Decays of the light hybrid meson 1 − ⁣ + 1^{\mathrm{-+}} G. Daylan Esmer, K. Azizi, H. Sundu, S. Türkmen Abstract The full width of the light isovector hybrid meson H V H_{\mathrm{V}} with spin-parities 1 − ⁣ + 1^{\mathrm{-+}} and content ( u ¯ ​ g ​ u − d ¯ ​ g ​ d ) / 2 (\overline{u}gu-\overline{d}gd)/\sqrt{2} is evaluated by considering the decays H V → ρ ± ​ π ∓ H_{\mathrm{V}}\to\rho^{\pm}\pi^{\mp} , b 1 ± ​ π ∓ b_{1}^{\pm}\pi^{\mp} , f 1 ​ ( 1285 ) ​ π f_{1}(1285)\pi , f 1 ​ ( 1420 ) ​ π f_{1}(1420)\pi , η ​ π \eta\pi , and η ′ ​ π \eta^{\prime}\pi . To calculate the partial widths of these channels, we use QCD three-point sum rule method which is necessary to determine strong couplings at the corresponding hybrid-meson-meson vertices. It turns out that the main contribution to the full width Γ ⁡ [ H V ] = ( 109.7 ± 16.0 ) ​ MeV \Gamma[H_{\mathrm{V}}]=(109.7\pm 16.0)~\mathrm{MeV} of the hybrid meson comes from the processes H V → ρ ± ​ π ∓ H_{\mathrm{V}}\to\rho^{\pm}\pi^{\mp} partial width of which amounts to ≈ 67 ​ MeV \approx 67~\mathrm{MeV} . The effects of the decays H V → b 1 ​ π H_{\mathrm{V}}\to b_{1}\pi and H V → f 1 ​ π , f 1 ′ ​ π H_{\mathrm{V}}\to f_{1}\pi,f_{1}^{\prime}\pi are also sizeable: Their partial widths are equal to 13 ​ MeV 13~\mathrm{MeV} and 20 ​ MeV 20~\mathrm{MeV} , respectively. The decays to η ​ π \eta\pi and η ′ ​ π \eta^{\prime}\pi mesons are subdominant reactions, nevertheless they form ≈ 9 % \approx 9\% of the full width Γ ⁡ [ H V ] \Gamma[H_{\mathrm{V}}] . Results obtained in this work may be interesting to unravel the tangle of predictions about H V H_{\mathrm{V}} existing in the literature, as well as useful in analyses of different resonances. I Introduction During last decades, the hadron spectroscopy was replenished with new particles and resonances discovered by experimental collaborations in different processes. Some of them were interpreted as excited states of mesons and baryons observed and identified already in the experiments. Others may be considered as evidences for conventional hadrons seen for the first time. But there are resonances which can not be included into the standard picture of q\overline{q}^{\prime} mesons and qq^{\prime}q^{\prime\prime} baryons. They may have diquark-antidiquark, pentaquark, glueball or hybrid structures and belong to the class of unusual or exotic hadrons. Various hybrid mesons containing apart from valence quarks also a valence gluon field are interesting objects for the experimental and theoretical studies. Actually such particles attracted interest of researches almost fifty years ago starting from the first days of the parton model and quantum chromodynamics (QCD) [ 1 ] . Theoretical investigations of hybrid states have rather long history and encompass numerous publications and results obtained in the framework of various methods and models [ 2 ] (and references therein). It is worth to emphasize that the hybrid mesons can carry quantum numbers which are forbidden for the conventional mesons. For instance, they can have spin-parities J^{\mathrm{PC}}={0^{--},~0^{+-},~1^{-+},~2^{-+}} which are not allowed for the mesons with quark-antiquark structure. Therefore, discovery of the particles with such quantum numbers provides valuable knowledge about the exotic mesons and features of the low energy QCD. Information on structures which can be considered as the potential candidates to hybrid mesons became available since the experiments reported in Ref. [ 3 ] , where {1^{-+}} meson with the mass (1406\pm 20)~\mathrm{MeV} and width (180\pm 30)~\mathrm{MeV} was observed in the exclusive reaction \pi^{-}p\rightarrow\pi^{0}n\eta . It was labeled as \pi_{1}(1400) and became the first member of the {1^{-+}} mesons’ family. The structure {1^{-+}} was explored in other experiments as well [ 4 , 5 , 6 , 7 , 8 ] . Data collected and critically analyzed during this process led to the picture: There are two isovector resonances \pi_{1}(1400) and \pi_{1}(1600) with J^{\mathrm{PC}}={1^{-+}} . But the coupled channel analyses carried out in Refs. [ 9 , 10 ] favor the existence of only one broad state \pi_{1}(1600) (see, also Ref. [ 11 ] ). The evidence for the next exotic meson \pi_{1}(2015) from this family was reported in Refs. [ 5 , 6 ] . The isoscalar vector particle \eta_{1}(1855) with the quantum numbers {1^{-+}} was seen quite recently by the BESIII collaboration in the radiative decay J/\psi\rightarrow\gamma\eta\eta^{\prime} [ 12 ] . The hybrid mesons of different contents and quantum numbers were explored in the context of numerous theoretical methods. Thus, they were considered using QCD sum rule (SR) method in the articles [ 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 ] . As a candidate to a lightest hybrid meson, the state \overline{q}gq with exotic spin-parities J^{\mathrm{PC}}={1^{-+}} was studied in a more detailed form. But available predictions for the mass of this particle are controversial and differ from each other considerably. Thus, in Ref. [ 14 ] the masses of the nonstrange and strange light hybrid states {1^{-+}} were found around of 1.5~\mathrm{GeV} and 1.6~\mathrm{GeV} , respectively. The result (1.7\pm 0.1)~\mathrm{GeV} was predicted in Ref. [ 15 ] . The comprehensive studies were carried out in the sum rules framework for the light hybrid mesons with J=0 , 1 and different combinations of the quantum numbers \mathrm{P} and \mathrm{C} as well [ 16 ] . The mass of the hybrid {1^{-+}} was found equal to (1.6-1.7)~\mathrm{GeV} , 1.81(6)~\mathrm{GeV} , (1.71\pm 0.22)~\mathrm{GeV} and (1.72-2.60)~\mathrm{GeV} in Refs. [ 17 , 18 , 21 , 22 ] , respectively. Recently, the SR studies of the light hybrid meson 1^{-+} were updated by including into analysis quark-gluon condensates up to ten dimensions: The mass of this particle amounts to 2.30_{-0.17}^{+0.18}~\mathrm{GeV} [ 26 ] . The parameters of the light hybrids, their production mechanisms and decay channels were explored using alternative theoretical approaches as well [ 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 ] . In fact, lattice computations predicted the existence of the hybrid meson {1^{-+}} with the mass (1.9\pm 0.2)~\mathrm{GeV} [ 27 ] . The two exotic states with masses 1439~\mathrm{MeV} and 1498~\mathrm{MeV} were found based on the Dyson-Schwinger equations in Ref. [ 30 ] . Relevant investigations were also carried out in the context of the flux tube [ 31 , 32 ] , the MIT bag [ 33 , 34 ] , a holographic QCD [ 35 ] and the constituent gluon models [ 36 , 37 ] . The decays of the hybrid state J^{\mathrm{PC}}={1^{-+}} and its full width is also the important problem which was addressed in the context of various methods. It is known that the meson \pi_{1}(1600) is the broad particle with the width 370_{-60}^{+50}~\mathrm{MeV} [ 11 ] . Having suggested a hybrid interpretation for \pi_{1}(1600) one should explain its decay modes as well. Theoretical analyses of this question led to interesting and sometimes to mutually exclusive conclusions. For instance, the lattice simulations and flux tube model predicted a prevalence of the decay to b_{1}\pi mesons as a main channel for the transformation of the hybrid state {1^{-+}} [ 28 , 29 ] , whereas in the SR computations the process {1^{-+}\rightarrow}\rho\pi was found to be the dominant one [ 19 ] . There are also other decay channels of the meson \pi_{1}(1600) reported in Ref. [ 11 ] and considered in the literature. In the current work, we explore decays of the isovector hybrid meson J^{\mathrm{PC}}={1^{-+}} by applying the QCD sum rule approach [ 38 , 39 ] . This method is one of the powerful and predictive nonperturbative tools for calculations of hadrons’ parameters. Originally elaborated to explore conventional mesons and baryons, it was successfully employed to study also the exotic hadrons [ 40 , 41 ] . In our analysis we rely on results for the mass and current coupling of this state obtained in Ref. [ 26 ] and, in what follows, denote it as H_{\mathrm{V}} . We are going to consider the decays H_{\mathrm{V}}\rightarrow\rho\pi , b_{1}\pi , f_{1}(1285)\pi , f_{1}(1420)\pi , \eta\pi , and \eta^{\prime}\pi of this particle. For simplicity of the presentation, we will employ notations f_{1} and f_{1}^{\prime} for the mesons f_{1}(1285) and f_{1}(1420) , respectively. Partial widths of these processes depend on the strong couplings g_{i} at the hybrid-meson-meson vertices. We determine them by employing the three-point sum rule method which allow us to find the sum rules for the corresponding form factors g_{i}(q^{2}) . At the mass shell of a final pion q^{2}=m_{\pi}^{2} these form factors give the required strong couplings. We divide the article into the following sections: In Sec. II , we consider the processes H_{\mathrm{V}}\rightarrow \rho^{-}\pi^{+} and H_{\mathrm{V}}\rightarrow\rho^{+}\pi^{-} . We evaluate the coupling at the vertex H_{\mathrm{V}}\rho^{-}\pi^{+} and find the partial width of this decay. The partial width of the second process H_{\mathrm{V}}\rightarrow\rho^{+}\pi^{-} is, evidently, equal to the width of the first one. In Sec. III , we concentrate on the channels H_{\mathrm{V}}\rightarrow b_{1}\pi , f_{1}\pi , and f_{1}^{\prime}\pi and calculate their partial widths. The decays to \eta\pi and \eta^{\prime}\pi mesons are studied in Sec. IV . Here, we also find the full width of the hybrid meson H_{\mathrm{V}} . In the last part of the article, we compare our predictions with results of various works obtained by means of SR or other methods. II Decays H_{\mathrm{V}}\rightarrow\rho\pi In this section, we explore the decays H_{\mathrm{V}}\rightarrow\rho^{-}\pi^{+} and H_{\mathrm{V}}\rightarrow\rho^{+}\pi^{-} of the hybrid isovector meson H_{\mathrm{V}} . We are going to present in a detailed form calculation of the process H_{\mathrm{V}}(p)\rightarrow\rho^{-}(p^{\prime})\pi^{+}(q) , where p , p^{\prime} and q are momenta of the particles involved into the decay. The coupling g_{1} which describes the strong interaction of particles at the vertex H_{\mathrm{V}}\rho^{-}\pi^{+} is equal to the form factor g_{1}(q^{2}) at the mass shell of the pion q^{2}=m_{\pi}^{2} . In its turn, the sum rule for g_{1}(q^{2}) can be obtained from analysis of the three-point correlation function. To this end, we consider the three-point correlator given by the expression \displaystyle\Pi_{\mu\nu}(p,p^{\prime}) \displaystyle= \displaystyle i^{2}\int d^{4}xd^{4}ye^{ip^{\prime}x+iyq}\langle 0|\mathcal{T}\{J_{\nu}^{\rho}(x) (1) \displaystyle\times J^{\pi}(y)J_{\mu}^{\dagger}(0)\}|0\rangle, where J_{\mu}(x) is the interpolating current for the vector hybrid state, \displaystyle J_{\mu}(x) \displaystyle= \displaystyle\frac{1}{\sqrt{2}}g_{s}\frac{\lambda_{ab}^{n}}{2}G_{\mu\theta}^{n}(x)\left[\overline{u}_{a}(x)\gamma^{\theta}u_{b}(x)\right. (2) \displaystyle\left.-\overline{d}_{a}(x)\gamma^{\theta}d_{b}(x)\right]. Here, g_{s} is the strong coupling constant, a,b=1,2,3 and n=1,2,\cdots,8 are color indices, \lambda^{n} and G_{\mu\theta}^{n}(x) are the Gell-Mann matrices and the gluon field strength tensor, respectively. The currents J_{\nu}^{\rho}(x) and J^{\pi}(x) for the conventional mesons \rho^{-} and \pi^{+} have the following forms J_{\nu}^{\rho}(x)=\overline{u}_{i}(x)\gamma_{\nu}d_{i}(x),\ J^{\pi}(x)=\overline{d}_{j}(x)i\gamma_{5}u_{j}(x), (3) with i,j=1,2,3 being the color indices. To derive the sum rule for the form factor g_{1}(q^{2}) one needs to write the correlator \Pi_{\mu\nu}(p,p^{\prime}) by employing the phenomenological parameters of the particles H_{\mathrm{V}} , \rho^{-} and \pi^{+} , and also express it using the quark and gluon propagators. The first expression \Pi_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime}) establishes the physical side of the sum rule equality, whereas the second one \Pi_{\mu\nu}^{\mathrm{OPE}}(p,p^{\prime}) is the QCD side of this equality. The \Pi_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime}) has the following form \displaystyle\Pi_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime})=\frac{\langle 0|J_{\nu}^{\rho}|\rho^{-}(p^{\prime},\varepsilon^{\prime})\rangle\langle 0|J^{\pi}|\pi^{+}(q)\rangle}{(p^{\prime 2}-m_{\rho}^{2})(q^{2}-m_{\pi}^{2})} \displaystyle\times\frac{\langle\rho^{-}(p^{\prime},\varepsilon^{\prime})\pi^{+}(q)|H_{\mathrm{V}}(p,\varepsilon)\rangle\langle H_{\mathrm{V}}(p,\varepsilon)|J_{\mu}^{\dagger}|0\rangle}{(p^{2}-m^{2})}+\cdots, (4) where m , m_{\rho} and m_{\pi} are the masses of the hybrid state H_{\mathrm{V}} and mesons \rho^{-} and \pi^{+} , respectively. Here, \varepsilon_{\mu} and \varepsilon_{\nu}^{\prime} are the polarization vectors of the particles H_{\mathrm{V}} and \rho^{-} . The term written down explicitly is the contribution of the ground-level particles: Effects of higher excited states and continuum are shown by the dots. The correlation function \Pi_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime}) can be rewritten in a more convenient form. For these purposes, we introduce the matrix elements \displaystyle\langle 0|J_{\nu}^{\rho}|\rho^{-}(p^{\prime},\varepsilon^{\prime})\rangle \displaystyle= \displaystyle m_{\rho}f_{\rho}\varepsilon_{\nu}^{\prime},\ \langle 0|J^{\pi}|\pi^{+}(q)\rangle=\frac{f_{\pi}m_{\pi}^{2}}{2m_{q}}, \displaystyle\langle 0|J_{\mu}|H_{\mathrm{V}}((p,\varepsilon))\rangle \displaystyle= \displaystyle mf\varepsilon_{\mu}, (5) and \langle\rho^{-}(p^{\prime},\varepsilon^{\prime})\pi^{+}(q)|H_{\mathrm{V}}(p,\varepsilon)\rangle=g_{1}(q^{2})\epsilon_{\sigma\tau\alpha\beta}q^{\alpha}p^{\prime\beta}\varepsilon^{\sigma}\varepsilon^{\prime\ast\tau}. (6) In Eq. ( 5 ), 2m_{q}=m_{u}+m_{d} , whereas f , f_{\rho} and f_{\pi} are the current coupling and decay constants of the hybrid H_{\mathrm{V}} and the mesons \rho^{-} and \pi^{+} , respectively. The matrix element \langle 0|J^{\pi}|\pi^{+}(q)\rangle can also be written using the parameter \mu_{\pi} \mu_{\pi}=\frac{m_{\pi}^{2}}{2m_{q}}=-\frac{\langle\overline{q}q\rangle}{f_{\pi}^{2}}. (7) The second equality in Eq. ( 7 ) is the relation between m_{\pi} , f_{\pi} , quark masses and the quark condensate \langle\overline{q}q\rangle arising from the partial conservation of the axial vector current (PCAC). Then the correlator \Pi_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime}) amounts to \displaystyle\Pi_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime}) \displaystyle= \displaystyle\frac{g_{1}(q^{2})mfm_{\rho}f_{\rho}f_{\pi}m_{\pi}^{2}}{2m_{q}(p^{\prime 2}-m_{\rho}^{2})(p^{2}-m^{2})} (8) \displaystyle\times\frac{\epsilon_{\alpha\beta\mu\nu}q^{\alpha}p^{\prime\beta}}{(q^{2}-m_{\pi}^{2})}\cdots. The function \Pi_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime}) contains only one term proportional to the structure \epsilon_{\alpha\beta\mu\nu}q^{\alpha}p^{\prime\beta} . The corresponding function \displaystyle\Pi^{\mathrm{Phys}}\left(p^{2},p^{\prime 2},q^{2}\right) \displaystyle= \displaystyle\frac{g_{1}(q^{2})mfm_{\rho}f_{\rho}}{2m_{q}(p^{\prime 2}-m_{\rho}^{2})(p^{2}-m^{2})} (9) \displaystyle\times\frac{f_{\pi}m_{\pi}^{2}}{(q^{2}-m_{\pi}^{2})}, is the invariant amplitude which will be employed in the following consideration. It can be presented through double dispersion integral [ 42 , 43 ] \displaystyle\Pi^{\mathrm{Phys}}(p^{2},p^{\prime 2},q^{2})=\frac{\widehat{g}_{1}(q^{2})}{\left(p^{2}-m^{2}\right)(p^{\prime 2}-m_{\rho}^{2})} \displaystyle+\int\int_{\Sigma}dsds^{\prime}\frac{\rho^{\mathrm{h}}(s,s^{\prime},q^{2})}{(s-p^{2})(s^{\prime}-p^{\prime 2})}+\cdots. (10) Here, \widehat{g}_{1}(q^{2}) contains input information from Eq. ( 8 ), whereas the second term is contributions of the excited and continuum states. The function \rho^{\mathrm{h}}(s,s^{\prime},q^{2}) is unknown hadronic double spectral density. The \Sigma is a region in the (s,s^{\prime}) plane and has the boundaries (s_{0},s_{0}^{\prime}) which depend on a process under consideration. The single dispersion integrals over s and s^{\prime} which vanish after the double Borel transformations are shown in Eq. ( 10 ) by the ellipses. The Eq. ( 10 ) can be obtained from the dispersion integral \Pi^{\mathrm{Phys}}(p^{2},p^{\prime 2},q^{2})=\int\int_{\Sigma}dsds^{\prime}\frac{\rho^{\mathrm{Phys}}(s,s^{\prime},q^{2})}{(s-p^{2})(s^{\prime}-p^{\prime 2})}, (11) using the double spectral density \displaystyle\rho^{\mathrm{Phys}}(s,s^{\prime},q^{2})=\widehat{g}_{1}(q^{2})\delta(s-m^{2})\delta(s^{\prime}-m_{\rho}^{2}) \displaystyle+\rho^{\mathrm{h}}(s,s^{\prime},q^{2})\theta(s-s_{0})\theta(s^{\prime}-s_{0}^{\prime}), (12) where, \theta(z) is the unit step function. To find the SR for the form factor g_{1}(q^{2}) , the same correlation function \Pi_{\mu\nu}(p,p^{\prime}) has to be expressed in terms of the light quark propagators and computed with some accuracy in the operator product expansion ( \mathrm{OPE} ). For \Pi_{\mu\nu}^{\mathrm{OPE}}(p,p^{\prime}) calculations lead to the result \displaystyle\Pi_{\mu\nu}^{\mathrm{OPE}}(p,p^{\prime})=\frac{g_{s}\lambda_{ab}^{n}}{2\sqrt{2}}i\int d^{4}xd^{4}ye^{ip^{\prime}x+iyq}G_{\mu\theta}^{n}(0) \displaystyle\times\left\{\mathrm{Tr}\left[\gamma_{\nu}S_{d}^{ij}(x-y)\gamma_{5}S_{u}^{jb}(y)\gamma^{\theta}S_{u}^{ai}(-x)\right]\right. \displaystyle\left.-\mathrm{Tr}\left[\gamma_{\nu}S_{d}^{ib}(x)\gamma^{\theta}S_{d}^{aj}(-y)\gamma_{5}S_{u}^{ji}(y-x)\right]\right\}. (13) In Eq. ( 13 ) S_{d(u)}^{ab}(x) are the propagators of the u and d quarks \displaystyle S_{u(d)}^{ab}(x)=i\delta_{ab}\frac{\not{x}}{2\pi^{2}x^{4}}-\delta_{ab}\frac{m_{q}}{4\pi^{2}x^{2}}-\delta_{ab}\frac{\langle\overline{q}q\rangle}{12} \displaystyle+i\delta_{ab}m_{u(d)}\frac{\not{x}\langle\overline{q}q\rangle}{48}-\delta_{ab}\frac{x^{2}}{192}\langle\overline{q}g_{s}\sigma Gq\rangle \displaystyle+i\delta_{ab}m_{u(d)}\frac{x^{2}\not{x}}{1152}\langle\overline{q}g_{s}\sigma Gq\rangle-i\frac{g_{s}G_{ab}^{\alpha\beta}}{32\pi^{2}x^{2}}\left[\not{x}{\sigma_{\alpha\beta}+\sigma_{\alpha\beta}}\not{x}\right] \displaystyle-i\delta_{ab}\frac{x^{2}\not{x}g_{s}^{2}\langle\overline{q}q\rangle^{2}}{7776}-\delta_{ab}\frac{x^{4}\langle\overline{q}q\rangle\langle g_{s}^{2}G^{2}\rangle}{27648}+\cdots, (14) where G_{\alpha\beta}^{ab}(x) is G_{\alpha\beta}^{ab}(x)=G_{\alpha\beta}^{m}(x)\lambda_{ab}^{m}/2. (15) In our calculations we adopt for Eq. ( 14 ) the approximation m_{u}=m_{d}=0 . The correlation function \Pi_{\mu\nu}^{\mathrm{OPE}}(p,p^{\prime}) contains the vacuum condensates generated by the terms in Eq. ( 14 ) which are not the full contributions. With expressions Eqs. ( 13 ) and ( 14 ) at hand, we can explain computation of the QCD side of the sum rule. As is seen, \Pi_{\mu\nu}^{\mathrm{OPE}}(p,p^{\prime}) contains three light quark propagators and the tensor G_{\mu\theta}^{n}(0) . This gluon field strength tensor contracted with the term -ig_{s}G_{\alpha\beta}^{ab}(x)\left[\not{x}\sigma^{\alpha\beta}+\sigma^{\alpha\beta}\not{x}\right]/32\pi^{2}x^{2} from the quark propagators gives rise to the matrix element of two-gluon fields sandwiched between the vacuum states \langle 0|G_{\alpha\beta}^{m}(x)G_{\mu\theta}^{n}(0)|0\rangle . This two-gluon matrix element is treated by two different manners. Thus, first we consider it as the full gluon propagator in coordinate space between two points 0 and x , and use the formula \displaystyle\langle 0|G_{\alpha\beta}^{m}(x)G_{\mu\theta}^{n}(0)|0\rangle=\frac{\delta^{mn}}{2\pi^{2}x^{4}}\left[g_{\beta\theta}\left(g_{\alpha\mu}-\frac{4x_{\alpha}x_{\mu}}{x^{2}}\right)\right. \displaystyle+(\beta,\theta)\leftrightarrow(\alpha,\mu)-\beta\leftrightarrow\alpha-\theta\leftrightarrow\mu]. (16) The matrix element \langle 0|G_{\alpha\beta}^{m}(x)G_{\mu\theta}^{n}(0)|0\rangle is also treated as the two-gluon condensate. To this end, the gluon field at point x is expanded around x=0 and the first term is taken into account. As a result, we get \displaystyle\langle 0|g_{s}^{2}G_{\alpha\beta}^{m}(x)G_{\mu\theta}^{n}(0)|0\rangle=\frac{\langle g_{s}^{2}G^{2}\rangle}{96}\delta^{mn}[g_{\alpha\mu}g_{\beta\theta} \displaystyle-g_{\alpha\theta}g_{\mu\beta}]. (17) After these manipulations, one multiplies these terms to remaining two light quark propagators and other factors and finishes calculations. The correlation function \Pi_{\mu\nu}^{\mathrm{OPE}}(p,p^{\prime}) has the same Lorentz structure as \Pi_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime}) . Having denoted the corresponding invariant amplitude by \Pi^{\mathrm{OPE}}(p^{2},p^{\prime 2},q^{2}) and applied the double dispersion relation, we get \Pi^{\mathrm{OPE}}(p^{2},p^{\prime 2},q^{2})=\int_{0}^{\infty}\int_{0}^{\infty}\frac{dsds^{\prime}\rho^{\mathrm{OPE}}(s,s^{\prime},q^{2})}{(s-p^{2})(s^{\prime}-p^{\prime 2})}+\cdots. (18) Note that the spectral density \rho^{\mathrm{OPE}}(s,s^{\prime},q^{2}) is equal to a double discontinuity (an imaginary part) of \Pi^{\mathrm{OPE}}(s,s^{\prime},q^{2}) . After equating the functions \Pi^{\mathrm{Phys}}(p^{2},p^{\prime 2},q^{2}) and \Pi^{\mathrm{OPE}}(p^{2},p^{\prime 2},q^{2}) , performing the double Borel transformations over the variables -p^{2} , -p^{\prime 2} and subtracting from the QCD side of this equality contributions of the excited and continuum states under the quark-hadron duality assumption \rho^{\mathrm{h}}(s,s^{\prime},q^{2})\simeq\rho^{\mathrm{OPE}}(s,s^{\prime},q^{2}) , we find the sum rule for g_{1}(q^{2}) \displaystyle g_{1}(q^{2})=\frac{2m_{q}(q^{2}-m_{\pi}^{2})}{mfm_{\rho}f_{\rho}f_{\pi}m_{\pi}^{2}}e^{m^{2}/M_{1}^{2}}e^{m_{\rho}^{2}/M_{2}^{2}} \displaystyle\times\Pi(\mathbf{M}^{2},\mathbf{s}_{0},q^{2}). (19) Here, the function \Pi(\mathbf{M}^{2},\mathbf{s}_{0},q^{2}) has the form \displaystyle\Pi(\mathbf{M}^{2},\mathbf{s}_{0},q^{2})=\int_{0}^{s_{0}}ds\int_{0}^{s_{0}^{\prime}}ds^{\prime}\rho^{\mathrm{OPE}}(s,s^{\prime},q^{2}) \displaystyle\times e^{-s/M_{1}^{2}}e^{-s^{\prime}/M_{2}^{2}}+\Pi(\mathbf{M}^{2}). (20) It contains the parameters \mathbf{M}^{2}=(M_{1}^{2},M_{2}^{2}) and \mathbf{s}_{0}=(s_{0},s_{0}^{\prime}) where the pairs (M_{1}^{2},s_{0}) and (M_{2}^{2},s_{0}^{\prime}) correspond to H_{\mathrm{V}} and \rho channels, respectively. The spectral density \rho^{\mathrm{OPE}}(s,s^{\prime},q^{2}) for the decay H_{\mathrm{V}}\rightarrow\rho\pi is given by the formula \displaystyle\rho^{\mathrm{OPE}}(s,s^{\prime},q^{2})=\frac{g_{s}^{2}m_{q}}{32\sqrt{2}\pi^{4}}\int_{0}^{1}d\alpha\int_{0}^{1-\alpha}\frac{d\beta}{\beta\alpha(1-\alpha-\beta)} \displaystyle\times\left[\beta^{2}+\alpha^{2}(\alpha-1)+\beta(\alpha^{2}+2\alpha-1)\right]\theta(N) (21) where \alpha and \beta are the Feynman parameters, and the argument N of the function \theta(N) amounts to N=(s\beta+s^{\prime}\alpha)(1-\alpha-\beta)+q^{2}\beta\alpha. (22) The nonperturbative term \Pi(\mathbf{M}^{2}) is determined directly from \Pi^{\mathrm{OPE}}(p^{2},p^{\prime 2},q^{2}) through double Borel transformations and has the form \displaystyle\Pi(\mathbf{M}^{2})=-\frac{\pi^{2}\langle\alpha_{s}G^{2}/\pi\rangle\langle\overline{q}q\rangle}{18\sqrt{2}} \displaystyle\times\int_{0}^{1}d\alpha\frac{1}{\alpha^{2}M_{1}^{2}}\delta\left(\frac{1}{M_{2}^{2}}-\frac{1-\alpha}{\alpha M_{1}^{2}}\right). (23) The SR given in Eq. ( 19 ) depend on the masses and decay constants of the particles involved into this decay. The parameters of the \rho and \pi mesons are well known quantities [ 11 ] : We use m_{\rho}=(775.11\pm 0.34)~\mathrm{MeV,} f_{\rho}=(216\pm 3)~\mathrm{MeV} and m_{\pi}=(139.57039\pm 0.00017)~\mathrm{MeV,} f_{\pi}=(130.2\pm 0.8)~\mathrm{MeV} , and m_{q}=(3.49\pm 0.07)~\mathrm{MeV} . The mass m= 2.30_{-0.17}^{+0.18}~\mathrm{GeV} and current coupling f=\sqrt{2}\cdot 0.38_{-0.04}^{+0.05}\cdot 10^{-1}~\mathrm{GeV}^{3} of the vector hybrid H_{\mathrm{V}} were evaluated in the sum rule framework in Ref. [ 26 ] . Besides, the correlation function \Pi(\mathbf{M}^{2},\mathbf{s}_{0},q^{2}) depends on the condensates \displaystyle\langle\alpha_{s}G^{2}/\pi\rangle \displaystyle= \displaystyle(0.012\pm 0.004)~\mathrm{GeV}^{4}, \displaystyle\langle\overline{q}q\rangle \displaystyle= \displaystyle-(0.24\pm 0.01)^{3}~\mathrm{GeV}^{3}, (24) which were extracted from analysis of the different hadronic processes [ 38 , 39 ] . For numerical calculations, one also needs to fix parameters \mathbf{M}^{2} , \mathbf{s}_{0} . They should meet standard requirements of the sum rule computations. Prevalence of the pole contribution to extracted quantities, as well as relative stability of numerical predictions under variations of \mathbf{M}^{2}=(M_{1}^{2},M_{2}^{2}) are among such important constrains. In the H_{\mathrm{V}} channel, we choose [ 26 ] M_{1}^{2}\in[2.5,3.5]~\mathrm{GeV}^{2},\ s_{0}\in[8,10]~\mathrm{GeV}^{2}. (25) These intervals coincide with corresponding working windows used in the calculation of the mass m and current coupling f of the hybrid H_{\mathrm{V}} , which allow us to avoid additional uncertainties in extracting SR data for g_{1}(q^{2}) . In the channel of the \rho meson, we employ M_{2}^{2}\in[1,2]~\mathrm{GeV}^{2},\ s_{0}^{\prime}\in[0.9,1.1]~\mathrm{GeV}^{2}. (26) The information about parameters permits us to carry out numerical computation of the form factor g_{1}(q^{2}) . At the pion’s mass shell q^{2}=m_{\pi}^{2} this function is equal to the strong coupling g_{1} which is necessary to determine the partial width of the decay H_{\mathrm{V}}\rightarrow\rho^{-}\pi^{+} . But the SR method can be applied for calculation of g_{1}(q^{2}) in the Euclidean region q^{2}<0 . To avoid this problem, we introduce the function g_{1}(Q^{2}) with Q^{2}=-q^{2} and use it in our analysis. As an example, in Fig. 1 we plot the SR result obtained for g_{1}(Q^{2}) at Q^{2}=10~\mathrm{GeV}^{2} , s_{0}=9~\mathrm{GeV}^{2} and s_{0}^{\prime}=1~\mathrm{GeV}^{2} as a function of the Borel parameters M_{1}^{2} and M_{2}^{2} . One can see, that there is residual dependence of g_{1}(Q^{2}) on these parameters. At this point, we find g_{1}(Q^{2}=10~\mathrm{GeV}^{2})=(3.35\pm 0.65)\ \mathrm{GeV}^{-1}, (27) where the theoretical uncertainties are equal to \pm 19\% of the central value and remain within limits acceptable for the SR computations. We perform the same analysis for g_{1}(Q^{2}) in the region Q^{2}=2-20~\mathrm{GeV}^{2} depicted in the left panel of Fig. 2 . Afterward, we use these SR data to find the extrapolating function \mathcal{F}_{1}(Q^{2}) which at the Q^{2}=2-20~\mathrm{GeV}^{2} coincides with them, but can be easily extended into a domain of positive Q^{2} . The function \mathcal{F}_{i}(Q^{2})=\mathcal{F}_{i}^{0}\mathrm{\exp}\left[c_{i}^{1}\frac{Q^{2}}{m^{2}}+c_{i}^{2}\left(\frac{Q^{2}}{m^{2}}\right)^{2}\right] (28) is one of convenient choices for these purposes. Here, \mathcal{F}_{i}^{0} , c_{i}^{1} , and c_{i}^{2} are parameters that are fixed from comparison with the SR data. Then, it is not difficult to extract the constants \mathcal{F}_{1}^{0}=1.03~\mathrm{GeV}^{-1} , c_{1}^{1}=0.78 , and c_{1}^{2}=-0.08 . The function \mathcal{F}_{1}(Q^{2}) is also plotted in Fig. 2 (right panel), in which is seen nice agreement between \mathcal{F}_{1}(Q^{2}) and the SR data. For the strong coupling g_{1} , we find g_{1}\equiv\mathcal{F}_{1}(-m_{\pi}^{2})=(1.03\pm 0.20)\ \mathrm{GeV}^{-1}. (29) The width of the process H_{\mathrm{V}}\rightarrow\rho^{-}\pi^{+} is given by the formula \Gamma\left[H_{\mathrm{V}}\rightarrow\rho^{-}\pi^{+}\right]=g_{1}^{2}\frac{\lambda_{1}^{3}}{12\pi}. (30) Here, \lambda_{1}=\lambda(m,m_{\rho},m_{\pi}) and the function \lambda(x,y,z) has the form \lambda(x,y,z)=\frac{\sqrt{x^{4}+y^{4}+z^{4}-2(x^{2}y^{2}+x^{2}z^{2}+y^{2}z^{2})}}{2x}. (31) As a result, we find \Gamma\left[H_{\mathrm{V}}\rightarrow\rho^{-}\pi^{+}\right]=(33.3\pm 10.1)~\mathrm{MeV}. (32) The width \Gamma\left[H_{\mathrm{V}}\rightarrow\rho^{+}\pi^{-}\right] of the second decay is also equal to Eq. ( 32 ). Therefore, the contribution of the processes H_{\mathrm{V}}\rightarrow\rho\pi to the full width of the hybrid meson H_{\mathrm{V}} amounts to \approx 67~\mathrm{MeV} . Figure 1: The SR result for the form factor g_{1}(Q^{2}) at Q^{2}=10~\mathrm{GeV}^{2} , s_{0}=9~\mathrm{GeV}^{2} and s_{0}^{\prime}=1~\mathrm{GeV}^{2} as a function of the Borel parameters M_{1}^{2} and M_{2}^{2} . Figure 2: The SR data for the form factor g_{1}(Q^{2}) with corresponding error bars (left panel). The fit function \mathcal{F}_{1}(Q^{2}) (right panel). The red diamond marks the point Q^{2}=-m_{\pi}^{2} , where the strong coupling g_{1} is evaluated. III Processes H_{\mathrm{V}}\rightarrow b_{1}\pi,\ f_{1}\pi , and f_{1}^{\prime}\pi This section is devoted to analysis of the decays H_{\mathrm{V}}\rightarrow b_{1}\pi , f_{1}(1285)\pi , and f_{1}(1420)\pi . The b_{1}(1235) is the axial-vector meson with spin-parity J^{\mathrm{PC}}={1^{+-}} , whereas f_{1}(1285) and f_{1}(1420) are mesons with the quantum numbers {1}^{++} . In the case of the b_{1} meson, we are going to study processes H_{\mathrm{V}}\rightarrow b_{1}^{-}\pi^{+} and H_{\mathrm{V}}\rightarrow b_{1}^{+}\pi^{-\text{ }} . The mesons f_{1} and f_{1}^{\prime} appear in the decays H_{\mathrm{V}}\rightarrow f_{1}\pi and H_{\mathrm{V}}\rightarrow f_{1}^{\prime}\pi . III.1 Decays H_{\mathrm{V}}\rightarrow b_{1}^{-}\pi^{+} and H_{\mathrm{V}}\rightarrow b_{1}^{+}\pi^{-} The partial widths of these processes are equal, therefore we investigate only one of them. Let us analyze the decay H_{\mathrm{V}}\rightarrow b_{1}^{-}\pi^{+} . To determine the coupling g_{2} responsible for the strong interaction of particles at the vertex H_{\mathrm{V}}b_{1}^{-}\pi^{+} , we start from the correlation function \displaystyle\widetilde{\Pi}_{\mu\nu}(p,p^{\prime}) \displaystyle= \displaystyle i^{2}\int d^{4}xd^{4}ye^{ip^{\prime}x+iyq}\langle 0|\mathcal{T}\{J_{\nu}^{b}(x) (33) \displaystyle\times J^{\pi}(y)J_{\mu}^{\dagger}(0)\}|0\rangle. The currents for the hybrid state J_{\mu}(x) and pion J^{\pi}(x) have been defined above. The interpolating current for the meson b_{1}^{-} is given by the formula J_{\nu}^{b}(x)=\overline{u}(x)\overleftrightarrow{D}_{\nu}(x)\gamma_{5}d(x), (34) where \overleftrightarrow{D}_{\nu} is \overleftrightarrow{D}_{\nu}(x)=\frac{1}{2}\left[\overrightarrow{D}(x)-\overleftarrow{D}(x)\right]. (35) Here, \displaystyle\overrightarrow{D}_{\nu}(x) \displaystyle= \displaystyle\overrightarrow{\partial}_{\nu}-i\frac{g_{s}}{2}\lambda^{n}A_{\nu}^{n}(x), \displaystyle\overleftarrow{D}_{\nu}(x) \displaystyle= \displaystyle\overleftarrow{\partial}_{\nu}+i\frac{g_{s}}{2}\lambda^{n}A_{\nu}^{n}(x), (36) where A_{\nu}^{a}(x) is the external gluon field. In the Fock-Schwinger gauge x^{\nu}A_{\nu}^{a}(x)=0 this field can be expressed in terms of the gluon field strength tensor A_{\nu}^{n}(x)=\frac{1}{2}x^{\beta}G_{\beta\nu}^{n}(0)+\frac{1}{3}x^{\alpha}x^{\beta}D_{\alpha}G_{\beta\nu}^{n}(0)+.... (37) The sum rule for the form factor g_{2}(q^{2}) is extracted from analysis of the correlator \widetilde{\Pi}_{\mu\nu}(p,p^{\prime}) . Then, the first component of the SR equality reads \displaystyle\widetilde{\Pi}_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime})=\frac{\langle 0|J_{\nu}^{b}|b_{1}^{-}(p^{\prime},\varepsilon^{\prime})\rangle\langle 0|J^{\pi}|\pi^{+}(q)\rangle}{(p^{\prime 2}-m_{b}^{2})(q^{2}-m_{\pi}^{2})} \displaystyle\times\frac{\langle b_{1}^{-}(p^{\prime},\varepsilon^{\prime})\pi^{+}(q)|H_{\mathrm{V}}(p,\varepsilon)\rangle\langle H_{\mathrm{V}}(p,\varepsilon)|J_{\mu}^{\dagger}|0\rangle}{(p^{2}-m^{2})}, (38) with m_{b}=(1229.5\pm 3.2)~\mathrm{MeV} being the mass of the meson b_{1}^{-} . To recast \widetilde{\Pi}_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime}) into a form suitable for further analysis, we make use of the following new matrix element \langle 0|J_{\nu}^{b}|b_{1}^{-}(p^{\prime},\varepsilon^{\prime})\rangle=m_{b}f_{b}\varepsilon_{\nu}^{\prime},\ (39) where f_{b}=181~\mathrm{MeV} is the decay constant of b_{1}^{-} . The gauge-invariant expression for the matrix element \langle b_{1}^{-}(p^{\prime},\varepsilon^{\prime})\pi^{+}(q)|H_{\mathrm{V}}(p,\varepsilon)\rangle in the case of the vector-axial-vector-pseudoscalar particles is \displaystyle\langle b_{1}^{-}(p^{\prime},\varepsilon^{\prime})\pi^{+}(q)|H_{\mathrm{V}}(p,\varepsilon)\rangle=g_{2}(q^{2})\left[(p\cdot q)(\varepsilon\cdot\varepsilon^{\prime\ast})\right. \displaystyle\left.-(p\cdot\varepsilon^{\prime\ast})(q\cdot\varepsilon)\right]. (40) This new and known matrix elements of the particles H_{\mathrm{V}} and \pi^{+} allow us to rewrite \widetilde{\Pi}_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime}) in the following form \displaystyle\widetilde{\Pi}_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime})=g_{2}(q^{2})\frac{mfm_{b}f_{b}f_{\pi}m_{\pi}^{2}}{2m_{q}(p^{\prime 2}-m_{b}^{2})(p^{2}-m^{2})} \displaystyle\times\frac{1}{(q^{2}-m_{\pi}^{2})}\left[\frac{m^{2}-m_{b}^{2}+q^{2}}{2}g_{\mu\nu}-p_{\mu}p_{\nu}\right. \displaystyle\left.+p_{\mu}^{\prime}p_{\nu}-\frac{m^{2}}{m_{b}^{2}}p_{\mu}^{\prime}p_{\nu}^{\prime}+\frac{m^{2}+m_{b}^{2}-q^{2}}{2m_{b}^{2}}p_{\mu}p_{\nu}^{\prime}\right]. (41) The QCD side of the required equality reads \displaystyle\widetilde{\Pi}_{\mu\nu}^{\mathrm{OPE}}(p,p^{\prime})=\frac{g_{s}\lambda_{ab}^{n}}{2\sqrt{2}}i\int d^{4}xd^{4}ye^{ip^{\prime}x+iyq}G_{\mu\theta}^{n}(0) \displaystyle\times\left\{\mathrm{Tr}\left[\gamma_{5}S_{d}^{ij}(x-y)\gamma_{5}S_{u}^{jb}(y)\gamma^{\theta}\overleftrightarrow{D}_{\nu}(x)S_{u}^{ai}(-x)\right]\right. \displaystyle\left.-\mathrm{Tr}\left[\gamma_{5}S_{d}^{ib}(x)\gamma^{\theta}S_{d}^{aj}(-y)\gamma_{5}\overleftrightarrow{D}_{\nu}(x)S_{u}^{ji}(y-x)\right]\right\}. (42) The SR for the form factor g_{2}(q^{2}) is obtained using the invariant amplitudes \widetilde{\Pi}^{\mathrm{Phys}}\left(p^{2},p^{\prime 2},q^{2}\right) and \widetilde{\Pi}^{\mathrm{OPE}}(p^{2},p^{\prime 2},q^{2}) that correspond to terms p_{\mu}p_{\nu} both in \widetilde{\Pi}_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime}) and \widetilde{\Pi}_{\mu\nu}^{\mathrm{OPE}}(p,p^{\prime}) . After standard Borel transformations and continuum subtraction, one gets \displaystyle g_{2}(q^{2})=\frac{2m_{q}(q^{2}-m_{\pi}^{2})}{mfm_{b}f_{b}f_{\pi}m_{\pi}^{2}}e^{m^{2}/M_{1}^{2}}e^{m_{b}^{2}/M_{2}^{2}} \displaystyle\times\widetilde{\Pi}(\mathbf{M}^{2},\mathbf{s}_{0},q^{2}). (43) The remaining operations are similar to ones explained in the previous section, therefore we omit these details. The difference here is connected with the choice of the region for the parameters in the b_{1}^{-} meson channel: It is introduced in the form M_{2}^{2}\in[1.5,2.5]~\mathrm{GeV}^{2},\ s_{0}^{\prime}\in[2,2.5]~\mathrm{GeV}^{2}. (44) The results of the SR computations are shown in Fig. 3 . The fit function \mathcal{F}_{2}(Q^{2}) employed to calculate the coupling g_{2} is fixed due to the parameters \mathcal{F}_{2}^{0}=0.40~\mathrm{GeV}^{-1},c_{2}^{1}=0.41,c_{2}^{2}=-0.02. (45) Then, it is easy to evaluate g_{2} g_{2}\equiv\mathcal{F}_{2}(-m_{\pi}^{2})=(4.0\pm 0.8)\times 10^{-1}\ \mathrm{GeV}^{-1}. (46) The width of the process H_{\mathrm{V}}\rightarrow b_{1}^{-}\pi^{+} is given by the expression \Gamma\left[H_{\mathrm{V}}\rightarrow b_{1}^{-}\pi^{+}\right]=g_{2}^{2}\frac{\lambda_{2}}{24\pi m^{2}}|M|^{2}. (47) Here, \displaystyle|M|^{2} \displaystyle= \displaystyle\frac{1}{4m_{b}^{2}}\left[m^{6}-2m_{\pi}^{2}m^{4}+2m_{b}^{2}(m_{b}^{2}-m_{\pi}^{2})^{2}\right. (48) \displaystyle\left.+m^{2}(m_{\pi}^{4}-3m_{b}^{4}+6m_{b}^{2}m_{\pi}^{2})\right], and \lambda_{2}=\lambda_{2}(m,m_{b},m_{\pi}). The partial width of this decay is equal to \Gamma\left[H_{\mathrm{V}}\rightarrow b_{1}^{-}\pi^{+}\right]=(6.4\pm 1.8)~\mathrm{MeV}. (49) The width of the second process H_{\mathrm{V}}\rightarrow b_{1}^{+}\pi^{-} is also given by Eq. ( 49 ). Figure 3: The sum rule’s data and the extrapolating function \mathcal{F}_{2}(Q^{2}) . The red square shows the position Q^{2}=-m_{\pi}^{2} . III.2 Processes H_{\mathrm{V}}\rightarrow f_{1}\pi and f_{1}^{\prime}\pi The treatment of the decays H_{\mathrm{V}}\rightarrow f_{1}\pi and H_{\mathrm{V}}\rightarrow f_{1}^{\prime}\pi due to mixing in the system f_{1}-f_{1}^{\prime} differs from analyses of the previous processes. Because the mesons f_{1} and f_{1}^{\prime} have the nonstrange and strange components, it is convenient to consider them in the quark-flavor basis. Originally, this basis was introduced to treat the mixing in the \eta-\eta^{\prime} system [ 44 ] , and later was used to explore different processes with \eta and \eta^{\prime} mesons [ 45 , 46 ] . But it can also be applied to the system of the axial-vector mesons f_{1} and f_{1}^{\prime} as well [ 47 ] . In this scheme the physical mesons f_{1} and f_{1}^{\prime} can be presented using the basic states |f_{1q}\rangle=(\overline{u}u+\overline{d}d)/\sqrt{2} and |f_{1s}\rangle=\overline{s}s through the formula \begin{pmatrix}f_{1}\\ f_{1}^{\prime}\end{pmatrix}=U(\phi)\begin{pmatrix}|f_{1q}\rangle\\ |f_{1s}\rangle\end{pmatrix}. (50) Here U(\phi)=\begin{pmatrix}\cos\phi&-\sin\phi\\ \sin\phi&\cos\phi\end{pmatrix}, (51) is the mixing matrix and \phi is the mixing angle. This assumption on the state mixing implies that the same pattern also applies to relevant currents, decay constants and wave functions. Then, in the quark-flavor basis the axial-vector currents for the mesons f_{1} and f_{1}^{\prime} acquire the following forms \begin{pmatrix}J_{\nu}^{f_{1}}(x)\\ J_{\nu}^{f_{1}^{\prime}}(x)\end{pmatrix}=U(\phi)\begin{pmatrix}J_{\nu}^{q}(x)\\ J_{\nu}^{s}(x)\end{pmatrix}, (52) where \displaystyle J_{\nu}^{q}(x) \displaystyle= \displaystyle\frac{1}{\sqrt{2}}\left[\overline{u}_{i}(x)\gamma_{\nu}\gamma_{5}u_{i}(x)+\overline{d}_{i}(x)i\gamma_{5}d_{i}(x)\right], \displaystyle J_{\nu}^{s}(x) \displaystyle= \displaystyle\overline{s}_{i}(x)\gamma_{\nu}\gamma_{5}s_{i}(x). (53) As is seen, the currents J_{\nu}^{f_{1}}(x) and J_{\nu}^{f_{1}^{\prime}}(x) have the \overline{q}q and \overline{s}s components. It is also clear that only \overline{q}q components of the mesons f_{1} and f_{1}^{\prime} contribute to the decays under consideration. The correlation function which should be explored is given by the formula \displaystyle\widehat{\Pi}_{\mu\nu}(p,p^{\prime}) \displaystyle= \displaystyle i^{2}\int d^{4}xd^{4}ye^{ip^{\prime}x+iyq}\langle 0|\mathcal{T}\{J_{\nu}^{f_{1}}(x) (54) \displaystyle\times J^{\pi 0}(y)J_{\mu}^{\dagger}(0)\}|0\rangle, where J_{\nu}^{f_{1}}(x) and J^{\pi 0}(x) are interpolating currents of the mesons f_{1} and \pi^{0} . They are defined by the expressions J_{\nu}^{f_{1}}(x)=\cos\phi J_{\nu}^{q}(x), (55) and \displaystyle J^{\pi 0}(x) \displaystyle= \displaystyle\frac{1}{\sqrt{2}}[\overline{u}_{j}(x)i\gamma_{5}u_{j}(x)-\overline{d}_{j}(x)i\gamma_{5}d_{j}(x)]. For the correlation function \widehat{\Pi}_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime}) , we get \displaystyle\widehat{\Pi}_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime})=\frac{\langle 0|J_{\nu}^{f_{1}}|f_{1}(p^{\prime},\varepsilon^{\prime})\rangle\langle 0|J^{\pi 0}|\pi^{0}(q)\rangle}{(p^{\prime 2}-m_{f_{1}}^{2})(q^{2}-m_{\pi}^{2})} \displaystyle\times\frac{\langle f_{1}(p^{\prime},\varepsilon^{\prime})\pi^{0}(q)|H_{\mathrm{V}}(p,\varepsilon)\rangle\langle H_{\mathrm{V}}(p,\varepsilon)|J_{\mu}^{\dagger}|0\rangle}{(p^{2}-m^{2})}, (57) where m_{f_{1}}=(1281.8\pm 0.5)~\mathrm{MeV} and \varepsilon_{\nu}^{\prime} are the mass and polarization vector of the meson f_{1} . To find \langle 0|J_{\nu}^{f_{1}}|f_{1}(p^{\prime},\varepsilon^{\prime})\rangle we introduce the matrix elements \displaystyle\langle 0|J_{\nu}^{q}|f_{1}(p^{\prime},\varepsilon^{\prime})\rangle \displaystyle= \displaystyle m_{f_{1}}f_{f_{1}}^{q}\varepsilon_{\nu}^{\prime}, \displaystyle\langle 0|J_{\nu}^{s}|f_{1}(p^{\prime},\varepsilon^{\prime})\rangle \displaystyle= \displaystyle m_{f_{1}}f_{f_{1}}^{s}\varepsilon_{\nu}^{\prime}, (58) where the decay constants f_{f_{1}}^{q} and f_{f_{1}}^{s} can be determined by means of the formula \begin{pmatrix}f_{f_{1}}^{q}&f_{f_{1}}^{s}\\ f_{f_{1}^{\prime}}^{q}&f_{f_{1}^{\prime}}^{s}\end{pmatrix}=U(\phi)\begin{pmatrix}f_{1q}&0\\ 0&f_{1s}\end{pmatrix}. (59) In other words, all four decay constants are expressed in terms of the two parameters f_{1q} and f_{1s} . The vertex \langle f_{1}(p^{\prime},\varepsilon^{\prime})\pi^{0}(q)|H_{\mathrm{V}}(p,\varepsilon)\rangle has the following form \displaystyle\langle f_{1}(p^{\prime},\varepsilon^{\prime})\pi^{0}(q)|H_{\mathrm{V}}(p,\varepsilon)\rangle=g_{3}(q^{2})\left[(p\cdot q)(\varepsilon\cdot\varepsilon^{\prime\ast})\right. \displaystyle\left.-(p\cdot\varepsilon^{\prime\ast})(q\cdot\varepsilon)\right]. (60) Then it is not difficult to write down \displaystyle\widehat{\Pi}_{\mu\nu}^{\mathrm{Phys}}(p,p^{\prime})=\frac{mfm_{\pi}^{2}m_{f_{1}}f_{1q}\cos^{2}\phi}{2m_{q}(p^{\prime 2}-m_{f_{1}}^{2})(p^{2}-m^{2})} \displaystyle\times\frac{1}{(q^{2}-m_{\pi}^{2})}\left[\frac{m^{2}-m_{f_{1}}^{2}+q^{2}}{2}g_{\mu\nu}-p_{\mu}p_{\nu}\right. \displaystyle\left.+p_{\mu}^{\prime}p_{\nu}-\frac{m^{2}}{m_{f_{1}}^{2}}p_{\mu}^{\prime}p_{\nu}^{\prime}+\frac{m^{2}+m_{f_{1}}^{2}-q^{2}}{2m_{f_{1}}^{2}}p_{\mu}p_{\nu}^{\prime}\right]. (61) The correlator \widehat{\Pi}_{\mu\nu}^{\mathrm{OPE}}(p,p^{\prime}) is given by the expression \displaystyle\widehat{\Pi}_{\mu\nu}^{\mathrm{OPE}}(p,p^{\prime})=i\frac{g_{s}\lambda_{ab}^{n}\cos\phi}{2\sqrt{2}}\int d^{4}xd^{4}ye^{ip^{\prime}x+iyq}G_{\mu\theta}^{n}(0) \displaystyle\times\left\{\mathrm{Tr}\left[\gamma_{\nu}\gamma_{5}S_{q}^{ib}(x)\gamma^{\theta}S_{q}^{aj}(-y)\gamma_{5}S_{q}^{ji}(y-x)\right]\right. \displaystyle\left.+\mathrm{Tr}\left[\gamma_{\nu}\gamma_{5}S_{q}^{ij}(x-y)\gamma_{5}S_{q}^{jb}(y)\gamma^{\theta}S_{q}^{ai}(-x)\right]\right\}, (62) where S_{q}(x) is the light d or u propagator. The sum rule for the form factor g_{3}(q^{2}) reads \displaystyle g_{3}(q^{2})=\frac{2m_{q}(q^{2}-m_{\pi}^{2})}{mfm_{\pi}^{2}m_{f_{1}}f_{1q}\cos\phi}e^{m^{2}/M_{1}^{2}}e^{m_{f_{1}}^{2}/M_{2}^{2}} \displaystyle\times\widehat{\Pi}(\mathbf{M}^{2},\mathbf{s}_{0},q^{2}). (63) Let us note that to extract the sum rule Eq. ( 63 ) we have used invariant amplitudes \widehat{\Pi}^{\mathrm{Phys}}\left(p^{2},p^{\prime 2},q^{2}\right) and \widehat{\Pi}^{\mathrm{OPE}}(p^{2},p^{\prime 2},q^{2}) that correspond to terms p_{\mu}p_{\nu} in the physical and \mathrm{OPE} expressions of the correlator \widehat{\Pi}_{\mu\nu}(p,p^{\prime}) . The numerical analysis are carried out after fixing the parameters of the f_{1}-f_{1}^{\prime} system. For the mixing angle \phi and decay constants f_{{}_{1}q} and f_{1s} we employ the values \displaystyle\phi \displaystyle= \displaystyle(24.0_{-2.7}^{+3.2})^{\circ},\ f_{1q}=193_{-38}^{+43}\ \mathrm{MeV,} \displaystyle f_{1s} \displaystyle= \displaystyle(230\pm 9)\ \mathrm{MeV}. (64) The mixing angle \phi in Eq. ( 64 ) was extracted by the LHCb collaboration from the process B_{d,s}^{0}\rightarrow J/\psi f_{1}(1285) [ 48 ] . The decay constants f_{1q} and f_{1s} were estimated in the framework of the covariant light-front quark model in Ref. [ 49 ] . Afterward, we apply the operations which have been explained above. Therefore, we omit details and write down only principal results. They have been found by employing the following Borel and continuum subtraction parameters: In the of the hybrid H_{\mathrm{V}} we have used the regions Eq. ( 25 ), whereas for the f_{1} channel employed M_{2}^{2}\in[1.5,2.5]~\mathrm{GeV}^{2},\ s_{0}^{\prime}\in[1.8,2.0]~\mathrm{GeV}^{2}. (65) Then the strong coupling g_{3} computed at the mass shell Q^{2}=-m_{\pi}^{2} amounts to g_{3}\equiv\mathcal{F}_{3}(-m_{\pi}^{2})=(5.6\pm 1.3)\times 10^{-1}\ \mathrm{GeV}^{-1}, (66) where \mathcal{F}_{3}(Q^{2}) is the fitting function with parameters \mathcal{F}_{3}^{0}=0.56~\mathrm{GeV}^{-1},c_{3}^{1}=0.46,c_{3}^{2}=-0.03 (see, Fig. 4 ). The partial width of the decay H_{\mathrm{V}}\rightarrow f_{1}\pi is \Gamma\left[H_{\mathrm{V}}\rightarrow f_{1}\pi\right]=(10.7\pm 3.6)~\mathrm{MeV}. (67) The decay H_{\mathrm{V}}\rightarrow f_{1}^{\prime}\pi can be studied in the same manner. The difference here is connected with \sin\phi and m_{f_{1}^{\prime}}=1428.4_{-1.3}^{+1.5}~\mathrm{MeV} in the relevant formulas. As a result, we find g_{3}^{\prime}\equiv\mathcal{F}_{3}^{\prime}(-m_{\pi}^{2})=(6.7\pm 1.4)\times 10^{-1}\ \mathrm{GeV}^{-1}. (68) In these calculations in the f_{1}^{\prime} channel we have employed the regions M_{2}^{2}\in[1.5,2.5]~\mathrm{GeV}^{2},\ s_{0}^{\prime}\in[2.9,3.0]~\mathrm{GeV}^{2}. (69) The width of the decay H_{\mathrm{V}}\rightarrow f_{1}^{\prime}\pi amounts to \Gamma\left[H_{\mathrm{V}}\rightarrow f_{1}^{\prime}\pi\right]=(9.6\pm 2.9)~\mathrm{MeV}. (70) Figure 4: QCD data and extrapolating functions \mathcal{F}_{3}(Q^{2}) (solid line) and \mathcal{F}_{3}^{\prime}(Q^{2}) (dotted line) The circle and square fix the point Q^{2}=-m_{\pi}^{2} . IV Reactions H_{\mathrm{V}}\to\eta\pi and H_{\mathrm{V}}\to\eta^{\prime}\pi The reactions H_{\mathrm{V}}\rightarrow\eta\pi and H_{\mathrm{V}}\rightarrow\eta^{\prime}\pi can be investigated by employing the technical tools described in the previous section. Here, because of U(1) anomaly we have to take into account the mixing in the \eta-\eta^{\prime} system. The physical mesons \eta and \eta^{\prime} , in general, can be formed using either the octet-singlet or quark-flavor bases of the flavor SU_{f}(3) group [ 44 ] . Mixing of the physical states, decay constants, leading and higher twist distribution amplitudes in the quark-flavor basis |\eta_{q}\rangle=(\overline{u}u+\overline{d}d)/\sqrt{2} and |\eta_{s}\rangle=\overline{s}s have very simple form. In the quark-flavor basis the physical mesons \eta and \eta^{\prime} are expressed in terms of the basic states |\eta_{q}\rangle and |\eta_{s}\rangle through the formula \begin{pmatrix}\eta\\ \eta^{\prime}\end{pmatrix}=U(\varphi)\begin{pmatrix}|\eta_{q}\rangle\\ |\eta_{s}\rangle\end{pmatrix}. (71) In Eq. ( 71 ) U(\varphi) is the mixing matrix U(\varphi)=\begin{pmatrix}\cos\varphi&-\sin\varphi\\ \sin\varphi&\cos\varphi\end{pmatrix}, (72) and \varphi is the corresponding mixing angle. Then, in the quark-flavor basis the currents for the pseudoscalar mesons \eta and \eta^{\prime} acquire the following forms \begin{pmatrix}J^{\eta}(x)\\ J^{\eta^{\prime}}(x)\end{pmatrix}=U(\varphi)\begin{pmatrix}J_{q}(x)\\ J_{s}(x)\end{pmatrix}, (73) where \displaystyle J_{q}(x) \displaystyle= \displaystyle\frac{1}{\sqrt{2}}\left[\overline{u}_{i}(x)i\gamma_{5}u_{i}(x)+\overline{d}_{i}(x)i\gamma_{5}d_{i}(x)\right], \displaystyle J_{s}(x) \displaystyle= \displaystyle\overline{s}_{i}(x)i\gamma_{5}s_{i}(x). (74) We start our studies from the process H_{\mathrm{V}}\rightarrow\eta\pi . In this case, we have to investigate the correlation function \displaystyle\Pi_{\mu}(p,p^{\prime}) \displaystyle= \displaystyle i^{2}\int d^{4}xd^{4}ye^{ip^{\prime}x+iyq}\langle 0|\mathcal{T}\{J^{\eta}(x)(x) (75) \displaystyle\times J^{\pi 0}(y)J_{\mu}^{\dagger}(0)\}|0\rangle. The contribution of the ground-level particles to the correlation function \Pi^{\prime}(p,p^{\prime}) has the form \displaystyle\Pi_{\mu}^{\mathrm{Phys}}(p,p^{\prime})=\frac{\langle 0|J^{\eta}|\eta(p^{\prime})\rangle\langle 0|J^{\pi 0}|\pi^{0}(q)\rangle}{(p^{2}-m^{2})(p^{\prime 2}-m_{\eta}^{2})} \displaystyle\times\frac{\langle\pi^{0}(q)\eta(p^{\prime})|H_{\mathrm{V}}(p,\varepsilon)\rangle\langle H_{\mathrm{V}}(p,\varepsilon)|J_{\mu}^{\dagger}|0\rangle}{(q^{2}-m_{\pi}^{2})} \displaystyle+\cdots, (76) where the dots indicate effects of higher resonances and continuum states. There are two new matrix elements in Eq. ( 76 ) which should be defined before computation of \Pi^{\mathrm{Phys}}(p,p^{\prime}) . To fix the first of them \langle 0|J^{\eta}|\eta(p^{\prime})\rangle it is convenient to introduce the matrix elements of the currents J_{q}(x) and J_{s}(x) \displaystyle\ 2m_{q}\langle 0|J_{q}|\eta(p^{\prime})\rangle \displaystyle= \displaystyle h_{\eta}^{q},\ \displaystyle 2m_{s}\langle 0|J_{s}|\eta(p^{\prime})\rangle \displaystyle= \displaystyle h_{\eta}^{s}, (77) with 2m_{q} being equal to m_{u}+m_{d} . In the |\eta_{q}\rangle- |\eta_{s}\rangle basis the twist-3 matrix elements h_{\eta}^{q} and h_{\eta}^{s} is obtained from the expression \begin{pmatrix}h_{\eta}^{q}&h_{\eta}^{s}\\ h_{\eta^{\prime}}^{q}&h_{\eta^{\prime}}^{s}\end{pmatrix}=U(\varphi)\begin{pmatrix}h_{q}&0\\ 0&h_{s}\end{pmatrix}, (78) where h_{s}=0.086~\mathrm{GeV}^{3} , and \varphi=39.3^{\circ}\pm 1.0^{\circ} were determined from a fit to experimental data [ 44 ] and h_{q}=0.0025~\mathrm{GeV}^{3} was found from theoretical analysis [ 50 , 51 ] . The next matrix element \langle\pi^{0}(q)\eta(p^{\prime})|H_{\mathrm{V}}(p,\varepsilon)\rangle is \langle\pi^{0}(q)\eta(p^{\prime})|H_{\mathrm{V}}(p,\varepsilon)\rangle=g_{4}(q^{2})\varepsilon\cdot p^{\prime}. (79) Having employed this input information, we get \displaystyle\Pi_{\mu}^{\mathrm{Phys}}(p,p^{\prime})=\frac{mff_{\pi}m_{\pi}^{2}h_{q}\cos^{2}\varphi}{4m_{q}^{2}(p^{2}-m^{2})(p^{\prime 2}-m_{\eta}^{2})(q^{2}-m_{\pi}^{2})} \displaystyle\times\left[\frac{m^{2}+m_{\eta}^{2}-q^{2}}{2m^{2}}p_{\mu}-p_{\mu}^{\prime}\right]+\cdots. (80) The correlator \Pi_{\mu}(p,p^{\prime}) amounts to \displaystyle\Pi_{\mu}^{\mathrm{OPE}}(p,p^{\prime})=-\frac{\cos\varphi g_{s}\lambda_{ab}^{n}}{2\sqrt{2}}i\int d^{4}xd^{4}ye^{ip^{\prime}x+iyq}G_{\mu\theta}^{n}(0) \displaystyle\times\left\{\mathrm{Tr}\left[\gamma_{5}S_{q}^{ib}(x)\gamma^{\theta}S_{q}^{aj}(-y)\gamma_{5}S_{q}^{ji}(y-x)\right]\right. \displaystyle\left.+\mathrm{Tr}\left[\gamma_{5}S_{q}^{ij}(x-y)\gamma_{5}S_{q}^{jb}(y)\gamma^{\theta}S_{q}^{ai}(-x)\right]\right\}. (81) The SR for the form factor g_{4}(q^{2}) can be obtained by means of amplitudes that correspond to structures p_{\mu} in the correlators \Pi_{\mu}^{\mathrm{Phys}}(p,p^{\prime}) and \Pi_{\mu}^{\mathrm{OPE}}(p,p^{\prime}) . Numerical computations of g_{4}(Q^{2}) have been carried out at Q^{2}=2-20~\mathrm{GeV}^{2} , using the working windows M_{2}^{2}\in[1,1.5]~\mathrm{GeV}^{2},\ s_{0}^{\prime}\in[0.8,1]~\mathrm{GeV}^{2}. (82) The extrapolating function \mathcal{F}_{4}(Q^{2}) to fit the SR data is determined by the parameters \mathcal{F}_{4}^{0}=1.29 , ~c_{4}^{1}=0.26,c_{4}^{2}=-0.01. This function and corresponding SR data are plotted in Fig. 5 . As a result, the coupling g_{4} is g_{4}\equiv\mathcal{F}_{4}(-m_{\pi}^{2})=(1.29\pm 0.27). (83) The width of the process H_{\mathrm{V}}\rightarrow\eta\pi is given by the formula \Gamma\left[H_{\mathrm{V}}\rightarrow\eta\pi\right]=g_{4}^{2}\frac{\lambda_{4}^{3}}{24\pi m^{2}}, (84) where \lambda_{4}=\lambda(m,m_{\eta},m_{\pi}) . The partial width of this channel is equal to \Gamma\left[H_{\mathrm{V}}\rightarrow\eta\pi\right]=(5.3\pm 1.6)~\mathrm{MeV}. (85) The similar calculations are carried out for the \eta^{\prime}\pi final-state as well. Here, differences are connected with necessities to replace in numerical analysis the factors \cos\varphi\rightarrow \sin\varphi , m_{\eta}\rightarrow m_{\eta^{\prime}} and also by the regions for M_{2}^{2}\in[1.5,2]~\mathrm{GeV}^{2},\ s_{0}^{\prime}\in[1.3,1.7]~\mathrm{GeV}^{2} . Our studies demonstrate that the extrapolating function \mathcal{F}_{5}(Q^{2}) has the parameters \mathcal{F}_{5}^{0}=1.49 , ~c_{5}^{1}=0.74,c_{5}^{2}=-0.04 (see, Fig. 5 ). Then the strong coupling g_{5} extracted from this function at Q^{2}=-m_{\pi}^{2} is g_{5}\equiv\mathcal{F}_{5}(-m_{\pi}^{2})=(1.49\pm 0.32). (86) The partial width of the process H_{\mathrm{V}}\rightarrow\eta^{\prime}\pi amounts to \Gamma\left[H_{\mathrm{V}}\rightarrow\eta^{\prime}\pi\right]=(4.7\pm 1.4)~\mathrm{MeV}. (87) Information gained in this and previous sections allows us to estimate the full width of the hybrid H_{\mathrm{V}} as \Gamma\left[H_{\mathrm{V}}\right]=(109.7\pm 16.0)~\mathrm{MeV}. (88) Figure 5: The SR data and the extrapolating functions \mathcal{F}_{4}(Q^{2}) for the strong coupling g_{4} (solid line) and \mathcal{F}_{5}(Q^{2}) for g_{5} (dot-dashed line). V Discussion and conclusions In the present article we have estimated the full width of the isovector hybrid meson H_{\mathrm{V}} with spin-parities J^{\mathrm{PC}}={1^{-+}} and content (\overline{u}gu-\overline{d}gd)/\sqrt{2} . We have considered the decays of this particle to \rho\pi , b_{1}\pi , f_{1}\pi , f_{1}^{\prime}\pi , \eta\pi , and \eta^{\prime}\pi mesons. The partial width of these channels have been calculated using QCD three-point sum rule method. This approach is necessary to evaluate the strong couplings at the corresponding hybrid-meson-meson vertices. The three-point sum rules for the form factors g_{i}(q^{2}) depend on the two independent Borel parameters and contains double integral over parameters s_{0} and s_{0}^{\prime} . We have used SRs for numerical computations of the QCD data which, by means of the fitting functions, have been later extrapolated to regions of negative Q^{2} to extract strong couplings of interest. It is worth to emphasize that we have performed these calculations without making any additional assumptions about the Borel parameters or imposing limits on the particles’ momenta. Our studies have demonstrated that aforementioned decays contribute to the full width of the hybrid H_{\mathrm{V}} in the proportions: \rho\pi : b_{1}\pi : f_{1}\pi : f_{1}^{\prime}\pi : \eta\pi : \eta^{\prime}\pi\rightarrow 61:12:10:9:5:3 . In other words, the dominant decay channel of H_{\mathrm{V}} is the process H_{\mathrm{V}}\rightarrow \rho\pi . Important are also decays H_{\mathrm{V}}\rightarrow b_{1}\pi and f_{1}^{(\prime)}\pi . The result \Gamma\left[H_{\mathrm{V}}\right]\approx 110~\mathrm{MeV} of the present article characterize the hybrid meson H_{\mathrm{V}} as a state with moderate full width. The processes with \eta\pi and \eta^{\prime}\pi mesons are subdominant modes, but their partial widths are of the same order as ones of the decays H_{\mathrm{V}}\rightarrow b_{1}\pi , f_{1}\pi . The total contribution of the decays H_{\mathrm{V}}\rightarrow\eta\pi , \eta^{\prime}\pi amounts to 9\% of the hybrid meson’s full width \Gamma\left[H_{\mathrm{V}}\right] . This is important result because in all existing calculations, widths of these decays were found negligibly small. The flux tube model even forbids decays to two ground-state mesons, i.e., to \eta\pi , \eta^{\prime}\pi pairs. This conclusion contradicts to experimental measurements, because structures \pi_{1}(1400) and \pi_{1}(1600) were also seen in the \eta\pi and \eta^{\prime}\pi channels [ 52 ] . The decays of H_{\mathrm{V}} were studied using the sum rule method in numerous publications. Thus, in Ref. [ 19 ] the authors used the three-point correlation function for the vertex H_{\mathrm{V}}\rho\pi and considered it in the limits q^{2}\rightarrow 0 ( q_{\mu} is a momentum of the pion) and m_{\pi}\rightarrow 0 . A pion-hybrid coupling was extracted from analysis of a singular term \sim 1/q^{2} in the correlator and used later to estimate the partial width of the decay H_{\mathrm{V}}\rightarrow \rho\pi . The width of this decay was found around 180~\mathrm{MeV} . The same approach allowed ones to evaluate widths of the decays H_{\mathrm{V}}\rightarrow . f_{1}\pi and b_{1}\pi with results 24~\mathrm{MeV} and 3~\mathrm{MeV} , respectively. The widths of the decays to \eta\pi and \eta^{\prime}\pi mesons were found equal < 1~\mathrm{MeV} . In the framework of the QCD light-cone sum rule approach the authors of Ref. [ 20 ] in the case m_{\mathrm{H}}=1.6~\mathrm{GeV} predicted: \Gamma[\rho\pi]=73\sim 120~\mathrm{MeV} , \Gamma\left[f_{1}\pi\right]=69\sim 122~\mathrm{MeV} , and \Gamma\left[b_{1}\pi\right]=0.14~\mathrm{MeV}. In the case of m_{\mathrm{H}}=2.0~\mathrm{GeV} the results obtained there read: \Gamma[\rho\pi]=216\sim 370~\mathrm{MeV} , \Gamma\left[f_{1}\pi\right]=109\sim 195~\mathrm{MeV} , and \Gamma\left[b_{1}\pi\right]=3.7~\mathrm{MeV} , respectively. Later these estimates were modified and for m_{\mathrm{H}}=2.0~\mathrm{GeV} became \Gamma[\rho\pi]=0.04~\mathrm{MeV} and \Gamma\left[b_{1}\pi\right]=52-152~\mathrm{MeV} [ 23 ] . The decay modes of the hybrid meson H_{\mathrm{V}} were also considered using alternative methods. In the lattice simulations [ 28 ] the mass of the hybrid meson {1^{-+}} was found equal to (2.2\pm 0.2)~\mathrm{GeV} . In this work the partial widths of the processes H_{\mathrm{V}}\rightarrow b_{1}\pi and H_{\mathrm{V}}\rightarrow f_{1}\pi were estimated \Gamma\left[b_{1}\pi\right]=(400\pm 120)~\mathrm{MeV} and \Gamma\left[f_{1}\pi\right]=(90\pm 60)~\mathrm{MeV} . In other words, this model implies fulfilment of the approximate ratio \Gamma\left[b_{1}\pi\right]/\Gamma\left[f_{1}\pi\right]\rightarrow 4/1 . The flux tube model gave, at the same time, \Gamma\left[b_{1}\pi\right]\approx 80~\mathrm{MeV} and \Gamma\left[f_{1}\pi\right]\approx 25~\mathrm{MeV} [ 29 ] , respectively. Comparing the mass 2.30_{-0.17}^{+0.18}~\mathrm{GeV} of the hybrid state H_{\mathrm{V}} from Ref. [ 26 ] with available experimental data, we see that it is considerably heavier than the structure \pi_{1}(1600) . Its mass is close to the mass of the resonance \pi_{1}(2015) . At the same time, the width of H_{\mathrm{V}} is smaller than the width \Gamma=(230\pm 32\pm 73)~\mathrm{MeV} of the hybrid candidate \pi_{1}(2015) presented in Ref. [ 6 ] . But \Gamma\left[H_{\mathrm{V}}\right] may be refined to reduce this gap by considering its other decay channels. Therefore the hybrid H_{\mathrm{V}} with some caution may be interpreted as the meson \pi_{1}(2015) . It is interesting that in Ref. [ 53 ] \pi_{1}(2015) was considered as radially excited state of the meson \pi_{1}(1600) . In Ref. [ 54 ] , the main decay modes and the decay ratios of light hybrid mesons including \pi_{1}(1600) are also investigated in the framework of the extended linear sigma model. Our detailed theoretical analyses in the present study show that the main decay channel of H_{\mathrm{V}} is the process H_{\mathrm{V}}\rightarrow\rho\pi . This is in accord with results of the sum rule analyses carried out in Refs. [ 19 , 20 ] though numerical predictions differ from our findings. In the flux tube model [ 29 ] and expectation of extended linear sigma model [ 54 ] , the main channel is decay to b_{1}\pi mesons. Even from this non exhaustive information about the dominant decay channels and full widths of the vector hybrid state H_{\mathrm{V}} obtained by means of different methods, it is evident that our knowledge about H_{\mathrm{V}} is far from being complete. Existing discrepancies between different approaches, or controversial results extracted in the context of the same method demonstrate that a lot of work should be done to reach a reliable conclusion about nature of this and similar exotic states. Further experimental studies of the processes where the states \pi_{1}(1600) and \pi_{1}(2015) were discovered are necessary. Such investigations are required to clarify a situation with the mesons \pi_{1}(1400) and \pi_{1}(1600) and may lead to observation of new hybrid candidates.