Siena 7th Topical Seminar: The Higgs in two jet modes

1
Siena 7th Topical Seminar The Higgs in two jet
modes

• W. J. Murray
• RAL

Standard Model results
2
LEP Accelerator
Thanks to Steve Myers and friends

• Very clean environment
• Excellent performance!

2.5 fb-1 _at_ Ecmgt189 GeV
3
Higgs Searches at LEP 1

• These were typically two-jet modes
• Many modes
• Stable,gg,ee,mm,pp,tt,bb
• Clean Z decays (ll, nn) used
• Prior to LEP only some patchy constraints
• The mass range to 0 now excluded, no holes.

Events expected at LEP1
0.0 ? mH ? 65 GeV/c2 Excluded at 95 C.L.
4
Closing in on the Higgs!
EW fits assume a Higgs Search looks for one
Bayesian
Frequentist
After A. Wagner, ICHEP 2000
5
Total LEP Higgs
Predicted LEP events, 4 experiments
17 _at_ 115GeV
6
LEP Higgs production HZ
Higgstrahlung process
Make a Higgs and a Z together So need Energy
greater than Higgs mass plus Z mass
7
Higgs and Z decay channels
Assuming Both Z H are made!
Z decay modes
Higgs decay modes
WW
nn
B quarks
Any quark
Marumi described green x green I get the
odds and ends!
8
Fusion WW and ZZ

• These diagrams allow non-resonant Higgs
  production
• WW always much larger than ZZ
• Has been suggested as way to beat kinematic
  threshold
• (Ebeam-Mz)

WW
n
e
Any quark
H
W
n
e-
9
Higgs ?? cross-section
fb
?s192GeV
WW
Any quark
MH

• Cross-section few fb at 115
• Not accessible with ½ fb-1 per experiment

10
LEP Higgs channels
Higgs decays, 115GeV
This talk
Search channels
11
How is the Higgs search done?

• Classify events into one category, pre-select
• Assign measured particles to four initial
  fermions quark, e/m, t, n
• tag b quark jets, mvtxleptons
• Decide which fermions are from Z, which from H
• Not obvious in t cases
• Require compatibility with Z mass for Z pair
• Fit for MH, assuming E, p conserved, Mz at least
  for Hnn channel

12
Higgs backgrounds
13
Channel comparison at 115GeV
14
Channel sensitivity (3s evidence)
Probability of 3s assuming MH 2 jets and 4 jets
comparable Need to combine all
45
20
15
B-tagging crucial step

LOG SCALE

• Tag of bs reduces many backgrounds
• Semi-leptonic Ws dangerous
• Mis-tag is well controlled

b c uds
FACTOR gt500 IN WW REJECTION REACHED
16
ZZ cross-checks

• B-tag important for ZZ selection
• Proof that this very similar channel can be
  measured
• A major background is under control

17
Significant high mass events
18
Top three candidates

• L3 Hnn candidate, s/b0.7
• 114GeV Higgs mass
• But very collinear too much?
• ALEPH Hee candidate, s/b0.6
• 118GeV mass
• Evidence for brehmstrahlungMass 99?
• ALEPH Htt candidate, s/b0.5
• 115GeV mass
• Kinematic fit poor

All well b-tagged
All have problems not reflected in their s/b
estimates
19
L3 Hnn candidate

• Two clear b-quarks
• Lots of energy missing
• Could be HZ, with Z decaying to invisible
  neutrinos
• Mass is 114.5GeV
• Or maybe it is just a pair of quarks.
• Importance dropped in publication

20
L3 Hnn characteristics
Mass and neural network output both at signal
peak
Used for discriminator
21
Collinearity of Hnn
Four fermion backgrounds acollinear Two fermion
collinear Total well modelled
We cannot add criteria in the light of the
data. Otherwise classical statistical analysis
is impossible.
L3 candidate
22
How is significance assessed?

• Maximum likelihood fit to observed distribution
• Done in 1D or 2D (e.g. Mass,Neural-Net)
• Each bin needs signal and background estimates,
  from simulation, dependent upon Ecms, channel etc.

Qi

• This is a weighted sum of events.
• L is compared with distributions expected for
  background and signal to quantify probabilities

23
Log-likelihood for two jets

• Positive
• background like
• Negative
• signal like
• Height gives power
• Green 1s
• Yellow 2s

Most powerful
24
Log-likelihood for two jet channels
Good agreement
Tiny excess at 117
25
PDFs for channels
No single leptonic channel has much
discrimination Results in each compatible with
either hypothesis.
26
Combined leptonic PDF

• Disfavours signal
• Excludes nothing

27
Comparison with Hqq PDF
Rather similar distributions (sensitivity)
results somewhat different
28
Combined two-jet limit
MHgt114.2GeV/c2 Exactly what would be expected
29
CLb for two jets combined
Good agreement with background expectation
30
Hqq alone 1-CLb
LEP 1-CLb
Marumi showed you this allready
Significance reduced to 2.1s
31
Conclusions from Two Jet search

• MH gt 114.2GeV/c2, nothing unusual seen
• Statistical analysis has its limitations
• Just as powerful as four jet search
• Results should only be used combined with four
  jets
• MHgt114.1GeV/c2

32
My own conclusions
Lancon will explain this

• Precision electroweak measurements DEMAND a
  Higgs
• MW agrees with Higgs predictions to 1 per mille,
• Mtop agreement to 10

mH 88 GeV/c2
53
- 35
Just a statistical fluke?

• Direct Searches (2.1s effect)

mH 115.6 GeV/c2
0.8
In about 4 years, TevatronLHC should decide
- 0.8
More Precision Measurements with Lepton
Colliders should follow after 2012
We need to build TESLA/NLC
33
Acollinearity of Hnn
Higgstrahlung
Higgstrahlung rather collinear Total much less so
WW fusion
total
L3 candidate