Two-Particle Azimuthal Correlations As a Probe of Collective Behaviour in Deep Inelastic Ep Scattering at HERA

Two-particle azimuthal correlations as a probe of collective behaviour in deep inelastic ep scattering at HERA

Dhevan Gangadharan On behalf of the ZEUS Collaboration

Universit¨atHeidelberg

DIS 2021, 14 April 2021

1/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 1 / 29 High-energy heavy-ion collisions: initial state

target Initial state: elliptical in X-Y plane Most spatial conﬁgurations in the initial state correspond to partially overlapping nuclei. Z axis Consequently, the initial scattering leaves behind an elliptically eccentric zone in the transverse plane. X axis

projectile b

2/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 2 / 29 High-energy heavy-ion collisions: ﬁnal state

Subsequently, a prominent stage of rescattering is expected that rapidly leads to a local thermal equilibrium. X Hydrodynamics of a QCD ﬂuid is used Rescattering (hydrodynamic stage) to describe the evolution of this matter. P Y This is a non-perturbative process that converts the initial-state spatial eccentricity into a ﬁnal-state momentum anisotropy.

3/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 3 / 29 Observations in heavy-ion collisions

Pb+Pb centrality

Two particle correlations show a clear double ridge, which is interpretated as a sign of ﬂuid-like behaviour. The ﬂuid of QCD matter is referred to as a Quark–gluon plasma (QGP).

CMS EPJC (2012) 72:2012

4/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 4 / 29 Observations in high-multiplicity p + p collisions

p+p

The LHC revealed a similar double-ridge in high multiplicity p + p collisions. The ﬁnding came as a surprise since a p + p collision was thought to be too small to produce a thermally equilibrated QGP. This motivated the search for similar eﬀects in even smaller ep DIS at HERA. ATLAS PRL 116 172301

5/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 5 / 29 Deep inelastic scattering (DIS)

DIS Scattered electron DIS is deﬁned by large virtualities: Incoming 2 2 electron Q ΛQCD. Transverse radius (Rt ) and longitudinal length (L) of the probed region are given by: 1 Rt ∼ Q 1 L ∼ PRD 95 114008 mproton x Neutral current (NC) DIS involves exchange of photon or Z boson.

6/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 6 / 29 Photoproduction (PhP)

Photoproduction PhP is deﬁned by small virtualities: 2 2 Q ΛQCD. Exchange photon may ﬂuctuate into quarks and gluons. Larger interaction regions are probed. Scattering is hadron-like. PhP is not presented here. A complementary ZEUS analysis in photoproduction on this subject will be published soon.

7/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 7 / 29 Two-particle correlation function cn{2}

Two-particle azimuthal correlations are Transverse Plane measured: Directed Elliptic asymmetry asymmetry

cn{2} = hhcos n(φi − φj )ii.

ϕi is the azimuthal angle of particle i. n is the harmonic.

The inner and outer brackets denote an Triangular Quadrangular average within an event and over all events, asymmetry asymmetry respectively.

Detector acceptance corrections described in the backup.

8/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 8 / 29 The HERA collider and experiments

Location: DESY, Hamburg, Germany Data taking: 1992 - 2007 27.6 GeV electrons/positrons 920 GeV protons √ → s = 318 GeV HERA I+II: 500 pb−1 per experiment

We present recently published measurements of two-particle correlations in NC DIS with ZEUS: JHEP 04 (2020) 070.

9/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 9 / 29 DIS event and track selection

Charged particles are tracked in the central tracking detector (CTD) and micro vertex detector (MVD) in a 1.43 T magnetic ﬁeld.

Depleted uranium calorimeters. The barrel and rear ones are used to help identify the scattered electron.

A fully contained event is characterized by P i (Ei − Pz,i ) = 55 GeV due to energy and momentum conservation.

10/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 10 / 29 DIS event and track selection

High Q2

ZEUS Event selection (46 M) FCAL BCAL RCAL DIS triggers Q2 = −(k − k0)2 > 5 GeV2 HAC HAC 0 k0 > 10 GeV r > 15 cm EMC L C o

M w θ > 1 rad

E e

Solenoid magnet Q P 2 47 < (Ei − Pz,i ) < 69 GeV CTD scattered electron |Vz | < 30 cm r MVD Track selection for correlation analysis electron proton Reject scattered electron −1.5 < η < 2.0

0.1 < pT < 5.0 GeV ≥ 1 MVD hit

DCAXY ,Z < 2 cm ∆R > 0.4 (cone around scattered electron)

11/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 11 / 29 ZEUS ZEUS NC DIS 366 pb•1 s = 318 GeV 2 2 {2} {2} Q > 5 GeV 1 LEPTO gen 2 c

c p < 5.0 GeV ARIADNE gen T •1.5 < η < 2.0 ARIADNE non•diff gen 0.2 p > 0.1 GeV 0.1 p > 0.1 GeV c1{2} &T c2{2} versus chargedT particle multiplicity Nch no |∆η| cut no |∆η| cut

0 0 0 10p > 0.5 GeV 20 30 0 10p > 0.5 GeV 20 30 0.1 T T {2} {2} 1 |∆η| > 2 2 |∆η| > 2 c1{2} is better described by the ARIADNE c c 0.05 generator. 0 c2{2} is better described by the LEPTO generator. −0.1 0 Neither model works well for both harmonics.

0 10p > 0.5 GeV 20 30 0 ARIADNE 10 non•diff gen 20 30 T 0.2 The diﬀractive component in ARIADNE only {2} {2}

1 ARIADNE non•diff gen |∆η| > 2 2 c c slightly inﬂuences c {2} . 0 p > 0.5 GeV 2 T |∆η| > 2 0.1 Massless jets were reconstructed from the generated hadrons with the kT algorithm and −0.2 ≤ 1 jet: LEPTO gen ≤ 1 jet: ARIADNE gen Et > 2 GeV, ∆R = 1. = 2 jets: LEPTO gen 0 = 2 jets: ARIADNE gen Jets can explain the observed correlations. 0 10 20 30 0 10 20 30 Nch Nch

12/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 12 / 29 ZEUS ZEUS NC DIS 366 pb•1 s = 318 GeV 2 2 {2} {2} Q > 5 GeV 1 LEPTO gen 2 c

c p < 5.0 GeV ARIADNE gen T •1.5 < η < 2.0 ARIADNE non•diff gen 0.2 p > 0.1 GeV 0.1 p > 0.1 GeV c1{2} &T c2{2} versus chargedT particle multiplicity Nch no |∆η| cut no |∆η| cut

0 0 Short-range (|∆η| ∼ 0) correlations are strongest 0 10p > 0.5 GeV 20 30 0 10p > 0.5 GeV 20 30 0.1 T T {2} {2} at low Nch 1 |∆η| > 2 2 |∆η| > 2 c c 0.05 Longe-range correlations 0 (|∆η| > 2, orange-black pairs) of the ﬁrst harmonic are negative 0 −0.1 and the largest in magnitude

0 10p > 0.5 GeV 20 30 0 ARIADNE 10 non•diff gen 20 30 T 0.2 {2} {2}

1 ARIADNE non•diff gen |∆η| > 2 2 c c 0 p > 0.5 GeV T |∆η| > 2 0.1

−0.2 ≤ 1 jet: LEPTO gen ≤ 1 jet: ARIADNE gen = 2 jets: LEPTO gen 0 = 2 jets: ARIADNE gen 0 10 20 30 0 10 20 30 Nch Nch

13/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 13 / 29 Ridge ﬁgures in DIS

0.5 < p < 5.0 GeV ZEUS T

s = 318 GeV 2 ≤ Nch < 10 15 ≤ Nch < 30 Q2 > 5 GeV2 ) )

ϕ ϕ 1.2 ∆ ∆

, 1 , η η ∆ ∆

C( C( 1 0.8

0.8 0.6

4 4 ∆ 3 ∆ 3 ϕ 3 2 ϕ 3 2 2 1 η 2 1 η 1 0 ∆ 1 0 ∆ 0 −1 0 −1 −1 −3−2 −1 −3−2 Jet peak centered at ∆ϕ ∼ ∆η ∼ 0. Away-side ridge at high Nch is expected from tilted dijets. No double ridge visible at high Nch

14/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 14 / 29 Ridge comparisons between the LHC and HERA DIS

Pb+Pb p+Pb centrality

CMS CMS EPJC (2012) 72:2012 PLB 718 (2013) 795

p+p e+p DIS

0.5 < p < 5.0 GeV ZEUS T

s = 318 GeV 2 ≤ Nch < 10 15 ≤ Nch < 30 Q2 > 5 GeV2 ) )

ϕ ϕ 1.2 ∆ ∆

, 1 , η η ∆ ∆

C( C( 1 0.8

0.8 0.6

4 4 ATLAS ∆ 3 ∆ 3 ϕ 3 2 ϕ 3 2 PRL 116 172301 2 1 η 2 1 η 1 0 ∆ 1 0 ∆ −1 −1 0 −2 0 −2 −1 −3 ZEUS −1 −3 JHEP 04 (2020) 070 15/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 15 / 29 Summary and outlook

Two-particle azimuthal correlations have been measured in ep neutral current deep inelastic scattering with ZEUS. Comparisons of the observed correlations with available models of DIS suggest that the measured correlations are dominated by contributions from jets. The correlations do not indicate the kind of collective behaviour recently observed in high-multiplicity hadronic collisions at the highest RHIC and LHC energies. Soon to be published results in photoproduction will shed more light on multi-particle production mechanisms and how they evolve from DIS to hadronic collisions.

16/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 16 / 29 Backup

17/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 17 / 29 Correcting for detector eﬀects

Applied weights from Monte Carlo simulations:

wi : Single-particle weights (1/eﬃciency) as a function of charge, pT , ϕ, and η.

w∆ϕ: Two-particle weights formed from the ratio of the number of generated to reconstructed pairs as a function of ∆ϕ, |∆η|, Nch, and relative charge.

Corrected event multiplicity is a weighted sum over all reconstructed tracks Nrec: N Prec Nch = wi i Measured two-particle correlation function: N "N # N "N # eventsP Prec eventsP Prec cn{2} = wi wj w∆ϕ cos n(ϕi − ϕj ) wi wj w∆ϕ e i6=j e i6=j e e 18/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 18 / 29 Systematic uncertainties

Considered sources of systematic uncertainties: Monte Carlo closure test (dominant source) Secondary contamination DIS event selection Eﬃciency corrections

Monte Carlo test uncertainty shown with boxes. The other systematic uncertainties are added bin-by-bin in quadrature and shown with thin capped lines.

19/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 19 / 29 cn{2} versus charged particle multiplicity Nch

ZEUS Short-range (|∆η| ∼ 0) correlations are strongest s = 318 GeV ZEUS NC DIS 366 pb•1 {2} {2}

1 2 2 Q > 5 GeV 2 p > 0.1 GeV at low Nch c

c T p < 5.0 GeV p > 0.1 GeV, | η∆ | > 2 T T •1.5 < η < 2.0 p > 0.5 GeV, | η∆ | > 2 0.2 0.1 T Longe-range correlations (|∆η| > 2, orange-black pairs) of the ﬁrst harmonic are negative 0 and the largest in magnitude 0

0 10 20 30 0 10 20 30 {2} {2} 3 4 c c 0.05 0.02

0 0

0 10 20 30 0 10 20 30 Nch Nch 20/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 20 / 29 cn{2} versus |∆η|

ZEUS The correlations with pT > 0.5 GeV (red) are ZEUS NC DIS 366 pb•1 s = 318 GeV {2} {2}

1 2 2 2 Q > 5 GeV more pronounced than those at low pT (blue) as c p > 0.1 GeV c 0.2 T p < 5.0 GeV p > 0.5 GeV T expected from particles in jet-like structures T 0.2 •1.5 < η < 2.0

15 ≤ Nch < 30 Negative (positive) c1{2} (c2{2} ) for pT > 0.5

0 GeV extend out to |∆η| ∼ 3 Large directed and elliptic anisotropy 0 → tilted dijet 0 1 2 3 0 1 2 3 {2} {2} 3 4 c c

0.1 0.05

0 0

0 1 2 3 0 1 2 3 |∆η| |∆η| 21/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 21 / 29 DIS event distributions

ZEUS ) s = 318 GeV 2 107 107 Q2 > 5 GeV2 Uncorrected multiplicity, Nrec , and rec, gen 0.1 < p < 5.0 GeV 2 6 T Q distributions. 10 •1.5 < η < 2.0 (1/GeV 2

dN/dN 105 ARIADNE MC simulations with ZEUS detector response gives a fair dN/dQ 104 ZEUS NC DIS 366 pb•1 ARIADNE gen description of the measured ZEUS ARIADNE non•diff gen 103 LEPTO gen data. ARIADNE rec hN i = 5 1.5 1.2 rec 1 ratio 1 ratio 2 2 ZEUS / ARIADNE rec Q = 30 GeV 0.5 0.8 0 10 20 30 2 10 10 2 2 Nrec, gen Q (GeV )

22/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 22 / 29 DIS track distributions

ZEUS Uncorrected pT and η track

10 •1 η ZEUS NC DIS 366 pb distributions. ARIADNE gen ARIADNE non•diff gen dN/d

LEPTO gen ev Reconstructed ARIADNE is (1/GeV) 2 T 1 ARIADNE rec

1/N compatible with the data to within about 10%. dN/dp −1 ev 10 s = 318 GeV 1 Q2 > 5 GeV2 Most of the proton fragments lie 1/N 0.1 < p < 5.0 GeV T outside of the ZEUS acceptance •1.5 < η < 2.0 10−2 0 near η = 4.

1.1 1

ratio ZEUS / ARIADNE rec ratio The proton remains intact for the 1 0.95 0 1 2 3 4 0 5 diﬀractive events as seen with the p (GeV) η T small peak in ARIADNE near η = 8.

23/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 23 / 29 c1{2} and c2{2} versus hpT i

ZEUS Correlations at low Nch were down-scaled by

0 s = 318 GeV hNch i / hNch i .

{2} {2} low high 1 2 2 2 c c Q > 5 GeV Scaling factor inspired by observations in 0.1 < p < 5.0 GeV T heavy-ion collisions where non-collective behaviour 0.05 •1.5 < η < 2.0 contributes to c2{2} as 1/Nch . |∆η| > 2 |∆η| > 2 The observed excess correlation at Nch wrt low −0.2 ZEUS NC DIS 366 pb•1 Nch is stronger for c1{2} and c2{2} . 15 ≤ Nch < 30 5 ≤ N < 10 (down•scaled) 0 ch Therefore, the 1/Nch scaling of non-collective 0 1 2 0 1 2 correlations may not be appropriate for ep 〈p 〉 (GeV) 〈p 〉 (GeV) T T scattering.

24/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 24 / 29 Model comparisons of c1{2} versus |∆η| and hpT i

ZEUS

ZEUS NC DIS 366 pb•1

{2} {2} 0.1

1 LEPTO gen 1

c ARIADNE gen c no |∆η| cut 0.1 ARIADNE non•diff gen

0.05 p > 0.1 GeV T 0

0 ARIADNE predictions describe the data 0 0 1 2s = 318 GeV 3 0 1 2 {2} {2} 1 Q2 > 5 GeV2 1 c c for the 1st harmonic reasonably well. 0.2 p < 5.0 GeV T •1.5 < η < 2.0

15 ≤ Nch < 30 |∆η| > 2 p > 0.5 GeV T 0 −0.2

0 1 2 3 0 1 2 〈p 〉 (GeV) |∆η| T 25/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 25 / 29 Model comparisons of c2{2} versus |∆η| and hpT i

ZEUS

ZEUS NC DIS 366 pb•1

{2} LEPTO gen {2} 2 2

c 0.1 ARIADNE gen c no |∆η| cut ARIADNE non•diff gen

0.2 0.05 p > 0.1 GeV T

0 0 LEPTO predictions describe the data for 0 1 2s = 318 GeV 3 0 1 2 {2} {2}

2 Q2 > 5 GeV2 2

c c the 2nd harmonic reasonably well. p < 5.0 GeV T 0.2 •1.5 < η < 2.0 0.1 15 ≤ Nch < 30 |∆η| > 2 p > 0.5 GeV T

0 0 0 1 2 3 0 1 2 〈p 〉 (GeV) |∆η| T 26/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 26 / 29 Raw c1{2} and c2{2} versus Nrec compared to models

ZEUS

•1 0.2 0.4 ZEUS NC DIS 366 pb s = 318 GeV {2} {2}

1 LEPTO rec 2 2 2 Q > 5 GeV c ARIADNE rec c p < 5.0 GeV ARIADNE non•diff rec T •1.5 < η < 2.0 rec rec 0.2 p > 0.1 GeV p > 0.1 GeV T T no |∆η| cut 0.1 no |∆η| cut LEPTO/ARIADNE provides a reasonable 0 description of the reconstructed data.

0.10 10p 20 > 0.5 GeV 30 0 10p 20 > 0.5 GeV 30 {2} T {2} T Utilization of tracking eﬃciency 1 2 c |∆η| > 2 c |∆η| > 2 corrections from such Monte Carlo data is 0.05 rec 0 rec reasonably justiﬁed.

0 −0.1

0 10 20 30 0 10 20 30 Nrec Nrec 27/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 27 / 29 Raw c1{2} versus |∆η| and hpT i compared to models

ZEUS

ZEUS NC DIS 366 pb•1

{2} {2} 0.1

1 LEPTO rec 1 c c 0.1 ARIADNE rec no |∆η| cut ARIADNE non•diff rec rec rec

0.05 p > 0.1 GeV T 0 ARIADNE provides a reasonable

0 description of the reconstructed data. 0 0 1 2s = 318 GeV 3 0 1 2 {2} {2} Utilization of tracking eﬃciency 1 Q2 > 5 GeV2 1 c c 0.2 p < 5.0 GeV T corrections from such Monte Carlo data is •1.5 < η < 2.0 rec rec 15 ≤ Nrec < 30 reasonably justiﬁed. |∆η| > 2 p > 0.5 GeV T 0 −0.2

0 1 2 3 0 1 2 〈p 〉 (GeV) |∆η| T 28/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 28 / 29 Raw c2{2} versus |∆η and hpT i compared to models

ZEUS

ZEUS NC DIS 366 pb•1 {2} {2}

2 LEPTO rec 2

c 0.1 ARIADNE rec c no |∆η| cut ARIADNE non•diff rec rec rec 0.2 0.05 p > 0.1 GeV T LEPTO provides a reasonable description of the reconstructed data. 0 0

0 1 2s = 318 GeV 3 0 1 2 {2} {2} Utilization of tracking eﬃciency 2 Q2 > 5 GeV2 2 c c p < 5.0 GeV T corrections from such Monte Carlo data is •1.5 < η < 2.0 rec 0.2 rec 0.1 15 ≤ Nrec < 30 reasonably justiﬁed. |∆η| > 2 p > 0.5 GeV T

0 0 0 1 2 3 0 1 2 〈p 〉 (GeV) |∆η| T 29/29 Dhevan Gangadharan On behalf of the ZEUS Collaboration (Uni Heidelberg) DIS 2021, 14 April 2021 29 / 29