic predictivity, and the success of its predictions relies on the quality of the training dataset and on the alignment between the current process conditions with the ones generating the previous process data.…”

Section: Introductionmentioning

confidence: 99%

Multiscale Simulations for Defect-Controlled Processing of Group IV Materials

et al. 2022

View full text Add to dashboard Cite

Multiscale approaches for the simulation of materials processing are becoming essential to the industrialization of future nanotechnologies, as they allow for a reduction in production costs and an enhancement of devices and applications. Their integration as modules of “digital twins”, i.e., a combined sequence of predictive chemical–physical simulations and trained black-box techniques, should ideally complement the real sequence of processes throughout all development and production stages, starting from the growth of materials, their functional manipulation and finally their integration in nano-devices. To achieve this framework, computational implementations at different space and time scales are necessary, ranging from the atomistic to the macro-scale. In this paper, we propose a general paradigm for the industrially driven computational modeling of materials by deploying a multiscale methodology based on physical–chemical simulations bridging macro, meso and atomic scale. We demonstrate its general applicability by studying two completely different processing examples, i.e., the growth of group IV crystals through physical vapor deposition and their thermal treatment through pulsed laser annealing. We indicate the suitable formalisms, as well as the advantages and critical issues associated with each scale, and show how numerical methods for the solution of the models could be coupled to achieve a complete and effective virtualization of the process. By connecting the process parameters to atomic scale modifications such as lattice defects or faceting, we highlight how a digital twin module can gain intrinsic predictivity far from the pre-assessed training conditions of black-box “Virtual Metrology” techniques.

show abstract

Next-generation virtual metrology for semiconductor manufacturing: A feature-based framework

Cited by 21 publications

References 12 publications

Review: Machine learning for advancing low-temperature plasma modeling and simulation

Review: Machine learning for advancing low-temperature plasma modeling and simulation

Soft analyzer for sulfur content in tail oil of hydrocracking process based on adaptively regularized dynamic polynomial PLS

Multiscale Simulations for Defect-Controlled Processing of Group IV Materials

Contact Info

Product

Resources

About