betaFeedback

WorksPhysics and Astronomy

ISR Reconstruction of Radiative C=+1C=+1 Charmonia as the Origin of the Belle–BESIII Line-Shape Puzzle

Abstract

A longstanding disagreement persists between Belle and BESIII over the line shapes of e+e−→Λc+Λˉc−e^+e^-\to Λ_c^+\barΛ_c^-, DsDs1(2536)D_sD_{s1}(2536), and DsDs2∗(2573)D_sD_{s2}^{*}(2573) near 4.63 GeV. Belle observes enhancements through initial state radiation that are absent or much weaker at BESIII. We propose that the extra Belle yield arises from radiative production of a C=+1C=+1 charmonium. Luminosity normalization increases this contribution to the extracted cross sections, so tens of fb can explain the enhancements. Belle II can test this mechanism through full reconstruction and searches for missing charmonia in open charm modes.

Introduction

The observation of charmonium-like XYZXYZ states has opened a window onto hadron spectroscopy and deepened our understanding of the nonperturbative behavior of the strong interaction. Among these states, the YY states—charmonium-like states produced in electron–positron annihilation—constitute the so-called YY problem [1, 2], which has motivated a variety of theoretical efforts [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]; see Refs. [14, 15, 16] for recent reviews.

A long-standing puzzle in this context concerns the Λc+Λˉc−\Lambda_c^+\bar{\Lambda}_c^- invariant mass spectrum. In 2008, the Belle Collaboration reported a higher YY state, Y(4630)Y(4630), in e+e−→Λc+Λˉc−e^+e^- \to\Lambda_c^+\bar{\Lambda}_c^- [17]. With the accumulation of experimental data, the BESIII Collaboration searched for Y(4630)Y(4630) through the same process [18, 19] but found no enhancement structure in this energy range. This discrepancy constitutes a persistent Belle–BESIII difference in the Λc+Λˉc−\Lambda_c^+\bar{\Lambda}_c^- line shape. Despite serious cross-checks by both collaborations, no explanation has been found.

Remarkably, this phenomenon is not unique. With the BESIII and Belle measurements of e+e−→DsDs1(2536)e^+e^- \to D_sD_{s1}(2536) and e+e−→DsDs2∗(2573)e^+e^- \to D_sD_{s2}^*(2573) [20, 21, 22], the same Belle–BESIII difference emerges in the line shapes of the corresponding invariant mass spectra. This suggests an underlying dynamical mechanism common to these channels, the identification of which is a timely task in this area.

Motivated by the strong coupling of higher ηc\eta_c states to Λc+Λˉc−\Lambda_c^+\bar{\Lambda}_c^- [23], we conjecture that the Λc+Λˉc−\Lambda_c^+\bar{\Lambda}_c^- spectrum measured by Belle contains a higher ηc\eta_c signal. The key point is the reconstruction method used in initial state radiation (ISR) analyses. Full reconstruction requires reconstructing both hadrons from their decay products, as in Belle’s DDˉD\bar{D} analysis [24]. Their measured momenta allow the photon momentum to be inferred and its angular distribution to be determined without photon detection.

ton detection. In the Λc+Λˉc−\Lambda_c^+\bar{\Lambda}_c^- channel, Belle uses partial reconstruction: one Λc\Lambda_c and an energetic photon are reconstructed, and the other baryon is identified through recoil constraints [17]. Radiative events can satisfy the photon energy and recoil mass requirements and enter Belle’s γΛc+Λˉc−\gamma\Lambda_c^+\bar{\Lambda}_c^- sample.

In this Letter, we extend this line of reasoning to the DsDs1(2536)D_sD_{s1}(2536) and DsDs2∗(2573)D_sD_{s2}^*(2573) channels. We show that the inclusion of higher charmonium contributions naturally accounts for the BESIII–Belle difference in the line shapes of these invariant mass spectra: higher charmonia lying above the DsDs1(2536)D_sD_{s1}(2536) and DsDs2∗(2573)D_sD_{s2}^*(2573) thresholds can couple to these channels and produce additional resonant contributions that may appear in the Belle measurements but be absent in those of BESIII. Normalization by the small effective ISR luminosity increases their contribution to Belle’s extracted cross sections, allowing radiative products of tens of fb to generate visible enhancements and account for the observed line shape differences. The Supplemental Material (SM) [25] gives the fitting methods and results of the spectroscopy calculations.

Production mechanism.—Figure 1 compares the two topologies yielding a photon γ\gamma and two hadrons h1,h2h_1,h_2. Here s0\sqrt{s_0} is the collision energy and m=sm=\sqrt{s} the hadron pair invariant mass. Panel (a) shows ISR at Belle; after photon emission, the virtual photon produces the hadron pair through the same transition measured by BESIII:

ISR production and radiative production diagrams

Figure 1. (a) ISR production of the hadron pair h1h2h_1h_2 at Belle. The circle is the transition γ∗(s)→h1h2\gamma^*(s)\to h_1h_2 also measured by BESIII. (b) Radiative production of RR at Belle’s energy, followed by R→h1h2R\to h_1h_2. The square is γ∗(s0)→γradR\gamma^*(s_0)\to\gamma_{\rm rad}R; the filled dot is the decay. Here ss is the squared invariant mass of h1h2h_1h_2 and s0s_0 the squared collision energy.

e+e−→γISRγ∗(s)→γISRh1h2.(1)e^+e^- \to\gamma_{\mathrm{ISR}} \gamma^*(s) \to\gamma_{\mathrm{ISR}} h_1h_2. \tag*{(1)}

Panel (b) is radiative production of a charmonium resonance RR at the full Belle energy,

e+e−→γ∗(s0)→γradR,R→h1h2.(2)e^+e^- \to\gamma^*(s_0) \to\gamma_{\mathrm{rad}}R,\qquad R \to h_1h_2. \tag*{(2)}

Both photons in panel (b) have C=−1C=-1, so charge conjugation fixes C(R)=+1C(R)=+1. Radiative production allows ηc\eta_c states with JPC=0−+J^{PC}=0^{-+} and χcJ\chi_{cJ} states with JPC=J++J^{PC}=J^{++}, J=0,1,2J=0,1,2; these labels include all states with the corresponding quantum numbers. At s0=10.58\sqrt{s_0}=10.58 GeV, a resonance near 4.63 GeV recoils against a photon whose energy in the collision center of mass is

Eγ∗=s0−s2s0≃4.28 GeV.(3)E_\gamma^*=\frac{s_0-s}{2\sqrt{s_0}}\simeq4.28\ \mathrm{GeV}. \tag*{(3)}

In a BESIII direct scan at the same hadron pair mass, the collision energy is used to produce the pair, leaving no energy for a hard photon.

The ISR and radiative processes have the same γh1h2\gamma h_1h_2 final state and can satisfy the same recoil requirements. In a fully reconstructed analysis, the squared mass recoiling against the two hadrons, Mrec2(h1h2)M_{\mathrm{rec}}^2(h_1h_2), is required to be close to zero, as expected for one missing photon. The three Belle analyses considered here use partial reconstruction: one hadron and an energetic photon are reconstructed, and the other hadron is identified through recoil constraints. Requiring full reconstruction of only one charm hadron improves efficiency, especially when the accessible decay branching fractions are small [17, 29]. For Λc+Λˉc−\Lambda_c^+\bar{\Lambda}_c^-, Belle reconstructs one Λc+\Lambda_c^+, requires the photon energy to exceed 3.5 GeV, and identifies the recoiling Λˉc−\bar{\Lambda}_c^- by Mrec(Λc+γ)M_{\mathrm{rec}}(\Lambda_c^+\gamma) together with an antiproton from its decay [17]. The two DsDsJD_sD_{sJ} analyses impose the same kind of recoil constraint and require a photon above 3 GeV in the collision frame [20, 21]. Equation (3) shows that an ISR photon and a radiative photon have the same energy at the same pair mass, so the cut on the photon energy and the recoil mass do not separate them.

In addition, full reconstruction allows the photon momentum and direction to be inferred from the two hadron momenta through energy and momentum conservation, without requiring photon detection. Belle’s DDˉD\bar{D} analysis used this method and observed the characteristic ISR peaks along the beams [24]. In the partially reconstructed samples considered here, the photon polar angle θγ∗\theta_\gamma^* relative to the electron beam in the collision center of mass is measured directly from the detected photon.

Its angular coverage is limited by the calorimeter acceptance [30], excluding the directions closest to the beams where ISR is most strongly peaked. The predicted photon distribution for radiative production has no sharp peaks along the beams [31]. Within the central window ∣cos⁡θγ∗∣<0.9|\cos\theta_\gamma^*|<0.9 used in several Belle ISR analyses [24, 32, 33, 34], the ISR [35] and radiative angular shapes are similar [25]. The restricted acceptance and present statistics therefore limit their separation. We add their yields because interference cancels when final states and their charge conjugates are combined with equal weight [35].

Belle normalizes the event yield in each hadron mass bin using the effective ISR luminosity,

LISR,j=LB∫mj−mj+2ms0W(s0,1−m2s0) dm,(4)\mathcal{L}_{\mathrm{ISR},j}=\mathcal{L}_B\int_{m_j^-}^{m_j^+}\frac{2m}{s_0}W\left(s_0,1-\frac{m^2}{s_0}\right)\,dm, \tag*{(4)}

where LB\mathcal{L}_B is the collected integrated luminosity, mj±m_j^\pm are the edges of mass bin jj, and WW is the ISR radiator, the probability density in the photon energy fraction 1−m2/s01-m^2/s_0 [20, 36]. In ISR, an incoming lepton emits a photon before annihilation, lowering the effective annihilation energy to the invariant mass mm of the hadron pair. Selecting a narrow hadron mass bin restricts this photon to a narrow energy interval. The small emission probability in that interval is the integral in Eq. (4), so LISR,j≪LB\mathcal{L}_{\mathrm{ISR},j}\ll\mathcal{L}_B.

Radiative production in Eq. (2) proceeds through annihilation at the full collision energy, with the photon produced together with RR. No preceding ISR emission is needed to lower the annihilation energy. For channel ii and mass bin jj, σijrad\sigma_{ij}^{\mathrm{rad}} is the radiative product cross section integrated over the bin and normalized with the full Belle luminosity LB\mathcal{L}_B. If these events enter the ISR sample, their contribution to Belle’s reported cross section is extracted using the smaller effective ISR luminosity LISR,j\mathcal{L}_{\mathrm{ISR},j}; we denote this contribution by σR,ij\sigma_{R,ij}.

The radiative product is the e+e−→γRe^+e^-\to\gamma R production cross section multiplied by B(R→h1h2)\mathcal{B}(R\to h_1h_2) and any daughter branching fractions retained in the measured observable. The published baryon cross section is corrected for the Λc\Lambda_c reconstruction and inclusive Λˉc→pˉX\bar{\Lambda}_c\to\bar{p}X branching fractions [17], while the meson products retain the DsJD_{sJ} decay branching fractions [20, 21]. The corresponding reconstruction efficiencies are ϵijISR\epsilon_{ij}^{\mathrm{ISR}} and ϵijrad\epsilon_{ij}^{\mathrm{rad}}. We multiply Belle’s dressed meson cross sections and uncertainties by the vacuum polarization correction factor 0.94 to express them in the Born convention used for the BESIII vector baselines [20, 21]. We quote σR,ij\sigma_{R,ij} and σijrad\sigma_{ij}^{\mathrm{rad}} with the same Born normalization. Normalizing the radiative contribution with the two luminosities gives

σR,ij=LBLISR,jϵijradϵijISRσijrad.(5)\sigma_{R,ij}=\frac{\mathcal{L}_B}{\mathcal{L}_{\mathrm{ISR},j}}\frac{\epsilon_{ij}^{\mathrm{rad}}}{\epsilon_{ij}^{\mathrm{ISR}}}\sigma_{ij}^{\mathrm{rad}}. \tag*{(5)}

We set ϵijrad/ϵijISR=1\epsilon_{ij}^{\mathrm{rad}}/\epsilon_{ij}^{\mathrm{ISR}}=1. Counting radiative events as ISR therefore multiplies their cross section by LB/LISR,j\mathcal{L}_B/\mathcal{L}_{\mathrm{ISR},j}.

Belle reports LB=695 fb−1\mathcal{L}_{B}=695\ \mathrm{fb}^{-1} for Λc+Λˉc−\Lambda_{c}^{+}\bar{\Lambda}_{c}^{-} and LB=921.9 fb−1\mathcal{L}_{B}=921.9\ \mathrm{fb}^{-1} for each meson channel [17, 20, 21]. In DsDs1D_{s}D_{s1}, the bin 4.61<m<4.63 GeV4.61<m<4.63\ \mathrm{GeV} has LISR,j=90.6 pb−1=0.0906 fb−1\mathcal{L}_{\mathrm{ISR},j}=90.6\ \mathrm{pb}^{-1}=0.0906\ \mathrm{fb}^{-1}, efficiency 0.132%0.132\% including the branching fractions of the reconstructed Ds+D_{s}^{+} decay modes, and a fitted signal yield of 21.4 events [20]. If these events were radiatively produced, their contribution in the Born normalization adopted for the meson fits would be

σR,ij=0.94×21.490.6×0.00132 pb≃168.2 pb.(6)\sigma_{R,ij}=\frac{0.94\times21.4}{90.6\times0.00132}\ \mathrm{pb}\simeq168.2\ \mathrm{pb}. \tag*{(6)}

The same yield corresponds to a radiative product of

σijrad=0.94×21.4921.9×0.00132 fb≃16.5 fb.(7)\sigma_{ij}^{\mathrm{rad}}=\frac{0.94\times21.4}{921.9\times0.00132}\ \mathrm{fb}\simeq16.5\ \mathrm{fb}. \tag*{(7)}

With LB/LISR,j=1.02×104\mathcal{L}_{B}/\mathcal{L}_{\mathrm{ISR},j}=1.02\times10^{4}, a radiative product of 16.5 fb in this bin would contribute 168.2 pb to the spectrum in the same Born normalization.

Fits and candidate analysis.—We describe the Belle invariant mass spectra as the sum of the vector and radiative contributions,

σ(s)=σV(s)+σR(s),(8)\sigma(s)=\sigma_{V}(s)+\sigma_{R}(s), \tag*{(8)}

where σ\sigma is the total Born cross section fitted to Belle’s ISR spectrum, σV\sigma_{V} the vector contribution common to Belle ISR and BESIII direct scans, and σR\sigma_{R} the extra contribution extracted when radiative events are normalized with the ISR effective luminosity.

For Λc+Λˉc−\Lambda_{c}^{+}\bar{\Lambda}_{c}^{-} we use the cross section of Ref. [37], including the ψ(4500)\psi(4500), ψ(4660)\psi(4660), ψ(4790)\psi(4790), and ψ(4900)\psi(4900) contributions and their interference. Keeping its resonance masses, width parameters, and fitted form factor parameters at their published values, we compare the baseline with the sixteen BESIII points from 4.575 to 4.951 GeV and obtain χ2/dof=8.19/16\chi^{2}/\mathrm{dof}=8.19/16 [18, 19].

For the meson channels, we adopt BESIII’s published total fits containing two interfering vector Breit–Wigner amplitudes near 4.6 GeV and 4.7–4.8 GeV. BESIII reports χ2/dof=4.0/8\chi^{2}/\mathrm{dof}=4.0/8 for DsDs1D_{s}D_{s1} and 6.2/76.2/7 for DsDs2∗D_{s}D_{s2}^{*} [22]. These vector baselines are fixed as background in the Belle fits.

We describe σR\sigma_{R} with a relativistic Breit–Wigner and the phase space for γh1h2\gamma h_{1}h_{2},

σR(s)=Φ(3)(s)∣CfL(q)RBW⁡(s)∣2.(9)\sigma_{R}(s)=\Phi^{(3)}(s)\left|C f_{L}(q)\operatorname{RBW}(s)\right|^{2}. \tag*{(9)}

Here, Φ(3)(s)\Phi^{(3)}(s) is the phase space density for γh1h2\gamma h_{1}h_{2} as a function of the hadron pair mass m=sm=\sqrt{s}. Radiative production of such highly excited charmonia may involve complex unquenched effects [16], making the production strengths difficult to calculate. We therefore determine the overall normalization C≥0C\geq0 from a fit to each Belle spectrum. In the fitted line shape, we treat the smooth luminosity conversion factor as constant near the resonance and absorb it into CC. For resonance mass MRM_{R} and width ΓR\Gamma_{R} at s=MR2s=M_{R}^{2}, the relativistic Breit–Wigner function is RBW⁡(s)=MRΓR/[MR2−s−iMRΓR(s)]\operatorname{RBW}(s)=M_{R}\Gamma_{R}/[M_{R}^{2}-s-iM_{R}\Gamma_{R}(s)]. Its width depends on the hadron momentum q(s)q(s) in the h1h2h_{1}h_{2} rest frame and relative orbital angular momentum LL as

ΓR(s)=ΓR[q(s)qR]2L+1MRs[BL(q)BL(qR)]2,(10)\Gamma_{R}(s)=\Gamma_{R}\left[\frac{q(s)}{q_{R}}\right]^{2L+1}\frac{M_{R}}{\sqrt{s}}\left[\frac{B_{L}(q)}{B_{L}(q_{R})}\right]^{2}, \tag*{(10)}

with qR=q(MR2)q_{R}=q(M_{R}^{2}) and Blatt–Weisskopf barrier factors BLB_{L} [38]. We use a barrier radius r=3 GeV−1r=3\ \mathrm{GeV}^{-1} [39]. The decay factor is fL(q)=(q/qR)LBL(q)/BL(qR)f_{L}(q)=(q/q_{R})^{L}B_{L}(q)/B_{L}(q_{R}) [39], with f0=1f_{0}=1. For a massless photon, standard kinematics [40] gives

Φ(3)(s)=q(s)32π3(1−ss0).(11)\Phi^{(3)}(s)=\frac{q(s)}{32\pi^{3}}\left(1-\frac{s}{s_{0}}\right). \tag*{(11)}

Below threshold, σR=0\sigma_{R}=0. We use L=0L=0 for Λc+Λˉc−\Lambda_{c}^{+}\bar{\Lambda}_{c}^{-}. For the allowed decays of χcJ\chi_{cJ} states, we use L=1L=1; for ηc\eta_{c} states decaying into DsDs2∗D_{s}D_{s2}^{*}, we use L=2L=2. The bin averages σˉR,ij\bar{\sigma}_{R,ij} in the Born convention used in the fits give the radiative product over the mass window [ma,mb][m_{a},m_{b}]:

σirad[ma,mb]≃∑j⊂[ma,mb]σˉR,ijLISR,jLBϵijISRϵijrad.(12)\sigma_{i}^{\mathrm{rad}}[m_{a},m_{b}]\simeq\sum_{j\subset[m_{a},m_{b}]}\bar{\sigma}_{R,ij}\frac{\mathcal{L}_{\mathrm{ISR},j}}{\mathcal{L}_{B}}\frac{\epsilon_{ij}^{\mathrm{ISR}}}{\epsilon_{ij}^{\mathrm{rad}}}. \tag*{(12)}

For the luminosity conversion, we treat all Belle data as taken at s0=10.58 GeV\sqrt{s_{0}}=10.58\ \mathrm{GeV}.

The Λc+Λˉc−\Lambda_{c}^{+}\bar{\Lambda}_{c}^{-} fit uses sixteen Belle bins of 20 MeV from 4.58 to 4.90 GeV [17]. The meson fits cover 4.51–4.95 GeV for DsDs1D_{s}D_{s1} and 4.54–4.94 GeV for DsDs2∗D_{s}D_{s2}^{*} [20, 21]; their uncertainty treatment is given in the SM [25]. For Λc+Λˉc−\Lambda_{c}^{+}\bar{\Lambda}_{c}^{-}, the error bars in Belle’s published figure show statistical uncertainties only. We include the 10% experimental systematic of Ref. [17] in each bin, and an estimated common 4.9% uncertainty based on recent Λc\Lambda_{c} branching fraction measurements [41, 42]. The data term with common normalization nn is

χdata2(n)=∑j[yj−nσˉj]2uj2.(13)\chi_{\mathrm{data}}^{2}(n)=\sum_{j}\frac{\left[y_{j}-n\bar{\sigma}_{j}\right]^{2}}{u_{j}^{2}}. \tag*{(13)}

Here, jj labels the fitted mass bins, yjy_{j} is the measured cross section, and σˉj\bar{\sigma}_{j} is the total prediction averaged over bin jj. The factor nn accounts for the common branching fraction uncertainty. The uncertainty uju_{j} combines the statistical and 10% experimental systematic errors in quadrature. The full χ2\chi^{2} also includes the normalization constraint given in Ref. [25]. Uncertainties on fitted parameters are the Δχ2=1\Delta\chi^{2}=1 profile intervals [40].

We denote the additional resonances in Λc+Λˉc−\Lambda_{c}^{+}\bar{\Lambda}_{c}^{-}, DsDs1(2536)D_{s}D_{s1}(2536), and DsDs2∗(2573)D_{s}D_{s2}^{*}(2573) by X1X_{1}, X2X_{2}, and X3X_{3}, respectively. Figure 2 shows the fits to these three spectra. Our previous B0→K0Λc+Λˉc−B^{0}\to K^{0}\Lambda_{c}^{+}\bar{\Lambda}_{c}^{-} analysis [23] studied an ηc(6S)\eta_{c}(6S) contribution to the Λc+Λˉc−\Lambda_{c}^{+}\bar{\Lambda}_{c}^{-} spectrum through an SS-wave decay. We therefore consider this state as a candidate for X1X_{1} and adopt the fixed input (M1,Γ1)=(M_{1},\Gamma_{1})= (4629, 24.2) MeV from that study. As shown in Figure 2(a), this radiative contribution can describe Belle’s enhancement near 4.63 GeV above the fixed vector baseline. The fit reduces χ2\chi^2/dof from 17.12/16 for the vector baseline alone to 6.81/15, with a local significance of 3.2σ3.2\sigma. We also allow the mass and width to vary, obtaining M1=4640.1−13.0+6.0M_1=4640.1^{+6.0}_{-13.0} MeV and Γ1=28.7−23.2+33.8\Gamma_1=28.7^{+33.8}_{-23.2} MeV, with χ2\chi^2/dof = 4.67/13. Both fixed inputs lie within the corresponding individual 1σ1\sigma intervals of this best fit [25].

Figure 2. Fits to the (a) Λc+Λˉc−\Lambda_c^+\bar{\Lambda}_c^-, (b) DsDs1D_sD_{s1}, and (c) DsDs2∗D_sD_{s2}^* spectra. Black circles and green open squares show Belle and BESIII data. Grey dotted curves show the fixed vector baselines H0=σVH_0=\sigma_V. Red solid and purple dash-dotted curves show the total fits HXiH_{X_i} and radiative contributions σR(Xi)\sigma_R(X_i), including the fitted normalization. The labels X1X_1, X2X_2, and X3X_3 correspond to (a)–(c). The mass and width of X1X_1 are fixed; those of X2X_2 and X3X_3 are fitted. Error bars include statistical and systematic uncertainties.

For X2X_2, the fit gives M2=4627.3−7.0+6.7M_2=4627.3^{+6.7}_{-7.0} MeV and Γ2=23.9−15.2+16.2\Gamma_2=23.9^{+16.2}_{-15.2} MeV, lowering χ2\chi^2/dof from 35.72/22 for the vector baseline to 21.61/19. For X3X_3, the fit gives M3=4625.1−9.5+20.3M_3=4625.1^{+20.3}_{-9.5} MeV and Γ3=35.9−28.3+66.7\Gamma_3=35.9^{+66.7}_{-28.3} MeV, with χ2\chi^2/dof = 8.87/17, compared with 17.12/20 for the baseline alone. With their masses and widths fixed at the fitted values, the local significances are 3.8σ3.8\sigma for X2X_2 and 2.9σ2.9\sigma for X3X_3. The fitted radiative products are σ1rad=76.9−24.5+25.2\sigma_1^{\mathrm{rad}}=76.9^{+25.2}_{-24.5} fb, σ2rad=42.9−13.4+14.1\sigma_2^{\mathrm{rad}}=42.9^{+14.1}_{-13.4} fb, and σ3rad=60.7−26.6+29.6\sigma_3^{\mathrm{rad}}=60.7^{+29.6}_{-26.6} fb [25]. These contributions of tens of fb can produce the Belle enhancements through the ISR luminosity normalization. The ηc(6S)\eta_c(6S) inputs used for X1X_1 lie within the corresponding individual 1σ1\sigma intervals of the free X3X_3 fit. Fixing (M3,Γ3)=(4629,24.2)(M_3,\Gamma_3)=(4629,24.2) MeV gives χ2\chi^2/dof = 9.69/19 [25], so an ηc(6S)\eta_c(6S) contribution can also describe this spectrum. Since ηc→DsDs1\eta_c\to D_sD_{s1} is forbidden, we consider χc0\chi_{c0}, χc1\chi_{c1}, and χc2\chi_{c2} states with allowed PP-wave decays to this channel as candidates for X2X_2.

To examine possible charmonium assignments, we calculate the spectrum with a screened potential and the strong decays with the quark pair-creation (QPC) model [1, 2, 10, 43, 44, 45], retaining the parameters of our previous study [23]. The central calculated masses of the χc0\chi_{c0}, χc1\chi_{c1}, and χc2\chi_{c2} states are [25]

M[χcJ(5P)]=(4552.4,4563.1,4572.4) MeV,M[\chi_{cJ}(5P)] = (4552.4, 4563.1, 4572.4)\ \mathrm{MeV},
M[χcJ(6P)]=(4678.1,4685.9,4693.0) MeV.M[\chi_{cJ}(6P)] = (4678.1, 4685.9, 4693.0)\ \mathrm{MeV}.

The fitted X2X_2 mass lies between the central predictions for these two multiplets. Orbital mixing has been considered in studies of vector charmonia [1, 10]. Among the χcJ(6P)\chi_{cJ}(6P) states, only the χc2\chi_{c2} has an FF-wave partner with the same JPC=2++J^{PC}=2^{++}. Possible mixing of χc2(6P)\chi_{c2}(6P) with χc2(5F)\chi_{c2}(5F), whose calculated mass is 4726.4−34.7+33.74726.4^{+33.7}_{-34.7} MeV, could lower the χc2(6P)\chi_{c2}(6P) mass into the fitted range. We discuss this mixing possibility for a χc2\chi_{c2} interpretation of X2X_2 in the SM [25]. In addition, the possibly mixed χc2(6P)\chi_{c2}(6P) candidate may also contribute to the DsDs2∗D_sD_{s2}^* channel. However, the calculated ηc(6S)\eta_c(6S) partial width into this channel is larger than those of the mixed tensor candidates, supporting ηc(6S)\eta_c(6S) as a candidate for X3X_3.

Summary.— We have proposed radiative production of higher charmonia as a possible common origin of the differences between the Belle and BESIII line shapes. Radiatively produced C=+1C=+1 charmonia can pass Belle’s partial reconstruction requirements, and normalization by the effective ISR luminosity increases their contribution to the extracted cross sections. The baryonic and DsDs2∗D_sD_{s2}^* enhancements can be described by ηc(6S)\eta_c(6S) contributions, while a χc2\chi_{c2} candidate can describe DsDs1D_sD_{s1}. Possible PP–FF mixing offers an interpretation of the tensor candidate in the latter channel. The spectrum and decay properties of highly excited charmonia remain incompletely understood, and further studies are needed to identify the contributing states.

With larger samples, comparing photon angular distributions in the enhancement region and neighboring mass intervals would help test whether the excess contains a radiative component. At Belle II, full reconstruction of both hadrons would allow the recoil photon direction to be inferred from their momenta, helping distinguish ISR from radiative production. Hadron decay angles would also provide information on the resonance spin. For example, angular correlations involving Ds2∗→DKD_{s2}^{*} \to DK in the DsDs2∗D_sD_{s2}^{*} channel could help distinguish the ηc(6S)\eta_c(6S) and χc2\chi_{c2} assignments.

Similar differences in the line shapes may also occur in other open charm channels. Many Belle points in D∗+D∗−D^{*+}D^{*-} between 4.2 and 4.4 GeV lie above the BESIII measurements, although the two data sets remain consistent within uncertainties [15]. In this region, well above the D∗+D∗−D^{*+}D^{*-} threshold, the large vector baseline means that an additional radiative contribution may produce only a modest relative increase in the total cross section. Channels opening closer to the resonance region, such as D∗Dˉ2∗(2460)D^{*}\bar{D}_{2}^{*}(2460), may therefore be more promising for observing a similar effect. Close collaboration between theorists and experimentalists will help test this mechanism and search for similar resonance structures in other channels.

Acknowledgments

This work is supported by the Natural Science Foundation of Gansu Province (No. 26RCKA012 and No. 25JRRA799), the National Natural Science Foundation of China under Grants No. 12335001, No. 12247101, and No. 12135005, the 111 Center under Grant No. B20063, the National Key R&D Program of China under Contract No. 2024YFA1610503, the Fundamental Research Funds for the Central Universities (lzujbky-2023-stlt01), and Lanzhou City High-Level Talent Funding. T.-L. G. is supported by the Gansu Province Postgraduate Innovation Star Program No. 2026CXZX-042. The authors used ChatGPT for language editing and for assistance with the numerical code, reviewed the output, and take responsibility for the content.

Data availability. The cross sections come from Refs. [17, 18, 19, 20, 21, 22] and their supplemental material. The Belle DsDs2∗D_sD_{s2}^{*} points are digitized from the published figure.

Fitting methods

Photon angular distributions

The photon polar angle θγ∗\theta_\gamma^* is measured relative to the electron beam in the collision center of mass. The radiative distributions follow the calculation at leading order in Ref. [31], and the ISR distribution follows Ref. [35]. Each curve is averaged over bins of width 0.05 in cos⁡θγ∗\cos\theta_\gamma^* and normalized to unit area within ∣cos⁡θγ∗∣<0.9|\cos\theta_\gamma^*| < 0.9.

The full angular range is shown to display the ISR peaks along the beams. When both hadrons are reconstructed, their momenta determine the recoil photon direction, allowing the ISR peaks along the beams to be studied without detecting the photon [24]. In the partially reconstructed samples considered here, the photon must be detected, and the calorimeter acceptance excludes the directions closest to the beams.

We use ∣cos⁡θγ∗∣<0.9|\cos\theta_\gamma^*| < 0.9 as the central comparison interval, following several Belle ISR analyses [24, 32, 33, 34]. Figure S1 shows that the ISR and radiative distributions have similar shapes in this interval, where the strongest ISR peaks are excluded. The restricted angular coverage and present statistics therefore limit separation of the two processes.

Figure S1. Theoretical photon angular distributions at s0=10.58\sqrt{s_0}=10.58 GeV and hadron pair mass m=4.63m=4.63 GeV. Steps show averages over 0.05 bins. Each distribution has unit area within ∣cos⁡θγ∗∣<0.9|\cos\theta_\gamma^*|<0.9, marked by the vertical gray dashed lines.

Data and fitting method

The fits use 20 MeV bins: 16 bins over 4.58–4.90 GeV for Λc+Λˉc−\Lambda_c^+\bar{\Lambda}_c^-, 22 over 4.51–4.95 GeV for DsDs1(2536)D_sD_{s1}(2536), and 20 over 4.54–4.94 GeV for DsDs2∗(2573)D_sD_{s2}^*(2573) [17, 20, 21]. Belle’s collected integrated luminosity LB\mathcal{L}_B is 695 fb−1^{-1} for the baryonic channel and 921.9 fb−1^{-1} for each meson channel [17, 20, 21]. The central values of the published cross sections are retained. Channels related by charge conjugation are included throughout. The DsDs2∗D_sD_{s2}^* points are digitized from the published figure [21].

For i=1,2,3i=1,2,3, corresponding to Λc+Λˉc−\Lambda_c^+\bar{\Lambda}_c^-, DsDs1D_sD_{s1}, and DsDs2∗D_sD_{s2}^*, respectively, σirad\sigma_i^{\mathrm{rad}} denotes the radiative product including the decay branching fractions B(Ri→h1h2)Bsub,i\mathcal{B}(R_i\to h_1h_2)\mathcal{B}_{\mathrm{sub},i}. Here RiR_i is the additional resonance and h1h2h_1h_2 the hadron pair in channel ii. These products use the same Born normalization as the fitted spectra.

For Λc+Λˉc−\Lambda_c^+\bar{\Lambda}_c^-, Bsub,1=1\mathcal{B}_{\mathrm{sub},1}=1, because Belle’s published cross section has been corrected for the Λc\Lambda_c reconstruction and inclusive Λˉc−→pˉX\bar{\Lambda}_c^-\to\bar{p}X branching fractions [17]. The measured products in the meson channels include Bsub,2=B(Ds1−→Dˉ∗0K−)\mathcal{B}_{\mathrm{sub},2}=\mathcal{B}(D_{s1}^-\to\bar{D}^{*0}K^-) for DsDs1D_sD_{s1} and Bsub,3=B(Ds2∗−→Dˉ0K−)\mathcal{B}_{\mathrm{sub},3}=\mathcal{B}(D_{s2}^{*-}\to\bar{D}^{0}K^-) for DsDs2∗D_sD_{s2}^*, while the branching fractions of the reconstructed DsD_s decay modes are included in the efficiencies [20, 21].

We average the fitted Born contribution σR,i\sigma_{R,i} over each experimental mass bin and denote the result by σˉR,ij\bar{\sigma}_{R,ij}. The radiative product over the mass window [ma,mb][m_a,m_b] is then

σirad[ma,mb]≃∑j∈[ma,mb]σˉR,ijLISR,jLBϵijISRϵijrad.(14)\sigma_i^{\mathrm{rad}}[m_a,m_b] \simeq\sum_{j\in[m_a,m_b]} \bar{\sigma}_{R,ij}\frac{\mathcal{L}_{\mathrm{ISR},j}}{\mathcal{L}_B}\frac{\epsilon_{ij}^{\mathrm{ISR}}}{\epsilon_{ij}^{\mathrm{rad}}}. \tag*{(14)}

The ISR luminosity LISR,j\mathcal{L}_{\mathrm{ISR},j} is defined in the Letter [20, 21, 36]. We set ϵijrad/ϵijISR=1\epsilon_{ij}^{\mathrm{rad}}/\epsilon_{ij}^{\mathrm{ISR}}=1, so each bin contribution is multiplied by LISR,j/LB\mathcal{L}_{\mathrm{ISR},j}/\mathcal{L}_B before summing over the mass window. For the luminosity conversion, we treat all Belle data as taken at s0=10.58\sqrt{s_0}=10.58 GeV.

We fit the bin averages of the total cross section described in the Letter, with the vector baselines fixed as background [18, 19, 22, 26, 37]. For each channel ii, the data contribution to χi2\chi_i^2 is

χdata,i2(ai,ni)=∑j[yij−niσˉij(ai)]2uij2.(15)\chi_{\mathrm{data},i}^2(a_i,n_i)=\sum_j\frac{[y_{ij}-n_i\bar{\sigma}_{ij}(a_i)]^2}{u_{ij}^2}. \tag*{(15)}

Here yijy_{ij} is the measured cross section, σˉij\bar{\sigma}_{ij} is the average total prediction in bin jj, and aia_i denotes the resonance parameters varied in the fit. The uncertainty uiju_{ij} combines the statistical error and systematic errors that are independent between bins. A single factor nin_i multiplies the total prediction in every bin and is fitted together with the resonance parameters. The total χi2\chi_i^2 also includes the Gaussian constraint (ni−1)2/δni2(n_i-1)^2/\delta_{n_i}^2, where δni\delta_{n_i} is the relative common normalization uncertainty. The upper and lower 1σ1\sigma limits for each parameter are the values at which χi2\chi_i^2 increases by 1 from its minimum. The other free parameters, including nin_i, are refitted when determining these limits [40].

Experimental uncertainties

Belle’s original Λc+Λˉc−\Lambda_c^+\bar{\Lambda}_c^- figure gives statistical errors only [17]. Its total systematic uncertainty of about 42% is dominated by branching fraction uncertainties. We retain the 10% experimental contribution in the error for each point and estimate the common branching fraction uncertainty from recent BESIII measurements.

BESIII measures B(Λc+→pK−π+)=(6.61±0.11±0.12)%\mathcal{B}(\Lambda_c^+ \to pK^-\pi^+) = (6.61 \pm0.11 \pm0.12)\% [41] and B(Λˉc−→nˉX)=(32.4±0.7±1.5)%\mathcal{B}(\bar{\Lambda}_c^- \to\bar{n}X) = (32.4 \pm0.7 \pm1.5)\% [42]. Including subsequent antihyperon decays, B(Λˉc−→pˉX)=1−B(Λˉc−→nˉX)\mathcal{B}(\bar{\Lambda}_c^- \to\bar{p}X) = 1 - \mathcal{B}(\bar{\Lambda}_c^- \to\bar{n}X) gives (67.6±0.7±1.5)%(67.6 \pm0.7 \pm1.5)\%. Using the precision of pK−π+pK^-\pi^+ as a reference for the reconstructed Λc\Lambda_c modes, we conservatively add the two relative errors, rBF≃2.5%+2.4%=4.9%r_{\mathrm{BF}} \simeq2.5\% + 2.4\% = 4.9\%, to estimate the common normalization uncertainty.

For DsDs1D_sD_{s1}, we estimate the common normalization uncertainty as 4.9% based on Belle’s efficiency variation, retaining the remaining systematic errors for each point [20].

Λc+Λˉc−\Lambda_c^+\bar{\Lambda}_c^-DsDs1D_sD_{s1}DsDs2∗D_sD_{s2}^*
Fixed MR,ΓRM_R,\Gamma_RFixed MR,ΓRM_R,\Gamma_RFree MR,ΓRM_R,\Gamma_RFree MR,ΓRM_R,\Gamma_RFree MR,ΓRM_R,\Gamma_R
MRM_R [MeV]46294640.1−13.0+6.04640.1^{+6.0}_{-13.0}4627.3−7.0+6.74627.3^{+6.7}_{-7.0}4625.1−9.5+20.34625.1^{+20.3}_{-9.5}
ΓR\Gamma_R [MeV]24.228.7−23.2+33.828.7^{+33.8}_{-23.2}23.9−15.2+16.223.9^{+16.2}_{-15.2}35.9−28.3+66.735.9^{+66.7}_{-28.3}
CiC_i [(pb/GeV)1/2^{1/2}]1220.1−212.6+185.51220.1^{+185.5}_{-212.6}1181.4−300.7+127.91181.4^{+127.9}_{-300.7}748.9−164.0+346.8748.9^{+346.8}_{-164.0}776.5−263.3+606.7776.5^{+606.7}_{-263.3}
σR(MR2)\sigma_R(M_R^2) [pb]435.7−138.6+142.6435.7^{+142.6}_{-138.6}447.1−205.6+1105.3447.1^{+1105.3}_{-205.6}241.2−93.3+277.2241.2^{+277.2}_{-93.3}218.4−115.9+451.4218.4^{+451.4}_{-115.9}
σirad\sigma_i^{\mathrm{rad}} [fb]76.9−24.5+25.276.9^{+25.2}_{-24.5}93.4−34.8+40.193.4^{+40.1}_{-34.8}42.9−13.4+14.142.9^{+14.1}_{-13.4}60.7−26.6+29.660.7^{+29.6}_{-26.6}
χ2/dof\chi^2/\mathrm{dof}6.81/154.67/1321.61/198.87/17

Table SM-I. Fit results for the three Belle spectra with fixed vector baselines. The vector baselines alone give χ2/dof=17.12/16\chi^2/\mathrm{dof} = 17.12/16, 35.72/2235.72/22, and 17.12/2017.12/20 in these channels, respectively. Radiative products use the same Born normalization as the fitted spectra and the mass intervals given above. The coefficient CiC_i is the fitted amplitude normalization for channel ii. Errors correspond to Δχ2=1\Delta\chi^2 = 1.

For DsDs2∗D_sD_{s2}^*, we treat the full 10.9% systematic uncertainty as a common normalization uncertainty and use the statistical error for each point [21].

Fit results

Panels (a) and (b) of Figure S2 compare the Λc+Λˉc−\Lambda_c^+\bar{\Lambda}_c^- fits with fixed and free mass and width. We adopt the fixed input (MR,ΓR)=(4629,24.2)(M_R,\Gamma_R) = (4629, 24.2) MeV from the previous analysis of the B0→KS0Λc+Λˉc−B^0 \to K_S^0\Lambda_c^+\bar{\Lambda}_c^- line shape [23]. Both fixed inputs lie within the corresponding individual 1σ1\sigma intervals of the free fit. The local significance with fixed mass and width is 3.2σ3.2\sigma.

Figure S2. Fits to Belle data (black circles) for e+e−→Λc+Λ‾c−e^{+}e^{-}\to\Lambda_{c}^{+}\overline{\Lambda}_{c}^{-} in (a,b), e+e−→DsDs1e^{+}e^{-}\to D_{s}D_{s1} in (c), and e+e−→DsDs2∗e^{+}e^{-}\to D_{s}D_{s2}^{*} in (d); green open squares show BESIII data. The resonance mass and width are fixed in (a) and fitted in (b)–(d). The resonance terms use L=0L=0 in (a,b), L=1L=1 in (c), and L=2L=2 in (d). Grey dotted, red solid, and purple dash-dotted curves show the fixed vector baseline, total fit, and radiative contribution. The total and radiative curves include the fitted normalization. Error bars include statistical and systematic uncertainties.

Panels (c) and (d) of Figure S2 show the meson fits. With the masses and widths fixed at their fitted values, the local significances are 3.8σ3.8\sigma for DsDs1D_sD_{s1} and 2.9σ2.9\sigma for DsDs2∗D_sD_{s2}^*. Fixing the mass and width in DsDs2∗D_sD_{s2}^* to (4629,24.2)(4629, 24.2) MeV gives χ2/dof=9.69/19\chi^2/\mathrm{dof} = 9.69/19 and a local significance of 2.7σ2.7\sigma; both inputs lie within the individual 1σ1\sigma intervals of the free fit.

Spectroscopy calculations

We calculate the charmonium masses and wave functions in the modified Godfrey–Isgur (MGI) model used in our previous study [23]. The linear scalar confinement br+cbr+c is replaced by a screened form, written before smearing as

Vconf(r)=c+bμ(1−e−μr),(16)V_{\mathrm{conf}}(r) = c + \frac{b}{\mu}\left(1-e^{-\mu r}\right), \tag*{(16)}

Here rr is the quark–antiquark separation, and μ\mu is the screening parameter. The color Coulomb and spin dependent interactions, Gaussian smearing, and momentum dependent relativistic corrections follow the GI formalism [43] as used in Refs. [2, 23].

We adopt the central parameters mc=1.65m_c = 1.65 GeV,

μ=0.12\mu= 0.12 GeV, b=0.238b = 0.238 GeV2^2, and c=−0.337c = -0.337 GeV from Ref. [23]. Following the same study, we estimate the mass uncertainties by varying μ=0.12±0.01\mu= 0.12 \pm0.01 GeV. The values (b,c)=(0.230 GeV2,−0.329 GeV)(b,c) = (0.230\ \mathrm{GeV}^2,-0.329\ \mathrm{GeV}) and (0.246 GeV2,−0.344 GeV)(0.246\ \mathrm{GeV}^2,-0.344\ \mathrm{GeV}) are used at μ=0.11\mu= 0.11 and 0.130.13 GeV, respectively, with all other parameters unchanged. The quoted errors give the shifts from the central masses to the largest and smallest masses obtained with these three parameter sets.

Table SM-II gives the calculated masses and model uncertainties of the nearby pseudoscalar and χcJ\chi_{cJ} states. The central masses of the χcJ(5P)\chi_{cJ}(5P) and χcJ(6P)\chi_{cJ}(6P) multiplets lie below and above the fitted meson masses, while ηc(6S)\eta_c(6S) lies close to the enhancement region.

The strong decays are calculated in the quark pair-creation (QPC) model [44, 45], in which a light qqˉq\bar q pair is created with JPC=0++J^{PC}=0^{++}. The strengths for creating uuˉ/ddˉu\bar{u}/d\bar{d} and ssˉs\bar{s} pairs are γu,d=5.84\gamma_{u,d}=5.84 and γs=5.84/3\gamma_{s}=5.84/\sqrt{3}, respectively, and the constituent quark masses are mu,d=0.22m_{u,d}=0.22 GeV and ms=0.55m_{s}=0.55 GeV. The wave functions of the final mesons are taken from [27, 28]. The final meson masses are (mDs,mDs1,mDs2∗)=(1.970,2.536,2.573)(m_{D_{s}},m_{D_{s1}},m_{D_{s2}^{*}})=(1.970, 2.536, 2.573) GeV.

StateMassStateMass
ηc(6S)\eta_{c}(6S)4606.6−24.9+24.44606.6^{+24.4}_{-24.9}χc2(5F)\chi_{c2}(5F)4726.4−34.7+33.74726.4^{+33.7}_{-34.7}
χc0(5P)\chi_{c0}(5P)4552.4−20.1+19.74552.4^{+19.7}_{-20.1}χc0(6P)\chi_{c0}(6P)4678.1−31.3+30.64678.1^{+30.6}_{-31.3}
χc1(5P)\chi_{c1}(5P)4563.1−20.8+20.44563.1^{+20.4}_{-20.8}χc1(6P)\chi_{c1}(6P)4685.9−32.0+31.34685.9^{+31.3}_{-32.0}
χc2(5P)\chi_{c2}(5P)4572.4−21.5+21.04572.4^{+21.0}_{-21.5}χc2(6P)\chi_{c2}(6P)4693.0−32.8+32.04693.0^{+32.0}_{-32.8}

Table SM-II. Calculated masses in MeV. The errors show the model variation under the three parameter sets specified in the text.

For Ds1(2536)D_{s1}(2536), we use the spin mixture ∣Ds1(2536)⟩=cos⁡ϕ ∣3P1⟩−sin⁡ϕ ∣1P1⟩|D_{s1}(2536)\rangle=\cos\phi\,|{}^{3}P_{1}\rangle-\sin\phi\,|{}^{1}P_{1}\rangle, with ϕ=−54.7∘\phi=-54.7^\circ. The decay calculations use the DsDs2∗D_{s}D_{s2}^{*} fit mass, 4625.1 MeV, for ηc(6S)\eta_{c}(6S). For the candidate formed by mixing χc2(6P)\chi_{c2}(6P) and

χc2(5F)\chi_{c2}(5F), we use the DsDs1D_{s}D_{s1} fit mass of 4627.3 MeV. The partial widths are calculated with the amplitude normalization of Ref. [23],

Γi(MR)=π4qiMR2∑J,L∣MiJL(qi)∣2.(17)\Gamma_{i}(M_{R})=\frac{\pi}{4}\frac{q_{i}}{M_{R}^{2}}\sum_{J,L}\left|M_{i}^{JL}(q_{i})\right|^{2}. \tag*{(17)}

Here MiJLM_{i}^{JL} is the QPC partial-wave decay amplitude for channel ii with final total spin JJ and relative orbital angular momentum LL. The two final mesons have momentum magnitude qiq_{i} in the resonance rest frame. Charge conjugates are included.

To illustrate possible mixing of χc2(6P)\chi_{c2}(6P) and χc2(5F)\chi_{c2}(5F), we denote the lower and higher mass states by RLR_{L} and RHR_{H}, respectively:

∣RL⟩=cos⁡θ ∣6P⟩−sin⁡θ ∣5F⟩,∣RH⟩=sin⁡θ ∣6P⟩+cos⁡θ ∣5F⟩,(18)\begin{aligned} |R_{L}\rangle&= \cos\theta\,|6P\rangle-\sin\theta\,|5F\rangle,\\ |R_{H}\rangle&= \sin\theta\,|6P\rangle+\cos\theta\,|5F\rangle, \tag*{(18)} \end{aligned}

Here ∣6P⟩|6P\rangle and ∣5F⟩|5F\rangle denote the unmixed χc2(6P)\chi_{c2}(6P) and χc2(5F)\chi_{c2}(5F) states, both with JPC=2++J^{PC}=2^{++}, and θ\theta is the mixing angle. Using the central 6P and 5F masses, we choose ∣θ∣=39.15∘|\theta|=39.15^\circ to place RLR_{L} at the fitted DsDs1D_{s}D_{s1} mass of 4627.3 MeV. This angle illustrates possible additional mixing.

StateMRM_R [MeV]DsDs1D_sD_{s1}DsDs2∗D_sD_{s2}^{*}
ηc(6S)\eta_c(6S)4625.101.4
RL(+39.15∘)R_L(+39.15^\circ)4627.34.20.1
RL(−39.15∘)R_L(-39.15^\circ)4627.35.10.1

Table SM-III. Partial widths in keV, using the DsDs2∗D_sD_{s2}^{*} fitted mass for ηc\eta_c and the DsDs1D_sD_{s1} fitted mass for RLR_L.

Table SM-III shows that the calculated ηc(6S)\eta_c(6S) partial width into DsDs2∗D_sD_{s2}^{*} is larger than those of the mixed χc2\chi_{c2} candidates. We therefore use the ηc(6S)\eta_c(6S) candidate for this channel and a tensor candidate for DsDs1D_sD_{s1}, where the ηc\eta_c decay is forbidden. Additional tensor contributions to DsDs2∗D_sD_{s2}^{*} remain possible; their relative yields depend on radiative production rates, full resonance widths, and reconstruction efficiencies. Further angular measurements can help separate the contributions.

References

References

  1. [1]J.-Z. Wang, D.-Y. Chen, X. Liu, and T. Matsuki, Constructing J/ψ family with updated data of charmoniumlike Y states, Phys. Rev. D 99, 114003 (2019), arXiv:1903.07115 [hep-ph].DOI
  2. [2]J.-Z. Wang, R.-Q. Qian, X. Liu, and T. Matsuki, Are the Y states around 4.6 GeV from e⁺e⁻ annihilation higher charmonia?, Phys. Rev. D 101, 034001 (2020), arXiv:2001.00175 [hep-ph].DOI
  3. [3]D.-Y. Chen, J. He, and X. Liu, Nonresonant explanation for the Y(4260) structure observed in the e⁺e⁻ → J/ψπ⁺π⁻ process, Phys. Rev. D 83, 054021 (2011), arXiv:1012.5362 [hep-ph].DOI
  4. [4]D.-Y. Chen, J. He, and X. Liu, Novel explanation of charmoniumlike structure in e⁺e⁻ → ψ(2S)π⁺π⁻, Phys. Rev. D 83, 074012 (2011), arXiv:1101.2474 [hep-ph].
  5. [5]D.-Y. Chen, X. Liu, X.-Q. Li, and H.-W. Ke, Unified Fano-like interference picture for charmoniumlike states Y(4008), Y(4260) and Y(4360), Phys. Rev. D 93, 014011 (2016), arXiv:1512.04157 [hep-ph].DOI
  6. [6]D.-Y. Chen, X. Liu, and T. Matsuki, Interference effect as resonance killer of newly observed charmoniumlike states Y(4320) and Y(4390), Eur. Phys. J. C 78, 136 (2018), arXiv:1708.01954 [hep-ph].DOI
  7. [7]L.-P. He, D.-Y. Chen, X. Liu, and T. Matsuki, Prediction of a missing higher charmonium around 4.26 GeV in J/ψ family, Eur. Phys. J. C 74, 3208 (2014), arXiv:1405.3831 [hep-ph].DOI
  8. [8]D.-Y. Chen, X. Liu, and T. Matsuki, Observation of e⁺e⁻ → χc0ω and missing higher charmonium ψ(4S), Phys. Rev. D 91, 094023 (2015), arXiv:1411.5136 [hep-ph].
  9. [9]D.-Y. Chen, X. Liu, and T. Matsuki, Search for missing ψ(4S) in the e⁺e⁻ → π⁺π⁻ψ(2S) process, Phys. Rev. D 93, 034028 (2016), arXiv:1509.00736 [hep-ph].DOI
  10. [10]J.-Z. Wang and X. Liu, Confirming the existence of a new higher charmonium ψ(4500) by the newly released data of e⁺e⁻ → K⁺K⁻J/ψ, Phys. Rev. D 107, 054016 (2023), arXiv:2212.13512 [hep-ph].DOI
  11. [11]J.-Z. Wang and X. Liu, Identifying a characterized energy level structure of higher charmonium well matched to the peak structures in e⁺e⁻ → π⁺π⁻D⁰D̄⁰, Phys. Lett. B 849, 138456 (2024), arXiv:2306.14695 [hep-ph].arxiv.org/abs/2306.14695
  12. [12]T.-C. Peng, Z.-Y. Bai, J.-Z. Wang, and X. Liu, How higher charmonia shape the puzzling data of the e⁺e⁻ → ηJ/ψ cross section, Phys. Rev. D 109, 094048 (2024), arXiv:2403.03705 [hep-ph].DOI
  13. [13]Z.-L. Man, S.-Q. Luo, Z.-Y. Bai, and X. Liu, Coupled-channel study of 4S-3D mixing dynamics in ψ(4220) and ψ(4380), Phys. Lett. B 868, 139644 (2025), arXiv:2502.08072 [hep-ph].DOI
  14. [14]H.-X. Chen, W. Chen, X. Liu, Y.-R. Liu, and S.-L. Zhu, An updated review of the new hadron states, Rept. Prog. Phys. 86, 026201 (2023), arXiv:2204.02649 [hep-ph].DOI
  15. [15]X. Wang, X. Liu, and Y. Gao, Colloquium: Hadron production in open-charm meson pairs at e⁺e⁻ colliders, Rev. Mod. Phys. 98, 021001 (2026), arXiv:2502.15117 [hep-ex].DOI
  16. [16]Z.-Y. Bai, D.-Y. Chen, Q. Huang, X. Liu, S.-Q. Luo, and J.-Z. Wang, Unquenched charmonium and beyond, Phys. Rept. 1204, 1 (2027), arXiv:2602.19887 [hep-ph].DOI
  17. [17]G. Pakhlova et al. (Belle), Observation of a near-threshold enhancement in the e⁺e⁻ → Λc⁺Λ̄c⁻ cross section using initial-state radiation, Phys. Rev. Lett. 101, 172001 (2008), arXiv:0807.4458 [hep-ex].DOI
  18. [18]M. Ablikim et al. (BESIII), Precision measurement of the e⁺e⁻ → Λc⁺Λ̄c⁻ cross section near threshold, Phys. Rev. Lett. 120, 132001 (2018), arXiv:1710.00150 [hep-ex].DOI
  19. [19]M. Ablikim et al. (BESIII), Measurement of Energy-Dependent Pair-Production Cross Section and Electromagnetic Form Factors of a Charmed Baryon, Phys. Rev. Lett. 131, 191901 (2023), arXiv:2307.07316 [hep-ex].DOI
  20. [20]S. Jia et al. (Belle), Observation of a vector charmonium-like state in e⁺e⁻ → D⁺sDs1(2536) + c.c., Phys. Rev. D 100, 111103 (2019), arXiv:1911.00671 [hep-ex].
  21. [21]S. Jia et al. (Belle), Evidence for a vector charmoniumlike state in e⁺e⁻ → D⁺sD∗−s(2573) + c.c., Phys. Rev. D 101, 091101 (2020), arXiv:2004.02404 [hep-ex].
  22. [22]M. Ablikim et al. (BESIII), Study of the decay and production properties of Ds1(2536) and Ds2(2573), Phys. Rev. Lett. 133, 171903 (2024), arXiv:2407.07651 [hep-ex].arxiv.org/abs/2407.07651
  23. [23]C.-X. Liu, J.-Z. Wang, and X. Liu, Exploring charmonia ηc(6S, 7S) in the Λ⁺cΛ̄⁻c invariant mass spectrum of B⁰ → K⁻sΛ⁺cΛ̄⁻c, Phys. Rev. D 114, 074050 (2026), arXiv:2608.15524 [hep-ph].
  24. [24]G. Pakhlova et al. (Belle), Measurement of the near-threshold e⁺e⁻ → D̄D cross section using initial-state radiation, Phys. Rev. D 77, 011103 (2008), arXiv:0708.0082 [hep-ex].DOI
  25. [25]See Supplemental Material for the fitting methods and results of the spectroscopy calculations. The Supplemental Material cites Refs. [26–28].
  26. [26]C. Chen, B. Yan, and J.-J. Xie, The electromagnetic form factors and spin polarization of Λ⁺c in the process e⁺e⁻ → Λ⁺cΛ̄⁻c, Chin. Phys. C 49, 023102 (2025), arXiv:2407.19445 [hep-ph].DOI
  27. [27]Q.-T. Song, D.-Y. Chen, X. Liu, and T. Matsuki, Higher radial and orbital excitations in the charmed meson fam ily, Phys. Rev. D 92, 074011 (2015), arXiv:1503.05728 [hep-ph].
  28. [28]Q.-T. Song, D.-Y. Chen, X. Liu, and T. Matsuki, Charmed-strange mesons revisited: Mass spectra and strong decays, Phys. Rev. D 91, 054031 (2015), arXiv:1501.03575 [hep-ph].DOI
  29. [29]V. Zhukova et al. (Belle), Angular analysis of the e+e− → D(*)±D(*)∓ process near the open charm threshold using initial-state radiation, Phys. Rev. D 97, 012002 (2018), arXiv:1707.09167 [hep-ex].DOI
  30. [30]A. Bashirian et al. (Belle), The Belle detector, Nucl. Instrum. Methods Phys. Res. A 479, 117 (2002).
  31. [31]H. S. Chung, J. Lee, and C. Yu, Exclusive heavy quarkonium +γ production from e+e− annihilation into a virtual photon, Phys. Rev. D 78, 074022 (2008), arXiv:0808.1625 [hep-ph].DOI
  32. [32]G. Pakhlova et al. (Belle), Observation of the ψ(4415) → D̄D*2(2460) decay using initial-state radiation, Phys. Rev. Lett. 100, 062001 (2008), arXiv:0708.3313 [hep-ex].
  33. [33]G. Pakhlova et al. (Belle), Measurement of the e+e− → D0D̄*0 + π± cross section using initial-state radiation, Phys. Rev. D 80, 091101 (2009), arXiv:0908.0231 [hep-ex].DOI
  34. [34]G. Pakhlova et al. (Belle), Measurement of e+e− → D(*)s+D(*)s− cross sections near threshold using initial-state radiation, Phys. Rev. D 83, 011101 (2011), arXiv:1011.4397 [hep-ex].
  35. [35]S. Binner, J. H. Kühn, and K. Melnikov, Measuring σ(e+e− → hadrons) using tagged photons, Phys. Lett. B 459, 279 (1999), arXiv:hep-ph/9902399.
  36. [36]G. Pakhlova et al. (Belle), Measurement of the near-threshold e+e− → D(*)±D∓ cross section using initial-state radiation, Phys. Rev. Lett. 98, 092001 (2007), arXiv:hep-ex/0608018.
  37. [37]C. Chen, B. Yan, and J.-J. Xie, e+e− → Λ+cΛ̄−c cross sections and the Λ+c electromagnetic form factors within the extended vector meson dominance model, Chin. Phys. Lett. 41, 021302 (2024), arXiv:2312.16753 [hep-ph].DOI
  38. [38]F. von Hippel and C. Quigg, Centrifugal-barrier effects in resonance partial decay widths, shapes, and production amplitudes, Phys. Rev. D 5, 624 (1972).
  39. [39]R. Aaij et al. (LHCb), Observation of J/ψp resonances consistent with pentaquark states in Λ0b → J/ψK−p decays, Phys. Rev. Lett. 115, 072001 (2015), arXiv:1507.03414 [hep-ex].
  40. [40]F. Takahashi et al. (Particle Data Group), Review of Particle Physics, Int. J. Mod. Phys. A 41, 2630011 (2026).
  41. [41]M. Ablikim et al. (BESIII), Measurements of the absolute branching fractions of the Λ+c hadronic decays, JHEP 06 (2026), 269, arXiv:2601.01503 [hep-ex].DOI
  42. [42]M. Ablikim et al. (BESIII), Measurement of the absolute branching fraction of the inclusive decay Λ̄−c → n̄ + X, Phys. Rev. D 108, L031101 (2023), arXiv:2210.09561 [hep-ex].DOI
  43. [43]S. Godfrey and N. Isgur, Mesons in a Relativized Quark Model with Chromodynamics, Phys. Rev. D 32, 189 (1985).pubmed.ncbi.nlm.nih.gov/9955999
  44. [44]L. Micu, Decay rates of meson resonances in a quark model, Nucl. Phys. B 10, 521 (1969).DOI
  45. [45]A. Le Yaouanc, L. Oliver, O. Pène, and J.-C. Raynal, “Naive” quark-pair-creation model of strong-interaction vertices, Phys. Rev. D 8, 2223 (1973).

Paper details

vlnc:W_3wj2dz8u
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC0-1.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

ChatGPT

View statement
Looking for related papers…
Contents