The Physicist as Chocolatier by Naveen Sinha
The history of chocolate is intertwined with the history of soft matter, as chocolate-makers have experimented with ways to create stable emulsions and well-structured lipid crystals. The Aztecs and Mayans first drank cocoa-based beverages as early as 600 AD. Columbus introduced the drink to Europe, where sugar was added to make the beverage more palatable. A fundamental problem with the chocolate drinks was the high fat content of the cocoa beans, which would not dissolve in the hot water, but would instead float unappetizingly to the surface. The Dutch figured out the soft matter problem of creating a stable dispersion, by removing much of the cocoa fat and alkanizing the resulting powder. Their process resulted in a stable beverage, but this created the problem of what to do with all the excess cocoa fat. An Englishman, John Fry, figured out how to use the cocoa butter to create a stable suspension of cocoa nibs and sugar: a chocolate bar.
I would like to continue this delicious collaboration between the culinary and scientific worlds by helping out two of my favorite chocolate companies, Taza Chocolates in Cambrdige, Massachusettes, and the Kakawa Chocolate House in Santa Fe, New Mexico, with their soft matter-related problems. Taza Chocolates is well-known for producing their own chocolate bars and other cocoa creations from raw cocoa beans. They are quite interested to know the particle size distribution in their chocolate, especially produced to mass-produced chocolates like Lindt. The Kakawa Chocolate House in Santa Fe is famous for their historical chocolate elixers and other cocoa-based confections. Despite their expertise in the field, their choclatier was having problems with creating a smooth coating around the locally-made truffles. Instead of a shiny, dark cover, the chocolate would form "bloom," leading to a dull, whitish appearance. I wondered whether insights from the world of soft matter could help solve his problems.
Particle Size Distribution
I began with the particle size measurement, a class problem in the field of soft matter, with several subtleties that might not be understood by all chocolate-makers. The particle size distribution has a major effect on whether chocolate tastes like the traditionally smoother European chocolates or somewhat rougher American chocolates. The optimal particle size for dark chocolate is less than 35 micron. The particle size distribution is effectively found using laser diffraction. In the traditional method, the dark chocolate is melted in vegetable oil and then placed under ultrasonic dispersion to break apart any aggregates. One commercial system available for making these measurements is the MasterSizer(R) Particle Size Analyzer for Chocolate, made by Malvern Instrume Ltd. . A schematic of a laser scattering apparatus from their website illustrates the basic principle:
Light from the laser scatters from a cloud of particles, resulting in an interference pattern of concentric circles. The spacing between these circles is related to the size of the particles, so by monitoring the light intensity over a range of concentrations, the particle size can be determined. The physical basis for this phenomenon, known as Mie scattering, assumes a dilute arrangement of spherical particles. Malvern has developed a patented technique for dealing with the multiple scattering complications that occur at high particle concentrations.
Emmanuel O. Afokawa is one of the leaders in the field of chocolate research (see a partial list of his papers below) and some of his typical results are shown below:
He observed a bimodal distribution, which food scientists have shown to have the most pleasing texture in the mouth. By having a range of scientists, it is possible that the particles are able to pack more efficiently.
To make my own measurements, I first used optical microscopy to compare three types of chocolate:
- Lindt 80% Chocolate (the company is named after Randolphe Lindt, the inventor of the conche. This sea-shell shaped device produced smoother, better tasting chocolate than anything previously)
- Taza 80% Chocolate
- Taza Mexican-style hot chocolate (vanilla)
I melted the chocolate in coconut oil, which I had leftover from earlier culinary experiments. To disperse the particles, I vortexed a small 1.5 mL Eppendorf tube for about 30 seconds. The resulting suspension contained about 0.3 grams of chocolate in 1.5 mL of total volume. This seemed to lead to the optimal concentration of particles when placed between a microscope and cover slip. I observed the samples under an upright microscope at 5X, 10X, and 40X. The lowest magnification actually showed the best differentiation between the chocolate varieties. The total field of view is slightly over 1mm x 1mm.
Sadly, the black-and-white camera did not capture the clear differences between brown cocoa particles and translucent sugar crystals, but the relatively homogeneous mixture of small particles is still apparent.
As expected by its somewhat gritty mouth-feel, the Taza chocolate did contain much larger particles, up to about 50 micron in size.
Amy Rowat, a post-doc in Prof. Weitz's lab suggested viewing the sample with crossed polarizers, which would make the sugar molecules clearly visible against the background, due to their birefringent nature. I tried this out with some 80% dark chocolate from Taza:
The crystals in the bright field image above are all comparable in brightness.
When viewed between crossed polarizers, the sugar crystals stand out clearly from the darker background.
These results provided important qualitative information about the differences among the various chocolates, but I wanted some more quantitative information about particle size distribution. I consulted Tom Kodger, a graduate student in Prof. Weitz's lab with extensive experience in particle size measurements. We used a dynamic light scattering measurement, which monitored the correlations in the scattered light intensity from a dilute suspension of chocolate particles in coconut oil. Smaller particles diffuse faster, so they have a shorter correlation time. Unfortunately, the larger particles, which were visible in the optical microscopy experiments, most likely settled out of the suspension. The remaining particles were all much smaller, less than 10 micron. As a result, we saw little difference between the Taza and Lindt chocolates:
In the legend, Series 1 is Taza chocolate and Series 2 is Lindt chocolate, but the differences are minimal. The technique is less sensitive for larger particles, so the slight differences that we did observe in the larger end of the size spectrum are probably not significant.
These results are promising, but leave much room for refinement. Difficulties arise in knowing whether the sizes measured are intrinsic to the chocolate or a result of the preparation process. For instance, if the chocolate suspension is dispersed for longer times, does the particle size continue to decrease? Also, the laser diffraction method employed by many chocolate researchers is not well suited for detailed information about particle size distribution: anything more specific that a bimodal distribution is largely a matter of interpretations.
For even greater resolution, I plan to use electron microscopy. Ellen Hodges, a post-doc at Harvard, took some images of Trader Joe's dark chocolate for Prof. Stone's Holiday lecture about the science of chocolate:
The image above is from frozen Trader Joe's dark chocolate.
The appearance changes dramatically when "bloom" forms on the surface of the chocolate. Mark Sciscenti at Kakawa chocolate told me about differences between fat bloom and sugar bloom, which would be interesting to study with the electron micrscope.
For some reason, the bloom takes on an entirely different appearance at room temperature. Unfortunately, there was not time to do repeat measurements to see if this is a regular occurance.
I will be meeting with Ellen in the near future to discuss extensions to these preliminary experiments.
A Phase Diagram of Chocolate
As a somewhat longer-term goal, I would like to help out the Kakawa Chocolate House with their truffle-making issues. The main difficulty lies with creating the proper crystal structure of the lipid molecules, known as tempering. This is often regarded as one of the most difficult aspects of chocolate-making, since improperly crystallized chocolates can be dull and crumbly.
I found a simple procedure on-line (at chow.com) for tempering chocolate at home, so I eagerly tried out the technique:
I prepared two batches: one in which I followed the procedure as best I could and another in which I added some water to "seize" the chocolate on purpose and ruin the texture. I took some photos using a borrowed IntelPlay microscope and a trial version of the imaging software. At 200x magnification, the differences were quite dramatic:
For a more complete understanding of how to temper chocolate, Dr. Morrison suggested creating a phase diagram. Such studies have already been done for complex materials like soap and metal alloys, so an equivalent study for chocolate is certainly feasible. For instance, is there a maximal temperature for tempering chocolate, above which the crystal structure will not form properly?
With the insights of soft matter physics, chocolate makers will be able to quantitatively understand the tempering process and perhaps develop new methods for making chocolate.
- E O Afoakwa et al. "Relationship between rheological, textural and melting properties of dark chocolate as influences by particle size distribution and composition." European Food Research Technology (2008) 227:121-1223.
- E O Afoawka, A Paterson, and M Fowler. "Factors influencing rheological and textural qualities in chocolate - a review." Trends in Food Science & Technology. 18 (2007) 290-298.
- E O Afoakwa, A Paterson, M Fowler, and J Vieira. "Effect of tempering and fat crystallization on microstructure, mechanical properties and appearance in dark chocolate systems." Journal of Food Engineering 89 (2008) 128-136.
- E O Afoakwa, A Paterson, M Fowler, and J Vieira. "Characterization of melting properties in dark chocolates from varying particle size distribution and composition using differential scanning calorimetry." Food Research International 41 (2008)751-757.
- S T Beckett. "The Science of Chocolate." Cambridge: The Royal Society of Chemistry. 2000.
- H McGee. "On Food and Cooking." New York: Scribner. 2004. pp. 694-712.
Naveen Sinha is currently studying biofilms in Prof. Michael Brenner's group. This class is changing the way he sees the world. On his morning runs he thinks about the viscoelastic properties of his Saucony shoes. At a cafe, he contemplates the physical properties of the artful foam on his latte. When he cooks dinner, he wonders if this class could lead to some consulting jobs for the food industry.