Graziano Venanzoni Laboratori Nazionali di Frascati KLOE results on Hadronic cross section Topical Workshop on (g-2) Glasgow, 25-26 Oct 2007 (for the

Graziano Venanzoni Laboratori Nazionali di Frascati

KLOE results on Hadronic cross section

Topical Workshop on (g-2)

Glasgow, 25-26 Oct 2007

(for the KLOE collaboration)

Outlook

• Motivation of (e+e- ) at low energy

• Reminder: KLOE published results on |F and awith 2001 data (normalization to Bhabha)

• Update of |F and a with 2002 data, using two selection schemes (large and small angle)

• Work in progress:– Extract |F via bin-by-bin normalisation to

– Use data collected at 1 GeV

Motivation: High Precision Test of the Standard Model

Anomalous magnetic moment of the muon

Fine structure constant at Z0-mass QED (MZ)

Hadronic Cross Section & Muon-Anomaly

Muon-Anomaly a= (g-2)/2 =

aaaaatheo QED had weak New physics

Dispersion-Relation Channel = (e+ e gives >70% contribution to a

had!

• K(s) = analytic kernel-function, • above typically 2…5 GeV, use pQCD

Alternative: Spectral function from decay ( Hadrons) taking intoaccount isospin breaking corrections

ahad had (s)

Hadronic Vacuum Polarization

2nd largest contribution, pQCD not applicableError of hadronic contribution dominates total error of a!

µ+

q

q

µ+

M. Davier

had and the muon anomaly

(e+e-+) with ISR:Since DANE runs at fixed energy (1020 MeV), we measure the cross section (e+ e- hadrons ) as a function of the hadronic c.m. energy M hadr by using ISR (”radiative return”):

This method is a complementary approach to the standard energy scan.

• Luminosity, acceptance (and other normalizations) enter only once !• Requires precise calculations of the radiator H

PHOKHARA MC NLO Generator

(H. Czyż, A. Grzelińska, J.H. Kühn, G. Rodrigo, Eur. Phys. J. C 27, 2003)

Neglecting FSR effects:4m

2 s = M m

2

S. Binner, J.H. Kühn, K. Melnikov, Phys. Lett. B 459, 1999

)(2

2 sdM

dM

H(s)

Rho - Resonanz

dds

0.1 0.3 0.5 0.7 0.9M

[GeV2]

a.u

.

Radiatorfunction H(s)

0.1 0.3 0.5 0.7 0.9M

[GeV2]

H. Czyz

DEAR,

e+e- - collider with =m1.0195 GeVs

DANE: A -Factory

Integrated Luminosity

Peak Luminosity Lpeak= 1.5 • 1032cm-2s-1

Total KLOE int. Luminosity:

L dt ~ 2500 pb1 (2001 - 05)

2006:• Energy scan with 4 points around m-peak• 220 pb-1 at = 1000 MeV s

This talk is based on 240pb-1 from 2002 data!

KLOE DetectorDrift chamber

p/p = 0.4% (for 900 tracks)

xy 150 m, z 2 mm

Excellent momentum resolution

p/p = 0.4% (for 900 tracks)

xy 150 m, z 2 mm

Excellent momentum resolution

KLOE DetectorElectromagnetic Calorimeter

E/E = 5.7% / E(GeV)

T = 54 ps / E(GeV) 100 ps(Bunch length contribution subtracted from constant term)

Excellent timing resolution

E/E = 5.7% / E(GeV)

T = 54 ps / E(GeV) 100 ps(Bunch length contribution subtracted from constant term)

Excellent timing resolution

a) Via absolute Normalisation to VLAB Luminosity (as in 2001 analysis):

€

dσππ +γ (γ )

obs

dMππ

2=

ΔNObs − ΔNBkg

ΔMππ

2⋅

1

εSel⋅

1

Ldt∫

€

s( ) ≈ Mππ

2dσ

ππ +γ (γ )

obs

dMππ

2⋅

1

H(s)

Relation between |F|2 and the cross section ee

Obtain from (ISR) - radiative cross section ddM2 via theoretical radiator function H(s):

ddM2 is obtained by subtracting background from observed event spectrum, divide by selection efficiencies, and int. luminosity:

Extracting |F|2 from eve:

1)

2)

3)

€

Fπ2 =

3s

πα 2βπ3

σ ππ s( )

b) Via bin-by-bin Normalisation to rad. Muon events (see later…)

Event Selection

Pion tracks at large angles 50o< <130o

a) Photons at small angles < 15o or > 165o

€

rp=

rpmiss=−(

rp+ +

rp−)

• High statistics for ISR photons• Very small contribution from FSR• Reduced background contamination

Photon momentum from kinematics:

Event Selection

b) Photons at large angles 50o < < 130o

Photon is explicitly measured in the detector!

• Threshold region accessible• Increased contribution from FSR• Contribution from f0(980)

Pion tracks at large angles 50o< <130o

a) Photons at small angles < 15o or > 165o

€

rp=

rpmiss=−(

rp+ +

rp−)

Photon momentum from kinematics:

• High statistics for ISR photons• Very small contribution from FSR• Reduced background contamination

Acceptance 0.3%

Trigger 0.3%

Tracking 0.3%

Vertex 0.3%

Reconstruction filter 0.6%

Particle ID 0.1%

Trackmass cut 0.2%

Background 0.3%

Unfolding effects 0.2%

Total experimental Error 0.9%

Acceptance 0.3%

Trigger 0.3%

Tracking 0.3%

Vertex 0.3%

Reconstruction filter 0.6%

Particle ID 0.1%

Trackmass cut 0.2%

Background 0.3%

Unfolding effects 0.2%

Total experimental Error 0.9%Luminosity ( LA Bhabhas ) 0.6%

Vacuum polarization 0.2%

FSR corrections 0.3%

Radiator function 0.5%

Total theoretical Error 0.9%

TOTAL ERROR 1.3%Statistical error negligible (1.5 Million events) Phys. Lett. B606 (2005) 12

Published result, 2001 data

a(0.35-0.95 GeV2) = (388.7 0.8stat4.9syst) · 10-

10

QuickTime™ and aTIFF (LZW) decompressor

are needed to see this picture.

(e+e- ) nbKLOE

2001 Data140pb-1

Published KLOE Result: Phys. Lett. B606 (2005) 12

• additional online software trigger level introduced to recover cosmic veto inefficiency in 2002 (that gave an inefficiency of up to 30% in 2001)

• Improved offline-event filter reduces its systematic uncertainty to <0.1%, was the largest contribution (0.6%!) to the error in published result

• New event generator BABAYAGA@NLO - theoretical error of Bhabha effective cross section decreases from 0.5% to 0.1% - Bhabha cross section lowered by 0.7%; therefore pion form factor decreases by 0.7%

• Larger data set allows for more refined evaluation of systematic errors associated with selection efficiencies; evaluate all contributions again

• 2002 data less effected by pile-up events from machine background

• Trigger issue in 2001 data resolved (see below)

C. M. Calame et al., Nucl. Phys. B758 (2006) 227

2002 data: improvements wrt 2001

Trigger 2001 update

Impact of update on trigger correction on 2001 cross section:


are needed to see this picture.QuickTime™ and a

TIFF (LZW) decompressorare needed to see this picture.

KLOE 2001 TRG updatedKLOE 2001 published

KLOE 2001 TRG updatedKLOE 2001 published

Changes published value on a by 0.4%



Small angle analysis 2002

€

dσ ππγ

dMππ2 =

Nobs−Nbkg

ΔMππ2 ×

1εSelect.

× 1 L

€

dσ ππγ

dMππ2 =

Nobs−Nbkg

ΔMππ2 ×

1εSelect.

× 1 L

Interference

M2

(GeV2)

Statistics: 242pb-1 of 2002-data 3.4 Million Events



Miss<150 or Miss>1650

Preliminary

M2

(GeV2)

QuickTime™ and aTIFF (Uncompressed) decompressor


Small Angle Analysis• Experimental challenge: Fight background from ee , , reduced by means of kinematical cuts in (trackmass), and likelihood function

• Background from LO-FSR negligible (reduced by acceptance cuts: small ), NLO-FSR not reduced and not a background, efficiency has to be known

• Normalization to integrated luminosity, which is obtained from large-angle- Bhabha events (clean exp. selection)

LO-FSR NLO-FSR

m

m

( ) 0)( 2221

222

222

1 ==+−+−+− γφ qppMpMpM trktrk

rrrr

m

• Surviving background evaluated by MC and subtracted

signal region

Background:Main backgrounds estimated from MC shapes fitted to data distribution in MMain backgrounds estimated from MC shapes fitted to data distribution in MTrkTrk

((eeee



0.42< M2 < 0.44 GeV2

DataMCee

MTrk [MeV]



DataMCee

0.68< M2 < 0.7 GeV2

• Bhabha contamination small • events only fitted below 0.6 GeV2 • -peak is cut by event streaming and elliptical Cut in MTrk

• Excellent agreement on TRK mass spectrum Between data and MC

MTrk [MeV]



Luminosity:At KLOE, Luminosity is measured using „Large angle Bhabha“ events (55o<e<125o)

KLOE is its own Luminosity Monitor!

Generator used for Bhabha cross section:

–BABAYAGA (Pavia group):

eff = (428.00.3stat) nbC. M. Calame et al., C. M. Calame et al., Nucl. Phys. B758 (2006) 227

The luminosity is given by the number of Bhabha events divided for an effective cross

section obtained by folding the theory with the detector simulation.

Systematics on Luminosity

Theory 0.10 %

Acceptance 0.25 %

Background () 0.08 %

Tracking+Clustering

Energy Calibration

0.13 %

0.10 %

Knowledge of s run-by-run 0.10 %

TOTAL 0.10 % theo 0.32% exp = 0.34 %

Quoted theor. accuracy:0.1%

New version of generator (BABAYAGA@NLO) gives 0.7% decrease in cross section

compared to previous version

F. Ambrosino et al. (the KLOE Coll.) Eur.Phys.J.C47:589-596,2006





F=1

(

) w

ith

F=

1 [n

b]

Radiative corrections

Radiator-Function H(s) (ISR):

i) Bare Cross Section divide by Vacuum Polarisation

from F. Jegerlehner: http://www-com.physik.hu-berlin.de/~fjeger/

ii) FSR - Corrections Cross section must be incl. for FSR

FSR corrections have to be taken into account in the efficiency eval. (small angle acceptance,

MTrk) and in the passage M2

FS

R c

on

tr. (

sQE

D)

Net effect of FSR is ca. 0.8%:

- ISR-Process calculated at NLO-level PHOKHARA generator (Czyż, Kühn et.al) Precision: 0.5%

× )(2

2 sdM

dM ππ

ππ

ππγππ σ

σH(s)

Radiative Corrections:

Correcting for FSR energy:

Go from M2 M2



[GeV2]

[

GeV

2 ]MonteCarlo

Use special version of PHOKHARA which Use special version of PHOKHARA which allows to determine whether photon comes allows to determine whether photon comes from initial or final state from initial or final state build matrix which build matrix which relates relates to .

The presence of FSR results in a shift of the measured quantity M2

towards lower values:

M2

M2

ISR only: M2

FSR photon present: M2

(FSR)

M2

Small angle result from 2002 data:



Offline Filter negligible

Background 0.3%

Trackmass/Miss. Mass 0.2% (prelim)

/e-ID 0.3%

Vertex 0.5%

Tracking 0.4%

Trigger 0.2%

Acceptance () negligible

M2 M (FSR corr.) 0.3% (prelim)

Software Trigger 0.1 %

Luminosity 0.3%

Acceptance (Miss) 0.1%

Radiator H 0.5%

Vacuum polarization negligible

Systematic errors on a:

Total= 1.1%

Preliminary

Comparison 2001-2002:



(updated)

Evaluating a with small angle

2001 published result (Phys. Lett. B606 (2005) 12):

Applying update for trigger eff. and change in Bhabha-cross section used for luminosity evaluation:

a(0.35-0.95GeV2) = (388.7 0.8stat4.9syst) · 10-10

€

aμππ =1/4π 3 ds σ (e+e− → π +π −) K(s)

0.35GeV2

0.95GeV2

∫

Dispersion integral for 2-channel in energy interval 0.35 <M2<0.95 GeV2

a(0.35-0.95GeV2) = (384.4 0.8stat4.9syst) ·

10-10

2002 preliminary:

a(0.35-0.95GeV2) = (386.3 0.6stat3.9syst) · 10-10a

(0.35-0.95GeV2) = (386.3 0.6stat3.9syst) · 10-10

Large angle analysis:

Threshold region non-trivial due to irreducible FSR-effects, to be computed from MC using phenomenological models (interference effects unknown)

ff

FSR f0

important! small!

important cross check with small angle analysisthreshold region is accessible photon is explicitly required (allows 4-momentum constraints)

lower statistics for signal FSR not negligible large background irreducible bkg. from decays

240 pb-1

2002 Data

M2 [GeV2]

Nr.

of

even

ts /

0.01

GeV

2

Preliminary



Large angle analysis (cont‘d):

2002 Data L = 240 pb-1

M2 [GeV2]

Apply specific selection cuts:

• Exploit kinematic closure of the event

Cut on angle btw. ISR-photon and missing momentum

• Kinematic fit in hypothesis using 4-momentum and 0-mass as constraints

• FSR contribution added back to cross section (estimated from PHOKHARA generator)

misspr

ãpr

Error on LA dominated by syst. uncertainty on f0 contr.

LA stat.+syst. error

• Reducible background from and µ+µwell simulated by MC

Preliminary

• Model dependence of irreducible background from f0 is the dominating uncertainty. Computed using different models for f0-decay and input from dedicated KLOE f0 analyses (with f0 decaying to charged and neutral pions).





Range of comparison

Comparison SA-LA 2002

Small angle:

a(0.50-0.85GeV2) =

(255.4 0.4stat 2.5syst) · 10-10

Large angle:

a(0.50-0.85GeV2) =

(252.5 0.6stat 5.1syst) · 10-10

Range of comparison

LA stat.+syst. error

LASASA

a between 0.5 - 0.85 GeV2:

Preliminary

(60% (=3.1×10-10) of systematical error due to f0-uncertainty)

Asymmetry:

Binner, Kühn, Melnikov, Phys. Lett. B 459, 1999

check the validity of the FSR model (Phokhara uses sQED, i.e. pointlike pions)

Czyz, Grzelinska, Kühn, hep-ph/0412239

radiative decays of the into scalar mesons decaying to + -contribute to the charge asymmetry

Possibility to study the properties of scalar mesons with charge asymmetry

€

A =N(θ+ > 90o) − N(θ+ < 90o)

N(θ+ > 90o) + N(θ+ < 90o)

€

A =N(θ+ > 90o) − N(θ+ < 90o)

N(θ+ > 90o) + N(θ+ < 90o)

In case of non-vanishing FSR contribution, the interference term

between ISR and FSR is odd under exchange + - .This gives riseto a non-vanishing forward-backward asymmetry:



KLOE DataMC [ISR + FSR]MC* [ISR+FSR+f0(980)]

*G. Pancheri et. al., arXiv:0706.3027

a KLOE summary

Jegerlehner (hep-ph/0703125):

Using the new KLOE result increases difference from 3.2to 3.4

€

Δaμ = aμexp − aμ

the = (28.7 ± 9.1) ⋅10−10



Preliminary



a Summary

Comparison with a from CMD2 and SND in the range

0.630-0.958 GeV :Phys. Lett. B648 (2007) 28 Preliminary

Conclusions:

KLOE has extracted a in the range between 0.35 - 0.95 GeV2

using cross section data obtained via the radiative return with photon emission at small angles.

• The preliminary result from 2002 data agrees with the updated result from the published KLOE analysis based on 2001 data

•Data from an independent and complementary KLOE measurement (Large angle analysis) of the cross section yelds a

in the range between 0.5 - 0.85 GeV2

• All three KLOE results are in good agreement

•KLOE results also agree with recent a results from the

CMD2 and SND experiments at VEPP-2M in Novosibirsk

Outlook:

• Refine the small angle analysis by unfolding for detector resolution, to release the cross section table

• Continue evaluation of resonance contributions in the large angle analysis to decrease model dependence from f0(980) contribution

• Measure the pion form factor via bin-by-bin ratio of pions over muons (Normalization to radiative muon pairs)

• Obtain pion form factor from data taken at s = 1000 MeV (off-peak data)

See next slides…

Extracting |F|2 from norm. to

)(/

/)( s

sdd

sdds Born

obs

obsBorn ′

′′

≈′

(exact only in the absence of FSR)

In presence of FSR:

Corr. due to FSR „Observed“ cross sections

€

Fπ s'( )2⋅ 1+η (s')( ) =

4 1+ 2mμ2 s'( )βμ

βπ3

⋅dσ ππγ ds'( )

ISR+FSR

dσ μμγ ds'( )ISR

Luminosity ( LA Bhabhas ) 0.6%

Vacuum polarization 0.2%

FSR corrections 0.3%

Radiator function 0.5%

Total theoretical Error 0.9%

Some Effects cancel out in the ratio:

x0.3%

needs to select µµ events with similar precision as

separation:separation is achieved in the analysis by a rigid cut in MTrk:

Data

N

o m

an‘s

lan

d

MTrk[MeV]

Additional tool used in evaluation of

efficiencies: ID by means of a Neural Network (NN):



MTrk< 115 MeV

MTrk> 130 MeV

NN >0.7 for NN <0.2 for

Data-- MC

The spectra of selected events for the small angle analysis from 242.62

pb-1 of data taken in 2002:


are needed to see this picture. QuickTime™ and aTIFF (LZW) decompressor


0.87 Mio. events3.42 Mio events

Spectra after SMA selection:

Pions Muons

DANE Off-peak data:

Run-Nr

s (

MeV

)

s = 1023s = 1023

s = 1030s = 1030

s = 1018s = 1018

s = 1010s = 1010

s = 1000s = 1000

• It allows to extract |F|2 down to thresh. in a background-free enviroment• 220 pb-1, competitive with VEPP-2M at threshold• Study model-dependence of f0(980) (in connection with on-peak data)

• Run at √s=1000 MeV between Dec. 2005 and March 2006 220 pb-1 on tape

• Energy scan around -peak

4 scan points, 10 pb-1 each

MC Phokhara

Data 2006 off-peakTagged photons

Preliminary

Forward-Backward-Asymmetry

MM22 (GeV (GeV22))





[email protected]

http://www.lnf.infn.it/conference/phipsi08/

http://www.lnf.infn.it/conference/phipsi08/

Backup slides

Reminder: Comparison with e+e- and - Data

s [GeVs [GeV22]]

Relative difference btw. publishedKLOE (2001) and CMD-2, SND



/

KL

OE -

1

interpolation of 60KLOE data points from

0.35 to 0.95 GeV2Plot from I. Logashenko using non-published data from CMD-2 and SND!

KLOE published

SND hep-ex/0605013CMD-2 published

CMD-2 prel. ALEPH -data

••••••

Relative difference btw. e+e Data and -spectral function from ALEPH

s [GeV]s [GeV]

KLOE confirmed the CMD-2 discrepancy with tau data and “...invalidates the use of -data until a better understanding of the discrepancies is achieved” (A. Höcker ICHEP-04)Some disagreement btw. KLOE and SND/CMD-2 seen at low and high masses aµ

had from all recent e+e- experiments agree now within 0.5 Waiting for new results from KLOE (this talk), Babar, and data (Belle)

Documents

Graziano Venanzoni Laboratori Nazionali di Frascati KLOE results on Hadronic cross section Topical Workshop on (g-2) Glasgow, 25-26 Oct 2007 (for the