*Article* **Recovery Potential of Rare Earth Elements (REEs) from the Gem Mining Waste of Sri Lanka: A Case Study for Mine Waste Management**

**Nimila Dushyantha 1,2,\*, I. M. Saman K. Ilankoon 3,\* , N. P. Ratnayake 1, H. M. R. Premasiri 1, P. G. R. Dharmaratne 1, A. M. K. B. Abeysinghe 1, L. P. S. Rohitha 1, Rohana Chandrajith <sup>4</sup> , A. S. Ratnayake <sup>2</sup> , D. M. D. O. K. Dissanayake <sup>1</sup> and N. M. Batapola <sup>1</sup>**


**Citation:** Dushyantha, N.; Ilankoon, I.M.S.K.; Ratnayake, N.P.; Premasiri, H.M.R.; Dharmaratne, P.G.R.; Abeysinghe, A.M.K.B.; Rohitha, L.P.S.; Chandrajith, R.; Ratnayake, A.S.; Dissanayake, D.M.D.O.K.; et al. Recovery Potential of Rare Earth Elements (REEs) from the Gem Mining Waste of Sri Lanka: A Case Study for Mine Waste Management. *Minerals* **2022**, *12*, 1411. https:// doi.org/10.3390/min12111411

Academic Editors: Eric D. van Hullebusch, Shenghua Yin and Leiming Wang

Received: 26 September 2022 Accepted: 5 November 2022 Published: 7 November 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Abstract:** Sri Lanka is one of the wealthiest countries in terms of gems. Therefore, gem mining is extensively carried out in many areas of Sri Lanka, including districts such as Ratnapura, Monaragala, Matale, and Kalutara. During the mining process, only valuable gemstones are collected, and the remaining gravel fraction with many heavy minerals is discarded. Therefore, the gem mining industry produces a large amount of waste that is mainly used only for backfilling. To sustainably manage this waste stream, gem mining waste collected from a gem pit at Wagawatta in the Kalutara district in Sri Lanka was investigated for value recovery, specifically for rare-earth elements (REEs). The gem-bearing alluvial layer contained 0.3% rare-earth oxide (REO) that could easily be upgraded up to 2.8% (LREEs = 94%) with wet sieving and subsequent density separation via a shaking table. Therefore, the concentrates of gem mine tailings with REE-bearing minerals have the potential to be a secondary source for LREEs. The organic-rich clay layer underlying the gem-bearing alluvial layer contained 0.6% REO with 49% HREEs, including Gd, Dy, Er, Yb, and especially Y enrichments. Detailed explorations are thus necessary to assess the REE potential in Sri Lanka's gem mining waste, and value recovery flowsheets should be subsequently developed to economically extract REEs. In addition, the presence of high U concentrations (800 mg/kg) in the concentrated samples could be alarming when considering the health and safety of the people engaged in gem mining. This aspect also requires detailed research studies.

**Keywords:** gem mining waste; rare-earth elements; Sri Lankan gems; tailing management; uranium

#### **1. Introduction**

Since the mining industry produces large volumes of tailings, mine tailings of past and present mining activities have become a potential secondary source for many critical raw materials, including rare earth elements (REEs). Several authors presented mine waste management and value recoveries for different mine waste materials, such as copper [1,2], vanadium [1], gold [3], and REEs [4].

REEs are a set of 17 elements comprising 15 lanthanides, Y, and Sc. They are further categorised into two subgroups, namely, light rare earth elements (LREEs) (from La to Eu) and heavy rare earth elements (HREEs) (from Gd to Lu, including Y). Despite the crustal abundance, these elements are typically not found in nature in high concentrations, and similar chemical properties complicate the separation processes [5]. Therefore, the mining and processing of REEs are complex and expensive [6]. These elements have

become indispensable ingredients in the global industrial revolution due to their unique properties such as high thermal stability, strong magnetism, and high electric conductivity. As a result, they are used in various applications in high-tech, green energy, military, and aerospace industries [5,7]. Therefore, the demand for REEs is continuously increasing, especially considering the developments of electric vehicles and wind turbines. However, the Chinese rare earth industry mainly fulfils the current global rare earth consumption, including value-added products such as rare earth magnets [5]. These factors have attracted significant attention to the exploration and extraction of REEs from new and secondary sources [8] to diversify global rare earth supply chains [5].

The enrichments of REEs were reported in uranium, iron-ore, and titanium mine tailings in Australia, Kazakhstan, and Sweden [9,10], and the authors in [11] investigated the economic viability of extracting REEs from iron-ore mine tailings in the Kiruna mine. In this context, the authors in [12,13] reported that gem-bearing sediments are enriched in rare earth-bearing minerals. However, in Sri Lanka, it has not been attempted to utilise gem mining waste to produce REE concentrates through physical separation techniques. This article thus addresses that research gap, considering the significance of the gem mining industry to the gross domestic product (GDP) in Sri Lanka. It is of paramount importance to design economic mineral processing techniques for waste materials such as gem mining waste in Sri Lanka while considering local socio-techno-economic factors.

Since 250 BCE, Sri Lanka has been famous for its wide variety of gems, gem industry, and international gem trading. Some gemstones have adorned crowns and thrones in the royal coronations and religious functions of ancient rulers [14]. The export value of precious and semiprecious stones was USD 145 million and 141 million in 2017 and 2018, respectively. This is equivalent to 1.3% and 1.2% of the total exports in 2017 and 2018, respectively [15]. This implies the requirement for the sustainable management and development of the gem mining industry in Sri Lanka, including the generated mining waste. Mitigating the damage to fragile ecosystems during gem mining is imperative, and artisanal gem mining activities in Madagascar [16] and Myanmar [17] have caused adverse environmental effects.

Secondary gem deposits are primarily found in Sri Lanka as sedimentary rocks and placers. During gem mining, gem-bearing sediments are piled up near the pit head and mechanically washed or manually panned using rattan baskets (i.e., storage vessels produced from rattan palm fibres) to separate the heavy fractions of sediments [14,18]. A fraction of the generated gem mining waste is only used for backfilling. Even though mining activities are regulated by the National Gem and Jewellery Authority, gem mining is a traditional industry in Sri Lanka [14]. Thus, a significant volume of gem mining waste has accumulated in gem pits. In addition, current gem mining activities in gem mining areas (Figure 1) produce additional volumes of gem mining waste. However, the waste volumes have not been estimated. Existing volumes typically result in environmental issues such as land degradation, river-bank and soil erosion, sedimentation, landslides, groundwater pollution and reduced irrigation efficiency. In addition, the gem pits with accumulated water become breeding grounds for mosquitoes; thus, this affects the health and safety of the people living around gem mining areas. These research gaps highlight the requirements of sustainable gem mining activities in Sri Lanka, including gem mining waste management and repurposing strategies. However, these are not explicitly addressed in the gem mining industry in Sri Lanka and other countries, such as Madagascar and Myanmar. This study was thus designed to address the identified research gaps. In this context, gem mining waste could also be a potential source for the economic extraction of REEs. Therefore, this study assesses the REE potential of gem mining waste in Sri Lanka by performing economic value concentration processes. Thus, the physical separation methods of wet sieving and density separation are investigated.

**Figure 1.** Location of the gem pit in Wagawatta on a simplified geological map of Sri Lanka (source: reproduced from [19] with permission from Precambrian Research) with the main gem mining areas [20].

#### **2. Study Area**

Geologically, the Precambrian basement rocks of Sri Lanka are divided into three main lithotectonic divisions: the Highland, Vijayan, and Wanni Complexes [19]. In addition, the island's northern and northwestern coastal belts are underlain by Miocene limestones (Figure 1). Among these complexes, gem deposits are abundantly found in the southern part of the Highland Complex (Figure 1). In addition, this area consists of REE-bearing mineral sources such as granites and pegmatites [14].

Kalu Ganga River has the second largest river basin in Sri Lanka, with a 2766 km<sup>2</sup> drainage area originating from the central hills [21,22]. Since Sri Lanka is a tropical country, a high yield of sediments is evident due to the extreme weathering conditions [23–25]. In this context, after the weathering and decomposition of the gem-bearing source rocks in the catchment area, the Kalu Ganga River transports these sediments and redeposits them as alluvial layers, especially in old river channels and flood plains where the gemstones are found [14]. A gem pit in Wagawatta in the Kalutara district was thus selected for this study. The gem layer is deposited in an old river channel with an organic-rich clay layer underneath the alluvial gem-bearing gravel layer (Figure 2).

In Sri Lankan gem mining, the alluvial layer is subjected to wet sieving or/and panning to find gemstones. Miners typically discard the remaining fractions, including gem gravel and fine tailing fractions containing heavy minerals, as waste. In this study, the authors employed such a fine tailing fraction with heavy minerals as the feed of the shaking table (Figure 3). In addition, the organic layer was mined to find the gem-bearing alluvial layers underneath. In this process, the mined organic layer is thus stored on site. Lastly, all the waste materials, including gem gravel, fine tailings with heavy minerals, and the organic

layer, are used for backfilling the gem mining pit [26]. The authors thus focused on all these waste materials in this work.

**Figure 2.** Sample locations with an illustration of the gem-bearing alluvial layer and organic-rich clay layers of the gem pit.

**Figure 3.** Physical separation flowsheet of gem-bearing alluvial samples.

#### **3. Materials and Methods**

#### *3.1. Sample Collection*

Five samples from the gem-bearing alluvial layer (AL-1 to AL-5) and five samples from the organic-rich clay layer (OCL-1 to OCL-5) were collected at different heights of the gem pit (Figure 2).

#### *3.2. Physical Separation*

Gem-bearing sediment samples in the alluvial layer (AL-1 to AL-5) were subjected to primary wet sieving using a 3 mm sieve. The oversized fraction (>3 mm) was obtained as the gem gravel samples (GG-1 to GG-5), and the undersized fraction (<3 mm) was further subjected to secondary wet sieving using a 500 μm sieve (Figure 3). After the secondary wet sieving, the oversized fraction (>500 μm) was obtained as the sand fraction. The undersized fraction (<500 μm or Samples F-1 to F-5) was further subjected to a density separation process via a shaking table (Figure 3). After the separation, the heavy fraction was considered to be the concentrate (CC-1 to CC-5). The lighter fraction was taken as the tailing (T-1 to T-5) (Figure 3). The density separation was conducted using a Wilfley shaking table employed at a deck angle of 8◦, stoke amplitude of 10 mm, and wash water flow rate of 3 L/min for 1 kg of feed (pulp density 25% *w*/*w*). Optimal deck angle and stroke amplitude values were employed on the basis of preliminary experiments.

In addition, particle size distributions (PSDs) of the feed (F-1 to F-5), concentrate (CC-1 to CC-5), and tailings (T-1 to T-5) were ascertained using a laser particle size analyser (AimSizer HMK CD-02). The PSDs are shown in Figure 4. The D50 values for the feed, concentrate, and tailings were 90.2, 35.9 and 65.5 μm, respectively. In addition, D10 values were 20.3, 10.6 and 10.3 μm for the feed, concentrate, and tailings, respectively. Furthermore, the determined D90 values of the feed, concentrate, and tailings were 324.1, 152.1 and 327.6 μm, respectively.

**Figure 4.** Particle size distributions of the feed (F-1 to F-5), concentrate (CC-1 to CC-5), and tailing (T1-T5) samples resulted from the physical separation process employed in Figure 3.

#### *3.3. Sample Analysis*

#### 3.3.1. REE and U Analysis

All the samples were oven-dried at 105 ◦C, and the dried samples were then powdered using a laboratory Tema mill. The powdered samples were sieved with a 63 μm sieve, and the representative samples were selected via coning and quartering for further analysis.

Subsamples (0.2 g) were digested with an HNO3 and HCl mixture (1 mL:3 mL) and 1 mL of H2O2 using a Mars-6 microwave digester (CEM; Mathews, NC) equipped with EasyPrep Plus high-pressure vessels. The digested samples were diluted with deionized water and then analysed with an inductively coupled plasma mass spectrometer (ICP-MS) (ICapQ-Thermo Fisher, Bremen, Germany). The instrument was calibrated according to multielemental ICP-MS standards (Sigma-Aldrich, Germany). Quality control in the analysis was maintained using certified international reference samples (San Joaquin NIST SRM 2709a from Sigma-Aldrich, Germany), replicate analysis, and blanks at every tenth sample.

#### 3.3.2. Mineralogical Analysis of Concentrated Samples

Mineralogical compositions of samples were determined with powder X-ray diffraction (XRD) using a BRUKER D8 advance eco X-ray diffractometer with normal operating conditions. Phase analysis was carried out using the International Centre for Diffraction database (ICDD). In addition, scanning electron microscopic (SEM) and energy-dispersive spectrometric (EDS) analyses were carried out using a Carl Zeiss EVO-18 instrument with an EDS element on the SEM.

#### **4. Results and Discussion**

#### *4.1. REE Concentrations of Alluvial and Gem Gravel Samples*

The average total rare earth element (TREE) concentrations of alluvial (AL) and gem gravel (GG) samples were 2188 and 819 mg/kg, respectively (Table 1). These results reveal that REEs are associated with the finer fraction of the alluvial layer (<3 mm), since the gem gravels contained a relatively low content of REEs. Both alluvial and gem gravel samples showed a higher fraction of LREEs (85% and 80%, respectively) over HREEs (12% and 16%, respectively). The rare earth oxide (REO) grades of alluvial (AL) and gem gravel (GG) samples were 0.3% and 0.1%, respectively (Table 1).

#### *4.2. REE Concentrations of Organic-Rich Clay Samples*

The average TREE concentration of the organic-rich clay layer (OCL) was 5348 mg/kg, whereas HREEs such as Gd, Dy, Er, Yb and Y showed higher average concentrations (191, 219, 167, 135 and 1766 mg/kg, respectively) compared to other HREEs. As a result, HREEs were 49%, almost the same as the LREE percentage of 48% (the remaining 3% was Sc) (Table 1). In addition, the REO percentage of this clay layer was 0.6. LREEs and HREEs were thus enriched in the organic-rich clay layer (OCL) compared to the alluvial layer (Figure 5).

**Figure 5.** Distribution of LREEs and HREEs in the vertical cross-section, including alluvial (AL) and organic-rich clay (OCL) layers.

*Minerals* **2022**, *12*, 1411


Economic concentrations of HREEs are abundantly found in ion-adsorption clay (IAC) deposits in South China and Madagascar in economic extractions. The REO percentage of IAC deposits in China varies from 0.05 to 0.2 [27], whereas the HREE concentrations of IAC deposits in Madagascar goes up to 1570 mg/kg [28]. Gd, Dy, Er, Yb, and especially Y in both deposits mainly contribute to the enriched HREE content. In this context, the organic-rich clay layer (OCL) in this gem deposit showed a similar variation, with a high content of Y, and possessed a potential source of HREEs, especially for Y. Therefore, the development of a suitable extraction method is necessary to extract HREEs from the organic-rich clay layer in this geochemical formation.

#### *4.3. REE Concentrations of Concentrated Samples*

The average REE concentration and REO grade of the concentrated samples (CC1- CC5) were 24,670 mg/kg and 2.8%, respectively (Table 2). This clearly showed that the wet sieving and subsequent density separation process via a shaking table significantly enriched the REE content in the concentrated samples, which upgraded the REO percentage by almost 10-fold compared to the original alluvial samples. The LREEs percentage in the concentrated samples accounted for about 94% of the TREE contents (Table 2). Therefore, these concentrated samples could be a secondary source for REEs, particularly for LREEs. In addition, the wet sieving and density separation process via a shaking table is a promising physical separation method to upgrade the REO grade of gem mining waste before further physical beneficiations and chemical leaching.

#### *4.4. REE Patterns of Collected Samples*

Figure 6 illustrates the chondrite-normalised REE patterns of the studied samples. Significant enrichments of LREEs in the concentrated samples (CC) and HREEs in the organic-rich clay (OCL) samples were evident in these normalised patterns. All samples showed positive Ce anomalies and significantly negative Eu anomalies. Comparing the chondrite-normalised patterns of IAC deposits in China, the organic-rich clay layer shows that HREEs such as Gd, Dy, Er and Yb were enriched in the organic-rich clay layer.

**Figure 6.** Chondrite-normalised REE patterns of the studied samples. Since Lu concentrations were below the detection limit, Lu values in alluvial and gem gravel samples are not plotted in this figure. Published REE values of IAC deposits in China and chondrite-normalised factors were adapted from [29,30], respectively.

*Minerals* **2022**, *12*, 1411


#### *4.5. REE Yield, Grade, and Recovery of Separation Experiments*

During the shaking-table tests, the average yield of the concentrate was 37%, whereas tailing accounted for 63% of the yield (Table 3). The TREO grades of feed, concentrate, and tailings were 1.4, 2.8 and 0.6, respectively. The recovery of TREO via the shaking table was 74%. The recoveries of LREO (light rare earth oxide) and HREO (heavy rare earth oxide) via the shaking table with respect to the feed were 74% (La2O3 = 76, Ce2O3 = 71, Pr2O3 = 87, Nd2O3 = 74, Sm2O3 = 73 and Eu2O3 = 75) and 90% (Gd2O3 = 91, Tb2O3 = 79, Dy2O3 = 81, Ho2O3 = 72, Er2O3 = 87, Tm2O3 = 65, Yb2O3 = 79 and Y2O3 = 95), respectively. The results reveal that the shaking-table separation could remove major gangue minerals in the feed and enrich REEs in the concentrate. Due to the high recoveries, grade, and enrichment factors of REEs, the concentrate (CC) possessed a high potential of being a secondary source for REEs, particularly for LREEs.

**Table 3.** Results of physical separation through Wilfley shaking table.


#### *4.6. REE-Bearing Minerals in the Concentrate*

REE-bearing minerals such as monazite, zircon, and xenotime were identified in the concentrate (Figure 7). In addition, the SEM/EDX spot analysis indicates the presence of LREEs, Y and phosphorus in the surface sediment samples (Figure 8). The presence of monazite and zircon justified the high concentrations of LREEs. In contrast, xenotime acted as the main HREE-bearing mineral, particularly for Y.

**Figure 7.** Concentrate's XRD patterns indicating REE-bearing and other heavy minerals.

**Figure 8.** SEM/EDX spot analysis of the concentrate.

In addition, heavy minerals such as ilmenite and garnet were enriched in the concentrate. The organic-rich clay layer showed a lower abundance of REE-bearing minerals. However, the REO was as high as 0.6% compared to the alluvial layer. Therefore, that REE minerals may have been dissolved into the groundwater in the form of REE ions due to the weathering and decomposition of the alluvial layer (Figure 2). Once groundwater penetrates through the alluvial layer, dissolved components can be migrated into the clay layer, and REE ions could be adsorbed onto the surface of the clay minerals. A similar formation could be found in IAC rare earth deposits in China [31]. Despite the low REO content, such geochemical formations render REE extraction easy and economically viable.

#### *4.7. Mine Waste, Value Recovery and Presence of U*

Considering recent trends in mine tailing management and critical metals [1,4], the investigation of REE potential in gem mining waste in Sri Lanka is crucial on two fronts, namely, value recovery from mine waste streams, and sustainable and cleaner mine waste management. In addition, significant levels of U were found in the studied samples that need to be addressed, considering the waste management aspects of gem mining in Sri Lanka. U concentrations ranged from 8.8 to 800 mg/kg in alluvial and concentrated samples, respectively (Tables 1 and 2). Since zircon, monazite and xenotime were the REEs' primary carrier minerals, U may also be enriched by these minerals in the concentrated samples. According to the soil quality guidelines (SQGs) recommended by the Canadian Council of Ministers of the Environment (CCME), the permissible U concentration for industrial land use is 300 mg/kg [32]. Therefore, the concentrated samples in this study exceeded the permissible U level.

In some gem mines in Sri Lanka, gem-bearing sediments are subjected to a density separation process using a shaking table to recover any available gold nuggets. The concentrates are thus piled up in gem mining sites without proper safe disposal procedures. Therefore, workers may be exposed to a high level of radiation, and their health could be at increased risk [33]. Detailed studies are recommended to investigate this aspect in the gem mining industry in Sri Lanka to ensure that the industry is cleaner and more sustainable.

#### **5. Conclusions**

Gem mining in Sri Lanka is intensively carried out and annually generates tonnes of mine tailings. The quantification of REE content and REO grade is essential to sustainably manage this waste stream and to assess the REE resource potential in the gem waste of Sri Lanka. In the Wagawatta gem pit, the alluvial layer contained 0.3% REO. A simple physical separation method consisting of wet sieving and subsequent density separation via a shaking table was employed to upgrade the REO up to 2.8% with a 94% LREE. Therefore, the concentrated samples are a high potential source for LREEs. The organic-rich clay layer

formed underneath the alluvial layer contained 0.6% REO and 49% HREEs. Therefore, the organic-rich clay layer could also be considered a potential HREE source.

Since this study was limited to a particular gem mine, a substantial amount of REEs could be extracted from the waste tailings in the gem mining industry. Therefore, it is necessary to assess the REE potential in gem mine waste in the entire country (this can be performed by estimating the gem mining waste generation at each mine for a given period). Subsequently, sustainable process flowsheets can be developed to extract REEs economically. In addition, it is recommended to investigate the background radiation levels in gem mining sites. Furthermore, adverse radioactive impacts must also be mitigated to maintain the sustainability of the REE extraction from the gem mining waste in Sri Lanka, fostering the United Nations Sustainable Development Goals (SDGs) such as responsible production and consumption (SDG 12), and life on land (SDG 15).

**Author Contributions:** Conceptualisation, N.D., I.M.S.K.I., N.P.R. and H.M.R.P.; data curation, N.D. and N.M.B.; formal analysis, N.D., I.M.S.K.I., N.M.B., R.C. and A.S.R.; funding acquisition, N.P.R., H.M.R.P., P.G.R.D., A.M.K.B.A., L.P.S.R. and D.M.D.O.K.D.; investigation, N.D., I.M.S.K.I. and N.M.B.; methodology, N.D. and N.M.B.; project administration, N.D., N.P.R., H.M.R.P., P.G.R.D., A.M.K.B.A., L.P.S.R., R.C., A.S.R. and D.M.D.O.K.D.; resources, N.D., N.P.R., H.M.R.P., P.G.R.D., A.M.K.B.A., L.P.S.R., R.C., A.S.R. and D.M.D.O.K.D.; supervision, N.P.R., H.M.R.P., P.G.R.D., A.M.K.B.A., L.P.S.R., R.C., A.S.R. and D.M.D.O.K.D.; validation, N.D., I.M.S.K.I., N.P.R., H.M.R.P., L.P.S.R.; writing original draft, N.D. and N.M.B.; writing—review and editing, I.M.S.K.I., N.D., N.P.R., H.M.R.P., P.G.R.D., A.M.K.B.A., L.P.S.R., R.C., A.S.R. and D.M.D.O.K.D. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors wish to acknowledge the financial and other support provided by the Accelerating Higher Education and Development (AHEAD) Operation of the Ministry of Higher Education of Sri Lanka, funded by the World Bank (AHEAD/DOR/6026-LK/8743-LK), to carry out this study. Additionally, this work was supported by the Engineering Research Node of "Advanced Waste Engineering and Management (A-WEM)" of the Faculty of Engineering, Monash University.

**Data Availability Statement:** Data available on request from I.M.S.K.I. (saman.ilankoon@monash.edu) and N.D. (nimila.dush@gmail.com).

**Conflicts of Interest:** The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

#### **References**

