Calibration of luminometers at Compact Muon Solenoid experiment

Abstract

Measuring luminosity, a quantity related to the amount of collected data, is essential in high energy physics experiments. It is used in measuring the absolute cross section of physical processes and in searches for new physics. In order to achieve precise results at the Large Hadron Collider (LHC) at the European Organisation for Nuclear Research (CERN), luminosity must be measured with high precision, which motivates studies to minimise all systematic effects that dominate the uncertainty of this measurement. The state-of-the-art studies are carried out using luminometers, which record a fraction of the number of collisions taking place at the LHC experiments. The rate recorded by these devices has to be normalised through a factor given by the visible cross section σ vis , to provide an estimate for the absolute luminosity. The Compact Muon Solenoid experiment (CMS) contains several luminometers capable of producing a real-time estimate of luminosity per colliding particle bunch. Three luminometers have been considered in this work: the Fast Beam Condition Monitor (BCM1F), the Pixel Luminosity Telescope (PLT), and the Forward Hadron Calorimeter (HF). The BCM1F and PLT detectors are operated entirely by the BRIL group, which also provides the software used to derive the calibration of the rates for the luminosity measurement.

The established calibration method used at LHC requires a specific scanning procedure, where the two particle beams are moved over each other producing a way to determine the beam overlap width, which in turn can be used to estimate σ vis of a luminometer. This scanning procedure is called a van der Meer scan. The methods used to measure this calibration factor and the luminosity during the commissioning period for Run 3 are discussed.

The software used for analysing van der Meer scans is called the vdM framework. It is capable of rapidly estimating the σ vis after a vdM scan is performed, and also applying a set of sophisticated corrections in the offline analysis to mitigate systematic effects and minimise the total uncertainty of the measurement. It has been updated throughout the long shutdown 2, a maintenance period between Run 2 (years 2015 – 2018) and Run 3 (from 2022 on) data taking periods. The most recent updates that have been implemented as a part of this thesis are described in detail and an overview of the current features of the program are presented. Recent improvements include establishing an automated testing procedure, automated figures, and a centralised versioning & documentation system implemented in GitLab. These new features upgrade the software tool to render it faster and simple to use.

A case study to correct for a source of inaccuracy of the σ vis measurement caused by time evolution of the beam overlap of colliding particle beams is considered. The data collected during the fill 6868 from Run 2 is studied, and it is found that the convoluted beam widths as well as the peak detector rate evolve in a linear fashion over time. This finding suggests that a multiplicative factor should be applied, in order to correct the effect of this phenomenon. The impact of this factor has been found to vary between 0.0 to 0.2% in fill 6868, and has to be taken into account considering the luminosity measurement aims to achieve an inaccuracy of no more than 2%.