Extreme conditions research using the large-volume press at the P61B endstation, PETRA III

The operation of the P61B endstation large-volume press and optics of P61 are reviewed. The instrumentation at P61B, including the large-volume press, detection systems and data acquisition for in situ high-pressure experiments are described.


S1. Data acquisition and processing
Custom in-house software is available to acquire X-ray diffraction data using 1 or 2 Ge-SSD and imaging data using the X-ray microscope. The beamline controls are scripted with a graphical user interface using python and Qt, calling the operation of the Tango servers for the various devices.
Users have access to the stage movements of the LVP and slit system to shape the beam for diffraction or imaging. Pre-sets can be set up during an experiment to move to various locations in the cell assembly under pressure into the beam (e.g. sample and pressure marker locations). In order to prevent user error and unintended damage, some functions are not accessible to the user, such as individual stage movement of the detectors to avoid collisions between the detector components and the LVP.
Data acquisition for ED-XRD is implemented using a custom graphical user interface as follows. Data are saved in the hdf5 (nxs) file format and can be acquired simultaneously using 2 Ge-SSDs. There is no limit on the acquisition time, although the maximum time per frame is 50s. So, if 100 s of acquisition is needed, two frames of 50 s are used. For each acquisition, several options can be included in the file name such as, a counter, a user comment, duration, detector positions, press-stage positions, etc. Additionally, a log file is saved with the same file name containing a time stamp and all the information of the state of the experiment (pressures and positions of the LVP rams, the furnace parameters if DC heating is used, motor positions of the slits and press stages, list of detected peaks and their intensities, and beam status with current and energy information). A final third option available for each acquisition is a direct conversion of the diffraction data in hdf5 nxs file data format to readable csv format saved separately in the 'processed' directory of the beamline gpfs filesystem.
Data acquisition for radiography imaging is similarly implemented using a custom graphical user interface. Data may be acquired in hdf5 format or uncompressed tif format. When huge quantities of frames are to be captured, hdf5 container format is preferred. The following available functions are separated for data saving mode and live view mode. In data saving mode, an image or a series of images can be saved with a chosen acquisition time and delay, typically 0 s. For live view mode, only acquisition time can be modified before starting the acquisition. For both modes, a region of interest (ROI) can be implemented in order to improve the refresh rate of the camera and reject any portion of the image (e.g. where no beam is available). The software additionally features a toggle switch for 5x and 10x magnification, as well as the option for incremental z-stage image acquisition if the entire length of the assembly or a sample larger than the beam in the vertical direction, is to be imaged.
Typically, during data acquisition, immediate feedback is required for the experiment. For instance, direct estimation of the pressure and temperature conditions in the cell assembly is needed, as well as the analysis of diffraction patterns for the occurrence of a phase transition or phase nucleation. To J. Synchrotron Rad. (2022). 29, https://doi.org/10.1107/S1600577522001047 Supporting information, sup-2 meet these needs, the control hutch is equipped with extra PCs for data processing with various freely available software installed. The main software used for ED-XRD data processing is called PDIndexer (Seto et al., 2010), modified by the author for data processing at P61B. The programme reads the hdf5 (nxs) file format written by the Tango device for the Xspress digital analyser using libraries from the HDF Group (https://www.hdfgroup.org/). Upon loading data files, a dialog opens with the pre-entered information of the channel-energy calibrations for the detectors and the calibrated 2θ angle of the detector. The software can be used to estimate pressure based on a range of materials with established equations of state (such as MgO and NaCl), and peak fitting is possible from crystal data of any material in the list or imported from a cif structure file. For more information about the software, please visit the author's software homepage (http://pmsl.planet.sci.kobe-u.ac.jp/~seto/). P61B offers additional python-based software tools to perform the following functions: 1) data conversion tool between hdf5 (nxs) format to csv, 2) precise determination of detector 2θ angle from MgO or NaCl peak positions manually fitted by other software (Fityk, Wojdyr, 2010), 3) ram oil pressure to sample pressure estimation for MA6-8 and MA6-6 compression modes for a variety of commonly used assemblies (cell size and anvil truncation), 4) simultaneous estimation of pressure and temperature from the experimental lattice constants (or volume) combined with established equation of states of commonly used materials (Fig. S3). This application uses various functions of the freely available Burnman tool (Cottaar et al., 2014). Lastly, for radiography, rapid feedback on sample length changes may be needed. The use of the freely available software ImageJ (or Fiji, a plugin supported version) is highly recommended (Schindelin et al., 2012).  To date, the calibrations agree very well and remained very consistent.

Figure S4
Selected diffraction patterns obtained from NaCl using 2 Ge-SSD simultaneously a) before the start and b) at the end of deformation. The conditions of the acquisitions are as indicated.
Peak fitting was carried out using the software PDIndexer (Seto et al., 2010). The positions of the peaks (i.e. d-spacing) in the diffraction patterns, were plotted against the azimuthal positions of the detectors (0° and 90°, respectively). The expected variations of d-spacing with azimuth are also shown by the wavy colored lines calculated using Equation 1 (see text). The hydrostatic pressures are calculated from d P (hkl) (Eqn. 1).