1. Studying FSR at threshold for  events O. Shekhovtsova, G.V. lnf 3/1/05 Radiative meeting Disclaimer: everything is preliminary!!!

2. Generalization of FSR amplitude: we need 3 form factors f i.  Limit of soft photon (what we call sQED) Each form factor depends by 3 independent variables (one is s) At threshold (very hard photon) this approximation could not work

3. A model for FSR, based on  PT f i =f i sQED +  f i (S. Dubinsky et al,hep-ph/041113) The contribution with  +,- (instead of a 1 ), turns out to be negligible MesonM(GeV)G V (GeV)F V (GeV)F A (GeV)  0.7750.00660.156- a1a1 1.23--0.122

4. The model is a first step in the direction of a generalization of the FSR amplitude. However, the calculation (at the moment) is approximate since: - it doesn’t take into account other mesons (like  and  ’) - it doesn’t include the      decay - …. What about      decay ? Recently H. Czyz et al. (hep-ph/0412239) showed that this decay can be important also at low Q 2 region. This contribution can be quite different depending by the model. Charge asymmetry can help.

5. How to include      decay in our calculation? For the moment we consider only the contribution of f 0 (no  meson). This could be too crude at low Q 2 !!!      is related to      by the same matrix element (a part a factor ½). We use the Achasov 4q parametrization with the parameters of the model taken from the fit of the kloe data      m  (MeV) dBR/dm x 10 8 (MeV -1 )

6. (Analitical) Comparisons (at s=m  2 ): = FSR sQED +  f Since M FSR *M   |M  | 2 at low Q 2  f can be relevant only for destructive. interference (we will consider only this case in the following) 0 o <    0 o 0 o <    0 o =      resonant cont. = FSR sQED =  f Q 2 (GeV 2 ) What happens for s<m  2 ? d  /dQ 2 (nb/GeV 2 )

7. The comparison at s=1 GeV 2 : = FSR sQED +  f 0 o <    0 o 0 o <    0 o = 100·      resonant cont. = FSR sQED =  f Q 2 (GeV 2 ) Multiplied by a factor 100 1/|D  (s)| 2 In this case the interference M FSR *M  is expected to be >>|M  | 2 d  /dQ 2 (nb/GeV 2 )

8. The following matrix element has been introduced in a MC, for e + e -       (based on EVA structure): We neglect the contributions from        found negligible in hep-ph/0411113) We consider destructive interference between FSR sQED and  We consider large angle analysis: 50 o <    0 o, 50 o <    0 o

9. Numerical results: differential cross section… 50 o <    0 o 50 o <    0 o = ISR+FSR sQED +  = ISR+FSR  PT +  = ISR+FSR sQED d  /dQ 2 (nb/GeV 2 ) d  TOT /d  sQED d  sQED  /d  sQED d  TOT /d  sQED  Q 2 (GeV 2 ) Effect al low Q 2 …however the contribution of  is not much accurate (no interference with  has been taken into account)

10. And asymmetry…. Q 2 (GeV) ISR+FSR sQED ISR+FSR sQED + 

11. Zoom on the threshold region: = ISR+FSR sQED +  = ISR+FSR  PT +  = ISR+FSR sQED d  /dQ 2 (nb/GeV 2 ) Q 2 (GeV 2 ) d  TOT /d  sQED d  sQED  /d  sQED d  TOT /d  sQED  Up to 30% of contribution beyond sQED at the threshold

12. And asymmetry… = ISR+FSR sQED +  = ISR+FSR  PT +  = ISR+FSR sQED Q 2 (GeV 2 )

13. Conclusions and outlook First MC results on a generalization of FSR using  PT have been presented. A sizeable effect can be seen on the cross section (at low Q 2 ). The situation on the asymmetry is less clear, but it strongly depends on the parametrization of the  direct decay For the near future: Improve the simulation: better parametrization of  (including also the  meson) consistency between the various parameters of the models in MC Try to disentangle the various contributions: Improve the knowledge on the phi decay (in particular at low Q 2 ): Constraint fit with the neutral channel asymmetry,and other kinematical variables (angular distributions) Work off resonance Model independent analysis of f i ?