Heavy Flavor Production and Energy Loss with Two-Particle Correlations at PHENIX
HHeavy Flavor Production and Energy Loss with Two-ParticleCorrelations at PHENIX
Tatia Engelmore for the PHENIX Collaboration
Columbia University, New York, New York 10027, USA and Nevis Laboratories, Irvington, New York 10533, USA
Abstract
Heavy quarks are a valuable probe of the hot, dense medium created in a heavy ion collision,and are an important test of proposed mechanisms of energy loss. It was discovered that singlenon-photonic electrons are suppressed at a similar level to light hadrons, implying a comparablelevel of energy loss between light and heavy partons. Because theory has had a di ffi cult timeexplaining the level of heavy quark energy loss, it is crucial to better understand charm andbottom suppression. Electron-hadron correlations have been used at PHENIX to study heavyflavor in both p + p and Au + Au collisions. In p + p the ratio of charm to bottom production hasbeen measured using mass correlations through a partial reconstruction of the D meson. Electron-hadron angular correlations have also been used to study medium modification of heavy flavor,and we see hints of energy loss e ff ects. A complementary study of correlated electron-muon pairsprovides a clean measurement of heavy flavor production in a rapidity range not yet studied.
1. Introduction
Two-particle correlation studies at RHIC have yielded valuable information on the interactionof jets produced in nucleus-nucleus collisions with the hot nuclear matter. These correlations,when compared against the baseline from proton-proton collisions, show powerful evidence ofthe modification of jet structure within the medium [1]. Until now, jet interactions at PHENIXhave been studied using correlations with light hadrons; a natural next step, then, is to understandthe medium modification of heavy-flavor jets. The suppression of single leptons from heavyflavor decay has been shown to be similar to the suppression of π [2], meaning that heavy quarkenergy loss in the medium is significant. Therefore it is of interest to determine the e ff ects of themedium on the structure of heavy flavor jets.PHENIX measures charm and bottom mesons mainly through the electron and muon decayproducts. In order to isolate double heavy flavor decays we have used correlations of electronsand hadrons coming from the same c ¯ c or b ¯ b pair. The heavy flavor electrons are isolated througha statistical subtraction of the background. Mass correlations are used to give us a measure of theratio of charm to bottom produced in the collision, while the correlation of e - h in azimuthal angleallows us to study jet properties of heavy flavor. Results from the PHENIX e - h correlations in p + p and in Au + Au are now presented. A complementary analysis involves electron-muon corre-lations where both leptons are charm or bottom meson decay products. Although the statistics arelow, there is a clean heavy flavor signal. Results are shown from the analysis in p + p collisions.Electron-muon correlations in d + Au will also be interesting in order to probe the modificationof the gluon structure function in heavy ions.
Preprint submitted to Nuclear Physics A October 30, 2018 a r X i v : . [ nu c l - e x ] S e p . Electron-Hadron Correlations Because the majority of electrons measured at PHENIX are due to photonic sources, theheavy flavor signal must be isolated through a statistical subtraction. What we measure is theinclusive yield of hadrons per trigger electron, Y e inc − h , which is proportional to both the heavyflavor and the photonic per-trigger yield: Y e inc − h = N e HF Y e HF − h + N e phot Y e phot − h N e HF + N e phot (1)From this we find the heavy flavor yield: Y e HF − h = ( R HF + Y e inc − h − Y e phot − h R HF (2)where R HF = N eHF N ephot which was found in [3].The important quantity to determine is Y e phot − h . This yield is primarily due to photonic elec-trons arising from the decay, π → γγ . These decay photons then convert to electrons in thedetector elements. Two methods are used to determine the photonic electron-hadron yield. Inthe first, the measured PHENIX inclusive photon spectrum is used as the input to a GEANTsimulation; these photons are predominantly from Dalitz decays. The electrons resulting fromconversions are then reconstructed, and the mapping between the photon p T and the electron p T is determined. For the second method, the measured π spectrum is input into a Monte Carlosimulation, and the correspondence between the decay γ p T and the p T of the electron resultingfrom the photon’s conversion is found. The relationships between the photons and their resultantelectrons are used to determine weights in order to construct the Y e phot − h from the Y γ inc − h : Y e phot − h ( p T , i ) = (cid:88) j w i ( p T , j ) Y γ inc − h ( p T , j ) (3)The e inc − h and γ − h correlations were then measured, and converted through the subtractionin Eq. 2 to the heavy flavor per-trigger yield. The angular correlations were then corrected forthe restricted PHENIX acceptance using mixed events, and the combinatoric background wasremoved using a ZYAM subtraction [4]. Fig. 1(a) shows the correlation functions for a trigger of2 . < p T < . / c with various associated p T bins. Also, Fig. 1(b) shows I AA , which is theratio of the per-trigger e − h pair yield in Au + Au to that in p + p . Results are shown for threeregions in ∆ φ , corresponding to (from top) the near side, the head and shoulder combined, andthe head alone (for definitions see [5]). A larger I AA for the head and shoulder region combined,rather than the head alone, gives a hint of a shoulder appearing in heavy flavor correlations.Larger statistics are needed, though, to make this result conclusive.While e - h angular correlations measure inclusive heavy flavor, mass correlations have beenused to separate charm from bottom. This was found by partially reconstructing D mesons, andthe ratio of charm to bottom as a function of p T was obtained [6].
3. Electron-Muon Correlations
Electron-muon correlations have long been proposed as a clean signal of charm pair produc-tion in heavy ion collisions [7]. Because the heavy flavor signal consists of opposite-sign e - µ !" ! " dd N t r i g N <1.0GeV/c T,h !" <1.5GeV/c T,h !" <2.0GeV/c T,h !" <2.5GeV/c T,h !" ! " dd N t r i g N <3.0GeV/c T,h !" <3.5GeV/c T,h !" <4.0GeV/c T,h < 3.0GeV/c
T,e - hadron: 2.0 < p ± heavy flavor e = 200GeVs Au+Au, 0-60%p+p (a) ( - . r a d ) AA I - h, Au+Au ± heavy flavor ePHENIX PRELIMINARY r a d ) ! ( . - AA I < 3.0 GeV/c T,e (GeV/c)
T,hadron p r a d ) ! ( . - AA I (b)Figure 1: Results for e - h correlations in Au + Au. At left, angular correlations in both p + p (black squares) and Au + Au(blue circles). At right, I AA (ratio of per-trigger yield in p + p to that in Au + Au) for 2 . < p T , e < . / c as a functionof hadron p T . Top row corresponds to near side, middle to combined head and shoulder region, and bottom to headregion alone. pairs rather than dielectrons or dimuons, there are no backgrounds from other physics processessuch as Drell-Yan, thermal production, or resonance decays. We are sensitive to the productionof back-to-back charm pairs, mainly via gluon fusion [8]. Since the signal is solely opposite-sign pairs, the majority of the background from combinatorics, light meson decays and photonicelectrons is removed with a like-sign subtraction.The remaining backgrounds that need to be accounted for are due to standard backgrounds inthe muon detectors [9]. These include hadrons that survive the front muon absorber and appearas muons, as well as muons that arise from light meson decay rather than from heavy flavor.While neither of these contribute significantly in the signal region ( ∆ φ = π ), they a ff ect the levelof flat background to the angular correlation and must be subtracted. Correlations of electronswith “punch-through” hadrons were measured using tracks that stop within the PHENIX muonidentifier. For these tracks a clear muon stopping peak in the longitudinal momentum spectrumwas seen. Hadron tracks are those that fall outside of this peak. To separate the correlationsinvolving decay muons, a comparison of muon tracks with an event vertex near to and far fromthe detector was made. This is because if light hadrons are created far from the detector, they havemore of a chance to decay. The contributions from both of these backgrounds were subtractedfrom the inclusive signal, and the result is shown in Fig. 2.
4. Conclusions and Outlook
The study of heavy flavor jet physics is important for understanding how heavy quarks propa-gate through the medium created in a heavy ion collision. Heavy quarks have been demonstratedto lose similar amounts of energy as light mesons and are also shown to have elliptic flow [10],so the modification of their jet structure is a crucial piece of information. It will be interesting3 igure 2: Electron-muon correlations from heavy flavor in p + p . Grey band corresponds to error from decay muonsubtraction, blue boxes to error from “punch-through” hadron subtraction. to compare the shape of these correlations to the structures found in hadron-hadron correlations,for example the shoulder and the ridge.The electron-hadron analysis requires a large background subtraction, so it was di ffi cult toextract a clear measure of the signal. With the addition of the silicon vertex upgrades to PHENIX,though, it will be easier to determine the heavy flavor signal and this measurement will becomemuch cleaner. Electron-muon correlations are limited by statistics, though this signal will beeasier to measure with increased luminosity. We have shown here a measurement of the p + p baseline, which is necessary for future studies of these correlations in heavy ion measurements.A study of the cold-matter e ff ects on e - µ correlations in d + Au is underway.
References [1] A. Adare, et. al.,
Phys. Rev. C
78, 014901 (2008)[2] A. Adare, et. al.,
Phys. Rev. Lett
98, 172301 (2007)[3] A. Adare, et. al.,
Phys. Rev. Lett
97, 252002 (2006)[4] J. Jia,
Nuclear Phys. A , 783, 501c-506c (2007)[5] A. Adare, et. al.,
Phys. Rev. C
77, 011901 (2008)[6] A. Adare, et. al., [arXiv:0903.4851][7] S. Gavin, et. al.,
Phys. Rev. C
54, 2606 (1996)[8] I. Vitev, et. al.,
Phys. Rev. D
74, 054010 (2006)[9] S. Adler, et al.,
Phys. Rev. D
76, 092002 (2007)[10] A. Adare, et. al,
Phys. Rev. Lett
98, 172301 (2007)98, 172301 (2007)