IR absorption coefficients for water in nominally anhydrous high-pressure minerals
IR absorption coefficients for water in nominally anhydrous high-pressure minerals
American Mineralogist (May 2010) 95 (5-6): 770-775
Infrared spectroscopic quantification of traces of OH and H (sub 2) O in minerals and glasses is based on the Beer Lambert law A=epsilon c.t, where epsilon is the molar absorption coefficient. Numerous experimental and theoretical studies show that epsilon generally increases with decreasing wavenumbers. However, this general trend seems to be valid only for hydrous minerals and glasses and should not be applied to water quantification in nominally anhydrous minerals (NAMs) that incorporate traces of water in their structures. In this study, we analyze epsilon -values from literature data and propose that within a polymorphic mineral series of the same composition epsilon negatively correlates with the molar volume and positively correlates with the density of the respective mineral phase. To test this hypothesis, we determined epsilon -values for synthetic hydrous ringwoodite samples ranging in composition from X (sub Mg) =0.0 to 0.6 by combining results of FTIR-spectroscopy with those of secondary ion mass spectrometry. The epsilon -values plot well below the general calibration curves but follow the same trend, i.e., they increase with decreasing wavenumbers of the OH bands from 59000+ or -6000 for the Fe end-member to 85800+ or -10000 L/mol(H (sub 2) O)/cm (super -2) for the Mg-richest sample. From this relation we can predict an absorption coefficient for the iron-free Mg end-member as 100000+ or -7000 L/mol(H (sub 2) O)/cm (super -2) . This value together with epsilon -values for forsterite and wadsleyite taken from literature confirm the proposed correlation with the molar volume and density within this polymorphic series. This allows us to predict absorption coefficients for some minerals, where coefficients for one or better two of their polymorphs, either high- or low-pressure, are available.