Hybridised production of technetium-99m and technetium-101 with fluorine-18 on a low-energy biomedical cyclotron

Abstract

New modes of production and supply of short-lived radioisotopes using accelerators are becoming attractive alternatives to the use of nuclear reactors. In this study, the use of a compact accelerator neutron source (CANS) was implemented to explore the production of ^99mTc and ¹⁰¹Tc. Irradiations were performed with neutrons generated from a 16.5 MeV cyclotron utilising the ¹⁸O(p, n)¹⁸F reaction during routine ¹⁸F-fluorodeoxyglucose (FDG) production in a commercial radiopharmacy. Natural molybdenum targets in metal form were employed for the production of several Tc isotopes interest via (n, γ) reactions on ⁹⁸Mo and ¹⁰⁰Mo. The production of ^99mTc and ¹⁰¹Tc under these conditions is considered and discussed.

1 Introduction

The capability of non-invasive, internal imaging, static or metabolic, of the various anatomical structures, processes, or the disease and detriment of these, has become the desideratum of diagnostics in the medical community. More than tens of millions of positron emission tomography (PET) and single-photon emission computed tomography (SPECT) procedures are administered every year in the United States of America (U.S.) alone. With the ever-growing demand for these procedures, the pursuit of innovative technologies and processes has become an incessant enterprise for establishing techniques and systems that are more work-efficient, cost-effective, and safety-conscious.

Technetium-99m (^99mTc, t$_{1/2} = 6.007$ h) has been widely used for radiodiagnostic purposes for decades, and it is still one of the most used radioisotopes worldwide constituting approximately 85% of all nuclear medicine procedures conducted. Tc-99m can be produced through various nuclear transmutation methods, but commercially speaking, it is generally derived from molybdenum-99 (⁹⁹Mo, t$_{1/2} = 65.925$ h) via ²³⁵U targets [1]. However, the current commercial production and distribution of ^99mTc rely on a complex supply chain that has proven itself prone to disruptions in years past, which was most recently observed during the SARS-CoV-2 pandemic [2]. Ultimately, this leads to delays in the diagnoses of patients due to postponed imaging procedures as well as the loss of material and capital.

Compact accelerator neutron sources (CANS) have presented themselves in the last few decades as a potential alternative for the decentralised production and distribution of radioisotopes [3]. In this regard, CANS refer to an array of fundamentally different accelerator sources such as cyclotrons, radiofrequency quadrupole (RFQ) accelerators, linear accelerators (LINACs) coupled with a photoneutron converter, electrostatic accelerators, laser-driven sources, and neutron generators. With many of these systems, neutron fluxes upwards of ∼10¹² n/s are achievable [4]. However, an under-utilised source of neutrons is those originating from (p, n) reactions during routine PET radioisotope production, such as for ¹⁸F-based radiopharmaceuticals [5].

For example, over the last few decades routine manufacturing of ¹⁸F-fluorodeoxyglucose (FDG), the most frequently implemented PET agent, has greatly evolved. From the benchtop synthesis of several doses at a time to hundreds of doses manufactured in a single production run with multiple runs daily, the evolution of ¹⁸F[FDG] has become the gold standard for in-house medical radioisotope production. At the present moment, there exists at least one of these facilities with a cyclotron in every single state in the U.S., where states with larger populations, and thus higher demand requirements, may have several. Therefore, the growth in ¹⁸F has signified a correlative increase in potentially available neutrons that are effectively not being utilised.

The use of complementary neutrons generated in a biomedical cyclotron employing the ¹⁸O(p, n)¹⁸F reaction has been proposed for radioisotope production [6], although its application for the production of ⁹⁹Mo / ^99mTc or other Tc isotopes is hardly mentioned in the literature. In one study, ⁹⁹Mo was implemented as a monitor for neutron flux in a Mo-containing multi-component flux wire measurement, although there was no targeted discussion for its production [7]. In another study, Link and Krohn report using neutrons generated during ¹³N production, i.e., ¹⁶O(p, α)¹³N, with an 11 MeV Siemens Eclipse at 30 μA over a duration of 0.5 h for producing ⁹⁹Mo / ^99mTc for teaching purposes. Although only small amounts were generated under these circumstances, the authors propose using (p, n) reactions and longer irradiations for higher output [8]. Considering the information provided in the literature pertaining to neutron production rates and energies, it was of interest to determine whether the simultaneous production ⁹⁹Mo / ^99mTc and ¹⁰¹Mo / ¹⁰¹Tc with ¹⁸F in a biomedical cyclotron was feasible and if it could be a viable option for production and distribution of these medically relevant radioisotopes.

2 Materials and methodology

Irradiations were performed with complementary neutrons generated on a biomedical cyclotron (Fig. 1) during routine ¹⁸F production with a ∼16.5 MeV proton beam operating between 75 μA and 85 μA. The ¹⁸F target material was ¹⁸O[H₂O] with a purity of >98.0% ¹⁸O. High-yield ¹⁸F targets (BTI Targetry) were used [9]. The target bodies are made of Al 6061-T6 with a niobium (Nb) insert and implement a Havar alloy foil (0.04 mm) as a target window. Cooling of the target is performed using water and liquid transfer and pressurisation is used with high-purity helium (He) gas. The ¹⁸O[H₂O] target volume is ∼3.5 mL.

Figure 1

(A) Experimental setup for irradiation of Mo metal samples on a biomedical cyclotron (1) using high yield ¹⁸F targets from BTI Targetry. White arrows indicate the position of the samples to be irradiated on the primary ¹⁸F target (2) located in position 5, vertically located above a second ¹⁸F target in position 2 (3). The top arrow is pointing to the Mo metal block attached directly behind the target, and the bottom arrow is the location of the Mo metal foils behind a Styrofoam moderator (4). (B) Close-up view of the ¹⁸F target and Mo sample configurations from the perpendicular perspective relative to the incoming beam

Several varying geometries including cubes and foils of Mo metal were used for irradiations to determine the production of ⁹⁹Mo. The Mo foil (10 cm × 10 cm × 0.01 cm) with ≥ 99.98% Mo was subdivided into 4-square pieces of dimension (5 cm × 5 cm × 0.01 cm) and approximate masses of ∼2.6–3.0 g. Mo target masses and dimensions are presented in Table 1. Foil samples were positioned underneath the ¹⁸F target in position 5 with a piece of Styrofoam moderator placed between the foils and the target. Mo cubes (volume ∼1 cm³, mass ∼10.3 g) with 99.95% Mo were used to determine production in thicker targets. Mo cubes were placed in small plastic bags and adhered to either the back of the ¹⁸F target directly (Cube 1) or behind the Styrofoam moderator (Cube 2) as shown in Fig. 1.

Table 1 Characteristics of Mo cube and foil samples used in irradiation experiments

Irradiation durations and conditions were logged throughout the course of the workweek. The values generated, such as run time, ¹⁸F yield, and beam current, were used to determine relative neutron production rates and relative total neutron output in order to compare with measured values of activity generated in the Mo samples. Scheduled production of ¹⁸F in the facility involves at least two to three production runs in a typical working day. The cyclotron is equipped with two ¹⁸F targets located in positions 2 and 5 approximately ∼28 cm apart at the rear of the targets as shown in Fig. 1. For each working day, the two ¹⁸F targets were alternated for each corresponding run. Samples were positioned on the back of the ¹⁸F target in position 5 with the sample faces oriented perpendicularly to the direction of the incoming proton beam, as well as underneath the ¹⁸F target with the sample faces oriented parallel to the beam. Several of the samples underneath the ¹⁸F target were also positioned behind two pieces of a Styrofoam moderator to determine the effect of moderation. Although both sample sets were located in the closest proximity to the neutron flux generated on the target in position 5 (Fig. 1), it is acknowledged that the samples were also within the bounds of the irradiation field generated by the ¹⁸F target in position 2 during operation, albeit at a lower neutron flux and different energy regime.

Radioactivity in the Mo foils was quantitatively determined using a NaI-type gamma spectrometer, taking measurements at various time intervals after end-of-bombardment (EOB). The NaI spectrometer was calibrated with a ¹³⁷Cs source from North American Scientific with an activity of 2118 Bq at the time of use. The detector efficiency was determined to be 10.814% for the ∼661 keV peak. It is acknowledged that at lower γ energies (Eγ) better detector efficiencies are achievable, however, due to time and materials constraints of this study an efficiency plot was not established to determine this. The activities of ⁹⁹Mo, ^99mTc, and ¹⁰¹Tc were determined using the gamma emissions at 181.1 keV (I$\gamma = 6.05$%), 140.5 keV (I$\gamma = 89.4$%), and 306.8 keV (I$\gamma = 89.4$%). The absolute activity (A) of ⁹⁹Mo in several samples was calculated as a function of integrated counts under the corresponding peaks (C) and adjusted for background counts (B) at the region of interest (ROI), detector efficiency (ε), decay time after EOB until the measurement ($T_{d}$), decay during the measurement time ($T_{c}$), and the associated γ-ray emission probability ($I_{\gamma}$) and decay constant (λ) with each distinct isotope using Eq. (1) [10]. Furthermore, the reaction rate of ⁹⁹Mo production in the irradiated samples was also calculated considering the irradiation duration ($T_{i}$) with Eq. (2) [10].

$$\begin{aligned} &A= \frac{\lambda (C-B)}{\varepsilon \mathrm{I}\gamma e^{-\lambda Td} (1- e^{-\lambda Tc} )}, \end{aligned}$$

(1)

$$\begin{aligned} &R_{Mo-99} = \frac{\lambda (C-B)}{\varepsilon \mathrm{I}\gamma ( 1- e^{-\lambda T_{i}} ) e^{-\lambda T_{d}} (1- e^{-\lambda T_{c}} )}. \end{aligned}$$

(2)

3 Results and discussion

3.1 Complementary neutron production on a low-energy cyclotron

At the fundamental level, the endoergic nuclear reaction of an impinging proton on ¹⁸O for the production of ¹⁸F can be represented by the equation:

$$\begin{aligned} ^{18} O + p \rightarrow ^{18} F + n,\qquad Q = -2.44\ \mathrm{MeV}. \end{aligned}$$

(3)

The reaction threshold energy is approximately 2.5 MeV with a reaction cross-section maximum of ∼550 mb occurring around 5 MeV. Although the ¹⁸O transmutation proceeds above and below the maximum cross-section energy, the construction of the target is devised so that the beam terminates within the water target, and it does so with a beam energy as close to the maximum cross-section. For a 16.5 MeV incoming proton beam, slowing of the beam, or beam degradation is performed within the Havar alloy window and controlled by its relative thickness. According to Eq. (3), the transmutation of ¹⁸O via proton capture is accompanied by the ejection of a neutron from the nucleus. Therefore, for every atom of ¹⁸F formed, a neutron is also generated. Thus, the degree of transmutation is dependent upon the number of impinging protons on the target, which can be calculated using Eq. (4).

Here, the elemental charge of a proton ($q_{\mathrm{proton}}$) is equal to $1.6 \times 10^{-19}$ Coulombs (C), and 1 microampere (μA) is equivalent to 10⁻⁶ C⋅s⁻¹. The proton production rate ($R_{\mathrm{proton}}$) for a beam current ($I_{\mathrm{beam}}$) can be calculated as:

$$\begin{aligned} R_{\mathrm{proton}} \biggl( \frac{\mathrm{proton}}{\mathrm{sec}} \biggr) = I_{\mathrm{beam}} (\mu A)* 10^{-6} \biggl( \frac{C}{\mathrm{sec}} \biggr) * \frac{1}{q} \biggl( \frac{\mathrm{proton}}{C} \biggr). \end{aligned}$$

(4)

Under standard operating beam currents, $R_{\mathrm{proton}}$ values up to ∼1 × 10¹⁴ protons⋅s⁻¹ are achievable. However, because of scattering events and interactions with the target body, window, and other components, the proton beam is not fully converted into ¹⁸F. One indicator of approximating beam conversion efficiency in the target is the saturation activity (SA), which is a function of ¹⁸F activity per unit of beam current, i.e., mCi⋅μA⁻¹. Using SA of the target and $I_{\mathrm{beam}}$ the number of ¹⁸F atoms produced per unit time can be determined, and thus the number of neutrons produced (R_neutron) can be inferred as:

$$\begin{aligned} R_{\mathrm{neutron}} \biggl( \frac{\mathrm{neutrons}}{\mathrm{sec}} \biggr) = I_{\mathrm{beam}} *SA \biggl( \frac{mCi}{\mu A} \biggr) * 3.7*10^{6} \biggl( \frac{s^{-1}}{mCi} \biggr). \end{aligned}$$

(5)

For determining the thermal neutron flux ($\Phi _{\mathrm{neutron}}$, neutrons/cm²⋅sec) from R_neutron as shown in Eq. (5), Patterson’s formula [11] shown in Eq. (6) can be applied, where K (=1.25) is a constant and I is the surface area exposed to the neutron field.

$$\begin{aligned} \Phi _{\mathrm{neutron}} \biggl( \frac{\mathrm{neutrons}}{\mathrm{sec}* \mathrm{cm}^{2}} \biggr) =K \frac{R}{ I}. \end{aligned}$$

(6)

Thus, the higher the saturation activity, the higher the efficiency of beam conversion is and production of neutrons from the target. Likewise, the total activity generated per batch of ¹⁸F is equivalent to the total number of neutrons manifested. However, Carroll observed computationally with ALICE9 that the correlation between neutron production and SA did not trend linearly when proton beam energies exceeded 12 MeV. This was attributed to other energetically accessible ¹⁸O(p, x) neutron-emitting channels that were not possible under 12 MeV [12]. Likewise, the neutrons resulting only from ¹⁸O transmutation do not completely account for the total production within the system. In fact, an array of (p, n) reactions occurs in the irradiation system, such as with nearby parts of the cyclotron or the ¹⁸O target including the Havar foil itself that interact with the stray proton beam. Accounting for the entirety of these possible interactions, the reported calculated flux produced from the production of ¹⁸F with a proton beam operating at 15 MeV and 75 μA approaches 1.3 × 10¹² n/s [13].

More accurate determinations of neutron fluxes produced in PET cyclotrons during ¹⁸F production have been conducted through experimental flux wire, neutron detection, and dosimetry measurements coupled with multi-system computational modelling, such as Monte Carlo, MCNP, FLUKA, etc. For example, Jeffries et. al. have reported the neutron flux generated from a GE PETtrace-800 employing BTI Targetry high yield ¹⁸F targets TS-1700 (80 μA) and TS-1650P (72 μA) [14]. Measurements were performed using a variety of activation flux wires with different neutron threshold energies along with comparative modelling via STAYS PNNL, MCNP6, and 3-Group programs. Results determined that the fast neutron flux density adjacent to the target was 1.8 × 10⁹ n/cm²⋅s to 3.0 × 10⁹ n/cm²⋅s with 1 MeV equivalent flux density from 2.4 × 10⁹ neutrons/cm²⋅s to 4.9 × 10⁹ n/cm²⋅s. Castillo analysed modelled data from a compilation of past studies concerning the neutron production on a 16.5 MeV proton beam on an ¹⁸O[H₂O] target normalised to a beam current of 75 μA [13]. A summary of the data shows that depending on the model employed and physical parameters integrated into the model, the $R_{\mathrm{neutron}}$ rate varied from 1.20 × 10¹² n/s to 1.73 × 10¹² n/s with neutron fluxes ranging from 2.18 × 10⁸ n/cm²⋅s to 1.45 × 10⁸ n/cm²⋅s. The determined neutron energy distribution spanned a broad energy range with maxima occurring in the epithermal to fast neutron regime, i.e., 0.01 to 5 MeV, as well as in the thermal region, ≤ 0.025 eV. Bosko [15] also reported on modelled neutron production rates and neutron energy distributions from a GE PETtrace-800. The $R_{\mathrm{neutron}}$ was determined for a 16.5 MeV proton beam on a thick ¹⁸O[H₂O] target to be 3.21 × 10¹¹ n/s assuming a 60 μA beam current. The neutron energy distribution ranged from 1 MeV with a relative flux of ∼1.1 × 10¹¹ n/s and extended to over 10 MeV with a gradual drop in flux to ∼1.1 × 10⁸ n/s; the majority of the flux focused in the neutron energy range from 1 MeV to 4 MeV. This $R_{\mathrm{neutron}}$ value is similar to the one reported by Horitsugi et al. from the GE Healthcare ALARA reports as 7.13 × 10¹¹ n/s at 80 μA during dual port irradiation [16, 17]. Figure 2 shows the correlation between proton beam current and the resulting neutron flux (n/s) for 16.5 and 18.5 MeV cyclotrons from reported values in the literature.

Figure 2

Reported neutron fluxes arising from the ${}^{18} O(p,n)^{18}$F transmutation reaction as a function of proton beam current for (A) 16.5 MeV and (B) 18.5 MeV energy cyclotrons [7, 12, 13, 15–18]

Another factor that should be considered is the neutron direction or angular distribution from the target. This is particularly important when considering the placement of a source to be irradiated, where neutron energy distribution and flux can be determined by sample position relative to the neutron source, i.e., ¹⁸O[H₂O] target. For example, in a nuclear reactor scenario neutrons are emitted isotropically, whereas for accelerators this is not necessarily the case and neutron angular distribution is anisotropic. For accelerators, some of the energy from the primary incoming beam will impart some of its energy on the secondary particle emitted, which can lead to scattering as well as backscattering of the secondary particles. Generally, the higher the energy of the incoming proton, the higher the energy of the neutrons produced.

For the transmutation of ¹⁸O to ¹⁸F with a high-energy proton beam, the majority of the generated neutrons are emitted in the same direction as the incoming proton beam. As mentioned previously, scattering is observed further out from the source origin in the ¹⁸O target, where the neutron field becomes more diffuse, although still mostly directionally forward. It is also noted that a lesser, yet significant amount of neutrons are emitted in the opposite direction, or essentially backwards from the target into the incoming proton beam. Specifically, the GE Site Planning Guide states that when compared to the total flux of neutrons in the forward direction, those perpendicular to the incoming particle beam will be 30% less of this value, and those in the backwards direction will be at least 10× lower in magnitude [18].

Irradiations were performed on several Mo metal samples (Table 1) with various physical characteristics, i.e., mass, thickness, surface area, and volume, using the complementary neutrons produced during the production of ¹⁸F with a low-energy cyclotron. As previously discussed, the proton-induced transmutation of ¹⁸O to ¹⁸F results in the liberation of a free neutron, which is directly correlated to the yield of ¹⁸F produced in the ¹⁸O target. Shown in Fig. 3 is an ¹⁸F production curve for a 20-minute irradiation at 75 μA yielding a total of ∼1.8 Ci of ¹⁸F at EOB. As every atom of ¹⁸F generated is equal to at least one neutron, then this would correspond to an average neutron flux of 5.2 × 10¹¹ n/s with a total neutron output of $2.3 \times 10^{13}$ in the measured time period. With routine production schedules consisting of 4 to 5 h of operational beam time, the total average neutron output on a system equivalent to the one tested would range between 7.5 × 10¹⁵ to 9.4 × 10¹⁵ neutrons in a working day based solely on ¹⁸F transmutation, not considering other avenues of neutron production, for example, (p, n) reactions with the Havar window. In comparison to literature-reported values for neutron fluxes in a 16.5 MeV cyclotron, the value inferred from the ¹⁸F activity, i.e., 5.2 × 10¹¹ n/s, is comparable to the one of Castillo [13], i.e., 6.0 × 10¹¹ n/s, at 75 μA, suggesting that the system is likely capable of outputting fluxes upwards of 1 × 10¹² n/s with all other neutron producing reactions considered.

Figure 3

Activity (mCi) of ¹⁸F generated as a function of time (s) in an enriched ¹⁸O target using the cyclotron and experimental setup described previously with an operating current of 75 μA

3.2 Hybridised production of ^99mTc and ¹⁰¹Tc with complementary neutrons

Although Mo is characterised by seven naturally occurring isotopes, i.e., ⁹²Mo (14.53%), ⁹⁴Mo (9.16%), ⁹⁵Mo (15.84%), ⁹⁶Mo (16.67%), ⁹⁷Mo (9.60%), ⁹⁸Mo (24.39%), ¹⁰⁰Mo (9.82%), only ⁹⁸Mo and ¹⁰⁰Mo provide direct routes to the appreciable formation of ⁹⁹Mo when interacting with neutrons. For ⁹⁸Mo, the fundamental interaction is via ⁹⁸Mo(n, γ)⁹⁹Mo. The neutron capture on ⁹⁸Mo occurs across a wide range of neutron energies. The thermal neutron capture, i.e., 0.025 eV, on ⁹⁸Mo has a cross-section (σ_thermal) of approximately 0.137 b. In comparison to the fission-based production of ⁹⁹Mo with thermalised neutrons, this value is nearly ${\sim} 270{\times}$ less for the equivalent irradiation using ²³⁵U as the fuel source. Furthermore, neutron capture reactions with thermal neutrons generally yield specific activities of hundreds of mCi/g at most depending upon the neutron flux, the amount of self-shielding effects, and the level of enrichment of ⁹⁸Mo in the target. It is noted that enrichment of ⁹⁸Mo can increase production yields up to 4× more than natural isotopic samples.

For higher energy neutrons, particularly in the resonance, i.e., 10-300 eV, and intermediate regions, i.e., 300 eV-0.05 MeV, ⁹⁸Mo exhibits multitudes of enhanced resonance capture maxima. The resulting averaged cross-section of this region (σ_resonance) approaches ∼7 b, which is more than 50× than that of σ_thermal for ⁹⁸Mo, i.e., 0.130 b. Likewise, it has been established that the self-shielding of epithermal neutrons from other Mo isotopes is negligible comparative to interactions with ⁹⁸Mo [19]. Because of the greater probability of interaction, it has been reported that specific activities up to ∼3.4 Ci/g and ∼15 Ci/g are achievable for natural and enriched targets, respectively, thus yielding significantly greater outputs in ⁹⁹Mo [20]. The neutron capture behaviour of ⁹⁸Mo is quite similar to ¹⁰⁰Mo, where σ_thermal for ¹⁰⁰Mo is 0.199 b and σ_resonance is 3.76 b.

The second accessible pathway for ⁹⁹Mo production with neutrons via ¹⁰⁰Mo(n,2n)⁹⁹Mo only occurs for fast neutrons, (≥ 1 MeV) [21, 22]. The threshold reaction energy is approximately 8 MeV with a cross-section (σ_fast) maximum of roughly 1.5 b befalling between 13 and 16 MeV. Relative to σ_thermal of ⁹⁸Mo, this value is nearly 11× larger, although, for the averaged σ_resonance of ⁹⁸Mo, it is about 5× less. However, because of the lower isotopic concentrations of ¹⁰⁰Mo in comparison to ⁹⁸Mo in natural samples, enriched materials can provide up to 10× the production. Due to the higher reaction threshold energy, this pathway is generally associated with production means where adequate fast neutron fluxes are present [23].

In order to determine the production of ⁹⁹Mo/^99mTc via neutron activation in a Mo target, several Mo metal samples with varying geometries, i.e., cube versus sheet, were subjected to the neutron field within the cyclotron. Samples were placed in the vicinity of the highest neutron fluxes around the ¹⁸F target. Typically samples were arranged prior to ¹⁸F production for the given day and either measured at the end of the production day or after several days.

Shown in Fig. 4 is the γ-ray spectrum of a Mo cube measured 1.4 h post-EOB following irradiation after 4.5 h in an un-moderated zone directly behind the ¹⁸F target, aligned with the incoming proton beam. The γ-ray spectrum shows the characteristic peaks for ⁹⁹Mo at 185 keV and 725 keV, ^99mTc at 146 keV, and ¹⁰¹Tc at 311 keV and 531 keV. These are in good agreement with the reported gamma energies for ⁹⁹Mo at 181.1 keV and 739.5 keV, ^99mTc at 140.5 keV, and ¹⁰¹Tc at 306.8 keV and 531 keV. The presence of ⁹⁹Mo is most likely attributed to the neutron activation of ⁹⁸Mo in the target, whereas ^99mTc present is a result of ⁹⁹Mo decay.

Figure 4

Gamma spectrum of Mo cube after ∼1.4 h post EOB showing characteristic peaks for ⁹⁹Mo, ${}^{99 m}$Tc, and ¹⁰¹Tc after being irradiated in a neutron field for ∼4.6 h

The radioisotope ¹⁰¹Tc is also a daughter product, however, from ¹⁰¹Mo, formed via (n, γ) on ¹⁰⁰Mo; the absence of any characteristic γ-ray peaks (E$\gamma = 229.1$ keV (2.20%), 257.1 keV (2.77%), 261.1 (20.6%), 359.1. keV (3.31%), 570.1 keV (6.62%)) from ¹⁰¹Mo is due to its shorter half-life (t$_{1/2} = 14.16$ min.). Unlike ¹⁰¹Mo, which is formed directly during the neutron irradiation, ¹⁰¹Tc being the daughter product still persists and can be seen as one of the more prominent peaks in the spectrum. Likewise, because the irradiation period (∼4.6 h) is sufficiently long compared to the half-life of ¹⁰¹Mo to achieve saturation, the concentration of ingrown ¹⁰¹Tc should be equivalent to that of ¹⁰¹Mo at EOB in the Mo target. Therefore, the total amount of ¹⁰¹Tc generated after EOB should be equal to the amount present at EOB plus the amount generated from residual ¹⁰¹Mo decay. The decay-corrected activity for the 311 keV peak of ¹⁰¹Tc in the sample was calculated to be ∼74 μCi at EOB.

Presented in Fig. 5 is the γ-ray measurement of the same Mo cube after 26 h post-EOB. The identified species in the spectrum were ⁹⁹Mo with peaks at 180 keV and 740 keV, and ^99mTc with a peak at 150 keV. At this point, there was no remaining ¹⁰¹Tc. The peak-to-peak ratio of ^99mTc to ⁹⁹Mo was determined to be 22.7, which is 6.6× greater than after 1.4 hr post EOB; this time post-EOB correlates to the near-maximum ingrowth of ^99mTc from ⁹⁹Mo decay.

Figure 5

(Left) Gamma spectrum of Mo cube after 26 h post EOB showing characteristic peaks for ⁹⁹Mo and ${}^{99 m}$Tc. (Right) Gamma spectrum (12 h count) of Mo foil acquired 24 h after EOB

In Fig. 5, the γ-ray spectrum acquired 24 h after EOB of an irradiated Mo foil is presented. The foil was positioned underneath the ¹⁸F target and behind several layers of Styrofoam and irradiated within the neutron field for a total duration of 4.6 h. The γ-ray spectrum shows identifiable peaks indicative of ⁹⁹Mo at 180 keV and 740 keV, and ^99mTc at 150 keV. As expected, no ¹⁰¹Mo and ¹⁰¹Tc were detectable at the time of/during measurement. The γ-ray spectrum here is also comparable with the one shown in Fig. 5 for the Mo cube.

The effect of various physical attributes, (i.e., volume, surface area) of the Mo targets and moderation / no moderation of the neutron field utilised in the production of ⁹⁹Mo / ^99mTc were of particular interest. Three different samples, i.e., two Mo cubes (Cube 1 and Cube 2) and one Mo foil (Foil 1), were compared (Table 2), where all samples were irradiated for similar times (4.3 to 4.6 h) over the course of two production runs, one in which the samples were attached to the ¹⁸F used for production (∼2.7 h) and a second run during which they were located adjacently to the ¹⁸F target being operated (1.6 to 1.9 h); between the two production runs was a pause of approximately 1.5 h, however, for simplicity, ⁹⁹Mo production only accounted for total irradiation time during operation. The effect of moderation was determined, where Cube 1 was placed in an un-moderated zone behind the ¹⁸F target, and Cube 2 was positioned underneath the target behind both a Styrofoam moderator and Foil 1. Final production activities and rates determined using Eq. (1) and Eq. (2), respectively, for ⁹⁹Mo were normalised for decay after EOB and mass of the targets.

Table 2 Comparison of activities and reaction rates of ⁹⁹Mo in various Mo metal samples after one production day of ¹⁸F in a low-energy cyclotron with neutrons generated through the ¹⁸O(p, n)¹⁸F reaction

As shown in Table 2, the production and reaction rates of the two Mo cubes are presented. Both cubes exhibited identical volumes and surface areas, however, Cube 1 was subjected to no moderation of the incoming neutron field, whereas the neutron field was relatively moderated prior to interacting with Cube 2. The specific activities of ⁹⁹Mo normalised to EOB and per gram of sample show that both samples yielded approximately 240 nCi/g to 340 nCi/g of ⁹⁹Mo produced at EOB after one routine production day of ¹⁸F, or ∼12 nCi/g⋅Ci-¹⁸F to 18 nCi/g⋅Ci-¹⁸F. The reaction rates of formation of ⁹⁹Mo in the samples were determined to be 5.0 μCi/g to 7.6 μCi/g, accounting for the irradiation time of each sample. From these values, it is seen that Cube 1 yielded slightly greater values of specific activity and reaction rates in comparison to Cube 2, although not significantly. When taking into consideration that the neutron flux generated perpendicular to the incoming proton beam is 70% of the beam generated in the same direction, the values of Cube 2 when adjusted for this are 343 nCi/g and 7.1 μCi/g for the specific activity and reaction rate, respectively, and are essentially identical for those of Cube 1. An explanation for this behaviour is as such: although Cube 2 experienced a lower neutron flux than that of Cube 1, the neutron energies due to moderation were more favourable for neutron capture reactions for Cube 2, thus allowing for equivalent production and reaction rates of ⁹⁹Mo. As discussed previously, neutron capture on ⁹⁸Mo is greatly favoured within neutron energies of 300 keV to 0.05 MeV, and it quickly diminishes with neutron energies outside of this region. Therefore, if the neutrons coming out of the target have energies greater than 0.05 MeV, which is likely the case, then there will be far fewer interaction probabilities for capture on ⁹⁸Mo.

After establishing the approximate yield of ⁹⁹Mo that could be generated in a single ¹⁸F production day, it was of interest to investigate and track production over the course of a several-day period. However, it is noted that because ¹⁸F production days were not all homogenous, where some days accounted for a single run and others multiple runs, attempting to extract any definitive behaviour of ⁹⁹Mo / ^99mTc production was beyond the limit of the study and only broad trends in the data were to be considered. Figure 6 shows the relative activity of ⁹⁹Mo produced over the course of several ¹⁸F production days. The starting day corresponds to a Friday, after which no irradiations were performed for the following two days over the weekend. Generally, the trend of the curve shows the relative build-in of ⁹⁹Mo as a function of successive irradiations. Furthermore, a similar trend was also observed when tracking the ingrowth of ^99mTc in the samples, as shown in Fig. 6.

Figure 6

(Left) Production of ⁹⁹Mo in a Mo foil tracked over several days of irradiation. Days 2 and 3 correspond to a weekend and no irradiations were performed. (Right) Presence of ${}^{99 m}$Tc (raw counts per second) in Mo foils 1–4 over the duration of the irradiation experiment

4 Conclusion

The purpose of the study was to determine whether it was feasible to co-produce ^99mTc and ¹⁰¹Tc in parallel with ¹⁸F using a hybridised system based upon liberated neutrons from the ¹⁸O(p, n)¹⁸F and other potential associated (p, n) reactions generated with a low-energy biomedical cyclotron. From the data, it was demonstrated that the co-production of both isotopes under the tested conditions was feasible. It is noted that Mo sample design and placement were extremely rudimentary—if much higher quantities of ⁹⁹Mo / ^99mTc or ¹⁰¹Mo / ¹⁰¹Tc were to be produced, possibilities for increasing yields, considering target mass/volume, irradiation time, and neutron flux could be applied for scaling. Linking a system like this with the appropriate separation platform for isolating Tc isotopes from irradiated low-specific activity (LSA) Mo targets, may become a distributed source of these and other medically relevant radioisotopes [3, 24]. Additionally, the possibility of leveraging at least some of the currently existing cyclotron infrastructures already in place could be seen as a huge advantage compared to building new reactors or facilities for radioisotope production.