Slag Substitution Effect on Features of Alkali-Free Accelerator-Reinforced Cemented Paste Backfill

Abstract

Cemented paste backfill (CPB) improves underground stability by filling mine voids, but the high cost of cement presents economic challenges for miners. While alternative binders and admixtures have been explored, the combined impact of slag substitution and alkali-free (AF) accelerators on CPB performance is not yet fully understood. This study investigates the influences of slag substitution and AF accelerators on the performance of CPB through a comprehensive experimental approach. CPB samples were prepared with slag substitution ratios of 25%, 50%, and 75%, maintaining a fixed AF accelerator content of 0.4%. Various test techniques, including unconfined comprehensive strength (UCS), mercury intrusion porosimetry (MIP), X-ray diffraction (XRD), and thermal analysis (TG/DTA), were employed to study their mechanical and microstructural properties. Monitoring tests were also conducted to thoroughly assess the performance of CPB, including suction (self-desiccation), electrical conductivity (EC), and volumetric water content (VWC) tests. The results showed that the PCI50–SL50–0.4AF sample exhibited 2.3 times higher strength than the control sample for 28 days, with this improvement attributed to enhanced pozzolanic reactions contributing to better microstructural compactness. Monitoring tests revealed accelerated hydration kinetics and reduced water content in slag-reinforced CPB, highlighting the significant role of AF accelerator in facilitating rapid setting and improving early-age mechanical strength. Microstructural findings revealed that porosity decreased and C–S–H gel formation increased in the specimen containing slag and AF accelerators, contributing to increased strength and durability. These findings highlight the potential usage of slag and AF accelerators to enhance CPB’s mechanical, microstructural, and hydration properties, offering significant benefits for mining operations by improving backfill performance, while contributing to environmental sustainability through reduced cement consumption and associated CO₂ emissions.

1. Introduction

Cemented paste backfill (CPB) constitutes a backfill material engineered by combining fine-grained mine tailings generated from mineral processing operations with water and binding agents [1]. Employed to fill voids resulting from ore extraction in underground mines, CPB facilitates the safe and environmentally responsible management of mine tailings while mitigating the risk of environmental damage [2,3]. Furthermore, CPB effectively stabilizes the ground, enhancing both the structural integrity of mining operations and the safety of the working environment [4,5]. A significant advantage of CPB is its contribution to sustainable mining practices by promoting the reuse of tailings materials [6,7,8]. The incorporation of tailings into engineered designs diminishes the requirement for disposal methods and facilitates the management of environmental impacts at mine sites [8,9,10].

Despite its advantages, CPB faces several critical challenges that limit its broader application, primarily stemming from economic, environmental, and technical factors. Economically, the reliance on ordinary Portland cement (OPC) as a binding agent significantly escalates backfill costs, constituting up to 75% of CPB production expenses and contributing nearly 20% to the overall operational costs in mining operations [11,12,13,14,15]. Environmentally, Portland cement production is a major contributor to CO₂ emissions, accounting for approximately 8% of global emissions, with each ton of cement produced releasing 0.9 tons of CO₂, thereby exacerbating climate change concerns [16,17,18]. Technically, CPB is vulnerable to sulfate and acid exposure, often arising from the oxidation of pyritic tailings. These reactions degrade the cement hydration products, weaken the backfill’s structural integrity, and precipitate ettringite, leading to significant strength and durability losses [19,20,21,22,23]. These multifaceted challenges underscore the necessity of developing and incorporating alternative binders to enhance CPB’s economic viability, environmental sustainability, and technical performance.

As seen, OPC within CPB is weak in terms of economic properties, environmental issues, strength, and durability, prompting miners to seek alternative products to cement. Therefore, using these materials with pozzolanic properties, such as metallurgical wastes, as binders in specific proportions is crucial to mitigate the disadvantages of the use of OPC and to explore alternative secondary materials [24,25,26]. Advancements in CPB technology have driven significant research into incorporating alternative materials such as slag to improve mechanical, microstructural, and rheological properties while addressing economic and environmental challenges. For instance, Xiao et al. [14] and Kou et al. [27] investigated the effects of active slag on the rheological properties of CPB, particularly focusing on the impact of sodium chloride and sulfate admixtures. Similarly, Zhang et al. [28] analyzed the mechanical behavior and rheological properties of foamed concrete backfill using hydrogen peroxide. Researchers such as Li et al. [29] and Cihangir et al. [30] examined the mechanical strength and microstructural properties of slag-blended CPB with the addition of activators like sodium sulfate and sodium silicate. Other studies, including those by Zhao et al. [31] and He et al. [32], evaluated the influence of slag-blended cement on the mechanical and rheological performance of CPB. These studies collectively demonstrate the effective utilization of slag both as a binding component and as an ingredient in blended cement for CPB, resulting in improved strength, rheology, and microstructure. However, a notable research gap exists in exploring the combined application of AF accelerators with slag in CPB. AF accelerators are environmentally friendly admixtures commonly used in shotcrete applications, particularly in mining and tunneling. Considering the well-documented benefits of AF accelerators in improving concrete and shotcrete properties, investigating their integration with slag in CPB formulations may offer a promising avenue for addressing traditional Portland cement’s economic, environmental, and technical limitations.

To address the research gap mentioned above, this study comprehensively investigates the combined effects of slag substitution and AF accelerators on the performance of CPB. In this context, a detailed experimental approach was utilized to evaluate CPB mixtures’ mechanical, microstructural, and hydration properties, incorporating various slag substitution levels and constant AF ratios. By integrating advanced analyses such as UCS, XRD, and MIP, alongside monitoring tests, this work aims to provide deeper insights into optimizing CPB formulations for enhanced performance and sustainability in mining operations.

2. Materials and Methods

2.1. Materials

CPB specimens were prepared using a binder (cement and slag), silica-based tailings (SBTs), water, and a setting accelerator as the admixture for the experimental investigations. The binder materials and SBTs were sourced from a cement plant located in Gümüşhane, Turkey. The chemical admixture was obtained from BASF Türkiye.

The grain size distribution of binders and SBTs was analyzed using a Malvern Hydro 2000 MU (Malvern Panalytical, Worcestershire, UK) laser particle size analyzer. This device determines particle size distribution by analyzing the laser light scattering pattern from particles in the sample. The analyses were conducted with a refractive index of 1.6 and an ultrasonic mixing time of 5 min, and the results were averaged over three replicates. The chemical compositions of the binders were also analyzed using an ARL 9900 Series XRF Spectrometer (ThermoFisher Scientific, Waltham, MA, USA). This device determines the sample’s chemical composition by measuring the energy of the incident X-rays and the characteristic X-rays emitted as a result of atomic excitation. The equipment used for these analyses is shown in Figure 1.

2.1.1. Hydraulic Binders

This study utilized Type I ordinary Portland cement (OPC, designated as PCI) and slag as binder materials.

Table 1. Physical properties and chemical content of binders (PCI and slag).

Element	PCI (%)	Slag (%)
SiO2	19.02	34.32
Al2O3	4.64	9.54
Fe2O3	3.09	-
CaO	62.98	40.84
MgO	2.89	9.79
SO3	3.03	3.91

presents the chemical compositions along with key physical properties, including density and specific surface area, of PCI and slag. Based on the chemical compositions, the hydration modulus (H_m) and the basicity coefficient (K_b) of the slag can be calculated using the relevant formulas (Equations (1) and (2), respectively) [10]. The equations are as follows:

$$H_m={\displaystyle \frac{\left(\mathrm{C}\mathrm{a}\mathrm{O}+\mathrm{M}\mathrm{g}\mathrm{O}+{Al}_2{O}_3\right)}{\left({SiO}_2\right)}}$$

(1)

$$K_b={\displaystyle \frac{\left(\mathrm{C}\mathrm{a}\mathrm{O}+\mathrm{M}\mathrm{g}\mathrm{O}\right)}{\left({SiO}_2+{Al}_2{O}_3\right)}}$$

(2)

It is reported that the hydration modulus (H_m) must exceed 1.4 to ensure favorable hydration properties in slags [33]. In the calculation made with the relevant formula, the H_m value of the slag used in the study was 1.76. This value indicates that the slag is highly active and has a high hydrolysis potential. It also shows that it can easily react with water and gain binder properties. The basicity coefficient of the slag was calculated as K_b = 1.15, which indicates that the slag is classified as basic slag.

The detailed grain size distribution data of solid materials are presented in Figure 2. The measurements revealed specific surface areas of 3.45 m²/g for PCI and 5.25 m²/g for slag, indicating that slag possesses a finer texture than cement. Additionally, slag, characterized by a high coefficient of uniformity (C_U), exhibits a broad particle size distribution, contributing to enhanced compaction and mechanical strength when utilized as a backfill material. Conversely, PCI cement, with a lower C_U indicative of a more uniform particle size distribution, enhances backfill stability through its binding properties. The coefficient of curvature (C_C) values for both materials, being close to 1, reflect a well-graded and balanced particle size distribution, which positively influences the homogeneity and structural integrity of the backfill.

2.1.2. Tailing

In this study, artificially produced silica-based tailings (SBTs) were preferred for preparing CPB mixtures. SBTs are used to avoid the potential effects of reactive components in natural tailings on cement hydration and, thus, uncertainties in test results. Natural tailings may interfere with cement hydration due to reactive minerals in their composition, which may reduce the reliability of the results. On the other hand, artificial silica-based tailings offer more homogeneous and controllable properties, increasing the consistency of test results [34,35]. In

Table 2. Chemical content of SBTs.

Element	SBTs (%)
SiO2	99.8
Al2O3	0.05
Fe2O3	0.035
K2O	0.02
TiO2	0.02
CaO	<0.01
MgO	<0.01
Na2O	<0.01

, the chemical composition of SBTs is presented as provided by the supplier. The chemical composition of ST is almost entirely quartz. The grain size distribution and some physical properties of SBTs are also provided in Figure 2. The grain size distribution of SBTs reveals a relatively broad distribution, as indicated by the higher Cu = 16.75, suggesting significant variability in particle sizes compared to PCI and slag.

2.1.3. Water and Accelerating Admixture

In this study, the accelerator was used as a liquid-form AF accelerator. Tap water was used as water. While accelerator admixtures are widely used in shotcrete applications, they are also preferred in underground backfilling in recent years [36,37]. AF admixtures, primarily comprised of alkanolamines and aluminum salts, mitigate issues such as durability concerns and reductions in long-term strength, making them more advantageous than alkali-containing accelerators. Some properties of the AF admixture with beige to white color are listed in

Table 3. Some properties of AF accelerator.

Form	Density at 20 °C	pH (1:1 Water Solution)	Viscosity	Chloride Content	Na2Oe, by Mass
Suspension	1.47 ± 0.30 kg/L	2.7 ± 0.5	750 ± 250 mPa.s	<0.1%	<1%

. However, the solid content of the accelerator admixture was 43.7%, its Al₂O₃ content was 7.5%, and its SO₄²⁻ content was 26.1%.

2.2. Sample Preparation

CPB batches, SBTs, binders, AF additives, and water were prepared by combining them in a mixer, and then, CPB samples were formed from these batches to be subjected to various tests (Figure 3). In all CPB batches prepared using the mixer, the binder proportion was set at 4.5% relative to the total mass of tailings, while the AF dosage was fixed at 0.4% based on the solid content. Previous studies have shown that the accelerator admixtures when used above 0.4% increase the porosity and reduce the strength of CPB [37,38,39]. Therefore, 0.4% was determined as an optimum value to ensure sufficient curing time and maintain the desired strength of the mixture [40,41,42]. Blast furnace slag was used in the mixtures by replacing cement at 25%, 50%, and 75% by weight. The target slump value for all CPB batches was set at 20 cm to guarantee flowability, and the materials were adjusted accordingly to achieve this.

Table 4. Mixture recipes used in the preparation of CPB batches.

Mix Name	Admixture Content	Binder Content	Slag Content	Solid Content	Water/Binder Ratio	Slump
	(wt. %)	(wt. %)	(wt. %)	(wt. %)	(w/b)	(cm)
Control	-	4.5	0	76.22	7.25	≈20
PCI100-SL00-0.4AF	0.4	4.5	0	75.98	7.25	≈20
PCI75-SL25-0.4AF	0.4	4.5	25	75.98	7.25	≈20
PCI50-SL50-0.4AF	0.4	4.5	50	75.98	7.25	≈20
PCI25-SL75-0.4AF	0.4	4.5	75	75.98	7.25	≈20

summarizes the mixing ratios of the CPB samples prepared in the laboratory.

The dry solid materials (tailing + binder) were first mixed in the mixer container to ensure homogeneity while preparing the CPB samples. Water was gradually added to the solid materials and blended in the mixer for 7 min to achieve a homogeneous CPB mixture. The setting accelerator admixture was added to the mixture at the end of 7 min and mixed for 9 min to form CPB samples. The blended materials in the mixer were transferred into cylindrical molds with dimensions of 5 cm in diameter and 10 cm in height. The specimens were subsequently kept in a controlled laboratory setting at ambient temperature (22 °C ± 2 °C) for 1-, 3-, 7-, and 28-day cure periods.

2.3. Testing and Property Monitoring

Fresh and hardened CPB specimens prepared in the mixer were subjected to several consistency tests and to monitoring, mechanical, and microstructure tests. Figure 4 illustrates the flowchart detailing the laboratory test procedure for analyzing the prepared specimens.

2.3.1. Slump Test

The workability of the freshly mixed CPB was assessed through a slump test. This test measures the flowability or stiffness of CPB mixtures after they have been prepared, thus evaluating their transferability properties. The CPB mixes were prepared to a constant consistency of 20 cm, commonly used and targeted to ensure easy transfer from the surface to the subsurface. Initially, the binders (cement and slag), tailings (SBTs), and water were combined and completely mixed in a blender. Subsequently, the slump was measured, and if the target value of 20 cm was not attained, additional water was gradually added until the desired consistency was achieved. The blending times for all batches were mentioned in the previous section. The slump test was performed according to the procedure in ASTM C143 [43] standards. The entire test procedure was carried out 3 times to provide accuracy and consistency in the results. The mixing times for all components were detailed in the previous section.

2.3.2. Monitoring Tests

Tests for suction monitoring (SM), volumetric water content (VWC), and electrical conductivity (EC) were performed on CPB mixtures made with a constant w/b ratio of 7.25, including samples containing 50% slag and 0.4% AF, along with control specimens free from any admixtures. SM values were measured using a Teros 21, a soil water potential sensor, while EC and VWC were recorded using a Teros 12, a soil moisture sensor. The testing began by embedding the sensors in the center of cylindrical molds with dimensions of 10 cm in diameter and 20 cm in height filled with freshly mixed CPB mixtures. The sensors were linked to a Meter ZL6 data logger using cables for real-time data collection.

2.3.3. Unconfined Compressive Strength Tests

After all batches were prepared in the mixer and cast into molds, they underwent curing for 1, 3, 7, and 28 days. Samples of each mixture corresponding to the respective curing period were subjected to the UCS test following the curing periods. A uniaxial compression test was performed on the specimens using a UTEST 50 kN universal testing machine equipped with a load cell and data logger. The deformation ratio was set to 1 mm/min to ensure loading conditions. A minimum of three replicate tests were conducted for each batch and curing period to minimize experimental error and enhance the precision of the strength data. When calculating the average strength value of the specimens from the same batch and curing period, only the values within ±10% of the mean were considered and reported. To ensure comparability with industry standards, all UCS tests were performed following ASTM C39 [44].

2.3.4. Microstructural Analysis

The porosity properties and microstructures of the samples obtained by adding slag and AF admixtures at different ratios to the CPB mixtures were analyzed in detail. While 100% cement mortars were used as the control group, in the other specimens, 50% of the cement was substituted with slag, and 0.4% AF admixture was used. Pore size distribution and total pore volume were characterized using mercury intrusion porosimetry (MIP). Additionally, X-ray diffraction (XRD) and thermogravimetric analysis (TG/DTG) were employed to investigate the hydration products in cement pastes with a water-to-binder ratio (w/b) of 2. Following a 28-day standard curing period, the specimens were subjected to testing.

The slag and AF accelerator-reinforced (PCI50-SL50-0.4AF) sample, which showed the best results in the mechanical test, and control specimens were dried in a compressed air oven at 50 °C until a constant weight was obtained to ensure complete removal of moisture before MIP analysis. These samples were then subjected to MIP analysis by Quantachrome PM 60/13 instrument without any physical damage.

Similarly, cement pastes with a w/b ratio of 2 were designed to mimic PCI50-SL50-0.4AF and control samples for XRD and thermal investigations. The proportions in Table 4 were followed to prepare both cement pastes. The samples were blended in the mixer at the specified proportions and then poured into cylindrical molds and kept cured for 28 days under the appropriate laboratory conditions described in the previous section. All samples were dried in a compressed air oven at 50 °C until a constant weight was obtained to ensure complete removal of moisture. The samples were then ground following the test procedure and subjected to XRD analysis using a Panalytical Empyreyan device. XRD patterns were acquired using a step size of 0.0200° and a scan rate of 1.0000°/min over a 2-theta range of 5–60°.

Thermal analyses were carried out in two stages: thermogravimetric (TG) and differential thermogravimetric (DTG). Using a Linseis TGA PT 1000 analyzer, the temperature was enhanced at a ratio of 10 °C/min, reaching a max. of 1000 °C. Both mass loss and heat flow were monitored simultaneously during the process.

3. Results and Discussion

3.1. Influence of Slag Substitution and Addition of AF Accelerator on the Electrical Conductivity of CPB

One of the important parameters for monitoring the physical and changes that arise throughout the hydration process of cementitious materials is the electrical conductivity value. The cementitious materials’ properties, such as durability, porosity, and ion transport capacity, can be effectively characterized through electrical conductivity (EC) measurements, particularly during the early stages of hydration [45,46,47]. The conductivity values of the samples were reported to develop because of the quick release of Ca²⁺, Na⁺, K⁺, SO₄²⁻, and OH⁻ ions, following the dissolution of gypsum (CaSO₄·2H₂O), alite (C₃S), and soluble alkali sulfates [48,49,50]. These values, which are related to the movement of charge carrier ions during hydration, provide important clues about the durability and longevity of cement paste. High conductivity indicates that the material has high porosity and permeability and, therefore, may be less resistant to environmental influences [45,51].

The analysis of the electrical conductivity (EC) values provides valuable insights into the hydration process and the combined effects of slag substitution and AF accelerators on CPB mixtures. As seen in Figure 5, the control sample and the PC50-SL50-0.4AF mixture exhibit distinct behaviors in their EC profiles, reflecting the differing contributions of the slag and the AF admixture to the hydration dynamics. The control sample, which lacks slag and AF admixture, exhibits a gradual increase in EC, reaching its peak at approximately 498 min. This behavior is consistent with the steady release of ions during the hydration of Portland cement. In contrast, the PC50-SL50-0.4AF sample shows a significantly accelerated peak EC value, occurring at 360 min. This suggests that the presence of AF accelerators induced a rapid hydration process, likely due to the early formation of ettringite and the acceleration of alite hydration (C₃S).

Interestingly, the EC peak value for the PC50-SL50-0.4AF sample (3.30 mS/cm) is considerably lower than that of the control sample (5.45 mS/cm). This can be attributed to the partial replacement of cement with slag, which reduces the availability of charge carrier ions by slowing the dissolution of cementitious compounds. Moreover, the accelerated setting caused by the AF admixture forms a denser gel structure in the early stages, thereby reducing ion mobility. The subsequent decline in EC values for both samples over the 7-day period reflects the precipitation of hydrated phases and the reduced ion mobility due to pore structure densification. Notably, the PC50-SL50-0.4AF sample experiences a steeper decline, reaching a final EC value of 1.33 mS/cm compared to 4.09 mS/cm for the control sample. This indicates a more pronounced reduction in porosity and permeability, potentially enhancing the durability of the CPB mixture.

These findings highlight the dual and somewhat contrasting roles of slag and AF accelerators in influencing the hydration process. While slag tends to slow early hydration by moderating the alkaline environment, AF accelerators counteract this effect by promoting rapid initial setting and hydration. The net result, as observed in this study, is a synergistic effect that balances these opposing mechanisms, leading to early ion release followed by a rapid decline in ion mobility due to the formation of a dense microstructure. Other monitoring studies including suction and volumetric water content assessments corroborate these results.

3.2. Influence of Slag Substitution and Addition of AF Accelerator on the Self-Desiccation of CPB

Suction tests are employed to assess the absorption and transport of water within the internal pore structure of cement paste backfills, providing critical insights into the hydration process. These tests evaluate the penetration of water into the cement matrix during hydration, facilitating the composition of binder structures such as C–S–H and influencing the pore structure of the mixture. Prior studies have indicated that the crystalline structures and chemical compounds formed during hydration play a key role in identifying the mechanical strength and durability of the backfill material [29,52].

Figure 6 shows the 28-day self-desiccation development of the control sample and the sample coded PC50-SL50-0.4AF. The figure shows that for both samples, the negative pore water pressure (suction) grows negatively with time. While the sample coded PC50-SL50-0.4AF exhibited a sudden decrease in suction value in the negative direction after 6–7 h, the control sample showed a negative decrease after approximately 3 days. Following the 28 days of curing, the control sample showed a suction value of −117 kPa, whereas the sample named PC50-SL50-0.4AF demonstrated a suction value of −178 kPa.

The setting accelerator used in the PC50-SL50-0.4AF specimen caused an early hardening of the cement and consumed the water in the mix faster. This caused other binders, especially slag, to react faster, resulting in premature drying of the mix. Wan et al. [53] reached a similar finding and stated that setting accelerators promoted early cement hydration, which generated heat and caused water loss. In addition, Wan et al. [53] indicated in their study that mineral admixtures shorten the time of decrease in the relative humidity in the cement mortar. Additionally, the reaction of the set accelerator admixture with the mixing water present in the CPB mixture led to the formation of a gel-like structure, resulting in an increased presence of gel within the mixture and contributing to a more solid structure. Cavusoglu et al. [36] also reported this phenomenon. It is also clear that the products derived from the hydration process fill the pores in the mixture and form a more impermeable material.

In addition to the water consumption associated with cement hydration, further water loss occurs through the pore network, contributing to overall mass loss [53]. This phenomenon is substantiated by volumetric water content (VWC) measurements conducted on both samples. Figure 7 illustrates the VWC test results for the control and PC50-SL50-0.4AF samples. According to the VWC versus time data obtained from 5TE sensors, the control sample exhibited a higher water content compared to the samples containing slag and AF admixture. This disparity is attributed to the denser microstructure developed in the sample containing slag and AF accelerator during the setting and hardening process. Moreover, the VWC values align with the degree of saturation and porosity of the samples, as previously established in the literature [54]. The MIP tests’ findings on both samples further corroborate this relationship.

3.3. Influence of Slag Substitution and Addition of AF Accelerator on the Strength Development of CPB

Figure 8 illustrates the impact of slag substitution ratios on the compressive strength development of AF accelerator-reinforced CPB samples. The data indicate a consistent increase in strength for all samples with extended curing durations, attributable to the advancement of the hydration period. Furthermore, it is evident that replacing cement with slag at specific ratios (25%, 50%, and 75%) in combination with an AF admixture significantly improved the development of compressive strength of CPB as curing time advances.

AF accelerators are known to increase early hydration by growing the rate of CH formation and reducing the setting time, thus significantly accelerating early strength. This effect can be seen in the early-age strengths of PCI100-SL00-0.4AF, which performed better than the control sample without any admixture after 1 day of curing time. Following the 1-day curing period, the strength value of the control sample was measured as 189.42 kPa, while the strength value of the sample with 0.4% AF accelerator (PCI100-SL0-0.4AF) was measured as 214.46 kPa. However, the long-term accelerator effect decreases as hydration progresses naturally. After 28 days of curing time, the strength value of the sample containing AF accelerator was measured as 787.19 kPa, while the strength value of the control sample was measured as 827.37 kPa. Previous research indicates that incorporating AF accelerators with Portland cement promotes more hydration products during the initial stages of curing; however, their influence on hydration diminishes significantly beyond approximately 28 days [55,56,57,58]. This is the reason why these admixtures have a more important effect on the initial hardening of the material rather than on the strength achieved after a full 28-day curing period.

Slag typically causes a decrease in the early mechanical strength of cement pastes, attributed to its slower hydration kinetics relative to Portland cement. For example, the PCI75-SL25-0.4AF and PCI50-SL50-0.4AF samples showed delayed strength gain at 1 and 3 days compared to the 100% cement systems (control). This is due to the slag relying on the activation of calcium hydroxide (CH) derived from the cement’s hydration, which takes time to accumulate. Additionally, the low initial strengths can be attributed to slag’s secondary role in the hydration process. However, this situation was different in the PCI25-SL75-0.4AF specimen. The 1- and 3-day early-strength values of this sample were about 12.8% and 67% higher than those of the control sample, respectively. This suggests that the high slag content has an early positive effect on the chemical bonding mechanisms. In particular, the AF accelerator may have increased the strength of this specimen by enhancing the early reactions. In cases where slag is more prone to a chemical reaction, an early strength increase can be achieved if sufficient alkali activation is provided. This is particularly evident in the PCI25-SL75-0.4AF specimen.

Figure 8 clearly demonstrates that the substitution of cement with slag in CPB mixtures markedly enhances the long-term compressive strength in the presence of AF accelerator, irrespective of the slag replacement ratio. The PCI25-SL75-0.4AF, PCI50-SL50-0.4AF, and PCI75-SL25-0.4AF mixtures achieved 28-day strengths of 938.66 kPa, 1901.15 kPa, and 1227.98 kPa, respectively, compared to 827.37 kPa for the control sample. Additionally, the samples PCI50-SL50-0.4AF and PCI25-SL75-0.4AF exceeded the reference compressive strength threshold of 1000 kPa, as established in previous studies [30], demonstrating their suitability for applications requiring high mechanical stability. The incorporation of slag as a partial cement replacement substantially improves long-term compressive strength due to secondary reactions that densify the paste matrix. This aligns with previous studies highlighting slag’s ability to enhance matrix compactness and reduce porosity over extended curing periods [36]. The MIP test results further corroborate this observation.

The strength results of slag-based CPB mixes reveal that a 50% slag replacement ratio achieves the highest long-term strength (28 days). The performance of this sample can be attributed to two key factors. First, mixtures with lower slag content (e.g., 25%) exhibit insufficient reaction between slag particles and the portlandite (Ca(OH)₂) produced during cement hydration. This limits secondary hydration reactions, which are crucial for strength gain. Second, in mixtures with excessive slag content (e.g., 75%), the binder becomes inadequately activated due to insufficient portlandite availability for such a high slag fraction. These findings emphasize the need for careful optimization of slag replacement ratios in CPB mixtures. Several studies corroborate this conclusion, showing that a 50% substitution balances the pozzolanic reactivity of slag and its activation, delivering superior strength and durability compared to other ratios.

In summary, secondary materials such as slag require activation by CH produced during the hydration of cement, which is facilitated by the action of the setting accelerator at the early stages. However, the initial CH content is low, leading to a delay in strength development. Over time, the secondary hydration of slag, activated by CH, significantly contributes to strength gain, resulting in higher values at 28 days in slag-rich mixtures such as PCI50-SL50-0.4AF. Notably, the combined use of slag and setting accelerators more than doubles the CPB’s long-term compressive strength compared to traditional cementitious systems. This enhancement not only enables the use of CPB with reduced cement content, offering significant economic benefits to mining operations, but also promotes environmental sustainability by decreasing CO₂ emissions associated with cement production. By substitution part of the cement with slag, a byproduct material, the carbon footprint of cement manufacturing is minimized, aligning with global efforts to lower greenhouse gas emissions and promoting the broader adoption of eco-friendly mining practices.

3.4. Effect of Slag and AF Accelerator on the Microstructure of CPB

To assess the effects of slag substitution for cement and its combined use with AF accelerator on the microstructure and to enhance the performance of CPB samples, MIP testing was conducted on both the control sample and the PCI50-SL50-0.4AF CPB samples. Cement pastes with a water-to-binder ratio of 2, similar to those used in the CPB samples, were subjected to XRD and DTG/DTA analyses to elucidate the influence of slag substitution on the microstructure and thermal behavior.

3.4.1. Effect on Pore Structure

The PCI50-SL50-0.4AF and the control samples were subjected to MIP analysis following 28-day cure periods. The MIP data are depicted in Figure 7, presented in both incremental (a) and cumulative (b) formats, enabling a comparative evaluation of the pore size distribution and total porosity values of the CPB samples.

As depicted in Figure 9a, the PCI50-SL50-0.4AF sample exhibited a lower porosity relative to the control. One of the major contributions to the interpretation of the total porosity of this type of material is the critical pore size (d_cr). The pore size distribution curve reveals a dominant peak at the critical pore size, indicating the most probable pore diameter within the material. The data presented in Figure 9a reveals that the sample coded PCI50-SL50-0.4AF has a tighter pore structure than the control sample. This is supported by reducing the critical pore size from 1.743 μm to 1.206 μm.

The analysis of total porosity (n_t) values indicated that the control sample demonstrated a higher porosity of 29.15 cc/g, in contrast to 27.71 cc/g observed for the PCI50-SL50-0.4AF sample (Figure 9b). This difference indicates that the control sample possesses a more extensive and interconnected pore network, resulting in higher permeability. Conversely, the reduction in total porosity observed in the PCI50-SL50-0.4AF sample suggests that voids within the structure have been partially filled or the pores have become finer. This densification effect contributes to a more compact microstructure, which is directly connected with the enhanced strength properties of the material.

3.4.2. Effect on Hydration Products

In this research, XRD and thermal analyses were carried out on two separate cement paste samples, each formulated with a w/b ratio of 2. The control sample was formulated without any admixtures, whereas the PCI50-SL50-0.4AF sample comprised 50% Portland cement, 50% blast furnace slag, and 0.4% alkali-free admixture. The results of the XRD analysis, conducted following 28-day cure periods, are exhibited in Figure 8, whereas the corresponding differential thermal analysis (TG) and differential thermogravimetric analysis (DTG) results are depicted in Figure 10.

It can be seen from the graph that portlandite (Ca(OH)₂), which is known as a basic phase formed during cement hydration, produces peaks at 2–theta (θ) ≈ 18°, 34°, and 47° in the control sample. In addition, C–S–H gels are also observed to produce peaks in the same sample. In the sample with slag and the addition of alkali-free accelerator, the intensity of portlandite peaks decreased significantly compared to the control sample. This indicates that portlandite is consumed because of the pozzolan reaction of slag. The amorphous C–S–H gel causes an increase in the diffusion band. A wider diffusion band between 25° and 35° in the PCI50-SL50-0.4AF graph supports this. Due to the addition of AF admixture, the ettringite phase is observed around 2θ ≈ 9°–11°. The XRD results show that the PCI50-SL50-0.4AF sample has a more complex microstructure and enhanced hydration products, emphasizing the positive effect of slag and AF admixture in the mixtures.

Figure 11 shows the graphs of the TG/DTG of the control and PCI50-SL50-0.4AF coded samples. These graphs provide the possibility to compare the mass losses and decomposition rates of the samples depending on temperature. Figure 11 illustrates that the initial, secondary, and tertiary mass losses and corresponding endothermic peaks are observed within the temperature ranges of 50–200 °C, 400–500 °C, and 600–700 °C, respectively. The mass loss and peaks observed between 50 and 200 °C are ascribed to the dehydration of cement’s hydration products, including C–S–H, carboaluminates, ettringite, and gypsum. In this temperature range especially, the decomposition of physically bound water and some early hydration products is observed. A significant mass loss within this temperature range indicates a high concentration of hydration products or reflects the material’s substantial suction capacity, which may result from hydration mechanisms. At 400–500 °C, CH begins to decompose. During the decomposition of CH, free water is lost, and this process causes both weight loss and a pronounced endothermic peak. In the 600–700 °C range, calcite (CaCO₃) decomposition occurs. In this temperature range, carbon dioxide (CO₂) is released by the decomposition of carbonate phases, resulting in weight loss and the formation of endothermic peaks.

The mass loss observed in the PCI50-SL50-0.4AF sample within the 50–200 °C range exceeds that of the control sample, indicating that the setting accelerator enhances hydration kinetics during the early stages of cement hydration. At 400–500 °C, the control sample exhibits a more pronounced CH peak, signifying a higher concentration of free calcium hydroxide. In contrast, the PCI50-SL50-0.4AF sample shows a reduced peak in this range, which can be attributed to the pozzolanic reactions consuming portlandite and promoting the generation of secondary products of hydration, such as C–S–H gel. Within the 600–700 °C range, the carbonate (CaCO₃) peak is less prominent in the control sample, suggesting limited carbonation. Conversely, the higher carbonate content observed in the PCI50-SL50-0.4AF sample implies that the setting accelerator not only influences hydration but also enhances carbonation processes. These findings underline the dual role of the setting accelerator in accelerating hydration and promoting the formation of carbonation products, which may contribute to improved long-term performance.

The TG graph shows that there is more mass loss in the control sample, especially after 450 °C. This indicates that the portlandite formed during hydration is denser in the sample. In the sample named PCI50-SL50-0.4AF, the mass loss after 450 °C is less than the control sample. This indicates that with 50% slag and 0.4% alkali-free admixture used, portlandite is consumed by pozzolanic reactions, and less free CH content is formed.

In conclusion, thermal analyses indicate that the PCI50-SL50-0.4AF sample exhibits lower portlandite content, higher C–S–H gel formation, and increased carbonate content due to the influence of pozzolanic reactions and the AF admixture. These findings suggest that the sample has the potential to enhance long-term durability performance, highlighting the synergistic effects of slag substitution and the setting accelerator.

4. Conclusions

This study evaluated the effects of slag substitution (25%, 50%, and 75%) and a fixed 0.4% AF dosage on CPB performance. Tests included unconfined compressive strength (UCS), microstructural analysis (MIP), and monitoring (EC, suction, and VWC), along with XRD and TG/DTG analyses of cement pastes. Key findings are summarized below:

• Replacing 50% of cement with slag and adding 0.4% AF accelerator shifted the EC peak 2 h earlier than the control sample and reduced the peak EC value from 5.45 mS/cm to 3.30 mS/cm. This indicates lower porosity and reduced ion mobility in the slag-reinforced mixture, which contributes to enhanced durability.
• The AF accelerator significantly accelerated hydration, as evidenced by a sharp decrease in suction values in slag-containing mixtures within 6–7 h compared to 3 days for the control. Lower volumetric water content (VWC) measurements in slag-containing mixtures further confirmed faster water consumption and hydration.
• For slag-free CPB, the AF accelerator improved early-age strength at 1 and 3 days but had minimal influence on the 28-day strength (827.37 kPa for control vs. 787.19 kPa for AF-containing mixtures). However, in slag-containing CPB, the combination of 50% slag and 0.4% AF accelerator dramatically enhanced 28-day strength, achieving 1901.15 kPa, which is 2.3 times higher than the control. This highlights the positive synergy between slag’s pozzolanic activity and AF accelerator in promoting hydration.
• MIP analysis revealed that the PCI50-SL50-0.4AF sample exhibited lower porosity and a finer pore structure compared to the control, contributing to a denser matrix and improved mechanical properties.
• XRD and TG/DTG analyses showed that slag and AF accelerator reduced portlandite content by promoting C-S-H gel formation and increased secondary hydration products such as carbonation compounds. These changes contributed to the long-term durability and stability of the CPB matrix.

These findings highlight the significant contribution of slag substitution and AF accelerators to enhancing the strength and durability of CPB. The integration of byproduct materials like slag, commonly generated by various industries, not only enhances the mechanical properties of backfill but also supports the adoption of sustainable and cost-effective mining practices. By reducing reliance on Portland cement, this approach aligns with global efforts to lower CO₂ emissions, promoting environmentally friendly and economically viable solutions for the mining sector.