Foods are frequently complex, with many coexisting phases of matter, and are deliberately created as metastable materials.
To take an extreme example, ice cream has a crystalline ice phase (so in some sense is a high-temperature ceramic) but also a sugar syrup, prevented from freezing by the osmotic pressure of solutes, which undergoes a liquid-liquid phase separation because of the dissolved polymers. There is also a fat phase (which is not only crystalline, but contains liquid oil at warmer temperatures), air phase, and a whole set of two-dimensional phases where proteins and diglycerides compete for space at the interfaces between other components. Margarines, dressings, soups and sauces present similar complexities.
In all cases, the lowest free energy state is nothing like the product that is sold, so, during processing, one needs to hit a tiny and unstable target in the space of possible structures – and to do this as fast as possible while using a minimum of energy. That structure needs to stay intact for perhaps months of storage under conditions similar to when it’s eaten, but, when it is finally consumed, the microstructure needs to fall apart just so, to give a nutritious and tasty experience.
All of the steps in this chain involve complex physics that needs to be understood to improve both foods and food processing, and which can be probed through advanced measurement techniques. No single technique is able to cover to cover all of the interesting length scales in evolving food structures, from nanometres (typical of macromolecules) to hundreds of microns; and light scattering, electron and confocal optical microscopy and X-ray tomography are all brought to bear. Magnetic resonance imaging may be needed to follow the breakdown of food during digestion, where the rapidly changing pH, dilution, and hydrolysis of polymers can push food materials into domains of fractal flocculation long beloved of statistical physicists. Imaging is not the only area of concern – structure is primarily important for the effect it has on texture, so it is also necessary to measure the mechanical properties both of the bulk food (rheology), and of the surfaces of air bubbles and fat droplets, since the coarsening of structure is often driven by the free-energy gains from eliminating interfaces.
In recent years, simulation has been playing an ever larger role in the food industry. For example, new processing equipment is often tested in silico before being trialled in reality, and computational fluid dynamics can be brought to bear on anything from reducing the energy usage of new freezer cabinets to choosing the best design of teabag (pyramidal) to enhance flavour infusion. Despite this, the simulation of complex, evolving microstructures in the large-scale flows typical of processing machinery remains a serious challenge for the industry.
Lastly, it’s tempting to imagine that sophisticated measuring equipment and high-powered silicon are the sine qua non of modern food research. However, as in so many fields, there is no substitute for good thinking. Even quite pure pencil-and-paper theoretical work can be used to address interesting practical questions in manufacturing. For example, many foods are packed particulate systems, whether emulsion drops in a mayonnaise or salt crystals in a dry bouillon, and recent work on sphere packing has led to a new heuristic for predicting random close packing fractions of polydisperse particles – an important parameter that determines many mechanical properties of these types of system.
- This an extract from the IOP report The Health of Physics in UK Food Manufacturing, due to be launched at PepsiCo on 21 October. Follow the launch event on Twitter.