1. Beauty and charm results from B factories
Boštjan Golob
University of Ljubljana, Jožef Stefan Institute & Belle Collaboration
Helmholtz International Summer School “Heavy Quark Physics”
Bogoliubov Laboratory of Theoretical Physics, Dubna, Russia, August 11-21, 2008
JINR
University
of Ljubljana
“Jožef Stefan”
Institute

2. Outline Beauty Lecture 1 leptonic Lecture 2 semileptonic
Introduction
B Oscillations
(Mostly) rare B decays
leptonic
semileptonic
b →sg
b →sll
Lecture 2
Charm and others
4. D0 mixing and CPV
decays to CP states
WS decays
t-dependent Dalitz
5. Ds leptonic decays
6. Spectroscopy
exotic states
exp. results
with some comments
on phenomenology
It is a curious fact that
people are never so
trivial as when they take
themselves seriously.
O. Wilde ( )
Part of B-factories lectures with A.J. Bevan;
division by topics, not by experiments

3. Introduction Experiments continuum production g* c
very diverse exp. conditions
We all live with the
objective of being happy;
our lives are all different
and yet the same.
on resonance production
e+e- → U(4S) → B0B0, B+B-
s(BB)  1.1 nb (~0.9x109 BB pairs)
Anne Frank ( )
continuum production g* c
s(c c)  1.3 nb (~109 XcYc pairs)
e+e- → y(3770) → D0D0, D+D-
(coherent C=-1 state);
~800 pb-1 of data
available at y(3770);
2.8x106 D0D0
3.5 fb-1 on tape
s(D0; pt>5.5 GeV,|y|<1)≈
≈13 mb
50x109 D0’s

4. B oscillations Time evolutionq1 q2 P0 P0 q2 q1
Time evolution
(also lectures by U. Nierste, A. Pivovarov)
flavour states ≠ Heff eigenstates:
(defined flavour) (defined m1,2 and G1,2)
P0 = K0, Bd0, Bs0 and D0
eigenvalues:
diagonal elem.: P0  P0
(including decays)
non-diagonal elem.: P0  P0
more

5. B oscillations Time evolution
D. Kirkby, Y. Nir, CPV in Meson Decays, in RPP
Time evolution
P1,2 evolve in time according to m1,2 and G1,2:
|P0(t)>, |P0(t)>
decay rates:
for easier notation:
Gt → t
U(4S) →B0B0: B meson pair
in quantum coherent state;
before 1st B decay: B0B0
1st B decay:
tag B0/B0;
mixing clock start, t →Dt
Decay time distribution of experimentally accessible states P0, P0
sensitive to mixing parameters x and y, depending on final state
more

6. B oscillations Time evolution more difficult to observe
probab. to observe an initially produced X0 as X0 after time t
~ Bs0
~ D0
more difficult to observe
oscillations within t
visually unobservable
deviation from pure
exponential

7. B oscillations Method l- similar to CPV, reconstruct
signal
B oscillations
Method
similar to CPV, reconstruct
flavor specific final states
signal
B0 or B0 m+ m-
J/y
fully reconstruct decay
to flavor specific final state p-
Bsig
K*0 K+
tag flavor
of other B
from
charges
of typical
decay
products
U(4S) l- K-
Btag
Dt=Dz/bgc
determined
B0(B0)
determine time between
decays

8. B oscillations Method Method reconstructed flavour specific decays
Belle, PRD71, (2005), 140 fb-1
DE signal region
Method
reconstructed
flavour specific
decays
measure
Dt distribution
Method
Dt distribution
Af=0, |y|<<1
more
w: wrong tag probability (reduces ampl. of oscillations)
R(Dt): resolution function
- intrinsic detector resolution on position of
both B vertices
- smearing due to non-primary tracks
- smearing due to B meson CMS momentum
saver(Dt)=1.43 ps
more

9. B oscillations Results flavour asymmetry Dmd=(0.511±0.005±0.006) ps-1
Belle, PRD71, (2005), 140 fb-1
Results
flavour asymmetry
Dmd=(0.511±0.005±0.006) ps-1
largest syst.: D** bkg.
Dmd=(0.507±0.005) ps-1
x=DmdtBd= 0.776±0.008
HFAG,

10. B oscillations Phenomenology P0-P0 transition →
P0: any pseudo-scalar meson;
specific example of Bd0
Phenomenology
(see also lectures by U. Nierste)
P0-P0 transition →
box diagram at quark level d b
u, c, t W+ W- B0
Vid
Vjd
Vjb*
Vib* d b
u, c, t W+ W- B0
if mi = mj  due to CKM unitarity:
no mixing
loop int., CKM unitarity 
considering CKM values and
q masses: largest contribution from t quark
more

11. B oscillations Phenomenology calculate M12, G12 from
A.J. Buras et al., Nucl.Phys.B245, 369 (1984)
Phenomenology
calculate M12, G12 from
box diagram; from that
calculate Dm, DG
must be calculated
to determine Vij;
theor. uncertainty
(LQCD)
q: d (Bd) or s (Bs);
and Dms also measured...
BBq: bag parameter, <Bq0|bgm(1-g5)q|Bq0>
fBq: decay constant
hB(‘): QCD corr. O(1)
S0(xt): known kinematic function
reduced theor.
uncertainty in ratio x2
M. Okamoto, hep-lat/

12. B oscillations Bs Dms /Dmd A amplitude method: instead of Dms
CDF, PRL97, (2006) A Bs
amplitude method: instead of Dms
fit A at different values of Dms;
A=1  oscillations at this Dms value
Dms=(17.77±0.10±0.07) ps-1
x=DmstBs= 25.5±0.6
Dms /Dmd
uncertainties on (r2+h2):
Dmd constraint ±13%
Dmd ±1%
fBdBBd ±12%
Dms /Dmd constraint ±6%
Dms /Dmd ±1.5%
x ±5%

13. Leptonic B decays l+ B → tn Method Q fP (H+) W+
G(B+ → t+n): G(B+ → m+n): G(B+ → e+n)=
1:4x10-3:10-7
fP → meas. VQq;
H±; q
VQq n
Method
fully reconstruct Btag in hadronic decays
(K+p-p+p-p+);
search for 1/3 tracks from Bsig→tn
(e-);
no additional energy in EM calorim.
(from p0, g, ...);
signal at EECL~0
EM calorim.
B → tn
candidate
event

14. Leptonic B decays Results BaBar: hadronic decays for Btag;
Belle, PRL97, (2006), 414 fb-1
Belle, ICHEP08, 600 fb-1
Results
largest syst. from signal and bkg. shape
semileptonic tag added
BaBar:
hadronic decays for Btag;
combined with semil. decays:
bkg.
Nsig=17 ± 5
3.5 s signif.
(-2lnL0/Lmax)
signal
BaBar, PRD77, (2008), 346 fb-1
BaBar, PRD76,
(2007), 346 fb-1
expected signal
Br=3x10-3
HFAG,
more

15. Leptonic B decays t Phenomenology b H+ u n
using fB=(216 ± 22) MeV, |Vub|=(3.9 ± 0.5)x10-3, tB
 BrSM(B+ →tn) = (1.25 ± 0.41)x10-4
new physics:
to make predictions/measure |Vub| →
fB (from LQCD) needed;
validate LQCD in charm sector
(better exp. precision) →
to be addressed later;
established method for decays with large Emiss;
to be exploited at SuperB (B→Knn, dark matter)
HPQCD, PRL95, (2005) b u B- t n H+
SuperB
50 ab-1
more

16. Semileptonic B decays l+ P →Pln P →Vln H± exchange measurementq1 q3 n l+ M1 q2 M2
skip
W±, H±
P →Pln
in G suppressed by ml2/mM12
negligible for e,m; not for t
P →Vln
3 form f. for e,m; 4 for t
HQS: relations among f.f.’s;
can be tested; for suppressed f.f.’s only by t
more
H± exchange
modified SM Br’s for t;
in P→ V only helicity=0 V possible
measurement
challenging due to multiple n’s;
more

17. Semileptonic B decays B0 →D*-t+nt method: D* reconstruction;
Bsig
e/p n
Btag
B0 →D*-t+nt
method:
D* reconstruction;
t →enn, pn
Bsig: D* and e/p
Btag: rest of event
control sample:
Bsig →D*p , check Btag reconstruction
signal sample:
requirements on Xmis, Evis
excl. Btag reconstruction
t →enn, mnn
Bsig: D/D* and e/m
mmis2=pmis2 t
Belle, PRL99, (2007), 480 fb-1
Bsig →D*p MC
data
BaBar, PRL100, (2008), 209 fb-1
related to missing mass (>0 for several n);
Evis < m(U(4S))
missing mass (>0 for several n);

18. Semileptonic B decays B0 →D*-t+nt results bkg. from B0 →D*en (peaking)
Belle, PRL99, (2007), 480 fb-1
B0 →D*-t+nt
results
bkg. from B0 →D*en (peaking)
t →rn
Nsig=60 ±12
6.7 s signif.
(-2lnL0/Lmax)
main systematics:
from signal and bkg shape (MC)
Btag reconstr. eff. (control sample)
BaBar, PRL100, (2008), 209 fb-1
D*-l+n
D*-t+nt
last uncertainty: normaliz. modes (Dln , D*ln)
main systematics:
from signal and bkg shape (MC)
D** contrib.
D-l+n
D-t+nt

19. Semileptonic B decays B →D(*)tn phenomenology limits on H±;
M. Tanaka, Z.Phys.C67, 321 (1995)
B →D(*)tn phenomenology
limits on H±;
inclusive B →Xctn predicted Br:
(2.30 ±0.25)%
sum of D*tn, Dtn:
(2.59 ±0.39)%
G(B →D*long.tn)
G(B →D*mn)|SM
Ba/lle average
(assuming no correl.
and 100% long. polariz.)
A.F.Falk et al., PLB326, 145 (1994)
BaBar
more

20. b → sg Motivation Difficulties FCNC process; theory:W±
u, c, t
Vqb
Vqs g b s
u, c, t X Y H± g
Motivation
FCNC process;
sensitive to NP in loop;
parton level: Eg ≈ mb/2;
determ. of mb, Fermi motion →
needed for Vub determ. from
inclusive semil. B decays;
Difficulties
theory:
parameter extraction from
partial Br(Eg>Ecut) →
extrapolation needed;
experiment:
measure low Eg
huge bkg. c± b s X Y g
more
signal
continuum p0
Your background and
environment is with
you for life. No question
about that.
S. Connery (1930)

21. b → sg Inclusive measurement (see also lectures by U. Heisch)
off on
Inclusive measurement
(see also lectures by U. Heisch)
only g reconstructed;
bkg. treatment
subtract lumin. scaled off-data
from on-data (continuum bkg.);
veto p0, h → gg;
rest bkg. from MC (control samples);
timing info for EM calorim. clusters
(overlapping evts.: hadronic + Bhabha)
inclusive B→ p0X, hX samples
reconstructed in data
(off- data subtraction) and MC;
5%-10% correction to MC bkg. normaliz. on
scaled off
Belle,
arXiv: ,
605 fb-1
subtracted
more
80% of
remaining bkg.
from p0, h → gg
after vetoing
p0, h → gg

22. b → sg Inclusive measurement Eg spectrum Br(B →Xsg)
Belle, arXiv: ,605 fb-1
consistent
with 0 above
B decay
threshold
Inclusive measurement
Eg spectrum
Br(B →Xsg)
deconvolution of Eg
(Egmeas → Egtrue; using
radiative di-muon evts);
boost to B rest frame;
b →dg contrib. (4%);
mb1S/2~2.3 GeV
last uncertainty due to boost;
largest system.: corr. factors in off-data subtraction;
bkg. g’s from B (other than p0, h)

23. b → sg Seminclusive measurement B reconstructed;
BaBar, PRD72, , 82 fb-1
Seminclusive measurement
B reconstructed;
(see also lectures by B. Pecjak)
sum of exclusive decay modes
Xs: no S-wave states in B→Xsg
22 final states K-(0)+(1-4)p
K-(0)+h+(0-2)p
K-(0)+(0-1)p
g + Xs  B
(better resol.)
bgk.: p0, h veto, NN from topological
variables for continuum;
not all final states reconstructed
→corr. for missing fraction
(from MC, checked with data in various
final state categories)
peaking bkg.: missing final states
reconstructed as one of signal decays;
signal decays with some particles
exchanged with other B
25% at low
M(Xs) from KL
at high M(Xs)
from K+ 5p

24. b → sg Seminclusive measurement fit in bins of M(Xs) Br(M(Xs));
BaBar, PRD72, , 82 fb-1
Seminclusive measurement
fit in bins of M(Xs) Br(M(Xs));
Eg spectrum (Eg>1.9 GeV);
moments of dG/dEg also determined;
mb (and other QCD parameters)
determined for use in b →uln;
e.g.
main systematics: from missing final states
K*(892)
more
more details at
HFAG,

25. b → sg Phenomenology average of results: comparison with limits from
M. Misiak et al., PRL98, (2007)
Phenomenology
average of results:
comparison with limits from
B →tn:
HFAG, winter 08,
95% C.L. lower
limit on m(H±),
all tanb
first error: stat.+syst.
second error: Eg spectrum (extrapol.)
m(H±)=300 GeV
Belle, PRL97,
(2006), 414 fb-1
For my part I know nothing with any certainty,
but the sight of the stars makes me dream.
V. van Gogh ( )

26. b → sll Motivation FCNC process; (see also lectures by E. Lunghi)
M expressed in terms C7,9,10;
Wilson coeff.’s
NP modifies C7,9,10 or/and adds
new operators
Wilson coeff.’s independent of
final state (C7 same for b→ sg and b→ sll);
|C7 |2 constrained by Br(B→ Xsg);
sign not known;
b→ sll: interference of amplitudes
additional information (also sign)
on C7,9,10 b s W±
u, c, t
Vqb
Vqs g
=VqbV*qs C7 x =
perturbative
(dependence
on MW, mt, MNP) non-perturbative
more

27. b → sll l+l- l+ l- exclusive B →K*ll qK distrib.  fractionp K* K qK
exclusive B →K*ll
qK distrib.  fraction
of long. polarized K* (FL);
ql distrib.  lepton
forward-backward asymmetry
(AFB);
prediction for AFB:
q2=m2(l+l-) B l- K* ql l+
veto
veto
low q2
high q2 SM
C7 = -C7SM
C9 C10 = -C9SM C10SM
C7 = -C7SM
C9 C10 = -C9SM C10SM q2

28. b → sll reconstruction e+e-, m+m-; K* →Kp, Kp0, Ksp; Mbc fit
BaBar,
arXiv: ,
350 fb-1
low q2
high q2
Ns=27.2 ±6.3
reconstruction
e+e-, m+m-; K* →Kp, Kp0, Ksp;
Mbc fit
combinatorial bkg.: e+m-;
misid. hadrons: h+m-;
peaking bkg.: D(→K*p)p
(mm sample only,
veto on m(K*p));
signal fraction
qK fit
FL free parameter;
ql fit
AFB free parameter;
Ns=36.6 ±9.6

29. b → sll results FL; consistent with SM and -C7SM; AFB;q2
average over interval SM
results
FL;
consistent with SM and -C7SM;
AFB;
-C9SM C10SM disfavored (>3 s);
stronger constraints;
C7 = -C7SM
BaBar, arXiv: , 350 fb-1
Belle, PRL96, (2006), 357 fb-1
Belle, ICHEP08, 600 fb-1 SM
C7 = -C7SM
C9 C10 = -C9SM C10SM
C7 = -C7SM
C9 C10 = -C9SM C10SM
more

30. b → sll semi-inclusive similar as b →sg; e+e-, m+m-; K-/Ks+(0-4)p;
Nsig=68 ±14
5.4 s signif.
(-2lnL0/Lmax)
Belle PRD72,
(2005), 140 fb-1
semi-inclusive
similar as b →sg;
e+e-, m+m-; K-/Ks+(0-4)p;
~30% missing modes;
charmonium sample provides
cross-check of bkg.;
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)
Belle PRD72, (2005), 140 fb-1
BaBar PRL93, (2004), 82 fb-1
dC9
dC10
J. Kamenik, arXiv:
dC7
dC9

31. B oscillations more Time evolution
D. Kirkby, Y. Nir, CPV in Meson Decays, in RPP
Time evolution
state initially produced as superposition
(n.b.: a(0)/b(0) can be 0)
will evolve in time as
if interested in a(t), b(t):
effective Hamiltonian and
t-dependent Schrödinger eq.:
eigenstates:
(well defined m1,2 and G1,2)
back

32. B oscillations more Time evolution eigenvalues:
diagonal elem.: P0  P0 (including decays)
non-diagonal elem.: P0  P0
P1,2 evolve in time according to m1,2 and G1,2:

33. B oscillations more
Time evolution
eigenvalues:

34. B oscillations more
Time evolution
eigenvalues:
back

35. B oscillations more Time evolution taking into account
we arrive at time evolution of P0, P0:
back

36. B oscillations more Time evolution decay rates:
for CP conjugated states: Af → Af, Af → Af

37. B oscillations more CPV |p/q|=1, y<<1  (well fulfilled for Bd)
|Af/Af|≠1
CPV in decay
|q/p| ≠1
CPV in mixing
I(lf) ≠ 0
CPV in interf.
back

38. B oscillations more Method reconstructed flavour specific decays; D*ln
Belle, PRD71, (2005), 140 fb-1
Method
reconstructed
flavour specific
decays;
D*ln
= known meas known meas known meas. meas.
total bkg
D** bkg.
back

39. B oscillations more Method tagging q=+(-)1 B0(B0) r: tag quality
H. Kakuno et al., NIM A533, 516 (2004)
Method
tagging
q=+(-)1
B0(B0)
r: tag quality

40. B oscillations more Method tagging single r bin: two r bins:
H. Kakuno et al., NIM A533, 516 (2004)
Method
tagging
single r bin:
two r bins:
back

41. B oscillations more Method resolution function
H. Tajima et al., NIM A533, 370 (2004)
Method
resolution function
Rful: vtx of fully reconstructed B meson
Rasc: vtx of tagging B meson
Rnp: non-primary tracks
Rk: kinematic smearing
back

42. B oscillations more Phenomenology P0-P0 transition →
back
P0: any pseudo-scalar meson;
specific example of Bd0
Phenomenology
P0-P0 transition →
box diagram at quark level d b
u, c, t W+ W- B0
Vid
Vjd
Vjb*
Vib* d b
u, c, t W+ W- B0
if mi = mj  due to CKM unitarity:
no mixing
simplified treatment based on dimension:
O. Nachtmann, Elem. Part. Phys., Springer-Verlag
for serious treatment see e.g.:
A.J. Buras et al., Nucl.Phys.B245, 369 (1984)

43. Leptonic B decays more Systematic checks Bsig decay modes
Belle, PRL97, (2006), 414 fb-1
Systematic checks
Bsig decay modes
check of EECL,
double tagged decays,
Bsig- →D*0 l- n, D*0 →D0p0
back

44. Leptonic B decays more Phenomenology additional Higgs doublet;
Type II Two Higgs
Doublets Models
(f1 gives masses to d-type
and charged l; f2 gives
masses to u-type; in Type I
models f1 is decoupled and
f2 generates all masses)
Phenomenology
additional Higgs doublet;
tanb=v1/v2, ratio of vacuum
expectation values;
H± coupling  ml  same factor as
helicity SM suppression
ratio independent of
H ± contribution:
W.S.Hou, PRD48, 2342 (1993)
if Gmeas>GSM  H± contribution dominant
back

45. Semileptonic B decays more
Form factors
P→P:
B(v) → B(v’):
for mb →  amplitude
can only depend on
g = v·v’; for v = v’ nothing
happens, z(1)=1;
B(v) → D(v’):
for mb, mc →  same (HQS)
z(v·v’): Isgur-Wise function
relates two in principle independent form factors for P → P transition
back

46. Semileptonic B decays more
Form factors
P→V: q2
one more f.f. if ml not small;
HQS: relations among f.f.’s for P→ P and P →V
back

47. Semileptonic B decays more
M. Tanaka, Z.Phys.C67, 321 (1995)
B →D*tn phenomenology
amplitude for W exchange:
lM=±,0; lt=±; lW=±,0;
D*, t, W helicity
amplitude for H±
exchange:
relation among H ±, W
exchange amplitudes:
H ± : no contribution of
transversely polarized
D* (HR,L±=0)
back

48. Semileptonic B decays more
U. Nierste et al., PRD78, (2008)
B →Dtn phenomenology
update of predictions:
measurement
BaBar, PRL100, (2008), 209 fb-1
mB2/mH2 tan2b (in 2HDM-II)
back

49. b → sg more inclusive semil. B decays semil. width: Operator Product
Expansion to O(1/mb2):
two parameters, l1, l2:
average p2 of b in B
hyperfine
interaction
back

50. b → sg more inclusive semil. B decays Fermi motion: new parameter L
same parameters
governing moments
of various distributions,
e.g. mass of hadronic system
in semil. decays:
or Eg moments in b →s g:
A.F.Falk, M.E.Luke,
PRD57, 424 (1998)
A.Kapustin, Z. Ligeti
PLB355, 318 (1995)
back

51. b → sg more off-data subtraction a: lumin. ratio;
ehadronic,B→XsgON,OFF:
efficiency of hadronic, signal
selection;
FN,E: corr. factor due to lower
mean E and multiplicity in off-data
back

52. b → sg more Eg resolution inclusive meas.: Eg measured in EM calorim.;
s(Eg;Eg=2 GeV) ~ 20 MeV;
semi-inclusive meas.:
Eg from
s(Eg) ~ 1-5 MeV;
back

53. b → sll more OPE, Wilson coeff. example of b →cdu
BaBar Physics Book, SLAC-R-504 W b c u
OPE, Wilson coeff.
example of b →cdu
almost point-like inter.:
series:
product of currents
expressed as series
of local operators (OPE);
such expansion valid if q2/MW2<<1;
in this range an effective theory
can be constructed, valid up to a
cut-off, in the above case up to MW;
back

54. b → sll more OPE, Wilson coeff. example of b →cdug d
BaBar Physics Book, SLAC-R-504 W b c u
OPE, Wilson coeff.
example of b →cdu
rad. corr. to lowest order:
operators
receive radiation corr. and must be
renormalized; they become dependent
on renormalization scale m;
physics must be independent of m 
operators receive m dependent coefficients
in order for Heff to satisfy:
(i,j: color indices)
back

55. b → sll more OPE, Wilson coeff. example of b →cdug d
BaBar Physics Book, SLAC-R-504 W b c u
OPE, Wilson coeff.
example of b →cdu
under renormaliz. set
of operators can be enlarged,
for the example under
consideration there is also
Heff is thus
Ci(m) are Wilson coeff., containing
information on short distance physics
down to (arbitrary) scale m; all heavy
masses (M>>m) dependence (mt, MW,
MNP) is contained in Ci(m)
(changed color indices)
back

56. b → sll more OPE, Wilson coeff. once Oi(m) dependence is
calculated, Ci(m) follow from
for b →cdu
division of energy scales between
Ci(m) and local operators
<f|Oi(m)|B> can be
schematically viewed as
Wilson coeff. Ci(m) are independent
of external states (f)
<f|Oi(m)|B> Ci(m)
back

57. b → sll more OPE, Wilson coeff. Br(b → sll ): AFB, RPV SUSY contrib.:
P. Gambino et al., PRL94, (2005)
OPE, Wilson coeff.
Br(b → sll ):
AFB, RPV SUSY contrib.:
Y.-G. Xu et al., PRD74, (2006)
|l1i3’l1i2’*|<4.7x10-5
back

58. b → sll more SM C7 = -C7SM C9 C10 = -C9SM C10SM C7 = -C7SM
Belle, PRL96, (2006), 357 fb-1 SM
C7 = -C7SM
C9 C10 = -C9SM C10SM
C7 = -C7SM
C9 C10 = -C9SM C10SM
back