The steel slag utilized in this study was sourced from Shandong Laiwu Steel Group Co., Ltd. in Shandong Province, China. Its appearance is shown in Figure 1. It was derived from a typical converter steelmaking process, with initial particle sizes ranging from 4.75 mm to 16 mm. Before pretreatment, the steel slag underwent preliminary crushing and sieving to eliminate impurities. Its fundamental physical and chemical properties were then assessed in accordance with relevant ASTM standards. The chemical composition was analyzed using X-ray fluorescence spectroscopy (XRF), while the content of f-CaO was determined using the ethylene glycol method (ASTM C114) (ASTM C114, 2014). Additionally, the physical properties of steel slag aggregates, including apparent specific gravity, water absorption, and crushing value, were evaluated. Key properties of the raw steel slag are summarized in Tables 1, 2.

The results indicated that the f-CaO content in the steel slag sample was 3.40%. According to the Chinese specification (JT/T 1086-2016) (JT/T 1086-2016, 2016), the f-CaO content in aggregates intended for pavement use should not exceed 3.0%. This clearly demonstrates that the steel slag material fails to meet the required standards. If utilized directly in pavement construction, it may result in significant quality issues, including volume expansion and cracking, thereby posing considerable engineering risks. Therefore, it is imperative to conduct pretreatment of the steel slag aggregates to reduce their f-CaO content, thereby enhancing the water stability of the steel slag asphalt mixture.

Styrene-butadiene-styrene (SBS) modified asphalt of the I-D type was utilized as the binder, in accordance with ASTM D6114 specifications (ASTM D6114, 2019). Its fundamental physical properties are presented in Table 3. The test results confirmed that all properties satisfy the requirements for high-performance asphalt pavements.

Limestone was utilized as the fine aggregate (0–4.75 mm) and mineral filler (<0.075 mm) to maintain consistency with conventional asphalt mixture design. The apparent specific gravity of the limestone fine aggregate and mineral filler was determined to be 2.727 and 2.700, respectively, with a water absorption of 0.3%. All properties of the limestone aggregates comply with ASTM D1073 (ASTM D1073-16, 2016) specifications for highway aggregates.

Three typical pretreatment methods were employed to enhance the performance of steel slag, with detailed parameters optimized based on preliminary experiments and relevant literature (Sun et al., 2023; Gan et al., 2022b).

The natural aging process involved the outdoor stockpiling of steel slag, utilizing atmospheric temperature and humidity conditions to promote the hydration reactions of its active components. This technique is recognized as the mainstream method for treating steel slag waste. The steel slag was openly stacked outdoors under ambient conditions, with temperatures ranging from −22 °C to 42 °C and relative humidity between 48% and 75%. To ensure uniform exposure to atmospheric moisture and CO₂, the steel slag was periodically turned every 2 weeks during the aging process. Samples were collected after 12 months for subsequent testing.

The water bath aging technique lies in promoting the reaction between f-CaO and water to generate calcium hydroxide. A systematic study was conducted to investigate the effects of water bath temperature and curing duration on the volume stability of steel slag. The steel slag was submerged in distilled water at temperature gradients of 30 °C, 60 °C, and 90 °C, controlled by a thermostatic water bath, over curing periods of 7 days. After aging, the steel slag was air-dried at room temperature to a constant weight to remove surface moisture before further processing.

Carbonation treatment was performed in a sealed high-temperature and high-pressure reactor. Before this process, the steel slag was pre-soaked in deionized water for 8 h to enhance pore accessibility, after which it was placed in the reactor. A review of the literature guided the optimization of carbonation parameters, which included a temperature of 600 °C, a pressure of 0.5 MPa, a CO₂ flow rate of 9 L/h, a water vapor flow rate of 9 L/h, and a reaction duration of 2.5 h. Research conducted by Ma, Mao, Zhu and Yao (Ma et al., 2025) indicated that this carbonation treatment effectively produced optimal performance in the steel slag. After the carbonation process, the steel slag was dried at 105 °C for 24 h until a constant weight was achieved.

Notably, discrepancies exist between controlled laboratory conditions and real-world field environments in engineering practice. Therefore, the present experimental findings should be regarded primarily as indicative guidance rather than direct predictors of field performance. A 12-month period of natural weathering simulated long-term atmospheric exposure, comprehensively accounting for the effects of temperature, humidity, and CO₂ levels. However, water bath aging at 30 °C, 60 °C, and 90 °C, as well as accelerated carbonation, may be influenced by factors such as temperature fluctuations, water purity, CO₂ concentration, energy consumption, and associated costs in practical engineering applications. Consequently, challenges arise regarding the repeatability and reproducibility of these experimental results. Future field trials are recommended to validate long-term durability under diverse climatic and traffic conditions.

The AC-13 dense-graded asphalt mixture, commonly used for pavement layers, was designed with a synthetic gradation that aligns with the median gradation specifications of AC-13 (JTG F40-2004, 2004). The gradation was controlled across a range of sieve sizes from 0.075 mm to 16 mm, and the detailed gradation curve is illustrated in Figure 2.

In this study, steel slag was utilized as a complete replacement for coarse aggregates larger than 4.75 mm. Due to the higher density of steel slag compared to limestone aggregates, a volume design method based on density conversion was employed to ensure that the overall volume of the specimens remained unchanged after the aggregate substitution. To determine the optimum asphalt content (OAC), the Marshall mix design method (ASTM D6926) was applied (ASTM D6926, 2020). The OAC was established at 5.4%, which was determined based on maximizing stability, minimizing voids in mineral aggregate (VMA), and achieving target air voids of 4%.

To comprehensively evaluate the performance of pretreated steel slag and corresponding asphalt mixtures, a series of systematic tests were conducted, covering the volume stability of steel slag aggregates, water stability and abrasion performance, of asphalt mixtures.

The volume stability of steel slag was evaluated by measuring the water immersion expansion rate at 80 °C. The results corresponding to the various pretreatment methods are presented in Figure 4.

The raw steel slag exhibited poor volume stability, exhibiting an immersion expansion rate of 2.26% after 10 days. This poses significant risks of pavement cracking and upheaval if used directly. Natural aging effectively mitigated this issue, reducing the 10-day expansion rate to 1.52%. The 12-month aging process consumed most of the reactive f-CaO, thereby minimizing the potential for residual expansion. Water bath aging further enhanced volume stability, with improvements observed at increasing temperatures and curing durations. At 30 °C for 7 days, the 10-day expansion rate was 2.17%, representing a 3.98% reduction compared to the raw steel slag. At 60 °C for 7 days, the expansion rate decreased to 1.62%, reflecting a 28.32% reduction. Finally, at 90 °C for 7 days, the rate dropped to 1.15%, a significant reduction of 49.11%. This improvement is attributed to thermal activation, which accelerated the complete hydration of f-CaO, converting it into stable Ca(OH)₂, resulting in negligible subsequent expansion. Carbonation treatment achieved optimal volume stability, with a 10-day immersion expansion rate of only 0.76%, indicating a 66.37% reduction. The chemical conversion of f-CaO to CaCO₃ eliminated the primary source of expansion, while the dense CaCO₃ layer acted as a physical barrier to prevent moisture penetration and subsequent hydration of internal components. This finding is consistent with the volume stability evaluation by Sun, Luo, Wang, Dong and Zhang (Sun et al., 2023), which emphasized that converting reactive oxides into stable carbonates is the most effective method for inhibiting steel slag expansion. The significant enhancement in volume stability effectively increases the applicability of carbonated steel slag in actual engineering projects (Pang et al., 2016). However, the large-scale implementation of carbonation treatment faces challenges in two primary areas: energy consumption and economic cost. Regarding energy consumption, accelerated carbonation typically necessitates heating, pressurization, CO₂ compression, and material pretreatment (such as grinding and stirring), all of which substantially elevate electricity and thermal energy demands (Liu et al., 2026). Furthermore, compared to basalt, naturally aged steel slag can reduce total construction costs by 14.5%, whereas carbonated steel slag results in an approximate 1.7% increase in pavement construction expenses (Wang et al., 2025).

The evolution of the chemical composition of steel slag following pretreatment primarily involves the reduction of reactive f-CaO, which is a critical factor influencing volume instability. As illustrated in Figure 5, all three pretreatment methods successfully decreased the f-CaO content; however, there were notable differences in their efficiency.

After 12 months of natural aging, the f-CaO content decreased to 1.85%, significantly below the specification threshold. This reduction is attributed to the slow yet continuous hydration of f-CaO through exposure to atmospheric moisture and CO₂ over the extended period, which gradually consumes reactive components and repairs minor structural defects. Water bath aging demonstrated a strong temperature dependence: at 30 °C for 7 days, the f-CaO content only decreased to 2.91%, representing a 14.41% reduction compared to raw steel slag, indicating insufficient thermal activation of the hydration reactions. Increasing the temperature to 60 °C for 7 days further reduced the f-CaO content to 2.03%, a reduction of 40.29%. At 90 °C for 7 days, the f-CaO content plummeted to 1.42%, reflecting a 58.23% reduction. This observation aligns with the findings of Gan, Li, Zou, Wang and Yu (Gan et al., 2022b), which indicated that the reaction rate of steel slag aggregates at elevated temperatures is more than twice that of ambient conditions. The thermal energy enhances the diffusion of water into the pores of the steel slag and promotes the reaction between f-CaO and water. Carbonation treatment achieved the most significant reduction in f-CaO, with the content decreasing to 1.07%, indicating a 68.53% reduction. The high-temperature and high-pressure environment facilitates the rapid reaction between f-CaO and CO₂, converting unstable calcium oxide into thermodynamically stable CaCO₃ (Wang et al., 2023). This was further corroborated by XRF analysis (Santos et al., 2013), which showed no significant changes in other major components, indicating that the pretreatments specifically target reactive f-CaO without altering the inherent mechanical properties of the steel slag.

The water stability of asphalt mixtures, assessed through the residual stability ratio and the TSR, is a vital indicator of pavement durability in humid or freeze-thaw environments. Building on the previous analysis of volume stability, Figure 6 presents the water stability results of steel slag asphalt mixtures subjected to the three pretreatment methods under optimal conditions.

The raw steel slag mixture exhibited the lowest water stability, with a residual stability ratio of 78.33% and a TSR of 74.08%. This is attributed to the hydration of unreacted f-CaO in the steel slag, which leads to internal cracking and weakens the interfacial adhesion between the aggregate and SBS asphalt. In contrast, all pretreated steel slag mixtures demonstrated significant improvements in water stability, with performance rankings consistent across both tests: carbonation treatment > water bath aging (90 °C for 7 days) > natural aging > raw steel slag. The carbonated steel slag mixture achieved the highest residual stability ratio of 92.07% and a TSR of 88.67%. The dense CaCO₃ layer formed on the steel slag not only inhibits hydration expansion but also enhances interfacial compatibility with SBS asphalt, thereby reducing water-induced debonding. The water bath aged mixture (90 °C for 7 days) exhibited strong water stability, with a residual stability ratio of 89.23% and a TSR of 83.59%. This improvement is due to the reduced f-CaO content and the compact surface layer, which minimize water penetration and freeze-thaw damage (Gan et al., 2022a). The naturally aged mixture displayed moderate water stability, with a residual stability ratio of 83.45% and a TSR of 80.45%. While this level of stability is adequate for most moderate-humidity regions, it is inferior to the other two methods due to the presence of residual reactive components and a relatively porous surface structure. These results confirm that pretreatments enhance water stability by reducing f-CaO content and optimizing aggregate surface properties, with carbonation treatment providing dual benefits through chemical stabilization and the formation of a physical barrier.

The abrasion performance of asphalt mixtures is closely linked to pavement skid resistance and overall service life. This performance is evaluated through the Los Angeles abrasion test to assess aggregate wear resistance, 2D angularity analysis to evaluate aggregate shape retention, and the BPN to measure surface friction.

As illustrated in Figure 7, the results of the Los Angeles abrasion test revealed that the raw steel slag had an abrasion value of 11.6%. Pretreatments significantly enhanced wear resistance, with natural aging reducing the abrasion value to 9.9%, representing a 14.7% reduction. Water bath aging at 90 °C for 7 days further decreased the value to 9.7%, reflecting a 16.4% reduction, while carbonation treatment resulted in an abrasion value of 8.8%, indicating a 24.1% reduction. These improvements can be attributed to enhanced aggregate hardness and structural integrity. The carbonation treatment creates a rigid CaCO₃ layer on the surface of the steel slag, while water bath aging increases particle compactness through hydration products, and natural aging helps to reduce internal defects.

The 2D angularity parameters of steel slag, assessed before and after the Los Angeles abrasion test, further elucidated the effects of pretreatments on shape retention, as presented in Table 4.

Raw steel slag exhibited significant angularity degradation after abrasion: convexity decreased from 0.90 to 0.74, representing an 17.8% reduction; roundness increased from 0.72 to 0.86, reflecting a 19.4% increase; and the axis ratio dropped from 1.35 to 1.20, indicating an 11.1% reduction. These changes suggest severe blunting of sharp edges and a decrease in elongation, as mechanical wear smoothed particle contours and diminished the differences between the major and minor axes. Natural aging and water bath aging effectively mitigated this degradation. Natural-aged steel slag exhibited a convexity reduction of 15.7%, decreasing from 0.89 to 0.75; roundness increased by 12.0%, rising from 0.72 to 0.81; and the axis ratio showed a decrease of 6.1%, dropping from 1.32 to 1.24. Water bath aged steel slag, treated at 90 °C for 7 days, displayed more moderate changes, with convexity reducing by 12.5%, roundness increasing by 10.3%, and the axis ratio decreasing by 3.8%. Carbonation treatment demonstrated the best retention of angularity across all three parameters, with convexity decreasing by only 10.8%, roundness increasing by 7.5%, and axis ratio declining by merely 3.0%. The minimal decay in axis ratio for carbonated steel slag ensures the preservation of particle elongation, which is advantageous for interlocking between aggregates and maintaining long-term skid resistance. This aligns with the findings of Hu, Peng, Wang, Zou, Tao, Liu and Yu (Hu et al., 2025), who emphasized that angularity retention is critical for the long-term skid resistance of steel slag asphalt mixtures.

The BPN results, illustrated in Figure 8, exhibited a similar trend to that of 2D angularity retention. The raw steel slag mixture exhibited an initial BPN of 55, which decreased to 38 after 80 h of accelerated wear, reflecting a 30.9% attenuation. The natural-aged mixture had an initial BPN of 56, which declined to 46, resulting in a 17.9% attenuation. The water bath aged mixture, treated at 90 °C for 7 days, displayed an initial BPN of 57 and a final BPN of 45, indicating a 21.1% attenuation. In contrast, the carbonated steel slag mixture recorded the highest initial BPN of 59 and the lowest attenuation rate of 10.2%, retaining a final BPN of 53. This enhanced performance can be attributed to the retained angularity of carbonated steel slag, which creates a rough surface texture that improves tire-pavement friction. Furthermore, the rigid CaCO₃ layer formed during carbonation effectively resists wear-induced smoothing. Consequently, the angularity and wear resistance of steel slag are critical factors in improving the skid resistance of asphalt mixtures, with carbonation treatment achieving optimal synergy between these two properties. However, the BPN value should be interpreted with caution, as it was derived under controlled indoor conditions that primarily reflect tire-induced polishing. In reality, skid resistance degradation is influenced by a wider range of factors, including seasonal climate variations (such as rainfall, snowfall, and freeze–thaw cycles), traffic composition, pavement age, and surface contamination. The attenuation rate observed in the laboratory may overestimate long-term field decay in various climatic regions. Therefore, full-scale field validation is essential to establish reliable life-cycle skid resistance performance.

It is important to note that, in order to maximize resource utilization, this study implemented complete replacement of coarse aggregate (≥4.75 mm) with steel slag, which may influence the overall load-bearing skeleton structure of the mixture. The research findings indicate that, compared to raw steel slag, the steel slag optimized through the three pretreatment methods exhibited higher angularity, lower Los Angeles abrasion values, and higher and more durable British Pendulum Number (BPN) values, with carbonated steel slag demonstrating the most favorable results. These properties contribute to the formation of a stronger interlocking skeleton, reducing deformation and maintaining long-term structural performance. Such improvements suggest that steel slag positively contributes to the stability of the overall load-bearing skeleton rather than introducing adverse effects. However, differences in particle density, shape distribution, and gradation packing behavior may still result in subtle changes in stress transmission under repeated heavy loading. Consequently, while the laboratory results are promising, it remains essential to verify the durability of the coarse skeleton structure under actual pavement service conditions through long-term field trials.

Furthermore, the composition of steel slag varies significantly among steel mills due to differences in raw materials, production processes, and cooling methods, particularly regarding the contents of f-CaO and f-MgO. This variability can impact the effectiveness of pretreatment methods. Natural weathering and water bath aging depend on hydration kinetics, which are slower for slags with lower reactivity, whereas carbonation is more effective for high f-CaO slags due to enhanced carbonate formation. In this study, the selected steel slag exhibited a moderate f-CaO content (typical of converter slag), and carbonation proved to be the most efficient method for suppressing expansion. To enhance the generalizability of the conclusions, future research should evaluate the effectiveness of pretreatment methods across a range of steel slag sources with varying compositions. This would aid in establishing guidelines for site-specific pretreatment optimization and broaden the applicability of carbonated steel slag in asphalt mixtures.