ATLAS measurement of dipolar flow (v 1 ) in Pb-Pb collisions Jiangyong Jia for the ATLAS Collaboration WWND 2012 April 7 th - 14 rd Based on results in.

ATLAS measurement of dipolar flow (v1) in Pb-Pb collisions

Jiangyong Jia for the ATLAS Collaboration

WWND 2012 April 7th- 14rd

Based on results in 1203.3087 (v1-v6 summary)https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/PAPERS/HION-2011-01/

Motivation 2

ε2

ε3

ε4

Initial geometry has multi-pole shape due to fluctuations.

~400 nucleons >20000 particles

Probe shape of initial geometry and transport properties

Alver, Roland etc

Fourier expansion of azimuthal distribution in momentum space

Also measure with two-particle correlation (2PC)

Eccentricity from Glauber model

Sizable eccentricities for all order

vn~εn in linearized hydro, but Complicated by dynamic mixing

during expansion, especially for n>3.

Higher order vn damped more by viscosity.

ε1 is smaller, but v1 is not affected by dynamic mixing and less affected by viscosity.

3

Two-particle correlation (2PC) method

Long range structure (“ridge”, “double-hump”) well described by v1,1-v6,6.

Factorization works for n=2-6 Soumya Mohapatra’s talk

Not for n=1.

4

|Δη|>2

v1 physics

v1(η) dependence has a rapidity-odd and a rapidity-even component rapidity-odd v1 reflect sideward bounce off, small at mid-rapidity

rapidity-even v1 is associated with the dipole asymmetry in initial geometry

v1 also affected by global momentum conservation Balance of pT of one particle by all other particles: N. Borghini nucl-th/0004026

Inversely proportional to multiplicity M, linear in pT.

5

Odd component: vanish at η=0 Even component: ~boost invariant in η

Fig from P. Stankus

Dipole in Cosmic Microwave Background The CMB is dominated by a dipole, representing the Doppler shift of

observer (600km/s)

6

Rapidity-even v1 and expected trend in v1,1

Values inferred from STAR 2PC data by estimating the second term.

7

Expected v1,1 contribution from rapidity-even v1

a,b both at high pT positive and increase with pTa,b (convex shape)

a,b both at vey low pT positive

a at low pT, b at high pT negative, more negative at higher pTb (concave shape)

Luzum et.al

Do we see these trends in the data?

Δη dependence of v1,1

Peripheral collisions(GMC dominated): v1,1 is always negative at large Δη. More negative at higher pT. Magnitude decrease at large Δη

Influence of jets and dijets

Central collisions(flow dominated): v1,1 is negative at low pT, become positive at large pT. Magnitude flat in Δη.

Consistent with a rapidity-even v1.

Integrate over 2<|Δη|<5 and look at the pT dependence

8

pT dependence of v1,1 data

Peripheral collisions (GMC dominated): v1,1 negative, linear in pTa,pT

b.

Central collisions (flow dominated): v1,1 becomes positive at 1.5-6 GeV range, but on top of a negative momentum conservation component

Cross each other at low pT where flow driven v1,1 ~ zero.

9

Can we account for both with a two-component fit?

Two-component fit Simultaneous fit of v1,1 data of each centrality with a function

Simple χ2 minimization

v1Fit(pT) defined at 15 pT, and interpolate in between. Total 16 parameters

Systematic checks: Interpolation form: Linear or cubic spline. Number of interpolation points (vary within 9-21 points) Vary pT range of fitting (0-5 to 0-10 GeV) Account for correlations between data points and fitting parameters.

10

Similar fit in arXiv:1203.0931

Fit for 0-5% centrality

Agrees with data within 1σ at pT<6 GeV. Slightly more deviation ~ 2σ in some higher pT bin.

11

Understanding v1,1=<cosΔϕ> in 2PC (0-5%) Correlation function well described by v2-v6 and v1,1

12

Most of v1,1 is due to momentum conservation

~1.5 : 1 ~3:1Most of v1,1 is due to dipolar flow

Fit for 40-50% centrality

Despite that the v1,1 is always negative, significant positive v1 can still exist.

13

Fit result vs pT and centrality

v1Fit(pT) peaks around 4-5 GeV, peak-value

increases with centrality by about 20%. Less viscosity damping, reflecting the

increase in ε1?

14

Glauber

Compare with v2 and v3

v1 comparable to v3 but peak at higher pT.

High pT v1 seems drop slower than v2,v3. Limitation of two-component assumption? Both L dependent eloss become important?

v1>v2 in jet absorption model calculation in central collisions.

15

1203.3265 L3

About momentum conservation component

The system that conserves momentum may only be a subset of the event

c dN/dη but decrease toward peripheral by 20-30%

For <pT2>=1 GeV2, M=5000 in 0-

5% events, about 3 units in rapidity

Increases for peripheral collisions, about ~4 unit for 40-50% centrality

16

?

M. Lisa 0807.3569

Comparison with AMPT model: arxiv:1203.3410

AMPT=HIJING +F.S scattering. Interaction strength controlled by αs and μ.

HIJING only need momentum conservation, while AMPT need both The complex pT dependence of v1,1 can naturally be generated from final state

interaction

17

v1,1 calculated for pairs with |Δη|>1.5

Arxiv1203.3410

Centrality and energy dependence

pT dependence is qualitatively similar to what is seen in data and hydro predictions

Weak dependence on centrality Increases from RHIC to LHC

18

Dependence on the strength of interaction

More sensitive to changing αs than changing screening mass μ Values from a larger screening mass and smaller coupling constant is closer to the data from

ATLAS: αs=0.33, σ=1.5mb

19

Summary The cos(Δϕ) component of the 2PC data suggests contributions from

rapidity-even dipolar flow and global momentum conservation. A two-component fit is used to extract the individual contribution from

these two components

Extracted v1 cross zero at pT~1 GeV, reaches a value of 0.1 (comparable to v3), and decreases at higher pT. The pT at which it reaches maximum is 1 GeV higher than other vn.

Extracted v1 shows a mild increase with centrality (~20%) The system conserving momentum only involves a subset of the event

AMPT transport model calculation confirmed qualitative trend at low pT. Dipolar flow is indeed associated with final state interaction Flow magnitude is sensitive to the strength of the interaction

20

Backup

21

Extracting the η dependence: 1203.3410

Extend the procedure to study rapidity dependence by using a simultaneous fit of the 4-D v1,1 data. Only v1,1 data satisfying a certain η gap is used (|Δη|>2)

The number of independent c values can be restricted by symmetry

Impose the constraint v1(η) = v1(−η)

22

η dependence of v1 and c from AMPT

Weak η dependence at RHIC energies but has a dip at mid-rapidity at LHC energy strong longitudinal flow?

c is not constant: contribution from momentum conservation is not constant across whole η and |∆η| range and shows a strong dependence on |∆η|

23

24

25

26