American Scientist

As a professor of chemistry at the University of North Carolina at Chapel Hill, Yosuke Kanai does research in theoretical and computational chemistry that is centered on developing quantum-mechanical computational methods and using them to study the dynamical behavior of complex chemical systems. His work on simulating quantum-mechanical processes spans a wide range of topics from solar energy conversion to proton beam cancer therapy. The overarching theme of his research is to develop predictive understanding of electronic excitation and dynamics phenomena that arise from the interplay among electrons and atoms, especially in condensed-phase and other extended systems. He is particularly interested in obtaining understanding at the molecular level by developing and applying computational methods based on first-principles electronic structure theory. He received his PhD in theoretical chemistry from Princeton University. Kanai gave the inaugural Sigma Xi Pariser Global Lectureship for Innovation in Physical Sciences in April 2022. After the event, he spoke about his research with American Scientist editor in chief Fenella Saunders. This interview has been edited for length and clarity.

How is quantum chemistry connected to quantum mechanics and theoretical chemistry?

In the early 1900s, the concept of theoretical chemistry emerged. Essentially, instead of performing experiments, we can use mathematics and physics to figure out how molecules behave under different conditions. Within theoretical chemistry, there are a couple of different subfields, including quantum chemistry. What quantum chemistry comes down to is using the fundamental physics of quantum mechanics to understand how molecules behave under different circumstances. If you go down to the scale of atoms or molecules, you have electrons and atomic nuclei. Instead of electrons behaving like classical particles, the electrons follow quantum mechanics. Also, some parts of atomic nuclei, such as protons, are also very small and need to be treated quantum mechanically. When quantum mechanics became rather mature in the 1920s, many scientists realized that they could actually use the laws of quantum mechanics to understand how electrons and atomic nuclei behave. Even though it started a long time ago, electronic computers made it possible to actually use quantum mechanics, to apply it to study molecules. That started out in the 1950s, and with all the advances in computers in the past 70 years, we’ve seen more and more interesting developments and advances in the field of quantum chemistry.

Is quantum chemistry more about the interactions between particles, whereas quantum physics is more about the behavior of fundamental particles?

Quantum mechanics is a very broad field. On the one hand, you have a lot of high-energy physicists applying quantum mechanics to make sense of subatomic particles that are fundamentally related to understanding the overall structure of the universe. But on the other hand, you go to the really tiny scale of electrons and molecules, and we’re using quantum mechanics in the context of understanding how electrons and atomic nuclei behave.

How have technological advances affected this field?

It started with the development of electronic computers in the early 1950s. In the 1990s people started having more access to personal computers. Certainly that motivated people to explore fields like quantum chemistry, because now they could start to perform quantum chemistry calculations on their personal computers. And then of course you also see a lot of excitement about supercomputers. The continuing advancement of computing technology is pushing back the frontiers for quantum chemistry. But that’s one side of chemistry as a whole field, because on the other side, it’s not like experimental chemistry has stopped advancing. In the past 50 years or so, on the experimental side, people have started having things like particle accelerators and beam lines for doing spectroscopic experiments. If you have a lot of advancement on the experimental side, then you’re able to obtain a lot more exciting results, and that motivates people like myself to understand what’s happening in experimental chemistry. In a way we all advance hand in hand.

Is it possible to produce an accurate model at the scale of particles?

Particularly because of that tiny scale, it’s difficult. We know how to use Newtonian physics very well. If you want to model how I throw a ball, you know exactly where that ball is going to go. When you go to a tiny scale, it’s no longer Newtonian physics but quantum mechanics that governs behavior. That makes it very complicated to understand and predict what happens on that tiny scale of molecules. We can’t obtain an exact solution. But thanks to advances in computing technology, we can be pretty accurate for some properties, such as predicting reaction energy. Some other properties, like the optical excitation energy, are more difficult to predict accurately. For organic molecules, we’re able to reliably model those optical excitation properties like transition energy. But for some other types of molecules, such as transition metal molecules that contain heavy mass elements, it’s difficult for us to model optical properties reliably. It’s also partly because of relativistic effects and those things becoming important. So there are certain properties we can model more reliably than others, but also certain groups of molecules we can model more reliably than others.

Why do you work on characterizing high-energy proton beams?

This physical phenomenon is important in cancer therapy. But it’s also important for space missions. If you go outside of the Earth’s atmosphere, you have these galactic cosmic rays of high-energy ions going by. If you’re an astronaut and you’re not protected, you’ll be hit by them. One of the big problems for space missions is that electronics must be protected from cosmic rays, because high-energy ions can pass through electronics and induce electronic excitation, so your electronics can fail. So understanding the behavior of high-energy ions becomes very important in different contexts.

How do simulations give insight into the behavior of these high-energy beams?

To do experiments, you want to produce these high-energy protons artificially, which can be done using a cyclotron, which can accelerate the proton into a charged particle using magnetic fields. You can measure things like electronic stopping power, which is essentially the rate of energy transfer from these high-energy protons to electrons in the target matter, like DNA. But what is difficult for experimenters to figure out is where this energy is transferring into the DNA. If you look at the structure of DNA, it’s a very complex macromolecule. The quantum mechanical simulations we do can show where the energy is going. For example, we showed in our recent work that a lot of energy is being transferred into the side chains of the DNA, as opposed to the DNA base pairs. This information is important because in the context of proton beam cancer therapy, for example, you try to induce damage to the DNA side chains, because that damage is very difficult for proteins to repair. If you can induce that side chain damage in a cancer cell, the proteins are less likely to be able to fix that DNA in the cancer cell. You can induce cell death in the cancer cells. That’s one example of where a quantum mechanical simulation can provide insights that it probably would be very difficult, if not impossible, to obtain from experiments because the energy transfer happens on the scale of femtoseconds. We can get a resolution of attoseconds in our simulation to understand how the energy is being transferred.

Why are the DNA side chains targeted in this energy transfer?

In DNA, especially when it’s under physiological conditions like in water, all of its side chains are missing hydrogen, missing protons. That means those side chains have exposed electrons. These electrons appear to be responsible for this high degree of energy transfer. We have also studied DNA by itself, in a vacuum. But DNA in a vacuum is protonated; you don’t have those exposed electrons on the side chains. Only when we studied DNA in water, like when it’s under physiological conditions, did we have these exposed electrons. Then we realized that that’s important in terms of absorbing energy from these high-energy protons. But it’s experimentally difficult to study the DNA under physiological conditions when responding to the proton beam. Even when you go to simulating DNA in water, you have more than 10,000 electrons to simulate, and that’s a very challenging task.

Do you think that these simulations will assist in finding a more efficient way of targeting cancer DNA?

If you really want to improve ion beam therapy, you need to understand how it works. We’re doing fundamental science to provide an understanding at the molecular level of how it actually works. Many people talk about things like going beyond protons in terms of the ion beams, things like using carbon ions for example. The rationale is that if you have a carbon ion, it has more charge, so you can deposit a larger amount of energy into the DNA of the cancer cell. That makes sense. A proton has a plus one charge. Carbon, if you strip away all the electrons, is plus six. But does anything else change? Those are the insights that our simulation can provide. Now we have an original simulation with other ions, to show how that behavior of energy transfer changes when you go to things like alpha particles and carbon ions. That can guide people’s thinking about improving or advancing the ion beam therapy beyond proton beams.

“When you simulate DNA in water, you have more than 10,000 electrons to simulate, and that’s a very challenging task.”

• Facebook
• Twitter

The simple rationale is, okay, you have more charge, so you can have more energy transfer. There’s a simple formula for this. But this is a very idealistic, simplistic idea. If you actually perform a simulation like we do, we can see the degree of ionization for these particles. The carbon is never really carbon plus six. Even with the protons, a proton is never completely a plus one charge. These quantum mechanical simulations can provide those insights into what actually are the charge states of these ions.

What is the importance of focusing in on the fundamental science?

It’s easy to appreciate the part of our research that can connect to everyday life, like cancer therapy, for example. But the reason we can perform these calculations and provide these important insights for, in this case, cancer therapy is because there’s a lot of fundamental work that goes in before we can perform these simulations. We spend a significant amount of our time developing computer simulation methodologies. It involves a lot of mathematics and a lot of programming. It’s not just because we have a supercomputer that we have these simulations. We do need the supercomputer, but we also have to have the computational methods development: a lot of programming, a lot of solving equations, a lot of mathematical derivation. That also has to be done before this simulation can be performed. It’s the work that people don’t see or read about very often. But still, it’s very important.

How are these computational derivations created?

In order for us to do these simulations, we have to create a computer program. To do that, you first have to understand how quantum mechanics works, and also you have to formulate the equations of quantum mechanics in such a way that they can be programmed. Supercomputers come with many processors working in parallel, and computational scientists like us have to program these equations so that the computer program we write can take advantage of all the processors simultaneously. If you want to do this type of work, you don’t just need to know math, physics, and chemistry. You also have to learn the computer science and programming aspect. I’m in quantum chemistry, but if you look closely at what’s happening in the research, you see a lot of math, physics, chemistry, and computer science. It’s quite interdisciplinary.