Why Big Data Is Transforming Environmental Nanoparticle Analysis

At the SciX 2026 Conference, which took place in Sparks, Nevada, we sat down with Kevin Wilkinson, a professor of environmental chemistry at the Université de Montréal, to talk about him receiving the Lester W. Strock Award and about his laboratory’s work on environmental nanoparticles and trace metals. Below is a text transcript of that conversation.

Wilkinson’s group combines many techniques, with elemental and isotope ratios proving especially useful in complex, heterogeneous systems. And as he stated in our conversation, this strategy is necessary for distinguishing engineered nanoparticles from naturally occurring ones.

Congratulations on winning the Lester W. Struck Award. What does winning this award mean for you and your research?

I am extremely honored to have won this award. It was completely unexpected. This is an award where we don't ask to be nominated or we do nothing to try and win it, so it was a very pleasant surprise. I consider it somewhat unearned, because my view is that I'm inspired by the work of others, by reading other papers, and I look for things I can do in other people's research. And then I take it and I assign it to some of my students, so they're doing the work and I'm just the intermediate. So, I don't know if it's really me that should be getting the award or the people on each side of me.

Distinguishing engineered nanoparticles from naturally occurring ones in real environmental samples is analytically difficult. What is the key signature or method that lets your laboratory tell them apart with confidence?

I wish there was a key signature or just one technique to be able to do this. Unfortunately, that's not the case yet. We're looking to combine a multitude of techniques and different signatures, so implying big data and suggesting that we're going to need many techniques and many ways of analyzing the results from the techniques that we're using. Putting all that together, I can't say there is a key signature, but as we go along with this, we see things that are more useful than others in the elemental and isotope ratios. These are helping us understand the systems and identify markers of these rare events, and that's been very helpful for us to progress with our knowledge of these very complex environmental systems that are entirely heterogeneous, polydisperse, and difficult to characterize.

According to your conference presentation, you detected up to 21 elements in individual particles and analyzed hundreds of thousands of them. What does the instrumentation need to do to achieve that, and what are the biggest challenges in handling that volume of data?

In regard to instrumentation, what's really changed in the past 10 years is the speed at which these techniques are able to acquire the data, and that's been entirely helpful for what we're trying to do. Then it's the big data treatment, and that's improved by orders of magnitude just over the past few years, in that initially I'd send someone to the lab, they would spend a day doing analysis and then about two weeks' worth of data crunching. And now, with some of the advances that we've got with some of the software, we're about one day of analysis and maybe 2–3 days of data analysis. Therefore, it's really helping us better understand what we're looking at so we can treat this data more efficiently, which is giving us better insights on what we're measuring.

Many particles containing arsenic, antimony, bismuth, indium, and tin had no copper at all. Why was that absence so telling, and what does it mean for how the region's arsenic and lead pollution is understood?

The fact that the particles had no copper was expected. We're outside a copper extraction facility. They're trying to take out the copper, and so it was perfectly normal to see particles with little copper left. The thing that's really interesting here is some of the combinations. They used a combination of recycling and ores as their inputs for these processes. When we see particles coming out of the refinery with elements from the recycling and from the ores, that can tell us it's coming either from the stack or from fugitive emissions, but it's really coming from the process.

That's giving us insight, and these particles tend to be smaller. They are the particles where the arsenic is often located, and these are the particles that tend to be transported away from the facility, so it's helpful for us to understand those particles through their chemical composition.

You detected airborne metalloids up to six kilometers downwind, and lead isotope signatures varied with particle size. What are the implications for the surrounding community and for regulators, and how might this approach be applied to other industrial sites?

The lead isotope work is interesting for some specific cases when the lead is an important part of the process, which is the case here and the case of other places that we're looking at. Lead tends to be a fairly common pollutant. That being said, we were able to measure 6 km away and actually look at the compositions to better understand the transport by following the differences in the composition. However, 20 years ago, there were other groups measuring arsenic particles 100 kilometers away in the snowfall. They relied on the fact that arsenic accumulates in the snow over several months, and so they let the snow concentrate their particles for them to increase their detection limits.

So perhaps we should learn from them and go a little further out or look at the composition of particles in the snow at these further distances. We know from their work that the particles are being transported very long distances and going a long way away from the source where they are being produced.

For a graduate student starting out in environmental analytical chemistry today, what instrumentation or skill set would you tell them is going to matter most over the next decade of nanoparticle and trace metal research?

I like to tell my students to learn how to learn. That's the only thing that I expect in a PhD student. If they can learn how to learn, they'll be able to adapt to different conditions. That being said, and in addition, the instrumentation is evolving so quickly right now that it's a little bit difficult to focus on one type of instrumentation. In three years, the instrumentation will evolve, and we'll have something much better.

In terms of instrumentation, I would avoid making any recommendations to my students. Instead, I would tell them to learn how to evaluate whether these instruments are going to do what they say they're going to do because salespeople will always tell us, "This is the best thing since sliced bread," and we have to be able to evaluate that.

The final point is that students need to learn how to analyze the data using big data techniques. Everything we're doing as analytical chemists now has evolved towards big data analysis, so whereas 20 years ago we'd make a few measurements and try and put that together in a report, now we're getting gigabytes or terabytes of data per sample or samples. Therefore, we've got to be able to analyze that efficiently. They should be able to pull the important results out of the data using these advanced data treatment and statistical techniques.

Further Reading

1. Wilkinson, K.; Workman, J., Jr.; Wetzel, W. Kevin Wilkinson on Tracing Airborne Metal Pollution Particle by Particle. Spectroscopy Online, 2026. https://www.spectroscopyonline.com/view/kevin-wilkinson-on-tracing-airborne-metal-pollution-particle-by-particle (accessed October 28, 2026)