NO n A at Caltech

NOA at Caltech

Leon Mualem

DOE Review

July 25, 2007

The NOA Detector~80 m

15.7 m

~16 kT total mass “Totally Active” granular

design Outstanding e pattern

recognition & measurement

Alternating X and Y views12 Extruded PVC Modules

per plane32 Individual cells per Module,

so 384 Cells per plane

Working to fit 260M AY$ TPC cap

NOA Tasks

Caltech Initiated or Responsible for many aspects of NOA

Hardware DAQ/Electronics Management APD Testing PVC Testing Fiber Testing Vertical Slice Tests

Software Framework Development Subshower Package Photon Transport simulation Supernova Sensitivity

Overview of Detector R&D

NOA Perform light output tests to understand the

components of the scintillator system [Ongoing] PVC extrusions, liquid scintillator, WLS fiber

Verification of scintillator system performance using a NOA APD [Ongoing]

Photon production and transport Monte Carlo [Ongoing]

Personnel – Jason Trevor, Leon Mualem + undergraduate

NOA Scintillator System Each cell an extruded TiO2 loaded PVC tube

with ID 60mm x 39mm x 15.7m long Cells are filled with mineral oil scintillator

which is read out at one end with a U-loop WLS fiber running to a multi-pixel APD

Kuraray 0.7 mm WLS Fiber Light output requirement determined by

achievable noise on the APD amplifier. The current estimate of minimum required Light Output is ~20-25 photoelectrons

One Cell

0.7mm WLS Fiber

R&D at Caltech Composition of the PVC cell walls Liquid scintillator composition Fiber diameter and dye concentration Fiber position Integration testing

Interface and Readout

Electronics Box

32 Pixel APD Photodetector Array

ManufacturerPixel Active Area 1.95 mm × 1.0 mmPixel Pitch 2.65 mmArray Size 32 pixelsDie Size 15.34mm × 13.64mmQuantum Efficiency (>525 nm) 85%Pixel Capacitance 10 pFBulk Dark Current (IB) at 25 C 12.5 pABulk Dark Current (IB) at -15 C 0.25 pAPeak Sensitivity 600 nmOperating Voltage 375 ± 50 voltsGain at Operating Voltage 100Operating Temperature (with Thermo-Electric Cooler)

-15ºC

Expected Signal-to-Noise Ratio (Muon at Far End of Cell)

10:1

APD channels per plane 384APD arrays per plane 12

APD Photodetector

Si Avalanche Photodiode

Custom design to match two-fiber aspect ratio

Bare die mounted to PCB via gold bump thermo-compression

Scintillator System R&D

Liquid Scintillator

Two competing mineral oil vendors (Parol, Ren) Competing vendors for additives Concentration of additives (pseudocumene,

PPO, Bis-MSB)

PVC Extrusions Three competing PVC formulations

(Prime, Aurora, NOvA collaboration) Two types of TiO2 (Anatase, Rutile)

Competing factors (extrudability, reflectivity, structural issues)

Test Setup

PMT

M16 PMT Box

“ Cell”

Lead

Scintillating Disc

1.2mm Clear Fiber

We are currently using a “NOA Cell” with internal dimensions 38.5mm x 60mm x 85cm.

Each end of the fiber “U” is connected to an individual pixel on an M16 phototube by 3.5m of 1.2mm clear fiber.

Fibers are held in a fixed position inside the cell by a pair of acrylic “spiders.” (More on this later)

For testing purposes I am using vertical muons from cosmic rays.

The cosmic ray muon telescope consists of two circular discs (~4 cm diam) separated by 14 cm and 1 inch of lead.

The “NOA Cell”

PVC

60mm

“NOA Cell”

38.5mm

The “NOA Cell” is actually a rectangular aluminum tube which is lined on the inside with the PVC we are testing.

More “NOvA Cell” Pictures

DataData from each fiber end is

collected and plotted separately

Because the rate is low (one event every 150 seconds), data is acquired over a long period of time (>72 hours) in order to obtain a statistically significant sample

Results: Scintillator Samples

Three scintillator samples, RenDix 517p, ParDix 517p, and RenAld517p.

Results show a 20% difference in light output between scintillators made with oil from Ren Oil and those made with oil from Parol.

Pseudocumene from different suppliers appear to perform similarly.

Measurements are highly repeatable.

47.546.1

57.2 57.5

0

10

20

30

40

50

60

70

ParDix517P1

ParDix517P2

RenDix 517P RenAld 517P

Summed Light Output(PE)

Results: Extrusions We have tested several extrusion

samples. Shown here are our baseline extrusion, our best extrusion made with rutile Ti02, and our best Anatase extrusion.

We have also included two other samples for reference: A MINOS Strip The duplicate “NOA Cell”

painted on the inside with BC-620, an acrylic based paint loaded with TiO2(Anatase).

All measurements were performed with RenDix 517P

Initial results show we can do better that the minimum light output specification.

8.2

44

51

57.2

62.3

0

10

20

30

40

50

60

70

Minos Strip Pet B(Baseline)

BB15R AnatasePaint

BB18A

Summed Light Output (PE)

Minimum Spec.

Upgraded Test SetupPMT

M16 PMT Box

Actual Cell

Lead

Scintillating strip

1.2mm Clear Fiber

Increased trigger sizes.More than triple the rate, no effect on precision.

Testing apparatus is otherwise unchanged

Increased throughput of system; limited by sample preparation time,instead of trigger rate.

Extrusion tests

More Extrusion Results Tests of recent

extrusions show high and consistent light output compared to previous recipes.

Recent extrusions have also extruded well mechanically.This is CRITICAL to integrity of the detector:

— the PVC is the structure 46

48

50

52

54

56

58

60Series BSeries CSeries D

Caltech Mini Mini Tracker

Minos Scintillator

Trigger

Minos Scintillator

Trigger

FRONTVIEW

SIDEVIEW

16.4 cm 30 cm

60cm

40c

m

150ppm1 2 3 4

300ppm

250ppm N N

1” Lead

“N” = Near length fibers

Actual Device

Not as photogenic, but: Uses prototype APD 33.4m fiber

(Actual length) It works

300 ppm results

Far Light Output vs. Concentration

Measurement Summary

150 ppm 24 pe’s Significant spread, 20-35

250 ppm 22 pe’s Very small spread (only 2 samples)

300 ppm 35 pe’s Range: 30-40 3 with ~10% spread

NOA Software at CaltechWe developed a set of light weight libraries

(“SoCal”) to allow people to access NOA data and information in C++/ROOT.

SoCal consists of:Data format for NOνANOvA geometry and electronics connection mapEvent display packageDetector & Electronics response simulation toolsFull (and up to date) documentation.Tools to help people write further packages.

Used by the collaboration to develop reconstruction and analysis used for TDR

SoCal

Caius Howcroft

Caius Howcroft

Decay chain

BrandingEvent Source

Reco'ed event

True Hits e/μπp

Colour = energy deposited

Event Display

Caltech Reconstruction:“Subshower” Code [HZ, CH]

3D shower & Track-like feature reco.

Now the standard in MINOS

Has been applied to NOA

Preliminary use by Bob Bernstein at Fermilab shows significant signal/background separation

Needs to be carried through to a complete analysis [e.g. Patterson]

“Raw Data”

Sub Showers

Caius Howcroft

Detector Simulation Detector simulation code that models the light output

of the scintillator, the collection of WLS fiber and the propagation to the APD, “PhotonTransporter”

Tracks individual photons and correctly deals with wavelength dependent absorption, reflection and emission coefficients.

Has been used to understand results from the Caltech test-stand and in production MC.

Accurately reproduces features of measured light collection in a cell

Caius Howcroft

Charged Particle

Simulated Cell

WLS FibersPhoton

Simulations of Light Output vs. Position

Fiber Position Results Light Yield Simulation

suggests light output decreases as fibers approach walls

Effect seen in test stand data, but magnitude smaller than predicted

Tune simulations with data to reproduce changes quantitatively

Background Studies

NOA is a search for a small signal

Understanding and correctly modeling the background is important

Work at Minnesota demonstrated the need for an overburden

This work also showed potential for Supernova detection with the overburden

Additional effort needed to determine sensitivity, and computing requirements to search for small signal events

Find the SuperNOvA

31Leon MualemNOvA Electronics

~15min of data With typical ~10s supernova signal

1s time binsNO OVERBURDEN

Find the SuperNOvA

~15min of data With typical ~10s supernova signal100ms time bins

1m OVERBURDEN

Software / AnalysisCreated the Framework used for NOvA software

development and TDR analysisThis base now being expanded to add features

Created the Subshower analysis package for MINOS, ported to NOvA frameworkShowing promising results by Bernstein@FNALNeeds to be carried through to a complete

analysisCreated Photon propagation code

Generally useful for understanding light collection and detector performance

Validation with actual test data continues

Caltech Work on NOA WLS Fiber R&D

Examined the effect of fiber diameter on light collection in the NOA geometry (The reduction to 0.7mm WLS Fiber saved $3 million+)

Measured light output for fibers with varying fluor concentration, ongoing optimization

Examine the effect of fiber position inside the NOA cell on light collection, ongoing input to simulation

Optical Readout System Verify scintillator system performance using the NOA

APD, original and prototype versions Quantify results of production variability using different

plastic samples, many cells and fibers Gain experience using the new NOA APD in small scale

prototype modules

SummaryNOA

Caltech has taken a leading role in NOA detectorhardware R&D Measurements done thus far at Caltech have

been, and continue to be instrumental in the detector design process

We will continue to make contributions central to the detector development effort throughout the next year

We will continue to define the detector performance and identify unique capabilities, such as supernova detection

The arrival of Ryan Patterson will add considerable strength, allowing us to build on our founding roles in NOA software and analysis