Thessaloniki, June 5, 2003

It takes glue to study the “Glue”

Zisis PapandreouDept. of Physics

University of Regina

Zisis PapandreouDept. of Physics

University of Regina

Canadians:Regina,Carleton

Refe

ren

ces

Design Report

Nov 2002

Sept/Oct 2000

Feb 2001

Sept 2000

JLab whitepaper

Cover story articleon exotics and Hall D

Articleon exotics and Hall D

All can bedownloaded fromthe Hall D website

Scientific Goals and Means• Definitive and detailed mapping of hybrid meson

spectrum• Quantitative understanding of confinement mechanism

in Quantum Chromodynamics (QCD).• Search for smoking gun signature of exotic JPC hybrid

mesons; these do not mix with qq states• Tools for the GlueX Project:

– Accelerator: 12 GeV electrons, 9 GeV linearly polarized photons with high flux

– Detector: hermiticity, resolution, charged and neutrals, R&D– PWA Analysis: each reaction has multi-particle final states – Computing power: 1 Pb/year data collection, data grids,

distributed computing, web services…

-

white

white

QCD is the theory of quarks and gluons

3 Colors3 Anti-colors

8 Gluons, each with a color and an anti-color charge.

)(21)(

GGtrmDiLQCD

(colorless objects)

confinement

qqq

qq

Jets at High EnergyDirect evidence for gluons come from high energy jets. But this doesn’t tell us anything about the “static” properties of glue. We learn something about s.

q

q

2-Jet

g

q

q

3-Jet

gluon bremsstrahlung gqqqq )()(

Deep Inelastic Scattering

As the nucleon is probed to smaller and smaller x, the gluons become more and more important. Much of the nucleon momentum and most of its spin is carried by gluons!

Glue is important to hadronic structure.

x = q2/2M, 0<x<1

QCD and confinement

Large DistanceLow Energy

Small DistanceHigh Energy

PerturbativeRegime

Non-PerturbativeRegime

High EnergyScattering

GluonJets

Observed

Spectroscopy

GluonicDegrees of Freedom

Missing

Strong QCD:See and systems.qq qqq

white

whiteNominally, glue isnot needed to describe hadrons.

Allowed systems: , , ggggg, gqq qqqqGlueballs Hybrids Molecules

Gluonic Excitations

u

d s

c

b

t

Six Flavors of quarks

GlueX Focus: “light-quark mesons”

Flux Tubes

Color Field: Gluons possess color charge: they couple to each other!

Dipole

Flux Tube

•1 electric charge carries no charge

•3 Color charges•g carry charge

Notion of flux tubes comes about from model-independentgeneral considerations: observed linear dependence of m2 on J.Idea originated with Nambu in the ‘70s.

Lattice QCD

Flux tubes realized

Flux

tube

forms

between

qq

linear potential

0.4 0.8 1.2 1.6

1.0

2.0

0.0

r/fm

Vo(

r)

[GeV

]

From G. Bali

Lattice “measurement”of quenched static potential

“Pluck” the Flux Tube

q

q

Normal meson:flux tube in ground state

q

q

Hybrid meson:flux tube in excited state

There are two degenerate first-excited transverse modes with J=1 – clockwise and counter-clockwise – and their linear combinations lead

to JPC = 1– + or JPC=1+ – for the excited flux-tube

How do we look for gluonic degrees of freedom in spectroscopy?

LS

S

1

2

S = S + S1 2

J = L + S

C = (-1)L + S

P = (-1)L + 1

Mass Difference

1 GeV/c2 mass difference

Hybrid mesons

Normal mesons

r - quark separation

m=0 CP=(-1) S+1

m=1 CP=(-1) S

Flux-tube Model

ground-state flux-tube m=0

excited flux-tube m=1

built on quark-model mesons

CP={(-1)L+S}{(-1)L+1} ={(-1)S+1}

S=0,L=0,m=1

J=1 CP=+

JPC=1++,1--

non exotic

S=1,L=0,m=1

J=1 CP=-JPC=0-+,0+-

1-+,1+-

2-+,2+-

exotic

normal mesons

Hybrid Mesons

1-+ or 1+-

,,

exotic

K

linear potential

ground-state flux-tube m=0

excited flux-tube m=1

Gluonic Excitations provide anexperimental measurement of the excited QCD potential.

Observations of exotic quantum number nonets are thebest experimental signal of gluonic excitations.

QCD Potential

Meson Map 1

Glueballs

Hybrids

Mas

s (G

eV)

1.0

1.5

2.0

2.5

qq Mesons

L = 0 1 2 3 4

S

I 3

K

– o

K o

Ko

K–

Pseudoscalar

S

I 3

K *o K *

K *oK *–

– o

Vector

JPC = 1--

JPC = 0-+

(L = qq angular momentum)

Each box correspondsto 4 nonets (2 for L=0)

(2S+1)LJ=3S1

(2S+1)LJ=1S0

Mas

s (G

eV)

1.0

1.5

2.0

2.5

qq Mesons

L = 0 1 2 3 4

Each box correspondsto 4 nonets (2 for L=0)

Radial excitations

(L = qq angular momentum)

exoticnonets

0 – +

0 + –

1 + +

1 + –

1– +

1 – –

2 – +

2 + –2 + +

0 – +

2 – +

0 + +

Glueballs

Hybrids

Meson Map

0++ 1.6 GeV

Lattice 1-+ 1.9 GeV

Production of Exotics

A pion or kaon beam, when scattering occurs,

can have its flux tube excitedor

beam

Quark spins anti-aligned

Much data in hand with some evidence for gluonic excitations(exotic hybrids are suppressed)

q

q

befo

req

qaft

er

q

q

aft

er

q

q

befo

re

beamAlmost no data in hand

in the mass regionwhere we expect to find exotic hybrids

when flux tube is excited

Quark spins aligned

__

__

Phenomenology

Model predictions for regular vs exotic meson prodution with pion (S=0) and photon (S=1) probes

Szczepaniak & Swat

Partial Wave Analysis (PWA)

A simple example - identifying states which decay into

Decay into implies J=L, P=(-1)L and C=(-1)L

Production and decay

point the way

N N

e

b

p n m p1 p2 2

t pb p1 p2 2

b e

C.M.S.

dNd cos

YJ0 2

Line shape and phase consistent with Breit-Wigner line shape

Low t

production

•Identify the JPC of a meson•Determine production amplitudes & mechanisms•Include polarization of beam, target, spin and parity of resonances and daughters, relative angular momentum.

f()= (1/k) (Cl/l)sinlPl(cos)

d = f()2d

coscos cos

Decay Angular Distributions

JPC=2++ JPC=3--

b e

C.M.S.

dNd cos

YJ0 2

JPC=1--

Place PWA tools on firm footing, years before experiment•effects of polarized and unpolarized beam•detector acceptance (hermetic)•Leakage of non-exotics into exotic signal

GlueX uses two parallel PWA codes for cross-checking!

500

400

300

200

100

0

1.81.61.41.2

PWA fit

Finding the Exotic Wave

PC The J exotic was in themix with other wavesat the level of 2.5%

The generated mass andwidth are compared withthose from PWA fits:

Mass

Input: 1600 MeV

Width

Input: 170 MeV

Output: 1598 +/- 3 MeV

Output: 173 +/- 11 MeV

500

400

300

200

100

0

1.81.61.41.2

Mass (3 pions) (GeV)

events/20 MeV generated

Statistics correspond to a few days of running

Double-blind MC exercise

p n

JPC=1-+1

What is Needed for the GlueX Experiment?

PWA requires that the entire event be kinematically identified - all particles detected, measured and identified. It is also important that there be sensitivity to a wide variety of decay channels to test theoretical predictions for decay modes.

The detector should be hermetic for neutral and charged particles, with excellent resolution and particle identification capability. The way to achieve this is with a solenoidal-based detector.

Hermetic Detector:

Linear Polarization is required by the PWA At least 12 GeV electrons Small spot size and superior emittance CW beam minimizes detector deadtime, permitting much

higher rates

Linearly Polarized, CW Photon Beam:

Why a 12 GeV Electron Beam?

The mass reach of GlueX is up to about 2.5 GeV/c2 so the photon energy must at least be 5.8 GeV. Energy must be higher than this so that:

1. Mesons have enough boost so decay products are detected and measured with sufficient accuracy.

2. Line shape distortion for higher mass mesons is minimized.

3. Meson and baryon resonance regions are kinematically distinguishable.

4. Linear polarization is optimal. But the photon energy should be low enough so that:

1. An all-solenoidal geometry (ideal for hermeticity) can still measure decay products with sufficient accuracy.

2. Background processes are minimized.

9 GeV photons ideal

flu

x

photon energy (GeV)

12 GeV electronsCoherent Bremsstrahlung

This technique provides requisite energy, flux and

polarization

collimated

Incoherent &coherent spectrum

tagged

with 0.1% resolution

40%polarization

in peak

electrons in

photons out

spectrometer

diamondcrystal

Linear Polarization

Linear polarization is:

Essential to isolate the production mechanism (M) if X is known

A JPC filter if M is known (via a kinematic cut)

Degree of polarization is directly related to required statistics

Linear polarization separates natural and unnatural parity

States of linear polarization are eigenstates of parity. States of circular polarization are not.

M

Experiment a2 yield Exotic Yield

SLAC 102 --BNL (published) 104 250

BNL (in hand - to be analyzed) 105 2500

GlueX 107 5x 106

More than104 increase

GlueX estimates are based on 1 year of low intensity running (107 photons/sec)

Even if the exotics were produced at the suppressed rates measured in -production, we would have 250,000 exotic mesons in 1 year, and be able to carry out a full program of hybrid meson spectroscopy

How GlueX Compares to Existing Data

Computational Challenge

• GlueX will collect data at 100 MB/sec or 1 Petabyte/year - comparable to LHC-type experiments.

• GlueX will be able to make use of much of the infrastructure developed for the LHC including the multi-tier computer architecture and the seamless virtual data architecture of the Grid.• To get the physics out of the data, GlueX relies entirely on an amplitude-based analysis - PWA – a challenge at the level necessary for GlueX. For example, visualization tools need to be designed and developed. Methods for fitting large data sets in parallel on processor farms need to be developed.

• Close collaboration with computer scientists has started and the collaboration is gaining experience with processor farms.

LRP

www.nscl.msu.edu/future/lrp2002.html

NSAC Long Range

Plan

Jefferson Lab FacilityNewport News, Virginia

A B C

Hall D will belocated here D

6 GeV CEBAF

11

CHL-2CHL-2

12Upgrade magnets Upgrade magnets

and power and power suppliessupplies

Two 0.6 GV linacs1.1

Beam Power: 1MWBeam Current 5 µA

Emittance: 10 nm-radEnergy Spread: 0.02%

GlueX Collaboration

US Experimental Groups

A. Dzierba (Spokesperson) - IUC. Meyer (Deputy Spokesperson) - CMUE. Smith (JLab Hall D Group Leader)

L. Dennis (FSU) R. Jones (U Conn)J. Kellie (Glasgow) A. Klein (ODU)G. Lolos (Regina) (chair) A. Szczepaniak (IU)

Collaboration Board

Carnegie Mellon University

Catholic University of America

Christopher Newport University

University of Connecticut

Florida International University

Florida State University

Indiana University

Jefferson Lab

Los Alamos National Lab

Norfolk State University

Old Dominion University

Ohio University

University of Pittsburgh

Renssalaer Polytechnic Institute

University of Glasgow

Institute for HEP - Protvino

Moscow State University

Budker Institute - Novosibirsk

University of Regina

CSSM & University of Adelaide

Carleton University

Carnegie Mellon University

Insitute of Nuclear Physics - Cracow

Hampton University

Indiana University

Los Alamos

North Carolina Central University

University of Pittsburgh

University of Tennessee/Oak Ridge

Other Experimental Groups

Theory Group

100 collaborators25 institutions

The impact of

• Chair of Collaboration board: George Lolos• Hardware working group co-leader: Zisis Papandreou• Software working group co-leader : Ed Brash• Most recent Hall D Collaboration meeting was held at UofR.• SPARRO has been tasked with the design and construction

of the full barrel calorimeter: weight of 30 tons, 5,000 km of fibers, and a price tag of US$4,000,000.

• Initial R&D is ongoing at Regina; JLab discretionary funds • Assembly and machining of full modules in Edmonton.

Subatomic Physics At Regina with Research Offshore

GlueX Detector

Lead GlassDetector

Superconducting Solenoid

Electron Beam from CEBAF

Coherent BremsstrahlungPhoton Beam

Tracking

Target

CerenkovCounter

Time ofFlight

BarrelCalorimeter

Note that tagger is80 m upstream of

detector

Barrel Calorimeter

Tracking Package

Rail System

Lead/scintillating fiber sandwich;1-mm diameter fibers; 0.5mm sheets

PMT Readout

PMT Readout

Scintillating fiber

Barrel Calorimeter Design Issues

• Length of device (> 4 m)• Shower containment/radiation lengths: ~16Xo• Azimuthal granularity vs. PMT geometry & light collection• Minimum energy deposition threshold: 20 MeV• Neutral energy resolution balance (vs. FCAL)

– Fiber diameter & Pb thickness: uniform throughout?

• Charged vs. Neutral energy resolution balance• Timing resolution vs. other subsystems• Mechanical issues: manufacturing, inner detector

mounting

R&D at Regina:•Monte Carlo•Hardware: Readout, fibers, Pb/SciFi Matrix

What’s a Hybrid?

• Proximity-focused vacuum tube• Fast rise- and fall-time (2, 12 ns)• High QE for UV, Gain is 105

• Insensitive to high magnetic fields

Fused silica window

HV: -8kV Bias: -80V Pre-amp Shaper

2 TeslaCentral Fieldmeans HPMTs

or20m fiber

light guides

Hybrid PMT

DEP PP0350G

DEP PP0100Z

UofR HV regulator

• R&D:– NaI detector– Scintillator

• Parameters measured

CREMAT CR-101D

• System works:– Pre-amp– R&D on circuit– Price issue

Optical Fiber Tests with 100MeV pions

• Properties:– Light collection

efficiency (cladding)– Scintillation light

production (doping)– Light attenuation

coefficient– Timing resolution

• Fibers Types (single-, double-clad)– Kuraray SCSF-81– Pol.Hi.Tech. 0046

– Bicron (too expensive) M11 Experimental Area at TRIUMFAugust 2001

Attenuation: ~3m (Kuraray)

Performance/Price: (Pol.Hi.Tech.)

Fiber Optical Properties• Quality Control of Fibers

– Quick, reliable, reproducible

• Commercial system was used• Components: SD2000 spectrometer,

ADC1000USB, laptop, fibers, neutral density filters, connectors, lenses, optical grease, light source.

• Kuraray, PHT fibers have been tested for transmission and attenuation.

Master Branch(Gold fibers)

Slave Branch(Test fibers)

Prototype Pb/SciFi Modules

Build in May 2002with help from

KLOE FolksP. CampanaF. Cervellie Technici

Prototype Module Construction

96 x 13 x 10 cm3

70 kg

B. Klein, J. Kushniryk, T.Summers, G.

WilliamsLead Sheet Swaging

Matrix gluing & pressing

Inspection and cleaning

1-m “Baby” Module

Finished Prototype

But…

Next: 2-m module just completed

Conclusions

• An outstanding and fundamental question is the nature of confinement of quarks and gluons in QCD.• Lattice QCD and phenomenology strongly indicate that the gluonic field between quarks forms flux-tubes and that these are responsible for confinement. • The excitation of the gluonic field leads to an entirely new spectrum of mesons and their properties are predicted by lattice QCD.• But data are needed to validate these predictions.• Only now are the tools in place to carry out the definitive experiment and JLab – with the energy upgrade – is unique for this search.• If exotic hybrids are there, we will find them! • In Regina, we will continue the R&D on MC, HPMTs and Pb/SciFi

We are eagerly awaiting CD-0 from the DOE.

The right people are in place - ready to make this happen

Backup Slidesbeyond this

point

Flux Tubes and Confinement

Color Field: Because of self interaction, confining flux tubes form between static color charges

mesons

Linear potential: infinite energyrequired to separate qq. -

Confinement arises from flux tubes !

Normal Mesons

LS

S

1

2

S = S + S1 2

J = L + S

C = (-1)L + S

P = (-1)L + 1

Spin/angular momentum configurations& radial excitations generate our knownspectrum of light quark mesons

Nonets characterized by given JPC

L=0, S=0: O-+ pseudoscalar mesonsL=0, S=1: 1-- vector mesonsL=1, S=1: O++ scalar, 1++ axial vector, 2++ tensor nonets

Hybrid Predictions

Lattice calculations --- 1-+ nonet is the lightest UKQCD (97) 1.87 0.20MILC (97) 1.97 0.30MILC (99) 2.11 0.10Lacock(99) 1.90 0.20Mei(02) 2.01 0.10

~2.0 GeV/c2

1-+

0+-

2+-

Splitting 0.20

Flux-tube model: 8 degenerate nonets 1++,1-- 0-+,0+-,1-+,1+-,2-+,2+- ~1.9 GeV/c2

S=0 S=1

N N

e

X

,,

1 IG(JPC)=1-(1-+)

’1 IG(JPC)=0+(1-+)

1 IG(JPC)=0+(1-+)

K1 IG(JPC)= ½ (1-)

1

1 b1 ,

’1

Couple to V.M + e

1-+ nonet

Exotics in Photoproduction

Hybrid Decays modes need to be measured

Current Evidence

Glueballs Hybrids

Overpopulation of thescalar nonet and LGT

predictions suggest thatthe f0(1500) is a glueball

See results fromCrystal Barrel

JPC = 1-+ states reported

1(1400) 1(1600)

See results fromBNL E852

Complication ismixing with conventional qq

statesNot withoutcontroversy

Have gluonic excitations already

been observed ?

f0(1500)

f2(1270)

f0(980)

f2(1565)+

700,000 000 Events

Crystal Barrel Results

f0(1500)

250,000 0 Events

Discovery of the f0(1500) f0(1500) ’, KK, 4f0(1370)

Crystal Barrel Results Establishes the scalar nonetSolidified the f0(1370)Discovery of the a0(1450)

p p 30

An Exotic Signal in E852

LeakageFrom

Non-exotic Wavedue to imperfectly

understood acceptance

ExoticSignal

1

Correlation ofPhase

&Intensity

M( ) GeV / c2 3 m=1593+-8+28

-47 =168+-20+150-12

’ m=1597+-10+45-10 =340+-40+-50

1(1600)

What Electron Beam Characteristics Are Required?

Coherent bremsstrahlung will be used to produce photons with linearpolarization so the electron energy must be high enough to allowfor a sufficiently high degree of polarization - which drops as the energy of the photons approaches the electron energy.

At least 12 GeV electrons

In order to reduce incoherent bremsstrahlung background collimation willbe employed using 20 µm thick diamond wafers as radiators. Small spot size and superior emittance

The detector must operate with minimum dead time

Duty factor approaching 1 (CW Beam)

Photon Beam Energy Figure of Merit

12

10

8

6

4

2

0

flux

of

phot

ons:

mill

ions

/sec

121110987

photon beam momentum (GeV)

endpoint energy = 12 GeV = 11 GeV = 10 GeV

Photon Flux inCoherent Peak

Total Hadronic RateConstant at

approx 20 KHz

0.6

0.5

0.4

0.3

0.2

0.1

0.0lin

ear

pola

riza

tion

121110987

photon beam momentum (GeV)

endpoint energy = 12 GeV = 11 GeV = 10 GeV

LinearPolarization

electron energy

photon energy

Electron energy

Photon energy

Given that 9 GeV photons are ideal - 12 GeV electron energy suffices; lower than this seriously degrades polarization and tagged hadronic rate.

BNLSLAC

18 GeV/c19 GeV

p n

p p

a2 a2

We will use for comparison – the yields for production of the well-established and understood a2 meson

Compare Pion and Photoproduction Data Rates

Photoproduction of 3π at SLAC and JLab

SLAC

similar cuts

p n

SLAC 19 GeVCLAS

5.7 GeV

a2 a2

10-2

1990 2000 2010

Lattice gauge theory invented

Quenched Hybrid Spectrum

Hybrid Mixing

Hybrids in Full QCD

Exotic candidate at BNL

First data from CEBAF @12 GeV

Tflop-year

100

10-1

10-4

10-6

1974

Lattice Meson spectrum agrees with Experiment.

101

102

10-3

10-5

Flux tubes between Heavy Quarks

FY03 Clusters 0.5 TFlop/sec

FY06 Clusters 8 Tflop/sec

LQCD

LRP

www.nscl.msu.edu/future/lrp2002.html

The Site and Beamline

Work at JLab on civilissues

At Glasgow and U Connecticut,progress on diamond wafers

Particle IdentificationTOF tests inRussia - Serpukhov

Magnetic shielding studies

Cerenkovcounter

Coils

Iron Yokes

FDC

Barrel Calorimeter

CDC

Vertex Detector

Target

GlueX Detector

Central Field:2 Tesla

Detector Platform

Top View of the HALL

PLATFORM

• The Platform is 8 ft from the base

• Supports the extracted components

• Assists in assembling the parts

All Dimensions in Meters

Solenoid Assembly

Hall D

Solenoid & Lead Glass Array

At SLACNow at JLab

MEGA magnet at LANL

Lead glass array

Magnet arrives in Bloomington

Tracking

R&D on central trackerat Carnegie Mellon

Start Counter

Computational Tools for Simulations

Energy resolution: Driving physics processes Detector subsystems

relation Timing resolution:

Detector subsystems relation

Longitudinal position requirement? Detector subsystems

relation

HDFast GEANT

Notes: use a2 decay or fake to look at particle distributions; use known smearing from FCAL and KLOE to do the first order resolution modeling. Is the missing mass resolution important?

Neutral and charged particle energy deposition profile

Readout granularity Lateral energy spread Verify back-of-the-envelope X0

Pb/SciFi/Glue detailed modeling Position resolution (vs. angle):

pos.res./N; shower algorithms

Notes: results of simulations have to be reconciled with mechanical constraints (HPMT outer and inner diameter; area reduction factor from Winston cones).

Tools are now in place at Regina

BCAL Readout Configuration

Longitudinal View

K. Wolbaum

Even with SciFi bundling,~540 HPMTs will be needed:

No of phi segments 54

Phi segment angle 0.116

HPD active area 5.07 cm2

HPD total area 20.3 cm2

Calorimeter thickness 20.27 cm

Outer calorimeter radius 89.71 cm

Calorimeter length 400 cm

No of HPMT’s 1080

o Length: 4.5 mo Outer Radius: 90 cmo Thickness: 20-25 cm

$$ cost issue $$

• Light sources: – NaI crystal and 241Am,

137Cs and 60Co sources– LED (pulsed): neutral

density filter, fiber lightguide

• Measurements: • HPD PP0350G with Multi-

Channel Analyzer (Tektronix 5200) and DAQ System

• Bottom line: system works, needs a bit more R&D

HPMT Innards

CREMAT CR-101D

Attenuation Length

~3m

”Cave” Optical System Setup

• Optical system

•Spectrometer•ADC•Laptop

Master (Gold)

Slave (tests)

Master Branch(Gold fibers)

Slave Branch(Test fibers)

L. Snook

Coupling Reproducibility

Kuraray Multi-Clad 2002 Batch4000

Wavelength (nm)300 1100

Inte

nsi

ty

(from Aug. 8-12, 2002)

4000

Inte

nsit

y

Wavelength (nm)1100

1m (#2)1m (#1)2m3m4m5m10m

Kuraray Multi-Clad 2002 Batch

Attenuation Determination

Lead Roll