Civil engineering researchers have demonstrated a multi-response statistical optimization approach that balances the mechanical strength, fresh workability, and cradle-to-gate embodied energy of ultra-high-performance concrete. Investigating a quaternary binder framework composed of ordinary Portland cement, ground granulated blast-furnace slag, microsilica, and fly ash, the investigation established that supplementary cementitious materials can replace substantial fractions of clinker without sacrificing structural capability. By applying a Box–Behnken experimental design coupled with response surface methodology, the study revealed how specific chemical reactions and granular packing mechanisms govern both physical densification and environmental burdens. Across the fourteen experimental runs, measured compressive strengths spanned from 117 to 132 megapascals, flexural strengths ranged between 25 and 30 megapascals, flow diameters varied from 220 to 238 millimeters, and cradle-to-gate embodied energy ranged between 6210.09 and 7144.74 megajoules per cubic meter of concrete.
Ultra-high-performance concrete represents an advanced class of cementitious composite distinguished by remarkable durability and compressive strength exceeding 120 megapascals alongside tensile strength over 6 megapascals under conventional formulations. These mechanical thresholds typically depend on a minimal water-to-binder ratio below 0.25 and dense particle packing that eliminates internal microdefects. However, conventional formulations typically consume two to three times more cement than standard concrete mixes, leading to high initial costs and elevated carbon dioxide emissions associated with clinker calcination and thermal processing. To mitigate these environmental impacts, the research team examined whether industrial by-products such as blast-furnace slag and fly ash could be systematically proportioned alongside ultrafine silica fume. The total binder content was held constant at 1200 kilograms per cubic meter, while water, brass-coated micro steel fibers, superplasticizer, and shrinkage-reducing admixture dosages were fixed across Phase I mixture trials.
The study utilized a 53-grade ordinary Portland cement, Class F fly ash, ground granulated blast-furnace slag, and condensed microsilica, examining their physical properties, chemical oxide proportions, and morphological features. Chemical characterization revealed that ordinary Portland cement contained 62.1 percent calcium oxide and 18.1 percent silicon dioxide, while microsilica possessed 95.0 percent silicon dioxide with a Blaine fineness of 22,000 square meters per kilogram. Laser diffraction and field-emission scanning electron microscopy highlighted sharp contrast among the constituent particles. Fly ash exhibited predominantly smooth, spherical particles ranging between 1 and 20 micrometers that served as micro-ball bearings to reduce friction and water demand. In contrast, slag and cement exhibited irregular, angular geometries providing active nucleation sites, whereas nanoscale microsilica formed agglomerated clusters capable of filling minute interstitial voids between larger cementitious grains and accelerating pozzolanic reactions.
The experimental matrix was established using a three-factor Box–Behnken design treating cement, slag, and microsilica as continuous numerical variables. Cement was varied between 50 and 60 percent of total binder, ground granulated blast-furnace slag between 0 and 20 percent, and microsilica between 15 and 25 percent. Fly ash functioned as a balancing binder constituent calculated as the remainder of the total binder fraction. Fine sand with particle sizes up to 2 millimeters, a specific gravity of 2.67, and a bulk density of 1780 kilograms per cubic meter was adjusted between approximately 755 and 856 kilograms per cubic meter to maintain an exact unit volume of one cubic meter. Brass-coated micro steel fibers measuring 13 millimeters in length and 0.22 millimeters in diameter, possessing a tensile strength of 2300 megapascals, were added at a fixed dosage of 118 kilograms per cubic meter, corresponding to approximately 1.5 percent of total mixture volume.
To rigorously assess environmental performance, the investigators conducted a cradle-to-gate embodied energy assessment conforming to ISO 14040, ISO 14044, and EN 15804 standards. The defined system boundary comprised product-stage modules A1 through A3, accounting for raw material extraction, constituent processing, and plant-level mixing electricity. Embodied energy coefficients were established from literature benchmarks, allocating 4.62 megajoules per kilogram to cement, 1.60 megajoules per kilogram to ground granulated blast-furnace slag, 0.10 megajoules per kilogram to microsilica, and 0.033 megajoules per kilogram to fly ash. Steel fibers represented the most energy-intensive component at 24.60 megajoules per kilogram, while chemical admixtures contributed 12.60 megajoules per kilogram and mixing electricity required 3.60 megajoules per kilowatt-hour. Downstream transport, construction placement, maintenance, and end-of-life phases were excluded due to the lack of uniform regional transport coefficients.
Analysis of variance demonstrated distinct physical mechanisms governing each material response. Compressive strength followed a significant linear model where cement contributed 54.0 percent and microsilica contributed 43.9 percent of the observed variance, while slag contributed an insignificant 2.1 percent due to slow latent hydraulic reactivity at early testing ages. Flexural strength similarly followed a linear model governed primarily by cement at 51.5 percent and microsilica at 27.9 percent, which enhanced interfacial transition zone bonding with steel fibers. In contrast, fresh workability measured via a mini slump cone followed a quadratic model with an adjusted coefficient of determination of 0.9604. Slag content dominated mixture flowability, explaining 73.0 percent of total variance because its glassy surface morphology reduced inter-particle friction up to a 10 percent replacement level, beyond which higher replacement decreased cohesiveness and microsilica additions above 20 percent markedly increased paste viscosity.
Mixture density and embodied energy were both accurately represented by linear regression models with high statistical reliability. Density yielded an adjusted coefficient of determination of 0.91, reflecting a direct relationship with the combined proportions of dense mineral powders and the systematic reduction of entrapped air voids. For embodied energy, the linear model achieved an adjusted coefficient of determination of 0.9666 and a predicted coefficient of determination of 0.9443, with ground granulated blast-furnace slag accounting for 60.7 percent and cement accounting for 37.4 percent of the model variance. Because fly ash served as the balancing component, increasing slag content proportionally decreased the ultra-low-energy fly ash fraction, demonstrating that regression coefficients reflect conditional binder substitutions within the constrained quaternary system rather than isolated material properties.
Multi-criteria optimization using a desirability ramp function identified an optimal binder proportion composed of 56.80 percent cement, 7.32 percent ground granulated blast-furnace slag, 20.07 percent microsilica, and 15.81 percent fly ash by weight, achieving a composite desirability score of 0.62. In experimental validation tests, this optimized mixture achieved an experimental compressive strength of 131.50 megapascals compared to a predicted 128.04 megapascals, representing an error of 2.70 percent. The experimental flexural strength reached 28.60 megapascals against a predicted 28.78 megapascals, exhibiting a 0.60 percent error. Fresh flow reached 220.00 millimeters compared to a predicted 223.32 millimeters, while measured dry density was 2364.00 kilograms per cubic meter and cradle-to-gate embodied energy was 6765.00 megajoules per cubic meter, matching model predictions within an absolute error of less than 3 percent.
In the secondary phase of the study, the structural stiffness, tensile strain capacity, and microstructure of five optimized formulations were evaluated in detail. Modulus of elasticity testing on cylindrical specimens yielded values between 45,000 and 60,000 megapascals, aligning closely with predictive equations from the American Concrete Institute building code. Direct tensile testing on dogbone specimens demonstrated peak tensile stresses ranging from 8.00 to 8.47 megapascals for fly ash-dominant mixes and 10.88 to 10.92 megapascals for slag-rich formulations. Although higher slag proportions increased peak tensile resistance, mixtures containing higher fly ash fractions demonstrated superior post-cracking ductility and pronounced strain-hardening behavior, facilitated by the controlled pull-out resistance and crack-bridging mechanics of the randomly distributed steel fibers.
Microstructural and phase examinations confirmed the nanoscale mechanisms supporting these macroscale properties. Field-emission scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy indicated that optimized mixtures formed a dense, homogeneous calcium-alumino-silicate-hydrate gel matrix with minimal unreacted portlandite. Diffractograms obtained through X-ray diffraction revealed dominant quartz reflections alongside broad amorphous halos corresponding to hydrated binding gels, while expansive ettringite peaks were completely absent. The investigators noted that their empirical response surface models remain conditional to the tested binder ranges and water-to-binder ratio of approximately 0.18, emphasizing that future research must incorporate broad durability assessments such as chloride permeability and freeze-thaw resistance alongside full cradle-to-grave life-cycle impact evaluations.
Subject of Research: Multi-response optimization of quaternary binder systems for ultra-high-performance concrete
Article Title: Influence of binder chemistry and mix optimization on the mechanical and environmental performance of ultra-high-performance concrete
Article References: Kothari, V., Dave, U., & Maaze, M. R. (2026). Influence of binder chemistry and mix optimization on the mechanical and environmental performance of ultra-high-performance concrete. Results in Chemistry, 32, Article 104003. https://doi.org/10.1016/j.rechem.2026.104003
Image Credits: AI Generated
DOI: 10.1016/j.rechem.2026.104003
Keywords: Ultra-High-Performance Concrete, Response Surface Methodology, Binder Chemistry, Supplementary Cementitious Materials, Embodied Energy, Compressive Strength, Microstructure, Materials Engineering, Influence, binder, chemistry, optimization
News Source: Bethany Barker. (November 7, 2026). Multi-Response Optimization Balances Strength and Embodied Energy in Ultra-High-Performance Concrete. Scienmag.



