laurence carson, beauty 20091 charmless hadronic b decays at lhcb laurence carson, university of...
Post on 24-Jan-2016
213 views
TRANSCRIPT
Laurence Carson, Beauty 2009 1
Charmless Hadronic B Decays at LHCbLaurence Carson, University of Glasgow
(on behalf of the LHCb collaboration)
12th International Conference on B Physics at Hadron Machines
Heidelberg, 7th – 11th September 2009
Laurence Carson, Beauty 2009 2
Outline
• Overview of LHCb detector
• Physics opportunities in charmless hadronic B decays at LHCb– Measurement of γ in decays involving penguins
• Using Bd→π+π- and Bs→ K+K-
• Using B+/-→K+/-π+π- and Bd→KSπ+π-
– Search for rare decay modes– Many other analyses not covered here
• e.g. α measurement with Bd→ρπ, ρρ and ππ
Laurence Carson, Beauty 2009 3
• Take advantage of B production at LHC– pairs in 2fb-1
(nominal year @ 2*1032cm-2s-1)
– Produce all B species: B+,Bd,Bs,Bc,Λb,…
• Use trigger(*) to cope with high event rate– Level 0: hardware trigger on high pT/ET
leptons or hadrons
– High Level Trigger: matches L0 object to track with displaced vertex, then runs inclusive & exclusive selections
The LHCb Detector
bbbbb
1210500
Unique angular coverage: 1.9 < η < 4.9
Single forward arm spectrometer
(*) see talk on Friday by L. de Paula
Laurence Carson, Beauty 2009 4
bkg
Hadronic Modes at LHCb• Vertex Locator (VeLo): Si strip
detector giving στ ≈ 40fs; allows displaced vertex trigger and time-dependent Bs studies
• Cherenkov Counters (RICH): Excellent p/K/π separation for 2<p<100GeV; can separate out final states
Without RICH PID
With RICH PID
PID example: Bs→K+K-
Bd→π+π- τ
residualBs→K+K-
oscillations with 2fb-1 width = 39fs
Mass resoln ≈ 17MeV
Laurence Carson, Beauty 2009 5
Current Status of γ• Least well constrained parameter of the “db” unitarity triangle.
• Indirect measurement from fit to apex gives γ = (66.8 +5.4-3.8)°.
• Direct measurements from time-independent analyses of B+/- → D0(*)K+/-(*) (“tree-only” decays) give γ = (70+27
-29)°
– Analysis method (ADS/GLW/GGSZ) depends on D decay mode.
• Such tree-only decays expected to give SM value for γ.
Following talk (S. Ricciardi) will discuss how LHCb can improve on current γ measurements using tree-only decays.
Laurence Carson, Beauty 2009 6
LHCb γ Strategy
• Measurements of γ from decays involving both tree and penguin amplitudes are sensitive to New Physics effects– Compare to SM “standard candle” γ value from tree-only decays.
– Significant disagreement would be unambiguous sign of NP.
• Prospects for two such measurements have been studied by LHCb– Measurement of time-dependent CP asymmetries in Bd → π+π- and Bs
→ K+K- (method: R. Fleischer, Phys. Lett. B 459 (306) 1999).
– Combined Dalitz analysis of B+/- → K+/-π+π- and Bd → KSπ+π- (method:
I. Bediaga et al, Phys. Rev. D 76 (073011) 2007).
Gluonic penguin in SM
Gluonic penguin in MSSM
Laurence Carson, Beauty 2009 7
Measuring γ with Bd → π+π- and Bs → K+K-
Method first proposed by R. Fleischer, Phys. Lett. B 459 (306) 1999
Laurence Carson, Beauty 2009 8
• εgeo acceptance~21%, εselection(no PID)~18%, εtrigger~36% εtotal~1.4%
B→hh at LHCb• Studies of two-body charmless hadronic (“B→hh”) decays at LHCb
summarised as part of forthcoming “physics book”.– Including selection/bkg studies, calibration/systematics, and γ extraction.
Expected total mass distribution (ππ hypothesis) with 0.2fb-1
Channel 2fb-1 Yield (no PID)
Bd → π+π- 59k
Bd → K+π- 217k
Bs → K-π+ 15k
Bs → K+K- 72k
Λb → pπ- 7k
Λb → pK- 11k
3-body bkg [76k,186k]@68% C.L.
Combinatorial bkg [106k,226k]@68% C.L.
Laurence Carson, Beauty 2009 9
Some Calibrations for B→hh
• Proper time resolution model is Gaussian with mean and width depending on per-event proper time error στ
– Model can be determined on data using Bs → K-π+ (if mistag is known).
• Proper time acceptance distorted by IP cuts in trigger and selection – Can be determined on data by “swimming” the B to build up event-by-event
acceptance (top hat).
mean = GM.στ width = GS.στ
swimB→hh acceptance after: offline selection offline selection + trigger
• PID likelihood calibrated using D*+→ D0(→K-π+)π+ and Λ → pπ-.
Laurence Carson, Beauty 2009 10
γ from Bd → π+π-, Bs → K+K-
• Direct and mixing induced CP asymmetries measured using tagged, time-dependent analysis:
)2/sinh()2/cosh(
)sin()cos(
)()(
)()(
tAt
mtAmtA
fBfB
fBfB mixCP
dirCP
• Have 4 asymmetries depending on 7 physics parameters: ,
, , .
),,(1 dfAdir ),,,(2 d
mix dfA ),','(3 dfAdirKK ),,','(4 s
mixKK dfA
• Assume U-Spin (d↔s) symmetry d=d’ and θ=θ’
• Take φd from J/ψKS system becomes soluble.
• d (d’) is related to the ratio of the magnitudes of the penguin and tree contributions for the Bd (Bs) decay, while θ (θ’) is related to the difference in their strong phases.
Laurence Carson, Beauty 2009 11
LHCb B→hh Fit Approach• Fit for branching ratios, lifetimes, CP asymmetries, and detector parameters
is made in all B→hh channels simultaneously (68 free parameters)– Mass and PID likelihoods used to separate modes.
• With 2fb-1 of data, each of the four CP asymmetries from Bd→π+π- and Bs→K+K- will be measured with a statistical uncertainty of ≈0.04.
• For the extraction of γ, Adir(KK) can be replaced with ACP(Bd→K+π-), since they are expected to be equal given U-spin symmetry, and the uncertainty on ACP(Bd→K+π-) with 2fb-1 will be <0.01 (systematics dominated).
Fit to Bs→K+K- mass distribution under ππ hypothesis (full MC)
PID distribution for 0.2fb-1 toy MC sample
π K p
Laurence Carson, Beauty 2009 12
LHCb Sensitivity to γ• Extraction of γ allows for some U-spin breaking, with 0.8 < d’/d < 1.2 and
-20° < θ’ - θ < 20°.• SM solution for γ is isolated by imposing the condition θ>90°.
68% C.L.95% C.L.
γ (input 70°) φs (input -0.04 rad)
• With 2fb-1, find σγ = 7°, σφs = 0.05 rad , σd = 0.13 and σθ = 7°.
Contours in (d,γ) plane without imposing θ >90°
• Alternative strategy would be to constrain φs (expect σφs≈ 0.03 rad from Bs→J/ψφ with 2fb-1) and
release one of the U-spin constraints d=d’ or θ=θ’.
Laurence Carson, Beauty 2009 13
Measuring γ with B+/-→K+/-π+π- and Bd→KSπ+π-
Method first proposed by I. Bediaga et al, Phys. Rev. D 76 (073011) 2007
Laurence Carson, Beauty 2009 14
• B+/-→K*0π+/- penguin only; Bd→K*+π- penguin and tree
– Extract penguin amplitude (VtbV*tsP) via standard Dalitz analysis of B+/- →K+/-π+π-
– Use this as input to untagged Dalitz analysis of Bd→KSπ+π- to extract tree amplitude (VubV*usT)
– Tagging not required as resonance interference regions for Bd and Bd are distinct
– Parameters extracted relative to contribution from χc0K (common contribution to Bd and B+/-)
γ from B+/-→K+/-π+π-, Bd→KSπ+π-
Combined: two distinct regions of resonance interference
Bd→KSπ+π- Bd→KSπ+π-
• Tree amplitudes in three-body charmless hadronic decays involve Vub
– Probe γ using interference between intermediate resonances
Resonances: K*(892)+/-, K*(1430)+/-, ρ(770)0, f0(980), χc0
m2(KSπ+) m2(KSπ+) m2(KSπ+)
m2 (
KSπ- )
m2 (
KSπ- )
m2(K
Sπ- )
Laurence Carson, Beauty 2009 15
B+/-→K+/-π+π-, Bd→KSπ+π- at LHCb
• Lower efficiency for Bd→KSπ+π- due to lower KS reconstruction efficiency– 25% decay in VeLo good efficiency/resolution– 50% decay between VeLo and TT Stations poorer efficiency/resolution– 25% decay downstream of TT Stations track lost
Channel B.R. εselection 2fb-1 yield (no HLT)
B+→K+π+π- 5.5*10-5 1.1% 494k
Bd→KSπ+π- 2.2*10-5 0.5% 90k
• Selection: want to avoid tight cuts on pT as this lowers efficiency in corner of Dalitz plot, where some of the resonances lie– Trade-off between minimising bias and reducing
combinatoric bkg
B+→K+π+π- with pT >200 MeV for
all daughters
• Sensitivity to γ : idealised toy MC study (no background or detector effects)
estimates σγ ~5° (stat. only) with 2fb-1.
Laurence Carson, Beauty 2009 16
• Baryonic mode Bd → pp unobserved SM predictions close to being ruled out
– If true B.R. close to current limit (1.1*10-7), expect evidence with ~0.2fb-1 (2010 run)
– Also expect to observe Bd → K+K- and Bs → π+π- (penguin annihilation and W exchange only) with ~1fb-1 (if SM B.R.s).
Rare Hadronic B Decays at LHCb
Specific bkgs to Bd→pp suppressed by RICH PID
• B+ → K-π+π+ and B+ → K+K+π- highly suppressed in SM– B.R.s O(10-14) and O(10-11) respectively.
• Can be strongly enhanced by NP effects– e.g. B.R.(B+→K-π+π+)~10-6 (≈ current B.R. limit) with Z’.
• If such enhancements are present, LHCb may be able to make an observation.
Laurence Carson, Beauty 2009 17
• Search for undiscovered rare decays– Expect to observe Bs → π+π-, Bd → K+K- and Bd → pp with early data, whether SM
branching ratio or not
– Can observe B+ → K-π+π+, B+ → K+K+π- if B.R. strongly enhanced by NP.
• Measure α using Bd → ρπ, ρρ, and ππ
– Expect σα(stat) ~ 5° after 2fb-1 (dominated by ρπ).
Conclusions• LHCb has a rich physics programme using charmless hadronic decays:• Measure γ from decays with penguins and compare to SM value from tree-
only decays– Expect σγ = 7° after 2fb-1 from analysis using Bd → π+π-, Bs → K+K-
– Measurement with B+/- → K+/-π+π- and Bd → KSπ+π- should be competitive (effect
• We are awaiting the first collisions…
of backgrounds and systematics to be checked).
Laurence Carson, Beauty 2009 18
Backup Slides
Laurence Carson, Beauty 2009 19
More Calibrations for B→hh
• PID calibrated using PID-unbiased samples of D*+→ D0(→K-π+)π+ and Λ → pπ-
– Dedicated stream in trigger
– Re-weight likelihood distribution for B using p and pT distributions.
• Tagging performance for Bd measured using flavour-specific Bd → K+π-
– For Bs tagging, Bs → K-π+ has less statistics and is sensitive to proper time resolution
– Can study Bd-Bs difference in other modes (e.g. Bd → D-π+ and Bs → Dsπ+)
– Can also calibrate using per-event mistag.
DLL distributions: Bd →π+π-, π from D → Kπ before reweighting, π from D → Kπ after reweighting.
Per-event mistag for Bd → K+π-
Laurence Carson, Beauty 2009 20
Measuring α with Bd → π+π- π0
• α can be measured via a tagged Dalitz analysis of Bd → π+π- π0
– Dominant submodes are ρ+π-, ρ-π+ and ρ0π0
– Isospin symmetry used to extract α, tree/penguin amplitudes and strong phases.
• Multivariate selection used, εtotal~0.07%. Expected 2fb-1 yield is 14k, with B/S~1.
Resolved π0
Merged π0
• 2fb-1 toy experiment gives α = (97+9-4)° (stat. only)
Sum of Dalitz coords
Effect of ET cut on π0
Time evolution of tagged Dalitz plotsbefore
after