Brewing  Research report.

Besflow Device Report


CS 2530

Chris Giles
Lesley Buggey 

June 2005

Evaluation of the BesFlow to extend the shelf life of cask products

Executive Summary

 Results gathered from the microbiological assessment and the flavour profiling appear to confirm that the cask fitted with the BesFlow device did show an improved beer quality, compared to the cask fitted with the traditional shive.
 The BesFlow device resulted in a reduced number of contaminating bacteria and non-brewing yeast present in the cask, which resulted in a lower level of acetic acid flavour being observed during the sensory profiling.


A new device for cask dispensed products, designed to replace the traditional shive. The new device fits into the traditional shive hole, but has a bag which inflates inside the cask as beer is drawn off. Due to this design feature, air from the pub cellar atmosphere is not permitted to mix with the cask product. This has two main advantages:

 Any bacteria or yeast contamination present in the cellar environment and drawn in to the cask is not permitted to mix with the product, so reducing the chances of off flavours being produced via contamination.
 Atmospheric oxygen entering the cask is unable to act upon and oxidise the product, thus the beer is protected from the negative flavour effects of oxidative staling.

This project is designed to expand upon previous work carried out by the company at a preliminary stage of development. The device has now been refined and the designers would like to have the product tested independently to show that it can extend the shelf life of cask ale. This work has been carried out by BRi and includes an evaluation of both the microbial stability and the flavour quality of cask ale subjected to storage utilising the device.

Scope of work

Brewing Research International has assessed the BesFlow over a four week period, this is to determine if the inclusion of the BesFlow device into the cask increases the shelf life of the product compared to a normal cask.

The two products chosen for evaluation were cask beers which had different flavour notes, with varying alcohol levels.

 Timothy Taylor’s Landlord at 4.3%ABV (BB 19/05/05); a pale ale with a well-balanced flavour with hoppy, citrus and nutty malt characters. This was chosen as it is a light, refreshing beer which should allow any off-flavours that develop to be readily identified.

 Jennings’ Snecklifter at 5.1%ABV (BB 13/05/05), a darker beer with more complex flavour characters. This was chosen as beers such as this tend to age very differently from lighter styles as their greater malt content and higher ABV can lend a protective 'anti-staling' effect.

The products were dispensed directly from the cask, this was in order to prevent external effects from any dispense system having an influence on the results. The casks all had 1 pint poured off per day, in order to introduce air into the cask from the cellar environment, thus simulating a low turnover outlet. Immediately prior to taking samples for analysis, a pint of beer was also drawn from each cask.

For each product evaluated, 2 casks were set up:

Cask A1: Landlord Control - a traditional shive was used
Cask A2: Landlord Test - the BesFlow device was used

Cask B1: Snecklifter Control - a traditional shive was used
Cask B2: Snecklifter Test - the BesFlow device was used

The beers were obtained from a reputable source and stored in an area which would replicate the worst practices of bad cellar management. This created an extreme level of evaluation and allowed differences in the cask to become more apparent.

Samples of the beers were taken over a 4 week testing period. For the microbiological assessment of the product, samples were taken twice per week. The sensory evaluation was carried out once per week. The BesFlow device was also sampled microbiologically to assess weather the manufacturing of the bag and shive might be itself a source of contamination in the cask.

The control beer provided a bench mark for the beer in the cask under test, being used to demonstrate the typical ageing profile of a cask ale in-trade and showing the point at which the beers’ characteristics changed. A comparison of the two sets of data allows an assessment of the action of the BesFlow device and has provided

information as to whether it improves the shelf-life of the cask ale over the period of evaluation.

Four weeks was considered a sufficient period of sampling as this represents a considerable extension to the normal shelf-life of cask ale in trade and, if successful, should demonstrate the effectiveness of the BesFlow device.

Experimental methods

There were two assessment methods used in this project; microbiological assessment and flavour profiling.

Microbiological Assessment

The casks during the course of the project were aseptically sampled twice per week, for the assessment of microbial contamination. Prior to sampling, the taps were cleaned using a 70% ethanol spray as a contact disinfectant. Once dry, a quantity of beer was drawn off the casks and discarded; samples were then aseptically taken for testing. The taps were then re-cleaned with 70% ethanol spray.
Beer samples were plated onto specific growth media, to enumerate and identify the presence of spoilage and non-spoilage bacteria and also non-brewing yeast.

Microorganisms, which are able to grow during the brewing process or in final beers in cask are considered to be spoilage organisms. Spoilage is usually manifested by abnormal flavours, due to the excretion of specific metabolites, and haze due to the presence of large numbers of microorganisms. Acetobacter species are the main bacteria contaminant for cask products in cellar environments. These are readily pulled into the cask as the beer is drawn off, leading to the production of acetic acid, thus acidifying the beer and affecting flavour quality. Non-brewing yeasts may also be present, either during the racking of the beer into cask or from the cellar environment. This can result in the production of phenolic (antiscptic or TCP-like flavours) and acidic off flavours, along with the production of hazes.

Flavour Evaluation

The BRi expert panel tasted each beer once per week during the course of the trial. The casks were equilibrated at cellar temperature (12°C) prior to tasting and the samples were presented anonymously in lidded dark glasses to prevent visual clues, such as hazes. The aroma, taste and linger (aftertaste) were considered. Any flavours perceived but not included on the profile form were recorded under the general term ‘other’. The tasters were encouraged to comment on the nature of individual notes and these comments are included in the written description for the samples. The intensity of each note was scored on a scale of 0 to 9, where 0 = absent and 9 = intense. The tasting results for each beer were combined and average scores calculated. The flavour evaluations are presented as a comparison of the aroma and taste profiles of each cask over the four-week period. In addition, the two casks for each beer have been compared and any differences highlighted.



Cask A – Landlord Bacterial Results

The microbiological counts observed for these casks showed that the numbers of bacteria were closely matched in the first two weeks, although the control cask recorded higher results (appendix 1). Following this initial period, the control cask continually recorded much higher bacterial counts than for the cask fitted with the BesFlow device.

The sample from the start of the fourth week for the BesFlow cask showed a result which was a dramatic increase on the previous week, however the level of contamination reduced the week after. This result was unexpected and may have been due to sampling error,

Cask A – Landlord Non-Brewing Yeast Results

The yeast population is not greatly affected by the BesFlow device, as non-brewing yeast strains are already present in the cask and they are able to survive without the presence of oxygen. Non-brewing yeast are also able to utilize the sugars not used by the brewing yeast during the fermentation, which offers a source of nutrition. Yeast will also be present in the environment of the cellar, and would naturally be drawn into the cask as the product is drawn off.

The results observed showed that the control cask recorded the higher non-brewing yeast cell counts (appendix 2), for the whole of the experiment. The non-brewing yeasts recovered were also considerably lower then the bacteria recorded over the same period, indicating that the product had low levels of non-brewing yeast present during racking in the brewery.

Cask B – Snecklifter Bacterial Results.

The bacterial results from the Snecklifter casks mirrored those from the Landlord casks. The bacterial numbers recovered from the control cask were consistently higher than the results recorded from the cask with the BesFlow device (appendix 1). In the first two weeks the casks were closely matched in bacterial numbers recovered; from week three the control cask recovered significantly more bacterial cells than the cask fitted with the BesFlow device.

Cask B – Snecklifter Non-Brewing Yeast Results

The wild yeast results for the Snecklifter also mirrored the result from the Landlord casks. The control cask again recorded higher non-brewing yeast cells than the cask fitted with the BesFlow device during the experiment (appendix 2). The numbers of cells recovered were significantly higher than for the non-brewing yeast cells from the Landlord casks.

Landlord - A1
The aroma of Landlord was initially sulphury and this decreased over the 4 weeks. Hop character remained similar and malty/toffee flavours increased, as did the winey/sherry and acetic flavour.
On taste, the beer lost some of its sweet and fruity characters and became wine-like, vinegary, metallic and very sour. Between weeks 1 and 4, there was a significant increase in winey flavour.

Landlord - A2
The aroma of A2 at week 1 was very sulphury and this masked other aromas. At weeks 2 and 3 the hoppy and fruity character of the beers were prominent and by week 4, some toffee, winey and papery notes were detected.
The taste of A2 lost the sulphur and fruity character over the course of the 4 weeks and the beer flavour increased in perceived acetic and sour notes. By week 4, the beer had significantly higher winey-sherry flavours.

Comparison of Flavour Profiles – Landlord
Compared to A1, A2 had greater sulphur character and lower overall aroma at week 1. By week 4, A1 had greater perceived malty/toffee aroma and greater winey and acetic notes.

On taste, both beers were considered to be sour with a harsh, metallic bitterness and A1 had slightly greater perceived winey and acetic notes.

Snecklifter – B1
The aroma of B1 was described as dark chocolate, burnt, black treacle and sulphury. These notes decreased over the 4 weeks and fruity, sweet, hoppy and sherry-like aromas increased.
On taste, the initial burnt character of the beer diminished and was replaced with sweet, fruity and malty notes. By week 4, the beer had greater sherry-like character and was very sour.

Snecklifter – B2
The aroma of B2 was initially burnt and sulphury, however this was gradually replaced by sweet,fruity and sherry-like aromas.
On taste, the beer decreased in burnt and hoppy flavours and increased in fruity
and sherry flavours. By week 4, the beer was alcoholic/solvent-like and sour.

Comparison of Flavour Profiles – Snecklifter
B1 and B2 had similar aroma profiles at week 1. B1 had enhanced perceived burnt
character and B2 more citrus and sherry-like notes. By week 4, both beers were
perceived as having greater fruit, sweet, sherry and alcoholic notes.
Both beers had similar taste profiles at week 1 and at week 4. Differences in flavour were perceived such that B1 had enhanced sour notes and B2 enhanced alcoholic, malty and sweet notes. Statistically, both B1 and B2 were more winey and sherry characters at week 4 compared to week 1.

The microbiological results showed that the BesFlow device, when fitted to a cask, did reduce the recoverable bacterial and non-brewing yeast strains.
Over the first two weeks, while the numbers of contaminating bacteria and yeast were closely matched between the control and test casks, the results from the control cask were always higher than the cask fitted with the BesFlow device.
Following the first two weeks however, the levels of bacteria recovered increased significantly, and this was especially noticeable in the Snecklifter control cask. The bacterial numbers in the casks with the BesFlow device did show increased numbers beyond this time; however this was a gradual increase over the following weeks, with no marked increases. The only exception to this was Landlord at the start of the fourth week, however the number of bacteria reduces for the following sampling point, indicating that this is most likely to be a result of sampling error.
The non-brewing yeast results also show that the control casks recorded higher results than the test casks with the BesFlow device. With Landlord, the levels increase after the first two weeks, in both casks, however the levels of non brewing yeast in the cask fitted with the BesFlow device drops off over the subsequent weeks, while the control cask stays at the elevated levels.
With the Snecklifter, the results closely replicate those for the bacterial contamination detected; the levels recorded for the control cask increased at weeks three and four, before dropping off at the end of week four, while the cask fitted with the BesFlow device had small increases from week two until the end of the experiment, with the results significantly lower that those recovered for the control cask.
The contamination detected from the BesFlow device was limited to a lower level of moulds and Bacillus sp bacteria. These are not recognised as capable of beer spoilage and should present no quality control issues (Appendix 3).

The sensory results indicate that there is a noticeable difference in the flavour of each beer over the four week period. By the end of the experiment aged notes were detected that were absent on the beer when fresh.
As a beer ages, flavour changes that can occur include the loss of initial eggy-like fresh sulphur notes, a general decrease in fresh hop and fruit character and a lowering of perceived bitterness. As the beer ages, it may also increase in toffee character and become sweeter. At some point, other aged notes, such as acetic, solvent-like, winey-sherry, cardboard, rancid, ribes, fatty/soapy and metallic may became apparent. The flavour of different styles of beer will age in varying ways and the state flavour characteristics of each beer are unique to any particular beer style, composition, package-type and storage conditions. Flavours such as acetic and rancid notes indicate contamination, whilst winey, metallic and cardboard notes, for example, are the result of oxidation of the beer.

The changes in flavour for each of the two brands indicated that the control cask of each beer showed higher levels of perceived contamination-related flavours compared to the cask fitted with the BesFlow device.
Landlord ale is particularly hoppy and bitter style of ale. Although this bitterness and hop persisted, the control cask showed perceptibly higher acetic acid character and enhanced winey notes compared to the cask fitted with the BesFlow device. These characters are indicative of greater staling in the control cask.
Snecklifter is a different style of ale in that it has a high burnt, liquorice, toffee and coffee character and some sherry-like flavours when fresh. As the beer aged, the sherry and toffee flavours increased for both casks. Although similar in final flavour profiles, the control cask was perceived as more sour compared to the cask fitted with the BesFlow device, although the presence of any specific acetic acid flavour was difficult to detect in this style of beer at this stage.
It is the conclusion of the sensory profiling that the control cask of each brand was more susceptible to flavour changes caused by spoilage (acetic acid) compared to the cask fitted with the BesFlow device. Both beer displayed changes due to oxidation during the course of the trial, however these were not as apparent in the cask fitted with the BesFlow device compared to the control cask.

Results gathered from the microbiological assessment and the flavour profiling appear to confirm that the cask fitted with the BesFlow device did show an improved beer quality compared to the cask fitted with the traditional shive. The BesFlow device results in a reduced number of contaminating bacteria and non-brewing yeast present in the cask, which resulted in a lower level of acetic acid being observed during the sensory profiling. Stale flavour characters were less apparent in the beer from the casks fitted with the BesFlow device. This was practically noticeable in the lighter flavoured style of beer.

Future Work

This project has demonstrated the success of the BesFlow device. BRi believes that a collaboration with a brewing company should be sought for further retail based trials, as part of an ongoing scheme of work which will result in the commercialisation of BesFlow’s unique product.

Appendix 1 – Bacterial Contamination