Banner
    The Plot Of The Week - Single Top Loud And Clear
    By Tommaso Dorigo | September 20th 2011 05:54 AM | 17 comments | Print | E-mail | Track Comments
    About Tommaso

    I am an experimental particle physicist working with the CMS experiment at CERN. In my spare time I play chess, abuse the piano, and aim my dobson...

    View Tommaso's Profile
    A well-known HEP rule says that yesterday's searched new processes will be tomorrow's annoying irreducible backgrounds; but since I am an optimist, I always see the glass half-full and feel compelled to add that today they are pleasing high-statistics signals. Take single top quark production: the Tevatron experimentalists (you can include me in the lot) banged their head for a decade trying to measure it; they finally succeeded, but the signal always remained a small excess of events in the tail of a highly-refined multi-variable discriminator.

    Nowadays, the 7-TeV collisions of the Large Hadron Collider produce single top quarks at rates much higher than that of the 2-TeV Tevatron, enabling high-statistics studies of those reactions which surpass the precision of CDF and DZERO despite the still smaller total integrated luminosity available at CERN. More important than the total number of events produced (which you can always compute by multiplying the production cross section S by the integrated luminosity L as N = S L) is the fact that at 7 TeV the signal to noise ratio is larger, enabling the signal to "stand out" more clearly.



    The figure above shows the single-top signal (in blue) in the reconstructed top-quark mass distribution for events with two hadronic jets (left) or three hadronic jets (right). The number of jets one should see when single top is produced by the Feynman diagram on the left, the so-called "t-channel" production also known as "W-gluon fusion", is three, since the top quark (in red) is produced together with a b-quark, then decays into lepton, neutrino, and a further b-jet, but there is another quark leg coming from the initial state (the line in the upper part of the diagram).

    However, this last "leg" of the process is typically produced at high pseudorapidity (i.e. close to the beam axis), and it sometimes is not reconstructed as an independent hadronic jet. Because of this, one expects to see a significant fraction of events featuring only two energetic central jets. A separate analysis of 2- and 3-jet events allows to best discriminate backgrounds and take advantage of the different kinematics of the two subsets.

    You should appreciate the high purity of this new single top quark selection: in the plot on the left, events with a reconstructed mass around 150-180 GeV are twice more likely to originate from the signal process than from any of the  backgrounds. In both plots you may also notice that the dominant background, in green, is coming from the production of W bosons together with heavy flavor jets ("W-bbbar" for friends).

    These events originate when a quark-antiquark pair from the colliding protons produce both a W boson and an energetic, virtual gluon (hard color radiation), and then the gluon "splits" in a bottom-antibottom quark pair. If the latter have enough energy, they emerge as honest-to-god jets. The separation of this process from single top production is then only possible by exploiting the different kinematic properties, but it is a tough job.

    A further reason for being interested in that green-painted background is that "W-bbbar" production is the main background to another extremely important signal we are looking at the LHC: associated production of a W boson and the Higgs. Since we by now know that the Higgs mass is lighter than 140 GeV or so, and actually we already have hints that it is in the 120-GeV ballpark (I venture to predict it will be 119+-4 GeV), we also know that over three quarters of the produced Higgs bosons decay to bottom quark pairs. Associated WH production thus mostly yields a W and two bottom quarks, a process that fights again with the same green background shown above.

    Note that CMS already published a search for WH production with Higgs bosons decaying to bottom quark pairs, setting a limit on the rate of the process. Unfortunately, the "Wbbbar background" is really nasty, so that we will need higher statistics to put that process in evidence: the discovery of the Higgs boson will mostly rely on the WW*, ZZ*, and gamma-gamma final states. 

    Comments

    Dear Tommaso,

    If I recall correctly, a few posts ago you said something about betting one dollar on a Higgs at three to one odds, which I take as a one-in-three probability. But now you are predicting it. Does this indicate you have changed your mind? (If so I hope it is not due to a political directive.)

    dorigo
    Hi David,

    no, I am just observing here that if the Higgs exists - which I believe - it is most probably in the indicated range.

    Cheers,
    T.
    Sorry this probably a dumb question, but why does the mass of the top quark vary along the graph's x-axis? I thought the top quark just had one rest mass. Is that the non-rest mass which varies due to variation of the velocity?

    dorigo
    Hi Anon,

    the "top quark mass" in the graph is not a constant of nature, but rather a reconstructed quantity measured from the energy-momentum of the objects observed in the events; it is a "measured mass in the hypothesis that the three objects come from the decay of the top quark".
    In other words, even in case there were only top quarks here, you would not expect to obtain always the same value, but rather a smeared bump of average close to Mt=173 GeV.

    Cheers,
    T.
    Thanks, I understand a bit better now. But, I was wondering what is the underlying physical source of this smeared bump having a non-zero width? Is it Heisenberg's uncertainty principle?

    All the algorithms and code that is used for this analysis - where is it published, stored, archived? Suppose after LHC, there is a hiatus, will this practical knowledge of QCD decay or will it be preserved somewhere? How?

    dorigo
    Hi Arun,

    there are projects to make data persistent and algorithms maintained across platforms. See e.g.
    http://www.dphep.org/

    Cheers,
    T.
    dorigo
    Hi Lubos,

    the value "MH=119 GeV" is written in stone.... literally, on a mountain side in the Italian Alps, since a while ago... With white stones on the green of the grass. Only available during late spring-early fall.

    Cheers,
    T.
    Hi Tommaso,

    Do you have some comment about the superluminal neutrinos at Opera?

    Just kidding, I already read yesterday your nice post on the subject. ;-)

    dorigo
    I will republish it once it does not raise more eyebrows...
    Cheers,
    T.
    Ah finally a nice result on this blog :-)

    (Just kidding - since I had the chance to be convener of the group that produced this very nice result).

    dorigo
    Hi Julien! Nice work! Let's have a beer together the next time we're both at CERN, it's long overdue.
    Cheers,
    T.
    Ciao Tommaso !
    Sure ! it will be with great pleasure !

    "A well-known HEP rule says that yesterday's searched new processes will be tomorrow's annoying irreducible backgrounds"

    It's true in gamma spectroscopy as well :)

    "A well-known HEP rule says that yesterday's searched new processes will be tomorrow's annoying irreducible backgrounds"

    It's true in gamma spectroscopy as well :)

    How about in dating? ;)

    dorigo
    Ha - good one JJ.
    T.