Simulation study of the impact of AGIPD design choices on X-ray Photon Correlation Spectroscopy utilizing the intensity autocorrelation technique

TL;DR

This study uses simulations to evaluate how different AGIPD detector design choices affect the quality of X-ray Photon Correlation Spectroscopy data, focusing on pixel size, aperturing, and noise effects at XFEL.

Contribution

It provides a comparative analysis of detector pixel sizes and aperturing strategies, revealing optimal configurations for XPCS experiments at XFELs.

Findings

Aperturing is not beneficial at low intensities but improves data quality at higher intensities.

Smaller pixels (100 μm) outperform larger ones (200 μm) when intensity exceeds 0.05 photons per (100 μm)^2.

Detector design influences signal-to-noise ratio and correlation accuracy, with differences less than a factor of 3.

Abstract

The European XFEL, currently under construction, will produce a coherent X-ray pulse every 222 ns in pulse trains of up to 2700 pulses. In conjunction with the fast 2D area detectors currently under development, it will be possible to perform X-ray Photon Correlation Spectroscopy (XPCS) experiments on sub-microsecond timescales with non-ergodic systems. A case study for the Adaptive Gain Integrating Pixel Detector (AGIPD) at the European XFEL employing the intensity autocorrelation technique was performed using the detector simulation tool HORUS. As optimum results from XPCS experiments are obtained when the pixel size approximates the (small) speckle size, the presented study compares the AGIPD (pixel size of (200 $\upmu$m)$^2$) to a possible apertured version of the detector and to a hypothetical system with (100 $\upmu$m)$^2$ pixel size and investigates the influence of intensity fluctuations and incoherent noise on the quality of the acquired data. The intuitive conclusion that aperturing is not beneficial as data is 'thrown away' was proven to be correct for low intensities. For intensities larger than approximately 1 photon per (100 $\upmu$m)$^2$ aperturing was found to be beneficial, as charge sharing effects were excluded by it. It was shown that for the investigated case (100 $\upmu$m)$^2$ pixels produced significantly better results than (200 $\upmu$m)$^2$ pixels when the average intensity exceeded approximately 0.05 photons per (100 $\upmu$m)$^2$. Although the systems were quite different in design they varied in the signal to noise ratio only by a factor of 2-3, and even less in the relative error of the extracted correlation constants. However the dependence on intensity showed distinctively different features for the different systems.