1. on top mass measurement based on B had decay length
Progress
on top mass measurement
based on B had decay length
ATLAS UK Top Physics Group
Birmingham
Thursday 25 May 2006
D. Typaldos

2. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
Outline
Brief description of the proposed method
General issues of the analysis
Analysis on systematic effects
Results
Summary - Plans
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

3. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
Reminder of the method
Method proposed by C.S Hill, J. R. Incandela and J. M.Lamb last year
(University of California Santa Barbara - hep-ex/ )
The drawback of the direct measurement of the top mass is its dependence upon the b jet energy measurement and affected so by the jet energy scale uncertainty
This method suggests an alternative approach on determining the top mass by focusing on the charged track reconstruction than dealing with calorimeter information
The decay length can be determined independently by using the average distance of the displaced vertices of the b hadrons (transverse component)
Mass estimator ~> correlation of the B hadron decay length to the top mass
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

4. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
Dilepton channel
The logical choice to set as ‘clear’ and as possible
and most controllable the environment for B hadron
decay vertex reconstruction
missing top’s hadronic decays :
reduced background and combinatorics
Disadvantage: BR of ~5%
(400K ttbar events per L= 10 fb-1)
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

5. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
Datasets
Private production of ~ 20 K ttbar (to e/mu):
PYTHIA ATHENA rel
Fast and full detector simulated data produced and used
Keeping the general prescription of Rome PYTHIA settings
top quark decay constrained to leptons only
(tau excluded, though not expected to contribute
significantly)
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

6. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
Preselection
Generally the initial cut are left as soft as possible to avoid biasing effects
(what and how will be restricted to be decided progressively)
At least 2 leptons within the acceptance range of Pt > 25 GeV && |Eta| < 2.5
(**)Discriminated events with 2 leptons having invariant mass close to Zmass (m=2 GeV)
(**)Accepted events with the 2 highest Pt leptons having opposite charge
Jets in within the acceptance range of Pt> 15 GeV && |Eta|<1.8
Application on various tagging algorithms (standard – not optimised cuts on weights)
-- For the b-signal study : b jets match a Bhadron in R-cone of radius=0.3
(**) Cuts applied on bkg samples (WW, WZ, ZZ)
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

7. Svbu tagger-bjet efficiency
Good agreement between
reconstructed secondary vertices and B had decay length (truth)
Reco sec vtx
Bhad decay
length (truth)
Bjet efficiency
weight
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

8. Systematics contribution
Possible systematic uncertainties :
B fragmentation
ISR/FSR
B hadrons lifetime
PDFs
Mixing of various B hadron flavours
Dependence on preselection cuts (sensitivity)
Jet algorithms
Vertex efficiency
Additional issues :
Background evaluation wrt signal
Optimisation to improve sensitivity
Further correlations ? (lepton pT)
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

9. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
B-fragmentation
contribution of di-boson pairs
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

10. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
b-fragmentation
B Fragmentation effect on decay length for fast simulated data (PYTHIA - 120K)
Standard top mass value (± 5 GeV/c2)
Variation of Peterson parameter around the default value (= )
Various fitting functions considered
The most promising one: exponential + 1/x
The preliminary results suggest further study
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

11. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
Fit on decay length (1)
Choosen fitting function :
Helps in isolating vertex mis-reconstruction effects near PV
from the  fitted parameter
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

12. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
Fit on decay length (2)
    
170
175
180
GeV
3.875
3.971
4.082
3.790
3.922
4.025
3.757
3.892
3.992

ATLAS UK Top Physics Group-Birmingham - 25 May 2006

13. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
Fitted parameters  B  A   
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

14. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
Mass uncertainty 
preliminary
Mass uncertainty (GeV)
Central region
uncertainty ~2GeV
Top Mass (GeV)
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

15. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
Di-boson samples
Preliminary study on background contribution using the next rome datasets :
WW 30K
ZW(+) 27K
ZW(-) 6K
ZZ [rome recov10.McAtNlo_ZZ_Jim_leplep] @ 20K
Unfortunately much lower statistics than expected
(one boson decays to everything in all cases)
WW->leplep = ± 0.17 pb
ZZ->leplep = 0.54 ± 0.01 pb
WZ+ -> leplep = 7.08 ± 0.09 pb
WZ- ->leplep = 4.38 ± 0.04 pb
tt->leplep= 22 pb
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

16. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
Results WW
28975
6273
2028
1980
1665
2022 49 67 72 3 65 21
1241/2 ZZ
19220
1157
858
453
293
520 12 19 16 8
56/2
ZW-
6231
1302
790
552
326
526 15 21 23 18 11
605/2
ZW+
26777
5937
3393
2290
1369
2107 59 63 61 51 25
607/2
Total Event Nb
Nb Events with W/Z->leptons
Nb Events with high Pt lept
Nb Events (Zmass)
Nb Events (l+ l-)
Nb Jets passing Pt/Eta cut
TAGGING
[Likelihood > 0.9 ]
[ Weight > 3 ]
IP2D
IP3D
SVBU
CMB
LF2D
Nb Events expected at 10 fb
Nb ttbar Events expected at 10 fb
200 K events
ATLAS UK Top Physics Group-Birmingham - 25 May 2006

17. ATLAS UK Top Physics Group-Birmingham - 25 May 2006
Summary-Plans
General aspects of the analysis
First preliminary analysis on 2 systematic effects
- b fragmentation (Fast Simulation data)
mass uncertainty ~ 2 GeV (default mtop)
- di-boson contribution (Rome datasets)
most significant encounted for WW (~ 10 fb-1)
limited statistics
Plans
Long list ofbkg potential bkg - focus on:
- effect of preselection cuts
- Drell-Yan
- B lifetime
- ISR
Extending to new datasets
New Athena release ( – )
New data (TopRex, McAtNlo)
New csc data available soon
Thank you for your attention!
ATLAS UK Top Physics Group-Birmingham - 25 May 2006