Low Temperature Magnesite Formation

Mg-carbonate formation is of interest for both ex situ and in situ CO₂ sequestration strategies, yet precipitation of magnesite [MgCO₃], the most stable form, is kinetically inhibited at low temperatures.  In the Mg-carbonate system, there are numerous metastable hydrated Mg-carbonate minerals, but magnesite is the most stable and has the greatest carbon density [1], making it a preferred product forsequestering CO₂.  However, rates of magnesite precipitation are severely limited at temperatures below 60 °C due to the strong hydration of Mg²⁺ ions in solution [2].  Thus, understanding the rates of and controls on magnesite formation at low temperature remains a challenge.  Here, we examine a natural playa environment to understand the long-term viability for storing CO₂within Mg-carbonate minerals and estimate the rates of magnesite formation at near-surface conditions.  In laboratory experiments at room temperature, we use carboxylated polystyrene microspheres to accelerate magnesite formation by several orders of magnitude in comparison to the natural environment. 

Secure carbon storage within Mg-carbonate minerals requires an understanding of fundamental geochemical processes related to fluid-mineral interactions at large spatial scales and over millennial timescales [3].  Hydromagnesite-magnesite playas (hectare-scale), having formed at the Earth’s surface since the last deglaciation, demonstrate the stability of Mg-carbonate minerals for long-term carbon storage [4].  Geochemical, physical and microbial processes have facilitated carbonate precipitation from Mg-rich waters produced from the weathering of ultramafic bedrock [5,6].  Over several millennia, sediment deposition has transitioned from siliciclastic to subaqueous Ca-Mg-carbonate precipitation to subaerial Mg-carbonate deposition [4].  As a consequence of these variable modes of deposition, a complex assemblage of carbonate minerals is present within the playas including metastable hydrated Mg-carbonate minerals and magnesite.  Mineralogical analysis in concert with scanning electron microscopy (SEM) show that hydrated Mg-carbonate minerals adhere to the Ostwald’s law of phases, with more hydrated phases (e.g., lansfordite, MgCO₃·5H₂O) transforming to less hydrated and more stable phases such as hydromagnesite [Mg₅(CO₃)₄(OH)₂·4H₂O].  At surface, magnesite abundance is up to 41 wt.%, while at depth it is present at up to 86 wt.%.  Stable, radiogenic, and clumped isotope[7] data demonstrate that the magnesite is modern (<10,000 years old) and has formed at low temperatures (~3-10 °C) through direct precipitation from shallow groundwaters.  Magnesite crystal morphology varies with depth suggesting that crystal growth mechanisms also vary along the sediment profiles.  Particle size analyses show that increases in magnesite abundance are positively correlated with increasing median particle size, indicating that magnesite formation is likely to be nucleation limited.  We estimate the rate of magnesite formation (nucleation + crystal growth) is in the range of 10^-17 to 10^-16 mol/cm²/s with magnesite saturation indices ranging from 1.5 to 3.2, respectively.

Similar to magnesite, dolomite [CaMg(CO₃)₂] precipitation is also kinetically inhibited at Earth’s surface temperatures making it particularly difficult to precipitate[8].  Previous studies on microbially mediated dolomite precipitation have yielded key insights into the mechanisms of low temperature precipitation of anhydrous Mg-carbonate phases [9-11].  Carboxyl functional groups (R-COOH), found on microbial cell walls, have been shown to cause desolvation of Mg²⁺ ions in aqueous solution and facilitate dolomite formation [10-11].  Using microcosm experiments containing polystyrene microspheres with a high density of carboxyl groups, we were able to precipitate magnesite at room temperature over 72 days.  Transmission electron microscopy (TEM) and selected area electron diffraction (SAED) of a thin section obtained using focused ion beam milling were used for mineralogical confirmation.  In comparison to natural magnesite formation, this is an acceleration of several orders of magnitude without requiring energy input.  Numerous industrial and natural waters (e.g., ocean water) are supersaturated with respect to magnesite and have the potential to serve as feedstock for sequestering CO₂.  We postulate that reactive membranes or bioreactors that utilize this reaction pathway could be designed for large-scale magnesite precipitation as part of a strategy for sequestering anthropogenic CO₂.

[1] Lackner et al. (1995) Energy 20: 1153-1170. [2] Hänchen et al. (2008) Chem. Eng. Sci. 63: 1012-1028. [3] Bickle et al. (2013) Rev. Mineral. Geochem. 77: 15-71. [4] Power et al. (2014) Sedimentology. 61: 1701-1733. [5] Power et al. (2009) Chem. Geol. 206: 302-316. [6] Power et al. (2007) Geochem. Trans. 8: 13. [7] Streit Falk and Kelemen, unpublished data. [8] Land (1998) Aquat Geochem, 4: 361-368. [9] Vasconcelos et al. (1995) Nature 377: 220-222. [10] Roberts et al. (2013) PNAS. 110: 14540-14545. [11] Kenward et al. (2009) Geobiology 7: 556-565.