MEETING of the LONDON SECTION HELD at the CHARING. CROSS HOTEL, LONDON, on 14th OCTOBER, 1941
Mr. H. Heron, in the chair
The following paper was read and discussed: —
THE ATTENUATION OF WORTS AND BEERS
By L. R. Bishop, Ph.D., F.I.C.
The two extreme groups of top fermentation yeasts are the strongly attenuative Burton type and the weakly attenuative Yorkshire type. The behaviour of these two well-known races is characteristic, not only in the primary fermentation but also in the secondary and attenuation limit fermentations. The cause of this difference in yeast behaviour is, as yet, unexplained. As well as by this racial effect the extent of attenuation is also controlled by nutrition, sediment action and the. different types of rousing. These are considered with the object of assisting in the systematisation of our knowledge in relation to the primary, secondary and limit fermentations.
The fermentation to the attenuation limit is considered to be a useful estimation which has been unduly neglected in this country. A comparison of the methods proposed for its determination and of the sources of error discovered by various workers has been used, in conjunction with the general considerations given above; to obtain evidence of the best conditions for carrying out the determination of the attenuation limit.
One way in which the Institute Research Scheme can help at the present time is in assisting in the solution of some’ of the outstanding technical problems facing brewers. Experience of these problems has pointed to yeast weakness, with its attendant difficulty in controlling attenuations, as one of the main sources of technical trouble in British breweries. If anything, war-time conditions seem somewhat to have increased the difficulty, and in consequence considerable attention has been given to it.
In approaching the problem of an abnormally high racking gravity the first question to ask is whether there is too high a proportion of wort dextrins. Although this is not usually the cause, nevertheless, the carbohydrate composition of wort must first be examined. Similarly in the case of a racking gravity which is too low it is necessary- for the same reason to know the proportion of fermentable sugars in the wort.
Without the estimation of the proportion of fermentable sugars, by means of the attenuation limit test, any problem involving these must be dealt with by rule of thumb methods which may result in serious difficulty. In other uses, besides the study of yeast weakness, this estimation is of value. For instance, from the information it gives on the proportion of fermentable sugars left in a beer, as we are now aware from recent research of the danger of infection, or of yeast haze, resulting from too much sugar. On the other hand, if too little fermentable matter has been left flat beers may result. It is possible to argue that, with the present rapid trade and with the use of priming and carbonating, many of these troubles are not nearly so serious as they once were. So it is primarily from the standpoint of our general knowledge and control of wort composition that a sound method for the determination of the attenuation limit should be available.
Some of the errors which may arise in this determination, as it is usually carried out, have been encountered incidentally in the Institute researches, and the subject has been further investigated. Before presenting the results, the Yeast Sub-Committee wishes to have the assistance of brewers and chemists in considering all aspects and possible methods.
This paper, therefore, deals with familiar phenomena and is confined to the consideration of published work, so that the object is to bring out relations between well known facts in order to promote discussion. The relation of attenuation limit to wort attenuation in general is also considered with the intention of indicating how the Institute’s results bear on the problem of estimating the attenuation limit.
Attenuation, in the special sense in which it is used in brewing, refers to the fall of specific gravity during fermentation and may be regarded as taking place in three stages: —
(1) The attenuation from the original gravity of the wort to the racking gravity—the primary fermentation.
(2) The attenuation of the beer during storage and conditioning—the secondary fermentation.
(3) The complete attenuation.
In stating these figures, it is usual to give the “apparent” specific gravity—that is, the gravity shown with the alcohol present. In the case of the complete attenuation, the results may be given as the “real” specific gravity, that is, the specific gravity of the remaining solids when the alcohol has been removed. In order to avoid confusion, it is preferable to give this latter figure as percentage of unfermented matter.
Between the three fermentations— the primary fermentation, the secondary fermentation and the complete fermentation, it is possible to notice a number of similarities which are helpful in the understanding of the brewing process in general and in meeting possible errors of the complete fermentation, in particular, so it is proposed to draw attention to these first.
I. The Primary Attenuation to the Racking Gravity
When a wort has been pitched with yeast and active fermentation has started, the evolution of carbon dioxide tends to stir up the yeast and maintain it in suspension. If the yeast is of a Burton type, the wort of 1,040°O.G. and the wort depth is, say, 4 ft., then some 2 litres of carbon dioxide is evolved per square centimetre without other effects on the yeast than the stirring. But suddenly at this stage the yeast changes its behaviour, adheres to the carbon dioxide bubbles, and is carried to the top on the remaining small evolution of gas. As a result, the active fermentation rapidly ceases, although a proportion of the yeast tends to remain in suspension or on the bottom of the vessel. With a Yorkshire type of yeast the behaviour is similar except that, unless prevented, the yeast rises more gradually and at a much earlier stage, leaving an incompletely fermented beer nearly free from yeast.
Each of these extremes of behaviour can, at least in part, be corrected by suitable rousing. The usual type of rousing is that for correcting the incomplete fermentation of the Yorkshire type. This may be the perforated bucket, the swan-necked fountain, the Scott rouser or the centrifugal pump and spray. The essentials of this type of rousing are that it is used after the yeast has begun to rise and removes beer from the main bulk in order to wash the yeast, from the head, back into the zone where active fermentation can take place, thus, encouraging a more complete fermentation. On the other hand, with the strongly attenuative type of yeast some form of rousing is necessary before or at the moment when yeast rising begins in order to stir up yeast over the whole of the bottom of the vessel. This assists it to rise completely and so checks the tendency to excessive fermentation and the continual rising of yeast from the bottom of the vessel which results in too. much yeast in suspension. The apparatus for “bottom” rousing may consist of an implement similar to the rubber floor squeegee, or of a jet of air distributed over the bottom of the vessel or of a propeller near the bottom. Improved devices are, however, possible.
Rousing is important also in connection with the systems of fermentation discussed later, but the behaviour of some types of yeast in aggregating more readily is also of importance. This aggregation is only partially explained as yet, but it is relevant to review the evidence available. Quantitative studies have shown that, in experimental brewery-type fermentations without rousing, the quantity of sugar fermented is independent of the original gravity apart from its effect on the quantity of fermentable sugars. In other words, in two worts of different original gravity but with the same number of degrees of total fermentable sugar then, in the primary fermentation, not the whole of the sugar, but an equal number of degrees would be fermented in both worts; that is, if other factors such as type of yeast and nutrient supply were equal. These other factors, however, will also play their part, for a Yorkshire type yeast, as is well known, would ferment a smaller proportion in both worts than would a Burton type yeast. Similarly, it has been shown that in a reduced supply of assimilable nitrogen less sugar is fermented than if a full nitrogen supply were available. With the less attenuative types of yeast it has also been shown that excess of wort sediment checks still further the amount fermented; not, as it is usually explained, by clogging the yeast cell surface, but by assisting the evolution of carbon dioxide as a gas on the bottom of the fermenting vessel and so sweeping out the yeast more effectively. That is, a more efficient type of “bottom” rousing is provided for the type of yeast for which this is a disadvantage. Incidentally, the present reduction in substitutes, elimination of foreign barley and high nitrogen content of the home barley make a nation-wide experiment in increased nitrogen supply, and it appears to the writer that the results are worth recording.
The proportion of sugars in the wort, the amounts of assimilable nitrogen and of wort sediment only play a part in controlling the racking gravity, the predominant effect is due to the type of yeast exemplified by the extremes of the Burton and Yorkshire types. This can be readily shown by controlled laboratory fermentations in portions of the same wort using the technique described by Burns (this Jour., 1041, 10) or by Whitley and myself (ibid., 1938, 73). This is also a suitable way of assessing the value of a yeast for use as a change.
II. The Secondary Attenuation
In the secondary conditioning fermentation difficulties may arise from infection; but, apart’ from this, many similarities to the primary fermentation may be traced. With yeast of the Burton type the amount left in suspension tends to be high, and, as in the primary fermentation more stays in suspension. There is, therefore, a strong tendency to a secondary fret, but this is counteracted by the fact that the primary fermentation has been vigorous and has removed most of the fermentable sugars. On the other hand, beers from a fermentation with yeast of the Yorkshire type tend to have more sugars left but less yeast in suspension to ferment it, and, furthermore, this yeast aggregates and settles out very rapidly just as it was carried out in the primary fermentation. Consequently, with such yeast some form of rousing is necessary also in the secondary conditioning—such .as rolling the cask or using propellers in the conditioning tank. This corresponds to the “bottom” rousing of the primary fermentation. It is this small amount of yeast in suspension, and not the fermentability of malto-dextrins which is used to explain the slowness of conditioning. This fermentation is dependent on the yeast behaviour and also on the shape of the storage vessel. It has, for instance, been found, more especially in lager brewing, that it is better to place a conditioning tank lying on its side rather than upright, as more surface with settled yeast is exposed and the thickness of beer above is less.
In the secondary conditioning, as in the primary and attenuation limit fermentations, it is possible that yeast nutrition plays a part. This is under investigation.
III. The Complete Attenuation
When the laboratory estimation of the complete attenuation is carried out it is again possible to notice similarities in behaviour to the two previous fermentations. In a stationary fermentation—e.g., in a bottle on a forcing tray—yeast of the Yorkshire type tends to settle out readily and not to ferment well accentuating the tendency for the Yorkshire type of beers to have more sugar remaining. It may be argued that such an estimation would conform to the practical behaviour of the beer; but against this it may be pointed out that, as in the primary and secondary fermentations, the extent of fermentation might be expected to depend on physical conditions such as the shape of the container in which the attenuation is performed, so that, for instance, a short broad bottle would give a different result from a tall narrow one. Furthermore, in all such fermentations the amount of yeast in suspension is small and there is a distinct danger that the true end point may not be reached either from incomplete fermentation or development of infection. To hasten the attainment of the end-point some form of rousing is necessary.
In an earlier paper attention has been drawn to the differences encountered when studying a yeast problem by the chemists’ method in which the yeast is continually shaken and by the brewing method in which, even if roused, the amount of artificial stirring is small. The results under the two sets of conditions may bear little relation to one another. This is because under brewing conditions the physical behaviour of the yeast is predominant and it appears essential to adopt shaking in order to attain a true end-point in the attenuation limit test. Furthermore, it is considered that the shaking should be mechanical because, if hand shaking is adopted and even if it is most assiduously carried out, it normally only takes place during some 8 or 9 hours out of every 24.
Another important factor to be borne in mind is the nutrition of yeast which is necessary if a true end-point is to be reached. It has been shown that assimilable nitrogen compounds are necessary for yeast growth, so are minerals; and in recent years the importance of “bios,” that is of certain vitamins essential for yeast, has been widely recognised. Without these the fermentation and growth of yeast are checked, if not completely inhibited. This is most clearly shown in the determination of the fermentability of invert sugars and of glucose syrups. Here there is little nutrient matter and this must be overcome in one of two ways. If a heavy yeast dose is used it is possible to ferment completely but it is considered that certain errors are introduced which are discussed later. The alternative is to supply the yeast with adequate nutrition and, in our experience, yeast water is a partial but not a complete source of nutrients. Earlier work by Whitley and my self has indicated that a partial shortage of assimilable nitrogen and of bios may exist in brewing worts and may affect the primary fermentation, so that this possibility must be considered also in dealing with the attenuation limit fermentation of worts as well as of sugars.
IV. Is There a True and Definite End-Point to the Attenuation Limit?
In previous paragraphs the assumption has been implied that a definite end-point exists to the process of attenuation and it is proposed to take this as a basis for discussion because it is considered to be the more helpful, even if in the future it is found to require some modification.
It appears to the writer that in the past an indefinite end-point has been assumed partly because of the results of inadequate shaking and of inadequate nutrition as already mentioned. A further reason might be that, if the forcing results were taken as evidence of the limit, then any development of infection would tend to obscure the true end-point. Other reasons are considered later in dealing with recent papers on the estimation of the attenuation limit.
The view that an indefinite end-point exists has been supported by the opinion that malto-dextrins are only fermented slowly and that there is a variation among yeasts in their ability to ferment the malto-dextrins; but recent evidence is against this. Haehn, Glaubitz and Grosz (Woch. Brau., 1937, 54, 335; this Jour., 1938, 109) have prepared from starch a dextrin containing six glucose units and have shown that it is not attacked by the following yeasts: top, bottom, Saaz, Frohberg, Logos and Schizosaccharomyces Pombe and octosporus, Similarly, Baker and Hulton (this Jour., 1938, 514), have prepared a malto-dextrin and have found it unattacked by Saccharomyces Jdere viaiae, festinans, ellipsoideua and intermedius. It is still possible that there may be differences among these yeasts in their ability to attack malto-dextrins with three to five glucose units, but in the view of Baker and Hulton the amount of these present in wort is small or negligible.
Certainly, most workers using the attenuation limit test have found that different yeasts all give the same limit for a given wort. This was found by Schdnfeld (Woch. Brau., 1906, 489; this Jour., 1906, 743) and by Isotti (Woch. Brau., 1933, 181; this Jour., 1933, 497) for bottom yeasts and by Whitley and the writer for top yeasts (this Jour., 1938, 125). The most extensive data is that of Isotti, who gives the results for each of 26 different brewery yeasts acting on three different worts. Unfortunately, duplicate results are not given, but from the size of the standard deviation of the 26 determinations from the respective means [Wort I, 81-52 + fl-13; Wort II, 83-37 ± 012 and Wort III, 82-29 ± 0-14] it would appear probable that the variation could be attributed to experimental errors. Pirataky (Woch. Brau., 1932, 59; this Jour., 1932, 211) assumed that different yeasts give different limits, but judging from an abstract (Gambrinus, 1934, 36; this Jour., 1934, 287) Rokita and Gefundner have attacked this claim and have provided further evidence showing that it is the wort composition and not the particular yeast which decides the attenuation limit.
V. Published Methods for the Determination of the Complete Attenuation
The complete attenuation is of more immediate concern to lager brewers and in recent years there has been a revival of interest in the method as is shown by a number of publications. They appear in the main to be concerned with the desire for speed rather than accuracy, and are concerned chiefly with the physical conditions such as time, temperature and yeast rate. The most effective way of reviewing these appeared to be in the form of a summary table giving the details for a number of essential conditions where these are given in the text.
Of these papers that by Bengtsson, Elmfeldt and Andersson deserves particular attention, as these workers were concerned primarily with obtaining an accurate estimate. They showed by repeated measurement that during the test the apparent sp. gr. of wort falls steadily to a low limit and then rises again owing to autolysis of the yeast. They showed also that high temperatures such as 25°-30° C. (760-86° F.) encourage yeast autolysis and an incorrect result and that a similar result is produced by high yeast rates. This had been shown also by Schönfeld, and the error produced by temperatures above 20° C. (68° F.) is confirmed by the work of Silbereisen. The result might be expected to be a little more complicated than that given by Bengtsson, Elmfeldt and Andersson from the knowledge of the behaviour of yeast on autolysis as shown by the papers of Harman and Oliver (this Jour., 1925, 353) and Hind (ibid., 1925, 336). These authors showed that autolysing yeast excretes mineral matter and alcohol, the latter being in predominant amount. So that heavy yeast doses might tend to lower the apparent attenuation (by excess of alcohol) and to raise the real attenuation (by excretion of mineral matter). With this criticism of the danger of yeast autolysis in view, it is possible to suggest that nearly all the recently proposed methods (those of Schreda and Stolpp, of Schild, of Silbereisen and of Schuster and Kärnbach) might be in error from the products of yeast autolysis. Naturally the danger is greater if stale yeast is used as is shown by Schild. These workers all take the apparent attenuation, but further support for the suggestion of a double source of error with heavy yeast rates is provided by Schuster and Kärnbach, who have tested their own method and that of Schild and of Silbereisen and have shown that by these methods the real and apparent attenuation do not agree. They deduce that there is a loss of alcohol, but it may be that the result is produced by the excretion of mineral matter from the high yeast rates which affects the “real” attenuation result. However, the methods of Schild and of Schuster and Kärnbach use aeration so that the danger of alcohol loss exists in these although not in that of Silbereisen. The method of Schild in which 200 c.c. of beer or wort is shaken mechanically in a 2 litre flask appears particularly liable to such a loss.
The ideal appears to be to use a small yeast rate which will grow actively and so postpone autolysis. Bengtsson, Elmfeldt and Andersson use 0·5 percent yeast rate for this reason, after testing other rates. Helle, on the other hand, determines the limit in beer stirred by a paddle but without a special yeast addition. Since he takes the precaution of sterilizing the containers to avoid acetic infection it may be that, with this very low yeast rate, such a danger is more immediate.
As far as can be ascertained, earlier methods did not employ mechanical agitation, but aeration or some form of mechanical agitation is used in the method of Schüler and in all recent methods (those of Schreda and Stolpp, of Schild, of Silbereisen, of Helle and of Schuster and Kärnbach). It may be pointed out that, besides hastening the determination, agitation should tend to help the completion of attenuation before autolysis sets in and so to separate these two actions more clearly.
Whilst most of the recently proposed methods aim at rapidity rather than accuracy, as is shown-by the determination of the end point by special accurate hydrometers rather than by the pycnometer or sp. gr. bottle, from the present stand point it seems desirable to aim at accuracy rather than speed. In this connection it is notable that none of the methods proposes the addition of nutrients. However, in the method given by W.A.Davis (Int.Sugar J., 1938,40,235), ammonium phosphate is added to molasses to assist in the determination of the attenuation limit. This method is Carefully worked out in other respects also. Pure culture yeast is used, the alcohol escaping with the carbon dioxide is caught in an acid trap and returned to the liquid before distillation. The distillate is then, made alkaline and redistilled to remove volatile acids. ‘The flask is, however, not shaken and the fermentation temperature 33° C. (91° F.) appears high by brewing standards.
The result of these considerations is to suggest that the ideal method should include the following conditions: —
(1) The yeast rate should be small—around 0·5 percent.
(2) The fermentation temperature should not be too high—around 20° C. (68° F.).
(3) Steps should be taken to avoid loss of alcohol, or to make it regular so that allowance can be made for it.
(4) The fermentation should be mechanically shaken.
(5) The yeast nutrition should be adequate.
Such a determination would occupy some considerable time but should not take excessively long in order to avoid autolysis. For those who desire a more, rapid estimate of the attenuation limit it is considered that a determination of the carbohydrate composition of the wort offers the most rapid means. One attempt in this direction has been made by Piratsky (loc. cit.) who used the copper reducing power of the worts as a measure, but general experience in this country suggests that this would not be an adequate measure.
Discussion
Mr. B. M. Brown said that there seemed to be two attenuation limits to be considered. One, an absolute limit which may be used as a measure of what a wort and a yeast can do with a plentiful supply of bios and nitrogen and with ample yeast kept in contact with the wort; the other, a practical limit which would apply to the worts and yeasts of any brewery and which could be used as a measure of day-to-day fermentations. He was not convinced that all yeasts gave the same limit figure. His experience was that a good many English pure culture yeasts gave the same figure, but there seemed to be differences among some outstanding types. The method he made most use of was a shaking procedure similar to Dr. Bishop’s but carried out at 60° F., with a pitching rate of 3·5 gms. of centrifuged yeast per litre of wort at 1040°, gravities being taken at 48 hours. The results were not always very consistent. He would like to know a little more of the method used by Isotti.
Mr. L. C. Thompson said his yeasts were giving lower finals than before the War and as his wort was brewed with an increased proportion of malt owing to the decreased amount of sugar available he supposed the conditions favourable to yeast weakness were decreased. He could understand a brewer of continental lager beer attaching great importance to the limit of fermentation because he would want to know the most favourable time for dropping his beers into his storage cellar; he would be anxious not to postpone that too late and leave no fermentable margin for his yeast during the long conditioning period of 2 to 3 months. English brewers were largely spared that anxiety, particularly when brewing running mild ales, which could be mashed, fermented, racked and drunk within a period of 10 days. He thought members were too prone to think of bottled ales in discussions and so ignore that which still formed by far the larger portion of the trade of the country—draught beer.
Mr. C. A. Kloss said he believed Dr. Bishop suggested that there was a definite limiting attenuation, which could be attained with any type of primary yeast, depending only on the character of the wort, and that if the usual attenuation achieved in a brewery was expressed as a percentage of the theoretical limit of attenuation a step towards correlation of the various idiosyncrasies of fermentations might be brought about. The measurement of the limit of attenuation was really a rapid means of arriving at the wort composition. If the method were sufficiently quick and reasonably accurate, some sort of control might be effected; that is, when. certain troubles arise, the percentage of the total possible fermentable matter might reveal some of the underlying causes.
Mr. P. G. Wilkinson said that Dr. Bishop’s work on the attenuation of worts and beers had been directed chiefly to demonstrating deficiencies of nitrogen and bios in the wort, but with the present day nearly all-malt worts, it was useful to know the actual amount of fermentable matter present. Dr. Slator has recently described such a method (this Jour., 1941, 207). The process that he (the speaker) used, was to ferment 100 c.c. of sugar or wort with 5 gms. of yeast and after fermentation at 70° F. to boil off the alcohol and take the residual sp. gr., the gravity imparted by the yeast being allowed for by a control. A more reliable procedure was to conduct the test fermentations against a control of pure cane sugar solution which was assumed to be completely fermented. With wort gravities between 1020° and 1060° using cane sugar controls at corresponding gravities, it was found that between 1030° and 1060° there were differences of only 1·5 per cent, in the fermentable matter. He had experimented with brewers’ yeast and laboratory growths of yeasts of various types, such as clumping, non-clumping, Burton union and Yorkshire, and he found that all gave much the same results for the fermentable matter in a wort. Under standardised conditions differences could be reduced. For example, a wort was sterilised in bottles and tested with the current- brewery yeast in July and again in September, the percentages of fermentable matter were respectively 63·9 and 63·0. Similar experiments at other times on different worts gave differences of 0·1 per cent, and 0·5 per cent. The test was designed to eliminate differences due to deficiencies in yeast nutrition, it was simple and the time taken was 3 days, but with a rocking tray the time could probably be shortened consider ably. The test did not need bios concentrates and gave a measure of the actual total fermentable matter in the wort irrespective of the brewing conditions.
Mr. McHugo said that he agreed with Dr. Bishop that it was very important that a yeast should be capable of fermenting wort down to a stable point, beyond which was little risk of further attenuation. He thought, however, that any method for deter mining the attenuation limit, if it was to be of much value to the brewer, would have to be one that was simple and rapid to carry out. He believed that if the supply of nutrient materials, both in the form of soluble nitrogen, mineral salts, etc., in the wort were adequate to maintain the yeast in a healthy condition, very little trouble would be experienced in respect of unsatisfactory or fluctuating attenuations. It did, of course, sometimes happen that when the yeast from one brewery was introduced into another, and even when nutrition was satisfactory in both cases, the yeast would function differently in the new environment, with the result that there were, at first, marked differences in the degree of attenuation. This phenomenon was, he thought, connected more with the enzymic capacity of the yeast, rather than its reproductive power. It was well known that the activity of the enzymes within the yeast cell were very sensitive to change and environment but when the yeast had acclimatised itself to its new conditions the differences in attenuative power became less, and ultimately both yeasts would behave the same in that respect.
Mr. J. A. Burns said that even in low gravity beers the health of the yeast and the regularity of the fermentations depended on wort composition; and since the attenuation limit was a fundamental character of a wort, the determination of that value might be useful in brewery control. Dr. Bishop had demonstrated the essential conditions and requirements of such a test.
Crescent City Brew Talk 