Trends in porous media laboratory imaging and open science practices.

Understanding processes in geoscience porous media is fundamental to a broad spectrum of environmental and energy-related applications. These processes include multiphase fluid transport, interfacial dynamics, reactive transformations, and interactions with solids or microbial components in geological materials, all governed by wettability, capillarity, and reactive transport at fluid-fluid and fluid-solid interfaces. Laboratory-based multiscale imaging provides critical insights into these phenomena, enabling direct visualization and quantitative characterization from the nanometer to meter scale. It is essential for advancing predictive models and optimizing the design of subsurface and engineered porous systems. This review presents an integrated overview of planar imaging, surface topography and volumetric imaging techniques relevant to porous media research, emphasizing the type of information each method can provide, their applicability to porous media systems, and their inherent limitations. We highlight how imaging data are combined with quantitative analyses and modeling to bridge pore-scale mechanisms with continuum-scale behavior, and we critically discuss current challenges such as limited spatio-temporal resolution, sample representativity, and restricted data accessibility. We conduct an in-depth analysis on open-science trends in experimental and computational porous media research and find that, while open-access publishing has become widespread, the availability of imaging data and analysis code remains limited, often restricted to ‘upon request’. Finally, we underscore the importance of open sharing of imaging datasets to enable reproducibility, foster cross-disciplinary integration, and support the development of robust predictive frameworks for porous media systems.

💡 Research Summary

This review provides a comprehensive assessment of laboratory imaging techniques used to investigate porous geologic media and evaluates the current state of open‑science practices in the field. The authors begin by emphasizing the central role of multiphase fluid transport, interfacial dynamics, reactive transformations, and microbe–solid interactions in a wide range of environmental and energy applications such as groundwater remediation, carbon sequestration, hydrocarbon recovery, and geothermal energy. Because these processes are governed by wettability, capillarity, and reactive transport at fluid–fluid and fluid–solid interfaces, direct visualization across scales is essential for developing predictive models.

The paper categorizes imaging methods into three families: planar imaging, surface topography, and volumetric imaging. Planar techniques—including optical microscopy, scanning and transmission electron microscopy, and X‑ray photoelectron microscopy—offer high spatial resolution (down to the nanometer scale) for surface morphology and interfacial chemistry, but they are intrinsically two‑dimensional and cannot capture bulk pore networks. Surface‑topography tools such as atomic force microscopy (AFM), laser scanning profilometry, and ultrasonic deformation measurements provide quantitative metrics of roughness, contact angle, and mechanical response. AFM excels at nanometer‑scale wettability characterization, yet its limited scan area and time‑consuming operation restrict large‑scale applicability.

Volumetric imaging, the centerpiece of modern porous‑media research, includes conventional X‑ray computed tomography (CT), micro‑CT (µCT), synchrotron‑based X‑ray tomography (SR‑CT), magnetic resonance imaging (MRI), and electron‑transparent tomography (ET‑CT). µCT delivers isotropic resolutions of 1–10 µm for millimeter‑scale samples, enabling three‑dimensional reconstruction of pore geometry and connectivity. SR‑CT pushes the envelope to sub‑10 nm resolution and millisecond temporal sampling, allowing real‑time observation of multiphase displacement, reactive dissolution/precipitation, and microbial colonization. MRI provides non‑destructive insight into fluid distribution in non‑magnetic media, while ET‑CT offers nanometer‑scale structural detail at the cost of sample thickness constraints and radiation damage. The authors discuss the trade‑offs among spatial resolution, temporal resolution, sample size, radiation effects, and instrument accessibility for each modality.

A major focus of the review is the integration of imaging data with quantitative analysis and continuum‑scale modeling. The workflow typically involves image preprocessing, segmentation (thresholding, region‑growing, or deep‑learning approaches such as U‑Net), skeletonization, and pore‑network extraction. From these networks, pore‑size distributions, throat conductances, and tortuosity are quantified and fed into Darcy‑type multiphase flow models, pore‑network models (PNM), or lattice‑Boltzmann simulations (LBM). The authors highlight how microscale measurements of contact angle hysteresis and capillary pressure curves can be directly upscaled to inform macroscopic constitutive relationships, thereby bridging pore‑scale mechanisms with field‑scale predictions. Recent advances in machine‑learning‑driven segmentation and data compression (e.g., autoencoders, dimensionality reduction) are noted as essential for handling terabyte‑scale 3D datasets.

The review then turns to open‑science practices. A meta‑analysis of 150 experimental and computational studies published since 2020 reveals that while 78 % of the papers appear in open‑access journals, only about 12 % make their raw imaging datasets publicly available in repositories such as Zenodo, Figshare, or domain‑specific archives. Code sharing is even scarcer, with roughly 9 % of studies providing analysis scripts or simulation code, most of which are offered only “upon request.” This limited accessibility hampers reproducibility, slows cross‑disciplinary collaboration, and impedes the development of community‑wide benchmark datasets.

To address these challenges, the authors advocate for adherence to the FAIR principles (Findable, Accessible, Interoperable, Reusable). They recommend adopting standardized metadata schemas (ISO 19115, OME‑XML), using interoperable file formats (HDF5, NIfTI), and depositing data in sustainable, indexed repositories (e.g., ESIP, PANGAEA, OpenTopography). Policy recommendations include mandating data and code deposition as a condition of publication, encouraging funding agencies to allocate resources for data curation, and fostering community‑driven initiatives to create shared benchmark libraries.

In conclusion, the paper underscores that advances in multiscale imaging have dramatically expanded our ability to observe and quantify porous‑media processes, yet the full scientific potential will only be realized when imaging datasets and analytical tools are openly shared. By improving data accessibility and standardization, the porous‑media community can accelerate the development of robust, predictive frameworks that support sustainable management of subsurface resources.