Accelerator Neutrino Oscillation Physics Lecture II

Accelerator Neutrino Oscillation Physics

Lecture II

Deborah Harris

Fermilab

SUSSP

August 16, 2006

16 August 2006 Deborah Harris Accelerator Neutrino Oscillation Physics Lecture II 2

Outline of this Lecture

• Introduction– What are the detector goals?

– Particle Interactions in Matter

• Detectors– Fully Active

• Liquid Argon Time Projection • Cerenkov (covered mostly in Atmospheric Lectures)

– Sampling Detectors• Overview: Absorber and Readout• Steel/Lead Emulsion • Scintillator/Absorber• Steel-Scintillator


For Each Detector

• Underlying principle

• Example from real life

• What do events look like?– Quasi-elastic Charged Current

– Inelastic Charged Current

– Neutral Currents

• Backgrounds

• Neutrino Energy Reconstruction

• What else do we want to know?

All detector questions are far from answered!

Warning: this lecture is a set of mini-lectures about different detectortechnologies…


Detector Goals

• Identify flavor of neutrino– Need charged current events!

– Lepton Identification (e,)

• Measure neutrino energy– Charged Current Quasi-elastic Events

• All you need is the lepton angle and energy

• Corrections due to – p,n motion in nucleus

– Everything Else• Need to measure energy of lepton and of X, where X is the hadronic

shower, the extra pion(s) that is (are) made..

E

LmP

4sin2sin

222

pln

nlp

lXN


Making a Neutrino Beam

2For each of these beams,

flux (Φ) is related to boost of parent particle ()

• Conventional Beam

• Beta Beam

• Neutrino Factory


Goals vs Beams

Conventional Beams (, %e)– Identify muon in final state– Identify electron in final state, subtract backgrounds– Energy regime: 0.4GeV to 17GeV

beams (all e )– Idenify muon or electron in final state– Energy regime: <1GeV for now

Neutrino Factories (, e)– Identify lepton in final state– Measure Charge of that lepton!

• Charge of outgoing lepton determines flavor of initial lepton

– Energy regime: 5 to 50GeV ’s


Next Step in this field: appearance!

13 determines

– If we’ll ever determine the mass hierarchy– The size of CP violation

• How do backgrounds enter?

– Conventional beams: → e• Already some e in the beam• Detector-related backgrounds:


Why do detector efficiencies and background rejection levels matter?

• Which detector does better (assume 1% -e oscillation probability)

– 5 kton of • 50% efficient for e

• 0.25% acceptance for NC

– 15kton of • 30% efficient for e

• 0.5% acceptance for NC events?

Assume you have a convenional neutrino beamline which produces:•1000 CC events per kton (400NC events)•5 e CC events per kton

Background: ( 5*.5 e + 400*.0025NC)x5=17.5

Signal: (1000*.01*.5)x5=25, S/sqrt(B+S)=3.8

Background: ( 5*.3 e + 400*.005NC)x15=52.5

Signal: (1000*.01*.3)x15=45, S/sqrt(B+S)=4.6


Now for a Factory…Assume you have a neutrino factory which produces:• 500 CC events per kton (200NC)• 1000 e CC events per kton (400NC)

Again, assuming 1% oscillation probability, but now the backgrounds are 10-4 (for all kinds of interactions), the signal efficiency is 50%, and again you have 15kton of detector (because it’s an easy detector to make)…

Background: ( .0001*2100(CC+NC))x15=3

Signal: (1000*.01*.5)x15=150, S/sqrt(B+S)=12

Get a “figure of merit” of 12 instead of 3 or 4…which is like getting a 12 result instead of a 4 result,

or being sensitive at 3 to a 10 times smaller probability!

Note: as muon energy increases, you get more /kton for a factory!


Particles passing through material

Particle Characteristic Length Dependence

Electrons Radiation length (Xo) Log(E)

Hadrons Interaction length (INT) Log(E)

Muons dE/dx E

Taus Decays first ct=87m

Material Xo

(cm)

INT(cm) dE/dx

(MeV/cm)

Density

(g/cm3)

Liquid Argon 14 83.5 2.1 1.4

Water 37 83.6 2.0 1

Steel 1.76 17 11.4 7.87

Scintillator (CH) 42 ~80 1.9 1

Lead 0.56 17 12.7 11.4


Liquid Argon TPC (ICARUS)

• Electronic Bubble chamber

• Planes of wires (3mm pitch) widely separated (1.5m) 55K readout channels!

• Very Pure Liquid Argon

• Density: 1.4, Xo=14cmINT =83cm

• 3.6x3.9x19.1m3 600 ton module (480fid)


View of the inner detector

Half Module of ICARUS


Liquid Argon TPC

• Because electrons can drift a long time (>1m!) in very pure liquid argon, this can be used to create an “electronic bubble chamber”

Raw

Data to R

econstructed Event


Principle of Liquid Argon TPC

Time

Drift direction

Edrift

•High density

•Non-destructive readout

•Continuously sensitive

•Self-triggering

•Very good scintillator: T0

dE/dx(mip) = 2.1 MeV/cmT=88K @ 1 barWe≈24 eVW≈20 eVCharge recombination (mip)@ E = 500 V/cm ≈ 40%

Readout planes: Q

Continuous waveform recording

Low noise Q-amplifier


dE/dx in Materials

• Bethe-Bloch Equation • x in units of g/cm2

• Energy Loss Only f()• Can be used for Particle ID in range of momentum

2

2ln

2

11 22

max222

22

I

Tcm

A

Zz

dx

dE e


Bethe-Bloch in practice

• From a single event, see dE/dx versus momentum (range)

AB

BC

K+

µ+

Run

939

Eve

nt 4

6

ABC

K+µ+

e+

Question: how wouldThis look different for a → event?


Examples of Liquid Argon Events

• Lots of information for every event…

e-, 9.5 GeV, pT=0.47 GeV/c

CNGS interaction, E=19 GeV

- e- + e + _

e-, 15 GeV, pT=1.16 GeV/c

Vertex: 10,2p,3n,2 ,1e-

CNGS e interaction, E=17 GeV

CourtesyAndré Rubbia

Primary tag:edecayExclusive tag:decayPrimary Bkgd: Beam e


0 identification in Liquid Argon

Preliminary

<dE/dx> MeV/cm

cut

1 π0 (MC)One photon

converts to 2 electrons before showering, so dE/dx for photons is higher…

Imaging provides ≈210-3 efficiency for single 0


Oustanding Issues

• Do Simulations agree with data (known incoming particles)

• Can a magnetic field be applied• Both could be answered in CERN

test beam program • Is neutral current rejection that

good?• How large can one module be

made? • What is largest possible wire plane

spacing?

Liquid Argon Time Projection Chamber

Several R&D Efforts world-wideworking to get >10kton detectors “on the mass shell”


Cerenkov Light

• For water, n(280-580nm)~1.33-6,so pthreshold≈1.3*mass

• Threshold Angle: 42o

• What is Threshold momentum for neutral pions?

As particles move faster than the speed of light in that medium, they emit a “shock wave” of light

n

1

c

v

1

))(/1(cos

2

1

n

mp

n

threshold

c

particle p (threshold)

e 660keV

137MeV

± 175MeV

K 650MeV

p 1300MeV


Event Reconstruction in Cerenkov Detector

• Vertex Point fit: time of flight should be as sharp as possible

• Define set of in-time tubes• Use Hough Transform to find rings• Look for rings until

you’re done• Particle ID • Corrections to

Vertex • Energy Reconstruction• Decay Electron Finding


Particle ID Using Cerenkov Light

Super-Kamiokande detector

50,000 ton water Cherenkov detector

(22.5 kton fiducial volume)

1000m underground (2700 m.w.e.)

11,146 20-inch PMTs for inner detector

1,885 8-inch PMTs for outer detector


Single-Ring Energy Resolution

• Tested in situ with LINAC at KEK


MiniBooNE detector: Cerenkov with Mineral Oil

total volume: 800 tons (6 m radius)fiducial volume: 445 tons (5m radius)1280 PMTs in detector at 5.5 m radius10% photocathode coverage 240 PMTs in veto

electron ring ring Events courtesy G. Zeller


What about water Cerenkov at High (>1GeV) Energies???

Cou

rtes

y M

ark

Mes

sier

: on

e is

e s

igna

l, on

e is

0 b

ackg

roun

d

VisibleEnergy

=2GeV:One

Ise-

One Is a 0


(E) of Water Cerenkov vs E

e

CC

CC

NC

Rec

onst

ruct

ed/T

rue

Ene

rgy


Oustanding Issues

• What is largest vessel that can be made? (48mx58mx250m?)

• What is highest energy regime that is possible, with better electronics,photo-detectors, etc?

• Water Cerenkov clearly the cheapest per kton

Cerenkov Detectors


recon for Water Cerenkov

• Again, courtesy Mark Messier, for Fermilab to Homestake Study

Probability of e CC Giving 1 e-like ring

Probability of CC Giving 1 -like ring

Probability of NC Giving 1 e-like ringGiving 1 -like ring

Reconstructed Energy (GeV)




From Fully Active to Sampling

• Advantages to Sampling:– Cheaper readout costs– Fewer readout channels– Denser material can be used

• More N, more interactions• Could combine emulsion with readout

– Can use magnetized material!• Disadvantages to Sampling

– Loss of information– Particle ID is harder (except emulsion for taus

in final state)

N

X

electronE

electronE

NhadronE

hadronE

samplesNNE

E

INT

0

)(

)(

)(

)(

,1


Sampling calorimeters

• High Z materials: – mean smaller showers,– more compact detector– Finer transverse segmentation needed

• Low Z materials:– more mass/X0 (more mass per instrumented plane)– Coarser transverse segmentation– “big” events (harsh fiducial cuts for containment)

Material Xo (cm) lINT

(cm)

Sampling (Xo) Xo

(g/cm2)

L.Argon 14 83.5 .2 (ICARUS) 20

Steel 1.76 17 1.4 (MINOS) 14

Scintillator 42 ~80 .2 (NOA) 40

Lead 0.56 17 .2 (OPERA) 6


detection (OPERA)

• Challenge: making a Fine-grained and massive detector to see kink when tau decays to something plus


Lead-Emulsion Target

Emulsion films (Fuji)Emulsion films (Fuji)production rate ~8,000m2/month

(206,336 brick ~150,000m2)

Lead plates Lead plates (Pb + 2.5% Sb)(Pb + 2.5% Sb)requirements:

low radioactivity level,emulsion compatibility,constant and uniform thickness

6.7m

Wall prototypeWall prototype

2 emulsion layers (44 m thick)glued onto a

200 m plastic base

52 x 64 bricks52 x 64 bricks

12.5cm

10.2cmPb1 mm

8.3kg

10 X0’s

BRICK: 57 emulsion foils +56 interleaved Pb plates

Total target mass : 1766 t


Particle ID in Emulsion

pionsprotons

Test exposure (KEK) : 1.2 GeV/c pions and protons, 29 plates

Grain density in emulsion is proportional to dE/dxGrain density in emulsion is proportional to dE/dx

By measuring grain density By measuring grain density as a function of as a function of

the distance fromthe distance from the stopping pointthe stopping point,,

particle identification particle identification can be performedcan be performed. .

Plo

ts c

ourt

esy

M. D

eSer

io


One cannot live by Emulsion alone

• Need to know when interaction has happened in a brick• Electronic detectors can be used to point back to which

brick has a vertex• Take the brick out and scan it

(don’t forget to put a new brick in!)

• Question: what can you use for the “electronic detectors” that point back to the brick?

• (Hint: you’ve used up most of the money you have to buy emulsion, you need something cheap that can point well anyway)

Track segments found in 8 consecutive plates

Connected trackswith >= 2 segments

Passing-through tracks rejection

Vertex reconstructin


Muon Spectrometer w/RPC

B=

1.5

5 T

base

slabs

coil

8.2 m

M ~ 950t

Drift tubes

12 Fe slabs(5cm thick)

RPC’s

2 cm gaps

identification: identification: > 95% (TT)> 95% (TT)

p/p < 20% ,p/p < 20% , p < 50 GeV/cp < 50 GeV/c

charge charge Mis-id prob.Mis-id prob.

0.1 0.1 0.3% 0.3%

21 bakelite RPC’s (2.9x1.1m2) / plane (~1,500m2 / spectrometer)

pickup strips, pitch: 3.5cm (horizontal), 2.6cm (vertical)

Inner Tracker: Inner Tracker: 11 planes of RPC’s11 planes of RPC’s

Precision tracker: Precision tracker: 6 planes of drift tubes6 planes of drift tubes diameter 38mm, length 8m

efficiency: 99%space resolution: 300µm

RPC: gives digital information about track: has been suggestedfor use in several “huge mass steel detectors” (Monolith)


detection (OPERA)

• Detection Efficiency


backgrounds

• Cut on invariant mass of primary tracks


events expected (OPERA)

Comparison: 4 events

over 0.34 background

atDONUT .27kton

mm22 ( x 10-3 eV2)Back-

ground 1.9 2.4 3.0

µ 2.2 3.6 5.6 0.23

e 2.7 4.3 6.7 0.23

h 2.4 3.8 5.9 0.32

3h 0.7 1.1 1.7 0.22

Total 8.0 12.8 19.9 1.0


Outstanding Issues

• If LSND signature is oscillations, appearance will be much more important in the future

• For future neutrino factory experiments, could study e→

• For either of these topics, need to understand if/how magnetic field can be made…

• Any way to make this detector more massive?

Emulsion Sampling


90 m

All Scintillator Detector

• PVC extrusions– 17m tallx17m widex90m long– 3.87 cm transverse, 6 cm wide in

beam direction (more light)– 17.5 m long vs. 48 ft (less light)

• All Liquid Scintillator– 85% scintillator, 15% PVC

• Previous design: inactive (particle board) plus active (scintillator) material, but was less efficient at rejecting backgrounds

APD readouton TWO edges

To Build:Glue Planes of Extrusions together

Rotate them from horizontal to vertical Fill Extrusions with Liquid Scintillator

Each box gets a WLS fiber loop (bent at far end)Instrument WLS fibers with Advanced PhotoDiodes


Scintillator Events (2GeV)

One unit is 4.9 cm (horizontal)4.0 cm (vertical)

Particle ID: particularly “fuzzy” e’s long track, not fuzzy () gaps in tracks ( 0 ?) large energy deposition (proton?)

+ A -> p + 3± + 0 +

e+A→p + - e-

+ A -> p +-


Detector Volume• Scaling detector volume is not

so trivial

• At 30kt NOvA is about the same mass as BaBar, CDF, Dzero, CMS and ATLAS combined…– want monolithic, manufacturabile structures

– seek scaling as surface rather than volume if possible

Figure courtesy J.Cooper


Detector Volume, continued (courtesy K.McFarland)

17m

• It took 700 years to complete• Fortunately construction technology has improved

• Left:

120m

• Consider the Temple of the Olympian Zeus (right)…

• 17m tall, just like NOvA!

– a bit over ½ the length


Energy Resolution

• For e CC events with a found

electron track

(about 85%),

the energy resolution is

10% / sqrt(E) Measured – true energy divided by square root of true energy

• This helps reduce the NC and CC backgrounds since they do not have the same narrow energy distribution of the oscillated e’s (for the case of an Off Axis beam)

N

X

electronE

electronE

samplesNNE

E

0

)(

)(

,1

Energy Resolution


All Scintillator / e separation

• This is what it means to have a “fuzzy” track– Extra hits, extra pulse height

• Clearly CC are separable from e CC

Average pulse height per plane Average number of hits per plane

muons

electronselectrons

muons


Outstanding Issues

• How cheaply can this be made?

• Do you need any passive absorber?

• What is best choice for readout?

• Must have confidence in ability to reduce Neutral Current Backgrounds

Fine GrainedScintillator


Steel/Scintillator Detector (MINOS)

• 8m octagon steel & scintillator calorimeter

• Sampling every 2.54 cm

• 4cm wide strips of scintillator

• 5.4 kton total mass

• 486 planes of scintillator

• 95,000 strips


Simulated MINOS Events

• Long muon track + hadronic activity at vertex

CC Event NC Event e CC EventUZ

VZ

3.5m 1.8m 2.3m

• Short showering event, often diffuse

• Short event with typical EM shower profile

E = Eshower + P

Shower energy resolution: 55%/√EMuon momentum resolution: 6% range; 13% curvatureC

ourt

esy

Chr

is S

mit

h, F

NA

L S

emin

ar


Real Beam Events at MINOS (Far) CC (left)

e CC candidate(bottom)

Remember:2.5cm thicksteel plates(~1.5X0)

(Courtesy C. Smith, FNAL seminar)

U-Z View

V-Z View

X-Y View

T-Z View

PH-Z View


Steel Scintillator Response

Response measured in CERN test beam using a MINI-MINOS (1mx1m)

MC expectation

Provides calibration informationTest of MC simulation of low energy hadronic interactionsQuestion: why might EM response be higher than hadronic response?


Hadron/Electron Comparison

• Electromagnetic response: photons always convert to electrons which deposit all their energy nearby

• Hadronic response: when neutrons are created in the shower, they don’t deposit energy nearby, and often just get absorbed!


Outstanding Issues

• For Neutrino factory Application: what transverse and longitudinal segmentation is needed?

• Any way to make this detector cheaper?

Steel/Scintillator


There are huge detector demands on the next generation of detectors1. Size*signal efficiency2. Background rejection (NC) 3. “Ability to do other physics” Water Cerenkov the most economical choice, but we need more to get to high

energies and matter effects!

Detector

Technology

Largest Mass to Date (kton)

Event by Event Identification +/-?

Ideal

Energy

Rangee

LAR TPC 0.6 Not yet huge

Water Cerenkov 50 <2GeV

Emulsion/Pb/Fe 0.27 >.5GeV

Scintillator++ 1 or less huge

Steel/Scint. 5.4 >.5GeV

Detector Scorecard


Oscillation Experiments: Detectors past, present, and near future…

Exp’t Energy

(GeV)

Detector Technology

K2K 1.4 Water Cerenkov

MINOS 2-6 Steel Scintillator

OPERA 15-25 Emulsion-Lead

ICARUS 15-25 Liquid Argon TPC

T2K 0.7 Water Cerenkov

NOvA 2 Segmented Scintillator

e-, 15 GeV, pT=1.16 GeV

Vertex: 10,2p,3n,2 ,1e-

e+A→p + - e- CC

e CC

OPERA

ICARUS

MINOSNOvA

Super-K

Accelerator Neutrino Oscillation Physics Lecture II

Documents

Transcript of Accelerator Neutrino Oscillation Physics Lecture II