Nuclear Experiments and Theory Priors

Laboratory experiments and nuclear theory do not provide one universal “nuclear prior.” Each constrains a different observable, composition, temperature, momentum scale, and density interval; translating it to cold beta-equilibrated neutron-star matter requires a model with correlated uncertainty. The useful organizing principle is a density ladder: finite nuclei and parity-violating scattering anchor sub-saturation and near-saturation isovector physics, chiral effective field theory controls a finite moderate-density domain, heavy-ion collisions probe transient hot matter, and astrophysical stars test colder, more neutron-rich matter at higher density. This page defines how those inputs enter an EOS likelihood, with evidence cutoff 2026-08-18.

Required background. Equivalence principles and gravitational measurement supplies the discipline of separating a measured observable from the model used to infer a material parameter.

Helpful background. Stress–energy and conservation supplies pressure and energy density; the Einstein equation maps an EOS to stellar structure; relativistic matter and compact objects supplies dense-matter thermodynamics; gravitational radiation and waveforms supplies tidal evidence; and computational relativity supplies emulator and uncertainty validation.

Near-saturation parameters and their domain

Write the energy per baryon of uniform matter at baryon density $n$ and isospin asymmetry $\delta=(n_n-n_p)/n$ as

$$e(n,\delta)=e_{\rm sat}(n)+S(n)\delta^2+O(\delta^4).$$

Near symmetric-matter saturation $n_0$, let $x=(n-n_0)/(3n_0)$ and expand

$$S(n)=J+Lx+\frac12K_{\rm sym}x^2+\cdots, \qquad L=3n_0\left.\frac{dS}{dn}\right|_{n_0}.$$

The pressure is $P=n^2\,\partial e/\partial n$ at fixed composition. Symmetric matter has zero pressure at its saturation minimum, $e'_{\rm sat}(n_0)=0$. Keeping only the quadratic isospin term and setting $\delta=1$ therefore gives pure-neutron-matter pressure $P(n_0)\simeq n_0L/3$. The approximation lies in neglecting higher powers of $\delta$, not in the symmetric-matter pressure. This relation explains why a neutron-skin observable can inform pressure near $n_0$; it does not make the finite-nucleus map exact. Surface gradients, shell correlations, Coulomb effects, and the chosen energy-density functional mediate the translation to uniform matter.

Likewise, a Taylor expansion about $n_0$ is not a high-density EOS. Its truncation error grows with $|x|$, and coefficients inferred within one functional family can be strongly correlated. An analysis that carries only independent Gaussian priors on $J$, $L$, and $K_{\rm sym}$ may create matter that was never represented by the calibrated nuclear models.

A density-aware evidence map

Input	Direct observable	Principal domain	Translation that must be modeled
Nuclear masses and charge radii	Binding energies and electromagnetic radii	Finite nuclei, chiefly sub-saturation surface plus near-saturation bulk	Energy-density functional, pairing, deformation, shell and covariance model
PREX-II and CREX	Parity-violating asymmetry and weak form factor at finite momentum transfer	Neutron distributions in $^{208}\mathrm{Pb}$ and $^{48}\mathrm{Ca}$	Electroweak radiative corrections and nuclear structure from weak form factor to neutron skin and symmetry energy
Dipole polarizability and giant resonances	Electromagnetic response moments	Finite-nucleus isovector response	Response theory, correlations, surface symmetry energy, experimental strength outside the measured window
Chiral effective field theory	Many-body energy and pressure from an order-by-order interaction	Cold nucleonic matter below a declared EFT breakdown density	Low-energy constants, regulator, many-body truncation, EFT truncation and correlated density error
Heavy-ion flow and particle production	Momentum anisotropy, yields, spectra and transport observables	Transient hot, compressed, less neutron-rich matter	Transport code, in-medium cross sections, mean fields, cluster production, centrality and detector acceptance

PREX-II and CREX constrain complementary nuclei and should not be treated as interchangeable estimates of one $L$. Adhikari et al. 2021, preprint pp. 4–5, Eqs. (2)–(3) and Tables II–III, PDF reports the $^{208}\mathrm{Pb}$ parity-violating asymmetry and weak form factor; Adhikari et al. 2022, Eqs. (1)–(4), Tables I–II, pp. 042501-2–042501-4 does the corresponding $^{48}\mathrm{Ca}$ measurement. Their different nuclei and momentum transfers test nuclear models in complementary ways. Comparing separate and joint fits tests whether a functional can describe both nuclei. Mapping either skin directly to a stellar radius without exposing nuclear-model covariance understates model dependence.

Chiral EFT is a theory calculation, not a data point. Its prior must identify interaction order, cutoff or regulator, fitted low-energy constants, three-nucleon forces, many-body method, basis and convergence, and the truncation-error model across density. Drischler, Holt, and Wellenhofer 2021, §§ 2–5 explains why order-by-order and many-body uncertainties are correlated, and why the usable density range must be declared rather than extrapolated until causality fails.

Heavy-ion inference has a different composition and thermal history. Danielewicz, Lacey, and Lynch 2002, pp. 1592–1595 and figures 2–5 illustrates the forward route from transport calculations of flow to a pressure band. Updated use must revisit beam species and energy, centrality selection, event and track cuts, detector acceptance and calibration, transport implementation, in-medium interactions, and the observable response to the trial EOS. The band cannot be inserted as if it were a direct cold beta-equilibrium pressure measurement.

Likelihoods with correlated theory error

Let $y$ collect $N$ measured weak form factors, nuclear observables, or collision summaries, and let $f(\vartheta)$ be a forward nuclear or transport model with EOS parameters $\vartheta$. Define the residual $r=y-f(\vartheta)$ and total covariance $C=C_{\rm exp}+C_{\rm th}(\vartheta)$. For positive-definite $C$, a normalized Gaussian approximation has

$$\ln\mathcal L(y\mid\vartheta) =-\frac12r^{\mathsf T}C^{-1}r -\frac12\ln\det C-\frac{N}{2}\ln(2\pi).$$

$C_{\rm exp}$ includes statistical uncertainty and correlated calibration: beam polarization and detector response for parity-violating scattering, energy scale and acceptance for nuclear response, or centrality and tracking for heavy ions. $C_{\rm th}$ contains emulator error, EFT truncation, many-body approximation, functional discrepancy, or transport-model discrepancy. Treating calculations from several related interactions as independent samples is not a justified covariance model; shared calibration data and common operator choices make them correlated.

The determinant term matters when the theory covariance depends on $\vartheta$. Omitting it can reward a model merely for claiming large uncertainty. Non-Gaussian experimental likelihoods and bounded nuisance parameters should be retained when available rather than compressed into symmetric errors.

Priors also require physical and statistical separation. Chiral-EFT calculations may define a conditional process for neutron-matter pressure over their validity interval. A flexible high-density continuation then enforces thermodynamic stability and causality without pretending that EFT controls the stellar center. Nuclear skins update the nuclear-model parameters that connect finite and uniform matter; they do not act as independent likelihoods if those same nuclear calibration data were already used to construct the EFT-informed ensemble. Kumar et al. 2024, §§ 2.3, 2.5–2.7 and 7.2 surveys these experimental and theoretical domains and their distinct assumptions.

Versioned reproduction and stress tests

Each experimental dataset retains its primary publication, version, and covariance metadata.

The named experimental datasets are the final PREX-II $^{208}\mathrm{Pb}$ release in Adhikari et al. (2021), the final CREX $^{48}\mathrm{Ca}$ release in Adhikari et al. (2022), and the individually cited heavy-ion observable tables—not a merged secondary “nuclear constraint” band. A reproducible input package records their exact published data vectors and covariance or systematic breakdown, spectrometer and beam-polarization calibration, kinematic selections, electroweak corrections, and the finite-nucleus model used for the weak-form-factor map. For chiral EFT it records interaction files or identifiers, order, regulator, low-energy constants and covariance, many-body code and numerical settings, density grid, and truncation model. For heavy ions it records beam system and energy, centrality and particle selections, detector response, transport-code revision, initialization, cross sections, mean fields, random seeds, event count, and emulator validation.

This page documents that package; it does not claim a new nuclear, EFT, or collision reanalysis. Minimum stress tests are: remove PREX-II and CREX in turn; fit them jointly; change the finite-nucleus functional family; widen the EFT breakdown model; compare correlated with diagonal theory covariance; exclude the highest-density EFT points; and repeat heavy-ion inference with an independently implemented transport model. The posterior should be displayed as pressure versus density with prior and likelihood separated, so readers can see which interval each input actually changes.

Check your understanding

Using the parabolic approximation, estimate pure-neutron-matter pressure at saturation for $n_0=0.16\,\mathrm{fm}^{-3}$ and $L=60\,\mathrm{MeV}$. State why this estimate cannot be extended unchanged to $4n_0$.

Solution

At saturation,

$$P(n_0)\simeq\frac{n_0L}{3} =\frac{(0.16\,\mathrm{fm}^{-3})(60\,\mathrm{MeV})}{3} =3.2\,\mathrm{MeV\,fm}^{-3}.$$

The result uses a derivative evaluated at $n_0$, a quadratic expansion in isospin asymmetry, and the near-saturation expansion of $S(n)$. At $4n_0$, higher density coefficients, new degrees of freedom or phase transitions, and EFT or Taylor-series truncation dominate. Extrapolating the local slope $L$ as a constant high-density law would replace those unknowns with an unjustified linear model.

References

• Adhikari, D., et al. “Accurate Determination of the Neutron Skin Thickness of $^{208}\mathrm{Pb}$ through Parity-Violation in Electron Scattering.” Physical Review Letters 126 (2021): 172502. DOI.
• Adhikari, D., et al. “Precision Determination of the Neutral Weak Form Factor of $^{48}\mathrm{Ca}$.” Physical Review Letters 129 (2022): 042501. DOI.
• Danielewicz, Paweł, Roy Lacey, and William G. Lynch. “Determination of the Equation of State of Dense Matter.” Science 298 (2002): 1592–1596. DOI.
• Drischler, Christian, Jeremy W. Holt, and Corbinian Wellenhofer. “Chiral Effective Field Theory and the High-Density Nuclear Equation of State.” Annual Review of Nuclear and Particle Science 71 (2021): 403–432. DOI.
• Kumar, Rajesh, et al. (MUSES Collaboration). “Theoretical and Experimental Constraints for the Equation of State of Dense and Hot Matter.” Living Reviews in Relativity 27 (2024): 3. DOI.