Summary form only given. Previous proof-of-principle experiments at NRL used the Mercury IVA facility to test a self-magnetic-pinch (SMP) diode in conjunction with a pulsed resistive heating treatment that cleaned the SMP's anode surface. This heating treatment was tested as a method for mitigating the negative effects of low-Z ions, such as post-shot activation and potentially reduced diode impedance. These low-Z ions can form in the diode region from contaminants on the diode hardware that require in situ cleaning to remove. A more extensive series of experiments are reported here that tested 3 diode configurations with and without the heat treatment. The heat treatment resulted in improvements in spot symmetry, spot size, dose, and radiation pulse length to varying degrees depending on the particular configuration tested. The most significant improvement in dose was achieved from an SMP with a large AK-gap-to-cathodediameter ratio compared with traditional ratios, which are close to 1:1. Large ratios typically result in premature termination of the radiation pulse in the absence of heating. The heat treatment made the large ratio SMP resemble a more stable 1:1 ratio configuration in terms of pulse length, and provides evidence that supports a low-Z ion mechanism as the cause for poor SMP diode impedance at large aspect ratios [1]. The heating circuit itself was a simple resistive pulse through the tantalum anode supplied by switched batteries that drove the anode to >2300 K for approximately 0.5 seconds. Detailed results from this series of experiments will be presented.

Kinetic, time-dependent, electromagnetic, particle-in-cell simulations of the inductive current divider are presented. The inductive current divider is a passive method for controlling the trajectory of an intense, hollow electron beam using a vacuum structure that inductively splits the beam's return current. The current divider concept was proposed and studied theoretically in a previous publication [Swanekamp et al., Phys. Plasmas 22, 023107 (2015)]. A central post carries a portion of the return current (I1), while the outer conductor carries the remainder (I2) with the injected beam current given by Ib = I1 + I2. The simulations are in agreement with the theory which predicts that the total force on the beam trajectory is proportional to (I2−I1) and the force on the beam envelope is proportional to Ib. Independent control over both the current density and the beam angle at the target is possible by choosing the appropriate current-divider geometry. The root-mean-square (RMS) beam emittance (εRMS) varies as the beam propagates through the current divider to the target. For applications where control of the beam trajectory is desired and the current density at the target is similar to the current density at the entrance foil, there is a modest 20% increase in εRMS at the target. For other applications where the beam is pinched to a current density ∼5 times larger at the target, εRMS is 2–3 times larger at the target.

A passive method for controlling the trajectory of an intense, hollow electron beam is proposed using a vacuum structure that inductively splits the beam's return current. A central post carries a portion of the return current (I1), while the outer conductor carries the remainder (I2). An envelope equation appropriate for a hollow electron beam is derived and applied to the current divider. The force on the beam trajectory is shown to be proportional to (I2-I1), while the average force on the envelope (the beam width) is proportional to the beam current Ib = (I2 + I1). The values of I1 and I2 depend on the inductances in the return-current path geometries. Proper choice of the return-current geometries determines these inductances and offers control over the beam trajectory. Solutions using realistic beam parameters show that, for appropriate choices of the return-current-path geometry, the inductive current divider can produce a beam that is both pinched and straightened so that it approaches a target at near-normal incidence with a beam diameter that is on the order of a few mm.

Intense pulsed active detection (IPAD) is a promising technique for detecting fissile material to prevent the proliferation of special nuclear materials. With IPAD, fissions are induced in a brief, intense radiation burst and the resulting gamma ray or neutron signals are acquired during a short period of elevated signal-to-noise ratio. The 8 MV, 200 kA Mercury pulsed-power generator at the Naval Research Laboratory coupled to a high-power vacuum diode produces an intense 30 ns bremsstrahlung beam to study this approach. The work presented here reports on Mercury experiments designed to maximize the photofission yield in a depleted-uranium (DU) object in the bremsstrahlung far field by varying the anode-cathode (AK) diode gap spacing and by adding an inner-diameter-reducing insert in the outer conductor wall. An extensive suite of diagnostics was fielded to measure the bremsstrahlung beam and DU fission yield as functions of diode geometry. Delayed fission neutrons from the DU proved to be a valuable diagnostic for measuring bremsstrahlung photons above 5 MeV. The measurements are in broad agreement with particle-in-cell and Monte Carlo simulations of electron dynamics and radiation transport. These show that with increasing AK gap, electron losses to the insert and outer conductor wall increase and that the electron angles impacting the bremsstrahlung converter approach normal incidence. The diode conditions for maximum fission yield occur when the gap is large enough to produce electron angles close to normal, yet small enough to limit electron losses.

A self-magnetic-pinch diode (SMP) driven by a pulsed power generator is an attractive radiographic source option for future flash radiography experiments. TW-scale flash radiography systems often must be fielded in confined spaces. This can place challenging constraints on the permitted size for the radiographic diode hardware, as well as on the transmission line that must couple the pulsed power generator to the diode. For conventional SMP diodes at 4-8 MV, driven by the coaxial, magnetically-insulated transmission line (MITL) of an inductive-voltage adder (IVA), the MITL's vacuum electron flow current is shed to a large (~1 m dia.) outer conductor chamber just before the load using a torus or knob on the inner conductor. The large diameters and electrostatic field shaping are required to suppress new electron flow from developing post-shedding, which can degrade radiographic performance in standard SMP diodes. Space constraints can prohibit use of this large diameter chamber, which makes preventing reestablishment of the vacuum electron flow current very difficult. Because of this constraint, a new approach for an SMP was investigated where electron flow would not be shed before the SMP diode, but instead would be suppressed or harnessed using an advanced power flow configuration. LSP simulations suggest such an approach may be possible at small MITL diameters. Therefore, a constant-impedance (40-Ω Z flow ), tapered MITL was fielded on NRL's Mercury pulsed power generator to bring the outer conductor from a nominal 38-cm diameter down to 8 cm. The power flow was diagnosed with B-dot monitors along the MITL to measure the impedance and any current losses. The work presented here focuses on these power flow measurements. Outer current losses were minimal through the taper into the load; however, the small diameter MITL operated 10-15% below Mercury's line-limited impedance. One possible cause for the reduced impedance is localized ion source formation at current joints in the outer conductor. Results from LSP simulations and Mercury experiments of the tapered MITL geometry will be presented for both benign and SMP diode loads.

For many applications, control and manipulation of the electron orbits in a high-current electron beam is desirable. This is especially true when a weakly-self-pinched, multi-MV electron-beam is used to make bremsstrahlung radiation. In this case, the radiation pattern is highly peaked along the direction that the electron beam makes when it strikes the x-ray target. Therefore, to maximize the number of photons in the forward direction, it is desirable that the electrons strike the x-ray target as close to normal with as little spread in the beam angles as possible. In this paper, a method for controlling the macroscopic angle of a high-power electron beam using a post-diode magnetic-field structure is presented. The idea is to extract the electron beam into a vacuum cavity through a thin, low-mass foil where a portion of the return-current flows through a central post. The amount of current that flows through the central post and therefore the amount of beam straightening is controlled by inductively splitting the return current so that a portion of it returns through the central post and a portion returns outside the beam. By adjusting the balance between these two currents one can alter the electron orbits and achieve a wide range of angles that the electron beam makes with the target without the need for plasma or an external pulser. 1 Particle-in-cell simulations have been performed to determine the parameters required to straighten an 8-MV, 200-kA, 23-cm-diameter hollow electron beam with an inward 20° macroscopic (average) angle so that it approaches the x-ray target at normal incidence. The simulations show an increase in the forward photon spectrum by up to a factor of 3. Experiments with similar beam parameters using the Mercury Inductive-Voltage Adder at the Naval Research Laboratory have shown an increase of a factor of two in the forward dose using this technique and are in good qualitative agreement with the simulations. Additional simulations and experiments are planned to optimize the forward dose and will be reported on during this talk.

Mercury, a 2-TW inductive voltage adder located at the Naval Research Laboratory in Washington, DC, had previously been converted from negative to positive polarity output by rotating each of the cells. Positive polarity was needed to field an ion-beam diode. However, rotating the cells takes about 2 to 3 weeks and is very labor intensive. So, when we next needed to operate in positive polarity, we considered two options to quickly turnaround Mercury to positive polarity; reverse charging the Marx and fielding a vacuum convolute.

Summary form only given. The self-magnetic-pinch (SMP) electron beam diode is being developed for 4 to 10 M V, 30 to 50 ohms, 50 ns, flash radiography of explosively driven objects by AWE, SNL, and NRL. The goal is a reproducible <; 2-mm FWHM diameter radiographic spot at several hundred Rads (Si) at 1 m. In this talk a new approach is proposed where the outer diameter of an ~50-ohm vacuum impedance MITL is tapered down to ~7 cm diameter to drive an ~3 cm diameter hollow cathode that will be adjusted to operate at <;40 ohm. Undermatching the diode impedance to the MITL helps reduce the MITL vacuum electron flow with only a small reduction of voltage. Recent 3-D PIC simulations have shown that a finite length small outer diameter MITL helps symmetrize and center the electron beam pinch location. A flat carbon anode with a centered <;2-mm diameter W rod extending a few mm from the surface will be placed 2-3 cm from the cathode. This geometry is similar to the negative-polarity non-reentrant rod pinch diode which PIC simulations suggested would operate well at 10 MV. However, they also show that the electron beam initially spreads out to a large diameter potentially leading to ion production from undesirable locations, and the electrons do not pinch to the end of the rod until the diode is close to 10 MV. To mitigate these problems it is suggested that a pulsed laser illuminate the end of the W rod through the center of the hollow cathode to generate low-density plasma before the power pulse arrives. The proper laser energy, timing, and rod tip material and shape must be chosen so that the resultant laser-ablated plasma will touch the tip of the hollow cathode just as the main power pulse arrives at the diode. This plasma density should be just high enough to initially short the gap and conduct the initial electron beam to the tip of the rod. It is speculated that, as the voltage increases and the low density plasma erodes from the cathode tip, the diode will reach critical current with most of the electrons still focused on the tip of the rod. Numerical simulations will be presented along with preliminary experimental results.

Introduction: Passive and active detection of hidden fissile material are subjects of intense interest. In active detection, the naturally emitted characteristic radiations associated with radioactive decay are externally enhanced by inducing fission, thus increasing the strength of these radiations compared with what could be measured passively. One approach to active detection uses the bremsstrahlung produced when an energetic electron beam strikes a high-atomic-number material. Bremsstrahlung photons of energy greater than about 6 mega-electron-volts (MeV) can induce photo-fission in fissionable material, the products of which are detected. Bremsstrahlung-based interrogation using linear accelerators (LINACs) is typically used for active detection, but the LINAC approach requires minutes of exposure to produce a detectable signal above background. The Pulsed Power Physics Branch of NRL’s Plasma Physics Division has pioneered and demonstrated an intense pulsed active detection (or IPAD) approach1–3 using single-pulse, terawatt-level, pulsed-power technology to achieve an inspection time of a few seconds. We have focused on the use of single intense pulses of bremsstrahlung (x-rays) to induce photo-fission, and have also effectively applied the IPAD approach using both neutrons and characteristic gamma rays generated using energetic ion-beam interactions with appropriate targets as the interrogation sources.2 The use of a single, very intense interrogating pulse of either high-energy photons or neutrons has several unique advantages over lower-power systems that repetitively pulse over hundreds of seconds. (1) The short irradiation time leads to a detection time of order seconds (vs minutes for the LINAC-based approach), minimizing the passive background, which is proportional to the detection time.1 Also, the short irradiation time-window (~100 ns) allows for the possibility to measure radiations associated with prompt, short-, and long-lived fission products. (2) Pulsed accelerators driving high-power diodes, when used with a ruggedized commercial off-the-shelf detection system, can rapidly scan moving targets of interest within seconds. (3) Bremsstrahlung can efficiently generate both high-energy photons and neutrons (via photo-neutron processes). This can be useful in scenarios where lowor high-atomic-number shielding is placed around the fissile material. (4) Unique alternative signatures (such as neutronbased activation analyses and delayed-prompt neutron emissions for differentiation of 235U and 238U) may be accessible following a single, intense pulse.

Flash x-radiography usually involves the creation of a pulse of intense bremsstrahlung whose duration is short compared to the motion of the object being radiographed. In many applications it is also desirable to have a very bright source that produces high-spatial-resolution images. For these applications a figure-of-merit (FOM) that provides a measure of the source brightness can be defined as, FOM = D/R2, where D is the dose in CaF 2 at a distance of 1 m from the source and R (usually referred to as the radiographic spot size) is a number that characterizes the spatial extent of the radiation source. One method of creating a high FOM source uses a self-magnetic-pinch (SMP) electron-beam diode. [1] In relevant experiments on RITS-6 [2] and Mercury [3] inductive voltage adders IVA's, magnetically-insulated electron-flow in the adder can be a significant fraction of the current available to drive the SMP. It is believed that the magnetically-insulated flow produces an undesirably large radiographic spot if it is allowed to strike the x-ray target. Therefore, the approach adopted on RITS-6 and Mercury has been to dump this electron flow prior to the x-ray target. This usually involves the use of very large hardware and a loss of up to 50% of the available current. In this talk, we will present results from 3D LSP [4] particle-in-cell simulations of the coupling of magnetically-insulated electron-flow to an SMP diode. These simulations show that asymmetries in the electrical power-flow can cause the radiographic spot to wander. These simulations further show that nearly all the electrical current can be coupled to the SMP diode. However, in this situation the radiographic spot size is larger than cases without electron-flow in the SMP diode. In this paper we use LSP to explore methods of stabilizing the position and reducing the size of the radiographic spot.

Summary form only given. A high-impedance, small spot size diode for flash radiography has been successfully tested on a 2.3-MV generator. The generator is an L-3 Communications, Pulse Sciences Division, Pulserad 43703B, the largest in a series of generators originally called Febetron. These generators are normally used with long-life, larger diameter anodes in a sealed glass envelope. Previous work with the 1-MV Pulserad 43710A generator has demonstrated that the sealed tube can be replaced with a stacked ring insulator and the large anode replaced with a small, 1-mm diameter anode in order to achieve a much smaller spot size for single shot use. Recently, the cathode design for this retrofitted diode has been improved for optimal performance. With only small changes, this vacuum stack and diode have been retrofitted to the 43703B 2.3-MV generator. Because of the higher voltage, extra rings were added to the insulator stack, and the outer conductor hardware was rearranged to provide a smooth surface. This new diode hardware has been successfully tested at the full, 35-kV charge voltage and preliminary results show the spot size reduced by about a factor of 5 to about 1 mm with only a modest decrease in dose. If a figure of merit for the radiography source is defined as the dose divided by the spot size squared, this retrofitted diode improves the figure of merit by more than a factor of 10. Future work will be to optimize the hardware design for field use and to optimize the performance of the diode with the aid of computer models.

The output voltage of the NRL Mercury generator has been successfully increased to 8 MV. Mercury was originally designed for 6 MV, 300 kA operation in negative polarity [1]. However, new experiments required a bremsstrahlung x-ray source with a higher endpoint voltage [2],[3]. The threshold energy for photofission in targets of interest is about 5.5 MeV and, given the roughly Gaussian shape of the Mercury output pulse, very few x rays above threshold could be generated by Mercury with a 6-MV peak output voltage. Also, the cross section for fission rises sharply above the threshold energy. It was therefore decided to increase the output voltage of Mercury to the maximum possible without imposing undo cost or risk. A new center conductor was designed for Mercury that increased the effective output impedance of the MITL from about 23 to 40 Ω, thereby increasing the maximum output voltage to 8 MV. The new center conductor has carefully designed tapers (steps down in diameter) at each stage of the adder so that the adder cell voltages are equally balanced at the design limit of 1.3 MV. One limiting factor is the breakdown probability of the insulator stack in the adder cells. However, simulations have shown that the new center conductor only increases the breakdown probability a small amount. The 40 Ω output impedance was selected because it allowed us to reuse several pieces of the existing center conductor. Aluminum was used for most of the new parts. These steps allowed us to save time and keep material costs to a minimum. PIC and circuit simulations were performed to validate the design. Mercury has now been successfully operated for over 200 shots at 8 MV.

A pinch-reflex ion diode is fielded on the pulsed-power machine Mercury (R. J. Allen, et al., 15th IEEE Intl. Pulsed Power Conf., Monterey, CA, 2005, p. 339), which has an inductive voltage adder (IVA) architecture and a magnetically insulated transmission line (MITL). Mercury is operated in positive polarity resulting in layered MITL flow as emitted electrons are born at a different potential in each of the adder cavities. The usual method for estimating the voltage by measuring the bound current in the cathode and anode of the MITL is not accurate with layered flow, and the interaction of the MITL flow with a pinched-beam ion diode load has not been studied previously. Other methods for determining the diode voltage are applied, ion diode performance is experimentally characterized and evaluated, and circuit and particle-in-cell (PIC) simulations are performed. Results indicate that the ion diode couples efficiently to the machine operating at a diode voltage of about 3.5 MV and a total current of about 325 kA, with an ion current of about 70 kA of which about 60 kA is proton current. It is also found that the layered flow impedance of the MITL is about half the vacuum impedance.

In intense, pulsed active detection, a single, intense pulse of radiation is used to induce photofission in fissionable material, increasing its detectability. The Mercury pulsed-power generator was converted to positive polarity (+3.7 MV, 325 kA, 50-ns FWHM) to drive an intense, pulsed radiation source based on the FIGARO active detection concept. The probing radiation source consisted of an ion-beam diode and a thick PTFE (Teflon) converter where 6-7 MeV γ-rays were produced via the 19F(p,αγ)16O reaction. A suite of radiation detectors was used both to detect the presence of irradiated fissionable material and to characterize the probing radiation source. Four types of detectors were used for the source characterization. Thermoluminescent dosimeters were used to measure the angular distribution of the dose associated with x-rays and γ-rays from the ion-beam diode. Plastic scintillator-photodiode detectors were used to characterize the time dependence of this dose. A plastic scintillator-photomultiplier detector was used to monitor the γ-ray intensity of the probing radiation source and to monitor changes in the production of background neutrons by the diode and PTFE converter. A set of rhodium foil activation counters was used to measure the absolute yield of these background neutrons. Two types of detectors with comparable sensitivities were used to measure delayed neutrons resulting from photofission: 3He proportional counters and a 6Li-loaded-glass-scintillator detector. The neutron detection rate from each detector following the probing radiation pulse was over 100 times higher with depleted uranium present than with lead.

A high-power, high-impedance ion beam diode has been fielded on the Mercury inductive voltage adder at typical parameters of 4 MV, 360 kA, and 70–90-kA ion current. These results are consistent with theory and with LSP calculations. This beam is focused onto a CF2 target to produce characteristic gammas via the 19F(p,αγ)16O reaction. Diagnostics characterize the ion emission from the anode, the beam distribution on the target, and the gamma yield. Shaping the anode surface improves beam focusing on the target. Our experimental arrangement allows us to resolve the gamma signal in the presence of the diode bremsstrahlung, and to minimize spurious neutron production. The calculated gamma yield exceeds 1011 gammas/sr.

After 616 shots in a negative polarity configuration, Mercury, a 6-MV and 300-kA inductive voltage adder (IVA), has been converted to positive polarity in order to extract ion beams. Conversion to positive polarity was achieved by rotating all six of the adder cells by 180°. In principle, we could have chosen to instead insert the center conductor from the other end of the adder to change polarity, but rotating the cells minimized the time required to make the transition. Although most of the same pieces were used, the center conductor had to be reconfigured in order to align the transition pieces with the cell feed gaps. Because the electron flow was anticipated to be very different in positive polarity, a result of emission from surfaces of different potential, a simple blade diode was fielded for the initial shots to gain a better understanding of operation in positive polarity. The blade diode consisted of the same cathode used as a dummy load in the first negative polarity shots on Mercury, but with a different carbon anode that just covered the end of the center conductor. After a few short circuit and initializing shots, a series of shots were taken where only the blade diode AK gap was varied in order to characterize self-limited and load-limited operation and to compare measurements with theory and simulation.

During the past decade, the rod-pinch diode has been used extensively for high-brightness, megavolt x-ray radiography. Success of the rod pinch derives from its small radiographic spot size for a range of driving voltages and x-ray spectra. Here, we examine two experimentally observed and more subtle phenomena that may impact radiographic performance for some systems: flat-field nonuniformity in the image plane due to x-ray absorption in the rod, and weak, large-scale structures around radiographic features due to cathode-scattered x-rays. X-ray imaging data from Cygnus and HRS are analyzed by the ITS Monte Carlo codes to determine if these phenomena are significant for Cygnus.