Carsten Hast, SLACExcursion into Tau Physics, November 2007 1 A rather abbreviated history of the Reconstruction of hadronic decays 3 hadrons (g -2) |v.

Carsten Hast, SLAC Excursion into Tau Physics, November 2007 1

• A rather abbreviated history of the • Reconstruction of hadronic decays

• 3 hadrons • (g-2)• |vus|

• Mass measurement• Lifetime• Lepton Universality• Lepton Flavor Violation• Outlook

An Excursion into Physics

Carsten HastStanford Linear Accelerator Center

DESY Hamburg and Zeuthen November 12th,13th 2007


• Back in the early 60’s there was no standard model. People were still trying to sort out why there are electrons and muons. Quarks were suggested in ’64. How to find more, for example more leptons?

• One way of doing that is pair production in e+e- scattering

• Accelerators want to be build, meaning:• need to know how, where, persuade people for money and

finally do it…• SPEAR was proposed in 1964 • MARK I was proposed in 1971• First events in 1974

• In between there was quite some time for people to make up, what the future standard model would be and what features a new, heavy lepton would have …

How to Find a Lepton

e+e- l+l-


The following is from Yung-Su Tsai (1971)• Assuming that the new lepton has the same electro-

magnetic and weak interaction as e and • Assuming it is heavy, so it will decay into hadrons

• From decays via CVC Tsai calculated the decay rates of the into different hadrons:

, K(*) , , a1 , cont. and partial and total decay width, lifetime and cross section

Characteristics of Lepton Decays

W

ee

u

d

d = cos c d + sin c sc the Cabibbo angle


Characteristics of Lepton Decays

lifetime sec)-1 = 310 fsec

re≈ r≈≈ rhad≈


In 1973 MARK I was ready at SPEAR at SLAC

• Look for events with only 2 charged tracks and missing energy

How to Find the First Pair Event

W

ee

u u

d se+

e-

M.L. Perl et al. 75 PRL 35 1489


…it worked out:Martin PerlNobel Laureate of 1995

Have to give it a name:“the third”

M. Perl, Frederick Reines, King Carl Gustaf

And then you have to convince yourself and your colleagues…

• In early 70th there were only 3 quarks (uds), but then in 1974 the J/ was discovered charm was born• 2 families of leptons, 2 families of quarks• Nice SU(2) U(1) symmetry• Who needs a 3rd generation!

How to Find the First Pair Event

… but it helps when you have nice colleagues who find strange electron-muon events as well, like Pluto in 1976, and then…


Production of Pairs

In the following let’s focus on e+e- machines

• CMS energy from threshold to the Z resonance and beyond


LEPClean event environmentLow systematic errorsExcellent for large branching fractions

B-Factories: cross section at (4s) is 0.919 nb ≈ 1M pairs/fbVery good for rare and forbidden decaysWorking hard to get the systematic errors down

Number of pairs per experiment:ARGUS 0.5MCleo 14MLEP 0.1M/experimentBelle/BaBar 650M/440M

Experimental Luminosities


• Initial and final state radiation

• E ≤ ECMS/2 • Taus are flying back to back in CMS• Each decays to an odd number of charged particles (1, 3, 5,

7…)plus neutral particles (0,…)

• Most (semi)hadronic decays proceed via resonances (a1, , K*…)• Need to identify e, , , K, , 0

• Reconstruction of decay vertex is always good• 4 detector coverage, high efficiency, etc.

Reconstruction of Pairs

e+

e-


• ECMS=(4s) = 10.46 GeV• Unfortunately with a boost ( E(e-)=9.0GeV and E(e+)=3.1GeV )

Typical Pair: PEP II and BaBar

SVT, DCH: charged particle tracking vertex &mom. resolution

EMC: electromagnetic calorimetry /0/DIRC, IFR, DCH: charged particle ID /K/p,

1.5 T Solenoid Electromagnetic Calorimeter

(EMC)Detector of Internally

Recflected Cherenkov

Light (DIRC)

Instrumented Flux Return

(IFR)

Silicon Vertex Tracker (SVT)

Drift Chamber (DCH)e- (9 GeV)

e+ (3.1 GeV)

“Typical Event”

Simulated

1-7 topology

event


Standard Event Selection of an 1- n topology:

1. In CMS: Demand 1 charged track vs. n charged tracks2. Net charge of 03. Get rid of converted photons4. Invent good ways to reduce background

1. Cuts on event thrust angle and magnitude2. Cuts on photons in the event3. Lepton identification on the tag side4. Vertexing5. Missing Energy (neutrinos)

5. Particle identification on signal side6. Reconstruct (or get rid of) 0

Pair Event Reconstruction

e-

e+

tag

rec


• The following is an example of how to do a analysis• Most of these decays are quite well measured, except

for the “rarer” ones

• • • • •

(Ian Nugent (BaBar) Tau 06 in Pisa, submitted to PRL in July 2007 )

How to: to Three Hadrons


Rely on superb particle identification of the B-Factory detectors

How to: to Three Hadrons II

--+ K--+ K--K+ K-K-K+--+ 97.40% 22.49% 4.73% 1.02%

K--+ 1.42% 74.87% 16.43 6.38%

K--K+ 0.01% 0.49% 59.63% 25.54%

K-K-K+ 0.26% 50.87%


How to: to Three Hadrons III ARGUS 1993


All modes are related to each other (signal and background)

How to: to Three Hadrons IV

2.8 3.1 3.5 3.9

BR (8.83±0.01±

0.13)%

(0.273±0.002±

0.009)%

(0.1346±0.0010±

0.0036)%

(1.58±0.10±

0.13) 10-5

BRDPG (9.02±0.08)% (0.333±0.35)% (0.153±0.10)% <3.5x10-5 90% CL

BaBar arXiv:0707.2981 [hep-ex] Submitted to PRL


How to: to Three Hadrons V--+

K--K+

K--+

K-K-K+

Mike Roney, Lepton Photon 2007


First observation of tau to phi decays

How to: to Three Hadrons VI

Br) =(3.49±0.55±0.32) x 10-5 BaBarBr)=(3.48±0.20±0.26) x 10-5

preliminary

Belle Phys.Lett.B643:5-10,2006

Br) =(4.05±0.25±0.26) x 10-5

Br) =(3.42±0.55±0.25) x 10-5 BaBar (July 2007)

Br)=(3.39±0.20±0.28) x 10-5 arXiv:0707.2981v1 (acc. PRL)


heavy enough to decay into hadrons

Hadronic Decays for QCD Tests

probes the hadronic V-A current

H

5(1 ) 0H d u

W

d

u

Hadrons

W

d

u

Hadrons

e

e

Hadronse

e

Hadrons

ee H0 probes thehadronic electromagnetic current

0 0q

qH Q q q

hadrons

Whadrons

e+

e –

Isospin


Hadronic vacuum polarisation is the largest uncertainty in calculating a

(ahad)= a few x 10-10 (others are only 0.3-04 x 10-10 )

Dominant part can be calculated from the 2 spectral functions

Hadronic Decays: and (g-2)

muonanomalous magnetic moment:

weakhadQED

22

aaaa

g

Need to measure Br) and the precise mass spectrum shape

Isospin


Fudjikava (Belle preliminary) Tau 06


BG subtractionDetector unfoldingRadiative correctionsFitting Pion form factor

Gounaris-Sakurai parameterization


Fudjikava (Belle preliminary) Tau 06


Interference between ’ and”

a2x 1010

Belle0.5(stat.)3.2(sys.)2.3(Iso.)

Aleph,Cleo3.2(exp.)2.3(Iso.)

CMD2, Kloe4.9(exp.)1.6(rad.)

(770)’(1400) ’’(1700)


1010 x ath = 11 658 471.81 0.02 QED Kinoshita-Nio, Passera 15.4 0.2 EW Czarnecki-Marciano-Vainshtein

698.2 9.7 hvp (711.0 5.8) , (690.9 4.4)eeDavier et al

9.8 0.1 hvp NLO Krause, Hagiwara et al

12.0 3.5 light-by-light Melnikov-Vainshtein, Knech et al

= 11 659 187.6 10.3 (11 659 200.4 6.8) , (11 659 180.3 5.6)ee

aexp - ath = 1.3

Davier/Pich Anomalous Magnetic Moment Summary

BNL - E821A= (11 659 208.06.0) 10-10


Hadronic Decays: |vus|

The route to world best measurements of |vus| and ms

Need detailed knowledge of all (strange) BRs and shapes of all (strange) mass spectra

|Vud|2 +|Vus|2 + |Vub|2 = 1

|Vud|2 = 0.97377 + 0.00027from nuclear decays and neutron decays (PDG 2006)

|Vub|2 = (4.40 + 0.34) x 10-3

from e.g. inclusive B Xul decays

|Vus|2 = 0.2275 + 0.0012Mike Roney, Lepton Photon 2007


1= e

ee

Br hadrons B BR

BBr e

1= e

ee

Br hadrons B BR

BBr e

Strange and non-Strange -Decays

Strangenon,Strange, RRR

0

,, 0

0

( ) ( ) ( )

s

ijwij

dRR s ds w s ij ud or us

ds

, 0 , 00 2 2

( ) ( )( )

| | | |

w ww ud us

ud us

R s R sR s

V V

From weighted spectral functions…

one can calculate the flavour breaking difference



Hadronic Decays: I.Nugent: BaBar preliminary Tau 06 and updated in arXiv:0707.2922 [hep-ex] submitted PRD-RC

qqbar

BG


Hadronic Decays: s0

B.Shwartz, Belle preliminary Tau 06 and updated in arXiv:0706.223 [hep-ex] submitted to PLB

K*0(800)+ K*(892)+ K*(1410)Fit with K*0(892) only


not in average

Br(-K0-) = (0.808 0.004stat 0.026syst)%

PDG 2007 (0.90 0.04) %

Hadronic Decays: s(cont.)


Belle preliminaryarXiv:0708.0733 [hep-ex]

Hadronic decays with

BR is dominated (100%) by


Hadronic Decays: |vus| Summary

Swagato Banerjee, Kaon 2007


Strange Spectral Functions: ALEPH (‘99) and OPAL (‘05)

|Vus|=0.2204±0.0028exp±0.0003th±0.0001msNo good solution Need better data to clarifyNeed B-factories to contribute

Moments k,l=0

DHZ05, hep-ph/0507078


Decay Branching Ratios

Br(-1-prong 0 neutrals ) = (85.330.08)%

Br(-e-ve) = (17.840.05)%

Br(--v) = (17.360.05)%

Br(-1h 0 neutrals ) = (50.130.11)%

Br(-(3h)- 0 neutrals ) = (14.590.08)%

Br(-(5h)- 0 neutrals ) = (0.1020.004)%

Br(-(7h)- 0 neutrals ) < 2.7 x 10-7

Sum = (100.02 0.11)%

148 allowed modes + 55 LF, or L, or B violating

Impressively well measured!


1977 MARK I m= 1.9±0.1 GeV

Mass Measurement

1996 BESm= 1776.96±0.31 MeV


Ecm (GeV)

B r.c.

obs

Mass Measurement

At threshold (BES 1996) Need detailed knowledge of production cross section, beam energy, luminosity


Mass Measurement

Pseudo-Mass Technique • Pseudo-mass was introduced by ARGUS in 1992• Assume neutrino is mass less • direction is approximated by the reconstructed tracks

m*2=2(Ebeam – Ehad)(Ehad – Phad)+mhad

2

MC for input masses:

1.767 GeV1.777 GeV1.787 GeV

MC input mass

Fit

res

ult

BaBar Belle

Mτ = 1776.77 ± 0.13(stat.) ± 0.32(sys.) MeV (prel.) Mτ+ – Mτ– = 0.050.27 MeV

|Mτ+ – Mτ–|/Mτ < 2.8 · 10-4 @ 90% CL (CPT-Test)


Mass Measurement

New preliminary mass from KEDR Collaboration at the VEPP-4M Collider in Novosibirsk (hep-ex/0611046)

0.29 20

0.25 20.23

.26177

1776.80 0.15

6.99 MeV/c (PDG06, BES dominated)

MeV/c (KEDR preliminary)

m

m

e,,,K,– vs –

e -pair events

keV


Lifetime

Very precise results from all 4 LEP experiments are publishedWill describe one measurement technique: “Decay Length Measurement”

A.Lusiani (BaBar preliminary) 80/fb• 1 vs. 3 topology • One prong tagged as lepton• Rigorous event selection criteria to reduce BG• Reconstruct the 3 prong vertex• Measure distance to beam spot


Lifetime

The needed precision makes this type of analysis all but trivial• Fighting detector effects in the real detector and in the simulation• Understanding BG sources and their bias towards the measurement

Decay Length


Lifetime

The Results

PDG 2006 290.6 ± 1.0 fs

DELPHI 2005


Test of Lepton Universality (charged current)

e

g

g

g

g

/ eg g

K K

W W

e

e

e

e

B B

B B

B B

B B

1.0000 0.0020

1.0017 0.0015

1.012 0.010

0.997 0.010

/g g

K K

W W

eB

B B

1.0004 0.0022

0.996 0.005

0.979 0.017

1.039 0.013

/ eg g

W W e

B

B B

1.0004 0.0023

1.036 0.014

A.PichTau06


Test of Lepton UniversalityNow, let’s put everything together and see whether it all works:

mass (m) life time() leptonic branching ratios

3 4 2( ) 1 8 8 12 logf x x x x x x

B=(17.36LEP

and

Be=(17.84CLEO

(B/Be)exp= 0.9725

(B/Be)the= 0.972564

A.Pich Tau06

Standard Model Prediction

m=1776.990.29-0.26 MeV


Quite tough for the current experimental sensitivity…

… but:

Lepton Flavor Violation

• Lepton Flavor is conserved in the Standard Model

• SM extended to include non zero mass and mixing predicts Lepton Flavor Violation


Observation of LFV would be a clear signature of new physics

Some examples:

Theoretical Predictions of SM Extensions

• Many SM extensions include LF violation

• Some models predict LFV up to existing experimental limits

• Neutrinoless 2 and 3 body decays have different sensitivity


Typical Event Selection

e-

e+

tag

rec

• Decay products of Taus are well separated in space (CM)

• Signal side (Neutrinoless decays) l, lhh (l = e, h = , K)

Mass reconstructed from particles = m and their energy = Ebeam

• Tag side (1-prong decays)

Bee

3-prong decays) hhhfor)

/e

ll

l

lh

h


Event Selection

E = E – Ebeam = 0 smeared by resolution and radiation

Signal simulation

MEC candidate invariant mass, with energy constrained to beam energy

M = MEC –M ~ 0

Signal Region is blinded while developing the analysis


• Signal from MC and background from Data (excluding ± 3)

• Common Input Variables:

• Event Missing Mass

• Event Missing PT

• Tag side missing mass

• Tag side momentum

• helicity angle

• MC Optimized for different tagging modes

• electron

• electron + gamma

• muon

• hadron

• hadron gamma

• 3 hadrons

Event Selection II


Results

207106 e+e→ events (L = 232 fb-1)

• Estimate background from linear fit to M sideband

• Nbkg = 6.2 ± 0.5 Nobs = 4 = 9.3 ± 0.6 %

B(→ ) < 6.8 10-8 at 90% C.L.BaBar

Phys. Rev. Lett. 95, 041802 (2005)

What a few 10-7 signal would look like


e Results207106 e+e→ eventsL = 232 fb-1 What a few 10-7 signal would look like

B(→ e) < 1.1 10-7 at 90% C.L.

Estimate background from linear fit to M sideband

1 event in a 22 signal box, Nbkg =1.9 ± 0.4, = 4.7 ± 0.3%

BaBar

Phys. Rev. Lett. 96, 041801 (2006)


Lepton Flavor Violation Results (BaBar + Belle)

S.Banerjee Tau06


Estimate background from 2-dimensional fit

Total expected number of events: 11.3 , Nbkg =0.1- 3.0 (depends on channel)

Total Observed: 10 events

197106 e+e→ events L = 221 fb-1

B(→ lhh) < (0.7-4.8) 10-7 at 90% C.L.

lhh Results (BaBar)

BaBar

Phys. Rev. Lett. 95, 191801 (2005)


Lepton Flavor Violation Results (Belle)Recent results on Belle’s LFV search (Upper Limit on Br at 90% CL)

42 decay modes

535

535

401

401

281

281

158

158158

86

86

154

154

87

This year’s publicationsT. Ohshima Tau06


LFV new prelim. Belle lll

Belle update to 535 fb-1 increase in lumi by 6.1 fold

Efficiency: 6.0-12.5% <Bkgnd>: 0.01-0.40 eventsTotal <Bkgnd>=0.590.31 events Nobs=0 events

90%CL limits: BR lll)<(2.0-4.1)x10-8

5-7 fold improvement over previous limit




LFV new prelim. BaBar lll

BaBar update to 376fb-1 increase in lumi by 4.3 fold

Efficiency: 5.5-12.4% <Bkgnd>: 0.3-1.3 eventsTotal <Bkgnd>=4.20.8 events Nobs=6 events

90%CL limits: BR (lll)<(3.7-8.0)x10-8

2-5 fold improvement over previous limit


New Belle Results on lV0 using 543fb-1

Efficiency:2.5-4.0%<Bkgnd>:0.0-0.2 eventsTotal <Bkgnd>=0.780.29evNobs=3 events

90%CL limits:BR(tlV0)<(6.1-18)x10-8


New BaBar Results l ’

hep-ex/0610067 PRL 98.061803 (2007)

expected background/channel ~ 0.1-0.3Total expected background=3.1, Observed=2

UL on Br. between 1.1 x 10-7 and 5.8 x 10-7 @ 90% CLLafferty (BaBar preliminary) Tau 2006

… and more Violations …


Summary, Conclusions and Outlook

The is a clean laboratory to test the standard model

• Some examples of hadronic tau decay analysis• (g-2) : resonance analysis of hadronic tau decays is

currently in a small contradiction with the production measurements

• More detailed analysis of the strange spectral functions at B-Factories is eagerly anticipated (|vus| and ms)

• Mass measurements are already very precise (0.1 per/mille) and will get even better

• Lifetime measurements from B-Factories are overdue• Lepton Universality is tested to the few per/mille level

but to e/ Br measurements of B-Factories are overdue • Lepton Flavor Violation: no signal at the 10-7 to -8 level


Summary, Conclusions and Outlook

Many thanks for your kind invitation!

B-Factories in 2008 will have (more/less) 1 ab-1 each• Lepton universality O(10-4) (maybe)• LFV O(10-8)• CPV and CPT test in more systematic fashion• |vus| and ms and s (even better)

Super-B Factory with 100 ab-1 • Lepton universality O(10-4) tough to get errors smaller• LFV O(10-10)• CPT test to O(10-4-5)• CPV in decays: case for (4s) running with polarized beams


Additional


BELLE (preliminary) Use -K-0sample to get- K*-) through the K*-(892)K-0modeusing BR(K*-(892)K-0

-K*- BELLE~1.9 lower than

PDG 07(CLEO)

BR (1.130.19 0.07)x10-4 (2.900.800.42)x10-4


Some intepretations within benchmark models

compilation by S. Banerjee (Tau06,hep-ex/0702017)

Carsten Hast, SLACExcursion into Tau Physics, November 2007 1 A rather abbreviated history of the Reconstruction of hadronic decays 3 hadrons (g -2) |v.

Documents

lepton e

pair event reconstruction

tau physics

pair event w e e uu

reconstruction of pairs

final state radiation

e machines cms energy

fsec r e r r