Diacetyl, also referred to as 2,3-butanedione, is a compound that yields the buttery or butterscotch flavor in beer and other food products like wine and cheese. It is one of two vicinal diketones (VDK) produced during fermentation, the other being 2,3-pentanedione, which gives a honey-like flavor.
At low concentrations, diacetyl contributes to the slippery mouthfeel of beer. But the butter taste it imparts is considered an off-flavor in finished beer. Its presence in low or moderate levels, however, is acceptable or even desirable in certain styles of beer (Strong, 2015). These include Bohemian Pilsner and some English ales (Krogerus & Gibson, 2015), whose bitter balance and well-hopped character mask the diacetyl taste. VDKs are more easily noticeable in lighter beers, where the taste is not masked by malt and hop flavors.
Flavor thresholds for diacetyl are relatively quite low, falling within the ranges of 0.1–0.2 ppm in lager and 0.1–0.4 ppm in ales. Pentanedione, on the other hand, has a higher flavor threshold of around 0.9–1.0 ppm (Krogerus & Gibson, 2015). But even with such low flavor thresholds, diacetyl will not be detected by one in five beer drinkers even at higher concentrations (Fix, 1993).
Amateurs may find it hard to tell between the buttery flavor of diacetyl and the caramel or toffee flavors of certain low-alcohol beers or stale beers (de Piro, 2005). A good way to start training your tongue to detect it is by spiking light lagers with drops of imitation butter flavor (Johnson, 2015). A flavor standard kit from Aroxa or FlavorActiv, while pricier, also offers an excellent way for a group of enthusiasts to self-train on the most common off-flavors and general flavor descriptors.
For obvious reasons, controlling the amount of diacetyl in beer has been a focus of intensive scientific research. These studies have typically focused on two things: Minimizing diacetyl production and/or maximizing diacetyl reduction. But first, it’s important to understand an important process involved in diacetyl formation — valine uptake.
Brewer’s yeast (Saccharomyces cerevisiae) and lager yeast (Saccharomyces pastorianus) produce diacetyl via a pathway that is involved in the the biosynthesis of valine. Valine is one of many amino acids – the basic building blocks of proteins – found in yeast. Diacetyl is a by-product of the valine pathway.
Two Ways Yeast Obtains Valine
Valine is an essential amino acid in humans, meaning it cannot be synthesized by the body and hence needs to be obtained from food. In yeast however, valine can be both produced intracellularly and absorbed from its environment, which is wort in the case of brewer’s yeast. You can think of these two ways of obtaining valine as complementary processes — cells synthesize valine when the wort doesn’t have enough of it, and stop making it when the wort provides the right concentration of amino acids.
Because of its direct involvement in diacetyl formation, valine uptake plays an important role in brewing. It is interesting to note that in Saccharomyces, amino acids – or more broadly, free amino nitrogen (FAN) – are rarely used to directly build proteins upon being taken up into the cell. They are instead broken down or catabolized into smaller components, which are in turn the ones used by the cells for making macromolecules (Krogerus & Gibson, 2013). It’s akin to a carpenter buying ready-made chairs, disassembling them, and creating new furniture out of the parts.
Valine is a Slow Amino Acid
A minimum amount of free amino nitrogen (FAN) in the wort is needed to have healthy yeast cells. The recommended concentration of FAN is at least 100 ppm to maintain the yeast cells and ensure good rates of fermentation (Krogerus & Gibson, 2013). When FAN concentration is below optimum, fermentation slows down or does not reach completion at all. Too much FAN can also lead to a different set of problems, including beer haze and production of higher alcohols like iso-butanol.
Amino acids are too large and negatively charged to just freely penetrate cell membranes. Unlike carbon dioxide or water, which are small enough to simply diffuse in and out of membranes, amino acids are taken up in a controlled process. The uptake rates in this controlled process vary among the different types of amino acids, which is why they are classified into four groups: A, B, C, and D, corresponding to fast, moderate, slow, and no absorption, respectively. As it turns out, valine is classified under group B for S. cerevisiae and under group C for S. pastorianus. Yeast cells take a long time to absorb valine, and the lag phase can be as long as 12 hours. As will be discussed in later, this relatively slow uptake rate of valine has been a problem that researchers have been trying to solve through various methods.
Membrane Proteins Control Valine Uptake
How does yeast control passage of amino acids? It is done with the use of protein channels that act as gates. These membrane proteins use cations (positively charged particles) as sort of a currency to allow proteins inside the cells. Several proteins have so far been discovered that transport valine across the yeast membrane. These proteins are not dedicated to valine transport alone, and can also move other amino acids. The differences in uptake rates among and within the four groups of amino acids are influenced by the affinity of the amino acid to the protein channel. Fast amino acids are better at competing for these entrance slots. Slow amino acids like valine have weaker affinities.
Studies on the expression of the genes that code for these membrane proteins have shed light on the dynamics of valine uptake, and consequently on the conditions that affect diacetyl formation. For example, it was discovered that some of these genes are transcribed (i.e., DNA is copied into mRNA, the first step in gene expression) when pH of certain culture media is lower and when yeast is pre-oxygenated. This suggests that if you want to increase the uptake rate of valine during fermentation, you need to drop wort pH and pre-oxygenate your yeast culture (Krogerus & Gibson, 2013).
Factors Affecting the Formation of Diacetyl
Contamination by Bacteria
Historically, the occurrence of diacetyl in a brew has been mostly associated with unsanitary brewing practices. Certain bacterial contaminants, such as species of Lactobacillus and Pediococcus, can produce diacetyl as a by-product of fermentation. Among the various stages in beer production, yeast pitching is the process that is the greatest source of bacterial contamination, since it occurs after boiling. Modern day aseptic techniques, however, have largely removed the bacterial problem. It is the consequent formation of diacetyl during valine synthesis that present day brewers are aiming to control.
Valine and FAN in Wort
As previously mentioned, the presence of valine in sufficient quantities in the wort stops valine synthesis in the cells. We now understand that this happens due to the inhibitory effect of valine on the enzyme that catalyzes the formation of α-acetolactate from pyruvate. Pyruvate or pyruvic acid is the product of breaking down glucose, and is a key intermediate in both fermentation and cellular respiration. α-acetolactate is the immediate precursor of diacetyl. The enzyme that converts pyruvate to α-acetolactate is called acetohydroxyacid synthase (AHAS). Inhibiting this enzyme is key to decreasing diacetyl production (Krogerus & Gibson, 2013).
It’s now easy to see why increasing wort valine concentrations decreases formation of diacetyl. Direct supplementation of wort with valine (100-300 ppm) results in greater valine uptake, lessening the need for intracellular production of valine, and thus decreasing the precursor for diacetyl formation (Krogerus & Gibson, 2013).
In terms of the effect of general FAN content of the wort on diacetyl formation, a different story emerges. Many studies have demonstrated a decrease in diacetyl formation with decrease in overall FAN levels. They speculate that this is due to the quick depletion of the preferred amino acids (Group A), leading to the faster and greater uptake of the non-preferred amino acids like valine (Krogerus & Gibson, 2013).
These outcomes can be explained by the differences in valine uptake. At very high FAN levels, yeast cells absorb valine slower because they prioritize uptake of the preferred amino acids. At very low FAN levels, yeast cells absorb valine faster because the wort runs out of the preferred amino acids rather quickly. For optimally low diacetyl production, FAN concentration of about 150 ppm is recommended, although this may vary with yeast strain and fermentation conditions (Krogerus & Gibson, 2013).
Another factor that influences valine uptake rates is yeast culture source. Rates differ between yeast cells pitched from rehydrated active dry yeast and those from freshly cropped yeast slurry (Krogerus & Gibson, 2013). It appears that the drying process does something to the yeast cell membrane that significantly lowers valine uptake rates. Slower uptake means greater chances for intracellular synthesis of valine, with the undesired diacetyl by-product.
Pre-oxygenating yeast cells, however, results in the expression of certain membrane proteins that transport valine. This is supported by other studies that showed pre-conditioning of harvested yeast cells in a glucose solution or aeration shortens the lag time before amino acid uptake. Increased valine transport and shorter lag phase means less opportunity for diacetyl formation.
Wort pH and Temperature
Lowering the pH of wort to <5 potentially reduces diacetyl formation, due to the expression of genes that code for valine transport proteins in acidic conditions. This is however not true for all strains of Saccharomyces. A certain strain of lager yeast showed a negative effect of an acidic environment on the expression of these transport proteins (Krogerus & Gibson, 2013).
Higher temperatures during fermentation encourages diacetyl production (Fix, 1993). Ale that is fermented at 20 degrees C (68 degrees F) will produce more diacetyl than a lager at 10 degrees C (50 degrees F). But complicating matters is the temperature effect on diacetyl reduction, which is also increased with increasing fermentation temperature. A solution to this is raising the temperature towards the end of fermentation (refer to the Diacetyl Rest discussion below).
Oxidation After Product Packaging
It has to be emphasized that diacetyl can also be formed via a process that can occur even after the yeast has been removed and the final beer product has been packaged. The process is the simple oxidation of leftover acetolactate in the beer. Most of the acetolactate is normally broken down by the yeast by the end of fermentation. However, there are instances when not all the acetolactate is metabolized, and it is non-enzymatically converted to diacetyl when exposed to an oxidizing agent. In this case, the headspace air is the oxidizing agent, and shaking the packaged product facilitates the conversion into diacetyl (Fix, 1993). Exposure to warm temperatures likewise promotes the oxidizing process.
Pitching Rate and Fermentation Time
It has been shown that higher yeast pitching rates raise diacetyl concentration of beer. One explanation for this is the greater acetolactate that accumulates outside the cells coupled with the much shorter fermentation time. Acetolactate would not have enough time to be decarboxylated back to diacetyl, which given enough time can then be reduced by yeast cells to acetoin and finally to 2,3-butanediol. These products have a much higher flavor threshold, and are usually not of great concern among brewers. The acetolactate that accumulates after fermentation will then be available for the chemical oxidative formation of diacetyl in the packaged beer. Other studies demonstrate that even with increased pitching rates, high-gravity lager can still end up with lower diacetyl when the number of days of fermentation was increased (Krogerus & Gibson, 2013).
Detection of Diacetyl
Forced Diacetyl Test
A relatively simple method of detecting the presence of diacetyl in beer is the “forced diacetyl test”. This involves taking two samples of the same beer, heating one sample in a water bath for 10-20 minutes at 60-70*C (~150*F) while keeping the other as an unheated control. The heated beer will smell and taste of butter popcorn or butterscotch if acetolactate is present. This test may be performed ~7 days after brew day as a fundamental quality control check. If no diacetyl is detected in the heated sample, it is safe to crash for clarification. If diacetyl is detected, it may require more time before crashing. If it is particularly bad, krausening (without carbonation) is an option. Actively fermenting wort may be added (~1-2% volume) to provide fresh yeast and process the remaining VDKs.
More sophisticated methods for greater accuracy in detecting and measuring actual diacetyl content of beer include Gas Chromatography (GC), High Performance Liquid Chromatography (HPLC), Spectrophotometry, and Colorimetric Assays (Landaud et al., 1998; Li et al., 2012; Mawer & Marti, 1978). The first two are separation methods that rely on the fact that diacetyl has a different solubility in liquids or affinity to solids than the other beer components. Hence, when forced to travel with a carrier (gas in GC and liquid ni HPLC), diacetyl will do so at a different speed than others, which can then be detected and measured by a computerized sensors. Spectrophotometry relies on diacetyl’s ability to transmit or absorb light differently when in solution, while colorimetry depends on chemicals that undergo measureable color changes in the presence of diacetyl. All of these methods need a considerable investment on equipment and training, and hence are only used in large scale brewing industries.
Reducing Diacetyl Concentrations
VDKs formed during fermentation can be reduced by yeast into diols after fermentation. 2,3-butanediol, the final product in diacetyl reduction has a much higher flavor threshold at 4500 ppm. Various enzymes catalyzing the reduction of diacetyl have been identified in different yeast strains.
What limits the rate of diacetyl removal from beer is not the speed at which these enzymes do their jobs. Yeast is capable of rapidly converting diacetyl to its reduced forms of acetoin and butanediol. It is actually the spontaneous decarboxylation of acetolactate back to diacetyl that limits the rate of diacetyl removal.
Obviously, any factor influencing yeast growth, such as temperature, pH, and oxygenation of yeast, will also affect the reduction rate of diacetyl. Lowering pH, raising temperature, and increasing conditioning time are some common methods to decrease residual diacetyl in beer.
Increasing Valine During Malting and Mashing
The resulting diacetyl content of beer can be controlled by malting and mashing at temperatures, pH, and times that increase the amount of valine in the malt and wort. However, this may increase production time and decrease efficiency, as well as potentially affect yields and other properties of the wort (Krogerus & Gibson, 2013).
A higher temperature may lead to higher diacetyl formation due to greater yeast growth, but greater yeast mass also means there are more of the cells to perform the reduction of diacetyl. In addition, oxidative decarboxylation of acetolactate to diacetyl, which is the rate-limiting step in diacetyl reduction, is also promoted at higher temperatures. Brewers take advantage of these facts in the so-called “diacetyl rest”, which is more often used with lagers due to their slower diacetyl reduction than ales. At the end of primary fermentation, temperature is slowly ramped up to increase decarboxylation of acetolactate and to reinvigorate the yeast cells to metabolize the VDKs (Krogerus & Gibson, 2013).
Germans have an age-old beer-making traditional practice called “krausening”. Its main purpose was originally to carbonate beers without using sugars or adjuncts, which were not in the list of allowable ingredients as stipulated in the German purity law (Smith, 2010). Krausening involves the addition of fresh wort from the most recent batch to fermented beer. The active yeast will then carbonate the beer and at the same time metabolize any remaining VDKs.
Higher valine in the wort inhibits the AHAS enzyme, which means less acetolactate is formed from pyruvate. The disadvantage of this method is the high cost and increased chances of producing higher alcohols (Krogerus & Gibson, 2013).
Addition of Bacterial Enzyme
Another method of reducing diacetyl in beer makes use of acetolactate decarboxylase (ALDC), an enzyme that can be isolated from certain bacterial strains of Bacillus, Enterobacter, Streptococcus and others (Choi et al., 2015). ALDC can directly convert α-acetolactate to acetoin by skipping diacetyl formation altogether. However, ALDC may also incur higher costs than other methods of diacetyl reduction (Krogerus & Gibson, 2013).
Diacetyl is one of the most common off-flavors found in beer. Due to its low flavor threshold and complexity in controlling formation and promoting reduction, diacetyl poses a significant challenge to brewers. Its formation can be minimized with good sanitation practices, by increasing wort valine uptake, lowering fermentation pH, and raising temperature. In addition, slower pitching rates and longer fermentation times also decrease diacetyl formation. Adopting appropriate reduction methods effectively removes diacetyl from beer. These methods include diacetyl rest, krausening, and the use of valine supplement and bacterial enzyme. With proper detection and prevention, diacetyl can be prevented from reaching markets and the consumers.
- Choi, E. J., Ahn, H. W., & Kim, W. J. (2015). Effect of α-acetolactate decarboxylase on diacetyl content of beer. Food Science and Biotechnology, 24(4), 1373-1380.
- De Piro, George. (2005). Beer Flavors #1: Diacetyl. Professor Beer. Retrieved July 11, 2016, from http://www.professorbeer.com/articles/diacetyl.html.
- Fix, George J. (1993). Diacetyl: Formation, Reduction, and Control. Brewing Techniques. Retrieved July 11, 2016, from http://www.morebeer.com/brewingtechniques/library/backissues/issue1.2/fix.html.
- Johnson, Bobby Don. (2015). Diacetyl in Beer. Winning Homebrew. Retrieved July 11, 2016, from http://www.winning-homebrew.com/diacetyl-in-beer.html.
- Krogerus, K., & Gibson, B. R. (2013). 125th anniversary review: diacetyl and its control during brewery fermentation. Journal of the Institute of Brewing, 119(3), 86-97.
- Landaud, S., Lieben, P., & Picque, D. (1998). Quantitative analysis of diacetyl, pentanedione and their precursors during beer fermentation by an accurate GC/MS method. Journal of the Institute of Brewing, 104(2), 93-99.
- Li, P., Zhu, Y., He, S., Fan, J., Hu, Q., & Cao, Y. (2012). Development and validation of a high-performance liquid chromatography method for the determination of diacetyl in beer using 4-nitro-o-phenylenediamine as the derivatization reagent. Journal of agricultural and food chemistry, 60(12), 3013-3019.
- Mawer, J., & Martin, P. (1978). A comparison of colorimetric methods and a gas chromatographic method for the determination of vicinal diketones in beer. Journal of the Institute of Brewing, 84(4), 244-247.
- Smith, Brad. (2010). Krausening Home Brewed Beer.BeerSmith. Retrieved July 11, 2016, from http://beersmith.com/blog/2010/03/22/krausening-home-brewed-beer/.
- Strong, Gordon (ed.). (2015). BJCP 2015 Style Guidelines. Beer Judge Certification Program. Retrieved July 11, 2016, from http://www.bjcp.org/stylecenter.php.