Search for dark boson in the decayB+ → K+χ(→ µ+µ−) at LHCb

Andrea Mauri

University of Zurich

PHD seminar, PSI, 27 August 2015

Motivation

• There is a long list of theoretical models that predict the existenceof new particles that couple to the SM sector by mixing with theHiggs(Hχ

)=

(cos θ − sin θsin θ cos θ

)(H′

χ′

)

�

H χ

• Inflaton[1], axion-like[2], dark matter mediator[3] models alsopredict the new boson to be light.

[1] JHEP 1005:010,2010

[2] Phys.Rev.D81:034001,2010

[3] Phys. Lett. B727 (2013) 506-510

2 of 15

The LHCb detector

1. LHCb is a forward spectrometer placed at LHC

◦ Pseudorapidity range: 2 < η < 5

◦ focused on the study of b and c decays

◦ 3 fb−1 of data collected in 2011 and 2012

3 of 15

Why at LHCb?

Main production via B meson:

If χ mixes with the Higgs and it is light enough

Γ (K → πχ) ∝ (m2t | V ∗tsVtd |)2 ∝ m4

t λ5

Γ (D → πχ) ∝ (m2b | V

∗cbVub |)2 ∝ m4

bλ5

Γ (B → Kχ) ∝ (m2t | V ∗tsVtb |)2 ∝ m4

t λ2

• Looking for dark boson decaying into muons

B+ → K+χ(→ µ+µ−)

�B+ K+

χ

W−

b̄

u

µ+

µ−

s̄

u

t̄

• B(χ→ µ+µ−):◦ dominant till the hadronic threshold (χ→ 2h, χ→ 3h)◦ always significant O(10−2) in the full mass range

4 of 15

Signal properties

Depending on the coupling to the SM/hidden sector, we can identifytwo lifetime regimes: [

detector resolutionστ ∼ 0.2 ps

]

Long lifetime: Short lifetime:

◦ Inflaton [JHEP1005(2010)010] ◦ Dark matter mediator [Phys.Lett.B727(2013)]

◦ Axion(like) [Phys.Rev.D81(2010)034001]

◦ Displaced vertex◦ Almost background free

◦ Prompt decay◦ Contamination from SM background

◦ Lower reconstruction efficiency

5 of 15

B+ → K+(χ→ µ+µ−): motivation

• Benchmark model: Inflaton [1]

◦ τχ = 10−8 ÷ 10−10 s,

◦ mχ < O(1 GeV),

◦ B(B → Kχ) ∼ 10−6

◦ natural width � detector resolution

◦ effective coupling to SM particles:

• gYmfvEW

, gY ≡ sin θ

Interesting parameter values,in the range we can test!

• All models predict dark boson with width ∼ 0.

Existing experimental limit

[JHEP 1503 (2015) 171]

Similar analysis in BaBar [1] and

Belle [2] with B → Kµµ decay

[1] Phys.Rev.D86(2012)032012

[2] Phys.Rev.Lett.103(2009)171801

6 of 15

The experimental search ...

7 of 15

Status of the analysis

1. The analysis is still blind

2. Several MC simulations are used

3. Background has been studied intensively

4. Sensitivity study currently underway to define the final selection.

8 of 15

Selection

1. Triggered on muons

2. Multivariate selection:◦ input variables include:• tracks quality reconstruction,• primary, B and di-muon quality vertex reconstruction,• B transverse momentum,• di-muon distance of closest approach,• muon and χ candidate isolation

3. Optimized maximizing Punzi figure-of-merit P = S52+√B

4. Factorize lifetime into two components: L = Lprompt · Ldisplaced◦ prompt• irreducible SM background B+ → K+µ+µ−

◦ displaced• (almost) background free

9 of 15

Signal simulation

• Unknown mass and lifetime◦ We tried to cover the mass & lifetime plane in a smart way

τχ (ps) mχ (MeV)

1 250010 2500

100 250 500 800 1000 1500 2000 2500 3000 3500 4000 45001000 2500

• For each mass different lifetimes can be reweighted analytically◦ the procedure is tested on the mχ = 2500 MeV samples.

(ps)χτ1 10 210

or

igin

alsi

gnal

ε/re

wei

ghte

dsi

gnal

ε

0.9

0.92

0.94

0.96

0.98

1

1.02

1.04

1.06

1.08

1.1

original

= 100 psχτreweighted from

10 of 15

Signal simulation: resolution & efficiency

• Total efficiency:

εsignal = εacc×εtrigger×εreco×εPID

• Di-muon mass resolution

)2 (MeV/cχm0 1000 2000 3000 4000

(%

)si

gnal

ε

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

(MeV)χm0 2000 4000

(M

eV)

χσ

2

4

6

8

1012

14

16

18

2022

11 of 15

Strategy of the search: prompt bin

Looking for di-muon resonance:

• Total invariant mass K+µ+µ− required to fall into the B window◦ mB ∈ [5230, 5330] MeV

• B+ mass is constrained to improved the di-muon resolution

• Scan mtest in steps of 1/2 σχ◦ 2mµ < mtest < mB −mK

12 of 15

Background

Background from B decays:

1. SM B+ → K+µ+µ−: irreducible!◦ main background in the prompt bin

2. SM di-muon resonances: φ, J/ψ, ψ(2S), ψ(3770): vetoed◦ B+ → K+J/ψ(→ µ+µ−)

3. mis-identified B decays:

◦ B+ → K+D0(→ Kπ)◦ B+ → π+D0(→ Kπ)

◦ B+ → 3h (KKK ,KKπ,Kππ, πππ): negligible

◦ B0 → K∗(→ Kπ)µ+µ−,one missed track: negligible

}rejected by more stringent PID cuts

Combinatorial background:

◦ does not peak in the signal region (B window [5230, 5330] MeV)◦ strongly reduced by the BDT selection◦ can affect the displaced region

13 of 15

Sensitivity

Expected sensitivity of the search for a selection with τdisplaced > 1 ps∗

Note: different y-axis range.

∗ The choise of the threshold at 1 ps is a pure example, the factorization in

prompt/displaced bins still has to be optimized.14 of 15

Conclusion

1. A search for a dark boson in the decay channelB+ → K+(χ→ µ+µ−) is studied◦ resolution, efficiency and background contamination are well

understood

2. LHCb is expected to improve (up to one order of magnitude) thelimit of the B factories in the allowed mass range.

3. Looking forward to completing the selection and to looking atevents in the (blinded) signal region.

15 of 15

Backup

16 of 15

Control channel: B0 → J/ψKS

As usual in LHCb analysis, we require a control channel to correctand validate the MC simulation of the signal:

• Same topology (displaced vertex)except for the final number of tracks

• well known branching ratio:BR(B0 → (J/ψ → µµ)(K0

S → ππ)) = 1.8 · 10−5

• ∼ 28.5 · 103 events observed at LHCb

Data-driven correction of the MC efficiency:

εsig = εsigMC

εnormDATA

εnormMC

17 of 15

Control channel: B0 → J/ψKSData/MC comparison

B0_DIRA_OWNPV0.99990.999920.999940.999960.99998 1 1.000021.000041.000061.000081.0001

0

0.1

0.2

0.3

0.4

0.5

0.6

data

MC

B0_IPCHI2_OWNPV0 2 4 6 8 10 12 14 16 18 20

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

data

MC

B0_ENDVERTEX_CHI20 5 10 15 20 25

0

0.02

0.04

0.06

0.08

0.1

data

MC

KS0_D1_isolation650 2 4 6 8 10 12 14 16 18 20

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7data

MC

B0_FDCHI2_OWNPV0 5000 10000 15000 20000 25000 30000

0

0.05

0.1

0.15

0.2

0.25 data

MC

KS0_DIRA_OWNPV0.99990.999920.999940.999960.99998 1 1.000021.000041.000061.000081.0001

0

0.1

0.2

0.3

0.4

0.5

0.6

data

MC

18 of 15

Control channel: B0 → J/ψKSData/MC comparison

KS0_doca0 0.05 0.1 0.15 0.2 0.25

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

data

MC

muplus_IPCHI2_OWNPV0 2000 4000 6000 8000 10000 12000 14000

0

0.1

0.2

0.3

0.4

0.5

data

MC

KS0_ISOLATION_BDT_Hard-2 -1.5 -1 -0.5 0 0.5 1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

data

MC

piplus_IPCHI2_OWNPV0 5000 10000 15000 20000 25000

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

data

MC

KS0_PT0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

0

0.02

0.04

0.06

0.08

0.1

data

MC

piplus_PT0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

data

MC

Very good agreement!!

19 of 15