Introduction The established methods for the production of 99Mo, based on fission in nuclear reactors, continue to present problems as a result of the plant’s aging and the significant investments needed for maintenance or for their renewal. Much research work is thus in progress on the study of alternative methods for the production of 99mTc in quantities and with the degree of purity required for the clinical use. Between them, the cyclotron production of 99mTc via the 100Mo(p,2n)99mTc reaction has turned out as the most attractive alternative. One critical aspect regarding the production of 99mTc with cyclotron is the need for a robust and reliable target production process. Several techniques have been indicated as extremely promising such has plasma spray and laser cladding; however these methods require specialized instrumentation and complex operations to be performed handling activated materials in order to recover irradiated Mo. In this work we report the development of the work done at the University of Bologna, as a part of a wider INFN project, as regards the methods of preparation of solid targets suitable for the production of 99mTc irradiating a target of 100Mo, employing a cyclotron for biomedical use, normally operated for the production of PET radionuclides. Material and Methods Irradiations were performed with a 16.5 MeV GE PETtrace cyclotron equipped with a solid target station previously developed by our group (1). In initial tests, a stack of 1–3 metallic foils, 100 μm thick, of natMo were irradiated with protons in the 15.9→9.8 MeV energy range. Foils were then dissolved in a HNO3-HCl solution and samples were analyzed with high resolution gamma-ray spectrometry (Canberra, including a HPGe detector with a 30% relative efficiency and a resolution of 1.8 keV at 1332 keV) using Genie2000 software; the measurement campaign lasted several weeks to take into account the different half-lives of the produced radionuclides. Results were extrapolated to a highly enriched 100Mo target and compared to Monte Carlo simulations previously performed with FLUKA and TALYS codes (2). In order to investigate a method of preparation of the target that would make easier the recovery of the enriched material and recycling for the preparation of a new target, it was subsequently studied the preparation of pellets of Molybdenum trioxide. MoO3 powder (Sigma Aldrich, 99.9% trace metals basis, particle size < 150 μm) was used to prepare pellets using a 10 ton press. Pellets obtained in this way were then sintered on a Platinum support using a CARBOLITE furnace under a controlled atmosphere; the temperature was ramped according to a controlled and reproducible temperature cycle. Sintered pellets were subjected to visual inspection, mechanical tests of resistance to loading and downloading in the cyclotron target station, thermal tests and then irradiated at increasing current. The irradiated targets were again visually inspected then weighed, dissolved and subjected to gamma-ray spectrometry analysis. Results and Conclusion The experimental saturation yield for 99mTc calculated on the basis of the gamma-ray analysis of irradiated metal foils, gave an extrapolated yield of 1.115 ± 0.015 GBq/μA for a 100 μm thick 100Mo enriched target, in accordance with the value of 1.107 ± 0.002 GBq/μA obtained in Monte Carlo simulations. On these bases, an irradiation of 3 h at 50 μA is expected to produce 16.3 ± 0.2 GBq of 99mTc; considering the use of an efficient purification system, a radionuclidic purity > 99.9 % 2 h after the EndOfBombardment and a specific activity comparable with the actual standards are expected as achievable. Experiments on sintering pellets are still on going at the time of writing this report; initial results showed that addition of proper aggregating materials allows for suitable pellets preparation. The sintering process allows to obtain pellets having sufficient mechanical strength to withstand loading and downloading operations. Initial irradiation tests with beam current up to 25 μA were performed successfully with no changes in mass and mechanical properties of the pellet. These encouraging results suggest that sintered pellets may be a relatively inexpensive and easy solution to prepare 100Mo targets for the cyclotron production of 99mTc. Further experimental tests at higher beam current will be performed in order to assess the maximum current achievable with no damage of the target. At the same time, a prototype automated module based on standard industrial components is in testing phase as regards performance in the separation and purification processes.

Introduction Technetium-99m, supplied in the form of 99Mo/99mTc generators, is the most widely used radioisotope for nuclear medical imaging. The parent isotope 99Mo is currently produced in nuclear reactors. Recent disruptions in the 99Mo supply chain [1] prompted the development of methods for the direct accelerator-based production of 99mTc. Our approach involves the 100Mo(p,2n)99mTc reaction on isotopically enriched molybdenum using small medical cyclotrons (Ep ≤ 20 MeV), which is a viable method for the production of clinically useful quantities of 99mTc [2]. Multi-Curie production of 99mTc requires a 100Mo target capable of dissipating high beam intensities [3]. We have reported the fabrication of 100Mo targets of both small and large area tar-gets by electrophoretic deposition and subsequent sintering [4]. As part of our efforts to further enhance the performance of molybdenum targets at high beam currents, we have developed a novel target system (initially de-signed for the GE PETtrace cyclotron) based on a pressed and sintered 100Mo plate brazed onto a dispersion-strengthened copper backing. Materials and Methods In the first step, a molybdenum plate is produced similarly to the method described in [5] by compacting approximately 1.5 g of commercially available 100Mo powder using a cylindrical tool of 20 mm diameter. A pressure between 25 kN/cm2 and 250 kN/cm2 is applied by means of a hydraulic press. The pressed molybdenum plate is then sintered in a reducing atmosphere (Ar/2% H2) at 1,700 oC for five hours. The resulting 100Mo plates have about 90–95 % of the molybdenum bulk density. The 100Mo plate is furnace brazed at ~750 oC onto a backing manufactured from a disperse on strengthened copper composite (e.g. Glidcop AL-15) using a high temperature silver-copper brazing filler. This process yields a unique, mechanically and thermally robust target system for high beam power irradiation. Irradiations were performed on the GE PETtrace cyclotrons at LHRI and CPDC with 16.5 MeV protons and beam currents ≥ 100 µA. Targets were visually inspected after a 6 hour, 130 µA bombardment (2.73 kW/cm2, average) and were found fully intact. Up to 4.7 Ci of 99mTc have been produced to date. The saturated production yield remained constant between 2 hour and 6 hour irradiations. Results and Conclusion These results demonstrate that our brazed tar-get assembly can withstand high beam intensities for long irradiations without deterioration. Efforts are currently underway to determine maximum performance parameters.

Introduction The goal of this work is to advance the production and use of 52Mn (t½ = 5.6 d, β+: 242 keV, 29.6%) as a radioisotope for in vivo preclinical nuclear imaging. More specifically, the aims of this study were: (1) to measure the excitation function for the natCr(p,n)52Mn reaction at low energies to verify past results [1–4]; (2) to measure binding constants of Mn(II) to aid the design of a method for isolation of Mn from an irradiated Cr target via ion-exchange chromatography, building upon previously published methods [1,2,5–7]; and (3) to perform phantom imaging by positron emission tomography/magnetic resonance (PET/MR) imaging with 52Mn and non-radioactive Mn(II), since Mn has potential dual-modality benefits that are beginning to be investigated [8]. Material and Methods Thin foils of Cr metal are not available commercially, so we fabricated these in a manner similar to that reported by Tanaka and Furukawa [9]. natCr was electroplated onto Cu discs in an industrial-scale electroplating bath, and then the Cu backing was digested by nitric acid (HNO3). The remaining thin Cr discs (~1 cm diameter) were weighed to determine their thickness (~ 75–85 μm) and arranged into stacked foil targets, along with ~25 μm thick Cu monitor foils. These targets were bombarded with ~15 MeV protons for 1–2 min at ~1–2 μA from a CS-15 cyclotron (The Cyclotron Corporation, Berkeley, CA, USA). The beamline was perpendicular to the foils, which were held in a machined 6061-T6 aluminum alloy target holder. The target holder was mounted in a solid target station with front cooling by a jet of He gas and rear cooling by circulating chilled water (T ≈ 2–5 °C). Following bombardment, these targets were disassembled and the radioisotope products in each foil were counted using a high-purity Ge (HPGe) detector. Cross-sections were calculated for the natCr(p,n)52Mn reaction. Binding constants of Mn(II) were measured by incubating 54Mn(II) (t½ = 312 d) dichloride with anion- or cation-exchange resin (AG 1-X8 (Cl− form) or AG 50W-X8 (H+ form), respectively; both: 200–400 mesh; Bio-Rad, Hercules, CA) in hydrochloric acid (HCl) ranging from 10 mM-8 M (anion-exchange) and from 1 mM-1 M (cation-exchange) or in sulfuric acid (H2SO4) ranging from 10 mM-8 M on cation-exchange resin only. The amount of unbound 54Mn(II) was measured using a gamma counter, and binding constants (KD) were calculated for the various concentrations on both types of ion-exchange resin. We have used the unseparated product for preliminary PET and PET/MR imaging. natCr metal was bombarded and then digested in HCl, resulting in a solution of Cr(III)Cl3 and 52Mn(II)Cl2. This solution was diluted and imaged in a glass scintillation vial using a microPET (Siemens, Munich, Germany) small animal PET scanner. The signal was corrected for abundant cascade gamma-radiation from 52Mn that could cause random false coincidence events to be detected, and then the image was reconstructed by filtered back-projection. Additionally, we have used the digested target to spike non-radioactive Mn(II)Cl2 solutions for simultaneous PET/MR phantom imaging using a Biograph mMR (Siemens) clinical scanner. The phantom consisted of a 4×4 matrix of 15 mL conical tubes containing 10 mL each of 0, 0.5, 1.0, and 2.0 mM concentrations of non-radioactive Mn(II)Cl2 with 0, 7, 14, and 27 μCi (at start of PET acquisition) of 52Mn(II)Cl2 from the digested target added. The concentrations were based on previous MR studies that measured spin-lattice relaxation time (T1) versus concentration of Mn(II), and the activities were based on calculations for predicted count rate in the scanner. The PET/MR imaging consisted of a series of two-dimensional inversion-recovery turbo spin echo (2D-IR-TSE) MR sequences (TE = 12 ms; TR = 3,000 ms) with a wide range of inversion times (TI) from 23–2,930 ms with real-component acquisition, as well as a 30 min. list-mode PET acquisition that was reconstructed as one static frame by 3-D ordered subset expectation maximization (3D-OSEM). Attenuation correction was performed based on a two-point Dixon (2PD) MR sequence. The DICOM image files were loaded, co-registered, and windowed using the Inveon Research Workplace software (Siemens).

Introduction We have developed a low-cost stopcock valve actuator for radiochemistry automation built using a stepper motor and an Arduino, an open-source single-board microcontroller. The con-troller hardware can be programmed to run by serial communication or via two 5–24 V digital lines for simple integration into any automation control system. This valve actuator allows for automated use of a single, disposable stopcock, providing a number of advantages over stopcock manifold systems available on many commercial radiochemistry rigs or over using solenoid valves. This actuator allows for the use a wide variety of stopcocks, ranging in size, shape and material, giving flexibility to be used in a large variety of applications. Material and Methods The actuated valve consists of two main parts, the actuator and the control electronics. The actuator consists of a stepper motor, an infrared ‘home position’ sensor, a stopcock backplate, and a coupler from the driveshaft to stopcock handle. The stepper motor is a NEMA-17 size that runs 200 steps/rotation with a 5mm drive shaft. The coupler is an interchangeable part, custom to each stopcock model, with each part drilled out to fit the motor drive shaft and milled out for a tight fit to the stopcock handle. The backplane consists of a plate offset from the motor body with 5 screws positioned to keep the stopcock body from rotating relative to the motor. A reflective optical sensor (Vishay TCRT1000) is used as a limit switch to determine a ‘home’ position for the stopcock. With a slight modification to most any stopcock in cutting off a tab that limits rotation, the handle can rotate 360°. This allows for opening all three ports to each other, which has been done to all stopcocks used with this actuator. The control electronics consist of an Arduino Uno board and a motor shield (add-on board), connecting to the actuator by an Ethernet cable. The motor shield functions to interface the low-power Arduino circuitry with a high power H-bridge motor driver circuit. The Arduino runs two sets of code, initialization and its loop. The initialization routine runs when power is first powered up, and then continues to run the loop. The initialization routine rotates the valve until the IR limit switch is activated, and rotates an-other 45° from position home, sealing off all ports on the stopcock. Following initialization, the Arduino enters its loop, which repeatedly compares its current position to its target posi-tion. When the target position and current posi-tion do not match, the stepper motor turns in the shortest direction towards its target position. The hardware can be interfaced by either serial communication or by two 5–24V digital signals defining positions 1–4. The wide range of allowed input signal voltages is realized by using an optocoupler that accepts 5–24 V inputs but outputs TTL signals compatible with the Arduino’s hardware. Results and Conclusion A photo of the implementation of the actuator is shown in FIGURE 1. It has overall dimensions of 3.5×1.75×2.5”, excluding a mounting bracket. Control electronics are housed in a compact box built for an Arduino, giving the control electron-ics a clean, professional look. Challenges in de-sign included determining a maximum motor speed where the motor would provide enough torque but yet move fast enough to be useful, finding that rotational speed of 6 seconds/full rotation is best.

Introduction The internal pressure experienced by a gas tar-get during irradiation is dependent on the beam energy deposited in the target, the beam cur-rent, and the thermal behaviour of the target. [1] The maximum beam energy deposited is a function of the cyclotron capabilities and the gas inventory within the target. The maximum beam current is limited by the pressure produced in the target and the ability of the target assembly to remain intact. This is also a function of the thermal behaviour of the target, which is difficult to predict a priori since it is dependent on such things as convection currents that occur during irradiation. We conducted bench tests with model gas targets with and without forced convection currents to observe the effect on thermal behaviour. Based on those results we constructed a prototype gas target, suitable for irradiation, with an internal fan assembly that is rotated via external magnets. Material and Methods Bench tests were conducted with cylindrical and conical target bodies of aluminum. A nickel-chromium heater wire was inserted into the gas volume through the normal beam entrance port (FIGURE 1) to heat the gas while water cooling was applied to the target body. The voltage and current of the heater coil was monitored along with the pressure inside the target and the water inlet and outlet temperature. In the case of tests with a driven fan blade either the voltage applied to the electric motor was monitored or the fan speed itself was recorded. By assuming the ideal gas law, the pressure gives the average bulk temperature and a global heat transfer coefficient can be calculated between the target gas and the cooling water. [2] A cylindrical target body was constructed that incorporated a fan blade driven by an external motor. This assembly used a simple o-ring seal on the rotating shaft. This seal was not robust enough for any tests under beam conditions. A prototype design suitable for in-beam operation employs a propeller mounted on a rotating disc housing two samarium cobalt magnets and spinning on two micro-bearings which are constructed to operate in high temperature environments. The micro-bearings are mounted on a pin projecting from a plate welded to the back of the gas target to allow assembly of the fan mechanism prior to attachment to the body (FIGURE 2).

Introduction Batch-style water targets used for F-18 production are known to operate under boiling conditions in the target irradiation chamber, but the distribution of vapor under steady-state conditions was previously unknown. Thermal performance of batch targets has been correlated to average void in the target [1], but the simplified assumptions of such models do not represent the true non-uniform boiling behavior. Visualization targets can be used to observe boiling inside of a target during operation [2–5]. Commercial BTI targets operate at 28–35 bar (400–500 psi) with heat inputs of 0.5 to 3 kW and fill volumes of 1 to 4 mL. Recently, a visualization target featuring two transparent viewing windows was used to observe boiling conditions for realistic operating beam power, target pressure, and fill volume [4]. The same methodology has been applied to three additional visualization targets to examine the effect of target geometry on observed boiling phenomena. Material and Methods The original visualization target featured an aluminum body with a 0.127 mm (0.005 inch) integral aluminum beam window and two viewing windows made of optically clear sapphire (Al2O3). It was operated on an IBA 18/9 cyclotron with 18 MeV protons at beam power up to 1.1 kW, for pressures of 5 to 21 bar (70 to 300 psi), and a fill volume of 2.5 mL. All of the new designs featured a wider chamber to allow for higher beam transmission and an increased chamber height, consistent with cur-rent trends in high power targets. One target featured a reduced chamber depth, and another had a ramp in the back of the chamber to reduce fill volume. Target pressure was limited to a maximum of 14 bar (200 psi) due to the larger diameter beam window. A video camera was used to record the boiling conditions observed for each target under several lighting conditions. During irradiation, the proton beam excites the water molecules, producing visible blue light emissions during de-excitation. These light emissions provide a good indication of beam distribution and penetration depth. A strong backlight can be used to produce clearer images of bubbles generated during boiling. Results and Conclusion Proton range and visible blue light emissions were recorded in dark ambient conditions. The width of the Bragg peak and natural circulation in the bulk fluid were visible with good ambient lighting. Size and distribution of vapor bubbles could be observed by using a strong backlight. The beam current was increased gradually to determine the thermal limit for each target for several fill volumes and pressures. Two thermal limits were observed which resulted in some beam penetration in the top region of the beam. For lower fill volumes, steam ac-cumulates in or around the helium overpressure bubble, causing the helium bubble to move into the upper region of the beam. For higher fill volumes, beam penetration occurs due to excessive voiding, when bubbles produced in the beam region cannot rise quickly enough out of the path of the beam.

Introduction The worldwide use of PET has proven beyond dispute the importance for both routine diagnosis and physiological, oncological and pharmacological research. In many ways the present success of PET relies on the mature technology of PET compact medical cyclotrons. As long time developers of new targets, isotopes and com-pounds, we have been inclined to look for new block-buster applications, high power targets and sustainable ways of embracing the GMP and regional distribution, but recent pioneering development [1] around very small cyclotrons and “embedded synthesis and qc” has pointed out an old, but important nuclear physics lesson now halfway forgotten: that many PET isotopes can be made in high yields with proton energies far below 10 MeV [2]. This has opened a new interest in small cyclotrons and their targets. We have been testing the first GE Healthcare Prototype for a 7.8 MeV negative ion, internal ion source cyclotron with 3 production targets mounted on a short beamline. Here we present the first experimental yields of some of the important PET radionuclides. Materials and methods The prototype cyclotron (FIG. 1) has been in-stalled and tested without self-shield in designated experimental area in order to establish the neutron field around accelerator and targets in order to qualify design calculations for a future integrated shield. The cyclotron energy is fixed by the radial position of the extraction foil, while the azimuth determines which of the 3 targets are being irradiated. The beam energy at front of target foil was determined on several occasions: 7.8 ± 0.1 MeV by a 2 copper-foil sandwich method (adopted from [3]). The available beam inside the cyclotron at extractor position is > 50 μA, and 35 μA are easily and long term reliably extracted (> 90 %) on to any of the 3 target positions. The prototype is capable of delivering more than 40 μA to target, but target current was limited to 35 μA under present unshielded conditions. Results 18F We have tested the prototype gridded (> 80 % transmission) niobium body target with 10μm Havar foil using 95 % 18O water and 35 μA on target + grid with yields given in TABLE 1. The observed yields corrected for stopping in foil, grid loss and water enrichment are 75 % of theoretical. One Fastlab FDG run using 2 h irradiation yielded 16 GBq FDG EOS, confirming the “usual” 18F activity. Results 11C Using gridded target and a 10μm foil with 99% N2 + 1 % O2 at 10 bar followed by trapping into ascarite gave EOB activity as shown in TABLE 2. Results 13N We know that the 16O(p,alpha)13N cross section is a very steep function of energy around 7.8 MeV. In the hope of using the simple water target route to 13N NH3 we have measured the 13N yields (corrected for 18F contribution). It is still unclear if these yields can be improved to make useful single doses of ammonia. Results for other isotopes We have used solid targets to make 45Ti, 64Cu, 68Ga and 89Zr. The development of these solid targets is still in progress, but especially the 68Ga yield looks promising (3 GBq EOB after 1 h on natural Zn will give > 15 GBq on enriched 68Zn).

Es wurde eine Elektron-Zyklotron-Resonanz-Ionenquelle mit einer Mikrowellenfrequenz von2,45 GHz für die Produktion intensiver Ströme einfach geladener Metallionen entwickelt. Deren Beladung mit Metalldampf erfolgt über ein integriertes zylindrisches Sputtermagnetron, welches speziell für diese Aufgabe entworfen wurde. Die entstandene MECRIS, engl. Magnetron Electron Cyclotron Resonance Ion Source, vereinigt die ECR-Ionenquellentechnologie mit der Magnetron-Sputtertechnologie auf bisher einzigartige Weise und verkörpert so ein neues Metallionen-Quellenkonzept. Unter Verwendung eines Al-Sputtertargets konnte die Funktionsfähigkeit der MECRIS an dem Beispiel der Al+-Ionenerzeugung erfolgreich demonstriert werden. Der extrahierbare Al+-Ionenstrom wurde über einen neuartigen, im Rahmen der Arbeit entwickelten, Hochstrom-Faraday-Cup gemessen. Auf Basis numerischer Berechnungen wurde das Gesamtmagnetfeld so ausgelegt, dass die Permanentmagnete des Magnetrons und die Spulen der ECR-Quelle eine Minimum-B-Struktur erzeugen, welche einen effektiven Elektroneneinschluss nach dem magnetischen Spiegelprinzip ermöglicht. Gleichzeitig wird durch eine geschlossene ECR-Fläche, mit der magnetischen Resonanzflussdichte von 87,5 mT, eine optimale Heizung der Plasmaelektronen realisiert. Die mithilfe einer Doppel-Langmuir-Sonde gemessene Elektronentemperatur steigt in Richtung Quellenmitte an und beträgt maximal 11 eV. Geheizte Elektronen erlauben die effiziente Stoßionisation der Al-Atome, welche mit einer Rate von über 1E18 Al-Atome/s eingespeist werden und eine höchstmögliche Dichte von 2E10 1/cm³ aufweisen. Die MECRIS erzeugt hauptsächlich einfach geladene Ionen des gesputterten Materials (Al+) und des Prozessgases (Ar+). Der Al+-Ionenextraktionsstrom ist über die Erhöhung der Prozessparameter Sputterleistung, Mikrowellenleistung, Spulenstrom und Extraktionsspannung um eine Größenordnung bis auf maximal 135 μA steigerbar, was einer Stromdichte von 270 μA/cm² über die Extraktionsfläche von rund 0,5 cm² entspricht. Dies steht im Einklang mit der Prozessparameterabhängigkeit der anhand der Sonde bestimmten Plasmadichte, welche einen größtmöglichen Wert von etwa 6E11 1/cm³ annimmt. Das Verhältnis von extrahiertem Al+- zu Ar+-Ionenstrom kann durch Optimierung der Prozessparameter von 0,3 auf maximal 2 angehoben werden. Sondenmessungen des entsprechenden Ionendichteverhältnisses bestätigen diesen Sachverhalt. Um möglichst große Extraktionsströme und Al+/Ar+-Verhältnisse zu generieren, muss die ECR-Fläche demnach in dem Bereich der höchsten Al-Atomdichte in der Targetebene lokalisiert sein. Gegenüber dem alleinigen Magnetronplasma (ohne Mikrowelleneinspeisung) können mit dem MECRIS-Plasma um bis zu 140 % höhere Al+-Ionenströme produziert werden. Aus Sondenuntersuchungen geht hervor, dass dies eine Folge der um etwa eine Größenordnung gesteigerten Plasmadichte und der um rund 7 eV größeren Elektronentemperatur des MECRIS-Plasmas ist. Das MECRIS-Plasma wurde außerdem mittels optischer Emissionsspektroskopie charakterisiert und durch ein globales sowie ein zweidimensionales Modell simuliert. Die gewonnenen Prozessparameterabhängigkeiten der Plasmadichte, Elektronentemperatur sowie Al+- und Ar+-Ionendichte stimmen mit den Sondenergebnissen überein. Teilweise treten jedoch Absolutwertunterschiede von bis zu zwei Größenordnungen auf. Die Erhöhung der Sputterleistung und Extraktionsspannung über die derzeitigen Grenzen von 10 kW bzw. 30 kV sowie die Optimierung der Extraktionseinheit hinsichtlich minimaler Elektrodenblindströme bietet das Potential, den Al+-Ionenstrom bis in den mA-Bereich zu steigern. / An electron cyclotron resonance ion source working at a microwave frequency of 2.45 GHz has been developed in order to generate an intense current of singly charged metal ions. It is loaded with metal vapor by an integrated cylindrical sputter magnetron, which was especially designed for this purpose. The MECRIS (Magnetron Electron Cyclotron Resonance Ion Source) merges ECR ion source technology with sputter magnetron technology in a unique manner representing a new metal ion source concept. By using an Al sputter target, the efficiency of the MECRIS was demonstrated successfully for the example of Al+ ion production. The extractable ion current was measured by a newly developed high-current Faraday cup. On the basis of numerical modeling, the total magnetic field was set in a way that the permanent magnets of the magnetron and the coils of the ECR source are forming a minimum-B-structure, providing an effective electron trap by the magnetic mirror principle. Simultaneously, optimal electron heating is achieved by a closed ECR-surface at resonant magnetic flux density of 87.5 mT. Electron temperature increases towards the center of the source to a maximum of about 11 eV and was measured by a double Langmuir probe. Due to the heated electron population, efficient electron impact ionization of the Al atoms is accomplished. Al atoms are injected with a rate of more than 1E18 Al-atoms/s resulting in a maximum Al atom density of 2E10 1/cm³. The MECRIS produces mainly singly charged ions of the sputtered material (Al+) and the process gas (Ar+). The Al+ ion extraction current is elevated by one order of magnitude to a maximum of 135 μA by increasing the process parameters sputter magnetron power, microwave power, coil current, and acceleration voltage. Related to the extraction area of about 0.5 cm², the highest possible Al+ ion current density is 270 μA/cm². A corresponding process parameter dependency was found for the plasma density showing a peak value of about 6E11 1/cm³, which was deduced from probe measurements. The ratio of the extracted Al+ ion current to the Ar+ ion current can be enhanced from 0.3 to a maximum of 2 by optimization of the process parameters. This was confirmed by probe investigations of the appropriate ion density ratio. In conclusion, the ECR-surface needs to be located in the area of the highest Al atom density in the target plane in order to improve the extraction current and Al+/Ar+ ratio. The MECRIS plasma produces an Al+ ion current, which is up to 140 % higher compared to that of the sole sputter magnetron plasma (without microwave injection). As revealed by probe measurements, this effect is due to the higher plasma density and electron temperature of the MECRIS plasma, leading to a difference of one order of magnitude and 7 eV, respectively. Additionally, the MECRIS plasma has been characterized by optical emission spectroscopy and simulated by a global and a two-dimensional model. Retrieved process parameter dependencies of plasma density, electron temperature, Al+ ion density, and Ar+ ion density coincide with probe findings. Although a discrepancy of the absolute values of partly up to two orders of magnitude is evident. Potentially, the Al+ ion current can be enhanced to the mA-region by optimizing the ion extraction system for minimal idle electrode currents and by rising sputter magnetron power as well as acceleration voltage above the actual limits of 10 kW and 30 kV, respectively.

Harte amorphe wasserstoffhaltige Kohlenstoffschichten (a-C:H) haben in den letzten Jahrzehnten stark an Bedeutung gewonnen. Diese Art von Hartstoffschichten wird zunehmend für die Reduzierung von Reibung und Verschleiß in unterschiedlichen Bereichen eingesetzt. In der Forschung, aber auch für Kleinserien, werden a-C:H-Schichten üblicherweise mit Hochfrequenzplasmaentladungen abgeschieden. Eine Alternative ist die Plasmaaktivierung mit einer asymmetrisch bipolar gepulsten Spannung im Mittelfrequenzbereich. Auf diese Weise wird eine homogene Beschichtung großer Substratflächen mit qualitativ hochwertigen Schichten ermöglicht. Die vorliegende Arbeit beschäftigt sich mit der plasmagestützten Abscheidung von harten a-C:H-Schichten mit mittelfrequenzgepulsten Entladungen. Zur Schichtabscheidung werden Ethin-Argon- und Isobuten-Argon-Gasgemische verwendet. Der Einfluss des Prozessdrucks auf den Abscheideprozess und die Schichteigenschaften wird untersucht. Dazu wurden Argonentladungen und Beschichtungsplasmen mittels optischer Emissionsspektroskopie charakterisiert. Zur Charakterisierung der Schichteigenschaften wurden unter anderem Nanoindentation-Messungen, elastische Rückstreudetektionsanalysen und thermische Desorptionsspektroskopie verwendet. Zur Untersuchung des Einflusses der Ionen auf das Schichtwachstum wird ein Modell zur Identifizierung von Ionenspezies in Beschichtungsplasmen vorgestellt. In Verbindung mit der Messung der Substratströme konnte der Ionenanteil am Schichtwachstum bestimmt werden. Ein weiterer Teil der vorliegenden Arbeit untersucht ein Hybridverfahren, in dem die mittelfrequenzgepulste Entladung mit einer zusätzlichen ECR-Entladung kombiniert wird. Es wird gezeigt, dass durch dieses Hybridverfahren eine deutliche Steigerung der Abscheiderate harter a-C:H-Schichten erreicht werden kann. Die abgeschiedenen Schichten wurden zusätzlich bezüglich ihrer Oberflächenstruktur und ihrer Verschleißfestigkeit untersucht.

Harte amorphe wasserstoffhaltige Kohlenstoffschichten (a-C:H) haben in den letzten Jahrzehnten stark an Bedeutung gewonnen. Diese Art von Hartstoffschichten wird zunehmend für die Reduzierung von Reibung und Verschleiß in unterschiedlichen Bereichen eingesetzt. In der Forschung, aber auch für Kleinserien, werden a-C:H-Schichten üblicherweise mit Hochfrequenzplasmaentladungen abgeschieden. Eine Alternative ist die Plasmaaktivierung mit einer asymmetrisch bipolar gepulsten Spannung im Mittelfrequenzbereich. Auf diese Weise wird eine homogene Beschichtung großer Substratflächen mit qualitativ hochwertigen Schichten ermöglicht. Die vorliegende Arbeit beschäftigt sich mit der plasmagestützten Abscheidung von harten a-C:H-Schichten mit mittelfrequenzgepulsten Entladungen. Zur Schichtabscheidung werden Ethin-Argon- und Isobuten-Argon-Gasgemische verwendet. Der Einfluss des Prozessdrucks auf den Abscheideprozess und die Schichteigenschaften wird untersucht. Dazu wurden Argonentladungen und Beschichtungsplasmen mittels optischer Emissionsspektroskopie charakterisiert. Zur Charakterisierung der Schichteigenschaften wurden unter anderem Nanoindentation-Messungen, elastische Rückstreudetektionsanalysen und thermische Desorptionsspektroskopie verwendet. Zur Untersuchung des Einflusses der Ionen auf das Schichtwachstum wird ein Modell zur Identifizierung von Ionenspezies in Beschichtungsplasmen vorgestellt. In Verbindung mit der Messung der Substratströme konnte der Ionenanteil am Schichtwachstum bestimmt werden. Ein weiterer Teil der vorliegenden Arbeit untersucht ein Hybridverfahren, in dem die mittelfrequenzgepulste Entladung mit einer zusätzlichen ECR-Entladung kombiniert wird. Es wird gezeigt, dass durch dieses Hybridverfahren eine deutliche Steigerung der Abscheiderate harter a-C:H-Schichten erreicht werden kann. Die abgeschiedenen Schichten wurden zusätzlich bezüglich ihrer Oberflächenstruktur und ihrer Verschleißfestigkeit untersucht.