LHC preparing to increase peak luminosity to 9/ub/s

The beam teams at the Large Hadron Collider are now getting ready to increase peak luminosities up to about 9/μb/s which is 0.28/fb/year. To do this they will switch from their current 25 bunch scheme to a new 49 bunch scheme. Unlike previous filling schemes the bunches will be placed in a pattern that provides less luminosity to the ALICE experiment than the others. ALICE does not need higher luminosity at this time because it is designed for separate experiments using heavy ion collisions that will take place in November.

The new filling scheme is called 1250ns_48b_36_16_36 or 1250ns_49b_36_16_36 if extra diagnostic bunches are included to make up the 49 bunches per beam. 36 is the number of bunch collisions per turn in the experiments ATLAS, CMS and LHCb. This can be compared with the present scheme Multi_25b_16_16_16 which had 16 collisions per turn and delivered peak luminosities of 4/μb/s. if all else is equal we can therefore expect peak luminosities of 4 x 36/26 = 9/μb/s.

Over the last two weeks the total integrated luminosity has increased from 1/pb to about 1.5/pb. The new filling scheme will push the total up faster. In preparation the physicists will do new bunch injection tests today, possibly leading to physics runs with 49 bunches this evening or tomorrow.

In fact the luminosity gathered this month has been a little less than expected but that is because they have been using some of the time to commission better processes. for example they have been testing faster ramping of the energies. That will ultimately improve the recycle time between physics runs. The higher luminosity with 49 bunches should enable them to make up the lost time before the end of the month.

Update: The first 48 bunch physics run was a great success. It delivered 240/nb over 15 hours, before being adjusted for end-of-beam studies. The luminosity started at around 6.6/nb/s and ended at 3.3/nb/s, so the luminosity lifetime was also 15 hours (intensity lifetime was more like 60 to 100 hours). The total integrated luminosity over all runs now stands at about 1.8/pb.

11 Responses to LHC preparing to increase peak luminosity to 9/ub/s

  1. Bill K says:

    Title should say 9/nb-s, not 9/ub-s (introducing a new unit, the nanobarn-second, or perhaps we should call it the nanomotl.) Anyway, it looks like they did get to 48 bunches, a week and a half early. Fabulous! Any bets on how soon they decide to go to 96 bunches?

  2. Philip Gibbs says:

    It’s the 9/μb/s which is correct. The other figures I have are wrong. I will fix.

    The last plan I saw says they will add 32 bunches at a time per week from now on instead of continuing to double, but I doubt they will stick to that for long. I mentioned this plan before at http://blog.vixra.org/2010/08/09/lhc-progress-and-plans/

    They just made stable beams now and hit about 6.7/μb/s . Better fills should take them nearer the 9/μb/s figure.

  3. Philip Gibbs says:

    The 48 bunch run is still going strong after 11 hours and has so far collected about 200/nb. Previous record for a single run was about 120/nb.

    They plan to run this until noon, 1.5 hours to go, then refill for another run.

  4. Luboš Motl says:

    Good info.

    It’s my feeling that all this gradual increase of the luminosity is just a sleight-of-hand to make them look important. If they increased the luminosity 1000-fold immediately, by methods they know very well, I would bet that everything would just work.

    9/ub/s is 9 x 10^{30}/cm^2/s, right? So that’s 2.5 times the current record at atlas.ch which says 3.9 times… instead of 9 times.

    Best wishes

  5. Bill K says:

    They’re taking it slow because from this point on there’s a very real potential for causing damage if the beam should go astray. They’re still seeing an array of problems in between the successes. The current plan is very ambitious, driven by the goal of 1/fb. But plans are often modified depending on circumstances, and I think they probably will fall somewhat short of this.

  6. Philip Gibbs says:

    Hi Lubos,

    I sometimes feel that way too, but from reading some of their morning reports I get some feel for just how complex it all is underneath.

    I think part of the answer to that is they don’t need to increase immediately because they will get there by the end of the year which is soon enough given how long it took to build the machine.

    Plus there are new things they need to do as they build up stored energy. They just started multibunch injection at 25 bunches, now they need to inject them closer together. At design spec they will inject 72 at a go.

    They are also tuning and adjusting parameters that improve the stability. That can only be done in a gradual buildup.

    Another reason to go slowly is that the detectors need to be calibrated and uderstood. Calibration starts at low luminosities with no trigger filtering out events. Then they have to repeat old physics until they are sure they understand the system well enough to know where systematic errors might be.

    They also have to train everyone everytime they work out how to do something new. There are different shift teams that have to be kept informed and shown what has happened. No doubt people come and go. This has been cited as a limiting factor at some points in the comissioning process.

    Finally there is just the risk element. The energy in the beams can do a lot of damage to the collider if it gets loose. The total energy in the protons circulating in the collider ring is already above the energy in the Tevatron. It has enough energy to melt a few kilograms of lead or bore a whole a few meters into solid steel. This total energy figure will increase by another factor of 100 by time they reach the designed parameters for the beams.

    I suppose people who really know about it could give lots more reasons.

    Factor increase is 2.25 due to increase from 16 collisions per turn to 36 collisions per turn in ATLAS, CMS and LHCb. Did you notice that on atlas.ch they now give integerated luminosity as well as number of collisions? Assuming these are simulataneous numbers it means you can get a better estimate for the pp collsions cross-section.

  7. Luboš Motl says:

    Phil: “which is soon enough given how long it took to build the machine”

    Well, that’s nice. It has some logic but we may also say that our monkey (and bacterium) ancestors have wasted tens of millions of years (or more) by not having built a 7 TeV collider earlier. Does it mean that we can always wait for a million of years?

    That’s nice but you can’t change the past anymore. The history is whatever it is. If something can be done more quickly, it is a waste of time to do it slowly. There may be new and unexpected challenges in the future which will add months or years and they will deserve the time. waiting for no reason doesn’t deserve the time even if the time is small compared to the protractions from the past.

  8. Philip Gibbs says:

    Well, it was a good question and I found some more information before you responded so I added it.

    I think the hardest part is going to be waiting for the collaborations to release positive results when they get them. They already have 5 times the data shown at ICHEP.

  9. Lawrence B. Crowell says:

    Phillip hit the answer. The incremental steps are taken to perform calibrations.

  10. Luboš Motl says:

    Cool. Around midnight, or 0:42 or so,


    CMS recorded about 9.3/ub/s and similarly for others. 😉

  11. Philip Gibbs says:

    Fantastic! The accumulated luminosity for this run has now passed 200/nb taking the LHC total beyond 2/pb delivered by about 10am.

%d bloggers like this: