Data:
Run II, 350 pb-1 (ele+tau). 322 pb-1 (muon+tau)
- Overview
- Search Strategy
- Backgrounds
- Systematic Uncertainties
- Limits
- Event failing MLS
This analysis searches for doubly-charged Higgs H++ bosons that occur in models that have Higgs triplets. In particular, left-right symmetric theories posit the existance of an SU(2)R which is broken at much higher energies than the electroweak scale. There are two kinds of H++: those that couple to left-handed particles (HL++) and those that couple to right-handed particles (HR++) corresponding to I3 states of 1 and 0 respectively. The (HL++) has a ~50% higher production cross section, due to the terms involving I3 in the matrix element associated with Higgs H++ prodcution. The H++ are pair produced at CDF, with the only significant mode being through a Z/gamma* in the s-channel. At masses below ~160 GeV/c2 The H++ couple to leptons only, though there is no theoretical restriction on the lepton flavor. We search for the lepton flavor violating decays H++->e+tau+, and H++->mu+tau+. In both cases, we assume exclusive decays to the desired channel. We also assume that the coupling of H++ to the leptons is strong enough such that the decay occurs within ~10 microns of the production vertex.
Previous H++ searches have looked for a like-sign dilepton signal. Our signal involves tau leptons, and is subject to higher hadronic backgrounds and poorer mass-resolution than those signals the feature prompt, like-sign electrons and muons. In order to suppress backgrounds like W+jets (e or mu from W, tauh faked from jet), three or four leptons are considered only. This results in lower acceptance, but strongly suppresses most Standard Model (SM) backgrounds. The data for the ele+tau search are collected with 'lepton+track' triggers, which require one good central electron or muon with PT > 8 GeV/c, and one central track with PT > 5 GeV/c and track isolation appropriate for a hadronic tau.
The lepton requirements for the ele+tau search are:
10 GeV lepton of any flavor (No hard cut on |eta|, track must pass through central tracking changer.)
If three leptons are found, we look for a fourth lepton of any flavor with ET>10GeV. If such a lepton is found, we call this event a 'four-track' event. If no such lepton is found, we call this event a 'three-track' event. The three-track and four-track channels are investigated separately, as different amounts and types of SM backgrounds apply to each signature.
In both the three-track and four-track regions, cuts are made to suppress Z->ll type events. In the four-track channel, this is all that is necessary, in the three-track channel, a cut on HT (HT=Sum ET of leptons plus MET) controls backgrounds. We look in the mass window 30 GeV/c2 < m(H++) < 125 GeV/c2 for the signal. The mass window is large enough to be fully efficient, even though during mass reconstruction, there is no effort to account for the neutrinos associated with tau decay. Based on the theoretical production cross section, the expected mass limit in the absence of background events is ~113.6GeV.
Most backgrounds are estimated with Pythia MC samples. The QCD, photon+jets, and cosmic ray backgrounds are determined from data, and found to be insignificant (~0.1 +/- 0.1 events) for both the muon+tau and electron+tau analyses. Estimates for W+Jets are taken from Pythia MC. The MC is normalized to data in a W-enriched control region. This is to account for any discrepancies in jet to tau fake-rates as well as NLO cross section. W+jets yields 0+/-0.15 events in both ele+tau and muon+tau analyses.
For both the electron+tau and muon+tau searches, Z+jets is the most significant background. The jet(s) can fake a hadronic tau, while the decay leptons from the Z supply the other two leptons. Events that have oppositely charged, same-flavor leptons which make an invariant mass in the Z's mass range (71-111) are suppressed by the requiremnt that such events have HT > 350 GeV. This HT rejects most Z-based events, yet retains some signal. Top and diboson estimates are also derived from MC. The main source of backgrounds in the four-track region is ZZ->leptons, and contributes about 0.1 event in both e+tau and mu+tau searches. The total SM backgrounds (combined for 3-track and 4-track channels) are 0.29 and 0.4 events for the e+tau and mu+tau channels respectively. Plots for some control regions of the '3-track' regions are posted below.
Electron+Tau analysis: Control Region with Isolation Cuts only applied to leading electron.
l
Muon+Tau analysis: Control Region with Isolation Cuts only applied to leading muon.
Electron+Tau analysis: HT distribution after all cuts.
Muon+Tau analysis: HT distribution after all cuts.
The systematic errors on acceptance are driven by a 1.5% systematic on the efficiency of the lepton track isolation cuts in data and MC. Since the main reason a lepton fails track isolation is due to 'underlying event', it is assumed that the track isolation efficiencies for all leptons in the event are maximally correlated, thus the error is additive. Correlations between tthe 3-track' and '4-track' signatures are accounted for (for example, 'more' ISR results in lower track isolation efficiencies, some events which previoulsly had 4 isolated tracks will have 3 isolated tracks in the deviated 'more ISR' sample. The 4-track efficiency drops, but there is partial compensation as some of the 'lost' events pass all event-level cuts associated with 3-track events.).
The systematic errors associated with PDF's are estimated using the 'reweighting' method. The errors on acceptance due to PDF's are small, presumably because any change in the kinematics of the H++ themselves have only a secondary influence on the kinematics of the final-state leptons, which are dominated by the isotropic decays of the massive doubly-charged higgs'.
Ele+Tau analysis systematics tally.
PS or
GIF,
Muon+Tau analysis systematics tally.
PS or
GIF,
We observed no events in either the three-track or four-track signatures for both ele+tau and muon+tau searches. Based on the level of backgrounds, this was the most likey outcome, with a probabilities of ~65% and ~79% for the mu+tau and e+tau searches respectively.
Limits are calculated using a Baysean method with a flat prior for signal cross section and Gaussian priors for uncertainties on signal and background acceptance and integrated luminosity. The 3-track and 4-track channels are treated as separate measurements, taking into account correlated systematic uncertainties. In the calculation of mass limits, we use NLO (7.5%) and PDF uncertainties for the signal cross section.
The mass limit obtained on the Double Charged Higgs
EPS or
GIF,
for the process is 113.6/112.1 GeV assuming exclusive ele+tau and muo+tau decays respectively.
In the muon+tau analysis, one event passes all cuts except for the like sign (LS) mass window. Here is a table of expected and observed numbers of events for various ranges of HT and LS Mass (MLS). The center column corresponds to the signal region. The left-most column is for the low MLS region, where the event lies.
| | Expected MLS<30 | Observed MLS<30 | Expected 30LS<125 | Observed 30LS<125 | Expected MLS>125 | Observed MLS>125 |
| All HT | 1.9+/-0.6 | 3 | 6.56+/-0.84 | 7 | 0.26+/-0.2 | 0 |
| HT>190 | 0.05+/-0.07 | 1 (HT=202) | 0.26+/-0.24 | 0 | 0.05+/-0.05 | 0 |
The following table describes the leptons in the event.
| | Charge | TIso (GeV/c) | ET(GeV) | Eta | Phi | Relative Error on Track Curvature | Si Hits | Lepton Type | Cal ET (GeV) | Trk PT (GeV/c) | Vis Mass (GeV/c2) | Pi0 Et (GeV) | N Pi0's | Emfr |
| Leading Muon | -1 | 0 | 101 | 0.87 | 0.82 | 0.04 | 3 | Tight muon | N/A | 101 | N/A | N/A | N/A | N/A |
| Leading Tau | 1 | 1.22 | 58 | -0.13 | -2.67 | - | 5 | 1-Prong Tau | 58 | 29 | 1.4 | 19 | 0.46 | 2 |
| Second Tau | 1 | 0.25 | 10.6 | -0.66 | -1.96 | - | 6 | 1-Prong Tau | 10.6 | 6.2 | 0.49 | 1 | 0.47 | 1 |
| METCorr | N/A | N/A | 34 | N/A | -1.63 | - | - | - | - | - | - | - | - | - |
Mass(Mu-,Tauleading+)=171
Mass(Mu-,Tausecond+)=85
Mass(Tauleading+,Tausecond+)=22
Mass(Mu-,METCorr)=110
No Displaced Vertex Tag (as for B jets )
No Tagged Conversions
There is a secondary vertex with a couple of 500MeV/c tracks.
COT view: All trks w/ PT>500 MeV/c are shown.
PS or
GIF,
LEGO view: Calorimeter E.
PS or
GIF,
RZ view.
PS or
GIF,
This event is kinematically compatible with W+Jets events. It is rather incompatible with most other SM processes.
Here is a table of expected contributions from various SM processes.
| Background | Expected Number of Events |
| W+jets | 0.03+/-0.03 |
| ZW | 0.01+/-0.0001 |
| ZZ | 0.01+/-0.01 |
| Z+jets | 0.0+/-0.06 |
Given the backgrounds and errors, the probability of observing this event is 5.6%.
We looked in the same regions for the e+tau analysis, and observed 0 events. The background expectations were similar the ones shown above.
Page page:
http://fcdfhome.fnal.gov/usr/baroiant/HPlusPlus/
was last updated April 28th, 2006 by Sasha Baroiant.
Contact,
Legal Notices, and
Fermilab Policy
on Computing
