Reactor Oscillation Experiment Basics
νe
νe
νe
νe
νe
Unoscillatedν flux
e
observed here
Well understood, isotropic source
of electron anti-neutrinos
Oscillations observed
as a deficit of νe
Detectors are located
underground to shield
against cosmic rays.
Probability νe
1.0
sin22θ13
πEν /2Δm213
P( ν e  ν e )  1  sin 2 2θ13 sin 2 (m132 L / 4 Eν )
Distance (L/E)
~1800 meters
(@ 3 MeV)
The Existing Limit on θ13
Come from the Chooz and Palo Verde reactor experiments
Neither experiments found evidence for ne oscillation
The null result eliminated nm→ne as the primary mechanism for
the atmospheric deficit
Remember the oscillation probability

P( ν e ν x )sin 2 2θ13sin 2 Δm132 L 4 Eν

So these experiments are sensitive to
sin22θ13 as a function of Δm213
sin22q13< 0.13 at 90% CL
CHOOZ
• Homogeneous detector
• 5 ton, Gd loaded, scintillating
target
• 300 meters water equiv. shielding
• 2 reactors: 8.5 GWthermal
• Baselines 1115 m and 998 m
• Used new reactors → reactor off
data for background measurement
Chooz Nuclear Reactors, France
Palo Verde
Palo Verde Generating Station, AZ
• 32 mwe shielding (Shallow!)
• Segmented detector:
Better at handling the
cosmic rate of a shallow site
• 12 ton, Gd loaded, scintillating
target
• 3 reactors: 11.6 GWthermal
• Baselines 890 m and 750 m
• No full reactor off running
CHOOZ and Palo Verde Results
• Neither experiments found evidence for ne oscillation.
• This null result eliminated nm→ne as the primary mechanism for
the Super-K atmospheric deficit.
• sin22q13< 0.18 at 90% CL (at
m2=2.0×10-3)
• Future experiments should try to
improve on these limits by at least
an order of magnitude.
Down to sin22q13 <~ 0.01
In other words, a 1%
measurement is needed!
Nuclear Reactors as a Neutrino Source
Nuclear reactors are a very intense sources of νe coming from
the b-decay of the neutron-rich fission fragments.
A typical commercial reactor,
with 3 GW thermal power,
produces 6×1020 νe/s
The observable ne spectrum is
the product of the flux and the
cross section.
Arbitrary
From Bemporad, Gratta and Vogel
Observable ν Spectrum
Reactor Neutrino Event Signature
The reaction process is inverse β-decay
ne p→ e+n
n capture
Two part coincidence signal is crucial for background reduction.
Minimum energy for the primary signal of 1.022 MeV from e+e−
annihilation at process threshold.
Positron energy spectrum implies the anti-neutrino spectrum
Eν = Ee + 0.8 MeV ( =mnmp+me1.022)
In pure scintillator the neutron would capture on hydrogen
n H → D g (2.2 MeV)
Scintillator will be doped with gadolinium which enhances capture
n mGd → m+1Gd g’s (8 MeV)
Why Use Gadolinium?
Gd has a huge neutron capture cross section. So you get faster
capture times and smaller spatial separation. (Helps to reduce
random coincidence backgrounds)
With Gd
Without Gd
~30 μs
With Gd
Without Gd
~200 μs
Also the 8 MeV capture energy (compared to 2.2 MeV on H)
is distinct from primary interaction energy.
Neutrino Interactions in the Detector
Inverse β-decay
makes a nice
coincidence
signal in the
detector.
νe
e+
e
np
157
158
64
Gd
First burst of
light from the
positron.
10’s of μs later…
Delayed burst of
light from
neutron capture.
How Do You Measure a Small Disappearance?
A Simple Sensitivity Model
R
N far L2far
2
near
N near L
 < 1– 3σR means an effect is observed
Where N is the number of observed signal events, L is the
baseline and ε is the relative efficiency (≈1). Then…
2
 R   stat
  2   bg2
Where…
 stat 
N far  N bg
N far
Statistics
 

pair
n 1
Relative
Normalization
 bg rate  N bg
 bg 
N far
Background
Statistics
Ways to optimize statistics…
• Reactor power
Daya Bay is one of the most powerful nuclear plants in the
world with 6 cores online by 2011
• Detector mass
With a total of 80 tons at the far site and no fiducial mass cut
Daya Bay will be an order of magnitude larger than any
previous short baseline reactor neutrino experiment
• Run time
Three years run time will be two years more than previous
experiments
• Optimized baseline for known value of Δm2
Relative Normalization
The use of a near detector eliminates the normalization
uncertainty due to
• Inverse β-decay reaction cross section
• neutrino production in the reactor core
• reactor power
Truly identical detectors would eliminate the remaining sources
of normalization uncertainty which are
• detector efficiency
• gadolinium fraction (neutron detection efficiency)
• free proton count (neutrino target size and density)
• geometric acceptance
Background
The vast majority of backgrounds are directly related to cosmic
rays
There are three types of background:
1. Random coincidence ─ two unrelated events happen close
together in space and time
1. Fast neutron ─ fast neutron enters
detector, creates prompt signal,
thermalizes, and is captured
2. β+n decays of spallation isotopes
─ such as 9Li and 8He with β+n
decay modes can be created in μ
12C spallation event
(1%)
Random Coincidence Background
The rate of coincident events can be determined by studying the
rates for positron and neutron capture like events in the detector
The singles rates from long lived spallation isotopes and the U, Th
and K decay chains is shown below.
Assuming KamLAND
concentrations of 40K,
232Th and 238U and
450 mwe
Calculated rates for
Braidwood.
Plot by Hannah
Newfield-Plunkett
Positron-like events are
between ~2 and 8 MeV
Neutron events are ~6 to ~10
MeV and include neutron
captures from muon induced
neutrons which are not shown
Hannah was a bright high school
student who worked with me for a
couple of summers and is now a
Cornell undergraduate student.
Fast Neutron Backgrounds
There are three main processes for the prompt “positron-like”
events
1. Two neutron captures from the same cosmic ─ This
should be tagged the vast majority of the time, but it
sets the tag window for tagged muons at 100 μs.
2. Proton recoil off fast neutron ─ dominate effect.
3. Fast neutron excitation of 12C ─ interesting, but not
significantly different than 2. Energy spectrum peaks
at particular values (like 4.4 MeV, first 12C excited
state)
Tagging Muons at Daya Bay
The basic idea is to tag muons that pass near the detector so
that we can reject the fast neutron background. Neutrons from
farther away should be mostly ranged out.
p
n
m
n
m
Correlated Spallation Isotopes
Isotopes like 9Li and 8He can
be created in μ spallation on
12C and can decay to β+n
KamLAND found that the
spallation is almost exclusively
9Li
This production is correlated
with μ’s that shower in the
detector
from the thesis of Kevin McKinny
Therefore we can account for these events by looking at the
separation in time of candidate events from energetic showers
muon showers.
Background Summary
(a)
(d)
(c)
(b)
(1%)
Total expected background rates:
far site < 0.4 events/det/day
Daya Bay site < 6 events/det/day
Ling Ao site < 4 events/det/day
Sensitivity To Shape Deformation
sin22θ13 Sensitivity
Assumes negligible background; σcal relative near/far energy calibration
σnorm relative near/far flux normalization
90%CL
at Δm2 = 3×10-3 eV2
Huber et al hep-ph/0303232
400
8000
Exposure (ton GWth year)
Daya Bay Projected Sensitivity
For three years of Daya Bay data and Δm2 ≈ 2.5×10-3 eV2
0.006
90% CL limit at sin22θ13 < 0.008
0.013
3 σ discovery for sin22θ13 > 0.015
With Swapping
Source of Uncertainty
Far Statistical per Det.
Near Statistical per Det.
Reactor Related
Relative Normalization
Background (Near)
Background (Far)
Sensitivity to sin22θ13
%
0.3
0.1
0.1
0.38
0.12
0.3
0.1
Non-Reactor Handles on θ13
The reactor oscillation experiment is what is known as a
disappearance experiment since it is only sensitive to the
original neutrino type (νe)
Electron neutrinos oscillate into νμ or ντ which can’t
produce a μ or τ at reactor energies
Therefore, when oscillations occur a fraction of neutrinos
seem to disappear
Another class of experiments, known as appearance
experiments, are sensitive to the new neutrino types
But the expression for the oscillation probability is much
more complicated in these experiments
Accelerator Based θ13 Oscillation Experiments
•
•
•
•
Appearance νμ→νe (or νμ→νe with separate running)
Off-axis beam results in a mono-energetic νμ beam
Long baseline (300 – 900 km)
Needs a very large detector
NOnA
The oscillation probability for νμ→νe is given by
MINOS
P(νμνe) = sin2θ23 sin22θ13 sin2(1.27 Δm132 L/E)
+ cos2θ23 sin22θ12 sin2(1.27 Δm122 L/E)
± J sin δ sin(1.27 Δm132 L/E) (CP Violating Term)
+ J cos δ cos(1.27 Δm132 L/E)
NOνA (Fermilab)
where J = cosθ23 sin 2θ12 sin 2θ23 ×
sin(1.27 Δm132 L/E) sin(1.27 Δm122 sin 2θ13 L/E)
T2K (Japan)
νe Appearance Probability
P (ν μ  ν e )  sin 2 θ 23 sin 2 2θ13 sin 2 (1.27m132 L E )
 cos2 θ 23 cos2 θ13 sin 2 2θ12 sin 2 (1.27m122 L E )
 cosθ 23 sin θ12 sin 2θ13 sin 2θ 23 sin δ
 sin 2 (1.27m132 L E ) sin(1.27m122 L E )
 cosθ 23 sin θ12 sin 2θ13 sin 2θ 23 cosδ
 sin(1.27m132 L E ) cos(1.27m132 L E ) sin(1.27m122 L E )
CP Asymmetry…
P (ν μ  ν e )  P (ν μ  ν e )  cosθ 23 sin θ12 sin 2θ13 sin 2θ 23 
sin δ sin 2 (1.27m132 L E ) sin(1.27m122 L E )
The θ23 Degeneracy Problem
Atmospheric neutrino measurements are sensitive to sin22θ23

2
P( ν μ ν x )sin 2 2θ 23sin 2 m23
L 4 Eν

But the leading order term in νμ→νe oscillations is

P( ν μ ν e )sin 2 θ 23sin 2 2θ13sin 2 m132 L 4Eν

No 2!
sin2
If the atmospheric oscillation is not exactly maximal
(sin22θ23≠1) then sin2θ23 has a twofold degeneracy
sin22θ23
sin2θ23
θ
45º
θ
2θ
90º
2θ
More On Degeneracies
There are additional degeneracies due to the unknown CP phase
and the unknown sign of the mass hierarchy
Combining experimental results can resolve these degeneracies
McConnel & Shaevitz
hep-ex/0409028
Daya Bay +T2K
T2K only (5yr,n-only)
Double Chooz+T2K
90% CL
Daya Bay (3 yrs) + Nova
Nova only (3yr + 3yr)
90% CL
Double Chooz + Nova
Need the more sensitive reactor experiment to resolve degeneracies
Sensitivity to CPV and Mass Hierarchy
McConnel & Shaevitz
hep-ex/0409028
Daya Bay
Daya Bay
?
The accelerator experiments may be sensitive to CP violation and
the mass hierarchy, but if Daya Bay sets a limit on sin22θ13 these
questions can not be resolved by Noνa and T2K.