In 2017, the New Mexico Environment Department needed help disposing of a barrel of radioactive waste that had been abandoned in a vacated commercial property. The Laboratory stepped up to help, and a request came in to my engineering group: was there a way to see inside the 80-gallon steel drum without opening it, so that a plan could be made to dispose of the contents properly?
“Yes!” we said. “We can definitely do that.”
We used computed tomography (CT), an x-ray–based imaging technique that provides detailed 3D images of an object. We knew that the contamination of concern was americium, a manmade radioactive metal that is used in smoke detectors and scientific instruments. But with the CT image, we were able to tell with confidence what sorts of items were in the barrel, how they were packaged, and—when combined with gamma imaging—where the hot spots were. A team was going to open the barrel to remove the items for proper disposal, and the images helped minimize their exposure to hazardous material. They now knew exactly what they were going to find and where.
CT imaging is more complicated than, say, the kind of x-ray imaging done at a dentist’s office, because CT collects views from 180 to 360 degrees around an object. This requires either the source and detector, or the item being examined, to rotate. Special computing resources to synchronize the data collection and digital tools to turn the data into 3D images are also needed. And in this case, the barrel was much larger than what is typically imaged by CT. But we could handle it, because earlier that same year, we just happened to have developed a technique to image a very different, but also oversized, barrel. That one contained a fossilized dinosaur skull.
The “Bisti Beast,” as the fossil is known, had been a tyrannosaur, an ancestor of Tyrannosaurus rex, who died in northwestern New Mexico roughly 74 million years ago. It was discovered in 1996 in the Bisti/De-na-zin Wilderness Area and taken to the New Mexico Museum of Natural History and Science.
The fossil was too big for medical imaging, and it had a mineral makeup that required higher-energy x-rays than what most medical facilities can provide. In partnership with the museum, the fossil was brought to Los Alamos, where resources in high-energy x-ray and neutron radiography suitable for an object of this size and composition are available. It was one of the largest things my group had ever imaged by CT, and we had to figure out how to handle both the sample itself and the large amount of data for image reconstruction.
My colleagues at the Los Alamos Neutron Science Center (LANSCE) also imaged the fossil with high-energy neutrons. The two imaging modes, x-rays and neutrons, revealed different, complementary information and provided highly detailed images without damaging the sample. It was the highest-resolution scan ever done of a tyrannosaur skull, and it provided insight into the cranial anatomy and sensory systems of the animal, helping to fill in some of the gaps in the evolutionary history of the family Tyrannosaurridae.
These are just two of the many cool projects that I feel lucky to have participated in at the Lab over the last 30 years. I’ve had the good fortune to be able to use a variety of imaging modalities, with applications to a wide assortment of different projects at Los Alamos, and I have developed a deep appreciation for how much we can learn from scientific imaging.
SQUID doctor
What drew me to physics is that you can use a few rules to explain just about everything. I like math, too, and I entered college thinking it might be my major, until I got into a class on proofs and realized I felt like I couldn’t prove anything. Math was too abstract for me. Physics was a better fit; it’s where math meets the real world.
I first visited Los Alamos in the summer of 1991, when I was on my way to graduate school at the University of Minnesota. Having grown up in Southern California, it is quite the understatement to say that I was unprepared for Minnesota weather. When, in the third year of my Ph.D. program, I was offered the chance to move to New Mexico, I enthusiastically accepted. I had completed my coursework, my advisor had collaborators at Los Alamos, and I was using data from the meson physics facility there (the precursor to today’s LANSCE). Proximally, I was chasing better weather, but I knew the move would be good for my education and career.
After completing my nuclear physics Ph.D. in 1996, I stayed in Los Alamos as a postdoc in the Applied Modern Physics group in the Physics Division. I started working on magnetoencephalography, or MEG, which is a method for detecting the very weak, naturally occurring magnetic signals within the human brain. MEG can reveal important information about brain activity and pathology, and the project was focused on the diagnosis and treatment of epilepsy.
The human brain’s magnetic field is more than a million times smaller than Earth’s magnetic field.
MEG relies on highly sensitive magnetometers called superconducting quantum interference devices (SQUIDs). Because the human brain’s magnetic field is so low, 10–1000 femtoteslas, which is a million or more times smaller than what one encounters just walking down the street, isolating the signal from the surrounding noise is one of the chief challenges for MEG. We worked both sides of the problem, trying to get our SQUIDs to be more sensitive while simultaneously trying to improve our noise shielding.
SQUIDs use the combined principles of superconductivity and quantum mechanics to detect tiny changes in magnetic flux. Because they include superconducting materials, SQUIDs typically must be kept exceptionally cold: just a few degrees above absolute zero. The epilepsy-detection device we built was a helmet, with 155 highly insulated, liquid-helium cooled SQUIDs on the inside, which was sensitive enough to detect the response of the human brain to external stimuli without touching the scalp. We also discovered that by using supercooled lead metal as a magnetic mirror, we could shield out most of the external magnetic noise, and by including reference SQUIDs on the outside of the helmet, we could gather the rest of the noise and subtract it from the measurement. We got MEG signals that were nearly noise-free.
Around the time I became a staff scientist in 1998, my team was also exploring the use of SQUIDs as detectors for ultra-low field magnetic resonance imaging (MRI). This got us thinking that we might be able to combine MEG with MRI. It would be a powerful combo because, while MEG can pinpoint the instant an epileptic seizure begins, the location is inexact. But precise location information is crucial for the surgical treatment of epilepsy, so a collocated MRI is needed to resolve the details of the surrounding soft tissue. But MEG couldn’t be combined with conventional MRI because the magnetic field used for MRI is far too strong for sensitive SQUIDs. Even moving a chair in the lab can ruin a SQUID measurement if one isn’t careful. So two separate, expensive measurements had to be taken and then somehow co-registered. But maybe, instead of the SQUIDs being a problem for conventional MRI, they could be the solution for very low-magnetic-field MRI.
MRI is an imaging technology that exploits the physical phenomenon of nuclear magnetic resonance (NMR). NMR relies on the observation that certain atoms can interact with, and be manipulated by, magnetic fields. Hydrogen is one of them, and MRI is most often done on hydrogen because it is fairly ubiquitous in the human body, primarily in water.
Conventional MRI involves magnetic fields up to several tesla in strength—strong enough to turn nearby ferromagnetic objects into projectiles if they get too close to the magnet bore. A hairpin, for example, can get pulled across an exam room and into the magnet at 40 miles per hour.
Ultra-low-field (ULF) NMR uses magnetic fields many times lower than what is typical of medical MRI, approaching fields as low as that of Earth, just 50 millionths of one tesla. In exploring ULF-MRI for collocated MEG, we developed a technique that uses pulsed magnetic fields from one-tenth of one tesla all the way down to 10 millionths of one tesla. The key to getting good images from these low fields was our supersensitive SQUIDs. We demonstrated, for the first time, that MEG and anatomical MRI could be acquired in the same instrument and the data combined into a single co-registered measurement.
Our airport screening technique can classify liquids as safe or unsafe in less than a minute.
Our success with ULF-MRI for the MEG collocation project led us to ask, what other problems might this solve? One of the answers was a field-hospital MRI for soldiers with potential traumatic brain injuries. Because of the strong magnets, robust noise shielding, and other permanent infrastructure involved, conventional MRI is limited in terms of where it can be done, and thus who it can be done on. We had the ULF part and the SQUID part squared away from our previous work, but we needed to find a field-deployable way of shielding a ULF-MRI from environmental magnetic noise. We found that instead of shielding, with something like the metal-clad rooms typical of medical MRI facilities, we could cancel the noise. We built a lightweight modular noise-cancellation system that emits signal at the same frequency but opposite phase from the noise—the same principle as noise-canceling headphones—so the noise is literally canceled out of the air surrounding our ULF-MRI machine.
There were several other advances made by my colleagues at the Lab, including using different types of sensitive detectors, such as atomic magnetometers, and a technique called pre-polarization to enhance the NMR signal. The pre-polarization concept has made its way into simple MRI machines, which are now being deployed in resource-poor locations. From this initial work, various technologies and applications have spun off, and I’m proud that I have helped to make MRI more accessible.
Branching out
Although medical imaging was a major driver of my ULF-NMR work, it was not the only one. During my time in the Physics Division, I also applied ULF-NMR to important non-medical problems. Two that come to mind are liquid screening for airport security and tree drought-tolerance studies for climate change adaptation.
After a foiled terrorist attack in 2006 that involved liquid explosives onboard a commercial flight, cumbersome new rules for liquids in carry-on luggage were enacted. We still see them today in the size restrictions on the liquids we are allowed to bring as carry-on items. Wouldn’t it be great, I thought, if there was a fast and reliable way to prove that a beverage was just a beverage?
We developed a screening technique called MagViz, which uses either ULF-NMR or ULF-MRI to classify liquids as safe or unsafe in less than a minute. Chemical-threat agents have certain chemical signatures that coffee and soda lack. It turns out that some of these signatures are really easy to detect using low magnetic fields. Based on that observation, we built a liquid-screening machine that works in conjunction with the x-ray detectors already in use at airports.
Our first public demonstration of MagViz in 2008 was particularly exciting for me. We had arranged a fairly high-profile demonstration at the international airport in Albuquerque, with people from the local media and the Department of Homeland Security (DHS) in attendance. By then I was the leader of the Lab’s SQUID team, so I was presenting the demonstration. I stood facing the audience, with the MagViz machine behind me, and I couldn’t see the green and red lights on the machine that would indicate a successful demo—green for a safe liquid, red for an unsafe one. I could only see the audience’s faces. I waited and smiled, gripped by the terror that our machine, which had worked flawlessly in preliminary tests, might somehow not work at the crucial moment. Then the crowd went “Ooooh!” and their faces lit up, and I knew that it had worked.
Although MagViz isn’t yet at every security checkpoint due to priority shifts within the DHS, I believe that subsequent iterations of the device remain among the fastest and most accurate ways to screen liquids. Spin-offs of MagViz are still capabilities at the Lab today, proving the strength of the approach for detection and classification of chemical threats.
In 2013, I was part of a project that used ULF-NMR to study water transport in plants. Because of how quickly and dramatically the climate is changing, scientists need to better understand how plants, trees, and forests transport and use water.
We used a portable ULF-NMR setup—many turns of wire wrapped directly around the trunk of a tree like a collar—to measure the water content in a cross section of the tree over the course of two weeks. Not only could we accurately measure water content at any given time, but we could also detect changes over time, such as day versus night, and during stress, such as when we temporarily induced drought. Some of these NMR measurements were combined with neutron measurements: by giving the plant deuterated water, that is, water containing an isotope of hydrogen that has an additional neutron, we altered the NMR signal and neutron cross section, effectively creating a tracer that could be visualized over time. In this way we were able to observe how various species of plants manage their water, and we could draw conclusions about why some species seem to fare better in the face of drought.
The change-up
In 2016, after 20 years of MEG, SQUIDs, NMR, and MRI, I made the radical-feeling choice to switch groups from physics to engineering. Physics had done nothing wrong—I still enthusiastically collaborate with colleagues from the physics division—but I was restless and looking for a change. Some of my projects were wrapping up, and others had brought me closer to engineering and the more mission-science side of the Lab. I had enjoyed working with colleagues in the Lab’s Nondestructive Testing & Evaluation group on MagViz, tree imaging, and other projects. So, when I learned there was a chance I could move there, I went for it.
It was a steep learning curve, and I will be eternally grateful to my colleagues for their patience and support as I found my feet in my new group. It was a lot like being a postdoc all over again. I was still working mainly on imaging, but it was now x-rays and radiography, instead of magnetic fields and MRI. Also, the work was less research-focused and more customer-focused, with a much shorter time frame.
At the time I made the switch, I was getting involved with a unique Los Alamos imaging modality, which had nothing to do with SQUIDs or NMR. Proton radiography, or pRad, is a technique first demonstrated at Los Alamos in 1995. Since then, it has been developed and refined here at the Lab and has become a scientific capability that brings scientists from around the world to our user facility.
The pRad project I was working on was trying to show that by using targeted gold nano-particles, one could achieve enough contrast for pRad imaging to guide proton-based cancer therapy. While we were tinkering with our collimation scheme, we realized we were perfectly set up to test a new application of pRad, now known as dark-field pRad. Dark-field light microscopy is a way to enhance contrast: rather than illuminating a specimen directly, it relies only on scattered photons. Dark-field pRad is essentially the same thing, but with protons instead of photons. In addition to a collimator to exclude the most highly scattered protons, we added a second collimator to exclude the least-scattered protons—the proton equivalent of direct illumination. Dark-field pRad is especially useful for very thin objects, and my team was the first to demonstrate it, in 2019, by resolving millionth-of-a-meter-thick gold leaf.
It’s mainly CT imaging for me now, with some pRad and some neutron radiography. Most of what I work on is classified, so I can’t provide specific examples. The Lab does a lot of nondestructive testing in support of its mission, and I image things at all stages of development, as well as before and after various tests. Someone will have an object, they’ll put it under some sort of stress like heat or pressure, and they’ll want to know what happened to it. Before the stress, my images confirm that the object is built correctly. After the stress, my images reveal what exactly happened to the object during the test.
When we have a new technique we need to be able to show it off and prove that it works.
I am also still very much involved with the development of new techniques—like big-barrel CT or dark-field pRad—that can improve our ability to do nondestructive testing for national security applications. This is where projects like the Bisti Beast skull come back around. When we have a new technique or a new tweak to an established technique, we need to be able to show it off and prove that it works, so we need non-classified data to present. While the museum scientists got data about brain cavity and sinus size to help validate hypotheses about the evolution of Tyrannosaurridae, we at the Lab came away with an invaluable, unclassified data set from the dual modality of CT and neutron imaging that we can use to showcase our techniques. Little did we know that the approach we developed for handling the fossil would so soon prove invaluable in the americium-mitigation situation. It was a great example of how partnership within New Mexico’s scientific community can pay off in unexpected ways.
Big picture
When I look back at my career, I can’t believe it has gone by so quickly! It really feels like I just got started. I’m proud of the collaborations I have participated in, and I believe that the science I do is crucial and important for national security. I feel tremendously proud and lucky to have gotten to do such a breadth of work, and for having the chance to have a long career doing hands-on science.
But if pressed, I would say my greatest point of pride is the people I have mentored. I feel very fortunate in the mentoring that I received, from college through grad school and even to this day, at the Lab. I was well mentored and well taken care of, and I feel it’s important to pay that forward. Everyone here is super smart and the work is meaningful, but I have also found satisfaction in helping students and postdocs find their way. I have had somewhere between 40 and 50 students and postdocs over the years. All of them are doing good things, and several of them are still at the Lab. One great example is Christina Hanson, who is now a colleague in the same engineering group as me. She was on my team in 2004 as a high school co-op student. She came back to my team as a postbac after her undergraduate work, and we worked together again when she was a postdoc. Now she leads her own team.
Although I very seldom encountered explicit “girls aren’t as smart as boys” sentiment, thanks to those who went before me, I have seen and experienced things that are difficult for women in science. Some hurdles that are deeply embedded in the system disproportionately fall on parents, particularly women. Some of the flexibility that has come from the COVID-19 pandemic, like hybrid work schedules and virtual meetings, have made it easier. And the Lab is better than a lot of places—my husband and I have been able to raise two great kids while having successful careers at the Lab. But I don’t want to sugarcoat it: it’s still hard. It’s not a reason not to do it, but it is hard.
In thinking about the future, I sometimes worry that the younger folks won’t have as much fun as I have had. My coworkers and I have a sort of running joke: I’ve published on such a motley assortment of things—like plants and dinosaurs and brains—but I haven’t published on food. I would love to publish a food study. The most likely angle is that sugar is sometimes used as a mock for high explosives, so maybe one day I’ll publish in a candy journal! That would be fun.
I was well mentored and well taken care of and I feel it’s important to pay that forward.
But perhaps the new scientists will have a different sort of fun. Artificial intelligence and machine learning (AI/ML), for example, is cracking open whole new worlds of scientific research. AI/ML is coming to everything, so I expect it will come to my work, too. One way that it might be used is that there is a lot of data involved in what I do, and at present it all has to be processed by human brains. I don’t think AI/ML can replace humans in this capacity any time soon, but a CT data set is tens of gigabytes of data, with thousands of images to go through, and AI/ML could probably be used as a first pass, to narrow the field, so the humans would have fewer images to look at.
I’m optimistic about the continued role of nondestructive testing to help the Laboratory complete its mission. I feel lucky to have been a part of it so far, and I look forward to what’s coming next. Perhaps another big barrel with another big surprise. Whatever it is, I can’t wait to see inside. LDRD