ZEUS PDF analysis A'M CooperSarkar, Oxford DIS2004

1
ZEUS fits-report April 3rd 2006 A.M.Cooper-Sarkar
and C. Gwenlan team-1 K.Nagano and S. Shimizu
team -2

• Report on fitting new data on e- CC and NC
  polarised
• MW fits, MW/ GF fits etc.
• Electroweak couplings au,ad,vu,vd and
  T3u,T3d,sin2?W plus T3ur, T3dr fits
• Improvement to PDF uncertainties

2
Electroweak fits MW/GF etc replace plot
New CC e- data with ve polarisation P
-0.268297 Lumi 78.8 pb-1 Vs fit predictions for
ZEUS-ONLY fit including these data
New CC e- data with ve polarisation P
0.328555 Lumi42.7pb-1 Vs fit predictions for
ZEUS-ONLY fit including these data
3
Now let MW be free in the propagator- remember SM
value 80.4
Value of MW Conditions
of fit
Considerable improvement in errors compared to
HERA-I data- Note normalisation error of new data
taken as 3.5 Note new data do not have
correlated systematics yet ..but experience has
shown that this makes little difference in the CC
samples
4
How does MW enter the fit? In the factor GF2
MW4/(Q2MW2)2 i.e. in normalisation as well as in
propagator First try a fit to BOTH GF and MW
remember GF SM 1.1163910-5 In this case the
difference from SM values has gone into the
estimation of GF Then fit g and MW in g2 / (Q2
MW2)2 where g2GF2 MW4 0.07542 for standard
model, but is a coupling in general -then MW
enters only into the shape
This is the most general way to present our data
5
Whereas neither of these plots is unacceptable, I
would NOT be inclined to show these contours or
possibly just the g/MW one
GF/MW
g/MW
6
Alternatively USE the standard model
relationship GF2 MW4 0.5 (p?/ (1 MW2/MZ2))2
(1/(1-?r))2 So that MW is the only parameter
entering into either shape or normalisation One
can even go further, the factors PZ Q2/(Q2
MZ2) 1/ sin2 2?W Which enter into the NC
cross-sections could also be written in terms of
sin2 2?W 4(1 MW2/MZ2)MW2/MZ2 (1-?r)
where ?r depends on mtop and mHiggs , take
mtop175GeV ?r ?a - 3a mtop2
?rem 16p (1 MW2/MZ2)2
MZ2 Where ?a is the fermionic part of the vacuum
polarisation and ?rem contains contributions that
depend on mtop and mHiggs logarithmically. H1
result using this technique on HERA-I data MW
80.8 0.21
7
Electroweak fits NC couplings replace plot
New NC e- data with ve polarisation P -0.268297
Lumi 78.8 pb-1
New NC e- data with ve polarisation P 0.328555
Lumi42.7pb-1
vs fit predictions from ZEUS-ONLY fit including
these data..but this time the fits for ve, ve
and zero polarisation are shown on both figures,
so you can appreciate what a small effect it is
nevertheless we seem to be getting somewhere with
electroweak couplings!
8

• For NC processes at LO the unpolarised structure
  functions are given by
• F20 ?i Ai0(Q2) xqi(x,Q2) xqi(x,Q2)
• xF30 ?i Bi0(Q2) xqi(x,Q2) - xqi(x,Q2)
• Aio(Q2) ei2 2 ei vi ve PZ
  (ve2ae2)(vi2ai2) PZ2
• Bio(Q2) 2 ei ai ae PZ 4ai ae vi ve
  PZ2
• PZ2 Q2/(Q2 M2Z) 1/sin22?W
• And the unpolarised cross section is ?0 Y F20
  Y- xF30
• Whereas the polarised structure functions are
  given by
  
• F2P ?i AiP(Q2) xqi(x,Q2) xqi(x,Q2)
• xF3P ?i BiP(Q2) xqi(x,Q2) - xqi(x,Q2)
• AiP(Q2) 2 ei vi ae PZ - 2aeve(vi2ai2) PZ2
• BiP(Q2) 2 ei ai ve PZ - 2ai vi (ve2ae2)
  PZ2
• And the polarised cross-section is ?P Y F2P
  Y- xF3P
• So that the total cross-section is ? ?0 P ?P
• Given that ve is very small and the interference
  ?/Z terms in PZ dominate the pure Z exchange
  terms in PZ2
• we have sensitivity to ai from xF30 unpolarised
  and sensitivity to vi from F2P polarised

9
ai T3i vi T3i -2ei sin2?W
2 and 3 parameter fits nice, 4 parameter fit
unconstrained Shimas values in blue
10
au/vu contour from 2-parameter EW PDF fit
au/vu contour from 3-parameter (au,vu,vd) EW
PDF fit
There is improvement in vu compared to H1
published 2-parameter results -as there should be
with polarisation information!
11
ad/vd contour from 2-parameter EW PDF fit
ad/vd contour from 3-parameter (au,vu,vd) EW
PDF fit (ad -au)
There is improvement in vd compared to H1
published 2-parameter results -as there should be
with polarisation information!
12
vu/vd contour from 2-parameter EW PDF fit
vu/vd contour from 3-parameter (au,vu,vd) EW
PDF fit (ad -au)
Displayed to illustrate the strong correlations
between the vs which come from polarised F2P
13
Should be able to do au ad and I cant! the job
wont give nice MNCONT contours. However the
T3u/T3d plot gives the same information, from a
job with 3 EW parameters free - T3u, T3d, sin2?W
MY opinion show the 2 parameter contour fits
dont show the 3 parameter contours BUT DO keep
the values in a TABLE
ad
au
14
Try another way of approaching the free
parameters fit T3u, T3d, sin2?W Also define
right handed couplings by ai T3i T3iR vi
T3i T3iR -2ei sin2?W And fit T3uR, T3dR
15
MY opinion put values in a Table but do NOT show
contours except possibly for T3u/T3d
T3u -T3d and sin2?W, - 2 EW PDF
parameter fit- standard model assumptions
3 EW PDF parameter fit- T3u, T3d,
sin2?W Standard Model assumptions Note this
contour T3u/T3d is essentially au/ad, clearly it
works better when these are also fed into the vu
and vd values via standard model formulae
16
3 EW PDF parameter fit- T3uR, T3dR,
T3u Standard Model assumptions
T3uR T3dR 2 EW PDF parameter fit- standard
model assumptions
MY opinion DO show a T3uR/T3dR contour BUT do
not show the others My favourite T3uR/T3dR
contour is the 4 parameter one.. See next page
17
This is my favourite T3uR/T3dR contour because If
we are deviating from standard model in allowing
right handed couplings we should probably also be
considering heavier W/Zs and so sin2?W would
also not be SM hence it should be free. One could
argue why not T3uR/T3dR and sin2?W why does T3u
also need to be free? And I agree I just havent
done it yet!
4 EW PDF parameter fit- T3uR, T3dR, T3u and
sin2?W
18
Replace plot
Improvement of PDF uncertainties due to new data,
central values change very little- full line.
NOTE xuv uncertainties improve significantly at
high-x ..as expected since NC and CC e-
cross-sections are both u dominated
19
Improvement in uncertainty remains at high Q2-
relevant to LHC!
20
PDFs from the new fit compared to ZEUS-JETS
21
Conclusions

• Nice electroweak results
• 2nd analysis also in agreement
• Lets go fro preliminary at DIS06
• Decide EXACTLY what to show