SM Higgs searches: CMS

(1)

Javier Cuevas

Universidad de Oviedo

IMFP13

Santander , 20-24 May 2013

J. Cuevas

U. Oviedo (Spain) TAE 2015,

th th

SM and BSM:

experimental techniques and results

BSM

(2)

Physics case for new High Energy Machines

Reminder: The Standard Model - tells us how but not why

3 flavour families? Mass spectra? Hierarchy? 19 parameters!

- needs fine tuning of parameters to level of 10 -30 ! - has no connection with gravity

- no unification of the forces at high energy

- Supersymmetry

- Extra space dimensions

Many other ideas: More symmetry and gauge bosons, composite Higgs models, L-R symmetry, quark & lepton substructure,

Little Higgs models, Technicolor, Hidden Valleys, 4 th generation…

Understand the mechanism Electroweak Symmetry Breaking Discover physics beyond the Standard Model

New Gauge Bosons?

Hidden Valleys?

(4)

• Events with five jets of particles and large missing energy which could come from a possible dark matter particle

• But a few events is not enough to prove we have something new

• No visible excess has been building up with time…

Some Interesting Collisions

…already in 2010…

(5)

Beyond the Standard Model

• Apart from the naturalness argument:

– Standard Model accommodates, but does not explain:

• EWSB

• CP-violation

• Fermion masses (i.e., the values of the Yukawa couplings to the Higgs field)

– It doesn’t provide natural explanation for the:

• Neutrino masses

• Cold dark matter

• Logical conclusion:

– Standard model is an effective theory – a low-energy approximation of a more complete theory, which ultimately explains the above

phenomena

– This new theory must take off at a scale of ~1 TeV to avoid significant amount of fine tuning

– Three classes of solutions:

• Ensure automatic cancellation of divergencies (SUSY/Little Higgs)

• Eliminate fundamental scalar and/or introduce intermediate scale Λ ~ 1 TeV (Technicolor / Higgsless models) - basically dead now

• Reduce the highest physics scale to ~1 TeV (Extra Dimensions)

(6)

BSM signatures

(7)

(8)

Search for Large Extra Dimensions

Mono-jet final state +Missing E T (ADD)

p T jet > 110 GeV MET > 200 GeV

Limits on M D between 3 and 4 TeV

Lower limit on the Planck Scale versus number of extra dimensions

arXiv:1408.3583

(9)

Extra Dimensions!

Look for the decay producs of an evaporating black hole

Define S T to be the scalar sum of all high p T objects found in the event

Look for deviations at high S T

Planck scale a few TeV?

Search for Micro Black Holes

arXiv:1202.6396

Nice events, eg a 10 jet event

(10)

Searches with Boosted Objects

BOOST dedicated meetings

http://boost2015.uchicago.edu/

(11)

Resonances Decaying into qV or VV

Heavy resonances decaying into qZ or qW, or VV jets only (CMS) or llqq (ATLAS) using boosted jets and jet substructure analysis

These type of analyses will get even more important at 13 TeV Jets start to merge for X = 700-900 GeV

CMS

arXiv:1405.1994 ATLAS

arXiv:1506.00962

Excess in WZ of 3.4σ

(2.5 with LEE)

(12)

Top: a window to BSM physics ?

(13)

Summary of Exotica Searches

(14)

SlideGreg Landsberg - Search for Exotics @ the LHC - Benasque 2014

The Hierarchy Problem

• Higgs boson mass receives corrections from fermion loops:

!

! !

• The size of corrections is ~ to the UV cutoff ( L ) squared:

!

• In order for the Higgs boson mass to be finite, a fine tuning

(cancellation) of various loops is required to a precision ~(M H / L ) 2 ~ 10 -34 for L ~ M Pl

• This is known as a “hierarchy problem” stemming from a large

hierarchy between the electroweak symmetry breaking and Planck scales, and it requires new physics at Λ ~ 1-10 TeV

9

∆ M H 2 = λ 2 f 4 ⇥ 2 (⇥

2 + M H 2 ) + ...

The Hierarchy Problem

• Higgs boson mass receives corrections from fermion loops:

• The size of corrections is ~to the UV cutoff (Λ) squared:

• In order for the Higgs boson mass to be finite, a fine tuning (cancellation) of various loops is required to a precision ~  (M H /Λ) 2 ~ 10 -34 for Λ ~ M Pl

• This is known as a “hierarchy problem” stemming from a large hierarchy between the electroweak

symmetry breaking and Planck scales, and it requires new physics at Λ ~ 1-10 TeV

S li d e G re g L a n d s b e rg - S e a rc h f o r E x o ti c s @ t h e L H C - B e n a s q u e 2 0 1 4

The Hierarchy Problem

• Higgs boson mass receives corrections from fermion loops:

!

! !

• The size of corrections is ~ to the UV cutoff ( L ) squared:

!

• In order for the Higgs boson mass to be finite, a fine tuning

(cancellation) of various loops is required to a precision ~(M H / L ) 2 ~ 10 -34 for L ~ M Pl

• This is known as a “hierarchy problem” stemming from a large

hierarchy between the electroweak symmetry breaking and Planck scales, and it requires new physics at Λ ~ 1-10 TeV

9

∆ M H 2 = λ 2 f

4 ⇥ 2 ( ⇥

2 + M H 2 ) + ...

(15)

Important properties of SUSY

• Elegant solution to the hierarchy problem (i.e., why the Higgs mass is not at the Planck scale)

• Dark matter candidate with the right abundance

• Gauge unification

• predicts a light Higgs m h < 130 GeV

• consistent with EW precision tests

(16)

Searches for BSM Physics

• First Searches at the LHC (2010-2012)

– Supersymmetry with MET plus jets, lepton(s), photons – Extra Dimensions and black holes, heavy resonances (in

electrons, muons, taus, jets), leptoquarks, excited leptons and quarks, 4 th generation, a few very exotic signatures (R-

hadrons)…

• Evolved Searches (2013-…)

– Supersymmetry on third generation squarks, compressed spectra, stealth SUSY, EWKinos, VBF processes…

– Higgs in decays or as study object, vector-like quarks, boosted objects, long lived particles, fractional charges…

– More dedicated Dark Matter searches!

• We are now facing a restart of the machine at 13/14 TeV…

Back to the basics or do we change paradigm?

(17)

MSSM and cMSSM

• SUSY is a renormalizable and calculable theory and has been thoroughly studied theoretically over the last four decades

• MSSM has just two Higgs doublets; nevertheless the number of parameters describing the model is still very large: 124

– 18 are the SM ones + Higgs boson mass (now known!) – 105 genuinely new parameters:

• 5 real parameters and 3 CP-violating phases in gaugino sector

• 21 squark/slepton masses and 36 mixing angle

• 40 CP-violating phases in the sfermion sector

• This makes it very challenging to search for generic SUSY, and simplifying assumptions are typically made

• One of these simplifications is constrained MSSM, or cMSSM, which assumes gaugino unification and degenerate squark/slepton masses at high energy (typical of gravity-mediated SUSY breaking)

• That results in just five parameters fixing all the SUSY interactions:  

common scalar and fermion masses M0, M1/2, ratio of the vacuum

expectations of the two Higgs doublets tanβ, sign of Higgsino mass

term sign(μ), and trilinear coupling A0

(18)

Detecting Supersymmetric Particles

Energy produced in the detector

 

Super-symmetric particles decay and produce a cascade of jets, leptons and

missing transverse energy (MET) due to escaping ‘dark matter’ particle candidates

Very prominent signatures in CMS and ATLAS

simulation

(19)

SUSY Searches: No signal yet to date

CMS-SUS-11-015

 So far NO clear signal of supersymmetric particles has been found

 We can exclude regions where the new particles could exist.

Searches will continue for the higher energy in 2015

Plenty of searches ongoing: with jets, leptons, photons, W/Z, top, Higgs, with and without large missing transverse energy

Also special searches for contrived model regions

excluded

allowed

Status in 2013

(20)

Limits on Squarks and Gluinos

Results depend on the topologies studies, assumed mass of the LSP etc.

Examples Popular presentation of data:

Simplified ModelS (SMS)

Combined limits typically > 1-1.3 TeV on sparticle masses

(21)

What is really needed from SUSY?

N. Arkani-Ahmed CERN Nov 2011 Papucci, Ruderman, Weiler arXiv:1110.6926 LHC data end 2011 Stops > 200-300 GeV Gluino > 600-800 GeV

Moving away from

constrained SUSY models to ‘natural’ models

Natural SUSY survived LHC so far, but we are getting close to push it to its limits!

End 2011: Revision!

(22)

Stop Searches

• Stop is special for “naturalness” → directly cancels the top loop

• Search depends on stop mass and decay channels – broad program...

• Focus on just two Feynman diagrams representing relevant production and decay: t → t+χ and t → b+χ

– Both result in the same signature: bbW + W - +MET

– This is the same signature as tt production (unless both W’s decay hadronically) - gives you an idea of the dominant

background

September 2015 Javier Cuevas, TAE 2015, Benasque 24

S li d e G re g L a n d s b e rg - S e a rc h f o r D ir e c t S to p P ro d u c ti o n i n C M S

Direct Stop Signatures

We will model the stop pair pr oduction via a “Simpliﬁed Model

Scenario”, i.e. zooming only on the light SUSY particles that matter for this process and assuming all other SUSY particles to be heavy Focus on just two Feynman diagrams representing relevant

production and decay: t → t+ c 0 and t → b+ c +

๏ Both result in the same signature: bbW + W - +ME T

๏ N.B. this is the same signature as tt production (unless both W’s decay hadronically) - gives you an idea of the dominant background

5 P 1

P 2

˜ t ⇤ t ˜

t ¯ t

¯ b

W −

˜ χ 0 1

˜ χ 0 1 W + b

P 1 P 2

˜t ∗

˜t

˜χ −

W −

˜χ +

W +

¯b

˜χ 0 1

˜χ 0 1 b

~ ~

(23)

Stop Searches

• Depending on the mass differences between the stop and neutralino (chargino), several kinematic regions are

defined:

• Different regions correspond to different challenges, so search strategy generally depends on the region

• Given that 4-body decays are enormously suppressed kinematically, the region ΔM < M W in the tχ0 mode is

usually covered by other channels, e.g. FCNC t → cχ decay

d e G re g L a n d s b e rg - S e a rc h f o r D ir e c t S to p P ro d u c ti o n i n C M S

Kinematic Regions

Depending on the mass differences between the stop and neutralino (chargino), several kinematic regions are deﬁned:

!

Different regions correspond to different challenges, so search strategy generally depends on the region

Given that 4-body decays are enormously suppressed

kinematically, the region ΔM < M W in the t c 0 mode is usually covered by other channels, e.g. FCNC t →c c 0 decay

6

ΔM = M

W

MVA Strategy!

3/15/13! Single Lepton Stop WG! 1

!!

200 300 400 500 600 700 ! 100!

200!

300!

400!

500!

600!

m

^~!_t

[GeV]!

ΔM = M

_top!

! t ! t ! !

₁⁰

m

!

[ G e V ]!

1!0!~!

2-body

3- bo dy 4-body

1

(low ΔM)"

200 300 400 500 600 700 ! 100!

200!

300!

400!

500!

600!

2 (m ed iu m Δ M )"

3 (h ig h ΔM )"

m

^~!_t

[GeV]!

ΔM = M

_W!

t ® b c

1

+

® bW c

1

0

(x=0.5)

m

!

[ G e V ]!

1!0!~!

4 (off-shell W)"

2-body 3-body

~

(24)

Stop searches: what’s the best final state to pursue the search

• The final state depends on the W boson decay channels

– All hadronic channel has  the highest branching  fraction, but backgrounds  are huge

– Dilepton channel is clean  but the branching fraction is tiny

– Tau channels are tough

• Single-lepton (e+jets,μ+jets) channels as a compromise between BR (30%) and purity

• Standard variable when dealing with MET: MT

• MT2 or stransverse mass (Lesters & Summers, hep- ph/9906349 ):

Transverse Mass

Standard variable when dealing with signatures containing ME T Classical example: W(l n )

Transverse mass is an approximation of the invariant mass in the case when the

longitudinal

momentum component is not available (e.g., due to a neutrino)

Has a sharp Jacobian peak with a sharp falling edge at the true invariant mass m W

Signal has different distribution in M T , as it contains three invisible particles and therefore doesn’t have a Jacobian peak at m W

1 9 J H E P 0 1 ( 2 0 1 1 ) 0 8 0

[GeV]

E

T

0 20 40 60

n u m b e r o f e v e n ts / 2 .5 G e V

0 1 2 3

103

´

data

n

® e W

t EWK+t QCD

CMS = 7 TeV s at 2.9 pb-1

(a)

[GeV]

M

T

0 20 40 60 80 100 120

n u m b e r o f e v e n ts / 4 G e V

0 0.5 1 1.5

103

´

data

n m W ®

t EWK+t QCD

CMS = 7 TeV s at 2.9 pb-1

(b)

F i gur e 1. T he W signal dist ribut ions: (a) E/

_T

dist ribut ion for t he select ed W ! e ⌫ sample; (b) M

T

dist ribut ions for t he select ed W ! µ⌫sample. T he point s represent t he dat a. Superimposed are t he result s of t he maximum likelihood ﬁts for signal plus backgrounds, in yellow; all backgrounds, in orange; QCD backgrounds, in violet . T he dashed lines represent t he signal dist ribut ions.

Smirnov t est . T he inclusive yield is N

_W

= 12 257 ± 111. T he charge-speciﬁc yields are N

_W+

= 7 445 ± 87 and N

_W−

= 4 812 ± 69. Here, we ﬁt simultaneously for the inclusive yield N

_W

and t he rat io N

_W+

/ N

_W−

so t hat , by const ruct ion, N

_W

= N

_W+

+ N

_W−

.

6.2 Z b oson sel ect ion

To ident ify Z ! `

⁺

`

⁻

decays, a pair of ident iﬁed leptons is required, with dilepton invariant mass in t he range 60 < M

_`+`⁻

< 120 GeV. Backgrounds are very low, including backgrounds from QCD processes. In t he Z ! e

⁺

e

⁻

channel, t he yield is obt ained by count - ing t he number of select ed event s and making a small correct ion for backgrounds. In t he Z ! µ

⁺

µ

⁻

channel, yield and lept on efficiencies are ﬁtted simultaneously. No correction is made for γ

^⇤

exchange.

6.2.1 El ect r ons

T he Z ! e

⁺

e

⁻

candidat e event s are required t o have two elect rons sat isfying t he same select ion crit eria as t he elect rons select ed in t he W ! e ⌫ sample. Bot h elect rons must have an ECAL clust er wit h E

_T

> 20 GeV in t he ECAL ﬁducial volume. The fraction of signal event s select ed in t he simulat ion is F

_Z

= 0.285 ± 0 .005.

T he Z mass peaks in t he dat a exhibit small shift s, on t he order of 1 t o 2%, wit h respect t o t he simulat ed dist ribut ions. From t hese shift s, we det ermine ECAL clust er energy scale correct ion fact ors of 1.015 ± 0 .002 and 1.033 ± 0.005 for barrel and endcap elect rons, respect ively. T he uncert aint ies on t hese correct ion fact ors are propagat ed as syst emat ic uncert aint ies on t he yield. Applying t hese correct ions t o elect ron candidat es in t he dat a, we select 677 event s, wit h t he dielect ron invariant mass shown in ﬁgure 2 (a), along wit h

– 11 –

J H E P 0 1 ( 2 0 1 1 ) 0 8 0

6.1 W b oson sel ect ion

T he W event s are charact erized by a prompt , energet ic, and isolat ed lept on, and signiﬁcant missing energy. T he main backgrounds are QCD mult ijet event s and Drell-Yan event s in which one lept on fails t he select ion. T he QCD background is reduced by requiring t he lep- t on t o be isolat ed; t he remaining event s do not have large E/ T and can be dist inguished from signal event s on a st at ist ical basis. T he Drell-Yan background is suppressed by reject ing event s wit h a second lept on candidat e.

To measure t he signal yields, we choose t o ﬁt the E/ T dist ribut ion in t he elect ron channel and t he M T dist ribut ion in t he muon channel, where M T = p

2 p T E/ T (1 − cos ∆ φ );

∆ φ is t he angle between t he missing t ransverse moment um and t he lept on t ransverse moment um. QCD backgrounds are est imat ed from dat a, as explained below. According t o t he simulat ion, W ! ⌧ ⌫ makes a small relat ive cont ribut ion; backgrounds from Z ! ⌧ + ⌧ − , t t , and diboson product ion are negligible in bot h elect ron and muon channels.

6.1.1 Elect r ons

T he W ! e ⌫ candidat e event s are required t o have one ident iﬁed electron with an ECAL clust er of E T > 20 GeV in t he ECAL ﬁducial volume. If a second electron candidate sat isfying looser crit eria and wit h E T > 20 GeV is present in t he event , t he event is reject ed.

T he fract ion of signal event s select ed in t he simulat ion is F W = 0 .446 ± 0 .006, wit h F W + = 0 .459 ± 0 .007 and F W − = 0 .428 ± 0 .008. T he number of event s select ed in t he dat a is 28 601, wit h 15 859 posit ive and 12 742 negat ive elect rons.

T he W ! e ⌫ signal is ext ract ed from an unbinned maximum likelihood ﬁt of the observed E/ T dist ribut ion t o t he sum of signal and background shapes. T he QCD background shape, which account s for bot h QCD mult ijet product ion and direct -phot on product ion wit h t he phot on convert ing in t he det ect or, can be modeled by a modi ﬁed Rayleigh dist ribut ion,

f ( E/ T ) = E/ T ⇥ exp − E/ 2 T

2( σ 0 + σ 1 E/ T ) 2

! .

T his funct ion can be underst ood as describing ﬂuctuations of the missing transverse mo- ment um vect or around zero due t o measurement errors; t he resolut ion t erm, σ 0 + σ 1 E/ T , increases wit h E/ T t o account for t ails in t he E/ T measurement . T his funct ion describes well t he QCD background shape in t he simulat ion, over t he full range of E/ T , as well as E/ T dist ribut ions from signal-free samples obt ained by invert ing t he ident iﬁcation or isola- t ion crit eria.

T he signal dist ribut ions are derived from simulat ion, separat ely for W + and W − , and receive an event -by-event correct ion in bins of t he W t ransverse moment um, det ermined from a st udy of t he hadronic recoil dist ribut ions of Z ! e + e − event s in t he dat a [14 ]. In ﬁts t o t he E/ T dist ribut ions, t he free paramet ers are t he W signal yield, t he QCD background yield, and t he shape paramet ers σ 0 and σ 1 .

We ext ract t he inclusive yield N W from a ﬁt where the expected ratio for σ W + / σ W − is assumed. It has been checked t hat t he result was insensit ive t o t his assumpt ion. Fig- ure 1 (a) shows t he E/ T dist ribut ion of t he inclusive W ! e ⌫ sample and t he result s of

– 9 –

6 4 Event selection

[GeV]

MT

0 50 100 150 200 250 300

Entries / 30 GeV

10 102

103

104

105

106

Data top l 1

®ll t t W+jets rare

(650/50) x1000

0

c~1

® t

~t

Preselection (a)

Ldt = 19.5 fb-1

ò

= 8 TeV, s CMS

Data/MC

0 0.5 1 1.5 2

[GeV]

miss

ET

100 150 200 250 300 350

Entries / 25 GeV

10 102

103

104

105

106

Data top l 1

(650/50) x1000

0

c~1

® t

~t

Preselection (b)

Ldt = 19.5 fb-1

ò

= 8 TeV, s CMS

Data/MC

0 0.5 1 1.5 2

[GeV]

W

MT2

100 150 200 250 300 350 400 450 500

Entries / 25 GeV

1 10 102

103

104

105

106

Data top l 1

(650/50) x1000

0

c~1

® t

~t

Preselection (c)

Ldt = 19.5 fb-1

ò

= 8 TeV, s CMS

Data/MC

0 0.5 1 1.5 2

c2

Hadronic top 0 2 4 6 8 10 12 14 16 18 20

Entries / 1

10 102

103

104

105

106

Data top l 1

(650/50) x1000

0

c~1

® t

~t

Preselection (d)

Ldt = 19.5 fb-1

ò

= 8 TeV, s CMS

Data/MC

0 0.5 1 1.5 2

ratio

HT

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Entries / 0.04

10-1

1 10 102

103

104

105

106

107 Data

top l 1

(650/50) x1000

0

c~1

® t

~t

Preselection (e)

Ldt = 19.5 fb-1

ò

= 8 TeV, s CMS

Data/MC

0 0.5 1 1.5 2

[rad]

miss)} , ET

(j2

f D

miss), , ET

(j1

f D min{

0 0.5 1 1.5 2 2.5 3

Entries / 0.2 rad

102

103

104

105

Data top l 1

(650/50) x1000

0

c~1

® t

~t

Preselection (f)

Ldt = 19.5 fb-1

ò

= 8 TeV, s CMS

Data/MC

0 0.5 1 1.5 2

) [GeV]

(b1

pT

50 100 150 200 250 300

Entries / 30 GeV

102

103

104

105

106

Data top l 1

(650/50/0.5) x1000

±

c~1

® b

~t

(250/100) x10

0

c~1

® t

~t

Preselection (g)

Ldt = 19.5 fb-1

ò

= 8 TeV, s CMS

Data/MC

0 0.5 1 1.5 2

1) , b l D R(

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Entries / 0.33

1 10 102

103

104

105

106

107

Data top l 1

(650/50/0.5) x1000

±

c~1

® b

~t

Preselection (h)

Ldt = 19.5 fb-1

ò

= 8 TeV, s CMS

Data/MC

0 0.5 1 1.5 2

) [GeV] l

T(

20 30 40 50 60 70 80p 90 100

Entries / 10 GeV

10 102

103

104

105

106

Data top l 1

(250/150/0.5) x10

±

c~1

® b

~t

Preselection (i)

Ldt = 19.5 fb-1

ò

= 8 TeV, s CMS

Data/MC

0 0.5 1 1.5 2

Figure 2: Comparison of data w ith M C simulation for the distributions of (a) M T , (b) E T miss , (c) M W T2 , (d) hadronic top c 2 , (e) H T ratio , (f) minimum Df betw een the E T miss vector and the tw o leading jets, (g) p T of the leading b-tagged jet, (h) DR betw een the leading b-tagged jet and the lepton, and (i) lepton p T , after the preselection. For the plots (a)-(f), distributions for theet ! t e c 0 1 model w ith m et = 650 GeV and m c e

0

1

= 50 GeV, scaled by a factor of 1000, are overlayed. We also show distributions of et ! t e c 0 1 w ith m et = 250 GeV and m c e

⁰

1

= 100 GeV for (g), scaled by 10, and of et ! b e c + w ith m et = 650 GeV, m c e

0

1

= 50 GeV, and x = 0.5 for (h) and (i), scaled by 1000, as w ell as of m et = 250 GeV, m c e

0

1

= 150 GeV, and x = 0.5 for (i), scaled by 10. In all distributions the last bin contains the overﬂow.

M

T

> 120 GeV

M T > 120 GeV requirement is used for signal selection

Introduction

• The s t r a n s v e r s e m a s s M T 2 is a generalization of the transverse mass for decay chains with two unobserved particles, typical in SUSY events

M T 2 = min

p c 1

T + p c 2

T = / p T

max m T (1) , m T (2)

• For the simpliﬁed case of no ISR and zero masses:

( M T 2 ) 2 ≃ 2p T v is (1) p T v is (2) ( 1 + c o s φ 12 )

• Multijet events divided into 2 massless pseudo-jets using a hemisphere algorithm

• M T 2 ≈ / E T for symmetric SUSY-like topologies

• M T 2 is a QCD killer

• M T 2 ≈ 0 for back-to-back events with no genuine MET

• M T 2 < / E T still highly suppressed for nearly back-to-back QCD mismeasurements

• M T 2 provides a very good discriminating power between SM and SUSY-like events, and in this analysis is used as a discovery variable

Bruno Casal (ETH) SUSY Search with MT2 02/04/2012 4 / 35

M T 2 vs E / T

Pascal Nef

M T2 - MET comparison

69 PFMET

0 200 400 600 800

T2M

0 200 400 600 800

0 10 20 30 40 50 LM4

PFMET

0 50 100

T2M

0 50 100

1 10 102

103

104

QCD 470-600

0 200 400 600

0 10 20

QCD W+jets Z+jets Top Other LM6 data

3 Jets

³

Events

[GeV]

MT2

after MET > 300 GeV

Bruno Casal (ETH) SUSY Search with MT2 02/04/2012 45 / 35

Introduction

• The s t r a n s v e r s e m a s s M T 2 is a generalization of the transverse mass for decay chains with two unobserved particles, typical in SUSY events

M T 2 = min

p c T 1 + p c T 2 = / p T

max m T (1) , m T (2)

• For the simpliﬁed case of no ISR and zero masses:

( M T 2 ) 2 ≃ 2p T v is (1) p v T is (2) ( 1 + c o s φ 12 )

• Multijet events divided into 2 massless pseudo-jets using a hemisphere algorithm

• M T 2 ≈ / E T for symmetric SUSY-like topologies

• M T 2 is a QCD killer

• M T 2 ≈ 0 for back-to-back events with no genuine MET

• M T 2 < / E T still highly suppressed for nearly back-to-back QCD mismeasurements

• M T 2 provides a very good discriminating power between SM and SUSY-like events, and in this analysis is used as a discovery variable

Bruno Casal (ETH) SUSY Search with MT2 02/04/2012 4 / 35

The M T2 Variable

M T2 : “stransverse mass” - a generalization of the transverse mass in case of a pair of

invisible particles

For a simpliﬁed case of no

extra jets and zero masses for visible and invisible systems:

!

๏ M T2 ~ ME T for symmetric SUSY-like topologies

M T2 kills QCD background very efficiently:

๏ M T2 ~ 0 for dijets

๏ M T2 < ME T in case of mismeasured dijets

2 0

M T 2 vs E / T

Pascal Nef

M T2 - MET comparison

69 PFMET

0 200 400 600 800

T2M

0 200 400 600 800

0 10 20 30 40 50 LM4

PFMET

0 50 100

T2M

0 50 100

1 10 102

103

104

QCD 470-600

0 200 400 600

0 10 20

QCD W+jets Z+jets Top Other LM6 data

3 Jets

³

Events

[GeV]

MT2

after MET > 300 GeV

Bruno Casal (ETH) SUSY Search with MT2 02/04/2012 45 / 35

Lesters & Summers, hep-ph/9906349

Signal

Background

(25)

Final optimization

• Cut-based approach, variables are treated independently putting a cutoff in each of them.

• Multivariate: all variables are combined in a likelihood, BDT or ANN reflecting how signal-like they are.

• Select events with MT > 120 GeV

• Several signal regions defined with a cut-based or MVA

approach (~ 40 % improvement)

(26)

stop: Combination of channels:

8!TeV!Summaries!

LHCP!2015!Z!K.!Ulmer! 20!

neutralino mass = chargino mass [GeV]

100 200 300 400 500 600 700 800

LSP mass [GeV]

0 100 200 300 400 500 600 700 800 900

1

c~0

= m

1

c±

m~

+mZ

1

c0

~

= m

1

c±

m~

+mH

1

c0

~

= m

1

c±

m~

ICHEP 2014 = 8 TeV s

CMS Preliminary

production

1

c~

± 2

- c~

0

1) c~0 1)(W c~0

® (H

1

c~± 2

c~0

1) c~0 1)(W c~0

® (Z

1

c~± 2

c~0

) )=0.5 l-

l+

, BF(

lL

(~

1

c~± 2

c~0

)

-)=1

+l l , BF(

lL

(~

1

c~- 1

c~+

t) n~

t~

1 ( c~± 2

c~0

)

-)=1

+l , BF(l lR

(~

1

c~± 2

c~0

SUS-13-006 19.5 fb-1

SUS-14-002 19.5 fb-1

Observed Expected

1

c~0

= m

1

c±

m~

gluino mass [GeV]

600 800 1000 1200 1400 1600

LSP mass [GeV]

0 100 200 300 400 500 600 700 800 900 1000

m(gluino) - m(LSP) = 2 m(top)

= 8 TeV s

CMS

1

c~

0

t

® t g ~ production, g ~

- g ~

) 19.5 fb-1

+HT

ET

SUS-13-012 0-lep (

6) 19.3 fb-1 jets³

SUS-13-007 1-lep (n

SUS-13-013 2-lep (SS+b) 19.5 fb-1

SUS-14-010 3-lep (3l+b) 19.5 fb-1

SUS-14-010 2-lep(OS) 19.5 fb-1

SUS-14-010 0+1+2(SS,OS)+>2-lep 19.5 fb-1

Observed

SUSY theory

s Observed -1 Expected

Start to filling holes with new ideas:

indirect search (from tt cross-section)

(27)

Prospects for run 2

• Discovery reach in stop mass will reach to 800 GeV in a

conservative scenario

• Crucial region for testing naturalness and whether SUSY has a role in

Electroweak

symmetry breaking

• Naturalness prefers m stop lighter than 700 GeV

• Higgs mass of ~125 GeV prefers m stop

heavier than 300 GeV

(28)

Summary of SUSY Searches (ATLAS)

In short: no sign of SUSY with the data collected so far

(30)

Summary of SUSY Searches (CMS)

In short: no sign of SUSY with the data collected so far

(31)

Dark Matter: Complementary Searches?

After the discovery of the Higgs particle @ the LHC:

Dark matter is the next important physics problems to tackle for the LHC

The search is complementary to other experimental techniques used.

(32)

Mono-X signatures

September 2015 Javier Cuevas, TAE 2015, Benasque 34 4

X + M issing T ransverse Energy

4

q

¯ q

χ

¯ χ

Figure 1: Dark matter production in association with a single jet in a hadron collider.

3.1. Comparing Various M ono-Jet Analyses

Dark matter pair production through a diagram likefigure 1 is one of the leading channels for dark matter searches at hadron colliders [3, 4]. The signal would manifest itself as an excess of jets plus missing energy (j+E/_T) events over the Standard Model background, which consists mainly of (Z ! ⌫⌫) +j and (W ! `înv⌫) +j final states. In thelatter casethecharged leptonìs lost, as indicated by the superscript“inv”. Experimental studies ofj+E/Tfinal states have been performed by CDF [22], CMS[23] and ATLAS[24, 25], mostly in thecontext of Extra Dimensions.

Our analysiswill, for themost part, bebased on theATLASsearch [25] which looked for mono- jets in 1 fb⁻¹of data, although we will also compare to the earlier CMS analysis [23], which used 36 pb⁻¹of integrated luminosity. The ATLAS search contains three separate analyses based on successively harderpT cuts, the major selection criteria from each analysis that we apply in our analysis are given below.³

LowPT Selection requiresE/T> 120 GeV, one jet withpT(j1)> 120 GeV,|⌘(j1)| < 2, and events are vetoed if they contain a second jet withpT(j2)> 30 GeV and|⌘(j2)| <4.5.

Hi ghPT Selection requiresE/_T>220 GeV, one jet withpT(j1)>250 GeV,|⌘(j1)| <2, and events are vetoed if there is a second jet with|⌘(j2)| < 4.5 and with either pT(j2) > 60 GeV or

∆φ(j2, /ET)< 0.5. Any further jets with|⌘(j2)| <4.5 must havepT(j3)<30 GeV.

ver yHi ghPT Selection requiresE/_T > 300 GeV, one jet withpT(j1)> 350 GeV,|⌘(j1)| < 2, and events are vetoed if there is a second jet with|⌘(j2)| <4.5 and with eitherpT(j2)> 60 GeV or∆φ(j2, /ET)<0.5. Any further jets with|⌘(j2)| <4.5 must havepT(j3)< 30 GeV.

In all casesevents arevetoed if they contain any hard leptons, deﬁned for electronsas|⌘(e)| <2.47 andpT(e)>20 GeV and for muons as|⌘(µ)| <2.4 andpT(µ)> 10 GeV.

The cuts used by CMS are similar to those of the LowPT ATLAS analysis. Mono-jet events are selected by requiringE/T >150 GeV and one jet withpT(j1)> 110 GeV and pseudo-rapidity

|⌘(j1)| <2.4. A second jet withpT(j2)> 30 GeV is allowed if the azimuthal angle it forms with the leading jet is∆φ(j1, j2)< 2.0 radians. Events with more than two jets withpT> 30 GeV are vetoed, as are events containing charged leptons withpT> 10 GeV. The number of expected and observed events in the various searches is shown in table I.

3Both ATLAS and CMS impose additional isolation cuts, which we do not mimic in our analysis for simplicity and since they would not have a large impact on our results.

4

q

¯ q

χ

¯ χ

3.1. Compar ing Various M ono-Jet Analyses

Dark matter pair production through a diagram likefigure 1 is one of the leading channels for dark matter searches at hadron colliders [3, 4]. The signal would manifest itself as an excess of jets plus missing energy (j +E/_T) events over the Standard Model background, which consists mainly of (Z ! ⌫⌫) +j and (W ! `înv⌫) +j final states. In thelatter casethecharged leptonìs lost, as indicated by the superscript “inv”. Experimental studies ofj+E/Tfinal states have been performed by CDF [22], CMS[23] and ATLAS[24, 25], mostly in thecontext of Extra Dimensions.

Our analysiswill, for themost part, bebased on theATLASsearch [25] which looked for mono- jets in 1 fb⁻¹of data, although we will also compare to the earlier CMS analysis [23], which used 36 pb⁻¹of integrated luminosity. The ATLAS search contains three separate analyses based on successively harderpT cuts, the major selection criteria from each analysis that we apply in our analysis are given below.³

LowPT Selection requiresE/T> 120 GeV, one jet withpT(j1)> 120 GeV,|⌘(j1)| < 2, and events are vetoed if they contain a second jet withpT(j2)> 30 GeV and|⌘(j2)| <4.5.

Hi ghPT Selection requiresE/T>220 GeV, one jet withpT(j1)>250 GeV,|⌘(j1)| <2, and events are vetoed if there is a second jet with|⌘(j2)| < 4.5 and with eitherpT(j2) > 60 GeV or

∆φ(j2, /E_T)<0.5. Any further jets with|⌘(j2)| <4.5 must havepT(j3)<30 GeV.

ver yHi ghPT Selection requiresE/_T > 300 GeV, one jet withpT(j1)> 350 GeV,|⌘(j1)| < 2, and events are vetoed if there is a second jet with|⌘(j2)| <4.5 and with eitherpT(j2)> 60 GeV or∆φ(j2, /ET)<0.5. Any further jets with|⌘(j2)| <4.5 must havepT(j3)< 30 GeV.

In all cases eventsarevetoed if they contain any hard leptons, deﬁned for electronsas|⌘(e)| <2.47 andpT(e)>20 GeV and for muons as|⌘(µ)| <2.4 andpT(µ)> 10 GeV.

The cuts used by CMS are similar to those of the LowPT ATLAS analysis. Mono-jet events are selected by requiringE/_T >150 GeV and one jet withpT(j1)> 110 GeV and pseudo-rapidity

|⌘(j1)| <2.4. A second jet withpT(j2)> 30 GeV is allowed if the azimuthal angle it forms with the leading jet is∆φ(j1, j2)< 2.0 radians. Events with more than two jets withpT> 30 GeV are vetoed, as are events containing charged leptons withpT> 10 GeV. The number of expected and observed events in the various searches is shown in table I.

3Both ATLAS and CMS impose additional isolation cuts, which we do not mimic in our analysis for simplicity and since they would not have a large impact on our results.

Monojet

χ χ qbar

q g

g

4

q q ¯

χ

¯ χ

3.1. Comparing Various M ono-Jet Analyses

Dark matter pair production through a diagram likefigure 1 is one of the leading channels for dark matter searches at hadron colliders [3, 4]. The signal would manifest itself as an excess of jets plus missing energy (j+E/T) events over the Standard Model background, which consists mainly of (Z ! ⌫⌫) +j and (W ! `înv⌫) +j final states. In thelatter casethecharged leptonìs lost, as indicated by the superscript“inv”. Experimental studies ofj+E/_T final states have been performed by CDF [22], CMS[23] and ATLAS[24, 25], mostly in thecontext of Extra Dimensions.

Our analysis will, for themost part, bebased on theATLASsearch [25] which looked for mono- jets in 1 fb⁻¹of data, although we will also compare to the earlier CMS analysis [23], which used 36 pb⁻¹of integrated luminosity. The ATLAS search contains three separate analyses based on successively harderp_T cuts, the major selection criteria from each analysis that we apply in our analysis are given below.³

LowPT Selection requiresE/_T>120 GeV, one jet withp_T(j1)>120 GeV,|⌘(j1)| <2, and events are vetoed if they contain a second jet withpT(j2)>30 GeV and|⌘(j2)| <4.5.

Hi ghPT Selection requiresE/_T>220 GeV, one jet withp_T(j₁)>250 GeV,|⌘(j₁)| <2, and events are vetoed if there is a second jet with|⌘(j2)| < 4.5 and with eitherpT(j2)> 60 GeV or

∆φ(j2, /E_T)<0.5. Any further jets with|⌘(j2)| <4.5 must havepT(j3)<30 GeV.

ver yHi ghPT Selection requiresE/_T> 300 GeV, one jet withpT(j1)>350 GeV,|⌘(j1)| <2, and events are vetoed if there is a second jet with|⌘(j2)| <4.5 and with eitherpT(j2)>60 GeV or∆φ(j2, /ET)<0.5. Any further jets with|⌘(j2)| <4.5 must havepT(j3)<30 GeV.

In all cases events arevetoed if they contain any hard leptons, deﬁned for electronsas|⌘(e)| <2.47 andpT(e)>20 GeV and for muons as|⌘(µ)| <2.4 andpT(µ)>10 GeV.

The cuts used by CMS are similar to those of the LowPT ATLAS analysis. Mono-jet events are selected by requiringE/_T>150 GeV and one jet withpT(j1)>110 GeV and pseudo-rapidity

|⌘(j1)| <2.4. A second jet withpT(j2)> 30 GeV is allowed if the azimuthal angle it forms with the leading jet is∆φ(j1, j2)<2.0 radians. Events with more than two jets withpT>30 GeV are vetoed, as are events containing charged leptons withpT>10 GeV. The number of expected and observed events in the various searches is shown in table I.

W/Z

MonoB MonoTop

MonoPhoton MonoZ

q q

MonoW (monoLepton) MonoW/Z (Hadronic) BBbar /TTbar Higgs Portal

Norraphat SRIMANOBHAS | Latest Results in Dark Matter Searches, NODITA 2014

(33)

No signal  limits on “traditional”

plane

[GeV]

WIMP mass m

c

1 10 10

²

10

³

]

2

W IM P -n u c le o n c ro s s s e c ti o n [ c m

10

-44

10

-42

10

-40

10

-38

10

-36

10

-34

10

-32

10

-30

10

-28

10

-26

truncated, coupling = 1 truncated, max coupling

5q

mg g cq g5

gm

c D8:

nq sm

cq

n

sm

c D9:

ATLAS

fb-1

TeV, 20.3

=8 s

90% CL

spin-dependent

COUPP 90% CL SIMPLE 90% CL PICASSO 90% CL Super-K 90% CL

90% CL W-

IceCube W+

CMS 8TeV D8

Comparison to DD: σ χ-N

Translate collider limits to bounds on σ χ -N

‣ Lower limit on M* translated to limits on σ χ -N

• e.g. Phys.Rev.D82:116010,2010

‣ Collider searches more sensitive at low M Χ

• And up to medium M Χ for spin-dependent interactions

‣ Complementarity to direct searches

[GeV]

m

c

WIMP mass

1 10 10

²

10

³

]

2

-46

10

-45

10

-44

10

-43

10

-42

10

-41

10

-40

10

-39

10

-38

10

-37

10

-36

10

-35

10

-34

10

c

-c

+l

®l c c

®Z pp

= 8.0 TeV, L = 19.7 fb-1

s

Spin Independent

Preliminary CMS

CMS mono-Z (D5) CMS mono-Z (C3)

q = 1

c = g truncated, g

q = 1

c = g truncated, g

® SS CMS invisible Higgs: H

LUX 2013

superCDMS 2014

CDMSlite

(D5)

CMS mono-g CMS mono-jet (D5)

SI

SD

Mono-Jet

[GeV]

WIMP mass m

c

1 10 10

²

10

³

]

2

10

-44

10

-42

10

-40

10

-38

10

-36

10

-34

10

-32

10

-30

10

-28

10

-26

truncated, coupling = 1 truncated, max coupling

5q

mg g cq g5

gm

c D8:

nq sm

cq

n

sm

c D9:

ATLAS

fb-1

TeV, 20.3

=8 s

90% CL

spin-dependent

COUPP 90% CL SIMPLE 90% CL PICASSO 90% CL Super-K 90% CL

90% CL W-

IceCube W+

CMS 8TeV D8

Comparison to DD: σ χ-N

Translate collider limits to bounds on σ χ -N

‣ Lower limit on M* translated to limits on σ χ -N

• e.g. Phys.Rev.D82:116010,2010

‣ Collider searches more sensitive at low M Χ

• And up to medium M Χ for spin-dependent interactions

‣ Complementarity to direct searches

SI

SD

Mono-Jet

Annapaola de Cosa page

Mono-Jet

Hard jet from initial state radiation (ISR) recoiling against DM

‣ Most sensitive channel

• 1 central, high p

T

jet and lepton veto - A 2

^nd

jet allowed if Δφ (j

1

, j

2

)<2.5 [CMS]

- Other jets allowed but Δφ (j1, E

Tmiss

)>1.0 [ATLAS]

• High energy imbalance in transverse plane

‣ Key variable: Missing transverse energy (E

Tmiss

)

• Definition of inclusive SRs with increasing E

Tmiss

threshold

‣ Dominant backgrounds

• Z(→ νν) + Jets and W(→(l

^miss

)ν) + Jets

• Data-driven techniques employed

8

Events / GeV

10-2

10-1

1 10 102

103

104

Data 2012 SM uncertainty

)+jets n n

® Z(

)+jets n

® l W(

Di-boson + single top t t Multi-jet

ll)+jets

® Z(

GeV

*=670 GeV, M M=100 D5

TeV

D=3 n=2, M ADD

-4eV

G~=10 TeV, M

g=1

,~ q~

gM /~ q~ + G~ ATLAS

=8 TeV, 20.3 fb-1 s

>150 GeV miss ET

[GeV]

miss

ET

200 400 600 800 1000 1200

Data/SM

0.501 1.52

CMS, EPJC75 (2015) 235/ CMS-EXO-12-055/

ATLAS, EPJC (2015) 75:299

Mono-V leptonic

W(→νl) + DM

‣ 1 isolated, high p

T

lepton and high E

Tmiss

‣ Final discriminating variable

• Transverse mass, M

T

12

CMS, PRD 91, 092005 (2015) - ATLAS, JHEP 09 (2014) 037

(GeV) MT

500 1000 1500 2000 2500

Events / 1 GeV

10-4

10-3

10-2

10-1

1 10 102

103

104

105 W ® (m,t) _n

, single t t t DY QCD Diboson Data Syst. uncer.

M = 2000 GeV n, m SSM W' ®

= 4000 GeV L ®mn, HNC CI

= +1 x = 300 GeV, Mc

= 200 GeV, L DM,

(8 TeV) 19.7 fb-1

CMS

miss + ET m

CMS-EXO-12-054 / ATLAS, PRD 90, 012004 (2014)

Z(→ll) + DM

‣ 2 opposite-charge same-flavour leptons

‣ m

^ll

consistent with m

Z

‣ Look for an excess in E

Tmiss

/ M

T

spectrum [ATLAS /CMS]

‣ ZZ process fakes signal

Mono-V leptonic

W(→νl) + DM

‣ 1 isolated, high p

T

lepton and high E

Tmiss

‣ Final discriminating variable

• Transverse mass, M

T

12

CMS, PRD 91, 092005 (2015) - ATLAS, JHEP 09 (2014) 037

[GeV]

miss

ET

Entries / 50 GeV

10-2

10-1

1 10 102

103

104

105

106

107

108

ATLAS

=200 GeV mc

L=20.3 fb-1

ò

^s^{=8 TeV}

=0.050 TeV D1, M*

=0.7 TeV , M*

g max.

c c ZZ

=1 TeV, f=6 Mediator, mh

h Data W/Z+jets WW/Top quark

WZ®llnn ZZ

Systematic Unc.

[GeV]

miss

ET

0 50 100 150 200 250 300 350 400 450 500

Data/MC

0.60.81 1.21.4

CMS-EXO-12-054 / ATLAS, PRD 90, 012004 (2014)

Z(→ll) + DM

‣ 2 opposite-charge same-flavour leptons

‣ m

ll

consistent with m

Z

‣ Look for an excess in E

Tmiss

/ M

T

spectrum [ATLAS /CMS]

‣ ZZ process fakes signal

monojet

Mono“W” MonoZ

(34)

Does the

SM Higgs searches: CMS

Javier Cuevas

3 flavour families? Mass spectra? Hierarchy? 19 parameters!

Little Higgs models, Technicolor, Hidden Valleys, 4 th generation…

New Gauge Bosons?

Some Interesting Collisions

• Logical conclusion:

• Ensure automatic cancellation of divergencies (SUSY/Little Higgs)

Limits on M D between 3 and 4 TeV

Planck scale a few TeV?

These type of analyses will get even more important at 13 TeV Jets start to merge for X = 700-900 GeV

• The size of corrections is ~ to the UV cutoff ( L ) squared:

• Higgs boson mass receives corrections from fermion loops:

The Hierarchy Problem

Important properties of SUSY

electrons, muons, taus, jets), leptoquarks, excited leptons and quarks, 4 th generation, a few very exotic signatures (R-

• We are now facing a restart of the machine at 13/14 TeV…

– 18 are the SM ones + Higgs boson mass (now known!) – 105 genuinely new parameters:

• One of these simplifications is constrained MSSM, or cMSSM, which assumes gaugino unification and degenerate squark/slepton masses at high energy (typical of gravity-mediated SUSY breaking)

Detecting Supersymmetric Particles

CMS-SUS-11-015

Also special searches for contrived model regions

N. Arkani-Ahmed CERN Nov 2011 Papucci, Ruderman, Weiler arXiv:1110.6926 LHC data end 2011 Stops > 200-300 GeV Gluino > 600-800 GeV

• Search depends on stop mass and decay channels – broad program...

Direct Stop Signatures

๏ N.B. this is the same signature as tt production (unless both W’s decay hadronically) - gives you an idea of the dominant background

usually covered by other channels, e.g. FCNC t → cχ decay

kinematically, the region ΔM < M W in the t c 0 mode is usually covered by other channels, e.g. FCNC t →c c 0 decay

– Dilepton channel is clean but the branching fraction is tiny

Transverse mass is an approximation of the invariant mass in the case when the

F i gur e 1. T he W signal dist ribut ions: (a) E/

= 4 812 ± 69. Here, we ﬁt simultaneously for the inclusive yield N

channel, t he yield is obt ained by count - ing t he number of select ed event s and making a small correct ion for backgrounds. In t he Z ! µ

> 20 GeV in t he ECAL ﬁducial volume. The fraction of signal event s select ed in t he simulat ion is F

6.1 W b oson sel ect ion

6.1.1 Elect r ons

> 120 GeV

• Multijet events divided into 2 massless pseudo-jets using a hemisphere algorithm

Bruno Casal (ETH) SUSY Search with MT2 02/04/2012 4 / 35

69 PFMET

PFMET

Events

after MET > 300 GeV

• M T 2 ≈ / E T for symmetric SUSY-like topologies

Bruno Casal (ETH) SUSY Search with MT2 02/04/2012 4 / 35

๏ M T2 ~ 0 for dijets

69 PFMET

PFMET

Events

after MET > 300 GeV

• Select events with MT > 120 GeV

® t g ~ production, g ~

conservative scenario

Summary of SUSY Searches (ATLAS)

The search is complementary to other experimental techniques used.

Monojet

MonoB MonoTop

Comparison to DD: σ χ-N

Mono-Jet

• And up to medium M Χ for spin-dependent interactions

• High energy imbalance in transverse plane

CMS, EPJC75 (2015) 235/ CMS-EXO-12-055/

CMS-EXO-12-054 / ATLAS, PRD 90, 012004 (2014)

CMS-EXO-12-054 / ATLAS, PRD 90, 012004 (2014)

– Dilepton channel is clean  but the branching fraction is tiny