Searches for Long-Lived Particles Using Displaced Vertices Matthew - - PowerPoint PPT Presentation
Searches for Long-Lived Particles Using Displaced Vertices Matthew Walker for the ATLAS and CMS Collaborations Rutgers University November 12, 2013 Outline Review of ATLAS and CMS Tracking Detectors Overview of primary vertexing techniques
November 12, 2013 Matthew Walker, Rutgers University
Review of ATLAS and CMS Tracking Detectors Overview of primary vertexing techniques Use of vertexing in the ATLAS displaced vertex search Use of vertexing in the ATLAS displaced jet search Use of vertexing in the CMS displaced dijet search Summary
2
November 12, 2013 Matthew Walker, Rutgers University
Inside 2T solenoid Insertable B-Layer (IBL) Added during LS1 r = 33 mm Pixel detector
3 barrels (r = 51, 89, 123 mm) 2x3 endcaps (z = 495, 580, 650 mm)
Silicon micro-strip tracker double sided modules
4 barrel layers (r = 300, 371, 443, 514 mm)
2x9 forward disks (z in 835-2788 mm)
Transition radiation tracker
Designed so tracks with pT > 0.5 GeV will cross 30 straws
3
November 12, 2013 Matthew Walker, Rutgers University
Inside 4T solenoid Pixel detector
3 barrels (r = 44, 73, 102 mm) 2x2 endcaps (z = 345, 465 mm)
Silicon Strip tracker
4 inner barrel layers ( 200 < r < 550 mm) 6 outer barrel layers (550 < r < 1160 mm) 3 inner disks (580 < z < 1240 mm) 9 outer disks (1240 < z < 2820 mm)
4
r (cm)
10 20 30 40 50 60 70 80 90 100 110
z (cm)
100 200 300
3.0 2.8 2.6 2.4 2.2 2.0 1.8 1.6
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
→ η →
− TEC TEC+ TOB TIB − TID TID+ PIXEL
November 12, 2013 Matthew Walker, Rutgers University
Review of ATLAS and CMS Tracking Detectors Overview of primary vertexing techniques Use of vertexing in the ATLAS displaced vertex search Use of vertexing in the ATLAS displaced jet search Use of vertexing in the CMS displaced dijet search Summary
5
November 12, 2013 Matthew Walker, Rutgers University
Select a list of tracks based on some quality cuts and compatibility with beam spot Clustering in CMS is done with Deterministic Annealing Minimizes a free-energy-like function to determine the best number
Changing “temperature” allows relaxation of the system - if the system goes through a phase-change, the relevant vertex is split in two Once the system reaches a minimum temperature, splitting is stopped, and only track assignment is changed An additional outlier rejection term is added to down-weight tracks that are not near any vertex
6
November 12, 2013 Matthew Walker, Rutgers University
7
F = −T
#tracks
X
i
pi log
#vertices
X
k
ρk exp − 1 T (zT
i − zV k )2
σz
i 2
c = 2
X
i
pipik σz
i 2
✓zT
i − zV k
σz
i
◆2 / X
i
pipik σz
i 2
November 12, 2013 Matthew Walker, Rutgers University
Adaptive Vertex Fitter applied to each vertex with at least two tracks that are incompatible with other vertices Least squares estimator that weights each track based on its compatibility with the vertex ndof = -3 + 2 * (sum of weights) Improves the robustness of the fit in the event of misassociated tracks or mismeasured track errors Iterative fitting stops when vertex position has not changed by more than 1 micron Vertex resolution tested by splitting tracks into two sets and refitting
8
November 12, 2013 Matthew Walker, Rutgers University
Review of ATLAS and CMS Tracking Detectors Overview of primary vertexing techniques Use of vertexing in the ATLAS displaced vertex search Use of vertexing in the ATLAS displaced jet search Use of vertexing in the CMS displaced dijet search Summary
9
November 12, 2013 Matthew Walker, Rutgers University
Searching for RPV, GGM, and split-SUSY scenarios Uses a retracking pass to find additional displaced tracks in the inner detector Searches for multitrack displaced vertices or dilepton displaced vertices
10
November 12, 2013 Matthew Walker, Rutgers University
Track requirements: pT > 1 GeV > 1 SCT hits, > 0 TRT hits, > 1 pixel hits d0 > 2mm Vertexing Seed with vertices made between all pairs of tracks if the vertex has chi2 < 5 (for 1 d.o.f.) Tracks in these vertices can’t have any hits with radius less than the vertex Also requirements on having hits in the first or second layer
11
November 12, 2013 Matthew Walker, Rutgers University
Additional vertexing procedure Define distance vector as the vector between the primary and secondary vertices Require > -20 mm Iterative track combination procedure Look for tracks in multiple vertices If the chi2 for the tracks and one of the vertices is greater than 6, remove it from the vertex If not, find the second vertex with the smallest distance significance If the significance of the distance < 3, merge the two vertices If not, remove the track from the vertex that it has the higher chi2 from
12
~ d = ~ rDV − ~ rP V ~ d · ˆ p
November 12, 2013 Matthew Walker, Rutgers University
Vertex selection chi2 / ndof < 5 rDV < 300 mm, |zDV| < 300 mm Transverse distance to any PV > 4 mm Veto vertices in material Require invariant mass of vertices > 50 MeV, also remove K0s with mass cut Number of tracks > 4 Divide the vertices in high/low mass bins using a cut of 10 GeV
13
Density of vertices with less than 5 tracks that are excluded by material veto
November 12, 2013 Matthew Walker, Rutgers University
Signal Regions DV + lepton Triggering muon with pT > 55 GeV, d0 > 1.5 mm, |eta| < 1.07, cosmic-ray rejection Triggering electron with pT > 125 GeV, d0 > 1.5 mm Lepton distance of closest approach to DV < 0.5 mm DV + jets or MET 4 Jets with pT > 90, 5 Jets with pT > 65 or 6 jets with pT> 55 GeV and all jets have to pass quality criteria MET > 180 GeV
14
November 12, 2013 Matthew Walker, Rutgers University
Background estimates Accidental crossing: a low mDV vertex is accidentally crossed by a high-pT track at large angle Low-mDV component modeled using the mDV distribution of vertices whose tracks are highly collimated then scaled High-mDV component modeled by mixing vertices with tracks from other events which have their momentum rotated so that azimuthal and polar angles with respect to the distance vector are the same Merged vertices: two low-mDV vertices that are less than 1 mm apart are reconstructed as a single vertex that passes selection criteria
15
November 12, 2013 Matthew Walker, Rutgers University
Results No events were seen in any of the signal regions Upper limits are set on signal yields and cross sections for a variety of models, taking into account vertex efficiency for given ctau values Limits are calculated in each model for a variety of values of model dependent parameters
16
November 12, 2013 Matthew Walker, Rutgers University
Review of ATLAS and CMS Tracking Detectors Overview of primary vertexing techniques Use of vertexing in the ATLAS displaced vertex search Use of vertexing in the ATLAS displaced jet search Use of vertexing in the CMS displaced dijet search Summary
17
November 12, 2013 Matthew Walker, Rutgers University
Searching for decays of Higgs or other scalar bosons to long- lived particles, Hidden Valley Z’ and Stealth SUSY Requiring there to be 2 displaced vertices in the event Looking for displaced vertices in both the inner tracker and in the muon spectrometer
18
November 12, 2013 Matthew Walker, Rutgers University 19
Inner tracker vertexing Use tracks with d0 > 10 mm Remove d0 and z0 significance requirements Loosen requirements on numbers of pixels and strip hits Remove beamspot constraint Apply material cut (signficance > 6) Require number of tracks > 7
November 12, 2013 Matthew Walker, Rutgers University 20
Muon spectrometer vertexing Specialized tracking algorithm Match tracklets from the two layers in a given chamber Cluster tracklets from all chambers using a cone algorithm Calculate phi line-of-flight from cluster Barrel approach All tracklets projected to r-z plane Tracklets back-extrapolated to lines of constant radius in this plane Clustering performed using z intercepts on each line Select the radius and z position with the most tracklets associated
November 12, 2013 Matthew Walker, Rutgers University 21
Endcap approach No momentum or charge measurement Tracklets back-extrapolated as straight lines Iterative vertex finding performed by least squares fit and dropping farthest tracklet until the distance to farthest tracklet is less than 30 cm Vertices required to match to at least 3 tracklets Because extrapolation starts outside magnetic field and goes into magnetic field region, radius and z positions are biased to larger values
November 12, 2013 Matthew Walker, Rutgers University 22
Barrel vertices are required to have a minimum radius of r = 3.5 m Endcap vertices have to be upstream of middle station (z = 14 m) but within the endcap muon fiducial volume
Passing Muon RoI Trigger
November 12, 2013 Matthew Walker, Rutgers University 23
Barrel vertices are required to have a minimum radius of r = 3.5 m Endcap vertices have to be upstream of middle station (z = 14 m) but within the endcap muon fiducial volume
November 12, 2013 Matthew Walker, Rutgers University
Review of ATLAS and CMS Tracking Detectors Overview of primary vertexing techniques Use of vertexing in the ATLAS displaced vertex search Use of vertexing in the ATLAS displaced jet search Use of vertexing in the CMS displaced dijet search Summary
24
November 12, 2013 Matthew Walker, Rutgers University
Searching for decays of a heavy neutral scalar to a pair of long- lived neutral particles that decays to a quark-antiquark pair Identify tracks associated to jets with pT > 60 GeV Tracks have pT > 1 GeV and deltaR < 0.5 to the jet momentum vector Differentiate between prompt and displaced tracks based on the transverse impact parameter to the primary vertex with a cut of 0.5 mm
25
November 12, 2013 Matthew Walker, Rutgers University
For all pairs of jets in the events, combine the sets of displaced tracks and fit a common secondary vertex Vertices required to have at least one track from each jet Chi2/ndof < 5 Distance between primary and secondary vertex should have a significance > 8 Invariant mass of tracks > 4 GeV Vector sum of tracks pT > 8 GeV Average number of missing hits between vertex and start of track < 2 Subsequently cluster tracks in Lxy to verify tracks are consistent with a single vertex
26
November 12, 2013 Matthew Walker, Rutgers University
Analysis then looks in two bins of Lxy split at 20 cm Also apply cuts on the number of prompt tracks in each jet and the energy fraction of the jet carried by prompt tracks
27
November 12, 2013 Matthew Walker, Rutgers University
28