Introduction The Los Alamos Isotope Production Facility (IPF) is actively engaged in the development of isotope production technologies that can utilize its 100 MeV proton beam. Characterization of the proton beam energy and current is vital for optimizing isotope production and accurately conducting research at the IPF. Motivation In order to monitor beam intensity during research irradiations, aluminum foils are interspersed in experimental stacks. A theoretical yield of 22Na from 27Al(p,x)22Na reactions is cal-culated using MCNP6 (Monte Carlo N-Particle), TRIM (Transport of Ions in Matter), and Andersen & Ziegler (A&Z) [1] computational models. For some recent experiments, experimentally measured activities did not match computational predictions. This discrepancy motivated further experimental investigations including a direct time-of-flight measurement of the proton beam energy upstream of the target stack. The Isotope Production Program now tracks the beam energy and current by a complement of experimental and computational methods described below. Material and Methods A stacked-foil activation technique, utilizing aluminum monitor foils [2] in conjunction with a direct time-of-flight measurement helps define the current and energy of the proton beam. Theoretical yields of 22Na activity generated in the Al monitor foils are compared with experimental measurements. Additionally, MCNP, TRIM, and A&Z computational simulations are compared with one another and with experimental data. Experimental Approach Thin foils (0.254mm) of high purity aluminum are encapsulated in kapton tape and stacked with Tb foils in between aluminum degraders. Following irradiation, the Al foils are assayed using γ-spectroscopy on calibrated HPGe detectors in the Chemistry Division countroom at LANL. We use the well-characterized 27Al(p,x)22Na energy dependent production cross section [3] to calculate a predicted yield of 22Na in each foil. Details of the experimental activity determination and associated uncertainties have been addressed previously [4]. The nominally stated beam parameters are 100 MeV and 100–120 nA for the foil stack irradiation experiments. Time-of-flight measurements performed in the month of January 2014 revealed beam energy of 99.1 ± 0.5 MeV. Computational Simulations Andersen & Zeigler (A&Z) is a deterministic method and also the simplest of the three com-putational methods considered. While the mean energy degradation can be calculated using the A&Z formalism, the beam current attenuation cannot. Consequentially, A&Z will also lack the ability to account for a broadening in the beam energy that a stochastic method affords. Additionally, A&Z does not account for nuclear recoil or contributions from secondary interactions. TRIM uses a stochastic based method to calculate the stopping range of incident particles applying Bethe-Block formalisms. TRIM, like A&Z, does not include contributions from nuclear recoil or contributions from secondary interactions. Computationally, TRIM is a very expensive code to run. TRIM is able to calculate a broadening in the energy of the beam; however, beam attenuation predictions are much less reliable. TRIM determines the overall beam attenuation in the whole stack to be less than one percent, whereas 7–10 % is expected. MCNP6 is arguably the most sophisticated approach to modeling the physics of the experiment. It also uses a stochastic procedure for calculation, adopting the Cascade-Exciton Model (CEM03) to track particles. The physics card is enabled in the MCNP input to track light ion recoils. Contributions from neutron and proton secondary particle interactions are included, although their contribution is minimal. For both MCNP and TRIM, the proton beam is simulated as a pencil beam. To find the current, an F4 volumetric tally of proton flux from MCNP simulation is matched to the experimental current for the first foil in the stack. Subsequent foil currents are calculated relative to the first foil based on MCNP predictions for beam attenuation. The equation used for calculating the current from the experi-mental activity is [5]: where: is the cross section for the process, [mbarns] is the atomic mass of the target [amu] is the is the number of product nuclei pre-sent at End-of-Bombardment is the average beam current, [μA] is the density of the target material, [g/cc] is the target thickness, [cm] is the decay constant, [s−1] is the irradiation time, [s] For each foil in the experimental stack, we also have a statistically driven broadening of the incident energy. The beam energy is modeled as a Gaussian distribution, with the tallies for each energy bin determining the parameters of the fit. TABLE 1 and FIG. 3 summarize the mean energy and standard deviation of the energy for each aluminum monitor foil. To address the energy distribution, we calculate an effective or weighted cross-section. It is especially important to account for energy broadening in regions where the associated excitation function varies rapidly. In the excitation function, we see a strong variation in the energy range from 30–65 MeV, the energy region cov-ered by the last 3 foils in the stack. Cross section weighting also accounts for the mean energy variation within each foil. The excitation function will overlay the Gaussian shaped flux distribution, giving rise to a lateral distribution where incrementally weighted values of the cross section are determined by the flux tally of the corresponding energy bin. With the effective cross section and the current at each of the foils, it is straight-forward to calculate the number of 22Na atoms created and the activity of each foil using the previously stated equation. Results and Conclusion The general trend in the amount of activity produced follows the shape of the excitation func-tion for the 27Al(p,x)22Na reaction. Small shifts in the incident energy upstream trickle down to produce much more pronounced shifts in the energy range of foils towards the back of the foil stack. The characteristic “rolling over” of the activity seen in the experimental foils indicates that the 6th foil must be in the energy region below 45 MeV, where the peak of the excitation function occurs. Conservatively, computational simulations are able to accurately determine the proton beam’s energy for an energy range from 100 to 50 MeV. As the beam degrades below 50 MeV, computa-tional simulations diverge from experimentally observed energies by over-predicting the energy. This observation has been noted in past studies [6,7] that compare the stacked foil technique with stopping-power based calculations. A complement of experimental and computational predictions allows for energy determinations at several points within target stacks. While this study focuses on an Al-Tb foil stack, the analysis of a similar Al-Th foil stack resulted in the same conclusions. Although we do not have a concurrent time-of-flight energy measurement at the time of the foil stack experiments, it is reasonable to assume that the energy at the time of the stacked foil experiments was also lower than the assumed energy of 100 MeV. Computational simulations developed in this work firmly support this assumption. Various computational models are able to predict with good agreement the energy as a function of depth for complex foil stack geometries. Their predictions diverge as the beam energy distribution broadens and statistical uncertainties propagate. A careful inspection of the codes reveals that these discrepancies likely originate from minute differences between the cross sections and stopping power tables that MCNP and TRIM/A&Z use respectively.

Introduction The increasing demand for [18F]FDG for clinical PET-CT and the efficiencies associated with large production runs have encouraged endeavors to increase the amount of 18F− produced by cyclotrons in a single run. The amount of 18F− is determined by the saturation yield of the nuclear reaction, the irradiation time and the beam current striking the target. The saturation yield is a function of beam energy (typically fixed for PET cyclotrons), the enrichment of the H218O (typically > 97 %) and the efficiency of the target design. Target design has already been optimized on current systems. Diminishing gains in activity are achieved by extending the irradiation time much beyond 3 hrs, so the main focus has been to increase beam current onto the targets. Increasing the beam current requires: i) a cyclotron capable of producing the increased beam current; ii) targets that tolerate the beam current without appreciable loss in saturation yield; iii) sufficient shielding of the cyclotron and hot cells to accommodate the proportionally larger radiation dose rates during higher current irradiation and from the larger activities delivered to the hot cells. We reported [1] that the self-shielded targets fitted to our cyclotron can accommodate 100 µA currents without appreciable loss in saturation yield. We also identified the potential of routine production at 200 A (100 A per target in dual target irradiation mode), but had not establish its long-term viability in routine use. We present our experience in using 200 µA for routine production of 18F- since September 2012. Material and Methods Our PETtrace cyclotron was installed in 2002 and has been used for routine production of various 18F and 11C tracers since January 2003. It has been upgraded incrementally so that it is now equivalent to a current generation PETtrace 880 cyclotron, which is specified at a total beam current of 130 µA. The only components on our cyclotron currently not part of the standard PETtrace 880 cyclotron configuration are the self-shielded targets and a license which allows total target beam current of 200 µA. The self-shielded targets utilize a W/Cu alloy for the main body of the target surrounding the Havar foil to provide shielding from the Havar foil by a factor of about 10 and shielding of any remnant 18F- activity in the targets by a factor of about 100 [1]. The niobium target chamber is the same size as used in the standard GE Nb25 targets. However, it dispenses with the He cooling and the vacuum foil. Only the water foil is used, which is directly exposed to the vacuum in the chamber. Foil cooling is through the water in the target chamber. One of the issues that we previously identified [1] is beam stripping by gas molecules in the vacuum tank. The amount of beam that is stripped and which impacts on components in the cyclotron is proportional to the beam current. At high currents, this can result in a runaway condition, where the effects of the stripped beam deteriorate vacuum; this then results in more beam stripping and more severe effects. The effect of diffusion pump maintenance on vacuum system performance and on the reduction of beam stripping was investigated as part of this study. We have previously found that running the ion source gas at a low flow rate (2 sccm) when cyclotron is not used greatly reduces deterioration of ion source performance over time and with use [1]. This gas flow also appears to have a beneficial effect on the vacuum. Ion source gas flow when cyclotron is off has been employed throughout the evaluation period. [18F]FDG was produced with TRACERlab MXFDG modules or FASTlab modules using both Phosphate and Citrate cassettes. Stability studies of [18F]FDG were performed to ensure it met specifications over the specified expiry time. Our current stabilization regime did not have to be adjusted for the higher activities produced with the higher beam currents. [18F]FDG yields were calculated using input activity estimates from saturation yield and beam time and current and the non-decay corrected [18F]FDG activity measured at the end of synthesis. Thus yield calculations include target yield variations and losses in the transfer lines and not just synthesis yield. Results and Conclusion The flip-in probe to extraction foil transmissions as a function of ion source gas flow are given in TABLE 1. Transmission decreases with increasing ion source gas flow, as expected for a system with an internal ion source. In addition, diffusion pump maintenance had a positive impact on the transmission and this is of particular benefit at the higher beam currents where minimising beam stripping becomes more critical. The ion source output, however, decreases with decreasing ion source gas flow; hence ion source gas flow is a compromise between ion source output and probe to foil transmission. We currently use a gas flow of 5.5 sccm for our 200 µA runs. Over the period from 1st September 2012 to end of March 2014, a total of 419 [18F]FDG produc-tions were performed at total target beam currents ranging from 160 µA to 200 µA, with 227 production runs being performed at 200 µA. Beam times were typically 90 to 120 min, with some productions up to 180 min. The [18F]FDG yields are summarized in TABLE 2. The yields for the FASTlab phosphate and citrate cassettes have been listed separately in TABLE 2 as they are known to be different [2,3]. The yields obtained with the TRACERlab MXFDG are also shown. The yields at 200 µA total target current are not appreciably different from those at < 200 µA current, irrespective of the synthesis method. Consistency of yield is also not adversely impacted by the higher beam current. For a 180 min, 200 µA test production, the [18F]FDG activity produced using the FASTlab phosphate cassette was 763 GBq (20.6 Ci). Clinical productions with the FASTlab phosphate were limited to 130 min maximum beam time for 200 µA and achieved a maximum [18F]FDG activity of 656 GBq (17.7 Ci). The tolerance to a reduction in performance of the critical components to achieve high current operation (RF, ion source output and vacuum system) is reduced at high beam currents. The requirements for routine maintenance of ion source, targets and extraction system, however, have not increased with the increase in beam current from 160 µA to 200 µA. Extraction foil life and ion source maintenance intervals have remained at about 2000 Ah and >120 µAh, respectively. As more experience has been gained with the self-shielded targets, service interval is actually being extended from about 10,000 µAh to 20,000 µAh, despite the higher beam currents. Diffusion pump maintenance is currently recommended every 5 years, but a 2 year maintenance interval may be advantageous for 200 µA, given the observed deterioration over a 5 year period and the improvement in performance post service (Table 1). The more frequent service is associated with the additional costs of diffusion pump oil and an extra day of scheduled down-time. Typically, vacuum is sufficiently well established 24 h after opening of the vacuum tank to run 200 µA beams with the vacuum and beam conditioning that we employ. The targets generally have coped well with the 100 µA per target current (200 µA total beam current for dual target irradiation) over this 18 month period. However, currents of 80 µA to µ100 A per target in dual target irradiation mode reduce the tolerance to sudden increases in one of the target currents. There were 4 occasions (2 test beams and 2 production beams) when there were sudden increases of target current from 90 µA and 100 µA to about 150 µA. The rapid increase in heat deposited on the foil and target chamber and the resultant rapid pressure rise in the target chamber could not be withstood by the foil and target foil rupture ensued. This compared to 1 target foil issue over a similar period of time (18 months) at lower beam currents on the standard Nb25 target. Three separate causes were identified for these overshoots in target current: 1) behavior of control system when beam is allowed to continue past the set time; 2) large changes of set current of one of the two targets irradiated during a dual irradiation test beam and 3) an issue with DEE voltage regulation caused by the mechanical flap controls. These issues have been addressed by procedural changes (issues 1 and 2) and by fitting an available upgrade of the mechanical flap control mechanism (issue 3). The two target foil ruptures during production did not cause cancellation or delays to patient scanning, as the demand could be met by multi-ple productions and deliveries from the unaf-fected target. No unscheduled down-days occurred during the evaluation period. We have been able to achieve routine operation at 200 µA beam current through careful optimization of the critical components and parameters and a maintenance regime that we have detailed previously [1]. This maintenance scheme has not changed for the routine 200 µA operation. The safety margin, however, is reduced and so careful monitoring of the system is required to ensure that issues in one of the subsystems do not cause major events such as target foil ruptures. Our [18F]FDG yields have been maintained at the higher current and 200 µA allows large quantities of [18F]FDG to be produced routinely in a single run with relatively short beam times.

Introduction Temperature modeling using Finite Element Analysis (FEA) is widely used by particle beam-line designers as a useful tool to determine the thermal performance of an irradiated target system. A comparison study was performed between FEA calculated temperatures on platinum with experimental results using direct thermocouple measurements. The aims are to determine the best beam model for future solid target design, determine the maximum target current for different target materials and the temperature tolerance for any modification to our existing solid targetry system. Material and Methods The theoretical temperature of the target sys-tem was determined using SolidWorks 2013 with Flow Simulation Analysis (FSA) module. The FSA module determines the maximum temperature inside the target material given the global conditions (material specification, flow rates, boundary conditions, etc.) for a given target current. The proton beam was modeled as a volumetric heat source inside the target material based on the distribution of energy loss in the material along the beam axis. The method used by Comor, et al1 was used in this study. The method segmented the target material into five individual layers, each layer being 50 m thick. The energy lost per layer was calculated using SRIM3 and converted into the power lost per layer. A thickness of 250 μm of platinum completely stops the impinging proton beam at 11.5 MeV with the highest deposition of power per layer corresponding to the Bragg peak. The target material used in the simulation reflects the physical target disk used for temperature measurements (platinum, dia. 25.0 mm, thickness 2.0 mm) with two K-type thermocouples (dia. 0.5 mm, stainless steel sheath) embedded in the platinum disk. One thermocouple is located in the geometric center, while the other is located at a radial position 8 mm from center. The outer thermocouple is to determine the peripheral temperature near the o-ring seal. Temperature was maintained below the melting point for the material (Viton®, melting point 220 °C) during the irradiation to ensure the integrity of the water cooling system. The solid targetry system used in this study is an in-house built, significantly modified version2 of a published design1. The solid target system is mounted onto an 18/18MeV IBA Cyclotron with dual ion source, on a 300mm beam-line with no internal optics or steering magnets. A graphite collimator reduces the beam to 10mm in diameter and a degrader is used to reduce the proton beam energy to 11.5 MeV, considered suitable for production of radiometal PET isotopes 89Zr and 64Cu. Temperature was measured with and without the 300 mm beam-line to compare the effects of beam divergence on the solid target (FIGS. 1 and 2). The experiment was conducted using both H− ion sources with different ion-to-puller extraction gaps (ion source 1 is 1.55 mm with ion source 2 at 1.90 mm). The setting of the ion-to-puller gap changes the focusing of the accelerated beam inside the cavity. Results and Discussion The segmented beam model was used to calculate the temperature on and within the target, as well as the maximum temperature of the bulk material. The first segment is the leading segment of the material irradiated by the incident proton beam. Results are shown in TABLE 2. Target temperatures were measured experimentally under two different conditions; target attached at the end of a 300mm beam-line and target attached directly to the cyclotron. The temperature was measured experimentally using the platinum disk with 2 thermocouples inside the bulk target material irradiated on the end of a 300mm beam-line. The measured temperature is shown in TABLE 2. The variation between ion source 1 and 2 for the temperature measured in the center was 11–15 %, while the variation on the radial position was 2–6 %. A smaller ion-to-puller extraction distance (ion source 1) reduces the cross-sectional area of the accelerated beam; the consequent high proton current density (10mm diameter collimated beam) increases the temperature inside the bulk material for a fixed target current. The highest observed radial temperature was 93 °C, with target current of 50 μA using ion source 1. This is well below the melting point for the o-ring seal. The temperature measured experimentally using the same platinum disk with no beam-line is shown in TABLE 4. A temperature difference of up to 7 % was measured between ion source 1 and 2 at the exit port without the beam-line, while the maximum variation on the radial position was 3 %. A comparison between the calculated theoretical and measured temperatures is shown in FIGS. 3 to 6. The temperatures calculated by the FEA model underestimate the temperature regardless of target position (with or without the beam-line) and for both ion sources. The temperature difference between the FEA model and the experimental results increases with increasing target currents. As shown in Figure 3, at the target center the FEA model underestimated the temperature by 22–32 % for ion source 1 and 13–22 % for ion source 2. This is consistent with the difference between the two ion sources due to the difference in the ion-to-puller gap size. With the target mounted at the exit port the theoretical and measured temperature for the center of the platinum disk is shown in FIGURE 4. The FEA model underestimates the temperature at the center of the platinum disc by 2–10 % for both ion sources. As shown with the previous experiment, the margin of error increases with increasing target current. Comparison between FIGS. 3 and 4 shows the measured temperature at the center of the platinum disk is significantly lower when the target is attached to exit port of the cyclotron. Localised area of high current density (hot spots) is not registered as higher temperature in the bulk material. True temperature inside the bulk material is highly dependent on the thermal conductivity of the target material and the resolution of the thermocouple. The cross-sectional area of the beam ‘hot-spot’ will be greater due to beam divergence at the end of the beam line compared with the exit port. The ‘hot’ area of the expanded beam becomes a significant portion of the overall collimated beam (collimator dia. 10.0 mm). A more uniform beam profile (less heterogeneity) evenly distributed the area of high current density across the disk surface, effectively increasing the temperature of the bulk material while decreasing the sensitivity required to measure the true temperature. As observed from this comparative study it appears that a more homogeneous current density leads to a higher temperature measurement at the target center. With the solid target at the end of the beam-line, target current lost on the collimator and beam-line was >55%. The effect of beam divergence is clearly observed in TABLE 5. With the target mounted directly at the exit port the current lost was reduced to < 40 %. Although the average proton current density is the same for any set target current, irrespective of target position, the contribution of the peripheral beam to the total target current should not be underestimated. A loss of ~40 μA on the collimator and beam-line places greater reliance on the center of the ‘hot’ beam to maintain the same target current. The temperature at the radial position (FIG. 5) observes the same trend as for the temperature measured in the center. The error increases for higher target currents and the FEA model underestimated the temperature by 19–40 %. The error at this location is due partly to the model’s assumption of a uniform heat source, applied to the material on a single axis (perpendicular to the material surface) and does not account for any scattering or divergence of the incident proton beam. FIGURE 6 shows that the FEA model underestimated the radial temperature by 16–37 %, when the target is connected to the exit port, for reasons discussed previously. Comparison with FIG. 5 (target on the beam-line) shows the same margin of error between the FEA and the experimental results (19–40 %). The temperature difference between the FEA model and measured temperature at the radial position is independent of the beam profile and beam divergence. The FEA model underestimated the temperature at the radial location with or without the beam-line and for both ion sources. The significance difference in temperature between the FEA model and the experimental is due to our model assumption that the maximum radial temperature is on the irradiated surface and not inside the material corresponding to the layer with the maximum energy lost. In addition, the FEA model does not ac-count for the divergence of the proton beam as it travels through the material. Given the temperature at 50 μA target current is > 90 °C (TABLES 3 and 4) we have capped the experi-ment below this point to prevent any damage the o-ring seal. Conclusion The segmented FEA model was inadequate in determining the temperature for the target at the end of a 300mm beam-line (> 30 % difference). A combination of beam divergence and greater uniform coverage of high current density beam resulted in a higher than predicted temperature reading. However, the segmented FEA model provides a good estimation (< 10 % difference) for the observed temperature of the bulk material at the exit port. The simplistic FEA model was unable estimate the temperature at the radial position (~ 40 % difference) regardless of ion source or target position. A comparison between the two ion sources with different ion-to-puller extraction gap, leading to different focusing of the accelerated beam yield minimal temperature difference. Although a 15% difference was observed between the ion sources at the end of the beam-line, a major contributing factor is beam divergence beyond the magnetic field rather than the beam size of the accelerated beam. Further studies are underway to determine the beam profile (quantitatively using radiographic film), quantify the contribution of the peripheral beam to the total beam current by comparing different size collimators and to investigate other FEA models by applying different beam models (heterogeneous and homogeneous beam) and different heat sources (surface vs. volumetric). Currently the RAPID Lab solid targetry is placed at the end of the beam-line for easy loading and unloading, since multiple target irradiations are performed per month2. However, RAPID is presently developing a new solid targetry sys-tem which eliminates the need for a beam-line and will be able to manage a maximum extracted target current of 150 μA.

Introduction Annealed Inconel 718 alloy was chosen for the beam window at the Los Alamos Neutron Science Center (LANSCE) Isotope Production Facility (IPF) [1]. The window was replaced after 5 years of operation. Mechanical properties and microstructure changes were measured to assess its expected lifetime. Material and Methods A cutting plan was developed based on the IPF rasterred beam profile (FIG. 1). 3-mm OD samples were cut out from the window and thinned to 0.25-mm thick. Shear punch tests were per-formed at 25 °C on 21 samples to quantify shear yield, ultimate shear stress, and ductility. From 1-mm OD, 0.25-mm thick shear punched out disks, 4 TEM specimens of ~30×10×2 μm were obtained using standard FIB lift-out techniques. TEM was performed on an FEI Tecnai TF30-FEG operating at 300 kV. Results and Conclusions TABLE 1 shows MCNPX tally results of accumulated dpa, He and H content from both protons and neutrons fluences and ANSYS steady-state irradiation temperature for the 3-mm OD samples [2]. These peak values are at the peak density of Typically increases in shear yield and shear maximum stress occur with increasing dose. In this case, highest shear yield and ultimate stress was on the lowest dose samples at the outer edge (FIG. 2). Optical microscopy images of the fracture surfaces on the shear punched out disks show no significant change in the fracture mode or reduction in ductility in the un-irradiated, high and low dose irradiated samples. One un-irradiated and 4 irradiated samples (5, E, 16 and 19) were selected for TEM analysis. Figure 3 shows bright field TEM images of an un-irradiated, high and low dose irradiated samples. Un-irradiated sample shows some dislocations and some large precipitates. The high dose sample #5 (~11 dpa, 122 oC) shows small loops and dislocations (left and center images) and no γ\' or γ\'\' precipitates in SAD from z = [011] (right image). Low dose sample #19 (~0.7 dpa, 40 oC) shows a high density of dislocation loops (left image), high density of H/He bubbles (center image) and presence of γ\'\' precipitates in SAD from z = [011] (right image). Radiation induced-hardening is highest at the low dose region in the outer most edge. The hardening from γ\'\' precipitates is determined to be more pronounced than that from trapped bubbles. The lack of significant hardening in the highest dose region is attributed to a lower dis-location density and no γ” precipitates or bubbles [3]. Identification of H or He bubbles and the higher accumulation of these bubbles in the low dose region (no direct beam hitting) warrant further studies. Despite the evidence of irradiation-induced hardening, this spent beam window appears to retain useful ductility after 5 years in service. At the conclusion of 2013 run cycle, the current in-service beam window had reached the same dpa as of the spent window. We plan to extend the service of the current in-service window until it reaches its intended design threshold limit of ~20 dpa (in the highest dose region). Additional measurements at higher dpa values will enable better decision-making in managing risks of the window failure.

In-beam Positron Emission Tomography (PET) is a system for monitoring high precision radiation therapy which is in the most cases applied to the tumors near organs at risk. High quality and fast availability of in-beam PET images are, therefore, extremely important for successful verification of the dose delivery. Two main problems make an in-beam PET monitoring a challenging task. Firstly, in-beam PET measurements result in a very low counting statistics. Secondly, an integration of the PET scanner into the treatment facility requires significant reduction of the sensitive surface of the scanner and leads to a dual-head form resulting in imaging artifacts. The aim of this work is to bring the imaging process by means of in-beam PET to optimum quality and time scale. The following topics are under consideration: - analysis of image quality for in-beam PET; - image reconstruction; - solutions for building, testing, and integration of a PET monitoring system into the dedicated treatment facility.

Entladungsbasierte Elektronenquellen mit Kaltkathode waren gegen Ende des 19. Jahr hunderts weithin genutzte Forschungswerkzeuge und ermöglichten die Entdeckung des Elektrons und der Röntgenstrahlung. In jüngster Zeit erfahren sie erneutes Interesse in Wissenschaft und Industrie, motiviert durch ihre Fähigkeit, Elektronenstrahlen hoher Leistung für Produktionsprozesse (wie das Schweißen, die Materialverdampfung in der Vakuum beschichtung oder die Vakuum-Schmelzveredlung in der Metallurgie) basierend auf einem robusten Design sowie einfachen Versorgungs- und Steuerungssystemen zu erzeugen. Entladungsbasierte Elektronenquellen könnten also eine wirtschaftlich attraktive Alternative zu den gegenwärtig noch etablierten Elektronenstrahlkanonen mit Glühkathoden bieten. Trotz der langen Geschichte und vieler empirischer Ansätze, Gasentladungen zur Elektronenstrahlerzeugung für diverse Anwendungen zu nutzen, sind die bestimmenden Mechanismen bei dieser Art von Elektronenquellen immer noch unzulänglich verstanden. Es war deshalb das Ziel der für die vorliegende Dissertation durchgeführten experimentellen und theoretischen Arbeiten, nicht nur die technologischen Potentiale und Limitierungen entladungsbasierter Elektronenstrahlquellen zu untersuchen, sondern auch die Kenntnis grundlegender physikalischer Effekte zu verbessern. Analysiert wurden zunächst verschiedene, im Fraunhofer FEP vorhandene Kaltkathoden-Strahlquellen, die - ungeachtet der Tatsache, dass sie für unterschiedliche Anwendungen konstruiert wurden - sämtlich auf demselben Funktionsprinzip beruhen: Innerhalb des Gerätes wird eine Hochspannungs-Glimmentladung (HSGE) unterhalten. Ionen erfahren im Kathodenfall einen Energiezuwachs, treffen auf die Kathode und setzen dort Sekundärelektronen frei. Diese Elektronen werden in Richtung des Plasmas be schleu nigt und verlassen schließlich die Strahlquelle, um am Prozessort die beabsichtigte Wirkung zu erzielen. Zur Optimierung der Stabilität der die Ionen produzierenden Entladung, der Effizienz der Strahlerzeugung sowie der Strahlleistungsdichte und Kathodenlebensdauer wurden verschiedene Kombinationen von Kathodenmaterialien und Plasma-Arbeitsgasen experimentell untersucht. Die Abhängigkeit der Ausdehnung des Kathodenfalls von Strom und Spannung der Entladung wurde gemessen und konnte durch ein analytisches Modell erklärt werden. Emittanz und Richtstrahlwert sind wichtige Kenngrößen zur Charakterisierung der Qualität von Elektronenstrahlen. Beide wurden in dieser Arbeit für den Elektronenstrahl einer HSGE-basierten Kaltkathoden-Schweißstrahlquelle bestimmt, wobei zwei Ansätze verfolgt wurden: Zum einen konnte die Emittanz aus der Randstrahlgleichung gewonnen werden, die den experimentell beobachteten Verlauf des Strahldurchmessers entlang der Ausbreitungsachse analytisch beschreibt. Zum anderen wurde die Emittanz anhand des aus der numerischen Simulation berechneten Phasenraumprofils ermittelt. Eine Kernaufgabe dieser Arbeit war es, Software-Werkzeuge zur Simulation der Strahl erzeugung in verschiedenen geometrischen Konfigurationen zu entwickeln und zu validieren, mit denen künftig die Konstruktion und Optimierung neuer entladungsbasierter Strahlerzeuger unterstützt werden sollte. Da kommerziell verfügbare Programme zur Simulation der Erzeugung und Führung von Elektronenstrahlen grundlegende Effekte plasma basierter Quellen, wie z. B. die Raumladung der Ionen oder die ioneninduzierte Sekundär elektronen-Freisetzung, nicht berücksichtigen, wurde für diese Arbeit eine neue Herangehensweise favorisiert: „Particle-in-Cell“ (PIC)-Algorithmen werden in der Plasma forschung üblicherweise zur Modellierung von Entladungen sowie zum Studium nichtlinearer Probleme, wie z. B. Instabilitäten, verwendet. Deshalb wurde nun eine PIC-Simulations umgebung zur Modellierung der HSGE und der damit verbundenen Strahlerzeugung entwickelt. Die Simulation reproduziert experimentelle Ergebnisse, wie etwa die Charakteristik der Entladung, die Emittanz des Strahls oder die Ausdehnung des Kathodendunkelraums, in befriedigender Weise. Schließlich wurde im Rahmen dieser Arbeit eine entladungsbasierte Elektronenstrahlquelle neuen Typs entwickelt und charakterisiert, die die Einfachheit der bekannten Kaltkathoden-Strahler und vorteilhafte Leistungsparameter, z. B. eine hohe Strahlleistungsdichte und niedrige Arc-Rate, wie sie bisher nur mit traditionellen Glühkathodenstrahlern erreichbar waren, in sich vereinigt. Die Kathode bestand aus LaB6 - einem Material, das sowohl eine hohe Sekundärelektronen-Ausbeute als auch eine niedrige Austrittsarbeit aufweist - und wurde gegen die Halterung thermisch isolierend montiert. Dadurch kann sie von Ionen aus einer HSGE auf hohe Betriebstemperaturen geheizt werden und in erheblichem Maße thermisch freigesetzte Elektronen emittieren. Neben technisch nützlichen Gebrauchs eigenschaften weist diese so genannte „Hybrid-Kathode“ auch ein physikalisch interessantes Verhalten auf. Einige neuartige Effekte, die von Entladungen mit kalten Kathoden nicht bekannt waren, konnten beobachtet und erklärt werden, wie z.B. die auffällige „N-förmige“ Druck-Strom-Charakteristik, die bei plötzlicher Abschaltung der Entladung nur langsam und ungleichmäßig abklingende Elektronenemission, die Limitierung des erreichbaren Strahl stromes und eine Fülle von Kathodenverschleiß-Mechanismen. Physikalische Modelle zur Beschreibung verschiedener Aspekte der Hybridkathoden-Entladung wurden erarbeitet und mit den experimentellen Befunden verglichen.:1 Einleitung 1.1 Hintergrund und Motivation 1.2 Aufgabenstellung und Gliederung 2 Grundlagen 2.1 Erzeugung freier Elektronen 2.1.1 Elektronenfreisetzung durch Glüheffekt und Feldemission 2.1.2 Elektronenfreisetzung durch Teilchenbeschuss 2.2 Hochspannungsglimmentladungen 2.3 Kathodenverschleiß 2.4 Hochspannungsüberschläge 2.5 Aspekte der Strahlphysik 2.5.1 Strahlgüte 2.5.2 Strahlinstabilitäten 3 Experimentelle Basis 3.1 Elektronenstrahlquellen 3.1.1 CCGD-5/30 3.1.2 CCDG-EXP 3.1.3 CCGD-60/30 3.1.4 CCGD-400/40 3.1.5 EasyBeam-60/40 3.2 Messmethoden 3.2.1 Messung von Strom und Spannung 3.2.2 Druck- und Gasflussmessung 3.2.3 Messung des Kathodendunkelraums 3.3 Strahlstromregelung 4 Experimente mit Kaltkathoden-Elektronenstrahlquellen 4.1 Vermessung und Modellierung der Ausdehnung des Kathodendunkelraums 4.2 Untersuchungen zur Effizienz der Strahlerzeugung 4.3 Untersuchungen zum Kathodenverschleiß 5 Untersuchung und Charakterisierung der Strahlqualität 5.1 Überblick über etablierte Messmethoden 5.2 Eigene Messmethode 5.2.1 Aufbau und Messprinzip 5.2.2 Datenauswertung 5.2.3 Bestimmung der Emittanz 5.2.4 Diskussion 6 Numerische Simulation von Kaltkathoden-Elektronenstrahlquellen 6.1 Grundlagen zur Simulation von Plasmen und Elektronenstrahlen 6.2 Literaturübersicht zur Simulation von Hochspannungsglimmentladungen 6.3 Plasmasimulation mit Particle-in-Cell-Programmen 6.3.1 Skalierungsregeln und Stabilitätskriterien 6.3.2 Eingesetztes Simulationsprogramm und implementierte Modelle 6.4 Vergleich von Simluationsergebnissen und Experimenten 6.5 Fazit 7 Strahlerzeugung mit einer entladungsgeheizten thermionischen Kathode 7.1 Motivation 7.2 Funktionsweise und Wahl der Materialien 7.3 Experimentelle Ergebnisse 7.4 Erarbeitung und Diskussion von Modellvorstellungen 8 Zusammenfassung und Ausblick A Häufig verwendete Abkürzungen und Symbole A.1 Abkürzungen und Indizes A.2 Symbole A.3 Konstanten Tabellenverzeichnis Abbildungsverzeichnis Literaturverzeichnis / Discharge-based, cold-cathode electron sources were routinely used as research tools at the end of the 19th century and facilitated then the discovery of the electron and of the x-rays. In recent time, they experience a renewed interest in science and industry due to their capability of generating high power electron beams for production processes (like welding, evaporation of materials for vapor deposition, and vacuum melt refining in metallurgy) relying on rugged mechanic designs as well as simple supply and control systems. Hence, discharge-based electron sources could provide an economically attractive alternative to the currently established electron beam guns with thermionic cathodes. Despite the long history and many empirical trials to utilize electron beam generation by gas discharges in several applications, the mechanisms governing this kind of electron sources are far from being well understood. Therefore, it was the purpose of the theoretical and experimental work performed for this thesis not only to investigate in the technological potentials and limitations of discharge-based electron beam guns but also to improve the knowledge of physical basic effects. At first, several cold-cathode beam sources existing at Fraunhofer FEP were analyzed. Regardless that they were designed for different applications, all were based on the same function principle: A high-voltage glow-discharge (HVGD) is sustained inside the device. Ions gain energy in the cathode fall, hit the cathode and release secondary electrons. These electrons will be accelerated towards the plasma then and can finally leave the beam source to perform the desired action at the process site. In order to optimize stability of the ions generating discharge, efficiency of the beam generation, beam power density and longevity of the cathode, different combinations of cathode materials and plasma forming gases have been investigated experimentally. The dependence of the cathode dark space width on current and discharge voltage was measured and could be explained by an analytic model. Emittance and brightness are important measures which quantify the quality of electron beams. In this work, both were determined for the beam originating from a HVGD based cold-cathode electron gun designed for welding following two approaches: First the emittance could be extracted from the envelope equation which analytically describes the evolution of the experimentally observed beam diameter along the propagation axis. Second the emittance was calculated from numerically simulated traces in the phase space. It was a core purpose of this work to develop and validate software tools capable of simulating the beam formation in various geometric configurations. This task was aimed at supporting the design and optimization of new discharge-based beam sources. Since commercially available software for modeling electron beam generation and transport do not consider the key mechanisms of plasma-based sources like the ion space charge or the ion-dependent production of free electrons, a new attempt was favored for this work: Particle-in-Cell (PIC) are being used in plasma research for studying nonlinear problems like instabilities. Therefore, a PIC simulation environment was utilized to numerically model the HVGD and the related beam generation. The simulation satisfactorily reproduces experimental findings, like the characteristics of the discharge, the emittance of the beam or the cathode dark space dimension. Finally, a discharge-based electron-beam sources of a new type was developed and characterized in the frame of this work. It merges the simplicity of known cold cathode devices with beneficial performance parameters, like high beam power density and low arcing rate, which have been reached so far with traditional thermionic electron sources only. The cathode of the new beam source consists of LaB6 - a material with a high secondary electron yield and a low thermionic work function - and was mounted thermally insulated against the holder. Then, an elevated operation temperature resulting in considerable thermionic emission was maintained by ions extracted from a HVGD. Besides to technically advantageous features, this so called “hybrid“ cathode mode of beam generation shows a physically interesting behaviour. Several new effects - not known from traditional cold-cathode discharges - could be observed, like a peculiar “N-shaped“ appearance of the pressure-current characteristic, the slowly and irregularly decreasing electron emission after a sudden discharge cutoff, a limitation of achievable beam current, and a multitude of possible cathode wear mechanisms. Physical models describing various features of the hybrid cathode discharge were elaborated and compared with the experimental findings.:1 Einleitung 1.1 Hintergrund und Motivation 1.2 Aufgabenstellung und Gliederung 2 Grundlagen 2.1 Erzeugung freier Elektronen 2.1.1 Elektronenfreisetzung durch Glüheffekt und Feldemission 2.1.2 Elektronenfreisetzung durch Teilchenbeschuss 2.2 Hochspannungsglimmentladungen 2.3 Kathodenverschleiß 2.4 Hochspannungsüberschläge 2.5 Aspekte der Strahlphysik 2.5.1 Strahlgüte 2.5.2 Strahlinstabilitäten 3 Experimentelle Basis 3.1 Elektronenstrahlquellen 3.1.1 CCGD-5/30 3.1.2 CCDG-EXP 3.1.3 CCGD-60/30 3.1.4 CCGD-400/40 3.1.5 EasyBeam-60/40 3.2 Messmethoden 3.2.1 Messung von Strom und Spannung 3.2.2 Druck- und Gasflussmessung 3.2.3 Messung des Kathodendunkelraums 3.3 Strahlstromregelung 4 Experimente mit Kaltkathoden-Elektronenstrahlquellen 4.1 Vermessung und Modellierung der Ausdehnung des Kathodendunkelraums 4.2 Untersuchungen zur Effizienz der Strahlerzeugung 4.3 Untersuchungen zum Kathodenverschleiß 5 Untersuchung und Charakterisierung der Strahlqualität 5.1 Überblick über etablierte Messmethoden 5.2 Eigene Messmethode 5.2.1 Aufbau und Messprinzip 5.2.2 Datenauswertung 5.2.3 Bestimmung der Emittanz 5.2.4 Diskussion 6 Numerische Simulation von Kaltkathoden-Elektronenstrahlquellen 6.1 Grundlagen zur Simulation von Plasmen und Elektronenstrahlen 6.2 Literaturübersicht zur Simulation von Hochspannungsglimmentladungen 6.3 Plasmasimulation mit Particle-in-Cell-Programmen 6.3.1 Skalierungsregeln und Stabilitätskriterien 6.3.2 Eingesetztes Simulationsprogramm und implementierte Modelle 6.4 Vergleich von Simluationsergebnissen und Experimenten 6.5 Fazit 7 Strahlerzeugung mit einer entladungsgeheizten thermionischen Kathode 7.1 Motivation 7.2 Funktionsweise und Wahl der Materialien 7.3 Experimentelle Ergebnisse 7.4 Erarbeitung und Diskussion von Modellvorstellungen 8 Zusammenfassung und Ausblick A Häufig verwendete Abkürzungen und Symbole A.1 Abkürzungen und Indizes A.2 Symbole A.3 Konstanten Tabellenverzeichnis Abbildungsverzeichnis Literaturverzeichnis

A recent method in aero engine production is to fabricate components from smaller pieces, rather than machining them from large castings. This has made laser beam welding popular, offering high precision with low heat input and distortion, but also high productivity. At the same time, the demand for automation of production has increased, to ensure high quality and consistent results. In turn, the need for sensors to monitor and control the laser welding process is increasing. In laser beam welding without filler material, the gap between the parts to be joined must be narrow. Optical sensors are often used to measure the gap, but with precise machining, it may become so narrow that it is difficult to detect, with the risk of welding in the wrong position. This kind of problems can cause severe welding defects, where the parts are only partially joined without any visible indication. This thesis proposes the use of an inductive sensor with coils on either side of the gap. Inducing currents into the metal, such a sensor can detect even gaps that are not visible. The new feature of the proposal is based on using the complex response of each coil separately to measure the distance and height on both sides of the gap, rather than an imbalance from the absolute voltage of each coil related to gap position. This extra information allows measurement of gap width and misalignment as well as position, and decreases the influence from gap misalignment to the position measurement. The sensor needs to be calibrated with a certain gap width and height alignment. In real use,these will vary, causing the sensor to be less accurate. Using initial estimates ofthe gap parameters from the basic sensor, a model of the response can be used to estimate the measurement error of each coil, which in turn can be used for compensation to improve the measurement of the gap properties.The properties of the new method have been examined experimentally, using a precise traverse mechanism to record single coil responses in a working range around a variable dimension gap, and then using these responses to simulate a two coil probe. In most cases errors in the measurement of weld gap position and dimensions are within 0.1 mm.The probe is designed to be mounted close to the parts to be welded, and will work in a range of about 1 mm to each side and height above the plates. This is an improvement over previous inductive sensors, that needed to be guided to the mid of the gap by a servo mechanism.

In this thesis six different models of IPE240 have been created in order to study their behavior undershear, bending and torsion. These models simulate IPE240 but differ in the boundary conditions, inthe loading and the length of the beam and in some connections which connect certain elements. Inthis study the modeling and simulation of the steel member is executed in ABAQUS Finite ElementAnalysis software with the creation of input files. When developing a model for the finite elementanalysis a typical analysis process is followed. All the parameters that are required to perform theanalysis are defined initially to geometry which is half the beam due to symmetry, and the materialproperties of each model are defined too. Then a mesh is generated for each model, the loads of eachmodel are applied which are expressed as initial displacement. Subsequently, the boundary conditionsfor each model are defined and finally the model is submitted to the solver when the kind of analysishas been defined. Namely, the analysis which is performed in this thesis is static stress analysis.When the ABAQUS has run the models, the contour plots for the von Mises stresses for each modelare studied. In these contour plots, a large concentration of stresses and problems which arise in eachone of the models are notified. As it has been observed in all models, the beam yields at the flangesof the mid-span and collapses at the mid-span. Therefore, the failure at the mid-span is more criticalthan the failure at the support. Moreover, the beams are weak in bending due to the fact that theytwist almost 60-90 degrees under a large initial displacement at the control node. Additionally, much localized failure and buckling occurred at the mid-span, and local concentrated stresses also occurredat the bottom flange at the support due to the boundary conditions details.Thereafter, a verification of the results of the ABAQUS through the simple analytical handcalculations is performed. It is concluded that the error appearing in most selected points is small.However, in some points in the web of the mid-span the error is greater. Additionally, whilecomparing the load-displacement curves of the two different plastic behaviors, it is observed that themodel with an elastic-plastic with a yielding plateau slope behavior has smaller maximum loadresistance than the model with a true stress-strain curve with strain hardening behavior.Finally, some errors and warning messages have occurred during the creation of the input files of themodels and a way of solving them is suggested.

The purpose of the project is to investigate the difference in performance between ballasted and ballastless railway bridges dedicated to high speed trains by taking into account both static and dynamic requirements. The main questions are: a) whether choosing a ballastless bridge results in a more slender section due to the lower self-weight b) if the design of bridges for high speed trains is governed by the static or by the dynamic requirements. The method followed was to first make a complete static design of a ballasted and a ballastless bridge, and then subject them to a 2D dynamic analyses in order to see if the cross section dimensions must be changed. Some of the bridges required a more thorough dynamic analyses, and for these, a 3D model was developed. The analysed bridge is a simply supported beam with a T section carrying one track. Some variations were also considered, namely a simply supported bridge with a double T section carrying two tracks, as well as a single track bridge in two spans. It was found that all of the analysed bridges are somewhat more slender for the ballastless alternative, and require a 10 -30% less reinforcement. Simply supported bridges carrying one track are governed by the dynamic requirements; the bridges in two spans are for shorter spans governed by the statics and for longer spans by the dynamics. Bridges in double T fulfilled all the requirements according to the 2D analyses, but were found to be greatly affected by the 3 dimensional effects and failed to satisfy the criteria when these were taken into account. Finally, the optimal design according to these analyses is a ballastless bridge in a simple T section. If the bridge constructed should carry two tracks, then it should be constructed as two T beams that are not connected to one another in order to avoid the unfavourable 3D effects.

Millimeter wave (mmWave) plays a critical role in the Fifth-generation (5G) new radio due to the rich bandwidth it provides. However, one shortcoming of mmWave is the substantial path loss caused by poor diffraction at high frequencies, and consequently highly directional beams are applied to mitigate this problem. A typical way of beam management is to cluster users based on their locations. However, localization uncertainty is unavoidable due to measurement accuracy, system performance fluctuation, and so on. Meanwhile, the traffic demand may change dynamically in wireless environments, which increases the complexity of network management. Therefore, a scheme that can handle both the uncertainty of localization and dynamic radio resource allocation is required. Moreover, since the localization uncertainty will influence the network performance, more state-of-the-art localization methods, such as vision-aided localization, are expected to reduce the localization error. In this thesis, we proposed two algorithms for joint radio resource allocation and beam management in 5G mmWave networks, namely UK-means-based Clustering and Deep Reinforcement Learning-based resource allocation (UK-DRL) and UK-medoids-based Clustering and Deep Reinforcement Learning-based resource allocation (UKM-DRL). Specifically, we deploy UK-means and UK-medoids clustering method in UK-DRL and UKM-DRL, respectively, which is designed to handle the clustering under location uncertainties. Meanwhile, we apply Deep Reinforcement Learning (DRL) for intra-beam radio resource allocations in UK-DRL and UKM-DRL. Moreover, to improve the localization accuracy, we develop a vision-aided localization scheme, where pixel characteristics-based features are extracted from satellite images as additional input features for location prediction. The simulations show that UK-DRL and UKM-DRL successfully improve the network performance in data rate and delay than baseline algorithms. When the traffic load is 4 Mbps, UK-DRL has a 172.4\% improvement in sum rate and 64.1\% improvement in latency than K-means-based Clustering and Deep Reinforcement Learning-based resource allocation (K-DRL). UKM-DRL has 17.2\% higher throughput and 7.7\% lower latency than UK-DRL, and 231\% higher throughput and 55.8\% lower latency than K-DRL. On the other hand, the vision-aided localization scheme can significantly reduce the localization error from 17.11 meters to 3.6 meters.