**Forecast of AMD Quantity by a Series Tank Model in Three Stages: Case Studies in Two Closed Japanese Mines**

#### **Chiharu Tokoro 1,\* , Kenichiro Fukaki <sup>2</sup> , Masakazu Kadokura <sup>2</sup> and Shigeshi Fuchida <sup>1</sup>**


Received: 8 April 2020; Accepted: 9 May 2020; Published: 11 May 2020

**Abstract:** There are about 100 sites of acid mine drainage (AMD) from abandoned/closed mines in Japan. For their sustainable treatment, future prediction of AMD quantity is crucial. In this study, AMD quantity was predicted for two closed mines in Japan based on a series tank model in three stages. The tank model parameters were determined from the relationship between the observed AMD quantity and the inflow of rainfall and snowmelt by using the Kalman filter and particle swarm optimization methods. The Automated Meteorological Data Acquisition System (AMeDAS) data of rainfall were corrected for elevation and by the statistical daily fluctuation model. The snowmelt was estimated from the AMeDAS data of rainfall, temperature, and sunshine duration by using mass and heat balance of snow. Fitting with one year of daily data was sufficient to obtain the AMD quantity model. Future AMD quantity was predicted by the constructed model using the forecast data of rainfall and temperature proposed by the Max Planck Institute–Earth System Model (MPI–ESM), based on the Intergovernmental Panel on Climate Change (IPCC) representative concentration pathway (RCP) 2.6 and RCP8.5 scenarios. The results showed that global warming causes an increase in the quantity and fluctuation of AMD, especially for large reservoirs and residence time of AMD. There is a concern that for mines with large AMD quantities, AMD treatment will be unstable due to future global warming.

**Keywords:** acid mine drainage; life cycle simulation; global warming; earth system model; RCP2.6; RCP8.5

#### **1. Introduction**

Japan has more than 5000 abandoned/closed mines, and about 100 of their sites produce acid mine drainage (AMD) due to the presence of sulfide mineralization [1]. The general treatment for AMD is neutralization and sedimentation by addition of a neutralizer, such as lime, calcium carbonate, and sodium hydroxide [2], and solid/liquid separation [3] of the produced sludge from the neutralized effluents. In these treatments, all toxic elements are concentrated into the sludge by precipitation [4–6] and adsorption [7–12], and the sludge is controlled in a tailing pond at a mine site or final disposal site. For these last several decades, AMD has been treated properly in Japan and has not caused severe pollution. However, since our results of the statistical calculation (details are shown below) suggested that some mines have required AMD treatment for over 150 years [13,14] and other groups suggested that more than 1000 years of treatment will be necessary [15] in the current situation, more sustainable treatment to reduce both AMD generation [16,17] and treatment cost [18] is needed. To reduce the treatment cost of the addition of chemicals and of sludge generation, for example, a passive treatment

that utilizes the natural environment of mines, such as topography, plants, and microorganisms, has attracted attention as a sustainable AMD treatment based on new concepts [19–21]. Several researchers are trying to successfully reuse this sludge as an industrial material [22,23].

Unlike industrial wastewater, AMD quantity and quality differ significantly in mines due to regional, geological, mineralogical, and biological factors. Therefore, it is necessary to customize an appropriate treatment for each mine. To select an optimal treatment method from the various treatment technologies, including those based on both active and passive concepts, an accurate understanding of the current potential for AMD generation [24,25] and the future forecast of AMD quantity and quality are essential.

The objective of this study was to determine a forecast for AMD quantity. To accomplish this, we constructed a model that reproduces the current AMD quantity using previous monitoring data of AMD, and then extrapolated it to the future. To reproduce AMD quantity, there are two ways: one is a hydraulic simulation [26–31] and the other is a tank model. A hydraulic simulation provides detailed information on the origin and distribution of AMD and can be a powerful tool for discussing AMD generation countermeasures, but it requires, in addition to meteorological data, detailed geological, mineralogical, and hydraulic data, which are generally difficult to obtain, especially for abandoned and closed mines. On the other hand, a tank model is a blackbox to determine the relationship between inflow and outflow, but just inflow data of rainfall and snowmelt and outflow of AMD quantity are sufficient for the model [32,33]. In this study, the tank model was selected for AMD quantity modeling, and rainfall and snowmelt were used as inflow. The rainfall data were corrected for elevation and adjusted using the statistical daily fluctuation model to suit each AMD site. Snowmelt was also estimated from rainfall by considering mass and heat balance by using temperature and sunshine duration data. We did not select a hydraulic simulation but chose a statistical model because our target mines are closed and it was difficult to obtain detailed monitoring, geological, and hydraulic data for this study.

For the AMD quality model, we previously reported the geochemical calculation with first-order elution kinetics of sulfide minerals [13,14]. In the model, sulfides that should be the source of AMD were selected from the quality data, and their first-order elution rate and initial AMD generation potential were estimated by fitting to the time change of their elution amount obtained from the AMD quantity and quality data. The AMD quality could be estimated by the coupling of the kinetics for sulfide elution and oxidation, and the geochemical code for the chemical equilibrium calculation of precipitation and adsorption [34–37]. This means that accurate estimation of AMD quantity is crucial for the AMD quality model.

In this study, the AMD quantity model was constructed using two case studies of underground mines: a closed sulfur mine (Mine A) and a closed black-ore copper, lead, and zinc mine (Mine B). Mine A has a large quantity of AMD, averaging 18 m<sup>3</sup> min−<sup>1</sup> with a small fluctuation, which is opposite to that from Mine B (1.5 m<sup>3</sup> min−<sup>1</sup> ). From these case studies, the parameters of the model were estimated and the future AMD quantity for the next few decades was predicted using forecast data for rainfall and temperature based on the MPI–SEM (Max Planck Institute–Earth System Model) [37]. For this, we selected two kinds of global warming scenarios proposed by IPCC (Intergovernmental Panel on Climate Change): a low-stabilization scenario of RCP (representative concentration pathway) 2.6 and a high-level greenhouse gas emission scenario of RCP 8.5 [37]. We further discuss the effects of global warming on the forecast of AMD quantity stemming from the closed sulfide mines that were examined.

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

#### *2.1. AMD Quantity Model*

#### 2.1.1. Tank Model

The AMD quantity model was constructed by using a series tank model in three stages, as shown in Figure 1. The first and second stages correspond to the nonpolluted water reservoir on the surface and inside the mine, respectively. The third stage corresponds to the polluted water reservoir in the ore deposit that causes AMD. Inflow, *r* (mm), is the summation of rainfall, *r*w, and snowmelt, *r*s. In each tank, a part of the inflow is distributed to the outflow (mm h−<sup>1</sup> ), *q*o*<sup>i</sup>* (*i* = 1, 2, 3), and seepage flow to the next tank, *q*s*<sup>i</sup>* (*i* = 1, 2), according to the water reservoir height (mm), *x<sup>i</sup>* (*i* = 1, 2, 3) and outflow height (mm), *b<sup>i</sup>* (*i* = 1, 2). The water balance in each tank is as follows: −

$$\text{ctx/dt} = q\_{\text{si-1}} - q\_{\text{si}} - q\_{\text{oi}\text{i}} \text{ ( $i = 1$  2,  $3$ )}\tag{1}$$

where *t* is time, *q*s0 = *r* = *r*<sup>w</sup> + *r*s, and *q*s3 = 0. The outflow is calculated from:

$$a\_{\rm oi} = a\_{\rm oi} \ (\mathbf{x}\_i - b\_i) \ (i = 1, 2, 3) \tag{2}$$

where *a*o*<sup>i</sup>* (*i* = 1, 2, 3) is the outflow coefficient and *b*<sup>3</sup> = 0. The seepage flow is also calculated from:

$$a\_{l\text{si}} = a\_{\text{si}} \propto\_{l\text{\prime}} (i = 1, 2) \tag{3}$$

where *a*s*<sup>i</sup>* (*i* = 1, 2) is the seepage coefficient.

**Figure 1.** Schematic of the tank model in three stages used for the acid mine drainage (AMD) quantity model in this study.

In this study, *q*o3 corresponded to AMD quantity. The inflow, *r*, was set using the following procedure. The Kalman filter and particle swarm optimization methods were used for fitting *q*o3 to the observed data of AMD quantity to estimate the *x<sup>i</sup>* , *b<sup>i</sup>* , *a*s*<sup>i</sup>* , and *a*o*<sup>i</sup>* parameters [38].

#### 2.1.2. Correction of Rainfall Data and Judgment of Snowfall

° Rainfall data near each mine were derived from AMeDAS (Automated Meteorological Data Acquisition System) provided by the Japan Meteorological Agency [39]. To suit each mine situation,

<sup>−</sup> °

the daily data of rainfall and temperature obtained from AMeDAS were corrected for elevation and adjusted by using the statistical daily fluctuation model, as shown in Figure 2. Also, snowfall was estimated according to the corrected temperature, and if the temperature was under 2 ◦C, it was judged to be snowfall and not rainfall. ° ° °

− −

−

<sup>−</sup> °

**Figure 2.** Procedural flow to obtain the inflow for the tank model (rainfall and snowmelt) from the Automated Meteorological Data Acquisition System (AMeDAS) data of rainfall, temperature, and sunshine duration.

Since the capture rate of rainfall particles by the rain gauge decreases as the wind speed increases, known as the Jebons effect, Kondo et al. proposed the following statistical correction for daily rainfall data obtained from AMeDAS [40]:

$$r\_{\rm w1} = (1.25 + 0.15 \cos(w(d - 20))) r\_{\rm w0} \tag{4}$$

$$w = 2\pi \text{\textquotedbl{}365} \tag{5}$$

<sup>−</sup> × <sup>−</sup> <sup>−</sup> where *r*w0 (mm) is the raw rainfall data, *r*w1 (mm) is the corrected rainfall data by the model, *d* (days) refers to the days from 1 January. Furthermore, the amount of rainfall near mines depends greatly on elevation due to rapid updraft and adiabatic expansion, and is corrected as follows [40]:

$$r\_{\mathbf{w}2} = (1 + c(h - h\_0))r\_{\mathbf{w}1} \tag{6}$$

where *r*w2 (mm) is the corrected rainfall data by elevation, *h* (km) is the elevation of mine site, *h*<sup>0</sup> (km) is the elevation of the AMeDAS observation point, and *c* is the coefficient (0.001 km−<sup>1</sup> for 5 ◦C or less and 0.00064 km−<sup>1</sup> for 5 ◦C or more). AMeDAS data of temperature, *T*, were also corrected according to following equation:

$$T\_1 = -6(h - h\_0) + T \tag{7}$$

where *T*<sup>1</sup> ( ◦C) is the corrected temperature.

As shown in Figure 2, if the temperature at the mine site was above 2 ◦C, it was assumed that there was no snowfall and the inflow of rainfall was set as *r*w2. On the other hand, if the temperature was below 2 ◦C, the rainfall data, *r*w2, were judged to be equivalent to snowfall, *r*<sup>f</sup> = *r*w2, and *r*w2 = 0.

#### 2.1.3. Estimation of Snowmelt and Snow Cover

Daily snowmelt, *r*s, and snow cover, *r*c, were also estimated by following the mass and heat balance of snow using the AMeDAS data of rainfall, temperature, and sunshine duration, as shown in Figure 2. If the temperature was below 0 ◦C, no snowmelt was assumed and *r*<sup>s</sup> = 0; otherwise, snowmelt was calculated according to the following procedure.

Snowmelt, *r*<sup>s</sup> (mm), was calculated from the ratio of fusion heat, *Q* (J), and latent heat, *L* (334 J kg−<sup>1</sup> ), the catchment area, *S* (mm<sup>2</sup> ), and the density of water, ρ (9.97 × 10−<sup>7</sup> kg mm−<sup>3</sup> ):

$$
\sigma\_s \mathcal{S} \rho = \mathcal{Q}/\mathcal{L} \tag{8}
$$

The fusion heat was calculated from the heat balance of snow exposure as follows:

$$Q = Q\_1 + Q\_2 + Q\_3 + Q\_4 + Q\_5 \tag{9}$$

where *Q*<sup>1</sup> is the short-wavelength radiation, *Q*<sup>2</sup> is the long-wavelength radiation, *Q*<sup>3</sup> is the sensible heat transfer, *Q*<sup>4</sup> is the latent heat transfer, and *Q*<sup>5</sup> is the transfer heat due to rainfall. Here, heat changes in the snow layer and heat transfer from the ground were assumed to be negligible [41].

The short-wavelength radiation was calculated from albedo, *r*, which is the ratio of reflected sunshine radiation to sunshine radiation on the earth's surface and the daily average of solar irradiance, *I* (W m−<sup>2</sup> ):

$$Q\_1 = (1 - r)I \tag{10}$$

The average of solar irradiance was the function of the ratio of sunshine duration, *N*, and astronomical sunshine duration, *N*<sup>0</sup> [42]:

$$\text{I/I}\_0 = 0.179 + 0.550 \text{N/N}\_0 \text{ for } 0 \le N \le N\_0 \tag{11}$$

$$\text{I/I}\_0 = 0.114 \text{, for N} = 0 \tag{12}$$

where *I*<sup>0</sup> is the solar irradiance at the top of atmosphere. The values of solar irradiance at the top of the atmosphere, *I*0, astronomical sunshine duration, *N*0, and albedo, *r*, are available from references [39,43], and sunshine duration data, *N*, are available from AMeDAS.

The long-wavelength radiation was the difference between the radiation from the atmosphere, *Q*a, and the radiation from the snow surface, *Q*s:

$$Q\_2 = Q\_\text{a} - Q\_\text{s} \tag{13}$$

$$Q\_{\rm a} = \sigma (T\_1 + 273.15)^4 (0.605 + 0.048e^{0.5}) \tag{14}$$

$$Q\_s = 0.9\sigma (T\_s + 273.15)^4 \tag{15}$$

$$e = 6.1078 \times 10^{7.57/(T + 273.3)}\tag{16}$$

where σ is the Stefan–Boltzmann constant (5.67 × 10−<sup>8</sup> W m−<sup>2</sup> K −4 ) and *e* is amount of saturated water vapor [44]. The temperature of the snow surface, *T*s, was calculated from [45]:

$$T\_s = 1.13T\_1 - 1.67\tag{17}$$

when *T*<sup>1</sup> ≤ 1.47.

The sensible heat was calculated from [46]:

$$Q\_3 = K(1 - 0.0065h/(T\_1 + 273.15 + 0.0065h))^{5.257}, \text{for } T\_1 \ge 0,\tag{18}$$

$$Q\_3 = 0, \text{for } T\_1 \le 0 \tag{19}$$

where *K* is the transfer coefficient of the sensible heat and the latent heat; 3.5 was proposed for the area near the mines that were modeled in this case study [46]. The latent heat was calculated from:

$$Q\_4 = 1.53 \text{K} (e-6.11), \text{ for } T\_1 \ge 7 \tag{20}$$

$$Q\_4 = 0, \text{for } T\_1 \le 7 \tag{21}$$

The transfer heat due to rainfall was calculated from:

$$Q\_{\overline{5}} = \rho c\_w (273.15 + T\_1) r\_\text{W} \,\text{S} \tag{22}$$

where *c<sup>w</sup>* is the specific heat of the water (4.186 J kg−<sup>1</sup> K −1 ).

The snowmelt, *r<sup>s</sup>* , calculated from Equations (8) to (22), should be less than the snow cover, *r*c. Snow cover was calculated from following summation of daily mass balance:

$$r\_{\rm c} = \sum r\_{\rm f} - \sum r\_{\rm s} \tag{23}$$

if *r*<sup>s</sup> > *r*c, then snowmelt should be *r*<sup>s</sup> = *r*c.

#### *2.2. Forecast Data of Temperature, Rainfall, and Sunshine Duration*

In the above-mentioned AMD quantity and quality models, daily data of rainfall, average temperature, and sunshine duration obtained from AMeDAS were used for model construction. Therefore, future forecasts of daily data of rainfall, average temperature, and sunshine duration were also necessary for the forecast of AMD quantity and quality in the future. For the AMD quality model, the geochemical calculation with first-order elution kinetics of sulfide minerals was used as mentioned above; the first-order elution rate and initial AMD generation potential were estimated by fitting to the time change of their elution amount obtained from the AMD quantity and quality data. The daily output data of MPI–ESM (Max Planck Institute–Earth System Model), which is an earth system model proposed by Max Planck Institute, were used in this study. Two kinds of IPCC RCPs for the greenhouse gas (GHG) concentration scenario were selected: RCP2.6 and RCP8.5. The former is the scenario with the lowest GHG emission to keep future temperature rise below 2 ◦C, and the latter is the scenario with the highest GHG emissions.

From the system, daily forecast data of rainfall, average temperature, and maximum and minimum temperatures were available. The daily forecast of the average of solar irradiance, *I*, was estimated from [47]:

$$I = 0.76 I\_0 (1 - \exp(-A \Delta T^{2.2})) \tag{24}$$

$$A = 0.036 \exp(-0.154 \Delta T\_{\text{ave}}) \tag{25}$$

where ∆*T* is the difference between the maximum temperature and the minimum temperature, and ∆*T*ave is the monthly average of ∆*T*.

#### *2.3. Case Studies in Two Closed Mines*

In this study, two closed underground mines in the northern part of Japan were selected as a case study; Mine A has a large quantity of AMD, with small fluctuation of quantity and quality, and Mine B has a small quantity of AMD, with large fluctuation of quantity and quality. Snow is observed in winter at both of the closed mines. The locations of the mines are shown in Figure 3, and the AMD characteristics are shown in Table 1.

**Figure 3.** Location of the two closed mines for case study.


**Table 1.** The AMD characteristics of Mine A and B.

N.D.: Not detected; Q: Quantity.

In Mine A, native sulfur and pyrite were mined during its operation. The size of the ore deposit was about 1500 m East–West, about 1500 m North–South, and 25–150 m of thickness, and ore reserves were about 230 million tons per year. The mine produced about 1 million tons of ore and one-third of Japan's sulfur demand, but it closed in 1971 due to the market influence of sulfur recovered from oil refining. The quantity of AMD is about 18 m<sup>3</sup> min−<sup>1</sup> on average annually, which is one of the largest AMD values in Japan [48]. The AMeDAS point is located 11 km east and 825 m below the mine.

In Mine B, copper, lead, and zinc were mined during its operation. The ore deposit was a black ore type, which has changed from lower to yellow ore, black ore, and quartz band. The mine produced a maximum of about 25,000 tons per year, but closed in 1985 due to ore depletion. The annual average of the AMD quantity was 1.72 m<sup>3</sup> min−<sup>1</sup> in 2017 and increased from 5 to 7 m<sup>3</sup> min−<sup>1</sup> during the snowmelt season [49]. The AMeDAS point is located 18 km northwest and 465 m below the mine.

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

#### *3.1. AMD Quantity Model Construction*

The relation between the input data of AMeDAS rainfall (upper side) and the observed data and the calculated value of AMD quantity (underside) are shown in Figure 4. In this calculation, the daily observed data of AMD quantity in the prior one year were used for fitting, and later one-year data were used for the model validation. The fitting period was also changed from half a year to two years and

the correlation coefficients between the observed and calculated values were compared, as shown in Table 2 and Supplementary Figure S1. Of course, the longer the fitting period, the higher the correlation coefficient in the validation period, but a fitting period of one year seemed to be generally sufficient for the reproduction of AMD quantity in the next one year. As shown in Supplementary Figure S2, when the correction for elevation and the statistical daily fluctuation and the snowmelt estimation were not conducted, the reproducibility of AMD quantity became worse, especially for Mine A. This is because Mine A is located at a higher elevation and the effects of elevation correction and snowfall are larger than for Mine B.

**Figure 4.** Observed and calculated AMD quantity by the tank model for (**a**) Mine A and (**b**) Mine B.


**Table 2.** Relationship between the fitting period and the correlation coefficients.

<sup>1</sup> 2015/10/1–2016/3/31; <sup>2</sup> 2015/4/1–2016/3/31; <sup>3</sup> 2014/4/1–2016/3/31; <sup>4</sup> 2016/10/1–2017/4/1; <sup>5</sup> 2016/4/1–2017/3/31; <sup>6</sup> 2015/4/1–2017/3/1; <sup>7</sup> 2016/4/1–2017/3/31 for Mine A; 2017/4/1–2018/3/31 for Mine B.

− − The parameters obtained for the tank model are shown in Table 3. Mine A has a smaller outflow height and a larger AMD reservoir than Mine B. Additionally, Mine A has the smaller value of outflow coefficient in the third stage, which directly affects AMD generation, compared to Mine B. This trend means that Mine A has the bigger reservoir and the longer residence time of AMD, which resulted in the smaller fluctuation of AMD, compared to Mine B. As we mentioned in the previous section, ore production was 230 million tons per year in Mine A and 25,000 tons per year in Mine B. This difference in scale should directly affect the difference in reservoir and residence time of AMD.


**Table 3.** Parameters of the tank model obtained for Mines A and B.

#### *3.2. Forecast of AMD Quantity*

The forecast of AMD quantity (underside) is shown in Figure 5, with the forecast of rainfall (upperside) proposed by MPI–ESM. According to MPI–ESM in RCP2.6, the forecast for temperature rise around the mines is about 2 ◦C by 2050, but remaining at about 1.0–2.5 ◦C after 2050. On the other hand, in RCP8.5, the forecast for temperature continues to rise and reaches +5.9 ◦C in 2100.

**Figure 5.** Forecast of AMD quantity calculated from the model until 2165, with the forecast of rainfall proposed by the Max Planck Institute–Earth System Model (MPI–ESM) for (**a**) Mine A and (**b**) Mine B.

In Mine A, the MPI–ESM shows that both rainfall and heavy rain frequency, which is the number of days per year with greater than 50 mm of rainfall, increase due to the temperature rise. In 2100, the forecast of rainfall is +21% for both RCP2.6 and RCP8.5, and forecast of heavy rain frequency increases 4 days/year for RCP2.6 and 8 days/year for RCP8.5, compared to the present. According to these trends, the AMD quantity calculated from the constructed model increases, as shown in Figure 5. The forecast for AMD quantity in 2100 is +27% for RCP2.6 and +31% for RCP8.5.

In Mine B, the MPI–ESM shows that the temperature rise of around 2 ◦C in the RCP2.6 scenario does not have much effect on rainfall and heavy rain frequency. In 2100, the forecast for rainfall decreases 1.5% and heavy rain frequency decreases 2 days/year, which results in a 0.55% increase for the forecast of AMD quantity, compared to the present. However, the temperature rise of 5.9 ◦C in the RCP8.5 scenario affects the forecast of rainfall and heavy rain frequency as much as for Mine A. In 2100, the rainfall forecast increases 22% and heavy rain frequency increases 5.5 days/year, which results in a 25% increase for the forecast of AMD quantity, compared to the present.

The forecast of the standard deviation of AMD quantity is shown in Figure 6. A comparison of the coefficient of variation of AMD quantity between the present and the future is shown in Table 4. Here, the coefficient of variation was calculated for 10 years from 2010 to 2020 for the present and from 2100 to 2110 for the future. The temperature rise due to global warming caused larger fluctuations in the AMD quantity for Mine A. In the case of Mine B, since the AMD reservoir is small and the AMD residence time is short, even if the rainfall fluctuation increases in the future due to global warming, the AMD fluctuation will remain largely as it is now. On the other hand, in the case of Mine A, since the AMD reservoir is larger and the AMD residence time is longer, AMD fluctuation tends to increase gradually in the future, affected by increases in rainfall fluctuation due to global warming. This trend suggests that AMD treatment might be unstable because of global warming in the future, especially for mines with larger AMD quantities.

**Figure 6.** Forecast of the standard deviation of AMD quantity for (**a**) Mine A and (**b**) Mine B.


**Table 4.** The coefficient of variation for AMD quantity at present and in the future.

In general, AMD quality tends to deteriorate as the AMD quantity increases. This is because when the AMD quantity increases, AMD comes in contact with a new pollution source in the mine. Actually, at present, the fluctuation of AMD quantity is larger in Mine B than in Mine A, and the fluctuation in AMD quality tends to be larger in Mine B as well. This suggests that fluctuations in AMD quantity due to global warming will cause large fluctuations in AMD quality.

#### **4. Conclusions**

The AMD quantity model was constructed for two closed mines in Japan. The model was constructed with a series tank model, and fitted by using daily data for one year, which were enough to obtain adequate parameters. The results showed that Mine B has a smaller AMD reservoir and a shorter AMD residence time than Mine A, resulting in a large fluctuation of AMD quantity in Mine B. The forecast of AMD quantity was also estimated based on the forecast of rainfall and temperature proposed by the MPI–ESM with IPCC RCP2.6 and RCP8.5 scenarios. The forecast results showed that temperature rise due to global warming will cause an increase in rainfall, resulting in increased AMD quantity. The fluctuation of rainfall will also increase due to global warming, increasing the fluctuation of AMD quantity in Mine A. The effect of global warming in Mine A will be bigger than in Mine B due to its larger reservoir and longer residence time of AMD.

In this study, it is expected that the quantity and fluctuation of AMD might increase due to global warming. This suggests that fluctuations in AMD quality might also increase. Therefore, when selecting future treatment methods, careful consideration should be given to whether or not the AMD fluctuation can be sufficiently dealt with in the future, especially for passive treatment.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2075-163X/10/5/430/s1, Figure S1: Observed and calculated AMD quantity by the tank model for (a) Mine A with half-year fitting, (b) Mine A with two-year fitting, (c) Mine B with half-year fitting, and (d) Mine B with two-year fitting; Figure S2: Observed and calculated AMD quantity by the tank model without correction of rainfall and temperature or considering snowmelt for (a) Mine A and (b) Mine B.

**Author Contributions:** Conceptualization, C.T.; methodology, K.F.; validation, M.K. and S.F.; investigation, K.F.; data curation, S.F.; writing—original draft preparation, C.T.; writing—review and editing, K.F. and S.F.; supervision, C.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was partially funded by JOGMEC (Japan Oil, Gas and Metals National Corporation) and the Center for Eco-Mining, Japan.

**Acknowledgments:** Part of this work was performed as the activities of the Waseda Research Institute for Science and Engineering and Research Organization for Open Innovation Strategy, Waseda University.

**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*

### **Immobilization of Lead and Zinc Leached from Mining Residual Materials in Kabwe, Zambia: Possibility of Chemical Immobilization by Dolomite, Calcined Dolomite, and Magnesium Oxide**

**Pawit Tangviroon 1,\*, Kenta Noto <sup>2</sup> , Toshifumi Igarashi <sup>1</sup> , Takeshi Kawashima <sup>3</sup> , Mayumi Ito <sup>1</sup> , Tsutomu Sato <sup>1</sup> , Walubita Mufalo <sup>2</sup> , Meki Chirwa <sup>4</sup> , Imasiku Nyambe <sup>4</sup> , Hokuto Nakata <sup>5</sup> , Shouta Nakayama <sup>5</sup> and Mayumi Ishizuka <sup>5</sup>**


Received: 22 July 2020; Accepted: 24 August 2020; Published: 28 August 2020

**Abstract:** Massive amount of highly contaminated mining residual materials (MRM) has been left unattended and has leached heavy metals, particularly lead (Pb) and zinc (Zn) to the surrounding environments. Thus, the performance of three immobilizers, raw dolomite (RD), calcined dolomite (CD), and magnesium oxide (MO), was evaluated using batch experiments to determine their ability to immobilize Pb and Zn, leached from MRM. The addition of immobilizers increased the leachate pH and decreased the amounts of dissolved Pb and Zn to different extents. The performance of immobilizers to immobilize Pb and Zn followed the following trend: MO > CD > RD. pH played an important role in immobilizing Pb and Zn. Dolomite in RD could slightly raise the pH of the MRM leachate. Therefore, the addition of RD immobilized Pb and Zn via adsorption and co-precipitation, and up to 10% of RD addition did not reduce the concentrations of Pb and Zn to be lower than the effluent standards in Zambia. In contrast, the presence of magnesia in CD and MO significantly contributed to the rise of leachate pH to the value where it was sufficient to precipitate hydroxides of Pb and Zn and decrease their leaching concentrations below the regulated values. Even though MO outperformed CD, by considering the local availability of RD to produce CD, CD could be a potential immobilizer to be implemented in Zambia.

**Keywords:** mine waste; contamination; batch experiments; lead; zinc; immobilization; remediation; Kabwe; Zambia

#### **1. Introduction**

Kabwe District was one of the most important mining regions in Zambia for almost a century (1902–1994). It was regarded as Southern Africa's principal lead (Pb)-zinc (Zn) producer, producing

over 1.8 and 0.8 Mt of Zn and Pb, respectively [1]. While in operation, no pollution laws were enforced to regulate the discharge from wastes of the mine; therefore, operations of the mine have left Kabwe with a massive amount of unattended mining residual materials (MRM), which still contain elevated amounts of heavy metals, particularly Zn, Pb, and iron (Fe). Weathering of MRM causes heavy metals to transport from the contaminated sites to the surrounding environments (groundwater, surface water, and soil) [2]. In particular, the redistribution of heavy metals through solute transport processes has been reported to be one of the most dangerous pathways, which invokes harmful effects on water sources of nearby ecosystems and health-threatening to the nearby residents [3–9]. Therefore, the remediation of heavy metals in and around the mine is necessary.

In our recent studies, several potential remediation techniques have been investigated to remediate the contaminated site in Kabwe. Silwamba et al. (2020) [10,11] have proposed the concurrent dissolution and cementation method. The method shows promising results in terms of Pb removal and recovery. However, Zn could not be effectively recovered from the extraction solution, and further investigation is needed. Biocementation by locally available bacteria has been studied by Mwandira et al. (2019) [12,13]. The results indicate that the biocemented material can be effectively used as a covering layer to prevent airborne contamination of metallic dust and infiltration of water into the waste. In the present study, chemical immobilization is introduced as an alternative and practical method to remediate the site.

Remediation techniques for heavy metals polluted sites can be classified into two main categories, in-situ and ex-situ. In general, ex-situ treatment has high efficiency; however, it is less cost-effective than in-situ remediation due significantly to the costs for excavation and transport of large quantities of contaminated materials. In-situ remediation avoids the excavation and transportation costs because of on-site treatments of contaminants. Various kinds of in-situ remediation techniques have been developed to immobilize or extract the heavy metals in the contaminated sites. Among them, chemical immobilization is cheap, easy to implement, and quick in execution [14,15]. Thus, this is the most promising technique, especially to be applied in one of the developing countries.

In in-situ chemical immobilization, the leaching potential of heavy metals from contaminated soils is reduced via sorption and/or precipitation processes by adding chemical agent (immobilizer) into the contaminated area. The performance of a variety of immobilizers, including carbonates, phosphates, alkaline agents, clay, iron-containing minerals, and organic matters, has been evaluated [14–23]. However, most of the studies have been conducted to remediate contaminated soil samples in which they generally contain much lesser metals contents than those in MRM. Moreover, because of the complex interactions between solutes and immobilizers, the definite efficiency of the immobilizer remains site-specific. In other words, there is no guarantee on the effectiveness of a particular immobilizer implemented on different contaminated sites. Therefore, the objective of this study was to evaluate and compare the performance of selected potential immobilizers (e.g., locally available, low-cost) to reduce the mobility of Pb and Zn leached from the highly contaminated sample (MRM).

It is necessary to apply immobilizers that is low cost and abundant in nature for remediating contaminated areas. Hence, in this study, raw dolomite (RD) was selected as one of the potential candidates because it is naturally available in a large quantity in Zambia [24]. It is a carbonate mineral; therefore, it can increase and buffer pH of MRM, leading to more adsorption and precipitation of cationic heavy metal ions [25–27]. In the present study, calcined dolomite (CD) was also used as an immobilizer. The heat treatment was performed to change the carbonate property of dolomite to be more alkaline [28]. As a result, the immobilizer was expected to strongly increase the pH of MRM, favoring the immobilization of heavy metals by hydroxide precipitation in addition to adsorption and precipitation of other secondary minerals. At the same time, commercially available alkaline-based agent, magnesium oxide (MO), was also tested to compare the ability of RD and CD on immobilizing heavy metals in MRM. The current study will provide meaningful information for the development of chemical immobilization to remediate heavy metals contaminated sites in Zambia.

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

#### *2.1. Solid Sample Collection, Preparation, and Characterization*

MRM was collected from the dumping site of Pb-Zn mine wastes in Kabwe, Zambia. The sampling was done using shovels at random points within the area shown in Figure 1. This leaching residual was selected as one of the representative wastes because the leaching concentrations of Zn and Pb from the waste were higher compared with the other wastes. Moreover, the storage size of this waste occupies more than 50% of the total dumping area. The sample was stored in vacuum bags and transported to the laboratory in Japan with permission by the Ministry of Agriculture, Forestry, and Fisheries of Japan. In preparation, it was air-dried under ambient conditions, lightly crushed, sieved using a 2 mm aperture screen, and kept in a polypropylene bottle before use. Particle sizes of less than 2 mm were chosen to follow the Japanese standard for the leaching test of contaminated soils [29].

□ ′ ″ ′ ″ ′ ″ ′ ″ **Figure 1.** Dumping site in Kabwe; () sampling area (top left: 14◦27′39′′ South, 28◦25′42′′ East; down right: 14◦27′55′′ South, 28◦26′16′′ East).

Three types of immobilizing agents were selected to immobilize Zn and Pb in the waste: raw dolomite, calcined dolomite, and magnesium oxide denoted as RD, CD, and MO, respectively. RD was taken from a dolomite quarry source near the MRM storage site, while CD was prepared by burning RD with particle sizes of less than 2 mm in a furnace at 700 ◦C for 2 h. MO was commercially available, purchased from Ube Industry, Japan. The same preparation procedure as that for MRM was also applied to these materials.

Chemical and mineralogical properties of all solid samples were characterized on the pressed powder of finely crushed samples (<50 µm) using an X-ray fluorescence spectrometer (XRF) (Spectro Xepos, Rigaku Corporation, Tokyo, Japan) and X-ray diffractometer (XRD) (MultiFlex, Rigaku Corporation, Tokyo, Japan), respectively. Sequential extraction was conducted to evaluate solid-phase heavy metals speciation of MRM. The procedure used in this study was modified from two well-known procedures, Tessier et al. (1979) [30] and Clevenger (1990) [31]. The modification was done by Marumo et al. (2003) [32], and it was widely used to extract the tailings sample, mineral processing wastes, and leaching residues [10,33–36]. The process can divide solid-phase heavy metals bounded to solid into five different phases, including exchangeable, carbonates, Fe-Mn oxides, sulfide/organic matter, and residual. The details of the extraction procedure are summarized in Table 1. The extraction was done on 1 g of the <2 mm MRM sample. Between each step, the extractant solution of the previous step was retrieved by centrifugation of the suspension at 3000 rpm for 40 min to separate the residue out of the leachate. The residual was then washed with 20 mL of deionized water. Finally, the washing and extractant solutions were mixed, diluted to 50 mL, filtrated through 0.45-µm Millex® filters, and kept in polypropylene bottles prior to chemical analysis.


**Table 1.** Sequential extraction for heavy metals speciation.

#### *2.2. Batch Leaching Experiments*

Batch leaching experiments were performed using 250 mL polypropylene Erlenmeyer flasks with a lateral reciprocating shaker (EYELA Multi Shaker MMS, Tokyo Rikakikai Co., Ltd., Tokyo, Japan). All batches were conducted under ambient conditions by mixing 15 g of solid sample to 150 mL of deionized water (1:10 solid-to-liquid ratio) at 200 rpm for 6 h. Five replications of the leaching tests of MRM were conducted, while a single run of every immobilization experiment was done. The reason is that in immobilization tests, we adjusted the addition of immobilizers and did not control the pH of the suspension. In other words, pH is determined by the complex chemical and physical interactions between immobilizer and MRM. Thus, a variation of pH can be easily observed even though the same mixing ratio between MRM and immobilizer and the conditions is employed. To avoid the uncertainty of the variation of pH at the same immobilizer:MRM mixing ratio, we varied the immobilizer:MRM mixing ratios at 1:100, 3:100, 1:20, and 1:10 to evaluate the performance of immobilizers. After 6 h of shaking, the pH, electrical conductivity (EC), redox potential (ORP), and temperature were immediately measured. The leachates were collected by first centrifuging the mixtures at 3000 rpm for 40 min to separate the suspended particles. The supernatants were then filtered with 0.45-µm Millex® filters (Merck Millipore, Burlington, MA, USA) and kept in air-tight polypropylene bottles prior to chemical analysis.

#### *2.3. Chemical Analysis of Liquid Samples*

Inductively coupled plasma atomic emission spectrometer (ICP-AES) (ICPE-9000, Shimadzu Corporation, Kyoto, Japan) and inductively coupled plasma atomic emission mass spectrometry (ICP-MS) (ICAP Qc, Thermo Fisher Scientific, Waltham, MA, USA) were used to quantify the dissolved concentrations of heavy metals and coexisting ions. The analyses were performed on the pretreated liquid samples in which 1% by volume of 60% nitric acid (HNO3) was added to the liquid samples. The acidification was done to make sure that all target elements were in a dissolved form. The non-acidified samples, on the other hand, were used to determine alkalinity or acid resistivity. This parameter is generally reported as bicarbonate concentration (meq/L), quantified by titration of a known volume of sample with 0.01 M sulfuric acid (H2SO4) until pH 4.8. All chemicals used were reagent-grade.

#### *2.4. Geochemical Modeling*

An aqueous geochemical modeling program, PHREEQC (Version 3, U.S. Geological Survey, Sunrise Valley Drive Reston, VA, USA) [37], was used to aid in the interpretation of the experimental results. The program can determine the parameters that may affect the mobility of heavy metals from MRM, such as stability of minerals and chemical species. The input data included temperature, pH, ORP, and concentrations of heavy metals and other coexisting ions. Thermodynamic properties were taken from the WATEQ4F database.

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

#### *3.1. Properties of Solid Samples*

The mineralogical and chemical compositions of MRM and immobilizers, including RD, CD, and MO, are listed in Tables 2 and 3, respectively. MRM was composed of anglesite (PbSO4) as a primary mineral; zinkosite (ZnSO4) and quartz (SiO2) as the second-highest; and goethite (FeOOH), hematite (Fe2O3), and gypsum (CaSO4·2H2O) as the miner minerals. Anglesite and zinkosite are commonly found as the weathering products of Pb- and Zn-sulfides under natural oxygenated environments [38,39]. Therefore, the presence of these two minerals indicates that MRM has already been exposed to the surface environment for a long time before the sampling was done. It can also be expected that other than goethite and hematite, MRM also contained amorphous iron-(hydr)oxides and iron-sulfate salts (e.g., melanterite, coquimbite) since goethite and hematite were found as the miner minerals but Fe2O<sup>3</sup> content was the highest among all compositions detected. The contents of Pb and Zn in MRM were 10.9% and 8.1%, respectively. The values of both metals were extremely high and exceeded the permissible limit in soil, 600 mg/kg for Pb and 1500 mg/kg for Zn [40]. However, this does not guarantee that MRM can release significant amounts of Pb and Zn since their mobility also depends significantly on the chemical speciation. Sequential extraction was then performed, and the result showed that around 40% of the total contents of Pb and Zn were in mobile fractions (exchangeable, carbonates, and oxidizable) under surface environments (Figure 2). This confirms that MRM could be a potential source contaminating the surrounding environment with Pb and Zn.

**MRM RD CD MO** Quartz ++ - - - Gypsum + - - - Anglesite +++ - - - Zinkosite ++ - - - Hematite + - - - Goethite + - - - Dolomite - +++ +++ - Calcite - - + - Magnesia - - + +++

**Table 2.** Mineralogical composition of solid samples.

+++: Strong; ++: Moderate; +: Weak; -: None.

**Table 3.** Chemical composition of solid samples (the unit is in wt%).


The most dominant mineral found in RD was dolomite (CaMg(CO3)2). Magnesium (Mg) and calcium (Ca) oxides contents accounted for more than 95% with a molar ratio of Ca to Mg of 1.2. This indicates that RD adequately consisted of pure dolomite. Burning RD at 700 ◦C for 2 h generated

the new type of immobilizer, CD. Calcite (CaCO3) and magnesia (MgO) were detected in addition to dolomite. With almost the same molar ratio of Ca to Mg in CD compared with that in RD, it clearly indicates that the calcination process transformed dolomite into calcite and magnesia. MO composed only of magnesia with 100% MgO content, which shows pure magnesia.

□ □ □ □ □ **Figure 2.** Solid-phase Pb and Zn speciation of mining residual materials (MRM); () exchangeable, () carbonates, () sulfide/organic matter, () Fe-Mn oxides, and () residual.

#### *3.2. Leaching Characteristic of MRM*

Table 4 shows the leaching characteristic of MRM. The experimental values were reproducible with high precision since the standard deviations of all parameters were quite small. Four heavy metals, Pb, Zn, cadmium (Cd), and copper (Cu), were leached at the concentrations falling within the instrument detection limits of ICP-AES and -MS. Among these heavy metals, the leaching concentrations of only Pb (2.1 mg/L) and Zn (365 mg/L) exceeded the effluent standard in Zambia (0.5 mg/L for Pb and 1 mg/L for Zn) [41]. Therefore, this study focused only on the mobility of these two metals. PHREEQC simulation on the saturation indices of all possible Pb- and Zn-minerals showed that with the given conditions and components in MRM leachate, only saturation index of anglesite fell within a common error interval used to indicate saturation equilibrium (±0.2) (Table 5). The result indicates that the low solubility of anglesite restricted the dissolved concentration of Pb, while no restriction by means of precipitation was observed on the leaching of Zn. This can explain why 4.51% of the total Zn content was leached from MRM, while only 0.02% was observed in the case of Pb leaching.



\* Regulated value in mg/L specified by the Environment Management Act (2013) [40].


**Table 5.** Calculated saturation indices of possible Pb- and Zn-minerals in MRM leachate.

Calcium ion (Ca2+) and sulfate ion (SO<sup>4</sup> <sup>2</sup>−) were the major ions in the leachate, accounting for more than 85% of the total dissolved ions. These ions were likely to be enriched by the dissolution of soluble phase minerals, such as gypsum and zinkosite [42,43]. However, the dissolution of these two minerals might not only be the sources of SO<sup>4</sup> <sup>2</sup><sup>−</sup> since the molar ratio of Ca2<sup>+</sup> and Zn to SO<sup>4</sup> 2− was lower than one. The sulfide fractions of both metals in MRM (Figure 2), together with the slightly acidic pH (5.2) and positive ORP (300 mV) of the leachate, suggest that the oxidation of sulfide minerals (e.g., pyrite, galena, sphalerite) and dissolution of iron-sulfate salts (e.g., melanterite, coquimbite) also occurred and attributed to the enrichment of SO<sup>4</sup> <sup>2</sup>−, even though they were not detected by XRD. This could also partly contribute to the enrichment of Pb and Zn in the leachate.

#### *3.3. Potential of Immobilizers*

#### 3.3.1. Effects of Addition of Immobilizers on pH and Coexisting Ions

Changes in the pH of the leachate as a function of the amounts of addition of RD, CD, and MO are shown in Figure 3. The pH increased from 5.2 to 6.7, 8.2, and 9.8 with increasing RD, CD, and MO addition from 0 to 10%. When the same amount of immobilizer was added, the performance of the immobilizers to increase the leachate pH followed the order: MO > CD > RD. The results clearly showed the improvement of the alkaline property of CD over RD. The variation in pH could mainly be attributed to the liming effect(s) of dolomite (Equation (1)) in RD treatments, of dolomite (Equation (1)), of calcite (Equation (2)), and of magnesia (Equation (3)) in CD treatments, and of magnesia (Equation (3)) in MO treatments [44–47].

$$2\text{CaMg(CO}\_3\text{)}\_2 + 2\text{ H}^+ \Leftrightarrow \text{Ca}^{2+} + \text{Mg}^{2+} + 2\text{HCO}\_3^-\text{ (}\text{ }\tag{1}$$

$$\rm CaCO\_3 + H^+ \Leftrightarrow Ca^{2+} + HCO\_3^- \tag{2}$$

$$\text{MgO} + \text{H}\_2\text{O} \Leftrightarrow \text{Mg(OH)}\_2 \Leftrightarrow \text{Mg}^{2+} + 2\text{OH}^-. \tag{3}$$

Figure 4a–c illustrates the leaching concentrations of major coexisting ions, Ca2+, magnesium (Mg2+), and SO<sup>4</sup> <sup>2</sup>−, as a function of the amount of immobilizer added. The dissolved concentrations of Ca2<sup>+</sup> and Mg2<sup>+</sup> in RD treatments were higher than those in MRM and increased with the increasing addition of RD, indicating the simultaneous leaching of Ca2<sup>+</sup> and Mg2<sup>+</sup> from the dissolution of dolomite (Equation (1)). In CD treatments, Ca2<sup>+</sup> and Mg2<sup>+</sup> were also leached at higher concentrations than those in MRM leachate. At the same amount of CD and RD addition, the leaching concentration of Mg2<sup>+</sup> in CD treatment was higher than that in RD treatment, while almost the same dissolved concentration of Ca2<sup>+</sup> was observed in both treatments. This, together with the result that CD contained less dolomite and more calcite compared to those in RD, suggest the occurrence of hydration of magnesia (Equation (3)) in addition to the dissolution of carbonate minerals (Equations (1) and (2)) in CD treatments. Moreover, the difference in the leaching concentration of Mg2<sup>+</sup> between CD and RD treatments became more significant as the addition of immobilizers increased, which indicates that as pH increased, the hydration of magnesia (Equation (3)) played a more important role in controlling pH. In MO treatments, the concentration of Mg2<sup>+</sup> increased with higher addition of MO, while almost no change in the concentration of Ca2<sup>+</sup> from that in MRM was observed. This result suggests the occurrence of hydration of magnesia (Equation (3)) in MO treatments. <sup>−</sup>

<sup>−</sup>

**Figure 4.** Leaching concentrations of major ions upon addition of immobilizers: (**a**) Ca2+, (**b**) Mg2+, and (**c**) SO<sup>4</sup> 2−.

−

−

At the same amount of immobilizer added, the leaching concentration of SO<sup>4</sup> <sup>2</sup><sup>−</sup> was the highest when treating MRM with MO, followed by CD and RD (Figure 4c). In consideration of the trace amount of SO<sup>3</sup> content in all immobilizers, the results suggest that MRM should be the source of SO<sup>4</sup> 2−, and the leaching of SO<sup>4</sup> <sup>2</sup><sup>−</sup> might be caused by the change in the parameter(s) of the leachate, triggered

−

−

−

− −

−

− −

− −

− −

by the addition of immobilizer. Figure 5 illustrates the correlation between the leaching concentration of SO<sup>4</sup> <sup>2</sup><sup>−</sup> vs. pH. The SO<sup>4</sup> <sup>2</sup><sup>−</sup> concentration exhibited a strong positive correlation with pH (correlation coefficient (*r*) = 0.92, *p* < 0.01), suggesting that SO<sup>4</sup> <sup>2</sup><sup>−</sup> level could mainly be influenced by pH due possibly to the following mechanisms: (1) desorption, (2) production by oxidation of sulfide minerals, (3) common-ion of between calcite, dolomite, and gypsum, and (4) dissolution of anglesite. The pH increase led to a higher negative surface potential of MRM, thereby decreasing the affinity of SO<sup>4</sup> 2− toward the surface of MRM. However, Tabatabai (1987) [48] reported that since the adsorption of SO<sup>4</sup> 2− was only flavored under strongly acidic conditions, the amount of adsorbed SO<sup>4</sup> <sup>2</sup><sup>−</sup> became almost negligible under weakly acidic pH. This means that the desorption might not be a viable explanation in this study since all leachates' pH ranged from weakly acidic (5.2) to moderately alkaline (9.8). Therefore, the fact that the oxidation rate of sulfide minerals, such as pyrite and galena, increases with pH becomes the potential reason contributing to the higher leaching concentration of SO<sup>4</sup> <sup>2</sup><sup>−</sup> [49,50]. However, the enrichment of SO<sup>4</sup> <sup>2</sup><sup>−</sup> could be restricted by the solubility of gypsum because the saturation index of gypsum was within the equilibrium condition range of ±0.2 in all leachates. This could explain why the leaching concentration of SO<sup>4</sup> <sup>2</sup><sup>−</sup> slightly increased with pH at lower pH region where the dissolution of dolomite containing in RD and calcite and dolomite containing in and CD tended to control the pH (Equations (1) and (2)). In other words, Ca2<sup>+</sup> produced from the dissolution of calcite and dolomite precipitated with SO<sup>4</sup> <sup>2</sup><sup>−</sup> to form gypsum, thereby restricting SO<sup>4</sup> <sup>2</sup><sup>−</sup> concentration. Meanwhile, as pH became more alkaline, the concentration of SO<sup>4</sup> <sup>2</sup><sup>−</sup> increased rapidly. This could be attributed to the less contribution of calcite and dolomite dissolution (Equations (1) and (2)) to control the leachate pH in the case of CD addition, conjointly in MO treatments, only hydration of magnesia (no production of Ca2+) (Equation (3)) was found to control the pH of the leachates. The pH-dependent solubility of anglesite could also be attributed to the rapid increase of SO<sup>4</sup> <sup>2</sup><sup>−</sup> concentration under alkaline conditions since anglesite was originally contained in MRM and is unstable under alkaline pH [51].

**Figure 5.** Leaching concentration of SO<sup>4</sup> <sup>2</sup><sup>−</sup> vs. pH.

− 3.3.2. Effects of the Addition of Immobilizers on Mobility of Heavy Metals

− − − In this study, the performance of immobilizers was evaluated based on the solubility of Pb and Zn. Figure 6a,b shows the changes in Pb and Zn concentrations as a function of the amount of addition of immobilizers. In general, the leaching concentrations of Pb and Zn exponentially decreased with an increase in the dose of immobilizers. Figure 7a,b illustrates the correlation between leaching concentrations of Pb and Zn vs. pH. The mobility of heavy metals was strongly influenced by pH, indicated by significant negative correlations of Pb-pH (*r* = −0.92) and Zn-pH (*r* = −0.87) at the 0.01 significance level (2-tailed). Coupled with the results of the leaching concentration of SO<sup>4</sup> <sup>2</sup><sup>−</sup> and the characteristic of immobilizers, as well as the leaching condition used in the current study, the major modes of Pb and Zn attenuation could be either one or more of the following

−

mechanisms: precipitations of metal-sulfate, -carbonate, and/or -hydroxide, co-precipitation of metal with iron-(oxy)hydroxides, and metal ion adsorption to immobilizer. The formation of low soluble anglesite under acidic conditions was expected since the leaching concentration of SO<sup>4</sup> <sup>2</sup><sup>−</sup> was high and increased with pH (Figure 5) [52,53]. The carbonate property of RD and CD might result in the precipitation of cerussite (PbCO3), hydrocerussite (Pb3(CO3)2(OH)2), smithsonite (ZnCO3), and hydrozincite (Zn3(CO3)2(OH)2) [35,54–56]. The carbonate precipitations of Pb and Zn are also expected to occur in MO treatments since the experiments were done under atmospheric conditions in which carbon dioxide (CO2) in the atmosphere was freely dissolved [57]. However, the simulation results by PHREEQC showed that except hydrozincite in 3% addition of MO treatment, the precipitations of anglesite, cerussite, hydrocerussite, smithsonite, and hydrozincite were thermodynamically unfavorable (saturation index < −0.2) regardless of the type and amount of immobilizer added (Table 6). Therefore, the possible immobilization mechanisms of Pb and Zn in all types of immobilizers could be narrowed down to the hydroxide precipitation, adsorption, and co-precipitation.

**Figure 6.** Leaching concentrations of heavy metals upon addition of immobilizers: (**a**) Pb and (**b**) Zn (dashed lines represent the effluent standards of Pb and Zn in Zambia).

**Figure 7.** Leaching concentrations of heavy metals vs. pH: (**a**) Pb vs. pH and (**b**) Zn vs. pH (dash lines represent the effluent standards of Pb and Zn in Zambia).

− − − − − To verify the dominant mechanism(s), pH-dependent solubility diagrams of Pb- and Zn-hydroxides were plotted (Figure 8a,b). Points in the figures represent the relationship between the logarithmic activity of divalent heavy metal and pH in each batch test. The solid lines demonstrate the solubility of heavy metal hydroxides. Therefore, any point located on or close to the line implies the hydroxide precipitation-controlled sequestration process.

− − − −

− − − −

− − − − − − − − − −

− − − − − − − − −

− − − − − − − − − − − − − − − − − − − − − − − − −

− − − − − − − − − − − − − − − − − − − −

− − − − − − − − − −

− − − − − − − − − −


**Table 6.** Calculated saturation indices of anglesite, cerussite, hydrocerussite, smithsonite, and hydrozincite in leachates with the addition of RD, CD, and MO.

**Figure 8.** pH-dependent solubility diagrams of (**a**) Pb-hydroxide and (**b**) Zn-hydroxide ([ ] represents activity).

− In the case of RD addition, a discrepancy from the equilibrium line for Pb and Zn was observed. This means that at pH from weekly acid to neutral, hydroxide precipitation was not the main mechanism controlling the mobility of Pb and Zn. Therefore, Pb and Zn were suspected to be immobilized by sorption and co-precipitation with iron-(oxy)hydroxides. The sorption was likely to occur since dolomite, major mineral in the immobilizer, can adsorb Pb and Zn [58,59]. Besides, adding more of this immobilizer induced the leachate pH increase. This could change the surface charge of goethite, hematite, and iron-(hydr)oxide compounds in MRM to more negative, thereby increasing their adsorption ability against cationic divalent Pb and Zn [58–63]. During the neutralization process under ambient conditions, iron-(oxy)hydroxides precipitate from the oxidative dissolution of pyrite and dissolution of iron-bearing salts [63–66]. The precipitations of iron-(oxy)hydroxides have been reported by many studies to induce co-precipitation of divalent metals, including Pb(II) and Zn(II) [50,67,68], and thus the co-precipitation of Pb and Zn with iron-(oxy)hydroxides was also expected in RD treatments. From the above explanations, adding more RD should reduce the leaching concentration of both metals. However, leaching concentration of Pb increased from 0.98 mg/L to 1.4 mg/L when RD rose from 5% to 10% (Figure 6a). This could be attributed to the stability of Pb(II) species as a function of pH. Theoretically, as pH increases under acidic region, more of free Pb(II) ion tends to complex with OH- and CO<sup>3</sup> <sup>2</sup>−, generating larger ion with lower charge (Pb(OH)<sup>+</sup> and PbCO3), which lowers the affinity of Pb to the surface of the potential adsorbents and inhibits the co-precipitation [69–71].

On the other hand, adding CD and MO made most of the logarithmic leaching activities of Pb2<sup>+</sup> and Zn2<sup>+</sup> to approach their solubility product lines. This means that hydroxide precipitation is the dominant mechanism of attenuating Pb and Zn. Regardless of the type of immobilizer, at low pH, the logarithmic activities of both metals were slightly lower than their equilibrium lines and then tended to stay on or be slightly higher than the lines afterward. This probably indicates that at low pH, adsorption and co-precipitation with iron-(oxy)hydroxides also occurred in addition to the precipitation, but as pH got higher, they diminished. There are two probable explanations for this phenomenon as follows: (1) competition with strong competing ion (Mg2+) and (2) change in specification of the dissolved metals. The pH alteration mechanisms of CD and MO appeared to generate Mg2<sup>+</sup> as a by-product (Equations (1) and (3)). Because of this, the concentration of Mg2<sup>+</sup> significantly increased with pH with a correlation coefficient of 0.97, *p* < 0.01(Figure 9). Therefore, as pH increased, high concentration of Mg2<sup>+</sup> could compete for Pb and Zn for adsorption sites and for co-precipitation with iron-(oxy)hydroxides, attributing to the less contribution of the adsorption and co-precipitation on the immobilization process. Increasing pH could also result in the redistributions of Pb(II) and Zn(II) species. The fraction of free Pb(II) and Zn(II) reduces as pH increases since they are thermodynamically preferable to be hydrolyzed forming -(OH)+, -(OH)2, and -(OH)<sup>3</sup> <sup>−</sup> [69–72]. Moreover, since the systems contained high dissolved carbonate, the formation of carbonate complexes of Pb(II) and Zn(II) was also expected [69,73]. Once these complexes are formed, their abilities to get adsorbed and co-precipitated are inhibited by the larger size and lower positive potential they become.

**Figure 9.** Leaching concentration of Mg2<sup>+</sup> vs. pH.

#### 3.3.3. Performance of Immobilizers

When the same amount of immobilizer was added, the dissolved concentrations of Pb and Zn were the highest in RD treatment, second highest in CD treatment, and the lowest in MO treatment (Figure 6a,b). As previously mentioned, adsorption, co-precipitation, and hydroxide precipitation were the major sink of Pb and Zn, and their mobilities depended strongly on pH. Because of the carbonate property of dolomite, RD could not raise the pH of MRM leachate to the value favoring the precipitations of Pb- and Zn-hydroxides. Therefore, RD treatments remediated Pb and Zn by adsorption and co-precipitation in which it was insufficient to reduce the leaching concentrations of Pb and Zn down below their regulated values. On the other hand, magnesia in CD and MO played a significant role in increasing the leachate pH of MRM into the alkaline region. Lead and Zn were then mainly immobilized by precipitation as hydroxides. Thus, metal concentrations as high as 368 mg/L for Zn and 2.1 mg/L for Pb released from MRM were reduced to the values below their regulated concentrations. Efficiency-wise, MO was the most effective immobilizer in immobilizing Pb and Zn since it contained the highest MgO content. However, CD could be the immobilizer of choice since it can be produced from the naturally abundant material in Zambia (RD), and its performance was almost the same as that of MO.

#### **4. Conclusions**

The leachate of MRM was slightly acidic (pH 5.2) and contained high concentrations of Pb (2.1 mg/L) and Zn (365 mg/L), exceeding those of Zambian regulation. When immobilizers were introduced, the leachate pH increased, and the leaching concentrations of Pb and Zn decreased. Lead and Zn immobilized by RD were interpreted by the adsorption and co-precipitation mechanisms. On the contrary, chemical immobilization using CD and MO suppressed Pb and Zn leaching mainly by the hydroxide precipitation. Of the immobilizers investigated, only CD and MO decreased the dissolved Pb and Zn concentrations to below their regulated values, in which MO had a higher performance than CD. The results show that heat treatment on RD to produce CD drastically improved the immobilizing performance of Pb and Zn. Even though MO provided the highest efficiency, Pb and Zn could also be effectively immobilized by giving an adequate amount of CD. Therefore, by considering the availability of CD in the local area, CD could be the most promising chemical agent to be implemented in Zambia.

**Author Contributions:** Conceptualization, P.T.; methodology, P.T., K.N., T.I., and T.K.; formal analysis, P.T. and T.I.; writing—original draft preparation, P.T.; writing—review and editing, P.T., T.I., M.I. (Mayumi Ito), T.S., W.M., M.C., and I.N.; supervision, P.T. and T.I.; project administration, M.I. (Mayumi Ishizuka), S.N., and H.N.; funding acquisition, M.I. (Mayumi Ishizuka), S.N., and H.N.; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by JST/JICA SATREPS (Science and Technology Research Partnership for Sustainable Development; No. JPMJSA1501) and aXis (Accelerating Social Implementation for SDGs Achievement; No. JPMJAS2001) funded by JST.

**Acknowledgments:** The authors would like to acknowledge JST/JICA SATREPS (Science and Technology Research Partnership for Sustainable Development; No. JPMJSA1501) and aXis (Accelerating Social Implementation for SDGs Achievement; No. JPMJAS2001) funded by JST for the financial support.

**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* **Genomic Analysis of a Newly Isolated** *Acidithiobacillus ferridurans* **JAGS Strain Reveals Its Adaptation to Acid Mine Drainage**

**Jinjin Chen 1,† , Yilan Liu 1,†, Patrick Diep <sup>1</sup> and Radhakrishnan Mahadevan 1,2,\***


**Abstract:** *Acidithiobacillus ferridurans* JAGS is a newly isolated acidophile from an acid mine drainage (AMD). The genome of isolate JAGS was sequenced and compared with eight other published genomes of *Acidithiobacillus*. The pairwise mutation distance (Mash) and average nucleotide identity (ANI) revealed that isolate JAGS had a close evolutionary relationship with *A. ferridurans* JCM18981, but whole-genome alignment showed that it had higher similarity in genomic structure with *A. ferrooxidans* species. Pan-genome analysis revealed that nine genomes were comprised of 4601 protein coding sequences, of which 43% were core genes (1982) and 23% were unique genes (1064). *A. ferridurans* species had more unique genes (205–246) than *A. ferrooxidans* species (21–234). Functional gene categorizations showed that *A. ferridurans* strains had a higher portion of genes involved in energy production and conversion while *A. ferrooxidans* had more for inorganic ion transport and metabolism. A high abundance of *kdp*, *mer* and *ars* genes, as well as mobile genetic elements, was found in isolate JAGS, which might contribute to its resistance to harsh environments. These findings expand our understanding of the evolutionary adaptation of *Acidithiobacillus* and indicate that *A. ferridurans* JAGS is a promising candidate for biomining and AMD biotreatment applications.

**Keywords:** *Acidithiobacillus*; acid mine drainage; biomining; comparative genomics

#### **1. Introduction**

With continually increasing concerns about acid mine drainage (AMD) contamination and the depletion of high-grade ores, innovative and sustainable methods to recover heavy metals from tailings and AMD as well as to treat AMD pollution are urgently needed [1]. Pyrometallurgical and hydrometallurgical routes are the conventional methods for metal recovery, but they are environmentally unsustainable, with a high cost in terms of operating on low-grade ores [2,3]. Though many techniques have been applied for AMD management, such as neutralization, adsorption, oxygen barriers, bactericides and so on, most of those options are unsustainable and unaffordable [4]. Compared with conventional and other emerging reprocessing techniques, bioleaching is considered as a simple, highly efficient, safe, low-cost, more easily managed and eco-friendly technique to facilitate sustainable mining and prevent AMD [3]. Bioleaching facilitates metal mobilization from solid metal sulfides into their water-soluble forms by different microorganisms via direct and indirect bioleaching [5]. Direct bioleaching can be summarized as:

$$\text{MeS} + 2\text{O}\_2 \rightarrow \text{MeSO}\_4 \tag{1}$$

**Citation:** Chen, J.; Liu, Y.; Diep, P.; Mahadevan, R. Genomic Analysis of a Newly Isolated *Acidithiobacillus ferridurans* JAGS Strain Reveals Its Adaptation to Acid Mine Drainage. *Minerals* **2021**, *11*, 74. https://doi.org/ 10.3390/min11010074

Received: 20 November 2020 Accepted: 11 January 2021 Published: 13 January 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/).

while indirect bioleaching can be described as:

$$\rm{MeS} + \rm{Fe\_2(SO\_4)\_3} \to \rm{MeSO\_4} + \rm{2Fe\_2SO\_4} + \rm{S}^0 \tag{2}$$

Microorganisms take part in and accelerate the oxidation of mineral sulfide to sulfate or the reoxidation of ferrous iron to ferric iron [6]. Although many factors affect the bioleaching process such as temperature, pH, dissolved oxygen, redox potential and formation of secondary minerals [7–9], indigenous bacteria play a crucial role in effective bioleaching [10]. Therefore, an increasing research effort has been placed on discovering and characterizing indigenous robust microbes that are resistant to high metal concentrations in extremely acidic environments [11,12].

*Acidithiobacillus* is a group of Gram-negative, chemoautotrophic, acidophilic aerobes which dominates in all types of extremely acidic habitats, suggesting its tremendous potential in the bioleaching process [13]. Genome sequencing and analysis of this genus, revealing its adaption to harsh environments, will help us to better understand its mechanisms and provide insights for future genetic engineering strategies to make bioleaching more efficient and versatile. Li et al. [14] analyzed and validated genomic information from this genus, including seven species and 37 strains, and revealed that *Acidithiobacillus* spp. recruited and consolidated novel functionalities via horizontal gene transfer, gene duplication and purifying selection to cope with challenging environments. Presently, there are 10 species reported in this genus: *A. ferrooxidans*, *A. ferridurans*, *A. ferrivorans*, *A. ferrianus*, *A. ferriphilus*, *A. albertensis*, *A. caldus*, *A. thiooxidans*, *A. sulfuriphilus* and *A. cuprithermicus*. The first five species were reported to generate energy by oxidizing ferrous iron, sulfur and hydrogen. Since soluble ferric iron produced from this ferrous iron oxidation can serve as a powerful oxidant to accelerate the dissolution of sulfidic minerals to release target metals, *Acidithiobacillus* species such as *A. ferrooxidans* have drawn focused attention [7]. Since the 1940s, more than 500 isolates of *A. ferrooxidans* have been reported and whole-genome sequencing has been performed for nine isolates. The strain *A. ferrooxidans* ATCC23270 was chosen as the model strain in this genus because of its extensive description in the literature and its wide usage [15]. It can oxidize ferrous iron, reduced inorganic sulfur compounds (RISCs) and hydrogen to generate energy as ATP. Additionally, it can fix atmospheric carbon dioxide and nitrogen as nutrition sources [16]. *A. ferrooxidans* has been successfully applied to recover metals such as copper, nickel, zinc, arsenic and uranium from low-grade ores, sewage sludge and contaminated sediments [13].

*A. ferridurans* is a species reclassified from *A. ferrooxidans* by Hedrich and Johnson [17] in 2013 because the species' DNA–DNA hybridization (63%) was lower than the threshold value (70%) used to delineate species. Moreover, when comparing *A. ferrooxidans* ATCC23270 with *A. ferridurans* JCM18981 (formerly called *A. ferrooxidans* ATCC33020), the latter showed better resistance to lower pH and higher concentrations of Fe2+, Ni2+ and Mg2+ [17,18]. Thus, the genus *A. ferridurans* might be a good candidate chassis for industrial applications in the field of biomining and bioremediation. For instance, a newly isolated *A. ferridurans* SBU-SH2 was used for flask and column bioleaching from lowgrade uranium ore, which generated 96% and 95.5% uranium extraction in 7 and 26 days, respectively [19,20]. However, at present, only one whole-genome sequence of *A. ferridurans* species, *A. ferridurans* JCM18981 isolated from uranium drainage water in Japan, is available [21]. The limited genome information of *A. ferridurans* species hinders our understanding of the mechanism and evolutionary history underpinning its unique metal and acid resistance.

In our previous study, a dominant strain belonging to the *Acidithiobacillus* genus was found based on 16S rRNA gene sequence analysis, which took up 92.6% of the enriched culture from acid mine drainage (AMD) in Sudbury, Canada [22]. After whole-genome sequencing and assembly, we reclassified and named it as *A. ferridurans* JAGS based on simple 16S rRNA gene and ANIb analyses and announced its genome [23]. To better ununderstand this strain and provide useful data for future research, a detailed genomic analysis was further performed in this study. First, a side-by-side comparison of iron, nickel

and low-pH tolerance between *A. ferridurans* JAGS and *A. ferrooxidans* ATCC23270 was conducted. To reveal the genetic traits associated with heavy metal and acid resistance in isolate JAGS, its genomic data were compared with the reported genomes of *Acidithiobacillus* strains. A pan-genome analysis was further conducted on these genomes to explore the metabolic features leading to the diversity of physico-biochemical traits. Functional genes and pathways responsible for heavy metal and acid resistance were analyzed and compared. A mobile genetic element analysis further suggested that gene transfers among these strains likely enabled adaptation to challenging environments. The insights gained in this study enhanced our understanding of the mechanism and evolutionary history of heavy metal and acid resistance in *A. ferridurans* and we suggest possible approaches for engineering *A. ferridurans* as a microbial chassis for biomining processes.

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

#### *2.1. Culture Media, Phenotypic and Growth Observations*

The strain *A. ferridurans* JAGS was isolated from acidic mine drainage in our previous study [22]. It was the dominant species and made up 92.6% of the enriched culture, based on 16S rRNA gene sequence analysis. The strain *A. ferrooxidans* ATCC23270 was purchased from American Type Culture Collection (ATCC). The phenotypic features of isolated JAGS were observed on a light microscope (Nikon Eclipse E400, Nikon, Melville, NY, USA) and a scanning electron microscope (FEI XL30 SEM, Philips, Eindhoven, Holland), separately. The strain JAGS was cultured with either 9K-Fe2+ (160 mM ferrous iron, pH 2.0) or 9K-S<sup>0</sup> (0.5% elemental sulfur, pH 3.0) at 30 ◦C with shaking at 180 rpm or with 2:2 solid medium in an incubator at 30 ◦C [24].

The abilities to tolerate elevated concentrations of ferrous iron and nickel and low pH were tested in a side-by-side comparison between *A. ferridurans* JAGS and *A. ferrooxidans* ATCC23270. Cultures grown in the 9K-Fe2+ medium were inoculated at a ratio of 20% into the same medium with Fe2+ (200 or 320 mM), Ni2+ (100 or 200 mM) or at pH 1.5 and then incubated at 30 ◦C for 22 h. The Fe2+ concentration was tested by the colorimetric ferrozine-based assay [25] and the ferrous iron oxidation rate was calculated by using consumed Fe2+ divided by its initial concentration, as in a previous study [26].

The growth features of *A. ferridurans* JAGS in 9K-Fe2+ or 9K-S<sup>0</sup> medium were further monitored by detecting pH, iron or sulfate concentrations and cell numbers during incubation by removing samples at intervals. The pH value was detected using a pH meter (Thermo Scientific®, Orion Star A211, Waltham, MA, USA). Ferrous and ferric iron concentrations were examined by the colorimetric ferrozine-based assay [25]. Sulfate was detected using a turbidimetric method [27]. Three cell-counting methods were tested: direct cell counting, optical density (OD600) measurements and plate counting. For the direct cell-counting method, samples were taken from media and cell numbers were estimated using a hemocytometer (Hausser Scientific, Horsham, PA, USA). For the OD<sup>600</sup> method, cells were harvested, washed twice with a basal salt buffer (4.5 g/L (NH4)2SO4, 0.15 g/L KCl and 0.75 g/L MgSO4·7H2O) and resuspended in 6% betaine prior to measurements. The plate counting was carried out by spreading proper diluted samples on 2:2 solid plates and colonies were counted after 7–10 days.

#### *2.2. Comparative Genomics*

Details of *A. ferridurans* JAGS genomic DNA extraction, sequencing, assembly and annotation are described in our previous paper [23]. The complete genome sequence of the isolate JAGS contains 2,933,811 bp with a GC (guanine-cytosine) content of 58.6%. The Similar Genome Finder service on the PATRIC website was used with the default parameters to find the other similar *Acidithiobacillus* genomes published and to calculate their Mash/MinHash distances with isolate JAGS [28]. For these genomes, average nucleotide identities based on BLAST (ANIb) and MUMmer (ANIm) were calculated in JSpeciesWS [29]. Genome alignment among four whole-genome sequences was achieved

using progressiveMauve within PATRIC [30]. Genes related to acid stress and metal resistance were analyzed using PATRIC and created with BioRender (https://biorender.com).

#### *2.3. Pan-Genome Analysis*

NCBI PGAP [31] was used to predict coding sequences for *A. ferridurans* JAGS and 8 other *Acidithiobacillus* genomes, and these amino acid sequences were used as the input for the Bacterial Pan-genome Analysis tool (BPGA ver. 1.2) to estimate core and pan genomes using the USEARCH program (ver. 9.0) available in BPGA, with a 50% cut-off of sequence identity [32]. The empirical power law equation f(n) = a × n <sup>α</sup> and the exponential equation f1(n) = c × e (d.n) were used for extrapolation of the pan and core genome curves, respectively. Core, accessory and unique genes defined in USEARCH were mapped into various cluster of orthologous group (COG) categories and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. The EggNOG (ver. 5.0.0) program [33] with default parameters was further used to cluster genes into functionally related groups and to analyze metabolic pathways.

#### *2.4. Prediction of Mobile Genetic Elements*

Insertion sequences (ISs) and transposases (Tn) distributed over the 9 *Acidithiobacillus* genomes were predicted and classified using the ISFinder platform with manual inspection of search hits (E-value ≤ 10−<sup>5</sup> ) [34]. IslandViewer (ver. 4), which has integrated the three most accurate and complementary genomic islands (GIs) prediction tools, IslandPath-DIMOB, SIGI-HMM and IslandPick [35], was applied for the computational identification of putative GIs. In addition, the web tool CRISPRFinder was mainly used to identify the Clustered Regularly Interspaced Short Palindromic Repeats-Cas protein (CRISPR-Cas) array.

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

#### *3.1. Phenotypic and Growth Features*

The genus of *Acidithiobacillus* is widely distributed in natural environments such as acid mine drainage (AMD) settings. We isolated an *Acidithiobacillus* strain from an AMD sample collected from Sudbury, Canada, and named it *A. ferridurans* JAGS in our previous study [23]. Orange-brown colonies of isolate JAGS formed on solid media, taking the shape of dots after 10 days of incubation (Figure 1A). The cells of isolate JAGS collected from 9K-Fe2+ liquid media showed single and paired rods, approximately 0.5–1.5 µm long and 0.3 µm wide (Figure 1B), which is slightly smaller than the reported *A. ferrooxidans* that is 1–2 µm long and 0.3–0.6 µm wide.

‐ *‐*

‐

‐ ‐

‐

‐

‐

‐

‐

‐ ‐

‐

*‐*

‐ **Figure 1.** Differentiation of *Acidithiobacillus ferridurans* JAGS from closely related species. (**A**) Colony morphologies of *A. ferridurans* JAGS observed under 40× optical microscopy for both top and bottom images. (**B**) Cell morphology observed under SEM. (**C**) Ferrous iron oxidation rates of *A. ferridurans* JAGS and *A. ferrooxidans* ATCC23270 under different pressures of Fe200 (200 mM Fe2+), Fe320 (320 mM Fe2+), Ni100 (100 mM Ni2+), Ni200 (200 mM Ni2+) and pH 1.5. \*\* indicates *p* < 0.01.

‐

‐

The species of *A. ferridurans* was reported to have a notably higher tolerance to many metals such as Fe2+, Ni2+ and Mg2+ and lower pH when compared with other *Acidithiobacillus* species [17,18]. In addition, there is 0.5–1% Ni existing in the pyrrhotite tailings of Sudbury [36]. Therefore, we compared the Fe2+ oxidation rate (%) between *A. ferridurans* JAGS and *A. ferrooxidans* ATCC23270 under elevated concentrations of Fe2+ and Ni2+ and low pH pressures (Figure 1C). The results showed that isolate JAGS had higher Fe2+ oxidation rates compared with ATCC23270 under high concentrations of Fe2+ and low pH values but similar levels of Fe2+ oxidation rates under high concentrations of Ni2+. These results indicated the adaptation of *A. ferridurans* JAGS to the acid mine drainage in Sudbury and suggested that it might be a great candidate as the ferrous oxidizer in low-pH bioleaching.

The growth features of *A. ferridurans* JAGS in culture media with 9K-Fe2+ and 9K-S<sup>0</sup> were investigated (Supplementary Material Figure S1). Figure S1A,B present a standard curve of OD<sup>600</sup> versus the cell count obtained by plate counting. When OD<sup>600</sup> = 1, we estimated that there were 8.8 × 10<sup>9</sup> cells/mL of *A. ferridurans* JAGS, which is slightly higher than the reported number of *A. ferrooxidans* (8.3 × 10<sup>9</sup> cells/mL) [37]. This may be due to the smaller cell size of isolate JAGS that causes it to absorb less light in the cuvette. We noted that counting cells of isolate JAGS was difficult due to its very low cell density in the lag phase and the interference of precipitate formation in the exponential phase. Therefore, monitoring growth required indirect tracking via changes in the pH and electron donor concentrations, but this was corroborated with data from OD<sup>600</sup> measurements and the plate count method. The growth behavior of isolate JAGS in 9K-Fe2+ (Figure S1C) and 9K-S0 culture media (Figure S1D) shared similarities with what was reported for *A. ferrooxidans* ATCC 23270 [38,39], which indicates a close relationship between the two species. However, the cell numbers examined by OD<sup>600</sup> were much lower than those obtained by the plate count method, which is likely due to the cells lost during the process of precipitate removal prior to OD<sup>600</sup> measurements.

#### *3.2. Genomic Features*

To better understand the isolated strain *A. ferridurans* JAGS, its genome was sequenced (GenBank: CP044411) and analyzed. The genome of *A. ferridurans* JAGS is a single circular chromosome comprising 2,933,811 bases with a GC content of 58.56%, which contains 3001 protein-coding sequences (CDSs), 46 tRNAs and 6 rRNAs [23]. The genomic features of *A. ferridurans* JAGS are quite similar to those of *A. ferridurans* JCM18981, which are 2,921,399 bases with 58.4% GC content, containing 3026 CDSs, 47 tRNAs and 6 rRNAs.

To explore the relationship of *A. ferridurans* JAGS with other *Acidithiobacillus* species, the Similar Genome Finder service from PATRIC was used to find similar genomes with isolate JAGS as the reference. There were eight *Acidithiobacillus* genomes found (Table 1). These strains were collected from different environments but mainly from acid mine waters. Their genomes varied in size from 2.7 to 3.2 Mb, with total CDS numbers ranging from 2634 to 3179. For the Mash/MinHash distances, *A. ferridurans* JCM18981 and *A. ferrooxidans* IO-2C showed the closest distance with *A. ferridurans* JAGS compared to the other six strains, which suggests that the IO-2C strain might be incorrectly classified. Based on the average nucleotide identity (ANI) relatedness analysis, it appears that the strains JAGS, JCM18981 and IO-2C are all *A. ferridurans* species as they shared ANI values >98% with each other (Table 1 and Table S1), which is larger than the reported threshold of ≥96% for classification [14].

Genome alignment among the complete-genome sequences of *A. ferridurans* JAGS, *A. ferridurans* JCM18981, *A. ferrooxidans* ATCC53993 and *A. ferrooxidans* ATCC23270 was performed using progressiveMauve [24] and is shown in Figure 2. Surprisingly, the genomic arrangement of *A. ferridurans* JAGS had better co-linearity with *A. ferrooxidans* ATCC53993 and *A. ferrooxidans* ATCC23270 than with *A. ferridurans* JCM18981. The samecolor blocks suggest high conservation of gene orders among multiple genomes that are likely inherited through vertical transfer, while *A. ferridurans* JCM18981 had a large

≥

number of gene rearrangements, insertions and/or deletions. This result indicates multiple recombination events and a rich evolutionary history of *A. ferridurans* species. ‐ ‐ ‐

**‐ ‐**

**Table 1.** General features and genomic comparison (pairwise mutation (Mash) distance, average nucleotide identity (ANI)) between *A. ferridurans* JAGS and selected representatives. ‐

‐


Note: Average nucleotide identity (ANI) based on BLAST+ (ANIb) and MUMmer (ANIm).

**Figure 2.** Whole-genome alignment of *A. ferridurans* JAGS, *A. ferridurans* JCM18981, *A. ferrooxidans* ATCC53993 and *A. ferrooxidans* ATCC23270. Locally collinear blocks (LCBs) identified by Mauve are color-coded; links between LCBs are indicated by the thin colored lines.

#### *3.3. Pan-Genome and Functional Gene Analysis*

Pan-genome analysis was carried out using the Bacterial Pan-genome Analysis (BPGA) tool to provide insights regarding genomic features, diversity and evolution [32]. It is well accepted that more than five genomes in a pan-genome analysis could provide suf-

ficient data for extrapolation of the information for species [40]. In this study, genomes of *A. ferridurans* JAGS and eight other strains of *A. ferridurans* and *A. ferrooxidans* were used for the pan-genome analysis since we wanted to investigate the genetic diversity and ecological adaption of these two species. As shown in Figure 3A,B, according to the Heaps' Law function (f(n) = 2819.13n0.22), the pan genome is open as the γ was calculated as 0.22, which means that the addition of new genomes will provide novel genes and it indicates evolutionary changes in these genomes [41]. The pan genome contains a total of 4601 genes, of which 1982 genes are in the core genome and 1064 genes are in the unique genome. The richness of unique genes in *A. ferridurans* JAGS, JCM18981, IO-2C and *A. ferrooxidans* ATCC23270 suggests that they may actively exchange genes with other genera to adapt to different environmental conditions.

‐ ‐ ‐ **Figure 3.** Pan-genome and EggNOG analysis. (**A**) Core-pan plot of studied *Acidithiobacillus* genomes; (**B**) Venn diagram of the pan genome; (**C**) pie charts of cluster of orthologous groups (COGs) of studied *A. ferridurans* and *A. ferrooxidans* strains and *E. coli* MG1655 (reference). Note: In Figure 3C, the numbers represent the percentage of each category. A, RNA processing and modification; B, Chromatin structure and dynamics; C, Energy production and conversion; D, Cell cycle control and mitosis; E, Amino acid metabolism and transport; F, Nucleotide metabolism and transport; G, Carbohydrate metabolism and transport; H, Coenzyme metabolism; I, Lipid metabolism; J, Translation; K, Transcription; L, Replication and repair; M, Cell wall/membrane/envelope biogenesis; N, Cell motility; O, Post-translational modification, protein turnover, chaperone functions; P, Inorganic ion transport and metabolism; Q, Secondary structure; T, Signal transduction; U, Intracellular trafficking and secretion; V, Defense mechanisms. The categories of "R, General functional prediction only" and "S, Function unknown" were omitted.

‐ The cluster of orthologous group (COG) distributions of the annotated genes for each studied *Acidithiobacillus* strain are illustrated in Figure 3C. These COGs fell into 20 COG classes, not including "General functional prediction only" and "Function unknown". *Escherichia coli* MG1655 was used as a reference, which showed that *E. coli* had the highest portion of genes corresponding to (K) Transcription, while *Acidithiobacillus* strains showed higher portions of functional genes related to (L) Replication, recombination and repair

‐

‐ ‐

‐

‐

‐

‐

‐ ‐

‐ ‐

(10.7–14%) and (M) Cell wall/membrane/envelope biogenesis (9.4–10.9%). This was not surprising given that these categories of proteins have been reported to be necessary for acid and heavy metal resistance and, likely, long-term adaptation mechanisms to extreme environments [42]. When the two *Acidithiobacillus* species were compared, *A. ferridurans* had more genes associated with functions supporting (C) Energy production and conversion (9.7–9.9%), while *A. ferrooxidans* had more genes related to the function of (P) Inorganic ion transport and metabolism (8.5–9.2%).

#### *3.4. Genetic Mechanisms of Acid Stress and Metal Resistance.*

In response to acidic heavy metal stress, acidophiles have developed different genetic mechanisms to survive and they are described, which can be very complex [43,44]. The metabolic diversity and adaptive mechanisms of *Acidithiobacillus* spp. responding to extremely acidic environments have been reviewed [45] and are beyond the scope of this study. Here, we focused on five major mechanisms (Figure 4A) for acid and heavy metal resistance in representative *Acidithiobacillus* strains by analyzing related gene clusters: (1) a membrane barrier created by outer membrane proteins (Omp40) and hopanoids; (2) maintenance of a membrane potential by influx of potassium and sequestration of metal ions intra-/extracellularly; (3) active removal by antiporters or exporters; (4) decarboxylation and detoxification; and (5) DNA and protein repair systems.

‐ ‐ ‐ **Figure 4.** Overview of adaptive strategies for acid and heavy metal resistance. Potential resistance mechanisms (**A**) and comparisons of operons for *kdp* (**B**), *mer* (**C**) and *ars* (**D**) genes among *Acidithiobacillus* strains. *Kdp*, a high-affinity K<sup>+</sup> transport system; *mer*, Hg2+-resistant genes; *ars*, As2+/3+-resistant genes.

‐

‐ ‐

*‐*

‐ ‐

‐ ‐

The mechanisms responsible for acid stress resistance are complex. We analyzed relevant genes in *A. ferridurans* JAGS and listed them in Table S2 (Supplementary Materials). OMP40 (Gene ID: F6A13\_00370) was discovered and it was previously reported as an anionic porin in the outer membrane to restrict the influx of protons in *A. ferrooxidans* [46]. A number of genes coding for hopanoid-synthesis proteins were identified in the isolate JAGS genome, including a glycosyltransferase (HpnB, Gene ID: F6A13\_01630) and a cluster for hopanoid-associated proteins, a squalene-hopene cyclase and a squalene synthase (HpnMHNKJIAG-Sch-Sqs, Gene IDs: F6A13\_09105–09160). Hopanoid is an important type of bacterial lipid that can alter membrane fluidity and permeability to restrict H<sup>+</sup> influx. It was reported to be synthesized from squalene by SHC, HpnG and other proteins enriched in the outer membrane by a transporter HpnN, although the functions of some factors are still unknown [47].

The Na+/H<sup>+</sup> antiporter (Gene IDs: F6A13\_09475, 04755) can export excess protons by coupling the uptake of Na<sup>+</sup> , while decarboxylases (SpeA, PanD and Psd; Gene IDs: F6A13\_06090, 14595, 04545) will consume excess protons in the cytoplasm. For most of the proteins, the interspecies identities ranged from 94% to 98%, while intraspecies identities were 100%. Interestingly, we found the *A. ferridurans* JAGS genome to process three *kdp* clusters (*kdpEDFABC*, Gene IDs: F6A13\_09590-09605; *kdpFABC*, Gene IDs: 11020–11005; kdpFA, GeneIDs: 11030–11025), similar to other *A. ferrooxidans* strains. However, *A. ferridurans* JCM18981 and IO-2C only had one *kdp* cluster (Figure 4B). The *kdp* clusters code for a high-affinity K<sup>+</sup> transport system, which could generate a reversed membrane potential through the active influx of K<sup>+</sup> to cope with acid resistance [48]. In Figure S2 (Supplementary Material), a neighbor-joining (NJ) phylogenetic tree is constructed based on the kdpA protein sequences. Even though *A. ferridurans* JAGS had a similar pattern of *kdp* clusters to *A. ferrooxidans* species (Figure 4B), its *kdpA* (I) showed higher sequence identity with the other *A. ferridurans* strains, suggesting that *kdp* clusters can be acquired more than once in these genomes and have redundancy.

We also tried to find genes involved in heavy metal resistance pathways in *A. ferridurans* JAGS (Figure 4 and Table S2). In total, eight genes (Gene IDs: F6A13\_04860, 04905, 04925, 04945, 08625, 10865, 10890, 11740) were predicted to code for p-type ATPases to transport Pb2+, Cd2+, Zn2+, Hg2+ and Cu2+ [49]. Similar numbers of these genes were detected in ATCC53993 (8) and JCM18981 (9), while only 5–6 genes were found in other strains. Some of these genes are highly similar, such as the genes F6A13\_10865 and F6A13\_04945, which indicates gene duplication within a genome. There are several other clusters belonging to the Resistance-nodulation-division (RND) transporter system, in the *czc* or *znu* families, which corresponds to Ni2+, Mn2+, Fe2+/3+, Mo2+, Co2+, Cd2+ and Zn2+ and CorAC for Mg2+/Co2+ export. We also noticed that these *Acidithiobacillus* strains possessed two types and several copies of *czcABC* clusters located at different sites, suggesting that *czc* clusters may be acquired more than once from different origins. Overall, this pattern of redundancy of resistance clusters in *Acidithiobacillus* suggests that it is part of the adaptive strategy for survival in acidic heavy metal conditions.

We also found genes coding for proteins related to mercury (Hg2+) and arsenic (As2+/3+) resistance. The operons we found for Hg2+ resistance in *Acidithiobacillus* strains are summarized in Figure 4C. There were two subgroups of *mer* clusters detected, *merRTPA* and *merCAD*. *A. ferridurans* JAGS and *A. ferrooxidans* CCM4253, Hel18, RVS1 and YQH-1 possessed both *mer* clusters, while *A. ferridurans* JCM18981 and IO-2C lacked the *merCAD* cluster. An intact *merRTPA* cluster was not detected in *A. ferrooxidans* ATCC23270 and ATCC53993, although ATCC53993 had two copies of *merR*. The *merA* in the *merCAD* cluster exhibited 100% identity in all selected strains, suggesting that *merCAD* might come from the same donor. Therefore, we further investigated the genetic context of the cluster *merCAD* and found it adjacent to *kdpCBAFAF* clusters. The DNA sequences of the *kdpCBAFAF– merCAD* clusters (11,906 bp) in *A. ferridurans* JAGS share 100% identity and 100% coverage with *A. ferrooxidans* ATCC53993, ATCC23270, RVS1 and CCM4253 and 100% identity with less coverage for *A. ferrooxidans* Hel18 and YQH-1, possibly due to incomplete sequencing. This indicates that the *kdpCBAFAF–merCAD* clusters might come from a same donor via a horizontal gene transfer (HGT) event [14]. Two predicted mobile element proteins (Gene IDs: F6A13\_10980, 10985) were found upstream of the *kdpCBAFAF–merCAD* clusters in isolate JAGS, supporting the HGT hypothesis.

Annotation of arsenic resistance clusters included *arsHBRCDA*, *arsHRBC*, *arsCDA*, *arsRC* and *arsM* in the studied genomes. *A. ferridurans* JAGS possessed all of these clusters, while the *arsCDA* and *arsRC* clusters were absent in *A. ferridurans* JCM18981 and the studied *A. ferrooxidans* strains. The largest abundance of *ars* clusters in isolate JAGS compared with other strains might contribute to its dominance in our metal-rich mine drainage sample. Besides, the analysis of *arsC* protein sequences further suggested that the *ars* clusters *arsHBRCDA*, *arsHBRC* and *arsRC* are likely acquired from different donors during evolution. Since it was also reported that gene copy number alterations can benefit microorganisms' survival under selective pressure [50], we hypothesize that *A. ferridurans* JAGS might have gained *ars* clusters from other species during adaption to the metal-rich environments.

In summary, compared to the other *Acidithiobacillus* strains that we studied, *A. ferridurans* JAGS had several genes and considerable redundancy that likely contributes to its acid and heavy metal resistance, which highlights its strong potential for usage in biomining processes, especially for cinnabar (HgS) or arsenopyrite (FeAsS) tailings. Additionally, we speculate that *A. ferridurans* JAGS might be an intermediate species between *A. ferrooxidans* and *A. ferridurans* based on the evidence of the gene cluster types and genomic structure (Figure 4) and the genome alignment result (Figure 2).

#### *3.5. Mobile Genetic Elements Analysis*

Mobile genetic elements (MGEs) play a great role in genome plasticity and evolution, shaping both genes and genomes to respond to drastic changes in environmental conditions [51]. MGEs, including insertion sequences (ISs) and genomic islands (GIs), are listed in Table 2. The number of ISs per strain ranged from 38 (*A. ferridurans* IO-2C) to 78 (*A. ferridurans* JAGS and *A. ferridurans* JCM18981), which might be due to the genome assembly level. High similarity regarding IS type was observed in all studied strains: IS1595, IS21, IS3, ISL3 and Tn families, which were the most common IS families. IS3 was the most abundant family. However, closer inspection demonstrated several differences. For instance, *A. ferridurans* JAGS has a much higher number of IS1595 when compared with other species. The nine genomes harbored 19–26 GIs ranging from 4 to 158 kb in size, representing many versatile gene pools. Several ISs, such as IS110 and IS66, were presented in the predicted GIs, suggesting that these putative GIs were likely acquired by horizontal gene transfer. In addition, GIs carrying mercury resistance genes (merRTPA) were found in all *A. ferridurans* species but not in the model strains *A. ferrooxidans* ATCC23270 and *A. ferrooxidans* ATCC53993, which might provide *A. ferridurans* with an adaptive advantage in mercury-rich environments.

Furthermore, we examined the clustered regularly interspaced short palindromic repeats (CRISPR) systems in all studied genomes using the CRISPRCasFinder [52]. Interestingly, *A. ferrooxidans* ATCC23270 was the only strain with a predicted CRISPR system. One unique type IV Cas cluster (csf4-1-2-3) and five spacers presented in the vicinity were found and presumed to function in conjunction with other CRISPR arrays [53].


**Table 2.** The prediction of mobile genetic elements including insertion sequences (ISs) and genomic islands (GIs) of the *Acidithiobacillus* strains studied.

Note: Since RVS1, Hel18 and YQH-1 have many contigs, ISFinder and IslandViewer cannot predict ISs and GIs exactly, respectively.

#### **4. Conclusions**

In the present study, we provide the growth characteristics and genomic insights of a newly isolated strain, *A. ferridurans* JAGS. The growth features of isolate JAGS in 9K-Fe2+ and 9K-S<sup>0</sup> liquid media are similar to the *A. ferrooxidans* type strain ATCC23270, but it shows a higher oxidation rate under elevated concentrations of Fe2+ and low pH. Genomic comparison and pan-genome analysis among nine strains of two species of *A. ferridurans* and *A. ferrooxidans* revealed obvious genetic differences between the two species. *A. ferridurans* JAGS showed a closer evolutionary relationship with other *A. ferridurans* species but a higher similarity of genomic structure with the *A. ferrooxidans* strains. This suggests that *A. ferridurans* JAGS might be an intermediary strain. Investigations of gene clusters (*kdp*, *mer* and *ars*) and mobile genetic elements indicated that there have been frequent gene transfers between their genomes during evolution. The high abundance of acid and metal resistance genes found in *A. ferridurans* JAGS points to its unique abilities to survive in harsh mining environments, which highlights its strong potential for applications in biomining processes. Further transcriptomic and proteomic analyses are required to find the exact genes, proteins and possible mechanisms that lead to the increased resistance of the isolate JAGS strain.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2075-1 63X/11/1/74/s1, Figure S1: Growth features of *A. ferridurans* JAGS. (**A**) Standard curve of optical density (OD600) versus cell numbers obtained by plate count method. (**B**) Colonies on 2:2 solid medium for cell count. (**C**) Growth on 9K-Fe2+ medium, (**C1**) pH value; (**C2**) Fe oxidation; (**C3**) cell numbers by OD600; (**C4**) cell numbers by plate count method. (**D**) Growth on 9K-S<sup>0</sup> medium, (**D1**) pH value; (**D2**) sulfur oxidation; (**D3**) cell count by OD600; (**D4**) cell numbers by plate count method. Figure S2: Neighbor-joining (NJ) phylogenetic tree of the kdpA protein sequences derived from nine *Acidithiobacillus* strains. Bootstrap values indicated at each node are based on a total of 500 bootstrap replicates. Table S1: Average nucleotide identity (ANI) (%) based on whole-genome alignments among *Acidithiobacillus* strains by JSpeciesWS. Table S2: Genes predicted to be involved in acid and heavy metal tolerance in *A. ferridurans* JAGS.

**Author Contributions:** Conceptualization, J.C. and R.M.; methodology, J.C. and Y.L.; software, formal analysis and writing—original draft preparation, J.C. and Y.L.; writing—review and editing, P.D. and R.M.; supervision, R.M.; project administration, R.M.; funding acquisition, R.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded through the Elements of Biomining Grant from the Province of Ontario through the ORF Research Excellence funding program.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** Vale is acknowledged for providing access to sample tailings at their Sudbury, ON, Canada, operations.

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

#### **References**


### *Article* **The Role of Mineral Assemblages in The Environmental Impact of Cu-Sulfide Deposits: A Case Study from Norway**

**Yulia Mun 1,\* , Sabina Strmi´c Palinkaš <sup>1</sup> and Kåre Kullerud <sup>2</sup>**


**Abstract:** Metallic mineral deposits represent natural geochemical anomalies of economically valuable commodities but, at the same time, their weathering may have negative environmental implications. Cu-sulfide mineral deposits have been recognized as deposits with a particularly large environmental footprint. However, different Cu deposits may result in significantly different environmental impacts, mostly depending on weathering conditions, but also on geological characteristics (mineralogy, geochemistry, host-rock lithology) of the Cu mineralization. This study presents new mineral and geochemical data from the Repparfjord Tectonic Window sediment-hosted Cu deposits and the Caledonian volcanogenic massive sulfides (VMS) deposits. The deposits share similar mineral features, with chalcopyrite and bornite as the main ore minerals, but they differ according to their trace element composition, gangue mineralogy, and host lithology. The studied sediment-hosted Cu deposits are depleted in most toxic metals and metalloids like Zn, As, Cd, and Hg, whereas the Røros Caledonian VMS mineralization brings elevated concentrations of Zn, Cd, In, Bi, As, and Cd. The conducted leaching experiments were set to simulate on-land and submarine weathering conditions. A high redox potential was confirmed as the main driving force in the destabilization of Cu-sulfides. Galvanic reactions may also contribute to the destabilization of minerals with low rest potentials, like sphalerite and pyrrhotite, even under near-neutral or slightly alkaline conditions. In addition, the presence of carbonates under near-neutral to slightly alkaline conditions may increase the reactivity of Cu sulfides and mobilize Cu, most likely as CuCO<sup>3</sup> (aq).

**Keywords:** Cu-sulfide ore; Nussir; Ulveryggen; Røros VMS deposit; leaching tests; submarine weathering conditions; on-land weathering conditions

#### **1. Introduction**

Copper is one of the most widely used mineral commodities in modern society, with a particular importance to electronics, electrical power generation, and the renewable energy sector, as well as in electric vehicle technologies [1–3]. In nature, Cu can be found in various types of mineral deposits, but in addition to the Cu-porphyry type (e.g., Chuquicamata, Chile [4]; El Teniente, Chile [5]; Ok Tedi, Papua New Guinea [6]), deposits of volcanogenic massive sulfides (VMS) and Cu-sediment hosted types represent the most important sources of Cu. Worldwide, 20 million tons (Mt) of copper was the total mine production in 2020 [7]. This number decreased from 24.5 Mt in 2019 due to COVID-19 lockdowns in April and May [8]. Chile remains the major copper producer (5.7 Mt) followed by Peru (2.2 Mt), China (1.7 Mt), DR Congo (1.3 Mt), and the US (1.2 Mt) [9]. A range from 60% to 75% of copper is mined from porphyry-copper deposits [10], 20% from sediment-hosted Cu deposits [11], and around 6% of Cu is mined from VMS deposits [12].

In this study, sediment-hosted Cu mineralization is represented by samples from the Nussir and Ulveryggen Cu deposits, from the Repparfjord Tectonic Window, while the Røros deposit, located within the Upper Allochthon of Scandinavian Caledonides, was

**Citation:** Mun, Y.; Strmi´c Palinkaš, S.; Kullerud, K. The Role of Mineral Assemblages in The Environmental Impact of Cu-Sulfide Deposits: A Case Study from Norway. *Minerals* **2021**, *11*, 627. https://doi.org/ 10.3390/min11060627

Academic Editors: Carlito Tabelin, Kyoungkeun Yoo and Jining Li

Received: 31 May 2021 Accepted: 9 June 2021 Published: 12 June 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/).

selected as representative of VMS mineralization (Figure 1). All three deposits are characterized by chalcopyrite and bornite as the main ore minerals, but they significantly differ in trace element composition, gangue mineralogy, and host lithology. The Ulveryggen Cu sediment-hosted deposit was mined in the period from 1972 to 1978/79, and tailings were deposited subaqueously in Repparfjorden. The Nussir deposit has not been mined yet, but there are plans for start-up mining of both the Nussir and Ulveryggen deposits in the near future [13]. The mine tailings from this operation are designated to be disposed of subaqueously in Repparfjorden as well. The Røros VMS deposit was mined from 1644 to 1977 [14], and similar to other historic VMS mines along the Scandinavian Caledonides, the mine waste material was disposed of on land, and still represents a significant environmental threat [15].

Mining activities may result in negative environmental impacts due to the accumulation of large quantities of mine waste material, generation of acid mine drainage (AMD), and dispersion of heavy metals in aquifers, streams, and marine sediments and soils. Copper has been recognized as a commodity with a particularly large environmental footprint (e.g., [16–18]); the environmental impact of Cu mines mostly depends on tailings disposal site conditions and the geological features of Cu mineralization, including mineral, geochemical, and host lithology characteristics (e.g., [19–22]).

AMD is a major problem associated with mineral deposits, in which Cu occurs in the form of sulfides (e.g., chalcopyrite, bornite, chalcocite, covellite) or if barren sulfides (e.g., pyrite, marcasite, pyrrhotite) represent asignificant component in the ore mineral assemblage (e.g., [23,24]). The consequences are often severe, leading to a lowering of the pH of contaminated aquifers, and release of metals and metalloids into the environment (e.g., [25–29]). Traditionally, tailings have been deposited in subaerial conditions, but several countries including Norway, practice subaqueous deposition [30]. Since subaqueous deposition in particular raises environmental concerns, we have tested the stability of representative Cu mineral assemblages from two of the most common types of Cu-sulfide deposits in Norway (sediment-hosted Cu deposits and VMS deposit) in a set of experiments that simulated subaerial and subaquatic weathering site conditions.

Kinetic leaching tests represent a powerful and relatively inexpensive tool to predict generation of AMD (e.g., [31]). They are designed to simulate sulfide-weathering processes in different physicochemical conditions. Kinetic leaching tests can be industrial or performed in the laboratory. Industrial tests are run in leaching columns, heaps, tanks, vats, dumps, large bins or drums [32]. They are placed in the field and subjected to meteoric waters, oxygen from the atmosphere, and changing temperature depending on the season. These tests can be conducted for several months to several years and are infrequently sampled for concentrations of dissolved metals and metalloids, sulfate, and changing pH and Eh parameters [33]. The tests can be accelerated by adding additional water [32].

However, more often the leaching tests are performed in miniature versions and run in laboratory size equipment—batch reactors, leaching columns, and humidity cells (e.g., [34–42]). The results are later extrapolated or mathematically modelled for larger volumes [33]. The tests are well-controlled and parameters such as water pH and Eh, metals and metalloids concentrations are continuously measured. The tests are often accelerated by increased temperature or the addition of hydrogen peroxide (e.g., [43]). The laboratory leaching tests also allow determination of an acid neutralizing capacity of gangue minerals and the acid producing potential of sulfides as well as to test remediation mechanisms, (e.g., [34,44–46]). However, many authors (e.g., [47]) argue that laboratory tests cannot be simply extrapolated to the field conditions. For example, a faster oxidation of pyrite and chalcopyrite from the Aitik site in Northern Sweden was observed in the laboratory compared to the field conditions [47].

The importance of characterization of ore parageneses and their host rocks was recognized as an important tool in the prediction of leaching of heavy metals from naturally contaminated rocks during anthropogenic activities e.g., underground constructions [31]. The tests are designed around primary ore mineralization to observe the oxidation potential

of major ore minerals, and aim to extrapolate the results of the study to apply to produced mine tailings. The reactivity of the tailings with the surrounding environment will be significantly higher due to particle size and an increase in surface energy.

Norway has a long shoreline, and the ore-bearing rocks are often subjected to weathering by seawater. In addition, Norway is one of few countries where subaquatic mine tailings deposition is permitted. In both cases, it is important to understand the role of salinity on weathering of sulfides. Therefore, during the experiments seawater was used together with meteoric water.

The aim of this study is to evaluate the potential environmental impact of the studied Cu mineralization, considering geological characteristics including mineralogy, geochemistry, as well as the main physicochemical features of subaerial and subaquatic disposal sites.

#### **2. Geological Settings**

#### *2.1. Sediment-Hosted Cu Deposits of Nussir and Ulveryggen, Repparfjord Tectonic Window*

The Repparfjord Tectonic Window, Northern Norway, is composed of mafic metavolcanics and carbonate-siliciclastic sequences that were compressed in a SE-NW direction during the Svecofennian Orogeny at ca. 1.84 Ga [48,49]. The rocks are metamorphosed under greenschist to lower amphibolite facies conditions, and [50] determined the age of host volcanics to be about 2.1 Ga, with Nussir mineralization around 1.765 Ga. The Repparfjord Tectonic Window contains numerous sites with Cu mineralization (e.g., [14,51]), of which the Nussir (26.7 Mt at 1.13% Cu) and the Ulveryggen (7.7 Mt at 0.81% Cu) deposits have the greatest implications for the local environment [14,52]. The Nussir deposit is hosted by a thin (no more than 5 m thick) metadolostone layer that can be traced for several kilometres (Figure 1A,C), intercalated with metasandstone, metasiltsone and metapelites. The metasedimentary complex is overlain by a several hundred meters thick metavolcanic sequence [49,50,53]. Despite the close geographical occurrence of the Nussir and Ulveryggen deposits, they have different lithologies. The Ulveryggen deposit is hosted by arkosic metasandstones, metasiltstones, and metaconglomerates with low carbonate content. Mineralization can be traced for more than 1 km in the northeast direction and is structurally confined to tight folds [48,49,54,55]. Chalcopyrite, bornite, sphalerite, and minor pyrite are the main ore minerals at both deposits.

The Nussir mineralization is confined to a thin (up to 5 m thick) dolomitic marble layer intercalated with metasiltstone, metapelites and metasandstones, localizing the mineralization within quartz-carbonate veins as well as disseminated in mafic metavolcanics [49,50,54,55]. The major ore minerals at the Nussir and Ulveryggen deposits are chalcopyrite, bornite, and chalcocite [49,50,54].

**Figure 1.** Geological maps showing the locations of the mines in (**A**) the Repparfjord Tectonic Window: Nussir and Ulveryggen sediment-hosted Cu deposits, and (**B**) the Røros area, drawn based on interactive online maps at [56] as well as modified after [49,55]. The map of Norway in (**C**) shows the locations of the Nussir, and Røros areas.

#### *2.2. Volcanogenic Massive Sulfide (VMS) Deposit Røros, the Upper Allochthon of Scandinavian Caledonides*

The Røros VMS mining area includes several hundreds of mineralizations in the southeastern part of the Upper Allochthon of the Scandinavian Caledonides [57], including the Røros deposit studied herein. The mineralizations are characterized by predominately chalcopyrite, sphalerite, pyrite, and galena hosted by interbedded metatuffite, metagraywacke, and gabbroic sills and dykes [14,58,59].

The Upper Allochthon of Scandinavian Caledonides extends for about 1500 km, from the Stavanger region in southern Norway to the Barents Sea region in northern Norway. This first-order tectonostratigraphic unit is dominated by sedimentary and magmatic rocks derived from the Iapetus Ocean, including ophiolite and island-arc complexes usually associated with VMS type mineralization (Figure 1B,C; [60–62]).

The Røros mining area in Trøndelag County, south-eastern Norway, hosts numerous occurrences of the VMS type (Figure 1B). Among the largest occurrences are found at the Storwartz and Olav mines in the eastern sector, and the Kongens mine in the

north-western sector (Figure 1), with average Cu and Zn contents of about 2.7% and 4.2–5%, respectively [59]. The mineralization is hosted by metagraywacke interbedded with metatuffites, metabasalts and gabbroic sills and dykes [14,58,59], with major ore minerals: chalcopyrite, pyrrhotite, sphalerite, and pyrite.

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

For this study, representative samples of the Cu sulfide mineralization were selected from the Nussir and Ulveryggen sediment-hosted Cu deposits, and from the Røros VMS deposit. Two main types of samples were analyzed: (1) Mineral assemblages composed of ore and gangue minerals and (2) Individual Cu-sulfide minerals.

Twenty-seven polished thin sections of representative mineral assemblages were prepared at the Department of Geosciences at UiT The Arctic University of Norway. Three thin sections represented reference samples for the respective deposits. The reference samples were studied under a reflective polarizing light microscope and a Zeiss Merlin Compact Field Emission Scanning Electron Microscope (FE-SEM) equipped with an energy-dispersive X-ray spectrometer (EDS) at UiT The Arctic University of Norway, to determine mineral and geochemical characteristics of the ore mineral assemblages prior to and after the leaching tests. In order to investigate the primary ore, the SEM was run in a high vacuum regime at 20 kV accelerating voltage, 20 s counting time, and with an aperture of 60 µm.

In order to simulate weathering processes corresponding to the tailings disposal site conditions, a set of leaching experiments were performed on the polished thin sections (Figure 2). The experiments were designed to simulate (Figure 3): (1) Subaquatic vs. onland disposal site conditions; (2) Oxidative vs. reductive conditions; (3) Carbonate buffered vs. carbonate free systems; and (4) Seawater infiltrated vs. meteoric water infiltrated sites (Table 1). Each thin section was placed in an individual beaker of 400 mL with a height of 10.5 cm and diameter of 8.5 cm. The thin sections were then covered with a 4 cm thick layer of quartz sand (200 g) for the simulations of on-land conditions, whereas natural marine sediments (200 g) were used for the simulations of submarine conditions.

**Figure 2. (A**) Photograph of the test setup; (**B**) backscattered electron image of chalcopyrite (Cpy) in assemblage with bornite (Bn) and inclusions of iron oxides (FeO) from unaltered reference sample E-N-1 (Nussir); (**C**) backscattered electron image of chalcopyrite (Cpy) with small inclusions of bornite (Bn) from unaltered reference sample E-U-1 (Ulveryggen); (**D**) backscattered electron image of the Røros reference sample E-R-1 showing the mineral assemblage of chalcopyrite (Cpy), sphalerite (Sph), pyrite (Py), and pyrrhotite (Po).

**Figure 3.** Schematic presentation of test settings #1–#8 conditions. Eh and pH are measured in sediments before the tests. See text for explanation.

The average organic content in the natural marine sediments was 0.82 wt.% (Supplementary Table S1), whereas the quartz sand was free of organic matter. To test the influence of redox potential, half of the beakers with quartz sand were doped with ~10 wt.% of organic matter and sealed with parafilm tape to prevent oxidation reactions. The other half of the beakers were left open during the entire experiment and refilled with circa 150 mL of meteoric water once per day to ensure oxidative conditions. Pure calcium carbonate was used to buffer relevant solutions. The experiments were run for three months, and in order to accelerate the reactions, the beakers were kept in a water bath at 50 ◦C.

At the end of the experiments, the samples were carefully removed and investigated under a reflective light microscope. The formation of secondary minerals was studied by Raman spectroscopy, conducted at UiT The Arctic University of Norway in Tromsø. A Renishaw inVia confocal Raman microscope equipped with a 532 nm (green) diode laser was used to identify the mineral phases in the studied ore samples, as well as the degree of weathering after simulation of weathering conditions under on-land and subaquatic conditions. The identification was based on Raman spectra published in the literature [63].

Individual grains of Cu sulfides were handpicked under a binocular microscope, washed in an ultrasonic bath and pulverized in an agate mortar. An amount of 0.5 g was analyzed for bulk trace element composition at the AcmeLabs (Vancouver, B.C. Canada), on an ICP MS instrument following the internal LF202 analysis code.


**Table 1.** Experimental setups. The cross mark corresponds to the ingredient added to the composition of the mixture. Sample last digits correspond to the experiment setup number.

<sup>1</sup> NS—Nussir; <sup>2</sup> Ulv—Ulveryggen; <sup>3</sup> RS—Røros, "+"—present in the experiment

#### **4. Results**

*4.1. Mineral Analyses*

4.1.1. Nussir and Ulveryggen

Mineral analyses show that typical ore assemblages from the Nussir and Ulveryggen sediment hosted Cu deposits consist of chalcopyrite, bornite, and chalcocite (Figure 2B,C). The Røros VMS mineralization is represented by massive sulfide bodies predominantly composed of chalcopyrite, sphalerite, pyrite and pyrrhotite (Figure 2D).

The Nussir dolomitic marble contains rhomboidal-shaped fragments of carbonates that are up to 5 mm in diameter. The mineralization is confined to crosscutting quartz-carbonate veins with euhedral to subhedral grains of vein carbonate, which are up to 0.1–0.3 mm in diameter. Quartz grains are anhedral and about 0.1 mm in diameter. Chalcopyrite is the dominant Cu mineral together with bornite; the minerals are intergrown and contain abundant inclusions of pyrite and sphalerite (Figure 2B).

The Ulveryggen arkosic metasandstone contains fragments of quartz and feldspar up to 0.2 mm in size. The ore minerals are disseminated and they have grown interstitially between fragments of quartz and feldspar together with muscovite. The main ore minerals are bornite and chalcopyrite (Figure 2C), with minor amounts of pyrite and chalcopyrite.

#### 4.1.2. Røros

Since the Røros samples were prepared from pieces of massive ore, only ore minerals were observed under the microscope. The main ore minerals are pyrrhotite, pyrite,

chalcopyrite and sphalerite, which show various intergrowth textures. Chalcopyrite, which is the most abundant mineral, occurs as individual grains that are several centimetres in diameter. Pyrite and pyrrhotite crystals also show large grain-sizes (several centimetres in diameter), while sphalerite forms small inclusions of less than 0.1 mm in diameter (Figure 2D).

#### *4.2. Leaching Tests*

Experimental conditions #1 (Marine sediments, TOC = 0.82 wt.%; carbonate buffered; infiltrated with seawater (Figure 3); Ehsed = 239.7 mV, pHsed = 7.64, where Ehsed and pHsed are values of redox potential and pH for pore water in sediments measured after initial stabilization of conditions, i.e. 60 h after the experiment started) did not affect stability of mineral assemblages from the Nussir and Ulveryggen deposits (Figures 4A and 5A). However, for the sample from the Røros VMS deposit, small grains of sphalerite were weathered while some pyrite grains remained well-preserved (Figure 6A; Supplementary Figures S3H and S4A,B).

Experimental conditions #2 (Marine sediments, TOC = 0.82 wt.%; carbonate buffered; infiltrated with meteoric water; Ehsed = 247.9 mV, pHsed = 7.31; Figure 3) resulted in partial oxidation of sulfides from the Nussir and Ulveryggen deposits (Figures 4 and 5B; Supplementary Figures S1A–C and S2B). As for the experimental conditions #1, small grains of sphalerite from the Røros VMS deposit were affected (Figure 6B; Supplementary Figure S4D). Pyrite was partly oxidized along rims, while other parts remained unaltered (Figure 6B; Supplementary Figure S4C–E). Pyrrhotite was weathered significantly (Supplementary Figure S4D), and chalcocite was weathered while chalcopyrite remained unaltered (Supplementary Figure S4C,E).

Experimental conditions #3 (Marine sediments, TOC = 0.82 wt.%; no added carbonates; infiltrated with seawater; Ehsed = 211.1 mV, pHsed = 7.29; Figure 3) did not affect the samples from the Nussir and Ulveryggen deposits (Figures 4C and 5C; Supplementary Figure S2D), but sulfides from the Røros VMS deposit went through extensive oxidation reactions (Figure 6C; Supplementary Figure S4F,G). Pyrrhotite was oxidized significantly, however chalcopyrite crosscutting pyrrhotite remained well-preserved (Figure 6C). Pyrite was partly oxidized along the rims. Covellite was observed along cracks in the pyrite.

Experimental conditions #4 (Marine sediments, TOC = 0.82 wt.%; no added carbonates; infiltrated with meteoric water; Ehsed = 231.2 mV, pHsed = 7.63; Figure 3) partly affected the Nussir and Ulveryggen samples: fine-grained fragments were significantly oxidized (Figures 4D and 5D; Supplementary Figures S1D–F and S2E), while larger grains were partly oxidized. Chalcopyrite obtained oxidized rims while bornite remained well-preserved (Supplementary Figures S1D–F and S2E). Chalcocite from the Nussir deposit was partly oxidized, while chalcocite in contact with pyrite from Ulveryggen remained well-preserved (Supplementary Figure S2E). The Røros samples were weathered significantly (Figure 6D; Supplementary Figure S4H). Notably, bornite was weathered to a higher degree than chalcopyrite. The latter was found in veins within bornite and remained well-preserved (Supplementary Figure S4H). Oxidized rims were formed around pyrite grains, whereas pyrrhotite and sphalerite had undergone significant weathering (Figure 6F).

Experimental conditions #5 (quartz sand, TOC = 0 wt.%; carbonate buffered, infiltrated with meteoric water; Ehsed = 222.6 mV, pHsed = 8.76; Figure 3) affected the sulfides from all three studied deposits to different degrees. Chalcopyrite and bornite from Nussir were extensively weathered (Figure 4E; Supplementary Figure S1G), whereas for the Ulveryggen sample, chalcopyrite and chalcocite were partly weathered while most of the pyrite and bornite remained well-preserved (Figure 5E; Supplementary Figures S2F–H and S3A). Sulfides from Røros were more oxidized in comparison to sulfides from Nussir and Ulveryggen. In contrast to pyrite and chalcopyrite, which remained well-preserved or only partly oxidized, pyrrhotite and sphalerite were significantly weathered (Figure 6E; Supplementary Figure S5A–C).

**Figure 4.** Microphotographs of samples under the reflected light microscope from the Nussir test runs after 90 days of the weathering experiment. The photographs are taken in crossed polars and correspond to test setups from 1 to 8 in Table 1. (**A**) reductive condition #1: well-preserved chalcopyrite (Cpy) grain (sample ENV-NS-1); (**B**) reductive condition #2: secondary minerals (SM) on top of chalcopyrite (Cpy) (sample ENV-NS-2d); (**C**) reductive condition #3: well-preserved chalcopyrite, secondary minerals are possibly forming in the cavities in the grain (Sample ENV-NS-3); (**D**) reductive condition #4: secondary minerals formed on the surface of chalcopyrite (Cpy, sample ENV-NS-4); (**E**) oxidative condition: secondary minerals on the surface of Cpy (ENV-NS-5); (**F**) oxidative condition #6: Cpy grain is partly oxidized (ENV-NS-6); (**G**) oxidative condition #7: wellpreserved assemblage of Cpy and bornite (Bn) (ENV-NS-7); (**H**) oxidative condition #8: intensively oxidized chalcopyrite grain (#ENV-NS-8).

Experimental conditions #6 (quartz sand, TOC = 0 wt.%; no added carbonates; infiltrated with meteoric water; Ehsed = 245.7 mV, pHsed = 7.54; Figure 3) resulted in extensive oxidation of the sulfides from the Nussir and Ulveryggen deposits. Chalcopyrite, bornite and chalcocite obtained a weathered rim around the grains (Figures 4F and 5F; Supplementary Figures S1H and S3B,C). Some grains of chalcopyrite were entirely covered with a thin film of weathering products (Figure 4F). Pyrite from Ulveryggen remained wellpreserved, but might have accelerated the oxidation of bornite (Supplementary Figure S3C). Pyrrhotite and sphalerite from the Røros deposit were significantly weathered, whereas pyrite remained relatively well-preserved with insignificant formation of iron oxides (Figure 6F; Supplementary Figure S5D). A weathered Cu-containing mineral was also observed within the pyrite (Supplementary Figure S5D).

**Figure 5.** Microphotographs of Ulveryggen samples after 90 days of experiments. The photographs are taken under a reflected light microscope, with crossed polarizers. Setup description is given in Table 1, (**A**–**H**) microphotographs correspond to #1–8 conditions. (**A**) Intergrowth of chalcocite (Cct) with bornite (Bn, ENV-Ulv-1); (**B**) well-preserved chalcopyrite (Cpy) with minor oxidation of fine grains (ENV-Ulv-2); (**C**) well-preserved chalcopyrite grains (ENV-Ulv-3); (**D**) unaltered chalcopyrite grains (ENV-Ulv-4); (**E**) partly oxidized chalcopyrite grain with secondary mineral and oxidation cover on the surface; (**F**) oxidized Cu sulfide, likely chalcopyrite with secondary minerals formed on the lateral parts; (**G**) well-preserved chalcopyrite grain; and (**H**) micro-assemblage of well-preserved bornite (Bn) with chalcopyrite.

**Figure 6.** Microphotographs of polished thin sections from the Røros deposit after the 90-day experiment. The photographs (**A**–**H**) are taken under crossed poles, reflected light microscope and correspond to test conditions #1–8 (Table 1). (**A**,**B**) well-preserved pyrite grains (Py) with partly oxidized sphalerite (Sph) triangles (blue); (**C**) partly oxidized pyrite with secondary minerals formed on the rim of Cu-sulfides; (**D**) oxidation of sulfide mineral assemblages: pyrite (Py), sphalerite (Sph), pyrrhotite (Po), and secondary minerals likely formed on Cu-sulfides; (**E**) Intensive oxidation and the formation of secondary minerals (Cu-ox) of chalcopyrite (Cpy); (**F**) relatively well-preserved pyrite and intensely oxidized sphalerite, with minor amount of iron oxides (Fe-ox); (**G**,**H**) micro-assemblages of well-preserved sphalerite (Sph), chalcopyrite (Cpy), pyrite (Py), and pyrrhotite (Po).

Experimental conditions #7 (quartz sand, TOC ≈ 10 wt.%, carbonate buffered, infiltrated with meteoric water; initial Ehsed = 191.5 mV, pHsed = 7.86; Figure 3) did not affect the Nussir and Ulveryggen sulfides. The minerals remained well-preserved (Figure 4G and Figure 5G; Supplementary Figure S3D,E). The Røros ore minerals also remained almost unaffected, however some pyrite grains were slightly tarnished with bright blue secondary covellite and brownish iron hydroxides (Figure 6G; Supplementary Figure S5E,F).

Experimental conditions #8 (quartz sand, TOC ≈ 10 wt.%, no added carbonates, infiltrated with meteoric water; Ehsed = 286.1 mV, pHsed = 5.43; Figure 3) resulted in well-preserved sulfides from the Ulveryggen and Røros deposits (Figures 5H and 6H; Supplementary Figures S3F,G and S5G). Chalcopyrite from Nussir was observed both as relatively well-preserved grains (Figure 4H) and significantly weathered grains (Supplementary Figure S2B).

#### *4.3. Raman Spectroscopy*

Raman spectra (Figure 7) obtained from chalcopyrite and pyrite from the Nussir, Ulveryggen, and Røros deposits suggest formation of new peaks 450 and 500 cm−<sup>1</sup> after exposing samples from the Nussir and Ulveryggen to experimental conditions #2 and #5, respectively (Figure 7A–D).

**Figure 7.** Raman spectrometry of chalcopyrite (Cpy) and pyrite (Py) from Nussir, Ulveryggen and Røros after 90-days of testing in selected conditions. Sample numbers correspond to condition numbers and can be found in Table 1. Cps—counts per second. **A**–**E**: right images are microphotographs of samples under reflected light microscope; left diagrams relate to Raman spot analyses. **A**,**B**—chalcopyrite (Cpy) from the Nussir deposit (sample Env-NS-2, experimental condition #2); **C**,**D**—chalcopyrite (Cpy) from the Ulveryggen deposit (sample Env-Ulv-5, experimental condition #5); **E**—pyrite (Py) and **F**—chalcopyrite (Cpy) from the Røros deposit (sample Env-RS-1, experimental condition #1).

#### *4.4. Mineral Chemistry*

In addition, to determine the concentration of potentially toxic elements including Cu, Zn, Ni, Hg, Cd, and As (e.g., [64]), the bulk chemical compositions of hand-picked Nussir, Røros chalcopyrite as well as Ulveryggen bornite were analyzed (Table 2). The Nussir and Ulveryggen results are taken from [53].

**Table 2.** Lithogeochemistry of hand-picked chalcopyrite from the Nussir deposit (NS-35-ccp), bornite from the Ulveryggen deposit (Ulv-2-bn) and chalcopyrite Røros Mine (RSL-ccp).


<sup>1</sup> NS—Nussir, <sup>2</sup> Ulv—Ulveryggen, <sup>3</sup> RSL—Røros; Bn—bornite, ccp—chalcopyrite, LD—limit of detection, LLD lower than limit of detection, ND—no data.

#### 4.4.1. Nussir and Ulveryggen

The Nussir chalcopyrite contains 100 ppm of Ni, 85 ppm of Co and 310 ppm of Zn. The content of Bi is 0.3 ppm, while As, Mo, Cd, Sb, Pb, and Hg contents are minor or below the detection limit. The Nussir chalcopyrite also contains Ba (12 ppm), most likely in the form of nearly insoluble inclusions of barite.

Ulveryggen bornite contains 2 ppm of Co and 5 ppm of Mo. Nickel, Zn, As, Sn, Sb, Cd, Pb, and Hg contents are minor or below the detection limit, while the Rb content is 24 ppm. The content of Ba in the Ulveryggen bornite is 527 ppm.

#### 4.4.2. Røros

Chalcopyrite was picked from the crushed Røros sample. The Co content is 535.9 ppm. The chalcopyrite also contains: Ni 7.8 ppm and Zn 2942 ppm. As content is 226.1 ppm, and Sn and Sb contents are 8 and 1.8 ppm respectively. Chalcopyrite contains 30 ppm of Cd and 18 ppm of Ba. The Bi content is 9.9 ppm and Hg content is 4.59 ppm.

#### **5. Discussion**

The Cu mineralization found in the Nussir and Ulveryggen sediment-hosted Cu deposits is characterized by predomination of chalcopyrite, bornite and chalcocite. Mine tailings from these Cu-sulfide deposits are associated with a high risk for the generation of acid mine drainage (AMD) because of the high Fe2+/Fe3+ and S2−/SO<sup>4</sup> <sup>2</sup><sup>−</sup> ratios in their mineral assemblages. Furthermore, the Ulveryggen deposit has a low carbonate content that additionally increases the risk. In contrast, a low content of potentially toxic elements such as As, Cd, Hg, and Zn in both, reduces their environmental threat [53].

The Røros Cu-Zn VMS deposit, similar to other VMS deposits worldwide [65,66], is characterized by a polymetallic composition (Table 2). The main ore minerals are sulfides, including chalcopyrite, bornite, pyrite, and pyrrhotite. High Fe2+ and S2<sup>−</sup> contents together

with an absence of carbonates from its mineral assemblages point to a high risk for the generation of AMD in this deposit (e.g., [23,27,46,67,68]). In addition, the enrichment in a wide spectrum of potentially toxic metals and metalloids (e.g., As, Bi, Cd, In, and Zn) magnifies the environmental risk associated with mining activities and/or processes of natural weathering in this type of ore deposits (e.g., Rio Tinto VMS deposits, the Iberian Pyrite Belt in Spain [26,28,69–71] and the Britannia Creek VMS deposit, Canada [48]).

Leaching tests are recognized as powerful tools to predict the behavior of sulfides in different conditions (e.g., [20,32,37,39,40]). Such tests play an important role in the initial phases of mining planning, while deciding the potential placement of tailings for future storage. The tests are usually performed in batch reactors specially equipped with sensors controlling temperature, pH, and the amount of dissolved oxygen (e.g., [39,47,72]). Otherwise, the tests can be performed in leaching columns (e.g., [32,73]) or in static conditions (e.g., [46,74]) for periods of several days to several years. The leaching experiments in this study were designed to test the stability of ore mineral parageneses from three different Cu deposits under diverse physicochemical conditions (Figure 3, Table 1), and predict the behavior of ore-bearing mineral assemblages disposed in on-land and submarine conditions. As expected, the high redox potential was the main driving factor in destabilization of sulfides (e.g., [75]):

$$4\text{ FeS}\_2 + 15\text{ O}\_2 + 14\text{ H}\_2\text{O} \leftrightarrow 4\text{ Fe(OH)}\_3 + 16\text{ H} + 8\text{ SO}\_4^{2-} \tag{1}$$

$$4\text{ FeS} + 10\text{ O}\_2 + 9\text{ H}\_2\text{O} \leftrightarrow 4\text{ Fe(OH)}\_3 + 8\text{ H} + 4\text{ SO}\_4^{2-} \tag{2}$$

$$4\text{ CuFeS}\_2 + 15\text{ O}\_2 + 14\text{ H}\_2\text{O} \leftrightarrow 4\text{ Cu}^{2+} + 4\text{ Fe(OH)}\_3 + 16\text{ H}^+ + 8\text{ SO}\_4^{2-} \tag{3}$$

The organic matter content of the natural marine sediments that were used (0.82 wt.%) was sufficient to prevent oxidation of the sulfides in experimental conditions in which the sulfide parageneses were exposed to seawater. In the setups with meteoric water, a different scenario was observed. A lower solubility of oxygen in seawater (4.6 mg/L at 50 ◦C) compared to meteoric water (5.6 mg/L at 50 ◦C) was probably one of the controlling factors [76]. When sediments were doped with an additional 10 wt.% of organic matter, the differences between seawater and meteoric water influence were not recorded.

The carbonate buffered conditions were mostly less reactive due to a slightly alkaline pH value of the infiltrating aqueous solutions. However, the samples from the Røros VMS deposit revealed that sphalerite can enter galvanic reactions and be extensively dissolved even in alkaline or near-neutral conditions. The prerequisite is that sphalerite (−0.24 V; [77]) occurs in direct contact with sulfides with greater rest potentials like pyrite (0.63 V; [78]) or chalcopyrite (0.54 V; [79]). In interactions with oxygen-rich solutions, sphalerite will act as an anode:

$$\text{ZnS} \leftrightarrow \text{Zn}^{2+} + \text{S}^{0} + \text{2e}^{-} \tag{4}$$

and prevent oxidation of pyrite, and Cu-sulfides reacting with oxygen adsorbed on their surface:

$$\frac{1}{2}\text{ O}\_2\text{ (aq)} + 2\text{H}^+ + 2\text{e}^- \leftrightarrow \text{H}\_2\text{O}\tag{5}$$

This reaction will not affect the pH of the aquifer, but it will promote leaching of Zn.

The galvanic reaction may represent a particularly large environmental issue in mineral deposits in which pyrrhotite (−0.24 V; [77]) is intergrown or occurs in direct contact with Cu-sulfides and/or pyrite:

$$\text{FeS} \leftrightarrow \text{Fe}^{2+} + \text{S}^{0} + 2\text{e}^{-} \tag{6a}$$

$$\text{Fe}^{2+} + 3\text{H}\_2\text{O} \leftrightarrow \text{Fe(OH)}\_3 + 3\text{H}^+ + \text{e}^- \tag{6b}$$

The anode reaction will result in dissolution of pyrrhotite, oxidation of ferrous to ferric ions, and consequently in acidification of the system.

In the experimental setups buffered with carbonates, the Cu sulfides showed an increased reactivity (Figure 4B,E, Figure 5B,E and Figure 6E) whereas pyrite did not show significant changes under these conditions (Figure 6B). Carbonates are often used for prevention of acid mine drainage (e.g., [80]), but the results of this study show that in near-neutral to slightly alkaline conditions Cu can be mobilized from sulfides, most likely in the form of CuCO<sup>3</sup> (aq) [53,81]. This reaction is more intensive in solutions with a higher redox potential (Figures 4E, 5E and 6E).

In addition, gangue mineralogy might also play a role in the rate and degree of oxidation. The Nussir mineralogy is hosted by dolomitic marble, while Ulveryggen Cu sulfides are hosted by arkosic sandstone, which retards the oxidation of sulfides [53]. Røros sulfides, studied here, were sampled from a massive ore with an absence of gangue minerals. Therefore, galvanic interaction was favoured to higher degree of oxidation and was not prevented or retarded by gangue mineralogy. Weathering in mine tailings will be accelerated due to higher surface area, however this is also true for gangue mineralogy which in the case of Nussir and Ulveryggen might play a buffering role. On the other hand Røros mineralization being hosted by massive mafic volcanic rocks, which are extremely soluble, will most probably not retard the oxidation reaction or buffer it to limited degree.

Raman spectroscopy confirmed the slight distortion of crystal lattices of pyrite and chalcopyrite (Figure 7), however, the signal was low. Optically, even when the blue tinge was observed in the Nussir and Ulveryggen samples, Raman analyses did not record any changes in crystal structure of sulfides from these two deposits. This is attributed to oxidation occurring in a thin layer of secondary minerals, which tends to be amorphous and is characterized by the absence of a detectable lattice. In addition, the signal from primary minerals is significantly higher.

#### **6. Conclusions**

The mineral assemblages from all three studied deposits point to a high risk for generation of acid mine drainage due their high Fe2+/Fe3+ and S2−/SO<sup>4</sup> <sup>2</sup><sup>−</sup> ratios. However, different ore-forming conditions of sediment hosted Cu deposits (Nussir and Ulveryggen) from conditions related to the formation of VMS deposits (Røros) resulted in contrasting behavior of the trace elements. As a consequence, the Nussir and Ulveryggen deposits are depleted in most potentially toxic elements, such as As, Cd, Hg, and Zn, whereas the Røros VMS mineralization is enriched in a wide spectrum of potentially toxic metals and metalloids, including As, Bi, Cd, In, and Zn.

The leaching experiments revealed the redox potential as the main factor that controls the stabilities of Cu-sulfides for both on-land as well as submarine conditions. Galvanic reactions may contribute to the destabilization of minerals with low rest potentials, like sphalerite and pyrrhotite, even under near-neutral or slightly alkaline conditions. The destabilization of pyrrhotite can have particularly negative environmental consequences due to the release of ferrous ions to an aquifer and acidification of the system as a result of oxidation of ferrous to ferric ions.

Although carbonates are often used for prevention of acid mine drainage, the presence of carbonates under near-neutral to slightly alkaline conditions may increase the reactivity of Cu sulfides and mobilize Cu, most likely in the form of CuCO<sup>3</sup> (aq).

More complex ore minerals assemblages lead to deeper weathering in given conditions. Thus, Røros chalcopyrite were notably more altered than chalcopyrite from the Ulveryggen and Nussir deposits.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/min11060627/s1, Figure S1: Microphotographs demonstrating the Nussir sulfides reaction after 90-day tests; Figure S2: Microphotographs taken under a reflected light microscope; Figure S3: Microphotographs taken under reflected light; Figure S4: Microphotographs of Røros samples under reflected light; Figure S5: Microphotographs of Røros samples under the reflected light microscope after 90-days of experimental tests; Table S1: Total organic carbon (TOC) of gravity core HH-12-002-MF-GC obtained from Repparfjord.

**Author Contributions:** Conceptualization, Y.M., S.S.P.; methodology, Y.M. and S.S.P.; validation, Y.M. and S.S.P.; formal analysis, Y.M. and S.S.P.; investigation, Y.M. and S.S.P.; resources, Y.M., S.S.P. and K.K.; data curation, Y.M., S.S.P. and K.K.; writing—original draft preparation, Y.M.; writing—review and editing, Y.M., S.S.P. and K.K.; visualization, Y.M. and K.K.; supervision, S.S.P. and K.K.; project administration, S.S.P. and K.K.; funding acquisition, K.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Tromsø Fylkes Kommune and SINTEF through the PhD project of the first author, grant number RDA12/167. The publication charges for this article have been funded by a grant from the publication fund of UiT The Arctic University of Norway.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We are thankful to Carlito Tabelin, the Academic Editor of the special issue "Novel and Emerging Strategies for Sustainable Mine Tailings and Acid Mine Drainage Management" for constructive comments and suggestions that significantly improved the text. The authors are thankful to Kai Neufeld for the assistance with the SEM work. The laboratory staff of the Department of Geosciences, UiT, Trine Merete Dahl, Karina Monsen, Ingvild Hald are acknowledged for the efficient work and assistance in samples preparation and TOC analysis. Nussir ASA and in particular Øystein Rushfeldt is thanked for providing this research with study material. The authors thank Calvin Shackleton for correction of text.

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

#### **References**


*Article*
