What can we say about what we’ve found?

Alan Barr 7 June 2007

What can we sayabout what we’vefound?

Was it reallySUSY?

How

can

we

disc

over

y

SUSY

at LH

C?

Just find SM Higgs

Alan Barr

Alan Barr 7 June 2007Extended higgs sector(2 doublets)

Your mission…SM SUSY

quarks (L&R)leptons (L&R) neutrinos (L&?)

squarks (L&R)sleptons (L&R)sneutrinos (L&?)

Z0

W±

gluon

BW0

h0

H0

A0

H±

H0

H±

4 x neutralino

2 x chargino

AfterMixing

gluino

Spin-1/2

Spin-1

Spin-0

Spin-1/2

Spin-0

BinoWino0

Wino±

gluino

~

~


Features of RP SUSY?

• RPV as a conserved QN:• Events build from blobs

with 2 “exotic legs”• A pair of cascade decays

results• Complicated end result

• RPV as a conserved QN:• Events build from blobs

with 2 “exotic legs”• A pair of cascade decays

results• Complicated end result

Time

standard

2 exotics

Production part

Time

standard

heavyexotic lighter

exotic

Decay part Time

Complete event

= exotic= standard


General features

“typical” susy spectrum(mSUGRA)

• Complicated cascade decays– Many intermediates

• Typical signal– Jets

• Squarks and Gluinos

– Leptons• Sleptons and

weak gauginos– Missing energy

• Undetected LSP

• Model dependent– Various ways of

transmitting SUSY breaking from a hidden sector

LHC Pt5


What do we see?

Lifetimes short -> look for Standard Model decay relics + missing energy


Example of a search topology

SIGNAL topology

q

q_

squarkq

LSP

q_

LSP(and similar)

BACKGROUND topology (QCD)

• No unique choice of sensitive topology– Complementary information/sensitivity

• Expect SM backgrounds with similar characteristics to signal– Need to search for excesses

• No unique choice of sensitive topology– Complementary information/sensitivity

• Expect SM backgrounds with similar characteristics to signal– Need to search for excesses


Practical Problems• See only SM decay products

– Expect short lifetimes• Lose information about order of decays

– Jets (other than b and t) indistinguishable• Loose flavour information for other squarks

• “Missing momentum” from neutralinos only determined perpendicular to beam– Individual LSP momenta not individually

measurable– Z-momentum of initial state unknown (PDFs)– Can’t reconstruct from final state

• Forward jets lost down beam pipe

– Can’t form “invariant masses” of sparticles• No “clean” mass peaks for resonances


Precise measurement of SM backgrounds:

the problem

• SM backgrounds are not small

• There are uncertainties in– Cross sections– Kinematical

distributions– Detector response

Lower backgrounds

Higher backgrounds

“Rediscover”

“Discover”

ZZ

WW


Just look for jets?

Big QCD background

Scalar sum of transverse energy / GeV


Add some missing energyLook for events with jets and missing energy

Cuts

at least two jets with: ET

Jet1,2 > 150,100 GeV

|Jet1,2| < 2.5

Meff = Jets pTi + MET

But with addition of some other cuts…

Missing transverse momentum > 100 GeV

cuts based on i = |(Jet,i)-(MET)|): R1 = (2

2+(-1)2) > 0.5 rad

R2 = (12+ (-2)2) > 0.5 rad

no jet with i < 0.5 rad

Kill events with missing energy from miss-measured jets

Kill events with missing energy from miss-measured jets

QCD dijets

“SUSY”


Two-JetTwo-Jet

No MT2

Dijet cuts + MET +

Scalar sum of transverse energy / GeV

Expect discovery distribution to be of something like this form:Excess of “some sort” of new physics about SM backgrounds.


Importance of detailed detector understanding

• GEANT simulation already shows events with large missing energy– Jets falling in “crack” region– Calorimeter punch-through

• Vital to remove these in missing energy tails

• Large effort in physics commissioning

Lesson from the TevatronEt(miss)

Rare occurrences hurt


Inclusive reach in mSUGRA parameter space

Map of discovery potential corresponding to a 5σ excess above background in mSUGRA m0 – m1/2 parameter space for the ATLAS experiment.

jets + ETmiss

channel

L = 1033 cm-2 s-

1

~1 year → ~2200 GeV

~1 month → ~1800 GeV

few days (< one week) → ~1300 GeV

Health warning: expecting SUSY discovery in a few days will seriously damage your credibility

Health warning: expecting SUSY discovery in a few days will seriously damage your credibility


Different searches• We will be looking in many different

channels– n jets + m leptons + missing energy– +- b-jets (common at large tan β)– +- tau-jets (“ “ “)– Charged stable particles– NLSP -> photon gravitino (GMSB)– R-parity violating modes– R-hadrons– …


What might we then know?• Assume we have MSSM-like

SUSY with m(squark)~m(gluino)~600 GeV

• See excesses in these distributions

• Can’t say “we have discovered SUSY”

• Can say some things:– Undetected particles produced

• missing energy– Some particles have mass ~ 600

GeV, with couplings similar to QCD • Meff & cross-section

– Some of the particles are coloured • jets

– Some of the particles are Majorana• excess of like-sign lepton pairs

– Lepton flavour ~ conserved in first two generations

• e vs mu numbers– Possibly Yukawa-like couplings

• excess of third generation– Some particles contain lepton

quantum numbers• opposite sign, same family

dileptons – …

Slide based on Polesello


Mapping out the new world

• Some measurements make high demands on:– Statistics (=> time)– Understanding of detector– Clever experimental technique

LHC Measuremen

tSUSY

Extra Dimensions

MassesBreaking

mechanismGeometry &

scale

SpinsDistinguish

from EDDistinguish from SUSY

Mixings,Lifetimes

Gauge unification?Dark matter candidate?


Constraining masses• Mass constraints• Invariant masses in

pairs– Missing energy– Kinematic edges

Observable: Depends on:

Limits depend on angles betweensparticle decays

Frequently-studieddecay chain


Mass determination

• Basic technique– Measure edges– Try with different SUSY

points– Find likelihood of fitting

data

• Event-by-event likelihood– In progress

Measureedges

Variety of edges/variables

Try variousmasses in equations

C.G. Lester

• Narrow bands in ΔM• Wider in mass scale• Improve using cross- section information


SUSY mass measurements• Extracting

parameters of interest– Difficult problem– Lots of competing

channels– Can be difficult to

disentangle– Ambiguities in

interpretation– Lots of effort has

been made to find good techniques

Tryvariousdecaychains

Tryvariousdecaychains

Look forsensitive variables

(many of them)

Look forsensitive variables

(many of them)

Extractmasses

Extractmasses


SUSY mass measurements:• LHC clearly cannot fully constrain all

parameters of mSUGRA– However it makes good constraints

• Particularly good at mass differences [O(1%)]• Not so good at mass scales • [O(10%) from direct measurements]• Mass scale possibly best “measured” from cross-

sections– Often have >1 interpretation

• What solution to end-point formula is relevant?• Which neutralino was in this decay chain?• What was the “chirality” of the slepton “ “ “ ?• Was it a 2-body or 3-body decay?


SUSY spin measurements

• The defining property of supersymmetry– Distinguish from e.g. similar-looking

Universal Extra Dimensions

• Difficult to measure @ LHC– No polarised beams– Missing energy– Indeterminate initial state from pp

collision

• Nevertheless, we have some very good chances…


Measuring spins of particles

• Basic recipe:– Produce polarised particle– Look at angular distributions in its decay

spinθ


Left Squarks-> strongly interacting-> large production-> chiral couplings

mass/G

eV

Revisit “Typical” sparticle spectrum

Some sparticles omitted

10

–> Stable-> weakly interacting

Right slepton(selectron or smuon)-> Production/decayproduce lepton-> chiral couplings

LHC point 5

20 = neutralino2

–> (mostly) partnerof SM W0

10 = neutralino1

–> Stable-> weakly interacting

Right slepton(selectron or smuon)-> Production/decayproduce lepton-> chiral couplings


Spin projection factors

Approximate SM particles as massless-> okay since m « p

Lq~Lq

02

~1

0Lq

P

S

Chiral coupling




Lq~Lq

02

~

1

0~ LqP

S

0

1~02 S

Σ=0

Spin-0

Produces polarised neutralino




(near) Rl

θ*p

SLq~

Lq

Rl

~02

~Rl

Scalar

Fermion

Polarisedfermion




(near) Rl

θ*p

S

Lq~Lq

Rl

~02

~Rl

mql – measureinvariant mass

1

0~ LqP

S


lnearq invariant mass (1)

m/mmax = sin ½θ*

Back to backin 2

0 frame

θ*

quark

lepton

Phase space -> factor of sin ½θ*Spin projection factor in |M|2: l+q -> sin2 ½θ* l-q -> cos2 ½θ*

l+

l-

Phase space

Pro

bab

ility

Lq~ Lq

Rl

~02

~Rl

Invariant mass


After detector simulation (ATLFAST)

l+

l-Change in shapedue to charge-blind cuts

parton-level * 0.6

-> Charge asymmetry survives detector simulation-> Same shape as parton level (but with BG and smearing)

detector-levelInvariant mass

Ch

arg

e a

sym

metr

y,

spin-0

Even

ts

detector effects cuts to greatly reduce SM

SUSY


Interesting questions• Can we test gaugino universality?

– Can we constrain the neutralino mass mixing matrix?

• Can we measure sparticle splittings?– JMR: Htt coupling interesting

• Can we “predict”/confirm dark matter density?

• Can we measure mass scale to better than ~10%– Precision measurement/prediction for cross-

sections?• Can we confirm spin(s)?


Extras


Standard Model backgrounds: measure from LHC DATA

• Example: SUSY BG– Missing energy + jets

from Z0 to neutrinos– Measure in Z -> μμ– Use for Z ->

• Good match– Useful technique

• Statistics limited– Go on to use W => μ

to improve

Measure in Z -> μμ

Use in Z -> νν R: Z

B: Estimated

R: Z

B: Estimated


W contribution to no-lepton BG

• Use visible leptons from W’s to estimate background to no-lepton SUSY search

Oe, Okawa,Asai


Normalising not necessarily good enough

Distributions arebiased by lepton selection

Distributions arebiased by lepton selection


Need to isolate individual components…


Then possible to get it right…

Similar story for other backgrounds – control needs careful selectionSimilar story for other backgrounds – control needs careful selection


Direct slepton spin determination

• Spin important in slepton production– Occurs through

s-channel spin-1 process only

– Characteristic angular distribution in production

q

q_

e+

e-

Z/γ0

1~

01

~

e+

e-~

~

l-

l+

θ*q q_

~

~


Distributions @ parton level

black=SUSY

Parallel ParallelPerpendicularto beam

• Spin-0– SUSY– Sleptons– “perpendicular”

to beam

• Spin-½– UED– KK leptons– “parallel” to

beam

l-

l+

θq q_

~

~

red=UED

blue=PS

σtotal not to scale


Sensitive variables?• cos θlab

– Good for linear collider– Not boost invariant

• Missing energy means Z boost not known @ LHC

• Not sensitive @ LHC

• Δη– Boost invariant– Sensitive– Not easy to compare with

theory

• cos θll*– 1-D function of Δη:

– All benefits of Δη– Interpretation as angle in

boosted frame– Easier to compare with

theory

l1l2

θ2lab

θ1lab

l1l2η2

lab

η1lab

Δη

l1Δη

l2

θl*θl

*

(A)

(B)

(C)

N.B. ignores azimuthal angle

boos

t

)tanh()tan2cos(cos 211* 2

1

ell


“Data” = inclusive SUSY after cuts

• “SPS5” point– Below: spectrum– Right: results– Good stat.

discrimination

Some results

Documents

What can we say about what we’ve found?