The uranium-bearing ore used in the experiment was collected from a granite uranium deposit in Guangdong Province, China. The ore sample was crushed and sieved to obtain five size fractions: 3–5 mm (26.81%), 1–3 mm (36.28%), 0.5–1 mm (15.96%), 0.1–0.5 mm (16.37%), <0.1 mm (4.57%). A representative sample was prepared by coning and quartering for mineral phase and chemical analysis. The mineral phase was analyzed by Mineral Liberation Analysis (MLA250, FEI, Czech). The MLA analysis showed that the main mineral was granite porphyry, and the main uranium-bearing minerals were brannerite (0.21%), coffinite (0.15%) and uraninite (0.02%). It was a refractory and complex uranium ore. Other metalliferous minerals included goethite 1.59%, pyrite 0.05%, zircon 0.08%, Rutile 0.25%, etc. The main gangue minerals were albite (69.72%), quartz (4.18%), apatite (15.65%), almandine (2.82%), muscovite (1.81%), jadeite (1.71%), calcite (0.83%), etc. The uranium grade of the ore sample analyzed by ammonium vanadate titration was 0.24%. The other chemical components analyzed by X-ray fluorescence spectrometer (XRF) showed SiO₂ 55.11%, SO₃ 0.17%, CaO 8.44%, Al₂O₃ 18.08%, Fe₂O₃ 2.79%, Na₂O 8.77%, P₂O₅ 4.40%, K₂O 0.34%, MgO 0.91%, TiO₂ 0.26%, ZrO₂ 0.06%, ZnO 0.03%, MnO 0.08%.

The bacterial strains A. ferrooxidans ATCC 23270, A. thiooxidans A01 and L. ferriphilum YSK were selected to construct a bioleaching consortium for bioleaching experiments. A. ferrooxidans ATCC 23270 (Selkov et al., 2000) was obtained from American type culture collection (ATCC). A. thiooxidans A01 (Yin et al., 2014) was isolated from a coal heap drainage in Pingxiang, Jiangxi province, China, while L. ferriphilum YSK (Gao et al., 2007) was isolated from a drainage in Dexing copper mine in Jiangxi province, China. The leaching organisms were activated by inoculating in 100 mL culture medium to an initial density of ~5 × 10⁵ cells/mL (determined by hemocytometry) in 250-mL Erlenmeyer flasks. The medium was the 0K basal salt medium [(NH₄)₂ SO₄ 3 g/L, KCl 0.1 g/L, K₂HPO₄ 0.5 g/L, MgSO₄·7H₂O 0.5 g/L, Ca (NO₃)₂ 0.01 g/L, pH 2.2 adjusted with H₂SO₄, sterilized at 120 °C for 20 min] (Selkov et al., 2000; Li et al., 2017). 44.7 g/L FeSO₄·7H₂O were added for A. ferrooxidans ATCC 23270 and L. ferriphilum YSK as energy source, while 10 g/L elemental sulfur for A. thiooxidans A01. Cells of A. ferrooxidans ATCC 23270 and A. thiooxidans A01 were cultivated at 30°C, while L. ferriphilum YSK at 40 °C aerobically with 180 rpm in an incubator shaker.

The activated strains were subjected to serial adaptation in the leach liquor adsorbed with resin (resin adsorption tail liquor) and with 24.8 g/L FeSO₄·7H₂O for A. ferrooxidans and L. ferriphilum or 5 g/L S⁰ for A. thiooxidans until a constant iron oxidation rate or growth rate was achieved. Each adapted strain was scale-up cultivated in an aeration tank, and then equally mixed for subsequent column bioleaching.

To assess the effect of sulfur enhancement for uranium bioleaching in column reactors, four groups with sulfur dosages of 0.5, 1, 2, and 4 g/kg ore were designed, while no sulfur addition as blank control. The schematic of column bioleaching reactors is shown in Figure 1.

The columns were fabricated from acrylic tubes with a 10 mm-thick-wall. These reaction columns were 100 cm in height and 5 cm in internal diameter. The adsorption columns were 70 cm high with an internal diameter of 1 cm and connected to each reaction column. The column leaching experiments included two phases, acidic pre-leaching and microbial leaching (Yang et al., 2022). The column leaching turned into bioleaching phase when pH of the leaching liquor was at approx. 3.0. Bioleaching microorganisms of each strain were individually inoculated at 20% (v/v) to get a cell density of approx. 10⁸ cells/mL. The oxidation rate of ferrous iron reached 90%.

pH, redox potential, levels of Fe²⁺, Fe³⁺ and uranyl ions in feed solution (input) and leach liquor (output solution) were measured each day. The experiment was terminated when the concentration of uranyl ions in the leach liquor was less than 20 mg/L. The ore residues were sampled and analyzed for surface characterization and uranium quantification. The genomic DNA of the planktonic microorganisms and the attached microorganisms were extracted for microbial community analysis.

The pH value was measured using a pH meter (PHBJ-260, INESA, China); the redox potential was measured by a platinum electrode with an Ag/AgCl reference electrode; total iron concentration in solution was assayed by titration with EDTA; the Fe²⁺ concentration was detected by titration with potassium dichromate (K₂Cr₂O₇) (Nemati and Harrison, 2000). The surface morphologies of the original ore and leaching residues were characterized by scanning electron microscopy (SEM) (Sigma300, Carl Zeiss AG, Germany). The chemical components of the raw ore and residues were analyzed using X-ray fluorescence spectrometer (XRF) (S4 pioneer, Bruker, Germany).

The uranium concentration in solution and solid phase was titrated volumetrically with the ammonium vanadate method (Furman et al., 1951). The solid samples for titration analysis were first ground, and then the powered samples were boiled in perchloric acid till dried up for digesting the organics. Subsequently, the boiled samples were dissolved in a mixed solution of 10 mL hydrochloric acid, 3 mL hydrogen peroxide (30%) and 1 mL hydrofluoric acid. The filtrate was used for titration analysis (Sun et al., 2020).

For DNA extraction from the planktonic microorganisms, 150 mL solution sample were centrifuged at 10,000 g for 10 min to pellet the cells. Attached cells from the ore surfaces were collected from ore residues by repeated vortex and elution. Briefly, 10 g ore samples mixed with 4 g glass beads (diameter of 0.5 mm) and 50 mL sterile water were put in a 250 mL centrifuge bottle. They were vortexed in a rotary shaker at 220 rpm for 10 min. Afterwards, the mixture was centrifuged at 2,000 g for 2 min to separate the ore residue from the solution. The separated solution was centrifuged at 10,000 g for 10 min to pellet the cells. The supernatant was used to wash the ore residue circularly. The procedure was repeated for five times to collect effectively the attached microorganisms. Both the genomic DNA of the attached microorganisms and planktonic cells was extracted using the TIANamp^® Bacteria DNA kit (Tiangen Biotech Co. Ltd., Beijing, China) (Li et al., 2017). The DNA samples were checked on 1% agarose gel, and DNA concentration and purity were determined with a NanoDrop 2000 UV-vis spectrophotometer (Thermo Scientific, Wilmington, USA). The hypervariable region V3–V4 of the bacterial 16S rRNA gene was amplified with the primer pairs 338F (5′-ACTCCTACGGGAGGC AGCAG-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) in an ABI GeneAmp 9700 PCR thermocycler (ABI, CA, USA) as described previously by Wang et al. (2018). Purified amplicons were pooled in equimolar and paired-end sequenced on an Illumina MiSeq PE300 platform/NovaSeq PE250 platform (Illumina, San Diego, USA) according to the standard protocols by Majorbio Bio-Pharm Technology Co. Ltd. (Shanghai, China) (Wang et al., 2018).

The mineralogy analysis by MLA showed that the uranium grade of the ore sample was 0.24%, and the uranium-bearing minerals were mainly brannerite [UO₂·(TiO₂)₂], coffinite [UO₂·(SiO₂)_0.9·(H₂O)_0.2] and uraninite (UO₂). Uranium was tetravalent in the ore, which was hardly soluble in the aqueous solution. Thus, the ore sample was a complex and refractory uranium ore, and the addition of oxidizing agent was necessary to oxidize the tetravalent uranium to hexavalent uranium for dissolution purpose. XRF analysis showed that the iron content (2.79%) of the raw ore was sufficient for leaching of uranium. However, it would be transferred into jarosite on the surface of ore particles if only using iron-oxidizers like L. ferriphilum, or using iron/sulfur-oxidizers like A. ferrooxidans without enough sulfur as growth substrate. When using the single microbial species or single energy substrate, it would lead to the passivation accumulation on the mineral surface, which acts as a barrier against the diffusion of ions and then inhibits the uranium dissolution (Li et al., 2017). Herein, mixed iron-oxidizing bacteria and sulfur-oxidizing bacteria assist metabolic activities cooperatively and decompose minerals more effectively when presented at an appropriate Fe/S ratio in bioleaching operations. The mineral components analysis by XRF showed that the total sulfur content in the ore sample was 0.17%, indicating that sulfur enhancement for mixotrophic bioleaching was one of necessary and feasible ways in this study.

Figure 2A shows that the pH of leach liquors in the initial 5 days in the acidizing phase ascended very fast but the solution pH was still <7.0. This phenomenon was ascribed to the newly exposed acid-consuming minerals that consumed a large amount of H⁺ (Ram et al., 2020), resulting in a sharp rise in the pH of the leachate. Expectedly, the pH of each group gradually decreased with the proceeding of acidification, and bioleaching stage started by inoculation after 21 days acid pre-leaching. In the initial 13 days of bioleaching stage, the pH of the leach liquor was slightly lower in the tests with sulfur dosages of 1, 2 and 4 g/kg than those of 0.5 g/kg or absent of sulfur. Sulfur can be as energy substance for the growth of A. thiooxidans and A. ferrooxidans, which is conducive to acid production [Reaction (1)]. Subsequently, the variation trend of pH slightly fluctuated and was relatively stable in the late-bioleaching phase. It was likely attributed to that the protons could be consumed owing to the Fe²⁺ oxidation by L. ferriphilum and A. ferrooxidans simultaneously, resulting in a slight increase of pH at intervals in the bioleaching process [Reaction (2)]. Although the pH deviations of the leach liquor with different sulfur enhancement in each group were tiny in the later stage of bioleaching, it can still be seen that the pH value was a little lower than the control absent of sulfur (Figure 2A).

The uranium extraction could be achieved at low pH and high redox potential (Qiu et al., 2011; Tavakoli et al., 2017). The variation in redox potential of the solution is mainly determined by the presence of iron as the Fe³⁺ or Fe³⁺/Fe²⁺ ratio in the leaching solution (Yue et al., 2014). The redox potentials in the acid pre-leaching phase were all less than 450 mV (Figure 2B), and the reaction between the leaching solution and some acid-consuming substances in the minerals were the main reactions during this phase, and the U (VI) in the ore was dissolved by H⁺ attack [Reaction (3)]. After inoculation, the level of ferric iron and redox potential of the leach liquor increased exponentially with time (Figures 2B, 3B–F). Unexpectedly, the redox potential of the leach liquor was a little lower at the sulfur dosage of 0.5 g/kg. Furthermore, the uranium dissolution also ascended rapidly (Figures 3C–F). Afterwards, the redox potential kept a little higher than 650 mV and remained relatively stable in the assays with 1–4 g/kg sulfur dosages.

The hexavalent uranium in the ore had been dissolved in the acid pre-leaching phase [Reaction (3)], while the tetravalent uranium is hardly soluble in the aqueous solution. In the acid pre-leaching stage, the uranium extraction with sulfur addition was approx. Ten percent lower than the control (Figure 3A). It was possibly ascribed into that the sulfur powder in the reaction column formed into passivation layer on the ore surface in acid pre-leaching stage, which would inhibit the ion diffusion and uranium dissolution to some extent (Pathak et al., 2017).

The sulfur-oxidizer A. thiooxidans can rapidly oxidize the sulfur layer on the ore surface (He et al., 2014; Kim et al., 2021) and reduce the pH in the leaching system, which was conducive to uranium bioleaching. Fe³⁺ concentration gradually increased due to the oxidation capability of A. ferrooxidans and L. ferriphilum (Figures 3B–F). The uranium concentration in the leach liquor was 600 mg/L with sulfur enhancement (0.5–4 g/kg), while it was less than 350 mg/L for the control. The Fe³⁺ generated by iron-oxidizers can oxidize U (IV) to UO22+ by an indirect mode [Reaction (4)] as described previously (Tributsch, 2001; Qiu et al., 2011). Extracellular polymeric substances (EPS) secreted by the bacteria can enrich Fe³⁺ and form EPS-Fe³⁺ complex to increase the oxidation efficiency of U (IV) to UO22+, namely indirect-contact mode [Reaction (5)] (Tributsch, 2001; Sand et al., 2001; Yu et al., 2011). Furthermore, the SO42- generated by A. thiooxidans could have complexation reaction with UO22+ in the leachate under acidic conditions (pH ≤ 4.5) [Reaction (6)] (Vercouter et al., 2008; Abhilash and Pandey, 2013b), which can promote the uranium dissolution kinetics. After 77 days of leaching, the total uranium extraction of the assays with a sulfur dosage of 0, 0.5, 1, 2, and 4 g/kg were 73.6, 82.5, 86.2, 84.5, and 86.3%, respectively (Figure 3A). The results indicated that the gross uranium extraction could be increased by approx. 12.6% with appropriate sulfur enhancement.

To find clearly out the differences of the sulfur enhancement in the bioleaching phase, the uranium extraction yield in acid pre-leaching can be neglected (the same starting point as average extraction yield in the acid pre-leaching period) (Supplementary Figure 1). Thus, the uranium extraction was 35.06% with sulfur enhancement of 1–4 g/kg in the initial 13 days, which was increased by approx. 11% vs. that of 0.5 g/kg sulfur or in the absence of sulfur. Notably, the uranium extraction with sulfur enhancement increased by approx. 20%, compared to these in the absence of sulfur. However, the excessive increase in sulfur to some extent inhibited the uranium dissolution, and thus, it maintained the highest at the sulfur dosage of 1 g/kg. After 27 days of bacterial oxidation, the uranium extraction with 0.5 g/kg sulfur was also higher than that of 2 g/kg. It showed that the uranium leaching was positively proportional to the sulfur dosage in the first 13 days of bioleaching. Afterwards, the leaching of the assays with sulfur addition of 0.5 and 1 g/kg was faster than that with 2 and 4 g/kg.

Furthermore, in order to evaluate the uranium extraction rate, the uranium dissolution kinetics in column bioleaching process was analyzed. The bulk or granular ores was subjected in the column bioleaching reactors, which is a typical dynamic process of liquid-solid multiphase reaction. If the uranium ore particles are regarded as spherical, the leaching kinetics of uranium ore can be described by the shrink kernel model (SCM). In this model, the uranium dissolution rate can be dependent on the following rate-limiting steps, which is the surface chemical reactions controlled kinetic model (Eq 1) or the internal diffusion through product layer controlled kinetic model (Eq 2) (Abdel-Aal, 2000; Sun et al., 2017).

Where, t is the reaction time (d); k₁ is the chemical reaction rate constant; k₂ is the diffusion rate constant, x is the fraction of uranium extraction.

To determine the uranium rate-limiting step, Eqs 7, 8 were used to fit the experimental data, and the fitting degree was evaluated by correlation coefficient (R²) values. The results of each model are plotted in Figure 4. The apparent reaction rate constants (k₁ and k₂) and correlation coefficients (R²) for the two model above are given in Table 1. Figure 4 and Table 1 showed that the R² fitted by the second model was >0.98. This indicates that the internal diffusion through the product layer-controlled model was more applicable to the uranium column bioleaching process with sulfur enhancement. Besides, it also showed that appropriate sulfur dosages could increase the chemical reaction rate (Figure 4A). However, the excessive sulfur was likely to inhibit the diffusion rate owing to the formation of a passivation layer.

SEM images showed a smooth surface of the raw ore (Figure 5a). A large smooth surface of the ore from control assays was also observed (Figure 5b). This indicates that the passivation layer would be generated on the ore surface. For those bioleaching residues with the sulfur enhancement of 0.5, 1, 2, and 4 g/kg, rougher surfaces with more porosity and erosion traces by bacteria were visible (Figures 5c–f). Moreover, Table 2 shows that the sulfur proportion of the ore residues in the upper and lower layer in the column was around 18% with sulfur enhancement. This was higher than that of the control. Thus, more sulfur in the residues would contribute to a better permeability of the ore layer. Consequently, a better uranium extraction was achieved. These results demonstrated that sulfur enhancement in the column reactors would strengthen the porosity of passivation layer, which could contribute to the improvement of the ore permeability and the uranium dissolution is thus promoted.

As shown in Figures 5b–f, the surface of the lower layer ore was much smoother than that of the upper one, indicating that a lesser extent of erosion occurred. This was also reflected by the higher uranium extraction of the upper layer. Furthermore, this observation was consistent with the lower uranium concentration in the residues (Table 2). The possible reasons for this phenomenon were as follows: firstly, the lower layer contained less oxygen than the upper layer and it endured greater pressure. These might inhibit the bacterial growth and activities (Yang et al., 2022). Moreover, bacteria turned into the decline stage at the bottom of the column reactors due to lack of nutrients in the late-bioleaching phase. The leaching process would lead to the accumulation of other metals beyond the bacterial tolerance, which was inhibitory to bacterial growth (Sasaki et al., 2009; Qiu et al., 2021). Additionally, passivation substances gradually accumulated and wrapped the ore surface at the later stage of uranium leaching. Thus, the contact between bacteria and minerals was impeded. Consequently, the leaching of uranium was restricted (Wei et al., 2020). As can been seen in Table 2, the elements Al, Na and Mn were dissolved into the leach liquor simultaneously, resulting in a decrease of the content in the leaching residues. The content of Ca, Fe and K on the residue surface was 2–3 times that of the raw ore. We infer that jarosite as the passivation layer was formed on the surface of the residues (Tuovinen and Bhatti, 1999; Yang et al., 2022). As the upper layer had higher uranium extraction, the passivation layer was most likely formed in the lower layer of the column reactors.

Figure 6 shows the effect of sulfur dosages on bacterial community composition. Sulfur-oxidizing bacterium A. thiooxidans was the dominant population both in the leach liquor and on the ore surface in the column reactors (Figure 6). It was reported that a higher proportion of A. thiooxidans could metabolize complex sulfur sources and further promote the acidification of the ores in a complicated and adverse environment. In this way, the leaching efficiency was improved. The contribution of A. thiooxidans to the leaching efficiency was especially obvious in the initial bioleaching phase (Brune and Bayer, 2012; Li et al., 2017). It is noteworthy that the iron-oxidizers (L. ferriphilum and A. ferrooxidans) and sulfur-oxidizer A. thiooxidans maintained a good balance (nearly 1:1) on the ore surface at the sulfur dosage of 1 g/kg (Figure 6). These assays with 1 g/kg sulfur showed the best uranium extraction (Figure 3A). Cells of A. ferrooxidans exerted iron oxidation ability if both of the iron and sulfur coexisted in the substrate. Therefore, the iron-oxidizing and sulfur-oxidizing bacteria with a good population balance on the ore surface would decompose ore more effectively if presented in a favorable sulfur enhancement in the uranium column reactors.

To better interpret the effects of sulfur enhancement on uranium bioleaching in column reactors, a possible mechanism model (Figure 7) was proposed based on the data from the uranium dissolution reactions and kinetics, ore residue characteristics and bacterial community structures (Figure 7).

As shown in Figure 7, U (VI) in the ore could be dissolved by H⁺ attack, and the U (IV) could be dissolved by the Fe³⁺ as indirect mode or by EPS-Fe³⁺ as indirect-contact mode. Thus, the uranium extraction could be achieved at low pH and high redox potential. The low pH can be maintained by the oxidation of A. thiooxidans to sulfur or reduced inorganic sulfur compounds, and the high redox potential can be achieved by the oxidation of A. ferrooxidans and L. ferriphilum to ferrous iron. The sulfur-oxidizer A. thiooxidans cannot only reduce the pH in the leaching system, but also rapidly oxidize the sulfur layer on the ore surface, which was conducive to uranium bioleaching. The Fe³⁺ generated by iron-oxidizers can oxidize U (IV) to UO22+ by an indirect mode [Reaction (4)]. Extracellular polymeric substances (EPS) secreted by the bacteria can enrich Fe³⁺ and form EPS-Fe³⁺ complex to increase the oxidation efficiency of U (IV) to UO22+, namely indirect-contact mode [Reaction (5)]. Furthermore, the SO42- generated by A. thiooxidans preferred to complex with UO22+ to form UO₂ (SO₄) n2-2n in the leachate under acidic conditions (pH ≤ 4.5) [Reaction (6)], which can promote the uranium dissolution kinetics. Meanwhile, A. thiooxidans can change the structure of the passivation layer via sulfur oxidation activities (Li et al., 2017) and improve the permeability of the ore layer. Admittedly, excessive Fe³⁺ is likely to result in the generation of jarosite, and excessive sulfur would lead to the accumulation of elemental sulfur or polysulfide on the mineral surface in the bioleaching process. Besides, the results showed a good balance of the iron-oxidizing bacteria and sulfur-oxidizing bacteria on the ore surface would be one of the important factors for uranium bioleaching performance. Therefore, these results indicate that the quantitative balance of Fe²⁺ and sulfur, and the balance of the iron-oxidizing bacteria and sulfur-oxidizing bacteria are two of the key factors in the uranium bioleaching process. Suitable amount of sulfur addition is critical to improve the leaching kinetics.