Photon polarization in radiative B decays · Gambino & Misiak, NPB 611 (2001) 338 Buras, Czarnecki, Misiak, Urban, NPB 631 (2002) 219 We need more observables to uncover new physics!

Photon polarization inradiative B decays

Thomas Schietinger

Laboratory for High-Energy PhysicsEPF Lausanne

NIKHEF colloquium, 6 April 2006

– one of the last battlegrounds in B physics

Thomas Schietinger 6 April 2006Photon polarization in radiative B decays 2

Contents

● Introduction: the key role of the b s transition

● The quest for new observables– Photon polarization

– Double-radiative decays

● Experimental (Belle, LHCb) and phenomenological program at Lausanne:– B K(), B K, B at Belle

– b (*) at LHCb

– New method to measure photon polarization using charmonium resonance interference


The key role of theb s transition

b s


b s: an ideal place to study the Standard Model and to look

for new physics!

b s

b

W –

s

u,c,t

b

s

squark

SM

SUSY?

The Standard Model amplitude isstrongly suppressed (2nd order)...


b s

SM

SUSY?

b s exclusion area in mass plot of universal sfermion mass vs. universal s-scalar mass.

mt = (150±30) GeV/c2

J.Ellis et al.,Phys. Lett. B 573, 162 (2003)

b s: an ideal place to study the Standard Model and to look

for new physics!


If you want to heara rare bird singing,go somewhere quiet!

Penguin amplitude

New physics

SM


Penguin amplitude

...not where all the other birds make noise!

Tree amplitudes

SM

SM

SM

New physics

SM

If you want to heara rare bird singing,go somewhere quiet!


b s inclusive branching fraction

New HFAG average of these measurements using a common shape function for the extrapolation to low photon energies and taking into account the correlated error from b d contamination:

BF(b s) = 355 ± 24 +9 ± 3– 10

fully inclusive semi-inclusive

(see http://www.slac.stanford.edu/xorg/hfag/rare for details)

Calculation of extrapolation factors byO. Buchmüller and H. Flächer, hep-ph/0507253

errors are: stat./syst./shape

stat./syst.combined

shape function

b d contamination

BF(b BF(b ss) = 357 ) = 357 ±± 30 30

Standard Model prediction (NLO):

BF(b s) = 336 ± 53 ± 42+50 – 54

PLB 511 (2001) 151

BF(b s) = 355 ± 32 +30 +11– 31 – 7

PRL 93 (2004) 061803 ● E()>1.8 GeV ● E()>2.24 GeV ● 16 modes

BF(b s) = 335 ± 19 +56 +4– 41 – 9 BF(b s) = 367 ± 29 ± 34 ± 29

hep-ex/0507001 PRD 72 (2005) 052004 ● Lepton-tagged● E()>1.9 GeV; BF not extrapolated below!

● E()>1.9 GeV ● 38 modes covering ~55%

BF(b s) = 321 ± 43 ± 27+18 – 10CLEO

PRL 87 (2001) 251807 ● E()>2.0 GeV

● E()>1.6 GeV

● E()>1.6 GeV Gambino & Misiak, NPB 611 (2001) 338Buras, Czarnecki, Misiak, Urban, NPB 631 (2002) 219 Depressing agreement between

theory and experiment!

(BF units: 10 – 6)

http://www.slac.stanford.edu/xorg/hfag/rare


b s Blues


Experiment:

b s

BF(b s) = 355 ± 26

BF(b BF(b ss) = 357 ) = 357 ±± 30 30Gambino & Misiak, NPB 611 (2001) 338Buras, Czarnecki, Misiak, Urban, NPB 631 (2002) 219


b s Blues


Experiment:

b s

BF(b s) = 355 ± 26

BF(b BF(b ss) = 357 ) = 357 ±± 30 30Gambino & Misiak, NPB 611 (2001) 338Buras, Czarnecki, Misiak, Urban, NPB 631 (2002) 219

We need more observables to uncover

new physics!


Additional Observables in b s

Photon PolarizationDouble-radiative

Decays⇒

b s

b s

Expected to be left-handed in the Standard Model (chiral structure of W boson coupling in loop), but could have a right-handed component in LR-symmetric models.

Valuable additional observables (at the expense of statistical power):● size of non-resonant contribution● diphoton invariant mass

➔ Any surprising resonances?● forward-backward asymmetry

Two approaches are pursued in Lausanne:


Photon polarizationin b s

b s

⇒


Why left-handed?

b

W –

sL

u,c,t

⇒

⇒

The W only couples to a left-handed s-quark.

The (back-to-back) emitted photon must be left-handed too in order to conserve angular momentum!

The exact argument only holds for massless quarks ⇒ small right-handed component of order (m

s/m

b)2≈ 0.1%

expected (up to 1% with QCD corrections).


Photon Polarization in b s

The (high-energy) photon is usually detected via its energy deposit in the form of an electromagnetic shower inside a crystal.⇒ Any information on the polarization of the photon is lost!⇒ Need indirect methods to determine photon polarization!

b s

shower by Sven Menke

Several such methods have been suggested:● e+e– conversion● Interference of higher K resonances● B-B interference● Polarized b-baryons


b s polarization measurement: e+e – conversion method

⇒

b s

e –

e+

● The geometry of the conversion electrons allows the inference of the photon polarization.

● Use B K*(K), and measure correlation between the K*K decay plane and the e+e– plane.

● Two variants:

– Virtual photon (the decay B K*e+e – ); get also interference from Z-exchange and W-box diagram.

– Real photon: conversion somewhere inside the detector material (e.g. beam-pipe).

● Either way needs lots of statistics – probably Super-B-factory?

D.Melikov, N.Nikitin, S.Simula, Phys. Lett. B 442, 381 (1998)

Y.Grossman, D.Pirjol, J. High Energy Phys. 06, 029 (2000)


b s polarization measurement: K resonance method

● While the spin information is lost in B K*(K) (strong two-body decay, no phase involved) it can be reconstructed in more complicated B K**(K) decays, if the strong phases governing the intermediate resonances are known.

● Somewhat messy analysis of the K Dalitz amplitude necessary, but very promising where one or two amplitudes dominate.

● For some time this was thought to be the most promising method experimentally, but...

M.Gronau, Y.Grossman, D.Pirjol, A.Ryd, Phys. Rev. Lett. 88, 051802 (2002)

B

K*⇒ K

⇐

B

K**⇒ K

⇐

M.Gronau, D.Pirjol, Phys. Rev. D 66, 054008 (2002)


BABAR: B K hep-ex/0507031

(232M BB)

Results:

B+ K+ – +

B0 K+ – 0

First obser-vations!

Confirmation of Belle results

PRL 89 (2002) 231801

K invariant mass (background subtracted)

Clear signals!

Less clear resonance structure...

K1(1270)

● Recent BABAR analysis of B K channelsshows clear signals for all charge combinations butmessy resonance structure.

● Detailed study of resonance structure still in progress.● K

1(1270) clearly visible

● Polarization analysis requires “ clean” K1(1400) in

K+0 modes.● We probably have to wait for a Super-B factory for

this measurement!


b s polarization measurement: B-B interference method

● Photon from b s expected to be left-handed ⇒ photon from b s expected to be right-handed

● Quantum mechanics: amplitudes can only interfere if initial and final states are identical.

● If we can measure the interference between b s and b s, we can learn something about the photon polarization (expect no interference for fully polarized photon).

● How on earth can we measure interference between b and b?⇒ B mixing makes it possible!

D.Atwood, M.Gronau, A.Soni, Phys. Rev. Lett. 79, 185 (1997)

⇒

b s

⇒

interference?

b s


Reminder: Measurement of the B mixing phase (“ sin2” )

B0

bW –

d

c

c

s

d

Vcb*

Vcs

J/

K0 KS

B0

bW+

d

c

c

s

d

Vcb

Vcs*

J/

K0 KS

● Both B0 and B0 can decayinto the CP eigenstate J/ KS.

● No phases are involved in the decay (only Vcs and Vcb).

Bigi and Sanda, Nucl. Phys. B 193, 85 (1981)Carter and Sanda, Phys. Rev. D 23, 1567 (1981)


B0

bW –

d

c

c

s

d

Vcb*

Vcs

J/

K0 KS

B0

bW+

d

c

c

s

d

Vcb

Vcs*

J/

K0 KS

B0

d

b

t

t

Vtd*

Vtd

⇒ The B0 has two ways to decay into J/ KS:

“ unmixed” decay: no phase

“ mixed” decay: phase of Vtd!

⇒ Time-dependent interference term proportional to:

Vtb* Vtd

Vtb Vtd*= e2i


Repeat for B K*(K)!Note: need CP eigenstate, henceonly K*0 KS0 useful.

“ unmixed” decay: no phase

“ mixed” decay: phase of Vtd!

B0

b

d s

d

K*0KS0

W –

t

Vts

Vtb*

B0

b

d

d

B0

d

b

t

t

Vtd*

Vtd K*0KS0

s

W+

t

Vts*Vtb

⇒ No interference if photon is polarized!⇒ A measurement of the B mixing phase (“ sin2” ) in this channel will show if (to what degree) the photon is polarized!

⇒

⇒


A technical problem...● At the B factories the measurement of time-dependent

asymmetries relies on the separation of the decays of the two simultaneously produced B mesons (signal and tag vertex).

● K*0 KS0 does not give a good signal vertex!✗ All neutrals, KS has long lifetime!

B0

B0 Boost gives better timeresolution and a reference “ t0 ” !

~200 μ m

e+e – (4S) B0B0 or B±B∓

(4S)

Two solutions to the problem:1. Intersect KS momentum vector with beam envelope.

● Method developed by BABAR, successfully applied by BABARand Belle (see results).

2. Use a different CP eigenstate that leaves a measurable vertex.● Prime candidates are B KS and B KS.


Indirect CP violation in B0 KS0

Beam intersection method

B0 K*0(KS0)

e –

e+

0

KS +

–

interaction region

● Both BABAR and Belle have performed measurements of S = “ sin2” in radiative B decays.

● First in B0 K*0(KS0) , but S parameter should be the same for all B0 KS0 events.[Atwood et al., PRD 71 (2005) 076003]

⇒ combine events from the full KS0 spectrum.

● Experiments indeed seem to favor small S, buterrors are still large!

● Currently our best handle on photon polarization, butthis measurement may need a Super-B-factory as well!

S(B0 K*0[KS0] ) = – 0.21 ± 0.40 ± 0.05

PRD 72 (2005) 051103

S(B0 K*0[KS0] ) = 0.01 ± 0.52 ± 0.11

S(B0 KS0) = 0.08 ± 0.41 ± 0.10

hep-ex/0507059(386M BB)

(232M BB)

(S = “ sin2” )


B K and B K

● Alternative and complementary approach pursued in Lausanne: instead of B0 KS0, measure B KS and B KS (replace 0 by and , which give a well-measurable decay vertex!)

● Main drawbacks:– Need much more statistics.– Need angular analysis (VVP decay)

● Definitely requires a Super-B-Factory! Current study with Belle is of exploratory character, nevertheless a lot of nice bread-and-butter physics:

– b s hadronization

– K resonances

– Non-resonant B KKK decays, etc.

B0 KS()

e –

e+

KS +

–

interaction region

K –

K+ (Thesis Christian Jacoby)


Experimental status of B K()

BF(B+ K+) = 3.4 ± 0.9 ± 0.4 (5.5)

BF(B0 K0) = 4.6 ± 2.4 ± 0.6 (3.3)

BF(B0 K0) < 8.3 @ 90% C.L.(95.8M BB)

PRL 92 (2003) 051801

Previous measurement by Belle based on 90 fb– 1:

Plan for Lausanne group (thesis C. Jacoby):● Updated measurement on 450 fb– 1 planned for the

summer conferences.● First measurement (or limit) for B K.● Depending on signal size, first angular analysis.

No measurements/limits for B K yet...


b s polarization measurement: b-baryon method

● Idea: if we can polarize the initial b, we can infer the photon polarization from angular correlations with the strange final state.

● Unfortunately, B mesons carry no spin. The method only works with b-baryons, such as b (e.g. b ).

● Specific asymmetries suggested for b at a Z factory, but may also be possible at the LHC, if the b's are sufficiently polarized.

● At LHCb, b * (strongly decaying *) may be more promising than . ('s leave the vertex detector before decaying!)

T.Mannel, S.Recksiegel, J. Phys. G: Nucl. Part. Phys. 24, 979 (1998)

G.Hiller, A.Kagan, Phys. Rev. D 65, 074038 (2002)

⇒

⇒

b s

⇒

⇒

b s

versus


b (*) at LHCb(Thesis Federica Legger)

b ( p) :

b *( pK) :

● Advantage: can use photon and protonasymmetries to probe photon polarization (weakly decaying “ remembers” its polarization!)

● Drawback: ’ s fly on average 3 m: most of them leave the the vertex detector before decaying! ⇒ difficult to trigger

⇒ difficult to reconstruct

● Advantage: use pK vertex for triggerand reconstruction.

● Drawbacks: – need to disentangle various pK resonances.

– less statistics to start with (w.r.t. b )

– some of the lowest resonances have spin 3/2,complicated extraction of photon polarization.

VELO RICH 1 TT

p

(1520)(J = 3/2)

(1670)(J = 1/2)

(1690)(J = 3/2)


b (*) at LHCb(Thesis Federica Legger)

Right-handed polarization fraction

One of the few windows of opportunity left open by the B factories for LHCb!!

Phenomenological study:● Spin 3/2 resonances are experimentally useful!

Ratio of m = 1/2 to m = 3/2 amplitudes can be determined by experiment.[F. Legger and T.S., to be submitted to Phys. Lett. B]

Complete gen-sim-rec-sel-fit studyfor b and b (1670):

● We can expect ~750 b and ~2500 b (1670) events (fully reconstructed)

● Sensitivity to photon polarization:– assume 20% for b polarization– despite smaller statistics, p still offers better prospects (can measure down to 20%)– if p reconstruction impossible for some reason, * pK channels can still probe photon polarization down to 25%!

“ Naive” SM predictionLHCb reach

(1 year)


Double-radiative decays

b s


Initial idea: exploration of the“ K landscape”

m [GeV]

rate

1 32 4

Non-resonant contribution+ resonances + interferences

0 '

c

'cAlso:K*(K) contribution(smeared over the whole spectrum)


Sort of like B K(*)ℓ+ℓ–...

mℓℓ [GeV]

rate

1 32 4

J/

No equivalent toK*(K) contribution!

'

Contributions from,, negligible

ℓ+ℓ– pole


A brief history of B K ● First calculations of the b s amplitude in the 1990s, mainly in the context of Bs .

● 1997: Reina, Ricciardi and Soni have a closer look at inclusive B Xs (including QCD corrections). Suggest to measure diphoton mass spectrum and forward-backward asymmetry.

● 1997: Singer and Zhang calculate BF(B+ K+) ≈ 5 x 10 – 8.

L.Reina, G.Ricciardi,A. Soni, Phys. Lett. B 396, 231 (1997)Phys. Rev. D 56, 5805 (1997)P.Singer, D.X.Zhang,

Phys. Rev. D 56, 4274 (1997)

● 2003: Choudhury et al. find 1.477 x 10– 6 ≤ BF(B+ K+) ≤ 1.748 x 10– 6 !! – Wow! Could this be measurable at Belle? Let's have a look!

● 2004: EPFL master student Mathias Knecht plugs Choudhury formulae into EvtGen event generator and normalizes to measured branching fractions: Choudhury non-resonant result is at least too orders of magnitude too large!

● 2005: Hiller and Safir publish a SCET calculation that gives BF(B+ K+) = O(10– 9)!

– Choudhury et al. acknowledge a typo in their mathematica script and publish an erratum.

● Further studies by us reveal that the resonant decay B+ K*+( K+) K+ eclipses the non-resonant B+ K+ everywhere in phase space.

b s

S.R.Choudhury et al.,Phys. Rev. D 67, 074016 (2003)

G.Hiller, A.S.Safir, J. High Energy Phys. 0502, 011 (2005)

S.R.Choudhury et al.,Phys. Rev. D 72, 119906(E) (2005)


B K*(K1)

2 B K*(K2)

1

interference term

What does this mean?

1) Non-resonant B K is inaccessible to experiment!

● Even at the centre of the Dalitz plot, the decayB K*( K) K still dominates the non-resonant contribution!

● But wait! A small non-resonant contribution has its advantages too! It allows us to observe interference between B K*(K) and charmonium channels such as B Kc() and Kc(), with known final state polarization!

● A new method to measure photon polarization in B K*?

2) Our only hope to access short-distance b s (or b d) is Bs (Bd ).

● Belle has new world-best limits on both decays!


B K: a new method to measure b s photon polarization

Mathias Knecht and T.S.,Phys. Lett. B 634, 403 (2006)

Non-resonant B K is tiny ⇒ rate is dominated by B K* (K* K) even at the centre of the Dalitz plot!● two amplitude contributions:

B K*(K1)2 and B K*(K2)1!

B K*(K1)

2 B K*(K2)

1

interference term

⇒ Clean interference with charmonium

decays B Kc() and Kc(), where the photon polarization is well known

(think positronium!)

by measuring the

m spectrum, can derive photon

polarization in B K*! (constraints on Wilson coefficients C7, C'7)

Main issues:● statistics! (needs Super-B factory, better Hyper-B) ● strong phase between B K* and B Kc...


Experimental study of B K at Belle

● Search for non-resonant B K makes no sense.

● Focus now on search for radiative charmonium decays in B K():

– Study of c , c ’ , c , etc. (charmonium spectroscopy in clean radiative channels)

– Search also for recently found new states such as X(3782) to help determine C-parity of these new states.

● Use peaks from B K() and B K ’ () for calibration and systematics.

● Analysis well advanced, results being will be released at summer conferences. Stay tuned!

(Thesis Jean Wicht)


Bd and Bs : new limits by Belle● B : the only way to probe short-distance b s(d)!● Lausanne group has performed a search for Bd based on 100 fb –

1 (Stefano Villa).– Cannot use full dataset because of missing timing information for calorimeter clusters

in early data ⇒ huge background from Bhabha’ s!

● Limit on Bs based on this analysis from the (5S) engineering run (1.86 fb – 1, Alexey Drutskoy).

BF(Bd ) < 0.62

PRD 73 (2006) 051107(111M BB)

BF(Bs ) < 56preliminary

BF(BF(BBdd ) = 0.03) = 0.03SM:

BF(BF(BBss ) = 1.2) = 1.2

Bs physics at Belle isalready competitive!

SM:


Conclusion● Photon polarization remains one of the last glaring Standard

Model predictions to be tested in the realm of B physics.– Various methods have been proposed for B factories, but probably all of them

will require Super-B statistics.– A window of opportunity for LHCb using b (*) decays!!

● Vibrant research program in Lausanne to search for new physics in b s() decays:– B K() at Belle: angular analysis in view of time-dependent CP

measurement.– B K at Belle: look for charmonium (and other?) resonances

– B at Belle: world’ s best limit recently published.

– Phenomenological study on possible measurement of b s photon polarization with charmonium resonance interference.

– Very promising feasibility study for polarization measurement with b (*) decays at LHCb.


Appendix:What I have done for LHCb


LHCbA forward spectrometer to exploit the forward-peaked production of B hadrons at

the LHC!

100 µb

230 µb


Level-1:● software ● 1 MHz 40 kHz● Uses:

vertices (Si) some tracking L0 objects

LHCb Trigger

High-Level:● software ● 40 kHz 2 kHz● Uses:

full event data

Level-0:● hardware ● 10 MHz 1 MHz● Uses:

calorimeters muon chambers pile-up veto (Si)

trigg

er ra

te [H

z]

102

104

103

106

105

108

107 totalcharmbottom


B hadrons are the elephantsof the particle zoo: they are

heavy and long-lived

Level-1 Strategy⇒ Approximation at trigger level:look for tracks with both

1) large impact parameter(relative to primary vertex)

Reconstruction of rz tracks in VELO: σIP ≅ 50 µm

and2) high transverse momentum (pT)

Need to extrapolate tracks to some measurement that is influenced by the magnetic field!Two possibilities:● Extrapolation to first tracking station before the magnet (TT)● Extrapolation to objects found by Level-0

(accessible to Level-1)


µ

e,γ

π,K

We must extrapolate tracks to some measurement that is influenced by the magnetic field!

Two complementary approaches:

1) Fringe field before the magnet:extrapolation to first tracking station, TT (= Trigger Tracker), situated between VELO and magnet⇒ coarse momentum resolution but high

efficiency

2) Full pT kick after the magnet:recycle calorimeter clusters and muon tracksegments found by Level-0, try to matchthem to VELO tracks!⇒ better momentum resolution but low over-all

efficiency and low purity

pT Measurement (Trigger)


pT Measurement (TT)

VELO RICH 1 TT

Re-optimized LHCb design: some magnetic field between VELO and TT

integrated Bdℓ ≃ 115 kG cm⇒10-GeV track is deflected by 3.4 mm at TT

Momentum resolution:

20– 40%


Level-1 Trigger Performance

Performance and timing within specifications!

Level-1 trigger line rate [kHz]

Generic (hadron) 29.2Single-muon 3.2Di-muon 1.4Di-muon (J/psi) 0.6Electron 2.3Photon 2.3

L1 processing time

Physics channel L1 efficiency

Bd +− 83%Bs Ds

+K− 81%Bd D0K* 85%Bd K* 67%Bs J/(+−) 87%

Photon polarization in radiative B decays · Gambino & Misiak, NPB 611 (2001) 338 Buras, Czarnecki, Misiak, Urban, NPB 631 (2002) 219 We need more observables to uncover new physics!

Documents