Partially Contained Atmospheric Neutrino Analysis

Andy Blake + John ChapmanCambridge University

January 2004

ν

νμ

μ

UPWARD-GOING MUONS DOWNWARD-GOING MUONS

“direction” problem “containment” problem

Two Categories of PC Event

main background: stopping muons with mis-reconstructed direction

main background: through-going muons that appear contained

shield helps this analysis

Selecting PC Events(1) CONTAIN DIGITS (2) CONTAIN TRACKS

Cambridge Demultiplexer Cambridge Track Reconstruction

Combine digits in adjacent views+ select events with non-contained

hits beside 1 detector edge

Select tracks with > 8 planes + 1 contained vertex

Select events with hits < 0.5m

from 1 detector edge

bottom vertexcontained

top vertex contained

upward-going candidate

downward-going candidate

Event Rates

( kT yr ) -1 MC atmos MC cosmic data

TOTAL FLUX 400 4,500,000

DIGITS > 9 160 2,500,000 3,900,000

PC DIGITS 90 450,000 550,000

PLANES > 8 50 320,000 380,000

TRACK RECO 25 240,000 240,000

PC TRACKS 15 48,000 43,000

PC UP 7 46,000 41,000

PC DOWN 7 2,100 1,800

200,000 MC events

1,000,000MC events

September 2003r18900-19800(~0.12 kT-yr)

stopping muons

Upward-Going PC Events

Upward-Going Muons – Aim

Track Direction Track Topology

Likelihood Analysis

Upward-Going Muons – Timing (1)

(1) Timing

S

CT U viewV view

1/β = -11/β = +1

• Fit S-CT with time slope ± 1• Calculate RMS for each fit• Consider RMSup - RMSdown

MC atmos 97.2 %

MC cosmic 99.8 %

data 99.1 %

Percentage success rate …

Upward-Going Muons - Timing (2)

Timing appears worse for data

Resolution~ 60cm ~ 2ns

• Can also make use of absolute values of RMS …

Upward-Going Muons - Timing (3)

RMSup / RANGE• RMS from fitting wrong time slope

fit

track

3

1

3

)2( 2

2/

0

2/

0

2

2

RANGE

RMSS

dx

dxxRMS

S

S

0

S

Upward-Going Muons - Timing Cuts

5.2 70 370

• RMSup – RMSdown < -0.2 m• RMSup < 2.0 m• RMSdown > 1.0 m• RMSup / RANGE < 0.5• -1.4 < 1/β < -0.6

• RMSup – RMSdown < 0.0 m

1st pass timing cuts

2nd pass timing cuts

3.6 9

(2 evts)

15

(2 evts)

MC atmos

MC cosmic

data

7.3 46,000 41,000

PC digits / tracks

BKG / SIG ~ 1.0

Upward-going Candidates (1)data events: (1) run 19135, snarl 72302

CRATE 15 !

Upward-Going Candidates (2)data events: (2) run 18902, snarl 36351

NEUTRINO CANDIDATE

Upward-Going Candidates (3)MC events: (1) run 231, snarl 44685

Large Angle Scattering

Eμ = 6 GeV

? ?

? ?

?

?? ?

? ?

Upward-Going Candidates (4)MC events: (1) run 242, snarl 65409

Large Scattering Again !!!

Upward-Going Muons – Showers (1)

(2) Vtx Showers• Mop up remaining hits in event• Calculate distance from each track vertex to centre of hits:

Δ VTX = VTXshw - VTXtrk

• Consider Δ VTXup - Δ VTXdown

MC atmos 89.6 %

MC cosmic 38.1 %

data 45.6 %

Percentage success rate …Δ VTXup

Δ VTXdown

for cosmics, showers are distributed

roughly evenly between track vertices,

but slightly more vertex showers

are found at BOTTOM of track

Upward-Going Muons – Showers (2)

Vertex Shower Reconstruction• showers reconstructed in two passes 1st pass – dense “primary” showers ( ≥ 4 planes, ≥ 10 strips) 2nd pass – diffuse “secondary” showers• select the dense showers• eliminates many “fake” muon showers … … but some still found on steep tracks with multiple strips per plane

Upward-Going Muons - Showers (3)

Δ VTXup

Δ VTX = VTXshw - VTXtrk

Quality Cuts

W

ρ ~ Nstrips / W3

shower vertex position: shower density:

Upward-Going Muons - Shower Cuts

BKG / SIG ~ 40.0• shower planes > 4• Δ VTXup < 0.7 m • density > 5.0 strips plane-3

(kT-yr)-1 MC atmos MC cosmic data

PC UP 7.3 46,000 41,000

UP-GOING SHOWER

3.6 12,000 10,000

DENSE SHOWER

1.4 150 160

QUALITY CUTS

1.0 30(6 evts)

50(6 evts)

Upward-Going Candidatesexample data events

neutrino

Downward-Going PC Events

Downward-Going Muons

• 0.5m containment cut on top track vertex removes 99% through-going muons

• most remaining events are steep muons that sneak between the planes

• remove clean background using trace/direction cuts

track containment

Downward-Going Muons – Direction Cuts

Py/P Pz/P

py/p < 0.9 pz/p > 0.2

Downward-Going Muons –Trace Cuts

TRACE Zextrapolate trackto detector edge + calculate Z component

Trace Z > 6 plns

Next Background LayerRemaining background dominated by very steep muons:

• muons travelling significant distances down a single plane • muons turning back on themselves

Downward-Going Muons –Very Steep Muons

Distances of digits from track vertex

• Combine digits in adjacent views around track vertex

• Calculate distance between track vertex and furthest digits

Charge around track vertex

• Plane with maximum charge close to track vertex

ΔR < 1.1 m Q < 500 PE

Downward-Going Muons –Timing Quality

Steep muon events have poor timing

0.5 < | 1/β | < 1.5

Downward-Going Muons -Containment Cuts

(kT-yr) -1 MC atmos MC cosmic data

PC DOWN 7.2 2,100 1,800

DOWN-GOINGTIMING

5.4 2,100 1,800

ANGLE + TRACE

5.0 200 220

Qmax , ΔR

+ TIMING

3.5 30(6 evts)

70(8 evts)

• Py/P < 0.9• Pz/P > 0.2• Trace Z > 6 plns

• Qmax < 500 PEs• ΔR < 1.1 m• 0.5 < 1/β < 1.5• RMSdown < 2.0 m BKG / SIG ~ 10.0

Downward-Going Candidates

6 MONTE CARLO EVENTS• 3 demux errors• 1 tracking error• 1 missing detector• 1 only just contained

8 DATA EVENTS• 1 demux error• 2 tracking errors• 3 missing detector• 2 coil hole• … 0 neutrinos

Demultiplexing Errorsthese hitsshould be

higher

Tracking Errors

Missing Detector

swallowed by LI ?

Conclusion• Analysis is progressing … … but still need to peel away some more layers of background• Need detailed MC/data comparison e.g. high muon scattering MC timing resolution tracks/showers• Need another round of tagging/fixing

reconstruction errors• … but it’s good that neutrinos can be

extracted from the data!