Inventing a modular approach to generate zero-carbon ammonia using renewable electricity would reduce our demands for fossil fuels. Through the application of electric field, it increases energy efficiency to 36.3 gNH3 kWh-1 and enhances reaction rates by up to a ten-fold compared to conventional heating. Currently, we are missing (i) the understandings of how electric fields accelerate reaction rates, and (ii) the effective deep learning to reduce the time for predicting field-induced catalytic performance. To fill these gaps, we first performed DFT calculations of electric field effects on ammonia synthesis over Ru various facets. Results show that a high positive electric field (&gt; 0.6 V/Å) shifts the most favorable thermodynamic reaction pathway to hydrogenation of N2* to form N2H* and further lowers the energetics of the most favorable reaction pathway over Ru(103) (Figure 1a, b). To accelerate the scan time for examining the catalytic performance under electric fields, we then developed a deep learning model, which employed a Graph Neural Network to capture the relationship among the geometries, followed by a shared multiple-layer perception for field-induced energetics prediction (Figure 1c). Our deep learning model gives us the optimal accuracy of field-dependent adsorption energies with the mean absolute error of adsorption energies, dipole moments, and polarizability with 5% outliers removed being within 0.08 eV, 0.09 eÅ, and 0.06 eÅ2/V, respectively (Figure 1d). This project will advance the fundamental chemical science and deep learning model of field-dipole effects on heterogeneous catalysis.

Deep learning-accelerated investigation of Field-Dipole interactions on ammonia synthesis

Fall 2022 abstracts, posters, and presentations will launch in mid-September.

ACS Fall 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.