Physics - Implementing Betatron Radiation For Beam Diagnostics Studies

Supervisor: Dr Debdeep Ghosal

Supervisor bio: 

Being an Experimental Physicist, I am an empathic and flexible person with a solid scientific background in Science, 9+ years of international experience, a creative attitude to problem-solving, and with an immense desire to make tomorrow's world a more sustainable one by novel approaches. Innovative, analytical, collaborative experimental physicist with proven facility expertise.

Understanding nature and matter with all of their phenomena has always spurred my curiosity. Experienced in scientific writing and diverse research projects related to Nuclear and Particle Physics (Hadron physics to be more precise).  Highly motivated with hardware, analytical, conceptual and interpersonal skills, ready to have a positive and proactive attitude as a collaborative team member.

So far, my Research & Development experience cover different fields of Experimental Physics: Nuclear-Particle Physics(Detector and analysis), Optics, Accelerator Physics, Beam diagnostic, Plasma Physics(Modelling & Characterization). It would be great and highly appreciable if my contributions helps to solve daily-life problems of mankind directly or indirectly.Being an Experimental Physicist, I am an empathic and flexible person with a solid scientific background in Science, 9+ years of international experience, a creative attitude to problem-solving, and with an immense desire to make tomorrow's world a more sustainable one by novel approaches. Innovative, analytical, collaborative experimental physicist with proven facility expertise. Understanding nature and matter with all of their phenomena has always spurred my curiosity. Experienced in scientific writing and diverse research projects related to Nuclear and Particle Physics (Hadron physics to be more precise). Highly motivated with hardware, analytical, conceptual and interpersonal skills, ready to have a positive and proactive attitude as a collaborative team member. So far, my Research & Development experience cover different fields of Experimental Physics: Nuclear-Particle Physics(Detector and analysis), Optics, Accelerator Physics, Beam diagnostic, Plasma Physics(Modelling & Characterization). It would be great and highly appreciable if my contributions helps to solve daily-life problems of mankind directly or indirectly.

Email: dghosal@liverpool.ac.uk

School: Physical Sciences

Department: Physics

Module code: PHYS001

Suitable for students of: Accelerator Physics, Particle Physics, Computational Physics, Radiation detection Study, Plasma Physics, Data analysis

Desired experience or requirements: 

· Basic understanding of accelerator physics and good knowledge of electromagnetism and interaction of plasma with matter

· Some experience with C++ or Python would be useful

· Strong interest to learn machine learning (prior knowledge is even better) and particle-in-cell codes

Places available: 2

Start dates: Session 1 (15th June 2026)

Project length: 4 or 8 weeks

Virtual option: Yes

Hybrid option: Yes

Project description:

Novel wakefield accelerators are a plasma-based compact, high-gradient particle accelerator which use a powerful laser or particle beam to create electric field-wakes in a plasma, which then accelerates a trailing beam of particles to very high energies. This technology offers a more compact and powerful alternative to conventional accelerators, with the potential for similar capabilities in a much smaller footprint.
As relativistic particles like- electrons travel through plasma in such wakefield accelerators, they experience strong transverse focusing forces, causing them to oscillate and emit betatron radiation. This radiation inherently encodes information about key beam characteristics, including emittance and transverse size that are essential for quality check and monitoring. 

Leveraging this phenomenon, this study aims to propose a machine learning framework, trained on synthetic datasets generated via advanced simulation tools, to infer such beam properties. Maximum Likelihood Estimation (MLE) technique was previously used as one of the attempted methods of extracting this rich information about beam parameters from measurements of BR, while machine learning (ML) approaches were applied to improve the accuracy of these measurements.

The dataset will include input variables like beam energy and charge, while outputs capture diagnostic quantities e.g. beam emittance and spot size derived from the simulated radiation profiles. Different ML models are aimed to be benchmarked for predictive accuracy, highlighting their reliability as fast, physics-informed surrogate predictors. This approach underscores the growing potential of integrating simulation-driven data with ML to enhance diagnostic capabilities in future plasma-based accelerator research.

Additional requirements: N/A