REVIEW 3 major objections 7 minor 17 references
Spatially Distributed Wettability Characterization in Porous Media
T0 review · 3 major / 7 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read Automated pore-by-pore contact-angle measurement reveals that a mixed-wet rock with a mean angle of 64.7 degrees contains 40.4 percent intermediate-wet pore space, arguing that averaged wettability metrics mislead pore-scale flow modeling.
desk verdict A well-validated contact-angle measurement algorithm is the real contribution; the headline heterogeneity claim is an interpolation output that needs sensitivity analysis. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing mechanism is an extrapolation-based normal-vector reconstruction around three-phase contact lines. Instead of reading the fluid–solid and fluid–fluid normals directly at the contact loop—where segmentation artifacts and partial-volume effects are worst—the algorithm extracts triangular meshes with marching cubes, smooths them with a volume-preserving two-step smoothing scheme, fits the contact loops with B-splines, and then extrapolates each local normal from nearby mesh faces within a slab of thickness ±0.5 voxels and a radius of 12 voxels, using second-order polynomial regression with a LOWESS fallback and a 30-degree quality-control cap. Contact angles follow from the arccosine of the dot product of the two extrapolated normals. To turn the sparse loop measurements into a volume field, the algorithm assigns 30 degrees to uninvaded defending-fluid regions, interpolates invading-fluid objects by inverse-distance weighting, and interpolates the remaining defending fluid only within 20 voxels of any measurement.
What would settle it
Recompute the mixed-wet spatial map with the uninvaded-pore default changed from 30 to, say, 60 degrees and see whether the sample still shows roughly 40% intermediate-wet pore space; or image the same Bentheimer sample at higher resolution and measure contact angles directly inside the presumed-uninvaded regions to check the 30-degree assignment.
Extended reading notes
Core claim
The central claim is that a spatial, pore-by-pore wettability map, rather than a mean contact angle, is the correct characterization of mixed-wet porous media, and that such a map can be obtained from segmented micro-CT images with an automated geometric algorithm. On the mixed-wet Bentheimer sample, the algorithm yields a mean angle of 64.7° ± 19.8°, which by itself suggests weak water-wetness; but the spatially distributed map shows 59.6% of the pore surface in the water-wet range (1–69°) and 40.4% in the intermediate-wet range (70–110°), with some pores at 100°. The paper argues that this hidden heterogeneity explains the experimentally observed nearly minimal fluid–fluid interfaces and the coexistence of different pore-filling mechanisms in the same sample, and that the same reasoning applies to any capillary-dominated system where wettability alteration is uneven.
Load-bearing premise
The map is not directly measured everywhere: all pores that the invading fluid never entered are assigned a contact angle of 30 degrees by assumption, and unmapped voxels are filled in by distance-weighted interpolation from a sparse set of measured contact lines.
Editorial extensions
If this is right
- Pore-scale simulations of mixed-wet rocks should take a spatially distributed contact-angle field as input, since a mean angle misrepresents the coexistence of water-wet and intermediate-wet pores.
- The 40.4% intermediate-wet fraction in the mixed-wet Bentheimer sample offers a mechanistic explanation for the previously reported near-minimal surface interfaces: those interfaces live in the intermediate-wet pores.
- Averaged wettability metrics are insufficient for classifying mixed-wet systems; samples with nearly identical means can have very different displacement behaviors depending on how the angles are arranged.
- The open-source implementation lets other researchers generate the same spatially resolved wettability maps for their own micro-CT datasets, including CO2-brine and hydrogen-brine systems.
Reading between the lines
- The headline 40.4% intermediate-wet figure is partly an artifact of modeling choices: recomputing the map with a different default angle for uninvaded regions (say, 50° or 90°) would shift the fractions and could weaken or strengthen the minimal-surface link, so a sensitivity analysis is the natural next check.
- The same interpolation machinery could be extended to use invasion-percolation rules or curvature information as priors instead of inverse-distance weighting, which would make the map less dependent on the 20-voxel cutoff.
- If the maps are truly more informative than means, then a direct test follows: network or direct simulations of the mixed-wet sample should reproduce measured capillary pressure and relative permeability better when given the spatially resolved field than when given the single 64.7° value.
- The method's reliance on clear fluid–fluid interfaces means the wettability map is only as good as the segmentation; applying the pipeline to the same voxel data segmented by different algorithms would quantify how much of the reported heterogeneity is real versus inherited from image processing.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper presents an automated geometric contact-angle measurement algorithm for segmented micro-CT images. It avoids direct use of noisy three-phase contact-region segmentation by extrapolating fluid-fluid and fluid-solid normal vectors from nearby interface faces, with polynomial/LOWESS fitting and quality control. Validation is performed on synthetic voxelized droplets on flat and curved substrates at multiple resolutions, on a Ketton limestone ganglion (compared with manual and published measurements), and on several experimental micro-CT datasets. The authors then propose a hierarchical scheme to interpolate sparse loop measurements into a full 3D contact-angle field, assigning 30 degrees to uninvaded defending-fluid regions and inverse-distance weighting elsewhere. Applying this to a mixed-wet Bentheimer sample, they report a mean contact angle of 64.7 degrees but 40.4% intermediate-wet pore space, and argue that averaged wettability metrics misrepresent mixed-wet systems. The paper concludes with open-source code availability.
Significance. The contribution is potentially significant for pore-scale multiphase flow modeling if the measurement method holds up. The core measurement step is well designed: it is benchmarked against analytic solutions for droplets on flat and curved surfaces across five resolutions, it reproduces the independently published Ketton limestone contact angle (41.1 degrees vs 42.2 degrees from Scanziani et al.), and the code is open source. These are concrete strengths. The limitation is that the spatial heterogeneity claim, which is the main advertised advance over prior work, relies on assumed values in uninvaded regions and an unvalidated interpolation model, so the paper's conclusions currently outrun its evidence.
major comments (3)
- [Section 3.3, Section 4.3, Fig. 12] The headline statistic that 40.4% of the mixed-wet Bentheimer sample is intermediate-wet is not a direct measurement: it is an output of the spatial interpolation scheme described in Section 3.3. That scheme assigns a fixed 30 degrees to all uninvaded defending-fluid regions, fills invading-fluid objects and remaining defending-fluid voxels using inverse-distance weighting of sparse contact-loop means, and truncates interpolation beyond 20 voxels from any measurement. No sensitivity analysis is provided for any of these assumptions, and the raw-versus-interpolated histograms in Fig. 12 are not compared quantitatively. Because the abstract and conclusions use this fraction as the key evidence that averaged wettability is inadequate, the central claim is not yet robustly supported. The authors should add a sensitivity study (varying the 30-degree assignment, IDW parameters, and truncation distance) and an uncertainty estimate, and should also resolve the inconsistency between '40.4% of the surface' in Section 4.3 and '40% of its pore space' in the abstract.
- [Section 2.1, Fig. 9] The curved-surface synthetic test is described as having an exact analytical solution, but no derivation or formula for this solution is given in the manuscript. Since this benchmark is the only validation of the method for non-planar solid-fluid interfaces, the reader cannot verify the claimed +/-5 degree accuracy or the RMSE values. The authors should provide the derivation or an explicit formula for the reference contact angle, or re-frame the result as a consistency check.
- [Section 3.2, Algorithm 1, Fig. 7] The normal-vector extrapolation and quality control rely on several free thresholds (d = 0.5 voxel slab thickness, 12-voxel maximum face distance, minimum 20 faces per estimate, 3-voxel polynomial fit range, 30-degree fallback deviation, and the cluster volume threshold Vmin). The paper states these values but gives no sensitivity analysis and no justification beyond the synthetic benchmarks. Because these parameters directly determine which contact nodes are retained and what angles are reported, and the authors note their method returns fewer valid points than AlRatrout et al., a one-at-a-time sensitivity study is needed to demonstrate that the reported accuracy and the measured heterogeneity are robust to the threshold choices.
minor comments (7)
- [Section 2.2.2, Fig. 2] The figure caption lists panels (a), (b), (c), (d), and (f) with no (e); the labeling should be corrected or the missing panel explained.
- [Section 4.2] 'Altratorut et al. (2017)' should be 'AlRatrout et al. (2017)'.
- [Section 2.2.1] 'formally the Digital Rocks Portal' should be 'formerly the Digital Rocks Portal'.
- [Section 3.3] The sentence beginning 'uninvaded pore regions are identified by connected component analysis with 6-connectivity identifies pore regions' is grammatically incomplete and should be rewritten.
- [Section 4.1] 'AlRatrout et al. s higher nominal precision' is missing an apostrophe and space before the possessive 's'.
- [Equation (2)] The dot product inside the arccos should be clamped to [-1, 1] to avoid numerical errors; please state this explicitly.
- [Section 3.2] The many user-set parameters (d, maximum distance, minimum face count, fit range, fallback threshold, Vmin) would be easier to audit if collected in a summary table of default values.
Circularity Check
No significant circularity: the measurement algorithm is validated against analytic synthetic geometries and independent manual/experimental values; the spatial map is a stated interpolation model, not a derivation that reduces to its inputs.
full rationale
The paper's core technical claim is an automated geometric contact-angle measurement validated through multiple independent routes: exact analytical solutions for a droplet on flat and curved surfaces (Fig 7 and Fig 9), comparison with manual measurements on Ketton limestone (Fig 8), and agreement with published experimental values (Scanziani et al., mean 42.2° vs. this work's 41.1°). These benchmarks are external to the paper's fitted assumptions and do not depend on the authors' prior work. The spatially distributed contact-angle maps (Section 4.3) are produced by the interpolation scheme in Section 3.3, which includes modeling choices: uninvaded regions assigned 30°, inverse-distance weighting for invaded objects, and a 20-voxel truncation for defending-fluid interpolation. Although the headline 40.4% intermediate-wet fraction is an output of this interpolation rather than a direct measurement, this is a transparent, disclosed modeling step rather than a circular reduction: the interpolated field is neither defined in terms of the final statistic nor fitted to produce it. The paper does cite works with overlapping authorship (AlRatrout et al. 2017, Blunt 2017, Foroughi et al. 2020, Raeini et al. 2022, Wang et al. 2024), but these citations are used for context, benchmarking, or background, not as load-bearing justification for the paper's own measurement accuracy or the heterogeneity claim. No equation in the paper defines its output as equivalent to its input, and no parameter is fitted to a subset of data and then renamed a prediction. The absence of sensitivity analysis for the interpolation choices is a robustness limitation, not circularity. Hence the derivation chain is self-contained and does not reduce to its own inputs.
Assumptions & free parameters
free parameters (9)
- uninvaded region contact angle =
30 degrees
- normal extrapolation search window thickness d =
0.5 voxel unit
- maximum extrapolation distance =
12 voxel units
- minimum adjacent faces for normal estimation =
20
- polynomial fit range =
3 units from contact point
- normal deviation fallback threshold =
30 degrees
- max interpolation distance for defending fluid =
20 voxels
- contact angle clipping range =
1 degree to 180 degrees
- cluster volume threshold Vmin =
not specified
assumptions (6)
- domain assumption Marching cubes surfaces from segmented voxels faithfully approximate true phase interfaces away from contact lines
- domain assumption Static steady-state fluid configurations have contact angles representative of intrinsic equilibrium wetting angles
- ad hoc to paper Uninvaded defending-fluid regions are strongly water-wet in primary drainage and correctly assigned 30 degrees
- ad hoc to paper Inverse-distance weighted interpolation of sparse loop means produces physically consistent contact angle fields
- domain assumption Watershed segmentation and NLM or unsharp pre-processing correctly isolate the three phases
- standard math Taubin smoothing preserves mean curvature and normals at the scale needed for contact angles
Cite this review
Pith. "Pith review of Spatially Distributed Wettability Characterization in Porous Media." pith.science (2026). https://pith.science/paper/Q2OUUIID
@misc{pith2026250701617,
author = {Pith},
title = {Pith review of: Spatially Distributed Wettability Characterization in Porous Media},
year = {2026},
howpublished = {\url{https://pith.science/paper/Q2OUUIID}},
note = {Machine review of arXiv:2507.01617}
}
read the original abstract
An enhanced geometric algorithm for automated pore-by-pore contact angle measurement from micro-CT images, is presented that achieves superior accuracy compared to existing methods through robust fluid-fluid and solid-fluid interface extrapolation. Using this high resolution data, we generate spatially distributed contact angle maps that reveal previously hidden wettability heterogeneity. Our analysis of mixed-wet systems demonstrates the severe limitations of averaged metrics: a sample with a mean contact angle of 64.7 degrees, conventionally classified as uniformly weakly water-wet, exhibits 40% of its pore space in the intermediate-wetting regime (70-110 degrees). This heterogeneity explains the presence of minimal surface interfaces and fundamentally different pore-filling mechanisms operating within the same sample. By providing open-source tools for spatially-resolved wettability characterization, this work enables more accurate predictions of multiphase flow behavior in heterogeneous porous materials, essential for optimizing subsurface energy storage and recovery processes.
Reference graph
Works this paper leans on
-
[1]
Introduction Geoengineering, specifically gas storage in geological formations (Krevor et al., 2023), and oil recovery, are essential for the development of sustainable energy resources. These processes are governed by immiscible multiphase flow in porous media, where displacement and trapping mechanisms are predominantly capillary dominated (Singh et al....
work page 2023
-
[2]
Image Data Used and Processing Steps We assembled two complementary datasets to evaluate our contact‐angle measurement code in porous media containing two immiscible phases: (i) synthetic voxelized images with known analytical solutions, and (ii) high‐resolution micro -CT scans of rock samples subjected to flooding experiments. 2.1. Synthetic Voxelized Da...
-
[3]
Isolated Oil Ganglia Datasets Two high-resolution micro-CT datasets were obtained from a strongly water-wet Ketton limestone sample following an oil injection (primary drainage) and subsequent waterflooding cycle (Scanziani et al., 2017; Singh et al., 2017) . These volumes, acquired at 2 μm voxel size, capture isolated oil ganglia trapped within brine-sat...
work page 2017
-
[4]
Brine was then injected to initiate oil recovery
Waterflooding Datasets A Bentheimer sandstone sample was initially saturated with brine, followed by oil injection to establish initial conditions. Brine was then injected to initiate oil recovery. A fractional-step waterflooding method was employed, in which brine and oil were sequentially injected in discrete steps with incrementally increasing brine fr...
work page 2018
-
[5]
Hydrogen Storage Dataset A recent experiment on a strongly water -wet Bentheimer sandstone investigated hydrogen injection into initially brine-saturated pores (Waleed et al., 2024). Hydrogen formed disconnected ganglia within larger pores, resulting in non -continuous, high - curvature fluid interfaces under multiphase conditions. This dataset provides a...
work page 2024
-
[6]
Measurement of Contact Angle 3.1. Manual method for contact angle estimation from 3D images To manually measure contact angles from real 3D images, we followed the method of Andrew et al., (2014), with some changes. The procedure consists of the following steps
work page 2014
-
[7]
From the segmented image, three -phase contact loops were extracted and smoothed (see Section 4.2)
-
[8]
For each node on the contact loop, a secant line was defined based on its neighboring nodes
Show all 17 references
-
[9]
A 2D grayscale slice, perpendicular to the secant line, was extracted from the raw image
-
[10]
In contrast to the method of Andrew et al., (2014), which used filtered images for fitting interface lines, we used the edge image directly to reduce user subjectivity
This 2D slice was then processed using the edge -detection steps described in Section 3.2.2. In contrast to the method of Andrew et al., (2014), which used filtered images for fitting interface lines, we used the edge image directly to reduce user subjectivity. The detected ed...
2014
-
[11]
loop" • else • ℓ.type←
The contact angle was measured from the resulting edge image by drawing two tangent lines at the three-phase contact point using the open -source software 3D Slicer. These tangent lines were drawn relative to the contact node itself, rather than following t he detected edge pa...
2017
-
[12]
Results and Discussion 4.1. Measurements Accuracy and Uncertainty To evaluate the accuracy and robustness of our automated contact -angle algorithm, we designed three distinct tests—each chosen to probe a different source of error in voxeli zed datasets. First, we simulated a ...
2017
-
[13]
Conclusions This study addresses a fundamental challenge in characterizing wettability of porous materials, where recent advances have highlighted that spatially distributed wetting information, together with pore space geometry, controls displacement and trapping mec hanisms ...
-
[14]
Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper
-
[15]
Acknowledgment The first author would like to thank Khalifa University of Science and Technology for providing the PhD scholarship that supported his research
-
[16]
Data Availability Code is available at https://github.com/ImperialCollegeLondon/geometricContactAngle.git
-
[17]
Pore -scale X-ray imaging with measurement of relative permeability, capillary pressure and oil recovery in a mixed-wet micro- porous carbonate reservoir rock
References Alhammadi, A.M., Gao, Y ., Akai, T., Blunt, M.J., Bijeljic, B., 2020. Pore -scale X-ray imaging with measurement of relative permeability, capillary pressure and oil recovery in a mixed-wet micro- porous carbonate reservoir rock. Fuel 268, 117018. https://doi.org/10...
2020
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.