Long-range two-particle correlations of strange hadrons with charged particles in pPb and PbPb collisions at LHC energies

Measurements of two-particle angular correlations between an identified strange hadron (K0S or Lambda/anti-Lambda) and a charged particle, emitted in pPb collisions, are presented over a wide range in pseudorapidity and full azimuth. The data, corresponding to an integrated luminosity of approximately 35 inverse nanobarns, were collected at a nucleon-nucleon center-of-mass energy (sqrt(s[NN])) of 5.02 TeV with the CMS detector at the LHC. The results are compared to semi-peripheral PbPb collision data at sqrt(s[NN]) = 2.76 TeV, covering similar charged-particle multiplicities in the events. The observed azimuthal correlations at large relative pseudorapidity are used to extract the second-order (v[2]) and third-order (v[3]) anisotropy harmonics of K0S and Lambda/anti-Lambda particles. These quantities are studied as a function of the charged-particle multiplicity in the event and the transverse momentum of the particles. For high-multiplicity pPb events, a clear particle species dependence of v[2] and v[3] is observed. For pt<2 GeV, the v[2] and v[3] values of K0S particles are larger than those of Lambda/anti-Lambda particles at the same pt. This splitting effect between two particle species is found to be stronger in pPb than in PbPb collisions in the same multiplicity range. When divided by the number of constituent quarks and compared at the same transverse kinetic energy per quark, both v[2] and v[3] for K0S particles are observed to be consistent with those for Lambda/anti-Lambda particles at the 10% level in pPb collisions. This consistency extends over a wide range of particle transverse kinetic energy and event multiplicities.


Introduction
Studies of multiparticle correlations provide an important insight into the underlying mechanism of particle production in high-energy collisions of protons and nuclei. A key feature of such correlations in ultrarelativistic nucleus-nucleus (AA) collisions is the observation of a pronounced structure on the near side (relative azimuthal angle | φ| ≈ 0) that extends over a large range in relative pseudorapidity (| η| up to 4 units or more). This feature, known as the "ridge", has been found over a wide range of AA energies and system sizes at both the Relativistic Heavy Ion Collider (RHIC) [1-5] and the Large Hadron Collider (LHC) [6-10] and is interpreted as arising primarily from the collective hydrodynamic flow of a strongly interacting, expanding medium [11,12].
Similar long-range correlations have also been discovered in proton-proton (pp) [13], proton-lead (pPb) [14][15][16], and deuterongold (dAu) [17] collisions with high final-state particle multiplicity. As the collision volume size is reduced, it is possible that the E-mail address: cms-publication-committee-chair@cern.ch. system will not be able to equilibrate and the hydrodynamic description will break down. As such, there has been no consensus on the origin of the particle correlation structure in these smaller systems. A variety of theoretical models have been proposed to interpret this phenomenon in pp [18], pPb, and dAu collisions. Besides hydrodynamic effects in a high-density system [19,20], an alternate model including gluon saturation in the incoming nucleons has also been shown to describe these data [21,22].
In hydrodynamical descriptions, the collective flow manifests itself as an azimuthal anisotropy in the distribution of final-state particles. An additional key consequence of these models is that the measured anisotropies will depend on the mass of the particle [23][24][25]. More specifically, for particles with transverse momentum below about 2 GeV, the anisotropy will be larger for lighter particles. The presence of this mass ordering is well established in AA collisions at RHIC and LHC energies [26][27][28][29][30]. This phenomenon has recently also been observed in pPb [31] and dAu [17] collisions, consistent with expectations from hydrodynamic models [32,33]. The analysis presented in this paper aims to further explore this effect by extracting anisotropies of identified strange mesons (K 0 S ) and baryons (Λ and Λ) in pPb and in PbPb collisions that produce similar final-state particle multiplicity.
The azimuthal correlations of emitted particle pairs are typically characterized by their Fourier components, dN pair d φ ∝ 1 + n 2V n cos(n φ), where V n are the two-particle Fourier coefficients and v n = √ V n denote the single-particle anisotropy harmonics [34]. In particular, the second and third Fourier components are known as elliptic (v 2 ) and triangular (v 3 ) flow, respectively [12]. In hydrodynamical models, v 2 and v 3 are directly related to the response of the medium to the initial collision geometry and its fluctuations [35][36][37]. As such, these Fourier components can provide insight into the fundamental transport properties of the medium.
In AA collisions at RHIC, a scaling of v 2 as a function of p T with the number of constituent quarks (n q ) has been observed in the range 2 < p T < 6 GeV [38]. Specifically, the values of v 2 /n q are found to be very similar for all mesons (n q = 2) and baryons (n q = 3) when compared at the same value of p T /n q . This empirical scaling may indicate that final-state hadrons are formed through recombination of quarks in this p T regime [39][40][41], possibly providing evidence of deconfinement of quarks and gluons in these systems. At lower p T (p T < 2 GeV), a similar scaling behavior is observed, although, according to perfect fluid hydrodynamics, v 2 /n q values must be compared at the same transverse kinetic energy per constituent quark (KE T /n q , where KE T = m 2 + p 2 T − m) to account for the mass difference of hadrons [42,43]. This paper presents an analysis of two-particle correlations with identified strange hadrons, K 0 S and Λ/Λ, in pPb collisions at a center-of-mass energy per nucleon pair ( √ s N N ) of 5.02 TeV. With the implementation of a dedicated high-multiplicity trigger, the 2013 pPb data sample gives access to multiplicities comparable to those in semi-peripheral PbPb collisions. Two-particle correlation functions are constructed by associating a K 0 S or Λ/Λ particle with a charged particle (pairs of K 0 S or Λ/Λ particles are not studied due to their limited statistical precision). In the context of hydrodynamic models, Fourier coefficients of dihadron correlations can be factorized into products of single-particle azimuthal anisotropies. Assuming that this relationship holds, v 2 and v 3 are extracted from long-range two-particle correlations as a function of strange hadron p T and event multiplicity. To examine the validity of constituent quark number scaling, v 2 /n q and v 3 /n q are obtained as a function of KE T /n q for both K 0 S and Λ/Λ particles. A direct comparison of the pPb and PbPb results over a broad range of similar multiplicities is presented.

The CMS experiment and data sample
A description of the CMS detector in the LHC at CERN can be found in Ref. [44]. The main detector component used in this paper is the tracker, located in a superconducting solenoid of 6 m internal diameter, providing a magnetic field of 3.8 T. The tracker consists of 1440 silicon pixel and 15 148 silicon strip detector modules, covering the pseudorapidity range |η| < 2.5. For hadrons with p T ≈ 1 GeV and |η| ≈ 0, the impact parameter (distance of closest approach from the primary collision vertex) resolution is approximately 100 μm and the p T resolution is 0.8%.
Also located inside the solenoid are the electromagnetic calorimeter (ECAL) and the hadron calorimeter (HCAL). The ECAL consists of 75 848 lead tungstate crystals, arranged in a quasiprojective geometry and distributed in a barrel region (|η| < 1.48) and two endcaps that extend to |η| = 3.0. The HCAL barrel and endcaps are sampling calorimeters composed of brass and scintillator plates, covering |η| < 3.0. Iron/quartz-fiber forward calorimeters (HF) are placed on each side of the interaction region, covering 2.9 < |η| < 5.2. The detailed Monte Carlo (MC) simulation of the CMS detector response is based on geant4 [45].
The data sample used in this analysis was collected with the CMS detector during the LHC pPb run in 2013. The total integrated luminosity of the data set is about 35 nb −1 [46]. The beam energies are 4 TeV for protons and 1.58 TeV per nucleon for lead nuclei, resulting in a center-of-mass energy per nucleon pair of 5.02 TeV. The direction of the proton beam was initially set up to be clockwise (20 nb −1 ), and was later reversed (15 nb −1 ). As a result of the energy difference between the colliding beams, the nucleon-nucleon center-of-mass in the pPb collisions is not at rest with respect to the laboratory frame. Massless particles emitted at η cm = 0 in the nucleon-nucleon center-of-mass frame will be detected at η = −0.465 (clockwise proton beam) or 0.465 (counterclockwise proton beam) in the laboratory frame. A sample of peripheral PbPb data at √ s N N = 2.76 TeV corresponding to an integrated luminosity of about 2.3 μb −1 , collected during the 2011 LHC heavy-ion run, is also analyzed for comparison with pPb data at similar charged-particle multiplicity ranges.

Online triggering and offline track reconstruction and selection
The online triggering and the offline reconstruction and selection follow the same procedure as described in Ref. [47]. Minimum bias pPb events are triggered by requiring at least one track with p T > 0.4 GeV to be found in the pixel tracker for a pPb bunch crossing. Because of hardware limits on the data acquisition rate, only a small fraction (∼ 10 −3 ) of all minimum bias triggered events are recorded. In order to collect a large sample of highmultiplicity pPb collisions, a dedicated high-multiplicity trigger is also implemented using the CMS Level 1 (L1) and high-level trigger (HLT) systems. At L1, two event streams were triggered by requiring the total transverse energy summed over ECAL and HCAL to be greater than 20 or 40 GeV. Charged tracks are then reconstructed online at the HLT using the three layers of pixel detectors, and requiring a track origin within a cylindrical region of 30 cm length along the beam and 0.2 cm radius perpendicular to the beam. For each event, the number of pixel tracks (N online trk ) with |η| < 2.4 and p T > 0.4 GeV is counted separately for each vertex. Only tracks with a distance of closest approach of 0.4 cm or less to one of the vertices are included. The online selection requires N online trk for the vertex with the most tracks to exceed a specific value. Data are taken with thresholds of N online trk > 100, 130 (from events with L1 threshold of 20 GeV), and 160, 190 (from events with L1 threshold of 40 GeV). While all events with N online trk > 190 are accepted, only a fraction of the events from the other thresholds are kept. This fraction is dependent on the instantaneous luminosity. Data from both the minimum bias trigger and high-multiplicity trigger are retained for offline analysis.
In the offline analysis, hadronic collisions are selected by the presence of at least one tower with energy above 3 GeV in each of the two HF calorimeters. Events are also required to contain at least one reconstructed primary vertex within 15 cm of the nominal interaction point along the beam axis and within 0.15 cm transverse to the beam trajectory. At least two reconstructed tracks are required to be associated with the primary vertex, a condition that is important only for minimum bias events. Beam related background is suppressed by rejecting events for which less than 25% of all reconstructed tracks pass the high-purity selection (as defined in Ref. [48]). The pPb instantaneous luminosity provided by the LHC in the 2013 run resulted in a 3% probability of having at least one additional interaction present in the same bunch crossing (pile-up events). The procedure used for rejecting pile-up events is described in Ref. [47] and is based on the number of tracks as-sociated with each reconstructed vertex and the distance between different vertices. A purity of 99.8% for single pPb collision events is achieved for the highest multiplicity pPb interactions studied in this paper. With the selection criteria above, 97-98% of the events are found to be selected among those pPb interactions simulated with the epos lhc [49] and hijing 2.1 [50] event generators that have at least one particle from the pPb interaction with energy E > 3 GeV in each of the η ranges −5 < η < −3 and 3 < η < 5.
In this analysis, high-purity tracks are used to select primary tracks (tracks originating from the pPb interaction). Additional requirements are applied to enhance the purity of primary tracks. The significance of the separation along the beam axis (z) between the track and the best vertex, d z /σ (d z ), and the significance of the impact parameter relative to the best vertex transverse to the beam, d T /σ (d T ), must be less than 3, and the relative p T uncertainty, σ (p T )/p T , must be less than 10%. To ensure high tracking efficiency and to reduce the rate of misreconstructed tracks, primary tracks with |η| < 2.4 and p T > 0.3 GeV are used in the analysis (a p T cutoff of 0.4 GeV is used in the multiplicity determination to match the HLT requirement). Based on simulation studies using geant4 to propagate particles from the hijing event generator, the combined geometrical acceptance and efficiency for primary track reconstruction exceeds 60% for p T ≈ 0.3 GeV and |η| < 2.4. The efficiency is greater than 90% in the |η| < 1 region for p T > 0.6 GeV.
For the event multiplicity range studied in this paper, no dependence of the tracking efficiency on multiplicity is found and the rate of misreconstructed tracks is 1-2%.
The entire pPb data set is divided into classes based on the reconstructed track multiplicity, N offline trk , where primary tracks with |η| < 2.4 and p T > 0.4 GeV are counted. Details of the multiplicity classification in this analysis, including the fraction of the full multiplicity distribution and the average number of primary tracks before and after correcting for detector effects in each multiplicity range, are provided in Ref. [47].
A subset of semi-peripheral PbPb data collected during the 2011 LHC heavy-ion run with a minimum bias trigger are also reanalyzed in order to directly compare pPb and PbPb systems at the same collision multiplicity. The reanalyzed events were in the range of 50-100% centrality, where centrality is defined as the fraction of the total inelastic cross section, with 0% denoting the most central collisions. This sample was reprocessed using the same event selection and track reconstruction algorithm as for the present pPb analysis. A description of the 2011 PbPb data can be found in Refs. [47,51].

Reconstruction of K 0 S and Λ/Λ candidates
The reconstruction technique for K 0 S and Λ/Λ candidates (generally referred to as V 0 s) at CMS was first described in Ref. [52]. To increase the efficiency for tracks with low momentum and large impact parameters, both characteristic of the K 0 S and Λ/Λ decay products, the standard loose selection of tracks (as defined in Ref. [48]) is used in reconstructing the K 0 S and Λ/Λ candidates.
Oppositely charged tracks with at least 4 hits and both transverse and longitudinal impact parameter significances greater than 1 (with respect to the primary vertex) are first selected to form a secondary vertex. The distance of closest approach of the pair of tracks is required to be less than 0.5 cm. The fitted vertex in x, y, z of each pair of tracks is required to have a χ 2 value normalized by the number of degrees of freedom less than 7. The pair of tracks is assumed to be π + π − in K 0 S reconstruction, while the assumption of π − p (π + p) is used in Λ (Λ) reconstruction. For Λ/Λ, the lower-momentum track is assumed to be the pion.
Due to the long lifetime of K 0 S and Λ/Λ particles, a requirement on the significance of the V 0 decay length, which is the three-dimensional distance between the primary and V 0 vertices divided by its uncertainty, to be greater than 5 is applied to reduce background contributions. To remove K 0 S candidates misidentified as Λ/Λ particles and vice versa, the Λ/Λ (K 0 S ) candidates must have a corresponding π + π − (pπ − ) mass more than 20 (10) MeV away from the PDG value of the K 0 S (Λ) mass [53]. The angle θ point between the V 0 momentum vector and the vector connecting the primary and V 0 vertices is required to satisfy cos θ point > 0.999.
This reduces the effect of nuclear interactions, random combinations of tracks, and Λ/Λ particles originating from weak decays of Ξ and Ω − particles. From MC simulations using Geant4 and the hijing event generator, it is found that the contribution of Λ/Λ particles from weak decays is less than 3% after this requirement. The K 0 S (Λ/Λ) reconstruction efficiency is about 6% (1%) for p T ≈ 1 GeV and 20% (10%) for p T > 3 GeV within |η| < 2.4.
This efficiency includes the effects of acceptance and the branching ratio for V 0 particle decays into neutral particles. The relatively low reconstruction efficiency of the V 0 candidates is primarily due to the decay length cut and the low efficiency for reconstructing daughter tracks with p T < 0.3 GeV or large impact parameters.
Examples of invariant mass distributions of reconstructed K 0 S and Λ/Λ candidates are shown in Fig. 1 for pPb data, with V 0 p T in the range of 1-3 GeV and event multiplicity in the range 220 ≤ N offline trk < 260. Since the results for Λ and Λ are found to be consistent, they have been combined in this analysis. The V 0 peaks can be clearly identified with little background. The true V 0 signal peak is well described by a double Gaussian function (with a common mean), while the background is modeled by a 4th-order polynomial function fit over the entire mass range shown in Fig. 1.
The mass window of ±2σ wide around the center of the peak is defined as the "peak region", where σ represents the root mean square of the two standard deviations of the double Gaussian functions weighted by the yields (with typical value of σ indicated in Fig. 1). To estimate the contribution of background candidates in the peak region to the correlation measurement, a "sideband region" is chosen that includes V 0 candidates from outside the ±3σ mass range around the V 0 mass to the limit of the mass distributions shown in Fig. 1.

Analysis of two-particle correlations
The construction of the two-particle correlation function follows the same procedure established in Refs. [6,7,14,47]. However, in this paper, reconstructed V 0 candidates from either the peak or sideband region are taken as "trigger" particles within a given p trig T range, instead of charged tracks as used in previous publications. The number of trigger V 0 candidates in the event is denoted by N trig . Particle pairs are formed by associating each trigger particle with the remaining charged primary tracks in a specified p assoc T interval (which can be either the same as or different from the p trig T range). The two-dimensional (2D) correlation function is defined in the same way as in previous analyses as 1 N trig where η and φ are the differences in η and φ of the pair. The same-event pair distribution, S( η, φ), represents the yield of particle pairs normalized by N trig from the same event, (2) The mixed-event pair distribution, is constructed by pairing the trigger V 0 candidates in each event with the associated charged primary tracks from 20 different randomly selected events in the same 2 cm wide range of vertex position in the z direction and from the same track multiplicity class.
Here, N mix denotes the number of pairs taken from the mixed events. The ratio B(0, 0)/B( η, φ) mainly accounts for the pair acceptance effects, with B(0, 0) representing the mixed-event associated yield for both particles of the pair going in approximately the same direction and thus having maximum pair acceptance (with a bin width of 0.3 in η and π/16 in φ). Thus, the quantity in Eq. (1) is effectively the per-trigger-particle associated yield. A pair is removed if the associated particle belongs to a daughter track of any trigger V 0 candidate (this contribution is negligible since associated particles are mostly primary tracks). The same-event and mixed-event pair distributions are first calculated for each event, and then averaged over all the events within the track multiplicity class. The range of 0 < | η| < 4.8 and 0 < | φ| < π is used to fill one quadrant of the ( η, φ) histograms, with the other three quadrants filled (for illustration purposes) by reflection to cover a ( η, φ) range of −4.8 < η < 4.8 and −π /2 < φ < 3π /2 for the 2D correlation functions, as will be shown later in Fig. 2. In performing the correlation analyses, each reconstructed primary track and V 0 candidate is weighted by a correction factor, following the procedure described in Refs. [6,7,14,47]. This correction is also applied in calculating N trig . This factor accounts for detector effects including the reconstruction efficiency, the detector acceptance, and the fraction of misreconstructed tracks. This correction factor is found to have a negligible effect on the azimuthal anisotropy harmonics.

Extraction of v n harmonics
Motivated by hydrodynamic models of long-range correlations in pPb collisions, azimuthal anisotropy harmonics of K 0 S and Λ/Λ particles are extracted via a Fourier decomposition of φ correlation functions averaged over | η| > 2 (to remove short-range correlations such as jet fragmentation), If the observed two-particle azimuthal correlations arise purely as the result of convoluting anisotropic distributions of single particles, then the V n coefficients can be factorized into the product of single-particle anisotropies [47], Following this assumption, the elliptic (v 2 ) and triangular (v 3 ) anisotropy harmonics of V 0 particles can be extracted as a function of p T from the fitted Fourier coefficients, , n = 2, 3.
Here, a fixed p ref T range for the "reference" charged primary particles is chosen to be 0.3 < p T < 3.0 GeV (the lowest p T region accessible by CMS and the same as was used in Ref. [47]), to minimize correlations from back-to-back jets at higher p T .
The v n values are first extracted for V 0 candidates from the peak region (which contains small contributions from background V 0 s) and sideband region, denoted as v obs n and v bkg n , respectively.
The v n signal of true V 0 particles is denoted by v sig n and is ob- assuming v sig n and v bkg n are independent from each other. Here, f sig represents the signal yield fraction in the peak region determined by the fits to the mass distribution shown in Fig. 1. This fraction exceeds 80% for Λ/Λ candidates at p T > 1 GeV and is above 95% for K 0 S candidates over the entire p T range.  The range of the V 0 mass distributions used in fitting the signal plus background (Fig. 1) is varied by 10%. This change, which could affect the value of f sig used in Eq. (7), yields a systematic uncertainty of less than 1% for the v As misalignment of the tracker detector elements can affect the V 0 reconstruction performance, an alternative detector geometry is studied. Compared to the standard configuration, this alternative has the two halves of the barrel pixel detector shifted in opposite directions along the beam by a distance on the order of 100 μm.

Systematic uncertainties
The values of v sig 2 found using the shifted configuration differed by less than 2% from the default ones.
To test the procedure of extracting the V 0 signal v 2 from Eq. (7), a study using epos lhc pPb MC events is performed to compare the extracted v sig 2 results with the generator-level K 0 S and Λ/Λ values. The agreement is found to be better than 4%. Other systematic uncertainties introduced by the high-multiplicity trigger efficiency (1%) and possible residual pile-up effects (1-2%) for pPb data are estimated in the same way as in Ref.
[47], and found to make only a small contribution. The various sources of systematic uncertainties are added together in quadrature to arrive at the final systematic uncertainties (6.9% for pPb and 6.6% for PbPb), which are shown as shaded boxes in Figs. 4-7.

Results
The 2D two-particle correlation functions measured in pPb collisions for pairs of a K 0 S (left) and Λ/Λ (right) trigger particles and a charged associated particle (h ± ) are shown in Fig. 2 in the p T range of 1-3 GeV. The 2D correlation functions are corrected for the background V 0 candidates, following the same approach of correcting v n in Eq. (7). The correction is negligible in this p T range because of the high signal yield fraction of V 0 candidates. For low-multiplicity events (N offline trk < 35, Fig. 2(a) and (b)), a sharp peak near ( η, φ) = (0, 0) due to jet fragmentation  (truncated for better illustration of the full correlation structure) can be clearly observed for both K 0 S -h ± and Λ/Λ-h ± correlations. Moving to high-multiplicity events (220 ≤ N offline trk < 260, Fig. 2(c) and (d)), in addition to the peak from jet fragmentation, a pronounced long-range structure is seen at φ ≈ 0, extending at least  (4) to the long-range region, which will be discussed in detail later. Following the standard zero-yield-at-minimum (ZYAM) procedure [47], each distribution is shifted to have zero associated yield at its minimum to represent the correlated portion of the associated yield. Selection of fixed p trig T and p assoc T ranges of 1-3 GeV is shown for the long-range region (top) and for the difference of the short-and long-range regions (bottom) in Fig. 3. As illustrated in Fig. 2, the near-side long-range signal remains nearly constant in η. Therefore, by taking a difference of 1D φ projections between the short-and long-range regions, the near-side jet correlations can be extracted. As shown in the bottom panels of Fig. 3, due to biases in multiplicity selection toward higher p T jets, a larger jet peak yield is observed for events selected with higher multiplicities. Because charged particles are directly used in determining the multiplic-ity in the event, this selection bias is much stronger for charged particles than K 0 S and Λ/Λ hadrons. For N offline trk < 35, no near-side correlations are observed in the long-range region for any particle species. The PbPb data show qualitatively the same behavior as the pPb data, and thus are not presented here.
Recently, the v 2 anisotropy harmonics for charged pions, kaons, and protons have been studied using two-particle correlations in pPb collisions [31], and are found to be qualitatively consistent with hydrodynamic models [32,33]. In this paper, the elliptic (v 2 ) and triangular (v 3 ) flow harmonics of K 0 S and Λ/Λ particles are extracted from the Fourier decomposition of 1D φ correlation functions for the long-range region (| η| > 2) in a significantly larger sample of pPb collisions such that the particle species dependence of v n can be investigated in detail. In Fig. 4, the v sig 2 of K 0 S and Λ/Λ particles are plotted as a function of p T for the three lowest multiplicity ranges in PbPb and pPb collisions. These data were recorded using a minimum bias trigger. The range of the fraction of the full multiplicity distribution that each multiplicity selection corresponds to, as determined in Ref.
[47], is also specified in the figure. In contrast to most other PbPb analyses, the present work uses multiplicity to classify events, instead of the total energy deposited in HF (the standard procedure of centrality determination in PbPb) [47,51]. By examining the HF energy distribution for PbPb events in each of the multiplicity ranges, the corresponding average HF fractional cross section (and its standard deviation) can be determined, which are presented for PbPb data in the figure.
In the low multiplicity region (Fig. 4), the v 2 values of K 0 S and Λ/Λ particles are compatible within statistical uncertainties. As there is no evident long-range near-side correlation seen in these low-multiplicity events, the extracted v 2 most likely reflects backto-back jet correlations on the away side. Away-side jet correlations typically appear as a peak structure around φ ≈ π , which contributes to various orders of Fourier terms.   cles. Note that the ratio of baryon to meson yield in pPb collisions is enhanced at higher p T , an effect that becomes stronger as multiplicity increases [54,55]. This should also be taken into account when comparing v n values between inclusive and identified particles. Comparing the results in Fig. 4 and Fig. 5, the dependence of v 2 on the particle species may already be emerging in the multiplicity range of 60 ≤ N offline trk < 120.
The scaling behavior of v 2 divided by the number of constituent quarks as a function of transverse kinetic energy per quark, KE T /n q , is investigated for high-multiplicity pPb events in the middle row of Fig. 5. After scaling by the number of quarks, the v 2 distributions for K 0 S and Λ/Λ particles are found to be in agreement. The middle row of Fig. 5 also shows the result of fitting a polynomial function to the K 0 S data. The bottom row of Fig. 5 shows the n q -scaled v 2 results for K 0 S and Λ/Λ particles divided by this polynomial function fit, indicating that the scaling is valid to better than 10% over most of the KE T /n q range, except for KE T /n q < 0.2 GeV where the deviation grows to about 20%. In AA collisions, this approximate scaling behavior is conjectured to be related to quark recombination [39-41], which postulates that collective flow is developed among constituent quarks before they combine into final-state hadrons. Note that the scaling of v 2 with the number of constituent quarks was originally observed as a function of p T , instead of KE T , for the intermediate p T range of a few GeV [38], and interpreted in a simple picture of quark coalescence [39]. However, it was later discovered that when plotted as a function of KE T in order to remove the mass difference of identified hadrons, the scaling appears to hold over the entire kinematic range [42,43]. However, this scaling behavior is not expected to be exact at low p T in hydrodynamic models because of the impact of radial flow. As the v n data tend to approach a constant value as a function of p T or KE T for p T 2 GeV, the scaling behavior in terms of p T and KE T cannot be differentiated in that regime. Therefore, the n q -scaled v n results in this paper are presented as a function of KE T /n q in order to explore the scaling behavior over a wider kinematic range.
The particle species dependence of v 2 and its scaling behavior is also studied in PbPb data over the same multiplicity ranges as for the pPb data, as shown in Fig. 6. The mean and standard deviation of the HF fractional cross section of the PbPb data are indicated on the plots. Qualitatively, a similar particle-species dependence of v 2 is observed. However, the mass ordering effect is found to be less evident in PbPb data than in pPb data for all multiplicity ranges. In hydrodynamic models, this may indicate a stronger radial flow is developed in the pPb system as its energy density is higher than that of a PbPb system due to having a smaller size system at the same multiplicity. Moreover, the n q -scaled v 2 data in PbPb at similar multiplicities suggest a stronger violation of constituent quark number scaling, up to 25%, than is observed in pPb, especially for higher KE T /n q values. This is also observed in peripheral AuAu collisions at RHIC, while the scaling applies more closely for central AuAu colli

Summary
Measurements of two-particle correlations with an identified K 0 S or Λ/Λ trigger particle have been presented over a broad transverse momentum and pseudorapidity range in pPb collisions at √ s N N = 5.02 TeV and PbPb collisions at √ s N N = 2.76 TeV. With the implementation of a high-multiplicity trigger during the LHC 2013 pPb run, the identified particle correlation data in pPb collisions are explored over a broad particle multiplicity range, comparable to that covered by 50-100% centrality PbPb collisions. The long-range (| η| > 2) correlations are quantified in terms of azimuthal anisotropy Fourier harmonics (v n ) motivated by hydrodynamic models. In low-multiplicity pPb and PbPb events, similar v 2 values of K 0 S and Λ/Λ particles are observed, which likely originate from back-to-back jet correlations. For higher event multiplicities, a particle species dependence of v 2 (p T ) and v 3 (p T ) is observed.
For p T 2 GeV, the values of v n for K 0 S particles are found to be larger than those of Λ/Λ particles, while this order is reversed at higher p T . This behavior is consistent with RHIC and LHC results in AA collisions and for identified charged hadrons in pPb and dAu collisions. For similar event multiplicities, the particle species dependence of v 2 and v 3 at low p T is observed to be more pronounced in pPb than in PbPb collisions. In the context of hydrodynamic models, this may indicate that a stronger radial flow boost is developed in pPb collisions. Furthermore, constituent quark number scaling of v 2 and v 3 between K 0 S and Λ/Λ parti-cles is found to apply for PbPb and high-multiplicity pPb events. The constituent quark number scaling is found to hold at the 10% (25%) level in pPb (PbPb) collisions, for similar event multiplicities. It will be interesting to see if this scaling law continues to hold for other particles. The results presented in this paper provide important input to the further exploration of the possible collective flow origin of long-range correlations, and can be used to evaluate models of quark recombination in a deconfined medium of quarks and gluons. [13] CMS Collaboration, Observation of long-range near-side angular correlations in proton-proton collisions at the LHC, J.