A.M. How did you decide to work in the field of accelerator physics?
J.J: Originally I studied mathematical physics, with a strong orientation towards particle physics, at Edinburgh and Cambridge universities. Following my PhD in Statistical Mechanics I wanted to do something with a more immediate motivation. I had been a summer student at CERN and become fascinated with the particle accelerators so decided to apply for a job opening in what was then the ISR Theory Group and that’s how I started working on LEP. It wasn’t as radical a change as it might appear as I was initially able to apply techniques from statistical mechanics to high-energy electron beams. And, of course, some rather beautiful mathematics is fundamental to the working of circular accelerators. I certainly never regretted redirecting my career as I enjoy the mix of activities that accelerator physics offers. We often work in small groups on interesting physics problems – although you have to learn to speak the languages of the wide range of specialists who are needed to get these complex machines built and running. It’s a particular privilege to work on the front- line colliders at CERN and there is the permanent imperative of delivering a precious and expensive product, the accelerator performance, to our “customers” in the experiments.
A.M: How long have you been working at CERN and specifically in exploring the role of the LHC as a heavy-ion collider?
J.J: I’ve been working at CERN since 1980, much of that time on LEP. My personal involvement with the LHC goes back to a period in 1984 when I contributed to the first feasibility study. In early 2003, I was asked at quite short notice to review the heavy ion performance at the LHC workshop in Chamonix. So I have been deeply involved in the LHC heavy ions for about 10 years.
Soon after, we had to prepare the Design Report for the LHC and write the chapter about lead-lead operation of the main rings. In those years, we had the “Ions for LHC” project mainly dedicated to the injector complex: building the ECR source, Linac3 and LEIR machines and preparing the PS and SPS to inject lead beams into the LHC. It was led by Karlheinz Schindl and later Stephan Maury with me as deputy, mainly responsible for the heavy-ion performance of the LHC itself.
John Jowett, accelerator physicist and ion-team leader at the LHC, CERN.
A.M: Which are the main differences in the collisions between protons and heavy-ions from your point of view?
J.J. I would say that the most fundamental difference is the electric charge although the mass is also important of course. Heavy ions carry much higher electric charge than protons, 82 times in the case of lead. Physical processes which depend on individual interactions between particles in the same or the opposing beam depend on high powers of the charge and can assume greater importance for heavy ions. One process occurring in the collisions depends on the seventh power and is a dominant limit for lead-lead collisions while being utterly negligible for protons.
A.M Could you describe these physical effects and how they depend on the larger electrical charge of the heavy-ions?
J.J. First of all, the energy of lead ions accelerated at LHC equals 82 times the energy of protons but this amount of energy is distributed among 208 nucleons and hence in the end the energy per nucleon is about 40% of a proton’s energy. But the lead beam is deflected by the same amount in a magnetic field. We have exploited this basic principle - having particles with the same “magnetic rigidity” – in the first two lead-lead runs, in order to commission the heavy-ion collisions as quickly as possible. We minimize the changes to the magnetic configuration although we always have to deal with changes in the collision optics and separation scheme for ALICE. There is no such equivalence for the electric fields applied by the RF cavities but that is taken care of by the adjustment of the RF frequencies to maintain synchronicity.
This is essentially a careful strategy for saving time at the start of the one-month lead-lead runs. Experience at past machines led many to believe that we would barely get started in that time. We just couldn't afford to spend a couple of weeks repeating the standard set-up operations if they were very different for heavy ions. Thanks also to the astounding quality and reproducibility of the LHC magnets, controls and other systems, the commissioning has been done over a weekend and we were up to full luminosity in a few days.
The quadrupoles have a dual opening configuration and are combined in the same magnetic and cryogenic structure. (Image: CERN)
A.M. Are there other phenomena related to the electric charge that appear once the heavy-ion beams start running at the LHC?
J.J: Yes indeed, then we have to start dealing with the effects that go with higher powers of charge, the real beam physics of heavy ions. One of the most serious effects is the so called intra-beam scattering which is multiple small angle Coulomb scattering between particles within the beam. This increases the emittance and size of the beam and hence results in the reduction of the luminosity. At the low injection energy, in particular, this effect can blow up the beam completely and we need to take it very seriously. The phenomenon is even stronger in the SPS, where the ions are prepared to get into the LHC ring. The data from 2011 run show that ions which spent different amounts of time at injection energy in the SPS have different intensities and sizes because of this effect.
Even during collisions of Pb-Pb in the LHC it is responsible for a slow blow-up and loss which contributes to the luminosity decay. Once the heavy ions are colliding another class of interesting effects comes into play. Virtual photons surrounding the nuclei - due to their large charge again - collide in two ways: either photons from one beam with photons from the other or we can have the photons from one beam interacting with the nuclei in the other. The rates are significantly greater than for hadronic interactions where the nuclei overlap.
Cross section of the LHC double dipoles.
Nuclei are removed from the beam by two dominant electromagnetic effects: electromagnetic dissociation and the bound-free pair production. Virtual photons surrounding the ions -due to their large charge- collide in two ways: either photons from one beam with photons from the other or we can have the photons from one beam interacting with the nuclei in the other.
To understand electromagnetic dissociation, consider for example that one nucleus is excited when it interacts with the photons from the other beam and emits a neutron turning 208Pb into 207Pb, a reduction in mass and decrease in magnetic rigidity. The new nucleus will follow a different orbit and it turns out that it will finally end up in the collimators of the LHC. This process happens at a high rate during heavy-ion collisions, reducing beam intensity and luminosity. These losses may also limit the peak luminosity.
The other process that removes particles at an even higher rate scaling with the seventh power of the electric charge is the bound-free pair production. This process becomes very important at the LHC compared to other heavy-ion colliders. Among the enormous number of electron- positrons pairs created by photon-photon interactions, a small fraction are created with the electron bound to a nucleus – forming a one electron-atom - which then has a charge of 81 instead of 82, increasing the magnetic rigidity. Again the particle will follow a different orbit and will be lost from the beam. One could say that Pb81+ formed through this process form a secondary “beam” which is eventually lost in a well-defined spot on the super-conducting magnets, just a few hundred metres from the interaction point.
Welding the magnets at LHC. (Image:CERN)
A.M. Can this process also harm the super-conducting magnets used at the LHC?
J.J: We have been worried about the creation of this stream of Pb81+ particles since 2003. Depositing energy in a magnet can quench the superconducting cable and prevent the machine from running. The heating from this process is proportional to the luminosity of the beam and imposes an upper limit which we have tried to estimate. Simulations have shown
that it should be close to the Ph-Pb design luminosity of 1027 cm-2s-1. Of course these simulations carry a level of uncertainty as there are approximations related to the nuclear physics, the tracking of the particles, the showering /fragmentation of these ions in the beam screen and superconducting coils and finally the heat transfer between the coils and the liquid helium flowing around and through them. I would say that, despite the uncertainties, we now have some measured data and can extrapolate the observed effects at higher energies. It seems the LHC magnets do not quench as easily as was expected but we need to confirm this once the LHC operates at full energies.
A.M. And what happens when heavy-ions hit the collimating system of the LHC?
J.J. Here again nuclear physics plays a crucial role. The LHC has a very sophisticated multi-stage collimation system which was designed to work for protons but cannot work in the same way for heavy ions.
The explanation for this is quite straightforward. When a proton interacts with the collimator it scatters but retains its identity as a proton. It has the same charge and mass and can be directed to secondary collimators where they can be absorbed safely. On the contrary, heavy ions break up by electromagnetic dissociation and nuclear fragmentation into a whole range of species. Again, these have different trajectories and can be lost in the magnets and cause quenching. The LHC is really the first machine where such effects have to be considered as a potential limit on beam intensity.
A collimater for the LHC. The powerful LHC collimation system protects the accelerator against damage due to unavoidable regular and irregular beam loss (Claudia Marcelloni, © CERN).
A.M. Did we have some experience from other machines on how to deal with these phenomena?
J.J. This is also the first time that the losses from bound-free pair production need to be considered as a problem. For example when RHIC collides gold, bound-free pair production takes place but, since their beam-pipe is larger, they don’t experience the same local losses and risk of magnet quenching. In fact, this was not the case when RHIC collided copper in
2005, so we set up an experiment with our BNL colleagues in which we managed to measure bound free pair production losses for the first time. However the effect was tiny compared to what we get at LHC. At LHC design energy and luminosity the beam of Pb81+ ions carries a power of 25 W on each side of each experiment.
Last year we tried one technique that would mitigate this phenomenon. We bumped the orbits of the beams in such a way that the impact of the BFPP ions would be at a shallower angle and would spread over a larger area. This should reduce the concentration of heat in the magnets. The results were rather encouraging and we will rely on this technique after Long Shutdown 1.
A.M. How do these effects limit the physics that we can do?
J.J: As I already mentioned collimation may set an upper limit on the intensity of the beams. Moreover, nuclear effects like the bound free pair production that we discussed can limit the luminosity which consequently limits the data that the experiments can record. In the long term, the continuous loss in the same part of the same magnet can cause radiation damage to the insulation between the superconducting cables. We have estimated that the magnets currently used will last in the lifetime of LHC. At the same time we have developed various solutions and techniques to reduce these effects in view of the ALICE upgrade in Long Shutdown 2.
The ALICE detector.
Specifically, we have thought certain ways in which the magnets around ALICE could be protected and we are planning to implement them during the long shutdown 2. In the Chamonix workshop in 2003 there was a similar discussion and the first answer that came
to our minds was; why not simply install special collimators that will intercept the well- defined beam of Pb81+ ions before hitting the magnets. However this was not an easy task since these collimators had to be installed in the super-conducting areas of the machine that were already constructed then couldn’t be easily modified. This solution has been rejected as practically infeasible. Following this workshop some changes were discussed that could allow the installation of these special collimators which might also be needed in some parts of the LHC for the proton operation. Firstly we would like to have them installed near ALICE in order to get higher luminosities without worrying about the damage in the magnets.
In order to install the collimators near ALICE we need more space. One idea was to move some magnets and open gaps between them. However, this solution would change the geometry of the machine which means that everything should be connected again which would be very complicated from an engineering point of view. We are currently testing new high field magnets that will produce the same bending in a shorter length thus allowing for opening up some space for the collimators. These magnets are under development and we hope that they will be ready for the long shutdown 2.
A.M Do you have to take into account all of the ALICE detectors when designing the collision configurations?
J.J. One of the ALICE detectors that we have to consider most is the Zero Degree Calorimeter. Every time we set up a heavy ion run we need to arrange the beams in order to cross at an angle that will not affect the performance of the ZDC. However, ALICE has a spectrometer magnet that deflects the beams and is compensated by some corrector magnets near the interaction point but this induces a crossing angle which has to be combined with the additional crossing angle that we apply from further out. This has to be done in quite a different way for protons and heavy ions.
In the future at higher energies the effect of the spectrometer will be reduced because it operates at constant magnetic field and higher energy means that the beam will bend less. This will allow us to put more bunches in the beam. Currently we have a minimum bunch spacing of 200 ns because we are trying to avoid any interactions between the beams close to the interaction point.
A.M. So what changes when we move to proton-lead collisions and which are the challenges from your point of view?
J.J. Proton-lead collisions have been one of the most exciting projects for us. It is fair to say that they weren’t really considered in the design of the LHC and they were not mentioned in the LHC Design Report. Nevertheless ALICE has been interested in these asymmetric collisions for quite a long time. The fact that the fundamental design of the LHC was based in a two-in-one magnet seemed to prohibit the collision of hybrid beams at the LHC. We first explored this possibility back in 2005 during a workshop that was organized in collaboration with the experiments and the TH division.
The concerns stemmed from the fact that when RHIC tried a similar mode of operation with equal magnetic fields in the two rings they had many difficulties. These had to do with the fact that at the injection energy the beams were circulating at different revolution frequencies because of their different mass and charge. So the encounter points, where bunches of the two beams have long-range interactions move around the ring and are no longer in fixed positions as is the case for two equal beams.
This modulation of the long range beam-beam effects caused severe losses and beam blow-up problems at RHIC. However RHIC could escape these problems since it has two independent rings and the magnetic fields in each could be modified in order to get equal
frequencies for the two beams. At the LHC we don’t have this option and we need to face up to these effects (although they are weaker) and operate a collider in an almost unprecedented way. The first successful test of injecting asymmetric beams was done on 31 October 2011. During that test we injected and re-phased the beams at top energies where you can make the frequencies equal by displacing the beams from their usual central orbits in the beam pipe by a fraction of a millimetre. This cannot be done at lower energies because it would require the beams to be separated so much that they would be outside the beam pipe. The question was whether we could inject the beams and survive the ramp with good enough intensity and a good enough beam size for physics results. We succeeded in doing that with a few hundred proton bunches and a few lead bunches. In 2011 we didn’t make any collisions because of a subsequent leak in the proton injection septum in the PS machine which ruled out any further proton injection at the end of that year.
Finally, we were able to have the now famous p-Pb pilot run last September and provide a few hours of data taking. It has been very gratifying for us to see the interesting physics results that ALICE and CMS have brought out so quickly from this special run.
The fill was done with only 13 bunches in each beam so it didn’t allow us to study the beam- dynamical effects I mentioned above. Because of some other problems we weren’t able to perform a test with a higher number of bunches. So we are not in a good position to predict the performance of the LHC during the coming p-Pb run in January when we should increase the luminosity by several hundred times. This is a little awkward and we need to be rather cautious.
A.M. Which are the main aims for the January proton-lead run?
J.J. The run will start at a moderate luminosity as ALICE opts to collect some minimum- bias data first before increasing the luminosity to about 1029 by gradually increasing the number of bunches. The intensity per bunch will be approximately 108 for the lead ions and we will have about 300 bunches of 1.5 1010 protons that we hope to gradually increase. Perhaps one of the most important things that is still uncertain is the value of the beta- function that will be applied in the interaction point. This time we will push the ALICE optics further than ever before with a value similar to those in ATLAS and CMS.
Another new thing is that we will be running the LHC beams in an “off-momentum” mode. The proton beam will be displaced - since it is faster than the lead beam- a little bit outward and the lead beam will be displaced slightly inwards. This equalises the revolution times of the two beams. However their magnetic rigidities will not correspond to those of central orbits and this is the so-called “off momentum” mode in the accelerator jargon, another new feature for the LHC.
The focusing will be perturbed, the beam sizes will be perturbed and we will implement some corrections for these effects. This also adds to the complexity of setting up the collimation of the machine and hence we will need to spend more time on that. The two beams will collide with different sizes and intra-beam scattering rates. So there are quite a few new features and in that we sense we are really exploring new regimes with this run.
A.M. How significant will be the effects that limit our luminosity in the case of lead-lead collisions during a proton-lead run?
J.J. The effects in the collisions will be reduced because of the much smaller cross-sections of p-Pb as compared to the Pb-Pb and hence a smaller bound free pair production. We expect the losses to be just a percentage of those in lead-lead.
A.M. Which were the special requests of ALICE for the upcoming p-Pb run?
J.J. ALICE requested two things for the upcoming run. The first is the requirement to reverse the direction of the beams half-way through. Proton-lead will become lead-proton. ALICE asked for this change for important physics reasons.
ALICE has also asked to reverse the polarity of the spectrometer. This should be done once for each direction of the beams. So the run will cut in four pieces corresponding to the four possible combinations of beam directions and spectrometer polarities.
These changes bring us again to the question of time which is a key-factor in these short one- month runs. Of course these changes are not fundamentally difficult or impossible but we need to do them as fast as possible in order to save time.