betaFeedback

WorksPhysics and Astronomy

Mini-jet Clustering Algorithm Using Transverse-momentum Seeds in High-energy Nuclear Collisions

Abstract

We propose an algorithm to detect mini-jet clusters in high-energy nuclear collisions, by selecting a high-transverse-momentum (pTp_T) particle as a seed and assigning a clustering radius (RR) in the pseudorapidity and azimuthal-angle space. Our PYTHIA simulations for pp+pp collisions show that a scheme with a seeding pTp_T of around 0.5 GeV/cc and RR of approximately 0.6 satisfactorily identifies mini-jet clusters. The correlation between clusters obtained in PYTHIA calculations using the algorithm exhibits the proper behavior of hard-scattering-like processes, suggesting its usefulness in isolating mini-jet-like clusters from non-hard-scattering soft processes when applied to actual nuclear-collision data, thereby allowing a closer examination of both the mini-jet and the soft mechanisms.

Introduction

The hard-scattering process was originally proposed as the primary mechanism for producing high-transverse-momentum (pTp_T) jet clusters, typically in the range of tens of GeV/cc [1, 2, 3, 4, 5, 6, 7]. However, the UA1 Collaboration found that the production of particle clusters with a total pTp_T of a few GeV/cc is an important production process in hadron-hadron collisions, representing a sizeable fraction of the particle production cross-sections [8]. Jet clusters in such a pTp_T range were coined “mini-jet” [9]. The dominance of jet production extends to even lower-pTp_T domains at higher collision energies (s\sqrt{s}) because (i) the fraction of particles produced by such a process increases rapidly with s\sqrt{s}, and (ii) the jet-production invariant cross-section at midrapidities varies inversely with pTp_T [9, 10, 11, 12]. In fact, the mini-jet dominance has been found to extend to the pTp_T region of a few tenths of a GeV/cc in high-energy p+pp+p and p+pˉp+\bar{p} collisions at s=0.9\sqrt{s}=0.9 to 7 TeV [13, 14].

While the pQCD-type hard-scattering process remains a key source of particle production, soft processes contribute to the majority of low-pTp_T particles with distinct outcomes such as rings, pancakes, granular droplets, hydrodynamical collective motions, quark-gluon plasma, and color-glass condensate [15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45]. To better understand the underlying reaction processes for such production mechanisms, it is necessary to devise a method to distinguish between the part produced by mini-jet clusters and the part produced by soft processes. Isolating the mini-jet clusters by such a method will not only benefit the study of the mini-jet dynamics but also help to identify the soft part for further examination.

Previously, as a first step to identify mini-jets, we studied the clustering properties of produced particles in high-energy p+pp+p collisions in the space of pseudorapidity (η\eta) and azimuthal angle (ϕ\phi). We devised an algorithm to find mini-jet-like clusters by using the kk-means clustering method in conjunction with a kk-number (cluster number) selection principle [46]. To study the clustering algorithm, we examined minimum-bias events of p+pp+p collisions at s=200\sqrt{s}=200 GeV generated by PYTHIA8.1 [47]. Our findings indicate that multiple mini-jet-like and mini-dijet-like clusters of low-pTp_T hadrons can arise in high-multiplicity events. However, comparable clustering behavior is also evident for randomly produced particles confined to a finite η\eta and ϕ\phi space. Therefore, the ability to discern azimuthally back-to-back correlated mini-jet-like clusters as bona fide mini-jets and mini-dijets will rely on further correlation requirements.

A mini-jet, as a cluster of particles originating from a high-pTp_T jet cascade, is likely to contain some high-pTp_T particles. Thus, it is reasonable to use a high-pTp_T particle as a seed for identifying mini-jets. To fix the mini-jet cluster number KK and also to eliminate the noise that may contaminate the mini-jet clusters, we add a supplementary condition that requires each mini-jet to contain at least one seed particle with pTp_T greater than a certain threshold value, pT0p_{T0}. The fixed cluster number and the knowledge of the mini-jet size in η\eta and ϕ\phi bring us closer to understanding the mini-jet and their correlations. The remaining particles can be attributed to non-jet and non-minijet processes, which form the bulk part of the dynamics.

Clearly, the higher pT0p_{T0}, the more the mini-jet cluster will resemble the conventional high-pTp_T jet, and the higher the chance of detecting its dijet partner. However, we focus on mini-jets in the very low-pTp_T domain because they are more abundant in the underlying events. We analyze the clustering properties of mini-jets as a function of the seed pT0p_{T0} value.

To ensure that the reconstructed clusters are mini-jet clusters, we rely on the inter-cluster correlations. Analytical expressions for the hard scattering process summarize its essential features and dependencies, making it easier to uncover dynamical effects wherever they may occur. These analyses have revealed insights into the dominance of the hard-scattering process in the high-pTp_T domain and have helped to locate the boundary between the hard-scattering and flux-tube fragmentation processes in p+pp+p collisions at high energies [14].

In Section II, we first derive the differential cross-section EcEκdσ(AB→cκX)/dc dκE_cE_\kappa d\sigma(AB \to c\kappa X)/dc\,d\kappa for the production of two massive partons cc and κ\kappa. This allows us to integrate out pTp_T to acquire the two-particle correlation function dσ/dΔϕ dΔyd\sigma/d\Delta\phi\,d\Delta y in Section III, where Δϕ=ϕκ−ϕc\Delta\phi=\phi_\kappa-\phi_c and Δy=yκ−yc\Delta y=y_\kappa-y_c. This correlation function exhibits the ”ridge” structure on the away side at Δϕ≈±π\Delta\phi\approx\pm\pi in the hard-scattering process. Next, in Section IV, we present the new clustering algorithm and demonstrate its application to PYTHIA data of p+pp+p collisions. In Section V, we compare the two-cluster correlations obtained from the algorithm with the two-particle correlations. Additionally, in Section VI, we discuss how the anti-kTk_T algorithm complements our new clustering algorithm. Finally, we draw brief conclusions in Section VII.

Hard Scattering Integral for EcEκdσ(AB→cκX)/dc dκE_cE_\kappa d\sigma(AB \to c\kappa X)/dc\,d\kappa

In the parton model, the hard-scattering cross-section for AB→cκXAB \to c\kappa X, the production of partons cc and κ\kappa with momenta cc and κ\kappa, is given by [48, 49]

dσ(AB→cκX)=∑ab∫KabdxadaTdxbdbT×Ga/A(xa,aT)Gb/B(xb,bT)dσ(ab→cκ)(1)d\sigma(AB \to c\kappa X)=\sum_{ab}\int K_{ab}dx_a da_T dx_b db_T \times G_{a/A}(x_a,a_T)G_{b/B}(x_b,b_T)d\sigma(ab \to c\kappa) \tag*{(1)}

where (xa,aT)(x_a,a_T) and (xb,bT)(x_b,b_T) represent the momenta, and Ga/AG_{a/A} and Gb/BG_{b/B}, the structure functions of the incident partons aa and bb, respectively. KabK_{ab} is a correction factor that can be obtained perturbatively [50] or approximated nonperturbatively [51]. The quantity dσ(ab→cκ)d\sigma(ab \to c\kappa) is the cross-section element for the process ab→cκab \to c\kappa,

dσ(ab→cκ)=14[(a⋅b)2−ma2mb2]1/2∣Tfi∣2×d3c(2π)32Ecd3κ(2π)32Eκ(2π)4δ4(a+b−c−κ).(2)d\sigma(ab \to c\kappa)=\frac{1}{4[(a\cdot b)^2-m_a^2m_b^2]^{1/2}}|\mathcal{T}_{fi}|^2 \times\frac{d^3c}{(2\pi)^3 2E_c}\frac{d^3\kappa}{(2\pi)^3 2E_\kappa}(2\pi)^4\delta^4(a+b-c-\kappa). \tag*{(2)}

Here, we normalize the Dirac fields by uˉu=2m\bar{u}u=2m. The quantity ∣Tfi∣2|\mathcal{T}_{fi}|^2 is related to dσ/dtd\sigma/dt by

∣Tfi∣2=16π[s^−(ma+mb)2][s^−(ma−mb)2]dσ(ab→cκ)dt.|\mathcal{T}_{fi}|^2=16\pi[\hat{s}-(m_a+m_b)^2][\hat{s}-(m_a-m_b)^2]\frac{d\sigma(ab\to c\kappa)}{dt}.

We consider the simplified case with ma=mb=0m_a=m_b=0 and treat aTa_T and bTb_T as small perturbations. The cross-section element is then

dσ(ab→cκ)=s^ab2πdσ(ab→cκ)dtd3cEcd3κEκδ4(a+b−c−κ).(3)d\sigma(ab\to c\kappa)=\frac{\hat{s}_{ab}}{2\pi}\frac{d\sigma(ab\to c\kappa)}{dt}\frac{d^3c}{E_c}\frac{d^3\kappa}{E_\kappa}\delta^4(a+b-c-\kappa). \tag*{(3)}

where s^=sab=(a+b)2\hat{s}=s_{ab}=(a+b)^2, which is different from s=sAB=(A+B)2s=s_{AB}=(A+B)^2. We obtain

EcEκdσ(AB→cκX)d3c d3κ=∑ab∫KabdxadaTdxbdbT×Ga/A(xa,aT)Gb/B(xb,bT)s^dσ(ab→cκ)2πdtδ4(a+b−c−κ).(4)\frac{E_cE_\kappa d\sigma(AB\to c\kappa X)}{d^3c\,d^3\kappa} =\sum_{ab}\int K_{ab}dx_a da_T dx_b db_T \times G_{a/A}(x_a,a_T)G_{b/B}(x_b,b_T)\frac{\hat{s}d\sigma(ab\to c\kappa)}{2\pi dt}\delta^4(a+b-c-\kappa). \tag*{(4)}

We consider a factorizable structure function with a Gaussian intrinsic transverse momentum distribution,

Ga/A(xa,aT)=Ga/A(xa)12πσ2e−aT2/2σ2,(5)G_{a/A}(x_a,a_T)=G_{a/A}(x_a)\frac{1}{2\pi\sigma^2}e^{-a_T^2/2\sigma^2}, \tag*{(5)}

where 2σ2∼0.9  (GeV/c)22\sigma^2\sim0.9\;(\mathrm{GeV}/c)^2 [52]. Upon integrating over the transverse momenta aTa_T and bTb_T, we reach

dσ(AB→cκX)dyc pTc dpTc dϕc dyκ pTκ dpTκ dϕκ=∑ab∫Kabdxadxb×Ga/A(xa)Gb/B(xb)e−(cT+κT)2/4σ22(2πσ2)s^2πdσ(ab→cκ)dt×δ(a0+b0−(c0+κ0))δ(az+bz−(cz+κz)).(6)\frac{d\sigma(AB\to c\kappa X)}{dy_c\,p_{Tc}\,dp_{Tc}\,d\phi_c\,dy_\kappa\,p_{T\kappa}\,dp_{T\kappa}\,d\phi_\kappa} =\sum_{ab}\int K_{ab}dx_a dx_b \times G_{a/A}(x_a)G_{b/B}(x_b)\frac{e^{-(\mathbf{c}_T+\mathbf{\kappa}_T)^2/4\sigma^2}}{2(2\pi\sigma^2)} \frac{\hat{s}}{2\pi}\frac{d\sigma(ab\to c\kappa)}{dt} \times\delta(a_0+b_0-(c_0+\kappa_0))\delta(a_z+b_z-(c_z+\kappa_z)). \tag*{(6)}

To perform the integration over xax_a and xbx_b, we spell out the momenta in the infinite momentum frame,

a=(xas2+a2+aT22xas, aT, xas2−a2+aT22xas),(7)a=\left(x_a\frac{\sqrt{s}}{2}+\frac{a^2+a_T^2}{2x_a\sqrt{s}},\,a_T,\,x_a\frac{\sqrt{s}}{2}-\frac{a^2+a_T^2}{2x_a\sqrt{s}}\right), \tag*{(7)}
b=(xbs2+b2+bT22xbs, bT, −xbs2+b2+bT22xbs),(8)b=\left(x_b\frac{\sqrt{s}}{2}+\frac{b^2+b_T^2}{2x_b\sqrt{s}},\,b_T,\,-x_b\frac{\sqrt{s}}{2}+\frac{b^2+b_T^2}{2x_b\sqrt{s}}\right), \tag*{(8)}
c=(xcs2+c2+cT22xcs, cT, xcs2−c2+cT22xcs),(9)c=\left(x_c\frac{\sqrt{s}}{2}+\frac{c^2+c_T^2}{2x_c\sqrt{s}},\,c_T,\,x_c\frac{\sqrt{s}}{2}-\frac{c^2+c_T^2}{2x_c\sqrt{s}}\right), \tag*{(9)}
κ=(xκs2+κ2+κT22xκs, κT, −xκs2+κ2+κT22xκs)(10)\kappa=\left(x_\kappa\frac{\sqrt{s}}{2}+\frac{\kappa^2+\kappa_T^2}{2x_\kappa\sqrt{s}},\,\kappa_T,\,-x_\kappa\frac{\sqrt{s}}{2}+\frac{\kappa^2+\kappa_T^2}{2x_\kappa\sqrt{s}}\right) \tag*{(10)}

where xcx_c and xκx_\kappa can be represented by ycy_c and yκy_\kappa, respectively

xc=mTceycs,xκ=mTκeyκs.(11)x_c=\frac{m_{Tc}e^{y_c}}{\sqrt{s}},\qquad x_\kappa=\frac{m_{T\kappa}e^{y_\kappa}}{\sqrt{s}}. \tag*{(11)}

The two δ\delta functions in Eq. (6) can be integrated to yield

dσ(AB→cdX)dyc cTdcTdϕc dyκ κTdκTdϕκ=∑abKabxaGa/A(xa)xbGb/B(xb)e−(cT+κT)24σ22π(4πσ2)×dσ(ab→cκ)dt,(12)\frac{d\sigma(AB\to cdX)}{dy_c\,c_Tdc_Td\phi_c\,dy_\kappa\,\kappa_Td\kappa_Td\phi_\kappa} =\sum_{ab}K_{ab}x_aG_{a/A}(x_a)x_bG_{b/B}(x_b)\frac{e^{-\frac{(c_T+\kappa_T)^2}{4\sigma^2}}}{2\pi(4\pi\sigma^2)} \times\frac{d\sigma(ab\to c\kappa)}{dt}, \tag*{(12)}

where

xa=xc+κ2+κT2xκs−b2+bT2xbs=mcTeycs+mκTe−yκs−b2+bT2xbs,x_a=x_c+\frac{\kappa^2+\kappa_T^2}{x_\kappa s}-\frac{b^2+b_T^2}{x_b s} =\frac{m_cT e^{y_c}}{\sqrt{s}}+\frac{m_{\kappa T}e^{-y_\kappa}}{\sqrt{s}}-\frac{b^2+b_T^2}{x_b s},
xb=xκ+c2+cT2xcs−a2+aT2xas=mκTeyκs+mcTe−ycs−a2+aT2xas.(13)x_b=x_\kappa+\frac{c^2+c_T^2}{x_c s}-\frac{a^2+a_T^2}{x_a s} =\frac{m_{\kappa T}e^{y_\kappa}}{\sqrt{s}}+\frac{m_{cT}e^{-y_c}}{\sqrt{s}}-\frac{a^2+a_T^2}{x_a s}. \tag*{(13)}

The above formula renders the cross-section for the production of cc and κ\kappa, when the elementary cross-section dσ(ab→cκ)/dtd\sigma(ab\to c\kappa)/dt is given explicitly in terms of its dependent variables.

The angular correlation d(AB cX)/ddy

To obtain the angular correlations between particles cc and κ\kappa, we transform (yc,yκ,ϕc,ϕκ)(y_c,y_\kappa,\phi_c,\phi_\kappa) to (Y,Δy,Φ,Δϕ)(Y,\Delta y,\Phi,\Delta\phi),

yc=Y+Δy/2,yκ=Y−Δy/2,ϕc=Φ+Δϕ/2,ϕκ=Φ−Δϕ/2.(14)y_c=Y+\Delta y/2,\quad y_\kappa=Y-\Delta y/2, \qquad \phi_c=\Phi+\Delta\phi/2,\quad\phi_\kappa=\Phi-\Delta\phi/2. \tag*{(14)}

We then have

dyc dyκ dϕc dϕκ=dY dΔy dΦ dΔϕ.(15)dy_c\,dy_\kappa\,d\phi_c\,d\phi_\kappa=dY\,d\Delta y\,d\Phi\,d\Delta\phi. \tag*{(15)}

(12) becomes

dσ(AB→cκX)dΔϕ dΔy dY dΦ=∑abKab∫xaGa/A(xa)xbGb/B(xb)×cTdcTκTdκTe−(cT+κT)24σ22π(4πσ2)dσ(ab→cκ)dt.(16)\frac{d\sigma(AB\to c\kappa X)}{d\Delta\phi\,d\Delta y\,dY\,d\Phi} =\sum_{ab}K_{ab}\int x_aG_{a/A}(x_a)x_bG_{b/B}(x_b) \times c_Tdc_T\kappa_Td\kappa_T\frac{e^{-\frac{(c_T+\kappa_T)^2}{4\sigma^2}}}{2\pi(4\pi\sigma^2)} \frac{d\sigma(ab\to c\kappa)}{dt}. \tag*{(16)}

The cross-section should be independent of the average value of ϕc\phi_c and ϕκ\phi_\kappa. The integration over Φ\Phi gives

dσ(AB→cκX)dΔϕ dΔy dY=∑abKab∫cTdcTκTdκT×xaGa/A(xa)xbGb/B(xb)e−(cT+κT)24σ2(4πσ2)dσ(ab→cκ)dt.(17)\frac{d\sigma(AB\to c\kappa X)}{d\Delta\phi\,d\Delta y\,dY} =\sum_{ab}K_{ab}\int c_Tdc_T\kappa_Td\kappa_T \times x_aG_{a/A}(x_a)x_bG_{b/B}(x_b)\frac{e^{-\frac{(c_T+\kappa_T)^2}{4\sigma^2}}}{(4\pi\sigma^2)} \frac{d\sigma(ab\to c\kappa)}{dt}. \tag*{(17)}

We focus on Y=0Y=0 and consider approximate boost invariance at midrapidities. The correlation function Eq. (12) from the process ab→cκab\to c\kappa at midrapidities becomes

dσ(AB→cκX)dΔϕ dΔy∣Y=0=KabC(Δϕ,Δy),(18)\left.\frac{d\sigma(AB\to c\kappa X)}{d\Delta\phi\,d\Delta y}\right|_{Y=0}=K_{ab}C(\Delta\phi,\Delta y), \tag*{(18)}

where

C(Δϕ,Δy)=∫0∞cTdcT∫0∞κTdκTxaGa/A(xa)xbGb/B(xb)×1(4πσ2)exp⁡{−cT2+2cTκTcos⁡Δϕ+κT24σ2}dσ(ab→cκ)dt.(19)C(\Delta\phi,\Delta y)=\int_0^\infty c_Tdc_T\int_0^\infty\kappa_Td\kappa_Tx_aG_{a/A}(x_a)x_bG_{b/B}(x_b) \times\frac{1}{(4\pi\sigma^2)}\exp\left\{-\frac{c_T^2+2c_T\kappa_T\cos\Delta\phi+\kappa_T^2}{4\sigma^2}\right\}\frac{d\sigma(ab\to c\kappa)}{dt}. \tag*{(19)}

The basic cross-section dσ/dtd\sigma/dt can be written in terms of relativistic invariant quantities (s′,t′,u′)(s',t',u') measured in the intrinsic system of a+b→c+κa+b\to c+\kappa. We have Eq. (10.169) in Ref. [53] for gg→gggg\to gg,

dσ(gg→gg)dt=9παs28(s′4+t′4+u′4)(s′2+t′2+u′2)s′4t′2u′2.(20)\frac{d\sigma(gg\to gg)}{dt}=\frac{9\pi\alpha_s^2}{8}\frac{(s'^4+t'^4+u'^4)(s'^2+t'^2+u'^2)}{s'^4t'^2u'^2}. \tag*{(20)}

From Eq. (2.12) of Ref. [54] for gg→qqˉgg\to q\bar q, and Eq. (A1) of Ref. [55], with quark mass mc=mκ=mm_c=m_\kappa=m, we have for gg→qqˉgg\to q\bar q,

dσdt(gg→qqˉ)=παs216s′2[12s′2(m2−t′)(m2−u′)+83(m2−t′)(m2−u′)−2m2(m2+t′)(m2−t′)2+83(m2−t′)(m2−u′)−2m2(m2+u′)(m2−u′)2−2m2(s′−4m2)3(m2−t′)(m2−u′)−6((m2−t′)(m2−u′)+m2(u′−t′))s′(m2−t′)−6((m2−t′)(m2−u′)+m2(t′−u′))s′(m2−u′)].(21)\frac{d\sigma}{dt}(gg\to q\bar q)=\frac{\pi\alpha_s^2}{16s'^2}\left[\frac{12}{s'^2}(m^2-t')(m^2-u')+\frac{8}{3}\frac{(m^2-t')(m^2-u')-2m^2(m^2+t')}{(m^2-t')^2} +\frac{8}{3}\frac{(m^2-t')(m^2-u')-2m^2(m^2+u')}{(m^2-u')^2} -\frac{2m^2(s'-4m^2)}{3(m^2-t')(m^2-u')} -\frac{6\left((m^2-t')(m^2-u')+m^2(u'-t')\right)}{s'(m^2-t')} -\frac{6\left((m^2-t')(m^2-u')+m^2(t'-u')\right)}{s'(m^2-u')}\right]. \tag*{(21)}

From Eq. (2.7) of Ref. [54], we have for qqˉ→ccˉq\bar q\to c\bar c

dσ(qqˉ→ccˉ)dt=4παs29s′2(m2−t′)2+(m2−u′)2+2m2s′s′2.(22)\frac{d\sigma(q\bar q\to c\bar c)}{dt}=\frac{4\pi\alpha_s^2}{9s'^2}\frac{(m^2-t')^2+(m^2-u')^2+2m^2s'}{s'^2}. \tag*{(22)}

The structure function can be expressed in the form xaGa/A(xa)∝(1−xa)gax_aG_{a/A}(x_a)\propto(1-x_a)^{g_a}, for which the two-particle angular correlation function becomes approximately

dσ(AB→cκX)dΔϕ dΔy∣Y∼0∼A[1−4mcTscosh⁡Δy+O((mcTs)2)]ga×C(Δϕ).(23)\left.\frac{d\sigma(AB\to c\kappa X)}{d\Delta\phi\,d\Delta y}\right|_{Y\sim0} \sim A\left[1-\frac{4m_{cT}}{\sqrt{s}}\cosh\Delta y+O\left(\left(\frac{m_{cT}}{\sqrt{s}}\right)^2\right)\right]^{g_a}\times C(\Delta\phi). \tag*{(23)}

This indicates that as mcT/sm_{cT}/\sqrt{s} is much less than unity, the correlation function is a weak function of Δy\Delta y, and

C(Δϕ)=∫0∞cTdcT∫0∞κTdκT×1(4πσ2)exp⁡{−cT2+2cTκTcos⁡Δϕ+κT24σ2}dσ(ab→cκ)dt(24)C(\Delta\phi)=\int_0^\infty c_Tdc_T\int_0^\infty\kappa_Td\kappa_T \times\frac{1}{(4\pi\sigma^2)}\exp\left\{-\frac{c_T^2+2c_T\kappa_T\cos\Delta\phi+\kappa_T^2}{4\sigma^2}\right\}\frac{d\sigma(ab\to c\kappa)}{dt} \tag*{(24)}

From Eqs. (19) and (24), we expect that the contribution to the correlated production of particles cc and κ\kappa is the greatest when Δϕ=±π\Delta\phi= \pm\pi, for which particles cc and κ\kappa are produced positioned back-to-back. There is no completely rigorous model to go from pQCD for produced particles in the low-pTp_T region because the mechanism of the non-perturbative hadronization is not fully known. One relies on semi-empirical fragmentation functions to describe the fragmentation of the partons to hadrons cc and κ\kappa in the hard-scattering model.

To grasp the gross feature of the correlation function, we use the simple parton-hadron duality and approximate dσ(ab→cκ)/dtd\sigma(ab\to c\kappa)/dt to be of the form as in gg→gggg\to gg collisions in Eq. (20). We have

s′∼4mcT2cosh⁡2Y,s' \sim4m_{cT}^{2}\cosh^{2}Y,
t′−mc2=−2cosh⁡Ye−YmcT2=(1+e−2Y)mcT2,t' - m_{c}^{2}=-2\cosh Y e^{-Y}m_{cT}^{2}=(1+e^{-2Y})m_{cT}^{2},
u′−mc2=−2cosh⁡YeYmcT2=−(1+e2Y)mcT2.(25)u' - m_{c}^{2}=-2\cosh Y e^{Y}m_{cT}^{2}=-(1+e^{2Y})m_{cT}^{2}. \tag*{(25)}

For gg→gggg\to gg with small mcTm_{cT}, we have dσ(gg→gg)/dt∼1/mcT4d\sigma(gg\to gg)/dt\sim1/m_{cT}^{4}, and

dσ(ab→cκ)dt∼1t′2∼A′[1+(mc2+cT2)/MT02]n/2.(26)\frac{d\sigma(ab\to c\kappa)}{dt}\sim\frac{1}{t'^{2}}\sim\frac{A'}{[1+(m_{c}^{2}+c_{T}^{2})/M_{T0}^{2}]^{n/2}}. \tag*{(26)}

where n∼4n\sim4 in pQCD and MT0M_{T0} is the regularizing mass on the order of 1 GeV/c21\ \mathrm{GeV}/c^{2} empirically in the hadron-parton duality approximation [14]. The numerical integration over cTc_T and κT\kappa_T in Eq. (24) gives the correlation function C(Δϕ)C(\Delta\phi) as shown in Figure 1. The correlation function for mparton=0.14 GeV/c2m_{\mathrm{parton}}=0.14\ \mathrm{GeV}/c^{2} has maxima at Δϕ≈±π\Delta\phi\approx\pm\pi and a minimum at Δϕ=0\Delta\phi=0 because of the elementary back-to-back production process of a+b→c+κa+b\to c+\kappa. It is relatively flat in Δy\Delta y because mcT/s≪1m_{cT}/\sqrt{s}\ll1 in Eq. (16) for high-energy collisions. This gives the feature of the correlation function in the form of a ridge structure on the away side at Δϕ≈±π\Delta\phi\approx\pm\pi, with a back-to-back azimuthal correlation in ϕ\phi, as expected from momentum conservation.

Image not extracted

Figure 1. The correlation function C(Δϕ)C(\Delta\phi).

pT-seeded Clustering Algorithm

Because the mini-jet manifests as a concentrated release of energy with a high-pTp_T value during its parton-cascading evolution, some products of its on-shell hadronized particles are likely to retain high pTp_T at the endpoints of the evolution. For this reason, we assume that every mini-jet has a high-pTp_T remnant. The magnitude of the pTp_T value, which allows such a characterization, will need to be determined semi-empirically. The mini-jet is also expected to evolve in a certain direction, with the cluster of its evolution products likely within a bundle in the pseudorapidity angles and azimuthal angles. Therefore, we develop the pTp_T-seeded clustering algorithm to idealize mini-jet as a cluster of hadrons in the (η,ϕ)(\eta,\phi) space with a radius previously estimated to be on the order of R=0.5−0.6R=0.5-0.6 [56]. For a given set of MM particles in an event with specified positions xi=(ηi,ϕi)x_i=(\eta_i,\phi_i) and transverse momenta pTip_{Ti}, where i=1,2,…Mi=1,2,\ldots M is the particle index, particles above a threshold pT0p_{T0} are selected as the seed particles. These seed particles serve as indicators of mini-jets and determine the initial number and locations of the clusters, while the low-pTp_T particles are used for the cluster location refinement in later steps. Particles outside the mini-jet cones are regarded as products of non-jet processes.

Each seed particle is initially treated as the center of a seeded cluster. To avoid overlapping, we merge clusters that are sufficiently close to each other to form a single cluster. The merging process calculates the distance between each pair of centers, ℓ=[(Δη)2+(Δϕ)2]1/2\ell=[(\Delta\eta)^2+(\Delta\phi)^2]^{1/2}, and finds the two closest centers. If this shortest distance is less than a certain threshold, 0.8160.816 in this algorithm, we merge the two clusters. With a minijet modeled as a cone with a given radius R=0.6R=0.6, we envisage that two seed candidates will have an 83.4%83.4\% likelihood of belonging to the same minijet if their ℓ\ell is less than 0.8160.816. The assigned value of merging threshold 0.8160.816 is derived from the hypothetical scenario wherein two points are randomly sampled from inside a circle with a radius of R=0.6R=0.6, and the distance between them adheres to ρ(ℓ)\rho(\ell), the distance distribution in a uniform disk [57]:

ρ(ℓ)=4ℓπR2arccos⁡(ℓ2R)−2ℓ2πR4R2−ℓ24.(27)\rho(\ell)=\frac{4\ell}{\pi R^{2}}\arccos\left(\frac{\ell}{2R}\right)-\frac{2\ell^{2}}{\pi R^{4}}\sqrt{R^{2}-\frac{\ell^{2}}{4}}. \tag*{(27)}

After calculating the mean μ\mu and the width σ\sigma of the distance distribution, we obtain μ+σ≈0.816\mu+\sigma\approx0.816. The range [0,0.816][0,0.816] corresponds to approximately 83.4%83.4\% probability of the distance distribution. The cumulant probability can be increased to 90%(95%)90\%(95\%) with the upper bound of 0.894(0.974). Various merging thresholds can be explored by examining the physics of the jet cascade, for example, in PYTHIA. The new center of the merged cluster is calculated as the weighted midpoint of the previous two seed particles, with the weight assigned based on their pTp_T. This merging process continues until the distance between any two clusters is greater than 0.816. At this point, the cluster positions are roughly determined, based on the seed particles.

To enhance the precision and reliability of the cluster positions, all particles are used to recalculate the cluster centers as a refinement. The final cluster centers are calculated using a formula:

Ck=∑xik∈Sk(pTi⟨pT⟩)γxik∑xik∈Sk(pTi⟨pT⟩)γ,(28)C_k = \frac{\sum_{x_i^k\in S_k}\left(\frac{p_{T_i}}{\langle p_T\rangle}\right)^\gamma x_i^k}{\sum_{x_i^k\in S_k}\left(\frac{p_{T_i}}{\langle p_T\rangle}\right)^\gamma}, \tag*{(28)}

where the subscript kk represents the cluster index, and SkS_k denotes the subset of particles located within a radius of 0.6 around the center of the corresponding seeded cluster. A power index of γ=2\gamma=2 is used in the algorithm. The resultant clusters are finally settled after the refinement of cluster positions.

Figure 2 illustrates examples of the pTp_T-seeded clustering algorithm using a threshold of pT0=0.5p_{T0}=0.5 GeV/cc. The high-pTp_T particles, denoted by red points, are initially assigned as independent cluster centers. Seeds that are close to each other are merged, as shown in the upper right corner of the third panel. The center positions of clusters are then refined by the surrounding low-pTp_T particles. In cases where no other particles are present around a seed particle, a cluster is still formed with the seed particle as the center, as demonstrated in the second and fourth panels.

Image not extracted

Figure 2. Examples with pT0=0.5p_{T0}=0.5 GeV/cc. The seed particles are shown as red points.

Correlations Between Particles and Between Clusters

Two-Particles Correlations

To assess the viability and performance of our pTp_T-seeded clustering algorithm, we conducted evaluations using PYTHIA8.1 simulations for particles produced in p+pp+p collisions at s=200\sqrt{s}=200 GeV, within the range of −1≤η≤1-1\leq\eta\leq1 and −π≤ϕ≤π-\pi\leq\phi\leq\pi. Figure 3 depicts the η\eta, ϕ\phi, and pTp_T distributions of the charged particles in the simulation. We have explored three different values of the threshold pT0p_{T0}: 0.5, 1, and 1.5 GeV/cc, and retained events with at least two particles having pT≥pT0p_T \geq p_{T0} for correlation analyses. The pTp_T distributions of all particles in events that passed the pT0p_{T0} cut of 0.5 and 1.5 GeV/cc are presented in Figure 4. The discontinuities observed in the pTp_T distributions at pT0p_{T0} are due to the threshold effect, as events comprising solely low-pTp_T particles are removed. Despite this, low-pTp_T particles from the surviving events still contribute to the distribution below pT0p_{T0}. Therefore, the distribution above the pT0p_{T0} thresholds maintains its similarity to the original uncut distribution, while the distribution below these thresholds experiences a reduction due to the removal of low-pTp_T events, thereby causing the observed discontinuity.

Image not extracted

Figure 3. Particle distributions in (a) η\eta, (b) ϕ\phi, and (c) pTp_T for unbiased PYTHIA events. The upper narrow panels of the η\eta and ϕ\phi distributions show an amplified view of the distributions.

Image not extracted

Figure 4. Particle pTp_T distributions for PYTHIA events selected with pT0=0.5p_{T0}=0.5 (a) and 1.51.5 GeV/cc (b).

For comparison with the PYTHIA data (same events), we construct the mixed events by reassigning ϕ\phi and η\eta values to particles in each PYTHIA event, while preserving the event’s particle multiplicity and each particle’s pTp_T value. The ϕ\phi and η\eta values are randomly sampled from the original particle’s ϕ\phi and η\eta distributions, shown in Figure 3. This procedure ensures that the resulting particles naturally have the same pTp_T values as the original particles, the mixed events have the same multiplicities as the original events, and the mixed data set has the

same ϕ\phi and η\eta distribution as the PYTHIA data set. In the following, we study the difference between the same events and the mixed events in terms of two-particle and two-cluster correlations.

Figure 5 shows the ratio of the two-particle correlation (Δϕ−Δη)(\Delta\phi-\Delta\eta) function for the same events to that for the mixed events with pT0=0.5p_{T0}=0.5 (upper panel) and 1.5 GeV/cc (lower panel). Before taking the ratio, the correlations for both the same events and the mixed events have been normalized with the total number of particle pairs. We restrict the analysis to the Δη\Delta\eta range of [0, 1.4] due to the limited number of events that pass the pT0p_{T0} cuts. Although the same events and the mixed events have very similar ϕ\phi and η\eta distributions, notable differences can be observed between the two data sets in the two-particle correlation. In each panel, the correlation ratio exhibits a near-side peak at Δη≈0\Delta\eta\approx0 and Δϕ≈0\Delta\phi\approx0, as well as an away-side ridge at Δϕ≈π\Delta\phi\approx\pi, which is consistent with momentum conservation and the analysis of Section III. The away-side ridge becomes more prominent for higher pT0p_{T0}, with a magnitude of approximately 1.04 (1.15) for pT0=p_{T0} = 0.5 (1.5) GeV/cc.

Image not extracted

Figure 5. Ratio of the two-particle correlation (Δϕ−Δη)(\Delta\phi-\Delta\eta) function for the same events to that for the mixed events with pT0=0.5p_{T0}=0.5 (upper) and 1.5 GeV/cc (lower).

Image not extracted

Figure 6. Number of clusters, KK, as a function of multiplicity, MM, for the same events (red circle) and the mixed events (blue triangle), and the ratio (black circle) of the same events to the mixed events for pT0=p_{T0} = (a) 0.5 GeV/cc and (b) 1.5 GeV/cc.

Two-Cluster Correlations

While the two-particle correlation can reveal some interesting properties, the two-cluster correlation is expected to amplify the correlation features related to mini-jets due to the use of the pTp_T-seeded clustering algorithm. Figure 6 shows the number of reconstructed clusters, KK, as a function of multiplicity, MM, for pT0=p_{T0} = (a) 0.5 and (b) 1.5 GeV/cc. We observe significant differences between the same events and the mixed events, as well as between the cases with different pT0p_{T0} values. When pT0=p_{T0} = 0.5 GeV/cc, the difference between the same events and the mixed events is not prominent at low MM, and mixed events show a slightly larger KK at higher MM. The ratio, Ksame/KmixedK_{\mathrm{same}}/K_{\mathrm{mixed}}, exhibits a slowly decreasing trend from 0.98 to 0.94. On the other hand, when pT0=p_{T0} = 1.5 GeV/cc, the mixed events yield significantly more clusters than the same events, with Ksame/KmixedK_{\mathrm{same}}/K_{\mathrm{mixed}} being roughly constant around 0.86. The difference between the same events and the mixed events can be attributed to the clustering property of PYTHIA particles, among which particles belonging to the same mini-jet are likely to be captured by our algorithm to form a cluster. Conversely, particles in the mixed events are uniformly distributed, and clusters are randomly formed. Thus, the algorithm is more effective in clustering mini-jets when using higher pT0p_{T0} values, resulting in a more prominent difference between the same events and the mixed events. pT0=p_{T0} = 0.5 GeV/cc may be insufficient to clearly distinguish the same events and the mixed events, especially at low MM.

To investigate the correlation between two clusters, we classify events based on the value of KK obtained from the algorithm. Specifically, we analyze the cases of K=2K = 2, 3, and 4 to demonstrate the dependence on KK and to provide adequate statistical data for further analysis.

Figures 7 and 8 display the two-cluster correlations (Δϕ\Delta\phi–Δη\Delta\eta) for the same events (upper) and the mixed events (lower) with K=2K = 2 for pT0=0.5p_{\mathrm{T0}} = 0.5 and 1.51.5 GeV/cc, respectively. Note that Δϕ\Delta\phi and Δη\Delta\eta now represent the distances between two cluster centers instead of two particles. The 2D distribution primarily follows the Δη\Delta\eta distribution that approximately takes the shape of a triangle, holds a ridge around Δη≈0\Delta\eta\approx0, and gradually decreases as Δη\Delta\eta increases. The holes around Δϕ≈0\Delta\phi\approx0 and Δη≈0\Delta\eta\approx0 in these distributions result from the exclusivity of the clusters since two clusters tend to merge in our clustering algorithm if their distance is small. The same-event data exhibit an away-side ridge at Δϕ≈π\Delta\phi\approx\pi, which is absent in the mixed-event data. Specifically, at a fixed value of Δη\Delta\eta, the number of entries around Δϕ≈π\Delta\phi\approx\pi is significantly greater than that around Δϕ≈π/2\Delta\phi\approx\pi/2 in the same events, whereas it remains nearly uniform along Δϕ\Delta\phi in the mixed events. The threshold pT0p_{\mathrm{T0}} also impacts these distributions. A significantly higher away-side peak appears in the same events for pT0=1.5p_{\mathrm{T0}} = 1.5 GeV/cc (in Figure 8)

Image not extracted

Figure 8. Two-cluster correlation (Δϕ\Delta\phi–Δη\Delta\eta) for the same events (upper) and the mixed events (lower) with pT0=1.5p_{\mathrm{T0}} = 1.5 GeV/cc and K=2K = 2.

than for pT0=0.5p_{\mathrm{T0}} = 0.5 GeV/cc (in Figure 7).

Image not extracted

Figure 7. Two-cluster correlation (Δϕ\Delta\phi-Δη\Delta\eta) for the same events (upper) and the mixed events (lower) with pT0=p_{T0} = 0.5 GeV/cc and K=2K = 2.

Figure 9 shows the ratio of the two-cluster correlation for the same events to that for the mixed events with K=2K = 2 for pT0=0.5p_{\mathrm{T0}} = 0.5 (upper) and 1.51.5 GeV/cc (lower). Before taking the ratio, we have normalized the 2D correlations for the same events and the mixed events with their corresponding total numbers of cluster pairs, respectively, similar to the approach used for the ratio in the two-particle correlations in Figure 5. Except for the presence of holes at Δϕ≈0\Delta\phi\approx0 and Δη≈0\Delta\eta\approx0, the two-cluster correlations qualitatively resemble the two-particle correlations in Figure 5 and reveal more prominent away-side ridges around Δϕ≈π\Delta\phi\approx\pi. Remarkably, the away-side ridge appears consistently and uniformly for the two-cluster correlations, even for pT0=0.5p_{\mathrm{T0}} = 0.5 GeV/cc, and is enhanced for the higher pT0=1.5p_{\mathrm{T0}} = 1.5 GeV/cc. The magnitude of the ratio within the away-side range is about 1.25 (1.4) for pT0=0.5p_{\mathrm{T0}} = 0.5 (1.5) GeV/cc.

Figure 10 illustrates the KK dependence of the ratio of the two-cluster correlation (Δϕ−Δη\Delta\phi- \Delta\eta) function for the same events to that for the mixed events. with pT0=1p_{\mathrm{T0}} = 1 GeV/cc. The upper, middle, and lower panels correspond to KK values of 2, 3, and 4, respectively. As KK increases, the magnitude of the away-side ridge in the ratio decreases. The away-side ridge is likely due to momentum conservation, which is diluted by increasing KK.

Image not extracted

Figure 10. Ratio of the two-cluster correlation (Δϕ\Delta\phi-Δη\Delta\eta) function for the same events to that for the mixed events with pT0=1p_{T0}=1 GeV/c, and K=2K=2 (upper), 3 (middle), and 4 (lower).

The effectiveness of the pTp_T-seeded clustering algorithm can be evaluated by comparing the two-particle and two-cluster correlations, particularly by examining the away-side ridge, and reducing the correlation function to 1D, which simplifies the assessment of the correlation strength. To eliminate bias, both types of correlations are calculated using events with similar multiplicities. Figures 11 and 12 present the ratio of the Δϕ\Delta\phi distribution

for the same events to that for the mixed events with pT0=0.5p_{T0}=0.5 and 1.5 GeV/c, respectively, for the multiplicity ranges of [5, 8], [9, 12] and [13, 16]. The upper panels display the two-particle correlations, and the lower ones show the two-cluster correlations. The two-particle correlation shows a near-side peak, which is not present in the two-cluster correlation due to the exclusivity of clusters, as shown by the holes in the 2D correlations in Figs. 7–10. For pT0=0.5p_{T0}=0.5 GeV/c, the away-side peak decreases with increased multiplicity, and the peaks in the two-cluster correlations are lower than those in the corresponding two-particle correlations. However, for pT0=1.5p_{T0}=1.5 GeV/c, the two-cluster correlations consistently show higher away-side peaks than the two-particle correlations. Hence, by selecting an appropriate pT0p_{T0}, the clustering algorithm is more effective in revealing the mini-jet properties compared with the two-particle correlations.

Difference with the anti-kTk_T algorithm

The anti-kTk_T algorithm [58] is a sequential recombination approach, based on the distance measure,

dij=min⁡(pTi−2,pTj−2)×Rij2R2,(29)d_{ij}=\min\left(p_{Ti}^{-2},p_{Tj}^{-2}\right)\times\frac{R_{ij}^{2}}{R^{2}}, \tag*{(29)}

in combination with a boundary condition:

diB=pTi−2,(30)d_{iB}=p_{Ti}^{-2}, \tag*{(30)}

where Rij2=(ηi−ηj)2+(ϕi−ϕj)2R_{ij}^{2}=(\eta_i-\eta_j)^{2}+(\phi_i-\phi_j)^{2}. The first component dijd_{ij} of Eq. (29) serves as a filter that selects particles suitable for inclusion in a cluster. The use of a power of −2-2 for pTp_T prioritizes the clustering of high-pTp_T particles. The second component of Eq. (30) sets the criterion for either initiating or halting the recombination process. To satisfy the recombination condition dij<diBd_{ij}<d_{iB}, the value of RijR_{ij} must be smaller than the cluster radius RR. If this condition is not satisfied, the recombination process is terminated. The algorithm iterates through the recombination process until only a few objects remain, and these remaining objects are then referred to as jets. While not strictly a cone algorithm, the anti-kTk_T algorithm exploits the existence of high-pTp_T particles to stabilize the shapes of the resulting clusters, producing outcomes that resemble a cone-shaped pattern. This is especially evident under typical circumstances where the high-pTp_T particles exhibit substantial differences from soft particles. Due to the considerable contrast in the pTp_T values between hard and soft particles, the clustering process is relatively insensitive to the inclusion of soft particles in the jet. However, the anti-kTk_T algorithm may not be ideally suited for the mini-jet condition, which features considerable presence of soft particles and the indistinct demarcation between high- and low-pTp_T particles. The presence of non-jet soft particles makes it hard to clearly define mini-jets when we use the anti-kTk_T algorithm. In contrast, the pTp_T-seeded clustering method is a cone algorithm that allows the distinct separation of the mini-jet with other soft particles using high-pTp_T seeds, providing advantages in the low pTp_T regime. The pTp_T-seeded clustering method is designed to exclude the influence of soft particles in the mini-jet environment. The effect of this method can be observed in cases with pT0p_{T0} values of 0.5 GeV/cc and 1.5 GeV/cc. In particular, clusters with pT0=0.5p_{T0} = 0.5 GeV/cc display minimal discrepancies in the two-cluster correlation between the same events and the mixed events (in Figure 9). However, increasing pT0p_{T0} to 1.5 GeV/cc reveals noticeable differences in the correlation functions, empha-

Image not extracted

Figure 9. Ratio of the two-cluster correlation (Δϕ\Delta\phi–Δη\Delta\eta) function for the same events to that for the mixed events for events selected with K=2K = 2 for pT0=0.5p_{\mathrm{T0}} = 0.5 (upper) and 1.51.5 GeV/cc (lower).

Image not extracted

Figure 11. Ratio of the (a) two-particle or (b) two-cluster correlation (Δϕ\Delta\phi) function for the same events to that for the mixed events with pT0=0.5p_{T0} = 0.5 GeV/cc for the multiplicity ranges of [5, 8] (left), [9, 12] (middle) and [13, 16] (right).

Image not extracted

Figure 12. Same as Fig. 11, except for pT0=1.5p_{T0} = 1.5 GeV/cc.

sizing the influence of soft particles in the mini-jet environment and demonstrating the enhancement provided by the pTp_T-seeded clustering algorithm. The pTp_T-seeded clustering method includes all particles in the vicinity of the seeded cluster center, effectuating a static clustering approach that accurately captures the original particle configuration. Unlike the dynamic, continual particle-merging process characteristic of the anti-kTk_T algorithm, the pTp_T-seeded method clusters particles based on their actual spatial distribution, and consequently, it generates intuitive results. Our method allows for an easier prediction of the jet position, facilitating straightforward visual identification of the jet’s location. By incorporating the pTp_T seed, the numerous low-pTp_T particles can be disregarded in the initial round, leading to a substantial enhancement in the compilation speed. Eventually, the pTp_T-seed method can complement the anti-kTk_T algorithm in the clustering of mini-jets, particularly in the low-pTp_T region where the application of the anti-kTk_T algorithm might not be suitable.

Conclusions

We have developed a pTp_T-seeded clustering algorithm, which utilizes a threshold pT0p_{T0} constraint on the seed to exclude low-pTp_T particles during the initial clustering steps so that jets that are solely composed of low-pTp_T particles are eliminated from the output. The high-pTp_T seed particles are merged based on their η\eta-ϕ\phi distance, and subsequently, the clustering process is refined by incorporating low-pTp_T particles. The design is aimed at minimizing the effects of random cluster noise in the low-pTp_T range. We have evaluated the performance of our algorithm using a PYTHIA data set. The clustering results indicate the expected hard-scattering-type correlations. Notably, the performance characteristics vary with pT0p_{T0}, with a pT0p_{T0} range of 1 to 1.5 GeV/c yielding a clear distinction between the same events and the mixed events.

We are excited about the potential to apply our seeded clustering algorithm to real data to investigate novel phenomena, such as collective flow or other mechanisms beyond hard scattering. In actual hadron-hadron collisions, the number of clusters may be significant, as demonstrated in Figure 13. This presents an exciting opportunity to explore exotic reaction mechanisms.

Image not extracted

Figure 13. An example PYTHIA event with M=17M = 17, K=6K = 6, and pT0=1p_{T0} = 1 GeV/c.

Acknowledgments

G. W. and H. Z. H. are supported by the U.S. Department of Energy under Grant No. DE-FG02-88ER40424 and by the National Natural Science Foundation of China under Contract No.1835002. C. Y. W. is supported in part by UT-Battelle, LLC, under Contract No. DE-AC05-00OR22725 with the US Department of Energy (DOE). H. J. is supported by the U.S. Department of Energy under Grant No. DE-FG02-86ER40681. N. Y. is supported by the U.S. Department of Energy under Grant No. DE-SC0023861.

References

  1. [1]R. Blankenbecler and S. J. Brodsky, Phys. Rev. D 10, 2973 (1974); R. Blankenbecler, S. J. Brodsky and J. Gunion, ibid. 12, 3469 (1975); I. A. Schmidt and R. Blankenbecler, ibid. 15, 3321 (1977), R. Blankenbecler, A. Capella, J. Tran, Thanh Van, C. Pajares, and A.V. Ramallo, Phys. Lett. B 107, 106 (1981).
  2. [2]A.L.S. Angelis et al. (CCOR Collaboration), Phys. Lett. B 79, 505 (1978).
  3. [3]R. P. Feynman, R. D. Field and G. C. Fox, Phys. Rev. D: Part. Fields 18, 3320 (1978).
  4. [4]J. F. Owens, E. Reya, and M. Glück, Phys. Rev. D 18, 1501 (1978); D. W. Duke, J. F. Owens, ibid. 30, 49 (1984).
  5. [5]J. Rak and M. J. Tannenbaum, High-pT Physics in the Heavy Ion Era (Cambridge University Press, Cambridge, 2013).DOI
  6. [6]T. Sjöstrand, Comput. Phys. Commun. 39, 347 (1986); T. Sjöstrand and M. Bengtsson, ibid. 43, 367 (1987).
  7. [7]T. Sjöstrand and M. van Zijl, Phys. Rev. D: Part. Fields 36, 2019 (1987).
  8. [8]C. Albajar et al. (UA1 Collaboration), Nucl. Phys. B 309, 405 (1988).
  9. [9]K. S. Eskola, K. Kajantie, J. Lindfors, Nucl. Phys. B 323, 37 (1989).
  10. [10]T. Sjöstrand, S. Mrenna, P. Skands, J. High Energy Phys. 05 (2006) 026.
  11. [11]X. N. Wang and M. Gyulassy, Phys. Rev. D: Part. Fields 44, 3501 (1991).
  12. [12]J. Adams et al. (STAR Collaboration), Phys. Rev. C 73, 064907 (2006).
  13. [13]C.Y. Wong, G. Wilk, Phys. Rev. D 87, 114007 (2013).
  14. [14]C.Y. Wong, G. Wilk, L. J. L. Cirto and C. Tsallis, Phys. Rev. D 91, 114027 (2015).
  15. [15]P. Steinberg, Proc. Sci. CPOD2006, 036 (2006).
  16. [16]J. D. Bjorken, Phys. Rev. D: Part. Fields 27, 140 (1983).
  17. [17]R. C. Hwa, Phys. Rev. D: Part. Fields 10, 2260 (1974).
  18. [18]L. D. Landau, Izv. Akad. Nauk Ser. Fiz. 17, 51 (1953).
  19. [19]S. Z. Belenkij and L. D. Landau, Usp. Fiz. Nauk 56, 309 (1955); Nuovo Cimento, 3, 15 (1956).
  20. [20]S. Amai, H. Fukuda, C. Iso, and M. Sato, Prog. Theor. Phys. 17, 241 (1957).
  21. [21]P. Carruthers and M. Doung-van, Phys. Rev. D: Part. Fields 8, 859 (1973).
  22. [22]F. Cooper and G. Frye, Phys. Rev. D: Part. Fields 10, 186 (1974).
  23. [23]S. Chadha, C. S. Lam, and Y. C. Leung, Phys. Rev. D: Part. Fields 10, 2817 (1974).
  24. [24]D. K. Srivastava, J. Alam, B. Sinha, Phys. Lett. B 296, 11 (1992); D. K. Srivastava, Jan-e Alam, S. Chakrabarty, B. Sinha, S. Raha, Ann. Phys. 228, 104 (1993); D. K. Srivastava, J. Alam, S. Chakrabarty, S. Raha, B. Sinha, Phys. Lett. B 278, 225 (1992).
  25. [25]B. Mohanty, and J. Alam, Phys. Rev. C 68, 064903 (2003).
  26. [26]C. E. Aguiar, T. Kodama, T. Osada, Y. Hama, J. Phys. G 27, 75 (2001); Y. Hama, T. Kodama, O. Socolowski Jr., Braz. J. Phys. 35, 24 (2005).
  27. [27]L. D. McLerran and R. Venugopalan, Phys. Rev. D: Part. Fields 49, 2233 (1994).
  28. [28]L. D. McLerran and R. Venugopalan, Phys. Rev. D: Part. Fields 49, 3352 (1994).
  29. [29]E. Iancu, A. Leonidov, and L. McLerran, arXiv:hep-ph/0202270.
  30. [30]E. Iancu and R. Venugopalan, Quark-Gluon Plasma 3, 249 (2004).
  31. [31]F. Gelis, E. Iancu, J. Jalilian-Marian and R. Venugopalan, Annu. Rev. Nucl. Part. Sci. 60, 463 (2010).
  32. [32]S. Pratt, Phys. Rev. C 75, 024907 (2007).
  33. [33]A. Bialas, R. A. Janik, and R. Peschanski, Phys. Rev. C
  34. [34]T. Csörgő, M. I. Nagy, and M. Csanád, Phys. Lett. B 663, 306 (2008).
  35. [35]G. Beuf, R. Peschanski, and E. N. Saridakis, Phys. Rev. C 78, 064909 (2008).
  36. [36]T. Osada and G. Wilk, Centr. Eur. J. Phys. 7, 432 (2009).
  37. [37]E. K. G. Sarkisyan and A.S. Sakharov, arXiv:hep-ph/0410324; E. K. G. Sarkisyan, A.S. Sakharov, AIP Conf. Proc. 828, 35 (2006).arxiv.org/abs/hep-ph/0410324
  38. [38]C. Y. Wong and T. A. Welton, Phys. Lett. B 49, 243 (1974).
  39. [39]M. Basile et al., Nuovo Cimento A 65, 400 (1981); M. Basile, ibid. 67, 244 (1982).
  40. [40]L. P. Csernai, Sov. Phys. JETP 65, 216 (1987).
  41. [41]V. K. Magas, L. P. Csernai and D. D. Strottman, Phys. Rev. C 64, 014901 (2001); Nucl. Phys. A 712, 167 (2002).
  42. [42]L.P. Csernai, J.I. Kapusta, Phys. Rev. D: Part. Fields 46, 1379 (1992).
  43. [43]W.N. Zhang, S.X. Li, C.Y. Wong, M.J. Efaaf, Phys. Rev. C 71, 064908 (2005).
  44. [44]C.Y. Wong, and W.N. Zhang, Phys. Rev. C 70, 064904 (2004).
  45. [45]T. A. Trainor, Phys. Rev. C 90, 024909 (2014).
  46. [46]C. Y. Wong, H. Jiang, N. Yao, L. Wen, G. Wang, H. Z. Huang, Phys. Rev. D 102, 054007 (2020).
  47. [47]T. Sjöstrand, S. Mrenna, P. Skands, Comput. Phys. Commun. 178, 852 (2008).
  48. [48]J. F. Owens, Rev. Mod. Phys. 59,465 (1987).
  49. [49]C.Y. Wong, J. Phys.: Conf. Ser. 668, 012097 (2016).
  50. [50]T. Sjöstrand et al., Comput. Phys. Commun. 191, 159 (2015).
  51. [51]L. Chatterjee and C.Y. Wong, Phys. Rev. C 51, 2125 (1995); C.Y. Wong and L. Chatterjee, Heavy Ion Phys. 4, 201 (1996).
  52. [52]C.Y. Wong and H. Wang, Phys. Rev. C 58, 376 (1998).
  53. [53]R. Gastmans and T.-T. Wu, The Ubiquitous Photon: Helicity Methods for QED and QCD (Oxford University Press, United Kingdom, 1990).
  54. [54]B. L. Combridge, Nucl. Phys. B 151, 429 (1979).
  55. [55]M. Glück, J. F. Owens, and E. Reya, Phys. Rev. D: Part. Fields 17, 2324 (1978).
  56. [56]C.Y. Wong, Phys. Rev. C 80, 034908 (2009).
  57. [57]Ricardo García-Pelayo, J. Phys. A: Math. Gen. 38, 3475 (2005).
  58. [58]M. Cacciari, G. Salam, and G. Soyez, J. High Energy Phys. 04 (2008) 063.

Paper details

vlnc:W_aabfhrhv
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents