Considerable uncertainty persists as to the conditions undergone by samples during high pressure, high temperature solid media experiments, despite decades of research using piston-cylinder and multi-anvil apparatus. In the piston-cylinder, the workhorse of experimental investigations at the conditions of most of the crust and of the uppermost part of the mantle, the contributions of competing effects such as friction and thermal expansion remain controversial, with little consensus regarding the conditions actually undergone by experimental samples.

The computer-controlled piston-cylinder presses developed in-house and commissioned at RSES in 2016 provide what is, to our knowledge, a unique facility, allowing for the conditions of piston-cylinder experiments to be controlled and monitored with a degree of precision that was previously impossible. This facility has been used to conduct several series of experiments to investigate the calibration of sample pressure, the results of which have shown some long-standing assumptions regarding sample pressure in the piston-cylinder to be inaccurate.

This understanding of the sample conditions in solid media apparatus is applied to the spinel/garnet transition, a phase transition that is crucially important for geochemical and geophysical models of the upper mantle. The existing experimental dataset covering this reaction is sharply divided, with a discrepancy equivalent to c. 10 km of depth in the mantle between the results of different workers in superficially similar experiments. The pressure and temperature of the spinel/garnet transition is simultaneously determined in MAS (MgO-Al₂O₃-SiO₂) and CMAS (CaO-MgO-Al₂O₃-SiO₂) systems and this long-standing discrepancy between experimental data is resolved. These new results are used to evaluate the performance of the most popular thermodynamic model covering phase relations in the upper mantle.