Developing in-situ, online monitoring tools for radioactive materials using fluorescence spectroscopy

Oak Ridge National Laboratory (ORNL) initially came into existence when it was tasked with building the world’s first continuously operating nuclear reactor as a part of the Manhattan Project. Since then, it has grown to a premier scientific institution outside of Knoxville, Tennessee.

Now, it is at the forefront of materials research, advanced manufacturing, neutron science, supercomputing, clean energy, biological science, and national security. More than 6,000 scientists, engineers, technicians, and support staff work on 4,470 acres at the site.

Two of those scientists are at the leading edge of deploying optical spectroscopy for real-time monitoring. Hunter Andrews, Ph.D., and Luke Sadergaski, Ph.D., study the development of in-situ, online monitoring tools for nuclear materials analysis in complex environments, using optical spectroscopy and chemometrics. They are part of the Radioisotope Science and Technology Division at ORNL.

“Our research together focuses on developing spectroscopy tools for online monitoring of various chemical processes,” Andrews said. “That can be looking at processing streams on advanced nuclear reactors. It could be monitoring radioisotope production to generate plutonium-238 for the Mars Rover or medical isotopes for treating cancer. Or even recycling nuclear fuel and closing the fuel cycle so that we can better utilize our used nuclear fuel supply.”

The hardest part is actually getting things to work in restrictive hot cell environments.

The bulk of their research and the hard part of incorporating a spectroscopy sensor in-line is building a model that takes a complicated spectrum and within a matter of seconds, transforms it into something useful to somebody who's not a spectroscopist. Operators or technicians running the chemical process need to know in real time what's going on so they can focus on optimizing that process.

For the last several years, the pair have been focused on how to build regression models that convert complicated spectra to useful information using as little of the precious radioactive material as possible.

“For many systems, we might be looking at a radioisotope that's incredibly rare. You can't find them nearly anywhere else in the world, and there isn’t much available to build a predictive regression model. We need to use as little as possible, so it can be used by those who need it,” Andrews said.

They’ve been focusing on using statistically optimal designs to minimize the number of samples needed to train the models.

In general, a big part of the focus is isotope production, specifically radioisotope production. To make these radioisotopes, ORNL does everything from fabricating targets to irradiating them with neutrons at the lab’s High Flux Isotope Reactor (HFIR).

“And then we bring the irradiated targets to the Radiochemical Engineering Development Center (REDC) for radiochemical processing,” Sadergaski said. “Radiochemical separations are used for the purification of the product and to recycle the unconverted target so we can use it to make more product.”

The separation takes place in a hot cell, a safe enclosed environment to work with radioactive materials, and sometimes they're separating elements that have similar properties. “We typically go through multiple rounds of solvent extraction and ion exchange chromatography for each radioisotope production program,” Sadergaski said. “So, there are actually multiple opportunities to monitor these species throughout the different processes.”

The example they like to use is the separation of neptunium from plutonium using anion exchange columns.

The columns contain small resin beads with functional groups that attract ions and stick to the column. The resin beads will bind neptunium and plutonium and reject all impurities. “Then we change the chemistry after the impurities are gone such that we can recover the neptunium and plutonium.”  The cut decisions involve sending the column effluent stream to various tanks or product bottles.

“We're making plutonium-238 for the plutonium-238 Supply Program to power NASA spacecraft and helping to optimize processing with online monitoring,” he said. “Neptunium and plutonium have similar binding affinities for the resin because they have similar chemical properties. There's a slight difference, kind of at a razor's edge, and if you have an optical technique monitoring the separation inline, you can make really good cut decisions.”

Appropriately rejecting the neptunium allows staff to recover a pure plutonium product more efficiently. Without online monitoring, the process would be more tedious and cumbersome, and more expensive.

“We are characterizing the column effluent in real-time, monitoring the concentrations of Np and Pu. This information can help guide the technicians as they make cut decisions. Technicians can actuate a switch valve to go from a waste stream to a plutonium product bottle based on the information we provide them,” Sadergaski said.

“We have a switch valve right after our inline flow cell with our fiber optics where we're taking our measurements, so we know what's coming out before it even gets to that switch valve. And then the technicians know exactly when to start collecting the plutonium product.”

It's a quality control technique in essence.

“The plutonium program is sort of a neptunium processing program. And the Californium-252 Program is a curium processing program. We're developing optical sensors to do remote fluorescence measurements all in a hot cell. All our measurements are in a hot cell. That's one of the reasons our work is so challenging and exciting. We can do spectroscopy in hot cells using fiber optic cables. We have more than 10 meters of fiber optic cables. So we send the light into the hot cell, excite the sample, and then measure the spectrum through fiber optics in our spectrometer from the control room.”

“When irradiated Np targets are removed from the nuclear reactor, they are highly radioactive, and so you can't handle them with your hands,” Andrews said. “So we process them in hot cells, a large shielded facility surrounded by concrete in all directions except for where the user is looking at it, and they look at it through four and a half feet of lead-plated glass and mineral oil.”

“That is where it becomes challenging to do traditional analytical measurements. Normally, you would just take a grab sample and dilute it and send it to your ICP-MS on-site and get a great value. They can still do that, but there are disadvantages. It involves taking a sample inside this hot cell, using robotic arms, and diluting it enough to be allowed to remove it because of its radioactivity and then transferring it to another laboratory.”

“In the process, you lose information and the whole process takes time. So, if you're trying to make process decisions in a split second, that’s just not going to do what you needed to do, yet it’s still an important part of our process.”

The information for the cut times would preferably be instantaneous.

Spectroscopy provides a real-time signature that might have less accuracy but is good enough for a game-time decisions for operators to switch the valve and say “collect” versus “toss”.

“All of this goes back to the quality control and assurance as we can really add a huge enhancement to the efficiency of our radiochemical processing by giving users that insight into what's going on, as opposed to operating somewhat blind. When you're looking at things after the fact, you can't really do anything except plan to do better next time.”

Fluorescence is a technique the pair is developing for the Californium-252 Program, a radioactive element, where they plan to monitor fluorescence trans-plutonium species as well as lanthanides. The researchers are synthesizing californium-252 as a neutron source for a variety of applications from prompt neutron activation analysis to restarting nuclear reactors. Targets containing curium are irradiated at HFIR to produce Cf and other heavy elements. Curium is typically accumulated as a byproduct of nuclear fuel-related activities.

“We’re developing a new capability, with some surrogates, such as europium(III) and samarium(III) in this initial phase,” Sadergaski said. “We've also done some curium work as well. Curium is highly fluorescent in most solvents. It’s something that isn't monitored in a way that is ideal during hot cell processes. They have radiometric options, and all of the actinide elements are alpha emitters. You don't always get a good snapshot of one or the other.” The team is also looking at applying fluorescence spectroscopy to berkelium and einsteinium, both radioactive heavy elements.

Sadergaski said they are hoping to help out with curium processing.

The researchers plan to use fluorescence techniques in the lab’s up-and-coming uranium science and technology center. That's outside the scope of isotope production but within the nuclear fuel cycle.

Again, there's the separation of curium from some of the other elements, and there are cleanup cycles. Spectroscopy offers a faster, cheaper way.

Their work with these materials spans the areas of basic research and actual applied science.

“We're doing a lot of research that bridges that gap of how you take the spectra and build some sort of multivariate model that can turn it into a concentration value that accounts for how these species interact with one another. And then the applied side of how do we implement this in a hot cell to help a process that's going on,” Andrews said.

Fluorescence spectroscopy techniques have been around for decades, but the two are bringing the techniques into an area where there are many more restrictions. Radiological constraints, and hot cell facilities. It makes traditional methods amenable to their specific applications and environment. Their research has helped support production and helps them get their techniques into that production environment.

The team uses a HORIBA Fluorolog-QM (FL-QM) modular research fluorometer for their studies. The FL-QM series includes research-grade spectrofluorometers and are the fourth generation of the HORIBA Fluorolog.

The FL-QM features optically perfect, all-reflective optics, combined with a multitude of light source and detector options, and sample handling accessories, and provides the highest sensitivity and greatest versatility of any spectrofluorometer. It can also be enhanced to suit a broad array of luminescence experiments, with an extensive list of optional accessories to expand capabilities and performance.

The two scientists also use a HORIBA iHR320 spectrometer for steady-state fluorescence measurements as well.

Their further research is from the production cycle. The easier and shorter the time frame to conceptualize and execute the experiment the better.

“If we are trying to answer more simplistic questions where we can use surrogates, especially ones that we refer to as ‘cold’, so non-radioactive, that makes it a lot easier,” Andrews said. “Because we can just do it on a benchtop, and we can usually source those materials from somewhere else and we can even sometimes get student involvement to help us do it.”

But when they start using radioactive material, they are looking at doing it either in a radiological fume hood or glove box labs where now they are dealing with some of the issues they have to deal with on hot cells too, where they are feeding fibers through ports on the back of the box.

And they have to meticulously prepare and bring everything into the box in a way that keeps them on the outside safe and keeps any contamination on the inside. And then after all of that, they still have to execute the data analysis side of these studies.

“So, a lot of our studies have a pretty big reliance on data analytics, machine learning, and chemometric models,” Andrews said.

NASA, the Department of Energy Office of Nuclear Energy, and the Department of Energy Office of Isotope R&D and Production are the main benefactor of Sadergaski and Andrews’s work.

"The plutonium program is sort of a neptunium processing program. And the Californium-252 Program is a curium processing program. We’re developing optical sensors to do remote fluorescence measurements all in a hot cell. All our measurements are in a hot cell. That’s one of the reasons our work is so challenging and exciting. We can do spectroscopy in hot cells using fiber optic cables. We have more than 10 meters of fiber optic cables. So we send the light into the hot cell, excite the sample, and then measure the spectrum through fiber optics in our spectrometer from the control room."

Follow Us