Why Lightweight Aggregates Float in Concrete and How to Stop Them

Lightweight aggregate concrete has long promised engineers a rare combination of benefits: structures that weigh less, insulate better, and carry respectable loads despite a porous skeleton. Yet the very property that makes lightweight aggregates useful—their low density—also undermines them. In fresh concrete, these porous particles tend to drift upward, segregating into weak, aggregate-rich zones near the surface while leaving denser paste below. A new study published in Case Studies in Construction Materials by Weihao Zhang, Chen Qian, Zhenbo Wang, Fujie Jia, and Shunzeng Zhao systematically dissects this floating behavior, quantifying how aggregate properties and paste chemistry jointly control where lightweight aggregates end up, and how that spatial distribution in turn dictates the strength, density, and shrinkage of the hardened material.

The team began at the source, manufacturing their own lightweight aggregates from a blend of clay and volcanic ash. By tuning the sintering regime—heating granules at 5 °C per minute, preheating at 400 °C for 20 minutes, then firing between 1100 and 1250 °C—they could dial in the pore structure, density, and strength of each batch. Raising the sintering temperature from 1100 to 1200 °C melted the glassy phase progressively, forming a continuous surface glaze and a uniform closed-pore structure that lowered both density and water absorption while boosting strength. Pushing the temperature to 1250 °C backfired: excessive softening of the glassy phase allowed bubbles to coalesce, enlarging pores and thinning their walls until the aggregates weakened. Holding time followed a similar pattern, with the best cylinder compressive strength of 10.7 MPa achieved after 20 minutes at temperature. Longer grinding of the raw materials made aggregates lighter but also more porous and fragile, since finer feedstocks foamed more vigorously during firing.

With a library of aggregates spanning density grades from 600 to 1200 kg/m³, the researchers turned to the fresh state. They developed a simple but powerful evaluation protocol: after vibrating a concrete specimen, partition plates divided it into upper, middle, and lower thirds, and the aggregates recovered from each layer were weighed. From these layer-by-layer mass fractions they computed a floating index, which measures the excess of aggregates in the top two layers relative to the bottom, and a coefficient of variation, which captures overall spatial non-uniformity. For a perfectly homogeneous mix, each layer would hold roughly one third of the aggregates; any deviation signals redistribution. Vibration duration proved critical—at 20 seconds of vibration both indices rose sharply as the paste’s flocculated structure broke down, so the team standardized on 10 seconds of vibration at 50 Hz to keep the comparisons meaningful.

The aggregate experiments delivered a clear hierarchy of influence. As density grade increased from 600 to 1200, both the floating index and the coefficient of variation fell substantially, confirming that the density difference between aggregate and paste is the primary driving force for upward migration. Particle size came next: larger aggregates, fewer in number at the same volume fraction and presenting less surface area to the paste, floated more readily. Morphology and gradation played secondary but measurable roles. Spherical aggregates, with their smooth surfaces and low rolling resistance, drifted upward more easily than irregular ones, while continuously graded mixes outperformed single-sized ones because smaller particles fill the gaps between larger ones, creating mechanical interlock that resists migration. The worst case—low-density, large, spherical, single-sized aggregates—produced floating indices as high as 47.9 percent.

Paste parameters told an equally instructive story. Increasing the water-to-binder ratio raised both segregation indices roughly linearly, because additional free water thins the paste and erodes its capacity to suspend particles. The researchers also introduced an elegant volumetric metric, the excess filling rate, defined as the volume of mortar beyond what is needed to fill the voids between packed aggregates, normalized by that void volume. At low filling rates, frequent aggregate-to-aggregate contacts provide mechanical restraint; as the rate climbs, those contacts thin out and floating intensifies. Most striking was the effect of supplementary cementitious materials. Fly ash, with its spherical ball-bearing particles, lubricated the paste and worsened segregation, while slag—with its high specific surface area and hydraulic activity—thickened the paste and suppressed floating. The binary fly ash–slag blend proved the champion, cutting the floating index to 5.68 percent and the coefficient of variation to 0.90 by balancing flowability with cohesiveness.

A crucial insight emerged from linking these results to concrete slump, the industry’s standard workability measure. Slump alone cannot predict segregation. Two mixes with identical slump can behave entirely differently depending on whether the flowability gain came from added water, extra paste volume, or mineral admixtures. Water addition destroys suspension capacity; extra paste mainly reduces particle contacts without sacrificing viscosity; slag raises flow while preserving cohesiveness. The fly ash–slag system achieved high slump with minimal floating, demonstrating that flowability and segregation resistance can coexist when the chemistry is right. This finding cautions against mix-design shortcuts that judge fresh concrete by a single workability number.

To explain these observations mechanistically, the team built a kinetic model based on the force balance on a rising aggregate particle: buoyancy drives it up, gravity pulls it down, and Stokes-type viscous drag resists motion, with an additional term accounting for interparticle restraint. In the steady state, the migration velocity scales with the square of particle radius and the density difference, divided by paste viscosity and modified by the restraint factor. This simple proportionality reproduced every major experimental trend—why bigger and lighter aggregates float fastest, why thinner pastes accelerate migration, and why gradation and shape act through particle contacts rather than viscosity. The model offers a predictive framework that can be calibrated for other lightweight aggregate systems, giving mix designers a quantitative tool rather than trial-and-error folklore.

The hardened concrete results tied distribution to performance in unexpected ways. Raising aggregate cylinder compressive strength from 2.4 to 18.7 MPa lifted 28-day compressive strength of the concrete from 23.0 to 59.5 MPa, because weak aggregates shift from the composite’s weakest phase to genuine load-bearers. Compressive strength correlated most strongly with the water-to-binder ratio (R² = 0.91) and aggregate strength (R² = 0.81), while flexural strength responded far more to paste parameters, with R² values of 0.98 for the water-to-binder ratio and 0.96 for the excess filling rate—consistent with flexural failure being governed by crack propagation through the matrix and interfacial transition zone rather than bulk crushing. Intriguingly, the floating index correlated only weakly with compressive strength (R² = 0.75) and barely at all with flexural strength (R² = 0.11), but strongly with specific strength—strength per unit density—with a correlation coefficient of −0.94. Segregation, in other words, is best detected not by strength loss alone but by the efficiency metric that couples load capacity to weight.

The practical payoff is substantial. The optimized mixes achieved specific strengths of 25 to 33 MPa·m³/t, roughly double the 8.5 to 16.5 typical of ordinary concrete, meaning structures can shed self-weight without proportional strength sacrifice. Drying shrinkage told a favorable story as well: all lightweight mixes shrank less than conventional counterparts, thanks to internal curing as porous aggregates release stored water into the hydrating paste, though very porous low-grade aggregates partially offset this benefit by restraining the skeleton less. Microstructural analysis of the best-performing fly ash–slag system showed pores below 20 nanometers accounting for about 95 percent of fine-pore volume at 28 days, with a continuous C-S-H gel network penetrating the open pores at aggregate surfaces and knitting paste to particle. For engineers racing to build lighter, taller, and more sustainable structures, the message is clear: controlling where lightweight aggregates sit in the fresh state—through density matching, gradation, paste viscosity, and blended binders—is as important as the aggregates themselves.

Subject of Research: Floating-driven segregation of lightweight aggregates and its effects on the hardened performance of lightweight aggregate concrete

Article Title: Floating-driven spatial distribution of lightweight aggregates and its effects on the hardened performance of lightweight aggregate concrete

Article References: Zhang, W., Qian, C., Wang, Z., Jia, F., & Zhao, S. (2026). Floating-driven spatial distribution of lightweight aggregates and its effects on the hardened performance of lightweight aggregate concrete. Case Studies in Construction Materials, 25, Article e06533. https://doi.org/10.1016/j.cscm.2026.e06533

Image Credits: AI Generated

DOI: 10.1016/j.cscm.2026.e06533

Keywords: lightweight aggregate concrete, aggregate floating, segregation, sintered aggregates, floating index, excess filling rate, specific strength, drying shrinkage, fly ash, slag, interfacial transition zone, internal curing