Innovative Target for Production of Technetium-99m by Biomedical Cyclotron

Technetium-99m (99mTc) is the most used radionuclide worldwide in nuclear medicine for diagnostic imaging procedures. 99mTc is typically extracted from portable generators containing 99Mo, which is produced normally in nuclear reactors as a fission product of highly enriched Uranium material. Due to unexpected outages or planned and unplanned reactor shutdown, significant 99mTc shortages appeared as a problem since 2008 The alternative cyclotron-based approach through the 100Mo(p,2n)99mTc reaction is considered one of the most promising routes for direct 99mTc production in order to mitigate potential 99Mo shortages. The design and manufacturing of appropriate cyclotron targets for the production of significant amounts of a radiopharmaceutical for medical use is a technological challenge. In this work, a novel solid target preparation method was developed, including sputter deposition of a dense, adherent, and non-oxidized Mo target material onto a complex backing plate. The latter included either chemically resistant sapphire or synthetic diamond brazed in vacuum conditions to copper. The target thermo-mechanical stability tests were performed under 15.6 MeV proton energy and different beam intensities, up to the maximum provided by the available GE Healthcare (Chicago, IL, USA) PET trace medical cyclotron. The targets resisted proton beam currents up to 60 µA (corresponding to a heat power density of about 1 kW/cm2) without damage or Mo deposited layer delamination. The chemical stability of the proposed backing materials was proven by gamma-spectroscopy analysis of the solution obtained after the standard dissolution procedure of irradiated targets in H2O2.


Introduction
Technetium-99m ( 99m Tc) is an extremely important radionuclide, used in the vast majority of traditional diagnostic SPECT (Single Photon Emission Computer Tomography) imaging examinations. The radioisotope is usually available in hospitals from portable generators containing the 99 Mo parent nuclide, a fission product of highly 235 U-enriched uranium. About 95% of the 99 Mo world production is provided by a few ageing nuclear reactors, whose unplanned outages have already caused shortages at the global level, compromising the availability of 99m Tc. In order to compensate potential 99 Mo shortages in the future, alternative accelerator-based routes for the production of 99m Tc have been intensively developed.

State-of-the-Art: Cyclotron Target for 99m Tc Production
Although research and development (R&D) activity on 99m Tc cyclotron production using oxide [4,5,[11][12][13] carbide [4,5] and even aqueous solution [4,5] targets has been extensive, metallic Mo solid targets are considered the most promising, as they are able to provide the highest production yield. That is due to the high level of heat dissipation associated with the higher thermal conductivity of the metal. Even traces of oxygen can drastically reduce the thermal conductivity of the target material.
Two approaches-preliminary Mo pellet preparation followed by subsequent bonding to a backing plate and direct deposition of Mo onto a backing-have been used. 100 Mo-enriched metallic molybdenum is usually available on the market in powder form. Thus, the transformation of refractory metal powders into dense pellet or foil has been studied by different research groups applying a range of different methods, including pressing, sintering, and rolling. The targets produced using only hydraulic pressing of powder [4,5,[14][15][16] are characterized by much lower density than their corresponding bulk material and can sustain relatively limited cyclotron currents. The sintering of Mo powders [4,9,17] followed by press bonding [4,5,[14][15][16] or vacuum brazing [4,9,17] processes were used to manufacture Mo targets, sustaining beam currents in the order of hundreds of µA. Targets for nuclear cross-section measurements were produced by press-rolling, starting from powders or from Mo beads produced by e-beam powder melting [4,5,18], otherwise by re-melting of Mo powders, followed by press-roller reshaping. However, this method is not applicable to routine production in which a thick target is required, as the method is unable to provide a good thermal contact among Mo stacked foils or between them and the backing during irradiation. Instead, using direct Mo deposition onto a backing plate, the second step (i.e., pellet bonding to a backing plate) is unnecessary. The easiest approach (i.e., Mo powders melting directly onto a backing plate [19]), has been also investigated, but the results have shown non-uniformity of the Mo layer.
Electroplating of metals from aqueous solutions is a well-known industrial process. However, refractory metals such as Mo are difficult to deposit using standard electrodeposition methods due to their high affinity for oxygen. Despite this, electroplating from particular alkaline [20] or acetate [21] solutions has been described. Notably, such deposits have thicknesses not exceeding 20 µm and high levels of oxidation. In addition, the described process was found to be very inefficient (<2%). The co-deposition of molybdenum with zinc [22] was reported to be a much more efficient approach. The electrodeposition from ionic liquids or molten salts [4,23,24] has produced better Mo layer quality, although the use of expensive equipment and implementation of a more complicated protocol were required. Electrophoretic deposition from a mixture of Mo powders and molybdate with additives, followed by subsequent sintering at high temperature, has produced sufficient Mo thickness and resistance up to 300 µA [5,25]. Further improvement of the process by sintering in an inert atmosphere [4] has provided targets resistant up to 500 µA. Notwithstanding the much better results in beam power tests, the accumulation of impurities from the electrolyte bath is, however, an important drawback of all electrochemical deposition methods that is hard to overcome in target manufacturing.
Considering the reference 100 Mo(p, 2n) 99m Tc production route, the standard target processing procedure includes dissolution of a Mo and Tc mixture in hot concentrated H 2 O 2 . Transition, post-transition, and refractory metals are not perfectly inert at such conditions. In the case of radiopharmaceutical production, even the presence of a very low amount of impurities is critical.
The idea to use an inert baseplate was presented by the group at the University of Alberta during the International Atomic Energy Agency (IAEA) meeting held in 2013 [5] and 2015 [4], devoted to the accelerator-based production of 99m Tc, including materials such as glassy carbon, quartz, and aluminum oxide. At the same meetings, a Japanese group proposed the use of an inert vessel (aluminum oxide [5], SiC [4]) for direct irradiation of Mo powders by a vertical cyclotron beam. Table 1 presents a comparative summary of the studies carried out in different countries on targetry and target preparation for 99m Tc production by a cyclotron, highlighting the main drawbacks of each approach.
Even if progress is being accomplished on Mo solid target design and optimization all over the world, the problem of the target sustaining elevated cyclotron currents and providing few impurities is still a challenge. In this work, we propose solving this problem using Mo direct deposition by magnetron sputtering onto a complex backing plate composed of chemically inert sapphire or synthetic diamond (only this part is exposed during dissolution to H 2 O 2 dissolution media), brazed to a high thermal conductivity metallic holder.
Copper disks 32 mm in diameter and 1.5 mm thick were used as simple target backing and substrates for direct Mo sputtering. Copper components, which were disks 32 mm in diameter and 1.5 mm thick with a cylindrical recess 0.5 mm deep and diameter corresponding to the size of sapphire/synthetic diamond plus brazing radial clearance, were used for the preparation of complex target prototypes.
The standard procedure for copper substrate preparation included: ultrasonic washing for 20 min with GP 17.40 SUP soap (NGL Cleaning Technology SA, Nyon, Switzerland) and deionized water, chemical etching with SUBU5 solution (5 g/L sulfamic acid, 1 g/L ammonium citrate, 50 mL/L butanol, 50 mL/L H 2 O 2, and 1 L deionized water) at 72 ± 4 • C in order to remove surface oxides, passivation in 20 g/L sulfamic acid, ultrasonic washing with water for 20 min, rinsing with ethanol, and drying with nitrogen.
IR-grade sapphire with C-axis orientation (Meller Optics Inc., Providence, RI, USA) discs 12.7 mm in diameter and 0.5 mm thick, and chemical vapor-deposited (CVD) synthetic diamond substrates (II-VI Advanced materials GmbH, Pine Brook, NJ, USA) 13.5 mm in diameter and 0.3 mm thick of thermal grade, with thermal conductivity 1500 W/(m·K) were used for target backing preparation For pre-treatment of all non-metallic substrates, including sapphire, synthetic diamond, and the silicon wafer, the same procedure was used: ultrasonic washing for 20 min with Rodaclean ®− (NGL Cleaning Technology SA, Nyon, Switzerland) soap and then deionized water for 20 min, rinsing with ethanol, and drying with nitrogen flux.
Copper disks 13 mm in diameter and 1 mm thick were used as substrates for Mo sputtering for the Scanning Electron Microscopy (SEM) analysis of Mo film. For the SEM cross-section analysis, the samples were cut by electro-erosion, treated with abrasive paper, and then chemically etched in a "base piranha" solution (H 2 O + NaOH + H 2 O 2 1:1:1 vol.) for 1 min at 50 • C.

Cyclotron
In this study, a GE PETtrace 800S cyclotron (GE Healthcare, Chicago, IL, USA), installed at S. Orsola-Malpighi Hospital in Bologna, was used for all irradiation tests. The PETtrace ( Figure 1a) is an isochronous cyclotron working at fixed energy (16.5 MeV for protons and 8.4 MeV for deuterons) and a maximum beam intensity of 100 µA. Note, the maximum current available depends on the source and tuning of the magnets.
The solid target station (prototype TEMA Sinergie S.P.A., Faenza, Ra, Italy) used in this work is shown in Figure 1b. The target "coin" is manually placed inside the target station before irradiation. The aluminum water cooling chamber is pressed to the back of the target coin using a~3 bar air pneumatic piston. After irradiation, the system allows an automatic target transfer outside the cyclotron bunker. Double cooling is used to increase the current allowed on the target: water-cooling from the back, and helium gas cooling from the front side of the target coin. A detailed description of this irradiation unit was reported previously [32]. The target prototypes were developed fitting the design of this target station. The typical target coin is a disk at most 32 mm in diameter and 2 mm thick. In the current solid target system, the integer target coin provides maximum heat exchange performance that is limited by the thermal resistance of both the deposited Mo layer and the backing plate, and the contribution due to the thermal contact resistance between them.
For this work, a specific 4-pin target holder was designed ( Figure 2) to house the sapphire and synthetic diamond disks for testing chemical inertness. The external dimensions of the target holder were 32 mm diameter and 2 mm thick once assembled. The target prototypes for thermomechanical stability control were irradiated for one minute (enough to reach thermal equilibrium) with increasing current. After each irradiation, the sample was unloaded and the integrity of the target and the adhesion of the Mo film on the backing were visually controlled. One of two CVD synthetic diamond-based target prototypes was irradiated for 30 min at the maximum current produced by the cyclotron in order to confirm the previously demonstrated performance during short irradiations.

Unique Vacuum System for Target Preparation
For target prototype preparation, a vacuum system consisting of four vacuum chambers connected through the central zone and separated by pneumatic gates was used in order to continue with both vacuum processes of interest in different chambers: sputter deposition and vacuum brazing ( Figure 3). The central zone was connected to a general pumping system including a 360 L/min Pfeiffer turbo molecular pump with a Pfeiffer DCU display and operating unit (Pfeiffer Vacuum, Asslar, Germany), and a 210 L/min Varian (now Agilent, Santa Clara, CA, USA) Tri Scroll Pump as a primary pump. The base vacuum pressure of about 1 × 10 −6 mbar was reached without additional backing before each experiment. The target prototypes were developed fitting the design of this target station. The typical target coin is a disk at most 32 mm in diameter and 2 mm thick. In the current solid target system, the integer target coin provides maximum heat exchange performance that is limited by the thermal resistance of both the deposited Mo layer and the backing plate, and the contribution due to the thermal contact resistance between them.
For this work, a specific 4-pin target holder was designed ( Figure 2) to house the sapphire and synthetic diamond disks for testing chemical inertness. The external dimensions of the target holder were 32 mm diameter and 2 mm thick once assembled. The target prototypes were developed fitting the design of this target station. The typical target coin is a disk at most 32 mm in diameter and 2 mm thick. In the current solid target system, the integer target coin provides maximum heat exchange performance that is limited by the thermal resistance of both the deposited Mo layer and the backing plate, and the contribution due to the thermal contact resistance between them.
For this work, a specific 4-pin target holder was designed ( Figure 2) to house the sapphire and synthetic diamond disks for testing chemical inertness. The external dimensions of the target holder were 32 mm diameter and 2 mm thick once assembled. The target prototypes for thermomechanical stability control were irradiated for one minute (enough to reach thermal equilibrium) with increasing current. After each irradiation, the sample was unloaded and the integrity of the target and the adhesion of the Mo film on the backing were visually controlled. One of two CVD synthetic diamond-based target prototypes was irradiated for 30 min at the maximum current produced by the cyclotron in order to confirm the previously demonstrated performance during short irradiations.

Unique Vacuum System for Target Preparation
For target prototype preparation, a vacuum system consisting of four vacuum chambers connected through the central zone and separated by pneumatic gates was used in order to continue with both vacuum processes of interest in different chambers: sputter deposition and vacuum brazing ( Figure 3). The central zone was connected to a general pumping system including a 360 L/min Pfeiffer turbo molecular pump with a Pfeiffer DCU display and operating unit (Pfeiffer Vacuum, Asslar, Germany), and a 210 L/min Varian (now Agilent, Santa Clara, CA, USA) Tri Scroll Pump as a primary pump. The base vacuum pressure of about 1 × 10 −6 mbar was reached without additional backing before each experiment. The target prototypes for thermomechanical stability control were irradiated for one minute (enough to reach thermal equilibrium) with increasing current. After each irradiation, the sample was unloaded and the integrity of the target and the adhesion of the Mo film on the backing were visually controlled. One of two CVD synthetic diamond-based target prototypes was irradiated for 30 min at the maximum current produced by the cyclotron in order to confirm the previously demonstrated performance during short irradiations.

Unique Vacuum System for Target Preparation
For target prototype preparation, a vacuum system consisting of four vacuum chambers connected through the central zone and separated by pneumatic gates was used in order to continue with both vacuum processes of interest in different chambers: sputter deposition and vacuum brazing ( Figure 3). The central zone was connected to a general pumping system including a 360 L/min Pfeiffer turbo molecular pump with a Pfeiffer DCU display and operating unit (Pfeiffer Vacuum, Asslar, Germany), and a 210 L/min Varian (now Agilent, Santa Clara, CA, USA) Tri Scroll Pump as a primary pump. The base vacuum pressure of about 1 × 10 −6 mbar was reached without additional backing before each experiment.
The base vacuum pressure was controlled by a full-range Bayard-Alpert (BA) vacuum-meter. The pressure during MS deposition was controlled with a capacitance vacuum-meter since it was not sensitive to plasma as BA. The vacuum-meters were connected to a MaxiGauge TM control box (Pfeiffer Vacuum, Asslar, Germany). The MKS (MKS Inc., Andover, MA, USA) multi-gas mass-flow controllers powered by the MKS 647C four-channel power supply/readout system was used for gas flow control during the deposition process. The entire system is shown in Figure 3b.

Magnetron Sputtering
Sputtering deposition occurs in a vacuum by means of inert gas plasma (Ar). The material to be deposited, called the sputtering target (not to be confused with the cyclotron target), is attached to the cathode. Plasma is created when a difference in potential is applied between the cathode and the substrate (anode). The positive ions of the inert gas are accelerated toward the cathode. When the ions collide with the atoms of the sputtering target, the energy transfer causes the detachment of some atoms, which are then deposited on the substrate. Magnetron sputtering is characterized by elevated plasma use efficiency due to its magnetic confinement. The process is schematically described in Figure 4. The entire system was controlled by a home-made human-machine interface LabVIEW (National Instruments Italy, Roma, Italy)-programmed PLC (Programmable Logic Controller). Figure 3a shows the four-chamber system layout used for PLC control. Chambers 2 and 4 were used for sputter deposition with 2 magnetron sources. Chamber 3 was used for vacuum brazing.
The base vacuum pressure was controlled by a full-range Bayard-Alpert (BA) vacuum-meter. The pressure during MS deposition was controlled with a capacitance vacuum-meter since it was not sensitive to plasma as BA. The vacuum-meters were connected to a MaxiGauge TM control box (Pfeiffer Vacuum, Asslar, Germany). The MKS (MKS Inc., Andover, MA, USA) multi-gas mass-flow controllers powered by the MKS 647C four-channel power supply/readout system was used for gas flow control during the deposition process. The entire system is shown in Figure 3b.

Magnetron Sputtering
Sputtering deposition occurs in a vacuum by means of inert gas plasma (Ar). The material to be deposited, called the sputtering target (not to be confused with the cyclotron target), is attached to the cathode. Plasma is created when a difference in potential is applied between the cathode and the substrate (anode). The positive ions of the inert gas are accelerated toward the cathode. When the ions collide with the atoms of the sputtering target, the energy transfer causes the detachment of some atoms, which are then deposited on the substrate. Magnetron sputtering is characterized by elevated plasma use efficiency due to its magnetic confinement. The process is schematically described in Figure 4. In the current work, the films were deposited by direct current (DC) sputtering with a 2" planar, unbalanced, II Type magnetron cathode source. The magnetron drive model MDX 1.5 kW (Advanced Energy, Fort Collins, CO, USA) was used to complete the sputtering experiments with the possibility of an automatic control in order to set up the pulsing mode to perform multilayer deposition.
Depositions were performed onto a planar substrate holder (SH), which had the ability to heat up the substrate by controlling the temperature with a K-type thermocouple inserted in the SH plate. The distance between the substrate and the cathode was 6 cm. For film deposition, we used the downtop deposition configuration with a magnetron source placed from the bottom of the cylindrical vacuum chamber and substrate holder with substrates from the top of the chamber in order to minimize film delamination caused by metallic dust particles.
Molybdenum was deposited on a spot 10 mm in diameter in the center of each substrate (backing plate) defined by an appropriate mask.

Thin Film Analysis
FEI (Philips, Amsterdam, The Netherlands) SEM XL-30 was used for sputtered coatings analysis.
The contact stylus profiler, model Dektak 8 (Veeco, Plainview, NY, USA), able to measure a surface texture below submicro-inch and film thickness up to 262 µm, was used for sputtered film thickness characterization.

Vacuum Brazing
The brazing process was performed in a small mobile homemade furnace placed inside a vacuum chamber. The standard vacuum ConFlat flange (nominal diameter 100 mm) with the assembled furnace and all electrical connections was placed downside in the vacuum chamber. The heating source in the furnace was a 450 W infrared (IR) lamp (Helios Italquarz, Cambiago-MI, Italy). The temperature was controlled by a K-type thermocouple placed inside the furnace with an automatic custom-made IR lamp backing control system. This system allowed us to control the heating temperature and the heating and cooling rate, since this aspect is extremely important for successful ceramics brazing. The small furnace reached the maximum temperature of 1050 °C.
In order to provide good thermal and mechanical contact during brazing, ceramic substrates were coated with 1-2 µm titanium deposited by DC magnetron sputtering at a 0.5-A DC current, 8.8 × 10 −3 mbar Ar pressure at a 6 cm target-substrate distance in 20 min. The appropriate sputtering In the current work, the films were deposited by direct current (DC) sputtering with a 2 planar, unbalanced, II Type magnetron cathode source. The magnetron drive model MDX 1.5 kW (Advanced Energy, Fort Collins, CO, USA) was used to complete the sputtering experiments with the possibility of an automatic control in order to set up the pulsing mode to perform multilayer deposition.
Depositions were performed onto a planar substrate holder (SH), which had the ability to heat up the substrate by controlling the temperature with a K-type thermocouple inserted in the SH plate. The distance between the substrate and the cathode was 6 cm. For film deposition, we used the down-top deposition configuration with a magnetron source placed from the bottom of the cylindrical vacuum chamber and substrate holder with substrates from the top of the chamber in order to minimize film delamination caused by metallic dust particles.
Molybdenum was deposited on a spot 10 mm in diameter in the center of each substrate (backing plate) defined by an appropriate mask.

Thin Film Analysis
FEI (Philips, Amsterdam, The Netherlands) SEM XL-30 was used for sputtered coatings analysis. The contact stylus profiler, model Dektak 8 (Veeco, Plainview, NY, USA), able to measure a surface texture below submicro-inch and film thickness up to 262 µm, was used for sputtered film thickness characterization.

Vacuum Brazing
The brazing process was performed in a small mobile homemade furnace placed inside a vacuum chamber. The standard vacuum ConFlat flange (nominal diameter 100 mm) with the assembled furnace and all electrical connections was placed downside in the vacuum chamber. The heating source in the furnace was a 450 W infrared (IR) lamp (Helios Italquarz, Cambiago-MI, Italy). The temperature was controlled by a K-type thermocouple placed inside the furnace with an automatic custom-made IR lamp backing control system. This system allowed us to control the heating temperature and the heating and cooling rate, since this aspect is extremely important for successful ceramics brazing. The small furnace reached the maximum temperature of 1050 • C.
In order to provide good thermal and mechanical contact during brazing, ceramic substrates were coated with 1-2 µm titanium deposited by DC magnetron sputtering at a 0.5-A DC current, 8.8 × 10 −3 mbar Ar pressure at a 6 cm target-substrate distance in 20 min. The appropriate sputtering pressure was chosen in order to minimize the intrinsic stress in the film and avoid titanium film peeling.
Homemade paste was used as the filler material, containing 63% Ag, 35.3% Cu, and 1.7% Ti powder (content corresponding to CuSil active brazing alloy, ABA) mixed by three-dimensional shaker-mixer, model TURBULA ® T2F (Glen Mills Inc., Clifton, NJ, USA), water, and a water-soluble glue that was partially hydrolyzed under heating potato starch solution as a binder.
The brazing process was realized in a vacuum at 920 • C in 15 min maintaining 5 • C/min heating and 3 • C/min cooling rate in order to minimize the thermomechanical stress in the final target assembly/system ( Figure 5).
Molecules 2018, 23, x FOR PEER REVIEW 14 of 24 pressure was chosen in order to minimize the intrinsic stress in the film and avoid titanium film peeling.
Homemade paste was used as the filler material, containing 63% Ag, 35.3% Cu, and 1.7% Ti powder (content corresponding to CuSil active brazing alloy, ABA) mixed by three-dimensional shaker-mixer, model TURBULA ® T2F (Glen Mills Inc., Clifton, NJ, USA), water, and a water-soluble glue that was partially hydrolyzed under heating potato starch solution as a binder.
The brazing process was realized in a vacuum at 920 °C in 15 min maintaining 5 ° C/min heating and 3 °C/min cooling rate in order to minimize the thermomechanical stress in the final target assembly/system ( Figure 5).
The general process diagram used to realize the target prototypes is shown in Figure 6.

Dissolution Test and Gamma-Spectroscopy
In order to prove the chemical inertness of the ceramic materials used in the target prototype construction (sapphire and CVD synthetic diamond), the targets were exposed to standard dissolution conditions (concentrated hydrogen peroxide at 70 °C) [10] after cyclotron irradiation. The The general process diagram used to realize the target prototypes is shown in Figure 6. pressure was chosen in order to minimize the intrinsic stress in the film and avoid titanium film peeling.
Homemade paste was used as the filler material, containing 63% Ag, 35.3% Cu, and 1.7% Ti powder (content corresponding to CuSil active brazing alloy, ABA) mixed by three-dimensional shaker-mixer, model TURBULA ® T2F (Glen Mills Inc., Clifton, NJ, USA), water, and a water-soluble glue that was partially hydrolyzed under heating potato starch solution as a binder.
The brazing process was realized in a vacuum at 920 °C in 15 min maintaining 5 ° C/min heating and 3 °C/min cooling rate in order to minimize the thermomechanical stress in the final target assembly/system ( Figure 5).
The general process diagram used to realize the target prototypes is shown in Figure 6.

Dissolution Test and Gamma-Spectroscopy
In order to prove the chemical inertness of the ceramic materials used in the target prototype construction (sapphire and CVD synthetic diamond), the targets were exposed to standard dissolution conditions (concentrated hydrogen peroxide at 70 °C) [10] after cyclotron irradiation. The

Dissolution Test and Gamma-Spectroscopy
In order to prove the chemical inertness of the ceramic materials used in the target prototype construction (sapphire and CVD synthetic diamond), the targets were exposed to standard dissolution conditions (concentrated hydrogen peroxide at 70 • C) [10] after cyclotron irradiation. The gamma spectroscopy analysis of an aliquot of the obtained solution was performed at the Research Laboratory of St. Orsola-Malpighi Hospital Medical Physic Department (Bologna).
The spectrometry system is based on a high purity Germanium detector with 30% relative efficiency and a resolution of 1.8 keV at 1332 keV. The samples were counted after a waiting time to allow for a decrease in radioactivity, sufficient to produce a total counting frequency <2000 counts per second in the 10-2000 keV energy range. The waiting time was typically within 1 h. The spectrometry system efficiency was calibrated in the 59-1836 keV range, using a multi radionuclide certified reference solution, obtained from an accredited Standardization Laboratory (LEA CERCA, Pierrelatte CEDEX, France). The calibration process was performed accordingly to the IEC 61452 standard, using Genie 2000 software (Version 3.2.1, Canberra Industries Inc., Canberra, Australia). A dual logarithmic polynomial efficiency curve was used. The propagation of uncertainties in the calibration considers the accuracy of the reference source (1-2% at 1 sigma level, depending on the peak in the mixture), the tabulated yield of peaks (typically <1%), the net peak area (<1% for calibration peaks), and the interpolation of the efficiency values. The latter is evaluated from the covariance matrix of the fitting of the efficiency curve (typically <3%). In total, using a quadratic propagation of the above terms, the calibration uncertainty was about 4-5% at the 1 sigma level.

Optimization of Magnetron Sputtering Parameters
The control of stress in physical vapor-deposited films is important due to its close relationship with material technological properties, adhesion strength to the substrate, and the limit of film thickness without cracking, buckling, or delamination [33]. In particular, for the purpose of this work, it was mandatory to avoid stressed films because poor adhesion between Mo films and the backing plate could drastically increase the thermal resistance of the contact, decreasing the heat exchange efficiency.
In this work, gas sputtering pressure and temperature of the holder were managed and optimized for the system configuration described above.
Theoretically, a certain gas pressure corresponds to the transition between the tensile and compressive stress. At relatively high pressure, the frequency of the gas phase collision increases, reducing the kinetic energy of the sputtered atoms and reflected neutrals bombarding the growing film exhibiting an open porous microstructure; the interatomic attractive forces produce tensile stress. At low pressure, the arriving atoms have high kinetic energy and the resulting film has a dense microstructure, experiencing compressive stress [34]. The optimal pressure was obtained experimentally by performing depositions of Mo onto flexible substrates. The radius of curvature taken by the Kapton is an indicator of the stress (Figure 7). The spectrometry system is based on a high purity Germanium detector with 30% relative efficiency and a resolution of 1.8 keV at 1332 keV. The samples were counted after a waiting time to allow for a decrease in radioactivity, sufficient to produce a total counting frequency <2000 counts per second in the 10-2000 keV energy range. The waiting time was typically within 1 h. The spectrometry system efficiency was calibrated in the 59-1836 keV range, using a multi radionuclide certified reference solution, obtained from an accredited Standardization Laboratory (LEA CERCA, Pierrelatte cedex, France). The calibration process was performed accordingly to the IEC 61452 standard, using Genie 2000 software (Version 3.2.1, Canberra Industries Inc., Canberra, Australia). A dual logarithmic polynomial efficiency curve was used. The propagation of uncertainties in the calibration considers the accuracy of the reference source (1-2% at 1 sigma level, depending on the peak in the mixture), the tabulated yield of peaks (typically < 1%), the net peak area (< 1% for calibration peaks), and the interpolation of the efficiency values. The latter is evaluated from the covariance matrix of the fitting of the efficiency curve (typically < 3%). In total, using a quadratic propagation of the above terms, the calibration uncertainty was about 4-5% at the 1 sigma level.

Optimization of Magnetron Sputtering Parameters
The control of stress in physical vapor-deposited films is important due to its close relationship with material technological properties, adhesion strength to the substrate, and the limit of film thickness without cracking, buckling, or delamination [33]. In particular, for the purpose of this work, it was mandatory to avoid stressed films because poor adhesion between Mo films and the backing plate could drastically increase the thermal resistance of the contact, decreasing the heat exchange efficiency.
In this work, gas sputtering pressure and temperature of the holder were managed and optimized for the system configuration described above.
Theoretically, a certain gas pressure corresponds to the transition between the tensile and compressive stress. At relatively high pressure, the frequency of the gas phase collision increases, reducing the kinetic energy of the sputtered atoms and reflected neutrals bombarding the growing film exhibiting an open porous microstructure; the interatomic attractive forces produce tensile stress. At low pressure, the arriving atoms have high kinetic energy and the resulting film has a dense microstructure, experiencing compressive stress [34]. The optimal pressure was obtained experimentally by performing depositions of Mo onto flexible substrates. The radius of curvature taken by the Kapton is an indicator of the stress (Figure 7).  The substrate temperature influences the kinetic energy of the particles that have already arrived at the substrate. Low temperature promotes the columnar voided microstructure that is associated with tensile stress. High temperature corresponds to an increase in adatom mobility that leads to a bulk-like structure [35]. In this work, we performed the deposition at the homologous temperature T h = T/T m = 0.2, where T is the temperature during vacuum deposition and T m is melting point of material deposited. The high temperature of the holder during the Mo sputtering process provided a microstructure with a density of more than 95% of the bulk material (Figure 8). The Mo film had a columnar microstructure with a grain size from several hundred nm to microns as confirmed by the SEM image in Figure 8a. The substrate temperature influences the kinetic energy of the particles that have already arrived at the substrate. Low temperature promotes the columnar voided microstructure that is associated with tensile stress. High temperature corresponds to an increase in adatom mobility that leads to a bulk-like structure [35]. In this work, we performed the deposition at the homologous temperature Th = T/Tm = 0.2, where T is the temperature during vacuum deposition and Tm is melting point of material deposited. The high temperature of the holder during the Mo sputtering process provided a microstructure with a density of more than 95% of the bulk material (Figure 8). The Mo film had a columnar microstructure with a grain size from several hundred nm to microns as confirmed by the SEM image in Figure 8a. A multilayer deposition technique was shown to reduce stress [36], thus the deposition of Mo thick films was fragmented in a thousand consecutive brief depositions of thin films using an automation program to control the power. Each deposition was followed by a "relaxation time" during which the film was annealed (80% duty cycle for a one-minute period).
The best sputtering process parameters to obtain unstressed Mo films onto copper, sapphire, and CVD diamond are listed in Table 2. nat Mo films of about 98% of bulk density with thickness over 100 µm were successfully deposited onto copper, sapphire, and synthetic diamond, which were further used either as ready target prototypes (Cu-1,2,3 and S-1,2) or for further vacuum brazing to create target prototypes (S-3,4,5 and D-1,2).

Irradiation
After irradiation, the targets were visually inspected, looking for any sign of melting, checking the overall integrity of the target, and the adherence of the Mo film to the backing materials. The results of the irradiation are summarized in Table 3 and visual details are shown in Figures 8-11.
The 110-125 µm of Mo sputtered directly onto copper backing targets were irradiated for 1 min at 15.6 MeV energy on the target (after the Havar ® foils) and currents 30, 50, 60, and 70 µA. No sign of target damage was observed after irradiation at 30 and 50 µA. The oxidation of the Mo and Cu A multilayer deposition technique was shown to reduce stress [36], thus the deposition of Mo thick films was fragmented in a thousand consecutive brief depositions of thin films using an automation program to control the power. Each deposition was followed by a "relaxation time" during which the film was annealed (80% duty cycle for a one-minute period).
The best sputtering process parameters to obtain unstressed Mo films onto copper, sapphire, and CVD diamond are listed in Table 2. nat Mo films of about 98% of bulk density with thickness over 100 µm were successfully deposited onto copper, sapphire, and synthetic diamond, which were further used either as ready target prototypes (Cu-1,2,3 and S-1,2) or for further vacuum brazing to create target prototypes (S-3,4,5 and D-1,2).

Irradiation
After irradiation, the targets were visually inspected, looking for any sign of melting, checking the overall integrity of the target, and the adherence of the Mo film to the backing materials. The results of the irradiation are summarized in Table 3 and visual details are shown in Figures 8-11.
The 110-125 µm of Mo sputtered directly onto copper backing targets were irradiated for 1 min at 15.6 MeV energy on the target (after the Havar ® foils) and currents 30, 50, 60, and 70 µA. No sign of target damage was observed after irradiation at 30 and 50 µA. The oxidation of the Mo and Cu surfaces (observed in Cu-1 and 2) was related to the previous leakage of a small amount of cooling water in the target system. The copper target test proved that sputtered Mo film resists at a 70 µA irradiation current, maintaining excellent contact between the Mo film and Cu backing (Figure 9). surfaces (observed in Cu-1 and 2) was related to the previous leakage of a small amount of cooling water in the target system. The copper target test proved that sputtered Mo film resists at a 70 µA irradiation current, maintaining excellent contact between the Mo film and Cu backing (Figure 9).   We place 90 µm of natural Mo on a sapphire coin (sample S-1 and S-2) inside the four-pin target holder for irradiation at 10, 20, 30, and 50 µA for 1 min. The Mo film on sapphire resisted a beam current of 10 µA and 20 µA and the sapphire remained intact. At 30 µA current, sapphire-backed plate cracked but the Mo film kept the pieces together, without evidence of delamination or oxidation (Figure 10). At 50 µA, the sapphire cracked into two pieces (one of them was used for the dissolution test described below). It is clear that the contact between sapphire and the copper four-pin holder was extremely poor. A better contact is essential in order to perform the production at higher proton beam currents.
Molecules 2018, 23, x FOR PEER REVIEW

of 24
We place 90 µm of natural Mo on a sapphire coin (sample S-1 and S-2) inside the four-pin target holder for irradiation at 10, 20, 30, and 50 µA for 1 min. The Mo film on sapphire resisted a beam current of 10 µA and 20 µA and the sapphire remained intact. At 30 µA current, sapphire-backed plate cracked but the Mo film kept the pieces together, without evidence of delamination or oxidation (Figure 10). At 50 µA, the sapphire cracked into two pieces (one of them was used for the dissolution test described below). It is clear that the contact between sapphire and the copper four-pin holder was extremely poor. A better contact is essential in order to perform the production at higher proton beam currents.

As prepared
After irradiation 10 µA 20 µA 30 µA Figure 10. Irradiation test of S01 Mo sputtered on sapphire target prototype inside a four-pin target holder.
The prototype S-3, composed of 110 µm Mo sputtered onto 12.7-mm-diameter and 0.5-mm-thick sapphire, brazed onto copper, was tested in several steps by increasing the current in each following run, from 30 µA to the maximum cyclotron current of 60 µA, always at maximum beam energy 15.6 MeV. Each irradiation test lasted for 1 min. No cracking of the sapphire piece occurred. The prototypes S-4 and S-5 were tested directly under the maximum current of 60 µA at 15.6 MeV for 1 min. The systems resisted the proton beam irradiation as shown in Figure 11. Notably, in the time between the prototype S-4 preparation and irradiation test, a defect between the sapphire and brazing filler occurred, which seems to be attributed to a problem with the Ti sputtered film adherence before the brazing process of the sapphire. For the following preparation of diamond-based prototypes, the Ti sputtering parameters were optimized and additional attention was paid to substrate cleaning procedures in order to avoid this problem.

S-3
S-4 S-5 As prepared After irradiation 60 µA 60 µA 60 µA The prototype S-3, composed of 110 µm Mo sputtered onto 12.7-mm-diameter and 0.5-mm-thick sapphire, brazed onto copper, was tested in several steps by increasing the current in each following run, from 30 µA to the maximum cyclotron current of 60 µA, always at maximum beam energy 15.6 MeV. Each irradiation test lasted for 1 min. No cracking of the sapphire piece occurred. The prototypes S-4 and S-5 were tested directly under the maximum current of 60 µA at 15.6 MeV for 1 min. The systems resisted the proton beam irradiation as shown in Figure 11. Notably, in the time between the prototype S-4 preparation and irradiation test, a defect between the sapphire and brazing filler occurred, which seems to be attributed to a problem with the Ti sputtered film adherence before the brazing process of the sapphire. For the following preparation of diamond-based prototypes, the Ti sputtering parameters were optimized and additional attention was paid to substrate cleaning procedures in order to avoid this problem.
Molecules 2018, 23, x FOR PEER REVIEW

of 24
We place 90 µm of natural Mo on a sapphire coin (sample S-1 and S-2) inside the four-pin target holder for irradiation at 10, 20, 30, and 50 µA for 1 min. The Mo film on sapphire resisted a beam current of 10 µA and 20 µA and the sapphire remained intact. At 30 µA current, sapphire-backed plate cracked but the Mo film kept the pieces together, without evidence of delamination or oxidation (Figure 10). At 50 µA, the sapphire cracked into two pieces (one of them was used for the dissolution test described below). It is clear that the contact between sapphire and the copper four-pin holder was extremely poor. A better contact is essential in order to perform the production at higher proton beam currents.

As prepared
After irradiation 10 µA 20 µA 30 µA Figure 10. Irradiation test of S01 Mo sputtered on sapphire target prototype inside a four-pin target holder.
The prototype S-3, composed of 110 µm Mo sputtered onto 12.7-mm-diameter and 0.5-mm-thick sapphire, brazed onto copper, was tested in several steps by increasing the current in each following run, from 30 µA to the maximum cyclotron current of 60 µA, always at maximum beam energy 15.6 MeV. Each irradiation test lasted for 1 min. No cracking of the sapphire piece occurred. The prototypes S-4 and S-5 were tested directly under the maximum current of 60 µA at 15.6 MeV for 1 min. The systems resisted the proton beam irradiation as shown in Figure 11. Notably, in the time between the prototype S-4 preparation and irradiation test, a defect between the sapphire and brazing filler occurred, which seems to be attributed to a problem with the Ti sputtered film adherence before the brazing process of the sapphire. For the following preparation of diamond-based prototypes, the Ti sputtering parameters were optimized and additional attention was paid to substrate cleaning procedures in order to avoid this problem.
The first CVD synthetic diamond-based prototype D-1 was tested in several steps by increasing the current in each following run from 20 µA to 40 µA, and finally at maximum cyclotron current of 60 µA. A maximum entrance beam energy 15.6 MeV was used. Each irradiation test lasted 1 min. No cracking or oxidation of the target prototype occurred during irradiation as shown in Figure 12.
One irradiation test on the D-2 CVD synthetic diamond-based prototype was performed under conditions closer to those used in actual production. The irradiation was performed for 30 min at 60 µA and 15.6 MeV. Similar to the first diamond-based prototype, the second one showed excellent performance during the long irradiation test without any signs of damage ( Figure 12).
The new cyclotron target prototype composed of nat Mo sputtered on synthetic diamond brazed to Cu backing produced excellent results from a thermomechanical point of view, remaining stable at heat power densities in the order of 1 kW/cm 2 .
It should be noted that the cyclotron current typically used for electrodeposited solid targets (for example, 64 Ni on gold backing for 64 Cu production [37][38][39] and 68 Zn on platinum backing for 68 Ga production [40]) ranges from 20 to 50 µA, corresponding to a maximum heat power density of about 0.5 kW/cm 2 . Hence, the use of the sputtering process for target preparation can increase the radiopharmaceutical production yield two-fold. This demonstrates the better performance of the sputtered targets in respect to those produced by electrodeposition. D-1 D-2

Dissolution Test and Chemical Inertness Prove
The test of the chemical inertness of the ceramic part of the backing involved dissolution followed by γ-spectroscopy analysis.
A piece of S-2 irradiated target was dissolved and 10 µL of solution were placed in the γspectrometer in order to control if any impurities were released by the irradiated sapphire. The range of the γ-ray energy was 80-4096 keV. According to the software, the peaks belonging to 92m Nb, 94 Tc, 95 Tc, 95m Tc, 96 Tc, 99 Mo, and 99m Tc were identified (Table 4). All radionuclides are the irradiation products of natural molybdenum. 92m Nb is produced by the reaction 95 Mo(p,α); all other radionuclides are produced by a set of (p,xn) reactions starting from the natural molybdenum One irradiation test on the D-2 CVD synthetic diamond-based prototype was performed under conditions closer to those used in actual production. The irradiation was performed for 30 min at 60 µA and 15.6 MeV. Similar to the first diamond-based prototype, the second one showed excellent performance during the long irradiation test without any signs of damage ( Figure 12).
The new cyclotron target prototype composed of nat Mo sputtered on synthetic diamond brazed to Cu backing produced excellent results from a thermomechanical point of view, remaining stable at heat power densities in the order of 1 kW/cm 2 .
It should be noted that the cyclotron current typically used for electrodeposited solid targets (for example, 64 Ni on gold backing for 64 Cu production [37][38][39] and 68 Zn on platinum backing for 68 Ga production [40]) ranges from 20 to 50 µA, corresponding to a maximum heat power density of about 0.5 kW/cm 2 . Hence, the use of the sputtering process for target preparation can increase the radiopharmaceutical production yield two-fold. This demonstrates the better performance of the sputtered targets in respect to those produced by electrodeposition.

Dissolution Test and Chemical Inertness Prove
The test of the chemical inertness of the ceramic part of the backing involved dissolution followed by γ-spectroscopy analysis.
A piece of S-2 irradiated target was dissolved and 10 µL of solution were placed in the γ-spectrometer in order to control if any impurities were released by the irradiated sapphire. The range of the γ-ray energy was 80-4096 keV. According to the software, the peaks belonging to 92m Nb, 94 Tc, 95 Tc, 95m Tc, 96 Tc, 99 Mo, and 99m Tc were identified (Table 4). All radionuclides are the irradiation products of natural molybdenum. 92m Nb is produced by the reaction 95 Mo(p,α); all other radionuclides are produced by a set of (p,xn) reactions starting from the natural molybdenum isotopes: 92 Mo, 94 Mo, 95 Mo, 96 Mo, 97 Mo, 98 Mo, and 100 Mo [8,41]. Therefore, any contaminant radionuclides produced by the sapphire backing were present in the solution. This proves the chemical inertness of sapphire in the chosen dissolution conditions. The thermal grade CVD diamond (D-0 sample), Ø 13.5 mm and 0.3 mm thick with thermal conductivity 1500 W/(m·K), was tested under the 15.6 MeV and 20 µA cyclotron accelerated proton beam for 1 min inside the four-pin target-holder. After irradiation, the synthetic diamond piece was placed into concentrated H 2 O 2 . No change in the mass with respect to the mass of the diamond before dissolution was observed. The piece of synthetic diamond was then used for the γ-spectroscopy analysis to identify the elements created inside the diamond piece under irradiation. The γ-spectrum contained only one peak at 511 keV. The 511 keV is the energy of the photons due to the electron-positron annihilation. Thus, the radioactive isotope produced in the synthetic diamond during the proton irradiation was a positron emitter. The only suitable radioisotope is 13 N decaying to 13 C, emitting a positron. In order to prove this, the activity from the sample was measured via dose-calibrator versus time to build a decay curve. The time corresponding to the moment when 50% of activity left was 10 min. This means that T 1/2 = 10 min. The radioisotope produced in CVD synthetic diamond was 13 N.
No other products besides 13 N were detected in the CVD synthetic diamond plate after 15.6 MeV proton beam irradiation. The synthetic diamond plate does not release any foreign radionuclides after irradiation according to γ-spectroscopy analysis.

Magnetron Sputtering Efficiency and Further Perspectives
The main defect of the magnetron sputtering technique is the low efficiency of the deposition, which means high losses of enriched material. The losses are attributed to two main factors: low sputtering target use in standard configuration and losses of the sputtering chamber. The first can be solved using an advanced technique with High Target Utilization plasma Sputtering (HiTUS) [42]. The method for ultra-thick film sputtering proposed here can be easily applied to different sputtering techniques since it is based on main PVD principles. For the other factor, an efficient method to recover enriched Mo material from the sputtering chamber after deposition is required. The LARAMED group of LNL-INFN has already developed a method for enriched Mo recovery based on dissolution in an ammonium hydroxide and peroxide, precipitation of MoO 3 , and further molybdenum oxide reduction in an overpressure hydrogen reactor.
Positive results obtained for nat Mo targets have created potential for the application of the magnetron sputtering technique for a high current target with non-enriched elements as precursors. As an example, the LARAMED group of LNL-INFN has successfully developed a preparation method for nat Y target for the production of 89 Zr.

Conclusions
In this work, we developed a sputtering deposition process of a target material, producing Mo films more than 100 µm thick with bulk-grade density, low oxidation level, and high adherence to the backing plate, in order to create solid targets for medical radioisotope production. Due to the versatility of the magnetron sputtering technique, target prototypes were created by direct deposition on both metallic (copper) and non-metallic (sapphire and CVD synthetic diamond) target backing.
The use of sapphire and CVD synthetic diamond as backing materials guarantees chemical inertness during the dissolution process after irradiation in order to minimize the impurities. In order to reduce the implementation costs, a composite backing plate made of a non-metallic part vacuum brazed to copper was suggested. In order to produce such prototypes, the vacuum brazing method of sapphire and CVD synthetic diamond to copper was successfully developed using home-made brazing paste.
Solid cyclotron target prototypes, both on simple copper backing and on complex backing based on sapphire or CVD synthetic diamond substrates, showed excellent thermomechanical stability under 15.6 MeV and 60 µA (maximum available from used PETtrace cyclotron) proton beam, corresponding to a heat power density of about 1 kW/cm 2 .

Patents
From the work reported in this manuscript, an Italian patent application No. 102017000102990, dep. ref. P1183IT00, inventors V. Palmieri, H. Skliarova, S. Cisternino, M. Marengo, G. Cicoria, title "Metodo per l'ottenimento di un target solido per la produzione di radiofarmaci", was applied by Istituto Nazionale di Fisica Nucleare on 14.09.17; and extended to the International patent application PCT/IB2018/056826, dep. ref. P1183PC00, on 07.09.18, title "Method for obtaining a solid target for radiopharmaceuticals production".