Particle physics

The LHCb experiment at the Large Hadron Collider at CERN aims to study the forces of nature - electromagnetic, strong and weak - at the smallest scales and highest energies. Especially in the early universe, right after the big bang, new particles and forces could have had a dramatic impact on the way the universe looks and behaves today. At LHCb, we specialize in measuring the matter-antimatter differences in heavy particles, called CP-violation, and making precision studies of the effects that potential new particles or forces have in very rare particle decays, through so-called quantum loops.

In Maastricht, we contribute to the physics analyses, the operation of the detector at the LHC, and to the algorithms used to reconstruct physics from our data. To face the future challenges of the large amounts of data produced by the particle collisions, we study the applications of machine learning, the use of graphics processing units in high-performance computing [and explore the potential that quantum computing can offer].

Staff

Jacco de Vries

Christopher Pawley

Prof. dr. Marcel Merk

Kurt Driessens

Daniel Hugo Cámpora Pérez

Research topics

CP violation

CP Violation is the study of differences between matter and antimatter. At the big bang, we believe an equal amount of both must have been created, but today we only find evidence of matter in our universe. Some asymmetry must therefore occur in the laws of physics at higher energies. We study this at LHCb by creating high-energy collisions, and measuring the rates and interactions of matter and antimatter. This way, we hope to learn about the forces at play in the early universe, and why our universe looks the way it does, today.

Rare decays

Rare particle decays are an excellent way to study the possible existence of new, fundamental particles. Such new particles can have an imprint on properties such as decay rates of rare processes, through so-called ‘quantum loops’. LHCb is the world-leading expert in the search for new particles using rare decays of beauty baryons.

Charm physics

Charm physics: LHCb has collected worlds’ largest samples of particles containing a charm quark. By studying the production and asymmetries of these particles in the high-energy collisions at the LHC, we can learn something about the strong nuclear force that binds these quarks together.

Pattern recognition

Pattern recognition: In order to do physics, the raw data of the LHCb detector needs to be reconstructed: finding true particle tracks from individual readout ‘hits’. As this is run in real-time during the particle collisions (~40 million per second), this falls into the area of high-performance computing. At UM, we are deeply involved with these algorithms, and studying increased performance by using new architectures such as graphical processing units or quantum computers.

Machine learning

Machine learning: Artificial Intelligence or Machine Learning is a great help in extracting useful information from the large amounts of data generated by the particle collisions at the LHC. At MSP, we study the use of various of these tools: deep neural networks, convolutional networks, anomaly detection, long-short term memory, and more. These studies are performed in collaboration with experts at DKE.

Publications

Bs2mumu combination with CMS: https://www.nature.com/articles/nature14474
Measurement of CP averaged observables in the B^0->K*\mu^+\mu^- decay: https://arxiv.org/abs/2003.04831
Measurement of CP asymmetry in Bs->DsK decays: JHEP03 (2018) 059, https://doi.org/10.1007/JHEP03(2018)059
Allen: https://www.researchgate.net/publication/338065027_Allen_A_high_level_trigger_on_GPUs_for_LHCb
asls: https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.117.061803
outer tracker performance: https://iopscience.iop.org/article/10.1088/1748-0221/12/11/P11016
Ds production asymmetry: https://link.springer.com/article/10.1007%2FJHEP08%282018%29008

Student participation

Improving electron reconstruction with machine learning

One of the most interesting hints for new particles comes from the rare decays including electrons. Due to their light mass, however, these electrons lose a lot of energy in our detector, in the form of photons. This study aims to write a machine learning algorithm that finds these photons, and results in a better reconstruction of the electron energy.

Student: Felix Quinque

Production rates of Xic baryons

Here we perform the full analysis of the production rate of Xic charm baryons at the high-energy collisions at the LHC. The resulting paper will be used by theorists to improve their models of the strong interaction.

Students: Simon Calo, Jonas Tjepkema, honours project

Anomaly detection in the early universe

In this study we explored the usage of anomaly detection algorithms to generically search for new, unexpected phenomena in the LHCb data.

Student: Melissa Lopez

Improving particle identification with neural networks

In this study we explored the addition of a new variable, the “time-of-flight”, to the neural network used for the identification of particles in LHCb.

Students: Lizzy Rieth, Nelson Waakop Reijers

Data selection strategies for CP-violation at the upgraded LHCb detector (2021)

In this study, the trigger strategy for the very important Bs -> j/psi phi decays in 2021 were studied. The inclusion of beauty hadron identification (aka ‘flavour tagging’) parameters in the trigger can significantly enlarge the signal data rates for the near-future data taking at the upgraded LHCb detector.

Student: Ali Sidley

Modelling the mass shape of charmed particles

In this study, students analysed recent data from the LHCb experiment that include charmed baryons. Various models were built and fitted to the invariant mass distributions in data, to study the effects of detector resolution and signal to background ratio in these reconstructed particles.

Students: Joyce Cobussen, Jord Muffels, Diyon Wickremeratne, Francesco Marangio

Projects

Student stories

I participated in a 3-week project where our group of 6 people researched whether it is possible to use machine learning to improve the identification of particles in the Large Hadron Collider (LHC) in CERN. I am very happy I got to do this project not only because I got to apply some of the knowledge I gained in courses, but also because it presented me with the opportunity to learn many new skills.

Machine learning in itself was completely new to me, but I also learned a lot about programming in Python, handling big sets of data and how measurements are done in the LHC. Aside from this, the project held real relevance to the research done in CERN. Our supervisor is a researcher there and we got to develop an idea he had to try and improve the particle identification methods currently in place. Under his guidance, we were able to produce useful results enabling us to help him in his research.

Nelson Waakop Reijers, 3rd-year MSP student

If you wonder what practical applications could come out of this, there is no straightforward answer. Estimating the direct impact that our research could have on society is very tricky and somewhat unpredictable, but that is often the case with fundamental research in physics. After all, who could have foreseen the impact of Einstein’s theory of General relativity back in the 1920s? And yet, without it, we would not have functioning GPS systems, wireless internet and so much more. That is exactly the most exciting aspect of fundamental research: we are building the basis for the technological revolution of the future!

For me, this project has been an extremely eye-opening experience. It has given me a new perspective on scientific research: being a structured and persistent endeavour rather than a series of serendipitous discoveries. This has reinforced in me the idea that passion and dedication are the real cores of science and it has motivated me throughout the rest of my bachelor studies, including my thesis. It has made me feel like I was part of the real scientific world, already as a bachelor student.

Simon Calo, 3rd-year MSP student