When academics go looking for funding it is common (some might even say required) for their grant proposals to play up the industrial applications of their research. However, bridging the gap between fundamental research and industry is often difficult. One of the most important aspects of scientific research is the way it explores the unknown. This comes with a significant level of risk. The most interesting problems are often the most challenging ones and even seemingly straightforward questions are never as simple as they initially appear.

Academic research embraces the uncertainty that comes with this risk, celebrating the discovery of new questions and, in some cases, finding answers that are unrelated to the original line of enquiry. In contrast, within a commercial enterprise the most critical aspects of research projects are specific “deliverables” and the particular business needs they serve. Most companies do not have the resources to exploit new discoveries in unrelated fields or sectors, and instead focus on generating tangible returns within their own space.

Best of both worlds

In an attempt to get the best out of both of these worlds, companies and academics sometimes form customer–provider relationships in which the industrial partner essentially pays for a research service. In this way, the company can retain all the desired intellectual property rights and clearly define the work plan and goals. The academic partner in turn receives much needed funding, as well as a valuable route towards applying their results. However, this relationship can become strained if the academic partner aims to develop new methods and build fundamental know­ledge while their industry counterpart is expecting a specific deliverable or product. Unfortunately, in many sectors fundamental research is seen as an extremely long-term investment, making it one of the first budgets to be cut during a downturn. This can be problematic for academic partners.

I have personally experienced situations like this from both sides, having been both the industrial partner and the academic at various points of my research career. While working as an industrial materials scientist at Chevron I could see my research turn into tangible advances in technology. However, I was often frustrated by the fact that studying fundamental mechanisms and method development was given low priority. I then decided to return to academia, joining Alexander Shluger’s group at University College London (UCL) to focus on the theoretical modelling of material properties. While I was now able to throw myself into studying fundamental mechanisms, it became difficult to see how my work developed into real-world products. So, when I took a step back and looked closely at what research and development means to me, and where I wanted to position myself, I decided to use my experience in both academia and industry to try to reconcile these two goals. My interest lies in pushing the boundaries of our knowledge, so becoming a contracted problem-solver was not an ideal arrangement. Instead, I decided to embark on an exciting journey: I started a new company, Nanolayers Research Computing, with a few like-minded colleagues.

People are often curious about this approach, and I am sometimes asked how I balance an academic life and an industrial one. In fact, this is quite demanding. I have given up a lot of nights, weekends and holidays, and even so, it would have been extremely difficult to stay motivated without the encouragement and support of close friends, family and my group at UCL. Another common question is “What does your company actually do?” The short answer is that we apply computational chemistry, physics and machine-learning techniques to design and develop new materials for a variety of industrial applications. However, what that statement actually means in practice is not transparent. How does a materials design firm – and a heavily theory-based one at that – fit in to a landscape of chemical, pharmaceutical and electronics companies?

Novel nanoparticles

For Nanolayers, part of the answer lies in the European Union’s Horizon 2020 framework. This framework incorporates a role for companies that are designated as “translators” because they help research groups connect with people in industry who might want to use the group’s software or methods. Our years spent in the theoretical physics community gave us an excellent network of potential university collaborators, while my past life as an industrial materials scientist provided several useful industry connections. Before long, we were invited to join a Horizon 2020 project that aims to replace certain critical, industrially-relevant catalyst materials with novel transition-metal nanoparticles.

Within this project, known as CritCat, our role is to apply machine-learning techniques to results and data collected by our academic partners. We then use our findings to develop catalyst materials that do not incorporate elements such as platinum-group metals, which are of critical importance in Europe due to their cost and scarcity. Our strategy for catalyst design is to figure out what features are relevant in describing these materials and then train neural networks to learn how these features correlate to catalytic activity. This allows us to learn the mechanisms behind what makes a good material, and thus design and control the properties of our materials. We then design new nanoparticles that are subsequently produced by our manufacturing partners and then validated in real-world trials.

As a small-to-medium-sized enterprise (SME) capable of interfacing not only between academia and industry but also between theory and experiment, we hold a unique position within the CritCat project. We have taken a leading role in the dissemination and the exploitation of our technology, and have also leveraged our expertise in computational chemistry and theory techniques to provide additional support services and method development for our theory partners, who are based at Finland’s Aalto and Tampere universities.

Beyond materials science

When I got the opportunity to network with other materials design-focused companies and projects such as NoMAD (novel mat­erials discovery), one of my take away messages was the importance of developing a marketable product along with a diverse skillset. Since Nanolayers’ core values involve performing exploratory research rather than commercializing something that has already been tested, we decided to take on more of a consulting or partner role and looked for an opportunity to apply our expertise and experience in other sectors. Our goal was to use our simulations and machine-learning techniques in an equal partnership with someone capable of producing marketable devices or software.

To this end we recently formed a partnership with two firms (GV Concepts in the US and eQuumSoft in Asia) to develop new technologies for monitoring vital signs and conducting medical pre-screenings remotely. By continuously monitoring patients’ vital signs, clinicians may be able to spot qualitative early-warning signals for a variety of potential illnesses, and intervene if the risk is deemed high enough. The challenge is to do this outside a clinical setting, so that patients – particularly those who are elderly, high-risk or suffering from chronic diseases – can record these vital signs in the comfort of their own homes. This is a complex task, one that involves digital devices, diagnostic tools and complementary vital-signs data-collection software. Our solution will make it possible for patients to monitor their own health in a personalized way using a set of patented diagnostic tools including a digital stethoscope, otoscope, blood-pressure monitor, oximeter, thermometer, ophthalmoscope and camera. These devices are all integrated with a smartphone-based software suite that not only enables patients to connect remotely with healthcare professionals, but also allows doctors to remotely control diagnostic tools during the “virtual visit”.

We use the data collected in this project in a way that is similar to the method we employ for designing novel materials. In this case, we are seeking symptom–disease relationships rather than structure–performance ones, but the strategy of using machine-learning techniques to identify relevant relationships is the same. For example, we use neural networks for image recognition and signal processing to help healthcare professionals interpret the collected data.

As we pick up more projects and partnerships that are structured in a similar way, Nanolayers continues to expand while focusing on the theme of bridging the gap between fundamental scientific knowledge and techniques and industrial applications. In this way, we can enjoy the best of both worlds by exploring new materials and techniques while making sure that the discoveries and advancements we make are applied in a meaningful way. Time spent on improving our own methods and gaining experience is not wasted. After all, one of our most important products will always be the research team itself.