InterPore2026

19 May 2026, 09:00 → 22 May 2026, 18:00 Europe/Paris

Join us for fascinating lectures, engage with fellow researchers from across the globe and discover cutting-edge exploration of porous media. 

Please consider making a donation to InterPore Foundation in any amount to enable the participation of your fellow researchers.

Topics and applications

Transport phenomena     Swelling and shrinking porous media     Multiphysics-multiphase flow     Reservoir engineering     Soil Mechanics and Engineering     Geothermal energy     CO2 sequestration     Constitutive modeling     Wave propagation     Energy Storage

Biotechnology Biofilms Thin and nanoscale poromechanics Fuel cells and batteries Food Paper and textiles Filters, foams, membranes Fibers and composites Ceramics and constructions materials Other porous media applications

 

---

Event Management

InterPore2026 Flyer.pdf

InterPore2026 Sponsor Exhibitor Prospectus

InterPore2026 Terms and Conditions.pdf

Visa Letter InterPore2026.docx

Tuesday 19 May

Plenary Lecture: Plenary 1

1 3D and 4D X-ray imaging of the behaviour of porous systems

X-ray imaging can provide detailed structural information in 3D non destructively across scales ranging from tens of centimetre samples to tens of nanometres spatial resolution over timescales ranging from milliseconds to many months. This, and the fact that 3D image sequences can be collected non destructively, mean that it can uniquely shine a light on a range of porous materials behaviours from transport phenomena and permeability to fuel cells, from granular flow to cementitious materials, and from our perception of foods to the collapse of energy absorbing structures.

I will start with a primer on 3D and timelapse (4D) imaging for those new to the technique looking at the basic principles, the attributes and limitations of the method and its complementarity to other characterisation methods such as mercury intrusion porosimetry.

I will then examine a number of applications covering a very wide range of length and timescales and applications. In particular I will consider transport behaviour through homogeneous and inhomogeneous media, particle transport through filter cakes, the infiltration of fibrous preforms in polymer and ceramic matrix composite manufacturing, the behaviour of granular solids, the microstructure of 3D printed concrete and the long term carbonation behaviour of low carbon cements. Through these examples I will look at the practical limitations of the method, image quantification and segmentation aspects and also cover image-based modelling and digital volume correlation. I will then conclude by looking at future developments.

Speaker: Philip Withers

Poster: Poster I

MS01: 1.1

MS02: 1.1

MS05: 1.1

MS07: 1.1

MS08: 1.1

MS09: 1.1

MS16: 1.1

MS20: 1.1

MS01: 1.2

MS05: 1.2

MS08: 1.2

MS09: 1.2

MS10: 1.2

MS12: 1.2

MS15: 1.2

2 Decoupling the Non-linear Influence of Pore Structure on CO₂ Saturation: An Explainable Data-Driven Approach based on Microfluidic Experiments

Geological CO₂ sequestration efficiency relies on pore-scale structural parameters. However, the complex, non-linear coupling among these parameters is difficult to quantify using traditional experimental correlations alone. In this study, we apply an explainable machine learning (ML) framework to uncover the dominant governing factors of CO₂ saturation, utilizing a high-fidelity dataset derived from our systematic microfluidic displacement-imbibition experiments. The dataset encompasses a wide range of topological scenarios, where pore-size distribution, pore-throat ratio, and coordination number were independently varied under different capillary numbers. We developed a multi-modal deep learning model that integrates Convolutional Neural Networks (CNN) for extracting topological features from experimental images and Multi-Layer Perceptrons (MLP) for processing numerical structural parameters. This hybrid architecture maps the inputs to initial and residual CO₂ saturation, achieving high predictive accuracy (R² ≈ 0.95) and robust stability across cross-validation folds (standard deviation < 0.05). Crucially, to move beyond "black-box" prediction, we employed SHAP (SHapley Additive exPlanations) analysis to decouple the interactions between topological features. The analysis reveals that pore-size distribution characteristics and structural heterogeneity are the primary predictors, exhibiting a non-linear influence that standard linear regression fails to capture. Furthermore, the ML-derived feature importance aligns with the physical mechanism of capillarity-connectivity competition, confirming that the coordination number and pore-throat ratio jointly dictate the capillary-viscous transition. This work demonstrates that applying explainable AI to experimental datasets provides a robust pathway for identifying critical sequestration criteria in heterogeneous porous media.

Speaker: 晗 葛 (浙江大学)

3 A machine learning method to automatically segment solid and multiple fluid phases in time-dependent 3D (4D) images

Capturing dynamic processes like pore-filling and snap-off using fast synchrotron X-ray micro-tomography enables time-resolved quantitative and qualitative analysis. However, time-resolved imaging often generates noisy, low-contrast images, and the resulting datasets are often large. These factors present challenges for effective and accurate 4D image segmentation. Frame-by-frame segmentation methods treat each time step as an independent 3D image without considering temporal consistency, which often results in flickering and physically implausible interface evolution.

To address this, we present Spatio-Temporal SwinUNETR (ST-SwinUNETR), a deep-learning technique that segments 4D images by modelling space and time jointly. We validate the method on dynamic synchrotron micro-CT datasets and evaluate performance using both image-based and physics-based criteria, including porosity and phase saturation over time. ST-SwinUNETR improves spatial accuracy while enhancing temporal consistency of the predicted segmentations over time.

Speaker: Zhuangzhuang Ma

4 Beyond Spectral Bias in Geothermal Heat Transport: A Comparative Analysis of Fourier Neural Operators and DeepONet Architectures in Heterogeneous Media

Deep geothermal energy exploitation relies heavily on predicting heat transport within highly heterogeneous porous formations. The multi-scale nature of subsurface geology (ranging from pore-scale variances to reservoir-scale fracture networks) coupled with the non-linear interaction between Darcy flow and advective-diffusive heat transfer, renders traditional numerical solvers computationally prohibitive for real-time optimization and many-query uncertainty quantification. Scientific Machine Learning, specifically Operator Learning, offers a promising path to overcome this bottleneck by learning mesh-independent solution operators.

In this work, we present a rigorous analysis of two leading operator learning paradigms: Fourier Neural Operators (FNO) and Deep Operator Networks (DeepONet). We utilize a high-fidelity benchmark of dipole flow and heat transport simulations in stochastically generated heterogeneous media to evaluate the capacity of these architectures to act as reliable surrogates for the management of geothermal reservoirs. We present an exhaustive comparative analysis of both architectures, focusing not just on global error metrics, but on the spatial and spectral distribution of residuals. Furthermore, we investigate the internal mechanisms of both models to understand their respective failure modes. By explicitly mapping how each architecture encodes physical heterogeneity, we propose novel strategies to mitigate spectral bias, enabling hybrid architectures that reconcile global spectral efficiency with the local resolution necessary for robust geothermal digital twins.

Our primary contribution is the demonstration and quantification of ``spectral bias'' in standard FNO architectures. While FNOs exhibit exceptional performance in diffusion-dominated regimes, our results reveal a structural inability to resolve high-frequency spatial features in advection-dominated scenarios. Specifically, the intrinsic frequency truncation in FNO layers acts as a low-pass filter, leading to significant localized errors around singularities (injection/production wells) and sharp thermal fronts. This smoothing effect compromises the physical fidelity required for operational decision-making in geothermal doublets.

Comparatively, Deep Operator Networks (DeepONet) utilize a dual Branch and Trunk structure that learns an adaptive basis, theoretically offering superior resolution for local singularities compared to the fixed Fourier basis of FNOs (Lu et al., 2021). However, recent analyses indicate that while DeepONets offer geometric flexibility, standard MLP-based trunks suffer from their own spectral bias, leading to slower convergence when resolving multimodal global fields compared to spectral methods (Wang et al., 2021; Rahaman et al., 2019). To reconcile these trade-offs, we propose a novel hybrid strategy inspired by recent `global-local' operator learning paradigms (Wen et al., 2022; Jiang et al., 2024). Our approach integrates FNOs to efficiently resolve the dominant global transport dynamics with a localized DeepONet correction module, specifically targeted to capture the high-frequency residuals at injection wells. This architecture aims to bridge the computational speed of spectral methods with the physical fidelity required for high-Péclet geothermal reservoir management.

Speaker: Antonio Ortiz Romero (IDAEA-CSIC)

5 Investigating Machine Learning Models for Pore-Scale Multiphase Flow Using Lattice Boltzmann Simulations

Machine learning is increasingly used to accelerate or replace pore-scale simulations of multiphase flow in porous media, yet a clear understanding of how different model classes perform—and fail—under realistic flow conditions remains limited. In particular, it is often unclear which modelling choices are most appropriate for capturing interfacial dynamics, geometry–flow coupling, and temporal evolution at the pore scale.

In this study, we present a systematic benchmark of representative machine learning architectures for pore-scale multiphase flow prediction using datasets generated from lattice Boltzmann method (LBM) simulations on micro-CT–derived pore geometries. The datasets span a range of flow conditions and multiphase configurations while retaining full access to velocity fields, phase distributions, and geometric information. This enables controlled evaluation of model behaviour under well-defined physical settings.

We compare several classes of ML models, including convolutional neural networks operating on voxelised domains, graph-based representations of pore networks, and autoregressive temporal models. Performance is assessed in terms of short-term prediction accuracy, stability under multi-step rollout, sensitivity to pore geometry, and generalisation across flow regimes. Beyond aggregate error metrics, we examine qualitative failure modes, such as loss of interfacial sharpness, accumulation of long-term drift, and reduced robustness near rapid interface rearrangements.

Based on insights from these benchmarks, we explore modest extensions to existing training strategies, including geometry-aware conditioning and rollout-consistent supervision, aimed at improving stability and interpretability rather than maximal accuracy. The results provide a practical reference for selecting and training ML models for pore-scale multiphase flow, and clarify the trade-offs involved when using data-driven surrogates alongside conventional LBM simulations.

Speaker: Chunyang Wang (Imperial College London)

6 Operator Learning for Multispecies Reactive Transport in Heterogeneous Media

Coastal protection structures face increasing challenges from sea level rise, extreme weather events, and material degradation [1]. Innovative approaches such as cathodic protection, which induces the precipitation of calco-magnesian deposits (e.g., brucite and aragonite), provide promising strategies for reinforcing marine infrastructures in a sustainable manner [2-3]. However, the long-term durability of these deposits depends strongly on their evolving porosity, mineral balance, and transport–reaction processes [4]. Accurate prediction of this evolution is therefore crucial for assessing future stability of coastal protection measures.
Simulating these processes requires solving coupled nonlinear partial differential equations describing multispecies diffusion-reaction kinetics in heterogeneous media. Traditional direct numerical approaches, such as finite-difference, finite-volume, finite-element, or spectral methods, deliver accurate results but are computationally expensive, particularly when applied to long-time evolution, parameter studies (with significant number of species) [5].
To address these limitations, recent advances in operator-learning neural networks offer surrogate modeling approaches that learn mappings between function spaces and enable generalization across heterogeneous porous media [6–9]. In this work, operator learning is investigated as a surrogate modeling strategy for multispecies reactive transport in heterogeneous media. The governing dynamics are described by coupled diffusion–reaction equations with spatially varying transport properties derived from heterogeneous porosity fields. Numerical simulations are used to generate reference datasets capturing nonlinear spatiotemporal concentration dynamics under reaction-dominated regimes. A class of neural operator models is evaluated, with integral-transform-based formulations operating in latent spaces, where the learned integral operators may be linear or nonlinear [9]. These models are trained to advance the system in time and subsequently rolled out for temporal prediction.
Using this framework, model performance is assessed through quantitative error metrics and qualitative comparisons of spatial concentration patterns over extended time horizons. The results indicate that integral-transform-based operator models can achieve performance comparable to Fourier-based neural operators for multispecies reactive transport problems. In particular, learned integral kernels provide additional flexibility for representing high-frequency spatial features, which are not efficiently represented by Fourier-based neural operator models. Similar performance is observed with a reduced number of model parameters, motivating continued efforts toward improving accuracy and stability through architectural and training refinements.
Taken together, these results indicate that integral-transform-based operator learning may complement existing neural operator models in reactive transport applications.

Speaker: Mrs Fatima Tokmukhamedova (La rochelle University)

MS16: 1.2

Poster: Poster II

Invited Lecture: Invited I

7 Conceptual challenges to model biochemical processes in aquifers

Pollution is arguably the worst environmental global challenge faced by society. While pollution problems are local in nature, current policies (e.g., spill it rivers the outflow of wastewater treatment plants, WWTPs) favor the broad spread of pollutants, to the point that many of them are becoming global threats (e.g., antimicrobial resistance, endocrine disruptors, microplastics). While this error is widely acknowledged by the scientific community, a misunderstood precautionary principle prevents the use of soil aquifer treatment to remove these pollutants from WWTP effluents. In this presentation, I will expand on the importance of pollution as a global challenge, on why regulations concerns are unfounded, and especially on the processes that govern pollutants removal. These processes rely on the fact that porous media display large specific surface areas. These surfaces host biofilms that tend to absorb many organic pollutants (specifically those that are toxic by accumulation), precisely in the locations hosting the microbial communities that degrade those pollutants (we conjecture that this is the result of natural selection). Retention and degradation processes are improved by the addition of a reactive layer, as it favors the growth of biofilms, ad the adsorption of ionic compounds. Research challenges of these processes are significant. Degradation occur at the sub-pore scale in biofilms, which host the vast majority of microorganisms that catalyze them. Therefore, solute transport and reactions become controlled by diffusive processes. But aquifer scale transport is controlled by dispersive processes and heterogeneity. The issue is further complicated by biofilm growth, which changes porosity and permeability and large-scale heterogeneity, as well as localized residence and reaction times in different portion of the medium. Addressing these pore scale processes, while acknowledging their impact on aquifer scale heterogeneity is challenging, but needed to understand how these “new” pollutants are removed during soil passage, in turn needed to help convince regulators.

Speaker: Jesús Carrera (IDAEA)

Invited Lecture: Invited II

8 Porous pathways to improve food functionality and sustainability

The complexity of food arises not only from their multicomponent chemical nature but also from the diverse molecular and supramolecular arrangements that form a complex matrix comprising both matter and voids. Porous regions, distributed across nano-, micro-, and macro-scales, are not merely empty spaces but critical features that influence food functionality and sustainability. Food porosity significantly increases surface area, driving chemical and biological reactivity at interfaces and enhancing the release or absorption/adsorption of food liquids (e.g., water, oil), volatile compounds (e.g., flavors, antioxidants), and bioactive molecules (e.g., vitamins and other micronutrients). The size, shape, and connectivity of food pores can affect food performance throughout its lifecycle—from processing and storage to final consumption and digestion in the gut—impacting food acceptability, sensory perception, nutrient release during digestion, shelf life, and the efficient use of natural, often plant-based, resources.

Although many foods with macroporosity have traditionally been produced through processes such as fermentation, frying, puffing, or extrusion, the development of novel micro- and nano-structured porous materials with diverse potential functionalities has only recently emerged. This progress is largely driven by the ability to produce cryogels and aerogels. Cryogelation exploits the pore-forming action of ice crystals during freezing, while aerogelation involves replacing the liquid phase in a biopolymer gel or biological tissue with air—often through supercritical carbon dioxide drying.

This presentation initially focuses on the basic approach for preparing highly porous food-grade materials from proteins (whey, pea and soy), polysaccharides (carrageenan, cellulose) and food residues (whey and plant residues). It then explores a range of advanced food applications for porous materials – used as monoliths or particles -including smart ingredients controlling nutrient release, delivery systems for active compounds, oil structuring agents to develop fat substitutes, sensory experience modulators, cell-growth scaffolds, and novel biodegradable and intelligent food packaging materials. These examples serve to analyse current research challenges and prospect future market opportunities.

Speaker: Lara Manzocco (University of Udine)

MS01: 1.3

MS03: 1.3

MS06: 1.3

MS10: 1.3

MS13: 1.3

9 Humidity-driven crystallization and deliquescence of salt in nanopores

Variations in relative humidity (RH) can drive phase transitions of salts: crystallization upon water evaporation, and deliquescence (spontaneous crystal dissolution) upon RH increase. In porous materials, these phenomena play a central role in various applications, e.g., in heritage preservation, civil engineering, energy conversion/storage, or water management. While bulk deliquescence and crystallization are well understood in bulk situations, understanding the impact of confinement on these transitions remains challenging, especially in nanoscale pores.

Here, we systematically investigate how sodium chloride (NaCl) solutions confined in synthetic mesoporous materials (3 to 20 nm in diameter) respond to controlled RH cycles, as a function of pore size and salt concentration. Using these model materials, we observe large, well-defined and reproducible shifts of the deliquescence and crystallization points relative to the bulk, which are more pronounced as the pore size is reduced. We rationalize our observations using a theoretical model coupling nanoscale capillary effects (Kelvin equation) with osmotic contributions and classical nucleation theory. Our results, while fundamental, also suggest design rules for composite materials with controllable water content as a function of RH, or tunable crystallization and dissolution conditions for the salt.

Speaker: Olivier Vincent (CNRS & Univ. Lyon 1)

10 Elastocapillary Fingerprints of Distinct Drying Regimes in Nanoporous Media

Drying of porous media proceeds through distinct dynamical regimes that reflect the evolving morphology of the pore-scale liquid distribution. Here we combine high-resolution dilatometry with gravimetry and optical imaging to resolve the coupled mechanical and transport response of nanoporous Vycor during water desorption. We show that the macroscopic strain encodes a quantitative elastocapillary fingerprint of the classical constant-rate and falling-rate drying regimes, enabling direct inference of internal capillary pressures and morphological transitions that remain hidden in conventional mass-loss measurements.
These findings connect naturally to our earlier studies which focused on spatially resolved magnetic resonance imaging of drying in Vycor [1] under controlled air-flux boundary conditions as well as imbibition-induced deformation of nanoporous media [2]. These measurements revealed homogeneous or gradient-driven desaturation profiles and validate a diffusion-like transport model derived from Kelvin-law-controlled liquid pressure gradients. Together, the two approaches establish a unified framework linking pore-scale transport, macroscopic strain, and predictive drying models for functional nanoporous materials.

[1] Diffusionlike Drying of a Nanoporous Solid as Revealed by Magnetic Resonance Imaging - B Maillet, G Dittrich, P Huber, P Coussot, Physical Review Applied 18 (5), 054027 (2022).

[2] Deformation dynamics of nanopores upon water imbibition -
J Sanchez, L Dammann, L Gallardo, Z Li, M Fröba, RH Meißner, HA Stone, P. Huber, Proceedings of the National Academy of Sciences 121 (38), e2318386121 (2024).

Speaker: Prof. Patrick Huber (Hamburg University of Technology and Deutsches Elektronen-Synchrotron DESY)

11 Moisture transport through nanoporous clay

Moisture transfers in clayey soils or earthen construction materials play an essential role on the integrity of the structures and the regulation of humidity of the environment. Concentrated clay systems are nanoporous materials through which moisture transfers can involve vapor or liquid water transport. Here, with the help of NMR relaxometry and MRI allowing to distinguish the different liquid populations in the medium, we provide a detailed description of the different stages of extraction of water from a compacted clay sample during drying. Free water is extracted first at a constant rate, driven by capillary effects. In the next stage the moisture transport results from the flow of adsorbed water films, along with vapor transport through the porosity and exchanges between the two populations, a scheme somewhat similar to that presented for cellulosic materials [1]. We show that the transport diffusion coefficient of the adsorbed water films alone may be determined through drying experiments of the sample with its porosity filled with oil, while the water vapor diffusion coefficient may be determined from the permeability to ethanol vapor (i.e., with limited interactions with the solid phase). The total moisture transport can then be described by a diffusion equation with a diffusion coefficient depending on these two coefficients and the sorption curve. This model, relying on parameters determined from independent tests, finally appears to well describe the characteristics of standard drying tests. The tools developed in this work can be generalized to any solid clayey system, the main parameters of the model varying with the porosity and clay type.

Reference
[1] Y. Zou, B. Maillet, L. Brochard, and P. Coussot, Unveiling moisture transport mechanisms in cellulosic materials: Vapor vs. bound water, PNAS nexus 3, pgad450 (2024)

Speaker: Philippe Coussot (Univ. Paris-Est)

12 NMR T_{1-T2} mapping of fluid mobility and pore structure alterations in Mowry shale

Unconventional shale reservoirs have become increasingly important in sustaining oil production in response to increasing energy demand. In the northern Rocky Mountain region, the Mowry shale is recognized as a key Cretaceous source for oil and gas. Fluid flow in shale porous media is strongly governed by pore architecture, which stimulation fluids can alter, potentially influencing fluid transport. Nuclear Magnetic Resonance (NMR) provides a robust framework to evaluate pore structure and fluid mobility. In this work, one- and two-dimensional NMR were used to assess fluid mobility in Mowry shale samples reacted with stimulation fluids of different ionic strengths. Additionally, Fast Field Cycling NMR (FFC-NMR) data were collected to evidence rock surface alteration. Mowry shale from northern Wyoming was crushed into chips of approximately 0.85 mm and saturated with synthetic formation water at reservoir temperature (84°C) and pressure (1100 psi). The aqueous phase had an ionic strength of 0.9080 mol/L and a pH of 6.3. Subsequently, the rock samples were reacted with stimulation fluid obtained by diluting the original formation water to 12.5, 25, 50, and 75% of its original ionic strength, with one sample remaining unreacted as a control. Fixed field T1 and T2 measurements were performed before and after exposure, as well as T1-T2 relaxation maps. Three distinct relaxation peaks were obtained for T1 at approximately 1, 100, and 500 ms and T2 at 1, 50, and 200 ms across all samples. The two shortest peaks were interpreted as representing two dominant pore-size domains, whereas the longest peak is associated with bulk fluid surrounding the samples, consistent with previous measurements on fully saturated plugs. Following exposure to the stimulation fluids, the two long-time peaks shifted toward longer relaxation times, suggesting modifications associated with the larger pore domains. T1-T2 relaxation maps were generated to illustrate changes in the samples after stimulation, providing qualitative indications of variations in fluid mobility within the pore space. Relaxation rates, via FFC-NMR, before and after exposure to stimulation were used to confirm rock alteration. This work aims to understand the effect of stimulation fluid on the relaxation times of the Mowry shale.

Speaker: Mrs Johanna Romero (University of Wyoming)

MS14: 1.3

MS19: 1.3

MS20: 1.3

Convener: Oleg Iliev

Wednesday 20 May

Invited Lecture: Invited III

13 Sol-Gel Chemistry with a Twist: Porous Materials from Unconventional Precurs

The design of porous materials with well-defined architectures is a central challenge in materials chemistry, since pore size, connectivity, tortuosity, and shape strongly determine their potential applications in catalysis, separation, energy storage, and sensing.

Conventional sol-gel approaches often lack the versatility to achieve such deliberate structural control, motivating the development of new synthetic strategies. In this contribution, we present sol-gel processing routes towards highly porous monoliths based on unconventional, glycolated precursors such as tetrakis(2-hydroxyethyl)orthosilicate, organically substituted and related metal derivatives.

The replacement of classical alkoxy groups by diols/ polyols alters the reactivity of the precursors, enabling new pathways to tailor porosity, surface chemistry, and material composition, while also introducing specific synthetic challenges. In combination with co-monomers, these systems provide access to functional and structurally complex networks that extend the scope of sol-gel chemistry. By highlighting both opportunities and limitations of these non-traditional precursors, this work outlines new perspectives for the rational design of porous materials with controllable architectures and advanced functionalities.

Speaker: Nicola Hüsing (Universität Salzburg)

Invited Lecture: Invited IV

14 Immiscible two-phase flow in geological fractures

In crystalline rocks of the Earth’s crust, most fluid flows are accommodated by networks of interconnected fractures. Immiscible two-phase flow in such geological fractures is relevant to various industrial contexts, including subsurface fluid storage and hydrocarbon recovery. The fractures are natural objects resulting from thermally- or mechanically-inducted fracturing of a geological formation, followed by mechanical and/or (bio-)chemical weathering over millions of years. Their geometry possesses an inherent stochastic disorder that is well-characterized statistically; the wall roughness is usually Gaussian-distributed while exhibiting a self-affine scale invariance, and the two walls’ topographies are matched with each other at length scales larger than a characteristic ‘correlation’ length.

As in porous media, primary displacement of a resident fluid by an injected one in such geometries is controlled by the joint effect of viscous forces, capillary forces arising from surface tension effects at fluid-fluid interfaces, and gravity. However, capillary forces act in a different manner in fractures as compared to porous media, because in porous media the two principal curvatures of fluid-fluid interfaces are constrained by the medium’s structural heterogeneity, whereas in fractures only the out-of-plane curvature is; the in-plane curvature, in contrast, depends on the history of the displacement.

We use a combination of numerical simulations and analogue experiments to study such displacement in geological fractures, focusing on configurations for which the injected fluid is non-wetting. The numerical simulations adopt a volume-of-fluid approach to either describe the three-dimensional (3D) flow in the fracture’s volume, or directly model the depth-averaged 2D flow along the fracture plane, the latter approach being much more computationally-efficient. The experiments rely on transparent rough walls obtained from realistic synthetic geometries; their position with respect to each other can be adjusted to modify the relative fracture closure. Various morphological features of the fluid phases’ occupation patterns in the fracture plane, as well the pressure drop across the fracture, are analyzed to characterize the flow regimes as a function of three geometric parameters, the viscosity ratio of the fluids, the capillary number and/or Bond numbers, and an additional, novel, non-dimensional number. Phase diagrams are proposed for such primary two-phase flows in geological fractures. Flow configurations which maximize trapping of the displaced fluid are also determined.

Speaker: Prof. Yves Méheust (Geosciences Rennes, CNRS SCTD, 2 rue Jean Zay, 54519 Vandoeuvre les Nancy)

MS01: 2.1

MS02: 2.1

MS05: 2.1

MS09: 2.1

MS13: 2.1

15 Coupled dynamics of imbibition, evaporation and precipitation in nanoporous media

Spontaneous imbibition driven by capillary forces in nanoporous media underpins a wide range of natural and engineered processes, including water transport in plants and soils, oil recovery in rocks, drug delivery, and nanofabrication. Classical porous-media theories predict that evaporation limits imbibition by establishing a dynamic balance between capillary inflow and evaporative outflow, that precipitation blocks pore connectivity and suppresses further liquid advance. Here, we show that imbibition in nanoporous media can depart markedly from these classical expectations. Using a combination of in situ characterization techniques, multiscale imaging, and modeling, we investigate the coupled roles of capillary flow, evaporation, and precipitation in governing liquid transport. Our results reveal previously unrecognized mechanisms controlling imbibition dynamics and interfacial evolution in nanoporous systems, providing new insight into fluid transport in porous media and suggesting opportunities for manipulating capillary-driven flows in energy, environmental, materials, and biomedical applications.

Speaker: Mr Bin Pan (China University of Petroleum (Beijing))

16 Water Dynamics in Porous Materials: What can we learn from Quasielastic Neutron Scattering?

Water confined in nanoporous materials is ubiquitous in many applications related to energy and environment. This includes porous solids for water purification, solid electrolytes, membranes for proton exchange fuel cells, nanofluidic devices and desalinization technology.
Under these conditions, the structure and dynamics of water molecules is significantly altered with respect to the corresponding bulk state.1,2 This is a direct consequence of spatial restriction and liquid-surface interactions which become more prominent the smaller the pore size is. These effects obviously depend on the pore surface chemistry and the morphology (shape) of the material porosity. Interestingly, the water dynamics also depend on the length scale that is probed. For instance, different translational diffusion can be expected if it is monitored along a trajectory that is smaller than the pore size, that exceeds the diameter or even the grain size of nanoporous powder.
To resolve this problem, a multi-scale experimental approach is an asset. In the present communication, we will discuss the opportunity offered by quasielastic neutron scattering methods to characterize the dynamics of confined water at the nanoscale, that is to say for a molecular displacement equal to or less than the pore size, which can therefore be considered as a complementary tool to NMR that accesses longer scales. Our talk will be illustrated by recent studies carried out on water-filled porous silicas and organosilicas with various surface chemistry, which allowed fine tuning of the surface hydrophilicity and ionic charge and results in significant change in the liquid water local dynamics.3,4

References

1 B. Malfait, A. Jani, J. B. Mietner, R. Lefort, P. Huber, M. Fröba, and D. Morineau, J. Phys. Chem. C, 125, 16864 (2021)
2 B. Malfait, A. Moréac, A. Jani, R. Lefort, P. Huber, M. Fröba, and D. Morineau, J. Phys. Chem. C, 126, 3520 (2022)
3 A. Jani, M. Busch, J. B. Mietner, J. Ollivier, M. Appel, B. Frick, J.-M. Zanotti, A. Ghoufi, P. Huber, M. Fröba, and D. Morineau, J. Chem. Phys., 154, 094505 (2021)
4 A. Mozhdehei, P. Lenz, S. Gries, S.-M. Meinert, R. Lefort, J.-M. Zanotti, Q. Berrod, M. Appel, M. Busch, P. Huber, M. Fröba, D. Morineau J. Phys. Chem. C, 129, 18311−18324 (2025)

Acknowledgements
Funding by ANR (FIDELIO ANR-22-CE50-0002), ANR-DFG (SolutinPore ANR-23-CE29-0028) and DFG, project number 492723217 (CRC 1585) is acknowledged.

Speaker: Dr Denis Morineau (CNRS - Institute of Physics of Rennes)

17 A new experimental protocol to investigate adsorption-transport coupling in microporous materials

Gas transport in porous materials is typically described using flow models that assume a fixed pore structure and constant transport properties [1-4]. However, in materials where gas adsorption induces deformation, such assumptions become invalid [5-6]. In microporous materials, adsorption-induced swelling may alter pore geometry, transport porosity, and permeability, resulting in a significant coupling among adsorption, deformation, and flow that is still poorly characterized experimentally.
In this study, we proposed a novel experimental protocol to characterize adsorption effects on the internal pressure of Illite clay samples during CO2 transport. The original combination of three axial permeameters enables the measurement of the internal pressure evolution during CO2 adsorptive transport, in comparison with the inert transport of helium gas. Helium transport exhibits linear pressure propagation consistent with Klinkenberg’s theory, whereas CO2 produces systematic deviations that increase with pressure and distance along the sample. These deviations reflect progressive changes in transport properties caused by adsorption-induced swelling and lead to a redistribution of the internal pressure field.
This work demonstrates that adsorption-deformation coupling is not a secondary effect but a controlling mechanism for gas transport in microporous materials and must be explicitly included in predictive modeling frameworks.

Acknowledgements:
This work was funded by the Investissement d’Avenir French programme (ANR-16-IDEX-0002) within the framework of the E2S UPPA hub Newpores and by the Institut Universitaire de France.

References:
[1] H. Darcy, Les fontaines publiques de la ville de Dijon, Dalmont, Paris (1856).
[2] J. L. M. Poiseuille, C. R. Acad. Sci., 11, 961–967 (1840)
[3] W. Steckelmacher, Rep. Prog. Phys., 49, 1083–1107 (1986).
[4] L. J. Klinkenberg, API drilling and production practice, American Petroleum, 200–213 (1941).
[5] L. Perrier, F. Plantier, D. Grégoire, Rev. Sci. Instrum., 88, 035104 (2017).
[6] L. Perrier, G. Pijaudier-Cabot, D. Grégoire, Int. J. Solids Struct., 146, 192–202 (2018).

Speaker: David GREGOIRE (UPPA/ISABTP/LFCR, France)

18 A Novel Situ Gas Content Measurement Method for Deep Coalbed Methane Reservoirs Using Pressure Build-Up Analysis

As a new type of unconventional resource, deep coalbed methane reservoir has demonstrated generally significant development potential. It is characterized by rapid gas breakthrough, high gas production rates, and high estimated ultimate recovery (EUR) per well. Given that gas content is a key parameter for reserve assessment and development planning, it is crucial to establish a novel gas content measurement procedure especially for situ deep coalbed methane reservoirs. However, the testing results from conventional USBM method often deviate significantly from actual well production performance and fail to do accurate evaluation. To address this limitation, a novel in situ gas content testing method was proposed in this paper for deep coalbed methane. Specifically, three parts were included in this method: Firstly freshly retrieved core samples were promptly placed into a high-pressure vessel for simulating reservoir temperature condition, and methane was injected to restore the pore pressure to the formation pressure, thereby closely replicating the downhole temperature and pressure conditions. Subsequently, an isobaric displacement experimental procedure was established, in which water is injected to displace the annular gas between the core and the high-pressure vessel. It was ensured that all measured gas originates solely from the core itself. Next, rapid valve switching is performed to achieve a slight pressure reduction, followed by pressure re-equilibration. Based on the experimental data, a mathematical model for charaterizeing gas flow within matrix-fracture system was developed to calculate the free gas pore volume under reservoir conditions. By combining the measured total gas content with the derived free gas volume, the contributions of free and adsorbed gas were accurately determined. Ultimately, we conducted gas content evaluation comparation between the proposed method and conventional approaches. Based on field pressure preserved coring data, the proposed method yields a 50 % higher total gas content, with the free gas fraction reaching nearly 50 % greater than values obtained by other conventional methods. The method presented in this study significantly enhances the accuracy of in situ gas content measurement by closely replicating original formation pressure and temperature conditions. It thereby establishes an experimental framework for precisely evaluating the proportions of adsorbed and free gas under actual reservoir conditions.

Speaker: Prof. Wei Xiong (Research Institute of Petroleum Exploration and Development，PetroChina)

MS14: 2.1

MS19: 2.1

MS20: 2.1

Poster: Poster Session III

MS02: 2.2

MS03: 2.2

MS04: 2.2

19 How laboratory experiments can help to understand the dependencies of microbial activity during hydrogen storage on the various environmental aspects of porous rock formations

The underground storage of hydrogen (H$_{2}$) in porous rock formations offers a possibility for large-scale energy storage. However, hydrogenotrophic microorganisms can oxidize hydrogen through various metabolic processes e.g. sulfate or iron reduction, methanogenesis or acetogenesis. Since microorganisms can occur naturally or may be introduced through operational processes at the storage site, microbial processes must be considered when storing hydrogen in geological formations. In addition to hydrogen loss, microbial oxidation of hydrogen can also lead to other undesirable reactions, such as the formation of hydrogen sulfide, methane, organic acids, biofilms or corrosion. These reactions can affect the quality of the hydrogen as well as the storage performance.
Since the activity of microorganisms is determined by the in situ environmental conditions, it is also essential to understand the dependencies of microbial activity during hydrogen storage on the geochemical and mineralogical properties of porous rock formations in order to assess the potential effects of microbial activity during hydrogen storage.
Laboratory experiments simulating hydrogen storage with fluids from porous rock reservoirs showed hydrogen consumption, underlining the possibility of microbial activity during hydrogen storage (Dohrmann & Krüger 2023). In addition, experiments with pure cultures in batch incubation with minerals can help to better understand how microbial activity may be affected by porous rock material. Recent laboratory experiments in batch cultures have shown that hydrogen consumption by the methanogenic archaeon Methanothermococcus thermolithotrophicus was enhanced in the presence of rock material (Khajooie et al. 2024). The surface area was found to have a stimulating effect on the activity and that a formation-specific effect requires further investigation. So far, it is still unknown what role the surface plays and what mechanism controls the observed effects of rock material on microbial activity, including whether these effects are only temporary and how widespread they are. Therefore, further research on this aspect is needed. Preliminary results with two other hydrogen-consuming microorganisms did not show enhanced hydrogen oxidation in the presence of rock material.
In addition, porous rock formations also provide a habitat in which microorganisms may survive and persist. At the same time, biological processes like biomass accumulation, biofilm formation or microbially induced mineral precipitation might pose further challenges, as such activity might affect porosity and permeability of the porous rock reservoir. However, research on the impact of microorganisms on rock porosity and permeability is limited, mainly due to technical challenges in this research field. To simulate more in situ-like conditions a low-pressure flow-through system was used. M. thermolithotrophicus was successfully introduced into porous rock plugs while the anaerobic microorganisms stayed alive and active. At the same time, the setup was sensitive enough to detect a permeability reduction induced by the introduced microorganisms. This experimental workflow, which is a combination of batch incubations and flow-through experiments, allows us to study microbiology in direct relation to mineralogy. It will be used to gain further insights into the mechanisms that control microbial activity in rocks, as well as how microbial activity could affect the performance of a storage site.

Speaker: Anja Bettina Dohrmann (Federal Institute for Geosciences and Natural Resources (BGR))

20 Evaluating Microfluidic Platforms for Pore-Scale Investigation of Sulfate-Reducing Bacteria under Hydrogen Storage Conditions

Microfluidic chips are increasingly used to study microbial processes at the pore scale due to their optical accessibility, low cost, and experimental controllability. However, the diversity of available microfluidic platforms raises critical questions regarding their suitability for investigating anaerobic microbial reactions relevant to subsurface energy storage. In this study, we systematically evaluate three different microfluidic chip types for microbial experiments, using hydrogen-driven sulfate reduction as a representative case study. The sulfate-reducing bacterium Oleidesulfovibrio alaskensis G20, an anaerobe capable of using hydrogen as an electron donor to produce sulfide, was selected as a model organism relevant to underground hydrogen storage [1]. Experiments were conducted in (i) silicon–glass microfluidic chips, (ii) polymer-based ibidi microchips, and (iii) natural-rock micromodels fabricated from sandstone, each offering distinct advantages and limitations.
Silicon microfluidic chips allow operation under elevated pressures (up to 150 bar) and temperatures representative of reservoir conditions [2]. Their gas-impermeable materials facilitate stable anaerobic environments and enable quantitative studies of hydrogen consumption, biofilm-induced bioclogging, wettability changes , and flow alterations through image analysis [3]. However, their highly idealized pore geometries and surface properties differ significantly from natural rocks, potentially biasing interpretations, and the thick glass cover limits in situ Raman spectroscopic analysis. Ibidi microchips operate at atmospheric pressure but are well suited for coupling with confocal microscopy and Raman spectroscopy. Using a stage-top incubator under continuous nitrogen flushing, microbial activity, biofilm development, and sulfate reduction processes were monitored under controlled anaerobic and thermal conditions [4]. In contrast, natural-rock micromodels incorporate realistic mineralogy, surface roughness, and grain-scale heterogeneity while preserving pore-scale optical access [5]. Their main limitations include hydrogen leakage due to bonding constraints and potential microbial inhibition caused by epoxy-based sealing materials.
By combining these three complementary microfluidic platforms with optical, confocal, and Raman-based techniques, this work provides a methodological framework for selecting and integrating micromodels to investigate bio-geochemical processes relevant to underground hydrogen storage at the pore scale.

Speaker: Na LIU (University of Bergen)

21 Impact of rock-microbe interactions on methanogenic conversion of hydrogen

Hydrogen-consuming microbial metabolisms are gaining increasing attention in the context of underground hydrogen storage (UHS), because hydrogen is a universal electron donor for a wide range of subsurface microorganisms. These processes can cause hydrogen loss and generate unwanted by-products, thereby compromising gas quality and storage integrity. Robust site assessment therefore requires a quantitative understanding of microbial activity and hydrogen consumption kinetics. Batch reactors are commonly used to quantify hydrogen-driven metabolisms using natural formation fluids [Dohrmann and Krüger, 2023], or pure cultures [Strobel et al., 2023] by supplying hydrogen to the headspace. However, recent studies suggest that microbial activity can be markedly enhanced in the presence of particles or rock fragments, which was considered due to the increased accessible surface [Khajooie et al., 2024]. This concept was further experimentally measured in column experiments by measuring hydrogen consumption rates in sand packs with different effective surface areas [Mushabe et al., 2025]. In addition, rock dissolution may supply essential major and trace elements that support enzymatic function and microbial growth [Dong et al., 2022].

Here we present an approach using batch incubations to quantify methanogenic activity with and without Buntsandstein sandstone. Bottles contained 25 mL of a pure culture of Methanothermococcus thermolithotrophicus and were charged with a CO2/H2 gas mixture (20/80 vol%) to an initial pressure of 2.5 bar. Experiments were conducted at 60 °C under three conditions: (i) bulk solution, (ii) solution + crushed sandstone (24.6 g), and (iii) solution + a cylindrical sandstone core (24.6 g, permeability: 70 mD, porosity: 17%). Headspace pressure was monitored continuously and used to calculate hydrogen consumption rates via the ideal gas law. When pressure decline ceased, the headspace was flushed and repressurized to ~2.5 bar, for up to four cycles. Element concentrations in the initial and post-incubation fluids (bulk solution, solution + crushed rock, solution + core) were measured by ICP-OES. The results indicate that adding sandstone did not substantially change the initial hydrogen consumption rate during the first two cycles, consistent with rocks being immersed in the solution and not strongly increasing the effective gas-liquid interfacial area. In contrast, rock-bearing assays sustained methanogenic activity considerably longer than the fluid-only controls, with crushed sandstone supporting the longest activity. Post-incubation fluids containing rock showed elevated concentrations of Mn, Ni, and Ca (and additional trace elements) than the bulk solution, indicating that rock-fluid reactions may replenish nutrients and/or metal cofactors required for methanogenesis. These results demonstrate that rocks influence methanogenic hydrogen conversion not only by providing colonization surfaces and potentially modifying gas-fluid interfaces, but also by supplying geochemically derived nutrients. Rock-microbe interactions and bio-geochemical processes should therefore be explicitly considered in UHS risk assessment and predictive models of hydrogen loss.

Speaker: Dr Chaojie Cheng (Institute of Applied Geosciences, KIT – Karlsruhe Institute of Technology)

22 Biogeochemical reactivity in carbonate reservoirs during underground hydrogen storage

Underground hydrogen storage (UHS) in deep geological reservoirs is a promising technology for large-scale renewable energy storage. Hydrogen injection into the subsurface alters the chemical potential, resulting in a reducing environment that may trigger geochemical and microbial reactivity. This can lead to hydrogen conversion and loss, introduction of impurities, and pore clogging, impacting storage efficiency. Carbonate reservoirs, which make up a quarter of the potential UHS sites in Europe, are theoretically more susceptible to these types of reactivity. This is also true for pyrite-containing reservoirs (1–3), as the latter can react with hydrogen in redox reactions. While several studies have addressed reactivity during UHS, the extent and interactions of these reactions in carbonate aquifers, under reservoir-relevant conditions, remain unclear.
Recently, a pilot hydrogen injection and storage test was conducted in a karstified carbonate aquifer in Loenhout, Belgium, showing a shift in the microbial community and indications of (limited) reactivity upon hydrogen injection. In order to increase our understanding of these observed results and the behavior of such systems, we present here the results of a series of long-term laboratory-scale ambient- and high-pressure (80 bar) batch experiments under reservoir temperatures (65°C) and salinities (120g NaCl/L), with groundwater and crushed rock sampled from the Loenhout reservoir. We tested combinations of growth media with varying nutrient richness, different headspace compositions (hydrogen or nitrogen), and the presence or absence of crushed rock to simulate a range of subsurface conditions, including potential worst-case scenarios.
Preliminary results show low microbial cell counts (~10^3 cells/ml) in the sampled groundwater, with microbial communities initially mainly consisting of previously undiscovered species of sulfate reducing bacteria. Gas-phase analysis also indicates slow microbial reactivity. Moreover, after 19 months of laboratory incubations, cells appear to have been largely adsorbed on the crushed rock phase, without necessarily forming biofilms, suggesting a complex interplay between the solid phase and the microbial community. This may be the result of strong salinity-induced surface charge interactions between micro-organisms and calcite grains. This indicates that mineral surfaces play an important and potentially diverse role in the overall behavior of these systems, impacting availability of reactive minerals dissolved in the groundwater as well as the transport and retention of microbial cells. While further taxonomic analyses are ongoing to gain insight into community composition, our other results suggest slow and thus favorable reaction kinetics during UHS under the tested conditions. This outcome is important to verify the economic viability of hydrogen storage in carbonate reservoirs, which can play a crucial role in enabling the clean energy transition.

Speaker: Soetkin Barbaix (Ghent University)

23 Experimental Micromodel Approaches for Capturing Biogeochemical Interactions in UHS Systems

Underground hydrogen storage (UHS) in porous geological formations represents a promising solution for large-scale energy buffering in renewable-based energy systems. However, interactions between injected hydrogen (H₂) and the subsurface environment can significantly influence storage integrity and efficiency through coupled biogeochemical processes involving native microorganisms.
H₂ acts as a strong electron donor, stimulating microbial activity and modifying redox conditions within the reservoir. These changes can trigger both abiotic (mineral–chemical) and biotic (microbially mediated) reactions in subsurface systems. In particular, sulfate-reducing bacteria (SRB), methanogens, acetogenic bacteria, and iron-reducing bacteria (IRB) play a key role in these processes. Microbial consumption of H₂ may lead to the formation of byproducts such as hydrogen sulfide and methane, posing risks related to corrosion and gas contamination. In parallel, hydrogen-driven reactions can promote cycles of mineral dissolution and precipitation, potentially altering the initial petrophysical properties of the porous medium. The extent of these interactions is strongly controlled by site-specific factors, including mineralogical composition and microbial community structure.
This study investigates how mineralogical composition influences hydrogen-driven microbial processes relevant to UHS using an experimental setup based on a microfluidic system. The micromodel is functionalized with representative mineral phases to isolate the roles of surface reactivity and electron-acceptor availability during H2 and bacterial exposure. Three mineralogical configurations are examined under identical operating conditions (P ≈ 10.2 bar, T ≈ 38 °C): (i) a carbonate-functionalized system, where CaCO₃ is present as the dominant mineral phase; (ii) a sulfate-functionalized system, where CaSO₄ provides sulfate as an electron acceptor for sulfate-reducing bacteria; and (iii) a mixed carbonate–sulfate system combining CaCO₃ and CaSO₄. Following mineral functionalization, the porous micromodel is saturated with H₂ and subsequently inoculated with a strain of SRB (Oleidesulfovibrio alaskensis) as biotic component. System evolution is monitored through timelapse micrograph acquisition over a seven-day period.
The combined presence of an electron donor (H₂) and mineral-based electron acceptors can modify microbial spatial organization and activity within the porous medium, resulting in variable hydrogen consumption and the formation of secondary mineral phases that affect flow behavior. Overall, this work highlights the need for advanced experimental frameworks capable of capturing the complexity of biogeochemical interactions in UHS systems using multimineral micromodel platforms.

Speaker: Frank Viveros Acosta (University of Bergen)

24 Environmental Impacts of Hydrogen Leakage from Deep Underground Storage into Shallow Aquifers: Insights from First Field and Laboratory Investigations

Geological storage of hydrogen (H₂) is now considered a major strategic pillar to support the energy transition. However, several questions remain regarding the risks associated, especially in the event of a slow H₂ leak toward shallow subsurface environments, which constitute the final natural barrier before surface emission. Improving our understanding of H₂ reactivity and its influence on microbial processes in aquifers is therefore essential. Reliable monitoring approaches are required to detect H₂ directly (H₂ concentrations in dissolved and gaseous phases) or indirectly (CO₂, O₂, N₂ in dissolved and gaseous phases, ionic balance, trace elements, redox conditions).

Between 2017 and 2021, Ineris conducted the first in situ experiments at the Catlab experimental site located in the Paris basin (Catenoy, France). A simulated H₂ leakage was created by injecting groundwater saturated with dissolved H₂ into the shallow chalky aquifer (~20 m deep). This unconfined aquifer contains groundwater of calcium–bicarbonate facies with a near-neutral pH. A network of eight piezometers and four dry boreholes enabled monitoring in both saturated and unsaturated zones. The site was equipped with advanced geochemical instrumentation, including a gas-completion well coupled to Raman and mid-IR probes. A total of 5 m³ of H₂-saturated groundwater was injected into the aquifer, following a tracer injection to track plume migration. Complementary sampling enabled characterization of ionic and trace-element responses associated with the simulated leakage. These initial tests revealed short-term physicochemical perturbations following H₂ injection (decrease of redox, O2, CO2, electrical conductivity, bicarbonate ions…) but too brief to allow a reliable evaluation of H₂ biodegradation or microbial community dynamics.

To overcome this limitation, a laboratory column experiment was designed to reproduce, over several weeks, the geochemical and microbial evolution of an aquifer exposed to H₂. Groundwater from the Catlab site was saturated with H₂ and circulated through a sediment-filled column. This aimed to evaluate the potential stimulation of hydrogenotrophic, denitrifying, or sulfate-reducing communities through a multi-scale monitoring strategy, combining measurements and sampling.

Daily measurements included continuous monitoring of outlet flow rate, dissolved H₂ concentration in the column outflow, and key physicochemical parameters. The column itself was weighed daily to detect possible variations in water content, microbial development or gas retention within the porous medium. Weekly monitoring focused on the hydrochemical and microbiological evolution of water samples and porous material, enabling quantification of major and minor ions, trace elements, as well as assessment of microbial abundance, viability, and shifts in community structure: we noted variations in nitrates, nitrites and bicarbonates. In addition, targeted analyses using quantitative PCR and high-throughput sequencing were performed. Two dedicated samples, taken at the beginning and end of the experiment, allowed the identification and quantification of functional microbial groups potentially involved in H₂ consumption.

Overall, this work provides new insights into hydrogen reactivity, associated geochemical perturbations, and microbial responses in shallow aquifer systems, combining results from in situ experiments and controlled laboratory column studies.

Speaker: Imen ZAIER (Institut national de l’environnement Industriel et des risques)

MS06: 2.2

MS09: 2.2

MS13: 2.2

25 Flow and Electrokinetic Transport in Nanoporous Media

Ion transport is ubiquitous in aqueous environments in biological, geological, chemical and environmental systems. Electrokinetics plays a very important and key role in some special cases where pore size is comparable to the screening length of electrical double layer. The applications include fuel cells and batteries, radiative waste disposal, high-quality water purification, and even ion channels in cells. This talk will present (1) electrokinetic and interface theories for ion transport in micro/nanoporous media; (2) a mesoscopic numerical framework for predictions and the validations by comparisons with theories and experimental data; (3) multiscale analysis in both spacial and temporal scales for special applications.

Speaker: Prof. Moran Wang (Tsinghua University)

26 Electrochemical Responses of Mesoporous Carbons in Aqueous Electrolytes

When mesoporous carbon materials come into contact with electrolyte solutions, interactions at their surfaces can lead to the spontaneous formation of electrical potentials. Even without applying an external voltage, differences in surface properties can drive charge separation and ion rearrangement at the solid–liquid interface. When two materials with distinct surface characteristics are combined, these effects can generate a measurable electrical response, offering potential for energy harvesting applications.
​This work presents a theoretical and experimental investigation of the factors influencing such spontaneous potential differences. A modeling approach is introduced and supported by experimental observations across different material treatments and electrolyte conditions. Synchrotron-based techniques are used to gain qualitative insight into ion distribution profiles within the porous structures during filling, and how this behavior relates to the observed electrical signals. The study is aimed at providing a broader understanding of ion–surface interactions in porous materials and exploring their relevance for emerging electrochemical energy concepts.

Speaker: Mariia Liseanskaia

27 Statistical Mechanics of Fluid Adsorption on Flexible Metal-Organic Frameworks

Recent Nobel prize recognizing the inventors of Metal-Organic Frameworks (MOF) has sparked renewed overarching interest in these fascinating materials. Built of metal complexes connected by organic linkers into three-dimensional porous crystals, MOFs can selectively adsorb and separate multicomponent fluid mixtures. Their intrinsic framework flexibility enables the design of compliant molecular sieves, membranes, sensors, and actuators. Gaining a fundamental understanding of the physical mechanisms that couple fluid adsorption and mechanical response is essential for advancing the design, fabrication, quality control, and practical deployment of MOF-based materials. In this talk, I will present a rigorous statistical mechanical approach, which links the framework poromechanics and adsorption thermodynamics. This approach constitutes the basis of molecular simulation modeling of fluid adsorption on compliant nanoporous materials. The proposed approach is illustrated by several examples from our recent research, including studies of the effects of framework flexibility on adsorption isotherms, non-monotonic adsorption-induced deformation and framework compressibility, as well as gate opening and breathing transformations.

• Parashar and Neimark, Understanding the Origins of Reversible and Hysteretic Pathways of Adsorption Phase Transitions in Metal-Organic Frameworks, JCIS, 2024, 673, 700.
• Parashar, Corrente, and Neimark, Unveiling Non-Monotonic Deformation of Flexible MOFs during Gas Adsorption: From Contraction and Softening to Expansion and Hardening, JCIS, 2025, 606, 88.
• Neimark, Corrente, and Coudert, Phase Transformations in MOFs Induced by Adsorbate Exchange. Langmuir, 2025, 41, 4720.

Speaker: Prof. Alexander Neimark (Rutgers University)

28 Periodic Mesoporous Organosilicas as Host Materials for Studying Surface Chemistry and Pore Size Effects on the Properties of Nanoconfined Water

Water is undoubtedly the most important substance on earth. It is ubiquitous in nature and a necessary liquid for the emergence of life. Although by far the most classic liquid encountered in everyday life, water presents many unusual physical properties, which are not yet fully understood. A large number of studies have highlighted the crucial role of hydrogen-bonding interactions between water molecules in determining the peculiar liquid structure and physicochemical properties of water. In most frequent situations, water is found as spatially confined or in an interfacial state rather than forming a bulk phase. From a fundamental point of view, confining water at the nanoscale in prototypical porous solids has turned out to be particularly adequate in order to better understand the unusual behavior of interfacial water.
Among several types of confinement, including clays or zeolites, the mesoporous SBA-15 and MCM-41 silicas are particularly suited hosts due to their well-defined porous geometry formed by ordered cylindrical channels. While MCM-41 and SBA-15 silica provided an adjustable pore size and can address the geometrical aspect of the nanoconfinement, the evaluation of the effect of surface interaction on the water properties is limited due to the unchanged chemical composition. In order to extend current knowledge, which has so far been based on a few studies on grafted silicas, we are contemplating new opportunities offered by the molecular scale imprint of the water−surface interaction. Periodic mesoporous organosilicas (PMOs) are particularly well-suited, though barely used in water studies so far. In contrast to the MCM-41 silica the PMOs can contain organic bridging units within the quasi-crystalline pore walls and therefore a periodically modulated surface polarity [1]. The chemistry of these bridging units can vary from hydrophilic to hydrophobic and can also contain surface ionic charge with localized cations and exchangeable anions. Unlike post-synthetically surface-grafted nanoporous silicas, PMOs allow a stoichiometric control of a periodically alternating surface chemistry along the pore channel.
Here, we present new insights into how surface chemistry and pore size influence the properties of nanoconfined water. We studied water in PMOs with pore diameters in the range of 2-5 nm. In these materials, the molecular mobility of water as well as its melting point and the properties of the non-freezable water layer (so-called t-layer) are influenced by the polarity of the organic moiety [2-6].
Surface hydrophilicity has little effect on melting point depression in larger pores but becomes increasingly influential as pore size decreases. In hydrophobic PMOs, water exhibited larger melting point depression, lower specific enthalpies, and thicker t-layers with lower average density than in hydrophilic ones. In contrast, charged PMOs behaved differently: despite higher hydrophilicity, confined water exhibited a larger melting temperature depression, lower specific enthalpy, larger critical pore radius, and comparatively thicker t-layers, likely due to higher disorder of the hydrogen-bonding network close to the surface [4,6]. Moreover, the t-layer density did not follow a simple trend based solely on hydrophilicity. These results highlight the complex interplay between pore size, surface chemistry, and interfacial water behavior.

Speaker: Prof. Michael Fröba (University of Hamburg)

29 Capillary Flow of Aqueous Solutions in Nanopores

Aqueous solutions confined within nanopores play a fundamental role in both natural and technological systems, governing processes such as ion regulation in cells, desalination, blue energy generation and the durability of construction materials. In this project, we aim to investigate the flow and phase behavior of aqueous solutions and the possible deviations from bulk behavior caused by nanoconfinement. Particular attention is devoted to hydrotropic compounds, which, beyond their role as green solvents, enable the modulation of the interactions among water, solutes, and pore surfaces.

Experimentally, we investigate the imbibition of water into nanoporous silica at different solute concentrations and relative humidities. In parallel, we employ molecular dynamics (MD) simulations to investigate the capillary flow of aqueous glycerol and ethylene glycol (EG) through a single nanopore at varying concentrations, and generalize the framework to describe any aqueous mixture flowing within a nanopore of given wettability properties, by tuning the mutual interactions among the solvent, solute and pore.

Experimental results show qualitative deviations from Lucas-Washburn behavior, with the square of the filling length exhibiting a non-linear trend except for water, highlighting the influence of the solute. A two-regime flow was observed in glycerol solutions which can be explained by a possible solute-solvent demixing. This hypothesis is supported by MD simulations, which show that glycerol and EG exhibit a slower filling rate and preferential adsorption onto the pore walls compared to water. These findings provide new insights into the role of solute-solvent-pore interactions in nanoconfined flows and provide a basis for predicting and controlling transport in nanoporous systems.

Speaker: Abir-Wissam Boudaoud (Institut Lumière Matière, UMR 5306 , CNRS et Université Claude Bernard Lyon 1)

30 Cavitation in Confined Fluid

Liquid under tension “breaks” by cavitation, forming a vapor bubble. It occurs in engineering (ultrasonic cleaning, erosion of ship propellers...) as well as in the natural sciences (gas embolism in trees, pistol shrimp...). In the cavities of a saturated porous material, the liquid is also under tension when it evaporates. In this case, it was long considered that evaporation occurs by recession of the menisci delimiting the saturated region but it is now recognized that evaporation can also be due to cavitation [Thommes 2006, Rasmussen 2010, Doebele 2020]. However, there are only few experimental data on the impact of the confinement on the cavitation threshold so that theoretical approaches [Rasmussen 2020, Morishige 2021] cannot be accurately tested.

In this work, we focus on evaporation of nitrogen in ordered mesoporous silica (SBA-16). We have designed a capacitive setup in order to perfom continuous measurement of the fraction $f$ of the pores filled with liquid, while decreasing the vapor pressure $P _V$ outside the pores at controlled rate $A$. This technique has two major advantages compared to usual volumetry. First, comparing the dependence of $f$ with $P_V$ at different rate $A$ provides a direct signature that cavitation is sensitive or not to the fluid confinement. Second, the pressure cavitation threshold $P^*$ can be unambiguously defined and precisely measured as a function of the rate $A$. This allows to determine the dependence $\alpha=dE_B/dP_V$ of the energy barrier $E_B$ with the pressure.

We have performed systematic measurements of $\alpha$ for temperatures ranging from 70 K up to $T_h$ at which adsorption hysteresis disappears, for a serie of SBA-16 with cage diameter in the range 5 – 9 nm. For the largest pores and the lower temperatures, that is when the critical nucleation radius is the smallest, we recover $\alpha$ values which are close to those obtained for bulk cavitation [Bossert 2023]. The departure from the bulk case increases when the critical bubble radius becomes closer to the cage radius.

In contrast with $P^*$ measurements, the determination of $\alpha$ can be easily compared with theoretical predictions, since neither the knowledge of the attempt frequency nor the number of nucleation sites is required. Following the semi-macrosopic approach of Bonnet and Wolf [B&W 2018] and Morishige [Morishige 2021], we have calculated the energy barrier for bubble nucleation in the sharp interface limit, taking into account the curvature dependence of the surface tension [Bossert 2023]. Whatever the type of the wall-fluid interaction potential (whose amplitude is fixed by the measured value of $T_h$), we find this simple model underestimates the observed effect of confinement. More sophisticated approaches such as Density Functional Theory could possibly yield a better agreement with measurements. However, the semi-macrosopic model can still be improved by breaking the spherical symmetry, that is taking into account the probability that nucleation does not occur at the center of the cage, as observed in Molecular Dynamics simulations.

Speaker: Etienne Rolley (LPENS)

MS15: 2.2

31 Surrogate Modeling of Particle Retention in Porous Media Enabled by a Massive Pore-Scale Simulation Dataset

The retention of suspended particles in porous media plays a critical role in a wide range of subsurface processes, including filtration, contaminant transport in environmental applications, and formation damage in subsurface energy applications. As flow with suspended particles flow through porous media, they may deposit or clog flow pathways, changing local porosity, and ultimately impacting large-scale hydraulic behavior (permeability). Although pore-scale computational fluid dynamics (CFD) coupled with discrete element models (DEMs) can resolve these mechanisms, their high computational cost prevents extensive sensitivity analyses. Moreover, the absence of large pore-scale datasets suitable for surrogate modeling represents a major research gap.
To address this, we systematically extended the pore-scale model of Sadeghnejad et al. (2022) to generate a large-scale dataset for machine-learning surrogate development. Key physical and geometric parameters, including particle size, concentration, injection velocity, and pore-space morphology, were varied across wide ranges. For each realization, the Eulerian-Lagrangian workflow (including Navier-Stokes flow simulation, individual particle tracking modeling, dynamic voxel-based deposition, and porosity/permeability updating) was executed until steady post-retention conditions were achieved. Approximately 130,000 simulation points were run, consuming ~49,000 CPU-hour, which is one of the largest particle-retention datasets reported to date. Moreover, outliers of the dataset were removed by the Isolation Forest algorithm. Seven machine learning models (i.e., Adaptive Gradient Boost (AGB), Decision Tree (DT), Extremely Randomized Trees (XRT), Extreme Gradient Boost (XGB), Gradient Boost Machine (GBM), Multi-layer Perceptron (MLP), and Random Forest (RF)) were trained on 80% of the dataset with standard hyperparameter values to predict the final porosity and permeability of the domain after particle deposition.
Initial evaluations identified XGB and XRT as the most promising surrogate candidates. Both models were subsequently refined through Bayesian hyperparameter optimization to enhance predictive robustness and generalization. Model performance was assessed using five-fold cross-validation and the metrics Mean Squared Error (MSE), Mean Absolute Error (MAE), and the coefficient of determination (R²). The optimized models achieved excellent predictive accuracy, with R² values exceeding 0.98 for porosity and 0.90 for permeability, respectively. In addition to their accuracy, these surrogates provide orders-of-magnitude faster inference than pore-scale simulations, underscoring their suitability for rapid assessment of particle-retention behavior. Comparative performance metrics and predictive outcomes are illustrated in the following figure.

Speaker: Dr Saeid Sadeghnejad (Institute for Geosciences, Applied Geology, Friedrich-Schiller-University Jena, 07749 Jena, Germany)

32 Residual-based PINN Modeling for Coupled Transport Phenomena in Porous Gas Diffusion Layers

Abstact:
The gas diffusion layer (GDL) of high-temperature proton exchange membrane fuel cells plays a critical role in regulating the coupled transport of species, heat, and charge. The simulation accuracy of these transport phenomena directly affects the predictive reliability of fuel cell performance. However, conventional computational fluid dynamics (CFD) simulations suffer from prohibitively high computational costs, while standard physics-informed neural networks (Pure PINNs) struggle to capture the complex field gradients within the GDL due to gradient vanishing in deep architectures. To address these challenges, this study proposes a residual physics-informed neural network (Res-PINN) framework for accurately modelling the multiphysics coupling processes within the hydrogen-side GDL. The proposed model embeds the governing equations of momentum, mass, and charge conservation directly into the loss function, thereby ensuring strict adherence to physical laws. To overcome the training limitations of deep Pure PINNs, a residual architecture with skip connections is introduced. By constructing identity-mapping pathways, this design effectively mitigates gradient vanishing during backpropagation and significantly enhances the network's ability to capture strong nonlinear gradient variations in porous media. The results indicate that the Res-PINN consistently outperforms the Pure PINN, achieving substantial improvements in predictive accuracy, with overall error levels reduced by 95% to 212% across different physical fields. In particular, the pressure field predictions exhibit near-perfect agreement with the reference solutions. In terms of computational efficiency, the proposed model achieves a 374-fold speedup compared with conventional CFD methods, reducing the inference time per evaluation from 1.0 s to 2.7 ms, whilst maintaining excellent generalization performance under previously unseen operating conditions. Overall, these findings demonstrate the superior capability of the Res-PINN architecture in handling complex multiphysics coupling problems. By preserving strong physical consistency while alleviating the training bottlenecks of deep PINNs, the proposed approach provides an efficient and reliable digital modeling tool for real-time simulation and engineering optimization of next-generation hydrogen-powered aircraft propulsion systems.

Keyword: Residual Physics-Informed Neural Network (Res-PINN), Gas Diffusion Layer (GDL), Multiphysics Coupling, Porous Media Simulation, Gradient Vanishing Mitigation

Speaker: Ms Hui Zhang (University of Bristol)

33 Glassy dynamics in steady-state two-phase flow in porous media

Immiscible two-phase flow in porous media exhibits different flow regimes depending on driving parameters like the capillary number and viscosity ratio. In the steady state, these regimes correspond to characteristic pore-scale flow patterns, such as ganglion flow and drop-traffic flow. By considering pairwise fluid-fluid correlations in the pores and maximizing the entropy, we derive a configurational probability distribution for steady-state two-phase flow that characterizes these pore-scale patterns. The energy function in the probability distribution resembles that of an Ising spin system. Using Boltzmann machine learning applied to configurational data from dynamic pore-network simulations, we estimate the coupling constants in the energy function. We find the couplings are disordered with both positive and negative values similar to those in a spin-glass system, and their distribution depends on the applied pressure drop. Such distributions introduce frustration in a spin-glass system. We investigate the implications of this frustration in the two-phase flow system by measuring magnetization, spin-glass order parameter and susceptibilities from pore-scale configurations. These quantities allow us to characterize the flow regimes and reveal a spin-glass like transition. While our analysis uses steady-state configurations from a dynamic pore-network model, the method is equally applicable to data from other computational approaches or experiments.

Speaker: Dr Santanu Sinha (PoreLab, Department of Physics, Norwegian University of Science and University, N-7491 Trondheim, Norway)

34 Transparent on-demand neural approximation of EOS-based thermodynamics for pore-scale gas-condensate flow

Accurate evaluations of thermodynamic equilibria are essential for pore-network modeling (PNM) of gas-condensate flow in porous media. However, repeated equation-of-state (EOS) calculations impose a significant computational burden, limiting the feasibility of large-scale, dynamic simulations. This work presents a neural network–based data-driven proxy framework, implemented using JAX, for efficiently approximating thermodynamic phase behavior required in PNM simulations of gas-condensate systems. A custom implementation of the full Peng–Robinson EOS was developed as part of the same framework, serving both as a high-fidelity alternative to the proxy and as the reference model for network training. The proxy network is trained on EOS-based thermodynamic data spanning a representative range of pressures, temperatures, and fluid compositions. To further improve ease of use and efficiency, training is performed on demand and the resulting network parameters are automatically cached for reuse across simulations with compatible thermodynamic conditions. The trained model predicts phase equilibrium with good accuracy while achieving a substantial reduction in computational cost compared to conventional EOS solving. Integration of the proxy network into a dynamic PNM framework enables efficient simulation of multiphase gas-condensate transport, including phase appearance and disappearance at the pore scale. Results demonstrate that the proposed approach preserves the fidelity of predictions while significantly accelerating simulations. The framework provides a scalable and flexible pathway for incorporating complex thermodynamics into pore-scale models, facilitating improved understanding and upscaling of gas-condensate flow in porous media.

The authors thank the technical and financial support of Petrogal Brasil S.A. (Joint Venture Galp | Sinopec) and the promotion of Research, Development and Innovation (R,D&I) in Brazil by the National Agency of Petroleum, Natural Gas and Biofuels (ANP) for the execution of this project.

Speaker: Dr Gabriel Gerlero (LMMP/PUC-Rio)

35 Shearlet, a Novel Operator Learning Model

High-fidelity pore-scale flow simulations are indispensable for characterizing transport phenomena in complex porous media. Techniques like the Lattice Boltzmann Method (LBM) and direct Stokes solvers explicitly resolve three-dimensional pore-space flow fields, capturing essential effects of pore connectivity, multiscale heterogeneity, and intricate boundary conditions. However, their prohibitive computational cost restricts application to large domains, high resolutions, or multiple flow scenarios. This limitation has spurred interest in surrogate models that can replicate pore-scale solutions at a dramatically reduced computational cost while preserving physical accuracy.
This work introduces the Shearlet Neural Operator (SNO), a novel neural operator based on shearlet representations, as an efficient surrogate for pore-scale flow solvers. In contrast to conventional Fourier-based neural operators—which rely on global sinusoidal bases and struggle with localized, anisotropic, or non-smooth features—the SNO harnesses the multiscale and directional properties of shearlets. This allows it to efficiently represent localized flow structures, sharp gradients, and anisotropic patterns, making it particularly well-suited for problems involving multiscale geometries and non-smooth solutions, including regimes with shocks or sharp transitions.
Formulated to learn mappings between function spaces, the SNO directly approximates the solution operator that maps pore geometry and boundary conditions to velocity fields. By operating in a multiscale shearlet domain, it naturally accommodates varying resolutions and captures both global flow behavior and fine-scale local features. This design overcomes key limitations of Fourier-based neural operators, whose globally supported basis functions hinder their ability to represent localized phenomena and scale-dependent structures effectively.
The methodology is first validated on a series of controlled benchmark problems designed to test its capability in representing multiscale features, anisotropy, and sharp spatial variations. These benchmarks highlight the SNO's robustness and expressiveness in regimes where Fourier-based operators exhibit degraded accuracy. Subsequently, the approach is applied to a physically relevant pore-scale flow problem: predicting three-dimensional velocity fields. Trained on simulation data, the resulting surrogate mode estimates the velocity fields while achieving orders-of-magnitude acceleration in computational speed.
The Shearlet Neural Operator offers a scalable, resolution-aware, and physically expressive surrogate. By integrating multiscale directional representations with operator learning, this framework provides a promising pathway toward fast, accurate simulation of flow in complex porous media, with potential extensions to broader classes of multiscale and non-smooth physical systems.

Speakers: Júlio de Castro Vargas Fernandes (LNCC), Dr Fabio Pereira dos Santos (LNCC)

36 Assessing the potential of physics informed neural networks for modeling groundwater flow in unconfined aquifers

Groundwater flow modeling in aquifers is a fundamental problem in hydrogeology, traditionally addressed using numerical or data driven models that require sufficient observational data and well-defined boundary conditions and high computational demands. However, in many real-world groundwater systems, available observation data are sparse, and boundary conditions are often poorly known or highly uncertain. These limitations motivate the exploration of alternative modeling approaches that can remain reliable under data scarcity and incomplete physical information. In this context, neural network (NN) models are receiving significant attention due to their reliability and high computational performance when trained on GPU cards. potential of physics-informed neural networks (PINNs) a recent approach that reduces the dependence of neural network (NN) models on data by explicitly incorporating physical processes into the training procedure. This study aims to assess the performance of PINNs for modeling groundwater flow in heterogeneous unconfined aquifers, and to compare it against conventional data-driven NN models.

In this work, PINN is implemented using a mixed formulation of the governing equations to improve training in highly heterogeneous domains. The results of PINN are compared to a purely data-driven NN model. Finite element solutions are used as reference for error assessment of PINN and data driven NN models. The comparison is carried out by decreasing the amount of observational data. When trained using a relatively dense set of observation data, the pure NN demonstrates excellent predictive performance and accurately reproduces the reference hydraulic head field. Where field observations are typically limited, the predictive accuracy of the NN model deteriorates significantly, highlighting the inherent limitations of purely data-driven models when observational data is insufficient. The results demonstrate that the inclusion of physical constraints, through PINNs, substantially improves model performance under limited data availability, leading to more accurate and stable hydraulic head predictions compared to the conventional NN.

In a more challenging scenario, all boundary condition information is removed from the model to simulate situations in which aquifer boundary conditions are unknown or highly uncertain. In this case, data-driven methods exhibit poor performance. In contrast, the PINN approach remains capable of producing physically reliable results, even in the absence of explicit boundary condition information. Overall, the findings of this study indicate that PINNs offer a robust and powerful alternative for groundwater flow modeling, particularly in applications characterized by sparse data and uncertain boundary conditions.

Speaker: Marwan Fahs (ENGEES-LHYGES)

MS20: 2.2

MS03: 2.3

MS04: 2.3

37 Model calibration and prediction of biogeochemical processes in porous hydrogen storage

Underground hydrogen storage (UHS) represents a promising solution for the temporal balancing of energy supply and demand in energy systems increasingly based on renewable sources. Suitable geological storage formations include both water-saturated porous media (aquifers) as well as former hydrocarbon reservoirs such as depleted gas or oil fields. For the planning, development, and operation of such storage systems, a detailed understanding of the coupled flow and reactive transport processes in porous media is essential.
In this work, a numerical simulation model is presented that consistently couples two-phase flow processes with biogeochemical reactive transport. Particular emphasis is placed on the representation of microbial growth and reaction kinetics, allowing for the description of both substrate-rich and substrate-limited conditions. The model captures the interactions between gas and liquid phases, diffusive and advective transport mechanisms, and microbially induced reactions.
Key model parameters were calibrated using laboratory-scale porous media experiments, including diffusion experiments on core samples and microfluidic studies. In addition, the model has been preliminarily calibrated and validated using field data. The results indicate that biogeochemical processes can measurably influence hydrogen transport, gas composition, and overall storage performance. The proposed modeling approach provides a practical framework for evaluating coupled physical and biogeochemical processes in underground hydrogen storage systems.

Speaker: Birger Hagemann (Clausthal University of Technology)

38 Pore Scale Mechanistic Transitions in Geo-Methanation

The European pursuit of a net-zero economy is increasingly defined by two parallel challenges: (1) the urgent mandate to mitigate energy-related greenhouse gas emissions and (2) the necessity of managing the inherent volatility of renewable energy sources. As weather-dependent power production expands, the resulting temporal mismatches between energy supply and consumer demand require the integration of flexible, large-scale seasonal storage solutions. Storing energy in the form of gaseous molecules within subsurface geological formations provides the systemic flexibility required to stabilize the power grid, while offering a transformative pathway to reduce fossil fuel dependence over time.

Geo-methanation represents a transformative technology for circular carbon utilization, enabling the in-situ conversion of hydrogen and carbon dioxide into methane within geological formations. Despite its strategic potential for hydrogen storage and carbon sequestration, the large-scale implementation of subsurface methanation is hindered by fundamental uncertainties regarding conversion efficiency and pore-scale transport dynamics. This research addresses these gaps by establishing a novel, high-resolution experimental-numerical framework designed to resolve the complex interplay between microbial kinetics and multiphase flow.

The originality of this work lies in the development of a microfluidic platform capable of emulating relevant subsurface conditions, integrated with direct numerical simulations (DNS) to bridge the gap between visual observation and mechanistic theory. Through a workflow encompassing micromodel colonization, anaerobic substrate introduction, and gas chromatography, we characterized biomass distribution and methane production kinetics under controlled anaerobic flow regimes.

Our findings reveal three critical insights that redefine the current understanding of subsurface bio-conversion. First, during substrate gas injection, we observed a significant behavioral shift in microbial aggregation, transitioning from a colony-dominated to a planktonic lifestyle. Second, the spatial analysis demonstrated that colony disintegration and subsequent cell migration toward gas–liquid interfaces are primary drivers for enhanced substrate uptake. This phenomenon was quantified by a measured methane evolution rate peaking at approximately 0.35 mmol/L·h, indicating that biomass mobility is essential for maintaining conversion efficiency. Third, through dimensionless analysis, we identified distinct transport regimes within the pore network, ranging from molecular diffusion-limited zones to advection-enhanced mixing areas.

This research demonstrates that the efficacy of geo-methanation in unsaturated environments is governed by a delicate balance of microbial activity, interfacial mass transfer, and advective nutrient supply. By reconciling experimental pore-scale data with calibrated numerical results, this work provides predictive insights necessary to optimize the competitiveness of subsurface environments for renewable energy storage and greenhouse gas mitigation. These results have significant implications for the design of future pilot-scale operations, ensuring that the evolution of hydraulic rock properties and microbial dynamics are accounted for in long-term storage strategies.

Speaker: Patrick Jasek

39 Detailed characterization of pore structure and transport properties of biomass particles during pyrolysis

Biomass pyrolysis involves strongly coupled structural evolution and transport processes that govern heat and mass transfer, yet these processes remain insufficiently understood at the pore scale. In particular, the roles of pore-scale anisotropy and heterogeneity in controlling gas transport and reaction progression are often neglected in continuum-scale models. In this study, we present an image-based pore-scale framework to quantify the evolution of pore structure and transport properties in wood particles during staged pyrolysis, and to bridge these effects toward representative elementary volume (REV)–scale descriptions.
High-resolution X-ray computed tomography images acquired at multiple pyrolysis temperatures were used to reconstruct three-dimensional pore structures. Image-based pore network models (PNMs) were extracted that explicitly preserve the inherent anisotropy and heterogeneity of the biomass pore space. Structural descriptors, including pore size, coordination number, and orientation statistics, were quantified to characterize the temperature-dependent evolution of pore morphology and connectivity. The results reveal a contraction–enlargement duality: while the total number of micrometer-scale pores decreases due to solid-phase decomposition and pore collapse, the remaining pores enlarge and become increasingly aligned, leading to pronounced anisotropy in the pore network.
Pore-scale transport simulations were conducted on the extracted PNMs and subsequently upscaled to REV-scale transport properties. Although porosity remains an important control, permeability is shown to be strongly governed by coordination number and directional alignment, resulting in preferential transport along specific orientations. REV-scale conductance maps further demonstrate that anisotropy persists across scales: radial conductances migrate inward with increasing temperature, whereas azimuthal and elevation conductances remain spatially heterogeneous due to local structural variations.
By coupling REV-resolved transport properties with layer-resolved carbon loss, we show that pyrolysis progresses radially from the particle exterior toward the interior, while maintaining significant within-layer anisotropy in both reaction intensity and gas flux. The extracted REV-scale source terms and directional conductances provide physically grounded inputs for continuum-scale reactive transport models. Overall, this work highlights the critical role of pore-scale anisotropy in biomass pyrolysis and provides a multiscale pathway for predictive upscaling of thermochemical conversion processes.

Speaker: Ninghua Zhan

40 Pore-scale modeling of coupled processes in biofilm-colonized porous media

Biofilm formation in porous media plays a central role in controlling flow, transport, and biogeochemical processes in natural and engineered systems, including groundwater environments, wastewater treatment, water quality management, and geological gas storage. In this contribution, we present recent advances in pore-scale modeling that elucidate how biofilm dynamics and structure jointly shape the transport properties of porous media.

We employ a micro-continuum framework in which biofilms are represented as lower-scale fluid-filled porous media, enabling the simulation of biofilm processes without explicitly tracking the biofilm-fluid interface. Pore-scale simulations reveal distinct biofilm growth regimes controlled by hydrodynamic conditions. Increasing flow rates enhance biofilm accumulation up to a critical threshold, beyond which hydrodynamic stresses induce biomass detachment. These regimes are interpreted using a dimensionless number quantifying the balance between drag forces and biomass cohesion. We further show that permeability reduction is not solely determined by total biomass but strongly depends on the spatial organization of biofilm within the pore space.

Beyond bioclogging, we investigate the impact of biofilms on solute transport by coupling the micro-continuum approach with Random Walk Particle Tracking. Our results demonstrate that biofilm heterogeneity, internal convective pathways, and reduced effective diffusivity lead to anomalous transport behaviors, including enhanced dispersion and pronounced tailing. Together, these findings highlight how biofilm structure and dynamics fundamentally alter porous media properties and provide mechanistic insights relevant for predicting and managing biofilm-driven processes in environmental and engineering applications.

Speaker: michele starnoni (Universitat Politecnica de Catalunya)

41 PINNs enhanced multi-resolution modeling of laminar vortex dynamic process in pore-scale MICP

Microbial mineralization is a novel bioremediation and consolidation technology. However, its mineralization process is influenced by a variety of complex factors (such as microbial species, urea concentration, and the evolution and distribution of pore vortex structures) at the pore scale, presenting highly nonlinear characteristics and a certain degree of uncertainty in distribution and evolution. Due to the difficulty in real-time observation of the reaction-flow coupling process and the dynamic changes of pore structure in the pore space, experimental studies are hard to deeply explore the micro-mineralization mechanism at the pore scale. Based on the lattice Boltzmann method (LBM), Eulerian finite element method (FEM), and cellular automata (CA), this study constructed a multi-physics coupling numerical model for pore-scale microbial mineralization. High-resolution numerical simulation in space was achieved by using the Physical Informed Neural Network (PINNs) method. Combining GPU parallel acceleration technology, a three-dimensional complex pore flow-reaction coupled universal multi-physics field solver was independently developed. The full-scale mineralization process simulation of three-dimensional microfluidic chips was successfully realized, and the experimental phenomena of the microfluidic chips were quantitatively reproduced. Based on the verified model, the distribution law of calcium carbonate precipitation and the influence of the initial pore structure on its evolution process were quantitatively analyzed, providing quantitative suggestions and prediction tools for optimizing the biological grouting strategy. The mechanism of the vortex phenomenon caused by the dynamic evolution of pore structure and its influence on the uniformity of mineralization were further explored. Through quantitative analysis of the vortex evolution distribution based on the Liutex vortex identification method, the correlation between vortices and the generation amount of calcium carbonate as well as the degree of solute mixing was studied. The dynamic coupling influence mechanism of vortices and reaction processes in the microbial mineralization evolution system of porous media was preliminarily and quantitatively revealed. This research provides predictive analysis methods and models for the refined application design and process control of microbial mineralization technology.

Speaker: Dianlei Feng (Tongji University)

42 Porous media modeling of macromolecule diffusivity in living cells

Transport phenomena in biological systems are essential for maintaining life sustaining functions. Notably, biological materials, including tissues and cells can be viewed as porous media. Here we will focus on the passive transport of macromolecules in the intracellular space, involved in many cellular functions such as cell migration, blebbing and apoptosis. While it is well established that intracellular crowding significantly impacts macromolecule mobility, the physical mechanisms by which cytoplasmic structures influence diffusion within the cell remain unclear.

We propose a multiscale diffusion model of the intracellular space based on the volume averaging method. The cytoplasm is treated as a hierarchical porous medium with nanometric and micrometric obstacles. Numerical solution of the model allows us to predict the effective cytoplasmic diffusion coefficient for various obstacle volume fraction. Model predictions are confronted to experimental measurements of the effective diffusion coefficient in live cells and highlight the importance of two key diffusion reduction mechanisms: tortuosity and hydrodynamic drag. Importantly, we find that the effective cytosolic diffusivity is not dependent on specific cellular region but rather on intracellular obstacle volume fraction. Additional model predictions of intracellular diffusivity as a function of the macromolecule size give excellent agreement with literature data.

Altogether, this work demonstrates the potential of porous media modeling approaches to better understand transport phenomena in heterogeneous biological systems all the way to the intracellular scale.

Speaker: Morgan Chabanon (CentraleSupelec, Paris-Saclay University)

MS05: 2.3

MS07: 2.3

MS09: 2.3

MS10: 2.3

MS18: 2.3

MS20: 2.3

Poster: Poster IV

Plenary Lecture: Plenary 2

43 Additive Manufacturing of Porous Ceramics from Precursors

Additive manufacturing of porous ceramics is somewhat limited by their high melting temperatures and the processing issues related to handling of feedstocks containing a large volume of particles. Processing slurry-based feedstocks, in fact, poses several challenges: a high amount of powder is required to promote densification and results in high viscosity, scattering and sedimentation phenomena in vat photopolymerization processes, as well as clogging problems at the nozzle for extrusion-based processes. Some of these issues can be solved or mitigated when using precursor-based feedstocks, when they are all liquid.
Our research activities have focused on the use of preceramic polymers solutions as feedstock for the production of porous ceramic components by additive manufacturing.
We also investigated the additive manufacturing of both geopolymer solutions and geopolymer powders, as precursors for different components of interest for absorption, catalysis or high temperature applications.
In this talk, our strategies for producing high quality ceramic components using a variety of precursor feedstocks will be presented. Different additive manufacturing techniques were used to fabricate components ranging in size from the sub-micron to the tens of centimeters, including direct ink writing, binder jetting, digital light processing, two photon polymerization, robotic arm manufacturing and volumetric additive manufacturing.

Speaker: Paolo Colombo

Thursday 21 May

Invited Lecture: Invited V

44 From Understanding to Practice: Confined Thermodynamics and Diffusion in Tight Hydrocarbon Reservoirs

Recovery from shale and tight oil reservoirs remains limited, with recovery factors often below 10% of the original oil in place despite horizontal drilling and hydraulic fracturing. Traditional reservoir models fail to capture the physics of nanometer-scale pores, where confinement, adsorption, and molecular diffusion dominate. This presentation examines how confined thermodynamics reshapes phase behavior and miscibility, supported by experimental core-flooding and CO₂ Huff-n-Puff studies with CT scanning. Results show diffusion and oil swelling as critical recovery mechanisms, and predictive models that couple thermodynamics with transport phenomena offer more realistic production forecasts. Beyond improving unconventional oil recovery, this work highlights broader implications for subsurface processes, including CO₂ storage and hydrogen containment.

Speaker: Maria Barrufet (Texas A&M University)

Invited Lecture: Invited VI

45 Porous Media as a Means to Promote Exchange Processes in Icy Worlds of the Outer Solar System

Beyond the orbit of Mars, most of the solid planetary bodies contain a large fraction of water ice. During the last three decades, a series of space missions to Jupiter’s system (Galileo 1995-2003, Juno (2016-2026), Saturn’s system (2004-2017), dwarf planets Ceres (Dawn (2014-2018) and Pluto (New Horizons 2015), have revealed that several of these icy worlds possess salty water oceans beneath their icy crust. Due to lower gravity and reduced hydrostatic pressure and temperature compared to the terrestrial context, porosity can be maintained over geological timescales and sustained active exchange processes between the different layers constituting their interior. Porous media processes therefore play a key role in promoting chemical and thermal transport in these extraterrestrial environments, including hydrothermal water flow in their porous rocky core, tidally-induced porous flow at the ocean interface and in partially melted layers, and vapor transport through the porous ice near the surface and in active faults. In this presentation, I will review the current knowledge about these icy worlds and highlight a series of active processes revealed by recent exploration, involving porous media.

Speaker: Gabriel Tobie (Laboratoire de Planétologie et Géosciences, UMR 6112, CNRS, Nantes Université)

MS01: 3.1

MS05: 3.1

MS07: 3.1

MS08: 3.1

MS09: 3.1

MS10: 3.1

MS13: 3.1

46 Coupled thermal-hydraulic-mechanical-chemical processes in nanoporous media

Various types of porous media (both unconsolidated and consolidated geological bodies and engineering materials, etc.) and fluids (water, gas, oil, supercritical carbon dioxide, etc.) are closely intertwined with multiple fields such as the environment, geology, and geotechnical engineering, involving soil contamination and groundwater remediation, high-level nuclear waste disposal, carbon dioxide storage, shale oil and gas extraction, hydrogen energy storage, and geothermal utilization. Nano-petrophysical studies focus on rock properties, fluid properties, and the interaction between rocks and fluids, especially for low-permeability geological and engineering media with a large number of nano-scale pores, as their microscopic pore structure (pore size distribution, pore shape and connectivity) controls the macroscopic fluid-rock interaction and the efficient development or preservation of various energy fluids. Such a subsurface system involves a wide range of nm-μm scale pore sizes, various pore connectivity and wettability, in addition to the coupled thermal-hydraulic-mechanical-chemical (THMC) processes of deep earth environments. This presentation showcases the development and application of an integrated and complementary suite of nano-petrophysical characterization approaches, including pycnometry (liquid and gas), porosimetry (mercury intrusion, low-pressure gas physisorption isotherm), imaging (Wood’s metal impregnation followed with field emission-scanning electron microscopy), scattering (ultra- and small-angle neutron and X-ray), and the utility of both hydrophilic and hydrophobic fluids as well as fluid invasion tests (imbibition, diffusion, vacuum saturation) followed by laser ablation-inductively coupled plasma-mass spectrometry imaging of different nm-sized tracers on porous materials. These methodologies have been extended into coupled THMC processes under reservoir-relevant setting, such as the small-angle neutron scattering (SANS) method developed and utilized for the direct observation of rock deformation behavior at a spatial resolution of 1 nm with stresses up to 164 MPa using a self-developed high-pressure cell for mechanistic studies of fluid-solid coupling.

Speaker: Prof. Qinhong Hu (China University of Petroleum (East China))

47 Thermal Maturity and Gas Loading Effects on Transport Properties of Kerogen from Molecular Simulations

Kerogen is the dispersed organic matter in sedimentary rocks from which natural gas and oil are generated over time by thermal maturation. There has been widespread interest in developing atomistic models of kerogen for numerical investigations of adsorption and diffusion behavior. Currently, the most popular kerogen models for use in molecular simulations are "molecular models," which consist in packing and annealing small macromolecules in order to create a 3D kerogen model. This method neglects the cross-linking that occurs as maturity increases, which can strongly control both the amount of pore space and the mechanical properties of kerogen, and is crucial for studying the transport properties of adsorbed fluids. Whereas, Leyssale and coworkers have pursued a different approach to kerogen modeling by using statistical mechanics-based methods to simulate the formation process of kerogen from organic precursors [1], [2], [3], [4]. This new generation of kerogen models, called "mimetic" models, capture the evolution of the cross-linking and chemistry with the maturity [5].

Here, we report on an exhaustive investigation of the self-diffusion coefficient of CH$_4$ in kerogen using eleven different mimetic models of kerogen derived from fatty acid precursors, spanning the range of maturity from immature to post-mature. Kerogen swelling and matrix flexibility must be considered in order to accurately estimate the self-diffusion coefficient for soft matrices [6]. It is well-established now that the collective effects on CH$_4$ (or CO$_2$) transport in kerogen are negligible even when flexibility matters [7], [8]. So, the self-diffusion coefficient can capture the impact of the adsorption and mechanical properties of kerogen on transport. Furthermore, CH$_4$ and CO$_2$ transport in flexible kerogen are known to be quite similar, as both can be modeled by the same free volume theory [8]. Therefore, gas loading was calculated at pressures between 0.1–50 MPa by using a hybrid method that alternates between hybrid grand canonical Monte Carlo and isothermal–isobaric ensemble molecular dynamics simulation in order to explicitly allow for adsorption-induced deformation of the kerogen matrix due to the presence of adsorbed fluid. Thermomechanical and chemical equilibrium are thus simultaneously maintained during adsorption. Molecular dynamics simulation are then performed at a constant temperature of 45 °C in the canonical ensemble starting from the fluid-loaded matrix.

A free volume model inspired by Fujita–Kishimoto theory can fit the observed trends in the self-diffusion coefficient of CH$_4$, with regard to both gas loading and kerogen maturity, in the kerogen models that display significant adsorption-induced swelling. Maturity influences transport in kerogen by both static and dynamic effects. On the one hand—consistent with the experimentally observed gradual stiffening of kerogen during maturation—the flexibility of kerogen matrices decreases with increasing maturity, which reduces the enhancement of diffusive transport due to the fluctuating microstructure. However, more mature kerogen is also more porous, which naturally allows for more efficient diffusion as mean free paths are lengthened due to greater pore connectivity. With regard to gas loading, the fluid content of kerogen mainly influences transport through swelling effects, which again depend on the maturity [9].

Speaker: Mr Alex Eduardo Delhumeau Lozano (Université de Bordeaux)

48 CH₄/CO₂/H₂ Storage and Transport in Nanoporous Media: Microscopic Mechanisms and Scale Effects

Shale reservoirs exhibit a wide distribution of nanopore sizes, ranging from ultrafine pores of roughly 5 nm to larger pores exceeding several hundred nanometers. At the smallest scales, methane adsorption becomes a dominant storage mechanism. To quantify this effect, molecular simulations coupled with an equation of state are employed to characterize CH₄ adsorption in nanopores of various sizes, and the results are incorporated into a lattice Boltzmann (LB) free-energy model via a calibrated fluid–wall interaction formulation. The simulations reveal that adsorption can enhance methane storage by 10–25% in pores smaller than ~20 nm, whereas its influence becomes minimal (<3%) in pores larger than approximately 40 nm.

This scale-dependent behavior allows a natural transition to the flow regime: pores larger than ~100 nm, which constitutes the primary connected flow pathways in shale, but exhibits negligible adsorption effects. Building on this insight, pressure-driven flow and displacement processes are simulated using a multiple-relaxation-time LB model with a combined bounce-back/specular-reflection boundary treatment and regularization algorithm. The model is applied to investigate CO₂ and H₂ transport and storage in depleted shale gas reservoirs, focusing on how these injected gases move through with slippage velocity and displace residual methane in the larger, flow-dominant pore networks. Simulations quantify velocity fields, mass fluxes, apparent permeability, pressure drop, and displacement efficiency, revealing distinct CH₄ displacement mechanisms driven by the contrasting molecular properties of CO₂ and H₂.

Together, these two complementary components (adsorption analysis in ultrafine nanopores and flow modeling in larger and connected pores) provide a coherent, scale-consistent framework for understanding CH₄/CO₂/H₂ storage and transport across the hierarchical pore structure of shale formations.

Speaker: Dr Saman Aryana (University of Wyoming)

49 Pore-fracture connectivity and pore accessibility in overmature marine shale: Insights into fluid transport mechanisms

Pore-fracture connectivity and nanoscale pore accessibility are critical factors influencing gas occurrence, transport, and fluid migration in shale reservoirs. However, due to the extremely low permeability and high heterogeneity of shale components, accurately characterizing these properties remains challenging. This study integrates advanced experimental techniques to investigate the controlling mechanisms of pore-fracture connectivity and pore accessibility in overmature marine shales from the Wufeng-Longmaxi and Niutitang Formations in South China.

To evaluate pore-fracture connectivity, small-angle neutron scattering (SANS) under vacuum and high-pressure conditions, repeated mercury intrusion capillary pressure (MICP), and field-emission scanning electron microscopy (FE-SEM) imaging after Wood’s metal (WM) impregnation were employed. The results revealed that the sealing of pore system by brittle minerals significantly reduces overall connectivity within the shale matrix, resulting in isolated pore networks. While brittle minerals preserve pores within organic matter and clay minerals, they hinder the connectivity between pore systems. This isolation effect has important implications for methane transport, as only 38–78% of pores within 100 nm were accessible to methane in the studied samples. Furthermore, confinement effects were observed to increase methane density in nanopores smaller than 20 nm. This phenomenon results in the formation of nanoscale methane clusters, with densities exceeding those of ideal gas states under equivalent conditions. The novel integration of repeated MICP measurements and FE-SEM imaging after WM impregnation provides a robust framework for evaluating pore-fracture connectivity in shale systems.

In parallel, pore accessibility was systematically investigated using contrast-matching small-angle neutron scattering (CM-SANS) and supplementary experiments, including air-liquid contact angle measurements and spontaneous imbibition. A novel accessibility index was developed to quantify the interaction of fluids with varying wettability in nanoscale pore networks and their temporal dynamics. CM-SANS results indicated that pores larger than 7 nm were predominantly filled with toluene, attributed to the development of organic pores and the connectivity between organic and inorganic pore systems. Conversely, smaller hydrophilic pores (<7 nm) were associated with clay minerals or clay swelling, making them accessible primarily to water. The integration of CM-SANS and MICP further demonstrated that pore accessibility to water and toluene is largely controlled by pore surface wettability and connectivity.

The combined insights from these methodologies link pore-fracture connectivity and pore accessibility, offering a comprehensive understanding of their roles in methane transport and hydrocarbon fluid migration. Pore-fracture connectivity determines the transfer of gas from matrix pores to fracture systems and significantly influences gas storage and transport pathways. Simultaneously, pore accessibility governs fluid migration within the pore network, impacting fracturing fluid imbibition and hydrocarbon recovery efficiency. Understanding the interaction between pore connectivity and wettability offers new perspectives for improving hydraulic fracturing strategies and unconventional reservoir stimulation.

Speaker: Mengdi Sun (Northeast Petroleum University)

MS18: 3.1

Poster: Poster V

MS01: 3.2

MS02: 3.2

MS05: 3.2

MS07: 3.2

MS12: 3.2

MS13: 3.2

50 Adsorption and Thermal Conductivity in Nanoporous Materials: Underlying Molecular Mechanisms and the Rattle Effect

Nanoporous materials are at the heart of numerous important applications: adsorption (gas sensing, drug delivery, chromatography), energy (hydrogen storage, fuel cells and batteries), environment (phase separation, water treatment, nuclear waste storage), Earth science (exchange between the soil and the atmosphere), etc. In this talk, While confinement and surface effects on fluids severely confined in their porosity are well documented, the thermal behavior of nanoporous solids subjected to fluid adsorption remains puzzling in many aspects. With striking phenomena such as the so-called rattle effect, through which fluid/solid collisions decrease the overall thermal conductivity, the solid thermal conductivity and, more generally, heat transfer and dispersion in these complex systems challenge classical approaches (e.g., mixing rules including effective medium approaches fail to capture such effects as shown here). In particular, a robust molecular framework to describe the crossover between the decrease in thermal conductivity through the rattle effect in very narrow pores and the increase in thermal conductivity when replacing vacuum with a fluid phase in larger pores is still missing. Here, using a prototypical model of fluid-filled nanoporous materials, we perform a molecular simulation study to shed light on the parameters that govern the rattle effect in nanoporous solids. First, by varying the fluid/fluid, fluid/solid, and solid/solid interaction strengths as well as the fluid number density and mass density, we unravel the ingredients that lead to the essential coupling between fluid adsorption and phonon transport. Second, despite this complex interplay, inspired by pioneering molecular approaches on the rattle effect, we show that all data obey a simple statistical physics model that relies on the change in the speed of sound due to the fluid adsorbed density and the decrease in phonon lifetime due to scattering by fluid molecules. This framework, which provides a simple formalism to rationalize the thermal behavior of this class of solid/fluid composites, points to a decrease in thermal conductivity upon fluid confinement (up to 30% in some cases). Such an effect paves the way for the design of novel applications involving fluids in interaction with nanoporous materials.

Speaker: Benoit Coasne (CNRS/University Grenoble Alpes)

51 Phase Behavior of CO2-Alkane Mixtures in Nanopores: Insights from Wang–Landau Transition-Matrix Monte Carlo Simulations

The phase behavior of CO2-alkane mixtures plays a central role in fluid transport, storage, and displacement in nanoporous media, with direct relevance to geological carbon sequestration, enhanced oil recovery, gas separation, and CO2 utilization technologies. Under nanoconfinement, phase equilibria, stability limits, and adsorption behavior can deviate substantially from bulk behavior due to strong fluid-surface interactions and restricted pore geometry. Capturing these effects reliably remains a major challenge for both experiments and simulations.
In this contribution, we summarize a series of studies employing the Wang-Landau Transition-Matrix Monte Carlo (WL-TMMC) method to investigate CO2-alkane phase behavior in bulk and nanoporous systems. Compared to conventional Monte Carlo approaches, WL-TMMC provides direct access to free energy landscapes, enabling robust determination of vapor-liquid equilibria, van der Waals loops, and phase stability limits under confinement, quantities that are often difficult or inefficient to obtain using standard techniques. Benchmark comparisons demonstrate that WL-TMMC yields accurate and consistent phase behavior predictions for CO2-alkane mixtures across a wide range of conditions.
We apply this framework to CO2-hexane mixtures confined in nanopores representative of shale inorganic minerals (calcite, quartz, and muscovite mica) and organic matter (graphite), revealing how surface chemistry controls confined phase behavior and adsorption trends. Furthermore, by combining WL-TMMC with free-energy interpolation, we extend simulations of CO2-methane mixtures in metal-organic frameworks and quartz nanopores from a limited set of temperatures to a broad range (273-473 K), enabling efficient prediction of temperature-dependent phase behavior and adsorption without exhaustive simulations.
Overall, this contribution highlights the importance of phase behavior in nanoconfined fluids and demonstrates WL-TMMC as a powerful and versatile tool for studying complex CO2-alkane systems in nanoporous media, providing mechanistic insights and practical guidance for subsurface and energy-related applications.

Speaker: Zhehui Jin (University of Alberta)

52 Exploring Helium Metastability Using Porous Systems

A liquid can sustain tensile stress due to intermolecular attractions, but only up to a critical value beyond which it breaks through the spontaneous formation of a vapor bubble. This process, known as cavitation, is observed for instance in the wake of ship propellers or during sap ascent in trees. Cavitation also occurs during the drying of porous materials, when liquid-filled cavities are connected to an external gas reservoir through narrow constrictions. In this so-called ink-bottle geometry, the liquid inside the cavity is driven into a deeply metastable state by lowering the vapor pressure in the reservoir. In this work, we use independent ink-bottle pores to study cavitation in a controlled and quasi-static manner.

Previous results have shown that Classical Nucleation Theory (CNT) [1–2] accurately describes cavitation in fluids such as nitrogen, provided that surface tension is corrected for nanometric bubbles and that the critical bubble remains small [3] compared to the pore size. In contrast, cavitation in helium is still debated at low
temperature, in the superfluid phase where quantized vortices may act as preferential nucleation sites: all previous experiments which have relied on focused ultrasonic waves to drive the liquid in a metastable state leads to inconsistent values for the cavitation pressure threshold.

To investigate cavitation in the bulk limit for this fluid, we use two model mesoporous systems. The first consists of porous alumina membranes fabricated by anodization of aluminum disks[1]. The second is based on newly designed porous silicon structures produced using nanolithography techniques. The latter system allows for finer control of the ink-bottle geometric parameters, such as the cavity radius, the constriction radius, and the constriction thickness. In both cases, cavitation evaporation can be reach only by reducing the pore apertures down to a few nanometers. This is obtained by atomic layer deposition (ALD).

The samples are subjected to condensation–evaporation cycles using helium at various temperatures while the state of the confined fluid is monitored using a capacitive measurement technique. We present the first helium measurements of the pressure dependence of the cavitation energy barrier and discuss the observed deviations from the predictions of classical nucleation theory (CNT).

Speaker: Paul Coutin (Institut Néel, Université Grenoble Alpes, CNRS)

53 From Molecular Fluctuations to Coupled Transport: A Space- and Time-Dependent Onsager Matrix

Understanding transport phenomena in confined fluids remains a central challenge in liquid-state theory. When liquids are restricted to nanometric dimensions—such as in porous materials, mineral interfaces, and synthetic or biological nanopores—the large surface-to-volume ratio amplifies interfacial interactions and molecular-scale inhomogeneities. As a result, transport becomes highly sensitive to local structure, dynamics, and external gradients, enabling controlled coupling between fluid flow, solute transport, heat transfer, and charge dynamics. These effects underpin a wide range of applications, including energy conversion and storage, water purification, and nanopore-based sensing.

While continuum descriptions of coupled transport are well established at mesoscopic and macroscopic scales, nanoscale confinement introduces dominant contributions from thermal fluctuations, adsorption, electrical double layers, and molecular friction that are not adequately captured by standard constitutive relations. Addressing this regime therefore requires a framework that explicitly accounts for spatial non-locality and temporal memory effects at the molecular level. Here, we introduce a unified approach based on a space- and time-dependent response matrix to characterize transport in confined fluids.

Our framework formulates a generalized linear response relation linking local fluxes of mass, solute, heat, and charge to their conjugate driving fields—pressure, chemical potential, temperature, and electric potential. The resulting coupled response kernel captures non-local and transient correlations arising from confinement. We extract this kernel from equilibrium molecular dynamics simulations using an extended Green–Kubo formalism, thereby establishing a direct connection between microscopic fluctuations and collective transport behavior. This methodology allows us to resolve fundamental processes such as molecular layering, coupled advection–diffusion of solutes and heat, and charge relaxation, and to examine how coupled transport emerges across spatial and temporal scales.

Beyond providing spatially resolved transport coefficients, the present framework offers a transparent bridge to extended continuum descriptions, including dynamical density functional theory and mode-coupling theory. By linking atomistic dynamics to macroscopic transport formulations, our results advance the understanding of solid–fluid interfaces in nanoporous materials and provide a robust basis for modeling coupled transport processes at the nanoscale, with implications for energy, environmental, and subsurface systems.

Speaker: Minh-Thê Hoang (Princeton University)

54 Dynamic migration and recovery mechanism of multi-component shale gas within intra-connected kerogen nanopores

The occurrence and transport mechanisms of methane (CH4) and ethane (C2H6) in organic nanopores are crucial for the efficient development of shale gas reservoirs. While prior studies have examined the adsorption and recovery behaviors of light hydrocarbons (e.g., CH4, C2H6, C3H8) in kerogen nanopores, most analyses have focused on equilibrium states, with limited attention to dynamic production processes. Moreover, existing work has predominantly relied on single slit-shaped nanopore models, overlooking the role of interconnected pore structures. In this work, we therefore construct a model of two interconnected slit-shaped kerogen nanopores with different apertures (2 nm and 4 nm) to investigate the adsorption and extraction of CH4 and C2H6 using coupled grand canonical Monte Carlo (GCMC) and molecular dynamics (MD) simulations. Results show that C2H6 exhibits stronger adsorption affinity than CH4, with smaller pores favoring higher adsorption selectivity. During pressure depletion, the transport partition ratio of CH4 from the dead-end pore to the channel and from the channel to the fracture region greatly exceeds the pore size ratio (~19/32 vs ~4). For C2H6, the transport ratio from dead-end pore to channel is comparable to the pore size ratio (~5.6 vs ~4), whereas from channel to fracture it is significantly higher (~19 vs ~4). During CO2 soaking, nearly all gas components are recovered through the larger pore toward the fracture region. CH4 and C2H6 in the smaller nanopore channel follow a more complex path: from channel to dead-end pore, then to the larger pore, and finally to fractures. The flow partition ratio of CO2 from the fracture into nanochannels matches the pore size ratio. However, CO2 entering the smaller nanopore tends to remain in the channel and does not migrate further into the dead-end pore, meaning the CO2 in dead‑end pores originates mainly from the larger channel. After equilibrium, CH4 shows decreases in both adsorbed and bulk phases, while the adsorbed phase of C2H6 is enhanced. During CO2 soaking, CO2 injection mobilizes mainly the adsorbed hydrocarbons, with little effect on the bulk phase, leading to higher displacement efficiency in smaller pores where adsorbed gas predominates. This work advances the understanding of gas recovery behavior from a dynamic and structurally heterogeneous perspective, providing theoretical insights and simulation‑based guidance for the efficient development of shale gas reservoirs.

Speaker: Mingshan Zhang (Yanshan University)

55 Gas Separation through Nanoporous Graphenes: insights from Molecular Simulations

In the context of energy transition and carbon dioxide emission reduction, the optimization and development of techniques for separating chemical species in the gas phase is a crucial challenge. Membrane separation and selective adsorption are attractive solutions due to their low energy costs compared to other processes (e.g., cryogenic distillation). In this context, innovative materials such as 2D membranes appear promising: in addition to their advantageous physicochemical properties, they significantly reduce the cost of gas compression. Whether to optimize their performance or guide their design, the theoretical prediction of their transport and separation properties is a goal of great importance.

This presentation summarizes work aimed at clarifying the mechanisms of gas adsorption and diffusion in this type of material, focusing on the example of nanoporous graphene membranes. The proposed methodology relies on molecular simulations to document key mechanisms for inclusion in tractable theoretical models, most often in the form of scaling laws or analytical formulas that highlight the link between performance and membrane structural properties [1]. The case of permeation and separation of small gas molecules is considered, and the importance of taking flexibility into account in graphene molecular models is highlighted [2].

Speaker: Dr Romain Vermorel (LFCR, E2S-UPPA)

MS17: 3.2

56 The role of porous media in durability and performance of fuel cells and electrolyzers

The gas diffusion layer (GDL) plays a key role in water management in the proton exchange membrane (PEM) fuel cell; this is now well-accepted in the field. However, the GDL and the porous transport layer (PTL) in electrolyzers are not the only porous materials and interfaces that should be considered for PEM fuel cells and electrolyzers. Together with the catalyst layer, microporous layer, and their interfaces – these porous materials have even deeper impacts on device performance, and even durability, than previously understood due to a variety of factors, such as the heterogenous nature of liquid and gas arrangement, compression behaviour, and nanoscale chemical speciation. This talk will discuss our recent work in this area and the challenges and opportunities ahead.

Speaker: Aimy Bazylak (University of Toronto)

57 Engineering Microporous Layers in Polymer Electrolyte Water Electrolyzers

Porous transport layers (PTLs) are pivotal components in polymer electrolyte membrane water electrolyzers (PEMWEs). At the anode, the PTL is placed between the bipolar plate and the polymer electrolyte membrane and must provide sufficient electrical and thermal conductivity, efficient contact with the catalyst layer (which is deposited on a membrane) to maximize catalyst utilization, mechanical support, and the ability to efficiently remove generated gas bubbles. Furthermore, the corrosive anodic electrochemical environment (oxygen-rich) motivates the use of titanium materials for the state-of-the-art PEMWE PTLs due to the excellent stability of Ti. These include thermally sintered titanium powders, titanium/stainless steel felts, titanium foams, and titanium meshes.1

In recent years, the introduction of microporous layers (MPLs), inspired by polymer electrolyte fuel cells, have further enhanced the device performance2-4. However, there is a lack of fundamental understanding on how to deterministically design these materials. Through a rigorous and systematic study, we aim to elucidate the relationships between the three-dimensional structure of the PTL-MPL, their wettability, and the resulting mass transfer properties and performance. By obtaining this structure-composition-performance relationships, we hope to guide the design of advanced PTL-MPLs from the bottom-up.

In this study we show how MPLs with different structural characteristics such as particle size, pore size and thickness can be produced using ultrasonic spraycoating. Particle size and thickness can be easily controlled using this method, but the porosity of the layer requires more in-dept study. Using a Design of Experiments (DoE) approach, we systematically investigate how spray-coating parameters influence the porosity of the microporous layer (MPL), including binder concentration, cosolvent ratio, and spraying temperature. Subsequently, MPLs with different characteristics can be produced and tested in a PEM electrolyzer to study which MPL properties give the optimal PEM water electrolysis performance.

1.Yuan, X.-Z. et al. The porous transport layer in proton exchange membrane water electrolysis: perspectives on a complex component. Sustain. Energy Fuels 6, 1824 1853 (2022).
2. Lettenmeier, P., Kolb, S., Burggraf, F., Gago, A. S. & Friedrich, K. A. Towards developing a backing layer for proton exchange membrane electrolyzers. J. Power Sources 311, 153 158 (2016).
3. Hasa, B. et al. Porous transport layer influence on overpotentials in PEM water electrolysis at low anode catalyst loadings. Appl. Catal. B Environ. Energy 361, 124616 (2025).
4. Liu, Y et al. Comprehensive Analysis of the Gradient Porous Transport Layer for the Proton-Exchange Membrane Electrolyzer. ACS Appl. Mater. Interfaces 16, 47357 47367 (2024)

Speaker: Rafaël Vos (TU Eindhoven)

58 Pore-Scale Characterization of Stress-induced Compression in Porous Gas Diffusion Layers Using X-ray Computed Tomography and Pore Network Modelling

The transport behaviour of porous electrodes is fundamental to the performance of polymer electrolyte membrane (PEM) fuel cells. As a promising clean energy technology, PEM fuel cells rely on porous media to facilitate the electrochemical conversion of hydrogen and oxygen into water, heat, and electricity. This process depends on the effective diffusion of reactants through porous gas diffusion layers (GDLs) to catalytic reaction sites. However, the multilayer structure of the fuel cell introduces significant electrical and thermal interfacial resistance, necessitating mechanical compression to ensure sufficient interfacial contact while still preserving favourable transport characteristics [1]. Although many studies have investigated the dependence of transport properties on compression, most electrochemical characterizations rely on strain-controlled assemblies, where deformation is defined by displacement rather than applied pressure [2]. Therefore, the effects of stress-controlled compression remain poorly understood, emphasizing the need for quantitative microstructural characterization under variable pressure conditions.

In this study, the relationship between stress-controlled compression, transport properties, and pore-scale characteristics of GDLs is investigated using a novel compression device. This device enables simultaneous X-ray transmission imaging while applying a range of industrially relevant compressive stresses to commercial GDL materials. Under applied compression, the three-dimensional GDL microstructures are captured and digitally reconstructed using X-ray computed tomography (CT). Pore network modelling (PNM) is subsequently employed to quantify the resulting transport properties across increasing compression levels [3]. Therefore, this study uses CT imaging and PNM to elucidate the influence of stress-controlled compression on the pore-scale characteristics of PEM fuel cell GDLs. This research will provide valuable insights for the design of industrial PEM fuel cell stacks, progressing the development of clean energy generation.

Speaker: Shayan Talebi Marand (Bazylak Group, Department of Mechanical and Industrial Engineering, University of Toronto)

59 Alkaline Water Electrolyzers: A Pore Network Modeling Approach

Electrochemical production of hydrogen is an integral part of achieving a sustainable future, offering critical solutions for decarbonized residential heating and energy storage. Alkaline water electrolyzers (AWEs) are particularly promising as they eliminate the need for scarce iridium-based catalysts; however, their efficiency at high current densities is severely limited by gas evolution. In zero-gap configurations, the accumulation of bubbles deteriorates performance by blocking electrode pores. In this work, we employ a Pore Network Modeling (PNM) approach to resolve the pore-scale interactions between the gas phase and electrochemical transport.

The developed framework explicitly couples the saturation of the gas phase with the transport of multicomponent ions and reaction kinetics. This allows for a quantitative analysis of how bubble accumulation drives a substantial increase in overpotential through the reduction of available active sites and the elongation of ionic pathways. Furthermore, the model is utilized to evaluate the potential of novel electrode architectures designed to mitigate gas blockage. This physics-based analysis provides a detailed understanding of transport limitations and establishes a robust computational platform for the future design and optimization of high-performance electrode microstructures.

Speaker: Mohammad Mehrnia (Department of Chemical Engineering, University of Waterloo)

60 Porous Transport Layer Optimization via Additive Manufacturing of Inconel 718 Lattice Structures

Hydrogen production via alkaline water electrolysis (AWE) is an important clean energy technology; however, its efficiency is challenged by poor gas-liquid transport, high ohmic losses, and material degradation. Additive manufacturing (AM), specifically laser powder bed fusion (LPBF), enables the fabrication of porous transport layers (PTLs) with precise control over porosity and feature resolution, thereby improving gas transport and overall system performance.

This research focuses on optimizing porous transport layer (PTL) structures by refining printing parameters for Inconel 718 and implementing intricate lattice designs. A diamond lattice with a unit cell size of 2x2x2 mm³ and wall thicknesses ranging from 0.1 mm to 0.5 mm is designed to investigate the ideal structure for improving bubble transport. Aside from lattice structures designed to enhance bubble removal, process-driven stochastic pores can further optimize gas-liquid interactions and increase the number of electrochemical sites by increasing the overall effective surface area. These stochastic pores are generated by adjusting hatch spacing (100-500 μm) and rotational angles (67, 60, and 90°) to create lack-of-fusion pores across the electrode. An investigation into optimal process parameter selection is conducted to achieve repeatable, high-resolution geometric fidelity across various pore structures, using advanced characterization techniques, such as X-ray computed tomography (XCT), to analyze porosity distribution and structural properties.

The combination of lattice geometries and process-driven porosity yields porosity ranges of 40-80%, hydraulic pore sizes of 0.1-0.9 mm, and tortuosity values of 1-4. These properties are expected to enhance mass-transport efficiency in anion-exchange water electrolysis (AWE) systems by enabling a diverse array of pore types, sizes, and shapes within the PTL structure. The performance of the pore network is evaluated through electrochemical testing, which includes linear sweep voltammetry, whereby at 1V, the achieved current ranged from 120 to 250 mA, while the double-layer capacitance varied from 500 to 1000 µF/cm². The resulting electrochemical performance validates the design's efficacy.

By refining design and manufacturing parameters, in tandem with electrochemical testing, this research will establish a repeatable method for producing high-resolution lattice structures with controlled porosity. The findings will inform manufacturing protocols and design guidelines that can be integrated into existing AWE systems, leading to improvements in efficiency, geometric precision, and gas transport performance in additively manufactured PTLs, thereby supporting the enhancement of clean hydrogen production technologies.

Speaker: Tomisin Oluwajuyigbe (University of Waterloo)

61 Processing–Structure–Performance Relationships in Pristine and Recycled Catalyst Layers for CO₂ Electrolysis

Global warming and the urgency of achieving net-zero greenhouse-gas emissions by 2050, as articulated by international frameworks such as the Paris Agreement (IPCC 2023) [1], require scalable electrochemical CO2 reduction (CO2R) technologies powered by renewable electricity [2]. A critical component of CO2R systems is the catalyst layer—a reactive porous medium in which coupled multiphase, multicomponent transport and electrochemical reactions occur—and whose physicochemical properties (e.g., catalyst dispersion, ionomer distribution, wettability, and porosity) directly govern activity, selectivity, and stability of the system [3]. Despite its importance, catalyst-layer fabrication remains a major bottleneck: conventional ink-based methods often suffer from poor reproducibility, as minor variations in formulation and processing strongly affect catalyst distribution, wetting behavior, and mass transport [4]. Moreover, catalyst layers frequently rely on resource-intensive materials that are difficult to reclaim at end-of-life, and recycled catalyst materials often exhibit degraded performance due to surface chemical modification and catalyst agglomeration [5].
Here, we examine how catalyst-ink preparation methods influence ink composition, dispersion state, and deposition method, towards decoupling intrinsic catalyst properties from processing-induced variability in CO2R electrodes. The produced catalyst layers are characterized using scanning and transmission electron microscopy (SEM, TEM), X-ray diffraction, and operando electrochemical diagnostics, to extract structure–transport–reaction descriptors. We will discuss how properties—including pore size distribution, tortuosity, ionomer coverage, catalyst agglomeration, and gas–liquid–solid interfacial accessibility—govern activity, selectivity, and stability of the system. Beyond pristine systems, we extend this methodology to inks formulated from reclaimed catalyst materials.

Speaker: Dr Ashkan Irannezhad (University of Toronto)

MS20: 3.2

MS02: 3.3

MS04: 3.3

62 Real Rock Microfluidics Investigation of Solute Diffusion in Biofilm-Rock Systems

Biofilms are nearly ubiquitous in both natural and engineered subsurface systems, with relevance to processes ranging from groundwater contamination to thief zone remediation. The interaction between biofilms and permeable media is well-understood to be bidirectional: just as biofilm accumulation is mediated by both mass transport considerations and the physical stresses associated with fluid flow, biofilms can also significantly impact mass transport and fluid flow. As such, understanding and predicting biofilm behavior in biofilm-rock systems requires us to capture both flow through the rock and the associated advective transport as well as diffusive transport within both the rock and, potentially, the biofilm. Microfluidic experiments and modeling studies have significantly advanced our understanding of such systems. At the same time, some attributes of natural systems, such as mineral surface properties and heterogeneity in pore structure, are challenging to capture with these tools.
Here, we illustrate how solute diffusion through natural rock matrices of different porosities can affect, and be affected by, biofilm growth. We also explore the impact of matrix porosity on the efficacy of fracture sealing via ureolytic microbially-induced carbonate precipitation (MICP). Building upon recent advances in real rock microfluidics, in which natural rock samples are incorporated into microfluidic devices, we position porous rock chips between two flow channels. This setup mimics two fractures separated by a porous rock matrix. Through the use of conservative tracers, we quantify the diffusive flux through the porous matrix before, during, and after biofilm cultivation in one channel. We combine this experimental setup with non-destructive X-ray computed tomography to qualitatively compare solute transport through different matrices and at different stages of biofilm growth. Biofilm morphology and resistance to shear stress are found to depend on both matrix porosity and heterogeneities inherent to the pore structure of natural rocks. When urea-hydrolyzing biofilms are used to carry out carbonate precipitation, these effects may be even more pronounced.

Speaker: Eva Albalghiti (The University of Michigan)

63 Monitoring hydrogenotrophic activities in deep underground reservoirs: from lab scale to case study

The injection of gases such as CO2 and H2 into deep geological formations is a key strategy for carbon sequestration and energy storage. However, the success of these operations depends on our ability to monitor and predict the microbial response to such perturbations. Indigenous microorganisms can trigger biochemical reactions leading to gas conversion, reservoir souring, or bioclogging. Investigating these processes requires tools capable of mimicking the extreme conditions of the deep subsurface (i.e. high pressure, salinity) while providing high-resolution data on metabolic activities.
To address this, we developed optically transparent high-pressure multiscale reactors that allow for the monitoring of autotrophic microbial growth via in situ and ex situ characterization. The primary advantage of this technology is the ability to maintain the system at pressure (up to 100 bar) throughout the entire process, avoiding decompression biases and enabling also direct optical access (UV-Vis).
In the first part of this study, we established a laboratory-scale baseline using the model methanogenic strain Methanothermococcus thermolithotrophicus. We investigated the impact of H2/CO2 partial pressures and hydrodynamic conditions (i.e. stirred vs. unstirred) on methane production. Results demonstrated that unstirred conditions favor biofilm formation, which significantly extends the range of gas partial pressures under which the strain remains metabolically active. This underlines the critical role of spatial organization and mass transfer in hydrogenotrophic processes.
In the second part, we applied this methodology to a real case study using brine samples from depleted gas reservoirs (potential UHS sites). Through metagenomic analysis, we characterized the indigenous community and enriched a hydrogenotrophic co-culture including sulfate-reducing bacteria. High-pressure millifluidic and microfluidic cultivations revealed a metabolic symbiosis within this co-culture, where hydrogen consumption and microbial resilience are governed by the interplay between pressure and local physical constraints.
Overall, combining model strains and real reservoir co-cultures demonstrates that hydrogenotrophic activities are not only governed by thermodynamics but are strongly influenced by the local physical environment. This dual approach using multiscale reactors offers a direct method to evaluate biogeochemical risks, such as gas loss and souring, by capturing microbial behavior under representative reservoir conditions.

Speaker: Dr Anaïs Cario (ICMCB-CNRS)

64 Raman spectroscopic detection and quantification of microbial reactions in the pore network within a microfluidic chip

In many contexts, microbial reactions are studied in batch-type reactors to identify conditions necessary for active microbial metabolism and to determine reaction rates or kinetics of selected reactions. One example is the microbial oxidation of hydrogen (e.g. Dohrmann & Krüger, 2023; Dopffel et al 2023) in the context of subsurface storage of hydrogen as energy carrier.
Within batch-type reactors (or serum bottle experiments), a single large gas-fluid interface may limit the replenishment of dissolved hydrogen by mass transfer from the gas phase (cf. Strobel et al 2023). In addition, the single static interface present in the batch-type reactors and the analysis of bulk fluid or gas samples only prevents investigation of spatial chemical gradients of e.g. dissolved hydrogen concentrations or dissolved redox-acceptor concentrations developing on the micrometre scale in subsurface porous rocks. There these gradients most likely will govern growth rates, overall rates of biofilm formation - and more important, its localization with respect to pore throats (Hassannayebi et al 2021). This in turn will affect the overall microbial growth, hence microbial oxidation of hydrogen and formation of by-products, and changes in permeability. First attempts to assess the importance of localized biofilm formation used either packed column experiments (cf. Mushabe et al. 2025) without spatial resolution or were confined to the spatially resolved optical observation of biofilm growth (Liu et al. 2025) without information on chemical gradients.
Therefore we started to develop methods combining optical and Raman spectroscopic techniques enabling us to quantify the concentrations of dissolved ions in the aqueous phase with microbial cells and the partial pressure of gases in adjacent gas phase in microfluidic chips on the micrometre scale. We present data for a first example, the spatially resolved observation of changes in sulphate concentration and hydrogen partial pressure due to microbial oxidation of hydrogen by sulphate-reducing bacteria (strain Oleidesulfovibrio alaskensis) inside a microfluidic chip. It was possible to quantify the decrease of the concentration of sulphate down to 5 mM and hence determine the localized rate of microbial sulphate reduction. In adjacent gas pockets in the pore space, the decrease of the hydrogen partial pressure could be quantified down to 0.01 MPa. The ability to constrain the chemical composition within the chip with high spatial resolution enables addressing the above-mentioned questions of governing effects of evolving chemical gradients on microbial growth, biofilm formation and localization in the pore space - even under (stopped) flow conditions. We outline the next steps towards assessing in chip effective microbial rates in the context of factors as local sulphate concentration, limitations of e.g. hydrogen supply, influence of fluid velocity etc. - necessarily including parallel pore-scale modelling of the systems investigated.

Speaker: Christian Ostertag-Henning (Federal Institute for Geosciences and Natural Resources)

65 Biofilm-functionalized pervious concrete: the first iteration of an engineered living material for removing microplastics from stormwater

Microplastic contamination (plastic particles < 5 mm) is a growing concern. One potential solution is to use biofilms to trap and remove microplastics from contaminated water. Naturally forming biofilms (for example, those growing on submerged surfaces in rivers) have been observed to collect microplastics within their sticky extracellular polymeric substance. This study aims to bio-mimic this observation by growing biofilm on the surface of pervious concrete with the intention of removing microplastics transported via stormwater and thereby creating an engineered living material. Pervious concrete is an excellent alternative pavement strategy for infrastructure such as sidewalks, parking lots, and driveways because it can manage stormwater by reducing runoff, recharging groundwater, filtering out pollutants, and minimizing flood risks. Briefly, engineered living materials modify existing materials with living organisms, thus providing the original material with additional functionality. In this study, Bacillus mojavensis biofilm was established on pervious concrete aggregates using a continuous flow reactor system. Once a robust biofilm was formed (~10^7 cfu/g-concrete), a solution of microplastics (1000 mg/L) was injected into the system, and the removal efficiency was calculated using FlowCam analysis. Microplastic solutions were initially passed through columns containing concrete without biofilms to determine any baseline particle capture with the concrete alone. Scanning electron microscopy was used to observe microplastics trapped within the biofilm matrix covering pervious concrete aggregates. These experiments represent the first step towards developing a system that inhibits microplastic transport from terrestrial to aquatic environments due to stormwater runoff.

Speaker: Kayla Bedey (Montana State University)

66 Investigations on the Reduction of the Porosity and Water Absorption Properties of Recycled Brick Aggregate by MICP Treatment

The use of mixed recycled aggregates (RMA) for concrete is limited according to current German standards (DIN 1045-2). The coarse natural aggregate is only allowed to be replaced proportionally. RMAs contain a high amount of brick material, which results in high porosity and water absorption properties. This primarily influences the consistency of fresh concrete. If recycled aggregate consists exclusively of crushed bricks or masonry construction and demolition waste, it is also referred to as recycled brick aggregate (RBA), which is not yet regulated for use in recycled aggregate concrete. For this reason, a biodeposition approach was chosen to optimize the properties of the RBA. There are various applications based on microbial-induced calcium carbonate precipitation (MICP), whose promising approaches in construction have already proven effective [1]. This study tested an MICP treatment designed to optimize the water absorption properties of RBA. A bacterial culture of Sporosarcina pasteurii DSM 33 was used in combination with urea and calcium chloride to precipitate calcium carbonate. The aim is to use the CaCO$_{3}$ precipitate to form a layer on the surface of the RBA, thereby filling the pore space and significantly reducing the porosity [2]. For the treatment of RBA, a process with multiple short immersion intervals and intermediate vacuum extraction was used to apply the liquid MICP components. Up to 5 treatment intervals were carried out, and the water absorption was determined according to DIN EN 1097-6:2022-05 after each step. The results show a trend toward a steady reduction in water absorption, depending on the number of MICP treatments, where the initial water absorption can be reduced by 40.6%. García-González et al. [2] found similar results and stated that ceramic aggregate may offer particular advantages for MICP treatment due to its high surface roughness. In addition, changes in bulk density and apparent grain density were determined, which are directly associated with a reduction in porosity. According to Sun et al. [3], the reduction in porosity primarily affects pores in the range of 10 – 300 nm, with capillary pores or large pores (>1000 nm) decreasing to a lesser extent. Mineralogical investigations (SEM and XRD) confirm the formation of CaCO$_{3}$ on the surface of the RBA, whereas mainly vaterite crystals could be detected. MICP treatment of recycled aggregate appears to be an effective approach for reducing porosity and water absorption. However, further research is needed to investigate the pore space filling mechanism with precipitated CaCO$_{3}$ in order to optimize the MICP treatment method.

Speaker: Ms Brigitte Nagy (Munich University of Applied Sciences HM, Department of Civil Engineering, Germany)

67 Spatio-temporal Characteristics Of A Proliferating Saccharomyces cerevisiae Clog

Bioclogging is a process that result from the separation of biological particles from a fluid by a membrane; it has many environmental and sanitary applications. It results in a reactive porous medium with emerging properties: cells are deformable, can proliferate, consume nutrients and oxygen, and die. These specific features affect the structure and behavior of the porous medium. The coupling between proliferation, clog growth, and nutrient consumption can lead to a nutrient-limited environment, altering the proliferation of the organisms [1]. Bioclogging can thus be used to study the dynamics of reactive porous media under environmental constraints. Our objective is to investigate the spatio-temporal features of cell proliferation within a yeast assembly perfused with nutrients at the microscopic scale.
The model organism is Saccharomyces cerevisiae. A quasi-2D microfluidic system was developed, in which yeast cells are retained by a pore and continuously perfused with culture medium [2]. Two distinct growth regimes are observed during clog formation, corresponding to different states of the clog. In the initial phase, clog growth is exponential, associated with uniform proliferation throughout the clog. After a few hours, the clog length evolves linearly with time. Two distinct regions emerge: one proliferative, the other quiescent – as demonstrated by biological marking. We are also able to quantify local proliferation rates within the clog using local displacements. These results highlight the coupling between bioreactive flow and proliferation: growth reduces the flow rate, which in turn reduces the proliferation rate.
A mathematical model has been developed to support the experimental observations. It relies on three key components: a Monod-type proliferation law dependent on nutrient concentration, an advection-diffusion-reaction nutrient transport equation, and a Darcy description of flow through the clog. These equations are coupled to capture the interplay between cell growth, nutrient depletion, and flow reduction. The model successfully reproduces the transition between the observed growth regimes, as well as the emergence of spatially differentiated zones within the clog.

Speaker: Mathieu Ghenni (Institut de Mécanique des Fluides de Toulouse)

MS06: 3.3

MS08: 3.3

MS09: 3.3

MS10: 3.3

MS15: 3.3

68 PCP-GAN: Property-Constrained Pore-scale Image Reconstruction

Accurate characterization of porous media at the pore scale is fundamentally challenged by two critical limitations: the scarcity of core data available only at discrete well locations, and the high spatial heterogeneity inherent in rock formations that renders small, randomly sampled sub-images non-representative of bulk core properties. This work introduces PCP-GAN, a tailored multi-conditional Generative Adversarial Network (cGAN) framework, designed to synthesize geologically accurate pore-scale images with precise and simultaneous control over multiple petrophysical properties.

The unified cGAN framework was trained on an integrated dataset of thin section imagery derived from four distinct geological depths (1879.50 m to 1943.50 m) within a marine carbonate formation. By simultaneously utilizing both sample depth and porosity as conditional inputs, the model was forced to learn both universal pore network principles and the unique, depth-specific geological characteristics of the sequence. This conditioning enabled the model to accurately capture a wide spectrum of pore architectures, ranging from high-porosity grainstone fabrics to complex, low-porosity crystalline lithologies with anhydrite mineral inclusions.

PCP-GAN demonstrated high precision in property generation, achieving an R-squared value of 0.95 for porosity control across all tested geological conditions, with mean absolute errors consistently below 0.02. Beyond quantitative metrics, visual fidelity analysis confirmed high mineralogy accuracy, specifically, the model successfully preserved features critical to geological interpretation, such as dolomite grain boundaries, angular crystal morphology, and the sharp delineation of non-porous anhydrite patches in the crystalline samples (Figure below). Furthermore, comprehensive morphological analysis confirmed that the generated images preserved critical pore network characteristics, including the average pore radius, specific surface area, and tortuosity, within standard geological tolerances.

Crucially, we developed a validation framework to benchmark the representativeness of the generated images against laboratory-measured core data (porosity and permeability). Optimized synthetic images were selected based on a dual-constraint error metric. These generated images exhibited a combined property deviation (dual-constraint error) of only 2–12% from the core targets. This performance stands in contrast to the high spatial variability observed in the real rock, where randomly extracted sub-images from the same cores showed significantly higher property deviations, ranging from 36–570%. This remarkable improvement indicates that the framework successfully addresses the core representativeness challenge in digital rock physics.

This breakthrough ability to produce synthetic rock images that are quantitatively more representative of bulk formation properties than natural, randomly sampled sub-volumes offers a powerful new tool. It significantly enhances the reliability and applicability of digital rock physics modeling and is a critical advancement for characterizing sparse-data environments relevant to energy storage, carbon capture and storage, and sustainable groundwater resource management.

Speaker: Dr Arash Rabbani (University of Leeds)

69 A Graph Neural Network Framework for Upscaling the Pore Network Modeling Calculations

This study proposes an artificial intelligence (AI)-based framework for upscaling single-phase and two-phase quasi-static simulation results from small subsamples to larger porous media domains. Several simulation methods, including direct numerical simulation (DNS) and pore-network modeling (PNM), are employed to elucidate the transport phenomena within the pore space. While in DNS, the pore space geometry is directly discretized, in PNM, the complex pore morphology is reduced to a simplified network of pores and throats with idealized geometries [1], drastically reducing the computational requirements [2]. Notwithstanding, it remains computationally demanding when applied to very large samples.
To address this challenge, we utilize graph neural networks (GNNs) for upscaling the pore pressure and capillary pressure results from small to large 3D samples. The GNNs are powerful machine learning frameworks capable of directly learning from graph-structured data, such as pore networks [3, 4]. The core principle of a GNN is the iterative aggregation and transformation of information exchanged between interconnected neighboring nodes (pores) [4].
Our framework begins with a binarized tomography of the porous medium, from which both a subsample and the full sample are selected (see Figure). Pore networks are extracted for each, but fluid flow simulations are performed only on the small subsamples to reduce computational expense. The extracted pore network of the subsample is used as input to the GNN, while the node-level fluid flow simulation results serve as the training targets. The GNN is thus trained to predict flow parameters directly from graph data. Once trained, the model is applied to the pore network of the full sample to predict the same flow parameters without additional simulations.
The framework was evaluated using three X-ray tomography images of sandstone samples, including Bentheimer, Castle Gate, and Berea. Results demonstrate that the proposed method achieves high accuracy in upscaling pore pressure and capillary pressure from subsamples to full rock volumes. For instance, the upscaling from the train image dimensions of 2003, 4003, 6003, and 8003 to a validation image of 10003 was conducted, yielding R-squared values of 0.83, 0.91, 0.96, and 0.98, respectively. The training took ~20 seconds, and the upscaling took ~3 seconds, indicating the very computational efficiency of the method. Further assessment indicated the model's ability for transfer learning. While the model was trained on the Bentheimer data, the capillary pressure of the Castle Gate sample is successfully predicted by an R-squared of 0.96.

Speaker: Mehdi Mahdaviara (Hydrogeology group, Utrecht University)

70 Generalizable 3D Multiphase Segmentation for Pore-Scale Micro-CT: A Mamba-Unet

Three-dimensional multiphase segmentation of pore-scale X-ray CT imagery in porous media faces a fundamental bottleneck that extends beyond achieving high in-domain accuracy on individual volumes. A key limitation lies in the absence of artificial intelligence methods that can function as unified segmentation models across multiple samples. Existing deep learning approaches for porous media segmentation often suffer from pronounced domain shift when variations arise in rock type, imaging system and acquisition parameters, or fluid-bearing conditions. Consequently, models typically require retraining or repeated fine-tuning for each new sample, which substantially increases both annotation effort and computational cost. This sample-specific training paradigm restricts the scalability and reusability of AI-based segmentation within digital rock analysis and pore-scale multiphase flow imaging workflows.
To address these challenges, we propose Mamba‑UNet, an efficient 3D segmentation framework built around State Space Models (SSMs), designed to improve cross-sample and cross-scanner generalization while maintaining computational efficiency. We develop a micro‑CT–specific augmentation strategy to better account for intrinsic noise and structural variability, and to emulate shifts in imaging conditions and intensity statistics. We further introduce a tri-orientated scan collaboration module to capture long-range spatial dependencies and global contextual information throughout the volumetric domain. In addition, an uncertainty estimation mechanism is incorporated to adaptively assess feature reliability during multi-scale fusion, enhancing fusion robustness under domain shift.
The proposed Mamba-UNet framework is evaluated on publicly available Bentheimer sandstone and Ketton carbonate datasets. Experimental results demonstrate that the model achieves competitive segmentation performance and efficient inference on these benchmarks, while also maintaining strong segmentation quality on an unseen Bentheimer sandstone dataset excluded from training. Furthermore, the method exhibits stable performance on fluid-bearing Bentheimer sandstone and Ketton carbonate volumes acquired using different imaging systems. These results highlight the reusability and scalability of the proposed approach for multi-sample digital rock workflows, providing more reliable 3D segmentation to support high-throughput pore-structure quantification and pore-scale multiphase flow studies.

Speaker: Rui Zhang (Imperial College London; China University of Petroleum Beijing)

71 AI-Assisted Upscaling from pore to continuum scale during particle deposition in fractured porous media

Abstract
Particle deposition in fractured porous media induces pore-scale alterations that substantially influence macroscopic transport characteristics, particularly fracture permeability. These processes are crucial in numerous natural and engineered systems, including subsurface flow, filtration, biofouling, reactive transport, and energy-related applications [1]. In fractured rocks, fluid flow is mainly controlled by fracture openings, so particle deposition can strongly alter flow paths and permeability. [2]. Pore-scale simulations offer significant insights into these processes; yet their direct application to large-scale fractured systems required considerable computational resources [3]. AI-assisted upscaling frameworks offer a powerful alternative by enabling the transfer of pore-scale information into continuum-scale models without repeated high demand simulations [4]. In this study, an AI-based multiscale framework is developed to upscale deposition-induced pore-scale changes in a fractured carbonate into a continuum-scale flow model. A representative fracture with an elongated geometry is extracted from 2-D images of a realistic sandstone sample, and particle injection is applied at the fracture inlet. Particle transport and retention are simulated using a Eulerian-Lagrangian approach developed in [5], allowing non-uniform deposition along the fracture to be resolved under different flow conditions.
After deposition, fracture aperture and permeability are measured at several locations along the fracture, allowing the effect of particle accumulation on flow properties to be quantified. These pore-scale measurements are used to create a dataset that describes how fracture properties change with distance from the injection point under different flow conditions. An AI model is trained using this dataset to predict how particle deposition alters fracture porosity and permeability, explicitly considering the spatial evolution of these parameters along the flow direction. The trained model delivers precise pore-scale predictions of property alterations generated by deposition, eliminating the need for supplementary pore-scale simulations. The AI-predicted properties (porosity and permeability) are subsequently integrated into a continuum-scale fracture model by allocating spatially variable characteristics to the Darcy-scale grid, so enabling pore-scale deposition effects to be accurately represented in the larger-scale flow simulation.
This method facilitates the computation of pressure drop, flow rate, and effective fracture transmissivity, while accounting for the effects of heterogeneous particle deposition. Validation is performed at the pore scale by comparing AI-predicted parameter values with those extracted from independent pore-scale simulation data. The findings indicate that the AI-driven framework attains significant prediction accuracy at the pore size and facilitates a robust and dependable upscaling of deposition-induced alterations to the continuum scale, serving as an effective instrument for modeling transport in fractured porous media.
Acknowledgment
The authors sincerely thank Johannes Gutenberg University Mainz for making the licensed GeoDict (Math2Market GmbH) available for this research.

Speakers: Javad Razavinezhad (Department of Petroleum Engineering, Faculty of Chemical Engineering, Tarbiat Modares University, Tehran, Iran), Dr Saeid Sadeghnejad (Institute for Geosciences, Applied Geology, Friedrich-Schiller-University Jena, 07749 Jena, Germany)

72 Physics-informed machine learning for estimating permeability and dispersivity distributions in three-dimensional heterogeneous porous media

Flow and reactive transport in porous media are very important to improve our understanding of physical and chemical processes related to various geoscience and environmental applications such as enhanced geothermal systems, in-situ critical mineral and element recovery, unconventional resources recovery, and environmental fate and transport. One of the overarching challenges in improving prediction of flow and transport processes in porous media is how confidently we can estimate heterogenous permeability (and porosity) fields and associated parameters. Recent advances in machine learning (ML) involving advanced architectures and learning methods show promising results to enhance our ability to estimate heterogeneous subsurface properties and improve inverse modeling approaches. However, most of these ML methods have been evaluated with relatively simple synthetic cases. In this work state-of-the-art 3D tracer concentration datasets collected from non-reactive tracer transport experiments in a 3D sandbox setting using magnetic resonance imaging are utilized. Various ML workflows including Inverse physics-informed neural operator and ensemble smoother-multiple data assimilation approach with deep generative prior models are trained and evaluated to estimate 3D permeability fields and dispersivity distribution using spatio-temporal tracer concentrations in 3D sandbox experiments. These estimated fields with uncertainty quantification will be compared with traditional inverse modeling results. This work will provide outstanding benchmark datasets that can be used for validation of machine/deep learning approaches. SNL is managed and operated by NTESS under DOE NNSA contract DE-NA0003525.

Speaker: Hongkyu Yoon (Sandia National Laboratories)

73 GeoSlicer a Platform for Digital Rock Physics: Integrated Machine Learning, Data Preparation, and Generative AI with SinGAN

The digital characterization of porous media is undergoing a profound transformation driven by Artificial Intelligence (AI). However, the adoption of deep learning in Digital Rock Physics (DRP) is often hindered by the fragmentation of scientific workflows requiring separate, disconnected tools for image visualization, data annotation, and model training. We present GeoSlicer, an open-source, multi-platform software based on the robust 3D Slicer architecture, designed to unify these critical tasks into a single, cohesive environment. GeoSlicer democratizes access to advanced AI by bundling industry-standard deep learning frameworks, including TensorFlow and PyTorch, directly within its Python environment. This integration eliminates the complex dependency management that typically challenges geoscientists, enabling the seamless deployment of neural networks for reservoir characterization.

GeoSlicer excels as a comprehensive workbench for machine learning data preparation, addressing the "ground truth" bottleneck that limits supervised learning. It offers a suite of advanced annotation tools, allowing users to rapidly generate high-quality semantic labels for 3D micro-CT and thin-section imagery. Features such as semi-automated segmentation (e.g., fast marching, region growing), logical masking, and interactive thresholding streamline the creation of training datasets. Once annotated, data can be efficiently processed using internal pipelines that leverage HDF5 and out-of-core handling of massive volumes (e.g., $3000^{3}$ voxels), ensuring that multiscale data,from microCT, coreCT, well logs and thin sections,can be analyzed on standard workstations. The platform further supports real-time training monitoring via integrated TensorBoard visualization, closing the loop between geological interpretation and model performance.

In the context of AI for generating multiscale images, integrating microCT and coreCT data, for example, we modified the SinGAN (Single Image Generative Adversarial Network) model by integrating 3D convolutional layers, enabling it to process volumetric data. To address the memory constraints inherent in the original architecture, we developed Early Cropping and Patched Inference techniques, enabling generating images of $10^{10}$ voxels. We have named this 3D rock generation model as RockSinGAN, which was integrated into the GeoSlicer ecosystem, marking a significant leap in digital rock generation. Unlike traditional deep learning models that require thousands of training examples, RockSinGAN allows for the training of a generative model using a single representative 3D reference image. This capability enables the synthetic generation of large, statistically equivalent 3D rock volumes from limited input data. The model has a pyramidal resolution architecture which allows the integration of rock images in different scales as conditioning data. By generating stochastic realizations of the pore structure, RockSinGAN facilitates rigorous multiscale analysis and uncertainty quantification, providing researchers with a new tool to assess the impact of heterogeneity on rock properties essential in reservoir models.

Speaker: Rafael Arenhart (LTrace)

MS17: 3.3

74 Rethinking Electrode Choice: Matching Porous Microstructures to Electrolyte Properties in Redox Flow Batteries

Porous electrodes are performance- and cost-defining components of redox flow batteries (RFBs), governing electrolyte transport, accessible surface area for electrochemical reactions, and mass, charge, and heat transport within the cell [1]. Yet, the carbon fiber electrodes most commonly used today were originally developed for fuel cells and are not tailored to the diverse kinetic and transport requirements of liquid-phase redox chemistries. As a result, electrode-electrolyte mismatches, arising from trade-offs among conductivity, reaction kinetics, surface area, thickness, and pore size distribution, can significantly limit RFB performance.
Our prior computational work underscored this challenge by demonstrating that optimal electrode architectures are highly electrolyte-specific. Using an in-house genetic algorithm coupled to a pore network model, we showed that different redox chemistries (VO²⁺/VO₂⁺ and TEMPO/TEMPO⁺) converge toward distinct microstructural optima [2]. For example, sluggish kinetic systems such as all-vanadium chemistries benefit from high surface area, whereas electrolytes with low ionic conductivity require high through-plane permeability. These insights motivated a systematic experimental investigation into how commercial electrodes perform across different chemistries.
In this study, which will be the main topic of this presentation, we evaluated three widely used porous electrodes, carbon cloth, paper, and felt, across three electrolyte systems: all-vanadium, all-iron, and an aqueous organic chemistry. Through combined half-cell and full-cell testing, we found that each electrolyte exhibits a unique optimal electrode configuration, and that, in several cases, asymmetric electrode selection between the two half-cells yields superior performance. These results highlight the strong coupling between reaction kinetics, ionic and electronic transport, and electrode architecture, demonstrating how pore-scale structure governs electrolyte-dependent transport regimes. Importantly, they show that even within the constraints of commercially available materials, substantial performance gains can be achieved by matching electrode microstructure to the electrolyte’s physicochemical properties.
Building on these insights, we explore additive manufacturing as a route to move beyond traditional fibrous electrodes [3]. Triply periodic minimal surface (TPMS) architectures offer deterministic, multiscale control over porosity, tortuosity, and surface area, enabling the design of electrode structures tailored to specific electrolyte chemistries and operating conditions. This work demonstrates the potential of additive manufacturing to fabricate customized porous electrodes with enhanced electrochemical performance and reduced hydraulic resistance, paving the way for purpose-built RFB materials.

Acknowledgments
The authors gratefully acknowledge funding from the Natural Sciences and Engineering Research Council of Canada (NSERC) through the Discovery grant program (RGPIN-2025-04132).

Speaker: Maxime van der Heijden

75 Porous electrode failure in spent LFP batteries: a multi-analytical investigation of surface degradation

Abstract

The widespread retirement of lithium iron phosphate (LFP) batteries from the electric vehicle sector necessitates an advanced understanding of the failure mechanisms within their complex porous composite electrodes (Yao et al., 2024). Although LFP is widely regarded as a structurally robust cathode material, significant performance degradation is frequently observed at the end of service life. The bulk crystal structure of spent LFP particles is often well-maintained, electrochemical performance is significantly impaired by surface-related deterioration and microstructural failure (Lv et al., 2025). This study provides multi-analytical evidence of these mechanisms using ICP-OES, XRD, FTIR, SEM, and TEM. ICP-OES analysis shows a fundamental Li deficiency in degraded cathodes, with Li/Fe molar ratios frequently dropping from approximately 1.0 to values as low as 0.67–0.88, indicating a substantial loss of active lithium inventory. XRD and Rietveld refinement demonstrate that this Li loss triggers an irreversible phase transition, resulting in the coexistence of the LFP phase and a degraded FePO₄ (FP) phase within the particles. These phase changes indicate incomplete lithiation and spatially heterogeneous electrochemical activity. Furthermore, FTIR indicates structural disorder consistent with Li–Fe antisite formation, which shifts characteristic P–O bond vibrations to higher wavenumbers, effectively blocking one-dimensional Li⁺ diffusion channels.
Together, these chemical and crystallographic alterations reduce the fraction of electrochemically accessible active material within the porous electrode. Rather than uniform bulk degradation, lithium loss and defect accumulation promote spatially heterogeneous reaction environments, leading to localized transport limitations and uneven utilization of the LFP framework across the electrode thickness.
At the microscopic level, imaging analysis indicates that porous media integrity is compromised by the growth of non-uniform cathode–electrolyte interphase (CEI) layers. These interphase products act as a pore-clogging barrier, effectively increase transport resistance, and constrict pore pathways (Zhang et al., 2024). Heterogeneous phase distributions are found, with disordered FP phases predominating at particle surfaces while the LFP bulk remains intact. SEM provides evidence of microstructural disconnection, manifesting as intergranular cracks, void formation, and fracturing of the protective conductive carbon coating, which isolate active material regions and increase interfacial resistance (Hou et al., 2024).
The results demonstrate that LFP degradation is primarily influenced by surface-driven kinetic obstruction instead of bulk collapse, with interphase growth and microcracking impeding Li⁺ penetration into a generally stable framework. Comprehending the "bulk-stable, surface-limited" characteristic is essential for formulating recycling strategies aimed at surface reconstruction and Li⁺ replenishment.

Keywords: Lithium iron phosphate; Failure mechanisms; Li deficiency; Structural disorder; Microcracking

Speaker: Dr Tannaz Naseri (Institute of Circular Resource Engineering and Management, Hamburg University of Technology, Blohmst. 15, 21079, Hamburg, Germany)

76 Microstructure-Resolved SEI Modelling and Surrogate Learning in Lithium-Ion Batteries

Electrochemical energy storage systems are characterized by porous electrodes in which transport and electrochemical reactions are tightly coupled across multiple length and time scales. At the microscale, local heterogeneities in pore morphology and connectivity can induce strong spatial variations in transport pathways and interfacial conditions, ultimately affecting degradation mechanisms and long-term performance.

Among these processes, the formation and growth of the Solid Electrolyte Interphase (SEI) play a central role in capacity fade and impedance increase. While SEI evolution is commonly described using homogenized continuum models, it is inherently a heterogeneous phenomenon governed by pore-scale transport limitations and local electrochemical environments. Direct experimental access to these scales remains limited, motivating the development of geometry-resolved numerical approaches.

In this work, a time dependent pore scale modelling framework is proposed to investigate SEI evolution within a porous electrode. A four dimensional (three spatial dimensions plus time) finite element model implemented in COMSOL Multiphysics is used to simulate the initial charge and discharge cycles of a half-cell. By explicitly resolving the electrode microstructure, the model captures local variations in lithium concentration and electric potential arising from pore-scale transport constraints. These variations result in spatially non-uniform SEI growth, highlighting degradation features that cannot be captured by models based on averaged descriptors such as porosity and tortuosity.

The SEI is modelled as a thin interfacial layer whose growth is driven by a parasitic reaction competing with lithium intercalation. Its evolution is described using effective properties, enabling the analysis of the coupled effects of transport, interfacial kinetics, and microstructural features without explicitly resolving the detailed chemical composition of the layer.

To mitigate the high computational cost associated with pore scale simulations, a data driven surrogate modelling strategy is also explored. Convolutional Neural Networks (CNNs) are trained to predict electrochemical behaviour directly from three dimensional representations of porous electrode geometries, while recurrent neural network architectures are used to capture the temporal evolution of the electrochemical response. The training dataset is generated from physics-based simulations considering different active materials and particle packing configurations.

ACKNOWLEDGEMENT

This work was funded by the European Commission within the Horizon Europe research and innovation programme through the GA no 101137725 (BatCat).

Speaker: Elisa Buccafusco (Politecnico di Torino)

77 Performance prediction of Solid Oxide Cells (SOC) by ex-situ characterization of electrodes and physical modelling

Achieving the full potential of hydrogen energy requires the use of highly efficient devices for its production and consumption such as Solid Oxide Cells (SOCs). In-situ and ex-situ characterization techniques can be applied to differentiate effective designs from less efficient ones. In-situ methods assess cells during operation, while ex-situ techniques analyse individual components. Complementing these techniques, physical modelling aids in understanding cell phenomena and predicting Performance. However, models in the literature often require parameter tuning. The robustness of these models improves as more parameters are independently defined. Yet, destructive tests and advanced facilities can only determine some key morphological parameters. This study provides a methodology for performance prediction of SOCs using an ex-situ characterization. First, a comprehensive dataset of microstructures is generated by the Plurigaussian method, and their morphological parameters are evaluated. Next, a surrogate model is developed to estimate the triple phase boundary (TPB) density and phase-specific tortuosities (𝜏) using easily measurable parameters, namely phase volume fractions (𝜀) and mean pore/particle radius (𝑟𝑝). Finally, a physical model is employed to predict cell performance. Results indicate that the ion volume fraction significantly impacts the cell performance. Additionally, reducing particle sizes, especially electron-conductive particles, enhances cell performance by increasing TPB density. For manufacturers, optimizing electrode design with finer electron-conductive particles and composition of 60% ion and 20% electron volume fractions can notably improve SOC performance in both fuel cell and electrolyser operational modes.

Speaker: Mohammadhadi Mohammadi

78 Modeling Drying of a Colloidal Dispersion in a Fibrous Porous Medium Using Full Morphology Approach

Proton Exchange Membrane Fuel Cell is considered as an attractive pollutant-free alternative to thermal engines, especially for Heavy Duty applications. In this context, the study focuses on one major fuel cell components: the gas diffusion layer (GDL). The GDL is a thin porous medium, made of graphitized carbon fibers. To increase performance, a treatment is performed to render the GDL hydrophobic. It consists in dipping it in a polytetrafluoroethylene (PTFE) colloidal dispersion. Then, the medium is dried and sintered [1]. As it can be seen in the image (Fig.1), the PTFE after the treatment does not coat evenly all the fibers, and preferentially accumulates where the fibers are close to each other. As the PTFE distribution impacts the GDL properties [2], it is of interest to simulate the PTFE treatment step to predict its 3D distribution and the corresponding GDL single and two-phase transport properties. This will contribute to better predict the cell performance and improve treatment parameters to increase performance.
To this end, Daino et al. simulated the PTFE addition on 3D microstructures of GDL by using morphological closure [3], which is an image treatment that fills holes and small crevices in the image. Inoue et al. solved two-phase transport equations for PTFE particles and for dispersion saturation given by a continuous model [4].
Our work is based on a full morphology approach. Developed to simulate two-phase transport in a porous medium in the quasi-static limit, the full morphology approach is also image-based and consists in determining which parts of the media are accessible to a certain phase, by combining Laplace law and geometrical considerations. To do this type of simulation, Schulz et al. used morphological operations on images called dilation and erosion [5], while Sabharwal et al. developed a method based on the evaluation of pore size distribution [6].
To predict the PTFE distribution after drying, monitoring of PTFE concentration is performed in conjunction with the full morphology approach. In other words, drying simulation is performed via full morphology, while also computing the increase in PTFE concentration resulting from the solvent evaporation, until there is no solvent left. The computations are performed on 2D and on 3D GDL images obtained by x-ray tomography. Results of both full morphology algorithms mentioned in the previous paragraph are compared. The PTFE structures obtained are then compared to SEM images of the treated GDL. Also, through-plane distribution of PTFE in the material is compared to the experimental through-plane distribution obtained from EDX analysis.

Acknowledgement: This research is part of the project “DECODE" which has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement N° 101135537. More information on the project can be found at www.decode-energy.eu.

Speaker: Pierluigi Arnelli (Univ. Grenoble Alpes, CEA, Liten, DEHT)

79 A Structure-Transport-Driven Framework for Optimizing Laser-Engineered 3D Porous Electrodes

Recent studies on electrochemical energy storage devices, such as electrodes (anodes and cathodes) for Li-ion batteries and supercapacitors, have increasingly emphasized the critical role of the pore network [1, 2]. It is now well recognized that pore structure can either facilitate or hinder charge/ discharge or redox processes. In this context, the three-dimensional porous architecture of an electrode plays a decisive role in fast-charging mechanisms. This raises key questions: does pore architecture directly control fast charging, and if so, how can it be optimized? What structural “recipe” leads to high-performance electrodes?
In this work, we investigate a range of porous architectures and, by explicitly elucidating the role of tortuosity, propose a more informative and physically grounded framework for characterizing and optimizing porous electrodes. Various laser-based strategies reported in the literature have been used to create engineered porous geometries consisting of conical or cylindrical wells arranged in linear, rectangular, triangular, or grid-like patterns [3]. Such laser-engraved architectures have demonstrated promising improvements in the electrochemical performance of electrodes. In this work, we compare these well-defined patterns with an alternative laser-scanning strategy in which only the upper portion of the electrode (approximately half of its thickness) is continuously modified, while the bottom region remains intact. The resulting structures are computationally reconstructed and analyzed in terms of pore-network complexity, including tortuosity, connectivity, anisotropy, and the presence of isolated or dead-end regions that may impede ionic transport.
Three-dimensional transport simulations are performed within these topologies to evaluate ion accessibility and effective charge-storage utilization. The results reveal strong anisotropy between in-plane and through-plane transport, with tortuosity differing substantially between directions. Under such conditions, classical models based on effective medium theory, such as the Bruggeman relation fail to accurately describe transport behavior. This breakdown arises from the highly irregular pore geometries, including slit-like pores and strongly disordered networks, characteristic of the nano-carbon slurry–based electrodes investigated here. By solving diffusion transport equations within the actual reconstructed geometries, we demonstrate pronounced discrepancies between theoretical predictions and structure-resolved transport, particularly at length scales of a few nanometers.
We propose a hierarchical design methodology in which porous architectures are first characterized geometrically using available imaging or visualization techniques and subsequently optimized at the computational level before being selectively implemented experimentally [4]. Within this framework, a library of three-dimensional porous geometries is generated using computer-aided design and analyzed numerically to extract key structural descriptors, including tortuosity, connectivity, anisotropy, and the fraction of inactive or dead-end pore regions. These descriptors are correlated with simulated transport performance, enabling the identification of favorable architectural features. A classification algorithm is then used to associate optimized geometries with experimentally accessible fabrication parameters, thereby linking the numerical design space to practical preparation routes.
By restricting experimental efforts to a reduced subset of pre-optimized architectures, this strategy minimizes experimental cost and time and enables efficient iteration toward high-performance porous electrodes. The proposed workflow thus provides a general and scalable approach for rational pore-architecture optimization that moves beyond porosity-based design rules.

Speaker: Nadia Bali (FORTH/ICE-HT)

Poster: Poster VI

Plenary Lecture: Plenary 3

80 Reactive transport modeling of soil-based carbon removal: from reactive interfaces to objective limits

Achieving the temperature goals of the Paris Agreement will require 100 to 300 gigatons of carbon dioxide removal (CDR) this century. As large-scale interventions become central to climate planning, distinguishing between temporary carbon fluxes and durable atmospheric removals is essential. Yet the absence of robust and efficient monitoring, reporting and verification (MRV) frameworks remains a critical barrier for investment, policy progress and market development. Reactive transport models (RTMs) are often viewed as too complex, uncertain or immature to underpin MRV, despite their unique potential to enable uncertainty quantification, data assimilation and harmonization of discrepant fluxes. This tension highlights a broader challenge in carbon markets: how should scientific models be incentivized, governed and trusted as part of financial and regulatory infrastructure?

Using enhanced weathering (EW) as a case study, this lecture examines how mechanistic models can illuminate the coupled physical and chemical processes that govern CDR. MRV for EW requires translating mineral dissolution into durable atmospheric drawdown, as a function of coupled gas and aqueous transport, surface pH buffering, and dissolution-precipitation processes in variably saturated porous media and over scales spanning soils to estuaries. For the soil zone, new frameworks for surface proton buffering and the development of “reaction tags” identify mechanistic limits to verifiable carbon sequestration that arise from inefficiencies in alkalinity generation and export. Model-based analysis also establishes a physical basis for reconciling discrepancies between feedstock dissolution inferred from solid-phase measurements and the lack of measurable aqueous carbon export, a harmonization critical for robust MRV. Together, these examples illustrate both the diagnostic power of mechanistic modeling and the current limitations in parameterization, data integration, and multiphysics representations that constrain the readiness of models for decision support.

The talk concludes by expanding to other soil-based CDR pathways and raising emerging questions around model governance: What constitutes “fit-for-purpose” modeling in carbon markets, and how should model-based evidence be evaluated when used to substantiate claims of durable CO₂ removal?

Speaker: Katharine Maher

Friday 22 May

Invited Lecture: Invited VII

81 Designing the nanoremediation of contaminated aquifers: from laboratory tests to field implementation

Nanoremediation is a promising in-situ remediation strategy based on the subsurface injection of reactive suspensions of engineered nanoparticles (NPs), aimed at promoting the degradation, transformation, or immobilization of a broad range of groundwater contaminants. The success of field-scale applications depends on the ability to characterize and predict NP transport, retention, and reactivity in complex hydrogeological and geochemical conditions.
This talk presents an integrated methodology combining laboratory-scale testing and numerical modelling to support the design of nanoremediation interventions. Column transport experiments are performed using natural porous media and controlled flow conditions to evaluate key processes governing NP mobility, including deposition onto collector surfaces, detachment, aggregation, and clogging. These tests are designed to systematically explore the effects of ionic strength, pore-water velocity, and carrier fluid rheology. Experimental results are interpreted using the MNMs, a numerical model developed for one-dimensional simulation of colloid transport in saturated porous media, which enables inverse modelling of column tests to derive deposition kinetics and constitutive transport relationships. The resulting parameters are then used as input to MNM3D, a three-dimensional colloid transport model that simulates NP behaviour under realistic field-scale conditions, accounting for site heterogeneity, variable flow regimes, and evolving geochemical environments.
The modelling framework enables the simulation of alternative injection scenarios, supporting the optimization of operational parameters such as NP dosage, injection flow rate, duration, and spatial well configuration. It also provides insights into NP retention profiles and long-term fate under natural groundwater flow conditions.
The approach has been successfully applied in several field-scale studies with iron-based NPs, demonstrating its robustness as a quantitative, process-based tool for the design and performance assessment of permeation-based nanoremediation applications.

Speaker: Tiziana Tosco (Politecnico di Torino)

Invited Lecture: Invited VIII

82 Scaling microbial processes in porous media

Many porous media processes of interest involve microorganisms such as bacteria, fungi and viruses; examples include bioremediation, bioclogging, nutrient cycling, plant-microbe interactions, and critical mineral recovery. Consider the life of a bacterium in a porous medium. The size of its home is measured in micrometers – typical soil/sediment pores range in size from a few micrometers (e.g., shales or clays) to a few hundred micrometers (e.g. coarse sands). Like human homes, soil bacterial homes vary quite a lot in terms of who lives there (microbial community), how well they get along (competition or syntrophy), and what resources are available to the occupants (food, air, water). The microbially-mediated biogeochemical transformations that will occur, the types of microbes that will perform them, and the rates at which they occur, can dramatically differ between individual pores separated by very small differences. Importantly, microbes can actively respond to and modify their environment through regulation of their metabolism and other functions, so are often not well represented by standard chemical reaction models. On the other hand, the measurements we can make at field scales, and the models we use to represent field-scale biogeochemical transformations, are at the bulk scale. That is, we combine huge numbers of soil pores, grains, and microbes into a single sample (for measurement) or a single grid cell (in a numerical model) and we measure or simulate bulk properties (e.g., concentrations) and processes (e.g., reaction rates). But what a microorganism or microbial community actually senses and responds to is the environment in their individual pore home. Because natural porous media are highly heterogeneous, and the key reaction substrates (for example, oxygen, organic matter, nitrate, metals) are not uniformly distributed, the bulk characteristics are very different from the actual environment in any given individual pore. Furthermore, biogeochemical reaction processes are typically non-linear, so they don’t readily average up in the way we might expect. As a result, modeled reactions do not adequately represent the actual experiences and responses of microorganisms, creating a significant barrier to the application of biological advances to understanding and prediction of reactive transport in porous systems. This presentation will discuss these challenges in greater detail and present some novel approaches that may help us to address this scaling challenge based on emerging technologies and a creative combination of biological, physical, and computational sciences.

Speaker: Tim Scheibe (Pacific Northwest National Laboratory)

MS01: 4.1

MS06: 4.1

MS08: 4.1

MS09: 4.1

MS12: 4.1

MS15: 4.1

83 Microstructure/permeability relation of porous ceramics through active learning assisted experimental campaign

Understanding the saturated and unsaturated flow in porous media by producing ceramic porous model samples with controlled morphology. By controlling the morphology over a large range of microstructure, the study aims to isolate the parameters influencing resin impregnation and permanent flow in porous media. This subject has been treated by the community with many different approaches [1]. Unfortunately, existing models often fail to predict flow behavior correctly in cases where the porous medium is unsaturated, particularly during infusion. Compared to deformable fibrous media, porous ceramic model samples allow limiting and controlling the geometric variability of the porous network. Aiding in isolating the parameters influencing resin impregnation regimes in the material. This study has applications in the medical field (ceramics/polymers).
The medium-term objective is to develop models to better understand fluid flow in complex and controlled porous media [2]. To support this goal, a comprehensive experimental database is currently being built based on the study of porous ceramics manufactured with the sacrificial template method. First, an active-learning algorithm based on Gaussian Process Classification (GPC) has been developed to efficiently identify the parameters and boundaries of the chosen porous ceramic manufacturing process, with a minimal number of trial iterations. This approach is particularly advantageous for processes involving multiple parameters, where classical experimental designs would require extensive testing. We demonstrate the predictive capability of the algorithm for a test case involving two varying parameters: porogen volume and size.
Second, instrumented infusion tests are performed with an in-house set-up able to measure samples permeability from 10^(-16) to 10^(-12) m². Based on these measurements, a regression model is developed to predict permeability from the porogen characteristics (volume fraction of 2 classes of porogen). In parallel, the samples are characterized to quantify their internal structure (e.g., pore-size distribution) [3], enabling the quantitative assessment of how these parameters influence the fluid flow behavior.
Finally, dedicated descriptors are used to represent the 2D pore morphological features extracted from image-based characterization. These features are projected into a latent space using dimensionality-reduction techniques to obtain a compact representation of the pore morphology. Thus, regression is performed between reduced descriptors and permeability to establish a quantitative pore structure–property relationship. The study could bring insight into the relevant features of porous geometry that affect the permeability.
References
[1] D. Lee, M. Ruf, N. Karadimitriou, H. Steeb, M. Manousidaki, E.A. Varouchakis, S. Tzortzakis, A. Yiotis, Development of stochastically reconstructed 3D porous media micromodels using additive manufacturing: numerical and experimental validation, Sci. Rep. 14 (2024) 9375. https://doi.org/10.1038/s41598-024-60075-w.
[2] L. Xie, Q. You, E. Wang, T. Li, Y. Song, Quantitative characterization of pore size and structural features in ultra-low permeability reservoirs based on X-ray computed tomography, J. Pet. Sci. Eng. 208 (2022) 109733. https://doi.org/10.1016/j.petrol.2021.109733.
[3] S. Nickerson, Y. Shu, D. Zhong, C. Könke, A. Tandia, Permeability of porous ceramics by X-ray CT image analysis, Acta Mater. 172 (2019) 121–130. https://doi.org/10.1016/j.actamat.2019.04.053.

Speaker: Jnanesh Gopale Gowda

84 Comparison of CNN and GAN-Based Super-Resolution Methods for 3D Porous Microstructures

Across a wide range of energy and engineering applications, the performance of porous materials is strongly governed by their microstructure. In batteries, fuel cells, and hydrogen storage systems, microstructural features control key transport pathways and thus critically influence overall functionality. Accurate characterization therefore requires high-resolution (HR) three-dimensional (3D) microstructural data, since transport behavior depends heavily on fine-scale features. However, imaging methods such as focused ion beam–scanning electron microscopy (FIB-SEM) and X-ray computed tomography (CT) are costly and time-consuming, particularly at high spatial resolution.
To address these challenges, this work explores deep learning based super-resolution methods for generating HR 3D microstructures from low-resolution data. We study several super-resolution architectures, including CNN-based models (SRCNN, SRResNet, and U-Net) and a GAN-based approach (SRGAN). These 3D models take low-resolution inputs and reconstruct HR 3D microstructures. For comparison, we consider both geometric and transport properties: geometric fidelity is quantified using the Structural Similarity Index Measure (SSIM) and Peak Signal-to-Noise Ratio (PSNR), while physical fidelity is evaluated by computing effective tortuosity and permeability via FEM solutions of the Laplace and Stokes equations, directly linking reconstruction quality to material functionality.
Deep learning based SR outperforms nearest-neighbor, bilinear, and bicubic interpolation; among the tested models, SRResNet best matches the ground truth in both structural and transport properties. SRGAN further shows that perceptual sharpness alone does not guarantee functional accuracy. Overall, evaluation on lithium-ion battery cathode materials indicates that deep learning models, particularly SRResNet, best preserve the key properties required for reliable HR microstructure reconstruction.

Speaker: Rishabh Saxena (Helmut-Schmidt-Universität - Universität der Bundeswehr Hamburg)

85 Machine Learning for Tailoring Microstructural Properties

Inverse microstructure design is a persistent challenge in materials engineering because structure-property relations are high-dimensional, stochastic, and expensive to evaluate. As a result, conventional optimization and surrogate-driven workflows often become impractical when the design space is large, and microstructures must satisfy multiple constraints. Here we present PoreFlow, a data-driven framework for high-throughput generation of porous microstructures using continuous normalizing flows (CNFs). PoreFlow conditions the generative process on target properties through a latent representation, enabling efficient sampling of microstructures that meet specified objectives while retaining a continuous, invertible mapping between latent variables and generated structures.

We validate PoreFlow on 3D porous media generation. The framework achieves coefficients of determination above 0.915 for reconstruction and above 0.92 when generating previously unseen samples that satisfy the prescribed targets. In contrast to GAN-based approaches that can suffer from training instability and mode collapse, the flow-based formulation provides stable likelihood-based training and supports more transparent analysis of the latent space. The architecture is modular, allowing the autoencoder component to be replaced to accommodate alternative microstructure parameterizations beyond voxelized images.

PoreFlow provides a scalable pathway for inverse design of porous materials with applications in energy storage, catalysis, and related transport-dominated systems, enabling faster and more reliable exploration of structure space under property constraints.

Speaker: Dr Serveh Kamrava (Colorado School of Mines)

86 Data-Driven Prediction of Relative Permeability: Applications to CO₂ and Hydrogen Storage

Relative permeability curves are one of the fundamental parameters in multiphase flow modelling, supporting applications that now extend into Carbon Capture and Storage and Underground Hydrogen Storage. These curves are traditionally obtained experimentally using sophisticated special core analysis instruments, resulting in a workflow that relies on a limited number of core plugs that cannot fully capture reservoir heterogeneity. As interest in subsurface storage increases, there is a clear shift towards data-driven approaches that can connect sparse, complex measurements with the continuous property fields required by reservoir simulators. Accordingly, this study aims to apply machine learning techniques to predict relative permeability curves for water (krw) and gas (krg) in sandstone cores during drainage experiments.

The work described here is built on a moderate-sized dataset of approximately fifteen hundred data points, each characterised by a set of features that includes temperature, pressure, porosity, absolute permeability, and key fluid properties such as gas and brine viscosities and their ratio. Along with normalised water saturation and irreducible water saturation, these variables offer a realistic testbed for modern data-driven petrophysical modelling in systems relevant to gas and brine. The complete analysis will include the modelling workflow, explore how predictions respond to other key inputs such as Interfacial Tension and wettability, and provide an initial investigation into how this framework can be extended to incorporate detailed rock and fluid characteristics and broader gas–brine systems, thereby enhancing the transferability and efficiency of relative permeability modelling for subsurface storage applications.

Across the literature, there is a trade-off between model flexibility and physical consistency. Conventional regressions and unconstrained neural networks fit the data but often violate key constraints, especially the fact that relative permeability lies between 0 and 1. Deep networks tend to overfit and break monotonic saturation trends, while tree ensembles like XGBoost and kernel methods like Gaussian Process Regression perform well, with GPR quantifying uncertainty. Building on these insights, we trained monotonic XGBoost models on CO₂–brine drainage experiments in sandstone, using the above features to predict four quantities at each point, namely irreducible water saturation, gas relative permeability at irreducible water saturation, and the normalised water and gas relative permeabilities.

Finally, the model is evaluated on a held-out test set that covers the full range of experimental conditions in temperature, pressure, permeability, and viscosity ratio, providing a direct assessment of its ability to interpolate within realistic conditions. Initial results for a CO2-brine system indicate that the monotonic XGBoost surrogate accurately reproduces the normalised water relative permeability, achieving an R² of 0.9829 and a mean squared error (MSE) of 0.001725, corresponding to a root mean squared error (RMSE) of approximately 0.0415 on a held-out test set. For the gas phase, the model achieves an R² of 0.9747, an MSE of 0.002670, and an RMSE of 0.0517. The close agreement with SCAL measurements (Figure 1) indicates that this method can serve as a reliable predictive tool when laboratory data are sparse or unavailable, therefore helping to reduce experimental workload and costs while still providing simulation-ready kr curves.

Speaker: Mr Abdolali Mosallanezhad (PhD Student, Research Centre for Carbon Solutions (RCCS), School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh, UK)

87 DimExDAM: A Diffusion–Adversarial Framework for 2D-to-3D Generation of Complex Porous Microstructures

Accurate three-dimensional (3D) representations of porous microstructures are essential for predicting transport, mechanical, and reactive behavior in natural and engineered porous media. However, acquiring 3D datasets remains costly, technically demanding, and often infeasible for fragile or fine-grained materials such as clay-based systems. Recent deep generative approaches attempt to infer 3D structures from two-dimensional (2D) images, yet existing methods face important limitations. Classical reconstruction algorithms rely on low-order statistics and struggle with heterogeneous media, while Generative Adversarial Network (GAN)-based models, such as SliceGAN, exhibit unstable training and difficulties reproducing complex multi-phase textures. Diffusion models, although promising, typically require full 3D training data or incur high computational cost.
This work introduces Dimensionality Expansion Diffusion Adversarial Model (DimExDAM), a hybrid generative framework designed specifically for 2D-to-3D microstructure generation using minimal training data. The approach integrates a 3D diffusion-based generator with a single 2D adversarial discriminator. Instead of using a conventional denoising loss, the method employs an adversarial objective computed on orthogonal slices, allowing the model to learn structural consistency without access to 3D ground truth. This formulation stabilizes training, mitigates vanishing-gradient issues common in multi-critic GAN architectures, and reduces sampling redundancy typically observed in diffusion-based reconstruction.
We evaluate DimExDAM on porous materials with increasing structural complexity, including clay, carbonate, and sandstone datasets. Generated volumes are assessed using phase fraction agreement, directional connectivity measures, and structural descriptors relevant to porous media characterization. The model demonstrates: (i) consistent recovery of anisotropic features, (ii) minimal slice artefacts compared with SliceGAN, and (iii) strong statistical alignment with reference descriptors while requiring as little as one 2D training image per orientation. Training exhibits smoother convergence behavior than traditional GAN approaches and avoids the heavy dependence on full 3D volumes inherent to other diffusion frameworks.
The results indicate that DimExDAM provides a robust pathway toward data-efficient 3D reconstruction of complex porous microstructures, enabling realistic synthetic datasets for simulation. Ongoing work explores conditioning strategies and physics-informed priors to further integrate transport-relevant constraints into the generative process.

Speaker: Ali Aouf

MS16: 4.1

MS20: 4.1

Poster: Poster VII

MS01: 4.2

MS02: 4.2

MS03: 4.2

MS05: 4.2

MS07: 4.2

MS09: 4.2

MS17: 4.2

88 Mass transport characterization in nanoporous polymer electrolyte membranes used in electrochemical systems.

Understanding and improving mass and ionic transport mechanisms within the nano-porous membrane used in polymer electrolyte membrane (PEM) water splitting electrolyzers is vital for achieving improved efficiencies that would enable the use of water electrolysis in sustainable energy infrastructures. To achieve this goal, microfluidics electrolyzers can serve as flexible platforms for operando PEM characterization. For example, Krause et al. [1,2] developed a microfluidic PEM electrolyzer with a Nafion membrane capped on top of the channels to probe operando the water content in PEM. The measurements of the PEM water content can then be carried out using imaging methods such as the IR transmittance.

This work aims to improve characterization methods for measuring PEM hydration to get a better understanding of the transport mechanisms in those nano-porous material used in electrochemical applications. An operating microfluidic PEM electrolysis chip is used for operando infrared (IR) spectroscopy [3]. The IR imaging is coupled with electrochemical impedance spectroscopy (EIS) and distribution of relaxation times (DRT) to elucidate the relationship between membrane hydration and ohmic, kinetic, and mass transport losses. IR imaging unveils water diffusion gradients across the PEM of the microfluidic water electrolyzer. Varied H2SO4 anolyte concentrations directly correlated with water diffusion through the PEM, where the highest anolyte concentrations accompanied the strongest water diffusion gradients. We show that tuning the anolyte concentration for visualizing water diffusion across the PEM came with a tradeoff, as the electrochemical performance of the electrolyzer became increasingly unstable. These findings showcase the potential of IR imaging when coupled with a microfluidic PEM electrolyzer and electrochemical characterization techniques, and the influence of anolyte concentration for manipulating the PEM water gradient .

References
[1] K. Krause, M. Garcia, D. Michau, G. Clisson, B. Billinghurst, J. Battaglia, S. Chevalier, Probing membrane hydration in microfluidic polymer electrolyte membrane electrolyzers via operando synchrotron Fourier-transform infrared spectroscopy, Lab Chip. 23 (2023) 4002–4009.
[2] K. Krause, A. Crête-Laurence, D. Michau, G. Clisson, J.-L. Battaglia, S. Chevalier, Water gradient manipulation through the polymer electrolyte membrane of an operating microfluidic water electrolyzer, J Power Sources. 623 (2024) 235297.
[3] S. Chevalier, J.-N. Tourvieille, A. Sommier, C. Pradère, Infrared thermospectroscopic imaging of heat and mass transfers in laminar microfluidic reactive flows, Chemical Engineering Journal Advances. 8 (2021) 100166.

Speaker: Stéphane Chevalier (ENSAM)

89 Coated Metallic Foams as Versatile Porous Substrate: From Hydrogen-Electrolysis Electrodes to Photocatalytic Water Treatment

Open-cell metallic foams offer a combination of high permeability and a large accessible surface area, as well as good thermal and electrical conductivity. This makes them a versatile substrate for functional porous-media devices. Their three-dimensional strut network enables efficient heat and mass transport at low pressure drop. However, practical performance depends heavily on how surface functionality is introduced without compromising pore accessibility.
This contribution discusses the use of coated metallic foams as general materials and design concept for engineered porous media. We highlight the main challenges associated with coating open-cell foams, such as generating a uniform, conformal layer across the three-dimensional strut network and achieving sufficient adhesion and long-term stability during operation. Another challenge is maintaining the foam’s effective porosity to ensure that its transport benefits are not lost.
To illustrate these ideas, we refer to two ongoing thematics in our group that employ coated foams in different operating regimes. The first theme focuses on electrochemically functionalised foam electrodes for hydrogen electrolysis, demonstrating how conductive porous architectures can be transformed into highly active interfaces without sacrificing favourable mass transport. The second thematic examines photoactive coatings on foams for use in flow-through water treatment, showing how light-responsive surface functionality can be integrated into a permeable backbone. Together, these examples offer a practical basis for discussing transferable coating strategies across electrochemical and photocatalytic porous media technologies.
Overall, coated metallic foams emerge as a robust, porous platform for technologies relevant to energy and the environment, as well as being a useful case study of how the interplay between microstructural design, transport and surface reactivity determines the macroscopic performance of porous materials.

Speaker: Felix Neupert (Fraunhofer IFAM Dresden)

90 A Novel Approach to Fabricating 3D PAN based Carbon Electrode Architectures

The production of renewable energy is gradually increasing as part of the global efforts to mitigate the global warming. However, the inherent intermittency of renewable energy sources creates a growing need for reliable large-scale energy storage devices. Flow batteries (FBs) are considered a promising candidate for large scale stationary energy storage, but their energy efficiency is limited by various losses, like mass transport, kinetic, ohmic, and pressure losses all of which are strongly influenced by the electrode material and porous structure [1], [2]. Carbon electrodes are currently the most promising electrode materials for FBs due to their chemical stability, high surface area, good electrical conductivity and their ability to suppress parasitic reactions such as the hydrogen evolution reaction [3]. Nevertheless, the range of available porous electrode designs remains narrow, with most studies relying on traditional architectures such as carbon felts, papers, and cloths.
In this work, we developed a novel method to fabricate porous 3D-structured carbon electrodes, using PAN as the primary precursor material, typically used in commercial felt, paper and cloth based electrodes. This new fabrication approach enables the creation of custom-designed 3D architectures, allowing precise control over the electrode structure. As a result, new possibilities emerge for enhancing mass-transfer characteristics and reducing pressure drop, ultimately lowering the pumping energy required during operation [4], [5]. This work focusses on the production method of these 3D structured carbon electrodes. For the fabrication, we combined the non-solvent induced phase separation (NIPS) method with dissolvable mold materials to create 3D PAN structures. These structures are subsequently carbonized under inert conditions, yielding 3D structured carbon electrodes. The effect of different mold designs, PAN:PVP:DMF ratio’s and oxidation temperatures were studied electrochemically and physically. The electrochemical testing was performed in a custom made flow battery system and compared with traditional felt electrodes. Being able to cycle the flow battery equipped with these 3D structured electrodes in the range of 100 mA/cm² without optimizing the structure, shows the promising nature of this method. Moreover, this method makes it possible to design various carbon electrodes, like Triple periodic minimal surface (TPMS) structures, static mixer and carbon meshes.

Speaker: Frederik Vandenbulcke

91 Towards "Lung-Inspired" Porosity

Fuel cells and electrolyzers represent promising sustainable energy storage technologies that harness redox reactions at electrodes. However, their efficiency is compromised by challenges including low power density and expensive catalysts.

Porous electrode architectures offer a potential solution to overcome mass transfer limitations, with hierarchical structures being on the spotlight for many researchers. However, the variety in design and complexity of these structures poses a challenge, as it leads to trial and error strategy in the implementation. In our approach, we decided to enhance the hierarchical structure by using bio-mimic inspired "lung design" in the catalyst layer. N-doped carbon electrodes are synthesized via a hard-template method with acrylonitrile precursor and ZnO "Lung-Inspired" templates.

By synthesizing ZnO templates through two-step hydrothermal methods, we were able to achieve a structure imitating lungs in its hierarchy. Materials were characterized using scanning electron microscopy, nitrogen adsorption isotherms, and Raman spectroscopy. We compare between carbons with different porous archetictures (spheres, rods, and "lungs") through electrochemical activity in oxygen reduction reaction, using cyclic voltammetry, ring-rotating disk electrode (RRDE) analysis, and fuel cell testings.

Our results reveal the enhancement of mass transfer limit through porosity, adding fundamental insights into the rational design of hierarchy in carbons. These findings contribute to addressing a critical limitation in fuel cell and electrolyzer technologies, advancing their practical implementation for sustainable energy conversion and storage.

Speaker: Mr Nicola Seraphim (Technion)

MS20: 4.2

Plenary Lecture: Special Session on Green Housing

MS01: 4.3

MS03: 4.3

MS05: 4.3

MS08: 4.3

MS09: 4.3

MS15: 4.3

92 ML-Assisted Topology Optimization of Thermochemical Heat Storage Reactors

Thermochemical energy storage (TCES), where thermal energy is stored in a reversible chemical reaction in a porous powder bed, is a promising technology for large-scale and long-term thermal energy storage. Extensive research has been conducted on the subject for potential applications, including the capture of excess heat from industrial processes and the storage of energy in concentrated solar power plants. This study investigates TCES in the SrBr2 system, which offers a high energy capacity and near-perfect reversibility for medium temperature applications.
However, the scaling up of these reactors is hindered by the limited heat transfer from the heat source, such as reactor walls, to the powder bed. To address this challenge, heat conducting structures, such as fins, are incorporated into the bed to enhance thermal contact and shorten transport paths. Moreover, the powder agglomerates to a porous solid medium which expands and contracts during water uptake and release, respectively. This deformation of the bed may result in its detachment from the heat conducting surfaces, as illustrated in Figure 1, further inhibiting heat transport.
In a previous presentation [1], we presented the use of machine learning techniques to enhance heat transfer within the reactor with a non-deforming bed, which is achieved through the design of optimized heat-conducting structures. Due to the prohibitive time requirements of direct simulations, an artificial neural network surrogate model was constructed. The method entails the training of a neural network utilizing simulated data, which was generated with randomly generated fin structures. Subsequently, the trained network is used to predict the progression of the reaction over time. In this presentation, we will present the most recent findings on the use of neural networks for surrogate modeling, employing architectures based on the SinGAN [2]. Furthermore, the methodology for extending the surrogate model by a mechanical model for the deformation of the porous powder bed will be demonstrated. This enables the estimation of the powder bed/wall detachment, the resultant transport resistance, and the consequent impediment to reactor performance (see Fig. 2).
However, the primary emphasis of the presentation will be on topology optimization. The presentation will show the methodology employed to couple the surrogate model with topology optimization algorithms, which are based on the brute force, level-set (for an illustration see Fig. 3), and stochastic optimization methods [3]. These methods are employed to calculate optimal geometries for heat-conducting structures minimizing an objective function, which encodes the desired reactor performance characteristics. Finally, we will demonstrate how different objective functions give rise to different optimal geometries.

Speaker: Dr Torben Prill (German Aerospace Center (DLR))

93 Fast-to-Long Acquisition Projection Learning for Denoising X-ray Microtomography

The need to increase experimental throughput and support time-resolved imaging of dynamic laboratory experiments motivates the reduction of acquisition time in X-ray microtomography. However, faster acquisitions inevitably lead to lower signal-to-noise ratios, since fewer photons contribute to each projection, resulting in reconstructions with increased noise levels and degraded structural definition. This limitation can be mitigated using deep learning methods trained on paired acquisitions of the same sample obtained under different exposure times.

In the acquisition protocol adopted here, scans of 2 minutes and 35 seconds (fast) and 60 minutes (long) were performed sequentially on each rock plug without removing or repositioning the sample in the scanner, ensuring spatial alignment between acquisitions. In this setting, the fast scan provides a noisy representation of the sample, while the long scan serves as the target image, forming well-defined input–target pairs for supervised learning in which fast acquisitions encode acquisition-related noise and artifacts and long acquisitions define the desired reconstruction quality. Microtomography data for both exposure times were acquired using a VTomex M system (Baker Hughes). The fast acquisition employed timing = 50, average = 1, and skip = 0, whereas the long reference acquisition applied timing = 100, average = 40, and skip = 1. In both time configurations, the number of two-dimensional projections was kept constant to enable paired datasets. Because the number of projections is a key factor for reconstructed volume quality, higher values are desirable. To achieve a fixed total of 801 projections under the fast setting, the acquisition parameters were adjusted. The fast acquisition employed timing = 50, average = 1, and skip = 0, whereas the long reference acquisition applied timing = 100, average = 40, and skip = 1. The X-ray source operated with energies between 140–150 keV and tube currents in the range of 220–250 μA.

Based on these paired datasets, a supervised machine learning approach was applied to a set of 12 Brazilian carbonate plug samples. The model was trained to map the two-dimensional projections from the fast acquisition to the corresponding projections from the long acquisition. Operating directly in the projection domain is advantageous since it avoid compounding artifacts introduced in the reconstruction step, as our goal is to reduce acquisition noise. To assess generalization, a leave-one-out validation strategy was adopted. In each iteration, projections slices from 11 samples were used for training, while no slice from the remaining sample was included in the training set. The held-out sample, unseen during training, was reserved exclusively for evaluation. This process was repeated until all samples had been used once as test cases.

Model performance was evaluated on the reconstructed volume obtained from the network-generated projections. Reconstruction used the same acquisition parameters as the fast scans. The leave-one-out validation strategy captured variability across samples, reflecting the heterogeneity of carbonate rocks. Quantitative signal-to-noise metrics showed consistent improvements over fast acquisitions, with reconstructed volumes closer to the long-exposure reference and exhibiting a more concentrated grayscale range, although some smoothing and blurring were observed.

Speaker: Luan Vieira (Universidade Federal do Rio de Janeiro)

94 Machine Learning Applications for Predicting Drilling Mud Loss in Fractured Formations

This paper investigates the application of machine learning (ML) to model and predict drilling mud loss in subsurface formations with conductive natural fractures. Mud loss during drilling is a complex and costly issue that disrupts operations and increases non-productive time. The goal of this study is to develop an ML-based tool that leverages type-curves and physics-informed models to predict key parameters such as cumulative mud loss volume, maximum mud loss duration, and equivalent hydraulic fracture aperture.

The study integrates a physics-informed approach to model mud loss using the Herschel-Bulkley fluid model, which accounts for non-Newtonian fluid behavior. A Latin Hypercube Sampling (LHC) method systematically varies uncertain parameters, such as yield stress, consistency factor, and hydraulic fracture aperture, to generate a robust training dataset. We introduce a novel concept of terminal mud loss volume (TMLV) and terminal mud loss time (TMLT) to measure and predict mud loss dynamics. An artificial neural network (ANN) is employed to predict mud loss behavior, using cumulative mud loss data as input. The model was trained and validated using both synthetic and field data to ensure accuracy and adaptability. Early mud loss trends are incorporated to improve predictions and refine estimates of fracture conductivity.

The developed ML-based model demonstrated high accuracy in predicting cumulative mud loss, maximum loss duration, and equivalent hydraulic fracture aperture under a range of conditions. It effectively captured the complex, nonlinear relationships governing mud loss behavior in fractured formations. The ANN model successfully integrated physics-informed equations, yielding predictions that are closely aligned with field observations. This streamlined approach reduces computational demands while maintaining reliability, offering practical solutions for real-time decision-making in lost circulation scenarios.

This study introduces a novel machine-learning framework for modeling and mitigating mud loss in naturally fractured formations. By combining physics-based models with ML techniques, the proposed tool enhances the predictive capabilities of traditional methods and provides actionable insights for managing lost circulation. The approach is adaptable to diverse scenarios, making it an accurate and efficient solution for addressing one of the oil and gas industry’s most persistent challenges.

Speaker: Dr Xupeng He (Saudi Aramco)

95 Physics informed neural network for modeling seawater intrusion in coastal aquifers

Physics-informed neural networks (PINNs) is a new approach designed to reduce the dependence of neural network models on data. This technique shows strong potential for groundwater applications, where data are often scarce. PINNs can be used for forward modeling, surrogate modeling, uncertainty quantification, and inverse modeling. For this reason, the groundwater-related applications of PINNs are significantly growing. However, PINNs remain a recent technique, and their transition into operational tools for groundwater management requires substantial effort, particularly to address associated challenges and to adapt the approach to diverse groundwater problems. To the best of our knowledge, PINNs have not yet been applied to seawater intrusion (SWI) in coastal aquifers. This represents a challenging application due to the presence of coupled, nonlinear, multi-physical processes. This study aims to fill this gap by applying PINNs to the well-known SWI benchmark, the Henry problem. Using this benchmark provides insights into the applicability of PINNs for SWI and demonstrates how PINNs can enhance the reliability of neural network models for simulating SWI under limited data conditions.

Speaker: Mrs Maryam Mansouri Bajgiran

96 ML-assisted design of porous monolithic reactors using pore-resolved CFD surrogate models

Porous monolith reactors are attractive supports for heterogeneous catalysis due to their high surface-to-volume ratios and intricate pore networks, which promote efficient contact between reactants and catalyst. However, performance indicators such as productivity, pressure drop, selectivity, and operational safety depend strongly on the underlying pore geometry and are often tightly coupled. These competing objectives, together with the complex mechanisms by which geometry influences transport and reaction phenomena, make the physics-based design of porous monolithic reactors challenging.

One major barrier to the broader adoption of engineered porous structures for heterogeneous catalysis is the efficient identification of reaction-dependent optimal geometries. On the one hand, a wide range of geometry generation methods enable exploration of a high-dimensional design space, including triply periodic minimal surface (TPMS) formulations, stochastic methods, and data-driven generators. On the other hand, advances in fabrication techniques such as additive manufacturing and high-precision etching increasingly enable the fabrication of complex monolith geometries, making performance evaluation prior to manufacture a critical bottleneck. High-fidelity approaches such as pore-resolved computational fluid dynamics (PRCFD) provide detailed access to flow, transport, and reaction mechanisms at the pore scale, but remain prohibitively expensive for use in iterative geometry optimisation or large-scale design space exploration.

In this work, we leverage machine learning (ML) models, specifically convolutional neural networks (CNNs) and multiscale extensions inspired by the MSNet architecture [1,2], to construct surrogate models capable of predicting pressure, velocity, and concentration fields in porous monoliths. By predicting physically meaningful fields rather than only scalar performance values, the surrogate models retain sensitivity to geometry-induced transport mechanisms that are known to play a central role in porous media. The models are trained on PRCFD-generated datasets and used to rapidly evaluate candidate geometries. We investigate trends associated with dataset size and geometric variability, and their influence on surrogate accuracy. We further examine how these factors affect the Pareto-optimal solutions obtained when the models are embedded within a surrogate-assisted geometry optimisation framework.

To generate the high-fidelity datasets required for surrogate training, we rely on PRCFD simulations performed using Lethe [3]. Lethe is an open-source multiphysics PRCFD software based on the finite-element library deal.II [4], which solves the Navier–Stokes, mass conservation, and advection-diffusion-reaction equations in complex porous geometries using a sharp immersed-boundary method [5]. This approach eliminates the need for body-fitted meshing and enables efficient simulation of digitally generated structures across a range of geometries and operating conditions.

Rather than focusing on finalized optimal designs, we provide a proof of concept and a modular, extensible framework for ML-assisted, multi-objective geometry optimisation of monolithic reactors. The proposed approach explicitly targets trade-offs between conflicting objectives such as pressure drop and conversion, and illustrates how PRCFD-trained ML surrogates can shift pore-scale simulation from a purely diagnostic role toward a design-oriented tool in porous media. While additive manufacturing is not addressed directly, the workflow is compatible with AM-ready geometry generators and provides a pathway for translating pore-scale physics into application-specific reactor design strategies.

Speaker: Olivier Guévremont (Polytechnique Montréal)

MS16: 4.3

MS20: 4.3

Poster: Poster VIII

Plenary Lecture: Plenary 4

97 Multi-physical transport in porous media for energy applications

Meso-structured, porous materials exhibit favorable charge, heat, and mass transport properties and are used as absorbers, heat exchangers, insulators, reaction sites, electrodes and/or reactants in a wide variety of applications ranging from chemical processing, (photo)electrochemistry, combustion, filtering, to concentrated solar reactor technology. The transport properties of these materials largely depend on the meso-structure of the material and significantly affect its combined transport and ultimately the performance of the device. For example, electrochemical reactors for CO2 reduction show significant variation in activity and selectrivity dependent on the (anistropic) mesostructure of the gas diffusion electrode or porous thermal storage devices made of phase change material show significant variation in capacity and discharge time dependent on the mesostructure. In-depth understanding of the structure-property relation followed by pore-engineering of the materials used in the applications is therefore of fundamental importance to further improvements in performance. I will discuss decoupled and coupled pore-level numerical approaches for transport characterization and estimation of the local heterogeneity, discuss the use of neural networks for rapid performance assessment and optimization, and inverse experimental-numerical approaches for the characterization of the transport in porous media in extreme conditions.

Speaker: Sophia Haussener