We are pleased to introduce the collection, 2021 Retrospective: Computational Materials Science. This collection showcases well-received spontaneous articles from the past couple of years, which have been specially handpicked by our Chief Editor, Prof. Douglas Soares Galvao. The work presented here highlights the broad diversity of research performed across the section, and aims to put a spotlight on the main areas of interest. All research presented here displays strong advances in theory, experiment and methodology with applications to compelling problems. This collection aims to further support Frontiers’ strong community by recognizing highly deserving authors.

State University of Campinas

Views

Article Views

Article Downloads

Frontiers in Materials

Computational Materials Science

2021 Retrospective: Computational Materials Science

Materials genome methods have played an essential role in accelerating the discovery of high-performance novel materials, and include high-throughput calculation, database construction, and machine learning. Over the past decades, these approaches have been increasingly used in lithium battery materials, solar cells, transparent conductors, and thermoelectrics. Thermoelectrics are functional materials that can directly convert electricity into heat and vice versa, offering new ideas for conventional power generation and refrigeration. The application of high-throughput methods can achieve more efficient screening of new thermoelectric materials and accelerate experimental development. This review summarizes the recent progress in the application of materials genome methods for different thermoelectric materials, such as half-Heuslers, diamond-like structures, oxides, and other materials. Finally, current advances in machine learning for thermoelectrics are discussed. The progress of the theoretical design of thermoelectrics has driven the development of high-performance thermoelectrics.

Materials genome methods have played an essential role in accelerating the discovery of high-performance novel materials, and include high-throughput calculation, database construction, and machine learning. Over the past decades, these approaches have been increasingly used in lithium battery materials, solar cells, transparent conductors, and thermoelectrics. Thermoelectrics are functional materials that can directly convert electricity into heat and vice versa, offering new ideas for conventional power generation and refrigeration. The application of high-throughput methods can achieve more efficient screening of new thermoelectric materials and accelerate experimental development. This review summarizes the recent progress in the application of materials genome methods for different thermoelectric materials, such as half-Heuslers, diamond-like structures, oxides, and other materials. Finally, current advances in machine learning for thermoelectrics are discussed. The progress of the theoretical design of thermoelectrics has driven the development of high-performance thermoelectrics.

Application of Materials Genome Methods in Thermoelectrics

This paper proposes a novel neural network architecture and its ensembles to predict the critical superconductivity temperature of materials based on their chemical formula. The research describes the methods and processes of extracting data from the chemical formula and preparing these extracted data for use in neural network training using TensorFlow. In our approach, recurrent neural networks are used including long short-term memory layers and neural networks based on one-dimensional convolution layers for data analysis. The proposed model is an ensemble of pre-trained neural network architectures for the prediction of the critical temperature of superconductors based on their chemical formula. The architecture of seven pre-trained neural networks is based on the long short-term memory layers and convolution layers. In the final ensemble, six neural networks are used: one network based on LSTM and four based on convolutional neural networks, and one embedding ensemble of convolution neural networks. LSTM neural network and convolution neural network were trained in 300 epochs. Ensembles of models were trained in 20 epochs. All neural networks are trained in two stages. At both stages, the optimizer Adam was used. In the first stage, training was carried out by the function of losses Mean Absolute Error (MAE) with the value of optimizer learning rate equal to 0.001. In the second stage, the previously trained model was trained by the function of losses Mean Squared Error (MSE) with a learning rate equal to 0.0001. The final ensemble is trained with a learning rate equal to 0.00001. The final ensemble model has the following accuracy values: MAE is 4.068, MSE is 67.272, and the coefficient of determination (R2) is 0.923. The final model can predict the critical temperature for the chemistry formula with an accuracy of 4.068°.

This paper proposes a novel neural network architecture and its ensembles to predict the critical superconductivity temperature of materials based on their chemical formula. The research describes the methods and processes of extracting data from the chemical formula and preparing these extracted data for use in neural network training using TensorFlow. In our approach, recurrent neural networks are used including long short-term memory layers and neural networks based on one-dimensional convolution layers for data analysis. The proposed model is an ensemble of pre-trained neural network architectures for the prediction of the critical temperature of superconductors based on their chemical formula. The architecture of seven pre-trained neural networks is based on the long short-term memory layers and convolution layers. In the final ensemble, six neural networks are used: one network based on LSTM and four based on convolutional neural networks, and one embedding ensemble of convolution neural networks. LSTM neural network and convolution neural network were trained in 300 epochs. Ensembles of models were trained in 20 epochs. All neural networks are trained in two stages. At both stages, the optimizer Adam was used. In the first stage, training was carried out by the function of losses Mean Absolute Error (MAE) with the value of optimizer learning rate equal to 0.001. In the second stage, the previously trained model was trained by the function of losses Mean Squared Error (MSE) with a learning rate equal to 0.0001. The final ensemble is trained with a learning rate equal to 0.00001. The final ensemble model has the following accuracy values: MAE is 4.068, MSE is 67.272, and the coefficient of determination (R2) is 0.923. The final model can predict the critical temperature for the chemistry formula with an accuracy of 4.068&#xb0;.

Deep Learning Approach for Prediction of Critical Temperature of Superconductor Materials Described by Chemical Formulas

γ-titanium aluminide (TiAl) alloys with fully lamellar microstructure possess excellent properties for high-temperature applications. Such fully lamellar microstructure has interfaces at different length scales. The separation behavior of the lamellae at these interfaces is crucial for the mechanical properties of the whole material. Unfortunately, quantifying it by experiments is difficult. Therefore, we use molecular dynamics (MD) simulations to this end. Specifically, we study the high-temperature separation behavior under tensile loading of the four different kinds of lamellar interfaces appearing in TiAl, namely, the <mml:math id="m1" xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mrow><mml:mi>γ</mml:mi><mml:mo>/</mml:mo><mml:mrow><mml:msub><mml:mi>α</mml:mi><mml:mn>2</mml:mn></mml:msub></mml:mrow></mml:mrow></mml:mrow></mml:math></inline-formula>, <mml:math id="m2" xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mrow><mml:mi>γ</mml:mi><mml:mo>/</mml:mo><mml:mrow><mml:msub><mml:mi>γ</mml:mi><mml:mrow><mml:mtext>PT</mml:mtext></mml:mrow></mml:msub></mml:mrow></mml:mrow></mml:mrow></mml:math></inline-formula>, <mml:math id="m3" xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mrow><mml:mi>γ</mml:mi><mml:mo>/</mml:mo><mml:mrow><mml:msub><mml:mi>γ</mml:mi><mml:mrow><mml:mi mathvariant="normal">TT</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:mrow></mml:mrow></mml:math></inline-formula>, and <mml:math id="m4" xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mrow><mml:mi>γ</mml:mi><mml:mo>/</mml:mo><mml:mrow><mml:msub><mml:mi>γ</mml:mi><mml:mrow><mml:mi mathvariant="normal">RB</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:mrow></mml:mrow></mml:math></inline-formula> interfaces. In our simulations, we use two different atomistic interface models, a defect-free (Type-1) model and a model with preexisting voids (Type-2). Clearly, the latter is more physical but studying the former also helps to understand the role of defects. Our simulation results show that among the four interfaces studied, the <mml:math id="m5" xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mrow><mml:mi>γ</mml:mi><mml:mo>/</mml:mo><mml:mrow><mml:msub><mml:mi>α</mml:mi><mml:mn>2</mml:mn></mml:msub></mml:mrow></mml:mrow></mml:mrow></mml:math></inline-formula> interface possesses the highest yield strength, followed by the <mml:math id="m6" xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mrow><mml:mi>γ</mml:mi><mml:mo>/</mml:mo><mml:mrow><mml:msub><mml:mi>γ</mml:mi><mml:mrow><mml:mi mathvariant="normal">PT</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:mrow></mml:mrow></mml:math></inline-formula>, <mml:math id="m7" xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mrow><mml:mi>γ</mml:mi><mml:mo>/</mml:mo><mml:mrow><mml:msub><mml:mi>γ</mml:mi><mml:mrow><mml:mi mathvariant="normal">TT</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:mrow></mml:mrow></mml:math></inline-formula>, and <mml:math id="m8" xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mrow><mml:mi>γ</mml:mi><mml:mo>/</mml:mo><mml:mrow><mml:msub><mml:mi>γ</mml:mi><mml:mrow><mml:mi mathvariant="normal">RB</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:mrow></mml:mrow></mml:math></inline-formula> interfaces. For Type-1 models, our simulations reveal failure at the interface for all γ/γ interfaces but not for the <mml:math id="m9" xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mrow><mml:mi>γ</mml:mi><mml:mo>/</mml:mo><mml:mrow><mml:msub><mml:mi>α</mml:mi><mml:mn>2</mml:mn></mml:msub></mml:mrow></mml:mrow></mml:mrow></mml:math></inline-formula> interface. By contrast, for Type-2 models, we observe for all the four interfaces failure at the interface. Our atomistic simulations provide important data to define the parameters of traction–separation laws and cohesive zone models, which can be used in the framework of continuum mechanical modeling of TiAl. Temperature-dependent model parameters were identified, and the complete traction–separation behavior was established, in which interface elasticity, interface plasticity, and interface damage could be distinguished. By carefully eliminating the contribution of bulk deformation from the interface behavior, we were able to quantify the contribution of interface plasticity and interface damage, which can also be related to the dislocation evolution and void nucleation in the atomistic simulations.

γ-titanium aluminide (TiAl) alloys with fully lamellar microstructure possess excellent properties for high-temperature applications. Such fully lamellar microstructure has interfaces at different length scales. The separation behavior of the lamellae at these interfaces is crucial for the mechanical properties of the whole material. Unfortunately, quantifying it by experiments is difficult. Therefore, we use molecular dynamics (MD) simulations to this end. Specifically, we study the high-temperature separation behavior under tensile loading of the four different kinds of lamellar interfaces appearing in TiAl, namely, the <math><mrow><mrow><mi>γ</mi><mo>/</mo><mrow><msub><mi>α</mi><mn>2</mn></msub></mrow></mrow></mrow></math>, <math><mrow><mrow><mi>γ</mi><mo>/</mo><mrow><msub><mi>γ</mi><mrow><mtext>PT</mtext></mrow></msub></mrow></mrow></mrow></math>, <math><mrow><mrow><mi>γ</mi><mo>/</mo><mrow><msub><mi>γ</mi><mrow><mi mathvariant="normal">TT</mi></mrow></msub></mrow></mrow></mrow></math>, and <math><mrow><mrow><mi>γ</mi><mo>/</mo><mrow><msub><mi>γ</mi><mrow><mi mathvariant="normal">RB</mi></mrow></msub></mrow></mrow></mrow></math> interfaces. In our simulations, we use two different atomistic interface models, a defect-free (Type-1) model and a model with preexisting voids (Type-2). Clearly, the latter is more physical but studying the former also helps to understand the role of defects. Our simulation results show that among the four interfaces studied, the <math><mrow><mrow><mi>γ</mi><mo>/</mo><mrow><msub><mi>α</mi><mn>2</mn></msub></mrow></mrow></mrow></math> interface possesses the highest yield strength, followed by the <math><mrow><mrow><mi>γ</mi><mo>/</mo><mrow><msub><mi>γ</mi><mrow><mi mathvariant="normal">PT</mi></mrow></msub></mrow></mrow></mrow></math>, <math><mrow><mrow><mi>γ</mi><mo>/</mo><mrow><msub><mi>γ</mi><mrow><mi mathvariant="normal">TT</mi></mrow></msub></mrow></mrow></mrow></math>, and <math><mrow><mrow><mi>γ</mi><mo>/</mo><mrow><msub><mi>γ</mi><mrow><mi mathvariant="normal">RB</mi></mrow></msub></mrow></mrow></mrow></math> interfaces. For Type-1 models, our simulations reveal failure at the interface for all γ/γ interfaces but not for the <math><mrow><mrow><mi>γ</mi><mo>/</mo><mrow><msub><mi>α</mi><mn>2</mn></msub></mrow></mrow></mrow></math> interface. By contrast, for Type-2 models, we observe for all the four interfaces failure at the interface. Our atomistic simulations provide important data to define the parameters of traction–separation laws and cohesive zone models, which can be used in the framework of continuum mechanical modeling of TiAl. Temperature-dependent model parameters were identified, and the complete traction–separation behavior was established, in which interface elasticity, interface plasticity, and interface damage could be distinguished. By carefully eliminating the contribution of bulk deformation from the interface behavior, we were able to quantify the contribution of interface plasticity and interface damage, which can also be related to the dislocation evolution and void nucleation in the atomistic simulations.

Quantifying the High-Temperature Separation Behavior of Lamellar Interfaces in γ-Titanium Aluminide Under Tensile Loading by Molecular Dynamics

Reinforced concrete is one of the most critical composite materials in the modern civil engineering and can improve the tensile resistance of concrete. Its passivation film plays an important role in the durability of concrete and the steel corrosion. But, due to the size limitations, the destruction of micro-scale steel bars has not been well studied. In this work, the reactive molecular dynamics simulation was employed to studying the mechanical properties of the steel and its passivation film. The uniaxial stretching of different compounds of γ-FeOOH, γ-Fe2O3 and Fe was performed. We found that the oxidation can reduce the tensile strength of steel. For the three compounds of γ-FeOOH, γ-Fe2O3 and Fe, the order of tensile strength from high to low is Fe > γ-Fe2O3 > γ-FeOOH. But, the ductility of γ-FeOOH under x direction is increased. The detail microstructure analysis shown that the difference of tensile strength is origin from the coordination in the materials. The two kinds of stretching processes of whole system stretching (in the Fe phase and x direction of γ-FeOOH phase) and partly area stretching (in the Fe2O3 phase and z direction of γ-FeOOH phase) were clarified. The external force is dispersed in whole system stretching but opposite in partly area stretching. This investigation leads to possible new direction for studying the tensile strength of materials, and the strategy of evaluating materials tensile strength can supply valuable information in evaluating and improving the mechanical properties of reinforced concrete.

Reinforced concrete is one of the most critical composite materials in the modern civil engineering and can improve the tensile resistance of concrete. Its passivation film plays an important role in the durability of concrete and the steel corrosion. But, due to the size limitations, the destruction of micro-scale steel bars has not been well studied. In this work, the reactive molecular dynamics simulation was employed to studying the mechanical properties of the steel and its passivation film. The uniaxial stretching of different compounds of γ-FeOOH, γ-Fe2O3 and Fe was performed. We found that the oxidation can reduce the tensile strength of steel. For the three compounds of γ-FeOOH, γ-Fe2O3 and Fe, the order of tensile strength from high to low is Fe &gt; γ-Fe2O3 &gt; γ-FeOOH. But, the ductility of γ-FeOOH under x direction is increased. The detail microstructure analysis shown that the difference of tensile strength is origin from the coordination in the materials. The two kinds of stretching processes of whole system stretching (in the Fe phase and x direction of γ-FeOOH phase) and partly area stretching (in the Fe2O3 phase and z direction of γ-FeOOH phase) were clarified. The external force is dispersed in whole system stretching but opposite in partly area stretching. This investigation leads to possible new direction for studying the tensile strength of materials, and the strategy of evaluating materials tensile strength can supply valuable information in evaluating and improving the mechanical properties of reinforced concrete.

Mechanical Properties and Microstructure of Reinforced Passivation Film: A Reactive Molecular Dynamics Study

Responsive materials, as well as active structural systems, are today widely used to develop unprecedented smart devices, sensors, or actuators; their functionalities come from the ability to respond to environmental stimuli with a detectable reaction. Depending on the responsive material under study, the triggering stimuli can have a different nature, ranging from physical (temperature, light, electric or magnetic field, mechanical stress, etc.), chemical (pH, ligands, etc.), or biological (enzymes, etc.) type. Such a responsiveness can be obtained by properly designing the meso- or macroscopic arrangement of the constitutive elements, as occurs in metamaterials, or can be obtained by using responsive materials per se, whose responsiveness comes from the chemistry underlying their microstructure. In fact, when the responsiveness at the molecular level is properly organized, the nanoscale response can be collectively detected at the macroscale, leading to a responsive material. In the present article, we review the enormous world of responsive polymers, by outlining the main features, characteristics, and responsive mechanisms of smart polymers and by providing a mechanical modeling perspective, both at the molecular as well as at the continuum scale level. We aim at providing a comprehensive overview of the main features and modeling aspects of the most diffused smart polymers. The quantitative mechanical description of active materials plays a key role in their development and use, enabling the design of advanced devices as well as to engineer the materials' microstructure according to the desired functionality.

Responsive materials, as well as active structural systems, are today widely used to develop unprecedented smart devices, sensors, or actuators; their functionalities come from the ability to respond to environmental stimuli with a detectable reaction. Depending on the responsive material under study, the triggering stimuli can have a different nature, ranging from physical (temperature, light, electric or magnetic field, mechanical stress, etc.), chemical (pH, ligands, etc.), or biological (enzymes, etc.) type. Such a responsiveness can be obtained by properly designing the meso- or macroscopic arrangement of the constitutive elements, as occurs in metamaterials, or can be obtained by using responsive materials per se, whose responsiveness comes from the chemistry underlying their microstructure. In fact, when the responsiveness at the molecular level is properly organized, the nanoscale response can be collectively detected at the macroscale, leading to a responsive material. In the present article, we review the enormous world of responsive polymers, by outlining the main features, characteristics, and responsive mechanisms of smart polymers and by providing a mechanical modeling perspective, both at the molecular as well as at the continuum scale level. We aim at providing a comprehensive overview of the main features and modeling aspects of the most diffused smart polymers. The quantitative mechanical description of active materials plays a key role in their development and use, enabling the design of advanced devices as well as to engineer the materials' microstructure according to the desired functionality.

Smart Polymers for Advanced Applications: A Mechanical Perspective Review

Though perovskite solar cells have good prospects, they also have some disadvantages, especially the impact of Pb on the environment and the use of expensive elements, which makes their production difficult to industrialize. Using first-principle density functional theory, we have investigated the geometric structures, electronic structures, and optical absorption coefficients of non-toxic and earth-abundant 1B-based perovskite solar cell absorbers. Our results show that Cs2AgAuI6, a toxin-free and inexpensive AgAu-based perovskite solar cell absorber, is suitable for use. It has a suitable HSE bandgap (1.289 eV) and a sharp absorption coefficient (~105 cm−1). Meanwhile, it is beneficial to the average electron and hole effective masses are 0.346 and 0.316 m0 respectively. The phonon spectra show that it is stable. Because the d-orbital energy of Cu is higher than those of Ag and Au, CuAu-based perovskite is not stable. This can be seen from the phonon spectra. Therefore, our calculations could provide strong evidence for the experimental synthesis of lead-free and low-cost perovskite solar cell absorber materials.

Though perovskite solar cells have good prospects, they also have some disadvantages, especially the impact of Pb on the environment and the use of expensive elements, which makes their production difficult to industrialize. Using first-principle density functional theory, we have investigated the geometric structures, electronic structures, and optical absorption coefficients of non-toxic and earth-abundant 1B-based perovskite solar cell absorbers. Our results show that Cs2AgAuI6, a toxin-free and inexpensive AgAu-based perovskite solar cell absorber, is suitable for use. It has a suitable HSE bandgap (1.289 eV) and a sharp absorption coefficient (~105 cm−1). Meanwhile, it is beneficial to the average electron and hole effective masses are 0.346 and 0.316 m0 respectively. The phonon spectra show that it is stable. Because the d-orbital energy of Cu is higher than those of Ag and Au, CuAu-based perovskite is not stable. This can be seen from the phonon spectra. Therefore, our calculations could provide strong evidence for the experimental synthesis of lead-free and low-cost perovskite solar cell absorber materials.