High-Density Free-Neutron Targets

• High-density free-neutron targets are specialized setups that concentrate neutrons in an interaction region, overcoming the instability of isolated neutrons.
• They encompass diverse designs such as moderated spallation targets, compact cyclotron-driven assemblies, and liquid-lithium systems, each optimized for specific experimental metrics.
• These configurations enable precise neutron diagnostics and scalable luminosity, improving reaction measurements and cross-section determinations in advanced nuclear studies.

A high-density free-neutron target is an arrangement that creates, concentrates, or experimentally approximates a neutron-rich interaction region for reaction studies that cannot be performed with conventional stationary samples. In the recent literature, the term encompasses several distinct but related realizations: a standing population of moderated neutrons inside a compact moderator for inverse-kinematics measurements, liquid-lithium neutron-producing targets in which a flowing free surface serves simultaneously as source and beam dump, storage-ring-compatible compact source–moderator assemblies, and quasi-free neutron proxies based on liquid deuterium. Complementary diagnostic work addresses how neutron-related observables are extracted from recoil spectra or reaction products when the target itself is not a literal free-neutron medium (Reifarth et al., 2017, Dellmann et al., 4 Mar 2026, Tarifeño-Saldivia et al., 21 Aug 2025, Halfon et al., 2013, Thulliez et al., 10 Oct 2025, Zhang et al., 17 Mar 2026, Sierra, 2023, Kumar et al., 2011).

1. Conceptual basis and performance metrics

The central difficulty is that isolated neutrons are unstable and cannot be fabricated into ordinary macroscopic targets. As a result, the field relies on figures of merit that quantify neutron occupancy in an interaction region rather than a static bulk material density. In moderated spallation concepts, the average number of neutrons inside the ion beam tube is written as

$$\bar{n}_{\mathrm{neutron}}=\frac{I_{\mathrm{proton}}}{e}\,\bar{t}_{\mathrm{neutron,tot}},$$

and the corresponding areal neutron density as

$$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$

In storage-ring designs based on compact cyclotrons, the relevant quantity is the integrated thermal neutron areal density,

$$A_{\mathrm{den}}=\sum_{j=1}^{100} A_{\mathrm{den},j}, \qquad A_{\mathrm{den},j}=n_{\mathrm{th},j}\,\ell,$$

with thermal neutrons defined as $E \leq 1\,\mathrm{eV}$. In internal cryogenic targets used for baryon scattering, the controlling quantity is the areal number density

$\mathcal{T}=\frac{\rho\,l}{M}N_A,$

which directly enters the effective luminosity (Reifarth et al., 2017, Tarifeño-Saldivia et al., 21 Aug 2025, Zhang et al., 17 Mar 2026).

These definitions reveal a common structure across otherwise dissimilar implementations. A neutron target may be realized as a moderated bath traversed by an ion beam, as a compact neutron-emitting surface coupled to a nearby secondary target or moderator, or as a quasi-free neutron proxy in deuterium. This suggests that “high-density” in this context is operational: it refers to a large neutron population or target thickness in the overlap region relevant to the reaction measurement, not necessarily to a macroscopic volume of unbound neutrons.

2. Standing neutron targets for inverse kinematics

The spallation-based inverse-kinematics concept uses a tungsten target bombarded by high-energy protons, surrounded by a large heavy-water moderator, with an ion beam pipe passing through the moderated neutron field. In the 2017 proposal, the geometry consists of a tungsten spallation target, a spherical D$_2$O moderator with radii of 0.5 m, 1.0 m, or 2.0 m, and an ion beam pipe placed perpendicular to the proton beam pipe and shifted by $x = 7.5$ cm. GEANT-3.21 with GCALOR was used for both primary-neutron transport and proton-on-tungsten spallation studies. The moderator increases the neutron residence time in the ion beam pipe by many orders of magnitude: for 800 MeV, small target, the average time grows from $0.0051\,\mu\text{s}$ without moderator to $245\,\mu\text{s}$ with a 2 m moderator, while for 20 GeV, large target, it rises from $0.057\,\mu\text{s}$ to $$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$0. Under LANL/LANSCE-like conditions—100 $$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$1A, 800 MeV, small tungsten target, ion beam pipe cross section 20 cm$$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$2—the areal neutron density reaches $$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$3 cm$$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$4 (Reifarth et al., 2017).

The same paper emphasizes that moderation changes the neutron field qualitatively as well as quantitatively. The raw spallation spectrum is broad and fast, but the spectrum inside the ion beam pipe becomes dominated by low-energy and thermal neutrons once the D$$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$5O sphere is present. Time-weighted spectra are even more strongly biased toward low energies, because fast neutrons traverse the pipe quickly whereas moderated neutrons remain in the interaction region. A further result is that shifting the target upstream to exploit forward-peaked spallation emission reduces the average neutron residence time, because repeated scattering in the moderator largely erases the original angular distribution (Reifarth et al., 2017).

The Los Alamos Neutron Target Demonstrator is a proof-of-principle experimental step toward this class of target. It tests whether a $$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$6 m$$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$7 graphite moderator containing a central neutron source can create a standing neutron target for inverse kinematics. The moderator is assembled from graphite blocks with density 1.62 g/cm$$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$8, and activation along the future ion-beam path is measured with gold wires using the $$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$9 reaction and the 412 keV gamma ray. Experiments with $$A_{\mathrm{den}}=\sum_{j=1}^{100} A_{\mathrm{den},j}, \qquad A_{\mathrm{den},j}=n_{\mathrm{th},j}\,\ell,$$0 at 1.95 and 2.5 MeV and $$A_{\mathrm{den}}=\sum_{j=1}^{100} A_{\mathrm{den},j}, \qquad A_{\mathrm{den},j}=n_{\mathrm{th},j}\,\ell,$$1 at 9 and 45 MeV show excellent agreement between measured and simulated spatial profiles for the full-cube moderator, while the half-cube configuration exhibits deviations at the wings that are attributed to room background. For the planned 800 MeV LANSCE spallation case, the expected neutron target density is $$A_{\mathrm{den}}=\sum_{j=1}^{100} A_{\mathrm{den},j}, \qquad A_{\mathrm{den},j}=n_{\mathrm{th},j}\,\ell,$$2 for a 10 $$A_{\mathrm{den}}=\sum_{j=1}^{100} A_{\mathrm{den},j}, \qquad A_{\mathrm{den},j}=n_{\mathrm{th},j}\,\ell,$$3A proton beam. The simulations further show that between about 0.1 ms and 0.1 s only thermal neutrons are present in the ion beam pipe, which is the essential physical basis of the standing-neutron-target idea (Dellmann et al., 4 Mar 2026).

3. Compact cyclotron-driven targets for storage rings

A later storage-ring proposal replaces the large spallation driver with a compact cyclotron-based source centered on the $$A_{\mathrm{den}}=\sum_{j=1}^{100} A_{\mathrm{den},j}, \qquad A_{\mathrm{den},j}=n_{\mathrm{th},j}\,\ell,$$4 reaction. The architecture integrates four subsystems: a compact cyclotron neutron source, a moderator/reflector assembly using either heavy water or BeO with a graphite reflector shell, a cryogenic liquid-hydrogen moderator placed close to the ion-beam path, and a beam-pipe geometry that preserves vacuum for ion circulation while allowing neutron diffusion into the interaction volume. The demonstrator is based on a commercial cyclotron such as the IBA Cyclone KEY, with proton energy $$A_{\mathrm{den}}=\sum_{j=1}^{100} A_{\mathrm{den},j}, \qquad A_{\mathrm{den},j}=n_{\mathrm{th},j}\,\ell,$$5 and beam current $$A_{\mathrm{den}}=\sum_{j=1}^{100} A_{\mathrm{den},j}, \qquad A_{\mathrm{den},j}=n_{\mathrm{th},j}\,\ell,$$6, and is designed for CRYRING-scale installation (Tarifeño-Saldivia et al., 21 Aug 2025).

The moderating materials are chosen by comparing the macroscopic slowing down power $$A_{\mathrm{den}}=\sum_{j=1}^{100} A_{\mathrm{den},j}, \qquad A_{\mathrm{den},j}=n_{\mathrm{th},j}\,\ell,$$7 and moderating ratio $$A_{\mathrm{den}}=\sum_{j=1}^{100} A_{\mathrm{den},j}, \qquad A_{\mathrm{den},j}=n_{\mathrm{th},j}\,\ell,$$8. Light water is rejected as the primary moderator because its absorption is too large. Heavy water and BeO exceed the optimization threshold of $$A_{\mathrm{den}}=\sum_{j=1}^{100} A_{\mathrm{den},j}, \qquad A_{\mathrm{den},j}=n_{\mathrm{th},j}\,\ell,$$9 for thermal areal density normalized to one primary neutron, while graphite is selected as the reflector shell. Two optimized geometries are retained: a D$E \leq 1\,\mathrm{eV}$0O + graphite configuration with moderator core $E \leq 1\,\mathrm{eV}$1 cm, $E \leq 1\,\mathrm{eV}$2 cm and graphite reflector thickness $E \leq 1\,\mathrm{eV}$3 cm, and a BeO + graphite configuration with $E \leq 1\,\mathrm{eV}$4 cm, $E \leq 1\,\mathrm{eV}$5 cm and $E \leq 1\,\mathrm{eV}$6 cm. A 20 K liquid-hydrogen moderator around the interaction region is then optimized with a thickness scan from 0 to 40 mm, yielding an optimum around $E \leq 1\,\mathrm{eV}$7 mm and a roughly 60% increase over the threshold (Tarifeño-Saldivia et al., 21 Aug 2025).

For the CRYRING demonstrator, the quoted thermal neutron areal density is $E \leq 1\,\mathrm{eV}$8. The same study outlines explicit upgrade paths: commercial compact cyclotrons in the $E \leq 1\,\mathrm{eV}$9 mA range could reach about $\mathcal{T}=\frac{\rho\,l}{M}N_A,$0, and future compact isochronous cyclotrons approaching $\mathcal{T}=\frac{\rho\,l}{M}N_A,$1 mA could push beyond $\mathcal{T}=\frac{\rho\,l}{M}N_A,$2. Combined with $\mathcal{T}=\frac{\rho\,l}{M}N_A,$3 stored ions and a revolution frequency of $\mathcal{T}=\frac{\rho\,l}{M}N_A,$4 kHz, the luminosity estimate is $\mathcal{T}=\frac{\rho\,l}{M}N_A,$5, corresponding to $\mathcal{T}=\frac{\rho\,l}{M}N_A,$6 events/day with $\mathcal{T}=\frac{\rho\,l}{M}N_A,$7 in mb under the stated assumptions. The paper is explicit that this concept is not yet experimentally demonstrated; the next steps are to build and validate a demonstrator, measure the real $\mathcal{T}=\frac{\rho\,l}{M}N_A,$8 source performance, and integrate the source with storage-ring hardware and detection (Tarifeño-Saldivia et al., 21 Aug 2025).

4. Liquid-lithium neutron-producing targets

Liquid-lithium systems realize a different form of high-density free-neutron target: the neutron source is localized to the first few micrometers of a flowing free surface, while the rest of the liquid simultaneously functions as a self-renewing beam dump. In LiLiT, developed at Soreq Nuclear Research Center as part of the SARAF program, the relevant reaction is

$\mathcal{T}=\frac{\rho\,l}{M}N_A,$9

with $_2$0 and threshold $_2$1. Near threshold, the emitted spectrum is epithermal rather than MeV-scale, with the thick-target angle-integrated neutron spectrum at $_2$2 peaking around 25–30 keV and closely approximating a Maxwellian distribution with effective thermal energy $_2$3. The target is a windowless transverse liquid-metal jet operating at about $_2$4 to $_2$5, with film thickness 1.5 mm, width 18 mm, and velocity up to 7 m/s. TRIM calculations for a 1.91 MeV, 1 mA beam give a proton range of $_2$6 in liquid lithium, but neutron production occurs only in the first $_2$7 of the film. The projected output is $_2$8, and a secondary target 3 mm downstream is predicted to see $_2$9 with most probable energy 28 keV and mean energy 46 keV. Electron-beam tests demonstrated dissipation of about $x = 7.5$0 areal power density and about $x = 7.5$1 volumetric power density at a lithium velocity around 4 m/s, with stable film and acceptable vacuum (Halfon et al., 2013).

SATELIT at CEA-Saclay extends the liquid-lithium approach toward long-duration operation of a compact accelerator-driven neutron source. It couples the IPHI accelerator to a closed liquid-lithium loop producing a 2 mm thick film in vacuum, bombarded by a 3 MeV proton beam at 11 mA nominal and 10 kW beam power. The loop contains 22 L, or 11 kg, of lithium at 220–250 °C. During a 2024–2025 campaign, the facility accumulated nearly 100 h of neutron-production running time and 840 kW·h of deposited beam power, including two continuous runs exceeding 11 h. The normalization yield is $x = 7.5$2 n/p, corresponding to $x = 7.5$3 fast neutrons/s at 10 kW. A polyethylene moderator of dimensions $x = 7.5$4 mm placed 45 mm from the nozzle produces a thermal neutron beam whose flux at 1.4 m from the extraction point is above $x = 7.5$5 n·cm$x = 7.5$6·s$x = 7.5$7. Without a sapphire filter, gold foils give a center thermal flux of $x = 7.5$8 n·cm$x = 7.5$9·s$0.0051\,\mu\text{s}$0·kW$0.0051\,\mu\text{s}$1; with the filter the thermal flux is reduced by about 30%, while epithermal and fast flux are suppressed by about a factor of 10 and the cadmium ratio rises from about 5 to about 25. The principal long-term limitation is the accumulation of $0.0051\,\mu\text{s}$2Be: after nearly 100 h at 10 kW, the total activity in the 22 L inventory is estimated at about 200 GBq, motivating mitigation strategies such as a working cold trap, remote maintenance, chemical distillation, and rinsing pipes with $0.0051\,\mu\text{s}$3Be-free lithium (Thulliez et al., 10 Oct 2025).

5. Quasi-free neutron targets and internal detector implementations

Not all practical neutron-target programs aim to produce literal unbound neutrons in free space. In the BESIII proposal for precision $0.0051\,\mu\text{s}$4 and $0.0051\,\mu\text{s}$5 measurements, the preferred neutron-target proxy is a dedicated liquid deuterium target installed between the beam pipe and the Cylindrical Gas Electron Multiplier Inner Tracker. The proposed annular vessel occupies radii $0.0051\,\mu\text{s}$6–75 mm, with beam pipe radius 33.7 mm and CGEM-IT outer radius 76.9 mm, so the radial thickness is $0.0051\,\mu\text{s}$7 mm. The vessel is double-walled, cylindrical, and made of thin Kapton; the quoted liquid densities are 0.071 g/cm$0.0051\,\mu\text{s}$8 for LH$0.0051\,\mu\text{s}$9 at 20.3 K and 0.164 g/cm$245\,\mu\text{s}$0 for LD$245\,\mu\text{s}$1 at 23.7 K. The paper is explicit that LD$245\,\mu\text{s}$2 is not a free neutron target in the literal sense, because the neutron remains bound inside deuterium, but it is presented as the “premier proxy” for neutron targets since nuclear-structure corrections are far smaller than for heavier nuclei. Monte Carlo studies with one million events per configuration show that detector impact is modest: the worst case is about a 7% efficiency loss for protons with the 40 mm LD$245\,\mu\text{s}$3 target, the 20 mm LH$245\,\mu\text{s}$4 target gives less than 2% efficiency loss, and momentum and angle resolution degradation remain below 6%. For $245\,\mu\text{s}$5, $245\,\mu\text{s}$6, $245\,\mu\text{s}$7, and $245\,\mu\text{s}$8 beams, the effective luminosity for scattering on free protons is expected to increase by a factor of 10–30 relative to the beam pipe; for $245\,\mu\text{s}$9 the gain is about 5 because of earlier decay. The same target program substantially reduces systematics associated with the existing composite beam-pipe materials, where BESIII previously reported a 9.5% total systematic uncertainty for $0.057\,\mu\text{s}$0, dominated by 6.1% from cooling pipe material and 3.6% from the nuclear scaling model (Zhang et al., 17 Mar 2026).

This neutron-target proxy framework is experimentally significant because it separates two questions that are often conflated. One is whether the interaction occurs on a strictly free neutron. The other is whether neutron-level cross sections can be extracted with sufficiently small and controlled nuclear corrections. The BESIII proposal addresses the second question directly: LH$0.057\,\mu\text{s}$1 provides the cleanest free-proton benchmark, while LD$0.057\,\mu\text{s}$2 gives the best practical access to neutron-induced or neutron-tagged processes inside the detector volume (Zhang et al., 17 Mar 2026).

6. Diagnostics, proxy observables, and limiting distinctions

High-density neutron-target work is closely connected to methods for characterizing neutron distributions and neutron-rich matter, but these methods are not themselves target technologies. In coherent elastic neutrino–nucleus scattering, the sensitivity arises because the weak charge is neutron dominated,

$0.057\,\mu\text{s}$3

with $0.057\,\mu\text{s}$4 and $0.057\,\mu\text{s}$5. For a one-tonne liquid argon pseudo-data set with 5 recoil bins and a 5 keV$0.057\,\mu\text{s}$6 threshold, fits based on the weak-charge Helm parametrization yield $0.057\,\mu\text{s}$7 at 90% CL and translate to $0.057\,\mu\text{s}$8, while the decomposition-based approach with $0.057\,\mu\text{s}$9-dependent nucleon form factors gives $$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$00. Both correspond to a $$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$01–15% determination of the point-neutron radius in the stated setup. With substantial reduction of beam-related neutron and steady-state backgrounds, the same study concludes that a $$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$02 precision extraction becomes feasible. By contrast, the model-independent moment expansion leaves the data insufficient to determine $$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$03 sharply and produces only the upper limit $$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$04 at 90% CL (Sierra, 2023).

A different type of proxy appears in heavy-ion collisions, where the target is not a neutron medium but the observable is treated as a probe of high-density neutron-rich matter. Using the Isospin Quantum Molecular Dynamics model for Sn+Sn systems from 50 to 600 MeV/nucleon, the relevant quantity is the neutron-to-proton ratio $$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$05 and, after cancellation of Coulomb effects and some systematic uncertainties, the double ratio

$$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$06

The central conclusion is that the double neutron-to-proton ratio from free nucleons is sensitive enough to probe the high-density behavior of nuclear symmetry energy, whereas the double ratio from light charged particles becomes almost insensitive above about 200 MeV/nucleon and intermediate mass fragments are not produced in the supra-saturation region. The paper therefore identifies the free-nucleon $$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$07 as the relevant high-density free-neutron-target proxy observable rather than a material target (Kumar et al., 2011).

These distinctions resolve a recurring misconception. A standing neutron bath in graphite or heavy water, a near-threshold lithium source, an LD$$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$08 quasi-free target, CE$$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$09NS neutron-radius extraction, and a heavy-ion free-nucleon $$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$10 ratio do not represent the same object. They address different parts of a common experimental problem: how to create, approximate, or infer a neutron-rich interaction region with enough effective density, luminosity, or interpretability to measure neutron-induced phenomena. The present literature shows that moderated spallation targets and compact cyclotron-based assemblies are the primary routes toward literal standing free-neutron targets for inverse kinematics; liquid-lithium systems are the most developed compact neutron-producing free-surface targets; LD$$\eta_{\mathrm{neutron}}=\frac{\bar{n}_{\mathrm{neutron}}}{A_{\mathrm{ion\;pipe}}}.$$11 is the preferred practical neutron proxy in detector-based scattering; and precision recoil or fragment observables provide complementary access to neutron distributions and high-density neutron-rich matter rather than target construction itself (Dellmann et al., 4 Mar 2026, Tarifeño-Saldivia et al., 21 Aug 2025, Halfon et al., 2013, Thulliez et al., 10 Oct 2025, Zhang et al., 17 Mar 2026, Sierra, 2023, Kumar et al., 2011).