Introduction

As a pivotal medical radioisotope, ^99mTc is extensively utilized in clinical diagnostics [1, 2], accounting for approximately 80% of global nuclear medicine radioisotope consumption [3]. ^99mTc is primarily derived from the β^- decay of molybdenum–99 (⁹⁹Mo, T_1/2≈65.92 h). Consequently, the stability of the ⁹⁹Mo supply directly determines the market availability of ^99mTc. According to estimates by the Organization for Economic Cooperation and Development/Nuclear Energy Agency (OECD/NEA), the global demand for ⁹⁹Mo reached 5.5×10⁵ Ci in 2020 and is growing at over 10% annually, attaining 8.8×10⁵ Ci by 2030 [4]. Furthermore, according to the “Medium-to-Long-Term Development Plan for Medical Isotopes (2021–2035)”, China’s annual demand for ⁹⁹Mo has exceeded 1.6×10⁴ Ci since 2019, with a steady 5% annual growth rate.

Currently, the global supply of medical ⁹⁹Mo is predominantly produced using research reactors in Europe, Canada, Australia, and South Africa, etc. [5, 6]. Owing to the significant nuclear proliferation risks associated with the use of highly enriched uranium (HEU, ²³⁵U enrichment >90%), its application is restricted [7, 8]. Therefore, most research reactors have now transitioned to irradiating low enriched uranium (LEU,²³⁵U enrichment <20%) [9]. Most research reactors have exceeded 40 years of operating time and will face shutdown and decommissioning over the next decade [2, 10], potentially triggering a severe “⁹⁹Mo shortage crisis” that would critically impact ⁹⁹Mo-importing nations. To sustain the supply of ⁹⁹Mo, the High Flux Reactor (HFR) in Petten, Netherlands, originally scheduled for shutdown in 2022, had its closure delayed [11]. However, in 2024, due to pipeline deformation above the reactor vessel, it failed to restart as planned after a routine maintenance shutdown, once again causing issues in the global ⁹⁹Mo supply chain [12]. Meanwhile, Belgium’s BR2 reactor and South Africa’s Safari-1 reactor are extending their operational cycles to address the shortage [13, 14].

The international community is committed to developing new sustainable ⁹⁹Mo production methods. In 1992, Chopela and Ball [15, 16] from Babcock & Wilcox (B&W) proposed a Medical Isotope Production Reactor (MIPR). Although this technology has attained commercial viability with small-scale industrial implementation [17-20], MIPR face significant power fluctuations during operation owing to void, chemical, and temperature effects, risking control instabilities [21].

Molten salt reactors (MSRs) use molten salt mixtures as both fuel and primary coolant, offering significant advantages for isotope production, such as high production efficiency and online extraction capabilities [22, 23]. Consequently, global research efforts have increasingly focused on MSR-based ⁹⁹Mo production with online extraction systems. Studies have confirmed the technical feasibility and promise of MSRs for isotope generation [24, 25], revealing unique behaviors of fission products (e.g., Mo): these products spontaneously migrate from molten salt phase into gas phase, with at least 50% of Mo fragments accumulating in the aerosol phase above the salt surface [26, 27]. This behavior forms the foundation for ⁹⁹Mo production in MSRs, leading to further evaluations of online ⁹⁹Mo extraction feasibility from MSRs [22, 28]. Nevertheless, MSR-based isotope production remains predominantly in the theoretical validation and exploratory stages, with critical scientific and technical challenges requiring resolution.

In 2009, the OECD/NEA convened a meeting in Paris to address global ⁹⁹Mo/^99mTc supply security, establishing the High-Level Group on Medical Isotopes to ensure sustainable global access to these isotopes [6]. Among the various technological approaches, accelerator-based production routes, which either generate ⁹⁹Mo or directly yield ^99mTc, are recognized as viable medium-to-long-term alternatives owing to their regulatory flexibility, non-proliferation benefits, and low infrastructure/operating costs [29].

Accelerator-based production of ⁹⁹Mo/^99mTc primarily employs two approaches: direct irradiation of production targets with primary particles (e.g., protons, electrons, deuterons, α-particles) or irradiation of converter targets to generate secondary particles (neutrons/photons) for bombarding production targets. Viable pathways include: ¹⁰⁰Mo(γ,n)⁹⁹Mo, ¹⁰⁰Mo(p,2n)^99mTc, ¹⁰⁰Mo(p,pn)⁹⁹Mo, ⁹⁶Zr(α,n)⁹⁹Mo, ⁹⁸Mo(n,γ)⁹⁹Mo, ¹⁰⁰Mo(n,2n)⁹⁹Mo, ²³⁵U(n,f)⁹⁹Mo, U/Th(p,f)⁹⁹Mo and ²³⁸U(γ,f)⁹⁹Mo, among others [2, 9, 30-32]. Figure 1 summarizes these accelerator-based production technologies [6].

Fig. 1Full-screen View

(Color online) Accelerator–based production technologies for ⁹⁹Mo/^99mTc

Proton beam irradiation of natural uranium or thorium-232 (²³²Th) can produce ⁹⁹Mo via a (p,f) reaction; however, the reaction cross-sections are low [2]. Alternatively, proton irradiation of ¹⁰⁰Mo generates ^99mTc directly through (p,2n) reaction or produces ⁹⁹Mo via (p,pn) reaction. This approach achieves high ^99mTc yields with lower proton energy requirements (E_p<30 MeV). However, the short half-life of ^99mTc (T_1/2≈6.01 h) restricts long-distance transportation, thus it can only be supplied to surrounding areas. Notably, the competing reaction ¹⁰⁰Mo(p,2n)^99gTc produces significant quantities of ^99gTc, compromising radiopharmaceutical purity and imaging efficacy [33, 34]. Secondary particle methods utilize protons or electrons hitting converter targets to generate neutrons/photons, which then irradiate ¹⁰⁰Mo or ⁹⁸Mo to produce ⁹⁹Mo. Nevertheless, molybdenum-based production routes (involving enriched ¹⁰⁰Mo, ⁹⁸Mo, or natural molybdenum targets) face three critical challenges: (a) low specific activity (0.35–15 Ci/g) owing to the molybdenum carrier component, requiring complex separation processes—significantly below clinical generator requirements (20–500 Ci/g) [35, 36]; (b) prohibitive target costs owing to low yields from natural molybdenum, requiring enriched targets (⁹⁸Mo or ¹⁰⁰Mo at 500–1500 USD/g) [29, 37]; (c) minuscule utilization efficiency (only 0.001% to 0.0001% of Mo atoms are transformed), mandating costly recovery systems and increasing operational expenditures [29, 35, 37, 38].

Building on the MIPR concept, the United States proposed an accelerator-driven subcritical solution reactor for isotope production [39, 40]. However, this method requires high-intensity, long-term stable accelerator neutron sources, posing significant challenges, such as considerable R&D complexity, prohibitively expensive accelerator infrastructure, and high neutron production costs. In 2021, Han et al. [40, 41] proposed an accelerator-based deuterium-tritium (D-T) neutron source subcritical system for producing ⁹⁹Mo. However, this method requires the use of tritium, which is not only costly but also presents challenges in obtaining possession and operation licenses for tritium. Consequently, Han et al. [9] introduced a novel design for ⁹⁹Mo production based on a LUE subcritical blanket system (SBS). This system is driven by a Gas Dynamic Trap-based Fusion Neutron Source (GDT–FNS) and utilizes neutrons from the deuterium-deuterium (D-D) fusion reaction to induce the fission of ²³⁵U. Alternatively, ⁹⁹Mo can be produced by α-particle irradiation of ⁹⁶Zr via (α,n) reaction. To increase the yields, enriched ⁹⁶Zr targets must be employed, further increasing the production expenses [42].

Ruth [43] proposed photofission of ²³⁸U targets using electron accelerators for ⁹⁹Mo production, emphasizing that while ⁹⁹Mo yield via ²³⁸U(γ,f) reaction is significantly lower than ²³⁵U(n,f) production, this is outweighed by the advantage of using safer materials. Naik et al. [44] also validated ²³⁸U(γ,f) as a viable ⁹⁹Mo production route. Concurrently, Naik [45], Thierens [46], and Schmitt [47] measured cumulative ⁹⁹Mo fission yields from ²³⁸U at bremsstrahlung energies of 10 MeV, 25 MeV, and 48 MeV as 4.835%, 6.480%, and 6.600%, respectively. The cumulative yield of ⁹⁹Mo increases with the incident electron energy. In contrast, fast–neutron–induced fission of ²³⁸U achieves a ⁹⁹Mo cumulative yield of 6.168% [48], demonstrating the comparable efficiencies between photofission and fast-neutron-induced fission. Collectively, these studies establish the technical viability of ⁹⁹Mo production via ²³⁸U(γ,f) reaction. With its inherently high ²³⁸U abundance (99.27%), natural uranium emerges as a strategically advantageous target material. The adoption of natural uranium eliminates the enrichment requirements for ²³⁸U, substantially reduces target fabrication costs, and minimizes proliferation risks. Naik et al. [44] irradiated a 1 g natural uranium target for 24 hours with bremsstrahlung photons generated from a 10 MeV, 4 kW electron beam (10 Hz repetition rate) on a tantalum foil, achieving ⁹⁹Mo production of (0.309±0.050) mCi-confirming natural uranium’s practical utility.

Current accelerator-based ⁹⁹Mo production predominantly employs solid targets, facing persistent challenges such as complex fabrication, cutting, and dissolution processes, high costs, inability for online extraction, and inadequate heat dissipation. To overcome these limitations, this study integrates the synergistic advantages of liquid molten salts and electron accelerators to propose a method for ⁹⁹Mo production via electron beam irradiation of natural-uranium-bearing molten salt targets. This approach utilizes the ²³⁸U(γ,f) reaction to produce ⁹⁹Mo, offering significant advantages such as the elimination of HEU, significantly reducing proliferation risks; avoidance of solid-target processing through fluid targets; online extraction and simplified reprocessing, yielding high specific activity of ⁹⁹Mo; continuous irradiation capability to enhance uranium utilization efficiency; direct irradiation without converter targets, reducing costs and structural complexity; inherent flow–based cooling, reducing cooling requirements and enabling higher beam intensities for increased yields; and lower target fabrication costs with economical molten salts.

A distinctive feature of molten salt targets, compared to solid targets, is their fluidity. This characteristic also introduces numerous new challenges for target system design, including target size and structural design, molten salt circulation system, thermal management and temperature control, as well as material corrosion and compatibility. To evaluate the feasibility of this approach, this study employs Monte Carlo methods to systematically simulate and analyze the critical operational parameters of the ⁹⁹Mo production system, including molten salt composition, target geometry, incident electron beam energy, temperature distribution, and online extraction efficiency. Based on the parameter analysis, the theoretical annual production capacity of ⁹⁹Mo was quantitatively projected and comparatively assessed against China’s demand. These findings will provide essential references for the engineering implementation of this production methodology.

Materials and methods

2.1

Monte Carlo model

FLUKA has been validated as a reliable tool for modeling radioisotope production [49-51]. Consequently, this study utilized FLUKA to simulate and analyze ⁹⁹Mo generation via electron irradiation of molten salts. Complementarily, coupled electron-photon transport and bremsstrahlung analyses were conducted using pyPENELOPE [52]. The temperature distribution of the molten salt was analyzed using ANSYS Fluent software [53].

A schematic of ⁹⁹Mo production via electron beam irradiation of a molten salt target is shown in Fig. 2, comprising two core components: an electron beam and a molten salt target assembly. The electron beam features a radius of 1 cm, an energy range of 20–160 MeV, and an intensity of 1 mA, incident along the positive Z-axis. The molten salt target assembly comprises two concentric layers: an inner molten salt layer and an outer cladding layer. The internal cavity was filled with molten salt (Fig. 2b), while the outermost cladding is a 0.5 cm thick nickel-based alloy. Furthermore, to conduct thermal analysis of the molten salt target, the model incorporates the flow process of the molten salt. Accordingly, the inlets and outlets for the molten salt were configured (Fig. 2c). Along the Y-axis, the molten salt enters from the bottom and exits from the top of the receiver.

Fig. 2Full-screen View

(Color online) Schematic of ⁹⁹Mo production via electron irradiated molten salt target: (a) Three-dimensional target, (b) XZ-section, (c) Three-dimensional target (with Inlet and Outlet)

2.2

Reaction models and cross-sections

Under electron beam irradiation, the incident electrons initially generate bremsstrahlung photons through interactions with the constituent atoms of the molten salt target. These photons subsequently induce direct (γ,f) reactions with uranium to produce ⁹⁹Mo or alternatively undergo photonuclear reactions (γ,xn) to generate secondary neutrons. The latter initiates (n,f) reactions on uranium atoms, thereby yielding additional ⁹⁹Mo. This reaction cascade is illustrated schematically in Fig. 3. Consequently, ⁹⁹Mo in molten salt primarily originates from fission of both ²³⁵U and ²³⁸U, encompassing: direct fission products and decay of short-lived precursor nuclei. Direct fission pathways include both photofission and neutron-induced fission of ²³⁵U and ²³⁸U, with key precursors nuclei ⁹⁹Nb (T_1/2=15.0 s) and ^99mNb (T_1/2=2.5 min). Given that their half-lives are negligible compared to ⁹⁹Mo, fission-generated precursors are assumed to completely decay to ⁹⁹Mo during irradiation.

Fig. 3Full-screen View

(Color online) Schematic of the physical process for ⁹⁹Mo production via electron irradiated molten salt

Figure 4 displays the (γ,f) reaction cross-sections for ²³⁵U and ²³⁸U extracted from the ENDF/B-VIII.1 database [54]. Within the 10–20 MeV photon energy range, both ²³⁵U and ²³⁸U exhibit significant (γ,f) cross-sections. ²³⁵U peaks at 0.33 b (Eγ=14 MeV), and ²³⁸U peaks at 0.16 b (Eγ=14.5 MeV), indicating feasibility with relatively low electron/photon energies. Neutron-induced fission also contributes to ⁹⁹Mo production, where secondary neutrons are predominantly generated via photonuclear reactions. Figure 5 illustrates the (γ,xn) cross–sections for ²³⁵U and ²³⁸U, both peaking within 10–20 MeV (²³⁸U: 0.73 b at 14 MeV; ²³⁵U: 0.52 b at 14.5 MeV). The comparable photofission and photonuclear reaction cross-sections between ²³⁵U and ²³⁸U demonstrate low ²³⁵U enrichment requirements, enabling the use of natural uranium.

Fig. 4Full-screen View

(Color online) (γ,f) reaction cross-sections for ²³⁵U and ²³⁸U

Fig. 5Full-screen View

(Color online) (γ,xn) reaction cross-sections for ²³⁵U and ²³⁸U

Irradiation Parameters Analysis

The ⁹⁹Mo yield is primarily determined by the electron energy, beam intensity, irradiation time, molten salt composition, and dimensions. Because the per-particle yield remains constant, the total yield increases proportionally with the beam intensity. Consequently, in subsequent simulations, the beam intensity was fixed at 1 mA to focus on analyzing the impact of other parameters. In Sect. 3.1, composition analysis reveals that the LiF-UF₄ salt achieves an optimal ⁹⁹Mo yield (5.33×10⁹ Bq) under 40 MeV electron irradiation for 1 h in a 5 cm × 6 cm × 10 cm volume. For LiF-UF₄ systems, electron energies of 40–80 MeV are most efficient. Then, Sect. 3.2 evaluates ⁹⁹Mo spatial distributions under 40 MeV electrons and LiF-UF₄ salt. The optimized geometry (74 cm × 74 cm × 46 cm) yielded 8.78×10⁹ Bq after 1 h of irradiation. Finally, to ensure that localized boiling does not occur during the operation of the molten salt target system, the maximum temperature of the molten salt was determined. The results indicate that when the flow velocity of the molten salt is increased to 5 m/s under 80 MeV energy conditions, the maximum temperature is only 808.9 K. This effectively avoids the risk of boiling and ensures safe and stable operation of the system.

3.1

Molten Salt Composition and Electron Energy Analysis

The molten salt composition is a critical factor influencing ⁹⁹Mo yield. Compared to enriched molybdenum targets, molten salts offer significantly lower costs, for example, base salts (e.g., LiF-BeF₂) cost 71.23 USD/kg [56], and natural uranium is approximately 200 USD/kg [57]. This cost advantage enables substantial reductions in the target fabrication costs. This section investigates ⁹⁹Mo yields across various molten salt compositions, with the parameters listed in Table 1. The system employs 99.95% ⁷Li enrichment, 97% ³⁷Cl enrichment, natural uranium, and an operating temperature of 923 K.

Table 1Full-screen View

Molten salt composition and physical properties

Molten salt	Composition (mol%)	Melting point (K)	Density (g/cm3)	U content (mol%)
LiF-UF4	72–28	763	4.79	9.86
LiF-BeF2-UF4	71.7–16–12.3	773	3.46	4.86
NaF-UF4	72–28	896	4.22	9.86
NaCl-UCl3	55–45	870	3.60	15.52
NaCl-KCl-UCl3	45–25–30	810	3.08	11.54
NaCl-MgCl2-UCl3	50–33.3–16.7	778	2.50	6.26

Show more

A 40 MeV electron beam irradiated the molten salt target assembly with molten salt dimensions of 5 cm × 6 cm × 10 cm. After 1 h of irradiation, ⁹⁹Mo yields for various salts are presented in Table 2. LiF-UF₄ salt achieves the highest yield at 5.33×10⁹ Bq, followed by NaF-UF₄ at 4.55×10⁹ Bq. NaCl-MgCl₂-UCl₃ demonstrates the lowest yield (1.89×10⁹ Bq). Notably, while the NaCl-UCl₃ salt contains the highest U-content, its ⁹⁹Mo yield remains comparatively low. Conversely, LiF-BeF₂-UF₄ salt, despite having the lowest U-content, achieves higher ⁹⁹Mo yields than both NaCl-KCl-UCl₃ and NaCl-MgCl₂-UCl₃ systems. Crucially, fluoride salts exhibit higher ⁹⁹Mo yields than chloride salts.

Table 2Full-screen View

⁹⁹Mo yields of different molten salts after 1 h of irradiation with a 40 MeV electron beam

Molten salts	99Mo yield (Bq)	Photofission fraction (%)	Energy loss rate (MeV/cm)
LiF-UF4	5.33×109	95.44	22.43
LiF-BeF2-UF4	3.31×109	95.23	12.72
NaF-UF4	4.55×109	96.09	18.59
NaCl-UCl3	4.14×109	94.95	17.05
NaCl-KCl-UCl3	3.16×109	96.90	13.30
NaCl-MgCl2-UCl3	1.89×109	96.75	8.89

Show more

As shown in Table 2, ~95% of ⁹⁹Mo originates directly from photofission (encompassing both direct fission and decay of fission products), indicating that the bremsstrahlung photon yield governs ⁹⁹Mo production. For electrons, the bremsstrahlung energy loss rate satisfies [58]: (10)where E denotes the electron energy, represents the material density, and is the radiation length: (11)For compounds or mixtures, the radiation length is calculated as (12)At an electron energy of 40 MeV, the radiation energy loss rates in different molten salts are listed in Table 2. Concurrently, bremsstrahlung photon spectra generated by 40 MeV electrons in various salts were simulated using pyPENELOPE (Fig. 6). The results revealed that the LiF-UF₄ salt exhibited the highest bremsstrahlung photon yield, whereas the NaCl-MgCl₂-UCl₃ salt exhibited the lowest. This divergence is directly correlated with the radiation energy loss rates, with LiF-UF₄ displaying the maximum radiation energy loss rate, followed by NaF-UF₄, whereas NaCl-MgCl₂-UCl₃ showed the minimum, thereby accounting for the superior ⁹⁹Mo production in the LiF-UF₄ salt.

Fig. 6Full-screen View

(Color online) Bremsstrahlung photon energy spectra produced by 40 MeV electrons in different molten salts

For identical photon energies, the photonuclear reaction cross-sections of uranium slightly exceed its photofission cross-sections (Figs. 4 and 5), indicating that bremsstrahlung photons not only induce photofission but also generate secondary neutrons, whereas uranium fission produces additional neutrons. The spatial distributions of the secondary neutrons generated by 40 MeV electrons in various molten salts are illustrated in Fig. 7. As the penetration depth Z increases, the electron energy gradually diminishes, causing cumulative growth in the bremsstrahlung photon yield and flux, which elevates secondary neutron production. Beyond a critical depth, complete electron energy depletion coupled with photon attenuation diminishes the neutron yield. Consequently, LiF-UF₄ exhibited the highest secondary neutron yield owing to the maximal photon production. Figure 8 displays the normalized energy spectrum of secondary neutrons in LiF-UF₄, wherein fast neutrons dominate with a mean energy of ~1.23 MeV. At this energy, the neutron fission cross-sections for ²³⁵U and ²³⁸U are low (1.21 b and 0.04 b, respectively), resulting in poor neutron utilization efficiency as most neutrons leak from the target, contributing only ~5% to the total ⁹⁹Mo production.

Fig. 7Full-screen View

(Color online) Spatial distribution of secondary neutrons from 40 MeV electrons in molten salts

Fig. 8Full-screen View

(Color online) Normalized energy spectrum of secondary neutrons from 40 MeV electrons in LiF-UF₄ and ^235,238U(n,f) cross-sections

The effects of electron energy on ⁹⁹Mo yield in LiF-UF₄ were further investigated (Fig. 9). Photofission cross-sections of uranium peak within the 5–20 MeV photon energy range. Bremsstrahlung photons exhibit a continuous spectrum: the maximum photon energy approaches the incident electron energy with low intensity, whereas the peak intensity occurs near half the electron energy [59, 60]. To optimize the photofission and photonuclear cross-sections, the electron energies must exceed 40 MeV. ⁹⁹Mo yield increases with electron energy, though growth rates progressively diminish. However, excessively high energies generate energetic γ-rays with reduced reaction cross-sections, increased shielding complexity, and undesirable side-reactions that produce radiochemical impurities. Consequently, the 40–80 MeV range balances the yield efficiency and engineering feasibility.

Fig. 9Full-screen View

⁹⁹Mo yield at varying electron energies

3.2

Molten Salt Dimensions and Temperature Analysis

With a fixed electron beam spot size (e.g., 1 cm), the molten salt target dimensions critically impact ⁹⁹Mo yield. Undersized targets underutilize the beam flux, whereas oversized targets increase salt consumption and costs. FLUKA simulations revealed ⁹⁹Mo spatial distributions via residual nuclei calculations under 40 MeV, 1 mA electron irradiation for 1 h (Fig. 10). The electron beam propagates along the positive Z-direction, perpendicular to the XY-plane. In the XY-plane, the residual nuclei exhibited a concentric distribution from the beam point, with the density decreasing radially. Along the XZ-plane, the residual nuclei diminish exponentially along Z, becoming negligible beyond 90 cm. Statistical analysis confirmed that 98% of the residual nuclei were concentrated within ±37 cm (X/Y) and 46 cm (Z). Thus, the optimal dimensions of 74 cm × 74 cm × 46 cm achieved a peak yield of 8.78×10⁹ Bq.

Fig. 10Full-screen View

(Color online) Spatial distribution of residual nuclei: (a) XY-plane, (b) XZ-plane

For accelerator targets, a core challenge is the effectively removal of the substantial heat generated by beam energy deposition to prevent local overheating of the target [61, 62]. To evaluate the temperature distribution of a molten salt target (LiF-UF₄, 74 cm × 74 cm × 46 cm), the three-dimensional energy deposition distribution of an electron beam (40/80 MeV, 1 mA) in the target was calculated using FLUKA. The energy deposition data were imported into ANSYS Fluent for thermal-hydraulic calculations to obtain a steady-state temperature field. During actual operation, beam energy deposition can form hot spots in local areas, posing a risk of molten salt boiling, thus making efficient heat removal critical. However, owing to the large volume of the molten salt target, the heat deposited by the beam is insufficient to maintain the entire region in a molten state, necessitating additional heating devices to prevent solidification. To simplify the computational model, a constant temperature boundary condition of 773 K was applied to the molten salt region in Fluent.

Figure 11 shows the steady-state temperature distribution at Z = 0.2 cm under 40 MeV and 80 MeV electron beam irradiation. The electron beam is incident vertically with a beam spot radius of only 1 cm, and owing to the short electron range, energy is primarily deposited near the injection point, forming a local hot spot. The maximum temperature point was located at approximately (X = 0 cm, Y = 0 cm, Z = 0.2 cm). The molten salt flowed upward (inlet at the bottom and outlet at the top) and continuously removed heat from the vicinity of the hot spot, resulting in a temperature gradient band along the flow direction. The variation in the maximum temperature within the molten salt target with the inlet flow velocity is shown in Fig. 12. As the flow velocity increased from 0.1 m/s to 0.5 m/s, the maximum temperature decreased from 977.3 K to 805.0 K (40 MeV) and from 1011.0 K to 808.9 K (80 MeV). At the same flow velocity, the 80 MeV beam induced a higher maximum temperature than the 40 MeV beam; however, this temperature difference gradually diminished as the flow velocity increased. Importantly, all maximum temperatures remained well below the boiling point (above 1773 K [63]), indicating that the temperature distribution remained within a safe range. The heat dissipation performance of the flowing molten salt target is significantly superior to that of solid targets, which is attributed to their distinct thermal-hydraulic mechanisms. Solid targets primarily rely on molecular thermal vibrations for heat conduction, which is a single and less efficient heat transfer pathway. In contrast, a flowing molten salt target incorporates an additional forced convection mechanism on this basis, significantly enhancing heat transfer through convective heat exchange. The continuous flow of molten salt efficiently carries away heat, thereby preventing local overheating and ensuring more stable and efficient heat transport [64, 65]. This allows the molten salt target to meet heat dissipation requirements through its flow, eliminating the need for additional cooling devices, simplifying the system structure, and reducing costs.

Fig. 11Full-screen View

(Color online) The steady-state distribution at Z = 0.2 cm within the molten salt target: (a) E_e = 40 MeV, (b) E_e = 80 MeV

Fig. 12Full-screen View

(Color online) Variation of the maximum temperature in the molten salt with the inlet flow velocity

Online extraction computational analysis

Current accelerator-based ⁹⁹Mo production predominantly employs solid targets (e.g., metallic molybdenum, molybdenum oxide, uranium oxide). Solid target fabrication involves homogenizing powders (synthesized via dissolution-oxidation processes) with a carrier matrix, followed by mechanical compaction of the mixture. Isotope extraction requires multiple steps: target disassembly, sectioning, dissolution, wet or dry separation, and purification, yielding a complex workflow [5]. Therefore, leveraging ⁹⁹Mo’s existence as insoluble particles in molten salt and the fluid properties of liquid molten salts, this study adopted a bubbling method for ⁹⁹Mo online extraction analysis.

Based on the optimization results in Sect. 3, a configuration with a 40 MeV electron energy, LiF-UF₄ molten salt (74 cm × 74 cm × 46 cm) was selected. Figure 13 illustrates the variation of ⁹⁹Mo yield with irradiation time. The ⁹⁹Mo yield increased progressively with the irradiation time, whereas the amount of decay also increased. Consequently, production and decay eventually reached equilibrium, saturating the yield at 8.53×10¹¹ Bq. Utilizing this saturated yield and Eq. (6), the ⁹⁹Mo production rate Y under these conditions is calculated as 2.49×10⁶ Bq/s.

Fig. 13Full-screen View

⁹⁹Mo yield variation with irradiation time

Figures 14 and 15 display the residual ⁹⁹Mo activity in molten salts, cumulative extraction quantity, equilibrium time, and saturated extraction yields at different extraction efficiencies. Higher extraction efficiencies reduce residual ⁹⁹Mo activity, accelerate equilibrium attainment, and lower the equilibrium activity levels. At an extraction efficiency of 10^-3 s^-1, equilibrium was achieved in ~65 min. Once the ⁹⁹Mo activity in the molten salt saturates, the extraction quantity per unit time also stabilizes, resulting in a linear increase in the cumulative yield. Saturated extraction yield initially increases with extraction efficiency and stabilizes: at 10^-4 s^-1, it stabilizes at 2.09×10¹¹ Bq/day; at 10^-3 s^-1, it increases to 2.15×10¹¹ Bq/day. Balancing equilibrium time against saturated yield, the optimal extraction efficiency was determined to be 10^-4 s^-1 with an equilibrium time of 10.57 h and a saturated yield of 2.09×10¹¹ Bq/day.

Fig. 14Full-screen View

(Color online) Variation of ⁹⁹Mo activity in molten salt (a) and cumulative extraction quantity (b) over time under different extraction efficiencies

Fig. 15Full-screen View

(Color online) Equilibrium time and saturated extraction quantity at different extraction efficiencies

⁹⁹Mo production estimation and comparison with China’s demand

The stable daily production of ⁹⁹Mo via electron accelerator irradiation of molten salts can support domestic demand. The proposed technology enables continuous operation through the flexibility of electron accelerators and the online extraction capabilities of molten salt targets. Furthermore, through systematic optimization of electron energy, target parameters, and temperature control, efficient supply demand matching and autonomous security for China’s ⁹⁹Mo supply chain can be achieved.

Using optimized parameters from Sect. 3 and Sect. 4 (40/80 MeV electrons, 1 mA beam intensity, LiF-UF₄ molten salt (74 cm × 74 cm × 46 cm), 923 K salt temperature, 10^-4 extraction efficiency), annual ⁹⁹Mo extraction yields were calculated. Under these conditions, the ⁹⁹Mo production rates reached 2.49×10⁶ Bq/s (40 MeV) and 6.35×10⁶ Bq/s (80 MeV). Accounting for cyclic accelerator maintenance (3 days of irradiation + 0.5 days of maintenance cycles), Fig. 16 illustrates the annual ⁹⁹Mo activity and cumulative extraction dynamics. Each cycle features rising ⁹⁹Mo activity during irradiation and decays rapidly during maintenance, with cumulative extraction increasing quasi-linearly. The annual yields differed significantly: 6.53×10¹³ Bq (1764.86 Ci) at 40 MeV versus 1.66×10¹⁴ Bq (4486.49 Ci) at 80 MeV.

Fig. 16Full-screen View

(Color online) Annual dynamics of ⁹⁹Mo activity in molten salt and cumulative extraction quantity: (a) Ee=40 MeV, (b) Ee=80 MeV

As shown in Table 3, compared with other ⁹⁹Mo production methods proposed by researchers, this study shows no significant advantage in yield, only exceeding the result reported by Villa et al. [66]. However, this study employs natural uranium as the raw material, which offers higher safety and easier availability than schemes requiring LEU. Furthermore, studies by Minato et al. [67], Gao et al. [32], Villa et al. [66], and Bondar et al. [68] all utilize highly enriched molybdenum targets and highly enriched zirconium targets, which entail high costs and complex production processes. Therefore, the notable advantages of this study in terms of safety, flexibility, and online production capability are sufficient to compensate for the lower yield.

In clinical practice, the administered activity of ^99mTc varies with specific radiopharmaceuticals, disease types, and patient conditions. For instance, ^99mTc-MDP (Technetium-99m methylene diphosphonate) employed in bone scintigraphy requires a single administered dose of 740–1110 MBq (20–30 mCi) of ^99mTc, which necessitates 34–51 mCi of ⁹⁹Mo [69-71]. In contrast, ^99mTc–DMSA (Technetium-99m dimercaptosuccinic acid) for renal imaging requires 11.1–111 MBq (0.3–3 mCi) of ^99mTc, corresponding to a ⁹⁹Mo requirement of 0.5–5 mCi [72]. Excluding other losses, based on the maximum (51 mCi) and minimum (0.5 mCi) ⁹⁹Mo requirements per dose, a 40 MeV facility can supply approximately 34,605 (high-dose scenario) to 3,529,720 (low-dose scenario) patient doses annually, while an 80 MeV facility can provide approximately 87,970 (high-dose scenario) to 8,972,980 (low-dose scenario) patient doses per year. Furthermore, projections from China’s “Medium-to-Long-Term Development Plan for Medical Isotopes (2021–2035)” indicate that the national ⁹⁹Mo demand in 2030 will reach 2.74×10⁴ Ci. A single 40 MeV accelerator and an 80 MeV unit could satisfy 6.44% and 16.37% of the national demand, respectively. An increase in beam intensity can further reduce the number of facilities required.

To minimize decay losses during transportation and leverage the deployment flexibility and inherent safety advantages of accelerator systems, deploying ⁹⁹Mo production facilities in major cities (e.g., Beijing, Shanghai, Shenzhen, Chengdu, and Wuhan) is recommended. Integrated with existing ¹⁸F distribution networks, this strategy would achieve rapid nationwide ⁹⁹Mo supply delivery, covering 95% of the population within the effective distribution radius [73].

Conclusion

This study proposes an approach for ⁹⁹Mo production via electron accelerator irradiation of natural-uranium-bearing liquid molten salts, primarily exploiting the ²³⁸U(γ,f) reaction pathway. The key advantages of this approach include low nuclear proliferation risk, online extraction capability, and reduced target fabrication costs. Through a systematic feasibility analysis and parameter optimization, the following principal conclusions were drawn:

This study establishes an innovative technical pathway for ⁹⁹Mo production, although it primarily focuses on feasibility analysis and partial critical parameters. Several limitations require further refinement, such as the cladding alloy thickness and molten salt target configuration. The 0.5 cm thick nickel alloy cladding adopted in this study necessitated a rigorous evaluation of its impact on electron beam attenuation. Concurrently, the tentatively proposed cuboid molten salt target configuration warrants a comparative assessment against alternative geometries (e.g., spherical and cylindrical geometries). Future work must integrate shielding design, thermal-hydraulic analysis, and insulation system optimization to achieve systematic structural refinement of molten salt targets. Additionally, the bubbling online extraction system requires an in-depth investigation of the physicochemical properties of ⁹⁹Mo in molten salts. These advancements will establish a more robust foundation for the industrial-scale application of this methodology.