Multiple innovations converge to enable practical, portable particle accelerators
August 1, 2020
“We tend to think of particle accelerators as these enormous facilities that take decades to build and are only used for research at the very edge of known physics,” says Los Alamos physicist Evgenya Simakov. “What people don’t realize is that smaller, lower-energy particle accelerators are needed all the time—for cancer treatments and medical sterilization, security screening and defense applications, and research into new materials, biological processes, and much more.
“These smaller accelerators exist,” she clarifies. “They’re just not small enough.”
Typically the size of a small room, the accelerators in question are extremely specialized and therefore expensive. As a result, their availability is limited. Often, they are only found at major metropolitan medical centers or large universities, for example. Simakov, however, has a new approach to dramatically cut size and cost, egalitizing access to these valuable tools and greatly expanding their uses.
For example, a small accelerator can be used to sterilize foods, similar to the pasteurization of milk: killing any bacteria and parasites but leaving the food itself unharmed. This could permanently eliminate most types of food poisoning and corresponding food product recalls, such as from E. coli in romaine lettuce. As things stand now, the scale of such an operation makes it thoroughly cost-prohibitive in most cases. Particle accelerators would be needed all over the place: at farms, distribution centers, grocery stores, and restaurants.
Simakov thinks, why not?
To build a beam
The basic premise for a small accelerator (small compared to the likes of Fermilab or CERN, say) goes like this: a stream of particles—electrons, in Simakov’s case—are pushed by an intense electromagnetic wave in a specialized conduit called a waveguide. The waveguide is what it sounds like; it’s a structure designed to channel waves. One familiar type is a fiber-optic cable, made from a type of dielectric plastic with optical properties that keep visible and infrared light trapped inside. Thus, the light travels down the cable, reflecting back into the plastic whenever it bumps against the side of the fiber, rather than leaking out.
But for an accelerator, the waveguide is not such a simple matter. For one thing, the light must drive the electron beam, meaning that both the light and the electrons must occupy the same channel. Light can propagate through optical fiber, but electrons can’t; they must be accelerated in a vacuum. Therefore, the waveguide has to be inverted: there must be an empty channel through the dielectric medium, with both the light waves and electrons confined within that empty region. That much is fairly straightforward to implement. However, there is another, more vexing constraint.
An electromagnetic wave carries oscillating electric and magnetic fields, the electric field being the one that a particle accelerator uses to accelerate its particles. As the light wave zigzags its way down the waveguide channel, its electric field points along various diagonals, which can be broken into lengthwise, up-down, and sideways components; the lengthwise component, directed through the channel, can usefully accelerate electrons. However, most optical fibers guide (reflect) waves with the opposite orientation: electric fields oscillating perpendicular to the channel only, with no lengthwise component. These are called transverse-electric, or TE, waves, as their electric fields contain no component directed along the waves’ direction of motion, but rather across it. Those with the forward-backward orientation suitable for accelerating particles, called transverse-magnetic, or TM, waves, are only confined in a special kind of fiber, called a photonic band gap (PBG) waveguide.
But even with a PBG waveguide—suitably hollowed to accommodate a colinear electron beam, of course—there’s another problem: Because the electric fields are oscillating as waves, their lengthwise components alternately push and pull on electrons. To make an accelerator, the electrons must be fired off in precisely-timed bunches that only appear in the electric field zones that push, not pull. That in turn means the speed at which the laser’s wave pattern, or phase, works its way down the length of the waveguide must be tailored to match the speed of the electrons—that is, with the laser light’s phase traveling slower than the light itself. That way, the wave phase pattern and the electrons travel together: electrons enter a “push phase” and stay with it all the way down the accelerator channel. Thus, the waveguide needs to reflect not just TM waves, but TM waves with the proper phase velocity. Ordinary PBG fiber material doesn’t accomplish this.
Until now, solving this problem has required a fairly serious concession: using microwaves instead of infrared or visible light. With microwaves, conducting waveguides—hollow metal ducts, essentially—will do the trick. The oscillation frequency for microwaves is slow enough that electrons in conductors are able to keep up and jiggle back and forth at the same frequency; this produces an effectively perfect reflection, so all the waves are kept inside the channel. The interior metal walls of a microwave oven, for example, reliably reflect outbound microwaves back into the food.
To miniaturize a functional accelerator, we needed to innovate on both the waveguide and the emitter.
But easily reflected low-frequency waves come at a cost. Microwaves have much larger wavelengths (centimeters or meters) than infrared (microns, or millionths of a meter) or visible light (fractions of a micron). To produce and channel such large waves, the waveguides and other necessary hardware, including the microwave source, must be sized in multiples of the wavelength, ranging from 10 centimeters to several meters. In large part, this is what forces “small” particle accelerators to be such large, cumbersome, and specialized machines. By contrast, infrared and visible-light waveguides—if they could be made to reflect TM waves—would be sized in mere microns. And infrared and visible-light lasers are not only vastly more compact than microwave sources; they are also vastly more powerful.
Mind the gap
So Simakov bucked the prevailing wisdom and set out to build an infrared-driven electron accelerator. If she could somehow invent a waveguide that reflects TM waves with a suitably slower-than-light phase velocity—well, that would be miracle number one. Miracle number two would be manufacturing waveguides and other tiny components with the tight tolerances required to obtain the phase-velocity match. If she could do all that, then the whole system would be both powerful and portable. It would serve the same range of applications as current microwave-based accelerators (and perhaps many others), while being easily carried around by hand, like a briefcase.
Right away, Simakov realized that 3D printing offered a means of manufacturing the tiny waveguides. The printers have the necessary micron-scale control, and with only minor modifications, 3D-print resin would probably be a suitable dielectric. The biggest challenge, she knew, would be to engineer the structure of the 3D-printed waveguide channel to reflect TM waves with the right phase velocity. For that, she decided to go back to basics.
While slower-frequency light, such as microwaves, can be reflected by the oscillating motion of electrons that are free to move inside conducting metals, higher-frequency light oscillates too fast for the electrons to match pace. Instead, what makes various materials reflect certain frequencies of infrared and visible light, the way a strawberry reflects red, lies in its molecular structure: the regular, repeating arrangement of atomic nuclei. The mathematical details of how this comes about are not particularly straightforward, but fundamentally, the nature and spacing of a series of tiny, distinct “cells” for electrons in the material to occupy, dictated by the repeating lattice of atomic nuclei, results in a pattern of allowed and disallowed electron energies. Allowed energy ranges, or bands, are separated by disallowed gaps.
When an electromagnetic wave strikes a material, the outcome depends on the energy of the wave, which is determined by its frequency. If the wave’s frequency corresponds to an energy within one of the material’s allowed energy bands, then the material can accommodate the wave passing through. If the frequency corresponds to one of the disallowed gaps, however, the wave is reflected back. Therefore, what Simakov needed to do was engineer a dielectric with some kind of regular, repeating pattern at the micron scale and adjust its pattern and spacing so that her infrared laser resides in the middle of a disallowed energy gap. Then with a little fine tuning, she could be sure to guide the waves she needs to guide: TM waves with the right phase velocity.
The accelerator system could easily be carried around by hand.
“We went about this by brute force,” Simakov explains. “At a scale too small to see by naked eye, our waveguides are made up of an alternating patchwork of 3D-print resin and empty space. In other words, we used precise physical gaps to make the precise energy gap.”
This approach worked wonders, with just one slight flaw: standard 3D-print resin isn’t quite up to the task. In order to give the resin the right PBG attributes, its optical properties would need a slight upgrade. So Simakov worked with materials scientist Robert Gilbertson, his postdoctoral researcher Ethan Walker, and others in the Los Alamos materials science and technology division to devise a solution. They decided to create a specialized nanoparticle infusion for the resin.
The effect they were after is similar to looking at the surface of a placid lake: Look straight down, or nearly so, and you’ll see what’s underwater; light crosses the water-air boundary. But look farther out, and you’ll see a reflection of the sky—that is, the light you see is kept on one side (the air side, in this case) of the boundary. The angle of incidence for the light striking an interface between two materials at which this shift from transmission to reflection occurs depends upon a property known as the index of refraction. For water, the index of refraction is 1.33, and for standard 3D-print resin, it’s about 1.2. Simakov calculates that she needs to get the resin’s index of refraction above 2, and so far, the researchers have tried an infusion of lead nanoparticles and achieved 1.98—close but no cigar. They also tried germanium nanoparticles and succeeded with 2.05, but germanium oxidizes in air and is difficult to work with, so it may be challenging to scale up the process. But Simakov believes tweaking the process for lead will ultimately work as well.
Diamond nanostructures are an accelerator scientist’s best friend
Overcoming the longstanding problem of channeling TM waves through a micron-scale waveguide is a major achievement. But that success brings with it a new challenge. A miniaturized waveguide requires a miniaturized emitter—the component that fires electrons into the waveguide.
“The wavelength of the laser determines the width of the waveguide,” Simakov says. “In turn, the width of the waveguide determines the size of the emitter.” Due to the electrons’ mutual electrostatic repulsion, they tend to spread out in flight. In practice, this means the emitter tip must be small enough that the beam expanding from it remains narrower than the waveguide channel when it reaches the waveguide. “Therefore, to miniaturize a functioning accelerator, we needed to innovate on both the waveguide and the emitter.”
The “we” Simakov refers to includes her Los Alamos accelerator-science colleague Heather Andrews, Simakov’s postdoctoral researcher Dongsung Kim, and other colleagues from the Los Alamos accelerator operations and technology division. Together, they sought a way to produce a strong electron-emitting material that could be fashioned into an extraordinarily narrow point. They were aware of a process pioneered by researchers at Vanderbilt University, and they were able to replicate and adapt it at Los Alamos to generate crystal-perfect diamond emitters.
The virtue of using diamond is its strength. The emitter must handle a very large current density—a large number of electrons funneled through its extraordinarily narrow tip—and a weaker material would literally melt. In fact, even diamond must be upgraded to accommodate the extreme current density; it is therefore doped with nitrogen for added conductivity.
In broad strokes, the process for making the emitters works like this: Pyramid-shaped holes are etched into silicon wafers and then filled with diamond nanocrystals that grow into a single, solid structure. Then the silicon is removed to uncover a sharply pointed diamond pyramid. The pyramid tip is only nanometers in size—a thousand times narrower than the waveguide opening—as required to accommodate the widening electron beam. An infrared laser, which can even be the same one that accelerates the electrons, provides the power source driving the emitter to eject electrons from its tip.
There will be amazing applications that don’t yet exist.
To assemble an actual accelerator system, an emitter and a series of waveguides are all aligned in a row. The acceleration occurs because the infrared laser transfers energy to the electrons, but in so doing, the laser itself loses energy. Therefore, it is important to inject additional laser light at the inlet of each waveguide. The more laser-boosted waveguides in the sequence, the greater the energy of the particles emerging in the final accelerator beam.
Spot on
Testing the electron beam a few centimeters off the cathode, Simakov produced a micron spot size with 40-kiloelectronvolt (keV) electron energies at a beam current of 50 nanoamps, or 50 billionths of an amp. (Normally amps are used to quantify electrical current in a wire or other device; for example, half an amp flows through a 60-watt light bulb. In the context of a particle accelerator, amps quantify the rate of charged-particle flow in the beam.) In assembling a complete accelerator system, both the current and electron energy would have to increase by a factor of 20–25, to about 1 microamp and 1 megaelectronvolt (MeV), respectively, for most practical applications.
“Increasing the energy means precisely stacking a series of accelerating waveguides, and the current is limited only by desired spot size, because the more electrons you have, the more they spread out in flight,” explains Simakov. Countering that will require some additional experimentation with devices called magnetic lenses, but it should be relatively straightforward. “What’s important here is that we’ve already shown that the emitters can handle up to around 10 milliamps, which is 10,000 times more current than we really need.”
In addition to seeking to tighten the focus of a beam from a single emitter, Simakov has been pursuing another approach to increase the current: simultaneously firing from a whole array of emitters and subsequently combining the beams. This approach is suitable for producing rapid-pulsing, high-energy electron beams in an alternate design known as a wakefield accelerator. Either way, the outcome is the same: compact, inexpensive particle accelerators available for widespread use.
“It’s an enabling technology that we’re pioneering here at Los Alamos,” says Simakov. Indeed, many existing applications—particularly in the medical, research, and national security arenas—will benefit tremendously from tabletop-accelerator technology. But according to Simakov, that’s only part of the story.
“I think there will also be amazing applications that don’t yet exist,” she adds. “I mean, there’s never been a particle accelerator you can carry around by hand. Not even close. But every time I look at my phone—an ingenious blend of computer, wireless communication, touch screen, camera, GPS receiver, and other components—I am reminded of just how much becomes possible whenever key technologies are miniaturized.” LDRD