LHCC

Searches for Long-Lived Particles at LHCbMatthieu Marinangeli, on behalf of the LHCb collaboration,École Polytechnique Fédérale de Lausanne (EPFL), Switzerland.

Dark photon decaying to di-muon:

Study uses 1.6 fb-1 of data collected at Ecm = 13 TeV:

• Prompt-like search ( 2"# < " %% < 70 GeV/c2 ), ()∗+,~1.

• Displaced-like search (214 < " %% < 350 MeV/c2 ), A’ long-lived when ε is small. ()∗+

,= ()∗

+, (", 5). Main background from photon conversions at the VELO region.

Dark sector scenario involving a massive dark photon A’ which mixes with the off-shell photon 7∗ by a factor 8. :; → %=%> rate can be normalized to ?∗ → %=%> rate:

@A

A

%=

%>

?∗/:;

Drell-Yan

Massive LLP decaying to μ + jets:

LLP mass = [ 20, 200 ] GeV/c2

LLP lifetime = [ 5, 100 ] ps

10�3 10�2

c⌧ [m]

10

20

30

40

50

60

70

80

m[G

eV/c2 ]

LHCbp

s = 8 TeV

LHCb constraints on different B(H0 ! �̃01�̃

01) at 95% CL

B > 5% excl.B > 10% excl.B > 25% excl.B > 50% excl.

Search uses 3 fb-1 of data collected at Ecm = 7 and 8 TeV. Number of candidates extracted from fit to LLP mass.

[EPJC (2017) 77:224]

No excess found. Model independent limits set on ℬ DEF → %AA ×H, results interpreted with limits on various production modes for instance pair production from a SM Higgs boson.

h0

�̃

�̃

310 410 510

210

310

410

510

610

710

m(µ+µ�) [MeV ]

Can

dida

tes/

�[m

(µ+µ�)]/2

LHCbps = 13TeV

prompt µ+µ�

µQµQ

hh+ hµQ

) isolationapplied

prompt-like sample

pT(µ) > 1GeV, p(µ) > 20GeV• IJK

)∗ is known from prompt di-muon spectrum.• IJK

+, is obtained from fits of di-muon mass (in bins of mass and lifetime).

No significant excess was found. Regions where upper limit onILM+, is less than INO+

, at 90 % C.L are excluded.

[PRL 120.061801 (2018)]

Search for LLPs decaying into a high PT muon and two quarks, interpreted as mSUGRA RPV neutralinos:

Massive LLP decaying to jet pairs:

Search for LLPs produced in pairs from a SM Higgs boson and decaying into two b-quarks, interpreted as Hidden Valley pions:

LLP mass = [ 25, 50 ] GeV/c2

LLP lifetime = [ 2, 500 ] ps

Search uses 2 fb-1 of data collected at Ecm = 7 and 8 TeV. Number of candidates extracted from fit to di-jet mass in 6 bins of the radial vertex position Rxy (0.4 – 50 mm):

1 10 102 103Lifetime [ps]

10−2

10−1

1

10

102

103

(σ/σ

SM gg→H0) ੁต( H0

→π vπ v)

LHCbmπv

= 25GeV/c2

mπv= 35GeV/c2

mπv= 43GeV/c2

mπv= 50GeV/c2

mπv= 35GeV/c2, πv → c c۽

mπv= 35GeV/c2, πv → s s۽

No excess found. Upper limits presented as a function of PQlifetime for different masses, including PQ → R ̅R and PQ → TT̅ when

m(PQ) = 35 GeV/c2 .

[EPJC (2017) 77:812]

The possibility that dark matter particles may interact via unknown forces, felt onlyfeebly by Standard Model (SM) particles, has motivated substantial e↵ort to search fordark-sector forces (see Ref. [1] for a review). A compelling dark-force scenario involvesa massive dark photon, A0, whose coupling to the electromagnetic current is suppressedrelative to that of the ordinary photon, �, by a factor of ". In the minimal model, thedark photon does not couple directly to charged SM particles; however, a coupling mayarise via kinetic mixing between the SM hypercharge and A0 field strength tensors [2–7].This mixing provides a potential portal through which dark photons may be producedif kinematically allowed. If the kinetic mixing arises due to processes whose amplitudesinvolve one or two loops containing high-mass particles, perhaps even at the Planckscale, then 10�12 . "2 . 10�4 is expected [1]. Fully exploring this few-loop range ofkinetic-mixing strength is an important goal of dark-sector physics.

Constraints have been placed on visible A0 decays by previous beam-dump [7–21],fixed-target [22–24], collider [25–28], and rare-meson-decay [29–38] experiments. Thefew-loop region is ruled out for dark photon masses m(A0) . 10MeV (c = 1 throughoutthis Letter). Additionally, the region "2 & 5⇥10�7 is excluded for m(A0) < 10.2GeV, alongwith about half of the remaining few-loop region below the dimuon threshold. Many ideashave been proposed to further explore the [m(A0), "2] parameter space [39–51], includingan inclusive search for A0

!µ+µ� decays with the LHCb experiment, which is predictedto provide sensitivity to large regions of otherwise inaccessible parameter space using datato be collected during Run 3 of the LHC (2021–2023) [52].

A dark photon produced in proton-proton, pp, collisions via �–A0 mixing inherits theproduction mechanisms of an o↵-shell photon with m(�⇤) = m(A0); therefore, both theproduction and decay kinematics of the A0

!µ+µ� and �⇤!µ+µ� processes are identical.

Furthermore, the expected A0!µ+µ� signal yield is given by [52]

nA0

ex [m(A0), "2] = "2"n�⇤

ob[m(A0)]

2�m

#F [m(A0)] ✏A0

�⇤ [m(A0), ⌧(A0)], (1)

where n�⇤

ob[m(A0)] is the observed prompt �⇤! µ+µ� yield in a small ±�m window

around m(A0), the function F [m(A0)] includes phase-space and other known factors, and✏A0�⇤ [m(A0), ⌧(A0)] is the ratio of the A0

!µ+µ� and �⇤!µ+µ� detection e�ciencies, which

depends on the A0 lifetime, ⌧ (A0). If A0 decays to invisible final states are negligible, then⌧(A0) / [m(A0)"2]�1 and A0

!µ+µ� decays can potentially be reconstructed as displacedfrom the primary pp vertex (PV) when the product m(A0)"2 is small. When ⌧(A0) issmall compared to the experimental resolution, A0

!µ+µ� decays are reconstructed asprompt-like and are experimentally indistinguishable from prompt �⇤

!µ+µ� production,resulting in ✏A0

�⇤ [m(A0), ⌧(A0)] ⇡ 1. This facilitates a fully data-driven search and thecancelation of most experimental systematic e↵ects, since the observed A0

!µ+µ� yields,nA0ob[m(A0)], can be normalized to nA0

ex [m(A0), "2] to obtain constraints on "2.This Letter presents searches for both prompt-like and long-lived dark photons produced

in pp collisions at a center-of-mass energy of 13TeV, using A0!µ+µ� decays and a data

sample corresponding to an integrated luminosity of 1.6 fb�1 collected with the LHCbdetector in 2016. The prompt-like A0 search is performed from near the dimuon thresholdup to 70GeV, above which the m(µ+µ�) spectrum is dominated by the Z boson. Thelong-lived A0 search is restricted to the mass range 214 < m(A0) < 350MeV, where thedata sample potentially provides sensitivity.

1

Also recast into an heavy neutral lepton scenario PLB 2017.09.057.

0 20 40 60 80Dijet mass [GeV/c2]

10−11

10102103104

Candidates/

(2GeV/c2 )

0.4 < Rxy < 1.0mmఅs = 8TeV

LHCb

QCD

backgroundsignal model

with % = 1

1

10

102

103

104

Data 7 TeVLV38 10ps

bb sim.–

Muon pT

En

trie

s/(3

GeV

/c)

[GeV/c]0 50 100 150

a)

LHCb

1

10

102

103

104

Data 7 TeVLV38 10ps

bb sim.–

Muon isolation

Entr

ies/

0.0

4

1 1.5 2 2.5

b)

LHCb

1

10

102

103

104

Data 7 TeVLV38 10ps

bb sim.–

LLP number of tracks

Entr

ies

0 10 20 30

c)

LHCb

1

10

102

103

104

Data 7 TeVLV38 10ps

bb sim.–

LLP mass

En

trie

s/(3

GeV

/c2)

[GeV/c2]

0 50 100 150

d)

LHCb

1

10

102

103

104

Data 7 TeVLV38 10ps

bb sim.–

LLP Rxy

Entr

ies/

(0.5

mm

)

[mm]0 5 10 15 20

e)

LHCb

1

10

102

103

104

Data 7 TeVLV38 10psLV38 50ps

bb sim.–

Entr

ies/

(0.5

mm

)

[mm]0 5 10

LLP Rxy

15 20

e)

LHCb

1

10

102

103

104

Data 7 TeVLV38 10ps

bb sim.–

LLP σR

Entr

ies/

(0.0

1 m

m)

[mm]0 0.1 0.2 0.3 0.4 0.5

f)

LHCb

1

10

102

103

104 Data 7 TeV

LV38 10ps bb sim.

–

LLP σz

Entr

ies/

(0.0

4 m

m)

[mm]0 0.5 1 1.5 2

g)

LHCb

Figure 2: Distributions for the 7 TeV dataset (black histogram) compared to simulated bb events(blue squares with error bars), showing a) transverse momentum and b) isolation of the muon, c)

number of tracks and d) reconstructed mass of the displaced vertex. The fully simulated signal

distributions for LV38 10 ps are also shown (red dotted histograms). The distributions from

simulation are normalised to the number of data entries.2

]2cLLP mass [GeV/20 40 60 80

)2 cEn

tries

/(1.5

GeV

/

2−10

1−10

1

10 Data 8 TeV(non-isolated region)Background fit

a)

LHCb

]2cLLP mass [GeV/20 40 60 80 100

)2 cEn

tries

/(2 G

eV/

1−10

1

10 Data 8 TeV(non-isolated region)Background fit

b)

LHCb

]2cLLP mass [GeV/20 40 60 80

)2 cEn

tries

/(1.5

GeV

/

1−10

1

10Data 8 TeVFit: total background signal

c)

LHCb

]2cLLP mass [GeV/20 40 60 80 100

)2 cEn

tries

/(2 G

eV/

1−10

1

10

Data 8 TeVFit: total background signal

d)

LHCb

Figure 3: Reconstructed mass of the LLP candidate from the 8 TeV dataset. The two top

plots correspond to events with candidates selected from the background region of the muon

isolation variable. They are fitted with the sum of two exponential functions. In the bottom

row the candidates from the signal region are fitted including a specific signal shape, added to

the background component. Subfigures a) and c) correspond to the analysis which assumes the

LV38 5 ps signal model, b) and d) are for LV98 10 ps.

3

m(E) = 38 GeV/c2

U

UVW

VW"XTYI

?∗/:;

%=

%>

Meson decay

m(PQ) = 35 GeV/c2

5(PQ) = 10 ps

% from best-fit

10�4 10�3 10�2 10�1 100 101 102

⇡V c⌧ [m]

10

20

30

40

50

60

70

80

m⇡

V[G

eV/c

2]

Regions where B(H0 ! ⇡V ⇡V ) > 50% is excluded at 95% CL

ATLAS 20.3 fb�1 at 8 TeVLHCb 2.0 fb�1 at 7-8 TeVCMS 18.5 fb�1 at 8 TeV

Currently working on Run II data exploring lower PQ masses and developing

tools to study multi-jets at lower masses.

?∗/A’ efficiency ratio

Predicted limits for future runs for Dark Photons from charm mesons decay (ee final state) for 15 fb-1, 50 fb-1 and 500 fb-1 [PRD 92 (2015) 115017] and from inclusive di-muon searches for 15 fb-1, 50 fb-1 and 500 fb-1 [PRL 116 (2016) 251803].

2-10 1-10 1 1012-10

11-10

10-10

9-10

8-10

7-10

6-10

5-10

4-10

m(A0) [ GeV ]

"2

LHCb

LHCb

Previous Experiments

90% CL exclusion regions on [m(A0), "2]

1

10

210

310

[mm]z

500− 0 500 1000

[mm

]r

(sig

ned)

20−

10−

0

10

20

LHCb

ZF

PQ

PQ

[

@[

[

@[