A Parameterized Lifecycle Assessment of Ex Situ CO₂ Mineralization, Using Olivine as a Case Study

Abstract

To achieve global climate goals, novel technologies are needed to efficiently capture and durably store atmospheric CO₂. Reacting CO₂ with silicate minerals in engineered systems, often called ex situ mineralization, is one such approach. The potential net climate benefit of ex situ CO₂ mineralization is unclear, however, as it depends on the technology pathway, location, energy and resource demands, and potential for avoided emissions via co-product utilization. Accurately quantifying these factors in site-specific scenarios is critical for assessing a project’s carbon footprint. Here, we present a parameterized lifecycle assessment (LCA) model to quantify the net CO₂e emissions of a range of ex situ mineralization scenarios, considering both direct and avoided emissions. Using a case study with olivine feedstock in Washington State, U.S.A., we show that ex situ mineralization can range from net CO₂ positive for the least optimal scenarios to highly CO₂ negative for the most optimal scenarios. Considering avoided emissions, some scenarios can be >100% efficient (>1 t of net CO₂ stored + avoided per t of gross CO₂ stored). Capture efficiency depends heavily on low-carbon electricity, short transport distances, technologies that facilitate the passive capture of ambient CO₂, and maximizing avoided emissions. Substantial avoided emissions can be achieved via efficient co-product use in the local cement industry, although the size of the industry may be a constraint to scaling. Modest avoided emissions are also possible via integrated metal recovery. To facilitate comparability with future analyses and to help guide project development, our model is technology and mineral agnostic and available for download as a supplement.

1. Introduction

Limiting global warming to well below 2 °C, as outlined by the Paris Agreement, requires CO₂ emissions to reach net zero by mid-century. Achieving net zero entails decarbonization of systems reliant on fossil fuels, as well as atmospheric CO₂ removal (CDR) to compensate for “hard to abate” emissions. CDR may also be needed to compensate for temperature overshoot, in which global CO₂ emissions exceed the remaining emissions budget. , Given current trajectories, CDR is anticipated to be needed on the order of 100–1000 cumulative Gt of CO₂ by 2100.

CO₂ mineralization is a CDR technology that reacts CO₂ with alkaline minerals to form stable carbonates. Mineralization can be achieved by injecting CO₂ into underground mafic rock (in situ) or by reacting CO₂ with mined or waste feedstock (ex situ). Ex situ mineralization has been investigated as a sequestration technology for point-source CO₂ capture and storage (CCS) for decades, , and more recent research has considered applications with direct air capture (DAC). ,

A variety of feedstocks can be employed for ex situ CO₂ mineralization. Both industrial wastes and mine tailings are potentially available at scales facilitating the mineralization of hundreds of Mt of CO₂ year^–1, however their composition and supply may be unreliable over time. − Deploying mineralization at larger scales may require mining fresh alkaline feedstock. Olivine [(Mg,Fe)₂SiO₄] is attractive for CO₂ mineralization given high Mg content and fast reaction kinetics. Mineralization with olivine can proceed either directly (reaction ), or indirectly, where it is first dissolved then reacted with CO₂ in subsequent steps. With indirect pathways, olivine can be converted to a stable intermediate mineral, brucite [Mg­(OH)₂] (reaction ), which subsequently reacts with CO₂ to form magnesite (reaction ). Brucite can also react with aqueous CO₂ to form dissolved bicarbonate salt [Mg­(HCO₃)₂] as an ocean alkalinity enhancement (OAE) technology (reaction ) , (this is not typically referred to as mineralization and is rather considered a form of weathering; for convenience, we refer to all pathways as “mineralization”). Brucite can react spontaneously with ambient CO₂, so the two latter pathways avoid the need for CO₂ capture technologies, potentially offering temporal and spatial flexibility.

$${\mathrm{Mg}}_2{\mathrm{SiO}}_4+2{\mathrm{CO}}_2\to 2{\mathrm{MgCO}}_3+{\mathrm{SiO}}_2$$ 1

$${\mathrm{Mg}}_2{\mathrm{SiO}}_4+2{\mathrm{H}}_2\mathrm{O}\to 2\mathrm{Mg}{(\mathrm{OH})}_2+{\mathrm{SiO}}_2$$ 2

$$\mathrm{Mg}{(\mathrm{OH})}_2+{\mathrm{CO}}_2\to {\mathrm{MgCO}}_3+{\mathrm{H}}_2\mathrm{O}$$ 3

$$\mathrm{Mg}{(\mathrm{OH})}_2+2{\mathrm{CO}}_2\to {\mathrm{Mg}}^{2+}+2{{\mathrm{HCO}}_3}^{-}$$ 4

Mineralization poses both opportunities and challenges compared to other CDR technologies. Key opportunities include effectively permanent CO₂ sequestration, high availability of geologic feedstock, and large-scale deployment potential in industries like mining and cement. Additionally, mineralization generates valuable co-products (MgCO₃ and SiO₂) that can be used as supplementary cementitious materials (SCMs) to displace cement, as well as critical metals necessary for decarbonization technolgies. − A major challenge of CO₂ mineralization, however, is slow reaction kinetics: although production of MgCO₃ and Mg­(HCO₃)₂ are thermodynamically favored at earth surface conditions, energy-intensive processes are required to increase reaction rates. These include feedstock pretreatment (e.g., grinding), capturing and concentrating CO₂, and/or optimizing reaction temperature, pressure, and pH. Another challenge is the necessary transport of materials between the mine, mineralization site, and downstream industries. This complex interplay between energy demands and avoided emissions from co-product use necessitates detailed life cycle assessments (LCAs) to understand the net climate benefit.

Early mineralization LCAs showed promising results, however unfavorable economics for integration with CCS slowed further progress. , An important metric in this context is CO₂ capture efficiency: the net CO₂ stored considering lifecycle emissions, divided by gross CO₂ stored. More recent LCAs have shown a wide range of efficiencies, from 90% efficient to net CO₂ positive (where lifecycle emissions exceed gross sequestration), depending on feedstock, mineralization technology, and the complexity of the study. − Given the early stage of mineralization technology, these studies often focus primarily on within-gate processes. Some studies, have also considered the use of co-products, showing that efficient use of MgCO₃ and SiO₂ as SCMs can significantly increase net efficiency. ,, The magnitude of CO₂ avoidance is limited, however, by the capacity of the cement industry to utilize SCMs, illustrating that both upstream energy consumption and downstream (site-specific) constraints affect CO₂ capture efficiency.

Our analysis builds on these studies in several ways. First, we consider site-specific constraints and resources at a single location, including feedstock and CO₂ source, transportation, grid carbon intensity, and capacity of the regional cement industry to absorb SCMs. We also consider a complete range of co-products, including not only Si and Mg end products, but also metals. While recovery of valuable metals is recognized as a possible revenue stream for mineralization projects, the impact on net CO₂ emissions has not yet been considered. Finally, we make our parameterized LCA model fully available, facilitating future comparisons across alternative scenarios and locations (Supplemental File 1). In this study, we demonstrate its utility with a case study of a hypothetical CO₂ mineralization system using olivine. By comparing different CCS and CDR scenarios, our goal is to highlight the parameters that have the greatest impact on lifecycle CO₂ capture efficiency, and how these parameters are affected by practical site limitations.

2. Materials and Methods

2.1. LCA Modeling Framework

We created a parameterized, Excel-based, LCA model aligned with ISO 14040/14044/14067 to assess the cradle-to-grave greenhouse gas (GHG) emissions of a CO₂ mineralization system. CO₂-equivalent GHG emissions (CO₂e) are estimated based on the IPCC AR5 report, with climate carbon feedback. The LCA system boundary includes emissions associated with mineral supply (mining, comminution, and transport), CO₂ capture for DAC and CCS options, the mineralization process, and transportation and disposal of co-products and waste (Figure ). The system boundary also includes avoided emissions resulting from the productive use of metals, MgCO₃, and/or SiO₂ co-products. Standard LCA practice typically embeds avoided emissions in net project impact; however, keeping in line with best practices for CDR, we are careful to distinguish between net project emissions and avoided emissions. The parameterized model allows for modification of all key process variables, and includes optional default emission factors from LCA databases (e.g., Sphera Thinkstep, 2023). The functional unit is one metric ton of gross CO₂ captured via mineralization, which is equal to gross CO₂ stored, given the durability of mineralized CO₂.

1.

Process flow diagrams of CO₂ capture and storage and co-product utilization via mineralization. MgCO₃ route (A): Mineralization technologies can be coupled to point-source capture or DAC to produce MgCO₃, SiO₂, and trace metals. Mg­(OH)₂ route (B): Silicate minerals can be converted into Mg­(OH)₂ and reacted with CO₂ passively, producing either solid MgCO₃ (with atmospheric CO₂) or dissolved Mg­(HCO₃)₂ (with dissolved CO₂), along with SiO₂ and trace metals. Dashed lines indicate LCA boundaries, and colors are consistent for emissions categories in Figure .

In the following sections, we describe the parameterizations and default options available in the model. We apply these options to a site-specific case study, exploring a range of scenarios for both 2022 and 2050, considering current technologies and projected decarbonization and efficiency gains. All scenarios presented maintain a comparable system boundary, functional unit, and assumptions to maximize comparability.

2.2. Case Study Location

The case study explores realistic scenarios for CO₂ mineralization systems in Washington State, U.S.A. (Figure S1). The Twin Sisters olivine deposit has an estimated 200 Gt reserves of unaltered dunite (>90% olivine), although the current activities at the permitted quarry are relatively small. In our study, olivine is mined and transported by truck 74 km to an industrial area near the Cherry Point petroleum refinery, the largest CO₂ point source in the state (2.3 Mt of CO₂ year^–1). Following mineralization, SiO₂ and MgCO₃ co-products are transported by truck to Heidelberg cement terminal (26 km), and any excess can be distributed within the regional cement industry. While we explore a range of scenarios, for ease of comparison in the Results and Discussion, we refer to six scenarios, one for fossil CCS and five CDR parameterizations, with 1 Mt of CO₂ year^–1 capture, as shown in Table .

1. Parameterizations for the Fossil CCS Scenario and Five CDR Scenarios .

	Fossil CCS	CDR
		1	2	3	4	5
	MgCO3:	MgCO3:	MgCO3:	MgCO3:	Mg(OH)2:	Mg(OH)2:
	CCS (fossil),	CCS (biogenic),	DAC,	DAC,	air	dissolved
parameter	direct	direct	direct	indirect
gross CO2 stored	1 Mt of CO2 year–1
CO2 capture method	point-source capture (fossil C)	point-source capture (biogenic C)	DAC	DAC	N/A	N/A
electricity grid	NWPP
feedstock source	Twin Sisters olivine
feedstock particle size	10 μm (P80)
feedstock transport	74 km; truck; 2% wetted; 2% loss; empty return
reaction efficiency	80%
co-product recovery	80%
mineralization technology	direct			indirect	indirect	indirect, CO2/Mg(OH)2 = 1.7
CO2 capture pathway	reaction				reaction	reaction
local cement production	1.5 Mt year–1
SiO2 and MgCO3 use rates in cement	4 and 10% (use case 4)
SiO2 and MgCO3 transport	26 km local; truck; 2% loss
landfill transport	100 km; truck

^aDefault parameters, shown under fossil CCS, are used across all scenarios, excepted as noted.

2.3. Feedstock

Emissions associated with feedstock sourcing include mining, comminution, and transportation (Supplemental Note 1). We assume comminution occurs at the mine site, and feedstock is transported by ship, barge, train, and/or truck, with emissions based on diesel sourcing for the U.S. A wetting factor and mass loss during transportation between the mine and mineralization site can be applied.

Feedstock elemental composition can be specified in the model. For our case study, we use major and trace elements for the Twin Sisters dunite compiled by the U.S. Geologic Survey. For comparison, we also consider a scenario wherein ground olivine is shipped from Aheim, Norway (assuming the same elemental concentrations). The Aheim mine is the largest supplier of olivine globally, and relies entirely on hydropower for comminution energy.

2.4. Electricity Grid

For processes requiring electricity (comminution, CO₂ capture, and mineralization), the model considers emission factors for 2022 subregional grid mixes, or a national average, for the continental U.S. These are based on fuel combustion in the EPA’s eGRID database, and also include upstream manufacturing across the energy mix (Sphera, 2023) and transmission and distribution losses by North American Electric Reliability Corporation (NERC) region. We also provide estimates for 2050 emission factors for each subregional grid, based on the U.S. Annual Energy Outlook projected fuel mix. For the case study, we use the Northwest Power Pool (NWPP) subregion.

2.5. CO₂ Source

The model includes several default options for CO₂ source, including (1) direct utilization of flue gas or passive CO₂ capture with no energy penalty, (2) purified CO₂ from point-source capture, or (3) direct air capture (DAC) (Supplemental Note 2). For scenarios utilizing point-sources, net negative emissions are only possible if the CO₂ is derived from biomass (note that LCA considerations of biomass sourcing and processing are outside the scope of this LCA and would need to be considered separately). Conversely, capture of fossil fuel-derived CO₂ cannot contribute to negative emissions. For point-source capture, we assume absorption with monoethanolamine (MEA), and consider requirements for material and energy use, compression energy, and pumping energy, assuming a short pumping distance. − Associated emissions are a function of the total amount of CO₂ captured and grid carbon intensity. For DAC scenarios, we assume solid adsorbents in a cyclic temperature–vacuum swing adsorption process comparable to the technology operated by Climeworks. Associated lifecycle emissions are calculated as a linear function of grid carbon intensity.

2.6. Mineralization Pathways

For all mineralization pathways, the mass of inputs (feedstock and CO₂) and co-products are calculated based on mineral elemental composition, reaction stoichiometry, and specified reaction efficiencies. For our case study, we assume 80% mineral reaction efficiency, which is relatively conservative. ,, Different CO₂ mineralization technologies vary widely in their carbon intensities. , We estimate emissions of the mineralization technology from two primary studies, as described below and in Supplemental Note 3.

2.6.1. MgCO₃ Route

For the MgCO₃ route (reaction ; Figure A), we considered both direct technologies (which react feedstock with CO₂ directly in high temperature and pressure reactors) and indirect technologies (which first dissolve feedstock in acid, then subsequently react it with CO₂ and base). For the direct technology, we use data from Ostovari et al., based on a continuously stirred tank reactor described in O’Connor et al. , For indirect mineralization, we reference Scott et al., who describe an electrochemical system in which olivine is first dissolved with HCl, followed by precipitation of Mg­(OH)₂ with NaOH, and subsequent reaction with CO₂. While the net reaction for this process is equivalent to reaction , it proceeds via reactions and . For both processes, we apply estimated plant construction emissions from Ostovari et al.

2.6.2. Mg­(OH)₂ Route

For Mg­(OH)₂ routes (reactions –; Figure B), we assume production of Mg­(OH)₂ via the indirect process described above from Scott et al. Proceeding through reaction is equivalent to the MgCO₃ route, with the potential for passive capture of atmospheric CO₂. If used instead to capture dissolved CO₂ via Mg­(HCO₃)₂ (reaction ), CO₂ captured per mol olivine is stoichiometrically doubled relative to the MgCO₃ route. The exact ratio depends on the pH of the receiving system. For our case study, we assume a CO₂/Mg­(OH)₂ ratio of 1.7, implying ultimate storage of sequestered carbon in seawater.

2.7. Co-product Displacement

2.7.1. SiO₂ and MgCO₃

The co-products MgCO₃ and SiO₂ have a range of potential uses in the construction industry (Supplemental Note 4). In our case study, we consider co-product use as SCMs to displace ordinary Portland cement (OPC), at a 1:1 mass replacement ratio, in the local and regional cement industries. We modeled five use cases with a range of SiO₂ and MgCO₃ utilization rates (Table S1). In the most pessimistic use case, neither co-product is utilized. Considering use of SiO₂ only, SiO₂ replaces cement at either 10% (low SiO₂ use case) or 40% (high SiO₂ use case), the upper limit for pozzolan blended cements (ASTM standard C595). We also consider two use cases with stoichiometric ratios of SiO₂ and MgCO₃, which may be advantageous for direct mineralization technologies where co-product separation may be costly: 4/10 and 10/25% addition of SiO₂/MgCO₃ (low and high SiO₂ + MgCO₃ use cases, respectively). For all scenarios, the default use case is 4/10%.

For all use cases, the total amount of SiO₂ and MgCO₃ used is constrained by the size of the local or regional cement industry. In our case study scenarios, we consider production at the Heidelberg Materials Cement Terminal and Grinding Plant in Bellingham, which processes approximately 1.5 Mt of cement year^–1. We also analyze use of excess material in the regional cement industry, estimated from cement grinding capacity in Oregon and Washington, which adds an additional 1.1 Mt of cement year^–1. We assume a transport distance of 300 km from the mineralization site to regional cement facilities, and 2% mass loss during transportation. Any excess MgCO₃ and/or SiO₂ is landfilled.

In our case study, we assume a carbon intensity of 783 kg of CO₂e/t of cement for avoided emissions, based on the environmental product declaration for OPC from the Heidelberg terminal.

2.7.2. Metals

We consider the displacement of metals recovered during the mineralization process. For olivine, these include elements with relatively high concentrations (Fe, Cr, and Ni) and/or high value [Co and platinum group elements (PGEs)]. Technologies for integrated metals recovery during CO₂ mineralization are actively being developed, and a range of metal compounds may be produced. ,, These compounds are typically intermediate metal products, and thus may displace emissions associated with upstream stages of conventional production (e.g., mining, leaching, etc.). In our case study, we assume emissions displacement from production of intermediate metal products for Fe (magnetite), Cr (as chromite ore), and use emission factors for pure metals for Ni, Co, and PGEs (Supplemental Note 5). The model can accommodate inputs for emissions associated with integrated metals recovery; however, these are currently not well quantified, so we assume no additional emissions beyond those from the mineralization process (section ).

2.8. Landfill

Any material lost to process inefficiencies or SiO₂ and/or MgCO₃ exceeding the capacity for cement additives are landfilled. Since all materials are inorganic, no landfill or incineration emissions are expected, and we include only emissions associated with transport to and deposition into local landfills. In our case study, we assume a transportation distance of 100 km.

2.9. Future (2050) Scenarios and Best Case Scenarios

For current scenarios, the model considers state-of-the art technologies for CO₂ capture and mineralization, and the carbon intensity of the electrical grid for 2022 fuel mixes. To assess 2050 scenarios, we use forecasted data for several parameters. For subregional electricity grids, we consider projected fuel mixes with higher contribution from renewable energy, which substantially reduces carbon intensity. For DAC, we consider projected improvements in efficiency and heat pump integration for 2050 (Supplemental Note 2). For point-source capture, we assume a 20% increase in efficiency and reduction in thermal energy needs. For avoided emissions via future SCM utilization, both the size of the cement industry and emission factor for OPC may change. In our case study, we estimate an increase in total SCM demand of approximately 23%, based on WA State population projections. Meanwhile, the carbon intensity of cement is anticipated to decrease by approximately 25% due to decarbonized raw materials, fuel switching, and energy efficiency gains. Given uncertainty on both values, we consider these to effectively cancel each other. For best case scenarios, we also consider electricity sourced from wind energy and the use of battery-electric trucks.

3. Results and Discussion

Depending on selected parameterizations in the case study, mineralization results in net positive CO₂ emissions for the least optimal scenarios, to highly net negative for the most optimal scenarios. For example, net emissions ranged from −489 kg of CO₂e/t of CO₂ stored (i.e., 49% efficient; excluding displacement) for CDR scenario 5, to +528 kg of CO₂e/t of CO₂ stored for scenario 3 (Figure ). Emissions were dominated by CO₂ capture technology (when utilized) and the mineralization process itself, with feedstock comminution a secondary contributor. As a result, Mg­(OH)₂ pathways that bypass CO₂ capture are more efficient. Because these processes are electricity-intensive, efficiency increases significantly when using low-carbon electricity (e.g., the projected 2050 grid mix). Avoided emissions from SCM utilization exceeded −400 kg of CO₂e/t of CO₂ stored for optimistic use cases, but could extend to −2 t of CO₂e/t of CO₂ stored in highly optimized systems (Figure ). Metal co-products were relatively less important, cumulatively displacing −155 kg of CO₂e/t of CO₂ captured. These results are explored in more detail in the following sections.

2.

Net emissions of mineralization system scenarios in (A) 2022 and (B) 2050. Values above net emissions bars exclude displacement, and values below bars are inclusive of displacement. In 2050, indirect mineralization technology and DAC both benefit substantially from grid decarbonization, given dependence on electrical energy (scenarios described in Table ).

3.

Net emissions as a function of CO₂ captured and SCM utilization rate for CDR scenario 2 (DAC: direct). Y-axis values are for 2022 (left) and 2050 (right). Solid lines indicate use in the local cement industry only; dashed lines include regional industry. Net negativity is maximized at smaller CO₂ capture rates, when all produced byproducts can be used to displace cement; at higher CO₂ capture rates, SCM use efficiency decreases. Vertical line indicates capture rate of 1 Mt of CO₂ year^–1 (scenarios described in Table and Table S1).

3.1. Feedstock and Transportation

3.1.1. Transportation

Ultramafic feedstock like olivine is geographically limited, so transportation is appropriately identified as a major consideration in mineralization feasibility studies. , For scenarios using Twin Sisters olivine, however, feedstock transport was a relatively minor contributor to CO₂ emissions (10 kg of CO₂e/t of CO₂ stored), given the proximity of the mine and mineralization site. In contrast, using feedstock from Norway contributed 208 kg of CO₂e transport emissions/t of CO₂ stored. Although transoceanic bulk shipping is nearly nine times more efficient than trucking, the long distance (nearly 20 000 km) of the Norwegian option resulted in a significant emissions penalty, which was only partially offset by lower comminution emissions in Norway from hydropower electricity. This transportation penalty limited other parameter choices; for example, using DAC and/or indirect mineralization technology (which both have high energy-related emissions) resulted in net positive emissions. In contrast, with local feedstock, more system parameterizations remain net CO₂ negative. Emissions from downstream transport of co/byproducts were more significant than local feedstock transport (up to 59 kg of CO₂e/t of CO₂ stored, depending on scenario), highlighting the importance of industrial ecosystem considerations in siting. ,

3.1.2. Feedstock Grain Size

For Twin Sisters olivine, grinding accounted for the largest fraction of feedstock-related emissions, emitting 97 kg of CO₂e/t of CO₂ stored (10 μm grain size). Using larger grained feedstock significantly reduces related emissions (e.g., 8 kg of CO₂e/t of CO₂ stored for 200 μm grain size). However, grain sizes <20 μm may be necessary to facilitate carbonation. Other CDR pathways utilizing olivine (e.g., enhanced rock weathering) are similarly constrained to grain sizes <100 μm. Comminution relies entirely on electricity, so emissions will decrease with grid decarbonization, and may also benefit from efficiency improvements.

3.2. CO₂ Source

The fossil CCS scenario was naturally net CO₂ positive, with 659 kg of CO₂e emitted/t of CO₂ stored, indicating a 34% efficiency for captured emissions. Using the same technology with biogenic C (CDR scenario 1), CDR efficiency was similarly 34% (341 kg of CO₂e/t of CO₂), excluding displacement (Figure ). For scenario 1, CO₂ capture was the largest single source of emissions (308 kg of CO₂e/t of CO₂ stored). Other studies of mineralization systems similarly found CO₂ sourcing to contribute disproportionately to total emissions. , Using DAC as a CO₂ source (scenarios 2 and 3) more than doubled capture-related emissions to 647 kg of CO₂e/t of CO₂ stored, moving the system toward net positive emissions. Given the high emissions penalty for DAC, achieving net negativity for a coupled DAC-mineralization system on the current grid would require minimizing other emission factors (e.g., feedstock grinding and/or mineralization technology) or maximizing avoided emissions (e.g., SCM use, as in Figure ).

3.3. Mineralization Pathway

Mineralization technologies vary widely in associated emissions, but direct mineralization with olivine is demonstrably one of the most efficient processes. , Of the two methods considered here, the direct process emitted 500 kg less CO₂e than the indirect process (Figure ). Indirect technologies merit further research, however, because they enable the sequential extraction of high purity co-products (e.g., SCMs and metals), which may be more efficiently recovered than with direct technologies. Given the large emissions penalty, co-products must be recovered and used at high rates to achieve net system CO₂ negativity. For example, with no co-product use, net emissions for CDR scenario 3 were positive (+528 kg of CO₂e/t of CO₂ stored), but the high SCM use case could decrease net emissions to −44 kg of CO₂e/t of CO₂ stored. Indirect technologies are also potentially cheaper to build and operate, and will become more efficient with grid decarbonization, given the greater dependence on electrical vs thermal energy (Figure A vs Figure B). ,

If the indirect technology is used for the production of Mg­(OH)₂ followed by passive capture of ambient CO₂ (CDR scenario 4), net emissions are reduced significantly (−154 to −356 kg of CO₂e/t of CO₂ stored, with and without avoided emissions). If Mg­(OH)₂ is used to sequester dissolved CO₂ as HCO₃ ^– (scenario 5), emissions decrease even further since more CO₂ is sequestered per t of feedstock (−627 to −489 kg of CO₂e/t of CO₂ stored, with and without avoided emissions). Note that avoided emissions in these scenarios assume default SCM use of only 4% SiO₂ in cement, since no MgCO₃ is produced directly. Higher SiO₂ use rates would increase avoided emissions significantly.

While Mg­(OH)₂ can be used for novel CDR applications, it can also be used in existing industries like wastewater treatment as a replacement for conventional caustic materials like NaOH. In substitution applications like this, it is important to consider that the resulting capture of CO₂ as dissolved HCO₃ ^– is not additional, as it was previously accomplished with NaOH. However, avoided emissions could potentially be claimed for displacement of NaOH (Supplemental Note 4). CDR, or net atmospheric removal, should only be claimed in these applications if excess Mg­(OH)₂ is used to capture additional CO₂. For CDR applications like OAE, where displacement of conventional products does not occur, only CDR (and not avoided emissions) should be considered. In all cases, counterfactuals should be considered carefully. In our analysis, we consider only applications with the intentional goal of CDR, and thus do not include scenarios with avoided emissions from Mg­(OH)₂.

3.4. Grid Carbon Intensity and Future Scenarios

For CCS and direct mineralization scenarios, natural gas was ultimately the largest contributor to CO₂ emissions, due to the high thermal energy demand of both point-source capture and the direct mineralization process; cumulatively, these accounted for 54% of total emissions. Grid electricity was the next highest contributor, accounting for 25% of total emissions. For scenarios using indirect mineralization and/or DAC, natural gas use is relatively less important, and electricity contributes more to emissions (given minimal thermal energy requirements, and demonstrated heat recovery, respectively). , The NWPP subregional grid has an emission factor less than half the U.S. national average, given significant hydropower, so emissions associated with electricity are minimized compared to other locations.

Electricity will decarbonize significantly over the coming decades, and carbon intensity of the NWPP subregional grid is expected to decrease by ∼30% in 2050. As a result, net CO₂ emissions for scenario 1 will decrease from −341 to −523 kg of CO₂e/t of CO₂ captured (−842 kg of CO₂e/t of CO₂ captured, including avoided emissions) (Figure ). For DAC, cleaner electricity has a more profound impact, fundamentally changing the feasibility of a coupled DAC-mineralization system (scenarios 2 and 3) as emissions decrease from net CO₂-positive to −617 and −510 kg of CO₂e/t of CO₂ captured between 2022 and 2050, respectively. Note that in our model, DAC benefits disproportionately from grid decarbonization, given our assumption of heat pump integration. If point-source capture similarly uses heat pumps, or shifts to solvents with lower thermal energy demands, mineralization systems coupled to point-source capture or DAC would be comparable. Indirect mineralization technologies similarly benefit from grid decarbonization, given higher electricity needs (scenario 3). Considering use of Mg­(OH)₂ for dissolved CO₂ capture (scenario 5), the indirect mineralization system can become highly efficient, reaching −798 kg of CO₂e/t of CO₂ captured (−936 kg of CO₂e/t of CO₂ captured, including avoided emissions).

For illustrative purposes, we consider “best case” DAC scenarios utilizing 100% wind power and/or battery-electric trucks. The emission factor for wind energy is similar to the current local grid carbon intensity, which is dominated by hydropower (Whatcom Public Utilities District). Using wind power for DAC and direct mineralization, the system is 74% efficient (−737 kg of CO₂e/t of CO₂ captured), excluding avoided emissions. This increases slightly to 76% in 2050 with technology efficiency gains. Considering maximum SCM use in the regional cement industry (options 3 or 5), net emissions with wind energy could be as low as −1.6 t of CO₂e/t of CO₂ captured. “Best case” Mg­(OH)₂ scenarios using wind power are 90% efficient (exceeding −900 kg of CO₂e/t of CO₂ captured) without displacement, and exceed −1.5 t of CO₂e/t of CO₂ captured with optimal SCM use. Incorporating battery-electric trucks into the wind-powered DAC system could achieve modest gains, increasing total efficiency from 74 to 79% (82% for 2050). Considering the local grid, using battery-powered trucks in 2050 would increase system efficiency to 67% (vs 62% using conventional diesel trucks).

3.5. CO₂ Avoidance

3.5.1. Cement Displacement

If avoided emissions are considered, utilization of SiO₂ and MgCO₃ co-products as SCMs has the largest single impact on net emissions. The magnitude of SCM use is a function of not only cement substitution rate, but also SCM production rate relative to the size of the cement industry. Large mineralization systems coupled to small cement industries that are incapable of absorbing the produced SCMs would result in inefficient SCM use and high landfill rates. Considering CDR scenario 2, maximally efficient systems capture between 0.05 and 1 Mt of CO₂ year^–1, depending on SCM use case. (Figure ). For larger systems, exceeding 5 Mt of CO₂ captured year^–1, net CO₂ emissions converge on the “No SCMs” scenario, as production dwarfs the local and regional cement industries. For systems of intermediate size, CO₂ avoidance from SCM use varies greatly, with SCM production rate and % cement substitution both being important factors. This large range in avoided emissions across scales of CO₂ capture highlights that the size of the local/regional cement industry is critical for assessing the net benefit of a mineralization project in a specific area.

The ideal size of a mineralization system, in terms of total negative emissions, is thus a trade-off between CO₂ captured and potential CO₂ avoided. Larger systems would have higher total CO₂ capture, but lower avoided emissions. For example, the default 1 Mt of CO₂ year^–1 scenario achieves a maximum of −0.9 t of CO₂e avoided/t of CO₂stored, whereas a smaller system could achieve −2.2 t of CO₂e avoided/t of CO₂ stored. Project economics likely follow a similar trend; depending on carbon and SCM pricing, revenue is maximized at a certain rate of SCM use efficiency.

The productive use of MgCO₃ can greatly increase total SCM utilization because (a) both co-products can be used additively, and (b) the mineralization process produces 2.5 times more MgCO₃ than SiO₂. Thus, while using SiO₂ alone results in −1.1 t of CO₂e/t of CO₂ stored, use of SiO₂+MgCO₃ can reach 2.2 t of CO₂e/t of CO₂ stored. This may be optimistic given limitations in our understanding of MgCO₃ in cement; further research is needed on high substitution rates (Supplemental Note 4). If MgCO₃ use is limited as an SCM, it could be useful instead as a cement filler or aggregate. For these uses, however, emission factors are much lower than for SCMs, and resulting avoided emissions would be trivial. Likewise, the economic value is lower.

Because SCM use has such a high potential impact on net project emissions, we note some important considerations. First, cement gradually sequesters atmospheric CO₂ over time, resequestering over 40% of the CO₂ emitted by calcination over its expected lifetime. By displacing cement, SCMs may limit this reuptake. While it is reasonable to account for avoided emissions from SCM use upon cement displacement, it may also make sense to apply a discount rate over time, as the counterfactual would include potentially higher rates of gradual CO₂ uptake. Second, the carbon intensity of cement is likely to decrease in the coming decades, as a result of industry-wide increase in SCM use, alternative fuels, and widespread adoption of CCS. Longer term, carbon intensity may decrease even further by revolutionary technologies that replace OPC altogether with cements that do not require the calcination of limestone. Such innovations will reduce the value of mineralization-derived SCMs for emissions avoidance. Currently, the global capacity for SCMs in blended Portland cement exceeds supply of conventional SCMs, so SCMs produced by mineralization can be considered fully additive. If novel SCMs become more commonplace, however, the relative benefits of mineralization-derived SCMs would need to be re-examined.

3.5.2. Metal Displacement

For most scenarios, avoided emissions from co-production of metals totaled −155 kg of CO₂e/t of CO₂ captured, comparable to avoided emissions from modest rates of cement displacement. Avoided emissions from metal production was dominated by Fe and Ni, due to their relatively high concentrations in olivine (6% and 0.2%, respectively); recovery of lower concentration metals resulted in trivial emissions avoidance. The total mass of valuable metal production was 0.12 Mt (90% of which was Fe), compared to 2.6 Mt of SCM co-products. Although SCMs are produced in greater mass, they are not fully utilized in the default use case, resulting in similar avoided emissions for metals and SCMs. We note that this avoidance should be considered a maximum, given the assumption of zero additional emissions for metal recovery processes; in reality, metals recovery will likely require energy for reagent recovery and postprocessing. Furthermore, selecting appropriate emission factors for intermediate metal products is challenging, and has a direct impact on estimated avoided emissions (Supplemental Note 5). Assuming more conservative emission factors could further reduce associated avoided emissions.

While metal recovery makes only a modest potential contribution to CO₂ avoidance, increasing demand for metals used in decarbonization technologies encourages further research in integrated metals recovery. For example, Ni demand is expected to increase by as much as 8-fold by 2035. As demand increases and ore grades decline, both the environmental and economic benefits of enhanced metal recovery from alternative processes may increase. − To highlight the potential value of metal co-products, we consider a 1 Mt of CO₂ year^–1 DAC scenario. Carbon credits could yield $100 million (at a target carbon price of $100/t of CO₂), and SCMs could be worth $8.5 to $27 million, if priced as steel slag or cement ($41 and $130 t^–1, respectively). Considering prices for pure metals as a maximum, recovered Ni would be worth up to $82 million and cobalt an estimated $13 million. Thus, metal co-products may be as important to the success of mineralization projects as carbon credits, although markets for both are rapidly evolving.

3.6. Comparison to Other CDR Pathways

When coupled to point-source capture of biogenic C or DAC, mineralization can be compared to alternative sequestration pathways like injection into subsurface geologic reservoirs. Geologic sequestration requires compression of CO₂ for transportation and injection, requiring approximately 100 kWh/t of CO₂. , For the U.S. grid average, this results in less than 50 kg of CO₂e emitted per t of CO₂ injected and stored. As shown here, emissions associated with mineralization are significantly higher. Even considering efficiency gains expected by 2050, mineralization-related emissions will likely exceed 200 kg of CO₂e/t of CO₂. Thus, using mineralization solely for sequestration of captured CO₂ is not an efficient option if subsurface storage options are available. However, if avoided emissions from co-product displacement are considered, mineralization can result in greater net negative emissions than CO₂ sequestration alone. For locations without suitable subsurface geology, mineralization may also be a desirable option. In addition, subsurface storage faces scale-up and permitting hurdles that may limit growth in the coming decades, encouraging exploration of alternative sequestration options.

Considering mineralization as a capture and sequestration technology, enhanced rock weathering (ERW) is a comparable alternative. ERW uses ground ultra/mafic feedstock to sequester CO₂ passively in open systems like agricultural fields. Open system applications are theoretically more efficient than mineralization because they do not require energy for CO₂ capture, so associated emissions are limited only to mining, comminution, and transportation. In addition, the stoichiometric advantage of storing carbon as dissolved HCO₃ ^– instead of MgCO₃ results in more carbon sequestration per unit olivine. As a result, current estimates for agricultural ERW are more efficient than many of the scenarios presented here, ranging from 70 to over 95% efficiency, depending on feedstock and grid carbon intensity. − It should be noted, however, that these studies assume ideal dissolution kinetic models and no downstream losses; actual CO₂ uptake in the environment is likely much lower, resulting in overestimated efficiency estimates. − Furthermore, if avoided emissions from co-product use are considered, mineralization can be more efficient, with systems that maximize SCM use exceeding 100%.

Determining which carbon capture and sequestration technology is best suited for a specific location depends on many factors. As shown here, CO₂ mineralization can be extremely efficient from a net emissions perspective, given low-carbon electricity, short transport distances, if Mg­(OH)₂ is used aqueously (reaction ), and if SCM use is maximized in the cement industry. Without the advantage of one or more of these factors, mineralization may be inefficient (and even net CO₂ positive) compared to other carbon management options. Furthermore, it is important to note that other impact categories not included in this study, such as air pollution, land use, and human or eco-toxicity, also need to be considered. Future studies should examine the potential trade-offs between GHG emissions and other environmental impact categories, as well as economic costs and benefits.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.5c02627.

• Additional methodological details and notes, including study location (Supplemental File 1) (PDF)

• Excel-based lifecycle assessment model (Supplemental File 2) (XLSX)

• Model results of all explored scenarios (Supplemental File 3) (XLSX)

The authors declare no competing financial interest.