Highlights from LEP

1
Highlights from LEP David J.
Miller University College London

• Why LEP?
• Machine constraints.
• Machine achievements
• Results 1. Z0 Lineshape and Asymmetries
• First Payoff
• More Payoffs
• Achievements on the way
• LEP energy measurement

• Results 2. LEP2
• W mass
• Electroweak fits
• COULD HAVE INCLUDED MUCH MORE
• Luminosity measurement
• QCD
• Gamma-gamma physics
• Heavy quark decays and taus
• Running of ?QED
• Triple gauge couplings
• Searches

Stripped version with gt1 MB bitmap pictures
removed to separate 1-frame files in the same web
directory http//www.hep.ucl.ac.uk/djm/lepslc/.
Complete version is lepslc03.ppt (minor
errors corrected 2/12/03)
2
Why LEP ?
To explore the Standard Model
JOB 2 at ? 161 GeV C. of M.
LEP already proposed when Z0 and W? discovered
(1982/83). 2 big jobs JOB 1 at 91 GeV C. of
M. (shared with SLC at SLAC)
Study the production of WW pairs, and their
decays, at and above threshold.
Study all the properties of the Z0 and its decay
products, on and near resonance.
Note much smaller cross section. Need much more
luminosity.
3
Job 2. Study the production of WW pairs, and
their decays, at and above threshold.
Why LEP ?
4
Constraints on the Machine
1. Space available ?
5
Constraints from- Geology
Politics Physics
All experiments in France
Get e and e- from existing CERN machines
Mountains, Lake, River Versoix
6
Constraints on the Machine

• Space available
• Accelerating technology.
• 1980s copper RF cavities gave 1.5 MV/metre.
• Synchrotron radiation loss/turn ? E4/radius
• Made radius as large as possible (also
  allowed LHC).
• Made straight sections long ( 1 Km)
• (could stick more cavities in for LEP2)
• Did RD on superconducting cavities
• ? 7.5 MV/m
• Lined beampipe with Pb
• ?E/turn 130 MeV at Z ? gt 1 MW S.R.
• ?E/turn 2 GeV at Z ? gt 30 MW S.R.

SEE EXTRA PICTURE 6a
7
LEP machine achievements
Job 1. 4.5 million Z0 events ? 4
experiments. Job 2. 500 pb-1 per experiment
above the WW threshold.

• What is a barn?
• 10-24cm2
• Machine luminosity L ? 2 ? 1031 events/cm2/sec
• Working year 107 seconds
• Expect ? 2 ? 1031 ? 10-24 ? 107
• i.e. 2 ? 1014 events per barn
• OR 200 events per picobarn ? 200 pb-1 per year

500 pb-1 ? (?WW ? 16pb) ? ? 8000 WW/experiment
? (?ZZ ? 1pb) ? a few hundred
ZZ/experiment
8
Results on Job 1
Concentrate on the lineshape and asymmetries in
the four major channels. What are they?
9
Z0 ? ee-
ALEPH
Very easy to recognise in an hermetic, uniform
detector. How? Clear full momentum charged
tracks in the inner trackers pointing at
full-energy E.M. calorimeter clusters. Nothing
significant beyond. What inefficiencies can
there be?
10
Z0 ? ??-
Very easy to recognise in an hermetic
detector with a thick hadron absorber.
How? Clear full momentum charged tracks in
the inner tracker pointing at track segments in
outer layers of hadron calorimeter or in muon
chambers. Only mips-like hits in
calorimeters. What inefficiencies can there be?
11
Z0 ? ??-
Not so easy to recognise. Why? How do ?s decay?
??? ?? (l ?l)? or ??? ?? (qq)? i.e. 1,3
or 5 ?? n?0 So 1, 3 or 5 prongs. 1 prongs maybe
e, ? or ?.
-
12
Separating the lepton pair channels
Taus are easiest to loose.
13
What is the usual clear signature? Two
back-to-back high- multiplicity jets of
charged and neutral hadrons. What inefficiencies
can there be?
L3
Radiated gluon ? extra jet(s) but still a
hadronic event. Require charged multiplicity gt5
per jet to eliminate ?s gives negligible loss of
signal.
14
More hadronic events
15
What determines the Lineshape?
16
What determines the Lineshape?
Full width
Breit Wigner resonance shape
On peak cross section
17
Angular distributions in lepton channels
Note taus and muons v. similar. Electrons very
different. Why?
18
s channel t channel contributions interference
?, ? and e
Pointlike collisions with low angular momentum
J?1, so isotropic or low powers of cos?.
s channel
extra graph for e only
Longer range peripheral collisions
with unlimited angular momentum. C.f.
diffraction from a large object ? forward
peaking.
19
Forward-backward Asymmetries AFB
Why rising thru 0 at Z peak for ? and ?? Z-?
interference in s-channel. Z has V and A ? P
violation Why different for e? t-channel
? dominates except at Z peak
20
Asymmetry formulas
Asymmetry for ? and ? on peak
N.B. SLC made extra constraints with 75
polarised electrons Left or Right handed, at
random. What asymmetry did it measure?
SEE EXTRA PICTURE 20a
21
Fitting the Lineshape and Asymmetries
Extract different physics according to number
of parameters allowed to vary. 1. Simple
description of data in 5 parameter fit
23 ppm!
Assumes quite elaborate QED calculations to
correct for t-channel and for Radiative Return.
Summer 2000 numbers final.
22
(No Transcript)
23
First Payoff from the Lineshape
Cannot see events with so
Everything else
from ?l And AFB
To reduce systematic errors we measure the ratio
24
First Payoff from the Lineshape
OPAL saw SM predicts
theory error depends on E-W H.O.
corrections including ?mtop, ?mHiggs
So number of light neutrinos is
Error still shrinking as experiments combine
25
Important result!

• Why?
• For Physics
• 3 generations of neutrinos
• match 3 of quarks and 3 of
• charged leptons.
• For Cosmology
• Primordial 2H/4He ratio etc.
• depends on number of neutrino
• species which decouple
• as temperature drops to
• 10s of MeV just before
• nuclei begin to bind.

SEE EXTRA PICTURE 25a
26
Second Payoff from the Lineshape
Instead of 5 parameter fit can use 9
parameters. Dont assume equal lepton couplings,
but fit all six of Rl and Al separately. What
does this mean? Lepton couplings ARE equal
UNIVERSALITY. Note all AFB at peak ?0
27
Second Payoff from the Lineshape, continued

• 9 parameters include
• separate V and A couplings
• for each lepton type.
• All 6 couplings consistent
• and with S.M.
• Nonzero peak AFB implies
• Nonzero Vector part even
• though it is smaller than Axial.
• Remember Vector gets messed-
• up by SSB.

(N.B. gVf bigger for neutrinos with Q f 0, so
bigger Z??? than Z?ll -)
28
More Payoffs
Fits constrain contributions from higher
order Electroweak and QCD diagrams, e.g.
mt modifies
Gave prediction of mt before discovery
29
Note much tighter constraint on vector coupling
when ALR from SLC polarised electron results
included. Higher statistics make LEP better
for most quantities except what? Rb, Ab. Why?
Brunel may know! Smaller beampipe diameter
CCD microvertex detector at SLC gave better
beauty i.d.
SEE EXTRA PICTURE 29a
30
Achievements on the way to the Lineshape

• 1. Luminosity Measurement 1 part in 2,000. DJM
  involved, though not prime
• mover in final tour de
  force. See Eur. Phys. J. C14 (2000) 373-425.
• 2. LEP Energy Measurement

2 MeV in 100 GeV. 1/50,000
As often, straightforward at lowest order. Very
hard to get the last reduction by 1/10 ( see
Zeits. Phys. C 66(1995)45) We should know the
magnetic field and the diameter of the LEP ring,
so energy of closed orbit should be calculable
from BUT the magnetic field in cycling magnets
is not known to 1/50,000. Nevertheless there is
a very accurate way of measuring
directly. What?
31
Spontaneous polarisation of beams
Compton back-scatter laser beam off circulating
e or e-. Get asymmetry which measures
polarisation. Why should beams be
polarised? Sokolov-Ternov effect (Doklady
8(1964) 1203 one of many fundamental
Russian contributions to accelerator theory).
Electron energy is less if mag. moment aligned
with B. Synchrotron radiation allows slow
transition to lower state. Ideal machine would
build up to 100 polarisation. Actual machine
has lots of perturbing B-fields (e.g. focussing
quadrupoles) not parallel to bending field. If
incoherent, limit peak polarisation. If
coherent, can resonate turn-after -turn with
precession of spins around main B direction.
Depolarises beam.
SEE EXTRA PICTURE 31a
32
Use of Polarisation
Use polarimeter to measure buildup of
polarisation. THEN HIT BEAM with RF perturbing
B-field. Step frequency by 22.2 Hz through 11.25
kHz orbital frequency. See polarisation
disappear. Frequency very precisely known ? 0.9
MeV error. So why final error 2 MeV?
SEE EXTRA PICTURE 32a
33
Limitations on ?E
Some were expected - Cannot do resonant
depolarisation during data taking use single
beam at end of fill to calibrate magnet
current vs. beam energy. - Needed to reduce
errors on current monitoring.
Put NMR probe into a special magnet in series
with the LEP magnets. But unexpected questions
came up A. Does LEP stay the same size all the
time? NO LEP stretches by order 1mm with
the earths surface in 27 Km circumference. 1/30
,000,000 effects amplified by closed orbit beam
optics to 1/10,000 Sensitive to tiny
environmental effects ?
34
Stretching LEP 1
SEE EXTRA PICTURE 34a
The varying weight of Lake Geneva distorts the
land around it.
35
Stretching LEP 2
What causes this stetching of the earths crust?
Note model curves shapes and phases fit data
exactly. Tidal distortions (solar and lunar)
in the solid earth.
SEE EXTRA PICTURE 35a
36
Limitations on ?E
Unexpected questions (contd) A. Does LEP stay
the same size all the time? B. Why does
the NMR reading jump about during the day,
but not during the night?
SEE EXTRA PICTURES 36a AND 36b
37
Vagabond currents
Escape from the TGV and SBB rails and seek
out- LEP beampipe opposite
directions (pink, blue) on two sides. River
Versoix Lake Geneva making a closed circuit.
SEE EXTRA PICTURE 37a
38
LEP 2 Job 2
Study the production of WW pairs, and their
decays, at and above threshold.
39
What sort of event?
WW?4q at threshold ? Both back-to back
40
Another WW
41
Are these WW _at_200 GeV?
What are the decays?
ZZ ?????
ZZ?4e
42
Measurement of W mass
2 ways at LEP2 a) Threshold scan from 159
to 163 GeV. Cross section gives mass
(c.f. top at LC next week) b) Reconstruct
final states at vs gtgt 161.
b)
a)
What are pros and cons?
a) 159 to 163 GeV Pro - Energy scale
from LEP - Less systematic error
from final state interactions.
Con - Low statistics on signal channel
- Cannot study W couplings -
Cannot push search limits for
Higgs etc.
b) 180 to 208 GeV Pro - Better
statistics. - More physics to do.
Con - Colour reconnection and
BEC systematics. - Energy scale
depends on detector calibration.
43
Main measurement at vs gtgt 2MW
1. Select events 47.5 WW?4q and 4qg
43 WW?2ql?
9.5 WW?l?l? 2. Use neural
nets and jet-finding likelyhood programs to give
best identification reduce combinatorial
background (whats that?)

• 4 jets give 3 alternative pairings to make Ws
• 5 jets give more combinations 2 jets 3 jets.

3. Fit to known beam energies and measurement
errors. e.g. 4
equations, 0 unknowns ? 4C fit or add the
constraint ? 5C fit. For
WW?2ql?, loose 3 constraints with the ?,
(but know ) so ? 1C fit
44
Reconstucted MW
4C, but jets badly measured, combinatorics.
Only 1C, but e, ? well measured and no
combinatorics
1C with extra missing ? from ?
45
Colour reconnection due to rapid W decay?
Worst understood of LEP MW systematics.
Only applies to doubly hadronic events WW?4q and
4qg
Reconnection of colour strings?
W decay lengths short
Hadron formation lengths longer, from QCD string
fragmentation
Not a problem for leptonic W decay, at LEP or at
Tevatron. Makes Tevatron measurement competitive
with LEP.
46
Fits to Standard Model
Ignoring the 2.9 s.d. effect and putting
available Electroweak measurements into a common
fit ?
The yellow band is the SM prediction, with
largest errors due to uncertainty in mH, but a
visible contribution due to the uncertainty in
the hadronic corrections to ?QED
largely dominated by the low mass resonance
region 1-3 GeV.
47
Summer 2003 status of MW (from EW working group
http//lepewwg.web.cern.ch/LEPEWWG/)
LEP1/SLD/mt is the result of the Standard Model
fit to Z0 lineshape, asymmetries and mt from
Tevatron. NuTeV is deduced from NC/CC ratio
in ?-nucleon scattering. Remember NC/CC?sin2?W
and (see e.g. Halzen and Martin).
Direct
Indirect
There is still a 2.9 S.D disagreement! New
physics?
48
Higgs Prediction.
Putting in measured mt and taking mH as the
only Unknown we get the Blue-Band plot.
The lowest chi2 gives the most likely value for
the mass, with 95 confidence that chi2 is more
than 2.7 which corresponds to mH lt 196
GeV. The yellow area is ruled out by the direct
search at LEP2.
http//lepewwg.web.cern.ch/LEPEWWG/plots/winter200
3/