An asymmetric ultrasonic interferometry high-pressure set-up for elastic wave velocity measurements under simulated Earth's mantle conditions has been developed.
The piston-cylinder experiments were performed in a system with a piston diameter of 22 mm. The pistons were designed to optimize ultrasonic transmission. Each piston was equipped with three lithium niobate transducers of 16.7 MHz natural frequency, one P-wave transducer and two S-wave transducers on one piston, and reverse on the other one. The configuration allows ultrasonic measurements in reflection, transmission, two and single transducer generation and detection for both compressional and shear wave velocities. In the symmetric configuration two platinum buffer rods were used inside the cell to transmit ultrasonic waves to the sample and as pressure transmitting medium. A similar configuration was used in multi-anvil experiments (MAX80 at HASYLAB, HPX97 at Universität Bonn). A strong increase in signal-to-noise ratio was achieved by using an asymmetric configuration with high impedance contrasts, using fused quartz and sodium chloride as buffer rods. The configurations were tested with aluminium, San Carlos olivine single crystals and artificial polycrystalline anorthite samples up to 3 GPa pressure.
The results of the first measurements with aluminium demonstrated the feasibility of ultrasonic interferometric methods in the multi-anvil pressure cell MAX80 and similar high-pressure systems. The experiments with a piston-cylinder apparatus, optimized and adapted to ultrasonic measurements, showed that the results are not influenced by the shape of the travel path. The experiments up to 3 GPa pressure with natural San Carlos olivine allowed a detailed comparison of the results with the published data of several authors. Artificial anorthite samples were used to check the feasibility of measurements with polycrystalline samples. Our data solved discrepancies between published high- pressure and normal-pressure data for polycrystalline anorthite leading to vp = 7.28 km/s, vs = 3.93 km/s at ambient conditions and dvp/dp = 0.027 km/s * GPa, dvs /dp = 0.001 km/s * GPa. In general all our data correspond to published results within the accuracy of the method.