- © The Mineralogical Society Of America
A great deal of mineralogy has focused on determining crystal structures of minerals. Amorphous materials are often compared to crystalline materials of the same composition in order to predict atomic structure in the disordered phase. However, there are many instances when such analogies will break down or no crystalline counterpart exists. Even mineral surface structures may be significantly different from the bulk mineral structure, particularly when the mineral surface has undergone reaction with aqueous solutions (e.g., Hellman et al. 1997). Although unambiguous determination of the atomic structure of an amorphous phase is generally more problematic than for crystalline phases, modern spectroscopic techniques (see Hawthorne 1988 for a review) can shed a great deal of light on the atomic structure of phases with long-range disorder. Vibrational (infrared and Raman) spectroscopy is a particularly useful tool for studying many types of phases (McMillan and Hofmeister 1988 and references therein). Unfortunately, infrared and Raman spectra may be subject to various interpretations. Molecular modeling of vibrational spectra can be useful for sorting out these debates.
For example, the mechanism of CO2 dissolution in albitic melts is difficult to understand. Fine and Stolper (1985) measured infrared spectra and Mysen and Virgo (1980) measured Raman spectra of CO2-bearing albite (NaAlSi3O8) glasses. Both studies indicated the presence of a carbonate species. Although carbonate readily forms in more basic melts through the combination of CO2 with a “free” oxygen atom (where “free” means not bonded to a tetrahedral cation (Mysen and Virgo 1980), there were thought to be no “free” oxygen atoms in fully-polymerized Na-aluminosilicates. Mysen and Virgo (1980) interpreted the spectra in terms of a depolymerization of the melt with formation of octahedral Al from the previously tetrahedral Al atoms. Fine and Stolper (1985) proposed a different mechanism whereby a …