MEASUREMENT OF CP OBSERVABLES IN B-D0K- DECAYS AT CDFKaren GibsonUniversity of PittsburghICHEP 200808/01/08
2
CP ASYMMETRIES IN B-D0K- DECAYS CP asymmetries in
B-D0K- decays can be used to gain information about CKM angle Use method devised by
Gronau-London-Wyler (GLW) to construct CP asymmetries based on decay rates
= arg((-VudV*ub)/(VcdV*cb))Note: Charge conjugates of B- and D0 decays are implied unless both are specifically discussed!
3
CURRENT STATUS OF As of FPCP 2007, B-
D0K- asymmetries have been measured by Belle and BaBaR Measurement is still
quite statistically limited
All additional measurements help significantly CDF’s large sample of B-
decays gives opportunity to contribute significantly!
4
GLW METHOD
Measure CP asymmetry between B+ DCP0K+ and B- DCP
0K-
Ambiguities in remain Measure sin two possible solutions
Relative phase between f.s. final states sinsin)()(2
)()(00
2020
----
--
-
KDBAKDBA
KDBAKDBA CPCP
5
B-D0K- DECAYS
Consider flavor-specific and CP-even D0 decays CP-even: D0K+K-, D0 p+p-
Flavor-specific: D0K-p Don’t consider CP-odd decays
D0Ks0p0, D0Ks
0r0, D0Ks0w, D0Ks
0j Can’t easily reconstruct p0, r0, w & statistics too low in Ks
0j
6
CP OBSERVABLES IN GLW METHOD Start with experimentally accessible
observables:
Construct CP observables:
,)()()()(
00
00
--
--
pp DBBRDBBRKDBBRKDBBRR
)()()()(
00
00
-
-
-
-
pp CPCP
CPCP
DBBRDBBRKDBBRKDBBRR
-
-
-
-
-
CP
CPCP
CPCPCP
RrKDBBRKDBBRKDBBRKDBBRA
/sinsin2)()()()(
00
00
,coscos21 2 rrRRRCP
)(
)(where
0
0
--
--
KDBA
KDBAr
7
MEASUREMENT OF B-D0K- AT CDF
Use data collected with with CDF’s displaced track trigger detector between February 2002 and February 2006
Measure relative BR using kinematic information and particle-flavor identification in simultaneous maximum likelihood fit Invariant mass, particle momenta, dE/dx First measurement at a hadron collider!
8
SELECTION TO REDUCE BACKGROUNDS
Optimize sensitivity to ACP+ Fit in mass window m(D0p) [5.17, 5.60] GeV/c2
Reduce backgrounds from decays other than B-D0(*)p-
Veto event in ±2s around B-J/K- to eliminate contamination in B-D0[ pp-]K- sample Must include background
events from B- K- K K- in likelihood fit
9
RECONSTRUCTED B- AFTER SELECTION
N(B-D0[ K-p]p-) ~ 7,500
N(B-D0[ K+K-]p-) ~ 1,000
N(B-D0[ pp-]p-) ~ 250
D0 K-p
D0 K+K- D0 p+p-
10
INPUTS TO LIKELIHOOD FIT Variables input to fit
m(D0p) ptot = pp/K + pD0
a = 1 - pp/K/pD0, pp/K pD0
-(1 - pD0/pp/K), pp/K pD0
k = (dE/dxmeas – dE/dxexp(p))/ (dE/dxexp(K) – dE/dxexp(p))
11
MAXIMUM LIKELIHOOD FIT Total likelihood written for each channel and charge
Simultaneously minimize all six likelihoods Use MC to model correlations between a and ptot in
FDK, FDp, and FD*p PDFs Use mass sidebands for background Fbg PDF
)),(,,())),(,,()),(,,(
)),(,,()1(()1((
0
0**
0
0*
kpakpakpa
kpa
pppp
pp
DmpFfDmpFfDmpFf
DmpFfffL
totbgbg
totDDtotDD
Nevents
ktotDKDDbg
---
12
RAW ASYMMETRIES
Correct raw numbers Use Monte Carlo to correct for relative efficiencies
between decay modes Also correct for small biases observed in pseudo-
experiments
13
SYSTEMATIC UNCERTAINTIESSource R RCP+ ACP+
dE/dx 0.0028 0.056 0.030Input B- mass 0.0002 0.004 0.002D0*p- mass model 0.0028 0.025 0.006Combinatoric mass model 0.0002 0.020 0.001(a, ptot) model of combinatoric bg.
0.0002 0.100 0.020
(a, ptot) model of D0p- 0.0001 0.002 0.001(a, ptot) model of D0*p- 0.0007 0.004 0.002(a, ptot) model of D0K- 0.0006 0.002 0.004Bias error 0.0001 0.005 0.003MC statistics 0.002 - -
Total 0.0045 0.12 0.04
14
RESULTS
First measurement of CP observables at a hadron collider will soon be submitted to PRD!
R = 0.0745 ± 0.0043 (stat) ± 0.0045 (syst)
RCP+ = 1.30 ± 0.24 (stat) ± 0.12 (syst)
ACP+ = 0.39 ± 0.17 (stat) ± 0.04 (syst)
15
BACK-UP
16
DATA SELECTION Optimize selection for sensitivity to ACP+
Use pseudo-experiments to choose value of cuts based on smallest expected error on ACP+
Table of selection requirements
Parameter Valuec3D
2 13pT(B)/(pT(B)+DR=1.0pT(trk))
> 0.65
Lxy(B)/sLxy(B) > 12
Lxy(D) > 0.04 cmLxy(BD) > 0.01 cm|d0(B)| 0.007 cmpT(p/K ) > 2 GeV/cDR(p,D0) 2
17
SIGNAL PDFS Signal probabilities FDK and FDp include correlations
between a and ptot
Shapes of m(a, ptot) and P(a, ptot) are determined from CDF Monte Carlo that includes realistic detector simulation
)(),()),()(()),(,,( 00 kaapkpa PpPpmDmGDmpF tottotitoti -
---
--
--
22
22
222222
))2/()1((
))2/()1((
))2/((2),( 0
aa
aa
aa
p
p
totK
tot
totDKBtot
pm
pm
pmmmmpmwhere
2222
22222
))2/(())2/((
))2/()1((2 0
aa
aa
p
p
---
---
totKtot
totDKB
pmpm
pmmmm
for a > 0
for a 0
18
BACKGROUND PDFS Background probabilities FD*p and Fbg also
include correlations between a and ptot
PD*p (a, ptot) determined from CDF Monte Carlo that includes realistic detector simulation
Pbg (a, ptot) determined from mass sidebands
)(),())(()),(,,( 00 kapkpa PpPDmBGDmpF tottoti
19
RAW YIELDS FROM FITDecay Raw Fraction Raw YieldB+ D0p+ 0.902 ± 0.006 3769 ± 68B- D0p- 0.902 ± 0.006 3763 ± 68B+ D0K+ 0.060 ± 0.005 250 ± 26B- D0K- 0.064 ± 0.005 266 ± 27B+ DCP
0p+ [K+K-]p+ 0.902 ± 0.017 498 ± 29B- DCP
0p- [K+K-]p+ 0.849 ± 0.017 509 ± 29B+ DCP
0K+ [K+K-]K+ 0.060 ± 0.017 33 ± 11B- DCP
0K- [K+K-]K+ 0.116 ± 0.017 70 ± 13B+ DCP
0p+ [p+p-]p+ 0.902 ± 0.017 109 ± 14B- DCP
0p- [p+p-]p+ 0.849 ± 0.017 139 ± 15B+ DCP
0K+ [p+p-]K+ 0.060 ± 0.017 7 ± 3B- DCP
0K- [p+p-]K+ 0.116 ± 0.017 19 ± 5
20
EXTRACTING ASYMMETRY RATES Use fractions measured in fit times total
number of B - events measured in each sample to determine number of events in individual components e.g. N(B- D0 K- [K+K-]K-) = NKKK(1-fbg)(1-fDp-fD*p
Correct for relative efficiencies between decay modes using realistic MC
21
CHECKS OF FIT Check goodness of
fit with relative likelihood
)()()(
backgroundpdfDKBpdfDKBpdfRL