12-Jan-02 W.A. Zajc
1
PHENIXPHENIX for Beginners for Beginners
W.A. ZajcColumbia University
12-Jan-02 W.A. Zajc
2
The PHENIX The PHENIX CollaborationCollaboration
A strongly international venture: 11 nations
Brazil, China, France, Germany, India, Israel, Japan, South Korea, Russia, Sweden, United States
51 institutions
12-Jan-02 W.A. Zajc
3
Outline and GoalsOutline and Goals
What is the QGP phase transition?
What is the best strategy for observing it?
How has PHENIX implemented that strategy? An introduction to PHENIX physics results A (very) brief introduction to PHENIX detector
technologies Goals:
Provide overview of the many observables already measured by PHENIX in RHIC Run-1
Understand the trade-offs made by experimenters in optimizing a detector
Understand vast potential of PHENIX in future RHIC runs
12-Jan-02 W.A. Zajc
4
Need better control of dimensional analysis:
Relevant Thermal PhysicsRelevant Thermal Physics
Q. How to liberate quarks and gluons from ~1 fm confinement scale?
A. Create an energy density
??densitynuclear Normal ~fm GeV /2.0~)fm1/(~ 34
42
30Tg
Energy density for “g” massless d.o.f
42
303222
8
782 Tcfasg
8 gluons, 2 spins;
2 quark flavors, anti-quarks, 2 spins, 3 colors
3
4
4 fm GeV /4.21
1212
fmT
“Reasonable” estimate
42
3037 T 37 (!)
12-Jan-02 W.A. Zajc
5
37 B,- 90
42
gTgPQGP Pressure in plasma phase with
“Bag constant” B ~ 0.2 GeV / fm3
42
903 TP
Pressure of “pure” pion gas at temperature T
Rough EstimateRough Estimate Compare
Phase transition at T ~ 140 MeV with latent heat ~0.8 GeV / fm3
-0.25
0
0.25
0.5
0 100 200
Temperature (MeV)
Pressure
(GeV / fm3)
Pion Phase
QGP Phase PQGP
P
Select system with higher pressure:
Compare to best estimates (Karsch, QM01) from lattice calculations:
T ~ 150-170 MeV latent heat ~ 0.70.3 GeV / fm3
12-Jan-02 W.A. Zajc
6
Hadron 'level' diagram
0
500
1000
1500
0 10 20 30 40
Degeneracy
Mass (MeV)
Kfo
Aside 1Aside 1 Previous approach
(using a “pure” pion gas) works because QCD is a theory with a mass gap
{ This gap is a
manifestation of the approximate SU(2)R x SU(2)L chiral symmetry of QCD with pions as the Nambu-Goldstone bosons
12-Jan-02 W.A. Zajc
7
Aside 2Aside 2The frightening density of hadronic levels led to concepts of
A “limiting temperature” TH (Hagedorn, 1965)
A phase transition(?) in hadronic matter
before quarks were understood as underlying constituents
Density of States vs Energy
0
50
100
150
200
250
0 500 1000 1500 2000
Mass (MeV)
Number of available
states
HTmaemdm
dnm /~)(
dmem
dmem
TTm
a
Tm
H
)11
(
/
~
)(
requires T < THFit to this form
with TH = 163 MeV
12-Jan-02 W.A. Zajc
8
Making Something from Making Something from NothingNothing
Experimental method:Energetic collisions of heavy nuclei
Experimental measurements:Use probes that are Auto-generated Sensitive to all time/length scales
Perturbative Vacuum
cc
MeV 200 ~)f 1(/~ etemperatur requires mT
Color Screening
cc
Explore non-perturbative “vacuum” that confines color flux by melting it
Non-perturbative Vacuum
Perturbative Vacuum
cc Particle production Our ‘perturbative’ region
is filled with gluons quark-antiquark pairs
A Quark-Gluon Plasma (QGP)
12-Jan-02 W.A. Zajc
9
Question: How to proceed with experimental design when
(Partial) answers: The QGP phase transition will not be “seen” at RHIC
Instead it will emerge as a consistent framework for describing the observed phenomena
Avoid single-signal detectors There are no* cross sections at RHIC
Except GEOM ~ few barns CENTRAL ~ (1-10)% GEOM
but QGP ~ CENTRAL ?? Preserve high-rate and triggering capabilities
Expect the unexpected High gluon density production of exotics? Color topology high anti-baryon production? New vacuum large isospin fluctuations? Maintain flexibility as long as $’s allow
Design Guidelines for QGP Design Guidelines for QGP DetectionDetection
( Guidance written in 1991 )
12-Jan-02 W.A. Zajc
10
Studying Complex Studying Complex PhenomenaPhenomena
Ample historical evidence for categorizing complex physics through correlation of observables For example, Hertzprung-
Russell in astronomy PHENIX will approach QGP
via as many channels as possible
12-Jan-02 W.A. Zajc
11
PHENIX Approach to QGP PHENIX Approach to QGP DetectionDetection
1. DeconfinementR() ~ 0.13 fm < R(J/) ~ 0.29 fm < R(’ )
~ 0.56 fm Electrons, Muons
2. Chiral Symmetry RestorationMass, width, branching ratio of to e+e-, K+K-
with M < 5 Mev: Electrons, Muons, Charged HadronsBaryon susceptibility, color fluctuations, anti-
baryon production: Charged hadronsDCC’s, Isospin fluctuations: Photons, Charged Hadrons
3. Thermal Radiation of Hot GasPrompt , Prompt * to e+e-, +- :
Photons, Electrons, Muons
4. Strangeness and Charm Production
Production of K+, K- mesons: HadronsProduction of , J/, D mesons: Electrons, Muons
5. Jet QuenchingHigh pT jet via leading particle spectra:
Hadrons, Photons
6. Space-Time EvolutionHBT Correlations of ± ±, K± K± :
Hadrons
Summary: Electrons, Muons, Photons, Charged Hadrons
12-Jan-02 W.A. Zajc
12
Measuring all TimescalesMeasuring all Timescales PHENIX can and will
do this program Early timescales in
collision typically probed by “hard processes”.
“Hard” high momentum transfer rare luminosity limited
Run-1 (Summer 2000): < ~1 b-1 recorded ~5M “minimum bias”
events Run-2 (2001-2)
~24 b-1 recorded ~200M events
“sampled”
Timescale Probe
Available
Run-1?
Available
Run-2?
Initial Collision Hard ScatteringSingle "jet" via leading particle Yes Yesphoton + "jet" No Yes?
Deconfinement High-Mass Vector Mesons
J / , ' screening No Observation (non)screening No No
Chiral Restoration Low-Mass Vector Mesons mass, width No Yes? branching ratios No Yes?
QGP Thermalization Photons 0
, ' 0 only Yescontinuum direct; very soft No Yes
QGP Thermalization Dileptonsnon-resonant: 1-3 GeV No Yes? soft continuum, <1 GeV No No
QGP Thermalization Heavy Quark Productionopen charm No Noopen charm via single lepton Yes Yes
Hadronization HadronsHBT Interferometry, /K Yes Yesstrangeness production: K, Yes Yesspectra of identified hadrons Yes Yes
Hydrodynamics Global VariablesET, dN/dy Yes Yes
12-Jan-02 W.A. Zajc
13
Copyright, 1996 © Dale Carnegie & Associates, Inc.
General Issues of Experimental General Issues of Experimental DesignDesign
One must deal with hadronic and electromagnetic interactions at ~all length scales (I.e., not just
the first 10 fm) Convenient
to represent on a “logarithmic light-cone”
More importantly, PHENIX faced severe design contraints: Central arms-
minimize material in aperture
Muon arms-maximize absorber in aperture
12-Jan-02 W.A. Zajc
14
PHENIX at RHICPHENIX at RHIC
2 central spectrometers
2 forward spectrometers
3 global detectors
West
EastSouth
North
12-Jan-02 W.A. Zajc
15
Central Magnet
East Carriage
West Carriage
Ring Imaging CerenkovDrift Chamber
Beam-Beam Counter
PHENIX CentralPHENIX Central
12-Jan-02 W.A. Zajc
16
ScheduleSchedule
2 central spectrometers
2 forward spectrometers
3 global detectors
1999
20002001
2002
12-Jan-02 W.A. Zajc
17
Run-1 ConfigurationRun-1 Configuration
Two central arms Mechanically
~complete Roughly half of aperture
instrumented Global detectors
Zero-degree Calorimeters (ZDCs)
Beam-Beam Counters (BBCs)
Multiplicity and Vertex Detector (MVD, engineering run)
12-Jan-02 W.A. Zajc
18
Run-1 Run-1 AccomplishmentsAccomplishments
First collisions:15-Jun-00 Last collisions: 04-Sep-00
During this period: Commissioned Zero-Degree Calorimeters Beam-Beam Counters Multiplicity and Vertex Counter Drift Chambers Pad Chambers Ring Imaging Cerenkov Counter Time Expansion Chamber Time-of-Flight Counters Electromagnetic Calorimeter Muon Identifier Minimum Bias Triggers Data Acquisition System
Recorded ~5M minimum bias events
12-Jan-02 W.A. Zajc
19
Topic 1Topic 1
Charged particle multiplicity How many particles are produced? What to they tell us about underlying reaction
mechanism?
Answered in
“Centrality Dependence of Charged Particle Multiplicity in Au-Au Collisions at sqrt(s) = 130 GeV”, K. Adcox et al, PRL 86 (2001) 3500, preprint nucl-ex/0012008.
12-Jan-02 W.A. Zajc
20
Tracking Detectors: Pad Tracking Detectors: Pad ChambersChambers
PC3
PC1
Pixel Pad Cathode Pattern
Min bias multiplicity distribution at mid-rapidity
Cathode wire chambers using fine granularity pixel pad readout 2-D hit position, x = y ~ O( mm) 173k channels total, ~ 100 m2 detector coverage
Low-mass , rigid honeycomb/circuit board construction
All signal digitization takes place on-board in detector active region. Solves interconnect problem.
12-Jan-02 W.A. Zajc
21
~2A
PHOBOS Result on dNPHOBOS Result on dNchch/d/dQ. How many charged particles per nucleon-nucleon collision?
A: Express using dNch/d ~ “generic” dNch/dy
Npart
Number of participating nucleons
= 2 for p-p collision~2A for central A-A collision
Plot multiplicity per N-N collision ( dNch/d ) / ( Npart/2 )
Does not (quite) distinguish between “Saturation” models,
dominated by gg g “Cascade” models,
dominated by gg gg, gg ggq
( X.N. Wang and M. Gyulassy, nucl-th/0008014 )
12-Jan-02 W.A. Zajc
22
Determining NDetermining NPARTPART
Best approach (for fixed target!): Directly measure in a “zero degree
calorimeter”
(for A+A collisions)
Strongly (anti)-correlated with produced transverse energy:
PerNucleon
ZDCPART E
EAN 2
ET
ET
EZDC
NA50
12-Jan-02 W.A. Zajc
23
b
Fixing Initial Fixing Initial ConditionsConditions
Event characterization Impact parameter b
is well-defined in heavy ion collisions Event multiplicity predominantly determined
by collision geometry Characterize this by global measures
of multiplicity and/or transverse energy correlated with
ZDC Zero Degree Calorimeter Goals:
Uniform luminosity monitoring at all 4 intersections Uniform event characterization by all 4 experiments
Process: Correlated Forward-Backward Dissociation tot = 11.0 Barns (+/- few %)
0-5%5-10%
10-15%15-20%
12-Jan-02 W.A. Zajc
24
0-5%
5-10%
10-15%15-20%
Determining N(participants)Determining N(participants)
Use combination of Zero Degree Calorimeters Beam-Beam Counters
to define centrality classes Glauber modeling
to extract Nparticipants
determines Multiplicity vs. Centrality
i.e
dNch/d vs.
Nparticipants
which is presented as “specific particle production”
multiplicity per N-N collision ( dNch/d ) / ( Npart/2 )
12-Jan-02 W.A. Zajc
25
“Specific multiplicity” is not flat Yields grow significantly faster
than Nparticipants
Evidence for term ~ Ncollisionscollpartch NBNAddN
0
PHENIX ResultsPHENIX Results
28.088.0 A12.034.0 B19.038.0/ AB
In this cartoon are there
5 N-N collisions 5 x Npp?
OR
6 participants 3 x Npp?
Qualitatively consistent with HIJING Inconsistent with
some saturation models One interpretation:
Evidence for both “soft ” ( ~ Nparticipants)
“hard” (~ Ncollisions ) processes
12-Jan-02 W.A. Zajc
26
Consistency between ExperimentsConsistency between Experiments Trend
incompatible with final-state gluon saturation model
Good agreement with model based on initial-state saturation (Kharzeev and Nardi, nucl-th/0012025)
dN/d
/ .5N
part
Npart
Excellent agreement between (non-trivial) PHENIX and PHOBOS analyses of this systematic variation with nuclear overlap.
12-Jan-02 W.A. Zajc
27
1 J.P Blaizot, A.H. Mueller, Nucl. Phys. B289, 847 (1987)
Saturation condition ~ 1
QS2 ~ S A xG(x,Q2) / R2 ~ A1/3
Gluon saturation at RHICGluon saturation at RHICWhen do the gluons overlap
significantly?
2R m/pz
/ xpz
2R
Longitudinal
dT = /Q
Transverse
D. Kharzeev, nucl-th/0107033
So for /mx ~ 2R , ~ all constituents contribute
Parton density ~ A xG(x,Q2) / R2
Parton cross section
~ S 2 / Q2
12-Jan-02 W.A. Zajc
28
Initial State Partons?Initial State Partons? Procedure:
Determine scale QS2 as function of nuclear
overlap QS
2 ~ A1/3 ~ Npart1/3
Assume final-state multiplicities proportional to number of initial-state gluons in this saturated regime: dN/d = c Npart xG(x, QS
2)
Note that DGLAP requires evolution of xG(x, QS
2) :
xG(x, QS2) ~ ln (QS
2 / QCD2)
Results: “Running” of the multiplicity yield
directly from from xG(x, Q2) +DGLAP
12-Jan-02 W.A. Zajc
29
Topic 2Topic 2
Energy Density What energy densities are achieved in
central collisions at RHIC? How does this compare to previous
experiments?
Answered in
"Measurement of the Midrapidity Transverse Energy Distribution from sqrt(s) = 130 GeV Au-Au Collisions at RHIC", K. Adcox et al., PRL 87 (2001) 052301, preprint nucl-ex/0104015, or from the PHENIX site: pdf, ps, ps.gz.
12-Jan-02 W.A. Zajc
30
dydz
0
0 .2
0 .4
0 .6
0 .8
1
1 .2
1 .4
1 .6
-6 -4 -2 0 2 4 6
y
dy
dn
Dynamics 101Dynamics 101
Q. How to estimate initial energy density?A. From rapidity density
“Highly relativistic nucleus-nucleus collisions: The central rapidity region”, J.D. Bjorken, Phys. Rev. D27, 140 (1983).
Assumes ~ 1-d hydrodynamic expansion
Invariance in y along “central rapidity plateau”(I.e., flat rapidity distribution)
Then
since boost-invariance of matter where ~ 1 fm/c
dy
dE
RdyR
dE
dzR
dE
V
E T
TT
T
T
T
222
1~~
12-Jan-02 W.A. Zajc
31
For the most central events:
PHENIX
EMCAL
R2
Determining Energy Determining Energy DensityDensity
Bjorken~ 4.6 GeV/fm3
~30 times normal nuclear density ~1.5 to 2 times higher than any previous experiments
Bjorken formula for thermalized energy density
time to thermalize the system (0 ~ 1 fm/c)~6.5 fm
What is the energy density achieved?
How does it compare to the expected phase transition value ?
dy
dE
RT
Bj0
2
11
dydz 0
12-Jan-02 W.A. Zajc
32
Topic 3Topic 3
High Transverse Momenta Particle Production What is the rate for producing high pT particles? How does it vary with centrality?
Answered in
"Suppression of Hadrons with Large Transverse Momentum in Central Au+Au Collisions at sqrt(s) = 130 GeV", K. Adcox et al., Phys. Rev. Lett. 88, 022301 (2001), preprint nucl-ex/0109003, or directly from Phenix
12-Jan-02 W.A. Zajc
33
Energy Loss of Fast PartonsEnergy Loss of Fast Partons Many approaches
1983: Bjorken
1991: Thoma and Gyulassy (1991)
1993: Brodsky and Hoyer (1993)
1997: BDMPS- depends on path length(!)
1998: BDMS
Numerical values range from ~ 0.1 GeV / fm (Bjorken, elastic scattering of
partons) ~several GeV / fm (BDMPS, non-linear interactions of
gluons)
222
2
2 4ln~
4ln
4
303
M
ETT
M
ET
dx
dESS
D
SF
ETC
dx
dE
ln3
4 22
2
2Tk
dx
dE
gg
DRS
LL
C
dx
dE
ln
8
2
24
2TC
s
kN
dx
dE
12-Jan-02 W.A. Zajc
34
Jet Physics at RHICJet Physics at RHIC Tremendous interest in hard scattering
(and subsequent energy loss in QGP) at RHIC Predictions that dE/dx ~ (amount of matter to be traversed) Directly due to non-Abelian nature of medium
But: “Traditional” jet methodology fails at RHIC Dominated by the soft background:
For a typical jet cone R = 0.33(R2 = 2 + 2) have
<nSOFT> ~ 64
<ET> ~ 25 GeV Fluctuations in this soft background
swamp any jet signal for pT < ~ 40 GeV:
Investigate by (systematics of) high-pT single particles
Jet Cross Section in R=0.33 for Au+Au at 200 A*GeV
1.0E-05
1.0E-04
1.0E-03
1.0E-02
1.0E-01
1.0E+00
1.0E+01
1.0E+02
1.0E+03
20 30 40 50 60
pT (GeV/c)
d/d
pT (
mb
/Ge
V)
Soft particle background (assuming Poisson fluctuations)
Jet Cross-section
RJet
Axis
12-Jan-02 W.A. Zajc
35
Shrinking the ConeShrinking the Cone Why not let R 0 ?
Good: Reduces fluctuations of “underlying event”
“Bad” : Doesn’t contain all the energy of the jet
TTT
m
TTT
TT
nTT
dp
jetd
dp
jetd
dp
particledzzzD
zPpzDdP
jetddP
dp
particled
p
A
dp
jetd
)(
100
1)()constant(
)(/)1(~)(
)()()()(
,~)(
Hadron Interactions, Collins and Martin
R < 1.0
R < 0.7
PYTHIA study by M. Chiu
R < 0.5
But: For “quenching”, maximal effect
expected on the “leading particle”(that is, “lost” energy probably stays in the cone, which confuses picture )
High pT single particles have a simple relation to parent parton dynamics
12-Jan-02 W.A. Zajc
36
As you like it:
More usefully:
Magnetic Magnetic MeasurementsMeasurements
Fu
c
e
d
dp
meterTesla
cGeV
c
eRB
c
ep
/3.0,||
s
STAR
]GeV/c[%)1(~%)1(~ pp
p
~stuff in aperture ~spatial accuracy
Bvc
e
dt
pd
|| p
B
ds
rd
c
e
ds
rd
ds
d
1 meter of 1 Tesla field deflects 1 GeV/c by ~17O
Real world:
12-Jan-02 W.A. Zajc
37
Drift ChamberDrift Chamber Jet -chamber anode/cathode structure
modified for HI high multiplicity Joint Russia/US design & construction All Titanium frame x = 120 m , two-track sep = 2mm
Tracks in DC from Central Au-Au collision
DC wires with kapton wire dividers
Identified particle spectra using tracking system and TOF
12-Jan-02 W.A. Zajc
38
Rare processes at RHICRare processes at RHIC
Both PHENIX and STAR have measured charged particle spectra out to “small” cross sections
12-Jan-02 W.A. Zajc
39
ExpectationsExpectations
Particle production via rare processes should scale with Ncoll, the number of underlying binary nucleon-nucleon collisions
FunctionThickness
),()( dzzddTA dz
d
FunctionOverlap
)2
()2
()( sdb
sTb
sTbT BAAB
b
INTINT
)(2AB
AB
INT
small""for
1
then is section cross TOTAL the
section cross withinteract which
tsconstituen B has B"" Nucleus and
tsconstituen Ahas A"" Nucleus If
BA
ebd bTABINT
Assuming no “collective” effects
Test this on production at high transverse momentum
12-Jan-02 W.A. Zajc
40
A*B scaling at RHIC?A*B scaling at RHIC? NO! Both STAR and PHENIX data see deficit of
particles at high transverse momenta,relative to expected A*B scaling:
Deficit opposite sign of enhancements previously seen in nuclear collisions (Cronin effect)
12-Jan-02 W.A. Zajc
41
PHENIX
0 reconstruction
pT > 2 GeV/c
Asymmetry < 0.8
A good example of a “combinatoric” background Reconstruction is not done particle-by-particle Recall: 0 and there are ~2000 s per unit rapidity
So: 0 1
0 2
0 3
0 N
Unfortunately, nature doesn’t use subscripts on photons
N correct combinations: ( ), ( ), … ( ),
N(N-1)/2 – N incorrect combinations ( ), ( ), … Incorrect combinations ~ N2 (!)
Solution: Restrict N by pT cuts use high granularity, high resolution detector
00 Reconstruction Reconstruction
p-p at 200 GeV
12-Jan-02 W.A. Zajc
42
Additional EvidenceAdditional Evidence PHENIX sees a larger effect in the 0
spectrum:
Consistent with crude estimatesof additional energy loss in adeconfined medium
12-Jan-02 W.A. Zajc
43
Discovery of “Jet Quenching”?Discovery of “Jet Quenching”? Some would
(have) claimed just that
To be done before this can be experimentally established: Improve pT
range (Run-2) Measure p-p
spectrum (Run-2)
Study in “cold nuclear matter” with p-A collisions (Run-2?)
Vary system size (Run-2??)
12-Jan-02 W.A. Zajc
44
Topic 4Topic 4
Particle Composition What are the relative yields of pions, kaons and
protons? How do their yields depend on transverse
momentum? How do their yields depend on centrality?
Answered in
"Centrality dependence of pi+-, K+-, p and pbar production from sqrt(s)=130 GeV Au + Au collisions at RHIC", K. Adcox et al., Submitted to PRL, preprint nucl-ex/0112006,
12-Jan-02 W.A. Zajc
45
Particle Identification by TOFParticle Identification by TOF
The most direct way Measure by distance/time Typically done via scintillators
read-out with photomultiplier tubes Time resolutions ~ 100 ps
224
22
s
s
t
t
p
p
m
m
Performance: t ~ 100 ps on 5 m flight path P/K separation to ~ 2 GeV/c K/p separation to at least 4 GeV/c
K
p
e
Exercise: Show
12-Jan-02 W.A. Zajc
46
Putting it all togetherPutting it all together Charged pions match
nicely with 0’s Supports conclusion of
preferential suppression of pions
“High” pT dominated by protons and kaons??
Summed hadrons consistent with our first result on multiplicity
12-Jan-02 W.A. Zajc
47
Agreement between Agreement between ExperimentsExperiments
Overlaps in detector capabilities a feature of RHIC program
Excellent agreement in common observables provides confidence
12-Jan-02 W.A. Zajc
48
For the first timein heavy ion collisions, more baryons are pair-produced than brought in from initial state
Approaching the Early Approaching the Early UniverseUniverse
Early Universe: Anti-proton/proton = 0.999999999
We’ve created “pure” matterapproaching this value
12-Jan-02 W.A. Zajc
49
Topics not coveredTopics not covered This is a sample of the many physics
topics addressed by PHENIX in Run-1.
Additional results available on Flow
Fluctuations in <pT> Charge
Centrality dependence of particle yields
HBT
Charm (next lecture)
12-Jan-02 W.A. Zajc
50
Summary of ResultsSummary of Results How many particles are produced?
More than ever produced previously Have we made contact with the
early universe?Closer than all previous heavy ion experiments
Have we made sufficiently high energy densities for the phase transition?Yes
Is there evidence that the dense matter behaves collectively?Yes
Are there results consistent with the formation of a new state of matter?Perhaps ??????!?
(1 = 1 experiment )