W Charge asymmetry

W Charge asymmetry

Cigdem Issever, Adam Scott, and David Stuart

UCSB

Preblessing, EWK 5/13/04

•CDF Note 6282, Measurement of W e Charge Asymmetry

•CDF Note 6108, Alignment of the Plug Calorimeter

•CDF Note 6412, QCD Background for Plug Ws

Motivation

W± charge asymmetry arises from u(x)/d(x) in proton.

Gives most sensitive u/d constraint at our x and Q2

But still not well constrained at high x (high . We don’t already know the answer.

Experimentally appealing.

Can improve many of our forthcoming measurements.

Damn the ’s. forward e ahead

Asymmetry in W production complicated by unknown pz

So, use lepton asymmetry

Which convolves production asymmetry with V-A decay.

Damn the ’s, forward e ahead

Forward electrons with charge id are the key.

Two points and a curvature

define a unique helix:• Primary vertex• Shower Max cluster

• ET |C|

Attach hits using OI tracking

Select track with better 2/dof

Phoenix Tracking

Phoenix Tracking

Plug Alignment (CDF6108)

Align plug (actually PES) with COT tracks:

allow offsets in x,y,z and a rotation in phi

Plug Alignment

Measure and correct, iteratively from largest to smallest

, x & y, Z

Internal Alignment

Residual misalignments attributed to internal octant offsets

Fit for and radial shifts in each octant.

Consistent internals between bpel08 and bpel09

Residuals look good after alignment

We performed various validity and systematic checks

and found it to be robust with small systematics.

Charge ID

Incorrectly identifying the charge dilutes the asymmetry.

We can correct for it if we know the rate, fQ # wrong / total

Atrue = Ameas / (1-2fQ)

So make fQ small and measure it.

MC predicts fQ ≈ 1%.

But MC doesn’t know about residual mis-alignments etc.

We must measure it in data and cannot assume any dependence.

Measured fQ() with Z e+e-

Z Event Selection

Baseline electron selection with COT or PHX track. ET>25 GeV

PHX track is slightly non-standard to optimize fQ.

# hits >= 4

2 < 8

2 > 0.5

Seed Pull < 0.4

80<mee<100 and compare central Qtag

Charge ID Measurement

Errors calculated with Bayesian prescription, CDF5894.

fQCOT fQ

Phx

CDFII Preliminary ∫Ldt=170/pb

W Event Selection

Baseline electron selection w/ PHX track.

ET>25 GeV

Missing ET > 25 GeV

50 < MT < 100 GeV/c2

No other EMO with ET>25 GeV

to suppress QCD and DY

Check kinematic distributions

Data

MC

CDFII Preliminary ∫Ldt=170/pb

Check kinematic distributions

Data

MC

CDFII Preliminary ∫Ldt=170/pb

Holy Phi

Data

MC

Raw Asymmetry

CDFII Preliminary ∫Ldt=170/pb

Curve is just a fit to guide the eye.

Raw Asymmetry

CDFII Preliminary ∫Ldt=170/pb

Curve is just a fit to guide the eye.

BackgroundWe correct the asymmetry for backgrounds from:

• W e Asymmetric, measured from MC

• Z e+e-

Asymmetric, measured from MC

• QCD

Symmetric, measured from data CDF 6412.

Use Iso vs MET.

QCD Background systematic

Check assumption of non-correlation

small systematic, as long as we don’t use the

PEM 3x3 2 or 5x9 cuts (they are correlated with Iso)

upper limit by about a factor of 2.

QCD BackgroundQ

CD

fra

ctio

n

We use 0.5 ± 0.25. Becomes significant for >1.8

CDFII Preliminary ∫Ldt=170/pb

Corrected Asymmetry

CDFII Preliminary ∫Ldt=170/pb

Compare to COT when it is available

CDFII Preliminary ∫Ldt=170/pb

Compare PDFs to data

CDFII Preliminary ∫Ldt=170/pb

CTEQ5L

Overlay CTEQ3,4,5L, GRV98LO, MRST+d/uCTEQ3,4 to 5 have less d/u from Run1 asym. We keep pulling in Run2.

CDFII Preliminary ∫Ldt=170/pb

Compare PDFs to data

CP checkCDFII Preliminary ∫Ldt=170/pb

A(||)CDFII Preliminary ∫Ldt=170/pb

A(||)

Compared to Run1 e+

FMU

CDFII Preliminary ∫Ldt=170/pb

Getting more sensitivity to AW

Getting more sensitivity to AW

25 < ET < 35

35 < ET < 45

A() with 25 < ET < 35 GeV

CDFII Preliminary ∫Ldt=170/pb

A(||) with 25 < ET < 35 GeV

CDFII Preliminary ∫Ldt=170/pb

A() with 35 < ET < 45 GeVCDFII Preliminary ∫Ldt=170/pb

A(||) with 35 < ET < 45 GeVCDFII Preliminary ∫Ldt=170/pb

Conclusion

W asymmetry provides new PDF constraints.