betaFeedback

WorksPhysics and Astronomy

Single pion-production and pion propagation in Achilles

Abstract

We extend the applicability of Achilles (A CHIcagoLand Lepton Event Simulator) by incorporating the single-pion production mechanism in a fully exclusive fashion. The electroweak interaction vertex is modeled by combining the state-of-the-art Dynamical Coupled-Channels approach with realistic hole spectral functions, which account for correlations in both the initial target state and the residual spectator system. Final-state interactions are treated using a semi-classical intranuclear cascade that leverages nuclear configurations to determine the correlated spatial distribution of protons and neutrons. The meson-baryon scattering amplitudes used in the cascade are computed within the Dynamical Coupled-Channels framework, consistent with the electroweak vertex. To model pion absorption, we employ the optical potential approach of Oset and Salcedo. As an alternative approach, we explicitly model the production and propagation of resonances which mediate pion-nucleon scattering and pion absorption. We validate out approach against pion-nucleon and pion-nucleus scattering data, and present comparisons with electron- and neutrino-nucleus measurements from e4νν, T2K, MINERννA, and MicroBooNE.

Introduction

Accurately modeling neutrino–nucleus scattering in the resonance-production region is critical for the success of accelerator-based neutrino oscillation programs. This region is particularly important for Deep Underground Neutrino Experiment (DUNE) [1], as the incoming neutrino flux spans energies from 1 to 10 GeV. A precise understanding of resonance production is also essential for experiments with lower-energy fluxes, such as the planned Hyper-Kamiokande (HK) [2], since pions produced in neutrino interactions can be reabsorbed within the nucleus, causing such events to be misidentified as quasi-elastic. This effect is especially relevant when comparing predictions from nuclear ab initio methods to flux-folded cross sections, as these misidentified events must be excluded from the measured cross section [3, 4].

Capitalizing on liquid-argon detectors—such as those used in the Short Baseline Neutrino (SBN) program [5] and DUNE—requires modeling neutrino–nucleus scattering in an exclusive manner. A complete quantum-mechanical treatment of the complex nuclear dynamics involved in this process will remain unfeasible in the foreseeable future, despite ongoing progress using quantum

computers to simulate simplified scenarios [6, 7]. Neutrino event generators make the problem tractable by separating the electroweak interaction vertex—treated at the quantum-mechanical level—from the subsequent intranuclear cascade (INC), which models the semi-classical propagation of hadrons through the nuclear medium [8, 9, 10, 11].

This work focuses on single-pion production on nuclei, a multi-scale problem that involves both hadronic and nuclear excitations in a region where quantum chromodynamics is strongly coupled. In addition, the momenta at play are much larger than the pion mass, which places it outside the radius of convergence of chiral effective field theory in this regime. Thus, the model developed by Hernández, Nieves, and Valverde (HNV) [12], which evaluates the non-resonant background amplitudes from a leading-order chiral Lagrangian, is limited at intermediate energies. To extend its applicability, particularly in the Δ\Delta region, the authors of Ref. [13] have partially unitarized the model by imposing Watson’s theorem to the dominant vector and axial multipoles.

Two of the most widely used event generators, NEUT and GENIE, rely on the Rein–Sehgal model [14] and its Berger–Sehgal extension [15]. These frameworks simulate pion production through the incoherent excitation of multiple nucleon resonances—primarily the Δ(1232)\Delta(1232), but also higher-mass N∗N^* and Δ\Delta states—followed by their decay into a nucleon and a pion. The Berger–Sehgal model improves upon the original formalism by accounting for the charged lepton mass and refining the pion angular distributions. Vector transition form factors are constrained by electroproduction data, while axial contributions are typically modeled with a dipole form and tuned to historical neutrino–deuteron data [16]. NuWro has long included only the Δ(1232)\Delta(1232) explicitly and treated higher invariant mass production via a smooth interpolation to a Deep-Inelastic-Scattering (DIS)-based hadronization model, invoking quark-hadron duality [17]. More recently, NuWro has adopted the Ghent hybrid model [18]. In the low-energy region, the resonant and non-resonant contributions are those of the HNV model. At higher energies, the description of the non-resonant background is instead based on Regge phenomenology [19, 20].

In this work, we implement in Achilles the Dynamical Coupled-Channels (DCC) model [21, 22, 23, 24] to compute the single-pion production amplitudes. The DCC approach provides a unified, unitary description of meson–baryon scattering and electroweak pion production by solving coupled integral equations for multiple reaction channels. This framework enables a consistent treatment of both resonant and non-resonant contributions, including their interference, across a broad range of energies and final states. We note that the recent implementation of the DCC model in NEUT [25] is fully inclusive. Therefore, it computes the hadronic response via the structure functions, and is therefore limited to providing only lepton kinematics, without access to the hadronic final state. In contrast, our implementation retains the full dynamical information of the DCC amplitudes, allowing for the exclusive modeling of the hadronic final state, including pion and nucleon kinematics.

To account for correlations in the initial target state and the residual spectator system, we employ realistic hole spectral functions of 12C^{12}\mathrm{C} [26] and 40Ar^{40}\mathrm{Ar} [27] nuclei. After the primary interaction, all produced hadrons are propagated through the nuclear medium using an INC that accounts for scattering, absorption, and charge exchange. To gauge the sensitivity to a range of in-medium effects, we implement and test two different cascade modes depending on whether the Δ\Delta and higher nucleon resonances are propagating degrees of freedom.

In the first approach, the relevant meson–baryon cross sections are computed from the standard decomposition of the scattering matrix into partial-wave amplitudes. Crucially, to ensure consistency with the electroweak interaction vertex, these amplitudes are computed within the Dynamical Coupled-Channels (DCC) framework. In addition to meson–baryon scattering, where a meson remains in the final state, absorption processes play a significant role in the INC. Here, we implement an approach commonly used in neutrino event generators, based on the optical potential model of Oset and Salcedo [28], in which mesons are absorbed via the imaginary part of their self-energy in nuclear matter. Crucially in this approach, nucleon resonances are not propagating degrees of freedom.

In the second approach, we follow the approach taken by GiBUU [29] and INCL [30] where we treat the Δ\Delta resonance as a dynamical degree of freedom that propagates throughout the nuclear medium. This approach provides a more microscopic description of three-body processes involved in pion absorption (e.g., πN↔Δ\pi N \leftrightarrow\Delta, ΔN↔NN\Delta N \leftrightarrow NN), and accurately accounts for the relatively long lifetime of the Δ\Delta resonance, compared to typical intranuclear propagation timescales. This provides a mechanism to capture re-scattering effects before the decay of the Δ\Delta and for medium modifications due to nuclear correlations.

This paper is organized as follows. In Sect. II, we describe both the electroweak interaction vertex—including how the DCC model is incorporated into the extended factorization scheme—and the INC, using the two different cascade modes. Sect. III presents a validation of our approach, beginning with inclusive electron-nucleus scattering, followed by comparisons to pion–nucleus scattering observables. Comparisons with exclusive electron- and neutrino-nucleus scattering data, including results from the e1νe1\nu, T2K, MINERvA, and MicroBooNE experiments, are presented in Section IV. Finally, in Section V, we summarize our conclusions and outline future directions for Achilles development.

Theoretical Treatment

Achilles relies upon a factorization of lepton-nucleus scattering into an incoherent product of the initial electroweak interaction, and the subsequent intranuclear cascade which transports particles out of the nucleus [31]. This amounts to approximating the squared matrix element as

∣M({k}→{p′})∣2≃∑p∣V({k}→{p})∣2×∣P({p}→{p′})∣2.(1)\left|\mathcal{M}(\{k\}\to\{p'\})\right|^2 \simeq\sum_p \left|\mathcal{V}(\{k\}\to\{p\})\right|^2 \times\left|\mathcal{P}(\{p\}\to\{p'\})\right|^2 . \tag*{(1)}

In Eq. (1), V({k}→{p})\mathcal{V}(\{k\}\to\{p\}) is the amplitude for the initial hard interaction process, which in Achilles is provided by the spectral-function formalism. The intranuclear cascade (INC) in Achilles provides P({p}→{p′})\mathcal{P}(\{p\}\to\{p'\}), the probability for particles {p′}\{p'\} to leave the nucleus when starting with particles {p}\{p\} inside the nucleus. The INC takes as input nuclear configurations and uses a semiclassical, impact-parameter-based algorithm to model the propagation of particles in the nuclear medium [32]. Below we describe the ingredients necessary for both the hard interaction and the INC.

ANL-Osaka dynamical coupled-channel approach

The ANL-Osaka dynamical model describes πN\pi N and γN\gamma N reactions in terms of hadron degrees of freedom by solving a coupled-channel scattering equation, which preserves unitarity of scattering amplitudes. More explicitly, the approach employs an energy-independent Hamiltonian of the form

H=H0+Γ+v.(2)H = H_0 + \Gamma+ v . \tag*{(2)}

The first term H0H_0 corresponds to the free Hamiltonian. The second term Γ\Gamma describes production and decay of “bare” excited-hadron states such as the N∗N^* and Δ\Delta from a meson-baryon channel (e.g., πN\pi N or ηN\eta N). The third term vv describes non-resonant interactions as a sum of energy-independent meson-exchange potentials which are derived by applying a unitarity transformation to a phenomenological Lagrangian [33, 34].

The Hamiltonian in Eq. (2) contains stable asymptotic particles as well as unstable particles. What is desired is an effective description of the scattering process for the former that consistently includes the influence of the latter. Standard Feshbach operator projection techniques [35] can be applied to produce an energy-dependent non-local Hamiltonian in terms of the stable particles only. The problem at hand is phrased most conveniently in terms of two-body scattering matrix TT. In Ref. [36], operator projection techniques were applied to Eq. (2) to yield a unitary set of coupled-channel equations, allowing the TT-matrix to be written as

⟨β∣T±(s)∣α⟩=⟨β∣tbg±(s)∣α⟩+⟨β∣tres±(s)∣α⟩(3)\langle\beta\lvert T^{\pm}(s) \rvert\alpha\rangle= \langle\beta\lvert t_{\mathrm{bg}}^{\pm}(s) \rvert\alpha\rangle+ \langle\beta\lvert t_{\mathrm{res}}^{\pm}(s) \rvert\alpha\rangle \tag*{(3)}

where ss is the square of the scattering center-of-mass energy and α,β\alpha,\beta denote meson-baryon channels. Partial contributions to the total TT matrix are denoted with lower-case tt. The superscript ±\pm indicates if the outgoing/incoming boundary condition is applied; in the following equations, we omit this superscript for brevity. We have separated the resonant contribution to the T±T^{\pm} matrix, tres±t_{\mathrm{res}}^{\pm}, from the non-resonant background contribution, tbg±t_{\mathrm{bg}}^{\pm}.

The non-resonant meson-baryon tt-matrix is obtained by solving the coupled-channel equation as

⟨β∣tbg(s)∣α⟩=⟨β∣V(s)+V(s)GMB(s)tbg(s)∣α⟩(4)\langle\beta\lvert t_{\mathrm{bg}}(s) \rvert\alpha\rangle= \langle\beta\lvert V(s) + V(s)G_{\mathrm{MB}}(s)t_{\mathrm{bg}}(s) \rvert\alpha\rangle \tag*{(4)}

where GMB(s)G_{\mathrm{MB}}(s) is the Green’s function of a meson-baryon state. The effective two-body interaction V(s)V(s) is the sum of the non-resonant interaction vv and the contribution of the so-called “ZZ-diagram” mechanisms which include ππN\pi\pi N intermediate states, for more details see Figure 8 in Ref. [36].

Figure 8. Left: Comparison of the CC1π⁺ differential cross sections with respect to the initial nucleon momentum pNp^{N} and Right: to the double transverse momentum imbalance δpTT\delta p_{\mathrm{TT}} measured by T2K [73] compared to predictions by ACHILLES.

The resonant tt-matrix is given by the sum of the N∗N^* or Δ\Delta resonances as

⟨β∣tres(s)∣α⟩=∑m,n⟨χβ∣Γ∣Nm∗⟩[D]m,n⟨Nn∗∣Γ∣χα⟩(5)\langle\beta\lvert t_{\mathrm{res}}(s) \rvert\alpha\rangle= \sum_{m,n} \langle\chi_\beta\lvert\Gamma\rvert N_m^* \rangle[D]_{m,n} \langle N_n^* \lvert\Gamma\rvert\chi_\alpha\rangle \tag*{(5)}

where [D]m,n[D]_{m,n} is the resonance propagator and ⟨Nm∗∣Γ∣χα⟩\langle N_m^* \lvert\Gamma\rvert\chi_\alpha\rangle describes the production vertex of the resonance state from a distorted meson-baryon state

∣χα⟩=[1+GMB(s)tbg(s)]∣α⟩,(6)\lvert\chi_\alpha\rangle= [1 + G_{\mathrm{MB}}(s)t_{\mathrm{bg}}(s)] \lvert\alpha\rangle, \tag*{(6)}

where the distortion on the right-hand side arises from the non-resonant interaction. All free parameters of the meson-exchange interaction vv are determined by fitting the data for πN→πN\pi N \to\pi N from the SAID database [37].

Ref. [21] gives a detailed discussion of how the parameters and masses entering in the N∗N^* transition to meson-baryon states are determined using different datasets.

Within the DCC model, the electroweak pion-production matrix element which provides the hard matrix element of Eq. (1), is a sum of non-resonant and resonant contributions

⟨α∣jμ∣N⟩=⟨α∣jbgμ∣N⟩+∑m,n⟨χα∣Γ∣Nm∗⟩[D]m,n⟨Nn∗∣jμ∣N⟩,(7)\langle\alpha\lvert j^\mu\rvert N \rangle= \langle\alpha\lvert j^\mu_{\mathrm{bg}} \rvert N \rangle + \sum_{m,n} \langle\chi_\alpha\lvert\Gamma\lvert N_m^* \rangle[D]_{m,n} \langle N_n^* \lvert j^\mu\rvert N \rangle, \tag*{(7)}

where ⟨α∣=⟨πN∣\langle\alpha\rvert= \langle\pi N \rvert is a pion-nucleon final state, and the first and second terms on the right hand side of Eq. (7) are the electroweak counterparts of vv and Γ\Gamma in Eq. (2), respectively. The current in the resonant amplitude ⟨Nm∗∣jμ∣N⟩\langle N_m^* \lvert j^\mu\rvert N \rangle can be further decomposed as

jμ=jresμ+Γ[1+tbg(s)GMB(s)]jbgμ.(8)j^\mu= j^\mu_{\mathrm{res}} + \Gamma[1+t_{\mathrm{bg}}(s)G_{\mathrm{MB}}(s)]j^\mu_{\mathrm{bg}}. \tag*{(8)}

The first term in the resonant amplitude comes from the direct coupling of the current to the resonance in the DCC model, while the second term arises because a re-scattering process can also excite the resonance.

The dynamical model was initially developed to study the electromagnetic structure of the Δ\Delta resonance. In this energy regime, only the πN\pi N channel is open. The vector transition form factors at finite momentum transfer Q2Q^2 are extracted from analyses of pion electroproduction in Ref. [38, 39]. To extend the model to neutrino-induced reactions, Ref. [39] introduced an axial-vector current component, guided by the partially conserved axial current (PCAC) hypothesis,1 while all parameters associated with the strong interaction and the vector current are kept unchanged.

The ANL-Osaka DCC amplitudes given in Eq. (7) are expanded in partial waves and provide the basis of our calculations for single pion production in electron- and neutrino-nucleus scattering. Amplitudes for electromagnetic (EM), charged current (CC), and neutral current (NC) processes are provided in tabularized form in the πN\pi N center-of-mass frame and boosted to the lab frame following the procedure given in Ref. [26]. A more thorough discussion of their implementation is given in the next section. The DCC model has been successfully extended to account for the production of ππN\pi\pi N final states [38]. However, discussion of these developments exceeds the scope of the present work.

Electroweak Interaction Vertex

As mentioned above and emphasized in Ref. [31], the theoretical treatment of the hard scattering begins with an explicit factorization ansatz for the hadronic final state ∣Ψf⟩|\Psi_f\rangle immediately following the initial electroweak interaction. In the quasi-elastic region, the nuclear current operator reduces to a sum of one-body terms and the nuclear final state factorizes as ∣ΨfA⟩=∣p⟩⊗∣ΨfA−1⟩|\Psi_f^A\rangle=|p\rangle\otimes|\Psi_f^{A-1}\rangle. Here, ∣p⟩|p\rangle denotes the final-state nucleon with momentum pp and energy e(p)=p2+mN2e(p)=\sqrt{p^2+m_N^2}, while ∣ΨfA−1⟩|\Psi_f^{A-1}\rangle describes the (A−1)(A-1)-body spectator system. Supposing a momentum transfer of qμ=(ω,q)q^\mu=(\omega,\mathbf{q}), the energy and momentum of the recoiling system are fixed by energy and momentum conservation

EfA−1=ω+E0−e(p),PfA−1=q−p,(9)E_f^{A-1}=\omega+E_0-e(p),\qquad\mathbf{P}_f^{A-1}=\mathbf{q}-\mathbf{p}, \tag*{(9)}

where E0E_0 is the ground state energy of the initial state nucleus. The incoherent contribution to the hadron tensor with one nucleon (1N) in the final state is given by

W1Nμν(p,q,sp,tp)=∑sk,tk∫d3k(2π)312e(k)dE Stk(k,E)×⟨k∣jμ∣p⟩⟨p∣jν∣k⟩δ(3)(p−k−q)×δ(ω−E+mN−e(k+q)),(10)\begin{aligned} W^{\mu\nu}_{1\mathrm{N}}(p,q,s_p,t_p)=\sum_{s_k,t_k}\int\frac{d^3k}{(2\pi)^3}\frac{1}{2e(k)}dE\,S_{t_k}(k,E) \\ &\times\langle k|j^\mu|p\rangle\langle p|j^\nu|k\rangle\delta^{(3)}(\mathbf{p}-\mathbf{k}-\mathbf{q}) \\ &\times\delta(\omega-E+m_N-e(\mathbf{k}+\mathbf{q})), \tag*{(10)} \end{aligned}

where the integration runs over the removal energy and initial momentum of the initial-state nucleon and where sk,tks_k,t_k (sp,tps_p,t_p) label the spin and isospin of the initial (final) nucleons, mNm_N is the rest mass of the nucleon. We note that, consistent with Ref. [31], the phase space integrals over the momenta and the spin-isospin summations of the final-state particles are performed outside the hadron tensor, (see App. A for details on the phase space integrals). ACHILLES performs the summation over these quantum numbers, convolutes the result with the one-body spectral function, and enforces energy and momentum conservation. The hole spectral function Stk(k,E)S_{t_k}(k,E) provides the probability distribution of removing an initial-state nucleon with momentum kk from the target nucleus, leaving the residual (A−1)(A-1)-nucleon system with an excitation energy EE; it is normalized to the number of protons ZZ and the number of neutrons A−ZA-Z, depending on tkt_k.

The factorization scheme readily generalizes to include a pion in the final state [26]:

∣ΨfA⟩→∣pπp⟩⊗∣ΨfA−1⟩,(11)|\Psi_f^A\rangle\to|p\pi_p\rangle\otimes|\Psi_f^{A-1}\rangle, \tag*{(11)}

where a nucleon and a pion with four-momenta pp and pπp_\pi are produced in the hard interaction, respectively. The incoherent contribution to the hadron response tensor with one nucleon and one pion (1N1π\pi) in the continuum can be written as [26]

W1N1πμν(p,q,sp,tp,pπ)=∑sk,tk∫d3k(2π)3dE14e(k)e(pπ)×Ptk(k,E)⟨k∣jμ∣pπp⟩⟨pπp∣jν∣k⟩δ(3)(p+pπ−k−q)×δ(ω−E+mN−e(k+q−pπ)−eπ(pπ)),(12)\begin{aligned} W^{\mu\nu}_{1\mathrm{N}1\pi}(p,q,s_p,t_p,p_\pi)=\sum_{s_k,t_k}\int\frac{d^3k}{(2\pi)^3}dE\frac{1}{4e(k)e(p_\pi)} \\ &\times P_{t_k}(k,E)\langle k|j^\mu|p_\pi p\rangle\langle p_\pi p|j^\nu|k\rangle\delta^{(3)}(\mathbf{p}+\mathbf{p}_\pi-\mathbf{k}-\mathbf{q}) \\ &\times\delta(\omega-E+m_N-e(\mathbf{k}+\mathbf{q}-\mathbf{p}_\pi)-e_\pi(p_\pi)), \tag*{(12)} \end{aligned}

where eπ(pπ)=pπ2+mπ2e_\pi(p_\pi)=\sqrt{p_\pi^2+m_\pi^2} is the energy of the outgoing pion and tπt_\pi its isospin. Again, the phase space integrals over the momenta and the spin-isospin summations of the final-state particles are performed outside the hadron tensor. The main difference between the above expression and Eq. (10) resides in the elementary amplitude. To describe the pion-production processes, we need matrix elements of the charged-current operator causing the transition from a bound nucleon ∣k⟩|k\rangle to a state with a pion and a nucleon ∣pπp⟩|p_\pi p\rangle.

The present work uses the ANL-Osaka DCC model discussed above to generate the required matrix elements of the charged-current operator describing the electroweak transition from a bound nucleon state ∣k⟩|k\rangle to a pion nucleon state ∣pπp⟩|p_\pi p\rangle. The necessary matrix elements are obtained from tables of amplitudes in the helicity-LSJLSJ mixed representation of Ref. [21]. This corresponds to a partial wave expansion of the final-state pion nucleon pair, as in Appendix B, while the initial state is described by two-particle helicity states with total angular momentum JJ. From these amplitudes the matrix elements ⟨pπ,p;sp∣jμ∣k;sp⟩\langle p_\pi,p;s_p|j^\mu|k;s_p\rangle, with sp(sk)s_p(s_k) the final (initial) nucleon spin, are applied in the center of momentum system. They are related to the matrix elements of Eq. (12) by a boost as described in Ref. [26]. In particular, the Wigner rotation of the spins is properly taken into account, thereby all angular dependence of the matrix element is included. The phase space is again handled using standard techniques developed for the LHC (see App. A). The results obtained neglecting FSI have been validated against inclusive electron scattering data on nuclei in Ref. [26, 40] and have been interfaced to ACHILLES using the Fortran90 wrapper discussed in Ref. [31].

Intranuclear Cascade

The ACHILLES INC begins by randomly sampling spatial distributions of protons and neutrons. For the 12C{}^{12}\mathrm{C} nucleus, these are obtained from Green’s function Monte Carlo calculations [41], thereby fully retaining correlation effects [32]. On the other hand, despite recent progress in utilizing artificial neural networks to compactly represent nuclear many-body wave functions [42], 40Ar{}^{40}\mathrm{Ar} is not yet amenable to continuum quantum Monte Carlo methods. In this case, the spatial positions of protons and neutrons are sampled from the single-proton and single-neutron density distributions, neglecting correlation effects altogether [32]. Particles produced from hard scattering events are propagated through the nucleus in straight-line trajectories in small steps of time δt\delta t. At every time step, propagating particles have the ability to interact with spectator nucleons based on their impact parameter and the available cross sections. The probability of an interaction occurring is based on a Gaussian or cylinder probability model which depends on both the impact parameter and the total cross section. In Refs. [31, 32] this approach was validated based on hadron-nucleus and electron-nucleus scattering data. Below we describe newly incorporated hadron-nucleon scattering models needed to accommodate pion propagation in the nucleus.

ANL-Osaka DCC Model

To incorporate meson-baryon interactions in the INC, we make use of the standard decomposition of the scattering matrix into partial-wave amplitudes (PWAs). For completeness, definitions of the PWA are collected in App. B. In addition to πN\pi N, the other zero strangeness meson-baryon channels (ηN\eta N, KΛK\Lambda, KΣK\Sigma) are included. For consistency with the pion production model, the amplitudes are all obtained from the ANL-Osaka DCC model.

The total meson-baryon cross section is the sum of the angle-integrated cross sections for all kinematically accessible channels. The cross section for each process i→fi \to f, where i,fi,f denote a meson-baryon state is given in terms of squared PWAs

σif(s)=4π∣ki∣2∑L(L∣τifL,−(s)∣2+(L+1)∣τifL,+(s)∣2),(13)\sigma_{if}(s)=\frac{4\pi}{|\mathbf{k}_i|^2}\sum_L\left(L|\tau_{if}^{L,-}(s)|^2+(L+1)|\tau_{if}^{L,+}(s)|^2\right), \tag*{(13)}

where s\sqrt{s} is the center-of-mass energy, and ki\mathbf{k}_i the center-of-mass momentum of the initial pair. The reduced PWA for given orbital angular momentum LL, and total angular momentum J=L±1/2J=L\pm1/2, τifL,±\tau_{if}^{L,\pm}, are defined in App. B. A particular meson-baryon scattering event is selected according to the weight given by the partial cross sections σif(s)\sigma_{if}(s). The final-state kinematics are then determined by the cross section in the center-of-momentum frame. The CM scattering angle z=k^i⋅k^f=cos⁡θz=\hat{\mathbf{k}}_i\cdot\hat{\mathbf{k}}_f=\cos\theta of the meson-baryon pair is generated according to the cross section

dσifdΩ=1∣ki∣2∣∑LPL(z)[LτifL,−(s)+(L+1)τifL,+(s)]∣2+1−z2∣ki∣2∣∑LPL′(z)[τifL,+(s)−τifL,−(s)]∣2,(14)\frac{\mathrm{d}\sigma_{if}}{\mathrm{d}\Omega}=\frac{1}{|\mathbf{k}_i|^2}\left|\sum_L P_L(z)\left[L\tau_{if}^{L,-}(s)+(L+1)\tau_{if}^{L,+}(s)\right]\right|^2+\frac{1-z^2}{|\mathbf{k}_i|^2}\left|\sum_L P'_L(z)\left[\tau_{if}^{L,+}(s)-\tau_{if}^{L,-}(s)\right]\right|^2, \tag*{(14)}

where PLP_L and PL′P'_L denote the Legendre polynomials of order LL and their derivatives. Validations of both the total and angular differential cross sections can be found in App. § 2.0.3. The distribution of the azimuthal angle in the CM frame is uniform.

While the interference between different partial waves cancels for the angle-integrated cross section, the angular distribution does include interference terms. We use the full angular distribution given by the polynomial of Eq. (14), including all interferences exactly. Since the

PWAs constitute a model-independent, complete basis for the scattering matrix, any approach for meson-baryon scattering in the resonance region, such as the Jülich-Bonn-Washington analysis [43], can be included straightforwardly.

Note that we choose to include this set of two-body channels explicitly, but the full DCC analysis includes additional final states, notably the ρN\rho N, σN\sigma N, and πΔ\pi\Delta channels which are doorway states for two-pion production. This subset of channels will underpredict the total cross section obtained from the optical theorem

σi→X=2π∣ki∣2∑L,J(2J+1)Im⁡[τiiL,J].(15)\sigma_{i\to X}=\frac{2\pi}{|\mathbf{k}_i|^2}\sum_{L,J}(2J+1)\operatorname{Im}\left[\tau_{ii}^{L,J}\right]. \tag*{(15)}

It is possible to rescale the total cross section based on this value, but to saturate this total rate with the correct particle content, one needs to include explicitly multipion final-states. This will be pursued in future work.

Oset model

Another common approach to pion-nucleon interactions in neutrino event generators involves the optical-potential model of Oset [28]. The model begins with the pion-nucleus optical potential, which is computed in infinite nuclear matter and corrected for finite nuclei by using the local density approximation. The model also includes certain in-medium effects such as Pauli-blocking and in-medium modifications to the Δ\Delta propagator. The p-wave probability for a pion to interact in the nuclear medium is obtained by computing the Δ\Delta self energy associated with the excitation of a Δ\Delta-hole pair. Interactions in the nuclear medium modify the Δ\Delta, giving rise to a self-energy function with imaginary part

Im⁡ΣΔ(Tπ)=−[CQ(ρ/ρ0)α+CA2(ρ/ρ0)β+CA3(ρ/ρ0)γ],(16)\operatorname{Im}\Sigma_{\Delta}(T_\pi)=-\left[C_Q(\rho/\rho_0)^\alpha+C_{A2}(\rho/\rho_0)^\beta+C_{A3}(\rho/\rho_0)^\gamma\right], \tag*{(16)}

where ρ0=0.16/fm3\rho_0=0.16/\mathrm{fm}^3 is the nuclear saturation density and ρ\rho is the local nuclear density.

The different terms in Eq. (16) arise from quasi-elastic scattering, and pion absorption on two and three nucleons respectively. The coefficient functions CQC_Q, CA2C_{A2}, CA3C_{A3}, and the corresponding exponents are parameterized in Ref. [44] in terms of the pion kinetic energy, TπT_\pi, up to 350 MeV. In this work evaluate Eq. (16) for pion kinetic energies beyond this value to compute absorption cross sections and comment on this effect in Sections III and IV. Oset et al. also provide a parameterization of the s-wave, non-resonant part of the optical potential which is important at Tπ<50T_\pi<50 MeV.

Though this model can be used to compute microscopic cross sections for πN\pi N quasi-elastic scattering, we utilize it only to compute pion absorption cross sections as follows. Consider first the case of absorption on two nucleons. When a candidate pion-nucleon pair is chosen for possible interaction, the absorption cross section is computed by searching for a nearby secondary nucleon to participate in the interaction. This search involves all isospin combinations, i.e., if the interacting pair is a π+n\pi^+n pair, then both a secondary neutron for π+nn→pn\pi^+nn \to pn and a secondary proton for π+np→pp\pi^+np \to pp are considered. To select the actual interaction experienced by the candidate pair, the cross sections are weighted according to the isospin dependence of the elementary cross sections [45]. In this way the absorption cross sections on asymmetric nuclei can be consistently computed as soon as separate proton and neutron densities are given. For simplicity we have only implemented the two nucleon absorption process, though we retain the three-nucleon absorption term to maintain the overall rate. We leave the inclusion of genuine 4-body initial states to future work.

GiBUU Delta model

An alternate approach to pion-nucleus interactions, used in the GiBUU event generator [29] and the Liège INC [30], treats the Δ\Delta as a propagating degree of freedom. Although explicit propagation of unstable particles introduces certain complications (namely, requiring models for their production, interactions, and decay), its utility extends beyond the simple case of the Δ\Delta resonance. For example, a framework for propagation of unstable particles provides the ability to decay final-state particles such as π0→γγ\pi^0 \to\gamma\gamma.

When the Δ\Delta is a propagating degree of freedom, pion interactions occurs via a two-step process, where first a πN\pi N interaction creates a Δ\Delta which propagates through the nucleus, and second this Δ\Delta either re-interacts or decays back to a πN\pi N. For example, pion absorption would occur via

πN→Δ(17)\pi N \to\Delta \tag*{(17)}
ΔN→NN.(18)\Delta N \to NN. \tag*{(18)}

Decays are handled probabilistically at each time step according to P(t)=1−e−δt/τP(t)=1-e^{-\delta t/\tau}, where δt\delta t is the time step for the cascade, and τ=γℏc/Γ\tau=\gamma\hbar c/\Gamma is the lifetime in the lab frame of the particle with a vacuum width of Γ\Gamma. When a decay occurs, the decay channel is selected according to the branching ratios.3 The decay is rejected if any of the decay products are Pauli blocked, thereby capturing some in-medium effects on the resonance width.

Additional interaction channels (like NN→NΔNN \to N\Delta and NΔ→NΔN\Delta\to N\Delta) are physically possible and indeed required for theoretical consistency. Detailed balance [46] relates the reactions NN→NΔNN \to N\Delta and NΔ→NNN\Delta\to NN:

σcd→ab=σab→cd(pabpcd)2(2Ja+1)(2Jb+1)(2Jc+1)(2Jd+1)SabScd,(19)\sigma_{cd\to ab}=\sigma_{ab\to cd}\left(\frac{p_{ab}}{p_{cd}}\right)^2\frac{(2J_a+1)(2J_b+1)}{(2J_c+1)(2J_d+1)}\frac{S_{ab}}{S_{cd}}, \tag*{(19)}

where pij=∣pi∣=∣pj∣p_{ij}=|\mathbf{p}_i|=|\mathbf{p}_j| is the center-of-mass three-momentum of particles ii or jj, JiJ_i is the total spin of particle ii, and SijS_{ij} is a symmetry factor for particles ii and jj, e.g., 1/2 for identical particles.

The cross section for the two-body NN↔NΔNN \leftrightarrow N\Delta scattering process can be generally parameterized as

dσNN→NΔdμΔ2dΩ=∣MNN→NΔ∣264π2spNΔpNNAΔ(μΔ2).(20)\frac{\mathrm{d}\sigma_{NN\to N\Delta}}{\mathrm{d}\mu_\Delta^2\mathrm{d}\Omega}=\frac{|\mathcal{M}_{NN\to N\Delta}|^2}{64\pi^2s}\frac{p_{N\Delta}}{p_{NN}}\mathcal{A}_\Delta(\mu_\Delta^2). \tag*{(20)}

The resonance spectral function AΔ(μΔ2)\mathcal{A}_\Delta(\mu_\Delta^2) for the Δ\Delta is related to the Breit–Wigner propagator and is given by

AΔ(μΔ2)=1πμΔΓ(mΔ2−μΔ2)2+μΔ2Γ2,(21)\mathcal{A}_\Delta(\mu_\Delta^2)=\frac{1}{\pi}\frac{\mu_\Delta\Gamma}{(m_\Delta^2-\mu_\Delta^2)^2+\mu_\Delta^2\Gamma^2}, \tag*{(21)}

where mΔm_\Delta is the on-shell mass, μΔ\mu_\Delta is the off-shell mass, and Γ=112 MeV\Gamma=112\ \mathrm{MeV} is the width. Ref. [47] (see also recent theses [48, 49]) has computed the process pp→nΔ++pp \to n\Delta^{++} at leading order using the interaction Lagrangian

L=F(t)mπ[fNNπ(Nˉγ5τaN)∂μΠa+fNΔπ2(ΔˉμTaN)∂μΠa],(22)\mathcal{L}=\frac{F(t)}{m_\pi}\left[f_{NN\pi}(\bar{N}\gamma_5\tau_aN)\partial^\mu\Pi_a+f_{N\Delta\pi}\sqrt{2}(\bar{\Delta}_\mu T_aN)\partial^\mu\Pi_a\right], \tag*{(22)}

where Δμ\Delta_\mu is a spin-32\frac{3}{2} Rarita–Schwinger field for the Δ\Delta-baryon, NN is a spin-12\frac{1}{2} field for the nucleon, and Πa\Pi_a is the pion field. Where τa\tau_a are the Pauli isospin matrices describing the 12→12\frac{1}{2}\to\frac{1}{2} transition while TaT_a represent the 2×42\times4 isospin coupling matrices which obey the relation

TaTb†=Iδab−13τaτb(23)T_aT_b^\dagger=\mathbb{I}\delta_{ab}-\frac{1}{3}\tau_a\tau_b \tag*{(23)}

for the 12→32\frac{1}{2}\to\frac{3}{2} processes. The dimensionless coefficients fNNπ=1.008f_{NN\pi}=1.008 and fNΔπ=2.202f_{N\Delta\pi}=2.202 are the effective couplings of the pion to the nucleon and Δ\Delta, respectively. The vertex form factor F(t)F(t) is taken to have the form

F(t)=Λ2−mπ2Λ2−t,(24)F(t)=\frac{\Lambda^2-m_\pi^2}{\Lambda^2-t}, \tag*{(24)}

with Λ=0.63 GeV\Lambda=0.63\ \mathrm{GeV} [47]. Each ΔNπ\Delta N\pi vertex also gives an additional form factor z\sqrt{z} of the form

z(t,μΔ)=λ(μΔ,mN,t)+κ2λ(mΔ,mN,t)+κ2,(25)z(t,\mu_\Delta)=\frac{\lambda(\mu_\Delta,m_N,\sqrt{t})+\kappa^2}{\lambda(m_\Delta,m_N,\sqrt{t})+\kappa^2}, \tag*{(25)}

where λ(a,b,c)=a2+b2+c2−2ab−2ac−2bc\lambda(a,b,c)=a^2+b^2+c^2-2ab-2ac-2bc is the Källén function, and κ=0.2 GeV\kappa=0.2\ \mathrm{GeV} is a parameter obtained from a fit to π+p→π+p\pi^+p\to\pi^+p scattering data for μΔ<1.4 GeV\mu_\Delta<1.4\ \mathrm{GeV} [47]. Additional details on the calculation of the matrix elements and a validation of the NN→NNπNN\to NN\pi cross section are given in App. § 2.2.4.

Initial StatesFinal states in different cascade modes
Virt. Resonances modeProp. Resonances mode
NNNNNNNN, NNπNN\piNNNN, NΔ(→NNπ,NNγ)N\Delta(\to NN\pi,NN\gamma)
πN\pi NπN,ηN,KΛ,KΣ\pi N,\eta N,K\Lambda,K\SigmaΔ(→πN,γN)\Delta(\to\pi N,\gamma N)
ηN\eta NπN,ηN,KΛ,KΣ\pi N,\eta N,K\Lambda,K\SigmaInitial state not present
ΔN\Delta NInitial state not presentNN,ΔNNN,\Delta N
πNN\pi NNNNNN (ss- and pp-wave)ΔN(→NN),NN\Delta N(\to NN),NN (ss-wave)

Table 1. Available initial and corresponding final states in the Virtual Resonances and Propagating Resonances modes of the Achilles cascade. The final row with initial state πNN\pi NN corresponds to pion absorption; the corresponding ss- or pp-wave contributions are computed using the Oset model. The ss-wave absorption in the Propagating Resonances mode also uses the Oset model to capture this process (see the text for more details). As indicated, not all initial states are present in each case.

In addition to Δ\Delta production, we also include elastic scattering of Δ\Delta's with nucleons (i.e. NΔ→NΔN\Delta\to N\Delta). The implementation in Achilles follows the calculation done in Refs. [29, 48, 49]. The differential cross section is given as

d2σdμΔfdΩ=∣MNΔi→NΔf∣264π2spNΔfpNΔiAΔ(μΔf2),(26)\frac{\mathrm{d}^{2}\sigma}{\mathrm{d}\mu_{\Delta_f}\mathrm{d}\Omega}=\frac{\lvert\mathcal{M}_{N\Delta_i\to N\Delta_f}\rvert^{2}}{64\pi^{2}s}\frac{p_{N\Delta_f}}{p_{N\Delta_i}}A_{\Delta}(\mu_{\Delta_f}^{2}), \tag*{(26)}

where pNΔip_{N\Delta_i} and pNΔfp_{N\Delta_f} are the center-of-mass momentum of the initial and final states respectively, and μΔf\mu_{\Delta_f} is the off-shell mass of the final state Δ\Delta. The matrix element is again obtained from the one-pion exchange model. See App. § 2.2.4 for further details.

Following the work of Ref. [50] where it was shown that in-medium changes to the NN→NΔNN \to N\Delta cross sections improve agreement with neutrino scattering data on 40Ar{}^{40}\mathrm{Ar}, we implement as an option the density suppression of this cross section following the work of Song and Ko [51]. This modifies the resonance excitation in NNNN collisions according to

σNN→NΔ(ρ)=exp⁡(−αρρ0)σNN→NΔ(0),(27)\sigma_{NN\to N\Delta}(\rho)=\exp\left(-\alpha\frac{\rho}{\rho_{0}}\right)\sigma_{NN\to N\Delta}(0), \tag*{(27)}

where α\alpha is a tunable parameter and ρ\rho is the local nuclear density. This affects both pion production via Eq. (27) as well as pion absorption through the reversed process NΔ→NNN\Delta\to NN. For reference, the authors of Ref. [50] use a value of α=1.2\alpha=1.2. We found that including this effect reduces overall agreement with experimental data, and therefore set α=0.0\alpha=0.0 for the results shown in this paper.

Finally, in addition to allowing the Δ\Delta to propagate in NN→NΔ→NNπNN \to N\Delta\to NN\pi, we offer the option to immediately decay the Δ\Delta. This allows us to compute the Δ\Delta resonance contribution to NN→NNπNN \to NN\pi in models where the Δ\Delta is not a propagating degree of freedom.

Achilles Cascade Modes

The above models contain all the ingredients necessary for pion propagation within the nucleus. In contrast to the propagating resonances which mediate πN\pi N interactions in the GiBUU Δ\Delta model, the πN\pi N interactions in the Oset and ANL-Osaka DCC model in the Virtual Resonances cascade are time independent point interactions (i.e. the production and decay vertices are not displaced). This is due to the fact that the virtual resonances are already integrated out when computing the interaction cross sections.

Given this, we have developed two cascade modes dubbed Virtual Resonances and Propagating Resonances. Table I summarizes the ingredients of these two modes. The Virtual Resonances mode relies on the ANL-Osaka DCC model to simulate meson-baryon to meson-baryon scattering (i.e. πN→πN\pi N \to\pi N), the Oset model for pion absorption, and utilizes the Δ\Delta resonance contribution of the NN→NNπNN \to NN\pi process where the intermediate Δ\Delta is immediately decayed. The Propagating Resonances mode mainly uses the GiBUU Δ\Delta model with a propagating Δ\Delta, but also includes the S-wave piece of the Oset model for pion absorption. This is included to reproduce the behavior of the pion absorption cross section at lower pion kinetic energies. The Propagating Resonance mode neglects interference effects and higher mass resonances between different interaction channels in the πN→πN\pi N \to\pi N chain that are included within the ANL-Osaka DCC model. In future work, we plan to include the higher resonances to fully handle the ANL-Osaka DCC model range of validity. The difference between these two choices gives an estimate of the uncertainty associated with modeling resonances during the intranuclear cascade.

Validation

Hard Scattering

The hard scattering model of § 2 was validated against (e,e′)(e,e') data on 12C{}^{12}\mathrm{C} at several beam energies and scattering angles in Ref. [26]. For these comparisons the authors of Ref. [26] used the spectral function of 12C{}^{12}\mathrm{C} computed using correlated-basis function theory [53], which is also used in Achilles [31]. Recently, high-precision (e,e′)(e,e') and (e,e′p)(e,e'p) data was recorded at Thomas Jefferson Lab (JLab) on 12C{}^{12}\mathrm{C}, 40Ar{}^{40}\mathrm{Ar}, and 48Ti{}^{48}\mathrm{Ti} [52, 54, 55]. These experiments were intended to test the scaling behavior of the (e,e′)(e,e') cross section and to study nuclear effects in (e,e′p)(e,e'p) scattering on 40Ar{}^{40}\mathrm{Ar} and 48Ti{}^{48}\mathrm{Ti}, relevant for the next generation of liquid-argon based neutrino experiments. To compare to data taken on 40Ar{}^{40}\mathrm{Ar}, we have incorporated the proton and neutron spectral function of 40Ar{}^{40}\mathrm{Ar} extracted from (e,e′p)(e,e'p) data on 40Ar{}^{40}\mathrm{Ar} and 48Ti{}^{48}\mathrm{Ti} [27, 54, 55]. Figure 1 shows the Achilles comparison to a selection of (e,e′)(e,e') data taken at Ebeam=2.222E_{\mathrm{beam}=2.222} GeV and θe′=15.541∘\theta_{e'}=15.541^\circ on 12C{}^{12}\mathrm{C} and 40Ar{}^{40}\mathrm{Ar} including contributions from quasi-elastic scattering and single-pion production.

Figure 1. Achilles comparison to inclusive electron scattering data on 40Ar{}^{40}\mathrm{Ar} (left) and 12C{}^{12}\mathrm{C} (right) from JLab at Ebeam=2.222E_{\mathrm{beam}} = 2.222 GeV and θe′=15.541∘\theta_{e'} = 15.541^\circ [52] as a function of energy transfer (ω\omega). The data and predictions have been normalized to the number of nucleons in each respective nucleus. The total prediction from Achilles is in blue, while the quasi-elastic component is in orange and the pion production component is in green. Gaps between the data and Achilles’ prediction for the differential cross section are expected to be largely filled by the eventual inclusion of contributions from meson exchange currents and deep inelastic scattering.

The quasi-elastic (orange) and Δ\Delta peaks (green) in both nuclei are qualitatively reproduced. The significant difference between the data and predictions in the dip region between the QE and Δ\Delta peaks is due to the lack of meson-exchange currents (MEC) in Achilles, known to fill this region [26]. Similarly, the deficit at high energy transfers beyond the Δ\Delta peak can be explained by the lack of multi-pion production and DIS in Achilles. The inclusion of these additional hard scattering processes will be the focus of future efforts. The small shift in the QE peak relative to the data is due to neglected final-state interactions in the hard-scattering matrix elements. The effect of final-state interactions is twofold: the dispersion relation of outgoing hadrons is modified5, and strength is redistributed over asymptotic final-state channels. The INC takes care of the latter; crucially it goes beyond the optical-potential approach by explicitly populating accessible final-states [27, 56]. The former affects the nuclear response integrated over final-states, and hence the hard scattering matrix element of Eq. (1). This can be incorporated within the factorized picture via a folding function approach [26, 57], by treating the outgoing particles as scattering states in a suitable real potential [58] or through the (relativistic) Green’s function approach of Refs. [59, 60, 61].

The effect in this kinematic region is a slight shift toward lower energy transfers and reduction of the height of the quasi-elastic peak [62]. Implementing this in a manner that is consistent with the intranuclear cascade exceeds the scope of the present work. We also compare the Achilles predictions for the total cross section for pion production on nucleons against the reanalyzed bubble chamber data from ANL and BNL [16]. These include νμp→μ−π+\nu_\mu p \rightarrow\mu^- \pi^+, νμn→μ−nπ+\nu_\mu n \rightarrow\mu^- n\pi^+, and νμn→μ−pπ0\nu_\mu n \rightarrow\mu^- p\pi^0, the latter being extracted from data on deuterium instead of free neutrons. Figure 2 shows that the implementation of the ANL-Osaka DCC model in Achilles reproduces the neutrino energy dependence of these elementary cross sections very well.

Figure 2. Comparison of total pion-production cross sections for νμ\nu_\mu scattering on nucleons. The data are from measurements at Argonne and Brookhaven National Laboratories reanalyzed in Ref. [16].

Pion-Nucleus scattering

We compute the π+\pi^+-carbon, and π+\pi^+-argon reaction and absorption cross sections. The results for the reaction cross sections, using both the Virtual and Propagating Resonances model are shown in Figure 3. We remind the reader that the DCC piece of the Virtual Resonances model uses only the contribution from explicit final-state meson-baryon channels (πN,ηN,KΛ,KΣ\pi N,\eta N,K\Lambda,K\Sigma) without in-medium modification of the cross sections. Both models undershoot the data for the reaction cross section on carbon in the high-pπp_\pi tail. This is to be expected since two-pion production and other channels are not included. Surprisingly, both model predictions exceed the data near the Δ\Delta peak, around pπ=300p_\pi= 300 MeV, which could point to the need to incorporate in-medium modifications of the πN\pi N cross sections in both models. Other explanations are also possible. For instance, the present treatment describes the nucleon momentum distribution using a local Fermi gas. A more realistic distribution would include some high-momentum nucleons, which would lead to additional smearing of features.

Figure 3. Validation of the pion modes in ACHILLES INC against experimental data. ACHILLES results are shown using both the Propagating Resonances model (solid line) and the Virtual Resonances model (dotted line). Left: Absorption cross section for π+\pi^{+} on carbon (top) and π+\pi^{+} on argon (bottom). Experimental data are for π+\pi^{+} absorption measurements on carbon (DUET [63] and Ashery [64]) and on argon (LADS [65]). Right: Total reaction cross section for π+\pi^{+} on carbon (top) and π+\pi^{+} on argon (bottom). Data for carbon are from Ref. [66].

Figure 3 compares the predictions of the absorption cross section to π+\pi^+ data on carbon and argon. Both models slightly underestimate the data for the absorption cross section on carbon at pion momenta around the Δ\Delta peak. For argon, both models give excellent agreement with the LADS data [65], except for the Propagating Resonances model at the highest pion momentum. We note that the direct NNΔ→NNNNN\Delta\rightarrow NNN process is missing in the Propagating Resonance model which when included could improve agreement with the data. At the moment the Achilles cascade does not handle three-body initial states. Additionally, we note that the strength at low pion momentum comes almost entirely from the Oset S-wave absorption process. Although experiments on the deuteron have shown this to be an important process at low energies [67], little is known about this process in heavier nuclei. We leave the investigation of these two aspects for future work.

As mentioned previously, we have extend the use of the Oset Model beyond its original limit of Tπ=350T_\pi= 350 MeV in the Virtual Resonances cascade. We have investigated the effect of this by freezing the various pion absorption cross sections at kinetic energies beyond this limit to their values at Tπ=350T_\pi= 350 MeV. As expected, this only affects the π\pi-carbon and π\pi-argon absorption cross sections above Tπ=350T_\pi= 350 MeV (pπ≈470p_\pi\approx470 MeV), where the absorption cross sections are reduced by roughly 30%. The agreement with the data is not qualitatively changed. We have confirmed that this change has very little effect on the various electron- and neutrino-nucleus observables we will compare to, and we leave it to further work to analyze this approximation for higher energy experiments.

Exclusive electron- and neutrino-nucleus scattering

Modern accelerator neutrino experiments utilize broad-band beams spanning the energy range from hundreds of MeV to 10s of GeV. This enables us to test neutrino-nucleus interaction models against multiple nuclear targets over a wide range of energies where multiple processes contribute to the observed event rate. Additionally, electron-nucleus scattering has become invaluable for constraining the vector part of the nuclear electroweak current [68, 69, 70]. Electrons probe the same nuclear ground state as neutrinos, and hadrons leaving the nucleus undergo similar final-state interactions. In this section we update the comparisons made in Ref. [31] with exclusive electron-carbon scattering data from the e4νe4\nu collaboration [70], as well as present comparisons with flux-folded neutrino scattering data from T2K, MINERvA and MicroBooNE. These experiments operate in complementary beam lines and probe targets composed of hydrocarbon (MINERν\nuA and T2K), and argon (MicroBooNE). A global comparison including these experiments highlights the need to capture both the energy and AA dependence of neutrino-nucleus cross sections.

Comparison to e4

The CLAS and e4νe4\nu collaborations have analyzed data on electron-nucleus scattering at similar energies- and using similar hadron detection thresholds as neutrino scattering experiments [70]. This analysis focused on the comparison of different neutrino energy estimators experiments used compared to the known electron beam energy. Although data for carbon, iron, and helium-4 are available at 1.159, 2.257, and 4.453 GeV, we focus on comparisons with carbon using the lowest electron beam energy, 1.159 GeV, as this energy range is the least impacted by DIS and multi-pion production. The choice of carbon is to compare against neutrino experiments utilizing similar targets.

Figure 4 shows comparisons to the 0π0\pi cross section measurement by e4νe4\nu differential in EQEE_{\mathrm{QE}}, defined as

Figure 4. Comparison of the electron-carbon 0π0\pi differential cross section with respect to EQEE_{\mathrm{QE}} measured by e4νe4\nu and CLAS [70] compared to predictions by ACHILLES. The black vertical dashed line shows the true incoming beam energy.

EQE=2mNEe′+2mNEe′2−me22(mN−Ee′+pe′cos⁡θe′)(28)E_{\mathrm{QE}} = \frac{2m_N E_{e'} + 2m_N E_{e'}^2 - m_e^2}{2(m_N - E_{e'} + p_{e'} \cos\theta_{e'})} \tag*{(28)}

where the average separation energy ϵ=21\epsilon= 21 MeV was determined in the experimental analysis. This quantity is an energy estimator used by Cherenkov detectors like Super-Kamiokande and MiniBooNE where outgoing hadrons are often under detection threshold, and each neutrino event is assumed to be produced via quasi-elastic neutrino-nucleus scattering. The only relevant measured kinematic variables are the outgoing lepton energy and scattering angle. Events with EQEE_{\mathrm{QE}} reconstructing near the incoming beam energy are dominated by quasi-elastic scattering, whereas those in which a pion was produced and then absorbed during the cascade or was under threshold for detection reconstruct to lower values of EQEE_{\mathrm{QE}}. We note that the deficit of events in ACHILLES at large EQEE_{\mathrm{QE}} might be partially caused by the neglect of interference effects in the cascade and the effect of final-state interactions on the hard scattering matrix element which lead to a shift to lower energy transfer. The latter cause a modest shift in the peak, and MEC will also contribute a tail at large EQEE_{\mathrm{QE}} [71]. The deficit at low EQEE_{\mathrm{QE}} will be improved once MEC and DIS are incorporated. Both cascade modes give similar results in the region dominated by resonance production, meaning that pions are absorbed at similar rates in both models at this beam energy.

Moving to more exclusive final states, Figure 5 and Figure 6 show comparisons to the 1p0π1p0\pi cross sections differential in EcalE_{\mathrm{cal}} as well as PTP_{T}. This more restrictive signal reduces the contribution from resonance and DIS events. The calorimetric reconstructed energy is the neutrino energy estimator used by detectors like MINERνA, MicroBooNE, and the future DUNE experiment, where the energies of outgoing charged hadrons (and neutral mesons) can be measured. EcalE_{\mathrm{cal}} is in general defined as

Figure 5. Comparison of the electron-carbon 1p0π1p0\pi differential cross section with respect to EcalE_{\mathrm{cal}} measured by e4ν and CLAS [70] compared to predictions by Achilles. The black vertical dashed line shows the true incoming beam energy.

Figure 6. Comparison of the electron-carbon 1p0π1p0\pi differential cross section with respect to PTP_T measured by e4ν and CLAS [70] compared to predictions by Achilles.

Ecal=Eℓ′+∑Tp+∑Eπ±+∑EK±+∑Eγ+ϵ,(29)E_{\mathrm{cal}} = E_{\ell^\prime} + \sum T_p + \sum E_{\pi^\pm} + \sum E_{K^\pm} + \sum E_\gamma+ \epsilon, \tag*{(29)}

where TpT_p is the kinetic energy of the measured outgoing protons, Eπ±E_{\pi^\pm} (EK±E_{K^\pm}) is the energies of the measured outgoing charged pions (kaons), the sum of photon energies, and ϵ\epsilon is the average nucleon separation energy, taken here as 21 MeV. In this particular e4ν analysis, since the signal is one proton and zero pions, EcalE_{\mathrm{cal}} is given by

Ecal=Ee′+Tp+ϵ.E_{\mathrm{cal}} = E_{e^\prime} + T_p + \epsilon.

We find in Figure 5 that the peak near the beam energy is well described by Achilles and is dominated by quasi-elastic events. Resonance events contribute to smaller values of EcalE_{\mathrm{cal}} as they lead to knocked out neutrons or other undetected/under-threshold particles. Their contribution is diminished compared to the 0π0\pi selection because of the single proton requirement. Figure 6 shows the comparison with the cross-section differential in the transverse momentum imbalance, PTP_T, defined by

PT=∣pTe′+pTp∣.(30)P_T = \left|\mathbf{p}_T^{e^\prime} + \mathbf{p}_T^p\right|. \tag*{(30)}

PTP_T belongs to a class of observables called Transverse Kinematic Imbalance (TKI) variables which offer sensitivity to a range of nuclear affects. In contrast with the energy estimators, PTP_T is more sensitive to the kinematics of the outgoing nucleons from πN\pi N interactions in the INC. Further information about PTP_T and other TKI variables can be found in App. § 3.3. Figure 6 shows small differences between the two cascade modes in the resonance contribution.

We expect the additional strength needed to match the data will be provided by MEC and a small contribution from DIS.

Comparison to T2K

This section presents comparisons to measurements by the T2K collaboration for neutrino scattering on plastic scintillator (C8H8\mathrm{C}_8\mathrm{H}_8) in the T2K ND280 near detector with Ian average beam energy of ⟨Eν⟩≈600\langle E_\nu\rangle\approx600 MeV. The measurements include both charged-current pionless (CC0π) [72] and charged-current single-pion (CC1π⁺) events [73]. Events were generated with Achilles using the T2K beam flux from Ref. [74]. We note that the nucleon axial form factor used in these and all following comparisons is the zz-expansion form factor fit to deuterium data [75]. A detailed study of uncertainties from the form factors is left for when Achilles includes all interaction channels [76]. Comparisons were carried out by writing events to the NuHEPMC file format [77] and processing them through the nuisance framework [78].

Figure 7 shows comparisons to CC1π0π cross-section measurements by T2K, differential in the TKI variables δpT\delta p_T and δαT\delta\alpha_T defined in App. § 3.3. As expected for CC0π events at the beam energies of T2K, the contribution from resonance scattering followed by pion absorption is small compared to the dominant quasi-elastic scattering. The resonance contribution to the 0π0\pi sample in the Virtual Resonances cascade mode is larger than that from the Propagating Resonances mode, which is a reflection of the differences in the pion absorption cross section shown in Figure 3. The Achilles predictions are in agreement with the data, except at large δpT\delta p_T and large δαT\delta\alpha_T. In these regions MEC producing a two-nucleon final state are expected to contribute at the 10% level. Their inclusion will be the subject of future work and is expected to increase the agreement with the data. DIS interactions are expected to contribute at the percent level or below.

Figure 7. Left: Comparison of the CC0π differential cross section with respect to the transverse momentum imbalance δpT\delta p_{\mathrm{T}} measured by T2K [72] compared to predictions by ACHILLES. Right: Comparison of the CC0π differential cross section with respect to the transverse boosting angle δαT\delta\alpha_{\mathrm{T}} measured by T2K [72] compared to predictions by ACHILLES.

Figure 8 shows comparisons to CC1π+\pi^{+} cross-section measurements by T2K differential in the TKI variables pNp^{N} and δpTT\delta p_{\mathrm{TT}}. The agreement with the data is excellent, with contributions from resonant scattering accounting for nearly the entire cross section. This data set is complementary, as pion absorption decreases the contribution of resonance events in this sample, whereas it leads to an increase of events in the 0π\pi sample. In the 1π\pi sample this is reflected in the larger cross section predicted by the Propagating Resonances vs. Virtual Resonances mode, again consistent with the different pion absorption cross sections predicted by these two modes. We note that the contribution from DIS is expected to be small at these energies and its inclusion will not dramatically alter our conclusions.

Additional comparisons to T2K data can be found in Section F.

Comparison to MINERν\nuA

This section presents comparisons to measurements by the MINERν\nuA collaboration for neutrino scattering on plastic scintillator with an average beam energy of ⟨Eν⟩≈3\langle E_{\nu}\rangle\approx3 GeV [79]. Ref. [79] focused on CC0π\pi events with one or more final-state protons but zero final-state pions. Events were generated with ACHILLES using the MINERν\nuA beam flux from Ref. [80]. Comparisons were carried out by writing events to the NuHepMC file format [77] and processing them through the nuisance framework [78].

The neutrino beam at MINERν\nuA’s experimental setup in Ref. [79] is at substantially higher energies than T2K in Section IV B. Figure 9 shows the cross sections for the transverse boosting angle δαT\delta\alpha_{\mathrm{T}} and the reconstructed initial neutron momentum pnreconp_{\mathrm{n}}^{\mathrm{recon}} defined in Appendix § 3.3. Although MINERν\nuA’s CC0π\pi measurement has a similar signal definition to T2K’s, resonance scattering plays a much more important role at these energies. For example, at large δαT\delta\alpha_{\mathrm{T}}, roughly one third to a half of events come from resonance scattering. The absolute fractions for pnrecon>0.3p_{\mathrm{n}}^{\mathrm{recon}}>0.3 GeV are similar. Interestingly at these energies the difference between the two cascade modes’ predictions of the resonance contribution is enhanced, with the Virtual Resonances mode providing twice the predicted contribution from pion absorption when compared to the Propagating Resonances. This highlights the additional power of neutrino-nucleus scattering over hadron-nucleus scattering for constraining pion-nucleus interactions.

Figure 9. Left: Comparison of the CC0π differential cross section with respect to the transverse boosting angle δαT\delta\alpha_T measured by MINERνA [79] compared to predictions by Achilles. Right: Comparison of the CC0π differential cross section with respect to the reconstructed initial neutron momentum pnreconp_n^{\mathrm{recon}} measured by MINERνA [79] compared to predictions by Achilles.

In the region where quasi-elastic scattering is expected to dominate with only small contributions from FSI, i.e. at low δαT\delta\alpha_{\mathrm{T}}, the ACHILLES prediction is in excellent agreement with the data. The under prediction outside this region is due to the lack of MEC which is predicted to give a larger contribution at MINERν\nuA beam energies than in T2K. MEC interactions shift the δαT\delta\alpha_{\mathrm{T}} distribution towards higher values, due to the primary outgoing proton having only a fraction of the transferred momentum [81]. A similar discrepancy occurs at values of δpT\delta p_{\mathrm{T}} and pnrecon>0.3p_{\mathrm{n}}^{\mathrm{recon}}>0.3 GeV. Additional comparisons to MINERν\nuA data can be found in Section F.

We refrain from showing comparisons to MINERν\nuA Nπ\pi data as the contributions from DIS are comparable to that from resonance scattering.

Comparison to MicroBooNE

This final section presents comparisons to measurements by the MicroBooNE collaboration for neutrino scattering on liquid argon with an average beam energy of ⟨Eν⟩≈0.8\langle E_{\nu}\rangle\approx0.8 GeV [82], similar to T2K. Ignoring detector related effects, comparisons between T2K and MicroBooNE can be used to isolate AA-dependent nuclear effects. We compare to two measurements from MicroBooNE: Ref. [83] which focused on CC1π\pi0pp events with one final-state proton and zero final-state pions, and Ref. [82] which focused on NC1π0\pi^{0} events with and without protons. Events were generated with ACHILLES using the MicroBooNE Booster Neutrino Beam (BNB) flux from Ref. [82]. All ACHILLES predictions have been folded with the corresponding smearing matrix, defined in Ref. [84] as ACA_{\mathrm{C}}, in order to bring our predictions into the regularized space used by MicroBooNE [82, 83].

Figure 10 shows comparisons to CC1π\pi0π\pi cross-section measurements by MicroBooNE, double-differential in δpT\delta p_{\mathrm{T}} and δαT\delta\alpha_{\mathrm{T}}. Measurements of δpT\delta p_{\mathrm{T}} where δαT\delta\alpha_{\mathrm{T}} is small have been used to isolate nuclear structure effects, while at large δαT\delta\alpha_{\mathrm{T}} FSI and inelastic interactions dominate. ACHILLES reproduces the qualitative shape of the data in each bin, while the normalization is underpredicted. This deficit is unsurprising for bins with moderate to large δαT\delta\alpha_{\mathrm{T}} given the lack of MEC in the ACHILLES predictions. The lack of strength in the lowest δαT\delta\alpha_{\mathrm{T}} bin is notable, as this bin is supposed to have the smallest contribution from inelastic interactions.

Figure 10. Comparison of CC1p0π double-differential cross section measurement with respect to the transverse momentum δpT\delta p_T and the transverse boosting angle δαT\delta\alpha_T measured by MicroBooNE [83] in units of 10−39 cm2/GeV/degree10^{-39}\,\mathrm{cm}^2/\mathrm{GeV}/\mathrm{degree} compared to predictions by Achilles.

Reference [27] studied the same type of events with several different INCs and hard-scattering models and found a similar under prediction. We comment on several possible causes. First is an increase in the axial form factor consistent with that found by Lattice QCD calculations [86, 87, 88], and a recent measurement by MINERν\nuA [89, 90]. Both point to an axial form factor which is over 20% larger at Q2>0.4Q^{2}>0.4 GeV2^{2} than that extracted from neutrino-deuterium scattering [75], leading to a 10-20% increase in CC0π\pi cross sections at BNB energies. However, this increase would have to be reconciled with high Q2Q^{2} data from other neutrino-nucleus cross section measurements. Second is a lack of interference between one- and two-body currents leading to one-nucleon knockout [91, 92, 93]. In Ref. [92] this interference was shown to lead to a 10% increase in the neutrino-nucleus cross section on carbon at MiniBooNE. On an asymmetric nucleus like argon with more npnp pairs than carbon, this enhancement could be even larger. Lastly, we note that migration between bins caused by the ACA_C matrix could mean that this low δαT\delta\alpha_{\mathrm{T}} receives a large contribution from MEC when measured in the regularized space reported by MicroBooNE. This effect will be investigated further once MEC are included in ACHILLES predictions.

Figures 11 and 12 show our last two comparisons to NCπ⁰Xp cross-sections from MicroBooNE, differential in π⁰ scattering angle and double-differential in π⁰ momentum. These events contain a single π⁰ and any number of additional hadrons. Neutral current π⁰ production is an important background in electron neutrino appearance measurements as the two photons from π⁰ decay

can mimic the electromagnetic shower from a single electron. We find that the ACHILLES prediction is in excellent agreement with the data for both the single and double differential cross sections. The Propagating Resonances cascade mode consistently predicts a cross-section 10 to 15% larger than the Virtual Resonances mode. This is consistent with the differences between the two cascade modes’ predictions for T2K CC1π⁺ cross sections shown in Figure 7. This is most likely from the larger absorption cross section predicted by the Virtual Resonances mode, but could also stem from the different charge-exchange cross sections predicted by the DCC and GiBUU Δ\Delta models.

The exclusive predictions provided by the ACHILLES DCC implementation are key to obtaining the correct kinematics for the outgoing π⁰. In the future, comparison of the cross-section split into 0p and Np final states will allow us to constrain the isospin dependence of the pion absorption models in Achilles. We note that Achilles does not currently include NC coherent pion production which contributes mostly at forward pion scattering angles, or DIS which is expected to be subdominant at BNB energies.

Conclusions

The present work has extended the Achilles Monte Carlo event generator by incorporating single-pion production in a fully exclusive fashion. The electroweak interaction vertex is modeled by combining the state-of-the-art Dynamical Coupled-Channels (DCC) approach with realistic hole spectral functions, which account for correlations in both the initial target state and the residual spectator system. We have validated the hard interaction vertex by comparing to high precision inclusive electron scattering data from JLab, as well as bubble-chamber data from the ANL and BNL experiments that used elementary nucleon targets. These comparison serve to validate both the implementation of the DCC model in Achilles as well as its ability to describe quantitatively experimental data from both vector and axial-vector probes.

The present work has also extended the Achilles INC to include pions using two different treatments of resonances to furnish an estimate of theoretical uncertainty. In the first model, resonances have been fully integrated out so that the only propagating degrees of freedom are nucleons and pions; the required πN\pi N scattering amplitudes are provided DCC and the Oset optical-potential models. In the second model, all πN\pi N interactions during the INC are mediated by propagating Δ\Delta resonances, which are produced and decayed as explicit propagating degrees of freedom along side pions and nucleons. The two cascade modes qualitatively reproduce hadron-nucleus scattering observables like the total pion reaction cross section on carbon and the pion-absorption cross section on carbon and argon. In some instances (e.g., the total reaction cross section of π+\pi^+ scattering on 12C^{12}\mathrm{C} for pπ≳0.5p_\pi\gtrsim0.5 GeV or the π+\pi^+--12C^{12}\mathrm{C} absorption cross section near the Δ\Delta peak), differences between the modes’ predictions as well as differences with the data suggest that additional in-medium effects and interaction mechanisms in the cascade will be important to improve agreement with data. Additional study of these effects will be the subject of future work.

Figure 11. Comparison of NC1π0\pi^0 double-differential cross section measurement with respect to the scattering angle cos⁡θπ0\cos\theta_{\pi^0} and momentum pπ0p_{\pi^0} of the final-state π0\pi^0 measured by MicroBooNE [85] in units of 10−39 cm2/GeV/nucleon10^{-39}\,\mathrm{cm}^2/\mathrm{GeV}/\mathrm{nucleon} compared to predictions by Achilles.

Figure 12. Comparison of NC1π0\pi^0 differential cross section measurement with respect to the π0\pi^0 scattering angle measured by MicroBooNE [85] compared to predictions by Achilles.

We compared Achilles predictions to exclusive electron- and neutrino-nucleus scattering data for a range of beam energies and targets. Section IV A presented comparisons to recent measurements from CLAS and e4ν\nu for electron-carbon scattering, focusing on quasielastic-like 1e1p0π\pi events. Similar comparisons were carried out previously in Ref. [31] including QE events only. As expected, including single-pion production has improved agreement between Achilles’ predictions and the experimental data. Section IV B presented comparisons to measurements of neutrino scattering on plastic scintillator at ⟨Eν⟩≈0.6\langle E_\nu\rangle\approx0.6 GeV by T2K including CC0π\pi and CC1π\pi events for TKI observables. Few-sigma qualitative agreement in essentially all kinematic bins was observed. Section IV C also compared to neutrino scattering on plastic scintillator, this time for measurements at higher energies (⟨Eν⟩≈3\langle E_\nu\rangle\approx3 GeV) from MINERvA. As expected, resonance events make up a larger fraction of the total cross section, but there is also significant contributions from MEC and DIS which our prediction is currently missing. At these energies, observables like the transverse boosting angle δαT\delta\alpha_T and reconstructed initial neutron momentum pnreconp_n^{\mathrm{recon}} were found to be sensitive to difference in the pion mode (virtual resonances vs propagating resonances) used in the INC. Finally, Section IV D gave comparisons to neutrino-argon scattering with ⟨Eν⟩≈0.8\langle E_\nu\rangle\approx0.8 GeV at MicroBooNE. Achilles under-predicts the CC0π\pip double-differential cross section dσ/dδpT dδαTd\sigma/d\delta p_T\,d\delta\alpha_T, a feature which has been noted previously in predictions from other generators [27]. Better agreement is observed for NCπ0\pi^0Xp cross sections, where ACHILLES reproduces the experimental data at the ≈1σ\approx1\sigma level in essentially all kinematics bins.

Overall, the inclusion of single-pion production in ACHILLES represents an important milestone toward our goal of including all physical processes relevant for the accelerator-based neutrino program. In the near future, we also plan to include support for MEC and DIS. The results of the present work (e.g., in comparisons to MINERν\nuA data) highlight the importance and utility of neutrino cross-section measurements to distinguish between different model treatments of nuclear effects. It will be interesting and important to revisit these effects in the future once ACHILLES also includes MEC and DIS.

Finally, once all interaction mechanisms are present, providing a complete error estimate will be important to determine necessary paths for improvement to meet the needs of the DUNE experiment.

Acknowledgments

We would like to thank Afroditi Papadopoulou for sharing the details of the e4νe4\nu and MicroBooNE 1p0π\pi results and Ben Bogart for providing comparisons with the MicroBooNE NCπ0\pi^0 measurement. We would like to thank Luke Pickering with help using the Nuisance analysis framework, and for help with the NuHepMC interface to Nuisance. We would like to thank Afroditi Papadopoulou for comments on the manuscript.

This manuscript has been authored by Fermi Forward Discovery Group, LLC under Contract No. 89243024CSC000002 with the U.S. Department of Energy, Office of Science, Office of High Energy Physics. The present research is supported by the U.S. Department of Energy, Office of Science, Office of Nuclear Physics, under contracts DE-AC02-06CH11357 (A. L.), by the DOE Early Career Research Program (A. L.), by the SciDAC-5 NeuCol program (J. I., A. L., and N. R.), by the Neutrino Theory Network (A. N.), and by the Nuclear Theory for New Physics Topical collaboration (A. L., N. R., and N. S.).

This work was supported in part by Colorado State University through high-performance computer time and resources provided by the Data Science Research Institute.

Feynman diagrams were produced through the use of Tikz-Feynman [94]. Figures were generated using matplotlib [95] and seaborn [96].

The data in this paper can be reproduces using version 0.3.0 of Achilles [97] with the run cards and Nuisance analyses found in Ref. [98].

Phase-space handling

The following is based closely on the TASI lecture notes from S. Höche [99]. In general, the differential final-state phase space element for a 2→(n−2)2 \to(n-2) scattering is

dΦn({p⃗})=[∏i=3nd4pi(2π)3δ(pi2−mi2)Θ(Ei)](2π)4δ(4)(p1+p2−∑i=3npi)(31)\mathrm{d}\Phi_n\left(\{\vec{p}\}\right)=\left[\prod_{i=3}^{n}\frac{\mathrm{d}^4p_i}{(2\pi)^3}\delta\left(p_i^2-m_i^2\right)\Theta\left(E_i\right)\right](2\pi)^4\delta^{(4)}\left(p_1+p_2-\sum_{i=3}^{n}p_i\right) \tag*{(31)}

where p1p_1 and p2p_2 are initial state momenta, mim_i are the on-shell masses of the outgoing particles. Generic techniques for handling high-dimensional phase-space integrals were proposed in Ref. [100]. The authors showed that it is possible to factorize the phase space into three components, which correspond to 2→22 \to2 scattering, 1→21 \to2 decay, and 2→12 \to1 annihilation processes. Eq. (31) factorizes as [101]

dΦn(p1,p2;p3,…,pn)=dΦn−m+1(p1,p2;p3,…,pn−m,P)dP22πdΦm(P;pn−m+1,…,pn),(32)\mathrm{d}\Phi_n(p_1,p_2;p_3,\ldots,p_n)=\mathrm{d}\Phi_{n-m+1}(p_1,p_2;p_3,\ldots,p_{n-m},P)\frac{\mathrm{d}P^2}{2\pi}\mathrm{d}\Phi_m(P;p_{n-m+1},\ldots,p_n), \tag*{(32)}

where PP denotes a virtual intermediate particle. While the particle PP has no direct physical interpretation, it may be associated with an ss-channel propagator formed by the set of external states {pn−m+1,…,pn}\{p_{n-m+1},\ldots,p_n\}. Leveraging knowledge of the Feynman diagrams involved in the matrix element allows one to efficiently map out the peak structure of certain diagrams squared. Therefore, the ability to match PP to a propagator in the Feynman diagram enables efficient Monte-Carlo sampling of the integral. However, each diagram can have many different combinations of propagator structures, leading to the need to use the multi-channel method to find the optimal integrator over the full matrix element squared [102]. Repeating the process in Eq. (32) allows a decomposition of the complete phase space into three elementary building blocks defined as

dΦ2(pa,pb;pi,pj)=λ(sab,si,sj)16π2sabdcos⁡θidϕi,dΦ2(pi;pi,pj)=λ(sij,si,sj)16π2sijdcos⁡θidϕi,dΦ1(pa,pb;pi)=(2π)4d4piδ(4)(pa+pb−pi),(33)\begin{aligned} \mathrm{d}\Phi_2(p_a,p_b;p_i,p_j)&=\frac{\lambda(s_{ab},s_i,s_j)}{16\pi^2s_{ab}}\mathrm{d}\cos\theta_i\mathrm{d}\phi_i,\\ \mathrm{d}\Phi_2(p_i;p_i,p_j)&=\frac{\lambda(s_{ij},s_i,s_j)}{16\pi^2s_{ij}}\mathrm{d}\cos\theta_i\mathrm{d}\phi_i,\\ \mathrm{d}\Phi_1(p_a,p_b;p_i)&=(2\pi)^4\mathrm{d}^4p_i\delta^{(4)}(p_a+p_b-p_i), \tag*{(33)} \end{aligned}

where λ\lambda is the Källén function. The different building blocks can be associated with tt- and ss-channel vertices, while the dP2/2π\mathrm{d}P^2/2\pi can be associated with a propagator. This enables the usage of techniques developed for automatic matrix element calculations to also be used for generating the phase space, as is the case in the automated matrix element calculator Comix [103].

Meson-baryon scattering amplitudes

We express the cross section for meson-baryon scattering

mi(ki)+Bi(−ki)→mf(kf)+Bf(−kf)(34)m_i(\mathbf{k}_i)+B_i(-\mathbf{k}_i)\to m_f(\mathbf{k}_f)+B_f(-\mathbf{k}_f) \tag*{(34)}

with ki\mathbf{k}_i, kf\mathbf{k}_f the momenta in the center-of-mass (CM) frame, in terms of standard partial-wave amplitudes (PWA) τi,fL,±,I\tau_{i,f}^{L,\pm,I}. The PWA have well-defined pion-nucleon orbital angular momentum LL, total angular momentum J=L±1/2J=L\pm1/2, and total isospin II. The indices ii, ff denote the initial, final meson baryon pair πN\pi N, ηN\eta N, KΛK\Lambda, KΣK\Sigma.

From isospin conservation the PWA for each channel ii, ff can be decomposed in terms of total isospin II of the meson-baryon pair. I.e. I=1/2I=1/2 or 3/23/2 for πN\pi N and KΛK\Lambda, and I=1/2I=1/2 for ηN\eta N and KΣK\Sigma. The PWA for a specific process mB→m′B′mB\to m'B', are then given by

τmB,m′B′L,±=∑I(Im,I3m;IB,I3B∣I,I3)(I,I3∣Im′,I3m′;IB′,I3B′)τi,fL,±,I(35)\tau_{mB,m'B'}^{L,\pm}=\sum_I \left(I^m,I_3^m;I^B,I_3^B\mid I,I_3\right)\left(I,I_3\mid I^{m'},I_3^{m'};I^{B'},I_3^{B'}\right)\tau_{i,f}^{L,\pm,I} \tag*{(35)}

where Im/BI^{m/B}, I3m/BI_3^{m/B} denote the total isospin and its projection of the meson/baryon. The Clebsch-Gordan coefficients (Im,I3m;IB,I3B∣I,I3)(I^m,I_3^m;I^B,I_3^B\mid I,I_3) couple the isospin of the meson-baryon pair in the initial and final state to total isospin II. In the following we express the cross section for a specific process in terms of PWA. We drop the subscript mB,m′B′mB,m'B' for ease of notation.

The differential cross section in the CM frame can be written in terms of two amplitudes

dσdΩ=1∣ki∣2ρiρf(∣T++(s,Ω)∣2+∣T+−(s,Ω)∣2)=∣kf∣∣ki∣(∣h(s,Ω)∣2+∣g(s,Ω)∣2).(36)\frac{\mathrm{d}\sigma}{\mathrm{d}\Omega}=\frac{1}{|\mathbf{k}_i|^2}\rho_i\rho_f\left(|T_{++}(s,\Omega)|^2+|T_{+-}(s,\Omega)|^2\right)=\frac{|\mathbf{k}_f|}{|\mathbf{k}_i|}\left(|h(s,\Omega)|^2+|g(s,\Omega)|^2\right). \tag*{(36)}

Here T++,T+−T_{++},T_{+-} are helicity flip and non-flip amplitudes, h,gh,g are spin-flip and non-flip amplitudes respectively. The phase-space factors for initial and final state ρi,ρf\rho_i,\rho_f are given by

ρ(k)=πkWEmEB(37)\rho(k)=\pi\frac{k}{W}E_mE_B \tag*{(37)}

with Em,EBE_m,E_B the CM energy of the meson and baryon, respectively, and kk the CM momentum. The helicity amplitudes are expanded in terms of the total angular momentum of the meson-baryon system JJ

Tλ,λ′(s,Ω)=∑J(J+12)Tλ,λ′J(s)ei(λ−λ′)ϕdλ,λ′J(θ).(38)T_{\lambda,\lambda'}(s,\Omega)=\sum_J\left(J+\frac{1}{2}\right)T^J_{\lambda,\lambda'}(s)e^{i(\lambda-\lambda')\phi}d^J_{\lambda,\lambda'}(\theta). \tag*{(38)}

All angular dependence of Eq. (36) is given by the Wigner small-dd functions. For helicities λ=±1/2\lambda=\pm1/2 they are related to Legendre polynomials PlP_l as

(J+12)d±12,12J(θ)=1±cos⁡θ2(PJ+1/2′(cos⁡θ)∓PJ−1/2′(cos⁡θ)).(39)\left(J+\frac{1}{2}\right)d^J_{\pm\frac{1}{2},\frac{1}{2}}(\theta)=\sqrt{\frac{1\pm\cos\theta}{2}}\left(P'_{J+1/2}(\cos\theta)\mp P'_{J-1/2}(\cos\theta)\right). \tag*{(39)}

where the primes denote the derivative with respect to cos⁡θ\cos\theta. The relation between the helicity amplitudes and amplitudes with definite parity, and hence LL, is

2TL,±(s)=T++J=L−1/2(s)±T++J=L+1/2(s).(40)2T^{L,\pm}(s)=T^{J=L-1/2}_{++}(s)\pm T^{J=L+1/2}_{++}(s). \tag*{(40)}

The (dimensionless) PWA are then defined as

τL,±(s)=−ρiρf TL,±(s).(41)\tau^{L,\pm}(s)=-\sqrt{\rho_i\rho_f}\,T^{L,\pm}(s). \tag*{(41)}

We have explicitly

ρiρf T++(s,θ)=∑J(J+12)d±12,12J(θ)(τJ−1/2,+(s)±τJ+1/2,−(s)).(42)\sqrt{\rho_i\rho_f}\,T_{++}(s,\theta)=\sum_J\left(J+\frac{1}{2}\right)d^J_{\pm\frac{1}{2},\frac{1}{2}}(\theta)\left(\tau^{J-1/2,+}(s)\pm\tau^{J+1/2,-}(s)\right). \tag*{(42)}

The differential cross section of Eq. (36) is then most conveniently expressed as

dσdΩ=1∣ki∣2∣∑LPL(cos⁡θ)[LτL,−(s)+(L+1)τL,+(s)]∣2+sin⁡2θ∣ki∣2∣∑LPL′(cos⁡θ)[τL,+(s)−τL,−(s)]∣2.(43)\frac{\mathrm{d}\sigma}{\mathrm{d}\Omega} = \frac{1}{\lvert k_i\rvert^2}\left\lvert\sum_L P_L(\cos\theta)\left[L\tau^{L,-}(s)+(L+1)\tau^{L,+}(s)\right]\right\rvert^2+\frac{\sin^2\theta}{\lvert k_i\rvert^2}\left\lvert\sum_L P'_L(\cos\theta)\left[\tau^{L,+}(s)-\tau^{L,-}(s)\right]\right\rvert^2 . \tag*{(43)}

Here the first and second terms are proportional to the spin-flip and non-flip amplitudes ∣h∣2,∣g∣2\lvert h\rvert^2,\lvert g\rvert^2 respectively.

Efficient sampling of the CM scattering angle is readily obtained via the inverse sampling transform using the cumulative distribution function. To do so, we express the probability density function for the scattering angle, given by the normalized angular distribution of Eq. (43), in a monomial basis. This is done using the expansion coefficients Gkl,nG_k^{l,n} and Hkl,nH_k^{l,n} defined as

∑k=0l+nHkl,nxk≡(1−x2)Pl′(x)Pn′(x),(44)\sum_{k=0}^{l+n}H_k^{l,n}x^k\equiv(1-x^2)P'_l(x)P'_n(x), \tag*{(44)}
∑k=0l+nGkl,nxk≡Pl(x)Pn(x).(45)\sum_{k=0}^{l+n}G_k^{l,n}x^k\equiv P_l(x)P_n(x). \tag*{(45)}

These coefficients are multiplied by the PWA as in Eq. (43), yielding the angular distribution. The cumulative distribution function is the integral of the resulting polynomial, given analytically in terms of the same coefficients.

From the orthogonality of the associated Legendre polynomials the angle-integrated cross section is given by squared PWA

σ(W)=4π∣ki∣2∑L(L∣τL,−(s)∣2+(L+1)∣τL,+(s)∣2).(46)\sigma(W)=\frac{4\pi}{\lvert k_i\rvert^2}\sum_L\left(L\lvert\tau^{L,-}(s)\rvert^2+(L+1)\lvert\tau^{L,+}(s)\rvert^2\right). \tag*{(46)}

Validation of Meson-Baryon Scattering

We validate the results for both the total cross sections, and angular distributions from the ANL-Osaka DCC model for meson-proton scattering by running the INC simulation for a stationary proton target. Results are shown in Figure 13 for incoming pions and eta mesons, validating the agreement between the Achilles INC and the original ANL-Osaka model. For the angular distributions in on the right of Figure 13, the pion isospin channels not shown have almost exactly the same angular distribution at this momenta, as this kinematic point is almost exactly on top of the Δ\Delta peak, where the different pion cross sections differ only by isospin factors. Results for neutron and Hyperon targets can be obtained in a similar manner.

Validation of the Achilles implementation of the ANL-Osaka DCC Model

Figure 13. Validation of the Achilles implementation of the ANL-Osaka DCC Model. Left: Angle integrated cross sections for π+\pi^{+} (red), π0\pi^{0} (blue), π−\pi^{-} (green), and η\eta (purple) scattering off the proton. Results only include the contributions for meson-baryon final-state channels πN\pi N, ηN\eta N, KΛK\Lambda, KΣK\Sigma. Results shown by histograms are computed from the Achilles INC, while the dashed lines show the calculation directly from Eq. (46). Right: Angular distributions in the CM frame for π+\pi^{+} (red), and η\eta (purple) scattering off the proton at pπ/η=300p_{\pi/\eta}=300 MeV. Results shown by points are computed from the Achilles INC, while the dashed lines show the calculation from the ANL-Osaka model. The angular distributions for π0\pi^{0} and π−\pi^{-} are identical to π+\pi^{+} at these kinematics.

As a final check, Figure 14 shows the total cross sections for charged and neutral pion production off of a hydrogen target using the results from the GiBUU Δ\Delta model. To compare with our predictions, we utilize the parameterizations of the total cross sections from GiBUU [29], which have been fit to global NN→NNπNN \to NN\pi scattering data. Here one can see that the predictions from Achilles lie slightly below the data as we only include contributions from the resonant process NN→NΔ→NNπNN \to N\Delta\to NN\pi. The missing strength is given by the non-resonant, background NN→NNπNN \to NN\pi contributions which are for now not included in Achilles.

Figure 14. Comparison of results between Achilles and GiBUU. Left: Pion production cross sections versus s\sqrt{s} and pbeamp_{\mathrm{beam}} for pp→pnπ+pp \to pn\pi^{+} (blue) and pp→ppπ0pp \to pp\pi^{0} (red) in Achilles compared to the parameterized results from GiBUU [29]. Right: Comparison of Δ\Delta cross sections described in Eqs. (19), (20) against the cross sections obtained from GiBUU [29].

Delta Matrix Elements

Figure 15 shows the diagrams in the tree-level calculation for pp→nΔ++pp \to n\Delta^{++}; other channels follow using isospin symmetry and Clebsch–Gordan coefficients. The cross section is sum of three contributions: a direct term, a crossed term, and their interference. The direct contribution is given by

Figure 15. Tree-level Δ\Delta production diagrams arising from the Lagrangian Eq. (22). A solid lines denotes a nucleon, the double lines denotes a Δ\Delta, and a dotted line denotes a pion.

∣M(direct)∣2=z(t,μΔ)[fπfπ∗F2(t)mπ2(t−mπ2)]24mN2t3μΔ2×[t−(μΔ−mN)2][(μΔ+mN)2−t]2.(47)\left|\mathcal{M}(\mathrm{direct})\right|^{2} = z(t,\mu_{\Delta})\left[\frac{f_{\pi}f_{\pi}^{*}F^{2}(t)}{m_{\pi}^{2}(t-m_{\pi}^{2})}\right]^{2}\frac{4m_{N}^{2}t}{3\mu_{\Delta}^{2}}\times\left[t-(\mu_{\Delta}-m_{N})^{2}\right]\left[(\mu_{\Delta}+m_{N})^{2}-t\right]^{2}. \tag*{(47)}
ProcessIsospin Factor (I‾\overline{I})
pΔ++→pΔ++p\Delta^{++}\rightarrow p\Delta^{++}9/49/4
pΔ+→nΔ++p\Delta^{+}\rightarrow n\Delta^{++}33
pΔ+→pΔ+p\Delta^{+}\rightarrow p\Delta^{+}1/41/4
pΔ0→pΔ0p\Delta^{0}\rightarrow p\Delta^{0}1/41/4
pΔ0→nΔ+p\Delta^{0}\rightarrow n\Delta^{+}44
pΔ−→pΔ−p\Delta^{-}\rightarrow p\Delta^{-}9/49/4
pΔ−→nΔ0p\Delta^{-}\rightarrow n\Delta^{0}33

Table 2. Isospin factors for the process NΔ→NΔN\Delta\rightarrow N\Delta. The neutron channels are given by isospin symmetry.

The crossed contribution follows from the exchange of the Mandelstam variables t↔ut \leftrightarrow u. Finally, the interference term is given as

2Re⁡[M(direct)M∗(crossed)]=[fπfπ∗F2(t)mπ2(t−mπ2)][fπfπ∗F2(u)mπ2(u−mπ2)]z(t,μΔ)z(u,μΔ)(2mN2μΔ2)×{[tu+(μΔ2−mN2)(t+u)−μΔ4+mN4][tu+mN(μΔ+mN)(μΔ2−mN2)]\begin{aligned} 2\operatorname{Re}[\mathcal{M}(\text{direct})\mathcal{M}^{*}(\text{crossed})] = \left[\frac{f_{\pi}f_{\pi}^{*}F^{2}(t)}{m_{\pi}^{2}(t-m_{\pi}^{2})}\right]\left[\frac{f_{\pi}f_{\pi}^{*}F^{2}(u)}{m_{\pi}^{2}(u-m_{\pi}^{2})}\right]\sqrt{z(t,\mu_{\Delta})z(u,\mu_{\Delta})}\left(\frac{2m_{N}^{2}}{\mu_{\Delta}^{2}}\right) \\ &\times\{[tu+(\mu_{\Delta}^{2}-m_{N}^{2})(t+u)-\mu_{\Delta}^{4}+m_{N}^{4}][tu+m_{N}(\mu_{\Delta}+m_{N})(\mu_{\Delta}^{2}-m_{N}^{2})] \end{aligned}
−13[tu−(μΔ+mN)2(t+u)+(μΔ+mN)4][tu−mN(μΔ−mN)(μΔ2−mN2)]}.(48)-\frac{1}{3}[tu-(\mu_{\Delta}+m_{N})^{2}(t+u)+(\mu_{\Delta}+m_{N})^{4}][tu-m_{N}(\mu_{\Delta}-m_{N})(\mu_{\Delta}^{2}-m_{N}^{2})]\}. \tag*{(48)}

The comparison between the implementation and the calculation within GiBUU can be seen in the right panel of Figure 14. The agreement between the two implementations are at an acceptable level and may differ from exact values of couplings and masses.

The matrix element for the process NΔ→NΔN\Delta\rightarrow N\Delta is

∣MNΔi→NΔf∣2=I18(fNNπfΔΔπmπ)2F4(t)(t−mπ2)2×16(μi+μf)2+mN2t9μi2μf2×(−μi2+2μiμf−μf2+t)×(μi4−2μi3μf+12μi2μf2−2μiμf3+μf4−2μit+2μiμft−2μf2t+t2).(49)\begin{aligned} \left|\mathcal{M}_{N\Delta_{i}\rightarrow N\Delta_{f}}\right|^{2}=I\frac{1}{8}\left(\frac{f_{NN\pi}f_{\Delta\Delta\pi}}{m_{\pi}}\right)^{2}\frac{F^{4}(t)}{(t-m_{\pi}^{2})^{2}}\times\frac{16(\mu_{i}+\mu_{f})^{2}+m_{N}^{2}t}{9\mu_{i}^{2}\mu_{f}^{2}}\times(-\mu_{i}^{2}+2\mu_{i}\mu_{f}-\mu_{f}^{2}+t) \\ &\times\left(\mu_{i}^{4}-2\mu_{i}^{3}\mu_{f}+12\mu_{i}^{2}\mu_{f}^{2}-2\mu_{i}\mu_{f}^{3}+\mu_{f}^{4}-2\mu_{i}t+2\mu_{i}\mu_{f}t-2\mu_{f}^{2}t+t^{2}\right). \tag*{(49)} \end{aligned}

In the above equation, F(t)F(t) is the same form factor defined in Eq. (24), μi\mu_{i} and μf\mu_{f} are the masses of the off-shell Δ\Delta resonance in the initial and final state, respectively. The NNπNN\pi coupling constant fNNπ=0.946f_{NN\pi}=0.946 and the ΔΔπ\Delta\Delta\pi coupling constant is given as fΔΔπ=9/5fNNπf_{\Delta\Delta\pi}=9/5f_{NN\pi}. Finally, the isospin factors (I‾\overline{I}) are given in Tab. II.

Transverse Kinematic Imbalance

Transverse kinematic imbalance (TKI) variables are are popular observables designed for sensitivity to nuclear effects in neutrino-nucleus scattering [104]. To establish notation, consider CCNp0πNp0\pi scattering events ν+A→ℓ+p+X\nu+A\rightarrow\ell+p+X, where ℓ\ell and pp are respectively the outgoing lepton and leading proton. The transverse momentum δpT\delta p_{T} is defined via the vector sum of the outgoing transverse lepton and leading proton momentum

δpT≡pTℓ′+pTN′.(50)\delta p_{T}\equiv p_{T}^{\ell'}+p_{T}^{N'}. \tag*{(50)}

The notation δpT\delta p_{T} is meant as a visual reminder that this transverse momentum would vanish in the absence of nuclear effects. As for any three-vector, δpT\delta p_{T} is completely specified by its magnitude and two angles,

δpT=∣pTℓ′+pTN′∣(51)\delta p_{T}=\left|p_{T}^{\ell'}+p_{T}^{N'}\right| \tag*{(51)}
δαT≡arccos⁡(−pTℓ′⋅δpTpTℓ′δpT)(52)\delta\alpha_{T}\equiv\arccos\left(\frac{-p_{T}^{\ell'}\cdot\delta p_{T}}{p_{T}^{\ell'}\delta p_{T}}\right) \tag*{(52)}
δϕT≡arccos⁡(−pTℓ′⋅pTN′pTℓ′pTN′),(53)\delta\phi_{T}\equiv\arccos\left(\frac{-p_{T}^{\ell'}\cdot p_{T}^{N'}}{p_{T}^{\ell'}p_{T}^{N'}}\right), \tag*{(53)}

where pTℓ′p_{T}^{\ell'} and pTN′p_{T}^{N'} are the projections of the momentum of the outgoing lepton and leading proton onto the transverse plane defined by the incoming neutrino momentum. The angle δαT\delta\alpha_{T} is referred to as the transverse boosting angle, while the angle δφT\delta\varphi_T is the transverse deflection angle. Ref. [104] discusses physical expectations for these observables and provides a diagram of the kinematic setup. Note that these same variables are used in electron-nucleus scattering as well, often denoted as pTp_T.

Doubly transverse variables [105, 106] can be defined analogously for CCN1π+\mathrm{CC}N1\pi^+ events ν+A→ℓ′+p′+π′+X\nu+ A \to\ell' + p' + \pi' + X. A double-transverse axis z^TT\hat{z}_{TT} is defined with respect to the plane of the initial- and final-state lepton momenta,

z^TT≡pν×pℓ′∣pν×pℓ′∣.(54)\hat{z}_{TT} \equiv\frac{\mathbf{p}^{\nu} \times\mathbf{p}^{\ell'}}{\lvert\mathbf{p}^{\nu} \times\mathbf{p}^{\ell'} \rvert}. \tag*{(54)}

Projecting the resonance momentum pN′+pπ′\mathbf{p}^{N'} + \mathbf{p}^{\pi'}, with pN′\mathbf{p}^{N'} the leading proton, onto the double-transverse axis gives the double-transverse momentum imbalance,

δpTT≡(pN′+pπ′)⋅z^TT.(55)\delta p_{TT} \equiv\left(\mathbf{p}^{N'} + \mathbf{p}^{\pi'}\right) \cdot\hat{z}_{TT}. \tag*{(55)}

The magnitude of the initial nucleon momentum pN\mathbf{p}^{N} is defined in terms of it’s components transverse and longitudinal to the initial neutrino momentum,

pN=(δpTCC1π)2+(pLCC1π)2.(56)p^{N} = \sqrt{\left(\delta p_{T}^{\mathrm{CC}1\pi}\right)^2 + \left(p_{L}^{\mathrm{CC}1\pi}\right)^2}. \tag*{(56)}

The transverse momentum in CCN1π+\mathrm{CC}N1\pi^+ events is given by

δpTCC1π=pTℓ′+pTN′+pTπ′.(57)\delta\mathbf{p}_{T}^{\mathrm{CC}1\pi} = \mathbf{p}_{T}^{\ell'} + \mathbf{p}_{T}^{N'} + \mathbf{p}_{T}^{\pi'}. \tag*{(57)}

The longitudinal component pLp_L can similarly be computed in terms of final-state observables, although the expression is somewhat longer and will not be needed here. Details can be found in Ref. [73].

Additional Experimental Comparisons

This appendix collects additional comparisons between experimental measurements and the predictions of ACHILLES. Figure 16 shows a comparison to the CC0π\mathrm{CC}0\pi differential cross sections measured by MINERν\nuA [79] with respect to several different kinematic variables. Figure 17 shows a comparison of measurements by T2K.

Figure 16. Comparison of CC0π differential cross sections measured by MINERνA [79] compared to predictions by Achilles. In the first row, the cross sections are differential in momentum ppp_p and scattering angle θp\theta_p. In the second row, the cross sections are differential in momentum pμp_\mu and scattering angle θμ\theta_\mu of the outgoing muon. In the third row, the cross sections are differential in the transverse momentum imbalance δpT\delta p_T and the transverse deflecting angle δϕT\delta\phi_T.

Figure 17. Left: Comparison of the CC1π+\pi^{+} differential cross section with respect to the transverse boosting angle δαT\delta\alpha_{T} measured by T2K [73] compared to predictions by ACHILLES. Right: Comparison of the CC0π\pi differential cross section with respect to the transverse deflecting angle δϕT\delta\phi_{T} measured by T2K [72].

References

References

  1. [1]B. Abi et al. (DUNE), Eur. Phys. J. C 80, 978 (2020), arXiv:2006.16043 [hep-ex].DOI
  2. [2]K. Abe et al. (Hyper-Kamiokande), (2018), arXiv:1805.04163 [physics.ins-det].arxiv.org/abs/1805.04163
  3. [3]A. Lovato, J. Carlson, S. Gandolfi, N. Rocco, and R. Schiavilla, Phys. Rev. X 10, 031068 (2020), arXiv:2003.07710 [nucl-th].DOI
  4. [4]B. Acharya, J. E. Sobczyk, S. Bacca, G. Hagen, and W. Jiang, Phys. Rev. Lett. 134, 202501 (2025), arXiv:2410.05962 [nucl-th].DOI
  5. [5]R. Alvarez Garrote (SBND), PoS ICHEP2024, 135 (2025), arXiv:2501.11349 [hep-ex].
  6. [6]A. Roggero, A. C. Y. Li, J. Carlson, R. Gupta, and G. N. Perdue, Phys. Rev. D 101, 074038 (2020), arXiv:1911.06368 [quant-ph].
  7. [7]R. Weiss, A. Baroni, J. Carlson, and I. Stetcu, (2024), arXiv:2404.00202 [quant-ph].DOI
  8. [8]C. Andreopoulos et al., Nucl. Instrum. Meth. A 614, 87 (2010), arXiv:0905.2517 [hep-ph].DOI
  9. [9]T. Golan, J. T. Sobczyk, and J. Zmuda, Nucl. Phys. B Proc. Suppl. 229-232, 499 (2012).
  10. [10]U. Mosel, Ann. Rev. Nucl. Part. Sci. 66, 171 (2016), arXiv:1602.00696 [nucl-th].DOI
  11. [11]Y. Hayato and L. Pickering, Eur. Phys. J. ST 230, 4469 (2021), arXiv:2106.15809 [hep-ph].DOI
  12. [12]E. Hernandez, J. Nieves, and M. Valverde, Phys. Rev. D 76, 033005 (2007), arXiv:hep-ph/0701149.DOI
  13. [14]D. Rein and L. M. Sehgal, Annals Phys. 133, 79 (1981).
  14. [15]C. Berger and L. M. Sehgal, Phys. Rev. D 76, 113004 (2007), arXiv:0709.4378 [hep-ph].DOI
  15. [16]C. Wilkinson, P. Rodrigues, S. Cartwright, L. Thompson, and K. McFarland, Phys. Rev. D 90, 112017 (2014), arXiv:1411.4482 [hep-ex].DOI
  16. [17]T. Golan, C. Juszczak, and J. T. Sobczyk, Phys. Rev. C 86, 015505 (2012), arXiv:1202.4197 [nucl-th].DOI
  17. [18]Q. Yan, K. Niewczas, A. Nikolakopoulos, R. González-Jiménez, N. Jachowicz, X. Lu, J. Sobczyk, and Y. Zheng, JHEP 12, 141 (2024), arXiv:2405.05212 [hep-ph].DOI
  18. [19]R. González-Jiménez, N. Jachowicz, K. Niewczas, J. Nys, V. Pandey, T. Van Cuyck, and N. Van Dessel, Phys. Rev. D 95, 113007 (2017), arXiv:1612.05511 [nucl-th].
  19. [20]M. Guidal, J. M. Laget, and M. Vanderhaeghen, Nucl. Phys. A 627, 645 (1997).
  20. [21]H. Kamano, S. X. Nakamura, T. S. H. Lee, and T. Sato, Phys. Rev. C 88, 035209 (2013), arXiv:1305.4351 [nucl-th].DOI
  21. [22]S. X. Nakamura, H. Kamano, and T. Sato, Phys. Rev. D 92, 074024 (2015), arXiv:1506.03403 [hep-ph].DOI
  22. [23]H. Kamano, S. X. Nakamura, T. S. H. Lee, and T. Sato, Phys. Rev. C 94, 015201 (2016), arXiv:1605.00363 [nucl-th].DOI
  23. [24]M. Döring, J. Haidenbauer, M. Mai, and T. Sato, (2025), arXiv:2505.02745 [nucl-th].arxiv.org/abs/2505.02745
  24. [25]K. Yamauchi, M. Ishitsuka, and Y. Hayato, PoS TAUP2023, 340 (2024).
  25. [26]N. Rocco, S. X. Nakamura, T. S. H. Lee, and A. Lovato, Phys. Rev. C 100, 045503 (2019), arXiv:1907.01093 [nucl-th].DOI
  26. [27]A. Nikolakopoulos, A. Ershova, R. González-Jiménez, J. Isaacson, A. M. Kelly, K. Niewczas, N. Rocco, and F. Sánchez, Phys. Rev. C 110, 054611 (2024), arXiv:2406.09244 [nucl-th].DOI
  27. [28]L. L. Salcedo, E. Oset, M. J. Vicente-Vacas, and C. Garcia-Recio, Nucl. Phys. A 484, 557 (1988).
  28. [29]O. Buss, T. Gaitanos, K. Gallmeister, H. van Hees, M. Kaskulov, O. Lalakulich, A. B. Larionov, T. Leitner, J. Weil, and U. Mosel, Phys. Rept. 512, 1 (2012), arXiv:1106.1344 [hep-ph].DOI
  29. [30]J. Cugnon, J. Vandermeulen, and D. L’Hote, Nucl. Instrum. Meth. B 111, 215 (1996).
  30. [31]J. Isaacson, W. I. Jay, A. Lovato, P. A. N. Machado, and N. Rocco, Phys. Rev. D 107, 033007 (2023), arXiv:2205.06378 [hep-ph].DOI
  31. [32]J. Isaacson, W. I. Jay, A. Lovato, P. A. N. Machado, and N. Rocco, Phys. Rev. C 103, 015502 (2021), arXiv:2007.15570 [hep-ph].DOI
  32. [33]T. Sato and T.-S. H. Lee, Phys. Rev. C 54, 2660 (1996), arXiv:nucl-th/9606009.DOI
  33. [34]M. Kobayashi, T. Sato, and H. Ohtsubo, Progress of Theoretical Physics 98, 927 (1997).
  34. [35]H. Feshbach, Theoretical nuclear physics: Nuclear reactions (1992).
  35. [36]A. Matsuyama, T. Sato, and T. S. H. Lee, Phys. Rept. 439, 193 (2007), arXiv:nucl-th/0608051.DOI
  36. [37]CNS Data Analysis Center, “Cns data analysis center, george washington university,” http://gwdac.phys.gwu.edu (2025), accessed: 25 April 2025.
  37. [38]H. Kamano, S. X. Nakamura, T.-S. H. Lee, and T. Sato, Phys. Rev. C 88, 035209 (2013).
  38. [39]S. X. Nakamura, H. Kamano, and T. Sato, Phys. Rev. D 92, 074024 (2015).
  39. [40]N. Rocco, C. Barbieri, O. Benhar, A. De Pace, and A. Lovato, Phys. Rev. C 99, 025502 (2019), arXiv:1810.07647 [nucl-th].DOI
  40. [41]J. Carlson, S. Gandolfi, F. Pederiva, S. C. Pieper, R. Schiavilla, K. E. Schmidt, and R. B. Wiringa, Rev. Mod. Phys. 87, 1067 (2015), arXiv:1412.3487 [nucl-th].
  41. [42]A. Gnech, B. Fore, A. J. Tropiano, and A. Lovato, Phys. Rev. Lett. 133, 142501 (2024), arXiv:2308.16266 [nucl-th].
  42. [43]D. Ronchen, M. Doring, F. Huang, H. Haberzettl, J. Haidenbauer, C. Han hart, S. Krewald, U. G. Meissner, and K. Nakayama, Eur. Phys. J. A 49, 44 (2013), arXiv:1211.6998 [nucl-th].DOI
  43. [44]E. Oset and L. L. Salcedo, Nucl. Phys. A 468, 631 (1987).
  44. [45]M. J. Vicente Vacas and E. Oset, Nucl. Phys. A 568, 855 (1994).
  45. [46]L. Landau, E. Lifshitz, and L. Pitaevskii, Physical Kinetics, Course of Theoretical Physics, Vol. 10 (Pergamon Press, 1981).
  46. [47]V. Dmitriev, O. Sushkov, and C. Gaarde, Nucl. Phys. A 459, 503 (1986).
  47. [48]O. Buss, Photon- and Pion-induced Reactions in a Transport Approach, Other thesis (2008).
  48. [49]M. Effenberger, Eigenschaften von Hadronen in Kernmaterie in einem vereinheitlichten Transportmodell, Ph.D. thesis, Giessen U. (1999).DOI
  49. [50]B. Bogart, K. Gallmeister, and U. Mosel, Phys. Rev. C 110, 044001 (2024), arXiv:2405.05921 [hep-ex].DOI
  50. [51]T. Song and C. M. Ko, (2014), arXiv:1403.7363 [nucl-th].arxiv.org/abs/1403.7363
  51. [52]M. Murphy et al., Phys. Rev. C 100, 054606 (2019), arXiv:1908.01802 [hep-ex].DOI
  52. [53]O. Benhar, A. Fabrocini, S. Fantoni, and I. Sick, Nucl. Phys. A 579, 493 (1994).
  53. [54]L. Jiang et al. (Jefferson Lab Hall A), Phys. Rev. D 105, 112002 (2022), arXiv:2203.01748 [nucl-ex].DOI
  54. [55]L. Jiang et al. (Jefferson Lab Hall A), Phys. Rev. D 107, 012005 (2023), arXiv:2209.14108 [nucl-ex].DOI
  55. [56]A. Nikolakopoulos, R. González-Jiménez, N. Jachowicz, K. Niewczas, F. Sánchez, and J. M. Udías, Phys. Rev. C 105, 054603 (2022), arXiv:2202.01689 [nucl-th].DOI
  56. [57]O. Benhar, Nucl. Phys. B Proc. Suppl. 159, 168 (2006), arXiv:hep-ph/0602108.DOI
  57. [58]R. González-Jiménez, A. Nikolakopoulos, N. Jachowicz, and J. M. Udías, Phys. Rev. C 100, 045501 (2019), arXiv:1904.10696 [nucl-th].
  58. [59]A. Meucci, C. Giusti, and F. D. Pacati, Nucl. Phys. A 739, 277 (2004), arXiv:nucl-th/0311081.DOI
  59. [60]Y. Horikawa, F. Lenz, and N. C. Mukhopadhyay, Phys. Rev. C 22, 1680 (1980).
  60. [61]A. Meucci, F. Capuzzi, C. Giusti, and F. D. Pacati, Phys. Rev. C 67, 054601 (2003), arXiv:nucl-th/0301084.DOI
  61. [62]R. González-Jiménez, M. Barbaro, J. Caballero, T. Donnelly, N. Jachowicz, G. Megias, K. Niewczas, A. Nikolakopoulos, and J. M. Udías, Phys. Rev. C 101, 015503 (2020), arXiv:1909.07497 [nucl-th].
  62. [63]E. S. Pinzon Guerra, S. Bhadra, S. Berkman, C. Cao, P. de Perio, Y. Hayato, K. Ieki, M. Ikeda, Y. Kanazawa, J. Kim, P. Kitching, K. Mahn, T. Nakaya, M. Nicholson, K. Olchanski, S. Rettie, H. A. Tanaka, S. Tobayama, M. J. Wikling, T. Yamauchi, S. Yen, and M. Yokoyama (DUET Collaboration), Phys. Rev. C 95, 045203 (2017).
  63. [64]D. Ashery, I. Navon, G. Azuelos, H. K. Walter, H. J. Pfeiffer, and F. W. Schlepütz, Phys. Rev. C 23, 2173 (1981).
  64. [65]B. Kotliński et al. (LADS), Eur. Phys. J. A 9, 537 (2000).
  65. [66]S. Dytman, Y. Hayato, R. Raboanary, J. T. Sobczyk, J. Tena Vidal, and N. Volonina, Phys. Rev. D 104, 053006 (2021), arXiv:2103.07535 [hep-ph].DOI
  66. [67]J. Vogelsang, B. L. G. Bakker, and H. J. Boersma, Nucl. Phys. A 452, 644 (1986).
  67. [68]A. M. Ankowski et al., J. Phys. G 50, 120501 (2023), arXiv:2203.06853 [hep-ex].DOI
  68. [69]A. Papadopoulou, A. Ashkenazi, S. Gardiner, M. Betancourt, S. Dytman, L. B. Weinstein, E. Piasetzky, F. Hauenstein, M. Khachatryan, S. Dolan, G. D. Megias, and O. Hen (e4v Collaboration), Phys. Rev. D 103, 113003 (2021).
  69. [70]M. Khachatryan et al. (CLAS, e4v), Nature 599, 565 (2021).
  70. [71]A. Nikolakopoulos, M. Martini, M. Ericson, N. Van Dessel, R. González-Jiménez, and N. Jachowicz, Phys. Rev. C 98, 054603 (2018), arXiv:1808.07520 [nucl-th].DOI
  71. [72]K. Abe et al. (T2K), Phys. Rev. D 98, 032003 (2018), arXiv:1802.05078 [hep-ex].DOI
  72. [73]K. Abe et al. (T2K), Phys. Rev. D 103, 112009 (2021), arXiv:2102.03346 [hep-ex].DOI
  73. [74]K. Abe et al. (T2K), Phys. Rev. D 91, 072010 (2015), arXiv:1502.01550 [hep-ex].
  74. [75]A. S. Meyer, M. Betancourt, R. Gran, and R. J. Hill, Phys. Rev. D 93, 113015 (2016), arXiv:1603.03048 [hep-ph].DOI
  75. [76]D. Simons, N. Steinberg, A. Lovato, Y. Meurice, N. Rocco, and M. Wagman, J. Phys. G 52, 065003 (2025), arXiv:2210.02455 [hep-ph].arxiv.org/abs/2210.02455
  76. [77]S. Gardiner, J. Isaacson, and L. Pickering, (2023), arXiv:2310.13211 [hep-ph].DOI
  77. [78]P. Stowell et al., JINST 12, P01016 (2017), arXiv:1612.07393 [hep-ex].DOI
  78. [79]X. G. Lu et al. (MINERvA), Phys. Rev. Lett. 121, 022504 (2018), arXiv:1805.05486 [hep-ex].DOI
  79. [80]O. Altinok et al. (MINERvA), Phys. Rev. D 96, 072003 (2017), arXiv:1708.03723 [hep-ex].DOI
  80. [81]W. Filali, L. Munteanu, and S. Dolan, Phys. Rev. D 111, 032009 (2025), arXiv:2407.10962 [hep-ex].DOI
  81. [82]P. Abratenko et al. (MicroBooNE), Phys. Rev. Lett. 134, 161802 (2025), arXiv:2404.10948 [hep-ex].DOI
  82. [83]P. Abratenko et al. (MicroBooNE), Phys. Rev. Lett. 131, 101802 (2023), arXiv:2301.03706 [hep-ex].
  83. [84]W. Tang, X. Li, X. Qian, H. Wei, and C. Zhang, JINST 12, P10002 (2017), arXiv:1705.03568 [physics.data-an].
  84. [85]P. Abratenko et al. (MicroBooNE), (2024), arXiv:2404.10948 [hep-ex].DOI
  85. [86]G. S. Bali, L. Barca, S. Collins, M. Gruber, M. Löffler, A. Schäfer, W. Sö l dner, P. Wein, S. Weißhaupt, and T. Wurm (RQCD), JHEP 05, 126 (2020), arXiv:1911.13150 [hep-lat].DOI
  86. [87]S. Park, R. Gupta, B. Yoon, S. Mondal, T. Bhattacharya, Y.-C. Jang, B. Joó, and F. Winter (Nucleon Matrix Elements (NME)), Phys. Rev. D 105, 054505 (2022), arXiv:2103.05599 [hep-lat].DOI
  87. [88]D. Djukanovic, G. von Hippel, J. Koponen, H. B. Meyer, K. Ottnad, T. Schulz, and H. Wittig, Phys. Rev. D 106, 074503 (2022), arXiv:2207.03440 [hep-lat].arxiv.org/abs/2207.03440
  88. [89]T. Cai et al. (MINERvA), Nature 614, 48 (2023).
  89. [90]O. Tomalak, R. Gupta, and T. Bhattacharya, Phys. Rev. D 108, 074514 (2023), arXiv:2307.14920 [hep-lat].arxiv.org/abs/2307.14920
  90. [91]T. Franco-Munoz, R. González-Jiménez, and J. M. Udías, J. Phys. G 52, 025103 (2025), arXiv:2203.09996 [nucl-th].
  91. [92]A. Lovato, N. Rocco, and N. Steinberg, (2023), arXiv:2312.12545 [nucl-th].arxiv.org/abs/2312.12545
  92. [93]T. Franco-Munoz, J. García-Marcos, R. González-Jiménez, and J. M. Udías, Phys. Rev. C 108, 064608 (2023), arXiv:2306.10823 [nucl-th].
  93. [94]J. Ellis, Comput. Phys. Commun. 210, 103 (2017), arXiv:1601.05437 [hep-ph].DOI
  94. [95]J. D. Hunter, Computing in Science & Engineering 9, 90 (2007).
  95. [96]M. L. Waskom, Journal of Open Source Software 6, 3021 (2021).DOI
  96. [97]J. Isaacson, W. Jay, A. Lovato, P. Machado, A. Nikolakopoulos, N. Rocco, and N. Steinberg, “Achillesgen/achilles: Link repository to zenodo for persistence,” (2026).
  97. [98]J. Isaacson, W. Jay, A. Lovato, P. Machado, A. Nikolakopoulos, N. Rocco, and N. Steinberg, “Achillesgen/arxiv-2508.19213: Release of data reproduction for 2508.19213,” (2026).DOI
  98. [99]S. Höche, in Theoretical Advanced Study Institute in Elementary Particle Physics: Journeys Through the Precision Frontier: Amplitudes for Colliders (2015) pp. 235–295, arXiv:1411.4085 [hep-ph].
  99. [100]E. Byckling and K. Kajantie, Nucl. Phys. B 9, 568 (1969).
  100. [101]F. James, (1968).
  101. [102]R. Kleiss and R. Pittau, Comput. Phys. Commun. 83, 141 (1994), arXiv:hep-ph/9405257.DOI
  102. [103]T. Gleisberg and S. Hoeche, JHEP 12, 039 (2008), arXiv:0808.3674 [hep-ph].DOI
  103. [104]X. G. Lu, L. Pickering, S. Dolan, G. Barr, D. Coplowe, Y. Uchida, D. Wark, M. O. Wascko, A. Weber, and T. Yuan, Phys. Rev. C 94, 015503 (2016), arXiv:1512.05748 [nucl-th].DOI
  104. [105]X. G. Lu, D. Coplowe, R. Shah, G. Barr, D. Wark, and A. Weber, Phys. Rev. D 92, 051302 (2015), arXiv:1507.00967 [hep-ex].DOI
  105. [106]X. G. Lu, JPS Conf. Proc. 12, 010034 (2016), arXiv:1512.09042 [hep-ex].DOI

Paper details

vlnc:W_ycfk3a8g
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