Effect of Mixing Water Temperature on the Thermal and Microstructural Evolution of Cemented Paste Backfill in Underground Mining

Abstract

Cemented paste backfill (CPB) gains strength through the hydration of the binder constituent of the CPB, where mix temperature is a key influencing factor. Both rate of strength development and ultimate strength are influenced by the overarching temperature conditions in which the binder hydration occurs. This study investigates the influence of mixing water temperature on the thermal behaviour, hydration kinetics, and microstructural development of CPB using a combination of thermal finite element modelling, thermogravimetric analysis (TGA), scanning electron microscopy (SEM), and energy-dispersive X-ray spectroscopy (EDS). Five CPB mixtures were prepared, with water temperatures ranging from 5 °C to 50 °C, and tested under controlled conditions to isolate the effects of the initial thermal input. Results show that moderate mixing water temperatures (20–35 °C) optimize hydration and mechanical strength, while excessive temperatures (≥50 °C) increase the risk of thermal cracking due to generation of excessive heat. The thermal modelling results demonstrated that the highest temperatures were observed in the bottom section of the fill mass, in contact with the surrounding rock, where the combined effects of mix-generated heat and rock conduction were most pronounced. The 50 °C mix reached a peak internal temperature of 85.6 °C with a thermal gradient of 40.5 °C, while the 5 °C mix recorded a much lower peak of 55.7 °C and a gradient of 16.8 °C. These results highlight that higher mixing water temperatures accelerate early hydration reactions and significantly influence the internal thermal profile during the first 21 days of curing. Based on these findings, the design of paste plants can be improved by incorporating a heating/cooling system for the mixing water tank—firstly, to ensure the water temperature does not exceed 50 °C and secondly, to maintain water within an optimal temperature range, potentially reducing binder consumption.

1. Introduction

Cemented paste backfill (CPB) is widely used in underground mining operations to provide structural support in mined-out stopes, while enabling the environmentally responsible disposal of tailings. The performance of CPB is primarily evaluated by its mechanical strength and long-term chemical stability, which are influenced by several factors, including binder content, tailings’ characteristics, solids content, dynamic loading, confinement, and curing conditions [1,2,3,4]. Among these, temperature plays a pivotal role in controlling the rate and extent of binder hydration, which is the primary mechanism behind the development of strength in CPB. The hydration of binder is a highly exothermic process, and the thermal environment significantly influences its kinetics. As mining operations delve into increasingly greater depths, the ambient rock mass temperature rises significantly due to inherent geothermal gradients. The exothermic hydration reactions of the binding agents within the CPB itself generate substantial internal heat, leading to a self-heating phenomenon and elevated temperatures within the CPB fill mass. These thermal conditions, both external (geothermal) and internal (hydration heat), can profoundly influence the early-age curing process, the rate at which strength develops, and, critically, the long-term mechanical performance and durability of CPB [5,6,7,8,9]. Elevated temperatures can accelerate the dissolution of clinker phases and the precipitation of hydration products such as calcium silicate hydrate (C-S-H) and ettringite, resulting in a faster strength gain at early stages [10]. Zhang and Li [11] showed that increasing the curing temperature significantly accelerates early-age development of strength by enhancing the rate of hydration reactions.

However, there is a critical threshold beyond which these benefits diminish or even reverse. Excessively high temperatures, particularly those sustained above 40 °C, can lead to a reduction in long-term strength [9]. For instance, specific studies have observed that at temperatures around 65 °C, CPB can become notably more brittle, a phenomenon attributed to the formation of a coarser and less dense pore structure within the hydrated matrix [9]. This coarsening of pores can compromise the material’s load-bearing capacity and increase its susceptibility to failure. Conversely, at the other end of the spectrum, low temperatures significantly slow down the hydration process, thereby impeding early strength gains [12].

Rapid hydration can lead to an inhomogeneous microstructure with increased porosity and potential for thermal cracking, which compromises the integrity of the backfill [8,9,10,11,12,13,14]. In large-volume CPB applications, internal temperatures can rise well above ambient levels due to the heat generated during hydration. Due to the small surface area-to-volume ratio, mass CPB has a high potential for thermal cracking, caused by the heat generation from binder’s hydration. This fundamental is not fully investigated for CPB; however, for example, in concrete, the temperature is limited to 70 °C (158 °F) during hydration [15]. If the temperature of the concrete during hydration is too high, it will cause the concrete to exhibit high early-strength development but subsequently gain less strength in the later stages, resulting in lower overall durability of the concrete [15]. Additionally, high temperature issues are of concern, especially in mass concrete pours, where the core temperature can be very high due to the mass effect, while the surface temperature is lower. This causes a temperature gradient between the surface and the core; if this differential in temperature is too large, it will cause thermal cracking. To mitigate such risks, construction guidelines, such as ACI 301, recommend limiting the internal temperature rise and maintaining a maximum differential of 19 °C between the core and surface of the mass [15].

Beyond uniaxial compressive strength, temperature influences other key mechanical characteristics that govern the mechanical behaviour of CPB. Studies show that higher temperatures can increase both the cohesion and internal friction angle of CPB, which are critical parameters in Mohr–Coulomb failure criteria, suggesting an initial strengthening effect under confined conditions [16]. However, these studies also reveal that elevated temperatures can alter the failure modes of CPB under triaxial compression, indicating a change in the material’s fundamental deformation and failure mechanisms [16]. Furthermore, the effect of curing temperature is captured using ultrasonic pulse velocity (UPV), a non-destructive method to monitor and understand the material’s evolving strength properties [17]. UPV can be directly correlated with materials’ density and stiffness, both of which are influenced by hydration kinetics and microstructure development under varying thermal conditions [18].

The complexity of CPB behaviour under thermal loading is further compounded by the interaction of temperature with other material constituents and environmental factors. For instance, the combined effects of curing temperature with various mineral admixtures, such as phosphorus slag, have been investigated, revealing how these additives can modify hydration kinetics and the resulting mechanical properties under different thermal regimes [19,20,21,22,23]. Similarly, the presence of sulphate concentrations alongside varying curing temperatures is known to collectively influence development of strength and the formation of specific mineral phases in slag–CPB systems, which can impact long-term durability [24].

Temperature also significantly affects the time-dependent rheological behaviour of fresh CPB [22,25]. Recent research indicates that both yield stress and the viscosity of the paste increase steadily as the curing time elapses, with this increase further accelerated by higher curing temperatures [22]. This rheological response is crucial for pipeline transport and placement efficiency in mining operations. Similarly, Zhao et al. [26] found that temperature changes significantly influence the rheological and microstructural properties of CPB, especially when combined with extreme pH levels [26].

At a microstructural level, advanced techniques like ¹H Nuclear Magnetic Resonance (NMR) relaxometry have revealed that increasing curing temperature from 10 °C to 60 °C systematically alters the content of various hydration products, such as calcium silicate hydrate (C-S-H), and the distribution of pore water within the cemented paste, directly influencing its macroscopic properties [27]. The kinetics of hydration, specifically the intensity and timing of heat peaks, are dramatically affected by temperature, with higher temperatures leading to a significantly more intense and earlier heat release, indicating faster reaction rates and denser microstructures [4,7].

As discussed so far, a substantial body of research has examined the effect of curing temperature on hydration kinetics, strength’s development, and microstructural evolution of CPB. Recent studies have also underscored the importance of advanced modelling techniques such as thermal and microstructural simulations in predicting CPB behaviour under complex conditions. In particular, research on CPB made from high-sulphide tailings and used in deep stopes has highlighted the interplay between binder chemistry, temperature management, and hydration kinetics [25,28].

However, the influence of mixing water temperature on CPB has been less extensively studied. Wu et al. [29] examined the combined effects of water temperature and pH on CPB’s performance, noting that higher water temperatures accelerate hydration but must be balanced against risks of thermal cracking. The initial backfill temperature is also noted to affect stiffness significantly. Higher initial temperatures tend to increase early stiffness and steepen the stress–strain curve, which can be beneficial for early support. However, sustained temperatures above 35 °C may lead to a reduction in long-term elasticity, suggesting that while early stiffness is improved, the material’s ability to deform elastically under load might be compromised over time, highlighting an optimal temperature range for overall mechanical performance [30].

Mixing water temperature is a controllable parameter at the paste plant, offering a practical means of influencing the initial thermal energy of the backfill slurry. This control becomes particularly important in deep underground mining scenarios, where CPB is often poured into large stopes with low surface-to-volume ratios. In such geometries, the dissipation of heat generated during hydration is limited, which exacerbates the risk of thermal accumulation and cracking. The ability to modulate the initial temperature of the slurry through adjusting the mixing water temperature provides a potentially effective method to manage internal heat’s development, enhance uniformity in hydration, and reduce the risks associated with thermal stresses. It also offers the potential to influence the hydration of the paste fill binder to meet strength requirements specific to the mining cycle of any given operation. For example, where early-age strength development is critical to a mining cycle, acceleration of the CPB’s hydration by increase mix water temperature may be attractive to the optimization of the mining sequence, even with a reduction in peak CPB strength.

This study aims to develop a methodology and experimental framework to investigate the isolated effects of mixing water temperature on the thermal and mechanical behavior of CPB mixtures, as well as to evaluate the resulting thermal gradients within the stope environment. A combination of laboratory testing and thermal finite element analysis (FEA) is used to evaluate the temperature distribution and thermal gradients during hydration. This study also evaluates how thermal variations influence the hydration process and the resulting material properties. Scanning electron microscopy (SEM) is used to characterize microstructural features, while thermogravimetric analysis (TGA) quantifies hydration products and evaluates thermal stability.

2. Laboratory Test Work

2.1. Raw Materials

All materials used in this study were sourced from an operating mine in Western Australia. The CPB mixtures were prepared using tailings, Low Heat Portland cement (Type LH), and process water from the site. Tailings were collected in a dewatered form and stored in sealed containers prior to mixing. The particle size distribution (PSD) of tailings is shown in Figure 1.

The LH cement was selected as the current site binder. The chemical composition of the cement is provided in

Table 1. Chemical composition of the low heat cement (wt%), as provided by the supplier.

SO3	SiO2	Al2O3	Fe2O3	MgO	CaO	Na2O	LOI
2.15	28.75	11.00	1.85	4.55	50.05	0.40	1.25

.

2.2. Mix Proportions and Preparation

All five (5) CPB mixtures were prepared with a solids content of 75.5 wt%, with 5 wt% LH binder as a percentage of total dry mass. The remaining 24.5 wt% was process water. Five mix designs were prepared using mixing water at controlled temperatures of 5 °C, 20 °C, 30 °C, 35 °C, and 50 °C, respectively. The mine process water was preheated or cooled to the target temperature before being incorporated into the mix.

Mixing was conducted in a laboratory mixer with controlled rotational speed to ensure homogeneity. Each batch was mixed for 5 min, consisting of 2 min of dry mixing followed by 3 min after the addition of water. The fresh CPB slurry was cast into cylindrical moulds (50 mm diameter × 100 mm height) and sealed with plastic film to prevent moisture loss. Compression tests were conducted on three identical samples for each mix. Other characterizations, including TGA, SEM, and isothermal analysis, were performed on a single representative sample.

2.3. Curing Conditions

All specimens were cured in a controlled environmental chamber maintained at a constant temperature of 37 °C and relative humidity of 90% to simulate the deep underground mine conditions. Curing was undertaken for 14 days for all test samples prior to test work being undertaken. The selected curing temperature and humidity levels are representative of the elevated thermal and moisture environments commonly encountered in backfilled stopes in deep mines.

2.4. Experimental Procedures

The following methods were employed to evaluate the thermal and mechanical behaviour of the CPB mixtures:

• Thermal Finite Element Modelling: Temperature profiles and heat generation within a stope model were simulated based on material properties, initial temperatures, and boundary conditions relevant to mine backfilling.
• Thermogravimetric Analysis (TGA): The degree of hydration was determined by analysing weight loss profiles up to 1000 °C at a heating rate of 10 °C/min under a nitrogen atmosphere.
• Scanning Electron Microscopy (SEM) and Energy-Dispersive X-Ray Spectroscopy (EDS): Microstructural and elemental analysis was conducted on gold-coated specimens. SEM provided insights into pore structure and hydration morphology, while EDS identified elemental composition.
• Uniaxial Compressive Strength (UCS): Strength testing was performed using a servo-controlled loading frame at a displacement rate of 1 mm/min, with results averaged from three replicate specimens per mix.
• Isothermal Calorimetry: Heat evolution profiles were recorded for the first 72 h after mixing to evaluate early-age hydration kinetics.

2.5. Heat Flow Measurement

To obtain accurate heat generation profiles for use in the thermal FEA of CPB, isothermal calorimetry testing was conducted in accordance with the guidelines of ASTM C1702 [31]. A TAM AIR isothermal calorimeter was employed due to its high sensitivity and stable thermal environment, enabling precise measurement of the exothermic heat released during the binder’s hydration.

Five distinct CPB mixtures were prepared, using mixing water at controlled temperatures of 5 °C, 20 °C, 30 °C, 35 °C, and 50 °C. Immediately after mixing, approximately 20 g of fresh CPB slurry from each batch was transferred into airtight glass ampoules to prevent moisture loss and minimize environmental interference. The calorimeter chamber was pre-calibrated and maintained at a constant ambient temperature of 25 °C throughout the test’s duration.

Each sample was tested in triplicate to ensure the reliability and reproducibility of the data. The heat flow, representing the rate of hydration-related heat generation, was recorded continuously over a period of 144 h (6 days). This duration was chosen to capture both the initial and secondary stages of binder hydration, including the dormant and acceleration phases.

The resulting time-dependent heat flow curves provided key quantitative input for the FEA simulation of CPB thermal behaviour. By capturing the influence of different initial water temperatures on heat evolution, the calorimetry data enabled a more realistic prediction of in-stope temperature gradients, maximum core temperatures, and the potential for thermal-induced damage in deep backfill applications.

3. FEA Modelling

3.1. Geometry

The stope was modelled based on the benchmark model of the mine site from which tailings, binder and water samples were obtained, with a nominal geometry of 25 m width, 30 m length (along strike), and 75 m height. The host rock mass surrounding the stope was modelled as an environment with almost five times the stope’s width. Particular attention was paid to ensure the mating of adjacent surfaces, which were treated as intersecting surfaces, to match the exact CPB fill mass volume poured into the stope per day. Small details (relatively) on the stope’s surface in contact with the rock mass were ignored. The FEA model is shown in Figure 2, with the modelling completed in SolidWorks 2024 and imported into ANSYS 2024 R2 for the FEA.

Filling of the CPB void was simulated using a typical “plug and body” filling strategy. The initial plug pour was completed to an initial height of 8.0 m prior to a 24 h rest period. The remainder of the fill mass was modelled based on 24 h production intervals, equal to a 15 m elevation per pour, until completion of the stope after 7 days of production (a total of 56,375 m³). The thermal simulation setup was designed to continue analysis up to day 21 for the curing cycle and heat generation to be completed and converge for a total mix volume of 56,375 m³.

3.2. Effect of Paste Reticulation

The impact of the paste delivery reticulation (retic) system was analyzed independently due to its considerable length as shown in Figure 3. Instead of modelling the entire 1.5 km pipeline, a 1 m segment under underground conditions was simulated using FEA software. This segment was assumed to represent a straight steel pipeline extending 1.5 km in total, specified as DN150 Sch.80 Carbon Steel. Due to the considerable length of paste reticulation in inter-level boreholes, the full pipeline was modelled as being in contact with the surrounding host rock mass. The simulation results from the 1 m model were used to estimate the total heat loss or temperature change along the full pipeline, with a travel time through the pipeline assumed to be 7 min.

3.3. Materials Data

The stope was filled using a cemented paste backfill mix with the following measured thermal and physical properties:

• Density: 2100 kg/m³;
• Isotropic thermal conductivity: 1.5 W/m·°C;
• Specific heat capacity: 1000 J/kg·°C.

The surrounding stope rock, identified as Coora Dolerite based on geological data [32], had the following properties:

• Density: 3500 kg/m³;
• Isotropic thermal conductivity: 2.5 W/m·°C;
• Specific heat capacity: 272 J/kg·°C.

3.4. Boundary Conditions

The stope was thermally and mechanically bonded to the surrounding rock using a bond contact condition. Each lift layer was activated sequentially according to a daily CPB fill schedule, with contact conditions established between the newly placed layer, the underlying layer, and the rock mass. A transient thermal simulation was performed, with the initial temperature of the paste set uniquely for each case. The model was run for six different initial paste temperatures: 5, 20, 30, 35, and 50 °C. The surrounding rock temperature was uniformly set to 37 °C, in accordance with site data.

3.5. Heat Loads

The transient thermal analysis incorporated the following loading and boundary conditions:

• Internal heat generation was applied to each lift increment using tabulated data derived from calorimetry tests.
• A fixed temperature of 37 °C was assigned to the external surface of the surrounding rock mass, acting as an infinite heat reservoir, as illustrated in Figure 4.
• As per site specifications, no air exposure was considered following the placement of each lift. Therefore, air thermal effects were excluded from the model.
• It was assumed that the thermal contact between the backfill and the surrounding rock was 100% efficient, representing an ideal heat transfer scenario.
• Heat generated by friction during CPB transport through the reticulation system was considered negligible
• The temperature of the soil and rock surrounding the pipeline was also maintained at a constant 37 °C, consistent with the site’s thermal conditions.

3.6. Analysis

The FEA model of the ground and the stope was meshed using Hex-element. The mesh was to be smooth and not distorted, and a smooth grading was applied at points of interest. Transient thermal analysis determined temperatures and other thermal quantities that varied over time. The transient thermal analysis followed procedures similar to those of a steady-state thermal analysis. The main difference was that the applied loads in the transient analysis were functions of time. To specify time-dependent loads, the heat generation load versus time curve was divided into the simulation load steps. For each load step, it was necessary to specify both heat generation values and time values, along with other load step options, such as stepped or ramped loads, and automatic time stepping. Then, each load step was written to a file, and all load steps were solved together.

The heat transfer from the ground to the mix throughout the pipe was also analyzed. After one hour of 5 °C mix flowing inside the pipe, by a transient thermal analysis, the rock temperature around the pipe reached almost 5 °C, and the heat transfer reached close to zero (or almost negligible). The total heat energy that was transferred to 1 m of mix length in the pipe for 1 h was multiplied by 1500 (1.5 km of pipe length), to obtain the total heat that was transferred from the ground to the mix. After one hour of mixing flowing inside the pipe, the temperature gradient between the ground and the mix reached close to zero, which means that the heat transfer could be negligible.

Table 2. Thermal and physical parameters used in the FEA simulation of CPB heat transfer.

Heat Transferred to 1 m of Mix Tube Volume, at 5 Degree Mix, in First Hour Joule: 3,607,900	1 m Length of Pipe DN150 Mix Volume (m3): 0.016741547	Density kg/m3: 2100
LIFT 1 INITIAL TEMP ADD THROUGH PIPING for LIFT 1:	q = mcΔT	q = 5.41 × 109 m = 12,862,500
	ΔT =	0.4

present the thermal and physical parameters used in the simulation.

The total calculated heat is assumed to be only applied to lift stage 1 (L1) and its relevant volume, since for this stage the mix flow starts, and the ground temperature was 37 °C, and after day one, there was a 24 h delay before pouring level 2. For day three, the same total heat was considered for the total volume of lift 2 (L2), as the initial ground temperature again started at 37 °C. For L3 to L7, the earth temperature around the pipe was considered to be 5 °C (due to nonstop mix flow), and the mix temperature rise was almost zero inside the pipeline.

As per the above description, the temperature gradient between the earth and the mix for 20, 30, 35, and 50 °C models was lower than 32 °C (37 − 5= 32). As a result, the temperature increase/decrease through the reticulation system was negligible throughout the study.

4. SEM and TGA Tests

For SEM and TGA tests, CPB cylinders were prepared for the uniaxial compressive strength (UCS) test and then crushed at 14 days of curing. The 14-day curing period was selected as it reflects a critical point in early strength development for CPB in mining operations, where backfill must achieve sufficient strength to support ongoing excavation. The crushed samples were then used for the SEM and TGA tests. The SEM investigation was conducted at the Flinders Microscopy and Microanalysis lab using an FEI Inspect F50 SEM, equipped with an EDS detector, for the microstructural evaluation of the five mixes. The PerkinElmer TGA 8000 analyzer was employed to conduct TGA on the five mixes. The hydration was stopped prior to the TGA test using a two-step method involving deionized water extraction followed by alcohol–acetone treatment. During the TGA test, 25 to 35 mg of ground samples was slowly heated up from 30 °C to 1000 °C at a heating rate of 10 °C/min, using nitrogen as the carrier gas, while its weight was constantly monitored.

Equation (1) was used to calculate the degree of hydration values based on the TGA curves [7].

$$\alpha \mathrm{T}\mathrm{G}\mathrm{A}(\mathrm{t})(\mathrm{\%})={\displaystyle \frac{{\mathrm{W}}_{\mathrm{n}}\left(\mathrm{t}\right)}{{\mathrm{M}}_{\mathrm{c}}\times {\mathrm{W}}_{\mathrm{n}}\left(\mathrm{\infty }\right)}}\times 100$$

(1)

where W_n(∞) is the non-evaporable water mass per gram of binder at the point when time approaches infinity (t→∞), which marks complete hydration of 1 g of binder. At this stage, W_n(∞) was estimated to be 0.2293 g based on the binder’s mineralogical composition. W_n(t) and M_c are the non-evaporable water mass at a specific time (t) and the initial mass of the anhydrous sample in grams, respectively. The non-evaporable water content was calculated by subtracting the weight loss between 600 °C and 800 °C, resulting from CO₂ release during calcite decomposition, from the weight loss between 145 °C and 1000 °C [7].

5. Results

5.1. Heat Flow

The heat flow measurements for each mix, reflecting internal heat generation during hydration, are presented in Figure 5.

As can be seen in Figure 5, the peak internal heat generation by the binder hydration for different mixes is as follows:

• 5 °C water: 407 W/m³;
• 20 °C water: 538 W/m³;
• 30 °C water: 660 W/m³;
• 35 °C water: 725 W/m³;
• 50 °C water: 884 W/m³.

Figure 5 also shows that mixes with warmer water reached their heat generation peak more rapidly. This is attributed to higher molecular kinetic energy at elevated temperatures, which accelerates binder hydration [33]. Enhanced dissolution of binder grains and increased reactivity of key clinker phases (especially C₃S) contribute to greater early heat output. Additionally, elevated water temperatures promote the dissolution of binder particles, which is crucial for the formation of hydration products, such as calcium silicate hydrate (C-S-H) and ettringite [34].

5.2. FEA Modelling Results

The results of the modelling are presented in this section as a temperature distribution at the final time step.

5.2.1. Temperature Distribution Across the Mixes and the Rock

Figure 6, Figure 7, Figure 8, Figure 9 and Figure 10 illustrate the maximum temperature within the stope. The peak was observed near the interface between the CPB and the stope wall. Additional modelling (Figure 11) depicts heat transfer from both the delivery pipe and surrounding rock into the backfill, highlighting the impact of external boundary conditions on early-stage thermal behaviour.

5.2.2. Peak Temperature on Day 21

The temperature variations for the different mixes are presented in Figure 12, Figure 13, Figure 14, Figure 15 and Figure 16. The coloured lines represent distinct thermal metrics: the green line indicates the maximum temperature recorded within the stope, the red line shows the minimum temperature, and the blue line represents the average temperature across the stope volume. The maximum temperature (green) is primarily driven by exothermic hydration reactions, which are more intense at higher initial water temperatures. The minimum temperature (red) reflects the coolest zones, typically near the stope boundaries or areas with lower hydration activity. The average temperature (blue) captures the overall thermal behaviour and is influenced by both internal heat generation and heat dissipation to the surrounding rock. Differences in the shape and magnitude of these curves across figures are due to the varying initial mixing water temperatures, which affect hydration kinetics and thermal gradients. As shown, the highest temperatures within the stope—resulting from internal heat generation, the initial rock temperature, and subsequent heat transfer—occurred near the base of the stope lifts, where the mix was in direct contact with the surrounding rock. These results correspond to the simulation period from Day 1 to Day 21. The simulation was conducted individually for each mix, with input data from the respective test conditions applied separately. A summary of the stope temperatures on Day 21 for all mixes is provided in

Table 3. Summary of thermal simulation results for CPB mixes with varying initial water temperatures after 21 days of curing.

Mix ID	Initial Temperature (°C)	Max (°C)	Min (°C)	Delta (°C)
5 °C	5	55.7	38.9	16.8
20 °C	20	72.3	42.2	30.1
30 °C	30	76.8	43.2	33.6
35 °C	35	74.6	43.1	31.5
50 °C	50	85.6	45.1	40.5

.

Based on the thermal modelling outcomes, the maximum temperature within the stope increased proportionally with the initial mixing water temperature. Elevated initial water temperatures accelerated binder hydration reactions, resulting in a greater exothermic heat release. This effect was most pronounced in the 50 °C mix, which reached a peak temperature of 85.6 °C.

Similarly, the minimum recorded temperatures within the stope also increased with the rising initial mixing water temperature. Due to the limited dissipation of heat within the backfilled stope, a portion of the generated heat was retained, thereby elevating the baseline temperature across all sections.

The temperature gradient—defined as the difference between the maximum and minimum temperatures—also increased with the initial water temperature. For instance, the 5 °C mix exhibited a gradient of 16.8 °C, while the 50 °C mix reached a gradient of 40.5 °C. This trend confirms that higher initial water temperatures lead to a more intense and spatially variable heat distribution across the stope. In contrast, lower water temperatures (e.g., the 5 °C mix) resulted in reduced heat generation and more uniform thermal profiles.

5.3. UCS Test Results

Figure 17 displays the cylindrical specimens used for UCS testing, and the results are summarized in

Table 4. Average UCS values for CPB samples cured for 14 days, prepared with different mixing water temperatures. Values in parentheses represent mean error from three replicate tests.

Mix ID	Compressive Strength (MPa)
5 °C	1.47 (±0.026)
20 °C	1.55 (±0.032)
30 °C	1.64 (±0.036)
35 °C	1.58 (±0.029)
50 °C	1.22 (±0.022)

. The strength values are average values between three identical samples. As shown, the compressive strength of the CPB samples increased with water temperatures up to 30 °C, reaching a peak strength of 1.64 MPa. However, a further increase to 50 °C led to a decline in strength, with the 50 °C mix yielding only 1.22 MPa.

These findings are consistent with the TGA results suggesting that excessively high mixing temperatures may impair complete hydration, potentially due to premature pore closure or microcracking, ultimately reducing mechanical strength.

5.4. SEM/EDS and TGA Results

Figure 18, Figure 19, Figure 20, Figure 21 and Figure 22 present the SEM micrographs at different magnifications of CPB samples prepared using mixing water at 5, 20, 30, 35, and 50 °C after compression tests. All samples exhibited relatively porous microstructures due to the low binder content and high water-to-binder ratio typical of CPB. However, visual analysis indicated a denser microstructure from the 5 °C to the 30 °C mixes, probably due to improved hydration and a denser matrix formation at moderate temperatures. Beyond this point, at 35 °C and especially at 50 °C, an increase in visible porosity was observed, likely due to incomplete hydration and thermal microcracking.

Figure 23, Figure 24, Figure 25, Figure 26 and Figure 27 present the EDS results for CPB samples prepared with 5, 20, 30, 35, and 50 °C mixing water, analyzed at selected locations. The results are qualitative and used for comparative purposes only. Across all mixes, the dominant elements identified were calcium (Ca), silicon (Si), chlorine (Cl), sodium (Na), and gold (Au), along with trace amounts of iron (Fe), magnesium (Mg), aluminium (Al), and sulphur (S).

A high calcium content is expected, as Ca is a principal constituent of binder, particularly in hydration products such as calcium silicate hydrate (C-S-H), portlandite (Ca(OH)₂), and ettringite. The presence of substantial Ca levels reflects active binder hydration and the formation of these compounds.

Silicon, which combines with Ca to form C-S-H—the primary strength-contributing phase in hydrated binder—was also detected in significant quantities. The abundance of Si in the EDS spectra suggests effective hydration, particularly in mixes where temperature conditions were favourable to C-S-H’s formation. A higher Si content typically correlates with the increased mechanical strength and structural integrity of the CPB.

Chlorine’s presence may result from impurities or chemical constituents present in the mixing water or tailings. Sodium, detected in all samples, likely originates from sodium-bearing salts or additives present in the binder or tailings. Na can contribute to the formation of sodium silicates, influencing both the hydration process and the evolution of pore structure.

Gold (Au) was consistently detected, attributable to the use of gold coatings for the sample preparation of SEM/EDS analyses. It should be noted that the variation in Au content across the EDS spectra could be influenced by both orientation effects and inconsistencies in the gold coating applied during sample preparation. Trace elements such as Fe, Mg, Al, and S are common in the raw materials used for binder and tailings. These elements can influence the formation of secondary phases, such as calcium aluminates, ettringite, and sulphate-bearing compounds, which may affect the setting behaviour and long-term durability of the CPB.

The sulphur content, in particular, is associated with the formation of sulphate-bearing phases, which can impact the dimensional stability and strength development of the mix [35]. Unlike microprobe analysis, the CPB samples were not polished or embedded in resin, which may have introduced variability due to surface roughness, particle orientation, and detector shielding effects. Consequently, the EDS data should be interpreted as indicative of elements’ presence and distribution rather than absolute concentrations. These constraints limit quantitative comparisons but remain useful for identifying trends in hydration-related phase development.

Figure 28 illustrates the TGA results, showing weight loss (%) versus temperature (°C) for the CPB mixes.

Table 5. Degree of hydration calculated from TGA data for CPB samples cured for 14 days.

Mix ID	Degree of Hydration (%)
5 °C	31
20 °C	32
30 °C	44
35 °C	36
50 °C	25

presents the corresponding degree of hydration values at 14 days of curing, derived from the TGA curves. All mixes exhibited progressive weight loss with increasing temperatures, a typical characteristic of cementitious systems. This weight loss is attributed to the decomposition of hydration products, such as free water, bound water, portlandite, and carbonation compounds. The calculated degree of hydration represents the proportion of binder that has undergone hydration to form products, such as calcium silicate hydrate (C-S-H).

The 5 °C mix exhibited moderate weight loss, particularly in the 0–200 °C range, reflecting the evaporation of free water. Limited weight loss in the 200–600 °C range indicates a reduced amount of bound water and portlandite decomposition, consistent with a lower degree of hydration due to slower reaction kinetics at lower temperatures.

The 20 °C mix showed a slightly higher weight loss, particularly in the 200–600 °C range, indicating greater hydration than the 5 °C mix. However, the overall degree of hydration remained relatively modest, suggesting that 20 °C, while more favourable, was still suboptimal for complete hydration.

The 30 °C mix demonstrated the highest total weight loss, especially in the 200–600 °C and higher ranges, corresponding with the highest degree of hydration (43.82%). This indicates optimal hydration conditions at 30 °C, resulting in the extensive formation of hydration products, such as C-S-H and ettringite.

In the 35 °C mix, significant weight loss was also observed, though the degree of hydration was slightly lower than at 30 °C. This suggests that while hydration was extensive, it may have been affected by accelerated kinetics, resulting in early setting or less stable hydration products.

The 50 °C mix showed considerable weight loss at low temperatures but exhibited the lowest overall degree of hydration (24.53%). This indicates rapid early-stage hydration, followed by inefficient or incomplete reaction of the binder particles. Excessive temperatures may have led to premature setting, reduced workability, and lower formation of long-term stable hydration products.

6. Discussion

The integrated findings from thermal modelling, TGA, SEM, and EDS analyses provide a comprehensive understanding of how initial mixing water temperature affects the hydration behaviour, microstructure, and thermal evolution of CPB.

Simulation results revealed that the highest temperatures within the stope occurred near the bottom lifting layers, where the paste was in direct contact with the rock. Increasing the mixing water temperature resulted in higher peak temperatures and steeper gradients within the stope. The 50 °C mix exhibited the highest peak temperature (85.6 °C) and the greatest temperature gradient (40.5 °C), whereas the 5 °C mix showed the lowest temperature (55.7 °C) and gradient (16.8 °C). These outcomes highlight the significant influence of initial water temperature on in-stope thermal behaviour during the curing period. The modelling was based on simplified boundary conditions, steady-state heat transfer assumptions, and generic material properties. Future studies could enhance its accuracy by utilizing CAD-defined geometries and more precise material data.

TGA results indicated that the degree of hydration improved with increasing water temperatures up to 30 °C, beyond which it declined. The 30 °C mix showed the most complete hydration (43.82%), while the 50 °C mix exhibited the lowest degree of hydration (24.53%). These results align with the thermal findings—higher temperatures enhance hydration kinetics but may lead to less effective long-term hydration when too high. It should be mentioned that although the 50 °C mix exhibited the highest early heat release, TGA results confirmed that its overall degree of hydration was lower, indicating that excessive temperatures may accelerate initial reactions but hinder complete C-S-H formation due to premature setting or microstructural disruption.

At 5 °C, the SEM images revealed poorly developed hydration products and an open, porous matrix. This suggests that the low temperature significantly slowed the hydration reactions, resulting in incomplete bonding and a weak microstructure. In contrast, the samples mixed at 30 °C and 35 °C exhibited denser matrices with reduced porosity and fewer unreacted particles. Interestingly, while the 50 °C sample initially demonstrated a more intense hydration reaction, the SEM images revealed microcracks and signs of early-age thermal stress. This is likely attributed to excessive heat generation during hydration, potentially due to incomplete or unstable hydration.

EDS confirmed the presence of major hydration elements, including Ca and Si, which form critical phases such as C-S-H. The presence of trace elements (Cl, Na, Fe, Mg, Al, and S) provided additional insights into the potential formation of secondary minerals. The data supported the conclusion that higher temperatures increase the formation of hydration products, but excessive heat may disrupt phase stability.

Collectively, the results demonstrate that mixing water temperature significantly affects the hydration, thermal behaviour, and microstructural development of CPB. Moderate temperatures (20–35 °C) promote optimal hydration kinetics and stable product formation. In contrast, low temperatures slow the reaction rate, while excessively high temperatures accelerate early reactions but compromise long-term hydration and material integrity. Therefore, maintaining a controlled water temperature during mixing is essential for achieving durable and thermally stable CPB in deep mining applications.

7. Conclusions

This study has examined the effect of mixing water temperature on the thermal, compressive strength, and microstructural evolution of cemented paste backfill (CPB) in deep underground mining environments. The findings highlight the critical role of initial thermal input in shaping hydration kinetics, generation of internal heat, and development of strength. Key conclusions include

• Moderate mixing water temperatures (20–35 °C) result in optimal hydration, denser microstructures, and higher compressive strength.
• Excessively high temperatures (≥50 °C) accelerate early hydration but lead to incomplete binder reactions, increased porosity, and reduced long-term strength due to premature setting and thermal microcracking.
• Thermal modelling revealed that higher initial temperatures significantly increase peak internal temperatures and thermal gradients, especially near the stope base, posing risks of thermal-induced damage.
• TGA, SEM, and EDS analyses confirmed that degree of hydration and microstructural integrity are temperature-dependent, with 30 °C yielding the most favourable results.
• UCS testing showed a strength peak at 30 °C, with a decline at 50 °C, aligning with microstructural and hydration trends.

These results underscore the importance of controlling the mixing water temperature in paste plant operations. By maintaining water within an optimal range, mining operations can enhance CPB’s performance, reduce binder consumption, and mitigate thermal risks. Future studies should explore X-ray diffraction and Fourier-transform infrared spectroscopy to further validate hydration products and enhance microstructural insights; long-term durability and deformation behavior under cyclic loading; field-scale validation of thermal modeling predictions; and integration of temperature control systems in paste plant design for operational optimization. Moreover, implementing temperature control strategies in paste plants offers a practical and economically viable solution to improve backfill’s quality, reduce material costs, and enhance stopes’ stability in deep mining operations.