Biostimulation of jarosite and iron oxide-bearing mine waste enhances subsequent metal recovery

Novel resource recovery technologies are required for metals-bearing hazardous wastes in order to achieve circular economy outcomes and industrial symbiosis. Iron oxide and co-occurring hydroxysulphate-bearing wastes are globally abundant and often contain other elements of value. This work addresses the biostimulation of indigenous microbial communities within an iron oxide/ hydroxysulphate-bearing waste and its effect on the subsequent recoverability of metals by hydrochloric, sulphuric, citric acids

• Enhanced recovery of metal postbiostimulation of iron oxide/hydroxysulphate waste.• Simple flow through reactors used for the study.• Increase in relative abundance of iron/ sulphate reducers in presence of glycerol.• Glycerol fed waste exhibited higher extractability of metals.• Chelation extraction with EDTA shows notable selectivity of metals over iron.

Introduction
The "technosphere" includes any material established by human agency in contrast to lithospheric stocks established by geological processes (Johansson et al., 2013).Iron oxide minerals include the wide variety of iron oxides, hydrous ferric oxides, hydroxides, oxyhydroxides, and co-occurring hydroxysulphate minerals found within many industrial and mining wastes (IMWs).Billions of tonnes of IMWs have been produced and continued global arisings are in the hundreds of millions of tonnes annually.As the world moves towards circular economies, iron oxide/ hydroxysulphate-bearing wastes (hereon in "iron-rich" waste) present a potentially exploitable technospheric iron resource, as well as a source of other elements of strategic and/or economic value that have been sequestered via sorption, coprecipitation, or occlusion with the iron-rich waste (Roberts et al., 2020).
Iron-rich wastes therefore comprise a large technospheric stock of iron and associated valuable elements and as such will likely become a promising target for resource recovery.The target of value recovery might conceivably be the iron or the other associated elements of value, or both.As well as an iron source for steel making, iron oxides once decontaminated to acceptable levels can find diverse applications including being used as pigments (Hedin, 2003;Legodi and de Waal, 2007;Ryan et al., 2017), ferrites (e.g., Novais et al., 2016;Tamaura et al., 1991;Wang et al., 1996), and water/wastewater treatment reagents (e.g., Sapsford et al., 2015;Xu et al., 2012;Zhong et al., 2006).For the associated metals present at concentrations considered un-economic for recovery, the decontaminated land may be considered a resource which may be used to offset the cost required for decontamination of the iron-rich wastes.
To date, studies into metal recovery from iron-rich wastes and ores have been largely focused on abiotic leaching (e.g., Das et al., 1997;Hernández et al., 2007).Biohydrometallurgical and biomining technologies hold promise for sustainable resource recovery from wastes, by reducing the energy and reagent usage (and hence carbon intensity) associated with conventional hydrometallurgical and pyrometallurgical processes.Existing biohydrometallurgical leaching (applied in situ or ex situ) are already extensively used in processing of sulphidic ores.Other proposed biomining approaches include the use of other acids, cyanogenic, or chelate-generating microorganisms to mobilise metals (Lee and Pandey, 2012).Bioreduction of iron oxide, where the oxidation of a suitable electron donor is coupled with the reduction under anaerobic conditions of Fe(III), has been explored as a mechanism for metal recovery from (particularly lateritic) ores (Esther et al., 2020;Papassiopi et al., 2010).Recovery of Ni, Co, and Cu from laterites has been explored extensively using dissimilative iron reducing microorganisms (DIRM) (Hallberg et al., 2011a;Ňancucheo et al., 2014;Smith et al., 2017) for ore biobeneficiation (Natarajan, 2015), but there are very few studies that focus on metal recovery from wastes (Hallberg et al., 2011b;Johnson et al., 2013;Johnson and du Plessis, 2015;Smith et al., 2017).To date, DIRM utilised in most biohydrometallurgy studies are specifically cultured and supplied with an optimal carbon source, nutrients, and conditions for growth, with all the commensurate implication for carbon intensity and costs should these ex-situ bioreactor processes be upscaled.
For the iron hydroxysulphate minerals jarosite and schwertmannite (Fe 16 O 16 (OH) y (SO 4 ) z •nH 2 O), the presence of structural sulphate presents an indirect route to iron reduction via reaction with sulphides produced by microbial sulphate reduction.Iron oxyhydroxides are highly reactive with respect to dissolved sulphide, the reaction producing FeS and elemental sulphur directly or ferric sulphide that can disproportionate to pyrrhotite (Fe (1− x) S (x = 0-0.2))/pyrite(FeS 2 ) and elemental sulphur.The majority of literature focuses on Fe(III) bioreduction within the iron hydroxysulphate structure (Bingjie et al., 2014;Castro et al., 2013;Jones et al., 2006;Smeaton et al., 2012) but bioreduction of structural sulphate to sulphide within iron hydroxysulphates has also been demonstrated (Gramp et al., 2009;Ivarson et al., 1976).
This study uses laboratory-scale flow-through column reactors to examine the effect of using glycerol (10% w/w) to stimulate the in situ microbial community in an iron oxide/ hydroxysulphate-bearing mine water precipitate to examine the impacts in terms of evolution of microbiological community, element mobility, and subsequent hydrometallurgical extractability of metals.

Iron oxide/hydroxysulphate-bearing waste sampling and characterisation
Material was sampled from the Parys Mountain legacy Cu mine, located ~2.5 km south of Amlwch in Anglesey, North Wales (53 • 23 ′ 13 N, 4 • 20 ′ 37 W).Samples were obtained from former copper cementation pond, now infilled with ochreous sediment, to the south of the site (Fig. 1).The material (Parys Mt.) was homogenised manually and left to settle for 48 h, after which excess water was decanted and stored at 4 • C. Post this, the sample was manually homogenised again for 10 mins by continuously stirring and folding within a bucket, left to settle for 1hour, and the supernatant discarded.After repeating this procedure, samples were taken aseptically, for mineralogical, chemical, and microbiological characterisation.Measurements of the dry solids content (% weight), bulk density, dry density, void ratio, particle size distribution, and specific gravity were performed in accordance with BS 1377-2:1990.The particle size distribution (PSD) of the − 150 µm sieve fraction was characterised with a Mastersizer 3000 laser diffraction particle size analyser (Malvern).
Chemical characterisation was undertaken on the Parys Mt material before and after column experiments.The waste was dried in an oven (105 • C) for 24 h.Approximately 0.15 g sample was digested in 6 ml aqua regia in a PTFE-lined ceramic vessel using a microwave digestor (Multiwave 3000 Anton Parr).The digested samples were analysed M. Roberts et al. using an inductively coupled plasma-optical emission spectroscopy (ICP-OES; Perkin Elmer Optima 2100).The pH of the wastes was measured using the paste pH methodology outlined in BS ISO 10390:2005, with Mettler Toledo InLab™ Expert Pro pH probe.Mineralogical characterisation was undertaken using fine powder X-ray diffractometer (XRD; Philips PW3830 X-ray generator, PW1710 diffractometer controller) with a cobalt (Co Kα) or copper (Cu Kα) radiation source, scan angle range of 2θ from 5 • to 90 • , a step size of 0.02 • , and scan step time of 0.5 s.The traces were analysed using the mineral database software, X-pert Hi Score plus analysis.Total carbon and sulphur analysis of the solid waste samples was performed using a Leco SC-144DR furnace.Total carbon (TC) and Total Organic carbon (TOC) were analysed using a Shimadzu SSM-5000A.Total inorganic carbon (TIC) was calculated by difference (TC -TOC) after phosphoric acid pretreatment and analysis.
A sequential extraction procedure adapted from (Poulton and Canfield, 2005), designed for iron rich sediments, was used to differentiate various phases.A solid-to-liquid ratio of 0.02 g/ml was employed.These phases are ascribed as carbonate associated iron (Fe Carb ); easily reducible oxides (Fe ox1 ) e.g., ferrihydrite; reducible oxides (Fe ox2 ) e.g., goethite and haematite; magnetite (Fe mag ); and residual fractions (Fe res ).Metal concentrations within each fraction was determined by ICP-OES.Initial extraction with deionised water was used to account for any metal present in the water-soluble phase (Fe wat ) (Dold, 2003).See Table 1 for details.

Column experimental set-up
Three sets of triplicate up-flow reactors were packed with: (i) unadulterated waste, to be later fed with 10% w/w glycerol in deionised water (DIW) (live); (ii) autoclaved waste fed with 10% w/w glycerol in DIW (autoclaved); and (iii) unadulterated waste fed with DIW (organic starved).Columns were constructed from 1 l high density polyethylene (HDPE) bottles fitted with an inlet/outlet valve at either end.Each bottle was washed 3 times with DIW and autoclaved before use.Column effluents were collected in 2 l Erlenmeyer flasks which had been purged with nitrogen and fitted with a 2-hole rubber bung.One hole was for the inflow tube from the column whilst the other was fitted with a brewer's airlock "bubbler" valve to limit ingress of air to the collection flask (Supplementary Figure 1).Columns were fed at a flow rate of 0.2 ml/ min, using a Watson Marlow 205 U peristaltic pump, giving a nominal residence time of ~ 25 h.Effluent samples were collected at 2-3-day intervals and physicochemical parameters were measured.Subsamples acidified with 20% (v/v) HNO 3 were stored in the refrigerator at 4 • C until analysis.The experiments were run for 91 days giving a total of 26.2 l of influent percolating per column.Iron oxide staining was noted in some collection vessels.This was recovered by an acid wash (37% HCl), with subsequent determination of metals (Section 2.3).After column decommissioning, sludge was extracted from the columns and subsampled for XRD analysis and sequential extractions followed by ICP-OES.

Analysis of effluents
Analysis of pH, oxidation reduction potential (ORP), electrical conductivity (EC), and dissolved oxygen (DO) were measured using calibrated Mettler Toledo probes: InLab™ Expert Pro pH and ORP-probes, LE703 conductivity probe, 605-ISM DO probe in conjunction with a Mettler Toledo SevenExcellence™ Multiparameter system.Effluent alkalinity was measured using a Hach 16900 digital titrator.Elemental (Fe, Cu, Pb, As, Zn, Al, Ni, and S) analyses were conducted by ICP-OES (Perkin Elmer Optima 2100 DV).Colourimetery (HACH DR900) was used for the determination of Fe(II) (phenanthroline method) and sulphate concentrations (SulfaVer® 4 turbidimetric method).Glycerol concentrations were measured in column influents and effluents using a fluorometric free glycerol assay kit (Sigma Aldrich, Gillingham, UK) and Infinite® F50 fluorometer with Magellan software.

Hydrometallurgical extractions
The pre-and post-experiment column material were subjected to leaching using either hydrochloric acid (HCl), sulphuric acid (H 2 SO 4 ) or citric acid (C 6 H 8 O 7 ) at concentrations of 1 M, 0.5 M and 0.25 M (cf.(Esther et al., 2015)) or 10 mM, 5 mM, and 2.5 mM of ethylenediaminetetraacetic acid (ETDA), amended to pH 8 with NaOH (cf.(Sun et al., 2001)).All tests were performed in duplicates at L:S of 10:1 on an orbital shaker (24 h at 120 rpm) within a glovebox under N 2 .Final extracts were filtered through a 0.2 µm cellulose acetate filter and analysed by ICP-OES for Fe, Zn, Pb, Al, Cu, As, and S. To quantify the impact of glycerol on the leachability of metals, additional tests were performed on pre-experiment wastes where 10 mM of glycerol was added to the lixiviant (Supplementary section 1.1).The presence of glycerol was shown to have no effect on the subsequent leachability.

Microbial community analysis
Total DNA was extracted from both the pre-and the postexperimentation waste using the Fast DNA® SPIN Kit for Soil (MP Biomedicals, Solon, OH, USA) as per manufacturer's instructions, with slight modifications according to (Webster et al., 2003).Qubit dsDNA assay kit (Invitrogen, Carlsbad, CA, USA) was used to quantify the extracted DNA.Next-generation sequencing (NGS) of bacterial and archaeal 16 S rRNA genes and analysis was done as described in (Srivastava et al., 2022).

Physicochemical characterisation of the Parys Mountain waste
The Parys Mt. material was well graded with a range of particle sizes present (Supplementary Figure 2a, b).58.2% of the mass of the waste was of gravel size or greater.These larger particles largely comprised of quartz, silicates, and highly weathered iron-rich gossan.34.3% of the waste comprised of sand sized particles and 7.57% of the waste was silt or clay.The 150 µm sieve material ranged from 2.4 µm (D 10 ) to 94 µm (D 90 ), with a D 50 of 11 µm.Parys Mt. waste was acidic (pH = 2.61), with a sulphur content of 3.08 wt% and an organic carbon content of 0.68 wt %.Table 2 shows the physicochemical properties of the waste.The elemental composition of the waste is shown in Fig. 1c.Metals of potential economic interest within the waste include Pb (0.27 wt%), Zn (0.17 wt%), and Cu (0.69 wt%).XRD analysis indicated the presence of quartz, jarosite, and goethite, correlating with the high Fe, S, and K concentrations found by total digest and Leco sulphur analysis (Supplementary Figure 2c).

Table 1
Summary of the sequential extractions used for iron rich sediments.Fig. 1d shows the distribution of Fe throughout the sequential extraction phases present in the waste.A minor proportion of iron was present in the two most reactive phases, viz., carbonate associated (carb; 0.04%), and easily reducible oxide (ox1; 2.2%).Fe was absent in the water-soluble (DI) phase.40.9% of the reducible oxide phase (ox2) comprised of Fe.Zn, like Fe, was also found mostly within the 3 least reactive phases (Fig. 1d).The largest of these phases is the residual phase (res) with 47.8% total Zn.6.4%, 1.4%, and 1% of total Zn was located within the carbonate associated, easily reducible oxide, and water-soluble phases, respectively.23.2% of total Zn was held within the reducible oxide phase while the remaining 21.3% was within the magnetite targeted extract (mag).Over half of the Pb was located within the magnetite targeted extract (54.1%) while a further 29.5% was located within the residual phase.The remaining 16.4% of total Pb was distributed between the three most reactive phases, which represent target phases for metal recovery via microbial reductive dissolution of iron.The relatively low amount of Pb within these target phases suggests that there is significantly less Pb associated with iron oxyhydroxides (either co-precipitated or adsorbed) or as easily leachable minerals.Although a greater amount of Pb was present within the waste it appears that a lower proportion of this waste may be available for recovery via reductive dissolution.
The distribution of Cu was like Pb where most of the Cu was held within the magnetite targeted (46.9%) and residual (46.5%) phases.Despite the elevated Cu concentrations within the waste, the sequential extraction results suggest that only minor quantities of Cu are likely accessible for recovery as only 3.1% and 3.5% of total Cu was contained within the carbonate associated and easily reducible oxide phases respectively.It is notable, however, that no Cu was extracted in the reducible oxide phase representing crystalline ferric oxides such as goethite.This is attributable to the sodium dithionite within this phase's extractant which causes solubilised Cu to precipitate out of solution (Chou et al., 2015).This effect causes Cu to be underrepresented in the reducible oxide phase and therefore potentially overrepresented in the subsequent phases.This makes predicting the amount of Cu that can be considered feasibly recoverable difficult and inaccurate (Chou et al., 2015).
Despite higher concentrations of the primary target metals of Zn, Pb, and Cu within the waste, sequential extraction data suggests that a smaller proportion of this metal value is potentially recoverable by reductive dissolution of iron.If conservative recovery targets, using only the easily reducible oxide and reducible oxide phases, are considered then ~24.5% Zn, ~13% Pb, and ~3.5% Cu can be considered feasibly recoverable.

Iron release and pH/ORP/EC
Fig. 2 shows the total Fe release and corresponding pH/ORP/EC for the column experiments.Fe was released from all columns but showed markedly different behaviour.The organic starved column (control) demonstrated that iron was leaching from the column in response to water flushing.The results indicate a typical flush curve suggesting that there is a pool of soluble iron (some of which is clearly Fe(II) -see Fig. 3a) that was being mobilised and removed from the column in response to water flushing, this is also supported by the decrease in EC with time (Fig. 2a).The proportion of Fe(II) did not increase with time.The pH (Fig. 2a) was around pH 3.0 reflecting the paste pH of the material and explaining the presence of soluble iron in the effluent, which later increased gradually to ~ pH 6.5.The ORP remained consistent at a value 420 mV ± 25 mV.DO was 5.44 mg/l which dropped to negligible indicative of anaerobic conditions in the column.The behaviour of the organic starved material contrasts with the columns where glycerol was introduced (Figs. 2 and 3).
Total Fe (Fe Tot ) concentrations in the live columns were much higher (369 mg/l on day 91) than that observed in control (101 mg/l on day 91), suggesting that addition of glycerol lead to enhanced release of iron.It decreased steadily from the peak of 1363 mg/l on day 21 to 367 mg/l on day 56.Thereafter, Fe Tot remained relatively constant until day 75, after which a further decrease was observed (Fig. 3b).Fe(II) increased from day 8 onwards indicating the onset of reducing conditions and iron reduction in response to glycerol addition (Bridge and Johnson, 1998), which was not observed in organic starved controls.Thereafter, Fe(II) in these column effluents displayed a similar pattern to that of Fe tot (Fig. 3b).pH increased from the initial low value of pH 3.15 on day 1 to a final pH of 6.68 on day 91(Fig.2b).The increase was coeval with the rapid increase in iron concentrations, indicating the onset of iron and/or sulphate reduction as these are both protons consuming/alkalinity producing reactions (Paper et al., 2021).The variability in releases between replicate columns may be attributed to the variations in flow path/residence time.DO within the effluents was initially high, but soon dropped to < 1 mg/l indicating anaerobiosis.Eh steadily decreased throughout the experiment, with the rate of decrease slowing towards the later stages of experimentation (Fig. 2b).Eh was ~70 mV when the pH was circumneutral, below the critical Eh (~100 mV) suggested for iron reduction (Petruzzelli et al., 2005) but above the critical Eh (− 150 mV) for sulphate reduction (Connell and Patrick, 1968).However, this reflects the effluent only and it is likely that a range of redox gradients existed within microenvironments of the waste.Fe Tot in the autoclaved column effluent remained constant at ~340 mg/l for around 14 days before increasing sharply (Fig. 3c).Similarly, the Fe(II) concentration remained between 10 and 14 mg/l until day 12, after which a sharp increase was observed (Fig. 2c).Autoclaving the waste led to a substantially longer "lag phase" as compared to live columns (Fig. 3).Although the Fe Tot exhibited an increase with time, the increase was lower than that observed in live columns; with Fe Tot increasing gradually at an average rate of 27.5 mg/ l/day to a peak of 791.0 mg/l on day 44 of the experiment.Unlike Fe Tot , the average peak Fe(II) concentration of 400 mg/l, was far closer to the live column equivalent (Fig. 3c).The solubility of Fe(II), as in the live columns, did not limit the recovery of iron from the columns.An average of 5.4% of total iron was recovered from the autoclaved columns.Both pH and alkalinity of the autoclaved column effluents increased coevally with the increase in Fe Tot over time (Fig. 2c) and exhibited a trend very similar to live column effluents.The significant reduction in the rate of pH increase, post-day 44 is likely a reflection of reduced microbial activity in the columns as also evidenced by the decreasing Fe tot concentrations.Autoclave resistant spores present in the waste may be responsible for the microbial activity that regulates the physicochemical parameters (O'Sullivan et al., 2015;Srivastava et al., 2022).
Eh remained constant for first 12 days, after which it decreased steadily to the end of the experiment ending at ~− 70 mV (Fig. 2c), just as in the live column.A simultaneous increase in both the Eh and Fe tot concentration was observed, which further indicates microbial activity.DO exhibited a rapid decrease to < 1 mg/l and remained so for the duration of the study indicating anaerobic conditions, ideal for microbial dissimilative iron reduction (Lovley, 1993;Weber et al., 2006).Similarly, EC exhibited the same trend as live columns indicating iron to be the dominant cation in the effluent (Fig. 2c).

Sulphate
Sulphate concentration in the organic starved column effluent remained consistently at ~ 240 mg/l and exhibited a trend similar to Fe Tot (Fig. 3d).The SO 4 2-released in the experimental column effluents fed with glycerol was higher than organic starved that can be attributed either to the SO 4 2-sorbed on to mineral surfaces or the dissolution of iron hydroxysulphates such as jarosites (Bridge and Johnson, 1998;Jones et al., 2006).Sulphate concentration in live column effluent exhibited a similar trend to that of Fe Tot concentrations and was significantly higher than organic starved column effluents (Fig. 3d).This suggests that jarosite is the target for bioreduction where, both Fe(II) and SO 4 2-are solubilised from jarosite, commensurately.A decrease in SO 4 2-concentration is observed from day 56 onwards with a concomitant increase in pH to circumneutral conditions.This may either be attributed to a decrease in jarosite bioreduction or sulphate reduction superseding iron reduction.The latter is thermodynamically favourable at circumneutral pH conditions (Chapelle et al., 2009;Lovley and Phillips, 1987).As with the live samples, the trend in average effluent sulphate concentration is similar to the Fe tot concentration within the autoclaved effluents, with a broad curve peaking at 1200 mg/l (day 42; Fig. 3d).Decrease in SO 4 2concentration from day 49 onwards is indicative of sulphate-reducing conditions.Iron oxides are very reactive towards HS -, a product of sulphate reduction, reacting to form iron monosulphides (Hansel et al., 2015;Riedinger et al., 2017).

Release of Cu, Pb, Zn
The effect of biostimulation on the release of metals was evaluated by determining the concentrations of metals present in the column effluents (Fig. 4).Organic starved control columns exhibited low recovery of Fe (1.6%), Zn (4.7%), Cu (0.7%), and Pb (0.1%).Zn and Cu exhibited an initial release typical of a 'wash off' of weakly sorbed or water-soluble fraction, however the recovery declined post day 8. Pb release remained consistently at ~ 0.3 mg/l through the early part of the experiment, though variability was observed post day 49 onwards (Fig. 4c).Live columns, on the other hand exhibited only 3.1%, 0.3%, and 0.1% of total Zn, Cu, and Pb, respectively.The recovery rate of metals from live columns was lower than organic starved control columns.Like in control columns, majority of the Zn and Cu were recovered in the initial wash off phase.Although minimal recovery of Pb was obtained, it exhibited a trend similar to Fe Tot .Metal mobility could be limited by the increasing pH (Fig. 2b) as higher pH typically leads to lower mobility of Zn> Cu> Pb (Sintorini et al., 2021).According to this, organic starved control should have exhibited a higher recovery of Zn and Cu, however it exhibited a similar low recovery as live columns, suggesting that pH did not play a role in the mobility of these metals.
Autoclaved columns exhibited a slightly higher recovery of Zn (5.8%), Cu (0.8%), and Pb (0.2%) as compared to organic starved control or live columns.As previously, majority of Zn and Cu was recovered in the early stages of the study which rapidly declined to negligible concentrations, suggesting "wash-off" (Fig. 4c).The elevated metal recovery may be due to the heat sterilisation that induces the transformation of amorphous ferrihydrite (or similar) to goethite thereby reducing the surface area and its ability to adsorb associated metals (Das et al., 2011).Effluent Pb concentrations displayed very little similarity to Fe tot trend.Pb concentrations decreased for the first 12 days, after which they exhibited an increase, peaking at 0.8 mg/l on day 16, followed by a further decrease and stabilising at ~0.1 mg/l for the remainder of the experiment.The decrease in Pb immediately after the Pb elevation on day 19, at acidic pH (3.5) of effluent, is unlikely due to the re-adsorption of Pb 2+ , as Pb 2+ is usually soluble under acidic conditions and the surface adsorption on minerals like iron oxyhydroxides is merely ~10% (Brown et al., 2008).The lack of relationship between Fe Tot and Pb in control effluent concentrations confirms the assumption that the increased concentrations observed in both the live and autoclaved columns (coeval with increasing Fe Tot ) is a consequence of the reductive dissolution of iron hydroxysulphates in the waste.It also supports the theory that the decline in Pb within these effluents is a result of increasing pH as Pb is observed to remain relatively constant in the control effluents which also exhibit a stable pH.

Glycerol utilisation
Glycerol was added as an additional carbon source in the live columns.Influent of the live columns contained 10 mM glycerol (Fig. 5).Concentrations of glycerol in the first few days remained between ~4.5 mM and ~6.5 mM, however, from day 12 onwards it decreased to ~1 mM and remained so until the end of the experiment.This is accompanied by a concomitant increase in iron concentrations suggesting microbial activity.This also suggests that the amount and/or rate of supply of glycerol could have been increased, an important area for further investigation and consideration for upscaling.The total glycerol added over the column test was 0.027 kg per kg of Parys Mt.

Visual, chemical, and mineralogical characterisation of post experiment waste
The Parys Mt. waste, which was originally orange-light brown in appearance, post experiment largely changed to olive-khaki green in colour.The cross-section of the column (Supplementary Figure 3) exhibited variability with light brown regions in the centre demonstrating the impact of preferential flow paths within the waste with discrete areas undergoing less extensive bioreduction.This would inherently limit the amount of target metal that could be extracted by the flow of extractant and in the latter hydrometallurgical extractions.
Mineralogical characterisation was done by XRD.The Parys Mt. waste exhibited characteristic peaks of quartz, jarosite, and goethite, which after being fed with organic starved influent post-experiment still exhibited a similar diffractogram (Supplementary Figure 2c).However, live column waste post-experiment exhibited differences from preexperiment waste, most significant being the absence of ~34.2 • 2θ peak corresponding to jarosite (Supplementary Figure 2c) suggesting the utilisation of jarosite as primary electron acceptor for DIRM.Both biotic and abiotic reduction of iron oxyhydroxides and hydroxysulphates result in generation of iron sulphides (Li et al., 2006;Riedinger et al., 2017).However, peaks corresponding to FeS were absent which may be attributed to their metastable nature and susceptibility to rapid oxidation (Rickard and Morse, 2005).Heat sterilisation of the Parys Mt. waste resulted in a slight modification of the X-ray diffractogram with two additional peaks observed at ~43 • 2θ and ~59 • 2θ, corresponding to lepidocrocite (γ-FeO(OH)).Post-experiment autoclaved waste XRD data was very similar to post experiment live waste.
Sequential extractions of organic starved waste exhibited very little redistribution of metals pre-and post-experiment (Fig. 6a).A small  amount of redistribution of Fe to reducible oxide (36.1%) and easily reducible oxide (11.1%) phase was observed (Fig. 6a).Zn exhibited negligible changes in the various extractive phases with very little redistribution to easily reducible oxide (5.7%), reducible oxide (17.5%), and residual (63.6%) phases.A small amount of Cu redistributed from residual and magnetite targeted phases to more reactive phases viz., easily reducible, and reducible phases (Fig. 6a).Pb on the other hand exhibited a minor redistribution from water soluble, carbonate, and easily reducible oxides phases to magnetite targeted and residual phases (Fig. 6a).The most notable change was the decrease of Pb proportion in the carbonate associated phase from 9.4% to 2.5%, suggesting the mobilisation of Pb to aqueous phase, however, < 0.1% lead was recovered from the aqueous phase, indicating redistribution to least reactive phases which include magnetite targeted (24.1%) and residual phase (29.1%).Similarly, very little redistribution of Fe was obtained (Fig. 6b) in live columns post experiment, despite the apparent bioreduction observed during effluent analysis.Most of the Fe post experiment was held in the least reactive magnetite-targeted (18.3%) and residual phases (48.2%).Unlike Fe, Cu redistributed from least reactive phases (res and mag; Fig. 6b) to more reactive phases (carb and ox1; Fig. 6b).Although the redistribution of Zn in post experiment live columns was only minor, the trend was similar to Cu. Pb exhibited most extensive redistribution from least reactive phases to more reactive phases with more than 50% present in the easily reducible oxide phase.
Autoclaving of Parys Mt. waste did not have a significant effect on the distribution of Fe, Cu, and Zn between sequential extraction phases when compared to live pre-experiment waste, except for Pb (Fig. 6c).Steam under pressure resulted in redistribution of Pb from more reactive phases to least reactive phases with majority of Pb partitioning into the residual (69.4%) phase.Fe distribution in post-experiment autoclaved columns was similar to their live counterparts, with majority of Fe being in the residual (45.3%), reducible oxide (27.4%), magnetite targeted (10.8%), and easily reducible oxide (8.1%) phases, in that order.Despite the lower Fe concentrations in the effluents from autoclaved columns, the impact of bioreduction in the live and autoclaved columns seem to be comparable as evident from the similarity of distribution within various phases.Bioreduction resulted in minor redistribution of Zn from residual phase to magnetite targeted (16.3%), reducible oxide (16.5%), and easily reducible oxide (7.3%).Although slight, an apparent increase in carbonate associated (1.9%) and water-soluble (2.4%) phases was also observed.Cu was predominantly present in the magnetite targeted (63.5%) and residual (24.6%) phases, which post-experimentation got redistributed to more reactive easily reducible oxide (22.5%) while majority remained within the residual phase (59.8%).Furthermore, just like Cu, an increase in carbonate associated (5.9%) and water soluble (5.6%) phases was observed.Pb redistribution in the post experiment autoclaved waste occurred from least reactive phase (res-69.4%) to more reactive phases (ox1-49.5%;carb-25.8%),because of biostimulation/bioreduction.This is similar to the trend exhibited by the live column.

Microbial community analyses
Bacterial and archaeal 16 S rRNA gene sequences generated by next generation Illumina sequencing were used to identify and characterise microbial communities within the wastes, both pre-and post-experiment.Sequences were classified from phylum to genus level and quantified by their relative abundance.A total of 14 different phyla were identified in the original waste sample, consisting of 12 bacterial phyla and 2 archaeal phyla (Fig. 7a).Of these, 10 of the bacterial and archaeal phyla were present at ≥ 1% abundance.Interestingly, there was no single phylum that clearly dominated the system with Planctomycetes being the most abundant at 23.9%, followed by Firmicutes (16.3%) and Proteobacteria (14.1%).Classes Bacilli (10.1%) and Clostridia (6.2%) represented the entirety of the Firmicutes while Alphaproteobacteria (4.9%) and Gammaproteobactaria (8.4%) represented the majority of the Proteobacteria.Fig. 7a shows all the bacterial phyla identified in preexperiment samples.
In total 21 known cultured genera were identified in Parys Mt. waste, of which 3 were from the domain Archaea (Fig. 7b).Many of the genera identified comprised of uncultured bacteria or archaea of higher order or classes.Of the known cultured genera identified, Acidobacillus (10.0%) was the most abundant, which are acidophilic bacteria known for both iron-oxidation and reduction and have been identified in acid mine drainages (AMD) (ñancucheo et al., 2016;Shiers et al., 2016).Members of this genera have also been used for recovery of Cu from mine tailings (Falagán et al., 2017) and Co from limonitic ores (Smith et al., 2017).Amongst the other genera present in relatively low abundances, species within Leptospirillum are known to oxidise iron (Ojumu et al., 2009), while those present in Desulfosporosinus (Spring and Rosenzweig, 2006), Acidiphilium (Falagán and Johnson, 2014), Metallibacterium (Ziegler et al., 2013), and Acidibacter (Falagán and Johnson, 2014) are known to be iron reducers.Previous studies on Parys Mt. waste had identified Leptospirillum spp. in the surface spoils (Bryan et al., 2004) and Acidiphilium in the AMD contaminated Afon Goch area (Jenkins et al., 2000).However, Acidithiobacillus, Thiobacillus, and Ferromicrobium, identified in the previous studies were not identified in the present study, which indicates that the waste investigated here has a microbial community distinct from the ore body of the AMD.
The microbial community composition of the organic starved columns exhibited a slight difference from pre-experiment samples.Planctomycetes exhibited an increase in relative abundance from 23.9% to 42.8%, on the other hand, Cyanobacteria and Firmicutes decreased in relative abundance (Fig. 7a).Within Planctomycetes uncultured bacterium belonging to CPLa-3 termite group was the predominant genus, exhibiting an increase from 18.5% to 40.7% (Fig. 7b).Other bacterial genera were present at low abundances of 4.0% or less.As expected, given the lack of iron reduction in the effluent analysis, iron reducing genera were not identified.Substantial changes were observed in the microbial community structure post-experiment in the live columns waste.Planctomycetes and Proteobacteria decreased significantly while Firmicutes became the predominant phyla (Fig. 7a).Similarly, at the genus level, 13 genera were identified of which 5 were present at ≥ 1% in the pre-experiment samples, while the other 8 increased in abundance from < 1% in the preexperiment samples (Fig. 7b).Methanosaeta was the only archaeal genera present at > 1% abundance.The predominant bacterial genus identified was Desulfosporosinus, members of which are strictly anaerobic and form endospores (Spring and Rosenzweig, 2006).This genus and Delsulfitobacterium (Villemur et al., 2006), also identified here, are known for their ability to reduce both Fe(III) and SO 4 2-, with the latter forming H 2 S and resulting in indirect reduction of Fe(III) (dos Santos Afonso and Stumm, 1992;Roberts et al., 2020).The second most abundant genus identified was Thermincola, below detection in pre-experiment samples, whose members are thermophilic, anaerobic, iron reducers, with very few cultured species.
The addition of glycerol to the autoclaved waste resulted in significant changes in the microbial community structure of the waste (Fig. 7) and showed some resemblance to the community structure observed in post-experiment live samples.The most obvious similarity in these samples is the dominance of the phylum Firmicutes.This phylum exhibited 6.5 times increase in abundance in post-experimental samples (78.8%) as compared to pre-experimentation autoclaved samples (12.0%).This was accompanied with a concomitant decrease in the abundance of Proteobacteria and an increase in Synergistetes, which was below detection in pre-experiment autoclaved samples.A similar effect of autoclaving was observed with iron oxyhydroxide waste fed with glycerol during a study on passive bioremediation of dye bearing effluents, in which the abundance of Firmicutes increased and Proteobacteria decreased post-experiment (Srivastava et al., 2022).Interestingly, Euryarchaeota exhibited an increase in relative abundance to 8.9% from below detection (Fig. 7), and sequences were related to members of the sulphur reducing genus Thermoplasma (Bonch-Osmolovskaya, 1994).Of the 12 recognised genera present at relative abundance ≥ 1% pre-experiment, only 9 of these remained in post experiment.As observed in post-experimentation live samples, Desulfosporosinus was the predominant genus in post-experiment autoclaved samples, exhibiting an increase to 62.1% from an initial 4.5%.Other genera observed at abundance > 1% were Alicyclobacillus (6.1%), Desulfitobacterium (4.5%), Acidibacillus (2.4%), and Escherichia/Shigella (1.2%).Except for Escherichia/Shigella, all other genera represent potential iron reducers.These spore-forming, iron and sulphate reducing bacteria explain the observations of iron reduction in autoclaved waste and the olfactory identification of hydrogen sulphide generation during experimentation.

Hydrochloric acid leaching
The leachability of Fe from post-experiment organic starved control waste was negligible with only 2.8%, 1.2%, and 0.6% Fe leached when 1 M, 0.5 M, and 0.25 M HCl were used, respectively (Fig. 8a).This is indicative of the lack of biostimulation and transformation of amorphous minerals to more crystalline phases.Bioreduction that occurred in live and autoclaved columns resulted in a slight increase in leachability of Fe as compared to pre-experiment and control column wastes.Approximately 4.2% iron was recovered from live waste postexperiment when leached with 1 M HCl (Fig. 8a), and the amount of leachable Fe decreased as: 1 M > 0.5 M> 0.25 M.Although 1 M HCl gave the highest yield of Fe from live column wastes, 0.5 M, and 0.25 M HCl resulted in a more substantial increase in Fe recovery as compared to pre-experiment waste (Fig. 8a).An increase by 3.57% and 2.82% was observed in recovery of Fe from post-experiment live column waste when leached with 0.25 M and 0.5 M HCl, respectively, while live column waste leached with 1 M HCl exhibited only a 1.47% increase in Fe leachability.Autoclaving led to a small increase in the leachability of Fe from the waste when compared to pre-experiment waste, with 1 M HCl leaching ~4.7% Fe.Post-treatment autoclave waste when leached with 1 M HCl exhibited a slight decrease in the total Fe yield.However, an increase in yield was observed when 0.25 M and 0.5 M HCl were used as lixiviants, suggesting an increase in readily leachable phases of Fe.
The leachability of Zn from the waste before treatment was lower than that observed for Fe (Fig. 8a).When compared to pre-treatment waste leachability, organic starved wastes exhibited a marginal increase in leachability by 1 M and 0.5 M, while 0.25 M had a negligible effect on the leachability of Zn.Post biostimulation, Zn leachability in 1 M, 0.5 M, and 0.25 M HCl increased to ~7.1%, 6.4%, and 5.9% respectively.Interestingly, from the perspective of processes economics, the highest recovery was obtained with 0.25 M HCl exhibiting an 8-fold increase in leachability.Autoclaving led to a slight increase in leachability of Zn in pre-experiment wastes.A 5-fold increase was observed in Zn leachability of post-experiment autoclaved wastes in response to 0.25 M HCl (Fig. 8a).
Leachability of Cu from waste followed the same trend as Zn and Fe.Negligible Cu was recovered from pre-treated waste in response to HCl leaching.Post-treatment organic starved waste exhibited a nonsignificant increase in Cu leachability (Fig. 8a).However, biostimulation substantially increased the leachability of Cu with a 15-fold increase in recovery for 0.25 M HCl (Fig. 8a).The majority of Cu is readily leachable by dilute acids as Cu is redistributed from recalcitrant mineral to more reactive phases.A minor increase in leachability of Cu from pre-treatment waste was observed in response to autoclaving.However, post-experiment autoclave wastes exhibit a comparable increase in Cu leachability as live columns.
Pb exhibited the highest leachability amongst all the metals, in response to HCl leaching.9.3%, 4.0%, and 2.0% Pb was recovered from pre-treated waste when 1 M, 0.5 M, and 0.25 M HCl were used as lixiviants, respectively.Absence of any microbial activity in organic starved columns resulted in a minimal increase in leachability of Pb (Fig. 8a).However, post biostimulation a significant increase in Pb leachability was obtained with 0.25 M HCl giving the highest recovery rate (34.73%; exhibiting 33.5 times increase in leachability), followed by 0.5 M HCl (30.9%) and 1 M HCl (29.9%).Autoclaving preexperiment waste exhibited a substantial increase in Pb leachability than any other metals, which increased further post-experiment because of bioreduction and glycerol addition (Fig. 8a).The results of Pb recovery of post-experiment autoclaved waste are comparable with their live waste counterparts.

Sulphuric acid leaching
H 2 SO 4 leaching resulted in higher Fe recovery than HCl, with the pre-test waste exhibiting 2%, 3.4%, and 2% Fe recovery from 0.25 M, 0.5 M, and 1 M H 2 SO 4 , respectively.Organic starved columns exhibited low leachability of Fe (Fig. 8b) due to lack of biostimulation as discussed above.However, Fe recovery for live biostimulated waste increased to 7.3%, 6%, and 5.1% when leached with 1 M, 0.25 M, and 1 M H 2 SO 4 , respectively.Contrary to HCl leaching results, autoclaving had negligible to marginal increase in Fe recovery for pre-test material.However, post-test the leachability increased, albeit slightly lesser than live samples (Fig. 8b).
Although leachability of Zn with H 2 SO 4 from the wastes exhibited a similar trend to that of Fe, the recovery is higher than for HCl leaching (Fig. 8b).In organic starved samples leached with H 2 SO 4 , like HCl, an increase in Zn recovery when compared to pre-experiment was observed.Zn recovery from live samples was found to increase from 1.5%, 1.1%, and 1.0% in the pre-test material, to 7.9%, 6.9%, and 6.8% post-test, when 1 M, 0.5 M, and 0.25 M H 2 SO 4 was used respectively.As discussed previously, lack of microbial activity in the organic starved columns resulted in lower recovery of Zn as compared to their live counterparts.Autoclaving the pre-experiment waste exhibited a marginal increase in Zn leachability, which post-experiment showed a further increase because of bioreduction.
In contrast to Fe and Zn, the leachability of Cu generally decreased in pre-experiment waste with 1%, 0.6%, and 0.5% total Cu recovered with 1 M, 0.5 M, and 0.25 M H 2 SO 4 , respectively.Organic starved waste exhibited a slight increase in Cu recovery (Fig. 8b), which further increased with the addition of glycerol in live samples.All the three concentrations of H 2 SO 4 gave almost the same recovery (~9.0%), suggesting that Cu partitioned into more reactive phases making it readily leachable by dilute acids.Autoclaving the pre-experiment sample resulted in a minor increase in Cu leachability (Fig. 8b), which postexperiment exhibited a substantial increase in recovery and is comparable to live counterparts.Pb recovery across all samples was minimal (~0.5%) and the concentration of lixiviant did not have any effect on the recovery, due to the inability of cold H 2 SO 4 to solubilise Pb and formation of insoluble PbSO 4 (Thornton et al., 2001).

Citric acid leaching
C₆H₈O₇ is a weaker acid than HCl and H 2 SO 4, and thus the recovery of metals from C₆H₈O₇ leaching is lower than that obtained by the other inorganic acids.Although C₆H₈O₇ is a known chelator of Fe in aqueous solutions, while other organic acids are not (Silva et al., 2009), the recovery of Fe from pre-experiment waste was lower than that achieved by HCl and H 2 SO 4 leaching (Fig. 8c).Organic starved control exhibited negligible leachability of Fe by C₆H₈O₇.The leachability of Fe from live samples exhibited an increase, with 0.5 M C₆H₈O₇ leachate showing ~9-fold increase in recovery when compared to pre-experiment.Much like the other acids, C₆H₈O₇ leaching of autoclaved pre-experiment waste exhibited an increased recovery yield, with ~1.0% recovered regardless of lixiviant concentration (Fig. 8c).In comparison to this, Fe leachability of post-experiment autoclaved waste showed an increase, but a decrease in recovery relative to live samples.Leachability of Zn from C₆H₈O₇ leached pre-experiment waste was almost negligible and remained so even after experiment for organic starved waste.Bioreduction in live columns increased the recovery of Zn to 4.3% 3.8%, and 3.6% in response to 1 M, 0.5 M, and 0.25 M C₆H₈O₇, respectively.Interestingly, autoclaved pre-experiment waste did not show any change in leachability of Zn, which contrasts with the general trend of enhanced leachability observed for other metals and/or lixiviants used.Post-experiment autoclaved samples exhibited an increase in recovery of Zn because of bioreduction, however the yield was lower than that obtained from live samples.
As with Fe and Zn, leachability of Cu from pre-experiment waste is lower than that obtained with the inorganic acids.An increase in leachability was observed in organic starved control columns with ~1.1% recovered from all the three concentrations of C₆H₈O₇.Recovery of Cu exhibited a further increase in live columns to 3.4%, 2.5%, and 1.7% with 1 M, 0.5 M, and 0.25 M C₆H₈O₇, respectively.No significant increase in leachability of Cu was obtained on autoclaving the preexperiment sample.However, post-experiment autoclaved waste exhibited an increase in Cu leachability that are commensurate with the live columns.Pb, like other metals exhibited negligible leachability which demonstrates the ineffectiveness of C₆H₈O₇ as a lixiviant for preexperimentation waste.An insignificant change in recovery of Pb was observed from organic starved waste in response to C₆H₈O₇ leaching.However, introduction of glycerol and subsequent bioreduction caused the leachability of Pb to substantially increase in live samples.15.1%, 14.1%, and 9.1% Pb was recovered from live samples when 1 M, 0.5 M, and 0.25 M C₆H₈O₇ was used as a lixiviant, respectively.Autoclaving of the pre-experiment waste exhibited a minor increase in Pb leachability (1.8% − 1 M; 1.4% -0.5 M; 0.9% -0.25 M), which post-experiment exhibited a significant increase to 11.6%, 11.2%, and 8.9% when leached with 1 M, 0.5 M, and 0.25 M C₆H₈O₇, respectively.

EDTA leaching
Leachability of Fe from pre-experiment wastes with the chelator, EDTA exhibited a similar trend to inorganic acid leaching, however the yield was lower (Fig. 8d).0.4%, 0.2%, and 0.2% Fe was recovered from pre-experiment wastes with 10 mM, 5 mM, and 2.5 mM EDTA, respectively.The recovery rates are commensurate with the rates obtained with C₆H₈O₇.Similarly, negligible recovery of Fe was obtained from organic starved waste when subjected to EDTA leaching.Live samples exhibited a slight increase in leachability of Fe however the magnitude of increase was much lower than that achieved by inorganic acids.A marginal increase in leachability of Fe was observed in autoclaved preexperiment samples (between 0.4%− 0.6%), which increased further post-experiment (between 0.9%− 1.1%).The increase in leachability of Fe in pre-experiment and organic starved columns is negligible due to their acidic pH that neutralises EDTA (pH 8.0), while the live and autoclaved post-experimentation samples were circumneutral thereby facilitating extraction of Fe in EDTA.Leachability of Zn also exhibited a similar trend to C₆H₈O₇ leaching, with ~0.6%, 0.4%, and 0.5% total Zn recovered from pre-experiment wastes using 10 mM, 5 mM, and 2.5 mM EDTA, respectively.Organic starved waste exhibited a similar trend in Zn leachability in response to EDTA (Fig. 8d).The recovery of Zn increased in live samples because of bioreduction (Fig. 8d).Autoclaving exhibited an insignificant increase in leachability of Zn from preexperimentation waste, which post-experimentation increased to ~1.8%.
Much like Fe and Zn, the recovery rates for Cu from pre-experiment waste with EDTA as the lixiviant were low (Fig. 8d).Leachability of Cu in organic starved samples showed a marginal increase as compared to the pre-experiment waste.Live samples, on the other hand, exhibited an increased leachability of Cu as compared to both pre-experiment and organic starved wastes, with 6.2%, 4.9%, and 12.4% recovered with 10 mM, 5 mM, and 2.5 mM EDTA, respectively.These recovery rates are higher than those obtained with C₆H₈O₇.Autoclaving resulted in a negligible increase in leachability of Cu in pre-experimentation waste, which increased in post-experimentation samples, but lower than live column material.Leachability of Pb from pre-experiment waste by EDTA gave the highest yield, with 7.4%, 3.2%, and 4.2% total Pb recovered using 10 mM, 5 mM, and 2.5 mM EDTA, respectively.Organic starved samples, on the other hand, exhibited a significant decrease in leachability.Bioreduction of the waste further increased the leachability of Pb between ~29% -~35% in live samples (Fig. 8d).Surprisingly, autoclaving the pre-experiment waste led to a decrease in leachability of Pb by 10 mM and 5 mM EDTA, which is in contrast with the leachability exhibited by EDTA for other metals, and for all metals by other acids.
2.5 mM EDTA, on the other hand, exhibited an increase in leachability of Pb when compared to its non-autoclaved pre-experiment equivalent.Much like live column material, post-experiment autoclaved column material also exhibited a significant increase in total Pb recovery (Fig. 8d).In general, EDTA exhibited a high rate of Pb leachability as compared to other metals, with Pb seemingly outcompeting other metals for EDTA binding sites.
Glycerol fed wastes exhibited a higher leachability highlighting the role of microbial activity in mobility of metals.In general, Fe leaching exhibited a slight increase post-test.Zn was majorly associated with residual phases within sequential extractions which didn't appear to change, yet notable increases in extractability were observed.Similarly, most of the Cu was associated with the residual phase within the waste, however some of it was also associated with the easily reducible oxide phases which would explain the increased leachability in response to lixiviant application.Pb was associated with more reactive phases and was clearly seen to move pre-and post-test.This corresponded to release of Pb concurrently with Fe, although this was later curtailed probably due to the increasing pH of the system.This suggests if pH were maintained at lower pHs it may be possible to extract Pb directly from the waste.Even though Pb was retained at later stages of the work, the fact that it had moved phases corresponded to it responding very well in the hydrometallurgical extractions, furthermore the chelation extraction was notable in the selectivity of metal removal over Feimportant in extracting value from the waste.

Conclusions
Significant changes in waste colouration, mineralogy, sequential extractions, Fe(II) release, pH, alkalinity, sulphate removal, glycerol utilisation, and microbial community structure were observed in columns lixiviated with 10 mM glycerol, as compared to organic starved control columns.This is strongly indicative of bioreduction of iron and sulphate in the media in response to biostimulation by microbes.Sequential extractions revealed a redistribution of metals within the various sequential extraction phases in the order Pb>Cu>Zn.Small changes were seen in the Fe distribution.Mineralogy indicated dissolution of jarosite on the X-ray diffractogram which is consistent with the observed iron and sulphate reduction.The hydrometallurgical extractions performed with HCl, H 2 SO 4 , citric acid, and EDTA all point to a marked increase in both the leachability of Zn, Cu, and notably Pb and selectivity over Fe.Surprisingly, the highest extraction values were sometimes found for weaker reagent strengths.The largest change was observed in Pb redistribution within various extraction phases, where it moved from least reactive magnetite-targeted and residual phases to more reactive carbonate-targeted and easily reducible oxide phases.This corelates well with the increased leachability of Pb in the period when the column effluent pH remained low enough to maintain mobility.Furthermore, the chelation extraction was notable in the selectivity of metal removal over Feimportant in extracting value from waste.Post-test analyses indicated the presence of preferential flow paths, indicating that this process needs further optimisation to achieve improved yields.This study can be considered as a foundational work in demonstrating that in situ biostimulation of iron-rich jarosite-bearing wastes can lead to iron and sulphate bioreduction and an increase in the subsequent leachability of target metals.This suggests that addition of only low-cost organic substrate followed by either diluted acids or chelators could be applied in-situ for metal recovery.Any initial organic carbon costs can be negated by using waste crude glycerol.This may open-up circular and industrial symbiotic opportunities to biologically manipulate these globally abundant wastes ahead of in situ or ex situ metal recovery.In case of metals considered un-economic for recovery, it may be that the value accrued from metals recovered can be used to offset remediation costs and the subsequent decontaminated land may instead be considered the resource.

Environmental implication
Large quantities of iron oxide/hydroxysulphate bearing wastes are generated every year which present a potential exploitable iron resource along with being suited for recovery of other co-occurring metals.Resource recovery from these 'jarosite-type' material renders them safe for application in various domains.This study exhibits the circular economy route of making metals present within the iron-rich waste more susceptible to acid/chelator leaching post biostimulation in presence of glycerol, a waste by-product of biodiesel manufacturing industry.The indigenous microorganisms present in the waste were responsible of movement of metals from less reactive to more reactive phases.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 1 .
Fig. 1.(a) Aerial view of Parys Mountain mine, N. Wales.The red arrow indicates the sampling location (53 • 23 ′ 13 N, 4 • 20 ′ 37 W).Image copyright Google Earth™ 2022 Google, Imagery; (b) Picture of the sampling location on Parys Mt.Note the algal mats on the waste surface; (c) The elemental composition of the Parys Mt. waste as determined by ICP-OES; (d) Distribution of metals in various extraction phases as determined by sequential extraction.(Res-residual phase; Mag-magnetite targeted phase; Ox2-reducible oxide phase; Ox1-readily reducible oxide phase; Carb-carbonate associate phase; DI-water soluble phase).

Fig. 2 .
Fig. 2. Changes in the physicochemical properties like pH, conductivity, and the oxidation-reduction potential in comparison to the total iron of the column effluents from (a) organic starved, (b) live, and (c) autoclaved column.Please note the difference in the axes.

Fig. 3 .
Fig. 3. Fe(II) released over time in comparison to the total iron present in the effluents from (a) organic starved, (b) live, and (c) autoclaved columns; (d) Changes in the sulphate concentration in the column effluents, with live columns exhibiting the highest release at around day 20.

Fig. 4 .
Fig. 4. Effect of biostimulation on the trend of release of metals (a) Zn, (b) Cu, and (c) Pb in column effluents.Please note the different axes.

Fig. 5 .
Fig. 5. Decrease in glycerol concentration in the effluent from live columns when compared to input extractant and its relationship with the changes in the total iron concentration.

Fig. 7 .
Fig. 7. Changes in microbial community structure exhibited by the Parys Mt waste pre-and post-experiment, in organic starved, live, and autoclaved columns at (a) phylum level and (b) genus level.(Auto-autoclaved; Org.starved and org.starv-organic starved; Pre-Exp-pre experiment).

Table 2
Physicochemical characteristics of the Parys Mountain waste.