LUMI’06 takes strides towards LHC upgrade

2 March 2007

The LUMI’06 workshop, which took place in Valencia, brought together experts from Europe, Japan and the US to review scenarios for increasing the LHC’s luminosity.

Over the past two years, studies to upgrade the LHC have made great progress under the joint auspices of the European CARE accelerator network on High-Energy High-Intensity Hadron Beams (HHH) and the US LHC Accelerator Research Program (US-LARP). These efforts recently culminated in the third topical workshop of the CARE-HHH-APD network, LUMI’06, which was held in Valencia on 16–20 October 2006. About 70 members of CARE and LARP and their associated institutes attended, including 13 participants from major US laboratories and two from KEK in Japan.


LUMI’06 was devoted to the beam dynamics of the LHC luminosity upgrade and to high-intensity effects limiting the performance of both the LHC accelerator complex at CERN and the Facility for Antiproton and Ion Research (FAIR) at GSI. More specifically, the double objective of LUMI’06 was to establish a forward-looking baseline scenario for the LHC luminosity upgrade and to concur on a scientific rating of alternative scenarios for the upgrade of the CERN accelerator complex, while also assessing the performance of the GSI FAIR synchrotrons.

The workshop concluded an exciting year of intense HHH networking activity, in which several other workshops and conferences were devoted to various LHC upgrade issues, treating topics such as crystal collimation and channelling, rapid switching devices, superconducting magnet design, magnet optimization, super-ferric storage-ring approaches and beam dynamics in high-brightness hadron beams. Throughout the year, in preparing for LUMI’06, there had also been great progress made on the development of a web repository for accelerator physics codes, code benchmarking and on the construction of a database for superconducting cables and magnets.

Accomplishing key goals

A highlight of experimental studies just before the workshop was the first successful test of crystal reflection with a 400 GeV proton beam at CERN in the SPS North Area by the H8-RD22 collaboration. The demonstration of an extremely high effective field, together with more than 95% extraction efficiency, opens up a new perspective for the upgrade of the LHC collimator system. Such an improvement is certainly welcome, in view of the known obstacles on the way to reaching the nominal LHC performance.


Several speakers at LUMI’06, including CERN’s Ralph Assmann, Rudiger Schmidt and Gianluigi Arduini, surveyed the various difficulties and limitations of the nominal LHC and of the existing CERN complex – related, for example, to collimation, machine protection and the injectors – and they pointed out the challenges that need to be overcome to reach the LHC design luminosity of 1034 cm-2s-1. Nevertheless, after five days of intense discussions, the workshop participants displayed great optimism about the upgrade goal of boosting the LHC peak luminosity by another factor of 10 beyond nominal towards 1035 cm-2s-1.

A key objective that LUMI’06 successfully accomplished was to select the most promising upgrade paths and, possibly, improve them or identify new ones. The workshop considerably reduced the number of alternative scenarios for the upgrade of the interaction region by arguing against all layouts with strong separation dipoles between the collision points and the low-beta quadrupoles closest to them. A primary argument in favour of the “quadrupole-first” solutions is the different level of difficulty and implied development timescale. In particular, at present nobody in the world is actively prototyping strong superconducting dipole magnets.

In considering the technology on which to base the new low-beta quadrupoles there are two alternatives – namely “pushed” NbTi and Nb3Sn – that the workshop decided to pursue in parallel until the first results become available from long Nb3Sn prototype magnets to be built in the US. This should be within the next two or three years. CERN’s Tom Taylor in particular proposed an intriguing “hybrid” solution, combining both NbTi and Nb3Sn technologies.


Two novel concepts that would greatly enhance the luminosity potential of an LHC upgrade foresee complementing the interaction-region upgrade with additional slim superconducting dipole magnets (DO) or quadrupole doublets (QO), which would be embedded deeply inside the upgraded detectors. Together with other measures, such elements may allow squeezing the beta functions at the collision point by a factor of seven, as opposed to two, beyond nominal, down to a ß* of around 8 cm. Extensive studies are needed for the accelerator and detectors before these novel schemes can be soundly judged for viability.

The compensation of long-range beam–beam effects by a current-fed metre-long wire running parallel to the beam is by now almost established as a valuable and inexpensive complementary tool for enhancing performance. At LUMI’06, Fermilab’s Vladimir Shiltsev proposed the additional use at the LHC of electron lenses both for head-on beam–beam compensation and as a halo collimator. Large-angle “crab” cavities for interaction-region layouts with large crossing angles were rejected in view of numerous technical challenges, which several speakers identified, including Brookhaven’s Rama Calaga and Ramesh Gupta, and CERN’s Rogelio Tomas and Joachim Tuckmantel. Participants appreciated the high risk involved with choosing a crossing geometry that would fully rely on their functionality. In contrast, simpler small-angle crab cavities were recognized as a potentially powerful tool for realizing very small beta functions in conjunction with the detector–integrated dipole D0. KEK’s Kazuhito Ohmi presented simulations of LHC emittance-growth with crab cavities and feedback. The results of the first-ever crab cavity operation in a collider at the KEKB electron-positron machine will be the next milestone. Expected soon, these results will have a big impact on the further pursuit of using crab cavities at hadron colliders.

Figure 1 shows two example layouts of an upgraded LHC interaction region, accommodating several of the advanced elements discussed during the workshop. Advantages of the first scheme, with a detector-integrated slim dipole located about 3 m from the interaction point, are the reduced number of long-range collisions and the absence of geometric luminosity loss. The second scheme relaxes the triplet quadrupole requirements and decreases the chromaticity. A combination of the two schemes – that is an interaction region layout containing both D0 and Q0 – is another possibility, which combines all the advantages.

Tackling the beam-parameter frontier

The workshop also made sigificant progress at the beam-parameter frontier. In the past, parameter sets suffered either from an unacceptable number of events per crossing or from an electron-cloud heat load that by far exceeded the available cooling capacity. LUMI’06 approved two compromise solutions with 25 ns and 50 ns bunch spacing, which the authors presented (table 1). For these new sets of beam parameters the number of events per crossing stays near the maximum acceptable value, while the predicted electron heat load remains safely below the projected cooling capability.

The 25 ns option is accompanied by an 8 cm ß*, which requires a D0 magnet inside the detector, Nb3Sn large-aperture quadrupoles and a low-angle crab cavity. The 50 ns option has ß* = 25 cm, for which optics solutions exist based on either technology for the quadrupoles. In addition it needs only the wire compensation of long-range beam-beam effects. Since LUMI’06, the two biggest LHC experiments, CMS and ATLAS, have indicated a preference for the scenario with 50 ns spacing. LUMI’06 rejected the original baseline upgrade scenario with 12.5 ns bunch spacing – half the nominal – since accelerator physicists, cryogenics experts and detector physicists now generally agree that this spacing will produce an insurmountable heat load. Indeed, at this bunch spacing the well-known heating from image currents in the resistive wall and from synchrotron radiation already require the entire local cooling capacity, leaving zero reserve for the electron cloud, which is predicted to be the dominant heat source.

For the LHC injector upgrade, LUMI’06 has endorsed the Linac4/Superconducting Proton Linac upgrade, as well as PS2, a normal-conducting replacement for CERN’s venerable Proton Synchrotron (PS) with twice the PS circumference. However, the workshop also made it clear that these new accelerators alone may not overcome existing intensity limits in the Super Proton Synchrotron (SPS) and that complementary SPS “enhancements” are likely to be required. Several participants challenged the alternative to the normal-conducting PS2, namely a fast cycling superconducting PS2+. Issues of concern here include the distributed beam losses in a cold machine, heating from the fast ramp, technological development risks, missing physics arguments and lack of human resources. In addition, preliminary simulations presented by Miguel Furman of LBNL indicate that the electron cloud could be a serious problem for new superconducting injector rings.

In summary, the LUMI’06 workshop developed novel scenarios for the upgrade of the LHC interaction regions, while eliminating a number of previous options and proposed novel sets of beam parameters better tailored to a higher-luminosity LHC. The workshop also discussed the supporting upgrades to the CERN accelerator complex, including replacement of the PS, which may be necessary for boosting the integrated LHC luminosity, as well as the peak luminosity. With a substantial participation from US-LARP, the European and US upgrade activities could successfully be re-aligned and a general consensus emerged on the future steps to be taken. According to the present schedule, the LHC interaction regions will be upgraded by around 2014. The interaction region and beam-parameter upgrades should increase the peak luminosity several times. However, harvesting the full gain in the integrated luminosity as well will almost certainly require accompanying upgrades to the CERN injector complex, improving turnaround time and removing intensity bottlenecks.

• The HHH Networking Activity is supported by the European Community Research Infrastructure Activity under the European Union’s Sixth Framework Programme “Structuring the European Research Area” (CARE, contract number RII3-CT-2003-506395).

• Dedicated to the memory of Francesco Ruggiero.

bright-rec iop pub iop-science physcis connect