Day Five: All good things must come to an end

April 9, 2007

The last day of the Chocolate Technology course was short, starting with a chocolate troubleshooting session with Ed Seguine, and a long lab session working on tempering and various ways of modifying chocolate. We also did blind testing of a few commercial milk chocolates, attempting to discern what milk “backbone” was used in each product.

Ed Seguine is the VP of R&D at Guittard Chocolate, and is a world-class expert on chocolate. (He wrote the US standard on measuring chocolate viscosity, and has written a tall stack of research papers.) He’s an inspirational speaker, combining an energetic love of chocolate with a limitless depth of knowledge. His first law of understanding chocolate and troubleshooting problems is: “Think Like the Fat.” Crystallized fat makes up the structure of chocolate, and the root of most chocolate confectionary problems is something disturbing this structure. Bloom is the result of liquid fat leaking out of the structure and crystallizing on the surface. Seized chocolate is a result of moisture altering fat crystallization. “Thinking like the fat” is his key to diagnosing problems in the kitchen or factory.

Most of the morning session was spent on small case studies. Some examples:

* How do you deal with chocolate that has been seized to mud by moisture? Consensus: throw it away. Even if you can save it, water presents more problems that just awful texture and viscosity. Chocolate is a super-safe food, because it is very, very dry. Chocolate has so little moisture in it that nothing can really live on it or in it. If you introduce moisture, this inherent safety disappears. Water entering chocolate causes crystallization around droplets, creating little potential microbial islands.

* How do you establish where problems originated when a customer is getting bloomed chocolate? You have to track possible points of temperature variation (what weather has the truck travelled through? has it sat on a loading dock?) Also, having an organized, temperature controlled program to retain samples of every shipped product can help check the status of the product.

The rest of the day was spent in the lab, with part of the class hand tempering, and the rest experimenting with the effects of various processes on chocolate. Some of the modifications that were investigated:

* Denaturing chocolate by overheating. Heating milk chocolate to 200F significantly alters its viscosity. Dark chocolate also changes, but less dramatically.

* Adding water to chocolate. This is quite dramatic. Adding even a very small amount of water to liquid chocolate makes it grainy, and basically turns the mass to a fudgelike consistency.

* Adding lecithin and PGPR to chocolate. Lecithin is a ubiquitous chocolate ingredient, used to lower the viscosity of chocolate without the need to add additional cocoa butter to get the same effect. PGPR (polyglycerol polyricinoleate) also lowers vicosity, but in a different way. To understand the difference, a short detour into “non-newtonian fluids” is needed. Ideally, fluids are just more or less viscous, that is, they are “thicker” or “thinner”, and some fluids do operate this way. Chocolate (and many other fluids) are different, in that they have some resistance to going from rest to motion (called Yield Value), and some resistance to flowing once they are moving (called Plastic Viscosity.) Lecithin and PGPR alter these values very differently. PGPR is used extensively by Hershey’s and other big manfacturers (look on the ingredient label for a Hershey’s Kiss.)

The last activity of the course was to blind taste three commercial milk chocolates, and attempt to discern if they were made with milk crumb, roller dried milk, spray dried milk, or skim milk powder with added anhydrous butter oil. All of these processes produce dried milk, but with different mechanical and temperature stresses. These different stresses produce different flavors and melting textures. The milk crumb chocolate was immediately evident, as it has a very strong caramel flavor from the heating of the milk sugars. Cadbury is famous for milk crumb chocolate. The Guittard chocolate was spray dried, which has buttery kind of flavor with an even melt. Spray-dried milk is the default way of producing dried milk. The third chocolate was a Dove chocolate, which uses a combination of spray dried skim milk powder with added butter oil. The milk flavor in this chocolate emerges later in the melt, and the butter is more evident. The class overall did a good job at correlating tastes with the different milk chocolate processes.

The notes I’ve blogged are hugely incomplete. Spending five days with noted chocolate scientists is a great way to learn how chocolate works, but also how complex this topic can be. I’ve learned a lot of valuable fundamentals, and also learned the several hundred ways that I was messing up my chocolate. I’m looking forward to getting closer to my goal of a really top-class dark chocolate!


Day Four: Faux-co-late

April 7, 2007

The themes for Thursday at the Chocolate Technology course were “coatings”, how fats work, and the fun of panning. Everyone is pretty familiar with fat, and it’s pretty important since it’s the basic structural element of chocolate and how it behaves determines how the chocolate behaves. Coatings and panning are slightly more esoteric, though everyone reading this has probably eaten a lot of coating and a lot of panned chocolate. One’s a little scary, and the other is the route to the goodness that is the malted milk ball and its cousins.

Let’s start with the scary part: coatings. There are times when a food manufacturer is making a product where they want some chocolate attributes, but normal chocolate is either not durable enough to survive the handling, or the manufacturing process is going to cause problems with persnickety cocoa butter. Real chocolate, if you take it over 94F, either in cooking or in handling, is going to lose crystallization, and when it cools, is going to break down and bloom. Visually, people will read bloom as “mold” and you are going to have an unsellable product. (Note that even products with “real” chocolate will modify it a little to make it more shelf stable. Looking at the Archer Farms Cherry-Pistachio-Dark Chocolate Granola bars in our pantry, the chocolate is real, but contains butter oil. Why? The butter oil forms a eutectic mixture with the cocoa butter, and prevents the chocolate from forming waxy Type VI crystals.)

So, how does the food industry solve this problem? Coatings! These are chocolate-like products that either supplement or replace cocoa butter with a more stable fat. The result is “engineered chocolate” that doesn’t break down, and doesn’t require tempering for stability and texture. For example, take a look at the ingredients in Kellogg’s Special K Chocolatey Delight. (The term “chocolatey” should be a clue that we’re not dealing with by-the-book chocolate here.) The FDA mandates that to be called “chocolate”, you can only contain the following ingredients: cocoa, cocoa liquor, sugar, cocoa butter, lecithin, vanilla (or vanillin ester), and dried milk. Looking at the ingredients, Kellogg’s is clearly painting outside the lines here. There is sugar, cocoa, lecithin, and milk, but they are hanging out with some strange neighbors: partially hydrogenated palm kernel oil, cocoa processed with alkali, and artificial flavors. No cocoa butter at all!

Kellogg’s has probably found that chocolate chunks floating around in a cereal box tend to have some serious problems (and that cocoa butter is a very expensive ingredient.) This “chocolatey” compound they are using is classified as a coating. The three non-chocolate ingredients are there for very specific reasons. The partially hydrogenated palm kernel oil is an easy to work with (and cheap) substitute for cocoa butter. It can be refined (fractionated) in such a way that it crystallizes with a very high melt point, so the “chocolatey chunk” doesn’t melt or degrade in a hot warehouse or truck, and it requires no tempering. The partial hydrogenation alters some of the fatty acids in the triglycerides that constitute the fat so that it crystallizes and melts in a way that’s closer to cocoa butter. The cocoa processed with alkali is cocoa that has been treated with potassium carbonate to remove acid (due to cheap beans), and to impart a “dutched chocolate” taste. The artificial flavors are there to help compensate for the sensory qualities lost in this engineering. These kind of engineering feats are very common in chocolatey baked goods, and can be hard to detect because the mouthfeel of crunchy cereal or a cookie can mask the texture difference between chocolate and a coating.

We spent the first part of lab today playing with a number of different coatings, seeing how they set up at various temperatures, and tasting them in their naked form. These coatings come in a lot of different grades, depending on their intended application. They range in taste from pretty nasty to pretty good approximations of real chocolate. The fats used to replace cocoa butter range from simple partially hydrogenated soybean and cottonseed oils to very sophisticated oils that have been fractionated, hydrogenated, and interesterified. The lecture after lab was all about what those terms mean….(in simplified terms, which I’m going to surely mangle and oversimplify even further.)

Thalia Hohenthal gave a lecture in the afternoon on basic fat chemistry, and how fats can be manipulated to make better cocoa better replacers or equivalents. Basically (high-school chemistry mode here for a sec) most fat comes as a triglyceride, which is a glycerol backbone that connects to three fatty acids. These fatty acids are carbon chains with acid ends that vary in length (12, 14, and 16 atoms are common), and bond-types (saturated fats have all carbons connecting to two hydrogens, while unsaturated fats have some carbons connected by energy rich double bonds that are less stable.) The type and order of the fatty acids attached to the backbone determines how the fat crystallizes (and determines melting point) and melts (how fast it melts at various temperatures, which determines mouthfeel.) The combination and variety of fatty acid types in chocolate determines how it crystallizes into it’s various forms. Other fats work differently, so engineers mess with them to make them more cocoa-butterish.

The first basic tool is hydrogenation, which eliminates double bonds and adds hydrogen. Generally, this straightens the shape of the fatty acid, which allows for tighter packing, better crystallization, and hence a higher melting point. For example, if you take vegetable oils, which are typically liquid at room temperature, and hydrogenate them, you get Crisco, which is solid at room temperature. Why do this? Liquid oils can’t really hold together structures, but solid ones can. This is why we have tamales, pie crusts, and all the great stuff you can make with butter. However, liquid oils are plentiful and often the byproduct of activities like soap-making. If you are making soap, you can hydrogenate your oil byproducts, get margarine and Crisco, and start “educating” the public about the supposed health threats of butter and lard.

The second tool is fractionation. Oils are very often mixes of fat types, some of which melt at higher temperatures than others. By cooling the oil, you can pour off or otherwise get rid of the liquid oil, and take the higher melting solid oil. This gives you a smaller amount of high meltpoint fat without hydrogenation. Fractionated palm kernel oil (which is different than palm oil) gives a fat with a melt profile that is pretty close to cocoa butter, and is the basis for the most common coating.

The third tool, which must be labelled in European ingredient lists, but not American ones, is interestification. The fatty acid-glycerol bond is an ester bond, and interestification swaps the order that the fatty acids join to the glycerol. So, a fat that might have a kinked oleic fat first on the chain, followed by two straighter saturated fatty acids can be “shuffled” to put the oleic fat in the middle, which yields a molecule that will crystallize much more tightly than the unaltered fat. Interestified coatings are used often in soft baked goods where a plastic, flexible fat is needed to coat without breaking.

This stuff definitely fell into the category of interesting, but not something I think I’ll be applying in any chocolate manufacturing I’m doing. Knowing all this trivia is great for reverse engineering ingredient labels, though! This concludes the mangled high-school science part of the post.

The last part of the day was spend doing panning, which is the process of coating a center like a peanut or malted milk ball with chocolate. Since hand painting chocolate onto individual peanuts with a paintbrush would make for a pretty expensive product, panning was invented. To do this, you use a big, semi-closed bowl (see phto in previous post) with a load of the center. The bowl is rotated continually, and you drizzle untempered chocolate in. The ingredients rotate over each other, and gradually build up a chocolate coating. The trickery comes in preventing the natural consequences of chocolate coated stuff bouncing around in a bowl. You use forced cool air to make the chocolate set up quickly, thus preventing the ingredients from sticking together or to the wall of the bowl. After many slowly drizzled ladles of 110F chocolate, a nice thick coating is achieved. The next steps are to add some starch and corn syrup, which polishes the outside of the candy and provides a protective layer so that you can put another shellac layer on the outside. The shellac armors the candy, and provides a nice shine. Do all this, and you get the world’s tastiest Whopper.

Ed Seguin, one of the lecturers visiting from Guittard, explained why malted milk balls are so fantastic right out of the panner (or up to 24 hours after.) Chocolate at room temperature contains ~20% liquid fat bound up in the solid fat matrix. When you eat chocolate, the solid fat breaks down, and the liquid fat is released quickly, taking chocolate flavor with it. It takes a while to reach the 20% liquid state, though. Fresh chocolate has significantly more liquid fat, and so, when you eat it, releases a boatload more chocolate goodness in that rush of liquid fat onto the tongue. This also makes the chocolate feel “meltier.”

Little do they suspect…

April 6, 2007

That soon this panning machine will turn on, and these innocent malted milk centers will be slowly smothered in 115F chocolate. To further seal their fate, they will be polished, have a thin layer of starch added, then finished with shellac. I wasn’t really intending on learning this stuff, but I have to confess that making these was a lot of fun. Over the course of the afternoon, I went from the example of how not to do panning to having the best coatings in class!

Behold the Enrober!

April 6, 2007

This is a very tiny version of an enrobing machine, used to cover things in chocolate. The PC in the background is used to control the belt speed and three temperature zones in the cooling tunnel.

The Instructors

April 6, 2007

Thalia Hohenthal (Senior Scientist at Guittard Chocolate) and Terry Richardson. Terry is the founder and head instructor. With over 35 years of solving chocolate problems, he probably knows more about making chocolate than anyone else on the planet.

No, I’m not going to share!

April 6, 2007

Yes, we made chocolate covered malted milk balls today!

Day Three: The Chocolate Factory

April 5, 2007

Today was 100% lab work, mainly working around a chocolate enrober. This is the machine that coats stuff in chocolate. It works by creating a curtain of tempered chocolate that the center passes through, then sends it down the line to a cooling tunnel that sets the chocolate up. The other ways of coating in chocolate are shell molding, which lets you create goodies with a liquid or semi-liquid center, and panning, which creates items like chocolate covered peanuts or raisins. Being that my dream is to create a chocolate on par with Pralus or Amadei, it’s not really a core concern for me, but it was fun to learn how to make chocolate sit up, beg, and roll over.

My lab station was to run a Tricor Tempermeter, which takes in liquid chocolate, thinks about it for a while, and tells you if it contains the right amount of crystallization. How? More temperature tricks. Crystals are more stable than amorphous mixtures, so when crystals are formed, some amount of heat is released (called, imaginatively enough, the heat of crystallization.) so, the machine takes the liquid sample to a reference temperature, then starts chilling it. At a certain point, crystals will start forming very quickly if the chocolate is tempered. This is going to create heat from the chocolate. So, if you graph the temperature of the mass as you chill it at a constant rate, tempered chocolate will get a curve with an inflection point. No inflection, no temper. Too much inflection, and you are overtempered. Because going out of temper can ruin a whole production run, you have to keep measuring temper, so I was a busy guy.

We were also running experiments in keeping a chocolate in temper for several hours, and “statically” tempering chocolate by just keeping it for a very long time at 91F. We had to measure the temper of these results every hour, while also tracking the temper of the chocolate in the enrober. Doing this enough times, I got to the point that I could predict temper by looking at the surface of the chocolate and how it set up.

We did a long comparison session in the lab also. This was divided into four parts: testing the previously made chocolate, tasting different roast levels, tasting different conching levels, and preparing for taste reverse engineering milk chocolate on Friday.

Previously made chocolate tasting. On Monday, we started making dark chocolate (at about 50% cocoa content, which is pretty sweet for my taste) with a conventional refiner/conch, dark chocolate in a ball mill, milk chocolate with spray dried milk, and milk chocolate with milk crumb. We tempered most of these yesterday, and tasted them today in the unrefined, unconched, and finished forms. All of them turned out pretty well, though the conventional dark was overconched, and hence pretty flat. Conching gets rid of bad stuff (acetic acid), but also all the good tasty notes. The milks were very good also. The interesting part for me was how well the ball milled chocolate turned out. This would actually be pretty easy to make in small batches! Basically, to make it, we put the basic, unrefined ingredients into a sphere formed by two welded together salad bowls. Along with the ingredients, the instructor put some stainless steel shot of various sizes, and hooked the whole assembly to a motor. The tumbling motion of the balls refined and conched the chocolate at the same time, and it turned out better than the conventional stuff. Building the mill looked pretty doable, and I’ll experiment with this when I get home.

Roast/conch levels. The instructors brought out liquid dark chocolate made with lightly roasted, medium roast, and burnt beans. The difference was pretty dramatic. The lightly roasted chocolate was pretty acidic, the medium roast pretty chocolately, and the burnt had huge grassy/seedy/beer notes. The former brewpub owner in the class actually preferred this roast, as he said the notes were reminiscent of the bitter parts of a really dark beer. I would have loved to do a much wider tasting, with a matrix of bean types and roast levels. We also tasted dark chocolate conched for 24 hours and 48 hours. The 48 hour, as predicted, had a lot of flavor removed. It was sweet, and a little chocolatey, but missing any of the bitter/acid/astringent/fruity notes from the 24 hour. An interesting note from the discussion after this tasting: Guittard (and others) age their high end chocolates! Aging allows the sugar particles in the chocolate to take up the flavor from the volatiles in the cocoa solids, leading to a a deeper, longer lasting taste. Sugar treatment is also critical. Since you want small sugar particles, you typically pulverize the sugar. Pulverized sugar has lots of surface area, and it’s not protected like crystallized sugar is, so it starts taking up any ambient flavors in the air. Different types of sugar also have different levels of crystallization, and also different levels of ash in the sugar. (Yes, that’s right, ash…it happens whenever you grind or heat stuff, apparently.) Crushing the sugar particles along with the cocoa in the refining process should lead to a cleaner taste, since the vulnerable surface area is protected.

Reverse engineering. We had a training panel also, using five different milk chocolates, leading up to a competition on Friday to see who can best decipher the contents of various commercial milk chocolates. The “milk” in milk chocolate comes in many different forms (as described in Monday’s notes) and we tasted five today: spray dried, skim milk powder with added butter oil, cream powder, roller dried, and crumb. They all melt differently in the mouth, and they all have different residual dairy tastes, ranging from cheesy to buttery to creamy to caramel. We’ll see how I do on Friday establishing that I’m not just imagining this. We also discussed Hershey’s chocolate, which is really, really wierd! If you actually taste Hershey’s chocolate, you’ll notice some very distinct dairy/acidy notes. This chocolate formulation is ONLY sold in the United States. Even Canada can’t deal with this stuff! Where does this come from? Milk fat and lipases. Lipases break down lipids (ie, fats), and Hershey’s is hypothesized to be full of the stuff. This breaks down the milk fat into other compounds that form that taste, well, kind of barnyardy. Why do Americans like it? Dunno, Why do the British like Marmite?

Lastly, did a little shell molding today and made chocolate covered cherries. I’ll skip the details for now.