Real-time electron clustering in an event-driven hybrid pixel detector

Area of Science:

• Physics
• Materials Science
• Instrumentation

Background:

• Event-driven hybrid pixel detectors offer nanosecond time resolution for advanced electron microscopy.
• Accurate single-electron event detection is challenging due to multi-pixel excitation by impinging electrons.

Purpose of the Study:

• To develop a robust and fast clustering algorithm for real-time processing of raw detector data.
• To enable efficient conversion of multi-pixel hits into single-electron events for ultrafast electron microscopy.

Main Methods:

• A novel clustering algorithm compares detector hits (position and time) against a buffer of previous events.
• The algorithm determines event mergers based on the probability of relevance, ensuring computational efficiency.
• The algorithm was tested using a 'timepix' camera in continuous and femtosecond laser-triggered electron microscopy modes.

Main Results:

• The developed algorithm achieves real-time processing of continuous data streams.
• Computation time is independent of electron arrival density, preventing operational bottlenecks.
• Successful demonstration in both continuous beam and femtosecond laser-triggered electron microscopy regimes.

Conclusions:

• The presented clustering algorithm is a robust solution for real-time single-electron event identification in ultrafast electron microscopy.
• This facilitates advanced applications such as hyperspectral electron-energy loss spectroscopy and free-electron quantum optics.
• The algorithm's efficiency and independence from data density ensure its broad applicability in high-speed electron detection.