Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Machine learning exploration of binding energy distributions of H2O at astrochemically relevant dust grain surfaces

Anant Vaishnav, Niels M. Mikkelsen, Mie Andersen

TL;DR

This work addresses how water BE distributions on dust-grain surfaces depend on both the underlying material (graphene vs forsterite) and the ice structure (crystalline vs amorphous) across submonolayer to few-layer regimes. The authors train PaiNN-based MLIPs to efficiently sample adsorption sites and BE distributions on graphene and Mg-terminated forsterite, leveraging GOFEE and ab initio MD to generate diverse ice morphologies and surfaces. They find that strong Mg–O bonding on silicate yields deep adsorption sites in the submonolayer regime, while hydrogen-bonding within ice dominates in monolayer and thicker coverings, reducing surface dependence; amorphous ice generally shifts BE distributions toward stronger binding and creates defect pockets. These results provide physically grounded BE inputs for next-generation astrochemical models incorporating surface heterogeneity and have implications for diffusion, desorption, and reaction pathways of water and related adsorbates in star- and planet-forming regions.

Abstract

Binding energies (BEs) of adsorbates on interstellar dust grains critically control adsorption, desorption, diffusion, and surface reactivity, and therefore strongly influence astrochemical models of star- and planet-forming regions. While recent computational studies increasingly report full distributions of BEs rather than single representative values, these distributions are typically derived for either bare grain surfaces or thick water-ice mantles. In this work, we bridge these regimes by systematically investigating the BE distributions of water on partially and fully ice-covered dust grain surfaces. We employ machine-learning interatomic potentials (MLIPs) based on graph neural networks to model water adsorption on graphene and on the Mg-terminated (010) surface of forsterite, representing carbonaceous and silicate grains, respectively. The models enable extensive sampling of adsorption sites on water clusters, monolayers, and bilayers generated under both crystalline (thermally processed) and amorphous (low-temperature) growth conditions. At submonolayer coverage, the chemical nature of the underlying grain strongly affects both ice morphology and binding energies, with Mg-O interactions on silicate surfaces producing particularly deep binding sites. From monolayer coverage onward, adsorption on both substrates is dominated by hydrogen bonding within the ice, reducing the influence of the grain material. Across all coverages, amorphous ice structures systematically shift the BE distributions toward stronger binding compared to crystalline ice, introducing highly stable defect and pocket sites. These results demonstrate that BE distributions in the submonolayer to few-layer ice regime are broad and highly surface dependent, and they provide physically motivated input for next-generation astrochemical models incorporating surface heterogeneity.

Machine learning exploration of binding energy distributions of H2O at astrochemically relevant dust grain surfaces

TL;DR

This work addresses how water BE distributions on dust-grain surfaces depend on both the underlying material (graphene vs forsterite) and the ice structure (crystalline vs amorphous) across submonolayer to few-layer regimes. The authors train PaiNN-based MLIPs to efficiently sample adsorption sites and BE distributions on graphene and Mg-terminated forsterite, leveraging GOFEE and ab initio MD to generate diverse ice morphologies and surfaces. They find that strong Mg–O bonding on silicate yields deep adsorption sites in the submonolayer regime, while hydrogen-bonding within ice dominates in monolayer and thicker coverings, reducing surface dependence; amorphous ice generally shifts BE distributions toward stronger binding and creates defect pockets. These results provide physically grounded BE inputs for next-generation astrochemical models incorporating surface heterogeneity and have implications for diffusion, desorption, and reaction pathways of water and related adsorbates in star- and planet-forming regions.

Abstract

Binding energies (BEs) of adsorbates on interstellar dust grains critically control adsorption, desorption, diffusion, and surface reactivity, and therefore strongly influence astrochemical models of star- and planet-forming regions. While recent computational studies increasingly report full distributions of BEs rather than single representative values, these distributions are typically derived for either bare grain surfaces or thick water-ice mantles. In this work, we bridge these regimes by systematically investigating the BE distributions of water on partially and fully ice-covered dust grain surfaces. We employ machine-learning interatomic potentials (MLIPs) based on graph neural networks to model water adsorption on graphene and on the Mg-terminated (010) surface of forsterite, representing carbonaceous and silicate grains, respectively. The models enable extensive sampling of adsorption sites on water clusters, monolayers, and bilayers generated under both crystalline (thermally processed) and amorphous (low-temperature) growth conditions. At submonolayer coverage, the chemical nature of the underlying grain strongly affects both ice morphology and binding energies, with Mg-O interactions on silicate surfaces producing particularly deep binding sites. From monolayer coverage onward, adsorption on both substrates is dominated by hydrogen bonding within the ice, reducing the influence of the grain material. Across all coverages, amorphous ice structures systematically shift the BE distributions toward stronger binding compared to crystalline ice, introducing highly stable defect and pocket sites. These results demonstrate that BE distributions in the submonolayer to few-layer ice regime are broad and highly surface dependent, and they provide physically motivated input for next-generation astrochemical models incorporating surface heterogeneity.
Paper Structure (12 sections, 1 equation, 8 figures)

This paper contains 12 sections, 1 equation, 8 figures.

Table of Contents

  1. Introduction
  2. Methods
  3. Overview
  4. DFT settings
  5. PaiNN architecture and model settings
  6. Data acquisition
  7. Generation of surfaces
  8. Binding energy sampling
  9. Results and Discussion
  10. Effect of grain surface type
  11. Effect of ice structure
  12. Conclusions and outlook

Figures (8)

  • Figure 1: Examples of surfaces generated using global optimization, covering clusters, monolayers and bilayers on the (010) surface facet of forsterite and graphene. Water molecules that are part of the surface are shown with the O atom in red, and a water adsorbate used to probe the BE of an exemplary site is shown with the O atom in blue. Color code for other atoms: H (white), Mg (green, with darker green representing higher height), Si (brown) and C (grey).
  • Figure 2: Comparison of BE distributions of H2O on graphene and silicate substrates covered by (a) a H2O cluster and (b) a H2O monolayer. Ice structures were generated with global optimization.
  • Figure 3: Bonding analysis for (a) cluster and (b) monolayer on silicate for ice structures generated with global optimization. For each H2O adsorption configuration, the analysis considers whether a Mg-O bond is present and quantifies the number of hydrogen bonds. The solid gray line is the total density.
  • Figure 4: Structures of selected H2O adsorption configurations for cluster (left) and monolayer (right). Ice structures were generated with global optimization. The labels indicate the adsorption energy, whether a Mg-O bond is present, and the number of hydrogen bonds of type acceptor (A) and donor (D). The color code is as in Figure \ref{['fig:example structures']}.
  • Figure 5: Bonding analysis for (a) cluster and (b) monolayer on graphene for ice structures generated with global optimization. For each H2O adsorption configuration, the analysis quantifies the number of hydrogen bonds. The solid gray line is the total density.
  • ...and 3 more figures