Brettanomyces secondary fermentation experiment

This page documents an ongoing collaborative experiment conducted by Lance Shaner of Omega Yeast Labs and Richard Preiss of Escarpment Laboratories to investigate the effect of Brettanomyces secondary/bottle pitch rate on the sensory characteristics and flavor metabolites of beer during bottle conditioning.

Introduction

The effect of pitch rate on flavor in Brettanomyces bottle/secondary fermentation has been a contentious point of discussion amongst brewers. The vast majority of available information is strictly anecdotal, which highlights the need to analyze a Brettanomyces secondary fermentation with different pitch rates in order to better assess the effect pitch rate may have. It has been suggested previously that stressing Brettanomyces cells by underpitching will result in higher production of funky metabolites, such as phenolic compounds and organic acids. It is known that Brettanomyces is capable of production of volatile esters, phenols and organic acids during secondary fermentation, but whether or not these metabolites are produced in a dose-dependent manner is unknown. Furthermore, the rate at which these reactions occur is also poorly characterized in the context of beer secondary fermentations. In this study, we aim to determine whether there is a dose/pitch rate-dependent effect of Brettanomyces on a secondary beer fermentation, and if so, whether this effect is reduced during prolonged beer aging.

Methods

The base beer consisted of 12 gallons of 1.053 wort (75% Pilsner, 15% Wheat & 10% Munich) which was fermented with saison yeast from the OYL-217 C2C American Farmhouse Blend. Final gravity was 1.008 (2ºP). Bottles and bulk aging samples were then inoculated with the Brettanomyces bruxellensis strain from the same blend. Bottle inoculation rates: 1) zero (control), 2) 50,000 cells/mL, 3) 240,000 cells/mL, 4) 1 million cells/mL, 5) 2.4 million cells/mL. Bottles were then conditioned at room temperature for 3 weeks, after which they were opened for sensory and metabolite analysis.

Bottles of the Control (no Brett added), Low (50k Brett cells/mL), and High (2.4 million cells/mL) were refrigerated for 24 hours prior to tasting. The bottles were coded so that the identities of the samples were concealed. The three participants were given a 2 oz pour of each sample and asked to identify the control, Low, and High. This analysis was performed at 2 and 4 weeks post-bottling.

Samples for flavor metabolite analysis were filtered with a 0.2µM syringe filter and stored in 15mL conical tubes and stored at 4ºC prior to flavor metabolite analysis. Metabolite analysis was performed via HS-SPME-GC-MS using the methods outlined in Rodriguez-Bencomo et al. (2012).

The beers will also be analyzed at a 3 month time point.

Results/Discussion

• 

  Figure 1 Measurement of phenolic compounds at 3 weeks.

• 

  Figure 2 Measurement of ethyl esters at 3 weeks.

• 

  Figure 3 Measurement of acetic acid and acetate esters at 3 weeks.

• 

  Figure 4 Measurement of phenethyl alcohol and phenethyl acetate at 3 weeks.

• 

  Figure 5 Measurement of volatile organic acids at 3 weeks.

Sensory Analysis

At the tasting two weeks after bottling, all three participants correctly identified the Control. The participants noted that the Control had more clove aroma and fuller mouthfeel compared to the two Brett samples. The Brett samples were described as fruity and slightly acidic. No taster was able to discern a significant difference between the Low and High samples.

At the tasting four weeks after bottling, all three participants again correctly identified the Control due to the clove aroma and fuller mouthfeel. The Brett samples were described as fruity but with a stronger medicinal flavor than at two weeks. There was still no discernible difference between the Low and High Brett samples.

Metabolite Analysis

At 3 weeks, conversion of 4-vinylguiacol (4-VG) to 4-ethylguiacol (4-EG) was observed in all samples dosed with Brettanomyces (Figure 1). This indicates that regardless of pitch rate, conversion of above-threshold amounts of 4-VG (clove) produced by Saccharomyces was metabolized to above-threshold levels of 4-EG (spicy, smoky). Additionally, de novo production of 4-ethylphenol (4-EP, barnyard/medicinal) was observed at around sensory threshold levels regardless of Brettanomyces pitch rate when compared to the Saccharomyces control. Taken together, these results indicate that volatile phenol production does not correlate to pitch rate in the range tested, as early as 3 weeks into secondary fermentation.

In general, we observed slightly elevated levels of ethyl esters with the addition of Brettanomyces (Figure 2). However, there was a high degree of variability among the different pitch rates. Identification and confirmation of trends with regard to ethyl ester production in this experiment would require multiple replicates in order to more accurately assess the data and identify trends. The most dramatic increases in ester production were seen with ethyl nonanoate (tropical fruit) and ethyl caprylate (pineapple) production.

Interestingly, we observed a potential dose-dependent decrease in isoamyl acetate (banana) concentration at 3 weeks (Figure 3). Therefore, higher pitch rates of Brettanomyces in secondary fermentation may metabolize isoamyl acetate faster, reducing the concentration of this ester. Thus, high pitch rates of Brettanomyces could be useful to quickly reduce isoamyl acetate in beers where this flavour is undesirable. We also observed a more subtle dose-dependent decrease in concentration of two other acetate esters at high Brettanomyces pitch rate, ethyl acetate (pear, solvent) and phenethyl acetate (honey, yeasty, floral) (Figures 3, 4). Interestingly, we observed higher concentration of these esters in the low pitch rate sample. Confirming this phenomenon and its potential sensory ramifications would require an expanded experimental design including replicates. Ultimately the differences are subtle, and it is likely that these do not have significant effects on the sensory character of the resultant beers. Concomitant with the slight reduction in acetate esters by addition of Brettanomyces, we also observe slight increase in volatile acetic acid concentration with pitch rate (Figure 5), but the concentration regardless is well below the sensory threshold for acetic acid. We await sampling at the 3 month timepoint to assess whether these trends continue and become clearer.

We were not able to identify any clear trends in levels of volatile organic acids (Figure 5). These compounds typically cause off-flavours in immature Brettanomyces fermentations.

Discussion

There are a number of limitations to this experimental design which must be considered. First, the metabolite analysis was conducted using one bottle from each pitch rate at each time point, and therefore does not result in statistically robust data. Further experimentation can use multiple biological replicates (individual growths of the same Brett strain, dosed into bottles and tested in parallel at each time point). Furthermore, the sample size of the sensory panel (n = 3) does not allow for the use of statistics on the data. Additionally, this experiment was conducted with a single Saccharomyces/Brettanomyces combination. The choice of strain for either organism could influence the results. For example, yeast strains that do not produce 4-VG (POF-) do not provide this precursor to Brettanomyces, theoretically leading to a decrease in 4-EG concentration in the resultant Brettanomyces secondary fermentation. Furthermore, the choice of Brettanomyces strain may influence the rate and amount of flavor metabolite production in secondary fermentation. As such, it would be important in further experiments to include multiple Brettanomyces strains in order to correlate the strain-dependent and dose-dependent levels of flavor metabolite production and generate a better overall picture of the effects of Brettanomyces secondary fermentation on beer.

These results challenge the notion that "stressing" Brettanomyces by underpitching leads to more or different levels of flavor-active compounds. The results also demonstrate that exceedingly small amounts of Brett cells are needed to quickly cause changes in a beer. The low end of the pitch rates tested represents approximately 1/5 of a White Labs vial pitched into a 5 gallon batch.

See Also

Additional Articles on MTF Wiki

• Nick Mader's analysis of Brettanomyces pitching rates
• Caroline Tyrawa's Masters thesis data on co-pitching vs staggered pitching and impact of S. cerevisiae strain selection.
• Brettanomyces

External Resources

• MTF Discussion regarding these results.
• Mark Trent's anecdotal experiment on using 75,000, 50,000, 25,000, or 10,000 cells per ml in secondary (inspired by this experiment).