Study of Rare Top Decays at the Tevatron

1
Study of Rare Top Decays at the Tevatron

• John Strologas
• University of New Mexico
• for the CDF and D0 collaborations

2
Does top always decay to bottom?

• According to the SM, the top quark almost always
  decays to a b quark
• B(t?Wb)1
• Most of the SM rare decays of the top are really
  rare
• B(t?Ws)lt0.18, B(t?Wd)lt0.02 (these are the
  larger ones!!)
• An observation of a B(t?Wb) considerably
  different than unity will be an indication of new
  physics
• Non-SM top decay
• Non-SM background to top decay
• Fourth generation
• ???

3
Current Rare Top Decay Results from the
Tevatron

• The RB(t ? Wb)/B(t ? Wq) measurement
• Run II results (CDF and D0)
• The t ? Hb
• Run II result (CDF)
• FCNC t ? Zq or t ? ?q
• Run I result (CDF)

4
The Detectors
D0
CDF
5
The ratio RB(t?Wb)/B(t?Wq)

• The top quark decays to a b-quark almost always,
  given
• The unitarity of CKM matrix with 3 flavors
• The small measured values of Vub and Vcb
• Vtb 0.999
• R Vtb2 / (Vtd2 Vts2 Vtb2) Vtb2
  99.8
• up to phase space factors, given 3 flavors (SM)
• Any significant deviation from R1, would be an
  indication of new physics!

6
How to measure RB(t?Wb)/B(t?Wq)?

• Just count the events with b-tagged jets (jets
  that are associated with b-quarks)
• The number of b-tagged jets we expect to see
  from ttbar decays depends on
• R (if low, fewer b-jets are produced)
• The tagging efficiency (if low, fewer b-jets are
  tagged)
• We classify the ttbar based on the number of
  b-tagged jets
• The relative rates of events with 0/1/2 b-tags is
  more sensitive to R

7
Simple relation between R and
tag-multiplicity
Assuming zero backgrounds and only b-tagging
In reality the relation is more involved and a
likelihood is used
8
Measurement of the ratioRB(t?Wb)/B(t?Wq) at D0
Run II

• Study tt ? Wq Wq ? l?q qqq (leptonjets
  events)
• Integrated luminosity of 169 pb-1 (ejets) and
  158 pb-1 (µjets)
• Isolated ETe(PTµ)gt20 GeV, METgt20(17) GeV,
  Df(???,lepton) cut.
• 3-jet and gt4-jet subsets are considered
• Two methods of b-tagging used
• CSIP (Counting Signed Impact Parameter)
• SVT (Secondary Vertex Tagger)
• The probability to observe n-tags is calculated
  for three possible decay modes of the t-tbar
  pair
• tt ? Wb Wb
• tt ? Wq Wb (where q is a non-b quark)
• tt ? Wq Wq

9
Measurement of the ratioRB(t?Wb)/B(t?Wq) at D0
Run II

• The overall probability to observe n b-tags in an
  event
• Pn-tag R2Pn-tag(tt?WbWb) 2R(1-R)Pn-tag(tt?WqWb
  ) (1-R)2Pn-tag(tt?WqWq)
• From that, calculate the expected number of
  events in 8 samples (e/µ, 3-jet/4-jet,
  1-tag/2-tag), which is a function of an input stt
• Construct a likelihood, consisting mainly of
  Poissons for the 8 samples. The stt is a floating
  input to the likelihood)
• Maximize the likelihood to extract R.

10
Measurement of the ratioRB(t?Wb)/B(t?Wq) at D0
Run II
Use observed events
(stt is an input)
Maximize Likelihood (for R and stt simultaneously)
Measure R
11
Measurement of the ratioRB(t?Wb)/B(t?Wq) at D0
Run II

• 68 and 90 CL contours in the (R,stt) phase space

CSIP
SVT
Central measurement
Central measurement
12
Measurement of the ratioRB(t?Wb)/B(t?Wq) at CDF
Run II

• Integrated luminosity of 162 pb-1
• At CDF we study both tt ? Wq Wq ? l?q qqq
  (leptonjets) and tt ? Wq Wq ? l?q l?q
  (dilepton) events
• Use SVX b-tagging (separate 0-tag, 1-tag and
  2-tag sets)
• stt independent measurement
• Leptonjets set require
• Isolated lepton with ETe(PTµ)gt20 GeV, METgt20 GeV
  and at least 4 jets with ETgt15 GeV
• Dilepton set require
• At least two leptons (ee, µµ, eµ) with
  ETe(PTµ)gt20 GeV, METgt20 GeV, and at least two
  jets with ETgt15 GeV.
• Greater statistical significance comes from the
  leptonjets sample

13
Measurement of the ratioRB(t?Wb)/B(t?Wq) at CDF
Run II

• At CDF, we use the 0-tag sample as well to
  further constrain R.
• This means that we have to measure the
  top-content in a sample that has no b-tags !
• We do that by utilizing a Neural Network (NN), to
  measure Ntop(0-tag)
• The QCD background is independently estimated
• We have also NN measurements of Ntop(1-tag) and
  Ntop(2-tag), but the statistics are not that
  great. We prefer to use an a-priori method
  (based on MC normalized to the leptonjets data)
  to estimate the 1-tag and 2-tag backgrounds

14
Measuring the n-tag top content with a NN at CDF
Run II
1-tag
0-tag
2-tag
15
Measurement of the ratioRB(t?Wb)/B(t?Wq) at CDF
Run II

• We first determine the b- c- and q- jet tagging
  efficiencies eb, ec and eq , defined
  as ( tagged jets/ taggable jets)
• using MC and correcting with scale factors
• We then determine the fraction of MC events with
  i-taggable b-jets, j-taggable c-jets and
  k-taggable ql-jets
• From the above, using combinatorics we determine
  the efficiency to have 0, 1 or 2 tags in a
  particular top event.
• We explicitly set the tagging efficiency for a
  jet coming from a top to ebR (1-R)eq
• Multiply the efficiency to the expected top
  events, given the estimated background, to get
  the expected top content in 0/1/2 tags.
• Compare the expected top with the observed top in
  the 0/1/2 tag subsets and extract R by maximizing
  the likelihood.

16
Measurement of the ratioRB(t?Wb)/B(t?Wq) at CDF
Run II
Calculate expected events as a function of R
Compare to observed and
Maximize Likelihood
Measure R
Set FC lower limit
, assuming 3 generations
17
Top Decay to a Charged Higgs

• If we assume two Higgs doublets, then EWK
  symmetry breaking produces 5 Higgs fields, three
  neutral and two charged.
• The top quark will couple to H if mt gt mHmb
• B(t?Hb) (mt2cotß mb2tanß) 4mt2mb2 at tree
  level
• tanß is the ratio of vev for the two Higgs
  doublets
• The coupling of top to H will be strong, if
  tanßgtgtsqrt(mt/mb) or tanßltltsqrt(mt/mb)
• If tanß is low
• H?cs is the dominant decay
• Unless the mH is high enough to dominantly decay
  as H ? tb ? Wbb
• If tanß is high
• H?t? is the dominant decay

18
H/top branching ratios
19
Search for t?Hb at CDF Run II

• Luminosity of 193 pb-1 Tree level analysis
• Utilizing leptonjets, dilepton, and leptonthad
  top cross section analyses
• For the leptonthad sample require
• An electron(muon) with ET(pT)gt20 GeV and also
  METgt20 GeV
• t cuts (track requirements in a jet, calorimetry
  e/µ vetos)
• Zveto, HTgt205, gt 2 jets
• t charge (determined from the tracks) opposite of
  that of e or µ
• Calculate the estimate number of top events
  decaying to H, with the charged Higgs decaying
  to any of the three modes.

20
Search for t?Hb at CDF Run II(tree-level
analysis)
Expected sensitivity (for expected 11 dilepton,
66 leptonjets and 2 leptont events)
21
Search for t?Hb at CDF Run II(tree-level
analysis)
Paremeterizing the likelihood as a function of BR
t?Hb, for tau final states
22
Search for t?Hb(Run I / Run II comparison)
23
Flavor changing neutral currentsFCNC t ? Zq or t
? ?q

• FCNC at tree level are forbidden by the SM
• always cancel if left-handed fermions appear in
  iso-weak doublets
• They are allowed in second-order processes, like
  penguin diagrams
• SM rate O(10-12). Any observation of top FCNC
  would be a strong indication of new physics.

24
Searching for t ? ?q and t ? Zq at CDF Run I

• Run I analysis, 110 pb-1
• F. Abe et al. (CDF), Phys. Rev. Lett. 80, 2525
  (1988)
• Normalization sample of leptonjets top
  candidates
• An electron(muon) with ET (pT) gt 20 GeV
• METgt20 GeV
• At least three jets with ETgt15 GeV
• 34 t-tbar candidates with an estimated
  background of 9 /- 1.5 in our data
• ISAJET MC is used for the calculation of relative
  acceptances (FCNC/leptonjets)

25
Searching for t ? ?q at CDF Run I

• In the search for t ? ?q, we assume that the
  other top quark decayed to Wb

• If the W decayed hadronically
• gt4 jets with ETgt15 GeV
• A photon with ETgt50 GeV,
• b-tag of a jet related to top decay
• a photon-jet mass consistent with a top (140-210)
• The rest of the jets should have total ETgt140 GeV
  (consistent with a top)
• This channel carries 40 of our acceptance

• If the W decayed leptonically
• A lepton with ET(pT)gt20 GeV and METgt20 GeV
• gt 2 jets with ETgt15 GeV
• A photon with ETgt20 GeV
• The jets should have total ETgt140 GeV (consistent
  with a top)
• This channel carries 60 of our acceptance

• Background of 0.5 events expected in both
  hadronic and leptonic channels.

26
Searching for t ? Zq at CDF Run I

• In the search for t ? Zq, we assume that the
  other top quark decayed to Wb
• We require 2 electrons or 2 muons, 4 jets with at
  ETgt20 GeV and dilepton mass between 75 and 105
  GeV
• Expected background is 1.2 events (Zjets (0.5),
  residual dilepton-t-tbar (0.6), diboson(0.1))

27
Relative Run I acceptances
28
t ? ?q and t ? Zq CDF Run I limits

• Do set conservative limits, the backgrounds are
  not subtracted
• We see one event in the leptonic t ? ?q sample
• Kinematically consistent with radiative t-tbar
  leptonjets
• B(t ? ?c) B(t ? ?u)lt3.2 at 95 CL
• We also see a dimuon t ? Zq event
• Kinematically consistent with Zjets
• B(t ? Zc) B(t ? Zu)lt33 at 95 CL

29
FUTURE Vtb reach (CDF) and
Charged Higgs/FCNC sensitivity (LHC)

• LHC Sensitivity (100 fb-1)
• B(t ? ?q) 10-4
• B(t ? Zq) 10-4
• B(t ? Hq) 5 10-4
• (ATLAS studies)

CDF II
Assuming R1 and 3 generations (same analysis)
30
Conclusions and Plans

• Status of rare top decays at the Tevatron
• New Run II RB(t?Wb)/B(t?Wq) results
• New Run II t?Hb result
• Run I FCNC result
• Current D0 analyses/plans
• New RB(t?Wb)/B(t?Wq) under review
• H search using s(tt ? dileptons)/s(tt ?
  leptonjets)
• Hope to start a FCNC analysis in the near future
• Current CDF analyses/plans
• Improved charged Higgs analysis
• Inclusion of all QCD, SUSY-EWK, SUSY-QCD
  corrections
• Separate 1-tag and 2-tag leptonjets analyses
  (more sensitivity in the low tanß)
• the analysis is close to blessing
• Top FCNC analysis in the t?Zq sector
• the analysis just started

31
Back-up
32
CDF charged Higgs analysis implicit assumptions

• There are at least four important assumptions
  implicitly taken in the method
• The tt production cross section is not affected
  by the inclusion of the MSSM.Claimed by CDF. No
  reason against that.
• Is the background in the XS measurements affected
  by the inclusion of the MSSM ? Those processes
  involving SUSY particles are ignored here.The
  Higgs sector is considered ahead.
• The efficiencies ei,j do not depend in MSSM
  parameters.This can be shown by analyzing the
  decay topologies and MSSM coupling constants.
• Other H decays, besides the three final states
  mentioned, have negligible branching ratios.True
  for large fraction of MSSM parameter space.
• Q Do the width of top and Higgs modify the
  efficiencies ?Yes, slightly, but they are
  corrected for that in the method.

33
H/top widths
34
Contributions to the Posterior probability
density(three charged Higgs CDF analyses)
35
D0 RunI Charged Higgs analyses

• Direct PRL 88, 151803 (2002)
• 62.2 pb-1 (multijetMET trigger)
• H ? t?
• Loose selection METgt25 GeV,
  gt4 jets with
  ETgt20 GeV
  (lt8 with ETgt8 GeV)
• Use of a neural network cut
• Background 5.2 /- 1.6 (observed 3)
• Indirect PRL 82, 4975 (1999)
• 110 pb-1
• Leptonjets ET(pT)gt20 GeV,
  METgt25 GeV, gt4 jets
  (ETgt15 GeV),
  HTgt180 GeV
• Leptonjetsµ-tagged jet ET(pT)gt20 GeV,
  METgt20 GeV, gt3
  jets (ETgt20 GeV),
  HTgt110 GeV
• Background 30.9 /- 4 (Observed 30)
  

36
CDF RunI Charged Higgs analysis

• Direct PRL 79, 357 (1997)
• 100 pb-1
• H ? t?
• Hadronic tau cuts, with ETgt20 GeV
• for 1 tau, or ETgt30 GeV for 2 taus.
• METgt30 GeV
• Z veto
• Background of 7.4 /- 2
• (7 events observed)
  
  
  

37
NN for the top content measurement(CDF R
measurement)
9 input Variables
, 10 hidden nodes
1 Output
Ranked based on KS significance
38
CDF top event tagging efficiencies(efficiencies
to tag 0/1/2-jets in a top event)
Fijk are the fraction of MC top events with
i-taggable q-jets, j-taggable c-jets and
k-taggable b-jets
eb,, ec, eq are the single jet tagging
efficiencies, defined as (tagged/taggable)
39
QCD background size estimation

• We estimate the QCD background using the ISO vs
  MET scatter plot of the actual data

• We define 4 regions
• Signal region METgt20 GeV ISOlt0.1
• A METlt10 GeV ISOlt0.1
• BMETlt10 GeV ISOgt0.2
• C METgt20 GeV ISOgt0.2
• The estimated QCD background fraction in the
  signal region is NANC/(NBNsignal)