betaFeedback

WorksPhysics and Astronomy

Δ(1232)Δ(1232) isobars and weak-interaction bulk viscosity in neutron-star merger matter

Abstract

Binary neutron star mergers probe nuclear matter at densities of a few times nuclear saturation density and temperatures of tens of MeV, conditions under which heavier baryons such as the Δ(1232)Δ(1232) isobars may appear. Previous calculations of the bulk viscosity of merger matter have considered nucleonic matter, with and without muons, and hyperonic matter, but not ΔΔ isobars. In this work, we extend these calculations to neutrino-transparent matter in which ΔΔ isobars are present in strong equilibrium. We derive the relativistic matrix elements for the charged-current Δ→NΔ\to N and Δ→Δ′Δ\toΔ' transitions, evaluate all electronic and muonic Urca processes over the full phase space using the nucleon width approximation, and include the in-medium Δ→NπΔ\to Nπ width with Pauli blocking. The bulk viscosity is obtained from the coupled linear response of the electron and muon fractions about the finite-temperature flavor equilibrium. We find that the Δ−Δ^- takes over the negative charge from the leptons, which raises the proton fraction and opens the nucleonic direct Urca process. Up to about 3~MeV this change in composition, rather than the ΔΔ reactions themselves, controls flavor equilibration. As a consequence, the resonant maximum of the bulk viscosity moves to lower temperatures for all oscillation frequencies we consider, and kHz density oscillations near resonance are damped several times faster than in nucleonic matter. We also show that, for nucleonic matter, the choice of the equilibrium state about which the rates are linearized shifts the resonance temperature by up to a few MeV at high density.

Introduction

Neutron stars contain the densest matter in the universe, and binary neutron star mergers in particular probe nuclear matter at densities of a few times nuclear saturation density n0n_0 and temperatures of tens of MeV. In the violent dynamics of a merger we expect high-amplitude density oscillations in the kHz frequency range that drive the matter out of flavor (β\beta) equilibrium. The weak-interaction (“Urca”) processes that restore equilibrium can then produce strong bulk viscous damping of these oscillations [1]. Most et al. [2] post-processed a merger simulation that did not itself include bulk viscosity and found that the bulk viscosity reaches large values in parts of the remnant; their estimates indicated that its effect on the global post-merger dynamics could become significant, whereas during the inspiral it is probably too small to be observable. Simulations that evolve the out-of-equilibrium composition have since shown how a bulk-viscous pressure emerges in the post-merger dynamics [3]. Whether such effects are observable is still open: Zhu et al. [4] estimated that the resulting shift of the post-merger gravitational-wave peak frequency could be detected by third-generation detectors, in optimistic scenarios, only for equations of state with a large slope of the symmetry energy. For a review of bulk viscosity in dense matter, see Ref. [5].

These estimates rely on the microscopic Urca rates, which have been calculated with increasing care for nucleonic matter. At the temperatures above about 1 MeV attained in mergers, the Fermi surface approximation is no longer reliable, and the flavor-equilibrium condition acquires a correction to the cold condition, in which the neutron chemical potential equals the sum of the proton and electron chemical potentials [6, 7]. The bulk viscosity must then be calculated by linearizing about the true, finite-temperature equilibrium [8]. The nucleon width approximation (NWA) [9] includes the collisional-broadening (“modified Urca”) contribution on the same footing as the direct Urca contribution, and it has been used to study the effect of magnetic fields on the Urca rates [10] and on the bulk viscosity [11]. When muons are present there are separate electronic and muonic Urca processes and hence two relaxation rates [12]; Hernández, Manuel and Tolos [13] found that this can produce a double-peaked frequency dependence of the bulk viscosity, because the electronic and muonic direct Urca processes open at different densities. Near saturation density, the equation-of-state dependence of the bulk viscosity is determined by the symmetry energy and its slope [14]. A common theme of these studies is that the bulk viscosity depends sensitively on the composition of the matter, and in particular on whether the direct Urca process is kinematically allowed.

The composition can change qualitatively if heavier baryons appear. In relativistic mean-field models in which the Δ(1232)\Delta(1232) couples somewhat more strongly to the scalar meson than the nucleon does, the Δ−\Delta^- appears at two to three times n0n_0, while the maximum mass of neutron stars still exceeds two solar masses [15, 16, 17, 18, 19]. Raduta [19] showed that in cold DDME2 matter the Δ−\Delta^- replaces electrons as the negative charge, which raises the proton fraction and opens the nucleonic direct Urca process. The Δ\Deltas can also undergo direct Urca processes of their own; their neutrino emissivity has been computed recently [20] and included in neutron star cooling simulations [21]. These studies concern neutrino emission from cold matter, and the response of hot Δ\Delta-admixed matter to compression has, to our knowledge, not been studied. For hyperons, the other commonly considered heavy

viscosity must then be calculated by linearizing about the true, finite-temperature equilibrium [8]. The nucleon width approximation (NWA) [9] includes the collisional-broadening (“modified Urca”) contribution on the same footing as the direct Urca contribution, and it has been used to study the effect of magnetic fields on the Urca rates [10] and on the bulk viscosity [11]. When muons are present there are separate electronic and muonic Urca processes and hence two relaxation rates [12]; Hernández, Manuel and Tolos [13] found that this can produce a double-peaked frequency dependence of the bulk viscosity, because the electronic and muonic direct Urca processes open at different densities. Near saturation density, the equation-of-state dependence of the bulk viscosity is determined by the symmetry energy and its slope [14]. A common theme of these studies is that the bulk viscosity depends sensitively on the composition of the matter, and in particular on whether the direct Urca process is kinematically allowed.

The composition can change qualitatively if heavier baryons appear. In relativistic mean-field models in which the Δ(1232)\Delta(1232) couples somewhat more strongly to the scalar meson than the nucleon does, the Δ−\Delta^- appears at two to three times n0n_0, while the maximum mass of neutron stars still exceeds two solar masses [15, 16, 17, 18, 19]. Raduta [19] showed that in cold DDME2 matter the Δ−\Delta^- replaces electrons as the negative charge, which raises the proton fraction and opens the nucleonic direct Urca process. The Δ\Deltas can also undergo direct Urca processes of their own; their neutrino emissivity has been computed recently [20] and included in neutron star cooling simulations [21]. These studies concern neutrino emission from cold matter, and the response of hot Δ\Delta-admixed matter to compression has, to our knowledge, not been studied. For hyperons, the other commonly considered heavy baryons, the strangeness-changing rates and the resulting bulk viscosity have been calculated under merger conditions [22], and more recent work has addressed their role in the inspiral [23, 24]. The Δ\Deltas differ from the hyperons in one important respect: strong processes such as NN↔NΔNN \leftrightarrow N\Delta keep the Δ\Delta population in equilibrium at all times, so the Δ\Deltas do not introduce a new slowly relaxing quantity. Instead, they modify both the thermodynamic response of the matter and the set of weak processes through which the lepton fractions relax. The main thrust of this paper is to explore these two effects.

To do this, we extend the full phase-space NWA calculation of Refs. [9, 11], whose nucleonic results we first reproduce, to matter containing the Δ\Delta quartet. We derive the relativistic spin-summed matrix elements for the charged-current Δ→N\Delta\to N and Δ→Δ′\Delta\to\Delta^\prime transitions from the Rarita–Schwinger spin sum and include all electronic and muonic channels. We calculate the in-medium Δ→Nπ\Delta\to N\pi width with Pauli blocking to check whether the populated Δ\Deltas are stable against strong decay. We then obtain the bulk viscosity from the coupled linear response of the electron and muon fractions about the finite-temperature flavor equilibrium, and evaluate it as a function of density, temperature and oscillation frequency. In our calculations we assume that the matter is transparent to neutrinos, which is a reasonable approximation below about 10 MeV [8, 11].

This paper is structured as follows. In Sec. II we describe the models of Δ\Delta-admixed nuclear matter. In Sec. III we derive the matrix elements of the Δ\Delta transitions and the in-medium Δ\Delta width, and in Sec. IV we present the two-variable linear response. § 1.4 summarizes our numerical checks, Sec. VI presents our results, and in Sec. VII we discuss their implications. Standard ingredients that we use without modification, namely the mean-field equations, the phase-space reduction, the NWA, the nucleonic matrix element, and the one-variable and damping-time relations, are collected in Appendices A–§ 1.5.4, and all parameter values are listed in Tables I–III.

Matter with Δ\Delta isobars

We use natural units, in which the reduced Planck constant, the speed of light and the Boltzmann constant are set to one, and the metric signature (+,−,−,−)(+, -, -, -) throughout this paper, and we convert to CGS units only for the bulk viscosity and the damping time. Baryon densities are given in units of the reference density n0n_0 of Table III.

Models

We describe nuclear matter by relativistic mean-field theories containing the nucleon doublet and the Δ(1232)\Delta(1232) quartet, with the Δ\Deltas described by Rarita–Schwinger vector-spinors [25]. The baryons interact through σ\sigma, ω\omega and ρ\rho exchange, which we treat in the mean-field approximation. Each Δ\Delta is then a spin-32\frac{3}{2} quasiparticle with four spin states, an effective mass and a vector self-energy, and its meson couplings are specified as fractions of the nucleon couplings. The Lagrangian, the field equations and the finite-temperature thermodynamics are given in Appendix A.

One of the most important features influencing flavor equilibration is the direct Urca threshold. To isolate the effect of the Δ\Deltas, we use two nucleonic models in which the nucleonic direct Urca process is kinematically forbidden in cold β\beta-equilibrated matter up to eight times n0n_0. QMC-RMF3 [26, 27] has constant couplings and an ωρ\omega\rho mixing term; it is the model without a direct Urca threshold used in Refs. [8, 11]. DDME2 [28] has density-dependent couplings, and the absence of a direct Urca threshold agrees with Refs. [12, 19]. Any opening of the direct Urca process in our calculations can therefore be attributed to the Δ\Deltas. The parameters and saturation properties of both models are listed in Table I. The saturation properties obtained with our solver agree with Ref. [26] for QMC-RMF3 and, for DDME2, with the symmetry energy and slope quoted in Ref. [19].

Strong and weak equilibrium

On strong-interaction timescales the baryon composition equilibrates through processes such as NN↔NΔNN \leftrightarrow N\Delta, which are many orders of magnitude faster than the weak processes we consider. At fixed baryon density and charge fraction the baryon chemical potentials therefore obey

μb=μB+qbμQ,μB≡μn,μQ≡μp−μn,(1)\mu_b = \mu_B + q_b\mu_Q,\qquad\mu_B \equiv\mu_n,\qquad\mu_Q \equiv\mu_p-\mu_n, \tag*{(1)}

that is,

μΔ−=2μn−μp,μΔ0=μn,μΔ+=μp,μΔ++=2μp−μn.(2)\begin{aligned} \mu_{\Delta^-} &= 2\mu_n-\mu_p, & \mu_{\Delta^0} &= \mu_n,\\ \mu_{\Delta^+} &= \mu_p, & \mu_{\Delta^{++}} &= 2\mu_p-\mu_n. \tag*{(2)} \end{aligned}

For given baryon density, charge fraction and temperature we solved the conditions of baryon number and charge,

∑bnb=nB,∑bqbnb=YQnB=ne+nμ,(3)\sum_b n_b = n_B,\qquad\sum_b q_b n_b = Y_Q n_B = n_e+n_\mu, \tag*{(3)}

together with the field equations, for the unknowns μB\mu_B, μQ\mu_Q, Φ\Phi, WW, BB (and ΣR\Sigma_R for DDME2) by a multidimensional Newton method, to a relative residual below 10−810^{-8}.

On longer timescales the weak interactions change the lepton numbers. In neutrino-transparent matter the electron and muon fractions YeY_e and YμY_\mu (lepton densities per baryon) change only through Urca processes, and, as in Ref. [12], we assume that purely leptonic conversion between muons and electrons is frozen. A state out of weak equilibrium is then specified by the baryon density, the temperature and the two lepton fractions, and the baryon sector depends only on their sum YQY_Q. The departures from cold β\beta equilibrium are

QMC-RMF3DDME2
MNM_N939.0939.0
mσm_\sigma, mωm_\omega, mρm_\rho491.5, 782.5, 763.0550.1238, 783.0, 763.0
gσg_\sigma, gωg_\omega, gρg_\rho8.32, 9.76, 11.0210.5396, 13.0189, 3.6836 (at nsatn_{\mathrm{sat}})
θb\theta_b in Eq. (46)t3bt_{3b}2t3b2t_{3b}
bb, cc, b1b_1 (Ref. [26])0.0063, −0.006-0.006, 5.87–
κ\kappa (MeV), λ\lambda, Λv\Lambda_v11.83, −0.036-0.036, 0.0616–
aσa_\sigma, bσb_\sigma, cσc_\sigma, dσd_\sigma–1.3881, 1.0943, 1.7057, 0.4421
aωa_\omega, bωb_\omega, cωc_\omega, dωd_\omega–1.3892, 0.9240, 1.4620, 0.4775
aρa_\rho–0.5647
nsatn_{\mathrm{sat}} in Eq. (43)–0.152
nsatn_{\mathrm{sat}}0.15690.1519
E/AE/A−16.06-16.06−16.14-16.14
KK229.8250.9
JJ, LL33.55, 49.032.30, 51.3
M∗/MNM^*/M_N0.7320.572

Table 1. Nucleonic parameters and saturation properties. QMC-RMF3: constant couplings of Ref. [26], Table I; the parameters bb, cc, b1b_1 of that reference translate into κ=2bMN\kappa= 2bM_N, λ=6c\lambda= 6c and Λv=b1/gω2\Lambda_v = b_1/g_\omega^2 of Eq. (42). DDME2: couplings at saturation and the density-dependence parameters of Eq. (43) [28]. The isospin factor θb\theta_b of the ρ\rho coupling in Eq. (46) differs between the two models. Masses in MeV. Lower block: saturation density (fm−3\mathrm{fm}^{-3}), binding energy per baryon, incompressibility, symmetry energy and its slope (MeV), and Dirac effective mass, computed with our solver; for QMC-RMF3 the symmetry energy and slope are from the difference between pure neutron and symmetric matter, for DDME2 from the quadratic term.

δμe=μn−μp−μe,δμμ=μn−μp−μμ.(4)\delta\mu_e = \mu_n - \mu_p - \mu_e,\qquad\delta\mu_\mu= \mu_n - \mu_p - \mu_\mu. \tag*{(4)}

Because of Eq. (1), every charged-current process that converts a baryon into one with one unit more charge involves the same chemical-potential difference between initial and final baryon, μn−μp\mu_n-\mu_p. This means that a single departure δμl\delta\mu_l drives all the processes involving a given lepton.

coupling sets

For each model we use two Δ\Delta coupling sets within the ranges discussed in Refs. [15, 17, 18, 19], which differ only in the scalar coupling ratio. The six resulting equations of state are listed in Table II, together with the Δ\Delta onset densities and stellar properties, and we refer to them by the names given there. All four Δ\Delta sets support maximum masses of at least two solar masses, and our nucleonic maximum masses agree with the published values [19, 26]; the radii are less reliable because we use a polytropic crust. A scalar ratio of 1.3 produces a pressure that decreases with density in QMC-RMF3, so we exclude it. All four remaining sets are stable against spinodal decomposition between one and six times $n_0 ({xref}Appendix A). This is consistent with {citet}Ref. [19]`, which found Δ\Delta-driven instabilities in DDME2 only for a larger excess of the scalar over the vector coupling ratio than in our sets.

Weak rates with isobars

Channels and rates

For each charged lepton, electron or muon, we included the decay and capture branches of

n→plνˉl,Δ−→nlνˉl,Δ0→plνˉl,Δ−→Δ0lνˉl,Δ0→Δ+lνˉl,Δ+→Δ++lνˉl,(5)n \to p l\bar{\nu}_l,\qquad\Delta^- \to n l\bar{\nu}_l,\qquad\Delta^0 \to p l\bar{\nu}_l, \qquad\Delta^- \to\Delta^0 l\bar{\nu}_l,\qquad\Delta^0 \to\Delta^+ l\bar{\nu}_l,\qquad\Delta^+ \to\Delta^{++} l\bar{\nu}_l, \tag*{(5)}

together with the inverse lepton-capture reactions. The net lepton production rate per unit volume is

Γl(nB,T,Ye,Yμ)=∑ch[Γch,ldec−Γch,lcap],(6)\Gamma_l(n_B,T,Y_e,Y_\mu)=\sum_{\mathrm{ch}}\left[\Gamma^{\mathrm{dec}}_{\mathrm{ch},l}-\Gamma^{\mathrm{cap}}_{\mathrm{ch},l}\right], \tag*{(6)}

and we defined weak equilibrium at temperature TT by the vanishing of both net rates,

Γe=Γμ=0,(7)\Gamma_e=\Gamma_\mu=0, \tag*{(7)}

which differs from the vanishing of the chemical imbalances by the thermal shift discussed in Ref. [7].

We calculate each rate as the full phase-space integral of Ref. [7], reduced exactly to five dimensions, and average it over Breit–Wigner spectral functions of both baryons following the NWA of Ref. [9]. We have generalized both steps to unequal baryon masses, arbitrary vector potentials and spin-32\frac{3}{2} baryons, so that the Δ\Deltas enter only through their dispersion relations, their spectral widths and the matrix elements. The reduction, the NWA and the nucleonic matrix element are given in Appendix B. For the collisional width of the Δ\Delta we take the nucleon value and add the in-medium Δ→Nπ\Delta\to N\pi width of Sec. III C; Appendix B gives the reasons for this choice, and Figure 6 its effect. What remains is to specify the matrix elements of the Δ\Delta transitions, which we derive below.

Figure 6. Change of the logarithm of the bulk viscosity at 1 kHz relative to the baseline (QMC-Δ\Delta2 and DD-Δ\Delta2 with the inputs of Table III) when one input is varied, as a function of density at 1, 3 and 5 MeV (columns), for QMC-RMF3 (top) and DDME2 (bottom); isothermal. The grey band marks changes below 0.1 dex; the hatched region in the DDME2 panels marks densities near the zero of the bulk viscosity, where the ratio is not meaningful. The collisional width of the Δ\Delta dominates near threshold at 1 MeV, the weak couplings matter at the 0.1–0.25 dex level, and the Δ\Delta channels themselves matter most at 5 MeV and high density.

Namexσx_{\sigma}xωx_{\omega}xρx_{\rho}nΔ−n_{\Delta^-}nΔ0n_{\Delta^0}nΔ+n_{\Delta^+}ndUn_{\mathrm{dU}}Mmax⁡M_{\max} [M⊙M_{\odot}]RR, ncn_c at Mmax⁡M_{\max} [km, n0n_0]RR, ncn_c at 1.4 M⊙M_{\odot}
QMC-N–––––––2.14710.64, 6.6712.00, 2.92
QMC-Δ\Delta11.11.11.02.304.70–3.252.0289.99, 7.5611.67, 3.53
QMC-Δ\Delta21.21.11.01.953.454.302.702.0129.66, 7.9211.04, 4.00
DD-N–––––––2.48912.05, 5.0513.01, 2.11
DD-Δ\Delta11.11.11.01.804.40–3.15–4.752.39211.47, 5.6412.86, 2.37
DD-Δ\Delta21.21.11.01.452.753.952.102.37911.22, 5.7712.25, 2.63

Table 2. The six equations of state. Left block: Δ\Delta coupling ratios of Eq. (41). Middle block: onset densities of the Δ\Delta species and the cold nucleonic direct Urca threshold in β\beta-equilibrated matter with muons at 1 MeV. Right block: maximum mass, its radius and central density, and the radius and central density of a 1.4 M⊙M_{\odot} star (cold, polytropic crust). Densities in units of n0n_0; a dash means not reached below six times n0n_0 (onsets) or eight times n0n_0 (direct Urca). For DD-Δ\Delta1 the direct Urca triangle is open only between the two densities given.

transition matrix elements

The effective charged-current interaction, the lepton current and the spin-summed lepton tensor are given in Appendix B; the hadronic tensors derived below are contracted with this lepton tensor.

Δ→N\Delta\to N. The leading term of the NΔN\Delta axial transition current at low momentum transfer and the resulting amplitude are [29]

⟨N∣Aμ∣Δ⟩=C5AuˉNuΔμ,M=G2IC5AuˉNuΔμlμ;(8)\langle N|A^{\mu}|\Delta\rangle= C_{5}^{A}\bar{u}_{N}u_{\Delta}^{\mu}, \qquad\mathcal{M} = \frac{G}{\sqrt{2}} \mathcal{I} C_{5}^{A}\bar{u}_{N}u_{\Delta}^{\mu}l_{\mu}; \tag*{(8)}

the vector form factors and the other axial form factors are suppressed by powers of the momentum transfer over the Δ\Delta mass, and we neglected them. In our conditions the momentum transfer is of the order of the electron Fermi momentum, about 100 MeV or less, so this ratio is below about 0.1. The isospin factors ∣I∣2|\mathcal{I}|^2 and the value of C5AC_{5}^{A} and its variation are given in Table III; the isospin factors are normalized to one for the nucleon. The RS spin sum is

\sum_s u_{\Delta}^{\mu}\bar{u}_{\Delta}^{\nu} = -(\slashed{p}_{\Delta}+M)\left[g^{\mu\nu}-\frac{1}{3}\gamma^{\mu}\gamma^{\nu}-\frac{2p_{\Delta}^{\mu}p_{\Delta}^{\nu}}{3M^2}+\frac{p_{\Delta}^{\mu}\gamma^{\nu}-p_{\Delta}^{\nu}\gamma^{\mu}}{3M}\right], \tag*{(9)}

with MM the Δ\Delta effective mass, and the hadronic tensor, with mm the nucleon effective mass, is

H^{\mu\nu} = \operatorname{Tr}\left[(\slashed{p}_{N}+m)\sum_s u_{\Delta}^{\mu}\bar{u}_{\Delta}^{\nu}\right]. \tag*{(10)}

It follows term by term from the trace identities

\operatorname{Tr}[\slashed{a}\slashed{b}] = 4\,a\cdot b,\qquad\operatorname{Tr}[\gamma^{\mu}\gamma^{\nu}] = 4g^{\mu\nu}
\operatorname{Tr}[\slashed{a}\slashed{b}\gamma^{\mu}\gamma^{\nu}] = 4[(a\cdot b)g^{\mu\nu}-a^{\mu}b^{\nu}+a^{\nu}b^{\mu}], \tag*{(11)}

the on-shell condition for the Δ\Delta and the vanishing of traces of odd numbers of γ\gamma matrices. The four terms of Eq. (9) give

\begin{aligned} H_{(1)}^{\mu\nu} &\equiv-\operatorname{Tr}[(\slashed{p}_{N}+m)(\slashed{p}_{\Delta}+M)]g^{\mu\nu} \\ &= -4(p_N\cdot p_{\Delta}+mM)g^{\mu\nu},\\ H_{(2)}^{\mu\nu} &\equiv\frac{1}{3}\operatorname{Tr}[(\slashed{p}_{N}+m)(\slashed{p}_{\Delta}+M)\gamma^{\mu}\gamma^{\nu}] \\ &= \frac{4}{3}(p_N\cdot p_{\Delta}+mM)g^{\mu\nu}+\frac{4}{3}(p_{\Delta}^{\mu}p_N^{\nu}-p_{\Delta}^{\nu}p_N^{\mu}),\\ H_{(3)}^{\mu\nu} &\equiv\frac{2p_{\Delta}^{\mu}p_{\Delta}^{\nu}}{3M^2}\operatorname{Tr}[(\slashed{p}_{N}+m)(\slashed{p}_{\Delta}+M)] \\ &= \frac{8}{3M^2}(p_N\cdot p_{\Delta}+mM)p_{\Delta}^{\mu}p_{\Delta}^{\nu},\\ H_{(4)}^{\mu\nu} &\equiv-\frac{1}{3M}\operatorname{Tr}[(\slashed{p}_{N}+m)(\slashed{p}_{\Delta}+M)(p_{\Delta}^{\mu}\gamma^{\nu}-p_{\Delta}^{\nu}\gamma^{\mu})] \\ &= -\frac{4}{3}(p_{\Delta}^{\mu}p_N^{\nu}-p_{\Delta}^{\nu}p_N^{\mu}). \tag*{(12)} \end{aligned}

In the fourth term the contributions proportional to the nucleon mass and to the product of two Δ\Delta momenta cancel between the two orderings. The antisymmetric pieces of the second and fourth terms cancel, and the sum is

Hμν=−83(pN⋅pΔ+mM)[gμν−pΔμpΔνM2].(13)H^{\mu\nu}=-\frac{8}{3}(p_N\cdot p_{\Delta}+mM)\left[g^{\mu\nu}-\frac{p_{\Delta}^{\mu}p_{\Delta}^{\nu}}{M^2}\right]. \tag*{(13)}

The hadronic tensor is thus symmetric and transverse to the Δ\Delta momentum, as required by the RS constraints, and the antisymmetric part of the lepton tensor does not contribute. The two contractions needed are

gμνLμν=−16 kl⋅kν,g^{\mu\nu}L_{\mu\nu}=-16\,k_l\cdot k_{\nu},
pμpνLμν=8[2(p⋅kl)(p⋅kν)−M2kl⋅kν],(14)p^{\mu}p^{\nu}L_{\mu\nu}=8[2(p\cdot k_l)(p\cdot k_{\nu})-M^2 k_l\cdot k_{\nu}], \tag*{(14)}

and they give

∑‾∣M∣2=323G2∣I∣2(C5A)2(pN⋅pΔ+mN∗mΔ∗)[kl⋅kν+2(pΔ⋅kl)(pΔ⋅kν)mΔ∗2].(15)\overline{\sum}|\mathcal{M}|^2=\frac{32}{3}G^2|\mathcal{I}|^2(C_5^A)^2(p_N\cdot p_{\Delta}+m_N^*m_{\Delta}^*)\left[k_l\cdot k_{\nu}+\frac{2(p_{\Delta}\cdot k_l)(p_{\Delta}\cdot k_{\nu})}{m_{\Delta}^{*2}}\right]. \tag*{(15)}

For baryons at rest the angular average of the bracket is three times the product of the lepton energies, so

∑∣M∣2→64G2∣I∣2(CAA)2mN∗mΔ∗ElEν.(16)\sum|\mathcal{M}|^2 \to64G^2|\mathcal{I}|^2(C_A^A)^2m_N^*m_\Delta^*E_lE_\nu. \tag*{(16)}

Per initial spin state this is the nucleon value of Eq. (71) multiplied by the isospin factor and the squared transition coupling, divided by the nucleon Gamow–Teller factor 1+3gA21+3g_A^2. We verified Eq. (13) with the symbolic program FORM, and Eq. (15) against an explicit sum over RS spinors built from Clebsch–Gordan products of Dirac spinors and polarization vectors, at random off-axis relativistic kinematics; the ratio was one to ten digits.

Δ→Δ′\Delta\to\Delta'. For transitions inside the quartet we used the RS current and its isospin factor

Jμ=−uˉΔ′αγμ(1−gAΔγ5)uΔα,J^\mu=-\bar{u}_{\Delta'\alpha}\gamma^\mu(1-g_{A\Delta}\gamma^5)u_\Delta^\alpha,
∣⟨Δ′∣T+∣Δ⟩∣2=I(I+1)−t3(t3+1),(17)|\langle\Delta'|T_+|\Delta\rangle|^2=I(I+1)-t_3(t_3+1), \tag*{(17)}

which gives the factors 3, 4 and 3 of Table III for the three transitions. The axial coupling was fixed by the quark model. In the state of maximal spin projection all quark spins are aligned, so the quark-model Gamow–Teller operator reduces to the isospin raising operator, while the RS axial current measures only the Dirac-spin part of the spin-32\frac{3}{2} state, which by the projection theorem is two thirds of the total spin projection:

⟨Δ′∣∑qσzqt+q∣Δ⟩=⟨T+⟩,⟨2sz⟩=23Sz.(18)\left\langle\Delta'\left|\sum_q\sigma_z^qt_+^q\right|\Delta\right\rangle=\langle T_+\rangle,\qquad\langle2s_z\rangle=\frac{2}{3}S_z . \tag*{(18)}

Normalizing to the quark-model nucleon value of 5/35/3 then gives

gAΔ=35gA,(19)g_{A\Delta}=\frac{3}{5}g_A, \tag*{(19)}

with the numerical value of Table III. Since ⟨Δ∣T3∣Δ⟩=32\langle\Delta|T_3|\Delta\rangle=\frac{3}{2} for the Δ++\Delta^{++}, this coupling corresponds to the Δ++\Delta^{++} axial charge g1(0)=3gAΔ=95gA≈2.27g_1^{(0)}=3g_{A\Delta}=\frac{9}{5}g_A\approx2.27 in the normalization of Refs. [30, 31]. Chiral perturbation theory and a relativistic constituent-quark model give about 2.25, the chiral quark-soliton model 2.13, and lattice QCD at pion masses of 213–432 MeV gives 1.7–2.0, as compiled in Ref. [31]. The lattice range corresponds to gAΔ≈0.57g_{A\Delta}\approx0.57–0.66, which lies inside the range 0≤gAΔ≤gA0\leq g_{A\Delta}\leq g_A of our sensitivity test (Sec. VI). The spin-summed hadron–lepton contraction, computed with FORM, is given in Appendix C; it reproduced explicit RS spinor sums to 10−1010^{-10}. For both Δ\Deltas at rest it reduces to

∑∣M∣2→643G2∣⟨T+⟩∣2(3+5gAΔ2)mΔ∗mΔ′∗ElEν.(20)\sum|\mathcal{M}|^2\to\frac{64}{3}G^2|\langle T_+\rangle|^2(3+5g_{A\Delta}^2)m_\Delta^*m_{\Delta'}^*E_lE_\nu. \tag*{(20)}

Per initial state the Fermi strength is therefore the same as for the nucleon, sixteen times the isospin factor, and the Gamow–Teller strength is the nucleon one multiplied by the mean squared Dirac spin of the spin-32\frac{3}{2} state relative to the nucleon, which is five ninths, and by the squared ratio of the axial couplings.

In-medium width

In vacuum the Δ\Delta decays strongly to NπN\pi. In the medium, energy conservation for the decay of a Δ\Delta of momentum p\mathbf{p} into a mean-field nucleon and a vacuum pion, the available four-momentum and the decay momentum q\mathbf{q} in its rest frame are

EΔ∗(p)+UΔ−UN=EN∗+ωπ,E_\Delta^*(p)+U_\Delta-U_N=E_N^*+\omega_\pi,
P=(EΔ∗+UΔ−UN,p),\mathcal{P}=(E_\Delta^*+U_\Delta-U_N,\mathbf{p}),
q2+mN∗2+q2+mπ2=s,s=P2.(21)\sqrt{q^2+m_N^{*2}}+\sqrt{q^2+m_\pi^2}=\sqrt{s},\qquad s=\mathcal{P}^2 . \tag*{(21)}

We used the PP-wave form

ΓΔmed(p)=∑NBR⁡NΓ0(qq0)3mN∗/(EN∗+ωπ)MN/MΔ×sP0⟨1−fN(ENlab)⟩,(22)\Gamma_\Delta^{\mathrm{med}}(p)=\sum_N\operatorname{BR}_N\Gamma_0\left(\frac{q}{q_0}\right)^3\frac{m_N^*/(E_N^*+\omega_\pi)}{M_N/M_\Delta}\times\frac{\sqrt{s}}{\mathcal{P}^0}\langle1-f_N(E_N^{\mathrm{lab}})\rangle, \tag*{(22)}

normalized to the vacuum Breit–Wigner width at the Breit–Wigner mass [32]; these values, the corresponding vacuum decay momentum and the isospin branching ratios are listed in Table III. The Pauli factor is averaged over the decay angle after boosting the nucleon to the matter frame, and the ratio of the invariant mass to the energy is the time dilation. The width entering the NWA is the Fermi-surface average with weight proportional to the squared momentum times the product of the occupation and its complement. For a Δ\Delta at rest the formula reproduces the rest-frame value, and in vacuum it returns a decay momentum of 226.5 MeV and a width of 116.3 MeV. Pion self-energies and Bose enhancement were neglected.

Bulk viscosity with two slow variables

Because strong processes keep the Δ\Deltas in equilibrium, they do not introduce a new slow variable, but they enter every ingredient of the response. The pressure and its derivatives are those of matter in strong equilibrium at each perturbed state, and the relaxation matrix contains all the processes of Eq. (5). We therefore formulate the response directly for the two lepton fractions, generalizing the one-variable treatment of Ref. [8].

Coupled response of electrons and muons

Consider a fluid element undergoing a small periodic compression with angular frequency ω\omega and fluid velocity v\mathbf{v},

nB(t)=nˉB+δneiωt.(23)n_B(t)=\bar{n}_B+\delta n e^{i\omega t}. \tag*{(23)}

Baryon and lepton number conservation give

n˙B+nB∇⋅v=0,n˙l+nl∇⋅v=Γl,(24)\dot{n}_B+n_B\nabla\cdot\mathbf{v}=0,\qquad\dot{n}_l+n_l\nabla\cdot\mathbf{v}=\Gamma_l , \tag*{(24)}
QuantityValue
Reference density n0n_00.16 fm−30.16\ \mathrm{fm}^{-3}
Fermi constant GFG_F1.166×10−11 MeV−21.166 \times10^{-11}\ \mathrm{MeV}^{-2}
Cabibbo angle θC\theta_C13.04∘13.04^\circ
Nucleon axial coupling gAg_A1.261.26
NΔN\Delta axial coupling C5AC^A_51.21.2 (varied to 0.870.87)
ΔΔ′\Delta\Delta' axial coupling gAΔ′g_{A\Delta'}, Eq. (19)0.7560.756 (varied to 00)
Isospin factor ∣I∣2\lvert I\rvert^2: Δ−→n\Delta^- \to n, Δ0→p\Delta^0 \to p3,13, 1
Isospin factor ∣⟨T+⟩∣2\lvert\langle T_+\rangle\rvert^2: Δ−→Δ0\Delta^- \to\Delta^0, Δ0→Δ+\Delta^0 \to\Delta^+, Δ+→Δ++\Delta^+ \to\Delta^{++}3,4,33, 4, 3
Lepton masses mem_e, mμm_\mu0.511,105.6580.511, 105.658
Δ\Delta mass MΔM_\Delta (PDG estimate [32])12321232 (12301230–12341234)
Vacuum width Γ0\Gamma_0 (PDG estimate [32])117117 (114114–120120)
Vacuum decay momentum q0q_0; pion mass mπm_\pi227;139.57227; 139.57
NπN\pi branching: Δ−;\Delta^-; Δ0\Delta^0 (p,np,n); Δ+\Delta^+ (p,np,n); Δ++\Delta^{++}1;13;23;23;13;11; \frac{1}{3}; \frac{2}{3}; \frac{2}{3}; \frac{1}{3}; 1
NWA width scale TWT_W, Eq. (B10)5 MeV5\ \mathrm{MeV}
Extra Δ\Delta width in sensitivity test1010
Oscillation frequency ff1 kHz; 0.51\ \mathrm{kHz};\ 0.5–5 kHz5\ \mathrm{kHz} in Figure 5
Damping-time contour in Figure 520 ms20\ \mathrm{ms}
Quasi-Monte Carlo points per rate2172^{17}
Energy window around Fermi energies±16T\pm16T (plus ∣δμl∣\lvert\delta\mu_l\rvert)
Gauss–Chebyshev nodes in zlz_l66
Finite-difference steps in YlY_l and nBn_B2×10−3Yl, 10−3nB2 \times10^{-3}Y_l,\ 10^{-3}n_B
Equilibrium tolerance (net/gross rate)10−510^{-5}
Density grid11–6 n06\ n_0, step 0.5 n00.5\ n_0; 0.25 n00.25\ n_0 for nucleonic matter at 1,2,3 MeV1, 2, 3\ \mathrm{MeV}
Grid of Figure 5 (isothermal)11–6 n06\ n_0, step 0.25 n00.25\ n_0; T=1,1.5,2T=1, 1.5, 2–4.54.5 (step 0.250.25), 5,5.5,6,7,8,10 MeV5, 5.5, 6, 7, 8, 10\ \mathrm{MeV}
Temperature grid1,2,2.5,3,3.5,4,4.5,5,7,101, 2, 2.5, 3, 3.5, 4, 4.5, 5, 7, 10; also 1.5,2.251.5, 2.25 for the Δ2\Delta2 sets

Table 3. Weak-interaction, Δ\Delta and numerical inputs. Masses and energies in MeV.

hence

nBdYldt=Γl(nB,T,Ye,Yμ),l=e,μ.(25)n_B \frac{dY_l}{dt} = \Gamma_l(n_B,T,Y_e,Y_\mu), \qquad l=e,\mu. \tag*{(25)}

We linearized about the weak equilibrium of Eq. (7), writing

Yl=Yˉl+δYleiωt,(26)Y_l = \bar{Y}_l + \delta Y_l e^{i\omega t}, \tag*{(26)}

and obtained

iωδYa=−∑bGabδYb+γB,anˉBδn,(27)i\omega\delta Y_a=-\sum_b \mathcal{G}_{ab}\delta Y_b+\frac{\gamma_{B,a}}{\bar{n}_B}\delta n, \tag*{(27)}

with the relaxation matrix and the density response

Gab=−1nB∂Γa∂Yb∣nB,Yc≠b,γB,a=∂Γa∂nB∣Ye,Yμ.(28)\mathcal{G}_{ab}=-\left.\frac{1}{n_B}\frac{\partial\Gamma_a}{\partial Y_b}\right|_{n_B,Y_{c\ne b}},\qquad \gamma_{B,a}=\left.\frac{\partial\Gamma_a}{\partial n_B}\right|_{Y_e,Y_\mu}. \tag*{(28)}

The solution is

δY=(iω1+G)−1γBδnnˉB.(29)\delta\mathbf{Y}=(i\omega\mathbb{1}+\mathcal{G})^{-1}\boldsymbol{\gamma}_B\frac{\delta n}{\bar{n}_B}. \tag*{(29)}

The departure of the pressure from the value it would have at frozen composition, and the bulk viscous pressure that defines the bulk viscosity, are

δPneq=∑a∂P∂YaδYa≡X(ω)δnnˉB,\delta P_{\mathrm{neq}}=\sum_a\frac{\partial P}{\partial Y_a}\delta Y_a\equiv X(\omega)\frac{\delta n}{\bar{n}_B},
Π=−ζ∇⋅v=ζiωδnnˉB,(30)\Pi=-\zeta\nabla\cdot\mathbf{v}=\zeta i\omega\frac{\delta n}{\bar{n}_B}, \tag*{(30)}

the latter being the component of the former in phase with the compression rate. With

X(ω)=∑a,b∂P∂Ya[(iω1+G)−1]abγB,b,(31)X(\omega)=\sum_{a,b}\frac{\partial P}{\partial Y_a}\left[(i\omega\mathbb{1}+\mathcal{G})^{-1}\right]_{ab}\gamma_{B,b}, \tag*{(31)}

the dissipative part is the imaginary part of XX, and

ζ(ω)=Im⁡X(ω)ω.(32)\zeta(\omega)=\frac{\operatorname{Im}X(\omega)}{\omega}. \tag*{(32)}

For a single slow variable, without muons, Eq. (32) reduces to

ζ=−∂P∂YeγBγ2+ω2,γ=Gee,(33)\zeta=-\frac{\partial P}{\partial Y_e}\frac{\gamma_B}{\gamma^2+\omega^2},\qquad\gamma=\mathcal{G}_{ee}, \tag*{(33)}

the form of Ref. [8] with the isospin fraction replaced by the electron fraction. The eigenvalues of G\mathcal{G} are the two relaxation rates of the coupled electron–muon system, and we call the inverse of the diagonal elements the electron and muon relaxation times,

τe=1/Gee,τμ=1/Gμμ.(34)\tau_e=1/\mathcal{G}_{ee},\qquad\tau_\mu=1/\mathcal{G}_{\mu\mu}. \tag*{(34)}

In the presence of Δ\Deltas the pressure derivatives and the density response are evaluated with the baryon composition in strong equilibrium at every state, so the response of the Δ\Delta populations to compression is included at the level of the susceptibilities.

The derivatives were taken by central differences, with the steps of Table III, using the same quasi-random point set for every rate evaluation (common random numbers), so that the finite differences of the rates carry no noise from independent samples. As a check of the linear-response machinery we verified, in matter with Δ\Deltas, the Maxwell-type identity

1nB∂P∂Ye∣nB,T=nB∂(μQ+μe)∂nB∣Ye,T.(35)\left.\frac{1}{n_B}\frac{\partial P}{\partial Y_e}\right|_{n_B,T} = \left.n_B\frac{\partial(\mu_Q+\mu_e)}{\partial n_B}\right|_{Y_e,T}. \tag*{(35)}

which follows from the Gibbs–Duhem relation

dP=s dT+∑ini dμi(36)dP = s\,dT + \sum_i n_i\,d\mu_i \tag*{(36)}

and the strong-equilibrium relations, to 3×10−83 \times10^{-8}.

Oscillation type, damping time and resonance

We evaluate Eq. (28) both at fixed temperature (isothermal oscillations) and at fixed entropy per baryon (adiabatic oscillations), and convert the bulk viscosity into the damping time of a density oscillation, Eq. (77). These standard relations, together with the resonance form that we use to interpret the frequency dependence, are given in Appendix § 1.5.4.

Validation

We first validated our rate code by reproducing the nucleonic results of Ref. [11] (and, through them, Refs. [8, 9]) for IUF and QMC-RMF3, with and without magnetic fields. Over the 88 published curves we compared with, the median deviations in the rates, the relaxation time and the bulk viscosity are a few hundredths of a dex. This comparison used the constant matrix element of Refs. [10, 11]. With the full matrix element of Eq. (70) the rates are lower by a factor of two to three, and all the results below use the full matrix elements.

We then tested the extension to Δ\Deltas in several ways. With the Δ\Deltas removed, both the equation of state and the full bulk-viscosity calculation reproduce the nucleonic code to numerical precision. The generalized rate kernel reproduces the nucleon NWA kernel, and for the Δ→N\Delta\to N processes the full phase-space rate approaches the Fermi surface limit of the same matrix element in degenerate kinematics. The matrix elements of Eqs. (15) and (74) agree with explicit Rarita–Schwinger spinor sums to ten digits, and the two-variable response reduces to the one-variable result when the muon variable is removed. In matter with Δ\Deltas the Gibbs–Duhem relation and the thermodynamic identity of Eq. (35) are satisfied to better than 10−710^{-7}. We accept a flavor-equilibrium state only when the net rates are below the tolerance of Table III relative to the gross rates. This criterion is not met at one grid point (QMC-Δ\Delta1 at n0n_0 and 10 MeV), where the net rate is a small difference between large decay and capture rates, and we exclude that point.

Results

Composition and the opening of direct Urca

In Figure 1 we show the composition of β\beta-equilibrated matter for QMC-Δ\Delta2 and DD-Δ\Delta2. Once the Δ−\Delta^- threshold is crossed, it is energetically cheaper to carry negative charge in a heavy baryon near the bottom of a shallow Fermi sea than in a relativistic electron at the top of a deep one. The Δ−\Delta^- therefore takes over the role of the leptons in maintaining charge neutrality: the electron fraction drops by more than an order of magnitude, the muons disappear, and the proton fraction roughly doubles. With the larger proton Fermi momentum, the neutron Fermi momentum no longer exceeds the sum of the proton and electron Fermi momenta, so momentum conservation allows the nucleonic direct Urca process in all four Δ\Delta sets (Table II), as Raduta found for cold DDME2 [19]. For our couplings the corresponding conditions for the Δ\Delta processes themselves are violated in cold matter, so these processes remain kinematically forbidden at the Fermi surfaces. This is because the four-fold spin degeneracy keeps the Δ\Delta Fermi momenta small, while the electron Fermi momentum is small as well. At finite temperature these processes nevertheless contribute through thermal blurring of the Fermi surfaces and collisional broadening.

Figure 1. Particle fractions of β\beta-equilibrated matter (equal electron and muon chemical potentials, both equal to the neutron–proton difference) at 1 MeV for (a) QMC-Δ\Delta2 and (b) DD-Δ\Delta2. Thick lines, labeled on the curves: matter with Δ\Deltas. Thin lines of the same color and style: proton, electron and muon fractions of the nucleonic sets QMC-N and DD-N, with the proton line marked p (N). Vertical dotted lines mark the Δ−\Delta^- onset and the density above which the cold nucleonic direct Urca triangle is open. The Δ−\Delta^- replaces electrons and muons as the negative charge and doubles the proton fraction, which opens the direct Urca triangle.

Stability of the against strong decay

Although the decay Δ→Nπ\Delta\to N\pi remains kinematically allowed at all densities, the nucleon it would produce has a momentum well inside the nucleon Fermi sea. Each Δ\Delta species is therefore Pauli blocked against strong decay as soon as it is populated. The in-medium width is appreciable only for Δ0\Delta^0 and Δ+\Delta^+ at low density, where these species are absent or only dilute (Figure 7 in Appendix § 1.5.5). In the mean-field description, the populated Δ\Deltas are thus stable quasiparticles whose spectral width is determined by collisions.

Figure 7. Fermi-surface averaged in-medium Δ→Nπ\Delta\to N\pi width, Eq. (22), for Δ−\Delta^{-}, Δ0\Delta^{0} and Δ+\Delta^{+} in β\beta-equilibrated matter for (a) QMC-Δ\Delta2 and (b) DD-Δ\Delta2, at 1 MeV (thick) and 5 MeV (thin). Dotted grey line, right axis: Δ−\Delta^{-} fraction at 1 MeV. The decay is Pauli blocked wherever the decaying species is populated; the nonzero widths at low density belong to species that are absent there.

Relaxation and the channels that carry it

Urca processes are the mechanism by which the lepton fractions relax to their equilibrium values. In Figure 2 we plot the electron relaxation time with and without Δ\Deltas. In nucleonic matter the direct Urca process is kinematically suppressed at low temperature, so equilibration proceeds through the collisional (“modified Urca”) contribution of the NWA; it is slow at 1 MeV and speeds up rapidly as the temperature rises. Above the Δ−\Delta^- threshold the relaxation time drops sharply, by more than two orders of magnitude at 1 MeV and by one to two orders at 3–5 MeV for the sets with the larger scalar coupling. At 1 MeV the drop coincides with the opening of the nucleonic direct Urca process and appears as a step in Figure 2. In nucleonic matter the muon relaxation time is longer than the electron relaxation time, consistent with Ref. [12]. With Δ\Deltas the muons are nearly absent above threshold, and the second mode of the coupled response plays no role.

Figure 2. Electron relaxation time (thick lines) and muon relaxation time (thin lines, shown where the muon fraction exceeds 10−310^{-3}), Eq. (34), as functions of baryon density at 1, 3 and 5 MeV (columns), for QMC-RMF3 (top) and DDME2 (bottom), in isothermal, neutrino-transparent matter at the thermal weak equilibrium. Line styles distinguish the nucleonic and the two Δ\Delta sets of Table II; triangles on the lower axis mark the cold nucleonic direct Urca threshold of each Δ\Delta set, and the dotted grey line the inverse angular frequency at 1 kHz. The Δ\Deltas shorten the electron relaxation time by one to almost three orders of magnitude above the threshold.

To see which processes are responsible, in Figure 3 we show how the equilibrium rate is shared among the channels. At 1 MeV almost all of it is carried by the nucleonic process, which the change in composition has opened. The Δ\Delta processes are forbidden in cold matter and open only through thermal blurring and collisional broadening, so their share grows with temperature; together they carry more than half of the rate above about 3.5 MeV in DD-Δ\Delta2 and 4.5 MeV in QMC-Δ\Delta2. Which Δ\Delta process dominates depends on the composition. In QMC-Δ\Delta2, where only the Δ−\Delta^- is abundant at this density, Δ−→n\Delta^- \to n provides most of the Δ\Delta contribution, whereas in DD-Δ\Delta2, where the Δ0\Delta^0 is already present, the Δ→Δ′\Delta\to\Delta' processes contribute as much or more. We see that the Δ\Deltas act in two stages: at low temperature through the composition, which opens the nucleonic process, and at higher temperature through their own weak processes.

Figure 3. Share of the gross equilibrium Urca rate (decay plus capture, electrons and muons) carried by the nucleonic channel, the Δ↔N\Delta\leftrightarrow N channels (Δ−↔n\Delta^- \leftrightarrow n, Δ0↔p\Delta^0 \leftrightarrow p) and the Δ↔Δ′\Delta\leftrightarrow\Delta' channels, as a function of temperature at three times n0n_0 for (a) QMC-Δ2 and (b) DD-Δ2; circles are the computed temperatures. The vertical dotted line marks the temperature above which the Δ\Delta channels together carry more than half of the rate. The Δ\Delta channels, closed in cold matter, take over as the temperature rises.

Bulk viscosity and the resonance

One of the physical manifestations of flavor relaxation is bulk viscosity, which damps density oscillations. In Figure 4 we plot the bulk viscosity for 1 kHz oscillations as a function of temperature. The bulk viscosity attains a resonant maximum when the electron relaxation rate matches the angular frequency of the oscillation (Appendix § 1.5.4). At lower temperatures the matter equilibrates so slowly that the lepton fractions are almost conserved during a cycle, and the bulk viscosity is small. At higher temperatures the matter equilibrates so quickly that there is little phase lag between pressure and density, and the bulk viscosity again becomes small.

Figure 4. Bulk viscosity at 1 kHz as a function of temperature at two, three and four times n0n_0 (columns) for QMC-RMF3 (top) and DDME2 (bottom). Lines with circles: isothermal oscillations at the computed temperatures; crosses: adiabatic oscillations (fixed entropy per baryon), computed on the coarser base grid. Arrows mark the resonance temperature, at which the electron relaxation rate equals the angular frequency. Line styles as in Figure 2. The Δ\Deltas move the resonance to lower temperature, raise the bulk viscosity below the nucleonic resonance and lower it above.

Since the Δ\Deltas speed up the relaxation, they shift the resonant peak to lower temperature, from about 4–6 MeV in nucleonic matter to about 2–3.7 MeV in the Δ\Delta sets. The shift is larger for the larger scalar coupling and at higher density, where the Δ\Delta fraction is larger. We obtain the resonance temperature by log–log interpolation of the electron relaxation rate between neighboring grid temperatures; removing the grid points at 2.5, 3.5, 4 and 4.5 MeV changes it by at most 0.03 MeV, so the interpolation error is negligible. The shift resembles the effect of crossing a nucleonic direct Urca threshold, which in the IUF model speeds up the relaxation above about 4n04n_0 [8]. Here, however, the threshold is set by the onset of the Δ−\Delta^- and therefore lies at lower density, between 1.45 and 2.3 times n0n_0 in our sets, and above about 3 MeV the Δ\Delta processes accelerate the relaxation further. The height of the peak is set by the thermodynamic strength of the response rather than by the rates, and it changes by less than about an order of magnitude in either direction. The Δ\Deltas therefore have opposite effects on the two sides of the resonance. At 2–3 MeV, below the nucleonic resonance, the relaxation rate is brought closer to resonance and the bulk viscosity increases by more than an order of magnitude. Above the resonance, and in particular at four times n0n_0 and 5–10 MeV, the faster equilibration moves the rate further away from resonance, and the Δ\Deltas reduce the bulk viscosity by one to almost three orders of magnitude.

At high density the composition can suppress the bulk viscosity altogether. In DD-Δ\Delta2 the derivative of the pressure with respect to the electron fraction and the density response of the rate change sign together near five times n0n_0, at all temperatures. According to Eq. (35), this is the density at which the lepton chemical potential becomes independent of density at fixed electron fraction. Compression then no longer drives the matter out of β\beta equilibrium, and the bulk viscosity vanishes regardless of the rates; QMC-Δ\Delta2 approaches the same behavior at the upper end of our grid. Close to this zero, and in DD-N at 10 MeV and high density, where the finite-temperature correction to the equilibrium condition is large, the linearized calculation yields small negative values, at 3 of the 703 points of our base grid and at 11 of the 2392 points of the denser grid used for Figure 5. Because the linearization is about a state with nonzero chemical imbalances, the density response of the rate is no longer strictly proportional to the density derivative of the imbalance, and positivity is not guaranteed. We treat these points as undamped.

Figure 5. Frequency dependence of the bulk-viscous damping in the density–temperature plane for nucleonic matter (left), QMC-Δ1\Delta1 and DD-Δ1\Delta1 (middle) and QMC-Δ2\Delta2 and DD-Δ2\Delta2 (right); QMC-RMF3 top, DDME2 bottom; isothermal. Blue shading and thin black contours with inline labels: resonance frequency, Eq. (80), in kHz. Thick colored contours, identified in the legend of panel (d): damping time of 20 ms, Eq. (77), for density oscillations at 0.5, 1, 2, 3 and 5 kHz; oscillations damp faster inside each contour. Dashed vertical lines: Δ−\Delta^{-} onset. The isothermal grid of Table III (step 0.25n0n_0, 19 temperatures) was interpolated bilinearly in the logarithms, in density and logarithmic temperature, for display; points with non-positive bulk viscosity are treated as undamped. The Δ\Deltas lower the resonance temperature at every frequency and extend the fast-damping regions to lower temperature.

Wherever the bulk viscosity is large, adiabatic and isothermal oscillations give nearly the same value (crosses in Figure 4). Appreciable differences appear only at 10 MeV and near the zero in DD-Δ\Delta2, where the bulk viscosity is small. This is because the temperatures involved are lower than the Fermi energies of the relevant particles, so the entropy contribution to the pressure is a small correction [8]. At every grid point the adiabatic modulus exceeds the isothermal one, as required by thermodynamic stability.

For nucleonic DDME2 our resonance temperature is 1–2 MeV higher than the temperature at which Ref. [12] found the electron relaxation rate to equal the oscillation frequency. In Appendix F we show that the rate model accounts for only a small part of this difference, and that the main source is the equilibrium state about which the rates are linearized. Reference [12] linearized about the cold condition of vanishing chemical imbalance, at which electron capture greatly exceeds neutron decay in hot matter, whereas we linearize about the finite-temperature equilibrium, following Refs. [7, 8]. When we adopt their prescription, our code reproduces their crossing temperature. The earlier study of the same model [33], which also used the cold equilibrium, placed the 1 kHz maximum at about 5 MeV. That value refers to the total bulk viscosity of npeμnpe\mu matter. For npeμnpe\mu matter Ref. [12] likewise places the 1 kHz minimum of the damping time at about 5 MeV; its 3–4 MeV value refers to the crossing of the electron relaxation rate of npenpe matter below threshold, which is the quantity we compare with. The two studies therefore do not conflict. Reference [33] also estimated that the finite-temperature correction to the equilibrium would move the maximum by about 1 MeV, which agrees with the shift we find (Appendix F). The Δ\Delta-induced shifts discussed in this paper are differences within a single framework and do not depend on this choice.

Frequency dependence and damping

The frequency of the dominant density oscillations varies between remnants and between the oscillation modes of one remnant, so we now evaluate the response for several frequencies; the stored relaxation matrices allow this without new rate calculations. In Figure 5 we show the resonance frequency, i.e., the frequency for which a given density and temperature are on resonance, together with the regions in which oscillations of 0.5 to 5 kHz are damped within 20 ms, a time comparable to the early post-merger phase.

We note the following features. (1) In the nucleonic models the resonance-frequency contours rise steeply and almost uniformly with temperature and only slowly with density, because there is no direct Urca threshold. (2) With Δ\Deltas the contours bend downward above the Δ−\Delta^- threshold, so that wherever Δ\Deltas are present a given frequency is resonant at a temperature one to three MeV lower; at higher density they flatten or turn upward again. Doubling the frequency raises the resonance temperature by about half an MeV.

frequency is resonant at a temperature one to three MeV lower; at higher density they flatten or turn upward again. Doubling the frequency raises the resonance temperature by about half an MeV. (3) The damping regions lie along the resonance contours of the corresponding frequency, since an oscillation is damped most efficiently where it is on resonance. The shortest damping time at a given frequency is then set by the thermodynamic strength of the response and scales inversely with the frequency (Appendix § 1.5.4), and our results follow this scaling closely. (4) In nucleonic matter no region damps 0.5 kHz oscillations within 20 ms, 1 kHz oscillations are damped this fast only near saturation density, and the regions for higher frequencies start near 3 MeV at saturation density and move to 4–8 MeV at higher density. With Δ\Deltas the damping regions extend from saturation density to three to five times n0n_0, reach down to 2–2.75 MeV and are present at every frequency we consider, with shortest damping times about 2.5 to 9 times below the nucleonic ones. At high density and temperature the damping regions close again, for the reasons discussed above.

Sensitivity to the Δ\Delta weak inputs

Finally, in Figure 6 we show how the bulk viscosity of QMC-Δ2 and DD-Δ2 responds when single inputs are varied. The weak couplings have little influence: lowering the NΔN\Delta axial coupling or setting the ΔΔ′\Delta\Delta' axial coupling to zero changes the bulk viscosity by at most a quarter of a dex. Removing the Δ\Delta weak processes while keeping the Δ\Deltas in the equation of state changes it by about half a dex or less up to 3 MeV, which confirms that the change in composition controls the response of moderately hot matter. At 5 MeV and high density, where Δ0\Delta^{0} and Δ+\Delta^{+} are abundant and the Δ\Delta processes carry most of the rate, the change reaches an order of magnitude. The in-medium Δ→Nπ\Delta\to N\pi width has no visible effect, as expected from Pauli blocking. The results are most sensitive to the collisional width of the Δ\Delta. Just above the Δ−\Delta^{-} threshold at 1 MeV, where the Δ\Delta processes are kinematically suppressed, a larger width opens more of their phase space and increases the bulk viscosity by more than an order of magnitude, whereas at 3 MeV and above the effect stays below a third of a dex. The collisional width of the Δ\Delta in dense matter is the least constrained input of our calculation, because no microscopic calculation exists for degenerate Δ\Deltas (Appendix B).

Conclusions

In this work, we have calculated flavor equilibration rates and the resulting bulk viscosity of neutrino-

transparent nuclear matter containing Δ(1232)\Delta(1232) isobars in strong equilibrium, at the densities and temperatures relevant to neutron star mergers. To do so we derived the spin-summed matrix elements for the charged-current Δ→N\Delta\to N and Δ→Δ′\Delta\to\Delta' transitions from the Rarita–Schwinger spin sum, included them together with the nucleonic processes in full phase-space rates in the NWA, treated the in-medium strong width of the Δ\Delta with Pauli blocking, and obtained the bulk viscosity from the coupled response of the electron and muon fractions about the finite-temperature flavor equilibrium. The calculations were performed for two equations of state without a nucleonic direct Urca threshold, QMC-RMF3 and DDME2, each with two sets of Δ\Delta couplings.

We found that the Δ\Deltas affect the bulk viscosity mainly through the composition. As the Δ−\Delta^{-} takes over the negative charge from the leptons, the proton fraction rises and the nucleonic direct Urca process opens in models where it is otherwise closed. This controls flavor equilibration at temperatures of a few MeV, while the Δ\Delta weak processes, which are forbidden in cold matter, become dominant only at higher temperature. The faster equilibration lowers the temperature at which the bulk viscosity reaches its resonant peak, for every oscillation frequency, with little change in the height of the peak. Consistently with this, the bulk viscosity increases at 2–3 MeV, where nucleonic matter is still well below resonance, and decreases in hotter and denser matter, which is pushed further above resonance. At high density the composition can make the bulk viscosity vanish altogether. For kHz oscillations the regions of fastest damping become broader in density, extend to lower temperature and appear even below 1 kHz, with damping times several times shorter than in nucleonic matter. This can lead to effective bulk viscous damping in cooler regions of a merger if Δ\Deltas are present. The populated Δ\Deltas are Pauli blocked against strong decay, so they behave as long-lived quasiparticles throughout.

For nucleonic DDME2 we found that linearizing about the cold rather than the finite-temperature equilibrium lowers the resonance temperature by up to a few MeV at high density, more than any refinement of the rate model. This accounts for most of the difference between our results and those of Ref. [12], and it should be kept in mind when different calculations are compared.

There are many directions in which further work is needed. (1) The main uncertainties lie in the Δ\Delta sector. The Δ\Delta–meson couplings are not constrained by data, and we used two sets per model with a common vector coupling ratio. The collisional width of the Δ\Delta was set equal to that of the nucleon; it is the only input that changes our results by more than an order of magnitude, near the Δ−\Delta^{-} threshold at low temperature, and a microscopic calculation of it would be valuable. The axial couplings of the Δ\Delta transitions, taken from the low-momentum-transfer form factor and from the quark model, have a minor effect. (2) Our most significant assumption about the environment is neutrino transparency, which we applied up to 10 MeV; at the higher temperatures an explicit treatment of neutrino transport will be needed. We also assumed that purely leptonic conversion between electrons and muons is frozen. (3) Our calculation assumes linear response. Large-amplitude oscillations, and coupled slow variables driven strongly, can take the matter far beyond this regime [24], and a direct implementation of the rate equations in merger simulations, of the kind performed in Ref. [3], would include these effects naturally and show whether the lower resonance temperature affects the post-merger evolution and its gravitational-wave signal [4].

Relativistic mean-field model

Lagrangian

The baryons were the nucleon doublet ψN\psi_N and the Δ(1232)\Delta(1232) quartet ψΔμ\psi^\mu_\Delta,

ψN=(ψpψn),ψΔμ=(ψΔ++μψΔ+μψΔ0μψΔ−μ),(37)\psi_N = \begin{pmatrix} \psi_p \\ \psi_n \end{pmatrix}, \qquad \psi^\mu_\Delta= \begin{pmatrix} \psi^\mu_{\Delta^{++}} \\ \psi^\mu_{\Delta^+} \\ \psi^\mu_{\Delta^0} \\ \psi^\mu_{\Delta^-} \end{pmatrix}, \tag*{(37)}

the latter described by Rarita–Schwinger (RS) vector-spinors [25]. They interacted through the isoscalar-scalar σ\sigma, isoscalar-vector ωμ\omega^\mu and isovector-vector ρμ\rho^\mu mesons. The Lagrangian density was

L=ψˉN(iγμDNμ−mN+gσNσ)ψN+ψˉΔμ(iγνDΔν−MΔ+gσΔσ)ψΔμ+12∂μσ∂μσ−12mσ2σ2−U(σ)−14ωμνωμν+12mω2ωμωμ−14ρμν⋅ρμν+12mρ2ρμ⋅ρμ+Lωρ+∑l=e,μψˉl(i∂̸−ml)ψl.(38)\begin{aligned} \mathcal{L} ={}& \bar{\psi}_N\left(i\gamma_\mu D_N^\mu-m_N+g_{\sigma N}\sigma\right)\psi_N \\ &+\bar{\psi}_{\Delta\mu}\left(i\gamma_\nu D_\Delta^\nu-M_\Delta+g_{\sigma\Delta}\sigma\right)\psi_\Delta^\mu\\ &+\frac{1}{2}\partial_\mu\sigma\partial^\mu\sigma-\frac{1}{2}m_\sigma^2\sigma^2-U(\sigma) \\ &-\frac{1}{4}\omega_{\mu\nu}\omega^{\mu\nu}+\frac{1}{2}m_\omega^2\omega_\mu\omega^\mu-\frac{1}{4}\rho_{\mu\nu}\cdot\rho^{\mu\nu} \\ &+\frac{1}{2}m_\rho^2\rho_\mu\cdot\rho^\mu+\mathcal{L}_{\omega\rho}+\sum_{l=e,\mu}\bar{\psi}_l(i\not{\partial}-m_l)\psi_l . \tag*{(38)} \end{aligned}

with the covariant derivative

Dbμ=∂μ+igωbωμ+igρbtb⋅ρμ,(39)D_b^\mu=\partial^\mu+ig_{\omega b}\omega^\mu+ig_{\rho b}t_b\cdot\rho^\mu, \tag*{(39)}

where tbt_b is the isospin operator of baryon bb; its third component is ±12\pm\frac{1}{2} for the proton and the neutron and 32,12,−12,−32\frac{3}{2},\frac{1}{2},-\frac{1}{2},-\frac{3}{2} for Δ++\Delta^{++}, Δ+\Delta^+, Δ0\Delta^0, Δ−\Delta^-. The RS field obeys the constraints

γμψΔμ=0,∂μψΔμ=0,(40)\gamma_\mu\psi^\mu_\Delta=0,\qquad\partial_\mu\psi^\mu_\Delta=0, \tag*{(40)}

and in the mean-field approximation used below each Δ\Delta is a free spin-32\frac{3}{2} quasiparticle with four spin states, an effective mass and a vector self-energy. The Δ\Delta–meson couplings were written as fractions of the nucleon couplings,

giΔ=xigiN,i=σ,ω,ρ,(41)g_{i\Delta}=x_i g_{iN},\qquad i=\sigma,\omega,\rho, \tag*{(41)}

with the ratios xix_i of each coupling set listed in Table II. We used two nucleonic models, whose parameters are listed in Table I.

QMCQMC-RMF3RMF3 [26, 27] has constant couplings, the scalar self-interaction and the ωρ\omega\rho mixing term

U(σ)=κ6(gσσ)3+λ24(gσσ)4,U(\sigma)=\frac{\kappa}{6}(g_\sigma\sigma)^3+\frac{\lambda}{24}(g_\sigma\sigma)^4,
Lωρ=Λvgω2gρ2(ωμωμ)(ρν⋅ρν),(42)\mathcal{L}_{\omega\rho}=\Lambda_v g_\omega^2g_\rho^2(\omega_\mu\omega^\mu)(\rho_\nu\cdot\rho^\nu), \tag*{(42)}

with the parameters of Table I of Ref. [26]; Table I gives both the original parameters and their translation into the notation of Eq. (42). In cold β\beta-equilibrated matter without Δ\Deltas, with or without muons, the nucleonic direct Urca triangle [34], i.e. the condition that the neutron Fermi momentum does not exceed the sum of the proton and electron Fermi momenta, stays closed up to eight times n0n_0 in this model (our calculation; Ref. [11] used QMCQMC-RMF3RMF3 as its model without a direct Urca threshold).

DDME2DDME2 [28] has no scalar self-interaction and no ωρ\omega\rho term, and density-dependent couplings

gi(nB)=gi(nsat)fi(x),x≡nB/nsat,g_i(n_B)=g_i(n_{\mathrm{sat}})f_i(x),\qquad x\equiv n_B/n_{\mathrm{sat}},
fσ,ω(x)=ai1+bi(x+di)21+ci(x+di)2,fρ(x)=e−aρ(x−1),(43)f_{\sigma,\omega}(x)=a_i\frac{1+b_i(x+d_i)^2}{1+c_i(x+d_i)^2},\qquad f_\rho(x)=e^{-a_\rho(x-1)}, \tag*{(43)}

with the parameter values of Ref. [28]. The nucleonic direct Urca triangle is also closed up to eight times n0n_0 in this model, in agreement with Refs. [12, 19]. Both models thus isolate the effect of the Δ\Deltas on a background in which the nucleonic direct Urca process is closed in cold matter.

Table I also lists the saturation properties computed with our solver. For QMCQMC-RMF3RMF3 they agree with Ref. [26] to the digits quoted there, with the symmetry energy and its slope defined from the difference between pure neutron matter and symmetric matter as in that reference. For DDME2DDME2 the quadratic symmetry energy and its slope agree with the values 32.332.3 MeV and 51.251.2 MeV quoted in Ref. [19].

Mean-field equations and thermodynamics

In uniform, static matter the meson fields are replaced by their expectation values, and we used the self-energies

Φ=gσσˉ,W=gωωˉ0,B=gρρˉ0,3.(44)\Phi=g_\sigma\bar{\sigma},\qquad W=g_\omega\bar{\omega}_0,\qquad B=g_\rho\bar{\rho}_{0,3}. \tag*{(44)}

Each baryon species bb is then a free Fermi gas with spin degeneracy ηb\eta_b (two for nucleons, four for Δ\Deltas), effective mass and single-particle energy

mb∗=Mb−xσbΦ,(45)m_b^*=M_b-x_{\sigma b}\Phi, \tag*{(45)}
Eb(k)=k2+mb∗2+Ub,E_b(k)=\sqrt{k^2+m_b^{*2}}+U_b,
Ub=xωbW+xρbθbB+ΣR,(46)U_b=x_{\omega b}W+x_{\rho b}\theta_b B+\Sigma_R, \tag*{(46)}

where the coupling ratios are one for nucleons, the isospin factor θb\theta_b is given in Table I, and the rearrangement self-energy ΣR\Sigma_R of Eq. (51) is present only for density-dependent couplings. With the effective chemical potential, the kinetic energy and the occupation

νb=μb−Ub,Ek∗=k2+mb∗2,\nu_b=\mu_b-U_b,\qquad E_k^*=\sqrt{k^2+m_b^{*2}},
fb=11+e(Ek∗−νb)/T.(47)f_b=\frac{1}{1+e^{(E_k^*-\nu_b)/T}}. \tag*{(47)}

the kinetic contributions of species bb at temperature TT are

nb=ηb∫d3k(2π)3fb,nbs=ηb∫d3k(2π)3mb∗Ek∗fb,n_b=\eta_b\int\frac{d^3k}{(2\pi)^3}f_b,\qquad n_b^s=\eta_b\int\frac{d^3k}{(2\pi)^3}\frac{m_b^*}{E_k^*}f_b,
Pbkin=ηb3∫d3k(2π)3k2Ek∗fb,P_b^{\mathrm{kin}}=\frac{\eta_b}{3}\int\frac{d^3k}{(2\pi)^3}\frac{k^2}{E_k^*}f_b,
ϵbkin=ηb∫d3k(2π)3Ek∗fb,\epsilon_b^{\mathrm{kin}}=\eta_b\int\frac{d^3k}{(2\pi)^3}E_k^*f_b,
sb=(ϵbkin+Pbkin−νbnb)/T.(48)s_b=\left(\epsilon_b^{\mathrm{kin}}+P_b^{\mathrm{kin}}-\nu_bn_b\right)/T. \tag*{(48)}

Antiparticles were neglected. The most exposed case is the electron at the smallest electron fraction we encountered, about 2×10−32\times10^{-3} at three times n0n_0, where the electron chemical potential is about 70 MeV70\ \mathrm{MeV}; at 10 MeV10\ \mathrm{MeV} the positron density is then suppressed by a factor of about 10−310^{-3}.

The field equations follow from the stationarity of the grand potential Ω\Omega with respect to Φ\Phi, WW and BB. For QMC-RMF3,

mσ2gσ2Φ+κ2Φ2+λ6Φ3=∑bxσbnbs,\frac{m_\sigma^2}{g_\sigma^2}\Phi+\frac{\kappa}{2}\Phi^2+\frac{\lambda}{6}\Phi^3=\sum_b x_{\sigma b}n_b^s,
mω2gω2W+2ΛvB2W=∑bxωbnb,\frac{m_\omega^2}{g_\omega^2}W+2\Lambda_vB^2W=\sum_b x_{\omega b}n_b,
(mρ2gρ2+2ΛvW2)B=∑bxρbt3bnb.(49)\left(\frac{m_\rho^2}{g_\rho^2}+2\Lambda_vW^2\right)B=\sum_b x_{\rho b}t_{3b}n_b. \tag*{(49)}

and the pressure is

P=∑bPbkin−mσ22gσ2Φ2−κ6Φ3−λ24Φ4+mω22gω2W2+mρ22gρ2B2+ΛvW2B2+Pℓ,(50)P=\sum_bP_b^{\mathrm{kin}}-\frac{m_\sigma^2}{2g_\sigma^2}\Phi^2-\frac{\kappa}{6}\Phi^3-\frac{\lambda}{24}\Phi^4+\frac{m_\omega^2}{2g_\omega^2}W^2+\frac{m_\rho^2}{2g_\rho^2}B^2+\Lambda_vW^2B^2+P_\ell, \tag*{(50)}

where PℓP_\ell is the lepton pressure. For DDME2 the couplings depend on the baryon density, so the derivative of the grand potential with respect to the baryon densities acquires the rearrangement term

ΣR=∑b[∂ln⁡gω∂nBxωbWnb+∂ln⁡gρ∂nBxρb2t3bBnb−∂ln⁡gσ∂nBxσbΦnbs],(51)\Sigma_R=\sum_b\left[\frac{\partial\ln g_\omega}{\partial n_B}x_{\omega b}Wn_b+\frac{\partial\ln g_\rho}{\partial n_B}x_{\rho b}2t_{3b}Bn_b-\frac{\partial\ln g_\sigma}{\partial n_B}x_{\sigma b}\Phi n_b^s\right], \tag*{(51)}

which enters every UbU_b in Eq. (46). The DDME2 field equations and pressure are

Φ=gσ2mσ2∑bxσbnbs,W=gω2mω2∑bxωbnb,\Phi=\frac{g_\sigma^2}{m_\sigma^2}\sum_bx_{\sigma b}n_b^s,\qquad W=\frac{g_\omega^2}{m_\omega^2}\sum_bx_{\omega b}n_b,
B=gρ2mρ2∑bxρb2t3bnb,B=\frac{g_\rho^2}{m_\rho^2}\sum_bx_{\rho b}2t_{3b}n_b,
P=∑bPbkin−mσ22gσ2Φ2+mω22gω2W2+mρ22gρ2B2+nBΣR+Pℓ.(52)P=\sum_bP_b^{\mathrm{kin}}-\frac{m_\sigma^2}{2g_\sigma^2}\Phi^2+\frac{m_\omega^2}{2g_\omega^2}W^2+\frac{m_\rho^2}{2g_\rho^2}B^2+n_B\Sigma_R+P_\ell. \tag*{(52)}

The term containing the rearrangement self-energy is required for thermodynamic consistency: it makes the pressure obtained from the grand potential coincide with the pressure obtained from the derivative of the energy per baryon at fixed composition and entropy per baryon. In both models the energy density is

ϵ=−P+Ts+∑iμini.(53)\epsilon=-P+Ts+\sum_i\mu_in_i. \tag*{(53)}

Stability and consistency

(ii) Spinodal stability. We computed the curvature matrix of the free energy at fixed temperature in the variables baryon density and lepton density,

C=∂(μB,μL)∂(nB,nL),μL=μQ+μe.(54)C=\frac{\partial(\mu_B,\mu_L)}{\partial(n_B,n_L)},\qquad\mu_L=\mu_Q+\mu_e. \tag*{(54)}

It was positive definite between one and six times n0n_0, at 1 and 5 MeV5\ \mathrm{MeV}, for all four Δ\Delta sets; the smallest eigenvalue, 144 MeV fm3144\ \mathrm{MeV\ fm^3}, occurred for DD-Δ\Delta2 at 1.51.5 times n0n_0. This is consistent with Ref. [19], which found Δ\Delta-driven instabilities in DDME2 only when the scalar ratio exceeds the vector ratio by 0.150.15 to 0.20.2; in our sets the excess is at most 0.10.1. The numerical matrix was symmetric, as the Maxwell relation requires, to 7×10−37\times10^{-3} or better.

The energy density of Eq. (53), evaluated with an independently computed pressure, was consistent to a relative accuracy of 8×10−98\times10^{-9}.

Urca rates: phase space, nucleon width approximation and nucleon matrix element

Phase-space integral

For a decay of baryon 1 into baryon 2, a charged lepton and an antineutrino, with mean-field single-particle energies EiE_i (kinetic energy Ei∗E_i^* plus vector self-energy), the rate is

Γdec=∫d3p1(2π)32E1∗d3p2(2π)32E2∗×d3kl(2π)32Eld3kν(2π)32Eν×(2π)4δ4(P1−P2−kl−kν)×∑∣M∣2f1(1−f2)(1−fl),(55)\Gamma^{\mathrm{dec}}=\int\frac{d^3p_1}{(2\pi)^3 2E_1^*}\frac{d^3p_2}{(2\pi)^3 2E_2^*} \times\frac{d^3k_l}{(2\pi)^3 2E_l}\frac{d^3k_\nu}{(2\pi)^3 2E_\nu} \times(2\pi)^4\delta^4(P_1-P_2-k_l-k_\nu) \times\sum|\mathcal{M}|^2 f_1(1-f_2)(1-f_l), \tag*{(55)}

where the four-vectors PiP_i carry the canonical energies EiE_i and the momenta pip_i, and ∑∣M∣2\sum|\mathcal{M}|^2 denotes the sum over the spins of all four particles. No average over the initial spin is taken, because the occupation f1f_1 already counts each spin state. The spinor normalization to twice the effective mass makes the kinetic energies Ei∗E_i^* appear in the flux factors, while the canonical energies enter energy conservation, because the vector mean fields shift the energies but not the Dirac structure. The capture rate follows from Eq. (55) by replacing the statistical factor by (1−f1)f2fl(1-f_1)f_2f_l and the outgoing antineutrino by an incoming lepton and outgoing neutrino.

We reduced Eq. (55) exactly to five dimensions. The neutrino momentum integral removes the momentum delta function through

∫d3kν2Eν=∫d4kν δ(kν2)θ(kν0),(56)\int\frac{d^3k_\nu}{2E_\nu}=\int d^4k_\nu\,\delta(k_\nu^2)\theta(k_\nu^0), \tag*{(56)}

which turns energy conservation into

δ(Eν2−∣p1−p2−kl∣2),Eν≡E1−E2−El.(57)\delta(E_\nu^2-\lvert p_1-p_2-k_l\rvert^2),\qquad E_\nu\equiv E_1-E_2-E_l. \tag*{(57)}

We chose the momentum of baryon 1 along the zz axis, so that its solid angle gives 4π4\pi, put the momentum of baryon 2 in the xzxz plane at polar cosine z2z_2, so that its azimuth gives 2π2\pi, and gave the lepton momentum polar cosine zlz_l and azimuth ϕl\phi_l. With

d3p=pE∗dE∗dΩ(58)d^3p=pE^*dE^*d\Omega \tag*{(58)}

for each massive particle, and trading the energy of baryon 1 for the neutrino energy at fixed baryon-2 and lepton energies,

∣p1−p2−kl∣2=R(z2,zl)−S(z2,zl)cos⁡ϕl,R=p12+p22+kl2−2p1klzl−2p1p2z2+2p2klz2zl,S=2p2kl1−z221−zl2.(59)\begin{aligned} \lvert p_1-p_2-k_l\rvert^2&=R(z_2,z_l)-S(z_2,z_l)\cos\phi_l,\\ R&=p_1^2+p_2^2+k_l^2-2p_1k_lz_l-2p_1p_2z_2+2p_2k_lz_2z_l,\\ S&=2p_2k_l\sqrt{1-z_2^2}\sqrt{1-z_l^2}. \tag*{(59)} \end{aligned}

The azimuthal integral has two roots, each with Jacobian

∣Ssin⁡ϕl∣=S2−(R−Eν2)2.(60)\lvert S\sin\phi_l\rvert=\sqrt{S^2-(R-E_\nu^2)^2}. \tag*{(60)}

Collecting factors,

Γdec=1128π6∫dE2 dEl dEν dz2 dzl p1p2kl∑∣M∣2S2−(R−Eν2)2×f1(1−f2)(1−fl)∣E1=E2+El+Eν.(61)\Gamma^{\mathrm{dec}}=\frac{1}{128\pi^6}\int dE_2\,dE_l\,dE_\nu\,dz_2\,dz_l\, \frac{p_1p_2k_l\sum|\mathcal{M}|^2}{\sqrt{S^2-(R-E_\nu^2)^2}} \times f_1(1-f_2)(1-f_l)\bigg|_{E_1=E_2+E_l+E_\nu}. \tag*{(61)}

For fixed z2z_2 the radicand is a quadratic in zlz_l with two roots given in closed form by the kinematics, so the zlz_l integral of the polynomial numerator is evaluated exactly by Gauss–Chebyshev quadrature. The remaining four integrals were evaluated by scrambled Sobol quasi-Monte Carlo over energy windows around the Fermi energies; the window width, the number of points and the number of Chebyshev nodes are given in Table III. The allowed range of z2z_2 is the one for which a real azimuth exists. This reduction is the one of [7], extended here to unequal baryon masses, arbitrary vector potentials and spin-32\frac{3}{2} baryons, which enter only through the matrix element.

Nucleon width approximation

Above about 1 MeV the collisional broadening of the baryon quasiparticles opens the reactions below the direct-Urca threshold that are usually treated as modified Urca. We used the nucleon width approximation (NWA) of [9]: each baryon of the Urca reaction is given a Breit–Wigner spectral function in its mass,

Ab(m)=1πWb/2(m−mb∗)2+Wb2/4,(62)\mathcal{A}_b(m)=\frac{1}{\pi}\frac{W_b/2}{(m-m_b^*)^2+W_b^2/4}, \tag*{(62)}

and the rate is the spectral average

ΓNWA=∫dm1 dm2 A1(m1)A2(m2)Γ(m1,m2),(63)\Gamma_{\mathrm{NWA}}=\int dm_1\,dm_2\,\mathcal{A}_1(m_1)\mathcal{A}_2(m_2)\Gamma(m_1,m_2), \tag*{(63)}

with Γ(m1,m2)\Gamma(m_1,m_2) the rate of Eq. (61) evaluated with masses m1m_1 and m2m_2 in the dispersion relations and the matrix element. Following Refs. [9, 11], the nucleon width grows quadratically with temperature,

WN=T2TW,(64)W_N=\frac{T^2}{T_W}, \tag*{(64)}

with the scale TWT_W of Table III. For the Δ\Deltas we added the in-medium Δ→Nπ\Delta\to N\pi width of Sec. III C,

WΔ=WN+ΓΔmed,(65)W_\Delta=W_N+\Gamma_\Delta^{\mathrm{med}}, \tag*{(65)}

so that the collisional broadening of the Δ\Delta was set equal to that of the nucleon. Calculations of the in-medium Δ\Delta width concern Δ\Deltas excited above the nucleon Fermi sea, in nuclei and in hot hadronic matter, where the Δ→Nπ\Delta\to N\pi channel and its medium modification dominate [35]. We are not aware of a calculation of the collisional width of degenerate Δ\Deltas near their own Fermi surface. For any degenerate fermion this width vanishes at the Fermi surface as T2T^2, which is the form of Eq. (64); its coefficient for the Δ\Delta is not known, and in Figure 6 we test it by adding 10 MeV to WΔW_\Delta. We sampled Eq. (63) jointly with the kinematics through the substitution

m=m∗+12Wsinh⁡t,A(m) dm=dtπcosh⁡t.(66)m=m^*+\frac{1}{2}W\sinh t,\qquad\mathcal{A}(m)\,dm=\frac{dt}{\pi\cosh t}. \tag*{(66)}

mixed with a uniform component to cover the tails. The resulting six-dimensional quasi-Monte Carlo integral agreed with the nested evaluation of Eq. (63) to 0.015 dex.

Lepton tensor and nucleon matrix element

The effective charged-current interaction and the lepton current are

LW=−G2Jhadμlμ+h.c.,G=GFcos⁡θC,\mathcal{L}_{W}=-\frac{G}{\sqrt{2}}J^\mu_{\mathrm{had}}l_\mu+\mathrm{h.c.},\qquad G=G_{\mathrm{F}}\cos\theta_{\mathrm{C}},
lμ=uˉlγμ(1−γ5)vν,(67)l_\mu=\bar{u}_l\gamma_\mu(1-\gamma_5)v_\nu, \tag*{(67)}

with the constants of Table Table III. The lepton tensor, summed over spins, is

Lμν=Tr⁡[kγμ(1−γ5)k′γν(1−γ5)]=8[kμkν′+kνkμ′−gμνk⋅k′−iϵμναβkαk′β],(68)L_{\mu\nu}=\operatorname{Tr}\left[\cancel{k}\gamma_\mu(1-\gamma_5)\cancel{k}'\gamma_\nu(1-\gamma_5)\right] =8\left[k_\mu k'_\nu+k_\nu k'_\mu-g_{\mu\nu}k\cdot k'-i\epsilon_{\mu\nu\alpha\beta}k^\alpha k'^\beta\right], \tag*{(68)}

where kk and k′k' are the charged-lepton and neutrino momenta; the lepton mass drops out because of the chiral projectors.

Nucleons. With the nucleon current

Jμ=uˉpγμ(1−gAγ5)un,(69)J^\mu=\bar{u}_p\gamma^\mu(1-g_A\gamma_5)u_n, \tag*{(69)}

the standard trace gives

∑‾∣M∣2=32G2[(1+gA)2(pp⋅kl)(pn⋅kν)+(1−gA)2(pn⋅kl)(pp⋅kν)+(gA2−1)mn∗mp∗(kl⋅kν)],(70)\overline{\sum}|\mathcal{M}|^2=32G^2\left[(1+g_A)^2(p_p\cdot k_l)(p_n\cdot k_\nu)+(1-g_A)^2(p_n\cdot k_l)(p_p\cdot k_\nu)+(g_A^2-1)m_n^*m_p^*(k_l\cdot k_\nu)\right], \tag*{(70)}

which for nucleons at rest averages over lepton angles to

∑‾∣M∣2→32G2(1+3gA2)mn∗mp∗ElEν.(71)\overline{\sum}|\mathcal{M}|^2\to32G^2(1+3g_A^2)m_n^*m_p^*E_lE_\nu. \tag*{(71)}

Δ→Δ′\Delta\to\Delta' spin sum

With the current and isospin factor of Eq. (17), the spin-summed squared amplitude is

∑‾∣M∣2=G22∣⟨T+⟩∣2H,\overline{\sum}|\mathcal{M}|^2=\frac{G^2}{2}|\langle T_+\rangle|^2\mathcal{H},
H=gαα′gββ′Tr⁡[ΓμP1α′β′ΓˉνP2βα]Lμν,Γμ=γμ(1−gAΔγ5),Γˉν=γ0Γν†γ0,(72)\mathcal{H}=g_{\alpha\alpha'}g_{\beta\beta'}\operatorname{Tr}\left[\Gamma^\mu P_1^{\alpha'\beta'}\bar{\Gamma}^\nu P_2^{\beta\alpha}\right]L_{\mu\nu}, \qquad \Gamma^\mu=\gamma^\mu(1-g_{A\Delta}\gamma_5),\qquad \bar{\Gamma}^\nu=\gamma^0\Gamma^{\nu\dagger}\gamma^0, \tag*{(72)}

where the PiP_i are the RS spin sums of Eq. (9) for the two Δ\Deltas. In terms of the invariants

a=p1⋅p2,b=p1⋅kl,c=p1⋅kν,a=p_1\cdot p_2,\quad b=p_1\cdot k_l,\quad c=p_1\cdot k_\nu,
d=p2⋅kl,e=p2⋅kν,f=kl⋅kν,(73)d=p_2\cdot k_l,\quad e=p_2\cdot k_\nu,\quad f=k_l\cdot k_\nu, \tag*{(73)}

and the effective masses M1M_1, M2M_2 of the two Δ\Deltas, FORM gave

H=H0+gAΔH1+gAΔ2H2,H0=2569[−af+a2(be+cd)M12M22−a2fM1M2+bcM2M1+deM1M2]+6409(be+cd−fM1M2),H1=5129a2(cd−be)M12M22+2563(cd−be),H2=2569[−af+a2(be+cd)M12M22+a2fM1M2−bcM2M1−deM1M2]+6409(be+cd+fM1M2).(74)\begin{aligned} \mathcal{H}&=\mathcal{H}_0+g_{A\Delta}\mathcal{H}_1+g_{A\Delta}^2\mathcal{H}_2,\\ \mathcal{H}_0&=\frac{256}{9}\left[-af+\frac{a^2(be+cd)}{M_1^2M_2^2}-\frac{a^2f}{M_1M_2}+\frac{bcM_2}{M_1}+\frac{deM_1}{M_2}\right]+\frac{640}{9}(be+cd-fM_1M_2),\\ \mathcal{H}_1&=\frac{512}{9}\frac{a^2(cd-be)}{M_1^2M_2^2}+\frac{256}{3}(cd-be),\\ \mathcal{H}_2&=\frac{256}{9}\left[-af+\frac{a^2(be+cd)}{M_1^2M_2^2}+\frac{a^2f}{M_1M_2}-\frac{bcM_2}{M_1}-\frac{deM_1}{M_2}\right]+\frac{640}{9}(be+cd+fM_1M_2). \tag*{(74)} \end{aligned}

The term odd in the axial coupling is the vector–axial interference, which changes sign between decay and capture because the roles of the charged lepton and the neutrino are exchanged. For both Δ\Deltas at rest with equal masses, and with θ\theta the angle between the lepton momenta,

H→1289M2ElEν[9(1+cos⁡θ)+5gAΔ2(3−cos⁡θ)],(75)\mathcal{H}\to\frac{128}{9}M^2E_lE_\nu\left[9(1+\cos\theta)+5g_{A\Delta}^2(3-\cos\theta)\right], \tag*{(75)}

whose angular average gives Eq. (20).

One-variable limit, oscillation type, damping time and resonance

Equation (28) was evaluated at fixed temperature (isothermal) and at fixed entropy per baryon (adiabatic). In the adiabatic case every perturbed state was assigned the temperature at which the entropy per baryon equals its equilibrium value, found by root finding, and the rates were re-evaluated at that temperature. The adiabatic modulus exceeded the isothermal one at every grid point, by at most 12%, as required by thermodynamic stability.

The energy stored in a density oscillation and the mean dissipated power per unit volume are

E=12κ−1(δnnB)2,E˙=12ζω2(δnnB)2,\mathcal{E}=\frac{1}{2}\kappa^{-1}\left(\frac{\delta n}{n_B}\right)^2,\qquad \dot{\mathcal{E}}=\frac{1}{2}\zeta\omega^2\left(\frac{\delta n}{n_B}\right)^2,
κ−1=nB∂P∂nB∣Ye,Yμ,(76)\kappa^{-1}=n_B\left.\frac{\partial P}{\partial n_B}\right|_{Y_e,Y_\mu}, \tag*{(76)}

so the damping time of the amplitude is

τdamp=κ−1ω2ζ,(77)\tau_{\mathrm{damp}}=\frac{\kappa^{-1}}{\omega^2\zeta}, \tag*{(77)}

equivalent to the form in terms of the incompressibility used in Ref. [12]. The oscillation frequency is given in Table Table III.

For a single slow variable, Eq. (33) can be written in the resonance form

ζ=Cγγ2+ω2,C≡−∂P∂YeγBγ.(78)\zeta=\mathcal{C}\frac{\gamma}{\gamma^2+\omega^2},\qquad \mathcal{C}\equiv-\frac{\partial P}{\partial Y_e}\frac{\gamma_B}{\gamma}. \tag*{(78)}

To the extent that the net rate depends on the state only through the chemical imbalance, the ratio of the density response to the relaxation rate in the strength C\mathcal{C} is a ratio of thermodynamic derivatives, and C\mathcal{C} does not depend on the magnitude of the rates. At fixed frequency the bulk viscosity is then largest on resonance,

γ=ω:ζmax⁡=C2ω,τdampmin⁡=2κ−1ωC,(79)\gamma=\omega:\qquad\zeta_{\max}=\frac{\mathcal{C}}{2\omega},\qquad\tau_{\mathrm{damp}}^{\min}=\frac{2\kappa^{-1}}{\omega\mathcal{C}}, \tag*{(79)}

so the height of the peak is set by thermodynamics alone and the shortest damping time at a given frequency is inversely proportional to the frequency. We define the resonance frequency of a state as

fR=Gee2π.(80)f_R=\frac{G_{ee}}{2\pi}. \tag*{(80)}

In-medium width

Figure 7 shows the Fermi-surface averaged width of Eq. (22) discussed in Sec. VI B.

Comparison with Ref. AHS2023

To locate the difference between our nucleonic DDME2 resonance temperatures and those of Ref. [12], we evaluated the electron relaxation rate of Eq. (28) in matter without muons in four ways, at one, three and five times the DDME2 saturation density used there, which we denote n0′n'_0 (0.152 fm−30.152\ \mathrm{fm}^{-3}). This rate is the same quantity as the product of the rate coefficient and the susceptibility used in Ref. [12],

Gee=λeA1,∂δμe∂Ye∣nB=−nBA1.(81)\mathcal{G}_{ee}=\lambda_e A_1,\qquad\left.\frac{\partial\delta\mu_e}{\partial Y_e}\right|_{n_B}=-n_B A_1. \tag*{(81)}

The four variants were: A, the prescription of Ref. [12], namely direct Urca rates without width, only the first term of Eq. (70), and the derivative taken at the cold equilibrium, where the imbalance of Eq. (4) vanishes; B, our NWA rates with the full matrix element, at the cold equilibrium; C, our NWA rates at the thermal equilibrium of Eq. (7), which is what the main text uses; D, direct Urca without width at the thermal equilibrium.

Figure 8 shows the four relaxation rates in units of the angular frequency. Variant A crossed the oscillation frequency at 3.7, 4.4 and 4.2 MeV, which agrees to within 0.4 MeV with the statement of Ref. [12] that the crossing lies at 3–4 MeV below threshold. The remaining difference is of the size expected from the normalization of the coupling: Ref. [12] quotes the combination

Figure 8. Electron relaxation rate in units of the angular frequency at 1 kHz for nucleonic DDME2 matter without muons at (a) one, (b) three and (c) five times n0′n'_0 (0.152 fm−30.152\ \mathrm{fm}^{-3}). A (solid black): the prescription of Ref. [12] reproduced with our code (direct Urca without width, first term of the matrix element, linearized about the cold equilibrium). B (dashed blue): NWA rates with the full matrix element, same linearization point. C (dash-dotted orange): NWA linearized about the thermal equilibrium, as in the main text. D (dotted green): direct Urca linearized about the thermal equilibrium. Open circles mark the crossings with the dotted grey line, which move mainly with the linearization point, not with the rate model.

G=GFcos⁡θc(1+gA2)(82)G=G_F\cos\theta_c(1+g_A^2) \tag*{(82)}

in its reduced rate formula, and a change of the rate by the corresponding factor of 1.31 relative to the first-term normalization would lower the crossing by about 0.2 MeV. Our variant A also reproduced two qualitative statements of Ref. [12]: at the cold equilibrium the electron-capture rate exceeded the neutron-decay rate by more than three orders of magnitude, and at 5 MeV it depended only weakly on density (2.7, 1.6 and 2.5×10−13 MeV42.5\times10^{-13}\ \mathrm{MeV}^4 at the three densities).

Variant B differs from A only in the rate model and crossed at 3.3, 4.2 and 4.1 MeV, whereas variant C, which differs from B only in the linearization point, crossed at 3.6, 5.1 and 6.5 MeV. The rate model thus accounts for at most 0.4 MeV of the difference and the linearization point for 0.3–2.4 MeV, the latter growing with density. For direct Urca alone (variant D) the thermal equilibrium lies at an electron imbalance of about 20–160 MeV, and above n0′n'_0 the relaxation rate remains far below the oscillation frequency up to 8 MeV. Below threshold, direct Urca without collisional broadening therefore cannot describe equilibration about the thermal equilibrium, which is one reason the NWA is required.

At the cold equilibrium the net rate does not vanish, so the background about which variant A is linearized is not stationary; Ref. [7] showed that using this naive condition at 3 MeV overestimates electron capture by an order of magnitude. For this reason we used the thermal equilibrium throughout, following Refs. [7, 8, 11]. For the same model, Ref. [33] reported a 1 kHz resonant maximum at about 5 MeV, also with the cold equilibrium condition, and estimated that the finite-temperature correction would move the maximum by about 1 MeV. That maximum is the one of the total npeμnp e\mu bulk viscosity. For npeμnp e\mu matter Ref. [12] likewise places the 1 kHz minimum of the damping time at about 5 MeV, whereas its 3–4 MeV value is the crossing of the electron relaxation rate of npenpe matter below threshold. The two published values therefore describe different quantities and do not conflict. The shift we find between variants B and C at three times n0′n'_0, 0.9 MeV, agrees with the estimate of Ref. [33].

References

References

  1. [4]Z. Zhu, M. Chabanov, and R. O’Shaughnessy, Detectability of bulk viscosity effects on post-merger gravitational wave signals from binary neutron star mergers, arXiv e-prints (2026), arXiv:2609.06374.arxiv.org/abs/2609.06374
  2. [5]S. P. Harris, Bulk viscosity in dense nuclear matter, in Nuclear Theory in the Age of Multimessenger Astronomy (CRC Press, 2024), arXiv:2407.16157.DOI
  3. [6]M. G. Alford and S. P. Harris, Beta equilibrium in neutron star mergers, Phys. Rev. C 98, 065806 (2018), arXiv:1803.00662.DOI
  4. [7]M. G. Alford, A. Haber, S. P. Harris, and Z. Zhang, Beta equilibrium under neutron star merger conditions, Universe 7, 399 (2021), arXiv:2108.03324.DOI
  5. [8]M. G. Alford, A. Haber, and Z. Zhang, Isospin equilibration in neutron star mergers, Phys. Rev. C 109, 055803 (2024), arXiv:2306.06180.DOI
  6. [9]M. G. Alford, A. Haber, and Z. Zhang, Beyond modified Urca: The nucleon width approximation for flavor-changing processes in dense matter, Phys. Rev. C 110, L052801 (2024), arXiv:2406.13717.arxiv.org/abs/2406.13717
  7. [10]P. Tambe, D. Chatterjee, M. Alford, and A. Haber, Effect of magnetic fields on Urca rates in neutron star mergers, Phys. Rev. C 103, 045810 (2021), arXiv:2009.05181. Phys. Rev. C 111, 035809 (2025), arXiv:2409.09423.
  8. [11]P. Tambe, D. Chatterjee, M. Alford, and A. Haber, Thermal and magnetic effects on bulk viscosity in binary neutron star mergers, Phys. Rev. C 113, 055808 (2026), arXiv:2510.09104.DOI
  9. [12]M. Alford, A. Harutyunyan, and A. Sedrakian, Bulk viscosity from Urca processes: npeμ matter in the neutrino-transparent regime, Phys. Rev. D 108, 083019 (2023), arXiv:2306.13591.DOI
  10. [13]J. L. Hernández, C. Manuel, and L. Tolos, Impact of muons on the bulk viscosity of neutron star matter, Phys. Rev. D 114, 063045 (2026), arXiv:2603.23904.DOI
  11. [14]Y. Yang, M. Hippert, E. Speranza, and J. Noronha, Symmetry energy dependence of the bulk viscosity of nuclear matter, arXiv e-prints (2025), arXiv:2504.07805.arxiv.org/abs/2504.07805
  12. [15]A. Drago, A. Lavagno, G. Pagliara, and D. Pigato, Early appearance of isobars in neutron stars, Phys. Rev. C 90, 065809 (2014).DOI
  13. [16]B.-J. Cai, F. J. Fattoyev, B.-A. Li, and W. G. Newton, Critical density and impact of δ(1232) resonance formation in neutron stars, Phys. Rev. C 92, 015802 (2015).DOI
  14. [17]E. E. Kolomeitsev, K. A. Maslov, and D. N. Voskresensky, Delta isobars in relativistic mean-field models with σ-scaled hadron masses and couplings, Nucl. Phys. A 961, 106 (2017).DOI
  15. [18]J. J. Li, A. Sedrakian, and F. Weber, Competition between delta isobars and hyperons and properties of compact stars, Phys. Lett. B 783, 234 (2018).DOI
  16. [19]A. R. Raduta, δ-admixed neutron stars: Spinodal instabilities and Urca processes, Phys. Lett. B 814, 136070 (2021), arXiv:2101.03718.DOI
  17. [20]Y. N., M. Sinha, and V. Parmar, Neutrino emissivity in δ-direct-Urca processes inside neutron stars, Phys. Rev. Lett. (accepted) 10.1103/nhl7-9j9z (2026), arXiv:2610.05244.DOI
  18. [21]Y. N, M. Sinha, and V. Parmar, Neutron star cooling with δ direct Urca processes, arXiv e-prints (2026), arXiv:2610.11435.
  19. [22]M. G. Alford and A. Haber, Strangeness-changing rates and hyperonic bulk viscosity in neutron star mergers, arXiv:2009.05181.DOI
  20. [23]E. M. Kantor, M. E. Gusakov, and K. Y. Kraav, Hyperon bulk viscosity effects in neutron-star inspirals, Phys. Rev. D (accepted) (2026), arXiv:2608.28000.arxiv.org/abs/2608.28000
  21. [24]M. Storozhenko, J. Wu, A. Haber, E. R. Most, J. Noronha-Hostler, M. R. Pelicer, N. Cruz-Camacho, and V. Dexheimer, Strangeness transport in binary neutron star mergers, arXiv e-prints (2026), arXiv:2608.15527.
  22. [25]W. Rarita and J. Schwinger, On a theory of particles with half-integral spin, Phys. Rev. 60, 61 (1941).
  23. [26]M. G. Alford, L. Brodie, A. Haber, and I. Tews, Relativistic mean-field theories for neutron-star physics based on chiral effective field theory, Phys. Rev. C 106, 055804 (2022), arXiv:2205.10283.DOI
  24. [27]M. G. Alford, L. Brodie, A. Haber, and I. Tews, Tabulated equations of state from models informed by chiral effective field theory, Phys. Scripta 98, 125302 (2023), arXiv:2304.07836.DOI
  25. [28]G. A. Lalazissis, T. Nikšić, D. Vretenar, and P. Ring, New relativistic mean-field interaction with density-dependent meson-nucleon couplings, Phys. Rev. C 71, 024312 (2005).DOI
  26. [29]E. Hernández, J. Nieves, and M. Valverde, Weak pion production off the nucleon, Phys. Rev. D 76, 033005 (2007).DOI
  27. [30]F.-J. Jiang and B. C. Tiburzi, Chiral corrections and the axial charge of the delta (2008), arXiv:0803.3316.DOI
  28. [31]Y.-S. Jun, J.-M. Suh, and H.-C. Kim, Axial-vector form factors of the baryon decuplet with flavor SU(3) symmetry breaking (2020), arXiv:2005.06824.DOI
  29. [32]S. Navas et al. (Particle Data Group), Review of particle physics, Phys. Rev. D 110, 030001 (2024).
  30. [33]M. Alford, A. Harutyunyan, and A. Sedrakian, Bulk viscosity of relativistic npeμ matter in neutron-star mergers, Particles 5, 361 (2022), arXiv:2209.04717.DOI
  31. [34]J. M. Lattimer, C. J. Pethick, M. Prakash, and P. Haensel, Direct URCA process in neutron stars, Phys. Rev. Lett. 66, 2701 (1991).pubmed.ncbi.nlm.nih.gov/10043594
  32. [35]H. van Hees and R. Rapp, Δ(1232) and nucleon spectral functions in hot hadronic matter (2004), arXiv:nucl-th/0407050.DOI

Paper details

vlnc:W_nrkh9jcx
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