QCD Corrections to the Color-singlet J/\psi Production in Deeply Inelastic Scattering at HERA

QCD Corrections to the Color-singlet Production in Deeply Inelastic Scattering at HERA

Zhan Sun School of Science, Guizhou Minzu University, Guiyang, 500025, P. R. China.    Hong-Fei Zhang hfzhang@ihep.ac.cn (corresponding author) Department of Physics, School of Biomedical Engineering, Third Military Medical University, Chongqing 400038, China.
July 13, 2019

We present the first study of the QCD corrections to the color-singlet (CS) production in deeply inelastic scattering at HERA. The -factor ranges from to in the kinematic regions we study. In low transverse momentum regions, the -factors is even smaller, and close to 1, which indicates good convergence of the perturbative expansion. With the QCD corrections, the CS cross section is still below the data. At least at QCD next-to-leading order, the color-octet mechanism is necessary to describe the data.

12.38.Bx, 12.39.St, 13.85.Fb, 14.40.Pq

Since the discovery of the meson Aubert et al. (1974); Augustin et al. (1974), the production mechanism of heavy quarkonia has been a hot issue in high energy physics. Before the first measurement of the and hadroproduction carried out by the CDF Collaboration Abe et al. (1992), the color-singlet (CS) model Einhorn and Ellis (1975); Ellis et al. (1976); Carlson and Suaya (1976); Chang (1980); Berger and Jones (1981); Baier and Ruckl (1981, 1982) was generally accepted as a natural description of the heavy quarkonia production and decay mechanism. In 1994, the nonrelativistic QCD (NRQCD) effective theory Bodwin et al. (1995) was proposed, which successfully filled the huge gap between the QCD leading order (LO) predictions via the CS model and the CDF measurement of the and hadroproduction Braaten and Fleming (1995); Cho and Leibovich (1996a, b). At QCD LO in the NRQCD framework, the dominant hadroproduction mechanism of the mesons is the gluon fragmentation into a color-octet (CO) pair, which produces a by emitting soft gluons in a long-distance process. Although the CS model also permits a gluon fragmenting into a with additional two hard gluons emitted, according to NRQCD, this mechanism substantially underestimates the production rate. In addition to the hadroproduction, NRQCD also worked well in many other processes, including the production in fusion Klasen et al. (2002), the photoproduction at HERA Butenschoen and Kniehl (2010), the meson hadroproduction Ma et al. (2011); Zhang et al. (2016), and etc. However, one cannot overlook the controversies NRQCD is facing, among which the polarization puzzle is the most well known and challenging one. Many independent studies Butenschoen and Kniehl (2012); Chao et al. (2012); Gong et al. (2013); Shao and Chao (2014); Shao et al. (2014); Bodwin et al. (2014); Shao et al. (2015); Bodwin et al. (2016) at QCD next-to-leading order (NLO) level have been performed. Those who achieved good description of the polarization data are generally consistent with the dominance picture. However, the theoretical studies on the hadroproduction indicate that Butenschoen et al. (2014); Han et al. (2015); Zhang et al. (2014), regarding the heavy quark spin symmetry, this picture violates the recent LHCb measurement Aaij et al. (2015). This paradox was remedied in Reference Sun and Zhang (2015) which found that, even without the dominance picture, the polarization data can also be understood within the NRQCD framework.

To solve these problems, a new factorization theory was proposed in Reference Ma and Chao (2017). From another angle of view, some researchers challenge the significance of the CO contributions, and seek the way of describing the data within the CS framework. The production in annihilation provides an example of success of this idea. With the QCD and relativistic corrections Zhang and Chao (2007); Ma et al. (2009); Gong and Wang (2009a, b); He et al. (2010), the CS contributions almost saturate the Belle data Pakhlov et al. (2009), while the inclusion of the CO ones will generally ruin the agreement between theory and experiment. Looking back at the hadroproduction, QCD NLO corrections Campbell et al. (2007) enhance the differential cross sections for the CS hadroproduction in medium and high transverse momentum () regions by one to two orders of magnitude, which reduces the discrepancy between theory and data, at the same time, change the polarization from transverse to longitudinal Gong and Wang (2008). Due to the lack of the complete next-to-next-to-leading order results, one cannot yet draw definite conclusions on the significance of the CO contributions implied by the hadroproduction data (as a review, see e.g. Reference Brambilla et al. (2011)). Such large -factors are due to the fact that at QCD LO, both the leading power (LP) and next-to-leading power (NLP) terms vanish, and at QCD NLO, the NLP behaviour arises; this new behaviour enhances the cross sections significantly, especially in high regions.

The production in deeply inelastic scattering (DIS), which is also called the leptoproduction, can serve as another test of the quarkonium production mechanisms. The CS photoproduction has been studied at in Refs. Kramer et al. (1995); Kramer (1996); Artoisenet et al. (2009); Chang et al. (2009); Li and Chao (2009), which found that with the QCD corrections the CS contributions are still below the data, especially when is large. For the leptoproduction, the deflection angle of the scattered lepton is larger, accordingly the virtuality of the incident photon, which is emitted by the incident lepton and will interact in the hadronic process, need to be taken into account in the perturbative calculation. We define , where is the momentum of the incident photon. Due to larger , the leptoproduction shows better features than the photoproduction. The of the yield data measured at HERA is relatively low, thus the ratio () of the proton momentum taken by the interacting parton might be very small for the photoproduction, and the gluon saturation effects Mueller and Qiu (1986) would be important. Larger can increase the value of , correspondingly suppress the gluon saturation effects. Moreover, as increases, the contributions from the resolved photons are greatly suppressed, and the perturbative expansion in will be better as well. Another interesting feature of the leptoproduction is that the NLP behaviour arises at QCD LO. Unlike the hadroproduction case, no new behaviour emerges in the production in DIS at QCD NLO, thus, the convergence of the expansion should be better.

However, the emergence of at the same time makes the computation much more complicated. We notice that although the HERA collaborations have published abundant data, the one-loop level phenomenological study of the production in DIS is still lacking. The production in DIS has been studied at QCD LO in many papers Duke and Owens (1981); Baier and Ruckl (1982); Korner et al. (1982); Guillet (1988); Merabet et al. (1994); Fleming and Mehen (1998); Yuan and Chao (2001); Kniehl and Zwirner (2002), however, as indicated in our recent work Zhang and Sun (2017), all those Fleming and Mehen (1998); Yuan and Chao (2001); Kniehl and Zwirner (2002) under the NRQCD framework are based on a formalism that will lead to wrong results when the ranges of the or rapidity () are not taken to cover all their possible values. For this reason, we provided a renewed QCD LO study of the production in Reference Sun and Zhang (2017). In this paper, we study the QCD corrections to the CS production in DIS, following the calculation framework provided in Refs. Zhang and Sun (2017); Sun and Zhang (2017).

Figure 1: Representative feynman diagram for the leptoproduction.

The process for the leptoproduction is illustrated in Figure 1. , , , and are the momenta of the incident and scattered electron, the proton, the parton generated from the proton, and the produced , respectively. The generally used invariants are defined as


In our calculation, all the processes up to are counted, including at both tree and one-loop level, and , in which runs over , , , , and . Here we use and to denote a gluon and a light quark (anti-quark), respectively. Since all the singularities cancel in the hadronic process, namely , where denotes a virtual photon, we will directly employ the form of the leptonic tensor given in Reference Zhang and Sun (2017), which eventually leads to the same results as those by adopting the -dimensional form of the leptonic tensor. The contraction between the leptonic tensor and the hadronic one can be carried out in 4-dimensions. Thus, the form of the leptonic tensor given in Reference Zhang and Sun (2017) can be directly employed without extension to its -dimensional form. If we use the two-cutoff phase space slicing method Harris and Owens (2002) to separate the divergences in the real-correction processes, the phase space for the scattered electron and the can also be written in 4-dimensions. Then the short-distance coefficient (SDC) for the CS production in DIS can be expressed as Sun and Zhang (2017)


where runs over all the partonic processes, is the fine structure constant, is the spin and color average factor, is the parton distribution function (PDF) of a parton in a proton with and being the fraction of the proton momentum taken by the parton and the factorization scale, respectively, and are the transverse momentum of the and the azimuthal angle of the lepton plane around the axis, respectively, and for the processes , , while for the real-correction processes, namely ,


Note that Equation 2 is valid in any frame of reference. Since has been integrated over to eliminate the function which keeps the energy conservation, the value of has been fixed in Equation 2 as .

The expressions for can be found in Reference Sun and Zhang (2017) and are duplicated in the following as






Hereinafter, we denote all the physical quantities in the center-of-mass frame by a superscript . are defined as


where is the hadronic tensor. Note that here the long-distance matrix elements (LDMEs) have been eliminated from the hadronic tensors. Then the cross section for the CS production can be expressed as


where the CS LDME for the production, , is abbreviated as .

All the singularities are contained in ’s and , thus, we evaluate them in -dimension.

We denote as the for the process at tree-level, as the virtual corrections to , and as the sum of all the ’s for the real-correction processes and define


One need to be more careful in identifying the soft and hard collinear regions, since the squared invariant mass of the incident photon is negative. The soft region is defined in terms of the energy of the final-state gluon, , in the rest frame by , while the collinear region is defined by the inequality that the inner product of two massless momenta is smaller than , where is defined as in Equation 1. Here, and are two arbitrary real numbers, however, small enough to make sure that integrals of a finite function in the soft and collinear regions are negligible. Under these definitions, most of the formulas in Reference Harris and Owens (2002) remain unchanged. To avoid double counting, the gluon-soft regions need to be excluded in the calculation of the cross sections in the hard collinear regions. Accordingly, the integral domain of the ratio defined in Reference Harris and Owens (2002) should be properly determined. To distinguish this ratio and the inelasticity coefficient, we assign another symbol, , to this ratio. When and are collinear, is defined as . For and , the integral domain of should be , where , and is the invariant mass of all the hadronic final states other than and in the partonic process. In the process we study, is the mass, which is set to be twice of the -quark mass, , namely . For , the integral domain of should be . When and is collinear, the definition of is different, which should be . The integral domain of should be for , and for , where . Here is the ratio of the parton momentum, , to the incident proton momentum. The expressions of and are different from those given in Reference Harris and Owens (2002) due to the nonzero . With the above configurations, the sum of , and are divergence free. Defining


one can rewrite the SDC for the CS leptoproduction as


where the subscript means the integral is carried out in the hard noncollinear region, and and are evaluated in the limit .

To evaluate the ’s for each process, we employ our new Mathematica package, Malt@FDC. This package can automatically reduce the loop amplitudes into linear combination of master integrals, which will be computed numerically with Looptools Hahn and Perez-Victoria (1999), and simplify the expressions of the squared amplitudes. Our final expressions contain the terms of the and functions, which is not given in the Looptools library. These functions are computed with a new FORTRAN package, which will be discussed elsewhere Zhang and Sun (tion). Before working on the current process, we have applied our Malt@FDC in tens of other processes. All our results are consistent with those obtained by the FDC system Wang (2004) and/or those given in published papers. As an indispensable check, we studied the limit and compared the results with the photoproduction. Setting to be zero, replacing the leptonic tensor associated with the virtual photon propagator by , and implementing proper phase-space integration, we can reproduce the photoproduction results in Refs. Chang et al. (2009); Li and Chao (2009). Another check to mention is that our results are independent of and and gauge invariant.

Abundant leptoproduction data Adloff et al. (1999, 2002); Aaron et al. (2010); Chekanov et al. (2005) have been measured at HERA. However, most of them lie in the kinematic regions where the perturbative calculation might not be good. In order to make the perturbation theory work better, we constrain our concerns in the regions where and are not too small and is not too large. Adopting the selection criteria, , , and , we find that only one set of data is availabe, which is presented in Reference Adloff et al. (2002). Therein, the differential cross sections with respect to , , , , and are measured. To be consistent with the HERA convention, the forward direction in the rest frame are defined as that of the incident virtual photon.

Due to the -parity and color conservation, cannot be produced at . In our calculation, we completely omit the feed down contributions from . The contributions to the production from the feed down has been estimated in Reference Adloff et al. (2002). The diffractive are produced basically in large regions. In the region , about of the events come from this resource. The cross sections of the inelastic production have the same behavior with that of the production. Throughout the kinematic region we study, the contributions from the inelastic feed down are estimated to be about . The production from decay is expected in low region, and is estimated to be about in the region  Adloff et al. (2002). All these contributions are not included in our calculation. We will address them when comparing our results to data.

To present the numerical results, we adopt the following parameter choices. The -quark mass , and the electromagnetic coupling constant . The energy of the incident lepton and proton in the laboratory frame are and , respectively. For the calculation of the production in DIS at QCD LO, the one-loop running and the parton distribution function (PDF) CTEQ6L1 Pumplin et al. (2002) are employed, while for the calculation up to QCD NLO, the two-loop running and the PDF CTEQ6M Pumplin et al. (2002) are employed. The value of at the boson mass is set to be . The renormalization and factorization scales are set to be as our default choice. The wave function at the origin is determined in terms of  Eichten and Quigg (1995), and the CS LDME can be evaluated as . We notice that there are many parallel extractions of the CS LDME with different strategies and data. To be consistent with our calculation, we list here the value obtained through at QCD NLO in Reference Braaten and Lee (2003), i.e. , that extracted from the hadroproduction data,  Zhang et al. (2014), and that obtained from the potential model,  Eichten and Quigg (1995). We will find that the uncertainties in the LDMEs do not affect our phenomenological conclusions.

Figure 2: The differential cross sections of the CS production in DIS with respect to , , , , and . The bands are obtained by varying from to . The data are taken from Reference Adloff et al. (2002)

Our numerical results are presented in Figure 2, including the differential cross sections with respect to , , , , and . In the kinematic regions we study, the -factor ranges from to . As increases, the NLO corrections become more significant. In the largest bin, , the -factor reaches its maximum value, . For the bins, , and , the -factors are , and , respectively. As we expected, the convergence of the perturbative expansion becomes better as increases. One can easily see from the plots that, when and is not too large, the -factor is quite close to 1, which indicates good convergence of the perturbative expansion in . To study the uncertainties brought in by the uncertainties of the LDMEs, we present the bands in Figure 2 covering the results for and . The upper bound of the band is generally less than 1.5 times of the central curve.

Since the feed down contributions are not taken into account in our calculations, we need to estimate their magnitude while comparing our theoretical results to data. According to our earlier discussions, the feed down contributions are about 30% of the total production cross sections. In contrast, the ratio of the central value of the experimental data to our NLO results range from 1.4 to 4.9, among which, only 5 out of the 21 data points are smaller than 2. We can conclude that even with the feed down contributions included, the theoretical results still cannot describe the data. At least at the NLO precision, the CO mechanism is important and necessary for understanding the production in DIS.

In summary, we studied the QCD corrections to the production in DIS. This process is much more complicated than the photoproduction and hadroproduction ones due to the nonzero . In the kinematic regions that HERA experiment concerns, we found that the -factors are close to 1, which indicates good convergence of the perturbative expansion. With the NLO corrections included, the CS contributions are still much smaller than the experimental data. To this end, our study iterated the importance of the color-octet mechanism.

We are indebted to Professor Geoffrey Bodwin for helpful discussions. This work is supported by National Natural Science Foundation of China (Nos. 11405268, 11647113 and 11705034).


Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description