Experiments in roasting and grinding cacao beans to make milk chocolate at home

(Return to my main page for more articles about chocolate or to browse 70 other topics: everything from Knitting Nancies and metal detectors to the strange world of lucid dreaming.)

WARNING Working with chocolate involves dealing with hot materials and fire. No one should do so without the supervision of a responsible adult experienced with all the processes and equipment involved. The following is provided for information only and is not intended as a recommendation to repeat these or any other experiments.

The best advice I can offer to anyone interested in trying to make their own chocolate at home is to log onto Mr. John Nanci's outstanding site titled Chocolate Alchemy ( There you will not only find everything you need to learn about all the steps needed to make your own designer chocolate, you will also find clear and practical descriptions of the different kinds of beans and how each affects the resulting chocolate. This excellent site is also a retail source for many different types of beans and equipment, all at reasonable cost. Even if you decide not to make chocolate, this site is interesting and enjoyable to read. I've found none better anywhere on the Internet.

(Because I ordered my beans from Mr. Nanci and conferred with him on several chocolate making issues, his name and website will be mentioned repeatedly on this page. This is done only to give credit where it's due. I am in no way connected with this site, own no stock in Mr. Nanci's company, and derive no financial gain from him in any way. Rather, it is I who own him a debt of gratitude for providing a site with the information I needed, offering to sell supplies and materials virtually impossible to get anywhere else, and taking time out of his busy schedule to exchange emails with a newcomer to get me started in this interesting hobby.)

Why Make Chocolate at Home?

The reason I wanted to try and make milk chocolate at home is that after tasting 76 brands I never found one that satisfied all of my preferences, which were a pure chocolate flavor without any fruity, mocha, or caramel notes that's smooth and mild. I also wanted to experiment with making chocolate using only commonly available kitchen implements rather than purchasing exotic and expensive equipment.

This page is a running chronicle and I intend adding to it regularly as the experiments progress so please check back often. I promise to record my failures as well as any successes so that people interested in making their own chocolate can profit by my mistakes.

Step One: Buying the beans

The first question that needs to be answered when making chocolate is: What type of beans to use? Since I knew what I wanted in the end product, I emailed Mr. Nanci at Chocolate Alchemy and asked for his advice. He suggested I use a base of Ghana Forastero with some Venezuelan Carenero Superior for flavor. I ordered two pounds of each as well as a pound of cocoa butter and an ounce of lecithin. The order arrived promptly, professionally wrapped and in perfect condition.

My order


A full-sized picture of the Carenero Superior beans.
Mostly brownish, the light powdery coating is residue
from the fermenting (a euphemism for composting)
process used to separate the beans from the pulp
that surrounds them in the pod.


The more lightly-composted Forastero beans didn't have
as much of the powdery residue from composting, either
that or they has been cleaned. As warned in Mr. Nanci's
bean description page, these may be difficult to
shell and winnow after roasting. This will be one
of the first experiments I tackle.


The soy lecithin is a soft granular material
that breaks up easily into a powder.


The block of cocoa butter had a waxy texture when handled and tasted. It had an odd aroma than I haven't been able to classify and almost no flavor. Both types of beans had a mild cocoa aroma overlaid with an musty, moldy smell left over from the composting process.

Even before roasting the cocoa beans can be broken easily with the fingers to show the randomly segmented interior structure of cocoa nibs.

An unroasted cocoa bean broken in half


The husk or hull is paper-thin, tough, and flexible. This gave me the idea that since the hull didn't appear to be firmly attached to the nibs the two could be separated by simply rolling each half of the broken bean between my thumb and forefinger. It worked and was very easy and quick to do.

Cocoa nibs expelled from an unroasted bean


This suggests that if someone where willing to spend a little time they could hull and winnow enough nibs before roasting. These could then be roasted to produce a chocolate uninfluenced by flavor elements resulting from the hull and the composting by-products stuck to it. This is an experiment I plan to do. It's possible that the results could be good or bad. The hull protects the nibs from some of the heat of the roasting process and in so doing protects them from being over roasted. It, or the compost residue sticking to the hull, may provide interesting and necessary flavor components. On the other hand, the hull and compost remnants may also add negative flavor elements. Only testing will tell. I would be interested in hearing from anyone whose already tried this.

Update: I roasted some unhulled beans and found it was hard to prevent them from getting burnt.

My First Roast:

Keeping with the philosophy of using only things available in the average kitchen, I decided to experiment with roasting a technique where the beans were simply strewn over a cookie sheet in a single layer. For these initial experiments I used the Forastero beans.

The first roast started off with a 425 degree oven for 7 minutes, then 325 for eight minutes, and finally 260 degrees for 10 minutes, as recommended by Mr. Nanci in Chocolate Alchemy. I had to stop the roast after only three minutes because the beans were popping all over the place and were already burnt. The problem was that I had misread Mr. Nanci's instructions, which recommended this roasting profile for larger quantities of beans than the single layer I was using. Mine were heating too fast.

My second roasting attempt held the oven at a constant 300 degrees. I removed samples of the beans at ten minutes (tasted raw and uncooked with a weak chocolate flavor), 20 minutes (good chocolate flavor with just the slightest hint of over roasting), 30 minutes (noticeable over-roasted flavor), and 40 minutes (strong burnt taste.) From this I decided that an 18 minute roast in a preheated 300 degree oven was good for a single layer of beans.

I roasted several handfuls of beans in this way for the next step.

IMPORTANT UPDATE!!! Before placing beans on a cookie sheet to roast them, take a look at the bottom of the sheet. If it's bright, clean, and shiny everything's okay. But, if it's dark, bewarned that the dark coloration, whether from burnt-on soot or made that way by the manufacturer, will collect heat much faster than a shiny sheet and could cause the beans to burn.

Hulling the beans:

The quickest way to hull a lot of beans was to place them on a board and lightly run over them with a rolling pin. The problem was that the nibs were shattered into such small pieces that separating them from the hulls was difficult.

Next I tried the much slower method of simply breaking a bean in half with my fingers and gently rolling the halves between my forefingers and thumbs. This expelled the nibs with very little breakage and kept the hulls together in mostly large pieces that were easy to winnow out. Still, there was a lot of cocoa lost no matter how I tried to blow the hulls away.

Finally, I discovered a very tedious but clean way to separate the beans from the hulls. (This is only good for very small batches of beans and even then only by someone who has a lot of patience.) I found that I could scratch away a small section of the hull using a small serrated paring knife then use my fingernail to peel away the rest of the hull in much the same way as peeling the shell off a hard boiled egg. The pieces of the hull were discarded and the cocoa bean was left (usually) intact in a single piece.

Roasted and shelled cocoa beans

It took two hours to shell 9 ounces of roasted beans, which provided 8 ounces of clean beans. (I later discovered a problem with this technique. Like walnuts, cocoa beans have thin paper membranes between some of their internal segments and even a little on the surface of the bean. Without some type of air-flow winnowing these bits of membrane, though much finer than the hull, may nonetheless cause some grittiness in the final chocolate.)

On to the grinding:

This is where I ran into trouble. Although Mr. Nanci warns that food processors do not do a good job, since the simple-technique-without-buying-any-new-appliances philosophy prevented me from purchasing the juicer he recommends I was forced to experiment with a food processor.

I assumed that the problem was that it would not grind the cocoa nibs fine enough. To help the processor along I got the idea of preheating it and the beans in a 160 degree oven, thinking that a higher temperature would aid in the liquefaction and grinding. It may have, but not enough.

Placing the beans in the food processor I turned it on and waited. After just a few moments the beans were reduced to the coffee-ground stage and it appeared that a lot of steam was being expelled from the beans. Sniffing indicated that it wasn't smoke from blade friction overheating the grind, so I continued. Five minutes later I noticed that the grind was breaking down and starting to take on a wet look. After another five minutes the cocoa had a mud-like consistency and the flow of steam was noticeably less. After ten minutes the cocoa had been reduced to a thin liquor and there was no steam coming from it. I waited a few more minutes but since no change was observed I turned to processor off. The temperature of the chocolate liquor was 130 degrees and it flowed easily.

Pouring out my first batch of chocolate liquor

The liquor was grainy but had a very pure and strong cocoa flavor. I noticed no burnt notes or fruitiness.

Mixing the milk chocolate:

This is where things really turned bad. I tried using a food processor, a blender, and even a small coffee mill but none of them could grind granulated sugar into a fine powder (the powdered sugar in stores has cornstarch in it which will make chocolate gritty.) The same problem was present with trying to convert the powdered milk into a finer powder. Since that was the best I could do I had to use them

Following Mr. Nanci's milk chocolate recipe in Chocolate Alchemy I measured out 6 ounces (by weight) of liquor, 17 grams (about 2 tablespoons) of melted cocoa butter,
1/4 teaspoon of lecithin, 2 and 3/4 ounces of powdered (almost) milk, and 4 ounces of powdered sugar to produce a milk chocolate that was 45 percent cocoa. All the ingredients were mixed in a food processor until as smooth as the equipment permitted, just a few minutes minutes.

My first batch of milk chocolate

The last step: conching:

The best, though admittedly not good, conching system I could devise using only stuff found in the kitchen used a mixer set at its lowest speed and fitted with one beater. The chocolate was poured into a tall glass and illuminated with a 25 watt spotlight to maintain its temperature. The entire Rube Goldberg setup was held together with duct take and let to go on its merry way. (The mixer tended to overheat so I alternated two hours on with half-hour off cycles to let it cool.)

When cooling and at night I covered the chocolate with insulating layers of aluminized bubble wrap to keep it warm. After approximately 48 hours of actual conching time I tempered the chocolate using the marble-board-and-spatula technique (please see my Tempering page) and poured it into a foil mold to cool. Here's the result: a 10-ounce bar of homemade milk chocolate.

My first chocolate bar

And how did it taste? It is undeniably the grittiest, nastiest chocolate I've ever had. The sugar was so coarse that at first I thought I'd bitten into a Hershey's Crunch bar. But, after forcing myself to ignore the texture, I could tell that the chocolate itself wasn't that bad: it had a very strong, almost bittersweet, flavor without any burnt notes. Adding a little more milk powder and sugar should mild it out to be more in line with my taste preferences.

The Final Analysis:

I calculate that this one bar of pitifully poor chocolate cost me $21.00 and eight hours of work. Valuing the labor as unskilled, certainly appropriate in my case, and worth $5.00 per hour, that means this chocolate cost $100 per pound. Was it worth it? YES! And I'm indebted to Mr. Nanci for providing the information needed to help my do it. Sure, the chocolate was a disaster, but I had a lot of fun making it, acquired a greater appreciation for professionally made chocolates, and look forward to figuring out ways to make better chocolate.


Home Chocolate Making Version 2.0:

I tried several new ideas the second time around, some worked... others where complete disasters.

I reduced the oven roasting temperature to 280 degrees and the roast time to 16 minutes. The resulting beans had a taste that wasn't as sharp as the first roast, suggesting that I had over roasted it. From 8 ounces by weight of raw beans I got 6 ounces of chocolate liquor.

As before I was limited by the charter of this page to using only commonly available items for making my chocolate. For the grinding stage that meant using a food processor. To eliminate the grittiness I had the idea of pouring the chocolate liquor onto a large piece of cloth, gather the edges together and twisting the neck to create a forced strain of the liquor. Failed miserably. I tried it with light cotton and a light knit and the liquor was still far too thick, even when still hot and thin enough to pour easily, to squeeze through the mesh. I twisted until my face turned blue and nothing came out.

I had slightly better luck creating finely ground powdered sugar and dried milk. After playing around with several food processors, blenders, and coffee mills, I discovered that the coffee mill produced the finest grind. I found that it worked best when there was only a little material to grind, one tablespoon. A steady grind of two minutes got it as fine as it was going to get. Passing this product through a super-fine mesh cooking screen removed the largest unground particles. It still wasn't as fine as store-bought powdered sugar but it was much better than that used in the first chocolate making attempt.

I used a rock tumbler half filled with marbles and the almost-powdered sugar from above and let it run for 24 hours thinking that it would act like a ball mill and grind the sugar up finer. Didn't work. I could detect no noticeable improvement.

Finally, since the rock tumbler didn't work as a ball grinder, I decided to use it to drive a conching machine thrown together out of odds and ends laying around my work bench. Now don't laugh... it worked.

The paddle that stirred the chocolate was made from the bottom of a rubbermade plastic food storage container. The chocolate holder was an 8-inch diameter Tupperware container. The spot light provides enough heat to keep the chocolate at 110 degrees. The container rotated at 1 revolution every 2 seconds. The rock tumbler is great for this sort of use because it is designed to run for weeks without attention. It's important to provide some sort of dust protection or the chocolate could get gritty from things floating in the air that you wouldn't want to be eating. However, the cover has to be of such a nature that it still permits free circulation of air so that the volatiles that need to be driven off from the chocolate are free to escape.

Final Results: The resulting chocolate wasn't much better than the first attempt. The main problems to be addressed are eliminating the grittiness in the chocolate and getting a finer grind on the powdered sugar and milk. It looks like it's back to the old drawing board!



Please click HERE to visit my main site and browse 90 other pages covering very thing from the lost art of knitting Nancies to metal detecting.