CDF Ghost Muons

By John Conway | November 2, 2008 2:06 pm

It’s either an unaccounted-for background or it’s new physics. In either case, it’s complicated, for sure.

The CDF Collaboration, at Fermilab’s Tevatron accelerator, has submitted for publication a new paper describing a subsample of proton-antiproton collision events in which there is at least one muon produced far from the primary proton-antiproton interaction. This subsample is not yet described by known processes, including the effects of detector/reconstruction failures, and is starting to cause somewhat of a sensation in the high energy physics community. As a member of CDF, I can tell you this analysis has gotten some rather intense scrutiny in the past several months!

The excess subsample is called a “ghost” sample in the paper, and is characterized by the fact that there can be several muons whose direction of travel lies within 37 degrees of the primary (highest energy) one, and the distribution of muon “impact parameters” has a long tail, out to several centimeters. The impact parameter is a measure of how far away from the main event vertex the particle was produced, and so these extra muons appear to come from the decay of some sort of particle with a lifetime much longer than that of the b quark.

Is this new physics? Or did CDF underestimate the rate of background processes leading to this sort of observation?

The subsample in question came to light in the course of the measurement of the b quark pair production cross section, which is proportional to the rate at which events with a b quark and a b antiquark are produced. This measurement is done in at least two ways. Unlike the lighter quarks, the b quark lifetime is long enough that it flies a few millimeters through space before decaying. So in one method, one looks for “secondary vertices”, distinct from the primary location in space where the proton and antiproton collided. These secondary vertices are where the b quarks decay to several hadrons, or possibly leptons such as muons.

The muon is the heavier cousin of the electron. It has a mass of 106 MeV, compared with 0.511 MeV for an electron. Being so much heavier, and having a lifetime of 2.2 microseconds, when it is produced, it tends to travel through very thick layers of material before stopping. Indeed, in modern large collider physics experiments, the muon detection systems lie outside the rest of the detector, effectively using it as shielding from all the other particles produced in the collisions, which are mainly pions. Pions are hadrons, consisting of a quark and an antiquark. Pions tend to “shower” and leave their energy in the heavy layers of material (typically lead and steel) in the calorimeters. But sometimes they can “punch through” and leave hits in the muon detection system, fooling us. That’s one clear component of this ghost sample. If we have badly underestimated this, it could account for all of it, but that is not too likely as far as we can tell.

The b quark decays to muons offer another way to measure not only the pair production cross section but the “mixing” of b quarks. Due to a subtlety in the weak interaction that we need not explore here, a b quark (or more properly speaking a B hadron) flying along can change spontaneously into its own antiparticle. This can lead to events where, if both b quarks decay to muons, they can have the same electric charge if one b quark has flipped to the charge opposite that with which it was produced.

In the analysis presented in the paper, the authors (a group from Frascati/Harvard led by Paolo Giromini) have ostensibly resolved a long standing disagreement between the results from two different methods to measure the b pair cross section. If one selects events with two muons, but tighten the previously used requirements on where the muons come from, demanding that they emanate from near the main event vertex, they find much better agreement between the two different measurements of the b pair cross section. But this implies that the previous measurements suffered from a large, unaccounted-for background. What is it?

The paper is an exploration of this ghost background, and the conclusion is that we can’t explain it at this point. So the collaboration has published what we’ve learned so far, in hopes that other experiments, especially our neighbors around the ring at D0, can look at their data and tell us if they see this sort of thing too. This is the most important first question to be answered, and if they do see it, I can tell you what will happen: all hell will break loose in the field. If they don’t, well, we have more work to do in figuring out just how this ghost background comes about.

If this is the first observation of some sort of new physics, then it is tremendously exciting and very, very weird. Though, oddly, not entirely unanticiapted. Neal Weiner and Nima Arkani-Hamed have a recent paper out where they predicted “lepton jets” not unlike what we are seeing. Kind of makes the hair stand up on the back of your neck…but let’s not get too far ahead of ourselves just yet. I am sure the theorists are already very busy!

If you want more details, there is a very nice and much more physicist-oriented post by Tommaso Dorigo available at his blog.

Stay tuned!

  • Pingback: Not Even Wrong » Blog Archive » Discovery of a New Particle?()

  • a quantum diaries survivor

    Hi John,

    nice post. I do think this is a very weird background, but indeed, the 1% or so chance that it instead is new physics (my very own gut feeling) makes this a very exciting thing.

    I am sorry that 33% of our colleagues chose to not sign this paper. I think it was due to dissatisfaction with the publication procedure rather than with the content of the paper. I do not know about you, but I have the feeling that many of your and my colleagues in CDF did not read the paper and the documentation closely enough. For me, the fact that scrupolous and knowledgeable physicists such as Kevin, Doug, and others have found no clear problem with the results is a guarantee that it deserved to be published.

    So, let’s look forward to very interesting times.


  • Pingback: CDF publishes multi-muons!!!! « A Quantum Diaries Survivor()

  • optimistic

    What’s this? A third of CDF physicists refused to be on the author list? Tommaso, can you be more specific about what they objected to in it? They’ll be gutted if it turns out to be something real, so they must have had strong reasons.

  • A Student


    Thanks for the additional info

  • Chris W.


    This sentence didn’t quite parse:

    In the analysis presented in the paper, the authors (a group from Frascati/Harvard led by Paolo Giromini) have ostensibly resolved a long standing disagreement between the results two different methods to measure the b pair cross section.

    Is “results” a stray word that should have been removed during editing?

    Somehow it’s appropriate that this work should be drawing attention right around Halloween. :)

  • Carl Brannen

    What I think is odd about the paper is that they never mention the word “electron” and only talk about their electromagnetic calorimeter when placing its position relative to the muon detectors.

    One would certain expect muon production to have at least some electrons, but more importantly, the cosmic ray people have been talking about Centauros (anti-Centauros) for years. In those events, a single incoming particle creates a shower of hadrons (electromagnetic) stuff which is out of balance with the usual expectation of pi+, pi- and pi0 in roughly equal abundance.

  • John

    Chris W. – thanks for pointing out the typo.

    Carl, we’ll look at electrons soon. Electrons in jets are a bitch to reconstruct, though. Muons are much easier.

  • Bob McElrath

    The model by Arkani-Hamed and Weiner absolutely cannot be responsible for these muons. Their light boson would be extremely narrow, and would show up in an opposite sign dilepton plot quite clearly (just as the J/Psi does in at least one of the plots).

    I think the answer is clear: some currently unknown B-baryon or B-meson (or maybe several) has an anomalously large lifetime. The lifetime is only a factor of 10 over the B meson, so a B-hadron is an obvious choice. These states are not in Pythia and will not show up in your MC. Most of the B-baryons are unknown. These things definitely are in the sample, the only question is whether there is anything in there that’s not B-hadrons, and why does the state have an anomalously long lifetime.

    A more specific analysis need to be done looking at exclusive final states to identify the B-hadrons. The paper has a lot of plots which clearly mix a lot of different signals coming from different mesons, so it’s extremely hard to interpret them. The only thing that’s clear is that there is no new state with a 2-body dimuon decay (which B-decays aren’t anyway). But discovering a pile of B-hadrons not in the PDG is exciting!

  • Neal Weiner

    Hi Bob
    Our statement about lepton jets wasn’t just about production of the vector itself, because typical decays through the G_dark sector don’t go right to the vector, but proceed through some intermediate states, hence the n>=2 rather than n=2 definition of the lepton jets…

    That said, getting this rate would be a challenge!


  • Lawrence B. Crowell

    The paper by Arkani-Hamed & Weiner makes these suggestions in figure 2. If I am right these ideas of “gauge-Moose” are a sort of twisted SUSY with a chain of gauge fields on the winding loop around some space. I’d have to read this paper more fully to get a better idea of where this is heading. Yet figure 2 does point to electrons as the lepton jet. One might expect this do dominate over muon production.

    This paper along with Sean’s recent paper and with other conjectures and hypotheses on DM do hint that DM might in some way reflect a weakly interacting ghost world that couples to our “gauge world” by gravity and very weakly by other interactions.

    Lawrence B. Crowell

  • a quantum diaries survivor

    Hi optimistic,

    a third of the collaboration objected about the publication process, the paper content, the personality of the main author, and the day of the week of the release.

    The thing is, the main author is a brilliant physicist, with zero PR skills. He alienated most of his colleagues during 25 years of work in CDF. This explains some of the reaction to his analysis.

    The analysis itself is complicated, as John states from the outset above. It entails loosening the criteria of identification of muon candidates, which have a good reason to be there. But still, we need to understand our muon candidates even with loose cuts. This leads us to this sample, which is very hard to reconcile with SM sources. Many in CDF thought that we should keep to ourself our dirty laundry, others thought we should release that information.


  • Pingback: News from the CDF and PAMELA experiments « Theorema Egregium()

  • Sili

    What is it they say? “New physics isn’t hailed with a ‘Heureka!”, but by someone going ‘Huh. That’s odd.'”

  • Maynard Handley

    “Due to a subtlety in the weak interaction that we need not explore here, a b quark (or more properly speaking a B hadron) flying along can change spontaneously into its own antiparticle.”

    Can you clarify this statement?

    Would you have a problem with the claim that the essential difference between a particle and an antiparticle is the sign of the rate of change with time. (So for an energy eigenstate, this rate of change, ie the frequency, is either positive or negative, giving us particle or anti-particle)? Or would you say such a characterization in fundamentally misguided?

    If we assume this characterization is correct, is the point then that a B meson
    – is not in fact an energy eigenstate AND
    – unlike most “simple” non-eigenstates. if we were to plot “phase” (in some vague sense) against time, rather than seeing some variety of sinusoid like a beat pattern, the plot is best done showing a complex point against time, with the point moving first clockwise, then slowing down, stopping, moving anti-clockwise, then the whole thing reversing?

  • Chris Austin

    Hi John,

    Thanks for the helpful explanations. What is your opinion on Bob McElrath’s suggestion above that the explanation will turn out to be unknown B-hadrons with anomalously long lifetimes, and do you have the necessary raw data that would be required to disentangle the contributions of different unknown B-hadrons, and analyse exclusive final states as he suggests?

    Best regards,

  • Pingback: Hello world! « How the World Works()

  • Pingback: New Physics?()


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Cosmic Variance

Random samplings from a universe of ideas.

See More

Collapse bottom bar