Open Science Conversations: An Interview with Hao Li, Tohoku University

As part of an ongoing series here on ACS Axial, we’re interviewing authors and librarians from around the world to find out more about their research, their published work, and the impact that open science is having on a changing landscape of research communication. This time, we spoke with Hao Li, Professor and Principal Investigator at Tohoku University, Japan.

What can you tell me about your research group's current focus?

My research group's focus is on the development and application of AI for chemistry and materials science. We have a closed-loop design strategy to drive the effective discovery of advanced materials for industry. We have developed many scientific databases for materials, which we use to inform our research — including catalysis, battery, and hydrogen storage. The aim is to understand the structure-performance relationships of those materials, so that we can make better predictions and understand the key factors that drive their performance.

Because we have access to such a large amount of data, we developed AI agents that can better digest the knowledge and identify new areas to explore. Using these AI tools and physical models, we can make predictions for new materials.

To close the loop, we do validation experiments — both those involving lab staff, but also autonomous experimentation using robotics. The data from these experiments is fed back into the database, closing the loop and building an even larger database which helps us develop more precise AI models and more precise physical models, so that we can provide improved predictions for experiments. Through this iteration, we can develop materials more and more effectively.

Tell me about the work in your paper, “The Key Steps and Distinct Performance Trends of Pyrrolic vs Pyridinic M-N-C Catalysts in Electrocatalytic Nitrate Reduction.”

I think this is an interesting work, and it's one of my favorites that we've done recently. We set out to better understand the mechanism behind the electrochemical nitrate reduction reaction. Many previous studies used oversimplified models to understand the reaction which did not consider some key factors such as pH effects, kinetic effects, and interfacial effects between the liquid phase and the solid phase. As a result, many previous models failed to capture the real structure-performance relationship in their experiments.

In our work, we derived the new model by integrating our secret weapons — the pH-field coupled simulations in microkinetic modeling. By considering electric field effects, potential of zero charges (PZC), and solvent effects, then we can derive some equations and physical frameworks that can capture the phenomenon of pH effects and demonstrate how these factors affect reaction performance.

Based on that, we developed a so-called volcano model to link simple descriptors with performance. To validate our model, we compared it against data from past experiments, and then undertook new experiments for some well-defined structures of single-atom catalysts. We used those well-defined structures because, in theory, we can precisely simulate these kinds of catalysts as well as using the same structure for physical experiments. This provides a perfect opportunity to compare the theoretical results from simulations against the results of real experiments. So we performed these experiments on materials used in the electrochemical nitrate reduction reaction and found consistent agreements between the two under different pH. That's the key conclusion of the paper: we developed advanced modeling methods to provide explanations for the structure-performance relationship in the nitrate reduction reaction.

This was published in the Journal of the American Chemical Society — tell me about the experience.

JACS is one of my favorite journals because the review process is very professional. It has a very high quality bar, and the editors can always find reviewers that can really understand our work.

This is the most important thing: finding reviewers that have similar expertise. If a reviewer doesn't really understand your work, then they can leave comments which are less helpful or fair. This isn't the case at all for JACS — the comments are always useful and have some good suggestions to improve our work. As well as this, the reviewing and decision process is fast. Every paper from our lab that has been published in JACS has had a good impact.

The article was published in 2025 — what kind of impact do you think it's had since it was published?

Every time we publish in top journals like JACS, we always receive good feedback from readers and people who see us present our work at conferences. In recent years, there are many researchers who follow our work and use the results to find good catalysts for their own work.

Also, in China there are some public platforms to help translate research articles into blog posts that can help the work find more readers. I know that this has been effective for us in how our findings are being used in follow-up research by other labs in China.

The article was published as open access — do you think that's helped in communicating the findings?

Our library at Tohoku University in Japan has open access agreements with many different publishers. Whenever we have a paper published in a journal that is included in an agreement, then the library pays the article publishing charge (APC). I always choose open access when this option is available, as I believe this expands the number of readers — particularly from developing countries.

How did you discover this institutional open access agreement was available to you?

Every year, the library sends an email to all the faculty members in the university which tells us which publishers and journals have their APCs covered by the library. They particularly focus on covering open access fees for younger and early-career researchers, and only for trusted journals — including those published by ACS.

Every time I have a paper accepted by a journal, I will check the library's internal list of supported journals and will make the paper open access by this method if it is possible.

Do you use any other forms of open science workflows, for example open data?

Yes, I encourage and support open data. In fact, our databases — we have more than ten types of materials database platforms like the digital catalysis platform (DigCat), digital battery platform (DigBat), and digital hydrogen platform (DigHyd), and so forth — stores data not only from my own group, but also data obtained through data mining techniques, and uploads from other researchers around the world. I frequently ask my colleagues to upload all their key data, including experimental and computational data, to the relevant database so that users can access it in the future. Due to these contributions, many of the databases are the biggest in their fields, with totally millions of entries so far. We also share our data on GitHub upon the request of reviewers or editors, to make assessment easier.

What do you think are the biggest recent developments in open science and open access?

I feel that some researchers are hesitant to choose open access because of the limits of their research budget. For example, in my case, if the university cannot pay the APC, I may not be able to choose open access because an APC for a single paper might represent a huge portion of my research grant. I think realistically it will be difficult for people to choose open access without their institution's support.

Another case I found very interesting is bias against open access in some regions. I'm a Chinese researcher living and working in Japan, but I know very well how Chinese academic society thinks — and many researchers in China choose not to publish open access partly due to budgets, but partly because some still view it as "pay to publish" with no quality filter. In fact, some administrators who aren't familiar with scientific publishing will see a request for APC payment and refuse it because they don’t understand the open access concept and have similar thoughts about paying to publish. Some of them don't know about the review process, or how many publishers have a high quality bar for both open access and non-open access papers. I think people still need time to learn about how open access works, what the benefits are for readership and citation, and then accept it — particularly in Asia.

Where do you see OA in 10 years' time?

I expect more people will accept open access, and more universities will work like ours where the university or its library covers some or all APCs, instead of expecting authors to pay for publishing from their grants. I am also the editor-in-chief of an AI-focused journal which is open access, and in the first 2 or 3 years we will publish papers without charging an APC. I feel like in time, more people will accept this kind of publishing: it's good for spreading knowledge further.

What do you think you'd be doing if you weren't a researcher?

If I wasn't involved in fundamental research at the university, solving those '0 to 1' problems, then I would be focused on the industrial '10 to 100' problems instead, to meet the challenges of industry and solve real-world problems. There are two different sides to this research, the fundamental and the applied, but the logic and philosophy behind them are the same. It's this that really drives me and I would continue doing it no matter what.