Flow and Centrality fluctuations in ATLAS Jiangyong Jia

Flow and Centrality fluctuations in ATLAS Jiangyong Jia for the ATLAS Collaboration Stony Brook University Based on ar. Xiv: 1904. 04808 QM 2019 11/3 -11/9/2019 1

Origin of flow/centrality fluctuations Many collisions Each with own evolution … … Event by event fluctuations: n Same N can have different shapes flow fluctuation n Same N can have different sizes centrality fluctuation 2

Observables for flow fluctuations n Cumulants for p(vn) n Mixed-harmonic cumulants for p(vn, vm). Not discussed here, see 1904. 04808 3 All results cross-checked with subevent method to ensure non-flow is negligible n Current paradigm: L. Yan, J. Ollitrault ar. Xiv: 1312. 6555 Fine splitting of vn{2 k} reflects the fine splitting of εn{2 k} between k=1, 2, 3…

v 2 and v 3 from 4 -particle cumulants n Clear p. T dependence in cumulant ratios n v 2{4}/v 2{2} decreases with p. T, v 3{4}/v 3{2} increases with p. T Role of initial-state sub-leading eccentricity or final state fluctuations? 4

v 2 from 6 -particle cumulants n Excellent agreement among experiments n but much better precision 5

v 2 from 6 -particle cumulants n Excellent agreement among experiments n n but much better precision Clear p. T dependence seen n n Increase in central region Decrease in peripheral region Role of initial-state sub-leading eccentricity or final state fluctuations? 6

Constraining the initial-state fluctuations v 2{6}/v 2{4} vs. v 2{4}/v 2{2} has weaker p. T dependence more related to IS geometry v 2{6}/v 2{4} n 7 5% Central 25% Peripheral 65% v 2{4}/v 2{2}

Constraining the initial-state fluctuations v 2{6}/v 2{4} vs. v 2{4}/v 2{2} has weaker p. T dependence more related to IS geometry n n 2 -component Glauber model agrees better with Pb+Pb data Glauber model fails in peripheral region, where data approach fluctuation-driven model CMS v 2{6}/v 2{4} n 8 5% Central 25% 65% Peripheral Fluctuation-driven model describes p. Pb v 2{4}/v 2{2}

v 4 from 4 -particle cumulants n c 4{4} changes sign around 25% centrality, also depends on p. T. n v 4 has non-linear contribution from v 2: , but not enough quantitatively ar ine n-l No G. Giacalone, L. Yan, J. Noronha-Hostler, J. Ollitrault, 1608. 06022 L a e n i r 9

v 4 from 4 -particle cumulants n c 4{4} changes sign around 25% centrality, also depends on p. T. n n 10 v 4 has non-linear contribution from v 2: , but not enough quantitatively v 4{4}/v 4{2}~0. 5 in central region, magnitude reach >1 in peripheral region v 4{4}/v 4{2} ar ine n-l No L a e n i r G. Giacalone, L. Yan, J. Noronha-Hostler, J. Ollitrault, 1608. 06022 Implies large non-Gaussian fluctuation in peripheral collisions!

Centrality in A+A collisions n 11 Many variables to quantify centrality/volume. n n At initial state: b, Npart , x. Npart+(1 -x) Ncoll, Nqp , … At final state: Nch, ET, Nneutron, … Initial state sources: V Final state particles: N Vb Experimental use: Nb |η|<1 Subevent B Measurement Va Na 3<|η|<5 η Subevent A Event selection

Cent 2 How to detect centrality fluctuation? If no fluctuation: Cent 1 <cent 2> in narrow slices of cent 1 <Cent 2> <Cent 1> <cent 1> in narrow slices of cent 2 Cent 1 12

Cent 2 How to detect centrality fluctuation? Now smears along cent 1 direction with fluctuation: <cent 2> increases more slowly with cent 1, due to poorer centrality resolution or more centrality fluctuation of cent 1. Cent 1 <cent 2> in narrow slices of cent 1 <Cent 2> <Cent 1> <cent 1> in narrow slices of cent 2 Cent 2 This relation remains linear Cent 1 Non-linearity expected in ultra-central region 13

5. 02 Te. V Pb+Pb |η|<2. 5 Observation of centrality resolution in the data 3<|η|<5 Cent in |η|<2. 5 has poorer centrality resolution than 3<|η|<5 14

|η|<2. 5 Impact centrality fluctuation for flow observables p 1(vn) p 2(vn) 3<|η|<5 Even if <ET> and <Nch> are same, p 1(vn) and p 2(vn) could still be different 15

Centrality fluctuation and v 2 -slope in ultra-central collisions |η|<2. 5 ar. Xiv: 1803. 01812 3<|η|<5 16

Centrality fluctuation and v 2 -slope in ultra-central collisions 17 at same Nch |η|<2. 5 ar. Xiv: 1803. 01812 3<|η|<5 Larger centrality fluctuation for Nch-bin(mid-η) than for ET-bin(forward-η) Significant centrality decorrelation along η

Effects of centrality fluctuation on higher-order cumulants ~0. 2 n Direct comparison: difference largest in UCC, persists to mid-central collisions nc 2{4, Nch} (ΣET) obtained by map Nch to <ΣET> n Sign change in UCC! CF influences cn{4} over a broad centrality range! ~0. 05 18

Effects of centrality fluctuation on higher-order cumulants 19 ~0. 2 ~0. 08 n Direct comparison: difference largest in UCC, persists to mid-central collisions nc 2{4, Nch} (ΣET) obtained by map Nch to <ΣET> n Sign change in UCC! CF influences cn{4} over a broad centrality range! Glauber study show multiplicity smearing change sign of eccentricity cumulants ~0. 05 ar. Xiv: 1803. 01812

Centrality/size fluctuation in central collision Modification of dynamic fluctuations in ultra-central collisions seen in several observables v 22 fluctuation v 22<p. T> correlation <p. T> fluctuation See Tomasz Bold Wed 10: 20 Provide a way to study the nature of centrality and particle production mechanism 20

Summary 21 n Space-time dynamics of HI collisions studied via flow fluctuations and centrality fluctuations n Flow cumulants show p. T dependence εn not the only source for flow fluctuations n Provide new constrains on the initial-state fluctuations. Intriguing sign change of c 4{4} in mid-central collisions: mixing between v 2 and v 4? n Flow fluctuations sensitive to centrality resolution or volume fluctuations n Reflected by sign-change of many cumulants observables in UCC Flow fluctuations can be used to elucidate nature of centrality & particle production mechanism

Mix-harmonics n n Anti-corr. between v 2&v 3 reflects anti-corr. between ε 2&ε 3 ; strong p. T dependence Correlation between v 2 & v 4 mode-mixing of ; weak p. T dependence 22

Centrality fluctuation on mix-harmonic cumulants n n Centrality fluctuation effects is large for v 2 -v 3 correlation CF effects much smaller for v 2 - v 4 correlation, still visible in UCC 23

Four-particle cumulant for v 1 n Rapidity-even v 1 from dipolar fluctuations v 1{2} changes sign at p. T~1. 2 Ge. V, v 1{4} measurable only at large p. T 24 1203. 3087

Four-particle cumulant for v 1 n Rapidity-even v 1 from dipolar fluctuations n c 1{4}<0 observed at high p. T. n v 1{4} increase from central to peripheral. n Ranges from 1 to 4% 25 1203. 3087 v 1{2} changes sign at p. T~1. 2 Ge. V, v 1{4} measurable only at large p. T v 1{4} c 1{4}