Magnet assemblies with integrated charged particle beam dumps

Abstract

Magnet assemblies, e.g., dipole magnet assemblies, with integrated charged particle beam dumps and systems containing the same are described. Examples of a dipole magnet assembly include a dipole bending magnet and a channel assembly configured with a charged particle beam dump. The dipole magnet can be configured to generate a uniform magnetic field within the channel assembly that bends a trajectory of a charged particle beam. The channel assembly can be configured to support a vacuum for the charged particle beam while absorbing secondary-state particles using the charged particle beam dump. The charged particle beam dumps can be releasably secured to the channel assembly, welded to the channel assembly, or inserted into the channel assembly. Each charged particle beam dump can utilize an active cooling system such as cooling tubes or an embedded cooling channel to mitigate excessive heat generation from secondary-state particle absorption. Materials are also described.

Description

[0001]Attorney Docket No.57383-0012WO1 MAGNET ASSEMBLIES WITH INTEGRATED CHARGED PARTICLE BEAM DUMPS FIELD This specification relates generally to magnet assemblies, e.g., dipole magnet assemblies, with integrated charged particle beam dumps and systems containing the same. BACKGROUND Boron neutron capture therapy (BNCT) is a modality of treatment for a variety of types of cancer, including some of the most difficult types. BNCT is a technique that selectively aims to treat tumor cells while sparing the normal cells using a boron compound. The boron compound allows for efficient uptake by multiple cell types and selective drug accumulation at target sites, such as tumor cells. Boron loaded cells can be irradiated with neutrons, e.g., in the form of a neutron beam. The neutrons react with the boron to eradicate the tumor cells. Neutron beams for BNCT can be generated through various techniques. One such technique involves irradiation of a suitable neutron-generating target with a charged particle beam, such as a proton or deuteron beam. The charged particles react with nuclei in the target to emit a beam of neutrons in a target energy range (e.g., an epithermal energy spectrum) that can be used for BNCT. In order to generate a maximum number of neutrons in this target energy range, controllable beam optics are used to ensure that the cross-section and phase space of the charged beam are maintained throughout a beamline. This allows the charged beam to be scanned onto the target with reasonable uniformity. Moreover, dipole magnets (e.g., dipole bending magnets) are often implemented into the beamline to alter the charged beam’s trajectory, such that it can be steered into a desired direction, e.g., toward a target that is proximate a treatment room for use in BNCT. However, a charged particle beam is generally not composed entirely of the same species of particle with the same kinetic energy. A non-negligible number of secondary-state particles may also be present. Secondary-state particles can include particles having different atomic numbers, atomic masses, charges, and / or kinetic energies than the primary-state particles of the charged beam, each of which generally results in a different momentum-to- charge ratio. Thus, when interacting with the magnetic field generated by a dipole magnet, these secondary-state particles may deflect at different angles and collide with components of the beamline. Such collisions can immediately, or eventually, cause damage due to excessive Attorney Docket No.57383-0012WO1 heat generation and / or ion implantation. This can cause failure of the beam system, e.g., due to destroying or contaminating a vacuum. SUMMARYThis specification describes magnet assemblies (e.g., dipole magnet assemblies) with integrated charged particle beam dumps and systems (e.g., beam systems) containing the same. Example dipole magnet assembly (DMAs) are described in an illustrative context of a BNCT system, where the DMAs are positioned in a beamline and configured to bend a charged particle beam at a specified bending angle while filtering secondary-state particles (SSPs) from the beam using a charged particle beam dump. For example, a DMA can be positioned at a branch point in the beamline to direct the charged beam along an axis toward a target. The charged particle beam may be manipulated or measured with other beam optics components implemented into the beamline, e.g., quadrupole magnets, scanning magnets, beam correctors, or current monitors, before the charged beam irradiates the target to generate a neutron beam for BNCT. The DMA can include various components based on location, function, dimension, and / or constituent material. The components can include a dipole bending magnet and a channel assembly configured with the charged particle beam dump. The dipole magnet can include a pair of electromagnetic coils and a back iron enclosing the coils, thereby providing a magnetic flux return path for the coils. The channel assembly can be positioned between the coils of the dipole magnet. The dipole magnet can be configured to generate a uniform magnetic field in the channel assembly to bend the trajectory of the charged beam at the specified bending angle. The channel assembly can be connected to multiple beam tubes and configured to support a vacuum for the charged particle beam. The channel assembly can employ multiple different types of charged particle beam dumps to absorb SSPs that follow altered trajectories from the primary-state particles (PSPs) of the charged beam. The charged particle beam dumps can include releasable, welded-on, or insertable beam dumps, that can each utilize an active cooling system such as cooling tubes or an embedded cooling channel to provide improved thermal management. The releasable and welded-on beam dumps can each form a wall of the channel assembly using a support piece composed of a material having a high thermal conductivity and a high melting point. The support piece can be further coated with a layer of refractory metal that includes one or more refractory metal elements, e.g., in pure or alloy form, providing improved resistance to heat and wear. The insertable beam dumps can be implemented as an insert piece that can be inserted through a port Attorney Docket No.57383-0012WO1 disposed on an outer surface of the channel assembly. Similar to the aforementioned support pieces, the insert piece can be composed of a material having a high thermal conductivity and a high melting point, as well as coated with a layer of refractory metal to facilitate high temperature resilience. An embedded cooling channel of the insert piece can be attached to a dual liquid feedthrough to circulate coolant within the insert piece when it is positioned within the channel assembly. The DMA can have multiple different configurations and bending angles depending on the implementation. Numerous examples of DMAs with integrated charged particle beam dumps are disclosed that perform some or all of these functions. Numerous exemplary materials are also disclosed with capabilities to perform one or more of these functions. Particular embodiments of the subject matter described in this specification can be implemented so as to realize one or more of the following advantages. The disclosed charged particle beam dumps can have relatively long lifecycles and / or a high degree of modularity due to various innovative design choices. The beam dumps utilize materials that are highly thermally conductive, have high melting points, and are resistant to decomposition. This allows the charged particle beam dumps to uniformly distribute a high heat load while being resilient to wear from charged particle (e.g., SSP) collisions. Moreover, the charged particle beam dumps use active cooling systems, e.g., cooling tubes or embedded cooling channels, that can maximize convective heat transfer with a coolant, extending their lifecycles further. Due to the high temperatures exhibited by the beam dumps (e.g., about 500 Kelvin (K) or more, about 750 K or more, about 1000 K or more), radiative heat transfer becomes a significant factor which can alter, in most cases undesirably, the temperature profile of the beam dumps, e.g., toward one with hot spots. The active cooling systems disclosed herein provide a means of reducing radiative heat transfer in favor of convective heat transfer which generally provides a more uniform temperature profile in the beam dumps, e.g., with minimal or lower temperature hot spots. The details of one or more embodiments of the subject matter of this specification are set forth in the accompanying drawings and the description below. Other features, aspects, and advantages of the subject matter will become apparent from the description, the drawings, and the claims. BRIEF DESCRIPTION OF DRAWINGS FIG. 1A is a schematic view depicting an example of a neutron beam system for use in boron neutron capture therapy (BNCT). Attorney Docket No.57383-0012WO1 FIG. 1B is a schematic view depicting a more detailed example of a neutron beam system configured for use in BNCT. FIG. 2A is a perspective view depicting an example of a neutron generating target. FIG. 2B is a side view depicting an example of an assembly for housing a neutron- generating target. FIG. 2C is a cross-sectional view depicting an example of an assembly for housing a neutron-generating target. FIG. 3 is a perspective view depicting an example of a dipole magnet assembly including a dipole magnet and a channel assembly. FIGs. 4A-4C are various views depicting an example of a dipole bending magnet. FIG. 5A is a cross-sectional view depicting an example of a channel assembly. FIG. 5B is a cross-sectional view depicting another example of a channel assembly. FIGs. 6A-6B are cross-sectional views depicting an example of a dipole magnet assembly including a dipole magnet and a channel assembly. FIGs. 7A-7B are various views depicting an example of a channel assembly having a releasable charged particle beam dump. FIG. 7C is a section view depicting an example of a dipole magnet assembly including a dipole magnet and a channel assembly having a releasable charged particle beam dump. FIG. 8A-8B are various views depicting an example of a channel assembly having a welded-on charged particle beam dump. FIG. 8C is a perspective view of a charged particle beam dump that can be welded onto a channel assembly. FIG. 8D is a section view depicting an example of a dipole magnet assembly including a dipole magnet and a channel assembly having a welded-on charged particle beam dump. FIGs. 9A-9B are various views depicting an example of a channel assembly having an insertable charged particle beam dump. FIG. 9C is a section view depicting an example of a dipole magnet assembly including a dipole magnet and a channel assembly having an insertable charged particle beam dump. FIGs. 10A-10C are schematic views depicting examples of neutron beam systems that include high-energy beamlines configured with one or more dipole magnet assemblies. Attorney Docket No.57383-0012WO1 FIG. 11 is a flow diagram of an example method for bending and filtering a charged particle beam using a dipole magnet assembly with an integrated charged particle beam dump. Like reference numbers and designations in the various drawings indicate like elements. DETAILEDDESCRIPTIONThe term “particle” is used broadly herein and, unless otherwise limited, can be used to refer to an electron, a proton (or H+ ion), or a neutron, as well as a species having more than one electron, proton, and / or neutron, e.g., other ions, atoms, and molecules. The term “primary-state particle”, abbreviated as “PSP”, is used herein and generally refers to the charged particles of a charged particle beam having a same atomic number, atomic mass, charge, and kinetic energy. In the context of a beam system, the PSPs constitute the majority, desired particles of a charged particle beam. Since the kinetic energy of a free particle is continuous, a “same kinetic energy” refers to particles with up to roughly 5 % deviation in kinetic energy from the mean kinetic energy of the charged particle beam. The term “secondary-state particle”, abbreviated as “SSP”, is used herein and generally refers to the charged particles of a charged particle beam having a different atomic number, atomic mass, charge, and / or kinetic energy than the PSPs of the charged beam. In other words, a SSP is a charged particle that is of a different species and / or has a different kinetic energy than the PSPs. In the context of a beam system, the SSPs constitute the remaining, undesired particles in a charged beam. Here, a “different kinetic energy” refers to particles with more than roughly 5 % deviation in kinetic energy from the mean kinetic energy of the charged particle beam. For instance, in a beam system, the particle distribution of a charged particle beam may vary similarly to a Maxwell-Boltzmann distribution, e.g., a multivariate Gaussian in phase space (e.g., velocity and position components). Deviations in the kinetic energy and injection of SSPs in the charged beam can arise from, for example, variations in the tandem voltage of a tandem accelerator, voltage ripples in the power supplies of the accelerator, inefficiencies in the charge-exchange process of the accelerator, scattering events in a low vacuum, temperature changes in the plasma of an ion source, Coulomb interactions in the charged beam, thermal non-equilibrium of the charged beam, among other effects. As one example, a proton beam generated by a beam system may include about 90% PSPs and 10% SSPs, where: (i) the PSPs are protons having with the same kinetic energy, Attorney Docket No.57383-0012WO1 and (ii) the SSPs are also protons but having roughly half the kinetic energy of the PSPs. In this example, although the SSPs are of the same species as the PSPs, they have different kinetic energies and therefore will have trajectories that deviate from the main trajectory of the PSPs when interacting with a magnetic field. Such situations may arise in a tandem accelerator if some proportion of protons lose their kinetic energy from the first acceleration stage, e.g., during the charge-exchange stage when protons are converted from negative hydrogen ions (H-). As another example, a deuteron beam generated by a beam system may include about 95% PSPs and 5% SSPs, where: (i) the PSPs are deuterons having the same kinetic energy, and (ii) the SSPs are protons having the same kinetic energy as the PSPs. In this example, although the SSPs have the same atomic number, charge, and kinetic energy as the PSPs, they have a different atomic mass and therefore will have trajectories that deviate from the main trajectory of the PSPs when interacting with a magnetic field. Such situations may arise when a negative deuterium ion source generates a small proportion of negative hydrogen ions (H-) in addition to negative deuterium ions (D-). Examples of dipole magnet assemblies (DMAs) are described herein, which can be implemented into a beamline of a beam system, e.g., including a reactor or a particle accelerator. The example DMAs described herein can be used with any type of beam system in which bending of a charged particle beam and absorption of SSPs throughout the beamline is desired. DMAs described herein can be used in numerous applications, an example of which is a neutron beam system for generation of a neutron beam for use in boron neutron capture therapy (BNCT). BNCT utilizes a beam of epithermal neutrons, e.g., with an energy spectrum between one electronvolt (eV) and thirty kiloelectronvolts (keV), for cancer treatment. In BNCT, the neutrons can be generated from nuclear reactions of charged particles (e.g., protons or deuterons) colliding with either a beryllium or lithium target. The neutrons are then administered to a patient for treatment. The charged particle beam can be generated, accelerated, focused, and directed towards the target using various devices in the beam system. For example, a DMA can bend a trajectory of the charged particle beam at a particular angle by applying a calibrated magnetic field. SSPs that exhibit a differing Lorentz force than the PSPs will generally follow different trajectories and can be absorbed with one or more integrated charged particle beam dumps positioned in one or more SSP strike-zones. Thus, the DMA can steer a charged particle beam, e.g., toward a target proximate a treatment room, while simultaneously filtering the charged beam of SSPs, thereby mitigating degradation or failure of the beam system due to frequent SSP collisions. Moreover, the Attorney Docket No.57383-0012WO1 filtered charged particle beam generally includes about 100 % PSPs which can be useful for BNCT applications. For example, SSPs that irradiate the target can have a higher probability of generating neutrons outside the epithermal energy range. A filtered charged particle beam that includes nearly 100 % PSPs can, in some cases, increase the proportion of neutrons generated in an epithermal energy range. The examples of dipole magnet assemblies (DMAs) described herein are not intended to be viewed in isolation from each other. All features, elements, components, and functions described with respect to any example DMA provided herein are freely combinable and substitutable with those from any other example DMA unless explicitly stated otherwise. Examples of Neutron Beam Systems For ease of description, the examples of DMAs described herein will be done so in the context of a neutron beam system that converts a charged particle beam to a neutron beam for use in BNCT, where the DMAs are positioned in a beamline of the neutron beam system and configured to steer the charged particle beam toward a target to generate the neutron beam. However, the DMAs are not limited to such. The DMAs can be used in other beam systems that generate and manipulate charged particle beams, even those outside of BNCT applications utilizing different energy ranges, e.g., particle colliders, irradiation tests, isotope production, crystallography, among others. For example, the DMAs can be implemented in beamlines configured for experiments in particle physics, materials science, life science, chemistry, and molecular biology, among other applications. FIG. 1A illustrates a schematic view of an example beam system 100A configured as a neutron beam system for use in BNCT. At a high level, beam system 100A is configured to create a charged particle beam 61 and propagate it to a target 60 to generate a neutron beam 70. Neutron beam 70 is then directed towards a patient body 80 to be irradiated, e.g., a head of the patient 80. Beam system 100A includes a charged particle source 20, a low-energy beamline (LEBL) 30, an accelerator 40, and a high-energy beamline (HEBL) 50. Source 20 is configured to generate the charged particle beam 61, which is output to LEBL 30. LEBL 30 is configured to transport the charged beam 61 from source 20 to accelerator 40. Accelerator 40 is configured to accelerate the charged particle beam 61 to a higher energy. HEBL 50 extends from the accelerator 40 to target 60 housed within a target assembly portion of HEBL 50, see FIGs. 2A-2C for example. HEBL 50 transfers the charged particle beam 61 from an output of accelerator 40 to target 60, where it is converted to neutron beam 70. Attorney Docket No.57383-0012WO1 FIG. 1B is a schematic view illustrating a more detailed example of a beam system 100B configured as a neutron beam system for use in BNCT. Beam system 100B includes a pre-accelerator system 26 forming at least a portion of LEBL 30, where pre-accelerator system 26 serves as a charged particle beam injector. Beam system 100B includes a high- voltage tandem accelerator 40 coupled to LEBL 30, and HEBL 50 extending from tandem accelerator 40 to a target 60, as described with reference to the beam system 100A of FIG. 1A. LEBL 30 transfers a negative ion beam (e.g., hydrogen ions (H-)) from ion source 20, through pre-accelerator 26 which boosts the energy level of the ion beam and converges the ion beam, to an input (e.g., an input aperture) of accelerator 40. Accelerator 40 is powered by a high-voltage power supply 42 coupled thereto. Accelerator 40 includes a vacuum tank, a charge-exchange tube, accelerating electrodes, and a high-voltage feedthrough. Accelerator 40 can, in some implementations, accelerate a hydrogen ion beam to produce a proton beam with an energy generally equal to twice the voltage applied to the accelerating electrodes positioned within accelerator 40. The energy level of the proton beam can be achieved by accelerating the beam of negatively charged hydrogen ions from the input of accelerator 40 to the innermost high-potential electrode, stripping two electrons from each hydrogen ion, and then accelerating the resulting protons downstream by the same voltages encountered in reverse order. HEBL 50 transfers the proton beam from the output of accelerator 40 to the neutron- generating target 60 positioned at the end of a branch 71 of the beamline extending into a patient treatment room. Beam system 100B is configured to direct the proton beam to one or more targets 60 and associated target areas. In some implementations, HEBL 50 includes multiple (e.g., three) branches 71, 81, and 91 configured to extend to multiple different patient treatment rooms, with each branch terminating in a target 60. HEBL 50 includes a pumping chamber 51, quadrupole magnets 52 and 72 to prevent de-focusing of the beam, dipole (or bending) magnets 56 and 58 (e.g., configured as DMAs 200) to steer the beam towards one or more targets 60, beam correctors 53, diagnostics such as current monitors 54 and 76, a fast beam position monitor 55 section, and a scanning magnet 74 for branch 71. Branches 81 and 91 can contain components similar to branch 71. The design of HEBL 50 depends on the configuration of the treatment facility (e.g., a single-story treatment facility, a two-story treatment facility, and the like). The proton beam can be delivered to target 60 (e.g., positioned near a treatment room having a patient 80) with the use of dipole magnet 56. Quadrupole magnets 72 can be included to then focus the proton Attorney Docket No.57383-0012WO1 beam to a certain size at target 60. The proton beam can pass one or more scanning magnets 74, which provide lateral movement of the proton beam onto the target surface in a desired pattern (e.g., spiral, curved, stepped in rows and columns, combinations thereof, and others). The beam lateral movement can enable generation of smooth and even time-averaged distribution of the proton beam on the target 60, preventing overheating of the target 60 and making the particle (e.g., neutron) generation as uniform as possible within the target 60 (e.g., neutron generating layer 121 of FIG. 2A). Scanning magnets 74 can be configured to direct the proton beam to a current monitor 76, which measures beam current. The beam current value can be used to operate a safety interlock. The target assembly 65 containing target 60 can be physically separated from the high-energy beamline volume with a valve 77. A function of valve 77 is to separate the vacuum volume of the beamline from the target 60 during removal of a used target and loading of a new target. In some implementations, instead of being bent by 90 degrees by a dipole magnet 56, the beam can be directed straight to one or more quadrupole magnets 52 located in the horizontal beamline. The beam could be bent by another bending magnet 58 to a preset angle, depending on a setting requirement (e.g., location of a patient or a room configuration). In some implementations, bending magnet 58 can be arranged at a branch point in the beamline and can be configured to direct the beam in one of two directions for two different treatment rooms located on the same floor of a medical facility, see FIGs. 10A- 10C for example. Beam systems 100 as described with respect to FIGs. 1A-1B are examples of different configurations that can be used to generate charged particle and neutron beams. Different configurations of beam systems 100 can utilize accelerators other than electrostatic tandem accelerators, as well as targets that are either fixed or rotating. The examples of DMAs 200 described herein are not limited to use with any one type of beam system or neutron beam generating system. FIG. 2A is a perspective view of an example neutron-generating target 60. In this example, target 60 has a neutron generating layer 121 with a charged particle receiving face surface 122. Neutron generating layer 121 is positioned on or in proximity to a substrate 123. In some cases, layer 121 is covered with one or more other layers for protection. Layer 121 can also have one or more underlying layers between layer 121 and substrate 123, e.g., to resist blister formation. A charged particle beam, such as a proton beam, incident upon face 122, passes into target 60 and causes layer 121 to undergo a reaction that generates neutrons. This is the Li-7(p,n)Be-7 nuclear reaction in the case where neutron generating layer 121 is Attorney Docket No.57383-0012WO1 composed of lithium-7. Neutron generating layer 121 can alternatively be beryllium-9, and neutrons can be generated with a proton beam (Be9(p,n)B9) or deuteron beam (Be9(d,n)B10) at different energies. Substrate 123 can be a material with excellent thermal conductivity, such as copper or aluminum, to assist in removal of the heat generated by the reactions. FIG. 2B is a side view of an example target assembly 65 that can form a terminal portion of HEBL 50. Target 60 (not shown) can be contained within assembly 65 at or near end 67. The charged particle beam enters assembly 65 at end 66 and travels to the opposite end 67 where it impacts target 60. Various cooling channels 68 are routed to and from end 67 for the insertion and removal of coolant used to regulate the temperature of target 60 during use. Numerous sensors can be included to monitor temperature and radioactivity of and around assembly 65. Also shown is valve 77 in the example form of a gate valve. FIG. 2C is a cross-sectional view of an example target assembly 65 omitting components such as valve 77, coolant channels and sensor connections for clarity. A sidewall 62 has a tubular shape and contains an interior space 64 at a vacuum or near vacuum level. Target 60 is positioned at end 67 and held in place by end cap 63. Variations of this construction are possible, such as with target 60 surrounded by sidewall 62. Charged particle beam 61 is directed through interior space 64 and scanned across target 60 by scanning magnet 74 located upstream on HEBL 50 (not shown). Neutrons produced by target 60 will be emitted at some level in virtually all directions from target 60, but the majority of the neutrons will be emitted in a disperse but generally forward directed path. This is depicted here as neutron beam 70 in raw form. Examples of Dipole Magnet Assemblies FIG. 3 is a perspective view depicting an example DMA 200 that includes a dipole bending magnet 205 and a channel assembly 300 positioned within dipole magnet 205. DMA 200 is an example of a magnet assembly that can be arranged in a beamline of a beam system to bend a trajectory of a charged particle beam. For example, DMA 200 can be implemented as dipole magnets 56 and / or 58 in the HEBL 50 of the neutron beam system 100B of FIG. 1B. Other example neutron beam systems utilizing one or more DMAs 200 are depicted in FIGs. 10A-10C. For clarity, a description of DMA 200 is provided with reference to FIGs. 4A-6B to explain how DMA 200 generates a uniform magnetic field for the charged particle beam using dipole magnet 205, while supporting a vacuum for the charged beam using channel assembly 300. A discussion on how SSPs of the charged beam can compromise vacuum Attorney Docket No.57383-0012WO1 integrity is also provided. Following this discussion, a description of DMAs 200 integrated with different types of charged particle beam dumps is outlined with reference to FIGs. 7A- 9C. The charged particle beam dumps offer an effective means for managing SSP collisions that can occur when the charged beam interacts with the uniform magnetic field generated by dipole magnet 205. FIGs. 4A-4C are various views depicting an example dipole bending magnet 205 that can be implemented in DMA 200. FIG. 4A is a perspective view of dipole magnet 205. FIG. 4B is a cross-sectional view of dipole magnet 205 along a horizontal bisecting plane (horizonal plane 214-1). FIG. 4C is a cross-sectional view (Section A-A’) of dipole magnet 205 along a vertical bisecting plane. As shown in FIG. 4A, dipole magnet 205 includes a first electromagnetic coil 220-1, a second electromagnetic coil 220-2, and a back iron 210 enclosing first 220-1 and second 220- 2 coils. While many different coil configurations can be utilized, in this example, coils 220 are configured similarly to a Helmholtz coil but with deformed loop shapes as opposed to circular loop shapes. Particularly, each coil 220 is substantially planar and includes a winding (e.g., a wire) with one or more turns around the coil 220’s deformed loop shape. Each coil 220 is configured to conduct current through its winding to generate a respective magnetic field. Coils 220 can include windings with any number of turns, e.g., 5 or more turns, 10 or more turns, 25 or more turns, 50 or more turns, 100 or more turns, etc. For a same current, a larger number of turns generally increases the strength of the magnetic field generated by each coil 220 due to the increased current density (current per unit area) through the coil 220. Back iron 210 is made of a ferrous metal that is magnetic and provides a magnetic flux return path for the magnetic fields generated by coils 220. For example, back iron 210 can be composed of cast-iron, iron in pure form (e.g., 99.8% pure Fe), iron in pure form annealed with hydrogen (e.g., 99.95% pure Fe), or a suitable iron alloy having a desired set of magnetic properties. Consequently, back iron 210 can have a relative magnetic permeability ^^about 5000 or more, e.g., about 10000 or more, about 20000 or more, about 50000 or more, about 100000 or more, about 200000 or more, depending on the choice of ferrous metal and / or other considerations, e.g., durability, machineability, cost, weight, size constraints, etc. Due to its relatively high magnetic permeability, back iron 210 can focus and redirect the magnetic fields generated by coils 220 into a specific region of dipole magnet 205 with a specific magnetic field configuration. As explained in more detail below, back iron 210 is configured to focus the magnetic fields generated by coils 220 to produce a Attorney Docket No.57383-0012WO1 uniform magnetic field that can be used to bend a trajectory of a charged particle beam at specific bending angle. In this example, an outer surface of back iron 210 is covered by metallic plates 206 that are secured to back iron 210 by releasable fasteners 208, e.g., nuts and bolts. Plates 206 can provide protection from electromagnetic interference into and / or out of dipole magnet 205, e.g., in the form of a Faraday shield. For example, plates 206 can confine the magnetic fields generated by coils 220 such that they circulate within back iron 210 instead of escaping out of dipole magnet 205. As another example, plates 206 can shield against stray electric and / or magnetic fields generated by other beam optics components, e.g., other dipole magnets, quadrupole magnets, scanning magnets, etc., that may also be implemented in a beamline. Mounting elements 207 are disposed on plates 206 which allow dipole magnet 205 to be positioned within a beamline, e.g., by mounting to a ceiling or other support structure. Although not shown in FIGs. 4A-4C, dipole magnet 205 may also be equipped with a utility panel that includes, for example, voltage taps for coils 220, current monitors for coils 220, safety disconnects (e.g., fuses or breaks) to protect coils 220 against voltage spikes from a power supply, and / or an active cooling system to provide liquid cooling for coils 220. As shown in FIGs. 4B and 4C, coils 220 and back iron 210 are arranged relative a first axis 305-1, a second, non-parallel axis 305-2, and a circular arc 306. First axis 305-1, second axis 305-2, and circular arc 306 are coplanar such that the first 305-1 and second 305- 2 axes intersect each other at an intersection point 308 and are tangents of the circular arc 306 at respective tangent points 307-1 and 307-2. Tangent points 307-1 and 307-2 designate the end points of circular arc 306 which has a radius of ^, subtends a bending angle of ^, and has a corresponding arc length of ^ = ^^. In general, circular arc 306 coincides with a desired trajectory for a charged particle beam when interacting with the uniform magnetic field generated by dipole magnet 205. In some implementations, the radius ^ can be about 500 mm to 1000 mm and the bending angle ^ can be about 10 degrees to 170 degrees depending on the particular configuration of dipole magnet 205. In this example, bending angle ^ is depicted as 90 degrees but bending angles of 45 degrees and 135 degrees are also common choices. First 305-1 and second 305-2 axes span a horizontal plane 214-1 that constitutes a plane of symmetry of dipole magnet 205. Each coil 220 is secured to back iron 210, e.g., by clasps 222, such that first coil 220-1 is positioned above horizontal plane 214-1 and second coil 220-2 is positioned below horizontal plane 214-1. This forms a magnetic field Attorney Docket No.57383-0012WO1 configuration similar to two stacked current loops, e.g., a Helmholtz coil. Particularly, first 220-1 and second 220-2 coils are positioned opposite each other along a vertical axis 214-2 within a cavity 213 formed by back iron 210. Cavity 213 has a cross-section resembling the letter “H” which allows back iron 210 to redirect fringing magnetic fields generated above first coil 220-1 and below second coil 220-2. Thus, back iron 210 provides a magnetic flux return path for coils 220. Back iron 210 forms a first pole 211-1 and a second pole 211-2 that each protrude vertically into the cavity 213. Each pole 211 acts as a magnetic north pole or a magnetic south pole to emanate or terminate the uniform magnetic field generated within cavity 213. First pole 211-1 and second pole 211-2 are aligned with each other along vertical axis 214-2 such that the uniform magnetic field ^ = ±^^̂ is generated between respective pole faces 212-1 and 212-2. Pole faces 212 are orthogonal to vertical axis 214-2, thus the uniform magnetic field ^ is generated in the positive or negative z-direction with little or no gradient. Generally, coils 220 conduct current of the same magnitude and in the same clockwise or counterclockwise direction so their magnetic fields superimpose constructively to generate the uniform magnetic field. Here, clockwise and counterclockwise are specified by the curl right-hand rule such that a thumb points in the positive z-direction. Hence, current conducted clockwise in coils 220 generates the uniform magnetic field in the negative z-direction ^ = −^^̂, while current conducted counterclockwise in coils 220 generates the uniform magnetic field in the positive z-direction ^ = +^^̂. With respect to FIG. 4C, a cross indicates a direction of current going into the page and a dot indicates a direction of current coming out of the page. Coils 220 and back iron 210 have geometries that follow circular arc 306 along its bending angle ^. That is, coils 220 each constitute a deformed loop around circular arc 306, whereas back iron 210 extends along circular arc 306. In this way, the uniform magnetic field ^ is generated between pole faces 212 along the entire curved trajectory of a charged beam traveling in the horizontal plane 214-1. This generally provides a compact geometry for dipole magnet 205 while continuously imparting a calibrated Lorentz force on the charged beam such that it follows a trajectory along circular arc 306. For ease of description, it is assumed that the charged beam is received along first axis 305-1 in the positive y-direction and is output along the second axis 305-2 in the negative x-direction. Hence, the charged beam experiences a Lorentz force ^ = ^^ × ^ that points toward an origin 309 of the circular arc 306, where ^ and ^ are the charge and velocity, respectively, of a charged particle of the Attorney Docket No.57383-0012WO1 beam. In the case of a positively charged beam, dipole magnet 205 generates the uniform magnetic field in the negative z-direction. In the case of a negatively charged beam, dipole magnet 205 generates the uniform magnetic field in the positive z-direction. The direction of the uniform magnetic field in each of these cases is reversed if the charged beam is received along second axis 305-2 in the positive x-direction and is output along first axis 305-1 in the negative y-direction. Note, while circular arc geometries are described in detail, non-circular arcs (e.g., generic arcs or open curves) can also be utilized with the example DMAs 200, dipole magnets 205, and channel assemblies 300 described herein. For instance, a non-circular arc can correspond to a trajectory of a charged particle beam under a non-uniform magnetic field, e.g., having a magnetic field gradient along the trajectory of the charged particle beam. In this case, coils 220 and back iron 210 of dipole magnet 205 may be configured with a more complex geometry, e.g., such that the distance between poles 211 in the z-direction increases (or decreases) along the trajectory of the charged beam. In some implementations, pole faces 212 may also have gradient portions in a direction orthogonal to the trajectory, e.g., in the A- A’ direction, which can provide focusing of the charged particle beam. FIGs. 5A-5B are cross-sectional views depicting two examples of channel assemblies 300 that can be implemented in DMA 200. FIG. 5A shows a first example of a channel assembly 300A. FIG. 5B shows a second example of a channel assembly 300B. Second channel assembly 300B is configured similarly to first channel assembly 300A, therefore the two examples are described together with the differences stated when appropriate. In the described examples, channel assembly 300 is a weldment, e.g., a vacuum weldment, although the examples are not limited to such. In some cases, channel assembly 300 may be formed using other fabrication techniques, e.g., three-dimensional (3D) metal printing techniques, although these approaches may be more expensive than conventional welding methods. In general, a weldment is an assembly of multiple metal pieces that are welded together. Channel assembly 300 can be welded together from metal pieces that have been cut, bent, or otherwise fabricated to an appropriate size and shape using any suitable welding technique, where such welding techniques may be based on the particular material selected for channel assembly 300. As one example, if channel assembly 300 is formed from stainless steel metal pieces, welding techniques such as metal inert gas (MIG) welding, tungsten inert gas (TIG) welding, or manual metal arc (MMA) welding may be used. Attorney Docket No.57383-0012WO1 Preferably, channel assembly 300 is made from non-magnetic metals (e.g., non- ferrous metals) that are transparent to magnetic fields, e.g., having a relative magnetic permeability close to unity ^^≈ 1. In this way, channel assembly 300 does not significantly distort the uniform magnetic field generated by dipole magnet 205, which can affect the trajectory of a charged particle beam propagating within channel assembly 300. Such non- magnetic metals can include, but are not limited to, austenitic stainless steel, copper, tin, aluminum, titanium, and each of their respective alloys. Austenitic stainless steels, copper, and copper alloys are particularly well-suited for channel assembly 300 due to having high melting points, being resistant to wear and corrosion, and being relatively inexpensive compared to alternatives, e.g., titanium and titanium alloys. Certain non-magnetic nickel alloys, e.g., Inconel alloys, are also suitable for extreme environments and can be used for channel assembly 300 in some implementations. Channel assembly 300 supports a vacuum when connected to beam tubes or other beam optics components in a beamline. Air can be evacuated from the beamline to generate a suitable vacuum within channel assembly 300 using one or more vacuum pumps, e.g., mechanical pumps, and / or high vacuum diffusion pumps. The vacuum can have a pressure from about 10^^Torr to about 10^^Torr depending on the configuration of a particular beam system. For example, a mechanical pump can generate a vacuum of about 10^^Torr and a high vacuum diffusion pump can reduce the vacuum to about 10^^Torr or 10^^Torr. Generally, a relatively high vacuum (having low pressures) is desirable to ensure high power components of a beam system, e.g., accelerator 40, do not arc during operation. Moreover, a relatively high vacuum, e.g., of about 10^^Torr or 10^^Torr, allows a charged beam to be accelerated, e.g., by accelerator 40, to a desired energy with little or no scattering with stray atoms. Scattering can cause diffusion and defocusing of the charged particle beam, as well as generation of unwanted SSPs. In more detail, channel assembly 300 is an enclosure that forms a respective channel about the first axis 305-1, the second axis 305-2, and the circular arc 306. Each channel provides a passage for a charged particle beam along the first axis 305-1, the second axis 305-2, or the circular arc 306 while under a vacuum. As used herein, a channel refers to a distinct path a charged beam can traverse through channel assembly 300 in the presence or absence of a magnetic field, where the channel extends between a pair of openings. A first channel extends from a first opening 310-1 to a second opening 310-2 and provides a passage for the charged beam along circular arc 306. The first channel includes: (i) Attorney Docket No.57383-0012WO1 a first linear portion 302-1 extending from first opening 310-1 along first axis 305-1, (ii) a second linear portion 302-2 extending to second opening 310-2 along second axis 305-2, and (iii) an arcuate portion 303 between the first 302-1 and second 302-2 linear portions, where the arcuate portion 303 extends along the circular arc 306. A second channel extends from first opening 310-1 to a third opening 310-3 and provides a passage for the charged beam along first axis 305-1. The second channel includes: (i) the first linear portion 302-1, and (ii) a third linear portion 302-3 extending to third opening 310-3 along first axis 305-1. A third channel extends from a fourth opening 310-4 to second opening 310-2 and provides a passage for the charged beam along second axis 305-2. The third channel includes: (i) a fourth linear portion 302-4 extending from fourth opening 310-4 along second axis 305- 2, and (ii) the second linear portion 302-2. Note, the channels are not completely separated as they intersect with one another and extend between common openings 310. As shown in FIGs. 5A and 5B, the configuration of the channels is somewhat different between the two example channel assemblies 300A and 300B although they perform the same function. First channel assembly 300A is generally simpler to fabricate than second channel assembly 300B and therefore can be used when cost is a constraint. Conversely, second channel assembly 300B is generally more compact and structurally sound than first channel assembly 300A and therefore can be used when space and performance are constraints. For the first channel assembly 300A depicted in FIG. 5A, channel assembly 300A encloses an area 330 bounded between the first axis 305-1, second axis 305-2, and circular arc 306. At least a portion of each channel is supported in area 330. In this case, the arcuate portion 303 of the first channel extends along the circular arc 306 and is bounded between: (i) a wall 301 of channel assembly 300A that forms an inner radius of arcuate portion 303, and (ii) a support wall 325 positioned in area 330 that forms an outer radius of arcuate portion 303. Multiple support columns 324 are also positioned in area 330 to provide structural support for channel assembly 300A when a vacuum is generated. The second channel is in area 330 between first linear portion 302-1 and third linear portion 302-3. Similarly, the third channel is in area 330 between fourth linear portion 302-4 and second linear portion 302-2. For the second channel assembly 300B depicted in FIG. 5B, the first, second, and third channels are substantially separated as area 330 is absent. In this case, the arcuate portion 303 of the first channel extends along the circular arc 306 and is bounded between: (i) a first wall 301-1 of channel assembly 300B that forms an inner radius of arcuate portion 303, Attorney Docket No.57383-0012WO1 and (ii) a second wall 301-2 of channel assembly 300B that forms an outer radius of arcuate portion 303. Here, arcuate portion 303 has a width of ^ between first 301-1 and second 301-2 walls. In some implementations, the width ^ of arcuate portion 303 can be about 40 mm to 100 mm. The second channel includes a first intermediate linear portion 314-1 between the first 302-1 and third 302-3 linear portions, where the first intermediate linear portion 314-1 extends along the first axis 305-1. The third channel includes a second intermediate linear portion 314-2 between the second 302-2 and fourth 302-4 linear portions, where the second intermediate linear portion 314-2 extends along the second axis 305-2. In some implementations, the second and / or third channels of channel assembly 300 can also provide a passage for an optical beam which allows measurements of the charged particle beam and / or other components of a beam system to be performed. For example, for charged beams received at first opening 310-1 and output at second opening 310-2 (and / or third opening 310-3), a beam tube configured with an optical camera can be connected to fourth opening 310-4. The optical camera can direct an optical beam along the second axis 305-2 to measure the charged beam and / or directly inspect a target. A respective connector 312 is positioned at each opening 310 to allow connection of a beam tube or other beam optics components. For example, each connector 312 can be a flange, e.g., a vacuum flange, to provide an appropriate vacuum seal. For ease of design, openings 310 are depicted herein with the same size and shape. For example, each opening 310 can have a circular or rectangular shape with a same diameter or lateral dimension ^. In some implementations, the diameter or lateral dimension ^ of the openings 310 is in a range between about 30 mm to 80 mm. However, in other implementations, openings 310 may have different sizes and shapes, e.g., polygonal shapes or elliptical shapes. Threads (or holes) 322 are arranged about channel assembly 300 to allow positioning of channel assembly 300 within DMA 200, or attachment of other components to channel assembly 300, e.g., a charged particle beam dump. While the described examples include three channels, channel assemblies 300 having more or less channels can also be utilized in DMA 200, such as channel assemblies 300 having one, two, four, five, six, or more channels. As one example, in some implementations, channel assembly 300 may include only the first channel, only the first and second channels, or only the first and third channels. As another example, in some implementations, channel assembly 300 can be configured for multiple different curved trajectories of a charged particle beam, e.g., corresponding to multiple different strengths of a magnetic field. In this Attorney Docket No.57383-0012WO1 case, channel assembly 300 may include multiple curved channels that extend along respective circular arcs of differing radii. Such situations can be appropriate when a DMA 200 is configured to direct a charged particle beam at multiple different bending angles. FIGs. 6A-6B are cross-sectional views depicting an example DMA 200A that includes a dipole magnet 205 configured according to FIGs. 4A-4C and a channel assembly 300A configured according to FIG. 5A. FIG. 6A is a cross-sectional view of DMA 200A along a horizontal bisecting plane (horizontal plane 214-1). FIG. 6B is a cross-sectional view (Section A-A’) of DMA 200A along a vertical bisecting plane. Channel assembly 300A is positioned in horizonal plane 214-1 within cavity 213, between first 220-1 and second 220-2 coils. Openings 310 are aligned with corresponding axes 305 and located exterior to dipole magnet 205 for connection to beam tubes (or other beam optics components) at respective connectors 312. Hence, when DMA 200 is secured in a beamline, channel assembly 300A can receive a charged particle beam along first axis 305- 1 at first opening 310-1. The charged beam then follows a trajectory 340 along circular arc 306 due to interacting with the uniform magnetic field ^ generated by dipole magnet 205. Since dipole magnet 205 has a geometry that coincides with circular arc 306, the uniform magnetic field is generated between pole faces 212 of back iron 210 within the entire arcuate portion 303 of channel assembly 300A. Therefore, the charged beam follows the trajectory 340 through the first channel of channel assembly 300A. The charged beam is thereafter output along second axis 305-2 at second opening 310-2. If the uniform magnetic field is not applied by dipole magnet 205, the charged beam instead travels through the second channel of channel assembly 300A and is thereafter output along first axis 305-1 at third opening 310- 3. Nevertheless, a charged particle beam is generally not composed entirely of PSPs of the same species of particle having the same kinetic energy. As mentioned above, a non- negligible portion of the charged beam can include SSPs of different species and / or having different kinetic energies than the PSPs. To reiterate a pertinent example, in a tandem accelerator that is configured to accelerate a hydrogen ion beam to produce a proton beam, the proton beam can include about 90 % PSPs (protons with the same kinetic energy) and 10 % SSPs (including protons with about half the kinetic energy of the PSPs) due to the charge- exchange process being about 90 % efficient. In the presence of the uniform magnetic field ^ = ±^^̂, these SSPs will exhibit cyclotron orbits with differing radii than the PSPs and therefore follow altered trajectories, e.g., trajectory 341. Attorney Docket No.57383-0012WO1 More precisely, for a given state ^ of a particle with absolute momentum!= |#!| and charge ^!, the radius $!of the particle’s cyclotron orbit is given by: 1 '!(1) where ' = / ^ is the m !! !omentum- are indexed by the particle state ^ = 0 and SSPs are indexed by particle states ^ ≠ 0. Note, for non-relativistic particles, the absolute momentum!may also be characterized relative the mass +!and the absolute velocity ,!= |^!| (or equivalently the kinetic energy -!) of the particle as!= +!,!= .2+!-!. For ease of description, it is assumed that the particles of a charged beam are in the non-relativistic limit ,! / 0 ≪ 1, where 0 is the speed of light. Nonetheless, the example DMAs 200, dipole magnets 205, and channel assemblies 300 described herein can also be utilized for charged particle beams composed of relativistic particles, e.g., having momentum!= +!2!,!and kinetic energy -!= (2!− 1)+!05, where 2!= Lorentz factor for a particular particle state. Generally, dipole magnet 205 generates the uniform magnetic field ^ such that the radius of circular arc 306 coincides with the radius of the PSPs cyclotron orbit ^ = $7. Hence, dipole magnet 205 generates the uniform magnetic field with a magnitude calibrated to ^ = '7 / ^. However, for the same magnetic field amplitude ^, the cyclotron orbits $!for SSPs deviate from that of the PSPs. This is because the SSPs have different momentum-to- charge ratios than the PSPs '!≠ '7. For example, for SSPs of the same species as the PSPs but having about half the kinetic energy -!≈ -7 / 2, their corresponding cyclotron radius is about $!≈ ^ / √2. As mentioned above, such SSPs will follow altered trajectories (e.g., trajectory 341) and thus collide with the inner surfaces of channel assembly 300A residing in strike-zone 351. Additional strike-zones 352, 353, and 354 indicate other possible locations of SSP collisions with inner surfaces of channel assembly 300A. Strike-zone 352 can include collisions from SSPs of the same species as the PSPs but having about three quarters the kinetic energy of the PSPs, e.g., such that $!≈ ^√3 / 2. Strike-zone 353 can include collisions from SSPs having the same charge and kinetic energy as the PSPs but having about one and half times the mass as the PSPs, e.g., such that $!≈ ^.3 / 2. Strike-zone 354 can include collisions from SSPs having opposite charge than the PSPs, which are bent in an opposite (in this case clockwise) direction under the same magnetic field. A high rate of SSP collisions at Attorney Docket No.57383-0012WO1 any of strike-zones 351-354 can damage channel assembly 300A, e.g., due to excessive heat generation and / or ion implantation, which may cause failure of a beam system, e.g., due to destroying or contaminating the vacuum supported by channel assembly 300A. However, SSPs incident at strike-zones 351 and 352 are generally most prevalent, particularly for beam systems implementing tandem accelerators that convert a negatively charged particle beam to a positively charged particle beam since this conversion is not 100 % efficient. Therefore, the wall 301 extending along the arcuate portion 303 of channel assembly 300A is highly susceptible to degradation. As described below with reference to FIGs. 7A-9C, wall 301 can be replaced (or configured) with a charged particle beam dump to absorb these incident SSPs, which can significantly extend the lifecycle of channel assembly 300A as well as mitigate safety concerns. While the example charged particle beam dumps are described in the context of mitigating SSP collisions in strike-zones 351 and 352, the beam dumps are not limited to such. The charged particle beam dumps can also be configured for strike-zones 353 and 354, as well as other configurations of channel assemblies 300, e.g., having a different number of channels (e.g., one, two, four, five, six, or more channels). In any of these cases, inner surfaces of channel assembly 300 that exhibit a high prevalence of SSP collisions can be configured with the example beam dumps described herein to mitigate degradation and improve vacuum integrity. Examples of Dipole Magnet Assemblies with Integrated Charged Particle Beam Dumps FIGs. 7A-7C are various views relating to an example DMA 200X that includes a dipole magnet 205 configured according to FIGs. 4A-4C and a channel assembly 300X having a releasable charged particle beam dump 360X. FIG. 7A is a perspective view of channel assembly 300X. FIG. 7B is a section view of channel assembly 300X. FIG. 7C is a section view of DMA 200X. Here, the wall 301 extending along the arcuate portion 303 of the first example channel assembly 300A has been replaced with beam dump 360X to absorb SSP collisions in strike-zones 351 and 352. Beam dump 360X can be removed and thereafter inspected, modified, or replaced if beam dump 360X experiences significant degradation from SSP collisions. Beam dump 360X therefore provides a high degree of modularity to channel assembly 300X. For example, beam dump 360X may be replaced without uninstalling channel assembly 300X from a beamline. Attorney Docket No.57383-0012WO1 In more detail, channel assembly 300X is an enclosure that is formed by a housing 361 and beam dump 360X. Beam dump 360X is releasably securable to housing 361. Housing 361 has the same configuration as the first example channel assembly 300A, absent the wall 301. Thus, in this example, housing 361 is a weldment, e.g., a vacuum weldment, that can be fabricated with the same welding techniques and made of the same materials (e.g., non-magnetic metals) as those described above for channel assemblies 300A and 300B. Beam dump 360X includes a support piece 362 that is configured to absorb SSPs and distribute a generated heat load, e.g., such that a temperature profile is fairly uniform within support piece 362. In this example, support piece 362 can be made of the same or different type of metal than housing 361 since beam dump 360X is not welded in place. Although different types of metals can be welded together, this generally involves more complicated welding techniques and may compromise the vacuum supported by channel assembly 300X. As one example, support piece 362 can be made of a non-magnetic metal such as copper, a copper alloy, or a non-magnetic nickel alloy (e.g., an Inconel alloy) that is resistant to heat and has relatively high thermal conductivity, whereas housing 361 can be made of a different non-magnet metal such as an austenitic stainless steel. Support piece 362 is attached to a top piece 364-1 and a bottom piece 364-2 to provide appropriate alignment of support piece 362 when beam dump 360X is secured to housing 361. Top 364-1 and bottom 364-2 pieces can be composed of the same or different material as support piece 362 depending on the implementation. For example, if top 364-1 and bottom 364-2 pieces are welded to support piece 362, they can be composed of the same non-magnetic metal as support piece 362. Beam dump 360X can be releasably secured to housing 361 using releasable fasteners attached to corresponding threads (or holes) 322 of housing 361. For example, the releasable fasteners can include bolts, e.g., vacuum bolts, that provide an appropriate vacuum seal when beam dump 360X is secured to housing 361. Similar to housing 361 and support piece 362, the releasable fasteners can be made of a non- magnetic metal such as any of those described herein. When beam dump 360X is secured, support piece 362 forms a wall of channel assembly 300X that extends along an inner radius of the arcuate portion 303. An inner surface 363-1 of support piece 362 faces the circular arc 306 and thus the expected direction of incoming SSPs. In this example, support piece 362 is flat such that the inner surface 363-1 has a surface normal at relatively equal angles with the first 305-1 and second 305-2 axes. However, in other implementations, support piece 362 can also be curved to follow the curvature of the circular arc 306, similar to the example beam dump 360Y of FIGs. 8A-8D. Attorney Docket No.57383-0012WO1 Here, the inner surface 363-1 and an outer (opposite) surface 363-2 of support piece 362 are substantially parallel to each other. In some implementations, support piece 362 can have a thickness : between the inner 363-1 and outer 363-2 surfaces of about 10 mm to 50 mm. In some implementations, the inner surface 363-1 of support piece 362 supports a layer of refractory metal that is highly resistant to heat and wear. This can significantly increase the lifecycle of beam dump 360X. For example, the layer of refractory metal can protect the support piece 362 from high velocity SSP collisions. The layer of refractory metal can be applied to the inner surface 363-1 of support piece 362 using a variety of techniques such as 3D printing techniques, explosion bonding (or welding) techniques, deposition techniques (e.g., sputtering or electroplating), among others. In some implementations, the layer of refractory metal has a thickness in a range from about 1 mm to 10 mm. In general, the layer of refractory metal includes one or more refractory metal elements. Particularly, the layer of refractory metal can be composed of a refractory metal element in pure form or as an alloy that includes different compositions of one or more refractory metal elements (and other additives). The one or more refractory metal elements can include: niobium, molybdenum, tantalum, tungsten, rhenium, titanium, vanadium, chromium, zirconium, hafnium, ruthenium, rhodium, osmium, or iridium. In some examples, the layer of refractory metal is made of molybdenum or a molybdenum alloy (e.g., titanium- zirconium-molybdenum (TZM)). Molybdenum and TZM are high performance refractory metals that are exceedingly resistant to heat and wear, have high thermal conductivities, while still being cost effective compared to alternatives, e.g., tungsten and tungsten alloys. As shown in FIG. 7B, the outer surface 363-2 of support piece 362 can support cooling tubes 365 (e.g., configured as a cooling coil) for active cooling of beam dump 360X. The cooling tubes 365 are configured to conduct a coolant to reduce (or regulate) the temperature of beam dump 360X. Examples of coolant can include, but are not limited to, water, deionized water, liquified gases (e.g., liquid nitrogen), cooled gases, polyalkylene glycol (PAG), mineral oils (e.g., polyphenyl ether oils), nanofluids, among others. The cooling tubes 365 are generally made of highly conductive materials to provide adequate heat transfer between support piece 362 and the coolant. Such materials can include any of those listed above for support piece 362, e.g., non-magnetic metals such as copper or a copper alloy. One or more inlets and one or more outlets of the cooling tubes 365 can be attached to a coolant pump to circulate the coolant through the cooling tubes 365. Circulation of coolant may be based on a thermal management system that is configured to regulate the temperature of beam dump 360X within a certain range. For example, the thermal Attorney Docket No.57383-0012WO1 management system can include a temperature sensor (e.g., an infrared thermometer) to monitor the temperature of beam dump 360X and use it as feedback to control the flow velocity and / or input temperature of the coolant. When installed, the cooling tubes 365 are in direct thermal contact with support piece 362, thus, the cooling tubes 365 can remove heat from support piece 362 by convective heat transfer. This can significantly improve the lifecycle and performance of beam dump 360X, e.g., due to minimizing radiative heat transfer that can cause the accumulation of hot spots. The cooling tubes 365 can be brazed to the outer surface 363-2 or pressure loaded to the outer surface 363-2 with a thermal interface material (TIM) therebetween. Preferably, the TIM is of relatively high thermal conductivity to provide improved thermal contact between the cooling tubes 365 and support piece 362. In some implementations, the TIM has a thermal conductivity of about 10 W / mK or more, e.g., about 11 W / mK, about 12 W / mK or more, about 13 W / mK or more, about 14 W / mK or more, about 15 W / mK or more, about 20 W / mK or more. Examples of TIMs can include one or more inorganic particulate fillers, e.g., aluminum oxide, magnesium oxide, aluminum nitride, boron nitride, and diamond powder. Examples of TIMs can also include one or more metal fillers, e.g., silver and indium alloys. While cooling tubes 365 are depicted in FIG. 7B, in some implementations, beam dump 360X can alternatively include a cooling channel formed in the outer surface 363-2 of support piece 362, similar to the example beam dump 360Y of FIGs. 8A-8D. Further details of the example beam dump 360Y are provided below. FIGs. 8A-8D are various views relating to an example DMA 200Y that includes a dipole magnet 205 configured according to FIGs. 4A-4C and a channel assembly 300Y having a welded-on charged particle beam dump 360Y. FIG. 8A is a perspective view of channel assembly 300Y. FIG. 8B is a section view of channel assembly 300Y. FIG. 8C is a perspective view of charged particle beam dump 360Y. FIG. 8D is a section view of DMA 200Y. Here, the wall 301 extending along the arcuate portion 303 of the first example channel assembly 300A has been replaced with beam dump 360Y to absorb SSPs in strike- zones 351 and 352. Channel assembly 300Y is configured similarly to the previous example of channel assembly 300X depicted in FIGs. 7A-7C, except beam dump 360Y is welded in place as opposed to being releasable. Particularly, channel assembly 300Y is an enclosure formed from housing 361 and beam dump 360Y. Beam dump 360Y includes a support piece 362 that is configured to absorb SSPs and distribute a generated heat load, e.g., such that a temperature profile is fairly uniform within Attorney Docket No.57383-0012WO1 support piece 362. In this example, support piece 362 is generally composed of the same non- magnetic metal as housing 361 since support piece 362 is welded to housing 361. As mentioned above, although different types of metals can be welded together, this generally involves more complicated welding techniques and may compromise the vacuum supported by channel assembly 300Y. Support piece 362 forms a wall of channel assembly 300Y that extends along an inner radius of the arcuate portion 303. An inner surface 363-1 of support piece 362 faces the circular arc 306 and therefore the expected direction of incoming SSPs. In this example, support piece 362 is curved such that the inner surface 363-1 follows circular arc 306 at a generally fixed distance. With a curved surface, support piece 362 has a generally compact form factor. However, in other implementations, support piece 362 can be flat similar to the first example beam dump 360X of FIGs. 7A-7C. Here, the inner surface 363-1 and an outer (opposite) surface 363-2 of support piece 362 are substantially parallel to each other. In some implementations, support piece 362 can have a thickness : between the inner 363-1 and outer 363-2 surfaces of about 10 mm to 50 mm. The inner surface 363-1 of support piece 362 supports a layer of refractory metal 370 that is highly resistant to heat and wear, which as mentioned above, can significantly improve the lifecycle of beam dump 360Y. The layer of refractory metal 370 can be applied to the inner surface 363-1 of support piece 362 using a variety of techniques, such as any of those listed above for the first example beam dump 360X. Similarly, the layer of refractory metal 370 can include any of the one or more refractory metal elements listed above for the first example beam dump 360X, e.g., in pure form or as an alloy. As one pertinent example, the layer of refractory metal 370 can be made of molybdenum or a molybdenum alloy (e.g., TZM). A cooling channel 366 is formed in the outer surface 363-2 of support piece 362 and is configured to conduct a coolant for active cooling of beam dump 360Y. Examples of coolant can include any of those listed above for the first example beam dump 360X. In some implementations, the cooling channel 366 has a depth into the outer surface 363-2 in a range of about 5 mm to 25 mm. For example, cooling channel 366 can be machined into the outer surface 363-2 of support piece 362, or otherwise formed by an appropriate fabrication method. A cover (not shown) can be disposed over cooling channel 366 to prevent leakage of coolant when it is circulated through cooling channel 366. Here, cooling channel 366 is connected to an inlet 367-1 to receive the coolant and an outlet 367-2 to output the coolant. The inlet 367-1 and outlet 367-2 can be attached to a Attorney Docket No.57383-0012WO1 coolant pump to circulate the coolant through cooling channel 366, which as described for the first example beam dump 360X, may be based on a thermal management system. Inlet 367-1 can also be configured as an outlet and outlet 367-2 can also be configured as an inlet depending on the desired direction of coolant flow through cooling channel 366. As shown most clearly in FIG. 8A, cooling channel 366 follows a tortuous path in the second surface 363-2 to maximize the duration and thermal contact of the coolant with support piece 362. Thus, cooling channel 366 can remove heat from beam dump 360Y by convective heat transfer and reduce the effects of radiative heat transfer. While a cooling channel 366 is depicted in FIGs. 8A-8D, in some implementations, beam dump 360Y can alternatively include one or more cooling tubes supported on the outer surface 363-2 of support piece 362, similar to the example beam dump 360X of FIGs. 7A-7C. FIGs. 9A-9C are various views relating to a DMA 200Z that includes a dipole magnet 205 configured according to FIGs. 4A-4C and a channel assembly 300Z having a port 384 for an insertable charged particle beam dump 360Z. FIG. 9A is a perspective view of channel assembly 300Z. FIG. 9B is a section view of channel assembly 300Z. FIG. 9C is a section view of DMA 200Z. Here, the wall 301 extending along the arcuate portion 303 of the first example channel assembly 300A is configured with beam dump 360Z to absorb SSPs in strike-zone 351. Beam dump 360Z can be removed and thereafter inspected, modified, or replaced if beam dump 360Z experiences significant degradation from SSP collisions. Beam dump 360Z therefore provides a high degree of modularity to channel assembly 300Z. For example, beam dump 360Z can be replaced without uninstalling channel assembly 300Z from a beamline. In this case, beam dump 360Z is an insert piece 380 that is configured to absorb SSPs and distribute a generated heat load, e.g., such that a temperature profile is fairly uniform within insert piece 380. Insert piece 380 can be made of a non-magnetic metal having a high thermal conductivity and a high melting point, which can include any of those described elsewhere herein. In some implementations, insert piece 380 is composed of a refractory metal that includes one or more refractory metal elements, e.g., in pure form or as an alloy, similar to the layer of refractory metal described above for beam dumps 360X and 360Y. In some implementations, insert piece 380 can have a thickness of about 5 mm to 25 mm. When inserted through the port 384 disposed on wall 301, insert piece 380 is positioned on an inner surface 302-1 of wall 301 that faces the circular arc 306. Hence, a surface 381 of insert piece 380 faces the expected direction of incoming SSPs. Insert piece 380 can be removed through port 384 for inspection, replacement, and / or modification. Port Attorney Docket No.57383-0012WO1 384 is welded to an outer (opposite) surface 302-2 of wall 301 which allows access to an interior of channel assembly 300Z when it is secured in a beamline. A flange 385, e.g., a vacuum flange, is mounted to port 384 and includes a wingnut 386 that can be loosened to access insert piece 380 and tightened to provide a vacuum seal for channel assembly 300Z. A cooling channel (not shown) is formed in a volume of insert piece 380 and is configured to conduct a coolant for active cooling of insert piece 380. Examples of coolant include any of those listed above for beam dumps 360X and 360Y. Thus, the cooling channel can remove heat from insert piece 380 by convective heat transfer. For example, insert piece 380 can be 3D printed with the embedded cooling channel using 3D metal printing techniques. A dual liquid feedthrough 382 is connected to insert piece 380 to supply and receive the coolant to and from the cooling channel. For example, the dual liquid feedthrough 382 can be a standard KF25 dual liquid feedthrough or other appropriately sized dual liquid feedthrough. When insert piece 380 is inserted into channel assembly 300Z through port 384, an inlet and outlet of dual liquid feedthrough 382 are external to channel assembly 300Z. Thus, the inlet and outlet of dual liquid feedthrough 382 can be attached to a coolant pump to circulate coolant through the cooling channel of insert piece 380 while insert piece 380 is positioned within channel assembly 300Z. In some implementations, the surface 381 of insert piece 380 can support a layer of refractory metal, e.g., if insert piece 380 is composed of copper or a copper alloy. The layer of refractory metal can be applied to the surface 381 of insert piece 380 using a variety of techniques, such as any of those listed above for beam dumps 360X and 360Y. As one example, if insert piece 380 is 3D printed, the layer of refractory metal can be printed onto its surface 381 during the same fabrication process. In a similar vein, the layer of refractory metal can include any of the one or more refractory metal elements listed above for beam dumps 360X and 360Y, e.g., in pure form or as an alloy. Additional Examples of Neutron Beam Systems FIGs. 10A-10C are schematic views of three example neutron beam systems 1000A, 1000B, and 1000C that are each configured for use in BNCT. In the described examples, a HEBL 50 of each beam system 1000 includes one or more DMAs 200 with integrated charged particle beam dumps 360. For instance, each DMA 200 can be configured according to any of the example DMAs 200X, 200Y, or 200Z of FIGs. 7A-9C. The neutron beam systems 1000 are examples of beam systems that can be utilized in a medical facility with multiple treatment rooms 400 for administration of BNCT to different patients 80. Attorney Docket No.57383-0012WO1 Referring to FIG. 10A, beam system 1000A is configured to create a charged particle beam 61 and propagate it toward one of twos targets 60-1 and 60-2, to generate a respective neutron beam 70-1 and 70-2. Neutron beams 70-1 and 70-2 are then directed towards respective patient bodies 80-1 and 80-2 to be irradiated. Patients 80-1 and 80-2 are situated in respective treatment rooms 400-1 and 400-2 that are proximate the targets 60-1 and 60-2. Beam system 1000A includes a charged particle source 20, a LEBL 30, an accelerator 40, and a HEBL 50. Source 20 is configured to generate the charged particle beam 61, which is output to LEBL 30. LEBL 30 is configured to transport the charged beam 61 from source 20 to accelerator 40. Accelerator 40 is configured to accelerate the charged particle beam 61 to a higher energy. As described elsewhere herein, in some implementations, accelerator 40 is a tandem accelerator that is configured to convert a negatively charged beam into a positively charged beam, e.g., a negative deuterium ion beam into a deuteron beam, or a negative hydrogen ion beam into a proton beam. HEBL 50 extends from the accelerator 40 to the targets 60-1 and 60-2 which can be housed in respective target assembly portions of HEBL 50, see FIGs. 2A-2C for example. HEBL 50 transfers the charged particle beam 61 from an output of accelerator 40 to one of the two targets 60-1 or 60-2, where it is converted to the corresponding neutron beam 70-1 or 70-2. In this example, HEBL 50 includes a DMA 200 positioned at a branch point between two branches 57-1 and 57-2. DMA 200 is configured to receive the charged particle beam 61 along a first axis 305-1, bend the charged particle beam 61 at an angle ^ by applying a uniform magnetic field with a corresponding magnitude, and output the charged beam 61 along a second axis 305-2 towards first target 60-1. Alternatively, the charged beam 61 is output along first axis 305-1 towards second target 50-2 when DMA 200 does not apply the magnetic field. Referring to FIG. 10B, beam system 1000B is configured similarly to the beam system 1000A of FIG. 10A but includes an additional DMA 200-2 positioned in HEBL 50, altering the direction of second branch 57-2. Here, second DMA 200-2 is configured to receive the charged particle beam 61 along the first axis 305-1 (that is output from first DMA 200-1), bend the charged particle beam 61 at a second angle ^5by applying a uniform magnetic field with a corresponding magnitude, and output the charged particle beam 61 along a third axis 305-3 toward second target 60-2. Referring to FIG. 10C, beam system 1000C is configured similarly to the beam system 1000B of FIG. 10B but includes an additional DMA 200-3 positioned in HEBL 50, forming a third branch 57-3. Here, third DMA 200-3 is configured to receive the charged Attorney Docket No.57383-0012WO1 particle beam 61 along the first axis 305-1 (that is output from second DMA 200-2), bend the charged particle beam at a third angle ^^by applying a uniform magnet field with a corresponding magnitude, and output the charged particle beam 61 along a fourth axis 305-4 toward a third target 60-3. In each of the example beam systems 1000A-1000C, SSPs are filtered from the charged particle beam 61 when the beam 61 is bent at any of the bending angles ^;, ^5, or ^^by a respective DMA 200-1, 200-2, or 200-3. Particularly, the SSPs are absorbed by a respective beam dump 360 of each DMA 200 which helps preserve the vacuum integrity of HEBL 50 by mitigating degradation from SSP collisions. While the beam systems 1000A- 1000C of FIGs. 10A-10C provide a few example configurations of DMAs 200 in a beamline, various other combinations of DMAs 200 can also be utilized that direct a charged beam 61 along multiple different trajectories to multiple different targets 60. FIG. 11 is a flow diagram of an example method 1100 for bending and filtering a charged particle beam using a DMA that includes a dipole bending magnet and a channel assembly configured with a charged particle beam dump. For instance, the DMA can be configured according to any of the example DMAs 200X, 200Y, or 200Z of FIGs. 7A-9C to implement the method 1100. Briefly, the method 1100 may begin with securing the DMA in a beamline of a beam system. For example, the method 1100 can include mounting the dipole magnet from a ceiling or other support structure via mounting elements and connecting the channel assembly to beam tubes via connectors positioned at respective openings. The method 1100 may further include evacuating air from the beamline, e.g., using one or more vacuum pumps, to generate a suitable vacuum within the channel assembly. The vacuum can have a pressure from about 10^^Torr to about 10^^Torr depending on the configuration of the beam system. The method 1100 may further include activating (or powering up) the beam system to generate, accelerate, and output the charged particle beam to the beamline, as described elsewhere herein. The method 1100 proceeds with propagating, along a first axis, the charged particle beam toward a first opening of the channel assembly (1102). The charged particle beam includes: (i) PSPs of a same species and having a same kinetic energy, and (ii) SSPs of different species and / or having different kinetic energies than the PSPs. For example, the charged particle beam can be a positively charged beam (e.g., a proton or deuteron beam) resulting from the conversion of a negatively charged beam (e.g., a negative hydrogen or Attorney Docket No.57383-0012WO1 deuterium ion beam) by a tandem accelerator. In the case of a proton beam, the PSPs are protons and the SPPs may include protons of different kinetic energies (e.g., roughly half to three quarters the kinetic energy of the PSPs), as well as other species of charged particles. The method 1100 continues with generating, using the dipole magnet, a uniform magnetic field within the channel assembly (1104). For example, the method 1100 can include conducting a respective current of a same magnitude and direction through each of a first and second coil of the dipole magnet. In general, the uniform magnetic field has a magnitude to cause: (i) the PSPs to follow a trajectory through a first channel of the channel assembly, and (ii) at least a portion of the SSPs to follow trajectories intersecting the charged particle beam dump. The first channel extends from the first opening to a second opening and includes an arcuate portion that extends along a circular arc. The magnitude of the uniform magnetic field is generally equal to a momentum-to-charge ratio of the PSPs divided by a radius of the circular arc. Since the magnitude of the uniform magnetic field is calibrated for the PSPs, the SSPs follow altered trajectories. As described elsewhere herein, the charge particle beam dump is configured to absorb at least the portion of the SPPs that follow trajectories intersecting the beam dump. The method 1100 continues with outputting, along a second axis, a filtered charged particle beam from the second opening of the channel assembly (1106). The filtered charged particle beam includes the PSPs. In some implementations, the method 1100 further includes circulating coolant through one or more cooling tubes or a cooling channel of the charged particle beam dump (1108). For example, the method 1100 can include activating a coolant pump to circulate the coolant, e.g., according to a thermal management system, which provides active cooling of the charged particle beam dump during operation of the beam system. As described elsewhere herein, heat is convected away from the beam dump via the coolant which can significantly extend the lifecycle of the beam dump as well as the operational time of the beam system. The method 1100 may further include deactivating (or powering down) the beam system and interrupting the vacuum. The method 1100 may further include removing, modifying, or replacing the charged particle beam dump and thereafter repeating the method 1100. In addition to the embodiments of the attached claims and the embodiments described above, the following numbered embodiments are also innovative. Attorney Docket No.57383-0012WO1 In a first group of embodiments, a channel assembly is provided, the channel assembly including: an enclosure including: (i) a housing, and (ii) a charged particle beam dump secured to the housing, where: the enclosure forms a channel extending from a first opening to a second opening, the channel including an arcuate portion, and the charged particle beam dump includes a support piece forming a wall of the enclosure that extends along at least some of the arcuate portion. In some embodiments of the first group, the charged particle beam dump further includes: a layer of refractory metal supported on a first surface of the support piece, the first surface facing the channel. The layer of refractory metal can include one or more of: niobium, molybdenum, tantalum, tungsten, rhenium, titanium, vanadium, chromium, zirconium, hafnium, ruthenium, rhodium, osmium, or iridium. For example, the layer of refractory metal can be made of molybdenum or a molybdenum alloy. The molybdenum alloy can be titanium-zirconium-molybdenum. The layer of refractory metal can have a thickness in a range from 1 mm to 10 mm. The layer of refractory metal can be explosion bonded to the first surface of the support piece. The layer of refractory metal can be printed on the first surface of the support piece. The layer of refractory metal can be deposited on the first surface of the support piece via sputtering or electroplating. The first surface of the support piece can be flat. The first surface of the support piece can be curved. In some embodiments of the first group, the charged particle beam dump further includes: one or more cooling tubes supported on a second, opposite surface of the support piece, the one or more cooling tubes configured to conduct a coolant. The one or more cooling tubes can be configured as a cooling coil. The one or more cooling tubes can be configured to convect heat away from the charged particle beam dump via the coolant. The one or more cooling tubes can be made of a non-magnetic metal. For example, the one or more cooling tubes can be made of an austenitic stainless steel, copper, or a copper alloy. The one or more cooling tubes can be brazed to the second surface of the support piece. The one or more cooling tubes can be pressure loaded to the second surface of the support piece with a thermal interface material therebetween. The thermal interface material can have a thermal conductivity of 10 W / (m K) or more. The thermal interface material can include aluminum oxide. In some embodiments of the first group, the charged particle beam dump further includes: a cooling channel formed in a second, opposite surface of the support piece, the cooling channel configured to conduct a coolant. The charged particle beam dump can further include: an inlet connected to the cooling channel to receive the coolant; and an outlet Attorney Docket No.57383-0012WO1 connected to the cooling channel to output the coolant. The cooling channel can be configured to convect heat away from the charged particle beam dump via the coolant. The cooling channel can follow a tortuous path in the second surface of the support piece. The cooling channel can have a depth into the second surface of the support piece of at least 8 mm. In some embodiments of the first group, the coolant is water. In some embodiments of the first group, the first and second surfaces of the support piece are parallel. In some embodiments of the first group, the support piece has a thickness between the first and second surfaces in a range between 10 mm and 50 mm. In some embodiments of the first group, the charged particle beam dump is releasably secured to the housing. The channel assembly can further include multiple releasable fasteners to releasably secure the charged particle beam dump to the housing. The releasable fasteners can include multiple bolts. The bolts can be vacuum bolts. In some embodiments of the first group, the charged particle beam dump is fixed to the housing. For example, the support piece can be welded to the housing. In some embodiments of the first group, the housing includes multiple metal pieces. The metal pieces can be welded together. In some embodiments of the first group, the housing and the support piece are made from one or more non-magnetic metals. In some embodiments of the first group, the housing is made of a same metal as the support piece. For example, the housing and the support piece can be made of an austenitic stainless steel, copper, or a copper alloy. In some embodiments of the first group, the housing is made of a different metal than the support piece. For example, the housing can be made of an austenitic stainless steel and the support piece can be made of copper or a copper alloy. In a second group of embodiments, a channel assembly is provided, the channel assembly including: an enclosure that forms a channel extending from a first opening to a second opening, the channel including an arcuate portion, where a wall of the enclosure extends along the arcuate portion; a port disposed on a first surface of the wall that is opposite the channel; and a charged particle beam dump that, when inserted through the port, is positioned on a second surface of the wall that faces the channel. For example, the port can be welded to the first surface of the wall. Attorney Docket No.57383-0012WO1 In some embodiments of the second group, the enclosure includes multiple metal pieces. The metal pieces can be welded together. In some embodiments of the second group, the enclosure is made of a non-magnetic metal. For example, the enclosure can be made of an austenitic stainless steel, copper, or a copper alloy. In some embodiments of the second group, the charged particle beam dump includes: an insert piece; and a cooling channel formed in a volume of the insert piece, the cooling channel configured to conduct a coolant. The cooling channel can be configured to convect heat away from the charged particle beam dump via the coolant. The coolant can be water. The charged particle beam dump can further include: a dual liquid feedthrough connected to the cooling channel, the dual liquid feedthrough including: (i) an inlet to receive the coolant, and (ii) an outlet to output the coolant. When the charged particle beam dump is inserted through the port, the inlet and outlet of the dual liquid feedthrough can be external to the enclosure. The insert piece can have a thickness in a range from 5 mm to 25 mm. In some embodiments, the insert piece has been three-dimensional (3D) printed. The insert piece can be made of a refractory metal. The piece can include one or more of: niobium, molybdenum, tantalum, tungsten, rhenium, titanium, vanadium, chromium, zirconium, hafnium, ruthenium, rhodium, osmium, or iridium. For example, the insert piece can be made of molybdenum or a molybdenum alloy. The molybdenum alloy cam be titanium-zirconium-molybdenum. The insert piece can be made of a non-magnetic metal. For example, the insert piece can be made of an austenitic stainless steel, copper, or a copper alloy. The charged particle beam dump can further include: a layer of refractory metal supported on a surface of the insert piece. The layer of refractory metal can include one or more of: niobium, molybdenum, tantalum, tungsten, rhenium, titanium, vanadium, chromium, zirconium, hafnium, ruthenium, rhodium, osmium, or iridium. For example, the layer of refractory metal can be made of molybdenum or a molybdenum alloy. The molybdenum alloy can be titanium-zirconium- molybdenum. The layer of refractory metal can have a thickness in a range from 1 mm to 10 mm. The layer of refractory metal can be printed on the surface of the insert piece. When the charged particle beam dump is inserted through the port, the surface of the insert piece can face the channel. In some embodiments of the second group, the channel assembly further includes a flange mounted to the port. The flange can include a wingnut configured to seal the port. Attorney Docket No.57383-0012WO1 In some embodiments of the first and second groups, the channel further includes: a first linear portion extending to the first opening along a first axis; and a second linear portion extending to the second opening along a second axis that intersects the first axis, where the arcuate portion is between the first and second linear portions, the arcuate portion extending along an arc having the first and second axes as tangents. The arc can be a circular arc. The circular arc can subtend an angle in a range from 10 degrees to 170 degrees. For example, the circular arc can subtend an angle of 45 degrees, 90 degrees, or 135 degrees. In some embodiments of the first and second groups, the channel is a first channel, and the enclosure forms a second channel extending from the first opening to a third opening, the second channel including: the first linear portion; and a third linear portion extending to the third opening along the first axis. In some embodiments of the first and second groups, the enclosure forms a third channel extending from a fourth opening to the second opening, the third channel including: a fourth linear portion extending from the fourth opening along the second axis; and the second linear portion. In some embodiments of the first and second groups, each opening has a same size and shape. For example, each opening can have a circular shape or a rectangular shape. Each opening can have a maximum lateral dimension in a range from 30 mm to 80 mm. In some embodiments of the first and second groups, the channel assembly further includes: a respective connector positioned at each opening and configured to secure the channel assembly in a beamline. Each connector can be a flange. For example, each flange can be a vacuum flange. In some embodiments of the first and second groups, the channel assembly is secured in the beamline and configured to receive a charged particle beam along the first or second axis. The channel assembly can be configured to support a vacuum for the charged particle beam. The charged particle beam can include: (i) primary-state particles of a same species and having a same kinetic energy; and (ii) secondary-state particles of different species and / or having different kinetic energies than the primary-state particles. The charged particle beam dump can be configured to absorb at least a portion of the secondary-state particles. In some examples, the primary-state particles are protons. The secondary-state particles can include protons. In a third group of embodiments, a dipole magnet assembly is provided, the dipole magnet assembly including: the channel assembly of any embodiment of the first or second Attorney Docket No.57383-0012WO1 groups; and a dipole magnet configured to generate a magnetic field within the channel assembly. In some embodiments of the third group, the magnetic field is a uniform magnetic field. The uniform magnetic field can have a direction orthogonal to the first and second axes. In some examples, the uniform magnetic field is generated within the arcuate portion of the first channel. In some embodiments of the third group, the dipole magnet includes: a first coil; a second coil; and a back iron enclosing the first and second coils. The channel assembly can be positioned between the first and second coils. In a fourth group of embodiments, a beam system is provided, the beam system including: an accelerator; a target; and a beamline extending from the accelerator to the target, the beamline including the dipole magnet assembly of any embodiment of the third group of embodiments. The beam system can be configured for use in boron neutron capture therapy (BNCT). In a fifth group of embodiments, a method for bending and filtering a charged particle beam using the dipole magnet assembly of any embodiment of the third group is provided, the method including: propagating the charged particle beam toward the first opening of the channel assembly, the charged particle beam including: (i) primary-state particles of a same species and having a same kinetic energy; and (ii) secondary-state particles of different species and / or having different kinetic energies than the primary-state particles; generating, using the dipole magnet, the magnetic field within the channel assembly to cause: (i) the primary-state particles to follow a trajectory through the first channel of the channel assembly; and (ii) at least a portion of the secondary-state particles to follow trajectories intersecting the charged particle beam dump of the channel assembly; and outputting a filtered charged particle beam from the second opening of the channel assembly, the filtered charged particle beam including the primary-state particles. In some embodiments of the fifth group, the charged particle beam dump is configured to absorb at least the portion of the secondary-state particles. In some embodiments of the fifth group, a magnitude of the uniform magnetic field is equal to a momentum-to-charge ratio of the primary-state particles divided by a radius of the circular arc. In some embodiments of the fifth group, the primary-state particles are protons. In some embodiments of the fifth group, the secondary-state particles include protons. Attorney Docket No.57383-0012WO1 In some embodiments of the fifth group, generating, using the dipole magnet, the magnetic field within the channel assembly includes: conducting a respective current through each of the first and second coils of the dipole magnet. In some examples, the respective current through each of the first and second coils has a same magnitude and a same direction. In some embodiments of the fifth group, the method further includes: circulating coolant through the one or more cooling tubes of the charged particle beam dump. In some embodiments of the fifth group, the method further includes: circulating coolant through the cooling channel of the charged particle beam dump. While this specification contains many specific implementation details, these should not be construed as limitations on the scope of what is being claimed, which is defined by the claims themselves, but rather as descriptions of features that may be specific to particular embodiments of particular inventions. Certain features that are described in this specification in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination. Moreover, although features may be described above as acting in certain combinations and even initially be claimed as such, one or more features from a claimed combination can in some cases be excised from the combination, and the claim may be directed to a subcombination or variation of a subcombination. Similarly, while operations are depicted in the drawings and recited in the claims in a particular order, this by itself should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. In certain circumstances, multitasking and parallel processing may be advantageous. Moreover, the separation of various system modules and components in the embodiments described above should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products. Particular embodiments of the subject matter have been described. Other embodiments are within the scope of the following claims. For example, the actions recited in the claims can be performed in a different order and still achieve desirable results. As one example, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results.

Claims

Attorney Docket No.57383-0012WO1 What is claimed is:

1. A channel assembly, comprising: an enclosure comprising: (i) a housing, and (ii) a charged particle beam dump secured to the housing, wherein: the enclosure forms a channel extending from a first opening to a second opening, the channel comprising an arcuate portion, and the charged particle beam dump comprises a support piece forming a wall of the enclosure that extends along at least some of the arcuate portion.

2. The channel assembly of claim 1, wherein the charged particle beam dump further comprises: a layer of refractory metal supported on a first surface of the support piece, the first surface facing the channel.

3. The channel assembly of claim 2, wherein the layer of refractory metal comprises one or more of: niobium, molybdenum, tantalum, tungsten, rhenium, titanium, vanadium, chromium, zirconium, hafnium, ruthenium, rhodium, osmium, or iridium.

4. The channel assembly of claim 3, wherein the layer of refractory metal is made of molybdenum or a molybdenum alloy.

5. The channel assembly of claim 4, wherein the molybdenum alloy is titanium- zirconium-molybdenum.

6. The channel assembly of any one of claims 2-5, wherein the layer of refractory metal has a thickness in a range from 1 mm to 10 mm.

7. The channel assembly of any one of claims 2-6, wherein the layer of refractory metal is explosion bonded to the first surface of the support piece.

8. The channel assembly of any one of claims 2-6, wherein the layer of refractory metal is printed on the first surface of the support piece.Attorney Docket No.57383-0012WO1 9. The channel assembly of any one of claims 2-6, wherein the layer of refractory metal is deposited on the first surface of the support piece via sputtering or electroplating.

10. The channel assembly of any one of claims 2-9, wherein the first surface of the support piece is flat.

11. The channel assembly of any one of claims 2-10 wherein the first surface of the support piece is curved.

12. The channel assembly of any one of the preceding claims, wherein the charged particle beam dump further comprises: one or more cooling tubes supported on a second, opposite surface of the support piece, the one or more cooling tubes configured to conduct a coolant.

13. The channel assembly of claim 12, wherein the one or more cooling tubes are configured as a cooling coil.

14. The channel assembly of any one of claims 12-13, wherein the one or more cooling tubes are configured to convect heat away from the charged particle beam dump via the coolant.

15. The channel assembly of any one of claims 12-14, wherein the one or more cooling tubes are made of a non-magnetic metal.

16. The channel assembly of claim 15, wherein the one or more cooling tubes are made of an austenitic stainless steel, copper, or a copper alloy.

17. The channel assembly of any one of claims 12-16, wherein the one or more cooling tubes are brazed to the second surface of the support piece.

18. The channel assembly of any one of claims 12-16, wherein the one or more cooling tubes are pressure loaded to the second surface of the support piece with a thermal interface material therebetween.Attorney Docket No.57383-0012WO1 19. The channel assembly of claim 18, wherein the thermal interface material has a thermal conductivity of 10 W / (m K) or more.

20. The channel assembly of any one of claims 18-19, wherein the thermal interface material comprises aluminum oxide.

21. The channel assembly of any one of claims 1-11, wherein the charged particle beam dump further comprises: a cooling channel formed in a second, opposite surface of the support piece, the cooling channel configured to conduct a coolant.

22. The channel assembly of claim 21, wherein the charged particle beam dump further comprises: an inlet connected to the cooling channel to receive the coolant; and an outlet connected to the cooling channel to output the coolant.

23. The channel assembly of any one of claims 21-22, wherein the cooling channel is configured to convect heat away from the charged particle beam dump via the coolant.

24. The channel assembly of any one of claims 21-23, wherein the cooling channel follows a tortuous path in the second surface of the support piece.

25. The channel assembly of any one of claims 21-24, wherein the cooling channel has a depth into the second surface of the support piece of at least 8 mm.

26. The channel assembly of any one of claims 12-25, wherein the coolant is water.

27. The channel assembly of any one of claims 12-26, wherein the first and second surfaces of the support piece are parallel.

28. The channel assembly of claim 27, wherein the support piece has a thickness between the first and second surfaces in a range between 10 mm and 50 mm.Attorney Docket No.57383-0012WO1 29. The channel assembly of any one of the preceding claims, wherein the charged particle beam dump is releasably secured to the housing.

30. The channel assembly of claim 29, further comprising a plurality of releasable fasteners to releasably secure the charged particle beam dump to the housing.

31. The channel assembly of claim 30, wherein the releasable fasteners comprise a plurality of bolts.

32. The channel assembly of claim 31, wherein the bolts are vacuum bolts.

33. The channel assembly of any one of claims 1-29 wherein the charged particle beam dump is fixed to the housing.

34. The channel assembly of claim 33, wherein the support piece is welded to the housing.

35. The channel assembly of any one of the preceding claims, wherein the housing comprises a plurality of metal pieces.

36. The channel assembly of claim 35, wherein the plurality of metal pieces are welded together.

37. The channel assembly of any one of the preceding claims, wherein the housing and the support piece are made from one or more non-magnetic metals.

38. The channel assembly of claim 37, wherein the housing is made of a same metal as the support piece.

39. The channel assembly of claim 38, wherein the housing and the support piece are made of an austenitic stainless steel, copper, or a copper alloy.

40. The channel assembly of claim 37, wherein the housing is made of a different metal than the support piece.Attorney Docket No.57383-0012WO1 41. The channel assembly of claim 40, wherein: (i) the housing is made of an austenitic stainless steel, and (ii) the support piece is made of copper or a copper alloy.

42. A channel assembly, comprising: an enclosure that forms a channel extending from a first opening to a second opening, the channel comprising an arcuate portion, wherein a wall of the enclosure extends along the arcuate portion; a port disposed on a first surface of the wall that is opposite the channel; and a charged particle beam dump that, when inserted through the port, is positioned on a second surface of the wall that faces the channel.

43. The channel assembly of claim 42, wherein the port is welded to the first surface of the wall.

44. The channel assembly of any one of claims 42-43, wherein the enclosure comprises a plurality of metal pieces.

45. The channel assembly of claim 44, wherein the plurality of metal pieces are welded together.

46. The channel assembly of any one of claims 42-45, wherein the enclosure is made of a non-magnetic metal.

47. The channel assembly of claim 46, wherein the enclosure is made of an austenitic stainless steel, copper, or a copper alloy.

48. The channel assembly of any one of claims 42-47, wherein the charged particle beam dump comprises: an insert piece; and a cooling channel formed in a volume of the insert piece, the cooling channel configured to conduct a coolant.

49. The channel assembly of claim 48, wherein the cooling channel is configured to convect heat away from the charged particle beam dump via the coolant.Attorney Docket No.57383-0012WO1 50. The channel assembly of any one of claims 48-49, wherein the coolant is water.

51. The channel assembly of any one of claims 48-50, wherein the charged particle beam dump further comprises: a dual liquid feedthrough connected to the cooling channel, the dual liquid feedthrough comprising: (i) an inlet to receive the coolant, and (ii) an outlet to output the coolant.

52. The channel assembly of claim 51, wherein, when the charged particle beam dump is inserted through the port, the inlet and outlet of the dual liquid feedthrough are external to the enclosure.

53. The channel assembly of any one of claims 48-52, wherein the insert piece has a thickness in a range from 5 mm to 25 mm.

54. The channel assembly of any one of claims 48-53, wherein the insert piece has been three-dimensional (3D) printed.

55. The channel assembly of any one of claims 48-54, wherein the insert piece is made of a refractory metal.

56. The channel assembly of claim 55, wherein the insert piece comprises one or more of: niobium, molybdenum, tantalum, tungsten, rhenium, titanium, vanadium, chromium, zirconium, hafnium, ruthenium, rhodium, osmium, or iridium.

57. The channel assembly of claim 56, wherein the insert piece is made of molybdenum or a molybdenum alloy.

58. The channel assembly of claim 57, wherein the molybdenum alloy is titanium- zirconium-molybdenum.

59. The channel assembly of any one of claims 48-54, wherein the insert piece is made of a non-magnetic metal.Attorney Docket No.57383-0012WO1 60. The channel assembly of claim 59, wherein the insert piece is made of an austenitic stainless steel, copper, or a copper alloy.

61. The channel assembly of any one of claims 48-54, wherein the charged particle beam dump further comprises: a layer of refractory metal supported on a surface of the insert piece.

62. The channel assembly of claim 61, wherein the layer of refractory metal comprises one or more of: niobium, molybdenum, tantalum, tungsten, rhenium, titanium, vanadium, chromium, zirconium, hafnium, ruthenium, rhodium, osmium, or iridium.

63. The channel assembly of claim 62, wherein the layer of refractory metal is made of molybdenum or a molybdenum alloy.

64. The channel assembly of claim 63, wherein the molybdenum alloy is titanium- zirconium-molybdenum.

65. The channel assembly of any one of claims 61-64, wherein the layer of refractory metal has a thickness in a range from 1 mm to 10 mm.

66. The channel assembly of any one of claims 61-65, wherein the layer of refractory metal is printed on the surface of the insert piece.

67. The channel assembly of any one of claims 61-66, wherein, when the charged particle beam dump is inserted through the port, the surface of the insert piece faces the channel.

68. The channel assembly of any one of claims 42-67, further comprising a flange mounted to the port.

69. The channel assembly of claim 68, wherein the flange comprises a wingnut configured to seal the port.

70. The channel assembly of any one of the preceding claims, wherein the channel further comprises:Attorney Docket No.57383-0012WO1 a first linear portion extending to the first opening along a first axis; and a second linear portion extending to the second opening along a second axis that intersects the first axis, wherein the arcuate portion is between the first and second linear portions, the arcuate portion extending along an arc having the first and second axes as tangents.

71. The channel assembly of claim 70, wherein the arc is a circular arc.

72. The channel assembly of claim 71, wherein the circular arc subtends an angle in a range from 10 degrees to 170 degrees.

73. The channel assembly of claim 72, wherein the circular arc subtends an angle of 45 degrees, 90 degrees, or 135 degrees.

74. The channel assembly of any one of claims 70-73, wherein the channel is a first channel, and the enclosure forms a second channel extending from the first opening to a third opening, the second channel comprising: the first linear portion; and a third linear portion extending to the third opening along the first axis.

75. The channel assembly of claim 74, wherein the enclosure forms a third channel extending from a fourth opening to the second opening, the third channel comprising: a fourth linear portion extending from the fourth opening along the second axis; and the second linear portion.

76. The channel assembly of any one of claims 70-75, wherein each opening has a same size and shape.

77. The channel assembly of claim 76, wherein each opening has a circular shape or a rectangular shape.

78. The channel assembly of claim 77, wherein each opening has a maximum lateral dimension in a range from 30 mm to 80 mm.Attorney Docket No.57383-0012WO1 79. The channel assembly of any one of claims 70-78, further comprising: a respective connector positioned at each opening and configured to secure the channel assembly in a beamline.

80. The channel assembly of claim 79, wherein each connector is a flange.

81. The channel assembly of claim 80, wherein each flange is a vacuum flange.

82. The channel assembly of any one of claims 79-81, secured in the beamline and configured to receive a charged particle beam along the first or second axis.

83. The channel assembly of claim 82, configured to support a vacuum for the charged particle beam.

84. The channel assembly of any one of claims 82-83, wherein the charged particle beam comprises: (i) primary-state particles of a same species and having a same kinetic energy; and (ii) secondary-state particles of different species and / or having different kinetic energies than the primary-state particles.

85. The channel assembly of claim 84, wherein the charged particle beam dump is configured to absorb at least a portion of the secondary-state particles.

86. The channel assembly of any one of claims 84-85, wherein the primary-state particles are protons.

87. The channel assembly of claim 86, wherein the secondary-state particles comprise protons.

88. A dipole magnet assembly, comprising: the channel assembly of any one of claims 1-81; and a dipole magnet configured to generate a magnetic field within the channel assembly.Attorney Docket No.57383-0012WO1 89. The dipole magnet assembly of claim 88, wherein the magnetic field is a uniform magnetic field.

90. The dipole magnet assembly of claim 89 when also dependent on claim 71, wherein the uniform magnetic field has a direction orthogonal to the first and second axes.

91. The dipole magnet assembly of claim 90, wherein the uniform magnetic field is generated within the arcuate portion of the first channel.

92. The dipole magnet assembly of any one of claims 88-91, wherein the dipole magnet comprises: a first coil; a second coil; and a back iron enclosing the first and second coils.

93. The dipole magnet assembly of claim 92, wherein the channel assembly is positioned between the first and second coils.

94. A beam system, comprising: an accelerator; a target; and a beamline extending from the accelerator to the target, the beamline comprising the dipole magnet assembly of any one of claims 88-93.

95. The beam system of claim 94, configured for use in boron neutron capture therapy (BNCT).

96. A method for bending and filtering a charged particle beam using the dipole magnet assembly of any one of claims 88-93, the method comprising: propagating the charged particle beam toward the first opening of the channel assembly, the charged particle beam comprising: (i) primary-state particles of a same species and having a same kinetic energy; andAttorney Docket No.57383-0012WO1 (ii) secondary-state particles of different species and / or having different kinetic energies than the primary-state particles; generating, using the dipole magnet, the magnetic field within the channel assembly to cause: (i) the primary-state particles to follow a trajectory through the first channel of the channel assembly; and (ii) at least a portion of the secondary-state particles to follow trajectories intersecting the charged particle beam dump of the channel assembly; and outputting a filtered charged particle beam from the second opening of the channel assembly, the filtered charged particle beam comprising the primary-state particles.

97. The method of claim 96, wherein the charged particle beam dump is configured to absorb at least the portion of the secondary-state particles.

98. The method of any one of claims 96-97 when also dependent on claim 90, wherein a magnitude of the uniform magnetic field is equal to a momentum-to-charge ratio of the primary-state particles divided by a radius of the circular arc.

99. The method of any one of claims 96-98, wherein the primary-state particles are protons.

100. The method of claim 99, wherein the secondary-state particles comprise protons.

101. The method of any one of claims 96-100 when also dependent on claim 92, wherein generating, using the dipole magnet, the magnetic field within the channel assembly comprises: conducting a respective current through each of the first and second coils of the dipole magnet.

102. The method of claim 101, wherein the respective current through each of the first and second coils has a same magnitude and a same direction.

103. The method of any one of claims 96-102 when also dependent on claim 12, further comprising:Attorney Docket No.57383-0012WO1 circulating coolant through the one or more cooling tubes of the charged particle beam dump.

104. The method of any one of claims 96-102 when also dependent on claim 21 or 48, further comprising: circulating coolant through the cooling channel of the charged particle beam dump.

Images