**E**ff**ective Treatment of Acid Mine Drainage with Microbial Fuel Cells: An Emphasis on Typical Energy Substrates**

#### **Chenbing Ai 1,2,3, Zhang Yan 2,4,5, Shanshan Hou 2,4, Xiaoya Zheng 2,4, Zichao Zeng <sup>2</sup> , Charles Amanze 2,4, Zhimin Dai 6,7, Liyuan Chai 1,3, Guanzhou Qiu 2,4 and Weimin Zeng 2,4,\***


Received: 10 April 2020; Accepted: 10 May 2020; Published: 15 May 2020

**Abstract:** Acid mine drainage (AMD), characterized by a high concentration of heavy metals, poses a threat to the ecosystem and human health. Bioelectrochemical system (BES) is a promising technology for the simultaneous treatment of organic wastewater and recovery of metal ions from AMD. Different kinds of organic wastewater usually contain different predominant organic chemicals. However, the effect of different energy substrates on AMD treatment and microbial communities of BES remains largely unknown. Here, results showed that different energy substrates (such as glucose, acetate, ethanol, or lactate) affected the startup, maximum voltage output, power density, coulombic efficiency, and microbial communities of the microbial fuel cell (MFC). Compared with the maximum voltage output (55 mV) obtained by glucose-fed-MFC, much higher maximum voltage output (187 to 212 mV) was achieved by MFCs fed individually with other energy substrates. Acetate-fed-MFC showed the highest power density (195.07 mW/m<sup>2</sup> ), followed by lactate (98.63 mW/m<sup>2</sup> ), ethanol (52.02 mW/m<sup>2</sup> ), and glucose (3.23 mW/m<sup>2</sup> ). Microbial community analysis indicated that the microbial communities of anodic electroactive biofilms changed with different energy substrates. The *unclassified\_f\_Enterobacteriaceae* (87.48%) was predominant in glucose-fed-MFC, while *Geobacter* species only accounted for 0.63%. The genera of *Methanobrevibacter* (23.70%)*, Burkholderia-Paraburkholderia* (23.47%), and *Geobacter*(11.90%) were the major genera enriched in the ethanol-fed-MFC. *Geobacter* was most predominant in MFC enriched by lactate (45.28%) or acetate (49.72%). Results showed that the abundance of exoelectrogens *Geobacter* species correlated to electricity-generation capacities of electroactive biofilms. Electroactive biofilms enriched with acetate, lactate, or ethanol effectively recovered all Cu2<sup>+</sup> ion (349 mg/L) of simulated AMD in a cathodic chamber within 53 h by reduction as Cu<sup>0</sup> on the cathode. However, only 34.65% of the total Cu2<sup>+</sup> ion was removed in glucose-fed-MFC by precipitation with anions and cations rather than Cu<sup>0</sup> on the cathode.

**Keywords:** acid mine drainage; copper recovery; microbial fuel cell; electricity generation; microbial community

#### **1. Introduction**

Acid mine drainage (AMD) is one typical pollutant of water in many countries that have historic or current mining activities. Sulfide minerals present in mining wastes (e.g., open pits, mining waste rock, and tailings) are inevitably oxidized to form AMD when exposed to water, air, and chemolithotrophic acidophiles [1–3]. AMD is characterized by a high acidity and high concentration of toxic heavy metals/metalloids [2]. Thus, if it is not managed properly, AMD can undoubtedly cause considerable water and soil contamination, massive biodiversity loss in the aquatic ecosystem, and severe health impacts on nearby communities [4]. In order to achieve the long-term environmental sustainability regarding mining activities, effective and efficient technologies that can tackle the remediation of AMD are highly required.

Alkaline neutralizing chemicals, such as limestone and slaked lime, are conventionally adopted to treat AMD by decreasing the extreme acidity and precipitating the dissolved various poisonous metals/metalloids as hydroxides [5]. Despite effective remediation, the large volumes of sludges resulted from precipitation containing heavy metals/metalloids, which are categorized as hazardous materials and need further safe disposal. Other active and passive remediation technologies, such as bioremediation, phytoremediation, electrodialysis, wetlands, and adsorption, are also commonly used to treat AMD [4]. However, all those technologies have the drawbacks of either low remediation efficiency or high cost. Besides, some of those technologies generally produce new wastes (e.g., sludge, brines, and spent media), which require further treatment.

In fact, the high concentration of dissolved metals in AMD can be recovered by the bioelectrochemical system (BES) as valuable products to offset the cost of treatment. Therefore, the bioelectrochemical system is a promising technology for the treatment of AMD. The bioelectrochemical system is a special biological treatment process of sewage wastewater, which mainly utilizes the catalytic activity of electroactive microorganisms [6]. Under anaerobic conditions, electroactive microorganisms degrade organic pollutants and transmit electrons through external circuits to generate electricity [7]. As a new form of biomass energy utilization and pollutant removal, bioelectrochemical systems have received extensive attention due to their non-polluting characteristics [8,9]. Compared with a single strain, the electrogenic microbial consortium has many more advantages, such as higher electricity generation efficiency, a wider range of organic substrate, and higher coulombic efficiency [10,11]. Therefore, the enrichment and acclimation of electrogenic microbial consortium from environmental samples is a conventional and effective way to increase the power density of bioelectrochemical systems. Previous studies have shown that different energy substrates used to enrich electrogenic consortium can modulate the microbial community of electroactive biofilms [12,13]. However, studies focusing on the effects of typical energy substrates on the capacities of AMD treatment and microbial communities of BES are not available.

The purposes of this study were to compare the impacts of four typical energy substrates on the performance, microbial communities, and capacities of AMD treatment of enriched electroactive biofilms. Here, single-chamber microbial fuel cells were inoculated with anaerobic sludge and fed with glucose, acetate, ethanol, or lactate, respectively, as energy substrates to enrich electroactive biofilms. The performance of enriched electroactive biofilms was evaluated after the maximum voltage output was reached. The microbial communities of enriched electroactive biofilms were analyzed by high throughput sequence technology. The AMD treatment capacities of enriched electroactive biofilms were evaluated in dual-chamber microbial fuel cells. In addition, the mechanism for copper removal on the surface of the cathode was explored. These results indicated that the effects of organic chemical (that is usually contained in organic wastewater) on the enrichment of electroactive biofilm should be first evaluated in order to obtain an efficiently simultaneous treatment of organic wastewater and AMD.

#### **2. Materials and Methods**

#### *2.1. The Configuration of Microbial Fuel Cell (MFC) Reactors*

A single-chamber MFC reactor was adopted to enrich electroactive biofilms (Figure 1A). The cube-shaped single-chamber MFC reactor with a cylindrical chamber (3 cm diameter × 4 cm length) was made of Perspex. Each MFC reactor (with a working volume 28 mL) consisted of a carbon brush (1.5 cm in radius × 3 cm in length) as anode and a carbon cloth with disk shape (projected surface area of 7.07 cm<sup>2</sup> ) as a cathode. To save cost, the expensive Pt catalyst usually used to coat cathode was not adopted in this study [14]. The anode and cathode were connected by an external resistance of 1000 Ω by titanium wire. In order to remove contaminants on the surface, both carbon brush and carbon cloth were soaked overnight in acetone, followed by washing with distilled water and baked in a muffle furnace at 450 ◦C for 30 min. The dual-chamber MFC reactor was adopted to treat the simulated acid mine drainage (Figure 1B). The dual-chamber MFC reactor consisted of an anode chamber (28 mL) and a cathode chamber (15 mL). The anodes enriched in these single-chamber MFCs were then used in the double-chamber MFC. The two chambers were separated by an anion exchange membrane (Hangzhou Grion Environmental Technology, Co., Ltd, Hangzhou, China). The cathode and the electroactive anode of the dual-chamber MFC reactor were connected by an external resistance of 10 Ω. The cathode of the dual-chamber MFC reactor was made of carbon cloth with a rectangle shape (2.5 cm in length × 0.9 cm in width) immersed in simulated AMD. 1000 Ω by titanium wire. In order to remove contaminants on resistance of 10 Ω. The cathode of

**Figure 1.** The single-chamber (**A**) and double-chamber (**B**) microbial fuel cell (MFC) reactors.

#### *2.2. Startup and Operation of MFC*

μL/L) and Wolfe's vitamins (0.5 mL/L) The single-chamber MFCs were inoculated with anaerobic sludge obtained from a municipal wastewater treatment plant. Abiotic single-chamber MFCs without the inoculum of anaerobic sludge were set up. Duplicate single-chamber MFC reactors were set up for each energy substrate. The medium used to enrich electroactive biofilms contained 20 mM energy substrate (glucose, acetate, ethanol, or lactate, respectively), trace element solution (100 µL/L), and Wolfe's vitamins (0.5 mL/L) in 50 mM phosphate buffer (4.56 g/L, Na2HPO4; 2.45 g/L, NaH2PO4; 0.31 g/L, NH4Cl; 0.13 g/L, KCl; 0.02 g/L, CaCl2), as modified from previous study [15]. The trace element solution contained the following chemicals per liter: 3.00 g MgSO4, 0.25 g FeSO4·7H2O, 0.15 g ZnCl2, 0.60 g MnSO4·H2O, 0.01 g H3BO3, 0.01 g CuSO4·2H2O, 0.03 g NiCl2·6H2O, 0.03 g Na2MoO4, 0.20 g CoCl2, 0.03 g Na2WO4·2H2O, and 0.15g KAl(SO4)2·12H2O. All the chemicals used in this study were analytic pure (Sinopharm Chemical Reagent Co., Ltd, Shanghai, China). These MFCs were operated in a fed-batch mode in a temperature-controlled incubator (30 ◦C). The medium was replaced once the output voltage of MFC

declined below 20 mV. The medium used to maintain the growth of enriched bioelectroactive biofilms in the anode chamber of these dual-chamber MFCs was identical with that used for the single-chamber MFCs. The cathode chamber was fed with the simulated AMD that was diluted from the leachate of chalcopyrite bioleaching with acid water (pH 1.80) [16]. The simulated AMD mainly contained 348.87 mg/L Cu2+, 45.06 mg/L Fe3+, and 7.03 mg/L Fe2<sup>+</sup> with a pH value of 1.80. The Cu2<sup>+</sup> and Fe3<sup>+</sup> could be served as terminal electron acceptors. Abiotic double-chamber MFCs without the enriched electroactive biofilm were set up. Duplicate double-chamber MFC reactors containing the electroactive biofilms enriched with each of these different energy substrates were set up to treat the simulated AMD.

#### *2.3. Analysis and Calculations*

The voltage across the 1000 Ω external resistance of single-chamber MFCs was recorded every 50 s by the data acquisition unit (ADAM-4017 Analog Input Model, Advantech Co., Ltd, Shenzhen, China) connected to the computer. The power density and polarization curve of single-chamber MFCs were analyzed and calculated, as described in a previous study [17]. The power density was normalized to the geometrical surface area of the anode. Coulombic efficiency of single-chamber MFCs was calculated according to a previous study [18]. Electrochemical impedance spectroscopy (EIS) was applied to determine the internal resistance of these single-chamber MFCs enriched with different energy substrates using a potentiostat (Gamry reference 600+ workstation, Philadelphia, Pennsylvania, USA). The EIS measurements were conducted using a three-electrode configuration, with a saturated Ag/AgCl reference electrode and the anode serving as the working electrode and the cathode as the counter electrode. For each experimental condition, the EIS measurement was conducted in the frequency range from 1000 kHz to 0.01 Hz with an AC amplitude of 5 mV and analyzed by the software of Zview. The concentration of Fe2<sup>+</sup> and Fe3<sup>+</sup> in the cathode chamber was determined using the phenanthroline method [19]. The concentration of Cu2<sup>+</sup> was quantified with bis-cyclohexanone oxalyldihydrazone (BCO) [20]. The pH value of catholyte was measured with a pH-meter (SJ-4A, Leichi, Shanghai, China).

Scanning electron micrograph (SEM, JSM-6490LV, JEOL, Tokyo, Japan) was adopted to observe the enriched electroactive biofilms and the structure of cathode surfaces. The energy dispersive X-ray spectrometry (EDXS; Elect super, EDAX AMETEK, Kleve, Germany) equipped for SEM was used to examine the morphologies and compositions of the deposits on cathode electrodes after the treatment of AMD. The products deposited on the cathode electrode were determined by the X-ray powder diffraction (XRD) (D8 Advance, Bruker Corporation, Karlsruhe, Germany), in which data were recorded in the 2θ range of 10 to 80 degree with a step of 0.02 degree.

#### *2.4. Genomic DNA Extraction and MiSeq Sequencing of Bioelectroactive Biofilms*

The electroactive biofilms enriched with different energy substrates in MFCs with stable output voltages were sampled to extract the total genomic DNA by the DNeasy PowerSoil DNA Isolation Kit (QIAGEN, Chatsworth, CA, USA). Illumina adapter sequence, together with the universal primer pair 515FmodF (5'-GTGYCAGCMGCCGCGGTAA-3') and 806RmodR (5'-GGACTACNVGGGTWTCTAAT-3'), were used to amplify the V4 region of the bacterial and archaeal 16S rDNA genes. PCR amplification was performed on Applied Biosystems GeneAmp® 9700 thermal cycler (ABI Inc., Foster City, CA, USA). PCR system (25 µL) consisted of 1 µL of template DNA, 1 µL (10 nM) of each primer, 9.5 µL of DNase-free deionized water, and 12.5 µL of 2× Taq PCR Master Mix (TransGen, Beijing, China). Triplicate amplifications for each genomic DNA sample were amplified and blended to minimize potential biases of amplification, which were separated by agarose gel electrophoresis (2%, w/v) and recovered using AxyPrep DNA gel extraction kit (Axygen Scientific Inc., Union City, CA, USA). The concentration of the recovered PCR products was measured using QuantiFluor™-ST Fluorometer (Promega Corporation, Madison, WI, USA). Sequencing libraries were prepared and sequenced by the Illumina MiSeq platform with the sequencing strategy PE250 (Shanghai Majorbio Bio-pharm Technology Co., Ltd, Shanghai, China).

The raw data of 16S rRNA gene sequences from MiSeq sequencing was in FASTQ format. The Illumina adapter and other specific sequences were trimmed before the following process. Then, the pair-end reads with at least 10 bp overlap, and lower than 5% mismatches were merged using the Fast Length Adjustment of SHort reads (FLASH) software [21]. The sequences shorter than 240 bp, chimeric sequences, and low-quality sequences were filtered, trimmed, and removed [22]. Operational taxonomic units (OTUs) were obtained based on the threshold of 97% similarity by using UPARSE [23]. The taxonomy of OTU representative sequences was phylogenetically assigned to taxonomic classifications by the Ribosomal Database Project (RDP) classifier at the threshold of 70% for confidence based on the Bayesian algorithm [24]. Community richness, Ace and Shannon indices, and Chao1 richness estimates were obtained by MOTHUR analysis [25].

#### **3. Results and Discussion**

#### *3.1. E*ff*ect of Di*ff*erent Energy Substrates on Single Chamber MFC Performance*

Different energy substrates (i.e., glucose, acetate, ethanol, or lactate) affected the startup, maximum voltage output, power density, and coulombic efficiency of single-chamber microbial fuel cells (Figure 2). The output voltage of the MFCs enriched with lactate as an energy substrate began to be detectable only 40 h after the inoculation with anaerobic sludge (Figure 2A). However, in order to generate a detectable output voltage, 120, 210, or 220 h was required, respectively, for the MFCs enriched with ethanol, acetate, or glucose. Compared with the maximum voltage output (55 mV) obtained by glucose-fed-MFC, much higher maximum voltage output (187 to 212mV) was achieved by MFCs fed individually with the other three energy substrates. Around 400 h after the initial inoculation, the output voltage of each MFC reached the maximum. Thereafter, the output voltage could rapidly increase to the maximum value immediately after the removal of planktonic microorganisms by replenishing with growth medium containing each energy substrate (Figure 2A). This rapid recovery of maximum output voltage indicated that the current was mainly generated by the sessile microorganisms on the surface of the anode. Acetate-fed-MFC showed the highest power density (195.07 mW/m<sup>2</sup> ), followed by lactate (98.63 mW/m<sup>2</sup> ), ethanol (52.02 mW/m<sup>2</sup> ), and glucose (3.23 mW/m<sup>2</sup> ) (Figure 2B). As indicated by the polarization test, the output voltage of acetate-fed-MFC was much higher than those of other MFCs at different external resistance (Figure 2C). On the contrary, the output voltage of glucose-fed-MFC was the lowest (Figure 2C). Coulombic efficiencies of these MFCs were dependent on the energy substrates. The MFCs enriched with lactate had the highest coulombic efficiency (33.34%), followed by the MFCs enriched with ethanol (14.30%), acetate (12.53%), and glucose (1.98%). The lowest coulombic efficiency obtained by the glucose-fed-MFCs was consistent with the previous studies because the glucose is a fermentable substrate that can be utilized by diverse microorganisms besides the exoelectrogens enriched in the electroactive biofilms under the anaerobic condition [12,26].

The ohmic resistance and charge transfer resistances of these MFCs were obtained by electrochemical impedance spectroscopy (EIS) (Figure 3). As described in the previous study, the impedance at the high-frequency limit is the ohmic resistance, and the diameter of the semicircle is the charge transfer resistance [27]. The ohmic resistance of MFC containing the bioelectroactive biofilms enriched with glucose was 19.55 Ω. However, the MFCs containing the bioelectroactive biofilms enriched with acetate (2.81 Ω), lactate (3.39 Ω), and ethanol (5.31 Ω) had much lower ohmic resistance (Figure 3). The charge transfer resistances of the MFCs containing different bioelectroactive biofilms were also dependent on the energy substrate used for enrichment. The charge transfer resistances of the MFCs containing different bioelectroactive biofilms enriched with glucose, acetate, ethanol, and lactate were 33.67 Ω, 7.39 Ω, 15.00 Ω, and 15.38 Ω, respectively. Discrepancy regarding startup, maximum voltage output, power density, coulombic efficiency, and charge transfer resistances of single-chamber MFCs enriched with different energy substrate implied that the electroactive biofilms enriched on the surface of anode were different in terms of the microbial community.

**Figure 2.** The output voltage (**A**), power density (**B**), and polarization curve (**C**) of MFCs containing bioelectroactive biofilms enriched with different energy substrates.

**Figure 3.** Electrochemical impedance spectroscopy (EIS) analysis of the MFCs containing the bioelectroactive biofilms enriched with different energy substrates.

#### *3.2. Microbial Community of Anodic Bioelectroactive Biofilms*

In contrast to the abiotic control, electroactive biofilms were enriched on the surface of the anode of MFCs when they reached the maximum output voltage, as revealed by SEM analysis (Figure 4). The existence of electroactive biofilms on the surface of anode demonstrated the importance of electroactive biofilms for the generation of electricity. These electroactive biofilms consisted of microorganisms with different cell morphologies. This indicated the diversity of electroactive biofilm regarding the microbial community.

**Figure 4.** SEM of the bioelectroactive biofilms enriched with different energy substrates. (**A**: abiotic control; **B**: glucose; **C**: acetate; **D**: ethanol; **E**: lactate).

In order to investigate the microbial community of electroactive biofilms enriched by different energy substrates (i.e., glucose, acetate, ethanol, or lactate), approximately 32,430 to 97,701 high-quality sequencing reads were obtained from each sample (Table 1). A total number of 710 OTU was detected in the inoculated anaerobic sludge (Figure 5). During the enrichment of the bioelectroactive biofilms process, there was a succession of microorganisms at the OTU level. After MFCs reached the stable maximum output voltage, there were 590,482,286 and 205 OTUs in the bioelectroactive biofilms enriched by acetate, lactate, ethanol, and glucose, respectively (Figure 5). Both the microbial abundance and microbial diversity of these electroactive biofilms enriched by different energy substrates were less than that of the inoculated anaerobic sludge, as indicated by the Shannon index and Simpson index listed in Table 1. These data indicated that the microbial abundance and microbial diversity of these electroactive biofilms were dependent on the energy substrate.


The α **Table 1.** The α-diversity of enriched bioelectroactive biofilms.

The most dominant phyla were *Proteobacteria*, *Bacteroidetes,* and *Saccharibacteria* in the inoculated anaerobic sludge (Figure 6). Both *Proteobacteria* and *Bacteroidetes* remained as the dominant phyla in these enriched electroactive biofilms. The proportion of *Proteobacteria* increased significantly, while the proportion of *Bacteroidetes* decreased remarkably in these electroactive biofilms (Figure 6A). *Firmicutes* was enriched as one of the dominant phyla in these electroactive biofilms. It was worth mentioning that *Euryarchaeota* was enriched in these electroactive biofilms, especially in the electroactive biofilms fed with ethanol as an energy substrate.

**Figure 5.** The number (**A**) and Venn diagram analysis (**B**) of the operational taxonomic unit (OTU) of these bioelectroactive biofilms enriched with different energy substrates.

**Figure 6.** Microbial community of the bioelectroactive biofilms enriched with different energy substrates at phylum (**A**), class (**B**), and genus (**C**) levels.

The major classes in the electroactive biofilms were different from that of the anaerobic sludge (Figure 6B). *Gammaproteobacteria*, *Deltaproteobacteria,* and *Betaproteobacteria* were the three major classes within the inoculated anaerobic sludge and the electroactive biofilms enriched, respectively, with acetate, ethanol, or lactate. However, only *Gammaproteobacteria* constituted as the major class of the electroactive biofilms enriched with glucose (87.70%). *Sphingobacteriia* (25.48%) and *norank\_p\_Saccharibacteria* (7.66%) were two major classes that existed in the inoculated anaerobic sludge, both of which were shifted as minor constituents in these electroactive biofilms.

The major genera in anodic electroactive biofilms were modulated by energy substrates (Figure 6C). The *unclassified\_f\_Enterobacteriaceae* (87.48%) was predominant in the glucose-fed-MFC, while *Geobacter* species only accounted for 0.63%. The genera of *Methanobrevibacter* (23.70%)*, Burkholderia-Paraburkholderia* (23.47%), and *Geobacter* (11.90%) were the major genera enriched in the ethanol-fed-MFC. *Geobacter* was most predominant in the MFC enriched by lactate (45.28%) or acetate (49.72%), which corroborated with a previous study [28]. Results showed that the abundance of classic exoelectrogens *Geobacter* species correlated to the electricity-generation capacities of electroactive biofilms. It is worth mentioning that the *Euryarchaeota* was enriched in these electroactive biofilms, especially in the electroactive biofilms fed with ethanol as an energy substrate (Table 2). Recent studies have shown that quorum sensing (QS) plays an important role in shaping the dynamics of microbial community structure and enhancing the electron transfer process in the anodic electroactive biofilms of MFCs [29,30].

**Table 2.** The ratio of archaea species in the bioelectroactive biofilms enriched with different energy substrates.


The expression of functional genes in either single strain or microbial consortium has been altered by various physicochemical parameters [16,31,32]. Therefore, it is necessary to identify and compare the important genes involved in the electron transfer for electricity generation of these electrochemical biofilms in MFCs by comparative metagenomic and transcriptomic analyses in the future. The extracellular polymeric substances (EPS) are important for the functional roles of single strain and consortium [30,33,34]. The EPS of electroactive biofilm contains proteins, glycoproteins, extracellular DNA, glycolipids, and humic substances [30]. Previous studies have shown that cytochrome proteins, pili, and nanowire in EPS are directly involved in electron transfer [30,35]. Characterization of the compositions and redox properties of the EPS of these enriched electrochemical biofilms will provide novel insights into the functional role of EPS in mediating electron transfer.

#### *3.3. Contribution of Electroactive Biofilms on Anolyte's Chemical Oxygen Demand Removal and Catholyte's Copper Recovery*

Different ratio of chemical oxygen demand (COD) was depleted in the anodic chamber for the electroactive biofilms enriched by glucose (51.32%), acetate (82.00%), ethanol (72.49%), or lactate (35.95%), respectively, in 53 h after replenishing with fresh growth medium for copper recovery in dual-chamber MFCs (Figure 7A). A high concentration of COD (1909 mg/L) was removed in the anolyte of MFC fed with glucose as an energy substrate. Considering the lowest electricity production in each batch, most of the COD removed in the anolyte of glucose-fed-MFC was ascribed to the anaerobic growth by non-electrogenic microorganisms. It is worth mentioning that the number of planktonic microorganisms in MFC fed with glucose was much higher than those in the MFCs fed with other energy substrates.

**Figure 7.** The changes in chemical oxygen demand (COD) in anode chamber (**A**), and cupric ion (**B**), ferric iron (**C**), ferrous iron (**D**), and pH value (**E**) of cathode chamber of the MFCs containing electroactive biofilms enriched with different energy substrates.

The dual-chamber MFCs containing the electroactive biofilms enriched with acetate, ethanol, or lactate, respectively, could effectively recover copper from the acid mine drainage (Figure 7B). The copper in the catholyte of these MFCs decreased significantly after the initiation of the treatment of AMD. At the 39th h, no detectable copper ion was found in catholyte of MFCs containing the electroactive biofilms enriched with acetate or lactate. At the 43rd h, the copper ion in the catholyte of MFC containing the electroactive biofilms fed by ethanol was also completely recovered. However, the dual-chamber MFC containing electroactive biofilms enriched with glucose was deficient in the recovery of copper (Figure 7B). Only part of the copper ion (34.65%) was removed at the 53rd h, with a high concentration of Cu <sup>2</sup><sup>+</sup> (228.00 mg/L) remaining in the catholyte. The high concentration of Cu 2+ (310 mg/L) remained in the catholyte of abiotic control at the end of this experiment. Iron ions in the stimulated AMD were mainly Fe <sup>3</sup><sup>+</sup> (Figure 7C,D). The decrease of Fe <sup>3</sup><sup>+</sup> concentration in the catholyte of MFCs containing electroactive biofilms was partially ascribed to the bioelectrochemical reduction at the cathode to Fe <sup>2</sup><sup>+</sup> (Figure 7D). The decrease of iron ions in the catholyte of abiotic control probably resulted from the elevated pH value (Figure 7E). The pH values in catholyte of all these MFCs with electroactive biofilms were increased during the treatment of AMD. The increase in pH value was

likely ascribed to the diffusion of anions from the anolyte across the anion exchange member and reacted with the protons in the catholyte. Therefore, the decrease of iron ions in the catholyte of MFCs with electroactive biofilms was also affected by the increased pH values.

#### *3.4. Morphologies of Electrode and XRD Analysis*

The color of cathodes of dual-chamber MFCs containing the electroactive biofilms enriched with acetate, ethanol, or lactate, respectively, turned from black to brown after 53 h of treatment of AMD (Figure 8). This phenomenon indicated the bioelectrochemical reduction of copper on the surface of the cathode. However, the color of the cathode of abiotic control and MFCs containing electroactive biofilms fed with glucose remained as black (Figure 8).

**Figure 8.** The cathodes of the MFCs after the treatment of simulated AMD. (**1**: abiotic; **2**: glucose; **3**: acetate; **4**: ethanol; **5**: lactate).

– In order to better understand the copper recovery mechanism, the cathodes of dual-chamber MFCs after the treatment of AMD for 53 h were analyzed with SEM and XRD. The SEM micrographs of cathode surfaces of these MFCs containing the electroactive biofilms enriched with acetate, ethanol, or lactate were similar in terms of structure and morphology, which were different from that of the cathodes of abiotic control and the glucose-fed-MFCs (Figure 9). No deposit was observed on the cathodic surface of abiotic MFCs, which was further confirmed by the EDS analysis (Figure 9A). There were many thin segregates on the surface of cathodes of glucose-fed MFCs. Further, EDS analysis of the composition of these segregates clearly showed the characteristic peaks of Cu signals at 0.98, 8.06, and 8.87 KeVs, which confirmed the formation of Cu products (Figure 9B). Besides the Cu, many other elements (i.e., P, S, Cl, Na and Ca) were detected as compared with the surface of cathodes of abiotic control MFCs. This indicated that part of the cupric ion was precipitated with other anions and cations on the surface of the cathode, which was not observed in previous studies. The EDS analysis showed that the deposits on the cathodic surface of MFCs containing the electroactive biofilms enriched with acetate, ethanol, or lactate mainly contained the element of Cu (Figure 9C–E).

The XRD patterns of the cathodic surface of MFCs containing the electroactive biofilms enriched with acetate, ethanol, or lactate clearly demonstrated the metal copper (Cu 0 ) with characteristic peaks at 43.3, 50.4, and 74.1 degrees in 2-Theta (Figure 10). However, these characteristic peaks for metal copper (Cu 0 ) were absent for the cathode from the abiotic control MFCs and the MFCs fed with glucose. This further indicated that no copper was deposited on the cathodic surface of these MFCs. The decrease of copper in the catholyte of the abiotic control MFCs and the MFCs fed with glucose was probably ascribed to the precipitation with other anions or cations.

**Figure 9.** SEM pictures and EDS analysis of the cathodes of MFCs after the treatment of simulated AMD. (**A**: abiotic; **B**: glucose; **C**: acetate; **D**: ethanol; **E**: lactate).

**Figure 10.** XRD analysis of the cathodes of MFCs after the treatment of simulated AMD (▼Hindicates characteristic peaks of Cu 0 ).

▼

#### *3.5. Comparison of this Study with Previous Studies*

The organic wastewater (individually simulated by four typical pure chemicals) and simulated AMD were simultaneously treated in dual-chamber MFCs in this study. The effect of different energy substrates on anodic electroactive biofilms enrichment, bioelectrochemical activity, microbial communities, and AMD treatment was compared. For the scale-up of the BESs to treat the real industrial AMD in mining sites, these pure organic chemicals should be replaced by the real organic wastewater available near the pollution site in order to greatly reduce the costs. Different sources of real organic wastewater usually contain different predominant organic chemicals (such as these typical chemicals used in this study). Therefore, it is necessary to evaluate the effects of different energy substrates on anodic electroactive biofilms enrichment (both bioelectrochemical activity and microbial communities) and AMD treatment.

It is a fact that various organics (either pure chemical or real organic wastewater) have been studied as an energy substrate for MFCs [36]. However, only some studies have focused on the comparison of the electrochemical performance of MFCs enriched with different organic substrates (Table 3). On comparing our study with these studies, we found out the following differences: (1). The organic substrates used were generally different in other studies; (2). The microbial community structure of the anodic electroactive biofilms enriched with different energy substrates was studied with high throughput sequencing technique in this study. However, two other studies analyzed the anodic electroactive biofilms with traditional culture-dependent technique or denaturing gradient gel electrophoresis (DGGE) [12,37]. In addition, simultaneous treatment of different organic wastewater and simulated AMD was analyzed in this study.

MFCs have been adopted to recover some heavy metals from wastewater [38,39]. In order to be consistent with the target heavy metal in our study, the studies that focused on treatment with Cu2<sup>+</sup> were selected and compared (Table 3). On comparing our study with these studies, we found out the following differences: (1). Only a single energy substrate was used in these studies; (2). The microbial community structures of the anodic electroactive biofilm in these studies were scarcely studied. There is a study that analyzed the microbial community structure of the anodic electroactive biofilm under the stress of different concentrations of Cu2<sup>+</sup> in municipal wastewater in single-chamber MFC [14]. However, the simulated AMD was treated in the cathode chamber, while the electroactive biofilm was in the anode chamber in this study (two chambers were separated by an anion exchange membrane). Therefore, the stress of the electroactive biofilm should be negligible. Collectively, results obtained in this study are insightful for the enrichment of electroactive biofilms for AMD treatment.


**Table 3.** Comparison of this study with other related studies.

#### **4. Conclusions**

This study showed that different energy substrates affected the startup, maximum voltage output, power density, coulombic efficiency, ohmic resistance, and the charge transfer resistance of MFC. The microbial community structures of these electroactive biofilms were modulated by energy substrates during the enrichment. The abundance of classic exoelectrogens *Geobacter* species correlated with the electricity-generation capacities of different electroactive biofilms. *Geobacter* species constituted

as the predominant components of the electroactive biofilms enriched with acetate, ethanol, or lactate, which existed as minor species in glucose-fed electroactive biofilms (0.63%). The MFCs containing the glucose-fed electroactive biofilms were deficient in the extraction of copper from AMD. On the contrary, the MFCs containing the electroactive biofilms enriched with acetate, ethanol, or lactate recovered almost all the Cu2<sup>+</sup> from the AMD by electrochemical reduction as metal copper (Cu<sup>0</sup> ) on the surface of the cathode. These results indicated that the effects of organic chemical (that is usually contained in organic wastewater) on the enrichment of electroactive biofilm should be first evaluated in order to obtain an efficient simultaneous treatment of organic wastewater and AMD. Further research works are needed to assess the technical feasibility of the bioelectrochemical system to treat AMD, such as scale-up the reactor and run in continuous mode.

**Author Contributions:** Data curation, Z.Z.; Formal analysis, C.A. (Chenbing Ai); Investigation, C.A. (Chenbing Ai), Z.Y., S.H., X.Z., and C.A. (Charles Amanze); Methodology, Z.Y.; Project administration, G.Q.; Resources, Z.D.; Supervision, W.Z.; Writing—original draft, C.A. (Chenbing Ai); Writing—review and editing, L.C., G.Q., and W.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the postdoctoral research funding of Central South University (Grant No. 207154), the National Natural Science Foundation of China (Grant No. 31470230, 51320105006, 51604308), the Youth Talent Foundation of Hunan Province of China (No.2017RS3003), Natural Science Foundation of Hunan Province of China (No.2018JJ2486), Key Research and Development Projects in Hunan Province (2018WK2012).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 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 (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Performance Evaluation of Fe-Al Bimetallic Particles for the Removal of Potentially Toxic Elements from Combined Acid Mine Drainage-Effluents from Refractory Gold Ore Processing**

**Elham Aghaei <sup>1</sup> , Zexiang Wang <sup>1</sup> , Bogale Tadesse <sup>1</sup> , Carlito Baltazar Tabelin <sup>2</sup> , Zakaria Quadir <sup>3</sup> and Richard Diaz Alorro 1,\***


**Abstract:** Acid mine drainage (AMD) is a serious environmental issue associated with mining due to its acidic pH and potentially toxic elements (PTE) content. This study investigated the performance of the Fe-Al bimetallic particles for the treatment of combined AMD-gold processing effluents. Batch experiments were conducted in order to eliminate potentially toxic elements (including Hg, As, Cu, Pb, Ni, Zn, and Mn) from a simulated waste solution at various bimetal dosages (5, 10, and 20 g/L) and time intervals (0 to 90 min). The findings show that metal ions with greater electrode potentials than Fe and Al have higher affinities for electrons released from the bimetal. Therefore, a high removal (>95%) was obtained for Hg, As, Cu, and Pb using 20 g/L bimetal in 90 min. Higher uptakes of Hg, As, Cu, and Pb than Ni, Zn, and Mn also suggest that electrochemical reduction and adsorption by Fe-Al (oxy) hydroxides as the primary and secondary removal mechanisms, respectively. The total Al3+ dissolution in the experiments with a higher bimetal content (10 and 20 g/L) were insignificant, while a high release of Fe ions was recorded for various bimetal dosages. Although the secondary Fe pollution can be considered as a drawback of using the Fe-Al bimetal, this issue can be tackled by a simple neutralization and Fe precipitation process. A rapid increase in the solution pH (initial pH 2 to >5 in 90 min) was also observed, which means that bimetallic particles can act as a neutralizing agent in AMD treatment system and promote the precipitation of the dissolved metals. The presence of chloride ions in the system may cause akaganeite formation, which has shown a high removal capacity for PTE. Moreover, nitrate ions may affect the process by competing for the released electrons from the bimetal owing to their higher electrode potential than the metals. Finally, the Fe-Al bimetallic material showed promising results for AMD remediation by electrochemical reduction of PTE content, as well as acid-neutralization/metal precipitation.

**Keywords:** acid mine drainage; gold processing effluents; Fe-Al bimetallic particles; electrochemical reduction

#### **1. Introduction**

Acid mine drainage (AMD) refers to acidic runoff rich in high concentrations of metal ions, such as iron (Fe), manganese (Mn), zinc (Zn), copper (Cu), lead (Pb), nickel (Ni), arsenic (As), cadmium (Cd), aluminum (Al), and mercury (Hg) [1–3]. AMD is associated with mining and mineral processing activities and comes from the natural oxidation of sulfide-bearing minerals (such as pyrite) exposed to water, oxygen, and microbes [4,5]. AMD is considered one of the most prevalent causes of environmental pollution which stems from its high acidity (pH < 3) and toxic metal content [6]. Tailings waste from

**Citation:** Aghaei, E.; Wang, Z.; Tadesse, B.; Tabelin, C.B.; Quadir, Z.; Alorro, R.D. Performance Evaluation of Fe-Al Bimetallic Particles for the Removal of Potentially Toxic Elements from Combined Acid Mine Drainage-Effluents from Refractory Gold Ore Processing. *Minerals* **2021**, *11*, 590. https://doi.org/10.3390/ min11060590

Academic Editor: Juan Antelo

Received: 19 April 2021 Accepted: 27 May 2021 Published: 31 May 2021

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

**Copyright:** © 2021 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/).

processing of refractory gold ores is one of the major areas of concern as it contains sulfide species and is very likely to produce AMD over time, especially in dry climates and high evaporation rates [7]. Therefore, parts of tailings with sulfide minerals content exposed to air will start to oxidize during summer to form AMD.

To tackle the issue of AMD, many attempts have been made to limit the generation and release of AMD by protecting sulfide minerals from air, water, and bacteria and minimizing their interactions [5,8–13]. However, due to practical constraints involved in the prevention strategy [14], the next available option is AMD treatment by either active or passive methods [2]. The most common active methods include neutralization using caustic soda (sodium hydroxide), calcium hydroxide (Ca(OH)2) or limestone (CaCO3), as well as adsorption, ion exchange, and crystallization [6,15]. However, the interest in improving the efficiency of AMD remediation techniques motivated researchers to develop passive methods, which involves biological and chemical treatment of AMD using wetlands [16], permeable reactive barriers [17,18], compost reactors, and bioreactors, and cost-effective materials such as recycled concrete aggregates [14], sulfur-reducing bacteria (SRB) [19], and fly ash [20]. Studies with AMD have focused on neutralizing the acidity and heavy metal removal. However, in the case of combined AMD-waste effluents resulting from refractory gold processing with Cl<sup>−</sup> and NO<sup>3</sup> content, it has not been considered anywhere before.

When exploring the most appropriate treatment techniques, it is crucial to consider the use of non-toxic, cost-effective, and high-performance materials with the lowermost potential of hazardous wastes/bi-products generation. Accordingly, zero-valent iron (ZVI) has been considered a promising element for removal of heavy metals and PTE from the aquatic environment [21] and the most common reactive material used in permeable reactive barriers (PRBs) for remediating AMD and contaminated groundwater [22]. Depending on the environmental conditions (pH, redox, and oxic-anoxic conditions), type and concentration of dissolved constituents; ZVI can remove heavy metals and PTE from solutions effectively through adsorption, surface complexation, reductive precipitation, and co-precipitation [3,22]. However, one major drawback with this kind of application is the decreasing reactivity and performance of ZVI in the long-term due to iron corrosion and surface passivation by an iron oxy-hydroxide film [21,22]. Recently, iron-based bimetallic materials have been developed aimed at improving the reactivity and efficiency of ZVI in removing PTE. In this regard, due to the synergistic effect of Fe and Al, the Fe-Al bimetal has shown remarkably improved reductive ability for the contaminants [23]. The potential difference between Fe and Al (E<sup>0</sup> (Al3+/Al<sup>0</sup> ) = −1.667 V and E<sup>0</sup> (Fe2+/Fe<sup>0</sup> ) = −0.44 V) promotes better electron transfer within the bimetallic system and slows the passivation of the Fe surface, resulting in a higher reducing capacity for target contaminants [24].

A number of studies have examined the performance of Fe-Al bimetallic particles for their ability to remove heavy metals, including Cr(VI) [25], As(III) [26], U(VI) [27] from waste solutions. Their findings demonstrated the high capacity, selectivity, and rapid removal rate of target metal ions by the bimetal, predominantly through electrochemical reduction. Moreover, in a study by Han et al. (2016) [28], a higher removal efficiency for aqueous heavy metal ions (Cr(VI), Cd(II), Ni(II), Cu(II), and Zn(II)) was achieved by acid-washed ZVAl/ZVI mixture in PRBs compared to acid-washed ZVAl or ZVI alone. One significant finding to emerge from this previous study was that the Fe-Al bimetal formation during the reaction has been identified as a major contributing factor to the high removal efficiency. Despite previous studies describing Fe-Al bimetal as a potential technique in wastewater remediation, the direct application of this bimetal in AMD treatment has not been reported to date. According to standard electrode potential of Fe, Al and metals found in AMD such as Pb, Cu, Hg, and Zn, it is clear that the Fe-Al bimetallic material is an effective medium for treating AMD. Compared to common passive treatment methods, which suffer from long processing time [29], Fe-Al bimetallic particles are fast and effective for metal removal. Moreover, both Al and Fe are among the most abundant elements on the earth, and the amount of required bimetal for AMD remediation, and waste generated, is very small. Therefore, this paper evaluates the performance of the Fe-Al bimetal for acid-

neutralization and removal of potentially toxic elements from simulated AMD combined with gold processing effluents by considering the influencing parameters including bimetal dosage and reaction time.

#### **2. Materials and Methods**

#### *2.1. Preparation of Combined AMD-Waste Effluent from Refractory Gold Processing*

The combined AMD prepared in this study represents the combination of AMD and effluents resulting from the processing of refractory gold ores containing sulfide minerals in the Goldfields region of Western Australia. Parts of tailings from these processing plants, exposed to atmospheric conditions, are very liable to generate AMD over time. The gold processing tailings dam contains Cl−, and NO<sup>3</sup> <sup>−</sup> ions because of using hydrochloric acid (HCl) or nitric acid (HNO3) in the acid washing stage [30] and lead nitrate in cyanidation [31,32]. Moreover, scaling up the mining and processing operations has risen the demand for groundwater sources. The available source of process water in Australia, especially in arid regions, is hypersaline groundwater with high Cl<sup>−</sup> content [33,34]. With regard to what mentioned above, the combined AMD solution was prepared using 1000 mg/L single-element standard solutions of Mn, Pb, As, Ni, Cu, Zn, and Hg in 2% nitric acid (Sigma-Aldrich), as well as calcium chloride (CaCl2), sodium chloride (NaCl) and ferrous sulfate (FeSO4). The initial pH of the prepared solution was adjusted to 2 using sodium hydroxide (NaOH). The synthetic AMD was with initial metal concentrations shown in Table 1.

**Table 1.** Initial solute concentrations (mg/L) in the synthetic AMD-gold processing effluents.


#### *2.2. Synthesis of Fe-Al Bimetallic Particles*

All the reagents used in this study were of analytical grade. The Fe-Al bimetallic particles were synthesized using ZVAl powder (D90 = 86.5 µm) obtained from Barnes (NSW, Australia), and ferric chloride (FeCl3·6H2O, >99% purity) purchased from Chemsupply (SA, Australia). Fe-Al bimetals were prepared by optimizing the procedure used by Chen et al. (2008) [24] and Fu et al. (2015) [25], which are based on the electrochemical reduction and deposition of iron on the ZVAl surface. The first step was to remove the unreactive layer of aluminum oxide from ZVAl using acid washing, in which 20 mL of 1 M hydrochloric acid (HCl) was added to flasks containing 3 gr ZVAl in a shaking incubator and agitated for 15 min at 40 ◦C and 110 rpm. This treatment was followed by the cementation step by adding 30 mL Fe3+ solutions with a certain concentration (to give 0.5 g Fe to 1 gr Al) to the flasks and agitating for 30 min. Then, the Fe/Al particles were recovered and rinsed with deionized water, and dried in a vacuum desiccator. The residual Fe and Al concentrations in the solution was measured (data not shown) to calculate the total Fe and Al content of recovered bimetallic particles in each preparation batch as 1 g Fe/2.1 g Al (±0.05 for 3 samples) (Equations (1) and (2)).

$$\mathbf{Fe\_T = Fe\_0 - Fe\_r} \tag{1}$$

$$\mathrm{Al}\_{\mathrm{T}} = \mathrm{Al}\_{0} - \mathrm{Al}\_{\mathrm{r}} \tag{2}$$

where

Fe<sup>T</sup> and Al<sup>T</sup> are the total Fe and Al content of the bimetal, Fe<sup>0</sup> and Al<sup>0</sup> are applied Fe and Al content, and Fe<sup>r</sup> and Al<sup>r</sup> are residual Fe and Al ions concentrations.

#### *2.3. Analytical Techniques*

The concentration of dissolved ions were analyzed using inductively coupled plasma optical emission spectroscopy (ICP-OES) and mass spectroscopy (ICP-MS). X-ray powder diffraction (XRD) of bimetallic particles was performed using an Olympus diffractometer (Olympus Scientific Solutions Americas, USA) with Co-Kα radiation source in the range between 5 and 55◦ (2θ). To characterize the size distribution of ZVAl powder and bimetallic particles, Mastersizer Malvern 3000 was used (Malvern Instruments Ltd., Malvern, UK). The structure and elemental mapping of the bimetal were determined using a Tescan Clara field emission scanning electron microscope (SEM) equipped with an energy dispersive spectrometer (EDS) manufactured by the Oxford Instrument, Oxfordshire, UK).

#### *2.4. Experimental Procedure*

To investigate the combined AMD treatment using Fe-Al bimetallic particles, batch experiments were conducted in an incubator shaker at 110 rpm and 25 ◦C with varying time intervals from 10 to 90 min. In each batch, a specific amount of Fe-Al bimetals (5, 10, and 20 g/L) was added to Erlenmeyer flasks containing 25 mL of the prepared waste solution. No acid or alkali was subsequently added to control the pH during the reaction. All experiments were conducted in duplicate, and average values were presented. After a specified time, the solution content of each flask was recovered by filtration and analyzed for heavy metal concentrations. The percentage of heavy metal removal (% R) was calculated using Equation (3):

$$\% \text{R} = \frac{\text{C}\_0 - \text{C}}{\text{C}\_0} \times 100 \tag{3}$$

where C0: Initial heavy metal ion concentration, mg/L, and C: Residual heavy metal ion concentration, mg/L.

Stabilities of pollutants were modelled by the Geochemist's Workbench® [35] with the THERMODDEM database [36] based on measured solute activities in the experiments.

#### **3. Results and Discussion**

#### *3.1. Characterization of the Fe-Al Bimetallic Particles*

The particle size distribution of the ZVAl and the synthesized Fe-Al bimetallic material is illustrated in Figure 1. The graph shows that there was an increase in the particle size of the Fe-Al bimetal compared to ZVAl. Ninety percent of the ZVAl distribution has a smaller particle size of 86.5 µm (D90) while this value increased to 134 µm after acid washing and loading with Fe. Moreover, the diffraction peaks at 45.0◦ for both Al and Fe and 52.5◦ for Fe in the XRD pattern (Figure 2) confirmed the presence of both Al and Fe in the bimetal structure.

In addition, the core-shell structure of the bimetal has been detected in the SEM mapping (Figure 3). From the EDS spectra shown in Figure 3, it can be seen that Al is mostly found in the core while Fe is the dominant element on the surface of Al.

#### *3.2. pH Monitoring*

The pH plays a vital role in the AMD treatment as increasing in pH can lead to the dissolved metal and hydroxides precipitation [37]. Figure 4 shows the experimental data for the solution pH at a different time and bimetal dosage. As illustrated in the graph, a clear trend of increasing pH with time from 0 to 30 min for all bimetal dosage at initial pH 2 was observed. However, from 30 to 90 min, a slight change in the pH was recorded. In addition, for the combined AMD treated with the greater bimetal dosages, the higher pH values were obtained. The pH of the solution containing 20 g/L of bimetal reached more than 5.5 after 30 min, although it exhibited a slight decrease from 60 to 90 min. The Eh of the solutions was 0.5 V at initial pH of 2 just before adding the bimetal and decreased to minimum value of around 0.21 V in 90 min for all bimetal dosages.

**Figure 1.** Cumulative particle size distribution of ZVAl and Fe-Al bimetallic material.

**Figure 2.** XRD pattern for the synthesized Fe-Al bimetal.

In the acidic aqueous system containing Fe-Al bimetallic particles and dissolved oxygen, the oxidation of ZVAl to Al3+ (E<sup>0</sup> (Al3+/Al<sup>0</sup> ) = −1.667 V) and ZVI to Fe2+ (E<sup>0</sup> (Fe2+/Fe<sup>0</sup> ) = −0.44 V) (Equations (4) and (5)) was accompanied by oxygen reduction in the presence of protons (H<sup>+</sup> ) and the generation of hydrogen peroxide (H2O2) (E<sup>0</sup> (O2/H2O2)= +0.695 V) [38] (Equations (6) and (7)) [38,39]. Hydrogen peroxide subsequently accelerated the ZVAl corrosion to Al3+ (Equation (8)) [40,41] and triggered a Fenton reaction, where Fe2+ and H2O<sup>2</sup> reacted to form Fe3+, hydroxyl radicals (OH˙), and hydroxyl ions (OH¯) (Equation (9)) [38]. In addition, more OH¯ released into the solution, where H2O in the solution picked up electrons. The evolution of H<sup>2</sup> gas resulted from H2O/H<sup>+</sup> reduction in the solution was also evident in the experiments (Equations (10) and (11)). To sum up, increasing the solution pH is obviously related to the release of OH¯ ions into the solution via several reactions in the solution, as discussed above.

$$\rm Al^0 \to Al^{3+} + \rm 3e^- \tag{4}$$

$$\text{Fe}^{0} \rightarrow \text{Fe}^{2+} + 2\text{e}^{-} \tag{5}$$

$$\rm O\_2 + H^+ + e^- \rightarrow HO\_2^\cdot \tag{6}$$

	- 2H2O + 2e<sup>−</sup> → H<sup>2</sup> + 2OH<sup>−</sup> (10)
		- 2H<sup>+</sup> + 2e<sup>−</sup> → H<sup>2</sup> (11)

**25 µm** 

−

**Figure 3.** SEM image and EDS spectra of Fe-Al bimetallic material with a core-shell structure.

**Figure 4.** Variation in the pH of the combined AMD solution treated by the Fe-Al bimetal at a different time and bimetal dosage.

Al → Alଷା + 3eି

Fe → Feଶା + 2eି O<sup>ଶ</sup> + Hା + eି → HO<sup>ଶ</sup>

Al + 3HଶO<sup>ଶ</sup> → Alଷା + 3OH. + OHି Feଶା + HଶO<sup>ଶ</sup> → Feଷା + OH. + OHି

2Hା + 2eି → H<sup>ଶ</sup>

. → HଶO<sup>ଶ</sup> + O<sup>ଶ</sup>

HଶO + 2eି → H<sup>ଶ</sup> + 2OHି

−

.

HO<sup>ଶ</sup>

#### *3.3. Metal Removal by the Fe-Al Bimetallic Material*

The percent removal of various metals from the synthetic combined AMD treated by the Fe-Al bimetallic material for 90 min was compared and illustrated in Figure 5. What stands out in this figure is the higher uptake of Hg, As, Cu and Pb at all bimetal levels compared with Zn, Ni, and Mn. Experiments with 20 g/L of the bimetal resulted in significant removal of Hg (99.74%), As (99.80%), Cu (98.20%), and Pb (95.50%), while it dropped to 69.50% removal for Zn, 22.34% for Ni, and <5% for Mn. As previously stated, the higher standard redox potential of the aqueous contaminants than the Al and Fe is the underlying cause of a greater removal rate. Table 2 displays the standard reduction potential of various aqueous species in the experiments at 25 ◦C. The data are arranged in the increasing order of E<sup>0</sup> , which means an increase in the tendency of species to get reduced. Therefore, under competitive conditions in the process, the loss of Hg, As, Cu, and Pb was higher than Ni, Zn, and Mn, as they have a greater attraction for electrons released from the bimetal.

**Figure 5.** A comparison of the removal of various PTE from the synthetic combined AMD by Fe-Al bimetallic particles after 90 min and initial pH 2.


**Table 2.** Standard reduction potential of different species in aqueous solution at 25 ◦C [24,42].

The variation of residual metal concentrations over time at different bimetal dosages (5, 10, and 20 g/L) and initial pH of 2 are shown in Figure 6. The initial Hg concentration in the solution (Figure 6a) dropped significantly in 10 min at all bimetal dosages, although the residual Hg(II) was slightly higher in the experiments with 5g/L bimetal (2.14 mg/L) compared to 10 and 20 g/L (<0.2 mg/L). In addition, ZVAl and ZVI on the bimetal surface, Fe2+ ions (E◦ (Fe3+/Fe2+) = +0.77) have also been considered as a reducing agent for Hg(II)

Hgଶା + 2eି → Hg(୪)

elimination from the solution [43]. Moreover, the increase in the solution pH to >4.5 in 20 min at various bimetal dosages, resulted in the precipitation of Fe (oxy)hydroxides on the bimetal surface (Figure 7a), which can sequester Hg(II) from the solution [43].

A similar trend to Hg was observed for the residual As and Cu concentrations (Figure 6b,c) within 90 min of AMD treatment using the bimetal. Despite that within 20 min of the process using 5 g/L bimetal As uptake was lower than that of 10, and 20 g/L, the removal rate was almost the same from 20 to 90 min. In addition, the initial Cu concentration of 53.44 mg/L went down to 3.5, 1.4, and 1 mg/L at 90 min for 5, 10, and 20 g/L bimetal, respectively. The higher bimetal concentrations performed more effectively for the Pb(II) reduction, so that the best result was obtained by using 20 g/L of the bimetal at 60 min (96% removal) (Figure 6d).

Prior studies [26,44] have reported the possible mechanisms for As removal by the Fe-Al bimetal as follows: (1) the adsorption of part of free As(III) by Fe-Al oxides on the bimetal surface at the initial stages of the process; (2) the oxidation of the majority of As(III) to As(V) by reactive oxygen species generated in the system, and the subsequent adsorption of As(V) by the Fe-Al (oxy)hydroxides on the bimetal surface; (3) the reduction of the adsorbed As(V) to As(III), and then to As(0) by Fe and Al (either directly or through the galvanic cell effect) in the anoxic inner layer of the bimetallic particles. Therefore, the observed discrepancy in the As uptake within 20 min of the process using different bimetal dosages corresponds most directly to the solution pH and the formation of the Fe-Al oxy-hydroxides on the bimetal (Figure 7a,b). For the AMD treated with 10 and 20 g/L bimetal, the pH values reached more than 4 in 10 min, while the same pH was recorded after 20 min for the experiment with 5 g/L bimetal (Figure 4). Considering that under acidic and circumneutral pH, the solubility, mobility, and toxicity of the As(III) is higher than As(V) species [45,46], the Fe-Al bimetal seems to be an effective material for the As remediation from the contaminated water.

According to the redox potential (Table 2), Cu(II) and Pb(II) can be easily reduced to Cu<sup>0</sup> and Pb<sup>0</sup> by both Fe and Al. Moreover, the reduction of Cu(II) to Cu(I) is also thermodynamically favored, resulting in the formation of insoluble Cu2O [42]. However, Igarashi, et al. (2020) [47] have shown that Cu precipitation at pH < 6 is unfavorable and its reduction is mostly attributed to co-precipitation with Fe and Al oxy-hydroxides. The metal contaminants with more negative redox potential than Fe, including Ni, Zn, and Mn, are hard to be reduced by Fe. The reduction by ZVAl and adsorption may be the predominant removal mechanism as the bimetal surface became more negatively charged with increasing OH<sup>−</sup> concentration, which enhanced the attraction between heavy metal ions and ZVI or ZVAl [28]. The higher removal of Zn, compared to Ni (Figure 6e,f), may be attributed to the adsorption by precipitated iron oxides on the bimetallic particles, which has been considered as a major Zn(II) removal mechanism by ZVI [48]. Moreover, due to the competitive effects, the uptake of Ni, Zn, and Mn by the bimetal may be hindered by Hg, As, Cu, and Pb ions. In addition, as can be seen in Figure 7c–e, the pH increase in the system was not sufficient to drive Mn, Zn, and Ni precipitation as they exist as aqueous species under the experimental conditions. As can be seen in Figure 6e–g, concentrations of Zn, Ni, and Mn decreased over the first 10 min of the reaction before increasing again from 10 to 30 min. The re-dissolution of these metal ions can be explained by the re-oxidation of deposited metals with the accumulated Fe(III) ((Fe3+/Fe2+) = +0.77). In addition, the concentration of heavy metal ions such as Zn, Ni, and Cu do not seem to be affected by the high concentration of Cl– anion in the pH range of the experiments (pH of 2-4) as it causes an increase in the metal's solubility [49–51]. However, in the case of Hg, calomel (Hg2Cl2) precipitation may contribute to its reduction from the solution [52,53].

Regarding the presence of nitrate ions in the system, it should be noted that nitrate has a higher electron affinity than metal ions in the process meaning it is more likely to gain electrons released from the bimetal in the competitive system. In the Fe-Al bimetallic system, both Al and Fe are able to reduce nitrate ion to nitrite (NO<sup>2</sup> <sup>−</sup>) (E◦= 0.965 V), then ammonia (NH<sup>4</sup> + ) (E◦= 0.897 V) or nitrogen gas [54,55]. Further research should be

undertaken to investigate the precise effect of NO<sup>3</sup> <sup>−</sup> and Cl<sup>−</sup> ions on heavy metal removal from AMD and on the bimetal performance.

**Figure 6.** Variation of residual metal concentrations (**a**) As, (**b**) Hg, (**c**) Cu, (**d**) Pb, (**e**) Zn, (**f**) Ni, (**g**) Mn, (**h**) Fe, and (**i**) Al over time at different bimetal dosage (5, 10, and 20 g/L) and initial pH of 2.

The Fe corrosion (Equation (5)) led to a rise in its concentration in the solution within 10 min of the process (Figure 6h). However, it decreased after 10 min, which may be attributed to the precipitation of Fe ions by increasing the pH (Figure 7a). The total Fe ions concentration after 90 min of the process using 5, 10, and 20 g/L of the bimetal are 884, 845, and 764 mg/L, respectively. The pH rise driven by increasing the bimetal dosage can be the main reason for the lower dissolved Fe ions concentrations. As can be seen in Figure 7a (the dashed rectangle) in the experimental Eh range (0.5–0.21 V) and pH > 4, Fe may precipitate as Schwertmannite or Magnetite. Moreover, depending on the Fe2+ , Fe3+, and Cl<sup>−</sup> concentrations, pH, and temperature of the reaction mixture, chloride ions may incorporate into the iron (oxy)hydroxide structure to form akaganetite [56,57], which has shown desirable sorption properties for PTE, such as As and Zn [58,59]. So, further research is needed to better understand the possibility of akaganetite formation under the experimental conditions of this study.

− − − − − − − − − **Figure 7.** Eh-pH predominance diagram of (**a**) Fe3+ (activity = 10−2.65), (**b**) Al3+ (activity = 10−4.53), (**c**) Mn2+ (activity = 10−3.99), (**d**) Zn2+ (activity = 10−3.9), and (**e**) of Ni2+ (activity = 10−3.7) at 25 ◦C, 1.013 bars and activities of SO<sup>4</sup> <sup>2</sup>−, Na<sup>+</sup> and Ca2+ equal to 10−1.6, 10−0.57 and 10−2.4, respectively. Carbonate was modelled in the system by equilibrating it with the average CO<sup>2</sup> in air (Fugacity = 10−3.5). The dashed rectangle refers to experimental conditions in this study.

Considering that the maximum recommended level of Fe in drinking water by the World Health Organization (WHO) is 2 mg/L [25], the release of Fe ions from the bimetal after the reaction with AMD is significant. However, Fe is less harmful compared to toxic heavy metals content in AMD and can be removed from the solution by a secondary neutralization process and converted to usable iron oxides as a raw material in pigments, ceramics, etc. [37]. The Al3+ concentrations in the final solutions are negligible, except for the experiment with 5 g/L of the bimetal within 10 min of the process in which the total dissolved AL3+ was 45 mg/L (Figure 6i). It is attributed to the solution pH, which is less than 3, and Al3+ is the predominant species (Figure 7b). In addition, the Al3+ concentration in the experiment using 10 g/L of the bimetal and the reaction time of 90 min meet the established limit in drinking water by WHO (0.2 mg/L) [25].

#### **4. Conclusions**

In this investigation, the synthesized Fe-Al bimetallic material has demonstrated high efficiency for a rapid removal of potentially toxic elements from the combined AMD-waste solutions resulting from refractory gold production. Owing to a greater tendency for electrons released from the bimetal, higher removal rate was obtained for Hg, As, Cu, and Pb than the Ni, Zn, and Mn. Experiments with 20 g/L of the bimetal resulted in significant removal of Hg (99.74%), As (99.80%), Cu (98.20%), and Pb (95.50%) in 90 min, while it dropped to 69.50% removal for Zn, 22.34% for Ni, and <5% for Mn. Therefore, the electrochemical reduction of PTE by the bimetal seems to be the major contributing mechanism. The findings of this study also indicate that the higher bimetal dosages result in the greater heavy metal uptake from the solution. Moreover, the corrosion of Fe and Al in the bimetallic system and consequently the release of Fe(III), Al(III), and OH<sup>−</sup> ions into the solution led to the formation of Fe-Al (oxy)hydroxides which could sequester PTE, such as Hg and Zn, via adsorption. In addition, with respect to electrode potential of metal species, Fe(III) ions engaged in the re-oxidation of deposited Zn, and more significantly Mn and Ni, led to an increase in their concentrations after 10 min. The increase in the initial pH of 2 to more than 5 in 90 min using Fe-Al bimetallic particles is promising in AMD remediation as it can reduce the amount of alkaline reagents. Nearly no Al ions were detected in the solutions at higher bimetal concentrations. Although the Fe release from the bimetal was high, it can be precipitated and converted to a valuable by-product such as iron pigments. However, more research on this topic needs to be undertaken to identify the influencing parameters, characterize and analyze the surface chemistry of bimetallic particles after reaction with PTE and measure the elemental composition, and chemical and electronic state of the elements on the bimetal. Chloride ions in the studied system may affect the process by akaganeite formation and changing the stability of PTE. Moreover, the higher electrode potential of nitrate compared to Fe, Al, and other metals in the process, could mean it has a higher tendency to gain electrons and get reduced. However, further studies regarding the precise effect of Cl<sup>−</sup> and NO<sup>3</sup> <sup>−</sup> on PTE removal and bimetal performance is strongly recommended. The reversibility of the process and reusability of the bimetal also warrant additional investigation.

**Author Contributions:** Conceptualization, R.D.A., C.B.T., E.A.; methodology, E.A. and R.D.A.; validation and analysis, E.A., Z.W., R.D.A. and Z.Q.; writing—original draft preparation, E.A. and Z.W.; writing—review and editing, R.D.A., B.T., C.B.T. and E.A.; supervision, R.D.A. and B.T.; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Data Availability Statement:** Data sharing is not applicable to this article.

**Acknowledgments:** Curtin University's Strategic Scholarship is gratefully acknowledged for the PhD scholarship granted to Elham Aghaei.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

