Each at different stages of development depending on their implementation schedules and operating conditions, the tracking systems of the ATLAS, LHCb and CMS experiments, like that for ALICE, are undergoing complete replacements to prepare for the extreme operating conditions of future LHC runs.
Towards the CMS phase-2 pixel detector
The original silicon pixel detector for CMS – comprising three barrel layers and two endcap disks – was designed for a maximum instantaneous luminosity of 1034 cm–2 s–1 and a maximum average pile-up of 25. Following LHC upgrades in 2013–2014, it was replaced with an upgraded system (the CMS Phase-1 pixel detector) in 2017 to cope with higher instantaneous luminosities. With a lower mass and an additional barrel layer and endcap disk, it was an evolutionary upgrade maintaining the well-tested key features of the original detector while enabling higher-rate capability, improved radiation tolerance and more robust tracking. During Long Shutdown 2, maintenance work on the Phase-1 device included the installation of a new innermost layer (see “Present and future” image) to enable the delivery of high-quality data until the end of LHC Run 3.
During the next long shutdown, scheduled for 2025, the entire tracker detector will be replaced in preparation for the High-Luminosity LHC (HL-LHC). This Phase-2 pixel detector will need to cope with a pile-up and hit rate eight times higher than before, and with a trigger rate and radiation dose 7.5 and 10 times higher, respectively. To meet these extreme requirements, the CMS collaboration, in partnership with ATLAS via the RD53 collaboration, is developing a next-generation hybrid-pixel chip utilising 65 nm CMOS technology. The overall system is much bigger than the Phase-1 device (~5 m2 compared to 1.75 m2) with vastly more read-out channels (~2 billion compared to 120 million). With six-times smaller pixels, increased detection coverage, reduced material budget, a new readout chip to enable a lower detection threshold, and a design that continues to allow easy installation and removal, the state-of-the-art Phase-2 pixel detector will serve CMS well into the HL-LHC era.
LHCb’s all-new VELO takes shape
LHCb’s Vertex Locator (VELO) has played a pivotal role in the experiment’s flavour-physics programme. Contributing to triggering, tracking and vertexing, and with a geometry optimised for particles traveling close to the beam direction, its 46 orthogonal silicon-strip half-disks have enabled the collaboration to pursue major results. These include the 2019 discovery of CP violation in charm using the world’s largest reconstructed samples of charm decays, a host of matter–antimatter asymmetry measurements and rare-decay searches, and the recent hints of lepton non-universality in B decays.
Placing the sensors as close as possible to the primary proton–proton interactions requires the whole VELO system to sit inside the LHC vacuum pipe (separated from the primary vacuum by a 1.1 m-long thin-walled “RF foil”), and a mechanical system to move the disks out of harm’s way during the injection and stabilisation of the beams. After more than a decade of service witnessing the passage of some 1026 protons, the original VELO is now being replaced with a new one to prepare for a factor-five increase in luminosity for LHCb in LHC Run 3.
The entirety of the new VELO will be read out at a rate of 40 MHz, requiring a huge data bandwidth: up to 20 Gbits/s for the hottest ASICs, and 3 Tbit/s in total. Cooling using the minimum of material is another major challenge. The upgraded VELO will be kept at –20° via the novel technique of evaporative CO2 circulating in 120 × 200 µm channels within a silicon substrate (see “Fine structure” image, left). The harsh radiation environment also demands a special ASIC, the VeloPix, which has been developed with the CERN Medipix group and will allow the detector to operate a much more efficient trigger. To cope with increased occupancies at higher luminosity, the original silicon strips have been replaced with pixels. The new sensors (in the form of rectangles rather than disks) will be located even closer to the interaction point (5.1 mm versus the previous 8.2 mm for the first measured point), which requires the RF foil to sit just 3.5 mm from the beam and 0.9 mm from the sensors. The production of the foil was a huge technical achievement. It was machined from a solid-forged aluminium block with 98% of the material removed and the final shape machined to a thickness of 250 µm, with further chemical etching taking it to just 100 µm (see “Fine structure” image, right).
Around half of the VELO-module production is complete, with the work shared between labs in the UK and the Netherlands (see “In production” image). Assembly of the 52 modules into the “hood”, which provides cooling, services and vacuum, is now under way, with installation in LHCb scheduled to start in August. The VELO Upgrade I is expected to serve LHCb throughout Run 3 and Run 4. Looking further to the future, the next upgrade will require the detector to operate with a huge jump in luminosity, where vertexing will pose a significant challenge. Proposals under consideration include a new “4D” pixel detector with time-stamp information per hit, which could conceivably be achieved by moving to a smaller CMOS node. At this stage, however, the collaboration is actively investigating all options, with detailed technical design reports expected towards the middle of the decade.
ATLAS ITk pixel detector on track
The ATLAS collaboration upgraded its original pixel detector in 2014, adding an innermost layer to create a four-layer device. The new layer contained a much smaller pitch, 3D sensors at large angles and CO2 cooling, and the pixel tracker will continue to serve ATLAS throughout LHC Run 3. Like CMS, the collaboration has long been working towards the replacement of the full inner tracker during the next long shutdown expected in 2025, in preparation for HL-LHC operations. The innermost layers of this state-of-the-art all-silicon tracker, called the ITk, will be built from pixel detectors with an area almost 10 times larger than that of the current device. With 13 m2 of active silicon across five barrel layers and two end caps, the pixel detector will contribute to precision tracking up to a pseudorapidity |η| = 4, with the innermost two layers expected to be replaced a few years into the HL-LHC era, and the outermost layers designed to last the lifetime of the project. Most of the detector will use planar silicon sensors, with 3D sensors (which are more radiation-hard and less power-hungry) in the innermost layer. Like the CMS Phase-2 pixel upgrade, the sensors will be read out by new chips being developed by the RD53 collaboration, with support structures made of low-mass carbon materials and cooling provided by evaporative CO2 flowing in thin-walled pipes. The device will have a total of 5.1 Gpixels (55 times more than the current one), and the very high expected HL-LHC data rates, especially in the innermost layers, will require the development of new technologies for high-bandwidth transmission and handling. The ITk pixel detector is now in the final stages of R&D and moving into production. After that, the final stages of integrating the subdetectors assembled in ATLAS institutes worldwide will take place on the surface at CERN before final installation underground.