Freeze–thaw damage, quantified at the microstructural scale
Freeze–thaw cycling is a major durability challenge for cementitious materials used in cold-region environments. While bulk tests only capture the consequences of loading at the macroscopic scale (e.g., stiffness loss, strength reduction, increased permeability), pore-scale observations of damage evolution, by contrast, make it possible to investigate the underlying mechanisms that drive degradation and ultimately lead to cracking.
A newly published, open-access paper in Materials & Design addresses this gap using time-lapse X-ray micro-computed tomography (micro-CT) and deep-learning segmentation to track damage evolution inside a polyester fiber-reinforced cementitious composite across 100 freeze–thaw cycles.
Study at a glance
Paper: Freeze-thaw-induced microstructural damage in polyester fiber-reinforced cementitious composites revealed by X-ray microtomography
Imaging approach: High-resolution micro-CT (11 µm voxel) repeated across multiple time points on the same specimen
Analysis approach: Deep-learning segmentation to separate phases in a heterogeneous composite where gray levels overlap
What the researchers found
Across freeze–thaw cycling under saturated condition in closed system , the authors report a clear increase in damage and important spatial trends:
- Pore + crack volume fraction increased from 10.0% to 21.0%, with local maxima reaching 24.8% after 100 cycles.
- Polyester reinforcement fibers locally amplified cracking, with stress concentration effects occurring around fiber–matrix interfaces.
- Pore network geometry and connectivity were shown to influence where damage initiates and how it propagates over cycling.
- A thermo-mechanical interpretation identified thermal expansion mismatch between ice and the cement matrix as the dominant cracking mechanism under the study’s fully saturated conditions.
Where Dragonfly 3D World supported the workflow
The team used Dragonfly 3D World to support the image analysis required for time-lapse micro-CT comparison, with a particular focus on segmentation and tracking damage progression through repeated scans.
A key challenge in this dataset is that different phases can overlap in grayscale intensity, which makes traditional threshold-only segmentation unreliable. The authors addressed this using a deep-learning segmentation approach to achieve reliable phase separation in a complex, heterogeneous composite.
In their provided notes, the user specifically highlighted 3D World’s value for:
- Segmenting heterogeneous materials where phase gray levels overlap
- Tracking and segmenting crack evolution across freeze–thaw cycles
- Building a workflow that supports repeatable analysis across time points
Why this matters
By combining time-lapse micro-CT with deep-learning segmentation, this work offers a practical template for studying freeze-thaw damage in 3D not only identifying that damage occurs, but quantifying how pore/crack networks evolve, and connecting those observations to physically grounded cracking mechanisms.
It’s also a strong example of how advanced CT image analysis can help translate high-resolution scans into mechanism-informed insight that materials and infrastructure teams can act on, especially in cold-region applications where freeze–thaw durability is a key constraint.
Read the paper
The full article is open access in Materials & Design: https://doi.org/10.1016/j.matdes.2025.115256
Acknowledgement
Sophie Jung1, Jordan Harvey2, Hubert Taieb3, Sam Bhat4, Nicolas Piche3, Mahya Roustaei5 & Pooneh Maghoul1
- Department of Civil, Geological and Mining Engineering, Polytechnique Montreal, Montreal, Quebec, Canada
- Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, Alberta, Canada
3. Comet Technologies Canada Inc., Montreal, Quebec, Canada
4. Titan Environmental Containment Ltd., Iles des Chenes, Manitoba, Canada
5. Department of Civil Engineering, Ghent University, Laboratory of Geotechnics, Technologiepark 68, 9052 Zwijnaarde, Belgium
Try Dragonfly 3D World
If your work involves micro-CT analysis, crack/porosity quantification, or segmentation in heterogeneous materials, you can explore Dragonfly 3D World with a free trial and evaluate workflows for segmentation, visualization, and quantitative analysis.