New Cross-Sections for ^natMo(α,x) Reactions and Medical ⁹⁷Ru Production Estimations with Radionuclide Yield Calculator

Abstract

The production of ⁹⁷Ru, a potential Single Photon Emission Computed Tomography (SPECT) radioisotope, was studied at ARRONAX. The cross-section of ^natMo(α,x)⁹⁷Ru reaction was investigated in the range of 40–67 MeV irradiating the ^natMo and Al stacked-foils. The activities of ⁹⁷Ru and other radioactive contaminants were measured via gamma spectroscopy technique. A global good agreement is observed between obtained cross-section results, previously reported values and TENDL-2017 predictions. Additionally, Radionuclide Yield Calculator, a software that we made available for free, dedicated to quickly calculate yields and plan the irradiation for any radioisotope production, was introduced. The yield of investigated nuclear reactions indicated the feasibility of ⁹⁷Ru production for medical applications with the use of α beam and Mo targets opening the way to a theranostic approach with ⁹⁷Ru and ¹⁰³Ru.

1. Introduction

The ⁹⁷Ru radioisotope was first acknowledged as medically interesting in 1970 [1] and is even studied in recent measurements [2,3]. It has a half-life of 2.9 d allowing non-local production and emits low-energy high-intensity gamma lines (see

Table 1. Nuclear data [22] of ⁹⁷Ru and observed radionuclidic contaminants as well as reactions contributing to their formation during the irradiation of ^natMo target*.

Radionuclide	T1/2	Decay Mode (%)	γ-Lines [keV] and Intensities** (%)	Contributing Reactions***	Q-Value [MeV]
97Ru	2.83 d	EC (100)	215.7 (85.8) 324.5 (10.8)	94Mo(α,n)97Ru	−7.9
				95Mo(α,2n)97Ru	−15.3
				96Mo(α,3n)97Ru	−24.5
				97Mo(α,4n)97Ru	−31.3
				98Mo(α,5n)97Ru	−41.6
				100Mo(α,7n)97Ru	−54.1
89gZr	78.4 h	β+ (23), EC (77)	908.96 (100)	92Mo(α,x)89totZr	−16.7
				94Mo(α,x)89totZr	−14.0
				95Mo(α,x)89totZr	−21.4
				96Mo(α,x)89totZr	−30.6
				97Mo(α,x)89totZr	−37.4
				98Mo(α,x)89totZr	−46.0
				100Mo(α,x)89totZr	−60.2
				92Mo(α,x)89totNb→89totZr	−21.1
				94Mo(α,x)89totNb→89totZr	−38.9
				95Mo(α,x)89totNb→89totZr	−46.2
				96Mo(α,x)89totNb→89totZr	−55.4
				97Mo(α,x)89totNb→89totZr	−62.2
				98Mo(α,x)89totNb→89totZr	−70.9
				100Mo(α,x)89totNb→89totZr	−85.1
96gTc	4.28 d	EC (100)	778.22 (100) 812.58 (82) 849.93 (98) 1126.97 (15.2)	94Mo(α,x)96totTc	−13.3
				95Mo(α,x)96totTc	−14.4
				96Mo(α,x)96totTc	−23.7
				97Mo(α,x)96totTc	−30.4
				98Mo(α,x)96totTc	−39.0
				100Mo(α,x)96totTc	−53.3
99Mo	65.9 h	β− (100)	140.51 (89.43) 739.50 (12.13)	97Mo(α,2p)99Mo	−13.7
				98Mo(α,x)99Mo	−14.7
				100Mo(α,x)99Mo	−8.3
95gTc	20.0 h	EC (100)	765.8 (93.82)	92Mo(α,n)95gTc	−5.7
				94Mo(α,x)95gTc	−14.9
				95Mo(α,x)95gTc	−22.3
				96Mo(α,x)95gTc	−31.4
				97Mo(α,x)95gTc	−38.3
				98Mo(α,x)95gTc	−46.9
				100Mo(α,x)95gTc	−61.1
				92Mo(α,n)95Ru→95gTc	−9.0
				94Mo(α,3n)95Ru→95gTc	−26.7
				95Mo(α,4n)95Ru→95gTc	−34.1
				96Mo(α,5n)95Ru→95gTc	−43.3
				97Mo(α,6n)95Ru→95gTc	−50.1
				98Mo(α,7n)95Ru→95gTc	−58.7
				100Mo(α,9n)95Ru→95gTc	−73.0

* ^natMo composition: ⁹²Mo (14.6%), ⁹⁴Mo (9.2%), ⁹⁵Mo (15.9%), ⁹⁶Mo (16.7%), ⁹⁷Mo (9.6%), ⁹⁸Mo (24.3%), ¹⁰⁰Mo (9.7%); ** lines with less than 10% intensities are not included; *** “tot”—the reaction produces the radionuclide directly and via decay of its metastable state.

) which have favorable characteristics for prolonged Single Photon Emission Computed Tomography (SPECT) examinations. It decays only by electron capture (EC) which lowers the contribution to the dose as compared to β⁺ decays. It has a theranostic matched pair in the form of ¹⁰³Ru (T_1/2 = 39.26 d) that decays to the short-lived Auger emitter ^103mRh (T_1/2 = 56.12 min), a promising gamma-free therapeutic agent. Moreover, ruthenium element has a rich chemistry associated with its various oxidation states (II, III, IV and VIII) and forms more stable compounds compared to the SPECT-standard ^99mTc [4]. Many radioactive Ru-labeled compounds have been studied and found applications as summarized recently by [5], in particular as the chemotherapy agents [6,7].

Due to these interesting characteristics, many studies on production of ⁹⁷Ru have been conducted. The reactor route via ⁹⁶Ru(n,γ)⁹⁷Ru was reported by [1] but it yields very low specific activity which may limit its use for some applications such as molecular imaging. To obtain high specific activity product, one can use charged projectile from accelerators. In case of cyclotron routes, the first and most used reaction is ¹⁰³Rh(p,spall)⁹⁷Ru with 200 MeV proton beam and natural rhodium target, as suggested by [8]. While producing high amount of activity of no-carrier-added (NCA) ⁹⁷Ru, this method requires high energy protons but no details about the impurity levels were reported. Another reaction route is the ¹⁰³Rh(p,x)⁹⁷Ru reaction using 60 MeV proton beam [9]; the ⁹⁷Ru production yield is very high but accompanied by Tc radioactive impurities which are difficult to discard even after the chemical separation step. A very feasible option is the ⁹⁹Tc(p,3n)⁹⁷Ru reaction suggested by [10] and studied later up to 100 MeV by [4,11,12] as it produces significant amounts of ⁹⁷Ru with very small amount of radioactive impurities. However, the availability of ⁹⁹Tc radioactive target is an issue. Later, experimental excitation functions were reported for ^natAg(p,x)⁹⁷Ru up to 80 MeV by [13] and for ^natPd(p,x)⁹⁷Ru up to 70 MeV by [14]. These two production routes have much smaller cross-section, hence ⁹⁷Ru production would require long irradiation time and would contain a substantial amount of radioactive impurities. In case of deuteron beam, the available reaction ⁹⁶Ru(d,x)⁹⁷Ru studied by [15] is favorable but would produce low specific activity as the target material is an isotope of the nuclide of interest. Some groups have also investigated more exotic projectiles such as helium-3 through ^natMo(³He,x)⁹⁷Ru [16], ⁹³Nb(⁷Li,3n)⁹⁷Ru [17] and ⁸⁹Y(¹²C,p3n)⁹⁷Ru [2,18]. In these cases, after chemical separation, low level of radioactive impurities can be achieved but the availability of these beams is scarce making these processes not suitable to launched clinical trials. Finally, the cross-sections for α-induced reactions on Mo were investigated by [19]. ^natMo(α,x)⁹⁷Ru production and impurities up to 40 MeV were thoroughly studied in [3,20].

In this work, we investigate the optimization of ^natMo(α,x)⁹⁷Ru production route and extend the available cross-section data to higher energy in coherence with commercially available cyclotrons, [21] which are able to deliver up to about 70 MeV alpha beam. We also report on the coproduction of the measured radioactive impurities (listed in Table 1) via ^natMo(α,x) and explore the possible commercial production of ⁹⁷Ru with the α beam on Mo target using the software Radionuclide Yield Calculator (RYC) that we developed and made freely available to the community.

2. Materials and Methods

2.1. Stacked-Foils Irradiations

Three experiments were performed at the ARRONAX facility [23], irradiating stacked-foils targets in vacuum with α beam of 67.4(5) MeV for about 1 h with beam currents of 40–60 nA. The stacked-foils technique and set-up in our facility have been described most recently in [24,25,26]. A typical stacked-foil target consisted of an Al monitor foil (~10 μm thick) in front, followed by the set of multiple ^natMo foils (~10 μm thick) and Al degraders (50–500 μm thick), arranged alternately. The order of the foils in the stacks were planned so that each ^natMo foil is activated with a different energy, all covering the energy range from 40 MeV to 67 MeV in about 3 MeV intervals (the projectile stopping-power in the stacks was calculated using SRIM software [27]). Certain foils were also used as catchers of the recoil atoms.

All foils were purchased from the GoodFellow© company with a purity of 99% for Al and 99.9% for ^natMo. Each foil was weighed before irradiation using an accurate scale (10⁻⁵ g) and scanned for area determination, allowing the precise thickness calculation (assuming the homogeneity over the whole surface).

As recommended by the International Atomic Energy Agency [28], the activity of the ²⁴Na radioisotope formed in Al monitor foil was used to calculate the beam current impinging the stack. Additionally, during the irradiations, the online beam current monitoring was performed using a Faraday’s Cup with an electron suppressor for precise measurement and located behind the stack. The two measurements were consistent with each other.

2.2. Gamma Spectroscopy and Data Analysis

After about 14 h of cooling time, the gamma ray spectra of irradiated samples were collected using a HPGe Canberra detector with efficiency 20% at 1.33 MeV equipped with low-background lead and copper shielding. Each foil was placed at a height of 19 cm from the detector to ensure the dead-time below 10%. The detector was calibrated in energy and efficiency at 19 cm with ⁵⁷Co, ⁶⁰Co and ¹⁵²Eu calibrated sources from LEA-CERCA (France) prior to the measurements. Gamma spectra were recorded using the LVis software from Ortec© while the activity of the radionuclides produced at the End of Bombardment (EOB) were derived using the FitzPeaks Gamma Analysis and Calibration Software (JF Computing Services). For the identification and activity estimation we used the γ-line and associated branching presented in Table 1. Knowing the activity of each isotope and the thickness of the foil in which they were observed, it was possible to calculate their production cross-section σ with the following formula:

$$\sigma =\frac{A_{EOB}MZe}{H{N}_{A}I\rho x\left(1-exp\left\{-\lambda t\right\}\right)}$$

where: A_EOB—activity of the radioisotope at the EOB, M—atomic mass of the target, Z—ionization number of the projectile, e—elementary charge, H—enrichment and purity of the foil, N_A—Avogadro’s number, I—beam current, ρ—target material density, x—thickness of the foil, λ—decay constant of the radioisotope, t—time of the irradiation. The similar formula, solved for I and with cross-section values from [28], was used to calculate the beam current from the monitor foils. The projectile energy in the middle of the foils was adopted to the corresponding cross-section value.

The errors of cross-section values were propagated from the uncertainty of thickness measurements (around 1%), uncertainty of the counts in the γ-line peaks in the spectroscopy measurements (around 5–10%) and the error of the calculated beam current (around 5–10%) while the corresponding energy errors were propagated with SRIM software [27] considering the beam energy straggling through the foils (the initial energy error estimated by the cyclotron operators was 0.5 MeV).

The obtained cross-section values are then compared with TENDL-2017 (TALYS-based evaluated nuclear data library) [29] and the experimental results from other research groups.

2.3. Radionuclide Yield Calculator

Given the cross-section values, one can calculate the Thick Target Yield (TTY) with the following formula [30,31]:

$$TTY\left(E\right)=\frac{H{N}_A\lambda }{ZeM}\underset{E_{min}}{\overset{E_{max}}{\int }}\frac{\sigma \left(E\right)}{dE/dx\left(E\right)}dE$$

where: E_max and E_min—maximal and minimal energy of the projectile penetrating the target (in case of TTY, E_min ≤ reaction threshold), dE/dx—stopping-power of the projectile in the irradiated target. To facilitate this calculation for ⁹⁷Ru, as well as any other radioisotope and cross-section, we developed a Radionuclide Yield Calculator, later named RYC.

RYC is graphical user interface software written in python programming language (version 2.7) [32] using the TKinter module and compiled with PyInstaller software (version 3.4) [33]. It uses the cross-section and basic target data inputs to instantly calculate TTY and activity produced in any irradiation scenario. Data points can be fitted using different type of function, gaussian-like and polynomial functions, using the least-squares method. Excitation functions from TENDL [29] can be easily imported to compare with experimental data and to look for potential radioactive impurities. RYC with its detailed documentation can be downloaded from the ARRONAX website [34]. In particular, RYC uses implemented SRIM module [27] for stopping-power calculation.

The validation of this software was performed using data from the literature. On Figure 1, we compare the RYC-calculated TTY with the values published by IAEA [28,35] based on the same cross-section values for ¹²⁷I(p,3n)¹²⁵Xe, ⁶⁴Ni(d,2n)⁶⁴Cu and ²⁰⁹Bi(α,2n)²¹¹At reactions on metallic targets. Data calculated by RYC are presented as points whereas the curve published by IAEA correspond to the lines. As can be seen, for the 3 types of projectiles and for the different target masses, a very good agreement is obtained. The same good results have been obtained for all our tests.

3. Results and Discussion

3.1. Cross-Section Measurements

On Figure 2, Figure 3, Figure 4, Figure 5 and Figure 6 we present the measured cross-sections for ^natMo(α,x) reactions producing the ⁹⁷Ru radioisotope as well as observed radioactive impurities: ^89gZr, ^95gTc, ^96gTc, and ⁹⁹Mo. The contributing reactions forming these radioisotopes are shown in Table 1. The experimental data are compared with previous experiments reported in literature [3,19,20] and the values from TENDL-2017 library [29]. Measured cross-section values are also listed in

Table 2. Measured cross-sections for ^natMo(α,x) reactions (with the uncertainties in the parenthesis).

E [MeV]	natMo(α,x) Cross-Section [mb]
	97Ru	89gZr	95gTc	96totTc	99Mo
41.80(75)	237(20)	ND*	81(11)	73(7)	7.5(1.0)
46.03(68)	225(20)	ND	127(14)	89(8)	10.1(1.2)
50.00(64)	199(18)	ND	163(17)	100(9)	11.4(1.3)
51.93(62)	166(14)	ND	170(16)	101(9)	12.8(1.3)
55.30(60)	159(13)	3.6(9)	177(17)	109(9)	13.5(1.4)
58.51(56)	176(15)	11.7(1.6)	205(17)	119(10)	ND
59.97(55)	176(15)	18(2)	174(24)	116(10)	14.0(1.5)
63.47(53)	180(16)	30(3)	188(16)	118(10)	15.0(1.7)
66.84(50)	173(14)	40(3)	203(17)	122(10)	15.6(1.3)

* ND = not detected.

(with corresponding energy errors, not visible on the graphs).

In the case of ⁹⁷Ru production (Figure 2), our measurements correspond well to the data at lower energies. Compared to the experimental data, TENDL shows similar structure but underestimates the cross-section by about 30 mb in the region 20–40 MeV. The subsequent fall of the excitation function and a bump seem to be shifted by 5–10 MeV with respect to the experimental data.

The ^89gZr excitation function (Figure 3) is measured for the first time. The predictions of TENDL shows a similar trend as our measurements but with a slightly shifted toward lower energies (5 MeV).

For ^95gTc (Figure 4), our measurements are consistent with the previously measured data at lower energies. The shape of TENDL calculations seems to be the same as obtained in the measurements, but again a shift in energy is observed. This shift is probably related to the code since 3 different sets of data acquired at different time, different laboratories and overlapping energy range are consistent with each other and shows the same shift with respect to TENDL.

The experimental data describes well the excitation function for ^96totTc (Figure 5). Additionally, our measurements preserve the trend of the ones reported earlier for lower energies. Here we do not observe any shift with respect to the TENDL calculations, as seen in the previous reactions.

The experimental cross-sections for ⁹⁹Mo production (Figure 6) are consistent. Our results show a continuous rise of the excitation function up to the maximum energy of our measurements. This is in obvious contrast to TENDL which predicts a maximum at around 30 MeV and then a decrease of the excitation function. Slight shift between TENDL and experimental results is observed at low energies.

3.2. Calculated Yield and Production

Using RYC, we calculated TTY for ^natMo(α,x)⁹⁷Ru reaction on metallic ^natMo target, based on our cross-section measurements above 40 MeV and the values reported by [3,20] below 40 MeV (Figure 7). The TTY values for other radioisotopes were also calculated in a similar way (not shown) to estimate the radioactive impurities.

The obtained experimental TTY values for ⁹⁷Ru and radioactive impurities were used to estimate the possible production of ⁹⁷Ru (

Table 3. Estimation of ⁹⁷Ru activity produced via the irradiation of ^natMo (based on experimental data) and enriched ^95,96Mo targets (based on TENDL-2017 [29]) with α beam in two energy ranges. The list of radioactive impurities is narrowed down to the long-lived ones and shows their activity relative to activity of ⁹⁷Ru at EOB.

α energy		30–15 MeV		67–15 MeV
target		natMo	95Mo (100%)	natMo	96Mo (100%)
thickness		100 mg/cm2	100 mg/cm2	540 mg/cm2	540 mg/cm2
97Ru yield		3.5 MBq/µAh	14 MBq/µAh	20 MBq/µAh	31 MBq/µAh
irradiation		1 h, 15 µA	1 h, 15 µA	1 h, 2.5 µA	1 h, 2.5 µA
97Ru AEOB		50 MBq (1.4 mCi)	200 MBq (5.4 mCi)	50 MBq (1.4 mCi)	80 MBq (2.2 mCi)
SA at EOB		350 GBq/µmol (9 kCi/mmol)	1300 GBq/µmol (36 kCi/mmol)	420 GBq/µmol (11 kCi/mmol)	630 GBq/µmol (17 kCi/mmol)
relative activity [%]	97Ru	100	100	100	100
	89gZr	0	0	3	0.04
	95gTc	95	1E−3	200	150
	96gTc	4	0.2	25	34
	103Ru	0.12	0	0.02	0
reference		[3], [20]	TENDL-2017	[3], [20] this work	TENDL-2017

) with ^natMo target and for two energies: 30 MeV and 67 MeV, which are the most common in commercially available cyclotrons. The ⁹⁷Ru production yields are 3.5 MBq/µAh and 20 MBq/µAh respectively. Although the yield is almost 6 times larger at 67 MeV than at 30 MeV, the latter energy of α beam offer for example the optimal production of ²¹¹At (summarized recently by [36]) and ⁴³Sc [37] medical radioisotopes. Additionally, in Table 3 we show the possible production of 50 MBq as this amount was proven SPECT-applicable in several clinical trials [38]. We have also calculated the yield and the number of produced stable Ru atoms (⁹⁶Ru, ⁹⁸Ru, ⁹⁹Ru, ¹⁰⁰Ru, ¹⁰¹Ru, ¹⁰²Ru) based on the TENDL cross-sections to estimate the specific activity (SA) of ⁹⁷Ru. The SA presented here assumes 100% successful chemical extraction of Ru isotopes from Mo target at EOB and hence is just an estimation used to compare different production routes.

It is worth mentioning that from the diagnostics point of view, the most dangerous impurity is ¹⁰³Ru. It is the only other radioactive Ru element with long half-life (T_1/2 = 39.26 d), which will contribute to the patient’s dose via high-intensity gamma-line (497 keV with 90.9% intensity) and Auger electrons from its daughter (^103mRh). During the irradiation of ^natMo with α beam it can be only formed via ¹⁰⁰Mo(α,x) reactions marked as the red line on Figure 2. Its contribution is rather small in our energy range and its activity was below our detection limit but we address it nevertheless (based on the measurements of [3,20]).

For the completeness of this study, we show the alternative production of ⁹⁷Ru with the use of 100% enriched ⁹⁵Mo and ⁹⁶Mo targets and α beams of 30–15 MeV and 67–15 MeV, respectively.

Further chemical separation would be required to extract Ru element from Mo target and separate it from formed radioactive and stable elements of Tc, Nb, and Zr. This can be done for example with either the solvent extraction or distillation methods with an efficacy better than 80% [16]. The SA should also be considered in further chemical research as each production route form additional stable atoms of Ru, which would chelate the labeling compound.

4. Conclusions and Summary

We have extended the available cross-section measurements of selected ^natMo(α,x) reactions up to 67 MeV. Our measurements preserve well the trend of the cross-section values reported previously below 40 MeV and are consistent in overlapping energy ranges. A reasonable agreement with TENDL is observed however in certain cases the shift of 5–10 MeV is visible with respect to the experimental data.

We have shown the feasibility of no-carrier-added ⁹⁷Ru production with α beam up to 67 MeV and thick ^natMo targets. The impurity of the only long-lived radioactive Ru radioisotope (¹⁰³Ru) is small, around 0.1%. An irradiation of 1 h with few µA α-beam should satisfy the need for SPECT imaging for the patient. Several doses could be produced with longer irradiations at higher currents or using enriched ^95,96Mo targets which will substantially increase the produced activity and SA.

The use of RYC [34] to calculate TTY based on cross-section data was also demonstrated.