Mixing and CP Violation at the

Gustaaf Brooijmans, on behalf of the CDF and DØ Collaborations

HCP 2008, May 29, 2008

Gustaaf Brooijmans Mixing and CP Violation 1 Outline

• Bs mixing • D mixing: mixing interference in decay • CP Violation: • B+ → J/ψ K+: CPV in decays • Semileptonic ACP: CPV in mixing • Bs → J/ψ φ: CPV in interference of mixing and decay • Conclusions

(Throughout this talk, conjugated modes are implicitly assumed.) Gustaaf Brooijmans Mixing and CP Violation 2 Mixing

Define x = ΔM/Γ, y = ΔΓ/2Γ, and

x’ = x cosδ +y sinδ, y’ = -x sinδ + y cosδ (δ is a strong phase) Gustaaf Brooijmans Mixing and CP Violation 3 Feynman Diagrams (B) • Processes (dominantly) mediated by box diagrams

• Since Vtb is large, diagrams with top dominant

• In M12, mass comes in as mtop2/ mW2 • High oscillation frequency was the first experimental indication of large top mass (1987) • Sensitive to new heavy particles in loops

Gustaaf Brooijmans Mixing and CP Violation 4 Analysis Strategy

Opposite Side Reconstructed Side

• Reconstruct Bs decay on one side, flavor at decay is given by charges of decay products, measure and (transverse) decay length LT (to determine proper lifetime) • Tag flavor at production using either “opposite side” b-hadron (most b’s are pair-produced), or “same side” hadron from fragmentation (for Bs, in principle a K)

Gustaaf Brooijmans Mixing and CP Violation 5 Reconstruction • Look for Bs → Ds X • X can be , 1 or 3 pions • Consider multiple Ds decay channels • Use NN (CDF) / likelihood (DØ) to increase S/B • CDF includes particle ID based on TOF, dE/dx in drift chamber • In semileptonic Bs decays do not reconstruct Bs momentum fully ➡ Apply a “K-factor” to rescale momentum

Gustaaf Brooijmans Mixing and CP Violation 6 Flavor Tagging • At the Tevatron, most b quarks are produced in bb pairs • Can use opposite-side b quark for tagging • Same-side hadrons also useful (but usually less pure, because sensitive to fragmentation fluctuations) • Both experiments combine tags in NN (CDF) or likelihood (DØ)

• Verify performance using Bd mixing

Gustaaf Brooijmans Mixing and CP Violation 7 Result • Maximizing sensitivity requires using all information (uncertainties) on an event-by-event basis ➡ Unbinned maximum likelihood fits

DØ: small contribution from CDF: hadronic channels dominant hadronic channels large L1 bandwidth, SVT & PID PRL 97, 242003 2006

ΔlnL = -17.26

-1 Δms = 17.77 ± 0.10 (stat) ± 0.07 (syst) ps

Gustaaf Brooijmans Mixing and CP Violation 8 u¯ u¯

cu¯ u¯s

c s W + d¯

+ ¯ W du u u¯ u¯ u¯ u¯ c d c d W + s¯ + s¯ D Mixing W u u u d, s, b c • D oscillations expected to be u d, s, b c slow in the SM W W W W Negligible! Box diagram (ΔC=2) very small c¯ d,¯ s,¯ ¯b u¯ • c¯ d,¯ s,¯ ¯b u¯ due to CKM & GIM suppressions u¯ u¯ u¯ s,u¯d c u¯ s, d W u¯ c s,¯ d¯ W W “Long range” (2 x ΔC=1) s,¯ d¯ c¯ • c s W u c¯ dominant, proportional to u u + ¯ W ud breaking of SU(3)f “Long Range” u

Cannot observe oscillation u¯ u¯ • 0 π- D c Mixing interference may be d • + s¯ visible in doubly Cabibbo- W K+ u suppressed decays DCS Decay u d, s, b c

Gustaaf Brooijmans Mixing and CP Violation 9 W W

c¯ d,¯ s,¯ ¯b u¯

u¯ u¯ 1 s,1d c W s,¯ d¯ W c¯ u u

1 Analysis Technique

"103 CDF II Preliminary (1.5 fb-1) • Look for mixing-induced time 2 800 dependence of D0→K+π- / Favored 600 D0→K-π+ ratio Events / 0.5 MeV/c 400

• Sample collected with IP trigger 200 Use D*+ → π+ D0, D0 → Kπ chain 0 10 20 30 • !m (MeV/c2) 0 to “tag” D flavor at production CDF II Preliminary (1.5 fb-1) 2 10000 • Include PID, reject Kπ candidates that fall into D0 mass window with

both mass assignments Events / 0.5 MeV/c 5000 Doubly Cabibbo Look at Δm = m(Kππ) - m(Kπ) - • Suppressed m(π): should be ~5 MeV 0 10 20 30 !m (MeV/c2)

Gustaaf Brooijmans Mixing and CP Violation 10 D Mixing!

CDF II Preliminary (1.5 fb-1) 0.01 R Largest lifetime range CDF II PreliminaryarXiv:0712.1567 (1.5 fb-1) )

explored so far -3 20

0.008 y’ (10

10 2 /ndf = 19.2/17 0.006 Parabolicχ Ratio ( With Mixing) 3.8σ! 0 1σ 0.004 Fixed Ratio (No Mixing) σ χ2/ndf = 36.8/19 -10 2 3σ 4σ 0 2 4 6 8 10 t (ps) -0.5 0 0.5 x’2 (10-3) • Dominant systematics: signal and background shapes, fraction of D*’s from B decays

Gustaaf Brooijmans Mixing and CP Violation 11 CP Violation • In the decay amplitudes: B+ → J/ψ K+ • Γ(B → f) ≠ Γ(CP(B) → CP(f)) • Only possibility in charged decays • In mixing: Same-Sign Dimuons • Observe through in charged-current neutral meson decays: ASL

• Interference between the two: Bs → J/ψ φ • Interference between decays without/with mixing: need common final state

Gustaaf Brooijmans Mixing and CP Violation 12 B+ → J/ψ K+ (π+) • Select sample with J/ψ decaying to a pair of • Very clean sample, 40k candidate events • Fit for : • Divide sample in 8 based on solenoid polarity, sign of J/ψ K+ pseudo-rapidity η, and K charge

• All detector & forward- backward asymmetries found compatible with 0

Gustaaf Brooijmans Mixing and CP Violation 13 B+ → J/ψ K+ (π+): Result • Need to correct observed asymmetry for K± reconstruction asymmetry • Due to asymmetry in interactions (detector is made of )

• Measure in D*+ → D0π+, D0 → μνμK- + + ACP(B → J/ψΚ ) = 0.0075 ± 0.0061(stat) ± 0.0027 (syst)

+ + ACP(B → J/ψπ ) = -0.09 ± 0.08 (stat) ± 0.03 (syst)

arXiv:0802.3299

(Dominant systematic is mass distribution model)

Gustaaf Brooijmans Mixing and CP Violation 14 Same-Sign Dimuons • Measure • Multiple techniques: • Use all dimuon events, and estimate contributions from all processes (incl. sequential decays, Drell-Yan, instrumentals, etc): DØ (2006) Phys. Rev. D 74 , 092001 (2006) • Use all dimuon events but use impact parameter distributions to unfold contributions from various sources: CDF

q • Then derive A SL from sample composition, or • Use flavor-specific decays, e.g. Bs → μ+νDs-X: DØ

Gustaaf Brooijmans Mixing and CP Violation 15 Inclusive Measurements -1 CDF Run II Preliminary, L = 1.6 fb

New CDF measurement uses X-projection • minus-minus Data [205158] PP [51723 +/- 620] BB [103449 +/- 711] 2-D (μμ) impact parameter 4 BP [36225 +/- 1079] 10 CC [13762 +/- 1433] significance distribution

Number of Events 103 • Asymmetry from K/π faking *+ measured in D decays, 102 + unfold direct and sequential 0 20 40 60

decay contributions d0/!(d0)

-1 -1 μμ CDF Run II Preliminary, L = 1.6 fb CDF Run II Preliminary, L = 1.6 fb A SL= 200 - 180 + 0.0080 ± 0.0090 (stat) ± 0.0068 (syst) 120 K K 160 100 140

80 120 100 Number of Events Number of Events 60 DØ 2006 fully inclusive 80 40 60 measurement using 40 20 20 “8 subsamples technique”: 0 0 1.82 1.84 1.86 1.88 1.90 1.92 1.82 1.84 1.86 1.88 1.90 1.92 μμ D0 Mass for K- faking µ (GeV) D0 Mass for K+ faking µ (GeV) A SL= -0.0053 ± 0.0025 (stat) ± 0.0018 (syst)

Gustaaf Brooijmans Mixing and CP Violation 16 AdSL & AsSL

μμ • A SL has contributions from both Bd and Bs: d • Use A SL from the B-factories, Bd and Bs production ratios, and mixing parameters to extract AsSL • From CDF impact parameter analysis: • AsSL = 0.020 ± 0.021 (stat) ± 0.016 (syst) ± 0.009 (inputs) • From DØ inclusive analysis: • AsSL = -0.0064 ± 0.0101 (uncertainties combined) Phys. Rev. D 76 , 057101 (2007 ) + - • From DØ exclusive measurement (Bs → μ νDs X): • AsSL = 0.0245 ± 0.0193 (stat) ± 0.0035 (syst) Phys. Rev. Lett. 98 , 151801 (2007 )

Gustaaf Brooijmans Mixing and CP Violation 17 Bs → J/ψ φ: CPV in Interference • Equivalent of Bd → J/ψ KS in Bd system • Access to the equivalent of sin(2β): sin(-φs) or sin(2βs) Bd J/ψ KS Bs J/ψ φ

Bd Bs

• But βs is much smaller in the SM than β:

Gustaaf Brooijmans Mixing and CP Violation 18 Bs → J/ψ φ Signal • CDF uses neural network, with variables including PID from TOF & dE/dx in drift chamber • DØ uses “square cuts” event selection (no PID) CDF Run II Preliminary, L = 1.35 fb-1 ~2000 signal 2 candidates data 200

fit

100 ~2000 signal candidates Candidates per 2.0 MeV/c

0 5.3 5.35 5.4 5.45 2 Mass(J/" !) (GeV/c ) Gustaaf Brooijmans Mixing and CP Violation 19 J/ψ φ CP? → Decay Angles • VV final state ⇒ both CP eigenvalues possible. • Disentangle by determining relative polarization of J/ψ and φ • Express angular dependence in J/ψ rest frame (“transversity basis”) in terms of polarization ⇒ extract CP eigenvalue

Gustaaf Brooijmans Mixing and CP Violation 20 Flavor Tagging

• Needed to constrain flavor of decaying Bs • Removes a twofold ambiguity in result • Both experiments use both opposite-side and same- side tags • DØ: combined tagging power εD2 = 4.68 % • CDF: • OST tagging power εD2 = 1.28 % • SST tagging power εD2 = 3.65 %

Gustaaf Brooijmans Mixing and CP Violation 21 Fit • Use unbinned maximum likelihood fit to maximize sensitivity

• Constrain Δms = 17.77 ± 0.12 ps-1 • (Constrain or float strong phases) • Extract average lifetime, ΔΓs, φs (=-2βs), magnitudes of polarization amplitudes, and strong phases • NB: detector efficiencies are not flat vs transversity angles, correct • CDF likelihood profiles not parabolic close to minima • Use Feldman-Cousins-like likelihood ratio ordering to build 2-D confidence region

Gustaaf Brooijmans Mixing and CP Violation 22 Result: CDF With BABAR constraints SM probability: 15% ✓ on strong phases from B0 →J/ψ K*0 -1 CDF Run II PreliminaryarXiv:0712.2397 L = 1.35 fb CDF Run II Preliminary L = 1.35 fb-1

2 log(L) = 5.99 ) 0.6 95% C.L. ) 0.6 # -1 -1 68% C.L. 2#log(L) = 2.30 0.4 SM prediction 0.4 SM prediction (ps (ps

" " (Frequentist contours, # 0.2 # 0.2 not CL)

0.0 0.0

-0.2 -0.2 constrain strong phases BaBar: -0.4 -0.4 2#log(L) = 5.99 2#log(L) = 2.30 -0.6 -0.6 -1 0 1 -1 0 1 (rad) (rad) !s !s

Gustaaf Brooijmans Mixing and CP Violation 23 Result: DØ arXiv:0802.2255 68 & 90% CL curves (With BABAR constraints on strong phases from B0 →J/ψ K*0)

SM probability: ~7% ✓

Gustaaf Brooijmans Mixing and CP Violation 24 Combination & Prospects • UTfit: arXiv:0803.0659 * • Strong phases derived from BABAR Bd → J/ψ K may not be completely valid, so try to unfold this from DØ result • (UTfit combination also uses asymmetries) • Some uncertainty with this, but any of the tried procedures lead to values of φs 3σ or more from the • For both CDF and DØ, the measurement uncertainty is completely dominated by statistics ➡ Result will still improve quite a bit • If confirmed, this would be the first evidence for CP violation outside the CKM mechanism.

Gustaaf Brooijmans Mixing and CP Violation 25 Conclusions • Evidence for D mixing seen by CDF • ACP(B+→J/ψK+) and AsSL compatible with 0 • Bs mixing parameters: • Δms: evidence by DØ, observation by CDF • ΔΓs: good agreement with Standard Model • φs: suggestion of CP violation beyond the CKM mechanism? To be continued....

Gustaaf Brooijmans Mixing and CP Violation 26 Bd,u and Bs • Many aspects of CP violation in Bd,u well measured • Lots from B-factories, and important Tevatron contributions • e.g. B+ → J/ψ K+ • Tevatron also has access to large Bs sample! • CP violation expected to be small in b-c transitions from Bs in the Standard Model

VCKM =

Gustaaf Brooijmans Mixing and CP Violation 27 New Physics in B Mixing • Supersymmetric contributions for example (See Bertolini et al., Nucl. Phys. B353, 591 (1991)), but note that you still need flavor violation • See for example Foster et al., Phys.Lett.B641:452-460,2006

Gustaaf Brooijmans Mixing and CP Violation 28 B Decay!

Gustaaf Brooijmans Mixing and CP Violation 29