Geometrical design of a crystal growth system guided by a machine learning algorithm

Yu, Wancheng; Zhu, Chen; Tsunooka, Yosuke; Huang, Wei; Dang, Yifan; Kutsukake, Kentaro; Harada, Shunta; Tagawa, Miho; Ujihara, Toru

doi:10.1039/d1ce00106j

Cited by 25 publications

(12 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In optoelectronics, Tsunooka et al used ANN and SVR for rapid (10 7 faster) prediction of the results of CFD simulations of semiconductor crystals (SiC), which can aid the design and control of crystal growth by optimizing the variables and operating conditions. Yu et al similarly used ANN to aid the geometrical design of a top-seed solution growth device, enabling fast global optimization of parameters that govern crystal growth in a system where high-quality large crystals production is important. Dropka and Holena fed the results of 2D traveling magnetic field CFD simulations to ANN training and Gaussian process modeling to optimize the interface flattening on the production of multicrystalline, quasi-mono silicon ingots for solar cell applications.…”

Section: Data-driven Monitoring Modeling and Control Of Crystallizati...mentioning

confidence: 99%

Applications of Artificial Intelligence and Machine Learning Algorithms to Crystallization

et al. 2022

View full text Add to dashboard Cite

Artificial intelligence and specifically machine learning applications are nowadays used in a variety of scientific applications and cutting-edge technologies, where they have a transformative impact. Such an assembly of statistical and linear algebra methods making use of large data sets is becoming more and more integrated into chemistry and crystallization research workflows. This review aims to present, for the first time, a holistic overview of machine learning and cheminformatics applications as a novel, powerful means to accelerate the discovery of new crystal structures, predict key properties of organic crystalline materials, simulate, understand, and control the dynamics of complex crystallization process systems, as well as contribute to high throughput automation of chemical process development involving crystalline materials. We critically review the advances in these new, rapidly emerging research areas, raising awareness in issues such as the bridging of machine learning models with first-principles mechanistic models, data set size, structure, and quality, as well as the selection of appropriate descriptors. At the same time, we propose future research at the interface of applied mathematics, chemistry, and crystallography. Overall, this review aims to increase the adoption of such methods and tools by chemists and scientists across industry and academia.

show abstract

Section: Data-driven Monitoring Modeling and Control Of Crystallizati...mentioning

confidence: 99%

Applications of Artificial Intelligence and Machine Learning Algorithms to Crystallization

et al. 2022

View full text Add to dashboard Cite

show abstract

“…and a thickness of 20 mm has been reported, [ 44 ] and the possibility of growing larger SiC crystals by TSSG was investigated. [ 5 ] As is the case with other materials processing operations, scaling‐up always encounters difficulties, and in some cases, the empirical knowledge gained from manufacturing small products cannot be transferred directly to a process for larger products. In our study on TSSG of 3‐in.…”

Section: Design Of Objective Functions For 6‐in Sic Crystal Growthmentioning

confidence: 99%

“…application of machine learning techniques to CFD simulations greatly reduces the calculation time, so that it is possible to efficiently optimize the conditions for materials processing. [3][4][5][6][7][8][9] For mathematical optimization of the conditions for materials processing, it is necessary to define the objective function representing the optimal conditions to achieve high-quality products. In most cases, local values near or on the boundary between the material and fluid areas, such as the flow vectors, temperature, and crystal growth rate, are used as the objective function and optimized.…”

Section: Doi: 101002/adts202200302mentioning

confidence: 99%

“…More details on the TSSG setup are found in the literature. [5,6,26] The CFD simulation was performed on a 2D steady-axisymmetric model by taking into account the heat transport, mass transport, and convection using CGSim software. [27] The temperatures inside the seed shaft (T), the rotation speeds of the seed shaft (𝜔 s ) and crucible (𝜔 c ), and the position of the crucible (z) were set as variable parameters.…”

Section: Cfd Simulationmentioning

confidence: 99%

See 1 more Smart Citation

Optimization of Flow Distribution by Topological Description and Machine Learning in Solution Growth of SiC

Isono

Harada

Kutsukake

et al. 2022

Advcd Theory and Sims

Self Cite

View full text Add to dashboard Cite

The macroscopic distribution of fluid flows, which affect the quality of final products for various kinds of materials, is often difficult to describe in mathematical formulae and hinders the implementation of empirical knowledge in scaling up. In the present study, the characteristics of the flow distribution in silicon carbide (SiC) solution growth are described by using the position of the saddle point and the solution growth conditions are optimized by computational fluid dynamics simulation, machine learning, and a genetic algorithm. As a result, the candidates of the optimal condition for the solution growth of 6-in. SiC crystals are successfully obtained from the empirical knowledge gained from 3-in. crystal growth, by adding the topological description to the objective function. The present design of the objective function using the topological description can possibly be applied to other crystal growth or materials processing problems and to overcome scale-up difficulties, which can facilitate the rapid development of functional materials such as SiC wafers for power device applications. . IntroductionFluid flow that arises in many materials processing techniques, such as crystal growth, casting, thin film fabrication, and polymer processing, is critical to the quality and characteristics of the final product and optimization of the system. [1] Recent advances in computational fluid dynamics (CFD) aid in the design of suitable conditions for materials processing. [2] Furthermore, the

show abstract

“…This study applies transfer learning to the prediction of the time evolution of unsteady crystal growth to reduce the required amount of data, since the geometric changes between time steps are not significant and the major features of the initial input data representation could be maintained. In the field of crystal growth engineering, although numerous studies have been conducted on the prediction and design of growth conditions using machine learning, [25][26][27][28][29][30][31][32][33][34] the strategy of reducing the required amount of training data remains unexplored.…”

Section: Introductionmentioning

confidence: 99%

A Transfer Learning‐Based Method for Facilitating the Prediction of Unsteady Crystal Growth

Dang

Kutsukake

Liu

et al. 2022

Advcd Theory and Sims

Self Cite

View full text Add to dashboard Cite

Real-time prediction and dynamic control systems that can adapt to an unsteady environment are necessary for material fabrication processes, especially crystal growth. Recent studies have demonstrated the effectiveness of machine learning in predicting an unsteady crystal growth process, but its wider application is hindered by the large amount of training data required for sufficient accuracy. To address this problem, this study investigates the capability of transfer learning to predict geometric evolution in an unsteady silicon carbide (SiC) solution growth system based on a small amount of data. The performance of transferred models is discussed regarding the effect of the transfer learning method, training data amount, and time step length. The transfer learning strategy yields the same accuracy as that of training from scratch but requires only 20% of the training data. The accuracy is stably inherited through successive time steps, which demonstrates the effectiveness of transfer learning in reducing the required amount of training data for predicting evolution in an unsteady crystal growth process. Moreover, the transferred models trained with relatively more data (no more than 100%) further improve the accuracy inherited from the source model through multiple time steps, which broadens the application scope of transfer learning.

show abstract

Geometrical design of a crystal growth system guided by a machine learning algorithm

Abstract: In the design of a crystal growth system, the ability to efficiently regulate intertwined geometrical parameters is crucial for its successful development and commercialization. However, the traditional experimental and computational...

Cited by 25 publications

References 36 publications

Applications of Artificial Intelligence and Machine Learning Algorithms to Crystallization

Applications of Artificial Intelligence and Machine Learning Algorithms to Crystallization

Optimization of Flow Distribution by Topological Description and Machine Learning in Solution Growth of SiC

A Transfer Learning‐Based Method for Facilitating the Prediction of Unsteady Crystal Growth

Contact Info

Product

Resources

About