Prasanna Venkataraman Balachandran
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Prasanna Venkataraman Balachandran has a total of 266 co-authors. Here are the top co-authors by number of shared publications:
Mukil V. Ayyasamy
Department of Materials Science and Engineering, University of Virginia, Charlottesville, USA
7 shared publications- Shunshun Liu
Department of Materials Science and Engineering, University of Virginia, Charlottesville, USA
6 shared publications Sunidhi Garg
Department of Materials Science and Engineering, University of Virginia, Charlottesville, VA, 22903, USA
6 shared publications- Timothy Q. Hartnett
Department of Materials Science and Engineering, University of Virginia 2 , Charlottesville, Virginia 22904, USA
5 shared publications - James M. Rondinelli
Department of Materials Science and Engineering, Northwestern University, Evanston, USA
4 shared publications Kyungtae Lee
Department of Materials Science and Engineering, University of Virginia, Charlottesville, USA
4 shared publications
Latest Publications
- 11 September 2026
- Published by Elsevier in Journal of the European Ceramic Society
- Vol. 47 (2)
Infrared absorption of lanthanide disilicate environmental barrier coating materials
- 11 September 2026
- Published by Elsevier in Journal of the European Ceramic Society
- Vol. 47 (2)
Infrared absorption of lanthanide disilicate environmental barrier coating materials
- 1 September 2026
- Published by Association for Materials Protection and Performance (AMPP) in Corrosion
- Vol. 82 (9) , 835-849
A Machine Learning Approach for Predicting Passive Current Densities for Cr-Passivated Fe-Based Alloys as a Function of Physical, Chemical, and Metallurgical Conditions
Stainless steel is widely used across industries due to its superior corrosion resistance, which is continuously optimized to enhance reliability and performance. This study uses machine learning (ML) to predict the passivation current density (ipass) of stainless steels in chloride-containing environments under varying pH and temperature conditions, aiming to establish a quantitative relationship between alloy composition and electrochemical passivation behavior. Unlike previous studies that focused on individual grades or limited environmental factors, this work integrates multialloy, multienvironment experimental and literature data into a unified dataset, enabling simultaneous analysis of compositional and electrochemical variables. A curated dataset comprising 52 refined literature data points and 202 new electrochemical measurements was used for model training. Ensemble support vector regression (eSVR) and extreme gradient tree boosting (eXGBoost) models demonstrated good generalization performance. Post-hoc model explanation identified pH as the most influential input, with Cr content, temperature, NaCl concentration, and Fe content also contributing substantially to the predictions. Increasing Cr content was associated with lower ipass, consistent with its established role as the principal passivating alloying element in Fe-Cr alloys. An interactive web application, CRSSI (corrosion-resistant stainless-steel informatics), was developed to facilitate accelerated screening of alloy compositions and environmental parameters, enabling real-time prediction of passivation current and serving as a practical resource for data-driven design of corrosion-resistant stainless steels. Thus, the primary recommendation is to use the developed ML framework to inform alloy screening and targeted data acquisition.
- 1 September 2026
- Published by Association for Materials Protection and Performance (AMPP) in Corrosion
- Vol. 82 (9) , 835-849
A Machine Learning Approach for Predicting Passive Current Densities for Cr-Passivated Fe-Based Alloys as a Function of Physical, Chemical, and Metallurgical Conditions
Stainless steel is widely used across industries due to its superior corrosion resistance, which is continuously optimized to enhance reliability and performance. This study uses machine learning (ML) to predict the passivation current density (ipass) of stainless steels in chloride-containing environments under varying pH and temperature conditions, aiming to establish a quantitative relationship between alloy composition and electrochemical passivation behavior. Unlike previous studies that focused on individual grades or limited environmental factors, this work integrates multialloy, multienvironment experimental and literature data into a unified dataset, enabling simultaneous analysis of compositional and electrochemical variables. A curated dataset comprising 52 refined literature data points and 202 new electrochemical measurements was used for model training. Ensemble support vector regression (eSVR) and extreme gradient tree boosting (eXGBoost) models demonstrated good generalization performance. Post-hoc model explanation identified pH as the most influential input, with Cr content, temperature, NaCl concentration, and Fe content also contributing substantially to the predictions. Increasing Cr content was associated with lower ipass, consistent with its established role as the principal passivating alloying element in Fe-Cr alloys. An interactive web application, CRSSI (corrosion-resistant stainless-steel informatics), was developed to facilitate accelerated screening of alloy compositions and environmental parameters, enabling real-time prediction of passivation current and serving as a practical resource for data-driven design of corrosion-resistant stainless steels. Thus, the primary recommendation is to use the developed ML framework to inform alloy screening and targeted data acquisition.
- Open Access
- 14 August 2026
- Published by Elsevier in SSRN Electronic Journal
The Infrared Properties of Ytterbium Disilicate
Ytterbium disilicate (Yb2Si2O7: YbDS) is a wide-bandgap, high-melting-temperature, polar material widely used as the primary thermal and environmental barrier coating for silicon carbide composites deployed in the hottest regions of advanced gas turbine engines. These coatings are exposed to a substantial radiative flux spanning the ultraviolet through the mid-infrared as well as to convective heating. To assess the significance of radiative heating from combustion gas emissions, the complex dielectric function of polycrystalline YbDS has been characterized from ambient to 1,000 °C using infrared spectroscopic ellipsometry and Fourier transform infrared spectroscopy. Density functional theory calculations of the zone-center phonon modes, combined with Born effective charge calculations, were used to identify the infrared-active phonon modes responsible for the observed spectral features. The measured intrinsic optical properties have enabled the fractions of incident flux that are reflected, transmitted, absorbed, and re-emitted by YbDS coating materials to be deduced. These fractions are strongly wavelength-dependent in the near and mid-infrared but are only weakly influenced by temperature. The experimental measurements reveal that YbDS has a low reflectance and is highly transparent in the 2-5 μm wavelength region, indicating that coating-protected SiC components are vulnerable to radiative heating by combustion gases. The phonon calculations show that the spectral location of the transparency window results from those of the longitudinal and transverse optical phonon modes. Their wavelengths are determined by displacements of the Si-O bonds whose stiffness and effective ion mass are intrinsic properties of the disilicate sublattice and effected little by the lanthanide ion.
- 14 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessThe Infrared Properties of Ytterbium Disilicate
Ytterbium disilicate (Yb2Si2O7: YbDS) is a wide-bandgap, high-melting-temperature, polar material widely used as the primary thermal and environmental barrier coating for silicon carbide composites deployed in the hottest regions of advanced gas turbine engines. These coatings are exposed to a substantial radiative flux spanning the ultraviolet through the mid-infrared as well as to convective heating. To assess the significance of radiative heating from combustion gas emissions, the complex dielectric function of polycrystalline YbDS has been characterized from ambient to 1,000 °C using infrared spectroscopic ellipsometry and Fourier transform infrared spectroscopy. Density functional theory calculations of the zone-center phonon modes, combined with Born effective charge calculations, were used to identify the infrared-active phonon modes responsible for the observed spectral features. The measured intrinsic optical properties have enabled the fractions of incident flux that are reflected, transmitted, absorbed, and re-emitted by YbDS coating materials to be deduced. These fractions are strongly wavelength-dependent in the near and mid-infrared but are only weakly influenced by temperature. The experimental measurements reveal that YbDS has a low reflectance and is highly transparent in the 2-5 μm wavelength region, indicating that coating-protected SiC components are vulnerable to radiative heating by combustion gases. The phonon calculations show that the spectral location of the transparency window results from those of the longitudinal and transverse optical phonon modes. Their wavelengths are determined by displacements of the Si-O bonds whose stiffness and effective ion mass are intrinsic properties of the disilicate sublattice and effected little by the lanthanide ion.
- Open Access
- 20 June 2026
- Published by Springer Nature in npj Computational Materials
- Vol. 12 (1)
A computational Dieke design map for near-infrared optical absorption in transition metal-substituted yttria-stabilized zirconia
- 20 June 2026
- Published by Springer Nature in npj Computational Materials
- Vol. 12 (1)
Open AccessA computational Dieke design map for near-infrared optical absorption in transition metal-substituted yttria-stabilized zirconia
- Open Access
- 11 June 2026
- Published by Royal Society of Chemistry (RSC) in Digital Discovery
- Vol. 5 (8) , 3212-3223
Towards “on-demand” van der Waals epitaxy with adaptive ensemble sampling atomistic workflows
A real-time bayesian sampling with computing-resource scalable ensemble molecular dynamics workflow uncovers crystallization pathways for on-demand target moire twisted MoS 2 bi-layers from amorphous precursors.
- 11 June 2026
- Published by Royal Society of Chemistry (RSC) in Digital Discovery
- Vol. 5 (8) , 3212-3223
Open AccessTowards “on-demand” van der Waals epitaxy with adaptive ensemble sampling atomistic workflows
A real-time bayesian sampling with computing-resource scalable ensemble molecular dynamics workflow uncovers crystallization pathways for on-demand target moire twisted MoS 2 bi-layers from amorphous precursors.
- Open Access
- 11 June 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
A physics-regularized machine learning approach for predicting time-temperature-transformation curves in alloys: application to uranium-based alloys
A physics-regularized machine learning (ML) approach is developed for predicting time-temperature-transformation (TTT) diagrams from alloy composition in uranium (U)-based systems. To ensure physically realistic C-curve shape while maintaining predictive accuracy, four loss function formulations are systematically evaluated: (1) baseline mean squared error (MSE), (2) MSE regularized with a semi-empirical nucleation-based model, (3) MSE regularized with shape constraints that enforce characteristic C-curve convexity, and (4) a hybrid approach that combines both semi-empirical regularization and shape constraints with MSE. Models that incorporate shape-constrained loss functions generate TTT predictions exhibiting a physically consistent C-curve form and demonstrate good agreement with experimental isothermal transformation data. Post hoc model explanation using global feature importance and partial dependence plots reveals systematic changes in the learned composition-TTT relationships as a function of the loss function formulation. The optimized ML model, which used a hybrid loss function, was applied to previously unexplored U-Mo-X ternary systems. This analysis identified U-Mo-Pt, U-Mo-Au, U-Mo-Ta, U-Mo-W, and U-Mo-Fe as candidate alloys that exhibit enhanced kinetic stability against γ-to-α phase decomposition during isothermal holds compared to the U-Mo binary baseline. Furthermore, TTT nose time predictions from the optimized ML model are compared with those from a Gaussian Process Regression (GPR) model using MSE loss. The developed physics-regularized ML framework offers potential for accelerating alloy discovery in U-based systems, where traditional TTT experimental determination is time and resource-intensive.
- 11 June 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
Open AccessA physics-regularized machine learning approach for predicting time-temperature-transformation curves in alloys: application to uranium-based alloys
A physics-regularized machine learning (ML) approach is developed for predicting time-temperature-transformation (TTT) diagrams from alloy composition in uranium (U)-based systems. To ensure physically realistic C-curve shape while maintaining predictive accuracy, four loss function formulations are systematically evaluated: (1) baseline mean squared error (MSE), (2) MSE regularized with a semi-empirical nucleation-based model, (3) MSE regularized with shape constraints that enforce characteristic C-curve convexity, and (4) a hybrid approach that combines both semi-empirical regularization and shape constraints with MSE. Models that incorporate shape-constrained loss functions generate TTT predictions exhibiting a physically consistent C-curve form and demonstrate good agreement with experimental isothermal transformation data. Post hoc model explanation using global feature importance and partial dependence plots reveals systematic changes in the learned composition-TTT relationships as a function of the loss function formulation. The optimized ML model, which used a hybrid loss function, was applied to previously unexplored U-Mo-X ternary systems. This analysis identified U-Mo-Pt, U-Mo-Au, U-Mo-Ta, U-Mo-W, and U-Mo-Fe as candidate alloys that exhibit enhanced kinetic stability against γ-to-α phase decomposition during isothermal holds compared to the U-Mo binary baseline. Furthermore, TTT nose time predictions from the optimized ML model are compared with those from a Gaussian Process Regression (GPR) model using MSE loss. The developed physics-regularized ML framework offers potential for accelerating alloy discovery in U-based systems, where traditional TTT experimental determination is time and resource-intensive.
- Open Access
- 10 June 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
An active learning workflow for predicting misfit volume in body-centered cubic refractory high-entropy alloys
Refractory high-entropy alloys (RHEAs) exhibit exceptional high-temperature mechanical properties. However, a mechanistic understanding of their yield strength requires accurate determination of the misfit volume descriptor (δ), which quantifies the local volume change due to the size and electronic heterogeneity of constituent elements in the solid solution. Traditional approximations, such as Vegard's Law, fail to capture local atomic relaxation and electronic structure effects in these compositionally complex systems. An active learning workflow is developed that integrates density functional theory calculations with ensemble machine learning (ML) to efficiently predict the δ-descriptor across 126 equiatomic quinary body-centered cubic (BCC) RHEAs. Partial dependence plots and symbolic regression reveal that atomic size mismatch directly influences the δ-descriptor, while electronegativity variations provide electronic structure compensation. Incorporating ML-predicted δ-values into the Maresca-Curtin mechanistic model achieves good correlation with experimental yield strengths across diverse BCC RHEA compositions.
- 10 June 2026
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
Open AccessAn active learning workflow for predicting misfit volume in body-centered cubic refractory high-entropy alloys
Refractory high-entropy alloys (RHEAs) exhibit exceptional high-temperature mechanical properties. However, a mechanistic understanding of their yield strength requires accurate determination of the misfit volume descriptor (δ), which quantifies the local volume change due to the size and electronic heterogeneity of constituent elements in the solid solution. Traditional approximations, such as Vegard's Law, fail to capture local atomic relaxation and electronic structure effects in these compositionally complex systems. An active learning workflow is developed that integrates density functional theory calculations with ensemble machine learning (ML) to efficiently predict the δ-descriptor across 126 equiatomic quinary body-centered cubic (BCC) RHEAs. Partial dependence plots and symbolic regression reveal that atomic size mismatch directly influences the δ-descriptor, while electronegativity variations provide electronic structure compensation. Incorporating ML-predicted δ-values into the Maresca-Curtin mechanistic model achieves good correlation with experimental yield strengths across diverse BCC RHEA compositions.
- Open Access
- 15 May 2026
- Published by Springer Nature in Integrating Materials and Manufacturing Innovation
- Vol. 15 (2) , 290-306
Expert-Grounded Automatic Prompt Engineering for Extracting Lattice Constants of High-Entropy Alloys from Scientific Publications Using Large Language Models
Large language models (LLMs) have shown promise for scientific data extraction from publications but rely on manual prompt refinement. We present an expert-grounded automatic prompt optimization framework that enhances LLM entity extraction reliability. Using high-entropy alloy lattice constant extraction as a testbed, we optimized prompts for Claude 3.5 Sonnet through feedback cycles on seven expert-annotated publications. Despite a modest optimization budget, recall improved from 0.27 to > 0.9, demonstrating that a small, expert-annotated dataset can yield significant improvements. The optimized prompt was applied to extract lattice constants from 2267 publications, yielding data for 1861 compositions. The optimized prompt transferred effectively to newer models: Claude 4.5 Sonnet, GPT-5, and Gemini 2.5 Flash. Analysis revealed three categories of LLM failure modes: contextual hallucination, semantic misinterpretation, and unit conversion errors, emphasizing the need for validation protocols. These results establish feedback-guided prompt optimization as a low-cost, transferable methodology for reliable scientific data extraction, providing a scalable pathway for complex LLM-assisted research tasks.
- 15 May 2026
- Published by Springer Nature in Integrating Materials and Manufacturing Innovation
- Vol. 15 (2) , 290-306
Open AccessExpert-Grounded Automatic Prompt Engineering for Extracting Lattice Constants of High-Entropy Alloys from Scientific Publications Using Large Language Models
Large language models (LLMs) have shown promise for scientific data extraction from publications but rely on manual prompt refinement. We present an expert-grounded automatic prompt optimization framework that enhances LLM entity extraction reliability. Using high-entropy alloy lattice constant extraction as a testbed, we optimized prompts for Claude 3.5 Sonnet through feedback cycles on seven expert-annotated publications. Despite a modest optimization budget, recall improved from 0.27 to > 0.9, demonstrating that a small, expert-annotated dataset can yield significant improvements. The optimized prompt was applied to extract lattice constants from 2267 publications, yielding data for 1861 compositions. The optimized prompt transferred effectively to newer models: Claude 4.5 Sonnet, GPT-5, and Gemini 2.5 Flash. Analysis revealed three categories of LLM failure modes: contextual hallucination, semantic misinterpretation, and unit conversion errors, emphasizing the need for validation protocols. These results establish feedback-guided prompt optimization as a low-cost, transferable methodology for reliable scientific data extraction, providing a scalable pathway for complex LLM-assisted research tasks.
- Open Access
Constrained Diffusion for Accelerated Structure Relaxation of Inorganic Solids with Point Defects
Point defects affect material properties by altering electronic states and modifying local bonding environments. However, high-throughput first-principles simulations of point defects are costly due to large simulation cells and complex energy landscapes. To this end, we propose a generative framework for simulating point defects, overcoming the limits of costly first-principles simulators. By leveraging a primal-dual algorithm, we introduce a constraint-aware diffusion model which outperforms existing constrained diffusion approaches in this domain. Across six defect configuration settings for Bi2Te3, the proposed approach provides state-of-the-art performance generating physically grounded structures.
Open AccessConstrained Diffusion for Accelerated Structure Relaxation of Inorganic Solids with Point Defects
Point defects affect material properties by altering electronic states and modifying local bonding environments. However, high-throughput first-principles simulations of point defects are costly due to large simulation cells and complex energy landscapes. To this end, we propose a generative framework for simulating point defects, overcoming the limits of costly first-principles simulators. By leveraging a primal-dual algorithm, we introduce a constraint-aware diffusion model which outperforms existing constrained diffusion approaches in this domain. Across six defect configuration settings for Bi2Te3, the proposed approach provides state-of-the-art performance generating physically grounded structures.
- Open Access
- 2 February 2026
- Published by Elsevier in SSRN Electronic Journal
Mechanistic insights into laser-generated surface nanomorphology from physics-guided, explainable machine learning and molecular dynamics simulations
Nanoscale surface morphologies generated by ultrashort laser pulses are products of a complex interplay of laser energy deposition, pressure generation, melt dynamics, and rapid resolidification. In this work, we investigate the physical factors controlling the formation of laser-induced surface nanostructures using a combined data-driven and physics-based approach. A machine learning (ML) model is developed to predict the number density of surface nanoprotrusions formed during single-pulse laser processing of metals, using a compact set of physically motivated dimensionless descriptors derived via dimensional analysis. These descriptors capture key stages of surface morphology generation, including energy deposition, pressure buildup, liquid filament breakup, and solidification. The ML model is trained on quantitative features extracted from experimental scanning electron microscopy images spanning 39 data points across nine metals. Post hoc model explanation analysis identifies the absorbed fluence and the Ohnesorge number as the most influential descriptors. The electron-phonon coupling strength emerges as a key material parameter through its control of the effective energy deposition depth and the characteristic length scale of transient liquid structures. These trends are further examined using large-scale two-temperature molecular dynamics simulations, which confirm the predicted sensitivity of surface morphology to electron–phonon coupling near the spallation threshold.
- 2 February 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessMechanistic insights into laser-generated surface nanomorphology from physics-guided, explainable machine learning and molecular dynamics simulations
Nanoscale surface morphologies generated by ultrashort laser pulses are products of a complex interplay of laser energy deposition, pressure generation, melt dynamics, and rapid resolidification. In this work, we investigate the physical factors controlling the formation of laser-induced surface nanostructures using a combined data-driven and physics-based approach. A machine learning (ML) model is developed to predict the number density of surface nanoprotrusions formed during single-pulse laser processing of metals, using a compact set of physically motivated dimensionless descriptors derived via dimensional analysis. These descriptors capture key stages of surface morphology generation, including energy deposition, pressure buildup, liquid filament breakup, and solidification. The ML model is trained on quantitative features extracted from experimental scanning electron microscopy images spanning 39 data points across nine metals. Post hoc model explanation analysis identifies the absorbed fluence and the Ohnesorge number as the most influential descriptors. The electron-phonon coupling strength emerges as a key material parameter through its control of the effective energy deposition depth and the characteristic length scale of transient liquid structures. These trends are further examined using large-scale two-temperature molecular dynamics simulations, which confirm the predicted sensitivity of surface morphology to electron–phonon coupling near the spallation threshold.