Sijie Fu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Sijie Fu has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- David D. Hebert
Carnegie Mellon University
2 shared publications - Ethan Tran
Department of Chemistry, Carnegie Mellon University Pittsburgh Pennsylvania 15213 USA
2 shared publications - Ilkwon Cho
Carnegie Mellon University
2 shared publications - Isaac Garcia-Bosch
Carnegie Mellon University
2 shared publications - Olexandr Isayev
Carnegie Mellon University
2 shared publications - Rajdeep Sarma
Department of Chemistry, Carnegie Mellon University Pittsburgh Pennsylvania 15213 USA
2 shared publications
Latest Publications
- Open Access
- 6 July 2026
- Published by Royal Society of Chemistry (RSC) in Chemical Science
Machine learning-accelerated screening of hydroquinone analogs for proton-coupled electron transfer.
Proton-coupled electron transfer (PCET) mediated by hydroquinone and related molecules is key to natural and artificial energy conversion. The reactivity of these molecules depends on their bond dissociation free energy (BDFE), but studying the relationship between structure and thermochemistry across this chemical space has been limited by challenging experimental setup and high computational expense. Here, we present the first use of the AIMNet2 neural network potential to calculate average BDFE (BDFEavg) values for the 2H+/2e- dehydrogenation of about 200 000 hydroquinone-like compounds, including vicinal diamines, diols, and dithiols. Benchmarking against DFT calculations for 168 substituted ortho-phenylenediamines (opda) shows good agreement (R 2 ∼ 0.84). Our analysis finds that the BDFEavg of diamines ranges from 50 to 80 kcal mol-1 and can be systematically tuned by modifying the backbone and N-substitution: electron-withdrawing groups raise BDFEavg by up to 15 kcal mol-1, while lower aromaticity in furan and thiophene backbones decreases BDFEavg by approximately 10 kcal mol-1 compared to the phenyl systems (∼65 kcal mol-1). Validation through cyclic voltammetry and reactivity studies with quinone oxidants for selected compounds supports the computational results. This extensive thermochemical database and a web-based prediction tool developed as a result of this work will offer valuable resources for designing PCET reagents for catalysis, energy storage, and biomedical uses.
- 6 July 2026
- Published by Royal Society of Chemistry (RSC) in Chemical Science
Open AccessMachine learning-accelerated screening of hydroquinone analogs for proton-coupled electron transfer.
Proton-coupled electron transfer (PCET) mediated by hydroquinone and related molecules is key to natural and artificial energy conversion. The reactivity of these molecules depends on their bond dissociation free energy (BDFE), but studying the relationship between structure and thermochemistry across this chemical space has been limited by challenging experimental setup and high computational expense. Here, we present the first use of the AIMNet2 neural network potential to calculate average BDFE (BDFEavg) values for the 2H+/2e- dehydrogenation of about 200 000 hydroquinone-like compounds, including vicinal diamines, diols, and dithiols. Benchmarking against DFT calculations for 168 substituted ortho-phenylenediamines (opda) shows good agreement (R 2 ∼ 0.84). Our analysis finds that the BDFEavg of diamines ranges from 50 to 80 kcal mol-1 and can be systematically tuned by modifying the backbone and N-substitution: electron-withdrawing groups raise BDFEavg by up to 15 kcal mol-1, while lower aromaticity in furan and thiophene backbones decreases BDFEavg by approximately 10 kcal mol-1 compared to the phenyl systems (∼65 kcal mol-1). Validation through cyclic voltammetry and reactivity studies with quinone oxidants for selected compounds supports the computational results. This extensive thermochemical database and a web-based prediction tool developed as a result of this work will offer valuable resources for designing PCET reagents for catalysis, energy storage, and biomedical uses.
- Open Access
- 5 November 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Machine Learning-Accelerated Screening of Hydroquinone Analogs for Proton-Coupled Electron Transfer
Proton-coupled electron transfer (PCET) mediated by hydroquinone and related molecules is key to natural and artificial energy conversion. The reactivity of these molecules depends on their bond dissociation free energy (BDFE), but studying the relationship between structure and thermochemistry across chemical space has been limited by computational expense. Here, we present the first use of the AIMNet2 neural network potential to calculate average BDFE (BDFEavg) values for the 2H+/2e− dehydrogenation of about 200,000 hydroquinone-like compounds, including vicinal diamines, diols, and dithiols. Benchmarking against DFT calculations for 168 substituted ortho-phenylenediamines (opda) shows good agreement (R² > 0.9). Our analysis finds that BDFEavg ranges from 50 to 80 kcal/mol and can be systematically tuned by modifying the backbone and N-substitution: electron-withdrawing groups raise BDFEavg by up to 15 kcal/mol, while lower aromaticity in furan and thiophene backbones decreases BDFEavg by approximately 10 kcal/mol compared to phenyl systems. We developed an additive "offset model" that allows separate investigation of backbone and sidechain effects. Validation through cyclic voltammetry and reactivity studies with quinone oxidants for selected compounds supports the computational results. This extensive thermochemical database and web-based prediction tool offer valuable resources for designing PCET reagents for catalysis, energy storage, and biomedical uses.
- 5 November 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessMachine Learning-Accelerated Screening of Hydroquinone Analogs for Proton-Coupled Electron Transfer
Proton-coupled electron transfer (PCET) mediated by hydroquinone and related molecules is key to natural and artificial energy conversion. The reactivity of these molecules depends on their bond dissociation free energy (BDFE), but studying the relationship between structure and thermochemistry across chemical space has been limited by computational expense. Here, we present the first use of the AIMNet2 neural network potential to calculate average BDFE (BDFEavg) values for the 2H+/2e− dehydrogenation of about 200,000 hydroquinone-like compounds, including vicinal diamines, diols, and dithiols. Benchmarking against DFT calculations for 168 substituted ortho-phenylenediamines (opda) shows good agreement (R² > 0.9). Our analysis finds that BDFEavg ranges from 50 to 80 kcal/mol and can be systematically tuned by modifying the backbone and N-substitution: electron-withdrawing groups raise BDFEavg by up to 15 kcal/mol, while lower aromaticity in furan and thiophene backbones decreases BDFEavg by approximately 10 kcal/mol compared to phenyl systems. We developed an additive "offset model" that allows separate investigation of backbone and sidechain effects. Validation through cyclic voltammetry and reactivity studies with quinone oxidants for selected compounds supports the computational results. This extensive thermochemical database and web-based prediction tool offer valuable resources for designing PCET reagents for catalysis, energy storage, and biomedical uses.
- Open Access
- 6 August 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Autonomous Determination of Hansen Solubility Parameters via Active Learning
Automated experimentation and self-driving laboratories (SDLs) have a transformative potential to accelerate and revolutionize scientific discoveries. However, their adoption by the general scientific community has been limited due to the cost of automation, such as instrumentation and laboratory infrastructure. In this work, we demonstrate the utilization of a centralized (semi-)automated lab as a cost-effective and accessible platform to democratize autonomous research. Specifically, by integrating a remote automated lab with artificial intelligence (AI), we developed a workflow called autoHSP, to demonstrate an autonomous and closed-loop determination of Hansen Solubility Parameters (HSPs) via batch-mode active learning (BMAL), without the need for any hands-on instrumentation setups. We developed algorithms to design experiments and analyze experiment outcomes, while outsourcing the experimentation to the automated lab remotely. A server was set up to handle communication between the automated lab and the AI framework securely with encryption. The lab received instructions and executed the experimental protocols to prepare the requested samples and report their images as characterization results. Then, a computer vision (CV) algorithm hosted on the server analyzed the images to determine the experiment outcome, and the BMAL algorithm selected solvents to optimize experiment design based on the most recent experimental knowledge until the end of experiment (EOE). Together, these demonstrate that this autonomous workflow can be adapted for general research where researchers can focus innovations on the AI-based software design.
- 6 August 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessAutonomous Determination of Hansen Solubility Parameters via Active Learning
Automated experimentation and self-driving laboratories (SDLs) have a transformative potential to accelerate and revolutionize scientific discoveries. However, their adoption by the general scientific community has been limited due to the cost of automation, such as instrumentation and laboratory infrastructure. In this work, we demonstrate the utilization of a centralized (semi-)automated lab as a cost-effective and accessible platform to democratize autonomous research. Specifically, by integrating a remote automated lab with artificial intelligence (AI), we developed a workflow called autoHSP, to demonstrate an autonomous and closed-loop determination of Hansen Solubility Parameters (HSPs) via batch-mode active learning (BMAL), without the need for any hands-on instrumentation setups. We developed algorithms to design experiments and analyze experiment outcomes, while outsourcing the experimentation to the automated lab remotely. A server was set up to handle communication between the automated lab and the AI framework securely with encryption. The lab received instructions and executed the experimental protocols to prepare the requested samples and report their images as characterization results. Then, a computer vision (CV) algorithm hosted on the server analyzed the images to determine the experiment outcome, and the BMAL algorithm selected solvents to optimize experiment design based on the most recent experimental knowledge until the end of experiment (EOE). Together, these demonstrate that this autonomous workflow can be adapted for general research where researchers can focus innovations on the AI-based software design.