Fundamentals and Practical Aspects of Acid Mine Drainage Treatment: An Overview from Mine Closure Perspective

Abstract

Acid mine drainage (AMD) is perhaps one of the most relevant challenges the mining industry has faced during the last few decades. This issue is particularly important in the scenario of mine closure where mining processes cease to be active, and the sustainability of the sites needs to be re-established. This chapter reviews the fundamentals behind the generation of AMD as well as a set of physicochemical phenomena (chemisorption, precipitation, neutralisation, etc.) usually considered by researchers to mitigate it. Mine closure conditions where human presence is seldom or frankly rare turn the wastewater treatment even more challenging as it cannot be intensive in the utilization of reagents, energy, or human resources. Therefore, from a practical standpoint, passive-like wastewater treatment strategies mimicking nature are preferred. Finally, insights with regards to the complexities behind the implementation of pilot plant and industrial wastewater treatment systems conformed by long-term reactive barriers and constructed wetlands are also revised.

1. Introduction

During last few decades, there has been an increasing awareness among the scientific community about the impact of carrying out mining activities [1, 2]. Before implementing standard ore exploitation activities, potential contaminant species remain restrained inside the original rock, however, such situation changes once mining activities kickoff and valuable material along with other toxic species are mobilised throughout the atmosphere or other media such as surface- or ground-waters. Among the latter, Acid Mine Drainage (AMD) has arisen as one of the most relevant multidisciplinary challenges in the mining industry [3]. The AMD corresponds to an aqueous stream which appears spontaneously from the natural contact, and therefore the natural interaction, between the surface of the rocks (or mineral particles) exhibiting at their surface primarily metal sulphide structures, and water either in the form of vapour or liquid in conjunction with other atmospheric gases (Figure 1) [4, 5].

Figure 1.

Scheme of acid mine drainage (AMD) production.

Perhaps the major difference between AMD and other sorts of pollution is that the former is not directly produced by mining activities. Mining activities would inevitably produce, to some extent, solid wastes and then, the environment in contact with them would eventually trigger the generation of AMD. In other words, the misplacing of solid wastes coming from anthropogenic mining activities in nature itself spontaneously transforms it into a different system with increased toxicity. In this context, the appearance of AMD depends largely on the local atmospheric conditions. For instance, higher humidity or rainy weather will favour the generation of AMD compared to dry conditions [4].

From a historical standpoint, one of the first reports indicating the generation of AMD was published in 1895 by F.G. Holman who glanced at the presence of a waterflow coming out from a small mine site in Forbestown, Sierra Nevada, California [6]. During AMD formation several physical, chemical, and biological phenomena are triggered while solid wastes and environment interact and are commonly summarised by many authors as simple as “weathering” [7, 8]. Weathering, though, is a wide concept that encompasses all the characteristics related to the environment including climate and biosphere. This makes it a bit too general to fully predict the specifics of AMD (timespan to appear, chemical composition, etc.) and its instantaneous or long-term impact on the mine site surrounding areas. The locations where mine sites are placed present a variety of different climates like desertic, Mediterranean, or other. Therefore, when carrying out any study on AMD prediction, prevention, treatment or other, the ambient conditions used will be crucial to get proper results [9].

Before examining the fundamentals behind the generation of AMD, a brief analysis of where AMD might be generated from a mineral processing perspective will be presented. There are many situations where AMD may appear across the mineral processing line, especially when solid wastes appear. For instance, it is well known that base metals occurrence covers a wide range of mineral structures such as oxides, sulphides, and intermediate phases [10]. Although oxide minerals bearing ores may also produce AMD due to their susceptibility to undertake leaching and metal hydrolysis steps in aqueous aerated conditions, sulphide-bearing minerals are considered the major ones responsible for it. In that context, AMD is mainly associated with base metals and coal beneficiation plants where sulphide minerals occur as valuable or gangue material [11]. Only for exemplification purposes, and given its worldwide relevancy, the copper sulphide pyrometallurgical beneficiation path will be discussed. The traditional copper sulphide line of process usually includes blasting, rock size reduction (crushing and milling) and mineral selective separation commonly froth flotation [12]. Figure 2 presents a block diagram of that line of ore processing, identifying the most relevant scenarios where AMD may be anticipated to take place.

Figure 2.

Possible ore processing situations where AMD may appear, and the maximum diameter of rocks/particles involved in each scenario. Particle diameter is referred from [13].

From a practical perspective, it is all about sulphide-bearing material in the form of particles with different diameters being piled up which generate a certain natural porosity that will determine the access of atmospheric gases (or material weathering) towards the interior of the porous material. Different particle size distributions can be observed across the process line characterised in the picture by the maximum particle size only. Another common and better way to estimate a mean diameter biased to larger diameter values of a set of rocks or particles is through the Sauter mean diameter (${\mathit{d}}_{\mathbf{32}}$) [14]. As expected, the Sauter mean diameters follow a similar trend to the maximum particle diameter [Eq. (1)].

$$d_{32,\mathit{ROM}}\ge d_{32,\mathit{\text{pebbles}}}\ge d_{32,\mathit{\text{tailings}}}$$E1

where,

$${\mathit{d}}_{\mathbf{32}}\mathit{=}\frac{\sum _{\mathit{i}}{\mathit{d}}_{\mathit{i}}^{\mathbf{3}}}{\sum _{\mathit{i}}{\mathit{d}}_{\mathit{i}}^{\mathbf{2}}}$$E2

Eq. (2) is normally preferred as it considers the whole population of rocks or particles which brings up the significance of a correct sampling procedure, another crucial aspect of AMD. The mean diameter of a population of rocks or particles is studied with different techniques depending on the relevant particle sizes. For instance, after blasting operation optical methods are widely used while for tailing materials sieving processes are preferred [15]. Nevertheless, particle size is not the only key variable to look at. The content of the valuable and gangue species present in the solid waste is also a variable to take into consideration. For instance, copper grades of sulphide minerals bearing ores fed to a concentrator usually contain around 0.8% copper leaving final tailings with about 0.1% of the metal [16]. Low-grade copper sulphide ores fed to dump leach operations contain copper grades between 0.1 and 0.3% and the extraction may reach values of around 50% [17]. The requirements for ore sorting vary from one case to another but it is common to impose a cut-off grade of around 0.2% [15].

2. Fundamentals: physical, chemical, and biological aspects involved in AMD generation

2.1 Some physical aspects

From a mineral processing standpoint, the bigger the particles or rocks the less the chances of sulphide minerals from getting exposed to the environment. Such exposition is commonly associated with the concept of liberation. The liberation of a specific mineral was originally defined as the particle size threshold allowing the generation of particles composed of only one mineral [18]. From this definition, the liberation of each mineral present in an ore should be different as the presence of each mineral (or occurrence) changes from one to another. Such definition has been modified in time to better explain the efficiency of processes involved in metallurgy or mineral separation stages where the composition of the surfaces of particles is critical for their success [19]. For any geological occurrence of sulphide minerals (or any other mineral) the smaller the particle size the higher the liberation expected. The higher the liberation, the larger the exposure to the environment of surfaces containing such minerals and therefore the higher the chances of producing AMD. Nevertheless, there is in a way a trade-off between AMD flow rates and the potential of producing AMD.

Figure 3 shows two sets of particles having significant differences in particle diameter. The group of bigger particles will exhibit larger pore sizes although the air hold up may be similar for both case scenarios [20]. The set of particles exhibiting coarser sizes presents higher permeability than that of smaller sizes. Permeability depends on both static and dynamic properties of the porous medium and fluid characteristics. In the case of finer particles, capillary forces are more relevant which allow retaining more volume of the aqueous phase inside the porous material leaving the fluid phase and dissolved species to be transported slowly across through diffusive mechanisms favouring the acidification of the aqueous phase [21]. Coarser particles can produce higher flow rates reducing the residence time of fluid in contact with the surface of the particles. Then, there might be situations where the porous structure is fulfilled or saturated with water and the permeability can be modelled using the Darcy equation, however, for most case scenarios unsaturation would be frequently observed. So, permeability would be better modelled by the Soil-Water Critical Curve (SWCC) curve (Figure 4) which can be evaluated using three ranges of pressure values [Eqs. (3)–(5)] [22].

Figure 3.

The impact of particle size distribution on mineral liberation and on permeability and its association with AMD generation.

Figure 4.

SWCC Curve used to model the water content in an unsaturated porous media.

The parameters ${\mathit{w}}_{\mathit{u}}$ and ${\mathit{w}}_{\mathit{aev}}$ represent the water content for 1 kPa suction and for air-entry value, respectively. The first section [Eq. (3)] close to the y-axis represents the zone where the porous media is fulfilled with the aqueous phase which could represent the case of tailings coming from flotation operations when they have been freshly disposed of.

$${\mathit{w}}_1\left(\mathit{\psi }\right)\mathit{=}{\mathit{w}}_{\mathit{u}}\mathit{-}{\mathit{S}}_1\mathit{log}\left(\mathit{\psi }\right)1\mathit{\le }\mathit{\psi }\mathit{<}{\mathit{\psi }}_{\mathit{aev}}$$E3

The second section [Eq. (4)] located immediately to the right-hand side represents the behaviour of the porous media when the air gets inside displacing the water present in the pores which could correspond to an intermediate case observed in tailings where, by simple syneresis, the water drains across the material.

$${\mathit{w}}_2\left(\mathit{\psi }\right)\mathit{=}{\mathit{w}}_{\mathit{aev}}\mathit{-}{\mathit{S}}_2\mathit{log}\left(\frac{\mathit{\psi }}{{\mathit{\psi }}_{\mathit{aev}}}\right){\mathit{\psi }}_{\mathit{aev}}\mathit{\le }\mathit{\psi }\mathit{<}{\mathit{\psi }}_{\mathit{r}}$$E4

And the last section [Eq. (5)] represents the behaviour of the porous material when it gets dry leaving a certain residual water content, which describes a porous media where humidity is present mainly by wetting the surface of the particles.

$${\mathit{w}}_3\left(\mathit{\psi }\right)\mathit{=}{\mathit{S}}_3\mathit{log}\left(\frac{{10}^6}{\mathit{\psi }}\right){\mathit{\psi }}_{\mathit{r}}\mathit{\le }\mathit{\psi }\mathit{<}{10}^6\mathit{kPa}$$E5

Eq. (5) corresponds to later stages of tailings, the case of larger particles still retaining some water content (i.e., material coming from dump leaching or ore sorting operations) or abandoned tailings. From a thermodynamic standpoint, the suction is a function of the partial pressure of the pore-water vapour and the density of the vapour which depends on the temperature and can be computed as in Eq. (6).

$$\mathit{\psi }\mathit{=}\left(\mathit{\text{pore}}\mathit{-}\mathit{air}\mathit{\text{pressure}}\right)\mathit{-}\left(\mathit{\text{pore}}\mathit{-}\mathit{\text{water pressure}}\right)\mathit{+}\mathit{\text{osmotic suction}}$$E6

And the permeability at residual water conditions can be computed from the matric (soil) suction as Eq. (7) [22].

$${\mathit{k}}_{\mathit{w}}\mathit{=}\frac{{\mathit{k}}_{\mathit{s}}}{{\left[\frac{{\mathit{u}}_{\mathit{a}}\mathit{-}{\mathit{u}}_{\mathit{b}}}{{\left({\mathit{u}}_{\mathit{a}}\mathit{-}{\mathit{u}}_{\mathit{b}}\right)}_{\mathit{b}}}\right]}^{\mathit{n}´}\mathit{+}1}$$E7

One of the strong capabilities of Eq. (7) is that it correctly describes porous media with small particle sizes exhibiting lower permeability and higher water retention capacity [22, 23]. The latter would reduce the chances of producing large flow rates of AMD under these conditions [20]. Large flowrates would only then be possible in these conditions when water flows over the external surface of the piled-up porous material which reduces radically the exposed surface of the particles to the aqueous phase flowing around. Certainly, the magnitude of the drying conditions will depend on the water table present in each system. As it can be observed such description is based on semi-empirical mathematical models which is an indication that this is still a quite fruitful field of research.

Finally, the transport of liquid in porous materials built up of smaller particle sizes will expose the sulphide minerals in greater extension but they usually exhibit a higher hydrophobicity due to the presence of sulphur produced by oxidation reactions. This is known as natural hydrophobicity which occurs with much lower significance in the case of mineral oxides with stronger wettability properties. The liquid phase transferred through the porous medium needs to fill the voids displacing the air. Such subprocess is commonly referred to as imbibition and will be inhibited by the presence of sulphide hydrophobic surfaces which provide a first glance of how physics and chemistry are linked in these systems, but it is usually not considered. Indeed, physics and chemistry are frequently addressed by researchers separately. The relevancy of the chemistry and biology behind this process will be examined in the next subchapter.

2.2 Some chemical and biological aspects

There are several documents describing how sulphide minerals produce the so-called AMD, and the reader could refer to them for more information [4, 24]. Gas-solid and liquid-solid interactions are the major ones responsible for the significant differences between the chemical composition and structure of the bulk of the solid phase and the outmost surface layer arising from such interaction [25]. There are specific minerals that due to their instability under aerated conditions, notably metal sulphide minerals, are likely to produce enough acidity to stabilise several metals in dissolved state. Firstly, metal sulphide minerals would directly produce hydronium ions from their oxidation produced by the oxygen present in the atmosphere [Eq. (8)].

$${\mathit{MS}}_{\mathit{m}\left(\mathit{s}\right)}\mathit{+}{\mathit{O}}_{2\left(\mathit{g}\right)}\mathit{+}\left(\mathit{m}\mathit{-}\mathit{n}\right){\mathit{H}}_2{\mathit{O}}_{\left(\mathit{l}\right)}\mathit{\leftrightarrow }{\mathit{M}}_{\left(\mathit{aq}\right)}^{\mathit{n}\mathit{+}}\mathit{+}\mathit{m}{\mathit{SO}}_4^{2\mathit{-}}\mathit{+}\left(\mathit{m}\mathit{-}\mathit{n}\right){\mathit{H}}_3{\mathit{O}}_{\left(\mathit{aq}\right)}^{\mathit{+}}$$E8

Such acidity is then enhanced by metal hydrolysis reactions occurring at the bulk of the aqueous phase. Hydrolysis can be represented by Eq. (9).

$${\mathit{M}}_{\left(\mathit{aq}\right)}^{\mathit{n}\mathit{+}}\mathit{+}\mathit{m}{\mathit{H}}_2{\mathit{O}}_{\left(\mathit{l}\right)}\mathit{\leftrightarrow }\mathit{M}{\left(\mathit{OH}\right)}_{\frac{\mathit{m}}{2}\left(\mathit{i}\right)}^{\left(\mathit{n}\mathit{-}\frac{\mathit{m}}{2}\right)}\mathit{+}\frac{\mathit{m}}{2}{\mathit{H}}_3{\mathit{O}}_{\left(\mathit{aq}\right)}^{\mathit{+}}$$E9

where the variable $\mathit{m}\mathit{=}0\mathit{,}2\mathit{,}4\mathit{,}6\mathit{,}8\mathit{,}\mathit{etc}\mathit{.}$ Depending on the concentration, $\mathit{i}$ may refer to a solid phase for $\mathit{n}\mathit{=}\mathit{m}$ leading to the precipitation of the metal hydroxide. If the dissolved metal is polyvalent, it may undergo subsequent oxidation stages due to the presence of dissolved oxygen in the system. The latter may raise other stronger mechanisms of oxidation of the sulphide minerals. That is the case of dissolved iron which can go from Fe(II) to Fe(III) in acidic aqueous solutions due to dissolved oxygen reduction. The latter increases the rate at which the metal sulphide dissolves producing more acidity simultaneously rising the concentration of sulphate ions in solution.

2.3 Application to pyrite and marcasite (${\mathit{FeS}}_2$)

Probably the most reported case study that exemplifies AMD generation is that of pyrite and marcasite (${\mathit{FeS}}_2$) which encompasses a series of processes that up to now are not fully understood, especially with regards to the state of surface or surface mediator being formed [26].

In any case, the oxidation reaction is usually described as Eq. (10).

$${\mathit{FeS}}_{2\left(\mathit{s}\right)}\mathit{+}{\mathit{Ox}}_{\left(\mathit{aq}\right)}\mathit{+}2{\mathit{H}}_2{\mathit{O}}_{\left(\mathit{l}\right)}\mathit{\leftrightarrow }{\mathit{Fe}}_{\left(\mathit{aq}\right)}^{2\mathit{+}}\mathit{+}{\mathit{H}}_3{\mathit{O}}_{\left(\mathit{aq}\right)}^{\mathit{+}}\mathit{+}{\mathit{SO}}_{4\left(\mathit{aq}\right)}^{2\mathit{-}}$$E10

In Eq. (10) the oxidant, represented by the symbol $\mathit{Ox}$, can be oxygen or ferric ions if the thermodynamic potentials are suitable. The ${\mathit{FeS}}_{2\left(\mathit{s}\right)}$ in the reactants is only an over-simplification of what is really happening. Indeed, Holmes and Crundwell (2000) succeed in describing the oxidation of pyrite using the classic mixed potential theory but claimed that the state of surface also plays a relevant role in the process and is still not well understood [27]. Figure 5 attempts to summarise the kinetics of the major electrochemical reactions taking place in the system. The oxidation reactions of pyrite in presence of sulphate ions start at potentials about 0.54 V vs SHE, which as mentioned later in this text becomes wider in presence of chloride ions. The oxygen reduction reaction in acidic conditions exhibits a Nernst potential of about 1.23 V vs SHE. Such reduction reaction is known to be from an electrochemical point of view rather irreversible, so currents are only observed below 0.61 V vs SHE which is explained by the authors in terms of the low conductivity n-type pyrite semiconductor properties [28]. The latter is not observed in Figure 5. Instead, due to the significant reduction in overvoltage of the oxygen reduction, the current density would reach a maximum value of about $38.5\mathit{mA}{\mathit{cm}}^2$ considering a layer thickness of 0.05 cm, a concentration of $8\mathit{mg}{\mathit{L}}^{\mathit{-}1}$, and a diffusion coefficient of the unstirred aqueous phase of $2\mathit{\bullet }{10}^{\mathit{-}5}{\mathit{cm}}^2{\mathit{s}}^{\mathit{-}1}$. The current density observed by the authors is about 2% or that maximum value which is an indication that the diffusion layer thickness is much higher than that assumed. This only allows an exiguous net current that can be barely detected (in fact, in Figure 5 it was enhanced for the reader to be able to see it!) at a mixed potential labelled as (1), which is controlled by the cathodic reaction [27]. The low current is also obtained in the situation presented by the authors where both cathodic and anodic reactions exhibit a thermodynamic potential close to each other. In fact, under these conditions, the reversibility of both reactions needs to be high to attain any relevant reaction rates. Other authors have indicated that the oxidation potential in presence of chloride ions, especially relevant in mineral processes implemented using seawater, ranges between −70 V and 530 V vs SHE [29]. Strikingly, under such a range of potentials, a series of reversible electrochemical adsorption/desorption steps take place whenever sweep rates of$30\mathit{mV}{\mathit{s}}^{\mathit{-}1}$ or higher is used. At sweep rates below $10\mathit{mV}{\mathit{s}}^{\mathit{-}1}$ the authors found that oxidation steps are triggered indicating that precursors need time to be formed to undertake oxidation reactions. Under these experimental conditions, the pyrite oxidation takes place as a sequence of electron losses promoting the formation of several sulphite-like precursors occurring at both solid surface and bulk of aqueous phase.

Figure 5.

Electrochemical scheme of the I vs E curve presenting the most relevant electrochemical reactions as well as the mixed potentials.

Once ferric ions are formed in the aqueous phase, the oxidation rate of pyrite increases significantly not only because of the increase of the anodic overvoltage but also because of the high reversibility of the reduction reaction of ferric ions exhibits. Indeed, the mixed potential moves to higher voltages represented as (2) in Figure 5. Then, an accumulation of ferrous ions in solution may arise. The chances of regenerating the oxidant only by introducing oxygen would not be enough since the latter reaction still is mass transfer controlled. This is one of the most crucial issues the leaching of copper sulphide minerals presents which has been partially solved by microorganisms. In effect, it has been proved that bacteria, specifically, thiobacillus ferroxidans, would catalyse this reaction [30]. For many years it was not clear whether the catalysis is achieved by direct bacteria adsorption and oxidation of the mineral sulphide surface or indirect reaction through oxidation of ferrous ions happening at the bulk of the aqueous phase. Nowadays, such matter has been sorted out and it is known that it is the indirect oxidation mechanism that governs the oxidation of ferrous ions to ferric ions, and it constitutes the foundations of bioleaching of sulphide minerals [31]. During such studies on bioleaching, it has been learnt that the thiobacillus ferroxidans have a relatively narrow pH range in which they may adapt at their best and the presence of chloride ions at 2M or higher inhibits its growth [32].

Figure 6 summarises the main role of the two major types of bacteria, thiobacillus ferrooxidans and thiobacillus thiooxidans. Eqs. (10) and (11) are commonly coupled and a global reaction is obtained indicating how oxygen can oxidise elemental sulphur. These two reactions may also be studied separately. Dissolved oxygen oxidises a number of reduced chemical species such as Fe(II) or any other sulphide more susceptible to be oxidised than pyrite while sulphur can be at least oxidised by Fe(III). A simple stoichiometric analysis of both routes of reaction, without considering the cycle Fe(II)/Fe(III), indicates the simultaneous oxidation of 1 mol of elemental sulphur and the reduction of 2 mols of molecular oxygen would neutralise the local pH. In acidic aqueous solutions (as it may happen with AMD) it would be desirable to have more than 2 mols of oxygen reacting per mol of sulphur being oxidise. This rather simplistic analysis attempts to prove that the generation of AMD is not the inevitable outcome coming from these systems. Understanding and tuning the relevancy of the reactions at a fundamental level may also lead to inhibiting its formation.

Figure 6.

Conceptual simplified model for generation of AMD.

$${\mathit{O}}_{2\left(\mathit{g}\right)}\mathit{+}4{\mathit{H}}_3{\mathit{O}}_{\left(\mathit{aq}\right)}^{\mathit{+}}\mathit{+}4{\mathit{e}}^{\mathit{-}}\mathit{\to }6{\mathit{H}}_2{\mathit{O}}_{\left(\mathit{l}\right)}$$E11

$${\mathit{S}}_{\left(\mathit{s}\right)}^0\mathit{+}12{\mathit{H}}_2{\mathit{O}}_{\left(\mathit{l}\right)}\mathit{\to }{\mathit{SO}}_{4\left(\mathit{aq}\right)}^{2\mathit{-}}\mathit{+}8{\mathit{H}}_3{\mathit{O}}_{\left(\mathit{aq}\right)}^{\mathit{+}}\mathit{+}6{\mathit{e}}^{\mathit{-}}$$E12

Another more realistic analysis would involve coupling these two reactions. In this case, 2 mols of elemental sulphur would react with 3 mols of molecular oxygen producing 4 mols of hydronium ions and 2 mols of sulphate which is also troublesome for AMD (to be discussed in Section 5). Simultaneous oxidation of 12 mols of ferrous ions using 3 mols of molecular oxygen would then neutralise the acidity provided by the overall sulphur oxidation reaction by oxygen.

Supposedly, in real systems oxygen is slowly transferred to surface sites inside the porous media where the interaction with sulphide minerals exposed would produce elemental sulphur, one of the most relevant products generated at the surface of the particles, which then is oxidized due to the presence of microorganisms. Additionally, bacteria require oxygen and eventually carbon dioxide for growth, which is also responsible for producing both sulphur and acid [33]. It looks like the key would lie in inhibiting the formation of elemental sulphur which is quite insoluble (about $0.6\mathit{ng}{\mathit{L}}^{\mathit{-}1}$) [34]. Nevertheless, the role of sulphur though is much more complicated than just the formation of sulphate ions [29]. Research studies have proved that thiosulphate would be an intermediate species that in acid media would be disproportionate to sulphite or sulphur dioxide and elemental sulphur, going back to the initial state. Then, handling better the presence of sulphite or sulphur dioxide then might be crucial. Thermodynamically, elemental sulphur could be avoided but to do that an increase in temperature or the reduction of total amount of sulphur in the system is needed which is something not easy to accomplish considering the scale at these systems are commonly implemented [35].

From all the above, it may infer that the chances of producing AMD cannot only be observed from a fluid dynamic or physical perspective. The gathering of key reactants needs to occur to produce it. Delays in the interaction between reactants given by the transport of oxygen, or carbon dioxide in less importance, will slow down the generation of acidity and therefore that of AMD.

It is precise because of this that many of the strategies to prevent AMD obey to block the reagents from coming into the porous material (Figure 7). For instance, with regards to the sulphide minerals present in the porous material authors have suggested removing it before piling up the solid wastes using froth flotation or any other selective separation method [36]. Bear in mind that this requires preparing the material for the separation process such as milling to certain particle diameter and the use of appropriate reagents at certain dosages to run froth flotation operations adequately. Another option would be to cover up only the surface of the particles exposing the sulphide mineral to the gaseous phase to avoid any contact with oxygen inhibiting the AMD formation [40]. A similar but more extensive blockage would entail forming a cap enclosing the whole porous material preventing oxygen from entering the system [41]. Even more, some researchers have suggested using materials known as not acid producers (or NAP) and directly carrying out some neutralisation of the AMD. It also has been suggested to introduce some positive pressure on inert gases to keep the oxygen from entering the system [42]. From the use of water, perhaps the most studied strategy for suspensions of mineral particles involves reducing the water content of slurries to dose lower amounts of neutralisation reagents [43]. And finally, with respect to the microorganisms, it has been recommended the use of some bactericides to prevent the acid-forming bacteria to appear in the system [44] while other authors have focused their attention in using some bacteria growth inhibitors [45].

Figure 7.

Scheme of various strategies to prevent or treat AMD [36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55].

However, having implemented any of these paths to prevent AMD from appearing does not secure that it will not take place and if it occurs, several actions have been studied to deal with it. Furthermore, in many cases, these solid wastes are not appropriately disposed of over geomembranes or other impervious materials which forces continuously monitor surface- and ground-waters at the mining location outskirts. Whenever these waters acquire any properties resembling AMD, the wastewater treatment must act as a barrier to bringing the parameters of water quality back to their usual values. Even using such geomembranes does not ensure that the eventual AMD produced will be appropriately contained as the properties of these covers may also detrimentally evolve in time [56].

3. AMD in mine closure conditions

Mine closure is one of the subjects in the field of mine management that has gained notoriety over the last few decades [57]. The people’s perspective of mining activities is negative when abandoned mine sites impact adversely the environment or the nearby communities [58]. Figure 8A presents the iconic case of the Grand Canyon, in the United States, where mining activities were developed between 1957 to early 1960s. It can be observed that a section of the original plant is still in place and the exploitation has increased the surface area exposed to the atmosphere. In the same picture the before and after of a gold exploitation mine site located in the northern part of Chile shows a similar situation. In this case, the open-pit mine, unfortunately, made the entire town of Churrumata move without procuring better conditions for the villagers (Figure 8B).

Figure 8.

Pictures of two abandoned mine sites. (A.1 and A.2) Iconic case of Grand Canyon, United States, Guano Point, Bat Cave Guano Mine active between 1957 and early 1960, nowadays it is a renown National Park, (B.1) Before and (B.2) after gold mining exploitation of an open pit mine built at the heart of the small town Churrumata in the northern part of Chile, the mine site was active between 1984 and 2010 approximately.

All mine sites have a definite lifetime. At the end of the mining exploitation, the site needs to be rehabilitated, ideally eluding any threat to the environment, or living organisms, vegetation, and nearby communities.

Behind this topic there are many concepts to address and it is difficult to summarise them in just a few lines. For instance, researchers have differentiated the terms mine closure and mine completion [59]. On one hand, mine closure is a procedure over a timespan where plant operation stops, and decommissioning is undertaken. On the other hand, mine completion refers to an aim of mine closure where the ownership is renounced by the mining lease and accepted by the next user of the land for a different purpose. These and other perspectives are still an ongoing theme for the whole society.

Despite the latter, for this quest to be successful, every government has stated a plan which considers not only technical requirements but also regulatory and legislation guidelines to prevent the sites to become hazardous reducing eventual further contamination in many years to come.

It was not necessary to go by a great deal of time before governments, mining companies and the whole society realised that the major threat behind the cease of a mining operation is the lack of planning. There are many reasons why mine sites close such as economics, geological, technical, regulatory, policy changes, social pressure, end of markets, etc. Therefore, it is not surprising that this stage of mine development requires a multidisciplinary set of actions. For instance, the actions relevant to the present subject might include:

1. Focused brainstorming aims at identifying environmental values, gains or losses inherent to AMD.

2. Preliminary evaluation of testing sulphide ore bodies for acid-based accounting and metals.

3. Potential use of overburden to cap potential generation of AMD.

As the Mine Closure conditions refer to the situation where the mining activities cease to take place, it is expected that energy, material, or personnel are not going to be available to deal with AMD. Therefore, unassisted wastewater treatment needs to be implemented.

4. Major physicochemical phenomena used to mitigate the generation of AMD

The AMDs characteristics vary from site to site, however, for simplicity and without loss of generality the pH usually ranges between acid (close to pH 1 to 3) and slightly acid (around pH 5) which after wastewater treatment needs to reach pH values between 6.5 and 9 to be adequately discarded [60]. The acidity stabilises a concentration of certain heavy metals in aqueous solutions such as copper, zinc, chromium among others and metalloids like arsenic. Several physicochemical phenomena used to remove ions from an aqueous phase are succinctly reviewed in this section from both thermodynamic and kinetic classic modelling standpoints.

4.1 Neutralisation: coupled or not with chemical precipitation

Perhaps this is the most straightforward strategy to treat AMD wastewaters coming from mine sites. The idea is to remove from the aqueous phase dissolved species that might be toxic to human beings and the environment by adding salts with high saturation or solubility product constant (Kps) values but very small ones for the ions to be withdrawn. Considering that the AMD is acid, one of the priorities is to bring the pH of the water close to neutral and that is usually accomplished by dosing with an alkali reagent. The release of hydroxyl ions (${\mathit{OH}}_{\left(\mathit{aq}\right)}^{-}$) to the aqueous solution enhances the hydrolysis of the metal ions promoting the generation of neutral species which will eventually precipitate. The most used reagent at an industrial scale is the quicklime which in presence of water is transformed into slaked lime [61]. In Figure 9 a usual AMD containing a high concentration of acid ($H_3{O}_{\left(\mathit{aq}\right)}^{+}$), heavy metals ($M_{\left(\mathit{aq}\right)}^{n+}$) and sulphate (${\mathit{SO}}_{4\left(\mathit{aq}\right)}^{2-}$) is presented. When adding quicklime, ultimately dissolved calcium ions would react with sulphate forming anhydrite-like structures while hydroxyl ions would form metal hydroxides and simultaneous adjust the pH to neutral values. The precipitation of sparingly soluble species would follow two sequential steps namely nucleation and crystal growth being the former, in many cases, more energy consuming than the latter. That is the reason why seeds and/or surface roughness of different materials would be desirable to improve the precipitation thermodynamic spontaneity and kinetics [62].

Figure 9.

Sketch of the role of different ions formed during the dosing of quicklime to acidic aqueous solutions.

4.1.1 Thermodynamics

An initial concentration of metal is $C_{\mathit{Me},0}$ and of a counter ion (anion, which can be sulphate) is $C_{\mathit{ci},0}$. The flow rate of liquid enriched in these two ions is $Q$. For computation purposes, the dose of leachate of a metal hydroxide (for simplicity, calcium hydroxide or slaked lime) of concentration $C_x$ will be considered at a flow rate of $q$. Usually, as $C_x$ is relatively high, the following flowrate relationship holds $q\ll Q$. The reagent with higher solubility compared to that of the salt to be precipitated, undertake an ionisation reaction such as that presented in Eq. (13).

$$M{\left(\mathit{OH}\right)}_{n´\left(s\right)}\leftrightarrow M_{\left(\mathit{aq}\right)}^{n´+}+n´{\mathit{OH}}_{\left(\mathit{aq}\right)}^{-}$$E13

The mathematical condition required for the precipitation to occur is given by Eq. (13).

$$\frac{C_x{Q}_r}{Q}{C}_{\mathit{ci},0}>{\mathit{Kps}}^{\left(1\right)}$$E14

Then, the efficiency of the treatment can be computed using Eq. (14).

$$\phi =100\left\{\frac{C_x{Q}_r}{2C_{\mathit{ci},0}Q}+\frac{1}{2}\left[\sqrt{{\left(\frac{C_x{Q}_r}{C_{\mathit{ci},0}Q}-1\right)}^2-\frac{{\mathit{Kps}}^{\left(1\right)}}{C_{\mathit{ci},0}^2}}-1\right]\right\}$$E15

The physical units of the different parameters in these equations need to be consistent. Plus, bear in mind that the efficiency of Eq. (15) is overestimated as only the major species were considered in this computation. Simultaneously, the precipitation of metals can be computed from Eq. (16). For the metal precipitation, Eq. (16) needs to be solved.

$${\mathit{Kps}}^{\left(2\right)}=\left(C_{\mathit{Me},0}-y\right){\left(n´C_x-\mathit{ny}\right)}^n$$E16

Eq. (16) does admit an analytical solution only in very specific conditions, so it is better to solve it numerically. Although it looks like this strategy is quite promising as it would be able to remove sulphate, heavy metals and even neutralise the acidic conditions of the AMD, it is somehow misleading for at least four reasons [63, 64]:

1. The precipitates being formed are not necessarily stable.

2. The dose of quicklime does not produce a perfect solution. It is most of the time leachate indicating that several particles are suspended and not instantaneously dissolved. This is to some point troublesome as the final pH will evolve in time reaching eventually much higher values.

3. Even more, such loss of control of pH may lead to an over increase of pH generating anionic species of the metals which will be redissolved increasing again their concentration in the aqueous solution. Even more, in many cases, the minimum concentration of heavy metals reached by using quicklime does not satisfy the maximum concentration permitted by current regulations.

4. Finally, there are many chances to get different quicklime dosing optimum points for metals and pH adjustment. Indeed, there are small chances that one single dosage of the reagent will exactly satisfy the minimum concentration of multiple heavy metals and the right pH at once. That is why this mechanism can only be used to approach best conditions for the correct discharge of the treated AMD.

During mine closure conditions this strategy is not directly recommended since the reagent dosing control is difficult to implement without personnel in place. However, the fundamentals behind this mechanism still hold. In the scenario of mine closure, the main idea would be to incorporate some sparingly soluble minerals with alkaline behaviour such as silicates, or others.

4.1.2 Kinetics

There are several studies focused on determining the dissolution rate of solids, especially that of quicklime in water. The first stage of dissolution is usually modelled by Eq. (17) [65].

$$\frac{d{a}_i}{\mathit{dt}}=k\left(a_{i,\mathit{sat}}-a_i\right)$$E17

where $a_i$ is the activity of the ion “$i$” as a function of time, $a_{i,\mathit{sat}}$ is the activity of the ion “$i$” in saturation conditions, $t$ represents the time, and $k$ is a specific rate constant obtained per unit of volume that has been defined as a function of the diffusion coefficient of the ion being transferred from the solid surface to the aqueous solution bulk ($D$), the thickness of the diffusion layer ($\delta$), and the specific surface $\left(\frac{S}{V}\right)$ as presented in Eq. (18) [66].

$$k=\left(\frac{S}{V}\right)\frac{D}{\delta }$$E18

The classic shrinking core model with reaction control can also be used (Eq. 18) [67, 68, 69].

$$1-{\left(1-\alpha \right)}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$3$}\right.}=k_{\mathit{cc}}t$$E19

where $\alpha$ represents the conversion of the reaction which corresponds to the volume fraction of solid that has been dissolved, and $k_{\mathit{cc}}$ is the kinetic rate constant when the process is governed by chemical control.

4.2 Adsorption: Chemisorption

This mechanism aims at removing pollutants from an aqueous phase by fixing them onto a surface of a solid which is stable when immersed in the wastewater. The adsorption mechanism is one of the preferred reactions for wastewater treatment not only because low-cost adsorbents consisting of by-products or wastes from other industries may be used, but also because it may reach high removing efficiencies of dissolved molecules with final concentrations of a few parts per billion [69]. Figure 10 presents several aspects to consider when picking up this mechanism. Different reactions between the adsorbent and the aqueous solution lead to the partial dissolution of the adsorbent affecting the local pH near its surface and its stability of the adsorbent suspension whenever forming small particles may occur [70].

Figure 10.

Diagram of adsorption processes used in wastewater treatment.

There are several drawbacks behind the implementation of adsorption-based technologies. This technology requires optimising the contact between the solid phase and the aqueous phase containing the species to be adsorbed. Usually, piling up of adsorbent material in a column disposed of vertically or horizontally is preferred [71] but maintaining the permeability of the porous medium with time could become a challenge. It is also desirable to implement technologies using chemisorption rather than physisorption. Chemisorption has many advantages such as its specificity exemplified in Figure 10 by the single and double binding shown for the metal and sulphate. That is, different adsorption sites would be used by different types of adsorbates reducing the competition for adsorption sites. Plus, the relatively high binding energy associated with the adsorption process turns it quite irreversible from a kinetic standpoint which reduces the chances of pollutants desorption. Nevertheless, the main disadvantage would be that the eventual saturation of the adsorbent may be reached needing to move forward to a desorption stage to regenerate the adsorbent [72].

4.2.1 Thermodynamics

The thermodynamics of the adsorption process is explained in terms of the adsorption isotherm [73]. The adsorption isotherm is usually plotted in a graph where the y-axis represents the maximum quantity of adsorbed species per unit of the dry mass of adsorbent (also known as specific adsorption) while the x-axis presents the concentration of the species in equilibrium with the specific adsorption measured. All the data is obtained experimentally at a constant temperature and solids percent. There are many mathematical models that can be used to describe the process having each of the conditions and assumptions that as much as possible must represent the specifics of the process under study. The most used adsorption isotherms cited by researchers are the Langmuir and Freundlich isotherms as Eqs. (20) and (21) [74].

$$q_{i,\mathit{eq},L}=\frac{S_m{K}_L{a}_{i,\mathit{eq}}}{1+K_L{a}_{i,\mathit{eq}}}$$E20

where $q_{i,\mathit{eq},L}$ corresponds to the volume (or mol) of adsorbate $i$ at the surface of the adsorbent in equilibrium conditions, $a_{i,\mathit{eq}}$ is the equilibrium real concentration of the adsorbate in $\mathit{mol}{L}^{-1}$, $K_L$ is the Langmuir isotherm constant commonly associated with the binding energy, $S_m$ es the amount of adsorbate required to form a monolayer.

$$q_{i,\mathit{eq},F}=K_F{a_{i,\mathit{eq}}}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$n$}\right.}$$E21

where $K_F$ is the Freundlich empirical constant usually associated with the sorption capacity, and $n$ is the sorption intensity.

4.2.2 Kinetics

One general case to model adsorption kinetics is presented in Eq. (23) [75]

$$r=\frac{\left(\mathit{\text{kinetic factors}}\right)\left(\mathit{\text{process potential}}\right)}{\left(\mathit{\text{adsorption factor}}\right)}$$E22

wherein the numerator there is a description of the classic law of mass action and in the denominator, the inhibition of the adsorption rate procured by the blockage of surface sites of other species in the system is incorporated. For example, the mathematical model for adsorption kinetics of one species labelled with the underscore “$i$” can be described as in Eq. (24) in the case of chemical reaction control.

$$r=k{\theta }_i=k\frac{K\left[a_i\right]}{1+K\left[a_i\right]+\sum _{j\ne i}{K}_j\left[a_j\right]}$$E23

where $k$ is the specific kinetic constant, ${\theta }_i$ is the fractional occupancy of adsorption sites by the main species “$i$”, $K$ and $a_i$ are the Langmuir adsorption constant referred to the main ion and its activity in the aqueous phase while $K_j$ and $a_j$ are their equivalent but for other ions competing for adsorption sites.

4.3 Redox reactions

This mechanism is highly valuable for certain ions which as product of the electron transfer type of reaction may directly precipitate or produce precursors for precipitation (Figure 11). The reduction of metals is somehow difficult to implement unless certain scrap of metallic wastes contain metals with low standard reduction potentials, also called less noble elements. If the metal is removed from the aqueous solution exhibits higher hydrolysis constants than the metal being released into the aqueous medium, the resulting pH of the solution should increase. One example of this would be the removal of polynuclear lead (II) ions by iron (II) ions [76, 77, 78].

Figure 11.

Redox reactions in wastewater treatment.

4.3.1 Thermodynamics

The idea is to couple two electrochemical half-reactions, one half-reduction reaction and one half-oxidation reaction, having the former a higher standard reduction potential than the latter. As a requirement, since this wastewater treatment must occur without any energy input, the reaction must evolve spontaneously. Such a condition is presented in Eq. (25) [79].

$$E_{\mathit{RED}}-E_{\mathit{OX}}>0$$E24

Using Nernst equation, Eq. (25) would be of use to assess a first analysis of the impact of varying activities of different ions, partial pressures, or temperature on the spontaneity of the process can be assessed.

4.3.2 Kinetics

Electrochemical kinetics of spontaneous redox reactions are commonly studied in terms of the mixed potential theory and the corresponding current density. The mixed potential which is not a thermodynamic parameter is obtained from equalising the anodic and cathodic reactions [Eq. (26)]

$$I_{\mathit{\text{anodic}}}=\left|I_{\mathit{\text{cathodic}}}\right|$$E25

In a complex system, there could be several reduction reactions and oxidation reactions occurring simultaneously in different locations within the system, therefore several mixed potentials may be installed. The open rest potentials, in this case, will attempt to follow such mixed potentials and depending on the conductivity of the species formed at the solid interface and at aqueous solution bulk such tracking down will be faster or slower [80].

5. Practical aspects of semi-passive wastewater treatment: coupling long-term permeable reactive barriers with horizontally constructed sub-surface wetlands

This section is devoted to the design of AMD treatment systems which can operate in mine closure situations, close to stand-alone and unassisted systems to treat wastewaters. These treatment paths attempt to gather many of the mechanisms previously revised acting simultaneously to clean up AMD streams. Since most of these mechanisms need to take place spontaneously, many of these wastewater treatment systems attempt to mimic nature. Particularly, this subchapter addresses some aspects of two of these systems: Long-Term Permeable Reactive Barriers (LTPRB) and Horizontally Constructed Sub-Surface Wetlands (HCSSW).

On one hand, wetlands are one of the preferred passive wastewater treatment strategies to be implemented as a tertiary wastewater treatment [81]. Although it is considered a mature technology by many authors, it is one of the most difficult systems to model and understand. The latter is not only because of the many physicochemical interactions simultaneously taking place between all the species belonging to the system but also because of the multiple roles the local biota may play. For instance, sulphate ions are difficult to reduce using inorganic species only [82]. Indeed, looking at the Pourbaix diagrams, sulphate ions are stable over the whole pH range either in their acid form or not. It has been pointed out, though, that such reduction can be accomplished at the surface of organic material where carbon has a pivotal role. Indeed, it is well known that anaerobic systems, as well as aqueous media set in contact with solid metals, promote the growth of sulphate-reducing bacteria [83]. Carbon, among all its functions such as respiration, fermentation, methanogenesis, denitrification, and iron reduction would have a key role in sulphate reduction [84]. Sulphate reduction reactions are summarised by Eqs. (27), (28).

$$\begin{array}{l}2{\mathit{CH}}_3\mathit{\text{CHOHC}}{\mathit{OO}}_{\left(\mathit{aq}\right)}^{\mathit{-}}\mathit{+}{\mathit{SO}}_{4\left(\mathit{aq}\right)}^{2\mathit{-}}\mathit{+}{\mathit{H}}_3{\mathit{O}}_{\left(\mathit{aq}\right)}^{\mathit{+}}\\ \mathit{\to }2{\mathit{CH}}_3{\mathit{COO}}_{\left(\mathit{aq}\right)}^{\mathit{-}}\mathit{+}2{\mathit{CO}}_{2\left(\mathit{g}\right)}\mathit{+}3{\mathit{H}}_2{\mathit{O}}_{\left(\mathit{l}\right)}\mathit{+}{\mathit{HS}}_{\left(\mathit{aq}\right)}^{\mathit{-}}\end{array}$$E26

$${\mathit{CH}}_3{\mathit{COO}}_{\left(\mathit{aq}\right)}^{\mathit{-}}\mathit{+}{\mathit{SO}}_{4\left(\mathit{aq}\right)}^{2\mathit{-}}\mathit{+}2{\mathit{H}}_3{\mathit{O}}_{\left(\mathit{aq}\right)}^{\mathit{+}}\mathit{\to }2{\mathit{CO}}_{2\left(\mathit{g}\right)}\mathit{+}4{\mathit{H}}_2{\mathit{O}}_{\left(\mathit{l}\right)}\mathit{+}{\mathit{HS}}_{\left(\mathit{aq}\right)}^{\mathit{-}}$$E27

These reactions, though, are not in total agreement with classic electrochemical fundamentals. The standard electrode potential of the sulphate reduction is −220 mV vs SHE which is not fully consistent with the stability region of sulphate ions declared in Pourbaix diagrams [35]. Authors have indicated that such reduction is complex and involves metastable products [35, 85]. The reduction reaction would then consist of at least two reactions in series which are triggered by the sulphate activation by ATP sulphurylase increasing the potential to about −60 mV where the reduction from sulphate to sulphite is achieved. However, the reduction of sulphite to sulphide is yet not fully understood [86]. In addition, another disadvantage of these two reactions is, in principle, the production of carbon dioxide identified as a greenhouse gas. Additionally, authors have pointed out that the low performance in eliminating phosphorous may also be observed for other contaminants which usually increase the requirements in terms of residence time and/or surface lands available to implement these systems [87]. Whenever these systems are not available naturally, constructed wetlands are engineering-designed which can be implemented vertically or horizontally [84]. The latter corresponds to the case study to be described in the next section.

On the other hand, long-term permeable reactive barriers have captured interest from the scientific community since it houses several materials to treat wastewaters securing the correct quality of groundwater resources. The phenomena embedded in this type of strategy are mainly chemical or biological degradation, precipitation, and adsorption to immobilise contaminants [88]. Due to the similarities in dealing with organic matter between reactive barriers and wetlands, sulphate reduction bacteria can also be promoted in these systems. Additionally, the permeability of these systems needs to be secured. Unreactive or low-reaction alkali materials are used as a fixed bed introducing more reactive materials inside the pores that can range from specifically designed materials to wastes from other industries such as ferrihydrite-bearing soils or nanostructured calcium silicate adsorbent, among others [89, 90]. Since the growth of vegetation is not present in these systems, the permeability may be designed to avoid dead volumes or volumes with low mixing capabilities. Long term reaction kinetics is still a matter to do research on. Considering that a few reactions are associated to oxidation mechanisms by oxygen, and given the relatively low concentration of the gas, particularly in low permeability media, atmospheric corrosion perspective could improve the knowledge on these matters [91].

6. Physicochemical complexities behind coupling long-term reactive barriers with horizontally constructed sub-surface wetlands

In a mine site located in the northern part of Chile, a pilot plant of AMD treatment designed by the company ISMP SpA consisting of LTPRB-HCSSW combined has been installed ideally to secure the water quality of surface waters for a timespan of a few years. It has been widely proved that different types of vegetation have absorption capabilities for different metals such as that shown in Figure 12 which corresponds to Phragmites Australis.

Figure 12.

Picture of the AMD treatment pilot plant system implemented before being covered (a), and close-up to one of the Phragmites Australis used in the HCSSW (b).

The inlet pH was 5.0 and the aqueous solution flowrate is $1\mathit{L}{\mathit{s}}^{\mathit{-}1}$. The pilot plant consisted of 30 m³ effective volumes of LTPRB-HCSSW combined. The system has been designed in a way to promote and enhance the rate at which sulphate reduction reaction [Eq. (29)] is produced triggering as a secondary mechanism the heavy metal precipitation as metal sulphides. To accomplish this, the acidity is provided by introducing metallic scrap into the LTPRB. The dissolution of the metallic scrap by oxygen reduction is inhibited and it is expected to especially be driven by complexation reactions.

$$2{\mathit{CH}}_2{O}_{\left(s\right)}+{\mathit{SO}}_{4\left(\mathit{aq}\right)}^{2-}+2H_3{O}_{\left(\mathit{aq}\right)}^{+}\leftrightarrow H_2{S}_{\left(g\right)}+2{\mathit{CO}}_{2\left(g\right)}+3H_2{O}_{\left(l\right)}{\mathit{xM}}_{\left(\mathit{aq}\right)}^{n+}+y{S}_{\left(\mathit{aq}\right)}^{2-}\to M_x{S}_y$$E28

with

$$\mathit{mx}-2y=0$$

Metal sulphide precipitation, though, is required to be formed as much as possible at pH values where hydrolysis of sulphide ions is low which could be accomplished by evaluating the competitiveness for sulphide complexation within the system. Otherwise, the acidification of the aqueous phase could again take place following Eq. (30).

$$2{\mathit{M}}_{\left(\mathit{aq}\right)}^{\mathit{n}\mathit{+}}\mathit{+}\mathit{n}{\mathit{HS}}_{\left(\mathit{aq}\right)}^{\mathit{-}}\mathit{+}{\mathit{H}}_2{\mathit{O}}_{\left(\mathit{l}\right)}\mathit{\to }{\mathit{M}}_2{\mathit{S}}_{\mathit{n}\left(\mathit{s}\right)}\mathit{+}\mathit{n}{\mathit{H}}_3{\mathit{O}}_{\left(\mathit{aq}\right)}^{\mathit{+}}$$E29

Preliminary results indicate that LTPRB removes sulphate at between 10 and 30 times the rate reported for sulphate removal observed using wetlands only [92, 93]. The HCSSW allowed stabilising of the pH between 6 and 8. Preliminary computations indicate that the volume of control used is about one or two orders of magnitude lower than classic wastewater treatments. All these systems are complex by nature, but they could be engineering-designed from the beginning to enhance/inhibit reactions to avoid AMD. Now, considering the residence time of the AMD flowing through the system, is there any chance to adjust all these mechanisms to act standing alone at the appropriate rates enabling a wastewater treatment to last for a few years by itself keeping as much as possible the permeability of the porous media? This is certainly an opportunity still to be accomplished.

7. Conclusions

Acid Mine Drainage (AMD) formation is yet a process ill-understood. The AMD occurs spontaneously, and it is highly dependent on the local atmospheric conditions which makes it difficult to predict any of its characteristics. Preliminary strategies aiming at forming caps around these solid wastes could be considered a good first step towards preventing the formation of AMD. Nevertheless, in cases where AMD is already formed new strategies for isolating the wastes need to be considered.

Although the precursors of AMD such as sulphide minerals, and notably pyrite, water and oxygen are known to be involved, the physical chemistry and the biology linked to its production need to be studied in more detail and integrated, particularly in long-term reaction kinetics of the different mechanisms taking place.

Several strategies have been suggested to treat AMD. The condition of mine closure takes this challenge to the next level requiring a solution that cannot be intensive in the use of personnel, energy, or reagents.

Strategies involving passive wastewater treatment technologies which attempt to somehow mimic natural systems look promising.

Nowadays, the difference between passive and active wastewater treatment has become a thin line. On one hand, even passive wastewater treatment strategies require to some point the involvement of human resources. On other hand, new long-term permeable reactive barriers have been pointed out as wastewater treatment strategies than can gather several aspects of passive treatment systems such as low maintenance requirements to work and, simultaneously, exhibit fast wastewater treatment kinetics. Some practical aspects associated with implementing long-term permeable barriers coupled with constructed wetlands were presented but improvements with regard to the efficiency of these strategies to remove sulphate, heavy metals and other contaminants are still a matter of study.

Finally, perhaps the most relevant conclusion that can be drawn from this chapter is that addressing AMD generation, prevention or treatment is in fact a multidisciplinary topic where the conjunction of many specialities occurs such as chemistry, physics, hydrology, microbiology, electrochemistry, among others.

Acknowledgments

The author would like to dedicate and acknowledge the contribution to this study of Professor Fernando Valenzuela Lozano. For his invaluable friendship, his continued mentorship and all the technical and inspiring discussions across many years already, regarding hydrometallurgy and especially wastewater treatment of AMD. I would like to specially acknowledge the helpful input of Dr. Marcelo Sepulveda and Mr. Cesar Arredondo to this work. In addition, many thanks to Mr. Mario Solari and Mr. Thomas Ph. Chirino for providing pictures from industrial case scenarios and real-life portraits which had significantly increased the value of this manuscript. And, last but not least many thanks to Ms. Ana Maria Rojo for her assistance and good ideas to put together this chapter.

Conflict of interest

The author declares no conflict of interest.

Notes/thanks/other declarations

Not Applicable.