Storing CO₂ in buried volcanoes

Keywords: CCS, CO₂, reservoirs, basalts, buried volcanoes.

Australia contains abundant natural gas resources and is a leading global exporter of liquified natural gas (LNG) (Geoscience Australia 2019). Many of Australia’s producing and undeveloped natural gas fields contain relatively high CO₂ contents, with particularly high values (>20%) reported from wells and fields in the Otway, Cooper-Eromanga, Gippsland, Bonaparte, and Carnarvon basins (Boreham et al. 2001). For some producing fields, most notably Gorgon (~7–27% CO₂: Maftei et al. 2013), CO₂ is separated and injected into deep saline aquifers (Michael et al. 2010). In many fields, however, reservoir CO₂ is not captured but directly released to the atmosphere, contributing significantly to national greenhouse gas emissions. The Ichthys Field in the Browse Basin contains high CO₂ contents in the Brewster (8%) and Plover (17%) reservoirs, and it is anticipated that ~96 Mt of reservoir CO₂ will be emitted over the 40-year lifetime of the Ichthys LNG project (INPEX 2010).

The most mature technology for the long-term carbon capture and storage (CCS) is underground sequestration of supercritical CO₂ in sedimentary rock formations such as deep saline aquifers or depleted hydrocarbon reservoirs deep enough (typically 1 km or deeper) for the CO₂ to remain in a supercritical state (Bunch et al. 2014; National Academies of Sciences Engineering and Medicine 2019). The critical requirements for CCS in sedimentary formations are (i) a thick reservoir (usually sandstone or carbonate) with sufficient porosity and permeability to contain large volumes of CO₂ at commercially meaningful injection rates, (ii) an overlying thick seal (usually shale) with sufficiently high capillary entry pressure and low permeability to retain the injected CO₂ over geological timescales, and (iii) a gas-trapping structure that, due to gas buoyancy, concentrates large volumes of CO₂ and prevent its migration and leakage (Dempsey et al. 2015; National Academies of Sciences Engineering and Medicine 2019). Additional important geological considerations include the absence of permeable faults and fractures through the seal and favourable stresses and pore pressures to avoid reservoir or seal fracturing during injection (Nicol et al. 2017; National Academies of Sciences Engineering and Medicine 2019). The combined underground storage capacity in saline aquifers and depleted hydrocarbon reservoirs is estimated between 5000 and 25 000 Gt CO₂ (Kelemen et al. 2019).

An alternative option for the secure, long-term underground storage of CO₂ is carbon mineralisation in mafic and ultramafic igneous rocks, which could potentially host up to 60 000 000 Gt CO₂ (Kelemen et al. 2019). Carbon mineralisation involves the formation of stable carbonate minerals (e.g. calcite, magnesite, dolomite) through the reaction of CO₂ (gas, liquid, dissolved in water or supercritical) with rocks that are rich in calcium or magnesium (Wolff-Boenisch and Galeczka 2018). Mg-rich, Ca-bearing rocks include ultramafic peridotites and mafic basalts, which contain highly reactive minerals including olivine and pyroxene (National Academies of Sciences Engineering and Medicine 2019). Whilst in situ carbon mineralisation in sandstone reservoirs can occur over timescales of thousands of years, peridotites and basalts may mineralise and sequester 90% of the injected CO₂ in a few months to decades (Fig. 1a; National Academies of Sciences Engineering and Medicine 2019).

Fig. 1.  (a) Comparison of CO₂-trapping mechanisms over time when injecting pure supercritical CO₂ into sedimentary basins (left) and water-dissolved CO₂ for mineralisation (right). Modified after Snæbjörnsdóttir et al. (2020). (b) Reactions rates of in situ mineralisation of CO₂ in peridotites, basalts and sedimentary rocks. Modified after Kelemen et al. (2019).

The rate of carbon mineralisation is influenced by factors including the available CO₂ dissolved in solution, temperature and variations in pH, with low pH promoting mineral dissolution and high pH accelerating carbonate precipitation (Kelemen et al. 2019). Olivine (Mg₂SiO₄), which is the major mineral in peridotite, has amongst the highest reaction rates with CO₂-bearing aqueous fluids, whilst plagioclase feldspar (the major constituent of basalts) has somewhat lower reaction rates, meaning that CO₂ mineralisation may be expected to occur more rapidly in peridotites than basalts (Fig. 1b; Kelemen et al. 2019; Seyyedi et al. 2020). However, an exception may occur where basalts contain amorphous glass horizons, which are thought to provide exceptionally good reactants for carbon mineralisation (National Academies of Sciences Engineering and Medicine 2019). Glass forms when basalt cools very quickly and is often abundant in submarine lavas and hyaloclastite breccias (National Academies of Sciences Engineering and Medicine 2019).

Basaltic rocks are extremely abundant; most of the ocean floor (comprising ~70% of Earth’s surface) and >5% of the continents is basaltic, with extensive flood basalt fields present in central India, Siberia, the United States, and Canada (Snæbjörnsdóttir et al. 2020). Counter to many petroleum geologists and engineers’ assumptions, volcanic rocks often have high porosity and permeability (Fig. 2), particularly when fractured (Massiot et al. 2017; Kelemen et al. 2019; Bischoff et al. 2020; Millett et al. 2020). Two recent pilot projects have demonstrated the potential for CO₂ storage through carbonate mineralisation in basalt reservoirs. In the Wallula project in Washington State, 977 tons of pure CO₂ was injected into high-permeability basalt flow tops at depths between 882 and 887 m over 3 weeks in August 2013 (McGrail et al. 2017). Whilst it is not yet known what proportion of injected CO₂ has formed carbonate minerals and how much remains within pore fluids, drill core from the injection zone indicates high rates of basalt dissolution and CO₂ mineralisation, consistent with laboratory experiments (National Academies of Sciences Engineering and Medicine 2019).

Fig. 2.  (a) Hand sample and (b) plain-polarised thin section of a trachybasalt crystal-rich lapilli tuff from Banks Peninsula volcanic complex, New Zealand. The rock mineral assemblage comprises olivine (ol) and feldspar (felds) immersed in a glassy matrix altered to palagonite (pal), which is a typical product of devitrification of basaltic magmas. The primary porosity (47%) and permeability (800 mD) are controlled by an intense degree of material fragmentation that forms intergranular pores (int) associated with high vesicles content (ves) and quenching fractures (frac). Secondary processes of mineral alteration creates dissolution pathways (dis), interconnecting primary pores.

In addition, during phase I of the CarbFix project in Iceland, ~200 tCO₂ (co-produced with SO₂ in geothermal fluids) was injected into highly permeable, fractured basalts with ~10% porosity at a depth of 500 m (ambient temperature ~20–50°C). Because CO₂ is not supercritical at this depth, CO₂ and H₂O were separately injected with proportions adjusted to ensure complete solubility of CO₂ into aqueous fluid at the target depth (National Academies of Sciences Engineering and Medicine 2019). Quantification of mineral carbonisation using reactive and non-reactive tracers and isotopes has demonstrated rapid mineralisation of the injected CO₂, with >95% of the injected gas mineralised within 2 years (Snæbjörnsdóttir et al. 2020). Following the success of CarbFix phase I, the project has considerably upscaled, injecting 10–20 ktCO₂/year into basaltic reservoirs at a depth of ~1500 m (ambient temperature ~250°C), with tracer results continuing to indicate nearly complete loss of carbon along a ~2000 m flow path (National Academies of Sciences Engineering and Medicine 2019).

Despite the success of these projects, a number of uncertainties remain, including determining the precise mineral assemblage (including alteration phases) that are reacting in the reservoirs, the passivation of reactive mineral surfaces over time, the potential clogging of pore space, and the spatial distribution of carbonate mineral precipitation versus dissolution (National Academies of Sciences Engineering and Medicine 2019). Today, the cost of storing CO₂ in basalts is estimated to be US$20–30/tCO₂ as compared to <US$20 for storage in sedimentary reservoirs (Kelemen et al. 2019). However, if the technology can be further proven, storage in basaltic reservoirs may become a preferred choice in volcanic provinces (Kelemen et al. 2019).

Buried basalt lava flows and volcanoes are commonly found at depths amenable to dissolved and/or supercritical CO₂ injection (i.e. >1 km) in many Australian sedimentary basins. These include the Browse and Carnarvon basins along the North West Shelf (Holford et al. 2013), and the Bight, Otway, Bass, and Gippsland basins along the southern Australian margin (Holford et al. 2012; Meeuws et al. 2016; Reynolds et al. 2017, 2018; Watson et al. 2019) and in the Eromanga Basin in central Australia (Hardman et al. 2019). In some of these basins, buried basaltic sequences are located in close proximity to gas-bearing reservoirs with high CO₂ content (Holford et al. 2012).

Well-preserved basaltic submarine volcanoes are common within the Miocene Torquay Group of the Bass Basin (Holford et al. 2017; Reynolds et al. 2018), with a large (~5 km diameter) volcano overlying the Yolla Gas Field, where reservoir fluids contain ~17–20% CO₂ (Holford et al. 2012; Watson et al. 2019). Core from the volcanic section penetrated by the nearby Bass-1 well, which penetrated the flank of a Miocene volcano, contains clast-supported conglomerate of dark grey microvesicular basalt interpreted to represent reworked hyaloclastite and volcaniclastic material, whilst log data and thin sections from the Tasmania Devil-1 well, which penetrated the crest of a volcano, indicate the presence of fine grained crystalline basaltic material interpreted as pillow lavas (Watson et al. 2019).

Volcanic rocks have been intersected by many wells in the hangingwall of the basin-bounding Rosedale Fault System in the northern Gippsland Basin, where numerous gas accumulations with high-CO₂ content (some exceeding 30%) occur (O’Brien et al. 2008; Holford et al. 2012). The majority of these volcanic rocks are Campanian-age basalts that have been subject to various degrees of alteration (Holford et al. 2012). In some cases (and in spite of the alteration), these basalts provide the top seals to underlying gas accumulations, most notably at the Kipper field where a 328 m gross gas column (~10% CO₂) within Golden Beach Group sandstones is sealed by ~98 m of volcanic rocks, interpreted to be basaltic lava flows (Sloan et al. 1992; O’Halloran and Johnstone 2001). Volcanic rocks also act as seals at Remora-1 (oil) and Tuna-4 (oil and gas) (Holford et al. 2012). Petrological and petrophysical studies of these rocks have been limited, though cuttings and sample descriptions indicate that they comprise fine-grained basaltic lavas and tuffs (McPhail 2000). However, our analysis of borehole image logs acquired through volcanic sequences in the Manta-2A, Basker-2, and Basker-5 wells demonstrates an abundance of conductive fractures (Fig. 3), raising the possibility of significant fracture-related permeability.

Fig. 3.  Image log-based observations of natural fractures in the (a) Basker-5, (b) Manta-2A, and (c) Basker-2 wells presented along with detailed lithologies determined from drill cuttings and petrophysical logs, indicating substantial fracturing within volcanic sequences. Natural fractures are plotted as tadpoles (conductive fracture – blue triangles and electrically resistive fractures – red circles) where marker location present the amount of dip on fracture plane, with the tick corresponding to the dip direction of the plane, followed by bias corrected fracture intensity, and dynamic UCS log. Rose diagrams are plotted natural fractures orientation on lower hemisphere, equal-area projection of poles to natural fractures (both conductive – blue colour and resistive – red).

Hardman et al. (2019) have recently described a province of Jurassic extrusive and intrusive rocks within the Eromanga Basin, overlying the Nappamerri Trough of the Cooper Basin, where the highest CO₂ contents (up to 40%) in gas samples are found (Boreham et al. 2001). 3D seismic interpretation coupled with log analyses indicates the presence of multiple mafic monogenetic volcanoes that extend into tabular basalt lava flows (Hardman et al. 2019). The Lambda-1 well intersected 283 m of basalt directly underlying the Lower Jurassic Birkhead Formation, with the upper 33 m comprising heavily weathered, fractured, and vesicular facies, and the remaining 250 m comprising fresh and crystalline rock (Hardman et al. 2019). In other parts of the province, basalts have been extensively altered; Kappa-1 intersected a 120 m succession of mostly fine-grained basalt with evidence for extensive chlorite replacement of ferromagnesian minerals (Hardman et al. 2019).

Given the promising results from both experimental studies and pilot projects, and the close proximity of basaltic sequences to many high-CO₂ content gas fields in onshore and offshore Australia, we propose that the possibility of storing CO₂ through in situ carbon mineralisation in buried volcanic rocks in sedimentary basins merits further consideration. In addition to the aforementioned uncertainties related to CO₂–rock–fluid interactions, further challenges specific to sedimentary basins include defining the first-order stratigraphic and permeability architecture of buried volcanic sequences in order to identify potential reservoir facies (e.g. Bischoff et al. 2021), quantifying the hydraulic and petrophysical properties of buried basaltic volcanoes and lava flows, and defining their petrological characteristics and diagenetic histories.

There are no conflicts of interest associated with this research.

AB thanks funding from the MBIE research grant UOCX1707.