Author: Karsten Knorr

Abstract
Since more than 30 years piston/cylinder cells are used in neutron powder-diffraction for high pressure studies above 5 kbar. Below that pressure, gas cells are well established. The piston/cylinder technique is applied to maximum pressures of about 30 kbar. The reason for that limit is the pressure capability of the cylinders, which are typically made of steel, CuBe alloy, aluminum, Ti/Zr zero scattering alloy, or sintered alumina. The diffraction patterns are contaminated by very intense Bragg peaks from the cylinder material, except for Ti/Zr that contributes a large incoherent signal.

The highest pressures (exceeding 300 kbar) are obtained using opposed anvil cells with single crystal anvils made of sapphire, moissanite or diamond. However, the sample volume is usually very small (less than 0.1 mm^3) such that investigations using these devices are restricted to the strongest scatterers and simple structures.

The use of toroidal anvils with the Paris-Edinburgh (PE) cell proved to be very successful for crystal structure investigations at spallation sources and phonon measurements at neutron reactors. At spallation sources the primary beam usually penetrates the anvils, while the diffracted radiation is collected in 90° geometry. The PE cells may reach 300 kbar of pressure with about 30 mm3 sample volume and temperatures ranging from several K to 2000 K.
The new VX-type of the PE presses has a 140° horizontal opening angle. Hence, it may be used with the incoming and the diffracted beams in horizontal geometry at diffraction instruments on reactor sources. That concept was successfully applied for the first time measuring NiO at SINQ [1].

Very recently, a VX4 cell equipped with single and double toroidal cubic boron-nitride anvils was tested at the SPODI diffractometer of the FRM2 [2]. The diffractometer was operated at 1.548 Å wavelength, in 155° back-scattering geometry, and open collimation between monochromator and sample. High-pressure measurements of magnetite, Fe3O4, prepared in an encapsulated Ti/Zr gasket with a mixture of deuterated methanol/ethanol as pressure transmitting medium were performed. The typical counting time was about 12 hours, measuring four overlaps with 100 mm readout height at the detector and a step width of 0.1°. The crystal structure data obtained from Rietveld refinement are compared to data obtained at ISIS earlier. The conclusions from that experiment are: (i.) high-pressure studies on simple systems are feasible up to pressures of about 100 kbar at SPODI, (ii.) cell parameters are of good quality, (iii.) crystal and magnetic structure refinements are possible. The improvement of the neutron flux for small samples (50 – 100 mm^3) is desirable.