Samantha Corapi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Samantha Corapi has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Alán Aspuru-Guzik
Department of Chemistry, University of Toronto, Toronto, Canada
3 shared publications - Jesse L. Prelesnik
University of Minnesota 207 Pleasant Street SE
2 shared publications - Joern Ilja Siepmann
Department of Chemical Engineering and Materials Science University of Minnesota Minneapolis Minnesota USA
2 shared publications - Randall Q. Snurr
Northwestern University
2 shared publications - Sergio Pablo-García
Instituto de Micro y Nanotecnología (IMN-CNM), CSIC, Isaac Newton 8, Tres Cantos, Madrid, Spain, 28760
2 shared publications - Benjamin Sanchez-Lengeling
University of Toronto
1 shared publication
Latest Publications
- Open Access
- 12 May 2026
- Published by American Chemical Society (ACS) in ACS Nano
- Vol. 20 (20) , 14514-14527
Adsorption Hysteresis Under Control: Tuning Host–Guest Interactions via a Genetic Algorithm
Mesoporous adsorbent materials offer a large volumetric capacity; however, cyclic adsorption/desorption processes in these systems often suffer from hysteresis and may require a significant pressure swing to access this capacity. To mitigate hysteresis, a proposed strategy is to include nucleation sites on the walls of the mesoporous material to facilitate droplet and bubble formation, lowering the free energy barriers to the respective phase transitions. It is unclear, however, what combination of adsorbate–adsorbent interactions and spatial patterning would be beneficial for a given application, considering that improvements to some sorption properties may come at the expense of other attributes. To understand these interconnected observables, we examine two model systems, planar-slit and cylindrical pores with tunable interaction sites, using GPU-accelerated transition matrix Monte Carlo simulations. The simulations provide a free energy map of the pressure–adsorption space in a matter of minutes, which we use to track adsorption isotherm characteristics as a function of adsorbent properties. We then leverage the rapid acquisition of simulation data to construct a genetic algorithm to iteratively modify interaction sites of the slit-pore wall to minimize the hysteresis of this system without sacrificing uptake. We find that the adsorption branch of the isotherm is easily modulated via the average host–guest interaction strength, but desorption is only adjustable if there is a suitable bubble nucleation site. Within the context of a slit-pore system, we identify relative interaction strengths and patch sizes required to gain control over both branches of the hysteresis loop.
- 12 May 2026
- Published by American Chemical Society (ACS) in ACS Nano
- Vol. 20 (20) , 14514-14527
Open AccessAdsorption Hysteresis Under Control: Tuning Host–Guest Interactions via a Genetic Algorithm
Mesoporous adsorbent materials offer a large volumetric capacity; however, cyclic adsorption/desorption processes in these systems often suffer from hysteresis and may require a significant pressure swing to access this capacity. To mitigate hysteresis, a proposed strategy is to include nucleation sites on the walls of the mesoporous material to facilitate droplet and bubble formation, lowering the free energy barriers to the respective phase transitions. It is unclear, however, what combination of adsorbate–adsorbent interactions and spatial patterning would be beneficial for a given application, considering that improvements to some sorption properties may come at the expense of other attributes. To understand these interconnected observables, we examine two model systems, planar-slit and cylindrical pores with tunable interaction sites, using GPU-accelerated transition matrix Monte Carlo simulations. The simulations provide a free energy map of the pressure–adsorption space in a matter of minutes, which we use to track adsorption isotherm characteristics as a function of adsorbent properties. We then leverage the rapid acquisition of simulation data to construct a genetic algorithm to iteratively modify interaction sites of the slit-pore wall to minimize the hysteresis of this system without sacrificing uptake. We find that the adsorption branch of the isotherm is easily modulated via the average host–guest interaction strength, but desorption is only adjustable if there is a suitable bubble nucleation site. Within the context of a slit-pore system, we identify relative interaction strengths and patch sizes required to gain control over both branches of the hysteresis loop.
- Open Access
- 17 November 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Adsorption Hysteresis Under Control: Tuning Host—Guest Interactions via a Genetic Algorithm
Mesoporous adsorbent materials offer a large volumetric capacity; however, cyclic adsorption/desorption processes in these systems often suffer from hysteresis and may require a significant pressure swing to access this capacity. To mitigate hysteresis, a proposed strategy is to include nucleation sites on the walls of the mesoporous material to facilitate droplet and bubble formation, lowering the free energy barriers to the respective phase transitions. It is unclear, however, what combination of adsorbate—adsorbent interactions and spatial patterning would be beneficial for a given application, considering that improvements to some sorption properties may come at the expense of other attributes. To understand these interconnected observables, we examine a model system, a planar slit-pore with tunable interaction sites, using GPU-accelerated transition matrix Monte Carlo simulations. The simulations provide a free energy map of the pressure—adsorption space in a matter of minutes, which we use to track adsorption isotherm characteristics as a function of adsorbent properties. We then leverage the rapid acquisition of simulation data to construct a genetic algorithm to iteratively modify interaction sites of the slit-pore wall to minimize the hysteresis of this system without sacrificing uptake. We find that the adsorption branch of the isotherm is easily modulated via the average host—guest interaction strength, but desorption is only adjustable if there is a suitable bubble nucleation site. Within the context of our model system, we identify relative interaction strengths and patch sizes required to gain control over both branches of the hysteresis loop.
- 17 November 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessAdsorption Hysteresis Under Control: Tuning Host—Guest Interactions via a Genetic Algorithm
Mesoporous adsorbent materials offer a large volumetric capacity; however, cyclic adsorption/desorption processes in these systems often suffer from hysteresis and may require a significant pressure swing to access this capacity. To mitigate hysteresis, a proposed strategy is to include nucleation sites on the walls of the mesoporous material to facilitate droplet and bubble formation, lowering the free energy barriers to the respective phase transitions. It is unclear, however, what combination of adsorbate—adsorbent interactions and spatial patterning would be beneficial for a given application, considering that improvements to some sorption properties may come at the expense of other attributes. To understand these interconnected observables, we examine a model system, a planar slit-pore with tunable interaction sites, using GPU-accelerated transition matrix Monte Carlo simulations. The simulations provide a free energy map of the pressure—adsorption space in a matter of minutes, which we use to track adsorption isotherm characteristics as a function of adsorbent properties. We then leverage the rapid acquisition of simulation data to construct a genetic algorithm to iteratively modify interaction sites of the slit-pore wall to minimize the hysteresis of this system without sacrificing uptake. We find that the adsorption branch of the isotherm is easily modulated via the average host—guest interaction strength, but desorption is only adjustable if there is a suitable bubble nucleation site. Within the context of our model system, we identify relative interaction strengths and patch sizes required to gain control over both branches of the hysteresis loop.
- Open Access
- 19 August 2025
- Published by AIP Publishing in APL Machine Learning
- Vol. 3 (3)
Ranking over regression for Bayesian optimization and molecule selection
Bayesian optimization (BO) has become an indispensable tool for autonomous decision-making across diverse applications from autonomous vehicle control to accelerated drug and materials discovery. With the growing interest in self-driving laboratories, BO of chemical systems is crucial for machine learning guided experimental planning. Typically, BO employs a regression surrogate model to predict the distribution of unseen parts of the search space. However, for the selection of molecules, picking the top candidates with respect to a distribution, the relative ordering of their properties may be more important than their exact values. In this paper, we introduce rank-based Bayesian optimization (RBO), which utilizes a ranking model as the surrogate. We present a comprehensive investigation of RBO’s optimization performance compared to conventional BO on various chemical datasets. Our results demonstrate similar or improved optimization performance using ranking models, particularly for datasets with rough structure–property landscapes and activity cliffs. Furthermore, we observe a high correlation between the surrogate ranking ability and BO performance, and this ability is maintained even at early iterations of BO optimization when using ranking surrogate models. We conclude that RBO is an effective alternative to regression-based BO, especially for optimizing novel chemical compounds.
- 19 August 2025
- Published by AIP Publishing in APL Machine Learning
- Vol. 3 (3)
Open AccessRanking over regression for Bayesian optimization and molecule selection
Bayesian optimization (BO) has become an indispensable tool for autonomous decision-making across diverse applications from autonomous vehicle control to accelerated drug and materials discovery. With the growing interest in self-driving laboratories, BO of chemical systems is crucial for machine learning guided experimental planning. Typically, BO employs a regression surrogate model to predict the distribution of unseen parts of the search space. However, for the selection of molecules, picking the top candidates with respect to a distribution, the relative ordering of their properties may be more important than their exact values. In this paper, we introduce rank-based Bayesian optimization (RBO), which utilizes a ranking model as the surrogate. We present a comprehensive investigation of RBO’s optimization performance compared to conventional BO on various chemical datasets. Our results demonstrate similar or improved optimization performance using ranking models, particularly for datasets with rough structure–property landscapes and activity cliffs. Furthermore, we observe a high correlation between the surrogate ranking ability and BO performance, and this ability is maintained even at early iterations of BO optimization when using ranking surrogate models. We conclude that RBO is an effective alternative to regression-based BO, especially for optimizing novel chemical compounds.