In this paper
AbstractHidden-color dynamicsReferencesWorksPhysics and Astronomy
Origin of Tensor Polarized Structure of the Deuteron
Abstract
The microscopic origin of the deuteron tensor polarized structure function remains unresolved. Conventional nuclear models attribute a nonzero mainly to the -wave component of the deuteron wave function, but generally fail to reproduce the magnitude and momentum-dependence observed by the HERMES experiment. Using a QCD-based light-front framework with degrees of freedom beyond the conventional nucleonic picture, in particular including hidden-color components and relativistic spin–momentum correlations, we find that a small but nonzero estimated hidden-color admixture can be sufficient to generate a substantial tensor response (), while it is essentially insensitive to the -wave component. The physical -wave probability of , fixed by the empirical quadrupole moment, changes by only , with the dominant contribution arising from the colored-cluster -wave sector. These results demonstrate that the tensor-polarized structure of the deuteron is driven primarily by hidden-color dynamics rather than by the -wave component.
Hidden-color dynamics
At hadronic resolution, the deuteron is conventionally described as a weakly bound system of two color-singlet nucleons. At the quark level, however, its six quarks can couple to an overall color singlet through both singlet–singlet and hidden-color channels. We write this structure schematically as
where represents the conventional color-singlet nucleon–nucleon channel. The state collectively represents configurations in which two colored subclusters are coupled to an overall color singlet. These configurations are referred to as “hidden color” because the individual colored clusters cannot appear as asymptotic physical states but entangled in their color degrees of freedom with the deuteron. They are expected to become increasingly relevant when the deuteron is resolved at short distances or high resolutions [10, 11].
Orbital and color dynamics represent distinct aspects of the deuteron wave function. The -wave describes its nonspherical orbital structure and is essential for the quadrupole moment, whereas hidden color describes non-nucleonic couplings among its six quarks. Including both effects allows us to test whether is controlled by orbital deformation or by the internal color dynamics.
The interplay between hidden-color and orbital dynamics can be studied using different QCD-motivated descriptions. Here, we employ a LF realization in which the hidden-color sector is represented by two effective colored clusters coupled to an overall color-singlet state [21]. This construction provides one concrete way to incorporate non-nucleonic color degrees of freedom while retaining a relativistic description of the internal spin and orbital dynamics.
The corresponding light-front wave function (LFWF) depends on the longitudinal momentum fraction , relative transverse momentum , deuteron LF spin projection , and cluster helicities and , and can be decomposed as
where is the Clebsch–Gordan which couples the orbital angular momentum and total cluster spin to the deuteron spin . We retain the and channels, corresponding to and , respectively. The function describes the relative cluster motion, while is obtained by applying the Melosh rotations to the canonical cluster-spin state [22, 23].
The complete deuteron LFWF is expressed as
where and denote the and components with the probabilities and , respectively.
In the mixed longitudinal-momentum and transverse-coordinate representation, the wave function factorizes as [24]
where , is the transverse separation between the clusters, , and is the azimuthal angle of . The transverse radial mode is obtained by solving the LF holographic Schrödinger equation [24], while the longitudinal mode is determined from the ’t Hooft equation [25]. The factor carries the azimuthal dependence associated with the orbital-angular-momentum projection .
The momentum-space wave function follows from the transverse Fourier transform
where is conjugate to . The LFWF is obtained by solving the LF Schrödinger-like equations for the transverse and longitudinal dynamics, as detailed in Ref. [21]. The deuteron mass eigenvalue and the corresponding LFWF are determined by three parameters: the effective cluster mass , the longitudinal confinement strength , and the transverse confinement scale . The parameter enters the ’t Hooft equation for the longitudinal mode, whereas determines the LF holographic potential for the transverse mode. Requiring the physical deuteron mass correlates these three parameters and fixes the corresponding LFWF.
The resolution of the effective two-cluster configuration is characterized by the probing energy , where is the deuteron mass. Small corresponds to the weakly bound nucleonic regime, whereas increasing probes shorter-distance dynamics for which colored-cluster configurations can contribute. Although does not directly quantify the hidden-color content of the deuteron wave function, it serves as an effective probe of these short-distance color dynamics.
Figure 1 displays the correlated combinations of , , and that satisfy the deuteron mass eigenvalue condition. Along the solutions shown, increasing reduces the confinement strengths required to reproduce , whereas decreasing requires larger values of both the effective confinement scales. This trend is consistent with the expected color structure of the deuteron at different resolutions. In the formal chiral limit , the system becomes more compact and strongly interacting; color recoupling then implies an octet–octet probability of about 80% [10, 11]. Conversely, as , the required interactions weaken and the system approaches the conventional weakly bound singlet–singlet nucleonic configuration, for which .

Figure 1. Combinations of the effective cluster mass and the longitudinal and transverse confinement strengths and , respectively that reproduce the deuteron mass. The highlighted point denotes the parameter set used in this work, GeV, corresponding to MeV.
For the consistency of our analysis approximating the cluster quark PDFs as the unpolarized nucleon PDFs from the NNPDF fits, we use GeV, . GeV, and GeV. This set corresponds to the highlighted point and to MeV [21]. The selected point, corresponding to MeV, stays close to the nucleonic limit and is well separated from the formal chiral limit (as shown in Figure 1). Although the mass eigenvalue condition does not directly determine the color amplitudes and , the hidden-color content may be estimated indirectly by interpolating between the two limiting regimes. Assuming a linear interpolation in between the compact six-quark limit, at , and the nucleonic limit, at , we obtain . This provides a qualitative estimate of the hidden-color admixture rather than a probability obtained through an explicit color projection. Within this qualitative interpolation, we estimate that the deuteron contains a hidden-color admixture of less than 10%.
To fix the orbital admixture, we require the full wave function to reproduce the empirical deuteron quadrupole moment, obtaining . This relatively small -wave probability should be contrasted with the () -wave admixture [26, 27, 28] commonly found in conventional descriptions of the deuteron as a weakly bound nucleon–nucleon () system. In the present approach, however, the deuteron is described in terms of strongly interacting color clusters rather than as a shallow bound state. In particular, the inclusion of hidden-color components, even though their overall probability is less than 10%, modifies the internal dynamics and leads to a substantially smaller -wave admixture. Thus, the small value of is a consequence of the underlying cluster structure of the deuteron with the entangled color degrees of freedom rather than an additional phenomenological assumption.
Despite its small probability, the -wave remains important for observables that are particularly sensitive to the orbital structure. In particular, the quadrupole moment receives a strong contribution from the interference between the - and -wave components. Consequently, the quadrupole moment is highly sensitive to the wave: the -wave result, , is brought to the empirical value [29] by the -wave admixture. We therefore use this independently determined value without further adjustment when calculating .
In contrast, the small admixture has only a minor effect on the charge and magnetic observables. The charge radius changes from fm to fm, to be compared with the experimental value fm [30]. The quoted theoretical uncertainties are obtained by varying the parameters , , and by 10% around their central values. The magnetic radius remains fm, consistent with the extracted value fm [31], while the magnetic moment changes only from to , close to the empirical value [29]. The unpolarized and helicity-dependent distributions are likewise essentially unchanged [21].
*Tensor polarized structure function.—*Using the well-established convolution framework, implemented for effective clusters in Ref. [21], we evaluate the quark contribution to the tensor structure function as
where labels the two effective clusters, is their longitudinal momentum fraction, and is the quark charge in units of the positron charge. The tensor-polarized longitudinal momentum distribution is , where denotes the density of a cluster with helicity in a deuteron with LF helicity , obtained from the LFWF by integrating over transverse momentum and summing over the spectator helicity. The cluster quark PDFs are approximated by unpolarized nucleon PDFs from the NNPDF fits at , as in Ref. [21].
As shown in Figure 2, the - and -wave results remain close throughout the range considered. Their maximum absolute difference, normalized to the maximum magnitude of the -wave result, is only 1.7%. The description of the HERMES data [1] is therefore essentially unchanged by the inclusion of the wave.
Figure 2. Deuteron tensor structure function at . The orange-shaded band shows the full result with , while the purple-shaded band denotes the pure -wave result. The bands reflect the 10% variation of , , and around their central values. The dashed and dotted curves show the conventional nucleonic calculations of Ref. [4]: basic convolution and the virtual-nucleon approximation, respectively. Open circles with error bars represent the HERMES data [1].
Figure Figure 2 also compares our result with the two conventional nucleonic calculations of Ref. [4], obtained using the basic convolution approach and the virtual-nucleon approximation. In the ordinary nucleonic picture, a pure -wave deuteron gives ; a nonzero tensor response arises from – interference and the pure -wave contribution. Their predictions nevertheless differ from the HERMES data in magnitude and dependence. By contrast, the colored-cluster -wave sector already produces a nonzero result consistent with the data. The wave then provides only a small correction, despite being required by the quadrupole moment. This comparison identifies hidden-color dynamics as the dominant source of the tensor response.
**Conclusion and outlook.—**Our results identify a source of the deuteron tensor structure function beyond the conventional mechanism. Within the present relativistic light-front description, hidden-color degrees of freedom produce a nonzero already in the -wave sector, consistent with the HERMES data. These results indicate that a small but nonzero estimated hidden-color admixture, below 10%, can be sufficient to generate a substantial tensor response. Adding a -wave component with probability , fixed independently by the empirical quadrupole moment, changes only slightly: the maximum absolute difference is 1.7% of the maximum magnitude of the -wave result. The same -wave admixture brings the quadrupole moment into agreement with its empirical value, while leaving the charge and magnetic observables and the unpolarized and helicity-dependent distributions essentially unchanged.
The wave needed to reproduce the quadrupole moment therefore does not drive the tensor partonic response. Instead, the persistence of in the colored-cluster -wave sector points to non-nucleonic color dynamics as an important ingredient in its microscopic origin. Precision measurements of can test this interpretation and help distinguish the roles of orbital motion and color dynamics in the deuteron.
We are aware of ongoing efforts to measure the transverse-momentum structure of the deuteron through tensor-polarized SIDIS measurements at CLAS12 [20]. The available projections for these observables remain approximate. In our forthcoming study, we will explore how the hidden-color dynamics essential for the inclusive tensor response manifest themselves in the transverse structure of the deuteron. Our preliminary results indicate that hidden-color dynamics are likewise essential for describing the transverse tensor-polarized distributions.
**Acknowledgements.—**We thank Siqi Xu, Gerald A. Miller and James P. Vary for insightful discussions. XZ is supported by National Natural Science Foundation of China, Grant No. 12375143, by Key Research Program of Frontier Sciences, Chinese Academy of Sciences, Grant No. ZDBS-LY-7020, by international partnership program of the Chinese Academy of Sciences, Grant No. 016GJHZ2022103FN, by the Strategic Priority Research Program of the Chinese Academy of Sciences, Grant No. XDB34000000. CRJ is supported in part by the U.S. Department of Energy (Grant No. DE-FG02-03ER41260) and acknowledges the National Energy Research Scientific Computing Center (NERSC) supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.
References
References
- [1]H. A. Airapetian et al. (HERMES), Phys. Rev. Lett. 95, 242001 (2005), arXiv:hep-ex/0506018.DOI
- [4]W. Cosyn, Y.-B. Dong, S. Kumano, and M. Sargsian, Phys. Rev. D 95, 074036 (2017), arXiv:1702.05337 [hep-ph].DOI
- [10]C. R. Ji, J. Phys. Conf. Ser. 543, 012004 (2014).
- [11]B. L. G. Bakker and C.-R. Ji, Prog. Part. Nucl. Phys. 74, 1 (2014).
- [20]J. Poudel, A. Bacchetta, J.-P. Chen, D. Keller, I. Fernando, E. Long, D. Ruth, N. Santiesteban, and K. Slifer, (2025), arXiv:2502.20044 [hep-ph].arxiv.org/abs/2502.20044
- [21]S. Kaur, C. Mondal, X. Zhao, and C.-R. Ji, Phys. Rev. D 113, 054008 (2026), arXiv:2507.09886 [hep-ph].arxiv.org/abs/2507.09886
- [22]H. J. Melosh, Phys. Rev. D 9, 1095 (1974).
- [23]C. R. Ji, P. L. Chung, and S. R. Cotanch, Phys. Rev. D 45, 4214 (1992).
- [24]S. J. Brodsky, G. F. de Teramond, H. G. Dosch, and J. Erlich, Phys. Rept. 584, 1 (2015), arXiv:1407.8131 [hep-ph].DOI
- [25]G. ’t Hooft, Nucl. Phys. B 75, 461 (1974).
- [26]R. Machleidt, Phys. Rev. C 63, 024001 (2001), arXiv:nucl-th/0006014.DOI
- [27]R. B. Wiringa, V. G. J. Stoks, and R. Schiavilla, Phys. Rev. C 51, 38 (1995), arXiv:nucl-th/9408016.
- [28]V. G. J. Stoks, R. A. M. Klomp, C. P. F. Terheggen, and J. J. de Swart, Phys. Rev. C 49, 2950 (1994), arXiv:nucl-th/9406039.DOI
- [29]M. Garcon and J. W. Van Orden, Adv. Nucl. Phys. 26, 293 (2001), arXiv:nucl-th/0102049.DOI
- [30]I. Sick and D. Trautmann, Nucl. Phys. A 637, 559 (1998).
- [31]A. Afanasev, V. D. Afanasev, and S. V. Trubnikov, (1998), arXiv:nucl-th/9808047.
