Frost Resistance of Fully Recycled Coarse Aggregate Concrete in Saline-Soil Regions: Seasonal Freezing

Abstract

With global sustainable construction growth, fully recycled coarse aggregate concrete (RCAC)—eco-friendly for cutting construction waste and reducing natural aggregate over-exploitation—has poor durability in seasonally freezing saline-soil regions (e.g., Tumushuke, Xinjiang): freeze-thaw and salt ions (NaCl, Na₂SO₄) cause microcracking, faster performance decline, and shorter service life, limiting its use and requiring better salt freeze resistance. To address this, a field survey of Tumushuke’s saline soil was first conducted to determine local salt type and concentration, based on which a matching 12% NaCl + 4% Na₂SO₄ mixed salt solution was prepared. RCAC specimens modified with fly ash (FA), silica fume (SF), and polypropylene fiber (PPF) were then fabricated, cured under standard conditions (20 ± 2 °C, ≥95% relative humidity), and subjected to rapid freeze-thaw cycling in the salt solution. Multiple macro-performance and microstructural indicators (appearance, mass loss, relative dynamic elastic modulus (RDEM), porosity, microcracks, and corrosion products) were measured post-cycling. Results showed the mixed salt solution significantly exacerbated RCAC’s freeze-thaw damage, with degradation severity linked to cycle count and admixture dosage. The RCAC modified with 20% FA and 0.9% PPF exhibited optimal salt freeze resistance: after 125 cycles, its RDEM retention reached 75.98% (6.60% higher than the control), mass loss was only 0.28% (67.80% lower than the control), and its durability threshold (RDEM > 60%) extended to 200 cycles. Mechanistic analysis revealed two synergistic effects for improved performance: (1) FA optimized pore structure by filling capillaries, reducing space for pore water freezing and salt penetration; (2) PPF enhanced crack resistance by bridging microcracks, suppressing crack initiation/propagation from freeze-thaw expansion and salt crystallization. A “pore optimization–ion blocking–fiber crack resistance” triple synergistic protection model was proposed, which clarifies admixture-modified RCAC’s salt freeze damage mechanism and provides theoretical/technical guidance for its application in extreme seasonally freezing saline-soil environments.

1. Introduction

Global production of construction and demolition waste (CDW) reaches substantial volumes, with concrete waste accounting for 70% of this total. Annual CDW generation varies significantly by country: China produces 2360 million metric tons (Mt), India 700 Mt, and the United States 600 Mt, while the Netherlands, the United Kingdom, Germany, and France report 102 Mt, 138 Mt, 225 Mt, and 240 Mt, respectively (Ahmad Alyaseen, 2023) [1]. Multiple nations have enacted legislation to advance sustainable CDW management practices. Recycled coarse aggregate (RCA) has emerged as a key research focus, though its performance limitations due to adhered mortar residues require mitigation through strategies such as water-cement ratio adjustment and silica fume incorporation. While RCA has demonstrated applicability in both structural and non-structural components, breakthroughs in high-performance applications and standardization systems remain pending (Ahmad Alyaseen, 2024) [2].

Recycled aggregate concrete (RAC) exhibits significant environmental benefits compared to natural aggregate alternatives. Studies indicate RCA utilization reduces greenhouse gas emissions by 65% and energy consumption by 58% relative to virgin aggregates, underscoring its critical role in accelerating low-carbon transitions within civil engineering (Md. Uzzal Hossain, 2016) [3].

According to the Carbon Peak Implementation Plan for Urban and Rural Construction in Xinjiang (2023–2030) [4] issued by the Xinjiang Uygur Autonomous Region Housing and Urban-Rural Development Department, new construction projects must limit construction waste generation to ≤300 tons/10,000 m² and achieve a resource utilization rate of 55%. Research indicates that substituting NCA with RAC can reduce carbon emissions during concrete production by 18–22% (Xiao et al., 2023) [5]. However, durability deficiencies remain the primary bottleneck hindering large-scale RAC application, necessitating urgent solutions.

Regarding RAC durability issues, the academic community has advanced modification technology from empirical regulation to model-driven quantitative optimization, establishing a “model-driven parameter quantification multi-factor synergy” technical framework. Xiao et al. (2011) [6] proposed a model-based quantitative optimization framework that enhances RAC performance through mortar strength matching and residual mortar regulation. Lin et al. (2023) [7] achieved elastic modulus recovery exceeding 90% of ordinary concrete by targeting residual mortar control (coverage 33–50%). For aggregate optimization, non-regular aggregates improve performance through reduced angularity and controlled porosity, while fractal theory-guided grading-shape synergy optimization enables elastic modulus approaching natural aggregate concrete levels (Fu et al., 2025) [8]. These advances provide quantitative support for RAC performance enhancement but lack targeted solutions for interfacial transition zone (ITZ) damage under combined salt freeze-thaw environments. Moreover, existing models focus on single-factor optimization without integrating the framework to establish quantitative relationships among “multi-factors-microstructure-macro-properties”.

Although studies on cementitious systems (e.g., alkali-activated concrete) have revealed modification mechanisms that inform RAC optimization, critical differences require attention: alkali-activated systems typically exclude RCA, while RAC exhibits more complex ITZ structures due to residual mortar, necessitating tailored modification mechanisms. Three key distinctions exist: (1) In degradation mechanisms, silica fume in alkali-activated systems reduces the Na/Si ratio (from 0.83 to 0.34) to suppress alkali leaching by physically filling and densifying gels to reduce ion migration pathways (Saludung et al., 2021; Wetzel et al., 2019) [9,10]; (2) Alkali-activated systems’ “single-composite” environmental evaluation gradients offer reference value but lack direct transferability due to the absence of RCA (Rostami et al., 2025) [11]; (3) For modification performance, Wetzel et al. (2019) [10] and Rostami et al. (2025) [11] confirmed that silica fume and PPF effects depend critically on dosage, but their dosage laws cannot be directly applied to RAC. Alkali-activated system research neither addresses ITZ synergistic repair after RCA incorporation nor extends to combined salt freeze-thaw environments. While providing methodological insights, these systems fail to resolve RAC-specific ITZ issues. Unlike alkali-activated systems’ simple structures, RAC’s ITZ is more vulnerable to combined salt freeze-thaw damage due to residual mortar, with this problem intensifying under freeze-thaw conditions, resulting in significantly greater frost resistance degradation than natural aggregate concrete (NAC), particularly in combined salt freeze-thaw environments.

Sami W. (2009) [12] observed through 100 freeze-thaw cycles that RAC compressive strength loss exceeded NAC by 5–22%, with the gap widening to 30% as RCA replacement increased from 50% to 100%. The primary cause is RAC’s ITZ porosity being 1.8–2.5 times higher than NAC, leading to ice expansion stress concentration and microcrack propagation [12]. This deficiency directly impedes RAC promotion in Northwest China’s high-cold saline soil regions (e.g., Xinjiang, Qinghai), where seasonal freeze-thaw and salt ion erosion synergistically accelerate degradation. Existing modification technologies fail to address this composite damage or establish quantitative relationships between “ITZ microstructure evolution-macro-property degradation”, creating a significant gap with the model-driven framework development trend.

To improve RAC frost resistance, the academic community has developed two approaches: “microstructural densification” and “macro-toughness enhancement”, both with notable limitations and failing to integrate the quantitative advantages of model-driven frameworks: in micro-optimization, Gao (2023) [13] demonstrated that 6% silica fume increases ITZ microhardness to 92% of NAC, but testing was limited to water-based freeze-thaw without considering salt ion erosion of densified structures. Wang et al. (2017) [14] found that fly ash generates low-calcium C-S-H gels to improve interfacial bonding, with these gels adsorbing Cl⁻ to block ion migration. However, single fly ash modification resulted in 12% mass loss after 50 freeze-thaw cycles due to insufficient toughness, failing to prevent cracking. Neither approach achieved quantitative matching between “densification effect-salt-freeze-thaw damage”.

In macro-reinforcement, Wilson Nguyen (2021) [15] reported that steel fibers (1.0–1.5% vol) enhance RAC crack resistance by 15–25%, but corrosion in 2–3.5% Cl⁻ environments causes long-term flexural strength degradation of 13–14% (confirmed by Xiao et al., 2017) [16]. Recent studies (Yang Li, 2024) [17] show aramid fibers further improve ductility, but fiber-mineral admixture synergies remain unexplored. While polypropylene fibers (PPF) resist salt corrosion, a 0.9% dosage only reduces dynamic elastic modulus loss by 18% due to insufficient rigidity to inhibit early ITZ microcracks (Yuan et al., 2024) [18]. Notably, Wang et al. (2017) [14] combined 50% RCA replacement, 20% fly ash, and 0.9% PPF but failed to quantify synergistic ratios in ITZ repair or test under combined salt freeze-thaw conditions. Current research indicates that fly ash’s low-calcium C-S-H gels adsorb Cl⁻ for ion blocking, silica fume refines ITZ porosity through physical filling, and PPF compensates for mineral admixtures’ toughness deficiency to inhibit freeze-thaw crack propagation. Their synergistic interaction may hold the key to resolving composite salt freeze-thaw damage, but the synergistic ratio and mechanism remain undefined.

To clarify research positioning differences,

Table 1. Quantitative comparison of concrete durability research.

Reference	Research Object	Environmental Conditions	Modification Method	Core Results	Limitations
Sami W. et al. (2009) [12]	RAC 100%	100 water-based freeze-thaw	No modification	Strength reduction 10–25% vs. NAC	No modification or salt erosion
Gao (2023) [13]	RAC/100%	Standard curing	Silica fume (3%, 6%, 9%), optimal 6%	28 d strength +9.5%; lowest ITZ porosity 21%	No salt erosion consideration
Wilson et al. (2021) [15]	RCA 29% (vol)	3.5% NaCl + freeze-thaw	0.2% PVA + 0.5% 30 mm steel + 0.8% 60 mm steel (1.5 vol% hybrid)	Crack resistance +15% −25%	Steel fiber corrosion in Cl−; no composite salt freeze-thaw
Rostami et al. (2025) [11]	Alkali-activated slag/No RCA	0.63% HNO3	10% silica fume + 0.3% PP fiber	210 d acid resistance 65 MPa	No RAC or salt freeze-thaw

Notes: Core results reflect key cycle indicators; ITZ porosity measured via MIP.

summarizes key parameters and results from recent related studies:

Table 1 shows that existing studies lack “multi-factor-microstructure-macro-property” model-based associations. For example, Wilson et al. (2021) [15] only verified fiber crack resistance without quantifying relationships with ITZ porosity.

In summary, current RAC modification technologies face two core limitations incompatible with model-driven framework development: (1) lack of synergistic mechanisms, with research limited to single mineral admixture or fiber effects. Even rare synergistic attempts (e.g., Wang et al. [14]) fail to clarify ITZ repair synergies (e.g., how silica fume’s pore refinement interacts with fiber’s crack bridging to suppress AFt expansion), preventing “multi-factor-microstructure-macro-property” quantitative associations; (2) insufficient environmental adaptability, with research focused on water-based or single-salt conditions without specialized composite salt freeze-thaw studies. Composite salt formulations often use equal concentrations, conflicting with Northwest China’s actual saline soil environment featuring unequal Cl⁻/SO₄²⁻ concentrations (e.g., Tumushuke, Xinjiang: NaCl 6.02%, Na₂SO₄ 7.04% average) [19]. This causes significant discrepancies between experimental results and engineering realities. Additionally, the “silica fume pore optimization + PPF deformation control” quantitative findings from alkali-activated systems (Wetzel et al., 2019; Rostami et al., 2025) [10,11] remain unapplied to RAC modification.

These limitations directly result in the paradox of “theoretically feasible but practically unusable” RAC applications in Northwest saline soil regions.

Engineering-wise, the lack of synergistic modification mechanisms makes single modification technologies inadequate for durability demands in Northwest saline soil projects. For example, fly ash-silica fume dual-doped concrete still experiences 25% strength loss after 50 composite salt freeze-thaw cycles, showing limited frost resistance improvement (Yao, 2018) [20]. Single additives cannot withstand multi-factor coupling damage from salt ions and freeze-thaw cycles. Theoretically, the ambiguity of synergistic mechanisms leaves modification dosage ratios without quantitative support, hindering complete association model development (Yao, 2018) [20]. The weakness of composite salt degradation research also leaves Cl⁻/SO₄²⁻ interactive damage mechanisms unclear. While SO₄²⁻ has been found to exert a “first promotion, then inhibition” dual effect on Cl⁻ diffusion under freeze-thaw conditions and drive AFt-Friedel’s salt conversion, ITZ damage remains without quantitative characterization (Yao, 2018; Xu, 2025) [20,21].

The extreme environment of Northwest saline soil regions exacerbates this contradiction.

The actual erosion environment in Northwest saline soil areas (e.g., Tumushuke, Xinjiang) further highlights research limitations. Jiang et al. (2023) [19] measured extreme values of NaCl at 11.44% and Na₂SO₄ at 4.31%, with averages of approximately 6.02% and 7.04%, respectively, demonstrating significant unequal concentration characteristics. Annual seasonal freeze-thaw cycles and salt ion erosion create a “freeze-thaw-salt erosion” coupling damage effect. Xu et al. (2025) [21] found through accelerated corrosion tests that Cl⁻/SO₄²⁻ interactions in composite salt environments exhibit “first inhibition, then promotion” characteristics: initially, SO₄²⁻ preferentially reacts with Ca(OH)₂ to form gypsum, temporarily inhibiting Cl⁻ diffusion. After 50 cycles, gypsum replaces Friedel’s salt in the ITZ, releasing Cl⁻ and generating AFt, which combined with ice expansion stress increases crack width by 0.3–0.5 mm compared to single-salt environments [21]. Chen (2022) [22] further confirmed that such ion interactions and product conversions accelerate RAC freeze-thaw damage rates by 1.5–1.8 times.

The extreme environment of Northwest saline soil regions intensifies these challenges.

Field investigations show that salt erosion has become the primary cause of concrete durability degradation in southern Xinjiang. Structures with over 10 years of service commonly exhibit cover spalling and steel corrosion. Typical engineering cases include 82 mm maximum corrosion depth at bridge piers contacting saline soil on the Tarim River First Bridge; 110–150 mm expansion joint cracks in the Alaer City Army Reclamation Road pavement due to salt-frost heaving (indicating insufficient toughness and crack resistance); and 62 mm wide through-cracks in a Kizilsu anti-seepage channel after just 3 years of service. Such early failures caused by inadequate adaptation to high-salt environments not only threaten structural safety and functionality but also significantly shorten service life and increase maintenance costs (Wang, 2017; Li, 2012) [23,24].

This study proposes the following innovations: (1) Establish a Cl⁻/SO₄²⁻ dynamic gradient salt solution system based on Jiang et al.’s (2023) [19] measured extremes of NaCl 11.44% and Na₂SO₄ 4.31%, approximating a 12%/4% concentration gradient to accurately simulate actual erosion conditions. (2) Combining scanning electron microscopy (SEM) for ITZ microstructure observation, X-ray diffraction (XRD) for quantitative analysis of ettringite (AFt) and Friedel’s salt contents, and mercury intrusion porosimetry (MIP) for pore structure evolution tracking, this study quantitatively linked crystallization behavior (e.g., AFt expansion) to ITZ porosity changes during compound salt freeze-thaw cycles. (3) Develop a “pore optimization-ion blocking-fiber crack resistance” triple protection model to achieve >60% relative dynamic elastic modulus retention for 100% RCA-based RAC after 200 composite salt freeze-thaw cycles, providing critical technical support for engineering applications in Northwest high-cold saline soil regions.

2. Experimental Design

2.1. Fundamental Properties of Raw Materials

All materials used in this study comply with the specifications outlined in the Chinese National Standards Common Portland Cement (GB 175-2020) [25] and Technical Code for Application of Mineral Admixtures (GB/T 51003-2014) [26]. Their critical properties are summarized in

Table 2. Basic material properties.

Material Name	Fundamental Properties
Ordinary Portland Cement (P·O 42.5)	Density: 3.10 g/cm3, specific surface area:349 m2/kg, loss on ignition (LOI): 3.57%, residue on 80 μm sieve: <10%, chemical composition (wt.%): CaO 58.62; SiO2 23.08; Al2O3 6.01
Fly Ash (FA)	Density: 2.10 g/cm3, bulk density: 1.10 g/cm3, moisture content: 0.40%, loss on ignition (LOI): 2.62% fineness: 16%, chemical composition (wt.%): SiO2: 45.1; Al2O3: 36.8; CaO: 4.5; SO3: 1.2; Fe2O3: 0.85; alkali content (as Na2O eq.): 0.75; Cl−: 0.015
Silica Fume (SF)	SiO2 content: 98.1%, Cl− content: 0.01%, alkali content: 0.18%, specific surface area: 22.1 m2/g, loss on ignition: 1.48%, water demand ratio: 112%, activity index (28-day): 105%
Fine Aggregate of Class II	Natural river sand, with a fineness modulus of 2.68, a bulk density of 1516 kg/m3, and an apparent density of 2550 kg/m3
Recycled Coarse Aggregate (RCA)	Continuously graded particles with a size range of 5 to 26.5 mm, an apparent density of 2356 kg/m3, a crushing value of 16.1%, and a water absorption rate of 4.5%
Polypropylene Fiber (PPF)	Monofilament polypropylene fibers produced by Shandong Jinhongyao Engineering Materials Co., Ltd. (Jinan City, China) are used, with a diameter of approximately 36 μm, a density of about 0.91 g/cm3, a length of approximately 12 mm, a tensile strength > 430 MPa, an elongation > 35%, a breaking strength of 455 MPa, an initial modulus of 4200 MPa, and resistance to acids and alkalis
Water Reducer	DFTR-PCE standard type, a high-performance polycarboxylate-based water reducer, is a brown liquid with a density of (1.09 ± 0.02) g/mL and a water reduction rate of 26% (Sichuan Dongrun Baisheng New Materials Co., Ltd., Chengdu, China)

.

2.2. Mix Design of Recycled Aggregate Concrete

This study design for targeted research uses 100% recycled coarse aggregate to focus on freeze-thaw conditions in Tumushuke, Xinjiang’s saline soil area. The experimental design incorporated three mix proportion schemes strictly adhering to Xinjiang’s local standard “Technical Specification for Recycled Aggregate Concrete” (XJJ 076-2017) [27] and referencing research outcomes from [13,14,28,29].

RCAC1(reference group): single-doped with 6% SF, utilizing the pozzolanic effect to optimize pore structure and enhance ITZ compactness.

RCAC2 (FA-PPF modified group): compound doped with 20% FA and 0.9% PPF, leveraging FA’s filling effect and PPF’s bridging-crack arresting synergy to improve freeze-thaw damage resistance.

RCAC3 (FA-SF modified group): compound doped with 20% FA and 6% SF, analyzing the pozzolanic-filling dual effect on salt resistance performance. During specimen preparation, based on ITZ evolution studies, the water-binder ratio was strictly controlled at 0.40, and workability was ensured through pre-wetted recycled coarse aggregate (RCA) using the water absorption compensation method. Detailed mix proportion parameters for each group are provided in

Table 3. Mix design for frost-resistant recycled coarse aggregate concrete.

Specimen Number	Cement (kg/m3)	FA (%)	SF (%)	PPT (%)	RCA (kg/m3)	Sand (kg/m3)	Water (kg/m3)	Water-Reducing Agent (kg/m3)
RCAC1	497.26	0	6	0	1085	535	185	5.29
RCAC2	418.44	20	0	0.9	1085	535	185	5.29
RCAC3	391.46	20	6	0	1085	535	185	5.29

. Each group comprised five replicate specimens (prisms and cubes) designed for standardized testing.

2.3. Specimen Preparation

To ensure consistent recycled concrete quality and minimize variability from unstable recycled coarse aggregate sources all recycled coarse aggregate (RCA) was purchased in one batch from Nanjing Shoujia Renewable Resources Co., Ltd., Nanjing, China. The aggregates were thoroughly washed before pouring to remove surface silt (Figure 1). To further ensure consistency, all specimens were cast in a single pour to prevent quality differences between them.

During the pouring phase, humidity was continuously monitored and regulated via a laboratory-grade constant temperature and humidity control system, ensuring full compliance with the 60% ± 5% RH range throughout the entire specimen preparation workflow—from raw material weighing, mixing, and pouring to vibration compaction.

During concrete pouring, the mix proportion design strictly followed the Standard for Mix Proportion Design of Ordinary Concrete (JGJ 55-2011) [30]. Insert-type vibrators were used to ensure proper compaction of the concrete.

Post-forming, specimens underwent a 4 h static curing period at 20 ± 2 °C and 60% ± 5% RH. Following demolding, they were transferred to a standard curing chamber (25 ± 2 °C, ≥95% RH) for 24 days of controlled curing to ensure complete hydration reaction progression.

Meanwhile, prism specimens and cube specimens were cured together to ensure uniform material properties across different specimen types.

2.4. Composite Salt Freeze-Thaw Cycling Test

To enhance the study’s relevance and practical applicability, a representative salt ion concentration prevalent in the Tumushuke region of Xinjiang (12% NaCl and 4% Na₂SO₄) [19] was selected to establish the freeze-thaw cycling environment. This configuration was used to investigate the degradation characteristics of concrete’s frost resistance under these conditions.

Specimen preparation strictly adhered to the “Standard for Test Methods of Long-Term Performance and Durability of Ordinary Concrete” (GB/T 50082-2009) [31]. Two specimen types were cast: 100 mm × 100 mm × 100 mm cubes (for compressive strength testing) and 100 mm × 100 mm × 400 mm prisms (for mass loss and relative dynamic elastic modulus measurement). Following casting, all specimens were cured for 24 days in a standard curing room (temperature 25 ± 2 °C, relative humidity ≥ 95%) to ensure complete hydration.

The experimental procedure comprised the following steps:

(1) Salt solution immersion: specimens cured for 24 days were fully immersed in the prepared composite salt solution (temperature 20 ± 2 °C). The solution level was maintained 20–30 mm above the specimen top surface for 4 days to achieve salt saturation.

(2) Freeze-thaw cycling: specimens underwent multiple freeze-thaw cycles using a TR-TSDR-3 rapid freeze-thaw testing apparatus. A single test period consisted of 25 complete freeze-thaw cycles. The temperature regime per cycle was freezing at −18 °C (maintained for 2 h), followed by thawing at +5 °C (maintained for 2 h), resulting in a total cycle duration of 4 h. Cooling and heating rates were controlled at ≤1 °C/min and 0.5 °C/min, respectively.

(3) Temperature monitoring and control: thermocouples (with an accuracy of ±0.5 °C) were embedded to continuously monitor the core temperature of the specimens, ensuring deviations from the setpoint remained ≤1 °C. A real-time temperature monitoring system was used to verify two key conditions: firstly, that the temperature profiles during each phase complied with preset parameters; secondly, that the contact state between the specimen surface and the composite salt solution remained stable.

After every 25 freeze-thaw cycles, both prism and cube specimens were removed from the apparatus. Following surface drying, their mass, relative dynamic elastic modulus, and compressive strength were measured to track the progression of performance degradation. The test was terminated upon meeting any of the following criteria: completion of the predetermined number of freeze-thaw cycles, reduction of the specimen’s relative dynamic elastic modulus to 60% of its initial value, and achievement of a 5% MLR.

To ensure systematicity and completeness of the study while avoiding omission of critical variables, this experiment employed a design of experiments (DoEs) framework (see

Table 4. DoE framework for recycled concrete frost resistance testing.

Test Group	Mix Proportion Scheme	Fixed Experimental Conditions (Identical Across Groups)	Independent Variables (Test Parameters)	Dependent Variables (Observation Metrics)
RCAC-1	See Table 2	1. RCA replacement rate: 100% 2. Water-cement ratio: 0.40; Sand ratio: 0.33 3. Specimen dimensions: Cube: 100 mm × 100 mm × 100 mm Prism: 100 mm × 100 mm × 400 mm 4. Curing: Standard conditions (20 ± 2 °C, RH ≥ 95%) 5. Freeze-thaw method: Rapid freezing-thawing (GB/T 50082—2009), fully immersed 6. Composite salt medium: 12% NaCl + 4% Na2SO4 (by mass)	0 (unfrozen control)	Macroscopic:  - MLR- RDEM - CSLR Microscopic:  - SEM - XRD   - MIP
			25/50/75/100/125	Same as above
RCAC-2	See Table 2	Same as above	0 (unfrozen control)	Same as above
			25/50/75/100/125	Same as above
RCAC-3	See Table 2	Same as above	0 (unfrozen control)	Same as above
			25/50/75/100/125	Same as above

Notes: (1) Core independent variables: mix proportion schemes (RCAC1/RCAC2/RCAC3) and freeze-thaw cycle counts (N), investigated through bivariate design to analyze frost resistance differences. (2) Dependent variables: macroscopic performance (MLR/RDEM/CSLR) and microscopic characterization (SEM/XRD/MIP). (3) Abbreviations: MLR = mass loss rate; RDEM = relative dynamic elastic modulus; CSLR = compressive strength loss rate; DoEs = design of experiments. (4) Freeze-thaw parameters: freezing temperature: −18 ± 2 °C; thawing temperature: 5 ± 2 °C; cycle duration: 2–4 h. (5) SEM (ITZ morphology/cracks), XRD (expansion products), MIP (ITZ porosity).

) to organize variables, levels, and control groups, establishing a holistic research structure.

2.5. Multiscale Evaluation Metrics

A comprehensive “macro-meso-micro” evaluation system was established to thoroughly assess the freeze-thaw durability of recycled aggregate concrete. The specific evaluation metrics are detailed below:

2.5.1. Macro-Scale Performance Metrics

Macro-scale evaluation focuses on changes in the overall physical and mechanical properties of the material under freeze-thaw cycling.

(1)

MLR

MLR serves as a key indicator for quantifying surface scaling and internal material loss in concrete. Exceeding the critical threshold (typically 5%) signifies protective layer failure, leading to reinforcement corrosion and accelerated structural deterioration. MLR is calculated as follows (1):

$$ML={\displaystyle \frac{m_0-m_{\mathrm{N}}}{m_0}}\times 100\%$$

(1)

where:

ML: MLR after N freeze-thaw cycles (%)

m₀: initial mass of specimen (g), measured to 0.1 g accuracy

m_N: oven-dried mass of specimen after N freeze-thaw cycles (g)

Testing strictly adhered to the “Standard for Test Methods of Long-Term Performance and Durability of Ordinary Concrete” (GB/T 50082-2009) [31]. To ensure data reliability, post-test processing mandated retention of at least three valid datasets per experimental group after outlier exclusion, with final results calculated as arithmetic means.

(2)

RDEM Loss Rate

RDEM is a critical non-destructive parameter characterizing the attenuation of material stiffness and the accumulation of internal damage. It is essential for evaluating the frost resistance of recycled aggregate concrete. This parameter is determined by measuring changes in the propagation velocity of ultrasonic waves through the specimen. In accordance with Standard [31], testing was performed using the non-metallic ultrasonic pulse method. A DJUS-05 dual-channel non-metallic ultrasonic detector (Shenzhen Shenborui Instrument Co., Ltd., Shenzhen, China) was employed. Changes in RDEM were quantified by measuring variations in the ultrasonic wave transit time through the specimen. Measurement Procedure: Following every 25 freeze-thaw cycles, RDEM was calculated using Equation (2):

$$E_{\mathrm{RN}}={\displaystyle \frac{E_{\mathrm{dN}}}{E_{\mathrm{d}0}}}={({\displaystyle \frac{t_0}{t_{\mathrm{N}}}})}^2\times 100\%$$

(2)

where:

E_RN: RDEM retention rate after N freeze-thaw cycles (%)

E_d0, E_dN: dynamic elastic modulus of the specimen in its initial state and after N freeze-thaw cycles, respectively (MPa)

t₀, t_N: ultrasonic longitudinal transit time before any freeze-thaw cycles (0 cycles) and after N freeze-thaw cycles, respectively (s)

To ensure data reliability, post-test processing mandated retention of at least three valid datasets per experimental group after outlier exclusion, with final results calculated as arithmetic means.

(3)

Cube Compressive Strength

Compressive strength represents the fundamental mechanical property of concrete and is significantly affected by freeze-thaw cycling. In accordance with the “Standard for Test Methods of Mechanical Properties of Ordinary Concrete” (GB/T 50081-2019) [32], the compressive strength of 100 mm × 100 mm × 100 mm recycled aggregate concrete cube specimens was determined using a YES-2000B compression testing machine (Wuxi Jianyi Instrument and Machinery Co., Ltd., Wuxi, China). This assessment evaluated the impact of freeze-thaw cycles on the load-bearing capacity of the concrete.

The compressive strength was calculated using Equation (3):

$$f_{\mathrm{c}}={\displaystyle \frac{P_{\max }}{A}}$$

(3)

where:

f_c: cube compressive strength of concrete (MPa)

P_max: ultimate load at specimen failure (N)

A: loading surface area of the specimen (mm²)

To ensure data reliability, post-test processing mandated retention of at least three valid datasets per experimental group after outlier exclusion, with final results calculated as arithmetic means.

2.5.2. Mesoscale Structural Evaluation Metrics

The mesoscale analysis focuses on the internal phase composition and morphological characteristics of the material to elucidate the physical mechanisms of freeze-thaw damage. This study systematically investigated the microstructural deterioration process of recycled aggregate concrete using the following two complementary analytical methods:

(1)

Phase Composition Analysis

Phase analysis was performed using a Rigaku SmartLab SE X-ray diffractometer (XRD) (Tokyo, Japan). Testing conditions were Cu-Kα radiation (λ = 1.5406 Å), an operating voltage of 40 kV, and a current of 30 mA. Continuous scanning mode was employed over a 2θ range of 5° to 80° at a scan rate of 10°/min. Phase identification and crystallinity calculations were conducted using Jade 9.0 software with the ICDD PDF-4+ 2023 database.

(2)

Micromorphological Characterization

The fracture surface micromorphology of specimens was examined using an FEG650 field emission scanning electron microscope (FESEM) (FEI Company, Hillsboro, OR, USA). Testing conditions included an accelerating voltage of 15 kV and a working distance of 10 mm. Analysis specifically targeted the ITZ between aggregates and the cement paste matrix, with particular focus on crack initiation, propagation, and microstructural alterations.

2.5.3. Microstructural Evolution Evaluation Metrics

The microscale investigation delves into the characteristics and evolution of internal pore structures, which are critical for understanding the physicochemical nature of freeze-thaw deterioration. Accordingly, pore structure analysis was conducted following ASTM D4404-18 (ASTM International, 2018) [33] using a Quantachrome POREMASTER 33 mercury intrusion porosimeter (MIP) (Boynton Beach, FL, USA). Specimens were oven-dried at 50 °C for 48 h under vacuum prior to testing. The maximum intrusion pressure was set to 414 MPa, corresponding to a minimum detectable pore diameter of 3 nm. Data analysis focused on three key parameters: cumulative porosity, median pore diameter, and the evolution of harmful pore (50–200 nm) volume fraction.

3. Analysis and Discussion of Macro-Scale Test Results

3.1. Evolution of Macro-Scale Properties in Recycled Aggregate Concrete

3.1.1. Visual Degradation of Prism Specimens

Figure 2, Figure 3 and Figure 4 present the visual manifestations of scaling and surface deterioration on three groups of recycled aggregate concrete prism specimens subjected to varying numbers of freeze-thaw cycles (0, 25, 50, 75, 100, and 125 cycles) within the composite salt solution.

Figure 2, Figure 3 and Figure 4 illustrate the surface deterioration characteristics of recycled aggregate concrete prism specimens after 125 composite salt freeze-thaw cycles. Comparative analysis reveals:

(1) RCAC1 (reference group) exhibited approximately 65% aggregate exposure ratio, with scaling depths ranging from 5 to 12 mm.

(2) RCAC2 (FA-PPF modified group) demonstrated significantly enhanced damage resistance, reducing aggregate exposure by 32% to 51% relative to the reference group and limiting scaling depth to approximately 2.0 mm (a 31% reduction).

(3) RCAC3 (FA-SF modified group) experienced performance degradation due to material incompatibility issues, resulting in increased aggregate exposure (~70%) and scaling depths up to approximately 15 mm.

Deterioration mechanisms: composite salt freeze-thaw damage is driven by synergistic physical and chemical processes. The physical mechanism involves ~9% volumetric expansion from pore water freezing (−7 to −1 °C), while the chemical mechanism originates from Cl⁻/SO₄²⁻ enrichment forming expansive crystalline phases (e.g., Friedel’s salt and AFt). Concurrently, crystallization pressure exceeding the concrete’s tensile strength threshold induces cracking within the ITZ of recycled aggregates.

Comparative findings: The RCAC2 mixture effectively suppressed crack propagation through fiber bridging effects. Conversely, the alkaline environment (pH 12–13) of the cementitious matrix in the RAC3 group creates essential conditions for chemical reactions between aggressive ions and cement hydration products. This facilitates the formation of two expansive crystalline products: AFt (from sulfate ions) and Friedel’s salt (from chloride ions). Specifically, sulfate ions first react with calcium hydroxide to form gypsum, which subsequently reacts with tricalcium aluminate (C₃A) to produce AFt. Chloride ions, meanwhile, react with monosulfate aluminoferrite (AFm) phases to form Friedel’s salt. Notably, the ITZ in RAC3 recycled aggregate concrete exhibits dual interface characteristics (“recycled aggregate-old mortar” and “old mortar-new paste”), resulting in significantly higher porosity compared to conventional concrete. This makes the ITZ a preferential site for expansive product accumulation. The local expansive stresses generated during crystallization exceed the ITZ’s tensile strength (1.5–4.0 MPa), ultimately accelerating structural degradation [34,35,36].

3.1.2. Evolution of MLR

Freeze-thaw responses in composite salt solution, demonstrated in Figure 5 show three distinct mass change phases in recycled concrete specimens: (1) salt absorption weight gain phase (N < Nc): salt solution enters pores through capillary absorption, resulting in an overall mass increase (negative loss rate); the critical cycle count Nc indicates salt-frost resistance. (2) Dynamic equilibrium phase (N ≈ Nc): microcrack growth balances corrosion product filling, and mass stabilizes (loss rate approaches zero). (3) Damage weight loss phase (N > Nc): synergistic salt crystallization pressure and ice-expansion stress increase, microcracks expand and connect, surface mortar spalls off, and mass decreases with a rapidly rising positive loss rate.

RCAC1 (reference group) results: below 67 freeze-thaw cycles (N < 67), specimens gained mass (negative MLR) during the salt absorption phase. At 76 cycles (N = 76), mass loss reached 0%—absorption/damage equilibrium point. Beyond 76 cycles (N > 76), mass loss progressively increased reached 0.87% at 125 cycles. Initially during freeze-thaw cycles, cementitious materials form expansive corrosion products—AFt and Friedel’s salt [37]. These products fill pores, dominating mass gain. First, as cycling progresses, salt solutions penetrate concrete cracks and crystallize (e.g., Na₂SO₄·10H₂O). Second, this causes volume expansion [38], accelerating microcrack growth and interconnection. Furthermore, structural failure occurs in surface mortar, leading to spalling and resulting in accelerated mass loss.

RCAC2 (FA-PPF modified group): this group showed significantly delayed critical damage onset at Nc = 92 cycles, later than other mixes. After 125 freeze-thaw cycles, mass loss was only 0.28% and 67.8% lower than the baseline. The 20% FA and 0.9% PPF synergy provided dual protection. First, FA’s pozzolanic reaction formed C-S-H gel, refining pores and slowing salt penetration. Second, PPF bridged cracks, and MLR distributed ice-expansion stress [14,39]. Furthermore, this extended the salt-absorption phase and suppressed mass loss.

After 100 freeze-thaw cycles, it reached only 0.20%, closely approaching the 0.132% observed in ordinary concrete with 20% FA under identical conditions [40]. Following 125 cycles, this rate (0.30%) outperformed ordinary concrete modified with 1% PPF and 10% silica fume [41], while remaining substantially lower than the 4.64% loss reported for 100% recycled coarse aggregate concrete containing silica fume and 0.6% PPF [14]. These findings conclusively validate the superiority of this modification approach in enhancing freeze-thaw damage resistance for recycled concrete.

The RCAC3 group (FA-SF modified) exhibited accelerated deterioration: the critical threshold Nc occurred earlier at 39 cycles, after which MLR increased sharply. By N = 125 cycles, the mass loss reached 2.3%, being 2.6 times higher than RCAC1 and 8.2 times higher than RCAC2. This is because while 20% FA optimized pore structure, the addition of 6% SF elevated pore solution alkalinity. This promoted the Na₂SO₄ reaction and expansive AFt formation. As revealed by Shi (2019) [42], high SF dosage alters hydration product formation through alkalinity elevation, causing silicate framework depolymerization and C-(A)-S-H gel structure degradation. Consequently, porosity increased along with harmful pore proportions (>50 nm). During advanced freeze-thaw stages, accumulated corrosion products combined with reduced hydration product cohesion, accelerating critical point onset and causing severe surface mortar spalling. This mechanism explains RCAC3′s inferior performance compared to other groups.

In summary, distinct failure mechanisms emerged among the groups: RCAC2 (delayed Nc) achieved prolonged durability through synergistic effects: FA ash pore refinement reduced permeability, while PPF crack-bridging dispersed stress concentrations. This dual action extended the serviceable life phase. RCAC3 (accelerated Nc) experienced premature deterioration as high-reactivity silica fume elevated pore solution alkalinity, triggering chemical attacks (AFt expansion, silicate depolymerization) and physical degradation (increased harmful pores >50 nm). These processes accelerated damage progression, with Nc occurring 26 cycles earlier than RCAC1 and 86 cycles earlier than RCAC2. RCAC1 (intermediate Nc) performed between the extremes, where inherent recycled aggregate defects (elevated porosity, microcrack networks) facilitated aggressive ion penetration. This group’s Nc value fell midway between RCAC2′s superior performance and RCAC3′s rapid decline. The contrasting behaviors highlight material-specific vulnerability thresholds where chemical-induced degradation dominates in silica fume systems, while physical pore optimization governs FA-based modifications. Recycled aggregate quality remains the baseline determinant for erosion resistance.

3.1.3. Degradation Pattern of RDEM

Under combined salt freeze-thaw cycling, all three recycled concrete groups exhibited continuous decline in RDEM retention rates (Figure 6), though with markedly different degradation rates.

RCAC1 group (reference): after 125 cycles, RDEM retention reached 71.30% ± 0.20% (p < 0.05). Primary degradation stemmed from salt solution infiltration into existing cracks during cycling. First, leading to salt crystallization and expansive corrosion products (AFt, Friedel’s salt). Second, the combined crystallization pressure and expansion stress accelerated microcrack propagation. Furthermore, reducing structural density [43].

RCAC2 group (FA-PPF modified): synergistic modification delivered superior durability. After 125 cycles, RDEM retention reached 75.98% ± 0.21% (6.6% improvement over RCAC1). Pozzolanic reaction of 20% FA generating C-S-H gel, blocking capillary pores and reducing salt solution penetration. First, 0.9% PPF bridges cracks and distributes ice expansion stress [44]. Second, this dual action delays salt solution ingress and erosion. Furthermore, it preserves internal structural integrity better than other groups—consistent with its lowest MLR. Comparative analysis shows Wang et al. (2015) [45] reported ordinary concrete with 20% FA reached 50.3% RDEM after only 40 freshwater cycles. Wang et al. (2017) [14] found 100% recycled aggregate concrete with SF + 0.6 kg/m³ PPF achieved only 33.15% RDEM after 125 freshwater cycles. Both outcomes validate RCAC2′s superior performance. An exponential decay model was developed to predict RDEM retention. This project predicts > 60% retention at 150–200 cycles.

RCAC3 group (FA-SF Modified) RDEM retention dropped sharply to 51.86% after 125 cycles (69.6% greater degradation than RCAC1). Critical failure occurred earlier at Nc = 39 cycles (Figure 6 dashed line). 6% SF increasing pore solution alkalinity. First, accelerating the Na₂SO₄ reaction and expansive AFt formation. Second, the high-alkali environment causes silicate framework depolymerization and C-(A)-S-H gel degradation [42].

This increased harmful pore proportions (>50 nm) and accelerated crack initiation/propagation, leading to severe structural disintegration. Conclusion FA-PPF modification (RCAC2) effectively delayed RDEM decline and enhanced salt freeze resistance. Conversely, FA-SF modification (RCAC3) accelerated degradation through alkali-induced chemical attacks (AFt expansion, product decomposition) and physical deterioration (accelerated cracking). RDEM trends correlate directly with mass loss observations, collectively demonstrating material performance evolution under salt freeze conditions.

3.1.4. CSLR

Compressive strength degradation under salt freeze-thaw cycling results in all three recycled concrete groups exhibiting continuous increases in compressive strength loss rates with rising freeze-thaw cycles under combined salt exposure (Figure 7), though with markedly different degradation kinetics.

RCAC1 group (reference): initial strength measured 42.5 MPa, declining to 23.4 MPa after 125 cycles (45.0% ± 0.20% loss, p < 0.05). During freeze-thaw cycling, synergistic expansion stresses arise from two sources. First, one source is salt crystallization pressure. Second, the other source is the expansive forces generated by corrosion products like AFt and Friedel’s salt. These combined stresses induce interconnected internal cracking, as documented by Zhang et al. (2017) [46] and Valenza and Scherer (2005) [47]. The cumulative mechanical-chemical interaction progressively weakens the concrete matrix, accelerating structural deterioration through crack coalescence. Additionally, accelerated Cl⁻ penetration during cycling exacerbated damage, corroborated by concurrent observations of 0.87% mass loss and 71.3% RDEM retention at 125 cycles.

RCAC2 group (FA-PPF modified): after 125 cycles, strength loss reached 38.0% ± 0.20% (15.6% lower than RCAC1). Degradation occurred at a slower rate, demonstrating superior freeze-thaw resistance. This is because 20% FA fills harmful pores (>100 nm), optimizing pore structure and 0.9% PPF bridging cracks and enhancing crack resistance. These findings align with Akbulut et al. (2024) [39], who reported 18–35% improved energy dissipation in fiber-reinforced RCAC. Wang et al. (2017) [14] further confirmed that 0.9 kg/m³ PPF increases compressive strength by 2.23 MPa compared to references, validating the “filling-crack-arresting” synergy in delaying strength loss.

RCAC3 group (FA-SF modified): strength plummeted to 19.6 MPa after 125 cycles (53.9% ± 0.20% loss, p < 0.01), 20% higher than RCAC1. Degradation occurred in two phases. First, was the early stage (N < 50) where 6% SF elevated pore solution alkalinity, accelerated OH⁻ erosion at the aggregate-paste interface (ITZ) and microcrack propagation. The second phase was the late stage (N > 75) where massive formation of expansive products (AFt, gypsum CaSO₄·2H₂O) combined with frost heaving stress, caused ~70% surface mortar spalling and severe structural disintegration.

Salt freeze damage in recycled concrete involves interconnected processes where existing cracks act as ion migration pathways, accelerating crystallization pressure buildup and swelling product formation (Friedel’s salt, AFt). Synergistic expansion from these stresses drives crack coalescence, manifesting macroscopically as rising strength/mass loss and RDEM decline. Performance comparisons clearly demonstrate RCAC2 (FA-PPF) outperforms both RCAC1 and RCAC3 under salt freeze conditions. This aligns with Cui et al. (2023) [48], who reported PPF’s effectiveness in reducing mass loss, RDEM degradation, and strength loss in RCAC, particularly at high recycled aggregate contents.

3.2. Analysis Based on Standard Deviations of Indicators

The standard deviation serves as a critical statistical metric for quantifying the dispersion of experimental data and evaluating the stability of material performance along with the reliability of test results. This study analyzed the experimental validity and performance stability of three recycled aggregate concrete (RAC) groups using standard deviations calculated from MLR, RDEM loss rate, and compressive strength loss rate data collected during the full freeze-thaw cycle period (0–125 cycles). Detailed results are presented in

Table 5. Summary of standard deviations for frost resistance indicators of three RAC groups.

Evaluation Indicator	RCAC1	RCAC2	RCAC3
MLR (%)	0.3754	0.2510	0.8699
RDEM Loss Rate (%)	9.6218	8.0940	16.5424
CSLR (%)	16.9108	13.6203	16.6714

.

As demonstrated in Table 4, the standard deviations of all indicators for the three groups exhibit a consistent stability gradient: RCAC2 shows the smallest standard deviations across all three metrics, indicating minimal performance fluctuations during freeze-thaw cycling and superior material homogeneity along with experimental reproducibility. RCAC1 ranks intermediate, with performance stability between RCAC2 and RCAC3. RCAC3 displays significantly higher standard deviations in MLR and RDEM loss rate compared to the other groups, though its CSLR standard deviation approaches that of RCAC1. This pattern indicates greater dispersion in surface damage and internal structural degradation for RCAC3, making it the least stable material overall.

3.3. Comprehensive Frost Resistance Evaluation Using Radar Charts

To integrate multi-dimensional information from MLR, RDEM loss rate, and compressive strength loss rate while avoiding the bias of single-indicator evaluations, this section employs radar chart visualization for quantitative comprehensive evaluation of the frost resistance of the three RAC groups.

3.3.1. Data Preprocessing and Normalization Method

Although MLR, RDEM loss rate, and compressive strength loss rate are all percentage-based metrics, their numerical ranges differ significantly (0–2.3%, 0–48.14%, and 0–53.9%, respectively). Direct comparison risks overemphasizing indicators with larger numerical ranges. Given that all three are negative indicators (smaller values indicate better frost resistance), an inverse extreme value normalization method was adopted to eliminate dimensional effects, scaling raw data to the [0, 1] interval (closer to 1 indicates better frost resistance).

$$\mathrm{The} \mathrm{normalization} \mathrm{formula} \mathrm{is}: S_{\mathrm{i}}={\displaystyle \frac{X_{\max }-X_i}{X_{\max }-X_{\min }}}$$

(4)

where S_i is the normalized value for group i; X_i is the raw indicator value for group i; M_max and M_min are the maximum and minimum values of the indicator across all groups during the 0–125 freeze-thaw cycles. The normalized results for the three groups after 125 freeze-thaw cycles are shown in

Table 6. Normalized results for three RAC groups after 125 freeze-thaw cycles.

Evaluation Indicator	RCAC1	RCAC2	RCAC3
Normalized MLR	0.5257	0.7426	0.0000
Normalized RDEM Loss Rate	0.4038	0.5010	0.0000
Normalized CSLR	0.1651	0.2430	0.0000

.

3.3.2. Radar Chart Construction and Evaluation Logic

(1)

Indicator Weight Assignment

The three macro-indicators reflect frost resistance from distinct dimensions: normalized MLR characterizes surface spalling and mass degradation; normalized RDEM loss rate reflects internal microstructural damage and stiffness degradation; normalized compressive strength loss rate represents core degradation in load-bearing capacity. Given their equal importance in frost resistance evaluation, equal weights (1/3 for each indicator) were assigned to avoid subjective bias and ensure objective evaluation.

(2)

Radar Chart Geometry

The radar chart uses the three normalized indicators as independent axes, with pairwise angles set to 120° (uniform distribution for geometric symmetry). Axis scales range from 0 to 1.0, with 0.2 intervals.

The three groups are distinguished by line styles and fill patterns:

RCAC1: black dashed line + red fill (60% transparency);

RCAC2: blue solid line + green fill (50% transparency);

RCAC3: gray dotted line + no fill

3.3.3. Radar Chart Results and Quantitative Area Analysis

The radar charts for the three RAC groups after 125 freeze-thaw cycles are shown in Figure 8.

As shown in Figure 8, the radar polygons of the three groups exhibit distinct morphological differences:

RCAC2: the most robust polygon, with the highest normalized values across all indicators (mass: 0.7426, RDEM: 0.5010, strength: 0.2430) and no significant performance weak link, indicating balanced frost resistance.

RCAC1: a “slender” polygon, with moderate normalized MLR (0.5257) and RDEM loss rate (0.4038), but a markedly low normalized compressive strength loss rate (0.1651), which becomes the primary constraint on comprehensive performance.

RCAC3: all normalized values are 0.0000, collapsing to the origin, indicating complete failure in surface integrity, internal structural degradation, and load-bearing capacity retention.

To quantitatively compare comprehensive frost resistance, the closed area of the radar polygon (calculated using a formula for polygons with 120° angles between axes) was employed (a larger area indicates stronger comprehensive frost resistance).

$$\mathrm{The} \mathrm{formula} \mathrm{is}: A=0.5\times (S_1{S}_2+S_2{S}_3+S_3{S}_1)\times \sin {120}^{\mathrm{o}}$$

(5)

where A is the polygon area; S₁, S₂, and S₃ are the normalized MLR, RDEM loss rate, and compressive strength loss rate, respectively. Combining Table 5 data, the areas and rankings are summarized in

Table 7. Comparison of single-indicator rankings and comprehensive radar chart rankings.

Evaluation Dimension	Single-Indicator Ranking (Descending Order)	Consistency with Comprehensive Ranking
Normalized MLR	RCAC2 > RCAC1 > RCAC3	Full Consistency
Normalized RDEM Loss Rate	RCAC2 > RCAC1 > RCAC3	Full Consistency
Normalized CSLR	RCAC2 > RCAC1 > RCAC3	Full Consistency
Radar Chart Closed Area	RCAC2 (0.2919) > RCAC1 (0.1584) > RCAC3 (0.000)

.

As shown in Table 7, the comprehensive ranking from the radar chart analysis is fully consistent with the single-indicator rankings, validating the reliability of the comprehensive evaluation. Through multi-indicator visualization and quantitative area analysis, the radar chart effectively integrates frost resistance information across appearance, structure, and mechanics, establishing the comprehensive frost resistance ranking as RCAC2 > RCAC1 > RCAC3.

Specifically, RCAC2 demonstrates optimal frost resistance, with minimal surface damage, slight internal stiffness degradation, and well-preserved load-bearing capacity. RCAC1 exhibits moderate frost resistance, with compressive strength degradation as the critical optimization target. RCAC3 performs worst, with complete loss of structural integrity and mechanical performance after freeze-thaw cycling.

To elucidate microscopic degradation mechanisms, subsequent analyses will employ MIP for pore structure evolution tracking. SEM for crack propagation visualization and corrosion product identification. XRD for phase composition analysis. These methods will systematically investigate salt freeze damage from the perspectives of pore refinement, crack development pathways, and corrosion product distribution.

4. Analysis and Discussion of Microstructural Test Results

4.1. Microstructural Phase Analysis of Recycled Aggregate Concrete (XRD Investigation)

XRD analysis was employed to investigate the phase evolution within the recycled aggregate concrete (RCAC) specimens following combined freeze-thaw and salt attack exposure (Figure 9).

Table 8. Changes in key parameters after freeze-thaw cycles.

Group	Friedel’s Salt Increase	AFt Increase	Gypsum Increase	Porosity Change	Dominant Reaction Pathway
RCAC1	+33.3%	+15.8%	-	+6.2%	Cl−/SO42− synergistic attack; competitive precipitation of expansive phases
RCAC2	+28.6%	+9.2%	+12.5%	−1.8%	Cl− preferential reaction; gypsum buffering expansion
RCAC3	+18.4%	+24.7%	-	+9.9%	SO42−-dominated reaction; enrichment of expansive phases

presents a comparative analysis of the changes in internal corrosion product content and porosity for the three specimen groups before and after freeze-thaw cycling. Analysis revealed a strong correlation between the competitive formation of expansive products (Friedel’s salt/AFt) and alterations in pore structure. These factors jointly govern the internal damage evolution process within the concrete.

4.1.1. RCAC1 Group (Reference Group): Cl⁻/SO₄²⁻ Synergistic Degradation Mechanism

In the RCAC1 reference group, competitive reactions between Cl⁻ and SO₄²⁻ ions led to simultaneous formation of Friedel’s salt and AFt (Table 3). During initial freeze-thaw cycles, Cl⁻ preferentially reacted with aluminate phases (C₃A), increasing Friedel’s salt content by 33.3%. This dense filling effect temporarily slowed erosion progression. However, as cycling continued, SO₄²⁻ penetration accelerated, causing sustained AFt formation with a 15.8% content increase. This triggered crack propagation in the ITZ.

Rietveld refinement analysis confirmed a significant rise in the peak intensity ratio of Friedel’s salt to AFt (I_11.3°/I_9.1°), indicating that competitive interactions between expansive products were the primary cause of material degradation (e.g., 28.7% RDEM loss rate) [46,47]. The conversion dynamics between AFt and Friedel’s salt during composite salt freeze-thaw cycling is illustrated in Figure 10.

4.1.2. RCAC2 Group (FA-PPF Modified): Synergistic Stabilization Effect

In contrast, combined incorporation of FA and PPF significantly optimized reaction pathways and microstructure. First, FA’s pozzolanic effect substantially reduced Ca(OH)₂ content, improving pore structure as evidenced by a 73% increase in non-harmful pore proportion and 1.8% porosity reduction (Table 8). Simultaneously, FA promoted preferential Cl⁻ binding into Friedel’s salt (content increased by +28.6%), while SO₄²⁻ primarily precipitated as less expansive gypsum (expansion rate only 3.7%) with a content increase of +12.5%. This reaction pathway partially mitigated AFt expansion stress [49]. Second, PPF’s three-dimensional network effectively constrained crack propagation and reduced the SO₄²⁻ diffusion coefficient, inhibiting sulfate ingress and blocking medium transport channels [50]. The synergistic interaction between FA and PPF resulted in the lowest AFt increment (Table 8), significant gypsum accumulation, and refined pore structure, collectively enhancing damage resistance under salt freeze-thaw conditions.

4.1.3. RCAC3 Group (FA-SF Modified): Competitive Deterioration of Expansive Phases

In contrast, combined incorporation of SF and FA induced dual detrimental effects. High SF dosage created a hyper-alkaline environment that exacerbated pore structure degradation. Under elevated alkalinity, SO₄²⁻ dominated the reaction pathway, causing a 24.7% increase in AFt content—exceeding Friedel’s salt accumulation (18.4%) and leading to massive expansion product aggregation. Studies confirm synergistic interactions between AFt’s volumetric expansion and freeze-thaw-induced frost heaving stress [47], significantly increasing crack density and forming a “loose-crushed” failure pattern that accelerated structural deterioration. Furthermore, while SF’s pore-refining effect initially aimed to improve microstructure, it paradoxically worsened pore connectivity and intensified alkaline dissolution, increasing porosity by 9.9% (Table 8). Huo et al. (2018) [49] demonstrated that such pore degradation accelerates crack propagation and surface spalling, exacerbating material performance decline.

In summary, the RCAC1 group (control) demonstrated competitive ingress of Cl⁻/SO₄²⁻, leading to competitive formation of Friedel’s salt and AFt, inducing progressive deterioration. Whereas in the RCAC2 group (FA-PPF), FA optimized the pore structure and promoted the formation of low-expansion gypsum, while PPF suppressed cracking. This dual synergy achieved efficient protection. In RCAC3 group (FA-SF) the highly alkaline environment promoted SO₄²⁻-dominated reactions and AFt enrichment, compounded by pore degradation, leading to accelerated damage.

4.2. MIP Analysis of Recycled Aggregate Concrete

The progression of freeze-thaw damage is intrinsically linked to pore structure evolution. MIP was performed to systematically analyze changes in pore parameters for the three specimen groups before (0 cycles) and after 75 freeze-thaw cycles (

Table 9. Comparison of key MIP parameters for the three specimen groups after 75 freeze-thaw cycles.

Core Parameter	RCAC1 Group (Reference)	RCAC2 Group (FA-PPF)	RCAC3 Group (FA-SF)
Porosity Change (%)	↓22.4% (9.8→7.6)	↓13.7% (6.5→5.6)	↑9.9% (8.2→9.0)
Median Pore Diameter (nm)	↑140.7% (52→125)	↓33.6% (48→32)	↓71.9% (68→19)
Permeability (mD) Variation (%)	↑99.3% (1.02→2.03)	↓99.96% (0.85→0.0003)	↓87.6% (2.15→0.27)
Seepage Fractal Dimension	↑6.7% (2.31→2.47)	↑1.5% (2.25→2.28)	↓0.1% (2.40→2.39)

Notes: (1) Permeability unit conversion: 1 mD ≈ 9.869 × 10⁻¹⁶ m² (≈1 × 10⁻³ μm²); (2) The arrow symbols ↑ indicate an increase, and ↓ indicate a decrease.

), revealing several key patterns:

4.2.1. RCAC1 Group (Reference)

Experimental results revealed a 22.4% porosity reduction (9.8%→7.6%) alongside a 140.7% surge in median pore diameter and a 99.3% permeability increase. This indicates that while pozzolanic reactions locally densified the matrix, they exacerbated macropore connectivity, creating invasion pathways for aggressive ions [44]. This aligns with the group’s poor freeze-thaw performance (71.3% RDEM retention, 45.0% compressive strength loss rate).

4.2.2. RCAC2 Group (FA-PPF Modified)

Synergistic FA-PPF modification achieved 13.7% porosity reduction, 33.6% median pore diameter reduction, and 99.96% permeability drop. Microscopic analysis confirmed a 73% increase in benign pore proportion (<50 nm), with harmful pore (>100 nm) volume decreasing from 15.2% to 0.40%. Cl⁻ diffusion coefficient reduction validated the “pore refinement-permeation suppression” synergy, making this the most freeze-thaw resistant group. Similar findings were reported by Wu (2011) [51].

4.2.3. RCAC3 Group (FA-SF Modified)

Despite a sharp 71.9% reduction in median pore diameter, porosity counterintuitively increased by 9.9%, with permeability remaining significantly higher than in RCAC2. This paradoxical outcome stems from structural instability caused by excessive SF, which induced coexistence of nanoscale pores (10–30 nm) and microcrack networks. This dual defect elevated total porosity, characterized by “pore refinement but connectivity deterioration”—a marked increase in harmful pore proportions. Consequently, RCAC3 exhibited accelerated surface spalling, a low RDEM retention rate (66%), and high compressive strength loss (33%), making it the group with the poorest freeze-thaw resistance.

RCAC2′s superior performance stemmed from combined porosity reduction (13.7%) and pore structure optimization (benign pore increase, harmful pore decrease), augmented by PPF’s crack-arresting effect. Conversely, RCAC3’s excessive SF disrupted pore connectivity control (porosity +9.9%, harmful pores increased), elevating permeability and accelerating damage. RCAC1′s paradoxical densification-permeability increase highlights the critical need for pore connectivity management beyond surface refinement.

4.3. SEM Morphology Analysis and Damage Mechanism of Recycled Concrete

Environmental SEM was employed to investigate microscopic damage progression in three recycled concrete groups after 0/75 freeze-thaw cycles, revealing distinct degradation mechanisms (Figure 11).

RCAC3 group (FA-SF modified): after 75 cycles, the ITZ exhibited through-thickness cracks averaging ~30 μm in width (Figure 11f), creating rapid percolation pathways. This accelerated mass loss (2.30%) and caused 54% compressive strength degradation—both significantly exceeding other groups. The deterioration originated from excessive SF inducing gel structure degradation and reducing ITZ bond strength.

RCAC2 group (FA-PPF modified): PPF effectively constrained crack propagation through bridging action, limiting widths to <21 μm (Figure 11d). Simultaneously, FA optimized pore structure via pozzolanic reactions, reducing harmful pore (>100 nm) proportions below 9.8% and suppressing aggressive ion penetration. These findings align with Zhang et al. (2022) [40], who confirmed FA’s effectiveness in reducing harmful pores and optimizing distribution.

RCAC1 group (reference): crack propagation severity fell intermediate between RCAC2 and RCAC3 (Figure 11b). Damage resulted from synergistic effects of salt crystallization pressure and expansive corrosion products (e.g., ettringite).

SEM analysis confirms that RCAC3’s excessive SF created brittle nanopore-microcrack networks, accelerating degradation. RCAC2’s FA-PPF synergy achieved optimal crack control and pore refinement. RCAC1’s intermediate performance stemmed from competing salt-induced expansion mechanisms.

4.4. Damage Evolution Pattern Under Coupled Salt-Freezing-Thawing Action

Based on macroscopic performance testing and microscopic observations, both RCAC1 and RCAC3 groups exhibited a two-stage damage evolution under composite salt freeze-thaw cycling:

(1)

Crystallization Phase Transition Stage

Periodic temperature fluctuations during freeze-thaw cycles drove repeated Na₂SO₄·10H₂O crystallization in pore solutions, accompanied by significant volumetric expansion [38]. This generated crystallization stress exceeding concrete’s tensile strength (3–4 MPa), initiating microcracks. Neville (2004) [34] confirmed Na₂SO₄ crystallization pressures up to 10–20 MPa, while Müllauer et al. (2013) [52] found ettringite formation in small pores (10–50 nm) could generate 8 MPa stress—both exceeding matrix tensile capacity. Cyclic crystallization-dissolution accelerated crack propagation, leading to cumulative initial damage.

(2)

Chemically-Dominated Corrosion Stage

Freeze-thaw cycles accelerated Cl⁻/SO₄²⁻ penetration into the ITZ. Cl⁻ bound into Friedel’s salt, while SO₄²⁻ reacted with Ca(OH)₂/C₃A to form expansive AFt. Synergistic expansion from these products increased ITZ porosity and accelerated crack growth. Material stiffness progressively degraded, with RDEM loss exceeding 40%.

(3)

RCAC2 Group (FA-PPF Modified) Damage Mitigation

PPF’s bridging action enhanced matrix fracture toughness, distributing stress concentrations and limiting microcrack coalescence. Chen (2022) [53] confirmed PPF networks improve compressive strength by 6.5–7.5% and tensile strength by 14.95%, reducing crack density. FA’s pozzolanic effect optimized pore structure, reducing harmful pores (>100 nm) to <9.8% and lowering Cl⁻ diffusion by 51.4% [51]. Additionally, low-expansion gypsum formation at crack tips alleviated stress concentration. This synergy limited RDEM loss to 24.02% and reduced compressive strength degradation by 23% compared to RCAC1.

In summary, the damage evolution follows distinct pathways: RCAC1/RCAC3: crystallization stress → chemical corrosion → accelerated degradation; RCAC2: pore refinement + crack bridging → suppressed stress accumulation → enhanced durability.

4.5. Triple Synergistic Protection Model Construction

Damage progression and performance optimization of recycled concrete in salt freeze-thaw environments in composite salt freeze-thaw conditions. Recycled concrete damage escalates through three sequential stages. First is physical frost heave where ice crystallization expands initial pores, creating pathways for ion migration. Second is chemical erosion where enlarged pores accelerate Cl⁻/SO₄²⁻ penetration, triggering expansive corrosion products. Third is crack coalescence where synergistic stresses from chemical expansion and freeze-thaw cycles lead to macroscopic cracking. Performance comparison of modified groups post-freeze-thaw evaluations (MLR, RDEM degradation, and microcrack distribution) confirmed RCAC2 (FA-PPF composite modification) as the most effective solution.

Based on the “pore → ion → crack” damage cascade, a graded modification strategy was developed for RCAC2 (FA-PPF), forming a technical pathway with specific parameters and a progressive protection concept:

1. Pore optimization (source control): FA generated nanoscale gels through pozzolanic reactions, filling capillary pores (<10 nm) and converting harmful pores (>100 nm) to benign pores (<50 nm). This increased the benign pore proportion to 73% (vs. 58% in RCAC1). Notably, while RCAC1 achieved 9.5% porosity reduction, its median pore diameter surged by 140.7%, paradoxically increasing permeability by 99.3%. RCAC2 achieved dual optimization: 33.6% median pore reduction and crack width control (<50 μm), resulting in “low porosity-low permeability” synergy.

2. Ion blocking (pathway interruption): FA-induced gels physically restricted Cl⁻/SO₄²⁻ migration toward the ITZ. PPF suppressed early-stage microcracks, preventing ion transport shortcuts. This combination reduced Cl⁻ diffusion coefficients by 51.4%, disrupting the “ion invasion → chemical expansion” cascade and alleviating fiber stress.

3. Fiber barrier (propagation control): when residual ions induced local expansion stress, PPF’s 3D network provided bridging/tensioning effects: fibers spanning microcracks redistributed stress concentrations at crack tips. Fiber constraint maintained crack integrity during freeze-thaw cycles, preventing macroscopic channel formation.

Synergistic model validation: the “pore optimization-ion blocking-fiber barrier” triple mechanism (Figure 12) effectively interrupted the damage cascade. By reducing initial pore vulnerabilities, suppressing ion-driven chemical stresses, and arresting crack propagation, RCAC2 demonstrated superior freeze-thaw resistance. These findings provide a theoretical and experimental basis for durable recycled concrete applications in harsh salt freeze environments.

5. Conclusions

This study comprehensively investigated the damage evolution of recycled coarse aggregate concrete (RCAC) under salt freeze coupling action, simulating the harsh saline soil conditions in Tumushuke, Xinjiang. Through multi-scale characterization and cyclic loading tests, key degradation mechanisms were identified, and a synergistic modification strategy was developed. Major conclusions include:

(1)

Significant Durability Enhancement

After 125 freeze-thaw cycles, the MLR of the FA-PPF modified group (denoted as RCAC2) exhibited an MLR of 0.28% ± 0.02%, significantly lower than the control group (0.87% ± 0.02%, p < 0.01) and the FA-SF modified group (2.30% ± 0.02%, p < 0.001). RCAC2 met the Class Ⅱ durability criteria (mass loss <1%), aligning with the macroscopic damage progression law. Critically, FA-PPF modification extended the durability threshold (defined as RDEM >60%) from 75.98% ± 0.21% at 125 cycles to 63.79% at 200 cycles, fully complying with the requirements of Class Ⅱ in China’s national standard GB/T 50082-2009 Standard for Test Methods of Long-Term Performance and Durability of Ordinary Concrete.

(2)

Pore Structure Optimization

After 75 cycles, the porosity of RCAC2 decreased by 13.7%, primarily attributed to corrosion product filling effects. Conversely, RCAC3 (FA-SF) showed a 9.87% porosity increase due to weakened ITZ. This contrast confirms that RCAC2′s optimized pore structure (refinement + filling) effectively suppressed crack initiation/propagation and reduced permeability, maintaining structural integrity under salt freeze coupling.

(3)

Triple Synergistic Protection Model

The proposed “pore optimization-ion blocking-fiber cracking resistance” model demonstrated remarkable efficacy. FA optimized pore distribution (harmful pores ≤ 9.8%), while PPF restricted microcrack width (<21 μm). This synergy significantly enhanced RCAC’s long-term performance in extreme saline-freeze environments, providing critical theoretical and experimental support for engineering applications in high-cold saline regions.

(4)

Future Work

Subsequent research will focus on two aspects: ① Establishing a graded pretreatment system for recycled coarse aggregates to address source variability, integrating moisture content monitoring and water adjustment. ② Optimizing field mixing, pumping, and vibration processes to facilitate practical implementation. These advancements lay a foundation for expanding RCAC applications in challenging saline soil environments.

Concurrently, throughout subsequent research phases, we will systematically strengthen the documentation of equivalencies between Chinese and international standards. This will involve explicit cross-referencing of test protocols in Chinese standards (e.g., GB/T 50081-2019 and GB/T 50082-2009) with equivalent international norms. Comparative analysis of core parameters (specimen geometry, curing conditions, loading rates, failure thresholds) to demonstrate methodological alignment. Contextual justification for protocol deviations arising from regional environmental requirements (e.g., composite salt solutions in freeze-thaw testing). These measures aim to enhance both the academic rigor of the manuscript and its accessibility to international audiences, ensuring cross-disciplinary comparability while preserving contextual relevance.