Recycled-aggregate thermal-insulating concrete (RATIC) is a functional recycled concrete that provides both thermal insulation and structural performance [1][2]. RATIC contains four aggregate types: glazed hollow beads (GHB) as thermal-insulating aggregate, natural coarse aggregate (NCA), natural fine aggregate (FA), and recycled coarse aggregate (RCA). RCA is produced by crushing and screening ordinary concrete. The adhered old cement mortar and associated pores and microcracks give RCA poorer physical properties than NCA, particularly higher water absorption and a higher crushing index [3][4]. These characteristics can limit the mechanical performance and durability of recycled concrete. Recent experiments also show that deterioration under combined chloride exposure and freeze-thaw cycling depends on aggregate replacement and exposure duration [5], emphasizing the need to evaluate durability under the relevant service conditions.

Several approaches have been investigated to improve recycled concrete durability, including nano-silica modification [6], CO₂ treatment of recycled aggregate [7][8], prewetting and water-control strategies [9], improved mixing procedures [10], and silane-based aggregate pretreatment [11]. Bao et al. [6] reported that nano-silica refined the pore structure and improved salt-frost resistance under their tested conditions. Peng et al. [7] found that CO₂ curing of recycled aggregate mitigated deterioration during NaCl freeze-thaw exposure and produced a more compact aggregate-matrix interface. Sapata et al. [8] examined carbonated fine and coarse RCA and observed lower aggregate water absorption and improved mechanical and water transport properties in conventional concrete; those results did not assess salt freeze-thaw resistance. Prewetting benefits depend on initial saturation and the basis used to control water content [9], while silane surface treatment has improved water impermeability and freeze-thaw resistance in RAC [11]. The effects of these treatments therefore need to be distinguished by material and exposure conditions. Previous research on RATIC demonstrated substantial damage under combined chloride attack and freeze-thaw cycling [12]. Šeps et al. [13] also reported severe damage, particularly surface scaling, in fiber-reinforced recycled concrete exposed to freezing and deicing salts. These studies motivate a specific assessment of carbonated RCA in thermal-insulating concrete under salt freeze-thaw exposure.

In this study, recycled coarse aggregate was treated in an accelerated carbonation chamber to obtain carbonation-treated recycled coarse aggregate (RCA-C). Its basic properties and the mechanical performance and salt freeze-thaw response of concrete incorporating it were compared with those of untreated RCA and the corresponding concrete.

The cementitious materials used in the RATIC were grade 42.5 ordinary Portland cement and silica fume; their compositions and basic properties are listed in Table 1. The two mixtures contained RCA or RCA-C, natural fine aggregate (FA), and glazed hollow beads (GHB) as thermal-insulating aggregate. Neither mixture contained natural coarse aggregate (NCA), as shown in Table 2. The coarse aggregate particle size ranged from 5 to 20 mm. River sand with a fineness modulus of 2.37 was used as the fine aggregate. The RCA was sourced from Handan, Hebei Province, China.

Table 1. Properties and composition of cementitious materials.

Glazed hollow bead thermal-insulating aggregate is an inorganic mineral material produced by expanding and calcining pretreated perlite ore sand. It has an irregular spherical cavity structure with a honeycomb-like interior. The thermal-insulating concrete used a high-performance polycarboxylate water-reducing admixture with a water-reduction rate of 35% - 40% at a dosage of 3%. The concrete mix proportions are given in Table 2. Before batching, the aggregates used in both mixtures, including RCA and RCA-C, were presoaked and brought to the saturated surface-dry (SSD) condition. No additional water was added to compensate for aggregate absorption. This SSD conditioning was used to limit differences in free mixing water caused by the different absorption capacities of RCA and RCA-C. Both mixtures were proportioned at a water-to-binder ratio of 0.50, with binder comprising 379 kg/m³ of cement and 28 kg/m³ of silica fume.

Table 2.Concrete Mix Proportions (kg/m³).

Before carbonation, the RCA was held at room temperature and had an initial degree of saturation of 50%. The aggregate was then treated for 7 d in the accelerated carbonation chamber shown in Figure 1, at a CO₂ concentration of 20%, standard atmospheric pressure, 25˚C, and 70% relative humidity. At intervals during treatment, a random portion of the aggregate was removed, and the old cement mortar adhering to its surface was ground into powder and sprayed with phenolphthalein solution. The operational treatment endpoint was the absence of a visible color response in the sampled mortar powder (Figure 2). This endpoint indicates reduced alkalinity in the tested material; it does not establish complete carbonation of all adhered mortar or every aggregate particle. RCA-C denotes aggregate treated to this phenolphthalein-based endpoint.

(a) Sampled mortar with no visible phenolphthalein color response; (b) sampled mortar with a visible color response

Mechanical properties were measured in accordance with the Standard for Test Method of Mechanical Properties on Ordinary Concrete [14]. Compressive-strength and splitting-tensile-strength specimens were 100 mm × 100 mm × 100 mm cubes. The specimens were demoulded 24 h after casting and cured at 20˚C and 95% relative humidity until testing at an age of 28 d. Three specimens per mixture were tested for each mechanical property, and the arithmetic mean was reported.

Figure 1.Schematic of the recycled-aggregate carbonation treatment.

Figure 2.Old cement mortar after spraying with phenolphthalein solution.

Salt freeze-thaw specimens were 100 mm × 100 mm × 400 mm prisms. They were demoulded 24 h after casting and cured in a chamber at 20˚C ± 3˚C and 95% relative humidity until an age of 24 d after casting. The prisms were then immersed for 4 d in a 3.5% NaCl solution by mass, and freeze-thaw exposure began at an age of 28 d. Three prisms were tested for each mixture.

Salt freeze-thaw testing, including mass-loss and relative-dynamic-modulus measurements, followed the rapid freezing method specified in GB/T 50082-2009 [15]. A 3.5% NaCl solution by mass was used as the exposure medium. The solution was not renewed during testing; a salinometer was used to monitor and control its salinity. The time-temperature profile is shown in Figure 3. The displayed temperature profile spans approximately 160 min, with the curve labelled Center ranging from approximately −17˚C to +5˚C and the edge and diagonal curves ranging from approximately −24˚C to +22˚C.

Mass loss and relative dynamic modulus were measured at 0, 25, 50, 75, and 100 cycles for RATIC and at 0, 15, 30, 45, 60, 80, and 100 cycles for RATIC-CO₂. The reported durability results are arithmetic means of three specimens per mixture. The 60% relative-dynamic-modulus value is used as a comparison benchmark in this study; measurements continued to 100 cycles, including after the reported modulus had fallen below this benchmark. Between-specimen variability is not quantified, so the statistical uncertainty of the reported means remains undetermined.

Figure 3. Temperature-control curve for freeze-thaw cycles.

Under the conditions tested, the mean compressive and splitting tensile strengths of concrete incorporating RCA-C were 21.2% and 7.1% higher, respectively, than those of RATIC. The tensile-strength contrast remains descriptive because between-specimen variability has not been quantified.Concrete incorporating RCA-C retained a higher relative dynamic modulus during the laboratory salt freeze-thaw test. After 100 cycles, the plotted values were approximately 53% for RATIC-CO₂ and 23% for RATIC. These results characterize the tested exposure and do not quantify field service life.After 100 salt freeze-thaw cycles, the mean mass-loss rates were approximately 3.6% for RATIC and 3.25% for RATIC-CO₂. Without the corresponding replicate variability, these results do not establish that carbonation had no appreciable effect on surface scaling.