Detective Quantum Efficiency of the Timepix4 Hybrid Pixel Detector and its Application to Parallel-Beam Diffraction
Zhiyuan Ding, Nina Dimova, Jonathan S. Barnard, Giulio Crevatin, Liam O’Ryan, Richard Plackett, Daniela Bortoletto, Angus I. Kirkland, Marcus Gallagher-Jones, [10pt] Department of Materials, University of Oxford, Oxford, UK · 2026
Read Detective Quantum Efficiency of the Timepix4 Hybrid Pixel Detector and its Application to Parallel-Beam Diffraction by Zhiyuan Ding, Nina Dimova, Jonathan S. Barnard, Giulio Crevatin, Liam O’Ryan, Richard Plackett, Daniela Bortoletto, Angus I. Kirkland, Marcus Gallagher-Jones, [10pt] Department of Materials, University of Oxford, Oxford, UK (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-06857.txt.
Introduction
Ding and colleagues measure detective quantum efficiency and noise spectrum of the Timepix4 detector in event-driven electron imaging at two voltages. In raw readout mode, zero-frequency score tops a high mark at both voltages. At Nyquist frequency, efficiency stays solid at the lower voltage but falls close to zero at the higher voltage. Early parallel-beam diffraction data from a gold nanoparticle sample shows the detector can capture weak diffracted spots beyond a wide angle at the higher voltage.
Catalog record
| CREATOR | Zhiyuan Ding, Nina Dimova, Jonathan S. Barnard, Giulio Crevatin, Liam O’Ryan, Richard Plackett, Daniela Bortoletto, Angus I. Kirkland, Marcus Gallagher-Jones, [10pt] Department of Materials, University of Oxford, Oxford, UK |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 40141 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.06857
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Detective Quantum Efficiency of the Timepix4 Hybrid Pixel Detector and its Application to Parallel-Beam Diffraction Zhiyuan Ding, Nina Dimova, Jonathan S. Barnard, Giulio Crevatin, Liam O’Ryan, Richard Plackett, Daniela Bortoletto, Angus I. Kirkland, Marcus Gallagher-Jones, [10pt] Department of Materials, University of Oxford, Oxford, UK Abstract The detective quantum efficiency (DQE) and normalised noise power spectrum (NNPS) of the Timepix4 hybrid pixel detector in event-driven mode in TEM have been measured at 100 kV and 200 kV. In a raw data readout mode, the zero-frequency DQE exceeds 0.9 at both 100 kV and 200 kV. At the Nyquist frequency, the DQE remains above 0.2 at 100 kV but drops close to zero at 200 kV. Initial parallel-beam diffraction data from a polycrystalline gold nanoparticle sample is reported which shows that at 200 kV Timepix4 can detect weak diffracted information beyond a 75 mrad half-angle. 1 Introduction The Timepix4 is the newest member of the Medipix family of readout ASICs [ 1 ] . As an event-based hybrid pixel detector with 195 ps event binning [ 2 , 3 ] , it has been widely used for the detection of high energy particles [ 4 ] , X-rays [ 5 , 6 , 7 ] and photons from the visible to gamma rays [ 8 , 9 , 10 ] . In applications for transmission electron microscopy (TEM), the Timepix4 detector offers clear advantages for high timing resolution data recording. This is particularly important for data collection in various imaging geometries, including 4DSTEM [ 11 , 12 , 13 ] , tomography [ 14 , 15 ] and microED/3DED [ 16 , 17 ] , at low fluences which minimise radiation damage to the sample. For practical applications in TEM, the detector response as a function of the electron beam energy and fluence requires characterisation. Quantitative measurements including the modulation transfer function (MTF) and detective quantum efficiency (DQE) are representative metrics that characterise the detector’s response to electron hits [ 18 , 19 , 20 ] . The MTF reflects the detector’s ability to resolve information as a function of spatial frequency [ 19 ] . The DQE reflects the transfer of information above noise across different spatial frequencies [ 21 , 22 ] . Using Timepix4 for TEM, Dimova et al. [ 23 ] have previously measured the MTF at 100 kV and 200 kV acceleration voltages and improved the MTF using a clustering algorithm to reduce the effect from charge sharing. Ghosh et al. [ 24 ] have calculated the MTF and DQE at 300 kV using Monte-Carlo simulations of the electron trajectories in the sensor and discussed the influence of sensor thinning and different event threshold settings on the DQE and MTF. In this work, the normalised noise power spectrum (NNPS) and DQE of a Timepix4 detector in TEM at 100 kV and 200 kV are measured. The NNPS is obtained from flat-field illumination datasets and subsequently the DQE is calculated from the measured NNPS and MTF using a slanted knife-edge as a deterministic input signal [ 23 ] . Initial experimental data using the Timepix4 detector to record data in a parallel-beam diffraction geometry from a sample of polycrystalline gold nanoparticles is also presented as an example application. 2 DQE Measurement 2.1 Detector System The Timepix4 ASIC was bump-bonded to a 300~\mu m planar silicon sensor. The detector assembly was installed on a JEOL CryoARM Z300FSC TEM with the Merlin T4 readout system, the only commercially available readout system for the Timepix4, providing 81.92 Gbps FireFly optical readout. The per-pixel threshold for signal detection was set at 1000~ electron-hole pairs in the sensor and a reverse-bias voltage of 100~ V was applied. In event-driven mode, the maximum detection speed is approximately 3.6~ Mhits/mm 2 /s, which is limited by the transfer bandwidth [ 3 ] . The physical pixel size (pixel pitch) is 55~\mu m and number of pixels is 512\times 448 . The theoretical maximum number of events that can be detected (calculated by the maximum detection speed) is approximately 1.1\times 10^{4} events/pixel/s. If the illumination is a uniform flat-field intensity across all pixels, the theoretical maximum current that can be measured on this detector is approximately 400 pA. 2.2 Methods 2.2.1 Flat-field Data Collection and Processing Four flat-field datasets at 200 kV and five datasets at 100 kV were recorded and processed for calculation of the DQE. Each flat-field dataset was recorded with a 50 s exposure and the datasets were saved in .h5 file format as event lists. An event is a record of the energy deposition caused by an incident electron within a single pixel. Each event contains the pixel index and a timestamp [ 3 ] and an event list includes all events recorded within the exposure time on the detector. The detector was evenly divided into two regions, top and bottom, each with 256\times 448 pixels. Sixteen .h5 files stored the event lists from the top region, and another sixteen files stored event lists from the bottom region. The exposure time for a single acquisition was limited to approximately 60 s, which corresponds to the time required for a dataset to fill 512 GB of RAM on the control PC. To extend the total exposure time, multiple independent exposures with the same settings were collected, stored, and preprocessed separately before being combined for the NPS calculation. The datasets were then processed using the following steps: 1. Event lists stored in .h5 files for each dataset were merged and events in the processing area were converted into frames, each containing a target number of events ( N_{tar} ). Here ‘frames’ are defined as 2D image-like representations of the data formed by accumulating events in a pixel array. 2. Frames at the end of datasets containing fewer than the target number of events ( N_{tar} ) were discarded. 3. Frames at the same accelerating voltage from multiple datasets were combined and used to calculate the NPS spectrum and zero-frequency DQE. The target number of events per frame ( N_{tar} ) was set to 5,000,000 (the summation of each frame generated from the event list is hence also 5,000,000 ). In theory, this value should be as low as possible, since a lower value makes the NPS calculation process closer to that handling single-electron events. A lower target number of events ( N_{tar} ) also results in a smoother NPS. However, if the value is too low, the number of frames increases, leading to higher computational time and storage requirements. As also discussed later pixels along the 448-pixel side of the detector exhibit a higher response than the central area. We therefore selected a central area of 448\times 448 pixels on the detector for the NPS calculation as this gives a more faithful representation of the detector’s true performance in practical applications. The frames converted from the event list can be represented as: \displaystyle c(f,x,y),\quad\text{where:} (1) \displaystyle f\in\{1,2,\dots,N_{f}\}, \displaystyle x\in\{1,2,\dots,N_{x}\}, \displaystyle y\in\{1,2,\dots,N_{y}\}, where f is the frame index, and x,y are the pixel indices. For the datasets considered in this work, as mentioned above, N_{x}=N_{y}=448 . The characteristic properties of c(f,x,y) are described by the parameters N_{e} and d_{n} . N_{e} represents the total number