Diapositive 1

1
THREE-BODY CHARMLESS B DECAYS WORKSHOP LPNHE -
PARIS 1-3 February 2005
Three-Body Charmless B Decays at LHC(b)

• Robert O. Deschamps
• LPC Clermont-Ferrand LHCb collaboration

2
1
MAIN PURPOSE
Flavour physics is an important topic of the LHC
program.
B physics will be studied by three experiments,
among them LHCb, which is dedicated to this
topic with the most complete program.
Concerning the 3-body charmless B decays,
several studies are aiming at the CKM metrology
g extraction
B? pp-p, B0 ?Kpp B0 ? pp-p0
a extraction
Talk mainly dedicated to the a extraction from
the Bd ? pp-p0 decay channel in the LHCb
experiment framework
3
2
TOWARDS THE NEXT GENERATION OF B FACTORIES
? Large Hadron Collider at CERN
Starting planed in mid-2007
L(2007) 1032 cm-2s-1 ? L(2010) 1034 cm-2s-1
(ATLAS CMS)
3 experiments with a B physics program
(dedicated)
4
3
B DECAYS STUDY WITH THE LHCb EXPERIMENT
? Dedicated detector to reconstruct a large
variety of (rare) B decay modes.
bb correlation
? The challenge lies in reconstructing and
extracting the B decay of interest within a
high multiplicity tracks environment
Lnom 2.10 32 cm-2s-1
R
One Bd ? pp-p0 decay every 2 seconds
z
5
4
THE Bd ? pp-p0 DECAY CHANNEL
Working hypothesis B0d ? rp ? ppp is dominant
spin-1 pseudo-scalar meson
S-
? r channel interferences
r0p0
rp-
Form factors fiFi.Yj0(cos(q))
r-p
S
In the corners of the Dalitz plot one the 3 pions
remains at rest in the B rest frame ? LHCb
acceptance versus kinematics will have an
impact on the signal original profile
6
THE Bd
? pp-p0 DECAY CHANNEL SNYDER and QUINN
PRESCRIPTION Phys. Rev. D48-2139 (1993)
5
a
a
a
B0 at t0
a
a
a
-
B0 at t0
Remove ambiguities between 0 and p
7
TOWARDS THE a EXTRACTION
6
11 observables ? 13 parameters but

• Flavor structure of local operators
• DI1/2, DI3/2
• Assume SU(2)F symetry.
• Neglect E-W penguins in this study.

Pij ? DI1/2
P00-1/2(P-P-)
Fit of 9 dynamical parameters
r lineshape model Introduce a Kuhn Santamaria
form
8
30
7
METHOD FOR SENSITIVITY ESTIMATION
Maximum likelihood fit ?
? The experimental inputs are estimated from
fully simulated MC data
9
8
FRAMEWORK OF THE STUDY
The data are not (yet) here

• Use of Monte Carlo samples of fully simulated
  events
• (electronic noise , pattern recognition )
• to estimate analysis performances whitin a
  framework as realistic as possible

MC statistics used for this study

• Signal (Bd ? pp-p0 ) events, N3p 1.106 events
  1 day of data taking.
• bb inclusive sample, Nbb 40.106 events 12 mn
  of data taking.
• Specific background model samples. Charmless
   cocktails  of B
• and B0, B ? Kpp decays.
• 100.106 Minimum bias events pp ?X 5 seconds of
  data taking.
• (Trigger performance studies)

10
EXPERIMENTAL CHALLENGE NEUTRAL PIONS
RECONSTRUCTION
9
(signal combinatorics)
(signal combinatorics)
sMgg 15 MeV/c2
sMgg 10 MeV/c2
Invariant mass (MeV/c2)
Invariant mass (MeV/c2)
lte(po)gt 53 Resolved 33 Merged 20
g Identification is based on calo-cluster
neutrality (anti-track matching) and
kinematics
Bd ? pp-p0

efficiency
2 kinds of reconstructed p0s

• Resolved p0 ? Pair of isolated photons
• Merged p0 ? Pair of photons merged within
• a single cluster
  (large PT)
• ? use of cluster
  shape

11
10

THE Bd ? pp-p0 EVENT SELECTION (1)
Multivariate Selection
Reconstructed and identified pp-p0
combinations from inclusive bb events

• Particule
• identification

signal
e,m,K,p
p,p-
bb
S(DLL)

• b momentum
• spectra

similar level of bkg over the r bands ? The
whole Dalitz plane is considered
Inclusive bb
signal

• Vertexing
• criteria

bb
signal
PDF distribution


12
11

THE Bd ? pp-p0 EVENT SELECTION (2)

• Selection efficiency

• Signal distorsions - resolutions
• (dominated by Ecal energy resolution)

• Expected annual yield

N3p 12.103 events/2 fb-1


s(t) ? 70 fs
s(Min(?s, ? s-)) ? 30 MeV/c2

• Signal distorsions acceptance functions

Phase space coordinates acceptance
Proper time acceptance
Depopulated region due to p/- large impact
parameter requirement
Depopulated region due to p0 large
energy requirement
13
12
THE Bd ? pp-p0 EVENT SELECTION (3)
Flavour tagging and background within an hadronic
framework

• Tagging efficiency e 40
  2
• Wrong tag fraction w 31
  2

• Assumption most dangerous
• source of background originates from b decays

Most contaminating specific bkgs
Combinatorial fragments from charmed B decays
B/S0.3 Bd ? rr-

B/S0.15 Bd ? Kp, Kg

B/S0.1
Considering B/S O(1) in the following
studies is probably a not too wrong guesstimate
14
13
BACKGROUND MODELS
Without real data background is hard to model
? Introduce 3 generic classes of background

• Bd -gt K(K p-) p0
• Bd -gt K(K p0) p-
• Bd -gt Kr-(p- p0)

r resonant
B ? Kpp
Flat
(CP violating - g)


signal
Assumptions of the sensitivity study

• Assume specific mixtures of the 3 classes of bkg
  (res) (flat) (Kpp)
• Assume same proper time distribution than for Bd
  ? pp-p0 events.
• Assume same acceptance functions, resolutions
  and tagging dilution
• than for Bd ? pp-p0 events.

15
14
PROSPECTIVE STUDY (1)

• 2 background classes 0.5(res), 0.5(flat)

s
Distribution of fit error
s
-s
-s

• 3 background classes 0.6(res), 0.3(flat),
  0.1(Kpp)

90 of toy experiments with s(a)lt10
With B/S1
16
15
PROSPECTIVE STUDY (2)
3 classes of background B/S1
Build a single toy averaging 100 toys
?Ldt2 fb-1
agen 98.1 (from CKMfitter EPS 2005)
(g value is supposed to be perfectly known in bkg
models)
17
16
SENSITIVITY TO QCD PARAMETERS
Assumed scenario
?Ldt2 fb-1 B/S1
Penguin strong phases
Tree strong phases
Imaginary part

• Highly asymmetric distributions

• With 2 fb-1 , poor sensitivity to R
• Difficult to see sizeable NP effects through R
  determination
• after 1 year

Real part
18
17

SYSTEMATICS EFFECTS
Preliminary study of the impact of an imperfect
knowledge of experimental or phenomelogical
ingredients in likelihood function definition
Da 4
Include g 5 uncertainties in Kpp model
Da 1
Non-uniform wrong-tag - averaged in fit
Da 0
L not accounting for proper time acceptance in fit
L not accounting for Dalitz acceptance in fit
Da 5
Da 0
L not accounting for r/w mixing in signal
L not accounting for r and r contribution in
signal
Da 7
arbitrary
L not accounting for r3 contribution in signal
(k0.2)
Da 12
A poor description of r lineshape and/or the
phase space acceptance leads to large bias on the
a measurement ? Strategy to obtain an accurate
knowledge is under development
19
CONCLUSION
18

• The Bd ? pp-p0 time dependent Dalitz plot
  analysis is ambitious.
• With 2 fb-1 LHCb may achieve sstat 10,
  assuming an accurate
• control of the r lineshapes and the
  experimental distorsions.
• This expected value is competitive with the
  current results
• of the B factories.
• Such a control is not trivial and strategies have
  to be developped to
• extract the relevant parameters in real data.

December 2005
Will be ready at day 1 !
20
SPARE SLIDE(S)
21
THE Bd ? pp-p0 DECAY CHANNEL
Two kind of topologies can contribute at the rp
level
W
p-
u,c,t
b
W
p-
b
Vub (g)
Vtb(b)
g
r
r
  Penguin (P)
  Tree (T)
2 components ? CKM unitarity
(ij0)
q/pe-2ibo(10-3)
abgp
22
NUMBER OF MISRECONSTRUCTED SIGNAL EVENTS
S-
Fraction 3 in upper zone Fraction 20 in
lower zone (but including MC truth matching
failure from soft p0)
S
23
NEW PHYSICS SENSITIVITY
Whithin SM framework
q/pe-2ib
cancel exactly the b ? d penguin contribution
NP can modify this picture (loops)
London and Page Phys. Rev. D (2004) 017501


Fit versus observables
? Assume Rij theoMS10 10
Check compatibility with qNP?0 ? with 5 years of
data taking Rij mes55 signs NP at 3s
Nominal years