🔬 materials science

Revealing the innate sub-nanometer porous structure of carbon nanomembranes with molecular dynamics simulations and highly charged ion spectroscopy

By combining molecular dynamics simulations with highly charged ion spectroscopy, this study reveals that terphenylthiol-based carbon nanomembranes possess an innate, reactive sub-nanometer porous structure stabilized by atmospheric hydrogen and oxygen groups, overcoming the limitations of traditional characterization methods for these radiation-sensitive materials.

Original authors: Filip Vuković, Anna Niggas, Levin Mihlan, Zhen Yao, Armin Gölzhäuser, Louise Fréville, Vladislav Stroganov, Andrey Turchanin, Jürgen Schnack, Nigel A. Marks, Richard A. Wilhelm

Published 2026-07-14

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Filip Vuković, Anna Niggas, Levin Mihlan, Zhen Yao, Armin Gölzhäuser, Louise Fréville, Vladislav Stroganov, Andrey Turchanin, Jürgen Schnack, Nigel A. Marks, Richard A. Wilhelm

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a carbon nanomembrane (CNM) as a super-thin, invisible sheet of carbon, so delicate it's only about one nanometer thick—roughly the width of a few atoms stacked up. Scientists have been trying to peek inside these sheets to see their secret architecture, but it's like trying to look at a soap bubble without popping it. If you shine a powerful electron microscope beam on them, the energy zaps the structure, turning it into something else entirely (graphite), ruining the view.

To solve this puzzle, the researchers in this paper acted like cosmic detectives. Instead of blasting the membrane with destructive beams, they fired tiny, highly charged "bullets" (ions) at it. Think of these ions as super-energetic bees. When a bee flies through a dense forest, it gets bumped around a lot and loses its energy (charge). But if it flies through a clearing or a gap, it zips right through, keeping most of its energy. By watching how these "bees" came out the other side—how much energy they lost and how much they bounced off—the team could map the invisible holes inside the membrane without ever touching it.

The Big Reveal: A Swiss Cheese of the Sub-Nanometer World

The main finding suggests that these carbon sheets aren't solid, flat pancakes. Instead, they look more like a sub-nanometer Swiss cheese. The simulations, which act like a virtual microscope, indicate that the membrane is riddled with tiny, innate holes (porosity) that are smaller than a nanometer.

Here is the twist: The carbon atoms holding this structure together aren't all holding hands in perfect, stable groups. A significant chunk of them are "under-coordinated," meaning they are missing a few neighbors. In the vacuum of the lab, this creates a reactive, slightly unstable network. The authors suggest that once this membrane is taken out of the vacuum and exposed to our normal atmosphere, it likely grabs onto hydrogen and oxygen molecules (like water vapor) to stabilize itself, kind of like how a dry sponge soaks up water to become heavy and stable.

What They Ruled Out

The team was very careful to knock down some popular ideas.

No Perfect Graphene: They explicitly ruled out the idea that these membranes are just perfect, flat sheets of graphite or graphene. If they were, the ions would have behaved differently, and the math wouldn't have added up.
No "Solid" Sheets: They also argued against the idea that the membrane is a solid, dense block of carbon with no gaps. The data shows that ions are passing through open spaces, proving the sheet is porous.
No "Perfect" Models: They tested a method where they tried to build the membrane by simulating the exact process of turning a precursor layer into a film (using "momentum transfer"). While this created some holes, the resulting material was too stiff. The paper suggests that this specific simulation method didn't quite capture the real-world softness of the membrane, so they had to try a different approach.

How They Did It: The Virtual Lab

Since they couldn't see the atoms directly, they built a virtual world using Molecular Dynamics (MD) simulations. Imagine a giant digital sandbox where they dropped thousands of carbon atoms.

The "Exclusion Cylinder" Trick: To force the creation of holes, they used a digital tool they called "exclusion cylinders." Think of these as invisible, ghostly poles standing up through the sandbox. The carbon atoms were programmed to bounce off these poles, never allowed to step inside them. This forced the atoms to arrange themselves around the poles, creating a network with built-in gaps.
The "Momentum Transfer" Test: They also tried a second method, mimicking the real-world process where electrons hit the material and push atoms around. This was like giving the sandbox a series of gentle (and not-so-gentle) shoves to see if holes would form naturally.

The Numbers and the Confidence

The researchers didn't just guess; they compared their virtual results to real experiments.

They fired ions with specific energies: 72 keV, 135 keV, and 180 keV.
They found that the best match for the real-world data came from a model with 150 exclusion cylinders (creating a hole density of 0.84 holes per nm²) that had been "annealed" (heated and cooled in the simulation) for 9 ps (picoseconds).
They measured the "tensile modulus" (how stretchy or stiff the sheet is). The real membranes are between 6–12 GPa. Their best simulation models landed right in that 5–12 GPa range, but only if the structure had a lot of those under-coordinated carbon atoms and sub-nanometer holes.

The Verdict

The paper doesn't claim to have solved the entire mystery of carbon nanomembranes forever. Instead, it suggests that the most likely structure is a reactive, porous network that stabilizes itself when it hits the air. The authors propose that this "passivation" (getting covered by air molecules) is a crucial step that makes the membrane usable.

So, the next time you think of a carbon nanomembrane, don't picture a solid, unbreakable sheet. Picture a delicate, reactive spiderweb of carbon atoms, full of tiny, invisible holes, waiting to grab a breath of air to hold its shape together. It's a structure that is as much about what's missing (the holes) as what's there.

Technical Summary: Revealing the Innate Sub-Nanometer Porous Structure of Carbon Nanomembranes

Problem Statement
Carbon nanomembranes (CNMs) are nanometer-thin, disordered carbon materials with significant potential for applications in energy storage, filtration, and electronics. However, establishing the relationship between their structure and functional properties has been hindered by the difficulty of characterizing free-standing membranes. Traditional techniques like high-resolution transmission electron microscopy (HRTEM) induce graphitization due to beam damage, while scanning probe methods are limited to membranes on solid substrates. Furthermore, existing theoretical models, such as momentum transfer simulations, have struggled to reproduce the experimentally observed low tensile moduli (6–12 GPa) of CNMs, often predicting values that are significantly higher. There is a critical need for characterization methods that can probe the disordered, free-standing structure without inducing damage, and for theoretical models that can reconcile structural porosity with mechanical properties.

Methodology
The study employs a combined approach of highly charged ion (HCI) spectroscopy and molecular dynamics (MD) simulations to infer the atomistic structure of terphenylthiol-based CNMs.

Experimental Characterization: HCI transmission experiments were conducted at TU Wien using Xe ions with charge states up to $q=44$ and kinetic energies ranging from 72 to 180 keV. To address the sensitivity of CNMs, a cleaning protocol was developed using low-charge-state ions ( $q \le 10$ ) to remove contaminants via electronic excitation without damaging the membrane. The exit charge state distributions and scattering angles of transmitted ions were measured to probe membrane thickness, density, and porosity.
Molecular Dynamics Simulations: Two distinct MD simulation approaches were utilized to generate candidate structural models:
1. Exclusion Cylinder Method: This approach enforced a distribution of voids (exclusion cylinders) within the carbon network during structural evolution. By varying the number of cylinders (0 to 150) and simulation annealing times, the authors generated 81 unique CNM structures with varying degrees of sub-nanometer porosity and $sp^2$ fractions (40%–90%).
2. Momentum Transfer Method: An extension of previous work, this method simulated the SAM-to-CNM conversion by imparting discrete momentum impulses to a carbon-only terphenyl SAM, mimicking electron irradiation effects to induce pore formation without explicit geometric constraints.
Simulation of HCI Transmission: The generated CNM structures served as targets for time-dependent potential (TDPot) simulations. These simulations modeled the neutralization and scattering of HCI trajectories through the atomistic structures, predicting angle-resolved exit charge state distributions to be compared directly with experimental data.

Key Results

Structural Models and Mechanical Properties: The exclusion cylinder simulations revealed that achieving a tensile modulus consistent with experimental values (5–12 GPa) requires either a highly disordered structure with significant under-coordinated carbon ( $\gtrsim 30\%$ $sp$-like character) or a high density of pores that suppresses in-plane stiffness. High-annealing structures with high pore densities tended to form nanotube-like or fullerene-like morphologies rather than planar graphitic sheets.
HCI Spectra Comparison: The TDPot simulations successfully reproduced the bimodal charge state distributions observed in experiments. The low charge states corresponded to ions traversing denser regions of the membrane, while high charge states (and low scattering angles) corresponded to ions passing through voids or thinned regions.
Best-Fit Structure: The simulation results that best matched the experimental charge exchange spectra (specifically for 72 keV Xe $^{8+}$ , 135 keV Xe $^{15+}$ , and 180 keV Xe $^{20+}$ ) were structures with a high number of exclusion cylinders (150) annealed for a short duration (9 ps). These structures exhibited a significant fraction of under-coordinated carbon and an innate sub-nanometer porous network.
Pore Analysis: Pore detection algorithms applied to the simulated structures indicated that the "exclusion cylinder" method produced pore sizes consistent with gas permeation data (diameters $\sim 5.5$ Å). The momentum transfer simulations yielded broader pore distributions, suggesting multiple pore formation mechanisms, though these could not be fully resolved by current HCI data.

Significance and Claims
The paper claims to provide the first quantitative insight into the atomistic structure of CNMs by coinciding experimental HCI spectroscopy with MD simulations. The primary finding is that the in-vacuum composition of terphenyl-based CNMs likely comprises a significant fraction of under-coordinated carbon with an open sub-nanometer porous structure.

The authors assert that such a reactive carbon network would be unstable in atmospheric conditions and is presumably stabilized by hydrogen and oxygen groups (passivation) upon exposure to air. This suggests that the functional properties of CNMs are intrinsically linked to this post-irradiation passivation. The study demonstrates that HCI transmission spectroscopy is a viable, non-destructive method for probing the sub-nanometer porosity of free-standing membranes, overcoming the limitations of electron microscopy. By linking specific structural models to experimental charge exchange data, the work establishes a framework for understanding the structure-property relationships of organic short-range ordered membranes.

More like this