In this post I’ll describe how we can use jet mass to identify a large jet containing the decay products of a boosted particle versus your every-day jet coming from a light quark (ie: any quark but the top) or a gluon.
Last time I introduced the top quark decay mode into jets and described why it’s necessary to consider these boosted particles in future LHC analyses. To underline the point, one of our recent ATLAS searches for a new heavy resonance decaying into a top/anti-top pair has included boosted tops, where the three hadronic decay products of one of the tops were unresolvable as individual jets.
We haven’t found new physics….yet….but we were able to show that the sensitivity to new physics increases by a lot once boosted tops are added in:
“Resolvable” hadronicly-decaying tops only:
Limit of Z'->ttbar mass in the semi-leptonic channel with "resolvable" jets only: M(Z') > 880 GeV
Including highly-boosted tops as well:
Limit of Z'->ttbar mass in the same channel when including boosted hadronic tops: M(Z') > 1.15 TeV
The lower-mass limit of a resonance like a new Z’ boson improved by a few hundred GeV! So what’s the trick?
Jet Mass
Until now, mass has been a relatively uninteresting property of a jet. When all the jets are separately reconstructed in the detector, we care a lot more about their energy and momentum….you just add them up and get back the invariant mass of the interaction that produced them, or in our case, the top quark that decayed into them. We don’t have such a luxury if the only thing we have to work with is one large jet containing indistinguishable decay products.
Ultimately, however, what we really want is the mass of the boosted top; the three individual jets don’t matter as much by themselves. So just like before when we could combine the energies of the three jets to infer the mass of the interaction, we can get the boosted top mass by adding up all the bits which compose the large jet, where in this case “bits” = calorimeter energy clusters.
Here, where E_i and p_i are the energy and three-momentum of the “ith” jet constituent. And it totally works! Using the jet mass is great for discriminating these from other normal jets in the detector; a light-quark or gluon jet will have a steeply-falling mass spectrum, but a jet from a boosted top will have a mass peaking around 173 GeV.
Comparison of the data and the Standard Model prediction for the large-R jet mass distribution after a background subtraction.
You may notice in this plot that the mass peaks somewhat higher than expected. This is because you never get a jet with just the energy deposits from the top decay products….it is an experiment after all and nothing is ever that clean. The detector is filled with a whole mess of other junk coming from the rest of the proton-proton interaction (called the “underlying event”), not to mention the high luminosity conditions at the LHC causing extra jets from other softer proton interactions in the same event (called “pile-up”).
How to deal with that is a whole topic in and of itself, so I’ll save that discussion for next time.
New jet substructure techniques on the horizon
Besides the mass, large jets which contain the decay products of a boosted object like the top quark are expected to have a much different internal structure than your typical jet from a light-quark or gluon. Jet substructure calculations represent the core of new developments in the field of boosted physics, and their discussion occupied a large chunk of the agenda at the BOOST conference in Valencia.
As an experimentalist, part of the fun attending the conference was getting to meet the theorists who are working on this. In high-energy particle physics, when performing a test on some theoretical model, the usual feedback time to go from theory paper to experimentalists announcing “I found it!” (or more likely, “I didn’t find it…”) is decades.
This is not the case when it comes to boosted physics. Some of the papers outlining new jet substructure models came out less than 5 years ago, and we already have experimental results! I get the impression that a lot of this has to do with the spirit of the BOOST conference series, which is a forum for open communication and productive discussion between theorists and experimentalists.
In fact, this HEP subfield has really exploded since the first BOOST conference was held in 2009. Theorists are coming up with all kinds of new ideas that go way beyond jet mass as a way of determining whether or not a jet is a good boosted candidate. There was a great Venn diagram by Matthew Schwartz of Harvard to describe the “happy medium” theorists are trying to achieve between models which were calculable by them, measurable by us, and/or useful to everyone.
….though some people argued a little about which models were actually “useful” or not. 😉
As you can see, there are quite a few ideas. Jet mass still turns out to be the most powerful tool for finding boosted particles, but there is still a lot of unused information inside a jet. Any extra help in telling apart jets with substructure versus those without is extremely valuable in extending our discovery reach for new physics.
Since this post is already getting pretty long I’ll come back to discuss some of these ideas in more detail in a couple of weeks.
