Microbial Oxidation of Fe²⁺ and Pyrite Exposed to Flux of Micromolar H₂O₂ in Acidic Media

Abstract

At an initial pH of 2, while abiotic oxidation of aqueous Fe²⁺ was enhanced by a flux of H₂O₂ at micromolar concentrations, bio-oxidation of aqueous Fe²⁺ could be impeded due to oxidative stress/damage in Acidithiobacillus ferrooxidans caused by Fenton reaction-derived hydroxyl radical, particularly when the molar ratio of Fe²⁺ to H₂O₂ was low. When pyrite cubes were intermittently exposed to fluxes of micromolar H₂O₂, the reduced Fe²⁺-Fe³⁺ conversion rate in the solution (due to reduced microbial activity) weakened the Fe³⁺-catalyzed oxidation of cubic pyrite and added to relative importance of H₂O₂-driven oxidation in the corrosion of mineral surfaces for the treatments with high H₂O₂ doses. This had effects on reducing the build-up of a passivating coating layer on the mineral surfaces. Cell attachment to the mineral surfaces was only observed at the later stage of the experiment after the solutions became less favorable for the growth of planktonic bacteria.

Introduction

Pyrite (FeS₂) commonly occurs in the Earth's surface environments¹. Great efforts have been made for the past decades to understand the mechanisms and kinetics of pyrite oxidation^{2,3,4,5,6,7,8,9,10,11,12,13,14}, which is the driving force leading to the widespread ecological degradation caused by the formation of acid mine drainage (AMD) and acid sulfate soils (ASS)^15,16,17. The established model to describe pyrite oxidation consists of the following overall chemical equations:

Initially, oxidation of pyrite by molecular oxygen results in generation of sulfuric acid and ferrous sulfate (Equation 1). This process only completes the conversion of pyrite-S from S⁻ to S⁶⁺. The aqueous Fe²⁺ of pyrite origin may further react with molecular oxygen to form Fe³⁺ (Equation 2), which subsequently acts as a more effective oxidant for pyrite oxidation (Equation 3) if the pH is sufficiently low to prevent Fe³⁺ hydrolysis. Abiotic oxidation of Fe²⁺ is very slow under acidic conditions but the process can be catalyzed by iron-oxidizing bacteria³.

Hydrogen peroxide (H₂O₂) is an aggressive oxidant and it is commonly present in rainwater^18,19,20, especially during thunderstorms²¹. H₂O₂ can also be generated through radiolysis of water in geological formations or industrial wastes containing radioactive materials such as uranium ores, spent nuclear fuel etc^22,23. Therefore, contacts between H₂O₂ and pyrite grain surfaces or/and aqueous Fe²⁺ are likely to occur in natural environments. Moreover, H₂O₂ has potential applications in hydrometallurgy for the extraction of base metals from sulfide ores^24,25,26 and in desulfurization of coal²⁷. There is therefore a possibility for the leaking/residual H₂O₂ to be in contact with pyrite in the non-target areas surrounding the heap leach piles. Furthermore, it was also suggested that H₂O₂ could be spontaneously generated when pyrite reacted with H₂O in the absence of oxygen²⁸ or during the oxidation of pyrite with molecular oxygen^29,30. From a planetary perspective, the presence of atmospheric H₂O₂ in other planets such as Mars may represent an important factor that has been affecting the acid sulfate-producing processes in Martian environments³¹.

It is well established that powerful oxidative intermediates (free radicals) are generated as a result of the chain reactions initiated by contact between H₂O₂ and Fe²⁺ i.e. so-called Fenton reactions^32,33. Therefore, H₂O₂ may have a dual role to play: (a) as a strong oxidant for abiotic oxidation of pyrite and aqueous Fe²⁺ and (b) as a toxicant (through the generation of free radicals) to cause oxidative stress in iron/sulfide-oxidizing bacteria and consequently affect their ability to catalyze Fe²⁺ oxidation.

Surface oxidation of pyrite by millimolar H₂O₂ (2–200 mmol L⁻¹) in molecular oxygen-free conditions has been examined by Lefticariu et al^34,35 in abiotic systems. The role of H₂O₂ spontaneously generated from oxidation of powdered pyrite by molecular oxygen was investigated by Schoonen et al³⁰. In AMD and ASS scenarios, it is hardly possible that iron/sulfide-oxidizing bacteria are not involved in the process of pyrite weathering. Therefore, while the abiotic oxidation research provides a fundamental basis for understanding the role of H₂O₂ in pyrite oxidation, it has some limitations in explaining complex systems involving microbially mediated processes.

We previously examined the oxidation of pyrite cubes that were exposed to an Acidithiobacillus ferrooxidans strain (an iron- and sulfide-oxidizing bacterial species commonly present in natural environments), ambient dissolved oxygen and intermittent fluxes of H₂O₂ at micromolar levels under circumneutral (initial pH = 6.8). It was found that Equation 3 did not operate observably due to low Fe³⁺ solubility. The more aggressive nature of H₂O₂ as an oxidant (as compared to molecular oxygen) caused marked surface corrosion of the pyrite cubes. It was observed that cell colonization on the mineral surfaces was inhibited at high H₂O₂ levels (initial concentration of H₂O₂ >100 μmol L⁻¹). However, at a H₂O₂ concentration of 50 μmol L⁻¹, cell attachment was enhanced, as compared to the no-added H₂O₂ treatment³⁶.

In this article, we report on the experimental results for the acidic scenario (initial pH = 2). This represents a step forward in simulating pyrite weathering involving H₂O₂ in field conditions e.g. pyrite grains in tailings dams, inundated mine waste impoundments and constructed wetlands that receive cyclic flux of micromolar H₂O₂ caused by intermittent rainfall events. Pyrite cubes instead of powdered pyrite were used in this study in order to minimize the possible interference from the spontaneously generated H₂O₂, as reported by Schoonen et al³⁰, who found that it was a challenge to achieve quantitative measurements of the spontaneously generated H₂O₂. To obtain further insights into the response of planktonic Acidithiobacillus ferrooxidans to toxicity of Fenton reaction-derived free radicals and the resulting impacts on the oxidation rate of aqueous Fe²⁺, a separate experiment was also conducted to observe the evolution of cell population, aqueous Fe²⁺ and other relevant chemical parameters following a single H₂O₂ flux. The objective of this study was to examine the effects of micromolar H₂O₂ flux on the microbially involved oxidation of aqueous Fe²⁺ and cubic pyrite by comparison between the control (no added H₂O₂) and the treatments (with varying dosage levels of H₂O₂) under the same preset experimental conditions.

Results

Aqueous Fe²⁺ oxidation experiments

For the abiotic experiments, the concentration of aqueous Fe²⁺ remained little change when exposed to molecular oxygen only (i.e. the control) for both the low and high Fe²⁺ scenarios. The injection of H₂O₂ (the treatments) caused a sudden decrease in Fe²⁺ within the first 5 min of the experiment. The aqueous Fe²⁺ then maintained at a stable level throughout the entire duration of the experiment after this initial drop. There was a tendency that the magnitude of Fe²⁺ drop increased with increasing dosage level of H₂O₂ (Fig. 1a and c).

Figure 1

Evolution of aqueous Fe²⁺ in the solution when exposed to micromolar concentrations of H₂O₂.

(a) abiotic oxidation experiment at an initial Fe²⁺ concentration of 55.8 mg L⁻¹, (b) biotic oxidation experiment at an initial Fe²⁺ concentration of 55.8 mg L⁻¹, (c) abiotic oxidation experiment at an initial Fe²⁺ concentration of 558 mg L⁻¹ and (d) biotic oxidation experiment at an initial Fe²⁺ concentration of 558 mg L⁻¹. C: control; T1: 50 μmol L⁻¹ H₂O₂; T2: 100 μmol L⁻¹ H₂O₂; T3: 300 μmol L⁻¹ H₂O₂; and T4: 1000 μmol L⁻¹ H₂O₂.

In contrast with the abiotic experiment, the concentration of aqueous Fe²⁺ in the control (no added H₂O₂) rapidly decreased in the presence of Acidithiobacillus ferrooxidans for both the low and high Fe²⁺ scenarios (Fig. 1b and d). For the low Fe²⁺ scenario, the initial rapid drop in Fe²⁺ was observed for all the treatments. The aqueous Fe²⁺ evolution patterns in T3 and T4 were similar to those in T3 and T4 of the abiotic experiment, respectively. However, the aqueous Fe²⁺ in T1 and T2 continued to decrease after the initial drop with T1 showing a generally rapider rate of reduction, relative to T2 (Fig. 1b). When the initial concentration of Fe²⁺ in the solutions was increased to 558 mg L⁻¹, there was no significant difference in the aqueous Fe²⁺ evolution pattern among the control, T1 and T2. T3 also had a very similar evolution pattern aqueous Fe²⁺ to the control except that it dropped much quickly during the first 5 min and took a slightly longer time to disappeared in the solution. T4 was the only treatment showing marked delay in aqueous Fe²⁺ depletion (Fig. 1d).

The curves of viable cell population during the incubation experiments for both the low and high Fe²⁺ scenarios are provided in Supplementary Fig. S1a and b. In general, the cell density decreased with increasing dosage level of H₂O₂. Fig. S2 is a selected SEM image showing the cellular damage in the affected microbes.

In the low Fe²⁺ scenario, there was a high degree of similarity in the curves of dissolved oxygen (DO) among the control, T1, T2 and T3. However, T4 showed much higher DO in the earlier stage of the incubation experiment, as compared to the others. This gap was closed at the 40^th h of the experiment (Supplementary Fig. S3). The curves of DO were highly consistent among the control and all the treatments for the high Fe²⁺ scenario (data not shown).

Toxic response of Acidithiobacillus ferrooxidans to H₂O₂ under various conditions

Figure 2 shows that when the bacteria were exposed to 50 μmol L⁻¹ H₂O₂ for 30 min (BH), the cell population in the solution significantly reduced, as compared to the control (BI, without added H₂O₂) and in the treatment (BHA) with a reactive oxygen species scavenger (ascorbic acid), which has the capacity to rapidly decompose H₂O₂. When comparing H₂O₂ only systems (i.e. BH and BIH) with H₂O₂-Fe²⁺ reaction systems (i.e. BIH and BIHA), it can be seen that the cell population was denser in a H₂O₂ only system than in its corresponding H₂O₂-Fe²⁺ system (Fig. 2), indicating that more severe oxidative stress in the cells in the latter than in the former.

Figure 2

Comparison of cell population among different treatments.

Acidithiobacillus ferrooxidans were exposed to H₂O₂ (50 μmol L⁻¹) for 30 min. BI: without added H₂O₂; BH: 50 μmol L⁻¹ H₂O₂ only; BIH: 50 μmol L⁻¹ H₂O₂ and 55.8 mg L⁻¹ Fe²⁺; BHA: 50 μmol L⁻¹ H₂O₂ and 50 μmol L⁻¹ ascorbic acid; and BIHA: 50 μmol L⁻¹ H₂O₂, 55.8 mg L⁻¹ Fe²⁺ and 50 μmol L⁻¹ ascorbic acid. Error bars represent standard deviation of five replicates.

Aqueous phase of the pyrite oxidation experiment

The viable cell count on the 6^th day after inoculation of Acidithiobacillus ferrooxidans was lower than the number of cells added at the time when the inoculum was introduced into the solution for the control and all the treatments. The total number of viable planktonic cells was in the following decreasing order: CP > TP1 = TP2 > TP3 > TP4 on this occasion. Therefore, approximately 35% and 91% of the added bacteria disappeared from the solution for the control and the highest H₂O₂ dosage treatment (TP4), respectively. After this initial drop in planktonic cell population, the number of planktonic cells in various treatments fluctuated with different evolutionary patterns during the period of the incubation experiment. In general, the cell count was higher in CP and TP1 than in the treatments with higher H₂O₂ doses (TP2, TP3 and TP4). TP4 had the lowest cell count at any sampling occasion during the period of the experiment. The cell population in the control and the treatments consistently dropped to the lowest level on the 74^th day and remained low until the end of the incubation experiment (Fig. 3a)

Figure 3

Changes in (a) planktonic cell population, (b) ferrous ion concentration, (c) ferric ion concentration, (d) sulfur concentration and (d) pH in the solution during the period of incubation experiment for the control and various treatments.

On the 3^rd day of the experiment, the concentration of aqueous Fe²⁺ was in the following decreasing order: CP > TP4 > TP1 > TP2 > TP3. The Fe²⁺ concentration in TP1, TP2 and TP3 maintained at a level <5 mg L⁻¹ during the entire period of the experiment. After the sudden drop from 9.6 mg L⁻¹ on the 3^rd day to 4.5 mg L⁻¹ on the 6^th day, the Fe²⁺ concentration in CP also maintained at a level <5 mg L⁻¹ for the remaining period of the experiment. In contrast, TP4 exhibited a different Fe²⁺ variation pattern during the period of the experiment; the concentration of Fe²⁺ sharply increased from 8.3 mg L⁻¹ on the 3^rd day to 26.6 mg L⁻¹ and 26.7 mg L⁻¹ on the 12^th and 15^th day, respectively and then rapidly decreased to 4.6 mg L⁻¹ on the 22^nd day; after this, the Fe²⁺ concentration maintained at a level <6 mg L⁻¹, with fluctuation, until the end of the experiment (Fig. 3b).

The aqueous Fe³⁺ had a similar temporal variation pattern for the control and all the treatments, showing a general trend to increase the concentration over time. However, Fe³⁺ concentration in TP4 was always much lower, relative to the control and other treatments on any sampling occasions and no Fe³⁺ was detected during the initial 19 days of the experiment. In the early part of the experiment, the aqueous Fe³⁺ concentration tended to be markedly higher in CP and TP1 than in TP2 and TP3. However, this gap was gradually reduced in the latter part of the experiment (Fig. 3c).

There was a clear tendency that the aqueous sulfur concentration increased with increasing incubation time. This was accompanied by an opposite trend for the solution pH. Closer examination found that on most sampling occasions, TP4 had a higher pH and lower S concentration, as compared to the control and other treatments (Fig. 3d and e).

Scanning electron microscope (SEM) observations and X-ray photoelectron spectroscopy (XPS) analysis of the pyrite cube surfaces

No attached cells were observed on the surface of pyrite cubes that were taken on the 28^th day of the experiment for the control and various treatments. Pitting corrosion was clearly observable for CP and TP1 but not for other treatments (Fig. 4a, d, g, j and m). The micro-morphological features of the reacted pyrite surfaces were markedly different at the end of the experiment, as compared with those observed on the 28^th day. Except for TP4, all the treatments and the control exhibited marked surface cracking and partially flaking corrosion (Fig. 4b, e, h, k and n). Attached cells were clearly observed to occur on the flakes that were still tightly or loosely attached to the pyrite substrate for TP1, TP2 and TP3 (Fig. 4f, i and l). Scattered cell-shaped objects were also observed on the mineral surfaces for CP and TA4 (Fig. 4c and o), but they were not as clear as those observed for TP1, TP2 and TP3.

Figure 4

SEM images showing the micro-morphological characteristics of the reacted surfaces of pyrite cubes taken on the 28^th day of the experiment.

(a) control (CP), (d) TP1, (g) TP2, (j) TP3 and (m) TP4; pyrite cubes taken at the end of the experiment (the 86^th day): (b) control (CP), (e) TP1, (h) TP2, (k) TP3 and (n) TP4; (f), (i) and (l) show the presence of attached cell clusters on the reacted mineral surfaces (the 86^th day) for TP1, TP2 and TP3, respectively and the arrows point to such example clusters in these three treatments; (c) and (o) point to the scattered cell-shaped objects on the reacted mineral surfaces (the 86^th day) for the control (CP) and TP4, respectively.

The XPS results showed that within a ~3–5 nm thick surface layer of the reacted mineral surface, oxygen accounted for a large proportion (52–79% on a molar basis) of the sum of iron, sulfur and oxygen. There was a trend that the oxygen percentage decreased with increasing dosage level of H₂O₂ in the solution (Fig. 5a). The Fe/S ratio of the reacted surfaces ranged from 0.183 to 0.257, which was much lower than the value (0.5) of the theoretical Fe/S ratio for pyrite. After treatment of the reacted pyrite cubes with the boiling HCl to remove the oxidized materials, the proportion of oxygen in the sum of iron, sulfur and oxygen markedly decreased for the outermost layer (top ~3–5 nm thick) of the corroded surface (Fig. 5b). The Fe/S ratio of the corroded surface ranged from 0.173 to 0.239, which was not markedly different from those prior to HCl treatment.

Figure 5

Comparison of chemical composition (normalized to oxygen, sulfur and iron) in (a) the reacted pyrite cube surface and (b) the corroded pyrite cube surface among the control and various treatments.

There was a trend that XPS peaks of Fe 2p_3/2 and S 2p shifted to lower binding energies with increasing dosage level of H₂O₂, indicating that higher level of H₂O₂ resulted in increased proportion of lower-valence species of Fe and S on the reacted pyrite surfaces (Table 1). CP and TP1 exhibited the dominance of Fe³⁺ (in Fe-OH bond) in the surface Fe species. There was a trend that the percentage of Fe³⁺ decreased with increasing H₂O₂ level. Fe²⁺ (in Fe-S bond) was not detected for CP and TP1but dominated the surface Fe species for TP2, TP3 and TP4. In agreement with surface Fe speciation, S⁻ dominated the surface S species for TP2, TP3 and TP4 while CP and TP1 only contained small amount of surface S⁻. The percentage of surface sulfate-S was greater in CP and TP1 than in TP2, TP3 and TP4. Small amount of thiosulfate-S was also detected for CP and TP1.

Table 1 Chemical state of iron and sulfur on the reacted surfaces of pyrite. BE and At% denote binding energy and atomic percentage, respectively. Peak assignment was done by comparison with literature values (Table S1 in the Supplementary Information)

After HCl treatment, the chemical states of Fe and S were highly consistent among the control and the treatments, showing the predominant presence of surface Fe²⁺ and S⁻ (in Fe-S bond). Polysulfides with varying valences (including the end product elemental S) were also present at significant amounts. Trace amounts of monosulfide (S²⁻) were also detected. Fe²⁺/Fe³⁺ (in Fe-O/Fe-S bonds) accounted for about 34% (±3%) (Table 2).

Table 2 Chemical state of iron and sulfur on the corroded surfaces of pyrite. BE and At% denote binding energy and atomic percentage, respectively. Peak assignment was done by comparison with literature values (Table S1 in the Supplementary Information)

Discussion

In the presence of dissolved oxygen alone (the control), no clear sign of Fe²⁺ oxidation was observed. This is consistent with the established theory that oxidation of aqueous Fe²⁺ by molecular oxygen under acidic pH was kinetically slow³. Consequently, the sudden drop in aqueous Fe²⁺ concentration for the treatments with the added H₂O₂ can be interpreted as the result of H₂O₂-driven oxidation of aqueous Fe²⁺. The “standstill” of aqueous Fe²⁺ level after this initial rapid oxidation indicates that the H₂O₂ was already depleted within a few minutes. The rapid consumption of H₂O₂ through Fe²⁺ oxidation can be expressed by the following chemical equations:

The decomposition of 1 mole of H₂O₂ leads to simultaneous oxidation of 1 mole of Fe²⁺ and generation of 1 mole of (Equation 4), which may further oxidize another mole of Fe²⁺ (Equation 5). If spontaneous decomposition of H₂O₂ takes place, elevated concentration of DO is expected according to the following equation:

The trend lines showing the evolution of DO were very simular among the control and T1–T3 for the low Fe²⁺ scenario, indicating that spontaneous decomposition of H₂O₂, if any, was not significant for these treatments (Supplementary Fig. S3). The higher DO in T4 suggests that spontaneous decomposition of H₂O₂ did occur for this high-dose treatment. This is supported by the fact that the ratio of the converted Fe²⁺ to H₂O₂ was less than 1, indicating that over half of the added H₂O₂ was not consumed through Fe²⁺ oxidation.

By comparison with the abiotic oxidation results, it is evident that microbes played a crucial role in further oxidation of aqueous Fe²⁺ following the initial H₂O₂-driven oxidation in the corresponding biotic systems. However, the ability of microbes to catalyze the oxidation of aqueous Fe²⁺ was markedly affected by the initial concentration of aqueous Fe²⁺. When the initial concentration of aqueous Fe²⁺ was 55.8 mg L⁻¹, the microbially mediated Fe²⁺ oxidation was impeded for all the treatments, as compared to the control. The oxidation rate of aqueous Fe²⁺ tended to decrease with increasing dose of H₂O₂. This was consistent with the observed decrease in cell population and cellular damage in the affected microbes (Supplementary Fig. S1a and Fig. S2) in response to the toxicity of hydroxyl radical that was generated through Fenton reaction (Equation 4). As shown in Fig. 2, H₂O₂ at an initial concentration of 50 μmol L⁻¹ could cause marked oxidative stress in the cells. However, the presence of hydroxyl radical did increase the degree of oxidative stress in the cells.

In comparison with the low initial aqueous Fe²⁺ scenario (55.8 mg L⁻¹), the microbial oxidation of aqueous Fe²⁺ was not retarded for the treatments except for T4 (H₂O₂ at 1000 μmol L⁻¹) when the initial concentration of aqueous Fe²⁺ was increased to 558 mg L⁻¹. The reduced microbial toxicity level due to the presence of higher initial concentration of aqueous Fe²⁺ can be attributed to the effect of excessive Fe²⁺ on scavenging , as shown in Equation 5.

Rapid disappearance of most added Acidithiobacillus ferrooxidans from the nutrient solution containing powdered pyrite was observed and attributed to the colonization of the bacteria on the pyrite mineral surfaces^37,38. However, the lack of evidence showing the presence of attached cells and cell-shaped corrosion pits on the surfaces of pyrite cubes taken on the 28^th day of the experiment suggests that this was not the case in this study. The general trend that the viable planktonic cell population decreased with increasing H₂O₂ dosage level suggests that the H₂O₂-derived resulted in the growth inhibition or even killing of some Acidithiobacillus ferrooxidans following inoculation. Ferrous ion was present in all the solutions at least after the 3^rd day of the experiment. Consequently, free radical generation was possible.

However, it cannot be excluded that insufficient supply of Fe²⁺ to feed all the bacteria also contributed to the decrease in the viable planktonic cell population during the initial stage of the experiment³⁹. This explains the initial drop in planktonic cell count for the control, which contained no added H₂O₂.

The marked recovery of the bacterial population after the initial drop for CP and TP1 can be attributed to (a) the adaptation of the bacteria to the H₂O₂-induced oxidative stress, (b) the increased supply of Fe²⁺ released from the mineral surfaces and (c) possibly, the reduced concentration of free radicals as a result of the increased release rate of Fe²⁺, which alleviated the oxidative stress in Acidithiobacillus ferrooxidans, as discussed above.

The decreasing trend in cell density at the later stage of the experiment for CP and TP1 was likely to be caused by the reduced rate of Fe²⁺ release from the mineral surface during this period (the reasons for this have been discussed previously), as evidenced by the low concentration of aqueous Fe²⁺ and negative to very slow increase in aqueous Fe³⁺ during the period from the 28^th day to the end of the experiment. This made the solutions become less and less favourable for the growth of the planktonic bacteria. Consequently, the bacteria had to seek alternative “food” sources by landing on the mineral surfaces and feeding on the sulfur-rich substrates. This is in contrast with the scenario at the earlier stage of the experiment (excluding the initial 6 days) when sufficient amount of Fe²⁺ was constantly released into the solution. The results obtained here indicates that the Acidithiobacillus ferrooxidans preferentially fed on free Fe²⁺ and possibly thiosulfate (S₂O₃²⁻) in the solution rather than the structurally bound Fe²⁺ and reduced S species.

The absence of Fe³⁺ in TP4 until the 19^th day of the experiment indicates that Fe²⁺ oxidation was negligible during this period. This resulted in the accumulation of Fe²⁺ in the solution. Clearly, the weak microbial activity resulting from the high dosage level of H₂O₂ was responsible for the inhibition of Fe²⁺ oxidation while the H₂O₂ itself as an oxidant had limited effect on the abiotic oxidation of the aqueous Fe²⁺ under the set experimental conditions. The aqueous Fe³⁺ concentration reflected the combined effect of the Fe³⁺ production from Fe²⁺ oxidation and the Fe³⁺ immobilization due to precipitation of iron compounds (precipitates were observed during the experiment but they were not produced in sufficient quantity to allow chemical and mineralogical analysis). Therefore, the steady increase in aqueous Fe³⁺ in CP and TP1 until the 28^th day of the experiment indicates a higher rate of the Fe²⁺-Fe³⁺ conversion, relative to the rate of Fe³⁺ immobilization during this period. This more or less corresponded with a period when the bacterial population rapidly increased, reflecting an interdependency relationship between the Fe²⁺-Fe³⁺ conversion rate and the bacterial density. The subsequent plateau and decline phase of aqueous Fe³⁺ evolution was likely to represent a period of reduction in Fe³⁺ production as a result of reduced rate of Fe²⁺ release. This can be supported by the fact that the bacterial population tended to decrease during this period. The high similarity of Fe³⁺ evolutionary pattern between CP and TP1 suggests that a H₂O₂ level of 50 μmol L⁻¹ was unlikely to cause a marked reduction in microbial oxidation of aqueous Fe²⁺ under the set experimental conditions. However, the increased dosage level of H₂O₂ tended to result in lower Fe³⁺ concentration in the solution. It is interesting to note that the Fe²⁺ concentration remained low during the entire period of the experiment for the H₂O₂ treatments except for the earlier stage of TP4. This indicates that the Fe²⁺-Fe³⁺ conversion rate more or less kept pace with the Fe²⁺ production rate. Therefore, the reduced rate of Fe²⁺ release from the mineral surface was a more upstream cause responsible for the lower solution Fe³⁺ concentration in these treatments.

The molar ratio of Fe to S in the solution ranged from 0.15 to 0.22, which were much smaller than the theoretical value of 0.5 assuming that equal amount of Fe and S was liberated from the pyrite cube surfaces during the experiment. In fact, it was likely that more Fe than S entered into the solution during the experiment, as evidenced by the presence of the Fe-deficient reacted pyrite surfaces (Fig. 5a). Therefore, much more Fe than S was removed from the solution through iron compound precipitation. Schwertmannite (Fe₈O₈(OH)₆SO₄) and jarosite (KFe₃(OH)₆(SO₄)₂) have a Fe/S molar ratio of 8 and 1.5, respectively. Therefore, the formation of basic sulfate minerals such as schwertmannite and jarosite explains the imbalance between Fe and S in the solution. This is consistent with the observed generation of H⁺ (as indicated by pH drop) as a result of Fe³⁺ hydrolysis leading to the formation of these minerals:

The Fe-deficient nature of the corroded surfaces (after HCl treatment to remove the coatings) indicates that the pyrite-Fe was preferentially liberated from the mineral surface, as compared to the pyrite-S. This can be explained by the relative easiness of the pyrite-Fe liberation reaction, as shown in the following chemical equations:

The completion of the above chemical reactions requires only one electron transfer from pyrite-Fe²⁺ to an electron acceptor (oxidant) in the solution, leaving structurally bound polysulfides (S₂⁻) on the mineral surfaces. In contrast, the liberation of pyrite-S₂²⁻ requires multiple steps of electron transfer from the pyrite surface to electron acceptors. The minimum number of electrons needed to be transferred to the external oxidants in order to liberate two pyrite-S atoms is 6 according to the equation below:

Consequently, the oxidation of pyrite-S₂²⁻ did not keep pace with the oxidation of pyrite-Fe²⁺, leaving sulfur species of intermediate oxidation states remained structurally connected with the pyrite substrate. The presence of a surfur-rich surface layer in oxidized pyrite crystals was also observed by others^40,41.

The XPS results indicate the presence of oxidation products containing Fe³⁺ and SO₄²⁻ within the ~3–5 nm thick outermost layer of the reacted pyrite surface. This suggests the formation of coating materials such as ferric (oxy)hydroxysulfates and (oxy)hydroxides. Xia et al⁴² observed the presence of jarosite in the microbially oxidized pyrite surfaces at acidic pH and suggested that the formation of surface-jarosite was the major cause responsible for the passivation of the sulfide mineral surfaces. The tendency that the proportion of surface-Fe³⁺ and -SO₄²⁻ decreased with increasing dosage level of H₂O₂ appears to suggest that the presence of H₂O₂ affected the build-up of the ferric (oxy)hydroxysulfate-containing coating layer on the mineral surfaces although the exact mechanisms are not clear. Possibly, the differential composition of oxidants in the different H₂O₂-treated systems was an upstream factor that needs to be considered for the interpretation of the observed phenomena. From CP to TP4, the relative importance of H₂O₂-driven surface oxidation was likely to increase as a result of the increase in the H₂O₂ concentration and the simultaneous decrease in Fe³⁺ concentration. The weakened role of Fe³⁺ as a driving oxidant for mineral surface oxidation might reduce the frequency of Fe³⁺-SO₄²⁻ contact at the solution-mineral interfaces and consequently disfavour the formation of ferric (oxy)hydroxysulfate coatings, as described by the following H₂O₂-driven oxidation reaction:

This is in contrast with the Fe³⁺-driven oxidation reaction, as shown in the following equation describing the formation of jarosite:

Since no surface-S⁻ was detected by XPS for CP and TP1, it is reasonable to believe that the thickness of the coating layer on the pyrite cube surfaces was >3–5 nm when the mineral crystals were exposed to H₂O₂ at a dosage level below 50 μmol L⁻¹. The coating layer became thinner with increasing dosage level of H₂O₂, resulting in the occurrence of bulk pyrite-S (S⁻) within the ~3–5 nm thick outermost layer of the original reacted pyrite surfaces, which was consistent with the XPS results for the treatments with high H₂O₂ doses.

The research findings obtained from this study shed some light on the possible complication of the biogeochemical processes associated with the weathering of pyrite and other sulfide minerals due to the presence of H₂O₂ in the concentration range that may be encountered in field conditions. Our recent work⁴³ also suggests that the Fe²⁺-H₂O₂ combinations that are likely to be encountered in natural waters resulted in significant degradation of agricultural herbicides. There is, therefore, a strong rationale for conducting additional laboratory-based investigations to obtain further insights into the exact chemical mechanisms and kinetics and the subsequent field-scale study, which can be used to better evaluate the environmental risk of the toxic surface runoff from acid sulfate soils and sulfidic mine sites and develop cost-effective management strategies and techniques to minimize its environmental impacts.

Methods

Pyrite specimens and pretreatment

The pyrite specimens used in this study were natural pyrite cubes purchased from the Anhui Tongling Siling Mineral Ltd. For this study, the purpose was to examine the effects of externally originated H₂O₂ on microbially involved oxidation of aqueous Fe²⁺ and pyrite grain surfaces. Therefore, powdered pyrite with high specific surface area was not appropriate due to its potential for causing the production of spontaneous H₂O₂³⁰. Pyrite cubes with similar size (approximately 1.5 × 1.5 × 1.5 cm³) were selected for the mineral-solution contact experiment. Pyrite surfaces are readily reactive when exposed to air. Pyrite oxidation products are expected to be present on the surfaces of any naturally occurring pyrite specimens. It is necessary to remove these coating materials prior to pyrite oxidation experiment. The surfaces of the mineral crystals were treated with a boiling 6 mol L⁻¹ HCl solution to remove any oxides, (oxy)hydroxides and (oxy)hydroxysulfates of iron that were possibly present on the original mineral surfaces⁴⁴. The “cleaned” pyrite cubes were immediately used for the experiment after washing with distilled water twice and acetone for three times to further remove elemental sulfur that is possibly present on the mineral surfaces.

Bacteria, culture conditions and inoculum preparation

A strain of Acidithiobacillus ferrooxidans was purchased from the China Marine Microbial Culture Collection Center (MCCC). The bacterial culture was maintained at 4°C in a 9 K nutrient medium³⁹ containing 3.0 g of (NH₄)₂SO₄, 0.01 g of Ca(NO₃)₂, 0.5 g of MgSO₄.7H₂O, 0.5 g of K₂HPO₄, 0.1 g of KCl and 44.3 g of FeSO₄.7H₂O in 1 L of distilled water with pH adjusted to 1.6 with a H₂SO₄ solution.

The inoculum was prepared prior to the experiment. An adequate amount of the bacteria required for the experiment was produced by facilitating bacterial growth in a sterile 9 K medium at 30°C, coupled with shaking (130 rpm) on a rotary shaker for 5–6 days. The cells in the enriched suspension were firstly separated from the iron precipitates (formed during the incubation) by centrifugation at 3000 rpm for 3 min to allow the settlement of the solid iron compounds. The cells remained in the suspension were then transferred into a new centrifuge tube and harvested by centrifugation at 5000 rpm for 10 min to allow the settlement of the cells. After washing twice with sterile distilled water (adjusted to pH 2 with a H₂SO₄ solution), the inoculum was formed by adding an appropriate amount of the same acidified distilled water into the centrifuge tube containing the cleaned cells. The cell concentration in the inoculum for each batch was determined by direct cell counting prior to addition of the inoculum into the experimental reactors in an experiment.

Aqueous Fe²⁺ oxidation experiment

Both biotic and abiotic experiments were conducted to observe the oxidation of Fe²⁺ following a single injection of H₂O₂. The abiotic experiment was to account for the effect of H₂O₂ as an oxidant on the chemical oxidation of aqueous Fe²⁺ while the biotic experiment was to examine the integrative effects of H₂O₂ (as an oxidant and a toxicant to iron-oxidizing bacteria) on Fe²⁺ oxidation. One control (C, without added H₂O₂) and four treatments with different initial H₂O₂ concentrations were established: (a) Treatment 1 (T1): 50 μmol L⁻¹; (b) Treatment 2 (T2): 100 μmol L⁻¹; (c) Treatment 3 (T3): 300 μmol L⁻¹; and (d) Treatment 4 (T4): 1000 μmol L⁻¹. For either biotic or abiotic experiment, two separate experiments with different initial levels of Fe²⁺ were conducted: (a) 55.8 mg L⁻¹ and (b) 558 mg L⁻¹.

For all the experiments, a Fe-free 9 K medium (without added FeSO₄·7H₂O) was used as the basal solution (i.e. the solution prior to the addition of various ingredients). A 150 mL conical flask was used as a batch reactor to contain 100 mL of a reacting solution with a pre-set H₂O₂-Fe²⁺ combination. The pH of the reacting medium was adjusted to 2 using a H₂SO₄ solution.

For the biotic experiments, the reacting medium was inoculated with Acidithiobacillus ferrooxidans at an initial cell concentration of 1.1 × 10⁷ cells mL⁻¹ for the lower Fe²⁺ treatments and 1.0 × 10⁷ cells mL⁻¹ for the higher Fe²⁺ treatments. All the conical flasks were loosely wrapped by aluminum foil to allow entry of air but not dust during the entire period of the experiment except at the time of sample collection. The reactors were shaken at 130 rpm on a rotary shaker with temperature set at 30°C. A small amount of sample was taken at different time to determine residual Fe²⁺ concentration, cell density and other relevant parameters. The experiment was performed in triplicate.

For the abiotic experiments, DO in the solutions was also monitored to examine whether marked spontaneous decomposition of H₂O₂ took place during the incubation experiment according to Equation 6.

Pre-experiment tests indicated that the added H₂O₂ was almost depleted within 5 min in the reaction systems investigated for the current study. Consequently, monitoring of H₂O₂ was not performed due to the following reasons: (a) the research objective (overall effects of H₂O₂ flux at the preset initial concentrations on the microbial oxidation of cubic pyrite and aqueous Fe²⁺) can be satisfactorily achieved without knowing the detailed evolution of H₂O₂ within the 5 min following the H₂O₂ injection; (b) the measured values of H₂O₂ for the different treatments are not reasonably comparable for such highly dynamic reaction systems since it is technically very difficult, if not impossible, to ensure the simultaneous analysis for the samples of all the treatments; and (c) there is a lack of highly reliable analytical methods for rapid measurements of micromolar level H₂O₂³⁰.

Pyrite oxidation experiment

The basal solution used was the same as the aqueous Fe²⁺ oxidation experiment. The “cleaned” pyrite cubes were exposed to H₂O₂ at various initial concentrations in the basal solution (with the pH adjusted to 2.0 by a H₂SO₄ solution) in the presence of the Acidithiobacillus ferrooxidans at an initial concentration of 2.7 × 10⁷ cells mL⁻¹. One control (CP, without added H₂O₂) and four treatments with different initial H₂O₂ concentrations were established: (a) Treatment1 (TP1): 50 μmol L⁻¹; (b) Treatment 2 (TP2): 100 μmol L⁻¹; (c) Treatment 3 (TP3): 300 μmol L⁻¹; and (d) Treatment 4 (TP4): 1000 μmol L⁻¹. A 250 mL conical flask was used as the reaction chamber. Five pyrite cubes were soaked in 90 mL of a respective solution for the control and each treatment. The conical flask was loosely wrapped using aluminium foil to allow entry of air but not dust and then kept in a biological incubator with the temperature set at 30°C during the entire period of the experiment except at the time of cyclic H₂O₂ injection and sample collection.

A time interval of 3–5 days was established for re-injection of H₂O₂, in-situ measurements of pH and solution sample collection for determinations of Fe²⁺, total Fe, S and cell density. After sampling, an equal amount of sterile Fe-free 9 K medium was added into the reactor to compensate the solution loss caused by the sample collection. To avoid markedly disturbing the solution equilibrium system, the volume of solution samples for chemical and microbial monitoring had to be minimized. Therefore only the above important parameters were measured for this experiment.

The experiment lasted for 86 days. On the 28^th day of the experiment, two of the five pyrite cubes for the control and each treatment were taken for surface characterization analysis: one was used for observing the original oxidized surface and another was treated with a boiling 6 mol L⁻¹ HCl solution to remove the oxidized materials for observing the corroded surface. At the end of the experiment, the pyrite cubes were harvested for surface characterization analysis to compare with those collected on the 28^th day of the experiment.

Toxic response of Acidithiobacillus ferrooxidans to H₂O₂ under various conditions

Because both aqueous Fe²⁺ and cubic pyrite oxidation experiments indicated that Acidithiobacillus ferrooxidans experienced stress when exposed to H₂O₂, a supplementary toxic response experiment was conducted to determine whether the bacterial oxidative stress was caused by H₂O₂ alone or/and by its deriving hydroxyl radical as well. The basal solution used was the same as that in the aqueous Fe²⁺ and cubic pyrite oxidation experiments. Five treatments were established: (1) BI: without added H₂O₂; (2) BH: 50 μmol L⁻¹ H₂O₂ only; (3) BIH: 50 μmol L⁻¹ H₂O₂ and 55.8 mg L⁻¹ Fe²⁺; (4) BHA: 50 μmol L⁻¹ H₂O₂ and 50 μmol L⁻¹ ascorbic acid (a scavenger of reactive oxygen species); and (5) BIHA: 50 μmol L⁻¹ H₂O₂, 55.8 mg L⁻¹ Fe²⁺ and 50 μmol L⁻¹ ascorbic acid. Only one single injection of H₂O₂ was performed for the experiment. Acidithiobacillus ferrooxidans were inoculated into 100 mL of a respective solution in a conical flask. The initial pH was 2 and the initial cell concentration was 3.15 × 10⁷ cells mL⁻¹. The reactors were shaken at 130 rpm on a rotary shaker with temperature set at 30°C. Samples were taken for cell counting after shaking for 30 min.

Analytical methods

In-situ measurement of pH and DO in the solution was made by a calibrated pH meter and DO meter, respectively. Fe²⁺ in the reacting solutions was measured by the potassium perchromate titration method⁴⁵. Total Fe and S in the solution were determined by the inductively coupled plasma–atomic emission spectroscopy (ICP-AES). The concentration of viable planktonic cell in the reacting solutions was determined by direct cell counting using a Neubauer hemocytometer.

A FEI-XL30 environmental scanning electron microscope coupled with energy dispersive X-ray spectrometer (ESEM/EDS) was used for surface imaging and determining the surface chemical composition of the pyrite cubes. The observed surfaces were coated with a 20 nm thickness of gold. The accelerating voltage was 15 kV and the working distance was 4.9 to 6.3 mm.

XPS was employed to determine the chemical composition and element state of the mineral surfaces (to a depth of ~3–5 nm). XPS analyses were performed with a Kratos Axis Ultra^DLD spectrometer using a monochromatic Al Kα X-rays source. Broad scan was conducted using 160 eV pass energy, while narrow high-resolution spectra of all major lines were obtained using a resolution function with a width of 0.1 eV for a pass energy setting of 40 eV. The charge effect was corrected using C 1 s from contamination at 284.6 eV. Spectra were analyzed using the CasaXPS software (Version 2.2.19). Assignment of iron and sulfur species was made by referring to the published documents (Table S1 in the Supplementary Information).