A Rewarding Life

At Pittcon 2015, I received the Dal Nogare Award from the Chromatography Forum of Delaware Valley. I was thrilled because, as I explained to the audience, all four of my chromatography mentors – J. Calvin Giddings, my post-doc adviser, Peter Carr who is also a dear friend, Jack Kirkland who has mentored me for 30 years and is a wonderful person, and the late Georges Guichon, to whom I was very close – were recipients of the Dal Nogare Award. Mary Ellen McNally read one of the seconding letters; a long list of my accomplishments written by Georges made the occasion very special.

I guess my “big-ticket item” is that I have more than 30 years of simulation science experience to bring to bear on chromatography and other separation problems. It’s proven to be powerful for solving problems that are far too difficult and complex for mathematical investigation and for which experimental investigation is not definitive. Simulation is the third paradigm of science and I was probably the first to bring these types of techniques, which require advanced computer power. But it’s not all about introducing new techniques, it’s about using them to unveil the mechanism of separations.

Let me elaborate. I’ve worked with two types of particle simulation. One is a molecular simulation technique called “Configurational Bias Monte Carlo in the Gibbs Ensemble”. This works very well for phase equilibria problems, which is essentially what chromatography is about. In chromatography, no one seems to be able to agree on the separation mechanism because it requires “further study” or it is very difficult, if not impossible, to elucidate experimentally. Ilja Siepmann from the University of Minnesota Department of Chemistry and I have been working on this problem for over 15 years.

Ilja had developed Configurational Bias Monte Carlo, which we believed to be the best option ever for obtaining equilibrium results. Other simulation techniques that looked at liquid chromatography (LC) used molecular dynamics. Having worked in that field for years, I knew you could not rely on molecular dynamics to provide good equilibrium results. However, the Configurational Bias technique coupled with the Gibbs Ensemble (developed by Athanassios Panagiotopoulas at Princeton University, Princeton, New Jersey, USA) worked well for phase equilibria. So, we used this approach to solve problems such as how does reversed-phase LC work. We knew we’d got it right because we compared our energetics – free energy calculations – to experimental results and they were nearly an exact match. This matching with experimental results is typically how you validate a simulation, and, in our case, the result was striking.

The other type of simulation I’ve worked on is Monte Carlo Transport; it is also known as Brownian dynamics and molecular dynamics, but it is a particle-based technique –  not molecular-based. It’s used to solve transport problems. One example of this is simulating the flow through a packed bed incorporating retention, convection and diffusion – all the elements of packed-bed LC. And we’d already solved the dispersion calculations you need for LC by running tracers through model packed beds under non-retention conditions. This is a very powerful simulation technique, but let’s be clear, it’s not molecular, it’s transport based. You begin with the flow field, send your tracers through the porous medium and watch their arrival times as they leave the column. It’s just like watching a physical solute leaving a chromatography column. This approach has been applied to packed-bed LC, pore diffusion in ion exchange chromatography, capillary electrophoresis and a host of other techniques used in separation science.

It also has applications for field-flow fractionation (FFF) and its various versions; in fact, we figured out a long-standing mystery of what Coriolis forces do in sedimentation FFF. From our simulations, we discovered that when people did sedimentation FFF, they were rotating the channel in the wrong direction. It made me realize how powerful these techniques were because you can figure out from the simulation what the experimentalist may not know. So, simulation is a closer match to the actual experiment than it is to theory-based investigation.

Keeping it practical

One thing I’ve also known is that it is important to keep practical applications in the headlights! Indeed, if you are doing anything – and this includes deep theoretical and deep simulation work – it needs a practical outlet. Let’s take reversed-phase LC: if you know how it actually works, you can play with the parameters, which gives you a better feel for what the experimentalist wants to do with the technique. It also gives you a test bed for further discovery and optimization.

That point really hit home with elucidation. Where do the solutes you want to separate go into the phase – do they embed in the bonded chains or do they sit on top? And, of course, you can do this without actually having to run a series of experiments. This is clearly something that interests industry and I’m all for sharing academic work with industry, having spent many happy years working in industrial settings. Indeed, Uwe Neue also had good academic connections and he knew the power of doing basic experiments.

For most of my industrial career, I’ve been a modeler and have had to deliver practical results. I could use any approach I thought would shed light on the problem and that generally entailed using a lot of computing power. For example, there was an ion exchange division at Rohm and Haas and I would often be asked by my colleagues to explain how things worked at the molecular level. I’ve also done many biomolecule calculations to investigate binding sites – the business recognized that asking me was the most practical way to get an answer, especially because computing power had increased a lot, relieving us of inexact simulations or the need to solve differential equations by mathematical methods that often gave broad-brush but not definitive results.

As computing became cheaper and more powerful, and people began building their own computer clusters, investigations became increasingly more practical. It encouraged them to predict that in 10 years the majority of their work would be simulation-based with people running powerful computer programs. The reality of the 10 year estimation is proving elusive and today we are still writing our own software for about half the work we do, even though there is more commercially-available problem-solving software on the market.

Industry meets academia

Uwe Neue understood the need to link industry with academia and he reached out to universities to augment his own knowledge and studies. For my own part, I’ve been an adjunct professor in the chemical engineering department of the University of Delaware for 20 years. The department has many bright people – faculty and students – and I’m fortunate to work with them. Additionally, I consult for Advanced Materials Technology in Wilmington, Delaware, USA, and the company willingly funds work-study students from the chemical engineering department. I currently have four people working with me – all honors students and exceptionally gifted.

I also have good access to other professors whenever I need their insights. It’s great to have such high caliber people to work with to augment my own knowledge and experience, which is important because separation science is interdisciplinary and you need input from academics who understand other issues that may not relate immediately to injecting a sample and seeing a chromatogram.

I’ve been lucky to have good co-workers at the University of Minnesota as well – that’s where Peter Carr is. Pete introduced me to Ilja Siepmann and continues to provide input to our studies, which have been funded by the National Science Foundation for the last 14 years.

The team approach

Academia is clearly different to the industrial world. In industry, you concentrate on product development and effective utilization of resources to execute a development process. This isn’t really a focus of university work. Nevertheless, universities like Delaware are becoming more team-oriented even though academia generally focuses on basic research. That said, I sometimes had to do basic research in industry as part of the product development cycle – in reality, there is not a very clear separation between the two. Indeed, the most successful people in industry are those who can work on both sides.

I said earlier that separation science is interdisciplinary – especially in theoretical separation science. I need to be a computer guy, a chemical engineer, a mathematician, and even an electrical engineer, which is one of my lesser-known specialties. And when you run out of ideas you have to find someone with a different knowledge base who will become part of your team – and that’s the case in both industry and academia.