Inclusive Radiative B meson decays at Belle b sg Boštjan Golob Belle & Belle II University of Ljubljana, Jožef Stefan Institute 1. Introduction 2. Measurement method background treatment corrected spectra 3. Results 4. Constraints 5. Summary

Introduction Motivation FCNC process; sensitive to NP in loop; also b s W± u, c, t Vqb Vqs g Motivation FCNC process; sensitive to NP in loop; also parton level: Eg ≈ mb/2; dG/dEg → determ. of mb, Fermi motion (mp2) →Vub,Vcb b s u, c, t X Y H± g c± b s X Y g MH± [GeV] b sg U. Haisch, arXiv:0805.2141 tanb

Introduction Difficulties theory: parameter extraction from I. Bigi, N. Uraltsev, Int.J.Mod.Phys. A17, 4709 (2002) fraction of b sg with Eg < E; Fg(E=1.9 GeV) ~ 7% Fg(E=1.7 GeV) ~ 3 % Difficulties theory: parameter extraction from partial Br(Eg>Ecut) → extrapolation needed; experiment: measure low Eg → huge bkg. continuum p0  gg h  gg b  sg 1.7 GeV signal

Measurement Inclusive measurement l± Background treatment Belle, arXiv:0907.1384, 605 fb-1, subm. to PRL Inclusive measurement only g on signal side reconstructed; two samples: tagged/untagged; similar sensitivity, largely stat. independent, profit from combination g l± 1.26 GeV < E < 2.20 GeV Bsig Xs Btag Fischer discriminant based on energy flow and event-shape for continuum supp. off on Background treatment subtract luminosity scaled Eg distribution for off-data from on-data (continuum bkg.); a: lumin. ratio, s dependence, ~8.8 Ce: efficiency; FE: lower Eg in off-data

Measurement Background treatment explicit veto p0, h → gg; Belle, arXiv:0907.1384, 605 fb-1, subm. to PRL after vetoing p0, h → gg Background treatment explicit veto p0, h → gg; individual remaining bkg. categories taken from MC; shape and yield corrected by data control samples; - inclusive B→ p0X, hX samples reconstructed in data (off- data subtraction) and MC; 5%-10% correction to MC yield; - veto eff. measured by partialy reconstructed D0 K-p+p0; - timing info for EM calorim. clusters (overlapping evts.: hadronic + Bhabha); - beam background from random trigger events; signal p0, h → gg other B decays 65% of remaining bkg. from p0, h → gg

Measurement corrected spectra raw spectra Belle, arXiv:0907.1384, 605 fb-1, subm. to PRL un-tagged raw spectra each bkg. (scaled by control samples) subtracted; corrected spectra correct for select. eff. (euntag  15%, etag ~ 2.5%); average tagged and un-tagged spectra taking stat. correlation into account (r  0.1); control region: -1629±4894 1245±4349 tagged average -745±889 292±410

Results Branching fraction unfolding Egmeas → Egtrue Egmin Belle, arXiv:0907.1384, 605 fb-1, subm. to PRL corrected spectra unfolding Egmeas → Egtrue calibrated using radiative di-muon evts (average Eg resolution 2%); correct for EM cluster detection eff.; correct for B → Xdg (Rd/s=4.5% ± 0.3%) boost to B meson rest frame Branching fraction main syst. uncertainties: from B background other than p0, h (varied by 20%, using various models); correction factors in off-data subtraction;

Results Moments moments (and Br’s) given for various Egmin Belle, arXiv:0907.1384, 605 fb-1, subm. to PRL Moments moments (and Br’s) given for various Egmin J.R. Andersen, E. Gardi, JHEP 0701, 029 (2007) (‘05) MB/2- Eg [GeV] Eg2 - Eg2 [GeV2] this meas. this meas. (‘05) Egmin [GeV] Egmin [GeV] @ Egmin =1.7 GeV

Constraints MH±  300 GeV @95% C.L. constraints on type-II 2HDM M. Misiak et al., PRL98, 022002 (2007) s(Br) [10-4] constraints on type-II 2HDM MH±  300 GeV @95% C.L. Br [10-4] constraints on Wilson coeff., MFV using Br(B →Xsg), Br(B →Xsll;q2), Br(K →pnn ), Br(Bs →mm), AFB(B →K*ll ;q2) ; using Br(B →Xsg) only T. Hurth et al., Nucl.Phys. B808, 326 (2009)

Summary lowest Egmin (=1.7 GeV); most precise measurement up to date; measured Br and dG/dEg of inclusive B  Xs g decays; lowest Egmin (=1.7 GeV); most precise measurement up to date; constraints on NP models and QCD parameters used in Vcb, Vub determination;

Detector more EM calorimeter CsI (16X0)

b→ cln more relation b→ cln /b→ sg in OPE Benson et al., Nucl.Phys. B665, 367 (2003)

b→ cln more relation b→ cln /b→ sg in OPE Benson et al., Nucl.Phys. B710, 371 (2005)

b→ cln more relation b→ cln /b→ sg in OPE fraction of I. Bigi, N. Uraltsev, Int.J.Mod.Phys. A17, 4709 (2002) fraction of b uln with MX < M high Eg: weight function f(Eg) small, dG/dEg spectrum less important; Eg  2 GeV: weight function f(Eg) ~ 1, importat to measure dG/dEg precisey to low Eg M=1.6 GeV M=1.7 GeV M=1.8 GeV

b→ sg OPE more b→ sg in OPE in b→ sg cut on Eg>Ecut used experimentally; this limits the theoretical accuracy of inclusive B decays treatment; Ecut  new scale D=MB-2Ecut; some non-perturbative effects described as series in 1/D; if Ecut high  questionable series expansion; biases for Ecut > 1.85 GeV, breakdown for Ecut > 2.1 GeV; further discussion in D. Benson et al., Nucl.Phys. B710, 371 (2005) Z. Ligeti et al., PRD78, 114014 (2008)

b→ sg Off-data scaling more Belle, arXiv:0804.1580,605 fb-1 bkg. treatment subtract lumin. scaled off-data from on-data (continuum bkg.); Lon = 605 fb-1 Loff = 68 fb-1 off on on-data scaled off-data subtracted a: lumin. ratio, including s dependence, ~8.8 ehadronic,B→XsgON,OFF: efficiency of hadronic and signal selection; FN,E: corr. factor due to lower mean Eg and multiplicity in off-data

b→ sg Moments more 1±0.12 1±0.06 1±0.023 1±0.008 1±0.48 1±0.12 Belle, arXiv:0804.1580,605 fb-1 1±0.12 1±0.06 1±0.023 1±0.008 1±0.48 1±0.12

b→ sg extrapol. more extrapol. factors to Eg=1.6 GeV used by HFAG: E. Barberio et al. (HFAG), arXiv:0808.1297

b→ sg tag/untag more MC study: same FOM (Nsig/dNsig) for tagged/un-tagged sample; data un-tagged sample: tagged: combined:

b→ sg MC sum of exclusive K*g and inclusive; inclusive: Xs BW J=1 resonance, m=2.4 GeV, G=1.5 GeV hadronized by JETSET and reweighted according to Kagan-Neubert Dressed Gluon Exponentiation Bosch-Lange-Neubert-Paz models Gambino-Giordano Benson-Bigi-Uraltsev

b→ s g decays outlook more for Egmin=1.6 GeV (comparison to theory) few x more data needed 95% C.L. lower limit on m(H±), all tanb Belle, 50 ab-1 + existing meas. s(Br) Belle, 0.6 ab-1 M. Misiak et al., PRL98, 022002 (2007) HFAG, ICHEP08 L.L. can get as high as 500 GeV Belle, 50 ab-1 + existing meas. Br

b → sg CPV more t-dependent CPV in SM: helicity structure of effective Hamiltonian bR →sLgL  mb (since W in loop couples to bL spin flip required) or bL →sRgR  ms interference mixing/no-mixing for bR →sLgL  mb bR → bR →sLgL  ms CPV in SM  ms/mb appropriate modes: K*0(Ksp0)g, Kshg, Ksfg (ang. analysis necessary), ... NP with heavy right-handed fermions in loop can enhance CPV; D. Atwood et al., PRL79, 185 (1997)

b → sg CPV more t-dependent CPV in SM: SCP(K*g)  (2ms/mb)sin2f1  0.04 Left-Right Symmetric Model: SCP(K*g)  0.67 sin2f1  0.5 SCP(Ksp0g) = -0.10 ±0.31 ±0.07 ACP(Ksp0g) = -0.20 ±0.20 ±0.06 for m(Ksp0) < 1.8 GeV (mainly K*g) largest syst. from signal fract. and resol. f. important additional improvements with upgraded SVD (Ks - IP vtx) D. Atwood et al., PRL79, 185 (1997) Belle, PRD74, 111104 (2006), 535M BB Belle, PRL101, 251601(2008), 657M BB similar sensitivity for Ksr0g (dilution from K*pg)

b → sg CPV more t-dependent CPV DCPV expectation: main syst. scales with luminosity s(S(Ksp0g))=0.09 @5 ab-1 0.03 @50 ab-1 (~SM value) +20% increase in Ks acceptance with SVD DCPV suppressed by |VubVus*/VtbVts*|, as(mb) (strong phase), (mc/mb)2 (GIM); OPE: ACP(B→Xsg) = (0.44±0.240.14)% semi-inclusive analysis: K+(1-4)p; KKK(p), KSKK(p); ACP(B→Xsg;MXs<2.1 GeV) = (0.2±5.0±3.0)% syst.: bias (detector charge asymmetry), possible bkg. asymmetry, uncertainty of M(Xs) shape Dp control sample measured asymmetries non-scaling in other decay modes T. Hurth et al., Nucl.Phys. B704, 56 (2005) Belle, PRL93, 031803 (2004), 140 fb-1

e for Ks reconst. Dz resolution b → sg SVD more to scale r [cm] SVD Belle SVD SuperBelle (if two SVD layers/p required) z [cm] sBelle Design Group, KEK Report 2008-7 e for Ks reconst. Dz resolution in B →K*0g

Constraints on Wilson coeff. more constraints on NP in Ci Br(B →Xsg), Br(B →Xsll), Br(K →pnn ) Br(Bs →mm), no Br(B →K*ll ) (large th. uncertainty) dC9 dC10 J. Kamenik, arXiv:0805.2363 dC7 dC9