betaFeedback

WorksPhysics and Astronomy

One-loop matching for gluon lattice TMDs

Abstract

Transverse-momentum-dependent parton distributions (TMDs) can be calculated from first principles by computing a related set of Euclidean lattice observables and connecting them via a factorization formula. This work focuses on the leading-power factorization formula connecting the lattice quasi-TMD and continuum Collins TMD for gluons. We calculate the one-loop gluon matching coefficient, which is known to be independent of spin and exhibits no mixing with quarks. We demonstrate that this coefficient satisfies Casimir scaling with respect to the quark matching coefficient at one-loop order. Our result facilitates reliable lattice QCD calculations of gluon TMDs.

Introduction

Protons are a fundamental building block of almost all the visible matter in the universe, yet over a century after their discovery, their internal dynamics still confound us. The behavior of quarks and gluons inside the proton are described by universal functions appearing in cross sections, such as parton distribution functions (PDFs), transverse-momentum-dependent PDFs (TMDs), and generalized parton distributions (GPDs). Various collider facilities, such as Jefferson Lab, CERN, DESY, and Fermilab, have provided initial data on these distributions, and many have dedicated plans to increase their precision. These efforts will be further bolstered by the planned Electron-Ion Collider (EIC) [1, 2] at Brookhaven National Lab. The imminence of high-precision data about parton distributions underscores the need for developing a corresponding first-principles understanding.

Parton distributions can be rigorously defined in quantum field theory. Unfortunately, they always contain important nonperturbative components that cannot be computed directly using lattice QCD, as they are typically described by matrix elements involving light-like Wilson line paths. Light-like observables depend on a real-valued time variable, which induces a lattice sign problem: a numerical difficulty suspected to be NP-hard in general. Absent direct analytical or numerical approaches, physicists have traditionally constructed models for such observables, which are then compared to experimental data using global fits. Alternatively, one can carry out an indirect first-principles calculation: one constructs a numerically-tractable lattice observable that shares the same IR physics as the target distribution, but may differ in the ultraviolet (UV). This lattice observable must then be perturbatively matched back onto the desired physical continuum distribution via a factorization formula.

Here, we focus on the physical-to-lattice factorization program for TMDs, an important class of observables describing the three-dimensional momentum distribution of quarks and gluons inside a proton. TMDs appear in many cross sections, including semi-inclusive deep-inelastic scattering (SIDIS), the Drell-Yan process, weak boson production, and Higgs production. Nonperturbative effects for TMDs dominate at small transverse momenta kT∼ΛQCDk_T \sim\Lambda_{\mathrm{QCD}}. For kT≫ΛQCDk_T \gg\Lambda_{\mathrm{QCD}}, the dependence of the TMD on kTk_T can be computed perturbatively [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]; but there are still important nonperturbative contributions [14, 15, 16, 17]. On the modeling front, high precision global fits with next-to-next-to-next-to-leading logarithmic accuracy on perturbative ingredients have been performed for the existing experimental data [18, 19]. On the lattice front, the calculation of TMD-like correlators was pioneered and developed over the course of a decade using the Musch-Hägler-Engelhardt-Negele-Schäfer (MHENS) scheme [20, 21, 22, 23, 24, 25]. Later on, quasi-TMDs were developed for the lattice [26, 27, 28, 29, 30, 31, 32, 33, 34, 35] using the framework of large momentum effective theory (LaMET) [36, 37, 38]. The leading-power factorization formula connecting quasi-TMDs to the continuum has been proposed in Refs. [28, 29, 31, 32], and recently derived to all orders in αs\alpha_s in [39]. First lattice results have been presented for various components of quasi-TMDs: the beam function and Collins-Soper kernel [40, 41, 42, 43, 44, 45], and soft function [43, 44].

At leading power, the factorization formula connecting the continuum limit of quasi-TMDs (lattice) and the physical Collins TMD reads [39]:

f~i/h[s](x,b⃗T,μ,ζ~,xP~z)=Cκi(xP~z,μ)exp⁡[12γζκi(μ,bT)ln⁡ζ~ζ]fi/h[s](x,b⃗T,μ,ζ)(1)\widetilde{f}^{[s]}_{i/h}(x,\vec{b}_T,\mu,\widetilde{\zeta},x\widetilde{P}^z)=C_{\kappa_i}(x\widetilde{P}^z,\mu)\exp\left[\frac{1}{2}\gamma_{\zeta}^{\kappa_i}(\mu,b_T)\ln\frac{\widetilde{\zeta}}{\zeta}\right]f^{[s]}_{i/h}(x,\vec{b}_T,\mu,\zeta) \tag*{(1)}

where ff is a continuum TMD, f~\widetilde{f} is the continuum limit of a lattice TMD, CκiC_{\kappa_i} is a perturbative matching coefficient, and the exponential term evolves the so-called Collins-Soper (CS) scale ζ→ζ~\zeta\to\widetilde{\zeta} using the CS kernel γζκi\gamma_{\zeta}^{\kappa_i}. The other parameters include xx, which is the fraction of the hadron hh’s longitudinal momentum PP that the parton of type ii carries; the Fourier-conjugate to the parton’s transverse momentum bTb_T; and a renormalization scale μ\mu. Here ii refers to either a gluon (i=gi=g) or specific quark flavor (i=u,d,s,…i=u,d,s,\ldots), where κi=q\kappa_i=q is universal for all quarks, but differs from κg=g\kappa_g=g for gluons. No quark-gluon or flavor mixing occurs in the factorization relation, which simplifies calculation of gluon TMDs from the lattice [39]. The definition of each of the TMDs in eq. (1) will be discussed in detail in section 1.1.

The quasi-to-Collins matching coefficient CqC_q is known for quarks at one loop [26, 28, 29], for logarithmic terms at two loops [32], and for all next-to-leading-logarithmic (NLL) terms [39]. A key missing ingredient is the matching coefficient for gluons. Gluons are known to be responsible for a large part of internal proton dynamics and would be useful to understand from first principles. In this work, we calculate the one-loop gluon matching coefficient CgC_g.

Definitions of lattice and physical continuum TMDs

A TMD fi/hf_{i/h} for gluon or quark flavor ii generally is a product of three components: a beam function (hadronic matrix element), a soft function (vacuum matrix element), and their renormalization. The beam and soft matrix elements are defined in terms of Wilson lines:

WR[γ]=Pexp⁡[ig∫γdxμAμa(x)TRa](2)W^{R}[\gamma]=P\exp\left[ig\int_{\gamma} \mathrm{d}x^\mu A_\mu^a(x)T_R^a\right] \tag*{(2)}

where γ\gamma is the Wilson line’s path and RR is its color representation. Here we focus on gluons, taking R=gR=g for the adjoint representation. Beam functions involve staple-shaped Wilson line paths as shown in Figure 1,

Figure 1. (a) Staple-shaped Wilson line defined in eq. (3). Edges may extend along the conjugate direction PP, which is not shown. (b) Wilson line structure of the soft function, eq. (6) for η,ηˉ<0\eta,\bar{\eta}<0. Figure (a) is adapted from ref. [39] and (b) is adapted from ref. [48].

W□g(b,ηv,δ)=Wg[b2→b2+ηv−δ2→−b2+ηv+δ2→−b2].(3)W_{\square}^{g}(b,\eta v,\delta)=W^{g}\left[\frac{b}{2}\to\frac{b}{2}+\eta v-\frac{\delta}{2}\to-\frac{b}{2}+\eta v+\frac{\delta}{2}\to-\frac{b}{2}\right]. \tag*{(3)}

A generic notation for gluon beam functions is given by

Ωg/hμνρσ(b,P,ϵ,ηv,δ)=⟨h(P)∣Gμν(b2)W□g(b,ηv,δ)Gρσ(−b2)∣h(P)⟩.(4)\Omega_{g/h}^{\mu\nu\rho\sigma}(b,P,\epsilon,\eta v,\delta)=\langle h(P)|G^{\mu\nu}\left(\frac{b}{2}\right)W_{\square}^{g}(b,\eta v,\delta)G^{\rho\sigma}\left(-\frac{b}{2}\right)|h(P)\rangle. \tag*{(4)}

Here, Gμν(b)G^{\mu\nu}(b) is the gluon field strength tensor. Gluon fields are spatially separated by bb, which is Fourier-conjugate to the struck parton’s momentum, hh denotes a parent hadron with momentum PP, ϵ\epsilon is the UV regulator, and ηv\eta v and δ\delta characterize the Wilson line path. See refs. [46, 47] for the decomposition of gluon TMDs into independent spin structures.

Note that the correlator Ω\Omega characterizes the TMDs in bb-space. When we define a full TMD in terms of beam and soft functions, we generally make a Fourier transform P⋅b→xP\cdot b\to x and write

Ω(b,…)→B(x,b⃗T,…),(5)\Omega(b,\ldots)\to B(x,\vec{b}_{T},\ldots), \tag*{(5)}

where BB indicates the correlator in these new coordinates.

The gluon soft function is defined as

S□g[b,ϵ,ηv,ηˉvˉ]=1Nc2−1⟨0∣Tr⁡[S□g(b,ηv,ηˉvˉ)]∣0⟩,(6)S_{\square}^{g}[b,\epsilon,\eta v,\bar{\eta}\bar{v}]=\frac{1}{N_{c}^{2}-1}\langle0|\operatorname{Tr}\left[S_{\square}^{g}(b,\eta v,\bar{\eta}\bar{v})\right]|0\rangle, \tag*{(6)}

where the trace is over color, and the soft Wilson line is

S□g(b,ηv,ηˉvˉ)=Wg[b2→b2+ηˉvˉ→−b2+ηˉvˉ→−b2→−b2+ηv→b2+ηv→b2],(7)S_{\square}^{g}(b,\eta v,\bar{\eta}\bar{v})=W^{g}\left[\frac{b}{2}\to\frac{b}{2}+\bar{\eta}\bar{v}\to-\frac{b}{2}+\bar{\eta}\bar{v}\to-\frac{b}{2}\to-\frac{b}{2}+\eta v\to\frac{b}{2}+\eta v\to\frac{b}{2}\right], \tag*{(7)}

as shown in Figure 1. The Wilson line S□S_{\square} consists of two beam function staples glued together at the points ±b/2\pm b/2; the long sides of the staples run along the ηˉvˉ\bar{\eta}\bar{v} and ηv\eta v directions. Note that the figure shows the case where b=b⃗Tb=\vec{b}_{T}, and the points are translated to 00 and −b⃗T-\vec{b}_{T}.

We can now define all off-lightcone gluon TMD schemes in terms of the exact same BB and SS:

fg/h[s](x,b⃗T,μ,P,v,vˉ,δ,…)=lim⁡lightcone& UV limitsZUVg(ϵ,μ,v,vˉ,δ,…)lim⁡η→∞Bg/hρσ(x,b⃗T,P,ϵ,ηv,δ)Sg[b,ϵ,ηv,ηvˉ](8)f^{[s]}_{g/h}(x,\vec{b}_{T},\mu,P,v,\bar{v},\delta,\ldots)=\lim_{\substack{\text{lightcone}\\ \&\ \text{UV limits}}}Z_{\mathrm{UV}}^{g}(\epsilon,\mu,v,\bar{v},\delta,\ldots)\lim_{\eta\to\infty}\frac{B^{\rho\sigma}_{g/h}(x,\vec{b}_{T},P,\epsilon,\eta v,\delta)}{\sqrt{S^{g}[b,\epsilon,\eta v,\eta\bar{v}]}} \tag*{(8)}

where the ellipses and ZUVgZ_{\mathrm{UV}}^{g} account for a scheme-dependent renormalization. 1 Each scheme is characterized by a specific choice of the parameters bμb^{\mu}, ηvμ\eta v^{\mu}, PμP^{\mu}, and δμ\delta^{\mu} appearing in the beam and soft functions, as well as the precise nature and order of the lightcone and UV limits in eq. (8). The longitudinal staple length η\eta distinguishes physical TMDs (η=±∞\eta=\pm\infty) from lattice TMDs (finite η\eta).

The remainder of § 1.1 specifies the order of limits and beam/soft function arguments that define three major TMD schemes: the continuum Collins TMD, the lattice quasi-TMD, and the continuum Large Rapidity (LR) TMD. We summarize this information in Table 1, which also includes values taken on by the four-vectors bμb^{\mu}, vμv^{\mu}, δμ\delta^{\mu}, PμP^{\mu}.

Collins TMD (continuum)Quasi-TMD (lattice)
TMDlim⁡ϵ→0ZUVκilim⁡yB→−∞Bi/hSκi\lim_{\epsilon\to0} Z_{\mathrm{UV}}^{\kappa_i}\lim_{y_B\to-\infty}\frac{B_{i/h}}{\sqrt{S^{\kappa_i}}}lim⁡ϵ→0ZUVκiB~i/hSκi\lim_{\epsilon\to0} Z_{\mathrm{UV}}^{\kappa_i}\frac{\widetilde{B}_{i/h}}{\sqrt{S^{\kappa_i}}}
Beam functionΩi/h[b,P,ϵ,−∞nB(yB),b−nb]→FTBi/h\Omega_{i/h}[b,P,\epsilon,-\infty n_B(y_B),b^{-}n_b]\xrightarrow{\mathrm{FT}}B_{i/h}Ωi/h(b~,P~,a,η~z~,b~z~)→FTB~i/h\Omega_{i/h}(\widetilde{b},\widetilde{P},a,\widetilde{\eta}\widetilde{z},\widetilde{b}^{\widetilde{z}})\xrightarrow{\mathrm{FT}}\widetilde{B}_{i/h}
Soft functionSκi[b⊥,ϵ,−∞nA(2yn),−∞nB(2yB)]S^{\kappa_i}[b_{\perp},\epsilon,-\infty n_A(2y_n),-\infty n_B(2y_B)]Sκi[b⊥,a,−η~nA(2yn)∣nA(2yn)∣,−η~nB(2yB)∣nB(2yB)∣]S^{\kappa_i}\left[b_{\perp},a,-\widetilde{\eta}\frac{n_A(2y_n)}{\lvert n_A(2y_n)\rvert},-\widetilde{\eta}\frac{n_B(2y_B)}{\lvert n_B(2y_B)\rvert}\right]
bμb^\mu(0,b−,b⊥)(0,b^{-},b_{\perp})(0,bTx,bTy,b~z~)(0,b_T^x,b_T^y,\widetilde{b}^{\widetilde{z}})
vμv^\mu(−e2yB,1,0⊥)(-e^{2y_B},1,0_{\perp})(0,0,0,−1)(0,0,0,-1)
δμ\delta^\mu(0,b−,0⊥)(0,b^{-},0_{\perp})(0,0,0,b~z~)(0,0,0,\widetilde{b}^{\widetilde{z}})
PμP^\mumh2(eyP,e−yP,0⊥)\frac{m_h}{\sqrt{2}}(e^{y_P},e^{-y_P},0_{\perp})mh(cosh⁡yP~,0,0,sinh⁡yP~)m_h(\cosh y_{\widetilde{P}},0,0,\sinh y_{\widetilde{P}})

Table 1. Definition of the Collins and quasi-TMDs for quarks and gluons. The Large Rapidity and Collins TMDs are identical, except for the order of the limits and expression for ZUVκiZ_{\mathrm{UV}}^{\kappa_i} in the first row.

Collins scheme [6]. The continuum Collins scheme uses spacelike Wilson lines with directions

nAμ(yA)≡naμ−e−2yAnbμ=(1,−e−2yA,0⊥),n_{A}^{\mu}(y_{A})\equiv n_{a}^{\mu}-e^{-2y_{A}}n_{b}^{\mu}=(1,-e^{-2y_{A}},0_{\perp}),
nBμ(yB)≡nbμ−e2yBnaμ=(−e2yB,1,0⊥),(9)n_{B}^{\mu}(y_{B})\equiv n_{b}^{\mu}-e^{2y_{B}}n_{a}^{\mu}=(-e^{2y_{B}},1,0_{\perp}), \tag*{(9)}

parametrized by the rapidities yAy_{A} and yBy_{B}. (See appendix § 5 for our lightcone coordinate conventions.) The Collins gluon TMD for a hadron hh moving along nan_{a} with rapidity yPy_{P} is

fg/h(x,b⃗T,μ,ζ)=lim⁡ϵ→0ZUVg(ϵ,μ,ζ)lim⁡yB→−∞Bg/hCσ(x,b⃗T,ϵ,yP−yB)Sg[b⊥,ϵ,−∞nA(2yn),−∞nB(2yB)](10)f_{g/h}(x,\vec{b}_{T},\mu,\zeta)=\lim_{\epsilon\to0}Z_{\mathrm{UV}}^{g}(\epsilon,\mu,\zeta)\lim_{y_{B}\to-\infty}\frac{B_{g/h}^{C\sigma}(x,\vec{b}_{T},\epsilon,y_{P}-y_{B})}{\sqrt{S^{g}[b_{\perp},\epsilon,-\infty n_{A}(2y_{n}),-\infty n_{B}(2y_{B})]}} \tag*{(10)}

where ∣b⊥∣=bT|b_{\perp}|=b_{T} and the beam function is defined as

Bg/hCρσ(x,b⃗T,ϵ,yP−yB)=∫db−2πe−ib−(xP+)xP+Ωg/h+ρ+σ[b,P,ϵ,−∞nB(yB),b−nb],(11)B_{g/h}^{C\rho\sigma}(x,\vec{b}_{T},\epsilon,y_{P}-y_{B})=\int\frac{db^{-}}{2\pi}\frac{e^{-ib^{-}(xP^{+})}}{xP^{+}}\Omega_{g/h}^{+\rho+\sigma}[b,P,\epsilon,-\infty n_{B}(y_{B}),b^{-}n_{b}], \tag*{(11)}

with Ωg/h+ρ+σ=(nb)μΩg/hμρνσ(nb)ν\Omega_{g/h}^{+\rho+\sigma}=(n_{b})_{\mu}\Omega_{g/h}^{\mu\rho\nu\sigma}(n_{b})_{\nu}. The CS scale ζ\zeta is defined as

ζ=2(xP+e−yn)2=x2mh2e2(yP−yn).(12)\zeta=2(xP^{+}e^{-y_{n}})^{2}=x^{2}m_{h}^{2}e^{2(y_{P}-y_{n})}. \tag*{(12)}

Quasi-TMDs [26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 39]. Collins TMDs cannot be directly discretized and computed on the lattice, as their Wilson lines’ explicit dependence on time induces a sign problem. Instead, one defines numerically-tractable quasi-TMD beam functions using finite-length Wilson lines lying on equal-time paths. The lattice gluon quasi-TMD takes the form

f~i/h[Γ](x,b⃗T,μ,ζ~,xP~z)=lim⁡η~→∞a→0ZUVg(μ,μ~)ZUVg(a,μ~,yn−yB)B~g/hαβρσ(x,b⃗T,a,η~,xP~z)Sg[b⊥,a,−η~nA(2yn)∣nA(2yn)∣,−η~nB(2yB)∣nB(2yB)∣](13)\widetilde{f}_{i/h}^{[\Gamma]}(x,\vec{b}_{T},\mu,\widetilde{\zeta},x\widetilde{P}^{z})=\lim_{\substack{\widetilde{\eta}\to\infty\\ a\to0}}\frac{Z_{\mathrm{UV}}^{g}(\mu,\widetilde{\mu})Z_{\mathrm{UV}}^{g}(a,\widetilde{\mu},y_{n}-y_{B})\widetilde{B}_{g/h}^{\alpha\beta\rho\sigma}(x,\vec{b}_{T},a,\widetilde{\eta},x\widetilde{P}^{z})}{\sqrt{S^{g}\left[b_{\perp},a,-\widetilde{\eta}\frac{n_{A}(2y_{n})}{|n_{A}(2y_{n})|},-\widetilde{\eta}\frac{n_{B}(2y_{B})}{|n_{B}(2y_{B})|}\right]}} \tag*{(13)}

where the bare quasi-beam function reads as

B~g/hαβρσ(x,b⃗T,a,η~,xP~z~)=Nαβρσ∫dz~2πeiz~(xP~z~)xP~z~Ωg/hαβρσ[b~,P~,a,η~z~,b~z~].(14)\widetilde{B}_{g/h}^{\alpha\beta\rho\sigma}(x,\vec{b}_T,a,\widetilde{\eta},x\widetilde{P}^{\widetilde{z}})=N^{\alpha\beta\rho\sigma}\int\frac{d\widetilde{z}}{2\pi}\frac{e^{i\widetilde{z}(x\widetilde{P}^{\widetilde{z}})}}{x\widetilde{P}^{\widetilde{z}}}\Omega_{g/h}^{\alpha\beta\rho\sigma}\left[\widetilde{b},\widetilde{P},a,\widetilde{\eta}\widetilde{z},\widetilde{b}^{\widetilde{z}}\right]. \tag*{(14)}

Here, α\alpha, β\beta, ρ\rho and σ\sigma are generic Lorentz indices, and NαβρσN^{\alpha\beta\rho\sigma} is a trivial normalization factor. The factor ZUVg(a,μ~,yn−yB)Z_{\mathrm{UV}}^{g}(a,\widetilde{\mu},y_n-y_B) renormalizes UV divergences from lattice regularization, and ZUVg(μ,μ~)Z_{\mathrm{UV}}^{g}(\mu,\widetilde{\mu}) matches the result to the MS scheme. The second equality in eq. (13) holds for large P~z\widetilde{P}^{z}, and the variable ζ~\widetilde{\zeta} is

ζ~=(xmheyP~+yB−yn)2≈(2xP~zeyB−yn)2.(15)\widetilde{\zeta}=\left(xm_he^{y_{\widetilde{P}}+y_B-y_n}\right)^2\approx\left(2x\widetilde{P}^{z}e^{y_B-y_n}\right)^2. \tag*{(15)}

Large Rapidity (LR) scheme To facilitate deriving the factorization relation between the quasi- and Collins TMDs, Ref. [39] introduced the Large Rapidity (LR) scheme,

fg/hLR(x,b⃗T,μ,ζ,yP−yB)=lim⁡−yB≫1lim⁡ϵ→0ZUVLR(ϵ,μ,yn−yB)×Bg/hCρσ(x,b⃗T,ϵ,yP−yB)Sg[b⊥,ϵ,−∞nA(2yn),−∞nB(2yB)].(16)\begin{aligned} f_{g/h}^{\mathrm{LR}}(x,\vec{b}_T,\mu,\zeta,y_P-y_B) &=\lim_{-y_B\gg1}\lim_{\epsilon\to0}Z_{\mathrm{UV}}^{\mathrm{LR}}(\epsilon,\mu,y_n-y_B)\\ &\quad\times\frac{B_{g/h}^{C\rho\sigma}(x,\vec{b}_T,\epsilon,y_P-y_B)}{\sqrt{S_g[b_{\perp},\epsilon,-\infty n_A(2y_n),-\infty n_B(2y_B)]}}. \tag*{(16)} \end{aligned}

which differs from the Collins scheme only by its order of ϵ→0\epsilon\to0 and yB→−∞y_B\to-\infty limits. Using Lorentz invariance and the large P~z\widetilde{P}^{z} expansion (−yB≫1-y_B\gg1), one can demonstrate that when ζ~=ζ\widetilde{\zeta}=\zeta or yP~=yP−yBy_{\widetilde{P}}=y_P-y_B, the quasi- and LR TMDs are equivalent at leading order under a large rapidity expansion [39], i.e.,

f~g/h(x,b⃗T,μ,ζ~,xP~z)=fg/hLR(x,b⃗T,μ,ζ~,yP−yB)+O(yBκBeyB).(17)\widetilde{f}_{g/h}(x,\vec{b}_T,\mu,\widetilde{\zeta},x\widetilde{P}^{z})=f_{g/h}^{\mathrm{LR}}(x,\vec{b}_T,\mu,\widetilde{\zeta},y_P-y_B)+\mathcal{O}(y_B^{\kappa_B}e^{y_B}). \tag*{(17)}

where O(yBκBeyB)\mathcal{O}(y_B^{\kappa_B}e^{y_B}) are exponentially suppressed contributions.

Calculation of one-loop gluon matching coefficient

We now compute the one-loop quasi-to-Collins matching coefficient CgC_g for gluons. To do so, we actually carry out a simpler equivalent matching calculation: between the LR and Collins schemes. In the course of proving the quasi-to-Collins factorization formula in eq. (1), [39] showed that the quasi-to-Collins and LR-to-Collins matching coefficients are identical. Note that this matching calculation focuses on beam functions: the quasi-soft function has been chosen to reproduce the Collins soft function as ∣η~∣→∞|\tilde{\eta}| \to\infty.

Let us compute the matrix element of the Collins beam function for a free external gluon state with momentum pμ=(p+,0,0⊥)p^\mu= (p^+,0,0_\perp), transverse polarization vectors ϵ⊥α\epsilon^\alpha_\perp and ϵ⊥β\epsilon^\beta_\perp, as well as color indices aa and bb. We work in Feynman gauge and dimensional regularization with space-time dimension d=4−2ϵd = 4 - 2\epsilon. As we are interested in the leading power gluon TMDs, the Lorentz indices ρ\rho, σ\sigma are both transverse, and for both the LR and Collins schemes μ=ν=+\mu= \nu= +. In our calculation we leave the choice of ρ\rho and σ\sigma unspecified, to make clear that our result applies to all spin-dependent gluon TMDs.

At tree-level, we must compute the diagrams in Figure 2a, which give

Figure 2. The tree-level (a) and one-loop level (b–h) diagrams contributing to gluon beam function in a free gluon state. Since the external gluons are transversely polarized, they cannot be attached to the Wilson lines in the nBn_B direction. While depicted in these diagrams as a straight line, the beam function Wilson line actually traverses an infinitely long staple-shaped path, as shown in Figure 1a.

Ωg/g+ρ+σ(0)(bμ)=δab(p+)2[gρβgσαeip⋅b+(α↔β,p→−p)],(18)\Omega_{g/g}^{+\rho+\sigma(0)}(b^\mu) = \delta^{ab}(p^+)^2 \left[g^{\rho\beta}g^{\sigma\alpha}e^{ip\cdot b} + (\alpha\leftrightarrow\beta, p \to-p)\right], \tag*{(18)}

where the second term in the square bracket comes from the exchange symmetry of bosonic particles. Taking a Fourier transform bz→xpzb^z \to xp^z, the tree-level gluon beam function becomes

xBg/gCαβ(0)(x,b⃗T)=δab[gρβgσαδ(1−x)+(α↔β,x→−x)],(19)xB_{g/g}^{C\alpha\beta(0)}(x,\vec{b}_T) = \delta^{ab}\left[g^{\rho\beta}g^{\sigma\alpha}\delta(1-x) + (\alpha\leftrightarrow\beta, x \to-x)\right], \tag*{(19)}

where we moved the normalization factor 1/x1/x to the left hand side for clarity. We emphasize that to find the matching coefficient, we need only examine terms proportional to δ(1−x)\delta(1-x) since the factorization relation does not involve a convolution in xx.

The one-loop Feynman diagrams that contribute to the beam function are shown in Figure 2. According to Ref. [39], only diagrams containing a rapidity divergence contribute to the matching, significantly simplifying our task. Only figures 2c and 2e could exhibit rapidity divergences, as the other diagrams have Lorentz-covariant loop integrals, which remain unaffected by the large rapidity limit yB→−∞y_B \to-\infty (except for the transformation of tensor structures). Using the Feynman rules in appendix § 6, we express figures 2c and 2e as

Ωg/g,c+ρ+σ(1)(bμ)=(nb)μ(nb)νℓϵ∫ddk(2π)d(−ikσgνδ+ikνgσδ)δfdeik⋅b2(ipρgμβ−ipμgρβ)δbceip⋅b2×(gffea)[gδγ(p−2k)α+gγα(k−2p)δ+gαδ(p+k)γ]×gfedc∫ds γγ′(s)e−i(p−k)⋅γ(s)ik2+i0i(p−k)2+i0+(nb)μ(nb)νℓϵ∫ddk(2π)d(ikρgμδ−ikμgρδ)δfceik⋅b2(−ipσgνα+ipνgσα)δadeip⋅b2×(gfbef)[gδγ(p−2k)β+gγβ(k−2p)δ+gβδ(p+k)γ]×gfedc∫ds γγ′(s)ei(p−k)⋅γ(s)ik2+i0i(p−k)2+i0.(20)\begin{aligned} \Omega_{g/g,c}^{+\rho+\sigma(1)}(b^\mu) ={}& (n_b)_\mu(n_b)_\nu\ell^\epsilon\int\frac{d^d k}{(2\pi)^d}(-ik^\sigma g^{\nu\delta}+ik^\nu g^{\sigma\delta})\delta^{fd}e^{ik\cdot\frac{b}{2}}(ip^\rho g^{\mu\beta}-ip^\mu g^{\rho\beta})\delta^{bc}e^{ip\cdot\frac{b}{2}} \\ &\times(gf^{fea})\left[g^{\delta\gamma}(p-2k)^\alpha+g^{\gamma\alpha}(k-2p)^\delta+g^{\alpha\delta}(p+k)^\gamma\right] \\ &\times gf^{edc}\int ds\,\gamma'_\gamma(s)e^{-i(p-k)\cdot\gamma(s)}\frac{i}{k^2+i0}\frac{i}{(p-k)^2+i0} \\ &+ (n_b)_\mu(n_b)_\nu\ell^\epsilon\int\frac{d^d k}{(2\pi)^d}(ik^\rho g^{\mu\delta}-ik^\mu g^{\rho\delta})\delta^{fc}e^{ik\cdot\frac{b}{2}}(-ip^\sigma g^{\nu\alpha}+ip^\nu g^{\sigma\alpha})\delta^{ad}e^{ip\cdot\frac{b}{2}} \\ &\times(gf^{bef})\left[g^{\delta\gamma}(p-2k)^\beta+g^{\gamma\beta}(k-2p)^\delta+g^{\beta\delta}(p+k)^\gamma\right] \\ &\times gf^{edc}\int ds\,\gamma'_\gamma(s)e^{i(p-k)\cdot\gamma(s)}\frac{i}{k^2+i0}\frac{i}{(p-k)^2+i0}. \tag*{(20)} \end{aligned}

and

Ωg/g,e+ρ+σ(1)(bμ)=(nb)μ(nb)νi∫ddk(2π)d(gfdae)(gναgσγ−gνγgσα)ei(p−k)⋅b2×(ipρgμβ−ipμgρβ)δbceeip⋅b2gfecd∫ds γγ′(s)e−ik⋅γ(s)ik2+i0+(nb)μ(nb)ν∫ddk(2π)d(gfeceb)(gμγgρβ−gμβgργ)e−i(k−p)⋅b2×(−ipσgνα+ipνgσα)δadeeip⋅b2gfecd∫ds γγ′(s)eik⋅γ(s)ik2+i0,(21)\begin{aligned} \Omega_{g/g,e}^{+\rho+\sigma(1)}(b^\mu) = (n_b)_\mu(n_b)_\nu i \int\frac{d^d k}{(2\pi)^d}(g f^{dae})(g^{\nu\alpha}g^{\sigma\gamma} - g^{\nu\gamma}g^{\sigma\alpha})e^{i(p-k)\cdot\frac{b}{2}} \\ &\times\left(ip^{\rho}g^{\mu\beta}-ip^{\mu}g^{\rho\beta}\right)\delta^{bce}e^{ip\cdot\frac{b}{2}}gf^{ecd}\int ds\,\gamma'_{\gamma}(s)e^{-ik\cdot\gamma(s)}\frac{i}{k^{2}+i0} +\left(n_{b}\right)_{\mu}\left(n_{b}\right)_{\nu}\int\frac{d^{d}k}{(2\pi)^{d}}\left(gf^{eceb}\right)\left(g^{\mu\gamma}g^{\rho\beta}-g^{\mu\beta}g^{\rho\gamma}\right)e^{-i(k-p)\cdot\frac{b}{2}} \times\left(-ip^{\sigma}g^{\nu\alpha}+ip^{\nu}g^{\sigma\alpha}\right)\delta^{ade}e^{ip\cdot\frac{b}{2}}gf^{ecd}\int ds\,\gamma'_{\gamma}(s)e^{ik\cdot\gamma(s)}\frac{i}{k^{2}+i0}, \tag*{(21)} \end{aligned}

where the Wilson line takes a staple-shaped path:

γ(s):b2→b2−b−nb2−∞nB→−b2+b−nb2−∞nB→−b2,(22)\gamma(s):\quad\frac{b}{2}\to\frac{b}{2}-\frac{b^{-}n_{b}}{2}-\infty n_{B}\to-\frac{b}{2}+\frac{b^{-}n_{b}}{2}-\infty n_{B}\to-\frac{b}{2}, \tag*{(22)}

and we have suppressed the contributions from diagrams with external gluon lines exchanged (α↔β,p→−p)(\alpha\leftrightarrow\beta,p\to-p) for conciseness. We also use a coupling in the MS‾\overline{\mathrm{MS}} scheme, which introduced the factors ιϵ=(eγE/(4π))ϵ\iota^{\epsilon}=(e^{\gamma_{E}}/(4\pi))^{\epsilon}.

Simplifying the above,

Ωg/g,c+ρ+σ(1)=−2ig2CAp+gασgβριϵ∫ddk(2π)dk+(eik⋅b−eip⋅b)k2(p−k)2[1nB⋅(p−k)+iδ+1nB⋅(p−k)−iδ]+2ig2CAe2yBp+ιϵ∫ddk(2π)deik⋅b−eip⋅bk2(p−k)2[kαkσgβρnB⋅(p−k)+iδ+kβkρgασnB⋅(p−k)−iδ]−2ig2CAp+gασgβριϵ∫ddk(2π)dk+k2(p−k)2(eik⋅b−eip⋅b),(23)\begin{aligned} \Omega_{g/g,c}^{+\rho+\sigma(1)} ={}&-2ig^{2}C_{A}p^{+}g^{\alpha\sigma}g^{\beta\rho}\iota^{\epsilon}\int\frac{d^{d}k}{(2\pi)^{d}}\frac{k^{+}(e^{ik\cdot b}-e^{ip\cdot b})}{k^{2}(p-k)^{2}} \left[\frac{1}{n_{B}\cdot(p-k)+i\delta}+\frac{1}{n_{B}\cdot(p-k)-i\delta}\right]\\ &+2ig^{2}C_{A}e^{2y_{B}}p^{+}\iota^{\epsilon}\int\frac{d^{d}k}{(2\pi)^{d}}\frac{e^{ik\cdot b}-e^{ip\cdot b}}{k^{2}(p-k)^{2}} \left[\frac{k^{\alpha}k^{\sigma}g^{\beta\rho}}{n_{B}\cdot(p-k)+i\delta}+\frac{k^{\beta}k^{\rho}g^{\alpha\sigma}}{n_{B}\cdot(p-k)-i\delta}\right]\\ &-2ig^{2}C_{A}p^{+}g^{\alpha\sigma}g^{\beta\rho}\iota^{\epsilon}\int\frac{d^{d}k}{(2\pi)^{d}}\frac{k^{+}}{k^{2}(p-k)^{2}}\left(e^{ik\cdot b}-e^{ip\cdot b}\right), \tag*{(23)} \end{aligned}
Ωg/g,e+ρ+σ(1)=ig2CAe2yBp+gασgβριϵ∫ddk(2π)d1k2(e−ik⋅b−1)eip⋅b[1nB⋅k+iδ+1nB⋅k−iδ],(24)\Omega_{g/g,e}^{+\rho+\sigma(1)} =ig^{2}C_{A}e^{2y_{B}}p^{+}g^{\alpha\sigma}g^{\beta\rho}\iota^{\epsilon}\int\frac{d^{d}k}{(2\pi)^{d}}\frac{1}{k^{2}}\left(e^{-ik\cdot b}-1\right)e^{ip\cdot b}\left[\frac{1}{n_{B}\cdot k+i\delta}+\frac{1}{n_{B}\cdot k-i\delta}\right], \tag*{(24)}

where we obtain the Wilson line propagators [nB⋅(p−k)±iδ]−1[n_{B}\cdot(p-k)\pm i\delta]^{-1} by integrating over the path parameter ss. Note that in the LR and Collins scheme that the staple Wilson line has infinite extent. We introduce the infinitesimal imaginary part ±iδ\pm i\delta to ensure that the path integral contribution from s=−∞s=-\infty properly vanishes in Feynman gauge.

In the yB→−∞y_{B}\to-\infty limit, the rapidity divergence is in the first line of eq. (23). The second line is suppressed by a factor of e2yBkT2/(p+)2e^{2y_{B}k_{T}^{2}/(p^{+})^{2}} compared to the first line, as kT=∣k⊥∣≪k+,p+k_{T}=|k_{\perp}|\ll k^{+},p^{+} in the power counting, and α,β,ρ,σ=⊥\alpha,\beta,\rho,\sigma=\perp. The last line in eq. (23) is Lorentz covariant and so is not affected by yB→−∞y_{B}\to-\infty; it thus does not contribute to the matching.

After taking a Fourier transform b−→xp+b^{-}\to xp^{+}, the first line of eq. (23) becomes

xBg/g,cρσ(1)(x,b⃗T,ϵ,yP−yB)(25)xB_{g/g,c}^{\rho\sigma(1)}(x,\vec{b}_{T},\epsilon,y_{P}-y_{B}) \tag*{(25)}
=−2ig2CAp+gασgβριϵ∫ddk(2π)dk+e−ik⃗T⋅b⃗Tk2(p−k)2[1nB⋅(p−k)+iδ+1nB⋅(p−k)−iδ]×[δ(k+−xp+)−δ(p+−xp+)]−2ig2CAgασgβρδ(1−x)ιϵ∫ddk(2π)dk+(e−ik⃗T⋅b⃗T−1)k2(p−k)2[1nB⋅(p−k)+iδ+1nB⋅(p−k)−iδ].\begin{aligned} ={}&-2ig^{2}C_{A}p^{+}g^{\alpha\sigma}g^{\beta\rho}\iota^{\epsilon}\int\frac{d^{d}k}{(2\pi)^{d}}\frac{k^{+}e^{-i\vec{k}_{T}\cdot\vec{b}_{T}}}{k^{2}(p-k)^{2}} \left[\frac{1}{n_{B}\cdot(p-k)+i\delta}+\frac{1}{n_{B}\cdot(p-k)-i\delta}\right]\\ &\qquad\times\left[\delta(k^{+}-xp^{+})-\delta(p^{+}-xp^{+})\right]\\ &-2ig^{2}C_{A}g^{\alpha\sigma}g^{\beta\rho}\delta(1-x)\iota^{\epsilon}\int\frac{d^{d}k}{(2\pi)^{d}}\frac{k^{+}(e^{-i\vec{k}_{T}\cdot\vec{b}_{T}}-1)}{k^{2}(p-k)^{2}} \left[\frac{1}{n_{B}\cdot(p-k)+i\delta}+\frac{1}{n_{B}\cdot(p-k)-i\delta}\right]. \end{aligned}

The first term of eq. (25) is a plus function in xx; it is unaffected by the yB→−∞y_{B}\to-\infty limit as long as x≠1x\ne1. The second term contains the contribution to the matching coefficient.

Likewise, after a Fourier transform b−→xp+b^- \to xp^+, the second line of eq. (23) becomes

xBg/eCρσ(1)(x,b⃗T,ϵ,yB−yB)=ig2CAe2yBgασgβρϵℓ∫ddk(2π)d1k2[1nB⋅k+iδ+1nB⋅k−iδ](eik⃗T⋅b⃗T−1)(26)xB_{g/e}^{C^{\rho\sigma}(1)}(x,\vec{b}_T,\epsilon,y_B-y_B) =ig^2C_A e^{2y_B}g^{\alpha\sigma}g^{\beta\rho}\epsilon^\ell\int\frac{d^d k}{(2\pi)^d}\frac{1}{k^2}\left[\frac{1}{n_B\cdot k+i\delta}+\frac{1}{n_B\cdot k-i\delta}\right]\left(e^{i\vec{k}_T\cdot\vec{b}_T}-1\right) \tag*{(26)}
×[δ(k+−(1−x)p+)−δ(p+−xp+)]+ig2CAe2yBgασgβρδ(1−x)p+ϵℓ∫ddk(2π)d1k2[1nB⋅k+iδ+1nB⋅k−iδ](eik⃗T⋅b⃗T−1).\qquad\times\left[\delta(k^+-(1-x)p^+)-\delta(p^+-xp^+)\right] +ig^2C_A e^{2y_B}g^{\alpha\sigma}g^{\beta\rho}\frac{\delta(1-x)}{p^+}\epsilon^\ell\int\frac{d^d k}{(2\pi)^d}\frac{1}{k^2}\left[\frac{1}{n_B\cdot k+i\delta}+\frac{1}{n_B\cdot k-i\delta}\right]\left(e^{i\vec{k}_T\cdot\vec{b}_T}-1\right).

The first term is again a plus function. The second term vanishes under parity transforms k±→−k±k^\pm\to-k^\pm. Therefore, the diagrams in Figure 2e do not contribute to the matching.

All that remains to compute is the term proportional to δ(1−x)\delta(1-x) in eq. (25). After a change of variables k→p−kk\to p-k, the next step is to evaluate the integral

I=ϵℓ∫ddk(2π)dp+−k+k2(p−k)2(eik⃗T⋅b⃗T−1)[1nB⋅k+iδ+1nB⋅k−iδ].(27)I=\epsilon^\ell\int\frac{d^d k}{(2\pi)^d}\frac{p^+-k^+}{k^2(p-k)^2}\left(e^{i\vec{k}_T\cdot\vec{b}_T}-1\right)\left[\frac{1}{n_B\cdot k+i\delta}+\frac{1}{n_B\cdot k-i\delta}\right]. \tag*{(27)}

To simplify the integral we can make a change of variable k−→e−yBk−k^-\to e^{-y_B}k^- and k+→eyBk+k^+\to e^{y_B}k^+ which corresponds to a boost to equal-time with kz=(k+−k−)/2k^z=(k^+-k^-)/\sqrt{2}, and yields

I=ϵℓ2∫ddk(2π)dp′+−k+(k2+i0)[(p′−k)2+i0](eik⃗T⋅b⃗T−1)[1kz−iδ+1kz+iδ],(28)I=\frac{\epsilon^\ell}{\sqrt{2}}\int\frac{d^d k}{(2\pi)^d}\frac{p^{\prime+}-k^+}{(k^2+i0)[(p^\prime-k)^2+i0]}\left(e^{i\vec{k}_T\cdot\vec{b}_T}-1\right)\left[\frac{1}{k^z-i\delta}+\frac{1}{k^z+i\delta}\right], \tag*{(28)}

where p′μ=(p+e−yB,0,0⊥)p^{\prime\mu}=(p^+e^{-y_B},0,0_\perp). Note that the result involves a principal value integral for kzk^z,

12(1kz+iδ+1kz−iδ)=PV⁡(1kz).(29)\frac{1}{2}\left(\frac{1}{k^z+i\delta}+\frac{1}{k^z-i\delta}\right)=\operatorname{PV}\left(\frac{1}{k^z}\right). \tag*{(29)}

We must evaluate the integral in eq. (28) with two different orders of limits.

Taking yB→−∞y_B\to-\infty before ϵ→0\epsilon\to0. We first perform a contour integration in the complex k0k^0 plane, then we integrate over kzk^z to obtain:

I=i4ϵℓ∫dd−2kT(2π)d−1(eik⃗T⋅b⃗T−1)[(2kT2+1pz′2)ln⁡(kT2+2pz′kT2+pz′2+2pz′2kT2)1+kT2/pz′2−4kT2].(30)I=\frac{i}{4}\epsilon^\ell\int\frac{d^{d-2}k_T}{(2\pi)^{d-1}}\left(e^{i\vec{k}_T\cdot\vec{b}_T}-1\right)\left[\left(\frac{2}{k_T^2}+\frac{1}{p_z^{\prime2}}\right)\frac{\ln\left(\frac{k_T^2+2p_z^\prime\sqrt{k_T^2+p_z^{\prime2}}+2p_z^{\prime2}}{k_T^2}\right)}{\sqrt{1+k_T^2/p_z^{\prime2}}}-\frac{4}{k_T^2}\right]. \tag*{(30)}

Here pz′2=p+2e−2yB/2p_z^{\prime2}=p^{+2}e^{-2y_B}/2. The integrand diverges in the pz′→∞p_z^\prime\to\infty or yB→−∞y_B\to-\infty limit, so by taking the limit first we carrying out the power expansion in pz′p_z^\prime before integrating over k⊥k_\perp, and keeping the ln⁡(pz′)\ln(p_z^\prime) singularity, which leads to

IyB→−∞=i4πϵℓ∫dd−2kT(2π)d−21kT2[−2+ln⁡4pz′2kT2](eik⃗T⋅b⃗T−1),=i(4π)2[1ϵ2+1ϵ(2+ln⁡μ24pz′2)−12ln⁡2(bT2μ2b02)+ln⁡(bT2μ2b02)(2+ln⁡μ24pz′2)−π212],(31)\begin{aligned} I_{y_B\to-\infty}=\frac{i}{4\pi}\epsilon^\ell\int\frac{d^{d-2}k_T}{(2\pi)^{d-2}}\frac{1}{k_T^2}\left[-2+\ln\frac{4p_z^{\prime2}}{k_T^2}\right]\left(e^{i\vec{k}_T\cdot\vec{b}_T}-1\right), \\ &=\frac{i}{(4\pi)^2}\left[\frac{1}{\epsilon^2}+\frac{1}{\epsilon}\left(2+\ln\frac{\mu^2}{4p_z^{\prime2}}\right)-\frac{1}{2}\ln^2\left(\frac{b_T^2\mu^2}{b_0^2}\right)+\ln\left(\frac{b_T^2\mu^2}{b_0^2}\right)\left(2+\ln\frac{\mu^2}{4p_z^{\prime2}}\right)-\frac{\pi^2}{12}\right], \tag*{(31)} \end{aligned}

where b0=2e−γEb_0=2e^{-\gamma_E}, and the 1/ϵ21/\epsilon^2 and 1/ϵ1/\epsilon poles are UV divergences. In the MS scheme we carry out the renormalization by simply subtracting all the 1/ϵ1/\epsilon poles.

Taking ϵ→0\epsilon\to0 before yB→−∞y_B \to-\infty. Here, we can directly integrate:

Iϵ→0=ℓϵ2∫ddk(2π)d2p′z−k0(k2+i0)[(p′−k)2+i0](eik⃗T⋅b⃗T−1)[1kz−iδ+1kz+iδ]−ℓϵ∫ddk(2π)d1(k2+i0)[(p′−k)2+i0](eik⃗T⋅b⃗T−1)=i(4π2)[1ϵ−12ln⁡2(4p′zbT)2b02+ln⁡(4p′zbT)2b0+ln⁡bT2μ2b02−2].(32)\begin{aligned} I_{\epsilon\to0}={}&\frac{\ell^\epsilon}{2}\int\frac{\mathrm{d}^d k}{(2\pi)^d}\frac{2p^{\prime z}-k^0}{(k^2+i0)[(p^\prime-k)^2+i0]}\left(e^{i\vec{k}_T\cdot\vec{b}_T}-1\right)\left[\frac{1}{k^z-i\delta}+\frac{1}{k^z+i\delta}\right]\\ &-\ell^\epsilon\int\frac{\mathrm{d}^d k}{(2\pi)^d}\frac{1}{(k^2+i0)[(p^\prime-k)^2+i0]}\left(e^{i\vec{k}_T\cdot\vec{b}_T}-1\right)\\ ={}&\frac{i}{(4\pi^2)}\left[\frac{1}{\epsilon}-\frac{1}{2}\ln^2\frac{(4p^{\prime z}b_T)^2}{b_0^2}+\ln\frac{(4p^{\prime z}b_T)^2}{b_0}+\ln\frac{b_T^2\mu^2}{b_0^2}-2\right]. \tag*{(32)} \end{aligned}

Again the 1/ϵ1/\epsilon pole is a UV divergence, and is subtracted for the MS renormalized result.

Finally we note that the one-loop Collins soft function is the same for quarks and gluons, so can be obtained from the quark result [29] with the replacement of CF→CAC_F\to C_A,

SCg,(1)(bT,ϵ,2yn,2yB)=αsCA2πℓϵΓ(−ϵ)4ϵ(bT2μ2)ϵ[2(yn−yB)(1+e2(yB−yn))1−e2(yB−yn)−2].(33)S_C^{g,(1)}(b_T,\epsilon,2y_n,2y_B)=\frac{\alpha_s C_A}{2\pi}\ell^\epsilon\frac{\Gamma(-\epsilon)}{4^\epsilon}(b_T^2\mu^2)^\epsilon\left[\frac{2(y_n-y_B)(1+e^{2(y_B-y_n)})}{1-e^{2(y_B-y_n)}}-2\right]. \tag*{(33)}

Here we use the shorthand SCg(bT,ϵ,yA,yB)=Sg[b⊥,ϵ,−∞nA(yA),−∞nB(yB)]S_C^g(b_T,\epsilon,y_A,y_B)=S^g[b_\perp,\epsilon,-\infty n_A(y_A),-\infty n_B(y_B)]. This result is not affected by the order of ϵ→0\epsilon\to0 and yB→−∞y_B\to-\infty limits, i.e.,

lim⁡ϵ→0−yB≫1SCg,(1)(bT,ϵ,2yn,2yB)=αsCA2π(1ϵ+ln⁡bT2μ2b02)[2−2(yn−yB)].(34)\lim_{\substack{\epsilon\to0\\-y_B\gg1}}S_C^{g,(1)}(b_T,\epsilon,2y_n,2y_B)=\frac{\alpha_s C_A}{2\pi}\left(\frac{1}{\epsilon}+\ln\frac{b_T^2\mu^2}{b_0^2}\right)[2-2(y_n-y_B)]. \tag*{(34)}

Therefore this result cancels out for the matching between the LR and Collins schemes.

Matching coefficient. The difference between the renormalized LR and Collins scheme results in eqs. (31) and (32) gives

Iϵ→0ren−IyB→−∞ren=−i(4π2)[12ln⁡2μ24pz′2+ln⁡μ24pz′2+2−π212],(35)I_{\epsilon\to0}^{\mathrm{ren}}-I_{y_B\to-\infty}^{\mathrm{ren}}=-\frac{i}{(4\pi^2)}\left[\frac{1}{2}\ln^2\frac{\mu^2}{4p_z^{\prime2}}+\ln\frac{\mu^2}{4p_z^{\prime2}}+2-\frac{\pi^2}{12}\right], \tag*{(35)}

which is purely of ultraviolet origin, and hence is perturbative at a scale μ2∼4pz′2\mu^2\sim4p_z^{\prime2}. Thus the one-loop matching coefficient between the LR and Collins TMD is given by

Cg(p′z,μ)=1−αs(μ)CA4π[ln⁡2μ24pz′2+2ln⁡μ24pz′2+4−π26].(36)C_g(p^{\prime z},\mu)=1-\frac{\alpha_s(\mu)C_A}{4\pi}\left[\ln^2\frac{\mu^2}{4p_z^{\prime2}}+2\ln\frac{\mu^2}{4p_z^{\prime2}}+4-\frac{\pi^2}{6}\right]. \tag*{(36)}

Note that here 4pz′2=2(p+)2e−2yB4p_z^{\prime2}=2(p^+)^2e^{-2y_B}. For a hadron with momentum PP, we must replace p+→xP+p^+\to xP^+, so that

4pz′2⟶2(xP+)2e−2yB≈4x2mh2sinh⁡2(yP−yB)=ζLR(37)4p_z^{\prime2}\longrightarrow2(xP^+)^2e^{-2y_B}\approx4x^2m_h^2\sinh^2(y_P-y_B)=\zeta_{\mathrm{LR}} \tag*{(37)}

in the limit of −yB≫1-y_B\gg1. Therefore, identifying ζLR=(2xP~z)2\zeta_{\mathrm{LR}}=(2x\widetilde{P}^z)^2, we find that the final gluon matching coefficient for the quasi-TMD is:

Cg(xP~z,μ)=1−αsCA4π[ln⁡2μ2(2xP~z)2+2ln⁡μ2(2xP~z)2+4−π26].(38)C_g(x\widetilde{P}^z,\mu)=1-\frac{\alpha_s C_A}{4\pi}\left[\ln^2\frac{\mu^2}{(2x\widetilde{P}^z)^2}+2\ln\frac{\mu^2}{(2x\widetilde{P}^z)^2}+4-\frac{\pi^2}{6}\right]. \tag*{(38)}

Interestingly this result is identical to the quark result CqC_q [29], up to changing Casimirs from CF→CAC_F\to C_A. This occurs because the entire contribution to the matching comes from the part of the Wilson line diagrams associated with the rapidity divergence. The well-known (generalized) Casimir scaling of the cusp anomalous dimension and CS kernel have been validated to four-loop order [49, 50, 51, 52, 53, 54]. The fact that the full one-loop matching and cusp-anomalous dimension obey Casimir scaling, implies that the full tower of next-to-leading-logarithms for CgC_g will obey Casimir scaling as well [39]. This hints at a relationship for quasi-TMD matching coefficients at higher orders.

Conclusion

In this paper, we calculated the one-loop gluon matching coefficient CgC_g in the factorization formula relating quasi and Collins TMDs, marking an important step towards calculating gluon TMDs from lattice QCD. As the matching coefficient is independent of spin structure and there is no quark-gluon mixing, our result enables lattice calculation of all eight leading-twist gluon TMDs. Given the encouraging first results for quark TMDs, it is reasonable to expect that lattice QCD will provide highly useful nonperturbative input for gluon TMDs and their Collins-Soper evolution, neither of which has yet been extracted from experiments.

Acknowledgments

This work is dedicated to Markus Ebert, who has become more industrious. This work was supported by the U.S. Department of Energy, Office of Science, Office of Nuclear Physics, from DE-SC0011090, DE-AC02-06CH11357 and within the framework of the TMD Topical Collaboration. I.S. was also supported in part by the Simons Foundation through the Investigator grant 327942. S.T.S. was partially supported by the U.S. National Science Foundation through a Graduate Research Fellowship under Grant No. 1745302. Y.Z. is partially supported by an LDRD initiative at Argonne National Laboratory under Project No. 2020-0020.

Lightcone coordinate conventions

We work in a frame where the hadron momentum PP is close to the lightlike unit vectors

naμ=12(1,0,0,1),nbμ=12(1,0,0,−1),(39)n_a^\mu= \frac{1}{\sqrt{2}}(1,0,0,1), \qquad n_b^\mu= \frac{1}{\sqrt{2}}(1,0,0,-1), \tag*{(39)}

which obey na2=nb2=0n_a^2 = n_b^2 = 0 and na⋅nb=1n_a \cdot n_b = 1. We can decompose a four-vector pμp^\mu is

pμ=(p+,p−,p⊥)=p+naμ+p−nbμ+p⊥μ,(40)p^\mu= (p^+,p^-,p_\perp) = p^+ n_a^\mu+ p^- n_b^\mu+ p_\perp^\mu, \tag*{(40)}

where p±=(p0±pz)/2p^\pm= (p^0 \pm p^z)/\sqrt{2} and p⊥μ=(0,px,py,0)=(0,p⃗T,0)p_\perp^\mu= (0,p^x,p^y,0) = (0,\vec{p}_T,0). Here p⊥μp_\perp^\mu is a Minkowski vector, and p⃗T\vec{p}_T is the corresponding Euclidean vector with magnitude pT≡(p⃗T 2)1/2=(−p⊥2)1/2p_T \equiv(\vec{p}_T^{\,2})^{1/2} = (-p_\perp^2)^{1/2}.

Feynman rules

The gluon field strength tensor is defined as igGaμντa=[Dμ,Dν]igG_{a}^{\mu\nu}\tau^{a} = [D^{\mu},D^{\nu}], where iDμ=i∂μ+gAaμτaiD^{\mu} = i\partial^{\mu} + gA_{a}^{\mu}\tau^{a} and τa\tau^{a} is the SU(3) generator in the fundamental representation. The Feynman rules for the local, three-gluon, and gluon-Wilson line vertices are:

local:Gaμν(b)=(−ikμgνα+ikνgμα)δabe−ik⋅b,three-gluon:Gaμν(b)=gfabc[gμαgνβ−gμβgνα]e−i(k+l)⋅b,gluon-Wilson line:=gfabc[gαβ(k1−k2)γ+gβγ(k2−k3)α+gγα(k3−k1)β],=gfabcγ(s)′μe−ik⋅γ(s).(41)\begin{aligned} \text{local:}\quad&G_{a}^{\mu\nu}(b) = (-ik^{\mu}g^{\nu\alpha} + ik^{\nu}g^{\mu\alpha})\delta^{ab}e^{-ik\cdot b},\\ \text{three-gluon:}\quad&G_{a}^{\mu\nu}(b) = gf^{abc}\left[g^{\mu\alpha}g^{\nu\beta} - g^{\mu\beta}g^{\nu\alpha}\right]e^{-i(k+l)\cdot b},\\ \text{gluon-Wilson line:}\quad& = gf^{abc}\left[g^{\alpha\beta}(k_{1}-k_{2})^{\gamma} + g^{\beta\gamma}(k_{2}-k_{3})^{\alpha} + g^{\gamma\alpha}(k_{3}-k_{1})^{\beta}\right],\\ & = gf^{abc}\gamma'^{\mu}_{(s)}e^{-ik\cdot\gamma_{(s)}}. \tag*{(41)} \end{aligned}

References

  1. [1]A. Accardi et al., Electron Ion Collider: The Next QCD Frontier: Understanding the glue that binds us all, Eur. Phys. J. A 52 (2016) 268 [1212.1701].
  2. [2]R. Abdul Khalek et al., Science Requirements and Detector Concepts for the Electron-Ion Collider: EIC Yellow Report, 2103.05419.
  3. [3]J. C. Collins and D. E. Soper, Back-To-Back Jets: Fourier Transform from B to K-Transverse, Nucl. Phys. B 197 (1982) 446.DOI
  4. [4]S. M. Aybat and T. C. Rogers, TMD Parton Distribution and Fragmentation Functions with QCD Evolution, Phys. Rev. D 83 (2011) 114042 [1101.5057].DOI
  5. [5]S. Catani, L. Cieri, D. de Florian, G. Ferrera and M. Grazzini, Vector boson production at hadron colliders: hard-collinear coefficients at the NNLO, Eur. Phys. J. C 72 (2012) 2195 [1209.0158].DOI
  6. [6]J. Collins, Foundations of perturbative QCD, Cambridge monographs on particle physics, nuclear physics, and cosmology 32 (2011) 1.
  7. [7]T. Gehrmann, T. Luebbert and L. L. Yang, Calculation of the transverse parton distribution functions at next-to-next-to-leading order, JHEP 06 (2014) 155 [1403.6451].DOI
  8. [8]M. G. Echevarria, I. Scimemi and A. Vladimirov, Unpolarized Transverse Momentum Dependent Parton Distribution and Fragmentation Functions at next-to-next-to-leading order, JHEP 09 (2016) 004 [1604.07869].DOI
  9. [9]T. Lübbert, J. Oredsson and M. Stahlhofen, Rapidity renormalized TMD soft and beam functions at two loops, JHEP 03 (2016) 168 [1602.01829].DOI
  10. [10]M.-X. Luo, X. Wang, X. Xu, L. L. Yang, T.-Z. Yang and H. X. Zhu, Transverse Parton Distribution and Fragmentation Functions at NNLO: the Quark Case, JHEP 10 (2019) 083 [1908.03831].DOI
  11. [11]M.-X. Luo, T.-Z. Yang, H. X. Zhu and Y. J. Zhu, Quark Transverse Parton Distribution at the Next-to-Next-to-Next-to-Leading Order, Phys. Rev. Lett. 124 (2020) 092001 [1912.05778].DOI
  12. [12]M.-x. Luo, T.-Z. Yang, H. X. Zhu and Y. J. Zhu, Unpolarized quark and gluon TMD PDFs and FFs at N3LO, JHEP 06 (2021) 115 [2012.03256].DOI
  13. [13]M. A. Ebert, B. Mistlberger and G. Vita, Transverse momentum dependent PDFs at N3LO, JHEP 09 (2020) 146 [2006.05329].DOI
  14. [14]A. Bacchetta, F. Delcarro, C. Pisano, M. Radici and A. Signori, Extraction of partonic transverse momentum distributions from semi-inclusive deep-inelastic scattering, Drell-Yan and Z-boson production, JHEP 06 (2017) 081 [1703.10157].DOI
  15. [15]I. Scimemi and A. Vladimirov, Systematic analysis of double-scale evolution, JHEP 08 (2018) 003 [1803.11089].DOI
  16. [16]G. Lustermans, J. K. L. Michel, F. J. Tackmann and W. J. Waalewijn, Joint two-dimensional resummation in qT and 0-jettiness at NNLL, JHEP 03 (2019) 124 [1901.03331].DOI
  17. [17]M. A. Ebert, J. K. L. Michel, I. W. Stewart and Z. Sun, Disentangling Long and Short Distances in Momentum-Space TMDs, 2201.07237.DOI
  18. [18]I. Scimemi and A. Vladimirov, Non-perturbative structure of semi-inclusive deep-inelastic and Drell-Yan scattering at small transverse momentum, JHEP 06 (2020) 137 [1912.06532].DOI
  19. [19]A. Bacchetta, V. Bertone, C. Bissolotti, G. Bozzi, F. Delcarro, F. Piacenza et al., Transverse-momentum-dependent parton distributions up to N3LL from Drell-Yan data, JHEP 07 (2020) 117 [1912.07550].DOI
  20. [20]P. Hagler, B. U. Musch, J. W. Negele and A. Schafer, Intrinsic quark transverse momentum in the nucleon from lattice QCD, EPL 88 (2009) 61001 [0908.1283].DOI
  21. [21]B. U. Musch, P. Hagler, J. W. Negele and A. Schafer, Exploring quark transverse momentum distributions with lattice QCD, Phys. Rev. D 83 (2011) 094507 [1011.1213].DOI
  22. [22]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 (2012) 094510 [1111.4249].DOI
  23. [23]M. Engelhardt, P. Hägler, B. Musch, J. Negele and A. Schäfer, Lattice QCD study of the Boer-Mulders effect in a pion, Phys. Rev. D 93 (2016) 054501 [1506.07826].DOI
  24. [24]B. Yoon, T. Bhattacharya, M. Engelhardt, J. Green, R. Gupta, P. Hägler et al., 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, 11, 2015, 1601.05717.arxiv.org/abs/1601.05717
  25. [25]B. Yoon, M. Engelhardt, R. Gupta, T. Bhattacharya, J. R. Green, B. U. Musch et al., Nucleon Transverse Momentum-dependent Parton Distributions in Lattice QCD: Renormalization Patterns and Discretization Effects, Phys. Rev. D 96 (2017) 094508 [1706.03406].DOI
  26. [26]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 (2015) 074009 [1405.7640].DOI
  27. [27]X. Ji, L.-C. Jin, F. Yuan, J.-H. Zhang and Y. Zhao, Transverse momentum dependent parton quasidistributions, Phys. Rev. D 99 (2019) 114006 [1801.05930].DOI
  28. [28]M. A. Ebert, I. W. Stewart and Y. Zhao, Determining the Nonperturbative Collins-Soper Kernel From Lattice QCD, Phys. Rev. D 99 (2019) 034505 [1811.00026].DOI
  29. [29]M. A. Ebert, I. W. Stewart and Y. Zhao, Towards Quasi-Transverse Momentum Dependent PDFs Computable on the Lattice, JHEP 09 (2019) 037 [1901.03685].DOI
  30. [30]M. A. Ebert, I. W. Stewart and Y. Zhao, Renormalization and Matching for the Collins-Soper Kernel from Lattice QCD, JHEP 03 (2020) 099 [1910.08569].DOI
  31. [31]X. Ji, Y. Liu and Y.-S. Liu, TMD soft function from large-momentum effective theory, Nucl. Phys. B 955 (2020) 115054 [1910.11415].DOI
  32. [32]X. Ji, Y. Liu and Y.-S. Liu, Transverse-momentum-dependent parton distribution functions from large-momentum effective theory, Phys. Lett. B 811 (2020) 135946 [1911.03840].DOI
  33. [33]M. A. Ebert, S. T. Schindler, I. W. Stewart and Y. Zhao, One-loop Matching for Spin-Dependent Quasi-TMDs, JHEP 09 (2020) 099 [2004.14831].DOI
  34. [34]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 (2021) 074005 [2011.13397].DOI
  35. [35]X. Ji and Y. Liu, Computing light-front wave functions without light-front quantization: A large-momentum effective theory approach, Phys. Rev. D 105 (2022) 076014 [2106.05310].DOI
  36. [36]X. Ji, Parton Physics on a Euclidean Lattice, Phys. Rev. Lett. 110 (2013) 262002 [1305.1539].pubmed.ncbi.nlm.nih.gov/23848864
  37. [37]X. Ji, Parton Physics from Large-Momentum Effective Field Theory, Sci. China Phys. Mech. Astron. 57 (2014) 1407 [1404.6680].arxiv.org/abs/1404.6680
  38. [38]X. Ji, Y.-S. Liu, Y. Liu, J.-H. Zhang and Y. Zhao, Large-momentum effective theory, Rev. Mod. Phys. 93 (2021) 035005 [2004.03543].DOI
  39. [39]M. A. Ebert, S. T. Schindler, I. W. Stewart and Y. Zhao, Factorization connecting continuum & lattice TMDs, JHEP 04 (2022) 178 [2201.08401].DOI
  40. [40]P. Shanahan, M. L. Wagman and Y. Zhao, Nonperturbative renormalization of staple-shaped Wilson line operators in lattice QCD, Phys. Rev. D 101 (2020) 074505 [1911.00800].DOI
  41. [41]P. Shanahan, M. Wagman and Y. Zhao, Collins-Soper kernel for TMD evolution from lattice QCD, Phys. Rev. D 102 (2020) 014511 [2003.06063].DOI
  42. [42]M. Schlemmer, A. Vladimirov, C. Zimmermann, M. Engelhardt and A. Schäfer, Determination of the Collins-Soper Kernel from Lattice QCD, JHEP 08 (2021) 004 [2103.16991].
  43. [43]LATTICE PARTON collaboration, Q.-A. Zhang et al., Lattice QCD Calculations of Transverse-Momentum-Dependent Soft Function through Large-Momentum Effective Theory, Phys. Rev. Lett. 125 (2020) 192001 [2005.14572].DOI
  44. [44]Y. Li et al., Lattice QCD Study of Transverse-Momentum Dependent Soft Function, Phys. Rev. Lett. 128 (2022) 062002 [2106.13027].DOI
  45. [45]P. Shanahan, M. Wagman and Y. Zhao, Lattice QCD calculation of the Collins-Soper kernel from quasi-TMDPDFs, Phys. Rev. D 104 (2021) 114502 [2107.11930].DOI
  46. [46]P. J. Mulders and J. Rodrigues, Transverse momentum dependence in gluon distribution and fragmentation functions, Phys. Rev. D 63 (2001) 094021 [hep-ph/0009343].DOI
  47. [47]M. G. Echevarria, T. Kasemets, P. J. Mulders and C. Pisano, QCD evolution of (un)polarized gluon TMDPDFs and the Higgs q_T-distribution, JHEP 07 (2015) 158 [1502.05354].DOI
  48. [48]Y. Li, D. Neill and H. X. Zhu, An exponential regulator for rapidity divergences, Nucl. Phys. B 960 (2020) 115193 [1604.00392].arxiv.org/abs/1604.00392
  49. [49]G. P. Korchemsky and A. V. Radyushkin, Renormalization of the Wilson Loops Beyond the Leading Order, Nucl. Phys. B 283 (1987) 342.DOI
  50. [50]S. Moch, J. A. M. Vermaseren and A. Vogt, The Three loop splitting functions in QCD: The Nonsinglet case, Nucl. Phys. B 688 (2004) 101 [hep-ph/0403192].
  51. [51]J. M. Henn, G. P. Korchemsky and B. Mistlberger, The full four-loop cusp anomalous dimension in N = 4 super Yang-Mills and QCD, JHEP 04 (2020) 018 [1911.10174].
  52. [52]A. von Manteuffel, E. Panzer and R. M. Schabinger, Cusp and collinear anomalous dimensions in four-loop QCD from form factors, Phys. Rev. Lett. 124 (2020) 162001 [2002.04617].DOI
  53. [53]I. Moult, H. X. Zhu and Y. J. Zhu, The Four Loop QCD Rapidity Anomalous Dimension, 2205.02249.DOI
  54. [54]C. Duhr, B. Mistlberger and G. Vita, The Four-Loop Rapidity Anomalous Dimension and Event Shapes to Fourth Logarithmic Order, 2205.02242.

Paper details

vlnc:W_hkhkekyy
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