A Neutrino Factory MIND at Large theta_13

A Neutrino Factory MIND at Large θ13

R. Bayes1, A. Bross3, A. Cervera-Villanueva2 , M. Ellis4,5, TapasiGhosh2 A. Laing1 , F.J.P. Soler1 , and R. Wands3

1University of Glasgow, 2IFIC and Universidad de Valencia, 3Fermilab, 4Brunel University,5Westpac Institutional Bank, Australia,on behalf of the IDS-NF collaboration

NUFACT 201224 July, 2012

R. Bayes (University of Glasgow) A Neutrino Factory MIND at Large θ13 NUFACT, July 2012 1 / 18

1 Introduction

2 Simulation Overview

3 Analysis

4 Sensitivity to δCP

5 Conclusions


Consequences of Large θ13 on Neutrino Factory

Physics priorities shift tomeasurement of CP violation.For measurement of θ13 used

Two baselines: 4000 km and7500 km25 GeV stored µ energy.

Re-optimization of baselineand beam energy requiredMeasurement of δCP achievedwith

Single 2000 km baseline.10 GeV stored µ energy.

0.65

0.65

0.7

0.7 0.7

0.75

sin22!13"10#1

GLoBES2010

0.7

0.75

0.8

sin22!13"10#2

GLoBES2010

0.45

0.45

0.55

0.65

sin22!13"10#3

GLoBES2010

0.3 0.35

0.4

sin22!13"10#4

GLoBES2010

10

15

20

25

E $!GeV"

1000 2000 3000 4000 5000 6000L !km"

5

10

15

20

25

E $!GeV"

2000 3000 4000 5000 6000L !km"

Figure 9. Fraction of ! for which CPV will be discovered (3" CL) as a function of L and Eµ for thesingle baseline Neutrino Factory. The di!erent panels correspond to di!erent true values of sin2 2#13,as given there. Here SF=1 (2.5! 1020 muons per year and polarity) is used with a 50 kt detector. Theoptimal performance is marked by a dot: (2200,10.00), (2288,13.62), (3390,20.00) and (4345,22.08).Figure and caption taken from reference [109].

1.4.2. Optimisation of a two baseline Neutrino Factory

The simultaneous optimisation of two Neutrino Factory baselines has been studied in reference [108],see figure 11 (shaded regions). One easily recovers the behaviour discussed above: sin2 2#13 (upperleft panel) and mass hierarchy (upper right panel) sensitivity prefer a long baseline, but the combi-nation with a shorter baseline is not significantly worse. For CP violation (lower left panel), however,one baseline has to be considerably shorter. If all performance indicators are combined (lower rightpanel), the optimum at 4 000 km plus 7 500 km is obtained. A very interesting study is performed inthe dashed curves: these curves illustrate the e!ect of a potential non-standard interaction $e! , whichis known to harm especially the appearance channels. While the absolute performance for all perfor-

25

√ ×

× ×

From IDS-NF-020, Interim Design Report

MIND simulation used to examine sensitivities with thesespecifications.


New Baseline Neutrino Factory IDS-NF updated baseline design

21/28

Proton Driver HARP: primary beam on production target Target, Capture and Decay MERIT: first create π and later decay into μ Bunching and Phase Rotation

Reduce the spread in energy (ΔE) of bunch Cooling

MICE: Reduce the transverse emittance Acceleration

EMMA: go from 130 MeV to 10 GeV with RLAs or FFAGs Decay Ring Store for roughly 1000 turns; long straight sections

A Staged Approach is conceivable with outstanding physics cases at each stage!

S. K. Agarwalla, 4th EUROnu Annual Meeting, APC, Paris, 13th June, 2012

One 2000 km baseline, 100 kt MIND, 10 GeV muons, 1021 useful decays/year

549 m Single decay ring.Stores both µ+ andµ−.10 GeV final muonenergy.Either RLA or FFAGwill be used in finaldesign.


Physics at a Neutrino Factory

Neutrino Factory provides neutrinos from µ+ and µ− decay.Will produce 5×1020 useful muons of both species per year.

ν Oscillation Channels

Store µ+ Store µ−

Golden Channel νe → νµ ν̄e → ν̄µνe Disappearance Channel νe → νe ν̄e → ν̄eSilver Channel νe → ντ ν̄e → ν̄τPlatinum Channel ν̄µ → ν̄e νµ → νeνµ Disappearance Channel ν̄µ → ν̄µ νµ → νµDominant Oscillation ν̄µ → ν̄τ νµ → ντ

MIND optimized for Golden Channel signal (wrong sign muon).


MIND Design for Neutrino Factory

100 kTon detector14 m×14 m×140 m.X and Y views from 2 cm thick lattice of1 cm×3.5 cm scintillator bars.~B field from 3 cm Fe plates, induced by120 kA current carried by 7 cm diameterSCTL

Superconducting Transmission Line


MIND SimulationEvents simulated with GENIE.

Full geometry & ~B field in GEANT 4

Realistic field map generated by BobWands at FNAL

default positive focussing.

| (Te

sla)

B|

00.20.40.60.811.21.41.61.822.22.4

X-position (m)-8 -6 -4 -2 0 2 4 6 8

Y-po

sitio

n(m

)

-6

-4

-2

0

2

4

6

Dimensions of detectoreasily altered for

optimization.testing variations.


MIND Event Reconstruction

Simulated events digitized.Hits positions smeared andenergy depositionattenuated.Edep clustered into3.5 cm×3.5 cm units.

Tracks identified by KalmanFilter or Cellular automata.Kalman fitting used todetermine momentum andcharge.Algorithms from RecPack.

supported byCervera-Villanueva et al.

50 ν̄µ CC events.

Z Position of Hit in Meters-30 -20 -10 0 10 20 30

Rad

ial P

ositi

on o

f Hit

in M

eter

s

0

1

2

3

4

5

6

7

8sqrt(pow(XPositions,2) + pow(YPositions,2))*0.001:0.001*ZPositions {abs(Evt - 51) < 25 && FittedNodes}

50 νµ CC events.

Z Position of Hit in Meters-30 -20 -10 0 10 20 30

Rad

ial P

ositi

on o

f Hit

in M

eter

s

0

1

2

3

4

5

6

7

8sqrt(pow(XPositions,2) + pow(YPositions,2))*0.001:0.001*ZPositions {abs(Evt - 51) < 25 && FittedNodes}

Fitted hits in red others in black.R. Bayes (University of Glasgow) A Neutrino Factory MIND at Large θ13 NUFACT, July 2012 8 / 18

Cuts Based Golden Analysis

Described in detail in IDR.Separates NC like from CClike events.CC backgrounds arereduced as they are partiallyNC like.

Departures from IDR AnalysisQuadratic and displacementcuts removed.Kinematic cuts replaced by auniform requirementQt > 0.15 GeV

µ+ or µ− Focussing Magnetic Field

Fraction of Events Passed by Cuts

-610 -510 -410 -310 -210 -110 1

Reconstruction Success

Fiducial

Max Momentum

Fitted proportion

Track quality

CC Selection

Kinematic

CC Signalµν

µνCC ID as µν

NC µν rec. from +µ

CC eν rec. from -µ

µνCC ID as τν

µνCC ID as τν

Consider νµ and ν̄µ appearance.







-610 -510 -410 -310 -210 -110 1


Fiducial

Max Momentum

Fitted proportion

Track quality

CC Selection

Kinematic

CC Signalµν

µνCC ID as µν

NC µν rec. from -µ

CC eν rec. from +µ

µνCC ID as τν

µνCC ID as τν








-610 -510 -410 -310 -210 -110 1


Fiducial

Max Momentum

Fitted proportion

Track quality

CC Selection

Kinematic

CC Signalµν

µνCC ID as µν

NC µν rec. from +µ

CC eν rec. from -µ

µνCC ID as τν

µνCC ID as τν








-610 -510 -410 -310 -210 -110 1


Fiducial

Max Momentum

Fitted proportion

Track quality

CC Selection

Kinematic

CC Signalµν

µνCC ID as µν

NC µν rec. from -µ

CC eν rec. from +µ

µνCC ID as τν

µνCC ID as τν



Charge Current Selection Efficiencies

Signal Efficiencies

True Neutrino Energy0 1 2 3 4 5 6 7 8 9 10

Fra

ctio

nal E

ffici

ency

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

CC Signalµν

CC Signalµν

Background

True Neutrino Energy (GeV)0 1 2 3 4 5 6 7 8 9 10

Fra

ctio

nal B

ackg

roun

d

-510

-410

-310

µνCC ID as µνNC µν rec. from -µCCeν rec. from -µ

µνCC ID as τνµνCC ID as µν

NC µν rec. from +µCCeν rec. from +µ

µνCC ID as τν

All reconstruction efficiencies at or above 50%.Background suppressed by parts in 103.NC backgrounds completely suppressed.Efficiency behaviours different for positive and negative focussing.


Energy Response in Octagonal MINDAssuming 10 GeV factory

νµ Appearance Signal

Energy (GeV)νReconstructed 0 2 4 6 8 10

Ene

rgy

(GeV

)ν

Tru

e

0

1

2

34

5

6

78

9

10

-510

-410

-310

-210

-110

νµ App. ν̄µ CC Background

-710

-610

-510

-410

-310


Ene

rgy

(GeV

)ν

Tru

e

0

1

2

34

5

6

78

9

10

ν̄µ Appearance Signal

-510

-410

-310

-210

-110


Ene

rgy

(GeV

)ν

Tru

e 0

1

2

34

5

6

78

9

10

ν̄µ App. νµ CC Background


Ene

rgy

(GeV

)ν

Tru

e

0

1

2

34

5

6

78

9

10

-710

-610

-510

-410

-310

Two ~B settings.

µ+ Focussing.

µ− Focussing.

Field settings runseparately.

Optimization stillrequired


Energy Response in Octagonal MINDAssuming 10 GeV factory

νµ Appearance Signal

-510

-410

-310

-210

-110


Ene

rgy

(GeV

)ν

Tru

e

0

1

2

34

5

6

78

9

10

νµ App. ν̄µ CC Background

-710

-610

-510

-410

-310


Ene

rgy

(GeV

)ν

Tru

e

0

1

2

34

5

6

78

9

10

ν̄µ Appearance Signal

-510

-410

-310

-210

-110


Ene

rgy

(GeV

)ν

Tru

e 0

1

2

34

5

6

78

9

10

ν̄µ App. νµ CC Background

-710

-610

-510

-410

-310


Ene

rgy

(GeV

)ν

Tru

e

0

1

2

34

5

6

78

9

10

Two ~B settings.

µ+ Focussing.

µ− Focussing.

Field settings runseparately.

Optimization stillrequired


Multi-Variate Analysis

A full multivariate analysis is under development.Use a set of correlated variables to select signal from background.

Should be able to achieve higher efficiency than existing analysis.Still a work in progress.

Variety of Methods Tested

Signal efficiency0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Bac

kgro

un

d r

ejec

tio

n

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

MVA Method:BDTGBDTKNNRuleFitMLPCutsDLikelihoodLikelihoodPCAFDA_GA

LDCuts

Background rejection versus Signal efficiency

Example KNN Method

Cut value applied on KNN output0 0.2 0.4 0.6 0.8 1

Eff

icie

ncy

(P

uri

ty)

0

0.2

0.4

0.6

0.8

1

Signal efficiency

Background efficiency

Signal puritySignal efficiency*purity

S+BS/

For 624 signal and 50059 background isS+Bevents the maximum S/

16.3564 when cutting at 0.9876

Cut efficiencies and optimal cut value

Sig

nif

ican

ce

0246810121416


Extracting Sensitivities from SimulationUse the simulation to produce "migration matrices"

Relates true neutrino energy to reconstructed energy.Contains efficiency, energy resolution, and response information.

Run pseudo-experiments with simulation package (ie. NuTS).

Data is composed of sum

ndatai = Msig

ij νsig(Ej) +∑

k

Mbkg,kij νbkg,k (Ej)

Define a fit of θ13 and δCP , simultaneously

χ2 = 2L∑

i=0

(AxN+,i(θ13, δCP)− ndata

+,i + ndata+,i ln

(ndata

+,iAxN+,i (θ13,δCP)

)+AN−,i(θ13, δCP)− ndata

−,i + ndata−,i ln

(ndata−,i

AN−,i (θ13,δCP)

)+ (A−1)2

σ2A

+ (x−1)2

σx

)

Systematic uncertainties —Number of Interactions & Known ν/ν̄ difference.


Preliminary Precision in δCP For Octagonal MINDAssuming 10 GeV Factory, 10 years Running, 0.5× 1021 µ+ + 0.5× 1021 µ− per year

Uses cuts-based analysis.Consider µ+ and µ− focussingSystematic variations shown(σA, σX )=(1%, 1%)→(2.5%, 3%).

χ2 Contours; Arbitrary δCP , θ13 = 9◦

13θ22sin0.07 0.08 0.09 0.1 0.11 0.12 0.13 0.14 0.15

(deg

)C

Pδ

-150

-100

-50

0

50

100

150 sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

Error in δCP from 1 σ curves

(Deg)CPδ-200 -150 -100 -50 0 50 100 150 200

(D

eg)

CP

δ∆

0

1

2

34

5

6

78

9

10

(Deg)δ∆0 2 4 6 8 10

Fra

ctio

nδ

0

0.1

0.2

0.30.4

0.5

0.6

0.70.8

0.9

1


Preliminary Sensitivity to δCPAssuming 10 GeV Factory, 10 years Running, 0.5× 1021 µ+ + 0.5× 1021 µ− per year

Contours determined using the expression

max(χ2(δCP = −180◦), χ2(δCP = 0◦), χ2(δ013 = 180◦))− χ2

min ≥ n2

Sensitivity curves

13θ22sin0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16

(deg

)C

Pδ

-150

-100

-50

0

50

100

150 sensitivityσ3

sensitivityσ5

sensitivityσ10

13θDaya Bay

5σ Discovery Potential

13θ22sin0 0.02 0.04 0.06 0.08 0.1 0.12 0.14

Fra

ctio

nC

Pδ

0

0.2

0.4

0.6

0.8

1

Both normal and inverted hierarchy cases are considered.


What About Dipole Geometry at 10 GeV?Assuming 10 GeV Factory, 10 years Running, 0.5× 1021 µ+ + 0.5× 1021 µ− per year

Achieved 10−4 backgroundrejection in IDRSimilar efficiency.Optimized for two baselines.

χ2 Contours; Arbitrary δCP , θ13 = 9◦

13θ22sin0.07 0.08 0.09 0.1 0.11 0.12 0.13 0.14 0.15

(deg

)C

Pδ

-150

-100

-50

0

50

100

150 sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

sensitivityσ1

sensitivityσ3

sensitivityσ5

Error in δCP from 1 σ curves

(Deg)CPδ-200 -150 -100 -50 0 50 100 150 200

(D

eg)

CP

δ∆

0

1

2

34

5

6

78

9

10

(Deg)δ∆0 2 4 6 8 10

Fra

ctio

nδ

0

0.1

0.2

0.30.4

0.5

0.6

0.70.8

0.9

1


What does this tell us?

δCP Sensitivity curves

13θ22sin0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16

(deg

)C

Pδ

-150

-100

-50

0

50

100

150 sensitivityσ3

sensitivityσ5

sensitivityσ10

13θDaya Bay

5σ Discovery Potential

13θ22sin0 0.02 0.04 0.06 0.08 0.1 0.12 0.14

Fra

ctio

nC

Pδ

0

0.2

0.4

0.6

0.8

1

δCP precision and sensitivity comparable between dipole andtoroidal analyses for the same experimental conditions.Reduction of backgrounds does not increase sensitivity.Will reduction of energy threshold help?


Outlook

There has been great progress in the past year.Complete change to GENIE event generator.Introduction of realistic octagonal geometryImprovement of reconstruction to allow toroidal field.

Still work in progressReconstruction of secondary, hadron tracks.Update to use multi-variate analysis.Optimize analysis for physics outcomes.

Early conclusionsPreliminary precision of δCP between 2◦ and 8◦ with toroidalmagnetic field.


A Neutrino Factory MIND at Large theta_13

Documents

Transcript of A Neutrino Factory MIND at Large theta_13