Engineered carbon mineralization in ultramafic rocks for CO2 removal from air: Review and new insights

Abstract Carbon Dioxide removal from air (CDR) combined with permanent solid storage can be accomplished via carbon mineralization in ultramafic rocks in at least four ways: 1. Surficial CDR: CO2-bearing air and surface waters are reacted with crushed and or ground mine tailings, alkaline industrial wastes, or sedimentary formations rich in reactive rock fragments, all with a high proportion of reactive surface area. This can be implemented at a low cost, but most proposed methods have a very large area footprint at the gigatonne scale. The area requirement can be greatly reduced by calcining (heating to produce pure CO2 for permanent storage or use) followed by recycling of MgO, CaO, Na2O, … Such looping methods have predicted costs that are as low or lower than for direct air capture with synthetic sorbents or solvents (DACSS), and a similar area footprint. 2. In situ CDR: CO2-bearing surface waters are circulated through rock formations at depth. These methods potentially have a cost similar to that of surficial carbon mineralization, and a giant storage capacity with reduced surface area requirements, but they involve uncertain feedbacks between permeability, reactive surface area, and reaction rate, providing a fascinating topic for fundamental research. Furthermore, the size, injectivity, permeability, geomechanics, and microstructure of key subsurface reservoirs for in situ CDR remain almost entirely unexplored. 3&4. Combined partial enrichment of CO2 using direct air capture with synthetic sorbents (DACSS) plus surficial carbon mineralization (3) or in situ carbon mineralization (4). Energy requirements and total costs for partial enrichment of CO2 are substantially lower than for enrichment to high purity. CO2 enriched air can be sparged through mine tailings at the surface, and/or through water to increase dissolved carbon concentrations prior to circulation through rock reactants. Such combined or hybrid approaches have not been investigated thoroughly, and offer many avenues for optimization.