Firefighters’ Choice of Coolant Shapes How Much Strength Concrete Keeps After a Blaze

When a building fire is finally knocked down, the damage to its concrete skeleton may not be over. A new experimental study suggests that the very substance firefighters use to quench the flames can determine how much load-bearing capacity the structure retains, with water — the most common firefighting agent — inflicting the greatest additional harm, and fire-suppression foam preserving the most strength. The findings, published in Case Studies in Construction Materials, offer a practical framework for engineers assessing fire-damaged structures and for fire services choosing suppression strategies.

The research team, led by Daniel Paul Thanaraj and colleagues including N. Anand and Éva Lublóy, heated M20-grade concrete specimens to 200, 400, 600 and 800 degrees Celsius in a programmable electric furnace, holding each target temperature for one hour at a slow heating rate of 10 degrees per minute. K-type thermocouples embedded at the surface and at the geometric core of 150-millimetre cubes confirmed that surface temperatures tracked the furnace closely while the core lagged behind, a gap that widened with higher target temperatures. After heating, the specimens were cooled in five different ways: passive natural air cooling, water sprayed from a hose, carbon dioxide gas from a standard extinguisher, sodium bicarbonate dry-powder extinguisher, and aqueous film-forming foam.

The choice of coolant mattered far more than the researchers expected. Foam-cooled specimens consistently retained the highest residual strength across all three mechanical tests — compression, split tension and flexure — while water-quenched specimens fared worst. After exposure to 800 degrees Celsius, foam-cooled cubes kept 73.35 percent of their original compressive strength of 25.52 megapascals, compared with only 47.49 percent for water-cooled specimens. Air cooling fell in between at 53.44 percent, while carbon dioxide and sodium bicarbonate cooling produced nearly identical intermediate results of roughly 67 percent. The spread between the best and worst coolant widened dramatically with temperature: a difference of just 7.6 percent at 200 degrees grew to 25.86 percent at 800 degrees.

The physics behind these differences is subtle. Water cooling, it turns out, is doubly damaging. When cold water strikes hot, low-permeability concrete, the steep thermal gradient generates tensile stresses that crack the surface, while simultaneously reintroducing moisture into a dehydrated pore network where vapour pressure drives further microcracking and spalling. Carbon dioxide cooling, by contrast, is cryogenic — the gas leaves the extinguisher at roughly minus 78 degrees Celsius — yet it outperformed water. The researchers attribute this to its dryness and speed: the hot surface reaches equilibrium with the gas quickly, shortening the cumulative time the concrete spends under differential thermal stress, and because no liquid is involved, there is no vapour-pressure damage pathway. Thermal damage, they note, is governed by both the magnitude of the temperature difference and how long that difference persists, not by the peak gradient alone.

Sodium bicarbonate powder cooling works through an entirely different, chemically active mechanism. When the powder hits the hot surface, it decomposes endothermically between about 87 and 177 degrees Celsius, breaking down into sodium carbonate, water vapour and carbon dioxide while absorbing heat. The team observed a visible white-grey residue on treated specimens, confirming the reaction had occurred. The resulting porous sodium carbonate layer acts as insulation, buffering the surface against temperature gradients, while the released gases provide additional cooling. Foam cooling, the top performer, relies on a gentler principle still: the air trapped in the thin liquid films of its bubble structure lowers the effective thermal conductivity of the layer, so heat is extracted smoothly rather than violently. Although foam’s average cooling rate was broadly comparable to water’s, the rate of change of the surface-to-core temperature difference was far smaller, and it is precisely that rate of change — not total heat extraction — that governs thermally induced cracking.

Non-destructive testing told the same story. Ultrasonic pulse velocity values, which drop as internal cracks and voids multiply, fell from an excellent 4.78 to 5.56 kilometres per second at 200 degrees to a doubtful 0.24 to 1.54 kilometres per second at 600 and 800 degrees. Water-cooled specimens recorded the lowest velocities at every temperature above 200 degrees, while foam and sodium-bicarbonate-cooled specimens scored highest. Regression analysis revealed a strong relationship between pulse velocity and compressive strength, with coefficients of determination between 0.93 and 0.98 across all cooling regimes — meaning engineers can use a simple ultrasonic survey to estimate residual strength in a fire-damaged structure with considerable confidence.

Statistical analysis reinforced the hierarchy of effects. A two-way analysis of variance showed that exposure temperature was overwhelmingly the dominant driver of strength loss, with partial eta-squared values approaching one for flexural strength, but cooling method remained a highly significant secondary factor. Tukey post-hoc comparisons confirmed that foam cooling significantly outperformed air cooling for all three strength measures, while the largest single drop in flexural strength occurred between 200 and 400 degrees — the range in which calcium silicate hydrate gel begins to dehydrate and microcracks initiate. Water cooling also introduced the greatest variability between specimens, with coefficients of variation exceeding 15 percent at the highest temperatures, a statistical fingerprint of the chaotic, non-uniform nature of thermal shock.

Physical inspection and microscopy made the damage visible. After exposure to 800 degrees, water-cooled specimens showed the widest thermal cracks, measuring up to 0.75 millimetres, followed by air-cooled specimens at 0.62 millimetres. Foam-cooled specimens showed the narrowest cracks at just 0.16 millimetres, with carbon dioxide and sodium bicarbonate in between. Scanning electron microscopy revealed severe networks of cracks and degraded interfacial transition zones in water-cooled samples, whereas foam-cooled samples retained a relatively dense calcium-silicate-hydrate structure with good bonding between aggregate and paste. Energy-dispersive X-ray analysis showed a pronounced depletion of calcium and oxygen peaks at 800 degrees, most severe in water-cooled specimens, consistent with the breakdown of calcium-bearing hydration products.

Even the aggregates inside the concrete suffered differently depending on the coolant. Recovered coarse aggregates from water-cooled specimens showed the greatest deterioration in specific gravity, water absorption and crushing value — at 800 degrees their water absorption was 4.85 percent, nearly double the 2.48 percent measured for foam-cooled counterparts — indicating that quenching damage penetrates deep into the mineral structure of the material. Mass loss followed the same pattern, with water-cooled specimens shedding 10.28 percent of their mass at 800 degrees against 6.06 percent for foam-cooled ones, reflecting spalling and the decomposition of hydration products.

The practical implications are significant. Flexural and tensile strengths proved more sensitive to cooling damage than compressive strength, since even small microcracks disproportionately degrade tension and bending performance — water-cooled beams retained only 15.33 percent of their flexural strength after 800-degree exposure, versus 29.32 percent for foam-cooled ones. Because the relative performance of coolants diverged most sharply at high temperatures, the choice of suppression agent matters most in severe fires. The authors suggest that for structures likely to survive a fire and be repaired rather than demolished, suppression strategies that moderate the cooling rate — such as foam — could preserve substantially more structural value, while ultrasonic pulse velocity surveys paired with the regression models developed here give assessors a fast, reliable tool for triage. The team cautions that continuous temperature instrumentation during cooling was not performed in this study, and recommends future work with embedded thermocouples and controlled flow rates to quantify cooling kinetics precisely.

Subject of Research: Effect of fire suppression and cooling methods on the residual mechanical properties of heat-damaged structural concrete

Article Title: Post-fire behaviour of structural concrete: Effect of different fire suppression methods on residual strength and damage recovery

Article References: Thanaraj, D. P., G, A., Anand, N., A, D. A., & Eszter, L. E. (2026). Post-fire behaviour of structural concrete: Effect of different fire suppression methods on residual strength and damage recovery. Case Studies in Construction Materials, 25, Article e06560. https://doi.org/10.1016/j.cscm.2026.e06560

Image Credits: AI Generated

DOI: 10.1016/j.cscm.2026.e06560

Keywords: concrete, fire, residual strength, cooling methods, foam, thermal shock, ultrasonic pulse velocity, compressive strength, fire suppression, structural assessment, microcracking, sodium bicarbonate