susy squark flavour studies w i th atlas

SUSY squark flavourSUSY squark flavour studies w studies wiith ATLASth ATLAS

Tommaso LariTommaso LariUniversità and INFN MilanoUniversità and INFN Milano

On Behalf of the ATLAS CollaboratOn Behalf of the ATLAS Collaborationion

T. LariT. Larisquark flavour studies with ATLASsquark flavour studies with ATLAS

Flavour WorkshopFlavour WorkshopCERN, 08/11/2005 CERN, 08/11/2005

OutlineOutline

IntroductionIntroduction SUSY mass scale from SUSY mass scale from

inclusive searchesinclusive searches Left-handed squarkLeft-handed squark Sbottom and gluino Sbottom and gluino Right-handed squarkRight-handed squark StopStop

ConclusionsConclusions

• Most of the ATLAS work since Physics TDR (1999) done on mSUGRA models.

A particularly extensive study is available for SPS1a point (fast simulation) – it will be used here to illustrate techniques to reconstruct the squark mass spectrum.

When available for a given signature, recent full simulation results will also be shown (obtained for other mSUGRA points)



IntroductionIntroduction

In mSUGRA: At the Unification scale, all

scalars have the same mass. At the EW scale:

First two generation almost mass degenerate, but m(qL) ≠ m(qR)

Large L-R mixing in 3° generation: b1,b2,t1,t2 mass eigenstates

Light b1 and t1

SPS1a mass spectrum (colored states)

mSUGRA free parameters: m0, m1/2, A0, tan sgn(

SPS1a: m0= 100 GeV, m1/2= 250 GeV, A0= -100 GeV,tan(β)=10 , μ>0

~ ~

~ ~ ~ ~

~ ~



mSUGRA events topologymSUGRA events topology

Strongly interacting sparticles (squarks, gluinos) dominate LHC production.

Cascade decays to the stable, weakly interacting lightest neutralino follows.

• Event topology: • high pT jets (from squark/gluino decay)

– Large ETmiss signature (from LSP)

– High pT leptons, b-jets, -jets (depending on model parameters)

If sbottom or stop quarks in the decay chain: b-jets

Charm tagging impossible?

A typical decay chain:

lqq

l

g~ q~ l~

~

~

p p



Inclusive signaturesInclusive signatures

First hint of the existence of non-First hint of the existence of non-SM physics will probably be an SM physics will probably be an excess of events with large excess of events with large missing energy and hard jetsmissing energy and hard jets

The peak of the distribution of the The peak of the distribution of the effective mass:effective mass:

if visible above the background, if visible above the background, is strongly correlated with the is strongly correlated with the mass of the SUSY particle mass of the SUSY particle produced (gluino or squark): first produced (gluino or squark): first estimate of SUSY mass scaleestimate of SUSY mass scale

jets

jetT

missTeff pEM

ATLAS 10 fb-1

10 fb-1

ATLAS fast simulation

SM background(ALPGEN+PYTHIA)

SUSY (1 TeV)

Effective mass (GeV)

Effective mass (GeV)SU

SY

mas

s sc

ale

(GeV

)



Inclusive searches (Full simulation)Inclusive searches (Full simulation)

Full simulation data, produced for seven different mSUGRA Full simulation data, produced for seven different mSUGRA points, confirms the good correlation between the Effective points, confirms the good correlation between the Effective Mass peak and the SUSY mass scale Mass peak and the SUSY mass scale

ATLAS PreliminaryFull Simulation

ATLAS PreliminaryFull Simulation

coannihilation point

SM background(fast simulation)

Stau coannihilation point: m0 =70 GeV m1/2 = 350 GeV A=0 tan>0



SPS1aSPS1a

A detailed study for SPS1aA detailed study for SPS1a ((bulkbulk

region of parameter space)region of parameter space) waswas

done in done in fast simulationfast simulation

Gjelsten, Lytken, Miller, Osland, Polesello, ATL-PHYS-2004-007

• qL mass from qL → q

→ q l l → q l l

• qR mass from qR → q

• b, g mass fromg → bb → bb

2 → bb ll → bb ll

~

~ ~

~~

~ ~~ ~ ~

SPS1a



Left squark cascade decayLeft squark cascade decay

qll edge qlmin edge qlmax edge qll threshold

ll edge

0 40

Invariant mass (GeV)0 600 600 600 6000 0 0

80

01

02

~ ~

~~ llqqllqq RL

2 SFOS lep., pT>20, 10 GeV ≥ 4 jets, pT>150,100,50,50 GeV Meff > 600 GeV ETmiss >max(100, 0.2 Meff)

The decay chain (note the two isolated leptons)

Allows to measure several kinematical endpoints, each providing one mass relation between the SUSY particles.

SPS1a100 fb-1

Fast simulation

Gjelsten, Lytken, Miller, Osland, Polesello,ATL-PHYS-2004-007



Mass reconstruction

Fit results (L = 100 fb-1)

Gjelsten, Lytken, Miller, Osland, Polesello, ATL-PHYS-2004-007

Δm(χ10)= 4.8 GeV, Δm(χ2

0)= 4.7 GeV,

Δm(lR) = 4.8 GeV, Δm(qL)= 8.7 GeV

m(χ10) = 96 GeV m(lR ) = 143 GeV

m(qL) = 540 GeV m(χ20) = 177 GeV

5 relations for 4 masses!

Left squark mass fitLeft squark mass fit

Can compare qll and bll thresholds, and measure m(b)-m(uL)Precision limited by systematics: error on jet energy

scale (1% expected)

~

~

~ ~

~ ~



Going up the decay chainGoing up the decay chain

Once the mass of the Once the mass of the is is known, known,

it is possible to get the momentum it is possible to get the momentum of the of the

using the approximate relation

p(p(= ( 1-m(

m(ll) ) pll

valvalid for lepton pairs with invariant mass near the edge.

The can be combined with b-jets

to reconstruct the gluino and sbottom mass peaks:

g → bb → bb

~

~

~ ~

~

~ ~ ~

ATLAS fast sim.ATL-PHYS-2004-007

Dilepton invariant mass

~~

b-tagging used to separate light and bottom squark decay chains



Gluino and sbottom reconstructionGluino and sbottom reconstruction

Good combinations. m(g) and m(b) correlated (Dominant error from

Momentum affects both)

Bad b combinations (b-jet

is from gluino decay)

~ ~

~

~

ATLAS fast simulationATL-PHYS-2004-007300 fb-1



Gluino mass reconstructionGluino mass reconstruction

~ ~m(g)-0.99m(

= (500.0 ± 6.4) GeV with 300 fb-1 ~ ~

Error is statistical plus (dominant) 1% uncertainty on jet energy scaleCentral value 10 GeV lower than nominal because no dedicated calibration for b-jet energy scale used in this (fast sim) study



M(bb) (GeV) M(LSP) (GeV)M

(bb

) (G

eV)

~



sbottom masssbottom mass

m(bb) and m(

b) strongly correlated : use the difference

With 300 fb-1 it should be possible to separate the b1 and b2 peaks

~ ~

~ ~

m(g)-m(b= (103.3 ± 1.8) GeV m(g)-m(b= (70.6 ± 2.6) GeV

With lower statistics only measure the average

~



b1

b2

m(bb)-m(b) (GeV) m(bb)-m(b) (GeV)

~ ~ ~

~ ~

~

~



Right-handed squarkRight-handed squark

qR does not couple to Wino

is nearly a Bino

is nearly a Wino

qR → q

m(qR)-m(= (424.2 ± 10.9) GeV

Combine the transverse momentum of two leading jets with missing transversemomentum as follows:

~~~

~~

~

~

~ ~

~

~ ~

qR qR production: two high pT jets and missing energy

ATLAS fast simulationATL-PHYS-2004-007SPS1a 30 fb-1

ATLAS full simulationstau coannihilation point

The different decay chains allow separation from qL

(veto additional jets and b-tagged jets)

~

~ ~



SPS1a: summarySPS1a: summary

The following masses would be measured by ATLAS for SPS1a:The following masses would be measured by ATLAS for SPS1a:

m(qR)-m(= (424.2 ± 10.9) GeV 30 fb-1

m(qL) = (444.0 ± 4.9) GeV 300 fb-1

m(g)-m(= (500.0 ± 6.4) GeV 300 fb-1

m(g)-m(b= (103.3 ± 1.8) GeV 300 fb-1

m(g)-m(b= (70.6 ± 2.6) GeV 300 fb-1

After a few years at LHC design luminosity, precision would be limited by systematic error on jet energy scaleThe analysis works in a large fraction of parameter space (when relevant decay chains exist) but results depend on specific point

~

~

~

~ ~

~ ~



Stop edgesStop edges

g → tt1 → tb±1 or g → bb → bt ±

1

tb invariant mass has a maximum function of the masses of g, b(or t) and ±1

Two closely spaced edges from the two decays: can measure a weighed average.

Selections:- total jet energy and missing energy- 2 b-jets- lepton veto- 4 to 6 non-b jets

Reconstruction:- m(jj) close to m(W)- m(jjb) close to m(t)- W-sidebands to estimate and subtract combinatorial background

~ ~ ~ ~ ~ ~

~ ~ ~

See J. Hisano et al., Phys. Rev. D68, 035007for details

ATLAS

~~ ~



ConclusionsConclusions

On a significant fraction of mSUGRA parameter space, the LHC On a significant fraction of mSUGRA parameter space, the LHC would be able to measure the mass of would be able to measure the mass of qqRR,q,qLL,b,b11,b,b22

Many measurements studied also with Full Simulation data for a variety of points – they confirm that fast simulation results are realistic.

A model-independent measurement of the A model-independent measurement of the stop massstop mass is is more more challengingchallenging (but kinematical decays related to stop decays would (but kinematical decays related to stop decays would be observed for most parameter values)be observed for most parameter values)

Measurement of mixing angles needs more studyCPV, FV, mixing and non-degenerate masses in first two generations

also poorly covered - - availability in HERWIG/PYTHIA/ISAJET ?- - criteria to choose parameters? flavour benchmarks?

~ ~ ~ ~

susy squark flavour studies w i th atlas

Documents

ql mass

mass relation

mass degenerate

squark mass spectrum

c01 qr mass

effective mass peak

t2 mass eigenstateslight

gev m12