Mineral carbonation of magnesium-rich rocks is a naturally occurring process that captures CO₂ from the atmosphere and produces magnesium carbonates, a stable and metastable group of minerals capable of storing carbon dioxide for long periods of time (hundreds and thousands of years). A wide variety of magnesium-rich rocks occur in nature. These include basalts, gabbros, a variety of peridotites, metamorphic rock series such as serpentinite and greenschists (chlorite-bearing rocks), and weathered Mg-rich laterites. Ultramafic rocks, peridotites (e.g., lherzolites, harzburgites, and dunites), are the Mg-rich rocks with the highest magnesium content (45–35 wt% MgO), followed by kimberlites (diamond-bearing volcanic rocks) with about 30 wt% MgO and gabbros with about 10 wt% MgO. The low-grade metamorphism of these rocks produces serpentinites and greenschists with Mg contents of 35–5 wt% MgO, depending on the type of metamorphosed rock (Haldar, 2020).

The region of the Balkan Peninsula, as part of the global Alpine-Himalayan collision belt, is known for its large peridotitic massifs that crop out along two large NW-SE trending ophiolitic belts of Jurassic age (e.g., Robertson et al., 2009; Schmid et al., 2020). During the evolution of the Tethyan Ocean, ultramafic and mafic rocks were metamorphosed and significant amounts of serpentinites were formed by almost complete serpentinization of ultramafic rocks. Serpentinites can be formed wherever ultramafic rocks are infiltrated by carbon dioxide-poor hydrothermal fluids. The final mineral composition of serpentinite is usually dominated by lizardite, chrysotile and antigorite (minerals of the serpentine subgroup), remnants of olivine and pyroxenes, spinels, and variable amount of magnetite. Lizardite, chrysotile and antigorite all have approximately the chemical formula Mg 3 Si 2 O 5 ( OH ) 4 but differ in minor components and crystal structure.

Exposure of peridotites and serpentinites to air and water results in their mineral carbonation, i.e., the precipitation of magnesium and calcium carbonate minerals. It is estimated that global natural weathering of ultramafic rocks contributes to the sequestration of 2.4 Gt of carbon dioxide per year (Bullock et al., 2021). The most famous example is a large peridotite massif in Oman, which covers an area of almost 1 × 10 5 km² and sequesters approximately 1 × 10 5 tons of CO₂ per year through a natural passive carbonation process (Kelemen et al., 2020). Serpentinites and dunites are the rocks with the highest estimated theoretical carbon sequestration potential, 722 kg and 690 kg CO₂ per ton of rock, respectively. Among minerals, forsterite (786 kg CO₂/t) and lizardite (663 kg CO₂/t) have the highest carbon sequestration capacity (Bullock et al., 2021).

Enhanced weathering is a carbon dioxide removal (CDR) technology based on accelerated rock dissolution that sequesters CO₂ through alkalinity production (Bullock et al., 2022). The essence of this strategy is to accelerate the process of rock dissolution, release Mg ions into solution, and then create the alkaline conditions necessary for the crystallization and precipitation of magnesium carbonates. This process can be illustrated by the chemical weathering of serpentinite minerals. These minerals dissolve in acidic environments according to the following chemical equation (Bullock et al., 2022):

In alkaline conditions, magnesium ions react with carbonate ions forming hydrated magnesium carbonates such as hydromagnesite that precipitates from the solution:

Geological carbon storage in, e.g., magnesium-rich rocks according to Kelemen and Matter (2008) is limited by a:

Bodenan et al. (2014) created a map of the global distribution of ultramafic massifs and mine tailings suitable for mineral carbonation, located less than 300 km from large industrial CO₂ emitters. Unlike countries such as Canada and Australia, where magnesium-rich rocks and mine tailings are often located hundreds or even thousands of kilometres away from any industrial source of CO₂, in Serbia—and the Balkans in general—magnesium-rich ultramafic rocks suitable for mineral carbonation are often located in close proximity to major industrial CO₂ sources. The most economical way to transport CO₂-rich flue gas from industry is by pipeline. For distances up to 300 km, recompression is not required, whereas longer pipelines need recompression stations (Bodenan et al., 2014). Many magnesium-rich rock deposits and mining tailings in this region are situated well within this 300 km threshold, creating a strong opportunity to develop facilities for CO₂ capture and storage from industrial flue gases.

The Zlatibor Mountain region, for example, lies less than 300 km from major Serbian CO₂ emitters such as coal-fired power plants and cement factories. That large ultramafic massif represents a promising location for development of carbon capture and storage infrastructure.

A large-scale mineral carbonation process in this context would involve extracting magnesium from ultramafic rocks and converting it into stable magnesium carbonate minerals, which can sequester CO₂ over geological timescales. Given the close proximity of major CO₂ sources, the availability of CO₂ is unlikely to be a limiting factor; the main challenge is optimizing magnesium extraction from local rocks.

We experimentally investigated two approaches for magnesium extraction:

Experiment 1 aimed to compare direct chemical leaching with 1 M sulfuric acid against bioleaching using approximately 1 mol of elemental sulfur (32 g). Theoretically, complete bacterial oxidation of 1 mol of sulfur produces 1 mol of sulfuric acid. While bioleaching proceeds more slowly, it could offer significant cost savings—critical for making large-scale CO₂ sequestration economically viable.

Experiment 2: simulated heap leaching, which requires coarse rock particles, thus avoiding the high energy costs of fine grinding necessary for chemical reactors or bioreactors. Although larger particle sizes generally reduce leaching efficiency, heap leaching offers lower capital and operational costs. The optimal technology choice must therefore balance extraction efficiency with energy and cost considerations.

To the best of our knowledge, this is the first study to investigate serpentinites bioleaching in bioreactors, providing a novel approach to coupling magnesium extraction with sustainable CO₂ removal strategies.

The following amounts of selected chemical elements were measured for the serpentinite sample (g/kg): Si 325; Mg 130.26; Fe 51.15; Ca 9.46; Al 5.08; Ni 1.99; Na 0.1. Mineralogical analysis of the sample by XRD (Table 1; Supplementary Figure 1) revealed typical serpentine mineral phases such as lizardite and chrysotile, as well as clinochlore (chlorite) and enstatite (pyroxene). More detailed mineralogical and petrographical analysis showed that the studied rock sample is lepidoblastic to blastogranular in texture and massive, locally schistose in fabric. The majority of the rock (around 70 vol%) is composed of serpentine that creates the rock matrix that encloses remnants of primary minerals – olivine, orthopyroxene, clinopyroxene and Cr-spinel (Supplementary Figures 2a,b). Olivine appears as oval relicts or cluster of relicts with high birefringence, whereas orthopyroxene is usually elongated and displays typical pyroxene cleavages. Some orthopyroxene grains exhibit exolutions of clinopyroxene along the cleavage planes, which show textural evidence of ductile deformations characteristic for tectonite peridotites (Supplementary Figure 2c). Clinopyroxene is present as rare and small (0.5–1 mm) remnants. Cr-spinel is disseminated as individual and irregularly shaped grains with characteristic dark-brownish pleochroism (Supplementary Figure 2d) in transmitted-light optical microscope. SEM-EDS analyses results of the above-mentioned minerals are given in Supplementary Table 1.

Crystallographic analysis of the solid residues after sulfuric acid dissolution in beakers and bacterially mediated mineral dissolution in bioreactors shows that chrysotile (present in the serpentinite sample) disappeared and four new crystal phases were detected (Table 1; Supplementary Figures 3, 4).

During the dissolution experiment with 1 M sulfuric acid in a beaker, the solution pH increased from negative values at the beginning of experiment to approximately 0.6 at the end of experiment (data not shown). Figure 1a shows changes in solution pH values during the experiments in percolators and bioreactors. Changes in magnesium concentrations during the experiments are presented in Figures 1b–d. The final concentration of magnesium in the beaker experiment with 1 M sulfuric acid reached approximately 24 g/L (24,000 ppm) corresponding to Mg extraction of 92 ± 2.8%. Final extraction values of Fe and Ca were 82.65 ± 6.9% and 28.53 ± 3.1%, respectively. Concentration of Si in final solution was approximately 110 mg/kg, which corresponds to 0.17 ± 0.06% of silica extraction.

The magnesium concentration in bioreactors reached approximately 10 g/L (10,000 ppm) corresponding to a Mg extraction of 95 ± 3.9%. Final extractions of Fe, Ca and Si were 60.29 ± 2.9%, 26.42 ± 1.4% and 0.5 ± 0.07%. A two-sample t-test showed no statistically significant difference between Mg extractions during bioleaching experiment in bioreactors and leaching with 1 M sulfuric acid in chemical reactors (t = 0.61, p = 0.59). Application of the two-sample t-tests comparing Mg leaching in (1) chemical reactors versus percolators and (2) bioreactors versus percolators showed that the differences in Mg extractions were statistically significant (p < 0.05). Final concentrations of magnesium after dissolution with sulfuric acid and microbially mediated dissolution in percolators were approximately 19 g/L (19,000 ppm) and 4 g/L (4,000 ppm), respectively, corresponding to magnesium extractions of 73 ± 2.9% and 16 ± 1.4%. Extractions of Fe, Ca and Si after sulfuric acid dissolution in percolators were 52 ± 4.7%, 10.57 ± 1.9% and 0.23 ± 0.4%, respectively. Dissolution with microbially generated sulfuric acid led to lower extractions of Fe, Ca and Si: 8.8 ± 1.1%, 5.28 ± 0.9 and 0.09% ± 0.01, respectively.

The data on extracted Mg, Fe, Ca and Si are summarized in Table 2.

The serpentinite sample was dominated by typical serpentine group minerals such as lizardite and chrysotile, and the latter was completely dissolved during the acid leaching experiments in which Mg was efficiently extracted from the rock in most cases. The very low extraction of silica in all experiments means that the silicate framework remained intact. This finding confirms the incongruent mineral dissolution of serpentine group minerals by acid treatment: Mg²⁺ ions are released into solution due to the preferential attack of H⁺ ions on Mg-OH bonds in the crystal lattice of the serpentine mineral phases (Lacinska et al., 2016). The efficiency of magnesium extraction from serpentine minerals follows the trend: chrysotile > lizardite > antigorite (Arce et al., 2017). Consistent with the results of our study, previous studies reported almost complete magnesium extraction from serpentinite after acid dissolution, with sulfuric acid identified as the most efficient leaching agent (Teir et al., 2007; Arce et al., 2017).

The percentage of Mg extracted by biogenic sulfuric acid leaching in bioreactors was comparable to that obtained by 1 M sulfuric acid leaching in beakers and percolators, although the microbial process was significantly slower. From an economic feasibility perspective, the main products of the carbon removal process are tons of sequestered CO₂. This “commodity” has a relatively low market value,¹ so it is imperative to develop a low-cost process. Sulfuric acid has been shown to be efficient for extracting magnesium from ultramafic rocks at temperatures between 70 °C and 90 °C (Yoo et al., 2009; Lacinska et al., 2016). In order to keep the energy consumption low, the application of sulfuric acid at 30 °C with extended reaction time was tested. The dissolution of ultramafic rocks consumes a significant amount of acid (chemical Equation 1), and high acid consumption could make a commercial process unprofitable. Therefore, we tested Mg extraction in bioreactors using elemental sulfur as a feedstock for the sulfur-oxidizing acidophilic bacteria Acidithiobacillus thiooxidans. The use of elemental sulfur instead of sulfuric acid should reduce the operating costs for Mg extraction—sulfur is produced in large quantities as a by-product of oil rafination process; advantages of using sulfur instead of sulfuric acid also include lower transportation costs, simpler storage and lower risk of environmental hazards. Production of one ton of biogenic sulfuric acid requires 0.327 tons of sulfur. If we apply the qualified presumption that the price of 1 ton of sulfur is 100 USD, and the price of one ton of sulfuric acid is 165 USD, it can be calculated that the cost of raw material for biogenic production of 1 ton of sulfuric acid is only 32.7 USD.

The potential disadvantage of using chemical reactors and bioreactors for mineral carbonation is the significant capital cost of purchasing this sophisticated equipment, the operational cost of maintenance, and the need for highly skilled engineers to control and optimize the process. For this reason, we additionally tested Mg extraction using laboratory-scale columns (percolators) that simulate a “heap leaching” process that is routinely used on a commercial scale for the production of metals such as copper, zinc, nickel, cobalt, and uranium (Ghorbani et al., 2016; Petersen, 2023). Heap leaching is based on the construction of a heap that is irrigated with an acidic leaching solution; the recirculated pregnant leaching solution is collected in a pond and then transported to downstream processing for metal extraction. Power et al. (2010) tested a bioleaching approach for the extraction of magnesium from chrysotile asbestos mine tailings using small columns. They reported a maximum Mg extraction of 14.3% in columns amended with sulfur and inoculated with the sulfur-oxidizing bacteria Acidithiobacillus thiooxidans. This result is comparable to the Mg extraction achieved in the glass column bioleaching experiments in this study (16 ± 1.4%). Substantially lower Mg extraction during bioleaching in percolators in comparison to bioleaching in bioreactors is not surprising. Much bigger particle size in percolators gives lower surface area for chemical reactions. Also, bioreactors provide controlled conditions (aeration, temperature, stirring, pH) that promote growth and activity of the bacterial culture. On the other side, percolators were running at room temperature, without additional aeration or stirring, which affects the activity of the bacterial culture and sulfuric acid production in percolators.

This study demonstrates that serpentinites from the Zlatibor ophiolite massif can be effectively processed for magnesium extraction, offering a viable pathway toward CO₂ sequestration via mineral carbonation. Both chemical leaching with 1 M H₂SO₄ and bioleaching with Acidithiobacillus thiooxidans in stirred tank bioreactors achieved high Mg recoveries (>90%), while coarse-particle percolator leaching with H₂SO₄ yielded 73% Mg recovery. Bioleaching in percolators, although less efficient (16%), confirms the feasibility of low-cost, biologically driven processes. In contrast to the bioleaching of copper sulfide ores, where oxidation of sulfide minerals like pyrite generates sulfuric acid, bioleaching of serpentine minerals does not produce acid. The process can be engineered to consume nearly all the sulfuric acid in the leaching solution, preventing acid accumulation. Moreover, the solid residues remaining after magnesium extraction are chemically inert and pose no risk of acid mine drainage. These characteristics indicate that magnesium bioleaching from serpentinites is inherently less environmentally hazardous, offering a more sustainable alternative compared to copper recovery from sulfide ores.

The results highlight that coarse-particle heap leaching could deliver magnesium concentrations in pregnant leach solutions comparable to those obtained from fine-particle chemical or bioreactor systems, while significantly reducing comminution energy requirements and capital expenditure. Integrating such low-energy extraction methods with downstream carbonation could enable scalable, cost-effective CO₂ removal, especially when coupled to the valorisation of by-products such as magnesium carbonates. The future research direction will include experimental testing of the Mg carbonation in pregnant leach solution in pressure reactors at mild temperature, in order to produce hydrated magnesium carbonate minerals, and test their potential application as supplements for production of low-carbon cement.