ALICE Data Analysis

The quark-gluon plasma (QGP), which is expected to be created in heavy-ion collisions, only exists for a very short time before it expands and cools down. Therefore, the only observables of the QGP state in heavy-ion collisions are the energy and the momentum of the particles created in the collision, which are measured by the detector.
Hence, our goal is to measure tracks in the detector and make a physics analysis with them!

In our group we study the production of open heavy-flavour hadrons containing a heavy quark (charm or beauty), which serve as unique probes of the QGP properties. Due to their mass, which is large compared to the QCD scale factor, heavy quarks are produced in the initial hard scattering processes with large momentum tranfer on a small generation time scale, which is smaller than the production time of the QGP. This leads to the fact that they experience the whole evolution of the QGP, interacting strongly with the medium and losing energy while traversing it. All these characteristics make heavy quarks unique probes to study and characterise the QGP, which is the main objective of our work.
An important mechanism, which is still not fully understood and which we address with our studies, is the process of hadronisation (the formation of hadrons from quarks and gluons) in the presence of a QGP and without.

A major part of the physics analysis of ALICE data is the reconstruction of the short-lived heavy-flavour particles from their decay products, which are measured as tracks in the detector. To reduce the large combinatorial background we exploit an advanced analysis technique which is based on machine learning techniques. The reconstructed particle candidates are classified using Boosted Decision Tree models for the binary or multiclass classification of signal and background. The machine learning models are optimised using different sets of features describing the topology of the particle decay or the particle identity of the measured daughter tracks. This analysis technique provides an advanced and more secure alternative for the selection of reconstructed particle candidates.
In order to extract the interesting physics from the data, we calculate the cross sections of these charm baryons and mesons.

Schematic: Decay of the Lambda-c baryon.
The charged, long-lived decay products are measured as tracks by the ALICE detector and are used to reconstruct the short-lived mother particle.

Mass peak of reconstructed low momentum Lambda-c baryon.
The invariant mass spectrum is calculated from the selected tracks and the short-lived Lambda-c baryon is reconstructed.

Tools for Data Analysis

For the data analysis procedure we are using the software ROOT which was developed at CERN for particle physics data analysis.
Furthermore, we utilise machine learning algorithms in order to optimise the signal extraction in our data analyses. The minimal heavy ion physics environment for Machine Learning (hipe4ML) together with XGBoost is used for the binary or multiclass classification.

Further Information on Heavy-Flavour Analysis in ALICE

If you want to get a better idea of open-heavy-flavour hadron analysis in ALICE you can read into the following two papers:

Charm baryon production in pp, p-Pb and Pb-Pb collisions with ALICE at the LHC. archiv link

First measurement of Xic0 production in pp collisions at sqrt(s) = 7TeV. archiv link

Here you can also find all concluded analysis theses in our group. Individual theses can be used as sources of basic information and to get a general idea of our analysis work and its methods.

Contact Persons in the Group

Carolina Reetz (PhD student)

Contact at GSI: Andrea Dubla