Study on the evolution of dissolved organic matter in the underground storage of mine water

Open-pit coal mining is one of the methods to obtain coal resources by manually stripping the overlying strata of coal seams. It is a primary mining method in major coal-mining countries such as the United States, India, Indonesia, Australia, and Russia. In 2018, the output of open-pit coal mines in all the coal-producing countries accounted for more than 50% of the total output, and it even reached more than 90% in some countries (Monjezi et al. 2009). The deepening coal exploration in China recently has led to the discovery of integrated coal fields suitable for open-pit mining (He et al. 2008). Most of these coal mines are concentrated in arid and semi-arid areas of China, such as Xinjiang, Inner Mongolia, and Shanxi Province. While coal is being obtained in large quantities, a large area of the land has been stripped and occupied, and the underground water system has been destroyed. This has led to compounded disasters and the pollution of solids, water, and air. Realizing the coordinated development of mines and the ecological environment while exploiting open-pit coal resources has become a recent research trend (Hou et al. 2019). In recent years, the ecological restoration technology for mines is mainly based on surface regreening. However, this technology also needs a large amount of water (Liu et al. 1999; Lu et al. 2021).

Table 1 shows the water demand data of the Baorixile open-pit mine of the national energy group. The water demand increased from 2005 to 2014 since the water requirement in open-pit mine production increased yearly with the annual amount of coal stripped. However, due to the climatic conditions in this area, the mine drainage in summer could not meet the overall water demand. On the other hand, the demand for water required for dust reduction and greening dropped sharply in winter due to the cold climate. Therefore, the mine drainage could not be utilized, and a high discharge fee had to be paid in winter. Following this, a technical framework for the underground reservoir of the open-pit coal mine was proposed. It recommended the use of the reconstructed sandstone aquifer in the open-pit coal mine waste dump for water storage (Fig. S1). This led to a reduction in ineffective evaporation, and the storage in winter and utilization in summer were achieved through winter pumping and summer injection, respectively. However, mine water contained some trace organic pollutants that pose a risk of deterioration in the water quality during long-term storage (Mayes et al. 2008; Sun et al. 2009). One of the major sources of dissolved organic matter (DOM) contamination in the raw groundwater was the infiltration by perennial rainwater. The other source was the pollutants leaching out in groundwater through the flow of mine water in the shaft (Wright & Burgin 2009). During the artificial recharge of mine water into the underground reservoir, there were changes in the chemical parameters, such as dissolved oxygen and CO₂ and the occurrence state of the overlying strata, thereby disturbing the steady state of the system and expediting the dissolution and release of DOM in mine water (Cook 2009). In addition, the mine water DOM and the mineral components in the stripped materials were prone to adsorption–chelation–ion exchange–precipitation reactions that lead to changes in the DOM composition and concentration in the water body. The roof of the strata above the coal seam consists of mainly the quaternary sand layer and mudstone. The difference in lithology was mainly manifested by the difference in mineral composition. Furthermore, the different lithology and surface characteristics can influence the quality of water during recharge. The major component of sandstone is SiO₂, which is relatively stable. During water infiltration, surface filtration, sedimentation, and other reactions were frequent (Lundegard & Trevena 1990). The mudstone was rich in clay minerals, such as kaolin, which can undergo ion exchange reactions with some cations in water under the influence of surface charges (Du et al. 2018).

Previous studies mostly considered the evolution processes of conventional inorganic ions under the influence of mixed materials. However, when compared with inorganic ions, organic compounds exhibited higher covertness, longer periods, and stronger pollution (Burrows & Whitton 1983). Considering the case studies of sandstone and mudstone, and mine water DOM as the research object, systematically designed underground mine water storage experiments were conducted to understand the evolution of mine water DOM from various water storage materials. This study aimed to understand the influence of the type of the water storage medium and storage time on the DOM composition and concentration in mine water, thereby evaluating the dissolution and release of DOM from sandstone and mudstone. Finally, a clean water storage method suitable for underground reservoirs of open-pit coal mines was introduced.

The waste materials were collected from 5 cm beneath the surface of the Baorixile waste dump with a circular knife. The gravel with very large particle sizes was removed, and the soil samples were divided into sandstone and mudstone using a sieve shaker.

From January to February 2020, pit water was collected from the bottom of the Baorixile open-pit mine, placed in a light-proof plastic bucket, and brought to the laboratory for further experimental work. Table 2 shows the results of total organic carbon (TOC) and UV254 values of water samples.

• (1)

  The raw materials were segregated and placed in two plastic buckets, then rinsed with ultrapure water, and dried in an oven at 105 °C.

• (2)

  In the first sampling batch, five clean 500-mL conical flasks were taken and numbered 1–1, 1–2, 1–3, 1–4, and 1–5. About 200 mL of mine water was added to each of the flasks. Furthermore, 200 g of sandstone segregated in step 1 was added to the flask numbered 1–1, 200 g of mudstone segregated in step 1 was added to the flask numbered 1–2, 200 g of 1 : 1 sandstone–mudstone mixture was added to the flask numbered 1–3, 200 g of 1 : 2 sandstone–mudstone mixture was added to the flask numbered 1–4, and 200 g of 2 : 1 sandstone–mudstone mixture was added to the flask numbered 1–5. Sampling was conducted every 7 days during the experimental period, and a total of five samples were taken in 35 days. In the second, third, fourth, and fifth rounds of sampling, the flasks were numbered 2–1, 2–2 … , 3–1, 3–2 … , 4–1, 4–2 … , and 5–1, 5–2 … , respectively (Figure 1).

• (3)

  The environmental factors such as sunlight and temperature were accurately simulated during underground storage. After sample loading, the conical flasks were placed in a constant-temperature reaction box with the temperature set at 5 °C. The flasks were wrapped in a black plastic bag to create a dark environment.

• (4)

  On days 7, 14, 21, 28, and 35, 30 mL of the water sample was pipetted from the upper one-third area of the conical flask and immediately analyzed using the three-dimensional fluorescence spectrometer test.

The chemical characteristics were measured using the Brunauer, Emmett and Teller (BET) Fully Automated Surface Area and the Porosity Analyzer (Micromeritics Inc., United States).

X-ray powder diffractometer (XRD) was used to test the surface properties of the medium (Tristar II 3020M at the XRD facility of North West Geological Institute of Non-Ferreous Metals, Xi'an.)

DOMs in actual wastewater were characterized by excitation–emission matrix (EEM) fluorescence on a Hitachi F-7000 fluorescence spectrophotometer (Hitachi).

Excitation wavelengths were scanned from 245 to 400 nm at 5 nm intervals, and emission was measured at each excitation wavelength from 300 to 550 in 2 nm increments. Blank EEMs were routinely collected using deionized water. Sample fluorescence intensity was adjusted by subtracting the intensity of the deionized water.

The algorithm model was represented by A/B/C, containing a_if/b_if/c_if elements. A/B/C was solved by the alternating least-squares method to minimize the residual sum of squares after fitting (Chen et al. 2003). The specific iterative solution process followed is given below:

• (1)

  Determination of the fluorescence factor F in the mixed sample.

• (2)

  Selection of the relative excitation spectrum matrix A and the emission spectrum matrix B randomly.

• (3)

  Calculation of row k in the concentration matrix C using the A/B values as follows:

In the equations above, diag represents the diagonal matrix, m is the number of iterations, T represents the transpose of the matrix, and ※ is the Hadamard product.

Table 3 shows the particle size distribution of the waste materials in the waste dump. The waste materials were mainly composed of sandstone having a larger particle size and mudstone having a smaller particle size. The sandstone had a D10 of 0.29 mm and a D50 of 2.29 mm, accounting for a higher fraction of the waste material.

The waste materials in the waste dump were characterized by an X-ray diffractometer to determine the mineral composition, and the test results are shown in Figure 3(b). It can be observed that there were strong diffraction peaks in both products, indicating good crystallinity. JADE6 analysis revealed that the two samples were similar, having SiO₂, KAlSi₃O₈, and CaAl₂Si₂O₈ as the major components (Caracas & Ballaran 2010).

S1, S2, S3, S4, and S5 represent the evolution of DOM in sandstone, sandstone:mudstone = 1 : 2, sandstone:mudstone = 1 : 1, sandstone:mudstone = 2 : 1, and mudstone, respectively (Figure 4). When the water-containing medium was sandstone, a slight fluorescence reaction at 260 and 460 nm was observed when the water was stored for 7 days. On day 14, the fluorescence intensity reached the minimum value, possibly due to the adsorption of some DOM in water on the surface of suspended solids (Xu et al. 2019). However, with a further increase in time, the fluorescence intensity gradually increased and reached the maximum value at 35 days. When the sandstone–mudstone ratio in the water-bearing medium reached 2 : 1, the DOM was similar to that in other water storage media. Therefore, the composition of DOM and fluorescence intensity exhibited no significant change on day 7 or 14. However, the fluorescence intensity increased significantly on day 21, and the DOM concentration gradually increased with time, reaching a maximum value of 5,870 AU at 35 days. When the sandstone–mudstone ratio in the water-bearing medium reached 1 : 1, the composition and fluorescence intensity of DOM showed no significant change on day 7 or 14, indicating that the increase in the mudstone fraction had no significant impact on the release of DOM in the water body over a short period. When the water storage time reached 21 days, there was a significant increase in the fluorescence intensity. Later, the fluorescence intensity gradually increased, reaching a maximum value of 6,896 AU on day 35. When the sandstone–mudstone ratio in the water-bearing medium was 1 : 2, the composition of DOM, i.e., fulvic acid (260 and 460 nm), was unchanged after continuous water storage for 7 days. However, the fluorescence intensity was significantly enhanced. The fluorescence intensity gradually increased with the increase in water storage time, reaching a maximum value of 10,350 AU on day 35.

Fig. S2 shows the temporal variations in fluorescence intensities of organic components in the artificial water storage medium. The total fluorescence intensity of DOM in the artificial water storage media was lower than that of the waste materials in the waste dump. When artificial zeolite was used as the water storage medium, the fluorescence intensities of the four components followed a trend of C4 > C3 > C2 > C1. When activated carbon was used as the water storage medium, the fluorescence intensities of all four DOM components were less than 100 AU for 7 days, indicating complete removal. This strong adsorption capacity was because of the high specific surface area of the activated carbon (Tan et al. 2008). However, with the water storage time, the fluorescence intensity of the C4 component showed a gradual increase, followed by a decrease. The fluorescence intensity of the C2 component increased gradually for 28 days and then increased rapidly, reaching the maximum value of 760 AU on day 35.

The fluorescence peaks at 270 and 410 nm (Figure 2) represent UV fulvic acid-like fluorescence, and the source was mostly from the raw underground water. The fluorescence peaks at 250 and 330 nm represent visible tryptophan-like fluorescence, which was mainly introduced by human activities and industrial sewage discharge, indicating that the mine water sample was polluted (Yamashita & Jaffé 2008).

Hysteresis was observed in the adsorption–desorption profile due to the weak adsorption between the adsorbate and the adsorbent (Figure 3(a)). Typical mesoporous adsorption was observed, i.e., the pore size was within 2–50 nm (Liang et al. 2009). According to the Langmuir formula, the waste material had a specific surface area of 14.5961 m²/g and demonstrated certain adsorption performance. The large particle size test indicated the presence of quartzite, anorthosite, and muscovite. Similarly, the small particle size test indicated the presence of minerals, such as montmorillonite and vermiculite, possibly because the component with a lower content in the waste material had a relatively large particle size (Flores et al. 2018).

The fluorescence intensity increased significantly when the water storage time reached 21 days, which indicated that humus in the mudstone was released at this stage. Therefore, the optimal water storage time for sandstone was 21 days. In other words, a period longer or shorter than 21 days would lead to an increase in DOM. After 7 days of continuous water storage, the composition of DOM remained unchanged, but the fluorescence intensity increased significantly, which indicated that there was a continuous increase in the humus concentration in water with the increase in the mudstone fraction. This is because a large amount of humus accumulates during the process of mudstone deposition. The composition of DOM and fluorescence intensity exhibited no significant change on day 7 or 14, indicating that the fractions of sandstone and mudstone in the water-bearing medium had little effect on the DOM composition and concentration for 14 days. When the mudstone is in contact with water, it rapidly disintegrates and releases a part of the humus. This is consistent with our previous finding that weak adsorption was observed between the adsorbate and the adsorbent during the BET test. The fluorescence intensity of DOM gradually increased with the increase in water storage time (Kida et al. 2021), reaching a maximum of 8,935 AU on day 35. When compared with that in the other four water storage media, the fluorescence intensity in mudstone increased on day 14, indicating that the increase in mudstone fraction reduced the duration of humic acid release. The fraction of mudstone in the water-bearing medium significantly affected the fluorescence intensity of DOM in the water body. When the water-bearing medium was sandstone, DOM in the mine water was removed by the sedimentation of suspended solids due to the high density of sandstone, and the optimal water storage time was 21 days. When the water-bearing medium was a mixed sandstone–mudstone medium, the fraction of mudstone was positively correlated with the DOM concentration, and the optimal water storage time was 14 days. When the water-bearing medium was mudstone, DOM was rapidly released during the initial stage of water storage.

Artificial zeolite can remove DOM to a certain extent. However, there was no significant change in the fluorescence intensity with time (Figure 5), indicating that the adsorption saturation was reached at the initial stage, and furthermore, the adsorption was stable over time with no desorption (Engel & Chefetz 2016). When the activated carbon was used as the water storage medium, DOM was not detected on day 7 as the activated carbon has a high specific surface area and can adsorb a large fraction of some organic components from water (Kruk et al. 1999). Peaks were observed on day 14, and this may be due to the desorption of small molecules into the water body because of the weak surface van der Waals forces of tyrosine protein (Volokitin & Persson 2002). Therefore, the optimal duration of water storage in the activated carbon medium was 7 days.

The results shown in Figure 7 validated the previous hypothesis of the release of humus from mudstone. In addition, the total fluorescence intensity exhibited a significant positive correlation with the water storage time in the natural water storage medium. On the other hand, lower total fluorescence intensity was observed in the artificial water storage medium, indicating that the artificial water storage medium was more advantageous for controlling the release of DOM.

A sudden increase was observed on day 28 (Figure 8), indicating that the critical period of humus release was 28 days, which was consistent with previous research results (Lee et al. 2018). At the initial stage of water storage, some small molecules of protein-like tryptophan are adsorbed on the surface of the suspended matter and, thus, transferred from the liquid to the solid phase accompanying natural sedimentation. However, these small molecules are released as the water storage time progresses, thereby increasing the C2 fluorescence component. When the medium contains both sandstone and mudstone, the duration of release of humus in the soil decreases. The reason for this phenomenon is that sandstone brings about a change in the agglomeration state of the original mudstone, thereby increasing its specific surface area and leading to an enhanced release of humus (Cao et al. 2020).

The results in Figure S2 (fluorescence intensity C4 > C3 > C2 > C1) again verified that the source of the C1 component of the waste materials in the waste dump was humus released from the mudstone. Furthermore, there was no significant change in the fluorescence intensities of the components with time, indicating that artificial zeolite had no physical or chemical effects on the four DOM components in mine water. In other words, the medium was neither able to remove the organic components from water by adsorption nor release such components into water (Turp et al. 2020). The fluorescence intensity of the C2 component was less than 10% of the value for the waste materials in the waste dump (i.e., 10,000 AU), as mentioned earlier in this study. This indicated that the risk of release of organic components in the artificial aqueous medium was lower than that of the waste materials in the waste dump.

The surface characteristics and chemical composition of the waste materials in the waste dump highlighted that the average particle size was 2.29 mm, and the specific surface area was 14.5961 m²/g. The major mineral components were examined to be quartz, calcium feldspar, and muscovite. The waste materials were found to be feasible for water storage due to relatively stable chemical properties.

The water storage medium enhanced the dissolution and release of DOM in mine water, and the fraction of mudstone was one of the key factors. The fraction of mudstone had a significant positive correlation with the concentration of humus in mine water. Therefore, the fraction of mudstone in the water storage media should be controlled.

During the mine water storage, the dissolution and release of DOM components were monitored over 28 days, which indicated that the DOM of various water-containing media increased significantly after 28 days. Therefore, the water pumping and injection for underground mine water storage should be controlled for 28 days.

Xiyu Zhang conducted field sampling and test analysis, and drafted the manuscript. Shuning Dong gave guidance on the outline of the article. Jidong Liang participated in the field test and analysis. Xiaoming Guo participated in the polishing and modification of the manuscript. Lingyun Huang participated in the polishing and modification of the manuscript. All authors read and approved the final manuscript.

This work was supported by the Tiandi Science and Technology Co. Ltd Science and Technology Innovation Venture Capital Special Project under Grant No. 2022-2-TD-ZD005 and Natural Science Basic Research Plan in Shaanxi Province of China under Grant No. 2023-JC-ZD-27.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.