New Thoughts on Muon ID Design and Performance

Talk Presented by P.L McGaughey at ORNL on 5/23/95 at the PHENIX Muon Arm Meeting

The current Muon ID design is very expensive
• A very large channel count, but very low # hits.
• Complicated pad readout structure.
• No systematic study of performance tradeoffs versus channel counts has been done.
• High cost may prevent us from instrumenting central arms with muon ID.
What is the Muon ID system required to do ?
• Provide level 1 trigger: Reject backgrounds from beam gas as well as most non-muons from vertex region.
• Provide level 2 trigger: Provide hit coordinates and depth of penetration for tracker masking.
• Offline reconstruction: Reduce backgrounds from hadron leakage below that of the decay muons.

Can a simple Muon ID consisting of only LST's in an an X-Y geometry do the job?

Comparison of North and South Arm Leakage

Conditions -

• Run PISA with 1 MeV thresholds on everything.
• Use JMM version of UA1 generator.
• Central Au-Au event with lOK multiplicity, 1/3rd each pi- pie piO.
• No neutron and Pb shields.
• New Cu absorber on South Arm.

Results for charged particles -

Notes -

• Rates for the two arms are nearly identical.
• Turning on neutron and lead shields decreases rates in tracker by 30%. Doesn't affect muID.
• Average of < 1 hit/plane on muID 3-5 !!!

Results for Neutral Particles

Notes -

• Turning on neutron and lead shields decreases rates in tracker by ~30%. Little effect on muID.
• Probably don't need these shields which degrade the mass resolution significantly.

Questions -

• With the very low occupancies observed in the muon ID, why are we building a system with almost 20,000 channels of readout ???
• What segmentation of muID is actually required to give near optimal performance?

Graphs

North Muon Arm Hit Distributions for a Au-Au Collision with Multiplicity 5000

Detector hits in the muon tracker and identifier are plotted versus radius and Z direction (along the beam axis). No real tracks are observed in the muon identifier. The track segments observed in the muon tracker are plentiful, but do not point back to the interaction vertex.

Z Vertex Distribution for Same Event Observed in South Muon Arm Tracker

The tracks originate downstream (toward the arm), not from the true interaction vertex. The distribution has a mean Z of about 100 cm and none of the tracks originate in the vertex region.

Z Vertex Distribution for Drell-Yan Dimuon Events

Z vertex distribution for Drell-Yan dimuon events which originated at the true interaction vertex, as observed by the South muon arm tracker. The sigma of the distribution is only 9 cm, demonstrating that the background tracks from the previous figure can all be rejected by a vertex cut.

Z Vertex Distribution for Particles Hitting Back of Muon Identifier Due to 100 Au-H2 Beam-Gas Events

Z vertex distribution for particles hitting the back of the muon identifier due to 100 Au-H2 beam-gas events which originated at 70 m away from PHENIX. Two contributions are evident, one being hadrons produced at the beam-gas vertex, and the second which is muons from hadron decay along the flight path.

Composition of Particles from Beam Beam-Gas Events

Composition of particles from beam beam-gas events which reached the back of the muon identifier. Mainly pions are observed, but a substantial number of protons and decay muons are also evident.

Z Vertex Distribution from X Position Readout of muID Layers 1 through 3

Z vertex distribution from X position readout of muID layers 1 through 3 for muons originating at the nominal interaction vertex. The distribution has a sigma of 112 cm for an infinite resolution in the X coordinate readout. Clearly most beam-gas particles are easily rejected with this vertex resolution.

Z Vertex Distribution with 10 cm Segmentation in the X Position

Z vertex distribution as in the previous figure but with 10 cm segmentation in the X position readout of muID layers 1 through 3. The sigma of the distribution has grown to 177 cm, which is still adequate to reject beam-gas events.

PISA Plot from a Beam-Gas Event Hitting the Muon Identifier

Plot from PISA showing tracks from a beam-gas event hitting the muon identifier. Most hadrons shower in the back, with only one or two particles per event reaching the front of the muon ID. The particle trajectories are of almost normal incidence.

Z Vertex Distributions for Tracks from 32 Beam-Gas Events Determined by X and Y Readout of Muon ID Layers 1 through 3

Z vertex distributions for tracks from 32 beam-gas events as determined by X and Y readout of muon ID layers 1 through 3. For infinite readout resolution no tracks are found within a 10 m by 10 m region around the nominal interaction vertex. The origin of the Au-H2 beam-gas events was at 70 m.

Z Vertex Distributions with 10 cm Segmentation of the X and Y Readout

Z vertex distributions as in the previous figure but with 10 cm segmentation of the X and Y readout. Again no tracks are found near the interaction vertex. Most tracks are off scale due to the normal incidence and size of the segmentation. It appears possible to design a level one trigger which can effectively reject beam-gas events using only X + Y readout. For the level 2 trigger, X + Y from level one would need to be mapped into theta and phi.

Z Vertex Distributions from South Arm Muon Tracker for Decay Muons from 2 x 10^6 UA1 p-p Events

Z vertex distributions from South arm muon tracker for decay muons from 2 X 10^6 UA1 p-p events. These are equivalent to about 2000 central Au-Au events. The distribution has a sigma of 46 cm compared to 9 cm shown in a previous figure for real dimuon events. Thus many of these can be removed by a vertex cut.

Z Vertex Distributions for Hadrons

Z vertex distributions as in the previous figure but for hadrons. Here the sigma is 52 cm and a vertex cut will be even more effective.

Correlation of Penetration Depth of Particles in the Muon ID

Correlation of penetration depth of particles in the muon ID with momentum measured in the muon tracker. The top panel is for decay muons while the bottom panel is for non-muons (mainly hadrons). If muon ID planes 1 through 3 are always required to fire, a simple correlation cut can be defined which eliminates most of the hadrons while preserving nearly all of the muons.

Mass Distributions for Decay Muon and Hadron Pairs from 2 X 10**6UA1 p-p Events

Mass distributions for decay muon and hadron pairs from 2 X 10**6 UA1 p-p events. The particles are required to penetrate to at least the 3rd plane of the muon ID. Above 3 GeV mass, hadron pairs dominate the mass spectrum, showing that additional mu / pi separation is required.

Mass Distributions with a 40 cm (sigma) Vertex Cut and Penetration Depth Versus Momentum Cut

Mass distributions as in the previous figure but adding a 40 cm (sigma) vertex cut and the penetration depth versus momentum cut shown earlier. The hadron background is almost eliminated with decay muons now dominating the mass spectrum everywhere. These cuts have been tested for dimuon events and were found to be ~95 % efficient.

Mass Distributions with the Addition of 5% Occupancy in the Muon ID Planes

Mass distributions as in the previous figure but with the addition of 5% occupancy in the muon ID planes. No real degradation of the hadron rejection occurs and decay muons still dominate the entire mass spectrum. Thus it appears that a simple muon ID system with only X and Y readout is capable of adequate performance.

to PHENIX Home Page

to PHENIX Muon Subsystem Home Page