The Precision Determination of |V | and |Vub
Transcript of The Precision Determination of |V | and |Vub
The Precision Determination of |Vcb| and |Vub|
Phillip UrquijoExperimental Particle Physics GroupThe University of Melbourne
EPFL LPHE Seminar Oct 2007
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 2
Why 3 sets (= generations) of particles? How do they differ? How do they interact with each other?
strong E&M weak
g γW ±
Z 0
µ− cνµ sτ − tντ b
leptons quarkse− uνe d
1st generation
2nd generation
3rd generation
It turns out there are two “extra”
copies of particles
Simplified Standard Model
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
Strong force is transmitted by the gluon
Electromagnetic force by the photon
Weak force by the W± and Z0 bosons
3
Simplified Standard Model
uu
g
dd
g
uu
γ
dd
γ
e−
γ
e−
uu
Z0
dd
Z0 Z0
e−e−
Z0
νeνe
du
W+
νee−
W−
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
Strong force is transmitted by the gluon
Electromagnetic force by the photon
Weak force by the W± and Z0 bosons
3
Simplified Standard Model
uu
g
dd
g
uu
γ
dd
γ
e−
γ
e−
uu
Z0
dd
Z0 Z0
e−e−
Z0
νeνe
du
W+
νee−
W−Note W± can “convert”
u ↔ d, e ↔ ν
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 4
Mass and the Generations
Fermions They differ only by the masses The Standard Model has no explanation
for the mass spectrum The masses come from the interaction
with the Higgs field ... whose nature is unknown
Par
ticle
mas
s (e
V/c
2 )
Q = −1 0 +2/3 −1/3
The origin of mass is one of the most urgent questions in particle physics today
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
Masses uniquely define the u, c, t, and d, s, b states We don’t know what creates masses
We don’t know how the weak eigenstates are chosen
couples quark charged currents to W± Mixes (qj=d,s,b) quark mass eigenstates to give weak eigenstates The Standard Model does not predict V
5
Weak decays
6
Cabibbo-Kobayashi-Maskawa matrix
matrix elements are fundamental parameters of the Standard Model, must be determined experimentally
quark mixing in the SM
d′
s′
b′
=
Vud Vus Vub
Vcd Vcs Vcb
Vtd Vts Vtb
dsb
CKM Workshop Nagoya 2006C K M
増川小林
6
quark mixing in the SM
d′
s′
b′
=
Vud Vus Vub
Vcd Vcs Vcb
Vtd Vts Vtb
dsb
€
1− λ2
2λ Aλ3(ρ− iη)
−λ 1− λ2
2Aλ2
Aλ3(1−ρ− iη) −Aλ2 1
+ O(λ4)Vij=
Wolfenstein ParameterisationCKM Workshop Nagoya 2006
One irreducible complex phase → CP violationThe only source of CP violation in the minimal Standard Model
C K M増川小林
6
quark mixing in the SM
d′
s′
b′
=
Vud Vus Vub
Vcd Vcs Vcb
Vtd Vts Vtb
dsb
€
1− λ2
2λ Aλ3(ρ− iη)
−λ 1− λ2
2Aλ2
Aλ3(1−ρ− iη) −Aλ2 1
+ O(λ4)Vij=
Wolfenstein ParameterisationCKM Workshop Nagoya 2006
One irreducible complex phase → CP violationThe only source of CP violation in the minimal Standard Model
C K M増川小林
Kobayashi and Maskawa winners of EPS High Energy and Particle Physics Prize, 2007
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 7
CP violation and New Physics
Are there additional (non-CKM) sources of CP violation?
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 7
CP violation and New Physics
The CKM mechanism fails to explain the amount of matter-antimatter imbalance in the Universe ... by several orders of magnitude
New Physics beyond the SM is expected at 1-10 TeV scale e.g. to keep the Higgs mass < 1 TeV/c2
Almost all theories of New Physics introduce new sources of CP violation
New physics effects in quarkflavour mixing can only appear as small corrections to the leading CKM mechanism
New sources of CP violation almost certainly exist
Are there additional (non-CKM) sources of CP violation?
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 8
The Unitarity Triangle
V†V = 1 gives us
Measurements of angles and sides constrain the apex (ρ, η)
This one has the 3 terms in the same order of magnitude
A triangle on the complex plane
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 8
The Unitarity Triangle
V†V = 1 gives us
Measurements of angles and sides constrain the apex (ρ, η)
This one has the 3 terms in the same order of magnitude
A triangle on the complex plane
Φ1
Φ2
Φ3
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 9
What can we measure in B decays?
φ1φ3
φ2
CPV in SM ∝ Area
€
VudVub*
€
VtdVtb*
€
VcdVcb*
€
VtbVts* +VcbVcs
* +VubVus* = 0
CP asymmetries
φ1φ3
φ2
€
(ρ,η)
€
(0,0)
€
(1,0)Measurements often plotted in (ρ,η) plane€
≈Vub
*
λVcb*
€
≈Vtd
λVcb*
€
φ3 ≈ argVub*,φ1 ≈ −argVtd ,φ2 ≈ π −φ1 −φ3
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 9
What can we measure in B decays?
φ1φ3
φ2
CPV in SM ∝ Area
€
VudVub*
€
VtdVtb*
€
VcdVcb*
€
VtbVts* +VcbVcs
* +VubVus* = 0
CP asymmetries
φ1φ3
φ2
€
(ρ,η)
€
(0,0)
€
(1,0)Measurements often plotted in (ρ,η) plane€
≈Vub
*
λVcb*
€
≈Vtd
λVcb*
€
φ3 ≈ argVub*,φ1 ≈ −argVtd ,φ2 ≈ π −φ1 −φ3
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 10
The sides
To measure the lengths of thetwo sides, we must measure|Vub| ≈ 0.004 and |Vtd| ≈ 0.008 The smallest elements – not easy!
Main difficulty: Controlling theoretical errors due to hadronic physics
γ
α
β
Vtd
Vub
Vcb
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 11
€
s =10.58GeV,Ee+ = 3.5GeV,Ee− = 8.0GeV
Υ(4S)→B/B(96%)
Data to date: >700 million B/B
Asymmetric e+e- collider running at the Υ(4S) resonance
KEK-B
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 11
€
s =10.58GeV,Ee+ = 3.5GeV,Ee− = 8.0GeV
Υ(4S)→B/B(96%)
Data to date: >700 million B/B
Asymmetric e+e- collider running at the Υ(4S) resonance
KEK-B
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 11
€
s =10.58GeV,Ee+ = 3.5GeV,Ee− = 8.0GeV
Υ(4S)→B/B(96%)
Data to date: >700 million B/B
Asymmetric e+e- collider running at the Υ(4S) resonance
KEK-B
Crab cavity commissioning
• Current precision
• sin2φ1 ~ 4%
• |Vcb| ~ 1-2%
• |Vub| ~ 8%
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 12
Significance of |Vub| and |Vcb| Measurement
∝|Vub|/|Vcb|
• The precise determination of |Vub|/|Vcb| provides a benchmark for testing new physics in other processes
• Precision of sin2β outstrips the other measurements
• We must make the green ring thinner!
• Large new physics contributions to penguins would have already been seen.
• New physics contributions to decays such as B → τν is still open (e.g. minimum flavour violation)
• Current precision
• sin2φ1 ~ 4%
• |Vcb| ~ 1-2%
• |Vub| ~ 8%
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 12
Significance of |Vub| and |Vcb| Measurement
∝|Vub|/|Vcb|
• The precise determination of |Vub|/|Vcb| provides a benchmark for testing new physics in other processes
• Precision of sin2β outstrips the other measurements
• We must make the green ring thinner!
• Large new physics contributions to penguins would have already been seen.
• New physics contributions to decays such as B → τν is still open (e.g. minimum flavour violation)
!-1 -0.5 0 0.5 1
"
-1
-0.5
0
0.5
1#
cbVubV
!-1 -0.5 0 0.5 1
"
-1
-0.5
0
0.5
1
tree level constraints
• Current precision
• sin2φ1 ~ 4%
• |Vcb| ~ 1-2%
• |Vub| ~ 8%
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 12
Significance of |Vub| and |Vcb| Measurement
∝|Vub|/|Vcb|
• The precise determination of |Vub|/|Vcb| provides a benchmark for testing new physics in other processes
• Precision of sin2β outstrips the other measurements
• We must make the green ring thinner!
• Large new physics contributions to penguins would have already been seen.
• New physics contributions to decays such as B → τν is still open (e.g. minimum flavour violation)
!-1 -0.5 0 0.5 1
"
-1
-0.5
0
0.5
1#
cbVubV
!-1 -0.5 0 0.5 1
"
-1
-0.5
0
0.5
1
tree level constraints
4% we’re aiming for
the task is quite challenging!
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
b
decay properties depend directly on |Vcb| & |Vub| and mb
perturbative regime (αsn)
tree level, short distance:
Semileptonic B decays
13
c
e
ν
b c e ν
W
b
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
b
decay properties depend directly on |Vcb| & |Vub| and mb
perturbative regime (αsn)
tree level, short distance:
Semileptonic B decays
13
But quarks are bound by softgluons: non-perturbative
long distance interactions of b quark with light quark
+ long distance:
B D e ν
W
e
ν
]Dc
B[ b
14
B-decay prediction
c
u
d
b → c d u
b
W
14
B-decay prediction
c
u
d
b → c d u
b
WW
b c
u
d]π]DB[
B → D π
Very difficult
14
B-decay predictionb → c e ν
W
e
ν
cbc
u
d
b → c d u
b
WW
b c
u
d]π]DB[
B → D π
Very difficult
14
B-decay predictionb → c e ν
W
e
ν
cbc
u
d
b → c d u
b
WW
b c
u
d]π]DB[
B → D π
Very difficult
B →D e ν
W
]DB[Still hadronic
14
B-decay predictionb → c e ν
W
e
ν
cbc
u
d
b → c d u
b
WW
b c
u
d]π]DB[
B → D π
Very difficult
B →D e ν
W
]DB[Still hadronic
B →τ ν
W τ
νB[b
simple!
u
14
B-decay predictionb → c e ν
W
e
ν
cbc
u
d
b → c d u
b
WW
b c
u
d]π]DB[
B → D π
Very difficult
B →D e ν
W
]DB[Still hadronic
B →τ ν
W τ
νB[b
simple!
u
..but helicity suppressed
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
b
15
Semileptonic Decays: Inclusive
One hadronic current.
Inclusive decays b→q l ν:
Weak quark decay + QCD corrections.
Γsl described by Heavy Quark Expansion in (1/mb)n and αs
k
Non perturbative parameters at each order in mb
need to be derived from data, i.e. from inclusive spectral moments of the semileptonic decay products.
Γ(B → Xc!ν) =G2
Fm5b
192π3|Vcb|
2[[1 + Aew]AnonpertApert]
mb is a renormalisation scheme dependent quantity
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
b
15
Semileptonic Decays: Inclusive
One hadronic current.
Inclusive decays b→q l ν:
Weak quark decay + QCD corrections.
Γsl described by Heavy Quark Expansion in (1/mb)n and αs
k
Non perturbative parameters at each order in mb
need to be derived from data, i.e. from inclusive spectral moments of the semileptonic decay products.
Γ(B → Xc!ν) =G2
Fm5b
192π3|Vcb|
2[[1 + Aew]AnonpertApert]
mb is a renormalisation scheme dependent quantity
Measurement of Vxbb x
Vxb
Wl
ν
! Measurement is very straightforward, use a relation
Γ(b → x!−ν) =G2F192π2
|Vxb|2m5b(
1 +)
! Only need to count the number of b→ x!−ν events, however in reality
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 16
Inclusive semileptonic decays Vcb Vs. Vub
Vub: @8% precision, dominated by theoretical uncertainties
Vcb: At the precision of the cabibbo angle, @~1%, with the inclusion of radiative decay measurements
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
B meson - “Beam”
tag - charge - momentum
B+ and B0 decays studied
separately
Fully reconstruct the tag-side B meson by searching the decay modes
B→D(*)π, B→D(*)ρ, and B→D(*)a1
Bsig→Xlν
π
π
Kγ
γ
l-
ν
B B00
Υ(4S)
Btag→DX
17
Low efficiency (~0.1%), but plenty of data at Belle!
Elepton: in rest frame of signal B
MX: all remaining particles (X),
pX = pbeam-pBtag-pl-pν
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
Particle identification in BelledE/dx from DriftchamberE/p from E.M.Calorimeter Shower Shape
ee eπ
ππ
1 < p < 2 GeV/c 0.8 < p < 1.2 GeV/cE/p > 0.5
0.8 < p < 1.2 GeV/c
Magnet Coil
Vertex Detector
π π
π
Light yield in Cerenkov Detector
Track and cluster matching
Range and transverse scattering in Muon/hadron
detector
@KEK-B
(GeV/c)e*Bp
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
En
trie
s p
er
0.1
Ge
V/c
0
100
200
300
400
500
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.40
100
200
300
400
500Belle data
+B
(GeV/c)e*Bp
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
En
trie
s p
er
0.1
Ge
V/c
0
100
200
300
400
500
600
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.40
100
200
300
400
500
600
Belle data
0B
! e c X"B
! e u X"B
Secondaries
Combinatorial
Continuum
FIG. 2: Measured electron momentum spectra from B+ and B0 decays before background subtrac-
tion, overlaid with the various backgrounds and the MC signal. The errors shown are statistical
only.
TABLE I: Electron yields for p∗Be ≥ 0.4 GeV/c. The errors are statistical only.
B candidate B+ B0
On Resonance Data 6423 ± 80 5403 ± 74
Scaled Off Resonance 249 ± 48 209 ± 39
Combinatorial Background 1244 ± 20 696 ± 13
Secondary 558 ± 11 1847 ± 22
B → Xueν 74 ± 5 57 ± 6
Background Subtracted 4297 ± 96 2593 ± 87
THE ELECTRON ENERGY SPECTRUM
Unfolding
To measure the moments of the electron energy spectrum, we need to determine the true
electron energy spectrum in the B meson rest frame, E∗Be . In this analysis we assume the
10
• beam background.
The photon spectra for ON and scaled OFF data samples along with the results ofsubsequent background subtractions are plotted in Fig. 1(a). The B → Xsγ photon energyspectrum that has been corrected for efficiency is shown in Fig. 1(b). The analysis measuredthe branching fraction,
B(B → Xsγ) = (3.55 ± 0.32+0.30+0.11−0.31−0.07) × 10−4, (1)
where the errors are statistical, systematic and theoretical, respectively. This result agreedwith the latest theoretical calculations [15, 16], as well as with previous measurements madeby CLEO [17] and Belle [18].
102
103
104
105
1.5 2 2.5 3 3.5 4 4.5 5
E*! [GeV]
Ev
en
ts/1
00
Me
V
On-resonance
Scaled off-resonance
On-Off Difference
Background subtracted
E*! [GeV]
Ev
en
ts/1
00
Me
V
-5000
0
5000
10000
15000
20000
25000
1.5 2 2.5 3 3.5 4
(a) (b)
FIG. 1: From [12]. (a) Photon energy spectra in the Υ(4S) frame. (b) Efficiency-corrected photon
energy spectrum. The two error bars show the statistical and total errors.
MOMENT MEASUREMENTS
We follow the procedure used in the published analysis [12], with some slight variations.No attempt is made to correct for the part of the spectrum that is not measured witha satisfactory precision i.e from energy below 1.8 GeV. We apply lower energy thresholdcuts as measured in the Υ(4S) rest frame (E∗
cut) to the efficiency corrected spectrum, fromwhich we obtain truncated first and second moments. Corrections are applied to recover themoments such that the lower energy thresholds correspond to quantities measured in theB-meson rest frame (Ecut).
A simple procedure is used to unfold the effects of detector resolution, the small B-mesonboost in the Υ(4S) frame, and that of the 100 MeV wide bins. We define the first momentas 〈Eγ〉 (mean) and the second moment as ∆E2
γ ≡⟨
E2γ
⟩
−〈Eγ〉2 (variance). The correctionsare as follows:
5
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
b
19
Inclusive decays- Big Picture
rate
shape
shape
|Vcb|
mb,mc
μ2G, μ2
πInclusive Mx spectrum
Semileptonic B decay
|Vub|
2|| ubV∝
€
Shape(B→ Xs γ)
(GeV/c)e*Bp
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
En
trie
s p
er
0.1
GeV
/c
0
100
200
300
400
500
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.40
100
200
300
400
500Belle data
+B
(GeV/c)e*Bp
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
En
trie
s p
er
0.1
GeV
/c
0
100
200
300
400
500
600
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.40
100
200
300
400
500
600
Belle data
0B
! e c X"B
! e u X"B
Secondaries
Combinatorial
Continuum
FIG. 2: Measured electron momentum spectra from B+ and B0 decays before background subtrac-
tion, overlaid with the various backgrounds and the MC signal. The errors shown are statistical
only.
TABLE I: Electron yields for p∗Be ≥ 0.4 GeV/c. The errors are statistical only.
B candidate B+ B0
On Resonance Data 6423 ± 80 5403 ± 74
Scaled Off Resonance 249 ± 48 209 ± 39
Combinatorial Background 1244 ± 20 696 ± 13
Secondary 558 ± 11 1847 ± 22
B → Xueν 74 ± 5 57 ± 6
Background Subtracted 4297 ± 96 2593 ± 87
THE ELECTRON ENERGY SPECTRUM
Unfolding
To measure the moments of the electron energy spectrum, we need to determine the true
electron energy spectrum in the B meson rest frame, E∗Be . In this analysis we assume the
10
Inclusive El spectrum
Belle0
200
400
600
800
1000
1200
1400
1600
0 5 10 15
M2X (GeV
2/c
4)
entr
ies
/ 0.3
33 G
eV2/c
4
Belle
E*l ! 0.7 GeV
data
fake/secondary
combinatorial
B ! Xul"
B ! Xcl"
0
200
400
600
800
1000
1200
1400
1600
0 5 10 15
M2X (GeV
2/c
4)
entr
ies
/ 0.3
33 G
eV2/c
4
Belle
E*l ! 1.1 GeV
0
200
400
600
800
1000
0 5 10 15
M2X (GeV
2/c
4)
entr
ies
/ 0.3
33 G
eV2/c
4
Belle
E*l ! 1.5 GeV
0
50
100
150
200
250
300
350
0 5 10 15
M2X (GeV
2/c
4)
entr
ies
/ 0.3
33 G
eV2/c
4
Belle
E*l ! 1.9 GeV
FIG. 1: Measured hadronic mass spectrum for different lepton energy thresholds. The pointswith error bars are the experimental data after subtraction of the continuum background. Thehistograms show the B → Xc!ν signal and the different background components, explained in moredetail in the text.
and B0 muon). From the unfolded spectrum, we calculate the first moment and its statisticaluncertainty squared,
〈M2X〉 =
∑i(M
2X)ix′
i∑i x
′i
, σ2(〈M2X〉) =
∑i,j(M
2X)iXij(M2
X)j
(∑
i x′i)2
. (3)
Here, x′ is the unfolded spectrum corrected for slightly different bin-to-bin efficiencies and Xis its covariance matrix, also determined by the SVD algorithm. (M2
X)i is the central valueof the i-th bin of the unfolded spectrum. The second central and non-central moments,
8
Experimental Challenge: go from
the measured shapes to the true shapes.
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 20
Moments of inclusive spectra
El>Ecut
Theory calculations predict these moments as a function of the minimum lepton energy in the B rest frame, Ecut :
higher moments are sensitive to 1/mb3 terms → reduce theory error on
| Vcb| and Heavy Quark parameters
comparison of results from different measurements and schemes provides a test of the consistency of predictions and of underlying assumptions
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 21
Electron energy spectrum
Boost into B rest frame
Add bremsstrahlung photons to the electron
Subtract background
Unfold detector effects and QED radiative effects
Sample preparations:
(GeV/c)e*Bp
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
En
trie
s p
er
0.1
GeV
/c
0
100
200
300
400
500
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.40
100
200
300
400
500Belle data
+B
(GeV/c)e*Bp
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
En
trie
s p
er
0.1
GeV
/c
0
100
200
300
400
500
600
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.40
100
200
300
400
500
600
Belle data
0B
! e c X"B
! e u X"B
Secondaries
Combinatorial
Continuum
FIG. 2: Measured electron momentum spectra from B+ and B0 decays before background subtrac-
tion, overlaid with the various backgrounds and the MC signal. The errors shown are statistical
only.
TABLE I: Electron yields for p∗Be ≥ 0.4 GeV/c. The errors are statistical only.
B candidate B+ B0
On Resonance Data 6423 ± 80 5403 ± 74
Scaled Off Resonance 249 ± 48 209 ± 39
Combinatorial Background 1244 ± 20 696 ± 13
Secondary 558 ± 11 1847 ± 22
B → Xueν 74 ± 5 57 ± 6
Background Subtracted 4297 ± 96 2593 ± 87
THE ELECTRON ENERGY SPECTRUM
Unfolding
To measure the moments of the electron energy spectrum, we need to determine the true
electron energy spectrum in the B meson rest frame, E∗Be . In this analysis we assume the
10
(GeV/c)e*Bp
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
En
trie
s p
er
0.1
GeV
/c
0
100
200
300
400
500
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.40
100
200
300
400
500Belle data
+B
(GeV/c)e*Bp
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
En
trie
s p
er
0.1
GeV
/c
0
100
200
300
400
500
600
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.40
100
200
300
400
500
600
Belle data
0B
! e c X"B
! e u X"B
Secondaries
Combinatorial
Continuum
FIG. 2: Measured electron momentum spectra from B+ and B0 decays before background subtrac-
tion, overlaid with the various backgrounds and the MC signal. The errors shown are statistical
only.
TABLE I: Electron yields for p∗Be ≥ 0.4 GeV/c. The errors are statistical only.
B candidate B+ B0
On Resonance Data 6423 ± 80 5403 ± 74
Scaled Off Resonance 249 ± 48 209 ± 39
Combinatorial Background 1244 ± 20 696 ± 13
Secondary 558 ± 11 1847 ± 22
B → Xueν 74 ± 5 57 ± 6
Background Subtracted 4297 ± 96 2593 ± 87
THE ELECTRON ENERGY SPECTRUM
Unfolding
To measure the moments of the electron energy spectrum, we need to determine the true
electron energy spectrum in the B meson rest frame, E∗Be . In this analysis we assume the
10
pe>0.4GeV/cP.Urquijo et al. Phys. Rev. D 75, 032001 (2007)
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 21
Electron energy spectrum
Boost into B rest frame
Add bremsstrahlung photons to the electron
Subtract background
Unfold detector effects and QED radiative effects
Sample preparations:
(GeV/c)e*Bp
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
En
trie
s p
er
0.1
GeV
/c
0
100
200
300
400
500
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.40
100
200
300
400
500Belle data
+B
(GeV/c)e*Bp
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
En
trie
s p
er
0.1
GeV
/c
0
100
200
300
400
500
600
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.40
100
200
300
400
500
600
Belle data
0B
! e c X"B
! e u X"B
Secondaries
Combinatorial
Continuum
FIG. 2: Measured electron momentum spectra from B+ and B0 decays before background subtrac-
tion, overlaid with the various backgrounds and the MC signal. The errors shown are statistical
only.
TABLE I: Electron yields for p∗Be ≥ 0.4 GeV/c. The errors are statistical only.
B candidate B+ B0
On Resonance Data 6423 ± 80 5403 ± 74
Scaled Off Resonance 249 ± 48 209 ± 39
Combinatorial Background 1244 ± 20 696 ± 13
Secondary 558 ± 11 1847 ± 22
B → Xueν 74 ± 5 57 ± 6
Background Subtracted 4297 ± 96 2593 ± 87
THE ELECTRON ENERGY SPECTRUM
Unfolding
To measure the moments of the electron energy spectrum, we need to determine the true
electron energy spectrum in the B meson rest frame, E∗Be . In this analysis we assume the
10
(GeV/c)e*Bp
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
En
trie
s p
er
0.1
GeV
/c
0
100
200
300
400
500
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.40
100
200
300
400
500Belle data
+B
(GeV/c)e*Bp
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
En
trie
s p
er
0.1
GeV
/c
0
100
200
300
400
500
600
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.40
100
200
300
400
500
600
Belle data
0B
! e c X"B
! e u X"B
Secondaries
Combinatorial
Continuum
FIG. 2: Measured electron momentum spectra from B+ and B0 decays before background subtrac-
tion, overlaid with the various backgrounds and the MC signal. The errors shown are statistical
only.
TABLE I: Electron yields for p∗Be ≥ 0.4 GeV/c. The errors are statistical only.
B candidate B+ B0
On Resonance Data 6423 ± 80 5403 ± 74
Scaled Off Resonance 249 ± 48 209 ± 39
Combinatorial Background 1244 ± 20 696 ± 13
Secondary 558 ± 11 1847 ± 22
B → Xueν 74 ± 5 57 ± 6
Background Subtracted 4297 ± 96 2593 ± 87
THE ELECTRON ENERGY SPECTRUM
Unfolding
To measure the moments of the electron energy spectrum, we need to determine the true
electron energy spectrum in the B meson rest frame, E∗Be . In this analysis we assume the
10
pe>0.4GeV/cP.Urquijo et al. Phys. Rev. D 75, 032001 (2007)
Measure up to 4th moment!
0.4 GeV electron energy threshold
electron to be massless, imposing E∗Be = p∗Be . The measured electron energy spectrum is
distorted by various detector effects. Hence, the true electron energy spectrum is extracted
by performing an unfolding procedure based on the Singular Value Decomposition (SVD)
algorithm [34]. The reliability of the unfolding procedure is dependent on the agreement
between data and MC simulation, both for the physics models and the detector response.
The unfolded spectrum is corrected for QED radiative effects using the PHOTOS algo-
rithm [33], as the OPE does not have O(α) QED corrections. The unfolded electron energy
spectrum and the bin to bin statistical covariance matrix calculated with the unfolding al-
gorithm are shown in Figure 3 (for illustrative purposes only, as the full error analysis is
performed on a moment measurement basis).
(GeV/c)e*BE
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4E
ntr
ies p
er
0.1
GeV
/c0
200
400
600
800
1000
Belle
Generated (GeV)e*BE
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
R
eco
nstr
ucte
d (
GeV
)e*B
E
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
FIG. 3: Unfolded electron energy distribution in the B meson rest frame(left), combining con-
tributions from B0 and B+ decays, and corrected for QED radiative effects. The errors shown
are statistical. On the right is the corresponding unfolded electron energy distribution covariance
matrix, where the filled boxes represent negative elements. (These plots are shown for illustrative
purposes only.)
Moments and Branching Fractions
We measure the first four central moments with nine electron energy threshold values
(Ecut = 0.4, 0.6, 0.8, 1.0, 1.2, 1.4, 1.6, 1.8, 2.0 GeV) combining the spectra from B+ and B0
11
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
Ge
V)
-3 (
10
1M
1400
1500
1600
1700
1800
1900
2000
2100
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)2
Ge
V-3
(1
02
M
0
20
40
60
80
100
120
140
160
180
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)3
Ge
V-3
(1
03
M
-25
-20
-15
-10
-5
0
5
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)4
Ge
V-3
(1
04
M
0
10
20
30
40
50
60
70
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
-2
10
!B
R
0
2
4
6
8
10 Belle
FIG. 4: First, second, third and fourth electron energy moments and branching fraction (M1,
M2, M3, M4, B), as a function of the electron threshold energy Ecut . The errors shown are the
statistical and systematic errors added in quadrature.
24
Electron energy moments and partial BR
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
Ge
V)
-3 (
10
1M
1400
1500
1600
1700
1800
1900
2000
2100
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)2
Ge
V-3
(1
02
M
0
20
40
60
80
100
120
140
160
180
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)3
Ge
V-3
(1
03
M
-25
-20
-15
-10
-5
0
5
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)4
Ge
V-3
(1
04
M
0
10
20
30
40
50
60
70
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
-2
10
!B
R
0
2
4
6
8
10 Belle
FIG. 4: First, second, third and fourth electron energy moments and branching fraction (M1,
M2, M3, M4, B), as a function of the electron threshold energy Ecut . The errors shown are the
statistical and systematic errors added in quadrature.
24
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
Ge
V)
-3 (
10
1M
1400
1500
1600
1700
1800
1900
2000
2100
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)2
Ge
V-3
(1
02
M
0
20
40
60
80
100
120
140
160
180
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)3
Ge
V-3
(1
03
M
-25
-20
-15
-10
-5
0
5
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)4
Ge
V-3
(1
04
M
0
10
20
30
40
50
60
70
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
-2
10
!B
R
0
2
4
6
8
10 Belle
FIG. 4: First, second, third and fourth electron energy moments and branching fraction (M1,
M2, M3, M4, B), as a function of the electron threshold energy Ecut . The errors shown are the
statistical and systematic errors added in quadrature.
24
Systematics: b→c model, background, electron detection
BR(B+)0.4GeV=(10.79±0.25±0.27)BR(B0)0.4GeV=(10.08±0.30±0.22)
x10-2
P.Urquijo et al. Phys. Rev. D 75, 032001 (2007)
22
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
Ge
V)
-3 (
10
1M
1400
1500
1600
1700
1800
1900
2000
2100
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)2
Ge
V-3
(1
02
M
0
20
40
60
80
100
120
140
160
180
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)3
Ge
V-3
(1
03
M
-25
-20
-15
-10
-5
0
5
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)4
Ge
V-3
(1
04
M
0
10
20
30
40
50
60
70
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
-2
10
!B
R
0
2
4
6
8
10 Belle
FIG. 4: First, second, third and fourth electron energy moments and branching fraction (M1,
M2, M3, M4, B), as a function of the electron threshold energy Ecut . The errors shown are the
statistical and systematic errors added in quadrature.
24
Electron energy moments and partial BR
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
Ge
V)
-3 (
10
1M
1400
1500
1600
1700
1800
1900
2000
2100
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)2
Ge
V-3
(1
02
M
0
20
40
60
80
100
120
140
160
180
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)3
Ge
V-3
(1
03
M
-25
-20
-15
-10
-5
0
5
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)4
Ge
V-3
(1
04
M
0
10
20
30
40
50
60
70
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
-2
10
!B
R
0
2
4
6
8
10 Belle
FIG. 4: First, second, third and fourth electron energy moments and branching fraction (M1,
M2, M3, M4, B), as a function of the electron threshold energy Ecut . The errors shown are the
statistical and systematic errors added in quadrature.
24
Decrease of truncated BR
Incr
ease
of t
he m
ean Decrease of the width
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
Ge
V)
-3 (
10
1M
1400
1500
1600
1700
1800
1900
2000
2100
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)2
Ge
V-3
(1
02
M
0
20
40
60
80
100
120
140
160
180
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)3
Ge
V-3
(1
03
M
-25
-20
-15
-10
-5
0
5
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
)4
Ge
V-3
(1
04
M
0
10
20
30
40
50
60
70
Belle
cut (GeV)*BeE
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
-2
10
!B
R
0
2
4
6
8
10 Belle
FIG. 4: First, second, third and fourth electron energy moments and branching fraction (M1,
M2, M3, M4, B), as a function of the electron threshold energy Ecut . The errors shown are the
statistical and systematic errors added in quadrature.
24
Systematics: b→c model, background, electron detection
BR(B+)0.4GeV=(10.79±0.25±0.27)BR(B0)0.4GeV=(10.08±0.30±0.22)
x10-2
P.Urquijo et al. Phys. Rev. D 75, 032001 (2007)
22
The hadronic XThe hadronic Xcc systemsystem
!"#
$!#
! !%#
??
&
&'
&''
HQET:
GroundGround
statesstatesBroadBroad
statesstatesNarrowNarrow
statesstates
for L=1 " jq=1/2, jq=3/2
A. OyangurenEPS ‘05 - Lisboa 4
21%
BR(B→Xcl ν)~10.5%
54%
~25%
The hadronic XThe hadronic Xcc systemsystem
!"#
$!#
! !%#
??
&
&'
&''
HQET:
GroundGround
statesstatesBroadBroad
statesstatesNarrowNarrow
statesstates
for L=1 " jq=1/2, jq=3/2
A. OyangurenEPS ‘05 - Lisboa 4
Grounds states Broad states Narrow states
Hadronic Xc system
23
Systematics: b→c model, background subtraction, unfolding
3.9
4
4.1
4.2
4.3
4.4
4.5
1 1.5 2
Emin
* (GeV)
<M
2 X>
(G
eV
2/c
4)
Belle
0
0.25
0.5
0.75
1
1.25
1.5
1.75
2
1 1.5 2
Emin
* (GeV)
<(M
2 X-<
M2 X>
)2>
(G
eV
4/c
8)
Belle
16
17
18
19
20
21
22
1 1.5 2
Emin
* (GeV)
<M
4 X>
(G
eV
4/c
8)
Belle
FIG. 2: Graphical representation of the results in Table II. The error bars indicate the statistical
and total experimental errors.
for B → D∗!ν (B → D!ν) [32]. For B → D∗!ν, we also vary the form factor ratios R1 and
R2 [33].
The LLSW model [18] predicts the relative abundance and the form factor shape of the
different components in B → D∗∗!ν only. To obtain the absolute branching fractions of
the B → D∗∗!ν components and of B → (D(∗)π)non−res.!ν, we use B(B+ → D01!
+ν) =
(5.6 ± 1.6) × 10−3 [31], the recent Belle measurement of B(B → D(∗)π!ν) [34] and the total
semileptonic branching fraction [31]. The uncertainty assigned to the B → D∗∗!ν branching
10
3.9
4
4.1
4.2
4.3
4.4
4.5
1 1.5 2
Emin
* (GeV)
<M
2 X>
(G
eV
2/c
4)
Belle
0
0.25
0.5
0.75
1
1.25
1.5
1.75
2
1 1.5 2
Emin
* (GeV)
<(M
2 X-<
M2 X>
)2>
(G
eV
4/c
8)
Belle
16
17
18
19
20
21
22
1 1.5 2
Emin
* (GeV)
<M
4 X>
(G
eV
4/c
8)
Belle
FIG. 2: Graphical representation of the results in Table II. The error bars indicate the statistical
and total experimental errors.
for B → D∗!ν (B → D!ν) [32]. For B → D∗!ν, we also vary the form factor ratios R1 and
R2 [33].
The LLSW model [18] predicts the relative abundance and the form factor shape of the
different components in B → D∗∗!ν only. To obtain the absolute branching fractions of
the B → D∗∗!ν components and of B → (D(∗)π)non−res.!ν, we use B(B+ → D01!
+ν) =
(5.6 ± 1.6) × 10−3 [31], the recent Belle measurement of B(B → D(∗)π!ν) [34] and the total
semileptonic branching fraction [31]. The uncertainty assigned to the B → D∗∗!ν branching
10
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 24
Hadron invariant mass moments
Moments of unfolded spectra.
decrease in higher mass final states
Measured Mx2
spectra for varying El
*cut
C.Schwanda, P.Urquijo et al., Phys. Rev. D 75, 032005 (2007)
0
200
400
600
800
1000
1200
1400
1600
0 5 10 15
M2X (GeV
2/c
4)
entr
ies
/ 0.3
33 G
eV2/c
4
Belle
E*l ! 0.7 GeV
data
fake/secondary
combinatorial
B ! Xul"
B ! Xcl"
0
200
400
600
800
1000
1200
1400
1600
0 5 10 15
M2X (GeV
2/c
4)
entr
ies
/ 0.3
33 G
eV2/c
4
Belle
E*l ! 1.1 GeV
0
200
400
600
800
1000
0 5 10 15
M2X (GeV
2/c
4)
entr
ies
/ 0.3
33 G
eV2/c
4
Belle
E*l ! 1.5 GeV
0
50
100
150
200
250
300
350
0 5 10 15
M2X (GeV
2/c
4)
entr
ies
/ 0.3
33 G
eV2/c
4
Belle
E*l ! 1.9 GeV
0
200
400
600
800
1000
1200
1400
1600
0 5 10 15
M2X (GeV
2/c
4)
entr
ies
/ 0.3
33 G
eV2/c
4
Belle
E*l ! 0.7 GeV
data
fake/secondary
combinatorial
B ! Xul"
B ! Xcl"
0
200
400
600
800
1000
1200
1400
1600
0 5 10 15
M2X (GeV
2/c
4)
entr
ies
/ 0.3
33 G
eV2/c
4
Belle
E*l ! 1.1 GeV
0
200
400
600
800
1000
0 5 10 15
M2X (GeV
2/c
4)
entr
ies
/ 0.3
33 G
eV2/c
4
Belle
E*l ! 1.5 GeV
0
50
100
150
200
250
300
350
0 5 10 15
M2X (GeV
2/c
4)
entr
ies
/ 0.3
33 G
eV2/c
4
Belle
E*l ! 1.9 GeV
Systematics: b→c model, background subtraction, unfolding
3.9
4
4.1
4.2
4.3
4.4
4.5
1 1.5 2
Emin
* (GeV)
<M
2 X>
(G
eV
2/c
4)
Belle
0
0.25
0.5
0.75
1
1.25
1.5
1.75
2
1 1.5 2
Emin
* (GeV)
<(M
2 X-<
M2 X>
)2>
(G
eV
4/c
8)
Belle
16
17
18
19
20
21
22
1 1.5 2
Emin
* (GeV)
<M
4 X>
(G
eV
4/c
8)
Belle
FIG. 2: Graphical representation of the results in Table II. The error bars indicate the statistical
and total experimental errors.
for B → D∗!ν (B → D!ν) [32]. For B → D∗!ν, we also vary the form factor ratios R1 and
R2 [33].
The LLSW model [18] predicts the relative abundance and the form factor shape of the
different components in B → D∗∗!ν only. To obtain the absolute branching fractions of
the B → D∗∗!ν components and of B → (D(∗)π)non−res.!ν, we use B(B+ → D01!
+ν) =
(5.6 ± 1.6) × 10−3 [31], the recent Belle measurement of B(B → D(∗)π!ν) [34] and the total
semileptonic branching fraction [31]. The uncertainty assigned to the B → D∗∗!ν branching
10
3.9
4
4.1
4.2
4.3
4.4
4.5
1 1.5 2
Emin
* (GeV)
<M
2 X>
(G
eV
2/c
4)
Belle
0
0.25
0.5
0.75
1
1.25
1.5
1.75
2
1 1.5 2
Emin
* (GeV)
<(M
2 X-<
M2 X>
)2>
(G
eV
4/c
8)
Belle
16
17
18
19
20
21
22
1 1.5 2
Emin
* (GeV)
<M
4 X>
(G
eV
4/c
8)
Belle
FIG. 2: Graphical representation of the results in Table II. The error bars indicate the statistical
and total experimental errors.
for B → D∗!ν (B → D!ν) [32]. For B → D∗!ν, we also vary the form factor ratios R1 and
R2 [33].
The LLSW model [18] predicts the relative abundance and the form factor shape of the
different components in B → D∗∗!ν only. To obtain the absolute branching fractions of
the B → D∗∗!ν components and of B → (D(∗)π)non−res.!ν, we use B(B+ → D01!
+ν) =
(5.6 ± 1.6) × 10−3 [31], the recent Belle measurement of B(B → D(∗)π!ν) [34] and the total
semileptonic branching fraction [31]. The uncertainty assigned to the B → D∗∗!ν branching
10
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 24
Hadron invariant mass moments
Moments of unfolded spectra.
decrease in higher mass final states
Measured Mx2
spectra for varying El
*cut
C.Schwanda, P.Urquijo et al., Phys. Rev. D 75, 032005 (2007)
|Vcb| DeterminationMulti-parameter Global fits of HQE to inclusive B decay spectra
INPUT Belle measured moments:
Ee moments: Phys. Rev. D 75, 032001 (2007)
Mx moments: Phys. Rev. D 75, 032005 (2007)
Eγ moments: hep-ex/0508005
P. Urquijo et al., Belle Conf 0669, hep-ex/0611047
25
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
Heavy quark parameters
−
Decay rate in terms of Operator Product Expansion up to 1/mb3
Expansions in terms of:
mkinb ,mkin
c (m1Sb) - mass of b and c quarks
ΛQCD2/mb2μπ2(λ1) - kinetic energy of b quark,μG
2(λ 2) - chromomagnetic coupling
ΛQCD3/mb3 ρD, ρLS (ρ1,τ1-3)
Two approaches:
... and |Vcb|2 dependence on partial branching fractions
P. Urquijo et. al. , Belle Conf 0669, hep-ex/0611047
Pole mass not used anymore: not well behaved: irreducible error on mb
1S mass (C.Bauer, Z.Ligeti, M.Luke,
A.Manohar, M.Trott PRD 70 094017)
Kinetic running mass(P. Gambino and N.Uraltsev, Eur.
Phys. J. C 34, 181 (2004))
26
Theory calculation for B→Xs γ is at 1-loop.
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 27
Experimental data incl. errors and correlaions
Theory (OPE)
|Vcb| and HQ parameters (kinetic scheme)
|Vcb| and HQ parameters (1S scheme)
HQ parameters (SF scheme)
Fit Strategyfit to set of equations
extract
convert
Without truncation of perturbation theory, any path to a given scheme would lead to the same result, e.g.:
[Fit in kinetic scheme] = [Fit in 1S scheme] ⊕ [Translation: 1S → kinetic]
In practice, results differ at finite order in αs.
(GeV)bm4.55 4.6 4.65 4.7 4.75 4.8 4.85 4.9
)2
(G
eV
1!
-0.4
-0.35
-0.3
-0.25
-0.2
-0.15
Belle1S scheme
(GeV)bm4.55 4.6 4.65 4.7 4.75 4.8 4.85 4.9
)2
(G
eV
1!
-0.4
-0.35
-0.3
-0.25
-0.2
-0.15
Belle1S scheme
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 28
Fit to Belle data: |Vcb| and HQ parameters Preliminary
χ2/d.o.f. =17.8/24
mb = 4.564 ± 0.076 GeV
mc = 1.105 ± 0.116 GeV
|Vcb| = ( 41.93 ± 0.65fit ± 0.07αs ± 0.63th ) × 10-3
P. Urquijo et. al. , Belle Conf 0669, hep-ex/0611047
|Vcb| = (41.55 ± 0.68(fit) ± 0.08(τB) ) × 10-3
mb1S = (4.723 ± 0.055) GeVλ1 = (-0.30 ± 0.05) GeV2 χ2/d.o.f. =6/17
(GeV)bm4.55 4.6 4.65 4.7 4.75 4.8 4.85 4.9
)2 (G
eV1!
-0.4
-0.35
-0.3
-0.25
-0.2
-0.15
Belle1S scheme1S scheme
kinetic scheme
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
4.2 4.4 4.6 4.8 5
M2X only
El only
E! only
combined
mb (GeV)
!2 " (
GeV
2)
-0.05
0
0.05
0.1
0.15
0.2
0.25
4.2 4.4 4.6 4.8 5
M2X only
El only
combined
mb (GeV)
#3 D (
GeV
3)
FIG. 3: Stability of the fit. The fit is repeated using lepton energy moments only, hadron mass
moments only and photon energy moments only. The ellipses are ∆χ2 = 1.
IV. SUMMARY
We have performed a fit to the Belle measured moments of the lepton energy and hadronic
mass spectrum in B → Xc!ν decays and the photon energy spectrum in B → Xsγ decays
using HQE expressions derived in the kinetic scheme. We have extracted |Vcb|, mb and other
heavy quark parameters.
It is expected that improved calculations of the perturbative corrections to the hadronic
mass moments could improve the quality of the fit and thus reduce the fit error.
[1] P. Urquijo, Moments of the Electron Energy Spectrum in B → Xc"ν decays at Belle, Belle
note #904.
[2] C. Schwanda, Moments of the hadronic invariant mass spectrum in B → Xc"ν decays at Belle,
Belle note #920.
[3] K. Abe et al. [Belle Collaboration], “Moments of the photon energy spectrum from B –¿ X/s
gamma decays measured arXiv:hep-ex/0508005.
[4] P. Gambino and N. Uraltsev, Eur. Phys. J. C 34, 181 (2004) [arXiv:hep-ph/0401063].
8
Kinetic
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
contours Δχ2=1
1S
Global fit with all available results (except the latest BaBar moments)
Kinetic Vcb (10-3) mb (GeV)
41.91±0.19±0.28±0.59
4.613±0.022+0.027
no b→sγ 41.68±0.39±0.58 4.677±0.053
1S Vcb (10-3) mb (GeV)
41.78±0.30±0. 08 4.70±0.03
no b→sγ 41.56±0.39±0.08 4.751±0.058
Vcb and HQ parameters (world fit)
c.f. mb1S=4.71±0.03 GeV (Hoang PRD 61 034005)
BaBar, Belle, CDF, CLEO & DELPHI measured moments
using upsilon sum rules29
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 30
Problem with using B→Xsγmoment data
“ B→XSγ should not be used to extract heavy quark parameters ” (M. Neubert Lepton-Photon 2007)
No model-independent treatment is possible ( Neubert & Becher, Phys. Rev. Lett. 98, 022003 (2007 )
Present analyses either use OPE directly (not fully justified), or use a model to apply a “bias correction” to the data before using the OPE
Tension in data, which lowers mb, and raises |Vcb| and |Vub|
contours Δχ2=1
1S
|Vub| DeterminationLimiting factor in CKM precision tests; known much less well than |Vcb|
CKM suppressed Vub~0.1xVcb- therefore harder to measure
The problem: b → clv decay
Tree level
How can we suppress 50× larger background?How can we reach 4% precision on |Vub|?
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
Cut away b → clv Lose a part of the b → ulv signal
We measure
Fraction of the signal that pass the cut
Requires the knowledge of the b quark’s motion inside the B meson
Figuring out what we see
Total b → ulv rate
Cut-dependentconstant predicted
by theory
Must be corrected for QCD
Main uncertainty (±5%) from mb5 → ±2.5% on |Vub|, correlated
between all measurements/experiments!
But we need the reasonable fraction (e.g., Eℓ > 2 GeV) of the rate.
32
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 33
Detecting b → ulν• Inclusive: Use mu << mc difference in kinematics
• Maximum lepton energy 2.64 vs. 2.31 GeV• First observations (1990) used this technique• Only 6% of signal accessible
• How accurately do we know this fraction?
2.64
2.31
b→c background
Other background
ΔBR x 10-4 % total phase space
Belle 27fb-1:Ee>1.9 GeV 8.47 ± 0.37stat ± 1.53sys ~24%
(a)
Ele
ctr
on
Can
did
ate
s / (
50 M
eV
/c)
Ele
ctr
on
Can
did
ate
s / (
50 M
eV
/c)
(b)
Momentum (GeV/c)
102
103
104
0
2500
5000
7500
2 2.2 2.4 2.6 2.8 3
Fig. 2. The electron momentum spectrum in the Υ(4S) rest frame: (a) ON data(filled circles), scaled OFF data (open circles), sum of scaled OFF data and esti-mated BB backgrounds (histogram). (b) ON data after subtraction of backgroundsand correction for efficiency (filled circles) and model spectrum of B → Xueνe
decays with final state radiation (histogram, normalised to the data yield in the1.9 − 2.6 GeV/c momentum range).
the size of the effect. The correction factors for each HI region are given inTable 2.
The values of fu for the different momentum intervals are determined withthe DeFazio-Neubert prescription, using three different forms of the shapefunction with the parameters, ΛSF and λSF
1 , determined from fits to the Bellemeasured photon energy spectrum in B → Xsγ decays [21,24]. The resultantvalues of fu are given in Table 2, they range from 3 − 32% as the lowermomentum limit is decreased. The statistical uncertainty, averaged over eachshape function form, is determined from the half-difference of maximum and
15
(a)
Ele
ctr
on
Ca
nd
ida
tes
/ (
50
Me
V/c
)E
lec
tro
n C
an
did
ate
s /
(5
0 M
eV
/c)
(b)
Momentum (GeV/c)
102
103
104
0
2500
5000
7500
2 2.2 2.4 2.6 2.8 3
Fig. 2. The electron momentum spectrum in the Υ(4S) rest frame: (a) ON data(filled circles), scaled OFF data (open circles), sum of scaled OFF data and esti-mated BB backgrounds (histogram). (b) ON data after subtraction of backgroundsand correction for efficiency (filled circles) and model spectrum of B → Xueνe
decays with final state radiation (histogram, normalised to the data yield in the1.9 − 2.6 GeV/c momentum range).
the size of the effect. The correction factors for each HI region are given inTable 2.
The values of fu for the different momentum intervals are determined withthe DeFazio-Neubert prescription, using three different forms of the shapefunction with the parameters, ΛSF and λSF
1 , determined from fits to the Bellemeasured photon energy spectrum in B → Xsγ decays [21,24]. The resultantvalues of fu are given in Table 2, they range from 3 − 32% as the lowermomentum limit is decreased. The statistical uncertainty, averaged over eachshape function form, is determined from the half-difference of maximum and
15
(a)
Ele
ctr
on
Can
did
ate
s / (
50 M
eV
/c)
Ele
ctr
on
Can
did
ate
s / (
50 M
eV
/c)
(b)
Momentum (GeV/c)
102
103
104
0
2500
5000
7500
2 2.2 2.4 2.6 2.8 3
Fig. 2. The electron momentum spectrum in the Υ(4S) rest frame: (a) ON data(filled circles), scaled OFF data (open circles), sum of scaled OFF data and esti-mated BB backgrounds (histogram). (b) ON data after subtraction of backgroundsand correction for efficiency (filled circles) and model spectrum of B → Xueνe
decays with final state radiation (histogram, normalised to the data yield in the1.9 − 2.6 GeV/c momentum range).
the size of the effect. The correction factors for each HI region are given inTable 2.
The values of fu for the different momentum intervals are determined withthe DeFazio-Neubert prescription, using three different forms of the shapefunction with the parameters, ΛSF and λSF
1 , determined from fits to the Bellemeasured photon energy spectrum in B → Xsγ decays [21,24]. The resultantvalues of fu are given in Table 2, they range from 3 − 32% as the lowermomentum limit is decreased. The statistical uncertainty, averaged over eachshape function form, is determined from the half-difference of maximum and
15
(a)
Ele
ctr
on
Can
did
ate
s / (
50 M
eV
/c)
Ele
ctr
on
Can
did
ate
s / (
50 M
eV
/c)
(b)
Momentum (GeV/c)
102
103
104
0
2500
5000
7500
2 2.2 2.4 2.6 2.8 3
Fig. 2. The electron momentum spectrum in the Υ(4S) rest frame: (a) ON data(filled circles), scaled OFF data (open circles), sum of scaled OFF data and esti-mated BB backgrounds (histogram). (b) ON data after subtraction of backgroundsand correction for efficiency (filled circles) and model spectrum of B → Xueνe
decays with final state radiation (histogram, normalised to the data yield in the1.9 − 2.6 GeV/c momentum range).
the size of the effect. The correction factors for each HI region are given inTable 2.
The values of fu for the different momentum intervals are determined withthe DeFazio-Neubert prescription, using three different forms of the shapefunction with the parameters, ΛSF and λSF
1 , determined from fits to the Bellemeasured photon energy spectrum in B → Xsγ decays [21,24]. The resultantvalues of fu are given in Table 2, they range from 3 − 32% as the lowermomentum limit is decreased. The statistical uncertainty, averaged over eachshape function form, is determined from the half-difference of maximum and
15
signalGeV
b → u
b → c
Phys.Lett.B621:28-40 (2005)
ELEP
(GeV)
A.U
.
0
0.025
0.05
0.075
0.1
0 1 2 3
mX
(GeV)
A.U
.
b ! u
b ! c
0
0.2
0.4
0 1 2 3
q2 (GeV
2)
A.U
.
b ! u
b ! c
0
0.02
0.04
0.06
0 5 10 15 20
P+ (GeV)
A.U
.
0
0.02
0.04
0.06
0.08
0 0.5 1 1.5 2
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
ELEP
(GeV)
A.U
.
0
0.025
0.05
0.075
0.1
0 1 2 3
mX
(GeV)
A.U
.
b ! u
b ! c
0
0.2
0.4
0 1 2 3
q2 (GeV
2)
A.U
.
b ! u
b ! c
0
0.02
0.04
0.06
0 5 10 15 20
P+ (GeV)
A.U
.
0
0.02
0.04
0.06
0.08
0 0.5 1 1.5 2E
LEP (GeV)
A.U
.
0
0.025
0.05
0.075
0.1
0 1 2 3
mX
(GeV)
A.U
.
b ! u
b ! c
0
0.2
0.4
0 1 2 3
q2 (GeV
2)
A.U
.
b ! u
b ! c
0
0.02
0.04
0.06
0 5 10 15 20
P+ (GeV)
A.U
.
0
0.02
0.04
0.06
0.08
0 0.5 1 1.5 2
34
Accessing a little more phase space
q2 = lepton-neutrino mass squared
mX = hadron system mass
El = lepton energy
b→u
b→c
b→u
b→c
b→ub→c
Not
to
scal
e!
u-quark turns into one or more hadrons
Best access to kinematics with B meson “beam” 41
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 35
Hadronic Mass
The background subtracted hadron
mass spectrum
Same recoil technique as for B->Xclν
ΔBR x 10-4 % total phase space
Belle 253 fb-1: mX < 1.7 12.4 ± 1.1stat ± 1.0sys ~47%
Belle 253 fb-1: mX < 1.7, q2 > 8 8.4 ± 0.8stat ± 1.0sys ~24%
Phys.Rev.Lett.95:241801,2005
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 36
Shape Function
Shape Function needs to be determined from experimental data
Theory doesn’t work everywhere in the phase space →resumming turns non-perturbative terms into a Shape Function F(K+)
0
F(K+)
Λ = MB -mb
K+
We can determine mean and width of shape function.
Shape not well constrained.
Two ways to solve the problem:
1. Cutoff-based “shape Function” approach2. Use something different : Dressed Gluon Exponentiation
Theory calculation for SF is at 2-loop.
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 37
Predicting b → ulν Spectra
Fit the b → sγ spectrum to constrain the shape function
Additional information from b → clv decays b-quark: mass and kinetic energy
Plug the SF into the b → ulv spectrum calculations
Ready to extract |Vub|
€
dxxicut
∫ dΓ(B→Xulυ )dxi
= dEγ∫ W(xi,cutEγ )dΓ(B→Xsγ )
dEγ
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 38
How Things Mesh Together
Inclusiveb → ulv
q2
b→sγ
ShapeFunction
Eγ
mb
Inclusive b → clv
mXEl
HQE Fit
mX
El
WA
duality
|Vub|
SSFs
|Vcb|
AKA: The HQE plumbing diagram
]-3 10×| [ub|V2 4 6
]-3 10×| [ub|V2 4 6
) eCLEO (E 0.44± 0.46 ±3.91
) 2, qXBELLE sim. ann. (m 0.36± 0.45 ±4.23
) eBELLE (E 0.38± 0.43 ±4.67
) eBABAR (E 0.39± 0.24 ±4.23
) hmax, seBABAR (E
0.49± 0.29 ±4.37
) XBELLE (m 0.32± 0.26 ±3.92
) XBABAR (m 0.39± 0.20 ±4.09
Average +/- exp +/- (mb,theory) 0.35± 0.17 ±4.31
HFAGLP 2007
OPE-HQET-SCET (BLNP)Phys.Rev.D72:073006,2005
moments! s " and b# c l " input from bbm
/dof = 6.1/ 6 (CL = 41 %)2$
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 39
Inclusive |Vub|: BLNP framework
Statistical ±2.0%
Expt. syst. ±3.3%
b → clν model ±1.8%
b → ulν model ±1.1%
HQ params. ±6.9%
Theory ±3.8%
WA ±1.7%
|Vub| world average as of Summer 07 |Vub| determined to ± 8.9%
CKM Workshop Nagoya 2006
LNP
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 40
Theory Errors1.7% Weak annihilation
Expected to be <2% of the total rate
3.8% Higher order non-perturbative corrections
Cannot be constrained with b → sγ6.9% HQ parameters
mainly mb
b
u
soft
B
From my |Vcb| fit!
only limits determined
All of these uncertainties can be reduced significantly if we increase the measured phase space.
The power dependence of mb and related theoretical uncertainties scale inversely with phase space coverage.
ubV0.003 0.0035 0.004 0.0045 0.005
)ub
(V!
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5-310×
0
1
2
3
4
5
6 !
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 41
The side versus the angleMost probable value of |Vub| from measurements of other CKM parameters in the Standard Model
|Vub| inclusive
!"#$%&'%()#*++, -.#-/(001#2%(3%(4 56
!"##$%&
! 7(890:8#48;8(<0&%;0/&#/=#>!"#>#9/<?@8<8&;:#:0&*! ;/#;8:;#;A8
B0&C9/<?@8;8&8:: /=#;A8#D;%&4%(4#-/48@
! "".,E#%99'(%9)#%9A08384#:/#=%(" FE#?/::0G@8H
! I@/:8#9/@@%G/(%;0/&#G8;J88&#;A8/()#%&4#8K?8(0<8&;#0:#9('90%@
! L%?04#?(/M(8::#0�&9@':038#>!"#>#0&#;A8#@%:;#*#)8%(:
! N<?(/38<8&;#0& $! #!$ =/(<#=%9;/(#0:#&88484
!
>!"# >
φ1
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
[GeV]0E2 2.1 2.2 2.3 2.4
]-3| [
10ub
|V
0
2
4
6
8
10
12 Calculation
LLRLNP
Graph
A novel approach for |Vub| determinationar
Xiv
:hep
-ph/0
702072v1 7 F
eb 2
007
SLAC-PUB-12336
Extraction of |Vub| with Reduced Dependenceon Shape Functions.
Vladimir B. Golubev, Yuri I. SkovpenBudker Institute for Nuclear Physics, Novosibirsk 630090, Russia.
Vera G. LuthStanford Linear Accelerator Center, Stanford, CA 94309, USA.
February 5, 2007
Abstract
Using BABAR measurements of the inclusive electron spectrum in B → Xueν decays andthe inclusive photon spectrum in B → Xsγ decays, we extract the magnitude of the CKMmatrix element Vub. The extraction is based on several theoretical calculations designed toreduce the theoretical uncertainties by exploiting the assumption that the leading shapefunctions are the same for all b → q transitions (q is a light quark). The current resultsagree well with the previous analysis, have indeed smaller theoretical errors, but arepresently limited by the knowledge of the photon spectrum and the experimental errorson the lepton spectrum.
1
Babar data
Novel method for |Vub| extraction theoretical calculations designed toreduce theoretical uncertainties by exploiting the assumption that the leading shape functions are the same for all b→q transitions.
Uncertainties due to non-perturbative power corrections increase rapidly near the kinematic end point (not included in LLR).
Note:
The results agree with analyses that rely on an assumed shape function form.
42
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
Rectangular cuts? A linear boundary? A nonlinear one?
• Exploit the many non-linear correlations between kinematic variables that separate b→u and b→c.
• Use it to reduce theoretical errors traditionally large due to limited phase space.
• No need to place stringent hard cuts that result in zero efficiency.
Insight into the futureMy latest approach for “hunting down” b→u
43
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
Almost fully inclusive: Shape Function not needed!
Insight into the future
|Vub| measurement (using BLNP)
Phys.Rev.Lett. 96 (2006) 221801 BaBar: Mx<2.5GeV
Phys.Rev.Lett. 96 (2006) 221801 Belle: Mx<1.7GeV
MVA “guestimate”
fu uncertainties (%)stat sys the tot
~90% 18.2 7.8 2.6 ~20
~47% 5.0 4.8 5.9 ~9
~90% ~5 ~4 ~2.6 ~6.9
|Vub| theory errors are highly correlated. Measurements of the full phase space are crucial to reducing the error on
|Vub| in the world average!
• Work is underway on a novel- non-linear approach to selecting signal events, allowing for ~90% phase space coverage without a loss in statistical power! Coming soon from Belle.
44
HQE param from S.L. and
radiativeResults for various cuts
3.64±0.18±0.3513.64.09±0.20±0.390.41BaBar, MX>1.55 GeV
3.91±0.42±0.339.34.23±0.45±0.360.24Belle, MX>1.7 GeV,
q2>8 GeV2
3.58±0.24±0.2910.53.92±0.26±0.320.47Belle, MX<1.7 GeV
3.88±0.26±0.4413.84.37±0.29±0.490.13BaBar, Ee>2.0 GeV,
shmax<3.5 GeV2
3.84±0.22±0.3511.44.23±0.24±0.390.19BaBar, Ee>2.0 GeV
4.31±0.40±0.339.54.67±0.43±0.360.24Belle, Ee>1.9 GeV
3.77±0.15±0.31
4.31±0.17±0.35
4.18±0.16±0.34
Average (HFAG)
My average
3.46±0.41±0.3914.43.91±0.46±0.440.13CLEO, Ee>2.1 GeV
|Vub| [10-3](mb=4.71±0.06 GeV)
a|Vub| [10-3]
(mb=4.63±0.06 GeV)fu
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
Results for various cuts
45
M.Neubert (LP07)
<
<
fraction of total phase space HQE param from S.L. only
HQE param from S.L. and
radiativeResults for various cuts
3.64±0.18±0.3513.64.09±0.20±0.390.41BaBar, MX>1.55 GeV
3.91±0.42±0.339.34.23±0.45±0.360.24Belle, MX>1.7 GeV,
q2>8 GeV2
3.58±0.24±0.2910.53.92±0.26±0.320.47Belle, MX<1.7 GeV
3.88±0.26±0.4413.84.37±0.29±0.490.13BaBar, Ee>2.0 GeV,
shmax<3.5 GeV2
3.84±0.22±0.3511.44.23±0.24±0.390.19BaBar, Ee>2.0 GeV
4.31±0.40±0.339.54.67±0.43±0.360.24Belle, Ee>1.9 GeV
3.77±0.15±0.31
4.31±0.17±0.35
4.18±0.16±0.34
Average (HFAG)
My average
3.46±0.41±0.3914.43.91±0.46±0.440.13CLEO, Ee>2.1 GeV
|Vub| [10-3](mb=4.71±0.06 GeV)
a|Vub| [10-3]
(mb=4.63±0.06 GeV)fu
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
Results for various cuts
45
M.Neubert (LP07)
<
<
fraction of total phase space
ubV0.003 0.0035 0.004 0.0045 0.005
)ub
(V!
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5-310×
0
1
2
3
4
5
6 !
HQE param from S.L. only
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 46
Conclusionsb→ clν
|Vcb| with 1% error in the 1S scheme (2% Belle only), mb @1% (kinetic and 1S schemes), mc @5% (kinetic scheme).
A precise mb is an essential ingredient for reducing the error on inclusive Vub.
b→ ulν Inclusive Vub ~8% error shared between theoretical and experimental. Went up in the past year!
|Vub| @ 4~5% possible? Intelligent techniques to hunt down b→ ulν will help, and work on the theoretical error is ongoing.
|Vub| Vs. sin2φ1 puzzle resolved when only model-independent information on mb is used.mb extracted from semileptonic moments yields |Vub|↓10% lower than if photon moments are included.
mb↑ |Vub|↓ |Vcb|↓
!"#$%&'%()#*++, -.#-/(001#2%(3%(4 56
!"##$%&
! 7(890:8#48;8(<0&%;0/&#/=#>!"#>#9/<?@8<8&;:#:0&*! ;/#;8:;#;A8
B0&C9/<?@8;8&8:: /=#;A8#D;%&4%(4#-/48@
! "".,E#%99'(%9)#%9A08384#:/#=%(" FE#?/::0G@8H
! I@/:8#9/@@%G/(%;0/&#G8;J88&#;A8/()#%&4#8K?8(0<8&;#0:#9('90%@
! L%?04#?(/M(8::#0�&9@':038#>!"#>#0&#;A8#@%:;#*#)8%(:
! N<?(/38<8&;#0& $! #!$ =/(<#=%9;/(#0:#&88484
!
>!"# >
φ1New physics hiding skillfully! (behind theoretical uncertainties)
|Vub| & |Vcb| @ Belle
Backup slides
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 48
Inclusive |Vub|: DGE framework
Statistical ±2.0%
Expt. syst. ±3.2%
b → clν model ±1.9%
b → ulν model ±1.0%
DGE calc. ±3.1%
WA ±1.9%
Theory ±4.4%
|Vub| determined to ± 6.8%|Vub| world average as of Summer 07
DGE calc.: matching scheme method and scale Theory: mb(MS): ± 1.3%
αs: ± 1.0%
total SL width: ± 3.0 %
Here, 〈X〉meas.i are the measured moments. 〈X〉kin
i are the corresponding kinetic scheme
predictions that depend on these free parameters. The covariance matrix is the sum of the
experimental and theoretical error matrices.
We determine the CKM element |Vcb| by treating it as an eighth free parameter in the
fit. |Vcb| is related to the semileptonic width Γ(B → Xc!ν) by [5]
|Vcb|0.0417
=
(Γ(B → Xc!ν)
1.55 ps
0.105
)1/2
× (1− 0.0018)× (1 + 0.30(αs − 0.22)) (10)
×(1− 0.66(mb − 4.6 GeV) + 0.39(mc − 1.15 GeV)
+0.013(µ2π − 0.4 GeV2) + 0.09(ρ3
D − 0.1 GeV3)
+0.05(µ2G − 0.35 GeV2)− 0.01(ρ3
LS + 0.15 GeV3) .
Using this expression, we calculate Γ(B → Xc!ν) from |Vcb| and add the following term to
the χ2 function,
χ′2 = χ2 + (BXc"ν
Γ(B → Xc!ν)− τB)2/σ2
τB. (11)
As µ2G and ρ3
LS are determined from B∗ − B mass splitting and heavy quark sum rules
and because the expressions depend only weakly on these parameters, we fix µ2G and ρ3
LS to
0.35± 0.07 GeV2 and −0.15± 0.1 GeV3, respectively, by adding the following terms to the
χ2 function,
χ′′2 = χ′2 + (µ2G − 0.35 GeV2)2/(0.07 GeV2)2 + (ρ3
LS + 0.15 GeV3)2/(0.1 GeV3)2 . (12)
The minimization of the χ′′2 is performed using MINUIT [15].
C. Fit results and discussion
The result of the kinetic scheme analysis is shown in Table IV and in Figs. 4 and 5. The
value of the χ2 function at the minimum is 17.76, compared to (31− 7) degrees of freedom.
All results are preliminary.
To assess the stability of the fit, we have repeated the analysis using lepton energy
moments only, hadron mass moments only and photon energy moments only. The result
is shown in Fig. 6. In general, changes are well covered by the fit uncertainty though the
B → Xsγ data seems to prefer lower values of mb.
Finally, the result for |Vcb| reads
|Vcb| = (41.93± 0.65(fit)± 0.48(αs)± 0.63(th))× 10−3 .
11
Here, 〈X〉meas.i are the measured moments. 〈X〉kin
i are the corresponding kinetic scheme
predictions that depend on these free parameters. The covariance matrix is the sum of the
experimental and theoretical error matrices.
We determine the CKM element |Vcb| by treating it as an eighth free parameter in the
fit. |Vcb| is related to the semileptonic width Γ(B → Xc!ν) by [5]
|Vcb|0.0417
=
(Γ(B → Xc!ν)
1.55 ps
0.105
)1/2
× (1− 0.0018)× (1 + 0.30(αs − 0.22)) (10)
×(1− 0.66(mb − 4.6 GeV) + 0.39(mc − 1.15 GeV)
+0.013(µ2π − 0.4 GeV2) + 0.09(ρ3
D − 0.1 GeV3)
+0.05(µ2G − 0.35 GeV2)− 0.01(ρ3
LS + 0.15 GeV3) .
Using this expression, we calculate Γ(B → Xc!ν) from |Vcb| and add the following term to
the χ2 function,
χ′2 = χ2 + (BXc"ν
Γ(B → Xc!ν)− τB)2/σ2
τB. (11)
As µ2G and ρ3
LS are determined from B∗ − B mass splitting and heavy quark sum rules
and because the expressions depend only weakly on these parameters, we fix µ2G and ρ3
LS to
0.35± 0.07 GeV2 and −0.15± 0.1 GeV3, respectively, by adding the following terms to the
χ2 function,
χ′′2 = χ′2 + (µ2G − 0.35 GeV2)2/(0.07 GeV2)2 + (ρ3
LS + 0.15 GeV3)2/(0.1 GeV3)2 . (12)
The minimization of the χ′′2 is performed using MINUIT [15].
C. Fit results and discussion
The result of the kinetic scheme analysis is shown in Table IV and in Figs. 4 and 5. The
value of the χ2 function at the minimum is 17.76, compared to (31− 7) degrees of freedom.
All results are preliminary.
To assess the stability of the fit, we have repeated the analysis using lepton energy
moments only, hadron mass moments only and photon energy moments only. The result
is shown in Fig. 6. In general, changes are well covered by the fit uncertainty though the
B → Xsγ data seems to prefer lower values of mb.
Finally, the result for |Vcb| reads
|Vcb| = (41.93± 0.65(fit)± 0.07(αs)± 0.63(th))× 10−3 .
14
Here, 〈X〉meas.i are the measured moments. 〈X〉kin
i are the corresponding kinetic scheme
predictions that depend on these free parameters. The covariance matrix is the sum of the
experimental and theoretical error matrices.
We determine the CKM element |Vcb| by treating it as an eighth free parameter in the
fit. |Vcb| is related to the semileptonic width Γ(B → Xc!ν) by [5]
|Vcb|0.0417
=
(Γ(B → Xc!ν)
1.55 ps
0.105
)1/2
× (1− 0.0018)× (1 + 0.30(αs − 0.22)) (10)
×(1− 0.66(mb − 4.6 GeV) + 0.39(mc − 1.15 GeV)
+0.013(µ2π − 0.4 GeV2) + 0.09(ρ3
D − 0.1 GeV3)
+0.05(µ2G − 0.35 GeV2)− 0.01(ρ3
LS + 0.15 GeV3) .
Using this expression, we calculate Γ(B → Xc!ν) from |Vcb| and add the following term to
the χ2 function,
χ′2 = χ2 + (BXc"ν
Γ(B → Xc!ν)− τB)2/σ2
τB. (11)
As µ2G and ρ3
LS are determined from B∗ − B mass splitting and heavy quark sum rules
and because the expressions depend only weakly on these parameters, we fix µ2G and ρ3
LS to
0.35± 0.07 GeV2 and −0.15± 0.1 GeV3, respectively, by adding the following terms to the
χ2 function,
χ′′2 = χ′2 + (µ2G − 0.35 GeV2)2/(0.07 GeV2)2 + (ρ3
LS + 0.15 GeV3)2/(0.1 GeV3)2 . (12)
The minimization of the χ′′2 is performed using MINUIT [15].
C. Fit results and discussion
The result of the kinetic scheme analysis is shown in Table IV and in Figs. 4 and 5. The
value of the χ2 function at the minimum is 17.76, compared to (31− 7) degrees of freedom.
All results are preliminary.
To assess the stability of the fit, we have repeated the analysis using lepton energy
moments only, hadron mass moments only and photon energy moments only. The result
is shown in Fig. 6. In general, changes are well covered by the fit uncertainty though the
B → Xsγ data seems to prefer lower values of mb.
Finally, the result for |Vcb| reads
|Vcb| = (41.93± 0.65(fit)± 0.07(αs)± 0.63(th))× 10−3 .
14
Here, 〈X〉meas.i are the measured moments. 〈X〉kin
i are the corresponding kinetic scheme
predictions that depend on these free parameters. The covariance matrix is the sum of the
experimental and theoretical error matrices.
We determine the CKM element |Vcb| by treating it as an eighth free parameter in the
fit. |Vcb| is related to the semileptonic width Γ(B → Xc!ν) by [5]
|Vcb|0.0417
=
(Γ(B → Xc!ν)
1.55 ps
0.105
)1/2
× (1− 0.0018)× (1 + 0.30(αs − 0.22)) (10)
×(1− 0.66(mb − 4.6 GeV) + 0.39(mc − 1.15 GeV)
+0.013(µ2π − 0.4 GeV2) + 0.09(ρ3
D − 0.1 GeV3)
+0.05(µ2G − 0.35 GeV2)− 0.01(ρ3
LS + 0.15 GeV3) .
Using this expression, we calculate Γ(B → Xc!ν) from |Vcb| and add the following term to
the χ2 function,
χ′2 = χ2 + (BXc"ν
Γ(B → Xc!ν)− τB)2/σ2
τB. (11)
As µ2G and ρ3
LS are determined from B∗ − B mass splitting and heavy quark sum rules
and because the expressions depend only weakly on these parameters, we fix µ2G and ρ3
LS to
0.35± 0.07 GeV2 and −0.15± 0.1 GeV3, respectively, by adding the following terms to the
χ2 function,
χ′′2 = χ′2 + (µ2G − 0.35 GeV2)2/(0.07 GeV2)2 + (ρ3
LS + 0.15 GeV3)2/(0.1 GeV3)2 . (12)
The minimization of the χ′′2 is performed using MINUIT [15].
C. Fit results and discussion
The result of the kinetic scheme analysis is shown in Table IV and in Figs. 4 and 5. The
value of the χ2 function at the minimum is 17.76, compared to (31− 7) degrees of freedom.
All results are preliminary.
To assess the stability of the fit, we have repeated the analysis using lepton energy
moments only, hadron mass moments only and photon energy moments only. The result
is shown in Fig. 6. In general, changes are well covered by the fit uncertainty though the
B → Xsγ data seems to prefer lower values of mb.
Finally, the result for |Vcb| reads
|Vcb| = (41.93± 0.65(fit)± 0.07(αs)± 0.63(th))× 10−3 .
14
Here, 〈X〉meas.i are the measured moments. 〈X〉kin
i are the corresponding kinetic scheme
predictions that depend on these free parameters. The covariance matrix is the sum of the
experimental and theoretical error matrices.
We determine the CKM element |Vcb| by treating it as an eighth free parameter in the
fit. |Vcb| is related to the semileptonic width Γ(B → Xc!ν) by [5]
|Vcb|0.0417
=
(Γ(B → Xc!ν)
1.55 ps
0.105
)1/2
× (1− 0.0018)× (1 + 0.30(αs − 0.22)) (10)
×(1− 0.66(mb − 4.6 GeV) + 0.39(mc − 1.15 GeV)
+0.013(µ2π − 0.4 GeV2) + 0.09(ρ3
D − 0.1 GeV3)
+0.05(µ2G − 0.35 GeV2)− 0.01(ρ3
LS + 0.15 GeV3) .
Using this expression, we calculate Γ(B → Xc!ν) from |Vcb| and add the following term to
the χ2 function,
χ′2 = χ2 + (BXc"ν
Γ(B → Xc!ν)− τB)2/σ2
τB. (11)
As µ2G and ρ3
LS are determined from B∗ − B mass splitting and heavy quark sum rules
and because the expressions depend only weakly on these parameters, we fix µ2G and ρ3
LS to
0.35± 0.07 GeV2 and −0.15± 0.1 GeV3, respectively, by adding the following terms to the
χ2 function,
χ′′2 = χ′2 + (µ2G − 0.35 GeV2)2/(0.07 GeV2)2 + (ρ3
LS + 0.15 GeV3)2/(0.1 GeV3)2 . (12)
The minimization of the χ′′2 is performed using MINUIT [15].
C. Fit results and discussion
The result of the kinetic scheme analysis is shown in Table IV and in Figs. 4 and 5. The
value of the χ2 function at the minimum is 17.76, compared to (31− 7) degrees of freedom.
All results are preliminary.
To assess the stability of the fit, we have repeated the analysis using lepton energy
moments only, hadron mass moments only and photon energy moments only. The result
is shown in Fig. 6. In general, changes are well covered by the fit uncertainty though the
B → Xsγ data seems to prefer lower values of mb.
Finally, the result for |Vcb| reads
|Vcb| = (41.93± 0.65(fit)± 0.07(αs)± 0.63(th))× 10−3 .
14
EPFL LPHE Seminar Oct 2007 Phillip Urquijo
Kinetic Scheme fit procedure
49
Belle Conf 0669, hep-ex/0611047
B-B* mass splitting constrained
Non-Perturbative Perturbative (vary αs) Bias correction ΓSL
vary 1/mb2 parameters by 20%
vary 1/mb3 parameters by 30%
lepton/photon: 0.220 +/- 0.040
hadron: 0.3 +/- 0.1
|Vcb|: 0.220 +/- 0.008
vary bias correction by 30% (uncorrelated
between Emin)
1.5% uncertainty due to the limited accuracy of the
ΓSL expression.
Theoretical uncertainties:
Constraints:=
=
BXc!ν (%) mb (GeV) mc (GeV) µ2π (GeV2) ρ3
D (GeV3) µ2G (GeV2) ρ3
LS (GeV3)
value 10.590 4.564 1.105 0.557 0.162 0.358 -0.174
σ(fit) 0.164 0.076 0.116 0.091 0.053 0.060 0.098
σ(αs) 0.006 0.003 0.005 0.013 0.008 0.003 0.003
BXc!ν 1.000 −0.023 0.003 0.229 0.192 −0.147 −0.024
mb 1.000 0.983 −0.729 −0.623 −0.024 −0.111
mc 1.000 −0.716 −0.633 −0.124 −0.033
µ2π 1.000 0.851 0.005 0.052
ρ3D 1.000 −0.046 −0.156
µ2G 1.000 −0.071
ρ3LS 1.000
TABLE IV: Results of the kinetic scheme fit. The error from the fit contains the uncertainties
related to the experiment, the non-perturbative corrections and the bias correction. σ(αs) is the
uncertainty related to the perturbative corrections. In the lower part of the table, the correlation
matrix of the parameters is shown.
The first error is due to all uncertainties taken into account in the fit (experimental error
in the moment measurements, non-perturbative corrections and bias correction to the mo-
ments). The second error is obtained by varying αs in the expressions for the moments and
in Eq. 10. In Eq. 10, we vary αs by ±0.04 around the nominal central value of 0.22. The
last error is a 1.5% uncertainty due to the limited accuracy of the theoretical expression for
the semileptonic width, assessed in Ref. [5].
V. SUMMARY
We have performed a fit to the Belle measured spectral moments of the lepton energy
and hadronic mass spectrum in charmed semileptonic B decays, and the photon energy
spectrum of inclusive radiative B decays using expressions for the moments in terms of
HQE parameters in the 1S mass and kinetic mass schemes. The fits produce values of |Vcb|
12
EPFL LPHE Seminar Oct 2007 Phillip Urquijo 50
Preliminary
Kinetic SchemeBelle Conf 0669, hep-ex/0611047
|Vcb| = ( 41.93 ± 0.65fit ± 0.07αs ± 0.63th ) × 10-3
Fit parameter correlations
mb & mc highly correlated
introducing extra terms in the χ2 function of the fit,
χ2param(mχ, Mχ) =
0 |O| ≤ m3χ ,
(|〈O〉|−m3
χ
)2/M6
χ |O| > m3χ ,
(7)
where (mχ,Mχ) are both quantities of order ΛQCD, and 〈O〉 represents any O(1/m3b) param-
eters. In the fit, we take Mχ = mχ = 500 MeV after Ref [2]. The parameter mχ may have
a value anywhere between 500 MeV and 1 GeV.
The overall form of the χ2 function used in the 1S fit is
χ2 =∑
i,j
(〈X〉meas.i − 〈X〉1Si )cov−1
ij (〈X〉meas.j − 〈X〉1Sj ) +
2∑
i=1
χ2ρi
+3∑
i=1
χ2τi
, (8)
where 〈X〉meas.i are the measured moments and 〈X〉1S
i are the corresponding 1S scheme
predictions.
C. Fit results and discussion
Minimizing the χ2 function in Eq. 8 using MINUIT [15], we find the following results for
the fit parameters,
|Vcb| = (41.49± 0.52fit ± 0.20τB)× 10−3 ,
m1Sb = (4.729± 0.048) GeV/c2 , and
λ1 = (−0.30± 0.04) GeV2 .
The first error is the uncertainty from the fit including experimental and theory errors, and
the second error (on |Vcb| only) is due to the uncertainty on the average B lifetime. The
correlations between these fit parameters are provided in Table II. Using the measurement of
the partial branching fraction at Emin = 0.6 GeV, we obtain for the semileptonic branching
fraction (over the full lepton energy range),
B(B → Xc#ν) = (10.62± 0.25)% .
The measured moments compared to the 1S scheme predictions are shown in Figs. 1 and 2.
We assess the stability of the fit in two ways (Table III): by repeating the fit only to
B → Xc#ν data (21 measurements), (a) to (c) ; and by releasing the mχ constraint on
6
In this analysis, we determine a total of seven parameters: |Vcb|, Λ, λ1, τ1, τ2, τ3 and ρ1.
One of the higher order parameters, τ4 is set to zero , and from available constraints, e.g.
B∗−B mass splitting, the remaining parameters in Eq. 1 are set to: λ2 = 0.1227−0.0145λ1
and ρ2 = 0.1361 + τ2, following advice from Ref. [2]. The parameter Λ is the difference
between the b-quark mass and the reference value about which it is expanded, i.e., Λ =
mΥ(1S)/2−m1Sb . We will present our results in terms of m1S
b in place of Λ.
B. The χ2 function
The fit takes into account both experimental and theoretical uncertainties. Following the
approach in Ref. [2], an element of the combined experimental and theoretical error matrix
is given by
σ2ij = σiσjcij, (3)
where i and j denote the observables and cij is the experimental correlation matrix element.
The total error on the observable i is defined as
σi =√
(σexpi )2 + (Afnm2n
B )2 + (Bi/2)2 for the nth hadron moment , (4)
σi =√
(σexpi )2 + (Afn(mB/2)n)2 + (Bi/2)2 for the nth lepton moment , (5)
σi =√
(σexpi )2 + (Afn(mB/2)n)2 + (Bi/2)2 for the nth photon moment , (6)
and f0 = f1 = 1, f2 = 1/4 and f3 = 1/(6√
3). Here, σexpi are the experimental errors, Bi =
X(16) are the coefficients of the last computed terms in the perturbation series (providing the
error on the uncalculated higher order perturbative terms). The dimensionless parameter A
contains various theoretical errors (uncalculated power corrections, uncalculated effects of
order (αs/4π)Λ2QCD/m2
b , and effects not included in the OPE, i.e., duality violation), and is
multiplied by dimensionful quantities. We fix A = 0.001 as in Ref. [2]. For B → Xsγ, the
accessible phase space is limited, and the theoretical extraction of mb is affected by shape
function effects. So, A is multiplied by the ratio of the difference from the end point relative
to Emin = 1.8 GeV.
As the fit does not provide strong constraints on the O(1/m3b) parameters, it is necessary
to provide constraints to ensure their convergence to sensible values. We achieve this by
8
1S scheme kinetic scheme
n = 0 Emin = 0.6, 1.0, 1.4 n = 0 Emin = 0.4, 0.8
Lepton moments 〈En! 〉Emin n = 1 Emin = 0.6, 0.8, 1.0, 1.2, 1.4 n = 1 Emin = 0.4, 0.8, 1.0, 1.2 1.4
n = 2 Emin = 0.6, 1.0, 1.4 n = 2 Emin = 0.4, 0.8, 1.0, 1.2 1.4
n = 3 Emin = 0.8, 1.2 n = 3 Emin = 0.4, 0.8, 1.0, 1.2 1.4
Hadron moments 〈M2nX 〉Emin n = 1 Emin = 0.7, 1.1, 1.3, 1.5 n = 1 Emin = 0.7, 0.9, 1.1, 1.3
n = 2 Emin = 0.7, 0.9, 1.3 n = 2 Emin = 0.7, 0.9, 1.1, 1.3
Photon moments 〈Enγ 〉Emin n = 1 Emin = 1.8, 2.0 n = 1 Emin = 1.8, 1.9, 2.0
n = 2 Emin = 1.8, 2.0 n = 2 Emin = 1.8, 1.9, 2.0
TABLE I: Experimental input used in the 1S and kinetic scheme analyses. The values of Emin are
given in GeV. The 1S (kinetic) scheme analysis uses a total of 24 (31) measurements.
III. 1S SCHEME ANALYSIS
A. Theoretical input
The inclusive spectral moments of B → Xc!ν decays have been derived in the 1S scheme
up to O(1/m3b) [2]. The theoretical expressions for the truncated moments have the following
form (where 〈X〉Emin represents any of the experimental observables):
〈X〉Emin = X(1) + X(2)Λ + X(3)Λ2 + X(4)Λ3 + X(5)λ1 + X(6)Λλ1 + X(7)λ2 + X(8)Λλ2 + X(9)ρ1
+X(10)ρ2 + X(11)τ1 + X(12)τ2 + X(13)τ3 + X(14)τ4 + X(15)ε + X(16)ε2BLM + X(17)εΛ .
(1)
The coefficients X(k), determined by theory, are functions of Emin. The non-perturbative
corrections are parametrized by Λ (O(mb)), λ1 and λ2 (O(1/m2b)), and τ1, τ2, τ3, τ4, ρ1 and
ρ2 (O(1/m3b)).
Predictions for the partial branching fractions are obtained using the following expression,
B(B → Xc!ν)Emin = 〈X〉B,Emin
ηQED|Vcb|2G2F m5
192π3τB, (2)
where 〈X〉B,Emin is an expression of the form of Eq. 1, m is the 1S reference mass, m =
mΥ(1S)/2, ηQED = 1.007, and G2F m5/(192π3) = 5.4× 10−11.
7
EPFL LPHE Seminar Oct 2007 Phillip Urquijo51
1S Scheme procedureBelle Conf 0669, hep-ex/0611047
Theoretical uncertainties:
Constraints:
mχ~500 MeV
In this analysis, we determine a total of seven parameters: |Vcb|, Λ, λ1, τ1, τ2, τ3 and ρ1.
One of the higher order parameters, τ4 is set to zero , and from available constraints, e.g.
B∗−B mass splitting, the remaining parameters in Eq. 1 are set to: λ2 = 0.1227−0.0145λ1
and ρ2 = 0.1361 + τ2, following advice from Ref. [2]. The parameter Λ is the difference
between the b-quark mass and the reference value about which it is expanded, i.e., Λ =
mΥ(1S)/2−m1Sb . We will present our results in terms of m1S
b in place of Λ.
B. The χ2 function
The fit takes into account both experimental and theoretical uncertainties. Following the
approach in Ref. [2], an element of the combined experimental and theoretical error matrix
is given by
σ2ij = σiσjcij, (3)
where i and j denote the observables and cij is the experimental correlation matrix element.
The total error on the observable i is defined as
σi =√
(σexpi )2 + (Afnm2n
B )2 + (Bi/2)2 for the nth hadron moment , (4)
σi =√
(σexpi )2 + (Afn(mB/2)n)2 + (Bi/2)2 for the nth lepton moment , (5)
σi =√
(σexpi )2 + (Afn(mB/2)n)2 + (Bi/2)2 for the nth photon moment , (6)
and f0 = f1 = 1, f2 = 1/4 and f3 = 1/(6√
3). Here, σexpi are the experimental errors, Bi =
X(16) are the coefficients of the last computed terms in the perturbation series (providing the
error on the uncalculated higher order perturbative terms). The dimensionless parameter A
contains various theoretical errors (uncalculated power corrections, uncalculated effects of
order (αs/4π)Λ2QCD/m2
b , and effects not included in the OPE, i.e., duality violation), and is
multiplied by dimensionful quantities. We fix A = 0.001 as in Ref. [2]. For B → Xsγ, the
accessible phase space is limited, and the theoretical extraction of mb is affected by shape
function effects. So, A is multiplied by the ratio of the difference from the end point relative
to Emin = 1.8 GeV.
As the fit does not provide strong constraints on the O(1/m3b) parameters, it is necessary
to provide constraints to ensure their convergence to sensible values. We achieve this by
8
In this analysis, we determine a total of seven parameters: |Vcb|, Λ, λ1, τ1, τ2, τ3 and ρ1.
One of the higher order parameters, τ4 is set to zero , and from available constraints, e.g.
B∗−B mass splitting, the remaining parameters in Eq. 1 are set to: λ2 = 0.1227−0.0145λ1
and ρ2 = 0.1361 + τ2, following advice from Ref. [2]. The parameter Λ is the difference
between the b-quark mass and the reference value about which it is expanded, i.e., Λ =
mΥ(1S)/2−m1Sb . We will present our results in terms of m1S
b in place of Λ.
B. The χ2 function
The fit takes into account both experimental and theoretical uncertainties. Following the
approach in Ref. [2], an element of the combined experimental and theoretical error matrix
is given by
σ2ij = σiσjcij, (3)
where i and j denote the observables and cij is the experimental correlation matrix element.
The total error on the observable i is defined as
σi =√
(σexpi )2 + (Afnm2n
B )2 + (Bi/2)2 for the nth hadron moment , (4)
σi =√
(σexpi )2 + (Afn(mB/2)n)2 + (Bi/2)2 for the nth lepton moment , (5)
σi =√
(σexpi )2 + (Afn(mB/2)n)2 + (Bi/2)2 for the nth photon moment , (6)
and f0 = f1 = 1, f2 = 1/4 and f3 = 1/(6√
3). Here, σexpi are the experimental errors, Bi =
X(16) are the coefficients of the last computed terms in the perturbation series (providing the
error on the uncalculated higher order perturbative terms). The dimensionless parameter A
contains various theoretical errors (uncalculated power corrections, uncalculated effects of
order (αs/4π)Λ2QCD/m2
b , and effects not included in the OPE, i.e., duality violation), and is
multiplied by dimensionful quantities. We fix A = 0.001 as in Ref. [2]. For B → Xsγ, the
accessible phase space is limited, and the theoretical extraction of mb is affected by shape
function effects. So, A is multiplied by the ratio of the difference from the end point relative
to Emin = 1.8 GeV.
As the fit does not provide strong constraints on the O(1/m3b) parameters, it is necessary
to provide constraints to ensure their convergence to sensible values. We achieve this by
8
Various Perturbative Shape function effects.
A contains: Uncalculated power corrections, uncalculated effects of order and
effects not included in the OPE (i.e. duality) (A=0.001)
Last term in pert. expansion gives error on higher
order pert. terms.
A is scaled up as Emin approaches the B→sγ end
point