Longkun Xu
Longkun Xu has a total of 34 co-authors. Here are the top co-authors by number of shared publications:
- Michelle Louise Coote
Institute for NanoScale Science and Technology, College of Science and Engineering, Flinders University, South Australia, 5042, Australia
5 shared publications - Qiang Wang
Samsung Research China - Beijing
3 shared publications Ting-Jun Bi
College of Chemical Engineering, Sichuan University, Chengdu 610065, China.
3 shared publications- Wei Shao
Samsung Research China-Beijing (SRC-B)
3 shared publications - Xiang-Yuan Li
School of Chemical Engineering, Sichuan University, Chengdu 610065, China
3 shared publications - Fan Wang
Sichuan Univ, Inst Atom & Mol Phys, Chengdu 610065, Peoples R China
2 shared publications
Latest Publications
- Open Access
- 10 October 2024
- Published by American Chemical Society (ACS) in ChemRxiv
TADF-GEN: An Iterative and Self-improvingMethod with Long-term Memory and DynamicSimilarity Weight for Generating ThermallyActivated Delayed Fluorescence (TADF)Molecules
Thermally Activated Delayed Fluorescence (TADF) molecules are expected to be used in emitting layer materials for the next generation of organic light-emitting diodes (OLEDs) in various display applications, but their high-throughput discovery/generation are still challenging due to their vast chemical space, high cost of quantum chemical calculations and the tricky exploration-exploitation trade-off. Besides, TADF populations are far away from existing open-source database, which hinders the use of these database and makes the direct use of deep generative models a challenge. To address these issues, in this work we present an iterative and self-improving workflow, TADF-GEN. We combine atom-wise and fragment-based morphing operations designed with domain knowledge, machine learning based property prediction methods, multiscale quantum chemical calculations, TADF-specific metrics, with our special implementation of long-term memory (LTM) and dynamic similarity weight (DSW). With TADF-GEN, we explored the chemical space including various types and sizes of TADF molecules for the first time and established a new dataset of over 1.3 M molecules among which over 39 K molecules with TD-DFT labelled data. Combined LTM with DSM, we find the improvement of molecular diversity inherently help generate molecules with better performance. Besides, our model can effectively navigate the chemical space across different TADF domains (from multiple resonance type to donor-acceptor type). With accurate double hybrid TD-DFT excited state calculations, our generated molecules are proved excellent in both TADF properties and diversity compared with their seed molecules. Our TADF-GEN model and generated database can be directly used for future TADF works. Our proposed workflow is expected to be also useful for molecular discovery/generation in other data-shortage domains such as batteries and semiconductor materials.
- 10 October 2024
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessTADF-GEN: An Iterative and Self-improvingMethod with Long-term Memory and DynamicSimilarity Weight for Generating ThermallyActivated Delayed Fluorescence (TADF)Molecules
Thermally Activated Delayed Fluorescence (TADF) molecules are expected to be used in emitting layer materials for the next generation of organic light-emitting diodes (OLEDs) in various display applications, but their high-throughput discovery/generation are still challenging due to their vast chemical space, high cost of quantum chemical calculations and the tricky exploration-exploitation trade-off. Besides, TADF populations are far away from existing open-source database, which hinders the use of these database and makes the direct use of deep generative models a challenge. To address these issues, in this work we present an iterative and self-improving workflow, TADF-GEN. We combine atom-wise and fragment-based morphing operations designed with domain knowledge, machine learning based property prediction methods, multiscale quantum chemical calculations, TADF-specific metrics, with our special implementation of long-term memory (LTM) and dynamic similarity weight (DSW). With TADF-GEN, we explored the chemical space including various types and sizes of TADF molecules for the first time and established a new dataset of over 1.3 M molecules among which over 39 K molecules with TD-DFT labelled data. Combined LTM with DSM, we find the improvement of molecular diversity inherently help generate molecules with better performance. Besides, our model can effectively navigate the chemical space across different TADF domains (from multiple resonance type to donor-acceptor type). With accurate double hybrid TD-DFT excited state calculations, our generated molecules are proved excellent in both TADF properties and diversity compared with their seed molecules. Our TADF-GEN model and generated database can be directly used for future TADF works. Our proposed workflow is expected to be also useful for molecular discovery/generation in other data-shortage domains such as batteries and semiconductor materials.
- 4 January 2024
- Published by American Chemical Society (ACS) in The Journal of Physical Chemistry Letters
- Vol. 15 (2) , 371-379
A Multiple-Fidelity Method for Accurate Simulation of MoS2 Properties Using JAX-ReaxFF and Neural Network Potentials
Reactive force field (ReaxFF) is a commonly used force field for modeling chemical reactions at the atomic level. Recently, JAX-ReaxFF, combined with automatic differentiation, has been used to efficiently parametrize ReaxFF. However, its analytical formula may lead to inaccurate predictions. While neural network-based potentials (NNPs) trained on density functional theory-labeled data offer a more accurate method, it requires a large amount of training data to be trained from scratch. To overcome these issues, we present a multiple-fidelity method that combines JAX-ReaxFF and NNP and apply the method on MoS2, a promising two-dimensional semiconductor for flexible electronics. By incorporating implicit prior physical information, ReaxFF can serve as a cost-effective way to generate pretraining data, facilitating more accurate simulations of MoS2. Moreover, in the Mo-S-H system, the pretraining strategy can reduce root-mean-square errors of energy by 20%. This approach can be extended to a wide variety of material systems, accelerating their computational research.
- 4 January 2024
- Published by American Chemical Society (ACS) in The Journal of Physical Chemistry Letters
- Vol. 15 (2) , 371-379
A Multiple-Fidelity Method for Accurate Simulation of MoS2 Properties Using JAX-ReaxFF and Neural Network Potentials
Reactive force field (ReaxFF) is a commonly used force field for modeling chemical reactions at the atomic level. Recently, JAX-ReaxFF, combined with automatic differentiation, has been used to efficiently parametrize ReaxFF. However, its analytical formula may lead to inaccurate predictions. While neural network-based potentials (NNPs) trained on density functional theory-labeled data offer a more accurate method, it requires a large amount of training data to be trained from scratch. To overcome these issues, we present a multiple-fidelity method that combines JAX-ReaxFF and NNP and apply the method on MoS2, a promising two-dimensional semiconductor for flexible electronics. By incorporating implicit prior physical information, ReaxFF can serve as a cost-effective way to generate pretraining data, facilitating more accurate simulations of MoS2. Moreover, in the Mo-S-H system, the pretraining strategy can reduce root-mean-square errors of energy by 20%. This approach can be extended to a wide variety of material systems, accelerating their computational research.
- 13 December 2023
- Published by American Chemical Society (ACS) in The Journal of Physical Chemistry C
- Vol. 127 (50) , 24106-24117
Data Efficient and Stability Indicated Sampling for Developing Reactive Machine Learning Potential to Achieve Ultralong Simulation in Lithium-Metal Batteries
- 13 December 2023
- Published by American Chemical Society (ACS) in The Journal of Physical Chemistry C
- Vol. 127 (50) , 24106-24117
Data Efficient and Stability Indicated Sampling for Developing Reactive Machine Learning Potential to Achieve Ultralong Simulation in Lithium-Metal Batteries
- Open Access
- 24 October 2023
- Published by American Chemical Society (ACS) in ChemRxiv
A multiple-fidelity Method for Accurate Simulation of MoS 2 Properties Using JAX-ReaxFF and Neural Network Potentials
Reactive force field (ReaxFF) is one of the most commonly used force field to model the chemical reactions on atomic level. Recently, JAX-ReaxFF, combined with auto- matic differentiation, has been used to efficiently parameterize ReaxFF. However, pre- dicted properties using parameterized ReaxFF may be inaccurate due to the inductive bias of its analytical formula. While neural network-based potentials (NNPs) trained on density functional theory (DFT)-labeled data offer a more accurate method, it re- quires a large amount of training data to be trained from scratch. To overcome these issues, we present a multiple-fidelity method that combines JAX-ReaxFF and NNP, and apply the method on MoS 2 , a promising two-dimensional (2D) semiconductor for flexible electronics due to its excellent mechanical, optical, and electronic properties. By optimizing ReaxFF for MoS 2 and incorporating implicit prior physical information in the functional forms, we show that ReaxFF can serve as a cost-effective way to generate pretraining data, facilitating more accurate simulations of MoS 2 properties, such as the convex hull diagram, sulfur vacancy formation, and interaction with S 8 using SchNet. Moreover, in the Mo-S-H multi-element system, the pretraining strat- egy can reduce root-mean-square errors(RMSE) of energy by 20%. This approach can be extended to a wide variety of material systems, accelerating their computational research.
- 24 October 2023
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessA multiple-fidelity Method for Accurate Simulation of MoS 2 Properties Using JAX-ReaxFF and Neural Network Potentials
Reactive force field (ReaxFF) is one of the most commonly used force field to model the chemical reactions on atomic level. Recently, JAX-ReaxFF, combined with auto- matic differentiation, has been used to efficiently parameterize ReaxFF. However, pre- dicted properties using parameterized ReaxFF may be inaccurate due to the inductive bias of its analytical formula. While neural network-based potentials (NNPs) trained on density functional theory (DFT)-labeled data offer a more accurate method, it re- quires a large amount of training data to be trained from scratch. To overcome these issues, we present a multiple-fidelity method that combines JAX-ReaxFF and NNP, and apply the method on MoS 2 , a promising two-dimensional (2D) semiconductor for flexible electronics due to its excellent mechanical, optical, and electronic properties. By optimizing ReaxFF for MoS 2 and incorporating implicit prior physical information in the functional forms, we show that ReaxFF can serve as a cost-effective way to generate pretraining data, facilitating more accurate simulations of MoS 2 properties, such as the convex hull diagram, sulfur vacancy formation, and interaction with S 8 using SchNet. Moreover, in the Mo-S-H multi-element system, the pretraining strat- egy can reduce root-mean-square errors(RMSE) of energy by 20%. This approach can be extended to a wide variety of material systems, accelerating their computational research.
- Open Access
- 16 August 2023
- Published by American Chemical Society (ACS) in ChemRxiv
Data Efficient and Stability Indicated Samplingfor Developing Reactive Machine LearningPotential to Achieve Ultra-long Simulation inLithium Metal Batteries
Modelling the formation of solid-liquid interphase (SEI) is challenging as its strict requirement with both simulation accuracy and length. Machine learning potential (MLP) based molecular dynamics (MD) simulation is expected to play a role in this field while currently its use is hindered by sampling efficiency and simulation stability. In this work, we tackle the two challenges together. We propose the stability-indicated-sampling (SIS) algorithm for efficiently sampling training data using physical information (temperature). Unlike previous strategies, our method does not need prior knowledge of reaction networks or training multiple MLPs for uncertainty estimation. Compared with the recent proposed methods HAIR and DP-GEN, our approach gives significant improvement of sampling efficiency with less requirements with the initial training data, to realize > 10 ns MLPMD simulation using ab initio MD (AIMD) trajectory of just a few ps. We introduce the concept underlying instability consistency by showing the accuracy of reaction mechanisms and radial distribution function (RDF) can be improved by SIS-MLPMD, although their information is not explicitly used in our sampling decision. Furthermore, we show that long-time MLPMD simulation of Lithium metal battery (LMB) can not only reproduce some well-known SEI components including LiF, Li2O, LiOH, LiS and the incomplete N-S breaking in high-concentration systems, but also ionic aggregation structures of LiF, which is not shown in our AIMD training data but matches previous results of electrochemical impedance spectroscopy. Our work is expected to help accelerate future investigations, especially for studying long-time (≥ ns scale) reaction dynamics in interfacial problems.
- 16 August 2023
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessData Efficient and Stability Indicated Samplingfor Developing Reactive Machine LearningPotential to Achieve Ultra-long Simulation inLithium Metal Batteries
Modelling the formation of solid-liquid interphase (SEI) is challenging as its strict requirement with both simulation accuracy and length. Machine learning potential (MLP) based molecular dynamics (MD) simulation is expected to play a role in this field while currently its use is hindered by sampling efficiency and simulation stability. In this work, we tackle the two challenges together. We propose the stability-indicated-sampling (SIS) algorithm for efficiently sampling training data using physical information (temperature). Unlike previous strategies, our method does not need prior knowledge of reaction networks or training multiple MLPs for uncertainty estimation. Compared with the recent proposed methods HAIR and DP-GEN, our approach gives significant improvement of sampling efficiency with less requirements with the initial training data, to realize > 10 ns MLPMD simulation using ab initio MD (AIMD) trajectory of just a few ps. We introduce the concept underlying instability consistency by showing the accuracy of reaction mechanisms and radial distribution function (RDF) can be improved by SIS-MLPMD, although their information is not explicitly used in our sampling decision. Furthermore, we show that long-time MLPMD simulation of Lithium metal battery (LMB) can not only reproduce some well-known SEI components including LiF, Li2O, LiOH, LiS and the incomplete N-S breaking in high-concentration systems, but also ionic aggregation structures of LiF, which is not shown in our AIMD training data but matches previous results of electrochemical impedance spectroscopy. Our work is expected to help accelerate future investigations, especially for studying long-time (≥ ns scale) reaction dynamics in interfacial problems.
- 16 October 2021
- Published by American Chemical Society (ACS) in Journal of the American Chemical Society
- Vol. 143 (42) , 17431-17440
Experimental Evidence of Long-Lived Electric Fields of Ionic Liquid Bilayers
Herein we demonstrate that ionic liquids can form long-lived double layers, generating electric fields detectable by straightforward open circuit potential (OCP) measurements. In imidazolium-based ionic liquids an external negative voltage pulse leads to an exceedingly stable near-surface dipolar layer, whose field manifests as long-lived (∼1–100 h) discrete plateaus in OCP versus time traces. These plateaus occur within an ionic liquid-specific and sharp potential window, defining a simple experimental method to probe the onset of interfacial ordering phenomena, such as overscreening and crowding. Molecular dynamics modeling reveals that the OCP arises from the alignment of the individual ion dipoles to the external electric field pulse, with the magnitude of the resulting OCP correlating with the product of the projected dipole moment of the cation and the ratio between the cation diffusion coefficient and its volume. Our findings also reveal that a stable overscreened structure is more likely to form if the interface is first forced through crowding, possibly accounting for the scattered literature data on relaxation kinetics of near-surface structures in ionic liquids.
- 16 October 2021
- Published by American Chemical Society (ACS) in Journal of the American Chemical Society
- Vol. 143 (42) , 17431-17440
Experimental Evidence of Long-Lived Electric Fields of Ionic Liquid Bilayers
Herein we demonstrate that ionic liquids can form long-lived double layers, generating electric fields detectable by straightforward open circuit potential (OCP) measurements. In imidazolium-based ionic liquids an external negative voltage pulse leads to an exceedingly stable near-surface dipolar layer, whose field manifests as long-lived (∼1–100 h) discrete plateaus in OCP versus time traces. These plateaus occur within an ionic liquid-specific and sharp potential window, defining a simple experimental method to probe the onset of interfacial ordering phenomena, such as overscreening and crowding. Molecular dynamics modeling reveals that the OCP arises from the alignment of the individual ion dipoles to the external electric field pulse, with the magnitude of the resulting OCP correlating with the product of the projected dipole moment of the cation and the ratio between the cation diffusion coefficient and its volume. Our findings also reveal that a stable overscreened structure is more likely to form if the interface is first forced through crowding, possibly accounting for the scattered literature data on relaxation kinetics of near-surface structures in ionic liquids.
- Top Cited PapersOpen Access
- 10 December 2020
- Published by Springer Nature in Nature Communications
- Vol. 11 (1) , 1-8
The corona of a surface bubble promotes electrochemical reactions
The evolution of gaseous products is a feature common to several electrochemical processes, often resulting in bubbles adhering to the electrode’s surface. Adherent bubbles reduce the electrode active area, and are therefore generally treated as electrochemically inert entities. Here, we show that this general assumption does not hold for gas bubbles masking anodes operating in water. By means of imaging electrochemiluminescent systems, and by studying the anisotropy of polymer growth around bubbles, we demonstrate that gas cavities adhering to an electrode surface initiate the oxidation of water-soluble species more effectively than electrode areas free of bubbles. The corona of a bubble accumulates hydroxide anions, unbalanced by cations, a phenomenon which causes the oxidation of hydroxide ions to hydroxyl radicals to occur at potentials at least 0.7 V below redox tabled values. The downhill shift of the hydroxide oxidation at the corona of the bubble is likely to be a general mechanism involved in the initiation of heterogeneous electrochemical reactions in water, and could be harnessed in chemical synthesis.
- 10 December 2020
- Published by Springer Nature in Nature Communications
- Vol. 11 (1) , 1-8
Top Cited PapersOpen AccessThe corona of a surface bubble promotes electrochemical reactions
The evolution of gaseous products is a feature common to several electrochemical processes, often resulting in bubbles adhering to the electrode’s surface. Adherent bubbles reduce the electrode active area, and are therefore generally treated as electrochemically inert entities. Here, we show that this general assumption does not hold for gas bubbles masking anodes operating in water. By means of imaging electrochemiluminescent systems, and by studying the anisotropy of polymer growth around bubbles, we demonstrate that gas cavities adhering to an electrode surface initiate the oxidation of water-soluble species more effectively than electrode areas free of bubbles. The corona of a bubble accumulates hydroxide anions, unbalanced by cations, a phenomenon which causes the oxidation of hydroxide ions to hydroxyl radicals to occur at potentials at least 0.7 V below redox tabled values. The downhill shift of the hydroxide oxidation at the corona of the bubble is likely to be a general mechanism involved in the initiation of heterogeneous electrochemical reactions in water, and could be harnessed in chemical synthesis.
- 1 July 2020
- Published by American Chemical Society (ACS) in Journal of the American Chemical Society
- Vol. 142 (29) , 12826-12833
Ordered Solvents and Ionic Liquids Can Be Harnessed for Electrostatic Catalysis
Herein, we employ classical molecular dynamics simulations using the Drude oscillator-based polarizable force field, quantum chemical calculations, and ONIOM multiscale calculations to study (a) how an external field orders the solvent environment in a chemical reaction and then (b) whether in the absence of this same applied field the ordered solvent environment alone can electrostatically catalyze a chemical reaction when compared with the corresponding disordered solvent. Our results show that a 0.2 V/Å external electric field, which is below the threshold for bond breaking of solvent molecules, leads to significant ordering of bulk methanol solvent and the ionic liquid [EMIM][BF4]. Importantly, in the absence of this same field, the ordered solvent lowers the activation energy of the hydrogen-transfer reaction of o-alkylphenyl ketones in excess of 20 kcal/mol when the solvent is methanol and by over 30 kcal/mol for [EMIM][BF4]. Even a 0.1 V/Å external field has effects of ca. 10 and 20 kcal/mol, respectively. This work suggests a possible strategy for scaling electrostatic catalysis by applying a pulsed external field to the reaction medium to maintain solvent ordering while allowing the reaction to proceed largely in the absence of an external field.
- 1 July 2020
- Published by American Chemical Society (ACS) in Journal of the American Chemical Society
- Vol. 142 (29) , 12826-12833
Ordered Solvents and Ionic Liquids Can Be Harnessed for Electrostatic Catalysis
Herein, we employ classical molecular dynamics simulations using the Drude oscillator-based polarizable force field, quantum chemical calculations, and ONIOM multiscale calculations to study (a) how an external field orders the solvent environment in a chemical reaction and then (b) whether in the absence of this same applied field the ordered solvent environment alone can electrostatically catalyze a chemical reaction when compared with the corresponding disordered solvent. Our results show that a 0.2 V/Å external electric field, which is below the threshold for bond breaking of solvent molecules, leads to significant ordering of bulk methanol solvent and the ionic liquid [EMIM][BF4]. Importantly, in the absence of this same field, the ordered solvent lowers the activation energy of the hydrogen-transfer reaction of o-alkylphenyl ketones in excess of 20 kcal/mol when the solvent is methanol and by over 30 kcal/mol for [EMIM][BF4]. Even a 0.1 V/Å external field has effects of ca. 10 and 20 kcal/mol, respectively. This work suggests a possible strategy for scaling electrostatic catalysis by applying a pulsed external field to the reaction medium to maintain solvent ordering while allowing the reaction to proceed largely in the absence of an external field.
- 28 October 2019
- Published by American Chemical Society (ACS) in Journal of Chemical Theory and Computation
- Vol. 15 (12) , 6958-6967
Improving the Accuracy of PCM-UAHF and PCM-UAKS Calculations Using Optimized Electrostatic Scaling Factors
The optimal electrostatic scaling factor (ESF) for the calculation of solvation Gibbs free energies with the Polarizable Continuum Model (PCM) was determined via extensive benchmarking against 1719 experimental solvation Gibbs free energies and transfer free energies taken from the MNSol-v2012 database, as well as 45 aqueous pKa values covering 9 solute types (amines, thiols, carbon acid cations, pyridines, alcohols, anilines, carboxylic acids, carbon acid neutrals, phenols) and 20 pKa values in acetonitrile covering 4 solute types (phenols, carbon acids, carboxylic acids, pyridines). Optimal values for the ESF in a range of solvents are reported. For example, in water, the optimal ESF value is 1.2 and this does not vary with solute charge, radii type or method. For acetonitrile, we recommend 1.1 for neutrals, 1.0 and 1.1 respectively for cations with UAHF and UAKS radii types, and 1.3 for anions for IEFPCM-UAHF and 1.4 for other anions. At the same time, we show that ESF optimization does not address all errors in PCM, and is thus not a substitute for the appropriate use of explicit solvent molecules, nor the use of isodesmic methods to enhance systematic error cancellation. For a representative subset of the data we show that the errors in PCM are somewhat larger than in SMD and somewhat smaller than in COSMO-RS, although the latter has not benefitted from cavity scaling.
- 28 October 2019
- Published by American Chemical Society (ACS) in Journal of Chemical Theory and Computation
- Vol. 15 (12) , 6958-6967
Improving the Accuracy of PCM-UAHF and PCM-UAKS Calculations Using Optimized Electrostatic Scaling Factors
The optimal electrostatic scaling factor (ESF) for the calculation of solvation Gibbs free energies with the Polarizable Continuum Model (PCM) was determined via extensive benchmarking against 1719 experimental solvation Gibbs free energies and transfer free energies taken from the MNSol-v2012 database, as well as 45 aqueous pKa values covering 9 solute types (amines, thiols, carbon acid cations, pyridines, alcohols, anilines, carboxylic acids, carbon acid neutrals, phenols) and 20 pKa values in acetonitrile covering 4 solute types (phenols, carbon acids, carboxylic acids, pyridines). Optimal values for the ESF in a range of solvents are reported. For example, in water, the optimal ESF value is 1.2 and this does not vary with solute charge, radii type or method. For acetonitrile, we recommend 1.1 for neutrals, 1.0 and 1.1 respectively for cations with UAHF and UAKS radii types, and 1.3 for anions for IEFPCM-UAHF and 1.4 for other anions. At the same time, we show that ESF optimization does not address all errors in PCM, and is thus not a substitute for the appropriate use of explicit solvent molecules, nor the use of isodesmic methods to enhance systematic error cancellation. For a representative subset of the data we show that the errors in PCM are somewhat larger than in SMD and somewhat smaller than in COSMO-RS, although the latter has not benefitted from cavity scaling.
- 6 August 2019
- Published by American Chemical Society (ACS) in The Journal of Physical Chemistry A
- Vol. 123 (34) , 7430-7438
Methods To Improve the Calculations of Solvation Model Density Solvation Free Energies and Associated Aqueous pKa Values: Comparison between Choosing an Optimal Theoretical Level, Solute Cavity Scaling, and Using Explicit Solvent Molecules
Many approaches have been used to improve the accuracy of implicit solvent models including solute cavity scaling, introducing explicit solvent molecules, and changing the level of theory for the solvation calculations. Here, we compare these strategies using a large test set of aqueous pKa values for amines, nucleobases, carboxylic acids, thiols, peptide carbon acids, alcohols, and anilines for the specific case of solvation model density (SMD) within the framework of a thermodynamic cycle in which the gas-phase component is consistently calculated via the accurate CBS-QB3 method. We show that the choice of theoretical level for solvation energies should be based on the original parameterization of the solvent model, with separate levels of theory for the solvation energies of neutrals, anions, or cations, outperforming the best compromise level of theory. However, when explicit solvent molecules are introduced, a higher level of theory is needed to describe the solute–solvent interactions. For the systems studied here, explicit solvation improved the results for acids (and hence anions) but not for bases, for which results deteriorated. Importantly, we find that solute cavity scaling does not significantly improve the SMD results for the CHNO compounds tested when the correct theoretical level is employed and explicit solvent effects are correctly treated.
- 6 August 2019
- Published by American Chemical Society (ACS) in The Journal of Physical Chemistry A
- Vol. 123 (34) , 7430-7438
Methods To Improve the Calculations of Solvation Model Density Solvation Free Energies and Associated Aqueous pKa Values: Comparison between Choosing an Optimal Theoretical Level, Solute Cavity Scaling, and Using Explicit Solvent Molecules
Many approaches have been used to improve the accuracy of implicit solvent models including solute cavity scaling, introducing explicit solvent molecules, and changing the level of theory for the solvation calculations. Here, we compare these strategies using a large test set of aqueous pKa values for amines, nucleobases, carboxylic acids, thiols, peptide carbon acids, alcohols, and anilines for the specific case of solvation model density (SMD) within the framework of a thermodynamic cycle in which the gas-phase component is consistently calculated via the accurate CBS-QB3 method. We show that the choice of theoretical level for solvation energies should be based on the original parameterization of the solvent model, with separate levels of theory for the solvation energies of neutrals, anions, or cations, outperforming the best compromise level of theory. However, when explicit solvent molecules are introduced, a higher level of theory is needed to describe the solute–solvent interactions. For the systems studied here, explicit solvation improved the results for acids (and hence anions) but not for bases, for which results deteriorated. Importantly, we find that solute cavity scaling does not significantly improve the SMD results for the CHNO compounds tested when the correct theoretical level is employed and explicit solvent effects are correctly treated.