Measurement of jet radial proﬁles in Pb–Pb collisions

The jet radial structure and particle transverse momentum ( p T ) composition within jets are presented in centrality-selected Pb–Pb collisions at √ s NN = 2 . 76 TeV. Track-based jets, which are also called charged jets, were reconstructed with a resolution parameter of R = 0 . 3 at midrapidity | η chjet | < 0 . 6 for transverse momenta p T , chjet = 30–120 GeV/ c . Jet–hadron correlations in relative azimuth and pseudorapidity space ( (cid:4) ϕ , (cid:4) η ) are measured to study the distribution of the associated particles around the jet axis for different p T , assoc -ranges between 1 and 20 GeV/ c . The data in Pb–Pb collisions are compared to reference distributions for pp collisions, obtained using embedded PYTHIA simulations. The number of high- p T associate particles (4 < p T , assoc < 20 GeV/ c ) in Pb–Pb collisions is found to be suppressed compared to the reference by 30 to 10%, depending on centrality. The radial particle distribution relative to the jet axis shows a moderate modiﬁcation in Pb–Pb collisions with respect to PYTHIA. High- p T associate particles are slightly more collimated in Pb–Pb collisions compared to the reference, while low- p T associate particles tend to be broadened. The results, which are presented for the ﬁrst time down to p T , chjet = 30 GeV/ c in Pb–Pb collisions, are compatible with both previous jet–hadron-related measurements from the CMS Collaboration and jet shape measurements from the ALICE Collaboration at higher p T , and add further support for

A unique way to characterize the properties of the QGP is to utilize jets as a probe of the medium. Hard scatterings are expected to occur early in the collision evolution, producing high transverse momentum (p T ) partons that propagate through the expanding medium and eventually fragment into jets of hadrons. High-p T partons lose energy in interactions with the medium due to elastic scattering and induced gluon radiation [18,19]. Besides a reduction of the jet energy, this can result in a broadening of the transverse jet profile and a softening of the fragmentation [20]. E-mail address: alice -publications @cern .ch. Jet quenching has been observed at RHIC [21][22][23][24][25][26][27][28][29][30][31][32][33][34] and at the LHC [8,16,17,35-47], e.g. via inclusive yield and correlation measurements of high-p T hadrons and reconstructed jets. These measurements provide insights into the mechanisms of parton energy loss in the medium and eventually into the medium itself.
More differential measurements of the jet modification in a medium, i.e. measurements of modifications of jet angular profile and particle composition, can provide complementary information to observables that focus on the overall yield change like nuclear modification factors. Measurements of correlated associated particle production relative to jets or high-p T particles allow for a detailed measurement of the redistribution of quenched energy around the jet. An excess of low-p T particles in and around the jet up to large distances, as well as a suppression of high-p T particles, have been reported [17,48-50]. Two-particle correlations and jet-hadron correlations show an angular broadening of low-p T particles below 3 GeV/c in heavy-ion collisions with respect to pp collisions [50]. For low-p T two-particle correlations, measurements also indicate an asymmetry in the shape of the near-side jet peaks: they are broader in η compared to ϕ [48,49]. The variables η and ϕ are the distance in pseudorapidity η and azimuth ϕ relative to the near-side jet. At the same time, measurements of the radial moment of jets point to a general collimation of jets in Pb-Pb collisions [ Using jets instead of high-p T particles as a reference (trigger) to study angular correlations-as done in this analysis-should have the advantage that jet properties better reflect the initial parton energy. This analysis extends the study of jet-hadron correlations into a regime of low track-based jet p T, ch jet not yet explored with these techniques at the LHC.
In this paper, we study the correlation of charged particles (associates) with the direction of reconstructed track-based jets (triggers) in the ϕ-η plane in the same event. The jets are reconstructed using charged particles above a certain transverse momentum p T,const . The analysis focuses on two aspects of the modification of jets within the medium created in Pb-Pb collisions compared to a PYTHIA [52] reference. First, the overall modification of the associated particle yield and its jet-energy dependence is studied. Second, the modification of the radial distribution of associated particles with respect to the jet axis is studied by comparing the Pb-Pb results to the PYTHIA reference. Both aspects are analyzed in detail for several jet transverse momenta p T, ch jet and low and high p T of associated charged particles. PYTHIA is used as vacuum baseline, because the size of the pp dataset at √ s = 2.76 TeV is insufficient for this analysis.
The paper is structured as follows. In Sec. 2, details on the detector and general data reconstruction will be given. The correlation analysis, which serves as basis for this paper, is presented in Sec. 3. Subsequently, jet reconstruction will be described in Sec. 4, followed by a discussion on the embedded PYTHIA reference in Sec. 5. Before the results will be presented in Sec. 8, the observables are introduced in Sec. 6 and systematic uncertainties are discussed in Sec. 7. A summary concludes the paper in Sec. 9.

Experimental setup
For a complete description of the ALICE detector and its performance see Refs. [53] and [54], respectively.
The data were recorded in 2011 for Pb-Pb collisions at √ s NN = 2.76 TeV using a set of centrality triggers based on the hit multiplicity measured by the V0 detector, which consists of segmented scintillators covering the full azimuth over 2.8 < η < 5.1 (V0A) and −3.7 < η < −1.7 (V0C). Events were selected with V0 multiplicities corresponding to the 0-50% most central events using the centrality determination as described in Ref. [55]. The accepted events, reconstructed as described in Ref. [56], were required to have a primary reconstructed vertex within 10 cm of the center of the detector along the beam axis. For this analysis, a total of 12M events were used. The analysis presented here relies mainly on the central ALICE tracking systems, which are located inside a large solenoidal magnet with a field strength of 0.5 T. They consist of the Inner Tracking System (ITS), a high-precision six-layer cylindrical silicon detector system with the inner layer at a radius of 3.9 cm and the outer layer at 43 cm from the beam axis, and the Time Projection Chamber (TPC) with a radial extent of 85-247 cm, which provides up to 159 independent space points per track.
To ensure a good track-momentum resolution for jet reconstruction, all reconstructed tracks were required to have at least three hits in the ITS. For tracks without any hit in the Silicon Pixel Detector (SPD), which provides the two innermost layers of the ITS, the location of the primary vertex was used in addition to the hits in the TPC and ITS. This improves the track-momentum resolution and reduces the azimuthal dependence of the track reconstruction efficiency due to the non-uniform SPD response. Accepted tracks were required to be measured with 0.15 < p T < 100 GeV/c in |η| < 0.9, and to have at least 70 TPC space-points and no less than 80% of the geometrically findable space-points in the TPC.
The single-track tracking efficiency was estimated from the detector response of HIJING [57] events reconstructed to detector level using GEANT3 [58] for the particle transport. In the 0-10% centrality class, it is about 56% at 0.15 GeV/c, about 83% at 1.5 GeV/c and then decreases to 81% at 3 GeV/c, after which it increases and levels off to about 83% at above 6.5 GeV/c. For the 10-30% most central collisions, the tracking efficiency follows a similar p T -dependence pattern, with absolute values of the efficiency that are 1 to 2% higher compared to the 0-10% most central collisions. The momentum resolution, which was estimated on a track-by-track basis using the covariance matrix of the track fit, is about 1% at 1 GeV/c and about 3% at 50 GeV/c. The contamination by secondary particles [59] produced in particle-material interactions, conversions, and weak-decay products of long-lived particles is on the level of few percent.

Correlation analysis
The two-dimensional associated per-trigger yield Y ( ϕ, η) measures the distribution of particles relative to the jet axes in bins of ϕ, η, event centrality, and trigger and associate transverse momenta p T, assoc [60]. This distribution serves as the basis of the analysis and is formed using so-called same and mixed event correlations. Correlations from the same event are the actual correlations of trigger jets and associated particles, calculated for each selected event. In the mixed event technique, jets are correlated with particles from a pool containing tracks from different events with similar trigger jet p T , vertex z, and centralities. For vertex z, there are six bins in this pool, whose boundaries are given by (−10, −5, −2, 0, 2, 5, 10) in cm.
The mixed-event-corrected associated per-trigger yield for given jet p T -range, associate p T -range, and centrality selection is defined as where the ratios in the sum are formed differentially in bins of centrality and vertex z.
The factor α in Eq. (1) is chosen such that the mixed-event correlations are normalized to unity in the region | η| < 0.2, | ϕ| < 0.2 around the near-side jet peak where the efficiency for pairs of parallel jets and associates is largest. The contribution of the statistical uncertainty of this normalization to the total statistical uncertainty is negligible. The finite tracking efficiency and the contamination by secondaries (see Sec. 2) are taken into account and a correction has been performed for associated tracks differentially in η, p T , centrality, and vertex z for same and mixed event correlations in Eq. (1). The efficiency maps were created using Monte Carlo simulations for the same track definition and detector conditions. However, this correction turns out to be negligible for all observables except for the absolute jet-associated yields, because its effect mostly cancels in the used relative observables, which will be defined in Sec. 6.
In addition to the correction for detector inhomogeneities and acceptance effects, the correlation also needs to be corrected for background. The underlying background for the chosen observables mainly consists of the uncorrelated particle background baseline from soft processes and the flow modulation in the correlation function. The background was found to be independent of within |η| < 0.9 [61] and is therefore estimated as a function of ϕ for the whole η-range as B( ϕ). To avoid including parts of the jet signal, B( ϕ) is calculated in 1.0 < | η| < 1.4, where the contribution from the jet is expected to be small, based on measurements in pp collisions.
The background is directly subtracted from the correlation function. The background-corrected per-trigger yield serves as a basis for all subsequent measurements. It is defined as To illustrate the impact of the background on the per-trigger yields, the uncorrected per-trigger yields can be found in Fig. 1 for high-and low-p T associates. The background is nearly negligible for high-p T associates and it is sizeable for low-p T associates. In the illustrated example for low-p T associates, the signal to signal+background ratio, i.e. the percentage of the signal in the measured observable, is roughly 0.1 within a radius of r < 0.3 around the near-side jet peak. Note also that the background correction removes all η-independent correlations, including the away-side ridge which is not investigated in the presented analysis.

Jet reconstruction
The measurement of jets in heavy-ion collisions is challenging since a single event can contain multiple, possibly overlapping, jets from independent hard nucleon-nucleon scatterings. In addition, low transverse momentum particles originating from soft processes lead to a fluctuating background which strongly influences the jet reconstruction. The relative effect is largest for low-p T jets and most central events. Consequently, jet reconstruction in heavyion collisions requires a robust jet definition, and a procedure to correct for the presence of the large background [62].
Jets were reconstructed using the anti-k T or the k T algorithms [63] in the FastJet package [64] with a resolution parameter of R = 0.3. Only those jets whose axis was reconstructed within |η| < 0.6 were kept in the analysis to assure the nominal jet cone is fully contained within the track acceptance of |η| < 0.9. This limits the effect of the acceptance boundaries on the measured jet spectrum. Jets reconstructed by the anti-k T algorithm were used to quantify signal jets, while jets reconstructed by the k T algorithm were used to quantify the contribution from the underlying event [65].
Two different jet definitions are used in this analysis: for measurements at high associate-p T , jets are measured with a constituent cut p T,const ≥ 0.15 GeV/c, measurements at low associate-p T are performed for jets measured with a constituent cut p T,const ≥ 3.0 GeV/c. Jets with p T,const ≥ 0.15 GeV/c are reconstructed using all charged particles available for jet reconstruction and, thus, the fragmentation bias is as small as possible. This bias is caused by only including certain particles of the jet and could lead to a sample of harder fragmenting jets when leaving out particles at low p T . On the other hand, using all charged particles available for jet reconstruction also includes particles in the correlation analysis which were already used in the jet finding process. The jet finding algorithm selects regions in momentum space with large energy flow. This implies that the distribution of charged particles inside the jet is biased. For example, the radial distribution of particles with respect to the jet axis will show a small depletion at distances just outside the jet cone radius R. This particularly affects the shape of the jet, i.e. how the constituents are distributed relative to the jet axis, leading to an autocorrelation bias.
Therefore, the jets themselves and in particular their shapes are intimately connected to the jet definition. For high-p T associates, the autocorrelation bias cannot be avoided and has to be accepted as a part of the jet definition.
Low-p T associates are broadly distributed up to large distances relative to the jet. Since the jet finding algorithm clusters the jets roughly into cones with a nominal radius of R = 0.3, it strongly affects the shape of the jet. When measuring properties of low-p T associates, we avoid the autocorrelation bias by adapting the jet definition: Trigger jets and associates can be decoupled by using jets with constituents above a certain threshold and associates below the threshold. Therefore, for measurements at low associate-p T , jets are reconstructed with p T,const ≥ 3 GeV/c. Using a geometrical matching procedure that is performed on two collections of the differently defined jets which are reconstructed in each event it was checked that the jet axes for both jet definitions do not strongly change. For instance, for jets with p T,const ≥ 3 GeV/c and p T, ch jet > 30 GeV/c the mean and standard deviation of the matched jet distance distribution are approximately given by 0.016 and 0.014, respectively. However, it must be emphasized that these jet definitions select two different jet samples and that the autocorrelation bias was avoided here at the expense of a possible fragmentation bias.
The transverse momentum of reconstructed jets including constituents as low as 0.15 GeV/c is affected by the contribution from the underlying event. In order to suppress the contribution of such background to the measurement of the jet momentum, we followed the approach described in Refs. [65,66], which addresses the average additive contribution to the jet momentum on a jet-by-jet basis. The underlying background momentum density ρ was estimated event-by-event using the median of p raw T, jet /A jet , where p raw T, jet Table 1 True jet populations p T, true in GeV/c corresponding to given p T, ch jet -ranges for different event centrality classes. The ranges are given such that they contain at least 68% of the jet population. The most probable values of the distributions are given in parentheses.  (52) is the uncorrected jet transverse momentum and A jet is the area of jets reconstructed with the k T algorithm.
The average raw background momentum density ρ decreases towards more peripheral collisions. It is ρ ≈ 110, 65, and 25 GeV/c in the 0-10%, 10-30%, and 30-50% most central Pb-Pb collisions, respectively. The background momentum density is a detector-level quantity that depends on the tracking efficiency and track definition. For signal jets reconstructed with the anti-k T algorithm and constituents above 0.15 GeV/c, the background density scaled by the area of the reconstructed signal jet was subtracted from the raw reconstructed transverse momentum (p raw T, jet ) of the signal jet according to p T, ch jet = p raw Due to region-to-region variations of the background, the background-corrected jet transverse momenta are affected by residual fluctuations. To give an estimate for these fluctuations for the jet definition used, cones with radius R = 0.3 are randomly placed in each event. In these cones, the track momenta are summed and the background is subtracted to calculate δ p T : where A is the area of the cone.
For the 0-10%, 10-30%, and 30-50% most central collisions, the standard deviation of the δ p T -distribution as a measure for the magnitude of the fluctuations has been evaluated to 6.7, 5.1, and 3.3 GeV/c, respectively. Since the δ p T -distribution also contains the jet signal, the standard deviation of the full distribution is impacted by it. A lower limit of these fluctuations is given by performing a Gaussian fit of the left-hand side of the δ p T -distribution.
The Gaussian widths were evaluated to 5.5, 4.0, and 2.3 GeV/c for the 0-10%, 10-30%, and 30-50% most central collisions. The sample of jets that only uses constituents above 3 GeV/c is not corrected for the underlying event as the constituent cut already strongly suppresses the contribution from the background such that it is negligible. In addition to background fluctuations, also the finite detector resolution and single particle efficiency influence the measurement. To quantify both effects, the ratio of reconstructed jet momentum p T, rec and true jet momentum p T, true was calculated taking into account the detector resolution by using a response matrix and background fluctuations given by the δ p T distributions.
The response matrix was created from Monte Carlo simulations for which the true jet momentum is known by geometrically matching particle-level PYTHIA jets with the corresponding detector-level jets reconstructed using a full detector model in GEANT3. More detailed studies have been performed for jets on the same dataset in Ref. [66].
There are two effects contributing to the jet momentum resolution: detector effects and underlying event fluctuations. The detector effects lead to a jet momentum response that is peaked at p T, rec = p T, true , but has a tail to lower values of detector level momentum due to tracking inefficiency. The tracking efficiency changes by only a few percent from peripheral to central events. Background fluctuations produce an approximately Gaussian response, with a width that depends strongly on centrality. The combined effect leads to a standard deviation in the jet momentum resolution of 30% (20%) for jets with p T, ch jet = 30 GeV/c and 27% (27%) for jets with p T, ch jet = 120 GeV/c for the 0-10% (10-30 and 30-50%) most central events.
It should be emphasized that p T, ch jet refers to the jet transverse momentum at detector level, corrected for background only. Since within-event fluctuations of the background are not corrected for, the mean of the given p T, ch jet -range is slightly higher than that of the underlying true p T distribution for more central collisions where fluctuations are dominant. Hence, due to the steeply-falling jet spectrum, fluctuations lead to a shift of the spectrum to larger values. For more peripheral collisions where detector effects are dominant, there is the opposite effect, i.e. the spectrum is shifted to smaller values. The fraction of purely combinatorial jets in the momentum ranges used in the analysis was found to be negligible.
To give a rough estimate of the true jet populations for a given reconstructed jet momentum range, projections of the response matrices, introduced above, are used [67]. For measured p T, ch jet -distributions, approximate ranges are given in Table 1 as a measure for the true jet momentum distributions. The true populations are defined as the smallest possible ranges around the p T, ch jet -range in which at least 68% of the jet population can be found.

Construction of PYTHIA baseline
In this analysis, reconstructed detector-level PYTHIA-jets serve as vacuum baseline, because the size of the pp dataset at √ s = 2.76 TeV is insufficient for this purpose.
To account for the fluctuations of the underlying event in Pb-Pb collisions, PYTHIA jets embedded in real Pb-Pb collisions are used as a reference. Jets reconstructed in this reference dataset still show the same baseline jet properties but also include the effect of background fluctuations from the Pb-Pb event. To create this reference dataset, the following procedure is applied. Events are simulated with PYTHIA6 (Perugia-0 [68], version 6.421) followed by transport in the detector using GEANT3 and full response simulation and reconstruction simulating the same detector conditions as in the Pb-Pb dataset. The reconstructed tracks are embedded into Pb-Pb events, i.e. they are combined with tracks from Pb-Pb events. In order to simulate the same conditions as in Pb-Pb, the tracking efficiency in pp is decreased to the level expected in Pb-Pb. Since the tracking efficiency in pp is higher than in Pb-Pb, 2% of the PYTHIA tracks are randomly discarded before they are embedded [54]. Jet finding algorithms are applied to the PYTHIA event and also to the combined PYTHIA + Pb-Pb event. Jets found in the combined event are only accepted for the reference dataset if they can be matched geometrically with those in the PYTHIA event. A matched embedded jet needs to be less than R = 0.3 away from a PYTHIA jet.
Due to the very high particle occupancy of the Pb-Pb collision system, the probability to reconstruct a PYTHIA jet in the embedded event is much lower than the probability to reconstruct a jet of same momentum by overlapping a jet that already existed in the Pb-Pb event, even after applying a geometrical matching procedure. Therefore, without any further intervention, the embedded jet sample would consist mostly of Pb-Pb-jets overlapping low-p T PYTHIA jets.
Two approaches have been tested which ensure that the jet sample shows Pb-Pb-event-like fluctuations of a PYTHIA jet, and not jets from the Pb-Pb event. The analysis baseline technique uses a cut on the fraction of the jet p T that originates from the matched jet in PYTHIA. The applied cut values are motivated by the underlying true jet distribution that shows two separated populations: jets mostly consisting of particles from PYTHIA or from Pb-Pb. The cut value was chosen to achieve the best separation of the two distributions. In the 0-10% most central collisions, it is required that at least 20% of the jet constituents' p T originate from the PYTHIA jet. For more peripheral collisions, this fraction is increased to 25%.
For jets with p T, const ≥ 3 GeV/c, which were measured down to 30 GeV/c, a cut of 50% is applied. However, this procedure might impose a bias on the implicitly accepted background fluctuations. Therefore, variations around these nominal values were considered for the evaluation of systematic uncertainties. Alternatively, a jet veto technique has been used: an embedded jet is not accepted if it overlaps with an already existing jet of sizeable transverse momentum p T, ch jet in the Pb-Pb event. Several veto cut values between 15 and 40 GeV/c were tested. Eventually, it turns out that both approaches yield very similar results. The reconstructed jets which survive the MC percentage cut serve as an input to the next analysis steps which are the same as in the data analysis.

Observables
In this analysis, two features of particle jets are probed in Pb-Pb collisions: changes in the particle p T composition of jets and their radial distribution relative to the jet axis.
To probe relative changes in the charged particle p T composition of jets in a surrounding cone with R = 0.3, the jet-associated yield ratio is measured. The ratio is formed from the integrated jet-associated per-trigger yields Y PbPb and Y emb which represent the integrals of the per-trigger yield in the jet cone for a given p T, assoc -range as introduced in Eq. (2). Technically, the integral is the sum over the entries of all ( η, ϕ)-bins whose center is within distances of up to R = 0.3 around the jet axis in the background-corrected per-trigger yield histogram.
The jet-associated yield ratio is defined by It directly compares integrated jet-associated per-trigger yields in Pb-Pb to the same yields for embedded PYTHIA jets. An enhancement or suppression in associated yields is directly seen as a deviation from unity in the ratio. The relative radial particle distribution around the jet is directly derived from the jet-associated yields. It shows the relative distribution of particle yields inside the jet cone. Thus, it is a measure for the broadening or collimation of constituents with certain momenta in or around the jet cone. As for the jet-associated yield ratio, this measurement is performed for high-and low-p T jetassociated yields. The radial shape is normalized to represent a probability distribution. It is defined in bins of r = η 2 + ϕ 2 , the distance to the jet axis, to exploit the radial symmetry of the jet peak. In Refs. [48,49], an asymmetric broadening of the nearside jet peak is observed in two-particle correlations. It is strongest for low associate and trigger momenta and vanishes for higher momenta. Therefore, in the analysis presented here, the influence of this asymmetry on jet-hadron correlations was tested to check the radial symmetry of the jet peak. Even for the lowest accessible jet and associated track momenta, no jet peak asymmetry was observed. Measurements in η and ϕ lead to the same conclusions within statistical precision, which justifies the presentation of the jet radial shape in bins of r. The correlation function which is used to obtain the shape is originally binned in η and ϕ. The binning was chosen fine enough to avoid significant binning effects.
For a given centrality-bin, and trigger and associate p T , it is defined by the following formula: where Y corr (r) represents the background-corrected per-trigger yield, r min and r max the bin edges, and A = r range 0 Y corr (r)dr the integral for the self-normalization of the radial shape. The upper limit in the integral used for the self-normalization is chosen to reflect the different ranges of the shown radial shape and is r range = 0.3 for the jets with p T, const ≥ 0.15 GeV/c and r range = 0.9 for jets with p T, const ≥ 3 GeV/c. The statistical uncertainty is calculated taking into account the self-normalization.

Systematic uncertainties
Several sources of systematic uncertainties contribute to the full uncertainty of the measurement and the evaluated individual uncertainties are combined using a quadratic sum, assuming they are uncorrelated. Uncertainties for the following analysis aspects have been taken into account: the non-jet-related background correction technique, the mixed-event correction, the selection of embedded jets, the tracking efficiency, and the impact of using a PYTHIA reference instead of a measured reference in pp at the same energy. The uncertainties are partly correlated point-to-point. The discussed uncertainties are summarized in Tables 2-4. To correct for the non-jet-correlated background in the correlation function, the background is evaluated on the sidebands and subtracted in ϕ, as described in Sec. 3. Different underlying background methods for the correlation functions have been tested: for systematic uncertainties, the definition of the sideband range was varied to 1.1 < | η| < 1.3 instead of 1.0 < | η| < 1.4. In addition, a simpler method that approximates the background by a constant baseline (B( ϕ = const)) has been used.
The mixed-event acceptance/inhomogeneity correction is a small correction. Two variations are considered for systematic uncertainties. First, the mixed-event correction is calculated inclusively for all ϕ. Second, the normalization of the mixed-event correlations is performed for | η| < 0.3 and full | ϕ| instead of using the plateau in | η| < 0.2 and | ϕ| < 0.2.
In the embedding, a cut motivated by studying the underlying true jet distributions is applied on the fraction of jet p T originating from the PYTHIA event, as described in Sec. 5. Instead of cutting at 20% for 0-10% centrality, and 25% for other centralities, the cut is varied to 15% and 25% for 0-10% centrality, and to 20% and 30% for other centralities. As described above, for jets with p T, const ≥ 3 GeV/c a baseline cut value of 50% is used. For systematic variation, the cut is performed at 15% and 60% for 0-10% centrality, 20% and 60% for other centralities.
The detector has a finite single track reconstruction efficiency, which is only known with finite precision. Since all observables are corrected for the tracking efficiency, they are all directly affected by its uncertainty. Detailed studies of the tracking efficiency uncertainty have been performed to evaluate the size of its systematic uncertainty [54,66]. The studies indicate that the (absolute) uncertainty is 4% for Pb-Pb collisions, mainly due to an imperfect Table 2 Table of systematic uncertainties for jet-associated yields in Pb-Pb, embedded PYTHIA, and their ratio for high-p T associates (4-20 GeV/c) and low-p T associates (1-2 GeV/c) and for the 0-10% most central collisions. Uncertainties are given as relative uncertainties in percentages.  Table 3 Table of systematic uncertainties for jet radial shapes for high-p T associates (4-20 GeV/c) in Pb-Pb and embedded PYTHIA for the 0-10% most central collisions. Uncertainties are given as relative uncertainties in percentages. Note that relative uncertainties grow for higher r values.  description of the ITS-TPC matching efficiency. Another uncertainty from the tracking efficiency correction enters this analysis due to the usage of PYTHIA simulations. The tracking efficiency of the PYTHIA data is artificially lowered by 2% before embedding to account for the lower tracking efficiency in Pb-Pb collisions. As a conservative estimate, a relative uncertainty of 100% is assigned to this value. Both components of the tracking efficiency uncertainty are taken into account as independent contributions to the uncertainty, i.e. added in quadrature to the full uncertainty. These uncertainties are directly used as uncertainties for the yields, see Table 2. For the jet-associated yield ratio, the uncertainty on the tracking efficiency in Pb-Pb cancels, because it is correlated in Pb-Pb and the embedded PYTHIA reference. For the radial shape distribution, a change in the tracking efficiency has no impact either, since these observables are relative quantities that do not depend on the global magnitude of the tracking efficiency. As an alternative approach to estimate the impact of these two uncertainties of the tracking efficiencies on the observables, the full analysis was redone using corrections that assume the above given lower tracking efficiencies. There was no significant impact on the presented results. Finally, an uncertainty is assigned since PYTHIA is used as a baseline instead of a measured pp reference. Including this uncertainty, the conclusions are also valid for a pp reference and not only for an embedded PYTHIA reference. In order to do so, the presented observables were calculated and compared for PYTHIA events and pp collisions at 7 TeV. Within the statistical precision of this comparison, it is only possible to give an estimate for the inclusive p T, ch jet -range. The relative deviations of each observable between both datasets enter directly as a systematic uncertainty and are on the level of a few percent, cf. Tables 2-4.

Results
Figs. 2 and 3 depict the jet-associated yields (left) and yield ratios (right) for high-p T and low-p T associated particles, respec-  tively. Both quantities are shown as a function of event centrality and for several selected jet transverse momenta.
The jet-associated yield ratio shows a suppression with a significance of several standard deviations in the centrality range 0-50% for the considered high-p T associated particles. In the probed jet momentum range, no significant p T, ch jet -dependence is observed. The centrality-dependent linear slope of the distribution for p T, ch jet = 40-60 GeV/c is more than one standard deviation away from zero, taking into account statistical and systematic uncertainties added in quadrature, indicating that there is a slightly stronger suppression for more central collisions in this case. As a cross check, the same observable was also measured for jets with several higher minimum p T, const -cuts, i.e. 1, 2, and 3 GeV/c, which are less affected by the underlying event. They lead to similar conclusions.
The jet-associated yield ratio for low-p T associates has much larger statistical and systematic uncertainties than the ratio of high-p T constituents, thus it is not possible to draw a definite conclusion.
The measured jet relative radial shapes are presented in Figs. 4 and 5. The top panels show the self-normalized distributions, the difference and the ratio of the shapes in Pb-Pb and embedded PYTHIA can be found in the two lower panels. The jet radial shapes of high-p T associates are measured for p T, ch jet = 40-60 GeV/c, 60-80 GeV/c, and 80-120 GeV/c. Shapes of low-p T associates are presented for jets with p T, ch jet = 30-40 GeV/c and p T, const >3 GeV/c for associates with p T, assoc = 1-2 GeV/c and p T, assoc = 2-3 GeV/c.
In general, the radial shape measurements indicate that all jetassociated yields are similarly distributed relative to the jet axis in Pb-Pb and embedded PYTHIA. The yields of high-p T associates appear to be slightly more collimated near the core for jets in Pb-Pb, though the absolute effect is small. While the shape is not significantly changed for jet transverse momenta between 40 and 60 GeV/c in Pb-Pb compared to the reference, there is a visible collimation for higher jet momenta above 60 GeV/c. This can be seen best in the difference distributions PbPb−emb of Fig. 4 which show that a larger fraction of the associated yield can be found near the core in Pb-Pb collisions.
The ratio distributions show that the collimation effect persists up to r = 0.2, which is best visible for jets with p T, ch jet = 60-80 GeV/c. In the CMS measurement [50], no significant change of the near-side jet peak width is observed in Pb-Pb for high-p T associates and jets above 120 GeV/c. However, the magnitude of the effect observed here is compatible with the observations within uncertainties. Also note that the CMS data hints as well to a small collimation of the peak for higher-p T associates (4-8 GeV/c). Possible effects which might lead to a collimation include a relative change in the quark/gluon content in Pb-Pb compared to the reference [69], as well as a suppression of large-angle soft radiation in the coherent jet energy loss picture [70,71]. Low-p T jet-associated yields presented in Fig. 5 are measured up to a distance of r = 0.9 relative to the jet since in this case the associates are decoupled from the trigger jets.
For p T, assoc = 1-2 GeV/c, a hint of a broadening of the radial shape is observed for jets with momenta between 30 and 40 GeV/c Fig. 4. Jet relative radial shape distributions, differences, and ratios for the 0-10% most central collisions for high-p T constituents, shown for different jet transverse momenta. Boxes represent systematic uncertainties, shaded boxes include uncertainties from PYTHIA/pp comparison, and error bars represent statistical uncertainties. Observables are corrected for acceptance and background effects.
for the given definition. The broadening is visible in the difference distribution of the left plot in Fig. 5: in Pb-Pb collisions, a smaller fraction of particles can be found directly next to the jet axis. For higher associate transverse momenta, i.e. p T, assoc = 2-3 GeV/c, there is no significant modification of the low-p T radial shape of jets in Pb-Pb collisions within the large current experimental uncertainties. A robust measurement of this observable for p T, ch jet = 40-60 GeV/c or higher momenta is not possible due to the insufficient size of the dataset. For higher jet momenta above 120 GeV/c, CMS measures a significant broadening of the near-side jet peak.

Summary
The presented results constitute the first attempt to study jethadron correlations with track-based jets down to transverse momenta of 30 GeV/c in Pb-Pb collisions -a challenging regime due to the large underlying event and its fluctuations. The jet ra-dial shapes and the change in the particle p T composition were measured in Pb-Pb collisions at √ s NN = 2.76 TeV for high-and low-p T associates and compared to embedded PYTHIA simulations. The number of high-p T associates in Pb-Pb collisions is suppressed compared to the reference by roughly 30 to 10%, depending on centrality. The radial particle distribution relative to the jet axis shows a moderate modification in Pb-Pb collisions with respect to PYTHIA. High-p T associate particles are slightly more collimated in Pb-Pb collisions compared to the reference.
For jets with p T, const ≥ 3 GeV/c, the radial distributions of low-p T associates were measured. A hint of a broadening of the low-p T radial shapes is observed for p T, assoc = 1-2 GeV/c. The shape for p T, assoc = 2-3 GeV/c does not show a significant modification within its large uncertainties. The results are in line with both previous jet-hadron-related measurements from the CMS Collaboration and jet shape measurements from the ALICE Collaboration at higher p T and add further support for the established picture of in-medium parton energy loss.

Acknowledgements
The ALICE Collaboration would like to thank all its engineers and technicians for their invaluable contributions to the construction of the experiment and the CERN accelerator teams for the outstanding performance of the LHC complex. The ALICE Collaboration gratefully acknowledges the resources and support pro-