How many protons collided in ATLAS in Run 2?

The large amount of Run-2 data (collected in 2015–2018) allows the LHC experiments to probe previously unexplored rare processes, search for new physics and improve Standard Model measurements. The amount of data collected in Run 2 can be quantified by the integrated luminosity – a number which, when multiplied by the cross section for a process, yields the expected number of interactions of that type. It is a crucial figure. The uncertainty of several ATLAS Run-1 cross-section measurements, particularly of W and Z production, was dominated by systematic uncertainty on the integrated luminosity. To minimise this, ATLAS performs precise absolute and relative calibrations of several luminosity-sensitive detector systems in a three-step procedure.

The first step is an absolute calibration of the luminosity using a van-der-Meer beam-separation scan under specialised beam conditions. By displacing the beams horizontally and vertically and scanning them through each other, it is possible to measure the combined size of the colliding proton bunches. Determining in addition the total number of protons in each colliding bunch from the measurement of the beam currents, the absolute luminosity of each colliding bunch pair can be derived. Relating this to the mean number of interactions observed in the LUCID-2 detector – a set of photomultiplier tubes located 17 m in either direction along the beam pipe that detect the Cherenkov light of particles which come from the interaction – the scale for the absolute luminosity measurement of LUCID-2 is set.

The second step is to extrapolate this calibration to LHC physics conditions, where the number of interactions increases from fewer than one to around 20–50 interactions per crossing, and the pattern of proton bunches changes from isolated bunches to trains of consecutive bunches with 25 ns spacing. The LUCID-2 response is sensitive to these differences. It is corrected with the help of a track counting algorithm, which relates the number of interactions to the number of tracks reconstructed in ATLAS’s inner detector.

The final step is to monitor the stability of the LUCID-2 calibration over time. This is evaluated by comparing the luminosity estimate of LUCID-2 to those from track counting in the inner detector and various ATLAS calorimeters over the course of the data-taking year (figure 1). The agreement between detectors quantifies the stability of the LUCID-2 response.

Using this three-step method and taking into account correlations between years, ATLAS has obtained a preliminary uncertainty on the luminosity estimate for the combined Run-2 data of 1.7%, improving slightly on the Run-1 precisions of 1.8% at 7 TeV and 1.9% at 8 TeV. The full 13 TeV Run-2 data sample corresponds to an integrated luminosity of 139 fb^–1 – about 1.1 × 10¹⁶ proton collisions.