Measurement of the Sensitivity of Two-Particle Correlations in $pp$ Collisions to the Presence of Hard Scatterings

A key open question in the study of multi-particle production in high-energy $pp$ collisions is the relationship between the"ridge'' - observed azimuthal correlations between particles in the underlying event that extend over all rapidities - and hard or semi-hard scattering processes. In particular, it is not known whether jets or their soft fragments are correlated with particles in the underlying event. To address this question, two-particle correlations are measured in $pp$ collisions at $\sqrt{s}=13$ TeV using data collected by the ATLAS experiment at the LHC, with an integrated luminosity of $15.8$ $\mathrm{pb}^{-1}$, in two different configurations. In the first case, charged particles associated with jets are excluded from the correlation analysis, while in the second case, correlations are measured between particles within jets and charged particles from the underlying event. Second-order flow coefficients, $v_2$, are presented as a function of event multiplicity and transverse momentum. These measurements show that excluding particles associated with jets does not affect the measured correlations. Moreover, particles associated with jets do not exhibit any significant azimuthal correlations with the underlying event, ruling out hard processes contributing to the ridge.

In heavy-ion collisions, two-particle correlations (2PC) in relative azimuthal angle with large pseudorapidity [1] separation show distinct long-range correlations [2][3][4][5][6][7][8][9][10][11][12].These long-range correlations are a simple manifestation of the single-particle anisotropies,   , that originate from the hydrodynamic expansion of the quark-gluon plasma (QGP) produced in these collisions.The   are defined by parameterizing the azimuthal distribution of produced particles as: where  is the azimuthal angle of the particle momentum and   and Ψ  are the magnitude and phase of the  th -order anisotropy, see Refs.[4,10] and references therein.
Because of their hydrodynamic origin in nucleus-nucleus (A+A) collisions, such long-range correlations were not expected in smaller colliding systems such as proton-nucleus (+A) or proton-proton ( ) collisions, where collective phenomena were not commonly expected to develop.However, measurements by CMS showed the presence of such long-range correlations, known as the "ridge," in high-multiplicity   collisions [13].Further investigations by ATLAS [9,14,15] have demonstrated that these long-range correlations in   collisions are produced from single-particle anisotropies similar to those in heavy-ion collisions.These long-range correlations have been interpreted as evidence of collective effects similar to those seen in heavy-ion collisions.However, some authors have proposed that the ridge primarily results from correlated production of partons in the presence of dense gluonic initial states (i.e. the GLASMA) [16][17][18][19][20], implying that much of the correlation structure associated with the ridge should be associated with hard-or semi-hard scattering processes.Previous measurements [21] have shown that the ridge is unmodified in   collisions producing a  boson, but no direct measurement in   collisions of the correlation between jets or their fragments and the underlying event has yet been performed, while such a correlation has been observed in +Pb collisions [22,23].
This Letter presents 2PC measurements in   collisions at a center-of-mass energy ( √ ) of 13 TeV, using the ATLAS detector at the LHC.The measurements are performed with two different particle-pair selections.The first case explores correlations between tracks that are not jet constituents, while the second case measures correlations between tracks that are constituents of jets and tracks that are well-separated from jets.Similar measurements in +Pb collisions have shown significant non-zero  2 for low [23] and high [22] transverse momentum ( T ) particles generated in hard processes.Correlations are also measured in events that are explicitly selected by requiring the presence or absence of low- T jets.These measurements can address whether or not the presence of jets affects the ridge, and if the particles from jets exhibit azimuthal correlations with particles from the underlying event and therefore contribute to the ridge.
The measurements presented here are performed using the ATLAS [24] inner detector (ID), minimum-bias trigger scintillators (MBTS), calorimeters and the trigger and data acquisition systems [25].The ID records charged-particle trajectories within the pseudorapidity range || < 2.5 using a combination of silicon pixel detectors including the "insertable B-layer" (IBL) [26,27], silicon microstrip detectors (SCT), and a straw-tube transition radiation tracker (TRT), all immersed in a 2 T axial magnetic field [1,28].The ATLAS calorimeter system consists of a liquid argon (LAr) electromagnetic calorimeter covering || < 3.2, a steel-scintillator sampling hadronic calorimeter covering || < 1.7, a LAr hadronic calorimeter covering 1.5 < || < 3.2, and two LAr electromagnetic and hadronic forward calorimeters (FCal) covering 3.2 < || < 4.9.The ATLAS trigger system [29] consists of a Level-1 (L1) trigger implemented using a combination of dedicated electronics and programmable logic, and a software-based high-level trigger (HLT).An extensive software suite [30] is used in data simulation, in the reconstruction and analysis of real and simulated data, in detector operations, and in the trigger and data acquisition systems of the experiment.
The data were collected during Run 2 of the LHC (2015-2018), with an average collision rate per bunch crossing () of less than 3, and an integrated luminosity of 15.8 pb −1 .The data used here were recorded using multiple minimum-bias, high-multiplicity, and jet triggers, which are described in Ref. [31].Additional offline requirements are imposed on the events selected by the triggers.The events are required to have a reconstructed vertex with || < 100 mm.To suppress events with more than one   collision in the same bunch crossing, events are required to have only one reconstructed vertex.Pileup events where the vertices from multiple collisions are sufficiently close such that they are reconstructed as a single vertex are not removed by the one vertex requirement.However, such merged events typically have a broader distribution for the longitudinal impact parameter of tracks relative to the vertex (| 0 sin()|).Such events are reduced by requiring that the standard deviation of | 0 sin()| for all tracks in an event is less than 0.25 mm.
The reconstruction and performance of tracks and primary vertices in the ID are described in Refs.[32][33][34].The specific track selection criteria can be found in Ref. [31].The track reconstruction efficiencies  (  T , ) are obtained using Monte Carlo (MC) generated events that are passed through a Geant4 [35] simulation [36] of the ATLAS detector and reconstructed using the procedures applied to the data.The efficiency varies between 69% and 87% as a function of  and  T .
Jets used in this analysis are reconstructed using the anti-  algorithm [37] with a radius parameter of 0.4.The inputs to jet reconstruction are "particle flow objects" as detailed in Ref. [38].Jets are calibrated to the hadronic scale using scale factors obtained from MC simulations specifically derived for low- data.Additional in situ corrections [39] are applied, which account for differences in the jet response between the MC samples and data.One issue in this analysis is that the modulation in the soft particles in the event (Eq.( 1)) biases the jet  T in a manner that depends on its orientation relative to the Ψ  .This affects the measurements of the correlations between jet-fragments and the underlying event (UE) particles (discussed in detail below).To mitigate this effect, instead of selecting jets based on their  T , selections are made on the following groomed quantity: where the sum runs over all the jet constituents with  T > 4 GeV, which considerably reduces the number of UE particles within the jet, and makes this bias negligible, as shown in Ref. [31].
In previous ATLAS measurements of 2PCs in +Pb [40,41] and   [14, 15, 21] collisions, events were quantified by  rec ch : the total number of reconstructed tracks with  T > 0.4 GeV, passing the track selections discussed above.In this analysis, a slight modification is made to ensure that the event activity is not biased by the presence of jets, and only reflects the soft multiplicity in the event.The number of constituent tracks in jets with  G T > 15 GeV is subtracted from the measured multiplicity, and the corrected quantity,  rec,corr ch , is used to represent the event activity.While counting the constituent tracks of jets, the  T > 4 GeV requirement is not imposed on the tracks.Additionally, this correction is offset by the average number of UE tracks within the jet cone.This offset is estimated by measuring the average number of tracks, as a function of  and , that are in a  = 0.4 cone in events with similar multiplicity and trigger conditions.
In 2PC measurements, the distribution of particle-pairs in relative azimuthal angle Δ =  a −  b are measured.The labels  and  denote the two particles in the pair.In evaluating the correlation functions, the tracks are weighted by the inverse of their reconstruction efficiency, 1/ (  T , ).To suppress short-range correlations, the particles are required to have a pseudorapidity separation of |Δ| > 2. In   collisions, back-to-back dĳets also make a significant contribution to the 2PCs.To remove this contribution, a template-fit method [14,15,21] is employed in which the measured 2PC is described by a fit having two components.The first component accounts for the dĳet contribution,  periph (Δ), which is measured using low-multiplicity events (called the "peripheral reference").This analysis uses the  rec,corr ch interval of 10-30 to build  periph .The second component accounts for the bulk contribution with a relative harmonic modulation,  ridge (Δ).The 2PC can then be described as: where  and  , are fit parameters and  is fixed by the requirement that the integrals of the fit and  (Δ) are equal.The Fourier moments,  , , obtained from the template-fit quantify the strength of the long-range correlation.It is demonstrated in Refs.[14, 15] that the  , in   collisions obtained from Eq. ( 3) factorize as , where   is the single particle anisotropy (Eq.( 1)).Thus, ).The tracks used in this analysis are categorized as follows: those that are separated from all  G T > 15 GeV jets by at least one unit in  [22] and having 0.5 <  T < 4 GeV are considered to be UE tracks (h UE ); tracks that are included as particle-flow constituents of jets having  G T > 40 GeV (called trigger jets henceforth) are considered to be jet constituents (h J ).Five classes of correlations are studied in this Letter: • h-h: Standard 2PC [14, 15] without applying any rejection of tracks around jets.
• h UE − h UE (AllEvents): 2PC where both tracks are h UE .About 14% of h-h 2PC pairs are removed by the abovementioned rejection.
• h UE − h UE (NoJets): 2PC using events with no jets with  G T > 15 GeV.• h UE − h UE (WithJets): 2PC using events with at least one jet with  G T > 15 GeV.• h UE − h J : 2PC performed between h UE and h J .These classes are not mutually exclusive.Specifically, the h UE − h UE (NoJets) and h UE − h UE (WithJets) classes add up to the h UE − h UE (AllEvents) class.The h UE − h J class has no overlapping particle-pairs with the ones in the h UE − h UE (AllEvents), h UE − h UE (NoJets), and h UE − h UE (WithJets) classes.The h-h class is identical to the measurements performed in the previous ATLAS publications [14, 15], and is used as a reference with which other classes are compared.
For the h UE − h J case, additional requirements are imposed on the trigger jets to avoid distortions of the 2PC.They must have no other jet with  G T > 15 GeV within Δ = √︁ Δ 2 + Δ 2 = 1 and they must have a balancing jet with  G T > 15 GeV and with |Δ| > 5/6.The first requirement removes distortions of the 2PC at smaller Δ while the second requirement ensures that fragments of the balancing jet are excluded from h UE .It may happen that some constituents of jets originate in the UE, leading to a contribution of combinatorial pairs in the 2PC.These combinatorial pairs, by construction, have the same correlation as those where both the tracks are from the UE.The contribution of such pairs is removed by the following technique.For each event that contributes to the h UE − h J correlation, a separate 2PC is made using another event with similar vertex position and multiplicity.In this event, one track is picked from an - region that is within  = 0.4 cone of the jet-axis and the other track is picked from the same  range as in the h UE − h J event.This combinatorial 2PC is then subtracted from the h UE − h J 2PC.
Statistical uncertainties in the measured 2PCs are evaluated using a bootstrapping procedure previously used in Ref. [42].Systematic uncertainties in the  2 measurements are estimated by varying different aspects of the analysis.For the template-fit procedure, the  rec,corr ch multiplicity range for the peripheral reference selection was varied from the nominal 10-30 to 10-40 and 20-40 [31] and the change in the  2 values is included as a systematic uncertainty.For the multiplicity dependence, this uncertainty for the  2 is 0.01 (absolute) for the h UE − h J class and is typically within 2% for the other classes.This uncertainty is fully correlated across all multiplicity intervals and is the dominant uncertainty for the h UE − h J class.Uncertainties in the tracking efficiency are propagated into the measured  2 .This uncertainty on the  2 is less than 0.5%, and is estimated by varying the efficiency up and down within its uncertainties (∼ ±3%) [43], and re-evaluating the  2 .The systematic uncertainty due to non-primary tracks is estimated by varying the selection criteria for transverse and longitudinal impact parameters, resulting in a 0.5% change in  2 .The 2PC analyses often use event-mixing [4,10] to estimate and correct the 2PCs for the detector's pair acceptance.This correction is quite small, and the full effect of the correction is included as a systematic uncertainty.As discussed previously, the events used in this analysis are required to have the standard deviation of | 0 sin()| for the tracks in an event to be smaller than 0.25 mm, to reduce pileup.Conservatively, the entire effect of this selection, which varies with multiplicity but is typically within 1%, is taken to be a systematic uncertainty associated with pileup effects.
Figure 1 compares the 2PCs for all classes, except the h UE − h UE (AllEvents) class.The figure also shows the template fits including the components of the fits.In general, the template fits describe the 2PCs quite well.A near-side ridge is visible for the h-h, h UE − h UE (WithJets) and h UE − h UE (NoJets) cases, while the  periph (Δ) appears to describe the full distribution in the h UE − h J case.
Figure 2 shows the multiplicity dependence of the  2 for all five 2PC classes.The  2 values for the h-h case vary weakly with multiplicity, as previously reported in Refs.[14,15].The  2 values in the h UE − h UE (AllEvents), h UE − h UE (NoJets), and h UE − h UE (WithJets) cases, are all consistent with the h-h result.This demonstrates that removing tracks associated with jets does not impact the long-range UE correlations, and nor does the presence (or absence) of jets in an event.Within uncertainties, the  2 values in the h UE − h J case are consistent with zero.The mean  2 for the h UE − h J correlations over the 40-150 multiplicity range is −0.009 ± 0.010(statistical) ± 0.014(systematic).This indicates that particles produced in hard scattering processes (with  G T > 40 GeV) do not contribute significantly to the long-range correlation observed in   collisions.Figure 3 shows the  T -dependence of the  2 .The differential  2 (  T ) values in the h UE − h UE (AllEvents), h UE − h UE (NoJets) and h UE − h UE (WithJets) cases are found to be consistent with the h-h case.Again, within uncertainties, the h UE − h J  2 values are consistent with zero, across the entire measured  T range.The findings drawn from the  T dependence are consistent with those from the multiplicity dependence, and similarly demonstrate that the presence or absence of jets has no influence on the flow of the UE and that there are no correlations between jet-fragments and the UE.The features of the  2 values discussed above do not show any systematic variation with the jet selections, for example the  G T thresholds used in the analysis, as discussed in Ref. [31].In conclusion, this Letter studies long-range 2PCs in   collisions when rejecting tracks in the vicinity of jets, and the correlations between jet constituent tracks and tracks from the UE.The 2PCs are analyzed using a template-fit procedure, previously developed by ATLAS [15], which extracts second-order Fourier coefficients ( 2 ) of the anisotropy.These results demonstrate that the magnitude of the  2 is not affected when removing tracks associated with jets, or by the presence or absence of jets in the event.The  2 measured with correlations between jet constituents with  T < 8 GeV and UE tracks are consistent with zero within uncertainties.These features are observed both in the  2 multiplicity and  T dependence.
The observation that fragments of high- T jets in   collisions do not have measurable long-range azimuthal correlations with the UE, and that the production of  bosons [21] or jets does not significantly influence the long-range correlations between UE particles, suggest a complete "factorization" between hard-scattering processes and the physics responsible for the ridge.Further studies are needed to extend this measurement to higher  T to compare with previous measurements in +Pb collisions [22] where such factorization is broken.This Letter provides important insights into the origin of the long-range correlations observed in   collisions and offers new fundamental input to theoretical models.Multiple triggers were used to record the data and can broadly be classified into three categories.The first category of triggers selects a set of minimum-bias events through a L1 trigger that requires a signal in at least one MBTS counter, and a second trigger that requires at least one reconstructed track with  T > 0.2 GeV at the HLT.The second category is a set of high-multiplicity triggers that apply a L1 requirement on either the transverse energy ( T ) in the calorimeters or a hit in at least one MBTS counter on each side, and an HLT requirement on the multiplicity of HLT-reconstructed tracks with  T > 0.4 GeV that is associated with the reconstructed vertex with the highest multiplicity in the event.The third is a set of triggers that enhance the rate of jets.These include a set of triggers that require a jet at L1 and at the HLT, and another set of triggers that require a minimum threshold on the total  T in the calorimeter at L1.

A.2 Detailed track selection criteria
The criteria used to select tracks include the requirements of  T > 0.4 GeV and || < 2.5, a hit in the IBL or a hit in the pixel layer next to the IBL, and a minimum of six hits in the SCT.Additionally, the transverse impact parameter of the track relative to the average beam position, and the longitudinal impact parameter of the track relative to the vertex are both required to be less than 1.5 mm.To remove tracks with mis-measured  T due to interactions with the material or other effects, the track-fit  2 probability is required to be larger than 0.01 for tracks having  T > 10 GeV.

A.3 Multiplicity distributions
Figure 4 shows the distribution of  rec,corr ch for the events in the h-h, h UE −h UE (AllEvents), h UE −h UE (NoJets), h UE − h UE (WithJets), and h UE − h J cases.The default peripheral reference in the template-fits is constructed using the events in the 10-30 multiplicity interval.The average minimum-bias multiplicity of reconstructed charged-particle tracks in 13 TeV   collisions in ATLAS is ∼ 20, and the default range for the peripheral reference is taken to be a ±10 window around this mean value.Alternative peripheral references built from the 10-40 and 20-40 multiplicity intervals are used to evaluate systematic uncertainties, where the lower and upper multiplicity ranges are increased by 10 from the default interval.The 0-10 multiplicity interval, is excluded from the peripheral references as 1) the h UE − h UE (WithJets) and h UE − h J cases have very few events at these low multiplicities, and 2) such low multiplicity are likely to contain a significant contribution from diffractive events.
, and h UE − h J cases.The discontinuities in the distributions correspond to different high-multiplicity trigger thresholds.

A.4 Pythia 8 Embedding
This section motivates the choice of the 4 GeV requirement used in Eq. ( 2) to remove the bias from the UE modulation on the jet selection.The effect of UE modulation on the jet selection is estimated by a toy Pythia 8 embedding study.The Pythia 8 events are simulated using the Monash 2013 tune with multi-parton interaction off and initial-state radiation on.Jet reconstruction is performed on these generated events, using the anti- T clustering algorithm with a radius parameter of 0.4.The jets thus produced are called generated-jets hereafter.The generated events are filtered by requiring the events to have a generated-jet with  T greater than 15 GeV and a balanced generated-jet with  T > 10 GeV and |Δ| > 5/6 relative to the first jet.The generated (stable1 ) particles are embedded onto minimum bias data events.After the embedding, jets are reconstructed using the embedded particles and the original particle flow objects [38] present in the data-event, using the anti- T clustering algorithm with a radius parameter of 0.4.These jets are called embedded-jets hereafter.Similar jet clustering is also done using only the particle flow objects in the minimum-bias data used in the embedding study.These jets are called data-jets hereafter.
The embedded-jets are required to match a generated-jet with Δ = √︁ Δ 2 + Δ 2 < 0.1.The embedded-jets are also required to be separated from data-jets by Δ > 0.8.The  T of the embedded-jets is always larger than the  T of the matched generated-jet as the particle flow objects in the data always push up the  T .The issue of concern for this analysis is that this increase in the  T is dependent on the azimuthal angle that the jet makes with the second-order event-plane, due to the modulation in the UE.To visualize the bias from the UE modulation, the difference between the azimuthal angle of the embedded-jet,  jet , and the second-order event plane angle, Ψ Data 2 , is plotted in Figure 5.The Ψ 2 angle is calculated using particles from the data events only, excluding particles within one unit in  around any data-jets with  T greater than 15 GeV.Since Pythia 8 events are uncorrelated with the data event onto which they are embedded, the distribution of  jet − Ψ Data 2 is, by construction, constant for the generated-jets.However, for the embedded-jets the corresponding distribution is modulated, as shown in Figure 5. Since the modulation effects are dominated by low  T particles, grooming the jets to remove soft particles can remove the UE bias.A groomed jet  T definition is proposed here, by summing the  T of jet constituents with  T greater than a particular threshold : Figure 6 shows the  jet − Ψ Data 2 using different values of the threshold  in Eq. ( 4).The modulation in the distribution systematically decreases with increasing , and is nearly gone for  = 4 GeV.for the embedded jets.The continuous line indicates a Fourier fit to the distribution that includes a 2 nd order modulation.From left to right, the three panels in each row correspond to increasing thresholds  (in Eq. ( 4)) of 2 GeV, 3 GeV, and 4 GeV.The top (bottom) row corresponds to the 50-60 (110-120) multiplicity interval.

A.5 Illustration of the procedure for selecting h UE − h UE and h UE − h J pairs
Figures 7 and 8 illustrate the procedure for selecting the h UE − h UE and h UE − h J pairs, respectively.The blue dots represent the tracks that are treated as UE tracks (see Figure 7).The green band indicates the region scanned to find the balancing jet for jet-.In this case the balancing jet is jet-.In the absence of a balancing jet, the event is not used for the h UE − h J correlation.

A.6 Cross checks on jet selections used in the analysis
This section discusses several checks that are made to test the sensitivity of the results on the jet selections used in the analysis.
• The sensitivity of the h UE − h J  2 measurements to the  G T selection is evaluated by varying the  G T selection threshold from the nominal value of 40 GeV to 35 GeV and 50 GeV.
• The sensitivity of the  2 to the  T threshold applied on the jet constituents in Eq. ( 2) is checked by raising it from its default value of 4 GeV to 4.5 GeV.
• A threshold of  G T = 15 GeV is applied to separate the hard and soft processes in the event.This threshold is an analysis choice as there is no  T value that fully separates hard and soft processes.The sensitivity of the results to the choice of the jet  G T is investigated by increasing the threshold from its nominal value of 15 GeV to 20 GeV.
• As stated in the Letter as isolation requirement is imposed on the  G T > 40 GeV jets in the h UE − h J correlations.This isolation requirement requires that there are no  G T > 15 GeV jets within Δ = 1 of the  G T > 40 GeV jets used in the h UE − h J correlation analysis.As a cross-check the measurements are repeated with this isolation requirement removed.
For all these cases no significant variation is observed in the measurements, and the results with these variations are consistent with the nominal ones within the quoted systematic and statistical uncertainties.

40 Figure 1 :
Figure 1: Template-fits to the two-particle correlations in Δ.Events with 10 ≤  rec,corr ch < 30 are used as the peripheral reference.The solid points indicate the measured 2PC, the open circles show the scaled and shifted peripheral reference, and the continuous line shows the fit.The dashed line shows the second-order harmonic component, and the dotted line shows the pedestal of the fit shifted up by  periph (0).The top row corresponds to different multiplicity intervals for the h UE − h J class.The left, center and right panels in the bottom row correspond to the h-h, h UE − h UE (NoJets), and h UE − h UE (WithJets) classes, respectively, for the 40-150 multiplicity interval.

Figure 2 :
Figure 2: The multiplicity dependence of  2 for 2 < |Δ| < 5. Events with 10 ≤  rec,corr ch < 30 are used as the peripheral reference.Jets with  G T > 15 GeV are used to classify the h UE − h UE (NoJets) and h UE − h UE (WithJets) samples.The data point for the h UE − h UE (WithJets) case has a particularly large statistical uncertainty in the 40-50 multiplicity interval and is not shown.The data-points for the h UE − h UE (AllEvents), h UE − h UE (NoJets), and h UE − h UE (WithJets) samples are slightly shifted along the -axis for clarity.The error bars and bands correspond to statistical and systematic uncertainties, respectively.

Figure 3 :
Figure 3: The  b T dependence of the  2 obtained for the 40-150 multiplicity interval for 2 < |Δ| < 5. Events with 10 ≤  rec,corr ch < 30 are used as the peripheral reference.Jets with  G T > 15 GeV are used to classify the h UE − h UE (NoJets) and h UE − h UE (WithJets) samples.The data-points for the h-h sample are drawn at the nominal values while the data-points for the h UE − h UE (AllEvents), h UE − h UE (NoJets), and the highest  b T point of h UE − h J samples are shifted slightly for clarity.The error bars and bands correspond to statistical and systematic uncertainties, respectively.

Figure 4 :
Figure 4: The distribution of  rec,corr ch

FitFigure 5 :
Figure 5: The distribution of  jet − Ψ Data 2 for the embedded jets.The continuous line indicates a Fourier fit to the distribution that includes a 2 nd order modulation.The different panels correspond to different multiplicity intervals.

Figure 7 :
Figure 7: Conceptual depiction of the procedure to measure the h UE − h UE correlations.The  and -axes depict  and , respectively.The ATLAS calorimeter system extends over ±4.9 in , over which range the jets are reconstructed.The vertical dashed lines at  = ±2.5 indicate the acceptance of the ATLAS ID, over which charged-particle tracks are reconstructed.The figure shows an event with three  G T > 15 GeV jets (grey circles).The vertical yellow bands indicate a ±1  window around the jets.The reconstructed tracks within the yellow bands (grey dots) are excluded from the h UE − h UE correlation study.The blue dots represent the tracks that are used in the h UE − h UE correlation.

Figure 8 :
Figure 8: Conceptual depiction of the procedure to measure the h UE − h J correlations.The  and -axes depict  and , respectively.The ATLAS calorimeter system extends over ±4.9 in , over which range the jets are reconstructed.The vertical dashed lines at  = ±2.5 indicate the acceptance of the ATLAS ID, over which charged-particle tracks are reconstructed.The figure shows an event with two  G T > 15 GeV jets (grey circles, labelled as  and ), and a  G T > 40 GeV jet (red circle, labelled as ).The red dots represent the tracks that are considered as jet constituents (h J ) of the  G T > 40 GeV jet.The vertical yellow bands indicate a ±1  window around the jets.The blue dots represent the tracks that are treated as UE tracks (see Figure7).The green band indicates the region scanned to find the balancing jet for jet-.In this case the balancing jet is jet-.In the absence of a balancing jet, the event is not used for the h UE − h J correlation.