Welcome to our second deep-dive into flavor. I’ve had more requests for chocolate than for all others put together, and I’m not surprised. Who doesn’t love it? And who hasn’t had problems with it? Chocolate’s up there with coffee when it comes to technical challenges, but it presents its own unique vexations.
Worldwied Chocolate
Chocolate also resembles coffee in that it’s become more like wine over the last decades, with the emergence of small artisanal producers, the availability of single origins, and a growing appreciation for its astonishing range of flavors. As with wine and coffee, these flavors are the product of different growing regions, different varietals, and different processing methods. The interplay of genetics, terroir, aging, and fermentation have led to complexities that science has barely begun to decipher.
This makes our investigations more complicated, as well as more interesting. Not so many years ago, a pastry chef could confidently say that brand x was the best chocolate. Today the question doesn’t even make sense. Best how? What flavors do you want? What textures? What other qualities? And how could you even sample all the chocolates that might be contenders?
We’ll discuss selecting chocolate a bit later. First let’s look at the basic challenges of making ice cream with good texture and a depth of chocolate flavors. If you
The Problems
With coffee, we faced the difficulties of extracting all of a great coffee bean’s flavors into the dairy. This isn’t a big deal with chocolate—you don’t even need to extract anything; just throw the chocolate into the mix. And chocolate’s flavors play well with sugar and dairy, and can handle a fair amount of dilution before the more subtle elements get obscured.
But getting the texture right with chocolate can be a bitch.
More specifically, if you want great chocolate flavor, with impressive intensity, it’s challenging to also get good texture. The culprit is cocoa butter, which typically makes up half or more of the cocoa mass. 100 grams of unsweetened chocolate will contain 50 to 60 grams cocoa butter. 100 grams of 70% bittersweet chocolate will contain 35 to 45 grams cocoa butter. Unlike milk fat, cocoa butter is solid at room temperature, and hard as a rock at freezer temperatures. Generally speaking, your ice cream’s texture will be better with less cocoa butter, and best with close to none.
One way to accomplish this is to just use less chocolate. But if this seems like a reasonable solution, you and I have nothing more to discuss. Please get out of my kitchen.
Another solution is to replace the couverture chocolate with cocoa powder. Cocoa is generally 10% to 25% cocoa butter—significantly lower than chocolate. The trouble is that great quality, distinctive cocoa powder is still a rarity. Even the best manufacturers, like Valrhona, Amedei, and Michel Cluizel, typcially make cocoa powder as a byproduct of cocoa butter production, and so pay little attention to origin or distinctiveness. These companies make dozens of varieties of single-origin couverture but will typically offer just one or two fairly generic cocoa powders.
There are a few exceptions. At least one large chocolate producer, Callebaut / Bensdof, offers a range of five single origin cocoas. But they don’t distribute these widely or sell at retail. There are a handful of smaller artisanal producers, some of which have a good reputation, but they don’t have the milling equipment to produce a fine powder; their cocoas will give ice cream a gritty consistency.
This is a shame. As with coffee, my interest has been in bringing out the most interesting and compelling flavors that chocolate has to offer. And to do so without any textural compromises. I believe that ultimately the solution will be found in top-quality single-origin cocoa powders, but we may have to wait for the industry to catch up and make this pursuit practical.
Our Solutions
I’ve formulated two recipes. One is a pure cocoa powder recipe that I’ll start testing when I find some worthy cocoa powder. The other is for right now, and represents a compromise—it gives the best chocolate flavor I’ve been able to manage without ruining the texture. This recipe uses a mix of couverture and cocoa powder. Let’s call it “double origin.” I don’t think most people will think it tastes compromised; so far it’s the best chocolate ice cream I’ve had. But I believe it falls short of what’s theoretically possible.
Here are some principles behind the “double origin” formula:
1. 90g / L single origin dark chocolate. Tested with Michel Cluizel Vila Gracinda (67% cocoa solids). This is the highest quantity before I start to see texture problems.
2. 70g / L cocoa powder
3. reduced cream to milk ratio, to keep total fat level to 15%, and the milk fat to 10%
4. no eggs
5. 14% total sugar. This is typical for most ice creams, but is much higher than my usual 11% level. Added sweetness compensates for the bitterness of cocoa.
6. larger proportion of dextrose and fructose, to counter the hardening effect of the cocoa butter
7. no skim milk powder. solids levels are already very high from the cocoa and chocolate
8. increase in guar and carrageenan, to compensate for the lack of thickening from the egg yolks
9. added lecithin, to compensate for the lack of emulsifier from the egg yolks
10. increased cooking temperature, to help hydrate the lecithin
Recipe 1: “Double Origin” Chocolate Ice Cream
(to make 1000g / 1.2L Updated 4-2019)
470g whole milk
70g cocoa powder (must be excellent quality. tested with Michel Cluizel. Valrhona and Pernigotti/ChefShop should be good too)
84g dextrose powder
29g fructose*
1.2g salt
2g soy lecithin
0.8g locust bean gum
0.6g guar gum
0.4g lambda carrageenan
90g bittersweet chocolate (67–72% cocoa solids)
(must be very high quality. tested Michel Cluizel Hacienda Los Ancones 67%)
240g heavy cream
10g vanilla extract
*or use 45g trimoline and decrease dextrose to 50g. add trimoline after blending solids.
*******
-set circulator to 80°C / 176°F
-thoroughly mix the dry ingredients (not the chocolate)
-measure milk into blender.
-turn on blender to lowest speed that makes a vortex. pour in dry ingredients.
-blend on high for one minute.
-add chocolate. blend on high until incorporated—probably 2 minutes to melt and emulsify chocolate.
-Add cream, trimoline (if using), and vanilla extract. Blend briefly.
-pour mixture into 1gal ziplock bag.
-add weight (recommended, to keep bag from floating) and evacuate the air.
-cook in water bath for 45 minutes to hydrate stabilizers and partially denature milk proteins.
-gently agitate bag after 5 and 10 minutes. if you see air accumulated in the bag after 15 minutes, release it, and carefully reseal bag.
-mix will be pasteurized (pasteurization time after reaching this temperature is under 2 minutes).
-remove bag from water bath. open and pour hot mix into clean blender container (or a square container if using a homogenizer or stick blender). remove weight (with tongs). use bag to squeegie off any mix. temporarily seal bag and keep handy.
-blend on highest speed for 60 seconds to homogenize.
-chill bag in ice water bath (use ice bath to evacuate the air when sealing bag). carefully agitate to cool. Try to cool to refrigerator temperature.
-refrigerate at least 8 hours, below 38°F / 3°C to age mix / pre-crystalize fat.
******
-snip off bottom corner of bag, and squeeze out mix as if using a pastry bag, into an open container (1.5L mixing bowl is ideal). mix will have formed a stiff gel from the emulsified and hardened cocoa butter. blend with a stick blender to thin texture (a whisk attachment will work best).
-scrape into ice cream machine; spin. With a mulitispeed machine, use a slow setting. Ideal drawing temperature is 23°F / -5°C or below.
-evaluate when surface texture of ice cream first looks dry. if it needs more overrun, continue on higher speed. if it needs to cool more, continue on lower speed.
-harden for several hours (preferably overnight) in a cold freezer. freezer should be set to -5°F / -20°C or lower. Ice cream will have to warm up several degrees before serving. 20 to 30 minutes in the fridge works well. Ideal serving temperature is 6 to 10° F / -14 to -12°C.
Analysis:
Total Fat: 15.5%
Milk Fat: 10.2%
Total Solids: 43.1%
Solids Nonfat: 27.6%
Milk Solids Nonfat: 5%
Acidity: 0.08%
Alcohol: 0.5%
Stabilizer/Emulsifier: 0.38%
Egg Lecithin: 0%
POD: 100 / 1000g
PAC: 227 / 1000g
Milk Fat: 10.2%
Total Solids: 43.1%
Solids Nonfat: 27.6%
Milk Solids Nonfat: 5%
Acidity: 0.08%
Alcohol: 0.5%
Stabilizer/Emulsifier: 0.38%
Egg Lecithin: 0%
POD: 100 / 1000g
PAC: 227 / 1000g
Tasting notes: I think this succeeds at presenting as a dark chocolate, rather than a milk chocolate, with little sense of anything between you and the chocolate itself. However, the flavor of the cocoa powder outweighs the flavor of the couverture by around 2 to 1. You can still taste the single-origin chocolate, but the flavor of the cocoa powder is the limiting factor here.
**********
Onto the single-origin cocoa recipe. When we can get our hands on the right cocoa powders, this will take us close to what’s possible with chocolate ice cream.
Here are differences from the “double origin” version:
1. more cocoa, no chocolate
2. higher cream to milk ratio, since there’s less cocoa butter to compensate for
3. closer to normal levels of dextrose and fructose, since there’s less cocoa butter to compensate for
Recipe 2: Single Origin Cocoa Ice Cream
395g whole milk
110g single-origin cocoa powder, best quality (very hard to find. ideally use non-Dutch process, which is even harder to find. Recipe assumes 15–20% cocoa butter content)
80g sugar
50g dextrose powder
12g fructose*
30g nonfat dry milk
2g soy lecithin
1.2g salt
0.8g locust bean gum
0.6g guar gum
0.4g lambda carrageenan
310g heavy cream
10g vanilla extract
*or use 30g trimoline and decrease dextrose to 38g. add trimoline after blending solids.
*******
-set circulator to 80°C / 176°F
-thoroughly mix the dry ingredients (not the chocolate)
-measure milk into blender.
-turn on blender to lowest speed that makes a vortex. pour in dry ingredients.
-blend on high for one minute.
-Add cream, trimoline (if using), and vanilla extract. Blend briefly.
-pour mixture into 1gal ziplock bag.
-add weight (recommended, to keep bag from floating) and evacuate the air.
-cook in water bath for 45 minutes to hydrate stabilizers and partially denature milk proteins.
-gently agitate bag after 5 and 10 minutes. if you see air accumulated in the bag after 15 minutes, release it, and carefully reseal bag.
-mix will be pasteurized (pasteurization time after reaching this temperature is under 2 minutes).
-remove bag from water bath. open and pour hot mix into clean blender container (or a square container if using a homogenizer or stick blender). remove weight (with tongs). use bag to squeegie off any mix. temporarily seal bag and keep handy.
-blend on highest speed for 60 seconds to homogenize.
-chill bag in ice water bath (use ice bath to evacuate the air when sealing bag). carefully agitate to cool. Try to cool to refrigerator temperature.
-refrigerate at least 8 hours, below 38°F / 3°C to age mix / pre-crystalize fat.
******
-snip off bottom corner of bag, and squeeze out mix as if using a pastry bag, into an open container (1.5L mixing bowl is ideal). mix will have formed a stiff gel from the emulsified and hardened cocoa butter. blend with a stick blender to thin texture (a whisk attachment will work best).
-scrape into ice cream machine; spin. With a mulitispeed machine, use a slow setting. Ideal drawing temperature is 23°F / -5°C or below.
-evaluate when surface texture of ice cream first looks dry. if it needs more overrun, continue on higher speed. if it needs to cool more, continue on lower speed.
-harden for several hours (preferably overnight) in a cold freezer. freezer should be set to -5°F / -20°C or lower. Ice cream will have to warm up several degrees before serving. 20 to 30 minutes in the fridge works well. Ideal serving temperature is 6 to 10° F / -14 to -12°C.
Total mass: 1002g
Milk Fat: 126g / 12.5%
Cocoa Butter: 22g? / ≤5%
Total fat: 148g? / 15%
Cocoa Solids Nonfat: 88g / 9%
nonfat milk solids: 82g / 8%
total sugars (non-milk) 142g /14%
total solids nonfat: 312g / 31%
total solids: 460g / 46%
stabilizer 0.18%
emulsifier 0.2%
Appendix 1. Some Chocolate Basics
 |
| Cocoa beans fresh out of the roaster. Thanks to Michael Laiskonis at the ICE Chocolate Lab. |
Cocoa % or cocoa solids %: In plain chocolates—unflavored and non-milk chocolates—this refers to everything besides sugar. It’s the cocoa mass from the cocoa pod. It will be very roughly half cocoa, half cocoa butter. So a 70% dark chocolate will be about 35% cocoa, 35% cocoa butter, 30% sugar. With some chocolates the cocoa butter can be as high as 60 or 65%. And with some specialty chocolates it can be as low as 45%. The best chocolate producers publish this information, so you don’t have to guess at what you’re working with.
Do not assume that higher cocoa % is always better. The best chocolate producers are striving for balance; they’re not chasing numbers. The best chocolates I’ve ever had have been around 67%. I’ve had some lousy 85% bars.
Origins: I haven’t written a tasting guide to the different regions because I just don’t know enough. And I suspect it would be pointless. There is so much variation from one small producer to another that the generalizations just don’t hold up very well.
When it comes to sheer pleasure and interest, though, I can easily say that the best chocolates I’ve had have been single origins. This term is surprisingly difficult to define. Does it mean from a single farm? A single Cooperative? A single region? A single country?
As with coffee, the precise definition varies from one place to the next. While there may be cases where the term is used in bad faith, by marketers who want to sell you a cheap blend, I’ve never had this experience with a quality chocolate maker. The better single origins taste like a distinctive expression of … something. Whether it’s a hillside, a region, or a nation seems less important.
Cocoa Powder Types: “Dutch” process cocoas are treated with an alkali, which alters the appearance and flavor. Dutched powders will be a darker, richer red, but the flavor will be milder, with less bitterness and astringency. Since dairy and sugar both take the edge off of chocolate’s flavors, you may find you can get a more intense flavor experience from natural process cocoas.
Cocoa Powder Types: “Dutch” process cocoas are treated with an alkali, which alters the appearance and flavor. Dutched powders will be a darker, richer red, but the flavor will be milder, with less bitterness and astringency. Since dairy and sugar both take the edge off of chocolate’s flavors, you may find you can get a more intense flavor experience from natural process cocoas.
That is, if all else is equal. Which it never is. Most European cocoas are only available as Dutch process. The quality of the individual powder is more important than any theoretical difference in its processing method.
In baking, the distinction is important; if you switch between Dutch and natural, you’ll change the pH, and will often have to compensate with changes to your leavening ingredients. This is one area where ice cream is more forgiving. We only worry about about the fat and the flavor.
Which brings us to the fat: check the cocoa butter percentage. Cocoa usually has more than you’d expect, and the high-end brands (annoyingly) usually have the most. Be prepared to compensate for high fat levels.
Appendix 1. Chocolate Variety Tasting Notes
 |
| Similar to the Coffee Wheel. Courtesy Barry Callebaut |
The Chocolate industry divides cocoa trees into three major species:
Criollo
Criollo cocoa beans account for less than 5% of the world’s production, due to their succeptibility to numerous blights. Partly because of its rarity, and partly because of its delicate flavor profile (which emphasizes fruit and other long-lingering secondary flavors over the more bitter and and astringent baseline chocolate notes) it’s considered a delicacy. Most blends that include criollo use it in small proportions.
Criollo is native to Central and South America, and the Caribbean island of Sri Lanka. Subspecies include Andino, Pentagana, and Porcelana.
Forestero
Much more robust than Criollo, Forastero cocoa comprises over 80% of the world’s production. It has strong bitter notes, basic chocolate flavor, and fewer of the fruity and acidic top notes than the other varieties. It tends to come on strong and have a short finish.
Forastero cocoa is native to the Amazon basin, and today is grown in Ecuador, Brazil, and much of equatorial Africa. Subspecies include Amelonado, Arriba, Cundeamor, and Calabacillo. While Forestero’s reputation is as a commodity chocolate, some subspecies and some farms present exquisite examples. Michel Cluizel’s Vila Gracinda, one of my favorite culinary chocolates, is 100% Amenolado.
Trinitario
Trinitario is a hybrid of Criollo and Forastero varieties, attempting to merge delicacy of the former with the robustness of the latter.
Trinitario is believed to have origninated in Trinidad, and is now grown in Mexico, the Caribbean, Colombia, Venezuela, and Southeast Asia. It’s a dominant componant in many fine chocolates.
In Conclusion Refutation
Rather than a conclusion, this appendix gets a refutation: almost everything I’ve written about these cocoa varieties is bullshit. I had to include it, because it’s the conventional industry knowledge, and without this information you won’t know what anyone’s talking about.
Which is to say, the conventional wisdom is mostly nonsense. Modern genetic testing tells us that there aren’t three cocoa varieties; there are at least eleven. They get grouped as three based on superficial resemblance, geographical accident, and lore. Most growers are dealing in hybrids, and genetically speaking, have no idea what the provenance of their trees might be.
Which is ok, if your not a botanist or agricultural anthropologist. Don’t worry about the labels. Worry about flavor. There are some fine chocolate review sites for the unfortunate instances when you can’t manage to taste everything yourself (see below).
For further reading:
Appendix 2. Chocolate Review Sites
Appendix 3. Where to buy Chocolate
(Their shipping prices for small quantities have become insulting.)
Appendix 4. The Future: Single Origin Cocoa Powders
 |
| Bensdorp / Callebaut Natural Process São Tomé |
Bensdorp / Callebaut
(Tanzania)
(Ecuador)
(Ghana)
(Honduras)
(Several origins)
And some halfway-there althernatives: Low Cocoa Butter Couvertures:
Permalink
What is the difference between cocoa powder and cocoa nibs? Can nibs be easily converted into suitable powder? I've noticed that single origin nibs are much easier to source.
Permalink
Permalink
My first chocolate batches were rock solid, and they remained so even after resting at room temperature for up to half an hour. What wasn't rock solid was either melted ice cream or the few chunks and scrapes I was able to break off of the primary mass. And those pieces could hardly be called "chewable", at least not in any normal sense of the word.
For my most recent batch, I followed your single-origin recipe closely, resulting in relatively identical values for each category. Of course, I make a few modifications to reduce carbohydrates, but it is otherwise quite similar. It is certainly not soft out of the freezer, but only a five to ten minute rest is required to get a decent piece. It is still hard, but not nearly as bad as the first batches.
What can I do to continue to improve this aspect of my chocolate ice cream? Ideally, I would like to be able to scoop it right out of the freezer, but I'm not sure what commercial makers do to make that possible. If I can't scoop it right out, I would like a better texture after a brief rest. My very first caramel ice cream was scoopable out of the freezer days after I made it. It had all the flavor of home made ice cream with the texture of a commercial batch. I wish I could replicate that with my chocolate ice cream.
Here's a brief look at my macros by percent for a 1016 gram batch:
Water Gums Yolks Fat Milk Fat MSNF Sugar Other solids Total solids POD PAC55.4 0.250 0.0 13.0 12.0 8.0 15.2 8.8 45.4 130 437
Any ideas would be more than welcome. Thanks! -Woody
Permalink
I'm having trouble reading the numbers you posted (not your fault; I'm having problems with the way WordPress messes with comment formatting). So my answers will be more general.
I'm assuming you've read the Sugars in Ice Cream post (https://under-belly.org/sugars-in-ice-cream/) and the Solids/Water/Ice post (https://under-belly.org/ice-cream-solids-water-ice/).
Excessive hardness in chocolate ice cream is either from too much water ice or too much cocoa butter.
We solve the water ice problem by increasing the solids (which reduces the water %), or by depressing the freezing point. We mostly use sugars for this (including the lactose in milk and skim milk powder). These methods reduce the ice fraction—the percentage of the mix made up by frozen water.
We solve the cocoa butter problem by either reducing the amount of it (more cocoa powder, less couverture) or by using freezing point depression and solids level to drop the ice fraction even lower—compensating for the hard fats by reducing the amount of ice.
You say you're reducing the total level of carbohydrates in your formulas. How are you doing this? Are you sure that you're calculating the freezing point depression correctly for the ingredients you're using?
A note on freezers: unless you're using a commercial dipping cabinet, your freezer is probably colder than what's ideal for serving ice cream. It ought to be. Dipping cabinets are typically around -14°C (7°F), or for Italian-style gelatos, as warm as -11°C (12°F). A standard freezer should be -18°F (0°F) or even a little colder.
So it's perfectly reasonable for ice cream to be too hard to scoop right out of the freezer. But it shouldn't be like concrete, as is the case with many home ice cream recipes or unbalanced commercial low-sugar ice creams.
Permalink
Permalink
Take a look at what happened at Jeni's Homemade several years ago: https://www.foodsafetynews.com/2016/08/130962/
Their kitchen was a certified milk processing facility, with a staff food safety expert and twice-monthly official health inspections. And still, they had two major incidents where listeria was found onsite, and ice cream had to be recalled nationally to keep anyone from getting sick or dying. The burden of dealing with raw milk safely ended up being too much for them, and they eventually offloaded most of their process to a commercial dairy.
If you're a small shop or pastry kitchen or are doing this at home, you are taking big risks for very little benefit. And unless you plan to invest in a homogenizer, your ice cream will be worse for it.
I'd recommend instead buying organic, homogenized, low-temperature pasteurized milk and cream from a local dairy or farm cooperative. You'll get all the benefits and none of the risks or problems.
As far as stabilizers, just use the right ingredients. There's nothing unnatural about gums. Locust bean gum is essentially a flour made from nuts of a carob tree. Carob is the same stuff hippies use as a chocolate substitute. Guar gum is just a flour made from guar beans. Carrageenan is an extract from seaweed. If you look hard enough, you can probably find certified organic versions of these, but my position is: why bother. It's all good stuff, and it will make up just a fraction of a percent of the ice cream.
For emulsifiers, lecithin is fine (get good stuff that has virtually no flavor. I have good luck with Will Powder). And don't overlook plain old organic egg yolks, for flavors that play well with egg.
For chocolate, use whatever you like, but do an honest blind taste comparison before you choose. Much of the certified organic chocolate and "cacao" powders don't taste very good. Meanwhile, almost all the highest quality, single-origin chocolate is organically produced. It's just not certified, because none of these plantations in the middle of the tropics is paying for US or European certification. But they're growing the cocoa the way they've done it for a hundred years, and there's very little about their process that's been modernized, for better and for worse.
As you can probably tell, I care much more about ice cream flavor, texture, and quality than I care about any kind of "label friendliness." I'm not going to sacrifice actual quality to satisfy an uneducated customer's misplaced ideas about quality. It's almost a moral stand—I want reason and knowledge to prevail over prejudice and ignorance. If we can't make this happen with ice cream, how can we expect it to happen in global politics?
Permalink
Permalink
I might be talked out of this position. Dana Cree, in her excellent book "Hello, My Name is Ice Cream," argues from a pastry chef's perspective. She correctly notes that people are hard-wired with cognitive biases, and that we taste things with our eyes as much as with our mouths and noses. She thinks its a cook's duty to tailor food to all the senses, and if this means adding color, so be it.
She briefly covers ways of using color ingredients like beet juice, annatto, turmeric, spirulina, blue majik, blueberry juice, activated charcoal, and cocoa powder. There are also ways to extract pure chlorophyl from herbs that she doesn't cover, but that you can find elsewhere online.
I may experiment with some of these ingredients, and I'll write about it if I do. In the mean time, check out Cree's book. I will definitely be writing a review of this book soon, along with a few others.
Permalink
Permalink
Permalink
Permalink
Permalink
Reading your post seems to make fat the biggest problem with chocolate ice cream. According to Corvitto it is cocoa butter and nonfat cocoaas the negative PAC of chocolate consists of the factor -1.8 for the nonfat solid and -0.9 for the butter. Your cocoa powder has 22g of butter and 88g of nonfat powder and therefore 22*-0.9=-19.8 and -158.4 => -178.2
I guess that your recipe is working anyway and not rock solid at a serving temperatur of -11 °C. So I wonder where the miscalculation in Corvitto's formula comes from. Could you explain?