Systematic studies on mw

WW meeting 28/11/01 D. BOUMEDIENE 1

Systematic studies on mwD. BOUMEDIENE – P. PEREZ – E. LANCON

CEA/ Saclay

O. BUCHMÜLLER

1- Reminder (cone, P cut, PF cut)

2- P Flow cut analysis

3- Cone analysis

4- No typ6 objects analysis

5- Charged only analysis

6- Mjet=0 analysis

7- Conclusion


ReminderReminder Leukerbad:

• Pcut analysis (cutting on all Eflow objects) is unstable for all channels

• The instability is due to some neutral objects (ECAL residuals : typ6 Eflow objects )

• the analysis is stable either :

• if typ6 objects are not considered at all,

• or if no cut is applied on typ6 objects

• However the two analyses lead to incompatible mw measurements

Pisa / W meeting 06-07-01:

• Particle Flow based cut analysis is not stable in the 4q channel (it motivated the test of Pcut on 4q and lqq)


AimAim

mw measurement stability is tested for the following algorithms

• Pcut (see Leukerbad meeting)

• PF cut (4q+lqq)

• Cone (4q+lqq)

Absolute ALEPH mw measurement scale is tested with

• Charged only (4q+lqq)

• Typ6 suppression (updated)

• Mjet=0 (4q+lqq) Mw is measured 23 times (4 channels)

New today


TechniquesTechniques 4q channels: 3D fit

lqq channels:

• The lepton is not used (because not concerned by the effect): 2D fit

• All the channels are used (except ll)

Parametrisations/jet corrections are systematically recomputed

Correlations are evaluated at 2 energies (4q+lqq merged)

shifts on mw derived from observed errors.


Pflow Cut AnalysisPflow Cut Analysis PF in 4q channel is given by Thomas's code

An equivalent distribution is defined in the 3 lqq channels

P Flow cut


Pflow cut analysis

Pflow cut analysis

tendency: slope is 2.0 far from 0 (jetset)

discrepancy wrt flat: 1.6


Pflow cut analysis: no cut on typ6 obj.

Pflow cut analysis: no cut on typ6 obj.

tendency: 1.2 from slope=0


tendency: 2.0 1.2

discrepancy wrt flat:

1.6


PFcut SummaryPFcut Summary• PF cut analysis (cutting on all objects) is unstable (2.0 )

• It is difficult to use it to measure CR

• The size of the effect is compatible with what was observed with the P cut.

-82 51 MeV/c2 (excluding [0.1,0.9], lqq+4q)

• The physical source of the problem in Pcut/Pfcut analyses comes from the same source: cutting on typ6 objects

• PF cut analysis without cutting on typ6 objects is stable


• Hugo’s cone Algorithm is used for the reconstruction (http://alephwww.cern.ch/~ruizh/aleph-only/cones/cones.html)

• jet parametrisations are recomputed

• Jet energy corrections are done ( Patrice talk)

• All energies, all channels (except ll) are used

• Correlations between analyses are computed with

2 x 200k evt (@ 2 energies)

The difference 4q / lqq and energy dependence are neglected

the cone analysisthe cone analysis

Rcone 0.6 0.75 1.0 1.25 2.0

(%) 73 78 83 90 95


Parametrisation (Rc=0.6)


b: angular resolutio

n




b: angular resolutio

n


mw versus RCmw versus RC

tendency: 2.6 far from expectations (jetset)


Durham analysis (RC=3.0) taken as

reference

better probability for the slope


4q/lqq contributions4q/lqq contributions

Probability of a linear behavior (with a slope) increases when combining 4q and lqq channels


mw versus s (Rc=0.75)

mw versus s (Rc=0.75)


Cone without typ6 objects

Cone without typ6 objects

tendency: FLAT (<1

discrepancy wrt flat: (<1

‘Cutting’ on typ6 objects caused the

instability

RC=2.0 taken as reference


Cone SummaryCone Summary• Cone analysis (with all measured energy) is unstable (2.6 ) as expected

• It cannot be used to measure CR

• The problem is not due to 1 energy only

• The effect is similar to what was observed with the P cut and PF cut analyses :

-85 31 MeV/c2 (Rc=0.6, lqq+4q)

• The physical source of the problem is the same

• Cone analysis without typ6 is stable


Standard Durham analysis without Typ6 objects

Standard Durham analysis without Typ6 objects

• All type6 (~5% energy) objects are removed from the jets.

• Statistical sensitivity loss is < 1% in both lqq and 4q channels

• Correlation with standard analysis =92%


Effect of Typ6 objects suppression

Effect of Typ6 objects suppression


Typ6 Suppression Summary

Typ6 Suppression Summary• mw measured without typ6 objects is

shifted wrt standard analysis by:

-112 20 MeV/c2

• The effect is compatible to what was observed with the P cut / PF cut / cone analyses


Charged only analysis

Charged only analysis

• we can test an analysis using only charged tracks. Such an analysis is reliable (all distributions for charged are in agreement between DATA/MC)

• but we expect a big loss of sensitivity.

Tested analysis: Charged objects only are used

correlation with std analysis: = 31 %

Sensitivity loss: ~ %


Charged only analysisCharged only analysis


Charged Only SummaryCharged Only Summary• mw measured with charged objects only is shifted wrt standard analysis by:

-120 65 MeV/c2

• The effect is compatible to what was observed with the P cut / PF cut / cone / pre typ 6 analyses


• Jet masses are biased by a bad simulation of multiplicities in particular for typ6 objects.

• It is known that the bias propagates to mw

• A measurement with a fixed Mjet should be insensitive to a bias

Tested analysis:

We fix Ejet = Pjet

( same jet mass in MC and DATA)

Nothing is removed, we just forget the jet mass

correlation with std analysis: = 74 %

Sensitivity loss: qlqq %

mjet=0 analysis

mjet=0 analysis


mjet = 0 analysis

mjet = 0 analysis


mjet=0 Summary

mjet=0 Summary

• mw measured with mjet=0 is shifted wrt standard analysis (2D for lqq) by:

-79 31 MeV/c2

• The effect is compatible with what was observed with

P cut / PF cut / cone / pre typ 6 / charged only analyses


CONE (Rc=0.6) - 85 31 (2.7

Pflow - 82 51 (1.6

(exclud. [0.1,0.9])

Pcut (cut 2GeV) -125 40 (3.1) 3.2

Pre TYP6 -112 20 (5.6

Charged only -120 65 (1.8

Mjet=0 - 79 31 (2.5

SummarySummaryAnalysis mw(MeV/c2) slope slope no typ6

we have compelling evidences that the aleph w mass is overestimated by 50-100 MeV/c2


ConclusionConclusion

We have compelling evidences that the aleph w mass is overestimated by 50-100 MeV/c2

We have developed low correlated analyses which are stable in the lqq channels (Minimum requirement for 4q Color

Reconnection sensitive analysis)

Those analyses will now be used to study Color Reconnection effect in the 4q channel

mw and width analysis

The code and ntuples are publicly available on

Castor for all energies and all MC

Systematic studies on mw

Documents

boumediene cone

q lnqq cone

boumediene pflow

boumediene dmw

analysis pf

boumediene techniques

typ6 objects pf

analysis tendency