Nuclear data for light charged particle induced production of emerging medical radionuclides

: Whatever the radionuclide to be used in nuclear medicine, it is essential to know the expected yield during the production process, but also of all the possible radio-nuclidic impurities coproduced, that can have an impact on the product final quality, as well as in the related waste management. The availability of the majority of emerging radioisotopes, including the theranostic ones or pairs, is mainly limited by the fact that, for most of them, the optimal production route still needs to be strengthened if not defined in some cases. The aim of this work is to present a review on the charged particle induced nuclear cross sections to produce some emerging radionuclides for medical applications to show that all types of projectiles should be considered in the quest of producing medical radionuclides. An accurate analysis of the production routes is presented for some radionuclides ( 67 Cu, 47 Sc, 89 Zr, 103 Pd, 186g Re, 97 Ru, 211 At) chosen as examples to highlight (i) how the quality of the ﬁ nal product strongly depends on the chosen target/projectile/ energy parameters set, (ii) how deuteron production routes may sometimes be more effective than the proton ones or lead to a different impurity pro ﬁ le and (iii) how α -particle beams may allow to bypass the limitations occurring when using Z = 1 beams. An overview of possible advantages and drawbacks of the cited production routes and of potential cross sections that stillneedto be measured, isalso reported.


Introduction 1.Radionuclides in nuclear medicine
Nuclear Medicine is a medical field that uses radionuclides both for diagnosis and therapy.The physical characteristics of radioactive decay of radionuclides determine their use in nuclear medicine [1][2][3].In most cases, radionuclides must be labelled to a vector molecule, forming what is commonly called a radiopharmaceutical.The vector molecule allows to efficiently target a dedicated cell type, due to its high affinity for it, or for a specific function of the human organ.In general, radionuclides employed in nuclear medicine both for diagnostic and therapeutic purposes, should meet the following specific requirements: appropriate physical properties for the selected application, in terms of half-life (T 1/2 ), decay mode and radiation emission energy; suitable chemical properties to allow their labelling to a vector molecule with high radiochemical yields through the labelling process; ability to deliver an appropriate radiation dose to the target tissue, sufficient to accomplish the intended clinical duty (diagnosis or therapy), but without causing damages to the patient; an appropriate purity, as the vector molecule can be expensive or the number of targeted sites limited; a reasonable price.
A new frontier of nuclear medicine is the so-called "theranostic approach" that combines therapy and diagnosis stages (preferentially PET imaging that can give a quantitative dosimetry) [4], and which includes a single radiopharmaceutical that: has both diagnostic and therapeutic properties, i.e. a bioactive molecule labelled with an isotope whose radiation decay pattern is suitable both for diagnosis and therapy, such as 47 Sc, 67 Cu, 67 Ga, 111 In, 117m Sn 123 I, 131 I, 153 Sm, 177 Lu, 186 Re, 213 Bi, etc.; as previously noted, the use of gamma-emitters instead of positron-emitters is recently attracting considerable interest [4]; can be labelled with different radioisotopes of the same element, one being utilized for diagnosis while the other one for therapy, such as 43/44 Sc/ 47 Sc, 61/64 Cu/ 67 Cu, 68 Ga/ 67 Ga, 86 Y/ 90 Y, 124 I/ 131 I, the terbium family 149/152 Tb/ 155/161 Tb, etc. [5]; can be labelled with different isotopes each belonging to a different element but having similar chemical properties (surrogate pair), where one radionuclide is utilized for diagnosis and the other for therapy, such as 68 Ga/ 177 Lu, 68 Ga/therapeutic lanthanide, etc.
The theranostic approach allows for collecting pre-therapy information with low-dose diagnostic imaging, aimed at optimizing the following high-dose therapeutic administration to the patient [6].This strategy is getting closer to the realization of personalized medicine, in which lowdose molecular imaging performed with SPECT/CT or PET/ CT is a necessary step to provide preliminary information about biodistribution, dosimetry on the tumor and limiting or critical organs or tissues, and the maximum tolerable dose [7].
Whatever the radionuclide to be used in nuclear medicine, it is thus essential to get knowledge of the possible presence of coproduced radionuclidic impurities that can have an impact on the final product quality and in the waste management during the production process or in the use at the hospital.Moreover, the availability of the majority of theranostic radionuclides is mainly limited by the fact that most of them are currently emerging and the optimal production route still needs to be improved or defined [8,9].

The use of accelerators for emerging medical radionuclides production
Most of the radionuclides used in nuclear medicine are artificially produced by using either nuclear reactors or accelerators delivering charged particle beams or inducing a γ-irradiation.This work is focused on light-charged particles induced reactions, even if there is an increasing interest on photo-production as for example for 47 Sc, 67 Cu, etc. [10][11][12].
In order to select the optimal way to produce a given radionuclide, it is mandatory to explore all possible nuclear reaction routes and thus to measure the nuclear cross sections for the concerned radioisotope, as well as of related main contaminants.Among the possible production routes, proton or light-charged particles (e.g.deuteron, 3 He and 4 He) are of great potential, mainly for positron emitters and some therapeutic radionuclides [13,14].In recent years, high energy (70 MeV) and multi particle accelerators have been made available from accelerators' manufacturers.This allows to widen the range of possible nuclear reaction routes which may be explored and radionuclides that may be thus produced (e.g. 211At or 82 Sr, 68 Ge, etc.).Examples in Europe are the ARRONAX facility [15] and the SPES cyclotron used in the LARAMED project [16].

The importance of accurate nuclear cross section determination and thick target yields (TTY) measurements in medical radionuclides production
By definition, the cross section σ(E) is the probability of a specific nuclear reaction to occur at the given energy E, i.e. the interaction probability of one bombarding particle and one target atom.This quantity is required to calculate the production yield of the radionuclide of interest.Indeed, the resulting activity at the End Of Bombardment (TTY EOB ) of a given radionuclide, produced by a particle beam hitting a thick target, which also depends on the irradiation time t IRR and the decay constant λ [s −1 ], is calculated considering the following formula [3,17]: where A T is the atomic weight of the target material [g⋅mol −1 ] and S T (E) is the mass stopping power of the particle beam through the target material at the energy E, expressed as dE/d(ρx).The term (1 − e −λt irr ) is the so-called saturation factor (SF) that takes into account both the production of nuclei due to the considered nuclear reaction and the decay of nuclei that have already been produced.In a thick target, the impinging projectile energy decreases as it penetrates into the target thickness, thus the energy variation against target thickness, and the related variation of the cross section versus energy, must be considered.The parameters that have an impact on the production yield (TTY) are the beam intensity, the irradiation time and the beam energy range: the beam intensity is limited by the maximum beam current available by the accelerator and also by the thermal-mechanical main features associated to the targetry; the irradiation time can be increased, keeping in mind that the SF trend keeps linear until the irradiation time remains below one T 1/2 ; the projectile energy range can be adapted to exploit the maximum cross section value, considering the limitation due to the coproduction of contaminants, by adjusting the incident energy and the target thickness.The presence of contaminants, especially radionuclides with close to or longer half-life than the desired radionuclide, affects the purity of the final product and may cause a dose increase to the patient.
To limit as much as possible the reaction routes which may be opened, and thus the coproduction of undesired radionuclides, it is often necessary to use expensive isotopically enriched materials.Even considering the ideal case of a fully monoisotopic target, it is foreseen the coproduction of isotopic contaminants, depending upon the beam energy range chosen.A good knowledge of production cross sections for all the occurring nuclear reaction routes is therefore a key point that has to be carefully considered.The production cross section of the concerned radionuclide, as well as those of main contaminants, the optimal irradiation parameters may be at last selected, i.e. the irradiation time and beam energy range, in order to ensure the achievement of the best final product purity given by the Specific Activity A S [Bq⋅g −1 ], the Radio Nuclidic Purity RNP [%] and the Isotopic Purity IP [%] parameters [2,18].

Experimental details on a nuclear cross section measurement
The use of a single thin homogeneous target foil, followed by a precise Faraday cup to measure the hitting beam current, is the best method to obtain an accurate measurement of a specific nuclear cross section.However, due to the limited access to beam time it is often preferred to use the well-known stacked-foils technique that allows, by the simultaneous bombardment of patterns composed of target and degrader foils, to get the experimental determination of some cross section values at different beam energies during the same irradiation run.In order to control the beam energy loss through the target stack, it is desirable to insert in the structure at least two monitor foils of an element for which the production cross section of a given reaction route is well known and that acts as a reference.Particular attention has also to be paid to the material used in the stack both for degrader and monitor foils, since the recoil effect, i.e. when some atoms jump out of one foil and are trapped in the following one, may also cause some additional interferences in the γ-spectrometry acquisitions.
In addition to the aforementioned parameters that have to be taken under control for a precise nuclear cross section measurement, it is important to remind the role of the target thickness (homogeneity and purity), the beam stability during the irradiation run and the beam energy precision (the possibility to use a tunable energy accelerator is favourable), the detector calibration (energy and efficiency) and the possible interferences in the emitted γ-lines or in the decay relations among metastable/ground state or parent/daughter radionuclides.

Theoretical aspects on nuclear cross sections assessment
Theoretical description of nuclear reactions, occurring below 100 MeV/amu, imply many different interaction processes, such as the compound-nucleus reaction, the direct nuclear reaction, and the pre-equilibrium nuclear reactions [19].Several nuclear reaction codes are available that include models that implement these different nuclear reaction mechanisms between a target nucleus and a hitting particle.The most used ones are TALYS [20,21], EMPIRE [22], FLUKA [23,24] and MCNPX [25].They are usually exploited to estimate the trend contributions for those nuclear reaction routes, whose cross sections data are not measured yet.The IAEA tool ISOTOPIA [26] is instead based on the IAEA data as well as on the TENDL library [27], a nuclear database built on the TALYS code.TALYS is used to simulate nuclear reactions involving protons, neutrons, deuterons, photons, tritons, 3 He and α-particles, in the 1 keV to 200 MeV energy range and covers target nuclides of mass 12 and heavier.TALYS includes a large number of different models and some of them are invoked in calculation by default.Concerning compound reaction contributions, TALYS 1.9, the last version available to date, performs default calculations in the Hauser-Feshbach approach with Moldauer width fluctuation corrections.Koning and Delaroche local parametrization for the optical model is used with different models to take into account direct reaction contributions: Distorted-Wave Born Approximation (DWBA) Coupledchannels (rotational and vibrational approaches).Since TALYS has been used in many works to benchmark the experimental determination of nuclear cross sections for many radionuclides of interest in medicine, we mainly refer to this code in this review study.We focused on TENDL-2021 which is based on the last version of the TALYS code.However, in some cases the TENDL-2021 cross sections are different from the earlier versions (and in few cases less in agreement with available data), underlining the need of additional work, both for new experimental data to constrain the code and in the optimal identification of the TALYS parameters that better describe the nuclear reactions for this kind of applications [28].

Charged-particles induced reactions
Nowadays, most accelerator-based radionuclides are produced using proton beams.This is mostly linked to the fact that there is a large installed network of low-energy cyclotrons, often called Small Medical Cyclotron (SMC) [29], and that protons can be accelerated as negative ions that are extracted with nearly 100% efficiency using stripper foils.Alpha particles may instead only be accelerated as positive ions and the extraction process is more challenging.With the larger availability of both medium and high energy accelerators, new linear accelerator like SPIRAL2 [30] or the one available at SOREQ [31] providing p, d, α-particles beams with mA intensities, and multiparticle cyclotrons allowing access to high energy deuterons and α particles beams, the production routes for established as well as emerging radionuclides may thus be re-investigated.In all cases, by changing the projectile type and/or its energy and/or the target material, it is possible to adapt the production yield and the profile of contaminants [32].In the following examples, we will present some cases showing that: the quality of the final product strongly depends on the chosen target/projectile/energy parameters set ( 67 Cu and 47 Sc cases); the deuterons-based production route may sometimes lead to a different impurity profile ( 67 Cu and 89 Zr examples) or be more effective than the protons-based one ( 103 Pd and 186 Re case).This latter property is basically due to the more complex deuteron-induced interaction process with respect to the proton one (i.e.involving the breakup mechanism, direct reaction stripping, and preequilibrium and compound-nucleus) that have been studied in different former works which details may be found in [33][34][35].Because of the weak binding energy of deuterons, B d = 2.224 MeV, a more complex interaction process indeed occurs, that involves different reaction routes started by neutrons and protons following the deuteron breakup that enhance the production route.Moreover, a reliable description of the deuteronnucleus interaction is a good test for the quality of the reaction mechanism modeling, as well as the evaluation of the nuclear data; -α-particle beams may allow to bypass the limitations occurring when using proton beams ( 97 Ru and 211 At examples).A complete description of potential of alpha beam can be found in [14].
Table 1 shows the decay characteristics, extracted from the NuDat 3.0 database [36], of the medical radionuclides considered in this review.When available, the recommended IAEA curves [37][38][39] will be used instead of the available data sets [40] for simplicity and readability.
2.1 67 Cu 200-300 μm range in soft tissue), while the 185 keV photons are suitable for pre-therapy low dose imaging (SPECT or SPECT/CT).These physical characteristics allow the use of 67 Cu for both therapy and associated diagnostic applications, thus making 67 Cu a promising theranostic radionuclide.It can be also used as a therapeutic counterpart of the 64 Cu (12.7 h half-life and β − /β + emitter), a PET radionuclide [5,41,42].Production techniques for 64 Cu are well known and are usually based on the 64 Ni(p,n) 64 Cu and 64 Ni(d,2n) 64 Cu reactions routes [43], even if there are other possibilities [43,44].The use of 67 Cu has been prevented by the lack of regular availability in enough amounts for preclinical and clinical studies, although it is possible to produce it through many production routes [9,45,46].There is currently not a clear consensus about the best nuclear reaction route to be used for a massive production, due to the impact of the coproduced isotopic 64 Cu that occurs with some specific routes [47].Indeed, the presence of the 64 Cu annihilation photons affects the total dose received by the patient and the staff, while the emitted Auger electrons, the positron and the electrons release their energy close to the decay point, adding their contribution to the dose in tissues.
Nuclear data are therefore important to help in defining benefits and drawbacks associated with each of these production routes.Among the feasible ones are the following: - 68 Zn(p,2p): this production method has quite a low cross section (maximum value of around 10 mb) [40] and requires intermediate energy proton-beams (30 < E < 100 MeV), although it can be exploited also at higher energies [48][49][50][51][52][53].In order to limit the coproduction of other Cu-radionuclides, which directly affects the radionuclidic purity of the final product, the use of enriched 68 Zn targets is mandatory, as well as the recycling of the target material to make this production route affordable.Among the variety of radionuclides produced during the bombardment of a zinc target, 67 Ga (half-life 3.2617 d) is particularly relevant: it presents the same γ-lines of 67 Cu since they both decay into 67 Zn [36].Moreover, they have a similar half-life, and therefore it is not possible to infer the precise activity determination of one radionuclide waiting for the decay of the other one.For these reasons a radiochemical procedure, aimed at the separation of copper from gallium elements, is thus mandatory to obtain a precise cross section for 67 Cu production [49,54].The separation procedure, as well as the use of different target materials (natural or enriched), selected monitor reactions or not up-to-date decay data, could explain the large discrepancies of the published data.The measurement carried out at the ARRONAX facility led to an estimated difference on the 67 Cu production yield of up to 15% in the 70-35 MeV energy range [54].This difference has a direct impact on the planning of sustainable production of 67 Cu for medical purposes.It should also be outlined that with this production route a large amount of 64 Cu is coproduced, as indicated by the 68 Zn(p,x) 64 Cu cross section recommended by the IAEA up to 100 MeV [37]. 64Cu can be limited in the final product by waiting for its decay up to the desired value of the 67 Cu radionuclidic purity, however loosing part of the 67 Cu activity. - 0 Zn(p,α): this nuclear reaction, already measured up to 35 MeV by Levkovskij (1991) [50] and Kastleiner et al.
(1999) [55], has a maximum cross section value of 15 mb at about 15 MeV.This production route can allow the supply of pure 67 Cu as there is no coproduction of 64 Cu below E thr = 23.8MeV (threshold energy associated with 64 Cu production).Considering the low cross section value this production route provides quite a low yield, i.e. 5.75 MBq/µAh for the 30-10 MeV energy range (corresponding to a 1.79 mm thick target of 100% enriched 70 Zn).Also in this case, enriched target material and recycling are mandatory. - 0 Zn(p,x): The use of 70 Zn targets at proton energies higher than 35 MeV allows to increase the 67 Cu yield (Figure 1 left), limiting the 64 Cu coproduction in comparison to the use of 68 Zn targets [56].The 67 Cu production cross section reaches a value around 22 mb at 70 MeV, thus it is almost a double value than the recommended cross section by the IAEA, based on 68 Zn target.In addition to that, the 64 Cu coproduction occurring with 70 Zn targets shows a decreasing trend versus the increasing energy, while the opposite trend can be noted when using 68 Zn targets (Figure 1 right).TENDL library values [27] seem to overestimate the 70 Zn(p,x) 67 Cu reaction, as well as the 68 Zn(p,2p) 67 Cu cross section (Figure 1 left).On the other hand, TENDL results of the 68 Zn(p,x) 64 Cu cross section slightly underestimate IAEA recommended values and show an energy shift of about 5 MeV for the peak below 30 MeV (Figure 1 right); the trend of the 70 Zn(p,x) 67 Cu cross section is well described by the TENDL-data, but there is an underestimation of the experimental values by about a factor of 2. - 70 Zn(d,x): this nuclear reaction has already been measured in 2012 [57] and in 2021 [58] up to 29 MeV.From the most recent data it was possible to identify the cross section peak value associated with this nuclear reaction (about 30 mb at 23 MeV).The comparison between the cross sections data obtained with 70 Zn as target material for proton or deuteron beams (Figure 2) shows that the deuteron production route is more efficient than the one using protons.Furthermore, since the threshold associated with 64 Cu production is 26.5 MeV, this isotope is not produced in significant amount up to 30 MeV, allowing for a "pure" 67 Cu production.The preferred energy range through this route is thus 16-26 MeV, that corresponds to a 70   64 Cu production is 23.7 MeV, this radioisotope is not produced in large amount up to 30 MeV allowing, also in this case, a "pure" production of 67 Cu [63].The main advantage of this production route is the fact that enriched 64 Ni target (natural abundance equal to 0.9255%) is well mastered as it is the target used for 64 Cu production.A pure 67 Cu production requires the use of enriched 64 Ni and recycling, making the enrichment level of the target material a crucial aspect (currently it is commercially available with an enrichment level >99%).
The values reported in Table 2 show that with a hypothetical fully enriched target material ( 68/70 Zn or 64 Ni) the   most convenient route to obtain 67 Cu (without 64 Cu coproduction) is by using deuteron beams and 70 Zn targets.The final 67 Cu quality using this route will be close to that obtained by photo-production [64,65].If some 64 Cu coproduction is acceptable for clinical application, the 68 Zn(p,2p) 67 Cu reaction seems to be a better option since it provides a larger 67 Cu yield in comparison with the 70 Zn(d,x) 67 Cu route mostly because more intense proton beams can be available (hundreds of µA as compared to tens of µA for deuterons).With the same motivation, also the high-energy 70 Zn(p,x) 67 Cu route seems to be interesting.The α-beam production route may benefit soon from the newly available multiparticle 30 MeV accelerators providing tens of µA.The first two machines have been installed in Europe (Jülich and Warsaw) and some others are available in Japan.In the future, linear accelerators that can provide very intense deuteron and α-beam (mA range) are soon foreseen, however in this case the major challenge will be the design of specific targets able to withstand such high currents.

47 Sc
47 Sc (half-life 3.3492 d) is one of the radionuclides of interest in theranostic applications thanks to its β − and γ-radiation (β − mean energy 162 keV, 100% intensity; E γ = 159.381keV, 68.3% intensity) which makes it suitable for both therapeutic and diagnostic purposes [41,42].The interest in 47 Sc is also due to the possibility to pair it with a β + emitter isotope, such as 44g Sc and 43 Sc, allowing to carry out also PET imaging [66]. 47Sc can be produced via different nuclear reaction routes by using cyclotrons, nuclear reactors, and electron linear accelerators [5]; moreover, 47 Sc may be produced either directly or by the decay of its parent radionuclide, 47 Ca, as shown in Table 3. Particular care must be paid to the other Sc-isotopes produced during irradiation, since they cannot be chemically separated from 47 Sc.Among the Sc-isotopic contaminants, the most critical one is 46 Sc, due to its long half-life (half-life 83.79 d) and γ emissions (889.277keV with 99.9840% intensity and 1120.545keV with 99.9870% intensity).
Considering the proton induced-reactions, the: -nat Ti(p,x) 47 Sc cross section has been widely studied by many research groups, but no energy range has been found in which the 47 Sc is produced without 46 Sc.Considering the composition of nat Ti (natural abundances: 50 Ti 5.18%, 49 Ti 5.41%, 48 Ti 73.72%, 47 Ti 7.44%, and 46 Ti 8.25%) it is not possible to extract from the whole cross sections the contribution of each titanium isotope to the production of the specific radionuclides of interest.Since no data are available for the 49 Ti(p,x) 47 Sc cross section and only a few data sets are published for the 50 Ti(p,x) 47 Sc [45] and the 48 Ti(p,2p) 47 Sc [50,67,68] reactions, both studied up to 85 MeV, it will be desirable to add new experimental data in order to evaluate if one of these production routes can be competitive. - 8 Ca(p,2n) 47 Sc excitation function was measured in 2019 up to 30 MeV [69] reaching a value of 1100 mb at about 18 MeV.However, it must be taken into account that, despite the large cross section value, the natural abundance of 48 Ca is only 0.187%, thus making this production route feasible with highly-enriched targets only (enrichment levels available up to 97%), that are anyway very expensive (recycling is mandatory).-nat V(p,x) 47 Sc cross section has been measured by many authors [28,50,[70][71][72][73][74] up to 200 MeV, reaching a  47 Sc and contaminants coproduction using nat V targets, compared with theoretical estimations obtained by varying the level densities using the TALYS code [28].Considering nat V targets it is possible to limit the coproduction of the 46 Sc by using the energy range 35-19 MeV.The theoretical 47/46 Sc production estimates based on the experimental results and the well-known DOTA-folate conjugate cm10 ([ 47 Sc]-cm10) biodistribution, allowed to evaluate the dose increase due to the presence of the 46 Sc contaminant [75].
The calculations showed that for E p < 35 MeV and specific irradiation conditions considered, the effective Dose Increase (DI) due to the presence of contaminants was maintained within the 10% limit required for clinical use [75,76].However, to limit the presence of the 46 Sc contaminant, the calculated 47 Sc yield in the 35-19 MeV energy range is rather low, i.e. 0.1 GBq (irradiation conditions: 1 µA proton beams and 24 h of irradiation time), still suitable for preclinical studies [75].
Regarding the indirect reactions induced by protons to obtain the 47 Ca → 47 Sc precursor system, only few data are available in the international database [40].The 51 V(p,x) 47 Ca cross section was measured up to 250 MeV with a maximum value of about 0.2 mb making this route unfeasible [69].No experimental data are present for the 50 Ti(p,x) 47 Ca and the 49 Ti(p,x) 47 Ca nuclear reactions, whose threshold energies are 24.5 and 21.2 MeV, respectively.On the other hand, the trend of the 48 Ca(p,x) 47 Ca cross section was measured up to 19 MeV in 2019 [69], showing a maximum value of about 240 mb at 17 MeV.The low natural abundance of the target material (0.187%) has to be reminded that would make this route very expensive.
Considering the deuteron-induced reactions to produce 47 Sc, the: -nat Ti(d,x) 47 Sc excitation function has been measured by several authors [77][78][79][80][81][82][83][84], showing a good agreement in the increasing trend of the reaction cross section from low energy up to 50 MeV.Also the co-production of 46 Sc has been measured by different research teams and it is not possible to indicate a deuteron energy range where there is a pure 47 Sc production [77][78][79][80][81][82][83][84][85][86]. -5 Ti(d,αn) 47 Sc cross section was measured in 1964 [87], showing an increasing trend up to about 65 mb at 17.5 MeV; unfortunately, there are no data on the coproduction of 46 Sc, whose production threshold energy is 4.6 MeV. - 49 Ti(d,α) 47 Sc excitation function was measured in the same experimental campaign in 1964 [87], depicting a peak of about 45 mb at around 10 MeV.In this case, the 46 Sc coproduction was measured in the same energy range (7-20 MeV), obtaining an increasing trend with a plateau value of about 25 mb from 17.5 MeV up to 20 MeV; the lowest 49 Ti(d,αn) 46 Sc cross section value is at 12 MeV, amounting to ca. 7 mb. -4 Ti(d,x) 47 Sc cross section has never been measured (E thr = 6.2 MeV).In the 1960s the 48   47 Sc and 46 Ca(d,n) 47 Sc cross sections were not measured, and also not for 46 Sc coproduction; the threshold energies are respectively 11.4 and 22.5 MeV for 47 Sc and 46 Sc production with 48 Ca targets, 0 and 4.8 MeV in case of 46 Ca targets.-nat V(d,x) 47 Sc cross section has been measured up to 90 MeV by three research teams, with agreeing results [90][91][92].The cross section starts at 25 MeV and it reaches the maximum value of about 30 mb at ca. 40 MeV, then it decreases to 20 mb at 60 MeV with almost a flat trend up to the maximum energy.Unfortunately, the co-production of 46 Sc can be avoided only for deuteron energies <30 MeV.Below 40 MeV there is a low coproduction of 46 Sc, since its production cross section is still at 5 mb, but it rapidly increases up to about 50 mb at ca. 55 MeV, keeping a quite flat trend up to 90 MeV.
Regarding the indirect reactions induced by deuterons to obtain the 47 Ca, it is possible to irradiate nat V targets: however, the production cross section is quite low (<0.4 mb) in the entire 30-90 MeV energy range investigated [91,92].
Considering the α-induced reactions, the: -44 Ca(α,p) 47 Sc cross section has been measured up to 45 MeV reaching a maximum value for 47 Sc production of about 120 mb at 14 MeV [50]; the same author also measured the coproduction of 46 Sc, finding a peak of about 400 mb at around 25-30 MeV.Although the natural abundance of 44 Ca is only 2.086%, the advantage in the use of α-beams lies in the short range of α particles in Ca targets, allowing the use of a limited amount of enriched 44 Ca target material, compared to the use of proton and deuteron beams.The work published by Minegishi et al. [93] also reports the 47 Sc yield of about 11 MBq at the end of preparation (approximately 1.5 h from the EOB), obtained by irradiating a [ 44 Ca]CaO ( 44 Ca = 97.0atom%, 40 Ca = 2.89 atom%) target of 200 mg in the energy range 28→0 MeV at 10 eμA, for 2 h irradiation.To obtain pure 47 Sc it is necessary to take into account the expected large coproduction of 46 Sc, not reported in the work by Minegishi et al. [93]. -4 Ca(α,t) 47 Sc cross sections are not known [40]; the threshold energies are respectively 12.3 and 23.9 MeV for 47 Sc and 46 Sc production.
In conclusion, some charged-particle induced reactions need to be measured to verify if they are feasible as 47 Sc production routes, e.g. 48/49/50 Ti(p,x) 47 Sc, 47 Ca → 47 Sc, 47/48/49/50 Ti(d,x) 47 Sc, 46 Ca/ 48 Ca(d,x) 47 Sc and 46 Ca(α,x).From all the data available it turns out that the only ways to get 47 Sc without 46 Sc using light charged particle induced reactions lead to low production yield, not suitable for clinical use.It is thus necessary to study how to add a mass separation step to remove 46 Sc.Also in this case further cross section measurements will be needed to identify a precise impurity profile associated with each possible production route and it will be necessary to carefully look at mass separation technique, presumably coupled to laser ionization to be able to extract pure 47 Sc with high efficiency [94,95].

89 Zr
Thanks to its nuclear characteristics (T 1/2 = 78.41h, 22.3% positron emission with a maximum energy of 900 keV), zirconium-89 ( 89 Zr) is a very promising radionuclide for immuno-PET [96].Different experimental results obtained to produce zirconium-89 from a 89 Y target [97] may be compared by using both proton and deuteron beams [50].The proton thick target yield is higher up to 26.5 MeV leading to higher production yield, whereas above 26.5 MeV the achievable yield with protons is lower than that obtained with deuterons [96,98].At the same time, a lower amount of radionuclidic impurities is produced by deuteron irradiation.For example, to keep a radionuclidic purity of 99.9%, it is possible to use a higher beam energy for deuterons (20.5 MeV) than for protons (16.2 MeV) which turns out in about 10% higher production yield for deuteron.
It is also possible to investigate other routes, such as the reactions 87,88 Sr(α,xn) 89 Zr [50] that can be the object of further investigation as they seem competitive in the production of 89 Zr with possibly high radionuclidic purity.

103 Pd
103 Pd (T 1/2 = 16.991d) decays almost exclusively (99.90%) by electron capture (EC) to 103m Rh (T 1/2 = 56.12min) which de-excites through internal transition (IT) [36].As a result of these processes (EC and IT) Auger-electrons and X-rays are emitted which are ideally suited for cancer therapy.Considering also the de-excitation of the "daughter" nuclide 103m Rh, every 100 decays of 103 Pd are accompanied by the emission of about 263 Auger electrons, 188 lowenergy conversion electrons and 97 X-rays.A comparison between the IAEA curve for the deuteron production and the IAEA curve for the proton production of 103 Pd [99] shows that the yields are comparable up to 12 MeV.For higher particle energies the achievable thick-target yield with deuterons is higher.
There are only two works reporting the production of 103 Pd by particles other than protons and deuterons [100][101][102].Some of these reactions, mainly 101 Ru(α,2n) 103 Pd and 103 Rh( 3 He,x) 103 Pd, may be worth the effort for further investigation in order to evaluate the achievable specific activity and the radionuclidic purities.

186g Re
186g Re is a β − /γ emitter radionuclide with a half-life of 90.64 h.The β − end-point energies of 1.07 and 0.93 MeV suggest that this radionuclide is a good candidate for therapy of cancers with small tumour dimensions (from few millimetres to a few centimetres) [103].Moreover, its γ -emission at 137.15 keV (very close to that of 99m Tc) is in the energy range suitable for SPECT imaging (i.e., sufficiently high energy to penetrate the body and sufficiently low to be collimated by thin collimators and detected by thin detectors) [41,42].
Very high specific activity 186g Re can be produced by either proton [104][105][106][107][108][109] or deuteron [77,[110][111][112][113][114][115][116] cyclotron irradiation [117], provided an enriched target is used.It is possible to compare the experimental results obtained to produce rhenium-186g from a nat W target using proton and deuteron beams.The cross section values for the deuteron beams are considerably greater than for the proton beams.The maximum value that is possible to obtain with deuterons is almost 9 times greater than the maximum value for the protons.This effect is also confirmed by the study of thick target yield [118].It is possible to compare the integrated thin target yields of 186g Re by using proton or deuteron beam: the two curves are coincident till energy of almost 8 MeV.From this energy up to 9 MeV, the yield of the proton bombardment is slightly larger than the deuteron one but starting from 9 MeV the deuteron yield grows much faster than that of the proton irradiation.
Even if there is a very short energy range in which proton irradiation is competitive with the deuteron one, there is no reason to prefer proton beams.In fact, the advantage of using deuterons over protons for the preparation of 186g Re is double: more 186g Re is formed and less 186 W is needed for the target, due to larger -dE/dx in the case of deuterons, by a factor of 4. This lower amount of target material is advantageous both to lower production cost and to simplify the dissolution of the target and separation from the produced 186g Re, with advantages also for waste and radioprotection purposes.This makes the deuteron route at facilities with medium-size cyclotrons the best choice for 186g Re production.
It is also possible to investigate other routes [119][120][121] among which the 192 Os(p,α3n) 186g Re [119] and 186 W(α,x) 186g Re [120,121] reactions may be of some interest.However, there is a high probability that other rhenium isotopes are produced and will affect the specific activity and the radionuclidic purity of the final product.

97 Ru
The radionuclide 97 Ru, with a half-life of T 1/2 = 2.83 d, emits low-energy high-intensity γ lines (215.7 keV with 85.8% intensity and 324.5 keV with 10.8%) which have favourable characteristics for SPECT imaging.This can be beneficial to select a patient that may respond to new chemotherapy agents developed with ruthenium [122].It decays by electron capture (EC) leading to Auger emission that can be used for Auger therapy.It can form a theranostic matched pair with 103 Ru (T 1/2 = 39.26 d) that decays to the short-lived Auger emitter 103m Rh (T 1/2 = 56.12min), a promising γ-free therapeutic agent.
Production of 97 Ru can be done by irradiation of 99 Tc with protons [123].Unfortunately, Tc as a radioelement (T 1/2 = 2.111⋅10 5 y) is not easily available and it requires a dedicated infrastructure to be used as target material.Other reactions with proton-beams can be measured [124][125][126], but it seems more attractive to explore the use of an α particle beam coupled to a molybdenum target that is easily available as natural molybdenum or as enriched material [127][128][129].Recently, new cross section values [130] were measured toward higher energy in coherence with commercially available cyclotrons [15] which are able to deliver α-beam up to 70 MeV.These data allow the optimization of nat Mo(⍺,x) 97 Ru production route and constrain the nuclear codes to explore the possibility to use enriched Mo targets.Figure 3 shows recent data reported in the literature for the cross sections of nat Mo(⍺,x) reactions to produce 97 Ru [123,[130][131][132].
A plateau is reached above 20 MeV coming from the fact that nat Mo is a mixture of six stable isotopes with comparable abundance.If enriched 95 Mo or 96 Mo targets are used instead, one can increase the production by a factor of five-six thanks to the natural abundances of these isotopes, that are respectively 15.84 and 16.67%, and considering that the beam energy can be defined adequately to favour the (α,2n) and (α,3n) reactions.Using a nat Mo target with a thickness of 100 mg/cm 2 and a 30 MeV beam energy, 3.5 MBq/µAh can be obtained.This example shows the benefit of having an α beam to allow for an alternative route when the target material is not easily available (see also [14]).

211 At
211 At (Z = 85) is an α emitter that is promising for targeted α therapy.It has a half-life T 1/2 = 7.214 h and it emits one α particle per decay, either directly decaying to 207 Bi (41.80%, Eα = 5.869 MeV) or indirectly through EC (58.2%) to 211 Po, that quickly (T 1/2 = 0.516 s) decays to 207 Pb (E α = 7.3695 MeV).In the emitted radiation by 211 At there are also X-rays that can be used for imaging purpose.As the heavier stable element in the radionuclide chart is 209 Bi (Z = 83) and there is no long lived radionuclide in the close vicinity of 211 At, it is not possible to use low or medium energy proton or deuteron beams to produce it.The only possibility is to rely on an α particles beam impinging on a bismuth target (Z = 83).Bismuth is by chance monoisotopic, which makes target fabrication not expensive and easy to manufacture by evaporation under vacuum.
The energy threshold for 211 At production through the (α,2n) reaction is 20.7 MeV.In Figure 4, the family curves of integrated thick target Yields Y(E,ΔE) and the loci of the maxima, corresponding to optimal irradiation conditions for production, are presented.From these curves, it is clear that increasing the incident energy allows increasing the 211 At production.Unfortunately, 210 At (T 1/2 = 8.1 h) can also be produced by irradiation of a bismuth target with an alpha beam starting at 28.613 MeV (energy threshold of this reaction).This radionuclide must be avoided as much as possible, as it decays to the bone seeker 210 Po (T 1/2 = 138 d) that decays by α emission and it is highly radiotoxic.One possibility is to restrict the beam energy below the energy threshold of 210 At production but this will limit the production yield of 211 At at the same time.The rise of the 210 At production cross section is slow and there is room to optimize the irradiation parameters if one can define the acceptable limit of 210 At in the final 211 At product.This point is of fairly high importance as beam delivered by accelerators have some energy dispersion.In Table 4 the measured 210 At/ 211 At activity ratio are presented [133][134][135], showing how the ratio evolves as a function of the beam energy close to 210 At production energy threshold.From these data, it is clear how the choice of an incident energy of 28.8 MeV, slightly higher than the threshold energy (28.613MeV) for the 209 Bi(α,3n) 210 At reaction, allows to produce limited quantities of the 210 At radionuclide impurity, remaining below 1%, considered as a limit value.
More precise cross section values close to the 210 At threshold may help optimize production yield as a function of 211 At purity.In parallel, some toxicology studies on the impact of 210 At content in the final 211 At product are needed to get a consensus on the acceptable level of 210 At for medical application.

Discussion and conclusion
Through the radionuclides discussed in this review ( 67 Cu, 47 Sc, 89 Zr, 103 Pd, 186g Re, 97 Ru, 211 At) it has been high-lighted how the accelerated beam type and/or its energy and/or the target material have a direct impact on the production yield, the purity and the contaminants profile of the final product to be used for radiopharmaceuticals labelling procedure.Since the quality of the final product strongly depends on the chosen target/projectile/energy parameters set, the accurate knowledge of the production cross section of the desired radionuclide and related contaminants is the key point for the irradiation parameters optimization.Considering the frequently need of using isotope-enriched materials to limit as much as possible the coproduction of contaminants, the possibility to recover and reuse the irradiated target materials is an additional key aspect for the economical sustainability of the production cycle.For this reason, it is also useful to study the coproduction of radioactive isotopes (especially the long-lived ones) that may affect the recovery process.
An interesting example of this work is the 67 Cu case, whose production is usually based on the well-known 4: Integrated thick target yield for 211 At as a function of energy and energy loss into the target, with the loci of maxima corresponding to the couples (E,ΔE) for optimal irradiation conditions to produce the radionuclide with higher as possible yield, radionuclidic purity and specific activity [134].(p,2p) reaction on 68 Zn targets; however, the high-energy 70 Zn(p,x) 67 Cu route seems to be an interesting option and even a favourable route above 56 MeV, not only for the higher 67 Cu production yield (almost double), but also for the lower 64 Cu coproduction.From a practical point of view, it has to be outlined that the price of the enriched 68 Zn and 70 Zn material differs by a factor of about four, considering that the natural abundances of these zinc isotopes are 18.45 and 0.61%, respectively.The expenses of the enriched 68/70 Zn isotopes can be amortised considering the possibility to recover and reuse the irradiated material, to manufacture new targets for a defined number of production cycles.In this case, it is important to know the production cross sections of the long-lived 65 Zn isotope (243.93 d half-life), to properly define the methods for the recovery process considering the activation of the material, minimizing as much as possible the dose imparted to the operator.With the development of high-energy high-intensity deuteron beams, the 70 Zn(d,x) production route will become attractive as it allows to obtain high-purity 67 Cu, as provided through the use of e-LINAC exploiting the 68 Zn(γ,x) 67 Cu reaction [65].
As previously shown, the use of deuteron beams instead of the proton ones can sometimes be more effective (as for examples the 103 Pd and 186 Re case) or lead to a different impurity profile (e.g. the 67 Cu and 89 Zr cases).The criticality of these production routes is linked to the practical availability of accelerators with deuteron beams providing beam energies larger than 9 MeV and reasonable intensity (≥50 μA).On the other hand, considering radionuclides with a mid half-life as for the 186g Re (3.7186 d) and 89 Zr (78.41 h) cases, nothing prevents to produce in centralized facilities with medium-size cyclotrons and ship them to the clinical centres over long distances.Also the increasing availability of intense α-beams may allow to bypass the limitations occurring when using protons or deuterons: this is the case of 97 Ru and 211 At production, and the possibility to investigate the new route for 186g Re based on the 186 W(α,x) 186g Re cross section.On the other hand, also considering proton beams some interesting production routes have still to be investigated: as examples it can be reminded the 47 Sc production and the use of enriched Ti-targets, such as 49 Ti and 50 Ti.
A paradigmatic case is linked to the supply of the medically relevant Tb-isotopes, i.e. 149 Tb, 152 Tb, 155 Tb and 161 Tb [66].These isotopes can be produced with different projectile/target/energy ranges, obtaining different yields and contaminant profiles; among the possible nuclear reactions to be exploited, some of them have still to be investigated.For the 155 Tb case [136][137][138][139][140][141][142], the main routes are based on the 155 Gd(p,n) and 156 Gd(p,2n) reactions, as outlined by a recent work [143], finding a purer 155 Tb with the (p,n) reaction and a higher yield with the (p,2n) one.On the other hand, 155 Tb can be also produced with a 155 Dy→ 155 Tb precursor system, exploiting the 155 Dy decay (9.9 h half-life).In this case, 155 Dy can be produced with high-energy proton beams and natural Tb-targets, using the 159 Tb(p,5n) 155 Dy reaction [138,144,145].Particular attention has to be paid to the coproduction of Dy-isotopes that decay to Tb and that may affect the final 155 Tb purity: the irradiation, decay, separation and elution times have thus to be carefully studied to maximize the 155 Tb radionuclidic purity.The advantage of the 159 Tb(p,5n) 155 Dy→ 155 Tb route relies on the use of natural targets, even if the separation and purification chemistry among lanthanides is known to be challenging.The therapeutic 161 Tb is instead produced in nuclear reactors, exploiting the 160 Gd(n,γ) 161 Gd→ 161 Tb route.Considering the charged particles induced reactions [136,139,140,146], only deuterons can be competitive via the 160 Gd(d,x) 161 Tb route (with a cross section of about 200 mb at ca. 10-15 MeV) [146], while the use of α-beams is possible, but curtailed by scarce availability and the low production yield of the nat Gd(α,x) 161 Tb reaction [140].The 149 Tb radionuclide, under the spotlight of the international community as α-emitter, can be produced with high-energy protons [147,148]: in this case the coproduction of Tb-isotopes cannot be avoided, requiring the use of a mass separation method (either on-line or off-line) to select the desired radionuclide.The 152 Tb case is interesting since it is possible to compare the cross sections measured using proton, deuteron and α-beams on 151 Eu and nat Gd targets [137,142,147].A promising production route may be the nat Gd(d,x) 152 Tb reaction, that was measured up to 50 MeV obtaining an increasing trend with the maximum value of ca. 100 mb; however, the coproduction of contaminants has to be carefully evaluated.As for the 47 Sc case, it is particularly useful for the Tb-isotopes to take into account the possibility to use the mass separation method, as proposed by the ISOLDE-MEDICIS collaboration [94] and the PRISMAP consortium [95], to maximize the production yield and the final radionuclidic purity.
Reliable nuclear codes are useful tools to describe the production of the radionuclides difficult to be measured (such as stable isotopes, long-lived or short-lived radionuclides, etc.) and to evaluate their impact on the final product quality.Most of the available nuclear codes reproduce quite well the low-energy region of proton induced reactions.However, in the estimation of more exotic reactions induced by either high-energy protons or heavier projectiles (such as deuteron and α particles) there is room for improvements.This is also partly due to the scarce availability of experimental data to constrain the nuclear codes in these regimes, another indication that further investigations are needed.
Considering the future availability of multi-particle accelerators and high-intensity beams, possibly coupled to mass separation techniques, it is foreseen that precise and complete set of cross section measurements will be needed to identify the best production route for the radionuclides of interest.
Author contribution: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission.

Figure 2 :
Figure2:67 Cu production cross sections induced by deuteron and proton beams on70 Zn targets.
Research funding: The National Institute of Nuclear Physics (INFN) of Italy has funded these research activities in the mainframe of projects approved by Interdisciplinary National Scientific Commission five since the eighties.The LARAMED project is funded by the Italian Ministry of Education, Universities and Research and it is based at the INFN Legnaro National Laboratories.The cyclotron Arronax is supported by CNRS, Inserm, INCa, the Nantes University, the Regional Council of Pays de la Loire, local authorities, the French government and the European Union.This work has been, in part, supported by a grant from the French National Agency for Research called "Investissements d'Avenir", Equipex Arronax-Plus noANR-11-EQPX-0004, Labex IRON noANR-11-LABX-18-01 and ISITE NExT no ANR-16-IDEX-007.

Table  :
Main decay data of the radionuclides of interest [].

Table  :
Cu and  Cu production yields obtained by using proton, deuteron and α-beams on  Zn,  Zn, and  Ni enriched target materials assuming I =  µA and T irr =  h.

Table  :
Direct and indirect reactions to produce  Sc with the nuclear reactions induced by charged particles or neutrons.Pupillo et al.: Nuclear data for emerging medical radionuclides maximum value of about 12 mb at 34 MeV and for energies higher than 80 MeV.A recent work reported new experimental data for Ti(d,αn)  Sc  Ti(n,p)  Sc (fast neutrons)  Ti(p,pn)  Sc  Ti(d,αp)  Ca→  Sc  Ti(p,p)  Ca→  Sc V  V(p,x)  Ca→  Sc  V(d,x)  Ca→  Sc Ca  Ca(p,pn)  Ca→  Sc  Ca(d,pn)  Ca→  Sc  Ca(n,γ)  Ca→  Sc (thermal neutrons)  Ca(d,p)  Ca→  Sc G.
[87] with the value reported by Hall et al.[89]very close to 0 mb and the value measured by Chen et al.[87]around 24 mb; considering the agreement in the values reported by Chen et al. and Anders et al. [88], one may conclude that the 24 mb value is more reliable.-47 Ti(d,2p) 47 Sc excitation function was also measured in 1964 by Chen et al. [87], showing an increasing trend from 10 MeV up to 20 MeV, with a maximum value of ca.45 mb.There are no data regarding the 46 Sc coproduction, whose threshold energy is 5.2 MeV.-48 Ca(d,3n)

Table  :
Experimental activity  At/  At ratio [].
a ) are reported in [, ].G. Pupillo et al.: Nuclear data for emerging medical radionuclides