Cold nuclear matter physics at forward rapidities from d+Au collisions in PHENIX
aa r X i v : . [ nu c l - e x ] S e p Cold nuclear matter physics at forward rapiditiesfrom d + Au collisions in PHENIX
M Chiu (for the PHENIX Collaboration ‡ ) Brookhaven National Lab, Upton, NY, 11973 USAE-mail: [email protected]
Abstract.
We present measurements by the PHENIX experiment at RHIC of di-hadron pair production in d +Au collisions where the particles in the pair are variedacross a wide range of pseudorapidity, out to η = 3 .
8. With di-hadrons, varyingthe p T and rapidity of the particles in the di-hadron pair allows studying any effectsas a function of partonic x in the nucleus. These di-hadron measurements mightprobe down to parton momentum fractions x ∼ − in the gold nucleus, where theinteresting possibility of observing gluon saturation effects at RHIC is the greatest. Ourmeasurements show that the correlated yield of back-to-back pairs in d +Au collisionsis suppressed by up to an order of magnitude relative to p + p collisions, and increaseswith greater nuclear path thickness and with a selection for lower x in the Au nucleus.PACS numbers: 21.65.Qr,25.75.-q,25.75.Bh
1. Introduction
Deuteron-gold collisions at RHIC provide a means to explore nuclear effects on theinitial-state parton densities in the nucleus, which is vitally important for understandingthe baseline production in heavy-ion collisions. RHIC experiments have shown thatsingle inclusive hadron yields in the forward (deuteron) rapidity direction for √ s NN =200 GeV d +Au collisions are suppressed relative to p + p collisions [1, 2, 3]. Themechanism for the suppression has not been firmly established. Many effects havebeen proposed for this suppression, such as gluon saturation [4, 5], initial state energyloss [6, 7], parton recombination [8], multi-parton interactions [9], and leading andhigher-twist shadowing [10, 11].One set of measurements that might help to distinguish between the competingmodels is forward azimuthally correlated di-hadron correlation functions, which directlyprobe di-jet production through their 2 → φ = π . This techniquehas been used extensively at RHIC and is described in detail elsewhere [12, 13, 14].The di-hadron results presented here were obtained from p + p and d +Au runs in 2008with the PHENIX detector and include a new electromagnetic calorimeter, the MuonPiston Calorimeter (MPC), with an acceptance of 3 . < η < . ‡ A list of members of the PHENIX Collaboration can be found at the end of this issue. orward di-hadron measurements in d + Au collisions with PHENIX < φ < π . Di-hadron measurements can probe more precise ranges of parton x in agold nucleus than do single hadron probes (e.g., R dA ). At forward rapidities, a singlehadron probe will cover a very broad range of x, 10 − < x Au < .
5, thus mixing togetherthe shadowing, anti-shadowing, and even EMC effects [10]. Azimuthally correlated di-hadron measurements also enhance the di-jet fraction in the event selection, since oneselects only the back-to-back hadrons. (pp) = 0.084 b (dA 60-88) = 0.134 b (dA 0-20) = 0.180 b ˜ (pp) = 0.086 b (dA 60-88) = 0.133 b (dA 0-20) = 0.185 b ˜ = 0.088 b = 0.135 b = 0.184 b p+pd+Au 60-88d+Au 0-20 0.5-0.75 GeV/c ˜ (rad) fD f w d h d T f w d ) dp fD N / d ( d m i d / N -1 0 1 2 3 400.020.040.060.080.1 = 0.059 b = 0.106 b = 0.142 b ˜ = 0.079 b = 0.127 b = 0.163 b ˜ = 0.095 b = 0.141 b = 0.176 b p+pd+Au 60-88d+Au 0-20 0.5-0.75 GeV/c ˜ (rad) fD ph d T p ) dp fD N / d ( d c l u s / N Figure 1.
Background-subtracted mid-forward rapidity π - π (top) and forward-forward cluster- π (bottom) per-trigger correlation functions, for p + p (open points), d +Au peripheral (60-88%, triangles), and d +Au central collisions (0-20%, solid points)at √ s NN = 200 GeV. The trigger p T ranges from 1.1 to 7 GeV/c and the associated π ’s have p T = 0 . − .
75 GeV/c. Systematic errors of up to 30% on the near side( | ∆ φ | < .
5) have not been shown. The subtracted pedestals, b , are shown for eachcase. By performing several correlation measurements with particles at different p T andrapidities, one can systematically scan different x ranges with an observable that isenhanced for the leading-order perturbative QCD component. Probing the x dependenceof the effect is an important test since most models predict that any effects should bestronger at smaller x . Particles at higher pseudorapidities are produced from smaller x ,so measuring hadrons from more forward rapidities should probe smaller x.
2. PHENIX MPC d + Au di-Hadron Correlations
For this analysis, back-to-back π - π or hadron- π pairs are measured with one particleat mid-rapidity, and the other at forward rapidity. Back-to-back cluster- π pairs are alsomeasured where both are in the forward rapidity region. The clusters are reconstructedfrom the energy deposit of photons in the MPC, and are estimated to be at least 80% orward di-hadron measurements in d + Au collisions with PHENIX π ’s, with the remainder coming from single photons from decays of η ’sand from direct photons. Further details of the analysis are available in [14].As shown in figure 1, the away-side peak for d +Au central collisions appearssignificantly suppressed compared to p + p collisions and peripheral d +Au collisions.This effect is large for the mid-forward di-hadron correlations and becomes even largerfor the forward-forward correlations. Within large errors, the Gaussian widths of theaway-side correlation peak for the mid-forward di-hadron correlations remain the samebetween p + p and central d +Au. For the forward-forward case, uncertainties in thepedestal level from the underlying event and the strong suppression of the away-sidepeak make extracting the width unreliable. For this case, the away side peak width incentral d +Au collisions is allowed to vary up to twice as much as in p + p when accountingfor this systematic uncertainty. Au or x Aufrag x -4 -3 -2 -1
10 1 A u g o r R d A J -1 Figure 2. J dA versus x fragAu for peripheral (60-88%, in red) and central (0-20%, inblue) d +Au collisions at √ s NN = 200 GeV, and compared to the EPS09 LO R Aug ( x Au )curves at a scale Q = 4 GeV [16]. The observed suppression is quantified by studying the relative yield, J dA [15], ofcorrelated back-to-back hadron pairs in d +Au collision compared to p + p collisions scaledwith the average number of binary nucleon collisions h N coll i , J dA = 1 h N coll i σ pairdA /σ dA σ pairpp /σ pp . (1)This is simply the analog of the usual nuclear modification factor R dA but for hadronpairs. The σ are the p + p or d +Au inelastic cross-sections, while σ pair is the cross-sectionfor di-hadron pair production. The J dA is calculated using the correlated away sidepeak after subtracting the pedestal b . J dA decreases with increasing number of binarycollisions, h N coll i , or equivalently with increasing nuclear thickness. The suppressionalso increases with decreasing particle p T and is significantly larger for forward-forwardhadron pairs than for mid-forward pairs. The observed suppression of J dA versus nuclear orward di-hadron measurements in d + Au collisions with PHENIX p T and η points to large cold nuclear matter effects arising at low partonmomentum fractions x in the nucleus probed by the deuteron. This trend is seen moreclearly in Fig. 2 where J dA is plotted versus x fragAu = ( h p T i e −h η i + h p T i e −h η i ) / √ s NN for all pair selections in η and p T . In the case of 2 → x fragAu is lower than x Au by the mean fragmentation fraction, h z i , of the struckparton in the Au nucleus. Since x fragAu is an entirely experimental defined quantity, itshould be reproducable in any theoretical framework.
3. Discussion
In a leading order pQCD picture, the variable J dA is J dA = σ pairdA /σ dA h N coll i σ pairpp /σ pp ≈ f ad ( x ad ) ⊗ f bAu ( x bAu ) ⊗ ˆ σ ab → cd ⊗ D ( z c , z d ) h N coll i f ap ( x ap ) ⊗ f bp ( x bp ) ⊗ ˆ σ ab → cd ⊗ D ( z c , z d ) (2)for partons a+b going to outgoing jets c+d, which then fragment to hadrons withlongitudinal fractions z c , z d . In the above convolutions over p+p and d+A, most of theterms are expected to be roughly similar between p+p and d+Au except for the nucleargluon pdf. Naively, J dA might be largely dominated by the modification to the nucleargluon parton distribution function (pdf’s), since most of the events with di-hadrons atforward rapidities consist of a high-x parton from the deuteron and a low-x gluon fromthe gold nucleus. Assuming this to be true then J dA ∼ R Aug = G Au ( x, Q ) /A G p ( x, Q )In figure 2 the J dA values are overlaid with the EPS09 R Aug curves [16]. The J dA valuesfor the peripheral bins are above the best fit EPS09, while the central bins are below.The EPS09 curves are taken largely from nuclear deep inelastic scattering and representan averaged value of R Aug over all centralities. The J dA values for the most central binat the lowest x are well below the EPS09 curves. This is qualititatively consistent withthe expectations for the Color-Glass Condensate [4], which posits an extreme form ofshadowing due to the onset of gluon saturation.If nature is kind and this data can be interpreted in terms of a simple LOpQCD picture, it may be possible to extract R Aug , which is extremely important forunderstanding the quark gluon plasma since it forms the baseline for production inheavy ion collisions. In addition, if the large suppression of J dA observed in central d +Au collisions is from gluon saturation, it may be possible to study the dependenceof that saturation on the thickness of the nucleus. One possible test of whether theseideas are correct would be to use extractions of R Aug from this data to predict J/ Ψ datain d +Au collisions from PHENIX. References [1] I. Arsene et al. [BRAHMS Collaboration], Phys. Rev. Lett. (2004) 242303[2] J. Adams et al. [Star Collaboration], Phys. Rev. Lett. (2006) 152302[3] S. Adler et al. [PHENIX Collaboration], Phys. Rev. Lett. (2005) 082302[4] L. D. McLerran and R. Venugopalan, Phys. Rev. D (1994) 3352[5] D. Kharzeev, E. Levin and L. McLerran, Nucl. Phys. A (2005) 627 orward di-hadron measurements in d + Au collisions with PHENIX [6] I. Vitev Phys. Rev. C (2007) 064906[7] L. Frankfurt and M. Strikman, Phys. Lett. B (2007) 412[8] R. C. Hwa, C. B. Yang and R. J. Fries, Phys. Rev. C (2005) 024902[9] M. Strikman and W. Vogelsang, Phys. Rev. D , 034029 (2011)[10] V. Guzey, M. Strikman and W. Vogelsang, Phys. Lett. B , 173 (2004)[11] J. w. Qiu and I. Vitev, Phys. Lett. B (2006) 507[12] S. Adler et al. [PHENIX Collaboration], Phys. Rev. C. (2006) 054903[13] S. Adler et al. [PHENIX Collaboration], Phys. Rev. Lett. (2006) 222301[14] A. Adare et al. [PHENIX Collaboration], Submitted to Phys. Rev. Lett. Preprint nucl-ex:1105.5112[15] A. Adare et al. [PHENIX Collaboration], Phys. Rev. C. (2008) 014901[16] K. J. Eskola, H. Paukkunen and C. A. Salgado, JHEP (2009) 065 [arXiv:0902.4154 [hep-ph]]. ufrag x -3 -2 d A J -1
10 1
Tfwd d+Au 60-88 p0.5-0.75 GeV/c0.75-1.0 GeV/c1.0-1.5 GeV/c