An Odd Path into Chemical Engineering
At the University of California Davis, I received Batchelor’s (1978) and Master’s (1980) degrees in Mechanical Engineering. In my master’s dissertation I was supervised by Harry Dwyer, an expert fluid mechanist with wide ranging interests and special expertise in finite difference solution of partial differential equations, and Brian Launder, a visitor to the department and a leading light in the field of turbulent flow modelling. When Brian took a position as department head at the University of Manchester (Institute of Science and Technology) I was enticed by the prospect of living in Europe and by a paid postgraduate position to work on a doctorate in the Mechanical Engineering department there. My sense of British engineers, such as G.I. Taylor, making powerful theoretical advances using quite basic but clever experiments to supply clues and provide verification was also an important draw.
Working on the statistical modelling of turbulent flows in which density is non-uniform (i.e. combusting flows, high Mach number flows or flows involving interaction of gases or miscible liquids of differing density) put me at a most difficult frontier of turbulence theory. Nearly all modelling before (and since) invokes the simplification of uniform fluid density. I managed to complete my Ph.D. (1985) without making any significant breakthroughs and came away painfully aware of, as I saw it, the poor predictive performance of turbulence models in relation to the great mathematical complexity involved. Sheffield University came into view for me during this period through research exchanges with Jim Swithenbank’s group in the Fuel Technology department and also through an elegant 1953 paper on self-similar characteristics of non-uniform density turbulent jet flows authored by an earlier professor in the department, Meredith Thring.
I joined Michigan Technological University (the Department of Mechanical Engineering and Engineering Mechanics) in the small ex-copper boom town of Houghton, there teaching Thermodynamics, Fluid Mechanics along with an advanced course on Computational Fluid Dynamics and Turbulence modelling, which I co-taught with George Huang who was a postgraduate student with me in Manchester and who went on to write codes for high Mach number flows over aircraft for NASA. While at Michigan Tech I spent two summers as a Research Fellow at NASA Lewis (since renamed Glen) Research Center and that led to a move to Princeton (1987) to work on turbulent spray combustion in the Engine Combustion Lab led by Frediano Bracco in the Mechanical and Aerospace Engineering department.
The lab was at the centre of efforts at that time to introduce better control of the combustion process in petrol and diesel engines and we participated every few months in working groups set up by the government to bring universities, government labs and vehicle companies together to make better engines. Fred also believed strongly in that sort of collaboration in the organisation of his lab, which was in a space divided down the middle with experiments on engines taking place in one half and computations of engine combustion processes carried out on banks of computers in the other half, the two groups meeting together each week. In the computational work, we used stochastic models to track the interaction of spray droplets with surrounding gas and with each other, so droplet collisions, coalescences, breakups, evaporation, heat exchange and, ultimately, combustion reaction could be represented in detail. Typically, many tens of thousands of sample droplets needed to be simultaneously computed along with the flow, temperature and gas species fields, with computations lasting days (and sometimes weeks). My work concentrated on modelling the droplet interaction with turbulent eddies and, along with papers presenting straightforward simulation studies of the spray flow process for certain commercial injection arrangements, a paper was published demonstrating a bias involved in usual methods used to select the gas fluctuation velocity that would determine particle motion at each instant along its path. One of the referees complained that a particular paper had not been referenced. I added that reference noting that the method described in it involved the sort of biased sampling of fluctuation velocity addressed in our paper, which would produce false accumulation of spray droplets at the edges of a spray jet.
A major presence on the experimental side of the lab was Phil Felton, master of experimental techniques and in particular the use of lasers and laser diagnostic techniques. Phil had moved from Sheffield where he had contributed to development of what became the Malvern particle size measurement instrument based on light diffraction.
In 1993 I moved with my family to Sheffield to take up a position as Lecturer in Fuel Technology in what has become the present Chemical and Biological Engineering Department. At that time the department was in transition from its past as a Department of Fuel Technology towards a then as yet unknown future under the temporary leadership of Jim Swithenbank. We were a joint administrative department with Mechanical Engineering with eleven staff and year-group student numbers in the 20s, but beginning to build back up towards a viable independent department. I was the second new academic hired, following Maria Vahdati (now at Reading University). Yajue Wu and Bruce Ewan soon joined and then the department began to enlarge more rapidly as first Ray Allen, then Mike Hounslow and then Philip Wright were recruited in succession, each bringing in several new academic posts. As the department unfolded from its Fuel Technology origins towards first process engineering with chemistry (Ray), then core Chemical Engineering, but with added emphasis on particle technology, (Mike) and then biology and genomics (Philip) I was kept busy working out what the department I had joined was about!
My teaching in the department began with leading a third year design project group, guided by the expert Kenneth Littlewood, and a Process Economics course, guided by some chaotic lecture notes of a pervious academic, a concise monograph on the subject produced by the I.Chem.E. and time spent working out mortgage equations when buying my first house, over in Manchester! Gradually, I progressed to teaching first and second year fluid mechanics courses, a technical drawing course (descriptive geometry), a second year computing element (Fortran programing, Matlab and Aspen Plus) and the Computational Fluid Dynamics course at fourth year level that I taught for some fifteen years.
Research directions developed slowly. Briefly, I continued with spray research, but conversations with UK colleagues at conferences in the area suggested there was slim hope of funding in that area. A couple of papers were published with Peter Foster who was a much appreciated mentor. These papers arose from work with an Erasmus student on self-similar combusting swirl jets (in effect extending the Thring non-uniform density jet work to confined swirling jets and also related to a paper I co-authored at Princeton comparing liquid versus gas fuel injection into an internal combustion engine). Briefly also I tried to link with local industry, building a scale physical model with BOC (in Holbrook, just east of Sheffield) to examine heat transfer when using lancing jets in electric arc furnaces. Jets of heated air were used to simulate the lance jets and, to ensure a well-defined boundary condition for later computations (which never took place), a ‘thermal’ mixer was developed based on the commonly employed strategy in aircraft and power plant combustors of using swirling flow for rapid mixing of fuel and oxidant streams.
Eventually, I began collaborating with Ray Allen in the area of microchannel flows in a series of cross-discipline EPSRC grants with chemistry departments, first at Hull University (Paul Fletcher and Steve Haswell) and then with Queens University, Belfast (Ken Seddon and Chris Hardacre). Ray generously developed the broad ideas, linked up with collaborators and won the funding, and allowed me to write a specific work package fitting with the project aims. These projects led to development of theoretical and, as always with good theory, practical advances in first DC electrokinetic flow networks and then in effective mixing of highly viscous liquids (ionic liquids were the subject of the work) in microchannels. In each case, first principles flow computations were linked with experimental measurement to maximise the understanding that can be achieved. In this approach, one is never fully sure whether the experiment is to check the computation and theory or whether the theory and computation are to check the experiment! But the net effect is that experiments can be far more targeted and theory more precisely verified.
Over the last dozen years or so, I have focused on an idea that arose from considering how a process like distillation might be carried out in microchannels. When first exploring Chemical Engineering, I was struck by the ‘disorganisation’ of the flows involved in the fluid phase contacting equipment that is widespread in chemical and biological processing. In the aerospace engineering field of my original training, all efforts were made to ‘clean up’ the flows to avoid the frictional losses associated with flow separation. I was aware that for heat and mass transfer (mixing) chaotic flows might be a necessary evil. But it also seemed clear that standard phase contacting arrangements such as distillation and extraction columns were ineffective. So, when Geof Priestman and I came upon the idea (in around 2004) of using a rotating spiral channel as a way to organise fluid phases to allow counter-current contacting and independent control of each phase flow rate and the diffusion scale in each phase, I sensed an important line of work had been found and I have worked to develop the fundamentals and practice of this general technology steadily ever since then. This has been achieved with a series of students supported by the governments of Mexico (Jaime Ortiz-Osorio), Libya (Mohamed Zambri) and Iraq (Ahmed Ayash).
Last year I lead my final group of MEng research project students through the trying initial months of the pandemic, as it happened. I last lectured a couple of years ago, deciding at about that same time to turn my efforts towards pushing forward application of rotating spiral contacting. After nothing developed from initial interest in using the spiral for membrane-less flow batteries, a chance hallway conversation with Esther Karunakaran has developed into a collaboration that attracted support from Unilever to develop a commercial bioreactor. The cells can be formed into a highly uniform biofilm layer by the rotation and the spiral allows both an adjacent sustaining media flow and a parallel flow of organic liquid to extract product species. Also, the mixing work I did some years ago is being applied, in collaboration with Xiubo Zhao, to form supersaturated liquid solutions for economic pharmaceutical nanoparticle production. By forming a uniform mixture sufficiently rapidly (in a time much smaller than particle formation time scales), the stochastic variations in particle properties that result from formation under the non-uniform concentration conditions during the mixing process are avoided. Success depends crucially on optimising the mixer design and operating conditions. Unlike in the early work, the liquids involved are relatively inviscid and the turbulent flow regime is accessible.
While it is true that I have retired, it looks like I may be around for just a little bit longer!