particle production from ee to aa and search for the qcd ... · particle production from ee to aa...
TRANSCRIPT
Particle production from ee to AAand search for the QCD Critical Point with PHENIX
Mate Csanad (Eotvos University, Budapest)for the PHENIX Collaboration
and for Roy L. et al (the authors of arXiv:1601.06001)
54th International School of Subnuclear Physics, Erice, Sicily, 2016
June 19, 2016
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 1 / 15
Particle production from ee to AA
Particle production mechanisms from ee to AA
A+A(B) collisions: frequently described with thermo/hydrodynamics
Model ingredients: macroscopic variables (temperature, entropy)see e.g. W. Kittel and E. A. DeWolf, Soft Multihadron Dynamics, (World Scientic, 2005)
or recent PHENIX, PHOBOS, STAR, ALICE, CMS and ATLAS papers
Microscopic phenomenology used in e− + e+, e± + p, p(p)+p or p+A
Perturbative gluon exchange, gauge fields, strings, partonssee e.g. Kharzeev et al., NPA747; Armesto et al., PRL94, Dusling et al., PRD87
and other references in arXiv:1601.06001
Similarity in particle production ⇔ available Eeff for part. productionFeinberg Phys Rept. 5, Albini et al. Nuovo Cim A32, Basile et al. PLB92&Nuovo Cim A67, Nyiri, IJMP A18
Sarkisyan and Sakharov, hep-ph/0410324
κ1
√see ≈ κ2
√spp ≈ κ3
√sep with κ1 ≡ 1 (1)
Role of quark participant pairs in pp, pA and AA describing initial stage
What do the measurements tell us?
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 2 / 15
Particle production from ee to AA
〈Nch〉 vs.√s scaling in e− + e+, p(p)+p and e± + p
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 3 / 15
Entropy ansatz: S ∼ (TR)3,dNch/dη and 〈Nch〉 ∼ S
Initial stage variable Npp number ofparticipant pairs (Npp = 1 here)
Global assumption: N1/3pp ∝ R
Simple result:[〈Nch〉/Npp]1/3 ∼ T ∼ 〈pT 〉Scaling versus κn
√s
Fit result:〈Nch〉 =
[b〈Nch〉 + m〈Nch〉 log(κn
√s)]3
b〈Nch〉 = 1.22± 0.01m〈Nch〉 = 0.775± 0.006See details and references
in Lacey et al., arXiv:1601.06001
Particle production from ee to AA
Quark participant scaling from pp through pA to AA
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 4 / 15
Initial stage size measure: Npp
Use quark participant pairs Nqpp
Recall [(dNch/dη)/Nqpp]1/3 ∼ T
Flat size (Nqpp) dependence
Strikingly similar√sNN trends for
p+p and A+A(B) collisions
Common production mechanism?
Small deviation for . 2 TeV
Smooth trend over full range
Details in Lacey et al.,
arXiv:1601.06001
When does QGP production start?
Phases of QCD?
Emergent QCD phenomena
Exploring the QCD phase diagram
QCD: fundamental theory mostly understood
Emergent phenomena (phases, nucleons, etc): hard to handle
Important part of the phase diagram: RHIC energies!
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 5 / 15
Emergent QCD phenomena
The PHENIX Experiment at RHIC
West Beam View
PHENIX Detector 2010
East
MPC RxNP
HBD
PbSc PbSc
PbSc PbSc
PbSc PbGl
PbSc PbGl
TOF-E
PC1 PC1
PC3
PC2
CentralMagnet TEC
PC3
BB
RICH RICH
DC DC
Aerogel
TOF-W 7.9 m =
26 ft
RHIC collisions: versatile in energy, 5-500 GeV/nucleon
Versatile in colliding nuclei: p, d, Cu, Au, Al, He, U
Tracking via Drift Chambers & Pad Chambers
PID via time of flight from TOF & EMCal
Momentum resolution: δp/p ≈ 1.3%⊕ 1.2%× p GeV/c
Charged pion ID from p ≈ 0.18 to 2 GeV/c
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 6 / 15
Emergent QCD phenomena
The RHIC Beam Energy Scan
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 7 / 15
Bose-Einstein correlations
Introduction to Bose-Einstein correlations
In case of free bosons, momentum correlation function:
C2(q) ' 1 +
∣∣∣∣∣ S(q)
S(0)
∣∣∣∣∣2
, S(q) =
∫S(x)e iqxd4x , q = p1 − p2 (2)
Final state interactions distort the simple Bose-Einstein picture
Coulomb interaction important, handled via two-particle wave function
Resonance pions: Halo around primordial Core
Halo part unresolvable
J.
Bolz et al, Phys.Rev. D47 (1993) 3860-3870
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 8 / 15
Bose-Einstein correlations
Levy shaped source and the Critical Point
Levy distributed source and resulting correlation function:
S(r)α,R =1
(2π)3
∫d3qe iqre−
12|qR|α ⇒ C2(q) = 1 + λ · e−(R|q|)α (3)
Parameters: λ,R, α, may depend on average pair momentumRelated to anomalous diffusion or generalized random walkMetzler, Klafter, Physics Reports 339 (2000) 1-77
α = 2: Gauss, α = 1: Cauchy; power-law tail for α < 2
Levy index α identified with critical exponent ηSpatial correlation function ∝ r−(d−2+η) → critical η exponentSymmetric stable distributions (Levy) → spatial corr. ∝ r−1−α
QCD universality class ↔ 3D IsingHalasz et al., Phys.Rev.D58 (1998) 096007, Stephanov et al., Phys.Rev.Lett.81 (1998) 4816
3D Ising: η = 0.03631(3), Random field 3D Ising: η = 0.50± 0.05Rieger, Phys.Rev.B52 (1995) 6659, El-Showk et al., J.Stat.Phys.157 (4-5): 869
Change in α↔ proximity of CEP
Motivation for precise Levy HBT
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 9 / 15
PHENIX Levy HBT analysis results
Example correlation function measurement result
Measured in 31 m2T = m2 + p2
T bins for π+π+ and π−π− pairs
Physical parameters: R, λ, α; measured versus pair mTMate Csanad (Eotvos University) 54th ISSP, Erice, 2016 10 / 15
PHENIX Levy HBT analysis results
Example correlation function measurement result
Measured in 31 m2T = m2 + p2
T bins for π+π+ and π−π− pairs
Physical parameters: R, λ, α; measured versus pair mTMate Csanad (Eotvos University) 54th ISSP, Erice, 2016 10 / 15
PHENIX Levy HBT analysis results
Levy scale parameter R versus pair mT
]2 [GeV/cTm0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
R [f
m]
3
4
5
6
7
8
9
10
freeα
= 200 GeVNNsMinBias Au+Au @
freeα
-π-π+π+π
PH ENIXpreliminary
Similar decreasing trend as usual Gaussian HBT radii
Hydro behaviour: 1/R2 ∼ mT , not incompatible
Interesting, since simple hydro would mean α = 2
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 11 / 15
PHENIX Levy HBT analysis results
Correlation strength λ versus pair mT
]2 [GeV/cTm0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
λ
0.6
0.8
1
1.2
1.4
1.6
freeα
= 200 GeVNNsMinBias Au+Au @
freeα
-π-π+π+π
PH ENIXpreliminary
From the Core-Halo model: λ = (NC/(NC + NH))2
Large and correlated systematic uncertainties
Observed decrease at small mT → increase of halo fractionMay be interesting phyics in λ, strange speculations:
Resonance effects, chiral symmetry restoration?Partially coherent pion production?Aharanov-Bohm effect?
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 12 / 15
PHENIX Levy HBT analysis results
Levy exponent α versus pair mT
]2 [GeV/cTm0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
α
1
1.05
1.1
1.15
1.2
1.25
1.3
1.35
1.4
1.45
1.5 = 200 GeVNNsMinBias Au+Au @
0.003± = 1.134 0α
/ndf = 159/612χ
-π-π+π+π
PH ENIXpreliminary
Far from Gaussian (α = 2) and Cauchy/exponential (α = 1)
Also far from 3D Ising value at CEP (α ≤ 0.5)
More or less constant (at least within systematic errors)
Although the constant fit is statistically not acceptable
Motivation to do fits with fixed α = 1.134
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 13 / 15
PHENIX Levy HBT analysis results
Newly discovered scaling parameter R
]2 [GeV/cTm0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
[1/fm
]R
1/
0
0.1
0.2
0.3
0.4
0.5
0.6
freeα
= 200 GeVNNsMinBias Au+Au @
freeα
)α(1+⋅λR = R
-π-π+π+π
PH ENIXpreliminary
Empirically found scaling parameter
Linear in mT
Physical interpretation → open question
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 14 / 15
Summary
Scaling of Nch from ee to AA data from GeV to TeV√s regions
Effective energy notation successfulQuark participants emergingSpecial signatures at special energies?
Seach for the QCD critical point at RHIC
New measurement method of Levy sources developed with PHENIX
Levy exponent α at√sNN = 200 GeV far from Gaussian or critical values
Correlation strength λ hints at interesting physicsNew empirically found scaling parameter R
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 15 / 15
Thank you for your attention!
Let me invite You to the 16th Zimanyi Winter School on Heavy Ion PhysicsBudapest, Hungary, Dec. 5 – 9, 2016
http://zimanyischool.kfki.hu/16/
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 16 / 15
Similarities from p+p through p+A to A+A(B)
Similar charged particle multiplicities (Nch)
Similar pseudorapidity densities (dNch/dη)
Azimuthal long range (|∆η| ≥ 4) angular correlations, “ridge”
Collective anisotropic flow in A+A collisions
Also in p+p, p+Pb, d+Au and He+AuALICE PLB719, ATLAS PRL110, CMS PLB718, PHENIX PRL114, PHENIX PRL115
Qualitative consistency achieved with hydroSee e.g. Bozek, PRC85, the Buda-Lund model from Csorgo et al., NPA661, JPhysG30, EPJA38, . . .
Common underlying particle production mechanism dominating?
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 17 / 15
Our framework to capture underlying physics
Macroscopic entropy (S) ansatz
S ∼ (TR)3 ∼ const. (4)
dNch/dη and 〈Nch〉 ∼ S (5)
Initial stage variable Npp number of participant pairsNpp = 1 for e− + e+, e± + p and p(p)+pNucleon or quark participant pairs (Nnpp, Nqpp) in p+A, A+A(B)
Further assumption: N1/3pp ∝ R ⇒ [(dNch/dη)/Npp]1/3 ∼ T ∼ 〈pT 〉
Monte Carlo Glauber calculations performed to obtain Nnpp and Nqpp.Lacey et al. PRC83, Eremin et al. PRC67, Bialas et al. PLB649, Nouicer EPJC49, PHENIX PRC89
Subset of initial particles become participants by an initial inelasticN+N or q+q interaction.
Nnp = 2Nnpp or Nqp = 2Nqpp
N+N (q+q) cross sections taken from literature Fagunders et al, J. Phys. G40
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 18 / 15
Nucleon participant scaling in A+A(B) collisions
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 19 / 15
All systems:[dNch/dη||η|=0.5
Nnpp
]1/3∼ T
(a): ∝ log(dNch/dη) ∼ log S ,
(b): ∝ N1/3npp ∼ R
Logarithmic S-dependenceLinear size dependence(at a given
√sNN)
〈pT 〉 increases with√sNN and log(dNch/dη)
Pseudorapidity density factorizes
into contributions depending on√sNN and N
1/3npp
Slope increaseses with beam energy
Lack of sensitivity to system type (Cu+Cu, Cu+Au,Au+Au, U+U), for fixed
√sNN.
PHENIX Levy HBT analysis
Dataset:√sNN = 200 GeV Au+Au, min. bias, ∼7 billion events
Goal: detailed shape analysis of two-pion correlation functions
Correlation functions measured in fine pair mT =√m2 + p2
T binsLevy source instead of Gaussian → better agreement with dataPrecision measurement of source parameters versus pair mT
Result: λLevy (mT ), αLevy (mT ),RLevy (mT )
Tracking: DCH & PC1, 2σ matching cuts in TOF & EMCal
Particle identification:
time-of-flight data from EMCal & TOF, momentum, flight length2σ cuts on m2 distribution
Pair-cuts:
Exclude particles hitting the same tower/slat/stripCustomary shaped cuts in ∆ϕ−∆z plane for TOF, EMCal, DCH
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 20 / 15
Example correlation functions
|k| [GeV/c]0 0.05 0.1 0.15 0.2 0.25 0.3
2C
1
1.1
1.2
1.3
1.4
1.5
1.6
Raw corr. function
Coulomb factor×Raw corr.
Coulomb factor
|k|)ε N(1+×;|k|) α,R,λ(LevyCoulC
|k|)ε N(1+×;|k|) α,R,λ(Levy0C
|k|)εN(1+
= 0.48-0.5 GeV/cT
, p-π-π = 200 GeV, NNsMinBias Au+Au @
0.06± = 1.14 λ 0.28 fm±R = 6.90 fm
0.03± = 1.09 α 0.001± = -0.025 ε 0.0001±N = 1.0015
/NDF = 295/2512χconf. level = 0.0293
|k|)ε N(1+×)) α |k|)⋅ R⋅ exp(-(2⋅λ=(1+Levy0C
|k| [GeV/c]0 0.05 0.1 0.15 0.2 0.25 0.3
(dat
a-fit
)/er
ror
-4-3-2-101234
PH ENIXpreliminary
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 21 / 15
Example correlation functions
|k| [GeV/c]0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45
2C
1
1.1
1.2
1.3
1.4
1.5
1.6
Raw corr. function
Coulomb factor×Raw corr.
Coulomb factor
|k|)ε N(1+×;|k|) α,R,λ(LevyCoulC
|k|)ε N(1+×;|k|) α,R,λ(Levy0C
|k|)εN(1+
= 0.72-0.74 GeV/cT
, p+π+π = 200 GeV, NNsMinBias Au+Au @
0.09± = 1.05 λ 0.30 fm±R = 4.84 fm
0.05± = 1.20 α 0.000± = -0.038 ε 0.0002±N = 1.0007
/NDF = 413/3802χconf. level = 0.1180
|k|)ε N(1+×)) α |k|)⋅ R⋅ exp(-(2⋅λ=(1+Levy0C
|k| [GeV/c]0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45
(dat
a-fit
)/er
ror
-4-3-2-101234
PH ENIXpreliminary
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 22 / 15
Levy scale parameter R versus pair mT
]2 [GeV/cTm0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
R [f
m]
3
4
5
6
7
8
9
10
freeα
= 200 GeVNNsMinBias Au+Au @
freeα
-π-π+π+π
PH ENIXpreliminary
]2 [GeV/cTm0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
]2 [1
/fm2
1/R
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
freeα
= 200 GeVNNsMinBias Au+Au @
freeα
-π-π+π+π
PH ENIXpreliminary
Similar decreasing trend as usual Gaussian HBT radii
Hydro behaviour: 1/R2 ∼ mT , not incompatible
Hard to say whether the 1/R2 scaling is linear or not
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 23 / 15
Correlation strength λ versus pair mT
]2 [GeV/cTm0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
λ
0.6
0.8
1
1.2
1.4
1.6
freeα
= 200 GeVNNsMinBias Au+Au @
freeα
-π-π+π+π
PH ENIXpreliminary
From the Core-Halo model: λ = (NC/(NC + NH))2
Large and correlated systematic uncertainties
Observed decrease at small mT → increase of halo fractionDifferent effects can cause change in λ, strange speculations
Resonance effects, partial chiral symmetry restoration?Partially coherent pion production?Aharanov-Bohm effect?
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 24 / 15
Levy exponent α versus pair mT
]2 [GeV/cTm0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
α
1
1.05
1.1
1.15
1.2
1.25
1.3
1.35
1.4
1.45
1.5 = 200 GeVNNsMinBias Au+Au @
0.003± = 1.134 0α
/ndf = 159/612χ
-π-π+π+π
PH ENIXpreliminary
Far from Gaussian (α = 2) and Cauchy/exponential (α = 1)
Also far from 3D Ising value at CEP (α ≤ 0.5)
More or less constant (at least within systematic errors)
Although the constant fit is statistically not acceptable
Motivation to do fits with fixed α = 1.134
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 25 / 15
Levy scale parameter R with fixed α = 1.134
]2 [GeV/cTm0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
R [f
m]
3
4
5
6
7
8
9
10
= 1.134α
= 200 GeVNNsMinBias Au+Au @
= 1.134α
-π-π+π+π
PH ENIXpreliminary
]2 [GeV/cTm0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
]2 [1
/fm2
1/R
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
= 1.134α
= 200 GeVNNsMinBias Au+Au @
= 1.134α
-π-π+π+π
PH ENIXpreliminary
More smooth trend
Hydro behaviour seems to be more valid
The linearity of 1/R2 holds
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 26 / 15
Correlation strength λ with fixed α = 1.134
]2 [GeV/cTm0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
λ
0.6
0.8
1
1.2
1.4
1.6
= 1.134α
= 200 GeVNNsMinBias Au+Au @
= 1.134α
-π-π+π+π
PH ENIXpreliminary
More smooth trend
Smaller systematic errors
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 27 / 15
Newly discovered scaling parameter R
α = 1.134 fixed
]2 [GeV/cTm0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
[1/fm
]R
1/
0
0.1
0.2
0.3
0.4
0.5
0.6
= 1.134α
= 200 GeVNNsMinBias Au+Au @
= 1.134α
)α(1+⋅λR = R
-π-π+π+π
PH ENIXpreliminary
Empirically found scaling parameter
Linear in mT
Physical interpretation → open question
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 28 / 15
Beam energy & system size dependence of HBT radii
Corr. func. in 3D, Gaussian fit (details: arXiv:1410.2559)
π+π+, π−π− data combined
Linear 1/√mT scaling of HBT radii
Interpolation to common mT , PHENIX and STAR consistent
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 29 / 15
Beam energy & system size dependence of HBT radii
quantities related to emissionduration and expansionvelocity
non-monotonic patterns
indication of CEP?
More precise mapping and further detailed studies required
Is there any other way to find the critical point?
Maybe Levy exponent α!
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 30 / 15
Generalized random walks
Random walk, step size with infinite varianceRandom Walk Generalized random walk
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 31 / 15
(Speculative) connection between λ(mT ) and UA(1)
Partial chiral UA(1) restoration ⇒ reduction of the η′ mass
Reduced η′ mass ⇒ larger abundance of η′
Decay: η′ → ηπ+π−, η → π+π−π0 (some probability)
In summary: UA(1) restoration ⇒ larger number of decay pions
Larger number of decay pions ⇒ larger halo, i.e. reduced λ
These decay pions have small momentum, as η′ has no availableenergy to give them momentum
λ reduces only at small momentum!
Hole in λ(mt) distribution!
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 32 / 15
(Speculative) connection: λ(mT ) and Aharanov-Bohm
Mixing of a→ A, b → B and a→ B, b → A
Correlation strength: 〈|Ψ2|2 − 1〉, with|Ψ2|2 = 1 + cos(∆x∆k) and∆x = x1 − x2, ∆k = k1 − k2
Average on local production probability and random phases at emission(c.f. incoherent production); if no random phases, no correlation!
Observation: closed loop in the above figure
Particles go through an expanding, fluctuating medium
Pairs picking up additional random phases φ,|Ψ2|2 = 1 + cos(∆x∆k + φ)
Assume φ distribution of exp[−φ2/(2σ2), average on possible φ values
〈|Ψ2|2 − 1〉φ = exp[−σ2/2] cos(∆x∆k)
Correlation strength λ reduced by exp[−σ2/2]
Suppose σ2 ∼ 1/m2T ⇒ “hole” in λ(mT )
Idea from Antal Jakovac (Eotvos University)
Mate Csanad (Eotvos University) 54th ISSP, Erice, 2016 33 / 15