status of w mass runiib analysis - in2p3 events directory (indico) · 2011. 5. 30. · patrice...
Post on 15-Nov-2020
7 Views
Preview:
TRANSCRIPT
Patrice Lebrun
Avec signature (utilisation « logo »)Pour tous supports de com(affiches, publications, posters, web…) et pages secondes des présentations power point (pastille)
Avec développé (utilisation « bloc marque »)Pour papeterie institutionnelle(courriers, cartes de visite) et « une » des présentations power point
Patrice Lebrun(For the D0 W Mass Group)
30/05/2011D0-France Meeting, LPNHE
Status of W Mass RunIIb Analysis
1
Patrice Lebrun
Groups and People Involved Near its critical Size or maybe «top heavy»
USA Stony Brook University
Bob McCarthy - Prof - Daniel Boline - post-doc, started Feb 2010- John Hobbs - Prof , Wmass convener - Rafael Lopes de Sa - grad student, started Oct. 2009 -
University of Mississippi Alex Melnitchouck - post-doc, EWK convener -
France LPSC Grenoble
Hengne Li - post-doc (atlas, d0) started 2010, Wmass convener - Jan Stark - CNRS -
LAL Pierre Petroff - CNRS -
IPNL Lyon Patrice Lebrun - CNRS - Tibor Kurca - IR info -
No more active in W mass Group Northwestern University
Heidi Schellman - Prof- Sahal Yacoob - gone when his PhD done (runIIb data) -
2
Patrice Lebrun
W Mass measurement and Higgs Mass Prediction Radiative Correction to Mw
For equal constraint on the Higgs mass uncertainty : ΔMw ≈ 0.006 ΔMt
The limiting factor on MH prediction is ΔMw not ΔMt
Direct search of Higgs bosons and compatibility with radiative correction is a powerful test of the standard Model. The accuracy on MW is crucial.LHC will be not able to improve it for a long time.
3
July 2010 World average : ΔMt = 1.06 GeVequivalent to : ΔMw = 6.4 MeV
currently we have ΔMw = 23 MeV
Brief Article
Patrice Lebrun
May 25, 2011
1 Section 1
1.1 subsection 1My formula :
∆r(Log(MH/MW ),Mt2) = 1− πα√
2Gf
1
sin2ΘWMW2 (1)
dN
dt= λN
2
1
Constraint
Constraint
Prediction
Diff. with Tree Level
DØ RunII 1fb-1
80.401 ± 0.021(stat.) ± 0.038(syst.) GeV80.401 ± 0.043 GeVthis new result is the
World average
single most precise measurementof the W boson mass to date.
For the first time the total uncertaintyof 31 MeV from Tevatron is smaller than
that of 33 MeV from LEPII(arXiv:/0908.0766 [hep-ex])
World average is now:
P. Pétroff Blois 2011 14 05/31/11
World average is now: 80.399 ± 0.023 GeV
Ref: Tevatron ElectroWeak Working Group : Combination performed with B.L.U.E. method L. Lyons et al, Nucl. Instrum. Methods in Phys. Res. A 500, 391 (2003). A. Valassi, Nucl.Instrum. Methods in Phys. Res. A 500, 391 (2003).
D0 :1 fb-1
CDF : 200 pb-1
Patrice Lebrun
SM Consistency
4
Constraining Standard Model
Since MW , Mtop , and MHiggs are all related via radiative corrections, we can constrain MHiggs with precision measurements of MW and Mtop
Measurements of MW and Mtop overlaid with theory predictions for the Higgs boson
!
direct searches
Higgs limit from E W fits
158GeV @95%CL
direct searches
coincides with low edge of Tevatron direct searches exclusion band
With the Z-Pole study at LEP1, SM have predicted the Top Mass very successfully
Extrapolation to 15 MeV(Same central value on Mw)
Light Higgs Boson Mass preferred
Patrice Lebrun
A Possible Scenario for next few years
5
Jan Stark D0 France donuts, November 10, 2010 3
A possible scenario for the next few years
ΔMW = 15 MeVΔMtop = 0.5 GeV
Patrice Lebrun
Experimental observables : PT, Missing ET, MT MC Simulation to predict the distribution of these observables for a given mass
hypothesis ( all MC are produced by Lyon at CC). RESBOS + Photos/WGRAD for W/Z production Parameterized detector model (fast MC)
Zee used for tuning
6
Experimental observables
Arbi
trary
line
ar s
cale
p (e) most affected by p (W)!
No PT(W)!PT(W) includedDetector Effects added
MT most affected by measurement
Arbi
trary
line
ar s
cale
pT(e) most affected by pT(W)!
Ref. hep-ex/0011009
05/31/11 P. Pétroff Blois 2011 6
MT most affected by measurement of missing transverse momentum
Arbi
trary
line
ar s
cale
Need Monte Carlo simulation to predict shapes of these observables for given mass hypothesis.use ResBos [Balazs, Yuan; Phys Rev D56, 5558] + Photos /WGRAD for W/Z production and decay, plus parameterised detector model.
Recoil calibrationFinal adjustment of free parameters in the recoil model is done in situ using balancing in Z ! e e events and the standard UA2 observables.
UA2 observables:in transverse plane, use a coordinate system defined by the bisector of the two electron momenta.of the two electron momenta.
recoil: uT = uTHard + uT
soft + uTElec + uT
FSR
05/31/11 P. Pétroff Blois 2011 10
W e Event: Theory and Analysis V iew
4
spectator quarks
additional ppbar collisions
FSR photon can be part of either system or none
hard component = recoil against W
recoilT
eT pp=
E lectron system + Recoil
T emplate F it (RunI I A-based method illustration)
7
m(W) = 80.401 0.023 G e V (stat)background model
best fit
template fit
no quantitative correspondence between three curves on the left plot and filled circles on the right plot is implied, this is a qualitative illustration
Patrice Lebrun
Current Uncertainties (1fb-1) and Projections
7
Jan Stark D0 France donuts, November 10, 2010 5
Projections!"#$%&%'()&%*+,-.#/"+#"-0%/.-%1/"+23%0#/#"0#",/2%4"+,2*5"+6%7030#-1/#",07%'.81%2"1"#-5%5/#/%,8+#.82%0/192-0:;%<-#70%-=#./982/#->
?#%-+5%8'%@*+%AAB%-=9-,#%#8#/2%*+,-.#/"+#3%8+%!%1/00%8'%&C%D-E%'.81%FG%/28+-;
H=9-,#%0"1"2/.%9-.'8.1/+,-%'.81%IFJB/+5%,81("+-5%-..8.%8'%&K%D-E;
L$"0%2-6/,3%1-/0*.-1-+#%M"22%(-%"+%#$-#-=#(88N0%'8.%5-,/5-0%#8%,81-;
I8*25%(-%/+%"198.#/+#%,8+#."(*#"8+%#8%6-##"+6#$-%0#/+5/.5%185-2%"+#8%#.8*(2-%"+%#$-%+-/.%'*#*.->
M"#$%!1!%O%&P%D-EB%!1
#%O%&%Q-E%
/+5%1!%O%RS;TSS%Q-E%>
4U;%@-+#8+B%AIVHU%KSSR:
cl 95% @ GeV 117 GeV 71 m 24
19-H<= +
Patrice Lebrun
W mass measurement vs PDFs The momenta of the initial state partons is unknown. We cannot measure the
longitudinal momentum of all final states either (neutrino !) Need accurate theoretical description of longitudinal momenta of initial state (i.e. PDFs) Prefer observables that are (almost) invariant under transverse boosts (e.g. transverse
mass).
We have three approaches (that are not mutually exclusive !) to reduce the PDF uncertainty in our m(W) measurement: Reduce the uncertainty on the PDFs themselves (theory + measurements like W
charge asymmetry) using the D0 published results of D0 , the PDFs errors on MW could be reduced by 25% or even more
with the last results.
Reduce sensitivity of m(W) measurement to longitudinal momenta by cutting less harshly on electron |eta| (so far use only central electrons with |eta| < 1.05) Using all calorimeter detector , we expect to reduce by a factor 2 the PDFs errors on MW
Reduce sensitivity of m(W) measurement to longitudinal momenta by using a "JES-corrected" recoil measurement to define m_T ("put electron and recoil on the same energy scale" to improve Lorentz invariance of m_T). Study is going on : a significant improvement is expected too.
8
Patrice Lebrun
Ongoing work: inclusion in PDF sets
9
Jan Stark Electronweak precision measurements at the LHC, CERN, April 4-5, 2011 22
Ongoing work: inclusion in PDF sets
CTEQ6.1 and CTEQ6.6 are in
excellent agreement with the
CDF measurement of
the W boson charge asymmetry ...
... but when you plot
the CDF data in terms
of lepton charge
asymmetry then they
do not really disagree
with the D0 data !
... they do not agree with the D0 lepton charge asymmetries
(neither with the electrons nor with the muons) ...
Patrice Lebrun
W charge asymmetry
10
Jan Stark Electronweak precision measurements at the LHC, CERN, April 4-5, 2011 25
W charge asymmetryPDFs CT10W : D0 data are preferred
Patrice Lebrun 11
More details tomorrow : Hengne’s talk
Current Analysis Issues
Hengne Li / LPCSConveners’ Meeting, May 27, 2011
W Mass Status
12
MWT P e
T
�ET
FullMC W -> e nu 62M events generated, 9.8M events after selection
Looks good for all the 3 observables: MT, ElecPT and Missing ET. And the fits of W Mass and Width close.
25 30 35 40 45 50 55 600
50
100
150
200
250
300
310!WCandMet_Spatial_Match_0
/ndf = 132.0/702"
FULL MC
FAST MC
WCandMet_Spatial_Match_0
25 30 35 40 45 50 55 60-4
-3
-2
-1
0
1
2
3
4
= 132.0 for 70 bins2" distribution with overall " = 132.0 for 70 bins2" distribution with overall "
0 5 10 15 20 25 300
100
200
300
400
500
600
310!WCandRecoilPt_Spatial_Match_0
/ndf = 744.9/302"
FULL MC
FAST MC
WCandRecoilPt_Spatial_Match_0
0 5 10 15 20 25 30-10
-5
0
5
10
= 744.9 for 30 bins2" distribution with overall " = 744.9 for 30 bins2" distribution with overall "
However, there is a small issue related to the Recoil Fine Tuning. This affects the MissingET a little.
Recoil PT
Two issues:- Modeling of PT(ee) in Zee (next slide).- Choice of parameterization of recoil fine
tuning need to be revisited. - Another approach is on going for the
Recoil Fine Tuning using the “Recoil Energy Flow” which has proven to be useful in our RunIIa analysis.
Hengne Li / LPCSConveners’ Meeting, May 27, 2011
Pt(ee) (GeV)0 5 10 15 20 25 30
Prj.
!Pt
(ee)
0
2
4
6
8
10
12
14
16
pZptEtaPrjVsZPtBins
FullMC
FastMC
pZptEtaPrjVsZPtBins
Pt(ee) (GeV)0 5 10 15 20 25 30
(GeV
)"
-0.2
-0.15
-0.1
-0.05
0
Diff Hists
FastMC-FullMC
Diff Hists
W Mass Status
13
FullMC Z -> eeEta-Imbalance is the reference for Recoil Fine Tuning. ηimb = (�pZT − �uT ) · η
η-axis: the bisector in R-φ plain of two electrons from Z decay
Mean of Pt(ee) projection to Eta-Direction vs. Pt(ee)
FastMC-FullMC for each bin
The Issue:The Pt(ee) projection to Eta-Direction has a mis-match when Pt(ee) is large (small fraction of events). See plots on the right.
Consequences:It will impact our Recoil fine tuning. If the Pt(ee) is wrong, this mistake will be transferred to the Recoil.
Investigations have shown, it is most likely caused by imperfections in our current description of the Phi Cracks. Remember, the two electrons from the Z decay are angularly highly correlated.
We observed that Data Zee has the same signature. It requires more follow-up.
3.2.3 Recoil System Reconstruction
The recoil !uT for W boson events is calculated from the !!ET and the electron(s) transversemomentum:
!uT = − !!ET −∑
e
!peT (4)
3.2.4 Useful Kinematic Variables
In Z → ee decays the di-electron momentum is given by !p(ee) = !p(e1) + !p(e2) and thedi-electron invariant mass is: m(ee) =
√
2E(e1)E(e2)(1 − cos ω) where ω is the openingangle between the two electrons.
When tuning the simulation and making comparisons with data, it is useful to definea coordinate system, first introduced by UA2 [2], in the plane transverse to the beamsthat depends only on the electron directions. Let us call the axis along the inner bisectorof the two electrons the η axis and the axis perpendicular to that the ξ axis. Fig. 1 (left)illustrates these definitions.
! et1
et2
^
^
Ptee
Ptee
Ptee
..
.. "
!
^
^
^
Ptrec
Ptrec
Ptrec
"
!"
p!T(e)
u!Tp!T(")
u||
u#
Figure 1: Left: definition of η and ξ axis for Z → ee events. Right: definition of u‖ andu⊥. u‖ can be positive or negative. As can be seen in this figure, u‖ is negative whenopposite to the electron direction. Note that !prec
T is the recoil measured by the calorimeter,and in general differs from !pee
T because the calorimeter resolution and response differs forhadrons (recoil) and electrons.
For W → eν decays useful quantities are the projection of the transverse recoil mo-mentum on the electron direction:
u‖ = !uT · pT (e) (5)
and the projection on the direction perpendicular to the electron:
u⊥ = !uT · [pT (e) × z] (6)
Fig. 1 (right) illustrates these definitions.
9
pour fitter le recoil
Hengne Li / LPCSConveners’ Meeting, May 27, 2011
25 30 35 40 45 50 55 600
10000
20000
30000
40000
50000
60000
WCandElecPt_Spatial_Match_0
/ndf = 261.1/702!DATAFAST MC
"#W->Z->eeQCD
WCandElecPt_Spatial_Match_0
25 30 35 40 45 50 55 60
-10
-8
-6
-4
-2
0
2
4
= 261.1 for 70 bins2! distribution with overall ! = 261.1 for 70 bins2! distribution with overall !
50 60 70 80 90 1000
5000
10000
15000
20000
25000
30000
WCandMt_Spatial_Match_0
/ndf = 95.8/1002!DATAFAST MC
"#W->Z->eeQCD
WCandMt_Spatial_Match_0
50 60 70 80 90 100-4
-3
-2
-1
0
1
2
3
= 95.8 for 100 bins2! distribution with overall ! = 95.8 for 100 bins2! distribution with overall !
W Mass Status
14
MWT P e
T
Data W -> e nu RunIIb 4.3 fb-1 1.7M events after selection
MT looks good, but ElecPT doesn’t.- We know the degradation of the
ElecPT Jacobian peak is due to the boost of the W boson. (see plot on the bottom)
- At the generator level, we tried to re-weight Resbos using Phi* measured from D0 Data (Vesterinen et., al.,), we found the impact is negligible.
- But, we do have a certain mis-modeling of the Recoil. This is reflected in the ElecPT distribution because of the cut at RecoilPT<15GeV
Jan Stark et al. Dzero Physics Workshop, Febrary 9-10, 2009 7
Experimental observables
! MT most affected by missing transverse momentum measurement.
Arb
itrary
linear
scale
Arb
itrary
linear
scale
• pT(e) most affected by pT(W).
Ref. hep-ex/0011009
No PT(W) PT(W) included Detector Effects added
Patrice Lebrun
Conclusion Results at 1 fb-1 : we are the «best» in the world :
Analysis on 6 fb-1 : is new compared to previous one due to the highest instantaneous luminosities Now the current work is on some details (see Hengne’s talk)
At the end the analysis with the last 4 fb-1 will be not too much different.
15
DØ RunII 1fb-1
80.401 ± 0.021(stat.) ± 0.038(syst.) GeV80.401 ± 0.043 GeVthis new result is the
World average
single most precise measurementof the W boson mass to date.
For the first time the total uncertaintyof 31 MeV from Tevatron is smaller than
that of 33 MeV from LEPII(arXiv:/0908.0766 [hep-ex])
World average is now:
P. Pétroff Blois 2011 14 05/31/11
World average is now: 80.399 ± 0.023 GeV
Ref: Tevatron ElectroWeak Working Group : Combination performed with B.L.U.E. method L. Lyons et al, Nucl. Instrum. Methods in Phys. Res. A 500, 391 (2003). A. Valassi, Nucl.Instrum. Methods in Phys. Res. A 500, 391 (2003).
Our goal : reach 15 MeV resolution on MW to be able to affirm SM is over.
top related