Research Interests

Research Background

Lyophilization, better known as freeze drying, is an increasingly important step in development and manufacture of pharmaceutical products—the number of FDA-approved lyophilized drug products has increased significantly in the last decades (from less than more than 10 new approvals each year before 1994 to more than 30 a year since 2016). Fundamentally, many modern drugs like mRNA vaccines and other biologics have a short lifetime in the aqueous state, and lyophilization is a useful separation step when drugs are also thermally sensitive. That same thermal sensitivity sets the fundamental limits of how quickly lyophilization can proceed, since drying rate is closely coupled to product temperature. Since the bulk of lyophilization costs is in amortized capital cost, an economical lyophilization process runs as quickly as possible within thermal limits, and this requires an effective mathematical model of the process to balance tradeoffs. My experience is in refining and developing new models, building on the state-of-the-art model (which is now 40 years old) for lyophilization in a vial. A major thrust of my work has been working on models which account for the addition of microwave heating, an up-and-coming technology with the promise of accelerating vial-on-shelf lyophilization but opening doors to many more geometries and dried product forms which are impractical in a typical lyophilizer. Another area has been in developing predictions of mass transfer resistance, which was the focus of my Fulbright grant spent collaborating with experts at the Politecnico di Torino in Italy. Though my interest lies especially in theory and computational methods development, models are of course only useful insofar as they predict reality, and I have had to not only gather experimental data but interpret, analyze, and use it to check validity and behavior of models. I am equipped to bring together experiment, theory, and simulation for the world of lyophilization and pharmaceutical manufacturing at large. https://younginstitute.research.purdue.edu/collaborative-accord/

Proposed Future Directions

Emphasis 1: make models easy to use for non-experts

My work has benefitted enormously from open source software—for example, although I understand the basics of high-order time integration tools for ordinary differential equations, being able to use cutting-edge algorithms implemented by performance experts is better than reimplementing everything myself. I try to make a point of contributing back to libraries I use, improving their usability and accessibility—experience shows that even the fastest and most flexible code is useless without good documentation and tutorials. In a similar vein, my experience collaborating with and training industry scientists and engineers has shown that new science is useful to industry insofar as it can be easily understood and implemented by the people closest to an industrial setting. My experience maintaining LyoPRONTO (a Python library with a web GUI for assisting lyophilization process design) has taught me that this does not happen automatically. Models need to have an easy-to-understand, accessible interface to see broad adoption and utilization, and these interfaces need to be easy to understand and easy to maintain. A major goal of my research group will be making models (both the mathematics and the implementations) accessible to non-experts, such as with a graphical interface that permits exploration and practical process design. Regulatory approval is a critical step for any pharmaceutical technology, which can involve not only stating the process design and conditions but justification. We will pay attention to creating robust interfaces for models, drawing on software engineering practices such as unit testing and continuous integration to increase the (actual and perceived) reliability of model implementations. This can facilitate technology adoption and regulatory approval processes.

Emphasis 2: end-to-end model for mass transfer in lyophilization

The mass transfer resistance of the partially-dried product is key to understanding the drying rate and therefore duration of lyophilization. The state of the art model for conventional lyophilization empirically describes this resistance with three empirical parameters that are known to depend on virtually every aspect of the process, from formulation composition to freezing and even the drying conditions themselves. This limits the usefulness of tabulated parameter values, and since each measurement of mass transfer resistance requires a full lyophilization cycle, process development can easily become very material-intensive and time-consuming. Part of my PhD work has been directed towards predicting the mass transfer resistance for lyophilization without requiring a new lyophilization cycle each time. This problem requires robust understanding of ice crystal nucleation and growth during freezing, mechanical behavior of porous product structures as ice crystals sublimate away, and rarefied vapor flow through porous structures—topics which have begun to get answers in scientific literature, but which will take probably years more of interdisciplinary effort to assemble into a coherent, generalizable model for lyophilization. My lab will work toward this goal, developing experimental expertise (such as in pore visualization), computational simulations (for porous flow), and collaborations with experts in neighboring fields (such as thermodynamics of crystallization and freezing).

Emphasis 3: expand to new and other drying technologies

As highlighted by the Lyo2040 technology roadmap, there are a variety of emerging drying technologies which address the weaknesses of traditional lyophilization, such as spray drying, spray freeze drying, microwave-assisted lyophilization, and “lyo beads”. The relevant mechanisms of heat and mass transfer differ from one process to another, and from a physics perspective each involves some unique phenomena (such as spray atomization or the microwave electromagnetism), but all of these techniques fundamentally deal with removing water from a temperature-sensitive aqueous solution. In this, the lessons learned from one drying approach may be useful to another—particularly as regards technology development and regulatory approval. My lab will work towards developing models with a unified framework for understanding these processes, not only in mathematical notation but numerical implementation, facilitating head-to-head evaluation of which methods might be most suitable for a new drug product, given some preliminary low-cost measurements (such as freeze-drying microscopy and direct scanning calorimetry).