• ytest250_wjet.eps: Test of y distributions include SUSY and W+Jet SM events

Things to note: There is a 250 GeV cut applied to all jets here, but nevertheless the transverse momentum distributions will be different. What I expected was:

• Scale for W splitting depends on the W mass so doesn't care what process the W comes from, or what the PT of the W is. This seems to be largely true, from the top plot.
• Scale for QCD jet splitting depends on the PT of the jet, and so can be rather different for QCD jets in SUSY events (which I think are generally higher PT). So I understand the top plot.
• y cut I would expect to be the same for QCD jets wherever they come from (modulo logarithmic corrections for running alpha_s). The differences between QCD jets in SUSY and W+jet look quite large to me, but maybe this is consistent with scaling. I might look into this further, maybe by reweighting the PT distributions and seeing if the y is then equal.
• I would have expected the y cut for the W's to be more different than it is. Perhaps it is so similar because both distributions are peaked around the cutoff at 250 GeV.

Some ideas to try on refining the selection...

• run the kt algorithm with R parameter < 1 (0.7 and 0.5 are two common values)? This may improve the W mass resolution but at some point the W's will stop being in a single jet and the method breaks.

• ytest250_wjet.eps: Test of y distributions include SUSY and W+Jet SM events

Are's initial attempts

Are's attempt at reproduction

• ytest250a.eps: Test of y distributions

• ytest250a.eps: Test of y variable

• ytesta.eps: Test of y variable

• ytesta.eps: Test of y variable

• ytesta.eps: Test of y distributions

Attempt at reproduction

I've trfied to reproduce John's test and failed miserably. My corresponding plots are found here:

• ytest250.eps: ytest250.eps

• ytest250.eps: ytest250.eps

* ytest.eps: Test of the y cut for W and other jets.

They show a marked difference (as expected) in both distributions, with the W jets peaking near to the W mass scale.

• ytest.eps: Test of the y cut for W and other jets.

Some tests from Jon

This is encouraging but not comparable directly to Are's plots since I haven't made any cuts on the jets at all and I can't yet pick out jets from squark decays (these will be included in the "QCD jet" sample at present). Nevertheless these are probably easy plots for you to reproduce Are? If we agree on these, that would at least remove one possible source of bugs.

I will look into this as a function of some other cuts (pT etc).

• deltahiggs.ps: Plot for higgs at \delta

• pTvsDeltaR.ps: pT of W versus \Delta R between quark decay products

I made another attempt by trying to find a Higgs in the decay chain ~q_L -> q + ~neutralino_2 -> q + h + ~neutralino_1, throught its decay to a bbbar pair. The conclusions are the same. While the jet pT and mass cuts work, the separation scale cut doesn't. You end up with the expected signal distribution, but with too much background. The relevant plots, following the same pattern as for the W is also attached. Here there might however be some more hope. No attempt at b-tagging the jet was made, which might give enought rejection on the background to see something.

• delta.ps: Some plots for \delta

• Top right: Plot of jet mass m_W as given from jet algorithm. Same color coding. Cut on jet pT: 150 < pT < 600.
• Bottom left: Plot of jet log(pT*sqrt(y)). Same color coding. Cut on jet pT: 150 < pT < 600. Cut on jet mass: 80 < m_W < 90.

As you can see there is no longer a top peak as in the first version, so this is no longer a problem. There are still too many background jets, and rather surprisingly the separation scale cut does not work very well. I

A short description of the plots follow:

• Top left: Plot of pT of jets. Cut on pT>15 and |\eta| < 2.5. In red is shown jets matched to a parton level W (by the usual distance \Delta R < 0.1).
• Top right: Plot of jet mass m_W as given from jet algorithm. Same color coding. Cut on jet pT: 150 GeV < pT < 600 GeV.
• Bottom left: Plot of jet log(pT*sqrt(y)). Same color coding. Cut on jet pT: 150 GeV < pT < 600 GeV. Cut on jet mass: 80 GeV < m_W < 90 GeV.
• Bottom right: Plot of invariant mass of w+jet system. We require two W's identified in each event from the cuts given above pluss a cut on separation scale: 1.3 < log(pT*sqrt(y)) < 2.0. The W candidates are combined with the two hardest jets in the event. In red (signal) events with two decays of the type mentioned above.

• Top left: pT of jets with pT>15. In red is shown jets matched to a parton level W (by the usual distance \Delta R).
• Top right: Same color coding. pTsqrt(y).
• Bottom left: Same color coding. Mass of jet as given from jet algorithm.
• Bottom right: In red (signal) events with two decays of the type q_L -> q + chi+-1 -> qW chi01. Invariant mass of w+jet after cuts demanding two jets with pT>150, 1.2<y<2, 70<m_j<110. The additional jet is picked from the two hardes jets of the event.

We have discrimination between the W's and other jets as seen from the first three plots, but the invariant mass plot shows there are too many W candidates around after cuts. Note the a part of the problem is the copious production of top quarks in this model, which can be seen as peaks in the invariant mass plot. I'm still working on the numbers for how efficient the cuts are and how much of the problem is caused by the real W's produced in other decays, and how much is being faked by other jets.

• delta.ps: Som start-up plots for \delta

• delta.xsec: Cross sections at NLO for \delta calculated with Prospino2

• delta_slha.dat: SLHA file for the \delta benchmark

• delta_slha.dat: SLHA (SUSY Les Houches Accord) file for the \delta benchmark point

Identifying massive hadronic decays in SUSY decay chains.

This page is intended to help discussion between Are, Erckan, John and Jon and isn't intended to be used by anyone else.

Project plan

We'll start by assuming that in a SUSY pair production, one decay chain has allowed us to remove all non-SUSY backgrounds, and what we are trying to do is interrogate/unpick the second decay chain to learn more about the SUSY spectrum. In particular we are looking at high ET jets and seeing they come from hadronic decays of W or Z.

John I suggest that we use the benchmarks postulated in hep-ph/0508198, particularly scenarios \alpha, \beta, \gamma and \delta. The first three have neutralino dark matter but non-universal Higgs mases (NUHM), with parameters chosen to get varying branching ratios for \chi_2^0 \to Z,H \chi_1^0, as seen in Fig. 2. I do not know offhand about the \chi^\pm \to W \chi decays.

The point \delta has gravitino dark matter, but the neutralino is the next-lightest sparticle and its decays are invisible. Hence, all four points have 'standard' missing energy signatures.

Since there are issues in treating NUHM models using ISAJET, we might want to start with point \delta. It is in any case the most interesting, as it has relatively large masses and hence is quite challenging using conventional techniques (see Fig. 9).

After our discussion at CERN on 19 July, we decided the immediate way ahead was as follows:

* Are to produce the PYTHIA config files from these and upload to this page. * Are to run the [http://hepforge.cedar.ac.uk/hzsteer HZSteer] analysis and show us the key distributions (mass, y, ET, eta etc... as shown previously). * Erckan to run similar jobs with Athena ATLFast and compare to Are's.

After this, we should have a good idea how much mileage there is in this kind of study, and hopefully be in a position to start looking at more challenging SUSY scenarios and treating non-SUSY background events.

-- JonButterworth - 19 Jul 2006

-- ErkcanOzcan