1. Study of e + e  collisions with a hard initial state photon at BaBar Michel Davier (LAL-Orsay) for the BaBar collaboration TM

2. 2 Initial state radiation (ISR) method High PEP-II luminosity at  s = 10.58 GeV  precise measurement of the e + e - cross section  0 at low c.m. energies with BaBar. Improved hadron spectroscopy Input to (g  -2) and  em calculations. Few previous data in the 1.4-3.0 GeV range. Comprehensive program at BaBar. Today: results for  +    0, preliminary results for 2  + 2  , K + K   +  , 2K + 2K  from 89.3 fb -1  ISR

3. 3 ee  0ee  0 Event selection: Isolated ISR photon with E CM > 3 GeV At least 2 good photons with E  > 0.1 GeV Two good, non-K tracks from IP Kinematic fit: Energy and momentum balance enforced Mass of two soft photons constrained to  0  2 < 40 for fit in      0  hypothesis selects signal events Reject events with extra photons if  2 < 40 for      0  0  hypothesis   J/  Excited  states ? Events/0.01 GeV/c 2

4. 4 Background for 3  Most dangerous backgrounds: e + e -  K + K -   , e + e -  qq          Other backgrounds: e + e -  2 , 4 , 5 , …,  +  -,      0 Two methods of background subtraction: 1. background mass distribution measured in data, subtracted bin-by-bin from signal mass distribution 2. taken from simulation, corrected to real experimental distribution Total background level: (0.5 - 1.5)% in ,  regions (15 - 50)% at higher masses accuracy in background level ~25% up to 2 GeV Data MC

5. 5 Detection efficiency for 3  The detection efficiency  (m) : determined from a Monte Carlo simulation that includes additional corrections extracted from special control event-samples quite uniform in 0.5<m<3 GeV/c 2 systematic error currently ~4% - will be improved with more data

6. 6 Fit of the      0 mass spectrum  3  (m) - Born cross section of e + e -   +  -  0 is a coherent sum of 4 resonances: , , ,  R(m)~1 - radiative correction function from calculation dL/dm - ISR luminosity taken from total integrated luminosity and photon radiator function W(s,x); (checked with  events) f(m,m) - detector resolution taken from simulation (with floating extra Gaussian smearing) Fix ,  widths to PDG values Fix  -  relative phase to experimental value (163  7)  Fix  -  relative phase to 180 ,  -  to 0 

7. 7 Fit results:  -  region  2 /d.f.=146/148 consistent with known properties of these resonances ( ,  widths fixed to PDG values)   The resolution is about 6, 7, 9 MeV/c 2 at , , J/  masses. Fitted resolution smearing is ~1 MeV/c 2 BaBar PDG B(  ee)B(  3  )=(6.70  0.06  0.27)  10 -5 (6.35  0.10)  10 -5 B(  ee)B(  ) =(4.30  0.08  0.21)  10 -5 (4.52  0.19)  10 -5

8. 8  Fit results: higher mass region BaBar PDG B(  ee)B(  3  )=(0.82  0.05  0.06)  10 -6 M(  )= 1350  20  20 MeV/c 2 1400 - 1450  (  )= 450  70  70 MeV 180 - 250 B(  ee)B(  3  )=(1.3  0.1  0.1)  10 -6 M(  )= 1660  10  2 MeV/c 2 1670  30  (  )= 230  30  20 MeV 315  35 Good fit obtained for the range up to 1.8 GeV/c 2. Extending the fit to masses above 1.8 GeV/c 2 may require a more complicated fitting function taking into account non-resonant 3  production. Mass and width parameters are dependent upon our assumed phases - interference effect is strong

9. 9 e + e -   +    0 cross section DM2 BaBar SND coverage of wide region in this experiment - no point-to-point normalization problems consistent with SND data E C.M. < 1.4 GeV inconsistent with DM2 results overall normalization error ~5% up to 2.5 GeV

10. 10 J/   3  decay The J/  meson is narrow - clean signal After sideband subtraction - N J/  = 920  34 Detection efficiency -  =(9.2  0.6)% The result:  (J/   ee)  B(J/   3  )= 0.122  0.005  0.08 keV We previously measured (  )  (J/   ee)= (5.61  0.20) keV Events/0.004 GeV/c 2 B(J/   3  ) = (2.18  0.19)% BaBar (1.50  0.20)% PDG (2.10  0.12)% BES 2003 Phys. Rev. D69,011103 (2004) sideband

11. 11 e  e   2   2   , K  K      , 2K  2K   Event selection: Isolated ISR photon with E CM > 3 GeV At least four good tracks from IP Kinematic fit: Energy and momentum balance enforced Energy and angles of hard ISR photon are not used - 1C fit Fit in 3 hypotheses: 4  for all events 2K2  if 1 or 2 identified kaons 4K if 2, 3 or 4 identified kaons Other ISR processes (5 , …) – using difference in  2 distributions e + e -  qq – using JETSET simulation Background subtraction:

12. 12 e  e   2   2   cross section Systematic errors: 12% for m 4  < 1 GeV, 5% for 1 < m 4  < 3 GeV, 16% for higher masses) best measurement above 1.4 GeV Intermediate states: a 1 (1260)  - dominant, structure which may be f 0 (1370)  final state is seen. For detailed study, a simultaneous analysis of     and       final states is required. BaBar Preliminary Coverage of wide region in one experiment No point-to-point normalization problems

13. 13 e  e   2   2   cross section Good agreement with direct e  e  measurements Most precise result above 1.4 GeV

14. 14         substructures BaBar preliminary MC generator: H.Czyz and J.H.Kuehn, Eur.Phys.J C18(2000)497-509 Includes a 1 (1260)  and f 0 (1370)  Does not include J/  a 1 (1260)  (770) f 0 (1370) J/ 

15. 15 e  e   K  K      Systematic error – 15% (model dependence, kaon identification) Substantial resonance sub-structures observed: K*(890)K  dominant  KK contribute strongly K* 2 (1430)K  seen. Much more precise than previous measurement J/ 

16. 16 K  K      substructures No studies in previous e + e - experiments! K*K  MC generators are not available yet   K* regions excluded K*(890)K  dominated  f 0 (980) ? Events from  band No signal from  f 0 (980) yet Connection to  f 0 (980)  ? BaBar preliminary

17. 17 e  e   2K  2K  First measurement Overall normalization systematic error – 25% (model dependence, kaon identification) No clear mass structure in the two- or three-body subsystems No  ’ s ! BaBar preliminary J/ 

18. 18 J/  and  (2S) decays BaBar preliminary PDG B(J/           =(3.61  0.26  0.26)  10 -3 (4.0  1.0)  10 -3 B(J/   K + K -      =(6.09  0.50  0.53)  10 -3 (7.2  2.3)  10 -3 B(J/   K + K - K + K - ) = (6.7  1.1  1.0)  10 -4 (9.2  3.3)  10 -4 B(  (2S)  J/   +  - )=0.361  0.015  0.028 0.317  0.011  +  - J/  +-+- 2+2-2+2- K+K-+-K+K-+- 2K + 2K -

19. 19 Summary Using ISR method the cross sections of e + e        ,    , K + K     , 2K + 2K  reactions have been measured from threshold to 4.5 GeV. These are the most precise measurements to date for c.m. energies greater than 1.4 GeV. Examples: contributions to a  had (  10  10 ) from 2  + 2   (0.56 – 1.8 GeV) from all e + e  exp. 14.21  0.87 exp  0.23 rad from all  data 12.35  0.96 exp  0.40 SU(2) from BaBar 12.95  0.64 exp  0.13 rad from  +    0 from all e + e  exp. 2.45  0.26 exp  0.03 rad from BaBar 3.31  0.13 exp  0.03 rad Several B(J/  -> X) measurements better than current world average More modes to come; aim for systematic errors  1% (in  +   ) Davier-Eidelman- Hoecker-Zhang 2003 total 696.3  7.2