| Summary: | The aqueous chemistry, precipitation, and crystallization of metal-carbonates comprises a vast field of research that underlies the urgency of CO2 sequestration, ocean-acidification, and biomineralization. The results of recent experimental and computational studies suggest that amorphous calcium and magnesium carbonates are precipitated from supersaturated aqueous conditions by non-classical aggregation of ion pairs, dimers, dynamically-ordered-liquid-likeoxypolymers (DOLLOPS), and prenucleation clusters (PNCs). We present the first high field (20 T) 43Ca and 25Mg NMR studies of amorphous calcium-magnesium carbonates (ACC, ACMC, AMC) materials. Direct integration of computational techniques with experimental NMR provides a novel step forward toward multi-scale integration of computational and experimental techniques. Supporting information is derived from X-ray diffraction (XRD), thermogravimetric/differential thermal analysis (TGA-DTA), and scanning electron microscopy—energy dispersive spectroscopy (SEM-EDS) and provides important comparison to the bulk structures and composition. High field NMR of amorphous carbonates demonstrates that amorphous carbonates contain various types of local disorder, but does not corroborate the theory of polyamorphism nor nano scale phase separations postulated by other workers. Carbon (13C) NMR of 13Cenriched materials indicates a degree of Ca-Mg solid solution in ACMCs, as ACMC 13C resonances cannot be adequately reconstructed from the pure ACC and AMC 13C resonances. However, with increasing Mg-content (and therefore H2O content) 13C NMR resonances are strongly influenced by water-carbonate hydrogen bonding, shifting to lower resonance frequency and broadening. The 13C-NMR are well-fit with single Gaussian distributions, suggesting that two-phase models of ACMCs are not required to explain our 13C NMR observations. Protoncarbon cross polarization indicates that there is a H population proximal to carbonate groups for all amorphous phases. 43Ca NMR yields line shapes that span the resonance frequency range of all known crystalline calcium carbonate polymorphs and is well fit with a single Gaussian distributions. 43Ca NMR does not support a theory of polyamorphisms, but rather suggests an unstructured, continuous distribution of local environments that is unlike any specific crystalline phase. The mean 43Ca chemical shifts vary 0.77 ppm from compositions x = 0 to 0.5 [x = Mg/(Mg + Ca)], demonstrating that Mg2+ has very little influence on the molecular-scale 43Ca environment in ACMCs. Through integration of quantum mechanical calculations, classical MD, and NMR we ascertain a maximum mean Ca-O bond distance in our ACCs/ACMCs of 2.45 ± 1 Å that is independent of composition. Unlike the indistinguishable local calcium environments, 25Mg NMR of amorphous material gives evidence for several distinct overlapping quadrupolar line shapes. These sites do not generate NMR resonances that are perfect matches for known crystalline polymorphs of magnesian carbonates and extend toward lower resonance frequencies far beyond the range of known equilibrium analogs. By comparison to the range of reference phases, the low frequency singularities of ACMC-AMC resonances are consistent with some population of Mg-O bond distances greater than 2.10 Å and/or some fraction of sites with high coordination numbers (up to 8). The local Mg environment of a protodolomite crystallization [x = Mg/(Mg + Ca) = 0.6] exhibits 25Mg NMR parameters most similar to the asymmetric Mg2+ coordination environment of lansfordite [Mg(CO3)2(H2O)4]2– or huntite. Although H-C cross polarization indicates no H-bonding with carbonate the XRD gives not longrange indications of huntite. The large effective radius of strongly hydrated Mg in the protodolomite likely provides a driving force for cation ordering in dolomite.
|
|---|