Enhanced Stability of Scorodite in Oxic and Anoxic Systems via Surface Coating with Hydroxyapatite and Fluorapatite

Abstract

With the objective of enhancing the stability of scorodite, its encapsulation with hydroxyapatite (Ca₅(PO₄)₃OH) (HAP) and fluorapatite (Ca₅(PO₄)₃F) (FAP) surface coatings, the two most stable of the calcium phosphates, inert to pH and redox potential variations, are presented in this work. The experimental work includes: (1) determination of the metastable zone for HAP and FAP precipitation, (2) the synthesis of crystalline scorodite under atmospheric conditions using hydrothermal scorodite seed and its characterization, (3) the coating of scorodite with hydroxyapatite and fluorapatite with supersaturation-controlled heterogeneous crystallization, and (4) the long-term stability of the encapsulated scorodite solids. Hydroxyapatite and fluorapatite were prepared with homogeneous precipitation from a metastable solution to which reagents were added at a controlled flow rate. Crystalline scorodite was produced with seeding precipitation and encapsulated with a direct apatite (HAP or FAP) deposition that was controlled by adjusting the pH and reagent addition. The stability tests in oxic and anoxic environments over the pH range of 5–9 showed the release of arsenic from the apatite-coated scorodite to be much lower than from naked scorodite, thereby demonstrating that apatite-based encapsulation of hazardous materials is technically feasible and merits further consideration for development into an arsenic stabilizing technology.

1. Introduction

The harmful presence of arsenic in waters from both natural and anthropogenic origins is widely known [1]. Despite the number of studies aiming to remove As from contaminated waters, there are still difficulties regarding the low stability of solid arsenic wastes, including the relatively insoluble crystalline scorodite compound (FeAsO₄·2H₂O) whose stability depends on the pH and redox potential (Eh) of the surrounding aqueous solution [2].

There have been increasingly strict environmental regulations [3] aimed at minimizing the concentrations of arsenic in mining and metallurgical industry wastewaters. Worldwide, the limiting value for arsenic in wastewaters discharged into the environment is 0.5 mg/L, and according to the World Health Organization [4] the level of arsenic in drinking water should be limited to 0.01 mg/L. On the other hand, the leachability limit for industrial arsenic wastes (TCLP) is set at 1 mg-As/L [5,6].

Depending on the As concentration in the solution, different approaches may be applied for its removal, including adsorption [7,8] (co-precipitation [8](Jia and Demopoulos 2005), precipitation of various metals (iron (II or III), copper, calcium and manganese), arsenates [9,10], svabite [11,12] and arsenate-hydroxyapatite [13]). More recently, electrochemical methods have also been used [14,15].

The most widely employed method to immobilize As from mineral processing industrial effluents is iron (III) arsenate co-precipitation by concurrent lime addition [16]. This process has been studied and optimized recently by De Klerk and co-workers [17,18], but it is better suited for low As-concentration solutions as it requires an iron (III) to arsenic (V) molar ratio of ≥3 adding to the cost, not to mention the voluminous tailings generated, composed of poorly crystalline ferric arsenate and ferrihydrite with adsorbed arsenate mixed with gypsum [6,19]). However, in the case of effluents, or arsenical wastes in general, that are rich in arsenic, a better option would be to immobilize arsenic in the form of scorodite [8], a crystalline ferric arsenate compound naturally occurring in nature [20]. Scorodite was successfully precipitated in a continuous system using industrial wastewater from the gold industry [21] while Min and co-workers [22] precipitated it from an alkaline liquor of leaching an anode lime rich within their research.

Scorodite has several advantages, such as low iron content since the iron-to-arsenic ratio is 1:1; and it can be easily dewatered, in contrast to amorphous sludge formed by co-precipitation. The product obtained is generally considered as satisfactorily stable for disposal purposes, since it can pass the TCLP limit of [As] <1 mg/L after 20 h of extraction period at pH 5 [5].

Conversely, scorodite undergoes incongruent dissolution at pH values over five, releasing arsenic into the solution as it converts to ferrihydrite (FeOOH) (see Equation (1)) to progressively larger extents as the pH is elevated into the alkaline region. In other words, its stability is strongly influenced by pH, meaning that for pH values in industrial tailings ponds higher than seven, the dissolution of scorodite may release arsenic in the surrounding water above the permissible levels.

FeAsO₄·2H₂O_(s) → FeOOH_(s) + H₃AsO_{4 (aq)}

(1)

Solubility results from Bluteau and Demopoulos [2] show that at pH 7, the equilibrium concentration of arsenic is 5.8 mg/L for hydrothermally prepared crystalline scorodite, which agrees with the value of 5.4 mg/L obtained by Paktunc and Bruggeman [23]) for nanocrystalline scorodite. At higher pH, arsenic concentration increases to unacceptable higher levels (e.g., >80 mg/L at pH 8), hence justifying the need to enhance the stability of scorodite. Under anoxic conditions, at a redox potential in reference to hydrogen (E_h) near or below 250 mV [24], scorodite undergoes reductive decomposition to Fe(II) and As(III), according to Equation (2), leading again to a significant release of arsenic.

FeAsO₄·2H₂O _(s) + 3H⁺ + 3e⁻ → Fe²⁺ + HAsO₃²⁻ _(aq) + 3H₂O

(2)

Recently, Das and Mandal [25] found high As concentrations (from 28.9 to 143 mg/L) in shallow aquifers (<50 m) with a reductive environment in West Bengal, India, as well as in sediments. Reductive conditions (low Eh) may prevail in cases where the deposition of hazardous solid wastes, as in the case of scorodite, is sub-aqueous with the presence of soluble organic carbon or at depths of approximately 2 m below the ground in an anaerobic environment, where reactive metal sulfides may act as a reducing agent [16]. Then, the encapsulation of scorodite with stable compounds may prevent contamination with the release of arsenic into the environment, both in anoxic and oxic conditions.

The enhancement of the stability of scorodite using aluminum phosphate, as well as aluminum hydroxy gels, to minimize arsenic release in alkaline (pH > 7) environments and reduce potential has been evaluated [26,27,28]. Aluminum hydroxy gels prepared by controlled hydrolysis of aluminum solutions with sodium hydroxide were mixed with scorodite particles and left to age. Scorodite particles were encapsulated in an aluminum oxy-hydroxide matrix, which resulted in a reduced arsenic release. A uniform coating of scorodite particles with aluminum phosphate as an encapsulation layer was obtained, but the long-term stability results showed that it was slowly dissolving over time, and therefore, after a certain period, the coating would not offer sufficient protection [28]. After encapsulation, the final solids can be safely disposed of in appropriate industrial landfills for hazardous waste, since they are not commercially used for scorodite in the volumes that are generated.

In this work, hydroxyapatite (HAP) (Ca₅(PO₄)₃OH) and fluorapatite (FAP) (Ca₅(PO₄)₃F), known for their high stability among the calcium phosphates, were used for the encapsulation of scorodite particles using direct crystallization of these phosphates on the scorodite surface. The improvement of the nature of HAP coating and the possibility of forming coatings of fluorapatite around the scorodite particles, taking advantage of the enhanced stability (lower solubility) of fluorapatite compared to hydroxyapatite was evaluated. Aqueous precipitation of hydroxyapatite and fluorapatite was based on supersaturation control of solutions [29] by adjusting the concentration of the ions, pH, and temperature of the system to be within the metastable zone. This metastable region is located between the solubility line and a line representing the critical solute concentration below which heterogeneous nucleation occurs, and growth is favored [29]. The effectiveness of the process is evaluated with appropriate solid characterization and the release of arsenic in oxic and anoxic waters.

2. Experimental Section

The experiments consisted of (i) scorodite synthesis (ii), evaluation of hydroxyapatite (HAP) and fluorapatite (FAP) metastable zones (iii), encapsulation experiments with heterogeneous deposition of hydroxyapatite or fluorapatite under atmospheric conditions of scorodite particles and (iv) stability evaluation of the naked and encapsulated scorodite. All reagents used were of analytical grade, As₂O₅·xH₂O, CaSO₄·2H₂O, HCl, Ca₅(PO₄)₃OH, Ca₅(PO₄)₃OH e Fe₂(SO₄)₃·xH₂O (Sigma Aldrich, St. Louis, MO, USA), NaF (Anachemia, Lachine, QC, Canada), NaOH, Na₂SO_3, H₂SO_4, CaCl₂·2H₂O and CaO (Fisher Scientific, Waltham, MA, USA).

2.1. Preparation and Characterization of Scorodite

Scorodite seeds produced hydrothermally were used to precipitate scorodite under atmospheric conditions, as described in previous work [28]. The preparation of scorodite consisted of mixing solutions of arsenic pentoxide of 40 g/L As and ferric sulfate with a molar ratio Fe/As of 1. The seeds were introduced to the solution and the temperature was kept stable at 85 °C, for 2 h. Further, the kinetics of the precipitation were sped up with the increase in temperature to 95 °C and the system was kept under stirring for 9 h. At the end of the experiment, the solution was left to cool down, filtered, washed, and dried. Both scorodite solids and seeds were characterized by X-ray diffractometry-XRD (Philips PW 1710 X-Ray Diffractometer) (Eindhoven, The Netherlands) with a copper target (Cu Kα1 radiation, λ = 1.5405 Å) and TGA (Thermogravimetric analysis) by Perkin-Elmer thermogravimetric analyzer model TA Instruments Q500 TGA (Waltham, MA, USA). The chemical groups of the solids were obtained by ATR-IR (attenuated reflectance infrared) spectra with a Perkin Elmer (Waltham, MA, USA) (Spectrum BX model) Fourier Transform Infrared (FTIR) spectrometer spectrum BX-model. The particle size distribution was obtained using Horiba LA-920 Laser (Kyoto, Japan) scattering analyzer and solids morphology was observed at SEM (scanning electron microscope Jeol JSM-840A) (Glen Ellyn, IL, USA). In addition, secondary electron microscopy analysis and elemental mapping were also obtained using a variable pressure scanning electron microscope (SEM-type Hitachi S-3000N) (Schaumburg, IL, USA). A Micromeritics TriStar 3000 was used to measure the BET surface area of the scorodite particles.

2.2. HAP and FAP Precipitation and Metastable Width Zone Determination

A series of homogeneous precipitation experiments of hydroxyapatite and fluorapatite were performed, so that the metastable zones of the precipitation reactions could be determined [27,29] For these experiments, solutions with different initial Ca and P concentrations were prepared by dissolving CaCl₂·2H₂O and anhydrous NaH₂PO₄ at a molar ratio of Ca/P = 1.67 and subsequently carefully neutralized with sodium hydroxide. The pH was increased gradually at a rate of 0.1 pH points per 5 min using 0.01 N sodium hydroxide solution. As soon as the solution became cloudy, the pH increase was stopped, and the solution was left to equilibrate for 24 h. The diagrams resulting from these experiments consist of two curves bounding the metastable region; the upper one for the pH and concentration at which precipitation (cloud formation) was first observed, and the lower one for the pH and concentration at pseudo-equilibrium (end of 24-h equilibration period). The experiments were carried out at 25 °C and 40 °C. The same procedure that was followed for hydroxyapatite was also adopted for fluorapatite; the only difference being that F⁻ was also present in the phosphate solution (added as NaF) at molar ratio Ca/F = 5.

2.3. Encapsulation Experiments

Scorodite encapsulation was carried out in the metastable region for the Ca-P-H₂O (HAP precipitation) and Ca-F-P-H₂O (FAP precipitation) systems, for Ca/P = 1.67 and Ca/F = 5, at 22 °C and at a stirring rate of 500 rpm. The process involved separate nucleation and growth steps, which were used in distinct combinations of conditions.

To prepare the surface of the scorodite particles for encapsulation, they were subjected to a pre-equilibration called the conditioning treatment (C) with (i) 5 g/L scorodite in aqueous solution of pH 7.6 (adjusted with NaOH), for 15–30 min under mild stirring on an orbital shaker or (ii) 15 g/L scorodite were conditioned in a solution containing 10 mmol/L of CaCl₂. Solids were separated with vacuum filtration in 22 µm membranes.

The heterogeneous nucleation of apatite coating on scorodite particles (N) was induced by contacting the 5.0 g/L of scorodite with a metastable solution for specific times. Nucleation of HAP was carried out at P: 0.9 mmol/L; Ca: 1.5 mmol/L (Ca/P: 1.67); pH: 7.6 to prevent homogeneous nucleation of calcium phosphates as indicated by the metastable region. Growth of 5.0 g/L of scorodite was carried out with 60 mmol/L P or 100 mmol/L Ca solutions. The stirring rate was varied from 85 to 350 rpm and the number and duration of nucleation cycles varied from 1 to 5 and from 1 to 24 or 48 h, respectively. Experiments were conducted at two temperatures: 22 °C and 40 °C. Between stages, the solids were waited to settle down; the solution was drained, and the subsequent stage was carried out.

The growth step (G) was carried out in two distinct modes (i) by adding the scorodite solids to the calcium phosphate metastable solution and after 30 min of contact, the addition of the two 60 mmol/L P or 100 mmol/L Ca at pre-determined fixed flow rates or (ii) scorodite particles were introduced to the reactor containing the calcium phosphate metastable solution, and after 30 min the Ca and P solutions were added, with flow rates regulated by the pH that was controlled to prevent overshooting the set value that could cause homogeneous precipitation. The concentrations of the feed solutions for the experiments with the flow rate regulated by pH were calculated according to the equations given in the work of Wu and Nancollas [30] as 10.8 mmol/L P and 18 mmol/L NaOH in one solution and 18 mmol/L Ca and 270 mmol/L NaCl in the other solution.

The fluorapatite encapsulation experiments were carried out with 3.5 mmol/L F added to the phosphate solution. The default experimental condition for FAP deposition was pH 7.4, supported by the metastable region obtained from Section 2.2, a temperature of 37 °C and a stirring speed of 500 rpm.

The encapsulation experiments consisted of:

(i)

Encapsulation A (ENCH-A): Scorodite in 9 nucleation steps followed by one step of growth carried out with flow rates of 20 mL/h of 60 mmol/L P plus 100 mmol/L Ca at 22 °C, pH 7.4 for 7 h.

(ii)

Encapsulation B (ENCH-B): Scorodite in 9 nucleation steps followed by growth according to the procedure suggested by Wu and Nancollas [30] at 37 °C pH 7.4; in this case, the addition of constituent elements is not performed at fixed rates but according to pH variation.

(iii)

Encapsulation C (ENCH-C): Previous conditioning of scorodite with 4.25 M CaCl₂ solution at pH 6.0 for 15 h, before the procedure of Wu and Nancollas [30] at 37 °C pH 7.4. Here, a stronger CaCl₂ solution was adopted for conditioning to populate the surface with Ca ions hence facilitating the subsequent nucleation-growth steps (not shown in this paper).

(iv)

Encapsulation D (ENCH-D): Scorodite in two stages of nucleation (N) the first one at pH 6.0–6.5, considering the metastable region of FAP and the second at pH 7.0, and two steps of growth.

(v)

Encapsulation E (ENCH-E): Scorodite was previously conditioned in calcium chloride solution and subjected directly to one step of deposition-growth.

2.4. Stability Evaluation

Evaluation of scorodite leachability under oxic or anoxic conditions at pH 8 and sometimes pH 7 and 9 was monitored over many days at room temperature (22 °C). The anoxic solutions (E_h = 150 or 200 mV) were monitored with a Cole-Parmer^® ORP platinum combination electrode (Ag/AgCl), set up to read directly the E_h, considering the unit conversion factor of +194 mV and a specified probe accuracy of ±40 mV. Additionally, two control stability tests of a physical mixture of scorodite and 15% hydroxyapatite were performed at pH 8 and E_h = 150 mV to verify if the presence of the hydroxyapatite powder in the suspension would affect the arsenic release from the scorodite. The stability tests were conducted in 250 mL Erlenmeyer flasks held in an orbital shaker using a 40:1 liquid-to-solid ratio. Ca(OH)₂ saturated deionized water slurry (14.76 g in 500 mL) and a 5% HCl solution were used to adjust pH, while a 1.4 mol/L Na₃₀₀O₃ solution was used to adjust the E_h to the required anoxic ORP.

For some tests, gypsum was added (4 g CaSO₄·2H₂O₄ in 200 mL) to evaluate possible interaction with the apatite-coated scorodite, such as the one reported by previous work [31] The pH was adjusted regularly within ±0.2 units of the target value and the E_h was measured and adjusted, if necessary, to be ±40 mV of the target value, 150 mV for HAP and 200 mV for FAP. The samples taken from the stability tests were filtered with 0.22 μm pore-size syringe filters and diluted with acidified HNO₃ (~5 wt.%) in deionized water and chemically analyzed by inductively coupled plasma atomic emission spectrophotometry (ICP-AES) for As, Fe, Ca and P concentrations.

In the case of F, Ion Chromatography (IC) was used. Ascorbic acid UV colorimetric method for P with a Perkin Elmer Lambda 20 UV-Vis Spectrometer (Waltham, MA, USA) was used to validate the ICP-AES results or when the on-the-spot analysis of the phosphorus levels was required to control supersaturation and avoid homogeneous precipitation during the encapsulation tests. The experimental strategy is graphically described in Figure 1.

3. Results and Discussion

3.1. Scorodite Preparation

The precipitated scorodite particles were highly crystalline, as indicated by the narrow peaks in the XRD pattern shown in Figure 2a. The FTIR spectroscopy (Figure 2b) was also as expected for scorodite based on previous work [32] The thermo-gravimetric analysis yielded 15.6% of weight loss between 100 °C and 200 °C which is equal to the theoretical value. The BET surface area was 0.755 m²/g and the average particle size was 19.8 μm.

3.2. HAP and FAP Metastable Width Zone Determination

Hydroxyapatite (HAP) is more soluble than fluorapatite (FAP). There is some variation in reported K_sp values due to the use of various methodologies and conditions (Wei et al. 2013). According to [33] K_sp values calculated by using PHREEQC for HAP varied from 10^−53.02 to 10^−53.51 and for FAP values ranged from 10^−55.18 to 10^−56.13 at 25 °C and pH 7.8 [32]. Using distinct initial pHs, [34] observed a reduction in solubility in acidic media. They obtained mean K_sp values for HAP of 10^−57.72 (10^−57.39~10^−58.05) and for FAP of 10^−59.08 (10^−58.65~10^−59.75) at 25 °C at initial pH 2.

In the present work, thermo-dynamic calculations were conducted using the OLI Systems electrolyte simulation software, analyzer version 3.0 (Parsippany, NJ, USA), and the respective solubility products were calculated as well. The values obtained are within the range observed in the literature. On the pre-supposition that these compounds have very low solubilities, the activity coefficients of the ions in an aqueous solution (pH = 7.0) were assumed to be very close to one, giving the K_sp for HAP and FAP as per Equation (3) and Equation (4), respectively.

K_{sp, 25 °C (HAP)} = [Ca²⁺]⁵ [PO₄³⁻]³[OH⁻] = 10^−56.55

(3)

K_{sp, 25 °C (FAP)} = [Ca²⁺]⁵ [PO₄³⁻]³[F⁻] = 10^−59.54

(4)

Additionally, the knowledge of the metastable zone width of HAP and FAP was essential to designing the precipitation experiments to favor heterogeneous nucleation instead of homogeneous nucleation. The metastable zones of the apatite precipitation compounds defined in terms of phosphorus concentration (concentration vs. pH) are presented in Figure 3. Based on these curves, the direct deposition of HAP/FAP on scorodite was attempted at either pH 7.6 with a starting P concentration of 0.9 mmol/L but not exceeding 1.2 mmol/L during the experiments, or at pH 7.0, with a starting P concentration of 2 mmol/L but not exceeding 2.5 mmol/L during the experiments. As depicted, the zones for HAP and FAP are similar, however, fluorapatite tends to precipitate in advance at a lower pH value for the same phosphorus concentration. Therefore, the conditions for FAP deposition experiments were kept as those used for HAP using three feed solutions, in this case, namely NaH₂PO₄ 30 mmol/L, CaCl₂ 50 mmol/L and NaF 10 mmol/L.

The nucleation step consisted of immersion of the scorodite particles into the metastable solution for several hours to induce the apatite nucleation on the foreign surface of the scorodite. As the induction period for such nuclei formations proved rather long, it was decided to immerse the same substrate in more than one fresh metastable solution, i.e., to try multiple nucleation steps with repeated contacts. In this way, the creation of an initial nucleation layer was hoped to form on the surface of the scorodite particles, making them amenable to encapsulation by the subsequent growth stage. The growth stage consisted of a semi-batch experiment where the apatite surface-nucleated scorodite particles were transferred to a metastable solution in which the constituent elements (Ca, P plus F in the case of FAP) feed solutions were added at a pre-determined flow rate. Note that prior to starting the addition of the feed solutions into the crystallization reactor, the solids were allowed 30 min for additional conditioning.

3.3. Encapsulation

3.3.1. Encapsulation of Scorodite with Hydroxyapatite

With the aim of preventing homogeneous nucleation, the flow rate of the reagents was set to 6 mL/h during the first 24 h and then increased to 20 mL/h (aleatory value) in the subsequent five steps. However, using this flow rate, the critical concentration of P of 1.2 mmol/L (see metastable zones in Figure 3) was exceeded, causing homogeneous nucleation and precipitation away from the scorodite particle surfaces. As a result, it was decided to abandon this type of nucleation procedure in favor of multiple contacts to the scorodite particles with fresh calcium phosphate solution.

Scorodite particles were brought into contact with a series of five calcium phosphate metastable solutions, for one hour at a time. A very small decrease in the concentration of P and Ca was noticed during these steps, however, the analysis of SEM images indicated that some precipitation had happened away from the scorodite particle surfaces, even at a longer time of 48 h. The effects of stirring speed and temperature were evaluated. During these tests, it was noticed that phosphorus and calcium were removed from the solution as soon as the scorodite was introduced with the amount removed, increasing with increasing temperature and decreasing stirring speed, as seen in Figure 4.

Low agitation speed and higher temperature (higher solubility, therefore lower supersaturation) stimulated the heterogeneous precipitation of calcium phosphate on the surface of scorodite and these conditions were adopted for the rest of the work. It must be noted, however, that when the solids were examined by SEM it was not possible to observe any surface layer nor any homogeneously precipitated particles. The latter, i.e., the absence of dispersed HAP particles was taken as evidence of successful heterogeneous nucleation although such surface layer could not be detected due to a very low amount of calcium phosphate coating.

In subsequent experiments, multiple nucleation steps (9) were used to obtain a better coverage of the scorodite particles by HAP, hence facilitating the subsequent growth stage. At this time, the operation point was changed to a higher supersaturation, 2 mmol/L P, 3.34 mmol/L Ca and pH 7.0, but still operating within the metastable zone, as shown in Figure 4. Therefore, to complete the HAP deposition, three approaches as described in Section 2.3 for growth were adopted: ENCH-A, ENCH-B and ENCH-C.

The SEM images of the resulting scorodite@HAP solids obtained with the encapsulation A and encapsulation B methods are presented in Figure 5. From ENCH A, the scorodite particles seem to be only partially covered by hydroxyapatite, while still a large amount of hydroxyapatite is seen to have formed away from the scorodite surface due to homogeneous precipitation. Additionally, there were small hydroxyapatite particles attached to the bigger scorodite particles. FTIR analysis of the solids (not shown in this paper) from the encapsulation A approach identified the characteristic features of hydroxyapatite and scorodite but could not distinguish between homogeneously and heterogeneously nucleated HAP. Particle size analysis was not conducted after encapsulation, hence the layer deposited was very narrow.

Encapsulation method B (ENCH-B) allowed for a constant composition environment to be maintained, i.e., it enabled a better control of supersaturation, and as a consequence, there was an improvement in the formation of scorodite@HAP solids. The SEM cross-section images of the encapsulated material revealed the majority of the scorodite particles to have been partially covered with hydroxyapatite; in addition, there was a significant occurrence of agglomerates of hydroxyapatite and scorodite sub-particles.

In the subsequent approach of encapsulation C, it was hypothesized that by conditioning the scorodite particles in a concentrated CaCl₂ solution (4.25 mol/L), the particle surface/inner side of the electrical double layer will be enriched with Ca²⁺ hence, creating a favorable localized environment for phosphate anions to approach (during the subsequent deposition test) and form in situ HAP nuclei. The affinity of Ca²⁺ ions to the surface of scorodite is supported further by the isoelectric point (IEP) of scorodite, determined by the zeta-potential measurements to be at pH = 3.9. That means the scorodite surface is negatively charged, which favors electrostatic attraction of Ca²⁺ at precipitation pH ~ 7.5.

The conditioning step involved equilibration for 15 h of 15 g of wet scorodite in 1 L of 4.25 M CaCl₂ solution. Indeed, the calcium concentration dropped to 3.77 M Ca at the end of this conditioning step, apparently due to adsorption or surface precipitation of a calcium arsenate as described in the literature [10,22] The subsequent deposition (growth) step (32 h) was found to have a short induction time in comparison to the deposition growth on naked scorodite, which had not been previously conditioned with a CaCl₂ solution other than immersing it in pH 7.6 water. After 32 h, the feed solutions were completely consumed. This means that conditioning with the calcium solution had effectively prepared the particle surface for extensive heterogeneous nucleation and growth of hydroxyapatite to take place, thus, favoring the formation of well-encapsulated scorodite@HAP material as confirmed by SEM examination, as shown by SEM in Figure 6. Clearly, heterogeneous deposition of HAP on scorodite took place. Elemental X-ray mapping of one of the cross-sections shows that the layer covering the scorodite particle is indeed a calcium phosphate compound. Additionally, the images reveal that the layer has not grown uniformly over the whole particle surface, and further studies to optimize the growth of the HAP coating are required.

3.3.2. Encapsulation of Scorodite with Fluorapatite

Having determined in the previous series of tests that conditioning of the scorodite particles in CaCl₂ solution is beneficial to subsequent heterogeneous deposition, the same approach was adopted to evaluate fluorapatite as a coating material. After scorodite conditioning in calcium chloride solution, two routes were investigated for fluorapatite precipitation: two stages of nucleation (N), the first one at pH 6.0–6.5, considering the metastable region of FAP in Figure 4, and the second at pH 7.0, as described in Experimental Section, and two steps of growth labeled as encapsulation D (ENCH-D). The decrease in the concentrations of Ca, P and F during the two nucleation steps of ENCH-D signalized that nucleation of FAP had taken place.

In the other route, nucleation was bypassed, and scorodite that was previously conditioned in calcium chloride solution was subjected directly to one step of deposition-growth; this strategy is specified as “Encapsulation E”. The successful coating of scorodite particles with FAP was evaluated with SEM with images presented in Figure 7. In this image, the FAP-coated scorodite particles after the growth step following conditioning with calcium chloride solution are depicted. The darker layers around some particles indicate coating formation, substantiated by the backscattered electron image and the corresponding elemental maps revealing phosphorus and calcium rims, confirming a core-shell type scorodite@FAP formation. The presence of fluorine was not detected due to limitations of the technique regarding light elements. In the meantime, due to small amounts of FAP coating, XRD and IR spectra could not detect the presence of FAP. The employment of the encapsulation E method that involved only conditioning in CaCl₂ solution and growth-deposition of FAP without an in-between nucleation stage proved successful as it could be validated with the SEM images, which revealed a well-developed phosphate coating on scorodite particles, thus, proving the effectiveness of calcium chloride conditioning.

3.4. Stability in Oxic and Anoxic Environments

The variables pH and E_h were measured and adjusted regularly to the set point during the stability tests. However, the variations in E_h were not easy to control. The value for anoxic experiments with HAP coating is 150 ± 40 mV and for FAP is 200 ± 40 mV. As they are related variables, a change in pH modifies E_h and vice-versa. In industrial practice, the pH is usually adjusted with time, and since sulfate ions are common in the wastewater, the presence of gypsum in the system must be considered. Due to this fact, the release of arsenic in the presence of gypsum was measured.

3.4.1. Stability of HAP-Encapsulated Scorodite

The release of arsenic in the oxic condition, as shown in Figure 8, was reduced significantly upon coating with HAP (ENCH-A) when compared to that of the naked scorodite. It is clear, therefore, that the surface deposition of a hydroxyapatite layer has a protective effect on scorodite, especially in the alkaline (pH > 7) region, where scorodite is the least stable. Thus, while at pH 9 the naked scorodite sample released 0.29 mmol/L (22 mg/L) of As, the coated material released only 0.014 mmol/L (1.1 mg/L) of As. This value is slightly lower in the presence of gypsum, giving a value of 0.011 mmol/L (0.8 mg/L) of As after 40 days of equilibration, apparently reflecting the common ion effect on the solubility of HAP. In addition to arsenic, the release of P was monitored, and it varied from 0.0098 mmol/L (0.28 mg/L) at pH 9 to 0.21 mmol/L (6.5 mg/L) at pH 7. This was slightly reduced in the presence of gypsum: 0.21 mmol/L (6.5 mg/L) vs. 0.076 mmol/L (2.3 mg/L) at pH 7. The Fe concentration was found to be very low, below the detection limit (0.1 mg/L) for all samples, which was expected, since at these pH values iron precipitates as ferrihydrite.

The results in anoxic conditions are plotted in Figure 9. After 20 days, in general, the release levels were stabilized, signaling near pseudo-equilibrium attainment. In an anoxic environment of 150 ± 40 mV, the naked scorodite sample was found to release 2.3 mmol/L As. Encapsulation A and B produced a scorodite@HAP that releases 60% less arsenic than the naked scorodite. These procedures were slightly better protected than ENCH-C, since the arsenic release at pH 8 was ~0.9 mmol/L for ENCH-A and ENCH-B versus 1.2 mmol/L As for ENCH-C. At pH 9.0, ENCH-A yielded a 30% decrease in As release, from ~1.3 mmol/L (110 mg/L) to 0.395 mmol/L (30 mg/L).

However, this level of protection is not considered adequate given that the environmentally acceptable level of the typical leachability tests, such as EPA’s TCLP, is required not to exceed 5 mg/L or, in other jurisdictions, 1.0 mg/L As. The underlying cause of the significantly different responses of the HAP-coated scorodite to the oxic and anoxic environments is not clear. It may be due to the incomplete growth of a non-porous continuous layer of HAP on the surface of all scorodite particles, allowing for the sulphite-reducing species to reach the surface and initiate degradation of part of the scorodite via the reduction in Fe(III) to Fe(II). This hypothesis seems to be supported by the relatively higher release of arsenic from the physical mixture of scorodite with hydroxyapatite (sample “Scoro+HAP ref”) vis-à-vis the coated samples.

Additionally, some degree of solubilization of the deposited HAP was observed, depending on the kind of deposition method used. The release of phosphorus was reduced following the order ENCH-A > ENCH-B > ENCH-C. This again points to the beneficial effect of the calcium chloride conditioning step in facilitating the deposition of a robust HAP coating layer.

3.4.2. Stability of FAP-Encapsulated Scorodite

A new batch of scorodite was created for FAP encapsulation tests, for this reason, the stability results were compared with the new sample, which had essentially similar release characteristics as the previous one.

Table 1. Summary of released (mmol/L and in mg/L) data after 6 and 10 days in oxic (O) and anoxic (A, 200 ± 40 mV) environment at 22 °C from naked scorodite and scorodite@FAP at different pHs.

	pH 7		pH 9
	Oxic	Anoxic	Oxic	Anoxic
Naked Scorodite
6 days (mmol/L)	0.094	0.291	0.210	1.009
6 days (mg/L)	7.04	21.80	15.7	75.59
10 days (mg/L)	8.99	23.44	nd	nd
20 days (mg/L)	10.24	nd	20.22	n.d.
ENCH-D-Sco@FAP
6 days (mmol/L)	0.057	0.177	0.057	0.138
6 days (mg/L)	3.99	13.28	4.26	10.38
10 days (mg/L)	4.24	16.74	5.10	11.22
20 days (mg/L)	3.21	20.66	6.67	12.83

summarizes the stability data for FAP-encapsulated scorodite and its comparison to that of naked scorodite. It is important to highlight that scorodite encapsulated with FAP according to the encapsulation method E has not been effective in protecting the scorodite, despite the layer of calcium phosphate being well evidenced (figures not shown in this paper). It is possible that, due to fast precipitation, the deposited layer was very porous due to poor heterogeneous deposition and inadequate protective FAP layer formation. In contrast, the encapsulation method D, which also involved preconditioning of scorodite in a concentrated calcium chloride solution, proved effective as it enabled the formation of a good FAP coating.

As can be verified with the data in Table 1, the FAP-encapsulated scorodite (obtained with the encapsulation D method) was released after 20 days, under oxic conditions and pH 9.0, with three times less arsenic, 0.089 mmol/L (6.67 mg/L) than the unprotected scorodite, 0.27 mmol/L (20 mg/L), and seven times less under an anoxic condition in 6 days (76 mg/L versus 10 mg/L). Thus, it is deduced that the fluorapatite layer effectively coated the surface of scorodite, and the observation is consistent with the elemental mapping shown in Figure 9.

The respective numbers after 6 days of testing in an oxic system at pH 7.0 were 0.094 mmol/L (7 mg/L) As for naked scorodite and 0.076 mmol/L (5.7 mg/L) As for the HAP encapsulated material, and a similar result of 0.053 mmol/L (4.0 mg/L) for scorodite@FAP. By comparison, in an oxic system at pH 9.0, the HAP-coated scorodite yielded 1.1 mg/L As, i.e., lower than that of the FAP-coated material (4.3 mg/L), apparently reflecting better continuous growth of the HAP layer. On the other hand, under an anoxic condition at pH 9.0, the concentration of released arsenic for the FAP-encapsulated materials is only 14% (0.14 mmol/L) of that of naked scorodite (1.0 mmol/L), confirming once again the effectiveness of the FAP layer. The corresponding value for the HAP-coated material was significantly higher (1.5 mmol/L). The origin of this discrepancy may be attributed more to the reducing E_h (150 vs. 200 mV), in which the stability was carried out for the HAP-coated scorodite.

In summary, both HAP and FAP materials bring a significant reduction in arsenic release when deposited on scorodite, especially under the oxic condition. The results at pH 9/oxic conditions with ≤5 mg/L As release (only ~1 mg/L for HAP vs. ~5 mg/L for FAP) vis-à-vis that of a value >15 mg/L for unprotected scorodite clearly demonstrate the encapsulation potential of these systems. In terms of anoxic condition stability, the FAP system looks more promising than the HAP system. Further work to optimize the coating on one hand and evaluate other reducing agents—both abiotic and biotic—should be pursued, to acquire further insight into the mechanism and build a totally robust encapsulation system.

4. Conclusions

The encapsulation of scorodite via direct precipitation of apatite (hydroxyapatite-HAP and fluorapatite-FAP) coatings and the stability of the products were studied under oxic and anoxic environments. Encapsulation experiments of scorodite were conducted by testing different combinations of procedures to enable the deposition of HAP/FAP on the scorodite particle surface via the control of supersaturation, and thus promote the formation of a protective coating layer. Heterogeneous nucleation of calcium phosphates proved to be difficult. After testing various procedures, prior conditioning of the scorodite substrate in a calcium solution created favorable nucleation conditions, both for HAP and FAP growth. Low agitation (500 rpm) and high temperature (37 °C) also proved beneficial, as did the adaptation of the constant composition procedure of Wu and Nancollas (1997) for the deposition of hydroxyapatite on titanium dioxide particles to the encapsulation of scorodite. The coating layers formed on the particles were not as thick and continuous as desired, an issue that should be addressed in future research. The long-term stability of the encapsulated solids, as well as the scorodite alone and a simple physical mixture with hydroxyapatite as controls, were tested, and the results clearly demonstrated that, in principle, protection of scorodite via coating with hydroxyapatite/fluorapatite surface layers is possible.

In all cases, the solids after the encapsulation experiments behaved much better than the scorodite control sample, either under oxic or under anoxic conditions in the alkaline region. By far the best results were obtained with the HAP-coated system under oxic conditions, yielding < 1 mg/L As release after 20 days at pH 9, compared to 6.6 mg/L for the FAP-coated material and 20 mg/L for the unprotected scorodite. In the meantime, the FAP-coated material was found to be more stable than the HAP-coated scorodite when in a reducing environment adjusted with sulfites, but further work with this system is required. Some of this work may involve the evaluation of biotic reducing agents that better simulate anoxic conditions in disposed waste sites. Finally, another positive feature of the tested encapsulation systems is the fact that the coatings released very low P levels, only 0.2–0.3 mg/L for both HAP and FAP; this level decreased even further in the presence of gypsum due to the common-ion effect.