A systematic analysis of transverse momentum spectra of $J/\psi$ mesons
in high energy collisions

Quantum chromodynamics (QCD) is the standard dynamics theory and an important part of the standard model, which is applicable in the study of heavy quark pair production and correlation 1; 2; 3; 4; 5; 6 such as the transverse momentum spectra, nuclear modification factor, azimuthal correlation, anisotropic flow, and so on. QCD is a kind of non-Abelian gauge field theory 6a, which implies that the strong interactions between quarks have three basic characteristics. Firstly, it can explain the asymptotic freedom characteristics proposed in the inelastic electron-proton and electron-deuteron scattering 6b. Secondly, it can explain the color confinement which shows quarks and anti-quarks cannot be separated due to very strong interactions. Lastly, it can explain the spontaneous break of the symmetry of the chirality. Understanding these characteristics is of great necessity for researchers to study the interactions among particles and their mechanisms of evolution, structure, and decay 6c; 6d; 6e; 6f; 6g.

The heavy flavor quarkonium is a bound state formed by the heavy flavor quark and anti-quark 7. It plays an important role in the theoretical research of QCD. In hadron induced high energy collisions, the generation of heavy quarkonium can be divided into two processes: One is the appearance of heavy quark pairs and the other one is the evolution of heavy quark pairs into hadrons. The former process can be calculated and analyzed by the perturbative QCD theory 9a; 9b; 9c. Particularly, due to the peculiar reason of the heavy quarkonium, in which the relativistic effect can be neglected, some special theories such as the non-relativistic QCD theory 9d; 9e; 9f can be used to calculate and analyze the production process.

As the basic theory of strong interactions of particles 1; 2; 3; 4; 5; 6, QCD predicts that the hadronic matter can be heated to a very high temperature when it experiences very strong interactions. Then, the system will go through a phase transition from the hadron matter to quark-gluon plasma (QGP) in the process 10; 11; 12. The experiment of relativistic heavy ion collisions is the only way of achieving a QCD phase transition in laboratory conditions 12a; 12b; 12c. Nevertheless, the lifetime of the produced QGP experiencing this phase transition can only reach the order of 10 fm/$c$ (from a few to dozens of fm/$c$) 12d; 12e, which cannot be directly observed in experiments. To detect QGP and study its properties, one has to use an indirect method. For example, one may study the spectrum properties of heavy quarkoniums to obtain the excitation degree (temperature) of emission source which is related to the information on QGP.

$J/\psi$ meson is the bound state of charm and anti-charm ($c\overline{c}$) quarks, where the constituent mass of charm quark is about 1.6 GeV/$c^{2}$ 7. As the first heavy quarkonium discovered experimentally, it has been extensively studied in high energy collisions. In addition, the constraint of $J/\psi$ is considered as an important signal for the generation of QGP 13; 14; 15; 15a; 15b. The yield of $J/\psi$ in electron-positron collisions is higher than that in nuclear collisions 15c; 15d, so the decay of $J/\psi$ in nuclear collisions is an ideal way and medium for studying the hadron spectrum and finding new particles. $\it\Upsilon$ meson is a bound state of bottom and anti-bottom ($b\overline{b}$) quarks, where the constituent mass of bottom quark is about 4.6 GeV/$c^{2}$ 7. The masses of both $J/\psi$ and $\it\Upsilon$ are very large, which leads to the change scale of energy (momentum) in the collision process to the order of GeV (GeV/$c$) when we study their structural properties. In addition, $c+\overline{c}$ and $b+\overline{b}$ can form new $J/\psi$ and $\it\Upsilon$ respectively, which can be researched by non-relativistic approach 9d; 9e; 9f.

We note that the theories and models based on QCD and related idea are complex in the calculation process 9a; 9b; 9c; 9d; 9e; 9f. The complex calculation limits the applications of these theories and models in comparison with experimental data. We hope that we could use a simple idea and formalism to describe uniformly the spectra of various particles, in particular the spectra of heavy quarkoniums such as $J/\psi$ due to its abundant early production in the collisions and wide transverse momentum distribution. In the framework of the multi-source thermal model 16; 17; 18; 18a, we have used the idea of quark composition describing tentatively and uniformly the spectra of various particles in our recent work 64; 64a. It is interesting for us to test further the idea systematically, but the idea of two contributor quarks or partons is used due to the fact that some particles such as leptons have no quark composition. Of course, the quark composition and two contributor quarks for heavy quarkoniums are the same.

In this paper, we test systematically the idea of two contributor partons which contribute to the transverse momentum spectrum of $J/\psi$. The experimental data are collected from gold-gold (Au-Au) 19; 20, deuteron-gold ($d$-Au) 21, lead-lead (Pb-Pb) 22; 23, proton-lead ($p$-Pb) 24; 25; 26; 27, and proton-(anti)proton ($p$-$p(\overline{p})$) collisions 28; 29; 30; 31; 32; 33; 34; 35; 36; 37; 38 over an energy range from dozens of GeV at the Relativistic Heavy Ion Collider (RHIC) to 13 TeV at the Larger Hadron Collider (LHC). Among the RHIC and LHC, there is the Tevatron Proton-Antiproton Collider from which we cited the data in $p$-$\overline{p}$ collisions at 1.96 TeV. These studies are useful for us to understand one of the three basic characteristics of QCD, the color confinement due to the very strong interactions among quarks and anti-quarks.

According to the multi-source thermal model 16; 17; 18; 18a, we may think that a few emission sources are formed in high energy collisions. For nuclear fragments from the projectile and target in nucleus-nucleus collisions, the sources can be nucleons and nucleon clusters. For produced particles such as pions, kaons, and $J/\psi$, the sources can be participant or contributor quarks or gluons, though the contributors $c+\overline{c}$ may be from gluon fusion at the first. The properties of sources can be described by different statistics such as the Boltzmann-Gibbs, Fermi-Dirac, Bose-Einstein, and Tsallis statistics. There are some relations among these statistics due to the fact that they may result in similar or different distributions while describing the spectra of particles.

The Tsallis distribution describes the transverse momentum ($p_{\rm T}$) spectra in wider range than the Boltzmann-Gibbs distribution, though the former is derived from the latter 39; 40; 41. Also, the latter is a special case of the former in which the entropy index $q=1$. Indeed, the former is widely used in high energy collisions from a few GeV to 13 TeV (the top LHC energy) to parameterize the $p_{\rm T}$ spectrum of final-state particles, which justifies its usage in the present work. The form of the Tsallis distribution is expressed as 42; 43; 44; 45; 46; 47; 48

Here, $E$, $p$, $N$, $y$, $m_{0}$, $m_{\rm T}$, $n$, and $T$ denote the energy, momentum, particle number, rapidity, rest mass, transverse mass, power exponent, and effective temperature, respectively. The transverse mass is given by $m_{\rm T}=\sqrt{p_{\rm T}^{2}+m_{0}^{2}}$ 49; 50; 51; 52; 53; 54. In particular, $n=1/(q-1)$, and the entropy index $q$ describes the degree of equilibrium. The closer the parameter $q$ to 1, the more equilibrated the emission source is.

According to the form of Tsallis distribution Eq. (1), as $p_{\rm T}\gg m_{0}$, $m_{0}$ can be ignored, followed by $Ed^{3}N/dp^{3}\propto p_{\rm T}^{-n}$. Then it can be observed that the particles are distributed in accordance with the inverse power law. This is the distribution type of particles produced by the hard scattering process in the high energy collision process 55; 56; 57; 58 and in high $p_{\rm T}$ region. In the non-relativistic limit $(p_{\rm T}\ll m_{0})$ condition, there is $m_{\rm T}-m_{0}=p_{\rm T}^{2}/2m_{0}=E_{\rm T}^{\rm classical}$, showing $Ed^{3}N/dp^{3}\propto e^{-E_{\rm T}^{\rm classical}/T}$, where $E_{\rm T}^{\rm classical}$ is the transverse energy in the non-relativistic limit. We call this distribution the thermodynamic statistical distribution, that is the Boltzmann distribution. Here, we have only discussed the two special cases ($p_{\rm T}\gg m_{0}$ and $p_{\rm T}\ll m_{0}$), though they are not used by us in the present work.

Usually, the empirical formula, the Tsallis–Pareto-type function, is adopted to outline the $p_{\rm T}$ spectrum 59; 60; 61; 62; 63. The general form of the mentioned function is

which is equivalent to Eq. (1) in the form of probability density function, where $C$ is the parameter dependent normalization constant. As the probability density function, Eq. (2) is normalized as $\int_{0}^{\infty}f(p_{\rm T})dp_{\rm T}=1$. In Eq. (2), $n$ and $T$ reflect the degrees of non-equilibrium and excitation of the source respectively. Larger $n$ corresponds to more equilibrium, and larger $T$ corresponds to higher excitation.

In the lower $p_{\rm T}$ range, due to the contribution of light flavor resonance decay, Eq. (2) cannot describe the spectra of light particles very well. For $J/\psi$, the feed-down contribution is more complicated, which is minimal at low $p_{\rm T}$ and grows with growing $p_{\rm T}$. This renders that Eq. (2) also fails to describe the $J/\psi$ spectra. As a result, we ought to empirically add a revised index $a_{0}$ on $p_{\rm T}$ to modify Eq. (2). Then the revised Eq. (2) becomes 64; 64a; 65

Both the normalization constants $C$ in Eqs. (2) and (3) are different, though we have used the same symbol. The two constants are also the parameter dependent. Compared to Eq. (2), Eq. (3) can be used to describe the spectrum in the entire transverse momentum range, having a broader application. For purpose of convenience, as in refs. 64; 64a; 65, we also call Eq. (3) the TP-like function in this work. In the TP-like function, the meanings of $n$ and $T$ remain unchanged as what they are in Eq. (2), though their values may be changed.

The discovery of $J/\psi$ provides a direct evidence for the existence of charm quarks, which makes the study of hadron structure theory presenting a new situation. We may think that in the formation of $J/\psi$ there are two participant or contributor (anti-)charm quarks taking part in the collisions. Let $p_{t1}$ and $p_{t2}$ denote the contributions of quarks 1 and 2 to the transverse momentum of $J/\psi$ respectively. The probability density function $f_{1}(p_{t1})$ ($f_{2}(p_{t2})$) obeyed by $p_{t1}$ ($p_{t2}$) is assumed to be Eq. (3). We have

Here, the two normalization constants $C_{1}$ and $C_{2}$ are parameter dependent. $m_{1}$ and $m_{2}$ are the constituent masses of quarks 1 and 2 respectively, both are 1.6 GeV/$c^{2}$ for charm and anti-charm quarks 7 used in this work. Because of the two quarks taking part in the same collisions, the parameters $n$, $T$, and $a_{0}$ in Eqs. (4) and (5) are separately the same.

The transverse momentum distribution of $J/\psi$ is given by the convolution of two TP-like functions 64; 64a; 65. We have

where the functions $f_{1}(x)$ and $f_{2}(x)$, as well as the parameters $n$, $T$, and $a_{0}$, are given in Eqs. (4) and (5). It should be noted that we have used two contributors. The convolution of two-parton contributions is applicable even for the spectra of particles with complex quark composition. For example, we may use the convolution of two-parton contributions fitting the spectra of various jets 65. As the probability density function, Eq. (6) is applicable in high energy collisions with small or large system, no matter what the density of produced particles is. In addition, Eq. (6) results in similar curve as Eqs. (4) and (5) with different parameters 64; 64a, though the form of Eq. (6) is more complex due to the convolution.

In the above discussions, we assume that $p_{\rm T}=p_{t1}+p_{t2}$, which is operated to describe the relationship among $p_{\rm T}$ of $J/\psi$, $p_{t1}$ and $p_{t2}$ contributed by quarks 1 and 2, respectively. This treatment assumes the azimuth angle $\phi_{1}$ of vector $\vec{p}_{t1}$ being equal to the azimuth angle $\phi_{2}$ of vector $\vec{p}_{t2}$. More generally 64; 64a; 66, if $\phi_{1}\neq\phi_{2}$, we have $p_{\rm T}=\sqrt{p_{t1}^{2}+p_{t2}^{2}+2p_{t1}p_{t2}\cos|\phi_{1}-\phi_{2}|}$. In particular, if $\vec{p}_{t1}$ is perpendicular to $\vec{p}_{t2}$, i.e. $|\phi_{1}-\phi_{2}|=\pi/2$, we have $p_{\rm T}=\sqrt{p_{t1}^{2}+p_{t2}^{2}}$. If $\vec{p}_{t1}$ is opposite to $\vec{p}_{t2}$, i.e. $|\phi_{1}-\phi_{2}|=\pi$, we have $p_{\rm T}=|p_{t1}-p_{t2}|$. Our explorations show that the relationship of $p_{\rm T}=p_{t1}+p_{t2}$ due to $\phi_{1}=\phi_{2}$, i.e. Eq. (6) based on Eqs. (4) and (5), is more easy to fit the data.

If we assume other relationships, Eq. (6) is not applicable. Thus, we need to explore new form of Eq. (3), which is also the topic for us 66; 67. If the analytical expression of Eq. (6) is not available for other relationships, the Monte Carlo method can be used to obtain $p_{t1}$, $p_{t2}$, and $p_{\rm T}$. The $p_{\rm T}$ distribution is then obtained from the statistics. For example, for the more general case of $\phi_{1}\neq\phi_{2}$, we have the expression of $p_{\rm T}=\sqrt{(p_{t1}\cos\phi_{1}+p_{t2}\cos\phi_{2})^{2}+(p_{t1}\sin\phi_{1}+p_{t2}\sin\phi_{2})^{2}}$. This expression can be extended to the case of three or more contributor partons if we add the third or more items in the components. For the special case of $|\phi_{1}-\phi_{2}|=\pi/2$, the analytical expression of Eq. (6) is changed 66; 67, and the form of Eq. (3) is also changed 67. For the special case of $|\phi_{1}-\phi_{2}|=\pi$, the analytical expression of Eq. (6) is applicable, though the form of Eq. (3) is not applicable. In the calculation, the conservations of energy and momentum should be satisfied naturally.

It is understandable that we use $p_{\rm T}=p_{t1}+p_{t2}$ in the present work. From the point of view of energy, we have the relationship among the energy $E$ of $J/\psi$, $E=E_{1}+E_{2}$, where $E_{1}$ and $E_{2}$ are energies contributed by quarks 1 and 2, respectively. In term of transverse mass and rapidity, we have the relationship $m_{T}\cosh y=m_{t1}\cosh y_{1}+m_{t2}\cosh y_{2}$, where $m_{t1}=\sqrt{p_{t1}^{2}+m_{01}^{2}}$, $m_{t2}=\sqrt{p_{t2}^{2}+m_{02}^{2}}$, $m_{01}$ denotes the mass of quark 1, $m_{02}$ denotes the mass of quark 2, $y_{1}$ denotes the rapidity of quark 1, and $y_{2}$ denotes the rapidity of quark 2. In the given narrow rapidity range and neglecting the mass, we have approximately $p_{\rm T}=p_{t1}+p_{t2}$.

Although the case of $\phi_{1}=\phi_{2}$ is a special one, the relationship of $p_{\rm T}=p_{t1}+p_{t2}$ and the convolution of two TP-like functions can fit easily the spectra of various particles and jets 64; 64a; 65. From our point of view, we do not think that the more general case of $\phi_{1}\neq\phi_{2}$ and the special cases of $|\phi_{1}-\phi_{2}|=\pi/2$ and $|\phi_{1}-\phi_{2}|=\pi$ are more easy to fit the spectra due to the calculation itself, though the idea is practicable. In particular, from the point of view of the contributor partons, but not the constituent quarks, the spectra of leptons and jets can be easily fitted from the relationship of $p_{\rm T}=p_{t1}+p_{t2}$ and the convolution of two TP-like functions 64; 64a; 65. This confirms the validity of Eq. (6) based on two contributor sources in the framework of multi-source thermal model 16; 17; 18; 18a.

We would like to emphasize here that we have used two contributor partons as the projectile and target particles/nuclei, no matter what the final-state products are 64; 64a; 65. For $J/\psi$, the two contributor partons (charm and anti-charm quarks) and the constituent quarks are coincidentally equal to each other. For baryons, the two contributor partons are not equal to the three constituent quarks. For jets, the two contributor partons are not equal to the sets of two or three constituent quarks, too. For leptons, the two contributor partons have no corresponding constituent quarks. We may consider that two light or slow contributor partons produce leptons and baryons, while two heavy or fast contributor partons produce jets, no matter what the structures of leptons, baryons, and jets are. In fact, the two contributor partons are regarded as two energy resources, but not the constituents.