Diamond sensor qualification for beam position and intensity monitoring by measurements at the MID instrument of the European XFEL
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Diamond and Related Materials |
| Published | 2026-02-23 |
| Authors | T. Conka-Nurdan, Wolfgang Freund, Jia Liu, Ulrike Boesenberg, Jörg Hallmann |
| DOI | 10.1016/j.diamond.2026.113480 |
Two electronic-grade single-crystal CVD diamond detectors measured pulse-resolved beam position and intensity at 2.25 MHz during qualification at the MID instrument of the European XFEL. The 20 μm and 40 μm sensors agreed with each other and with gas and imaging references. Their position uncertainty was below 1% of the 500 to 600 μm beam size.
Sensor construction and operating principle
Section titled “Sensor construction and operating principle”The detectors used 4 × 4 mm² electronic-grade single-crystal CVD diamond membranes produced by Element Six. One membrane was polished to 20 μm by Delaware Diamond Knives, and the other to 40 μm by Almax easylab. Electronic-grade material typically contains less than 5 ppb nitrogen, less than 0.03 ppb nitrogen vacancies, and less than 1 ppb boron. The authors state that this material quality can reach electron mobility up to 1700 cm²/Vs.
At 20 μm thickness, more than 90% of beam intensity passes through above 5 keV. Silicon of the same thickness reaches similar transmission only above 11 keV. This transmission result favors diamond at lower photon energies and reduces absorbed beam power. Lower absorption also reduces detector heating at XFEL beam powers.
Carbon implantation created a resistive surface layer with a sheet resistance of about 50 Ω/□. Sputtered low-atomic-number aluminum electrodes formed two parallel strips on each face. The rear pair was orthogonal to the front pair, creating a duo-lateral layout. One face supplied the horizontal coordinate, and the other supplied the vertical coordinate.
X-ray photoexcitation generated electron-hole pairs, which drifted under the applied electric field. Total intensity was proportional to the charge summed across opposing electrodes. Position was calculated from four electrode amplitudes by the center-of-gravity method. The electrode spacing was approximately 3.2 mm. Each membrane normally operated at 80 V, producing an electric field of 2 to 4 V/μm.
Readout and MID test conditions
Section titled “Readout and MID test conditions”The membrane was mounted on an alumina carrier with printed signal and bias circuits. An aluminum housing was designed for vacuum compatibility and provided four SMA sockets for bias and readout. SMA attenuators reduced the large signals generated by FEL pulses. Bias-tees separated the applied bias from each signal line.
The DAQ used commercial reartransition modules for analog shaping. Sixteen-bit ADCs then sampled the shaped signals at 108 MS/s. Each pulse train triggered the ADCs synchronously.
MID receives photons with energies from 5 to 25 keV. Each pulse contains about 10¹² photons. Up to 27,000 pulses arrive each second in 10 trains, with intra-train repetition rates up to 4.5 MHz. Each train lasts 600 μs, and the ten trains repeat at 10 Hz. The qualification result used a 2.25 MHz repetition rate. The 40 μm sensor was upstream, and the 20 μm sensor was downstream on a hexapod.
Pulse-resolved intensity agreement
Section titled “Pulse-resolved intensity agreement”At 12 keV and a 1.1 MHz repetition rate, each diamond signal was compared with absolutely calibrated XGM energy in microjoules. The reported regression explained 98% of the observed variation. The two diamond sensors also correlated with each other over the first five pulses shown in a train. The 40 μm sensor produced more ADC counts because photons traversed twice the diamond thickness.
Solid-state CVD and silicon attenuators were used to vary beam intensity. The unattenuated beam corresponded to 200 ± 14 μJ. The diamond-XGM correlation retained the same form across the attenuation configurations. This supported linearity over the tested intensity range.
At the lower XGM readings, single-shot pulse energy fluctuated more against averaged photon flux. The authors linked this behavior to reduced accuracy of the XGM fast-to-slow calibration during that run.
Beam position agreement
Section titled “Beam position agreement”Horizontal and vertical position came from center-of-gravity estimates for each X-ray pulse. The authors averaged 100 pulses per train at 12 keV, then compared the result with the 10 Hz MID imager. Every pulse in each selected train entered the mean calculation. The beam was about 500 to 600 μm across, and the reported position uncertainty was below 1% of beam size. Horizontal and vertical scans used different step sizes, so their plots showed different precision.
The 2D comparison combined the imager center of mass, single-shot diamond positions, a Gaussian beam ellipse, and each train’s diamond mean. Both membrane thicknesses agreed similarly with the imager. The authors report better position resolution than in their earlier FXE measurements. They attribute this comparison cautiously to more stable beam positions during the MID measurements.
Comparison with reference detectors
Section titled “Comparison with reference detectors”The paper reports pulse-resolved beam-position and intensity monitoring at 2.25 MHz with both sensor thicknesses. This was the first such demonstration at the European XFEL, according to the authors. The same duo-lateral structure supplied a total intensity signal and two position coordinates. Agreement with the upstream XGM and the MID imager served as the qualification result.
The XGM supplied an independent pulse-energy reference, while the imager supplied a slower position reference. The 98% explained variation and sub-1% position uncertainty are separate metrics for intensity and position. Both sensors retained MHz operation and high transmission above 5 keV. The authors regard these results as support for non-invasive hard X-ray diagnostics. Hard X-rays above 20 keV are the target because gas interaction cross sections become small.
The paper identifies beam-based alignment and intra-train position feedback as intended applications. It does not report a closed-loop feedback test, so control performance remains outside the demonstrated result.
Limits and unresolved questions
Section titled “Limits and unresolved questions”The detailed intensity regression and position comparison used 12 keV photons. The paper does not give numerical detector efficiency or a quantified radiation-dose tolerance for the tested membranes. The claimed wide dynamic range is not bounded by stated lower and upper detection limits. Low-intensity comparisons also inherit uncertainty from the XGM fast-to-slow calibration.
The authors say the better MID position resolution than at FXE may result from more stable beam positions. This is a proposed explanation rather than a controlled comparison. The effect of high atomic number platinum electrodes on wavefront and beam characteristics still needs analysis. The quadrant electrode configuration also remains to be analyzed. Neither issue was quantified in this study.
Implications for diamond sensor selection
Section titled “Implications for diamond sensor selection”Thickness controls both transmission and the charge generated in the diamond. The 20 μm membrane transmits more than 90% above 5 keV, while the 40 μm sensor generated higher ADC counts. The thinner membrane reduces absorption, while the thicker membrane provides more count at the tested conditions.
The tested position-sensitive design combined thin low-Z material, a resistive layer, patterned electrodes, and vacuum-compatible packaging. The paper’s 50 Ω/□ layer, aluminum electrodes, and 80 V bias provide a measured starting point. The tested readout handled large pulses with attenuation, shaping, synchronized 16-bit acquisition, and 108 MS/s sampling. Neither detector was tested with the high-Z platinum electrodes discussed as an open question.
For XFEL beamlines above 20 keV, the low atomic number and expected radiation hardness make single-crystal CVD diamond a candidate. That expectation is not a measured radiation-dose result. Integration into routine photon diagnostics is the authors’ proposed next step. The demonstrated result supports pulse-resolved monitoring at MHz rates and provides a baseline for feedback-system development.
Materials and support
Section titled “Materials and support”The paper’s beam-monitoring application needs a thin electronic-grade single-crystal diamond detector membrane with patterned electrodes and vacuum-compatible mounting. 6CCVD’s diamond materials catalog lists MPCVD-grown single-crystal diamond spanning 0.1 μm to 500 μm, so both tested thicknesses fall within the listed range. Although the paper used sputtered aluminum electrodes, a related build should specify polishing and electrode choices through engineering support.
View Original Abstract
X-ray Free-Electron Lasers (XFELs) such as the European XFEL (EuXFEL) generate extremely brilliant, ultra-short X-ray pulses at megahertz repetition rates. Because the pulse energy, position, and temporal propertiesfluctuate strongly from pulse to pulse, precise, non-invasive, pulse-resolved diagnostics are essential for optimalbeam delivery and advanced experiments. This work demonstrates the use of electronic-grade single-crystalCVD (scCVD) diamond detectors in a duo-lateral electrode configuration as a high-speed, radiation-hard beam-position monitor and intensity detector for hard X-rays (> 20 keV), where conventional gas detectors loseefficiency. Two diamond detectors (different thicknesses) were tested at the Materials Imaging and Dynamics(MID) instrument of the EuXFEL. They successfully performed pulse-resolved beam position and intensitymeasurements at 2.25 MHz repetition rate. The position uncertainty was < 1 % of the beam size. Excellentagreement was observed both between the two diamond sensors and with reference detectors (gas ionizationchamber and X-ray imager). These results establish duo-lateral scCVD diamond detectors as a powerful,compact, and reliable diagnostic tool for high-repetition-rate hard X-ray FELs, enabling real-time beam-basedalignment and intra-pulse-train position feedback.
Boron Doped Diamond Electrode On Silicon
View Specs →
Boron Doped Single Crystal Diamond Plate
View Specs →
Single Crystal Cvd Diamond Plate
View Specs →References we cover
Section titled “References we cover”Original source
Section titled “Original source”References
Section titled “References”- 2020 - A MHz-repetition-rate hard X-ray free-electron laser driven by a superconducting linear accelerator [Crossref]
- 2019 - X-ray photon diagnostics at the European XFEL [Crossref]
- 2019 - Operation of X-ray gas monitors at the European XFEL [Crossref]
- 2019 - Operation of photon diagnostic imagers for beam commissioning at the European XFEL [Crossref]
- 2024 - Hard X-ray operation of X-ray gas monitors at the European XFEL [Crossref]
- 2019 - Diamond power devices: state of the art, modelling, figures of merit and future perspective
- 2000 - Diamond detectors in particle physics. Reports on progress in physics [Crossref]
- 2015 - Diamond particle detectors systems in high energy physics [Crossref]
- 2016 - Impact of impurities and crystal defects on the performance of cvd diamond detectors [Crossref]