WorksPhysics and Astronomy
Novel method for measuring the isovector partner of X(3872)
Abstract
The predicted isovector partner of the , denoted as , is expected to produce only a relatively weak structure near the charged threshold, making its experimental identification challenging. In this work, we explore the triangle singularity (TS) mechanism as a means of enhancing the sensitivity to the in the decays . We investigate the triangle rescattering mechanism, incorporating the coupled-channel scattering amplitudes obtained from the Lippmann-Schwinger equation. The interplay between the TS and the coupled-channel dynamics produces characteristic structures in the invariant mass distributions associated with the and its isovector partner . In particular, the charged-channel contribution exhibits a distinct structure sensitive to the nearby pole. Our results demonstrate that the TS mechanism provides a promising way to probe the otherwise weak manifestation of the , offering a complementary experimental strategy for searching for this predicted state at Belle II and LHCb.
Introduction
In hadron spectroscopy, some peak structures in invariant mass distributions originate from dynamical poles. Kinematic singularities, such as the two-body threshold cusp and triangle singularity (TS), can also produce observations. The former arises from a square-root branch point at an -wave threshold [1], while the latter develops when three intermediate particles in a triangle diagram can simultaneously go on shell, leading to a logarithmic singularity [2]. Under the Coleman-Norton conditions [3], a TS can approach the physical boundary and generate pronounced structures in invariant mass distributions. Unlike dynamical poles, its location is governed by kinematics. Although such structures may mimic genuine resonances, the strong kinematic sensitivity of a TS can also be exploited to probe nearby dynamical singularities [4], potentially making otherwise weak signals more accessible experimentally.
In recent years, numerous exotic hadrons have been discovered near two-hadron thresholds, suggesting that hadronic molecular components may play an important role in their structure (see the reviews [5, 6, 7, 8, 9, 10, 11, 12, 13, 14]). Among them, the , discovered in decays [15] with quantum numbers [16, 17], is one of the most prominent hadronic molecular candidates. Its mass lies remarkably close to the threshold at approximately 3871.69 MeV. This proximity, together with its strong coupling to the channels, suggests that its properties are strongly influenced by near-threshold dynamics. In particular, the mass splitting between the neutral and charged thresholds makes coupled-channel interactions and isospin-breaking effects essential for understanding the and its possible molecular partners.
An intriguing consequence of the coupled-channel dynamics is the predicted existence of an isovector partner of the , denoted as , with quantum numbers [18]. Recently, the connection between the isospin breaking of the and the existence of the has been further clarified [19, 20]. The neutral-charged threshold splitting provides the seed of isospin breaking, while its substantial amplification requires a sizable interaction in the isovector channel, which also generates a nearby companion pole. However, unlike the pronounced signal, the neutral is predicted to produce only a relatively weak cusp-like structure near the charged threshold [18]. This makes its experimental identification challenging and motivates the search for production mechanisms that can enhance the sensitivity to the predicted pole.
Triangle singularities offer a promising possibility in this regard. A method for precisely measuring the mass of the through the invariant mass distribution was proposed in Ref. [21]. The sensitivity arises from the triangle loop, where the narrow widths of the intermediate mesons allow the TS to approach the physical boundary, making lineshape highly sensitive to the mass. Related mechanisms have been investigated in and production [22, 23, 24, 25, 26], as well as in the production of the double-charm tetraquark candidate [27]. These studies demonstrate the usefulness of TS mechanisms for probing near-threshold states, particularly through their sensitivity to the masses of the produced hadrons.
In Ref. [28], a TS mechanism involving the triangle loop in decays was proposed to determine the mass of the , with the transition described by a constant coupling . While this approach exploits the kinematic sensitivity of the TS to the mass, it does not explicitly incorporate the energy dependence of the coupled-channel scattering amplitude, particularly the charged-threshold dynamics associated with the predicted pole. In the present work, we replace the constant coupling with a two-channel matrix obtained by solving the Lippmann–Schwinger equation with contact interactions.
This simplified framework generates both the and poles and allows us to investigate how their associated near-threshold dynamics modify the TS-induced invariant-mass distributions. By varying the isovector contact interaction and examining finite-width effects, we explore the potential of this mechanism to enhance the visibility of the predicted in experimental searches.
In this work, we investigate the decays and through the triangle rescattering mechanism. We replace the constant coupling used in Ref. [28] with the coupled-channel -matrix obtained from the Lippmann–Schwinger equation. The neutral-channel elastic amplitude and the charged-to-neutral transition amplitude enter the two decay processes, respectively, allowing the near-threshold dynamics associated with the and its predicted isovector partner to be incorporated into the triangle amplitudes. We find distinct structures in the invariant mass distributions, with the charged-channel contribution exhibiting a characteristic structure sensitive to the nearby pole. These results suggest that the interplay between the TS and coupled-channel dynamics provides a promising approach for experimentally searching for the at Belle II and LHCb.
This paper is organized as follows. In Sec. II, we introduce the triangle rescattering mechanism and the coupled-channel -matrix formalism. In Sec. III, we present the numerical results and discuss the predicted invariant mass distributions. A summary and discussion are given in Sec. IV.
the triangle diagram and coupled channel scattering
We investigate the decays and through the triangle rescattering mechanism, as shown in Figure 1. The first process involves the intermediate state followed by neutral-channel elastic rescattering, whereas the second proceeds through and the charged-to-neutral transition . The kinematic conditions for the triangle singularity (TS) in this mechanism have been analyzed in Ref. [28]. For reference, the corresponding TS regions are

Figure 1. Triangle rescattering mechanisms for (neutral intermediate channel) and (charged intermediate channel). In each case is the four-momentum of the system.
The two regions probe different parts of the coupled-channel amplitude. In particular, the charged-channel triangle provides access to the dynamics near the threshold, where the predicted may affect the rescattering amplitude.
To incorporate the near-threshold dynamics associated with the and its predicted isovector partner, we replace the constant coupling used in Ref. [28] with a two-channel matrix constructed from contact interactions by
solving the Lippmann–Schwinger equation. The contact couplings are chosen to reproduce the pole positions reported in Ref. [29]. This approximation retains the energy dependence and coupled-channel threshold structure relevant to the present study, while not incorporating the full dynamical ingredients of Ref. [29]. Denoting the neutral and charged channels by indices 1 and 2, respectively, the on-shell amplitude is written as
where denotes the regularized nonrelativistic two-body loop functions, and and are the isoscalar and isovector contact couplings, respectively, which are regularized with by fitting to the poles in Ref. [29]. Here is the energy relative to the neutral threshold; represents neutral-channel elastic scattering and the charged-to-neutral transition. Their energy dependence, rather than a constant coupling, is essential for retaining the charged-threshold structure in the production amplitude.
We consider the decay amplitudes for the processes and , which receive contributions from the two triangle rescattering diagrams shown in Figure 1. Although the final states have the same hadronic content, the intermediate channels differ: the first process involves elastic rescattering in the neutral channel, , whereas the second involves the transition from the charged channel, .
The relevant vertex factors are specified as follows. The production vertex is parametrized as
where denotes the polarization vector of and is the pion momentum. In the center-of-mass frame, the total three-momentum of the pair vanishes, so that . The coupling is estimated from the relevant experimental branching fractions; its numerical value is taken to be approximately in the present calculation.
The vertex is given by
where is determined from experimental data [30]. The is a narrow state, with [30], making its width potentially important for the TS-induced lineshape. The broader contribution is not included in this work.
The rescattering amplitudes are parametrized as
with the polarization vector of the incoming/outgoing , where and are the coupled-channel -matrix elements in Eq. (3).
The vertices are given above. Summing over the polarizations of the internal and mesons, and using the nonrelativistic approximation , the internal polarization structures contract to leave only the polarization vector of the external . Since the adopted vertex contains no explicit momentum dependence, the remaining loop integral is scalar. The two decay amplitudes can therefore be expressed in a unified form,
where and denote the production and couplings, respectively, for each charge channel. The amplitudes and correspond to the neutral- and charged-channel triangle mechanisms, with and describing the corresponding rescattering processes.
The scalar triangle integral is defined as
where and . The kinematic variables are , and . Thus, the two mechanisms differ in both the rescattering amplitude and the intermediate-particle masses entering the triangle loop. Following Ref. [28], we evaluate the loop integral using LOOP-TOOLS [31].
The observable distributions are obtained from
where includes the polarization sum and
with . Integrating Eq. (10) over
yields , which we use to study how the TS mechanism makes the charged-channel dynamics associated with the predicted accessible in the two decay modes.
Numerical results and discussion
We now examine the invariant-mass distributions in the two decay modes introduced in Sec. II. The purpose is not to redetermine the or pole positions, but to investigate whether the triangle mechanism provides an experimental useful probe of the charged-channel dynamics associated with the predicted . We use the contact-interaction approximation to the coupled-channel amplitude, with and the low-energy constants adopted from Ref. [29]. Accordingly, the lineshapes presented below are predictions conditional on this particular scattering input, rather than an independent determination of the existence of the .
For each process, the invariant mass is integrated over its kinematically allowed range to obtain . The neutral intermediate state contributes to through , whereas the charged intermediate state contributes to through . In both cases, the production amplitude contains the product of a triangle loop and the energy-dependent rescattering amplitude. In particular, , so the charged-channel threshold dependence is retained rather than replaced by a constant transition vertex.
We firstly investigate the finite-width effects of the intermediate and mesons in the triangle diagrams. These effects are relevant for the invariant-mass distribution, whose lineshape can be sensitive to the mass and width of the [32]. We adopt two prescriptions to account for the width, while describing the mass distribution using a Breit–Wigner form in both cases. In Case I, the mass is varied between and , where and are the PDG values. In Case II, the is assigned a complex mass, , to investigate how its finite width smears the structures generated by the triangle mechanism. The complex-mass prescription is used to study finite-width effects in hadron spectroscopy [28, 33, 34].
For Case I, the results obtained with the two limiting values of are shown in Figure 2, where the bands correspond to the uncertainties of and . The neutral-channel contribution exhibits two prominent structures. The lower-mass peak is associated with the near-threshold pole, while the higher-mass peak is associated with the TS. In the charged-channel contribution, the cusp near the charged threshold reflects the virtual-state dynamics associated with the pole, whereas the higher-energy enhancement is primarily driven by the TS. The detailed lineshapes result from the interplay between the triangle mechanism and the corresponding coupled-channel amplitudes. Varying shifts the structures, illustrating the sensitivity of the predicted distributions to the intermediate-particle mass. In particular, the charged-channel lineshape may provide a useful probe of the predicted dynamics.

Figure 2. Predicted distributions from the triangle mechanisms, including the finite widths of the intermediate particles. The blue and red curves correspond to with neutral-channel rescattering () and with charged-to-neutral rescattering (), respectively. The relative heights of the two distributions depend on the channel-dependent production amplitudes and should not be interpreted as a model-independent event-yield ratio.
For Case II, the complex-mass prescription smears the sharp structures found in Case I. As shown in Figure 3, the nearby structures associated with the near-threshold dynamics and the TS merge into a shoulder-like enhancement because the width is comparable to their separation. Nevertheless, the enhancement in the charged-to-neutral rescattering contribution remains visible. Thus, the finite width modifies the detailed lineshape and reduces the sharpness of the structures, but does not eliminate the TS-related enhancement in the charged-channel distribution. The blue and red curves correspond to different decay channels and exhibit structures in different kinematic regions. Their relative heights depend on the channel-dependent production amplitudes and should not be interpreted as a model-independent event-yield ratio.

Figure 3. Predicted distributions from the triangle mechanisms, including the finite widths of the intermediate particles. The blue and red curves correspond to with neutral-channel rescattering () and with charged-to-neutral rescattering (), respectively. The relative heights of the two distributions depend on the channel-dependent production amplitudes and should not be interpreted as a model-independent event-yield ratio.
This result motivates a further investigation of the sensitivity of the TS-induced lineshapes to the isovector interaction. In the following, we vary the isovector contact coupling while keeping fixed, and examine how the resulting motion of the pole affects the observable invariant-mass distributions.
It is instructive to compare the present analysis with existing determinations of the isovector interaction. The dispersive analysis of Ref. [35] constrains the isospin-breaking dynamics of the through its decay, extracting the coupling ratio and updating the predictions for the poles. In particular, it reports a neutral pole near 3881.7 MeV, approximately 1.8 MeV above the charged threshold on an unphysical sheet. A more comprehensive coupled-channel analysis was performed in Ref. [29], which simultaneously fits BESIII and LHCb data and incorporates the energy-dependent self-energy, the associated three-body dynamics, one-pion exchange, and inelastic channels. These studies provide important constraints on the isovector interaction, but their dynamical descriptions are more elaborate than the contact-interaction model adopted here.
In the present work, we use an energy-independent contact interaction parametrized by and to reproduce the poles obtained in Ref. [29]. This simplified description does not explicitly include one-pion exchange, the full three-body dynamics, or inelastic channels, and therefore cannot capture all the dynamical effects incorporated in the more comprehensive analysis. To explore how such model simplifications may affect the predicted pole and its observable manifestations, we rescale the isovector coupling as while keeping fixed. We then track the resulting pole trajectory and examine its impact on the TS-induced invariant-mass distributions.
To investigate the pole trajectory, the rescaled isovector contact coupling is adopted. The resulting pole positions are shown in Figure 4. As increases, the isovector interaction becomes stronger and the pole approaches the charged threshold for . For larger values of , the pole moves onto the sheet, which is physical with respect to the charged channel. The pole thus acquires a bound-state-like character relative to the charged threshold. We restrict the scan to ; beyond this range, the rescaled isovector interaction becomes stronger than the isoscalar interaction in the parametrization adopted here.

Figure 4. Real and imaginary parts of the pole as functions of the scaling parameter , defined by with fixed. The stands for the first/second Riemann sheet of the corresponding channel.
This pole trajectory provides an estimate of the interaction dependence within the present contact-interaction model. The real part of the pole relative to the threshold determines its location with respect to the threshold, while its imaginary part and residue influence how the pole manifests itself in the physical scattering amplitude. Consequently, the pole position alone does not uniquely determine the observable lineshape, and its impact must be examined by calculating the corresponding production amplitudes.
We now examine how the variation of affects the observable invariant-mass distributions. Figure 5 shows the spectra obtained for representative values of , using the same

Figure 5. Predicted distributions for representative values of , with and fixed. The same finite-width treatment is used for all curves. The variation illustrates how changes in the isovector interaction and the associated pole affect the TS-induced invariant-mass distributions.
finite-width treatment. A nontrivial enhancement appears in the 4402–4406 MeV region. For , the virtual pole remains relatively far from the charged threshold, and the cusp in is comparatively mild, so the resulting lineshapes largely overlap. As increases, the pole approaches the charged threshold, leading to a more pronounced threshold enhancement around 4405 MeV. For , the pole develops a bound-state-like character with respect to the charged channel, and the enhancement becomes more prominent. At , the dominant enhancement shifts toward 4403 MeV. This behavior reflects the interplay between the evolving near-threshold structure of and the TS kinematics, indicating that the proposed production mechanism can be sensitive to the qualitative nature of the isovector interaction. The magnitude of this sensitivity, however, depends on the adopted interaction model.
The predicted lineshapes should be regarded as illustrative rather than direct predictions of experimental event yields. The relative yields of the two decay channels depend on their respective weak-production amplitudes and cannot be inferred from the normalized spectra alone. A realistic assessment of the experimental sensitivity to the therefore requires accounting for non-TS backgrounds, interference effects, detector resolution, and event-selection efficiencies.
Beyond the specific case of the , the present strategy may be extended to other near-threshold hadronic states whose signals are sensitive to coupled-channel dynamics. By combining suitable triangle-singularity kinematics with energy-dependent scattering amplitudes, one can investigate how nearby poles and threshold effects manifest themselves in experimentally accessible invariant-mass distributions. This approach may be particularly useful for hadronic molecular candidates, for which the interplay between kinematic singularities and near-threshold poles can produce distinctive, channel-dependent lineshapes.
The key requirement is to identify production processes in which the triangle kinematics enhance the sensitivity to the relevant coupled-channel dynamics. Such applications would provide a complementary means of probing near-threshold poles and exploring the nature of exotic hadrons, e.g. [32, 36, 37, 38, 39] and [39, 40, 41, 42].
Summary
In this work, we have investigated the decays and via the triangle rescattering mechanism. By replacing the constant coupling used in Ref. [28] with an energy-dependent two-channel matrix constructed from contact interactions, we investigate how the near-threshold dynamics associated with the and its predicted isovector partner affect the TS-induced invariant-mass distributions. We find that the TS-related enhancements remain visible under the finite-width prescriptions considered in this work. Furthermore, varying the isovector contact interaction reveals that the lineshapes are relatively stable when the pole remains in the virtual-state regime, but change more substantially when the pole develops a bound-state-like character with respect to the charged channel. These results demonstrate how the interplay between triangle kinematics and coupled-channel dynamics can influence experimentally accessible invariant-mass distributions, and suggest that the proposed mechanism may provide a complementary approach to searching for the . A quantitative assessment of its experimental sensitivity will require a more complete treatment of the coupled-channel dynamics, non-TS backgrounds, interference effects, and detector resolution. The present study therefore provides a proof of principle for this search strategy, rather than an independent determination of the pole parameters or evidence for its experimental observation.
Acknowledgments
This research is supported by the National Natural Science Foundation of China under Grants No. 12305096, No. 12235018 and No. 12675106, the Fundamental Research Funds for the Central Universities under Grant No. SWU-XDJH202304, No. SWU-KQ25016 and Chongqing Natural Science Foundation under Project No. CSTB2025NSCQ-GPX0516.
References
- [1]E. P. Wigner, Phys. Rev. 73, 1002 (1948).
- [2]L. D. Landau, Zh. Eksp. Teor. Fiz. 37, 62 (1960).
- [3]S. Coleman and R. E. Norton, Nuovo Cim. 38, 438 (1965).
- [4]F.-K. Guo, X.-H. Liu, and S. Sakai, Prog. Part. Nucl. Phys. 112, 103757 (2020), 1912.07030.
- [5]H.-X. Chen, W. Chen, X. Liu, and S.-L. Zhu, Phys. Rept. 639, 1 (2016), 1601.02092.
- [6]A. Esposito, A. Pilloni, and A. D. Polosa, Phys. Rept. 668, 1 (2017), 1611.07920.
- [7]R. F. Lebed, R. E. Mitchell, and E. S. Swanson, Prog. Part. Nucl. Phys. 93, 143 (2017), 1610.04528.
- [8]S. L. Olsen, T. Skwarnicki, and D. Zieminska, Rev. Mod. Phys. 90, 015003 (2018), 1708.04012.
- [9]A. Ali, J. S. Lange, and S. Stone, Prog. Part. Nucl. Phys. 97, 123 (2017), 1706.00610.
- [10]F.-K. Guo, C. Hanhart, U.-G. Meißner, Q. Wang, Q. Zhao, and B.-S. Zou, Rev. Mod. Phys. 90, 015004 (2018), [Erratum: Rev.Mod.Phys. 94, 029901 (2022)], 1705.00141.
- [11]Y. S. Kalashnikova and A. V. Nefediev, Phys. Usp. 62, 568 (2019), 1811.01324.
- [12]Y.-R. Liu, H.-X. Chen, W. Chen, X. Liu, and S.-L. Zhu, Prog. Part. Nucl. Phys. 107, 237 (2019), 1903.11976.
- [13]N. Brambilla, S. Eidelman, C. Hanhart, A. Nefediev, C.-P. Shen, C. E. Thomas, A. Vairo, and C.-Z. Yuan, Phys. Rept. 873, 1 (2020), 1907.07583.
- [14]Z.-Y. Bai, D.-Y. Chen, Q. Huang, X. Liu, S.-Q. Luo, and J.-Z. Wang, Phys. Rept. 1204, 1 (2027), 2602.19887.
- [15]S. K. Choi et al. (Belle), Phys. Rev. Lett. 91, 262001 (2003), hep-ex/0309032.
- [16]R. Aaij et al. (LHCb), Phys. Rev. Lett. 110, 222001 (2013), 1302.6269.
- [17]R. Aaij et al. (LHCb), Phys. Rev. D 92, 011102 (2015), 1504.06339.
- [18]Z.-H. Zhang, T. Ji, X.-K. Dong, F.-K. Guo, C. Hanhart, U.-G. Meißner, and A. Rusetsky, JHEP 08, 130 (2024), 2404.11215.
- [19]X.-K. Dong, F.-K. Guo, C. Hanhart, T. Ji, and U.-G. Meißner (2026), 2608.16266.
- [20]P.-P. Shi, M. Albaladejo, F.-K. Guo, and J. Nieves (2026), 2608.01237.
- [21]F.-K. Guo, Phys. Rev. Lett. 122, 202002 (2019), 1902.11221.
- [22]S. Sakai, H.-J. Jing, and F.-K. Guo, Phys. Rev. D 102, 114041 (2020), 2008.10829.
- [23]E. Braaten, L.-P. He, and K. Ingles, Phys. Rev. D 100, 031501 (2019), 1904.12915.
- [24]E. Braaten, L.-P. He, and K. Ingles, Phys. Rev. D 101, 014021 (2020), 1909.03901.
- [25]E. Braaten, L.-P. He, and K. Ingles, Phys. Rev. D 100, 074028 (2019), 1902.03259.
- [26]S. Sakai, E. Oset, and F.-K. Guo, Phys. Rev. D 101, 054030 (2020), 2002.03160.
- [27]E. Braaten, L.-P. He, K. Ingles, and J. Jiang, Phys. Rev. D 106, 034033 (2022), 2202.03900.
- [28]M.-J. Yan, Y.-H. Ge, and X.-H. Liu, Phys. Rev. D 106, 114002 (2022), 2208.03943.
- [29]T. Ji, X.-K. Dong, F.-K. Guo, C. Hanhart, and U.-G. Meißner, Sci. Bull. 71, 4416 (2026), 2502.04458.
- [30]F. Takahashi et al. (Particle Data Group), Int. J. Mod. Phys. A 41, 2630011 (2026).
- [31]T. Hahn, Nucl. Phys. B Proc. Suppl. 89, 231 (2000), hep-ph/0005029.
- [32]W.-J. Wang, M.-L. Du, F.-K. Guo, and B. Wu (2026), 2609.24231.
- [33]H.-J. Jing, S. Sakai, F.-K. Guo, and B.-S. Zou, Phys. Rev. D 100, 114010 (2019), 1907.12719.
- [34]M.-J. Yan, C.-S. An, and C.-R. Deng (2025), 2512.02655.
- [35]J. M. Dias, T. Ji, X.-K. Dong, F.-K. Guo, C. Hanhart, U.-G. Meißner, Y. Zhang, and Z.-H. Zhang, Phys. Rev. D 111, 014031 (2025), 2409.13245.
- [36]F.-K. Guo, H.-J. Jing, U.-G. Meißner, and S. Sakai, Phys. Rev. D 99, 091501 (2019), 1903.11503.
- [37]S. X. Nakamura, Phys. Rev. D 103, 111503 (2021), 2103.06817.
- [38]F.-Z. Peng, M.-J. Yan, M. S. Sánchez, and M. Payan Valderrama, Phys. Lett. B 846, 138207 (2023), 2211.09154.
- [39]Z.-Y. Yang, F.-Z. Peng, M.-J. Yan, M. Sánchez Sánchez, and M. Pavon Valderrama, Phys. Rev. D 111, 014012 (2025), 2211.08211.
- [40]T. J. Burns and E. S. Swanson, Phys. Lett. B 838, 137715 (2023), 2208.05106.
- [41]S. X. Nakamura and J. J. Wu, Phys. Rev. D 108, L011501 (2023), 2208.11995.
- [42]M.-J. Yan, F.-Z. Peng, M. Sánchez Sánchez, and M. Pavon Valderrama, Phys. Rev. D 107, 074025 (2023), 2207.11144.