Activation energy of magnesite (MgCO 3 ) precipitation: recent insights from olivine carbonation studies

We present two new activation energies for magnesite precipitation during forsteritic olivine (Mg 2 (cid:1) x Fe x SiO 4 ; 0.18 # x # 0.26) carbonation in high-pressure carbon dioxide. These new activation energies of 89 (cid:3) 6 and 85 (cid:3) 1 kJ mol (cid:1) 1 are consistent with the literature for magnesite precipitation in aqueous media and extend the temperature range to encompass 90 (cid:4) C to 50 (cid:4) C. These insights will help improve understanding of mineral transformation kinetics in the subsurface, including carbon storage in ma ﬁ c-ultrama ﬁ c environments, and aid in the development of carbon dioxide removal (CDR) and net negative-emissions technologies. Olivine is a key constituent of reactive geologic formations and industrial wastes that are targets for permanent carbon storage via mineralization. The relative paucity of kinetic parameters for olivine transformation to magnesite via coupled dissolution and carbonate precipitation hinders e ﬀ orts to predict rate and design e ﬃ cient mineralization strategies. Our calculations of two new olivine carbonation activation energies help address these knowledge gaps relevant to natural and engineered environmental carbon-management processes.

The concept of carbon dioxide removal (CDR) through carbon capture and sequestration is an integral component of current climate mitigation strategies and pursuit of net-negative emissions technologies. A promising CDR approach involves injection of carbon dioxide (CO 2 ) into reactive mac and ultramac rocks to form stable carbonate minerals, enabling rapid permanent carbon storage. [1][2][3][4][5][6][7][8] In this context, understanding rates of mineral carbonation is crucial for predicting fate and transport of subsurface CO 2 .
Olivine (Mg 2Àx Fe x SiO 4 ) is a key reactive component of mac and ultramac rocks, and its dissolution, hydration, and carbonation rates have received considerable scrutiny (c.f., ref. [9][10][11][12][13]. The recent quantitative kinetics analyses and compilations of Miller et al. 11 and Sendula et al. 12 t the Avrami model 14 and shrinking particle model (SPM), 12,15-17 respectively, to the broad olivine carbonation literature. The more recent and comprehensive study of Sendula et al. 12 provided 35 new experiments, nearly doubling the amount of available datasets, and the SPM proved most exible and adaptable for the diverse olivine carbonation literature. The goal of the present Communication is to extract carbonation activation energy parameters from recently compiled olivine carbonation studies. 11,12 To do so we critically reviewed the datasets to identify two 12,18 suitable internally-consistent collections of reaction rate vs. temperature data for magnesite precipitation during olivine carbonation. These datasets were suitable as they included reaction kinetics for at least three distinct temperatures.
The San Carlos olivine used in Sendula et al. 12 has $88-91% of the divalent metal sites occupied with Mg 2+ ( Plots of the Sendula et al. 12 (Se21, 50-150 C) and Gadikota et al. 18 (Ga14, 90-150 C) carbonation rates on Arrhenius plots ( Fig. 1a and b) illustrate the linear relationships needed to calculate apparent activation energies. The linearity of the Arrhenius plots indicates that temperature is the dominant control, and other possible variations in chemical affinity and pressure 12 (Fig. 1c) are negligible, at least for these far-fromequilibrium high-pressure carbonation studies. The olivine to magnesite activation energy values are "apparent" as they encompass contributions from all elementary reactions involved in the complex dissolution-precipitation processes. The calculations revealed the apparent activation energies of 89 AE 6 (Se21) and 85 AE 1 (Ga14) kJ mol À1 . These newly-determined  (Table 1). This present analysis extended the temperature range of the Table 1 dataset down from 90 C to 50 C. Although the studies compiled in Table 1 span a range of aqueous-mediated processes, including olivine carbonation, hydromagnesite transformation, and step advancement on magnesite, all values are presented given the paucity of literature data. Our group at Pacic Northwest National Laboratory has also studied the inuence of adsorbed water nanolm thickness on the activation energy of forsterite to magnesite carbonation, demonstrating a linear relationship between reported monolayer H 2 O thickness and activation energy, from $34 to $130 kJ mol À1 . [24][25][26] Given the occurrence of multiphase CO 2 -H 2 O uids, it is vital to understand the barriers to magnesite precipitation in aqueous media to predict and interpret experiments conducted in nonaqueous regimes (e.g., water lms).
In summary, this Communication presents two new robust activation energies for the olivine to magnesite carbonation reaction. These types of monomineralic studies are important for delineating controlling reaction mechanisms and kinetic interpretation of mac-ultramac rock carbonation studies (e.g. ref. 22,[27][28][29][30][31][32][33][34][35]. Further insights from dynamic kinetic model 36 and reactive force-eld 37,49 development, along with additional carbonation kinetics studies, 12,16,[38][39][40] are vital for clarifying the multiscale mechanisms and rates of silicate carbonation Fig. 1 Arrhenius plots using the carbonation rate results of (a) Sendula et al. 12 (Se21) and (b) Gadikota et al. 18 (Ga14), showing the variation of the natural logarithm of the olivine to magnesite transformation rates (J, mol m À2 s À1 ) as a function of 1000 times the reciprocal absolute temperature (T) of the experiments. Temperature ( C) is labelled on the upper x-axis for reference. The calculated apparent activation energies, coefficient of determination, and uncertainties are given next to the linear best fits. Red and dark cyan curves denote 95% prediction band and 95% confidence bands, respectively. In panel (c), the Arrhenius trends have both been plotted on the C vs. ln J plane, while the Sendula et al. 12 and Gadikota et al. 18 rates used to construct the Arrhenius plots are shown in the context of pressure and temperature conditions. The reference drop lines from the points to the P-T plane help clarify the 3D perspective. transformations. Our analysis provides a basis for focusing future work on key mechanistic and kinetic unknowns that could improve understanding of mineral transformation kinetics in the subsurface, including carbon storage in mac-ultramac rocks, and aid in the development of carbon dioxide removal and net negative-emissions technologies.

Conflicts of interest
There are no conicts of interest.