1. Higgs Searches at the Tevatron

2. Outline Introduction Standard Model (SM) Higgs Non-SM Higgs
Tevatron
Standard Model (SM) Higgs
Results
Low & high mass
Combination
Non-SM Higgs
Prospects & Conclusions
(See parallel sessions for full details)
[ Thanks to all my Tevatron colleagues ]
Gavin Davies – SUSY08

3. Many thanks to Accelerator Division
Tevatron Performance
Over 4fb-1 delivered per experiment. Over 0.2fb-1 recorded last month.
Expect ~6fb-1 by Apr 2009, and 8fb-1 by Oct 2010.
Average data-taking efficiency > 85%
Results presented here use up to 2.4fb-1
(3fb-1 results at ICHEP)
Many thanks to Accelerator Division
05/ / / /08
Gavin Davies – SUSY08

4. Standard Model Higgs Higgs mechanism Direct searches at LEP2
Additional scalar field in SM Lagrangian
 mass to W,Z & fermions
Predicts neutral, spin 0 boson
But not its mass
Direct searches at LEP2
mH > GeV
Improved mt & mw tighten indirect
constraints:
mH < 160 GeV (EW fit)
mH < 190 GeV if LEP2 limit included
 A light (SUSY?) Higgs is favoured
Gavin Davies – SUSY08

5. SM Higgs Production & Decay
Small production cross-sections pb
Branching ratio dictates search
mH < 135 GeV
gg → H → bb overwhelmed by multijet (QCD) background
Associated WH & ZH production
with H → bb decay
Main backgrounds: W+bb, Z+bb,
W/Z jj, top, di-boson, QCD
mH > 135 GeV
gg → H → WW
Main background: WW
Excluded
Gavin Davies – SUSY08

6. SM Higgs Topologies Low mass: Intermediate mass: High mass:
Leptonic decay of W / Z boson
provides ‘handle’ for event.
H  bb reduce SM background
High PT opposite sign leptons.
Missing ET
Gavin Davies – SUSY08

7. SM Higgs - snapshot Higgs searches at a hadron machine challenging
Requires
Efficient triggering, lepton ID,
b-tagging, jet resolution..
Sizeable data-sets
Well established
Analyses use multivariate techniques
Neural net (NN), matrix element (ME)..
Regular Tevatron combinations
Approaching the SM expectation
Rapid improvement well beyond √L gain
Even better to come
Gavin Davies – SUSY08

8. Highest associated production cross-section Selection Backgrounds
Use electron and muon channels
Selection
Isolated lepton, missing ET (MET), 2 jets
Backgrounds
W+jets, QCD, top, di-boson
Analyses
Separate channels: 1 &
2 b-tags, e or m, central
or forward
NN (and ME) based analyses
using kinematic variables
Gavin Davies – SUSY08

9. Cross section limits derived from NN distributions
Compare observed (obs) or
expected (exp) limit to SM prediction
Recent CDF analysis
Including isolated tracks
Increases acceptance by 25%
mH=115GeV
Lum
Obs/SM
Exp/SM DØ
1.7 fb-1
10.9
8.9
CDF
1.9 fb-1
8.2
7.3
Exp/SM = 6.4
(14% improvement)
Gavin Davies – SUSY08

10. Large cross-section & acceptance but hard Selection Backgrounds
Two jets, MET (not aligned in f with jets)
Backgrounds
Physics: W/Z + jets, di-boson, top – from Monte Carlo
Instrumental: Mis-measured MET with QCD jets – from data
Analyses
Optimised b-tagging, NN or decision tree (DT) discriminant
Exp/SM ~8 (both expts.)
Gavin Davies – SUSY08

11. And…. Not forgetting .. ZH → llbb
1 and 2 b-tags, NN event selection
Exp/SM ~ at 115GeV (1fb-1)
Sensitivity of existing channels rapidly improving
Add additional channels DØ
Exp/SM ~ 45 at 120GeV
CDF
Exp/SM ~ 40 at 120GeV
Exp/SM ~ 25 at 115GeV
Adds 10% to CDF combination (≡ 20% more data)
Gavin Davies – SUSY08

12. Intermediate mass range Selection Backgrounds Analyses
2 like-sign high PT leptons (ee, em, mm)
Backgrounds
WW,WZ, charge flips (from data)
Analyses
DØ 1fb-1 and CDF 1.9fb-1
Exp /SM ~25 at 140GeV
Exp /SM ~20 at 160GeV
Gavin Davies – SUSY08

13. Dominant for mH >135GeV Selection Backgrounds Analyses:
2 isolated, opposite sign leptons, MET
Backgrounds
Drell-Yan, W+jets/QCD, tt, SM WW
Analyses:
WW from spin 0 Higgs
Leptons prefer to point in same direction
Use NN, using kinematic variables and ME as inputs
Gavin Davies – SUSY08

14. Input ME information to NN
Matrix element
Use observed leptons & missing ET (xobs)
Integrate over LO theory predictions
for WW, ZZ, W+g, W+jet, Higgs
Construct LR discriminant from probabilities
Input ME information to NN
CDF – 4 ME’s and 6 kinematic variables
DØ – 1 ME and more kinematic variables
Cross section limit derived from NN output distribution
LR (H  WW) 
Gavin Davies – SUSY08

15. Closing in on the SM Expected Limit / SM ~ 2.5
Observed Limit / SM ~ 2.1
Expected Limit / SM ~ 2.5
Observed Limit / SM ~ 1.6
Closing in on the SM
Gavin Davies – SUSY08

16. Combination Compare background and signal plus
background hypotheses using Poisson
likelihoods
Systematics included in likelihoods
Combine across channels within an
experiment
Combine across experiments
Since 2005 sensitivity improved
by a factor of 1.7 beyond √L gain
Obs/SM
Exp/SM
115GeV
3.7
3.3
160GeV
1.1
1.6
Gavin Davies – SUSY08

17. Non-SM Higgs MSSM f → tt and bf → bbb or bf → btt topologies
2 Higgs doublets, ratio of vev’s = tanb
2 charged (H) and 3 neutral Higgs particles: f = h, H, A
At large tan b
Coupling to d-type fermions enhanced by tanb → s α tan2b
sf  2 sA
For low & intermediate masses
Br (f → bb) ~90%, Br (f → tt) ~10% f b f
f → tt and bf → bbb or bf → btt topologies
Gavin Davies – SUSY08

18. Neutral MSSM Higgs  tt Smaller BR but cleaner
Signal: (eτhad) (μτhad) (eμ) pairs + MET
Background: Z→tt, jet fakes
Use visible mass:
No significant excess in visible mass, so proceed to set limits
Initially on s x Br then interpret in MSSM (FeynHiggs – arXiv: v2)
arXiv:
Gavin Davies – SUSY08

19. Neutral MSSM Higgs  tt MSSM interpretation
Width included
Limits very similar for m<0
Expect to reach tanb~20 at low mA by end 2010
Most powerful constraints on tanb
Gavin Davies – SUSY08

20. Neutral MSSM Higgs  bb + b[b]
Signal
At least 3 b-tagged jets
Look for peak in dijet mass
Challenge - Large multijet background
Estimate from 2 b-tagged data and simulation
Composition: Sec. vertex mass (CDF) & fit to multiple b-tagging criteria (DØ)
No significant excess, so set limits, initially on s x Br then interpret in MSSM
arXiv:
Gavin Davies – SUSY08

21. Neutral MSSM Higgs  bb + b[b]
MSSM interpretation
Width included
Gavin Davies – SUSY08

22. Going further… Charged Higgs And beyond MSSM
Separate t  W+b & t  H+b
(H+  cs) via mass templates
And beyond MSSM
Fermiophobic Higgs
Doubly charged Higgs
arXiv:
arXiv:
HR > 127 GeV
HL > 150 GeV
CDF also looked for e/t final states
Gavin Davies – SUSY08

23. Prospects – SM Higgs Rapid evolution For 2010 expect, beyond √L gain
Significantly better than √L
Increased trigger efficiency, lepton acceptance, improved b-tagging, NN discriminants
For 2010 expect, beyond √L gain
x 1.4 improvement at high mass and x 2.0 improvement at low mass
mH =160GeV
mH =115GeV
Gavin Davies – SUSY08

24. Prospects – SM Higgs (cont)
End 2009
Very limited
allowed range
End 2010
Full exclusion
3 evidence possible
over almost entire
range
Assumes two experiments
Gavin Davies – SUSY08

25. Conclusions New SM analyses keep coming in ICHEP08?
Tevatron and CDF/ DØ experiments performing very well
Over 4fb-1 delivered, with 8fb-1 expected by end of 2010
New SM analyses keep coming in
Sensitivity almost improving linearly
with luminosity
Well established, common effort across
the Collaborations
Closing in on the SM
go beyond at 160GeV ?
Have a clear roadmap to reach desired sensitivity
Wide range of BSM searches, ready for discovery
Expect to reach tanb ~20 at low mA
Exciting times ahead – will only get better!
ICHEP08?
Gavin Davies – SUSY08

26. And… Thank you for your attention
Please see the parallel sessions for full details
Search for the SM Higgs Boson at DØ - Suyong Choi
Search for Charged Higgs Bosons at DØ - Yvonne Peters
Search for Charged Higgs in Top Quark Decays at CDF - Geumbong Yu
Search for BSM Higgs Bosons at DØ - Ingo Torchiani
Gavin Davies – SUSY08

27. Backup slides
Gavin Davies – SUSY08

28. CDF and DØ experiments Both detectors extensively upgraded for Run IIa
New silicon vertex detector
New tracking system
Upgraded muon chambers
CDF: New plug calorimeter & ToF DØ
New solenoid & preshowers
Run IIb: New inner tracking layer & L1 trigger
Gavin Davies – SUSY08

29. DØ data taking
Gavin Davies – SUSY08

30. ZH → llbb
WH → lnbb
H → WW → lvlv
Total
Gavin Davies – SUSY08

31. B-tagging Critical for low mass H bb Use lifetime information
Improves S/B by > 10
Use lifetime information
Correct for MC / data differences
Measured at given operating points
Analyse separately (“tight”)
single & (“loose”) double tags
CDF: Secondary vertex reconstruction
Neural Net - improves purity
Inputs: track multiplicity, pT, vertex
decay length, mass, fit
Loose = 50% eff, 1.5 % mistag
Tight = 40% eff, 0.5 % mistag
DØ: Neural Net tagger
Secondary vertex & dca based inputs,
derived from basic taggers
High efficiency, purity
Loose = 70% eff, 4.5% mistag
Tight = 50% eff, 0.3% mistag
Gavin Davies – SUSY08

32. DØ B-tagging Several mature algorithms used: 3 main categories:
- Soft-lepton tagging
- Impact Parameter based
- Secondary Vertex reconstruction
Gavin Davies – SUSY08

33. B-tagging – (DØ) Certification
Have MC / data differences – particularly at a hadron machine
Measure performance on data
Tag Rate Function (TRF) Parameterized efficiency &
fake-rate as function of pT and η
Use to correct MC b-tagging rate
b and c-efficiencies
Measured using a b-enriched data sample
Fake-rate
Measured using QCD data
Gavin Davies – SUSY08

34. Tau ID CDF: Isolation based
Require 1 or 3 tracks, pT > 1GeV in the isolation cone
For 3 tracks total charge must be ±1
pThad > 15 (20) GeV for 1 (3) prongs
Mhad < 1.8 (2.2) GeV
Reject electrons via E/p cut
Validated via W/Z measurements
Performance
Efficiency ~ 40-50%
Jet to tau fake rate ~
DØ: 3 NN’s for each tau type
Validated via Z’s
Efficiency (NN>0.9):
t ~0.65, jets ~0.02
Gavin Davies – SUSY08

35. Cleanest channel, but low cross section Selection: Backgrounds:
Loose lepton ID, 2 jets
Backgrounds:
Z+jets, top, WZ, ZZ, QCD
Analyses
Constrain dijet mass
NN event selection
Exp/SM ~ 16
Gavin Davies – SUSY08

36. Limit Setting LEP: Low background, small systematics
Tevatron/LHC: High background, large systematics
Systematics, including correlations, taken into account
Main systematics (depending on channel):
- Luminosity and normalisation
- QCD background estimates
- Input background cross-sections
- Jet energy scale and b-tagging
- Lepton identification
- K-factors on W/Z + heavy flavour
Systematic uncertainties included in likelihood via gaussian smearing of expectation (‘profile likelihood’)
Background constrained by maximising profile likelihood (‘sideband fitting’)
Tevatron experiments use LEP CLs (modified frequentist) and Bayesian methods
Limit setting approaches agree to within 10%
Gavin Davies – SUSY08

37. Spring 08 SM Combination Channel Lumi /Technique Final state
WH→lbb
fb—1 / NN+NN
e/, 1b/2b
ZH→ll bb
fb—1 / NN+NN
ZH→ bb
fb—1 / NN+DT
Z→, W→ł (1b/2b)
H→ (gg,VBF,WH,ZH)
2.1 + ….. fb-1 / NN
t + 2 jets
H→
… fb-1 / Di- mass 
H→WW*
fb—1 / NN&ME
ee, e, 
WH→WWW*
… fb—1 /2D LHood
Total of 28 CDF + DØ channels combined
Gavin Davies – SUSY08

38. Spring 08 SM Combination Channel CDF Obs (exp) DØ Obs (exp WH→lbb
8.2 (7.3)
(8.9)
ZH→ll bb
16 (16)
(20.4)
ZH→ bb
8.0 (8.3)
7.5 (8.4)
H→
30.5 (24.8) -
H→
~45
H→WW*
1.6 (2.5)
2.1 (2.4)
WH→WWW*
24 (18)
Gavin Davies – SUSY08

39. Spring 08 SM Combination
Gavin Davies – SUSY08

40. MSSM benchmarks Five additional parameters due to radiative correction
MSUSY (parameterizes squark, gaugino masses)
Xt (related to the trilinear coupling At → stop mixing)
M (gaugino mass term)
 (Higgs mass parameter)
Mgluino (comes in via loops)
Two common benchmarks
Max-mixing - Higgs boson mass
mh close to max possible value
for a given tan
No-mixing - vanishing mixing in
stop sector → small mass
for h
Gavin Davies – SUSY08

41. Neutral MSSM Higgs  tt Limits: s x Br (f tt) Evolution: Width:
Full 2010 dataset (2 expts.)
Width:
Gavin Davies – SUSY08