May 2, 2025 0:28 cozma˙trautmann˙arxiv

NEUTRON STAR RADII FROM LABORATORY EXPERIMENTS

arXiv:2505.00390v1 [nucl-th] 1 May 2025

M. D. COZMA
IFIN-HH, Reactorului 30, 077125 Mǎgurele-Bucharest, Romania
[email protected]

W. TRAUTMANN
GSI Helmholtzzentrum für Schwerionenforschung GmbH
Planckstr. 1, D-64291 Darmstadt, Germany
[email protected]

Received 07 March 2025

Our present knowledge of the nuclear equation of state is briefly reviewed in this ar-
ticle intended for a wider readership. Particular emphasis is given to the asymmetric-
matter equation of state required for modeling neutron stars, neutron-star mergers, and
r-process nucleosynthesis. Recent analyses based on combining information obtained
from nuclear theory, heavy-ion collisions and astrophysical observations confine the ob-
tained radii of the canonical 1.4-solar-mass neutron star to values between 12 km and 13
km. The remaining uncertainty is primarily related to missing information in the density
interval between nuclear saturation density and about twice that value which, however,
is accessible with laboratory experiments.

Keywords: Nuclear equation of state; symmetry energy; heavy-ion collisions.

1. Introduction
The first observation of a merger of two neutron stars by the gravitational wave
detectors of the LIGO (Laser Interferometer Gravitational-Wave Observatory) and
Virgo Collaborations in August 2017, event GW170817, marked the beginning of
a new era of multi-messenger astronomy including gravitational wave detection.1–3
The analysis of the observed gravitational wave signal and of the associated electro-
magnetic emissions has provided us with the first clear demonstration that r-process
nucleosynthesis occurs in neutron star mergers4 and, at the same time, with a new
type of constraint for the nuclear equation of state. The tidal effects inscribed into
the inspiral part of the gravitational waves emitted before contact reflect the de-
formabilities of the stars and are related to their radii.5, 6 The radii, in turn, are
governed by the balance between the gravitational forces compressing the stars and
the internal pressure forces counteracting them.7, 8
The Neutron Star Interior Composition Explorer (NICER) was installed at the
International Space Station in the same year 2017. From precise measurements of
the pulse profiles of the two millisecond pulsars J0030+0451 and J0740+6620, two

1
May 2, 2025 0:28 cozma˙trautmann˙arxiv

2 M. D. Cozma and W. Trautmann

independent analysis groups based in Amsterdam and Maryland extracted values

for their radii with accuracies of approximately ±1 km.9–12 In particular, the simul-
taneous knowledge of the radius and mass of the heavy pulsar PSR J0740+6620
(PSR for Pulsating Source of Radio) serves as a strong constraint for models con-
structed to describe neutron-star matter.

Fig. 1. Schematic representation of the temperature versus asymmetry regimes associated with
the sources of information available for the study of the equation of state of nuclear matter and
of neutron-star matter. The quantities nn and np represent the neutron and proton densities,
respectively. The arrows are meant to indicate the dynamical processes of neutron star formation
and of neutron star (NS) collisions with other compact objects (from Ref.,13 © 2021 AAS reprinted
with permission).

The sources of information now available for confining the nuclear equation of
state as a function of temperature and neutron excess (equal to the isotopic asym-
metry δ = (nn − np )/(nn + np )) are distributed over all four corner sections of the
schematic Fig. 1 taken from the work of Raithel, Özel, and Psaltis.13 The figure
emphasizes the different regimes of temperature and asymmetry probed with ter-
restrial nuclear experiments and astrophysical observations. The abscissa, reaching
from symmetric atomic nuclei to asymmetric neutron stars, in addition, stretches
over the 55 orders of magnitude in mass that distinguish these objects under study.
It will be seen how amazingly well our techniques for describing nuclear forces can
be applied to the different regimes indicated in the figure.
Nuclear structure deals with atomic nuclei in their ground and excited states
and is thus limited to nuclear matter near and around the nuclear saturation
density ρ0 = 2.7 × 1014 g cm−3 (corresponding to a nucleon number density of
nsat = 0.16 fm−3 ; in the following, both n and ρ will be used as notation for densi-
ties given in numbers of nucleons per fm3 ). Because of the contributions of surface
effects, information provided by nuclear structure studies is most precise for densi-
ties around two thirds of that value. Nuclear reactions probe nuclear matter also at
May 2, 2025 0:28 cozma˙trautmann˙arxiv

Neutron Star Radii from Laboratory Experiments 3

the higher and lower densities that parts of the collision system may pass through
for a short time. In heavy-ion collisions, densities of several times the saturation
value may be reached in the central volume of the collision zone, depending on the
collision energy.14 To exploit these opportunities, measurable signatures have to
be found that characterize the modified matter at this particular instant. Neutron
star properties, on the other hand, are the result of the balance between the grav-
itational forces and the counteracting pressure forces generated by the interacting
constituents inside the star. The masses, radii or deformabilities of neutron stars
are the result of the combined effects at all densities present inside the star. The so-
called mass-radius relation is, therefore, a representation of the neutron-star-matter
equation of state that is equivalent to the usual form of the pressure as a function
of density.7, 15, 16
The arrows in Fig. 1 are meant to indicate the trajectories in the temperature
versus asymmetry plane of the dynamical processes of neutron star formation17, 18
following the explosion of massive stars and of mergers of neutron stars with other
neutron stars or black holes, as now observed with gravitational waves.19 To identify
their role as sites for nucleosynthesis, a preeminent goal of nuclear and astro-physics,
knowledge of the equation of state is required over virtually the full regime of
temperature and asymmetry covered in the figure. A framework for treating the
temperature dependence was proposed by Raithel et al.13
This article will focus on the cold matter equation of state. Ultimately, as
one hopes and believes, combining information from all sources with their rela-
tive weights will lead to a unique solution determined with the statistically robust
method of Bayesian inference.20 The individual ingredients will be introduced with
brief characterizations of the methodical steps. All four domains indicated in Fig. 1
contribute significantly and their mutual consistency is, in fact, quite impressive.
Improvements regarding the accuracy of this type of results are necessary and pos-
sible. As it will become evident, a major role can be played by new information
covering the density interval between once and twice the saturation value which is
accessible with laboratory experiments.

2. Neutron Stars and Neutron Star Mergers

An example of the equation of state of neutron-star matter as it can be obtained
from astrophysical observations alone is shown in Fig. 2. The figure is taken from
the work of Legred et al.21 which was completed after the results of the radius mea-
surement for J0740+6620, the presently heaviest pulsar (abbreviated from pulsating
radio source) with a precisely known mass, have become available.22, 23 It represents
the obtained equation of state in the form of the pressure versus density relation
and illustrates the information provided by the masses (labeled PSR in Fig. 2), the
radii, and the tidal deformability (GW) of the observed neutron stars. The magenta
contours located at densities around 1015 g/cm3 , roughly four times the saturation
value, indicate the 50% and 90% credible intervals of the central pressure-density
May 2, 2025 0:28 cozma˙trautmann˙arxiv

4 M. D. Cozma and W. Trautmann

1037
Prior
PSR + GW + J0030
1036 PSR + GW + J0030 + J0740
P (dyn/cm2)

1035

1034

1033

2ρnuc

6ρnuc
ρnuc
1032
1014 1015
3
ρ (g/cm )

Fig. 2. Pressure-density relation of neutron-star matter: the shaded regions enclose the 90% sym-
metric credible intervals. Magenta contours give the 50% and 90% level of the central pressure-
density posterior for the pulsar J0740+6620 inferred from all astrophysical data used for the
analysis (reprinted with permission from Ref.,21 copyright © 2021 by the American Physical
Society).

posterior for J0740+6620 inferred from all data used in the analysis. Note that 1035
dyn/cm2 corresponds to 1029 bar, evidently a high pressure.
The importance of the radius measurement for the heavy pulsar J0740+6620 is
highlighted by indicating, in addition, the contours obtained without it (shown in
green). The contours including it (in blue) provide a lower bound for the pressure at
densities near and above twice the saturation value that is significantly higher than
that obtained by only using the radius measured for the ≈1.4 solar-mass pulsar
J0030+0451.9, 10 At lower densities, the astrophysical observations do not contain
information exceeding that of the model-agnostic process used for constructing the
adopted prior distribution of trial solutions for the equation of state.21
The most precise masses of neutron stars in our galaxy have been obtained
for binary systems consisting of two neutron stars and are narrowly distributed
around a value of 1.4 solar masses.7 It has, therefore, become customary to express
the stiffness of a neutron-star-matter equation of state with its result for a 1.4
solar-mass neutron star. Larger radii indicate stiffer solutions. Legred et al. report
a radius R1.4 = 12.6 ± 1.1 km with the error representing the 90% confidence
limit.21 The value R1.4 = 11.9 ± 1.4 km resulting from the revised analysis of the
LIGO/Virgo Collaborations,5 mainly based on the observed tidal deformability, is
smaller due to the lower pressures indicated in Fig. 2 but consistent within errors.
Evidently, stiffer solutions were favored by the observation of neutron stars with
masses close to and above two solar masses. Their existence is known since 2010
when Demorest et al. reported a value of 1.97 ± 0.04 solar masses for J1614-2230
which represented the first case of a precisely measured mass of this magnitude.24
May 2, 2025 0:28 cozma˙trautmann˙arxiv

Neutron Star Radii from Laboratory Experiments 5

With the method of measuring Shapiro delay, used also for J0740+6620 with 2.08 ±
0.07 solar masses,22, 23 the gravitational effect of the compact companion star by
delaying the radio signals of the pulsar is determined. The mass of the companion
star deduced therefrom then permits resolving the ambiguity of the mass of the
neutron star in the orbital equations. For the third pulsar J0348+0432 in this
class with a mass of 2.01 ± 0.04 solar masses, the method of phase-resolved optical
spectroscopy applied to its white-dwarf companion was used.25
A radius R1.4 = 12.45 ± 0.65 km (68% credibility), similar to that of Legred
et al.21 but with a slightly smaller uncertainty, was obtained by Miller et al. by
including information related to lower densities.12 The authors argue that up to
about half the saturation density, roughly corresponding to the core-crust transition
density in neutron stars, the equation of state is well known.26–28 The contribution
to the radius uncertainty related to modeling the crust should not exceed 300 m (we
note here that the structure of the crust itself, containing extremely neutron-rich
nuclei29 and unusual forms of nuclear matter, and its possible excitations in the
form of vibrations are very interesting topics in their own right16, 30 ). For densities
above half the saturation value, several carefully constructed models were used in
this analysis which, however, have not been validated by laboratory measurements.
How strongly the conclusions depend on the class of equation-of-state models that
is employed and how these choices influence the results and their errors is taken
into account.12
The highest densities that may be reached in the core of heavy neutron stars
(Fig. 2) has attracted particular attention because of the possibility of a phase
transition to quark matter in the central core of the star. The density at which this
might occur and whether it is likely to be confirmed by observation is presently a
matter of debate. It is being argued that the maximum density may still be too
low31 but it is as well demonstrated that current observational data are compatible
with quarkyonic matter existing in the core of massive stars.32 (for further reading
on this subject see, e.g., Refs.33–36 ).

3. Precise Information from Nuclear Structure

The ground-state properties of atomic nuclei and the structure of their excitations
document the nuclear equation of state at zero or very low temperatures. Because
of the finite volumes of the surface regions, the average densities probed are below
the saturation value encountered in the center of heavy nuclei.37 Perot, Chamel
and Sourie, in their report on the Brussels-Montreal equations of state,38 cite five
references with explicitly given values for the density to which the value obtained for
the symmetry energy applies.39–43 These results are listed in Table 1 and displayed
in Fig. 3 reprinted from their paper.
The symmetry energy for nuclear matter, Esym (ρ), can be defined as the co-
efficient of the quadratic term in an expansion of the energy per particle in the
asymmetry δ = (ρn − ρp )/ρ, where ρn , ρp , and ρ represent the neutron, proton, and
May 2, 2025 0:28 cozma˙trautmann˙arxiv

6 M. D. Cozma and W. Trautmann

ρ (fm−3 ) Esym (ρ) (MeV) Observable Ref.

39
0.10 24.1 ± 0.8 giant dipole resonance
40
0.10 23.3 ± 0.6 giant quadrupole resonance
41
0.11 26.2 ± 0.7 Fermi-energy difference
42
0.11 26.7 ± 0.1 binding energy difference
43
0.10 25.5 ± 1.0 binding energy and radius

Table 1. Results for the symmetry energy Esym (ρ) with 1-σ uncertainties at the specified nucleon
densities ρ as obtained from the analysis of nuclear structure data using selected force models.

total densities, respectively,

E/A(ρ, δ) = E/A(ρ, δ = 0) + Esym (ρ)δ 2 + O(δ 4 ). (1)

The symmetry energy is a function of the density ρ and, in the quadratic approxi-
mation,44 equal to the difference between the energies of symmetric matter (δ = 0)
and neutron matter (δ = 1). The densities ρ listed in the first column of Table 1
are close to 2/3 of the saturation density and the values for Esym (ρ) scatter around
25 MeV. A most probable value for density ρ = 0.10 fm−3 may be calculated
by scaling the values listed in lines 3 and 4 of the table to this density with the
slope parameter L(0.11 fm−3 ) = 47±8 MeV (Ref.45 ), leading to 24.8±0.7 MeV and
25.2±0.3 MeV, respectively. A χ2 analysis yields Esym (0.10 fm−3 ) = 24.8±0.5 MeV,
documenting that the symmetry energy at this density is consistently determined
by nuclear structure, as obtained with these studies based on different observables
and analysis methods.
The fact that nuclear structure offers precise information for a narrow density
interval has been known for many years (see, e.g., Refs.46, 47 ) and has served as
motivation for studies aiming at other density regimes. Zhang and Chen report that
their result for the electric dipole polarizability of 208 Pb provides precise information
for densities up to about one half of the saturation value.48 It is shown in Fig. 3
together with results for the symmetry energy S(n) obtained from the observation of
isospin diffusion in heavy-ion reactions49 and from analyzing isotopic analog states50
that, in this form, have been presented in the review of Horowitz et al.51 in 2013
(here, the alternative notation S(n) is used for the symmetry energy Esym (ρ)). All
three colored bands cover the value 24.8 MeV at 0.10 fm−3 rather comfortably and,
together, constitute a consistent result for the symmetry energy at subsaturation
density.
The phenomenon of isospin diffusion49 during the fragmentation process follow-
ing energetic heavy-ion collisions gives access to the isospin equilibration governed
by the strength of the symmetry potentials as the fragments form and detouch
from the excited collision system. These processes proceed through non-equilibrium
stages at high temperatures and require analyses with transport models to deduce
properties of the acting forces. Isobaric analog states are characterized by mutual
May 2, 2025 0:28 cozma˙trautmann˙arxiv

Neutron Star Radii from Laboratory Experiments 7

40

35
BSk19
30 BSk22
BSk24 ■
◆
●
25 BSk25
S(n) [MeV]

✚
▼
BSk26
20

15

10

0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16
-3
n [fm ]

Fig. 3. Variation of the nuclear-matter symmetry energy S(n) as a function of the nucleon den-
sity n for the Brussels-Montreal functionals. The shaded areas are experimental constraints from
the electric dipole polarizability of 208 Pb (green, Ref.48 ), heavy-ion collisions49 (blue), and from
isobaric-analog states and neutron skins50 (purple). The symbols represent the results given in
line 1 to 5 of Table 1 in the sequence cross,39 triangle,40 circle,41 diamond,42 and square,43 re-
spectively (reprinted with permission from Ref.,38 copyright © 2019 by the American Physical
Society).

exchanges of neutrons and protons in specific orbits. Their energy differences thus,
in principle, consist of the differences of the symmetry term and of changes in the
Coulomb term which have to be evaluated. Danielewicz and Lee concluded that
the deduced constraints for the symmetry energy are valid for the density interval
0.04 ≤ ρ ≤ 0.13 fm−3 (Ref.50 ).
The energy-density functionals of the Brussels-Montreal series BSk19-BSk26,
developed for the use in studies of neutron star physics, are also displayed in Fig. 3.
Their parameters were primarily determined by fitting to the 2353 measured masses
of atomic nuclei with Z > 8 from the 2012 Atomic Mass Evaluation52 and adjusted
to reproduce individual neutron-matter equations of state.38 The five members of
the series differ slightly in their predictions for Esym (ρ0 ), there assuming values
between 29 and 32 MeV, but differ considerably in their predictions at higher
densities.53 At density ρ = 0.10 fm−3 , they agree with each other within half a
MeV and very precisely with the common value obtained above (Fig. 3).
The tendency of mean-field models to approach S(n) = 0 MeV at very low
density is no longer realistic once the clustering degree of freedom of symmetric
nuclear matter is considered.54–56 Heavy-ion reactions in the Fermi energy regime
proved to be useful for determining clustering probabilities and provided constraints
for equation-of-state models for low-density matter with particular applications in
supernova physics.57, 58 A detailed account of how to calculate the equation of state
in the different regions of a neutron star, and in particular for the low-density
May 2, 2025 0:28 cozma˙trautmann˙arxiv

8 M. D. Cozma and W. Trautmann

crust, is given by Perot, Chamel and Sourie in their report38 (for reviews see, e.g.,
Refs.59, 60 ).

4. The Symmetry Energy at Saturation

It has been noticed by Li and Han several years ago that the many results ob-
tained for the nuclear symmetry energy from terrestrial nuclear experiments and
astrophysical observations are amazingly compatible, even though individual re-
sults may be afflicted with considerable uncertainties.61 The compilation of Oertel
et al. based on 53 laboratory results and astrophysical observations has arrived at
weighted averages Esym (ρ0 ) = 31.7 ± 3.2 MeV for the symmetry energy at satura-
tion density ρ0 and L = 58.7 ± 28.1 MeV for the slope parameter L describing its
density dependence.62 The parameter L is the coefficient of the linear term of the
Taylor expansion of the symmetry energy with respect to the density relative to ρ0

   2
L ρ − ρ0 Ksym ρ − ρ0
Esym (ρ) = Esym (ρ0 ) + + + ... (2)
3 ρ0 18 ρ0
and Ksym is the still fairly unknown coefficient of the quadratic or curvature
term.8, 63, 64
New information has been provided by terrestrial experiments aiming at deter-
mining the nuclear equation of state at densities closer to the saturation density
encountered in the center of heavy nuclei. The thickness ∆Rnp of the neutron skin of
208
Pb nuclei is considered a particularly valuable observable because it reflects the
pressure experienced by neutrons in the interior of this neutron-rich nucleus which
directly relates to the slope parameter L of the symmetry energy.65 Brown has
demonstrated that the knowledge of the neutron skin of heavy nuclei permits the ex-
trapolation from ρ = 0.10 fm−3 to saturation density.43 Present experimental values
of ∆Rnp ≈ 0.16 fm for 208 Pb, obtained from measuring the electric dipole reponse
of heavy nuclei and reported to be accurate within ≈ ±0.03 fm,66, 67 correspond
to uncertainties of more than ±2 MeV for the extrapolated Esym (ρ0 ) ≈ 32 MeV
according to Brown.43 With the unexpectedly large central value of the thickness
of the neutron skin of 208 Pb, ∆Rnp = 0.28 ± 0.07 fm, reported by the PREX-2
Collaboration,68, 69 the same extrapolation will lead to values in the neighborhood
of 40 MeV for Esym (ρ0 ).
The PREX results were awaited with high expectation because the measured
asymmetry of parity-violating electron scattering on 208 Pb is interpreted without
the aid of nuclear theory.70 Reed et al. deduce Esym (ρ0 ) = 38.1 ± 4.7 MeV from
this measurement.69 A result closer to currently favored values, Esym (ρ0 ) = 33.0 ±
2.0 MeV, has been obtained by Essick et al.71 by combining astrophysical data with
PREX-2 and chiral effective field theory constraints (here and in the following, if
reported error margins are not very different in positive and negative directions,
only the larger value is quoted.)
May 2, 2025 0:28 cozma˙trautmann˙arxiv

Neutron Star Radii from Laboratory Experiments 9

A new situation arose when the small value ∆Rnp = 0.121 ± 0.026(exp) ±
0.024(model) for 48 Ca was reported by the CREX Collaboration using the same
method.72 Theory very consistently maintains that the two values should be corre-
lated, with each of the models predicting approximately equal skin thicknesses for
48
Ca and 208 Pb.72, 73 This puzzle is presently not resolved and a wide range of opin-
ions have been expressed. Within the quoted errors which are mainly statistical, the
two results were shown to be not necessarily contradicting each other or theoretical
expectations74–76 whereas, on the other hand, modified energy density functionals
are being proposed to more closely accomodate the constraints imposed by the
PREX and CREX measurements.77 New data with the high resolution predicted
for the proposed MREX experiment78 at Mainz will be very useful here.
Because of the valuable information carried by a precise value for the thick-
ness of the neutron skin of 208 Pb, alternative methods are of particular interest.
Hu et al. reported on ab-initio calculations leading to ∆Rnp = 0.17 ± 0.03 fm,79
whereas Giacalone et al. deduced a value ∆Rnp = 0.22 ± 0.06 fm from the analysis
of the 208 Pb+208 Pb collisions at ultrarelativistic energy performed at the Large
Hadron Collider.80 Both measurements are compatible with the earlier nuclear
structure data66, 67 but do not necessarily reduce the overall uncertainty. Smaller
errors within ±0.02 fm were predicted for measurements of neutron-removal cross
sections in high-energy nuclear collisions of 0.4 to 1 GeV/nucleon, provided that the
applied reaction theories can be sufficiently tested and constrained in a necessary
series of separate measurements.81, 82 For further reading on the nuclear symmetry
energy see, e.g., the review of Baldo and Burgio83 and the Topical Issue on Nuclear
Symmetry Energy published by The European Physical Journal A in 2014.84

5. Heavy-Ion Collisions
Heavy-ion collisions (HIC) at sufficiently high energy provide the possibility of ex-
ploring the density interval between once and twice the saturation value ρ0 , even
though only for short times of the order of 30 fm/c (10−22 s) or less.14, 85, 86 Infor-
mation regarding the nuclear equation of state is extracted from observed strengths
of collective flows and from cross sections for meson production. Collective particle
motion refers to anisotropies of particle yields caused by pressure gradients within
the compressed and heated matter whereas meson yields may reflect density de-
pendent collision rates. To the extent that these observations can be related to
early dynamical stages of the collision, they will provide access to pressure and
high density, the quantities of interest. Theoretical reaction models are required for
quantitative interpretations.
Heavy-ion collisions have been important for symmetric matter by establishing
the soft nature of the equation of state (nuclear incompressibility K0 ≈ 200 MeV)
and the momentum dependence of the acting nuclear forces.87, 88 The collective
flows of nucleons in 197 Au+197 Au collisions observed at the Bevalac accelerator at
Lawrence Berkeley National Laboratory and the Alternating Gradient Synchrotron
May 2, 2025 0:28 cozma˙trautmann˙arxiv

10 M. D. Cozma and W. Trautmann

Fig. 4. Pressure in symmetric nuclear matter as a function of density. The lines represent pre-
dictions for the Skxs15 (solid blue line), MSK1 (dashed red line), and SKX (dotted black line)
interactions. The analysis of Wang et al.93 favors the interval between Skxs15 and MSK1. The
shaded regions represent the results obtained by Danielewicz et al.86 and Le Fèvre et al.92 (from
Ref.93 ).

(AGS) at Brookhaven National Laboratory at energies up to 11.5 GeV/nucleon

were analyzed by Danielewicz et al.86 and the results converted into a pressure
band for symmetric nuclear matter at densities reaching up to 4.5ρ0 . The covered
large density interval made the data useful for comparisons with astrophysical re-
sults.89 A comprehensive data set of the collective observables directed and elliptic
flow has been presented by the Four-Pi (FOPI) Collaboration for a variety of col-
lision systems in the SIS energy range up to 1.5 GeV/nucleon.90, 91 The maximum
densities are between 2 and 3ρ0 . The recent analyses of the elliptic flow results
measured for 197 Au+197 Au collisions performed by Le Fèvre et al.92 and Wang et
al.93 match very well the AGS-deduced pressures at 2ρ0 = 0.32 fm−3 (Fig. 4).
The forces used are conventionally characterized by the incompressibility K0
describing the variation of the pressure for small amplitudes in density but contain
higher-order terms. In the case of Le Fèvre et al.,92 the reported incompressibility
is K0 = 190 ± 30 MeV. If the third-order term is assumed to be responsible for
the pressures of 7.5–13 MeVfm−3 obtained for 2ρ0 , its value is of the order of
Q0 ≈ −260 ± 60 MeV.
Information on the asymmetric matter equation of state is expected from differ-
ential observables that are specifically sensitive to the pressure and density differ-
ences experienced by protons and neutrons in the high-density zone of the collisions.
Different pressures acting on neutrons and protons in asymmetric matter at high
density may thus be recognized as isotopic variations of collective flows94–96 and
yields.97 Meson production at threshold in two-step processes via ∆ production
is favored at higher densities because of the higher collision rates among the con-
May 2, 2025 0:28 cozma˙trautmann˙arxiv

Neutron Star Radii from Laboratory Experiments 11

stituents in the high-density zone. To the extent that the isotopic composition there
is influenced by the asymmetric matter equation of state, differential effects can be
expected for observables as, e.g., kaon or pion ratios.98 The interpretation of mea-
sured data relies on transport theory accounting for the non-equilibrium dynamics
of the reaction.

Fig. 5. Left panel: charged-pion yield ratios from the SπRIT experiments104 as a function of
the combined N/Z of the studied Sn+Sn collision systems at 270 MeV/nucleon. The data are
displayed as red crosses with the size of the open symbols inside representing the experimental
errors. The results of the calculations are shown as colored boxes for the different codes identified
by their color in the right panel. The height of the boxes is given by the difference of predictions
for the soft and stiff choices of the symmetry energy. The dashed blue line is a power-law fit
with the function (N/Z)3.6 , whereas the dotted blue line represents (N/Z)2 of the system. Right
panel: double ratios (DR) of pion yields for 132 Sn+124 Sn and 108 Sn+112 Sn. The data and their
uncertainty are given by the red horizontal bar and the results of the transport models are shown
as colored boxes, in a similar way as in the left panel (please see Refs.104, 110 for descriptions of
the codes and references; reprinted from Ref.,110 with permission from Elsevier).

Pion production in heavy-ion collisions has been extensively investigated by the

FOPI Collaboration and a comprehensive set of production yields and yield ratios
of charged pions has been collected.99 More recent studies indicate, however, that
the interpretation of the integrated pion ratios cannot be considered as conclusive
at the present time (see, e.g., Refs.100–103 and references given therein). The SπRIT
Collaboration has more recently continued the investigation of pion production and
employed radioactive 108 Sn and 132 Sn beams for extending the isotopic composi-
tions of the studied Sn+Sn reaction systems on the neutron poor and neutron rich
side.104 Central collisions at an incident energy 270 MeV/nucleon close to the pion
production threshold were selected.
Also here the conclusion was reached that the differences of model predictions
for the momentum-integrated yield ratios and double ratios are larger than their
May 2, 2025 0:28 cozma˙trautmann˙arxiv

12 M. D. Cozma and W. Trautmann

Fig. 6. Single pion spectral ratios for 132 Sn+124 Sn (left) and 108 Sn+112 Sn (right) reactions. The
curves are dcQMD predictions using different L and ∆m∗np values listed in the right panel (adapted
from Ref.,109 copyright © 2021 by the American Physical Society).

sensitivities to the density dependence of the symmetry energy as illustrated in

Fig. 5, an extension of the figure reported in the original paper.104 The right panel,
in particular, presents the results of nine transport model predictions, finalized be-
fore the experimental results were known, for the double ratio (DR) of charged pions
from the neutron rich and neutron poor reaction systems. The height of the colored
symbols represents the difference of the predictions for the soft and stiff choices
of the symmetry energy. Deviations from the experimental results are in nearly all
cases larger than the calculated sensitivity to the stiffness of the symmetry energy.
Even in the case of the χBUU105 code which comes closest to the experimental re-
sult, the latter is not covered by the sensitivity interval of the model. Investigations
of possible sources of differences among transport models106 and between several
transport models and experiment have appeared subsequently.107, 108
The investigation of spectral instead of integrated yield ratios has shown, how-
ever, that the high-energy part of the spectra responds reliably to changes in the
assumptions for the isovector part of the forces used in the transport description,
even though it remains dependent on the chosen value for the effective mass splitting
in the asymmetric medium (Fig. 6).109 The obtained result 42 MeV < L < 117 MeV
is compatible with previous knowledge of the the density dependence of the symme-
try energy but carries a large uncertainty, to a large part also related to the fading
experimental accuracy at high transverse momentum. The description of pion pro-
duction near threshold touches at the limit of what is presently possible with the
semi-classical transport approach for describing the temporary evolution of the col-
lision including resonances. The Transport Model Evaluation Project (TMEP) is
May 2, 2025 0:28 cozma˙trautmann˙arxiv

Neutron Star Radii from Laboratory Experiments 13

a community-wide effort to enhance the reliability of model predictions. A review

containing short descriptions of the codes of the participating groups has appeared
recently.110

6. Elliptic Flow Ratios

Results obtained by studying collective elliptic flows in heavy-ion collisions were
included in the first study of the neutron-star-matter equation of state that com-
bined information from astrophysical observations of neutron stars, neutron-star
mergers, and from collisions of gold nuclei at relativistic energies with microscopic
nuclear theory calculations.111 In particular, information from the FOPI91, 92 and
the Asymmetric-Matter EOS (ASY-EOS) experimental campaigns112 were used to
obtain new constraints for neutron-rich matter at densities around saturation up
to 2ρ0 .
The FOPI results, together with the Bevalac and AGS data, were discussed in
the previous section (Fig. 4). The ASY-EOS experiments had been motivated by
the analysis of the earlier FOPI-LAND data113, 114 which demonstrated the use-
fulness of elliptic-flow ratios and elliptic-flow differences of neutrons versus protons
and neutrons versus charged particles for gaining information on the density depen-
dence of the nuclear symmetry energy.95, 96, 115, 116 Systematic effects influencing the
collective flows of neutrons and charged particles in similar ways should cancel in
the ratios or differences and thus enhance the relatively small asymmetry effects in
the moderately asymmetric 197 Au+197 Au collision system. It was also shown that
the model dependence of the obtained results is small.117, 118
Motivated by these findings, an attempt was made to improve the accuracy with
a new experiment conducted at the GSI laboratory in 2011 (ASY-EOS experiment
S394). The experimental setup112 followed the scheme developed for FOPI-LAND
by using the Large Area Neutron Detector (LAND119 ) placed at θlab ≈ 45◦ as the
main instrument for neutron and charged particle detection. The strength of elliptic
collective flows is quantitatively described with the second Fourier coefficient v2 of
the azimuthal distributions near mid-rapidity with respect to the event-by-event
determined reaction plane after corrections for the finite dispersion of the latter are
applied.112, 120
As in the FOPI-LAND experiment, the reaction 197 Au+197 Au was studied at
400 MeV/nucleon, close to the energy at which collective emissions perpendicular
to the reaction plane, also called squeeze-out, reach a maximum. Correspondingly,
the v2 parameter is negative and passes through a minimum.120, 121 Constraints
for the symmetry energy were determined by comparing the ratios of the elliptic
flows of neutrons and charged particles (ch), v2n /v2ch , with the corresponding model
predictions for soft and stiff assumptions (Fig. 7). For the analysis, a version of
the UrQMD transport model adapted to the study of intermediate energy heavy-
ion collisions was employed.122 Different options for the dependence on isospin
asymmetry were implemented, expressed as a power-law dependence of the potential
May 2, 2025 0:28 cozma˙trautmann˙arxiv

14 M. D. Cozma and W. Trautmann

1.3
1.2 γ =0.75±0.10
1.1
1
vn2/vch
2 0.9 stiff
0.8
0.7
0.6
0.5 soft
0.4
0.3 0.4 0.5 0.6 0.7
pt/A (GeV/c)

Fig. 7. Elliptic flow ratio of neutrons over all charged particles for central collisions of 197 Au+197 Au
at 400 MeV/nucleon as a function of the transverse momentum per nucleon pt /A. The black
squares represent the experimental data, the green triangles and purple circles represent the
UrQMD predictions for stiff (γ = 1.5) and soft (γ = 0.5) power-law exponents of the potential
term, respectively. The solid line is the result of a linear interpolation between the predictions,
weighted according to the experimental errors of the included four bins in pt /A, and leading to
the indicated γ = 0.75 ± 0.10. (reprinted with permission from Ref.,112 copyright © 2016 by the
American Physical Society).

part of the symmetry energy on the nuclear density ρ according to

pot kin
Esym = Esym + Esym = 22 MeV(ρ/ρ0 )γ + 12 MeV(ρ/ρ0 )2/3 . (3)

Here γ = 0.5 and γ = 1.5 correspond to a soft and a stiff density dependence, respec-
tively, as shown in Fig. 7. For the symmetry energy at saturation, the parametriza-
tion assumes Esym (ρ0 ) = 34 MeV.
Applying the necessary corrections and their uncertainties converts the fit result
of the acceptance-integrated elliptic-flow ratio (Fig. 7) into a power-law coefficient
γ = 0.72 ± 0.19. This is the result displayed as a red band in Fig. 8 as a function of
the reduced density ρ/ρ0 . It has a considerably smaller uncertainty than the earlier
FOPI-LAND result shown in yellow. The chosen parametrization appears compat-
ible with the low-density behavior of the symmetry energy from Refs.,42, 43, 49, 50
discussed in Section 3 and included in the figure. The slope parameter describing
the variation of the symmetry energy with density at saturation is L = 72±13 MeV.
The corresponding curvature term in the Taylor expansion with respect to density
is in the range Ksym = −70 to -40 MeV.
The range of densities probed with the neutron versus charged particle elliptic-
flow ratio was explored with the Tübingen version of the QMD model (TüQMD,
Ref.117 ) by studying the effect of systematically modifying the asymmetry term at
different density intervals in a series of test calculations. It was found that, at the
given energy of 400 MeV/nucleon, the sensitivity curve is centered at saturation
density but forms a broad distribution from low densities up to about 2.5 times
May 2, 2025 0:28 cozma˙trautmann˙arxiv

Neutron Star Radii from Laboratory Experiments 15

80
Brown n/ch flow
Zhang
(MeV) 60 HIC Sn+Sn
IAS
FOPI-LAND
ASY-EOS
40
Esym

20

0
0 0.5 1 1.5 2
ρ/ρ0

Fig. 8. Constraints deduced for the density dependence of the symmetry energy from the ASY-EOS
data112 in comparison with the FOPI-LAND result of Ref.95 as a function of the reduced density
ρ/ρ0 . The low-density results of Refs.42, 43, 49, 50 as reported in Ref.51 and included in Fig. 3 are
given by the symbols, the grey area (HIC), and the dashed contour (IAS). For clarity, the FOPI-
LAND and ASY-EOS results are not displayed in the interval 0.3 < ρ/ρ0 < 1.0 (reprinted with
permission from Ref.,112 copyright © 2016 by the American Physical Society).

the saturation value.112 The same TüQMD transport model was also used for a
comprehensive study of the parameter dependence of the calculated predictions.117

7. Interpretation of Flow Data

The nature and quality of the measured elliptic flow ratios (Fig. 7) does not permit
fits with more than one free parameter which, according to Eq. (3), is the power law
parameter γ. The value Esym (ρ0 ) = 34 MeV and the power-law functional form of
the potential term were assumptions and kept fixed. A choice Esym (ρ0 ) = 31 MeV,
close to the lower end of the presently favored interval of Esym (ρ0 ), lowers the slope
parameter to L = 61 ± 9 MeV but the functional form of a power-law dependence
on density of the potential term remains as an assumption.112
Further analyses of the flow data appeared, some of them addressing the short-
comings just mentioned. The work of Cozma63 started from the so-called MDI force
(MDI stands for momentum dependent interaction) originally designed to produce
a fixed Esym (ρ0 ) = 31.6 MeV and containing a parameter x which controls the den-
sity dependence of the symmetry energy.8, 123 By adding a density dependent term,
a new force MDI2 with an additional parameter y was generated which, together
with the original x, permitted independent choices for the slope and curvature pa-
rameters L and Ksym of Eq. (2). By lifting the correlation of these two parameters,
May 2, 2025 0:28 cozma˙trautmann˙arxiv

16 M. D. Cozma and W. Trautmann

present in most models, uncertainties in the short-range three-body force are ad-
mitted and thereby taken into account.63 The second modification concerned the
density at which the symmetry energy assumes a predetermined value. It was moved
from the saturation density to the lower value 0.10 fm−3 at which uncertainties are
small (cf. Section 3), and the value Esym (0.10 fm−3 ) = 25.5 MeV obtained by
Brown43 was adopted.

1.5

1.25

1.0
p
v2 /v2

0.75
n

vac cs + sPBA
K0=210 MeV
K0=245 MeV
0.5 K0=280 MeV
K0=245 MeV
vac cs + gPBA
0.25 med cs + sPBA

0.0
-2 -1 0 1 2
x
Fig. 9. Neutron-to-proton elliptic flow ratio as a function of the stiffness parameter x in comparison
with the FOPI-LAND data. Experimental results as reported in Refs.116, 117 are indicated by the
dashed and full horizontal lines, respectively, with only the error band corresponding to the latter
data set being displayed (hashed). See text for explanations of the symbols and connecting lines
(from Ref.,63 reprinted with permission from Springer Science+Business Media).

The TüQMD, equipped with the MDI2 force and supplemented by a phase-space
coalescence model adjusted to qualitatively describe FOPI experimental multiplici-
ties of free nucleons and light clusters,90 was used to compare the theoretical predic-
tions with the FOPI-LAND neutron-to-proton and ASY-EOS neutron-to-charged-
particles elliptic flow ratios. Additional corrections that were applied are motivated
and described in the paper.63 The final result presented there is L = 85 ± 22(exp) ±
20(th) ± 12(sys) MeV and Ksym = 96 ± 315(exp) ± 170(th) ± 166(sys) MeV.63 The
systematic errors listed separately take, e.g., into account that the experimental
light-cluster-to-proton multiplicity ratios were underestimated by the calculations.
It was the achievement made in this work to obtain a solution up to the quadratic
term in Eq. (2) without assumptions, however at the cost of a larger uncertainty.
The largest contribution of the error is experimental and, as it turns out, mainly
May 2, 2025 0:28 cozma˙trautmann˙arxiv

Neutron Star Radii from Laboratory Experiments 17

related to the neutron-proton flow ratio presently only available from the FOPI-
LAND data95 and thus affected by the limited statistics collected in that experi-
ment. This can be improved with new measurements, so that much more precise
values for the slope and curvature terms can be expected in the future. The example
depicted in Fig. 9 illustrates this point, there for calculations with the original MDI
force. It shows that various choices made for the symmetric-matter incompressibility
K0 (colored symbols triangle, circle and square), the cross section parametrizations
vac cs (vacuum cross sections) or med cs (in-medium cross sections), or the algo-
rithms simulating Pauli blocking (sPBA or gPBA), which all affect the calculated
individual flows, have only small impact on flow ratios. Smaller experimental errors
would considerably improve the constraint on x and thus on the density dependence
of the symmetry energy.
Very recent work with an improved version of the Tübingen QMD model (dc-
QMD124 ) has particularly addressed open questions regarding the in-medium elas-
tic nucleon-nucleon cross-sections, accounting for their dependence on density and
isospin asymmetry, and the momentum dependence of the interaction adjusted by
means of fixing nucleon effective masses at saturation density. The study uses only
stopping and flow data of the FOPI Collaboration90, 91, 125, 126 without making use
of the differential observables neutron versus charged particles flow ratio or charged-
pion ratio. It builds on previous efforts demonstrating the existence of a hierarchy
of in-medium effects on nucleonic observables,86 the largest being due to in-medium
modifications of cross-sections, followed by those due to the momentum dependence
of the interaction and the density dependence of the equation of state.

Fig. 10. Pressure of cold neutron-star matter (asymmetry δ = 0.93) as a function of the baryon
number density as obtained in Ref.124 with the dcQMD model (red) in comparison with the results
of Legred et al. (blue),21 Huth et al. (magenta),111 and Tsang et al. (green)132 (adapted from
Ref.,124 copyright © 2024 by the American Physical Society).
May 2, 2025 0:28 cozma˙trautmann˙arxiv

18 M. D. Cozma and W. Trautmann

It was recognized in this study that the dependence of in-medium cross-sections

on isospin asymmetry has a non-negligible impact and, for that reason, comparisons
with experimental data for systems with a wide range of neutron-to-proton ratios
were crucial. Incident energies were restricted to Elab ≤ 800 MeV/nucleon to reduce
uncertainties related to resonance production and propagation. As in the previous
analysis,63 the symmetry energy at two thirds saturation density is kept fixed as
Esym (0.10 fm−3 ) = 25.5 MeV but, to reduce the number of free parameters, Ksym
and higher order terms were taken as correlated with L. The obtained pressure
versus density contour of cold neutron-star matter with 95% credibility is given in
Fig. 10 (in red) and compared to three combined analyses to be introduced in the
section below. It represents a rather narrow constraint for densities up to around
the saturation value. Besides the result K0 = 230 ± 11 MeV for the compression
modulus of symmetric matter, a constraint L = 63 ± 13 MeV for the slope param-
eter describing the density dependence of the symmetry energy is obtained. As an
important result, the study documents that valuable and consistent information is
contained in the isotopic dependences of charged particle cross sections and flow
data. Their sensitivity to density is found to be strongest near and below saturation
but to also extend beyond 2ρ0 .

8. Combined Analyses
Results that combine the information from astrophysical observations with nuclear
physics results including heavy-ion collisions have been reported recently. Bayesian
inference has become the standard framework for considering the available sources
of information with their proper statistical weights. In each step of a sequential
procedure, new information is used to inform the so far obtained distribution of
candidate equations of state, the prior of that step, with the new input. The result-
ing posterior distribution then serves as the new prior in the following step.
The procedure starts with an initial prior distribution which has to be carefully
chosen because it affects the final outcome. So-called agnostic priors are obtained by
choosing large intervals with wide open borders for the parameter set of the chosen
equation-of-state model. The term metamodeling127 applies to models based on
the Taylor expansion given in Eq. (2). A distribution of candidate equations of
state is generated by systematically varying the expansion coefficients. It remains a
challenge to keep the computing demands within realizable limits because, for each
of the candidates, the probability of their reproducing the set of observations will
have to be calculated.
Chiral effective field theory (χEFT) is an alternative choice for generating a
prior distribution, based on the expectation that it represents our knowledge of
the nuclear forces at low energies.128 χEFT is a systematic expansion in powers
of a typical momentum scale p over the χEFT breakdown scale Λb . For infinite
matter, p is of the order of the nucleon Fermi momentum pF . Provided pF < Λb ,
χEFT calculations of strongly interacting matter can, in principle, be improved to
May 2, 2025 0:28 cozma˙trautmann˙arxiv

Neutron Star Radii from Laboratory Experiments 19

any desired accuracy. The errors at a given order are calculated by estimating the
magnitude of the terms of the next higher order. In this way, a prior distribution
with probabilities assigned to each of its members can be obtained.

2
10 Prior
pressure P [Mev fm−3]

Astro+HIC

1
10

0
10

0.5 1.0 1.5 2.0 2.5 3.0

number density n [nsat]

Fig. 11. Constraints on the equation of state of neutron-star matter represented as the evolution of
the pressure as a function of baryon number density when combining data from multi-messenger
neutron-star observations (Astro) and from heavy-ion collisions (HIC). The shading corresponds
to the 95% and 68% credible intervals. The dashed lines represent the 95% credible interval of the
prior generated with χEFT and the requirement that a maximum neutron-star mass of at least
1.9 solar masses is supported by all candidate equations of state (adapted from Ref.111 ).

In the work of Huth et al.,111 chiral effective field theory is used to cover the den-
sity interval between the crust-core transition at about 0.5nsat and 1.5nsat which is
argued to be safely below the breakdown limit. The extrapolation of the prior distri-
bution to higher densities is achieved with parameter-free extension schemes,129, 130
constrained by requiring that the speed of sound cs obeys causality (cs < c with
c denoting the speed of light) and guarantees stability (cs ≥ 0 with c2s = ∂p/∂ϵ,
i.e. the increase of the pressure with energy density). With the additional require-
ment that neutron stars with at least 1.9 solar masses are supported, a radius
R1.4 = 11.96 ± 1.18 km (95%) for a 1.4 solar-mass neutron star is predicted by the
prior distribution.
Using Bayesian inference, the results available for the masses and radii of
neutron stars, results obtained from the observation of the neutron star mergers
GW170817 and GW190425 and from the kilonova AT2017gfo are combined with
the information obtained from heavy-ion collisions (Figs. 4 and 8). Partly new soft-
ware is used to deduce the information on the equation of state contained in the
gravitational wave signal of GW170817 and its electromagnetic counterpart. The
density distributions of the sensitivity of the elliptic flow results and the dependence
of the ASY-EOS result on the assumed Esym (ρ0 ) were taken into account.
The resulting equation of state given in the form of pressure versus density is
May 2, 2025 0:28 cozma˙trautmann˙arxiv

20 M. D. Cozma and W. Trautmann

displayed in Fig. 11. The distribution is significantly narrower than that obtained
by only using astrophysical observations (Fig. 2), in particular at densities below
2nsat . The final result for the radius of a 1.4 solar-mass neutron star is R1.4 =
12.01±0.78 km (95% credibility). The importance of the choice of the prior becomes
evident in the result R1.4 = 12.56 ± 1.07 km obtained with a lower breakdown
density of 1.0nsat . The uncertainty is increased as expected but the value for the
radius is larger by ≈ 500 m if the χEFT prediction for the interval 1.0 to 1.5nsat is
ignored. It is also apparent in Fig. 11 that the final posterior distribution touches
the upper limit of the prior distribution in this interval, indicating that information
provided by astrophysical observations and heavy-ion collisions at supra-saturation
densities deviates from the χEFT prediction as used in this work. As noted earlier
by Essick et al., NICER observations suggest that the equation of state stiffens
relative to χEFT predictions at or slightly above nuclear saturation density.131
An alternative approach was presented by Tsang et al.132 The method of meta-
modeling is used for generating the prior distribution of equations of state. For
constraining the parameters, the χEFT of Drischler et al.133 and a selected list of
laboratory data is employed, including the results obtained with PREX-2,68 the
SπRIT data presented by Estee et al.,109 and the ASY-EOS elliptic-flow result.112
Astrophysical results from NICER and LIGO/Virgo are then used to inform the so
obtained prior distribution. The radius R1.4 = 12.9 ± 0.5 km (68%) calculated from
the resulting posterior appears at the upper end of values obtained from recent
combined analyses. Upper limits reported earlier, after the discovery of ≈ 2 solar-
mass neutron stars, reach further up to nearly 14 km.129 There is a clear correlation
between the apparently large radius R1.4 and the laboratory constraints used for
generating the prior which all favor larger slope parameters L.
Figure 10 presents a comparison of the pressure versus density relations obtained
in three studies, Legred et al.,21 Huth et al.,111 and Tsang et al.132 Two of them
were already shown and introduced in Figs. 2 and 11, and their agreement at high
density and the effect of additional constraints at lower densities were addressed.
We observe that the contours of the two combined studies111, 132 (magenta and
green) are practically mutually exclusive up to 1.5 times saturation density, even
at the 95% confidence limits displayed in the figure. The different constructions of
the priors and the choices of additional information relating to lower densities are
responsible for this significant effect. This observation was already reported in the
White Paper of Sorensenet al.134 in which it was also pointed out that the χEFT
result of Drischler et al.133 stiffens with density and, at 1.5nsat , is considerably
higher than the combined result of Huth et al. who, for generating the prior dis-
tribution, used the softer χEFT results of Lynn et al.135 up to density 1.5nsat . It
is evident that χEFT predictions at these densities are not yet sufficiently robust
(see, e.g., Refs.128, 136, 137 ) and that the mode of using them has a strong influence
on the finally obtained posterior distributions of equations of state.
The work of Huth et al. has been extended by Koehn et al. who have added more
data sets of astrophysical observations and laboratory measurements.138 Altogether,
May 2, 2025 0:28 cozma˙trautmann˙arxiv

Neutron Star Radii from Laboratory Experiments 21

20 different types of input are used for informing a prior constructed with the
technique of metamodeling. χEFT is used as information for densities extending
up to between 1 and 2nsat . The outcome R1.4 = 12.20 ± 0.50 km (95%) is slightly
larger than that of Huth et al.111 and may be considered as resulting from the
presently most complete analysis. Each of the 20 sources of input information is
critically discussed in the report.
The analysis of Pang et al. does not make use of information from heavy-ion
collisions.139 With a prior obtained with χEFT and exclusively astrophysical inputs,
the value R1.4 = 11.98 ± 0.40 km at 90% credible interval is obtained. It carries a
comparatively small error and, at present, constitutes the lower limit of the interval
spanned by combined analyses. It is evident that ignoring information coming from
heavy-ion collisions and relying on χEFT up to 2nsat leads to softer results for the
neutron-star-matter equation of state.

9. Discussion and Outlook

When this article was near completion, several interesting new reports have ap-
peared. The analysis groups within the NICER Collaboration published improved
results for the stars under observation which had already been announced in their
earlier publications.11, 12 Salmi et al.140 and Dittmann et al.141 were able to include
additional observational data in their analyses and profited from improvements in
the light-curve analysis of the heaviest known neutron star J0740+6620. Their re-
sults for the equatorial radius of this star, R = 12.49+1.28 +1.49
−0.88 km and R = 12.76−1.02 km
(both 68%), respectively, are fully consistent with each other and with the expec-
tation of smaller errors following from additional observational input. They will
serve as strong anchor points for the nuclear equation of state at high density, in
particular also excluding significant hyperon components in neutron stars.142–145
Most recently, a very interesting observation was published by the NICER Col-
laboration,146 reporting a radius R = 11.36+0.95 −0.63 km (68%) for the millisecond
pulsar J0437-4715 whose mass M = 1.418 ± 0.044 is known with high precision.147
PSR J0437-4715 is the brightest pulsar on the NICER list of candidates for pulse
profile modeling and with a distance of about 500 light years much closer than
PSR J0030+0451 (1060 light years), the first star whose radius was reported by the
NICER Collaboration.9, 10, 148 With its mass of 1.4 solar masses, PSR J0437-4715
qualifies as a canonical neutron star and its radius may thus serve as a direct mea-
surement to be compared to existing analysis results for this quantity. The central
value of 11.4 km is lower than the interval 12 to 13 km appearing as favored by
the combined analyses discussed in the previous section. With its error of nearly
1 km (68%) it extends, however, deeply into this region. In a first combined analy-
sis149 including the new results,140, 141, 146 radii R1.4 between 12.01 and 12.28 km,
depending on the chosen extension model, were obtained with errors of ≈ 0.7 km
(95%). In Refs.150, 151 it is found that the new NICER measurement, if included,
leads to an average radius reduction of ≈ 0.1 km to 0.2 km in the posterior of the
May 2, 2025 0:28 cozma˙trautmann˙arxiv

22 M. D. Cozma and W. Trautmann

BHF N3LO∆/N2LO∆
SCGF N3LO

CC N2LOsat
MBPT N3LO
SCGF N3LO
BHF N3LO∆/N2LO∆

Fig. 12. Constraints obtained in Ref.124 (dcQMD FOPI, red lines and contours) for the density
dependence of the equation of state of symmetric nuclear matter (left panel) and for the symmetry
energy (right panel) in comparison with other analyses of data from heavy-ion collisions and ab-
initio calculations (black lines); see text for more explanations and Ref.124 for references to the
selected set of theoretical predictions (reprinted with permission from Ref.,124 copyright © 2024
by the American Physical Society).

obtained mass-radius relationships. The reported radius of J0437-4715 is thus not

in contradiction with previous combined analyses and their results.
The densities in the interior of the canonical 1.4-solar-mass neutron stars extend
to ≈ 3 − 4nsat 5 and their radii are thus dominated by the pressure at lower densities
near and around 2nsat . This is the upper end of the density range accessible with
heavy-ion collisions which emphasizes their role in clarifying the pressure versus
density relation in this important interval of 1 − 2nsat (Fig. 10). Transport models
are indispensable for the theoretical description of heavy-ion collisions and for ex-
tracting information from their high-density stages; continuously improving them
and controlling their consistency is thus an important task.
As an illustration of the current situation, we present in Fig. 12 the results re-
ported in Ref.124 in comparison with selected previous work. For symmetric nuclear
matter (left panel), the energy per nucleon deduced from the FOPI stopping and
flow data and characterized by K0 = 230 ± 11 MeV (in red) touches the result
of Le Fèvre et al.92 (IQMD FOPI, K0 = 190 ± 30 MeV) from above and is in
good agreement with the values K0 = 220 ± 40 MeV of Wang et al. (Fig. 4) and
K0 = 200 ± 25 MeV as adopted in the combined analysis of Huth et al.111 (all
errors represent 68% confidence limits). The symmetric-matter incompressibility of
this magnitude seems fairly robust and is found to be favored also by recent similar
studies152, 153 based on a wider selection of data from heavy-ion collisions, including
the high-statistics data of the HADES Collaboration.154, 155
The obtained density dependence of the symmetry energy (red line and contours
in the right panel) is of particular interest because it reflects the information con-
tained in the isotopic dependences of light charged-particle cross sections and flows
as discussed in Section 7. With increasing density, it follows closely the original
ASY-EOS result112 as well as the central value obtained in the subsequent anal-
May 2, 2025 0:28 cozma˙trautmann˙arxiv

Neutron Star Radii from Laboratory Experiments 23

ysis63 of the FOPI-LAND and ASY-EOS data. In the latter case, the uncertainty
grows rapidly toward 2nsat because the contribution of the less well determined
Ksym parameter becomes important (Eq. (2)). This work had demonstrated that it
is possible to determine L and Ksym independently but that additional experimental
data will be needed for achieving more satisfactory results.
New results may be expected from presently performed experiments at the
RIKEN and GSI/FAIR laboratories. Data taking by the SπRIT Collaboration at
RIKEN has been completed in 2024, with emphasis given to collecting the higher
statistics found necessary for analyzing pion ratios at high transverse meomentum
(Fig. 6 in Section 5). The density regime covered with this method is estimated
to be around 1.5nsat .132 The ASY-EOS Collaboration is scheduled for data taking
at the GSI laboratory in 2025 with the aim to extend the probed density interval
up to about 2nsat by measuring an excitation function of elliptic flows for incident
energies up to 1 GeV/nucleon.156 The expected mass resolution for light charged
particles and improved count rate statistics will enhance the significance of the
measured flow ratios (Fig. 9 in Section 7).

10. Conclusion
Recent analyses based on combining information obtained from nuclear theory,
heavy-ion collisions and astrophysical observations have confined the obtained radii
of the canonical 1.4-solar-mass neutron star to values between 12 and 13 km. Typical
errors are still of the order of 0.9 km (95%) but can be as low as 0.5 km (95%) or
0.4 km (90%) as reported in the most recent analyses of Koehn et al.138 and Pang
et al.,139 respectively.
At the time of its observation, it was not expected that GW170817 will remain
the only neutron-star merger of its kind for such a long time.6 Observing run 04b
of LIGO/Virgo/KAGRA which started in April 2024 is scheduled to continue until
June 2025.157 The LIGO and Virgo observatories are presently online with their ex-
panded reach due to higher sensitivity. The Kamioka Gravitational Wave Detector
KAGRA is expected to join as well. The consistent picture that has emerged from
the GW170817 multi-messenger observations and their interpretations makes it un-
likely, however, that any new observation of comparable significance will be severely
contradicting existing results. Their confirmation and possible improvement is still
very important, and a reduction of remaining uncertainties can be expected.
Altogether, one may expect that the precision of the result of Legred et al.21
(Fig. 2) will be improved without changing the overall behavior. In particular, the
data now available for PSR J0740+6620 represent a very strong reference point
for neutron-star matter up to high densities. Many other sources of astrophysi-
cal information as used by Koehn et al.138 may profit from continuing analyses
and observations. The statistical weight of information resulting from astrophysical
observations is thus expected to increase.
At low densities, the precise knowledge of the symmetry energy at 2/3 satura-
May 2, 2025 0:28 cozma˙trautmann˙arxiv

24 M. D. Cozma and W. Trautmann

tion density may be considered as the second robust reference point. Open issues at
lower densities are the precision of the symmetry energy at saturation density and
the pressures in the interval up to twice that value which, however, are accessible
with laboratory experiments. We may, therefore, expect that heavy-ion experiments
presently prepared and conducted at the RIKEN and GSI/FAIR laboratories have
the potential of reducing the still existing uncertainties in this density interval. Con-
tinuing activities in the field of transport theory110, 124 will be essential for achieving
this goal.

Acknowledgment
The authors would like to thank C. A. Raithel, R. Essick, N. Chamel, Yongjia Wang,
W. G. Lynch, M. B. Tsang and P. T. H. Pang for providing graphic material and
for stimulating communications. The continuing fruitful scientific exchange within
the ASY-EOS Collaboration156 is gratefully acknowledged.

References
1. B. P. Abbott et al., Phys. Rev. Lett. 119, 161101 (2017).
2. Focus Issue on the Electromagnetic Counterpart of the Neutron Star Binary Merger
GW170817, Astrophys. J. Lett. 848, Number 2 (2017).
3. J. M. Lattimer, Annu. Rev. Nucl. Part. Sci. 71, 433 (2021).
4. D. Watson et al., Nature 574, 497 (2019).
5. B. P. Abbott et al., Phys. Rev. Lett. 121, 161101 (2018).
6. A. G. Chaves and T. Hinderer, J. Phys. G 46, 123002 (2019).
7. J. M. Lattimer and M. Prakash, Phys. Rep. 442, 109 (2007).
8. Bao-An Li, Lie-Wen Chen and Che Ming Ko, Phys. Rep. 464, 113 (2008).
9. T. E. Riley et al., Astrophys. J. Lett. 887, L21 (2019).
10. M. C. Miller et al., Astrophys. J. Lett. 887, L24 (2019).
11. T. E. Riley et al., Astrophys. J. Lett. 918, L27 (2021).
12. M. C. Miller et al., Astrophys. J. Lett. 918, L28 (2021).
13. C. A. Raithel, F. Özel and D. Psaltis, Astrophys. J. 875, 12 (2019); Erratum: Astro-
phys. J. 915, 73 (2021).
14. Bao-An Li, Nucl. Phys. A 708, 365 (2002).
15. J. M. Lattimer, Annu. Rev. Nucl. Part. Sci. 62, 485 (2012).
16. G. Baym, T. Hatsuda, T. Kojo, P. D. Powell, Yifan Song and T. Takatsuka, Rep.
Prog. Phys. 81, 056902 (2018).
17. Madappa Prakash, I. Bombaci, Manju Prakash, P. J. Ellis, J. M. Lattimer and R.
Knorren, Phys. Rep. 280, 1 (1997).
18. J. A. Pons, S. Reddy, M. Prakash, J. M. Lattimer and J. A. Miralles, Astrophys. J.
513, 780 (1999).
19. The LIGO Scientific Collab., the Virgo Collab., the KAGRA Collab.: R. Abbott et
al., Phys. Rev. X 13, 011048 (2023).
20. T. Dietrich et al., Science 370, 1450 (2020).
21. I. Legred, K. Chatziioannou, R. Essick, Sophia Han and P. Landry, Phys. Rev. D 104,
063003 (2021).
22. H. T. Cromartie et al., Nature Astronomy 4, 72 (2019).
23. E. Fonseca et al., Astrophys. J. Lett. 915, L12 (2021).
24. P. B. Demorest, T. Pennucci, S. M. Ransom, M. S. E. Roberts and J. W. T. Hessels,
May 2, 2025 0:28 cozma˙trautmann˙arxiv

Neutron Star Radii from Laboratory Experiments 25

Nature 467, 1081 (2010).

25. J. Antoniadis et al., Science 340, 6131 (2013).
26. F. Douchin and P. Haensel, Astron. Astrophys. 380, 151 (2001).
27. I. Tews, Phys. Rev. C 95, 015803 (2017).
28. G. Baym, S. Furusawa, T. Hatsuda, T. Kojo and H. Togashi, Astrophys. J. 885, 42
(2019).
29. R. N. Wolf et al., Phys. Rev. Lett. 110, 041101 (2013).
30. N. Chamel and P. Haensel, Living Rev. Relativity 11, 10 (2008).
31. L. Brandes, W. Weise and N. Kaiser, Phys. Rev. D 108, 094014 (2023).
32. Shao-Peng Tang, Jin-Liang Jiang, Ming-Zhe Han, Yi-Zhong Fan and Da-Ming Wei,
Phys. Rev. D 104, 063032 (2021).
33. L. R. Weih, M. Hanauske and L. Rezzolla, Phys. Rev. Lett. 124, 171103 (2020).
34. R. Somasundaram, I. Tews and J. Margueron, Phys. Rev. C 107, L052801 (2023).
35. E. Annala et al., Nature Communications 14, 8451 (2023).
36. P. T. H. Pang et al., Phys. Rev. C 109, 025807 (2024).
37. E. Khan and J. Margueron, Phys. Rev. Lett. 88, 034319 (2013).
38. L. Perot, N. Chamel and A. Sourie, Phys. Rev. C 100, 035801 (2019).
39. L. Trippa, G. Colò and E. Vigezzi, Phys. Rev. C 77, 061304(R) (2008).
40. X. Roca-Maza et al., Phys. Rev. C 87, 034301 (2013).
41. Ning Wang, Li Ou and Min Liu, Phys. Rev. C 87, 034327 (2013).
42. Zhen Zhang and Lie-Wen Chen, Phys. Lett. B 726, 234 (2013).
43. B. A. Brown, Phys. Rev. Lett. 111, 232502 (2013).
44. Pengsheng Wen and J. W. Holt, Phys. Rev. C 103, 064002 (2021).
45. Zhen Zhang and Lie-Wen Chen, Phys. Rev. C 90, 064317 (2014).
46. B. A. Brown, Phys. Rev. Lett. 85, 5296 (2000).
47. C. Fuchs and H. H. Wolter, Eur. Phys. J. A 30, 5 (2006).
48. Zhen Zhang and Lie-Wen Chen, Phys. Rev. C 92, 031301(R) (2015).
49. M. B. Tsang et al., Phys. Rev. Lett. 102, 122701 (2009).
50. P. Danielewicz and J. Lee, Nucl. Phys. A 922, 1 (2014).
51. C. J. Horowitz et al., J. Phys. G 41, 093001 (2014).
52. G. Audi et al., Chin. Phys. C 36, 1287 (2012).
53. S. Goriely, N. Chamel and J. M. Pearson, Phys. Rev. C 88, 024308 (2013).
54. C. J. Horowitz and A. Schwenk, Nucl. Phys. A 776, 55 (2006).
55. S. Typel, G. Röpke, T. Klähn, D. Blaschke and H. H. Wolter, Phys. Rev. C 81, 015803
(2010).
56. J. B. Natowitz et al., Phys. Rev. Lett. 104, 202501 (2010).
57. M. Hempel, K. Hagel, J. Natowitz, G. Röpke and S. Typel, Phys. Rev. C 91, 045805
(2015).
58. R. Bougault et al., J. Phys. G 47, 025103 (2020).
59. Bao-An Li, P. G. Krastev, De-Hua Wen and Nai-Bo Zhang, Eur. Phys. J. A 55, 117
(2019).
60. G. F. Burgio, H.-J. Schulze, I. Vidaña and Jin-Biao Wei, Prog. Part. Nucl. Phys. 120,
103879 (2021).
61. Bao-An Li and Xiao Han, Phys. Lett. B 727, 276 (2013).
62. M. Oertel, M. Hempel, T. Klähn and S. Typel, Rev. Mod. Phys. 89, 015007 (2017).
63. M. D. Cozma, Eur. Phys. J. A 54, 40 (2018).
64. Bao-An Li and M. Magno, Phys. Rev. C 102, 045807 (2020).
65. C. J. Horowitz and J. Piekarewicz, Phys. Rev. Lett. 86, 5647 (2001).
66. D. M. Rossi et al., Phys. Rev. Lett. 111, 242503 (2013).
67. A. Tamii, P. von Neumann-Cosel and I. Poltoratska, Eur. Phys. J. A 50, 28 (2014).
May 2, 2025 0:28 cozma˙trautmann˙arxiv

26 M. D. Cozma and W. Trautmann

68. D. Adhikari et al., Phys. Rev. Lett. 126, 172502 (2021).

69. B. T. Reed, F. J. Fattoyev, C. J. Horowitz and J. Piekarewicz, Phys. Rev. Lett. 126,
172503 (2021).
70. C. J. Horowitz, J. Piekarewicz and B. Reed, Phys. Rev. C 102, 044321 (2020).
71. R. Essick, I. Tews, P. Landry and A. Schwenk, Phys. Rev. Lett. 127, 192701 (2021).
72. D. Adhikari et al., Phys. Rev. Lett. 129, 042501 (2022).
73. P.-G. Reinhard, X. Roca-Maza and W. Nazarewicz, Phys. Rev. Lett. 129, 232501
(2022).
74. J. M. Lattimer, Particles 6, 30 (2023).
75. Zhen Zhang and Lie-Wen Chen, Phys. Rev. C 108, 024317 (2023).
76. Jia Zhou and Jun Xu, Science China Physics, Mechanics & Astronomy 67, 282011
(2024).
77. B. T. Reed, F. J. Fattoyev, C. J. Horowitz and J. Piekarewicz, Phys. Rev. C 109,
035803 (2024).
78. D. Becker et al., Eur. Phys. J. A 54, 208 (2018).
79. Baishan Hu et al., Nature Physics 18, 1196 (2022).
80. G. Giacalone, G. Nijs and W. van der Schee, Phys. Rev. Lett. 131, 202302 (2023).
81. T. Aumann, C. A. Bertulani, F. Schindler and S. Typel, Phys. Rev. Lett. 119, 262501
(2017).
82. L. Ponnath et al., Phys. Lett. B 855, 138780 (2024).
83. M. Baldo and G. F. Burgio, Prog. Part. Nucl. Phys. 91, 203 (2016).
84. Topical Issue on Nuclear Symmetry Energy, edited by Bao-An Li, À. Ramos, G. Verde
and I. Vidaña, Eur. Phys. J. A 50, 9 (2014).
85. Bao-An Li, Phys. Rev. Lett. 88, 192701 (2002).
86. P. Danielewicz, R. Lacey and W. G. Lynch, Science 298, 1592 (2002).
87. C. Fuchs, A. Faessler, E. Zabrodin and Yu-Ming Zheng, Phys. Rev. Lett. 86, 1974
(2001).
88. P. Danielewicz, Nucl. Phys. A 673, 375 (2000).
89. A. W. Steiner, J. M. Lattimer and E. F. Brown, Astrophys. J. Lett. 765, L5 (2013).
90. W. Reisdorf et al., Nucl. Phys. A 848, 366 (2010).
91. W. Reisdorf et al., Nucl. Phys. A 876, 1 (2012).
92. A. Le Fèvre, Y. Leifels, W. Reisdorf, J. Aichelin and Ch. Hartnack, Nucl. Phys. A
945, 112 (2016).
93. Yongjia Wang, Chenchen Guo, Qingfeng Li, A. Le Fèvre, Y. Leifels and W. Traut-
mann, Phys. Lett. B 778, 207 (2018).
94. Bao-An Li, Phys. Rev. Lett. 85, 4221 (2000).
95. P. Russotto et al., Phys. Lett. B 697, 471 (2011).
96. M. D. Cozma, Phys. Lett. B 700, 139 (2011).
97. M. Kaneko et al., Phys. Lett. B 822, 136681 (2021).
98. G. Ferini, T. Gaitanos, M. Colonna, M. Di Toro and H. H. Wolter, Phys. Rev. Lett.
97, 202301 (2006).
99. W. Reisdorf et al., Nucl. Phys. A 781, 459 (2007).
100. Jun Hong and P. Danielewicz, Phys. Rev. C 90, 024605 (2014).
101. Taesoo Song and Che Ming Ko, Phys. Rev. C 91, 014901 (2015).
102. M. D. Cozma, Phys. Lett. B 753, 166 (2016).
103. Qingfeng Li and Zhuxia Li, Phys. Lett. B 773, 557 (2017).
104. G. Jhang et al., Phys. Lett. B 813, 136016 (2021).
105. Zhen Zhang and Che Ming Ko, Phys. Rev. C 97, 014610 (2018).
106. Jun Xu et al., Phys. Rev. C 109, 044609 (2024).
107. M. D. Cozma and M. B. Tsang, Eur. Phys. J. A 57, 309 (2021).
May 2, 2025 0:28 cozma˙trautmann˙arxiv

Neutron Star Radii from Laboratory Experiments 27

108. N. Ikeno and A. Ono, Phys. Rev. C 108, 044601 (2023).

109. J. Estee et al., Phys. Rev. Lett. 126, 162701 (2021).
110. H. Wolter et al., Prog. Part. Nucl. Phys. 125, 103962 (2022).
111. S. Huth et al., Nature 606, 276 (2022).
112. P. Russotto et al., Phys. Rev. C 94, 034608 (2016).
113. Y. Leifels et al., Phys. Rev. Lett. 71, 963 (1993).
114. D. Lambrecht et al., Z. Phys. A 350, 115 (1994).
115. W. Trautmann and H. H. Wolter, Int. J. Mod. Phys. E 21, 1230003 (2012).
116. Yongjia Wang, Chenchen Guo, Qingfeng Li, Hongfei Zhang, Y. Leifels and W. Traut-
mann, Phys. Rev. C 89, 044603 (2014).
117. M. D. Cozma, Y. Leifels, W. Trautmann, Qingfeng Li and P. Russotto, Phys. Rev.
C 88, 044912 (2013).
118. P. Russotto et al., Eur. Phys. J. A 50, 38 (2014).
119. Th. Blaich et al., Nucl. Instrum. Methods Phys. Res. A 314, 136 (1992).
120. A. Andronic, J. Lukasik, W. Reisdorf and W. Trautmann, Eur. Phys. J. A 30, 31
(2006).
121. A. Le Fèvre, Y. Leifels, C. Hartnack and J. Aichelin, Phys. Rev. C 98, 034901 (2018).
122. Qingfeng Li et al., Phys. Rev. C 83, 044617 (2011).
123. C. B. Das, S. Das Gupta, C. Gale and Bao-An Li, Phys. Rev. C 67, 034611 (2003).
124. M. D. Cozma, Phys. Rev. C 110, 064911 (2024).
125. A. Andronic et al., Phys. Rev. C 64, 034907 (2003).
126. W. Reisdorf et al., Phys. Rev. Lett. 92, 232301 (2004).
127. J. Margueron, R. H. Casali and F. Gulminelli, Phys. Rev. C 97, 025805 (2018).
128. C. Drischler, J. W. Holt and C. Wellenhofer, Annu. Rev. Nucl. Part. Sci. 71, 1
(2021).
129. K. Hebeler, J. M. Lattimer, C. J. Pethick and A. Schwenk, Astrophys. J. 773, 11
(2013).
130. I. Tews, J. Margueron and S. Reddy, Phys. Rev. C 98, 045804 (2018).
131. R. Essick, I. Tews, P. Landry, S. Reddy and D. E. Holz, Phys. Rev. C 102, 055803
(2020).
132. C. Y. Tsang, M. B. Tsang, W. G. Lynch, R. Kumar and C. J. Horowitz, Nature
Astronomy 8, 328 (2024).
133. C. Drischler, Sophia Han, J. M. Lattimer, M. Prakash, S. Reddy and Tianqi Zhao,
Phys. Rev. C 103, 045808 (2021).
134. A. Sorensen et al., Prog. Part. Nucl. Phys. 134, 104080 (2024).
135. J. E. Lynn et al., Phys. Rev. Lett. 116, 062501 (2016).
136. S. Huth, C. Wellenhofer and A. Schwenk, Phys. Rev. C 103, 025803 (2021).
137. I. Tews et al., arXiv:2407.08979 [nucl-th] (2024).
138. H. Koehn et al., Phys. Rev. X 15, 021014 (2025).
139. P. T. H. Pang et al., Nature Communications 14, 8352 (2023).
140. T. Salmi et al., Astrophys. J. 974, 294 (2024).
141. A. J. Dittmann et al., Astrophys. J. 974, 295 (2024).
142. D. Lonardoni, A. Lovato, S. Gandolfi and F. Pederiva, Phys. Rev. Lett. 114, 092301
(2015).
143. D. Logoteta, I. Vidaña and I. Bombaci, Eur. Phys. J. A 55, 207 (2019).
144. Jun-Ting Ye, Rui Wang, Si-Pei Wang and Lie-Wen Chen, arXiv:2411.18349 [nucl-th]
(2024).
145. W. Weise, EPJ Web Conf. 291, 01007 (2024).
146. D. Choudhury et al., Astrophys. J. Lett. 971, L20 (2024).
147. D. J. Reardon et al., Astrophys. J. Lett. 971, L18 (2024).
May 2, 2025 0:28 cozma˙trautmann˙arxiv

28 M. D. Cozma and W. Trautmann

148. S. Vinciguerra et al., Astrophys. J. 961, 62 (2024).

149. N. Rutherford et al., Astrophys. J. Lett. 971, L19 (2024).
150. T. Malik, V. Dexheimer and C. Providência, Phys. Rev. D 110, 043042 (2024).
151. L. Brandes and W. Weise, Phys. Rev. D 111, 034005 (2025).
152. L. A. Tarasovičová, J. Mohs, A. Andronic, H. Elfner and K.-H. Kampert, Eur. Phys.
J. A 60, 232 (2024).
153. V. Kireyeu et al., arXiv:2411.04969 [nucl-th] (2024).
154. J. Adamczewski-Musch et al., Eur. Phys. J. A 59, 80 (2023).
155. B. Kardan, J. Phys. Conf. Ser. 742, 012008 (2016).
156. P. Russotto et al., arXiv:2105.09233 [nucl-ex] (2021).
157. The LIGO Scientific Collab., the Virgo Collab., the KAGRA Collab.: for a recent
report see, e.g., A. G. Abac et al., Astrophys. J. Lett. 970, L34 (2024).