-
Mixing and Lifetime from Partially Reconstructed B0 D* l ν
1)Analysis Strategy; 2)Tagging Lepton Sample description: Prompt
Leptons; Cascade Decay Side; Cascade Tag Side 3)Preliminary BLIND
Results; Comparison w.r.t. Similar Analyses 4)Study of the Analysis
Bias 5)Conclusions
M. Margoni, Time Dependent Analysis Forum, 5/5/20041
-
Analysis Strategy
Signal Vertex:
Tag Vertex:
lπ* + beam spot (x,y)
l (Elbatag) + beam spot (x,y)
l+Signal SideTag Side
π∗−
l−
2
Fit to ∆t to determine simultaneously τ, ∆m and dilution D,
constrained to the fraction of mixed events:
Nmix
/Ntot
= χd D + (1−D)/2; χ
d =x2/(1+x2)2; x=∆m τ
Tag Lepton alone:No D0 bias in tag Vertex
-
The Signal PDF
Tagging Lepton Sample:
b−> lb−>c−>lD0 −> l
{ From tag BFrom decay B
Cascade Lepton Tag Side
Cascade Lepton Decay Side
Tag B Decay B
π∗Prompt Lepton
Decay Lepton
3
D0
-
Prompt Leptons"MC: D
P~ 0.98
Floated in the fit;"DT: D
P~0.96
"bN,
bW
compatible with zero
τΒ
= 1.539± 0.003 ps
∆m = 0.465± 0.001 ps−1
FU/MP= e−|∆t|/τ (1 ±D
P cos(∆m∆t))
4
Unmixed
Mixed
-
Cascade Decay Side
l+Signal SideTag Side
π∗−
l−
D0−>l−
Time independent dilution D0 leptons mostly populate unmixed
sample; No information on mixing
τDeffective
free parameter
in the PDF
∆z meas
-
Cascade Decay Side Cascade Fraction α(θ) for B0 and peaking B+
determined in three steps as a function of the angle (tag
lepton−π∗) in the Υ frame:1) Float the relative amount of
cascade/prompt for the B0 sample with fixed B+;2) Rescale the B+
cascade/prompt yields according to the corrections obtained for the
B0;3) Fix the B+ and float again the B0
0
500
1000
1500
2000
2500
3000
3500
4000
4500
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
:3/12/23 17.37:3/12/23 17.37:3/12/23 17.37:3/12/23 17.37:3/12/23
17.37:3/12/23 17.37
B0 Btag→leptB0 D0→leptB+ Btag→leptB+
D0→leptcombinatoriccontinuum
0
0.2
0.4
0.6
0.8
1
1.2
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
60.45 / 49A0 0.9989
alpha B0
0
0.1
0.2
0.3
0.4
0.5
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
6.715 / 16A0 -0.3526E-01 0.2889E-02A1 0.3761 0.1702E-01A2 -1.264
0.3708E-01A3 1.417 0.2848E-01
:4/01/26 16.39
α(cos(θ))
" ρ = Fraction of mixed /(mixed+unmixed) Cascade ~ 5%
6cos(θ) cos(θ)
-
Cascade Decay Side
"τD" = 0.35± 0.01 ps
bN=b
W = −0.25± 0.01 ps
FU/MD
=e −|∆t|/τD
7
-
Cascade Tag Side
∆z = Zreco
− Ztag
Signal Lepton
π*
∆z true
∆z meas.
If ∆z > 0: ∆zmeas
< ∆ztrue
Signal LeptonTag lepton
Tag lepton
∆z true
∆z meas.π*
If ∆z < 0: |∆zmeas
| > |∆ztrue
|
Btag
c Btag
c
8
"bN, b
W
-
FU/MC= e−|∆t|/τ (1 −+D
C cos(∆m∆t))
"DC : due to wrong flavor charm from upper vertex + hadron
misidentification
"PDG gives:
"MC truth : DC= 0.536
"Correction added o(%) due to fake leptons (as for the prompt
lepton component)
Cascade Tag Side
Fixed in the fit on Data
9
-
Cascade Tag Side
τΒ
= 1.555± 0.045 ps
∆m = 0.431± 0.022 ps−1
DC= 0.522± 0.016
bN = −0.35± 0.15 psbW=−3 ± 4 ps
10
Unmixed
Mixed
-
B0 Signal: Prompt+Cascade (Tag+Decay)
τ=1.548± 0.003 ps; ∆m=0.466± 0.001 ps−1 ; ρ=−4%11
-
Preliminary ResultsAdding in the PDF:"Signal"B+: pure
lifetime;"BB Bkg: empirical function (oscillating+lifetime+cascade
tag+decay tag)"Continuum: pure lifetime;"Sample Composition from
Mν2 fit; "Resolution: 2 Gaussians with floating offsets and
pulls+outliers
From ~65 K evts: τ=xxx ±0.010± 0.016 ps; ∆m=xxx± 0.005 ±0.008
ps−1
0
1000
2000
3000
4000
5000
6000
7000
8000
-10 -8 -6 -4 -2 0 2
:3/12/08 20.25:3/12/08 20.25:3/12/08 20.25:3/12/08 20.25:3/12/08
20.25:3/12/08 20.25:3/12/08 20.25
Mν2 (GeV2/c4)
entri
es /
0.1
data
B0
other
B+
combinatoric
continuum
12
-
Comparison with other similar analyses"Old" τB0 analysis (Run1,
no tagging), use now only lepton−tagged events:
−Determination of fr(B+) from the global fit (B+ events do not
mix); −Improved signal/Bkg ratio; −Tag−lepton alone in the Tag
Vertex: no D0 bias; Dilepton analysis:
−Lower Efficiency... but: −Separation between Decay Side (no
dilution) and Tag Side; −Smaller Bkg Fraction; −∆t distribution of
Combinatorial using also side−bands in the fit; Simultaneous meas.
of τB0 and ∆m
Exclusive D* l ν: − Bigger amount of combinatorial Bkg; − Higher
Efficiency: 65K evts in 81 fb−1 vs 14K evts in ~20 fb−1 ; − Not
sensitive to K or NN tags; ∆m τB0
13
-
Study of the Analysis Bias"Old" Problem:From the fit to the MC
Pure Signal we observe a bias (BAD 287, v11):
δ∆m=−0.006±0.001 ps−1
The mixed event fraction is underestimated: δχ
d = −0.003±0.001
But......The fraction of MC truly mixed events when just a
single π*l pair/event is reconstructed is correct!
χd=0.1744 ±0.0005 (w.r.t. 0.174 MC truth)
Bias induced by the events with more than one π∗l pair...
14
-
L−
Decay B
Tag B
...Why?
π*+π*+
In the case of mixed events with two D* from different Bs, a
second π* l "true" pair can be reconstructed with the Right Charge
Correlation.If the 2nd pair is chosen by the selection algorithm,
the event can fall in the Side Band region and it is classified as
"Combinatorial Background"
Reduction of the measured mixed event fraction.
1th Pair
2nd Pair
Tag Lepton
15
-
Mult
04/05/04 19.10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0.5 1 1.5 2 2.5 3 3.5 4 4.5
Mult
04/05/04 19.06
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.5 1 1.5 2 2.5 3 3.5 4 4.5
"Number of π*l candidates / event (Signal MC) Right Charge
Correlation R.C + Wrong Charge Correlation
R. C. Mixed event sample shows higher fraction of multiple
candidatesFraction of Mixed Events:
Strong χd dependence vs number of reconstructed candidates for
R.C. Correl.
Mult
04/05/04 19.12
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.5 1 1.5 2 2.5 3 3.5 4 4.5
Mult
3.459 / 3A0 0.1747
04/05/04 19.08
0.15
0.155
0.16
0.165
0.17
0.175
0.18
0.185
0.19
0.195
0.2
0.5 1 1.5 2 2.5 3 3.5 4 4.5 16
Mixed All
-
Event Tag
23.14 / 1A0 0.1675
04/05/04 19.12
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0.5 1 1.5 2 2.5 3 3.5 4 4.5 Event Tag
2.683 / 3A0 0.1747
04/05/04 19.09
0.15
0.155
0.16
0.165
0.17
0.175
0.18
0.185
0.19
0.195
0.2
0.5 1 1.5 2 2.5 3 3.5 4 4.5
Mixed Event Fraction χd vs Event Tag
Event Tag:1: just one π* l candidate;2: one additional π∗ l
candidate (π∗ not from D*);3: one additional π*l candidate (π* from
D*);4: two or more π* l candidates (at least one from D*)Right
Charge Correlation R.C + Wrong Charge Correlation
17
-
How to manage this effect on the data?Three possible
strategies:
1) Use only the event sample with just one π*l candidate, (
ε~80% for R.C+W.C);
2) Determine the fraction of events with more then one D* in
Data and MC, tune the simulation and compute the expected bias;
3) Use two separate analysis streams for the two subsamples:
−single candidate; −two candidates from D* from different Bs :
"golden events" with two B D* l ν and lowest dilution
18
... Approach to be chosen...
-
Work going on...
04/05/04 19.54
0
20
40
60
80
100
120
140
160
2 2.01 2.02 2.0302468
101214161820
2 2.01 2.02 2.03
05
10152025303540
2 2.01 2.02 2.030
10
20
30
40
50
60
2 2.01 2.02 2.03
2) Rescale the MC to the DATA result3) Compute the expected
bias.
R=N(D* (Kπ)π∗) Side Band, Wrong Charge
/N(D* (Kπ)π∗) Mass Band, Right Charge
independent from efficiency/ mixing effects
D* (Kπ) π∗
Mass BandRight Charge
Mass BandWrong Charge
Side BandRight Charge
Side BandWrong Charge
Strategy n. 2: Determination of the Fraction of Events with more
then one D* (Data vs MC)
1) Compute the ratio
19Μ
-
Conclusions
"Systematic errors still preliminary"Statistical errors to be
validated (Toy) .... but a very promising measurement...
20