Diamond Annual Review 2020/21

122 123 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 2 0 / 2 1 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 2 0 / 2 1 Detector Group Nicola Tartoni, Detector Group Leader L ast year sawthe Detector Group focusing on the scientific exploitation of the time-resolvedTristan10M, an innovative detector system developed entirely by the group. The Tristan project was recently completed with the delivery of a Tristan10M system to Diamond’s Small-Molecule Single-Crystal Diffraction beamline (I19) and commissioned with a time-resolved experiment in March 2021. The group’s other major development project is the Arc-detector, a cadmium telluride (CdTe) photon-counting detector for Diamond’s X-ray Pair Distribution Function (XPDF) beamline (I15-1). The Tristan10M detector The Tristan10M detector is a photon counting hybrid pixel area detector based on the read-out ASIC (application specific integrated circuit) Timepix3 1 , a photon counting ASIC with the format of a square matrix of 256 x 256 pixels with 55 μm pixel pitch. Sixteen Timepix3 ASICs are flip-chip bump bonded to monolithic silicon sensors 500 μm thick. Ten modules arranged in a matrix of five rows and two columns make up the detector (Fig. 1). Unlike conventional photon counting ASICs, Timepix3 places a time stamp to each event detected and immediately sends the information about the time of arrival and location of the event to the data bus. The notional accuracy of the time stamp is 1.56 ns and the measured accuracy is down to 8 ns r.m.s. 2 . This mode of operation opens the possibility of achieving unprecedented time resolution without gating the detector, which results in a much better overall detection efficiency. Because ofTimepix3, the data produced byTristan10M consists of a stream of events instead of an array of images.The data are recorded as a list of events in an HDF5 file. To obtain images from the list of events, appropriate post- processing codes were developed. This gives the scientist great flexibility in the choice of the time slicing, meaning they can choose the most appropriate time intervals between images. The commissioning at I19 EH2 on the Newport 4-circle diffractometer used a wavelength of 0.4859 Å (25.25 keV). The Tristam10M detector gives a resolution of >0.6 Å at a detector distance of 111 mm. The beam centre and detector distance was quickly calculated from analysis of LaB6 powder pattern standard (Fig. 2). The powder also showed good alignment of the modules. A single crystal rotation dataset was obtained for [NiDppeCl2] at RT 300 K. The event data was then converted into a series of 0.1 deg images and processed using DIALS, a data analysis package for Macromolecular Crystallography (MX). The resulting structure has good merging and refinement statistics. The Tristan M10 detector was then tested in a time-resolved experiment by monitoring the laser induced heating recovery of platinum (Pt) powder at various repetition rates of a pulsed laser (100 Hz to 50 kHz). During the final stages of the beam test, a Europium complex (organic electroluminescent device) was studied using a full time-resolved single crystal rotation procedure. The setup was pushed to its limits with the ultra-long four-hour data collection. The data analysis is ongoing and the results will be published at future conferences. The way the data are stored and processed in Tristan inspired a novel use of the Xspress4 pulse processor 3 for quick Extended X-ray Absorption Fine Structure (EXAFS) experiments. The quick EXAFS technique usually uses germanium (Ge) multi-element fluorescence detectors and takes frames (multi-channel analyser (MCA) spectra sets) in quick succession, up to 1 kHz in the current generation of experiments and up to 20 kHz or more in the future upgrades.The current pulse processors associated with spectroscopy detectors calculate the MCA spectra in real time; however, it becomes more and more difficult to allocate the events to the right time slot as the frame rate increases. This is due to the unpredictable latency of the event processing when building the MCA spectra that results in time jitter of the events. A mode of operation known as list mode was developed by using the time stamp capability of Xspress4. In this mode of operation, a record for each event, which consists of the time stamp, the detector channel, and the photon energy of the event, is sent to the data storage. The data structure is then a list of events rather similar to the Tristan data structure. This enables the reconstruction of MCAs as a function of time with the desired accuracy. External trigger events are also time stamped and recorded and provide the reference to the position of the monochromator or other events in the sample environment. A first test of this mode of operation proved effective and it could become the standard mode of operation for quick EXAFS experiments at Diamond. The Arc-detector The development of the Arc-detector has been progressing at a good pace, recently meeting a few important milestones: the CdTe sensors have been hybridised and tested and the front-end electronics are finished. Atotalof39CdTesensorswerehybridised.Eachofthemwasbump-bonded to three Medipix3RX ASICs, then attached and wire-bonded to the sensor carrier board. The hybridisation was performed in different batches so that process performance could be fed back to the manufacturers to allow process improvement if necessary.The CdTe hybrids were tested and characterisedwith X-rays to evaluate the image performance, leakage current and to debug the sensor carrier PC board, which had been developed the previous year. These tests were carried out using the Merlin readout system 4 and a readout PC board, which was purposefully developed for the ARC detector project. Simultaneously, the ARC detector readout boards were designed and tested. A Merlin-to-fibre optic interface board was also designed to initially test and debug the readout board and attached sensors, prior to full back-end firmware development. The Merlin-to-fibre interface board may also benefit other beamlines using the Merlin system, as it could allow the data acquisition system to migrate from a 68-way very-high-density copper cable interconnect to fibre optic, which extends physical separation possibilities and in due course, may allow increased frame rate. The development of the mechanics took place in parallel; the super- modules have been designed and delivered. Figure 3 shows one of the super- modules populated with four CdTe sensors, two readout boards and all the connections in place (chiller pipes, high voltage bias, low voltage power and fibre optic cables), which are currently being tested in the Detector Group laboratories. Nextstepsarefocussedon:developmentoftheback-endfirmwaretargeted at two FEM II boards where work is in progress; the delivery of a new batch of 30 CdTe sensors and associated sensor carrier boards, in order to hybridise more modules; and the design of back-end mechanics and development of the associated boards (optical fanout and midplane printed circuit boards (PCBs)) that will be required to complete the data acquisition system. References: 1. Poikela T. et al. Timepix3: A 65K channel hybrid pixel readout chip with simultaneous ToA/ToT and sparse readout. J. Instrum. 9 , C05013–C05013 (2014). DOI: 10.1088/1748-0221/9/05/C05013 2. Yousef H. et al . Proceedings, 14th Vienna Conference on Instrumentation (VCI 2016): Vienna, Austria, February 15-19, 2016. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 845 , pp. 639 – 643 (2017). 3. Dennis G. et al. First results using the new DLS Xspress4 digital pulse processor with monolithic segmented HPGe detectors on XAS beamlines. AIP Conf. Proc. 2054 , 60065 (2019). DOI: 10.1063/1.5084696 4. Plackett R. et al. Merlin: A fast versatile readout system for Medipix3. J. Instrum. 8 , C01038–C01038 (2013). DOI: 10.1088/1748-0221/8/01/ C01038 Figure 1: The assembly of Tristan10M has just been completed in the Detector lab and the detector is ready to be delivered to I19. The photograph shows the 10 modules of the detector Tristan10Mmounted on the detector chassis. Figure 2: LaB6 powder diffraction pattern measured by Tristan10M at I19 for calibration purposes during the detector commissioning. The image has been reconstructed in the post processing from the list of events. Figure 3: A super-module with the four sensor modules and the two readout boards (on the right-hand side) mounted on the mechanical support. Each CdTe sensor is flip chip bump-bonded to three Medipix3RX ASICs. X-ray Technologies