Shielding design for IECF devices: ensuring safety through material analysis

1 Introduction

An essential component in the installation of any radiation source is the provisioning of sufficient radiation shielding to protect operators, research staff, and members of the public. High flux neutron sources, commercial power reactors and research reactors, are capable of producing neutron flux in excess of 1 0 14 n c m − 2 s − 1, requiring large amounts of multi layered shielding, often consisting of metres of concrete doped with different elements designed to moderate the neutrons energy over multiple scattering events until they are ultimately absorbed. In contrast, Inertial Electrostatic Confinement Fusion (IECF) devices operate at much lower rates. When operating with a DD fuel, devices can reach neutron production rates on the order of 1 0 8 n s − 1 in continuous mode, and 1 0 9 n s − 1 in a pulsed mode (Wulfkühler et al., 2025; Yoshikawa et al., 2007; Bakr et al., 2019).

Key motivations for the use of IECF devices are their compact size, portability, and lower neutron yield compared to research reactors. This makes them highly desirable for laboratory use and in-field environments, with recent research showing an increased use in the areas of medical isotope production (Cipiti, 2004; Cipiti and Kulcinski, 2003; Weidner et al., 2003; Cipiti and Kulcinski, 2005) and active interrogation of materials (Takahashi et al., 2014; Roy et al., 2021; Masuda et al., 2024; Archambault et al., 2017). These situations are conducted in setups where space and shielding may be limited. Consequently, an optimised shielding design that ensures radiation safety while maintaining the system’s compactness is of significant practical importance.

The source proposed is based on an IECF device (Hirsch, 1967), consisting of a 6 cm diameter concentric grid molybdenum cathode enclosed within a 17 cm diameter titanium anode. The anode is electrically grounded and encapsulated in a 1 cm wide water jacket used for surface cooling. A diagram of the simulated neutron source is provided in Figure 1.

The operation of an IEC device involves evacuating the vacuum chamber and removing any ambient gas that will disrupt the fusion process. The chamber is then backfilled with high-purity deuterium gas to a pressure of a few millitorr. By applying a negatively biased voltage of the order of 10’s of keV between the central cathode and the grounded anode, an electric field is established. The high negative bias induces electron field emission in the cathode. The electrons are then accelerated outwards in the electric field and used to ionise the deuterium gas, generating a plasma (Miley and Sved, 1997). The ionisation process is shown below.

e − + D 2 → D 2 + + 2 e − e − + D 2 → D + + D + 2 e −

The D + ions are accelerated towards the cathode, producing an ion current of a few milliamps. Due to the grid structure, a large fraction of the ions will miss the cathode, travelling through the holes in the grid, generating a localised high-density ion region in the centre, where they undergo fusion reactions, yielding neutrons and protons from the following interactions:

D + + D + → T + p 3.02 MeV D + + D + → 3 He + n 2.45 MeV

Both reactions occur with approximately equal probability. However, protons will not leave the containment vessel as they will likely strike the cathode or collide with the wall of the vacuum chamber.

A key advantage of the IECF device is its portability. The current device has a diameter, height, and weight of 20 cm, 60 cm, and 35 kg. It can operate in either pulsed or continuous modes, with the neutron rate adjustable based on the applied voltage, current, and pressure. Additional information can be found in the reference paper on operation parameters (Bakr et al., 2024a).

The effect of shielding material composition on neutron attenuation has been the subject of extensive research (Bakr et al., 2024b; Bardhan Paul et al., 2023; Gokul et al., 2023; Ali et al., 2023; Zhou et al., 2025). Building on this foundation, the present work first simulates H * ( 10 ) ambient dose coefficients for neutrons and photons to validate the accuracy of the Geant4 simulation framework against tabulated reference data. The paper then proceeds to model a proposed radiation facility operating with an IECF utilising a DD fuel at 2.45 MeV, evaluating the shielding requirements necessary to reduce ambient dose rates to acceptable regulatory limits.

Simulations are developed using the Geant4 Monte Carlo toolkit (Agostinelli et al., 2003) to evaluate the effect of shielding thickness and material type to determine the minimum shielding width to reduce the ambient dose rate below the radiation protection limits, as specified by the UK Ionising Radiations Regulations 2017 (IRR17) (Expert Participation, 2017).

This paper is structured as follows: Section 2 outlines the methods and setup of the simulation, including material compositions, definitions for dose scoring, and simulation setup. Section 3 provides the results and a discussion of the simulation results. Finally, Section 4 offers a conclusion and next steps.

2 Methods

2.1 H * ( 10 ) definition

The ambient dose equivalent, H * ( 10 ), is defined as “The dose equivalent at a point in a radiation field that would be produced by the corresponding expanded and aligned field in the ICRU sphere at a depth of 10 mm on the radius vector opposing the direction of the aligned field” (ICRP International Commission on Radiological Protection, 1996). In this definition, the concept of a radiation field that is expanded and aligned is a theoretical construct and can be thought of as a field in which all field lines are perpendicular and evenly distributed throughout. The ambient dose can then be determined by calculating the dose deposited into a sphere of radius 15 cm constructed of ICRU Tissue, with a small scoring volume 10 mm into the sphere.

This research defines the scoring volume as a 5 mm radius cylinder with a length of 2 mm located 1 cm into the ICRU sphere, as shown in Figure 2.

H * ( 10 ) is obtained by multiplying the deposited energy, E dep, by a radiobiological quality factor, Q ( L ), which depends on the particle linear energy transfer (LET). The value of Q ( L ) is outlined in ICRP Publication 60 (ICRP International Commission on Radiological Protection, 1991), and refined in ICRP 103 (ICRP International Commission on Radiological Protection, 2007), and is given by Equation 1:

Q L = 1 , L < 10 keV / μ m , 0.32 L − 2.2 , 10 ≤ L ≤ 100 keV / μ m , 300 L , L > 100 keV / μ m , (1)

Gamma rays and electrons have a quality factor of 1, and only neutrons and heavy charged particles (e.g., protons, α-particles, or recoil nuclei) will have higher LET values.

2.2 ICRU tissue

The ICRU provides the chemical makeup of a reference human tissue, known as the ICRU soft tissue (ICRP International Commission on Radiological Protection, 2007), shown in Table 1.

TABLE 1. Elemental composition of ICRU soft tissue (ICRU-4) (GEANT4 Collaboration, 2026a).

This is used as the reference material for the Geant4 for H * ( 10 ) coefficient simulations.

2.3 Simulating H * ( 10 ) coefficients

Geant4 Neutron and photon transport through shielding materials such as iron, polymers, and concrete relies on evaluated nuclear data libraries and material-specific scattering models. The performance of these models has been benchmarked against shielding configurations and experimental shielding studies (Nunnenmann and Fischer, 2019; Nunnenmann et al., 2020; Shin et al., 2014), and comparisons against reference gamma attenuation coefficients (Alabsy et al., 2021; Aladailah et al., 2022). This work will focus on the verification of the fluence-to-dose conversion coefficient method.

To verify the implementation of the fluence-to-ambient dose equivalent, Geant4 was used to calculate the H * ( 10 ) conversion coefficients for neutrons and photons. These values were then compared to tabulated reference data published by the ICRP.

To do this, a circular source surface emitting particles towards the ICRU sphere was constructed in Geant4. The emitted particles are fired towards the ICRU sphere. Neutrons and gammas are scattered within the sphere, depositing energy as they step, ionising atoms within the tissue. A cylindrical volume of dimensions r = 5 m m & l = 2 m m was centred at a depth of 10 mm into the sphere on the axis of the radiating plane with the circular face directed towards the incoming radiation. Energy deposited into this volume was extracted and used to determine the deposited dose.

Particle steps were tracked through the sensitive region. For each step, the step length, L step, and E dep were recorded. These values were used to determine the LET of that step and derive the corresponding Q ( L ). E dep was then weighted by the Q ( L ) and summed over the simulation run.

The total deposited energy and the number of events were extracted and processed by a Python script. The script calculated the total fluence of the test run, normalising the number of simulated particles by the radiation field area, to determine a fluence in p a r t i c l e s c m − 2. The total deposited energy was converted to Joules and divided by the mass of the sensitive volume to determine the absorbed dose. Using a scaling factor, the value was returned in p S v c m 2.

There is a low probability for simulated particles to interact within the small cylindrical scoring volume, the region where H * ( 10 ) is evaluated, under conventional Monte Carlo sampling statistics. To increase the probability of interaction, biasing techniques can be employed, which focus the incoming particle around the sensitive region. To ensure results are accurate and unbiased, a statistical weight is applied to each event to account for the biased sampling and preserve an unbiased estimate of the dose. This is what is known as variance-reduction, and has the impact of improving the statistics (variance in results) without increasing the number of particle histories.

This work utilises a technique known as the Ferrari biasing method, which has previously been used to calculate H * ( 10 ) coefficients for gamma dose simulation studies (Casanovas et al., 2016) and has been repeated here for both gammas and neutrons.

The implementation of the technique is as follows. The radial position, r, of emitted particles is sampled according to Equation 2:where R is the beam radius, ξ ∈ U ( 0,1 ) is a uniform random number, and α = 1 2 is a constant parameter controlling the peaking of the source intensity toward the axis.

r = R ξ 1 1 − α , (2)

To maintain an unbiased dose estimate, each event is assigned a statistical weight, w, given by Equation 3:

w = 2 1 − α r 1 + α R 1 + α . (3)

This method increases the number of interactions contributing to the dose tally in the small scoring volume, thereby improving the convergence of H * ( 10 ) with fewer simulated histories. The biasing was used only for the monoenergetic beam validation to verify the Geant4 implementation of H * ( 10 ) conversion coefficients. All subsequent facility shielding simulations were performed without any variance-reduction biasing to preserve the true fluence and dose distributions.

The same process is repeated for gamma and neutron sources, with results shown below.

2.3.1 Gamma coefficients

Gamma simulations were run in batches of 20 × 1 0 6 and required fewer histories to achieve good statistical certainty. The simulation results are shown in Figure 3 and summarised in Table 2.

TABLE 2. Comparison of gamma ICRP74 H * ( 10 ) ambient dose equivalent conversion coefficients, in ( p S v c m 2 ), to Geant4.

2.3.2 Neutron coefficients

Neutron simulations were run in batches of 40 × 1 0 6 event histories, each run uniquely seeded, until the relative standard error (RSE) of the simulated H ^*(10) fell below 10%. The RSE quantifies the stochastic uncertainty of the Monte-Carlo simulation and is independent of the level of agreement with reference data. At this level of statistical precision, Monte-Carlo noise does not significantly impact the comparison with reported values. The results are shown in Figure 4 and summarised in Table 3.

TABLE 3. Comparison of neutron ICRP74 H * ( 10 ) ambient dose equivalent conversion coefficients, in ( p S v c m 2 ), to Geant4.

The Geant4-calculated neutron coefficients exhibit their largest deviations from the ICRP 74 reference values in the epithermal energy range, where neutron transport is highly sensitive to thermal scattering effects, resonance structures, and nuclear interactions.

The H * ( 10 ) fluence-to-dose conversion coefficients for both neutrons and photons showed acceptable agreement with the ICRP 74 reference data, with deviations of less than < 15 % in the energy range 0.01 − 3MeV for photons and 0.1 − 10MeV for neutrons. At lower neutron energies, 1 × 1 0 − 5 − 0.01MeV, the mean conversion coefficient deviates by 20 − 40 % of the tabulated data. At gamma energies above approximately 3MeV, the derived H * ( 10 ) coefficients start to plateau and slightly decrease. This behaviour is consistent with full charged-particle transport calculations that arise from the increasing range of the secondary electrons relative to the 10 mm reference depth used in the definition of H * ( 10 ). The observed differences are known limitations of the H * ( 10 ) as an operational quantity at high gamma energies (Sato et al., 1999).

Because of this, dose estimates in sensitive volumes were calculated using both the derived fluence-to-dose conversion coefficients and the tabulated ICRP 74 coefficients to assess the impact of the conversion methodology on the dose measurements.

2.4 Radiation safety limits

National and international radiation protection guidelines define permissible dose limits, which vary according to the regulatory framework, the degree of control of the area, and the classification of the exposed personnel.

An overview of some of these limits is provided below.

Dose limits are quoted per year, which can be approximated based on a 2,000-h work year to be equivalent to 10 μ S v h − 1 for workers, 3 μ S v h − 1 for students and 500 n S v h − 1 for members of the public.

Table 4 shows that the US NRC Dose limits differ from the ICRP, EU, and UK IRR17 limits, which are themselves consistent. The US NRC allows a higher annual effective dose limit 50 msV/yr compared to 20 msV/yr), and a slightly smaller student limit (5 msV/yr compared to 6 msV/yr). These differences are largely due to historical regulatory development (NPA, 2022). It is worth noting that all the regulatory bodies mentioned are aligned with the dose limit of 1 m s V / y r for members of the public.

TABLE 4. Regulatory dose limits (Vennart, 1991; Expert Participation, 2017; NRC, 2026).

The classification of an individual as a “worker” or a member of the “public” depends on the employer’s role designation and an adequate assessment of the potential radiation dose to which the person may be exposed. This paper will distinguish between worker and public dose limits when analysing ambient dose rates.

2.5 Facility setup

The simulated facility represented a rectangular laboratory space measuring 4 m × 3 m x 5.5 m (Bakr et al., 2024b), with a 1 m concrete floor, as shown in Figure 5.

The shielding walls were modelled as a boundary with a fixed central position. The layout was designed to emulate a realistic experimental environment, maintaining sufficient workspace around the neutron source, while constraining dose rates to within permissible limits.

2.6 Shielding materials

The elemental composition of concrete varies depending on its use case, the manufacturer, and the chemical composition of the surrounding geography from which the materials are extracted. For this paper, reference concrete constitutions will use compositions as defined by PNNL (Detwiler et al., 2021).

The simulations will also test the effectiveness of standard water G 4 _ W A T E R (GEANT4 Collaboration, 2026a), as shown in Table 5.

TABLE 5. Elemental composition (mass%) of reference concretes (rounded to 2 d.p.).

2.7 Radiation attenuation

2.7.1 Gamma attenuation

Gamma attenuation can be described by the exponential attenuation law I ( x ) = I 0 e − μ ( E ) x where I 0 is the initial gamma flux, I ( x ) is the transmitted intensity after a distance x, and μ ( E ) is the energy-dependent linear attenuation coefficient.

2.7.2 Neutron attenuation

Neutrons do not interact with atoms in materials through standard electromagnetic processes. Instead, a neutron loses energy through a range of nuclear processes, including elastic scattering, inelastic scattering, and neutron capture, depending on the material.

The fractional energy loss of a neutron in an elastic scattering event is governed by classical kinematics, ranging from 0 to E Max depending on the scattering angle, θ, and the mass of the target nucleus, A, with maximum energy transferred occurring at a scattering angle of π, as given by Equation 4:where E n is the incident neutron energy.

Δ E = E n 1 − A − 1 2 A + 1 2 (4)

The relationship shows that to moderate the energy of the neutrons and ultimately capture them, the most effective material would contain atoms with low Z numbers. Common moderators include concrete, water, and high-density polyethylene (HDPE). To subsequently capture the thermalised neutrons, neutron-absorbing additives such as boron are often incorporated into the shielding material.

An example of the shielding process is shown in Figure 6.

The figure illustrates the generation of neutrons in a setup similar to that described previously, with neutrons initially having an energy of 2.45 M e V, E D D. These travel through the containment vessel and may undergo scattering in the device structure, resulting in partial attenuation and an energy reduction. After leaving the device, the neutrons traverse the surrounding air with a reduced energy of E 1, until they interact with the shielding material. The shielding causes the neutrons to undergo multiple scattering interactions, further moderating the neutrons and increasing the likelihood of neutron capture, with some fraction of the neutrons making it through.

A similar exponential attenuation law can be used to approximate the decrease in flux of neutrons with a specific energy. However, because neutrons undergo many scattering events, changing their direction and energy, the resultant energy spectrum is more complex. Modern neutron transport theory, based on Boltzmann’s steady-state neutron transport equation, treats this process statistically, describing how neutrons are transferred between energy groups and spatial regions according to material and energy-dependent cross sections. In this study, these transfer functions were not derived analytically; instead, their effects were realised through Monte Carlo transport simulations, which numerically reproduce the full scattering and absorption processes.

2.8 Simulation dose measurement

The process of calculating the ambient dose for a specific shielding type and shielding thickness is summarised in the process flow shown in Figure 7.

The fluence, Φ, is determined using the track length and is defined as the summation of particle track lengths per unit volume, as given by Equation 5:where L i is the i-th chord length traversed by a particle, V is the volume of interest and Φ is the total flux given in units of n cm − 2. A schematic of the track-length estimation method is shown in Figure 8.

Φ = 1 V ∑ i L i n/cm 2 (5)

This interpretation is often preferred over surface fluence when conducting Monte Carlo simulations and ambient dose assessments in complex geometries, as it avoids reliance on surface crossings and provides a geometry-independent measure of particle exposure within a volume.

Track-length fluence was measured using the G4PSCellFlux scorer (GEANT4 Collaboration, 2026b), with separate filtering for neutrons and photons. Fluence spectra were extracted by integrating the total track length in logarithmically spaced energy bins.

H * ( 10 ) can then be calculated using the measured fluence and the tabulated ICRP H * ( 10 ) conversion coefficients, as given by Equation 6 (ICRP International Commission on Radiological Protection, 2007; Casanovas et al., 2016; IAEA, 2021):where

H * 10 = ∫ 0 ∞ ϕ E h * 10 ; E d E , (6)

• ϕ ( E ) is the differential fluence spectrum (n cm − 2 MeV − 1),
• h * ( 10 ; E ) is the fluence-to-ambient dose equivalent conversion coefficient (pSv cm 2),
• H * ( 10 ) is the ambient dose equivalent (pSv).

In practice, this can be achieved by converting this into a discretised form, as shown in Equation 7:where

H * 10 ≈ ∑ i ϕ E i h * 10 ; E i Δ E i , (7)

• ϕ ( E i ) is the differential fluence in the ith energy bin (n cm − 2 MeV − 1),
• h * ( 10 ; E i ) is the interpolated fluence-to-ambient dose equivalent conversion coefficient for the ith energy bin (pSv cm 2),
• Δ E i is the width of the ith energy bin (MeV),
• H * ( 10 ) is the ambient dose equivalent (pSv).

2.9 Simulation setup

All simulations were conducted using the F T F P _ B E R T _ H P physics list, which utilises high-precision (HP) neutron data for energies 0–20 MeV and thermal scattering laws ( S ( α , β ) ) where available (e.g., water, polyethylene) for neutron energies 0–4 eV, where applicable. This dataset samples data from various nuclear cross-section databases to statistically determine the process and physics for a particle’s interactions as it traverses the simulation.

The objective of the setup was to determine the ambient dose delivered by a DD neutron source operating at a neutron production rate of 1 × 1 0 5 s while varying the amount of shielding between the source and the measurement volumes at a stand-off distance of 1.6 m.

The simulation is based on the previously described 4 m × 3 m × 5.5 m room, with a 1 m thick concrete floor.

A neutron source is placed at the centre of the containment chamber. The source is surrounded by shielding walls as depicted in Figure 9.

Surrounding the wall are sensitive volumes, configured to record the track length flux of neutrons and photons for H * ( 10 ) calculations at floor and torso heights (1 m), around the containment chamber.

The sensitive volumes are in a fixed position between runs, and the distance from the neutron source to the central point of the walls is fixed. Shielding widths are varied between 10 cm, 20 cm, and 30 cm, with a 0 shielding run for reference.

2.9.1 Neutron source

The neutron source was simulated as a 17 cm inner vacuum chamber, surrounded by a 3 mm thick inner titanium chamber wall, 1 cm water jacket, and 3 mm outer titanium wall as outlined in (Bakr et al., 2024a). A diagram of the simulation is shown in Figure 10.

The neutron source was simulated as an isotropic point source of 2.45 MeV monoenergetic neutrons. Because the process of DD fusion is not modelled, there will be no tritium buildup, and therefore, no DT interactions modelled or proton interaction. The neutrons were partially moderated by the titanium chamber and water shell, resulting in a slightly attenuated spectrum with lower energy components. An example of the spectrum is shown in Figure 11.

The laboratory floor was modelled as a solid block of concrete, with a chemical composition as defined by the NIST Database G4_Concrete. Neutrons will elastically scatter off the lighter atoms in concrete, such as oxygen, hydrogen, and carbon, causing them to lose energy. A significant fraction of neutrons will scatter from the floor back up into the experimental area. The remaining neutrons will either leave the simulation volume or continue to scatter deeper into the concrete, slowly thermalising until they are completely stopped.

3 Results

Each shielding configuration was run with 10 independent seeds for 12 million event histories (resulting in a total of 120 million histories) to reduce statistical uncertainty.

A heat map of the dose received at floor level when unshielded and with 30 cm of Borated Concrete is shown in Figures 12, 13.

The image illustrates the impact of a 1.6m stand-off distance on the dose, showing that a source operating at 1 × 1 0 5 n s − 1 results in a total dose of 560 n S v h − 1.

The attenuation of the concrete significantly reduces the dose from 560 to 93 n S v h − 1.

Figure 16 illustrates the dose distribution in the vicinity of the shielding enclosure exits. In shielded configurations, multiple scattering off the surrounding walls moderates neutrons and produces a diffuse neutron field within the room. In the example shown (30 cm of Borated concrete), this results in an increase of 600 n S v h − 1 to 876 n S v h − 1 (46% increase) in areas approaching the exit, consistent with the room-return effects. Additional monitoring volumes placed at the room exit show that, although dose levels are observed inside the enclosure, the ambient dose equivalent at the room exit is still significantly reduced.

Excluding the exit, the largest dose is deposited in sensitive volume number 43 as denoted in Figure 9. The neutron and gamma doses for this volume are extracted and shown in Figures 14, 15, which compare the unshielded case with those obtained by varying the shielding type and thickness.

The impact of the shielding can be seen in the reduction of high-energy fluence. For thinner shielding widths of 10 cm and 20 cm, there is a noticeable shift of the neutrons to more epithermal energies. The epithermal fluence then drops off with further increases in shielding.

The results show that hydrogen-dense materials such as water and ordinary concrete (OC1) are the most effective neutron moderators tested. This can be seen in the sharp reduction of higher-energy, 2.45 MeV, neutrons achieved through elastic scattering on the light hydrogen nuclei, shifting the spectrum toward lower energies where capture reactions dominate. Additionally, the results show a reduction in thermal neutron flux when working with borated concrete, due to the large neutron capture cross-section of boron. In contrast, high-density materials such as Heavy Concrete (high Ba composition) and magnetite concretes attenuate photons well, but moderate neutrons less effectively. This is due to neutrons losing less energy per collision when colliding with the heavier atoms.

Furthermore, an increased peak can be seen in the gamma flux when employing a water shield due to the neutron capture interaction ¹ H ( n , γ ) ² H generating a 2.22 MeV gamma-ray.

All these factors, along with source strength, need to be taken into consideration when designing an effective shielding solution.

Limits of safe public dose limits are shaded green, safe for students are yellow, safe for workers are orange, and above the worker safety limit red.

Figure 16 shows that at a radiation facility, workers can avoid potentially harmful radiation at a standoff distance of 1.6 m if the neutron intensity is kept low 1 × 1 0 5 n s − 1 - 1 × 1 0 6 n s − 1. Higher neutron intensities of > 1 × 1 0 7 n s − 1 can likely be made safe by introducing shielding of at least 30 cm for most concrete compositions. The results also so show that some amount of shielding should be present if operating in around members of the public.

Table 6 shows the ambient dose equivalent values comparing ICRP 74 fluence-to-dose conversion coefficients with those derived from Geant4 for neutrons and photons. The neutron dose results within approximately 6 % of the ICRP values, with a trend of decreasing difference with increasing shield width, indicating that the large deviations in the epithermal conversion coefficients have less effect on the integrated dose. Additionally, as shielding thickness increases, the neutron contribution to the total dose decreases rapidly, whereas the gamma contribution is less impacted, leading to an increasing importance of monitoring the gamma dose. This trend is shown more clearly in water and OC1, where the gamma dose component contributes approximately 30 % and 10 % of the total ambient dose equivalent, respectively.

TABLE 6. Comparison of neutron and gamma dose equivalents, in n s V h − 1, for different shielding types and widths.

Water remains an extremely cost-effective shielding material for neutron sources, although it will generate additional gamma rays through neutron capture reactions with hydrogen atoms. It would be impractical to construct a full chamber and ensure no undesired shine paths are present. Therefore, a composite approach may be the best.

4 Simulation limitations

This study assumes a monoenergetic 2.45 MeV DD neutron source. It does not account for the angular spread of the energy range, typically ± 40 keV, or the small high-energy component that can arise from DT buildup in long-running IECF devices. The laboratory geometry is idealised, with uniform shielding walls and no holes, cable ducts, or other shine paths. The sensitive regions represent a limited set of points of interest. Material compositions are based on reference concrete materials rather than site-specific aggregates. These simplifications are suitable for comparative shielding analysis, but may result in small deviations from measured dose rates in actual facilities.

5 Conclusion and future work

In this paper, the methods section outlines the processes used in the simulation software Geant4 to recalculate the H * ( 10 ) Ambient dose coefficients for gamma and neutrons, and the results are compared to tabulated values provided by the ICRU. The paper then documented regulatory requirements, discussed the concepts of radiation attenuation and fluence measurements, and defined a fluence-to-dose conversion methodology. Shielding materials were defined, and the simulation setup was discussed.

The results section first presents the impact of no shielding at a standoff distance of 1.6 m. Then it showcases the effects of increasing shielding on fluence and dose measurements in specific cells, accompanied by a full dataset in the supplementary material.

The data is then aggregated into a graph that shows the impact of neutron production rate and shielding, highlighting where the ambient dose would pass regulatory limits.

The results show that the dose received from a DD neutron generator can be deemed safe and well below regulatory for worker limits when operated at neutron yields of 1 × 1 0 6 n s − 1 with a sufficient standoff distance (1.6 m) or at higher rates of 1 × 1 0 7 n s − 1 if appropriate levels of shielding are present. The model can now be used to provide a basis for future optimisation of IECF shielding, including mixed DD/DT operation and real laboratory geometries.

This work focused on single-material homogeneous shielding materials. However, practical IECF installations may benefit from composite shielding concepts that rely on multiple layers of materials. A layered approach could look to combine hydrogen-rich materials to moderate neutrons and boron-rich concrete layers for thermal neutron absorption and gamma attenuation of secondary photons. This could lead to a reduction in overall shielding thickness while maintaining low dose levels.

Future work will explore the optimisation of material selection and layer thicknesses to balance neutron moderation, absorption, and gamma attenuation for compact IECF shielding systems. As well as extending the model to more complex laboratory geometries and benchmarking simulated dose rates against experimental measurements. These developments will further enhance confidence in simulation-based design guidance for IECF shielding applications.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. Generative AI was used for grammar refinement and to improve clarity of expression. All scientific content, analysis, and conclusions were developed and verified by the authors. Generative AI assisted in LaTeX formatting. Generative AI was also used to debug simulation code.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Gokhan Kilic, Eskişehir Osmangazi University, Türkiye

Reviewed by: Hesham M. H. Zakaly, Al-Azhar University, Egypt

Frederic Stichelbaut, Ion Beam Applications, Belgium