1. SUSY Particle Mass Measurement with the Contransverse Mass Dan Tovey, University of Sheffield1

2. Motivation Aim to constrain sparticle masses with short decay chains:
Examples:
Tool already on the market: ‘stransverse mass’ mT2 (Lester & Summers, Phys.Lett.B463:99-103,1999) used in CSC SUSY-5 and SUSY-6.
Maximum transverse mass of two decays, minimised over all possible partitions of ETmiss between decays.
Requires mass of LSP as input
Some sensitivity in principle to individual masses
New proposal: ‘contransverse mass’ (JHEP 0804 (2008) 034, arXiv:  [hep-ph]) 2

3. Contransverse Mass
Consider identical pair-produced SUSY-particles D1 and D2, each decaying to LSP and a visible particle v.
In centre-of-mass of D1 and D2, v1 and v2 boosted with equal magnitude but opposite direction relative to rest frames of D1 and D2.
Invariant quantity:
In transverse plane:
If v1 and v2 massless then
MCT maximised when v1 and v2 co-linear 3

4. Contransverse Mass Q: Why is MCT interesting?
Phase-space distribution (scalar decays)
Q: Why is MCT interesting?
It does not represent mass of a particle decaying to give v1 and v2
A: Value in lab frame (if no transverse boost) equal to that calculated from transverse momenta measured in rest frames of D1 and D2.
These momenta constrained by the masses of the sparticles involved via two-body kinematics.
So e.g. for right-squark pair events with massless quark decay products expect end-point: 4

5. QCD Background Rejection
New basis for ETmiss provided by MCT for dijet events
MCT measures contribution to ETmiss from event topology:
Jet mis-measurement mainly modifies asymmetry but topology less so: powerful discrimination tool …
Asymmetry
Topology
SU4
QCD: J5,J6,J7
~400pb-1
~26pb-1
ETmiss=constant
ETmiss=constant 5

6. 2-Jet Analysis Test right-squark case with rel12 CSC data (PYTHIA)
Hard 2-jet exclusive selection cuts (standard CSC SUSY object definitions)
Pass xE70 trigger
Exactly 2 reconstructed jets, each with pT > 50 GeV
No electrons or muons
ETmiss > 100 GeV
pT of residual activity (i.e. excluding jet pT) < 50 GeV
Removes SUSY background and also SM:
QCD di-jet events strongly suppressed: jet mis-measurement maintains co-linear topology. Requires high pT 3rd jet to be completely missed.
W+1 jet and ttbar similarly suppressed by kinematics (e.g. MCT(ttbar) < mtop).
Znn + 2 jets and Wln + 2 jets dominate remainder.
Aim to use track ETmiss to constrain residual activity further. 6

7. Z/W+Jets Background Control samples: Limited statistics
Wln + jets from tagged lepton +ETmiss + 2 jet events with mT(l,ETmiss)< 100 GeV. Reconstruct W momentum from W-mass constraint, redecay assuming W unpolarised.
Znn +jets (1) from tagged lepton +ETmiss + 2 jet events with mT(l,ETmiss)< 100 GeV. Add lepton pT to ETmiss.
Znn +jets (2) from tagged 2 lepton + 2 jet events with |m(ll)-mZ|<15 GeV. Add leptons’ pT to ETmiss.
Limited statistics
Control samples ~100 pb-1, ~5 times less than background samples
Error bars indicate 1fb-1 expectation
Running out of MC statistics in tail … 7

8. 2-Jet Analysis
Normalise exponential fit with smallest gradient (conservative) to lowest bin
Subtract fit, normalised with signal present  clear end-point
Limited resolution of jets closely separated in f acts to suppress events near end-point.
Modified ARGUS function fit gives end-point ~600 +/- 50 GeV (615 GeV expected)
Small non-exponential tail to background – may be visible in control given equivalent MC stats  use e.g. double-exp fit. 8

9. 3-Jet analysis 3-jets exclusive:
more complicated background
edge more smeared
BUT larger signal acceptance.
Reject QCD (jet mis-measurement) by requiring all jets to lie in one half-circle of the transverse plane.
Use 2 leading jets for MCT y j1 j2 x j3 9

10. 4-Jet Analysis Repeat as for 3-jets Less background
End-point more smeared 10

11. SU4 2-jet 3-jet 4-jet Repeat for SU4 point
End-point expected at 396 GeV
Normalisation to data more difficult – signal contaminates norm region.
Plots show background estimate normalised without signal present.
2-jet
3-jet
4-jet 11

12. Independent Masses?
Possibility of measuring individual masses by combining products of multi-step chains
Problems with ISR and clusterisation – needs more work ~
g decays ~
q decays
SPS1a point, taken from JHEP 0804 (2008) 034 (non-ATLAS simulation)
10 fb-1, SPS1a signal only, exclusive 4-jet analysis
Without ISR
With ISR 12

13. Summary
Simple new variable proposed for measuring sparticle masses with short decay chains
Measures simple analytical combination of masses
Potential sensitivity to individual sparticle masses, assuming ISR problems can be over-come 13

14. Back-Up 14

15. Transverse Mass
Consider decay products v1 and v2 of massive particle D.
If both v1 and v2 fully observable then obtain mass of D from:
If z-component of v1 or v2 (or both) not observable then use:
where
If v1 and v2 massless then this simplifies to:
Transverse mass maximised when v1 and v2 back-to-back 15

16. Transverse Mass
‘Invariant mass’ m is invariant under Lorentz boosts of particle D.
Equivalent statement: m is invariant under co-linear, equal magnitude boosts of v1 and v2 .
Use of mT maintains invariance under co-linear equal magnitude boosts along beam direction. 16