Zooming in on jet substructure could pay off!

13 July 2009


In certain variations of Supersymmetry a symmetry called ‘R-Parity’ can be violated. R-Parity is a multiplicative quantum number present in any supersymmetric model that incorporates both baryon and lepton number conservation in order to preserve the proton lifetime. Whether or not R-Parity is conserved will have important consequences for experiments. If baryon or lepton number is violated then supersymmetric particles may be able to decay into final states containing only Standard Model particles. An example of this would be the decay of the lightest neutralino directly into three jets. But it will not be easy to pick out this decay in a hadronic collider environment...


Jet mass distributions before event selection cuts for R-Parity Violating SUSY events (left) and SM backgrounds. The QCD background must be significantly reduced in order to make the signal peak observable.


Inspired by progress in Higgs searches, Jon Butterworth, John Ellis, Are Raklev and Gavin Salam recently proposed that we might be able to separate neutralino three-jet decays from the anticipated large QCD backgrounds if we can use properties of jet substructure to identify those (highly boosted) neutralinos whose three decay products (sub jets) have "merged" into a single big jet (arXiv:hep-ph/0906.0728). A small group of people from ATLAS took up the challenge of testing, with fully simulated data, whether or not their method really can work in practice at ATLAS. Excitingly the outcome of the efforts was very promising and we have been able to demonstrate that this approach might actually work at ATLAS (ATL-PHYS-PUB-2009-076). The work was completed in June, just in time for SUSY09, the 17th International Conference on Supersymmetry and the Unification of Fundamental Interactions.

Jets are formed from clusters of energy in the hadronic calorimeters. A variety of jet algorithms can be used to take these clusters and form the jets. In the analysis we use the kT jet algorithm. This algorithm introduces a definition of the distance between the various calorimeter clusters and combines closest pair objects until it reaches a point, defined by the parameters of the algorithm, where it is told to stop. If the size of the jets made is large enough then all three of the jets from the neutralino's decay will be bundled up into a single jet.

After forming these large jets we can look at their masses. If our jets were large enough we should find a mass peak at the mass of the neutralino (96.1 GeV in our test). Looking at the jet masses in our SUSY events and in our background events we see that we are swamped by QCD background. Jon Butterworth and his collaborators suggest using a variable calculated from the distances between the components of the jet to pick out which large jets look like they are actually formed from three sub-jets. By looking at jet substructure in this way we can delve into the swamp and pick out the neutralino jets. We ask for two very boosted jets (which is asking for each event to have two potential neutralino candidates) both of which must pass selection criteria based on the distances between the components of the last and next-to-last mergings when the jet was made. Requiring two additional less-boosted jets, a common feature of supersymmetric decays, removes a significant proportion of the remaining background.


Jet mass distributions after event selection cuts. The QCD background has been reduced by a factor of the order 103 across the jet mass range leaving comparable numbers of background and signal events. The shape of the peak has been nicely preserved.


After all of this, the number of background events ends up being comparable to the number of signal events. In the analysis we considered a centre of mass energy of 10 TeV and 1 fb-1 of integrated luminosity which means that although this is not an early data analysis it may not require vast quantities of data. The success or failure of this analysis will depend upon how well we can understand the underlying background and how smooth it turns out to be, which makes understanding the detector and the QCD background a very crucial part of this analysis.

On a personal note, although a good proportion of our time spent in Boston before the SUSY09 conference was spent making final checks of the analysis before presenting it, we still found time to take in some of the sights and sounds. We never felt too far from home with the 'ATLAS Convenience Store' located just a minutes walk away from the hotel, the 'Athena Spa' only five minutes away and the 'other' Cambridge a short subway ride away. We look forward to the switch on of the Large Hadron Collider and the many challenges and puzzles its data will bring.

 

 

 

Sky French

University of Cambridge