Radiative transitions between $0^{-+}$ and $1^{--}$ heavy quarkonia on the light front
Abstract
We present calculations of radiative transitions between vector and pseudoscalar quarkonia in the light-front Hamiltonian approach. The valence sector light-front wavefunctions of heavy quarkonia are obtained from the Basis Light-Front Quantization (BLFQ) approach in a holographic basis. We study the transition form factor with both the traditional “good current” and the transverse current (in particular, ). This allows us to investigate the role of rotational symmetry by considering vector mesons with different magnetic projections (). We use the state of the vector meson to obtain the transition form factor, since this procedure employs the dominant spin components of the light-front wavefunctions and is more robust in practical calculations. While the states are also examined, transition form factors depend on subdominant components of the light-front wavefunctions and are less robust. Transitions between states below the open-flavor thresholds are computed, including those for excited states. Comparisons are made with the experimental measurements as well as with Lattice QCD and quark model results. In addition, we apply the transverse current to calculate the decay constant of vector mesons where we obtain consistent results using either or light-front wavefunctions. This consistency provides evidence for features of rotational symmetry within the model.
I Introduction
Radiative transitions offer insights into the internal structure of quark-antiquark bound states through electromagnetic probes. The radiative transition between (pseudoscalar) and (vector) mesons via the emission of a photon is characterized as the magnetic dipole (M1) transition. This transition mode is known to be sensitive to relativistic effects Godfrey and Rosner (2001), especially for those between different spatial multiplets, such as .
Heavy quarkonium is often dubbed as the “hydrogen atom” of quantum chromodynamics (QCD). It provides an ideal testing ground for various investigations to understand QCD. States below the open flavor threshold ( for charmonium and for bottomonium) have very narrow widths since they cannot decay via any Okubo-Zweig-Iizuka allowed strong decay channels Okubo (1963); ?; ?. Electromagnetic transition rates are therefore important. Comparing the theoretical and experimental rates for radiative transitions then provides guidance to improve our understanding of the internal structure of heavy quarkonia. Recently, some of us proposed a model based on the light-front holographic QCD and the one-gluon exchange with a running coupling Li et al. (2016, 2017); Tang and et al. (). In this model, charmonia and bottomonia are solved as relativistic quark-antiquark bound states using Basis Light-Front Quantization (BLFQ), a nonperturbative Hamiltonian approach within light-front dynamics Vary et al. (2010). The model provides a reasonable description of the mass spectrum and other properties that were studied. Observables including decay constants, r.m.s. radii Li (2017), distribution amplitudes and parton distributions as well as diffractive vector meson productions Chen et al. (2017) have been directly calculated from the light-front wavefunctions (LFWFs), and are in reasonable agreement with experiments and with other approaches (see also Ref. Adhikari and et al. ()). Therefore, we are motivated to investigate radiative transitions within this model. On the one hand, we hope to further test the model by comparing with the existing experimental results. On the other hand, results calculated for transitions that have not yet been measured provide predictions for future experiments.
In this work, we derive the formulae for radiative transitions between and mesons on the light front, using both the traditional “good current” and the transverse current . Though, in principle, these two choices should be equivalent due to Lorentz covariance, adoption of certain approximations in the model may lead to violation of the Lorentz symmetry that would be evident through inequivalent results. In nonrelativistic quantum mechanics, magnetic moments and transitions can only be extracted from the current density rather than the charge density . Therefore one may expect that for the M1 transitions in nonrelativistic systems such as heavy quarkonia, the transverse current density could be better than the charge density . Specifically, as we will see later, the transverse current allows us to extract the transition form factor through the state of the vector meson, which is not accessible with . The calculation with the state provides a more robust result by employing the dominant spin components of the two mesons (spin-triplet for the vector and spin-singlet for the pseudoscalar) in the transition, while that with always requires subdominant components of the LFWFs with relativistic origins. So in this work, we obtain the transition form factors with the states of the vector mesons through the transverse current. The results from , though less robust, are also presented for comparison. In addition, as a cross check, we revisit the decay constants by utilizing the transverse current to compare with the previous calculation using in Ref. Li et al. (2017). It follows that different components of the vector mesons are involved. This provides us with a different yet pertinent perspective to understand the degree of Lorentz symmetry manifestation in the current model.
The layout of the paper is as follows. In Sec. II, we introduce the formalism and methods to calculate the M1 transition form factor on the light front. In Sec. III, we apply the formalism to heavy quarkonia in the BLFQ approach. Sec. IV presents the numerical results of transition form factors. In Sec. V, we perform the calculations of vector meson decay constants with different magnetic projections. We conclude the paper in Sec. VI.
Ii Transition form factors on the light front
ii.1 Transition form factor and decay width
The matrix element for the radiative transition between a vector meson () with four-momentum and polarization and a pseudoscalar () with four-momentum via emission of a photon, can be parametrized in terms of the transition form factor as Dudek et al. (2006),
(1) |
where we define , with representing the four-momentum of the photon. and are the masses of the pseudoscalar and the vector, respectively. is the polarization vector of the vector meson. is the current operator.
In the physical process of , the photon is on shell (). The transition amplitude is:
(2) |
where is the polarization vector of the final-state photon with its spin projection . The decay width of follows by averaging over the initial polarization and summing over the final polarization. In the rest frame of the initial particle, it reads,
(3) |
The momentum of the final photon is determined by the energy-momentum conservation, . is the spin of the initial vector meson. To calculate the width of , exchange and , and replace with for the initial pseudoscalar in Eq. (3).
ii.2 Light-front dynamics
In principle, the Lorentz invariant function defined in Eq. (1) can be extracted from any of the four sets of hadron matrix elements, (, see definitions of the light-front variables in the Appendix). However, results from different current components may be different due to violations of Lorentz symmetry introduced by the Fock sector truncation as well as by the modeling of systems. These approximations have led to extensive discussions in the literature Carbonell et al. (1998); de Melo et al. (1998); Brodsky and Hwang (1999); Suzuki et al. (2013); Melikhov and Simula (2002). The “+” component, known as the “good current”, is typically used, together with the Drell-Yan frame (), to avoid contributions from pair production/annihilation in vacuum. The transverse components have been shown to be consistent with the “+” component in the limit of zero momentum transfer in certain theories, such as the theory Brodsky and Hwang (1999) and the spin-0 two-fermion systems Melikhov and Simula (2002). Another option, the “-” component, is known as the “bad current”, due to its association with the zero-mode contributions.
Here, we present formulae for the transition form factor, for both and , along with different magnetic projections () of the vector meson. Note that when the rotational symmetry on the transverse plane is preserved, which is usually the case, using or component or even combinations of the two are equivalent. In particular, we use to carry out the calculation in the case of the transverse current. For any transverse vector , which is expressed as in the Cartesian coordinate or in the polar coordinate, we will write its complex form as and . From the vector decomposition in Eq. (1),
(4) | ||||
(5) |
where we have introduced two variables and , which are invariant under the transverse Lorentz boost specified by the velocity vector ,
(6) |
This boost is kinematic and survives the Fock space truncation, whereas the full Lorentz transformation does not. The two sets of hadron matrix elements in Eqs. (4) and (5) can be related through such a boost,
(7) |
By applying the above relation to Eqs. (4) and (5), we find that for , and should give the same . For , on the other hand, cannot be extracted from , but can be extracted from transverse currents, such as .
(8) | |||
(9) |
Note that for the purpose of comparison, we label the transition form factors with their corresponding current components and the values of the vector wavefunctions. In practice, the different prescriptions of extracting the same transition form factor could provide a test of violation of the Lorentz symmetry in the calculation. In the covariant light-front dynamics, the transition form factor is extracted from combinations of several hadron matrix elements Carbonell et al. (1998).
ii.3 Impulse approximation
In the impulse approximation, the interaction of the external current with the meson is the summation of its coupling to the quark and to the antiquark, as illustrated in Fig. 1. The vertex dressing as well as pair creation/annihilation from higher order diagrams are neglected.
The hadron matrix element can be written accordingly as a sum of the quark term and the antiquark term:
(10) |
The current operator is defined as where is the quark field operator with flavor (). and are the normal ordered pure quark () and antiquark () part of , respectively, where () is the quark (antiquark) annihilation operator. The dimensionless fractional charge of the quark is, for the charm quark and for the bottom quark. The electric charge . For quarkonium, due to the charge conjugation symmetry, the antiquark gives the same contribution as the quark to the total hadronic current. So, for our purpose, we calculate the hadron matrix element for the quark part. As such, we compute which is related to by
The hadron matrix element can be written explicitly in terms of the convolution of LFWFs. To begin with, the valence Fock space representation of quarkonium reads:
(11) |
where the color index , and the number of quark colors . is the LFWF written in relative coordinates. is the longitudinal light-front momentum fraction, is the relative transverse momentum, where is the single-particle 4-momentum of the quark. represents the fermion spin projection in the direction.
The hadron matrix element in the Drell-Yan frame (, i.e. ) follows,
(12) |
We could then obtain the transition form factor from such hadron matrix elements according to Eqs. (8) and (9). Ideally, is independent of the spin projection and the current components. Nevertheless, one needs to carefully choose the proper matrix elements to evaluate certain quantities, when approximations break the Lorentz symmetry Li et al. (2018). For instance, there are different ways of choosing matrix elements to calculate the spin-1 electromagnetic form factors when the angular condition is violated Grach and Kondratyuk (1984). Among those, some are preferred in the sense that unphysical terms could be partially or entirely suppressed Karmanov (1996); Melikhov and Simula (2002).
In the case of the M1 transition form factor , using the combination of the transverse current with the polarization of the vector meson, according to the expression in Eq. (9), would give the LFWF representation as,
(13) |
where we define . Note that in deriving Eq. (13), we have taken advantage of symmetries in LFWFs, and .
The traditionally used good current is also worth looking at. As we have discussed, with this current component, the transition form factor can be extracted only from the polarizations of the vector meson. We present the expression for according to Eq. (8), while the expression for is similar. It is evident from this expression that the overlapped spin components of the two wavefunctions indicate no spin-flip (between spin-triplet and spin-singlet), which may appear counter-intuitive for the M1 transition. Indeed, this calculation relies on subdominant terms and is less robust, as we will discuss in the following section for heavy quarkonia.
(14) |
ii.4 The nonrelativistic limit
In the nonrelativistic limit, the M1 transition with the same radial or angular quantum numbers (e.g. ), is often referred to as allowed, for which the transition amplitude is large and as a result of the similarity between the spatial wavefunctions of the vector and the pseudoscalar mesons with the same spatial quantum numbers; whereas the transition between states with different radial or angular excitations is referred to as hindered, for which the transition amplitude is zero and at leading order due to the orthogonality of the wavefunctions Godfrey and Rosner (2001); Brambilla et al. (2006); Lewis and Woloshyn (2011); Eichten et al. (2008). The deviations of experimentally measured results from those nonrelativistic limits indicate relativistic effects Patrignani et al. (2016). For a heavy quarkonium system, which is close to the nonrelativistic domain, such deviations are expected to be small but nonzero.
The wavefunctions of heavy quarkonia, treated as relativistic bound states, are dominated by those components that are non-vanishing and reduce to the nonrelativistic wavefunction in the nonrelativistic limit. These wavefunction components are therefore referred to as the dominant components. It is necessary to emphasize that despite the correspondence between the dominant spin components and the nonrelativistic wavefunctions, the former carries relativistic contributions when solved in a relativistic formalism. There are also wavefunction components of purely relativistic origin, which vanish in the nonrelativistic limit and are therefore subdominant. In practice, the dominant components tend to be better constrained, while the subdominant ones are more sensitive to the model and numerical uncertainties. For the pseudoscalar states resembling S-waves (in particular ), and , their dominant components are the spin-singlets , while relativistic treatments would also allow them to have subdominant components, such as . Analogously, for the vector states of heavy quarkonia resembling S-waves (in particular ), and , the dominant components are the spin-triplets, , and . For those vector states identified as D-waves, and , where orbital excitations occur, all the spin-triplet components , and exist in the nonrelativistic limit and are considered dominant, and only the spin-singlet components are subdominant. In detail, the spin components with larger orbital angular momentum projection , and , have the largest occupancy. The less occupied components, , and , could also exist in the nonrelativistic limit. Moreover, the spin components with admit the admixtures of S-waves, though the actual amount of such admixtures is small and sensitive to both the model parameters and the truncation. For example, the [] state, though primarily a state, has contributions from states (notably ) Richard (1980); Rosner (2005); Eichten et al. (2004, 2006), and these S-wave admixtures are responsible for the transitions Eichten et al. (1978); Kwong and Rosner (1988); Rosner (2001); Eichten et al. (2008). In order to have a more intuitive view of the dominant and subdominant spin components for those states, we take the LFWFs from Ref. Li et al. (2017) to show in Fig. 2 the proportions of those dominant and subdominant components of heavy quarkonia. For all those pseudoscalar and vector states, the dominant terms could each occupy of the whole LFWF, suggesting that the heavy quarkonium indeed resembles a nonrelativistic system. The comparison between the same states of the charmonium and those of the bottomonium also reveals that the dominant component is more pronounced in the heavier, and less relativistic, system.
It follows that in calculating the transition form factors, we can examine the two procedures, presented in Eq. (13) and presented in Eq. (14), in terms of their proximities to the nonrelativistic domain. The result of mainly comes from the overlap of the dominant components, , whereas even the major part of involves the subdominant components, such as and . In heavy quarkonium, the dominant components tend to be better constrained than the subdominant ones which suggests that is more robust than .
The nonrelativistic limit can be achieved for by adopting nonrelativistic wavefunctions where only the dominant spin components exist. However, with , simply taking the nonrelativistic wavefunction would always lead to zero since the expression in Eq. (14) involves the vanishing subdominant terms. To be specific, we examine the transition form factors at , where they can be interpreted as the overlaps of wavefunctions in coordinate space [], shown in Eqs. (15) and (16). Though equivalent to Eqs. (13) and (14) at respectively, Eqs. (15) and (16) do not have the troubling factor of , and are therefore more intuitive for the purpose of illustration.
(15) | |||
(16) |
Note that in the nonrelativistic limit, the wavefunctions of the respective pseudoscalar and vector states with the same radial and angular numbers are identical in their spatial dependence, and they only differ in their spin structures. For the allowed transition, due to the normalization of the spatial wavefunctions, which can be seen from Eq. (15) along with taking Li et al. (2017) and small hyperfine splitting . For the hindered transition, due to the orthogonality of the two wavefunctions. Such a nonrelativistic reduction that takes advantage of the near orthonormality of wavefunctions is reminiscent of the nonrelativistic quark model (see Refs. Godfrey and Rosner (2001); Brambilla et al. (2006); Lewis and Woloshyn (2011)). However, for , the realization of the nonrelativistic limits depends strongly on the details of the subdominant wavefunctions which are less constrained in the parameter fitting. For the hindered transition, where cancellation occurs, this leads to a strong sensitivity to the model parameter and potentially to the truncation. Fig. 3 presents the integrands (those inside ) of and according to Eqs. (15) and (16) for an allowed () as well as a hindered () transition. In the left panel of Fig. 3, the integrands of the allowed transition have no nodes resulting from the coherent overlaps of the two wavefunctions. On the other hand, the right panel of Fig. 3 shows significant cancellations of contributions from the integrands which change sign due to nodes in the wavefunctions.
Based on these lines of reasoning, we take , using the transverse current, to evaluate the transition form factors for heavy quarkonia. The less robust , using the plus current, which has strong sensitivity to the violation of rotational symmetry will also be presented for comparison.
Iii Calculation in Basis Light-front Quantization
We adopt wavefunctions of heavy quarkonia from recent work Li et al. (2016, 2017) in the BLFQ approach Vary et al. (2010). The effective Hamiltonian extends the holographic QCD Brodsky et al. (2015) by introducing the one-gluon exchange interaction with a running coupling Wiecki et al. (2015). The starting point for the model of Refs. Li et al. (2016, 2017) is transverse light-front holography, inspired by string theory, which approximates QCD at long distance. As a complementary part, the longitudinal confining potential is introduced to allow a more consistent treatment of the mass term and the longitudinal excitation. The one-gluon exchange implements the short-distance physics and determines the spin structure of the mesons. The mass spectrum and LFWFs are the direct solutions of the eigenvalue equation, and are obtained by diagonalizing the Hamiltonian in a basis representation. The spectrum agrees with the PDG data with an r.m.s mass deviation of 30 to 40 MeV for states below the open flavor thresholds. The LFWFs have been used to produce several observables and are in reasonable agreement with experiments and other theoretical approaches Leitão et al. (2017). We now use these same LFWFs to calculate radiative transitions with the formalism described in Sec. II.
The LFWFs are solved in the valence Fock sector using a basis function representation:
(17) |
In the transverse direction, the 2D harmonic oscillator (HO) function is adopted as the basis. In the longitudinal direction, we use the modified Jacobi polynomial as the basis. is the orbital angular momentum projection, related to the total angular momentum projection as , which is conserved by the Hamiltonian. The basis space is truncated by their reference energies in dimensionless units:
(18) |
As such, the -truncation provides a natural pair of UV and IR cutoffs: , , where is the oscillator basis energy scale parameter. represents the resolution of the basis in the longitudinal direction. See Ref. Li et al. (2017) for details on basis functions and parameter values. The LFWFs are calculated at and . Transition form factors are computed at each of these basis truncations. Fig. 4 shows the convergence trends of as functions of . The left panel compares three different fitting functions to extrapolate our results obtained at finite basis sizes to the complete basis by taking the limit . The extrapolations using these three functions agree to within of each other. For the remainder of this paper we adopt the second-order polynomial in for our extrapolations, fitted to the 4 basis sizes, with an extrapolation uncertainty given by the difference between the result in the largest basis () and the extrapolated value. (Note that this uncertainty does not include any systematic uncertainty coming from the model for the interaction or from the Fock space truncation.) In the right panel of Fig. 4 we show our results for all allowed transitions at finite basis sizes, together with our extrapolation to the complete basis, including our extrapolation uncertainty estimate. Note that they are all close to the nonrelativistic limit 2, which is expected according to our discussion in Sec. II.4. For the hindered transitions, the uncertainties from such basis extrapolations are comparatively larger, since their calculations are more sensitive to the details of wavefunctions.
Iv Results
Here we present our results for selected pseudoscalar-vector transition form factors for charmonia and bottomonia below their respective open flavor thresholds. Fig. 5 shows our numerical results of the transition form factors in three groups, the allowed transition , the radial excited transition and the angular excited transition , through a progression from upper to lower panels. As already discussed in the previous section, for the allowed transitions we find , whereas for the hindered transitions involving either radial or angular excitations we have . Transitions involving higher radial excited states feature more wiggles in the curve, which is especially evident in the transitions as increases. This is because the radial excited states have transverse nodes. As a result, the transition form factors, in the form of their convolutions [see Eq. (13)], are not monotonic. The comparison between charmonia and bottomonia is also of interest. For comparable transition modes, the transition form factors show similarity in their patterns as well as their behaviour as a function of . Furthermore, as illustrated in the second row of panels in Fig. 5, one observes that the transition form factors are very similar to the form factors for .
Comparisons of from this work, with experiments (compiled by PDG Patrignani et al. (2016)) and with other models (Lattice QCD Dudek et al. (2009); Bečirević and Sanfilippo (2013); Bečirević et al. (2015); Donald et al. (2012); Hughes et al. (2015), Quark Model Ebert et al. (2003); Barnes et al. (2005); Godfrey and Moats (2015)) are collected in Table 1 and visualized in Fig. 6. Most calculations, as well as available experimental data, give a value of the the order of 2 for the allowed transitions : all such data in Table 1 are between 1.5 and 2.5 with only one exception, the relativistic quark model calculation of . This is in agreement with the vector and the pseudoscalar mesons possessing very similar spatial wavefunctions, but different spin structures. On the other hand, there is a significant spread in the theoretical results of the hindered transitions. This is expected because the hindered transitions involve changes in radial quantum numbers and/or orbital angular motions and are sensitive to delicate cancellations as discussed above. Considering the fact that only two free parameters are employed by the model for quarkonia in Ref. Li et al. (2017) and the fact that we did not adjust any parameters in our calculation for the transitions, the agreement to within an order of magnitude is encouraging.
The results with the “+” current, in combination with the vector meson wavefunctions, , are presented as a ratio to in Fig. 7. As already mentioned, we expect these calculations to be much less robust. This is because the calculation of according to Eq. (14), depends on subdominant components of the wavefunctions, which is less constrained from the model. Indeed, the dependence of these calculations on the basis truncation is much larger, resulting in significantly larger extrapolation uncertainties. Furthermore, the hindered transitions have a much larger fluctuation than the allowed transition, due to their sensitivity to the subdominant components in one of the two spatial wavefunctions with different radial quantum numbers and/or different orbital motions. Our results with the “+” component of the current , differ by up to 2 orders of magnitude from our more reliable results with the transverse component of the current .
PDG Patrignani et al. (2016) | BLFQ | Lattice Dudek et al. (2009); Bečirević and Sanfilippo (2013); Bečirević et al. (2015); Donald et al. (2012); Hughes et al. (2015); Lewis and Woloshyn (2012) | Quark Model | |||||
() | Dudek et al. | Bečirević et al. | HPQCD | NRQCD | rQM Ebert et al. (2003) | GI Barnes et al. (2005); Godfrey and Moats (2015) | ||
1.21 | 1.82 | |||||||
0.099 | 0.20 | |||||||
0.097 | 0.31 | |||||||
2.01 | 2.18 | |||||||
1.48 | 1.87 | |||||||
0.11(1) | 0.050 | 0.12 | ||||||
0.078(10) | 0.036 | 0.061 | ||||||
0.062(10) | 0.050 | 0.17 | ||||||
2.17 | 1.99 | |||||||
0.057 | 0.099 | |||||||
0.025(13) | 0.035 | 0.084 | ||||||
0.056 | 0.65 | |||||||
1.99 | 2.06 | |||||||
V Decay Constants
In the discussion of the transition form factors above, we saw that differences could arise when different magnetic projections of the vector mesons are used, in combination with different components of the electromagnetic current operator. These differences are linked with violations of rotational symmetry in the model. We argue that the violation of rotational symmetry is not a major factor by checking two representative observables. The first one is the meson masses. It has been shown in Ref. Li et al. (2017) that the mass eigenvalues associated with different magnetic projections are in reasonable agreement, with a mean mass spread of 17 MeV (8 MeV) for charmonia (bottomonia). The second observable is the decay constant, which we present in this section. The decay constant is defined with the same current operator as the transition form factor. For a vector meson, its value, too, could be extracted from different magnetic projections. On the other hand, it features the simplicity of involving only one LFWF instead of convoluting two LFWFs. Therefore, the decay constant provides a pathway for examining the rotational symmetry of LFWFs.
Decay constants for vector mesons are defined as the local vacuum-to-hadron matrix elements:
(19) |
In Ref. Li et al. (2017), the “+” current is used. Since , with the “+” component of the current the decay constant can only be extracted from , using
(20) |
However, in analogy to the transition form factor, we can also use the transverse current. In the case of , we get exactly the same expression for the decay constant as with the “+” current, namely Eq. (20). This should not come as a surprise. Recall that the “+” and the transverse matrix elements with the same can be related through a transverse Lorentz boost [see Eq. (7)]. Furthermore, with the transverse current we can also calculate the decay constant from the components of the vector meson. The expression for the decay constant with follows as
(21) |
Here is the quark mass which is one of the model parameters determined in Ref. Li et al. (2017). Note that using would lead to an equivalent expression considering the symmetry between the LFWFs.
As is the transition form factor, the decay constant is also Lorentz invariant and thus it should be independent of the polarization . Therefore in practice, the difference between and provides another measure of the violation of rotational symmetry by our model. For vector meson states identified as S-wave states, both and arise primarily from the dominant spin components of LFWFs, which relate to the nonrelativistic wavefunctions. Moreover, the two expressions reduce to the same form in the nonrelativistic limit, where Li et al. (2017) and . That is,