The Potential to Replace Cement with Nano-Calcium Carbonate and Natural Pozzolans in Cemented Mine Backfill

The eﬀectiveness of mine backﬁlling depends on the properties of its constituents. The high cost of cement, which is commonly used as a binder in mine backﬁll, has led researchers to seek alternatives to partially replace it with other binders. This study investigated the potential to use nano-calcium carbonate (NCC) and natural pozzolans (zeolite and pumice) along with Portland cement (PC) in mine backﬁll. Two types of experimental samples were prepared: (1) gold tailings and silica sand to investigate the eﬀect of NCC and (2) nickel tailings to investigate the eﬀect of natural pozzolans. The unconﬁned compressive strength (UCS) was measured for samples cured for up to 56 days. Moreover, selected samples were subject to mercury intrusion porosimetry to investigate microstructural properties. Results show that addition of NCC did not improve the UCS of backﬁll prepared with gold tailings and cured for 28 days, whereas a dosage of 1% NCC in backﬁll samples prepared with silica sand improved UCS by 20%, suggesting that the gold tailings negatively aﬀected strength development. Natural pozzolans, in particular, 20% zeolite, had 24% higher UCS after 56 days of curing compared to samples prepared with PC and thus have the potential to partially replace cement in mine backﬁll.


Introduction
Mine backfilling is the practice of filling underground voids created during ore extraction in underground mining operations to (1) provide stability for the underground working environment [1], (2) reduce negative environmental impacts from tailings' storage on the surface [2][3][4], and (3) maximize extracted ore volumes, such as in room and pillar mining [5,6]. ree common mine backfill materials are rock fill (RF), hydraulic fill (HF), and paste fill (PF) [1,[7][8][9]. To make RF, waste rocks from the surface or underground are crushed (typically to 16 mm) and possibly cemented, depending on the required strength [10][11][12]. By decreasing waste rock piles on the surface, adverse environmental impacts and rock subsidence are avoided [11][12][13][14]. However, to ensure backfilling efficiency, binder consumption is high [15]. HF is produced during the classification of mill tailings using hydrocyclones: coarse, dense (>70 wt.% solids) underflow is obtained [16][17][18][19]. e high water content of HF facilitates transport through pipelines directly into stopes [19,20]. According to Adiansyah et al. [21], drainage is the most critical factor in designing HF because ineffective drainage can pose a risk to workers if porous barriers are not used, and piping and liquefaction problems are common. PF consists of mill tailings with 15% of the material <20 μm in size and solid concentrations of 70-85 wt.% [15,22,23]. e high solids and low water content can cause issues for transport through pipelines and necessitate addition of a superplasticizer to enhance flowability [24,25].
Mine tailings are waste products from ore concentration in mineral processing plants [26]. eir chemical and physical properties must be considered during backfill design because they can affect backfill mechanical properties [27]. For instance, tailings that contain sulfide minerals are prone to sulfate attack, which can decrease the unconfined compressive strength (UCS) over the long term due to ettringite formation [28]. us, binders that are compatible with the tailings are added to backfill at low concentrations to enhance UCS and other mechanical properties [29]. Cement is typically the main binder, but it is expensive. erefore, alternative binders are required [30].
Recent advances in nanotechnology have presented the opportunity to use nanomaterials as supplementary cementitious materials [31]. e fine particles act as nuclei for cement to accelerate hydration and improve UCS at early stages [32]. Nano-silica, carbon nanotube, and nano-calcium carbonate (NCC) have been shown to increase cement base material strength [33][34][35]. For example, adding NCC to cement and silica fume accelerated the pozzolanic reaction at an early age and compensated for the low initial strength [36,37].
Supplementary cementitious materials such as natural (e.g., volcanic pumice and zeolite) and artificial pozzolans (e.g., fly ash) could partially replace cement to prevent sulfate attack [38,39]. Pozzolanic materials in mine backfill can reduce binder costs by more than 15% on an annual basis [30,[40][41][42]. At concentrations ranging from 10 to 30%, natural pozzolans improved concrete mechanical properties, permeability, durability, and transport properties [43][44][45][46]. Volcanic pumice properties vary by source and location due to different ash formation conditions, mineral components, and grain size characteristics [47][48][49]. Addition of zeolite enhanced the mechanical properties, durability, and performance of concrete [50][51][52][53]. Many industrial wastes also have pozzolanic properties [54]. For example, clay pozzolans, produced through calcination at temperatures ranging from 700 to 900°C [2,55], are inexpensive, have low CO 2 emissions compared to cement, and improve backfill mechanical properties [56,57]. Addition of 10% clay pozzolans to mine backfill produced maximum UCS and an annual savings of >7% [2,58]. e application of NCC and natural pozzolans has primarily focused on concrete production: possible effects on backfill mechanical properties have not yet been widely investigated. erefore, this paper investigates partial replacement of cement by NCC, pumice, or zeolite in cemented mine backfill. To investigate the effect of NCC, backfill samples were prepared with gold tailings and a superplasticizer to aid nanoparticle dispersion: various dosages of NCC were compared with reference samples containing only Portland cement (PC). To investigate the effect of natural pozzolans, backfill samples were prepared with nickel tailings and various dosages of pumice and zeolite were compared against reference samples containing PC only or a combination of PC and fly ash. Samples were subject to UCS tests after up to 56 days of curing and selected samples were subjected to mercury intrusion porosimetry to investigate microstructural properties.

Materials.
Tailings from gold and nickel mines and silica sand were used to make experimental backfill samples (Table 1). e particle size distribution (PSD) of these materials is shown in Figure 1.
e chemical composition of Type 10 PC with a specific gravity of 3.15, pumice (provided by Hess Pumice Company) with a specific gravity of 2.35, zeolite (provided by Bear River Zeolite), and fly ash are shown in Table 2. e chemical and physical properties of the NCC (provided by US Research Nanomaterials, Inc.) are shown in Table 3.
NCC is characterized by high energy and surface area; thus, surface reactions occur, and particles are prone to adhere on direct contact via magnetic, electrostatic, and Van der Waals forces [37]. To aid NCC dispersion, a superplasticizer (sodium dodecyl sulfate, ermo Scientific Company) was used, with the chemical and physical properties listed in Table 4.
Tap water (pH 7.5) was used to prepare experimental backfill samples. Compounds in mixing water can influence the mechanical properties of backfill and cement hydration [59]. For instance, dissolved calcium and magnesium enhance cement hydration, whereas salty brine decreases binder workability [60].

Experimental Design.
Backfill samples containing NCC were prepared with 7 wt.% (by weight of dry materials) PC as a binder and 78% solids' concentration using gold tailings ( Table 5). Mixtures of PC and 1, 3, and 5% NCC (by dry weight of PC, equation (1)) were initially dry mixed to break up agglomerated nanoparticles that were visible to the naked eye. Dried tailings were then added, and water was gradually added to the mixture to make a paste. Finally, superplasticizer was added (by dry weight of NCC, equation (2)). Samples were poured into 5.08 × 10.16 cm cylindrical molds and cured for 7, 14, and 28 days at 25 ± 2°C and 90 ± 2% relative humidity: where M NCC is the mass of NCC (g), D NCC is the dosage of NCC (%), M PC is the mass of PC (g), M SP is the mass of superplasticizer (g), and D SP is the dosage of the superplasticizer (%). Backfill samples containing 10 and 20% pumice, zeolite, or fly ash were prepared using nickel tailings and PC with a solid concentration of 80% (Table 6; equations (3) and (4)). Samples were cured for 7, 14, 28, and 56 days at 25 ± 2°C and 90 ± 2% relative humidity: where D PC is the dosage of PC (%), M T is the mass of tailings (g), M POZ is the mass of pozzolans (g), and R is the percentage of PC replaced by pumice, zeolite, or fly ash (%).  (Figure 2). UCS values are reported as means.

Mercury Intrusion Porosimetry.
Mercury intrusion porosimetry (MIP after ASTM D4404-18 [62]) was used to describe the pore structure of select backfill samples made with 20% pumice, zeolite, and fly ash after 56 days of curing (Table 6). e macrosize and volume distribution of pores inside materials are commonly measured using MIP, which is a well-known technique, although it has adverse environmental effects.

NCC Backfill Samples.
e mean UCS of backfill samples increased with curing time at all NCC dosages and was consistently higher in reference samples than those made with 1% NCC and 10-100% superplasticizer (Figure 3), 3% NCC and 10-80% superplasticizer (Figure 4), or 5% NCC and 10-30% superplasticizer ( Figure 5). us, the maximum UCS was observed for reference samples on day 28 (∼0.80 MPa). e differences between treatment and reference sample means became more pronounced with NCC content, such that, on day 28, the UCS of the 30% superplasticizer sample was 5% (Figure 3), 25% (Figure 4), and 36% ( Figure 5) lower for the 1, 3, and 5% NCC samples, respectively. At 1 and 5% NCC, addition of superplasticizer had a negative effect on UCS on each curing date ( Figures 3  and 5), but at 3% NCC, it had a positive effect on UCS (Figure 4). Figure 6 shows the samples with the maximum UCS on each curing day for 1, 3, and 5% NCC. Relative to the reference sample, increasing NCC from 1 to 5% decreased the UCS of backfill samples by 24-56% on day 7, 3-32% on day 14, and 5-30% on day 28.
Since adding NCC did not improve the UCS of backfill samples, the gold tailings were replaced with a more neutral material to see if the loss of strength was due to an interaction between the tailings and the NCC. Samples prepared with silica sand augmented with 7% PC or 7% PC + 1% NCC were tested for UCS. e UCS increased with curing time as above (Figure 7). Addition of 1% NCC increased the UCS from 0.71 to 0.82 MPa by day 7, did not affect UCS on day 14, and by day 28, increased the UCS    (Figure 1), which might contribute to the effectiveness of NCC. Previous studies have shown that NCC increased the UCS of concrete mixtures. Concrete has larger and relatively neutral in-earth components compared to backfill tailings. It is believed that physical (PSD) and chemical (components) differences contribute to NCC efficiency early in UCS development.

Natural Pozzolan Backfill Samples.
As with the NCC experiments, the UCS increased with curing time from 7 to 56 days (Figure 8), whereas reference samples had the highest UCS on day 14, and backfill samples prepared with 20% zeolite or 10% fly ash after 56 days of curing had the highest UCS (0.66 MPa), 24% higher than the reference UCS on that day. e porosity and pore size distribution of samples containing 20% pumice, zeolite, and fly ash are shown in Figures 9 and 10, respectively. Samples containing 20% fly    Table 5, for mixture designs. 4 Advances in Civil Engineering  Table 5, for mixture designs.  Table 5, for mixture designs.  Table 5, for mixture designs. ash had higher porosity (∼33%) compared to samples containing natural pozzolans (∼31%). is could explain the lower UCS value obtained for samples containing fly ash after 56 days of curing ( Figure 8). Samples containing fly ash and natural pozzolans had similar range of pore size (∼0.006-10 μm), as shown in Figure 10. Adding 20% natural pozzolans decreased the number of pores, especially in the range of 0.01-0.02 μm (Figure 10).  Table 6, for mixture designs.   Table 5, for the NCC mixture design.

Conclusions
is study investigated the potential replacement of cement by NCC and natural pozzolans in cemented mine backfill. Key conclusions are summarized as follows: (i) Addition of NCC did not improve UCS relative to reference samples prepared with PC and gold tailings (ii) 1% NCC without superplasticizer yielded the highest UCS among the experimental samples (iii) Samples containing silica sand instead of gold tailings improved UCC, suggesting that the gold tailings had a negative effect on strength development (iv) Addition of superplasticizer to NCC samples did not improve the UCS (v) Natural pozzolans may have some potential to partially replace cement in mine backfill (vi) Addition of 20% zeolite or 10% fly ash improved UCS the most on day 56 (24%) relative to the reference sample (vii) Samples prepared with 20% natural pozzolans had a similar pore size distribution (∼0.006-100 μm) and porosity (viii) Additional studies are required to examine the effect of particle size and chemistry on the UCS of backfill samples prepared with NCC

Data Availability
e data used to support the findings of the study are available within the article.

Conflicts of Interest
e authors declare no conflicts of interest.