Current and future challenges when operating the ATLAS Liquid Argon Calorimeter

The Liquid Argon (LAr) Calorimeters are employed by ATLAS for precision electromagnetic calorimetry and for hadronic and forward calorimetry. They also provide inputs to the first level of the ATLAS trigger system. The LAr Calorimeters have been operated during regular ATLAS data taking since 2008. All critical parameters, like LAr purity and temperature, have been kept at stable values resulting in reliable signal pulses. In 2022, the LHC collider has increased the luminosity leading to a pile-up of up to 70 interactions per bunch crossing. In the high luminosity phase, HL-LHC, planned to be fully functional in 2029, the pile-up will increase up to 200 interactions per bunch crossing. This contribution will highlight the present and future challenges in the operation of such a detector.


Introduction
The ATLAS detector [1] is one of two general multi-purpose detectors at the LHC.ATLAS started taking data from proton-proton collisions in 2010 at a center-of-mass energy of 7 TeV.The centerof-mass energy was increased in several steps to 13.6 TeV, that was reached during the currently ongoing Run 3 that had started in 2022.

The LHC and the ATLAS experiment
The cumulative luminosity delivered by the LHC to ATLAS from 2011 to 2023 is displayed in figure 1.In proton-proton collisions, ATLAS recorded 5.08 fb −1 at a center-of-mass energy, √ , of 7 TeV in 2011, 21.3 fb −1 at 8 TeV in 2012, 147 fb −1 at 13 TeV in 2015-18 and 66 fb −1 at 13. 6 TeV in 2022-23 [2].A bunch spacing of 50 ns was used from 2011-12 and changed to 25 ns from 2015 on.The high pile-up environment, visualized by the mean number of interactions per LHC bunch crossing, ⟨⟩, and shown in figure 2 for the years 2011-2023, and the high peak luminosity of up to 23.9 × 10 33 cm −2 s −1 create extremely challenging conditions for the experiments.In spite of this challenging environment, ATLAS announced the discovery of the Higgs boson in July 2012.The Higgs boson was the last missing piece of the Standard Model of Particle Physics and the corresponding Higgs field is giving mass to the elementary particles.An essential role was played by the Liquid Argon (LAr) Calorimeter which provided precise measurements of photons, electrons and jets.

The Liquid Argon Calorimeter design principles
The ATLAS calorimeters are sampling calorimeters, which means that the material producing the particle shower is distinct from the active material measuring the deposited energy.Liquid argon was chosen as the active material due to its radiation hardness and fast and uniform response.The design of the calorimeter was largely motivated by the needs of the searches for Higgs boson decays in final states containing photons, electrons, jets and missing transverse energy.
For reaching the required resolution for Higgs searches, the sampling term of the energy resolution of the electromagnetic calorimeter needs to be less than 10 %/ √︁ /GeV.Additionally, the constant term, which is dominating the calorimeter resolution at high energy, has to be below 0.7 %.Both requirements have been verified to be fulfilled in test beam studies [1].
There is one cryostat containing the EMB and one cryostat for each end-cap hosting the EMEC, HEC and FCal.Both EMB and EMEC use an accordion design, with lead as the absorbing material and copper and kapton electrodes.This geometry ensures a full azimuthal coverage with no cracks.The calorimeter is further segmented into three longitudinal layers for || < 2.5 (two coarser-granularity layers for || > 2.5) and a presampler up to || < 1.8, providing an estimate of the energy loss upstream of the calorimeter.The HEC is separated into four layers constructed in a more conventional parallel-plate geometry.It uses copper as absorbers and copper and kapton electrodes.The FCal is 1The pseudorapidity is defined as  = − ln(tan(/2)) with  the angle between particle momentum and beam axis.
-2 -split into three modules which are sitting in a metal absorber matrix and use electrode rods parallel to the beam pipe.To cope with the higher particle flux in the forward regions, the LAr gaps are smaller than in the rest of the LAr detector.The first module has copper absorbers and is used for electromagnetic measurements.The other two modules contain tungsten absorbers and are used for hadronic measurements.
Figure 3.The Liquid Argon Calorimeter subsystems: the hadronic end-cap with its parallel plate design, the electromagnetic end-caps and barrel with an accordion geometry design and the forward calorimeter with its rod matrix structure [1].

The Liquid Argon Calorimeter readout
When a charged particle from a calorimeter shower crosses a LAr gap, it ionizes the argon.The electrons drift in the electric field (around 1 kV/mm) towards the electrodes.The resulting ionization current has a fast rise time (less than 1 ns) and a slow linear decay while the charges are collected.This creates a triangular signal which is sent to the front-end electronics to be pre-amplified, shaped and digitized at 40 MHz.Both pulses are illustrated in figure 4. From there, the digitized samples (typically 4) are sent to the off-detector electronics at a frequency of about 100 kHz.Readout drivers compute the cell energy and pulse time.In total, the LAr Calorimeters comprise 182,468 readout channels [3].
The HEC has a special setup as it is the only calorimeter with active cold electronics.The boards are directly mounted on the outside of each HEC wheel within the cryostat.Each of them consists of 16 GaAs chips, containing eight preamplifiers and two summing amplifiers each [4].After the first six years of operation only five dead channels were found (< 0.1 %.)Additional test have shown that the HEC cold electronics are robust enough to withstand the more challenging conditions of the High Luminosity LHC (HL-LHC) and no exchange is needed [5].

Liquid argon purity
Impurities in the LAr such as O 2 can capture electrons and thus degrade the signal.To measure the impurities, 30 purity monitors are immersed in the liquid argon.The measurement results for Run 1 and Run 2 are shown in figure 5 [6].Values in the range of 200 ppb are obtained, which is well below the impurity requirement of < 1000 ppb [7].

Liquid argon temperature
Changes in the LAr temperature have been measured to have a −2 %/K influence on the calorimeter signal.This arises from density (−0.45 %/K) and drift velocity variations (−1.55 %/K).To keep the contribution to the calorimeter energy resolution constant term negligible, a stability and uniformity of below 100 mK is required.A temperature measurement system with 192 probes in the barrel and 158 probes in each end-cap cryostat has been installed to constantly monitor the status [6].
The temperature of the LAr for EMB and EMEC in the years 2010-12 is shown in figure 6.The stability of the average temperature in 2012 has been measured to be well below the 100 mK requirement.An RMS of 2 mK has been observed for the barrel and 5 mK for the end-caps.
The temperature uniformity measurement for the barrel is displayed in figure 7. The RMS of the difference of the average temperature of one sensor in a given year to the average temperature of the whole cryostat was measured to be 61 mK.Similar results were obtained for the end-caps, where 50 mK and 55 mK were observed.All values are well below the 100 mK requirement.

The Liquid Argon Calorimeter upgrade
Already in 2012, the LHC was running close to design luminosity.It was exceeded during the following Run 2 and will be pushed even further for future runs.To be able to cope with these more challenging conditions, the readout of the LAr detector has to undergo several upgrades.The Phase-I upgrade was installed during the LS2 (long shutdown 2) in 2019-22.The Phase-II upgrade will be installed during LS3 in 2026-28 before the start of the HL-LHC.

Phase-I upgrade
During the Phase-I upgrade a new digital trigger path was installed to replace the analog trigger and enhance the trigger performance.The new trigger has an up to factor 10 finer granularity, which helps to obtain a better trigger energy resolution and a higher efficiency in selecting physics objects.To achieve this, new electronics have been installed both on the front-end and off-detector side [8].The -4 -  performance comparison of the legacy analog level-1 trigger and the new digital Phase-I trigger can be found in figure 8 and figure 9 for the single electron trigger.The Phase-I trigger shows a sharper turn-on curve and a lower rate, i.e. a better background rejection.

Phase-II upgrade
For the HL-LHC it is expected that up to 7 times the design luminosity of the LHC will be reached and up to 200 collisions will happen simultaneously per bunch crossing.The LAr calorimeters themselves are expected to operate reliably in these conditions.However, the readout requires a complete replacement to cope with an increased trigger rate of 1 MHz, higher pile-up and radiation exposure.New front-end boards (FEBs) will be installed to provide digitization of all signals at a rate of 40 MHz for improved trigger decisions.This requires the modification of the back-end electronics as -5 -well.It will help to drastically improve trigger bandwidth and latency [10].The application of neural networks are expected to further improve on the cell energy reconstruction of the LAr Calorimeters.

Summary
The ATLAS LAr Calorimeters have been in operation for more than 10 years.They have achieved excellent performance and stability and shown the reliability of the LAr technology in the harsh environment of the LHC.The first major upgrade of the LAr calorimeters, the Phase-I upgrade, has been successfully installed in the LHC shutdown before 2022.The new digital trigger readout path provides significantly improved triggering performance on electromagnetic objects and is used as the primary trigger now.
During the next LHC shutdown starting in 2026, the complete readout will be exchanged during the Phase-II upgrade to cope with the even higher luminosities and increased pile-up.Prototypes of all electronics boards are already in preparation and being tested.

Figure 1 .
Figure 1.Cumulative luminosity versus day delivered to ATLAS during stable beams and for high energy p-p collisions for the years 2011 to 2023 [2].

Figure 2 .
Figure 2. The luminosity-weighted distribution of the mean number of interactions per crossing for p-p collisions in Run 1 (orange), Run 2 (green) and Run 3 (purple) [2].

Figure 4 .
Figure 4. Liquid argon ionization pulses, before (triangle) and after shaping.The circle markers represent the samples from digitization.Both pulses are normalized to 1 [3].

Figure 5 .
Figure 5. Measured impurity in oxygen equivalent of the liquid argon in the barrel cryostat during Run 1 and Run 2 (2009-18).The shaded areas mark the shutdown periods [6].

Figure 6 .
Figure 6.Liquid argon temperature as a function of time.The shaded areas mark the winter shutdowns.The spikes to lower temperatures in the end-caps observed outside the shaded areas are due to short periods with HEC cold electronics switched off[6].

Figure 7 .
Figure 7. Liquid argon temperature homogeneity in the barrel in the years 2011 and 2012.The difference of the average temperature of one sensor in a given year to the average temperature of the whole cryostat in that year is shown [6].

Figure 8 .
Figure 8. ATLAS Level One Calorimeter single electron trigger efficiencies for the legacy system (red) and the Phase-I system (magenta and green) as a function of the offline electron transverse momentum 2  T[9].

Figure 9 .
Figure 9. Dependence of the rate for the legacy L1_EM22VHI and Phase-I L1_eEM26M electron trigger items on the luminosity.The Phase-I system lowers the rate of the single electron trigger down to around 80 % [9].