betaFeedback

WorksPhysics and Astronomy

Flavour decomposition of the chiral-odd form factors of the nucleon

Abstract

We compute the chiral-odd form factors HTH_T, ETE_T and H~T\tilde{H}_T of the nucleon within light-cone QCD sum rules, separately for the uu and dd quarks, over 1≤Q2≤101\le Q^2\le 10 GeV2^2, and organise their values at Q2=0Q^2=0 into the tensor monopole, dipole and quadrupole moments gTg_T, κTκ_T and QTQ_T. The flavour components of the quadrupole moment and the isoscalar channel have not previously been obtained in this framework. The form factors are read from a basis of Lorentz structures that becomes independent only after a canonical ordering of the Dirac strings, and three of the eight surviving structures give the three form factors separately. Two statements follow from the analytic expressions. The dd-quark contributions to ETE_T and H~T\tilde{H}_T are equal and opposite in the chiral limit, the breaking being proportional to the quark mass and to the twist-six amplitude V6\mathcal{V}_6. At the level of the least suppressed momentum-fraction integral, the isoscalar tensor charge receives no leading-twist contribution, the twist-three terms of the two flavours cancelling. Neither statement is visible without resolving the flavours. At μ2=1μ^2=1 GeV2^2 we obtain, for the two distribution-amplitude sets, gTu−d=1.27(7)g_T^{u-d}=1.27(7) and 1.08(6)1.08(6), κTu−d=1.35(62)κ_T^{u-d}=1.35(62) and 1.58(59)1.58(59), QTu−d=−4.40(83)Q_T^{u-d}=-4.40(83) and −4.20(73)-4.20(73), gTu+d=0.39(7)g_T^{u+d}=0.39(7) and 0.38(6)0.38(6), κTu+d=4.47(62)κ_T^{u+d}=4.47(62) and 4.32(59)4.32(59), and QTu+d=2.41(53)Q_T^{u+d}=2.41(53) and 1.56(30)1.56(30). The mean-field relation 2H~Tu−d=−ETu−d2\tilde{H}_T^{u-d}=-E_T^{u-d} holds in sign and in order of magnitude, the ratio being 0.64(16)0.64(16) and 0.67(15)0.67(15). In the impact-parameter plane the two flavour distributions are displaced in opposite transverse directions by nearly equal amounts, ⟨by⟩=0.37\langle b_y\rangle=0.37 and −0.37-0.37 fm, and both are elongated across the polarisation axis, the dd quark some five to six times more strongly than the uu quark.

Introduction

The internal structure of the nucleon is encoded in the matrix elements of local QCD operators. Those of the vector and axial-vector currents have been mapped in detail; the third leading-twist structure, the one probed by the tensor current, remains the least constrained. It measures how quarks are polarised transversely to the nucleon spin, and both the experimental and the theoretical determinations of it are still sparse. The present work computes the tensor form factors of the nucleon within QCD light-cone sum rules, resolved into uu- and dd-quark contributions, and organises them into the three multipole moments that carry the physical content: the tensor charge, the anomalous tensor magnetic moment and the tensor quadrupole moment. Resolving the flavours is not a refinement here but the substance of the calculation, since each of the structural results reported below lives in one flavour channel or in the interplay of the two.

The tensor current qˉiσμνq\bar{q}i\sigma_{\mu\nu}q is odd under charge conjugation, so its forward matrix element, the tensor charge δq\delta q, measures the difference between the transverse polarisation of quarks and that of antiquarks: sea quark–antiquark pairs cancel in it to the extent that quarks and antiquarks are equally polarised, and δq\delta q is in this sense a valence quantity rather than a measure of the full quark content of the nucleon [1, 2]. The axial-vector current behaves in the opposite way: it is C-even, sea-quark helicities survive, and the singlet axial charge is correspondingly difficult to interpret. The tensor charge is the first moment of the transversity distribution h1h_1, the last of the three leading-twist collinear distributions to be mapped [1, 3, 4, 5, 6, 7]: being chiral-odd, h1h_1 decouples from inclusive deep inelastic scattering and must be paired with a second chiral-odd object, which has restricted its extraction to semi-inclusive processes and to dihadron production, and it is constrained from above by the Soffer inequality [8]. Interest in the tensor charge is not confined to hadron structure. The flavour-diagonal tensor charges are precisely the matrix elements that convert a quark electric dipole moment into a neutron one [9, 10, 11], so a bound on the neutron electric dipole moment translates into a bound on CP-violating physics beyond the Standard Model only to the accuracy with which gTqg_T^q is known [12, 13]. The same matrix elements control the sensitivity of precision β\beta-decay measurements to BSM tensor interactions [14]. Experimentally, transversity and the tensor charge have been extracted from Collins asymmetries in semi-inclusive deep inelastic scattering measured by HERMES [15] and COMPASS [16], combined with e+e−e^{+}e^{-} data from Belle [17], and from dihadron production [18, 19]; global analyses of increasing scope have followed [20, 21, 22, 23]. The precision remains well below that of the helicity sector, and improving it is among the goals of the Electron-Ion Collider [24, 25]. On the theory side the tensor charge has been computed in lattice QCD [26, 27, 28, 29, 30], in the chiral quark-soliton model [31, 32, 33, 34], in the bag model [35], in light-front and covariant quark models [36, 37, 38, 39], in Dyson–Schwinger approaches [11, 40], and in QCD sum rules [41, 42, 43]. The tensor form factors of other hadronic systems have also been studied [44, 45, 46, 47, 48].

At leading twist the matrix element of the tensor current between spin-1/21/2 states is parametrised by three form factors, HTH_T, ETE_T and H~T\widetilde{H}_T, the first Mellin moments of the chiral-odd generalised parton distributions [49, 50]. As such they are the natural variables in momentum space, but for the spatial interpretation the physical content is carried by particular combinations of them. Expanding the matrix element in irreducible tensors built from the momentum transfer separates a monopole, a dipole and a quadrupole term, an enumeration carried out systematically for the chiral-odd operator in Ref. [51], where the quadrupole structure was identified as a term that earlier analyses had omitted. The monopole is the tensor charge. The dipole moment κT\kappa_T measures the shift of the average transverse position of quarks polarised in the transverse direction, in an unpolarised target [52]. That distortion, combined with final-state interactions, has been argued to generate the Boer–Mulders function [52], so that the sign of κT\kappa_T carries a phenomenological consequence; the chiral-odd distributions moreover allow the quark angular momentum to be decomposed with respect to quarks of definite transversity, giving access to the correlation between quark spin and orbital angular momentum in an unpolarised nucleon [52]. The quadrupole moment QTQ_T characterises the deformation of the corresponding distribution when the nucleon is transversely polarised as well, and is peculiar to the chiral-odd sector: at leading twist the chiral-odd operator carries L3=±1L_3 = \pm1, and combined with the rank-two tensor the quadrupole is the highest structure that can be formed, whereas the chiral-even operator has L3=0L_3 = 0 and cannot reach it at all [50]. The chiral-odd generalised parton distributions themselves have been studied on the lattice [53] and in a range of models [35, 37, 39, 54]. Their measurement meets the same obstacle as that of h1h_1, a second chiral-odd object being required, and exclusive processes in which this role is played by the distribution amplitude of a transversely polarised ρ\rho meson have been proposed, in the diffractive electroproduction of two vector mesons [55] and in the photoproduction of a γρ\gamma\rho pair at large invariant mass [56].

Within light-cone QCD sum rules the coverage is uneven. The isovector form factors were obtained in Ref. [42]; Ref. [43] derived the sum rules for both isovector and isoscalar currents using the general interpolating field, but quoted no numbers, having found that the form factors change sign with the mixing parameter. Neither study formed the multipole combinations. The isovector value of 2H~T(0)2\widetilde{H}_T(0) is implied by the H~Tμ−d(0)\widetilde{H}_T^{\mu-d}(0) of Ref. [42], and we construct it for the comparison of Sec. IV, but the quadrupole moment is not discussed there, and neither its flavour components nor its isoscalar combination have been obtained in this framework. The present work differs from these studies in four respects. The form factors are computed separately for the uu and dd quarks, and both isospin combinations are evaluated, the isoscalar one included. The flavour resolution requires no new technique, each flavour component of the tensor current being inserted into the correlator separately; what makes it informative is that, with the Ioffe current written with the two uu quarks adjacent, the uu- and dd-quark contractions draw in the chiral limit on disjoint sets of distribution amplitudes, and the two structural results of Secs. IV A and IV B follow from that separation. The three multipole moments, the quadrupole included, are formed in every flavour channel. The form factors are read from a basis of Lorentz structures that is independent only after a canonical ordering of the Dirac strings; before ordering the structures are linearly related, so the coefficient assigned to any one of them depends on how the expression happens to have been written. And the results are used to test a relation derived in the mean-field picture, from a framework that makes no assumption about NcN_c. The moments are finally carried to the impact-parameter plane, where the three of them fix the displacement and the deformation of the transverse quark distribution.

The paper is organised as follows. Section II sets up the multipole decomposition, fixes the independent basis and derives the sum rules. Section III describes the input, the working region of the auxiliary parameters and the extraction procedure. Section IV presents the results and discusses them. § 1.4 summarises. The QCD-side expressions are collected in the Appendix.

Formalism

Form factors

The matrix element of the tensor current between nucleon states is parametrised by three form factors,

⟨N(p′)∣qˉiσμνq∣N(p)⟩=uˉ(p′)[iσμνHTq(t)+γμqν−γνqμ2mNETq(t)+P~μqν−P~νqμ2mN2H~Tq(t)]u(p),(1)\langle N(p')|\bar{q}i\sigma_{\mu\nu}q|N(p)\rangle=\bar{u}(p')\left[i\sigma_{\mu\nu}H_T^q(t)+\frac{\gamma_\mu q_\nu-\gamma_\nu q_\mu}{2m_N}E_T^q(t)+\frac{\widetilde{P}_\mu q_\nu-\widetilde{P}_\nu q_\mu}{2m_N^2}\widetilde{H}_T^q(t)\right]u(p), \tag*{(1)}

with P~=p′+p\widetilde{P}=p^{\prime}+p, q=p−p′q=p-p^{\prime} and t=q2=−Q2t=q^{2}=-Q^{2}; qq is the momentum flowing into the tensor current, so that p′=p−qp^{\prime}=p-q. A fourth structure is allowed by Lorentz covariance, but the corresponding form factor E~T\widetilde{E}_{T} vanishes by T-invariance [43, 49]. The three are the first Mellin moments of the chiral-odd generalised parton distributions.

Multipole structure

As first Mellin moments of the chiral-odd generalised parton distributions the three form factors are the natural variables in momentum space, but they do not coincide one by one with the multipoles of the transverse spatial distribution: H~T\widetilde{H}_{T} enters both the dipole and the quadrupole, and ETE_{T} enters the dipole only in combination with it. The multipole content is exposed by expanding the matrix element in irreducible tensors built from the momentum transfer, which separates a monopole, a dipole and a quadrupole term. Expanding in the transverse plane at fixed light-front time, which is where the density interpretation is available, gives [51]

monopole:gTq=HTq(0),dipole:κTq=ETq(0)+2H~Tq(0),quadrupole:QTq=2H~Tq(0).(2)\begin{aligned} \text{monopole}:&\quad g_{T}^{q}=H_{T}^{q}(0),\\ \text{dipole}:&\quad\kappa_{T}^{q}=E_{T}^{q}(0)+2\widetilde{H}_{T}^{q}(0),\\ \text{quadrupole}:&\quad Q_{T}^{q}=2\widetilde{H}_{T}^{q}(0). \tag*{(2)} \end{aligned}

The monopole gTg_{T} is the tensor charge, the net number of transversely polarised quarks in a transversely polarised nucleon. The dipole κT\kappa_{T} measures the shift of the average transverse position of quarks polarised along a fixed transverse direction, in an unpolarised target [52] — the one multipole that survives when the nucleon polarisation is averaged over; it plays for the tensor current the role that the anomalous magnetic moment plays for the vector one. The quadrupole QTQ_{T} characterises the deformation of the transverse distribution when the nucleon is polarised as well [50], a configuration developed in Sec. IV D. A second, three-dimensional expansion in the Breit frame is also in use [34]; the two sets do not coincide, differing by terms of order gTg_{T}, and every result below refers to Eq. (2).

The expansion stops at the quadrupole. At leading twist the chiral-odd operator carries L3=±1L_{3}=\pm1; combined with the rank-two tensor built from the momentum transfer, the quadrupole is the highest structure that can be formed. The chiral-even operator has L3=0L_{3}=0 and cannot reach the quadrupole at all, so this term is peculiar to the chiral-odd sector.

Light-cone QCD sum rules

The starting point is

Πμν(p,q)=i∫d4x eiqx⟨0∣T{JN(0) qˉσμνq(x)}∣N(p)⟩.(3)\Pi_{\mu\nu}(p,q)=i\int d^{4}x\,e^{iqx}\langle0|T\{J_{N}(0)\,\bar{q}\sigma_{\mu\nu}q(x)\}|N(p)\rangle. \tag*{(3)}

The interpolating field is not unique: the general local three-quark operator with the quantum numbers of the nucleon carries a free mixing parameter, and the resulting form factors depend on it. We fix it to the Ioffe value throughout, so that

JN=2ϵabc[(uaTCdb)γ5uc−(uaTCγ5db)uc]=ϵabc(uaTCγλub)γ5γλdc.(4)J_{N}=2\epsilon^{abc}\left[(u^{aT}C d^{b})\gamma_{5}u^{c}-(u^{aT}C\gamma_{5}d^{b})u^{c}\right]=\epsilon^{abc}(u^{aT}C\gamma_{\lambda}u^{b})\gamma_{5}\gamma^{\lambda}d^{c}. \tag*{(4)}

Three reasons. First, this is the choice for which the coupling of the current to the nucleon ground state is largest relative to the continuum in the two-point function, the criterion originally proposed in Ref. [57]. Second, it is the current adopted in the analysis from which the distribution amplitudes used here are taken [58], so that the current and its nonperturbative input are matched. Third, and specific to the present calculation, only at this value can the current be written with the like-flavour pair adjacent, as in the second form of Eq. (4), and it is in that form that the selection rule derived below becomes manifest.

The two earlier light-cone QCD sum rule studies of the tensor form factors work at other values of the mixing parameter. Writing it as β=tan⁡θ\beta=\tan\theta, the Ioffe current corresponds to cos⁡θ≃0.71\cos\theta\simeq0.71, whereas Ref. [42] reports stability near cos⁡θ≃0\cos\theta\simeq0 and Ref. [43] works in the window −0.5≤cos⁡θ≤0.3-0.5\leq\cos\theta\leq0.3.

In the second writing of Eq. (4) the two uu Dirac indices sit on the same object and the distribution amplitude carries them together, so the chain closes and each contraction is a Dirac trace multiplying an open string. The tensor current can be contracted with either of the two uu quarks of the diquark uaTCγλubu^{aT}C\gamma_{\lambda}u^{b}, which gives the contractions labelled (A) and (B) below, or with the dd quark, which gives (C); in (B) the contracted field is the transposed one, which is why the conjugated propagator S^u\widehat{S}_u appears there. Using CΓTC−1=ηΓΓC\Gamma^{T}C^{-1}=\eta_{\Gamma}\Gamma, C2=−1C^2=-1 and CσμνTC=σμνS^C\sigma_{\mu\nu}^{T}C=\sigma_{\mu\nu}\widehat{S} with S^q=CSqTC−1\widehat{S}_q=CS_q^{T}C^{-1}, every factor of CC cancels and

(A)=ηXTr⁡[Γ~XγλSuσμν][γ5γλΓ3X](5)(A)=\eta_X\operatorname{Tr}\left[\widetilde{\Gamma}_X\gamma_\lambda S_u\sigma_{\mu\nu}\right]\left[\gamma_5\gamma^\lambda\Gamma_3^X\right] \tag*{(5)}
(B)=Tr⁡[σμνS^uγλΓ~X][γ5γλΓ3X](6)(B)=\operatorname{Tr}\left[\sigma_{\mu\nu}\widehat{S}_u\gamma_\lambda\widetilde{\Gamma}_X\right]\left[\gamma_5\gamma^\lambda\Gamma_3^X\right] \tag*{(6)}
(C)=ηXTr⁡[γλΓ~X][γ5γλSdσμνΓ3X](7)(C)=\eta_X\operatorname{Tr}\left[\gamma_\lambda\widetilde{\Gamma}_X\right]\left[\gamma_5\gamma^\lambda S_d\sigma_{\mu\nu}\Gamma_3^X\right] \tag*{(7)}

where Γ~X\widetilde{\Gamma}_X is the diquark Dirac structure of the amplitude XX, Γ3X\Gamma_3^X the structure carrying the third quark index, and ηX=+1\eta_X=+1 for the SS, PP and AA amplitudes and −1-1 for VV and TT. In S^q\widehat{S}_q the x̸\not{x} term of the propagator flips sign while the mass term does not, so the two must be kept as separate objects once mq≠0m_q\ne0.

The dd-quark contraction (C) contains Tr⁡[γλΓ~X]\operatorname{Tr}[\gamma_\lambda\widetilde{\Gamma}_X], four times the vector component of Γ~X\widetilde{\Gamma}_X, which vanishes unless Γ~X\widetilde{\Gamma}_X is p̸\not{p}, γκ\gamma_\kappa or x̸\not{x}. Only the six VV amplitudes therefore contribute to the dd-quark channel; SS, PP, AA and TT drop out. This selection rule is used in Secs. IV A and IV B.

The light-quark propagator is kept with its mass term,

Sq(x)=ix̸2π2x4−mq4π2x2+⋯(8)S_q(x)=\frac{i\not{x}}{2\pi^2x^4}-\frac{m_q}{4\pi^2x^2}+\cdots \tag*{(8)}

the condensate pieces being removed by the Borel transformation and the gluon pieces feeding four- and five-particle amplitudes, which are small [58, 59] and are neglected. The nonperturbative input is the three-quark matrix element, whose decomposition into distribution amplitudes of definite twist we take unchanged from Refs. [58, 60]; we follow their notation throughout.

On the hadronic side the nucleon contribution is isolated by inserting a complete set of states and keeping the ground-state pole,

Πμν=λN̸p′+mNmN2−p′2Γμν(q)u(p)+continuum,p′=p−q(9)\Pi_{\mu\nu}=\lambda_N\frac{\not{p}^{\prime}+m_N}{m_N^2-p^{\prime2}}\Gamma_{\mu\nu}(q)u(p)+\text{continuum},\qquad p^{\prime}=p-q \tag*{(9)}

with ⟨0∣JN∣N(p′)⟩=λNu(p′)\langle0|J_N|N(p^{\prime})\rangle=\lambda_Nu(p^{\prime}) and Γμν\Gamma_{\mu\nu} the decomposition of Eq. (1).

The two sides are now matched structure by structure, and this is where a choice has to be made that is usually left implicit. The Dirac strings generated by the operator product are not independent as they stand; the Clifford algebra relates them, for instance

q̸σμν=σμνq̸+2i(qμγν−qνγμ)(10)\not{q}\sigma_{\mu\nu}=\sigma_{\mu\nu}\not{q}+2i(q_\mu\gamma_\nu-q_\nu\gamma_\mu) \tag*{(10)}

so that q̸σμν\not{q}\sigma_{\mu\nu}, σμνq̸\sigma_{\mu\nu}\not{q} and qμγν−qνγμq_\mu\gamma_\nu-q_\nu\gamma_\mu cannot be treated as three separate objects. The coefficient one assigns to any one of them therefore depends on how the expression happens to have been written, not only on its content, and the same ambiguity affects σμν\sigma_{\mu\nu}, qμγνq_\mu\gamma_\nu, γμqν\gamma_\mu q_\nu and gμνg_{\mu\nu}. A projection performed before the strings are brought to a common arrangement is thus not well defined, whichever side of the sum rule it is applied to.

We remove the ambiguity by reducing every string to the canonical arrangement

γμγνq̸p̸(11)\gamma_\mu\gamma_\nu\not{q}\not{p} \tag*{(11)}

through repeated use of

…yz…⟶2(y⋅z)…1…−…zy…,y≠γ5(12)\ldots yz\ldots\longrightarrow2(y\cdot z)\ldots1\ldots-\ldots zy\ldots,\qquad y\ne\gamma_5 \tag*{(12)}

applied so as to move q̸\not{q}, then γν\gamma_\nu, then γμ\gamma_\mu successively to the left, each substitution generating a scalar product and a reordered string, and eliminating p̸\not{p} through p̸u(p)=mNu(p)\not{p}u(p)=m_Nu(p). What remains is a set of eight mutually irreducible antisymmetric structures with scalar coefficients built from qμq_\mu, qνq_\nu, pμp_\mu, pνp_\nu and gμνg_{\mu\nu}. Applying the same reduction to the right-hand side of Eq. (9) gives

Πμν=λNmN2−p2{HT[2(qμγν−qνγμ)−2(pμγν−pνγμ)+2mNiσμν−iσμνq̸]+ET2mN[(qνγμ−qμγν)q̸−2(qμpν−qνpμ)]+H~T[−2mN(qμpν−qνpμ)+1mN2(qμpν−qνpμ)q̸]}u(p).(13)\begin{aligned} \Pi_{\mu\nu}={}&\frac{\lambda_N}{m_N^2-p^2}\Bigg\{H_T\left[2(q_\mu\gamma_\nu-q_\nu\gamma_\mu)-2(p_\mu\gamma_\nu-p_\nu\gamma_\mu)+2m_Ni\sigma_{\mu\nu}-i\sigma_{\mu\nu}\not{q}\right]\\ &+\frac{E_T}{2m_N}\left[(q_\nu\gamma_\mu-q_\mu\gamma_\nu)\not{q}-2(q_\mu p_\nu-q_\nu p_\mu)\right]\\ &+\widetilde{H}_T\left[-\frac{2}{m_N}(q_\mu p_\nu-q_\nu p_\mu)+\frac{1}{m_N^2}(q_\mu p_\nu-q_\nu p_\mu)\not{q}\right]\Bigg\}u(p). \tag*{(13)} \end{aligned}

Of the eight structures, (pμγν−pνγμ)q̸(p_\mu\gamma_\nu-p_\nu\gamma_\mu)\not q does not appear at all and qμpν−qνpμq_\mu p_\nu-q_\nu p_\mu mixes, receiving −(ET+2H~T)/mN=−κT/mN-(E_T+2\widetilde{H}_T)/m_N=-\kappa_T/m_N and so giving access to the dipole moment rather than to either form factor separately. The other six are clean, but four of them return HTH_T; the three that give one form factor each are

qμγν−qνγμ:2HT,(qνγμ−qμγν)q̸:ET2mN,(qμpν−qνpμ)q̸:H~TmN2.(14)\begin{aligned} q_\mu\gamma_\nu-q_\nu\gamma_\mu&: 2H_T,\\ (q_\nu\gamma_\mu-q_\mu\gamma_\nu)\not q &: \frac{E_T}{2m_N},\\ (q_\mu p_\nu-q_\nu p_\mu)\not q &: \frac{\widetilde{H}_T}{m_N^2}. \tag*{(14)} \end{aligned}

and these are the structures we project onto. Of the four that return HTH_T alone we read it from qμγν−qνγμq_\mu\gamma_\nu-q_\nu\gamma_\mu, which carries the lowest Dirac structure of the four; the remaining three are available as a consistency check.

Earlier light-cone QCD sum rule work extracts HTH_T from q̸ σμν\not q\,\sigma_{\mu\nu} [42, 43], which by Eq. (10) is not one of the eight. We do not claim this leads to an incorrect result — both pieces on the right of Eq. (10) happen to carry HTH_T alone — but the assignment of a QCD-side term to that structure is not unique before ordering, and the basis used here is therefore different from theirs, independently of the interpolating current or the fit range. This is one of the differences to bear in mind when the results are compared in Sec. IV.

Equating the two representations, Borel transforming in p′2p^{\prime2} and subtracting the continuum with the standard substitution rules [58] then gives

HTq(Q2)=12λNemN2/M2BM2,s0[ΠHTq],ETq(Q2)=2mNλNemN2/M2BM2,s0[ΠETq],H~Tq(Q2)=mN2λNemN2/M2BM2,s0[ΠH~Tq].(15)\begin{aligned} H_T^q(Q^2)&=\frac{1}{2\lambda_N}e^{m_N^2/M^2}\mathcal{B}_{M^2,s_0}[\Pi_{H_T}^q],\\ E_T^q(Q^2)&=\frac{2m_N}{\lambda_N}e^{m_N^2/M^2}\mathcal{B}_{M^2,s_0}[\Pi_{E_T}^q],\\ \widetilde{H}_T^q(Q^2)&=\frac{m_N^2}{\lambda_N}e^{m_N^2/M^2}\mathcal{B}_{M^2,s_0}[\Pi_{\widetilde{H}_T}^q]. \tag*{(15)} \end{aligned}

where BM2,s0\mathcal{B}_{M^2,s_0} denotes the Borel transformation together with the continuum subtraction and ΠFq\Pi_F^q is the coefficient of the corresponding structure of Eq. (14) on the QCD side, given in Appendix A. The prefactors follow directly from Eq. (14) and are independent of Q2Q^2; every sign in Eq. (15) follows from the convention p′=p−qp^\prime=p-q, and with the opposite choice the three would be reversed.

Numerical analysis

Input parameters

We use mN=0.94m_N=0.94 GeV, mu=2.16−0.26+0.49m_u=2.16^{+0.49}_{-0.26} MeV and md=4.67−0.17+0.48m_d=4.67^{+0.48}_{-0.17} MeV [61], which enter the expressions of Appendix A through the ratios rq=mq/mNr_q=m_q/m_N. The residue λN\lambda_N is taken from the two-point sum rule [42].

Unlike the vector and axial-vector currents the tensor current is renormalisation-scale dependent already at one loop. The distribution amplitudes are normalised at μ2=1\mu^2=1 GeV2^2 and all results are quoted at that scale; results obtained elsewhere are evolved using [6]

X(μ2)=(αs(μ2)αs(μi2))4/27[1−337468π(αs(μi2)−αs(μ2))]X(μi2).(16)X(\mu^2)=\left(\frac{\alpha_s(\mu^2)}{\alpha_s(\mu_i^2)}\right)^{4/27}\left[1-\frac{337}{468\pi}\left(\alpha_s(\mu_i^2)-\alpha_s(\mu^2)\right)\right]X(\mu_i^2). \tag*{(16)}

with

αs(μ2)=4π9ln⁡(μ2/ΛQCD2)[1−6481ln⁡ln⁡(μ2/ΛQCD2)ln⁡(μ2/ΛQCD2)].(17)\alpha_s(\mu^2)=\frac{4\pi}{9\ln(\mu^2/\Lambda_{\mathrm{QCD}}^2)}\left[1-\frac{64}{81}\frac{\ln\ln(\mu^2/\Lambda_{\mathrm{QCD}}^2)}{\ln(\mu^2/\Lambda_{\mathrm{QCD}}^2)}\right]. \tag*{(17)}

ΛQCD=0.25\Lambda_{\mathrm{QCD}}=0.25 GeV.

We use two distribution-amplitude parameter sets, listed in Table I: Set I, in which all eight nonperturbative parameters are estimated from QCD sum rules, and Set II, the asymptotic set in which the higher conformal-spin contributions are required to vanish [58, 62]. The spread between them is carried in the error budget as the distribution-amplitude systematic.

ParameterSet-ISet-II
fNf_N(5.0±0.5)×10−3 GeV2(5.0 \pm0.5) \times10^{-3}\ \mathrm{GeV}^{2}
λ1\lambda_1(−2.7±0.9)×10−2 GeV2(-2.7 \pm0.9) \times10^{-2}\ \mathrm{GeV}^{2}
λ2\lambda_2(5.4±1.9)×10−2 GeV2(5.4 \pm1.9) \times10^{-2}\ \mathrm{GeV}^{2}
A1uA_1^u0.38±0.150.38 \pm0.1500
V1dV_1^d0.23±0.030.23 \pm0.031/31/3
f1df_1^d0.40±0.050.40 \pm0.051/31/3
f2df_2^d0.22±0.050.22 \pm0.054/154/15
f1uf_1^u0.07±0.050.07 \pm0.051/101/10

Table I. Nonperturbative parameters entering the nucleon DAs [58]. The normalisation constants fNf_N, λ1\lambda_1 and λ2\lambda_2 are common to the two sets.

The three amplitudes V1V_1, A1A_1 and T1T_1 carry additive x2x^2 mass-correction partners of twist five, V1→V1+x2mN24V1MV_1\to V_1+\frac{x^2m_N^2}{4}V_1^M and likewise for A1A_1 and T1T_1, with the same Dirac structures [58]. These are included throughout. Switching them off changes nothing in ETE_T and H~T\widetilde{H}_T, which do not receive contributions from the Lorentz structures these amplitudes reach, and changes HTH_T by 0.2%. They are therefore not a source of uncertainty here.

Auxiliary parameters

The results depend on the Borel mass M2M^2 and the continuum threshold s0s_0, and must be insensitive to both. The working regions are fixed before any form factor is extracted, so that the criteria do not depend on the values they produce. The upper limit of M2M^2 follows from requiring convergence of the light-cone expansion, that is, that the highest-twist terms retained remain a small fraction of the total; the lower limit follows from requiring that the highest power of 1/M21/M^2 does not dominate, which is equivalent to demanding a sufficiently large pole contribution. The threshold is set from the mass of the first excitation and varied within a physically motivated interval. This gives

1.0 GeV2≤M2≤2.0 GeV2,2.0 GeV2≤s0≤2.5 GeV2.(18)1.0\ \mathrm{GeV}^{2} \le M^{2} \le2.0\ \mathrm{GeV}^{2}, \qquad2.0\ \mathrm{GeV}^{2} \le s_{0} \le2.5\ \mathrm{GeV}^{2}. \tag*{(18)}

Across this window the pole contribution, defined as the ratio of the ground-state term to the total, lies between 0.55 and 0.74, while the highest-twist terms retained in the expansion enter below the 0.2% level, so both criteria are met comfortably. The dependence on the Borel mass is shown in Figure 1–Figure 2. Within the window the results are stable to the point that the M2M^2 variation is negligible against the other sources, and we do not carry it in the error budget; the uncertainty is dominated by the s0s_0 variation and by the spread between the two distribution-amplitude sets.

Figure 1. Dependence of the isoscalar chiral-odd form factors on the Borel mass, shown over M2∈[1.0,4.0]M^2 \in[1.0,4.0] GeV2^2 so that the approach to the plateau is visible; the working window of Eq. (18) is the subinterval [1.0,2.0][1.0,2.0] GeV2^2. The curves are at Q2=1Q^2=1 GeV2^2 and for the three values of the continuum threshold s0=2.0s_0=2.0, 2.25 and 2.5 GeV2^2. Rows, from top to bottom: HTH_T, ETE_T and H~T\widetilde{H}_T. Left column: Set I distribution amplitudes; right column: Set II. The flatness across the working window is what justifies neglecting the Borel-mass systematic in the error budget; the spread between the three curves is the s0s_0 uncertainty quoted in Table II.

Figure 2. Same as Figure 1, but for the isovector chiral-odd form factors.

Extraction of the moments

The sum rules are reliable for Q2≳1 GeV2Q^2 \gtrsim1\ \mathrm{GeV}^{2}, so the moments at Q2=0Q^2 = 0 are obtained by extrapolation with

F(Q2)=f0(1+Q2/m2)p,(19)F(Q^{2}) = \frac{f_{0}}{\left(1+Q^{2}/m^{2}\right)^{p}}, \tag*{(19)}

in which the exponent is left free rather than fixed to a monopole or dipole form, so that the mass scale and the fall-off are determined separately by the computed points rather than tied to one another by the choice of ansatz. The fit range is fixed by the region in which the sum rules are valid and is not varied: narrowing it discards valid data rather than probing a systematic, and the sensitivity it would expose is already carried by the s0s_0 variation.

The extrapolation is performed separately at each of the three continuum thresholds, and the spread of the resulting f0f_0 is quoted as the uncertainty. Treating that spread instead as a point-by-point error would understate it, since the three curves shift coherently rather than fluctuating independently. The distribution-amplitude systematic is not folded into that number but displayed by quoting both parameter sets throughout, and the M2M^2 dependence is negligible as noted above.

The dipole moment requires a choice that the other two do not. Both GTG_T and QTQ_T are a single form factor times a constant, so forming the combination before or after the extrapolation is the same operation. For κT\kappa_T it is not: one may extrapolate ETE_T and H~T\widetilde{H}_T separately and add the static limits, or form ET+2H~TE_T+2\widetilde{H}_T pointwise in Q2Q^2 and extrapolate that. The two target the same number at Q2=0Q^2 = 0 and differ only in which object the multipole ansatz is applied to. We take the second. The combination is not an arbitrary sum: as Eq. (13) shows, it is the coefficient of a single structure of the ordered basis, qμpν−qνpμq_{\mu}p_{\nu}-q_{\nu}p_{\mu}, so within this calculation it is as much a form factor of the correlator as ETE_T and H~T\widetilde{H}_T are individually. A further practical gain is that the s0s_0 dependence cancels between the two before the fit rather than after it, which roughly halves the uncertainty.

The two routes agree. Summing the static limits of Table II instead gives κTu=4.28±1.19\kappa_T^u=4.28 \pm1.19 and κTd=1.84±0.58\kappa_T^d=1.84 \pm0.58 for Set I and κTu=3.62±1.02\kappa_T^u=3.62 \pm1.02 and κTd=1.51±0.46\kappa_T^d=1.51 \pm0.46 for Set II, the uncertainties of ET(0)E_T(0) and 2H~T(0)2\widetilde{H}_T(0) being added in quadrature. These uncertainties are about twice those of the combination fit because the quadrature sum treats the two static limits as independent, whereas their s0s_0 variations are correlated and partly cancel when ET+2H~TE_T+2\widetilde{H}_T is formed pointwise. Compared with the values of Table III, the differences are about 1σ1\sigma and 0.4σ0.4\sigma for Set I and 0.6σ0.6\sigma and 0.3σ0.3\sigma for Set II. The agreement is closest in the dd channel, where the fitted combination returns κTd=−ETd\kappa_T^d=-E_T^d to four digits in both sets, exactly as Eq. (25) requires; this confirms the identity of Eq. (24) at the level of the Q2Q^2 dependence rather than of the moment alone.

uuddu+du+du−du-d
Set I HT(0)H_T(0)0.83±0.060.83 \pm0.06−0.44±0.03-0.44 \pm0.030.39±0.070.39 \pm0.071.27±0.071.27 \pm0.07
ET(0)E_T(0)5.28±1.085.28 \pm1.08−1.56±0.29-1.56 \pm0.293.72±1.123.72 \pm1.126.84±1.126.84 \pm1.12
H~T(0)\widetilde{H}_T(0)−0.50±0.25-0.50 \pm0.251.70±0.251.70 \pm0.251.20±0.261.20 \pm0.26−2.20±0.42-2.20 \pm0.42
Set II HT(0)H_T(0)0.73±0.040.73 \pm0.04−0.35±0.04-0.35 \pm0.040.38±0.060.38 \pm0.061.08±0.061.08 \pm0.06
ET(0)E_T(0)4.94±0.944.94 \pm0.94−1.37±0.23-1.37 \pm0.233.57±0.973.57 \pm0.976.31±0.976.31 \pm0.97
H~T(0)\widetilde{H}_T(0)−0.66±0.20-0.66 \pm0.201.44±0.201.44 \pm0.200.78±0.150.78 \pm0.15−2.10±0.36-2.10 \pm0.36

Table II. Chiral-odd form factors at Q2=0Q^2=0 and μ2=1\mu^2=1 GeV2^2.

uuddu+du+du−du-d
Set I gTg_T0.83±0.060.83 \pm0.06−0.44±0.03-0.44 \pm0.030.39±0.070.39 \pm0.071.27±0.071.27 \pm0.07
κT\kappa_T2.91±0.542.91 \pm0.541.56±0.301.56 \pm0.304.47±0.624.47 \pm0.621.35±0.621.35 \pm0.62
QTQ_T−1.00±0.49-1.00 \pm0.493.40±0.493.40 \pm0.492.41±0.532.41 \pm0.53−4.40±0.83-4.40 \pm0.83
Set II gTg_T0.73±0.040.73 \pm0.04−0.35±0.04-0.35 \pm0.040.38±0.060.38 \pm0.061.08±0.061.08 \pm0.06
κT\kappa_T2.95±0.542.95 \pm0.541.37±0.231.37 \pm0.234.32±0.594.32 \pm0.591.58±0.591.58 \pm0.59
QTQ_T−1.32±0.39-1.32 \pm0.392.88±0.392.88 \pm0.391.56±0.301.56 \pm0.30−4.20±0.73-4.20 \pm0.73

Table III. Tensor multipole moments, Eq. (2). The κT\kappa_T entries come from the combination fit of Sec. III C, the QTQ_T entries from 2H~T(0)2\widetilde{H}_T(0) of Table II; the two therefore differ in the dd channel by the reconstruction systematic discussed in the text.

The flavour components are obtained directly from their own sum rules and the isospin combinations are constructed from them, for all three moments alike. Fitting the isospin combinations directly instead gives κTu−d=1.31(23)\kappa_T^{u-d}=1.31(23) and κTu+d=4.50(84)\kappa_T^{u+d}=4.50(84) for Set I and 1.46(27)1.46(27) and 4.38(78)4.38(78) for Set II, which reproduce the entries of Table III to better than 0.2σ0.2\sigma in all four cases.

An entry is quoted as a number only if the fitted shape parameter is stable against s0s_0 and lies in a physically sensible range and the total uncertainty is small compared with the central value; entries failing these conditions are reported together with the reason rather than omitted.

Results and discussion

Figure 3 shows the Q2Q^2 dependence of the three form factors over the region in which the sum rules apply. These curves are the direct output of the calculation, and the two structural results of Secs. IV A and IV B are properties of the sum rules themselves, holding at every Q2Q^2 in that region; the values at Q2=0Q^2 = 0, which Table II collects for both distribution-amplitude sets, and the moments built from them rely in addition on the extrapolation of Sec. III C. For HTH_T and ETE_T the flavour entries are the direct sum-rule fits and the isospin combinations are derived from them; for H~T\widetilde{H}_T the reverse, because the direct H~Tu\widetilde{H}_T^u fit does not meet the quotability conditions of Sec. III C while both isospin combinations do.

Figure 3. Q2Q^2 dependence of the chiral-odd form factors of the nucleon over the interval Q2∈[1,10]Q^2 \in[1, 10] GeV2^2, the region in which the sum rules are reliable. Rows, from top to bottom: HTH_T, ETE_T and H~T\widetilde{H}_T. Left column: Set I distribution amplitudes; right column: Set II. Each panel shows the uu- and dd-quark contributions together with the isoscalar and isovector combinations u+du+d and u−du-d, with bands from the s0s_0 variation.

That reversal makes the dd-quark entry a check on the extrapolation itself. Equation (24) makes H~Td=−ETd\widetilde{H}_T^d=-E_T^d exact on the QCD side, so reconstructing H~Td\widetilde{H}_T^d from the two isospin fits ought to return 1.56 and 1.37; it returns 1.70 and 1.44, a departure of 9% and 5%. This is the price of extrapolating the two isospin combinations independently rather than the dd channel itself, and it sets the size of the extraction systematic carried by every H~T\widetilde{H}_T and QTQ_T entry in Tables II and III. It is also why QTdQ_T^d in Table III is 3.40 rather than the 2κTd=3.122\kappa_T^d=3.12 that the identity would give: κTd\kappa_T^d comes from the combination fit of Sec. III C and therefore respects the identity exactly, while QTd=2H~Td(0)Q_T^d=2\widetilde{H}_T^d(0) inherits the reconstruction systematic just quoted.

Table IV places the results next to the other determinations, all at a common scale. For the isovector tensor charge most of the other determinations cluster tightly. Four independent lattice calculations give 1.00–1.11 once brought to a common scale, the external-field QCD sum rule of Ref. [41] gives 1.00(71) at the scale we work at, and the quark-model average of Ref. [35] gives 1.20(24). The earlier light-cone QCD sum rule is the exception: it spans 1.15–2.11 across the three fit regions quoted there, a factor of two, which already indicates how sensitive the extrapolation to Q2=0Q^2 = 0 is in this channel. Set II, at 1.08(6), falls inside the lattice band, and Set I, at 1.27(7), lies just above it while agreeing with the quark-model average; both sit in the lower part of the earlier sum-rule range. The global fit of Ref. [23] gives 0.56(13) from data alone and 0.99(2) once lattice input is imposed as a prior, so the phenomenological determination is the one entry lying clearly below the lattice band; that tension is between data and lattice rather than with the present work.

The isoscalar tensor charge is where we are clearly low. Our 0.38–0.39 is to be compared with 0.59–0.61 from two lattice calculations and 0.55(2) from the lattice-constrained global fit, three determinations that agree among themselves; the same shows up in the scale-independent ratio ∣δd/δu∣\lvert\delta d/\delta u\rvert, for which we obtain 0.53 and 0.48 against 0.30 in the chiral quark-soliton model, 0.25 in the bag model and 0.28 in the lattice-constrained global fit. These are one statement, not two, and the reason is structural: as shown in Sec. IV A, the isoscalar tensor charge receives no leading-twist contribution at the level of the least suppressed integral.

The contrast between the two isovector columns is worth noting on its own. The tensor charge has been computed four times on the lattice with a total spread of about 10% and is further constrained by a global fit to data, whereas the anomalous tensor magnetic moment has been computed on the lattice once, has no phenomenological determination at all, and the theoretical values for it range over a factor of four and a half.

The isovector dipole moment shows the widest spread of any entry in the table: 0.85 in the bag model, 1.14 on the lattice, 1.17 for the quark-model average, 1.52 and 1.66 in the two soliton calculations, 1.72 in the mean-field calculation with 1/Nc1/N_c corrections, 1.82–3.86 across the three fit regions of the earlier sum-rule work, and 1.35–1.58 here. Our values sit in the middle of that spread, between the lattice and quark-model determinations on the one side and the soliton results on the other. The reason the spread is so wide is structural rather than numerical: as discussed in Sec. IV C, κTu−d\kappa_T^{u-d} is the residual of a cancellation between two much larger quantities, enforced to two powers of NcN_c, so it inherits both their uncertainties and is intrinsically the least well-determined of the three moments in any approach. The width of the literature spread is itself evidence of that.

For the isoscalar dipole our 4.32–4.47 falls just below the two soliton values, 4.72 and 4.86, and well below the bag model. For H~Tu−d\widetilde{H}_T^{u-d} and QTu−dQ_T^{u-d} we agree in sign and in order of magnitude with the mean-field calculation while disagreeing in sign with the only other sum-rule determination. That determination also violates the large-NcN_c relation of Sec. IV C in sign rather than magnitude. The extraction basis is one difference to bear in mind here: there H~T\widetilde{H}_T and ETE_T are read from structures which, as shown in Sec. II C, are not part of the independent set once the Dirac strings are ordered. In the flavour decomposition both components of the dipole moment come out positive, κTu=2.91(54)\kappa_T^u = 2.91(54) and κTd=1.56(30)\kappa_T^d = 1.56(30) for Set I, as they do on the lattice and in the soliton models; this is the sign pattern required by the connection between the transverse distortion and the Boer–Mulders function [52].

The comparison with the soliton calculations also raises the role of sea quarks. In the chiral quark-soliton model the matrix elements are the sum of a contribution from the discrete valence level and one from the polarised Dirac continuum, and the latter has been found to be important for the tensor dipole structure [34]. Being odd under charge conjugation, the tensor current measures the difference between quark and antiquark contributions, so that quark–antiquark pairs from the sea enter only to the extent that the quark and the antiquark differ in transverse polarisation or in transverse position. The separation of the soliton model into valence level and Dirac continuum is moreover not the same as the partonic separation into valence and sea quarks, the continuum contributing to antiquark as well as to quark distributions. On the side of the method, the nucleon enters the sum rules only through its three-quark distribution amplitudes up to twist six, and Fock components with an additional quark–antiquark pair are not included, so a valence–sea decomposition analogous to that of the soliton model has no counterpart here. Whether the residual difference between our dipole moments and the soliton values reflects such components, the different treatment of NcN_c, or the truncation of the sum rule cannot be decided from the present comparison.

Finally, the quadrupole columns are sparsely populated. Apart from the present work, only the leading-order mean-field calculation in Table IV reports QTQ_T as such; the isovector entry of the earlier sum-rule work is our construction from the H~Tu−d(0)\widetilde{H}_T^{u-d}(0) quoted there, and differs from ours in sign, as discussed above. None of the determinations compared gives the isoscalar quadrupole.

μ2\mu^2gTu−dg_T^{u-d}gTu+dg_T^{u+d}κTu−d\kappa_T^{u-d}κTu+d\kappa_T^{u+d}QTu−dQ_T^{u-d}QTu+dQ_T^{u+d}
This work, Set I1.01.01.27(7)1.27(7)0.39(7)0.39(7)1.35(62)1.35(62)4.47(62)4.47(62)−4.40(83)-4.40(83)2.41(53)2.41(53)
This work, Set II1.01.01.08(6)1.08(6)0.38(6)0.38(6)1.58(59)1.58(59)4.32(59)4.32(59)−4.20(73)-4.20(73)1.56(30)1.56(30)
LCSR [42]1.01.01.151.15–2.112.11⋯\cdots1.821.82–3.863.86⋯\cdots0.860.86–1.941.94⋯\cdots
QCD sum rules [41]1.01.01.00(71)1.00(71)1.00(71)1.00(71)⋯\cdots⋯\cdots⋯\cdots⋯\cdots
Lattice QCD [27, 28, 29, 30]4.04.01.001.00–1.111.110.590.59–0.610.611.14(10)1.14(10)⋯\cdots⋯\cdots⋯\cdots
Global fit, no lattice [23]4.04.00.56(13)0.56(13)0.52(16)0.52(16)⋯\cdots⋯\cdots⋯\cdots⋯\cdots
Global fit, with lattice [23]4.04.00.99(2)0.99(2)0.55(2)0.55(2)⋯\cdots⋯\cdots⋯\cdots⋯\cdots
χ\chiQSM SU(3) [31, 32]0.360.361.231.230.670.671.521.524.724.72⋯\cdots⋯\cdots
χ\chiQSM SU(2) [32]0.360.361.251.250.660.661.661.664.864.86⋯\cdots⋯\cdots
χ\chiQSM, LO [33]0.350.350.860.86⋯\cdots⋯\cdots6.636.63−6.11-6.11⋯\cdots
χ\chiQSM, 1/Nc1/N_c [34]0.350.35⋯\cdots0.710.711.721.72⋯\cdots⋯\cdots⋯\cdots
MIT bag model [35]0.490.491.281.280.760.760.850.855.615.61⋯\cdots⋯\cdots
Quark-model average [35]4.04.01.20(24)1.20(24)⋯\cdots1.17(31)1.17(31)⋯\cdots⋯\cdots⋯\cdots

Table IV. Comparison with other determinations. Every entry, including its renormalisation scale, is taken from the original reference rather than from a compilation. The column μ2\mu^2 gives that scale in GeV2^2; every row is then brought to μ2=1\mu^2=1 GeV2^2 with Eq. (16), errors scaling with the same factor. The factors are 0.871 from μ2=0.35\mu^2=0.35, 0.876 from 0.36, 0.925 from 0.49 and 1.085 from 4.0. Lattice results are in the MS scheme at 2 GeV. For Refs. [33, 34] the scale is not displayed in the expressions but is fixed by the inverse average instanton size, ρˉ−1=0.59\bar{\rho}^{-1}=0.59 GeV; for Ref. [35] the initial scale is μ0=0.7\mu_0=0.7 GeV, and the range 0.5–0.9 GeV explored there moves the evolution factor by up to 15% downward and 6% upward, so that row carries an extra normalisation uncertainty of that size. For Ref. [42] the range spans the three fit regions quoted there, Q2∈[2,10],[1.5,10]Q^2 \in[2,10], [1.5,10] and [1,10][1,10] GeV2^2. Where a reference resolves the flavours, the isospin combinations are constructed from them. Errors are in the last digits; ⋯\cdots means the quantity is not given in that reference.

One caveat on the comparison itself. Reference [34] defines the computed isovector mean-field quantity as κTu−d+gTu−d\kappa_T^{u-d} + g_T^{u-d} in its Eq. (38b), yet quotes the resulting number as κTu−d\kappa_T^{u-d}. If the tensor charge were not subtracted, the corresponding κTu−d\kappa_T^{u-d} would be smaller by gTu−d≃1g_T^{u-d} \simeq1. We quote the value as published.

Why the isoscalar tensor charge carries no leading twist

With the Ioffe current the two flavour channels draw on disjoint sets of distribution amplitudes. In the chiral limit the uu channel is fed by S1S_1, S2S_2 and T1,…,T8T_1,\ldots,T_8 alone — every VV amplitude reaches it only through the mass term of the propagator, and so carries a factor rur_u — while the dd channel, by the selection rule of Sec. II C, is fed by V1V_1, V2V_2 and V3V_3 alone. No amplitude contributes to both.

Order the contributions by the momentum-fraction integral they carry. The least suppressed is I1(2)[1]I_1^{(2)}[1] of Eq. (32), whose coefficient is

u: −2mN(S1+T1)+12mNT7,d: 2mNV1,(20)u:\ -2m_N(S_1+T_1)+12m_NT_7,\qquad d:\ 2m_NV_1, \tag*{(20)}

the dd entry carrying the isospin factor that distinguishes the two combinations. Of these, T1T_1 and V1V_1 are of twist three and S1S_1, T7T_7 of twist four. The nucleon distribution amplitudes obey [58]

2T1(x1,x2,x3)=[V1−A1](x1,x3,x2)+[V1−A1](x2,x3,x1),(21)2T_1(x_1,x_2,x_3)=[V_1-A_1](x_1,x_3,x_2)+[V_1-A_1](x_2,x_3,x_1), \tag*{(21)}

and in the asymptotic limit A1A_1 vanishes while V1V_1 is symmetric, so that T1=V1T_1=V_1. The two twist-three terms in Eq. (20) are then equal in magnitude and opposite in sign: they cancel in the isoscalar combination and add in the isovector one.

The consequence is that gTu+dg_T^{u+d} has no leading-twist contribution at the level of I1(2)I_1^{(2)} and is built from twist four and above, together with whatever residue the departure of T1T_1 from V1V_1 leaves behind, whereas gTu−dg_T^{u-d} receives twice the leading-twist piece. Both channels retain a substantial higher-twist content: switching off V2V_2 and V3V_3, which reach the dd channel only through I2(4)[α]I_2^{(4)}[\alpha], removes about two thirds of HTdH_T^d at Q2=1 GeV2Q^2=1\ \mathrm{GeV}^2, so the dd channel is by no means saturated by its leading-twist amplitude either.

Two independent observations support the picture. First, the isoscalar charge is almost unchanged between the two distribution-amplitude sets, 0.39 against 0.38, while the isovector charge moves from 1.27 to 1.08: the two sets differ in the higher conformal-spin parameters that shape the twist-three amplitudes, but share fNf_N, λ1\lambda_1 and λ2\lambda_2, so a quantity built purely from twist four and above is expected to be the stable one. Second, and more directly, rescaling the normalisation of the twist-three amplitudes moves the isovector charge while leaving the isoscalar charge unchanged at the 0.01% level, which is the cancellation of Eq. (20) seen in the output rather than in the coefficients.

This also explains why the isoscalar entry is the least reliable in Table IV. A quantity with no leading-twist contribution is precisely the one for which a sum rule evaluated at leading order in αs\alpha_s, with the light-cone expansion truncated at twist six, is weakest.

An identity in the d-quark channel

Only the six VV amplitudes reach the dd-quark channel, by the selection rule of Sec. II C, and their contributions organise themselves into the calligraphic amplitudes of Ref. [58], 2(p⋅x)V2=V1−V2−V32(p\cdot x)\mathcal{V}_2=V_1-V_2-V_3 and 4(p⋅x)2V6=−V1+V2+V3+V4+V5−V64(p\cdot x)^2\mathcal{V}_6=-V_1+V_2+V_3+V_4+V_5-V_6:

ETd=16mN(p⋅x)I2(4)[1]V2+64mNrd(p⋅x)2I3(6)[1]V6,(22)E_T^d=16m_N(p\cdot x)I_2^{(4)}[1]\mathcal{V}_2+64m_Nr_d(p\cdot x)^2I_3^{(6)}[1]\mathcal{V}_6, \tag*{(22)}
H~Td=−16mN(p⋅x)I2(4)[1]V2,(23)\widetilde{H}_T^d=-16m_N(p\cdot x)I_2^{(4)}[1]\mathcal{V}_2, \tag*{(23)}

in the notation of Eqs. (32) and (A3), the full expressions being Eqs. (38) and (A9) of Appendix A. Hence

ETd(Q2)+H~Td(Q2)=64mNrd(p⋅x)2I3(6)[1]V6.(24)E_T^d(Q^2)+\widetilde{H}_T^d(Q^2)=64m_Nr_d(p\cdot x)^2I_3^{(6)}[1]\mathcal{V}_6. \tag*{(24)}

The dd-quark contributions to ETE_T and H~T\widetilde{H}_T are exactly equal and opposite in the chiral limit. The breaking is triply suppressed — by the quark mass, by being of twist six, and by the higher power of the denominator — and is numerically a 4×10−44\times10^{-4} effect: at Q2=1 GeV2Q^2=1\ \mathrm{GeV}^2 we obtain ETd=−0.2994E_T^d=-0.2994 against H~Td=+0.2993\widetilde{H}_T^d=+0.2993, and setting md→0m_d\to0 gives ∓0.29931\mp0.29931 to every digit. It follows that

κTd=−ETd+O(md),QTd=−2ETd+O(md),(25)\kappa_T^d=-E_T^d+\mathcal{O}(m_d),\qquad Q_T^d=-2E_T^d+\mathcal{O}(m_d), \tag*{(25)}

so the dd-quark sector carries a single independent function. The identity holds to this accuracy on the QCD side, at every Q2Q^2 in the working region; what degrades it in Table III is the extrapolation, as quantified at the start of this section. Because Eq. (24) isolates V6\mathcal{V}_6 multiplied by the quark mass, it also offers a direct window on the size of the twist-six contribution in the tensor channel.

Test of the large-NcN_c relation

In the mean-field picture ETu−dE_T^{u-d} and H~Tu−d\widetilde{H}_T^{u-d} are both O(Nc3)\mathcal{O}(N_c^3) while their combination is O(Nc)\mathcal{O}(N_c), so the cancellation 2H~Tu−d=−ETu−d2\widetilde{H}_T^{u-d}=-E_T^{u-d} is enforced to two powers of NcN_c [33]. Light-cone QCD sum rules make no assumption about NcN_c, so our results provide an independent test. Using Eq. (24) the relation reduces to 2H~Tu=−ETu−ETd2\widetilde{H}_T^u=-E_T^u-E_T^d, so any deviation resides entirely in the uu channel.

We find

2H~Tu−d−ETu−d=0.64(16)(Set I),0.67(15)(Set II),(26)\frac{2\widetilde{H}_T^{u-d}}{-E_T^{u-d}}=0.64(16)\quad(\mathrm{Set\ I}),\qquad0.67(15)\quad(\mathrm{Set\ II}), \tag*{(26)}

against the value of unity that the relation predicts. The sign and the order of magnitude come out right, with the central values lying about a third below unity, and the deviation is 2.2σ2.2\sigma in both sets. The two sets agree with each other to within three per cent, so whatever is responsible is systematic rather than a fit artefact; but at 2.2σ2.2\sigma, and with the quoted uncertainties covering the s0s_0 spread alone, this is a mild tension rather than a violation. It is worth reading together with the fact that a sum rule evaluated at leading order in αs\alpha_s and truncated at twist six is itself least reliable for combinations in which a cancellation occurs, which is exactly the situation here.

Expressed through the uu channel with the help of Eq. (24), the relation requires H~Tu=−1.86(56)\widetilde{H}_T^u=-1.86(56) and −1.79(48)-1.79(48) for the two sets, against the −0.50(25)-0.50(25) and −0.66(20)-0.66(20) we obtain, a 2.2σ2.2\sigma difference in both cases; the dd channel satisfies its part exactly, by construction. Those two values are the isospin reconstruction of Table II; fixing H~Td\widetilde{H}_T^d from Eq. (24) instead would shift them to −0.64-0.64 and −0.73-0.73, well inside the quoted uncertainty, so the conclusion does not depend on which route is taken. Whatever the status of the relation, the resulting flavour asymmetry is pronounced and comfortably survives the uncertainties: ∣H~T/ET∣|\widetilde{H}_T/E_T| is unity for the dd quark, exactly, and 0.09(5)0.09(5) and 0.13(5)0.13(5) for the uu quark in the two sets, more than fifteen standard deviations away from unity. This is the direct reason the tensor quadrupole moment sits almost entirely on the dd quark, as Table III shows.

Impact-parameter shapes

The multipole moments of Table III have a direct interpretation in the impact-parameter plane. Fourier transforming the matrix element of the tensor current with respect to the transverse momentum transfer,

ρq(b⊥)=∫d2Δ⊥(2π)2e−iΔ⊥⋅b⊥12P+MFFs[iσ+j],(27)\rho_q(b_\perp)=\int\frac{d^2\Delta_\perp}{(2\pi)^2}e^{-i\boldsymbol{\Delta}_\perp\cdot\boldsymbol{b}_\perp}\frac{1}{2P^+}\mathcal{M}_{\mathrm{FFs}}[i\sigma^{+j}], \tag*{(27)}

gives the transverse distribution of quarks of a definite transverse polarisation s⊥\boldsymbol{s}_\perp in a nucleon of transverse polarisation S⊥\boldsymbol{S}_\perp. Here v±=(v0±v3)/2v^\pm=(v^0\pm v^3)/\sqrt{2} are light-cone components, j=1,2j=1,2 is a transverse index, P=(p′+p)/2P=(p'+p)/2 so that P+P^+ is the average light-cone momentum, Δ⊥\Delta_\perp is the transverse momentum transfer with t=−∣Δ⊥∣2t=-|\Delta_\perp|^2, and MFFs[iσ+j]\mathcal{M}_{\mathrm{FFs}}[i\sigma^{+j}] denotes the matrix element of Eq. (1) with the tensor indices taken as μν=+j\mu\nu=+j. Both polarisations are needed. In the two-dimensional multipole expansion of Ref. [51] the monopole and the quadrupole sit on nucleon spin operators and the dipole on the unit operator, so in an unpolarised nucleon only the dipole survives; that limit is the Burkardt distortion [52] and it carries κT\kappa_T alone. To display all three moments we therefore take the nucleon polarised as well, along the same axis as the quarks.

Contracting the transverse index jj of the operator with the quark polarisation and taking s⊥=S⊥=x^\boldsymbol{s}_\perp=\boldsymbol{S}_\perp=\hat{x}, the three structures of [51] reduce to an isotropic term, a term odd in byb_y, and a cos⁡2ϕb\cos2\phi_b term, so that

ρq(b⊥,ϕb)=ρ0(b⊥)[1+a1(b⊥)sin⁡ϕb+a2(b⊥)cos⁡2ϕb],(28)\rho_q(b_\perp,\phi_b)=\rho_0(b_\perp)\left[1+a_1(b_\perp)\sin\phi_b+a_2(b_\perp)\cos2\phi_b\right], \tag*{(28)}

with ϕb\phi_b measured from the polarisation axis. No higher harmonics occur, since the expansion stops at the quadrupole (Sec. II B). The radial profiles ρ0\rho_0, a1a_1 and a2a_2 require the form factors at all Q2Q^2 and are not determined here. Two moments of the distribution are determined, however, and they depend on the profiles only through integrals. With ⟨O⟩q=∫d2b⊥ O ρq/∫d2b⊥ ρq\langle O\rangle_q=\int d^2b_\perp\,O\,\rho_q/\int d^2b_\perp\,\rho_q, the angular integrations in Eq. (28) give

⟨by⟩q=∫0∞db⊥ b⊥2 ρ0a12∫0∞db⊥ b⊥ ρ0,⟨bx2−by2⟩q=∫0∞db⊥ b⊥3 ρ0a22∫0∞db⊥ b⊥ ρ0,(29)\langle b_y\rangle_q=\frac{\int_0^\infty db_\perp\,b_\perp^2\,\rho_0a_1}{2\int_0^\infty db_\perp\,b_\perp\,\rho_0},\qquad \langle b_x^2-b_y^2\rangle_q=\frac{\int_0^\infty db_\perp\,b_\perp^3\,\rho_0a_2}{2\int_0^\infty db_\perp\,b_\perp\,\rho_0}, \tag*{(29)}

so that the displacement is fixed by a1a_1 alone and the deformation by a2a_2 alone. The same moments follow directly from Eq. (27): weighting ρq\rho_q with byb_y or with bx2−by2b_x^2-b_y^2 and integrating over the plane amounts to differentiating the matrix element once or twice with respect to Δ⊥\boldsymbol{\Delta}_{\perp} at Δ⊥=0\boldsymbol{\Delta}_{\perp}=0, where only the form factors at t=0t=0 survive. The normalisation is set by the monopole, the first derivative picks out the dipole term and the traceless part of the second the quadrupole term, so that

⟨by⟩q=κTq2mNgTq,⟨bx2−by2⟩q=QTq2mN2gTq,(30)\langle b_y\rangle_q=\frac{\kappa_T^q}{2m_N g_T^q},\qquad\langle b_x^2-b_y^2\rangle_q=\frac{Q_T^q}{2m_N^2g_T^q}, \tag*{(30)}

independently of the profiles. The first is the displacement of the centroid perpendicular to the polarisation axis, the light-front analogue of the shift generated by the anomalous magnetic moment in the vector channel; the second is the quadrupole deformation, positive when the distribution is elongated along the polarisation axis and negative when it is elongated across it. The monopole gTg_T sets the normalisation and appears in both denominators, so it is the relative, not the absolute, size of κT\kappa_T and QTQ_T that shapes the distribution.

With the moments of Table III we obtain

Set I:⟨by⟩u,d=0.37, −0.37 fm,⟨bx2−by2⟩u,d=−0.027, −0.170 fm2,Set II:⟨by⟩u,d=0.42, −0.41 fm,⟨bx2−by2⟩u,d=−0.040, −0.181 fm2.(31)\begin{aligned} \text{Set I}:&\quad\langle b_y\rangle_{u,d}=0.37,\,-0.37\ \mathrm{fm},\qquad\langle b_x^2-b_y^2\rangle_{u,d}=-0.027,\,-0.170\ \mathrm{fm}^2,\\ \text{Set II}:&\quad\langle b_y\rangle_{u,d}=0.42,\,-0.41\ \mathrm{fm},\qquad\langle b_x^2-b_y^2\rangle_{u,d}=-0.040,\,-0.181\ \mathrm{fm}^2. \tag*{(31)} \end{aligned}

Figure 4 shows the corresponding distributions. For display we take a Gaussian isotropic profile, ρ0∝exp⁡(−b⊥2/2R2)\rho_0\propto\exp(-b_\perp^2/2R^2), and the lowest profiles compatible with regularity at the origin, a1=c1b⊥a_1=c_1b_\perp and a2=c2b⊥2a_2=c_2b_\perp^2: for the density to be smooth at b⊥=0b_\perp=0, the sin⁡ϕb\sin\phi_b and cos⁡2ϕb\cos2\phi_b terms must vanish there as b⊥b_\perp and b⊥2b_\perp^2, respectively. Equation (29) then fixes c1=⟨by⟩/R2c_1=\langle b_y\rangle/R^2 and c2=⟨bx2−by2⟩/(4R4)c_2=\langle b_x^2-b_y^2\rangle/(4R^4), so that the displayed distributions reproduce the moments of Eq. (31) exactly, whatever the value of RR. We take R=0.60 fmR=0.60\ \mathrm{fm}, of the order of the transverse size of the nucleon. The width affects the appearance of the figure, but neither the moments nor the conclusions drawn from them below, which rest on Eq. (30) alone.

Figure 4. Impact-parameter distributions of uu (left) and dd (right) quarks polarised along +x+x in a nucleon polarised along the same direction, from Eq. (28) with the moments of Eq. (31), a Gaussian profile of width R=0.60R = 0.60 fm and the profiles a1a_1 and a2a_2 described in the text. Top row: Set I; bottom row: Set II. The colour map shows the density on a common scale, the grey curves are isodensity contours, the dashed circle is the undisplaced, undeformed reference and the arrow marks ⟨by⟩\langle b_y\rangle in fm. The two flavours are displaced in opposite directions by nearly equal amounts, while both distributions are elongated across the polarisation axis, the dd quark some five to six times more strongly. The overall normalisation of each panel is set by gTqg_T^q, which is negative for the dd quark, and is not displayed; only the shapes are compared. The density is shown within ∣b⊥∣≤0.85|b_\perp| \le0.85 fm. Beyond ∣b⊥∣≃0.9|b_\perp| \simeq0.9 fm the linearly and quadratically rising profiles a1a_1 and a2a_2, which are a display choice, begin to outweigh the isotropic term, and the displayed density would no longer be positive.

The two flavours are displaced in opposite directions by nearly equal amounts. This follows from the ratios κT/gT\kappa_T/g_T, which come out 3.5 and 4.0 for the uu quark in the two sets and −3.5-3.5 and −3.9-3.9 for the dd quark: gTdg_T^d is negative while κTd\kappa_T^d is positive, and it is that sign mismatch, rather than any difference in the magnitude of the distortion, that reverses the shift. The deformations, in contrast, have the same sign for both flavours, both distributions being elongated across the polarisation axis, but they differ by a factor of five to six: QT/gTQ_T/g_T is −1.2-1.2 and −1.8-1.8 for the uu quark against −7.7-7.7 and −8.2-8.2 for the dd quark. This is the impact-parameter form of the flavour hierarchy of Sec. IV C, where ∣H~T/ET∣|\widetilde{H}_T/E_T| is unity for the dd quark and an order of magnitude smaller for the uu quark. The two distribution-amplitude sets give the same picture throughout.

Summary

The tensor current is odd under charge conjugation, so the quantities it defines measure the difference between quark and antiquark contributions, and its three chiral-odd form factors organise themselves into a monopole, a dipole and a quadrupole. We have computed all three for the nucleon within light-cone QCD sum rules, resolved by flavour, over 1≤Q2≤10 GeV21\leq Q^2\leq10\ \mathrm{GeV}^2, and assembled the corresponding multipole moments from their values at Q2=0Q^2=0. The flavour components of the quadrupole moment and the isoscalar channel had not previously been obtained in this framework.

Three results stand apart from the numbers themselves. The first is methodological. The Lorentz structures generated by the operator product are not independent until the Dirac strings are brought to a canonical order. Eight independent antisymmetric structures survive the reduction; one of them does not appear on the hadronic side at all, one mixes two form factors, and of the six that are clean, four return the tensor charge, so that exactly three suffice to give the three form factors separately. The structure from which the tensor charge has usually been extracted is not among the eight. Before ordering the structures are linearly related, so the coefficient assigned to any one of them is not fixed by the content of the expression alone; the basis used here is therefore different from that of earlier sum-rule work, independently of the interpolating current or the fit range.

The second is an identity. Only the vector distribution amplitudes reach the dd-quark channel with the Ioffe current, and their contributions to ETE_T and H~T\widetilde{H}_T are equal and opposite in the chiral limit, the breaking being proportional to the quark mass and to the twist-six amplitude V6\mathcal{V}_6. Numerically the breaking is a 4×10−44\times10^{-4} effect, and setting the quark mass to zero reproduces the cancellation to every digit. The dd-quark sector therefore carries a single independent function, κTd=−ETd\kappa_T^d=-E_T^d and QTd=−2ETdQ_T^d=-2E_T^d, on the QCD side, up to the quark mass. The extraction confirms this: fitting ET+2H~TE_T+2\widetilde{H}_T as a function of Q2Q^2 returns κTd=−ETd\kappa_T^d=-E_T^d to four digits in both distribution-amplitude sets. The 9% and 5% by which the quoted QTdQ_T^d departs from 2κTd2\kappa_T^d is not a failure of the identity but the cost of reaching H~Td\widetilde{H}_T^d through the two isospin combinations, and it sets the extraction systematic on the H~T\widetilde{H}_T and QTQ_T entries.

The third concerns the isoscalar channel. The two flavour channels draw on disjoint sets of distribution amplitudes, and their leading-twist contributions to the tensor charge are equal and opposite, so that they cancel in the isoscalar combination and add in the isovector one. The isoscalar tensor charge thus has no leading-twist contribution at the level of the least suppressed integral. That explains both why it comes out small and why it is the least reliable entry in the comparison: a quantity built essentially from twist four and above is precisely the one for which a sum rule evaluated at leading order in αs\alpha_s, with the light-cone expansion truncated at twist six, is weakest. Two independent checks support the mechanism, the stability of the isoscalar charge against the choice of distribution-amplitude set and its insensitivity to the twist-three normalisation. None of these three would have been visible without resolving the flavours: each is a statement about one channel, or about the relation between the two.

The same is true of the picture the moments produce. Carried to the impact-parameter plane, and taken with the quark and the nucleon polarised along the same transverse axis, they displace the two flavour distributions in opposite directions by nearly equal amounts, ⟨by⟩=0.37\langle b_y\rangle= 0.37 and −0.37-0.37 fm, and elongate both across that axis, the dd quark some five to six times more strongly. The opposite displacements do not come from a difference in the distortion itself but from the sign of gTg_T, which is negative for the dd quark while κT\kappa_T is positive for both; and the disparity in the deformations is the impact-parameter form of the flavour asymmetry in ∣H~T/ET∣|\widetilde{H}_T/E_T|, which is unity for the dd quark and an order of magnitude smaller for the uu quark. Both distribution-amplitude sets give the same picture.

Against the other determinations the isovector tensor charge sits inside the existing band and the isovector dipole moment in the middle of it, while the isoscalar charge lies below for the reason just given and the isoscalar dipole falls just under the two soliton values. The dipole moment shows the widest spread of any quantity in the comparison, which is itself a consequence of its being the residual of a cancellation enforced to two powers of NcN_c; it has been computed once on the lattice and has no phenomenological determination at all. Among the determinations compared, only the mean-field calculation gives the quadrupole moment as such, and none gives the isoscalar quadrupole, so this is the entry that would benefit most from an independent calculation. Tested against our results, the mean-field relation 2H~Tu−d=−ETu−d2\widetilde{H}_T^{u-d}=-E_T^{u-d} holds in sign and in order of magnitude, the ratio coming out 0.64(16) and 0.67(15) against the predicted unity, a 2.2σ2.2\sigma deviation that resides entirely in the uu channel.

QCD-side results

The invariant functions multiplying the independent structures of Eq. (14), before the Borel transformation and the continuum subtraction, are collected here. It is convenient to abbreviate the momentum-fraction integrals as

I1(n)[w]=mNn−2∫01dξw(q−ξp)n,(32)\mathcal{I}_1^{(n)}[w] = m_N^{n-2}\int_0^1 d\xi\frac{w}{(q-\xi p)^n}, \tag*{(32)}
I2(n)[w]=mNn−2∫01dξ∫0ξdαw(q−αp)n,(33)\mathcal{I}_2^{(n)}[w] = m_N^{n-2}\int_0^1 d\xi\int_0^\xi d\alpha\frac{w}{(q-\alpha p)^n}, \tag*{(33)}
I3(n)[w]=mNn−2∫01dξ∫0ξdα∫0αdβw(q−βp)n,(34)\mathcal{I}_3^{(n)}[w] = m_N^{n-2}\int_0^1 d\xi\int_0^\xi d\alpha\int_0^\alpha d\beta\frac{w}{(q-\beta p)^n}, \tag*{(34)}

the weight ww being 11, α\alpha, β\beta or β2\beta^2. The factor mNn−2m_N^{n-2} is included so that all three carry the same mass dimension; with this normalisation, and writing rq=mq/mNr_q=m_q/m_N, every term below is an integer multiple of mNm_N times one of them. The distribution amplitudes are those of Refs. [58, 60], and the quark masses enter only through the mass term of the light-quark propagator, Eq. (8). The uu-channel amplitudes are evaluated at x2x_2 and the dd-channel ones at x3x_3, as required by the slot structure of Sec. II C; the arguments are suppressed below. The isoscalar and isovector combinations are obtained by adding the two channels with a relative sign +1+1 and −1-1 respectively.

Tensor form factor HTH_T

ΠHTu=2mN{−[I1(2)[1]+I2(4)[α]]S1+I2(4)[α]S2−[I1(2)[1]+4I2(4)[α]+3I3(4)[1]]T1+I1(4)[1]T1M+[3I2(4)[α]+4I3(6)[β2]+4I3(4)[1]](T2+T5)−[2I2(4)[α]+4I3(6)[β2]+I3(4)[1]]T3−[4I3(6)[β2]+I3(4)[1]]T4−3I3(4)[1]T6+[6I1(2)[1]+10I2(4)[α]+4I3(6)[β2]+7I3(4)[1]]T7+[4I3(6)[β2]+7I3(4)[1]]T8+ru(−[I2(4)[1]+2I3(6)[β]]V1+2I3(6)[β](V2+V3−V6)+[I2(4)[1]+2I3(6)[β]](V4+V5))},(35)\Pi_{H_T}^{u}=2m_N\Bigl\{-\left[\mathcal{I}_1^{(2)}[1]+\mathcal{I}_2^{(4)}[\alpha]\right]S_1+\mathcal{I}_2^{(4)}[\alpha]S_2-\left[\mathcal{I}_1^{(2)}[1]+4\mathcal{I}_2^{(4)}[\alpha]+3\mathcal{I}_3^{(4)}[1]\right]T_1+\mathcal{I}_1^{(4)}[1]T_1^M+\left[3\mathcal{I}_2^{(4)}[\alpha]+4\mathcal{I}_3^{(6)}[\beta^2]+4\mathcal{I}_3^{(4)}[1]\right](T_2+T_5)-\left[2\mathcal{I}_2^{(4)}[\alpha]+4\mathcal{I}_3^{(6)}[\beta^2]+\mathcal{I}_3^{(4)}[1]\right]T_3-\left[4\mathcal{I}_3^{(6)}[\beta^2]+\mathcal{I}_3^{(4)}[1]\right]T_4-3\mathcal{I}_3^{(4)}[1]T_6+\left[6\mathcal{I}_1^{(2)}[1]+10\mathcal{I}_2^{(4)}[\alpha]+4\mathcal{I}_3^{(6)}[\beta^2]+7\mathcal{I}_3^{(4)}[1]\right]T_7+\left[4\mathcal{I}_3^{(6)}[\beta^2]+7\mathcal{I}_3^{(4)}[1]\right]T_8+r_u\left(-\left[\mathcal{I}_2^{(4)}[1]+2\mathcal{I}_3^{(6)}[\beta]\right]V_1+2\mathcal{I}_3^{(6)}[\beta](V_2+V_3-V_6)+\left[\mathcal{I}_2^{(4)}[1]+2\mathcal{I}_3^{(6)}[\beta]\right](V_4+V_5)\right)\Bigr\}, \tag*{(35)}
ΠHTd=2mN{[I1(2)[1]+I2(4)[α]]V1−I2(4)[α](V2+V3)−I1(4)[1]V1M+rd([I2(4)[1]+2I3(6)[β]](V1−V4−V5)−2I3(6)[β](V2+V3−V6))}.(36)\Pi_{H_T}^{d}=2m_N\Bigl\{\left[\mathcal{I}_1^{(2)}[1]+\mathcal{I}_2^{(4)}[\alpha]\right]V_1-\mathcal{I}_2^{(4)}[\alpha](V_2+V_3)-\mathcal{I}_1^{(4)}[1]V_1^M+r_d\left(\left[\mathcal{I}_2^{(4)}[1]+2\mathcal{I}_3^{(6)}[\beta]\right](V_1-V_4-V_5)-2\mathcal{I}_3^{(6)}[\beta](V_2+V_3-V_6)\right)\Bigr\}. \tag*{(36)}

Two features of Eqs. (35) and (A5) are used in Sec. IV A. The uu channel receives the VV amplitudes only through rur_u, so in the chiral limit it is fed by SS and TT alone, while the dd channel is fed by VV alone. And the coefficients of the least suppressed integral I1(2)\mathcal{I}_1^{(2)} are −(S1+T1)+6T7-(S_1+T_1)+6T_7 in the uu channel against V1V_1 in the dd channel.

Tensor form factor ETE_T

ΠETu=8mN{I2(4)[1](S1−S2+2T1+T2−3T5−4T7)+4I3(6)[β](−T2+T3+T4−T5−T7−T8)+2ruI3(6)[1](V1−V2−V3−V4−V5+V6)},(37)\Pi_{E_T}^{u}=8m_N\Bigl\{\mathcal{I}_2^{(4)}[1](S_1-S_2+2T_1+T_2-3T_5-4T_7)+4\mathcal{I}_3^{(6)}[\beta]\left(-T_2+T_3+T_4-T_5-T_7-T_8\right)+2r_u\mathcal{I}_3^{(6)}[1](V_1-V_2-V_3-V_4-V_5+V_6)\Bigr\}, \tag*{(37)}
ΠETd=8mN{I2(4)[1](V1−V2−V3)−2rdI3(6)[1](V1−V2−V3−V4−V5+V6)}.(38)\Pi_{E_T}^{d}=8m_N\Bigl\{\mathcal{I}_2^{(4)}[1](V_1-V_2-V_3)-2r_d\mathcal{I}_3^{(6)}[1](V_1-V_2-V_3-V_4-V_5+V_6)\Bigr\}. \tag*{(38)}

Tensor form factor H~T\widetilde{H}_T

ΠH~Tu=16mNℓ2(4)[1](−T1+T3+T7),(39)\Pi_{\widetilde{H}_T}^{u}=16m_N\ell_{2}^{(4)}[1]\left(-T_1+T_3+T_7\right), \tag*{(39)}
ΠH~Td=−8mNℓ2(4)[1](V1−V2−V3).(40)\Pi_{\widetilde{H}_T}^{d}=-8m_N\ell_{2}^{(4)}[1]\left(V_1-V_2-V_3\right). \tag*{(40)}

The identity of Sec. IV B can now be read off directly. Both H~Td\widetilde{H}_T^d and the mass-independent part of ETdE_T^d are proportional to V1−V2−V3=2(p⋅x)V2V_1-V_2-V_3=2(p\cdot x)V_2 with the same prefactor and opposite sign, so they cancel term by term. What survives is the mass term of ETdE_T^d, whose amplitude combination is V1−V2−V3−V4−V5+V6=−4(p⋅x)2V6V_1-V_2-V_3-V_4-V_5+V_6=-4(p\cdot x)^2V_6, giving Eq. (24).

The same combination appears in ETuE_T^u, with the opposite overall sign and with rur_u in place of rdr_d. In the isospin limit the two therefore cancel in the isoscalar combination and add in the isovector one, although the effect is numerically negligible, being both mass-suppressed and of twist six.

References

  1. [1]R. L. Jaffe, X.-D. Ji, Chiral odd parton distributions and polarized Drell-Yan, Phys. Rev. Lett. 67 (1991) 552–555. doi:10.1103/PhysRevLett.67.552.DOI
  2. [2]R. L. Jaffe, X.-D. Ji, Chiral odd parton distributions and Drell-Yan processes, Nucl. Phys. B 375 (1992) 527–560. doi:10.1016/0550-3213(92)90110-W.DOI
  3. [3]J. P. Ralston, D. E. Soper, Production of Dimuons from High-Energy Polarized Proton Proton Collisions, Nucl. Phys. B 152 (1979) 109. doi:10.1016/0550-3213(79)90082-8.DOI
  4. [4]X. Artru, M. Mekhfi, Transversely Polarized Parton Densities, their Evolution and their Measurement, Z. Phys. C 45 (1990) 669. doi:10.1007/BF01556280.DOI
  5. [5]J. L. Cortes, B. Pire, J. P. Ralston, Measuring the transverse polarization of quarks in the proton, Z. Phys. C 55 (1992) 409–416. doi:10.1007/BF01565099.DOI
  6. [6]V. Barone, A. Drago, P. G. Ratcliffe, Transverse polarisation of quarks in hadrons, Phys. Rept. 359 (2002) 1–168. arXiv:hep-ph/0104283, doi:10.1016/S0370-1573(01)00051-5.DOI
  7. [7]V. Barone, F. Bradamante, A. Martin, Transverse-spin and transverse-momentum effects in high-energy processes, Prog. Part. Nucl. Phys. 65 (2010) 267–333. arXiv:1011.0909, doi:10.1016/j.ppnp.2010.07.003.DOI
  8. [8]J. Soffer, Positivity constraints for spin dependent parton distributions, Phys. Rev. Lett. 74 (1995) 1292–1294. arXiv:hep-ph/9409254, doi:10.1103/PhysRevLett.74.1292.DOI
  9. [9]M. Pospelov, A. Ritz, Electric dipole moments as probes of new physics, Annals Phys. 318 (2005) 119–169. arXiv:hep-ph/0504231, doi:10.1016/j.aop.2005.04.002.DOI
  10. [10]T. Bhattacharya, V. Cirigliano, R. Gupta, H.-W. Lin, B. Yoon, Neutron Electric Dipole Moment and Tensor Charges from Lattice QCD, Phys. Rev. Lett. 115 (21) (2015) 212002. arXiv:1506.04196, doi:10.1103/PhysRevLett.115.212002.DOI
  11. [11]M. Pitschmann, C.-Y. Seng, C. D. Roberts, S. M. Schmidt, Nucleon tensor charges and electric dipole moments, Phys. Rev. D 91 (2015) 074004. arXiv:1411.2052, doi:10.1103/PhysRevD.91.074004.DOI
  12. [12]T. Liu, Z. Zhao, H. Gao, Experimental constraint on quark electric dipole moments, Phys. Rev. D 97 (7) (2018) 074018. arXiv:1704.00113, doi:10.1103/PhysRevD.97.074018.DOI
  13. [13]A. Courtoy, S. Baßler, M. González-Alonso, S. Lüti, Beyond-Standard-Model Tensor Interaction and Hadron Phenomenology, Phys. Rev. Lett. 115 (2015) 162001. arXiv:1503.06814, doi:10.1103/PhysRevLett.115.162001.DOI
  14. [14]T. Bhattacharya, V. Cirigliano, S. D. Cohen, A. Filipuzzi, M. Gonzalez-Alonso, M. L. Graesser, R. Gupta, H.-W. Lin, Probing Novel Scalar and Tensor Interactions from (Ultra)Cold Neutrons to the LHC, Phys. Rev. D 85 (2012) 054512. arXiv:1110.6448, doi:10.1103/PhysRevD.85.054512.DOI
  15. [15]A. Arapietian, et al., Effects of transversity in deep-inelastic scattering by polarized protons, Phys. Lett. B 693 (2010) 11–16. arXiv:1006.4221, doi:10.1016/j.physletb.2010.08.012.
  16. [16]C. Adolph, et al., Experimental investigation of transverse spin asymmetries in muon-p SIDIS processes: Collins asymmetries, Phys. Lett. B 717 (2012) 376–382. arXiv:1205.5121, doi:10.1016/j.physletb.2012.09.055.DOI
  17. [17]A. Vossen, et al., Observation of transverse polarization asymmetries of charged pion pairs in e+e− annihilation near √s = 10.58 GeV, Phys. Rev. Lett. 107 (2011) 072004. arXiv:1104.2425, doi:10.1103/PhysRevLett.107.072004.DOI
  18. [18]A. Bacchetta, A. Courtoy, M. Radici, First extraction of valence transversities in a collinear framework, JHEP 03 (2013) 119. arXiv:1212.3568, doi:10.1007/JHEP03(2013)119.DOI
  19. [19]M. Radici, A. Courtoy, A. Bacchetta, M. Guagnelli, Improved extraction of valence transversity distributions from inclusive dihadron production, JHEP 05 (2015) 123. arXiv:1503.03495, doi:10.1007/JHEP05(2015)123.DOI
  20. [20]M. Anselmino, M. Boglione, U. D’Alesio, S. Melis, F. Murgia, A. Prokudin, Simultaneous extraction of transversity and Collins functions from new SIDIS and e+e− data, Phys. Rev. D 87 (2013) 094019. arXiv:1303.3822, doi:10.1103/PhysRevD.87.094019.DOI
  21. [21]Z.-B. Kang, A. Prokudin, P. Sun, F. Yuan, Extraction of Quark Transversity Distribution and Collins Fragmentation Functions with QCD Evolution, Phys. Rev. D 93 (1) (2016) 014009. arXiv:1505.05589, doi:10.1103/PhysRevD.93.014009.DOI
  22. [22]H.-W. Lin, W. Melnitchouk, A. Prokudin, N. Sato, H. Shows, First Monte Carlo Global Analysis of Nucleon Transversity with Lattice QCD Constraints, Phys. Rev. Lett. 120 (15) (2018) 152502. arXiv:1710.09858, doi:10.1103/PhysRevLett.120.152502.DOI
  23. [23]C. Cocuzza, A. Metz, D. Pitonyak, A. Prokudin, N. Sato, R. Seidl, Transversity Distributions and Tensor Charges of the Nucleon: Extraction from DiHadron Production and Their Universal Nature, Phys. Rev. Lett. 132 (9) (2024) 091901. arXiv:2306.12998, doi:10.1103/PhysRevLett.132.091901.DOI
  24. [24]A. Accardi, et al., Electron Ion Collider: The Next QCD Frontier: Understanding the glue that binds us all, Eur. Phys. J. A 52 (9) (2016) 268. arXiv:1212.1701, doi:10.1140/epja/i2016-16268-9.
  25. [25]R. Abdul Khalek, et al., Science Requirements and Detector Concepts for the Electron-Ion Collider: EIC Yellow Report, Nucl. Phys. A 1026 (2022) 122447. arXiv:2103.05419, doi:10.1016/j.nuclphysa.2022.122447.DOI
  26. [26]M. Gockeler, P. Hagler, R. Horsley, D. Pleiter, P. E. L. Rakow, A. Schaefer, G. Schierholz, J. M. Zanotti, Quark helicity flip generalized parton distributions from two-flavor lattice QCD, Phys. Lett. B 627 (2005) 113–123. arXiv:hep-lat/0507001, doi:10.1016/j.physletb.2005.09.002.DOI
  27. [27]T. Bhattacharya, V. Cirigliano, S. Cohen, R. Gupta, A. Joseph, H.-W. Lin, B. Yoon, Iso-vector and Iso-scalar Tensor Charges of the Nucleon from Lattice QCD, Phys. Rev. D 92 (9) (2015) 094511. arXiv:1506.06411, doi:10.1103/PhysRevD.92.094511.DOI
  28. [28]R. Gupta, B. Yoon, T. Bhattacharya, V. Cirigliano, Y.-C. Jang, H.-W. Lin, Flavor diagonal tensor charges of the nucleon from (2+1+1)-flavor lattice QCD, Phys. Rev. D 98 (9) (2018) 091501. arXiv:1808.07597, doi:10.1103/PhysRevD.98.091501.DOI
  29. [29]C. Alexandrou, S. Bacchio, M. Constantinou, J. Finkenrath, K. Hadjiyiannakou, K. Jansen, G. Koutsou, A. Vaquero Aviles-Casco, Nucleon axial, tensor, and scalar charges and σ-terms in lattice QCD, Phys. Rev. D 102 (5) (2020) 054517. arXiv:1909.00485, doi:10.1103/PhysRevD.102.054517.DOI
  30. [30]C. Alexandrou, et al., Moments of the nucleon transverse quark spin densities using lattice QCD, Phys. Rev. D 107 (5) (2023) 054504. arXiv:2202.09871, doi:10.1103/PhysRevD.107.054504.DOI
  31. [31]T. Ledwig, A. Silva, H.-C. Kim, Tensor charges and form factors of SU(3) baryons in the self-consistent SU(3) chiral quark-soliton model, Phys. Rev. D 82 (2010) 034022. arXiv:1004.3612, doi:10.1103/PhysRevD.82.034022.DOI
  32. [32]T. Ledwig, A. Silva, H.-C. Kim, Anomalous tensor moments and form factors of the proton in the self-consistent chiral quark-soliton model, Phys. Rev. D 82 (2010) 054014. arXiv:1007.1355, doi:10.1103/PhysRevD.82.054014.DOI
  33. [33]N.-Y. Ghim, H.-Y. Won, J.-Y. Kim, H.-C. Kim, Nucleon tensor form factors at large Nc, Phys. Rev. D 111 (7) (2025) 074024. arXiv:2501.12241, doi:10.1103/PhysRevD.111.074024.DOI
  34. [34]N.-Y. Ghim, H.-Y. Won, J.-Y. Kim, H.-C. Kim, Multipole structure of nucleon tensor form factors, Phys. Rev. D 114 (3) (2026) 034016. arXiv:2605.24435, doi:10.1103/wtdr-92yh.DOI
  35. [35]K. Tezgin, B. Maynard, P. Schweitzer, Chiral-odd GPDs in the bag model, Phys. Rev. D 110 (5) (2024) 054028. arXiv:2404.11563, doi:10.1103/PhysRevD.110.054028.
  36. [36]I. Schmidt, J. Soffer, Melosh rotation and the nucleon tensor charge, Phys. Lett. B 407 (1997) 331–334. arXiv:hep-ph/9703411, doi:10.1016/S0370-2693(97)00737-5.DOI
  37. [37]B. Pasquini, M. Pincetti, S. Boffi, Chiral-odd generalized parton distributions in constituent quark models, Phys. Rev. D 72 (2005) 094029. arXiv:hep-ph/0510376, doi:10.1103/PhysRevD.72.094029.DOI
  38. [38]I. C. Cloet, W. Bentz, A. W. Thomas, Transversity quark distributions in a covariant quark-diquark model, Phys. Lett. B 659 (2008) 214–220. arXiv:0708.3246, doi:10.1016/j.physletb.2007.09.071.
  39. [39]T. Maji, D. Chakrabarti, Transverse structure of a proton in a light-front quark-diquark model, Phys. Rev. D 95 (7) (2017) 074009. arXiv:1702.04557, doi:10.1103/PhysRevD.95.074009.DOI
  40. [40]S.-X. Xu, C. Chen, I. C. Cloet, C. D. Roberts, J. Segovia, H.-S. Zong, Contact-interaction Faddeev equation and, inter alia, proton tensor charges, Phys. Rev. D 92 (11) (2015) 114034. arXiv:1509.03311, doi:10.1103/PhysRevD.92.114034.DOI
  41. [41]H.-x. He, X.-D. Ji, The Nucleon’s tensor charge, Phys. Rev. D 52 (1995) 2960–2963. arXiv:hep-ph/9412235, doi:10.1103/PhysRevD.52.2960.DOI
  42. [42]G. Erkol, A. Ozpineci, Tensor form factors of nucleon in QCD, Phys. Lett. B 704 (2011) 551–558. arXiv:1107.4584, doi:10.1016/j.physletb.2011.09.089.DOI
  43. [43]T. M. Aliev, K. Azizi, M. Savci, Nucleon tensor form factors induced by isovector and isoscalar currents in QCD, Phys. Rev. D 84 (2011) 076005. arXiv:1108.2019, doi:10.1103/PhysRevD.84.076005.
  44. [44]U. Özdem, Isovector and isoscalar tensor form factors of N(1535) → N transition in light-cone QCD, Phys. Rev. D 102 (1) (2020) 014001. arXiv:2004.12312, doi:10.1103/PhysRevD.102.014001.DOI
  45. [45]U. Özdem, Tensor form factors of N(1535) state via light-cone QCD, Chin. J. Phys. 72 (2021) 93–99. doi:10.1016/j.cjph.2021.04.018.DOI
  46. [46]A. kucukarslan, U. Ozdem, A. Ozpineci, Tensor form factors of the octet hyperons in QCD, Phys. Rev. D 94 (9) (2016) 094010. arXiv:1610.08358, doi:10.1103/PhysRevD.94.094010.DOI
  47. [47]Z. Asmaee, N. Hajirasouliha, K. Azizi, Tensor form factors of the Δ+ baryon induced by isovector and isoscalar currents in QCD, Phys. Rev. D 113 (5) (2026) 054002. arXiv:2511.12360, doi:10.1103/mlt1-gvnw.DOI
  48. [48]Z. Asmaee, K. Azizi, Tensor form factors of decuplet hyperons in QCD, Phys. Rev. D 113 (9) (2026) 094020. arXiv:2601.17948, doi:10.1103/zwcs-5qx3.DOI
  49. [49]M. Diehl, Generalized parton distributions with helicity flip, Eur. Phys. J. C 19 (2001) 485–492. arXiv:hep-ph/0101335, doi:10.1007/s100520100635.DOI
  50. [50]M. Diehl, P. Hägler, Spin densities in the transverse plane and generalized transversity distributions, Eur. Phys. J. C 44 (2005) 87–101. arXiv:hep-ph/0504175, doi:10.1140/epjc/s2005-02342-6.DOI
  51. [51]J.-Y. Kim, C. Weiss, Chiral-odd generalized parton distributions in the large-Nc limit of QCD: Spin-flavor structure, polynomiality, and sum rules, Phys. Rev. D 111 (7) (2025) 074007. arXiv:2411.17634, doi:10.1103/PhysRevD.111.074007.
  52. [52]M. Burkardt, Transverse deformation of parton distributions and transversity decomposition of angular momentum, Phys. Rev. D 72 (2005) 094020. arXiv:hep-ph/0505189, doi:10.1103/PhysRevD.72.094020.DOI
  53. [53]M. Göckeler, P. Hägler, R. Horsley, Y. Nakamura, D. Pleiter, P. E. L. Rakow, A. Schäfer, G. Schierholz, H. Stüben, J. M. Zanotti, Transverse spin structure of the nucleon from lattice QCD simulations, Phys. Rev. Lett. 98 (2007) 222001. arXiv:hep-lat/0612032, doi:10.1103/PhysRevLett.98.222001.DOI
  54. [54]S. Kaur, S. Xu, C. Mondal, X. Zhao, J. P. Vary, Spatial imaging of proton via leading-twist non-skewed GPDs with basis light-front quantization, Phys. Rev. D 109 (1) (2024) 014015. arXiv:2307.09869, doi:10.1103/PhysRevD.109.014015.DOI
  55. [55]D. Y. Ivanov, B. Pire, L. Szymanowski, O. V. Teryaev, Probing chiral odd GPD’s in diffractive electroproduction of two vector mesons, Phys. Lett. B 550 (2002) 65–76. arXiv:hep-ph/0209300, doi:10.1016/S0370-2693(02)02856-3.
  56. [56]R. Boussarie, B. Pire, L. Szymanowski, S. Wallon, Exclusive photoproduction of a γ ρ pair with a large invariant mass, JHEP 02 (2017) 054, [Erratum: JHEP 10, 029 (2018)]. arXiv:1609.03830, doi:10.1007/JHEP02(2017)054.DOI
  57. [58]V. M. Braun, A. Lenz, M. Wittmann, Nucleon Form Factors in QCD, Phys. Rev. D 73 (2006) 094019. arXiv:hep-ph/0604050, doi:10.1103/PhysRevD.73.094019.DOI
  58. [59]M. Diehl, T. Feldmann, R. Jakob, P. Kroll, Linking parton distributions to form-factors and Compton scattering, Eur. Phys. J. C 8 (1999) 409–434. arXiv:hep-ph/9811253, doi:10.1007/s100529901100.DOI
  59. [60]V. Braun, R. J. Fries, N. Mahnke, E. Stein, Higher twist distribution amplitudes of the nucleon in QCD, Nucl. Phys. B 589 (2000) 381–409, [Erratum: Nucl.Phys.B 607, 433–433 (2001)]. arXiv:hep-ph/0007279, doi:10.1016/S0550-3213(00)00516-2.DOI
  60. [61]S. Navas, et al., Review of particle physics, Phys. Rev. D 110 (3) (2024) 030001. doi:10.1103/PhysRevD.110.030001.DOI
  61. [62]A. Lenz, M. Gockeler, T. Kaltenbrunner, N. Warkentin, The Nucleon Distribution Amplitudes and their application to nucleon form factors and the N → Δ transition at intermediate values of Q², Phys. Rev. D 79 (2009) 093007. arXiv:0903.1723, doi:10.1103/PhysRevD.79.093007.DOI

Paper details

vlnc:W_85p9t7hr
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 10

downloads · 0

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

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents