Fly ash and sand ratio reveal hidden trade-offs in joint-filling mortar

Joint-filling mortars face a stubborn engineering dilemma: they must flow freely into place, then harden into a load-bearing material that is not excessively stiff. A new systematic study published in Frontiers in Materials maps exactly where that balance breaks down, showing that equal steps in sand content do not buy equal gains in performance. The research, led by Zhengxin Wu of CCCC-FHDI Engineering Co. Ltd. with colleagues in Guangzhou and Xi’an, China, combined a 15-mixture composition grid with a replicated validation experiment under a fixed admixture dosage to resolve how fly ash replacement and sand-to-binder ratio jointly shape workability, strength, and stiffness.

The experimental core was a 5 × 3 full-grid program. Sand-to-binder ratio (S/B) took five values from 0.35 to 0.55, fly ash (FA) replacement took three values from 40% to 60% of total binder mass, and the water-to-binder ratio was fixed at 0.30 throughout. Polypropylene fiber was held at 2.0 vol.% of total mortar volume, corresponding to 18.2 kg/m³ using a nominal fiber density of 0.91 g/cm³, with a nominal fiber length of 12 mm and a maximum sand size of 0.6 mm. The binder system combined 42.5-grade ordinary Portland cement with Grade I ultrafine fly ash, and river sand was hand-sieved so that only the fraction passing the 0.6-mm sieve entered the mixes. All 15 mixtures were prepared with a paste-mediated staged mixing sequence in which cementitious materials were first dry-mixed, an intermediate paste was formed at a water-to-binder ratio of roughly 0.26, fibers were then worked in portions, and sand, remaining water, and admixtures were added last.

Processing behavior was measured as free-spread diameter on a mortar flow table without jolting, because the fiber-rich mortar could obstruct a flow-cone outlet. Flexural and compressive strengths were tested at 7 and 28 days on 40 mm × 40 mm × 160 mm prisms under standard curing at 20 °C ± 2 °C and relative humidity of at least 90%. The 28-day stiffness response was quantified as a compressive secant modulus following the GB/T 7897-2008 procedure, defined over the loading response up to 40% of axial compressive strength, with repeated measurement cycles until the strain difference between successive cycles did not exceed 20 × 10⁻⁶. One important caveat shaped the analysis: water-reducer dosage was adjusted within a ceiling of 0.8% of binder mass during full-grid preparation, but mixture-specific dosages were not retained, so the full-grid free-spread values were treated as descriptive realized responses rather than directly comparable composition-only data.

The full-grid results revealed a clear opposing pair of gradients. Averaging across the five sand levels, increasing fly ash from 40% to 60% reduced mean 28-day compressive strength from 48.36 to 38.08 MPa and mean secant modulus from 13.6 to 8.6 GPa, a 20-percentage-point increase in replacement associated with losses of 10.28 MPa in strength and 5.0 GPa in modulus. Increasing sand content moved the material in the opposite direction through most of the tested range: from S/B = 0.35 to 0.50, mean 28-day compressive strength rose from 34.17 to 49.37 MPa and mean secant modulus from 10.33 to 12.33 GPa. Empirical response models fitted to the grid quantified these directions, with a coefficient of +7.53 MPa in 28-day compressive strength for a 0.10 increase in S/B and −5.14 MPa for a 10-percentage-point increase in fly ash, and corresponding modulus coefficients of +1.53 GPa and −2.50 GPa. Cross-validated R² values reached 0.800 for strength and 0.832 for modulus, indicating that the fitted surfaces reproduced the principal measured patterns.

Age-dependent differentiation added a second layer to the picture. All 15 mixtures were stronger at 28 days than at 7 days, but the magnitude depended on composition. At the fly-ash group level, mean 7-day and 28-day compressive strengths were 28.62 and 48.36 MPa at 40% replacement, 26.52 and 41.98 MPa at 50%, and 21.32 and 38.08 MPa at 60%. The 60% fly-ash group showed the largest 28-day-to-7-day strength ratio in the dataset, reaching 1.79, and also the largest mean 7-to-28-day flexural-strength increment at 1.86 MPa. Within the 60% series, 7-day compressive strength climbed from 10.2 MPa at S/B = 0.35 to 27.9 MPa at S/B = 0.50, a gain of 17.7 MPa that dwarfed the corresponding increases of 10.1 MPa at 50% fly ash and 4.6 MPa at 40%. The fitted 7-day strength model captured this through a positive S/B × fly-ash interaction coefficient of +3.45 MPa, although its cross-validated R² of 0.580 was lower than for the 28-day models, consistent with greater local variability in the early-age field.

The pivotal finding emerged at the upper end of the sand range. From S/B = 0.35 to 0.50, mean 28-day compressive strength across the three fly-ash levels increased by 15.20 MPa. But the final equal increment, from S/B = 0.50 to 0.55, produced a mean strength decrease of 1.23 MPa while mean secant modulus continued to rise by 0.67 GPa. At the individual fly-ash levels, the 0.50-to-0.55 step changed 28-day strength from 53.2 to 51.3 MPa at 40% replacement, from 49.8 to 50.6 MPa at 50%, and from 45.1 to 42.5 MPa at 60%, so the final increment did not deliver a consistent strength gain. In other words, equal sand increments did not retain the same marginal response: the mechanical benefit of added sand ran out near S/B ≈ 0.50 within this tested system.

Because full-grid free-spread values could not support composition-only inference, the team ran a targeted validation experiment under a fixed high-range water reducer dosage of 0.70% of binder mass. Five mixtures were retested in a randomized complete block design with three blocks, using independent mixing date as the blocking factor. M10, M11, and M12 formed a fly-ash axis at S/B = 0.50 with 40%, 50%, and 60% replacement, while M9, M12, and M15 formed a sand axis at 60% fly ash spanning S/B = 0.45, 0.50, and 0.55. Each block contained one independently weighed and mixed batch of each mixture, with randomized mixing order, and specimen-level determinations were averaged before block-level analysis. Mixing date contributed statistically detectable variation to all four validation responses, with block p-values at or below 0.0022, but treatment ordering was identical across all three blocks for every response.

The validation confirmed both the fly-ash gradient and the change in marginal sand response with quantified uncertainty. At S/B = 0.50, raising fly ash from 40% to 60% increased block-adjusted mean free spread from 174.33 to 194.17 mm, an effect of +19.83 mm with a 95% confidence interval of +14.94 to +24.73 mm, while 28-day compressive strength fell by 7.46 MPa and secant modulus by 5.70 GPa. At 60% fly ash, the first sand increment from S/B = 0.45 to 0.50 reduced free spread by 7.17 mm while increasing 28-day compressive strength by 5.20 MPa, with a modulus change of +0.11 GPa whose confidence interval included zero, meaning no resolved modulus change. The next equal increment to S/B = 0.55 told a different story: free spread dropped by 20.50 mm, 28-day strength decreased by 2.53 MPa, and secant modulus increased by 1.92 GPa. Second-difference contrasts quantified the shift in marginal response at −13.33 mm for free spread, −5.24 MPa for 7-day strength, −7.73 MPa for 28-day strength, and +1.81 GPa for modulus, with all four 95% confidence intervals excluding zero.

To translate these coupled responses into practical guidance, the researchers applied Pareto screening to the five fixed-dosage validation mixtures using three objectives: maximize free spread, maximize 28-day compressive strength, and minimize secant modulus. Four mixtures emerged as non-dominated alternatives: M9, M10, M11, and M12. Their treatment means spanned free-spread diameters of 174.33 to 201.33 mm, 28-day strengths of 39.78 to 52.43 MPa, and moduli of 8.97 to 14.78 GPa. M9 occupied the higher-spread, lower-stiffness end with 201.33 mm of spread, 39.78 MPa of strength, and 8.97 GPa of modulus. M10 delivered the highest strength at 52.43 MPa alongside the highest modulus at 14.78 GPa and the lowest spread at 174.33 mm. M11 and M12 occupied intermediate trade-off states, and M15 was dominated by M12, which was more favorable in all three objectives. No single mixture won across the board, which is precisely the point: the non-dominated set preserves the conflict among workability, strength, and stiffness rather than collapsing it into one ranking.

The authors interpret the two-stage sand response through granular packing and paste availability. At lower-to-intermediate sand contents, additional fine aggregate strengthens the granular load-bearing framework while enough paste remains to lubricate and bind particles, which fits the strong strength gain and small stiffness change observed from S/B = 0.45 to 0.50. Beyond that range, particle crowding becomes more influential: higher fine-aggregate contents raise plastic viscosity through hydrodynamic and direct particle-contact interactions, and reduced excess-paste-layer thickness is associated with higher yield resistance. The larger spread penalty, negative strength increment, and continued stiffening over S/B = 0.50 to 0.55 are compatible with a shift toward a more contact-dominated material state. Crucially, the team cautions that S/B ≈ 0.50 marks an observed change in marginal response specific to this material and processing window, not a universal threshold. The findings apply to the tested conditions of water-to-binder ratio 0.30, 2.0 vol.% polypropylene fiber, 40–60% fly ash, S/B from 0.35 to 0.55, the selected sand and binder system, and the adopted mixing and admixture conditions, and engineering implementation would still require verification of bond, shrinkage, repeated joint movement, durability, and field placement. Within that window, however, the study gives mixture designers a quantified map of where additional sand stops paying for itself in strength and starts charging for it in workability.

Subject of Research: How fly ash replacement and sand-to-binder ratio control workability, strength, and stiffness trade-offs in fiber-reinforced joint-filling mortar

Article Title: Fly ash replacement and sand-to-binder ratio shape trade-offs among workability, strength, and stiffness in joint-filling mortar

Article References: Wu, Z., Han, B., Zhu, Z., & Han, L. (2026). Fly ash replacement and sand-to-binder ratio shape trade-offs among workability, strength, and stiffness in joint-filling mortar. Frontiers in Materials, 13, Article 1984896. https://doi.org/10.3389/fmats.2026.1984896

Image Credits: AI Generated

DOI: 10.3389/fmats.2026.1984896

Keywords: joint-filling mortar, fly ash replacement, sand-to-binder ratio, compressive strength, secant modulus, free-spread workability, polypropylene fiber, Pareto screening, randomized complete block design, mixture design, replacement, sand-to-binder