1 Introduction

Underground mining plays an important role in the mineral resource recovery, a large number of untreated voids have been left by underground mining, which have become one of the major hazard sources in mine (Zhou et al. 2013). There are lots of accidents caused by the collapse of voids all around the world. While the disposal and treatment of mining waste (mainly, tailings and waste rocks) produced from mine and mills have become a serious problem in many countries (Yilmaz et al. 2014a). In recent years, research has been focus on the tailings treatment technology to minimize the harmful effect to the underground environment. Backfilling technology is a simple and effective method, which can not only protect the land subsidence, but also decrease surface tailings storage requirements (Cao et al. 2018a, b). As a relatively new mine waste management technology, CCTB is widely applied in underground mine operations around the world because of its technical and economic advantages (Ke et al. 2015). Lots of cementitious materials are used as replacement of clinker to reduce cost, such as gangue (Wu et al. 2015; Franks et al. 2015; Cruz et al. 2015), fly ash (Hamzaoui et al. 2016; Chien-Hsing et al. 2015; Nawaz et al. 2016; Chousidis et al. 2015), water quenched slag (Jiang et al. 2015; Li et al. 2013; Feng et al. 2012). Cement is the most widely used construction materials for its high mechanical strength. Uniaxial compressive strength (UCS) is the most important property of CTB and concrete, other mechanical parameters such as compressive strength, tensile strength, elasticity modulus and passion rate all are related to the compressive tests closely. Plentiful experimental tests have been finished about compressive strength development of concrete and other building materials (Liu et al. 2015; Wang and Park 2015). Many researchers have studied the mechanical and physicochemical properties of cemented paste backfill (CPB). Kesimal et al. (2002) found that the chemical and mineralogical changes with time that affect the mechanical strength of the fill. Yilmaz et al. (2003, 2012, 2014b, 2018) also found that the factors affecting the strength and stability of paste backfill are the characteristics of binder, solids, and water and their mix rates in the filling mass. Ghirian and Fall (2014) and Cui and Fall (2015) had been studied the evolution of coupled thermal-hydraulic-mechanical-chemical (THMC) properties of underground CPB. Yin et al. (2012) investigated the effect of solid components on the rheological and mechanical properties of CPB. Yılmaz et al. (2014) also investigated the effect of binder type/dosage, water to cement ratio (w/c) and fines content (< 20 um) of the tailings on ultrasonic pulse velocity (UPV) of cemented paste backfill (CPB) samples. At the same time, other scholars had done lots of researches on backfill (Morteza Sheshpari 2015; Guo Lijie et al. 2015; Li et al. 2016a, b).

However, reports about the quantitative relationship between c/t, SC and the UCS values are scarce. In this paper, an original laboratory investigation was undertaken to characterize the SC and c/t effect on the mechanical properties of CCTB. The major objective of this paper is to highlight the influence of the factors of c/t and SC on the tailings backfill strength acquisition. The main objectives of this study are as follows:

  • To obtain the effect of the c/t and SC on the strength of CCTB.

  • To investigate the relationship between c/t and the UCS of CCTB.

  • To obtain the relationship between SC and the UCS of CCTB.

The results of the study can provide significant reference for the backfill design of underground mining.

2 Experimental Program

2.1 Materials

2.1.1 Tailings

The coarse tailings used in this study are provided by a gold mine in southeast of China . The particle size distributions of the coarse tailings are measured using a Winner2000 particle size analyzer in accordance with ASTM D421 and D4222 procedures. The particle size distribution (PSD) is shown in Fig. 1. The chemical compositions of coarse tailings used are given in Table 1.

Fig. 1
figure 1

Particle size distributions of coarse tailings

Full size image
Table 1 Chemical compositions of the coarse tailings in the tests
Full size table

From Fig. 1, it can be easily found that tailings contain sand-sized and silt-sized particles of 6.82 and 93.18%, respectively. We can also find that the most abundant mineral with in coarse tailings was SiO2 (63%), Al2O3 (16.4%), CaO (1.89%) and MgO (0.48%).

2.1.2 Binder and Water

Water, binders and coarse tailings with different mix proportions are used to prepare the CCTB slurry specimens. Ordinary Portland cement (OPC) 32.5 R with granulated blast-furnace slag (GBFS), fly ash and silica fume was used to as ingredient. Tap water was used to mix cement and coarse tailings. The pH of the tap water is 7.01, which is basically neutral.

2.2 Mixture Preparation

Tap water was used to mix the binder and coarse tailings. Five types of SC values with 63 wt%, 65 wt% wt%, 68 wt%, 70 wt%, 72 wt% and four types of c/t values with 1/4, 1/6, 1/8 and 1/10 were prepared. The required amounts of coarse tailings and cement were weighed and mixed in a metal container. The calculating volume of tap water was weighed by a measuring cup, and the mixture was working until a homogeneous slurries was obtained. The CTB sample used for UCS testing is a cube with a side length of 7.07 cm in accordance with the ASTM C109/C109 M-16a standard procedure (C109/C109 M-16a 2016). The CCTB slurries were added into 70.7 mm × 70.7 mm × 70.7 mm metal cubes. Setting the curing environment was 20 ± 1°C with a relative humidity of 92 ± 5%. The curing periods was 3 days. Figure 2 shows the maintenance process of CCTS samples, while Fig. 2a shows the curing box, Fig. 2b is the metal cubes and Fig. 2c is the produced samples.

Fig. 2
figure 2

Maintenance process of CCTS samples, a curing box, b metal cubes, c produced samples

Full size image

2.3 Uniaxial Compressive Testing and Procedure

Uniaxial compressive strength (UCS) can directly be used to measure the compressive value of different kinds of materials. To determine the strength of CCTB gained based on the GB/T1767-1999 Chinese standard procedure, the SANS electronic universal testing equipment (having a maximum pressure of 300 kN) was used at a loading rate of 0.5 mm/min (Xue et al. 2018). The compressive device (as shown in Fig. 3) was used to investigate the relationship between compressive strength and axial displacement. This device was mainly composed by a loading plate, a pressure head, a base head.

Fig. 3
figure 3

Unconfined compressive device

Full size image

The uniaxial compressive strength is defined by the ratio of the axial peak force F (N) and the cross section area A (m2).

$$UCS = {F \mathord{\left/ {\vphantom {F A}} \right. \kern-0pt} A}$$
(1)

2.4 Scanning Electron Microscope Testing

In this test, the S-360 scanning electron microscopy (SEM) system was used to analysis the microstructure and texture of CTB samples. The SEM method provides some advantages, such as high resolution, large magnification time, and 3D image views (Cao et al. 2018).

3 Results Analysis

Different SC values (63 wt%, 65 wt%, 68 wt%, 70 wt% and 72 wt%) and c/t values (1/4, 1/6, 1/8 and 1/10) of the CCTB specimens were used in this experiment to analyze the relation among UCS, SC and c/t. Table 2 presents the result of UCS under different c/t, SC and curing period values.

Table 2 Result of UCS under different c/t, SC and curing period values
Full size table

3.1 The Effect of the c/t on the Development of UCS

The effect of the c/t on the compressive strength of CCTBs is shown in Fig. 4. The c/t is an additional factor that should be considered to be a contributor to the strength of CCTB, which plays an important role in the development of the mechanical performance. The UCS values of the CCTB changed from 0.28 to 0.54 MPa, from 0.30 to 0.59 MPa, 0.41 to 0.82 MPa, 0.48 to 1.06 MPa and 0.57 to 1.31 MPa for 3 days curing period, with SC of 63–72 wt%. The results show that the compressive strength of CCTB increased with increasing of c/t. We also found that the cement content plays an important role in increasing UCS of CTB specimens. The influence of cement content on UCS was easy to understand, since increasing cement can speed up the binder hydration. The calcium hydroxide (CH), calcium silicate hydrate (C–S–H) increase with the cement content, which result in a corresponding increment in the CTB strength (Xu et al. 2018).

Fig. 4
figure 4

Relationship between UCS and c/t

Full size image

To study the relationship between UCS and c/t, different kinds of curve fitting (linear, logarithmic, exponential and power) are calculated (as illustrated in Table 3). In order to improve the accuracy of the calculation results, there are 5 groups of CCTB samples at 3 days curing period employed in this calculation. As demonstrated in Table 3, multiple correlation coefficients (R2) of four kinds of fitting methods are 0.8489, 0.9186, 0.7899 and 0.8760, respectively. It is found that the logarithmic fitting has the highest correlation coefficient. Thus, the relationship between UCS and c/t of the CCTB samples prepared at different SC values are in a logarithmic relationship. Figure 5 shows the fitting result between UCS and c/t for CCTB, as exhibited by the following equation:

$$UCS = a_{1} \ln (r) + b_{1}$$
(2)

where UCS is the uniaxial compressive strength, and r is c/t value. Also \(a_{1}\) and \(b_{1}\) are parameters related to c/t and UCS values.

Table 3 Results of the curve fitting
Full size table
Fig. 5
figure 5figure 5

Correlation of UCS and c/t for CCTB samples under different c/t values: a linear fitting, b logarithmic fitting, c exponential fitting, d power fitting

Full size image

3.2 The Effect of the SC on the Development of UCS

Figure 6 demonstrates the effect of SC (63 wt%, 65 wt%, 68 wt%, 70 wt% and 72 wt%) on the evolution of UCS development versus curing period at the same c/t. It is evident that the UCS value of CCTB samples increase with the rise of SC. In addition, Fig. 6 illustrates that the UCS of CCTB samples increase with the c/t values increase. From the microscopic point, the hydration products (C–S–H gel and enttringite) wrap up the tailings particles and link them together. These products fill the pore space and improve the connecting stress between particles of tailings. Consequently, this can lead to an increase in the mechanical performance. The reason for this higher strength of CTB with higher SC is because higher SC increases the compactness of tailings, and the hydration products are more closely bonded (Fall and Pokharel 2011; Claude et al. 2009).

Fig. 6
figure 6

Relationship between UCS and SC

Full size image

To find the relationship between the UCS of CCTB and SC, different kinds of curve fitting (linear, logarithmic, exponential and power) are calculated (as illustrated in Table 4). In order to improve the accuracy of the calculation results, there are 5 groups of CCTB samples at 3 days curing period employed in this calculation. As demonstrated in Table 4, multiple correlation coefficients (R2) of four kinds of fitting methods are 0.9513, 0.9445, 0.9801 and 0.9766, respectively. It is conducted that, the relationship between UCS and SC of CCTB samples prepared at different c/t values are in an exponential relationship with the highest multiple correlation coefficient (R2). Figure 7 shows the fitting result between UCS and SC for CCTB, as exhibited by the following equation:

$$UCS = {\text{a}}_{2} \cdot {\text{e}}^{{b_{2} \cdot w}}$$
(3)

where UCS is the uniaxial compressive strength, and wis SC value. Also \({\text{a}}_{2}\) and \(b_{2}\) are parameters related to SC and UCS values.

Table 4 Results of the curve fitting
Full size table
Fig. 7
figure 7figure 7

Correlation of UCS and c/t for CCTB samples under different SC values. a Linear fitting, b logarithmic fitting, c exponential fitting, d power fitting

Full size image

According to this fitting result, the UCS values of the CCTB samples can possibly be obtained.

Although the highest multiple correlation coefficients (R2) were found for the obtained equations, it does not mean the goodness-of-fit of these equations. Therefore, t and F tests were used to verify the validity of these equations. The t test compares the computed values with tabulated values using null hypothesis. A 95% level (p < 0.05) of confidence was selected in the t and F tests. If the computed t and F are greater than the tabulated ones, the null hypothesis is rejected. The results of t and F tests are shown in Table 5. As shown in Table 5, the computed t and F values are both greater than the tabulated ones, indicating the validity of those equations.

Table 5 Results of t and F tests for Eqs. (2) and (3)
Full size table

3.3 Hydration Products and Microstructure Analysis

In this test, the authors used S-360 to scan the fracture of CCTB to figure out the microstructure features of CCTB after a long period of curing time. The samples were prepared according to the size of 10 mm × 10 mm × 10 mm and are subjected to a carbon spray treatment for two or more times for each sample. Then, every sample is analyzed under different magnification factors and thus the SEM and EDS results are obtained, which are shown in Fig. 8.

Fig. 8
figure 8

SEM images and EDS spectra of CCTB

Full size image

It was found that the hydration products are calcite (aft) and C–S–H gel, and the elements of hydration products mainly include C, O, Mg, Al, Si, K and Ca. There are several voids on the surface of the samples, it can be explained that the hydration is not enough because of the short time hydration.

4 Conclusions

Backfill slurry solid content (SC) and cement-to-tailings ratio (c/t) are critical to the quality of the cemented tailings backfill. Accurate control of slurry’s SC and c/t is a big problem for mining industry all around the world. In this study, the effect of c/t and SC on the early age strength of CCTB are investigated. CCTB samples made at different c/t values (1/4, 1/6, 1/8 and 1/10) and SC values (63, 65, 68, 70, 72 wt%) are subjected to UCS measurements at 3 days of curing period. The UCS values of CCTB samples increase with the increase of c/t and SC. It is shown that the UCS and corresponding c/t values of CCTB are in a logarithmic function with each other. Also the relationship between the UCS and SC meets an exponential distribution.

The laboratory results obtained in this paper is based on a gold mine in southeast of China. However, the parameters such as tailings type, mineralogy, and cement are different for each mine. Thus, this study can only provide a basis for the backfilling design of the studied gold mine.

In future research works, the authors will consider applying the results obtained in this test to a metal mine backfilling. We also demonstrate the results of the laboratory research based on the technical problems that may arise during the backfilling process.