Precision measurement of B0 meson oscillation frequency

Basem KhanjiOn behalf of the LHCb collaboration

Milano-Bicocca, INFN, CERN

22-March-2016

LHC Seminar, 22 March 2016

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 1 / 40

CKM matrix

LW ±int = − g√

2(uL, cL, tL)γµVCKM

dL

sL

bL

W+ + h.c

VCKM =

Vud Vus Vub

Vcd Vcs Vcb

Vtd Vts Vtb

b c

W −

Vcb

• |Vij |2: probability of transition from one flavour to another

• Physics states 6= flavour states

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 2 / 40

CKM matrix

VCKM =

Vud Vus Vub

Vcd Vcs Vcb

Vtd Vts Vtb

• Strength of the transition between up-type and down-type quarks

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 3 / 40

CP violation

• VCKM is complex & unitary

• Can be parametrized using 4 parameters: 3 real + 1 complex

VCKM ∼

1− λ2/2 λ Aλ3(ρ− iη)

−λ 1− λ2/2 Aλ2

Aλ3(1− ρ− iη) −Aλ2 1

6= V ∗CKM

• Nature discriminates(!) between matter and anti-matter

• η is small → where did the anti-matter go?

• Test the consistency of CKM picture within SM experimentally

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 4 / 40

Unitarity Triangle

VudV ∗ub + VcdV ∗

cb + VtdV ∗tb = 0

• One example of unitarity triangles

• There are six more

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 5 / 40

Unitarity Triangle

VudV ∗ub + VcdV ∗

cb + VtdV ∗tb = 0

γ

α

α

dm∆ Kε

sm∆ & dm∆

ubV

βsin 2(excl. at CL > 0.95)

< 0βsol. w/ cos 2

α

βγ

ρ

­0.4 ­0.2 0.0 0.2 0.4 0.6 0.8 1.0

η

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

exclu

ded

are

a h

as C

L >

0.9

5EPS 15

CKMf i t t e r

• On the other hand: sides + angles are related to physics observables

• ... most of them are accessible through b-hadron decays

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 6 / 40

How to fish for b-hadrons

• 25% of bb pairs are produced forward

0/4π

/2π/4π3

π

0

/4π

/2π

/4π3

π [rad]1θ

[rad]2θ

1θ

2θ

b

b

z

LHCb MC = 7 TeVs

• → LHCb: forward spectrometer with unique η coverage (2 < η < 5)

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 7 / 40

LHCb detector

Int. J. Mod. Phys. A 30 (2015) 1530022• VELO : σ(IP) ∼ 20µm for high pT tracks• Tracking system : δ(p)/p = (0.5− 1)%, reversible magnet polarity• RICH system : ε(K) ∼ 95%, 5% π → K mis-id probability• Calorimeter : Energy measurement, identify π0,γ• Muon detector : ε(µ)∼ 97%, (1− 3)%, π → µ mis-id probability

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 8 / 40

Luminosity

Run I(2011 + 2012): 3 fb−1 Run II (2015): 300 pb−1

• + increase in bb cross section (7TeV→ 13 TeV)

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 9 / 40

A precise measurement of the B0 meson oscillation frequency,LHCB-PAPER-2015-031 1

1Paper is in preparationB. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 10 / 40

Theory

• Flavour oscillation(mixing) through electroweak interaction in neutral B mesons

�B0 B0

W−

u, c, t

W+

u, c, t

d b

b d

�B0 B0

u, c, t

W± W±

u, c, t

d b

b d

• Physics states |BH,L〉 = p|B0〉 ∓ q|B0〉 (p2 + q2 = 1)

• Mass difference between the two physics states: ∆md = mH −mL

• Flavour at production is the flavour of the state at t = 0

• Flavour at decay the flavour of the state at decay time t

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 11 / 40

Theory

• Box diagram calculation leads to:

∆md =G2Fm2

WMB0

6π2 S0(xt)η2B|V ∗tdVtb|2f 2

B0 BB0

• → ∆md constrains the side of theunitarity triangle

• But needs theory input

• B0-B0 system is described by Schrödinger equation

i ddt

(|B0(t)〉|B0(t)〉

)=(

M̂d − i2 Γ̂d)( |B0(t)〉

|B0(t)〉

)

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 12 / 40

Theory

• Mixing probability for neutral B0 mesons:

P(t) ∝ e−tτ (1 + qmixing cos(∆md t))

• Unmixed events: qmixing = 1flavour at production = flavour atdecay

• Mixed events: qmixing = −1flavour at production 6= flavour atdecay

0 1 2 3 4

Inte

nsity

0

0.5

1

y = 0.997x = -0.946: 0K(a)

0 1 2 3 4

-610

-410

-210

1

y = 0.0075x = 0.0063: 0D(b)

tΓ0 1 2 3 4

Inte

nsity

0

0.5

1

y = 0.005x = 0.773: 0B(c)

tΓ0 1 2 3 40

0.5

1

y = 0.046x = 25.194: sB(d)

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 13 / 40

State of the art

0.4 0.45 0.5 0.55

∆md (ps

-1)

World averagefor PDG 2015

0.510 ± 0.003 ps-1

CLEO+ARGUS(χ

d measurements)

0.498 ± 0.032 ps-1

Average of aboveafter adjustments

0.510 ± 0.003 ps-1

LHCb *

(3 analyses) 0.514 ± 0.005 ± 0.003 ps

-1

BELLE *

(3 analyses) 0.509 ± 0.004 ± 0.005 ps

-1

BABAR *

(4 analyses) 0.506 ± 0.006 ± 0.004 ps

-1

D0 (1 analysis)

0.506 ± 0.020 ± 0.016 ps-1

CDF1 *

(4 analyses) 0.495 ± 0.033 ± 0.027 ps

-1

OPAL (5 analyses)

0.479 ± 0.018 ± 0.015 ps-1

L3 (3 analyses)

0.444 ± 0.028 ± 0.028 ps-1

DELPHI *

(5 analyses) 0.519 ± 0.018 ± 0.011 ps

-1

ALEPH (3 analyses)

0.446 ± 0.026 ± 0.019 ps-1

* HFAG average

without adjustments

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 14 / 40

Main components

• Need: "flavour specific" decays with high yields• B0→ D−µ+νµ (D−→ K+π−π−): ∼ 1.6M events (full Run I data: 3 fb−1)• B0→ D∗−µ+νµ (D∗− → D0(→ K+π−)π−): ∼ 0.8M events (full Run I data: 3 fb−1)• µ charge is correlated with B0 flavour at decay

• Need: Flavour Tagging (qmixing)

• Need: Decay time reconstruction (t)

B0 → D−µ+νµ B0 → D∗−µ+νµ

B0

D-

π-

K+

μ+

υμ

π-

B0

D*-

D0

π-

K+

μ+

υμ

π-

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 15 / 40

Background: combinatorics

• First type of backgrounds: Combinatorics, D0 from B decays• Not sharing D(∗)− with the signal in their final state

• Apply sPlot technique to subtract those backgrounds (Nucl. Instrum. Meth. A555 (2005) 356)

• B0→ D∗−µ+νµ: mD0 , δm(= mD∗− −mD0 ) distributions• B0→ D−µ+νµ: mD− distribution

]2c [MeV/π πK m1800 1850 1900

)2 cE

vent

s /

( 1.

4 M

eV/

50

100

310× Data Total fit Signal Comb.

LHCb

]2c [MeV/mδ140 145 150 155

)2 cE

vent

s /

( 0.

125

MeV

/

10

20

30

40

310×

LHCb Data Total fit

*- D0 D

Comb.

LHCB-PAPER-2015-031 (preliminary)

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 16 / 40

Background: B+ background

• B+→ D(∗)−µ+π+νµX decays: share same characteristics of signal but do not exhibitany oscillation

• Fraction of these B+→ D(∗)−µ+π+νµX decays correlated with ∆md• Need to fix it from somewhere

• B of B+→ D(∗)−µ+π+νµX has 10% uncertainty → could amplify systematic effect

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 17 / 40

Closer look at B+ background

• B+→ D∗−µ+π+νµX(Left), B0→ D∗−µ+νµ(Right)

PV

B+D*-

D~0K+

pi-

pi-

mu+

pi+

nu

PV

B0D*-

D~0K+

pi-

pi-

mu+

nu

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 18 / 40

B+ background: Isolation variables

• Conduct a search over all tracks to find if at least one of them originates from yourB candidate

• If your candidate is signal you will find nothing• If your candidate is background you will find one

PV

B+D*-

D~0K+

pi-

pi-

mu+

pi+

nu

PV

B0D*-

D~0K+

pi-

pi-

mu+

nu

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 19 / 40

B+ background: busted!

iso(D)-0.5 0 0.5 1

Ent

ries

(arb

. uni

ts)

Signal

Background

iso(all)-0.5 0 0.5 1

LHCb

iso(B)-0.5 0 0.5 1

iso(sum)-3 -2 -1 0

Ent

ries

(arb

. uni

ts)

]2c [MeV/corrm4000 6000 8000 10000

) [mm]D, πIP(5 10

LHCB-PAPER-2015-031 (preliminary)

• Merge these quantities into one MVA classifier to maximize the separationB. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 20 / 40

B+ background

• Apply a cut on the MVA classifier → reduce B+→ D(∗)−µ+π+νµX by ∼ 70%

-1 -0.5 0 0.5 1

Even

ts /

0.05

5

10

310×

LHCb (e)

-1 -0.5 0 0.5 10

5

10

15

20

310×

DataTotal fit

signal0B bkg.+B

(g)-1 -0.5 0 0.5 1

Even

ts /

0.05

5

10

310×

(f)

BDT output-1 -0.5 0 0.5 1

0

5

10

15

20

310×

(h)

3 10×

LHCB-PAPER-2015-031 (preliminary)

• Bonus: use the MVA output to estimate the contribution of B+→ D(∗)−µ+π+νµX• → avoid large uncertainty(∼ 10%) from B of B+→ D(∗)−µ+π+νµX in PDG

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 21 / 40

Flavour tagging: qmixing

• Mixing state qmixing: flavour at decay × flavour at production = ±1• Flavour at decay in B0→ D(∗)−µ+νµ is determined by µ charge• Flavour at production?

3/20 Ulrich Eitschberger | Updates on Flavour Tagging | 72nd LHCb week | June 19th, 2014

Flavour Tagging: Determine B production flavours SS Pion SS Kaon Signal Decay

Same Side

Opposite Side

OS Vertex Charge OS Muon OS Electron

OS Kaon

PV

• Flavour at production (qprod : ±1, 0), Tagging efficiency (ε): how often we tag andMistag probability (ω): how often we make a mistake

• Tagging Power = ε× (1− 2ω)2

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 22 / 40

Flavour tagging: qmixing

• Lepton taggers are clean but rare

• Cascade and vertex taggers are abundant but less pure

ω

Eur. Phys. J. C72 (2012) 2022

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 23 / 40

Flavour tagging: qmixing

[%]2)ω (1-2∈0 1 2 3 4 5 6 7 8

)d

m∆(

σ

0.002

0.004

0.006

0.008

0.01

0.012

0.014

0.016

• Sensitivity scales with 1/√ε× (1− 2ω)2

• Higher tagging power is equivalent to higher luminosityB. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 24 / 40

Flavour tagging: qmixing

• Mixing state qmixing: flavour at decay × flavour at production = ±1

N±(t) ∝ e−tτ (1 + qmixing(1− 2ω) cos(∆md t))

• Events are Grouped in 4 categories in increasing mistag probability• This increase the tagging power → increases sensitivity to ∆md• Tagging power for B0→ D(∗)−µ+νµ ∼ 2.5%

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 25 / 40

Determination of B0 decay time: Problem

t =LB0 MB0

pB0

• Missing neutrino → pB0 cannot be known in data (only in simulation)• Decay time accessible in data is:

tdata =LB0 MB0

pDµ

B0 → D−µ+νµ B0 → D∗−µ+νµ

B0

D-

π-

K+

μ+

υμ

π-

B0

D*-

D0

π-

K+

μ+

υμ

π-

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 26 / 40

Determination of B0 decay time: Solution

• To correct the decay time in data per event:• Get correction (k) from simulation:

k =pDµ

pB0

• Improve the correction of the momentum: k-factor as a function of visible mass

tcorrected = tdata × k(MDµ)

]2c mass [MeV/µD*B3500 4000 4500 5000

-facto

rk

0.30.40.50.60.70.80.9

11.11.2 )2 c

Even

ts/(0

.004

)/(11

MeV

/

0

5

10

15

20

25

30

35LHCb

LHCB-PAPER-2015-031 (preliminary)B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 27 / 40

Determination of B0 decay time: Resolution

• Detector resolution on the flight distance of the B0 (R(L))

• k → k(mDµ) → additional resolution function (F (k))

• Momentum resolution introduces the largest effect in time resolution

N±(t) ∝ e−tτ (1 + qmixing(1− 2ω) cos(∆md t))⊗ R(L)⊗ F (k)

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 28 / 40

Tagged time-dependent fit

• Use binned likelihood sFit (arXiv:0905.0724) to extract ∆md

P(t, qmixing) = (1− fB+ )S(t, qmixing) + fB+B+(t, qmixing)

• S(t, q),B+(t, q): decay rates of signal & background• fB+ : fraction of the B+→ D(∗)−µ+π+νµX in the sample

t [ps]5 10 15

Even

ts / (

0.14

7 ps

)

0

10

20

30

40

50

310×

DataTotal fit

signal0B bkg.+B

LHCb

t [ps]5 10 15

Even

ts / (

0.14

7 ps

)

0

5

10

15

20

25

30

35310×

DataTotal fit

signal0B bkg.+B

LHCb

B0 → D−µ+νµ B0 → D∗−µ+νµ

LHCB-PAPER-2015-031 (preliminary)B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 29 / 40

Time dependent asymmetries

5 10

-0.5

0

0.5 LHCb

(a)

5 10

-0.5

0

0.5

(c)

5 10

-0.5

0

0.5

(b)

5 10

-0.5

0

0.5

(d)

)t(A

[ps]t

5 10

-0.5

0

0.5 LHCb

(e)

5 10

-0.5

0

0.5

(g)

5 10

-0.5

0

0.5

(f)

5 10

-0.5

0

0.5

(h)

)t(A

[ps]t

B0 → D−µ+νµ B0 → D∗−µ+νµ

LHCB-PAPER-2015-031 (preliminary)

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 30 / 40

Systematic uncertainties

• Several sources of systematic uncertainties• k-factor method, B+ and other background sources & Other sources (time acceptance,

detector resolution, momentum and length scale)

• Large number of parameterized simulation to evaluate systematic uncertainties

Source of uncertaintyB0→ D−µ+νµ [ns−1] B0→ D∗−µ+νµ [ns−1]

Uncorrelated Correlated Uncorrelated CorrelatedB+ background 0.4 0.1 0.4 –Other backgrounds – 0.5 – –k-factor distribution 0.4 0.5 0.3 0.6Other fit-related 0.5 0.4 0.3 0.5Total 0.8 0.8 0.6 0.8

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 31 / 40

... and ∆md is

• Fits performed separately for per channel and per year

• Combination across the two channel using weighted average

• Results per decay channel (LHCB-PAPER-2015-031):

Mode 2011 sample 2012 sample Total sample∆md [ ns−1] ∆md [ns−1] ∆md [ns−1]

B0→ D−µ+νµ 506.2± 5.1stat 505.2± 3.1stat 505.5± 2.7stat ± 1.1syst

B0→ D∗−µ+νµ 497.5± 6.1stat 508.3± 4.0stat 504.4± 3.4stat ± 1.0syst

• Combination of ∆md measurement across the two channels using Run I data:

∆md = 505.0± 2.1 (stat)± 1.0 (syst) ns−1(preliminary)

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 32 / 40

Theory catching up

∆md =G2Fm2

WMB0

6π2 S0(xt)η2B|V ∗tdVtb|2f 2

B0 BB0

• arXiv:1602.03560: Fermilab Lattice and MILC Collaborations

• Improvements on uncertainties related to hadronic matrix elements:

f 2B0 BB0 = 0.0526± 0.0042GeV/c2

∆md (Theory) = 0.639(50)(36)(5)(13) ps−1

∆md/∆ms(Theory) = 0.0323(9)(9)(0)(3)

• PDG 2015: ∆md = (0.510± 0.003) ps−1, ∆ms = (17.757± 0.021) ps−1

• Theory prediction for:∆md , ∆md/∆ms are 2.1, 2.9σ away from experiment

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 33 / 40

Conclusion

• LHCb measure ∆md inB0→ D(∗)−µ+νµ channels using 3 fb−1

∆md = 505.0± 2.1 (stat)± 1.0 (syst) ns−1

(preliminary)

• LHCb provides the most precisemeasurement of ∆md

• Better than the world average:(510± 3) ns−1(PDG 2015)

• Recent development from the lattice:∆md prediction is 2.1σ away fromworld average

]-1 [nsdm∆450 500 550

ALEPH (3 analyses) 19± 6 ±446

DELPHI (5 analyses) 11± 18 ±519

L3 (3 analyses) 28± 28 ±444

OPAL (5 analyses) 15± 18 ±479

CDF1 (4 analyses) 27± 33 ±495

D0 (1 analysis) 16± 20 ±506

BABAR (4 analyses) 4± 6 ±506

BELLE (3 analyses) 5± 4 ±509

LHCb (3 analyses) 3± 5 ±514

<0.66) +BABAR (P 0.25 + 0.26 - 0.25±4.15

This measurement 1.0± 2.1 ±505.0

BABAR (p*>1.3GeV) 0.27 + 0.19 - 0.21±4.32

Average (w/o this meas.) 3±510

HFAGSummer 2014

Bosch, Lange, Neubert and Paz (BLNP)Phys.Rev.D72:073006,2005

/dof = 9.2/11 (CL = 61.00 %)2χ

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 34 / 40

Backups

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 35 / 40

Time projections in 2012 data

t [ps]5 10 15

Eve

nts

/ (0.

147

ps)

0

10

20

30

40

50

310×

DataTotal fit

signal0B bkg.+B

LHCb

t [ps]5 10 15

Eve

nts

/ (0.

147

ps)

0

5

10

15

20

25

30

35310×

DataTotal fit

signal0B bkg.+B

LHCb

LHCB-PAPER-2015-031 (preliminary)

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 36 / 40

D0 mass distributions for 2012

]2c [MeV/πK m1840 1860 1880 1900

)2 cE

vent

s / (

0.8

MeV

/

0

2

4

6

8

10

310×

LHCb Data Total fit

*- D0 D

Comb.

]2c [MeV/πK m1840 1860 1880 1900

)2 cE

vent

s / (

0.8

MeV

/

0

5

10

15

20

25

310×

LHCb Data Total fit

*- D0 D

Comb.

LHCB-PAPER-2015-031 (preliminary)

• D0 mass fits in B0→ D∗−µ+νµ decay

• Components with D∗−D0π can be only distinguished using δm(= mD∗− −mD0 )

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 37 / 40

Asymmetries in 2011

5 10

-0.5

0

0.5 LHCb

(a)

5 10

-0.5

0

0.5

(c)

5 10

-0.5

0

0.5

(b)

5 10

-0.5

0

0.5

(d)

)t(A

[ps]t

5 10

-0.5

0

0.5 LHCb

(a)

5 10

-0.5

0

0.5

(c)

5 10

-0.5

0

0.5

(b)

5 10

-0.5

0

0.5

(d)

)t(A

[ps]t

LHCB-PAPER-2015-031 (preliminary)

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 38 / 40

MVA classifier

-1 -0.5 0 0.5 1

Even

ts /

0.0

33

5

10

15

20

310×

LHCb (a)

-1 -0.5 0 0.5 10

50

100

310×

DataTotal fit

signal0B bkg.+B

(c)-1 -0.5 0 0.5 1

Even

ts /

0.0

33

5

10

15

20

310×

(b)

BDT output-1 -0.5 0 0.5 1

0

50

100

310×

(d)

3 10×

-1 -0.5 0 0.5 1

Even

ts /

0.0

5

5

10

310×

LHCb (e)

-1 -0.5 0 0.5 10

5

10

15

20

310×

DataTotal fit

signal0B bkg.+B

(g)-1 -0.5 0 0.5 1

Even

ts /

0.0

5

5

10

310×

(f)

BDT output-1 -0.5 0 0.5 1

0

5

10

15

20

310×

(h)

3 10×

B0→ D−µ+νµ B0→ D∗−µ+νµ

LHCB-PAPER-2015-031 (preliminary)

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 39 / 40

0.35 0.4 0.45 0.5 0.55 0.6 0.65

∆md (ps

-1)

World averageSummer 2015

0.5055 ±0.0020 ps-1

CLEO+ARGUS(χ

d measurements)

0.498 ±0.032 ps-1

Average of 32 above 0.5055 ±0.0020 ps-1

LHCb D(*)

µ/OST(3 fb

-1, prel)

0.5036 ±0.0020 ±0.0013 ps-1

LHCb Dµ/OST,SST(1 fb

-1)

0.503 ±0.011 ±0.013 ps-1

LHCb B0d(full)/OST,SST

(1 fb-1

)0.516 ±0.005 ±0.003 ps

-1

LHCb B0d(full)/OST(0.036 fb

-1)

0.499 ±0.032 ±0.003 ps-1

BABAR D*lν/l,K,NN(23M BB

−)

0.492 ±0.018 ±0.013 ps-1

BABAR D*lν(part)/l

(88M BB−

)0.511 ±0.007 ±0.007 ps

-1

BELLE B0d(full)+D

*lν/comb

(152M BB−

)0.511 ±0.005 ±0.006 ps

-1

BELLE l/l(32M BB

−)

0.503 ±0.008 ±0.010 ps-1

BELLE D*π(part)/l

(31M BB−

)0.509 ±0.017 ±0.020 ps

-1

BABAR l/l(23M BB

−)

0.493 ±0.012 ±0.009 ps-1

BABAR B0d(full)/l,K,NN

(32M BB−

)0.516 ±0.016 ±0.010 ps

-1

D0 D(*)

µ/OST(02-05)

0.506 ±0.020 ±0.016 ps-1

CDF1 D*l/l

(92-95)0.516 ±0.099

+0.029 ps

-10.516 ±0.099

-0.035

CDF1 l/l,Qjet(94-95)

0.500 ±0.052 ±0.043 ps-1

CDF1 µ/µ(92-95)

0.503 ±0.064 ±0.071 ps-1

CDF1 Dl/SST(92-95)

0.471 +0.078

±0.034 ps-1

0.471 -0.068

OPAL π*l/Qjet

(91-00)0.497 ±0.024 ±0.025 ps

-1

OPAL D*/l

(90-94)0.567 ±0.089

+0.029 ps

-10.567 ±0.089

-0.023

OPAL D*l/Qjet

(90-94)0.539 ±0.060 ±0.024 ps

-1

OPAL l/Qjet(91-94)

0.444 ±0.029 +0.020

ps-1

0.444 ±0.029 -0.017

OPAL l/l(91-94)

0.430 ±0.043 +0.028

ps-1

0.430 ±0.043 -0.030

L3 l/l(IP)(94-95)

0.472 ±0.049 ±0.053 ps-1

L3 l/Qjet(94-95)

0.437 ±0.043 ±0.044 ps-1

L3 l/l(94-95)

0.458 ±0.046 ±0.032 ps-1

DELPHI vtx(94-00)

0.531 ±0.025 ±0.007 ps-1

DELPHI D*/Qjet

(91-94)0.523 ±0.072 ±0.043 ps

-1

DELPHI l/l(91-94)

0.480 ±0.040 ±0.051 ps-1

DELPHI π*l/Qjet

(91-94)0.499 ±0.053 ±0.015 ps

-1

DELPHI l/Qjet(91-94)

0.493 ±0.042 ±0.027 ps-1

ALEPH l/l(91-94)

0.452 ±0.039 ±0.044 ps-1

ALEPH l/Qjet(91-94)

0.404 ±0.045 ±0.027 ps-1

ALEPH D*/l,Qjet

(91-94)0.482 ±0.044 ±0.024 ps

-1

Heavy FlavourAveraging Group

∆md (world average 2015) = 505.5± 2.0 ns−1

B. Khanji, LHCb (Milano-Bicocca, INFN, CERN) ∆md at LHCb 22-March-2016 40 / 40