Morocco’s mining sector plays a crucial role in driving economic and social growth and development (Thurber and Morse 2015; Taib 2017). This is primarily due to the country’s diverse geological features, which include a vast variety of formations from different geological epochs including the Precambrian and Tertiary periods (Michard et al. 2017). As a result of this geological variety, Morocco has become a prominent global player in the mining industry, with anthracite coal being one of its most significant minerals (Bentaibi et al. 2021; Australien trade and investment commission 2023). This type of coal boasts the highest carbon content and energy density of all coal varieties and is predominantly extracted from mines located in the northern parts of the country (World Coal Institute 2005). However, the mining process generates substantial amounts of sulfide minerals during the extraction and treatment phases, which could lead to significant environmental challenges if not properly managed (Dai et al. 2012; Finkelman and Tian 2018). Nevertheless, with appropriate regulations and practices in place, the mining sector in Morocco may continue to grow while having a reduced negative impact on the environment.

The produced mining waste can be classified into two main categories. The first category is waste rock, which are composed of removed rocks during the process of extracting ore and are managed in large surface piles or dumps. The second category is slurry impoundments, which contain fine particles, water, and chemicals that are generated during ore treatment, concentration and other mining operations kept in tailings ponds (Aubertin et al. 2002; Bussière et al. 2004; Amos et al. 2015; Elghali et al. 2019a). Coal waste can be further classified into three types: solid coal waste, which is a mixture of coal, soil, and rock generated during coal processing; liquid coal waste, which is created during the washing treatment process and contains water filled with toxins; and coal ash, which is produced during the burning phase, released into the smokestack, and contains toxins (Woodruff 2013; Matsyuk et al. 2020). Coal mine waste rock (CMWR) are highly toxic, containing high levels of mercury, arsenic, and nickel, and can pose significant risks to both the environment and human health (Fecko et al. 2013; Dutta et al. 2017; Islam et al. 2021). These CMWR can cause acidic mine drainage (AMD) if the content of acid generating minerals (sulfide) is higher than that of acid consuming minerals (carbonates). Acid mine drainage causes a decrease of pH levels and an increase of the leachability of metal(loid)s (Benzaazoua et al. 2004; Nordstrom et al. 2015; Dutta et al. 2020). AMD is caused by the oxidation of ferruginous phases such as pyrite and pyrrhotite, which can be oxidized in two ways: direct reaction between oxygen and pyrite to create sulfate and acid, or ferric iron taking the place of oxygen as the primary oxidizing agent (Jamieson et al. 2015; Dold 2016; Parbhakar-Fox et al. 2016).

Improper waste management and disposal not only harms environment by allowing leaching of iron, manganese, and aluminum residues into waterways, as well as creating dust hazards and AMD, but also increases greenhouse gas emissions, a factor that contributes to climate change (U.S. Energy 2022). It is therefore essential to manage and dispose of waste properly to prevent further damage to the environment and protect human health. Recent studies have shown that environmental risks associated with sulfide oxidation can be minimized through the application of clean coal technology (CCT) (Melikoglu 2018). Additionally, proposed solutions include restricting oxygen and water entry (Bussière et al. 2004; Maqsoud et al. 2021), or adding neutralizing agents to reduce the mobility of elements (Elghali et al. 2019b; Igarashi et al. 2020; Qureshi 2021). Another effective approach is pyrite passivation, which prevents the oxidation of sulfides and facilitates the safe and sustainable use of mine waste rock in various fields (Chopard 2017; Tabelin et al. 2017; Roy 2019; Fan et al. 2021). Effective management of waste rock involves several measures, including the reclamation of abandoned mine sites, proper disposal of waste rock piles, and their re-treatment to reduce contaminants. To address these challenges, researchers have explored the possibility of waste rock valorization as a valuable alternative resource by re-mining, taking advantage of the non-negligible of ore still present in the waste material (Falagán et al. 2017; Araya et al. 2020; El Aallaoui et al. 2024b) and the possible presence of significative levels of rare earth elements (REE) (Seredin 1996; Finkelman et al. 2019; Zhao et al. 2019). This approach can be particularly effective when it does not require costly regrinding processes. In addition to re-mining, mine waste rock can be used in other fields as a raw material, either partially or entirely substituting primary resources in building materials and civil engineering (Hoffmann and Huculak-m 2012; Helser et al. 2022). For instance, CMWR can be used as plugging material in road embankments or in cement to save energy due to the exothermic phenomenon during coal burning (Frías et al. 2012, 2016; Addou et al. 2019; Kragovic et al. 2021). Additionally, CMWR can be reprocessed by extracting iron oxides that are concentrated in fine particles, using sifting, gravimetry, and flotation processes for use in painting, building materials as additional ingredient and bio-construction (Darmane et al. 2009; Addou et al. 2017).

The Jerada anthracite mine (JM), located in the Oriental region of Morocco, serves as a prominent example of abandoned mines in the country, leaving behind vast quantities of CMWR estimated at 15–20 million tons since its closure in 2001 (Taha et al. 2016; El Aallaoui et al. 2024a). These CMWR piles lack any rehabilitation plan, posing a significant environmental threat. Particularly concerning the elevated sulfate concentrations in groundwater resulting from the oxidation of pyrite present within these waste materials (Addou et al. 2017). Additionally, the proximity of coal piles to urban areas contributes to the degradation of the aesthetic surroundings (Bendra et al. 2011; Battioui 2013). Previous studies have highlighted the potential for the sustainable reuse of these CMWR in building materials, cement, or road construction due to their high granulometric evolution, sensitivity to water, and low resistance to degradation by abrasion and wear (Belkheiri 2016; Taha et al. 2016; Amrani et al. 2020).

The mining of hard coal can result in the significant production of CMWR, with quantities varying based on several factors such as deposit type, mining method, and geological conditions (Fecko et al. 2013). While this waste may seem like a burden, it contains valuable minerals that can be recovered economically. For instance, pyrite, often viewed as a waste mineral due to its link with AMD, serves as a source of sulfur and may contain recoverable metals like gold and copper. Therefore, it is important to recognize that this waste should not be treated as only a byproduct, but rather as a potential source of valuable resources (Amrani et al. 2020; Taha et al. 2022). To make the most of this opportunity, it is necessary to perform a thorough physical, chemical, and mineralogical characterization of the CMWR. The physical characterization involves assessing the particle size distribution and fineness of the CMWR, which can then be used as raw materials in other industries. The chemical characterization provides information about the proportion of remaining coal ore in the CMWR piles (Little 2015; Dai and Finkelman 2018; Gómez-Arias et al. 2021). The mineralogical assessment plays a crucial role in environmental studies as it elucidates the textural relationships among various minerals and determines the element deportment through comprehensive mineralogy analysis. Moreover, it helps to assess the degree of mineral liberation, which is vital in selecting the most appropriate treatment processes for improved efficiency (Dai et al. 2013; Blannin et al. 2021). The characterization of the spatially distributed carbon content can significantly contribute to creating a digital representation of the CMWR piles using 3D modeling techniques. It is possible then to identify regions with the highest concentration of minerals of interest. This information can be used to guide the extraction process, which will increase the efficiency and effectiveness of mineral recovery (El Aallaoui et al. 2024b). To further enhance the accuracy of assessing mineral resources, 3D modeling can be combined with grade estimation and interpolation methods. Numerous studies have utilized simple estimating tools, such as geophysical surveys, in conjunction with metal grades, to estimate metal tonnages (Barago et al. 2021; Bevandić et al. 2022). To generate a continuous grade model by interpolating grades across sample points, methods like inverse distance weighting (IDW) and kriging are employed (Bargawa 2022; Pereira et al. 2022). These techniques operate under the assumption that interpolation grades of neighboring sample points exhibit greater similarity compared to those situated at a greater distance (Handayani et al. 2019). Incorporating the 3D geological model into the grade estimation and interpolation process can increase the reliability and accuracy of the grade estimates (Kaplan and Topal 2020). Many researchers have employed one or more of these methods to model elements at various depths (Tripodi et al. 2019; Yasrebi et al. 2020; Blannin et al. 2023), or to create geo-metallurgical models by combining multiple estimation methods with other characterization methods (Louwrens 2016; Nwaila et al. 2021). In our study context, the 3D geo-metallurgical characterization can provide insights into the spatial variation of carbon content within Jerada’s CMWR piles. This fills a gap left by previous studies that solely focused on the horizontal variation of carbon through horizontal drilling studies. Our investigation also extends to analyzing and modeling of carbon content across the layers of CMWR piles, considering their layered deposition process. This 3D geo-metallurgical assessment aims to pinpoint areas of abundant carbon resources within the CMWR. By discerning the elemental distribution and consequently identifying the minerals hosting these elements, we can prioritize areas for recovery efforts. The feasibility of such endeavors will be assessed through a comprehensive study of mineral liberation.

This study aims mainly to: i) characterize Jerada CMWR and evaluate the potential for their recovery; ii) investigate the chemical distribution of the remaining carbon content and sulfur in these piles using a 3D block model; and iii) propose guidelines for CMWR piles reprocessing based on their resources’ estimation, chemical and mineralogical properties. This study provides a valuable understanding and novel insights into the (JM) CMWR, which have not been previously modeled for remaining coal or estimated using topo-photogrammetric images obtained by drone surveys. Traditional waste management methods have not been effective, which has led to negative environmental impacts and unsustainable use of limited natural resources. The CMWR in Jerada presents an opportunity to revitalize the region by extracting and utilizing valuable minerals. Transforming these CMWR piles into a source of prosperity is a promising prospect.

2.1 Mine site description

Jerada coal is a historic coal mine in Morocco that is considered to be an anthracite deposit. It is located 50 km from the Algerian border and 60 km away from Oujda city in the Moroccan eastern region. Spanning 25 km from east to west and 4 km in width, it is known for its high-quality coal. The basin of Jerada is part of the Paleozoic massifs of the meseta and is characterized by early syn-metamorphic folds, which support the Viseo-Namurien series in an unconformity (Hoepffner et al. 2005; Michard et al. 2011, 2017). The basin is composed of volcano-sedimentary series from the lower Carboniferous and a Westphalian coal series (A, B, and C) that is covered by discordant Mesozoic series from the Atlasic Highlands (Owodenko 1976; Essamoud and Courel 1998; Chellai et al. 2011). The Westphalian C coal bed, also known as the “productive coal” comprises of eight layers of coal, labeled A through H. However, only layers A, B, C, and F are currently being exploited. The coal basin is divided into a northern basin, where shallow layers A, B, C, and D are found, and a southern basin, where all layers exist and accompanied by several faults and fractures (Owodenko 1976; El Gout et al. 2015; Boufkri et al. 2021) (Fig. 1a). The three exploited layers have a dip angle of 20° to 45°, indicating that the eastern portion of the basin is sinking and offering the potential for further coal resources (Owodenko et al. 1968; Jerada Archives 1990) (for additional details on the history of the JM, please refer to supplementary data Sect. 1). The Jerada coal has a relatively low volatile matter variability (approximately 5%) and low sulfur content, generally below 3%, which reduces the risk of AMD from the waste, though it does not eliminate the possibility (Chellai et al. 2011). The studied zone climate is distinguished by its relatively dry conditions compared to the Atlantic coastal zone at similar latitudes. The annual rainfall averages around 200 mm across the region, with a noticeable decrease from north to south (Thauvin et al. 1969; Battioui 2013).

a Geological context of Jerada province. b, c Real images of piles T08 and T01 respectively, highlighting the substantial height of these piles in Jerada city

During its operation, the JM created 10 CMWR piles labeled T01 to T10 (Fig. 1a). The CMWR piles in Jerada can be classified into two groups: the older group, represented by T08, and the younger group, consisting of T01 to T10. Among these, T08 is notable as the largest and oldest pile, spanning a surface area of 15 hectares and reaching a height of 95 m. It accrued over a 50-year period, starting from the commencement of production in 1936, with CMWR transported to the surface via wagons and conveyed to the summit using conveyor belts (Jerada Archives 1990). The other piles are considered more recent and less elevated piles averaging between 6 and 24 m. These piles consist of a mixture of CMWR from mining operations, which underwent initial mechanical processing and screening, along with loose overburden waste displaying coarse grading, visibly present in the pile’s structure. This study focuses on analyzing three out of the ten piles (T01, T02 and T08) due to limitations in accessibility and sampling, in order to compare the carbon grades between the T08, and the younger ones (T01 and T02). The 3D modeling will primarily focus on T08 due to its accessibility, size, and availability of data for the interpolation methods. The ultimate goal is to apply the sampling and 3D modeling approach to investigate the recoverable coal potential of all the other CMWR piles.

2.2 Sampling process and raw material

Sampling is crucial in environmental and resource estimation, ensuring representative samples validate conclusions. Our study employs a multifaceted approach including photogrammetry with georeferenced drone imagery to construct topographic models of CMWR piles and calculate their volume. Using Datamine RM software, 3D modeling highlights morphology and interpolates high-density point cloud data (Lidar/las). The wireframe-volume tool estimates tonnage by multiplying volume and density (Datamine Corporate Limited 2013). Photogrammetric surveys guide drilling planning, executed with down-the-hole hammer (DTH) techniques at lengths from 2 to 74 m, sampled at 1-m intervals using casing for safety, and each sample weighed around 20 kg. 25 drillings were conducted with a total distance of 622 m across the entire CMWR piles. Table 1 presents the boreholes and their distances in T01, T02 and T08 (for detailed sampling procedure see supplementary data Sect. 2).

T01 and T02 are CMWR piles on the western edge of former JM in SW and NW of Jerada’s urban sector. T01 is a flattened cone with a moderate slope (3° to 4°), covering approx. 154.900 m², extending 820 m N-S with an average width of 200 m and a maximal height of 25 m. Its southern slope is around 66% (Fig. 2a). T02 has a similar shape and orientation but smaller, approximately 450 m long, 180 m wide, and 30 m high, with a steep southern slope (~ 70%) and an area of approximately 69.570 m² (Figs. 2a, b). They rank second in volume after T08. T08 is the most prominent feature, a conical pile covering approximately 147.900 m², with a length of around 425 m, a mean width of 395 m, and a max height of ~ 100 m. It has a relatively broad top (~ 100 m) and a dominant slope of ~ 14° in the southern part. Its composition varies in lithology, particle size distribution (micrometric to ~ 45 mm), and color (black and red due to iron oxides dominance) along with a dominant asymmetrical shape and steep slopes (~ 55%) on the northern, eastern, and western sides (Fig. 2a).

a Topo-Photogrammetric survey via drone for T01, T02 and T08. b On-site drill hole surveys (T02)

To prevent contamination, the samples were stored in sealed plastic bags and shipped to the laboratory. To ensure representativeness, the samples were divided into smaller pieces using an adjustable channel sample divider. Then, the samples were dried at 105 °C for 12 h and then crushed them with a jaw crusher. Additionally, 50 g of each sample were ground using an oscillating disc mill until the particle size was less than 100 μm, preparing the samples for further analysis.

2.3 Mineralogical evaluation

The CMWR petrographic analysis was conducted at Mohammed IV Polytechnic University’s GSMI laboratory using an Olympus BH2-UMA optical microscope equipped with a Leica EC4 camera. Thin sections prepared from borehole samples were examined under transmitted and reflected light to analyze petrographic features. X-ray Diffraction (XRD) analysis, performed with a Cu Kα (λ = 1.540593 Å) source on samples, identified major crystalline phases using an X’pert3 Powder Smart Lab diffractometer. Data collection spanned 5° to 70° 2θ range at a 2°/minute step size. Additionally, scanning electron microscope (SEM) characterization utilized a ZEISS GEMINI 1530, VEGA 3 SCAN with 1 nm resolution and 55.4 µm view field. SEM was coupled with semi-quantitative chemical analysis by energy dispersive X-ray spectroscopy (EDX), detecting below 0.1% and maintaining 2%~4% relative precision.

A refined mineralogical investigation was conducted at XPS Consulting & Testwork Services, Glencore, Canada, using Quantitative Evaluation of Minerals by Scanning Electron Microscope (QEMSCAN) with a Bruker SDD EDS on three high-carbon samples (T02-CT2, T08-C2 + C3, T08-C6 + C8). 30 mm polished sections were prepared for mineral texture, elemental distribution, and liberation analysis. Particle mineral analysis (PMA) was performed with a resolution of 3–4 µm. Major chemical compositions were analyzed using an electron probe micro-analyzer (EPMA). The analysis employed a Cameca SX-100 electron microprobe equipped with wavelength-dispersive spectrometry (WDS). Each thin section underwent analysis at 64 points with a 20 kV acceleration voltage and 20 nA beam current. Detection limits for base metals (Cu, Zn, Pb) ranged from 200 to 300 ppm in silicates and sulfides. Elements like F, Si, P, S, Cl, K, Ca, and Fe were measured using a concentrated electron beam. Quantitative mapping utilized a 1 µm step size and 0.1 s dwell time. Laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) was used to perform a highly sensitive multi-element and isotopic analysis. A stationary laser with a 30 µm spot size was used, and the results were analyzed with in-house software. LA-ICP-MS vaporizes materials with a laser beam diameter of 4–400 µm. Average Fe, Pb, or S concentrations measured by EPMA were used for internal calibration (Detailed methods are available in supplementary data Sect. 3).

The mineralogical reconciliation process was employed to estimate the distribution of various elements (such as S, C, Fe, Ti, Ca, etc.). This involved aligning the mineral composition with chemical assays (inductively coupled plasma mass spectrometry (ICP-MS), x-ray fluorescence (XRF), and induction furnace S/C) by adjusting mineral weight percentages iteratively until the chemical assay matched the calculated composition based on mineral stoichiometry from EPMA analysis. Calculated chemical concentration (C_m) was determined using Eq. (1), multiplying each element’s percentage (c_i) by its mineral abundance (x_i). Excel solver ensured the total mineral phases equaled 100%. XRF and S/C analyses validated ICP-MS results, while XRD and OM observations confirmed mineral presence. Modal mineralogy and mineralogical deportment were then quantified using Eq. (2). (Elghali et al. 2018; Frenzel et al. 2019; Pereira et al. 2021).

C_m: Chemical concentration

x_i: Mineral abundance

c_i: Elemental chemistry of the phases of interest

N: Number of phases in the sample

Md_m: Mineralogical deportments

x_i: Mineral abundance

c_i: Elemental chemistry of the phases of interest

N: Number of phases in the sample

2.4 Physical and chemical assessments

The particle size distribution was assessed using humid sieving and a laser diffraction analyzer (Mastersizer Malvern Bettersize S). The laser diffraction analyzer covered particles from 0.01 to 1000 μm, while sieving handled particles between 50 μm and 2 mm (Expert Committee USP29–NF24 2013). Three composite samples (T02-CT2, T08-C2 + C3, T08-C6 + C8) were prepared for fractions smaller than 50 µm. T02-CT2 represents the T02 CMWR pile, C2 + C3 comes from drillholes Sd-J-02 and Sd-J-03 within the T08 pile, which may be less oxidized and retain the black color of coal. T08-C6 + C8, from drillholes Sd-J-06 and Sd-J-08, shows higher oxidation, indicated by a red color due to iron oxides. Understanding mineral grain size distribution is crucial for process efficiency, cost, and effectiveness in mineral processing. The density was measured by dividing samples mass by their equivalent volume (Yoro and Godo 1989) (For details of density measurements, see supplementary data Sect. 4).

Chemical analysis covered major elements (SiO₂, Al₂O₃, TiO₂, Fe₂O₃, CaO, MgO, MnO, Na₂O, K₂O), loss on ignition (LOI), and metals (Zn, Pb, Cu, Ag), alongside total sulfur and carbon (S/C) measurements. X-ray fluorescence (XRF) was conducted with a Malvern Panalytical Epsilon 4 XRF-EDS, calibrated against the WROXI standard, featuring a detection limit of 0.1 wt%. LOI was determined by mass change post-calcination at 950 °C for 2 h. Trace elements were analyzed via inductively coupled plasma mass spectrometry (ICP-MS) at XPS Consulting & Testwork Services, Glencore, CANADA, following HCl/HNO₃/HClO₄ digestion and filtration. Total sulfur and carbon (S/C) were analyzed using an induction furnace (ELTRA CS-2000) with a detection limit of 0.09 wt% (ELTRA elemental analyzer 2019). Sulfates in solid samples were evaluated post-digestion with 40% HCl (Sobek et al. 1978) (Detailed methods are available in the Supplementary data Sect. 4).

2.5 3D Modeling and geo-metallurgical assessment

The imported database into Datamine RM software for 3D visualization includes essential topographical and chemical data in four files: survey, collars, assays, and litho. The collar table contained the borehole collar coordinates (X, Y, Z) and drilling length. The survey table included the azimuth, dip, and drilling depth, while the assay table primarily provided sample assay results, detailing the grades of total carbon, sulfur, LOI, major, and trace elements. The lithology table detailed rock types, facies, strata, minerals, and alterations. This chemical database was verified using Datamine Studio RM software, leading to the creation of a 3D chemical distribution model for carbon. Mineralized envelopes were represented as wireframes (Machuca 2010; Espinoza 2011) and filled with 6 m × 6 m × 2 m blocks, based on their storage configuration during operations. Pile layers, approximately 50 cm thick, were regularized using sub-blocks and split functions to minimize support effects in statistical estimations. A composite length of 1 m was adopted for all CMWR piles to assess spatial continuity of carbon content through variograms calculated in specific directions (Table 2). Block models were created using the “Grade” tool for basic grade interpolation and the “Estimate” tool, offering varied options like search volumes, variogram models, and estimation types (Datamine Corporate Limited 2014; Mousavi et al. 2022). Interpolation methods nearest neighbor (NN), inverse distance weighting (IDW), and ordinary kriging (OK) were employed, each evaluated for their suitability and compared for accuracy. The NN method assigned a grade to an interpolated position based on the grades of the closest surrounding points. While commonly used, the NN method has limitations related to outlier points and data density, where multiple points may be very close to the interpolated point (Cressie 1994; Caruso and Quarta 1998; Beutel et al. 2010) (for 3D modeling method details see supplementary data Sect. 5).

G (x) is the grade of the sample whose grade is to be interpolated

Grade (xi): is the grade of the nearest sample whose grade is known

IDW estimated values at unsampled locations using a weighted average of nearby sample points, with closer points having more influence. The interpolated values converged towards the nearest sample. While IDW assumed equal contribution from neighboring points with influence decreasing by distance, it does not account for spatial correlation. IDW is suitable for tabular, stratified, homogeneous, and continuous deposits (GMS 2021; Pereira et al. 2022; Qu et al. 2022).

G: interpolated value from a given bloc “b”

D: given distance (measurement operator) from the interpolation point to the point to be interpolated

Ordinary Kriging (OK) is a widely used spatial interpolation method known for its reliability and precision. It accounted for spatial correlation through a covariance function and minimized standard deviation. OK used a variogram to predict grade values at specific points by calculating a weighted average of nearby known values, independent of the actual variable value. This method is especially suitable for complex deposits with non-stationary trends like linear or quadratic patterns (Jalloh et al. 2016; Rezaei et al. 2019).

G(t) is the estimated grade at bloc t

t is the overall mean grade of the variable being interpolated

λ(i) is the weight assigned to the grade value at the neighboring sample location t(i)

t(i) is the location of the neighboring sample point i

ε(t) is the estimation error or residual at location t, which is assumed to have a mean of zero and a constant variance σ².

The performance of each of the three interpolation methods used for carbon grades estimation were compared using many parameters, the first one is the estimation variance (EV) that measures the variability of estimated values in relation to real values. It indicated the extent to which estimated values can vary around the mean and gave an idea of the precision of the estimate (Goovaerts 1997). The estimation variance (EV) can be calculated as follows:

V_estim is the estimated value for a given location

V_real represents the real value for the same location.

N is the total number of locations where we have made estimates.

Estimation variance was integrated with additional evaluation criteria, such as cross-validation, a widely used method in statistical modeling and machine learning to assess model performance and interpolation capacity. In this study, random samples at varying depths were selected, and the outcomes of different carbon grade interpolation methods were compared with actual measured values (Christensen 2015; Trippa et al. 2015; Moscovich and Rosset 2022). This approach aided in selecting the optimal model configuration and identifying issues like overfitting or underfitting (Matheron 1963; Wiley 2012) (details of the cross-validation method are described in the supplementary data Sect. 5).

3.1 Mineralogical properties

The petrographic analysis of CMWR piles composite samples under optical microscope using reflected light revealed that it consists of different forms of anthracite. Some of it were present as small inclusions dispersed within the aluminosilicate gangue (Fig. 3a). Other anthracite minerals were present as large crystals with a length ranging from 15 to 700 µm (Fig. 3b). The observations confirmed the presence of small often deformed amounts of quartz. Pyrite is present in the form of inclusions, often xenomorphic indicating high levels of deformation (Fig. 3c), with some trace of inclusions in automorphic form. Over time, due to oxidation and alteration conditions most of primary sulfides (such as pyrite) have been oxidized and transformed into hematite, goethite or other minerals forming a halo of iron oxy-hydroxides (Fig. 3d). SEM observations of sample surfaces combined with energy-dispersive X-ray spectroscopy (EDX) analysis, revealed an irregular distribution of anthracite and an abundance of gangue minerals (Figs. 3e, f). The high levels of Mg, Al and Si phases indicate a rich abundance of aluminosilicates specifically chlorite and muscovite (Fig. 3g).

a-d Mineralogical characterization under OM review, Chl: chlorite, Anth: anthracite, Py: pyrite, Hm: Hematite and Mag: magnetite. e and f SEM imagery of T08 composite sample. g results of gangue examination under EDX analyze

The XRD study was conducted to accurately identify the existing mineral phases. The mineral phase predominantly consists of typical minerals found in shale-based CMWR piles, including kaolinite (32%~45%), quartz (25%~42%), muscovite (14%~16%), and chlorite (0%~2%). Albite (0%~3%) and anatase/rutile (0%~2%) were also detected. Ankerite (2%~4%), gypsum (0%~1%), and pyrite (0%~2%) were identified as well. Additionally, small amounts of hematite, goethite, and magnetite (2%~4%) were observed, indicating the oxidation of the iron sulfide (Fig. 4).

XRD analysis revealing mineralogical phases in studied CMWR piles; (Q: Quartz; M: Muscovite; Py: Pyrite; H: Hematite; Kl: Kaolinite; R: Rutile; Ab: Albite; Chl: Chlorite; G: Gypsum)

QEMSCAN analysis revealed that the three studied samples mainly consist of anthracite, pyrite, quartz, feldspar, chlorite, micas, iron oxides, clay minerals (mainly kaolinite), and trace amounts of zircon, calcite, and gypsum. Quantitatively, the abundance of anthracite ranged from 5 wt% (T08-C6 + C8) to 15 wt% (T02-CT2). T08-C6 + C8 exhibited a higher abundance of silicate minerals, with muscovite accounting for 28 wt%, quartz for 25 wt%, feldspar for 19 wt%, and clay for 13 wt%. Iron minerals, represented by pyrite (1 wt%) and hematite (5 wt%), were also present. Rutile (1 wt%) and ankerite (1 wt%) were observed as well (Fig. 5a). QEMSCAN imagery demonstrated the presence of abundant iron oxides and partial liberation of anthracite (Fig. 5b). In the T08-C2 + C3 sample, similar mineral concentrations were observed, with 7 wt% anthracite. Silicate minerals, including quartz (27 wt%), muscovite (22 wt%), and feldspar (16 wt%), constituted the main gangue minerals, along with pyrite (0.5 wt%) and hematite (1 wt%). The clay matrix accounted for 16 wt% of the mineralogy in this sample (Fig. 5c). In contrast, the T02-CT2 sample was characterized by a higher abundance of clay fraction (27 wt%), followed by quartz (18 wt%), muscovite (13 wt%), biotite (8 wt%), and chlorite (5 wt%). The iron phase was represented by pyrite (1 wt%), which oxidized to form hematite (5 wt%). Gypsum was present in low amounts (0.2 wt%) (Fig. 5e). QEMSCAN imaging indicated a low concentration of iron oxides comparatively to T08-C6 + C8 (Figs. 5d, f). The abundance of silicate minerals in these CMWR piles could pose challenges for their valorization, as these minerals cannot be easily milled to liberate the ore. The analysis of mineral distribution in relation to oxidation levels unveils distinct patterns. Within the T08 pile, there is a notable increase in the presence of pyrite and hematite, accounting for 1.5 wt% and 5 wt% respectively. In contrast, the newer piles exhibit lower quantities of pyrite (0.3 wt%) and a reduced proportion of hematite (1.6 wt%). These findings, in conjunction with the assessment of sulfur sulfate, carry significant implications for potential AMD generation from the CMWR samples. The elevated presence of pyrite and hematite in T08 zone suggests a greater susceptibility to sulfide mineral oxidation, potentially leading to AMD formation within this pile.

Modal mineralogy for a T08-C6 + C8; c T08-C2 + C3; e T02-CT2, and QEMSCAN imagery for b T08-C6 + C8; d T08-C2 + C3; f T02-CT2

The results of the anthracite liberation degree analysis (Fig. 6b), indicated a generally moderate to low degree of liberation, with 23 wt% to 32 wt% of the anthracite liberated. Among the CMWR piles, the sample T08-C6 + C8 stood out with the highest degree of liberation, with approximately 30 wt% of the anthracite completely liberated. The sample T08-C2 + C3 exhibited less than 30 wt.% liberation and around 55 wt% mid-liberated anthracite. In contrast, the sample T02-CT2 showed poor liberation, with a degree of liberation below 30 wt% for the anthracite. The results of QEMSCAN analysis indicated that the liberation degree of anthracite was also influenced by particle size, which, in turn, could affect the effectiveness of separation methods like flotation and gravity separation (Jameson 2012; Leißner et al. 2013; Sousa et al. 2018). Specifically, finer fractions (less than 300 µm) are expected to exhibit higher liberation degrees compared to coarser fractions (greater than 300 µm). These liberated minerals are more conducive to separation through flotation processes due to their increased surface area, facilitating stronger interaction with flotation reagents and enhancing recovery efficiency during beneficiation. Conversely, grains larger than 300 µm may pose challenges in the mining process. These larger mineral fragments may require more energy for breakdown and processing, potentially leading to increased costs and reduced efficiency. Moreover, larger grains may present difficulties in separation from one another, necessitating additional processing steps in the flow sheet (as observed in the case of T08-C6 + C8 minerals).

a T08-C6 + C8, T08-C2 + C3 and T02-CT2 elemental deportment. b Anthracite liberation degree

The EPMA results were compared to the standard deviation of each element to determine the mineral composition. This analysis unveiled the sole presence of pyrite as a sulfide component, constituting of 38.12 wt% (Fe), 2.2 wt% (Si), 1.88 wt% (Al), and traces of Ti (0.07 wt%). Iron was found to be hosted in various minerals including pyrite, iron oxides, chlorite, and micas. Iron oxides primarily consisted of 38.12 wt% (Fe), accompanied by 2.2 wt% (Si), 1.88 wt% (Al), 1.16 wt% (Ca), and 0.82 wt% of Mg. Chlorite was composed of 25.59 wt% (Si), 14.4 wt% (Al), 6.9 wt% (Fe), 2.63 wt% (K), 2.43 wt% (Mg), and traces of Ti (0.3 wt%) and Na (0.7 wt%). Micas predominantly contained 22.21 wt% (Si), 12.38 wt% (Al), 1.41 wt% (Mg), and 7.38 wt% (Fe), alongside traces of Mg (0.88 wt%), Ti (0.78 wt%), Na (0.71 wt%), and Ca (0.13 wt%). Anthracite coal, conversely, exhibited 5.37 wt% (C), 0.25 wt% (Fe), 0.05 wt% (Ti), and a low sulfur content of 0.61 wt% (Table 3).

The mineral deportment of the three samples revealed the distribution of sulfur and carbon among different minerals. Sulfur was primarily associated with pyrite (90 wt%) and anthracite (10 wt%). Carbon was mainly present in anthracite (over 70 wt%), with the remaining portion found in ankerite for T08-C2 + C3 and T02-CT2. However, for T08-C6 + C8, the residual carbon was distributed among calcite (10 wt%) and gypsum (1 wt%). The primary iron carrier is hematite, constituting 80 wt% in T08-C6 + C8 (the zone most affected by oxidation) and approximately 40 wt% –50 wt% in T08-C2 + C3 and T02-CT2. The remaining iron concentration was found in pyrite (10 wt% –18 wt%), chlorite (3 wt% –20 wt%), biotite (1 wt% –10 wt%), and ankerite (3 wt%–12 wt%). Rutile accounted for the majority of titanium (over 90 wt%), while chlorite, quartz, and biotite contained smaller amounts. Calcium was associated with feldspar (27 wt% –76 wt%), calcite (specifically in T08-C6 + C8), and ankerite (30 wt% –58 wt%) (Fig. 6a).

3.2 Physical and chemical characteristics

The distribution of particle sizes in the studied CMWR piles varied irregularly from one to another. When comparing T01 and T02 with T08, it becomes apparent that the granulometric distribution of T08 was centered around 160 µm, while T01 and T02 have approximately a similar distribution, with a median particle size of approximately 190 µm. The D₃₀ values, which correspond to the particle size at which 30% of the sample passing on the cumulative particle size distribution curve, are 5 µm, 12.2 µm and 13 µm for T01, T02, and T08, respectively. Meanwhile, the D₈₀ values, which correspond to the particle size at which 80% of the sample passing on the cumulative particle size distribution curve, are approximately 163.77 µm, 222.27 µm and 301.68 µm for T01, T02, and T08, respectively (Fig. 7a). These CMWR piles are the mixture of the waste rock that have already undergone initial mechanical preparation and screening to achieve a granulometry of less than 12 mm expressing the large granulometry exposed in these CMWR. The density measures shows that all the CMWR piles have a density coefficient of 1.6 g/cm³.

a Analysis of the T01, T02, and T08 grain sizes. b Mineral’s grain size distribution

Size-by-size chemistry of the studied samples revealed that the dominant minerals consist of finely textured silicates (such as feldspars, biotite, chlorite, and quartz) and clay-like phases (kaolinite), with a grain size ranging between 6 µm and 33 µm. Anthracite, which was present in varying grain sizes between 14.08 µm and 22.94 µm, has a larger size in T08-C6 + C8. Similarly, pyrite grain size varied between 8 µm and 18 µm, with larger grain sizes found in T08-C6 + C8. The iron oxides also have a similar grain size distribution, ranging from 10 µm in T08-C2 + C3 to 33 µm in T08-C6 + C8 (Fig. 7b). The results indicated that pile T02 exhibited smaller mineral grain sizes compared to pile T08. This finer particle structure can result in a greater surface area, which could potentially amplify reactivity and AMD. Conversely, larger pyrite grains in T08 pile may impede the oxidation process compared to less oxidized zones. The results suggested that there is a probability that the fine fraction, ranging from 25 µm to 50 µm, could potentially be used as feedstock for the flotation process to extract high-quality anthracite coal. However, processing this fine fraction will potentially demand a significant amount of grinding energy compared to coarser ore.

The broad particle size distribution of our samples can impact the void space within the material, thereby influencing its air permeability. Larger particles, characterized by wider void spaces, facilitated easier air flow through the material. The low density of our materials implied a relatively higher permeability, potentially enabling easier penetration of oxygen. Moreover, coarser particles with larger void spaces facilitated faster percolation.

The XRF analysis revealed varying percentages of different elements in our CMWR samples. SiO₂ was found to be abundant, ranging from 12 wt% to 70 wt%, Al₂O₃ content varied between 5 wt% and 33 wt%. Fe₂O₃ ranged from 3.5 wt% to 11 wt%, and MgO varied from 0.2 wt% to 3 wt%. These findings indicated the presence of aluminosilicate gangue minerals in the samples, with the existence of iron oxides yet to be confirmed through XRD analysis. Minimal quantities of other oxides, such as TiO₂ and P₂O₅, were detected. The correlation between organic and mineral carbon percentage was assessed through a LOI measure. LOI values represented the amount of organic matter burned and volatilized in the form of CO₂, leaving behind mineral fractions. The results indicated a positive correlation between 0.75 and 0.86, with an approximate relationship of %LOI = 2 × C (wt%) ± 2 (Fig. 8).

Depth-Dependent variation in Carbon, Sulfur, and LOI grades. a T01; b T02 and c T08

The results of the total S/C analysis revealed a relatively higher carbon content in T01 and T02 compared to T08. The carbon grades ranged from 8 wt% to 17 wt% for T01 (Fig. 8a) and from 10 wt% to 22 wt% for T02 (Fig. 8b), while grades in the older pile T08 varied from 5 wt% to 12 wt% (Fig. 8c). The residual carbon grades support the possibility of reprocessing these CMWR piles through a re-exploitation approach. It was observed that the carbon grades inside each drillhole increased progressively with depth. For T08 we can notice obviously that within the depth, the oxidation increased as it typically involved the reaction of the material with oxygen in the air or in another oxidizing agent such as water. This reaction can result in the removal of carbon atoms from the material’s molecular structure: C + O₂ → CO₂. The sulfur content was considered low, with grades ranging from 0.01 wt% to 0.08 wt% for T01 (Fig. 8a) and from 0.01 wt% to 0.13 wt% for T02 (Fig. 8b). However, T08 exhibited comparatively higher sulfur content, with grades reaching 2.3 wt% at drillhole Sd-J-06 (Fig. 8c). The Sulfur sulfate content was generally found to be less than 1 wt% across all piles. It’s crucial to highlight that although the sulfur content in these CMWR may be low, it doesn’t guarantee immunity against AMD generation during oxidation processes. To comprehensively evaluate their potential for acidic generation, additional analyses such as acid–base accounting and net acidic generation (NAG) tests are imperative. These tests can provide insights into the AMD potential of the material and help anticipate environmental impacts. Considering the environmental concerns associated with CMWR piles, underscores the urgency for reprocessing. Reprocessing not only offers an opportunity to recover valuable resources but also mitigates potential environmental risks by reducing the likelihood of AMD formation and its detrimental effects on surrounding ecosystems.

The ICP-MS analysis revealed significant grades of REE and trace elements compared to their abundance in the Earth’s crust (Clarke) (Haynes 2016). Vanadium (V) concentrations ranged from 134 ppm to 163 ppm, Arsenic (As) grades ranged from 39.6 ppm to 61.3 ppm, and Chromium (Cr) content varied between 90 ppm and 163 ppm. Additionally, minimal grades of lead (Pb) were observed, possibly indicating contamination from the nearby Touissit zinc-lead mine. These elements, besides being valuable, also have the potential to raise environmental concerns during leaching processes. This highlighted the necessity for a comprehensive investigation into their leaching potential. The REE grades showed variability with Lanthanum (La) ranging from 43 ppm to 50 ppm, Cerium (Ce) from 84.1 ppm to 97.9 ppm, and Neodymium (Nd) from 36.6 ppm to 43.2 ppm. The highest concentrations of REEs are observed at T08-C6 + C8. These REE concentrations were comparatively higher than their Clarke, suggesting their potential as a source of valuable elements that can be utilized alongside coal (Table 4).

3.3 3D Modeling approach

The 3D topo-photogrammetric model blocks provided initial data for estimating resources and highlighting the substantial amount of stored material, totaling approximately 7 Mm³. The 3D topo-photogrammetric models (Fig. 9) accurately measured the volume of each residual CMWR piles, faithfully representing the actual state in the area. The estimated total volume of the CMWR piles is 9.5 Mm³, with T08 alone accounting for 4.6 Mm³, while T01 and T02 have volumes of 1.9 Mm³ and 1.4 Mm³, respectively (Table 5).

a, b and c High density and high precision point cloud (Lidar/las) of T01, T02 and T08 CMWR piles respectively. d T08 Created DTM using cloud points

The geostatistical interpretation of T08 through the carbon grades plot in function of depth showed that these grades reach a maximum of 11.61 wt% with a mean of 8.58 wt%, reflecting the fairness of the value distribution. The variance and standard deviation were two measure of C dispersion within T08. They were, respectively, of 1.65 wt%² and 1.28 wt.%, indicating the relatively low dispersion of carbon values in T08 (Fig. 10a). The spatial continuity analysis using the experimental variogram revealed a nugget effect of 0.44 wt%, representing the variation between measurements taken at extremely close locations, potentially caused by instrumental or parameter fluctuations. The Sill value was of 1.28 wt%, which indicated the limit where the variogram flattened off as the lag distance tended to infinity. The range was 32 m along the X-axis and 35 m along the Y-axis, signifying the distance beyond which there was no relationship between observed values (Fig. 10b).

a Scatter plots depicting carbon concentration variability in T08. b Experimental variogram of T08 mean dip direction of mineralized bodies. Histograms showing interpolated carbon grades using: c NN method d IDW method and e OK method

The histograms generated by the three interpolation methods (NN, IDW, OK) offered valuable insights. The NN method (Fig. 10c) exhibited a maximum carbon grade of 11.61 wt%, a range of 11.59 m, a mean of 8.31 wt%, a variance of 0.5 wt%², and a standard deviation of 0.71 wt%. These values indicated relatively dispersed carbon values compared to the other methods. The 25th, 50th, and 75th percentiles also reflected this dispersion, with over 75% of samples having more than 8.74 wt% of C. The skewness was 0.55, indicating a distribution that spread towards higher grades of C. The kurtosis, with a value of 13.88, signified a pointed distribution with more frequent abnormal values. The IDW method (Fig. 10d) compared to NN, showed a maximum carbon grade of 10.21 wt%, a range of 3.38 m, and a mean of 8.35 wt%. The dispersion measured indicate lower variability in carbon values compared to NN, with a variance of 0.22 wt%², a standard deviation of 0.47 wt%, and a coefficient of variation of 0.05. The 75th percentile demonstrated that over 75% of samples has more than 8.66 wt% of C. The skewness was 0.79, indicating a spread around medium grades of C between 8 and 10 wt%. The kurtosis, with a value of 0.24, suggested a leptokurtic distribution, with samples mainly distributed around the mean. Ordinary kriging (Fig. 10e) provided more precise interpolation of carbon grades, with a maximum of 8.77 wt%, a range of 1.41 m, a variance of 0.15 wt%², a standard deviation of 0.39 wt%, and a coefficient of variation of 0.04. The 75th percentile indicated that over 75% of samples has more than 8.44 wt% of C. The skewness was 0.68, highlighting a spread around the mean. The kurtosis has a negative value of − 0.87, indicating a platykurtic distribution with a relatively flat shape.

The estimation variance provided valuable insights into interpolation technique effectiveness. Notably, NN yielded a significant estimation variance of 13.53 wt%². In contrast, both IDW and OK exhibited much lower variance values, around 0.09 wt%² and 0.04 wt%² respectively. These markedly reduced variance values, nearly reaching zero for IDW and OK, indicated enhanced precision in carbon estimation. As a result, the IDW and OK methods demonstrated superior accuracy in the interpolation of carbon content within T08.

The final model block obtained by “Grade” function (Fig. 11a) illustrated the spatial distribution of coal ore within the T08, represented by sub-blocks measuring 6 m × 6 m × 2 m. Each sub-block was characterized by its coal content. Based on a density of 1.6 g/cm³, the total estimated tonnage of the T08 was 7.3 Mt. The grade-based block model demonstrated an irregular distribution of carbon throughout the T08, with moderate grades averaging around 6 wt% in the central region. This model exhibited lower carbon content interpolation due to the use of basic methods.

a 3D Model block generated by grade function in cross sections. b NN interpolated block. c IDW created block model. d OK interpolated block model. e Spatial grade information obtained by clicking on each point of the block. f Model block T08 with research ellipsoid extended to the sphere in cross section

In contrast, the “Estimate” function applied a more accurate interpolation techniques, resulting in a more precise block models of carbon distribution. The tonnage was further divided into five intervals (0–4, 4–6, 6–8, 8–10, and 10–12). The bloc model of carbon generated by the NN method showed a distribution of grades ranging from 6 wt% to 10 wt%, with an average of 8.49 wt% resulting in a total ore tonnage of 8.7 Mt, with fewer grades at the periphery of the CMWR pile (Fig. 11b). Using the IDW interpolation, the average interpolated carbon content was 8.37 wt%, corresponding to a residual total ore tonnage of 7.3 Mt. Further examination of the carbon distribution in the T08 revealed abundant zones with carbon content ranging from 6 wt% to 10 wt% (Fig. 11c). Mapping surveys conducted on the studied CMWR piles indicated well-organized, oblique to sub-horizontal artificial stratification resulting from the layering of coal-rich materials. The OK block model highlighted a carbon distribution centered around 8 wt% and 10 wt% (Table 6). The kriging results in this case showed an incomplete model due to the small sill proposed by the experimental variogram. The total ore tonnage was mainly a sterile with 44 Kt content ranging between 6 wt% and 8 wt%, and 33 Kt of metal tonnage with grades between 10 wt% and 12 wt% (Fig. 11d).

The Statistical comparison among the three interpolation methods revealed that OK provided more precise results for carbon estimation. It exhibited less dispersed values, medium asymmetry, and concentration of interpolation around the mean value. The negative kurtosis value suggested a lower presence of abnormal values. In our case study, given the relatively simple geometry of the CMWR piles, the IDW interpolation method proved to be sufficient and provided reliable results for the carbon distribution. Field observations revealed a gradation of materials from coarse to fine, with the base of the CMWR piles containing the coarsest materials and the top containing the finest.

3.3.1 Cross-Validation

The cross-validation results of the three interpolation methods indicated that the NN method has high empirical error of 7.24 wt% and a moderate correlation between the measured values and the estimated ones of 0.84 (Fig. 12). The IDW also showed a performant result with an error of 1.16 wt% and a high correlation coefficient of 0.99. In the other hand, the OK has the lowest error of 0.2 wt%, with a high correlation coefficient 0.95. fundings supported the use of IDW interpolation as it provided less error and a high positive correlation between the measured values and the predicted ones (Table 7).

Cross-validation diagram for assessing the precision of actual vs. estimated carbon values in T08 using IDW method

Applying the principles of circular economy to Jerada region could offer effective solutions for addressing CMWR disposal challenges. The objective was to shift from simply discarding CMWR to effectively using them primary or secondary materials. These CMWR comprised various components, including aggregates, shale, and residual coal, with varying particle sizes. Embracing material circularity within the mining and construction domains presented a substantial opportunity to reduce their environmental impact. The construction and building materials sectors are recognized as major consumers of raw resources like sand, gravel, and cement (Pullen et al. 2012; Laiblová et al. 2019; Taha et al. 2021). Conversely, the mining industry stood as a substantial waste producer globally (Hakkou and Benzaazoua 2015; Macías et al. 2017; U.S. Geological Survey 2023). By promoting collaborative efforts between these industries, not only can we mitigate their environmental impacts but also ease the pressure on finite natural resource reserves. Our research study was dedicated to thoroughly characterizing these CMWR and providing solution for their recycling. Our overarching goal is to establish a zero-waste concept, ensuring a sustainable and responsible approach to waste management.

Detailed examination of CMWR unveiled prominent particle size distributions at a D₈₀ that reached 300 µm. This emphasized the importance of additional processing and refining for minerals with particle sizes exceeding 300 µm. The chemical assessment highlighted a reprocessing potential of these CMWR piles with a relatively high carbon content ranging from 8 wt% to 17 wt% for T01, from 10 wt% to 22 wt% for T02 and from 5 wt% to 12 wt% for T08. Sulfur content generally remained below 1 wt%, although a deeper investigation is needed to assess potential AMD and associated environmental concerns. Clay minerals, alumino-silicates, residual anthracite, and iron oxides constituted the main mineralogical assemblage present in these CMWR. The targeted minerals (anthracite, pyrite…) showed a low liberation degree suggesting the requirement of a flotation process at sizes below 25 µm for fine anthracite recovery. Flotation residues > 25 µm could be used as a natural clays substitute for producing high-quality fired bricks (Taha et al. 2018). Generally, the Jerada CMWR piles coarser particles could undergo crushing and screening to yield various categories: coarse aggregate (16–25 mm), medium aggregate (5–16 mm), and sand (0–5 mm). Aggregates can serve for concrete production (Frías et al. 2012; Addou et al. 2017; Taha et al. 2022). The road construction sector presents a feasible avenue for valorizing these mining wastes in backfill applications (Amrani et al. 2020). The application of the 3D geo-metallurgical approach in this investigation offered valuable insights into the substantial accumulation of CMWR within these piles. An estimated volume of roughly 7 Mm³ and a total tonnage of approximately 11.2 Mt were delineated by this approach. The carbon distribution, as depicted through block models, highlighted zones abundant in carbon content, spanning from 6 wt% to 10 wt%. Notably, an oblique to sub-horizontal artificial stratification emerged from the layering of coal-rich materials. The utilization of the 3D model aided in identifying regions rich in metals, facilitating future targeting for re-exploitation purposes.

Characterizing Jerada CMWR piles involved inherent uncertainties impacting accuracy and reliability. Spatial and vertical variability of CMWR properties within piles pose challenges in obtaining representative samples. A rigorous sampling program addressed this, utilizing drone-obtained topographic data for precise site representation and aiding in volumes estimation and borehole sampling planning. Down-the-hole hammer (DTH) drilling techniques ensured fidelity and high recovery rates (> 90%), minimizing contamination. The heterogeneous nature of CMWR adds complexity, necessitating comprehensive analysis of each pile level to understand carbon distribution and mineralogical composition. Analytical uncertainties were addressed by reanalyzing 238 pulverized samples across multiple labs, revealing minimal variance in carbon content. Environmental factors like weathering and groundwater interactions introduce temporal variability in CMWR properties, affecting long-term behavior predictions. Ongoing research investigates AMD potential and leaching risks, enhancing understanding for effective CMWR pile management strategies.

In summary, the findings from the characterization of the CMWR piles strongly support their potential valorization. This can be achieved through various means, such as the re-exploitation of high-quality carbon with a calorific value exceeding 7500 kcal/kg, either for coal briquette production or direct use in the adjacent thermal power plant. Alternatively, integration into the construction sector is also a viable option. Implementing these solutions could yield employment opportunities on abandoned mine sites, mitigate the environmental impact of CMWR, free up land, and contribute to a more sustainable construction industry (Fig. 13).

An overview of the potential approaches for managing CMWR piles

The mineralogical examination revealed the presence of clay minerals, shales, and residual anthracite, with an abundance of iron oxides surrounding pyrite. Both pyrite and anthracite exhibited low release rates, suggesting the need for a froth flotation process to recover fine anthracite coal from the CMWR. Coarse particles, on the other hand, would benefit from additional crushing and sieving treatments. The physical analysis revealed predominant particle size distribution around 165 µm and 300 µm, with larger mineral sizes, particularly for anthracite. These findings highlighted the need for additional re-crushing and grain size separation processes to reduce particle size and enhance ore enrichment, preparing them for further utilization. Carbon content ranged from 6 wt% to 20 wt%, while sulfur content was generally below 1 wt%. Further assessment is needed to evaluate AMD generation potential and environmental risks. Geo-metallurgical results provided a 3D depiction of the CMWR piles, estimating a total volume of approximately 7 Mm³ and a total tonnage of about 11.2 Mt. The coal metal content was estimated at 2.3 Mt for grades between 6 wt% and 8 wt% carbon, and around 4.7 Mt for carbon grades ranging from 8 wt% to 10 wt%.

In conclusion, the Jerada CMWR piles could be re-mined for coal extraction due presence of carbon grades, reaching up to 20 wt%, particularly in the form of anthracite, which is rich in carbon and holds potential for energy recovery. It is important to note that certain CMWR piles are currently being mined, either by cooperatives or clandestinely mining. The residues resulting from the froth flotation process could be utilized as replacements for natural clays in the production of high-quality fired bricks or as additions in concrete samples, due to their ability to lower the sintering temperature in order to reduce their environmental impact.