Table Of ContentMeasurement of B Oscillations and CP Violation
s
9 Results from DØ
0
0
J. Ellison, for the DØ Collaboration
2
University of California, Riverside, CA 92521, USA
n
a
J
0
2 Abstract
] We present a measurement of the B0 −B¯0 oscillation frequency, ∆m , using a combination
x s s s
of semi-leptonic and hadronic B decay candidates selected from data collected by the DØ
e s
ExperimentattheFermilabTevatron.WealsopresentseveralresultsonCPviolation,including
-
p an improved measurement of the B CP-violating phase from a flavor-tagged analysis of B0 →
s s
e J/ψ+φ decays.
h
[
2 Key words:
v PACS:
5
3
1 1. Introduction
2
.
1 Among the primary goals of the Tevatron is the observation of B oscillations and the
s
0
searchforCPviolationinB-mesondecays.Inthispaperwedescribesomerecentresults
9
in these areas. The data used for the results described here are based on approximately
0
: 1.2−2.8 fb−1 of data collected by the DØ Experiment at the Fermilab Tevatron. The
v
DØ detector consists of a central tracking system surrounded by a uranium liquid-argon
i
X calorimeter and an outer muon detection system and is described in [1].
r
a
2. Measurement of B Oscillations
s
InthispaperwereportapreliminarymeasurementoftheB0−B¯0 oscillationfrequency
s s
∆m based on the analysis of five decay channels: (1) B0 → D−µ+νX,D− → φπ−;
s s s s
(2) B0 → D−e + νX,D− → φπ−; (3) B0 → D−µ+νX,D− → K∗0K−; (4) B0 →
s s s s s s s
D−µ+νX,D− →K0K−;(5)B0 →D−π+,D →φπ−.Thedatasetswereapproximatley
s s s s s s
2.4 fb−1 for all channels except (4) whihc used a dataset of 1.2 fb−1. DØ has previously
reporteddirectlimitsonB oscillationsusing1fb−1 ofdatainthefirstdecaychannel[2].
s
PreprintsubmittedtoElsevier January21,2009
After selection and reconstruction of the B0 decays, the flavor of the B0 at decay is
s s
determined from the charge of the muon, electron, or pion in the final state. The flavor
at production is determined using a combined flavor tagger based on opposite-side and
same-side tagging [3]. The effectiveness of the tagging algorithm is characterized by the
tagging efficiency (cid:15) and the dilution D defined by D = (N −N )/(N +N ),
RS WR RS WR
where N and N are the number of right-sign and wrong-sign tags, respectively.
RS WR
The overall effectiveness of the combined tagger, obtained using B → µνD∗− data and
B →J/ψφ data and Monte Carlo, is (cid:15)D2 =4.49±0.88%.
s
The determination of the proper decay length takes into account possible missing
momentum, e.g. due to the presence of a neutrino in the semileptonic decay modes. The
properdecaylengthisgivenbycτ =xMK,wherexM isthemeasuredorvisibleproper
B0
s
decay length, given by
(cid:34)d(cid:126)Bs0 ·p(cid:126) (cid:96)Ds−(cid:35) pµDs−
xM = T T cM with K = T
(p(cid:96)Ds−)2 Bs0 pBs0
T T
The K-factor, K, is determined from Monte Carlo.
Themeasurementof∆m isaccomplishedusinganunbinnedlikelihhodfittothedata.
s
The likelihood function is constructed using the probabilities for mixed and unmixed
decays,
(cid:18) (cid:19)
1 K Kx
P (t)= exp − (1−Dcos∆m Kx/c)
mix 2cτ cτ s
Bs Bs
(cid:18) (cid:19)
1 K Kx
P (t)= exp − (1+Dcos∆m Kx/c)
nomix 2cτ cτ s
Bs Bs
and accounting for detector resolutions and background contributions. The combined
amplitude scan, where the likelihood is modified to include an amplitude term, L ∝
1±ADcos(∆m Kx/c), is shown in Fig. 1(a). The likelihood as a function of ∆m is
s s
shown in Fig. 1(b). The measured value of ∆m is obtained from a fit to the likelihood
s
vs. ∆m in the region of the minimum shown in Fig. 1(b), and yields ∆m = 18.53±
s s
0.93(stat.)±0.30(syst.) ps−1. The significance of this result is 2.9 σ.
The∆m measurementcanbetranslatedintoaconstrainton|V /V |,andweobtain
s td ts
(cid:12) (cid:12)
(cid:12)(cid:12)Vtd(cid:12)(cid:12)=0.2018±0.005(exp.)+0.0081(theor.)
(cid:12)V (cid:12) −0.0060
ts
3. Analysis of B0 → J/ψ+K∗0 and B0 → J/ψ+φ Decays
d s
Before discussing CP Violation in B0 →J/ψ+φ, we present results of an analysis of
s
B0 →J/ψ+K∗0 andB0 →J/ψ+φdecays,inwhichCPviolationisneglected.Theaim
d s
istomeasuretheparameterswhichdescribethetime-dependentangulardistributionsof
thesedecayswheretheinitialB mesonflavorisnotdetermined.Therelevantparameters
are the linear polarization amplitudes, the strong phases, and the lifetimes. This allows
ustoobtainthelifetimeratio,andtestSU(3)flavorsymmetryandfactorization.Thisis
achievedbyfittingtothemass,lifetime,andangulardistributionsofthedecayproducts.
For details of the analysis see [4].
2
Fig.1.(a)Bs0 oscillationamplitudeasafunctionof∆ms and(b)Bs0 oscillationlikelihoodasafunction
of∆ms.
The preliminary results are shown in Table 1. The B results are competitive and
d
consistentwithpreviousmeasurementsbyCDF,BaBar,andBelle.Themeasurementsof
the strong phases indicate the presence of final-state interactions for B0 → J/ψ+K∗0,
d
since δ is 3.5 σ away from zero. Also, comparison of the polarization amplitudes and
1
strong phases shows no evidence for violation of SU(3) flavor symmetry.
Table1
Preliminaryresultsforthelinearpolarizationamplitudes,strongphases,lifetimeandwidthdifference
forBd0→J/ψ+K∗0 andBs0→J/ψ+φ.A0 andA(cid:107) arethelinearpolarizationamplitudes,δ1,δ2,and
δ(cid:107) are the strong phases, τ is the lifetime, and for the Bs decay, ∆Γs is the width difference between
thelightandheavymasseigenstates.
4. Flavor-Tagged Analysis of B0 → J/ψ+φ Decays
s
The B0 → J/ψ φ decay involves a final state that is a mixture of CP-even and CP-
s
oddstates.InordertoseparatetheCP-evenandCP-oddstates,weperformamaximum
likelihoodfittothemass,lifetime,andtime-dependentangulardistributionsoftheB0 →
s
J/ψ(→µ+µ−)φ(K+K−) decay. The fit yields the CP-violating phase φ and the width
s
difference∆Γ ≡Γ −Γ .Thedecaycanbedescribedbythreedecayanglesθ,φ,andψ
s L H
definedin[5].Weemployinitialstateflavortaggingwhichimprovesthesensitivitytothe
3
Fig.2.(a)Resultsofthefitsintheφs−∆Γs planefortheDØBs→J/ψφanalysisbasedon2.8fb−1
ofdata.(b)FitresultsfromthecombinationoftheDØresultswithCDFresultsbasedon1.35fb−1 of
data.
CP-violating phase and removes a sign ambiguity on φ for a given ∆Γ present in our
s s
previous analysis [6]. In the fit, ∆M is constrained to its measured value (from CDF)
s
and the strong phases are constrained to values measured for B at the B-factories,
d
allowing some degree of violation of SU(3) symmetry. The B flavor at production is
s
determined using a combined opposite-side plus sameside tagging algorithm. Confidence
level contours in the φ −∆Γ plane are shown in Fig. 2(a).
s s
The fit yields a likelihood maximum at φ =−0.57+0.24 and ∆Γ =0.19±0.07 ps−1,
s −0.30 s
where the errors are statistical only. As a result of the constraints on the strong phases,
the second maximum, at φ = 2.92+0.30, ∆Γ = −0.19±0.07 ps−1, is disfavored by a
s −0.24 s
likelihood ratio of 1:29.
Fromthefitresultsandstudiesofthesystematicerrorsweobtainthewidthdifference
∆Γ ≡ Γ −Γ = 0.19±0.07(stat)+0.02(syst) ps−1 and the CP-violating phase, φ =
s L H −0.01 s
−0.57+0.24(stat)+0.07(syst). The allowed 90% C.L. intervals of ∆Γ and φ are 0.06 <
−0.30 −0.02 s s
∆Γ <0.30 ps−1 and −1.20<φ <0.06, respectively. The probability to obtain a fitted
s s
value of φ lower than -0.57 given SM contributions only is 6.6%, which corresponds to
s
approximately 1.8 σ from the SM prediction.
The results have been combined with CDF results by the Heavy Flavor Averaging
Group. The CDF results are based on a data set of 1.35 fb−1 and the combination was
performed with no constraints on the strong phases. The results are shown in Fig. 2(b).
The fit yields two solutions as follows:
φ =−2.37+0.38 rad, ∆Γ =−0.15+0.066 ps−1
s −0.27 s −0.059
φ =−0.75+0.27 rad, ∆Γ =0.15+0.059 ps−1
s −0.38 s −0.066
The p-value assuming only SM contributions is 3.1%, corresponding to 2.2 σ from the
SM prediction.
4
5. Search for Direct CP Violation in B± → J/ψ K±(π±) Decays
DØ has performed a search for direct CP violation by measurement of the charge
asymmetry in B± →J/ψ K±(π±) decays [7]. The charge asymmetry is defined by
N(B− →J/ψ K−(π−))−N(B+ →J/ψ K+(π+))
A (B+ →J/ψ K+(π+))=
CP N(B− →J/ψ K−(π−))+N(B+ →J/ψ K+(π+))
Direct CP violation in these decays leads to a non-zero charge asymmetry. In the SM
there is a small level of CP violation: A (B+ →J/ψ K+)≈0.003 [8] and A (B+ →
CP CP
J/ψ π+)≈0.01 [9]. New physics may significantly enhance A .
CP
After selection of B± → J/ψ K±(π±) candidates, the invariant mass of the J/ψK is
constructed and fit to the sum of contributions from B± → J/ψ K±, B± → J/ψ π±,
B± →J/ψ K∗,andcombinatorialbackground.Fromthefitwefindapproximately40,000
B± → J/ψ K± candidates and about 1,600 B± → J/ψ π± candidates. To extract the
charge asymmetry the sample is divided into 8 subsamples according to the solenoid
polarity, the sign of the pseudorapidity of the J/ψK system, and the charge of the K
candidate. A χ2 fit to the number of events in each subsample yields the integrated raw
chargeasymmetry.Thisisthencorrectedfortheasymmetryofthekaoninteractionrate
on nucleons to obtain the final results:
A (B+ →J/ψ K+)=+0.0075±0.0061(stat.)±0.0027(syst.)
CP
A (B+ →J/ψ π+)=−0.09±0.08(stat.)±0.03(syst.)
CP
Theresultsareconsistentwiththeworldaverageresults[10].TheprecisionforA (B+ →
CP
J/ψ π+) is comparable with the current world average, while for A (B+ → J/ψ K+)
CP
the precision is a significant (factor of 2.5) improvement over the current world average.
6. Search for Direct CP Violation in Semileptonic B Decays
s
InthissectionwereportanewsearchforCPviolationinthedecayB0 →D−µ+νX, (D− →
s s s
φπ−, φ → K+K−) by measurement of the charge asymmetry using a time-dependent
analysis with flavor tagging. The technique used is similar to that used in the DØ B
s
oscillation analysis (see Section 2). A fit to the invariant KKπ mass of the selected data
is shown in Fig. 3.
The sample is divided into 8 subsamples according to solenoid polarity, the sign of
the pseudorapidity of the D µ system, and the muon charge. An unbinned likelihood
s
fit is used to extract the asymmetry. The systematic uncertainties are mainly due to
uncertaintiesinthecc¯contribution,uncertaintiesintheefficiencyvsvisibleproperdecay
length, and uncertainties in the B →D(∗)µν branching fractions. Accounting for these
s s
yields the final result:
as =−0.0024±0.0117(stat.)+0.0015(syst.)
sl −0.0024
This result is consistent with the SM prediction and is the most precise measurement
to date.
5
Fig.3.TheKKπ invariantmassdistributionshowingtheµ+D− andµ+D− signalstogetherwiththe
s
fitresults.
7. Measurement of Br(B0 → D(∗)D(∗))
s s s
FortheB0 system,thewidthdifferencebetween∆Γ ≡Γ −Γ isrelatedtothewidth
s s L H
difference between the CP eigenstates, ∆ΓCP ≡ Γeven −Γodd, by ∆Γ = ∆ΓCP cosφ .
s s s s s s
The quantity ∆ΓCP can be estimated from the branching fraction Br(B0 →D(∗)D(∗)).
s s s s
DØhasperformedapreliminarymeasurementofBr(B0 →D(∗)D(∗))basedon2.8fb−1
s s s
ofdatausingeventswhichcontaindecaysD+ →φπ+ andD− →φµ−ν (andtheircharge
s s
conjugates).Thesignalisextractedfroma2-dimensionallikelihoodfittothedatainthe
m(D φπ)−m(φµ) plane. The results are [12]:
s
Br(B0 →D(∗)D(∗))=0.035±0.010(stat.)±0.011(syst.)
s s s
∆ΓCP
s =0.072±0.021(stat.)±0.022(syst.)
Γ
s
Assuming CP violation in the SM is small, the results are in good agreement with the
SM prediction.
References
[1] V.M. Abazov et al., (The DØ Collaboration), “The Upgraded D0 Detector,” Nucl. Instr. and
MethodsA565,463(2006).
[2] V.M.Abazovetal.,(TheDØCollaboration),Phys.Rev.Lett.97,021802(2006).
[3] V. M. Abazov et al., (The DØ Collaboration), Phys. Rev. D 74, 112002 (2006); V. M. Abazov et
al.,(TheDØCollaboration),DØConferenceNote5210.
[4] V.M.Abazovetal.,(TheDØCollaboration),arXiv:0810.0037,submittedtoPhys.Rev.Lett.
[5] V.M.Abazovetal.,(TheDØCollaboration),arXiv:0802.2255,submittedtoPhys,Rev.Lett.
[6] V.M.Abazovetal.,(TheDØCollaboration),Phys.Rev.Lett.98,121801(2007).
[7] V.M.Abazovetal.,(TheDØCollaboration),Phys.Rev.Lett.100,211802(2008).
[8] W.-S.Hou,M.Nagashima,A.Soddu,arXiv:hep-ph/0605080.
[9] I.Dunietz,Phys.Lett.BB316,561(1993);W.-S.Hou,arXiv:hep-ph/9905541(1999).
[10]C.Amsleretal.,PhysicsLettersB667,1(2008).
[11]V.M.Abazovetal.,(TheDØCollaboration),Phys.Rev.Lett.97,021802,(2006).
[12]V.M.Abazovetal.,(TheDØCollaboration),arXiv:0811.2173,submittedtoPhys.Rev.Lett.
6
