Crystal collimation brings HL-LHC into focus

7 November 2022

A crucial upgrade of the LHC collimation system to cope with the challenges of High-Luminosity LHC operation is being put to the test during LHC Run 3, write Stefano Redaelli, Mario Di Castro and Roderik Bruce.

Crystal primary collimators
On target The installation in November 2021 of crystal primary collimators, which allow the crystal to be moved to the desired distance from the circulating beam. Credit: CERN SY/STI

The start of LHC Run 3 in 2022 marked an important milestone for CERN: the first step into the High-Luminosity LHC (HL-LHC) era. Thanks to a significant upgrade of the LHC injectors, the Run 3 proton beams are more intense than ever. Together with the raised centre-of-mass collision energy from 13 to 13.6 TeV, Run 3 offers a rich physics programme involving the collisions of both proton and heavy-ion beams. This is made possible thanks to several important upgrades involving HL-LHC hardware that were carried out during Long Shutdown 2 (LS2), ahead of the full deployment of the HL-LHC project during LS3, around four years from now. 

The HL-LHC aims to operate with 2.3 × 1011 protons per bunch (compared to the goal of 1.8 × 1011 protons per bunch at the end of Run 3), producing a stored beam energy of about 710 MJ (compared to 540 MJ in Run 3). Lead–ion beams, on the other hand, will already reach their HL-LHC target intensity upgrade in Run 3. This is thanks to the “slip stacking” technique currently implemented at the Super Proton Synchrotron, which uses complex radio-frequency manipulations to shorten the bunch spacing of LHC beam trains from 75 to 50 ns. Equating to a stored beam energy of up to 20.5 MJ at 6.8 TeV (compared to a maximum of 12.9 MJ achieved in 2018 at 6.37 TeV), the full HL-LHC upgrade needed to handle these more intense ion beams must be available throughout Run 3.

When the LHC works as a heavy-ion collider, many specific challenges need to be faced. Magnetically, the machine behaves in a similar way as during proton–proton operation. However, since the lead–ion bunch charge is about 15 times lower than for protons, a number of typical machine challenges – such as beam–beam interactions, impedance, electron-cloud effects, injection and beam-dump protection – are relaxed. Mitigating the nuisance of beam halos, however, is certainly not one of the tasks that gets easier. 

Absorber collimators

These halos are formed by particles that stray from the ideal beam orbit. More than 100 collimators are located at specific locations in the LHC to ensure that errant particles are cleaned or absorbed, thus protecting sensitive superconducting and other accelerator components. Although the total stored beam energy with ions is more than 30 times lower than it is for protons, the conventional multi-stage collimation system at the LHC (see “Multi-stage collimation” figure) is about two orders of magnitude less efficient for ion beams. Nuclear fragmentation processes occurring when ions interact with conventional collimator materials produce ion fragments with different magnetic rigidities without producing transverse kicks sufficient to steer these fragments onto the secondary collimators. Instead, they travel nearly unperturbed through the “betatron” collimation system in interaction region 7 (IR7) responsible for disposing safely of transverse beam losses. This creates clusters of losses in the high-dispersion regions, where the first superconducting dipole magnets of the cold arcs act as powerful spectrometers, increasing the risk of quenches whereby the magnets cease to become superconducting. 

The ion-collimation limitation is a well-known concern for the LHC. Nevertheless, the standard system has performed quite well so far and provided adequate cleaning efficiency for the nominal LHC ion-beam parameters. But the HL-LHC targets pose additional challenges. In particular, the upgrade does not allow sufficient operational margins without improving the betatron collimation cleaning. Lead–ion beam losses in the cold dipole magnets downstream of IR7 might reach a level three times higher than their quench limits, estimated at their 7 TeV current equivalent.

Various paths have been followed within the HL-LHC project to address this limitation. The baseline solution was to improve the collimation cleaning by adding standard collimators in the dispersion-suppressor regions that would locally dispose of the off-momentum halo particles before they impact the cold magnets. To create the necessary space, two shorter dipoles with a stronger (11 T) field would replace a standard, 15 m-long 8.3 T LHC dipole. This robust upgrade, which works equally well for proton beams, was planned to be used in Run 3. However, due to technical issues with the availability of the new dipoles, which are based on a niobium-tin rather than niobium-titanium conductor, the decision was taken to defer their installation. The HL-LHC project now relies on an alternative solution based on a crystal collimation scheme that was studied in parallel.

Crystals in the LHC

The development of crystal applications with hadron beams at CERN dates back to the activities carried out by the UA9 collaboration at the CERN SPS. Crystal collimation makes use of a phenomenon called planar channelling: charged particles impinging on a pure crystal with well-defined impact conditions can remain trapped in the electromagnetic potential well generated by the regular planes of atoms. If the crystal is bent, particles follow its geometrical shape and experience a net kick that can steer them with high efficiency to a downstream absorber. Crystal collimation was tested at the Tevatron, and in 2018 a prototype system was used for protons at the LHC in a special run at injection energy. The scheme is particularly attractive for ion beams as it was demonstrated that the existing secondary collimators can serve as a halo absorber without risking damage. 

Silicon swerve

At the LHC, a total of four bent crystals are needed for the horizontal and vertical collimation of both beams. During Run 2, a test stand for crystal-collimation tests was installed in the LHC betatron cleaning region of IR7 with the aim of demonstrating the feasibility of this advanced collimation technique at LHC energies. Silicon crystals with a length of just 4 mm were bent to a curvature radius of 80 m to produce a 50 μrad deflection – much larger than the few-μrad angles typically experienced by proton interaction with the 60 cm-long primary collimators (see “Silicon swerve” image). Indeed, to produce such a kick with conventional dipole magnets would require a field of around 300 T in the same volume of the crystal. The crystals were mounted on an assembly (see “On target” image) that is a jewel of accelerator technology and control: the target collimator primary crystal (TCPC). This device allows the crystal to be moved to the desired distance from the circulating beam – typically just a few millimetres at 7 TeV – and its angular orientation to be adjusted to better than 1 μrad. While the former is no more demanding than the control system of other LHC collimators, the angular control demands a customised technology that is the heart of LHC crystal collimation. 

Crystal channelling can only occur for particles impinging on the crystal surface with well-defined impact conditions. For a 6.8 TeV proton beam, they must have an angle of 90° with angular deviations of at most ±0.0001° (around ±2 μrad) – which is similar to aiming at a 10 cm-wide snooker pocket from a shooting distance of 25 km! If this tiny angular acceptance is not respected, the transverse momentum is sufficient to send particles out of the potential well produced between the planes of the crystal lattice, thus losing the channelling condition. Both the beam-impact conditions and the accuracy of the crystal’s angle must therefore be kept under excellent control. 

High precision

The crystal collimators are steered remotely using a technology that is unique to the CERN accelerator complex. It relies on a high-precision interferometer that provides suitable feedback to the advanced controller, and a precise piezo-actuation device that drives the crystal orientation with respect to impinging halo particles with unprecedented precision. During Run 2, the system demonstrated the sub-microradian accuracy required to maintain crystal channelling at high beam energy (see “High precision” figure, top). A recent feature of the newly installed devices is that the interferometer heads (which enable the precise control of the angle) are located outside the vacuum with the laser light coupled to the angular stage by means of viewports. This means that any fibre degradation due to motion or radiation, which was observed on the prototype system, can be corrected during routine maintenance. Using this setup in 2018, an improvement in ion-collimation cleaning by up to a factor of eight was demonstrated experimentally with the best crystal, paving the way for crystal collimation to become the baseline solution for the HL-LHC. 

The test devices used during Run 2 served their purpose well, but they do not meet the standards required for regular, high-efficiency operation. An upgrade plan was therefore put in place to replace them with a higher performing new design. This has been developed in a crash programme at CERN that started in November 2020, when the decision to postpone the installation of the 11 T dipoles was taken. Two units were built and installed in the LHC in 2021 (see “On target” image) and another four are nearing completion: two for installation in the LHC at the end of 2022 and the others serving as operational spares. The first two installed units replaced the two prototype vertical crystals that showed the lowest performance. The horizontal prototype devices remain in place for 2022, since they performed well and were tested with a pilot beam in October 2021. 

Improved ion-collimation cleaning has paved the way to adopt crystal collimation as the baseline of the HL-LHC

The start of Run 3 in April this year provided a unique opportunity to test the new devices with proton beams, ahead of the next operational ion run. One of the first challenges is to establish the optimal alignment of the crystals, to make sure stray particles are channelled as required. While channelled, the impinging particles interact with the crystal with the lowest nuclear-interaction rate: halo particles travel preferentially in the “empty” channel relatively far from the lattice nuclei. Optimum channelling is therefore revealed by the orientation that has the lowest losses, as measured by beam-loss monitors located immediately downstream of the crystal (see “High precision” figure, bottom). Considering the large angular range possible (more than 50 μrad, compared with the full angular range of 20 mrad), establishing this optimum condition is a bit like finding a needle in a haystack. However, following a successful campaign in dedicated operational beam tests in August 2022, channelling was efficiently established for both the new and old crystals, allowing the commissioning phase to continue. 

Looking forward

The LHC collimation system is the most complex beam-cleaning system built to date for particle accelerators. However, it must be further improved to successfully face the upcoming challenges from the HL-LHC upgrade which, for heavy-ion beams, begins during Run 3. Crystal collimation is a crucial upgrade that is now being put into operation to improve the betatron cleaning in preparation for the upgraded ion-beam parameters, mitigating the risks of machine downtime from ion-beam losses. The collimation cleaning performance will be established experimentally as soon as Run 3 ion operation begins. Initial beam tests with protons indicate that the newly installed bent crystals perform well. The first measurements demonstrated that the crystals can be put into operation as expected and showed the specified channelling property. We are therefore confident that this advanced technology can be used successfully for the heavy-ion challenges of the HL-LHC programme.

bright-rec iop pub iop-science physcis connect