The roles of physics in foods

modelling, computer simulation and things you probably never knew about

David Pink^1,2  | Michael Nickerson³

Before Alberta Einstein made it big, he served as a clerk in a Swiss Patent Office in Bern Switzerland around 1903, during which time he logged in the patent for Toblerone bar for its characteristic triangular prism shape resembling that of the Swiss Alps. Since then, the link between physics and foods has grown. Physicists have been helping the food industry around the world, although less so in Canada, with product re-formulation, process optimization and ensuring food structure and quality – all through the use of mathematical equations and, now, computer simulation. In the latter, one can inspect visually the results of modelling a food and so actually see structures that one might not have guessed at. In order for Canada’s food industry to remain competitive in a global marketplace, we need to start emulating what’s being done in Europe, especially in The Netherlands and Germany. We need to begin making use of the wealth of physics expertise and computer power available in Canada to model, and thereby point the way to optimizing and enhancing their products and processes for quality, functionality and economic value. This article highlights a few instances of how collaboration between theoretical physicists and food scientists played out.

The Edible Oil Identity
Imagine that you’re a manufacturer who produces shortenings, and you have been asked to reformulate your product, to replace some of the solids with “some other ingredient”, for reasons of cost, ingredient availability, consumer demand or simply to make it healthier. As a product developer, you must ensure that your consumers are totally unaware of any changes in taste or mouth-feel – all they must know is that it is less expensive or healthier. However the problem with edible oils is that they are extremely complicated and generally, by trying to guess how to modify your edible oil for a given process, you won’t succeed. The reason is that you are conducting, at best, an “intelligent” search in a multi-dimensional space of the oil components and, mathematics says that your search will frequently not succeed. A better way to try to modify your product is to discover how it is constructed – what structures exist inside your oil so that you know, up front, what you have to maintain so that a consumer cannot tell that you have made changes.

Can you discover what you need to know by experiments alone? Unfortunately, not really. In order to understand what the experiments are telling you, you need a (mathematical) model. If you think about it, everything that you experience, you interpret in terms of some sort of “mental model”. And the best mathematical modellers are theoretical physicists and chemists: experts who can model BIG complicated systems, for LONG times. As proof of concept, we have been modelling edible oils over the past two years.¹ Initially, we began our work with what we knew about edible oils on a nanometer scale (one billionth of a meter). In 2009, Nuria Acevedo and Alejandro Marangoni² showed that the smallest stable solid structures in edible oils at room temperature were (flat) Crystalline NanoPlatelets (CNPs) with sides of about 100-1000 nanometers in length and about 10-100 nanometers thick. We figured out how to construct a simple but adequate mathematical model of them, figured out how to calculate the attractive van der Waals interactions between two such CNPs, and carried out computer simulations of the model to see what would happen as the model CNPs interacted with each other, as they rotated and moved around in the liquid oil. This simulates a simple food processing operation in which shear does not play a dominating role, though mixing and simple dynamics are involved. Over the last 2 years, we predicted that CNPs would aggregate into 1-dimensional multilayer sandwich-like structures, which, in turn, would aggregate into 10-100 micron-scale fractal (i.e., self-similar) structures. This is summarized in Figure 1. Apart from the structures shown, we predicted that, if the system is cooled rapidly and the experiment carried out, then only A and B would be observed, while, if the cooling is slow, then A–D will be observed.

We performed experiments using ultra small angle X-ray scattering at the Argonne National Laboratory (Lemont, IL, USA), and our predictions were entirely confirmed.³ With this success, we extended our models to include CNPs in multi-component liquid oils.⁴ We predicted that these CNPs could become coated with minority-components of the oil and so bind together in different ways so as to create environments that would enable the large resulting fractal aggregates to bind liquid oil within their structure. In this way, we like to think that we have made a contribution to understanding the basis for an edible oil to possess the necessary stable oil-binding capacity so that it functions as it should. These results are all in the public domain. However, our colleagues at the University of Guelph with direct connections to industries have taken note of them and, together with other information that we have provided, are exploring, in collaboration with us, targeted modifications of edible oils.

The Honey Connection
In 2014, David Pink carried out a 6-month Engage grant project with Cornect Family Farms, a small rural Nova Scotia company engaged in producing flavoured creamed honeys. They had formulation concerns and wanted precise affordable answers concerning solubilization of natural additives into their products. This was clearly an “Physics” problem instead of a “Chemical” problem and it was undertaken in collaboration with Dr. E. Papp-Szabo in Guelph. The project required an understanding of the phase separation of natural additives, and how the physics of solvation was applicable to a complex system like honey, and was a huge success. Another joint project with the same company has just been funded thus expanding the application of physics into a honey-related field.

Four horsemen of the New Apocalypse: Chicken breasts, Oil migration, Cheese and Dough
The link between physicists and the food industry, especially in Western Europe, has been strong for some decades. This is not an exhaustive review and we can only touch upon some recent examples. One should read the articles by Athene Donald.^5,6 Ruud van der Sman has employed mathematical models to understand heat and mass transport in cooking chicken breast.⁷ These are based upon differential equations which, in some cases, have to be solved using numerical integration. This work is important for one must balance heat and mass flow in a soft complex structure against minimal output of time and energy in order to minimize production costs while ensuring that the product is safe to eat and resembles what a consumer thinks a chicken breast should be. Others have used atomic scale molecular dynamics to study protein restructuring and triacylglycerol oil migration. One area that has been modelled to some extent is casein micelle aggregation and degradation. This can take place via (a) rennet-induced enzymatic activity, (b) pH change via acidification or (c) application of heat treatments. These processes underlie those used in the manufacture of cheese or cheese-based products (cream cheese, processed cheese, etc.). Are not these processes already well-known? Why involve physicists? Because, although the chemistry is understood, details of the physical structures that emerge are not. Industries are realizing that the knowledge of cheese product structures on the scale of ~1–100 microns is likely to be very important in (a) designing new products and (b) making manufacturing processes more efficient without sacrificing quality. What happens to casein micelles and their components when they are acid- or heat-treated? What structures are formed on nano- and micro-scales? And, it is not only such cheese products that pose a challenge to understanding their nanscale structure. Dough, a product that has been around for thousands of years, carries its own mysteries. Nano- and micro-scale structural details and dynamics, which are important for some processes in bread production, are hidden under a cloud of unknowing. These are a few examples of fascinating challenges to any thinking physicist who works in the field of “soft condensed matter”.

La Nouvelle Vaguee and the Doorway to the Future
Why is it that we can actually create mathematical models of food structures? Why is it that we are able to understand old phenomena and predict new phenomena, thereby pointing the way towards how to manipulate the particular food? The reason is that we have realized that food is a class of materials known as “soft condensed matter”. Theoretical physicists and chemists have been dealing with such soft materials for many decades and have developed techniques to model them.^8,9,10 The Royal Society of Chemistry organized in 2012 a conference, “Soft Matter approaches to Structured Foods” at Wageningen in The Netherlands.¹¹ And in 2015 a Gordon Conference will introduce a new scientific forum for researchers of Nanotechnology for Agriculture and Food Systems.¹² These activities show there is a tide that is now flowing and those who do not make use of it might be left behind. Turning towards the future, one growth area is involved with calculating the effects of shear upon food structures. Shearing a system is like stirring it: you want parts of a fluid to move more rapidly than other parts. If one thinks about it one realizes that shearing occurs more often than not. But, to model shear properly one must satisfy certain conditions. If one wants to cause a fluid to move in a model oil or a water-based system, then one must ensure that the procedure used is in accord with the Navier-Stokes equations, the fundamental equations describing fluid flow. But, generally, food systems are so complicated that one must use computer simulation techniques to predict new phenomena. One must therefore find a way of applying the Navier-Stokes equations to complex computer simulations¹⁰. The approach enables us to emulate shear that occurs within fluid flow within a processing plant or during mastication.

I haven’t mentioned “kinetics”: how a system changes as time goes by. This is very important to many processes and one must take care in modelling such dynamical aspects correctly. Mechanisms of relaxation, the appearance and importance of metastable states, the onset of phase transitions and changes are all processes that are important in understanding the complex materials that make up “food”.

Riding the Tornado of Change
I have twice mentioned “prediction” of new phenomena. Let’s face it: in science Prediction Rules. If your model only “explains” phenomena without predicting the outcome of new experiments, then it is not a useful model. With the advent of high-performance computers and GPUs, mathematical modelling has come of age. Aircraft, cars and ships are designed that way, so why not food? The Airbus 380 was “flown” on a computer before it ever left the ground – and it flew exactly as its computer model predicted. Drones, autodrive, Mars landers, you name it… are all designed or manipulated via computer software.

The Times They Are a-Changin’.¹³ With an ever changing food industry in Canada, a company’s competitiveness should be at the forefront of any strategic vision. Physicists can play key roles in facilitating greater process optimization, ingredient re-formulation and product quality, such that the Canadian Food Industry remains competitive in an increasingly-competitive global food environment.

References

[1] Pink, D.A. et al.(2013) J. Applied Phys. 114: 234901

[2] Acevedo, N.C. & Marangoni, A.G. (2010) Crystal Growth Des. 10: 3327.

[3] Peyronel, F. et al. (2013) J. Applied Phys. 114: 234902.

[4] Quinn, B. et al. (2014) J. Phys. Condens. Matter. In press.

[5] Donald, A.M. (1994) Rep. Prog. Phys., 57: 1081.

[6] Donald, A. (2004) Nature Materials, 3: 579.

[7] van der Sman, R.G.M. (2013) Meat Science, 95: 940.

[8] Mezzenga, R. et al. (2005) Nature Materials, 4: 729.

[9] Pink, D.A. et al. (2013) Physics in Canada 69: 93.

[10] Pink, D.A. et al. (2014) Computer Simulation Techniques for Modelling Statics and Dynamics of Nanoscale Structures, in Edible Nanostructures: A Bottom-up Approach (eds. A.Marangoni, D.Pink) Roy. Soc. Chem, Oxford, Chapter 9.

[11] Royal Society of Chemistry, Faraday Discussions 158, Soft Matter Approaches to Structured Foods, Wageningen, The Netherlands, 2-4 July 2012.

[12] Gordan Research Conference. Nanoscale Science & Engineering for Agriculture & Food Systems. June 7-12, 2015

[13] Dylan, B. (1964) Columbia Records, 13 January.

 

 

 

¹Physics Department, St. Francis Xavier University, Antigonish, NS

²Food Science Department, University of Guelph, Guelph, ON

³Department of Food and Bioproduct Sciences, University of Saskatchewan, Saskatoon, SK

 

(*Corresponding author email: scorpiocarla@gmail.com)