University of Alberta
C. M. Scarfe Laboratory of Experimental Petrology
Multianvil Observations
Ken Domanik and Bob Luth - June 1999

Multianvil Observations
The following are some observations we've made over the years about multianvil assemblies (i.e. part of our lab folklore). We'd be interested in seeing how these compare with the experiences of other labs. Some points may be common knowledge, some may not, some may be just plain wrong, but they are what we currently think.

1) The best pressure efficiency we've obtained from any gasket design has been using pyrophyllite gaskets backed by cardboard with a layer of teflon tape (not cut back) on the opposing cube (see attached diagram). What do other labs see?

2) We use Cr-doped MgO endcaps (the same material as the octahedron) on runs at the high-pressure limits of any given assembly. We calculate that this prevents extrusion of the ZrO2 insulating sleeve out into the gasket area and thus decreases the chance of blowouts in high-pressure runs. Are there any other labs making similar assemblies?

3) By substituting Mo foil strip contacts and solid MgO endcaps for Mo contact rings in our 14M assemblies we find that we reduce the amount of power needed to reach any given temperature by about 35%. We believe that these assemblies also provide a more hydrostatic pressure medium and better protection from thermal gradients. Metal foil contacts are commonly used on castable octahedra. Our experience indicates they also work well with solid MgO octahedra. What is the experience in other labs with this type of design?

4) We observe a long term upward drift in our calibration curves over the last 10 years. This is very pronounced in our Bi calibrations and other electrical conductivity based calibrations. During these calibrations the ram is constantly moving upward and the transition is recorded as we pass through it. The long-term upward drift is much smaller for heated, phase equilibria runs where the ram is stationary. We interpret this data as indicating that the upward drift in the electrical conductivity calibrations over time is mainly the result of wear of the press hydraulic system and piston rings.

Do other labs observe this upward trend in calibration curves over time, particularly for electrical conductivity calibrations? If so, how do they interpret this? Also if we are correct in our interpretation, how does this affect calibration for unheated runs such as deformation experiments?

5) We try to disturb the gaskets as little as possible after we assemble the cubes and glue G10 directly to the WC cube faces. However, we have heard of other labs taping their cube assemblies together and rotating the entire assembly each time a side is taped. After all sides are taped, G10 laminate is glued over it (leaving space for electrical contacts for the furnace). The procedure is supposed to allow the gaskets to settle in and make a more robust assembly. How many labs use this tape method and do they get better results with this method?

6) Our supplier tells us that G10 laminate will no longer be manufactured except by special order. The replacement material FR-4 is similar except that it is also fire resistant. Has anyone noticed any differences in performance using FR-4?