Making reference glasses for measuring volcanic gases

A summary of Moussallam, Towbin, Plank et al. (2024), ND70 series basaltic glass reference materials for volatile element measurement, Geostandards and Geoanalytical Research
Published

July 2024

Our paper on the ND70 reference glasses came out in Geostandards and Geoanalytical Research: Moussallam, Towbin, Plank et al. (2024). It describes a set of basaltic glasses with known amounts of water, CO₂, sulfur, chlorine and fluorine, made so that labs can calibrate their instruments over the full range found in nature.

Title block of the paper: ND70 Series Basaltic Glass Reference Materials for Volatile Element (H2O, CO2, S, Cl, F) Measurement and the C Ionisation Efficiency Suppression Effect of Water in Silicate Glasses in SIMS, by Yves Moussallam, William Henry Towbin, Terry Plank and co-authors

A gap in the calibration data

Much of what we know about volcanic gases comes from tiny glass inclusions trapped in crystals, measured with an ion probe (SIMS). SIMS is a relative method: it only gives concentrations if you calibrate it against glasses whose composition is already known.

The available basaltic standards did not cover the range of real samples. Water calibrations above about 1 wt% were extrapolated, and the CO₂ standards stopped near a few hundred ppm, well below many natural melts. Any measurement outside that range depended on a straight line drawn past the last data point.

Making the glasses

The ND70 series was made to fill that gap. We started from a single basalt, added known amounts of water, CO₂, sulfur, chlorine and fluorine, and melted each mix at 1 to 1.5 GPa in a piston-cylinder press so the gases stayed dissolved. The series spans 0 to 6 wt% water, 0 to 1.6 wt% CO₂, and 0 to 1 wt% sulfur, chlorine and fluorine. Older standards from earlier studies were included so the labs could cross-check them against the new ones.

Labs in the US, France and Japan then measured the glasses by several independent methods: elastic recoil detection, nuclear reaction analysis, an elemental analyzer, infrared spectroscopy, electron microprobe, and SIMS at four different ion-probe facilities.

Seven scatter plots of measured water content against loaded water content, each close to a one-to-one line

Measured water content against the amount loaded into each glass, for seven different methods and labs. Points fall close to the 1:1 line. Figure 1 from Moussallam et al. (2024), Geostandards and Geoanalytical Research.

What we found

The amounts we loaded and the amounts measured agree well across methods and labs, so the glasses can serve as reference materials.

We also found a problem with how SIMS measures carbon. The more water a glass contains, the weaker its carbon signal. Measuring a glass with 4 wt% water against standards with less than 2 wt% water could underestimate its CO₂ two- to three-fold, and the reverse mismatch overestimates it. The fix is to calibrate with standards whose water content matches the samples, which the ND70 series now makes possible.

Four scatter plots showing the SIMS carbon calibration ratio decreasing as water content increases, with dotted best-fit curves

The SIMS carbon signal per unit of CO₂ falls as the water content of the glass rises, in every session and at every lab. Figure 8 from Moussallam et al. (2024), Geostandards and Geoanalytical Research.

Where the glasses are now

Pieces of each glass went to ion-probe labs in the US, Europe and Japan. The rest are held at the Smithsonian National Museum of Natural History, where any researcher can borrow them.

I helped start the project after finding gaps in the basaltic glass standards available for calibrating water and CO₂ by SIMS and FTIR. I ran the piston-cylinder experiments at Lamont-Doherty and helped coordinate the analyses across the labs.

Figures reproduced unchanged from Moussallam et al. (2024), Geostandards and Geoanalytical Research 48(3), 637–660, under a CC BY-NC-ND 4.0 license.