Nuclear Cross-Section of Proton-Induced Reactions on Enriched 48Ti Targets for the Production of Theranostic 47Sc Radionuclide, 46cSc, 44mSc, 44gSc, 43Sc, and 48V

The cross-sections of the 48Ti(p,x)47Sc, 46cSc, 44mSc, 44gSc, 43Sc, and 48V nuclear reactions were measured from 18 to 70 MeV, with particular attention to 47Sc production. Enriched 48Ti powder was deposited on an aluminum backing and the obtained targets were characterized via elastic backscattering spectroscopy at the INFN-LNL. Targets were exposed to low-intensity proton irradiation using the stacked-foils technique at the ARRONAX facility. Activated samples were measured using γ-spectrometry; the results were compared with the data int he literature and the theoretical TALYS-based values. A regular trend in the new values obtained from the different irradiation runs was noted, as well as a good agreement with the literature data, for all the radionuclides of interest: 47Sc, 46cSc, 44mSc, 44gSc, 43Sc, and 48V. 47Sc production was also discussed, considering yield and radionuclidic purity, for different 47Sc production scenarios.


Introduction
This work was carried out in the framework of the Production with Accelerator of Sc-47 for Theranostic Applications (PASTA) [1] and Research on Emerging Medical Radionuclides from the X-sections (REMIX) [2] projects, funded by INFN in 2017/2018 and 2021/2023, respectively.With the 70 MeV proton cyclotron installed at the INFN-LNL, the research activities carried out within Laboratory of Radionuclides for Medicine (LARAMED) are focused on the production of emerging radionuclides with proton beams [2].Among the radionuclides of major interest for our team is 47 Sc, a theranostic radionuclide that presents suitable decay characteristics for SPECT imaging and β − therapy (Table 1), which can be also paired with the β + emitter counterparts 43 Sc and 44 Sc for PET applications [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20].The lack of 47 Sc production is limiting its use in preclinical and clinical trials; for this reason, the 47 Sc proton-based production routes have been investigated within the PASTA and REMIX projects [1,2].First, nat V targets have been considered [21,22], then the nuclear reactions induced on isotopically enriched 48 Ti, 49 Ti, and 50 Ti targets (natural abundances of 73.72%, 5.41%, and 5.18%, respectively [23]) were studied.This paper presents the new The literature on proton-induced reactions with Ti-enriched targets is scarce; Gadioli et al. [30] and Levkovski [31] published data, respectively, in 1981 and 1991, using enriched 48 TiO 2 samples.Mausner et al. (1998) measured the relative cross-sections of 46c Sc, 44m Sc, 48 Sc normalized to 47 Sc, in the energy range 48-150 MeV using enriched 48 TiO 2 targets (99.81%); however, it is not possible to extract these absolute cross-section values [32].Some experimental data are also available for the 48 Ti(p,n) 48 V cross-section by rescaling the low energy (p,n) values obtained with nat Ti targets for the case of fully enriched material [24].
Enriched metallic 48 Ti powder was used in this work and deposited with the high energy vibrational powder plating (HIVIPP) technique [33,34], developed within the E_PLATE project (INFN in 2018/2019), on a substrate [35,36].A complete characterization of the 48 Ti-enriched targets was performed with the elastic backscattering (EBS) method using the proton beam available at the AN2000 accelerator at INFN-LNL.The EBS technique allowed the measurement of the amount of 48 Ti deposited (µg/cm 2 ) and its homogeneity.The nuclear cross-section measurements were performed at the ARRONAX facility [37], exploiting the available proton beam with tunable energy ranging from 35 to 70 MeV.

Discussion and Results
The nondestructive EBS technique was used to quantify the composition, the Ti deposited amount in µg/cm 2 , and the lateral homogeneity of the manufactured 48 Tienriched targets.The spectra acquired on the same target at three different points can overlap; the corresponding values of the Ti amount are thus similar, and the uniformity of the depositions along the diameter is therefore confirmed by the EBS analysis.The final value of the thickness in µg/cm 2 used for the nuclear cross-section calculations is the mean of the values measured at the three points.The mean value and the standard deviation for each sample are reported in Table 2.The targets prepared using the HIVIPP technique presented no modification after irradiation; the 48 Ti deposit remained adherent to the Al substrate.
The maximum value of the beam energy uncertainty, calculated with SRIM2003 code [38], was 875 keV.The major contribution to the cross-section uncertainty was always the monitor cross-section (max.5%) [39,40].The monitor reaction values used in the data analysis are reported in Table 3.The new experimental cross-section values, referring to a 100% enriched 48 Ti target, are reported in Table 4 and are plotted in Figures 1-5.A comparison with the literature data and the TALYS results (represented with a dashed line) is also given [29].TALYS simulations were performed using default parameters; additional information on the reaction, level density, and optical models used by the TALYS code can be found in [27,28,41].The results show a regular trend for all six radionuclides, 47 Sc, 46 Sc, 44g Sc, 44m Sc, 43 Sc, and 48 V.  48 Ti samples.
The data by Levkovski were corrected by a factor of 0.8 due to the monito used in 1991 [42]; for this reason, the data presented in the plots (Figures 1, 3, 4 have a star in the legend to indicate the applied rescaling factor.There is a gene agreement of our new results with the literature data, even though our expe values are about 20% lower than the data measured by Gadioli et al. (1981) in th range between 32 and 50 MeV [30].
A possible explanation could be the targets (e.g., enrichment level, composit target manufacturing) or on the monitor reactions and decay data used.The enr level of the target material was 99.1% for Gadioli et al., and that for Levkovsk from 95% to 98%.Gadioli et al. mixed 48 TiO2 powder with nat CuO (in a 2:1 ratio) to the beam intensity with 63 Cu(p,n) 63 Zn and 65 Cu(p,x) 64 Cu reactions, respectively, MeV and for the 25-85 MeV energy range.Nowadays, the IAEA recommends mo the nat Cu(p,x) 63 Zn reaction up to 100 MeV [39,40]; thus, it is possible to rescale th data for the low-energy region, up to the 63 Cu(p,n) 63 Zn channel (EP < 22 MeV) compare the actual monitor values with the ones reported by Gadioli et al. 40 ye Up to 20 MeV, there is a very good agreement on the monitor reaction values, s discrepancy is lower than 5%.However, the IAEA does not recommend the 65 Cu( or the nat Cu(p,x) 64 Cu reactions.It is worth noting that the literature data were o considering old values for the decay characteristics for the radionuclides of (discrepancies of up to 5%); however, it is not possible to correct these cross-sect considering the present values reported in Table 1 or to predict if the eventual co would lead to an increase or decrease of the published values in 1981 and 1991.  4Ti(p,2p) 47 Sc nuclear reaction [30,31].
have a star in the legend to indicate the applied rescaling factor.There is a gene agreement of our new results with the literature data, even though our expe values are about 20% lower than the data measured by Gadioli et al. (1981) in th range between 32 and 50 MeV [30].
A possible explanation could be the targets (e.g., enrichment level, composi target manufacturing) or on the monitor reactions and decay data used.The enr level of the target material was 99.1% for Gadioli et al., and that for Levkovsk from 95% to 98%.Gadioli et al. mixed 48 TiO2 powder with nat CuO (in a 2:1 ratio) to the beam intensity with 63 Cu(p,n) 63 Zn and 65 Cu(p,x) 64 Cu reactions, respectively, MeV and for the 25-85 MeV energy range.Nowadays, the IAEA recommends mo the nat Cu(p,x) 63 Zn reaction up to 100 MeV [39,40]; thus, it is possible to rescale t data for the low-energy region, up to the 63 Cu(p,n) 63 Zn channel (EP < 22 MeV compare the actual monitor values with the ones reported by Gadioli et al. 40 y Up to 20 MeV, there is a very good agreement on the monitor reaction values, discrepancy is lower than 5%.However, the IAEA does not recommend the 65 Cu or the nat Cu(p,x) 64 Cu reactions.It is worth noting that the literature data were considering old values for the decay characteristics for the radionuclides of (discrepancies of up to 5%); however, it is not possible to correct these cross-sec considering the present values reported in Table 1 or to predict if the eventual co would lead to an increase or decrease of the published values in 1981 and 1991.   Ti(p,x) 44g Sc cross-section: there is a perfect agreement results obtained by Levkovski up to 30 MeV [31]; at higher energies, the new about 30% lower than the values measured by Gadioli et al. [30].In general, values seem to be in agreement with the trend described by the TALYS estimation however present a slight energy shift toward lower energies, especially for EP > 5   48 Ti(p,x) 44g Sc cross-section: there is a perfect agreement results obtained by Levkovski up to 30 MeV [31]; at higher energies, the new about 30% lower than the values measured by Gadioli et al. [30].In general, values seem to be in agreement with the trend described by the TALYS estimation however present a slight energy shift toward lower energies, especially for EP > 5 The 48 Ti(p,x) 44m Sc cross-section is plotted in Figure 4: there is a general ag with the results obtained by Levkovski [31], while the TALYS results s underestimate the peak at around 30 MeV by a factor of two, but the general tren nuclear reaction is properly described.  4Ti(p,x) 43 Sc nuclear reaction [30].
Figure 5 reports the 48 Ti(p,x) 43 Sc excitation function: a good agreement results obtained by Gadioli et al. [30] can be noted.In this case, the TALYS result a visible energy shift and a general overestimation of the cross-section.  4Ti(p,x) 43 Sc nuclear reaction [30].
A possible explanation could be the targets (e.g., enrichment level, composition, and target manufacturing) or on the monitor reactions and decay data used.The enrichment level of the target material was 99.1% for Gadioli et al., and that for Levkovski ranged from 95% to 98%.Gadioli et al. mixed 48 TiO 2 powder with nat CuO (in a 2:1 ratio) to monitor the beam intensity with 63 Cu(p,n) 63 Zn and 65 Cu(p,x) 64 Cu reactions, respectively, up to 20 MeV and for the 25-85 MeV energy range.Nowadays, the IAEA recommends monitoring the nat Cu(p,x) 63 Zn reaction up to 100 MeV [39,40]; thus, it is possible to rescale the IAEA data for the low-energy region, up to the 63 Cu(p,n) 63 Zn channel (E P < 22 MeV), and to compare the actual monitor values with the ones reported by Gadioli et al. 40 years ago.Up to 20 MeV, there is a very good agreement on the monitor reaction values, since the discrepancy is lower than 5%.However, the IAEA does not recommend the 65 Cu(p,x) 64 Cu or the nat Cu(p,x) 64 Cu reactions.It is worth noting that the literature data were obtained considering old values for the decay characteristics for the radionuclides of interest (discrepancies of up to 5%); however, it is not possible to correct these cross-section data considering the present values reported in Table 1 or to predict if the eventual correction would lead to an increase or decrease of the published values in 1981 and 1991.
Figure 2 presents the 48 Ti(p,x) 46c Sc cross-section measured up to 70 MeV, which is compared with the data of Gadioli et al. [30] and TALYS results.There is a general very good agreement on the description of the 48 Ti(p,x) 46c Sc nuclear reaction in the entire energy range investigated.  4Ti(p,x) 43 Sc nuclear reaction [30].
Figure 5 reports the 48 Ti(p,x) 43 Sc excitation function: a good agreement w results obtained by Gadioli et al. [30] can be noted.In this case, the TALYS results a visible energy shift and a general overestimation of the cross-section.Figure 6 reports the 48 Ti(p,x) 48 V cross-section up to 90 MeV (A) and in the range investigated in this work, i.e., 20-70 MeV (B).There are several experimen available on the EXFOR database [24,25], and our new values are in very good ag both with the estimations in the literature and those of TALYS, which properly the reaction even if an overestimation of the peak value at ca. 13 MeV can be note

Discussion on 47 Sc Production
The new cross-section data provided in this paper can be compared with the based production route investigated with nat V targets [1,21,22], especially in the Figure 6.Cross-section of the 48 Ti(p,x) 48 V nuclear reaction [24,25,30,31]  Figure 3 reports the 48 Ti(p,x) 44g Sc cross-section: there is a perfect agreement with the results obtained by Levkovski up to 30 MeV [31]; at higher energies, the new data are about 30% lower than the values measured by Gadioli et al. [30].In general, our new values seem to be in agreement with the trend described by the TALYS estimations, which however present a slight energy shift toward lower energies, especially for E P > 50 MeV.
The 48 Ti(p,x) 44m Sc cross-section is plotted in Figure 4: there is a general agreement with the results obtained by Levkovski [31], while the TALYS results seem to underestimate the peak at around 30 MeV by a factor of two, but the general trend in this nuclear reaction is properly described.
Figure 5 reports the 48 Ti(p,x) 43 Sc excitation function: a good agreement with the results obtained by Gadioli et al. [30] can be noted.In this case, the TALYS results present a visible energy shift and a general overestimation of the cross-section.
Figure 6 reports the 48 Ti(p,x) 48 V cross-section up to 90 MeV (A) and in the energy range investigated in this work, i.e., 20-70 MeV (B).There are several experimental data available on the EXFOR database [24,25], and our new values are in very good agreement both with the estimations in the literature and those of TALYS, which properly describe the reaction even if an overestimation of the peak value at ca. 13 MeV can be noted.

Discussion on 47 Sc Production
The new cross-section data provided in this paper can be compared with the protonbased production route investigated with nat V targets [1,21,22], especially in the energy region below 45 MeV, where the 47 Sc cross-sections present a peak value, as shown in Figure 7.It can be noted that the 47 Sc excitation function for 48 Ti targets is larger than the one induced with nat V targets.As previously reported, in the energy range up to 30 MeV, the 47 Sc yield with nat V targets is calculated as 41.5 MBq/µA and 111 MBq/µA for 24 h and 80 h irradiation runs, respectively [1]; on the other hand, considering enriched 48 Ti targets, the 47 Sc yield for E P < 30 MeV is ca.200 MBq/µA and 530 MBq/µA for 24 h and 80 h, respectively.From these calculations, performed using the ISOTOPIA tool [43] made available by the IAEA, it can be inferred that 47 Sc production is about five times larger when using 48 Ti targets instead of nat V.
Pharmaceuticals 2024, 17, x FOR PEER REVIEW 7 of 15 region below 45 MeV, where the 47 Sc cross-sections present a peak value, as shown in Figure 7.It can be noted that the 47 Sc excitation function for 48 Ti targets is larger than the one induced with nat V targets.As previously reported, in the energy range up to 30 MeV, the 47 Sc yield with nat V targets is calculated as 41.5 MBq/µA and 111 MBq/µA for 24 h and 80 h irradiation runs, respectively [1]; on the other hand, considering enriched 48 Ti targets, the 47 Sc yield for EP < 30 MeV is ca.200 MBq/µA and 530 MBq/µA for 24 h and 80 h, respectively.From these calculations, performed using the ISOTOPIA tool [43] made available by the IAEA, it can be inferred that 47 Sc production is about five times larger when using 48 Ti targets instead of nat V. Particular attention has to be paid to the coproduction of Sc contaminants, i.e., 46 Sc, 44m Sc, 44g Sc, and 43 Sc, whose contribution in the final 47 Sc-labeled radiopharmaceutical has to be carefully assessed for each energy range considered.Table 5 reports the Sc radionuclide activities produced in different scenarios considering proton beams and enriched 48 Ti targets; in the calculation, all the experimental data available from the EXFOR database and the new ones measured in this work were considered to fit the nuclear cross-sections.Radionuclidic impurities can have undesirable effects on the patient's overall radiation dose, as well as on the image quality, so the European Pharmacopoeia established limits of radionuclides impurities for each radiopharmaceutical in the individual monographs to guarantee safe clinical application [44].In general, this limit is set to lower than 1%, but if very-high-purity products are technically achievable, it can be drastically reduced, up to 0.1% [45].
Table 5. Sc radionuclide yields calculated for several scenarios, considering 1 µA proton beam current and enriched 48 Ti targets.
Ep on 48   Particular attention has to be paid to the coproduction of Sc contaminants, i.e., 46 Sc, 44m Sc, 44g Sc, and 43 Sc, whose contribution in the final 47 Sc-labeled radiopharmaceutical has to be carefully assessed for each energy range considered.Table 5 reports the Sc radionuclide activities produced in different scenarios considering proton beams and enriched 48 Ti targets; in the calculation, all the experimental data available from the EXFOR database and the new ones measured in this work were considered to fit the nuclear cross-sections.Radionuclidic impurities can have undesirable effects on the patient's overall radiation dose, as well as on the image quality, so the European Pharmacopoeia established limits of radionuclides impurities for each radiopharmaceutical in the individual monographs to guarantee safe clinical application [44].In general, this limit is set to lower than 1%, but if very-high-purity products are technically achievable, it can be drastically reduced, up to 0.1% [45].
Table 5. Sc radionuclide yields calculated for several scenarios, considering 1 µA proton beam current and enriched 48 Ti targets.E p on 48  Figure 8 shows the radionuclidic purity (RNP) of 47 Sc considering 48 Ti targets and different scenarios for 24 h (A) and 80 h (B) irradiation, respectively.For all the cases, the RNP initially rapidly increases due to the decay of the short-half-time 43 Sc and 44g Sc impurities.After about 30 h, the rise in the RNP occurs more slowly due to the decay of 44m Sc.
impurities.After about 30 h, the rise in the RNP occurs more slowly due to the 44m Sc.
For irradiations performed at energies equal or larger than 35 MeV, the RNP a maximum (of the order of 50 and 55%, for E < 40 MeV and E < 35 MeV, respectiv then it decreases due to the contribution of the long-half-life 46 Sc impurity.Fo beam-energy irradiations (i.e., EP < 30 MeV), the RNP instead continuously increas in these scenarios, 46 Sc is not produced.The limit of 47 Sc RNP = 99% can be reache 1500 h after the EOB, corresponding to almost 20 times the half-time of 47 Sc.I concluded that the use of 48 Ti targets provides, at the EOB, a larger 47 Sc yield t targets but with a much lower RNP [22].For irradiations performed at energies equal or larger than 35 MeV, the RNP reaches a maximum (of the order of 50 and 55%, for E < 40 MeV and E < 35 MeV, respectively) and then it decreases due to the contribution of the long-half-life 46 Sc impurity.For lower-beamenergy irradiations (i.e., E P < 30 MeV), the RNP instead continuously increases since, in these scenarios, 46 Sc is not produced.The limit of 47 Sc RNP = 99% can be reached about Pharmaceuticals 2024, 17, 26 9 of 14 1500 h after the EOB, corresponding to almost 20 times the half-time of 47 Sc.It can be concluded that the use of 48 Ti targets provides, at the EOB, a larger 47 Sc yield than nat V targets but with a much lower RNP [22].

Enriched 48 Ti Targets
Thin deposits of enriched 48 Ti metallic powder (99.32%, purchased from Trace Sciences International Inc., Wilmington, DE, USA) onto a natural high-purity Al foil (99%, 25 µm thick, Goodfellow Cambridge Ltd., Huntingdon, UK) were obtained using the HIVIPP technique [35,36].Briefly, the deposition process was based on the application of an electrostatic field of 15 kV/cm between two Al substrates, used as electrodes, to start the superficial charging and the motion of the powder closed inside a quartz cylinder.The process took place in a vacuum of about 1•10 −7 mbar and lasted about 30 h. Figure 9A shows the experimental set up of HIVIPP deposition used in this study.The 48 Ti deposit had a diameter of 14 mm, which was cut then with punches (diameter of 12 mm) to fit the target holder used for the irradiation runs.A typical target foil is shown in Figure 9B.The peculiarities of this technique are the possibility of (i) realizing two substrates simultaneously, for which the target areal thickness of 0.2-2 mg/cm 2 was achieved, and (ii) recovering undeposited enriched 48 Ti powder, limiting the losses of this expensive material.More details about the technique and HIVIPP deposit characteristics are described in Refs.[35,36].The EBS analysis on 48 Ti targets was performed at the AN2000 Van the Graaff accelerator using a collimated 1800 keV proton beam with an approximate size of 1 mm 2 .The backscattering angle (θ out ) and the incidence angle with respect to sample normal (θ o ) were θ out = 160 • and θ o = 0 • .The measurements were made using a standard charged particle spectroscopy system consisting of a Si detector and NIM electronics.

2024, 17, x FOR PEER REVIEW 9 of 15
Thin deposits of enriched 48 Ti metallic powder (99.32%, purchased from Trace Sciences International Inc., Wilmington, DE, USA) onto a natural high-purity Al foil (99%, 25 µm thick, Goodfellow Cambridge Ltd., Huntingdon, UK) were obtained using the HIVIPP technique [35,36].Briefly, the deposition process was based on the application of an electrostatic field of 15 kV/cm between two Al substrates, used as electrodes, to start the superficial charging and the motion of the powder closed inside a quartz cylinder.The process took place in a vacuum of about 1•10 −7 mbar and lasted about 30 h. Figure 9A shows the experimental set up of HIVIPP deposition used in this study.The 48 Ti deposit had a diameter of 14 mm, which was cut then with punches (diameter of 12 mm) to fit the target holder used for the irradiation runs.A typical target foil is shown in Figure 9B.The peculiarities of this technique are the possibility of (i) realizing two substrates simultaneously, for which the target areal thickness of 0.2-2 mg/cm 2 was achieved, and (ii) recovering undeposited enriched 48 Ti powder, limiting the losses of this expensive material.More details about the technique and HIVIPP deposit characteristics are described in Refs.[35,36].The EBS analysis on 48 Ti targets was performed at the AN2000 Van the Graaff accelerator using a collimated 1800 keV proton beam with an approximate size of 1 mm 2 .The backscattering angle (θ out ) and the incidence angle with respect to sample normal (θ o ) were θ out = 160° and θ o = 0°.The measurements were made using a standard charged particle spectroscopy system consisting of a Si detector and NIM electronics.  4Ti-7 sample, with a typical 48 Ti deposition onto Al, as realized and after punching (B).
The total energy resolution of the spectrometer was 13 keV.To ascertain coating uniformity, EBS measurements were performed for at least three positions on each sample along the sample's diameter.The aluminum backings were also characterized using proton induced X-ray emission (PIXE) analysis to determine the presence of impurities.It turned out that 0.4 at% (±0.1)Fe was present in the Al substrates.The EBS experimental spectra were simulated using SimNRA 7.03 software [46].The individual elements' stopping powers were deduced from SRIM2003 code [38], and Bragg's rule was used for the compounds.The non-Rutherford oxygen and aluminum backscattering cross-sections were deduced from the IAEA Ion Beam Analysis Nuclear Data Library database [47].All other heavier relevant elements were assumed to have Rutherford cross-sections.The simulations took into account the significant coating roughness, which determined a long tail of the 48 Ti signal toward the low-energy region of the spectra [48], as shown in Figure 10.  4Ti-7 sample, with a typical 48 Ti deposition onto Al, as realized and after punching (B).
The total energy resolution of the spectrometer was 13 keV.To ascertain coating uniformity, EBS measurements were performed for at least three positions on each sample along the sample's diameter.The aluminum backings were also characterized using proton induced X-ray emission (PIXE) analysis to determine the presence of impurities.It turned out that 0.4 at% (±0.1)Fe was present in the Al substrates.The EBS experimental spectra were simulated using SimNRA 7.03 software [46].The individual elements' stopping powers were deduced from SRIM2003 code [38], and Bragg's rule was used for the compounds.The non-Rutherford oxygen and aluminum backscattering cross-sections were deduced from the IAEA Ion Beam Analysis Nuclear Data Library database [47].All other heavier relevant elements were assumed to have Rutherford cross-sections.The simulations took into account the significant coating roughness, which determined a long tail of the 48 Ti signal toward the low-energy region of the spectra [48], as shown in Figure 10.Simulation parameters were chosen to allow spectral fitting of the elements characterized by elastic scattering cross-section with a fine structure [49].In all the analyzed samples titanium resulted oxidized.The determination of the 48 Ti content was estimated by considering the Ti EBS simulated spectrum made of two contributions: the high energy part (characterized by low measurement error) and the Ti spectrum region tailing into the lighter elements, to which a higher uncertainty must be attributed due to the errors of the stopping powers and of the non-Rutherford cross sections.The results of the simulations are reported in Table 2.

Irradiation Runs, γ-Spectrometry, and Data Analysis
Fifteen 48 Ti targets, assembled with the well-known stacked-foils technique, were irradiated in eight irradiation runs at the ARRONAX facility [37] to cover the energy range of 18-70 MeV.The irradiation runs had a duration of 50-90 min, with a current of about 100-130 nA, monitored during the bombardment using an instrumented beam dump.The beam line was under vacuum, closed with a 75 µm thick Kapton foil; the stacks were located about 10-15 cm downstream in air; this distance was precisely measured for each irradiation run.
Close to each 48 Ti target, a nat Ni monitor foil was inserted (10 or 25 µm thick) in order to measure the effective beam flux by considering the reference reaction recommended by the International Atomic Energy Agency (IAEA) [39,40] for 57 Ni production (half-life 35.60 h, Eγ = 1377.63keV, Iγ = 81.7%).To catch the eventual 57 Ni recoil atoms, a thin aluminum foil (10 µm thick) was inserted after the nat Ni monitor foil.To decrease the beam energy, some thicker Al foils (500 µm thick) were used in the stacked structure, as shown in Figure 11.All the materials used in the stacks were high-purity foils (≥99%, Goodfellow Cambridge Ltd., Huntingdon, UK).
The proton beam energy in each layer of the stacked target was calculated using SRIM2013 code [50], considering the extracted proton beam energy from the cyclotron, and the energy losses in the Kapton foil, in the air and in each foil of the stacked targets.The uncertainty on the proton beam energy was obtained by considering the uncertainty of the energy extracted from the cyclotron (±500 keV) and calculating the energy straggling through each layer of the stacked target using SRIM.
Given that the 48 Ti powder was deposited on an Al substrate, the stacked-foil structure was always assembled in order to have the proton beam impinging on the 48 Ti powder first; in this way, recoil atoms were trapped by the Al support.Simulation parameters were chosen to allow spectral fitting of the elements characterized by elastic scattering cross-section with a fine structure [49].In all the analyzed samples titanium resulted oxidized.The determination of the 48 Ti content was estimated by considering the Ti EBS simulated spectrum made of two contributions: the high energy part (characterized by low measurement error) and the Ti spectrum region tailing into the lighter elements, to which a higher uncertainty must be attributed due to the errors of the stopping powers and of the non-Rutherford cross sections.The results of the simulations are reported in Table 2.

Irradiation Runs, γ-Spectrometry, and Data Analysis
Fifteen 48 Ti targets, assembled with the well-known stacked-foils technique, were irradiated in eight irradiation runs at the ARRONAX facility [37] to cover the energy range of 18-70 MeV.The irradiation runs had a duration of 50-90 min, with a current of about 100-130 nA, monitored during the bombardment using an instrumented beam dump.The beam line was under vacuum, closed with a 75 µm thick Kapton foil; the stacks were located about 10-15 cm downstream in air; this distance was precisely measured for each irradiation run.
Close to each 48 Ti target, a nat Ni monitor foil was inserted (10 or 25 µm thick) in order to measure the effective beam flux by considering the reference reaction recommended by the International Atomic Energy Agency (IAEA) [39,40] for 57 Ni production (half-life 35.60 h, E γ = 1377.63keV, I γ = 81.7%).To catch the eventual 57 Ni recoil atoms, a thin aluminum foil (10 µm thick) was inserted after the nat Ni monitor foil.To decrease the beam energy, some thicker Al foils (500 µm thick) were used in the stacked structure, as shown in Figure 11.All the materials used in the stacks were high-purity foils (≥99%, Goodfellow Cambridge Ltd., Huntingdon, UK).As soon as possible after the end of bombardment (EOB), a first γ-spectrometry measurement of the irradiated 48 Ti sample was carried out to estimate the activity of shortliving radionuclides: this acquisition was typically 15 min long, and it was performed about 2-4 h after the EOB.Each 48 Ti sample was also measured overnight to check for The proton beam energy in each layer of the stacked target was calculated using SRIM2013 code [50], considering the extracted proton beam energy from the cyclotron, and the energy losses in the Kapton foil, in the air and in each foil of the stacked targets.The uncertainty on the proton beam energy was obtained by considering the uncertainty of the energy extracted from the cyclotron (±500 keV) and calculating the energy straggling through each layer of the stacked target using SRIM.
Given that the 48 Ti powder was deposited on an Al substrate, the stacked-foil structure was always assembled in order to have the proton beam impinging on the 48 Ti powder first; in this way, recoil atoms were trapped by the Al support.
As soon as possible after the end of bombardment (EOB), a first γ-spectrometry measurement of the irradiated 48 Ti sample was carried out to estimate the activity of shortliving radionuclides: this acquisition was typically 15 min long, and it was performed about 2-4 h after the EOB.Each 48 Ti sample was also measured overnight to check for lower-activity products with a longer acquisition time (about 8-14 h).To follow the decay of the radionuclides of interest and to check for eventual γ-interferences, the γ-spectrometry measurements were repeated for all 48 Ti targets each day up to 5 days after the EOB (these acquisitions were typically 1.5-3 h long).All samples were measured with the same highpurity germanium (HPGe) detector (10% relative efficiency, FWHM 1.0 keV at 122 keV, Canberra GC1020), previously calibrated with a 152 Eu and an 241 Am point-like solid sources (purchased to Cerca-Lea, Tricastin, France).All 48 Ti samples were measured with the 48 Ti deposit in the direction of the HPGe detector in order to avoid the γ attenuation due to the Al support.The sample-detector distance was fixed at 19 cm to reduce the dead time during measurements, which was always kept below 10%.The γ spectra were analyzed using software jRadView, developed at the INFN-LNL for nuclear physics experiments.A typical γ spectrum obtained for a 48 Ti target is shown in Figure 12.As soon as possible after the end of bombardment (EOB), a first γ-spectrometry measurement of the irradiated 48 Ti sample was carried out to estimate the activity of shortliving radionuclides: this acquisition was typically 15 min long, and it was performed about 2-4 h after the EOB.Each 48 Ti sample was also measured overnight to check for lower-activity products with a longer acquisition time (about 8-14 h).To follow the decay of the radionuclides of interest and to check for eventual γ-interferences, the γspectrometry measurements were repeated for all 48 Ti targets each day up to 5 days after the EOB (these acquisitions were typically 1.5-3 h long).All samples were measured with the same high-purity germanium (HPGe) detector (10% relative efficiency, FWHM 1.0 keV at 122 keV, Canberra GC1020), previously calibrated with a 152 Eu and an 241 Am point-like solid sources (purchased to Cerca-Lea, Tricastin, France).All 48 Ti samples were measured with the 48 Ti deposit in the direction of the HPGe detector in order to avoid the γ attenuation due to the Al support.The sample-detector distance was fixed at 19 cm to reduce the dead time during measurements, which was always kept below 10%.The γ spectra were analyzed using software jRadView, developed at the INFN-LNL for nuclear physics experiments.A typical γ spectrum obtained for a 48 Ti target is shown in Figure 12.The nuclear data extracted from the NuDat 3.0 database (Table 1) were used in the data analysis, which was carried out following the methods of Otuka et al. [51], which were also used for the uncertainty calculations.In the calculation of the 46c Sc cumulative cross-section, only the γ line at 889 keV was used, since the 1120 keV line had an interference with the background 214 Bi emission from the natural 238 U decay chain.The recoil effect for the monitor 57 Ni activity was taken into account, and it was about 1%.The results of the 48 Ti(p,x) 47 Sc, 46c Sc, 44m Sc, 44g Sc, 43 Sc, and 48 V cross-sections are given for a 100% The nuclear data extracted from the NuDat 3.0 database (Table 1) were used in the data analysis, which was carried out following the methods of Otuka et al. [51], which were also used for the uncertainty calculations.In the calculation of the 46c Sc cumulative crosssection, only the γ line at 889 keV was used, since the 1120 keV line had an interference with the background 214 Bi emission from the natural 238 U decay chain.The recoil effect for the monitor 57 Ni activity was taken into account, and it was about 1%.The results of the 48 Ti(p,x) 47 Sc, 46c Sc, 44m Sc, 44g Sc, 43 Sc, and 48 V cross-sections are given for a 100% enriched target.New data were compared with the few experimental values available and with the results obtained from the TALYS code run with the default parameters (version 1.96 released in December 2021) [29].

Conclusions
The new cross-section data provided in this paper for the 48 Ti(p,x) 47 Sc, 46c Sc, 44m Sc, 44g Sc, 43 Sc, and 48 V reactions are generally in a good agreement with the literature data.The TALYS results give a satisfactory description of the trend in the nuclear reactions (e.g., 46c Sc and 44g Sc), even if there is a considerable energy shift and an overestimation of the experimental values in the case of 43 Sc.Theoretical studies to find the best TALYS parameters to properly describe the nuclear reactions are ongoing in the framework of the REMIX collaboration [21,52].
The calculations showed that the 48 Ti(p,x) 47 Sc route provides a larger 47 Sc yield compared to the use of nat V targets, but with an RNP not suitable for medical applications.To provide a comprehensive overview of the proton-induced routes for 47 Sc production, nuclear cross-section measurements using enriched 49 Ti and 50 Ti targets are ongoing within the REMIX project.These new data will be compared with those based on the use of nat V and 48 Ti targets to select the most suitable irradiation parameters for a reliable 47 Sc supply.

Figure 3
Figure3reports the48 Ti(p,x) 44g Sc cross-section: there is a perfect agreement results obtained by Levkovski up to 30 MeV[31]; at higher energies, the new about 30% lower than the values measured by Gadioli et al.[30].In general, values seem to be in agreement with the trend described by the TALYS estimation however present a slight energy shift toward lower energies, especially for EP > 5

Figure 3
Figure3reports the48 Ti(p,x) 44g Sc cross-section: there is a perfect agreement results obtained by Levkovski up to 30 MeV[31]; at higher energies, the new about 30% lower than the values measured by Gadioli et al.[30].In general, values seem to be in agreement with the trend described by the TALYS estimation however present a slight energy shift toward lower energies, especially for EP > 5

Figure 7 .
Figure 7.Comparison of the p-induced cross-sections for 47 Sc production on48 Ti and nat V targets.The dashed lines represent the fit of all the experimental data for both production routes.

Figure 7 .
Figure 7.Comparison of the p-induced cross-sections for 47 Sc production on48 Ti and nat V targets.The dashed lines represent the fit of all the experimental data for both production routes.

Figure 8 .
Figure 8.Comparison of the 47 Sc RNP for 24 h (A) and 80 h (B) irradiation, using proton b 48 Ti targets.

Figure 8 .
Figure 8.Comparison of the 47 Sc RNP for 24 h (A) and 80 h (B) irradiation, using proton beams and 48 Ti targets.

Figure 9 .
Figure 9. Photograph of the HIVIPP set up inside the vacuum chamber (A).Picture of the48 Ti-7 sample, with a typical48 Ti deposition onto Al, as realized and after punching (B).

Figure 9 .
Figure 9. Photograph of the HIVIPP set up inside the vacuum chamber (A).Picture of the48 Ti-7 sample, with a typical48 Ti deposition onto Al, as realized and after punching (B).

Pharmaceuticals 2024 , 15 Figure 10 .
Figure 10.Experimental spectrum of 48 Ti coating deposited onto Al substrate.Arrows indicate the 16 O and 48 Ti contributions.

Figure 10 .
Figure 10.Experimental spectrum of 48 Ti coating deposited onto Al substrate.Arrows indicate the 16 O and 48 Ti contributions.

Pharmaceuticals 2024 , 15 Figure 11 .
Figure 11.Scheme of a typical SRIM calculation for the stacked-foil target.

Figure 11 .
Figure 11.Scheme of a typical SRIM calculation for the stacked-foil target.

Pharmaceuticals 2024 , 15 Figure 11 .
Figure 11.Scheme of a typical SRIM calculation for the stacked-foil target.

Figure 12 .
Figure12.Typical γ spectrum obtained with an HPGe detector for an irradiated48 Ti target.In blue, the 47 Sc γ-peak is highlighted.

Figure 12 .
Figure12.Typical γ spectrum obtained with an HPGe detector for an irradiated48 Ti target.In blue, the 47 Sc γ-peak is highlighted.

Table 1 .
Nuclear data associated with the radionuclides studied in this work[23]; the uncertainty is reported in brackets.

Table 2 .
Results of the EBS analysis on the enriched