# Carbon dioxide storage through mineral carbonation

## Abstract

Carbon capture and storage (CCS) has a fundamental role in achieving the goals of the Paris Agreement to limit anthropogenic warming to 1.5–2 °C. Most ongoing CCS projects inject CO2 into sedimentary basins and require an impermeable cap rock to prevent the CO2 from migrating to the surface. Alternatively, captured carbon can be stored through injection into reactive rocks (such as mafic or ultramafic lithologies), provoking CO2 mineralization and, thereby, permanently fixing carbon with negligible risk of return to the atmosphere. Although in situ mineralization offers a large potential volume for carbon storage in formations such as basalts and peridotites (both onshore and offshore), its large-scale implementation remains little explored beyond laboratory-based and field-based experiments. In this Review, we discuss the potential of mineral carbonation to address the global CCS challenge and contribute to long-term reductions in atmospheric CO2. Emphasis is placed on the advances in making this technology more cost-effective and in exploring the limits and global applicability of CO2 mineralization.

## Key points

• Carbon capture and storage has a key role in achieving the goals of the Paris Agreement.

• CO2 storage through mineral carbonation extends the applicability of carbon capture and storage by enabling storage in areas previously not considered feasible.

• The rapid mineralization of CO2 through injection into reactive rock formations increases storage security.

• Carbon mineralization in basaltic rocks offers a global storage potential that exceeds anthropogenic emissions.

• The method can be used for the subsurface storage of CO2, and potentially other environmentally important gases, through water capture, although this approach is water-intensive.

• Considerable efforts are needed to accelerate the deployment of CO2 storage through mineral carbonation, including more widespread operation in diverse conditions.

## Introduction

Global concentrations of atmospheric CO2 are higher than at any other time in human history. From a pre-industrial baseline of 280 ppm, CO2 concentrations reached 407 ppm in 2018. Growth rates of 2.3 ppm per year have been observed over the past decade, approximately 100 times faster than natural changes1. Increasing CO2 concentrations can largely be attributed to the use of fossil fuels, which, through an enhanced greenhouse effect, raise global average temperatures and, in turn, affect climatic, ecological and social systems2. To limit these effects, the 2015 Paris Agreement aims to constrain anthropogenic warming to 1.5–2 °C (ref.3). Solutions are, therefore, needed to promote a substantial and sustained reduction in the net flow of CO2 into the atmosphere, while, at the same time, ensuring energy needs are still met. According to the International Energy Agency4, the goals of the Paris Agreement can be achieved by applying and pushing already available climate-mitigation technologies to their maximum practical limits.

Carbon capture and storage (CCS) is one such climate-mitigation strategy2,4, describing a range of processes for CO2 capture, separation, transport, storage and monitoring, usually from large point sources. For example, CCS is considered the key technology for reducing emissions from fossil-fuel power plants while these are still operational; limiting emissions from many industrial processes, such as steel, aluminium and cement production; and to deliver ‘negative emissions’ by removing and sequestering CO2 directly from air by the second half of the century. Indeed, all proposed pathways to limit global warming to 1.5 °C require some degree of CO2 removal5, with an estimated total of 190 GtCO2 needing to be stored3. However, large-scale deployment of CCS has yet to be achieved; present facilities capture and store only ~40 MtCO2 annually6. Thus, a considerable scale-up in technology is required to meet the aims of the Paris Agreement, specifically through the operation of >2,500 large-scale CCS facilities by 2040 (ref.6).

Currently, 14 out of the 18 operational large-scale CCS projects worldwide (each capturing >400,000 tonnes of CO2 annually) are based on enhanced oil recovery6, whereby CO2 is used to obtain the last remains of an oil field. These projects deploy one of the most common approaches of carbon storage, that is, injection of gaseous, liquid or supercritical CO2 into subsurface reservoirs, such as sedimentary basins. In such systems, CO2 can be physically trapped in porous rocks below an impermeable cap rock (structural trapping), some of which becomes trapped in small pores (residual trapping) and, over time, dissolves in groundwater (solubility trapping) and reacts with the subsurface rocks to form stable carbonate minerals (mineral trapping). As the storage progresses from structural to mineral trapping, the CO2 becomes more immobile, increasing the security of storage (Fig. 1a) and decreasing the reliance on the efficacy of the cap rock.

Several key challenges face the storage of CO2 in sedimentary basins. First, mineral trapping may be limited by the absence of the silicate-bound divalent metals needed for carbonate formation and can take thousands of years, owing to low rock reactivity7 (Fig. 1a). Second, the large volume of carbon that needs to be stored to achieve the climate goals requires the identification and characterization of many new storage reservoirs. Third, as most of the injected CO2 will likely remain in the gaseous, liquid or supercritical phase long term, it will tend to migrate back to the surface if not adequately stored. Even though the likelihood of such leakage from a well-regulated site is predicted to be negligible8, the lack of long-term storage feasibility has inhibited the broader application of sedimentary storage of CO2.

Alternative CCS methods have, therefore, been proposed to overcome the limitations of sedimentary injection, including in situ mineral carbonation9,10. In situ mineral carbonation aims to accelerate a natural process — the vast storage of carbon in rocks over millions of years — at a rate fast enough to contribute to climate-change mitigation. With this approach, the captured carbon is stored through its injection into reactive rocks, such as mafic or ultramafic lithologies, which contain high concentrations of divalent cations, such as Ca2+, Mg2+ and Fe2+, for rapid mineralization to calcite (CaCO3), dolomite (CaMg(CO3)2) or magnesite (MgCO3). Mineral carbonation can be further promoted by the dissolution of CO2 into water before or during its injection, achieving solubility trapping immediately11 and mineral trapping within 2 years (Fig. 1b) at 20–50 °C (ref.12). In situ mineralization results in a negligible risk of the CO2 migrating back to the atmosphere both over the short term (due to the dissolution of CO2 and the density-related inhibition of surface migration) and the long term (due to conversion into carbonate minerals).

Supercritical, liquid or gaseous CO2 can also be injected into reactive rock formations under favourable geological conditions, including offshore locations, where the risks of potential leaks are mitigated by the overlying seawater and the presence of a low-permeability sediment layer near the seafloor–seawater interface13,14. In such cases, mineralization may be slowed by the need for the injected CO2 to dissolve into the formation waters before it is mineralized. Mineral carbonation therefore provides an attractive alternative or addition to the more common sedimentary injection CCS techniques: it offers an expanded geographic range of onshore and offshore storage reservoirs (including large flood basalt provinces, large massifs of mantle peridotite and oceanic ridges) and, in turn, increases the opportunities for pairing carbon sources and sinks, and, thus, possibly reduces costs.

In this Review, we discuss the past and present state of in situ subsurface mineral CO2 storage in addressing the global CCS challenge. Emphasis is placed on assessing the recent advances that could make mineral carbonation a viable alternative for future CCS activities and exploring the limits of this technology in terms of potential sites and gas compositions. We recognise that ex situ mineralization — whereby CO2 interacts with a reactive mineral feedstock at the surface near emission sources — is an alternative CCS strategy (Box 1). However, given its limitations in regards to cost and scale, we constrain discussion to subsurface in situ mineralization.

## The mineral-carbonation process

### The geochemistry of mineral carbonation

Mineral carbonation proceeds through the reaction of water containing dissolved CO2 with rocks, notably mafic or ultramafic rocks. Water charged with CO2 is acidic, with a typical pH of 3–5, depending on the partial pressure of CO2, water composition and temperature of the system. This acidic solution promotes the dissolution of silicate minerals, such as pyroxene, a common mineral in basalt and peridotite:

$$2{{\rm{H}}}^{+}+{{\rm{H}}}_{2}{\rm{O}}+({\rm{Ca}},{\rm{Mg}},{\rm{Fe}}){{\rm{SiO}}}_{3}\leftrightharpoons {{\rm{Ca}}}^{2+},{{\rm{Mg}}}^{2+},{{\rm{Fe}}}^{2+}+{{\rm{H}}}_{4}{{\rm{SiO}}}_{4}$$
(1)

Such reactions promote CO2 mineralization in two ways: first, protons are consumed, neutralizing the acidic gas-charged water and facilitating the precipitation of carbonate minerals as the pH of the water increases; and, second, they provide cations (Eq. 1) that can react with the dissolved CO2 to form stable carbonate minerals. The degree to which the released cations form minerals depends on the element, pH and temperature.

Dissolved calcium readily reacts with CO2 in aqueous solution at temperatures below ~300 °C, forming calcite (CaCO3) and/or aragonite once the solutions are supersaturated15,16. Dissolved magnesium precipitates as the carbonates magnesite (MgCO3) and dolomite (CaMg(CO3)2) at temperatures above ~65 °C (refs17,18,19); at lower temperatures, the precipitation of these minerals is kinetically inhibited. Under such conditions, less stable hydrous Mg-carbonate minerals, such as hydromagnesite (Mg5(CO3)4(OH)2·4H2O), dypingite (Mg5(CO3)4(OH)2·5H2O) and nesquehonite (MgCO3·3H2O), can form20. Dissolved Mg2+ is also readily incorporated into clay minerals such as smectite, limiting its availability to promote carbonation. The degree to which dissolved iron (Fe2+) combines with dissolved injected CO2 to form carbonates in the subsurface remains unclear. Under oxic conditions, Fe2+ oxidizes before it can react to form a divalent metal carbonate, and, thus, the mineral siderite (FeCO3) is only rarely observed in modern sedimentary and basaltic rocks21. The formation of the mineral ankerite (CaFe(CO3)2), however, may be favoured under certain conditions at ambient to moderate temperatures when the pH of the solution is low enough to prevent Fe2+ from oxidizing to Fe3+ (refs22,23).

Other important factors for efficient subsurface carbon mineralization are the permeability and/or active porosity of the host rock formation24. The pores and fractures provide pathways for migrating fluids, access to mineral surfaces that contribute cations to the mineralization and space for the carbonate precipitates. The overall mass of carbon-bearing precipitates is also affected by the formation of other secondary minerals, most importantly clay minerals, but also minerals such as zeolites25 and anhydrites in seawater systems26. These minerals compete with carbonates for the divalent cations liberated from dissolving primary minerals and for the available pore space.

Carbonate and associated secondary mineral precipitates typically have a larger volume than their primary source minerals27 and, therefore, can clog flow pathways, where they precipitate and/or seal reactive surfaces28. However, the volume expansion during precipitation reactions can lead to cracking and the opening of fractures, which increase permeability and expose new surfaces to the fluid phase and, hence, may promote subsurface carbonation29,30. It should also be noted that, owing to the acidity of water–CO2 solutions, they tend to dissolve minerals during injection, opening up pore space and flow paths near the injection well. The precipitation of pore-filling secondary minerals is expected to occur only at a distance from the injection well after sufficient dissolution of the host rock has neutralized the acidic CO2-rich injection fluids27,31.

### Natural mineral carbonation

Natural mineral carbonation is most efficient in mafic and ultramafic rocks, owing to their high reactivity and the abundance of divalent metal cations contained in silicates32. Of these, basaltic rocks are the most abundant: most of the ocean floor, ~70% of the Earth’s surface and >5% of the continents is basaltic (Fig. 2). Abundant flood basalt fields are found in central India, Siberia, the United States, Canada and Yemen. Although the occurrence of basalts onshore is limited, the weathering of basaltic rocks on the continents and volcanic islands is responsible for ~30% of the natural drawdown of CO2 from the atmosphere attributable to continental silicate weathering, demonstrating the relative advantage of these rocks compared with other terrestrial rocks for mineral carbonation33.

Natural analogues for large-scale CO2 mineralization are found in various environments. In terms of scale, one of the most substantial natural analogues for CO2 mineralization is the carbon uptake of the oceanic crust; basalts in volcanic submarine geothermal systems receive substantial amounts of CO2 from the degassing of magma intrusions located in their roots. The oceanic crust is typically 6–7 km thick and has a remarkably consistent stratigraphy on a global scale34,35. Hydrothermal circulation through ridge flanks is focused in the uppermost 1 km of the ocean crust, resulting in extensive CO2–water–basalt interaction. This hydrothermally active crust mineralized ~40 MtCO2 annually36,37.

The storage potential of hydrothermally active geothermal systems located onshore in Iceland, the largest land mass above sea level along the mid-oceanic ridges, has been estimated by direct measurements of CO2 bound in carbonates in drill cuttings from three basalt-hosted geothermal fields. Although these carbonates are precipitated over long timescales (10,000–300,000 years), the results provide insight into the permeability and active porosity of natural systems and indicate that young and fresh basalts can naturally store >100 kg of CO2 per m3 (ref.38). On the basis of this estimate, the theoretical storage capacity of the ocean ridges is on the order of 100,000–250,000 GtCO2 — orders of magnitude larger than the amount of CO2 that would be derived from the burning of all fossil fuel24. This value agrees with other estimates that verify the enormous storage capacity of both sub-ocean13,14,39 and onshore40,41 basalts.

Evidence for natural carbonation is apparent at various other onshore locations. For example, in Oman, tectonically exposed mantle peridotites remove CO2 directly from the atmosphere, with this CO2–water–rock interaction resulting in the formation of travertines. Owing to the reactivity of the Omani peridotites, they are estimated to consume in the range of 10–100 ktCO2 per year through in situ carbonation42. The complete natural carbonation of the peridotites leads to the formation of listwanites, with the ultramafic rocks transforming through a series of reactions: serpentine + olivine + brucite → serpentine + magnesite → magnesite + talc → magnesite + quartz. This mineralogical transformation serves as a geological analogue for in situ CO2 mineralization43. Similarly, in West Greenland, extensive carbonate mineralization in basalts associated with petroleum migration was documented21, suggesting that CO2-bearing fluids may be readily mineralized, even in extensively altered rocks. The natural carbonation of mafic and ultramafic mine-waste tailings is also commonly observed44,45,46.

## Laboratory-based experimental studies

Experimental studies of mineral dissolution and precipitation rates provide insight into the fundamental processes behind mineral carbonation and are vital for the design of field-scale operations. Numerous experimental studies have verified the capacity of silicate minerals to release cations and, thereby, promote the formation of carbonate minerals in the presence of CO2-rich fluids. Many of these studies have focused on the dissolution rates of the primary silicates that contain the divalent metals necessary for carbonate precipitation. Emphasis on mineral dissolution stems from the commonly accepted view that mineral dissolution is the rate-limiting step in the mineral-carbonation process47,48.

An extensive number of dissolution studies of divalent-metal-bearing silicates have been reported on minerals such as olivine49,50,51,52,53,54, pyroxene55 and plagioclase56,57,58,59,60,61,62,63, as well as volcanic glasses64,65,66. The rates of Ca2+ and Mg2+ release from mafic minerals and rocks are often highly dependent on the fluid pH (Fig. 3). The mineral dissolution rates increase dramatically with a decrease in pH at the acidic conditions likely encountered near the CO2 injection point (Fig. 3). The dissolution rates for Al-bearing minerals, such as plagioclase (labradorite), and volcanic glasses are slowest at neutral pH and then increase at higher pH. By contrast, the rates of Al-free minerals, such as olivine (forsterite) and pyroxene (diopside), typically decrease continuously with increasing pH, such that the dissolution of these minerals is sluggish at the conditions at which carbonates tend to precipitate (Fig. 3). The large decrease in the reaction rates of most of these minerals with increasing pH suggests that the dissolution of reactive host rocks near the injection well of CO2-charged water will be rapid but the subsequent precipitation of secondary phases will be slower, more diffuse and at some distance from the injection site.

The minimum in the dissolution rates of many minerals at near-to-neutral pH has led to the proposal of a two-step ex situ carbonation process, in which divalent-metal-bearing silicates are first dissolved in strong acid, followed by the formation of carbonate minerals at more basic conditions. Olivine dissolution exhibits the fastest Mg2+ release rates at acidic conditions67. This observation justifies the interest in this mineral, or rocks rich in this mineral (such as peridotites), as a feedstock for ex situ mineral carbonation (Box 1) because smaller volumes of materials will be required to supply the required cations for the mineralization process. Note that the dissolution of the common sedimentary rock minerals quartz, kaolinite and the alkali feldspars does not liberate the divalent cations needed to promote the formation of common carbonate minerals.

Direct carbonation to form divalent-metal carbonates has also been demonstrated. Many of these studies focused on the carbonation of olivine and serpentines, owing to their widespread availability and abundance in Mg. Investigations of forsterite carbonation at 30 °C and 95 °C at varying partial pressures of CO2 revealed that magnesite formed only at 95 °C and 100 bar of CO2 pressure68. At temperatures below 65 °C, less stable hydrous Mg-carbonate minerals, such as nesquehonite, have been commonly observed to form through the carbonation of ultramafic rocks69,70. The precipitation of magnesite at lower temperatures, however, has been observed in experiments with water-saturated supercritical CO2 and in which the mineralization was enhanced by the addition of ligands to bind and dehydrate the Mg2+ ions, thereby accelerating magnesite precipitation71.

Most experimental studies on serpentine have explored mechanical or thermal treatment methods to accelerate the carbonation of this mineral72. The carbonation of Ca-bearing phases, such as glassy and crystalline basaltic rocks, has also been verified experimentally22,73,74.

One of the main limitations of laboratory-based experimental studies is their timescales; experiments are usually run for days, weeks or months, rather than years. The spatial scale is also limited, complicating the transfer of experimental results to field-scale operations. To extend the results over longer timescales and spatial scales, geochemical modelling is commonly used25,48,75,76. This approach is based on the construction of conceptual models of the processes of interest, linking observations from natural systems with results from experimental studies. Such calculations have been widely applied for simulating the dominant processes of mineral carbonation, with results suggesting that the carbonation of mafic and ultramafic rocks through water–CO2–rock interactions should lead to substantial mineralization of the injected carbon over decade-long timespans75,76. These results are a good indicator of the general processes involved in mineral carbonation but are limited owing to the natural variability and complexity of field-scale operations.

## Field mineral-carbonation studies

### Subsurface mineral carbonation

The results of experiments, geochemical modelling and natural analogues have motivated two larger-scale field projects. These projects have demonstrated the potential of in situ carbon mineralization by injecting CO2 into underground basaltic reservoirs for its rapid conversion into carbonate minerals, providing a safe, permanent storage solution for the captured carbon (Fig. 4).

#### The CarbFix project

The pilot phase of the CarbFix project, funded by the European Union, was undertaken in 2012 near the Hellisheiði geothermal power plant in southwest Iceland. A total of 230 tonnes of pure CO2 and a CO2–H2S gas mixture from the geothermal plant were fully dissolved in locally sourced groundwater during their injection to a depth of ~500 m into basaltic rocks (Fig. 4a). The temperature of the target basaltic reservoir was 20–50 °C. The method requires large amounts of water: ~25 tonnes of water for each tonne of gas injected to fully dissolve the CO2 at depth. However, as the gas-charged water is denser than fresh water, solubility trapping occurs immediately11. Injection of the acidic gas-charged water accelerates metal release from the bedrock and, hence, the formation of carbonate minerals12,77. The carbonation process was quantified using reactive and non-reactive tracers and isotopes, which revealed the rapid mineralization of the injected CO2 (refs12,78), with >95% of the injected gas mineralized within 2 years (Fig. 1b). The injection also led to an increase in the mass of the subsurface biota79, but carbon isotope measurements demonstrate that this biota increase contributed negligibly to the subsurface carbon fixation.

Following the success of the initial CarbFix project in Hellisheiði, the project was upscaled starting in 2014 in a hotter, deeper reservoir, with a stepwise increase in the amount of gases injected80,81. The acidic gases (CO2 and H2S) were captured directly from the power plant exhaust stream by their dissolution into pure water (condensed steam from the power plant turbines) in a scrubbing tower. The resulting gas-charged water was injected to a depth of ~800 m into the basaltic reservoir at temperatures of ~250 °C. Owing to the acidity of the injected solution, it is strongly undersaturated with respect to the primary and secondary minerals of the basaltic reservoir near the injection well31. Mineral dissolution gradually increases the pH of the gas-charged fluid to a range suitable for carbon mineralization, which occurs at a distance from the injection well. As a consequence, there is no sign, to date, of a decrease in the system injectivity since the initiation of the carbon injection in 2014. At present, >50% of the injected carbon is fixed as carbonate minerals within months of its injection in this upscaled system80.

The CarbFix project currently captures and stores ~33% of the CO2 emissions from the Hellisheiði power plant81, or ~12,000 tonnes annually, with the aim to increase injection to ~90% of the CO2 from the plant before 2030. Moreover, injection of CO2 from a second geothermal plant operated in the area will begin in 2021.

#### The Wallula basalt pilot demonstration project

In 2013, nearly 1,000 tonnes of pure liquid CO2 were injected into the Columbia River flood basalts near Wallula, Washington, USA82. The CO2 was provided in tanks, and the gas stream heated and pressurized before injection. The resultant liquid CO2 stream was injected into two brecciated basalt zones at a depth of 800–900 m and the reactions between the CO2, formation waters and basaltic subsurface were monitored (Fig. 4b).

Geophysical surveys indicated that a portion of the injected CO2 was trapped as free-phase CO2 in the interflow zones below the massive basalt cap rock. However, side-wall cores retrieved from the storage formation 2 years post-injection revealed the presence of carbonate nodules, composed primarily of ankerite, located in pores and fractures of the basalts23,83, confirming that some of the CO2 was mineralized within 2 years of injection.

Mass-balance calculations to determine the quantity of the injected CO2 that had been carbonated could not be performed and no quantitative data have been reported to date to evaluate what fraction of the injected CO2 was mineralized or remained as free-phase CO2 (ref.82). Nevertheless, carbon and oxygen isotope analysis of the ankerite nodules collected after the injection confirmed that the carbon in these minerals originated from the injected CO2 (refs23,83). Energy-dispersive X-ray spectroscopy of the side-wall cores also showed a progressive enrichment in Fe from the centre of the ankerite nodules, indicating that the mineralization derives material from the dissolution of the host basalt as the nodules form, rather than from reprecipitated carbonates. Extensive surface and well-bore monitoring several years after injection confirmed that no substantial leakage of CO2 has occurred from the storage formation since its injection84.

### Risks and challenges

Many of the main risks associated with the conventional injection of supercritical CO2 into sedimentary basins also apply to carbon-mineralization projects, including the contamination of water resources and induced seismicity. The main risks are discussed next.

#### CO2 buoyancy versus water penalty

The two in situ mineral-carbonation projects adopted different methods for carbon injection: the injection of water charged with CO2 (Fig. 4a) or the injection of liquid or supercritical CO2 (Fig. 4b).

Liquid and supercritical CO2 are routinely stored and transported commercially, and their compositional integrity can be assured. Efficient transportation of CO2 is possible given its low viscosity, high density and low critical temperature (31 °C) and pressure (74 bar). However, owing to its buoyancy, it is essential to ensure that the storage reservoir is securely sealed during and after the injection of a pure CO2 fluid phase. This requirement is especially challenging during onshore injection into volcanic rocks because these formations are often highly fractured85.

The injection of an aqueous solution of CO2 into the subsurface offers both advantages and challenges. The method is both simple and cost-effective80, and the risks of leaks are mitigated as CO2-charged water is denser than the corresponding CO2-free water. The major drawback of this approach is that a large quantity of water is needed to dissolve the CO2 gas. However, this water could be sourced from the target reservoir, therefore providing access for monitoring the chemistry of the injected gas-charged fluid and preventing pressure build-up in the reservoir owing to injection. As an example, during the CarbFix pilot injections, the amount of water pumped continuously from the two nearest monitoring wells was equal to the amount of water co-injected with the CO2 and, hence, no net change in water pressure was observed in the storage formation. The mass of water needed to dissolve 1 tonne of CO2 is a function of the capture pressure, as shown in Fig. 5a; the mass required decreases from >100 tonnes of water per tonne of CO2 at a pressure of 3 bar to <35 tonnes of water at a CO2 pressure of 25 bar. Abundant fresh water is not readily available for this process in many parts of the world, but in such cases, seawater may provide an adequate alternative.

One additional advantage of CO2 injection as a water-dissolved gas is that it might considerably lower the cost of CCS. Carbon dioxide and other acidic gases can be captured directly from CO2-rich exhaust streams in fresh water or seawater by mixing with the exhaust gas in a scrubbing tower80. The energy to pressurize CO2-charged water to pressures of up to 25 bar at 25 °C (the injection conditions used in the pilot CarbFix project) is less than that required to pressurize pure CO2 to the liquid state at this temperature (Fig. 5b). However, the injection of liquid CO2 is less energy-intensive than the injection of CO2 dissolved in water at pressures greater than 25 bar (Fig. 5b).

Although dissolving CO2 in water prior to its injection limits the risk of leakage, injecting supercritical CO2 may offer advantages in sub-ocean and/or other environments, because fewer wells may be needed to inject an equivalent volume of CO2 for storage. The low-permeability sediment cover of sub-ocean reservoirs also provide primary seals to trap CO2 injected in a buoyant, supercritical state14. However, this environment also offers vast seawater resources for the dissolution of CO2 prior to injection and could, thus, increase the efficiency of this alternative approach and minimize energy needs for CO2 capture. The optimal approach at potential sites likely depends on factors such as the location and depth of the target reservoir, as well as the economics and efficiency of an ocean-based storage system; thus, site-specific feasibility studies are required. Safety is also a consideration, and as offshore sites are far from inhabited areas, the risk posed to humans from potential CO2 leaks and other related liabilities is reduced13.

#### Induced seismicity

One of the main risks of CO2 injection of any kind is induced seismicity. The potential for damage to infrastructure and public concern over seismic events can result in CCS projects being shut down. Therefore, utmost care should be taken to minimize these risks.

There are >180,000 active or formerly active injection wells in the United States for both waste-water disposal and enhanced oil recovery. Induced seismic events have been associated with ~10% of these wells, with high injection rates being the dominant trigger for induced seismicity86. High injection rates also seem to have triggered the induced seismicity observed during commissioning at the Hellisheiði power plant in Iceland87.

Three years prior to the scaled-up gas injection, the CarbFix target reservoir was taken into service as a reinjection zone for waste water from the power plant. Reinjection started on 1 September 2011 with a high flow rate of ~500 kg s−1, resulting in an excess injection pressure of ~28 bar. Microseismicity increased immediately in the area north of the injection sites, with the largest seismic events being a sequence that included two magnitude 4 earthquakes on 15 October 2011 (ref.88). This problem was addressed by introducing a workflow through which preventive steps, including the adjustment of the injection rates, are taken to minimize the risk of induced seismicity87,89. The CarbFix project applies this framework and, following its implementation, the annual number of seismic events greater than magnitude 2 in the area has decreased from 96 in 2011 to one in 2018 (ref.90), which is considered satisfactory and demonstrates that the project is being operated within its regulatory boundaries.

The CarbFix example suggests that, prior to injection, a site-specific study of the regional seismicity must be performed to determine the seismic risk and should include a thorough characterization of depths, times, locations and magnitudes of seismic events. The risk of induced seismicity might be greater when injecting dissolved CO2, owing to the large volume of fluid that would need to be injected into the subsurface. For example, the injection of 1 MtCO2 per year (which is approximately the injection rate of CO2 into a single well at the Sleipner CCS project)91 as a dissolved water form would require ~20 injection wells with depths of ~500 m and injection rates of ~50 l s−1, which is a common injection rate for such shallow wells. Such a project could be fed with water from the target injection reservoir to prevent excessive pressure build-up in the system and to minimize the need for an additional water source. The risk of induced seismicity affecting infrastructure and the general public can also be mitigated by injecting CO2 into offshore reservoirs, away from populated areas and infrastructure that could be affected92.

#### Groundwater contamination

Another potential risk of injecting CO2 is groundwater contamination. During the initial phase of CO2 injection — prior to substantial carbon mineralization — a metal-rich plume can form, causing the concentrations of certain elements to exceed proposed drinking-water limits. The main metals of concern are Ni, Al and Cr, but other metals, such as Fe and Mn, can become toxic to biota at high concentrations93. This risk is highest before the precipitation of carbonate and other secondary minerals, such as clays, Al-oxide and Fe-oxide, and hydroxides, which can scavenge potentially toxic metals. The potential of carbonates to scavenge a metal plume has been described in detail in studies of natural carbonation during volcanic eruptions94,95. Ground contamination is not a major concern at offshore locations, given the lack of underground drinking-water sources.

Deployment of CCS technologies is not yet at the rate and scale needed to address the current global-warming challenge on a global scale. To date, the CarbFix and Wallula projects are the only examples of field-scale demonstrations of mineral carbonation. The results of these projects have demonstrated the rapid mineralization of CO2 when it is injected at rates of ktCO2 per year. Further upscaling is essential to obtain a more comprehensive understanding of the mineralization process under diverse conditions. There are, however, planned advances that might help accelerate mineral carbonation to the global scale necessary to limit atmospheric carbon concentrations (Fig. 6).

### Combining with direct air capture

Favourable conditions for subsurface carbon mineralization (for example, the availability of subsurface reactive rocks) are not equally distributed across the globe. Therefore, the pairing of large emission sources and suitable rock formations may be held back by the long distances involved and high costs of pipeline infrastructure. As CO2 mixes rapidly, resulting in approximately equal atmospheric concentrations everywhere, direct air capture (DAC) installations could be effective when installed near suitable reactive rock formations for carbon storage (Fig. 2) and appropriate energy sources92,96.

DAC technologies capture CO2 directly from the atmosphere, in contrast to point-source CCS technologies, which capture CO2 from more concentrated industrial streams. In a DAC system, large volumes of air flow through a filtering unit containing either a solid or liquid that can selectively remove CO2 from the atmosphere by processes such as absorption or adsorption. The resulting concentrated CO2 stream can either be stored underground or used as a feedstock for industrial applications. DAC may be an important approach in managing emissions that are challenging or costly to eliminate at the source (such as aeroplane emissions), and by combining DAC and CCS, it may be possible to create negative emission pathways during the latter part of the century5.

DAC technologies are, however, still immature and, to date, have only been demonstrated on the scale of ktCO2 per year. The primary limitation in the implementation of DAC is the high cost, currently estimated in the range of approximately US$90–900 per tonne of CO2 (refs84,97). The high cost is mostly due to the energy requirements, especially the thermal energy required for CO2 desorption: at present, ~3.4–10.7 GJ of energy is required for every tonne of CO2 captured84. Strategies that minimize the thermal requirements of DAC processes will be crucial for reducing the operating costs of these systems, enabling their scale-up to MtCO2 or GtCO2 per year. The efficiency of the DAC process is inevitably directly linked with the CO2 emissions generated by its energy source; for maximum efficiency, the energy demand of the DAC process should be supplied by available low-grade industrial-waste heat or low-CO2-emitting geothermal heat98. The coupling of DAC and mineral carbonation is being explored in the CarbFix2 project (which started in August 2017), and a DAC unit has been installed at Hellisheiði. The air-derived CO2 gas stream is then dissolved into water, injected and mineralized, achieving a negative emission pathway98. Scale-up of this DAC technology in combination with the CarbFix injection technology is being explored at various sites in Iceland. Looking towards the future, the Solid Carbon project in British Columbia is assessing DAC technologies with a similar in situ mineralization approach at an offshore site in the Cascadia Basin on the Juan de Fuca plate99. ### Moving offshore The increased storage security of subsurface mineral storage and, hence, diminished need for monitoring could enable the coupling of CCS with DAC for carbon storage in remote locations. For example, it has been proposed that CO2 could be extracted from the atmosphere using renewable wind energy and coupled to sub-ocean basalt storage in a remote offshore environment near the Kerguelen Islands in the southern Indian Ocean92. Although increased operational experience under different conditions is needed to enable remote operation at a reasonable cost, upscaling could provide high-capacity CO2 storage through in situ mineralization in basalt reservoirs with low risk of leakage, few environmental effects and minimal human inconvenience. Two projects are currently exploring offshore injection for carbon mineralization. In the north-eastern Pacific Ocean, the potential of injecting CO2 into oceanic basalts below marine sediments is being assessed offshore from Washington state and British Columbia. A pre-feasibility study was conducted under the US Department of Energy’s CarbonSAFE programme to evaluate technical and non-technical aspects of storing ~50 million tonnes of CO2 at an offshore storage complex located in the Cascadia Basin99. The study concluded that there are sufficient reservoir storage capacity, transport options and CO2 sources to meet such targets. A total of ~40 MtCO2 is generated per year onshore from nearby stationary sources; this could provide a steady stream of CO2 to the offshore reservoir, and DAC technology could be added to these carbon sources. Conditions at the Cascadia Basin site might favour CO2 injection as a supercritical fluid, owing to the low risk of degassing at ocean depths of >800 m. The CarbFix2 project is preparing for the injection of CO2 dissolved in seawater into submarine basalts. Initial results indicate that seawater, despite its high concentration of dissolved elements, is a suitable medium for carbon mineralization. Its applicability is, however, more limited than that of fresh water in terms of its temperature and pH range, owing to the formation of various secondary minerals that could affect the efficiency of the injection100,101. ### Mineral carbonation in other rock types The chemistry and reactivity of different rock types, the porosity and permeability of the potential reservoir, and the pressure and temperature of the reservoir during CO2 injection all strongly influence the rate of mineral carbonation reactions32,66,102,103. The most feasible formations for carbon mineralization are ultramafic and mafic rocks, owing to their reactivity, and, in the case of basalts, abundant pore space available for carbon storage. Other rock formations have the potential for mineral carbonation, including less reactive rock types, but it is yet to be established which formations contain high enough concentrations of the required cations for efficient CCS through carbon mineralization. The identification of suitable formations could extend the applicability of the method already established in basalts and peridotites. ### Water capture of impure CO2 gas streams The most widely applied and mature capture technology for CCS to date is amine capture, whereby an amine solvent is used to absorb CO2 from flue gas104. The CO2 is then recovered by heating the solvent and subsequently compressed and injected. A potential advantage of mineral carbonation is the possibility of using water capture instead of amine capture. The energy demand for pressurizing pure CO2 gas is far greater than that for water pressurization (Fig. 5b). However, the energy demand of any capture process and, thus, the energy cost depends on the purity of the CO2 stream. This issue is less significant for the amine-capture process, because most of the energy demand is for supplying heat for CO2 recovery, rather than the compression of the gas stream104. The estimated energy required to capture CO2 from air and from a gas stream containing ≤20% CO2 by water capture is compared with that required using an amine solvent in Fig. 7a. In this example, the efficiency of the water-capture process, assuming inefficient compression (80% efficiency) and that some CO2 (20 vol%) remains in the gas phase after passing through the capture plant, was set at 64%. This is approximately the efficiency of the ongoing injection operations at Hellisheiði, Iceland. The energy cost is based on an average global electricity price of US$0.13 kWh−1.

These cost estimates depend on the specific infrastructure requirements and energy prices at potential storage sites and, thus, could vary considerably from site to site. However, the energy cost of the capture process is by far the largest share of the operational cost of any CCS activity. Therefore, these comparisons serve to illustrate the applicability of the different capture processes for dilute CO2 streams (<20 vol%).

The energy cost of the amine-capture method is highly dependent on the energy supply for the heat used to recover the CO2. For gas streams containing <20 vol% CO2, the energy costs of amine capture are, to date, less than those for water capture when the energy can be supplied for free (waste heat) or by using natural gas. However, when electricity is the only source of energy for the amine-capture process, the cost of energy for water capture is less than that for amine capture for gas streams containing ≥15 vol% CO2 (Fig. 7a).

The water demand for the water-capture process at 30 bar is presented in Fig. 7b. At low CO2 concentrations, the water demand is high but decreases with increasing CO2 concentration. Owing to the low partial pressure of CO2 in air, ~55,000 tonnes of water would be needed to capture 1 tonne of CO2 from the atmosphere (which contains ~400 ppm CO2) at 30 bar, whereas only 22 tonnes are needed to capture 1 tonne of pure (100%) CO2.

These analyses demonstrate the feasibility of the water-capture method: for CO2 streams containing <15 vol% CO2, amine capture might be a more suitable method, owing to the lower energy cost and the high water demand of the alternative capture method. However, the water-capture method could be a better option for more concentrated (≥15 vol%) CO2 gas streams. The high demand for water and how it is sourced should be considered; in some cases, waste water from other processes at a site could be used. A great advantage of this method is the direct injection of the CO2-charged water into the subsurface without requiring additional separation or purification steps. Moreover, this approach enables the co-injection of other environmentally important gases, such as H2S (ref.80).

## The future of mineral carbonation

Recent advances have positioned in situ mineral carbonation as a viable technology for limiting at least some of the global anthropogenic CO2 emissions. In cases in which injection into sedimentary basins is not feasible, in situ mineral carbonation could offer a safe and economic alternative. Laboratory and field experiments, specifically the CarbFix and Wallula projects, have demonstrated the potential of the method. However, the only ongoing industrial-scale injection is at Hellisheiði, Iceland.

Considerable effort is therefore needed to accelerate the deployment of CO2 storage through mineral carbonation to gain experience of operating such projects under diverse conditions. Demonstrated successes will further increase public confidence in the carbon-mineralization process and motivate more widespread application of the method, including the construction of large-scale CCS systems.

For example, the long-term storage security of in situ mineral carbonation offers the possibility of injection into offshore locations, where the largest masses of mafic rocks exist. Storage at such sites could be facilitated by pairing DAC with mineral carbon storage and an appropriate energy source to enable extensive carbon storage beneath the basaltic ocean floor (Fig. 2). However, the regulatory framework for offshore storage projects will depend strongly on the nature and location of the specific CO2 sources and reservoirs, their corresponding jurisdictions and the governing laws99,105.

The biggest hurdle in applying in situ carbon mineralization on a greater scale is not technological, but the financial incentive for CCS. The current cost of capturing, injecting and storing CO2 still exceeds the cost of emitting, leading to the shutdown of many projects before they were in operation. For example, since the launch of the European Union’s carbon trading scheme in 2005, the cost of the rights to emit CO2 in Europe has been too low to motivate the development and application of CCS106. This lack of incentive rather than scientific and technical barriers has resulted in the slow deployment of CCS technologies.

The rising prices of CO2 emission quotas over the past 2 years (from ~€4.5 in May 2017 to ~€27 per tonne of CO2 as of 31 June 2019)106 are likely to change the CCS environment in the near future; the quota price of €27 in June 2019, for example, is higher than the current on-site costs of the CarbFix injection. In the United States, recent tax legislation provides financial incentives for CO2 storage of up to US$50 per tonne for saline aquifer injection and up to US$35 per tonne for enhanced oil-recovery injection107. Nevertheless, this financial incentive may still be too low to motivate many new carbon-storage projects.

The competitive cost of in situ mineral carbonation, when operated under favourable conditions, compared with other CCS methods may, however, lead to increased implementation in the future. In cases in which CO2 is captured from emission gases by dissolution into water at the source and the resulting gas-charged water directly injected into reactive subsurface rocks, the energy costs may be substantially reduced compared with alternative technologies (Fig. 7a). This approach could also provide additional value by co-capturing and storing other acidic gases, such as H2S and SO2, through subsurface mineralization, as demonstrated at Hellisheiði.

Human CO2 emissions already exceed global silicate weathering uptake by at least an order of magnitude108. Although anthropogenic emissions may decrease as fossil fuels are replaced by renewables, numerous industries, including the cement and metal industries, cannot readily avoid carbon emissions. Thus, there is a need for some form of CCS. The technology and advances discussed in this Review suggest that carbon mineralization can help to address this issue. To maximize the potential of mineral carbonation, it will be key to define the practical constraints on the application of existing technology (such as rock and gas compositions), to fill the knowledge gaps in our understanding of the fundamental processes underlying efficient mineral carbonation and to streamline the processes to limit their cost. These advances require road maps for the most probable reaction paths in natural systems, including studies of reaction rates of different rock types in the subsurface. Determining the effect of the composition of the injected fluid and the potential use of chemicals or catalysts to alter reaction rates would help to quantify the extent and consequences of host-rock carbonation at various sites. The most prominent options for carbon mineralization can already be identified, but only through increased understanding of the limitations and potential of the method will its wider applicability be realized.

## References

1. 1.

Blunden, J. & Arndt, D. S. State of the climate in 2018. Bull. Am. Meteorol. Soc. 100, Si–S305 (2019).

2. 2.

Intergovernmental Panel on Climate Change. Global warming of 1.5°C (IPCC, 2018).

3. 3.

United Nations Framework Convention on Climate Change. Report of the conference of the Parties on its twenty-first session, held in Paris from 30 November to 13 December 2015 (UNFCCC, 2016).

4. 4.

International Energy Agency. World energy outlook 2017 (IEA, 2017).

5. 5.

Rogelj, J. et al. in Global warming of 1.5°C. Ch. 2 (IPCC, 2018).

6. 6.

Global CCS Institute. The global status of CCS 2018 (Global CCS Institute, 2018).

7. 7.

Benson, S. et al. in IPCC Special Report on Carbon Dioxide Capture and Storage (eds Metz, B. et al.) Ch. 5 (Cambridge Univ. Press, 2005).

8. 8.

Alcalde, J. et al. Estimating geological CO2 storage security to deliver on climate mitigation. Nat. Commun. 9, 2201 (2018).

9. 9.

Seifritz, W. CO2 disposal by means of silicates. Nature 345, 486–490 (1990).

10. 10.

Lackner, K., Wendt, C. H., Butt, D. P., Joyce, E. L. & Sharp, D. H. Carbon dioxide disposal in carbonate minerals. Energy 20, 1153–1170 (1995).

11. 11.

Sigfusson, B. et al. Solving the carbon-dioxide buoyancy challenge: the design and field testing of a dissolved CO2 injection system. Int. J. Greenh. Gas Control 37, 213–219 (2015).

12. 12.

Matter, J. M. et al. Rapid carbon mineralization for permanent disposal of anthropogenic carbon dioxide emissions. Science 352, 1312–1314 (2016). Reports on the results from the rapid mineralization of the CarbFix pilot injection.

13. 13.

Goldberg, D. S., Kent, D. V. & Olsen, P. E. Potential on-shore and off-shore reservoirs for CO2 sequestration in Central Atlantic magmatic province basalts. Proc. Natl Acad. Sci.USA 107, 1327–1332 (2010).

14. 14.

Goldberg, D. S., Takahashi, T. & Slagle, A. L. Carbon dioxide sequestration in deep-sea basalt. Proc. Natl Acad. Sci. USA 105, 9920–9925 (2008).

15. 15.

Ellis, A. J. The solubility of calcite in carbon dioxide solutions. Am. J. Sci. 257, 354–365 (1959).

16. 16.

Ellis, A. J. The solubility of calcite in sodium chloride solutions at high temperatures. Am. J. Sci. 261, 259–267 (1963).

17. 17.

Saldi, G. D., Jordan, G., Schott, J. & Oelkers, E. H. Magnesite growth rates as a function of temperature and saturation state. Geochim. Cosmochim. Acta 73, 5646–5657 (2009).

18. 18.

Johnson, N. C. et al. Olivine dissolution and carbonation under conditions relevant for in situ carbon storage. Chem. Geol. 373, 93–105 (2014).

19. 19.

Gadikota, G., Matter, J., Kelemen, P. & Park, A. H. Chemical and morphological changes during olivine carbonation for CO2 storage in the presence of NaCl and NaHCO3. Phys. Chem. Chem. Phys. 16, 4679–4693 (2014).

20. 20.

Turvey, C. C. et al. Hydrotalcites and hydrated Mg-carbonates as carbon sinks in serpentinite mineral wastes from the Woodsreef chrysotile mine, New South Wales, Australia: controls on carbonate mineralogy and efficiency of CO2 air capture in mine tailings. Int. J. Greenh. Gas Control 79, 38–60 (2018).

21. 21.

Rogers, K. L., Neuhoff, P. S., Pedersen, A. K. & Bird, D. K. CO2 metasomatism in a basalt-hosted petroleum reservoir, Nuussuaq, West Greenland. Lithos 92, 55–82 (2006).

22. 22.

Gysi, A. P. & Stefánsson, A. CO2-water–basalt interaction. Low temperature experiments and implications for CO2 sequestration into basalts. Geochim. Cosmochim. Acta 81, 129–152 (2012).

23. 23.

McGrail, B. P. et al. Field validation of supercritical CO2 reactivity with basalts. Environ. Sci. Technol. Lett. 4, 6–10 (2017). Reports the results from the rapid carbonation during the Wallula pilot injection.

24. 24.

Snæbjörnsdóttir, S. Ó. et al. CO2 storage potential of basaltic rocks in Iceland and the oceanic ridges. Energy Procedia 63, 4585–4600 (2014).

25. 25.

Gysi, A. P. & Stefánsson, A. CO2–water–basalt interaction. Numerical simulation of low temperature CO2 sequestration into basalts. Geochim. Cosmochim. Acta 75, 4728–4751 (2011).

26. 26.

Voigt, M., Pearce, C. R., Baldermann, A. & Oelkers, E. H. Stable and radiogenic strontium isotope fractionation during hydrothermal seawater-basalt interaction. Geochim. Cosmochim. Acta 240, 131–151 (2018).

27. 27.

Snæbjörnsdóttir, S. Ó., Gislason, S. R., Galeczka, I. M. & Oelkers, E. H. Reaction path modelling of in-situ mineralisation of CO2 at the CarbFix site at Hellisheidi, SW-Iceland. Geochim. Cosmochim. Acta 220, 348–366 (2018).

28. 28.

Godard, M., Luquot, L., Andreani, M. & Gouze, P. Incipient hydration of mantle lithosphere at ridges: a reactive-percolation experiment. Earth Planet. Sci. Lett. 371, 92–102 (2013).

29. 29.

Kelemen, P. B. & Hirth, G. Reaction-driven cracking during retrograde metamorphism: Olivine hydration and carbonation. Earth Planet. Sci. Lett. 345, 81–89 (2012).

30. 30.

Zhu, W. L. et al. Experimental evidence of reaction-induced fracturing during olivine carbonation. Geophys. Res. Lett. 43, 9535–9543 (2016).

31. 31.

Clark, D. E. et al. The chemistry and potential reactivity of the CO2-H2S charged injected waters at the basaltic CarbFix2 site, Iceland. Energy Procedia 146, 121–128 (2018).

32. 32.

Wolff-Boenisch, D., Gislason, S. R. & Oelkers, E. H. The effect of crystallinity on dissolution rates and CO2 consumption capacity of silicates. Geochim. Cosmochim. Acta 70, 858–870 (2006).

33. 33.

Dessert, C., Dupre, B., Gaillardet, J., Francois, L. M. & Allegre, C. J. Basalt weathering laws and the impact of basalt weathering on the global carbon cycle. Chem. Geol. 202, 257–273 (2003). Assesses the natural weathering potential of basalts worldwide and their contribution to the carbon cycle.

34. 34.

Kappel, E. S. & Ryan, W. B. F. Volcanic episodicity and a non-steady state rift valley along northeast Pacific spreading centers: evidence from Sea MARC I. J. Geophys. Res. 91, 13925–13940 (1986).

35. 35.

Karson, J. A. Geologic structure of the uppermost oceanic crust created at fast-to intermediate-rate spreading centers. Annu. Rev. Earth Planet. Sci. 30, 347–384 (2002).

36. 36.

Alt, J. C. & Teagle, D. A. H. The uptake of carbon during alteration of ocean crust. Geochim. Cosmochim. Acta 63, 1527–1535 (1999).

37. 37.

Coogan, L. A., Parrish, R. & Roberts, N. M. W. Early hydrothermal carbon uptake by the upper oceanic crust: insight from in-situ U-Pb dating. Geology 44, 147–150 (2016).

38. 38.

Wiese, F., Fridriksson, T. & Ármannsson, H. CO 2 Fixation by calcite in high-temperature geothermal systems in Iceland (ÍSOR, 2008).

39. 39.

Marieni, C., Henstock, T. J. & Teagle, D. A. H. Geological storage of CO2 within the oceanic crust by gravitational trapping. Geophys. Res. Lett. 40, 6219–6224 (2013).

40. 40.

Callow, B., Falcon-Suarez, I., Ahmed, S. & Matter, J. M. Assessing the carbon sequestration potential of basalt using X-ray micro-CT and rock mechanics. Int. J. Greenh. Gas Control 70, 146–156 (2018).

41. 41.

McGrail, B. P. et al. Potential for carbon dioxide sequestration in flood basalts. J. Geophys. Res. 111, B12201 (2006).

42. 42.

Kelemen, P. B. & Matter, J. M. In situ carbonation of peridotite for CO2 storage. Proc. Natl Acad. Sci. USA 105, 17295–17300 (2008). Reviews the carbonation of peridotites and their potential for CO 2 storage.

43. 43.

Hansen, L. D., Dipple, G. M., Gordon, T. M. & Kellett, D. A. Carbonated serpentinite (listwanite) at Atlin, British Columbia: a geological analogue to carbon dioxide sequestration. Can. Mineral. 43, 225–239 (2005).

44. 44.

Wilson, S. A. et al. Carbon dioxide fixation within mine wastes of ultramafic-hosted ore deposits: examples from the Clinton Creek and Cassiar chrysotile deposits, Canada. Economic Geol. 104, 95–112 (2009).

45. 45.

Wilson, S. A. et al. Offsetting of CO2 emissions by air capture in mine tailings at the Mount Keith Nickel Mine, Western Australia: Rates, controls and prospects for carbon neutral mining. Int. J. Greenh. Gas Control 25, 121–140 (2014).

46. 46.

Oskierski, H. C., Dlugogorski, B. Z. & Jacobsen, G. Sequestration of atmospheric CO2 in chrysotile mine tailings of the Woodsreef Asbestos Mine, Australia: quantitative mineralogy, isotopic fingerprinting and carbonation rates. Chem. Geol. 358, 156–169 (2013).

47. 47.

Power, I. M. et al. in Geochemistry of Geologic CO 2 Sequestration Vol. 77 Ch. 9 (eds DePaolo, D. J., Cole, D. R., Navrotsky, A. & Bourg, I. C.) 305–360 (Mineralogical Society of America, 2013). Reviews all aspects of carbon mineralization.

48. 48.

Marini, L. Geological Sequestration of Carbon Dioxide: Thermodynamics, Kinetics, and Reaction Path Modeling (Elsevier, 2007). A compilation of the theoretical background of CO 2 storage.

49. 49.

Blum, A. E. & Lasaga, A. C. Role or surface speciation in the low-temperature dissolution of minerals. Nature 331, 431–433 (1988).

50. 50.

Oelkers, E. H. An experimental study of forsterite dissolution rates as a function of temperature and aqueous Mg and Si concentrations. Chem. Geol. 175, 485–494 (2001).

51. 51.

Pokrovsky, O. S. & Schott, J. Kinetics and mechanism of forsterite dissolution at 25°C and pH from 1 to 12. Geochim. Cosmochim. Acta 64, 3313–3325 (2000).

52. 52.

Rosso, J. J. & Rimstidt, J. D. A high resolution study of forsterite dissolution rates. Geochim. Cosmochim. Acta 64, 797–811 (2000).

53. 53.

Golubev, S. V., Pokovsky, O. S. & Schott, J. Experimental determination of the effect of dissolved CO2 on the dissolution kinetics of Mg and Ca silicates at 25 °C. Chem. Geol. 217, 227–238 (2005).

54. 54.

Olsen, A. A. & Rimstidt, J. D. Oxalate-promoted forsterite dissolution at low pH. Geochim. Cosmochim. Acta 72, 1758–1766 (2008).

55. 55.

Brantley, S. L. & Chen, Y. in Chemical Weathering Rates of Silicate Minerals Vol. 85 (eds White, A. F. & Brantley S. L.) 1767–1783 (Mineralogical Society of America, 1995).

56. 56.

Carroll, S. A. & Knauss, K. G. Dependence of labradorite dissolution kinetics on CO2(aq), Al(aq), and temperature. Chem. Geol. 217, 213–225 (2005).

57. 57.

Metz, V., Amram, K. & Ganor, J. Stoichiometry of smectite dissolution reaction. Geochim. Cosmochim. Acta 69, 1755–1772 (2005).

58. 58.

Oelkers, E. H. & Cole, D. R. Carbon dioxide sequestration: a solution to a global problem. Elements 4, 305–310 (2008).

59. 59.

Kampman, N., Bickle, M., Becker, J., Assayag, N. & Chapman, H. Feldspar dissolution kinetics and Gibbs free energy dependence in a CO2-enriched groundwater system, Green River, Utah. Earth Planet. Sci. Lett. 284, 473–488 (2009).

60. 60.

Pham, V. T. H., Lu, P., Aagaard, P., Zhu, C. & Hellevang, H. On the potential of CO2–water–rock interactions for CO2 storage using a modified kinetic model. Int. J. Greenh. Gas Control 5, 1002–1015 (2010).

61. 61.

Munz, I. A. et al. Mechanisms and rates of plagioclase carbonation reactions. Geochim. Cosmochim. Acta 77, 27–51 (2012).

62. 62.

Hellevang, H., Pham, V. T. H. & Aagaard, P. Kinetic modelling of CO2–water–rock interactions. Int. J. Greenh. Gas Control 15, 3–15 (2013).

63. 63.

Gudbrandsson, S., Wolff-Boenisch, D., Gislason, S. R. & Oelkers, E. H. Experimental determination of plagioclase dissolution rates as a function of its composition and pH at 22°C. Geochim. Cosmochim. Acta 139, 154–172 (2014).

64. 64.

Gislason, S. R. & Oelkers, E. H. Mechanism, rates, and consequences of basaltic glass dissolution: II. An experimental study of the dissolution rates of basaltic glass as a function of pH and temperature. Geochim. Cosmochim. Acta 67, 3817–3832 (2003).

65. 65.

Oelkers, E. H. & Gislason, S. R. The mechanism, rates and consequences of basaltic glass dissolution: I. An experimental study of the dissolution rates of basaltic glass as a function of aqueous Al, Si and oxalic acid concentration at 25°C and pH = 3 and 11. Geochim. Cosmochim. Acta 65, 3671–3681 (2001).

66. 66.

Wolff-Boenisch, D., Gislason, S. R., Oelkers, E. H. & Putnis, C. V. The dissolution rates of natural glasses as a function of their composition at pH 4 and 10.6, and temperatures from 25 to 74°C. Geochim. Cosmochim. Acta 68, 4843–4858 (2004).

67. 67.

Rimstidt, J. D., Brantley, S. L. & Olsen, A. A. Systematic review of forsterite dissolution rate data. Geochim. Cosmochim. Acta 99, 159–178 (2012).

68. 68.

Giammar, D. E., Bruant, R. G. & Peters, C. A. Forsterite dissolution and magnesite precipitation at conditions relevant for deep saline aquifer storage and sequestration of carbon dioxide. Chem. Geol. 217, 257–276 (2005).

69. 69.

Power, I. M., Kenward, P. A., Dipple, G. M. & Raudsepp, M. Room temperature magnesite precipitation. Cryst. Growth Des. 17, 5652–5659 (2017).

70. 70.

Hänchen, M., Prigobbe, V., Baciocchi, R. & Mazzotti, M. Precipitation in the Mg-carbonate system — effects of temperature and CO2 pressure. Chem. Eng. Sci. 63, 1012–1028 (2008).

71. 71.

Qafoku, O. et al. Formation of submicron magnesite during reaction of natural forsterite in H2O-saturated supercritical CO2. Geochim. Cosmochim. Acta 134, 197–209 (2014).

72. 72.

Maroto-Valer, M. M., Fauth, D. J., Kuchta, M. E., Zhang, Y. & Andrésen, J. M. Activation of magnesium rich minerals as carbonation feedstock materials for CO2 sequestration. Fuel Process. Technol. 86, 1627–1645 (2005).

73. 73.

Gislason, S. R., Veblen, D. R. & Livi, K. J. T. Experimental meteoric water-basalt interactions: characterization and interpretation of alteration products. Geochim. Cosmochim. Acta. 57, 1459–1471 (1993).

74. 74.

Xiong, W. et al. CO2 mineral trapping in fractured basalt. Int. J. Greenh. Gas Control 66, 204–217 (2017).

75. 75.

Harrison, A. L., Tutolo, B. M. & DePaolo, D. J. The role of reactive transport modelling in geologic carbon storage. Elements 15, 93–98 (2019).

76. 76.

Aradóttir, E. S. P., Sonnenthal, E. L., Björnsson, G. & Jónsson, H. Multidimensional reactive transport modeling of CO2 mineral sequestration in basalts at the Hellisheidi geothermal field, Iceland. Int. J. Greenh. Gas Control 9, 24–40 (2012).

77. 77.

Snæbjörnsdóttir, S. Ó. et al. The chemistry and saturation states of subsurface fluids during the in situ mineralisation of CO2 and H2S at the CarbFix site in SW-Iceland. Int. J. Greenh. Gas Control 58, 87–102 (2017).

78. 78.

Pogge von Strandmann, P. A. E. et al. Rapid CO2 mineralisation into calcite at the CarbFix storage site quantified using calcium isotopes. Nat. Commun. 10, 1983 (2019).

79. 79.

Trias, R. et al. High reactivity of deep biota under anthropogenic CO2 injection into basalt. Nat. Commun. 8, 1063 (2017).

80. 80.

Gunnarsson, I. et al. The rapid and cost-effective capture and subsurface mineral storage of carbon and sulfur at the CarbFix2 site. Int. J. Greenh. Gas Control 79, 117–126 (2018).

81. 81.

Sigfússon, B. et al. Reducing emissions of carbon dioxide and hydrogen sulphide at Hellisheidi power plant in 2014–2017 and the role of CarbFix in achieving the 2040 Iceland climate goals. Energy Procedia 146, 135–145 (2018).

82. 82.

McGrail, B. P., Spane, F. A., Amonette, J. E., Thompson, C. R. & Brown, C. F. Injection and monitoring at the Wallula basalt pilot project. Energy Procedia 63, 2939–2948 (2014).

83. 83.

McGrail, B. P. et al. Wallula Basalt Pilot demonstration project: post-injection results and conclusions. Energy Procedia 114, 5783–5790 (2017).

84. 84.

The National Academies of Sciences, Engineering, and Medicine. Negative emissions technologies and reliable sequestration: a research agenda (National Academies Press, 2018). A comprehensive review on carbon mineralization prepared by the US National Academy of Sciences.

85. 85.

Zakharova, N. V., Goldberg, D. S., Sullivan, E. C., Herron, M. M. & Grau, J. A. Petrophysical and geochemical properties of Columbia River flood basalt: implications for carbon sequestration. Geochem. Geophys. Geosyst. 13, Q11001 (2012).

86. 86.

Weingarten, M., Ge, S., Godt, J. W., Bekins, B. A. & Rubinstein, J. L. High-rate injection is associated with the increase in US mid-continent seismicity. Science 348, 1336–1340 (2015).

87. 87.

Thorsteinsson, H. & Gunnarsson, G. Induced seismicity – stakeholder engagement in Iceland. GRC Trans. 38, 879–882 (2014).

88. 88.

Juncu, D. et al. Injection-induced surface deformation and seismicity at the Hellisheidi geothermal field, Iceland. J. Volcanol. Geotherm. Res. https://doi.org/10.1016/j.jvolgeores.2018.03.019 (2018).

89. 89.

Zoback, M. D. & Gorelick, S. M. Earthquake triggering and large-scale geologic storage of carbon dioxide. Proc. Natl Acad. Sci. USA 109, 10164–10168 (2012).

90. 90.

Aradóttir, E. S. & Hjálmarsson, E. CarbFix – public engagement and transparency. Energy Procedia 146, 115–120 (2018).

91. 91.

Eiken O. in Geophysics and Geosequestration (eds Davis, T. L., Landrø, M. & Wilson, M.) 209–234 (Cambridge Univ. Press, 2019).

92. 92.

Goldberg, D. S., Lackner, K. S., Han, P., Slagle, A. L. & Wang, T. Co-location of air capture, subseafloor CO2 sequestration, and energy production on the Kerguelen plateau. Environ. Sci. Technol. 47, 7521–7529 (2013).

93. 93.

World Health Organization. Guidelines for drinking-water quality, 4th edition (WHO, 2011).

94. 94.

Flaathen, T. K., Gislason, S. R., Oelkers, E. H. & Sveinbjörnsdóttir, Á. E. Chemical evolution of the Mt. Hekla, Iceland, groundwaters: a natural analogue for CO2 sequestration in basaltic rocks. Appl. Geochem. 24, 463–474 (2009).

95. 95.

Olsson, J., Stipp, S. L. S., Makovicky, E. & Gislason, S. R. Metal scavenging by calcium carbonate at the Eyjafjallajökull volcano: a carbon capture and storage analogue. Chem. Geol. 384, 135–148 (2014).

96. 96.

Goldberg, D. & Lackner, K. Creating negative emissions at remote CO2 sequestration sites. Greenh. Gas Sci. Technol. 5, 238–240 (2015).

97. 97.

Keith, D. W., Holmes, G., St. Angelo, D. & Heidel, K. A process for capturing CO2 from the atmosphere. Joule 2, 1573–1594 (2018).

98. 98.

Gutknecht, V., Snæbjörnsdóttir, S. Ó., Sigfússon, B., Aradóttir, E. S. & Charles, L. Creating a carbon dioxide removal solution by combining rapid mineralization of CO2 with direct air capture. Energy Procedia 146, 129–134 (2018).

99. 99.

Goldberg, D. et al. Geological storage of CO2 in sub-seafloor basalt: the CarbonSAFE pre-feasibility study offshore Washington State and British Columbia. Energy Procedia 146, 158–165 (2018).

100. 100.

Wolff-Boenisch, D. On the buffer capacity of CO2-charged seawater used for carbonation and subsequent mineral sequestration. Energy Procedia 4, 3738–3745 (2011).

101. 101.

Luhmann, A. J. et al. Whole rock basalt alteration from CO2-rich brine during flow-through experiments at 150°C and 150 bar. Chem. Geol. 453, 92–110 (2017).

102. 102.

Schaef, H. T., McGrail, B. P. & Owen, A. T. Carbonate mineralisation of volcanic province basalts. Int. J. Greenh. Gas Control 4, 249–261 (2010).

103. 103.

Aradóttir, E. S. P., Sonnenthal, E. L. & Jónsson, H. Development and evaluation of a thermodynamic dataset for phases of interest in CO2 mineral sequestration in basaltic rocks. Chem. Geol. 304–305, 26–38 (2012).

104. 104.

Husebye, J., Brunsvold, A. L., Roussanaly, S. & Zhang, X. Techno economic evaluation of amine based CO2 capture: impact of CO2 concentration and steam supply. Energy Procedia 23, 381–390 (2012).

105. 105.

Webb, R. M. & Gerrard, M. B. Overcoming impediments to offshore carbon dioxide storage: legal issues in the U.S. and Canada. Environ. Law Rep. 49, ELR10634 (2019).

106. 106.

European Commission. Report from the Commission to the European Parliament and the Council. Report on the functioning of the European carbon market (European Commission, 2019).

107. 107.

Folger, P. Carbon capture and sequestration (CCS) in the United States (Congressional Research Service, 2018).

108. 108.

Gaillardet, J., Dupré, B., Louvat, P. & Allégre, C. J. Global silicate weathering and CO2 consumption rates deduced from the chemistry of the large rivers. Chem. Geol. 159, 3–30 (1999).

109. 109.

Sanna, A., Uibu, M., Carmanna, G., Kuusik, R. & Maroto-Valer, M. M. A review of mineral carbonation technologies to sequester CO2. Chem. Soc. Rev. 43, 8049–8080 (2014). Reviews the potential of in situ and ex situ storage through mineral carbonation.

110. 110.

Mervine, E. M. et al. Potential for offsetting diamond mine carbon emissions through mineral carbonation of processed kimberlite: an assessment of De Beers mine sites in South Africa and Canada. Mineral. Petrol. 112, 755–765 (2018).

111. 111.

Harrison, A. L., Power, I. M. & Dipple, G. M. Accelerated carbonation of brucite in mine tailings for carbon sequestration. Environ. Sci. Technol. 47, 126–134 (2012).

112. 112.

Power, I. M. et al. Strategizing carbon-neutral mines: a case for pilot projects. Minerals 4, 399–436 (2014).

113. 113.

Hartmann, J. & Moosdorf, N. The new global lithological map database GLiM: a representation of rock properties at the Earth surface. Geochem. Geophys. Geosyst. 13, Q12004 (2012).

114. 114.

Whittaker, J. M. et al. Long-term interaction between mid-ocean ridges and mantle plumes. Nat. Geosci. 8, 479–483 (2015).

115. 115.

Müller, R. D. et al. Ocean basin evolution and global-scale plate reorganization events since Pangea breakup. Annu. Rev. Earth Planet. Sci. 44, 107–138 (2016).

116. 116.

Johansson, L., Zahirovic, S. & Müller, R. D. The interplay between the eruption and weathering of Large Igneous Provinces and the deep-time carbon cycle. Geophys. Res. Lett. 45, 5380–5389 (2018).

117. 117.

Bales, R. C. & Morgan, J. J. Dissolution kinetics of chrysotile at pH 7 to 10. Geochim. Cosmochim. Acta 49, 2281–2288 (1985).

118. 118.

Gudbrandsson, S., Wolff-Boenisch, D., Gislason, S. R. & Oelkers, E. H. An experimental study of crystalline basalt dissolution from 2 ≤ pH ≤ 11 and temperatures from 5 to 75°C. Geochim. Cosmochim. Acta 75, 5496–5509 (2011).

119. 119.

Critelli, T. et al. Dissolution rates of actinolite and chlorite from a whole-rock experimental study of metabasalt dissolution from 2 ≤ pH ≤ 12 at 25°C. Chem. Geol. 390, 100–108 (2014).

120. 120.

Gebald, C., Repond, N. & Wurzbacher, J. A. Steam assisted vacuum desorption process for carbon dioxide capture. US Patent 2017/0203249 (2015).

121. 121.

Jackson, S. & Brodal, E. A comparison of the energy consumption for CO2 compression process alternatives. IOP Conf. Ser. 167, 012031 (2018).

## Acknowledgements

S.Ó.S., B.S., C.M., S.R.G. and E.H.O. work on the CarbFix project funded by the European Union’s Horizon 2020 research and innovation programme under grant agreement 764760 (CarbFix2), 818169 (GECO) and 764810 (S4CE). D.G. works on the Solid Carbon project funded by the Pacific Institute for Climate Solutions. The authors thank E.S. Aradóttir Pind, R.B. Bragadóttir, K. Helgason, M. Voigt and R. Þrastarson for discussions and assistance in developing Fig. 2.

## Author information

The authors contributed equally to all aspects of the article.

Correspondence to Sandra Ó. Snæbjörnsdóttir.

## Ethics declarations

### Competing interests

The authors declare no competing interests

### Peer review information

Nature Reviews Earth & Environment thanks G. Gadikota and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

### Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## Rights and permissions

Reprints and Permissions