betaFeedback

WorksPhysics and Astronomy

Power Counting to Saturation

Abstract

We present a description of saturation in small xx deep inelastic scattering from power counting in a top-down effective theory derived from QCD. A factorization formula isolates the universal physics of the nucleus at leading power in xx. The onset of saturation is then understood as a breakdown in the expansion in an emergent power counting parameter, which is defined by the matrix element of a gauge invariant operator. We identify a new radiation mode, which enables us to extend previous literature by distinguishing the appearance of the saturation scale from the transition to non-linear evolution.

One of the primary objectives of a future Electron Ion Collider is an investigation of the partonic structure of hadrons or large nuclei at high energies. The partonic density at such high energies is enhanced due to quantum radiative corrections leading to the phenomenon called saturation [1, 2, 3]. The current understanding of saturation is via the Color Glass Condensate (CGC) [4, 5, 6] which treats the hadron/nucleus as a classical source for small xx gluons. This is often thought of as a bottom-up effective theory [7], whose degrees of freedom consist of color sources ρ\rho at large xx with momenta above a cutoff, k+>Λ+k^+ > \Lambda^+, and gauge fields AμA^\mu at small xx with k+<Λ+k^+ < \Lambda^+. The input is a gauge invariant distribution WΛ+[ρ]W_{\Lambda^+}[\rho] which determines the probability of a configuration ρ\rho, and at lowest order one solves classical Yang-Mills equations for AμA^\mu with ρ\rho as a source. This formalism incorporates soft gluon radiative corrections via the non-linear Balitsky-JIMWLK or the Balitsky-Kovchegov (BK)(at large NCN_C) [8, 9, 10, 11, 12, 13, 14, 15] equations which describe the evolution in the cutoff Λ+\Lambda^+.

Our understanding of deep inelastic scattering (DIS) at large xx rests on the notion of non-perturbative factorization which enables us to define operators which separate the universal physics of the hadron from process dependent short distance physics [16, 17]. This can be thought of as a top-down effective field theory derived from QCD [18], where the power counting expansion parameter is ΛQCD/Q\Lambda_{\mathrm{QCD}}/Q. Here Q2=−q2>0Q^2=-q^2>0 where qμq^\mu is the momentum transfer of the virtual photon in the electron-proton collision, x=Q2/(2P⋅q)x=Q^2/(2P\cdot q) where PμP^\mu is the initial proton’s momentum, and we define H=(P+q)2H=(P+q)^2, the photon-nucleus invariant mass.

The goal of this letter is to derive an effective field theory description of small xx saturation physics from the top down, and connect it to the CGC formalism. Starting from QCD we setup a factorization framework for DIS at small xx at all orders in the strong coupling and leading power in xx. This leads to a definition for the universal physics of the nuclear medium as an operator matrix element in the nuclear state. Next we show that we can understand the appearance of the saturation scale in terms of a breakdown in an emergent power counting parameter that appears due to multiple probe-medium interactions.

Finally we give a distinct power counting argument to understand how the form of the factorization changes as we enter the regime with non-linear evolution in ln⁡x\ln x, as we approach the Froissart bound. Apart from the kinematical scales that we will discuss in the next section, we foreshadow the importance of three emergent length scales that are crucial for this description. The first is LDL_D, the average distance beyond which color sources in the medium are uncorrelated. The second is the time scale tct_c over which the partons in the probe maintain quantum coherence. Finally we have the mean free path of the dipole probe λmfp\lambda_{\mathrm{mfp}} which is the average distance between successive interactions of the probe with the medium.

Kinematics and Factorization: We want to describe the small x≪1x\ll1 region or equivalently ΛQCD2≪Q2≪2P⋅q\Lambda_{\mathrm{QCD}}^2\ll Q^2\ll2P\cdot q, which is the high energy limit of QCD where H≈2P⋅qH\approx2P\cdot q. Without loss of generality, we can orient our axes so that qμq^\mu has no transverse momentum and work in a frame with (q−,q+,q⊥)=(−Q,−Q,0⊥)(q^-,q^+,q_\perp)=(-Q,-Q,0_\perp), where we decompose qμ=q−nμ2+q+nˉμ2+q⊥μq^\mu=q^- \frac{n^\mu}{2}+q^+\frac{\bar n^\mu}{2}+q_\perp^\mu, with nμ=(1,0,0,1)n^\mu=(1,0,0,1), nˉμ=(1,0,0,−1)\bar n^\mu=(1,0,0,-1), and ⊥\perp refers to momenta in the xx-yy plane. For a proton of mass MpM_p, the momentum components in this frame are (Mp2Q/H,H/Q,0)(M_p^2Q/H,H/Q,0).

The dominant interaction is the photon fluctuating into a quark-anti quark dipole which then scatters off the hadron/nucleus [19]. The medium partons exchange a transverse momentum of order QQ with the dipole and scatter by an angle θ∼qT/E∼Q2/H=x≪1\theta\sim q_T/E\sim Q^2/H=x\ll1 which is the power counting parameter of the Effective Field Theory(EFT).

In terms of this expansion parameter, the partons in the dipole scale as the soft mode ps∼HQ(x,x,x)p_s\sim\frac{H}{Q}(x,x,x) while the partons in the medium scale as an nˉ\bar n collinear mode pnˉ∼HQ(x2,1,x)p_{\bar n}\sim\frac{H}{Q}(x^2,1,x). Both these modes are propagating degrees of freedom with an invariant mass p2∼Q2p^2\sim Q^2. This stage of the EFT is therefore designed to describe IR physics at the scale QQ while separating out the physics at the scale H/QH/Q. The small angle scattering between these modes is mediated by a Glauber mode whose scaling is fixed by the constraint that both the dipole and medium partons must retain their scaling in any interaction. Therefore, we can write pG∼HQ(x2,x,x)p_G\sim\frac{H}{Q}(x^2,x,x). These same EFT modes are used in [20] to derive small-xx factorization for DIS with applications to ln⁡x\ln x resummation for DIS coefficient functions and anomalous dimensions.

The QCD action can be systematically expanded in θ=x\theta= x within the Soft Collinear Effective Theory (SCET) [18, 21, 22, 23, 24], as shown in [25]. This allows us to write an effective Hamiltonian at leading power in xx

Heff=Hs+Hnˉ+HnˉsG+eqqˉsγμqsLμ+…(1)H_{\mathrm{eff}} = H_s + H_{\bar n} + H_{\bar n s}^{G} + e_q \bar q_s \gamma_\mu q_s L^\mu+ \ldots \tag*{(1)}

This has an effective hard interaction mediated by a photon between the electron current LμL^\mu and the soft quark current, where eqe_q is the electric charge of the quark. Soft and collinear modes of SCET are decoupled in the soft HsH_s and collinear HnˉH_{\bar n} Hamiltonians, but couple through the Glauber Hamiltonian density HnˉsGH_{\bar n s}^{G} that mediates forward scattering and has the form

HnˉsG=CG∑i,j∈{q,g}Onˉsij,Onˉsij=OnˉiB1P⊥2OsjB,(2)H_{\bar n s}^{G} = C_G \sum_{i,j\in\{q,g\}} \mathcal{O}_{\bar n s}^{ij}, \qquad\mathcal{O}_{\bar n s}^{ij} = \mathcal{O}_{\bar n}^{iB} \frac{1}{\mathcal{P}_{\perp}^{2}} \mathcal{O}_{s}^{jB}, \tag*{(2)}

where CG(μ)=8παs(μ)C_G(\mu) = 8\pi\alpha_s(\mu) and P⊥μ\mathcal{P}_{\perp}^{\mu} is a derivative operator in the ⊥\perp direction. The gauge invariant nˉ\bar n collinear and soft operators are

OnˉqB=χˉnˉTB̸nˉ2χnˉ,OnˉgB=i2fBCDBnˉ⊥μCn̸2(P+P†)Bnˉ⊥D,\mathcal{O}_{\bar n}^{qB} = \bar\chi_{\bar n} T^B \frac{\not{\bar n}}{2}\chi_{\bar n}, \qquad\mathcal{O}_{\bar n}^{gB} = \frac{i}{2} f^{BCD} \mathcal{B}_{\bar n\perp\mu}^{C} \frac{\not n}{2} \left(\mathcal{P}+\mathcal{P}^{\dagger}\right)\mathcal{B}_{\bar n\perp}^{D},
OsqB=χˉsTB̸nˉ2χs,OsgB=i2fBCDBs⊥C̸nˉ2(P+P†)Bs⊥D.(3)\mathcal{O}_{s}^{qB} = \bar\chi_s T^B \frac{\not{\bar n}}{2}\chi_s, \qquad\mathcal{O}_{s}^{gB} = \frac{i}{2} f^{BCD} \mathcal{B}_{s\perp}^{C} \frac{\not{\bar n}}{2}\left(\mathcal{P}+\mathcal{P}^{\dagger}\right)\mathcal{B}_{s\perp}^{D}. \tag*{(3)}

which are built out of the gauge invariant building blocks

χnˉ=Wnˉ†ξnˉ,Wnˉ=FT Pexp⁡{ig∫−∞0ds n⋅Anˉ(x+ns)},\chi_{\bar n} = W_{\bar n}^{\dagger}\xi_{\bar n}, \qquad W_{\bar n} = \mathrm{FT}\,\mathbf{P}\exp\left\{ig\int_{-\infty}^{0} ds\,n\cdot A_{\bar n}(x+n s)\right\},
χs=Snˉ†ξs,Snˉ=FT Pexp⁡{ig∫−∞0ds nˉ⋅As(x+snˉ)},\chi_s = S_{\bar n}^{\dagger}\xi_s, \qquad S_{\bar n} = \mathrm{FT}\,\mathbf{P}\exp\left\{ig\int_{-\infty}^{0} ds\,\bar n\cdot A_s(x+s\bar n)\right\},
Bnˉ⊥CμTC=1g[Wnˉ†iDnˉμWnˉ],Bs⊥CμTC=1g[Snˉ†iDs⊥μSnˉ].(4)\mathcal{B}_{\bar n\perp}^{C\mu}T^C = \frac{1}{g}\left[W_{\bar n}^{\dagger}iD_{\bar n}^{\mu}W_{\bar n}\right], \qquad\mathcal{B}_{s\perp}^{C\mu}T^C = \frac{1}{g}\left[S_{\bar n}^{\dagger}iD_{s\perp}^{\mu}S_{\bar n}\right]. \tag*{(4)}

FT stands for Fourier transform. These operators encode bare quark and gluons dressed by Wilson lines. Since the dipole probe can interact multiple times with the medium, we can think of it as an open quantum system evolving in a nuclear environment. We therefore follow

the evolution of the reduced density matrix for our probe. The initial state density matrix can be written as

ρ(0)=∣e−⟩⟨e−∣⊗ρA(5)\rho(0) = |e^-\rangle\langle e^-| \otimes\rho_A \tag*{(5)}

where the electron is disentangled from the nuclear medium (ρA\rho_A). We will evolve this with HeffH_{\mathrm{eff}} and measure the properties of the electron, being inclusive over the hadronic final states. In DIS we write the cross section differential in Q2Q^2 and xx, which we indicate with a measurement M\mathcal{M} which acts on the phase space of the final state e−e^-. Since the Glauber mode couples the soft and collinear sectors, we have to expand out the density matrix element order by order in the Glauber Hamiltonian, HnˉsGH_{\bar n s}^{G} in order to carry out soft-collinear factorization,

Σ≡lim⁡t→∞Tr⁡[e−iHefftρ(0)eiHefftM]=∑iΣ(i)(6)\Sigma\equiv\lim_{t\to\infty}\operatorname{Tr}\left[e^{-iH_{\mathrm{eff}}t}\rho(0)e^{iH_{\mathrm{eff}}t}\mathcal{M}\right] = \sum_i \Sigma^{(i)} \tag*{(6)}

This is an expansion in the number of interactions of the probe with the medium and is often referred to as the opacity expansion [26]. Taking the confinement length scale ΛQCD−1\Lambda_{\mathrm{QCD}}^{-1} and boosting to obtain a scale relevant for the size of a nucleon in the y−y^- direction, LD∼MpQ/(HΛQCD)L_D \sim M_p Q/(H\Lambda_{\mathrm{QCD}}). In reality LDL_D might be substantially larger due to inter-nucleon interactions, especially in larger nuclei. We also consider the mean free path of the probe through the medium, λmfp∼xQ/(αs(Q)2Γ)\lambda_{\mathrm{mfp}} \sim xQ/(\alpha_s(Q)^2\Gamma), which determines how often the probe interacts with the medium, and depends on the interaction strength and medium density. Here Γ\Gamma is the nucleon density per unit area, where Γ∼ΛQCD2\Gamma\sim\Lambda_{\mathrm{QCD}}^2 for nucleons packed next to one another. We work in the regime λ2=LD/λmfp∼Γ/Q2≪1\lambda_2 = L_D/\lambda_{\mathrm{mfp}} \sim\Gamma/Q^2 \ll1 where successive interactions of the probe with the medium happen with color uncorrelated partons. The assumption λ2≪1\lambda_2 \ll1 dictates that only even powers of HnˉsGH_{\bar n s}^{G} contribute so as to have the same number of Glauber exchanges on both sides of the cut. λ2≪1\lambda_2 \ll1 implies that each successive Glauber exchange happens with an independent scattering center displaced in the longitudinal direction by λmfp≫LD\lambda_{\mathrm{mfp}} \gg L_D. Therefore at the amplitude squared level, each Glauber exchange is accompanied by its complex conjugate, and odd number of Glauber exchanges are excluded. The factorization formula at O(2m)O(2m) in this series is

Σ(2m)=∣CG(μ)∣2mQ4[∫d4peM]Lαβ∫yˉ∈Nucleusdyˉ+d2yˉ⊥Im⁡{[∏i=1m∫0LNdyˉi− Θ(yˉi−−yˉi+1−)∫d2ki⊥(2π)2Bn(ki−,yˉi−,y+,yˉ⊥;ν)]Smαβ(k1⊥,…,km⊥;yˉ1−,…,yˉm−;μ,ν)}.(7)\Sigma^{(2m)} = \frac{|C_G(\mu)|^{2m}}{Q^4}\left[\int d^4p_e\mathcal{M}\right]L_{\alpha\beta}\int_{\bar y\in\mathrm{Nucleus}}d\bar y^+d^2\bar y_\perp \operatorname{Im}\left\{\left[\prod_{i=1}^{m}\int_0^{L_N}d\bar y_i^-\,\Theta(\bar y_i^- - \bar y_{i+1}^-)\int\frac{d^2k_i^\perp}{(2\pi)^2}\mathcal{B}_n(k_i^-,\bar y_i^-,y^+,\bar y^\perp;\nu)\right]\mathcal{S}_m^{\alpha\beta}(k_1^\perp,\ldots,k_m^\perp;\bar y_1^-,\ldots,\bar y_m^-;\mu,\nu)\right\}. \tag*{(7)}

It is written in terms of mm copies of an nˉ\bar n-collinear medium structure function Bn\mathcal{B}_n and a soft function Smαβ\mathcal{S}_m^{\alpha\beta} describing the evolution of the Dipole probe. The λ2≪1\lambda_2 \ll1 expansion leads to the product of Bˉn\bar{B}_n functions, rather than a single multi-interaction nˉ\bar{n} correlation function. ∫d+pe\int d^+p_e is the phase space integral over the final state electron and LαβL_{\alpha\beta} is the leptonic tensor. The coordinate yˉ\bar{y} indicates the position inside the nucleus. The physical picture is that the successive interactions of the probe dipole occur with color uncorrelated partons in the nucleus at a single value of yˉ+\bar{y}^+, yˉ⊥\bar{y}^{\perp} but at increasing values of yˉ−\bar{y}^- as shown in Figure 1. The nucleus extends over a length LN−L_N^- in the y−y^- direction. ki⊥k_i^\perp is the transverse momentum exchanged in the ithi^{\mathrm{th}} interaction. The separation of collinear-soft physics leads to rapidity divergences, for which we use the dimensional regularization like rapidity regulator of [27] with the corresponding renormalization scale ν\nu. Bˉn\bar{B}_n is a function of yˉ\bar{y} so that it can also describe an inhomogeneous medium. It is given by a current-current correlator in the nuclear background

Figure 1. Interaction of the probe with the medium.

Bˉn(k⊥,yˉ;ν)=121k⊥2(NC2−1)∫d2y^⊥dy^−e−ik⊥⋅y^⊥×Tr⁡[OnA(12(−y^+yˉ))OnA(12(y^+yˉ))ρA].(8)\bar{B}_n(k_\perp,\bar{y};\nu)=\frac{1}{2}\frac{1}{k_\perp^2}(N_C^2-1)\int d^2\hat{y}_\perp d\hat{y}^-e^{-\mathrm{i}k_\perp\cdot\hat{y}^\perp}\times\operatorname{Tr}\left[O_n^A\left(\frac{1}{2}(-\hat{y}+\bar{y})\right)O_n^A\left(\frac{1}{2}(\hat{y}+\bar{y})\right)\rho_A\right]. \tag*{(8)}

This captures the universal physics of the medium to all orders in perturbation theory and is independent of any measurements that are made on the probe. OnAO_n^A can be thought of as the operator analog of the color source function ρ\rho in the CGC approach [7]. The medium structure function does not have any UV divergences but obeys a renormalization group equation in rapidity which is the Balitsky–Fadin–Kuraev–Lipatov (BFKL) [28, 29] equation at one loop.

dBˉn(k⊥,yˉ;ν)dln⁡ν=−∫d2u⊥KBFKL(k⃗⊥,u⃗⊥)Bˉn(u⊥,yˉ;ν)=−αsNCπ2∫d2u⊥[Bˉn(u⊥,yˉ;ν)(u⃗⊥−k⃗⊥)2−k⊥2Bˉn(k⊥,yˉ;ν)2u⊥2(u⃗⊥−k⃗⊥)2].(9)\begin{aligned} \frac{d\bar{B}_n(k_\perp,\bar{y};\nu)}{d\ln\nu}=-\int d^2u_\perp K_{\mathrm{BFKL}}(\vec{k}_\perp,\vec{u}_\perp)\bar{B}_n(u_\perp,\bar{y};\nu) \\ &=-\frac{\alpha_sN_C}{\pi^2}\int d^2u_\perp\left[\frac{\bar{B}_n(u_\perp,\bar{y};\nu)}{(\vec{u}_\perp-\vec{k}_\perp)^2}-\frac{k_\perp^2\bar{B}_n(k_\perp,\bar{y};\nu)}{2u_\perp^2(\vec{u}_\perp-\vec{k}_\perp)^2}\right]. \tag*{(9)} \end{aligned}

Its natural rapidity scale corresponds to ν∼H/Q\nu\sim H/Q for which logarithms in this function are minimized. By RG consistency, we have

dSmαβ(k1⊥,…km⊥,{yˉi};μ,ν)dln⁡ν=∑i=1m∫d2u⊥KBFKL(u⊥,ki⊥)×Smαβ(k1⊥,…ki−1⊥,u⊥,ki+1⊥,…km⊥,{yˉi};μ,ν).(10)\begin{aligned} \frac{dS_m^{\alpha\beta}(k_{1\perp},\ldots k_{m\perp},\{\bar{y}_i\};\mu,\nu)}{d\ln\nu}=\sum_{i=1}^{m}\int d^2u_\perp K_{\mathrm{BFKL}}(u_\perp,k_{i\perp}) \\ &\times S_m^{\alpha\beta}(k_{1\perp},\ldots k_{i-1\perp},u_\perp,k_{i+1\perp},\ldots k_{m\perp},\{\bar{y}_i\};\mu,\nu). \tag*{(10)} \end{aligned}

The natural rapidity scale for this function is ν∼Q\nu\sim Q. We can use the rapidity RG equation to resum the large logarithms in xx by solving the BFKL equation, running the function Bˉn\bar{B}_n from ν≃s/Q\nu\simeq s/Q to ν≃Q\nu\simeq Q while setting ν≃Q\nu\simeq Q in the Dipole function, leading to an RG improved function BˉnR(k⊥,yˉ,x)\bar{B}_n^R(k_\perp,\bar{y},x). At leading log, we can resum the series in Eq. (6) to all orders in the Glauber interactions

Σ=∫d2yˉ+dyˉ−∫d2b F(b⃗,Q){1−P⁡e−∫0LN−dyˉ−dyˉ−λmfp(b⃗,Q,yˉ)}.(11)\Sigma=\int d^2\bar{y}^+d\bar{y}^- \int d^2b\,\mathcal{F}(\vec{b},Q)\left\{1-\operatorname{P}e^{-\int_0^{L_N^-}d\bar{y}^-\frac{d\bar{y}^-}{\lambda_{\mathrm{mfp}}(\vec{b},Q,\bar{y})}}\right\}. \tag*{(11)}

which is a path ordered exponential with λmfp\lambda_{\mathrm{mfp}} given by a RG improved medium structure function BˉnR\bar{B}_n^R,

λmfp−1(b⃗,Q,yˉ)=∫d2k⊥(2π)2k⊥2∣CG(k⊥)∣2BˉnR(k⊥,yˉ,x)×[eib⃗⋅k⃗⊥−1].(12)\begin{aligned} \lambda_{\mathrm{mfp}}^{-1}(\vec{b},Q,\bar{y})=\int\frac{d^2k_\perp}{(2\pi)^2k_\perp^2}|C_G(k_\perp)|^2\bar{B}_n^R(k_\perp,\bar{y},x) \\ &\times\left[e^{\mathrm{i}\vec{b}\cdot\vec{k}_\perp}-1\right]. \tag*{(12)} \end{aligned}

Here λmfp\lambda_{\mathrm{mfp}} is an emergent length scale, which can be interpreted as the mean free path of the probe in the medium in the y−y^- direction. It depends on the local properties of the medium at yˉ\bar{y}. The function F(b⃗,Q)\mathcal{F}(\vec{b},Q) includes the transverse component of the dipole function SnαβS_n^{\alpha\beta} at tree level,

F(b⃗,Q)=1Q3[∫d+peM]Lαβg⊥αβ∫d2p⊥d2pˉ⊥eib⃗⋅(p⃗⊥+pˉ⃗⊥)×∫01dz(z2+(1−z)2)[p⃗⊥Q2z(1−z)+p⃗⊥ 2+pˉ⃗⊥Q2z(1−z)+pˉ⃗⊥ 2]2.\begin{aligned} \mathcal{F}(\vec{b},Q)=\frac{1}{Q^3}\left[\int d^+p_e\mathcal{M}\right]L_{\alpha\beta}g_\perp^{\alpha\beta}\int d^2p_\perp d^2\bar{p}_\perp e^{\mathrm{i}\vec{b}\cdot(\vec{p}_\perp+\vec{\bar{p}}_\perp)} \\ &\times\int_0^1dz\left(z^2+(1-z)^2\right)\left[\frac{\vec{p}_\perp}{Q^2z(1-z)+\vec{p}_\perp^{\,2}}+\frac{\vec{\bar{p}}_\perp}{Q^2z(1-z)+\vec{\bar{p}}_\perp^{\,2}}\right]^2. \end{aligned}

with the transverse size of the dipole b⃗∼1/Q\vec{b}\sim1/Q, as this is the only scale in this function. The emergence of the mean free path automatically defines a new power counting parameter in the EFT

λ1=∫0LN−dyˉ−λmfp(b⃗,Q,yˉ).(13)\lambda_1=\int_0^{L_N^-}\frac{d\bar{y}^-}{\lambda_{\mathrm{mfp}}(\vec{b},Q,\bar{y})}. \tag*{(13)}

The appearance of the saturation scale is just the breakdown of the power expansion in this parameter, when λ1∼1\lambda_1\sim1. Given the fact that the transverse size of the dipole b⃗\vec{b} is conjugate to the scale QQ, we can define the saturation momentum scale Qs(yˉ)Q_s(\bar{y}) via

∫0LN−dyˉ−λmfp(∣b⃗∣=1/Qs,Qs,yˉ)=1.(14)\int_0^{L_N^-}\frac{d\bar{y}^-}{\lambda_{\mathrm{mfp}}(|\vec{b}|=1/Q_s,Q_s,\bar{y})}=1. \tag*{(14)}

This corresponds to the regime where the dipole is likely to have multiple interactions with the medium.

Eq. (13) will be equivalent to the Glauber–Gribov–Mueller (GGM) formula [2]

λ1=Γ(yˉ⊥)αsπ2Ncb⊥2xGN(x,1/b⊥),(15)\lambda_1 = \Gamma(\bar{y}_{\perp})\frac{\alpha_s\pi^2}{N_c}b_{\perp}^2 xG_N(x,1/b_{\perp}) , \tag*{(15)}

if we model the nucleus as a bunch of non-interacting nucleons with some effective Glauber gluon density xGN(x,1/b⊥)xG_N(x,1/b_{\perp}) and a density of nucleons per unit area Γ(yˉ)\Gamma(\bar{y}). Rescaling k⊥→k^⊥/∣b⊥∣k_{\perp} \to\hat{k}_{\perp}/|b_{\perp}| in Eq. (12) using Eq. (8), we find an overall factor λ1∝b⊥2\lambda_1 \propto b_{\perp}^2. We can then equate the k^⊥\hat{k}_{\perp} and yˉ−\bar{y}^{-} integral of the rescaled B\mathcal{B} with Γ(yˉ⊥)xGN(x,1/b⊥)\Gamma(\bar{y}_{\perp})xG_N(x,1/b_{\perp}); the number density of small xx gluons per unit area. GNG_N absorbs an αsNc\alpha_sN_c, leaving αsπ2/Nc\alpha_s\pi^2/N_c as the prefactor. This yields Eq. (15). In the CGC formalism xGN(x,1/b⊥)xG_N(x,1/b_{\perp}) is obtained by treating the nucleus as a source for small x gluons with radiative corrections from the BFKL/BK equations. The resummation modifies the saturation scale from its tree level value Qs0Q_{s0} to have a power law scaling with xx; Qs(x)=Qs0x−2.44αsNc/π2Q_s(x)=Q_{s0}x^{-2.44\alpha_sN_c/\pi^2}. The boundary condition for this resummation, in our case B(k⊥)\mathcal{B}(k_{\perp}), is described in the CGC by the MV model [30, 31, 32].

Role of the medium size: From Eq. (7), we observe potential sensitivity to another length scale LN−L_N^{-}; the extent of the medium in the y−y^{-} direction. The relevant radiation sensitive to this scale will have momenta p+∼1/LN−p^+ \sim1/L_N^{-}, and coherence time tc∼LN−t_c \sim L_N^{-}. If NN is the typical number of nucleons seen by the probe along its path, then the rest frame size of the nucleus is LR∼N/MpL_R \sim N/M_p, and boosting yields LN−∼LRMpQ/H∼QN/HL_N^{-} \sim L_RM_pQ/H \sim QN/H. Retaining a common virtuality ∼Q2\sim Q^2, this is a collinear-soft mode with

pcs∼HQ(Nx2,1N,x).(16)p_{\mathrm{cs}} \sim\frac{H}{Q}\left(Nx^2,\frac{1}{N},x\right). \tag*{(16)}

We can consider two important regimes of the consequent EFT as we dial down the value of xx with fixed QQ, keeping NN to be large but fixed as shown in Figure 2. We see that at not so small x∼1/Nx \sim1/N, the soft and collinear-soft modes are the same so that the two mode factorization formula discussed above, Eq. (7), is accurate.

Figure 2. Small xx regions for large NN nuclei (top panels) and small NN nuclei (bottom panel).

This means that while in a dense enough medium, λ1≳1\lambda_1 \gtrsim1 puts us in the saturation regime, it does not necessarily lead to a non-linear evolution. If the medium is large (LN−∼1/QL_N^{-} \sim1/Q) we still have a linear BFKL evolution.

For smaller x≪1/N≪1x \ll1/N \ll1, the collinear-soft and soft modes decouple, while still being well separated from the collinear function. Even though we will leave the explicit derivation of the factorization formula in this regime for a future paper, we can still say a lot about the physics. The dipole needs to interact at least twice with the nucleus to know its size, hence Eq. (7) is modified for m≥2m \ge2. In particular, the soft function SmμνS_m^{\mu\nu} must be further factorized into a yˉ−\bar{y}^{-} (medium size) insensitive soft function S^mαβ\hat{S}_m^{\alpha\beta} (with LN−→1/νL_N^{-} \to1/\nu) and a collinear-soft function sensitive to LN−νL_N^{-}\nu. The mm copies of the function Bn\mathcal{B}_n remains unchanged and still obey the BFKL equation with the same natural rapidity scale ν∼Q/x=H/Q\nu\sim Q/x = H/Q. The refactorization will modify the rapidity RG equation for the soft function away from BFKL with a natural rapidity scale ∼Q\sim Q. However, by RG consistency the combination of collinear-soft and soft function must still obey the BFKL equation. The natural rapidity scale for the CS function is ν∼H/(QN)\nu\sim H/(QN), so evolution to the soft function sums logs of Q2N/H∼xNQ^2N/H \sim xN. Thus ln⁡x\ln x summation will now have two contributions: the linear BFKL piece from Bn\mathcal{B}_n; and that from the CS function. At leading order the latter modifies terms with at least two copies of Bn\mathcal{B}_n, so we expect it to reproduce the non-linear Balitsky–Kovchegov (BK) [8, 15]/JIMWLK equation [9, 12, 33]. As a bonus, the CS function’s RG evolution also sums ln⁡N\ln N terms. The linear to non-linear evolution refers to the transition from pure BFKL to a nonlinear evolution which allows for both splitting and merging of what are referred to in literature as BFKL pomeron which in our case is the BFKL resummed B\mathcal{B} function. The interference generated between successive interactions with the medium through a collinear soft mode, would allow this merging and splitting thereby leading to terms in the evolution equation with a higher powers of the B\mathcal{B} function, going beyond linear BFKL proportional to just a single B\mathcal{B}.

Finally, as we approach N=1N=1, the collinear soft mode merges with the collinear mode. The O(m)O(m) glauber exchange yields a two mode factorization involving the same yˉ−\bar{y}^{-} insensitive Dipole function S^mαβ\hat{S}_m^{\alpha\beta} mentioned in the previous paragraph. However, in this case the independent scattering approximation breaks down, so rather than mm copies of Bn\mathcal{B}_n two-point correlators, we now have

mm point correlators Bnm\mathcal{B}_n^m. This dependence on distinct boundary conditions for each mm seems to complicate the resummation. To determine from our EFT whether the mm dependence of these boundary conditions is calculable at QQ with a small number of nonperturbative functions at ΛQCD\Lambda_{\mathrm{QCD}} requires a further scale separation ΛQCD2≪Q2\Lambda_{\mathrm{QCD}}^2 \ll Q^2 that we have not performed here. This assumption has been implemented when the CGC formalism has been applied to a single hadron [34, 35].

We setup a top-down EFT for small xx DIS on a large inhomogeneous nucleus, with specific power counting parameters that define its range of validity. We gave a factorization formula which has an operator definition for the universal physics of the medium, and separates out the probe and small-xx evolution. The structure at all orders in the number of independent probe-medium interactions leads us to define the mean free path of the probe by an operator matrix element in the nuclear state. This yields a power counting parameter that controls the importance of multiple interactions with the medium, and a means to define the saturation scale. The case x≪1/N≪1x \ll1/N \ll1 requires the introduction of a new soft-collinear radiation mode, providing a distinct condition for understanding the transition to non-linear evolution in small-xx resummation.

Acknowledgments

We thank Ira Rothstein and Raju Venugopalan for helpful discussions. This work was supported by the U.S. Department of Energy, Office of Science, Office of Nuclear Physics, from DE-SC0011090 and in part by the Simons Foundation through the Investigator grant 327942. V.V is supported by startup funds from the University of South Dakota.

References

References

  1. [1]L. V. Gribov, E. M. Levin, and M. G. Ryskin, Phys. Rept. 100, 1 (1983).
  2. [2]A. H. Mueller, Nucl. Phys. B 335, 115 (1990).
  3. [3]A. H. Mueller and J.-w. Qiu, Nucl. Phys. B 268, 427 (1986).
  4. [4]E. Iancu and R. Venugopalan, “The Color glass condensate and high-energy scattering in QCD,” in Quark-gluon plasma 4, edited by R. C. Hwa and X.-N. Wang (2003) pp. 249–3363, arXiv:hep-ph/0303204.DOI
  5. [5]F. Gelis, E. Iancu, J. Jalilian-Marian, and R. Venugopalan, Ann. Rev. Nucl. Part. Sci. 60, 463 (2010), arXiv:1002.0333 [hep-ph].DOI
  6. [6]Y. V. Kovchegov and E. M. Levin, Quantum Chromodynamics at High Energy, Cambridge Monographs on Particle Physics, Nuclear Physics and Cosmology (33), Vol. 33 (Cambridge University Press, 2022).
  7. [7]L. D. McLerran and R. Venugopalan, Phys. Rev. D 59, 094002 (1999), arXiv:hep-ph/9809427.DOI
  8. [8]I. Balitsky, Nucl. Phys. B 463, 99 (1996), arXiv:hep-ph/9509348.DOI
  9. [9]J. Jalilian-Marian, A. Kovner, A. Leonidov, and H. Weigert, Nucl. Phys. B 504, 415 (1997), arXiv:hep-ph/9701284.
  10. [10]J. Jalilian-Marian, A. Kovner, A. Leonidov, and H. Weigert, Phys. Rev. D 59, 014014 (1998), arXiv:hep-ph/9706377.DOI
  11. [11]A. Kovner, J. G. Milhano, and H. Weigert, Phys. Rev. D 62, 114005 (2000), arXiv:hep-ph/0004014.DOI
  12. [12]E. Iancu, A. Leonidov, and L. D. McLerran, Nucl. Phys. A 692, 583 (2001), arXiv:hep-ph/0011241.DOI
  13. [13]E. Iancu, A. Leonidov, and L. D. McLerran, Phys. Lett. B 510, 133 (2001), arXiv:hep-ph/0102009.DOI
  14. [14]E. Ferreiro, E. Iancu, A. Leonidov, and L. McLerran, Nucl. Phys. A 703, 489 (2002), arXiv:hep-ph/0109115.DOI
  15. [15]Y. V. Kovchegov, Phys. Rev. D 60, 034008 (1999), arXiv:hep-ph/9901281.DOI
  16. [16]G. F. Sterman, Nucl. Phys. B 281, 310 (1987).
  17. [17]S. Catani and L. Trentadue, Nucl. Phys. B 327, 323 (1989).
  18. [18]C. W. Bauer, S. Fleming, D. Pirjol, I. Z. Rothstein, and I. W. Stewart, Phys. Rev. D 66, 014017 (2002), arXiv:hep-ph/0209088.
  19. [19]V. N. Gribov, Zh. Eksp. Teor. Fiz. 57, 1306 (1969).
  20. [20]D. Neill, A. Pathak, and I. Stewart, (2023), arXiv:2303.13710 [hep-ph].arxiv.org/abs/2303.13710
  21. [21]C. W. Bauer, D. Pirjol, and I. W. Stewart, Phys. Rev. D 67, 071502 (2003), arXiv:hep-ph/0211069.DOI
  22. [22]C. W. Bauer, D. Pirjol, and I. W. Stewart, Phys. Rev. D 68, 034021 (2003), arXiv:hep-ph/0303156.DOI
  23. [23]C. W. Bauer, S. Fleming, D. Pirjol, and I. W. Stewart, Phys. Rev. D 63, 114020 (2001), arXiv:hep-ph/0011336.DOI
  24. [24]C. W. Bauer and I. W. Stewart, Phys. Lett. B 516, 134 (2001), arXiv:hep-ph/0107001.DOI
  25. [25]I. Z. Rothstein and I. W. Stewart, JHEP 08, 025 (2016), arXiv:1601.04695 [hep-ph].DOI
  26. [26]M. Gyulassy, P. Levai, and I. Vitev, Phys. Rev. Lett. 85, 5535 (2000), arXiv:nucl-th/0005032.DOI
  27. [27]J.-Y. Chiu, A. Jain, D. Neill, and I. Z. Rothstein, JHEP 05, 084 (2012), arXiv:1202.0814 [hep-ph].DOI
  28. [28]E. A. Kuraev, L. N. Lipatov, and V. S. Fadin, Sov. Phys. JETP 45, 199 (1977).
  29. [29]I. I. Balitsky and L. N. Lipatov, Sov. J. Nucl. Phys. 28, 822 (1978).
  30. [30]L. D. McLerran and R. Venugopalan, Phys. Rev. D 49, 2233 (1994), arXiv:hep-ph/9309289.DOI
  31. [31]L. D. McLerran and R. Venugopalan, Phys. Rev. D 49, 3352 (1994), arXiv:hep-ph/9311205.DOI
  32. [32]L. D. McLerran and R. Venugopalan, Phys. Rev. D 50, 2225 (1994), arXiv:hep-ph/9402335.DOI
  33. [33]H. Weigert, Nucl. Phys. A 703, 823 (2002), arXiv:hep-ph/0004044.DOI
  34. [34]E. Iancu, J. D. Madrigal, A. H. Mueller, G. Soyez, and D. N. Triantafyllopoulos, Phys. Lett. B 750, 643 (2015), arXiv:1507.03651 [hep-ph].DOI
  35. [35]G. Beuf, H. Hänninen, T. Lappi, and H. Mäntysaari, Phys. Rev. D 102, 074028 (2020), arXiv:2007.01645 [hep-ph].DOI

Paper details

vlnc:W_8m6t6jjq
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