Great progress has been made in the ATLAS operation since the last EP news article. There we presented the fantastic Phase-I upgrades of the ATLAS detector installed during LS2, described the steps towards their commissioning, and summarized the experience of the first data-taking year of Run 3. What have we been up to in the ATLAS Control Room and ATLAS operations since then?

A continuing feature throughout these years has been the excitement before any startup of the LHC, coupled with high expectations of what the new data-taking year will bring (see Figure 1).

Figure 1: Impressions from the ATLAS Control Room from the years 2023 and 2024. You may find a higher-resolution version HERE.

In 2023, and continuing into 2024, emphasis has been placed on the completion of the commissioning and the full utilization of the upgraded detector components, significantly enhancing the Run-3 ATLAS Trigger system and its capabilities. Specifically, these include the higher granularity information from the Phase-I Liquid Argon digital trigger to the Phase-I Level 1 (L1) Calorimeter (L1Calo) trigger system and the Muon New Small Wheel (NSW), the two innermost stations of the Endcap Muon spectrometer that can be integrated in the L1 muon trigger. These upgrades are designed to reduce the first-level trigger rate, which limits the amount of pileup and luminosity ATLAS can sustain, by effectively suppressing background triggers while maintaining high signal efficiency. Due to the early termination of high-intensity pp data taking following the ITL8 incident in July 2023, the commissioning of the Phase-I systems with beams was halted and needed to be completed in 2024. However, during the LHC downtime, commissioning activities and preparations for the heavy-ion data-taking continued. Thanks to the upgraded systems, ATLAS achieved a rate reduction of up to 19 kHz in 2024, corresponding to 19% of the total L1 rate budget available (see Figure 2), while improving the efficiency turn-ons for the desired signal events.

In 2023, the move to the Phase-I L1Calo single electron trigger resulted in a rate reduction of 5 kHz. Since April 17, 2024, coincidences between the Big Wheel TGC, Tile calorimeter, and NSW detectors have been activated in several steps to suppress background noise. At the end of May, the TGC-NSW coincidence was fully activated for all trigger sectors, resulting in an additional rate reduction of about 14 kHz.

Figure 2: (left) ATLAS L1Calo single electron trigger efficiencies for the legacy system (red) and the Phase-I system (magenta and green) as a function of the offline electron pT. (right) ATLAS level-1 muon trigger rate of the single muon trigger (L1_MU14FCH, 14 GeV threshold) as a function of date. The rate for each run is normalized to an instantaneous luminosity of 2×10³⁴ cm⁻² s⁻¹.

While in 2022 ATLAS was limited to a pile-up of ⟨μ⟩ = 54 due to readout limitations, stable operations were achieved up to ⟨μ⟩ = 61 in 2023, thanks to the L1 rate reductions and the replacement of readout system computers during the 2022–23 YETS. In 2024, ATLAS has successfully operated at a leveling target of ⟨μ⟩ = 64, corresponding to an increase of the peak instantaneous luminosity up to 2.15×10³⁴ cm⁻² s⁻¹ as shown in Figure 3.

Figure 3: The evolution of the peak luminosity in ATLAS over the course of Run 3. Each point represents one data-taking fill.

Owing to the excellent LHC performance in 2024, we have, at the time of writing this newsletter, recorded 100 fb⁻¹ out of the 106 fb⁻¹ delivered, corresponding to a recording efficiency of 94.2%. It is interesting to see how we improve our data-taking efficiency throughout the years of a run period. As data-taking extends, the efficiency improves. At certain points, Run 3 outperformed Run 2 at the same delivered luminosity, as shown in Figure 4. Although the Run-3 recording efficiency eventually dropped below that of Run 2, it’s important to consider the context: In 2018, the luminosity continuously decayed throughout a run, while in 2024, approximately half of the fill time is spent at the highest leveled instantaneous luminosity. During the luminosity decay, the L1 rate decreases, reducing the protective deadtime for detector readout, which allows ATLAS to acquire data more efficiently.

Moreover, as time progresses, we gain operational experience and can address small inefficiencies to make our system more robust. This progress is achieved thanks to the dedication and tireless support of our experts at Point 1 and in the ATLAS institutes!

Figure 4: The cumulative recording efficiency of ATLAS as a function of the delivered luminosity during Run 2 and Run 3.

Data-taking with high-intensity beams hasn’t been our only focus. ATLAS has successfully recorded data from several special run configurations that have proven highly beneficial. These included the annual van der Meer luminosity calibration programme and the 2023 high ß* data-taking period, during which strongly de-focused beams collide at the interaction points to study elastic proton scattering. Although these special runs often require custom configurations of the detector subsystems and trigger menu, they open up unique scientific opportunities that wouldn’t be possible otherwise.

Now we’re looking forward to completing the 2024 proton–proton physics data-taking in just a few weeks and gearing up to collect data for our exciting heavy-ion program! We will share more about the past and upcoming heavy-ion data-taking program at the end of the year in another EP newsletter, so we won’t spoil it here!

As you can imagine, not everything always goes smoothly. Over the past two years, we have encountered various challenges that caused inefficiencies in our data taking. These issues ranged from hardware failures during stable beams and software bugs that had remained dormant for years, to problems due to the aging of our detector, power cuts, or failures in our rack facilities (visible in the drops of efficiency in Figure 4).

As operations experts are replaced, longtime specialists who designed or built the systems retire, and others move on, any issues that arise provide great opportunities for knowledge transfer, ensuring the continuity of a skilled community capable of effectively operating the experiment.

Finally, no matter what problems we faced, it has always been impressive to witness the collaborative spirit of our operations teams in the ATLAS control room, the satellite control rooms, and more remote sites. Always with a smile, they are committed to achieving the highest efficiency and quality data-taking for ATLAS (see Figure 5). The excellent physics that our collaboration produces with these data is our greatest reward!

Figure 5: Some impressions from the ATLAS control room during the first stable beams in 2023 and 2024. You can find a higher-resolution version HERE.