BINP Novosibirsk Luminosity measurement with CMD3 detector at

BINP, Novosibirsk Luminosity measurement with CMD-3 detector at the VEPP-2000 e+e- collider G. V. Fedotovich on behalf of the CMD-3 collaboration WG 13 -14 September LNF, Frascati, Italy

Motivation As a rule all hadronic cross sections in experiments at e+e- colliders are normalized on the integrated luminosity For luminosity determination it is necessary to use well known QED processes which have the large cross sections and a simple signature in detector e+e-, , + - (cross check capability to believe 1% presic. ) Measurement of the cross section e+e- hadrons in the low energy range is interesting for: ✗ measurement of parameters of light vector mesons ρ, ω, φ, ρ’’, ω’’ ✗ Test of QCD sum rules, Ch. PT, VDM, . . . etc, search of exotics (light hybrids and glueballs) Luminosity determination at least better than 0. 5% is required. Hadron contribution to (g-2) of muon is about 60 ppm: 60 0. 005 = 0. 3 ppm The aim new FNAL experiment for (g-2)/2 is to improve BNL result by a factor of 4! and to achieve accuracy 0. 15 ppm - good test of SM 2

VEPP-2000 storage ring ILU 3 Me. V Linac CMD-3 B-3 M VEPP-2000 BEP 250 Me. V synchrobetatron e+, e booster 825 Me. V SND e e+ converter • Up to 2 Ge. V c. m. energy Status: • VEPP-2000 uses special optics Plans: ≈100 pb-1 per detector per year 2010 – start of experiments providing “round beams” give additional gain in luminosity and will provide: L 1032 cm-2 c-1, √s=2. 0 Ge. V

VEPP-2000 SND CMD-3 4

VEPP-2000 with round beams Integrated luminosity ~ 100 pb-1 per detector/year revolution time – 82 ns beam length – 3. 3 cm circumference – 24. 4 m L = 1032 cm-2 s-1 at 2. 0 Ge. V beam current – 200 m. A energy spread – 0. 7 Me. V beta function in IP x= z =4. 3 cm Lpeak = 1031 cm-2 s-1 at 1 Ge. V CMD-3 SND 5

3 D view CMD-3 detector DC – 1218 hexagonal cells with sensitive wires, W-Re alloy, 15 in diameter. Z-chamber – start FLT, precise determine z-coordinate ~ 500 (acceptance) LXe calorimeter thickness 5 X 0, 264 towers & 2112 strips. Spatial resolution 1 – 2 mm Calorimeter with Cs. I crystals ( 3, 5 t), 8 octants, number of crystals - 1152, 8 X 0. TOF – 16 counters, time resolution ~ 1 ns Muon range system – 8 octants (cosmic veto, time resolution ~ 1 ns ) Value 1. 3 T is achieved 6

Collected Luminosity Averaged over run The 1031 cm-2 c-1 luminosity at √s=2. 0 Ge. V Currently the luminosity at high energy is limited by a deficit of positrons and maximum energy of the booster (825 Me. V now) After upgrade it will gain factor x 10 Collected L ~ 60 pb-1 per detector 8. 3 pb-1 ω - region 9. 4 pb-1 < 1 Ge. V (except ω ) 8. 4 pb-1 φ - region 34. 5 pb-1 > 1. 04 Ge. V 7

First run: winter-spring 2011 event e+e- R-z plane R- plane 8

Event selection (e+e-) Two collinear tracks in DC, π-1. 0>( θ 0 +(π- θ 1))/2 > 1. 0 E 2, Me. V E 2 vs E 1 m. i. p. e+ e- → e + e- E 1, Me. V Energy cuts: Ebeam/2 < E 0, E 1< 3*Ebeam/2 9

Event e+e- -> γγ in CMD-3 detector R-z plane R- plane 10

Events selection e+e- -> γγ Back-to-back cluster in calorimeter with: angle cuts π - 1. 0 > ( θ 0 +(π-θ 1))/2 > 1. 0 energy cuts Ebeam/2 < E 0, E 1< 3*Ebeam/2 E 2, Me. V E 2 vs E 1 e+e- -> γγ E 1, Me. V 11

First step – collinear events selection 1 2 3 4 Hit points on Total charge Accolinearity track in DC > 5 (max 19) = 0 angle in R - plane: ||φ1 -φ0|-π|<0. 15 rad angle in R – z plane: |θ 1+θ 0 -π|<0. 25 rad Second step – Bhabha events selection Event is Bhabha, if: 1 2 3 4 Number of cluster in LXe calorimeter is 2 exactly Angle (π - 1. 0) < ( θ 1 lxe- θ 0 lxe + π )/2 < 1. 0 rad Energy of every cluster E 1, 2 > Ebeam/2 Number of hitted sectors in ZC > 1 12

Luminosity determination L = Ne+e-/(σBorn rad ε²dc ε²cl tr) Ne+e-- number of the detected Bhabha events σBorn - Born cross section rad - radiation correction MCGPJ ( 95 ± 0. 2% ) εDC - track reconstruction effic. in DC (99. 2 ± 0. 08)% Events are selected using calorimeters information only: Two back-to-back clusters in calorimeters with a good quality and look for tracks in DC which belong to cluster εcl - cluster reconstruction effic. in calorim. (99. 5± 0. 2)% Two back-to-back tracks in DC and look for clusters in LXe calorimeter which belong to tracks tr - charge trigger efficiency was based on neutral 13 trigger information

Track reconstruction efficiency in DC 1 lxe-φ0 lxe -π, rad ||φ1 lxe-φ0 lxe-π| - φrot | < 0. 03 Ebeam/2 < E 0, E 1< 3/2*Ebeam θ 1 lxe + θ 0 lxe -π, rad | θ 1 lxe + θ 0 lxe -π| < 0. 3 εdc = N 2 tr_coll / Ntotal ~ 99. 6 % 14

Collinear γγ events in CMD-3 no tracks in DC back-to-back clusters in calorimeter | | φ1 -φ0|-π | < 0. 2 | θ 1 + θ 0 -π| < 0. 6 15

Ratio of luminosities Red points – scan up Number of points – 39 Blue points – scan down Integr. Luminos. ~ 22 pb-1 Previous analysis Current analysis 0. 2 +- 0. 32 Only statistical errors are shown 16

Events separation based on momentum distribution Е = 160 Мэ. В Е = 250 Мэ. В 17

e/ / separation using particles momentum Can measure N( )/N(ee) & compare with QED CMD-2 CMD-3: -1 -2 0. 5 1. 3 % % CMD-3 -1 0. 5 % CMD-2: -2 1. 3 % 18

Ratio of luminosities First scan low energy region-this season Number of points – 10 Integrated Luminosity ~ 3 pb-1 Only statistical errors are shown 19

Ratio of luminosities First scan low energy region-this season Number of points – 28 Integrated Luminosity ~ 6 + 8 pb-1 Only statistical errors are shown 20

e+e- π+π- by CMD 3 Clean collinear events (mostly without background) Plans to reduce systematic error from 0. 6% -> 0. 3%: ✗ Event separation will be checked by different methods 0. 2% ✗ More proof of Radiative corrections 0. 2% -> 0. 1% ✗ Determination of fiducial volume controlled independently by LXe and ZC subsystems (0. 1%) Statistical precision of cross section measurement Collected L ~ 60 pb-1 per detector 8. 3 pb-1 ω - region 9. 4 pb-1 < 1 Ge. V (except ω) 8. 4 pb-1 φ - region 34. 5 pb-1 > 1. 04 Ge. V

Separation efficiency vs energy Important – overlap energy region Pi. Pi by momentum Mu. Mu by momentum By energy 22

Summary Peak L~2 x 1031 cm-1 sec-2 has been reached, but limited positrons production rate. “Unlimited” positron source is in preparation. Hope for X 10 in luminosity and statistics Luminosity was measured using two processes: e+e- → e+e-, . The currant systematic error for integrated luminosity is estimated as 1% Both detectors are taking data: first scan from 1 Ge. V to 2 Ge. V with 25 Me. V step with ~0. 5 pb-1/point was completed in 2011. This region was scanned again with smaller step and 1 pb-1/point in 2012 The new scan with precision energy measurement is done during this season (2 E < 1 Ge. V) First preliminary results for the ratio (N /Nee)/( / ee) are demonstrated: QED check (– 1 0. 5)%. Need 0. 2% or better (MCGPJ) 23

Short second part of my talk

BINP, Novosibirsk MCGPJ for the processes e+e- hadrons for experiments with CMD-3 detector at the VEPP-2000 collider G. V. Fedotovich WG 13 -14 September LNF, Frascati, Italy

VEPP-2000 with round beams Integrated luminosity ~ 100 pb-1 per detector/year revolution time – 82 ns beam length – 3. 3 cm circumference – 24. 4 m L = 1032 cm-2 s-1 at 2. 0 Ge. V beam current – 200 m. A energy spread – 0. 7 Me. V beta function in IP x= z =4. 3 cm Lpeak = 1031 cm-2 s-1 at 1 Ge. V CMD-3 SND 26

3 D view CMD-3 detector DC – 1218 hexagonal cells with sensitive wires, W-Re alloy, 15 in diameter. Z-chamber – start FLT, precise determine z-coordinate ~ 500 (acceptance) LXe calorimeter thickness 5 X 0, 264 towers & 2112 strips. Spatial resolution 1 – 2 mm Calorimeter with Cs. I crystals ( 3, 5 t), 8 octants, number of crystals - 1152, 8 X 0. TOF – 16 counters, time resolution ~ 1 ns Muon range system – 8 octants (cosmic veto, time resolution ~ 1 ns ) Value 1. 3 T is achieved 27

Motivation Hadronic contribution to (g-2)/2 of muon is about 92 % in the VEPP-2000 energy range One of the aim of experiments with CMD-3 is to measure the main multihadrons CS with systematic uncertainly smaller than 3% ( 0. 1 ppm) Systematic accuracy HCS at least better than 0. 5% is required. Hadronic contribution to AMM of muon is about 60 ppm: 60 0. 005 = 0. 3 ppm The aim new FNAL experiment for (g-2) is to improve BNL result by a factor of 4! and to achieve the accuracy 0. 15 ppm - good test of SM Obviously that the systematic uncertainly of RC for the MHCS should be smaller than 1% Measurement of the cross sections e+e- hadrons is interesting for: ✗ measurement of parameters of light vector mesons ρ, ω, φ, ρ’’, ω’’ ✗ Test of QCD sum rules, Ch. PT, VDM, . . . etc, search of exotics light 28 hybrids and glueballs, forbidden and rare decays and so on…

Motivation Number of channels with many hadrons in FS is more 15! with cross sections bigger 1 nb Results of analyses for six pion channel is published. Analyses for the channel K+K- + - is on progress. At least four intermediate states are seen in both cases. To build MCGPJ for the collinear events (e+e-, + -, ) it was taken about 3 year collaboration with our colleagues at Dubna (Kuraev & Arbuzov) We have not enough man-power to construct MCGPJ with similar accuracy Previous experience for 3 pion channel – FSR contributes to CS at the level 0. 3% only Obviously the FSR will contribute to CS smaller than 0. 3%! In the scale of 1% we can neglect by it and consider only ISR 29

Current approach In frame the SF approach the photon jets radiated in collinear regions inside narrow cones around electrons and positrons There are not known narrow resonances in the energy band from 1 to 2 Ge. V After some discussions with our experts in Dubna we chose the next strategy: Energy E between soft and hard photons is chosen as big as possible ( 10 – 100 Me. V) E is large enough – probability to radiate two hard photons is smaller than 1% (depends on CS behavior vs energy) For > E only one photon is considered but with known angular distribution which we assign by hand Total CS vs E shown stability (constant value) inside corridor 0. 5% It was checked for two and three pion channels Iteration approach: first step vis input 1 2 and so on… New ideas and approaches are welcome!!! 30

Collected Luminosity Averaged over run The 1031 cm-2 c-1 luminosity at √s=2. 0 Ge. V Currently the luminosity at high energy is limited by a deficit of positrons and maximum energy of the booster (825 Me. V now) After upgrade it will gain factor x 10 Collected L ~ 60 pb-1 per detector 8. 3 pb-1 ω - region 9. 4 pb-1 < 1 Ge. V (except ω ) 8. 4 pb-1 φ - region 34. 5 pb-1 > 1. 04 Ge. V 31

Preliminary results for the cross section of the process e+e- -> K+K- + Integr. Lum. 23 pb-1 N(K+K- + -) 30 000 Ba. Bar CMD-3

Radiative correction 33

Conclusion and Outlook CMD-3 collected data in a broad energy region. Particularly inside energy band from 1 to 2 Ge. V integrated luminosity 35 pb-1 L~2 x 1031 cm-1 sec-2 has been reached and limited while by positron rate production. “Unlimited” positron source is in preparation. Hope for X 10 in luminosity and statistic For some hadronic channels statistical errors are smaller 5% and will be improved in future runs. We can meet statistical accuracy 1% for some MH channels MC generators for these channels with RC desirable should has systematic uncertainly smaller 1% MCGPJ based on SF approach is in progress now New ideas how to build MC generators for MH channels and to reach systematic accuracy smaller than 1% are WELCOME!!! 34

Back slides 35

Results of the pion separation 36

SM prediction for muon g-2 aμexp = (g-2)μ/2 = (11 659 208. 9 ± 6. 3)x 10 -10 a (Exper. – Theor. ) ~ 3. 3 -3. 6 δaμ Main channels which contribute to precision at √s<1. 8 Ge. V π+π− 505. 65 ± 3. 09 π+π− 2π0 18. 62 ± 1. 15 π+π−π0 47. 38 ± 0. 99 (mostly from omega region) 2π+2π− 13. 64 ± 0. 36 (Ba. Bar) K+K 22. 95 ± 0. 26 (Ba. Bar) Isospin relations for √s<2 Ge. V : 12. 46 ± 0. 76 for not measured KK , KK 2π, 2π4π0, 2π3π0 It is crucial in case of isospin violation Rqcd[2 -11 Ge. V] LBL Theory TOTAL 41. 19 ± 0. 82 10. 5 ± 2. 6 Prades, de Rafael & Vainshtein need more theory ± 4. 9 37

R(s) measurements at low s At low s R(s) has to be measured in each channel. The value and the error of the hadronic contribution to muon (g-2) are dominated by low energy R(s) (<2 Ge. V give 92%). a. QED(MZ) – half of error comes from 2 E < 4 Ge. V 38

Accuracy of energy measurement (current status) 0. 1 Мэ. В 0. 3 % systematic 0. 16 Мэ. В 0. 1 % systematic 39

SM prediction for muon g-2 aμexp = (g-2)μ/2 = (11 659 208. 9 ± 6. 3)x 10 -10 aμth = aμQED + aμEW + aμhadr (Exp. – Theor. )~3. 3 -3. 6 δaμ QED contribution EW contribution NLO hadronic 11 658 471. 808 ± 0. 015 Kinoshita & Nio, Aoyama et al 15. 4 ± 0. 2 Czarnecki et al − 9. 8 ± 0. 1 HLMNT 11 694. 1 ± 4. 3 x 10 -10 HLMNT 11 Main channels which contribute to precision at √s<1. 8 Ge. V π+π− 505. 65 ± 3. 09 π+π− 2π0 18. 62 ± 1. 15 π+π−π0 47. 38 ± 0. 99 (mostly from omega region) 2π+2π− 13. 64 ± 0. 36 (Ba. Bar) K+K 22. 95 ± 0. 26 (Ba. Bar) Isospin relations: 5. 98 ± 0. 42 for not measured KK , KK 2π, 2π4π0, 2π3π0 12. 46 ± 0. 76 for √s<2 Ge. V) (1. 5 - 3σ of total error - crucial in case of isospin violation) Rqcd[2 -11 Ge. V] LBL Theory TOTAL 41. 19 ± 0. 82 10. 5 ± 2. 6 Prades, de Rafael & Vainshtein need more theory ± 4. 9 40

Accuracy of energy measurement 41

Results of the muon separation 42

e+e- -> π+π- by CMD 3 |Fπ|2 Ve ry e/ / separation using particles momentum p re lim i na ry e/ / separation using energy deposition in calorimeter 43

Event selection efficiency is based on only energy deposition in calorimeters en = N / Ntot 0. 8 Ebeam < E 1 < Nee real = Nee / ε 2 en 1. 2 Ebeam en ~ 99% 44

45

46

47

48

49

50

Back slides 51