Analysis of an altered simple silicate glass using different mineral and glass standards

Quantitative analyses of alteration products formed during the aqueous corrosion of glass were performed using four different sets of standards: relevant mineral standards, an NBS glass standard, unreacted simple silicate glass, and the unreacted center of the reacted glass. A simple silicate glass (containing Na, Mg, Al, Si, and Ca) was reacted in water vapor at 200° C for 14 days. Up to eight alteration phases, including a Mg‐rich smectite clay and a zeolite intermediate in composition between Ca‐harmotome and phillipsite, formed on the glass surface. A set of energy‐dispersed spectra (EDS) of the bulk glass, the clay, and the zeolite were collected from a polished cross‐section of the reacted sample. The sum of the elemental weight percents varied by as much as 48% using the different standards. These differences are due to poor electrical conductivity through the clay layer. When the analyses were normalized to 100%, the differences between analyses using the standards were small (less than 10 wt% for individual cations and <5 total wt%). Normalization is inappropriate in the analysis of hydrated alteration phases, which contain a significant amount of water. The inability of the EDS system to detect water requires that absolute analyses be used to estimate the water content of the alteration phases. The bulk glass standard analyses were best suited for this purpose. In addition, use of the bulk glass standard is preferable in our application because spectra of the altered layer and bulk glass can be acquired from the same specimen. Analytical errors that may occur due to poor conductivity through the reacted layer, drifting probe currents, or varying thickness of the carbon coating are also reduce.