A Review of: “The chemistry of beer aging – a critical review”

Every once in awhile I plan to review  a scientific journal article which I’ve found particularly interesting or noteworthy.  I’ve got quite a few to choose from, so there’s no lack of material.

One of the most informative articles I’ve seen is from Food Chemistry 95, 2006, titled “The chemistry of beer aging – a critical review”, by Vanderhaegen, et al.  This beast of a paper is 24 pages long, with 5 pages of references (roughly 220 of them).  It covers a vast amount of the research that has been put towards the processes behind beer aging, mostly focusing on the evolution of staling and oxidation flavors.   Here’s the abstract:

Currently, the main quality problem of beer is the change of its chemical composition during storage, which alters the sensory properties.  A variety of flavours may arise, depending on the beer type and the storage conditions.  In contrast to some wines, beer aging is usually considered negative for flavour quality.  The main focus of research on beer aging has been the study of the cardboard-flavoured component (E)-2-nonenal and its formation by lipid oxidation.  Other stale flavours are less described, but may be at least as important for the overall sensory impression of aged beer.  Their origin has been increasingly investigated in recent years.  This review summarizes current knowledge about the chemical origin of various aging flavours, and the reaction mechanisms responsible for their formation.  Furthermore, the relationship between the production process and beer flavour stability is described.

This article kicks off with an interesting (and at times frustrating) observation of beer drinkers: “de gustibus et coloribus non est distputandum” – or “there’s no accounting for taste”.  I’ve seen this more times than I can count, where oxidized flavors are either not noticed or, even worse, expected to be there as part of the intended flavor of the beer.  I suppose you’d have to admit that, by now, the lightstruck flavor of Corona and Rolling Rock [ed: apparently Rolling Rock is no longer skunky due to use of reduced hop extracts] would now be considered a ubiquitous and expected flavor of those beers.  Do people who drink Corona out of a bottle like the flavor of Corona from a can?  I certainly like the canned version better.  The Paper, henceforth to be referred as such, even points to a study by the always entertaining Charlie Bamforth which demonstrated that aging flavors are not always regarded as off-flavors.  What dictated acceptance of the beer was whether they recognized the flavor of the brand, only so much as it remains consistant.

In the introduction section we are given an overall view of how various aspects of beer flavor change as it ages, as described by a few early studies.

  • Bitterness decreases linearly, harshness of bitterness and astringency increase
  • Sweet aroma increases linearly
  • Ribes aroma (catty, blackcurrant) increases to a maximum before falling
  • Cardboard flavor increases gradually
  • Increase in caramel, burnt, toffee-like, wine, whiskey, and leathery aromas
  • Decrease in fruity, estery, floral, and fresh flavors

It is noted that these are just general trends and that each beer will respond differently depending on its ingredients and process variables, as well as the manner in which the beer ages; whether it’s been old and cold, or it’s young and hot, these treatments will produce different effects.  Somewhat enlightening is another cited Bamforth study which posits that, with modern manufacturing processes giving us the ability to produce beers with ridiculously low oxygen levels, beer staling must be a partly non-oxidative process.  In fact, since it’s apparent that there are a variety of reactions which occur, then it follows that each of these has a different “energy of activation”, meaning each one proceeds at a different rate depending on the temperature and also that the rates that they proceed at do not increase equally as the temperature rises.  Evidence of this is shown in studies which demonstrate that papery and cardboard notes increase more when beer is stored at 30C rather than 20C, while caramel-character staling flavors arise more when beer is kept at 25C compared to 30C or more.

The rest of the paper is broken into two general sections which cover some of the same topics in different ways, but reviewing the paper in the order it’s presented would result in a disjointed mess. I’ll assemble the information about each topic into sections so it flows better.

Reactive Oxygen Species (ROS):

Oxidation reactions, by definition, involve the loss of electrons from a particular molecule. Whenever one molecule is oxidized (losing electrons) the other molecule involved is being reduced (gaining electrons). This is the essence of “redox” reactions; the reduction/oxidation of two reacting molecules. The main culprits in beer oxidation are reactive oxygen species (ROS). These are a small variety of oxygen-based molecules which are in various reduction states and are highly reactive oxidants. Molecular oxygen (in its ground state) is actually quite stable and doesn’t react in very many circumstances. However, introduce ferrous iron (Fe^2+) or copper ions (Cu^+) and oxygen easily captures an electron from them which creates the superoxide anion (O2^-). This quickly acquires a proton (hydrogen ion) and becomes a more reactive oxygen species, the perhydroxyl radical (*OOH). Since the pKa if this reaction is 4.8, most of the ROS’ are in this more reactive form when at typical beer pH. Additionally, the original superoxide anion can also be reduced by Fe^2+ or Cu^+ to the peroxide anion (O2^2-), which is readily protonated to hydrogen peroxide (H2O2), which can then create hydroxyl radicals (OH*). All of these species have have high reactivity, which increases with reduction status (superoxide anion < perhydroxyl radical < hydroxyl radical). In fact, these hydroxyl radicals are so reactive, that it’s been suggested that they indiscriminately react with ethanol because it’s a good radical scavenger and is the second-most common compound in beer. This has been supported by studies which have found the resulting 1-hydroxyethyl radicals as being the most common radical found in beer.

These ROS’ are more predominant as metal ion, oxygen, and temperature levels increase. It seems clear that the kick-start of this process is dependent on metal ions as initiators of the ROS production, and there have been numerous ways in which breweries have attempted to limit its presence in beer. It then seems plausible that, if levels of metal ions and oxygen are carefully controlled and the beer is kept cold in the supply chain, a decent shelf-life could be achieved, possibly approaching the 180 day Holy Grail of shelf life.


Carbonyls attracted the most attention early in beer staling research, with acetaldehyde and particularly (E)-2-nonenal (“trans-2-nonenal”) becoming early poster-children for oxidation. It notes, however, that the increase of (E)-2-nonenal during storage is not a given: beers aged at 40C showed above-threshold levels for (E)-2-nonenal after a few days, while beer aged at 20C for four months showed no (E)-2-nonenal. So it appears that (E)-2-nonenal is not a symbol of old beer, so much as it is a symbol of abused beer. Also of note is the study which shows that (E)-2-nonenal formation occurs independent of oxygen concentration, although it is clear that there are somewhat conflicting results in the literature. A number of reactions have been studied in model systems which produce carbonyls, but it nevertheless remains unknown the extent to which the various reactions occur in vivo. Oxidation of fusel alcohols (alcohols of higher molecular weight than ethanol) occurs leading to detected levels of the corresponding aldehydes. Strecker degredation of amino acids has been suggested as a source of carbonyl compounds, but these reactions seem to only be significant at amino acid levels far beyond those found in beer. Similarly, aldol condensation is thought to possibly play a role in carbonyl formation, but it’s unclear if reaction products reach threshold levels. Degradation of the bitter iso-alpha acids also can produce carbonyl compounds which have been shown to be precursors in the formation of staling esters which will be discussed below.

But since the beginning, enzymatic and non-enzymatic oxidation of unsaturated fatty acids have received more attention than any other reactions. After (E)-2-nonenal was discovered as the source of cardboard flavors, it was assumed that lipid oxidation was the source partly because of the amount of research going on in the food industry relating to lipid oxidation and rancidity. In beer, the only fatty acids of note are linoleic and linolenic acid, which are barley-originated and released by lipases during mashing. It’s been found, however, that lipid oxidation does not occur in bottled beer, so it’s generally agreed that a “(E)-2-nonenal potential” is created by lipase action during malting and mashing which allows (E)-2-nonenal to be produced later in various ways downstream. Higher initial mashing temperatures as well as lowered mash pH (about 5.0) have the effect of limiting oxidation of fatty acids during mashing by slowing the action of lipoxygenase enzymes, but these solutions may not be applicable in many brewing situations.

A number of other carbonyl compounds are briefly mentioned as developing to near- or above-threshold levels during staling, including: linear aldehydes like butanal, pentanal, and hexanal (sweaty, grassy); benzaldehyde (almond, maraschino cherry); phenylacetaldehyde (sweet, honey); and methional (cooked potato). Some of these aldehydes are Strecker aldehydes, products of Maillard browning reactions between amines and reducing sugars. Viscinal diketones (diacetyl and 2,3-pentanedione – buttery, butterscotch) have also been seen to increase during beer aging, a phenomenon that we have demonstrated in the lab, which I could probably show in the future. Beta-damascenone is another carbonyl compound which has been shown to increase during aging. It’s production is likely caused by non-enzymatic breakdown of glycosides (aromatic compounds attached to sugar molecules). When these glycosides are hydrolyzed, the two portions of the molecule break apart which frees the aromatic groups and allows it to become part of the flavor of the beer. It is believe that glycosides play an important part, not only in the flavor of aged beer, but also in the fruity and floral character of hop aromas.


Changes in ester levels is also discussed. Overall, these changes lead to a loss in the freshness character of beer. Isoamyl acetate (banana) is considered a marquee ester and it, along with many others, undergo degradation during beer aging. In contrast, certain fruity/sweet esters are actually created during aging (such as ethyl formate, ethyl cinnamate, diethyl succinate) and some of these are associated with winy and whiskey flavors (like ethyl 3-methyl butyrate). Some of the compounds associated with these reactions are produced by degradation reactions of iso-alpha acids mentioned previously. Enzymatic activity (in this case, esterases) can also be left over from dying yeast cells, which can further lead to changes in the ester profiles of beer. Increases in gamma-nonalactone (peach, fruity) and other cyclic esters have been associated with staling of beer as well.

Sulfur compounds:

As for sulfur compounds, dimethyl trisulfide (DMTS, fresh onion) has been known to increase during storage, as can the ribes (catty) character. DMTS is produced by a reaction between hydrogen sulfide and methanesulfenic acid which is derived from hop constituents, and its production is enhanced at low pH.

Heterocyclic compounds:

Many of these types of compounds result from the Maillard reaction, the reaction causing the browning of your bread when it’s toasted. This highly complicated set of reactions are initiated by the reaction between the reducing end of a sugar and the amine group from proteins. The a variety of compounds result from these reactions (even when only one sugar and one protein type are present to begin with), but even the most common ones in beer (furfural and 5-hydroxymethyl furfural) remain below their flavor threshold. They do have the potential, however, to undergo further reactions which can produce high-molecular weight brown pigments known as melanoidins.

Non-volatile changes:

Non-volatile (non-aromatic) changes that occur during aging, and the main reactions involve development of astringency and loss of bitterness. As oxidation proceeds, polyphenols (being potent antioxidants) are among the first to oxidize, and once doing so randomly polymerize into complex structures which elicit an astringent mouthfeel and perhaps a harsh bitterness as well. Polyphenols, once oxidized and polymerized, also have the ability to interact with proteins in the beer and cause turbidity and haze complexes. Additionally, iso-alpha acids can break down in a few different ways, leading to a potent skunky aroma (we’ll discuss that later) or possibly a harsh bitterness, but certainly an overall loss of bitterness. Trans-isomers are reportedly more susceptible to degradation than the cis-isomers, and monitoring their ratio has been proposed as a tool for measuring beer aging.

Conclusion and Discussion:

I’ll conclude this review with a good summary quote from the article and some processes that have been proposed to deal with shelf life.

From the previous considerations, it becomes clear that oxygen triggers the release of free radicals, which can easily react with many beer constituents, leading to rapid changes in the flavour profile. Among these processes are the oxidation of alcohols, hop bitter compounds, and polyphenols. … Modern filling equipment can achieve total oxygen levels in the bottle of less than 0.1mg/L.   At such low oxygen levels, it is debatable whether the formation of ROS is the determining factor in the aging of these beers. Indeed, other molecules present in beer have enough reactivity to interact and form staling compounds. Beer staling is often regarded as only result of oxidation, but non-oxidative processes may be just as important, especially at the low oxygen levels reached in modern breweries.
Non-oxidative reactions causing flavor detorioration are esterifications, etherifications, Maillard reactions, glycoside and ester hydrolysis. Even (E)-2-nonenal, a compound long suspected to be the main cause of oxidized flavour, paradoxically appears to arise by non-oxidative mechanisms in beer. This explains why staling is possible in the absence of oxygen. On the other hand, although some compounds result from oxidation reactions, it is at present not really clear with compound(s) is/are responsible for the oxidation off-flavour of beer.

Some steps that this paper recommends for maximizing the freshness of beer:

  • Minimize formation and activity of ROS, by limiting molecular oxygen pickup and copper and iron levels.  Chelation of metal ions and use of anti-oxidants are mentioned.
  • Sulfites are mentioned to be among the most potent anti-oxidants used in brewing.  Some is formed by yeast, but most of it must be added by the brewer (must be labeled in US), if so desired.

If you can find a copy of this paper in a library somewhere, and can comprehend the oodles of chemistry contained within, this paper is highly informative and a great place to begin any research which might focus on the changes involved in beer aging, particularly from the standpoint of changes in flavor.


Vanderhaegen, B., Nevin, H., Verachtert, H., Derdelinckx, G., “The chemistry of beer aging – a critical review”, Food Chemistry 95 (2006), 357-381

About these ads

7 responses to “A Review of: “The chemistry of beer aging – a critical review”

  1. The conclusion of this one is quite interesting, in stating that though the general knowledge has always associated certain flavors, including cardboard, with oxidation, it appears those flavors develop non-oxidatively.

    Now, this paper is aimed at the commercial beer industry, and thus gives the conclusion that they need to continue to lower oxygen levels in bottled product, but perhaps it is worth speculating that the same cannot be said for high gravity craft designed for aging? Intentional oxidation has already been proven to be beneficial under certain circumstances in wine/fortified wine industry (madeira, sherry), and we have a whole host of very old bottles of beer that have aged in ways I haven’t yet experienced with more modern beers (admittedly, they aren’t old enough yet, but there are signs); I do wonder if the almost complete removal of oxygen from these newer bottles is actually harming their long-term aging potential.

  2. Pingback: LIVE: Onsdag 17. november

  3. perceptivebrew

    Do you know where I can find the oxidation beer wheel? I have heard of it, but I can’t track it down.

  4. Pingback: Has anyone oxidized their beer? - Page 3 - Home Brew Forums

  5. Nice review of the paper. Being an electrical engineer and not having much chemistry or biochem background I do find it a bit difficult to follow everything. I have been planning to write a “dummies guide” version of this paper for about two weeks now and would really appreciate your input when I’m done and it’s up on my blog. Thanks for a great blog, I find it exceptionally interesting and well written.

  6. This review paper will henceforth officially be known as “my guilty pleasure.” Fantastic summary of flavor chemicals as well.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s