Perovskite alteration in kimberlites and carbonatites: the role of kassite, CaTi2O4(OH)(2)

A microbeam (electron microprobe, X-ray diffraction and Raman) study of pseudomorphs after magmatic perovskite from kimberlite (Iron Hill, Wyoming, USA) and carbonatite (Prairie Lake, Ontario, Canada) showed that the early product of perovskite replacement in these samples is kassite, a monoclinic (space group P2(1)/a) polymorph of CaTi2O4(OH)(2). This mineral can be readily distinguished from its dimorph cafetite (space group P2(1)/n) based on the presence of strong signals at similar to 120, 300, 330, 450, 470 and 690 cm(-1), and the absence or very low intensity of signals at similar to 250, 420, 600, 800 and 825 cm(-1) in its Raman spectrum. The strongest X-ray diffraction lines, measured for the Prairie Lake material, are [d (obs) in (I) hkl]: similar to 3.29 (100) 022; 112, ; 1.764 (61) ; 2.284 (45) 132; 2.601 (24) 130; 2.050 (17) 222; 4.81 (16) 002; 2.034 (15) 042; 2.308 (14) 202; 1.778 (14) . Diffraction lines at 3.60, 2.99, 2.79, 2.57, 2.56 and 1.91 , characteristic of cafetite, are not observed. The electron-microprobe analyses of kassite give formulae close to the stoichiometric composition. Progressive Ca leaching leads to replacement of kassite by anatase + calcite, which are also commonly observed as direct products of perovskite alteration in silica-undersaturated igneous rocks. Raman spectroscopy is the fastest and most reliable technique to identify submicroscopic anatase-calcite intergrowths that can be easily mistaken for kassite (cafetite) based on electron-microprobe data. Thermodynamic calculations indicate that conversion of perovskite into kassite and, subsequently, anatase requires initially high levels of f(H2O) in the system, followed by an increase in f(CO2) at either decreasing or constant T and f(H2O). The implications of perovskite-kassite-anatase phase relations for deciphering the late-stage evolution of kimberlites and carbonatites are discussed.