Effect of Ice Charging on the Astrochemistry of Interstellar Sulfur Bearing Species on Amorphous Solid Water

arXiv:2607.09364 · astro-ph.GA · Submitted 2026-07-10 · Read on arXiv

Listen

Transcript

Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.

Vera: Next we'll be talking about the paper "Effect of Ice Charging on the Astrochemistry of Interstellar Sulfur Bearing Species on Amorphous Solid Water".

Jocelyn: The paper was written by T. Vorsselmans, I. Grubova, K. Verhagen, T. Guldentops, R. Buimer et al. from University of Antwerp and University of Montana.

Vera: Stay tuned as we take you through the paper and discuss its implications.

Title: Vera: We're looking at 'Effect of Ice Charging on the Astrochemistry of Interstellar Sulfur–Bearing Species on Amorphous Solid Water' by T. Vorsselmans and a team from the University of Antwerp.

Jocelyn: That title is quite a mouthful, Vera, but I'm guessing the 'ice charging' part is the real hook here?

Vera: It definitely is, because they're moving past the idea of ice being just a neutral, passive layer.

Jocelyn: Most people think of ice in space as just a frozen coating on a rock.

Vera: But these authors, Vorsselmans and his colleagues, are treating it as a charged, active surface.

Jocelyn: Is that because of cosmic rays or something?

Vera: Exactly, things like cosmic rays and energetic particles can leave these grains with an excess negative charge.

Jocelyn: Does this relate to that weird problem where we can't find enough sulfur in the gas clouds?

Vera: It's the sulfur depletion problem, which has been bothering us for decades.

Subrahmanyan: The missing sulfur is a massive headache for us because it's one of the essential elements for life, yet it's just... gone from our observations.

Jocelyn: So, are these authors suggesting the sulfur is just hiding on these charged ice mantles?

Subrahmanyan: That's the hypothesis they're testing, and it could fundamentally change how we track the chemical evolution of a molecular cloud.

Vera: They're specifically looking at how this charge might act as a trap for sulfur-bearing molecules.

Jocelyn: And they're using 'amorphous solid water,' which I assume isn't just a perfect ice cube?

Vera: No, it's much messier than that, with a huge variety of different binding sites across the surface.

Jocelyn: How does that messiness affect things like binding energy?

Vera: Binding energy is basically how much effort it takes to pull a molecule off the ice and back into the gas.

Jocelyn: So if the energy is high, the molecule stays put?

Vera: Exactly, and that's what's at stake here.

Jocelyn: I want to see if the actual data supports that idea of it being a trap.

Summary: Jocelyn: So, Vera, what did the team actually find when they ran these simulations on the different sulfur species?

Vera: The results were quite dramatic, especially for the sulfur atom and SO2.

Jocelyn: How dramatic are we talking?

Vera: For the sulfur atom, the binding energy jumped from about one thousand five hundred seventy-one Kelvin on neutral ice to over forty-three thousand Kelvin on the charged surface.

Jocelyn: That's a massive leap; it's like going from a light dusting to being glued down.

Vera: And SO2 saw a similar jump, going from around three thousand one hundred forty-five Kelvin to nearly thirty-five thousand Kelvin.

Subrahmanyan: The physical reason is that the SO2 molecule actually accepts an extra electron, becoming a negatively charged anion.

Jocelyn: What about H2S, since that's usually the big candidate for the sulfur reservoir?

Vera: H2S only saw a slight increase, from two thousand two hundred seventy-six Kelvin to about two thousand seven hundred seventeen Kelvin.

Subrahmanyan: The H2S anion is just too unstable to be a major player in this specific charged interaction.

Jocelyn: And OCS? I remember seeing that mentioned in the abstract as being a bit of an outlier.

Vera: OCS was mostly indifferent to the charge, staying around one thousand five hundred Kelvin, unless it lands right next to the electron.

Subrahmanyan: It's a very proximity-dependent situation for OCS, because it has to bend its structure to accept that electron.

Jocelyn: So the charge doesn't affect every molecule in the same way?

Vera: Not at all; it's completely dependent on the electronic properties of the molecule itself.

Jocelyn: They even showed spin density plots where you can see the extra electron sitting in a pocket between the H-atoms for H2S.

Vera: Yeah, it's a very specific way of visualizing where that charge actually lives in the complex.

Jocelyn: I'm curious how they actually managed to calculate these specific energy jumps with such precision.

Improvements: Vera: They used density functional theory, or DFT, specifically with the ORCA software to model these amorphous solid water clusters.

Jocelyn: Modeling amorphous ice sounds like a nightmare because it's so messy and irregular, doesn't it?

Vera: It is, so they used five different independent ice clusters to make sure their binding energy distributions were statistically robust.

Jocelyn: Did they have to double-check that their DFT math was actually accurate for these sulfur species?

Vera: They did a massive benchmark against much more expensive coupled-cluster calculations to pick the best settings for each molecule.

Subrahmanyan: That level of rigor is necessary because even a small error in binding energy can lead to massive mistakes in gas-grain models.

Jocelyn: So, if we've been using neutral ice in our models all this time, how much have we been missing?

Vera: We might have been missing the entire mechanism that locks sulfur into the solid phase.

Subrahmanyan: The paper basically describes a one-way valve where sulfur hits a charged grain and gets stuck almost immediately.

Jocelyn: That would certainly explain why our surveys aren't picking up the gas-phase abundance we expect.

Vera: They even modeled it as a random hit-and-stick process, which is how molecules actually land on grains in space.

Subrahmanyan: It's that site heterogeneity and the availability of dangling hydrogen atoms that makes the whole process so complex.

Jocelyn: It sounds like we need to overhaul our current chemical templates.

Vera: We definitely do, and this paper is a huge step in that direction.

Conclusion: Vera: We've covered a lot of ground with 'Effect of Ice Charging on the Astrochemistry of Interstellar Sulfur–Bearing Species on Amorphous Solid Water.'

Jocelyn: It really feels like a missing piece of the puzzle for the sulfur depletion problem.

Vera: It certainly does, and it's a reminder that even the 'empty' spaces between stars are chemically complex.

Subrahmanyan: If this holds up, we have to rethink the entire lifecycle of sulfur from diffuse clouds to protoplanetary disks.

Jocelyn: I'm looking forward to seeing the next generation of models that actually include these charged surfaces.

Subrahmanyan: The theoretical implications for how we interpret future JWST data are going to be massive.

Vera: Well, that's all the time we have for this one, but we'll be back with more soon.

Jocelyn: Goodbye for now!

University of Antwerp · University of Montana

astro-ph.GA

Submitted: 2026-07-10

Updated: 2026-09-10

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 87/100

The gist: The provided data consists of computational results detailing benchmarks for interstellar sulfur-bearing species, specifically focusing on SO 2 and related species (BE0, BE-1, and their difference

Key concepts

Ice Charging: Interstellar ice grains can gain an excess negative charge from cosmic rays and energetic particles. The authors argue that this charge makes the ice a 'charged, active surface,' which fundamentally changes how molecules interact with it and increases binding strength.
Sulfur Depletion Problem: This decades-old puzzle concerns the apparent lack of observable sulfur in gas clouds. Researchers test the hypothesis that instead of being missing, sulfur is locked away or trapped within solid ice mantles on dust grains.
Amorphous Solid Water: This refers to interstellar ice that lacks a perfect crystal structure. It is described as 'messier,' featuring a huge variety of different binding sites across its surface, which affects the energy required to remove molecules from it.

Terminology

Benchmark — 22× in this paper · explained in 2 episodes
Abundance — 11× in this paper · explained in 1 episode
Amorphous Solid Water — 8× in this paper · explained in 1 episode
Density Functional Theory — 5× in this paper
Density Functional Theory (DFT) — 4× in this paper · explained in 2 episodes
Accuracy — 3× in this paper · explained in 1 episode
Core — 2× in this paper · explained in 2 episodes
Analysis of Difference in Binding Energy — 1× in this paper
DFT Calculations — 1× in this paper · explained in 2 episodes
Trace — 1× in this paper · explained in 2 episodes

Summary

The provided data consists of computational results detailing benchmarks for interstellar sulfur-bearing species, specifically focusing on SO 2 and related species (BE0, BE-1, and their difference BE). The analysis utilizes various theoretical methods and basis sets to calculate electronic binding energies (BE) in Kelvin (K).

Computational Methodology:

The study employs a comprehensive set of computational models, as summarized in Table A.4. These methods include high-level correlated calculations such as CCSD(T), alongside Density Functional Theory (DFT) approaches: M06-2X, omega B97 M, BHANDHLYP, omega B97 X, B1LYP, X3LYP, REVPBE38, B3PW91, PBE0, revPBE0, and O3LYP.

The basis sets used throughout the analysis are consistently reported as aug-cc-pVTZ and def2-TZVPPD. Dispersion corrections are systematically applied using various models, including D3(BJ), VV10, D4REV, D4, and combinations thereof (e.g., D3(BJ)/VV10, D4/D3(BJ), etc.).

Benchmark Results for BE0 (K):

The calculated binding energy for the BE0 species shows a range of values depending on the method and basis set. For instance, using the omega B97 M functional with a specific dispersion correction, the calculated energy is 4365 K. The error associated with this calculation is reported as 23.66%.

Across different methods, significant variations are observed. For example, using the CCSD(T) method yields a value of 3632 K (with an error of 35.01%), while using the BHANDHLYP functional results in a value of 4932 K (with an error of 21.50%). The calculated energy for BE0 varies from a minimum reported value of 2205 K (using def2-TZVPPD/D3(BJ)) to a maximum reported value of 4989 K (using omega B97 X/).

Benchmark Results for BE-1 (K):

The binding energy for the BE-1 species shows values ranging from 29461 K to 46135 K. Using the omega B97 X/ method yields a value of 30827 K, corresponding to an error of 23.66%. The highest reported value for BE-1 is 46135 K (using def2-TZVPPD/ with the first dispersion correction listed), and the lowest is 28883 K (using /).

Analysis of Difference in Binding Energy (BE) and Error:

The difference in binding energy, BE, provides a critical comparison between the two species. The calculated values for BE range from 24036 K to 28101 K.

The error associated with the difference (Error) demonstrates a systematic trend, generally decreasing as the computational method changes. For instance, using omega B97 X/, the calculated BE is 26388 K, with an associated error of 2.16%. Conversely, using the initial combination (CCSD(T) with), the value is 25829 K, and the error is 8.80%.

The most accurate results, indicated by the lowest percentage errors, are observed in the later calculations for BE. Specifically:

Using B3PW91/, BE is 24560 K, with an error of-3.83%.
Using PBE0/, BE is 24567 K, with an error of-4.88%.
The final listed calculation yields BE = 24036 K with an error of-6.94%.

In summary, the data provides a detailed comparative analysis of electronic binding energies for interstellar sulfur-bearing species, demonstrating how various computational methods and dispersion corrections impact the calculated values for BE0, BE-1, and their difference (BE).

Improvements for AI systems

The data provided constitutes an exceptionally rich benchmark dataset in computational chemistry—specifically, a systematic comparison of various Density Functional Theory (DFT) functionals and dispersion corrections (D3, VV10) against high-level reference calculations (CCSD(T)).

Since the core challenge in this field is the enormous computational cost of obtaining accurate reference data, any AI improvement must focus on reducing the required sampling space or guaranteeing predictive accuracy with minimal computation.

Here are three specific, high-impact improvements for AI systems derived from this data:

Improvement: Development of a meta-learning regression model designed to predict the Systematic Error (BE) of any combination of computational parameters (Method, Basis Set, Dispersion Correction) relative to the gold standard (CCSD(T)).

Mechanism:

The input features for this model are not just categorical (e.g., B3LYP, aug-cc-pVTZ), but must be encoded features representing the physical characteristics of these methods (e.g., range-separation parameters, empirical scaling factors, functional form complexity). The target variable is the measured percentage error (Error).

We would employ an advanced architecture like a Graph Neural Network (GNN) or a specialized Ensemble Transformer trained on this dataset. The model learns the interaction effects between components (e.g., how D3(BJ) interacts with PBE0 specifically for SO 2).

What the Improved AI System Can Do:

Rapid Method Screening: Given a novel molecule and a required accuracy threshold (e.g., predict bond energy within 5% of CCSD(T) ), the DFTPP can instantly screen thousands of uncalculated functional/basis set combinations to output a ranked list of the most likely optimal method, saving months of supercomputer time.
Error Mapping: It can generate a quantitative Error Landscape Map for a given chemical class, identifying systematic failure modes (e.g., For SO 2 systems, DFT methods consistently underestimate the BE-1 by X% when using basis sets smaller than def2-TZVPPD).

Improvement: Implementation of an Active Learning loop guided by Bayesian Optimization (BO) to strategically select the next most informative calculation, thereby minimizing total computational cost while maximizing predictive convergence.

Improvement: Integrating Explainable AI (XAI) techniques, specifically SHAP (SHapley Additive exPlanations) values, to move beyond mere prediction and provide physical hypotheses for why a method succeeds or fails.

Abstract

We aim to derive statistically robust and physically interpretable BE distributions for atomic S and the sulfur-bearing molecules H 2 S, SO 2, and OCS on neutral and negatively charged ASW, and assess how excess negative charge alters their retention and potential role in the sulfur reservoir of cold dense molecular clouds. Basis set superposition error (BSSE) and zero-point energy (ZPE) corrected BEs of S, SO 2, OCS and H 2 S, are calculated by density functional theory (DFT), using the ORCA software. Molecule-specific DFT levels of theory were first selected from benchmark calculations on neutral and charged small water complexes, against coupled-cluster reference energies. The selected protocols were then applied to study adsorption on neutral and charged ASW clusters. A range of adsorption sites were sampled on five independent amorphous ice clusters, yielding BE distributions that account for the site heterogeneity of ASW. Neutral ASW yields broad, site-dependent BE distributions consistent with previous water-ice estimates. On charged ASW, three general cases are identified: the BE will either always increase due to electron transfer (S-atom, SO 2), the BE increases slightly without any electron transfer (H 2 S) or the BE remains the same unless an electron transfer occurs under specific conditions (OCS). These findings are inherently linked to the molecular properties of these molecules.