Investigation into the individual effects of temperature and sulphate ions on the fracture toughness of waste-based cemented paste backfill containing superplasticizer

1 Introduction

In 2024, the world produced 9.24 billion tons of coal, of which China contributed 4.78 billion tons (Energy Institute, 2025). Coal mining and coal washing processes generate massive coal gangue, accounting for 10%–25% of raw coal output (Wang et al., 2025). Meanwhile, coal combustion discharges a large volume of fly ash. By 2025, China’s cumulative stockpiles of coal gangue and fly ash had reached 7 billion tons and 3 billion tons respectively, with annual newly-accumulated volumes of both wastes exceeding 200 million tons (Junying et al., 2026). The stacking of these wastes on the ground surface not only occupies expansive land, but the spontaneous combustion of gangue will also result in the generation of carbon monoxide, sulfur dioxide, and hydrogen sulfide (Zhang et al., 2022; Deng J. et al., 2017). Another severe issue caused by the high-intensity exploitation of coal resources is the widespread surface subsidence, which affects the stability of man-made structures and destroys the environment (e.g., soil and water loss) (Zhang et al., 2016). It is recorded that the area of the surface subsidence induced by coal mining activities has reached 2.89 × 10⁴ km² (Zhou et al., 2020). Therefore, how to manage the wastes environmentally friendly and mitigate the surface subsidence simultaneously is an important issue.

As an effective way to deal with the wastes and surface subsidence, solid backfill mining technology has been adopted by many coal mines in China (Qiang et al., 2012). However, owing to its intrinsic drawbacks, especially the competing effects between the optimum strata movement controlling effect (which requires repetitive compaction and loading processes) and the demand for high productivity of mines, the application is still limited. For certain mines with specific geological conditions, it is necessary to manage the wastes underground for the sake of safety, but the solid backfilling and mining technology is just not applicable. For example, there is a coal mine called Gonggeyingzi, located in Inner Mongolia, with an average coal seam thickness of 21 m. Besides, a river known as Laoha River flows above the coalfield, and the aquifer is located 52.3 m above the coal seam (Deng X. et al., 2017). The major challenge is to ensure the mining-induced fracture does not propagate to the aquifer, and this requires an extremely high backfilling ratio in terms of solid backfill mining technology, which is basically impractical. Inspired by the cemented paste backfill (CPB), WCPB is thereafter proposed, and it is consists of crushed gangue, fly ash, binder, and water (Li et al., 2022). The significant difference between WCPB and solid backfilling and mining technology is that the former can ensure a higher backfilling ratio, and in return, a better strata movement controlling effect is achieved. The Gonggeyingzi coal mine divided the coal seam into several layers and finally exploited the coal resource with WCPB safely.

Currently, extensive research has been conducted on the effects of graphene, nano-silica, and steel fiber content on the mechanical strength of cement-based concrete (Jahami et al., 2024; Jahami, , 2026; Ali et al., 2023; Hama et al., 2024). However, in the context of backfill mining, WCPB must be pumped into goafs through long-distance pipelines (Kun and Mamadou, 2018). Therefore, when determining the mix ratio of WCPB, its flowability and mechanical strength should be considered comprehensively (Fang and Fall 2019). To enhance one property of CPB without affecting another property of it, many investigations have been carried out to explore how chemical compounds influence the mechanical behavior of backfilling materials. (Ercikdi et al., 2010; Ouattara et al., 2018; Huynh et al., 2006; Mangane et al., 2018; Guo et al., 2021). Among these, superplasticizers have proven to be an effective additive, which can allow for a 12%–30% decrease in the amount of water added without impairing the material’s consistency and workability (Haruna and Fall 2020; Al-Moselly et al., 2022). The main mechanism is that superplasticizers increase the repulsive force between various particles, thereby improving the workability of CPB (Yang et al., 2018). On the other hand, it is concluded that the mixing of superplasticizers into CPB during the production process benefits the strength acquisition as well (Mangane et al., 2018; Haruna and Fall 2021), and an improvement of 20%–50% in strength is observed after the addition of superplasticizers in the research carried out by Ercikdi et al. (Ercikdi et al., 2010).

Despite the contribution made by the aforementioned scholars to the understanding of the superplasticizers on the strength development of CPB, the effects of superplasticizers on the mechanical properties of WCPB may vary, considering the different ingredients. Besides, the mechanical property in previous studies usually refers to the uniaxial compressive strength (UCS) of the backfilling body. There is still a lack of research on the effect of superplasticizers on the fracture toughness of WCPB. Actually, fracture toughness is a key characteristic to assess the stability of structures subjected to dynamic loading. Crack initiates and propagates when the stress intensity factor (SIF) reaches the fracture toughness, which includes three modes (mode I, II, and III) depending on the loading conditions (Kataoka et al., 2015; Camanho et al., 2003). Considering the failure modes of backfilling structures usually encountered underground (tensile and in-plane shear failure), K _IC and K _IIC are the two major mechanical parameters. In addition, due to the existence of cementitious components, K _IC and K _IIC of WCPB are time-dependent. Moreover, by affecting the cement hydration, the complex underground conditions also make the evolution of fracture toughness complicated (Helinski et al., 2010; Ghirian et al., 2017).

To date, numerous scholars have performed experimental investigations on the fracture toughness of CPB considering various crack notch parameters, fiber content and cement content (Libos and Cui, 2020; Xu and Cao, 2018; Fang and Cui, 2022; Fang and Cui, 2023; Cui and McAdie, 2023). However, the respective influence mechanisms of temperature and sulfate on K _IC and K _IIC of superplasticizer-incorporated WCPB remain to be systematically clarified. Therefore, the present study implemented a range of macro- and micro-scale trials to improve understanding of the evolutionary characteristics of K _IC and K _IIC in WCPB containing superplasticizer.

2 Experimental program

2.1 Materials used

The WCPB samples for this study were prepared using coal-based solid waste, cement, water, superplasticizer, and ferrous sulphate heptahydrate (FeSO₄.7H₂O). The respective influences of dosage of superplasticizer and FeSO₄.7H₂O, as well as the temperature, on the K _IC and K _IIC of WCPB were examined systematically.

2.1.1 Binder and coal mine waste

For this study, Portland cement (P.O. 42.5) produced by Jiuqi Building Materials Co., Ltd., Zhucheng City, was used as the binder. In the solid phase system, the mass fractions of the raw materials were designed as follows: the cement content was 8 wt%, the fly ash content was 20 wt%, and the remaining 72 wt% consisted of coal gangue, which was crushed to less than 2 mm using a jaw crusher. The particle size distribution (PSD) of the fine gangue (0–1 mm fraction), fly ash, and cement was determined using a laser particle size analyzer. Figure 1 displays the PSDs generated as a result of the above analysis. The chemical components of each raw material are listed in Table 1.

TABLE 1. The chemical components of each raw material.

2.1.2 Chemical admixture

Superplasticizers produced in Laiyang, Shandong, can improve the fluidity of the slurry and enhance the strength of the hardened backfill. To investigate their effect on the fracture toughness of WCPB, different dosages of superplasticizer were added to the slurry before the mixing process, and the superplasticizer contents were designed as 0%, 0.05%, 0.125%, and 0.25% (Fang et al., 2026).

FeSO₄.7H₂O was selected as the source of sulfate ions (SO₄ ^2-) due to its high solubility. Another key reason for this choice is that the divalent cation (Fe²⁺) in FeSO₄.7H₂O aligns with the oxidation products of sulfide-rich minerals, thereby effectively simulating underground chemical environments (Fang et al., 2026). Based on preliminary test results, the specimens failed to bind when the FeSO₄.7H₂O concentration exceeded 10,000 ppm. Therefore, the sulfate concentrations for this experiment were established at 0 ppm, 2,500 ppm, 5,000 ppm, and 10,000 ppm.

2.2 Sample preparation

The semi-circular bending (SCB) method was used to measure the fracture toughness of the materials. A 3D model of the molds was first designed using SolidWorks software, as shown in Figure 2. This model consisted of a base plate, a collar, and dividing plates designed for specific modes: Mode I (with the short plate perpendicular to the long plate) and Mode II (with the short plate angled at 45° to the long plate). Key dimensions included: the thickness of the plates is 2 mm, the length of the short plate a is 20 mm, the radius R and thickness T of specimen are 50 mm and 20 mm, respectively.

To prepare the samples, a mixer was used to blend gangue, fly ash, cement, and a defined amount of superplasticizer. Thereafter, distilled water with varying sulfate concentrations was added to the blend to adjust the slurry concentration to 67 wt%. The mixture was then stirred for 7 min before being poured into the molds. Subsequently, manual vibration was applied to remove the trapped air, and the whole molds was wrapped with plastic film to prevent moisture evaporation. The prepared samples were cured in a curing chamber (with a humidity of 98%) for 3 days, 7 days, 14 days, and 28 days. As for the effect of temperature on the fracture toughness of WCPB, the thermal condition of the chamber was adjusted to 2 °C, 20 °C, and 35 °C (Table 2), with three replicates prepared for each testing condition.

TABLE 2. Experiment details.

2.3 Testing and monitoring program

2.3.1 Setup of SCB test and match ID measuring system

Before the SCB test, a viscous slurry was prepared by mixing calcium carbonate with water. This slurry was then painted on the surface of samples to create a white background (Figure 3A). Thereafter, irregular speckles (monitoring points) were marked on the specimens (Figure 3B) using a black marker pen. These speckles served as the tracking points for the Match ID full-field strain measurement system. The test equipment used was a WDW-300 universal testing machine manufactured by Jinan Xinshijin Testing Machine Manufacturing Co., Ltd. A representative specimen after testing is shown in Figure 3C, and the schematic testing diagram is illustrated in Figure 3D. While loading the WCPB (at a loading rate of 0.1 mm/min), the Match ID full-field strain measurement system was used to monitor crack propagation. The camera was synchronized to move vertically in tandem with the loading platform.

The distance between the two loading points is marked as 2 S. It is well-established that when measuring K _IC, specimens must possess a notch orientation angle α′ = 90. Conversely, K _IIC is achieved under different notch angles, which are closely related to S/R, and a/R, as documented in numerous studies (Cui and McAdie, 2023; Ayatollahi and Aliha, 2007). In this study, the notch angle, S/R, and a/R were specifically designed as 45°, 0.5, and 0.4, respectively. During the test, as shown in the figure, a normal load (which increases until failure occurs) is applied to the top of the specimen, while two corresponding point loads are applied at the bottom supports. The formulas for calculating K _IC and K _IIC of WCPB are given by (Ayatollahi and Aliha, 2007; Wong et al., 2019), and the mean and standard deviation of three parallel specimens were calculated for each condition.

Where Y _I and Y _II are the normalized SIF for Mode I and Mode II, respectively. For the geometric configuration adopted in this study, Y _I = 2.9050 and Y _II = 0.9256.

The geometric parameters of the present SCB test and the calculation method of normalized SIF refer to the studies by Fang et al. and Ayatollahi et al. (Fang and Fall 2020; Ayatollahi and Aliha, 2004).

2.3.2 Microstructural analyses and monitoring program

To reveal the mechanism by which various factors affect the K _IC and K _IIC of WCPB, a range of microscale experiments was implemented on the samples. These included: Electrical conductivity (EC), volumetric water content (VWC) monitoring, Scanning electron microscopy (SEM), Thermogravimetric (TG) and derivative thermogravimetry (DTG) analysis, Mercury intrusion porosimetry (MIP). For all microstructural specimens except those subjected to EC and VWC monitoring, a drying oven was used to evaporate all moisture from the samples, thereby halting the cement hydration process.

EC and VWC monitoring were performed using 5 TE sensors to measure the electrical conductivity and volumetric water content of the slurry, respectively. Data acquisition was accomplished with an EM50 data logger. The VWC values serve as indicators of the degree of cement hydration, while EC values provide critical metrics for assessing the rate of cement hydration reactions. MIP was performed using a Micromeritics AutoPore IV 9520 instrument. The measurement principle relies on the differential intrusion of mercury into pores under varying pressure, where pore size distribution is derived from applied pressure-intrusion volume relationships. SEM analysis (ZEISS, Sigma 300) enabling direct visualization of the pore structure in WCPB and qualitative assessment of hydration product distribution. TG was carried out on a Netzsch STA 449 F⁵ simultaneous thermal analyzer. Mass changes were monitored in real-time while heating specimens from ambient temperature to 1,000 °C (10 °C/min, nitrogen atmosphere). By referencing established decomposition temperatures of hydration products, quantitative analysis of hydration products under different conditions was performed.

3 Results and discussions

3.1 Evolution of the load-displacement curves of WCPB containing superplasticizers

Figure 4 illustrates the evolution of load-displacement curves from three-point bending tests for sulphate-free specimens containing 0.125% superplasticizer (with a curing temperature of 20 °C). An increasing load is observed with time. Specifically, peak loads of Mode I and Mode II samples increase from 28 N to 38 N at 3 days to 86 N and 118 N after being cured for 28 days. This is mainly due to cement continuously undergoing a hydration reaction over time, thereby continuously producing calcium silicate hydrate (C-S-H), as confirmed by TG results for 0.125% superplasticizer specimens cured for 3 and 28 days (Figure 4). Progressive mass loss of specimens during heating is widely acknowledged., with three distinct mass loss rate peaks evident in the DTG curves occurring at 30 °C–150 °C, 400 °C–600 °C, and 620 °C–740 °C. These peaks correspond to the decomposition of C-S-H, Ca(OH)₂, and CaCO₃, respectively (Xin et al., 2022; Wang et al., 2016). The mass loss rate at 30 °C–150 °C for 28-day specimens significantly exceeded that of 3-day specimens, confirming higher C-S-H content in mature samples, which directly enhances the fracture toughness. Secondary strengthening arises from pore-filling effects, hydration products such as CaCO₃ and ettringite (AFt, 3CaOAl₂O₃.3CaSO₄.32H₂O) particles occupy voids between raw materials, densifying the microstructure while providing additional nucleation sites for C-S-H (Fang and Fall 2020). Finally, as free water within the specimen gradually diminishes over time, the effective stress in the WCPB increases (Fang et al., 2022). This causes a boost in the brittleness of the WCPB, where a small loading displacement is sufficient to make the SIF reach K _IC and K _IIC. Consequently, the critical displacement corresponding to specimen failure reduces as the curing time increases.

Moreover, in Figure 4, the peak load of Mode II specimens is significantly higher than that of Mode I specimens, and the former requires more time to reach the peak load. However, the load-displacement curves for both modes exhibit a linear increase to the peak load, followed by a sudden drop. This phenomenon can be explained by the contour maps of WCPB for the two modes shown in Figures 5C, 6C. Usually, a crack initiates upon reaching the peak load, splitting the specimen into two-halves. In terms of the Mode I sample, the fracture initiates at the pre-existing notch apex and propagates toward the loading position situated at the top of the specimen. Due to material heterogeneity (Wong et al., 2019; Sun et al., 2021), the propagating crack in Mode II exhibits deflection. Relative to the Mode I specimen, a longer path length is observed in Mode II fractures, and this phenomenon demonstrates that additional energy accumulation is required to bring about material failure during Mode II loading. Additionally, the generation of a Mode II crack involves partial load dissipation to overcome energy loss due to potential frictional sliding along the crack propagation surface (Nejati et al., 2021; Lin et al., 2019). Crack initiation occurs when the energy accumulated within the specimen exceeds the material’s capacity to withstand it (Nejati et al., 2021). The crack initiation energy is calculated based on the area enclosed by the load-displacement curve prior to the peak load (Cui and McAdie, 2023). Therefore, WCPB for Mode II measurement exhibits a higher peak strength and a longer loading displacement prior to failure.

To more intuitively observe the crack propagation behavior in WCPB, a strain measurement line was established using the Match ID system, extending from the notch tip to the loading point (Figures 5A, 6A). Figures 5B, 6B present real-time strain data along the measurement lines. During the monitoring period from pre-fracture to post-fracture, the notch tip strain of the Mode I sample increased from 9.24347 × 10⁻⁴ to 3.6 × 10⁻², while that of the Mode II sample increased from 8.03 × 10⁻⁴ to 4.8 × 10⁻². Notably, the strain at loading point P remained virtually unchanged in both modes. Due to the heterogeneous property of WCPB, the Mode II fracture propagated starting at the man-made notch tip and extending toward the loading point (Figure 6C), and this accounts for the comparatively lower strain observed in Region A of Figure 6B.

3.2 Effect of superplasticizer content on the fracture toughness of WCPB

Equations 1 and 2 were employed to calculate the K _IC and K _IIC of WCPB specimens with the obtained experimental results. Figure 7 illustrates the evolution of K _IC and K _IIC of sulphate-free samples containing different amounts of superplasticizer with time (with a curing temperature of 20 °C). The evolution trends of K _IC and K _IIC over time are consistent with the changes in peak load. However, K _IIC is generally lower than K _IC. This stems from the fact that Mode II shear loading induces a magnified local tensile stress field ahead of the notch tip; given the intrinsically low tensile strength of WCPB, this stress concentration triggers premature tensile crack initiation rather than pure shear sliding (Nejati et al., 2021; Lin et al., 2019; Feng et al., 2018).

It can also be observed that K _IC and K _IIC decrease with increasing dosage of superplasticizer within the first 14 days of WCPB curing. This reduction originates from superplasticizer molecules adsorbing onto cement particle surfaces. This process generates a steric hindrance effect, which enhances the mutual dispersion of cement particles and consequently slows down the cement hydration reaction (Zhu et al., 2021; Matsuzawa et al., 2019). This assertion is substantiated by TG results for specimens with different superplasticizer dosages (Figure 4). At a curing age of 3 days, the mass loss corresponding to the dehydration and decomposition of C-S-H in the sample without superplasticizer was 1.222%, while that of the sample incorporating 0.125% superplasticizer was 1.101%. Taking the group without superplasticizer as the reference, the relative amount of generated C-S-H in the superplasticizer-containing sample decreased by approximately 9.90%.

However, the K _IC and K _IIC of WCPBs containing 0.05% and 0.125% superplasticizer surpassed those of the superplasticizer-free WCPBs after being cured for 28 days. This enhancement can be attributed to the more uniform spatial distribution of cement particles within the superplasticizer-modified WCPB, facilitated by the steric hindrance effect. This uniform distribution promotes the generation of additional cement hydration products, ultimately leading to a denser microstructure. This assertion is substantiated by MIP analysis results from specimens with varying superplasticizer dosages (Figure 8). As illustrated in Figure 8A, when the pore diameter exceeds 226.9 nm, the cumulative porosity of the WCPB specimen modified with 0.125% superplasticizer is slightly higher than that of the control (superplasticizer-free) specimen. However, for pore diameters smaller than 226.9 nm, the cumulative porosity of the control specimen becomes substantially larger than that of the 0.125% superplasticizer-modified specimen. To quantitatively evaluate the pore refinement effect induced by the superplasticizer, the pores were classified into four ranges according to literature (Zhang et al., 2026): highly harmful pores (>0.2 µm), harmful pores (0.05–0.2 µm), slightly harmful pores (0.02–0.05 µm), and harmless pores (<0.02 µm). The comparative analysis is presented in Figure 8B. Both specimens exhibit nearly identical volumes of highly harmful pores. Notably, the volume fraction of harmful pores in the 0.125% superplasticizer-modified specimen is only 23%, which is markedly lower than that of the control specimen. Furthermore, the 0.125% superplasticizer-modified specimen exhibits a reduced volume of slightly harmful pores and a higher proportion of harmless pores compared to the control. These results quantitatively demonstrate that an appropriate dosage of superplasticizer effectively refines the pore network of WCPB, thereby enhancing its long-term fracture toughness.

3.3 Contribution of temperature to the fracture toughness of WCPB

Figure 9 shows how K _IC and K _IIC evolve over time for sulphate-free specimens containing 0.125% superplasticizer, under different curing temperatures (2 °C, 20 °C, 35 °C). When the curing age was extended from 3 to 14 days, the fracture toughness of WCPB exhibited an increasing trend due to enhanced cement hydration, regardless of the curing temperatures. Furthermore, the rate of fracture toughness improvement accelerated with increasing temperature. For example, the K _IIC increased from 0.12 kPam^0.5, 0.18 kPam^0.5, and 0.25 kPam^0.5 to 0.17 kPam^0.5, 0.30 kPam^0.5, and 0.45 kPam^0.5 for samples cured at 2 °C, 20 °C, 35 °C, with an improvement of 41.7%, 66.7%, and 80%, respectively.

The mechanisms underlying the influence of increased temperature on the cement hydration process are well-documented in the literature (Lothenbach et al., 2007). First and foremost, higher temperatures intensify ion movement. This accelerates the hydration reaction, leading to increased gel production and greater consumption of free water. Consequently, this generates higher effective stress within the material. Secondly, the elevated temperature induces a higher osmotic pressure differential, and this differential promotes the transport of water molecules through the interstitial spaces within the hydration product layers, thereby facilitating the continuation of the cement hydration (Liu et al., 2023). This assertion is substantiated by the variations in VWC and EC observed in specimens cured at different temperatures (Figure 10). The VWC curves characterize how quickly cement hydration reactions proceed by reflecting the consumption rate of free water within the specimen, while the EC curves indicate both how fast cement particles dissolve and how rapidly hydration products form. In Figure 10, both VWC and EC curves exhibit distinct rising and declining phases. The rising phase of VWC results from percolation of free water toward the sensor, whereas the declining phase is owing to free water being consumed during cement hydration. Concurrently, the rising phase of EC is caused by cement particle dissolution, and the declining phase is caused by the formation of cement hydration products. The sequential occurrence of VWC peak times (35 °C, 20 °C, 2 °C) and progressively faster decline rates directly demonstrates accelerated hydration under elevated temperatures. This is further confirmed by EC peak times at 4.2 h (35 °C), 8.3 h (20 °C), and 21.6 h (2 °C), providing additional evidence for the temperature-dependent enhancement of hydration kinetics.

As the curing age of WCPB increased from 14 days to 28 days, the fracture toughness of specimens cured at 2 °C and 20 °C continued to increase. For example, for samples cured at 20 °C, K _IC and K _IIC rose from 0.62 kPam^0.5 and 0.30 kPam^0.5–1.16 kPam^0.5 and 0.53 kPam^0.5, respectively. In contrast, the fracture toughness of 35 °C -cured WCPB decreased with prolonged curing and was surpassed by the 20 °C specimens. Similarly, when Fang et al. investigated the fracture toughness of WCPB across various sulfate concentrations, a consistent trend was observed: the fracture toughness was higher for early-age specimens cured at 35 °C, whereas it became higher for late-age specimens cured at 20 °C (Fang et al., 2026). Based on their SEM and MIP analyses, this phenomenon can be attributed to the non-uniform distribution of hydration products induced by excessively rapid early hydration, which consequently generated internal microcracks. Furthermore, the localized accumulation of dense hydration products encapsulated unreacted particles, thereby hindering subsequent cement hydration. Furthermore, Lothenbach et al. (Lothenbach et al., 2008) demonstrated that high curing temperature coarsens the pore structure of specimens, leading to reduced fracture toughness.

3.4 Effect of sulphate ions on the fracture properties of WCPB

Figure 11 presents the evolution of fracture toughness of differently sulphated samples containing 0.125% superplasticizer (with a curing temperature of 20 °C). For specimens cured for less than 14 days, sulphate-free WCPB exhibited higher fracture toughness than sulphated specimens, and the fracture toughness decreased with increasing sulphate concentration. When the curing age of specimens reaches 28 days, the fracture toughness of specimens containing 2,500 ppm sulphate exceeds that of sulphate-free specimens. The K _IC values of the 2,500 ppm and 0 ppm specimens are 1.33 kPam^0.5 and 1.16 kPam^0.5, respectively. This reversal occurs because sulphate ions in early-age specimens preferentially react with C₃A (3CaOAl₂O₃) to form AFt (Emmanuel et al., 2022), which in turn leads to relatively limited formation of C-S-H. Moreover, the generated AFt usually accumulates at C₃A active sites, and this further inhibits C-S-H generation (Minard et al., 2007). However, with the passage of time, cement and free water progressively deplete in sulphate-free specimens. Consequently, the growth rate of fracture toughness slows down. In contrast, the amount of C-S-H produced in sulphated specimens gradually catches up to that in sulphate-free specimens. Moreover, owing to the refinement of the microstructures by pre-formed AFt, samples with sulphate content of 2,500 ppm depict higher fracture toughness.

This phenomenon is corroborated by the SEM results presented in Figure 12. As observed, the sulphate-free specimens with an age of 28 days exhibit substantial C-S-H but minimal AFt, resulting in numerous interparticle pores (red-circled areas). On the contrary, 10,000 ppm specimens display extremely dense microstructure yet less C-S-H. However, the specimens with 2,500 ppm sulfate contained both abundant AFt and C-S-H. Specifically, AFt refined their pore structure, while C-S-H enhanced the bonding force between particles—two effects that collectively contributed to their highest fracture toughness. Therefore, the fracture toughness of WCPB is affected by a dual mechanism of sulfate ions (enhancement or reduction, determined by the amount of sulphate ions). Adding an appropriate amount of sulfate to the material can enhance the fracture toughness of advanced-age specimens. In contrast, excessive sulfate ions generate surplus AFt, which increases the number of microcracks and thereby reduces the fracture toughness (Jun et al., 2024).

4 Summary and conclusions

With a set of SCB tests and microscale experiments, this study investigated how sulphate concentration and temperature respectively affect the K _IC and K _IIC of superplasticizer-containing WCPB, and the following conclusions were made.

• With the extension of curing time, the fracture toughness usually depicted an increasing trend owing to the strengthened cement hydration, except for the 28-day-cured specimen at 35 °C.
• For specimens cured for less than 14 days, the dispersive effect of superplasticizer inhibited the cement hydration, and K _IC and K _IIC decreased with increasing dosage of superplasticizer. However, at 28 days, an optimal dosage (0.05%, 0.125%) contributed to a uniformly distributed cement hydration product and a denser microstructure, leading to a higher fracture toughness.
• When the curing duration of WCPB prolonged from 3 days to 14 days, elevated temperatures accelerated the process of cement hydration. However, upon reaching 28 days of curing, specimens cured at 35 °C exhibited reduced fracture toughness because of the uneven distribution of cement hydration products, encapsulation of unreacted cores, and coarsened pore structure.
• For early-age specimens, sulphate ions reacted with C₃A to form AFt, inhibiting hydration progress and consequently reducing the fracture toughness. However, after 28 days of curing, the fracture toughness of the 2,500 ppm specimen surpassed the sulphate-free samples, as it produces more AFt to fill the pores.

This study primarily investigated the evolutionary patterns of fracture toughness in WCPB containing superplasticizers under the respective influences of temperature and sulfate ions within a 28-day curing period. While this research provides a valuable reference for the mix design of WCPB exposed to saline mine water, the long-term service performance of WCPB remains paramount for practical engineering applications. Therefore, future studies incorporating extended curing ages (e.g., 90 and 180 days) are highly warranted to fully validate the durability and long-term evolutionary mechanisms of WCPB under these complex environmental conditions. Furthermore, it is necessary to establish systematic mathematical correlations between TG and MIP test results and fracture toughness in future investigations.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Bolin Xiao, University of Science and Technology Beijing, China

Reviewed by: Penghui Wen, Chang’an University, China

Ali Jahami, University of Balamand, Lebanon