�

Search for Exotics with LHCb Bastien Muster

on behalf of the LHCb collaboration

Theoretical motivation Several models for Physics beyond the Standard Model predict the existence of new massive particles decaying into Standard Model particles. Two examples : • neutralinos in mSUGRA with baryon or lepton number violation • neutral π0ν particles in the Hidden Valley (HV) framework with a non-abelian gauge symmetry (benchmark model)

Massive particles with a lifetime between 1 ps and 1 ns can be identified by their displaced decay vertex. In particular, π0ν candidates are identified by two hadronic jets originating from a displaced vertex.

LHCb has a unique coverage for long-lived particles with small mass and lifetime

Results The shape of the background invariant mass distribution, the selection efficiency and the background yield depend on the radial displacement Rxy. Therefore, the signal yield is extracted by a fit of the invariant mass distribution in five bins of Rxy (0.4-4.8 mm) and with a minimal mvtx.

Event selection and analysis Search for π0ν with a mass between 25 and 50 GeV/c2 and a lifetime between 1 and 200 ps for a SM-like Higgs particle mass of 120 GeV/c2. Measurement performed with 0.62 fb-1 of data collected in 2011.

Displaced vertex (DV) selection: • DV are identified by combining charged tracks. • DV are selected with a minimal distance Rxy from the beam line and a minimal invariant mass mvtx

• vertices are rejected when reconstructed in the detector material (Matter Veto MV)

• at least 6 tracks per vertex, with a sum of the scalar pT of tracks > 3 GeV/c

Jet reconstruction: Anti-kT algorithm with a cone size of 0.7. • two jets with pT > 5 GeV/c •Only candidates with mcorr/m > 1/0.7 and a distance between the two jets of ΔR < 2.2 are selected

Hidden Valley

Hidden sector at a low mass scale that is weakly coupled to the SM gauge sectorthrough heavy mediators [Strassler, Zurek]

Long lifetime of the unstable ⇡v , due to the high mass of the mediator Z 0 and theweak coupling to SM

πv0

πv0

h ϕ

b

b

Z’

b

b

Z’

Long-lived particles can be produced inpairs through Higgs decay, e.g.H ! ⇡v⇡v ! qq̄qq̄

SM

LEPhiddenvalley

LHC

Other options:

R-parity violating MSSM [De Campos, Eboli, et al]

mSUGRA with baryon number violation [Carpenter, Kaplan, Rhee]

More examples in talk by Matt Strassler at the implications workshop 2013.

Veerle Heijne (Nikhef) Long-Lived Exotics October 2014 3 / 25

Theory

Hidden Valley

Hidden sector at a low mass scale that is weakly coupled to the SM gauge sectorthrough heavy mediators [Strassler, Zurek]

Long lifetime of the unstable ⇡v , due to the high mass of the mediator Z 0 and theweak coupling to SM

πv0

πv0

h ϕ

b

b

Z’

b

b

Z’

Long-lived particles can be produced inpairs through Higgs decay, e.g.H ! ⇡v⇡v ! qq̄qq̄

SM

LEPhiddenvalley

LHC

Other options:

R-parity violating MSSM [De Campos, Eboli, et al]

mSUGRA with baryon number violation [Carpenter, Kaplan, Rhee]

More examples in talk by Matt Strassler at the implications workshop 2013.

Veerle Heijne (Nikhef) Long-Lived Exotics October 2014 3 / 25

Theory

Long-lived particles can be produced in pairs through Higgs decay H-> π0ν π0ν, π0ν ->qqbar

PV

θ

DVd

p

Use the pointing of the candidate to thePV, with corrected massmcorr

m =q

1 +� p

m sin ✓�2

+ pm sin ✓

Found that m/mcorr is more independentof mass and boost than ✓ itself

Require m/mcorr > 0.7

Veerle Heijne (Nikhef) Long-Lived Exotics October 2014 13 / 25

Selection using jets - pointing

PV

θ

DVd

p

Use the pointing of the candidate to thePV, with corrected massmcorr

m =q

1 +� p

m sin ✓�2

+ pm sin ✓

Found that m/mcorr is more independentof mass and boost than ✓ itself

Require m/mcorr > 0.7

Veerle Heijne (Nikhef) Long-Lived Exotics October 2014 13 / 25

Selection using jets - pointing

Data compatible with the expected background from bbbar production

L0:L0Hadron |L0Electron | L0Photon | L0Muon(89% efficient on offline selected signal)

HLT1:Hlt1TrackAllL0 | Hlt1TrackPhoton | Hlt1TrackMuon(94% efficient on offline selected signal)! efficiency drops above Rxy > 5mm due to VELO segmentrequirement

HLT2:Hlt2DVSingleHighMassPS | Hlt2DVSingleHighFDPS |Hlt2Topo2BodyBBDT | Hlt2Topo3BodyBBDT |Hlt2Topo4BodyBBDT

(91% efficient on offline selected signal)

material veto:

HLT2 displaced vertex selection

PV-like algorithm tuned for displaced vertices.

Line Rxy > m > Ntr � material[mm] [GeV] veto

Longlived HighFD 2.0 4.5 5 yesLonglived HighMass 0.4 10 4 yes

Veerle Heijne (Nikhef) Long-Lived Exotics October 2014 9 / 25

Trigger

[mm]xyR0 2 4 6 8 10

Can

dida

tes /

(0.1

mm

)

1

10

210

310

410

510Data

2c = 35 GeV/vπm

LHCb

Figure 2: Distribution of the distance of the displaced vertex to the interaction region in thetransverse plane for data and for a hidden valley model with m⇡v = 35GeV/c2 and ⌧⇡v = 10psafter the full selection. For visibility, the simulated signal is scaled to 0.62 fb�1 assuming a Higgscross-section of 10 nb and branching fractions of 100% for B(H ! ⇡v⇡v) and B(⇡v ! bb). Theboundaries of the intervals used in the fit are indicated by the dotted lines. The generated Rxy

distribution is approximately exponential with an average of about 2mm.

convoluted with a bifurcated Gaussian function. The parameters of the background modelare fitted to data, independently in each Rxy bin. The signal is modelled by a bifurcatedGaussian function, whose parameters are determined from simulated events in bins of Rxy.The e↵ect of the uncertainty on the jet-energy scale is included by a scale parameter forthe mass, which is common to all bins and constrained using the Z +jet data, as explainedin Section 5. Additional nuisance parameters are added to account for the finite statisticsof the simulated samples and the systematic uncertainties on the signal e�ciency andthe luminosity. The fit model is implemented using the RooFit [31] package. Figure 3shows the fit result in the five radial bins for a signal model with m⇡v = 35 GeV/c2 and⌧⇡v = 10 ps.

The fit procedure is performed for a ⇡v mass of 25, 35, 43 and 50 GeV/c2 and for severalvalues of the lifetime in between 1 and 200 ps. No significant signal is observed for anycombination of ⇡v mass and lifetime. Upper limits are extracted using the CL

s

method [32]with a frequentist treatment of the nuisance parameters described above, as implementedin the RooStats [33] package.

Limits are set on the Higgs production cross-section multiplied by the branching fractioninto long-lived particles �(H)⇥B(H ! ⇡v⇡v). In the simulation it is assumed that both ⇡v

particles decay to the same final state. If the decay width of the ⇡v particle is dominatedby final states other than qq, the limits scale as 1/(Bqq(2 � Bqq)) where Bqq is the ⇡v ! qqbranching fraction. The obtained 95 % CL upper limits on �(H) ⇥ B(H ! ⇡v⇡v), underthe assumption of a 100 % branching fraction to bb, are shown in Table 4 and in Fig. 4.

8

0 10 20 30 40 50 60 70 80

c

1

10

102

103

c

Rxy

0 10 20 30 40 50 60 70 80

c

1

10

102

103

c

Rxy

0 10 20 30 40 50 60 70 80

c

1

10

102

103

c

Rxy

0 10 20 30 40 50 60 70 80

c

1

10

102

103

c

Rxy

0 10 20 30 40 50 60 70 80

c

1

10

102

103

c

Rxy

Figure 3: Di-jet invariant mass distributions for each of the five Rxy bins, superimposed with thefits for a hidden valley model with m⇡v = 35GeV/c2 and ⌧⇡v = 10 ps. The blue line indicates theresult of the total fit to the data. The black short-dashed line is the background-only contribution,and the red long-dashed line is the fitted signal contribution.

9

• Background is modelled by an exponential convolved with a bifurcated Gaussian.

• The signal is modelled by a bifurcated Gaussian (parameters obtained from simulation)

Limits The systematic uncertainty on the selection efficiency and integrated luminosity for the different models is in the range 12.6% to 13.3% . Upper limits on the production cross-section at 95% CL are extracted using the CLs method.

1 10 102

1

10

102

103

104

�(H

)⇥

B(H

!⇡

v⇡

v)

m⇡v c

m⇡v c

m⇡v c

m⇡v c

m⇡v c ⇡v ! cc̄

m⇡v c ⇡v ! ss̄

Figure 4: Observed 95% CL cross-section upper limits on a hidden valley model [3] for various⇡v masses, as a function of ⇡v lifetime, for ⇡v decays into bb unless specified otherwise. Thelimits for ⇡v decay to uu and dd are expected to be the same as for ss.

10

The results cover an unexplored region at the LHC and the limites are more restrictive than results from the Tevatron in the same mass and lifetime range. Branching fractions of greater than 25% are excluded for a SM Higgs to pair produce π0ν particles that decay to two hadronic jets.

LHCb-PAPER-2014-062. Search for long-lived particles decaying to jet pairs. submitted to EPJC, 9 December 2014. arXiv:1412.3021 [hep-ex]

]2cmass [GeV/0 20 40 60 80 100

)2 cC

andi

date

s / (1

GeV

/

0

1000

2000

3000

4000

5000

6000Data

2c = 25 GeV/vπm

2c = 35 GeV/vπm

2c = 50 GeV/vπm

LHCb(a)

]c [GeV/Tp

0 20 40 60 80 100

)cC

andi

date

s / (1

GeV

/

0500

10001500200025003000350040004500

Data2c = 25 GeV/

vπm2c = 35 GeV/

vπm2c = 50 GeV/

vπm

LHCb(b)

Figure 1: Di-jet (a) invariant mass and (b) pT

distributions for data and for hidden valley modelswith 25, 35 and 50GeV/c2 ⇡v masses and 10 ps lifetime. For visibility, the simulated signal isscaled to 0.62 fb�1 assuming a Higgs cross-section of 10 nb and branching fractions of 100% forB(H ! ⇡v⇡v) and B(⇡v ! bb).

select a ⇡v particle, for an illustrative mass of 35 GeV/c2 and lifetime of 10 ps, togetherwith the yield in the data after the most important selection steps. The total e�ciency forother masses and lifetimes, as well as for the decays to light quark jets, is shown in Table 2.The e�ciencies listed in Tables 1 and 2 represent the number of selected candidates dividedby the number of generated events. As the selection e�ciencies for the two ⇡v particles inan event are practically independent, the fraction of selected events with more than onecandidate is less than a few percent in simulated signal. In data no events with more thanone ⇡v candidate are found.

Figure 1 shows the mass and pT

distributions for data and for simulated signal events,assuming a ⇡v particle with a mass of 25, 35 or 50 GeV/c2. The turn-on at low valuesin the mass distribution of events observed in data (Fig. 1a) is caused by the minimump

T

requirement on the jets. The rest of the distribution falls o↵ exponentially. The pT

distribution shown in Fig. 1b illustrates that long-lived particles with a higher mass havelower p

T

as there is less momentum available in the Higgs decay. Studies on simulatedevents have shown that both the shape and the normalisation of the mass distribution indata are compatible with the expected background from bb production. It is not possibleto generate su�ciently large samples of bb events to use these for a quantitative estimateof the background after the final selection. Therefore, the signal yield is extracted by afit to the invariant mass distribution assuming a smooth shape for the background, asdiscussed in Section 6.

5

di-jet invariant mass

Motivation

Many new-physics models feature massive long-lived exotics( �̃0,⇡0

v , ...) which decay into Standard Model particles

A decaying long-lived particle produces a displaced vertex

Standard Model background is lowUnique coverage compared to ATLAS and CMS

Trigger on lifetimes of order 1psSensitive to low pT (mass) signals

CMS (7 TeV)ATLAS (7 TeV)

LHCb (7 TeV)

0c τ [cm]

1 10 100 1000

80

60

40

20

m(π

0 v ) [G

eV]

mh0 > 200 GeV

mh0 = 120-140 GeV

D0 (1.96 TeV)CDF (1.96 TeV)mh0 = 130-170 GeV

mh0 = 90-200 GeV

mh0 = 120 GeV

Veerle Heijne (Nikhef) Long-Lived Exotics October 2014 2 / 25

No significant signal for π0ν particles

Background

Background + Signal

Signal