by Scott Bickham (Brewing Techniques)
The relative levels of esters and alcohols are critical to a beer’s balance. Though sometimes difficult for brewers to control, the formation of these compounds can be affected by close monitoring of fermentation conditions.
So far, the Focus on Flavor column has emphasized flavors derived from ingredients, either directly or through changes that occur in the brewing process. Yeast plays a secondary role in the formation of most of these compounds, but it now emerges into the spotlight as we consider the aromas and flavors of esters and alcohols.
Although I described the nutty and grassy flavors arising from fusel alcohols in the last Focus on Flavor, I will revisit them here because their levels are affected by most of the factors that influence ester production. In fact, it is actually the ratio between esters and alcohols, not the absolute levels, that determines whether a beer is balanced. For this reason, I will discuss the relationship between alcohols and esters, but the emphasis will be on understanding and controlling the latter.
The first-tier descriptors for estery flavors are “fruity,” “estery,” or “solventlike.” Fruity flavors have eight secondary descriptors that refer to different fruits. The second-tier descriptors for estery flavors are the names of the three most common esters found in beer. Solventlike flavors are associated with high concentrations of esters, and these are further broken down into “plasticlike,” “can-liner,” and “acetonelike” flavors. Beer tasters often confuse these characteristics with alcoholic flavors, which is the final group of descriptors in this class. Alcoholic flavors refer to ethanol and higher alcohols, some of which may also be characterized by the second-tier descriptors “spicy” or “vinous.”
In this installment, I will identify the most common compounds responsible for fruity, estery, solventlike, and alcoholic flavors. The box “How Alcohols and Esters Interact” below summarizes Meilgaard’s sensory analysis experiments that indicate how combinations of these compounds are often perceived. I will then outline the chemistry of ester and alcohol formation, along with the factors that lead to elevated or depressed levels. The article will conclude with several examples of commercial beers that exemplify the flavors and describe doctoring methods that enable one to become acquainted with estery and alcoholic flavors.
The fruity and estery flavors in the first-tier groups 0140 and 0130 are primarily caused by esters, but be aware that other compounds often contribute similar flavors.*
Citrusy flavors: Citrusy notes are commonly produced by the ester isoamyl octanoate, which also has orange, pear, and spicy notes. Its concentration is usually low compared with its flavor threshold, but it can be a tertiary flavor constituent (at 0.1–0.5 times its flavor threshold or 0.1–0.5 flavor units). Citrusy notes can also originate directly from citric acid, which is a natural byproduct of the metabolism of carbohydrates by yeast. The factors governing the production of citric and other organic acids are not completely known, but wort composition and yeast strain play a significant role. Most brewers are also aware of the citrusy notes that are possessed by some American hop varieties, but although esters are important components of hop oil, the citrusy notes in hops instead are attributed to the alcohols linalool and geraniol and a wide variety of heterocyclic compounds.
Apple flavors: Apple flavors are primarily associated with ethyl hexanoate, which also has a hint of aniseed. This ester is primarily found at 0.5–2.0 flavor units, which makes it a secondary flavor constituent. Other esters associated with apple flavors include ethyl butyrate, ethyl octanoate, ethyl decanoate, and 2-phenyl ethyl acetate. This last compound is an alicyclic ester that is better known for its roselike characteristics, although rose character is more commonly due to the alcohol 2-phenylethanol.
Banana flavors: The next fruity descriptor listed in Table I is “banana,” which is associated with isoamyl acetate. Other esters, in particular active amyl acetate and n-butyl acetate, also contribute bananalike flavors, but they are typically present in beer at levels below their flavor thresholds and are therefore not listed in the “big three” estery compounds. In addition, the alcohol precursors of the amyl esters also have bananalike flavors and are often found at concentrations of 0.5 flavor units or greater. Several aldehydes, notably isobutanal, pentanal, and isopentanal, also contribute banana and fruit flavors to beer. These aldehydes are normally only present in green or oxidized beer, so it is usually safe to assume that banana flavors originate from the amyl esters and alcohols.
Black currant flavors: The “black currant” descriptor is used to label a group of dark fruit flavors produced by esters such as active isoamyl formate (plum) and heptanol butyrate (red currant). Surprisingly, beta-damascenone, a hop oil derivative, also has black currant, raspberry, and menthol flavors. The concentrations of this compound can be as high as 0.2 flavor units, which would make it a tertiary flavor constituent.
Pearlike and melony flavors: Pearlike flavors are produced by the esters n-heptyl acetate and isoamyl octanoate, whereas melony flavors are usually associated with the aldehydes butanal and isobutanal. Another source of melony flavors is 5-hydroxy pentanal, which is an aldehyde with a cantaloupelike aroma.
Berrylike flavors: Berrylike flavors are generally not associated with esters, the exception being ethyl lactate, which has perfumy (some say creamy) notes compared to artificial raspberry flavorings. It has a rather high flavor threshold, so it is typically a background flavor constituent. The hop oil derivatives alpha-ionone and beta-ionone, which are complex ketones, may also contribute low levels of artificial raspberry flavors, along with woody and “violet” notes. Another potential source of fruity and raspberry flavors is the aldehyde 2-pentanal, which is often present at near-threshold levels.
Strawberry flavors: The final fruit flavor listed in this group is strawberry, which may be among the spectrum of fruity tones produced by esters, but is more commonly associated with the di-ketone 2,3-hexanedione.
Other fruity flavors: Although the flavors described above exhaust the fruits listed on the Beer Flavor Wheel, it does not preclude the use of others as descriptors. For example, papaya flavors are produced by ethyl butyrate and ethyl valerate, whereas methyl isovalerate and ethyl methylbutyrate have grapelike flavors. Apricot and peach flavors are associated with gamma-decalactone, which can occur at concentrations as high as 0.5 flavor units. Several of the long-chain esters are compared to tropical fruits such as mango, papaya, and guava. A general trend noted by Meilgaard is that temperate-zone fruit flavors (apple, plum) are often produced by short-chain esters, subtropical fruit flavors (banana, pineapple) by medium-chain esters, and tropical fruit flavors by long-chain esters.
It is often accurate to describe an ester as simply being fruity without identifying a specific fruit. The most common ester, isoamyl acetate, which has typical concentrations ranging from 0.25 to 2.0 times its flavor threshold, has a distinctive fruity, sweet aroma that is not that different from some of the sweetened, fruit-flavored breakfast cereals (or bubble gum). Another source of unspecified estery flavors is oxidized melanoidins such as methylpyridine and 2,6-dimethylpyrazine. These compounds have estery and perfumy notes in addition to papery and cardboard flavors.
Solventlike flavors: The flavors in group 120 of Table I describe solventlike flavors, which are generally perceived as a harsh, burning sensation in both the aroma and taste. These sensations typically arise from high concentrations of esters; “can-liner” is an exception, but it is rarely an issue in small-scale brewing. The esters most often associated with solvent flavors are ethyl acetate, 2-butyl acetate, isoamyl acetate, active amyl acetate, and ethyl decanoate. Another source of solventlike flavors with varnish or acetone characteristics is the ketones. The simplest compounds in this group are acetone and butanone, but typical concentrations of these compounds are well below their flavor and toxicity thresholds. Longer-chain ketones such as 2-heptanone occur at similar levels, but their lower threshold means that they can have a more significant impact on beer flavor. A third source of solventlike flavors is the aldehydes butanal and isobutanal and the hydroxy aldehyde aldol. These compounds also possess fruity notes, but because they normally appear in conjunction with other staling aldehydes, they can be distinguished from esters by noting the presence or absence of oxidation flavors in the beer.
Alcoholic flavors: The last group of descriptors in Class 1 of the Beer Flavor Wheel denotes alcoholic flavors, which often have spicy or vinous characteristics. These flavors should be distinguished from the perception of alcoholic warmth, which is a mouthfeel sensation that will be described in the next Focus on Flavor column. Alcoholic flavors are generally produced by shorter-chain alcohols with fewer than six carbon atoms. These flavors are different from the coconut and walnut flavors of longer-chain alcohols, which I discussed in the November/December issue. Vinous flavors often arise from the amyl alcohols (in addition to their more distinctive banana notes), but another source of these flavors is the aldehydes hexanal and heptanal. These compounds also produce stale notes that allow their vinous flavors to be distinguished from those derived from alcohols. Spicy alcoholic flavors characterized as almond or cinnamonlike are usually associated with aromatic alcohols such as benzyl alcohol. These compounds have higher flavor thresholds than the aliphatic alcohols listed above, so they are usually not significant flavor constituents.
The formation of esters is a complex process that depends on many variables, some of which are not easy to control. The basic mechanism involves the esterification of fatty acids and lipids by ethanol and, to a lesser extent, fusel alcohols. This process is not spontaneous, but requires the presence of catalysts and the alcoholysis of acyl CoA compounds. Acyl CoAs are derived from several different sources, including the activation of wort fatty acids, the oxidative decarboxylation of keto-acids, the catabolism of lipids, and the biosynthesis of fatty acids. The predominant acyl CoA molecule is acetyl CoA, which combines with ethanol and isoamyl alcohol to form ethyl acetate and isoamyl acetate, respectively.
The supply of acyl CoA molecules is closely coupled to the oxygen supply of the yeast, because oxygen affects fat synthesis as well as the transition from respiration to fermentation. For example, decreasing the level of dissolved oxygen from 8 to 3 mg/L leads to a two- to fourfold increase in the levels of ethyl acetate, isoamyl acetate, and ethyl caproate. The rate of ester formation also increases substantially after the synthesis of fatty acids stops. So as long as yeast cells continue to produce lipids and fatty acids during the growth phase, the synthesis of esters is inhibited. Conversely, excess levels of fatty acids present in the trub will inhibit fatty acid synthesis and favor ester production.
The yeast strain and levels of dissolved oxygen and fatty acids in the wort play critical roles in ester formation. Other factors that tend to increase the ester level include wort gravities above 13 °P, high wort attenuation limits (in other words, increasing either the original gravity or the final gravity), fermentor geometry (for example, increasing the fermentor height, which decreases CO2 pressure), and agitation. The effect of temperature is not as clear-cut, with some references citing an increase in estets with a reduced fermentation temperature and others a decrease in ester production. The former analysis is accurate for the primary fermentation; higher temperatures are better during the growth phase because they tend to stimulate yeast growth as well as amino acid synthesis; ester production, however, continues during beer maturation, and can be controlled by reducing the fermentation temperature because higher temperatures encourage the ester-producing reactions to continue.
The method and duration of the fermentation also affect the ester level in the finished beer. In top fermentation, the ester level typically rises linearly during the first 36 hours and then reaches a plateau. In bottom fermentation, the ester level increases linearly during the primary fermentation and then gradually tapers off during the secondary and maturation phases. Ales typically have ester levels as high as 80 mg/L (for Doppelbocks), and lagers contain up to 60 mg/L. Abnormally high levels can result from very high fermentation temperatures or contaminations by wild yeast of, in lambics, enteric bacteria.
In contrast to ester production, which occurs during both the fermentation and maturation phases, most (80%) of the fusel alcohols are formed during the primary fermentation. In most fermentation systems, about half the higher alcohols come from amino acid metabolism, and the other half, from carbohydrate metabolism. These alcohols cannot be removed by normal measures, so it is important to control their levels by adjusting the fermentation conditions.
Higher alcohols are formed by deamination, decarboxylation, and reduction. For example, the amino acid leucine is first converted to alpha-oxoisocaproic acid, then isovaleraldehyde, and finally to isoamyl alcohol. This is known as the Ehrlich pathway. Note that each amino acid in this process can be linked to a specific fusel alcohol, and furthermore, these amino acids do not have to be present in the wort, but can be synthesized by yeast during the metabolism of carbohydrates.
The most prevalent fusel alcohols in beer are propanol, isobutanol, isoamyl alcohol, and the phenol alcohol tyrosol, but the amyl and phenol alcohols are the only ones that are commonly found at concentrations comparable to their thresholds.* As with ester production, the formation of fusel alcohols is to a large extent determined by the yeast strain; wild yeast and respiratory-deficient mutants produce extremely high levels. Other factors that favor fusel alcohol production are high fermentation temperature (above 70 °F [21 °C] for most ale strains and above 55 °F [13 °C] for most lager strains), agitation, low amino acid content in the wort, intensive aeration, high pitching temperatures, and wort gravities above 13 °P. Repeated repitching of yeast can also dramatically increase the amount of fusel alcohols, particularly 2-phenyl ethanol.
Comparing the factors that influence alcohol and ester formation, we see that the relative levels of these substances can be adjusted by increasing the level of dissolved oxygen. Intensive aeration stimulates yeast growth, and this in turn favors amino acid synthesis and alcohol formation over ester production. Of course, there are practical limits on the amount of oxygen that can be dissolved in wort, and the results will also be dependent on the viability and characteristics of the yeast strain. The ratio of alcohols to esters in a so-called balanced beer is approximately 2.5–3 to 1 (1†), so brewers usually select yeast strains and fermentation temperatures that favor ester production. Typical concentrations of fusel alcohols in top- and bottom-fermented beer usually are below their flavor thresholds, but abnormally high levels can result from wild yeast or bacterial contaminations.‡
The ester and alcohol content of beer varies tremendously, even within a given style. For example, the acetate ester level in beers fermented with four different Weizen strains ranged from 25 to 46 ppm in one particular experiment, and it varied from 26 to 45 ppm with eight bottom-fermenting strains. The fusel alcohol level varied from 160 to 202 ppm and 72 to 102 ppm with the Weizen and lager strains, respectively. Beers fermented with Alt and Kölsch strains fell somewhere between these two extremes, as would most other top-fermented styles.
Because the ester and alcohol levels both tend to increase with original gravity, good examples of beers that exhibit both flavors are strong ales such as Anchor Old Foghorn, Hair of the Dog’s Adam, and the Belgian Brasserie Dubuisson’s Scaldis. Esters are also an important part of the profile of strong lagers; the Bavarian lager EKU 28 has 68 ppm of ethyl acetate, for example, which is more than many ales. Ethanol and higher alcohols should be perceptible in all of these beers, but the latter should be restrained. Note that it is also possible to produce estery beers without high alcohol levels. German Weizens, Belgian pale ales, and most English bitters and milds all fall into this category.
Beers can be doctored in several ways to exaggerate these flavors. The American light lager style recommended as a reference beer typically has 4% (v/v) ethanol, or 14 mL per 12-oz bottle. An increase of at least 2% (v/v) ethanol is needed for most people to detect a difference, and this level can be achieved by adding 2 tsp (10 mL) of grain alcohol. Vinous, or winelike, flavors can be produced by adding 30 mL of fresh Chablis wine to a 12-oz sample of the reference beer. For estery flavors, the artificial banana flavoring sold at most supermarkets is a good source of isoamyl alcohol and other esters. It is a potent flavoring agent, so 4–5 drops per 12-oz bottle is enough to produce a detectable banana flavor. Larger additions may be perceived as solventlike, which is another flavor that should be recognizable. George Fix has pointed out that corn spirits have a much higher concentration of esters than artificial banana flavoring has, so this option offers a method of increasing the ester level without favoring isoamyl acetate.
To summarize, in balanced beers, the ratio of fusel alcohols to esters should be less than 3:1, but higher levels often result from poor yeast management, poor trub separation, or improper aeration. These compounds are the most important by-products of fermentation and often distinguish one beer style from another, so it is important to recognize their signatures and learn how they can be controlled. Table II summarizes the most common sources and remedies for these flavors.
The next “Focus on Flavor” will discuss the body and mouthfeel of beer, which is the final stop on our journey around the Beer Flavor Wheel. These characteristics are not tasted, but they produce physical sensations that can dramatically affect how a beer is perceived. The next installment will also summarize the beer doctoring experiments that have been presented in this column.
All contents copyright 2019 by MoreFlavor Inc. All rights reserved. No part of this document or the related files may be reproduced or transmitted in any form, by any means (electronic, photocopying, recording, or otherwise) without the prior written permission of the publisher.