Ctirad Červinka (MSc., PhD) from UCT Prague’s Department of Physical Chemistry, received a five-year Czech Science Foundation JUNIOR STAR grant for his project entitled Making ab initio modelling possible for disordered molecular semi-conductive materials. In this interview, he explains project goals and the broader context of his research.
Where did your journey to chemistry begin?
The beginning can clearly be traced to the summer training camps in Běstvina, which I repeatedly participated in as a high school student. My mentors’ (now-Professor Slavíček and Dr. Holzhauser) enthusiasm at the time and the unforgettable atmosphere of the chemistry camp inspired me to continue my education in chemistry at the university level at UCT Prague, where I gained a broad overview of the various chemical disciplines. Because I was interested in the interactions between molecules and their connections with the macroscopic behaviour of matter, I chose physical and computational chemistry as I progressed. I soon started working in Professor Fulem’s group, where I was introduced to the realm of thermodynamics, the phase behaviour of materials, and the world of basic research in general. He involved me in his grant projects, but he concurrently gave me great freedom of study and the right conditions for learning to model the thermodynamic properties of different forms of matter. Over time, I was especially drawn to the field of molecular crystal cohesion modelling, which also became the topic of my UCT Prague doctoral studies.
You also studied abroad. Did it somehow influence the trajectory of your professional activities?
During my doctoral studies, I conducted internships in Aachen and Clermont-Ferrand, where I had the opportunity to see how the academic community works at traditional Western universities. The positive attitude and willingness of my foreign supervisors to share their great know-how in the fields of theoretical chemistry and molecular simulations encouraged my thoughts regarding an academic career. Immediately after defending my PhD, I decided to return to Clermont-Ferrand for a postdoctoral research stay, where Professor Costa Gomes and Professor Pádua put a lot of trust in me and offered me the opportunity to work on their breakthrough project on modelling porous ionic liquids. I think I did not disappoint them, and I gained the courage to go on an even bigger adventure in life. The Fulbright Commission enabled me to spend almost a whole wonderful year in Riverside, California, where I collaborated with Professor Beran, one of the world’s greatest experts in the quantum-chemical modelling of the electronic structure and spectroscopic properties of molecular crystals. Fond memories of my prior studies and strong ties to home then brought me back to UCT Prague as assistant professor at the Department of Physical Chemistry, years ago already.
And how does the story go on upon returning home?
I started dealing intensively with the development of computational methodology for the ab initio prediction of polymorphism, i.e. the existence of different crystal forms of active pharmaceutical materials and crystal sublimation of ionic liquids. I managed to get my first grant funding for these research topics and also gradually attracted several students interested in computational chemistry to working with me. Because of this, I have the feeling that my work has meaning and that it is also interesting to someone other than just me.
What does the Czech Science Foundation’s JUNIOR STAR grant mean for your project?
In the Czech environment, one cannot rely only on institutional funding if one wants to build a research group with significance beyond our borders. This applies both in terms of financing the operation and investments necessary for conducting research as to personnel costs. The current basic salaries at our university are astronomically far from being competitive even on the Czech labour market. If you are dedicated, you can still cope with this yourself. However, if we want to even have any chance at recruiting young qualified and motivated researchers before they disappear abroad or to the private sector, we have to secure longer-term funding for them with grant projects.
Success in such kinds of competitions brings financial stability to a research group and, for a certain period of time, also brings relief from constant grant proposal writing. Logically, my first choice was to apply for very competitive and prestigious ERC grants. Unfortunately, almost a year’s work in this direction did not bring an immediate reward. However, I now see that this process, including the stages of initial enthusiasm, subsequent disappointment, learning, and reworking of the original proposal, were the key to success in this competition for a Czech JUNIOR STAR grant.
An environment in which a young aspiring scientist in the role of breadwinner has to stress out every year or two regarding grant competition success—where the purely statistical chance of success is often less than 15%—is not very inspiring and it does not encourage creative ideas leading to breakthrough discoveries. In this context, I greatly appreciate our success in the JUNIOR STAR competition and take it personally as a great responsibility towards the Czech Science Foundation, my entire team, and the ordinary taxpayer.
What would be benefits of the development of ab initio simulation methods?
In the title of the project, I pointed out the development of ab initio simulation methods for the description of amorphous phases of molecular materials, i.e. liquids or glasses. Existing computational simulation methods are, in principle, applicable to simulations of crystals, liquids, and gases. But if we want to go down the path of pure prediction of the macroscopic properties of materials, in my eyes, we should only rely on quantum-chemical postulates and minimize the amount of empirical information entering our models. Modern computer technologies allow us to study gaseous and crystalline states from the point of view of quantum chemistry very efficiently due to the practical absence of molecular interactions or due to a well-defined regular structure. On amorphous phases, quantum chemistry is more or less breaking its teeth, because it is not entirely easy or cheap to create a sufficiently large molecular chaos from first principles, to bring dynamics into this chaos, and moreover to maintain it long enough for us to learn something about such a system.
Why is this important?
If we want to simulate the properties of amorphous systems, or even just crystals of large molecules, we often have to renounce quantum-chemical tools and settle for semi-empirical approaches or models based on classical physics. This dramatically reduces the computational complexity of our simulations but at the same time deprives us of the very valuable attributes of the quantum approach. By this I mean the practically unlimited transferability of computational methods to different materials and the immense predictive power of calculations that are practically independent of empirical inputs. In the realm of classical molecular simulations, we always start from an empirical model of a force field that someone (or ourselves) has to create for us. Since we try to use a limited number of parameters describing our system, we usually run the risk of losing detailed insight into individual interaction sites on the surface of molecules in classical simulations. All these facts were at the root of my thinking about putting together a proposal that would target the development of computationally feasible quantum-chemical molecular simulations for amorphous molecular materials such as liquids, glasses, but also their crystalline counterparts and their mutual transformations.
Where, for example, could the new methodology be used?
The project should result in an efficient computational methodology that combines advanced methods of quantum chemistry for describing molecular interactions, random processes in the form of Monte Carlo simulations, and fragmentation of a large set of randomly-distributed molecules into smaller molecular clusters. While ab initio methods ensure sufficient accuracy together with the wide applicability of such simulations, the decomposition of a complex problem into a set of simpler ones will make it possible to fully benefit from a modern, highly parallel computing architecture.
In the development of new formulations for pharmaceutic ingredients, great efforts are being invested in increasing the water solubility of very non-polar molecules typical for many drugs. Amorphous solid phases always dissolve better than their crystalline alternatives due to thermodynamic rules. A second example, emphasized even more in the grant proposal, is the need to study the amorphous phases of many organic molecules acting as semiconductors due to the limited ability of such large molecules to crystallize. Electronic components often use thin films made of these active substances, which can be arranged at most into some transitional semi-crystalline or purely amorphous structures. Such applications, I think, well illustrate the need to reliably model the properties of amorphous forms of molecular materials.
Molecular materials are characterized by their extreme willingness to exist in many modifications of both crystalline and amorphous nature. At the same time, these individual forms often differ quite a bit in their properties. By this I mean, for example, different electrical conductivities of individual forms of a certain organic substance having the potential to act as a semiconductor, or different solubilities of polymorphs of an active pharmaceutical substance. For a given dream application, we then would want to choose a form of material that exhibits optimal properties and is stable for a sufficiently long time. Because experimental development by trial and error is somewhat expensive and time-consuming, material engineers would very much like to hear from computational chemists about what conditions they need to create the desired form of a given material and how they should handle that material. Or at least, some stability ranking and comparison of properties of similar modifications for a given chemical substance. Now imagine that most quantum-chemical simulations still produce results such as that water ice sinks in liquid water, which contradicts the knowledge acquired by the majority of the general public when they’re in kindergarten. Even such a banal problem is still difficult to solve with modern tools. The comparison of two similar forms of the same substance suffers from the curse of computational chemistry, i.e. making very precise differences of somewhat inaccurately calculated values of some quantity, such as density, entropy, or electrical conductivity. Therefore, if my group succeeds in accurately describing the interactions in a chaotic ensemble of molecules, imitating, for example, an amorphous drug grain or a semiconductor layer in an optoelectronic device, we should obtain a robust computational tool for predicting the properties of molecular materials, which would be very useful in many areas of materials research. The key will be to avoid the kind of failures that have occurred with the case of water ice for many years in the computational chemists’ community.
What kind of team does this project require?
The project was designed to involve one post-doctoral fellow, approximately four doctoral students, and several undergraduate students. Individual members should be able to work as a team in such a way that individual research tasks follow each other smoothly. A positive team spirit and singleness of purpose should, at the same time, induce each team member to be motivated and responsible for their work, similar as with team sports. A key attribute of a functioning team is ensuring continuity and the sharing of skills by older team members with younger colleagues.
Have you managed to find such people?
The team core was formed relatively quickly thanks to my previous research and teaching activities at UCT Prague. I was thus able to offer project involvement to existing PhD students and close colleagues immediately. I am currently looking for two PhD students interested in computational chemistry and the modelling of molecular interactions and macroscopic properties of molecular materials. Starting this September, they would take up the still- vacant sub-areas of the project, which concern, for example, the use of machine learning to describe the interactions of large molecules, or the incorporation of first principles into Monte Carlo simulations of phase equilibria. Master and Bachelor students will also be welcome on an ongoing basis, since participation in such a long-term project represents, in my opinion, a great opportunity for broadening one’s horizons in computational chemistry and getting an idea of which direction to take later in one’s career. I expect from every student an interest in learning more; from older ones also the motivation and will to work hard on an assigned task and to manage the goals set together.
Do you have an idea where your scientific journey will lead next?
Today, I think I will remain faithful to the pursuit of intermolecular interactions in materials for some time to come. At least until quantum computers and nuclear fusion become commercially available. Then we would have so much computing power and energy available to perform calculations that I can’t even imagine what the job of computational chemists actually would be. On a more serious note, I have been dealing with molecular crystals in the pharmaceutical context for a long time. The new project should guide me further to the world of organic semiconductors, but maybe we will come up with other interesting applications. Let’s see what our team achieves in the following years.
