betaFeedback

WorksPhysics and Astronomy

Transverse-momentum-dependent pion structures from lattice QCD: Collins-Soper kernel, soft factor, TMDWF, and TMDPDF

Abstract

We present the first lattice quantum chromodynamics (QCD) calculation of the pion valence-quark transverse-momentum-dependent parton distribution function (TMDPDF) within the framework of large-momentum effective theory (LaMET). Using correlators fixed in the Coulomb gauge (CG), we computed the quasi-TMD beam function for a pion with a mass of 300 MeV, a fine lattice spacing of a=0.06a = 0.06 fm and multiple large momenta up to 3 GeV. The intrinsic soft functions in the CG approach are extracted from form factors with large momentum transfer, and as a byproduct, we also obtain the corresponding Collins-Soper (CS) kernel. Our determinations of both the soft function and the CS kernel agree with perturbation theory at small transverse separations (b⊥b_\perp) between the quarks. At larger b⊥b_\perp, the CS kernel remains consistent with recent results obtained using both CG and gauge-invariant TMD correlators in the literature. By combining next-to-leading logarithmic (NLL) factorization of the quasi-TMD beam function and the soft function, we obtain xx-dependent pion valence-quark TMDPDF for transverse separations b⊥≳1b_\perp \gtrsim 1 fm. Interestingly, we find that the b⊥b_\perp dependence of the phenomenological parameterizations of TMDPDF for moderate values of xx are in reasonable agreement with our QCD determinations. In addition, we present results for the transverse-momentum-dependent wave function (TMDWF) for a heavier pion with 670 MeV mass.

Introduction

In high-energy scatterings, transverse-momentum-dependent parton distribution functions (TMDPDFs) provide a fundamental description of the transverse momentum and polarization degrees of freedom of quarks and gluons within hadrons [1]. These distributions play a crucial role in understanding the intricate dynamics of quark-gluon interactions and the phenomenon of color confinement. Knowledge of TMDPDFs is essential for predicting observables in multi-scale, non-inclusive high-energy processes, such as semi-inclusive deep-inelastic scattering (SIDIS) and Drell-Yan (DY), based on QCD factorization. Experiments at facilities including the Jefferson Lab 12 GeV Program [2], the Electron-Ion Collider (EIC) [3] [4], and the Large Hadron Collider (LHC) [5] rely heavily on accurate knowledge of the TMDPDFs.

Significant progress has been made in the phenomenological parameterizations of TMDPDFs, particularly for the nucleon, through global analyses of experimental data [6, 7, 8, 9, 10, 11, 12, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26] [13]. While these studies have significantly improved our understanding of the transverse-momentum structure of quarks and gluons within the nucleon, they remain in their early stages due to the limited availability of experimental data sensitive to the non-perturbative region. Compared to nucleons, much less is known about the transverse momentum structure of the pion [27, 28, 29]. As the lightest QCD bound state, the pion plays a fundamental role in hadron structure and non-perturbative QCD. Its TMDPDFs provide crucial insights into the internal dynamics of quarks and gluons in a strongly interacting relativistic system, with direct implications for hadronization processes and the underlying mechanisms of non-perturbative QCD. However, the scarcity of experimental data on pion TMDPDFs, combined with the inherent nature of the parameterization dependence of global analyses, leaves significant gaps in our non-perturbative first-principles understanding. Lattice QCD calculations provide a natural approach to gain insight into non-perturbative TMD structures of the pion starting from first-principles QCD.

Early lattice QCD studies primarily focused on computing the moments of TMDPDFs [30, 31, 32, 33]. More recently, large-momentum effective theory (LaMET) [34, 35, 36, 37] has provided a framework for directly calculating the xx-dependence of parton distributions [38], including TMDPDFs [39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57]. In the large-momentum limit, quasi-TMDs, defined as equal-time correlators in highly boosted hadron states, can be related to their light-cone counterparts through perturbative factorization, making first-principles calculations of TMDPDFs possible within lattice QCD.

This framework has driven substantial progress in recent years. One of its key achievements is the determination of the Collins-Soper (CS) kernel, which governs the scale dependence of TMDPDFs [58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68]. Another crucial development is the proposal to extract the intrinsic soft function [44, 53], which accounts for the soft gluon radiation and plays a vital role in TMD factorization, from the meson form factors with large momentum transfer [59, 60, 65]. This step completes the factorization framework that connects quasi-TMDs to light-cone TMDs in the large-momentum limit. These advances have significantly enhanced first-principles lattice calculations of TMDPDFs. Recent achievements include studies of the pion TMD wave function (TMDWF) [69], the extraction of the unpolarized nucleon TMDPDF [70], and investigations of the Boer-Mulders functions of the pion and nucleon [71, 72].

Despite these advancements, a major challenge in lattice calculations of gauge-invariant (GI) quasi-TMDPDFs arises from the structure of the quark-bilinear operators, which are separated in both longitudinal and transverse directions and connected by a staple-shaped Wilson link to maintain gauge invariance. The presence of Wilson lines introduces linear divergences, requiring careful renormalization [43, 66, 73, 74, 75, 76, 77, 78], and leads to an exponential suppression of the signal-to-noise ratio (SNR) as the quark field separation increases [68]. This suppression presents a significant obstacle to precise lattice determinations of TMDPDFs, particularly in the low transverse momentum region or at large spatial separations, where high precision is crucial. Furthermore, controlling the discretization effects and power corrections at finite momentum remains a challenge in ensuring reliable extrapolations to the continuum and physical limits.

To address this issue, a new approach based on LaMET has been proposed [79, 80] to extract TMDPDFs from Coulomb gauge (CG) fixed quark correlators. Unlike traditional GI correlators, CG correlators do not require Wilson lines. Despite this difference, they fall within the same universality class [36, 81] as GI correlators with staple-shaped Wilson lines, since both reduce to light-cone TMD correlators in the infinite boost limit. Consequently, CG quasi-TMDPDFs can be matched to light-cone TMDPDFs through a perturbative factorization [80]. The absence of Wilson lines simplifies renormalization and significantly enhances the SNR of boosted correlators, particularly at large transverse separations.

In this work, we present the first lattice QCD calculation of the unpolarized valence-quark pion TMDPDF using the CG method. We compute the quasi-TMDPDF and quasi-TMDWF matrix elements of the pion in CG. These matrix elements enable the extraction of the CS kernel, the intrinsic soft function, the TMDWF, and ultimately the unpolarized valence TMDPDF of the pion. Furthermore, we improve the perturbative accuracy of the matching procedure, particularly in the extraction of the intrinsic soft function, by resumming logarithmic terms in the Sudakov kernel, an aspect that has not been accounted for in previous lattice QCD studies. Finally, the comparison of our results with global fits of the experimental data shows encouraging consistency, demonstrating the bright potential of the CG approach for TMD physics from lattice QCD.

The paper is organized as follows: we first introduce the theoretical framework for calculating the TMDPDF from the CG quasi-TMDPDFs in Sec. II, which also includes the CS kernel and the intrinsic soft factor; the lattice setup is presented in Sec. III; then we present the detailed analysis of the quasi-TMD in Sec. IV: the CS kernel, the intrinsic soft factor, the full pion TMDWF and TMDPDF are analyzed in § 1.4 to be compared with the phenomenological results; finally, we conclude in Sec. VI.

Theoretical framework

Light-cone TMDPDF from CG quasi-TMDPDF

As proposed in Ref. [80], the light-cone TMDPDF can be derived from the CG quasi-TMD beam function defined as

f~Γ(x,b⊥;Pz;μ)=Pz∫dz2πeiz(xPz)h~Γ(z,b⊥;Pz;μ)(1)\widetilde f_\Gamma(x,b_\perp;P^z;\mu)=P^z\int\frac{dz}{2\pi}e^{iz(xP^z)}\widetilde h_\Gamma(z,b_\perp;P^z;\mu) \tag*{(1)}

where the matrix elements are given by

h~Γ(z,b⊥;Pz;μ)=12PΓ⟨P∣qˉ(z,b⊥)Γq(0)∣∇⋅A=0∣P⟩(2)\widetilde h_\Gamma(z,b_\perp;P^z;\mu)=\frac{1}{2P_\Gamma}\langle P|\bar q(z,b_\perp)\Gamma q(0)|_{\nabla\cdot A=0}|P\rangle \tag*{(2)}

where PΓP_\Gamma is the kinetic factor for normalization. In this work, we choose Γ=γt\Gamma=\gamma^t for quasi-TMD beam function and PΓ=PtP_\Gamma=P^t. The parameter μ\mu represents the MS‾\overline{\mathrm{MS}} renormalization scale, and the space-like separation between the quark and the antiquark is denoted as b⃗≡(z,b⊥)\vec b\equiv(z,b_\perp). The hadron state carries momentum P=(Pt,0,0,Pz)P=(P^t,0,0,P^z) and satisfies the relativistic normalization condition ⟨P∣P⟩=2Pt(2π)3δ(3)(0⃗)\langle P|P\rangle=2P^t(2\pi)^3\delta^{(3)}(\vec0). When the hadron moves at large momentum PzP^z, the quasi-TMD beam function factorizes as

f~Γ(x,b⊥;Pz;μ)=HΓ(x,Pz;μ)B(x,b⊥,xP+;μ,ν)×SC0(b⊥;μ,ν)(3)\widetilde f_\Gamma(x,b_\perp;P^z;\mu)=H_\Gamma(x,P^z;\mu)B(x,b_\perp,xP^+;\mu,\nu) \times S_C^0(b_\perp;\mu,\nu) \tag*{(3)}

where B(x,b⊥,xP+;μ,ν)B(x,b_\perp,xP^+;\mu,\nu) is the light-cone beam function with zero-bin subtraction, following the procedure outlined in Refs. [82, 83]. The term SC0(b⊥;μ,ν)S_C^0(b_\perp;\mu,\nu) represents the zero-bin contribution, and the operator definitions of both functions are detailed in Ref. [80]. The parameters μ\mu and ν\nu correspond to the renormalization scales associated with the ultraviolet (UV) and rapidity divergences, respectively. The function Hf(x,Pz;μ)=∣CTMD(xPz;μ)∣2H_f(x,P^z;\mu)=|C_{\mathrm{TMD}}(xP^z;\mu)|^2 represents the hard kernel that matches the QCD quark field operator to Soft-Collinear Effective Theory (SCET) [46].

The light-cone TMDPDF can be defined from the light-cone beam function BB and the soft function SS as

f(x,b⊥;μ,ζ)=B(x,b⊥,xP+;μ,ν)S(b⊥;yn;μ,ν)(4)f(x,b_\perp;\mu,\zeta)=B(x,b_\perp,xP^+;\mu,\nu)S(b_\perp;y_n;\mu,\nu) \tag*{(4)}

where the dependence on the rapidity renormalization scale ν\nu cancels out on the right-hand side, leaving only a dependence on the Collins-Soper scale ζ≡2(xP+)2e−2yn\zeta\equiv2(xP^+)^2e^{-2y_n}. Combining Eqs. (3) and (4), the factorization formula connecting the quasi-TMDPDF and the light-cone TMDPDF is given by [41, 42, 44, 45, 80]

SI(b⊥,yn;μ)⋅f~Γ(x,b⊥,Pz;μ)=f(x,b⊥;μ,ζ)Hf(x,Pz;μ)+p.c.,(5)\sqrt{S_I(b_\perp,y_n;\mu)}\cdot\widetilde{f}_\Gamma(x,b_\perp,P^z;\mu) =f(x,b_\perp;\mu,\zeta)H_f(x,P^z;\mu)+\mathrm{p.c.}, \tag*{(5)}

where p.c. indicates the power corrections, and the intrinsic soft function SIS_I is defined as

SI(b⊥,yn;μ)≡(S(b⊥,yn;μ,ν)SC0(b⊥;μ,ν))2.(6)S_I(b_\perp,y_n;\mu)\equiv\left(\frac{S(b_\perp,y_n;\mu,\nu)}{S_C^0(b_\perp;\mu,\nu)}\right)^2. \tag*{(6)}

Here SC0S_C^0 is the CG quasi-soft function, as defined in Ref. [80], which exactly cancels the rapidity divergences in SS. The above factorization formula is of the same form as that for the GI quasi-TMDPDFs [41, 42, 44, 45], while the method to calculate SIS_I was first proposed in Ref. [44], which enables a complete determination of the TMDs from the lattice.

The intrinsic soft function satisfies the evolution equation

ddynln⁡SI(b⊥,yn;μ)=−2γ‾MS(b⊥;μ),(7)\frac{d}{dy_n}\ln S_I(b_\perp,y_n;\mu)=-2\overline{\gamma}^{\mathrm{MS}}(b_\perp;\mu), \tag*{(7)}

where γ‾MS(b⊥;μ)\overline{\gamma}^{\mathrm{MS}}(b_\perp;\mu) is the Collins-Soper kernel. Therefore, the intrinsic soft function at any yny_n satisfies

SI(b⊥,yn;μ)=SI(b⊥,yn=0;μ)⋅exp⁡(−2yn⋅γ‾MS(b⊥;μ)).(8)S_I(b_\perp,y_n;\mu) =S_I(b_\perp,y_n=0;\mu)\cdot\exp\left(-2y_n\cdot\overline{\gamma}^{\mathrm{MS}}(b_\perp;\mu)\right). \tag*{(8)}

One can redefine the intrinsic soft function as SI(b⊥;μ)≡SI(b⊥,yn=0;μ)S_I(b_\perp;\mu)\equiv S_I(b_\perp,y_n=0;\mu), allowing us to rewrite Eq. (5) in a more explicit form,

SI(b⊥;μ)⋅f~Γ(x,b⊥,Pz;μ)=f(x,b⊥;μ,ζ)×Hf(x,Pz;μ)exp⁡[12ln⁡(2xPz)2ζγ‾MS(b⊥;μ)]+O(ΛQCDxPz,1b⊥(xPz)).(9)\begin{aligned} \sqrt{S_I(b_\perp;\mu)}\cdot\widetilde{f}_\Gamma(x,b_\perp,P^z;\mu) ={}&f(x,b_\perp;\mu,\zeta)\\ &\times H_f(x,P^z;\mu)\exp\left[\frac{1}{2}\ln\frac{(2xP^z)^2}{\zeta}\overline{\gamma}^{\mathrm{MS}}(b_\perp;\mu)\right]\\ &+\mathcal{O}\left(\frac{\Lambda_{\mathrm{QCD}}}{xP^z},\frac{1}{b_\perp(xP^z)}\right). \tag*{(9)} \end{aligned}

This factorization framework provides a systematic approach to relating quasi-TMDPDFs extracted from lattice QCD to their light-cone counterparts, ensuring a well-defined separation between perturbative and nonperturbative contributions. However, power corrections of order O(ΛQCDxPz,1b⊥xPz)\mathcal{O}\left(\frac{\Lambda_{\mathrm{QCD}}}{xP^z},\frac{1}{b_\perp xP^z}\right) [84] introduce systematic uncertainties, particularly in the small-xx region. Despite these limitations, the framework offers valuable first-principles constraints in the large-xx regime, where power corrections are expected to be relatively small.

Quasi-TMDWF and CS kernel

Similar to the quasi-TMD beam function defined in Eqs. (1) and (2), the light-cone TMDWF can be derived from the CG quasi-TMDWF, which is defined as

ϕ~Γ(x,b⊥,Pz;μ)=Pz∫dz2πeiz(xPz)ϕ~Γ(z,b⊥,Pz;μ),(10)\widetilde{\phi}_\Gamma(x,b_\perp,P^z;\mu)=P^z\int\frac{dz}{2\pi}e^{iz(xP^z)}\widetilde{\phi}_\Gamma(z,b_\perp,P^z;\mu), \tag*{(10)}

with the quasi-TMDWF matrix elements given by

ϕ~Γ(z,b⊥,Pz;μ)=e−izPz/2fπPΓ⟨Ω∣q‾(z/2,b⊥)Γq(−z/2,0)∣∇⃗⋅A⃗=0∣P⟩,(11)\widetilde{\phi}_\Gamma(z,b_\perp,P^z;\mu)=\frac{e^{-izP^z/2}}{f_\pi P_\Gamma} \langle\Omega|\overline{q}(z/2,b_\perp)\Gamma q(-z/2,0)|\vec{\nabla}\cdot\vec{A}=0|P\rangle, \tag*{(11)}

which corresponds to a pion-to-vacuum matrix element, the ∣Ω⟩|\Omega\rangle represents the QCD vacuum state, the fπf_\pi is the pion decay constant, and PΓP_\Gamma is the kinetic factor for normalization. In this work, we choose Γ=γzγ5\Gamma=\gamma^z\gamma^5 for quasi-TMDWFs with nonzero momenta and PΓ=iPzP_\Gamma=iP^z, while for zero-momentum quasi-TMDWFs, we set Γ=γtγ5\Gamma=\gamma^t\gamma^5 and PΓ=PtP_\Gamma=P^t.

In the large-momentum limit, the quasi-TMDWF can be matched to the light-cone TMDWF via the factorization formula

SI(b⊥,yn;μ)⋅ϕ~Γ(x,b⊥,Pz;μ)=ϕ(x,b⊥,yn;μ,ζ,ζˉ)Hϕ(x,xˉ,Pz;μ)+p.c.,(12)\sqrt{S_I(b_\perp,y_n;\mu)}\cdot\widetilde{\phi}_\Gamma(x,b_\perp,P^z;\mu) =\phi(x,b_\perp,y_n;\mu,\zeta,\bar{\zeta})H_\phi(x,\bar{x},P^z;\mu)+\mathrm{p.c.}, \tag*{(12)}

with xˉ=1−x\bar{x}=1-x, ζˉ≡2(xˉP+)2e−2yn\bar{\zeta}\equiv2(\bar{x}P^+)^2e^{-2y_n}, and the hard kernel

Hϕ(x,xˉ,Pz;μ)=CTMD(xPz;μ)⋅CTMD(xˉPz;μ),(13)H_\phi(x,\bar{x},P^z;\mu)=C_{\mathrm{TMD}}(xP^z;\mu)\cdot C_{\mathrm{TMD}}(\bar{x}P^z;\mu), \tag*{(13)}

calculated up to NLO in Ref. [80].

An essential component of quasi-TMD factorization is the CS kernel γ‾MS(b⊥;μ)\overline{\gamma}^{\mathrm{MS}}(b_\perp;\mu), which governs the rapidity evolution of TMDs and is crucial to achieve consistent matching between quasi-TMDPDFs and light-cone TMDPDFs. The CS kernel can be extracted from the ratios of quasi-TMD matrix elements computed at different hadron momenta [39, 41, 44]. In this work, we employ the ratio of CG quasi-TMDWFs [44, 49] as

γ‾MS(b⊥;μ)=γ‾MS(b⊥,P1,P2;μ)+p.c.=1ln⁡(P2/P1)ln⁡Hϕ(x,xˉ,P1;μ)ϕ~γzγ5(x,b⊥,P2;μ)Hϕ(x,xˉ,P2;μ)ϕ~γzγ5(x,b⊥,P1;μ)+O(ΛQCDxPz,1b⊥(xPz),ΛQCDxˉPz,1b⊥(xˉPz)).(14)\begin{aligned} \overline{\gamma}^{\mathrm{MS}}(b_\perp;\mu) ={}&\overline{\gamma}^{\mathrm{MS}}(b_\perp,P_1,P_2;\mu)+\mathrm{p.c.}\\ ={}&\frac{1}{\ln(P_2/P_1)} \ln\frac{H_\phi(x,\bar{x},P_1;\mu)\widetilde{\phi}_{\gamma^z\gamma^5}(x,b_\perp,P_2;\mu)} {H_\phi(x,\bar{x},P_2;\mu)\widetilde{\phi}_{\gamma^z\gamma^5}(x,b_\perp,P_1;\mu)}\\ &+\mathcal{O}\left(\frac{\Lambda_{\mathrm{QCD}}}{xP^z},\frac{1}{b_\perp(xP^z)},\frac{\Lambda_{\mathrm{QCD}}}{\bar{x}P^z},\frac{1}{b_\perp(\bar{x}P^z)}\right). \tag*{(14)} \end{aligned}

This method offers a key advantage over direct extractions from quasi-TMDPDFs, as quasi-TMDWFs naturally peak around x=0.5x=0.5, a region where power corrections are significantly suppressed. On the other hand, the quasi-TMDPDFs usually decrease quickly at moderate to large xx, making the extraction of the CS kernel more susceptible to power corrections.

Intrinsic soft function

Another essential component of the factorization of quasi-TMDs is the intrinsic soft function, which accounts for soft gluon radiation in the process. Its calculation is particularly challenging for two reasons. First, as a non-perturbative quantity, it cannot be computed using standard perturbative techniques. Second, because the soft function involves correlations along two light-cone directions, it cannot be directly simulated on the lattice, even with a large-momentum boost.

A practical approach to determining the intrinsic soft function, as proposed in Refs. [44], is to use the TMD factorization of a pseudoscalar light-meson form factor, which is defined as

F(b⊥,P1,P2,Γ,Γ′)≡−4Nc⟨P2∣qˉ(b⊥)Γq(b⊥)qˉ(0)Γ′q(0)∣P1⟩fπ2(P1⋅P2)(15)F(b_{\perp},P_{1},P_{2},\Gamma,\Gamma^{\prime}) \equiv-4N_{c}\frac{\langle P_{2}|\bar{q}(b_{\perp})\Gamma q(b_{\perp})\bar{q}(0)\Gamma^{\prime}q(0)|P_{1}\rangle}{f_{\pi}^{2}(P_{1}\cdot P_{2})} \tag*{(15)}

where ∣P1⟩|P_{1}\rangle and ∣P2⟩|P_{2}\rangle denoting the pion states with momentum P1P_{1} and P2P_{2}. The coefficient −4Nc=−12-4N_{c}=-12 is a normalization factor [53]. This form factor can be extracted using the following ratio

RF(b⊥,P1,P2,Γ,Γ′)≡−4Nc⟨P2∣qˉ(b⊥)Γq(b⊥)qˉ(0)Γ′q(0)∣P1⟩⟨0∣qˉ(0)γμγ5q(0)∣P1⟩⟨P2∣qˉ(0)γμγ5q(0)∣0⟩.(16)R_{F}(b_{\perp},P_{1},P_{2},\Gamma,\Gamma^{\prime}) \equiv-4N_{c}\frac{\langle P_{2}|\bar{q}(b_{\perp})\Gamma q(b_{\perp})\bar{q}(0)\Gamma^{\prime}q(0)|P_{1}\rangle}{\langle0|\bar{q}(0)\gamma^{\mu}\gamma_{5}q(0)|P_{1}\rangle\langle P_{2}|\bar{q}(0)\gamma_{\mu}\gamma_{5}q(0)|0\rangle}. \tag*{(16)}

To extract the leading-twist contribution, one can choose the Dirac matrices as Γ=Γ′∈{I,γ5,γ⊥,γ⊥γ5}\Gamma=\Gamma^{\prime}\in\{I,\gamma^{5},\gamma^{\perp},\gamma^{\perp}\gamma^{5}\} [53, 65]. In this work, we choose Γ=Γ′∈{γ⊥,γ⊥γ5}\Gamma=\Gamma^{\prime}\in\{\gamma^{\perp},\gamma^{\perp}\gamma^{5}\}. The denominator of the ratio RFR_{F} consists of two-point functions used to normalize the form factor, and the index μ\mu should be summed. The momenta in the external pion states are taken as P1=(Pt,0,0,Pz)P_{1}=(P^{t},0,0,P^{z}) and P2=(Pt,0,0,−Pz)P_{2}=(P^{t},0,0,-P^{z}), moving back-to-back to achieve a large momentum transfer Q2=(2Pz)2Q^{2}=(2P^{z})^{2}. In this kinematic regime, the form factor can be factorized in terms of the TMDWF ϕ(x,b⊥,yn;μ,ζ)\phi(x,b_{\perp},y_{n};\mu,\zeta) [44, 49, 53] as

F(b⊥,Pz)=∫dx1dx2HF(x1,x2,Pz;μ)ϕ†(x1,b⊥,yn;μ,ζ1,ζˉ1)ϕ(x2,b⊥,−yn;μ,ζ2,ζˉ2).(17)F(b_{\perp},P^{z})=\int dx_{1}dx_{2}H_{F}(x_{1},x_{2},P^{z};\mu) \phi^{\dagger}(x_{1},b_{\perp},y_{n};\mu,\zeta_{1},\bar{\zeta}_{1})\phi(x_{2},b_{\perp},-y_{n};\mu,\zeta_{2},\bar{\zeta}_{2}) . \tag*{(17)}

where HFH_{F} is the hard kernel encapsulating the short-distance dynamics of the process. The yny_{n}-dependence cancels between ϕ\phi and ϕ†\phi^{\dagger}, and the hard kernel can be expressed as the product of two Sudakov kernels [53]:

HF(x1,x2,Pz;μ)=CSud(x1,x2,Pz;μ)⋅CSud(xˉ1,xˉ2,Pz;μ),(18)H_{F}(x_{1},x_{2},P^{z};\mu)=C_{\mathrm{Sud}}(x_{1},x_{2},P^{z};\mu)\cdot C_{\mathrm{Sud}}(\bar{x}_{1},\bar{x}_{2},P^{z};\mu), \tag*{(18)}

where CSudC_{\mathrm{Sud}} represents the Sudakov kernel, which has been computed at one-loop order in the literature [53, 85]. The renormalization group (RG) resummed results for CSudC_{\mathrm{Sud}} at next-to-leading logarithmic (NLL) accuracy are provided in App. A.

Intrinsic soft function

By combining the factorization of the quasi-TMDWF in Eq. (12) with the form factor factorization in Eq. (17), the intrinsic soft function at yn=0y_{n}=0 can be extracted as

SI(b⊥;μ)=F(b⊥,Pz)∫dx1dx2HF(x1,x2,Pz;μ)Φ~†(x1)Φ~(x2),(19)S_{I}(b_{\perp};\mu)=\frac{F(b_{\perp},P^{z})}{\int dx_{1}dx_{2}H_{F}(x_{1},x_{2},P^{z};\mu)\widetilde{\Phi}^{\dagger}(x_{1})\widetilde{\Phi}(x_{2})}, \tag*{(19)}

where the reduced quasi-TMDWF Φ~(x)\widetilde{\Phi}(x), is defined as

Φ~(x)≡Φ~Γ(x,b⊥,Pz;μ)Hϕ(x,xˉ,Pz;μ).(20)\widetilde{\Phi}(x)\equiv\frac{\widetilde{\Phi}_{\Gamma}(x,b_{\perp},P^{z};\mu)}{H_{\phi}(x,\bar{x},P^{z};\mu)}. \tag*{(20)}

Using different renormalization schemes for quasi-TMDWF, one can get different intrinsic soft functions, which can be perturbatively converted to each other at small b⊥b_{\perp}. However, the scheme dependence will eventually cancel between the renormalized quasi-beam and intrinsic soft functions. The details of scheme conversion can be found in App. B.

Lattice setup

We perform a numerical lattice QCD calculation using a 2+12+1-flavor gauge ensemble generated by the HotQCD Collaboration [86]. The ensemble employs the Highly Improved Staggered Quark (HISQ) action [87] with a lattice spacing of a=0.06a=0.06 fm and a volume of Ls3×Lt=483×64L_{s}^{3}\times L_{t}=48^{3}\times64. The valence sector is treated using tadpole-improved clover Wilson fermions on a hypercubic (HYP) smeared [88] gauge background. The clover coefficient is set to cSW=u0−3/4c_{\mathrm{SW}}=u_{0}^{-3/4}, where u0u_{0} is the average plaquette after HYP smearing. For this ensemble, we use cSW=1.0336c_{\mathrm{SW}}=1.0336 in both the time and spatial directions. The valence quark masses are tuned to yield a valence pion mass of 300 MeV, with a corresponding hopping parameter of κ≈0.12623\kappa\approx0.12623.

The key requirement for the factorization of quasi-TMDs is a sufficiently large hadron momentum. To achieve a higher boost factor, we take advantage of the three-dimensional rotational symmetry of the CG approach and adopt off-axis momenta for the quasi-TMD, choosing momentum directions along n⃗=(nx,ny,0)\vec{n}=(n^{x},n^{y},0). The hadron momenta on the lattice are given by Pz=2π∣n⃗∣LsaP^{z}=\frac{2\pi|\vec{n}|}{L_{s}a}. In this study, we consider nx=ny∈{3,4,5}n^{x}=n^{y}\in\{3,4,5\}, which allows us to reach a maximum hadron momentum of Pz=3.04P^{z}=3.04 GeV for the quasi-TMD calculations. To optimize the signal-to-noise ratio and enhance overlap with large momentum hadron ground states, we employ boosted Gaussian smearing [89], using the same setup as in Ref. [79]. To extract the ground-state contribution, we compute quasi-TMD three-point functions for multiple source-sink separations, choosing tsep/a=6,8,10,12t_{\mathrm{sep}}/a=6,8,10,12. Calculations are performed on 553 gauge configurations and we apply the All-Mode Averaging (AMA) technique [90] to further improve the signal. The stopping criteria for the sloppy and exact inversions are set to 10−410^{-4} and 10−1010^{-10}, respectively, aligning with the settings in Ref. [91]. For the smaller separation tsep/a=6t_{\mathrm{sep}}/a = 6, we compute 1 exact source and 32 sloppy sources per configuration, while for the larger separations tsep/a=8,10,12t_{\mathrm{sep}}/a = 8, 10, 12, we compute 4 exact sources and 128 sloppy sources.

In addition, we compute the large-momentum pion form factor following the strategy outlined in Ref. [59]. Specifically, we evaluate the two-current three-point function

CF(tsep,τ)=⟨πw(−P⃗,tsep)ju(b⊥,τ)jd(0,τ)πw†(P⃗,0)⟩(21)C_F(t_{\mathrm{sep}},\tau) = \left\langle\pi_w(-\vec{P},t_{\mathrm{sep}})j_u(b_\perp,\tau)j_d(0,\tau)\pi_w^\dagger(\vec{P},0)\right\rangle \tag*{(21)}

where the two currents ju(b⊥,τ)≡uˉ(b⊥,τ)Γu(b⊥,τ)j_u(b_\perp,\tau) \equiv\bar{u}(b_\perp,\tau)\Gamma u(b_\perp,\tau) and jd(0,τ)≡dˉ(0,τ)Γd(0,τ)j_d(0,\tau) \equiv\bar{d}(0,\tau)\Gamma d(0,\tau) are separated by a transverse distance b⊥b_\perp, inserted at the same time slice τ\tau. The Dirac structures are chosen as Γ∈{γ⊥,γ⊥γ5}\Gamma\in\{\gamma^\perp,\gamma^\perp\gamma_5\} to extract the leading-twist contribution [53, 65]. Unlike the two-point and quasi-TMD correlator calculations, where Gaussian-smeared sources are used, here we adopt wall sources πw\pi_w for the pions. Given the significant computational demands of this calculation and the universality of the soft factor, insensitive to the choice of hadron species, we choose to employ a heavier valence pion mass of 670 MeV. Consequently, our analysis of the pion form factor is limited to a subset of 100 gauge configurations using wall sources in Lt=64L_t = 64 time slices. We compute the form factor at a single momentum n⃗=(0,0,6)\vec{n} = (0, 0, 6), corresponding to Pz=2.58P^z = 2.58 GeV. To ensure consistency in the extraction of the intrinsic soft function, we also compute the quasi-TMDWF using the same valence pion mass. For quasi-TMDWF, we use momenta in the zz-direction with n⃗=(0,0,nz)\vec{n} = (0, 0, n^z) and nz∈{8,9,10}n^z \in\{8, 9, 10\}, reaching a maximum momentum of Pz=4.30P^z = 4.30 GeV. This part of the calculation is performed on the same 553 HotQCD gauge configurations used in the main quasi-TMD analysis, with 1 exact source and 32 sloppy sources per configuration.

quasi-TMDPDF matrix elements

Two-point function and dispersion relation

To extract the ground-state quasi-TMD matrix elements, it is essential to first determine the energy spectrum associated with the pion interpolator π\pi. The two-point correlator is defined as

C2ptSS(tsep,Pz)=⟨πs′(P⃗,tsep)πs†(P⃗,0)⟩(22)C_{2\mathrm{pt}}^{SS}(t_{\mathrm{sep}},P^z) = \left\langle\pi_{s'}(\vec{P},t_{\mathrm{sep}})\pi_s^\dagger(\vec{P},0)\right\rangle \tag*{(22)}

where πs†\pi_s^\dagger and πs′\pi_{s'} represent the pion source and sink defined by

πs(x⃗,tsep)=dˉs(x⃗,tsep)γ5us(x⃗,tsep),\pi_s(\vec{x},t_{\mathrm{sep}}) = \bar{d}_s(\vec{x},t_{\mathrm{sep}})\gamma_5u_s(\vec{x},t_{\mathrm{sep}}),
πs(P⃗,tsep)=∑x⃗πs(x⃗,tsep)e−iP⃗⋅x⃗.(23)\pi_s(\vec{P},t_{\mathrm{sep}}) = \sum_{\vec{x}}\pi_s(\vec{x},t_{\mathrm{sep}})e^{-i\vec{P}\cdot\vec{x}.} \tag*{(23)}

In this work, we use a Gaussian-smeared source and sink (s=s′=Ss = s' = S), following the setup in Ref. [92]. The superscript SSSS will be removed in the following text for simplicity.

By inserting a complete set of states, the two-point correlator can be expressed as a sum over energy eigenstates:

C2pt(tsep)=∑n=0Ns−1∣zn∣22En(e−Entsep+e−En(Lt−tsep)).(24)C_{2\mathrm{pt}}(t_{\mathrm{sep}}) = \sum_{n=0}^{N_s-1}\frac{|z_n|^2}{2E_n}\left(e^{-E_nt_{\mathrm{sep}}} + e^{-E_n(L_t-t_{\mathrm{sep}})}\right). \tag*{(24)}

The overlap amplitude zn=⟨n∣π†∣Ω⟩z_n = \langle n|\pi^\dagger|\Omega\rangle quantifies the projection of the pion interpolator onto the nnth energy eigenstate, and NsN_s denotes the number of excited states with the same quantum numbers as the pion, which are considered within the fit function.

To investigate the asymptotic behavior of C2ptC_{2\mathrm{pt}} at large tsept_{\mathrm{sep}}, we define the effective mass as

meff(tsep)=ln⁡(C2pt(tsep)C2pt(tsep+a)).(25)m_{\mathrm{eff}}(t_{\mathrm{sep}}) = \ln\left(\frac{C_{2\mathrm{pt}}(t_{\mathrm{sep}})}{C_{2\mathrm{pt}}(t_{\mathrm{sep}}+a)}\right). \tag*{(25)}

The effective masses for different momenta are shown in the upper panel of Figure 1. As tsept_{\mathrm{sep}} increases, the effective masses approach plateaus, which align with the dashed lines computed from the relativistic dispersion relation

Image not extracted

Figure 1. Upper panel: effective mass plot of C2ptSSC_{2\mathrm{pt}}^{SS} with various momenta. The dashed lines are the ground state energies of pion calculated using the dispersion relation E0=m02+(Pz)2E_0 = \sqrt{m_0^2 + (P^z)^2}, where the valence pion mass is m0=300m_0 = 300 MeV. Lower panel: the ground-state energies E0E_0 extracted from two-state fit of the two-point functions. The red line represents the exact dispersion relations with m0=300m_0 = 300 MeV.

E0=m02+(Pz)2.E_0 = \sqrt{m_0^2 + (P^z)^2}.

This agreement indicates that the ground state is effectively isolated from the excited-state tower for tsep≳8at_{\mathrm{sep}} \gtrsim8a.

In practical analysis, the upper bound NsN_s must be truncated, as higher excited-state contributions decay rapidly with increasing tsept_{\mathrm{sep}}. In this work, we perform a two-state fit by setting Ns=2N_s = 2, which allows us to efficiently extract the ground-state contribution while accounting for the leading excited-state contamination. The lower panel of Figure 1 presents the extracted ground-state energies as a function of the hadron momentum. To test the validity of the relativistic dispersion relation, we fit the data points using the functional form

E=m02+c1P2+c2a2P4,(26)E = \sqrt{m_0^2 + c_1 P^2 + c_2 a^2 P^4}, \tag*{(26)}

where EE is the ground-state energy, PP is the hadron momentum, and aa is the lattice spacing. The coefficients c1c_1 and c2c_2 parameterize the possible discretization effects. As shown by the fit bands in Figure 1, the extracted ground-state energies exhibit excellent agreement with the expected relativistic dispersion relation up to P≈3.6P \approx3.6 GeV. This agreement shows that discretization effects remain small in the dispersion relation within the momentum range studied in this work.

Bare quasi-TMD beam function matrix elements of pion

The quasi-TMD beam function is extracted from the three-point correlator

C~h(tsep,τ)=⟨πs(P⃗,tsep)qˉ(z,b⊥,τ)γtq(0⃗,τ)πs†(P⃗,0)⟩,(27)\widetilde{C}_{h}(t_{\mathrm{sep}},\tau) = \left\langle\pi_s(\vec{P},t_{\mathrm{sep}})\bar{q}(z,b_\perp,\tau)\gamma^t q(\vec{0},\tau)\pi_s^\dagger(\vec{P},0)\right\rangle, \tag*{(27)}

which can be expressed in terms of a spectral decomposition:

C~h(tsep,τ)=∑n,m=0Ns−1zn†Onmzm4EnEme−En(tsep−τ)e−Emτ,(28)\widetilde{C}_{h}(t_{\mathrm{sep}},\tau) = \sum_{n,m=0}^{N_s-1} \frac{z_n^\dagger O_{nm}z_m}{4E_nE_m} e^{-E_n(t_{\mathrm{sep}}-\tau)}e^{-E_m\tau}, \tag*{(28)}

where Onm=⟨n∣qˉ(z,b⊥)γtq(0)∣m⟩O_{nm} = \langle n|\bar{q}(z,b_\perp)\gamma^t q(0)|m\rangle is proportional to the matrix elements of the quasi-TMD operator. To extract the ground-state matrix element O00=2E0h~γtO_{00} = 2E_0\widetilde{h}_{\gamma^t}, we employ a two-state chained fit using the two-point correlator C2ptC_{2\mathrm{pt}} and the ratio

Rh~(tsep,τ)=C~h(tsep,τ)C2pt(tsep).(29)R_{\widetilde{h}}(t_{\mathrm{sep}},\tau) = \frac{\widetilde{C}_{h}(t_{\mathrm{sep}},\tau)} {C_{2\mathrm{pt}}(t_{\mathrm{sep}})}. \tag*{(29)}

In the limit tsep→∞t_{\mathrm{sep}} \to\infty, the ratio Rh~R_{\widetilde{h}} asymptotically converges to the bare quasi-TMD matrix element h~γt\widetilde{h}_{\gamma^t}. The chained fit procedure first fits the two-point correlator C2ptC_{2\mathrm{pt}}, then uses the posterior distributions of E0E_0, E1E_1, z0z_0, and z1z_1 as priors in the subsequent fit of Rh~R_{\widetilde{h}}.

For illustration, six examples of the chained fit applied to the quasi-TMD matrix elements are shown in Figure 2. The error bars represent the data points of the ratio Rh~R_{\widetilde{h}}, the colored bands depict the results of the two-state fit, and the gray band indicates the extracted ground-state matrix element. Additional details on the ground-state fit can be found in App. § 1.11.

Image not extracted

Figure 2. From left to right, the real part of the ratios Rh~R_{\tilde{h}} between three-point and two-point functions for the pion quasi-TMD, corresponding to hadron momenta Pz=1.83P^z = 1.83, 2.432.43 and 3.043.04 GeV, are shown as functions of tsept_{\mathrm{sep}} and τ\tau. The upper and lower panels are for the cases with (b⊥,z)=(1,3) a(b_\perp,z)=(1,3)\,a and (b⊥,z)=(3,3) a(b_\perp,z)=(3,3)\,a, respectively. The colored bands are two-state fit results while the gray band is the estimated ground-state matrix element.

The extracted bare matrix elements of the quasi-TMD function in the coordinate space are presented in Figure 3. The transverse separation b⊥b_\perp is plotted up to 1.2 fm for three different hadron momenta: Pz=1.83P^z = 1.83, 2.43, and 3.04 GeV. In all cases, the quasi-TMD function decreases as b⊥b_\perp increases and asymptotically approaches zero in the large-zz regime, consistent with expected physical behavior.

Image not extracted

Figure 3. The bare matrix elements of quasi-TMD in the coordinate space are plotted as the function of b⊥b_\perp and λ=zPz\lambda=zP^z. These matrix elements are extracted from the two-state chained fit of C2ptC_{2\mathrm{pt}} and Rh~R_{\tilde{h}}. From left to right, the three panels correspond to hadron momenta of Pz=1.83P^z=1.83, 2.432.43 and 3.043.04 GeV, respectively. It is observed that for all three hadron momenta, the quasi-TMD decays as the transverse separation b⊥b_\perp increases and asymptotically approaches zero in the large λ\lambda regime.

Renormalization and extrapolation

As discussed in Ref. [79], the absence of Wilson lines in the CG correlator eliminates linear divergences, allowing the renormalized operator to be defined as

qˉ0(z,b⊥)Γq0(0)=Zq(a)[qˉ(z,b⊥)Γq(0)],(30)\bar{q}_0(z,b_\perp)\Gamma q_0(0) = Z_q(a)\left[\bar{q}(z,b_\perp)\Gamma q(0)\right], \tag*{(30)}

where ZqZ_q is the CG quark wave function renormalization factor. This renormalization is independent of both the external hadron states and the spatial separation of the quark bilinear operator. Consequently, an appropriate ratio can be constructed to cancel out the renormalization factor.

Following the approach in Ref. [68], we renormalize the quasi-TMD matrix elements using the ratio

h~γt(z,b⊥,Pz;μ)=h~γt0(z,b⊥,Pz;a)φ~γtγ50(z=0,b⊥,Pz=0;a),(31)\widetilde{h}_{\gamma^t}(z,b_\perp,P^z;\mu) = \frac{\widetilde{h}_{\gamma^t}^{0}(z,b_\perp,P^z;a)} {\widetilde{\varphi}_{\gamma^t\gamma^5}^{0}(z=0,b_\perp,P^z=0;a)}, \tag*{(31)}

and similarly for the quasi-TMDWF, it is renormalized as

φ~γtγ5(z,b⊥,Pz;μ)=φ~γtγ50(z,b⊥,Pz;a)φ~γtγ50(z=0,b⊥,Pz=0;a).(32)\widetilde{\varphi}_{\gamma^t\gamma^5}(z,b_\perp,P^z;\mu) = \frac{\widetilde{\varphi}_{\gamma^t\gamma^5}^{0}(z,b_\perp,P^z;a)} {\widetilde{\varphi}_{\gamma^t\gamma^5}^{0}(z=0,b_\perp,P^z=0;a)}. \tag*{(32)}

Here, φ~0\widetilde{\varphi}^{0} denotes the bare quasi-TMDWF matrix elements defined in Eq. (11) and the details of extraction can be found in App. C.

The renormalized matrix elements in the coordinate space are presented in Figure 4. As shown, the quasi-TMD matrix elements decay rapidly as a function of λ=zPz\lambda= zP^z, reaching approximately zero for λ≳5\lambda\gtrsim5. However, at large distances, while the values remain statistically consistent with zero, the statistical uncertainties persist at a constant level, which will lead to non-physical fluctuations in the direct Fourier transform. Due to the finite correlation length of spatial correlators in QCD [93], the quasi-TMD matrix elements in coordinate space are expected to exhibit exponential decay once the coordinate separation zz is large. Moreover, as demonstrated in Ref. [93], the extracted quasi-distributions in momentum space within the moderate xx region remain largely insensitive to the choice of extrapolation strategy. Therefore, we apply a non-fit extrapolation method to smooth the uncertainties of the renormalized quasi-TMD matrix elements at long distances. The extrapolation is performed using the

Image not extracted

Figure 4. The extrapolation of the renormalized quasi-TMD in coordinate space for various hadron momenta is shown in three panels. From left to right, they correspond to hadron momenta Pz=1.83P^z=1.83, 2.432.43 and 3.043.04 GeV, respectively. The regions between the two red dashed lines indicate the extrapolation range using Eq. (33). For different hadron momenta, the same starting point of z=0.78z=0.78 fm is chosen for the extrapolation.

following form:

h~ext=w⋅h~+(1−w)⋅0 ,(33)\tilde{h}^{\mathrm{ext}} = w \cdot\tilde{h} + (1-w)\cdot0 \,, \tag*{(33)}

where ww is a weight function that transitions linearly from 1 to 0 within the range λ∈[zextPz,zmax⁡Pz]\lambda\in[z_{\mathrm{ext}}P^z,z_{\max}P^z], the zext=0.78 fmz_{\mathrm{ext}} = 0.78\ \mathrm{fm} is the starting point of extrapolation and zmax⁡=1.2 fmz_{\max} = 1.2\ \mathrm{fm} is the largest longitudinal separation of quasi-TMD. It is expected that z>zextz > z_{\mathrm{ext}} is large enough to see the exponential decay behavior of quasi-TMD. In addition, the comparison of TMDPDF using different zextz_{\mathrm{ext}} can be found in App. F. The extrapolation range is indicated by the two red dashed lines in Figure 4. After applying this extrapolation, the uncertainty bands smoothly converge to zero, mitigating non-physical fluctuations.

Pion valence quark TMDPDF

As shown in the factorization formula Eq. (9), the computation of the unpolarized pion valence-quark TMDPDF relies on three key inputs: the Collins-Soper (CS) kernel, the intrinsic soft function, and the quasi-TMD beam function matrix elements discussed in Sec. IV. In this section, we present the numerical results for each of these components and, ultimately, the extracted unpolarized valence TMDWF and TMDPDF of the pion.

The Collins-Soper kernel

The CS kernel can be extracted from the ratio of quasi-TMDWFs with different momenta, as described in Eq. (14). The bare matrix elements of the quasi-TMDWFs are extracted in App. C, which follows the same strategy as in Ref. [68]. For the matching procedure, we use the fixed-order one-loop results for the CG matching coefficient CTMDC_{\mathrm{TMD}} and the corresponding hard kernel HϕH_\phi, as provided in Ref. [80]. Furthermore, we account for large logarithmic terms in the hard kernel by applying renormalization group evolution to improve the accuracy of the perturbative matching up to NLL; details on this resummation can be found in App. A.

Fig. 5 shows the ratio γMS‾(b⊥,P1,P2;μ)\gamma^{\overline{\mathrm{MS}}}(b_\perp,P_1,P_2;\mu), as defined in Eq. (14), which is extracted from different combinations of momentum n1z/n2zn_1^z/n_2^z. The results are consistent across different momentum ratios within the uncertainty bands. However, a slight xx dependence and variations between different momentum ratios suggest the presence of power corrections, since the pion mass mπ=670 MeVm_\pi= 670\ \mathrm{MeV} is quite heavy, contributing to systematic uncertainties. To estimate these uncertainties, we select two sets of closely spaced momentum values: n1z/n2z=8/9n_1^z/n_2^z = 8/9 and n1z/n2z=9/10n_1^z/n_2^z = 9/10. Within each Jackknife sample, we collect two momentum pairs and include all data points over the interval x∈[0.34,0.66]x \in[0.34, 0.66]. The mean value ⟨γ⟩i\langle\gamma\rangle_i and the standard deviation σi\sigma_i are then calculated for each Jackknife sample ii. The final mean value, along with the corresponding statistical and systematic uncertainties, is estimated as follows:

Mean=Avg[⟨γ⟩i]Stat=Std[⟨γ⟩i]⋅Ns−1Sys=Avg[σi],(34)\begin{aligned} \mathrm{Mean} &= \mathrm{Avg}[\langle\gamma\rangle_i] \\ \mathrm{Stat} &= \mathrm{Std}[\langle\gamma\rangle_i]\cdot\sqrt{N_s-1} \\ \mathrm{Sys} &= \mathrm{Avg}[\sigma_i], \tag*{(34)} \end{aligned}

where “Avg” means taking the average and “Std” means taking the standard deviation, the factor Ns−1\sqrt{N_s-1} arises from the Jackknife resampling procedure.

After incorporating both statistical and systematic uncertainties, Figure 6 presents our final results for the CS kernel at μ=2\mu= 2 GeV. In the small b⊥b_\perp region, our results align well with the three-loop perturbative calculations from Refs. [94, 95], labeled N3^3LO and N3^3LL. Moreover, our calculation remains reliable in the large b⊥b_\perp region, where perturbative methods break down.

Image not extracted

Figure 6. The CS kernel at μ=2 GeV\mu= 2\ \mathrm{GeV} in this work is presented using red points with error bars. The 3-loop perturbative results [94, 95] for the CS kernel are denoted as N3^3LO and N3^3LL in the figure. The CS kernels from recent phenomenological parameterizations of experimental data are shown from MAP22 [19], IFY23 [20], ART23 [23], MAP24F1 [25], EEC24 [96], PB24 [97], MAPNN25 [98] and ART25 [26]. In addition, some recent lattice calculations are presented from LPC23 [65], ASWZ24 [67] and DWF24 [68].

To further contextualize our findings, we compare them with CS kernels extracted from recent global fits of experimental data, including MAP22 [19], IFY23 [20], ART23 [23], MAP24FI [25], EEC24 [96], PB24 [97], MAPNN25 [98] and ART25 [26]. Additionally, we include recent lattice QCD results from LPC23 [65], ASWZ24 [67], and DWF24 [68].

A particularly important observation is the agreement between our calculation and the DWF24 result from chirally symmetric domain-wall fermion configurations, both of which employ the CG framework despite using different lattice actions. Moreover, it is observable that the two most recent global analysis results (MAP24FI and ART25) exhibit a deviation from their prior results in different directions. Nevertheless, both remain consistent with this work due to the large uncertainty in our CS kernel, primarily stemming from power corrections at such a heavy pion mass, which resulted in non-flat curves in Figure 5. In our future study, the precision of our CS kernel could be refined by adopting a smaller valence pion mass.

Image not extracted

Figure 5. The ratio γMS‾(b⊥,P1,P2;μ)\gamma^{\overline{\mathrm{MS}}}(b_\perp,P_1,P_2;\mu), as defined in Eq. (14), of combinations of momentum n1z/n2zn_1^z/n_2^z. The results at μ=2 GeV\mu= 2\ \mathrm{GeV} for b⊥=2ab_\perp= 2a (upper panel) and b⊥=8ab_\perp= 8a (lower panel) are presented.

Intrinsic soft function

The intrinsic soft function can be extracted from the analysis of the form factor with a large momentum transfer, aided by the quasi-TMDWF, as discussed in Sec. II C. The details of extracting the bare quasi-TMDWF and the form factor can be found in Apps. C and § 1.4.4.

To extract the intrinsic soft function using Eq. (19), we integrate the quasi-TMDWF over x1x_1 and x2x_2. However, near the endpoint regions (x≈0x \approx0 and x≈1x \approx1), quasi-distributions suffer from significant power corrections. To mitigate this issue, we extrapolate the reduced quasi-TMDWF Φ~\widetilde{\Phi} using the functional form

Φ~(x)=cxn(1−x)n(35)\widetilde{\Phi}(x) = c x^n (1-x)^n \tag*{(35)}

where cc and nn are fitting parameters. A comparison of the reduced quasi-TMDWF before and after extrapolation is shown in Figure 7, where the three panels correspond to hadron momenta Pz=3.44P^z = 3.44, 3.87 and 4.30 GeV, respectively. All three cases are evolved to PFF=2.58P_{\mathrm{FF}} = 2.58 GeV using the rapidity evolution of quasi-TMDWF. The red shaded bands indicate the regions that constitute the input for the extrapolation fit, with the selected fit range of x∈[0.3,0.45]x \in[0.3, 0.45] and x∈[0.55,0.7]x \in[0.55, 0.7], where the results of three different momenta show consistency. The extrapolation form derived from the fitting process is applied to the endpoint regions with x<0.3x < 0.3 or x>0.7x > 0.7. Figure 7 shows that the extrapolated results outside the fit range are in alignment with the observed trends. Furthermore, since the Sudakov kernel diverges near the endpoints (x≈0x \approx0) after the RG resummation, the integral in Eq. (19) is restricted to the range x1,x2∈[0.05,0.95]x_1,x_2 \in[0.05, 0.95]. This cutoff has a small impact on the results of the intrinsic soft function, since the integrand in the denominator of Eq. (19) converges near the endpoint regions.

Image not extracted

Figure 7. Extrapolations of reduced quasi-TMDWF using Eq. (35). From left to right, the three panels correspond to hadron momenta Pz=3.44P^z = 3.44, 3.873.87 and 4.304.30 GeV, respectively. All of three cases are evolved to PFF=2.58P_{\rm FF} = 2.58 GeV using the rapidity evolution of quasi-TMDWF. The shaded red bands indicate the regions constitute the input for the extrapolation fit.

To assess systematic effects of power corrections, we performed integration with three different momentum pairs, formed by the momentum of the form factor PFF=2.58P_{\mathrm{FF}} = 2.58 GeV and three momenta of the quasi-TMDWF PWF=3.44P_{\mathrm{WF}} = 3.44, 3.87 and 4.30 GeV. In order to bridge the gap between PFFP_{\mathrm{FF}} and PWFP_{\mathrm{WF}}, the quasi-TMDWF is evolved to PFFP_{\mathrm{FF}} by solving the rapidity evolution equation. The extracted intrinsic soft function, according to Eq. (19), is shown in Figure 8, where comparisons across different momentum pairs indicate that power corrections have only a minor impact. Note that the superscript “ratio” of SIratioS_I^{\mathrm{ratio}} indicates the ratio scheme in the renormalization of the quasi-TMDWF.

Image not extracted

Figure 8. Intrinsic soft function in CG at μ=2\mu= 2 GeV in the ratio scheme, for different hadron momentum pairs of the quasi-TMDWF, PWFP_{\rm WF}, and form factor, PFFP_{\rm FF}. The final results are shown as red points with error bars (see text for details). The corresponding one-loop fixed-order perturbative results are the red dashed line, and the one-loop RG-resummed (RGR) perturbative results are the blue dashed line.

For the final results of SIratioS_I^{\mathrm{ratio}}, the central values and statistical uncertainties are determined from a correlated average over the three momentum pairs formed by PFFP_{\mathrm{FF}} and PWFP_{\mathrm{WF}}. The systematic uncertainties are determined by the spread of the mean values between these three momentum pairs, quantified as half the absolute difference between the maximum and minimum values. The extracted intrinsic soft function at μ=2\mu= 2 GeV, shown as red points with error bars in Figure 8, is compared to perturbative predictions of fixed order and RG resummed (RGR) at leading logarithmic (LL) precision. Further details on RG resummation can be found in App. A. In particular, at small b⊥b_\perp, our lattice results exhibit reasonable agreement with perturbative predictions, highlighting the robustness of our approach. In the large b⊥b_\perp region, where perturbation theory becomes unreliable, our lattice results are expected to provide a more accurate description of the intrinsic soft function, offering valuable non-perturbative insights.

Pion TMDWF in xx space

By integrating the CG kernel, intrinsic soft function, and the renormalized quasi-TMDWF, the light-cone TMDWF can be extracted by employing the factorization formula found in Eq. (12). Choosing scales yn=0y_n = 0, ζ=(2xPfix)2\zeta= (2xP_{\mathrm{fix}})^2, and ζ=(2xˉPfix)2\zeta= (2\bar{x}P_{\mathrm{fix}})^2, the factorization formula can be rewritten in a more explicit form:

SI(b⊥;μ) Φ~Γ(x,b⊥,Pz;μ)=ϕ(x,b⊥;μ,ζ,ζˉ)\sqrt{S_I(b_\perp;\mu)}\,\widetilde{\Phi}_\Gamma(x,b_\perp,P^z;\mu) = \phi(x,b_\perp;\mu,\zeta,\bar{\zeta})
×Hϕ(x,xˉ,Pz;μ)exp⁡[ln⁡(PzPfix)γMS(b⊥;μ)]{}\times H_\phi(x,\bar{x},P^z;\mu)\exp\left[\ln\left(\frac{P^z}{P_{\mathrm{fix}}}\right)\gamma^{\mathrm{MS}}(b_\perp;\mu)\right]
+O(ΛQCDxPz,1b⊥(xPz)2,ΛQCDxˉPz,1b⊥(xˉPz)2)(36){}+\mathcal{O}\left(\frac{\Lambda_{\mathrm{QCD}}}{xP^z},\frac{1}{b_\perp(xP^z)^2},\frac{\Lambda_{\mathrm{QCD}}}{\bar{x}P^z},\frac{1}{b_\perp(\bar{x}P^z)^2}\right) \tag*{(36)}

where the CS scale is evolved from ζ0=(2xPz)2\zeta_0 = (2xP^z)^2 and ζˉ0\bar{\zeta}_0 to ζ=(2xPfix)2\zeta= (2xP_{\rm fix})^2 and ζˉ\bar{\zeta} using the CS kernel extracted from quasi-TMDWFs.

The final results for the TMDWF at yn=0y_n = 0, Pfix=4.30P_{\rm fix} = 4.30 GeV, and μ=2\mu= 2 GeV are shown in Figure 9 as a function of the momentum fraction xx for three different hadron momenta. Two representative transverse separations, b⊥=4ab_\perp= 4a (upper panel) and b⊥=8ab_\perp= 8a (lower panel), are selected for illustration. As can be seen, the variation between different momenta remains mild in the moderate xx region, demonstrating the validity of power expansion in large PzP^z within the quasi-TMD framework, where power corrections are small. The shaded gray bands (x<0.11x < 0.11 and x>0.89x > 0.89) indicate the endpoint regions where the estimated combined systematics are greater than 30%. The combined systematics include the variation of scales and power corrections, which are estimated by the variation of the initial scale μ0\mu_0 of the RGR procedure by a factor of 2\sqrt{2} and the spread in the central values of the TMDWFs with different momenta, respectively. More detailed discussion can be found in App. G.

Image not extracted

Figure 9. The light-cone pion TMDWF at yn=0y_n = 0, Pfix=4.30P_{\rm fix} = 4.30 GeV and μ=2\mu= 2 GeV of different hadron momenta, PzP^z, as the functions of momentum fraction, xx, and for two transverse separations, b⊥=4ab_\perp= 4a (upper panel) and b⊥=8ab_\perp= 8a (lower panel). The shaded gray bands (x<0.11x < 0.11 and x>0.89x > 0.89) indicate the endpoint regions where the estimated combined systematics are larger than 30%. The detailed estimation of the systematics is explained in App. G.

Pion TMDPDF in b⊥b_\perp and k⊥k_\perp space

The combination of the CG kernel, intrinsic soft function, and renormalized quasi-TMD beam functions enables the extraction of the light-cone TMDPDF using the factorization formula in Eq. (9). The final results for the TMDPDF at ζ=μ2=4\zeta= \mu^2 = 4 GeV2^2 are shown in Figure 10 as a function of the momentum fraction xx for three different hadron momenta. Two representative transverse separations, b⊥=4ab_\perp= 4a (upper panel) and b⊥=8ab_\perp= 8a (lower panel), are selected for illustration. As can be seen, the variation between different momenta remains mild in the moderate xx region, demonstrating the validity of power expansion in large PzP^z within the quasi-TMD framework, where power corrections are small. The shaded gray bands (x<0.28x < 0.28 and x>0.81x > 0.81) indicate the endpoint regions where the estimated combined systematics are greater than 30%, with details provided in App. G.

Image not extracted

Figure 10. The unpolarized light-cone pion TMDPDF at ζ=μ2=4 GeV2\zeta= \mu^2 = 4\ \mathrm{GeV}^2 of different hadron momenta are shown as the function of momentum fraction, xx, and for transverse separations, b⊥=4ab_\perp= 4a (upper panel) and b⊥=8ab_\perp= 8a (lower panel). The shaded gray bands (x<0.28x < 0.28 and x>0.81x > 0.81) indicate the endpoint regions where the estimated combined systematics are larger than 30%. The detailed estimation of the systematics is explained in App. G.

In Figure 11, we examine the b⊥b_\perp dependence of the TMDPDF at x=0.3x = 0.3, x=0.4x = 0.4, and x=0.5x = 0.5. The results for different hadron momenta exhibit good consistency, which is improved as the xx value approaches x=0.5x = 0.5, thereby confirming our expectation of the behavior of power corrections. For comparison, we include recent global fits of the pion TMDPDF from AV19 [27], MAP22 [28] and JAM23 [29]. Although slight deviations in magnitude are observed, the lattice results exhibit a qualitative agreement with global fits, especially at the smaller xx values, where the experimental data give a better constraint.

Image not extracted

Figure 11. Unpolarized light-cone TMDPDF of pion at x=0.3x = 0.3, x=0.4x = 0.4 and x=0.5x = 0.5 in b⊥b_\perp space up to b⊥≳1b_\perp\gtrsim1 fm, and for different hadron momenta, PzP^z. The lattice results are illustrated alongside the phenomenological results from AV19 [27], MAP22 [28] and JAM23 [29]. In addition, the collinear PDF from lattice calculation in CG (PDF Lat) [79] and global fit (PDF JAM21) [99] are plotted for comparison.

Since the collinear PDF and the b⊥→0b_\perp\to0 limit of the unpolarized TMDPDF differ only by a perturbative expansion in b⊥b_\perp [23], we also compare our results to collinear PDFs. Specifically, we include the collinear PDF extracted from lattice calculations in CG (PDF Lat) [79] and the global fit result from JAM21 (PDF JAM21) [99].

tracted from lattice calculations in CG (PDF Lat) [79] and the global fit result from JAM21 (PDF JAM21) [99]. As shown in the plots, the collinear PDFs closely match the TMDPDF in the small b⊥b_\perp region, supporting the expected theoretical behavior.

An important advantage of the CG method is the absence of linear divergence, which allows us to probe large b⊥b_\perp regions with a reasonable SNR. At b⊥≳1b_\perp\gtrsim1 fm, the TMDPDF smoothly decays to values consistent with zero, facilitating a feasible Fourier transform into the k⊥k_\perp space. In this study, we combined our lattice result for the collinear pion PDF (PDF Lat) with the pion TMDPDF to interpolate the distribution at short distances around b⊥≈ab_\perp\approx a, and subsequently applied a Gaussian form to extrapolate the large b⊥b_\perp regime up to 3 fm, the extrapolation form is

f(b⊥)=Ae−m⋅b⊥2,(37)f(b_\perp)=Ae^{-m\cdot b_\perp^2}, \tag*{(37)}

where AA and mm are fit parameters. This extrapolation form is inspired by the confinement in the transverse plane, while the model dependence of the results will be investigated in future works with improved data precision. The extrapolated unpolarized light-cone TMDPDF of three momenta are plotted in Figure 12.

Image not extracted

Figure 12. Unpolarized light-cone TMDPDF of pion at x=0.3x=0.3, x=0.4x=0.4 and x=0.5x=0.5 in b⊥b_\perp space after extrapolation to b⊥=3b_\perp=3 fm. The collinear PDF from lattice calculation in CG (PDF Lat) [79] is plotted at b⊥=0b_\perp=0 fm. The point at b⊥=0.06b_\perp=0.06 fm is determined by cubic interpolation.

After extrapolation, we performed a discretized Fourier transform, with the results presented in Figure 13. Although the extrapolation in b⊥b_\perp introduces minor model dependencies, the resulting distributions exhibit qualitative consistency with global fits, especially at the smaller xx values, where the experimental data give a better constraint. This demonstrates the potential of lattice QCD to explore the transverse momentum structure of hadrons from first principles. In future studies with higher precision, the extrapolation modeling uncertainties, as well as other systematic effects, will be investigated more comprehensively to further improve the robustness of such calculations.

Image not extracted

Figure 13. Unpolarized light-cone TMDPDF of pion at x=0.3x=0.3, x=0.4x=0.4 and x=0.5x=0.5 in k⊥k_\perp space. The lattice results are illustrated alongside the phenomenological results from AV19 [27], MAP22 [28] and JAM23 [29].

Conclusion

In this work, we have presented the first lattice QCD calculation of the pion unpolarized valence-quark TMDPDF within the framework of LaMET. The calculation was performed on a 2+1 flavor QCD ensemble with a lattice spacing of a=0.06a=0.06 fm and a pion mass of 300 MeV. By leveraging high statistics, off-axis momenta, a well-tuned boosted Gaussian smeared pion source, and the novel CG approach, we attained significant hadron momenta reaching 3 GeV. We computed the quasi-TMD beam function, extracted the CS kernel from quasi-TMDWF, and derived the intrinsic soft function. To enhance perturbative accuracy, we performed RGR of the Sudakov kernel at NLL order throughout our analysis. These advancements collectively enabled the determination of both the pion TMDWF and TMDPDF from the first principles.

Our results of the CS kernel and intrinsic soft function demonstrate consistency with perturbative calculations in the small-b⊥b_\perp region, where perturbative methods remain valid. Additionally, the extracted CS kernel agrees with both previous lattice QCD studies and phenomenological fits of the experimental data.

By combining the renormalized quasi-TMDPDF, the CS kernel, and the intrinsic soft function, we obtained the pion valence TMDPDF across a range of b⊥b_\perp. Our results extend beyond b⊥≳1b_\perp\gtrsim1 fm, allowing for a feasible Fourier transform into momentum space. The extracted TMDPDFs in the b⊥b_\perp and k⊥k_\perp space exhibit qualitative agreement with phenomenological fits. This agreement demonstrates the potential of lattice QCD to validate and complement global analyses.

The outcome of this study highlights the efficacy of the CG quasi-TMD approach in probing the transverse momentum structure of hadrons. Our results provide a valuable first-principles complement to global fits, especially in the non-perturbative region where experimental data remain scarce. Future improvements, such as the incorporation of finer lattice spacings, larger momenta, and improved control over systematic uncertainties, will further enhance the precision of lattice QCD determinations of TMDPDFs. Additionally, extending this methodology to other hadrons, including the nucleon, will provide deeper insights into the partonic structure of QCD bound states.

Acknowledgments

We thank Min-Huan Chu and Hai-Tao Shu for sharing the pion form factor data. We thank Lorenzo Rossi for sharing the pion TMD results of MAP22 and Patrick C. Barry for sharing the pion TMD results of JAM23. We thank Matteo Cerutti for sharing the CS kernel results of MAPNN25, Congyue Zhang for sharing the CS kernel results of EEC24, and Aleksandra Lelek for sharing the CS kernel results of PB24. We thank Yushan Su for a valuable discussion on the RG resummation.

This material is based upon work supported by the U.S. Department of Energy, Office of Science, Office of Nuclear Physics through Contract No. DE-SC0012704, No. DE-AC02-06CH11357, within the framework of Scientific Discovery through Advanced Computing (Sci-DAC) award Fundamental Nuclear Physics at the Exascale and Beyond, and under the umbrella of the Quark-Gluon Tomography (QGT) Topical Collaboration with Award DE-SC0023646. The work of YZ is also partially supported by the Laboratory Directed Research and Development (LDRD) funding from Argonne National Laboratory, provided by the Director, Office of Science, of the U.S. Department of Energy under Contract No. DE-AC02-06CH11357.

This research used awards of computer time provided by the INCITE program at Argonne Leadership Computing Facility, a DOE Office of Science User Facility operated under Contract DE-AC02-06CH11357, the ALCC program at the Oak Ridge Leadership Computing Facility, which is a DOE Office of Science User Facility supported under Contract DE-AC05-00OR22725, and the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under Contract DE-AC02-05CH11231 using NERSC awards NP-ERCAP0028137 and NP-ERCAP0032114. Computations for this work were carried out in part on facilities of the USQCD Collaboration, which is funded by the Office of Science of the U.S. Department of Energy. We gratefully acknowledge the computing resources provided on Swing, a high-performance computing cluster operated by the Laboratory Computing Resource Center at Argonne National Laboratory. This research also used resources of the Argonne Leadership Computing Facility, a U.S. Department of Energy (DOE) Office of Science user facility at Argonne National Laboratory and is based on research supported by the U.S. DOE Office of Science-Advanced Scientific Computing Research Program, under Contract No. DE-AC02-06CH11357.

The measurement of the correlators was carried out with the Qlua software suite [100], which utilized the multigrid solver in QUDA [101, 102]. Data analysis was performed using the Python package LaMETLat, which is specifically implemented to support the analysis of Lattice QCD data within the LaMET framework. The package is available at https://github.com/Greywy-HJC/LaMETLat and is expected to undergo further development for prospective research applications.

Renormalization group resummation

RGR for TMD hard kernel

As mentioned in Sec. II, the quasi-TMD beam function and quasi-TMDWF can be matched to their light-cone counterparts using the TMD matching coefficient CTMDC_{\mathrm{TMD}}. In CG, the matching coefficient is calculated up to NLO fixed-order perturbation theory as [80]

CTMD(xPz,μ)=1+αsCF4π[−Lp22−3Lp−12+7π212],(38)C_{\mathrm{TMD}}(xP^z,\mu)=1+\frac{\alpha_s C_F}{4\pi}\left[-\frac{L_p^2}{2}-3L_p-12+\frac{7\pi^2}{12}\right], \tag*{(38)}

in which Lp=ln⁡μ2(2xPz)2L_p=\ln\frac{\mu^2}{(2xP^z)^2} and CF=4/3C_F=4/3. The RG evolution (RGE) of the matching coefficient is given in the literature [103] as

ddln⁡μln⁡CTMD=γH=−ΓcuspLp+γC.(39)\frac{d}{d\ln\mu}\ln C_{\mathrm{TMD}}=\gamma_H=-\Gamma_{\mathrm{cusp}}L_p+\gamma_C . \tag*{(39)}

Combining with the fixed-order results in Eq. (38), we can solve to get the anomalous dimensions up to NLO

Γcusp(1)=2CF⋅αs(μ)4π,(40)\Gamma_{\mathrm{cusp}}^{(1)}=2C_F\cdot\frac{\alpha_s(\mu)}{4\pi}, \tag*{(40)}

and

γC(1)=(−6CF)⋅αs(μ)4π.(41)\gamma_C^{(1)} = (-6C_F) \cdot\frac{\alpha_s(\mu)}{4\pi}. \tag*{(41)}

Employing the defined QCD beta function β\beta, one can incorporate the evolution of the coupling constant as described in dln⁡μ=dαsβd\ln\mu= \frac{d\alpha_s}{\beta}, and subsequently perform the integration as

ln⁡CTMD(μ)−ln⁡CTMD(μ0)=∫αs(μ0)αs(μ)dαsβ(αs)[−2Γcusp∫αs(μ0)αsdαs′β(αs′)+γC(αs)],(42)\ln C_{\mathrm{TMD}}(\mu) - \ln C_{\mathrm{TMD}}(\mu_0) = \int_{\alpha_s(\mu_0)}^{\alpha_s(\mu)} \frac{d\alpha_s}{\beta(\alpha_s)} \left[-2\Gamma_{\mathrm{cusp}}\int_{\alpha_s(\mu_0)}^{\alpha_s}\frac{d\alpha_s'}{\beta(\alpha_s')}+\gamma_C(\alpha_s)\right], \tag*{(42)}

where we can choose the physical scale μ0=2xPz\mu_0 = 2xP^z. Due to the double logarithmic term LP2L_P^2 in Eq. (38), the next-to-leading logarithmic (NLL) result needs the tree-level CTMD(μ0)C_{\mathrm{TMD}}(\mu_0), the one-loop γC\gamma_C and the two-loop Γcusp\Gamma_{\mathrm{cusp}}.

RGR for Sudakov kernel

The NLO fixed-order Sudakov kernel with gamma structure in the insertion current taking the form Γ=γ⊥\Gamma= \gamma_\perp or Γ=γ5γ⊥\Gamma= \gamma_5\gamma_\perp is given in Refs. [53, 85, 104]

CSud(μ)=1+αsCF4π[−ln⁡2μ2Q2−3ln⁡μ2Q2−8+π26],(43)C_{\mathrm{Sud}}(\mu) = 1 + \frac{\alpha_s C_F}{4\pi}\left[-\ln^2\frac{\mu^2}{Q^2}-3\ln\frac{\mu^2}{Q^2}-8+\frac{\pi^2}{6}\right], \tag*{(43)}

where Q2=−q2=−(p1−p2)2Q^2=-q^2=-(p_1-p_2)^2. The NLO RGE of Sudakov kernel can be derived as

ddln⁡μln⁡CSud(μ)=γ1(μ),(44)\frac{d}{d\ln\mu}\ln C_{\mathrm{Sud}}(\mu)=\gamma_1(\mu), \tag*{(44)}

with the anomalous dimension

γ1(1)(μ)=−αs(μ)CF4π[4ln⁡μ2Q2+6].(45)\gamma_1^{(1)}(\mu)=-\frac{\alpha_s(\mu)C_F}{4\pi}\left[4\ln\frac{\mu^2}{Q^2}+6\right]. \tag*{(45)}

Similarly, incorporating the evolution of coupling constant via dln⁡μ=dαsβd\ln\mu=\frac{d\alpha_s}{\beta}, one can do the integral as

ln⁡CSud(μ)−ln⁡CSud(μ0)=∫αs(μ0)αs(μ)dαsβ(αs)[−αsCF4π(8∫αs(μ0)αsdαs′β(αs′)+6)],(46)\ln C_{\mathrm{Sud}}(\mu)-\ln C_{\mathrm{Sud}}(\mu_0) =\int_{\alpha_s(\mu_0)}^{\alpha_s(\mu)}\frac{d\alpha_s}{\beta(\alpha_s)} \left[-\frac{\alpha_s C_F}{4\pi}\left(8\int_{\alpha_s(\mu_0)}^{\alpha_s}\frac{d\alpha_s'}{\beta(\alpha_s')}+6\right)\right], \tag*{(46)}

where we can choose the physical scale μ0=Q\mu_0=Q. Due to the double logarithmic term in Eq. (43), the NLL result needs the tree-level CSud(μ0)C_{\mathrm{Sud}}(\mu_0) and the one-loop γ1\gamma_1.

RGR for intrinsic soft function

In MS‾\overline{\mathrm{MS}} scheme, the NLO fixed-order intrinsic soft function in CG is [80]

SI(b⊥,μ)=1−αsCFπLb,(47)S_I(b_\perp,\mu)=1-\frac{\alpha_s C_F}{\pi}L_b, \tag*{(47)}

where Lb=ln⁡μ2b⊥2e2γE4L_b=\ln\frac{\mu^2b_\perp^2e^{2\gamma_E}}{4}. The RGE can be derived as

ddln⁡μln⁡SI(b⊥,μ)=−2αs(μ)CFπ.(48)\frac{d}{d\ln\mu}\ln S_I(b_\perp,\mu)=-2\frac{\alpha_s(\mu)C_F}{\pi}. \tag*{(48)}

Similarly, incorporating the evolution of coupling constant via dln⁡μ=dαsβd\ln\mu=\frac{d\alpha_s}{\beta}, one can do the integral as

ln⁡SI(b⊥,μ)−ln⁡SI(b⊥,μ0)=−2CFπ∫αs(μ0)αs(μ)dαsβαs.(49)\ln S_I(b_\perp,\mu)-\ln S_I(b_\perp,\mu_0) =-2\frac{C_F}{\pi}\int_{\alpha_s(\mu_0)}^{\alpha_s(\mu)}\frac{d\alpha_s}{\beta}\alpha_s. \tag*{(49)}

The leading logarithmic (LL) result is given with the combination of the tree-level SI(b⊥,μ0)S_I(b_\perp,\mu_0) and the one-loop anomalous dimension.

Scheme conversion

As discussed in Sec. II, using different renormalization schemes for quasi-TMDWF, one can get the intrinsic soft function in the corresponding renormalization schemes. This section derives the scheme conversion between the MS‾\overline{\mathrm{MS}} scheme and the ratio scheme as defined in renormalization Eqs. (31) and (32).

Using the ratio scheme, the renormalized quasi-TMDWF is defined as

ϕ~ratio(z,b⊥,Pz;μ)=ϕ~0(z,b⊥,Pz;a)ϕ~0(z=0,b⊥,Pz=0;a)=ϕ~MS‾(z,b⊥,Pz;μ)ϕ~MS‾(z=0,b⊥,Pz=0;μ),(50)\widetilde{\phi}^{\mathrm{ratio}}(z,b_\perp,P^z;\mu) =\frac{\widetilde{\phi}^{0}(z,b_\perp,P^z;a)}{\widetilde{\phi}^{0}(z=0,b_\perp,P^z=0;a)} =\frac{\widetilde{\phi}^{\overline{\mathrm{MS}}}(z,b_\perp,P^z;\mu)}{\widetilde{\phi}^{\overline{\mathrm{MS}}}(z=0,b_\perp,P^z=0;\mu)}, \tag*{(50)}
φ~ratio(x,b⊥,Pz;μ)=Pz∫dz2πeiz(xPz)ϕ~ratio(z,b⊥,Pz;μ)=∫dλ2πeiλxϕ~MS‾(z,b⊥,Pz;μ)ϕ~MS‾(z=0,b⊥,Pz=0;μ).(51)\widetilde{\varphi}^{\mathrm{ratio}}(x,b_\perp,P^z;\mu) =P^z\int\frac{dz}{2\pi}e^{iz(xP^z)}\widetilde{\phi}^{\mathrm{ratio}}(z,b_\perp,P^z;\mu) =\int\frac{d\lambda}{2\pi}e^{i\lambda x} \frac{\widetilde{\phi}^{\overline{\mathrm{MS}}}(z,b_\perp,P^z;\mu)} {\widetilde{\phi}^{\overline{\mathrm{MS}}}(z=0,b_\perp,P^z=0;\mu)}. \tag*{(51)}

in which the denominator can be approximately replaced by the short distance factorization coefficient C0C_{0} in CG, then we got

ϕ~ratio(x,b⊥,Pz;μ)=ϕ~MS‾(x,b⊥,Pz;μ)C0(b⊥;μ).(52)\widetilde{\phi}^{\mathrm{ratio}}(x,b_{\perp},P^{z};\mu)=\frac{\widetilde{\phi}^{\overline{\mathrm{MS}}}(x,b_{\perp},P^{z};\mu)}{C_{0}(b_{\perp};\mu)} . \tag*{(52)}

Combining with the Eq. (19), we have the scheme conversion of intrinsic soft function as

SIratio(b⊥;μ)=SIMS‾(b⊥;μ)⋅C02(b⊥;μ).(53)S_{I}^{\mathrm{ratio}}(b_{\perp};\mu)=S_{I}^{\overline{\mathrm{MS}}}(b_{\perp};\mu)\cdot C_{0}^{2}(b_{\perp};\mu) . \tag*{(53)}

The NLO fixed-order result of C0C_{0} can be found in Ref. [79] as

C0(b⊥;μ)=1+αsCF2π(12−Lb2).(54)C_{0}(b_{\perp};\mu)=1+\frac{\alpha_{s}C_{F}}{2\pi}\left(\frac{1}{2}-\frac{L_{b}}{2}\right) . \tag*{(54)}

One can do the RGR in the similar way as in App. A to get the LL result of C0C_{0} as

C0(b⊥;μ)=exp⁡(CFβ0ln⁡αs(μ)αs(2e−γE/b⊥)).(55)C_{0}(b_{\perp};\mu)=\exp\left(\frac{C_{F}}{\beta_{0}}\ln\frac{\alpha_{s}(\mu)}{\alpha_{s}(2e^{-\gamma_{E}}/b_{\perp})}\right) . \tag*{(55)}

Bare matrix elements of quasi-TMDWF

The quasi-TMDWF is extracted from the non-local two-point correlator

Cϕ~(tsep)=⟨OΓCG(P⃗,0)πs†(P⃗,0)⟩=⟨qˉ(z/2,b⊥,tsep)Γq(−z/2,0,tsep)πs†(P⃗,0)⟩,(56)C_{\widetilde{\phi}}(t_{\mathrm{sep}})=\left\langle O_{\Gamma}^{\mathrm{CG}}(\vec{P},0)\pi_{s}^{\dagger}(\vec{P},0)\right\rangle =\left\langle\bar{q}(z/2,b_{\perp},t_{\mathrm{sep}})\Gamma q(-z/2,0,t_{\mathrm{sep}})\pi_{s}^{\dagger}(\vec{P},0)\right\rangle, \tag*{(56)}

which can be expressed in terms of a spectral decomposition as

Cϕ~(tsep)=∑n=0Ns−1zn2En⟨Ω∣OΓCG∣n⟩(e−Entsep+e−En(Lt−tsep)),(57)C_{\widetilde{\phi}}(t_{\mathrm{sep}})=\sum_{n=0}^{N_{s}-1}\frac{z_{n}}{2E_{n}}\langle\Omega|O_{\Gamma}^{\mathrm{CG}}|n\rangle\left(e^{-E_{n}t_{\mathrm{sep}}}+e^{-E_{n}(L_{t}-t_{\mathrm{sep}})}\right) , \tag*{(57)}

where ∣Ω⟩|\Omega\rangle is the interactive vacuum state, ∣n⟩|n\rangle is the nn-th energy eigenstate. When Γ=γzγ5\Gamma=\gamma^{z}\gamma^{5}, the matrix element of the ground state is ⟨Ω∣Oγzγ5CG∣0⟩=ifπPzϕ~γzγ50\langle\Omega|O_{\gamma^{z}\gamma^{5}}^{\mathrm{CG}}|0\rangle=if_{\pi}P^{z}\widetilde{\phi}_{\gamma^{z}\gamma^{5}}^{0}, where fπf_{\pi} is the bare pion decay constant, and ϕ~γzγ50\widetilde{\phi}_{\gamma^{z}\gamma^{5}}^{0} is the bare matrix element of quasi-TMDWF in the coordinate space. When Γ=γtγ5\Gamma=\gamma^{t}\gamma^{5}, the matrix element of the ground state is ⟨Ω∣Oγtγ5CG∣0⟩=fπE0ϕ~γtγ50\langle\Omega|O_{\gamma^{t}\gamma^{5}}^{\mathrm{CG}}|0\rangle=f_{\pi}E_{0}\widetilde{\phi}_{\gamma^{t}\gamma^{5}}^{0}, where E0=mπ2+(Pz)2E_{0}=\sqrt{m_{\pi}^{2}+(P^{z})^{2}}.

In order to determine the energy spectrum and facilitate the extraction of the quasi-TMDWF, the local two-point functions of the pion are additionally computed with a valence pion mass of mπ=670m_{\pi}=670 MeV for the corresponding hadron momenta. The ground-state energies from the two-state fit of the two-point functions are illustrated in Figure 14, in comparison with the exact dispersion relations with m0=670m_{0}=670 MeV. Good consistency is found up to the hadron momentum Pz=4.30P^{z}=4.30 GeV.

Image not extracted

Figure 14. The ground-state energies E0E_{0} extracted from two-state fit of the two-point functions. The red line represents the exact dispersion relations with the valence pion mass m0=670m_{0}=670 MeV.

For correlators with nonzero momenta, we employ the two-state joint fit of C2ptC_{2\mathrm{pt}} and Cϕ~C_{\widetilde{\phi}}, while for zero-momentum correlators, the one-state fit is adopted to fit Cϕ~C_{\widetilde{\phi}}. Some of the two-state fits are selected as examples in Figure 15. As a means of illustration, both data and fit results are plotted as the ratio of Cϕ~C_{\widetilde{\phi}} and C2ptC_{2\mathrm{pt}}, which is defined as

Image not extracted

Figure 15. From left to right, the ratios of quasi-TMDWF to two-point functions Rϕ~R_{\widetilde{\phi}} with hadron momentum Pz=3.44P^{z}=3.44, 3.873.87 and 4.304.30 GeV are shown as functions of tsept_{\mathrm{sep}}. The upper and lower panels are for the cases with with (b⊥,z)=(1,4) a(b_{\perp},z)=(1,4)\,a and (b⊥,z)=(3,5) a(b_{\perp},z)=(3,5)\,a, respectively. The colored bands are joint fit results.

Rϕ~(tsep)=Cϕ~(tsep)C2pt(tsep).(58)R_{\widetilde{\phi}}(t_{\mathrm{sep}})=\frac{C_{\widetilde{\phi}}(t_{\mathrm{sep}})}{C_{2\mathrm{pt}}(t_{\mathrm{sep}})} . \tag*{(58)}

This ratio will converge to ⟨Ω∣OΓCG∣0⟩/z0\langle\Omega|O_{\Gamma}^{\mathrm{CG}}|0\rangle/z_{0} in the tsep→∞t_{\mathrm{sep}}\to\infty limit. More details on the ground-state fit can be found in App. E.

The bare matrix element of quasi-TMDWF in the coordinate space can be found in Figure 16, different momenta are normalized using the mean value of the local matrix element at (z,b⊥)=(0,0)(z,b_{\perp})=(0,0), so that some artifacts such as discretization effects can be canceled out. Due to the good convergence in the large-λ\lambda region, the renormalized matrix elements φΓ\varphi_{\Gamma} can be Fourier transformed directly to the momentum space.

Image not extracted

Figure 16. Bare matrix elements of quasi-TMDWF in the coordinate space. The non-zero-momentum correlators extracted from two-state joint fits of C2ptC_{2\mathrm{pt}} and Cϕ~C_{\widetilde{\phi}}, while the zero-momentum correlators are extracted from one-state fit of Cϕ~C_{\widetilde{\phi}}. All bare matrix elements are normalized using the mean value of the local matrix element.

Figure 17 illustrates the quasi-TMDWF in momentum space, where it can be seen that the dependence on momentum is not significant, implying that the CS kernel is not a large quantity. The small bumps in the right panel are caused by the small non-zero tails in the coordinate space, primarily manifest at large b⊥b_{\perp}.

Figure 17. Quasi-TMDWF of pion in the momentum space. From left to right, the hadron momenta are Pz=3.44P^z=3.44, 3.873.87 and 4.304.30 GeV. The small bumps in the right panel are caused by the small non-zero tails in the coordinate space, primarily manifest at large b⊥b_\perp.

Bare matrix elements of form factor

The bare pion form factor is extracted from the ratio defined in Eq. (16). When the hadron momentum is large, the leading-twist contribution in the denominator of Eq. (16) comes from the two-point functions with structures γtγ5\gamma^{t}\gamma^{5} and γzγ5\gamma^{z}\gamma^{5}. According to the suggestion in Ref. [105], we chose the two-point function with structure γzγ5+γxγ5\gamma^{z}\gamma^{5}+\gamma^{x}\gamma^{5}, which has the same leading-twist contribution as that of γzγ5\gamma^{z}\gamma^{5}, to enhance the SNR and suppress excited-state contamination when the hadron momentum is large. Compared to the two-point functions with γ5\gamma^{5} for both source and sink, the effective mass from the new interpolating operator is plotted in Figure 18. It can be seen that the new interpolating operator leads to better SNR and faster convergence to the ground state. The difference between the two-point function of γtγ5\gamma^{t}\gamma^{5} and γzγ5\gamma^{z}\gamma^{5} is estimated by a factor of Pt/PzP^{t}/P^{z} in the denominator of Eq. (16). Thus, the ratio in Eq. (16) can be expressed in terms of a spectral decomposition as

The effective mass of two-point function

Figure 18. The effective mass of two-point function, where (5,5)(5,5) means using γ5\gamma^5 for both source and sink, and (Z5−X5,Z5+X5)(Z5-X5, Z5+X5) means using γzγ5−γxγ5\gamma^z\gamma^5-\gamma^x\gamma^5 and γzγ5+γxγ5\gamma^z\gamma^5+\gamma^x\gamma^5 for source and sink, respectively. The latter choice of the interpolating operator was proposed in Ref. [105] to enhance the SNR and suppress excited-state contamination when the hadron momentum is large.

RF(tsep,τ)=−4Nc1+(Pt/Pz)2CF(tsep,τ)∣C2pt(tsep/2)∣2=−4Nc1+(Pt/Pz)2∑n,mznFnmzm∗4EnEm⋅e−En(tsep−τ)e−Emτ∣∑nznzn∗2EnAn⋅e−Entsep/2∣2 .(59)\begin{aligned} R_{F}(t_{\mathrm{sep}},\tau)=\frac{-4N_{c}}{1+(P^{t}/P^{z})^{2}}\frac{C_{F}(t_{\mathrm{sep}},\tau)}{\left|C_{2\mathrm{pt}}(t_{\mathrm{sep}}/2)\right|^{2}} \\ &=\frac{-4N_{c}}{1+(P^{t}/P^{z})^{2}}\frac{\sum_{n,m}\frac{z_{n}F_{nm}z_{m}^{*}}{4E_{n}E_{m}}\cdot e^{-E_{n}(t_{\mathrm{sep}}-\tau)}e^{-E_{m}\tau}}{\left|\sum_{n}\frac{z_{n}z_{n}^{*}}{2E_{n}}A_{n}\cdot e^{-E_{n}t_{\mathrm{sep}}/2}\right|^{2}} \, . \tag*{(59)} \end{aligned}

Here zn=⟨n∣π†∣Ω⟩z_n=\langle n|\pi^\dagger|\Omega\rangle is the same overlap amplitude defined in Sec. IV, Fnm=⟨n∣qˉ(b⊥)Γq(b⊥)qˉ(0)Γq(0)∣m⟩F_{nm}=\langle n|\bar q(b_\perp)\Gamma q(b_\perp)\bar q(0)\Gamma q(0)|m\rangle is the matrix element, and An=⟨Ω∣qˉ(0)(γzγ5+γxγ5)q(0)∣n⟩A_n=\langle\Omega|\bar q(0)(\gamma^z\gamma^5+\gamma^x\gamma^5)q(0)|n\rangle is the amplitude with a local insertion current. When both the tsept_{\rm sep} and τ\tau are large enough, we can keep the first two energy states and approximate the ratio as

RF(tsep,τ)≈−4Nc1+(Pt/Pz)2F00∣A0∣21+c1⋅(e−ΔE⋅(tsep−τ)+e−ΔE⋅τ)1+c2⋅e−ΔE⋅tsep/2(60)R_F(t_{\rm sep},\tau)\approx\frac{-4N_c}{1+(P^t/P^z)^2}\frac{F_{00}}{|A_0|^2}\frac{1+c_1\cdot\left(e^{-\Delta E\cdot(t_{\rm sep}-\tau)}+e^{-\Delta E\cdot\tau}\right)}{1+c_2\cdot e^{-\Delta E\cdot t_{\rm sep}/2}} \tag*{(60)}

where the amplitude can be expanded as A0=ifπPzA_0=if_\pi P^z. According to Eq. (15), the ground state matrix element of the form factor correlator is F00=F0fπ2((Pz)2+(Pt)2)/(−4Nc)F_{00}=F^0 f_\pi^2((P^z)^2+(P^t)^2)/(-4N_c), where F0F^0 is the bare form factor. In the limit tsep→∞t_{\rm sep}\to\infty, the ratio RF(tsep,t)R_F(t_{\rm sep},t) asymptotically converges to the bare form factor F0F^0. Therefore, we employed the two-state fit of RFR_F to extract the bare form factor F0F^0, with more details provided in App. § 1.11. Note that the fit results of F0F^0 need to be multiplied by an extra volume factor Vol=Lt3\mathrm{Vol}=L_t^3 due to the lattice setup of wall source.

To suppress the higher-twist effects, we applied the Fierz rearrangement [60] to combine different Γ\Gamma structures in Eq. (15). Ignoring the renormalization factor ZA/ZVZ_A/Z_V, the final results of the form factor is given as

F(b⊥,Pz)=14[F(b⊥,Pz,Γ=γxγ5)+F(b⊥,Pz,Γ=γyγ5)+F(b⊥,Pz,Γ=γx)+F(b⊥,Pz,Γ=γy)](61)F(b_\perp,P^z)=\frac{1}{4}\left[F(b_\perp,P^z,\Gamma=\gamma^x\gamma^5)+F(b_\perp,P^z,\Gamma=\gamma^y\gamma^5)+F(b_\perp,P^z,\Gamma=\gamma^x)+F(b_\perp,P^z,\Gamma=\gamma^y)\right] \tag*{(61)}

which is plotted in Figure 19. The form factor results calculated on MILC ensembles with lattice spacing 0.12 fm and HISQ action [87] in Ref. [65] are plotted in the figure for comparison. The small deviation can be caused by the systematics of the lattice in different ensembles, such as discretization effects, as well as the different choices of gamma structure for the two-point function.

Figure 19. The final results of form factor FF is given by averaging over form factors of four gamma structures following the Fierz rearrangement [60]. The form factor results calculated on MILC ensembles with 2+1+1-flavor with HISQ action [87] in Ref. [65] are plotted in the figure for comparison.

Ground state fit

We have performed a fully correlated Bayesian analysis of the two-point and three-point correlation functions to extract the bare matrix element h~Γ0(x,b⊥,Pz;μ)\tilde h^0_\Gamma(x,b_\perp,P^z;\mu) defined in Eq. (2), F0(b⊥,P1,P2,Γ,Γ′)F^0(b_\perp,P_1,P_2,\Gamma,\Gamma') defined in Eq. (15) and ϕ~γtγ50(x,b⊥,Pz;μ)\widetilde{\phi}^{0}_{\gamma_t\gamma_5}(x,b_\perp,P^z;\mu) defined in Eq. (10). The correlation functions across different data sets are taken into account by performing a Bayesian least-squares fit on each sample of Jackknife resampling. The parameter settings for the ground state fit are collected in Table I.

FitPartsNsN_stsept_{\mathrm{sep}} rangeτ\tau range
ϕ~γtγ50\widetilde{\phi}^{0}_{\gamma_t\gamma_5}Cϕ~(tsep)C_{\widetilde{\phi}}(t_{\mathrm{sep}})1tsep∈[12,15]t_{\mathrm{sep}}\in[12,15]/
F0F^0RF(tsep,τ)R_F(t_{\mathrm{sep}},\tau)2tsep∈{6,8,10,12}t_{\mathrm{sep}}\in\{6,8,10,12\}τ∈[3,tsep−3]\tau\in[3,t_{\mathrm{sep}}-3]
ϕ~γzγ50\widetilde{\phi}^{0}_{\gamma_z\gamma_5}C2pt(tsep)C_{2\mathrm{pt}}(t_{\mathrm{sep}})2tsep∈[3,10]t_{\mathrm{sep}}\in[3,10]/
Cϕ~(tsep)C_{\widetilde{\phi}}(t_{\mathrm{sep}})2tsep∈[5,9]t_{\mathrm{sep}}\in[5,9]/
h~γt0\widetilde{h}^{0}_{\gamma_t}C2pt(tsep)C_{2\mathrm{pt}}(t_{\mathrm{sep}})2tsep∈[3,12]t_{\mathrm{sep}}\in[3,12]/
R~h(tsep,τ)\widetilde{R}_h(t_{\mathrm{sep}},\tau)2tsep∈{6,8,10,12}t_{\mathrm{sep}}\in\{6,8,10,12\}τ∈[3,tsep−3]\tau\in[3,t_{\mathrm{sep}}-3]

Table 1. Collection of ground state fit settings. Ns=1N_s=1 means that only one ground state is included in the fit functions. The tsept_{\mathrm{sep}} range of Cϕ~C_{\widetilde{\phi}} varies in some sets of (Pz,b⊥)(P^z,b_\perp) due to the different behaviors of excited states.

For the convenience of the readers, the fit functions corresponding to each part of Table I are listed below. The function Cϕ~C_{\widetilde{\phi}} in the ground state fit of zero-momentum ϕ~γtγ50\widetilde{\phi}^{0}_{\gamma_t\gamma_5} is

Cϕ~(tsep)=ϕ~γtγ50(e−E0tsep+e−E0(Lt−tsep)),(62)C_{\widetilde{\phi}}(t_{\mathrm{sep}})=\widetilde{\phi}^{0}_{\gamma_t\gamma_5}\left(e^{-E_0t_{\mathrm{sep}}}+e^{-E_0(L_t-t_{\mathrm{sep}})}\right), \tag*{(62)}

where the fit results of ϕ~γtγ50\widetilde{\phi}^{0}_{\gamma_t\gamma_5} are normalized using the mean value of the local matrix element at z=b⊥=0z=b_\perp=0.

The function RF(tsep,τ)R_F(t_{\mathrm{sep}},\tau) in the ground state fit of bare form factor F0F^0 is

RF(tsep,τ)=F01+c1⋅(e−ΔE⋅(tsep−τ)+e−ΔE⋅τ)1+c2⋅e−ΔE⋅tsep/2,(63)R_F(t_{\mathrm{sep}},\tau)=F^0\frac{1+c_1\cdot\left(e^{-\Delta E\cdot(t_{\mathrm{sep}}-\tau)}+e^{-\Delta E\cdot\tau}\right)}{1+c_2\cdot e^{-\Delta E\cdot t_{\mathrm{sep}}/2}}, \tag*{(63)}

where the fit results of F0F^0 need to be multiplied by an extra volume factor V=L3V=L^3 due to the lattice setup of wall source. The fit function of C2pt(tsep)C_{2\mathrm{pt}}(t_{\mathrm{sep}}) in the ground state fit of ϕ~γzγ50\widetilde{\phi}^{0}_{\gamma_z\gamma_5} is

C2pt(tsep)=∑n=01∣zn∣22En(e−Entsep+e−En(Lt−tsep)),(64)C_{2\mathrm{pt}}(t_{\mathrm{sep}})=\sum_{n=0}^{1}\frac{|z_n|^2}{2E_n}\left(e^{-E_nt_{\mathrm{sep}}}+e^{-E_n(L_t-t_{\mathrm{sep}})}\right), \tag*{(64)}

and the function Cϕ~(tsep)C_{\widetilde{\phi}}(t_{\mathrm{sep}}) in the ground state fit of ϕ~γzγ50\widetilde{\phi}^{0}_{\gamma_z\gamma_5} is

Cϕ~(tsep)=z02E0E0ϕ~γzγ50(e−E0tsep+e−E0(Lt−tsep))+z12E1O01(e−E1tsep+e−E1(Lt−tsep)),(65)\begin{aligned} C_{\widetilde{\phi}}(t_{\mathrm{sep}})={}&\frac{z_0}{2E_0}E_0\widetilde{\phi}^{0}_{\gamma_z\gamma_5}\left(e^{-E_0t_{\mathrm{sep}}}+e^{-E_0(L_t-t_{\mathrm{sep}})}\right)\\ &+\frac{z_1}{2E_1}O_{01}\left(e^{-E_1t_{\mathrm{sep}}}+e^{-E_1(L_t-t_{\mathrm{sep}})}\right), \tag*{(65)} \end{aligned}

where the fit results of ϕ~γzγ50\widetilde{\phi}^{0}_{\gamma_z\gamma_5} are normalized using the mean value of the local matrix element at z=b⊥=0z=b_\perp=0.

The function C2pt(tsep)C_{\mathrm{2pt}}(t_{\mathrm{sep}}) in the ground state fit of h~γt0\tilde{h}^{0}_{\gamma_t} is

C2pt(tsep)=∑n=01∣zn∣2(e−Entsep+e−En(Lt−tsep)),(66)C_{\mathrm{2pt}}(t_{\mathrm{sep}})=\sum_{n=0}^{1}\lvert z_n\rvert^2\left(e^{-E_n t_{\mathrm{sep}}}+e^{-E_n(L_t-t_{\mathrm{sep}})}\right), \tag*{(66)}

and function Rh~(tsep,τ)R_{\tilde{h}}(t_{\mathrm{sep}},\tau) in the ground state fit of h~γt0\tilde{h}^{0}_{\gamma_t} is

Rh~(tsep,τ)=∣z0∣2h~γt0e−E0tsep∑n=01∣zn∣2(e−Entsep+e−En(Lt−tsep))+z0z1O01e−E0(tsep−τ)e−E1τ∑n=01∣zn∣2(e−Entsep+e−En(Lt−tsep))+z1z0O10e−E1(tsep−τ)e−E0τ∑n=01∣zn∣2(e−Entsep+e−En(Lt−tsep))+∣z1∣2O11e−E1tsep∑n=01∣zn∣2(e−Entsep+e−En(Lt−tsep)),(67)\begin{aligned} R_{\tilde{h}}(t_{\mathrm{sep}},\tau)={}&\frac{\lvert z_0\rvert^2\tilde{h}^{0}_{\gamma_t}e^{-E_0t_{\mathrm{sep}}}}{\sum_{n=0}^{1}\lvert z_n\rvert^2\left(e^{-E_n t_{\mathrm{sep}}}+e^{-E_n(L_t-t_{\mathrm{sep}})}\right)}\\ &+\frac{z_0z_1O_{01}e^{-E_0(t_{\mathrm{sep}}-\tau)}e^{-E_1\tau}}{\sum_{n=0}^{1}\lvert z_n\rvert^2\left(e^{-E_n t_{\mathrm{sep}}}+e^{-E_n(L_t-t_{\mathrm{sep}})}\right)}\\ &+\frac{z_1z_0O_{10}e^{-E_1(t_{\mathrm{sep}}-\tau)}e^{-E_0\tau}}{\sum_{n=0}^{1}\lvert z_n\rvert^2\left(e^{-E_n t_{\mathrm{sep}}}+e^{-E_n(L_t-t_{\mathrm{sep}})}\right)}\\ &+\frac{\lvert z_1\rvert^2O_{11}e^{-E_1t_{\mathrm{sep}}}}{\sum_{n=0}^{1}\lvert z_n\rvert^2\left(e^{-E_n t_{\mathrm{sep}}}+e^{-E_n(L_t-t_{\mathrm{sep}})}\right)}, \tag*{(67)} \end{aligned}

which has a different form from Eqs. (28) and (29) due to the cancellation of energy factors in the numerator and the denominator. In the fit functions, all overlap factors znz_n are set as real numbers, while any potential complex phases are absorbed into the matrix elements O01O_{01} and O10O_{10}.

To evaluate the overall quality of the ground state fits, the density distribution of χ2/d.o.f.\chi^2/\mathrm{d.o.f.} and the cumulative distribution function (CDF) of the pp-value are plotted in Figure 20 and Figure 21, corresponding to the matrix elements of the quasi-TMDWF and the quasi-TMD beam function, respectively. The figures reveal that χ2/d.o.f.\chi^2/\mathrm{d.o.f.} is predominantly distributed within the expected range, with pp-values exceeding 0.05 for the majority of fits, thereby suggesting a high level of overall quality. In addition, as examples, the fit results of the quasi-TMD and quasi-TMDWF are plotted alongside the data points in Figure 2 and Figure 15, respectively.

Figure 20. The evaluation of joint fits of quasi-TMDWF matrix elements. The left panel is the density distribution of χ2/d.o.f.\chi^2/\mathrm{d.o.f.} and the right panel is the cumulative distribution function (CDF) of p-value.

Figure 21. The evaluation of chained fits on R~h\widetilde{R}_h of quasi-TMD matrix elements. The left panel is the density distribution of χ2/d.o.f.\chi^2/\mathrm{d.o.f.} and the right panel is the cumulative distribution function (CDF) of p-value.

Stability of extrapolation

As mentioned in Sec. IV, the light-cone TMDPDF in momentum space within the moderate xx region is insensitive to extrapolation strategies. The unpolarized light-cone pion TMDPDF with different zextz_{\mathrm{ext}} are plotted in Figure 22 for comparison.

Image not extracted

Figure 22. The unpolarized light-cone pion TMDPDF of different zextz_{\mathrm{ext}} are plotted in comparison as the function of momentum fraction xx. Two hadron momenta Pz=2.43P^z = 2.43 GeV (left panel) and Pz=3.04P^z = 3.04 GeV (right panel) are selected as examples. The variations among different zextz_{\mathrm{ext}} are mild in the moderate region near x=0.5x = 0.5. The shaded gray bands (x<0.28x < 0.28 and x>0.81x > 0.81) indicates the endpoint regions where the estimated combined systematics are larger than 30%. The detailed estimation of the systematics is explained in App. G.

It shows that the variations of the TMDPDF among the different zextz_{\mathrm{ext}} are mild in the moderate region near x=0.5x=0.5, which shows the stability of the extrapolation.

Estimation of systematics

As mentioned in Sec. II, the results of TMDWF and TMDPDF are only reliable in the moderate xx region due to various systematics such as uncontrolled power corrections in the endpoint regions. In this section, we will give an approximate estimation of the systematics based on the results of light-cone distributions, with the objective of delineating the reliable range within the xx-space.

The estimation of the systematics is separated into two parts, the systematics of varying the initial scale of RGR, and the variation of light-cone distributions calculated at different pion momenta. Taking TMDWF and TMDPDF at b⊥=6 ab_{\perp}=6\,a as representatives, the relative systematics are plotted in Figure 23. The systematics of scale variation is estimated by varying the initial scale μ0\mu_0 of the RGR procedure by a factor of 2\sqrt{2}. Note that the systematics of scale variation is dependent on the hadron momentum. For estimation purposes, the maximum hadron momentum is selected as the representative. The momentum variation is quantified as the ratio of the spread in the central values of the light-cone distributions, evaluated at three different momenta, to their mean. These two sources of systematic uncertainty are regarded as independent, so they are combined using the root-sum-square method to provide an estimation of the combined systematics. A threshold of 30% on the combined systematics is used to define the reliable region, yielding x∈[0.11,0.89]x\in[0.11,0.89] for the TMDWF and x∈[0.28,0.81]x\in[0.28,0.81] for the TMDPDF.

Image not extracted

Figure 23. The estimation of the relative systematics of TMDWF (left panel) and TMDPDF (right panel). The scale variation is estimated by varying the initial scale μ0\mu_0 of the RGR procedure by a factor of 2\sqrt{2}. The momentum variation is quantified as the ratio of the spread in the central values of the light-cone distributions, evaluated at three different momenta, to their mean. The combined systematics are estimated by the root-sum-square of two variations. A threshold of 30% on the combined systematics is used to define the reliable region, yielding x∈[0.11,0.89]x \in[0.11, 0.89] for the TMDWF and x∈[0.28,0.81]x \in[0.28, 0.81] for the TMDPDF. These regions are indicated by the red markers in the figure.

References

References

  1. [1]R. Boussarie et al., TMD Handbook, (2023), arXiv:2304.03302 [hep-ph].arxiv.org/abs/2304.03302
  2. [2]J. Dudek et al., Physics Opportunities with the 12 GeV Upgrade at Jefferson Lab, Eur. Phys. J. A 48, 187 (2012), arXiv:1208.1244 [hep-ex].DOI
  3. [3]A. Accardi et al., Electron Ion Collider: The Next QCD Frontier: Understanding the glue that binds us all, Eur. Phys. J. A 52, 268 (2016), arXiv:1212.1701 [nucl-ex].
  4. [5]S. Amoroso et al., Snowmass 2021 Whitepaper: Proton Structure at the Precision Frontier, Acta Phys. Polon. B 53, 12 (2022), arXiv:2203.13923 [hep-ph].DOI
  5. [6]C. T. H. Davies, B. R. Webber, and W. J. Stirling, Drell-Yan Cross-Sections at Small Transverse Momentum, 1, I.95 (1984).DOI
  6. [7]G. A. Ladinsky and C. P. Yuan, The Nonperturbative regime in QCD resummation for gauge boson production at hadron colliders, Phys. Rev. D 50, R4239 (1994), arXiv:hep-ph/9311341.DOI
  7. [8]F. Landry, R. Brock, P. M. Nadolsky, and C. P. Yuan, Tevatron Run-1 Z boson data and Collins-Soper-Sterman resummation formalism, Phys. Rev. D 67, 073016 (2003), arXiv:hep-ph/0212159.DOI
  8. [9]A. V. Konychev and P. M. Nadolsky, Universality of the Collins-Soper-Sterman nonperturbative function in gauge boson production, Phys. Lett. B 633, 710 (2006), arXiv:hep-ph/0506225.DOI
  9. [10]P. Sun, J. Isaacson, C. P. Yuan, and F. Yuan, Non-perturbative functions for SIDIS and Drell-Yan processes, Int. J. Mod. Phys. A 33, 1841006 (2018), arXiv:1406.3073 [hep-ph].
  10. [11]U. D’Alesio, M. G. Echevarria, S. Melis, and I. Scimemi, Non-perturbative QCD effects in qT spectra of Drell-Yan and Z-boson production, JHEP 11, 098, arXiv:1407.3311 [hep-ph].
  11. [12]A. Bacchetta, F. Delcarro, C. Pisano, M. Radici, and A. Signori, Extraction of partonic transverse momentum [13] I. Scimemi and A. Vladimirov, Analysis of vector boson production within TMD factorization, Eur. Phys. J. C 78, 89 (2018), arXiv:1706.01473 [hep-ph].
  12. [14]V. Bertone, I. Scimemi, and A. Vladimirov, Extraction of unpolarized quark transverse momentum dependent parton distributions from Drell-Yan/Z-boson production, JHEP 06, 028, arXiv:1902.08474 [hep-ph].
  13. [15]I. Scimemi and A. Vladimirov, Non-perturbative structure of semi-inclusive deep-inelastic and Drell-Yan scattering at small transverse momentum, JHEP 06, 137, arXiv:1912.06532 [hep-ph].DOI
  14. [16]A. Bacchetta, V. Bertone, C. Bissolotti, G. Bozzi, F. Delcarro, F. Piacenza, and M. Radici, Transverse-momentum-dependent parton distributions up to N^3LL from Drell-Yan data, JHEP 07, 117, arXiv:1912.07550 [hep-ph].DOI
  15. [17]F. Hautmann, I. Scimemi, and A. Vladimirov, Non-perturbative contributions to vector-boson transverse momentum spectra in hadronic collisions, Phys. Lett. B 806, 135478 (2020), arXiv:2002.12810 [hep-ph].DOI
  16. [18]M. Bury, F. Hautmann, S. Leal-Gomez, I. Scimemi, A. Vladimirov, and P. Zurita, PDF bias and flavor dependence in TMD distributions, JHEP 10, 118, arXiv:2201.07114 [hep-ph].DOI
  17. [19]A. Bacchetta, V. Bertone, C. Bissolotti, G. Bozzi, M. Cerutti, F. Piacenza, M. Radici, and A. Signori (MAP (Multi-dimensional Analyses of Partonic distributions)), Unpolarized transverse momentum distributions from a global fit of Drell-Yan and semi-inclusive deep-inelastic scattering data, JHEP 10, 127, arXiv:2206.07598 [hep-ph].DOI
  18. [20]J. Isaacson, Y. Fu, and C. P. Yuan, Improving Res-Bos for the precision needs of the LHC, (2023), arXiv:2311.09916 [hep-ph].arxiv.org/abs/2311.09916
  19. [21]F. Aslan, M. Boglione, J. O. Gonzalez-Hernandez, T. Rainaldi, T. C. Rogers, and A. Simonelli, Phenomenology of TMD parton distributions in Drell-Yan and Z^0 boson production in a hadron structure oriented approach, (2024), arXiv:2401.14266 [hep-ph].arxiv.org/abs/2401.14266
  20. [22]A. Bacchetta, A. Bongallino, M. Cerutti, M. Radici, and L. Rossi, Exploring the three-dimensional momentum distribution of longitudinally polarized quarks in the proton, (2024), arXiv:2409.18078 [hep-ph].arxiv.org/abs/2409.18078
  21. [23]V. Moos, I. Scimemi, A. Vladimirov, and P. Zurita, Extraction of unpolarized transverse momentum distributions from the fit of Drell-Yan data at N^4LL, JHEP 05, 036, arXiv:2305.07473 [hep-ph].arxiv.org/abs/2305.07473
  22. [24]K. Yang, T. Liu, P. Sun, Y. Zhao, and B.-Q. Ma, First Extraction of Transverse Momentum Dependent Helicity Distributions, (2024), arXiv:2409.08110 [hep-ph].DOI
  23. [25]A. Bacchetta, V. Bertone, C. Bissolotti, G. Bozzi, M. Cerutti, F. Delcarro, M. Radici, L. Rossi, and A. Signori (MAP), Flavor dependence of unpolarized quark transverse momentum distributions from a global fit, JHEP 08, 232, arXiv:2405.13833 [hep-ph].DOI
  24. [26]V. Moos, I. Scimemi, A. Vladimirov, and P. Zurita, Determination of unpolarized TMD distributions from the fit of Drell-Yan and SIDIS data at N^4LL, (2025), arXiv:2503.11201 [hep-ph].arxiv.org/abs/2503.11201
  25. [27]A. Vladimirov, Pion-induced Drell-Yan processes within distributions from semi-inclusive deep-inelastic scattering, Drell-Yan and Z-boson production, JHEP 06, 081, [Erratum: JHEP 06, 051 (2019)], arXiv:1703.10157 [hep-ph].
  26. [28]M. Cerutti, L. Rossi, S. Venturini, A. Bacchetta, V. Bertone, C. Bissolotti, and M. Radici (MAP (Multi-dimensional Analyses of Partonic distributions)), Extraction of pion transverse momentum distributions from Drell-Yan data, Phys. Rev. D 107, 014014 (2023), arXiv:2210.01733 [hep-ph].
  27. [29]P. C. Barry, L. Gamberg, W. Mel nitchouk, E. Moffat, D. Pitonyak, A. Prokudin, and N. Sato (Jefferson Lab Angular Momentum (JAM)), Tomography of pions and protons via transverse momentum dependent distributions, Phys. Rev. D 108, L091504 (2023), arXiv:2302.01192 [hep-ph].DOI
  28. [30]P. Hagler, B. U. Musch, J. W. Negele, and A. Schafer, Intrinsic quark transverse momentum in the nucleon from lattice QCD, EPL 88, 61001 (2009), arXiv:0908.1283 [hep-lat].DOI
  29. [31]B. U. Musch, P. Hagler, M. Engelhardt, J. W. Negele, and A. Schafer, Sivers and Boer-Mulders observables from lattice QCD, Phys. Rev. D 85, 094510 (2012), arXiv:1111.4249 [hep-lat].DOI
  30. [32]B. Yoon, T. Bhattacharya, M. Engelhardt, J. Green, R. Gupta, P. Hagler, B. Musch, J. Negele, A. Pochinsky, and S. Syritsyn, Lattice QCD calculations of nucleon transverse momentum-dependent parton distributions using clover and domain wall fermions, in 33rd International Symposium on Lattice Field Theory (SISSA, 2015) arXiv:1601.05717 [hep-lat].arxiv.org/abs/1601.05717
  31. [33]B. Yoon, M. Engelhardt, R. Gupta, T. Bhattacharya, J. R. Green, B. U. Musch, J. W. Negele, A. V. Pochinsky, A. Schäfer, and S. N. Syritsyn, Nucleon Transverse Momentum-dependent Parton Distributions in Lattice QCD: Renormalization Patterns and Discretization Effects, Phys. Rev. D 96, 094508 (2017), arXiv:1706.03406 [hep-lat].DOI
  32. [34]X. Ji, Parton Physics on a Euclidean Lattice, Phys. Rev. Lett. 110, 262002 (2013), arXiv:1305.1539 [hep-ph].DOI
  33. [35]X. Ji, Parton Physics from Large-Momentum Effective Field Theory, Sci. China Phys. Mech. Astron. 57, 1407 (2014), arXiv:1404.6680 [hep-ph].arxiv.org/abs/1404.6680
  34. [36]X. Ji, Y.-S. Liu, Y. Liu, J.-H. Zhang, and Y. Zhao, Large-momentum effective theory, Rev. Mod. Phys. 93, 035005 (2021), arXiv:2004.03543 [hep-ph].DOI
  35. [37]X. Ji, Euclidean effective theory for partons in the spirit of Steven Weinberg, Nucl. Phys. B 1007, 116670 (2024), arXiv:2408.03378 [hep-ph].arxiv.org/abs/2408.03378
  36. [38]X. Ji, Large-Momentum Effective Theory vs. Short-Distance Operator Expansion: Contrast and Complementarity, (2022), arXiv:2209.09332 [hep-lat].arxiv.org/abs/2209.09332
  37. [39]X. Ji, P. Sun, X. Xiong, and F. Yuan, Soft factor subtraction and transverse momentum dependent parton distributions on the lattice, Phys. Rev. D 91, 074009 (2015), arXiv:1405.7640 [hep-ph].DOI
  38. [40]X. Ji, L.-C. Jin, F. Yuan, J.-H. Zhang, and Y. Zhao, Transverse momentum dependent parton quasisdistributions, Phys. Rev. D 99, 114006 (2019), arXiv:1801.05930 [hep-ph].
  39. [41]M. A. Ebert, I. W. Stewart, and Y. Zhao, Determining the Nonperturbative Collins-Soper Kernel From Lattice QCD, Phys. Rev. D 99, 034505 (2019), arXiv:1811.00226 [hep-ph].
  40. [42]M. A. Ebert, I. W. Stewart, and Y. Zhao, Towards Quasi-Transverse Momentum Dependent PDFs Com- putable on the Lattice, JHEP 09, 037, arXiv:1901.03685 [hep-ph].DOI
  41. [43]M. A. Ebert, I. W. Stewart, and Y. Zhao, Renormalization and Matching for the Collins-Soper Kernel from Lattice QCD, JHEP 03, 099, arXiv:1910.08569 [hep-ph].DOI
  42. [44]X. Ji, Y. Liu, and Y.-S. Liu, TMD soft function from large-momentum effective theory, Nucl. Phys. B 955, 115054 (2020), arXiv:1910.11415 [hep-ph].DOI
  43. [45]X. Ji, Y. Liu, and Y.-S. Liu, Transverse-momentum-dependent parton distributions from large-momentum effective theory, Phys. Lett. B 811, 135946 (2020), arXiv:1911.03840 [hep-ph].DOI
  44. [46]A. A. Vladimirov and A. Schäfer, Transverse momentum dependent factorization for lattice observables, Phys. Rev. D 101, 074517 (2020), arXiv:2002.07527 [hep-ph].DOI
  45. [47]M. A. Ebert, S. T. Schindler, I. W. Stewart, and Y. Zhao, One-loop Matching for Spin-Dependent Quasi-TMDs, JHEP 09, 099, arXiv:2004.14831 [hep-ph].DOI
  46. [48]X. Ji, Y. Liu, A. Schäfer, and F. Yuan, Single Transverse-Spin Asymmetry and Sivers Function in Large Momentum Effective Theory, Phys. Rev. D 103, 074005 (2021), arXiv:2011.13397 [hep-ph].DOI
  47. [49]X. Ji and Y. Liu, Computing light-front wave functions without light-front quantization: A large-momentum effective theory approach, Phys. Rev. D 105, 076014 (2022), arXiv:2106.05310 [hep-ph].DOI
  48. [50]M. A. Ebert, S. T. Schindler, I. W. Stewart, and Y. Zhao, Factorization connecting continuum & lattice TMDs, JHEP 04, 178, arXiv:2201.08401 [hep-ph].DOI
  49. [51]S. Rodini and A. Vladimirov, Factorization for quasi-TMD distributions of sub-leading power, JHEP 09, 117, arXiv:2211.04494 [hep-ph].arxiv.org/abs/2211.04494
  50. [52]S. T. Schindler, I. W. Stewart, and Y. Zhao, One-loop matching for gluon lattice TMDs, JHEP 08, 084, arXiv:2205.12369 [hep-ph].DOI
  51. [53]Z.-F. Deng, W. Wang, and J. Zeng, Transverse-momentum-dependent wave functions and soft functions at one-loop in large momentum effective theory, JHEP 09, 046, arXiv:2207.07280 [hep-th].DOI
  52. [54]R. Zhu, Y. Ji, J.-H. Zhang, and S. Zhao, Gluon transverse-momentum-dependent distributions from large-momentum effective theory, JHEP 02, 114, arXiv:2209.05443 [hep-ph].DOI
  53. [55]O. del Río and A. Vladimirov, Quasitransverse momentum dependent distributions at next-to-next-to-leading order, Phys. Rev. D 108, 114009 (2023), arXiv:2304.14440 [hep-ph].DOI
  54. [56]X. Ji, Y. Liu, and Y. Su, Threshold resummation for computing large-x parton distribution through large-momentum effective theory, JHEP 08, 037, arXiv:2305.04416 [hep-ph].DOI
  55. [57]X. Ji, Y. Liu, Y. Su, and R. Zhang, Effects of threshold resummation on large-x PDF in large momentum effective theory, JHEP 03, 045, arXiv:2410.12910 [hep-ph].DOI
  56. [58]P. Shanahan, M. Wagman, and Y. Zhao, Collins-Soper kernel for TMD evolution from lattice QCD, Phys. Rev. D 102, 014511 (2020), arXiv:2003.06063 [hep-lat].DOI
  57. [59]Q.-A. Zhang et al. (Lattice Parton), Lattice-QCD Calculations of TMD Soft Function Through Large-Momentum Effective Theory, Phys. Rev. Lett. 125, 192001 (2020), arXiv:2005.14572 [hep-lat].DOI
  58. [60]Y. Li et al., Lattice QCD Study of Transverse-Momentum Dependent Soft Function, Phys. Rev. Lett. 128, 062002 (2022), arXiv:2106.13027 [hep-lat].DOI
  59. [61]M. Schlemmer, A. Vladimirov, C. Zimmermann, M. Engelhardt, and A. Schäfer, Determination of the Collins-Soper Kernel from Lattice QCD, JHEP 08, 004, arXiv:2103.16991 [hep-lat].DOI
  60. [62]P. Shanahan, M. Wagman, and Y. Zhao, Lattice QCD calculation of the Collins-Soper kernel from quasi-TMDPDFs, Phys. Rev. D 104, 114502 (2021), arXiv:2107.11930 [hep-lat].DOI
  61. [63]M.-H. Chu et al. (Lattice Parton (LPC)), Nonperturbative determination of the Collins-Soper kernel from quatransverse-momentum-dependent wave functions, Phys. Rev. D 106, 034509 (2022), arXiv:2204.00200 [hep-lat].DOI
  62. [64]H.-T. Shu, M. Schlemmer, T. Sizmann, A. Vladimirov, L. Walter, M. Engelhardt, A. Schäfer, and Y.-B. Yang, Nonperturbative determination of the Collins-Soper kernel in lattice calculations, Phys. Rev. D 108, 074519 (2023), arXiv:2302.06502 [hep-lat].arxiv.org/abs/2302.06502
  63. [65]M.-H. Chu et al. (Lattice Parton (LPC)), Lattice calculation of the intrinsic soft function and the Collins-Soper kernel, JHEP 08, 172, arXiv:2306.06488 [hep-lat].DOI
  64. [66]A. Avkhadiev, P. E. Shanahan, M. L. Wagman, and Y. Zhao, Collins-Soper kernel from lattice QCD at the physical pion mass, Phys. Rev. D 108, 114505 (2023), arXiv:2307.12359 [hep-lat].arxiv.org/abs/2307.12359
  65. [67]A. Avkhadiev, P. E. Shanahan, M. L. Wagman, and Y. Zhao, Determination of the Collins-Soper Kernel from Lattice QCD, Phys. Rev. Lett. 132, 231901 (2024), arXiv:2402.06725 [hep-lat].arxiv.org/abs/2402.06725
  66. [68]D. Bollweg, X. Gao, S. Mukherjee, and Y. Zhao, Non-perturbative Collins-Soper kernel from chiral quarks with physical masses, Phys. Lett. B 852, 138617 (2024), arXiv:2403.00664 [hep-lat].DOI
  67. [69]M.-H. Chu et al. (Lattice Parton), Transverse-momentum-dependent wave functions of the pion from lattice QCD, Phys. Rev. D 109, L091503 (2024), arXiv:2302.09961 [hep-lat].arxiv.org/abs/2302.09961
  68. [70]J.-C. He, M.-H. Chu, J. Hua, X. Ji, A. Schäfer, Y. Su, W. Wang, Y.-B. Yang, J.-H. Zhang, and Q.-A. Zhang (Lattice Parton Collaboration (LPC)), Unpolarized transverse momentum dependent parton distributions of the nucleon from lattice QCD, Phys. Rev. D 109, 114513 (2024), arXiv:2211.02340 [hep-lat].arxiv.org/abs/2211.02340
  69. [71]L. Walter et al. (LPC), Quark Transverse Spin-Momentum Correlation of the Pion from Lattice QCD: The Boer-Mulders Function, (2024), arXiv:2412.19988 [hep-lat].arxiv.org/abs/2412.19988
  70. [72]L. Ma et al. (LPC), Quark Transverse Spin-Momentum Correlation of the Nucleon from Lattice QCD: The Boer-Mulders Function, (2025), arXiv:2502.11807 [hep-lat].arxiv.org/abs/2502.11807
  71. [73]M. Constantinou, H. Panagopoulos, and G. Spanoudes, One-loop renormalization of staple-shaped operators in continuum and lattice regularizations, Phys. Rev. D 99, 074508 (2019), arXiv:1901.03862 [hep-lat].DOI
  72. [74]P. Shanahan, M. L. Wagman, and Y. Zhao, Nonperturbative renormalization of staple-shaped Wilson line operators in lattice QCD, Phys. Rev. D 101, 074505 (2020), arXiv:1911.00800 [hep-lat].DOI
  73. [75]J. R. Green, K. Jansen, and F. Steffens, Improvement, generalization, and scheme conversion of Wilson-line operators on the lattice in the auxiliary field approach, Phys. Rev. D 101, 074509 (2020), arXiv:2002.09408 [hep-lat].
  74. [76]K. Zhang, X. Ji, Y.-B. Yang, F. Yao, and J.-H. Zhang ([Lattice Parton Collaboration (LPC)]), Renormalization of Transverse-Momentum-Dependent Parton Distribution on the Lattice, Phys. Rev. Lett. 129, 082002 (2022), arXiv:2205.13402 [hep-lat].DOI
  75. [77]Y. Ji, J.-H. Zhang, S. Zhao, and R. Zhu, Renormalization and mixing of staple-shaped Wilson line operators on the lattice revisited, Phys. Rev. D 104, 094510 (2021), arXiv:2104.13345 [hep-ph].DOI
  76. [78]C. Alexandrou et al., Nonperturbative renormalization of asymmetric staple-shaped operators in twisted mass lattice QCD, Phys. Rev. D 108, 114503 (2023), arXiv:2305.11824 [hep-lat].DOI
  77. [79]X. Gao, W.-Y. Liu, and Y. Zhao, Parton distributions from boosted fields in the Coulomb gauge, Phys. Rev. D 109, 094506 (2024), arXiv:2306.14960 [hep-ph].arxiv.org/abs/2306.14960
  78. [80]Y. Zhao, Transverse Momentum Distributions from Lattice QCD without Wilson Lines, (2023), arXiv:2311.01391 [hep-ph].arxiv.org/abs/2311.01391
  79. [81]Y. Hatta, X. Ji, and Y. Zhao, Gluon helicity ΔG from a universality class of operators on a lattice, Phys. Rev. D 89, 085030 (2014), arXiv:1310.4263 [hep-ph].DOI
  80. [82]A. V. Manohar and I. W. Stewart, The Zero-Bin and Mode Factorization in Quantum Field Theory, Phys. Rev. D 76, 074002 (2007), arXiv:hep-ph/0605001.DOI
  81. [83]I. W. Stewart, F. J. Tackmann, and W. J. Waalewijn, Factorization at the LHC: From PDFs to Initial State Jets, Phys. Rev. D 81, 094035 (2010), arXiv:0910.0467 [hep-ph].DOI
  82. [84]Y. Liu and Y. Su, Renormalon cancellation and linear power correction to threshold asymptotics of space-like parton correlators, JHEP 2024, 204, arXiv:2311.06907 [hep-ph].
  83. [85]A. V. Manohar, Deep inelastic scattering as x → 1 using soft collinear effective theory, Phys. Rev. D 68, 114019 (2003), arXiv:hep-ph/0309176.DOI
  84. [86]A. Bazavov et al. (HotQCD), Equation of state in (2+1)-flavor QCD, Phys. Rev. D 90, 094503 (2014), arXiv:1407.6387 [hep-lat].DOI
  85. [87]E. Follana, Q. Mason, C. Davies, K. Hornbostel, G. P. Lepage, J. Shigemitsu, H. Trottier, and K. Wong (HPQCD, UKQCD), Highly improved staggered quarks on the lattice, with applications to charm physics, Phys. Rev. D 75, 054502 (2007), arXiv:hep-lat/0610092.DOI
  86. [88]A. Hasenfratz and F. Knechtli, Flavor symmetry and the static potential with hypercubic blocking, Phys. Rev. D 64, 034504 (2001), arXiv:hep-lat/0103029.DOI
  87. [89]G. S. Bali, B. Lang, B. U. Musch, and A. Schäfer, Novel quark smearing for hadrons with high momenta in lattice QCD, Phys. Rev. D 93, 094515 (2016), arXiv:1602.05525 [hep-lat].DOI
  88. [90]E. Shintani, R. Arthur, T. Blum, T. Izubuchi, C. Jung, and C. Lehner, Covariant approximation averaging, Phys. Rev. D 91, 114511 (2015), arXiv:1402.0244 [hep-lat].DOI
  89. [91]X. Gao, L. Jin, C. Kallidonis, N. Karthik, S. Mukherjee, P. Petreczky, C. Shugert, S. Syritsyn, and Y. Zhao, Valence parton distribution of the pion from lattice QCD: Approaching the continuum limit, Phys. Rev. D 102, 094513 (2020), arXiv:2007.06590 [hep-lat].DOI
  90. [92]T. Izubuchi, L. Jin, C. Kallidonis, N. Karthik, S. Mukherjee, P. Petreczky, C. Shugert, and S. Syritsyn, Valence parton distribution function of pion from fine lattice, Phys. Rev. D 100, 034516 (2019), arXiv:1905.06349 [hep-lat].DOI
  91. [93]X. Gao, A. D. Hanlon, S. Mukherjee, P. Petreczky, P. Scior, S. Syritsyn, and Y. Zhao, Lattice QCD Determination of the Bjorken-x Dependence of Parton Distribution Functions at Next-to-Next-to-Leading Order, Phys. Rev. Lett. 128, 142003 (2022), arXiv:2112.02208 [hep-lat].DOI
  92. [94]Y. Li and H. X. Zhu, Bootstrapping Rapidity Anomalous Dimensions for Transverse-Momentum Resummation, Phys. Rev. Lett. 118, 022004 (2017), arXiv:1604.01404 [hep-ph].DOI
  93. [95]A. A. Vladimirov, Correspondence between Soft and Rapidity Anomalous Dimensions, Phys. Rev. Lett. 118, 062001 (2017), arXiv:1610.05791 [hep-ph].DOI
  94. [96]Z.-B. Kang, J. Penttala, and C. Zhang, Determination of the strong coupling constant and the Collins-Soper kernel from the energy-energy correlator in e+e− collisions, (2024), arXiv:2410.21435 [hep-ph].arxiv.org/abs/2410.21435
  95. [97]A. B. Martinez, F. Hautmann, L. Keersmaekers, A. Lelek, M. Mendizabal Morentin, S. Taheri Monfared, and A. M. van Kampen, Soft-gluon coupling and the TMD parton branching Sudakov form factor, (2024), arXiv:2412.21116 [hep-ph].arxiv.org/abs/2412.21116
  96. [98]A. Bacchetta, V. Bertone, C. Bissolotti, M. Cerutti, M. Radici, S. Rodini, and L. Rossi (MAP), A Neural-Network Extraction of Unpolarised Transverse-Momentum-Dependent Distributions, (2025), arXiv:2502.04166 [hep-ph].arxiv.org/abs/2502.04166
  97. [99]P. C. Barry, C.-R. Ji, N. Sato, and W. Melnitchouk (Jefferson Lab Angular Momentum (JAM)), Global QCD Analysis of Pion Parton Distributions with Threshold Resummation, Phys. Rev. Lett. 127, 232001 (2021), arXiv:2108.05822 [hep-ph].DOI
  98. [100]A. Pochinsky, Qlua lattice software suite, https://usqcd.lns.mit.edu/qlua (2008–present).
  99. [101]M. A. Clark, R. Babich, K. Barros, R. C. Brower, and C. Rebbi (QUDA), Solving Lattice QCD systems of equations using mixed precision solvers on GPUs, Comput. Phys. Commun. 181, 1517 (2010), arXiv:0911.3191 [hep-lat].DOI
  100. [102]R. Babich, M. A. Clark, B. Joo, G. Shi, R. C. Brower, and S. Gottlieb (QUDA), Scaling Lattice QCD beyond 100 GPUs (2011), arXiv:1109.2935 [hep-lat].DOI
  101. [103]I. W. Stewart, F. J. Tackmann, and W. J. Waalewijn, The Quark Beam Function at NNLL, JHEP 09, 005, arXiv:1002.2213 [hep-ph].DOI
  102. [104]J. Collins and T. C. Rogers, Connecting Different TMD Factorization Formalisms in QCD, Phys. Rev. D 96, 054011 (2017), arXiv:1705.07167 [hep-ph].DOI
  103. [105]R. Zhang, A. V. Grebe, D. C. Hackett, M. L. Wagman, and Y. Zhao, Kinematically-enhanced interpolating operators for boosted hadrons, (2025), arXiv:2501.00729 [hep-lat].arxiv.org/abs/2501.00729

Paper details

vlnc:W_n62p26kw
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

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

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents