•Report on fitting FINAL new data:

• e- CC (175pb-1: P=0.30, 71pb-1, P=-0.27, 104pb-1)(DESY-08-177)

• e- NC (169pb-1, P=+0.29, P=-0.27)(DESY-08-202)

•Now add CCe+ (134pb-1: P=0.32, 78pb-1, P=-0.36, 56pb-1)(R Ciesielski)

•Then add NCe+ ( 113pb-1 : P=0.32, 71pb-1, P=-0.36, 42pb-1) (M Wlasenko)

•Improvement to PDF uncertainties

•Electroweak parameter fits: MW , MW/ GF ,MW/ g

•Electroweak NC couplings: au ad vu vd

ZEUS fits-report -High Q2 sessionApril 2009

A.M.Cooper-Sarkar and C. Gwenlan

PDFs from the new fit compared to published

ZEUS-JETS fit – central values of the PDFs do not change much…..

Adding New NC and CCe- data to ZEUS-JETS fit → ZEUS-pol fit

(NEW)

This takes 577 data points up to 828 data points

(NEW)

Dataset 2/ndp ndp

CC-ve poln. 1.33 37

CC+ve poln. 0.96 34

NC-ve poln. 0.99 90

NC+ve poln. 0.96 90

Quadratic 2

Remember we have shown this before with preliminary e- data- this is now final

The xuv PDF uncertainties are reduced at high-x ….. as expected since NC and CC e- cross-sections are both u dominated

However some other uncertainties increase marginally- this can always happen when OFFSETing correlated errors

(NEW)

The d-flavour sector is less well known- NOTE the scale difference-

hence we look forward to CCe+ data

PDFs from the new fit compared to

ZEUS-pol fit – central values of d-valence shift a little…..

Adding New CCe+ data to ZEUS-pol fit

This takes 828 data points up to 893 data points

Dataset 2/ndp ndp

CC e-p P=-. 1.33 37

CC e-p P=+. 0.96 34

NC e-p P=-. 0.99 90

NC e-p P=+. 0.96 90

CC e+p P=-. 1.00 32

CC e+p P=+. 1.75 33

Quadratic 2

Note only statistical errors so far on CCe+p

The xdv PDF uncertainties are reduced at high-x ….. as expected since CC e+ cross-sections are d dominated

Very marginal improvement in high-x Sea as well

The d-flavour sector is improved with the addition of CCe+ data

PDFs from the new fit compared to

All that came before – central values shift very little

Adding New NCe+ data to ZEUS-pol +Cce+p

This takes 893 data points up to 1073 data points

Dataset 2/ndp ndp

CC e-p P=-. 1.33 37

CC e-p P=+. 0.96 34

NC e-p P=-. 0.99 90

NC e-p P=+. 0.96 90

CC e+p P=-. 1.00 32

CC e+p P=+. 1.75 33

NC e+p P=-. 1.53 90

NC e+p P=+. 1.68 90

Quadratic 2

Note only statistical errors so far on CCe+p

And only stat. +uncorreated sys. On NCe+p

The xuv PDF uncertainties are reduced at high-x ….. as expected since NC e+ cross-sections are u dominated

The u-flavour sector is further improved with the addition of NCe+ data

ALL New NC/CC e- and e+ compared to

ZEUS -JETS

Electroweak parameter fitsNow let MW be a free parameter of the fit, together with the PDF parameters

How does MW enter the fit?In the factor GF

2 MW4/(Q2+MW

2)2

77.3 ± 0.7 ± 1.0

78.06 ± 0.66 ± 0.95

ZEUS-jets RTVFN including NEW CC/NC e-data ZEUS-jets RTVFN including NEW CC/NC e-data and NEW CC/NC e+ data

77.6 ± 1.4 ± 2.5

78.9± 2.0 ±1.8+2.2-1.8(PDF)

82.87 ± 1.82(exp) +0.32 0.16(model)

ZEUS HERA-I data done by this method

ZEUS DESY-03-093 fit to dσ/dQ2 data

H1 HERA-I data done by this method

Value of MW (=80.4 SM) Specifications of the fit

Little correlation between EW and QCD sector of fit

Now consider fits to electroweak NC couplings as well as PDF parameters

F20 = Σi Ai

0(Q2) [xqi(x,Q2) + xqi(x,Q2)]

xF30= Σi Bi

0(Q2) [xqi(x,Q2) - xqi(x,Q2)]

Ai0(Q2) = ei

2 – 2 ei vi ve PZ + (ve2+ae

2)(vi2+ai

2) PZ2

Bi0(Q2) = – 2 ei ai ae PZ + 4ai ae vi ve PZ

2

The unpolarised cross-section is given by σ0 = Y+ F20 + Y- xF3

0

The polarised cross-section is given by σP = Y+ F2P + Y- xF3

P

PZ >> PZ2 (γZ interference is dominant)

ve is very small (~0.04).unpolarized xF3 ai,polarized F2 vi

The total cross-section : σ = σ0 + P σP

)(2sin

122

2

2 QM

QP

ZZ

F2P = Σi Ai

P(Q2) [xqi(x,Q2) + xqi(x,Q2)]

xF3P= Σi Bi

P(Q2) [xqi(x,Q2) - xqi(x,Q2)]

AiP(Q2) = 2 ei vi ae PZ - 2 ve ae (vi

2+ai2) PZ

2

BiP(Q2) = 2 ei ai ve PZ - 2 ai vi (ve

2+ae2) PZ

2

SM formalism ai = T3i , vi = T3i – 2ei sin2ƏW

Let ai and vi be free as well as PDF parameters. Perform fits with 4 NC EW parameters free

NC and CCe+p data now included. Improvement is NOT due to extra statistics but due to NOT trying to compromise between two close by mimina anymore. Contours show that errors are now much less asymmetric wrt central values of the fit.

au ad vu vdSM value 0.5 -0.5 0.196 -0.346

4 EW param 0.50±0.13±0.22 -0.46±0.45±0.75 0.10±0.12±0.13 -0.53±0.28±0.39

4 EW param 0.60±0.06±0.06 -0.28±0.20±0.17 0.24±0.07±0.03 -0.24±0.15±0.11

au/vu ad/vd au/ad vu/vd contours: Final HERA-II e-

We look forward to NC e+

polarised data to improve

this.

au/vu ad/vd au/ad vu/vd contours: Final HERA-II e- plus Prel CC and NCe+

NC e+ polarised data do

improve this - it is not

just statistical it is

resolving a double

minimum problem

–hopefully the

systematics will cover

the SM!

Comparison with other experiments

Summary

• We have now repeated our preliminary work of Summer 2007 using the FINAL e- CC and NC data

• CCe+ improves the d-valence• And MW• NCe+ improves the u-valence further• And the NC EW couplings• For DIS re-do with just the CCe+ latest version and don’t

include NC EW contour update (and probably not MW update either at this stage.

Can also fit BOTH GF and MW remember GF SM= 1.11639×10-5

Or we can fit a more general formalism: fit g and MW in g2 / (Q2 + MW2)2

such that g2=GF2 MW

4 = 0.07542 for standard model

GF=1.141 ± 0.012 ± 0.022 ×10-5

GF=1.175 ± 0.010 ± 0.016 ×10-5

MW=79.5 ± 1.3 ± 2.6 MW=77.2 ± 1.1 ± 2.0

g= 0.0722 ± 0.0018 ± 0.0035 g= 0.0701 ± 0.0016 ± 0.0020

MW=79.5 ± 1.3 ± 2.6 MW=77.2 ± 1.1 ± 2.0

This is the most general way to present our data

Purple is with new NC/CC e- Red adds prel.NC/CCe+