NOºA
The NuMI Offaxis ºe Appearance Experiment
Andrew Norman for the NOºA Collaboration
NuFACT07, Okayama Japan
August 11, 2007
l
a
 Over 140 Physicists and Engineers from 28 institutions
b
o
r
a
t
i
o
n
Introduction
 Physics Motivation
 Sensitivities
 Detector Design
 R&D Progress
Ash River
International
Falls
Duluth
 Status
Minneapolis
Fermilab
Overview: NOvA Today (Baseline)
 NOvA is an 18kTon far detector +





218Ton near detector + NuMI beam
upgrade project.
Both detectors are “totally active” liquid International
Falls
scintillator designs
The detectors are 14mrad off the
primary beam axis to achieve narrow º¹
energy spectrum, peaked at 2GeV.
The far detect sits on a 810km baseline
between Chicago and Northern
Minnesota at the first oscillation maximum
Designed to use the 320KW then 700KW
NuMI beam with final upgrade to the
1.2MW “super-NuMI” beam from the
Fermilab main injector.
Integrate ultimately 10£1020 pot/yr
Ash River
Duluth
Minneapolis
Fermilab
MoTivation & sensitivities
Part I
Physics Program
 The NOºA experimental program goals:
 Observe º¹ ! ºe oscillations
 Measure µ13
 Or improve the current limit on µ13 by a factor of 20.
 Measure sin2(2µ23) to a precision of 0.5-1%.
 Resolve the neutrino mass hierarchy
 Measure the CP violating phase ±
 Measurement of NC cross section at 2 GeV
 Detection of near galactic supernova
The Off-Axis Beam
7
The Effect of Going Off-Axis
 By going off-axis, the neutrino
flux from ¼ ! ¹ + º is reduced at
a distance z to:
2 A
F = ( 1+ 2°
)
2
2
° µ
4¼z 2
 But the energy narrows as µ2:
Eº =
0:43E ¼
1+ ° 2 µ2
 For NOνA, moving 14 mrad off
axis makes the NuMI beam
energy
 peak at 2 GeV
 Eº width narrows to 20%
 This corresponds to the first the
oscillation maximum
1st Osc. Max
8
The Effect of Going Off-Axis
• This suppresses the high
energy tail (NC background)
• Significantly reduces the
Kaon background
contribution by shifting the
neutrino energy away from
the signal band
• Energy spectrum in the signal
region becomes almost
insensitive to the /K ratio
• Results in a neutrino peak
primarily from  decays
20
On Axis (π)
18
º’s from K’s well
above signal band
On Axis (K)
16
π 14mrad
14
K 14mrad
12
Eν(GeV)
Eº K =
0:96E K
1+ ° 2 µ2
Offaxis νμ Energies from π/K
10
8
6
º’s from ¼’s in 2GeV band
4
2
0
0
9
5
10
Eπ/K GeV
15
20
P(º¹ ! ºe) and Ue3
The Ue3 contribution to the third mass
state is small, requiring a precision
measurement of ºe appearance
 Measuring a ºe excess in the NuMI º¹
beam will give evidence for º¹! ºe
transitions and a non-zero Ue3 component
to ¢ m232
 This is done through the ºe CC channel
º3
 The º¹ NC is the dominant background,
 Controlled through the identification of
initial vertex and displaced shower
conversion point.
 NOºA’s energy (2GeV) and baseline
º2
(810km) and segmentation (0.15X0) are
chosen to maximize the physics reach of
accessing these transitions
º1
 Electron neutrino’s role in the º¹ flavor
transitions is given an upper bound by
CHOOZ limit at 5-10% of the total state.
¢ m2at m
ºe Charged Current Channel
Event Parameters
Reaction:
º ep ! p¼+ e¡
Primary Vertex
Eº = 2.5GeV
Ep = 1.1GeV
E¼ = 0.2GeV
Ee = 1.9GeV
Shower spans »65
of the 1178
planes
Localized E&M Shower
º¹ Neutral Current Background
Event Parameters
Reaction:
º eN ! p¼0 º ¹
Primary Vertex
Eº = 10.6GeV
Ep = 1.04GeV
E¼ = 1.97GeV
Suppressed by
vertex/shower
displacement
identification
Localized E&M Shower
Displacement
Sensitivity at 3¾ for µ13 from º¹ ! ºe
 For the current 18kTon
detector, with 700kW
(dashed) and 1.2MW
(solid) beam.
 The physics reach for
µ13 is shown for 3
years of running each
on º and º-bar.
 The reach of the
1.2MW is almost an
order of magnitude
beyond the CHOOZ
bound, giving
improvement over
the 3¾ MINOS reach
Sensitivity to sin2(2µ23)
 NOºA can perform the




disappearance measurement to
a precision of 0.5-1%
Proceeds as a parameterized
analysis of quasielastic º¹ CC
events
If 1, then resolve quadrant
(µ23 > ¼/4 or µ23 < ¼/4, )
Measure if º3 couples more to
º¹ or º¿
Resolve ambiguity by
comparing NOºA to Daya Bay.
Mass Ordering
 From solar and atmospheric data we
know:
¢ m223
¢ m212
• This leads to two possible mass
hierarchies
• A “natural” order which follows the
lepton mass ordering
• An “inverted” order where m3 is
actually the lightest
• NOºA can solve this by measuring the
sign of m23 using the MSW effect over
the 810km baseline
¢ m2sol ar
¢ m2at m
Matter Effect
 The forward scattering amplitudes for neutrinos and anti-
neutrinos through normal matter differ due to the inclusion of
the extra diagram for interactions off electrons
 This difference breaks the degeneracy in the neutrino mass
spectrum and modify the oscillation probability
e
ºe
W
Pm at (º ¹ ! º e) 6
= Pm at (¹º ¹ ! º¹ e)
 If the experiment is performed at the first peak in the
oscillation then the matter effects are primarily a function of the
beam energy and approximated by:
Pm at (º ¹ ! º e) ¼ (1 +
ER =
¢ m 223
p
2 2G F N e
E
ER
ºe
)Pvac (º ¹ ! º e)
¼ 11GeV
• In the normal hierarchy this matter effect enhances the transition probability for
neutrinos and suppresses the probability for antineutrinos transitions
• With an inverted hierarchy the effect is reversed
• For the 2 GeV neutrino beam used for NOºA, the matter effect gives a 30%
enhancement/suppression in the transition probability.
e
Sensitivities for P(º¹ ! ºe) = 0.02
 Some CP phases
create an ambiguity
in the resolution of the
mass hierarchy.
 Combine with a
second measurement
to break ambiguities
CP Violation
 Large Mixing Angle (LMA) solution gives sensitivity in º¹ ! ºe
transitions to the CP violating phase ±.
 In vacuum, the transition probability is shifted with ±. At the first
oscillation maximum the shift is:
q
j¢ P± (º ¹ ! º e)j » 0:06%
 Since the shift is proportional to
p
si n 2 2µ1 3
0:05
si n2 2µ13 the importance of the
2
sub-leading terms grow, as si n 2µ13 gets small.
 The ultimate sensitivity of
NOνA for resolving the CP
ambiguities in matter depend
on both sin θ13 and ±
 Combining the NOºA result
with other experiments lifts
the ambiguities in some
regions.
Resolution of Mass Hierarchy
(95% CL)
Normal
Inverted
Supernova Neutrinos
 Neutrinos and Antineutrinos are produced via:
N N ! N N º º¹ ;
e+ e¡ ! º º¹ ; ::::
 The neutrinos are trapped in core collapse,
reach thermal equilibrium and then escape in
a burst
 Duration of the neutrino burst: 1-10s
 The neutrino luminosity is upwards of 100
times greater than the optical luminosity
 Neutrino flash proceeds primary photons by
5-24 hours.
 Each flavor takes away the same energy
fraction
 Different neutrino temperatures are due to
allowed reaction channels
SuperNOºA Detection
 Primary SuperNOºA Signal:
º¹ e + p ! e+ + n
 For a supernova at 10kpc the total signal is
expected to contain:
 9664 total interactions over a time span of ¼
10s
 4800 interactions in the first second
 Energy peaks at 20MeV and falls off to »
60MeV
Signal: » 9000 event excess
Sig. to Noise: 1:3 (first second)
Expected Energy Spectrum (10Kpc SN)
 Challenge is triggering in real time
 Need data driven open triggering
 Long event buffering (» 30sec)
 Time window correlation & merging
Signal+Bkg
Cosmic Bkg
(25kHz)
 NOºA – farm 192 trigger/buffer PCs
(min 30s total event buffering)
Signal
The NOvA detector
Part II
The NOνA Detectors





Far Detector
18 kTons
1178
alternating X-Y
planes
Grouped into
38 modular half
kiloton “blocks”
Over 450,000
independent
detection cells
> 70% of total
mass is active
18,000 tons
218 tons
88 tons
45 tons
4.2 m
Far
Detector
4.2 m
Near
Detector
How big is this?
The NOvA far detector is big
enough to fit Atlas, CMS, DØ, and
CDF Inside it’s active volume.
IPND
Humpback
Whale
IPND
The “Integration Prototype Near
Detector” will be built in Q1 2008 to
show technological integration of all
the NOvA subsystem
NuMI Beam Options
 Two intensity options
 700 kW (ANU baseline)
 1.2MW (super-NuMI)
 1 year of beam = 44 weeks of running
 Duty factor = 0.6 (accel + NuMI down time)
 Run 3 years each on º and º-bar
 For 700kW this is 36£1020 pot
 For 1.2MW this is 60£ 1020 pot
NuMI Accelerator Upgrade
New extraction line from
Recycler to MI
•
•
•
•
Beamline Upgrade
Proton source upgraded from
320kW to 700kW
NuMI will deliver 4.9×1012
protons per pulse
1.33s rep-rate.
This results in 6×1020 pot/yr.
New Injection line from
MI to Recycler
&
Recycler
New RF stations added
and acceleration rate
switched to 240GeV/s
Changes
•Recycler runs proton not anti-protons
•New injection/extraction lines for Recycler to Main
Injector transfers
•Main Injector cycle time reduced from 2.2s to 1.5s
(stack in the recycler)
•Cycle time reduced again to 1.33s with 2 more RF
stations at MI-60 and with transition of the MI from 204
GeV/s to it’s design acceleration rate of 240 GeV/s.
•NuMI target redesign for high flux
Detector Modules
NOνA Modules
 The NOνA detector module forms
the base unit for the detector.
 Each module is made from two 16
cell high reflectivity PVC extrusions
bonded into a single 32 cell module
 Includes readout manifold for fiber
routing and APD housing
Two 16 cell extrusions
 Combined 12 module wide X or Y
measuring planes.
 Each module is capped, and filled
with the liquid scintillator.
 These are the primary containment
vessel for the 3.9 million gallons of
scintillator material.
 There are 14,136 detector modules
with a total of 452,352 separate
detection cells in the NOνA Far
Detector.
PVC Extrusions
Single Cell
15.7m
One Module
1.3m
6cm
3.9cm
Each extrusion is a single 15.7m (51.5ft) long set of 6x3.9cm cells.
Two extrusions are joined to form a single 1.3m wide module
Detector Cell ̶ Fiber and Scintillator

To APD Readout


Scintillation Light
15.5m
Particle Trajectory
Waveshifting
Fiber Loop
3.9cm 6.6cm
There are 452,352
cells in NOνA


NOºA Detection Cell
The base detector unit 15.5mx3.9x6.6cm cell filled with a
mineral oil based liquid scintillator.
Emission spectrum tune to match
reflectivity of PVC walls.
0.7mm wave shifting fiber loop
captures the light
Both fiber ends are read out by a
single pixel of the APD.
Light Collection
The scintillator/fiber detection cell is
measured to delivers 30-39 p.e. for a
muon traversing the far end of the cell
 NOºA uses 18,000km of fiber
Fiber
The fiber is a double clad
0.7mm fiber doped with
300ppm of a K27
fluorescent dye. The fiber
exhibits a long
attenuation length for the
peak emission spectrum
with tests at 300ppm of
dye concentration over
15m for the 550nm peak
of the spectrum.
0.6 mm diameter
25.00
Attenuation length (m)
Single Cell
20.00
150 ppm
300 ppm
15.00
10.00
5.00
0.00
500
550
600
Wavelength (nm)
650
Avalanche Photodiodes ̶ APDs





Require 452,352 optical readout channels
Custom designed 32 channel APD (Hamamatsu)
High 85% Q.E. above 525nm
Cooled to -15° to achieve noise rate < 3 p.e.
Thermal load » 2W
Pixel Geometry
Each pixel is designed
to have an active
area 1.0mm by
1.9mm to
accommodate the
aspect ratio of the
two 0.7mm fiber ends
from a single cell
0.15mm
0.25mm
1.9mm
The NOνA Readout
1.0mm
 Operated at gain of 100 for detection of 30-39
photon signal from far end readout
 Signal to noise at far end 10:1
The bare APD is bonded to the carrier board to
provide the optical mask and electrical interface
The 32 pixel APD array for the NOνA Readout
System
FEB
32 Module
Fiber Feed (64 ends)
APD Module
“Clam Shell”
Optical Housing
APD
Module
Carrier Board
w/ APD chip
64 fiber
bundle
Heat Sink
Interface
Front End Electronics & DAQ
 Use a continuous digitization and readout scheme
 APDs are sampled at a 2MHz and a dual correlated




sampling procedure is used for signal
recognition/zero suppresion
Done real time on the FPGA, the signals are then
dispatched to Collector nodes as “time slices”
Data Concentrator Modules assemble/order the data
and dispatch macro time windows to a “buffer farm”
of 192 compute nodes
Provides minimum 30sec full data buffer for trigger
decision
Dead-timeless system with software based
micro/macro event triggering
ASIC ADC FPGA
I/O
Power
Data Concentrators (DCM)
The digitized data streams from 64
front end boards are broadcast over
8B/10B serial links to an associated
data concentrator module which orders,
filters and buffers the data stream,
then repackages the data into an
efficient network packet and
rebroadcasts it to a specific buffer
node for trigger decisions.
FEB ASIC
a low noise device
The custom
with expected noise
designed ASIC is an < 200 e. rms.
integrator/shaper
with multiplexer
running at 16MHz.
The channels are
Muxed at 8:1 and
sent to a 40MHz
quad ADC for
digitization. For the
higher rate near
detector the
channels are muxed
at 2:1 and sent to 4
quad ADCs. ASIC is
Time Sync
 Require all FEB local counters and DCM master time stamps
be synced to § 62ns system wide (micro time window)
 Used for event building and hit ordering
 Additional absolute sync to FNAL beam signal
GPS Antenna
(Outside)
 Use a continous self
calibrating delay
loop with GPS
master time
stamp/clock
 Compensates for all
DCM and FEB path
differences
 Sync/RST to FEBs
NOvA Global Timing Distribution
Distribution
Unit 1
Distribution
Unit 2
FPGA W/
PLL
Based
Serdes
FPGA w/
PLL
Based
Serdes
Distribution
Unit 9
Distribution
Unit 10
Incident Timing Packet
Master Timer
GPS Clock Generator
163.84Mhz
VCXO
FPGA w/
PLL
Based
Serdes
FPGA w/
PLL
Based
Serdes
Loopback
Return Timing Packet
NOvA Local Timing to DCMs
Gobal
Upstream
`
Gobal
Downstream
FPGA w/ 4 internal Serdes
Serdes
163.84-Mb
DCM1
Serdes
163.84-Mb
DCM2
Serdes
163.84-Mb
DCM23
Serdes
163.84-Mb
DCM24
Project Status
Part III
NOºA Approval Status
 CD-1 completed and signed.
 Total project capped at $260M
 Includes Accel. Upgrade to 700kW
 FNAL Director’s review (Part 1) June ’07
 Baseline detector size 18kton
 Physics & detector sub systems reviewed -- PASSED
 Costs rolled up
 Initial estimate placed detector over project cap
 Included many obvious accounting errors
 July/Aug costs for project “scrubbed” to remove anomalous
entries, redundancies and errors…
 FNAL Director’s review (Part 2) Aug. ‘07
 Primarily cost accounting review
 DoE CD-2/CD-3a review scheduled for Oct. ’07
NOºA Time Line
 Now Electronics, PVC, Scint, DAQ, DCS, SIM R&D





ongoing and continues through 2008/9
Q1/Q2 2008 IPND installation in MINOS assem.
Building
Q1/Q2 2008 Far site infrastructure start
2010 Collider shutdown, NuMI upgrades
2011 Beneficial Occupancy at Far Site
2011-13 Installation/Commisioning/Production data
taking occur in parallel
Backup
Additional Details and References
The Far Site - Ash River, MN
Ash River is chosen as the site for the massive 18kTon far detector because it is the farthest site
from Fermilab that is still inside the United States and yet accessible by roadway. The site is
810.5km from Fermilab, 1.5 miles south of Voyageurs National Park, and 45minute away from
the town of International Falls, known as “The Icebox of the Nation” for it’s record breaking
winter time temperatures.
N
0
25
50 km
The NOνA Site Today
Voyageurs
National Park
Ash River
Off Axis?
By placing the detector at Ash
River, 14.6mrad off of the NuMI
beam axis, we obtain a sharp
peak at 2GeV in the neutrino
energy spectrum.
NuMI
at NOνA
gives a
2GeV peak
The Far Detector Hall
The NOνA far detector at Ash River is a surface
detector positioned to catch the neutrino beam from
FNAL as it exits the earth’s crust. The detector hall
has a 63ft wide by 67ft high cross section to the
neutrino beam direction capable of housing and
shielding the 18kTon totally active detector in all
directions. The hall extends back in the z direction to
295ft to accommodate the 38 full detector blocks
And has additional room for
module construction
Containment
The detector is contained in a
40ft deep excavation
surrounding by solid granite to
provide secondary containment
for the 12.6kTons of liquid
scintillator
40ft
NOνA
Detector
Overburden
The overburden for the detector is enhanced by 6
inches of barite (X0=3.8cm) to provide additional
stopping power for photons which are one of
primary sources of cosmic-ray background. The
expected background rate is less than 0.1 event
over a six year run period.
OverBurden Material
Barite
Cast Concrete
Concrete Planks
Excavated granite with voids
Detector Access
The detector is 7 levels high
with electronics racks on the
top catwalks. These racks
provide the power and
readout infrastructure for
the readout of over
450,000 detector cells
Sensitivities: ¢ m2
at other P(º¹ ! ºe) values
0.05
No ambiguity
0.02
0.005
0.01
Significant
sensitivity only
for ¢ m2 < 0
Supernova Backgrounds
 Backgrounds (triggering):
 97 kHz through-going muons
 24 kHz stopping muons
 Real background (analysis) dominated by:
 Cosmic ¹ (100Hz, fiducial cut)
 Michel electrons (25Hz, tagged)
 ¹ capture (12B decay, small >20MeV)
NuMI Long Baselines
Noa
18-25 kTons
810km
12km off-axis
Surface Detector
Minos
5.5 kTons
735km
On beam axis
Underground
(½ mile)
39
 Mixing and µ13
 The MNS Mixing matrix can be broken into terms which explicitly
depend on the  ! , e! , e!  mixing angles, as well as
the CP violating phases  and i
 The sensitivity of NOºA is explicit in the third term giving access to
both sin2 2µ13 and ± through the ºe appearance measurement.
 There is additional NOºA sensitivity to µ23 through the º¹ osc.
measurements
Solar and Reactor
(KamLAND)
40
Atmospheric
Neutrinos
(Super-K)
NOºA: CP-Violating
Phase and µ13
The IPND, 2008 & Beyond
The next milestone for NOνA is the construction of the Integration Prototype Near Detector (IPND) which will bring
together all the discrete subsystems into a single unified detector with operational electronics and detection cells.
The overall footprint of the IPND will be 3m wide by 4m high and extend 8.5m in length. The IPND will be constructed of
4 blocks of prototype extrusions, with 31 planes each of alternating vertical and horizontal measuring modules. Two of
the blocks will be constructed in the “A” configuration of 16 vert. and 15 horiz. planes, while the other two blocks will be
in a “B” configuration of 15 vert. and 16 horiz. Planes. The detector will be supported on the front and back by two
“bookends”, and placed within a secondary containment enclosure.
The IPND will be constructed in early 2008 at Argonne National Lab and assembled in the MINOS Service Building. The
IPND will be an operational detector and will allow for the final phases of R&D and integration testing to be completed
during 2008 and early 2009.
IPND Parameters
Blocks:
PVC Modules per Block:
X-Y Detector Planes:
Detector Cells:
Front End Boards:
Data Concentrators:
Detector Mass:
4
2
124
7,936
248
5
88tons
"B" Block
"A" Block
Bookend