The influence of thermodynamic and kinetic factors on the formation and dissolution of calcium phosphates in complex geochemical systems
Mineralogy, Petrology and Geochemistry
Oceanography
Final Report Abstract
Formation of calcium phosphate (Ca-P) is an important process in removing phosphate (P) from natural waters. An understanding of pathways of Ca-P formation in sediment, including the influence of inhibitory and enhancing factors, is essential to assess P diagenesis and related P budgets. Main objectives of this study were (i) to investigate apatite formation in porewaters with special attention paid to the identification and characterization of a transient precursor mineral phase, its solubility and its transformation into apatite using precipitation assays employing model seawater and model brackish water solutions; (ii) to investigate differential dissolution of Ca-P reference minerals and of Ca-P minerals in marine sediment using a dissolution assay; and (iii) to use the gained insights for analyzing Ca-P speciation of a diverse series of marine sediments in order to investigate early P diagenesis in Baltic Sea sediments in a comparison with early P diagenesis in global marine sediments. The results of the study demonstrate apatite formation within few months following initial precursor precipitation in model seawater solutions matching the ionic strength and composition of seawater at supersaturations that can be reached in porewaters. Apatite formation was affected by the Mg/Ca dissolved ratio and by seeding with hydroxyapatite and fluorapatite confirming a large nucleation barrier caused by a strong inhibitory effect of magnesium on the growth of apatite and an effect of seed surfaces on the nucleation rate of apatite. Results of SEM-EDX analysis and of the dissolution assay unambiguously confirmed that the precursor mineral phase initially precipitating from seawater and brackish water is a magnesium-containing variety of octacalcium phosphate (OCP) and does not contain significant amounts of other Ca-P minerals. However, the magnesium-OCP undergoes rapid alteration because it subsequently transforms into a more stoichiometric, less substituted OCP. Generally, crystal impurities, as here magnesium, can be incorporated into the crystal lattice or adsorbed at the crystal surface. However, the Ca/P ratio of particles precipitated varied largely as their (Ca+Mg)/P ratio remained constantly close to 1.33, the Ca/P ratio of OCP, clearly indicating that the determined magnesium was not just adsorbed at the surface, but a structural part of the crystal unit cell of the Mg-OCP. This suggests a possible pathway for subsequent apatite formation because magnesium, which consequently occupies calcium sites in OCP and likely inhibits the transformation into apatite by structural deformation of the crystal nuclei, is exchanged for calcium over time. The less substituted OCP exhibited a decreased solubility relative to that of the high-magnesium OCP precipitated initially. The dependence of the OCP solubility on the degree of magnesium substitution for calcium can be explained by magnesium-induced lattice deformation. Maximum dissolved P concentrations were about two-fold higher in normal seawater (I = 0.7M) compared to brackish water (I = 0.35 M; e.g. 500µM vs. 242µM at pH 8.1) because magnesium-OCP precipitated from brackish water contained far less magnesium than magnesium-OCP precipitated from seawater. The dissolution assay utilized 13 different buffers of adjusted ionic strength supplemented with 0.01 M CaCl2 and 0.05 M MgCl2 to obtain dissolution profiles in the pH range 2.5-8.0 for an extensive set of qualified Ca-P reference minerals. Using the obtained reference profiles, the mineral phase initially precipitating from seawater and brackish water was identified (see above). Based on the dissolution profiles, differential dissolution of FAP, CFA and OCP occurred in the pH ranges < 4.0, 4.0-5.2 and > 5.2, respectively. Thus, best separaration of these minerals was achieved using a pH 4 acetate buffer solution with supplements and a pH 5.2 MES buffer solution with supplements. The dissolution assay was subsequently used in combination with conventional sediment extraction to separate detrital FAP, authigenic CFA and authigenic OCP of marine sediments (coupled method termed CONVEX method). The profiles of the CONVEX method showed that sediments of nutrient-rich sites of coastal upwelling regions and sites with high, fertilizer-derived P-loads had the highest authigenic Ca-P contents (> 60% and > 50% of the total Ca-P, respectively) and the highest OCP contents (> 12% and > 30% of the total Ca-P, respectively). Profiles of Baltic Sea sediments revealed unusual, contrasting characteristics: Low authigenic Ca-P contents (5% and 14% for the Eckernförde Bay and Gotland deep, respectively), low OCP contents (< 1% and 7% for the Eckernförde Bay and Gotland deep, respectively) and unusual P adsorption capacities. Due to the low authigenic Ca-P concentrations and exceptionally high organic P levels, Baltic Sea sediments exhibited uniquely low authigenic Ca-P / organic P ratios. The analytical results of the CONVEX method were subsequently compared with analytical results of a conventional sediment extraction method (SEDEX), which successfully separates authigenic Ca-P and detrital FAP, but does not distinguish authigenic CFA versus authigenic OCP. The comparison was performed using a compilation of SEDEX data from scientific studies (n = 23, 146 cores, 1442 samples) in order to cross-validate the analytical results and to identify possible characteristic differences of early P diagenesis in Baltic Sea sediments compared to global marine sediments including sediments from sites of the Atlantic Ocean, Pacific Ocean, Northern Ocean, Southern Ocean, Mediterranean Sea, Arabian Sea, Black Sea, East Sea and coastal basins, estuaries and lagoons. The crossvalidation showed a good agreement between analytical results of the CONVEX versus those of the SEDEX method including authigenic Ca-P (CFA + OCP). SEDEX data mirrored the unusual authigenic Ca-P / organic P ratios of Baltic Sea sediments determined by the CONVEX method. The organic P concentration was on average ca. 3.5 times higher in Baltic Sea sediments (M = 13.0 µmol/g, SD = 4.7) compared to sediments of non-Baltic Sea sites (M = 3.8 µmol/g, SD = 4.3, t(90) = 6.4, p < .001, t-test), whereas the authigenic Ca- P concentration was on average ca. 2 times lower in Baltic Sea sediments (M = 4.8 µmol/g, SD = 4.2) compared to sediments of non- Baltic Sea sites (M = 9.5 µmol/g, SD = 7.9, t(18) = -3.0, p < .005, Welch’s t-test). If data of Bothnian Sea sediments are excluded from the analysis, organic P concentrations are ca. 4 times higher (M = 14.6 µmol/g, SD = 3.6, t(88) = 6.9, p < .001) and authigenic Ca-P concentrations are ca. 3 times lower (M = 3.5 µmol/g, SD = 3.0, t(19) = -4.4, p < .001). The ca. 10 times lower authigenic Ca-P / organic P ratios indicate that the major permanent burial sink in the Baltic Sea is organic P instead of authigenic Ca-P. This sink switching suggests that the transformation of organic P into OCP and apatite is kinetically retarded on a sea-basin scale by a critical, hitherto unrecognized controlling factor. Based on known inhibitory effects of humic substances on OCP and apatite formation and on the fact that common factors, e.g. salinity and redox potential, do not explain such a pronounced P sink anomaly on a sea-basin scale, I suggest that the high organic matter input to sediments of the Baltic Sea is responsible for both, the considerable accumulation of organic P and the almost stopped transformation of organic P into Ca-P, and that the link between organic matter and retarded OCP and apatite formation is most likely humic material as the primary controlling factor. The herewith proposed mechanism is additionally supported by sedimentation rates, concentrations of iron-bound P and Ca-P and dissolved P concentrations across sediment cores.
Publications
- Cretaceous oceanic anoxic events prolonged by phosphorus cycle feedbacks, Clim. Past, 16, 757–782
Beil, S., Kuhnt, W., Holbourn, A., Scholz, F., Oxmann, J., Wallmann, K., Lorenzen, J., Aquit, M., and Chellai, E. H.
(See online at https://doi.org/10.5194/cp-16-757-2020)
