Chocolate flavour chemistry


What is flavour?

Flavour is essentially caused by the combined effects of aroma, plus tastes (tastes being: bitter, acidic, sweet, salty, and umami).

Scientists have so far identified more than 600 volatile chemicals in cocoa and chocolate [1], which collectively make up the classic aroma of chocolate.

The tastes that occur naturally in cocoa are primarily bitterness, and (especially after fermentation) acidity. Sweetness is added to cocoa, usually in the form of sugar. And sometimes, poor fermentation, drying, or roasting techniques might result in the production of umami-type tastes, such as hamminess.

The flavour of chocolate - while incredibly complex, and practically impossible to synthesise - is not quite as mysterious as most people seem to believe. Below, I've outlined what is known about chocolate flavour development.

Chocolatey and nutty flavours

Classic chocolate and nut flavours in cocoa are produced predominantly by aldehydes and pyrazines, respectively. Aldehydes and pyrazines are volatile chemicals which are produced in cocoa during roasting, as a result of Maillard reactions.

To back-track one step: during fermentation, sucrose in the cocoa bean is broken down by enzymes into simple sugars, namely fructose and glucose. Simultaneously, a cocoa protein is broken down by enzymes into amino acids and peptides. These smaller molecules are known as "aroma precursors".

During roasting, the aroma precursors react with the simple sugars under the influence of heat to create new aroma chemicals, including a broad array of aldehydes and pyrazines.

The aroma precursors and aromas in Criollo and Forastero beans are essentially identical. However, because Forastero beans are normally fermented for longer than Criollos, they typically contain more aroma precursors prior to roasting, and more aroma chemicals after roasting [2]. Which brings me to my next point ...

Mild flavour (as traditionally associated with Criollo cocoa)

Many experts (such as Beckett; Cook; Wood and Lass) have observed that Criollo cocoa beans are characterised by a "mild" flavour. Wood and Lass actually call it a "mild or weak chocolate flavour" [3]. Presilla calls it a "neutral" flavour [4].

Criollo beans are traditionally fermented for a much shorter time than Forastero beans. As a result of short fermentation, Criollo beans are typically less acidic than Forasteros (see notes below about acidity), and they end up with fewer aroma precursors, and hence, fewer aromas after roasting.

Less acidity + fewer aroma chemicals = "mild", "weak", "neutral" flavour.

Low Acidity

As one of the last steps in chocolate manufacturing, conching allows acetic acid to evaporate out of chocolate. But there are other important causes of low acidity that can prevent a chocolate manufacturer from having to conch acidity out in the first place, such as:

Gentle drying of beans after fermentation.

If the cocoa beans are allowed to dry gently, either in the sun, or with skillfully-managed forced-drying, then acetic acid will be able to evaporate and escape through the shell during drying. (On the other hand, rapid forced-drying is widely accepted to be the most significant cause of excessive acidity, which I mention below).

Storage of harvested pods for up to 2 weeks before fermentation.

This practice (which I don't personally advocate) causes a decrease in the volume, as well as the water- and sugar-content of the pulp. The drier, less sugary pulp reduces the microbial activity, meaning that less acid is produced, and the resulting beans are therefore less acidic. The same result has been achieved by spreading beans out to partially dry prior to fermentation.

Excessive Acidity

During fermentation, the cocoa cotyledons become more acidic (i.e. their pH drops) as they are penetrated by acid (primarily acetic acid).

Acetic acid is produced by bacteria belonging to the Acetobacteraceae family. These bacteria dominate the second (aerobic) phase of fermentation. Acetic acid production peaks about 72 hours after fermentation begins [5], and Biehl has noted that diffusion of acetic acid through the whole bean is slow [6].

However, it's quite common for Criollo cocoa beans to be fermented for less than 72 hours - sometimes for only 48 hours. Hence, a short fermentation time typically produces less acidic beans than a longer fermentation. (A short fermentation also has major implications relating to taste and flavour development, as mentioned under the "mild flavour" heading above).

The pH of freshly-harvested beans is typically about 6.3. By the end of fermentation and drying, the beans are more acidic, with a lower pH of about 4.5 - 5.8.

Excessive acidity in dried cocoa beans is most often caused by force-drying the beans too quickly, at too high a temperature. This causes the shells to dry and harden very quickly, thereby becoming impermeable to the acetic acid, which is then trapped inside the bean [7]. (In practice, water is still able to permeate the dried shell and escape ... I assume this is because water has a higher vapor pressure than acetic acid).

Astringency

Polyphenols (such as tannins) are a major source of astringency in cocoa.

During fermentation, polyphenols are converted (with the help of enzymes such as polyphenol oxidase) into different chemicals within the bean, which are much less astringent. Some of these reactions are essentially the same as the aging process in red wine, whereby harsh-tasting polyphenols polymerise into different chemicals which feel more pleasant and "softer" on the tongue.

The fact that dried (unroasted) cocoa beans are brown is proof of the presence of tannins in the beans: the browning that takes place during fermentation is caused by the enzymatic oxidation of tannin-based molecules. (Not to be confused with the non-enzymatic browning that takes place during roasting, courtesy of Maillard reactions).

Bitterness

Polyphenols are not only astringent (as mentioned above), they're also bitter.

Some bitterness can also be attributed to the alkaloids found in cocoa, namely theobromine and caffeine.

But it's also worth noting that cocoa is naturally bitter. Contrary to what some people believe, it's really perfectly normal for cocoa to be somewhat bitter.

For instance, some people say that Criollo cocoa doesn't have any bitterness at all, which is simply not true. For example, see the study by Counet et al (2004): "Relationship between procyanidin and flavor contents of cocoa liquors from different origins". Counet's study is available in PDF format here: http://www.uclouvain.be/cps/ucl/doc/inbr/documents/relation(1).pdf

Also, if Criollo cocoa had no bitterness, then you wouldn't find premium Criollo chocolate makers adding 30% sugar to their products. Thirty percent is really a lot of sugar - enough to make any product sickly sweet if, indeed, the base ingredient was "naturally sweet" to begin with (which Criollo beans are not).

Caramel flavor

Roasting can produce caramel flavours - although not technically via caramelization, but rather via Maillard reactions.

Also, conching has been found to substantially increase the levels of some aroma chemicals, including furaneol, in chocolate. Furaneol has a caramel aroma, and interestingly, it is also known as "strawberry furanone", because it occurs naturally in strawberries, and is an important contributor to the classic strawberry aroma.

Some premium chocolates are said to have a "strawberries and cream" character. Lengthy conching is likely to be a major contributor to this particular characteristic.

Creaminess

A "creamy" character can be produced by long conching (see my comments above regarding the "strawberries and cream" hypothesis).

Floral notes

Roasting can play a role in producing floral notes (for example, the aroma compound phenylacetaldehyde is formed during roasting, and it has a floral/honey aroma). But there are also undoubtedly other factors at work in producing floral aromas in cocoa. Precisely what those factors are remains something of a mystery.

For example, one of the most common floral-smelling chemicals found in cocoa is linalool, a terpene which occurs naturally in many flowers and fruits. Exactly where the linalool in fermented cocoa beans comes from is uncertain, but it's definitely there after fermentation, and prior to roasting [8].

Linalool can be naturally synthesised by plants, and it can also be synthesised by some yeasts, including Saccharomyces and Kluyveromyces, both of which are associated with cocoa fermentation [9], [12]. This means that the fermentation process itself is a possible source of linalool in fermented cocoa beans.

Numerous different factors can affect the synthesis of linalool, as well as other fragrant, floral-smelling chemicals. These factors include:

In 1963, in his book "Chocolate Production and Use", L. Russell Cook wrote that the distinctly floral aroma of Ecuador's Arriba cocoa beans had been "deteriorating slowly for the past several decades" (p.63). Cook also wrote that when Arriba trees were planted in countries other than Ecuador, the trees produced "not a trace of the floral aroma for which Arriba cocoa is sought".

If Cook was correct in his observations, then it seems clear that Arriba's floral aroma is strongly influenced by environmental factors. If terroir is partially or wholly responsible for Arriba's distinct floral aroma, then one or more of the factors listed above might be responsible.

An alternate theory is that the Arriba flavour has disappeared because the Arriba trees have been replaced with different varieties, most notably CCN51. But Cook described the gradual diminution of the Arriba flavour before CCN51 even existed. (The first CCN51 clone was produced in 1965, two years after Cook's book was published).

To digress briefly on the subject of soil nutrients: many chocolate lovers are probably unaware that about 25kg of nitrogen per hectare of cocoa is removed from plantations in the form of dried cocoa beans every year. That nitrogen (and other nutrients) must be replaced somehow, or the plantation will inevitably yield less and less over time. And it is possible that a net-export of nutrients might have contributed to the gradual loss of floral aromas in Ecuador's cocoa.

Fruity flavors

Fruitiness in cocoa is strongly correlated with acidity, and it is well known that fermentation increases the acidity (and therefore lowers the pH) of cocoa beans (for reasons mentioned above).

Also, according to Wood and Lass, cocoa beans naturally contain small amounts of citric acid (0.5% after fermentation and drying), which may be enough to provide a citrus zing.

If you have any doubt about the relationship between acidity and fruitiness, try tasting some Dutched (i.e. alkalized) cocoa powder, which is notoriously lacking in fruity characteristics.

As Alice Medrich writes in her book, BitterSweet:

"The best quality natural [i.e. nonalkalized] cocoa powder is complex and fruity and wonderfully flavorful. [...]In reducing acidity, Dutching strips some of the natural fruit flavors, high notes, and complexity from cocoa." (pp.48,49)

Fruity (and floral) aromas can also come in the form of higher alcohols, produced by yeasts during fermentation.

Then there's the effect (mentioned above) of conching, which can substantially increase the amount of some fruity aroma chemicals, such as strawberry furanone, in chocolate.

Clearly, the contributors to fruity flavours in cocoa are many and varied!

Hamminess

Hammy flavours in cocoa are sometimes associated with smokiness. Ham is a cured meat, which may - or may not - be smoked. Smoke contamination does undoubtedly cause a smokey flavour in cocoa, but I'm not convinced that "smokiness" and "hamminess" necessarily go hand in hand, or that hamminess is caused by smoke.

Over-fermentation is known to cause "hammy" flavours ... but the important question is, why?

Conflicting theories exist on this subject, so I'm not certain about the precise cause (or causes) of "hamminess", but I do have a theory ...

Endoprotease (a type of enzyme that breaks down the globulin storage protein during cocoa fermentation) is highly stable, meaning that it is not de-activated during fermentation. Hence, in beans left to ferment for too long, endoprotease (and another enzyme, carboxypeptidase) will eventually get around to breaking down the albumin storage protein, too.

When cocoa albumin is broken down into its constituent parts by these enzymes, it predominantly releases the following amino acids:

Those amino acids all have an umami (i.e. savoury) taste, and are associated with foods like seaweed, asparagus, eggs, and meats.

Hence, my theory is that the "hamminess" associated with over-fermentation comes from the breakdown of the cocoa albumin into savoury-tasting amino acids.

I also believe that deamidation of these amino acids plays a role, producing ammonia, and yet more meaty flavours, but my chemistry is on shaky ground here!

Heavy metal contamination

Cadmium in cocoa is likely to come from the soil that the tree was grown in - but what about lead?

Lead contamination is a serious problem in chocolate products, but above-ground contamination from leaded fuels and machinery seem to be the major culprits.

Here's a relevant study that readers might find useful and interesting: Lead Contamination in Cocoa and Cocoa Products On this subject, does anyone know for sure what the source of Dagoba's lead contamination turned out to be? I've read in several places that the problem was traced to some lead solder used on one of their grinding machines.

Smokey flavour

Using wood fires to dry beans can cause a smokey flavour in the beans - but only if the dryer is of a poor design (for example, it has no chimney), or is poorly maintained (for example, the heat pipes can develop holes over time). In other words, not all wood fire dried beans end up tasting smokey.

Smoke contamination often happens because growers can't afford to maintain or repair their dryers (not because they're lazy or careless).

Sweetness

I disagree with the popular notion that certain cocoa varieties are "naturally sweet", although the issue here might come down to what an individual's definition of sweetness is. For example, people quite often describe the aroma of our 100%, unsweetened chocolate as "sweet" (and then they get a surprise when they start munching on a mouthful of it!)

There's nothing that I'm aware of in fermented cocoa that would trigger the sweetness receptors in human taste buds. For example, cocoa liquor contains 0% sugars.

But there seems to be a surprising amount of confusion surrounding the innocent little word "sweet".

I define sweetness as a taste, which is detected by the taste buds. Hence, by my definition, sweetness (as caused by sucrose, for example) has no aroma, and is not inherently linked to aroma at all. But many people expand the concept of sweetness to include the aroma of "sweet-smelling" things like soap, perfume, flowers, and unsweetened chocolate - all of which may taste anything but sweet!

There's always a more accurate and illuminating term for any given aroma than "sweet" - such as floral, honey, caramel, or chocolate. But sometimes it can be really hard to figure out precisely what it is you're smelling, and "sweet" seems to be a natural fall-back descriptor.

Finally, as I mentioned earlier, if cocoa was sweet, there'd be no need to add sugar to it to make it palatable. Some manufacturers do sell chocolate with little or no sweetener, but the majority of people - even dedicated chocolate connoisseurs - find such chocolate inedible. And even Amedei's Porcelana bar was 30% sugar, last time I checked.

References:

1 Counet C.; Callemien, D.; Ouwerx C.; Collin S. (2002)
Use of Gas Chromatography−Olfactometry To Identify Key Odorant Compounds in Dark Chocolate. Comparison of Samples before and after Conching
Journal of Agricultural and Food Chemistry, Vol. 50, pp.2385-2391
2 Counet C.; Ouwerx C.; Rosoux D.; Collin S. (2004)
Relationship between procyanidin and flavor contents of cocoa liquors from different origins
Journal of Agricultural and Food Chemistry, Vol. 52, pp.6243-9
3 Wood and Lass (1985)
Cocoa, 4th ed.
Longman, London.
4 Maricel E. Presilla (2001)
The New Taste of Chocolate
Ten Speed Press, California.
5 R.F. Schwan (1998)
Cocoa Fermentations Conducted with a Defined Microbial Cocktail Inoculum
Applied and Environmental Microbiology, Vol.64 Issue 4, pp.1477-1483
6 B. Biehl, D. Passern, W. Sagemann (1981)
Effect of acetic acid on subcellular structures of cocoa bean cotyledons
Journal of the Science of Food and Agriculture, Vol. 33 Issue 11, pp.1101-1109
7 Haslam, E. (1998)
Practical Polyphenolics: From Structure to Molecular Recognition and and Physiological Action
Cambridge University Press
8 Ziegleder, G. (1990)
Linalool contents as characteristic of some flavor grade cocoas
Lebensmitteluntersuchung und -Forschung, Vol. 191, No. 4, pp.306-309
9 Carrau; Medina; Boido; Farina; Gaggero; Dellacassa; Versini; Henschke (2004)
De novo synthesis of monoterpenes by Saccharomyces cerevisiae wine yeasts
FEMS Microbiology Letters, Vol. 243 Issue 1, pp. 107-115
10 Lawler, I.R.; Foley, W.J.; Woodrow, I.E.; Cork, S.J. (1997)
The effects of elevated CO2 atmospheres on the nutritional quality of Eucalyptus foliage and its interaction with soil nutrient and light availability
Oecologia, Vol. 109, pp.59-68
11 Vilanova, M.; Massneuf-Pomarede, I. (2005)
Characterization of yeast strains from rías baixas (NW Spain) and their contribution to the fermentation of albarino wine
Annals of microbiology, vol. 55, no. 1, pp.23-26
12 Leal, G.A; Gomes, L.H; Efraim, P; de Almeida T; Flavio C; Figueira, A. (2008)
Fermentation of cacao (Theobroma cacao L.) seeds with a hybrid Kluyveromyces marxianus strain improved product quality attributes
FEMS Yeast Research, Vol. 8, No. 5, August 2008 , pp.788-798