LHCb expected physics performance with 10 fb –1

Olivier Schneider

LHCb Upgrade WorkshopEdinburgh, January 11–12, 2006

LHCb expected physics performance with 10 fb–1

[email protected]

Laboratoire de Physique des Hautes Energies (LPHE)

O. Schneider, Jan 11, 2007LHCb Upgrade Workshop, Edinburgh 2

The whole idea Given the fact that

—LHC will be the facility producing the largest number of b hadrons (of all types), by far, and for a long time

—the Tevatron experiments have demonstrated the feasibility of B physics at hadron machines

perform a dedicated B-physics experiment at the LHC,but with a new challenge:—exploit the huge bb production

in the not-well-known forward region, despite the unfriendly hadronic environment (multiplicity, …) for B physics

• ~ 230 b of bb production in one of the forward peaks (400 mrad),corresponding to nearly 105 b hadrons per secondat a low luminosity of 21032 cm–2s–1

bb angular correlation in pp collisions at s = 14 TeV

(Pythia)

(unchanged since more than a decade)


LHCb spectrometer (before this upgrade workshop)LHCb-1 spectrometer (after this upgrade workshop)

VELO

VELO: Vertex Locator (around interaction point) TT, T1, T2, T3: Tracking stations RICH1-2: Ring Imaging Cherenkov detectorsECAL, HCAL: CalorimetersM1–M5: Muon stations

proton beam

proton beam

Dipolemagnet

1.9 < < 4.9 or 15 < < 300 mrad


LHCb pit (in December 2006)


Pileup and luminosity LHC machine, pp collisions at s = 14 TeV:

—design luminosity = L = 1034 cm–2s–1, bunch crossing rate = 40 MHz—average non-empty bunch crossing rate = f = 30 MHz (in LHCb)—Pileup:

• n = number of inelastic pp interactions occurring in the same bunch crossing• Poisson distribution with mean <n> = Linel/f, with inel = 80 mb• <n> = 25 at 1034 cm–2s–1 not good for B physics

At LHCb:—L tuneable by adjusting final beam focusing—Choose to run at <L> ~ 21032 cm–2s–1

(max. 51032 cm–2s–1)• Clean environment: <n> = 0.5• Less radiation damage • Expected to be available from first physics run

—2 fb–1 of data in 107 s (= nominal year)

pp interactions/crossing

LH

Cb

n=0

n=1


MC studies Technical proposal (1998):

— Rough detector description— No trigger simulation— No pattern recognition in tracking— Parametrized PID performance

Re-opt. Technical Design Report (2003)— Final detector design — Simulation of L0 and L1 trigger only— First version of full pattern recognition

“DC04” MC datasets (2004–2005):— Detailed material description— First simulation of High-Level Trigger

“DC06” MC datasets (2006–2007):— “Final” geometry and material description— Redesigned High-Level Trigger— “Final” reconstruction algorithms

REMINDER of important requirements for B physics

—Flexible and efficient trigger • final states with leptons • fully hadronic final states

—Excellent tracking:• Track finding efficiency• Momentum and mass resolution• Vertexing, proper time resolution

—Particle identification (p/K///e)

Background estimates:– based on a sample of inclusive bb

events equivalent to a few minutes of data taking !

– sometimes can only set limits

Today’s numbers: mostly from DC04 MC, at <L> = 21032 cm–2s–1


VELO

TT

T1 T2 T3 RICH2

RICH1

Magnet

PYTHIA+GEANT full simulation

Expected tracking performance High multiplicity environment:

— In a bb event, ~30 charged particles traverse the whole spectrometer

Track finding:— efficiency > 95%

for long tracks from B decays(~ 4% ghosts for pT > 0.5 GeV/c)

— KS+– reconstruction 75% efficient for decay in the VELO, lower otherwise

Average B-decay track resolutions:— Impact parameter: ~30 m — Momentum: ~0.4%

Typical B resolutions:— Proper time: ~40 fs (essential for Bs physics)— Mass: 8–18 MeV/c2

Mass resolution

Bs 18 MeV/c2

Bs Ds 14 MeV/c2

Bs J/ 16 MeV/c2

Bs J/ 8 MeV/c2

* with J/ mass constraint

*


Particle ID performance Average efficiency:

—K id = 88% mis-id = 3%

Good K/ separation in 2–100 GeV/c range—Low momentum

• kaon tagging —High momentum

• clean separation of the different Bd,shh modes

• will be best performance ever achieved at a hadron collider

invariant mass K invariant mass

With PID With PID

invariant mass

No PID


Flavour tagging

Performance assessed on full MC, after trigger and reconstruction Kaon tags are the most powerful, e.g. opposite K (from bcs) All tags combined with neural network Tagging performance depends on how event is triggered !

— will be measured in data using control channels

Tag D2=(1–2w)2

Opposite 0.7%–1.8%

Opposite e 0.4%–0.6%

Opposite K 1.6%–2.4%

Opposite Qvtx 0.9%–1.3%

Same side (B0) 0.8%–1.0%

Same side K (Bs) 2.7%–3.3%

Combined (B0) 4%–5%

Combined (Bs) 7%–9%

Qvtx

BsB0

D

l-K–

K+PVSV


Trigger performance & rates Algorithms and performance:

—Level-0 trigger algorithms mature, 1 MHz output rate—High-Level Trigger (HLT) under development

• Prototype available within time budget for a limited set of channels

—L0*HLT efficiencies:• Determined using detailed MC simulation

• Typically 30%–80% for offline-selected signal events, depending on channel

HLT output rates:—Indicative rates

(split between streams still to be determined)

—Large inclusive streams to be used to control calibration & systematics (trigger, tracking, PID, tagging)

Output rate

Event type Physics

200 Hz Exclusive B candidates

B (core program)

600 Hz High mass di-muons J/, bJ/X (unbiased)

300 Hz D* candidates Charm

900 Hz Inclusive b (e.g. b) B (data mining)


Physics performance vs L

Rough and quick study:—small DC04 MC samples generated at <L> = 51032 cm–2s–1

for a few representative signal channels—backgrounds not investigated yet (but will be possible with DC06 samples)

Preliminary overall conclusion (for L = 2–51032 cm–2s–1):—Significant gain for dimuon channels

• yield L

—“Statu quo” for hadronic channels • yield ~ constant

—Tagging performance seems ~constant (at least for Bs DsK)

same-side Kopp-side K

vertexelectron

muoncombined

L = 21032

L = 51032


Integrated luminosity scenario 2007 (end):

—Short pilot run at 450 GeV per beam with full detector installed—Establish running procedures, time and space alignment of the detectors—Integrated luminosity for physics ~ 0 fb–1

2008:— LHC reaches design energy—Complete commissioning of detector and trigger at s=14 TeV—Calibration of momentum, energy and particle ID—Start of first physics data taking, assume ~ 0.5 fb–1

2009–:—Stable running, assume ~ 2 fb–1/year

Availability of physics results:—with 0.5 fb–1 in ~2009 —with 2 fb–1 in ~2010—with 10 fb–1 in ~2014


Bs +–

Very rare loop decay, sensitive to new physics:—BR ~3.510–9 in SM,

can be strongly enhanced in SUSY—Current 90% CL limit from CDF+D0

with 1 fb–1 is ~20 times SM Main issue is background rejection

—with limited MC statistics, indication that main background is b, b

—assume background is dominated by b, b LHCb expected performance:

—with 0.5 fb–1: exclude BR values down to SM value—with 2 fb–1: 3 evidence of SM signal—with 10 fb–1: > 5 observation of SM signal


Bs +–

LHCb limit on BR at 90% CL (only bkg is observed)

LHCb sensitivity(signal+bkg is observed)

Integrated luminosity (fb–1)

5 observation

3 evidence

SM prediction

BR

(x1

0–9)

Integrated luminosity (fb–1)

BR

(x1

0–9)

Uncertainty in bkg prediction

Expected final CDF+D0 limit

SM prediction


sin(2) with B0J/KS

Expected to be one of the first CP measurements:—Demonstrate (already with 0.5 fb–1) that we can keep under control the main

ingredients of a CP analysis• in particular tagging extraction from control channels

—Sensitivity (from TDR, improved since):

• ~ 216k signal events/2 fb–1, B/S ~ 0.8 stat(sin(2)) = 0.02

With 10 fb–1:—Should be able to reach (sin(2)) ~ 0.010

• to be compared with 0.017 from final BaBar+Belle statistics

—Can also push further the search for direct CP violating term cos(mdt)

AC

P(t

)

€

ACP(t) =N B 0 → J /ψKS( ) − N B0 → J /ψKS( )

N B 0 → J /ψKS( ) + N B0 → J /ψKS( )

background subtracted, 2 fb–1 (toy MC)


BsDs+ sample and Bs mixing

Measurement of ms:— CDF observed Bs oscillations in 2006:

ms =17.77 ± 0.10 ± 0.07 ps–1 [hep-ex/0609040]compatible with SM

— LHCb expectation with 0.5 fb–1:• ~35k Bs Ds

+ signal events with average t ~ 40 fs and Bbb/S < 0.05 at 90% CL stat(ms) = ± 0.012 ps1, i.e. 0.07%

• will be completely dominated by systematics on proper time scale, but at most ((B0))/(B0) = 0.5%

Importance of BsDs+ sample:

— Normalization channel for all Bs branching fraction measurements• First absolute measurement from Belle, BR(BsDs

+) = (0.68 ±0.22 ±0.16)% [hep-ex/0610003], expect soon ~10% measurement

— Control channel for all time-dependent analyses with Bs decays• Measurement of dilution on cos(mst) and sin(mst) terms

— Important step towards measurement of other Bs mixing parameters • e.g. mixing phase or CP violation in mixing

Reconstructed proper time [ps – 1]

QuickTime™ and aTIFF (LZW) decompressor

are needed to see this picture.Ent

ries

per

0.0

2 ps

Full simulation 0.5 fb–1 (signal only, ms = 20 ps – 1)


Bs mixing phase s with bccs s is the strange counterpart of d=2:

s very small in SM s

SM = –arg(Vts2) =–22 = –0.036 ± 0.003 (CKMfitter)

— Could be much larger if New Physics runs in the box

Golden bccs mode is Bs J/:— Angular analysis needed to separate

CP-even and CP-odd contributions— Expect ~130k Bs J/() signal events/2fb–1

(before tagging), S/Bbb= 8

Add also pure CP modes such as J/(’), c, DsDs — No angular analysis needed, but smaller statistics

Combined sensitivity after 10 fb–1:

— dominated by BsJ/— systematics (tagging, resolution) need to be tackled— hopefully >3 evidence of non-zero s, even if only SM

€

W

€

W

€

b

€

Bs0

⎧ ⎨ ⎪

⎩ ⎪

€

s

€

b€

s

€

⎫ ⎬ ⎪

⎭ ⎪ B s

0

€

t

€

t

QuickTime™ and aTIFF (LZW) decompressor

are needed to see this picture.

Statistical sensitivities on s for 2 fb–1

stat(s) = 0.010 [LHCb-2006-047]


Constraints on New Physics in Bs mixing from s measurement

In April 2006, including CDF’s first measurement

of ms >90% CL

>32% CL>5% CL from hep-ph/0604112

After LHCb measurement of s with (s)= ±0.1(~ 0.2 fb–1)

After LHCb measurement of s with (s)= ±0.1(~ 0.2 fb–1)

After LHCb measurement of s with (s)= ±0.03(~ 2 fb–1)

€

New physics in Bs mixing parametrized with hs and σ s : M12 = 1+ hs exp(2iσ s)( ) M12SM

from hep-ph/0604112

courtesy Z. Ligeti

2009

2009

2010


bsss hadronic penguin decays Time-dependent CP analysis of penguin decays to CP eigenstates

B0 KS:— 800 signal events per 2 fb–1, B/S < 2.4 at 90% CL— After 10 fb–1: stat(sin(2eff)) = 0.14— Similar to a B factory experiment

Bs :

— CP violation < 1% in SM (Vts enters both in mixing and decay amplitudes) significant CP-violating phase NP would be due to New Physics

— Angular analysis required— 4k signal events per 2 fb–1 (if BR=1.410–5), 0.4 < B/S < 2.1 at 90%CL— After 10 fb–1: stat(NP) = 0.042


B0 K*0

s = (m)2 [GeV2]

AFB(s), theory

Sensitivity (ignoring non-resonant K evts for the time being)

—7.7k signal events/2fb–1, Bbb/S = 0.4 ± 0.1—After 10 fb–1:

zero of AFB(s) located to ±0.28 GeV2 determine C7

eff/C9eff with 7% stat error (SM)

Suppressed loop decay, BR ~1.210–6

—Forward-backward asymmetry AFB(s) in the rest-frame is sensitive probe of New Physics:• Predicted zero of AFB(s) depends on Wilson

coefficients C7eff/C9

eff

—Other sensitive observables based on transversity angles are accessible (cf A. Golutvin)

AFB(s), fast MC, 2 fb–1

s = (m)2 [GeV2]


Other rare decays B+ K+l+l– decays

/ee ratio equal to 1 in SM:

—New Physics can have O(10%) effect—After 10 fb–1: stat(RK) = 0.043

Radiative decays:—K*:

• ACP < 1% in SM, up to 40% in SUSY• Can measure at <% level

:• No mixing-induced CP asymmetry in SM,

up to 50% in SUSY :

• Right-handed component of photon polarization O(10%) in SM

• Can get 3 evidence down to 15% (10 fb–1)

Hiller & Krüger, hep-ph/0310219

€

RK = dsdΓ(B → Kμμ)

ds4 mμ

2

q max2

∫ dsdΓ(B → Kee)ds

4 mμ2

q max2

∫ =1.000 ± 0.001

Decay 2 fb–1 yield Bbb/S

B+ K+ 3.8k ~1

B+ K+ee 1.9k ~5

Bd K* 35k < 0.7

Bd 40 < 3.5

Bs 9k < 2.4

b 0.75k < 42

b 4.2k < 10

b 2.5k < 18

b 2.2k < 18


Two tree decays (bc and bu), which interfere via Bs mixing:—can determine s + , hence in a very clean way—similar to 2+ extraction with B0 D*, but with the

advantage that the two decay amplitudes are similar (~3) and that their ratio can be extracted from data

from Bs DsK

Bs Ds–+ background

(with ~ 15 larger BR)suppressed using PID: residual contamination

only ~ 10%

Expect 5400 signal events in 2 fb–1

Bbb/S < 1 at 90% CL

(to be updated soon with DC04 MC)

m = 14 MeV/c2

€

s

€

s

€

b

€

c

€

u

€

s

€

Bs0{

€

}Ds−

€

}K+

€

}K–

€

s

€

s

€

b

€

u

€

c

€

s

€

Bs0{

€

}Ds+


Fit the 4 tagged time-dependent rates:—Extract s + , strong phase

difference , amplitude ratio—Bs Ds also used in the fit

to constrain other parameters (mistag rate, ms, s …)

() ~ 13 with 2 fb–1 —expected to be

statistically limited

from Bs DsK

Both DsK asymmetries 10 fb–1, ms = 20 ps–1)

Ds–K+: info on + ( + s)

Ds+K–: info on – ( + s)


from B± DK±

“ADS+GLW” strategy:—Measure the relative rates of B– DK– and B+ DK+ decays with neutral D’s

observed in final states such as: K–+ and K+–, K–+–+ and K+–+–, K+K–

—These depend on:• Relative magnitude, weak phase and strong phase between B– D0K– and B– D0K–

• Relative magnitudes (known) and strong phases between D0 K–+ and D0 K–+,and between D0 K–+–+ and D0 K–+–+

—Can solve for all unknowns, including the weak phase :

—Use of B± D*K± under study

€

u

€

b

€

u

€

s

€

c

€

u

€

B– ⎧ ⎨ ⎩

€

}D 0

€

}K–

colour-suppressed

€

u

€

u

€

b

€

c

€

u

€

s

€

B–{

€

}D0

€

}K–

colour-allowed

Weak phase difference = Magnitude ratio = rB ~ 0.08

Decay 2 fb–1 yield Bbb/S

B– (K–+)D K– 28k ~0.6

B+ (K+–)D K+ 28k ~0.6

B– (K+–)D K– 180 4.3

B+ (K–+)D K+ 530 1.5

() = 5–15 with 2 fb–1


Treat with same ADS+GLW method—So far used only D decays to K–+, K+–, K+K– and +– final states

from B0 D0K*0

€

}D0

€

d

€

b

€

d

€

s

€

c

€

u

€

B0 ⎧ ⎨ ⎩

€

}K*0

colour-suppressed

Weak phase difference = Magnitude ratio = rB ~ 0.4

€

d

€

b

€

d

€

s

€

u

€

c

€

B0 ⎧ ⎨ ⎩

€

}D 0

€

}K*0

colour-suppressed

Decay mode (+cc) 2 fb–1 yield Bbb/S

B0 (K+–)D K*0 3400 <0.3

B0 (K–+)D K*0 500 <1.7

B0 (K+K–, +–)D K*0 600 <1.4

() = 7–10 with 2 fb–1


Sensitivities to from BDK decays

All channels combined (educated guess): () = 4.2º with 2 fb–1

() = 2.4º with 10 fb–1

B mode D mode Method (), 2 fb–1

B+ DK+ K + KK/ + K3 ADS+GLW 5º–15º

B+ D*K+ K ADS+GLW Under study

B+ DK+ KS Dalitz 8º

B+ DK+ KK 4-body “Dalitz”

15º

B+ DK+ K 4-body “Dalitz”

Under study

B0 DK*0 K + KK + ADS+GLW 7º–10º

B0 DK*0 KS Dalitz Under study

Bs DsK KK tagged, A(t) 13º

Signal only, no accept. effect


Sensitivity to SU(2) analysis of B0 +–, ±0, 00:

—Main LHCb contribution could be B0 00

Time-dependent Dalitz plot analysis of B0 +–0 (Snyder & Quinn)

—14k signal events/2fb–1, B/S ~1

00

–+

+–

average

gen

70 expts superimposed (2fb–1)

2 vs Distribution of fit error

stat() < 10º in 90% of the cases (2 fb–1)

15% (< 1%) fake solutions with 2 (10) fb–1


Impact of LHCb on UT

LHCb (2 fb–1, 10 fb–1):— LHCb: (sin(2)) = 0.02, 0.01 () = 4.2º, 2.4º () = 10º, 4.5º

Lattice QCD (2010, 2014):— 40, 1000 Tflop year ()/ = 2.5%, 1.5%

Central values:— SM assumed

(just for illustration)

%8.1/)(

%6.3/)(

==

10 fb–1 (2014)2 fb–1 (2010)

%9.3/)(

%1.7/)(

==

LHCb + LQCD only

From V. Vagnoni, CKM workshop, Dec 2006


from B and BsKK

Penguin decays, sensitive to New Physics Measure CP asymmetry in each mode:

—Adir and Amix depend on mixing phase, angle , and ratio of penguin to tree amplitudes = d ei

Exploit U-spin symmetry (Fleischer):—Assume d=dKK and =KK —4 measurements and 3 unknowns

(taking mixing phases from other modes) can solve for

With 2 fb–1:—36k B, B/S ~ 0.54

36k BsKK, B/S < 0.14—Sensitivity to Adir and Amix

~ twice better than current world average

€

ACP(t) = Adir cos(Δmt) + Amix sin(Δmt) stat() = 4

If perfect U-spin symmetry assumed

stat() = 7–10 + fake solution

If only 0.8<dKK/d<1.2

assumed

2 fb–1

2 fb–1


Charm physics Foresee dedicated D* trigger:

—Huge sample of D0h+h– decays

—Tag D0 or anti-D0 flavor with sign of pion from D*D0

Performance studies not as detailed as for B physics—just started

Interesting (sensitive to NP) & promising searches/measurements:—Time-dependent D0 mixing with wrong-sign D0K+– decays—Direct CP violation in D0K+K–

• ACP 10–3 in SM, up to 1% (~current limit) with New Physics

• Expect stat(ACP) ~ O(10–3) with 2 fb–1

— D0+–

• BR 10–12 in SM, up to 10–6 (~current limit) with New Physics

• Expect to reach down to ~510–8 with 2 fb–1

Potentially usable statistics in 10 fb–1

D* D0(hh) 500M

D*-tagged D0K+K– from b-hadrons

25M

D*-tagged WS D0K+– from b-hadrons

1M


Summary LHCb can chase New Physics in loop decays:

—couple superb highly-sensitive bs observables• Bs, Bs mixing phase

– expect interesting results with 0.5 fb–1 and 2 fb–1 already– can measure down to SM with 10 fb–1 (in case of no New Physics)

—several other exciting windows of opportunity:• Exclusive b sss Penguin decays (limited, even with 10 fb–1)• Exclusive bsll and bs• B hh Penguins• High statistics charm physics

LHCb can improve significantly on from tree decays:—use together with other UT observables to test CKM even more

But …—this is only MC, performance not demonstrated in real life yet

another 2 years to go !—while thinking about upgrade, please make sure LHCb-1 will work

bs

sb

b s


spares


from BDK Dalitz analyses

B± D(KS+–)K±:—D0 and anti-D0 contributions interfere in Dalitz plot—If good online KS reconstruction: 5k signal events in 2 fb–1, B/S < 1—Assuming signal only and flat acceptance across Dalitz plot:

() = 8 with 2 fb–1

B0 D(KS+–)K*0:—Under study

B± D(KK)K±:—Four-body “Dalitz” analysis—1.7 k signal events in 2 fb–1

—Assuming signal only and flat acceptance across Dalitz plot: () = 15 with 2 fb–1


Unitarity triangle in 2014

%7.1/)(

%5.3/)(

==

%8.2/)(

%8.4/)(

==

Without LHCb With LHCb at 10 fb–1

From V. Vagnoni, CKM workshop, Dec 2006

LHCb expected physics performance with 10 fb –1

Documents