Thursday, August 29, 2013

Two-sigmaish CMS multilepton excesses with a \(\tau\)

A possible hint of third-generation superpartners

Matt Strassler mentioned an interesting anomaly reported by CMS at a SUSY conference this week:
A Discrepancy to Keep an Eye On (Prof Dr RNDr Matt Strassler PhD CSc DrSc Dot COM)
It's small enough so that you may assume that it's just another example of a fluctuation that will go away with more data. But it's large enough for many of us to gain the right to be intrigued. ;-)

The excesses have something to do with multileptons. If you search this blog for multileptons, you find many articles, mostly from late year 2011 and early year 2012. The words "year" were inserted for you to notice that there were many hyperlinks in the previous sentence. It's plausible that those flukes have gone way during the 1.5-2.0 years.

What are the overrepresented events this time?

The CMS folks have performed a search for some signatures – at this point, it's not quite clear what comprehensive theory beyond the Standard Model they are testing but multileptons of course do naturally appear in long enough decay chains of supersymmetric particles (they are a strong hint that at least a pair of new particles was produced because they're almost absent in the Standard Model and in the decay of one new particle) – for events with four leptons and some of the subsets of these events exhibit modest but tantalizing excesses.

We're talking about events with four leptons including
  • at least one \(\tau^\pm\) that decays to hadrons; I start with that to point out that the signal could have something to do with the third generation of fermions and their partners
  • one light (\(e^\pm\) or \(\mu^\pm\)) lepton-antilepton pair (this is what OSSF1, opposite-sign-same-flavor-one, stands for) that seems not to originate from a Z-boson decay (because the Z-boson is assumed to be understood and boring; the condition is called "off-Z")
  • the total energy in the lepton pair plus a few other jets quantified by a variabled called \(H_T\) should be surprisingly small relatively to expected LHC energies; this condition indicates a threshold above which a new massive particle was barely created
  • another charged lepton, \(e^\pm\), \(\mu^\pm\), or \(\tau^\pm\)
  • some missing momentum – which surely gets a contribution from the neutrino(s) that is (are) produced along with the \(\tau\) lepton but may also include the LSP etc.
A table that Matt reproduces for us has about 48 entries and 3 of them contain noticeable excesses. All of them are in the OSSF1, \(H_T\lt 200\GeV\), off-Z, \(N(\tau_h)=1\), \(N_{b\rm -jets}=0\) – bins with other options for these variables are really consistent with the Standard Model predictions.

The three bins with an excess only differ in the missing transverse energy:\[

E_T^{\rm miss} & {\rm predicted} & \text{observed}\\
& \text{events}& \text{events}\\
(0,50) & 7.5\pm 2 & 15\\
(50,100) & 2.1\pm 0.5& 4 \\
(100,\infty)& 0.6\pm 0.24 & 3\\
{\rm total} & 10.2\pm 2.1 & 22\\

\] Sorry if the error margins shouldn't have been added in quadrature; even if it is roughly correct, it's not the totally accurate way to account for the error margins. I guess that the right error should be between 3 and 4 events. I suspect that the errors in the table are just the systematic ones and the Poisson-type statistical ones \(\sqrt{N}\) must be added in quadrature manually. If you can clarify these issues, it would be appreciated.

At any rate, when you combine these three similar channels – naturally ignoring the magnitude of the transverse energy – you get something that may look like a greater than (or at least approximately equal to) \(3\sigma\) excess (because \(10.2\) and \(22\) are really far from each other). I don't want to say \(5\sigma\) which the sloppy calculation above could hint at because I don't believe this can be the right result.

I would say that the missing energy may be arbitrarily low in the events above so if an LSP is created, it should be a light one, safely below \(50\GeV\) – possibly the \(8.6\GeV\) dark matter particle suggested by CDMS II-silicon and others. And the superpartners created at the beginning of the reaction are likely to have something to do with the third generation – like staus. Or sbottoms...

This is not an excess you should think about every night at 3 a.m. so far. But it's an excess that you may return to at some point in the future and describe by the words that you already knew about it on August 29th, 2013, before the revolution got started.


  1. Taking into account that Matt Strassler thinks this is worth to mention at least adds an additional sigma ... ;-P


  2. LOL, but maybe you're only allowed to add his personal sigma in quadrature which would still make a small difference.

  3. My thoughts exactly!

    Just as a follow-up: "Today (as I sit in a waiting room for jury service)"

    What are the odds he'd think there's enough evidence to convict? ;)

  4. Ha ha, LOL have you made that one up?

    I think it is time that the deeper picture how nature works starts finally to unfold, it least would I higly enjoy observing how things get finally clarified to a consistent picture that nicely fits with all present and future observations :-)

  5. I would really hope this is not an error!

    Third generation superpartners, so fast?!

  6. Dear Dimensions10, there's no error in this major point. It's been known for years that 3rd generation superpartners have very good reasons to be more accessible than the first two generations.

    Another reason is that the first two generations are easier to produce and if they were really light, we would have already found them.

    For an example of the first point, note that the superpartners that are light enough help to explain the lightness, 126 GeV, of the Higgs boson. But not all superpartners are equally important. Because the top quark - and its partner stop - are more strongly coupled to the Higgs, it's primarily the stop that has to be light enough for the Higgs mass to be natural enough. Higgsinos and partly gluinos are also somewhat important, but the top is the most important particle.

    So it's been actually usual to construct models where the mass ordering of the superpartners' generations is reverted relatively to the known generations of fermions: 3rd generation is lightest etc.

  7. That's certainly good to know! : ) And thanks for the great explanation about the stop - higgs coupling..

  8. Ha. I got it before reading the second link by the "thoroughbred" in the wording of the link.