Improving the Z line shape modeling Nikola Makovec

Improving the Z line shape modeling Nikola Makovec Group meeting

Generator level § Z line shape described by the relativistic Breit-Wigner. § In proton-proton collisions, the mass spectrum of the Z boson differs from the Breit-Wigner shape of the partonic process cross section. The probability that a quark and antiquark in the interacting pp system produce an object of mass M falls with increasing mass. In order to take this into account, the Breit. Wigner is multiplied by an ad-hoc parametrization L(M). § Photon propagator and the photon-Z interference term not taken into account in the parametrization the parton luminosity term contains also these 2 terms. Nikola Makovec Without L(M) : chi 2/ndf=3. 96 With L(M) : chi 2/ndf=1. 09 2

Resolution § Finally, in order to take into account the finite resolution of the electromagnetic calorimeter, the Breit-Wigner multiplied by the parton luminosity term is convoluted with a Gaussian: Nikola Makovec 3

Results L=200 pb-1 Residual distribution Truth Fit Mean : 0. 0033+/-0. 0003 Sigma : 0. 0049+/-0. 0002 Nikola Makovec 4

Bremsstrahlung Energy/mass resolution haven’t a Gaussian behaviour Need to take into the effect of the bremsstrahlung Nikola Makovec 5

New resolution function Analytical solution =0. 01 =0. 5 =1 Difficult to give a physical interpretation to the parameters Sigma is related to the detector resolution but it is not. Nikola Makovec 6

§ § R(x) depends on electrons eta § Stochastical term depends on eta § Bremsstrahlung effect depends also on eta The calorimeter was divided in 24 regions in eta § 24*24 R(x)’s Nikola Makovec 7

Statistics and chi 2/ndf § For couple of regions with more 50 events, a fit is performed § Otherwise a default parameterization is used corresponding to the whole statistics Nikola Makovec 8

Sigma and beta § Beta is related to the bremsstrahlung effect § Sigma is related to the resolution Nikola Makovec 9

R(x)’s Nikola Makovec 10

Results Mean : -0. 0004+/-0. 0002 Sigma : 0. 0040+/-0. 0002 Mean : 0. 0033+/-0. 0003 Sigma : 0. 0049+/-0. 0002 Improvement on the resolution and on the energy scale Nikola Makovec 11

Results : |eta|<1. 4 Mean : 0. 0028+/-0. 0003 Sigma : 0. 0039 +/- 0. 0002 Mean : -0. 0008 +/-0. 0002 Sigma : 0. 0034 +/- 0. 0002 Improvement on the resolution and on the energy scale Nikola Makovec 12

R(x)’s from ideal 0 § A perfect knowledge of the amount of § material in front of the calorimeter were assumed since the R(x)’s were computed with misal 1 data. Now, R(x)’s determined with ideal 0 data. No cut on eta Mean : -0. 0003 0. 0002 Sigma : 0. 0043 0. 0002 Almost same results Nikola Makovec 13

Final cross-check Before corrections After corrections Crack regions removed Nikola Makovec 14

Conclusion § The Z line shape modeling has been improved § Bremsstrahlung effect is taking into account § Eta dependences are taking into account § Improvement on the systematical uncertainties on the absolute energy scale : § Few per mil less than one per mil § Resolution is also improved § 0. 49 0. 43 Nikola Makovec 15

Back up Nikola Makovec 16

Roo. Keys Ideal 0 Misal 1 Nikola Makovec 17

Fit ideal 0 with pdf from ideal 0 Mean : 0. 0122 0. 0003 Sigma : 0. 0059 0. 0003 Nikola Makovec 18

Fit misal 1 with pdf from ideal 0 Mean : 0. 0123 0. 0003 Sigma : 0. 0047 0. 0003 Nikola Makovec 19

Final checks Nikola Makovec 20

ideal 0 Misal 1 Nikola Makovec 21

Bad chi 2/ndf Nikola Makovec 22