Palm Oil Waste Clinker Weakens Concrete by Coarsening Its Pore Structure

Concrete made with palm oil clinker, a porous byproduct of the palm oil industry, loses a substantial share of its mechanical performance as the clinker replaces natural aggregates, and a new laboratory study traces that deterioration directly to changes in the material’s internal pore architecture. Researchers prepared fifteen concrete mixtures in which crushed granite and river sand were replaced with coarse and fine palm oil clinker at controlled volume ratios, then combined mechanical testing, microscopy, mercury intrusion porosimetry, fractal analysis and grey entropy correlation to link the strength losses to specific features of the pore system. Their central finding is that coarse clinker, not fine clinker, drives the most damaging transformation: it shifts the pore size distribution from mainly unimodal to distinctly bimodal and raises the fractal dimension of the pore network, a quantitative signature of increasing structural complexity.

The experimental program used a fixed water-to-cement ratio of 0.48 and a cement content of 450 kg/m³ across all mixtures. The coarse palm oil clinker, with a particle size range of 5 to 20 mm and an apparent density of 1450 kg/m³, replaced granite at 20, 50, 70 and 100 percent by volume. The fine clinker, ranging from 0.15 to 4.75 mm with an apparent density of 1950 kg/m³, replaced sand at 20 and 30 percent. Combined mixtures paired coarse replacement levels from 20 to 100 percent with fine replacement at 20 or 30 percent. Because the clinker is far more absorptive than natural aggregate, with 24-hour water absorption of 5.52 percent for the coarse fraction and 10.82 percent for the fine fraction, the team added compensation water calculated from a 0.60 absorption factor and allowed the aggregates to stand wet for roughly 30 minutes before cement was introduced.

The physical results followed a consistent pattern. As coarse clinker content rose, dry density fell monotonically and water absorption climbed. The control mixture reached a 28-day dry density of 2470 kg/m³ with 1.5 percent water absorption, while the mixture containing 100 percent coarse and 30 percent fine clinker dropped to 1777 kg/m³, 28.1 percent lower, with absorption rising to 3.7 percent. Across all fifteen mixtures, compressive strength increased strongly with density, yielding a linear regression coefficient of R² = 0.977. Water absorption showed a weaker and more scattered relationship with strength, at R² = 0.886, which the authors interpret as evidence that bulk absorption alone cannot capture the pore characteristics that govern mechanical deterioration.

Mechanical testing at 1, 7, 14 and 28 days confirmed that every clinker-containing mixture underperformed the control. At 28 days, the fine clinker series showed compressive strength reductions of 9.4 to 13.6 percent relative to the control, whereas the coarse clinker series lost 10.6 to 44.1 percent, with the largest losses occurring at full coarse replacement. The combined mixture with 100 percent coarse and 30 percent fine clinker recorded a 44.6 percent reduction in compressive strength. Splitting tensile strength followed the same hierarchy, with fine-series losses of 8.8 to 11.1 percent at 28 days against 17.7 to 35.0 percent for the coarse series, and flexural strength showed reductions of 9.1 to 11.3 percent for fine replacement versus 17.1 to 34.5 percent for coarse. Notably, at 20 percent coarse replacement, adding 20 percent fine clinker caused only a small further strength reduction, and the combined mixture POC-20–20 performed comparably to its coarse-only counterpart.

Microscopy helped explain why. Handheld digital microscope images of fracture surfaces revealed a narrow, continuous aggregate-paste interface in the control concrete, while the mixture with 20 percent coarse clinker showed visible cavities and locally separated zones along the porous clinker surface. In contrast, the mixture combining 20 percent coarse with 20 percent fine clinker, and the mixture with 30 percent fine clinker alone, both displayed comparatively continuous bonding interfaces and paste coverage. Scanning electron microscopy at 15 kV corroborated these observations: cementitious material was seen occupying surface irregularities of fine clinker particles, and hydration products appeared along the rough clinker surface in the low-replacement combined mixture, conditions consistent with mechanical interlocking between aggregate and paste. At full coarse replacement, the images showed discontinuous paste coverage, interfacial gaps and local voids, alongside a less compact and more heterogeneous cement matrix.

Mercury intrusion porosimetry on three representative 28-day mixtures quantified the bulk pore structure behind these images. The control concrete had a total accessible pore volume of 0.0435 mL/g and a most probable pore diameter of 40.31 nm with a narrow unimodal distribution. Replacing 30 percent of the fine aggregate with clinker raised total pore volume by 86.0 percent to 0.0809 mL/g, yet the distribution remained mainly unimodal with only limited large-pore intrusion, and the natural coarse aggregate skeleton stayed intact. Pushing coarse replacement from 0 to 100 percent at the same fine replacement level increased pore volume a further 40.4 percent to 0.1136 mL/g, but the more consequential change was qualitative: the distribution became distinctly bimodal, with increased intrusion in the micrometer range, which the authors associate with internal clinker pores and a coarsened interfacial transition zone. This pore coarsening accompanied the compressive strength drop from 50.4 MPa for the fine-only mixture to 32.3 MPa for the fully coarse-replaced mixture, compared with 58.3 MPa for the control.

Fractal analysis converted these intrusion curves into a single complexity metric. Using a thermodynamic method that relates cumulative mercury intrusion energy to cumulative intrusion volume, the team calculated fractal dimensions of 2.6901 for the control, 2.7411 for the fine-replaced mixture and 2.9681 for the fully coarse-replaced mixture, with regression R² values of 0.995 or higher. The increment from fine replacement was only 0.0510, whereas the subsequent coarse replacement step added 0.2270, roughly 4.5 times larger, identifying coarse clinker and its interfacial deterioration as the dominant contributor to pore surface irregularity and multiscale complexity. Across the three mixtures, fractal dimension and 28-day compressive strength followed the linear relation y = -89.333x + 297.111 with R² = 0.984, meaning that within this dataset an increase of 0.1 in fractal dimension corresponded to approximately 8.9 MPa lower strength. The authors stress that this three-point regression is a mixture-specific mechanistic trend, not a universal design equation.

Grey entropy correlation analysis, applied to 60 observations from the fifteen mixtures at four curing ages, ranked the factors associated with mechanical performance. Curing age showed the highest correlation degree, between 0.868 and 0.892, followed by density at 0.829 to 0.841, the coarse clinker replacement ratio at 0.743 to 0.765, the fine clinker ratio at 0.712 to 0.723, and water absorption lowest at 0.691 to 0.708. The authors caution that the method measures sequence similarity rather than causality, but the ordering is consistent with the independent mechanical, interfacial and pore-structure evidence: coarse clinker exerts the stronger influence, and bulk water absorption is the least informative single indicator of pore-related strength deterioration.

The study’s limitations frame how far these conclusions travel. The fractal and porosimetry analyses covered only three representative mixtures, the strong fractal-strength correlation may partly reflect their shared dependence on clinker content, and the individual contributions of aggregate porosity, interfacial characteristics and particle packing could not be separated within the experimental design. Mercury intrusion records pore throat diameters rather than direct pore geometry, and the ink-bottle effect limits direct interpretation of connectivity. The authors recommend presaturating or coating coarse clinker, controlling aggregate pre-wetting and compaction, and optimizing fine clinker grading as potential corrective routes, while calling for validation across intermediate replacement ratios, replicate porosimetry tests and other porous, recycled and lightweight aggregates before the framework is generalized beyond palm oil clinker concrete.

Subject of Research: Effects of coarse and fine palm oil clinker aggregate replacement on concrete mechanical performance, pore structure and fractal dimension

Article Title: Mechanical performance and pore structure fractality of palm oil clinker concrete: Effects of fine and coarse aggregate replacements

Article References: Tan, Z., Fong, C. S., Mwafy, A., Su, T., & Yap, S. P. (2026). Mechanical performance and pore structure fractality of palm oil clinker concrete: Effects of fine and coarse aggregate replacements. Case Studies in Construction Materials, 25, Article e06650. https://doi.org/10.1016/j.cscm.2026.e06650

Image Credits: AI Generated

DOI: 10.1016/j.cscm.2026.e06650

Keywords: palm oil clinker, concrete, compressive strength, mercury intrusion porosimetry, fractal dimension, aggregate-paste interface, water absorption, lightweight aggregate, grey entropy correlation, pore size distribution, construction materials, solid waste recycling