DISTILLATION OF WINES
The fermented must, generally called wine in France and grappe in the West Indies, contains, in addition to ethyl alcohol, the proportion of which in the musts of the rhummerie usually varies between 4 and 8%, fixed matters and volatile matters. Among the fixed products, some are suspended in the liquid: bagasse particles, yeast cells, albuminoid substances, gums, waxes, etc. ; the others in solution: unfermented sugar, organic salts and organic acids (succinic, lactic, tricarballylic, etc.). Volatiles are organic acids (formic, acetic, butyric, etc.), homologues of ethyl alcohol (methyl, propyl, butyl, amyl etc.), esters (formate, acetate, ethyl butyrate, etc.), aldehydes (formaldehyde acetaldehyde, furfurol etc.), terpenes, etc. Various gaseous products, especially carbon dioxide and, in some cases, ammonia, hydrogen and hydrogen sulphide, are also present in the soluble state.
We give, hereinafter the physical constants of the main volatile constituents of musts:
The density is given, as far as possible, at 20° or 15°; m = miscible or soluble in all proportions; p = poorly soluble; i = insoluble; the number mentioned for the solubility expresses the mass dissolved in 100 gr. of solvent. The data relating to the physical constants of certain present bodies, according to the authors, has large divergences.
In the manufacture of industrial alcohol, one seeks to obtain pure and concentrated ethyl alcohol, completely rid of anything which may accompany it. In the distillation of spirits on the contrary, bad tasting impurities are eliminated, while those which give the alcohol a pleasant odor and taste are preserved. The way in which the separation of volatile impurities is carried out has a considerable influence on the quality of the resulting eau-de-vie.
Theory of distillation (1)
(1) Mariller (Ch) – Distillation et rectification des liquides industriels. Paris, 1935.
The separation of the components from a liquid mixture, obtained by subjecting the latter to partial vaporization and separately collecting the vapors and the residue, is known as fractional distillation. The most volatile elements accumulate in the liquid from the condensation of vapors (distillate) and the less volatile in the waste liquid.
When the liquids are immiscible (water and higher alcohols, for example), the boiling temperature of the mixture is lower than that of the various constituents taken alone. It is reached when the sum of the vapor contents of the elements is equal to the atmospheric pressure. The composition of the evolved vapors depends solely on the relative volatility of the constituents, whatever the proportions of these in the mixture. It is given, in the case of two liquids, by the following relation which closely resembles Dalton’s law on the mixture of gases:
where a and b are the weights of liquids, m and m’ their molecular weights, F and F’ their vapor pressures.
It is not the same when liquids are miscible or soluble in one another (water and ethyl alcohol). In this case, the vapor pressure of each of the constituents is diminished by the presence of the other, and, consequently, the sum of their tensions is lower than the tensions of the two liquids taken separately. The boiling point of the solution depends on the proportion and nature of the constituents: it may be lower, intermediate or higher than that of the constituents. The composition of the vapors is deeply influenced by that of the mixture.
Water-ethyl alcohol mixture.
The first important study of alcoholic distillation is attributed to Duclaux (1). This scientist distilled a known volume of an alcoholic liquid of known composition and he collected the distillate in measured and equal portions, which were then analyzed. He established the following relationship:
(1) Ann. Phys, China, 1878.
in which a = alcohol % by volume in the generating liquid
e = water
da = alcohol % in the vapors
de = water
This relation is represented by a hyperbole having as equation:
Duclaux determined the values of m, for mixtures of various alcohols with water, and obtained the following figures:
It follows that the vapors are richer in alcohol than the generating liquid, and that this enrichment is all the more pronounced as one rises in the series of alcohols.
This is so at least within the concentration limits indicated by the author. For very high alcohol liquids, there are azeotropic or eutectic mixtures, with isothermal boiling under constant pressure, which give vapors having the same composition as the generating liquid. If the alcohol concentration is increased, the vapors become depleted.
In the case of ethyl alcohol, the azeotropic mixture contains 97.15% alcohol and boils at 78.15°C under normal pressure (Dorosewski). It is therefore not possible to produce absolute alcohol by simple rectification under 760 mm.
Gröning and Sorel experimentally determined the alcoholic richness of the vapors emitted by boiling liquids, containing a given proportion of alcohol. We reproduce below an extract of the tables established by these authors. The last 3 columns indicate the values of the enrichment coefficient:
deduced from the works of Duclaux, Gröning and Sorel:
Duclaux and Gröning operated with a retort radiating freely in the air, in which, consequently, condensations could be produced as a cause of error. Sorel, on the other hand, used a retort completely immersed in a bath, thus preventing any disturbing radiation. Also, his figures, which have been confirmed by Bergstrom, are more accurate scientifically.
Under the conditions of practice, the results obtained may, however, depart significantly from those of Sorel, the mode of execution of the distillation having a significant influence on the richness of the vapors. Slow distillation, which favors radiant condensation, increases the alcohol content, which is reduced by the violence and the speed of boiling. The presence of certain salts more soluble in water than in alcohol (alkaline chlorides and alkaline earth, for example) also increases the richness of the vapors. This property has been used to produce absolute alcohol (Mariller process).
Duclaux obtained the percentages of alcohol below (in volume), in the successive tenths collected by the fractional distillation of liquids of various concentrations. These figures show the pace of exhaustion.
According to the results above, it would be necessary, to completely exhaust a liquid at 10% alcohol for example, distill 60% of the latter. Practically, if the distillation is slow the condensation by the radiation plays an important role and makes it possible to obtain a faster exhaustion. Thus, with the still Charentais, it is enough to distil a third of the volume of a wine to 10 ° to exhaust it. [Fascinating, I’ve never seen this spelled out as such.]
Complex alcoholic mixtures.
Fermented musts contain, in addition to alcohol and water, various volatile impurities, in generally very small proportions. The vapors may have a higher content of volatile products, equal to or less than that of the mixture, depending on the substance in question.
According to Sorel, the proportion of the impurity in the vapor released depends mainly on its solubility in boiling alcohol. This solubility can be represented by the relation:
s = KS
s being the relatively small weight of the impurity contained in 1 kg of steam, S the weight of the same impurity in the generating liquid, K a constant coefficient for each mixture (solubility coefficient) and variable depending on the alcohol content of the generator fluid.
Sorel has experimentally calculated the values of K for various alcoholic strength and for 8 different bodies, constantly keeping s less than 2%. We reproduce below the results obtained:
When K is less than unity, the impurity remains in the generating liquid, and when it is greater than unity, it escapes and concentrates in the vapors.
Guinot (1930) resumed the evaluation of the coefficient K, by operating on liquids containing 1 gr per liter of the impurity considered. For a liquid at 10°, the content was 1% alcohol and for a liquid at 100° 0.1%. At these concentrations, much lower than those used by Sorel, the author obtained the following results:
Instead of the coefficient K, the ratio of the impurity level in the vapor to the level of the impurity in the liquid, Barbet considers a purification coefficient, K’, ratio of the impurity of alcohol in steam to the impurity of the alcohol of the generating liquid, which translates in a more accurate and more precise way over the course of purification for the product under consideration. The enrichment coefficient of the mixtures of water and alcohol being da, K ‘ is given by the relation: [SOS there may be errors in the working here.]
The variations of K’ are, according to Barbet, the result of a certain number of physical properties, among which the boiling temperature proper to each impurity plays an important part. The presence of water in various proportions also intervenes, by modifying the solubility of the impurities in the mixture, by changing the proportions of the vapor tensions, and so on.
Barbet has established, starting from the figures of Sorel, a graph giving in the form of curves, the various values of K’ according to the alcoholic richness for various impurities. He drew the line of the abscissae at the point K’ = 1, where the purification changes sign. The row of ordinates is divided into regular intervals K’ = 2, K’ = 3, etc., above the origin; and in intervals of equal length K’ = 1/2, K’ = 1/3, etc. (1) below the origin. We added the curve of methyl alcohol, according to Zaharia.
(1) The ratio of the impurity in the two fluids is indeed double, or triple, etc., when K’ is less than unity, just as this ratio is double, or triple, but in the opposite direction when K’ is greater than 1.
FIG. 23. — Graph of tornary mixtures
It can be seen that the impurities studied are almost always represented by downward curves.
According to Mariller, the volatile products of musts, very soluble in alcohol and little or insoluble in water, form, in the case of low alcoholic degrees, a mixture of two immiscible liquids and are very quickly entrained by the vapors. But, as the proportion of alcohol increases, the products become more and more soluble, and their tension, already reduced by the lowering of the temperature, is more soluble: as a result, the impurity becomes less and less volatile. [I think this describes some fusel alchols during double distillation of low ABV mashes. They can appear in the heads of a the first distillation then the tails of the second as alcohol concentration increases.]
To obtain an ascending curve, the impurity must be more soluble in alcohol than in water: this is the case of ammonia, for which the coefficient K’ increases with the alcoholic richness (Mariller). It would be the same for methyl alcohol, according to Zacharia, Angelescu and Motoc (2), who, by operating on a mixture of methyl alcohol (1%) and diluted ethyl alcohol, have obtained the following results: [This is why it is so hard to separate methanol and therefore it must be limited at material selection and fermentation.]
(2) Bull. Ass. Chim. LIII, 243, 1936.
As long as K’ remains greater than unity, the vapors are richer than the generating liquid in impurities: they pass into the first portions of the distilled liquid (head products). When, on the contrary, K’ is smaller than 1, the impurities tend to accumulate in the liquid and pass into the distillate at the end of the operation (bottoms).
If the purification coefficient is very high as is the case for methyl formate, ethyl formate, and ethyl acetate, for example, the products are clearly in the heads: they pass entirely into the distillate provided that the boiling is sufficient. [Keep in minds phenomenons like entrainment, particular to ever still design, can push congeners further back along the run than you’d think. If you have tons of ethyl acetate, it appear well into the hearts which is why the grand arôme rums cannot cut away their very large surplus.]
For various impurities (higher alcohols, amyl acetate), K’ is greater than 1 in low degree of ethanol and smaller than 1 in high degrees. These bodies will be eliminated quickly by the boiler, but they will be stopped if they encounter sufficiently concentrated alcohol on their route: low head products (at 40°), they become tail products at high degree (above 65 – 70°). In a column apparatus, amyl alcohol, for example, concentrates on the first plates up to the plateau at 40° where the level of the impurity reaches its maximum (K’ = 1); then it is retained more and more (K’ <1). For propyl and butyl alcohols, the high concentration plateau (70 – 75°) is much closer to the top of the column: also these products will arrive faster than amyl alcohol to the sampling glass. [Esters of higher alcohols are more likely to form was these dense zones of higher alcohols accumulate in the column. Some of these esters can stick out and be overly salient when there is not a broad amount of esters and is considered a flaw, however it is embraced by many.]
Finally, when K’ is smaller than 1 and increases with the alcoholic richness (methyl alcohol), the impurity can not be eliminated during the simple distillation, since once evaporated, it passes through the concentrated alcohol. It is only through the use of pasteurization that it will be possible to get rid of it. [I have never heard of a term associated with Pasteur in this context.]
We reproduce below, by way of example, the composition of the products collected on the various plates of a wine distillation column (after Wilkins and Margaret) (1):
(1) The Wine Review VII, N° 7, 7, 1939.
When the distillation is carried out by means of a batch still, the nature and the quantity of impurities passing in the good-tasting alcohol are regulated by separating a greater or less proportion of head and tail products from the distillate. If a continuous column is used, the impurities can be extracted from the trays where they have the maximum concentration. Most often though, in the manufacture of alcohol, no extraction is made at the column: it is avoided that the brandy is soiled by tail products with an unpleasant taste, by collecting it at a rather high degree (at least 55°). [I chose not to mess with this paragraph too much. I think Kervegant simply means there is one extraction point used and not attempts to be very selective with the trays.]
The distillation processes can be classified into 3 groups: discontinuous distillation with fraction recycling, discontinuous distillation without fraction recycling and continuous distillation.
Distillation with recycling.
In this mode of work, the liquid to be distilled is placed in a boiler heated by open fire or steam and connected to a cooling coil. The alcoholic vapors released in the boiler are immediately condensed in the refrigerant.
The distillate obtained has a relatively low alcohol content. In order to have complete exhaustion, it is important, according to Duclaux’s experiments, to distil 40% of the wine if it contains 5% alcohol and 60% if it contains 10%. Although, under the conditions of practice, these rates can be substantially lowered, as a result of the vaporization and fractional condensation phenomena which occur when heating relatively slowly, it remains however that the product of the distillation can not measure more than 25 – 30°. [Why do those first figures seem backwards to me?]
Ces eaux alcoolisées, appelées petites eaux ou brouillis, doivent être soumises à une seconde distillation, ou repasse, pour donner l’eau-de-vie marchande. L’opération doit être conduite avec beaucoup de soins, de façon à conserver dans le distillat les produits aromatiques agréables au goût et à l’odorat et à éliminer ceux qui peuvent diminuer la qualité de l’eau-de-vie. On recueille à part les produits de tête et de queue, que l’on ajoute généralement aux brouillis suivants ou que l’on soumet parfois à une distillation spéciale. On ne retient comme
This section will only get us a few pages shy of page 250 of 500, but then I’ll switch to going by chapters as they shorten. Maybe we can be done by the end of the summer then repackage it by chapter so its easier to reference.
[First half of paragraph below carried over from the previous section.]
These alcoholic waters, called small waters or brouillis, must be subjected to a second distillation or re-passing, to yield brandy. The operation must be conducted with great care, so as to keep in the distillate aromatic products pleasant to taste and smell and eliminate those that can reduce the quality of the brandy. The head and tail products are collected separately, usually added to the following brouillis or sometimes subjected to special distillation. One only retains as good tasting alcohol the “heart” of the operation. In general, the bad tastes of the tail appear, when the degree falls at about 40° and sometimes even at 50°.
In order to save time and fuel, the heat generated by the condensation of the alcoholic vapors with the wine is often used, which will be introduced into the boiler for the following distillation. For this purpose, there is placed between the boiler and the condenser a wine pre-heater, containing the same capacity as the boiler, which receives the fermented must and which is crossed by a single pipe or having a small number of turns. The condensed alcoholic vapors are directed to the condensing coil.
The distillation heads are rich in aldehydes, esters and higher alcohols: the tails still contain acids, esters, furfurol and a few higher alcohols. [The higher alcohols are in the heads either because they were stuck in the condensor and left over from the last run or because the proof is very low (first stage of double distillation). As the proof increases, their volatility will change and they will migrate to the tails.]
Baudoin (1) gives the following composition of the various fractions obtained in the distillation of Charentais wines:
(1) J. Pharm. Chim. (6) XXI. 449. 1905
Rocques, by submitting to distillation, by the method charentaise, an artificial wine consisting of a mixture of pure alcohol, acetic acid, and acetaldehyde. The following are the results of the esterification of acetic ester, amyl alcohol and furfurol (in grams per hectolitre of alcohol at 100°): [This is exciting data and essentially describes aroma compounds created in the still.]
These figures show that, except for acetic acid, which is rapidly removed as the higher alcohol content is distilled, the various volatiles pass almost completely through distillation.
According to Ordonneau (2), the quantity of esters of the distillate would be all the greater as the distillation is slower, the prolonged boiling of the wine favoring the esterification of the volatile acids. The volatile acidity and the total acidity of the wine would thus have a great influence on the ester content of the eau-de-vie.
(2) Prac. 7. Int. Cong. Appl. Chem. London 1909.
Graham (1) concludes tests with batch distillation units that the esters and aldehydes concentrate in the first fractions of the distillate: the former decrease more rapidly as the operation continues. As for the volatile acids, they are not separated in a clear way during the distillation. The duration of this has a great influence on the respective composition of the various fractions of the distillate. When the distillation is rapid, the alcoholic richness of the first portions drops very slowly, but the volatile acid content remains relatively high. When done slowly, the first fractions have a lower ester content and the volatile acids tend to accumulate in the tails. According to the author, the volatile acidity of wines seems to have little effect on the richness of esters.
(1) Australian Brewing and Wine J. LVIII, N° 6, 40 ; N° 7, 31 ; N° 8, 26. 1940. [I recovered and digitized this whole series of student projects many years ago.]
Batch distillation by re-passing gives high quality eaux-de-vie. The judicious fractionation of the ethereal parts of the head and the empyreumatic tail oils, separated at the beginning and at the end of each heating, allows aromatic products of value to be collected at their maximum of finesse and delicacy. Operating slowly, the heating ensures the extraction of certain aromatic species with high boiling point (cognac oil, rum oil, etc.). (2) Finally, the duration of the operations gives the cooked and mellow character, so appreciated by connoisseurs. However, in order to obtain quality eaux-de-vie, it is important, firstly, that the fermented must is itself well-constituted and does not contain products that are in bad taste, requiring too much rectification, and that the operation is conducted by an informed distiller. The “brûleur” is mainly based on tasting to perform the fractioning. The way of pacing the heating must be slow and steady, the tasting talent of the operator influence much more than the still model used on the quality of the eau-de-vie obtained.
(2) Cognac eaux-de-vie, obtained by re-passing, contain a very aromatic oil that is not found in California wine spirits, produced using continuous stills. (Valaer).
[This seems to be in Kervegant’s personal language and almost like fatherly advice to new young impatient distillers.]
This distillation method also has the disadvantage of being slow and expensive. In the Charentes, the first distillation lasts from 8 to 10 hours, and the rectification from 16 to 18 hours. A 10-liter capacity still will only supply 200 liters of brandy per 24 hours with an 8° wine. The expenditure of fuel is high. Mariller calculated that for the production of a 25 ° brouillis, the quantity of heat theoretically required reaches 270,895 calories, corresponding to about 90 kgs of coal per hl, of pure alcohol. By counting fraction recycling, the production of a 68° eau-de-vie requires 150 to 180 kg of coal per hectolitre of pure alcohol. In practice, the above figures are to be increased by around 20%. By the use of a wine pre-heater, it is possible to achieve a saving of about 20 kg of coal per hl. pure alcohol.
[This information is amazing and we should try to do more with it to create modern price estimates for fine spirits produced from natural gas.]
Also distillation by re-passing, which was the only practiced until the beginning of the eighteenth century, is no longer used today for the manufacture of some eaux-de-vie, such as cognac in France (3) and cider brandy (applejack) in the United States. It has been almost completely abandoned in rhummerie, where it could however be of interest for the preparation of certain types of fine rums (de vesou or syrup) or cutting rums.
(3) The decree of 29 June 1937 stipulates that only the controlled appellation of “Cognac” may be entitled to spirits produced at an alcoholic strength of not more than 72° C. at a temperature of 15° C and obtained by means of re-passing stills or stills said as first jet with successive loads, except for continuous feeders. [SOS This is a bit of a translation nightmare and I don’t know enough context to sort it out.]
[This information needs some attention because Kervegant is somewhat writing in the commodity era. Cognac may have reverted but Applejack is far less likely to be pot stilled.]
However, in Haiti, distillation would still be carried out in two stages, according to Pairault. In the first operation, an alcoholic liquid with an average concentration of 50° GL is obtained. This product, which is called tafia, is redistilled and brought to 60° GL. This is done even when continuous stills are used. Quite often, the two distillations are made by different manufacturers: the “guildivier” produces the tafia and sells it to the “distiller”, who rectifies it to obtain clairin at 60°, which after aging becomes rum.
(3) The decree of 29 June 1937 stipulates that only the controlled appellation of “Cognac” may be entitled to spirits produced at an alcoholic strength of not more than 72° C. at a temperature of 15° C and obtained by means of re-passing stills or stills said as first jet with successive loads, except for continuous feeders. [SOS This is a bit of a translation nightmare and I don’t know enough context to sort it out.]
Batch distillation without fraction recycling.
In order to obtain a marketable spirit at first, it is important to subject the alcohol vapors from the boiler to partial condensation, which increases their alcohol content. This condensation can be achieved by passing the vapors in a condenser cooled by a stream of cold water, or in a retort containing an alcoholic liquid. [What he is describing here is common batch column still reflux. What he calls a bubbler, I translate as retort.]
In the first case, the most aqueous vapors, in contact with the cold wall, condense and retrogress in the boiler, while non condensed alcoholic vapors are enriched. The condenser is therefore an analyzer, which gives rise to a steam richer in alcohol and a less rich retrogradation than the latter (1). At the same time, there is a purification of the alcoholic vapors, by the condensation of impurities with a high boiling point. [Cryptic language but one important purification to remember is the transition of higher alcohols from the front to the rear of the distilling run.]
(1) It is at least so with low-concentration alcoholic vapors. In the case of high-concentration alcohols, the condensed and uncondensed portions have an almost identical concentration of alcohol and impurities (Barbet, Sorel). The effect of condensation, which is sensitive to low and medium concentrations, is almost nil at high concentrations.
To obtain satisfactory results, it is important, according to Sorel, to circulate the cooling water and the vapors in the opposite direction, to increase the contact surface, so as to reduce as much as possible the quantity of the coolant, and to arrange the apparatus so that the reflux has a measure and a temperature as close as possible to those of the liquid of the boiler. [SOS not sure if I translated the word measure correctly from “titre”.]
The importance of retrogradation and, as a result, the amount of heat to be transferred to the condenser increases as the wine runs out … If it is desired to obtain at the end of the distillation a relatively high alcohol, the expenditure of energy becomes so great that the practice considers it to be inadmissible. It is preferred to stop the distillation of alcohol and complete the operation by collecting a poor alcoholic liquor (small waters), which is passed back into the next distillation. The bad tastes of the tails also begin to appear when the alcoholic measure falls to around 40, forcing a separation from good tasting alcohol to what distills below this degree. To hinder the end of distillation, the flow of water to the condenser is reduced or eliminated.
In the retort apparatus, the alcohol vapors leaving the boiler go into a container containing the small waters of the preceding operation, where they condense by bubbling. Those which escape concentration are significantly enriched in alcohol and stripped of some of the high-boiling volatile impurities. On the other hand, being richer in alcohol than that of the boiler, the liquid in the retort is soon brought to the boil and vaporizes in turn, giving alcoholic vapors richer than those of the boiler. This device plays the same role as the condenser. [I think when he says condenser, he may not means it as the final condenser so much as a brandy ball or reflux coil. When I say retort, he is saying bubbler.]
Condensations in the retort do not fall back, usually at least into the boiler. It is not possible, therefore, to obtain in one single run a high alcohol strength on average. We have to stop distilling the brandy, when there is still in the boiler and retort a relatively large proportion of alcohol, and collect it apart, to be reprocessed later, as large quantities of small waters. Condenser and retort are associated in some models of stills. [What I think this means is that after you make your cut and turn off the still, there can still be significant volumes of spirit in the column or retort. You don’t want to drain them back into the pot so you drain them back as part of your tails fraction. I never really considered this because I’ve only worked with fairly small columns.]
Batch distillation without re-passing does not allow obtaining fractioning as neat as the method by re-passing. Some high-boiling aromatic products are also stopped before reaching the refrigerant, or only in small proportions. With must of very aromatic bouquets not presenting unpleasant tastes, it gives consequently products of less good quality than re-passing. However, if the wine is low in aroma or if it requires extensive rectification, it allows for finer eaux-de-vie.
Ordonneau (2) writes on this subject: “The many distilleries that have been created for some time in the Cognac area use almost all the ordinary alembics, because neither of the one pass stills, nor those with trays, nor those of semi-continuous operation have been able to provide so far, with a quality wine from the Cognac region, an eau-de-vie equal in aroma and finesse to that produced by the ordinary alembic. The stills with multiple bubble trays are completely discarded, and this because the Cognac eau-de-vie contains odorous bodies whose boiling point exceeds 300° and which can not reach the coil, because they are eliminated by the bubbling and return to the vinasse, so that these stills, suppressing a part of the tail aroma, provide a drier and shorter eau-de-vie, that is to say whose perfume is less persistent, and this is all the more sensitive as the still provides a higher alcohol concentatrion… The continuous apparatus made a short appearance in the Charentes, and that on a trial basis; it has been realized, by examining the products they supply with the best wines of the country, that they are excellent for extracting alcohol quickly and economically, but bad for removing the aromatic part, which is the most valuable. By using these devices, we obtain a vinous eau-de-vie, so to speak rectified and without vital essence, reminiscent of Gers brandy”. [SOS this passage is extraordinary enough that it deserves the attention of a better translator. I think here we are talking bubble trays and not retorts.]
(2) Alcools et eaux-de-vie. Paris, 1885.
And the author advises, in conclusion, to choose simple stills with re-passing for very aromatic wines that do not have bad tastes or unpleasant terroirs, the one pass appliances for the wines with little bouquet or having more terroir, and then plates and bubble trays for wines with strong terroir or diseased wines, that is to say those whose eaux-de-vie must be energetically deflegmated. [This is a different use of the terroir word and probably represents ordinary winemaking flaws rather than a sense of place.]
Note that many models of stills can be used at will for distillation without re-passing or with re-passing; it is sufficient, in the latter case, to stop the arrival of water to the condenser. [remember, this condenser is more like a dephlegmator]
Regarding the heating mode of the still, fire or steam (by barbotage or steam coil), Rocques writes: “The old practitioners claim that the first mode is much better than the second, because, they say, the wine undergoes a greater cooking. It seems, theoretically, that there can be no difference between the two heating modes. It should be noted, however, that furfurol, which exists in large quantities in fine champagne spirits, occurs mainly by boiling with naked fire. Why would cooking limit its role and produce other chemicals that could influence the aroma of distilled products? We can not say that and we must wisely limit ourselves to recording the opinions of experienced distillers”.
We give below, by way of example, the composition of the products of head, heart and tail obtained by distilling a must of cane juice in a batch still of the P. Labat type (according to Zizine):
The duration of the distillation, with reflux stills, generally varies from 5 hours to 10 hours, depending on the type of apparatus used. The fuel costs are a little lower than in the case of appliances operating by re-passing: according to Mariller, they would reach for a spirits of 68° theoretically using 128 kgs and almost 150 – 170 kgs of coal per hl. pure alcohol. The quantity can be reduced by about 20 kg, by the addition of a wine pre-heater.
The intermittent distillation without re-passing, almost abandoned in the French colonies, is still very much used in the English colonies of Jamaica and Demerara.
Continuous distillation, or methodical distillation, derives from the principle which we have just indicated for intermittent bubble tray and condenser apparatus. The vapors emitted by the boiler condense partially by bubbling in a richer liquid in alcohol, contained in a tray; the vapors from the latter go in their turn into a second tray, placed above the first and contain a liquid even richer in alcohol than the previous one, and so on.
The successive vases, or trays, are placed one above the other, so as to form a vertical column. The wine to be distilled is introduced to the upper part of it and then goes down from plateau to plateau, to the boiler, where it arrives exhausted. The alcoholic vapors circulate inversely, progressively enriching themselves, by successive condensations and evaporation, and by stripping themselves of certain impurities with a high boiling point.
Instead of dividing the vapor in the liquid (sparge columns), the liquid can instead be divided into the vapor-streaming columns. In this case, the wine circulates in thin layers on very large surfaces of baffle plates, porcelain balls, Raschig rings, etc …), contacted by the vapors, and the partitions of the column are removed.
Continuous appliances are much inferior to alambic stills in splitting impurities. If the rectification is weak, together with the desired perfumes, impurities are collected which, because of their nauseous or austere odor and because of their disagreeable taste, depreciate the eau-de-vie. When one wants, by means of a more thorough purification, entirely eliminate the undesirable products, one loses at the same time a part of the bouquets and aromas which make the quality of the eau-de-vie. At the same time, since the wine only stays in the column for a short time (10-15 minutes instead of several hours in the still), the production of the esters by the chemical route is very small, while the extraction of the oils and high-boiling esters, which contribute significantly to the formation of natural bunches, are poorly performed.
Finally, continuous distillation has less flexibility than batch distillation. One cannot, as in the case of the latter, vary the quality of the eau-de-vie, by modifying the speed of the distillation, the proportions of the products of head and tail, etc.
With regard to the particular case of rums, Arroyo concludes comparative tests which he has carried out, with the clear superiority of batch distillation. The rums thus obtained have an aroma of more accentuated origin, more persistence, either after a long exposure to the air or by the treatment with the sulfuric acid, finally a flavor superior to that of the rums distilled using continuous devices.
[Exposure to the air reference Arroyo’s test of puting roughly 1.0 ml in a tasting glass with a watch glass top to let it slowly evaporate and fill the head space for evaluation. Sulfuric acid references another Arroyo test where sulfuric acid is used to rapidly oxidize and destroy all the ordinary congeners leaving only the rum oil if it is present. I don’t like the latter test and find it a challenge because a lot of brimstone aroma can also be inexplicably formed.]
Some of the disadvantages mentioned above are being tackled by extending the heating of the wine, by the use of powerful boilers with open fire and by high-capacity wine boilers; on the other hand, by using low-grade columns, producing alcohol at 55-60° C (1) and having few or no concentration trays, which block the road for heavy esters. Often the diameter of the columns is exaggerated and the number of trays is reduced.
(1) Certain observations made in Martinique indicate that the alcoholic strength which makes it possible to obtain with the continuous columns the most aromatic rums would be 57 ° G.L.
Despite the devices, the eaux-de-vie obtained have, however, significant differences with those from discontinuous stills, detectable not only to taste, but also to chemical analysis. By way of example, the composition (in grams per 100 liters of alcohol at 100°) of two rums prepared from the same molasses must, the first with a continuous apparatus, the second by means of a batch still of the Privat and Roussel type (according to Bonis):
The continuous apparatus gave more aldehydes and volatile acids than the intermittent apparatus, which supplied more higher alcohols and furfurol.
Harrison found in Demerara rums obtained using continuous apparatus an average of 44.9 gr. of esters and 18.9 gr. of acids per 100 liters of pure alcohol, and in those from batch appliances (pot still type) 66.9 gr. of esters and 33.1 gr. acid.
Arroyo, in Puerto Rico, obtained the following results by distilling the same must of cane molasses, on the one hand in a continuous apparatus (A), and on the other hand in a batch apparatus (B), both constructed by the Lummus Company of New York. Must from 3 different fermentations was successively experimented.
Rums from the discontinuous apparatus are poorer in higher alcohols, the richest in aldehydes and especially in esters. The latter, as shown by a fractional distillation test with the birectifier, belong for the most part to the group of high molecular aldehydes and esters, while in the case of rums supplied by the continuous apparatus, they are for the most part low molecular weight and low boiling point. This explains why the rums of the intermittent apparatus are more full-bodied with a more pronounced and more persistent aroma, less modified by sulfuric acid, than those of the continuous apparatus.
[Sulfuric acid again refers to Arroyo’s classic test.]
The differences between the products of batch stills and continuous appliances are all the greater as the alcoholic strength is higher. If one can obtain with the alambic stills very high-potency eau-de-vie, as for example the grand arômae rum of Jamaica and the “cœur de chauffe” of Martinique, which are collected at 80° – 82° GL, it is not the same in the case of the columns, where the coefficient of impurities lowers rapidly when the alcoholic strength exceeds 70°. This explains why the legislation of certain countries provides for a maximum distillation proof, which must not be exceeded. In the United States, for example, New England rums must be distilled at less than 80° G.L. In France, rums grading more than 65° are considered cane alcohols.
Fuel costs are much lower for continuous equipment than for alambic stills. Mariller found, for a short French column, fed with a wine at 7° or 8 and producing alcohol at 65 – 70°, an expenditure of 15.2 kgs of steam (at 1 kg pressure) per hl. of wine (or 194.7 kg per hl of alcohol at 100 °), corresponding to about 36 kg of coal. In Martinique, a consumption of 210 to 240 kg of steam per hectolitre is generally reported for rum at 60°
The continuous distillation apparatus gives, in a small space, daily efficiencies which considerably exceed those of intermittent appliances. Columns in use in the French West Indies, for example, have a production that varies, depending on the model, between 2,000 and 18,000 liters of rum at 60-65° per 24 hours.
They no longer require constant supervision that requires to be entrusted to conscientious men, knowing their trade thoroughly. The quality of the eau-de-vie no longer depends, as in the case of intermittent appliances, on the value or good will of the worker.
[May scientific pursuits in distillation revolved somewhat upon labor. Pushing ferments to higher degrees of alcohol concentration was as much about lasting through a workers strike as it was about efficiency.]
Continuous appliances are therefore remarkably suitable for large-scale industry and tend to replace discontinuous stills almost everywhere. It should be noted, however, that some distilleries of Demerara, producing annually more than 500,000 liters of rums, are still equipped only with these. In the industrial rhummeries installed at Saint-Pierre (Martinique) before 1902, intermittent apparatuses and columns were generally used concurrently, the former giving more odorous products, which were mixed with those of the continuous apparatus in order to obtain certain marks of rum.
[What is not to commonly known is the extent this system is still practiced in Jamaica with most all rums being a blend of grand arôme and rum from continuous stills.]
Recently, Arroyo has insisted for the distilleries to show interest, of a certain importance, by this combination of intermittent and continuous appliances, which would by the execution of judicious cuts, to have rums of much higher quality.
Distillation under reduced pressure.
Reich (1), for example, proposed a few years ago to reduce steam expenditure and to recover by concentration the fertilizers contained in the vinasse, distillation by means of evaporation chambers with triple or quadruple stages. In the first, there is evaporation, which exhausts the wine completely in alcohol. The alcoholic vapors are directed into the 2nd stage, or they condense, giving a brouillis at about 30°, which is directed on an ordinary distillation column. The last traces of alcohol condense in the 3rd stage, giving small waters to 1-1.5% of alcohol, which are re-joined to the wine.
(1) Chem. Metall. Eng. XLIV, 131, 1938.
Distillation under reduced pressure has the disadvantage of greatly reducing the proportion of aromatic principles contained in the distillate. Barbet (2), for example, obtained the following results by treating the same must in a same apparatus, firstly at atmospheric pressure and, on the other hand, under partial vacuum:
(2) Bull. Ass. Chim. LI, 437, 1934.
Distillation to clear must and cloudy must.
Should we distill the must in the presence of yeasts that have formed during fermentation (lees) or after clarification or filtration? This question is quite discussed.
In the Charentes, healthy wines are always distilled in the presence of lees. It is admitted that these give the eau-de-vie its mellowness and its “fat”. On the other hand, the distillation of concentrated lees produces products that are probably lacking in finesse, but which are extremely fragrant and have a lot of couverture, that is enough to perfume large quantities of neutral alcohol. “This experimental fact proves,” writes Barbet (1), “that the yeast cell does not easily diffuse into the wines the aromatic secretions it produces. It must boil, destroying the cell, to put the perfumes in freedom.”
(1) C. R. 4 Cong. Int. Chim. Appl. II, 253, 1903.
Pacot and Guittonneau have carried out numerous distillation tests, adding more and more of lees to light wines. They came to the conclusion that by increasing the proportions of the lees, the spirits of clear wines were dry in the palate and of light bouquet but gradually become fatter with a very complex bouquet, varying according to the yeast races used. However, when the amount of yeast reaches 10-20%, the aroma of the product is nauseating. This aroma is also unpleasant, if the lees come from sick wines (pricked, turned, etc.) or wine with too much character (wines of foxed grape varieties). In this case, there is interest in clarifying the wine.
According to Büttner (2), wines distilled in the presence of yeasts give more aromatic eaux-de-vie, but less fine than light musts. On the other hand, the amount of heart alcohol obtained is less, the bad tastes of tail appearing more quickly. [This is the classic rule of thumb I’ve heard most often.]
(2) Z. Unters. Lebensen. LXIX, 463, 1935.
Arroyo considers that 75% of the bad tastes and odors that exist in the newly distilled rums come from the decomposition by heat of nitrogenous organic matter suspended in the must. And as well, the troubled musts have the disadvantage of dirtying the distillation apparatus, requiring frequent cleaning, the author strongly advises to clarify the wine before distillation, by decantation, by filtration or better by centrifugation. We can leave the fermented must placed in special containers with lids to prevent loss of alcohol by evaporation. Then, we decant, the liquid. But the most expeditious process is to pass the fermented must in a centrifuge, Alfa Laval type for example, which ensures the removal of 99 to 100% of yeasts suspended in the liquid. This operation has, moreover, the advantage of eliminating the gases in suspension or in emulsion in the must and which can pass into the distillate, imparting a unpleasant smell in certain cases (presence of H2S, NH3 …).
[Alfa Laval is a brand and the type of centrifuge is often called a Westphalia type]
According to Arroyo, this clarification of the wort has the advantage not only of allowing more regular operation of the distillatory apparatus, of improving the yield of the latter and of reducing cleaning expenses, but also of providing a finer rum without bad tastes (3), likely to mature much more quickly.
(3) As early as the end of the 18th century, Charpentier de Cossigny already strongly advised, in order to prevent rums from becoming burnt, to pass the musts through a sieve for the first time before sending them to the fermentation tanks and a second time immediately before distilling them.
Practically, in the rhummerie, the wines are generally distilled after a decantation of short duration in the same tanks of fermentation, and the bottoms of tank thrown in the river. Exceptionally, in the method of Melle for example, the fermented must is clarified before distillation. [Not good environmental stewardship!]
It could, however, be interesting in some cases distilling the lees when one wishes to get a very aromatic rum for blending. But it is important to take certain precautions. If a batch still is used, it must be avoided that the organic materials do not become attached to the bottom of the boiler and caramelize, giving the brandy a burnt taste and too much in furfurol. To overcome this disadvantage, preferably use steam heating. If operating with a continuous column, choose a model of apparatus suitable for the distillation of turbid substances. Finally, avoid distilling the lees of the musts having undergone a faulty fermentation.
[The birectifier isolates furfurol in the later fractions and you can see that it is very stale tasting.]
[In many instances of this translation I’ve changed discontinuous to batch to make things clearer.]
The still originally used for the distillation of rum was simply a boiler connected to a coil placed in a barrel containing cold water. It was usually done by the method of re-passing. Towards 1800, an attempt was made to obtain a fractionation of the products of the distillation and a commercial spirit in the first pass, by bubbling the vapors into an intermediate receptacle, according to the principle of the Woulfe bottle. Devices of this type are still used today (pot still of Jamaica). Subsequently, various models of rectifiers or dephlegmators (lentille de Pistorius, tubular condenser, etc.) were imagined.
Labat describes the distillatory apparatus used in the French West Indies at the end of the 17th century as follows:
“The boilers are red copper, about two and a half feet in diameter by four feet in height. Their bottom is flat, it is pierced next to an opening in which a hose is soldered with a tap or valve which serves to empty the liquor which remains after the spirits have been extracted. The top of the boiler is in a dome, with a round opening a foot in diameter and a rim about two inches in height. It is by this opening that the boiler is charged, that is to say that it is filled with the liquor which has fermented in the vats. It is mounted on a masonry stove whose mouth is inside the building and the vent that gives passage to the smoke is outside. Masonry encloses the boiler, up to a third of the height.”
“When the boiler is full, close its opening with a red copper capital which must fit tightly just in the rim of the top of the boiler and lute with clay; it is good that it is tinned so as not to be subject to the verte de gris. It has a beak eighteen to twenty inches long, which is brought into the extremity of a coil of copper or tin, which is placed in a purpose made barrel, well hooped with iron, placed near the boiler. The more the coil has convolutions, the more the brandy is good.”
“The barrel where the coil is always has to be filled with water, to refresh it, because the spirits that the heat did raise from the boiler in the still head, circulating in the coil where they were led by the beak of the big top that is joined and well luted. They warm the water up extraordinarily and dissipate through the pores of the metal, if they are not soaked by the coldness of the water. That is why it is good that there is always new water in the barrel, which must flow through a hole left in the bottom so much proportion to the quantity that falls there that it always remains full. We put a refinery pot of or a large jug at the end of the coil to receive the liquor that comes out. When you notice that the fire is no longer raising spirits and that it does not coalesce anything in the pitcher, you empty the boiler by the tap at the bottom and fill it with new liquor.” [Its hard to translate this old language.]
“The first liquor that comes from a boiler is called small water; indeed, it does not have much strength. We keep everything we get from small water during the first five days of the week, and then fill one or two boilers for re-passing. The spirit that comes out of it is really eau-d-vie, taffia or guildive, which is very strong and very violent. In the sugar factories where there are two boilers for eau-de-vie, one must make a week or six hundred pots or about measure of Paris”.
This apparatus, similar to the alambic used at the same time in France and as described for example by the abbot Rozier, in his “Complete Course of Agriculture” (Paris, 1786), was the only one used in the West Indies until the early nineteenth century, where it was replaced by the retort still.
However, in the eighteenth century, the size of the capital and the gooseneck was increased, so that brandy was obtained on the first pass. Ducourjoly describes the stills in the second half of the century as follows:
“The capacity of the stills most suitable for making good rum must be, according to experience, about 300 gallons, which is 1,200 pints, measured in Paris. They must be 4 and a half feet high; their bottom will be of a good thickness, as well as the parts which surround this bottom. The stills, moreover, will be roughly in line with those used for eau-de-vie, except that it is necessary to give them a little more thickness in totality. The collar of these stills will be about 16 inches high, so that the distillation is more prompt, and that the stillage does not boil over with the spirits.”
“The capital will be 3 times larger than those used in the French distilleries, proportionately kept from elsewhere; its shape will be a little more crushed. The collar of this capital will be about a foot in height, so that it can be easily and firmly adapted to the stills, and the beak, instead of being copper, according to custom, will be good tin, allied with a little copper, to give it a good consistency. We will adapt the beak of the capital to the top of this same capital, to facilitate the ascent of spirits; and this beak will be curved in the shape of a goose neck. The coil, which must be good tin, will be 3 inches and a half or 4 inches in diameter, and at least 6 large convolutions.”
This still resulted in 300 gallons of grappe becoming 80 to 84 gallons of rum grading 20 to 30 degrees below proof (46-40° GL) and 40-50 gallons of small water at 40-30° G.L. Distillation was cut off, for rum, when the measure of the distillate fell to 40° GL, and for the small water, when it fell to 30° GL. The small water was distilled separately and provided a spirit of 47° GL., and used to raise the level of rum, which was generally sold under 42-43° GL (exceptionally, 44°5 on the London market).
It was so at least in the English countries. “The capitals of our stills,” says Ducourjoly, speaking of the French colonies, “have too little capacity, and their mouths, or the snares by which they adapt to those of the stills, are too short; so that the vapors which are sublimated, in spite of the best kept fire, having not enough space to circulate, there is only a mediocre segregation of spirits of the aqueous parts of the grappe, mingled with these spirits, hence the bad qualities of tafia. The coils of our stills have neither enough matter nor enough convolutions; which is opposed to the goodness of the liquor. The English have, for a long time, felt the defects of our guildiveries; so they have perfected them, while we remain enslaved to the first ideas and the first habits on this interesting object.”
[SOS this is old French and needs a more skilled translator.]
In addition, French settlers mismanaged the separation of the tail products. “We must be very careful,” says the author, “to prevent the small water from mixing with the rum. The French, who do not do enough to distinguish exactly the time of the distillation where the rum ends and where the small water must begin, which also has a fixed point where it marks that one must stop, they spoil one and the other, and at the same time impoverish the stillage too much, that they are no longer suitable for forming good grappes. It is mainly from this mixture of rum with the small water that our tafia results. ” [I’ve never seen it mentioned that you can spoil dunder by boiling too long.]
Charente alembics — The distillation by re-passing may be of interest for the preparation of certain types of rum, we will give some details on the still and its operation.
The apparatus consists of a copper boiler, often wider than it is tall, with a capacity usually ranging from 500 to 2,500 liters, placed on a brick stove. The boiler, which has a manhole for cleaning, is capped with a capital, sometimes rounded like a head (Moorish head apparatus), but more often extended by a “gooseneck” to avoid entraining foams with alcoholic vapors. The pipe that continues the gooseneck is terminated by a coil plunging into a tank of 4,500 liters of capacity (refrigerant), full of cold water that is constantly renewed from bottom to top during the distillation. Generally, between the boiler and the refrigerant, a tin-plated copper wine pre-heater, often in the form of an urn, is fitted with the same capacity as the boiler. To prevent the loss of alcohol, it prevents the wine from boiling, either by making the pipe turn a turn in the wine heater, or by introducing the wine only towards the end of distillation. The alcoholic vapors condensed in the wine heater are directed to the refrigerant. The still is usually heated by open fire, with wood or coal, rarely by a steam coil or by direct steam. The distilled liquid is collected in a special test tube, equipped with an alcohol meter, which then directs the liquid into a drum.
Distillation is conducted with great care. The attention of the worker-distiller, or “brûleur”, focuses on heating, which must be moderate and regular, so as to obtain a slow and uniform distillation. However, it is possible to push the fire a little at the beginning, until the moment when the wine boils, and towards the end of the distillation. The water of the refrigerant should not be too cold, so that the condensation of the alcoholic vapors is not too abrupt; it is sufficient that the inner third of the refrigerant is cold, the medium being warm and the upper part hot. The optimum pouring temperature is considered to be 18° C: at 20° C and above, there is loss of alcohol by evaporation, whereas below 15° C the resulting eau-de-vie is dried.
[This last comment is unique and I have not seen these numbers given by anyone else but Arroyo. I have never seen the sensory descriptor dried. We know fatty acids and other high boiling point congeners get stuck to the sides of condenser and being too cold likely exacerbates that effect.”
The first alcohol that passes measures 53°, and it is only after 9-16 hours of heating that a load of wine at 10° G.L. is exhausted. 1/3 of the original liquid was collected. The distillate (brouillis) contains 25-30% alcohol. The distiller then moderates the heat of the fire, covering it with wet charcoal; he removes his vinasse from the boiler and refills it with the liquid from the wine pre-heater.
There is nothing absolutely rigorous in the separation of head, heart and tail products. The proportion of these various elements is very variable, according to the quality of the wines and also the way in which the fire was paced (the more the distillation was slow and the more important the proportion of eau-de-vie of the hearts). [Remember Cognac wines have to wait in a queue to be distilled so wines at the back of the line see more risk of infection.]
There are, in fact, slight differences between subregions producing cognac. “In the Grande Champagne,” Nottin points out, “(1) the quality of the wine is so good for the production of a perfect brandy that the stills have to do the least reflux possible. The boiler is capped with a capital, either rounded or frustoconical, from which a tube is slightly inclined towards the refrigerant. In other regions, the capital is extended to the upper part by a gooseneck, in which a retrogradation occurs; with this device, the classic type of the Charentais still, we do some reflux, as well to the distillation of the brouillis as to the bonne chauffe… In the wines with terroir, we even push the rectification further, using either lenticular rectifiers, either spherical rectifiers or even small flat columns surmounting the boiler; we thus eliminate the taste of the terroir”. [Keep in mind this is a different meaning of terroir that may be ordinary wine making flaws and problems like tufo.]
(1) Bull. Ass, Chim. XLVII, 405, 1939.
In Armagnac, they proceed differently. The distillation of the hearts eau-de-vie is continued until the degree falls to 24 and even 20°; the mixture marks 52°. The head fractions are added to the wine, while the tails are used in the care of the casks. [I’m not sure exactly what care of the casks implies, but there are many options.]
Some manufacturers of distillatory equipment still manufacture simple stills, specially adapted for the distillation of rum. The Deroy House, in France, for example, is building a model with an elephant trunk and another with a retort cap. Pontifex and Wood, in England, also make a still boiler still with a retort cap. But these devices seem to be used today only in a very exceptional way. [It is hard to say what the retort cap implies exactly.]
Retort Alambics. [Alambics à barboteurs.]
Jamaican still — Jamaica pot still consists of a flat copper boiler with a capacity of 1,000-2,000 gallons (4,500-9,000 liters) heated by open fire or by steam coil. It is surmounted by a huge marquee in the shape of a cornue. From the top of this last part a large pipe, about 20 cm in diameter, which goes to the bottom of a copper cylinder, the bubbler or retort (cornue), containing alcoholic water from a previous operation. The cylinder, which has a capacity equal to a quarter of that of the boiler, is equipped at its upper part with a dropping funnel, for the introduction of small water, and at its base with a discharge valve for the vinasses. The first retort is often followed by a second, in all respects similar, from which the alcoholic vapors go to the coil condenser, sometimes after passing through a wine pre-heater. [cornue apparently translates to retort but I’m not sure about “huge marquee in the shape of a cornue” if that is similar enough to use the word retort. Everything is depicted in the photo…]
Cylinders are charged, the first with small waters at 10-15% (low wines), and the second with alcohol at 60-70% (high wines), so that the tube of supply of alcoholic vapors plunges into the liquid. When there is only one retort, the high wines are introduced into the boiler with the wine to be distilled, and the retort loaded with the low wines. At the beginning of the campaign, where no small water is available, ordinary water is used.
When the liquid in the boiler begins to boil, one is warned by the characteristic rumble that occurs in the retorts. The fire is reduced a little bit. The first liquid that passes is set aside, to be mixed with small water: it represents about 1 per 1,000 of the volume of wine treated, or 1 gallon for a load of 1,000 gallons. The tasteful liquor that flows then has a high degree (88°); it is collected until the mixture marks about 80°. The tail products are fractionated, so as to obtain on the one hand high wines at 60-70°, and on the other hand low wines at 10-15°. The proportions of these products per 100 marketable rum are on average 55-60% high wines and 150-175% low wines.
[Fascinating, Kervegant describes a demisting test of only 1 gallon for a heads cut and then two tales cuts! The last being incredibly large relative to the hearts fraction. No wonder they can fill 250 gallon retorts.]
This still, which is an application of the apparatus with flasks invented by Clauber in 1648, was already widespread in the English colonies during the first half of the nineteenth century. Soleau (1) wrote, in 1835, in a travelogue in English Guiana:
(1) Ann Marit. Col. 1835, t. 2, 39.
“Above all, there was one thing I wanted to observe, it was the method of distilling the English, which gives their rum a quality so superior to ours. This is what I noticed in the distillery. The stills they use are very simple and are not nearly as perfect as the Derosne stills. Theirs consist of 2 retorts, into which the grappe is introduced and which are exposed to fire; the result of the distillation of these two retorts arrives in a third, or a condensation occurs and gives rise to a new distillation to increase probably the degree of the rum that comes, after having crossed a coil plunging into the cold water, it flows into the vessel destined to receive it. This method of distilling, very imperfect, could not affect the quality of the rum. There was only one thing to remark in this distillery, that the liquid of the beginning, the middle, and the end of the distillation was collected separately. The middle rum of the operation is the only saleable rum; that of the beginning and the end are united and form a rum of inferior quality, which is consumed in the country. ”
[This is one mans observation and one reason it is flawed is it does not recognize the qualities of their ferment that can justify the distillation. Secondly, the practice of selling inferior rum (probably horribly tainted by tufo) eventually went away and greater fraction recycling was practiced.]
Wray, in 1848, made the following statement: “But of all these different provisions of the distillatory apparatus, I have never known one that outweighed the ordinary double retort-pot still. I point it out as unparalleled as a distillatory device, particularly suited to the needs of planters, from the point of view of simplicity, solidity and economy. ”
These qualities, combined with the excellence of the products obtained, have made the double-retort still in Jamaica today, and several English houses, including Blairs, Campbell and McLean of Glasgow, Fawcett, Preston and Liverpool Co., continue to manufacture it.
The apparatus was introduced into the French West Indies around 1895. Lavollée (1) reports that in 1838 the still used in Martinique consisted of two retorts heated by open fire; the alcoholic vapors passed into a third retort before reaching the refrigerant. The daily output of the apparatus was 100-120 gallons of taffia per day.
(1) Notes sur les cultures et la production de la Martinique et de la Guadeloupe. Paris, 1841.
Alambic P. Labat — Modifications were made, however, to the still, which, towards the end of the century, was known in Martinique as the P. Labat apparatus.
The boiler, fire heated or rarely by direct steam, and of a capacity not exceeding 1,500 liters, generally preserved its primitive silhouette, with its huge retaining cap. But we replaced the “retorts” in copper with a closed frustoconical wooden tank, equipped at 15 cm. from the bottom, an overflow tubing allowing the demotion of condensed spirit towards the boiler. The retort was filled up to the height of the overflow with the small waters of the previous distillation. Often, between the retort and the refrigerant, was arranged a wine pre-heater, wooden tub of the same capacity as the boiler. Finally, in some cases, the apparatus was completed by the addition of a rectifier, a copper cylinder cooled by water runoff outside; the condensed vapors retrogressed towards the retort by a pipe situated at a third of the height of the receptacle. The marketable rum was collected until the alcoholic strength of the mixture was brought to 55 and even 50°, then the fire would be stopped to get the wine out.
The P. Labat still, which was widely used in Martinique towards the end of the last century, both in the grand industrial rhummeries, where it was used concurrently with continuous equipment, and in small agricultural distilleries, has completely disappeared today. It was replaced in the French West Indies by the apparatus Privat and Roussel, that we often continue to designate as the apparatus P. Labat (1).
(1) This term is virtually synonymous with a batch device in Martinique. It is scarcely necessary to point out that Father Labat never suspected the existence of the pot still attributed to him. [Labat, for the record, might have been brilliant, but was also a terrible person.]
Condenser stills. [column stills]
Alambic Shear. — Condenser stills or de-phlegmators have closely followed those with retorts. Wray describes the Corty apparatus, modified by Shear and Sons, and was very much used in his time in the West Indies.
This still used the condenser invented by Pistorius in 1817 for the rectification of alcoholic vapors. The vapors arrive in a box, where they strike a metal plate that forces them to divide along the walls of the box, cooled below by air and above by water. The condensed aqueous vapors return to the boiler, and the more concentrated alcohol passes into one or more similar boxes, where it concentrates more and more.
[Valaer remember was the super star IRS chemist who gave us the best surveys of spirits.]
In the simplified still of Shear, the cylindrical capital is surmounted by 3 boxes of Pistorius, cooled by water coming from the refrigerant by means of the tube E and traversing the outer surfaces of the boxes by distribution tubes C. A still containing 400 gallons allows to distill 4 or 5 loads per 12 hour day, giving a eau-de-vie at 78° G.L. [This refers back to the image two pictures up.]
In another device, the previous apparatus is added to an ordinary still with a retort cap. The alcoholic vapors emitted by the latter bubble into the liquid of B, whose contents are poured into the boiler, once the must has been exhausted of alcohol. Only one hearth heats the 2 alambics. These double boiler units were quite common in the English colonies, especially at Demerara, where there were some with a capacity of 1,200 gallons. They gave rise to a much lower fuel consumption than simple stills and allowed for an excellent rum.
These devices are no longer used today, at least in their primitive form. In Louisiana, however, a still is used which has a certain resemblance to the Shear apparatus. At the outlet of the boiler, heated by direct steam, the alcoholic vapors pass through a gooseneck in a Pistorius condenser with 3 or 4 boxes. The products of the condensation are directed backwards into the boiler. [Each box here is sort of a plate in a still column.]
Demerara vat stills. — The intermittent apparatus (vat still) employed in English Guiana is constituted as follows. The boiler is a wooden tank A, of variable capacity according to the importance of the installation (3,000 gallons in general), heated by direct steam: it arrives by pipe D which almost reaches the bottom of the tank. The alcoholic vapors are driven by a copper gooseneck into another tank, smaller in size, the bubbler (retort), which receives the small waters (low wines) of the previous operation. The retort is surmounted by a rectifier, consisting of a copper cylinder containing numerous small diameter tubes, between which circulates cold water. The condensations of the rectifier fall back into the retort, while the non-condensed vapors go to the coil condenser. The boiler and the retort are sometimes made of copper. [The diagram seems to be missing the letters that Kervegant refers to.]
The facilities usually include 2 vat stills, working alternately. A 3,000 gallon still runs 3 charges of 2,500 gallons, or one and a half tanks, per 15-hour day. [I wonder if they have to alternate because of the high demand for steam?]
The alcohol which flows at the beginning (below 83 ° G.L.) and at the end (below 77 °) of the operation is excepted: the products thus obtained constitute the low wines. The commercial rum collected between these two sections marks 87-89° G.L.; it is brought back for export at 80-81° by dilution with water. The cutting points vary depending on the composition of the must, so as to obtain a product of taste as constant as possible.
Alambic Privat and Roussel. — The still known under the name Privat and Roussel (Martinique) or Wendelken (1) (Guadeloupe) was, a few years ago, widely used in the agricultural distilleries of the French West Indies. Today, it has been almost replaced everywhere by continuous apparatus; it is only found in a few small rhummeries working with cane juice.
(1) The name of a manufacturer of distillatory devices, located in Saint-Pierre (Martinique).
This still consists of a copper cylindrical boiler of 1,500-3,000 liters capacity, heated by open fire and surmounted by a small column containing 2 to 4 trays. From the top of it a gooseneck goes to the serpentine. At 20-25 cm, from the top is arranged a circular gutter, which goes up flaring along the column, which it exceeds a little height. The gutter receives, throughout the duration of the distillation, a stream of water brought by a special pipe. The water is evacuated by 2 discharge tubes located at different levels, which makes it possible to vary the importance of the water tank constituted at the top of the column. A lateral tube with funnel and tap is used to introduce on the trays, to a desired height, the small waters of the previous operation. This height is sometimes adjusted by a level tube, which also serves as a pressure indicator during the operation of the device. Between the column and the refrigerant, is sometimes arranged a brise-mousse, small cylindrical box in which the particles of entrained liquids are separated from the alcoholic vapors, to roll back to the column. Finally, the device is sometimes supplemented by the addition of a wine pre-heater. [I have seen this use of water on the outside of the still (instead of a reflux coil inside) called a “brandy ball” in other spirits. I’m having trouble visualizing the pressure indicator, but the u-tube in the birectifier also indicates pressure which can be useful for pacing the still. I did not translate brise-mousse which I guess is simply a trap for entrained foam.]
The first alcohol that passes measures 50-60°, then the degree rises to 85-90°. In the manufacture of ordinary rum, head products are not separated and the brandy is collected until the degree falls to 40° and even sometimes to 30° (in the latter case, the product often has a bad taste): the mixture is 55-60°. Small waters collected then serve, mixed with the head products, to load the column during the next distillation. They are also in small quantities (10-15 liters at 25-30°, for 1,500-2,000 liters of must). At the end of the distillation, they are evacuated at the same time as the vinasse: they contain only traces of alcohol (0.4% according to Pairault). The duration of the heating is 8-10 hours. [Notice the tails fraction is very small because this is a column still distilling at a very high proof.]
In the manufacture of the type of rum in Martinique called cœur de chauffe and which is characterized by a delicate scent of genièvre, a more careful fractionation is carried out. After having separated the heads (1-2 liters), the spirits are collected as long as the alcoholic strength of the mixture is maintained at 80° G.L. : reduced to 60° with good spring water, the product is the “cœur de chauffe”. Only 20 to 25 liters of this product are removed per 2,000 liters of must. The liquid which then passes, until the average alcoholic measure falls to 55°, is delivered for consumption as an ordinary “grappe blanche”; the small waters are re-introduced into the column. [I left genièvre untranslated but it likely translates to juniper over genever as we see in the next paragraph and note.]
The skill of the distiller, which is based mainly on the tasting to effect the separation of the head and tail products, has a great importance in the manufacture of the “cœur de chauffe”. It is prepared exclusively from cane juice subjected to spontaneous fermentation, without the addition of sulfuric acid or vinasse (only a little Ammonium sulfate is added). Great importance is attached to the quality of the water used, which must be very pure, and to the meticulous cleanliness of the various appliances and vessels. Not all “grappes” are likely to give a good “cœur de chauffe”, but only an average of one in two (1).
(1) Sometimes a “cœur de chauffe” of lower quality is also prepared by adding berries and crushed juniper to the must before distillation (6 liters per 200 liters of must).
Deroy stills, Egrot, etc. — Most of the major French manufacturers fabricate batch appliances, equipped with rectifiers and suitable for the distillation of rum.
The Deroy house in Paris, for example, is produces a still with a hydraulic double point rectifier and a wine pre-heater, specially designed for making rums. The boiler, heated by open fire, has the shape of a wide cylinder and low height. The rectifying capital is placed as an ordinary lid and fits freely in the gutter of the upper rim of the boiler. Part of the overflow water of the refrigerant flows into the center of the capital, flows into the gutter and is a first hydraulic seal. A second internal seal is formed by the condensation of water vapors which, following the walls of the capital, fall into a second gutter and prevent the alcoholic vapors from condensing in the water of the first seal. Any loss of alcohol is thus stopped, and the closure perfectly hermetic. The degree of the eau-de-vie which one wishes to obtain is regulated by opening more or less a tap placed on the water supply pipe to the rectifying capital. When one wishes to operate by repassing, it is enough to close this tap. One can obtain with this apparatus in a single operation, with an insignificant production of small waters an eaux-de-vie varying at will from 50 to 70 ° GL. [Very clever with those seals! Don’t forget, re-passing here implies double distillation a la Cognac instead of single pass batch distillation.]
In another model of the Deroy house, the rectifying capital has, instead of the double hydraulic joint, a clamping joint: light levers, distributed around the joint create a compression on the circle of the capital, which comes to rest itself on a rubber band embedded in a groove formed at the upper part of the boiler. The capital is surmounted by a lenticular rectifier, which is a modification of the Pistorius condenser. This device consists essentially of a copper lens placed horizontally, inside which is located a vapor retarder. The outer surface is covered with a sheet which is always kept wet, by allowing a trickle of water to collect in a central collar. As it evaporates, it removes from the inside the lens a certain amount of heat, all the more significant as water runs in larger quantities. This rectifier, which can be applied to different models of batch stills, makes it possible to obtain eaux-de-vie at a higher degree (60-75°) and also to operate more quickly. [This “lens” and “vapor retarder” may again be another variation of the “brandy ball”.]
Let’s also mention the immediate re-passing capital from Deroy, that can be adapted to all kinds of stills and that allows one to carry out in a single operation the distillation and the re-passing of the eau-de-vie.
In the Egrot and Grangé still, the rectifier consists of two concentric hollow spheres. Inside the internal sphere comes a current of water, which flows out to the upper part, spreading on the surface of the outer sphere, which is covered with a large mesh. The alcoholic vapors circulate in the space left free between the two spheres and partially condense there, before passing in the wine pre-heater or the refrigerant. The power of the rectification is adjusted by circulating the cold water more or less rapidly. The device makes it possible to obtain rum at 60-65° on the first pass.
Although they give superior results to the previously reported stills, Deroy devices and Egrot, etc. are rarely used in the colony’s rhummeries.
The continuous distillation apparatus, which is derived from the multi-boiler still invented by Adam in 1801, appeared in rhummeries around 1840. But the first results obtained were not satisfactory. Wray wrote in 1843: “The stills of Blumenthal, Laugier, Coffey, although excellent and a very good service assuredly, are nevertheless much better suited to European distilleries than to the colonies. I have seen several, some with various modifications, working on plantations in India and in the settlements of the Malay Straits; I did not see any of them whose owner had reason to be happy, probably because they lacked the skillful and careful workers to make them work whom it is easy to obtain in Europe”.
We also read in the Lavollée Report on the productions of Martinique and Guadeloupe: “Several distillery machines from M.C. Derosne, in Paris, were tested at Basse-Terre and Pointe-à-Pitre, and this considerably improved the quality of the rums, but it was not enough to make it reach the English perfection; besides, these apparatuses require repairs which the workmen of the colonies are incompetent to execute, they will not be long in being put aside as expensive and impossible to preserve a good state”.
Thanks to the improvements made to the apparatuses, as well as to the improvement of the colonial labor, the columns, which present appreciable advantages as regards both the economy and the speed of the distillation, have, however, progressively replaced almost everywhere batch stills, for the production of rum.
From the point of view of interior construction, Mariller divides continuous distillation apparatus into: surface-heated appliances; direct steam apparatus; contact apparatus, comprising horizontal mechanical movement columns and trickle columns; solid columns. The direct steam columns, which give the best performance with the minimum of steam, are almost the only ones used in the rhummerie. However, there are some solid columns, including the Ilges (Argentina) and Collette (Indochina). There are many models of direct steam columns, among which the major French brands (Savalle, Barbet, Egrot, Fives-Lille, Crepelle-Fontain, Deroy) are the most widespread. The Coffey apparatus is also very common in the English colonies and the Bohm column in the United States. [The nomenclature here is tricky and I think I substituted direct steam for bubbler correctly. The solid columns may be like falling film evaporators which are common now in laboratory vacuum distillation?]
From the point of view of their operation, the continuous apparatuses can be classified in low degree columns and high degree columns. The first provide alcohols whose richness corresponds substantially to the degree of the vapors released by the wine which feeds them, according to Gröning’s table (generally 50-60 °): they are the most used in the French colonies producing rum. The second, very common in the English and Spanish countries, give dephlegmation at 80° and often 90-95°. They include, in addition to depletion trays, concentration trays, more or less large depending on the measure of the alcohol dired to obtain. [I think the difference between the two trays described here is that one is for separating the beer and the second is for after the beer has been separated.]
A low-strength column has the following general arrangement. Wine, contained in a tank A kept at a constant level, first crosses the wine pre-heater C, where its arrival is regulated by the tap r, and where it warms up at the expense of the alcoholic vapors. It then enters the column E, by a curved siphon pipe, of sufficient length that the internal pressure of the device does not force the liquid into the wine pre-heater. The wine descends from plateau to plateau, gradually becoming exhausted, and arrives in the boiler of the column in the state of vinasse, containing only traces of alcohol. The vinasse is evacuated by a siphon pipe.
The boiler is heated by direct steam; a regulator R allows one to regulate the admission thereof. The vapors produced rise in the column, enrich themselves more and more in alcohol, while contacting the wine which goes down, and arrives at the wine pre-heater, where they partly condense. The condensate liquid retrogrades in the column by a pipe and crosses some trays before coming to join the wine. The alcoholic vapors leaving the wine pre-heater are condensed in the refrigerant F, which receives the water from the tank H by a pipe provided with a control valve R, and the cooled eau-de-vie arrives at the test tube G. A thermometer T makes it possible to control the running of the column.
The arrangement of the high-strength column is similar. However, in this case, a series of trays is inserted between the depletion column and the condenser, which form the concentration or rectification column, generally placed at the top of the depletion column. Concentration trays receive the retrogradation of the condenser. The alcoholic vapors are enriched by bubbling in this liquor rich in alcohol, especially as the number of trays is large and the retrogradation is abundant.
Organs of a distillation column (1).
(1) According to Mariller, loc. cit.
The following components can be distinguished in a distillation column: trays, the heating device, the steam regulator, the wine cooler and the refrigerant, the vinasse heat exchanger, the test piece. [I think the test piece or l’éprouvette de controle is a spirits safe.]
Trays.—Modern appliances are generally made up of exhaustion vessels, or trays, placed one above the other, so as to form a vertical column. These trays of circular or sometimes rectangular form (Savalle, Golley) communicate with each other by overflow pipes. which establish on each of them a constant level, by pouring the excess of liquid on the lower plate. The thickness of the liquid layer is generally 4 to 5 cm.
The devices adopted for dividing the vapors in the liquid are numerous. Sometimes the trays are simply constituted by perforated plates of small holes (Coffey trays, trays of concentration), through which the steam passes and bubbles in the liquid, maintained by the pressure. Excellent with regard to the use of steam, this device has serious drawbacks from the practical point of view: the holes get clogged with troubled musts, gradually growing with acidic liquids, which create irregularities of operation. In addition, the column may be discharged at any time, when pressure drops as a result of changes in the power supply. [I think acidic liquids here eat the copper, expanding the holes and changing the kinetics of operation.]
Also, one prefers in general the plates with bubble caps. In these, the steam inlet ports are covered by a cap, which forms a hydraulic seal with the liquid on the plate; the vapor accumulates beneath the cap, and soon, by its pressure, represses the liquid in which it bubbles.
There may be only one central cap (Bohm plateau), sometimes with several star branches with serrated edges to ensure better contact of the steam with the liquid (plateau Champonnois). More often, however, the trays have multiple caps. These have smooth or serrated edges, full or pierced with clearance holes. When the teeth are rectangular, the caps are called comb caps and, when the holes have the same shape, fenestrated caps. The Creole column, widely used in rhummeries of the French West Indies, has trays with 5 caps, full of hemispherical and smooth edges. The Barbet trays carry a large number of equidistant small comb caps, with a great power of exhaustion. The Savalle rectangular plateau has two large rectangular openings, covered with roof-shaped caps.
The large diameter Egrot trays have a particular disposition, which makes it possible to use the wine methodically on a very small number of trays (4 or 5). The wine arrives by a discharge pipe on the periphery of the plateau and goes, in a series of concentric circles, towards the center, from where it flows through the overflow on the plateau below. After having traversed the first ring, it meets a vertical wall which compels it to make, in the second circle, a rotation in opposite direction of the first, and so on to the center. Throughout its journey, the wine receives steam through a large number of small caps, which ensures an intimate contact between steam and the alcoholic liquid.
We sometimes combine the plate with caps and the perforated plate: in the column Crépelle-Fontaine for example, the liquid is maintained at constant level by the usual overflow tube and the steam is brought by a big cap under a false perforated bottom. [notice we’re interchangeably saying plate and tray.]
To achieve perfect bubbling, it is necessary that the vapor is distributed in extremely fine bubbles in the liquid, and that this distribution is as regular as possible. When the tray only has one or two large caps, the steam escapes in a big plume, which turns into a huge bubble passing through the liquid without condensing. In the case of multiple caps, the bubbles are less large, but still remain voluminous if the edges of the caps are smooth. The caps pierced with holes give thin streams; however, when the holes are multiplied too much, the streams meet and reform large bubbles.
The best results are obtained with the serrated caps, especially when the teeth are long and thin. The steam is then divided into thin horizontal streams, which do not immediately rise and follow a fairly long course in the liquid; the boiling is done regularly and all the points of the plate are subjected to the same bubbling. Rectangular plates with multiple caps (Savalle type) also provide excellent bubbling, but star caps, like Champonnois, have the serious drawback of not regularly distributing steam, the intervals between the branches not subject to any splashing. Multiple caps and small dimensions have the disadvantage of getting clogged easily, with troubled musts (cane juice).
It is essential that the mounting of the trays is correct. If the column is not vertical, the trays are inclined and some caps are not submerged or almost not in the liquid, the will vapors pass without producing a useful effect.
The number of trays in the column varies according to the type of tray, the alcoholic richness of the wine, the degree of alcohol to obtain, etc. In the French West Indies, the appliances used in rhummerie (wine at 5-6°, rum at 55-65°) generally have 14 to 20 trays, sometimes up to 32 trays (large distilleries). The total surface of the trays varies, according to the type of apparatus, between 3.7 and 7 mq, for a production of 100 liters of rum per hour. In the columns used at the beginning of the century, the number of trays was generally 12-15. The trays are joined into sections: each of these usually comprises two trays, one fixed, the other moving between two sections, so as to facilitate cleaning.
The distillation columns are most often made of copper, a metal that is resistant to the acids that exist in musts and that is easily adapted to the construction of the most complicated trays. Coffey appliances are often made of wood. This has the disadvantage of absorbing a lot of alcohol, which evaporates to the outer surface of the wall.
Heating of the column. — The columns can be heated by direct fire or by steam. The heating with direct fire is hardly used in rhummeries other than for small apparatuses. However, it has the advantage of giving more fruity products than steam heating.
The Midi apparatus, still very much used in the south of France for the distillation of wine and which is the old Derosne and Cail column, a little modified, is heated by a boiler with two compartments superimposed. The apparatus designated in the English columns under the name of French column is closely related to the preceding one. [Two compartments? Sort of like a 3 chambered still?]
In the rhummeries existing in St-Pierre of Martinique, at the end of the last century, heating with direct fire was also the general rule. Most often, there were 2 (sometimes 3) large capacity boilers, the first forming the base of the column, the second having the shape of a vertical or horizontal cylinder connected to the first by a gooseneck. The vinasse leaving the last tray flowed into the first boiler, by means of a tube plunging to the bottom of it. It then went to the second boiler placed a little below and heated by direct fire. This arrangement allowed a prolonged boiling of the vinasses, intended to extract as much odorous products as possible. In the columns currently used in Martinique, the second boiler has been removed and the first boiler, which is fortunately continued to give in general a large capacity, is heated by direct steam.
[This is a fascinating concept and attempt to collect “tail waters” possibly full of rum oil. This concept is going to need more attention.]
Steam heating can be by bubbling, coil or tubular bundle. The most frequently used is that by bubbling, which has the advantage of being simple and economical, but the disadvantage of diluting the vinasses. On the other hand, if we use the exhaust steam of the machines, it happens that it causes particles of oil or fat, which contaminate the brandy and give it a very unpleasant taste. [The exhaust steam here would be of the cane crushing machines. In many cases I have translated bubbling to direct steam because it is for frequently used in contemporary literature.]
Coil heating is expensive. The coil occupies a rather considerable space, which makes it necessary to give a larger capacity to the boiler, and moreover it gets dirty frequently.
In the tubular heater adopted by the Savalle house, the hot vinasse coming out of the appliance circulates inside a bundle of tubes heated externally by the steam. The vapors produced pass through the column through a pipe located at the top of the heating box. The vinasse flows continuously through a tap at the bottom of the unit. [This simply sounds like a heat exchanger.]
In another system sometimes called “bouillisseur” and used in some factories in Martinique, heating steam is admitted inside the tubular fiber, outside which the vinasse circulates. This arrangement not only avoids mixing with the wine condensation water heating vapors, which often contain traces of lubricating oil, but also allows a prolonged boiling of vinasse and better extraction of aromatic principles. [I’m not sure if fiber from “faiceau” should be better translated as coil or winding. I also wasn’t sure if boillisseur should simply be boilers. This does seem like another deliberate way to exaggerate time under heat for the benefit of aroma.]
For the heating of the columns, the exhaust steam of the steam engines is generally used. This is often stored, at the pressure of 1.5 kg for example, in an exhaust balloon, also equipped with a live steam intake valve, which opens automatically when the pressure falls below 1.5 kg. The steam arriving at the base of the column must have a pressure a little higher than the sum of the pressures represented by the liquid layers which are on the different plates. For low-pressure columns, a pressure of 200 to 250 gr. above atmospheric pressure is usually adopted. [so that intake opens automatically to maintain pressure]
Regulators of steam, feed supply, etc.— In order for the column to work in good conditions, by ensuring the complete exhaustion of the vinasses with the minimum of steam expenditure, it is important that the wine supply and the steam supply are well proportioned. If the quantity of steam is too great, the degree of alcohol produced is lowered, and if it is too weak, the exhaustion is imperfect.
In some appliances (Champonnois), the adjustment is obtained only by means of taps placed on the arrival pipes of the wine and steam. The setting is then very difficult, the pressure of the generators and, to a lesser degree, the alcoholic richness of the fermented must varying within quite large limits. [You gotta watch a lot of thermometers to gauge the exhaustion at different points.]
The solution generally adopted is to act on the heating steam, using an automatic Savalle type regulator. This device is a water gauge, consisting of 2 containers joined by a central tube, plunging to the bottom of the inner tank. The latter contains water and communicates with the steam chamber of one of the trays of the column. The pressure which prevails in the apparatus is exerted in the upper reservoir, or there is a float that acts, by levers, on the steam inlet valve in the column, on a special valve or throttle valve. If the pressure rises, the float is lifted by the water and the valve closes partially; if it decreases, the float lowers and the valve opens, admitting more steam. Various improvements have been made by the manufacturers (Barbet, Egrot, etc.) to the regulator of Savalle, to make it more sensitive and more regular in operation.
It may be advantageous to modify the course of the column, so as to vary the flow of the apparatus or the richness of the alcohol produced. The variable speed is obtained, according to the manufacturers, by moving the upper tank of the regulator by any process (rack, chain, telescopic tubes, etc.), by varying the level of the water in the lower tank by means of taps overlapping (Barbet), or taking the pressure by valves superimposed on different floors of the column.
The feeding of the wine column, contained in a constant level tank, is usually set by hand. However, for direct fire appliances, Egrot builds an automatic regulator, consisting of a copper container, communicating through a pipe with the liquid contained in the boiler. The pressure which is established in the latter during the run raises the vinasse in the container of the regulator where is located a float. The movements of this float are transmitted by a rod articulated to the wine inlet valve.
Beside the adjustment of the pressure, it has often been advocated that of the temperature. The pressure being adjusted, it is necessary, in order to be in conditions of maximum yield, that the temperature be maintained at a certain degree. Usually, a thermometer is placed at a certain stage of the column in the steam chamber of a plate, and, according to the indications of this thermometer, the arrival of the wine is manually regulated: when the temperature drops the supply valve is closed in a small quantity and when it rises, it is opened.
Finally, to regulate the arrival of the water (or wine) to the condenser, often placed at the top of the columns a manometric bottle or a manometer graduated in centimeters of water. If the pressure exceeds the normal, there is lack of water: in the opposite case, there is excess of water. It is therefore possible to easily correct, according to the manometer indications, the flow of water to the condenser, so as to maintain the pressure potential (1) of the column corresponding to the maximum yield.
(1) This pressure potential is different from the total pressure, delimited by the pressure regulator. It is given by the difference between the pressure at the bottom of the column and that at the outlet, before the vapors enter the condenser or the wine heater. The total pressure remaining unchanged, the pressure potential varies with the amount of retrograding liquid on the concentration trays, which itself depends on the amount of water arriving at the condenser. [I think these are very slight pressure differences but enough to impact operating balance.]
The discharge of the vinasse is usually obtained by means of a curved siphon pipe, which ensures a constant level of liquid in the boiler and allows the excess to come out automatically. Sometimes a special regulator is used, consisting of a tank containing a float controlling an outlet valve: when the level of the liquid reaches the float, it floats up and opens the valve.
Condensers, wine heaters, refrigerants, etc.—The condenser is generally constituted by a tubular bundle into which the cooling liquid arrives. This passes into the tubes of the apparatus, around which the alcoholic vapors circulate. Condensation retrogrades to the column.
It is economical to use heat from the condensation of alcoholic vapors, for the heating of the wine that feeds the column. But this has the disadvantage of linking the condensation to the power supply, which is a cause of imbalance. As well, there may be losses of alcohol by the release of the incondensable gases of the wine heater and if the section of the discharge tube is too small, a return of the wine in the tank under load, under the action of the counter-pressure which is established. [I think this is only referring to continuous stills.]
In the rhummeries of the French West Indies, the wine passes successively through 2 tube-bundled wine heaters or, more rarely, a serpentine. The heating surface of each of these is on average 4.5-5 mq, for a production of 100 liters of rum per hour. The wine is carried in the first wine heater at the temperature of 65-75° and in the second, generally known as the analyzer, at 90-95°. [That 90-95° may sound high, but keep in mind that is only a 6% abv wine.]
The refrigerant is constituted by a tubular system or, more often, by a coil. The coolant is sometimes the wine itself, but it is much better to use water. In fact, as the temperature of the fermented must hardly descends, in hot countries, below 30-35°, it is impossible to lower the condensed vapor below 35-40°, hence appreciable losses of alcohol by evaporation. The cooling surface is about 3.5-4 mq, for a production of 100 liters of rum per hour.
Arroyo drew attention to the need for a gradual condensation of alcoholic vapors, to obtain a soft, “non-dislocated” rum. He advises the use of a long and narrow refrigerant, with well-marked temperature zones: 25-30° at the base, 45-50° at the middle and 60-65° at the top, the difference in temperature water at the inlet and outlet of the refrigerant to be at least 30° C.
We can avoid the inconveniences inherent to the use of the wine pre-heater, by using a vinasse heat exchanger, to bring the wine to the desired temperature, This apparatus is generally constituted by a tubular bundle: vinasse, which leaves the column boiling arrives at the top, passes between the tubes and flows to the lower part. The wine enters from below, circulates inside the tubes and comes out warm at the top of the device. This device, which is particularly valuable when practicing the return of vinasses, but is rarely used in the French West Indies. [This quickly gets the dunder to a workable temperature so a new ferment can be set.]
Eprouvettes. — It is necessary to be able to control at any rate the alcoholic degree of the distilled liquid. A test-tube is most often used for this purpose. The simplest test-tube model consists of an expanded copper tube, into which an alcoholometer plunges: the alcohol arrives at the bottom and flows from above. A glass bell covers the whole. [Eprouvette did not translate the best and I left test-tube as its rough approximation even though we call them parrots or spirit safes. Its really nice to distilleries investing in gorgeous spirit safes.]
The Savalle eprouvette-gauge allows the distiller to realize at the same time the flow velocity of the liquid. This arrives at the lower part of a glass globe, in which is located a central tube pierced at its base with a small diameter flow orifice. The level of the liquid in the globe is in direct relation with the flow: a graduation inscribed on the central tube indicates the quantity of alcohol passed by hour. By plunging an hydrometer and a thermometer into the test tube, the degree of alcohol and the temperature can be checked at any moment.
In a second type of eprouvette (Barbet), the flow rate is measured by timing. The liquid arrives through a central tube and pours into the glass globe, which it fills until overflow, by a concentric tube to the one of arrival. To know the flow, the specimen is emptied by means of a special draining tap, then the stopwatch is counted to the time necessary to fill a certain volume read on the glass globe, which is graduated.
In English countries, small glass ampoules (bubbles), which are set so that they remain immersed in alcoholic liquids of different densities, are most often used to determine the alcohol content. When it is desired to have, for example, alcohol at 45° over proof (83° 3 GL), 3 bubbles are placed in the eprouvette (test case), one which floats in alcohol at 42 O.P. , the other in alcohol at 45 O.P. and the third in alcohol at 48 O.P. If the 3 ampoules rise at the same time, the alcohol is too weak, and if 3 of them sink, it is too strong. [I have a rare complete set of these and they have often called philosophical bubbles. I think they were only used in the early 19th century before the modern hydrometer. We think of the hydrometer as always the go-to tool for measuring SG, but keep in mind these are in-line continuous measurements. For stationary measurements, the pycnometer often is more accurate before the modern era of u-tube densitometry.]
In some devices, the flow rate with the test specimen is held constant. The alcoholic vapors first pass through a first wine heater E, whose condensation is retrograde on the column, then in a second refrigerating wine heater, where condensation ends. Condensate comes out through tubing with an air instruments K, and goes down to the test tube N. It carries at its entry a control valve, allowing to let exactly the quantity of alcohol that one wants to obtain. The excess liquid rises in the tube K and returns through an overflow pipe R at the bottom of the condenser.
Some columns (Barbet) are also provided with a device for the pasteurization of condensate. The alcohol is taken, in whole or in part, on one of the concentration trays by the tap H and, after being cooled in the refrigerant G, flows to the test tube P, which is provided with a control valve. The violent boiling of the retrograde liquid expels the most volatile products. The eau-de-vie obtained is purified and undergoes quick aging. [What I think they are doing is drawing off a single plate where predictable congeners exist then refining it. I don’t think this is fusel oil separation, but rather separation of ethyl acetate and acetaldehyde.]
Description of some continuous devices.
Creole column.— The Creole column, or du pays, is very much used in the agricultural and industrial rhummeries of Martinique. It is built on site and has been little changed since the end of the last century. It processes wine at 4-5° and gives rum at 55-60° G.L., sometimes 65-70°, if the number of trays is quite large.
It consists of a boiler heated by direct steam or, more rarely, by a coil, and surmounted by a column with circular plates, having 0.80-1 meter in diameter, provided with 5 hemispherical caps. There are usually 12-16 depletion trays and 0 to 4 trays of concentration. [depletion trays for the beer (vinasse) and concentration trays for the first distillate (no vinasse).
The alcoholic vapors leaving the column pass through 2 wine heaters, the first with tube bundle (analyzer) (1), the second with tube bundle or coil. The condensations of the analyzer demote to the column. The second wine heater plays, depending on the case, the role of a condenser with demotion condensation on the column, or a simple refrigerant. It is usually doubled by a water cooled condenser. The vapors emitted by the wine from the wine heater pass through a foam breaker: the entrained liquid particles return to the column through the supply or the retrogradation pipe, while the incondensable gases are released into the atmosphere, sometimes after crossing a coil placed in the wine rack, or more often go to the condenser coil. [Incondensible gases may mostly be CO2.]
(1) At the end of the last century, the analyzer consisted of a small cylinder without tubular bundle, externally contacted by alcoholic vapors.
The adjustment of the steam and the supply is done by hand. There is no heat recovery of vinasses. When these enter the composition of the must, they are cooled by passing through a water cooler, with a tubular bundle or with a coil.
Savalle Column. — There are different models of Savalle columns, some with circular trays, others with rectangular trays, with rectangular caps. The number of trays varies from 15 to 25, depending on the proof degree to be obtained. Heating is usually done by direct steam.
Column model E. 7 No 12, which can supply 18,000 liters of rum at 70° per 24 h., consists of 13 trays of exhaustion and 5 trays of concentration. It does not have a boiler: the vinasse is evacuated by a siphon tube. The alcoholic vapors coming out of the column, after passing through a foam breaker, pass through two superposed horizontal tubular wine pre-heaters joined by a pipe equipped with an adjustable diaphragm. The condensations of the second wine pre-heater partially retrograde back into the column and pass partly to the eprouvette, after having passed through a tubular-bundled water cooler. The controller with the eprouvette regulates the degree of the retrogradation. The vapors emitted by the wine from the wine pre-heater pass through a foam breaker: the mosses retrogresses in the wine heater and the incondensable gases pass through a coil placed in the wine tank before being released into the atmosphere. The admission of the wine is regulated by a hand-held tap and the arrival of the heating steam by a Savalle regulator. The apparatus can give alcohol at 85°, but in the rhummeries of the French colonies, one does not exceed the measure of 70°.
The Savalle columns are very common in rhummeries. Already by 1884, the number of these apparatuses active in the distilleries of cane molasses amounted to 122, of which 82 produced rum and 40 gave high proof alcohol (J. Paul-Roux). They are found not only in the French colonies (Martinique, Guadeloupe and especially Réunion), but also in South America (Brazil, Argentina), the Greater Antilles (Cuba, Puerto Rico) and so on.
Barbet column.—Barbet has imagined, for the distillation of wine spirits, a model of a continuous apparatus, in which the wine and the vinasse are subjected to prolonged boiling. Although this device is not known to us in rhummeries, we believe it worth mentioning, because of the interest it has for the production of very aromatic eaux-de-vie.
The fermented must, when leaving the feed tank P, first goes into a heat recuperator of the vinasses M: the flow is regulated by means of a tap. Then, the wine enters the bottom of a large wine heater G, containing enough liquid to feed the column for 3 or 4 hours. The liquid is brought to the temperature of 92-93 °, by a coil receiving steam taken from the boiler of the apparatus. It then enters the exhaustion trays of column E (comb-cap trays), which it descends while gradually exhausting alcohol. The vinasse, from the last tray, flows in a boiler D of large capacity, heated by a coil, where it is subjected to a boiling of 3-4 hours before being evacuated.
In order to recover volatile acids and heavy aromatics contained in the vinasse which could not, because of their high boiling point, rise in the column, the condensed vapor in the coil of the wine pre-heater G, is collected in double condensor H. The cooled liquid exits by the eprouvette R, where a very sensitive alcoholometer makes it possible to control in a permanent way the exhaustion of the vinasse in alcohol, and is sent by a pump to the tank O. From this tank, it goes in the section of the concentration column, above the arrival of the wine, where it encounters alcohol at 40-50°. The acids in the water combine with the alcohol to give esters.
[Notice there are multiple path ways out of boiler D. What I’d be interested to know is if D is ever under pressure to generate super heated steam. What they are hoping for is high value tale waters that are free of tufo. This idea may be a lot more feasible now that you can centrifuge the ferment a la Arroyo. Other designs were known to recycle fractions to the top of the column to generate esters, but possibly not the same was as here. I never found R in the diagram and and wondering if there is a typo. I suspect this diagram was taken from Mariller.]
The alcoholic vapors emerging from the exhaustion column pass into the concentration section situated above and then go to 2 water condensers J and K, the condensations of which completely retrogress on the concentration tray F, with the exception from 2 to 5% of overhead products (composed of aldehydes and gases with an unpleasant odor) taken from the second condenser. These bad tastes from the heads, which measure 90° and even 95°, are collected in eprouvette T. The high-grade alcohol coming from the condensers descends the concentration column, decreasing progressively. 3 taps, placed at different levels, allow extracting the eau-de-vie on the tray where it is the finest. With the wines of good quality, it is generally on the tray of 60-70°, that the product has the maximum of quality.
The spirits coming out of the column pass to the refrigerant H and exit to the eprouvette S. The column is provided with a Barbet vapor regulator. A very fine eau-de-vie, already aged by the prolonged boiling of the alcohol (pasteurization), is obtained with this apparatus.
In the case of wines of inferior quality, containing sulfur dioxide, hydrogen sulphide (which, when combined with alcohol, gives rise to foul-smelling ethers, which can not be eliminated later), ammonia, etc., Barbet performs a purification of the alcoholic vapors prior to concentration. The vapors, on leaving the depletion trays, pass into a special vessel, called a reagent vessel, which can overcome or be separated from the depletion column. They first cross a compartment containing pieces of marble, which retains the sulfurous anhydride and sulfuric acid, then they bubble in a solution of metal salts to absorb H2S and NH3. The purified vapors then go to the concentration or rectification column.
For the distillation of rum, Barbet constructs a column resembling the one previously described, but simpler. The fermented wine passes successively in the refrigerant H and the wine heater G. From there, it arrives at D in a first small column intended to expel the bad fermentation gases. The boiling takes place by means of a little alcoholic vapor taken from the top of the distillatory column C-C ‘, by the valve S. The mixture of gases and light vapors goes into the water condenser E. The gases and a few aldehyde vapors escape into the atmosphere, while the rest condense and enter into D. [I think this refers to diagram two pictures ago…Fig 41. Diplomatico apparently uses a simplified Barbet column.]
The wine, purified from its bad gases, descends into the exhaustion trays C-C’ and from there passes into the boiler B, which is heated by direct fire or by steam coil. The boiler is of sufficient capacity so that the vinasse can be cooked for a long time. The alcoholic and acidic vapor passes to the base of the second plate column F, which in turn receives the reflux of the tubulars G and H. The pasteurized, softened rum, is extracted in the liquid state and cooled in the refrigerant I. It goes from there to the eprouvette P. At the same time, a very small proportion (1 or 1/2%) of aldehydes is extracted at the head, which cools in a second coil I and flows from there to the eprouvette T.
[We’re seeing multiple eprouvettes and I’m still not positive how he is using the word pasteurization.]
Various other models of French columns are used in rhummeries. We can point out in particular the Crépelle-Fontaine appliances, quite numerous in the industrial distilleries of Martinique, Mollet-Fontaine, Deroy, Egrot, Wauquier, Fives-Lille.
Coffey device. —Very common in the English colonies, the Coffey column is often built of wood and in rectangular form. In modern appliances, however, circular plates, made entirely of copper, are preferably adopted.
The apparatus, which produces alcohol at a high degree (84-85°), consists of 2 large twin columns, one serving for the exhaustion of the must (analyzer), the other for enrichment and the purification of alcoholic vapors (rectifier). These columns are formed by the assembly of trays with rectangular wooden frames, each carrying a copper plate pierced with small diameter holes. Each plate is further provided with a copper valve opening from the top and an overflow tube. However, the top plates of the rectifier are full and provided only with an opening at one or the other of their ends alternately. The rectification column is furrowed by a continuous coil, passing successively through all the steam chambers and traversed by the wine to be distilled.
This still, usually working with a low alcohol charge, arrives in the rectifier, descends the trays by the serpentine pipe and, once reaching the bottom of the column is pumped to the analyzer top plate. It crosses the various trays of the latter, through the over flow tubes, gradually running out of alcohol, to flow in the state of vinasse, by a siphon tube, at the base of the apparatus.
Steam, admitted at 350-700 gr. pressure at the base of the analyzer, flows from bottom to top, through the perforations of the trays and, if the pressure is too high, through the valves. The alcoholic vapors escape from the analyzer through the pipe f, to penetrate the lower part of the rectifier. In contact with the cold wine contained in the serpentine pipe, the vapors partially condense by heating the liquid. The water, the high-boiling substances and a certain amount of alcohol condensed at the bottom of the column are discharged through the hose h: they constitute the hot faints, which are mixed with the wine later. The alcohol is largely condensed in the 5 upper chambers of the rectifier; it is collected in the liquid state on the plate n, in the receptacle o, where it is taken by the pipe m, which leads it to a water cooler. The non-condensed vapors are discharged through a pipe G and condensed in a condenser. They are cold feints, which are either mixed with the wine or sent by a pump to the top of the analyzer. In 1, there is a water coil that adjusts the amount of overhead products collected.
The wine comes out of the rectifier at a temperature of 88-93° C. The measure of the alcohol obtained can be adjusted by acting on the temperature. When it rises excessively, the condensation is bad, the measure decreases and the alcohol passes largely in hot feints. On the contrary, a low temperature causes condensation of the vapors above the alcohol extraction tray and an increase in the amount of cold feints. The best results are obtained, in the distillation of rum, by maintaining a temperature of about 81° C. in the steam chamber of the second or third tray situated above the extraction tray, which gives an alcohol containing 85-86° GL.
The column is regulated by acting on the wine supply or on the steam, by means of taps placed on the supply pipes. To be aware of the degree of alcohol, a tubing, which passes through a special refrigerant, draws a small amount of liquid on the extraction tray and leads it into a test tube containing spirit bubbles. Thermometers placed at different stages of the column provide additional control. [They use spirit bubbles or philosophical bubbles because their volumes are likely very low. Possibly they would also very very precise because of their narrow range?]
A column capable of handling 1,000 gallons (4,546 liters) of must per hour is approximately: 24 X 8 X 3 feet rectifier; analyzer 42 X 8 X 3 feet; number of trays of the analyzer and rectifier 27; Serpentine pipe surface of the analyzer 416 square feet.
Treatment of vinasses
Evacuation of vinasses and bottoms, which is easy to achieve when distilleries are placed at the seaside, presents some difficulties for settlements located inland. The dumping of these products into watercourses is, in fact, a cause of real inconvenience and economic damage for residents: the fish die; the water, cloudy and blackish, gives off an unpleasant smell, becomes unfit for animal feed, washing clothes, etc. (1). In most countries, therefore, the discharge of distillery waste water into rivers which has not undergone preliminary purification has been forbidden. [Before you ever make rum, you must figure out how to properly dispose of the vinasse.]
(1) It is so at least when the flow of the river is not important. The regulations currently in force in Martinique provide for vinasse to be kept in watertight pits. Until the floods of wintering. At this time, the contents of the pits can be poured into the streams.
The volume and composition of the latter vary greatly according to the raw materials used (molasses or cane juice), the density of the musts put in fermentation, the importance of the return of the vinasses, etc. Vinasse from distilleries processing vesou generally contain 30 to 40 gr of dry extract (including 3-4 gr of mineral matter) and have an acidity of 3-6 gr sulfuric acid per liter. Their volume is usually 10 to 15 hl per hl. rum at 55°, or per ton of cane handled. They are a little more concentrated than beet vinasse, which generally contains 18 to 25 gr. dry extract per liter.
The vinasses supplied by the distilleries working the molasses are more concentrated and of smaller volume. The dry extract, generally between 70 and 100 gr. per liter (of which 20-25 grams of mineral matter, rich in potash), can reach 200 gr. and more, in establishments working with dense musts, with a strong return of vinasse (making grand arôme rum). Acidity, of 4-10 gr. usually, rises in the latter case up to 30 gr. per liter. The volume of the vinasses produced is then very low: 1 to 1.5 hl, instead of 10-12 hl, approximately per hl. rum at 55 °, when the musts are not dense and the return of the vinasse weak or null. Between these extremes, obviously all the intermediaries exist. Beet molasses contain on average 100-150 gr. dry extract per liter. [As it becomes increasingly difficult to process effluent it may make sense to pursue the production of rum styles that recycled dunder. They are also higher value so you can get a lot more value per area of land.]
The bottoms, formed mainly by yeast cells mixed with fine bagasse, are much more loaded with organic matter than the actual vinasse: the rate of dry extract can vary from 100 to 300 gr. per liter and that of nitrogenous materials from 20 to 70 gr. They constitute an important element of the nuisance of waste water from distilleries.
Many studies have shown that most often the power of a wastewater to produce a nuisance is proportional to its power of deoxygenation of the water of the river into which it is discharged. Yeast cells, which exist in high proportions in the bottom of the tank, are a nitrogen-rich and a very putrescible material, requiring large quantities of oxygen for its decomposition.
Hoover and Burr (1) indicate the average constitution of the distillery vinasses of cane molasses (in gr, per liter):
(1) Ind, Eng. Chem. XXVIII, 33, 1936.
Various processes have been applied for the treatment of distillery wastewater: concentration of vinasses for the production of fertilizers, spreading on cultivation grounds, chemical or biological purification before their discharge into watercourses. The use of one or other of these processes is especially a case in point.
Concentration of vinasses.
The concentration of vinasses, with aim of recovering the fertilizers contained therein, is frequently practiced in beet molasses distilleries in France. In the past, the concentration of vinasse was used successively until about 21° Baumé, then its auto-incineration, in the potash furnace at Porion. The dry residue obtained, or saline, is rich in potassium carbonate (40 to 60%). Nowadays, evaporation is done in multiple-effect units, which require lower fuel costs. On the other hand, the calcination having the disadvantage of destroying the nitrogenous organic matters, it is preferred to mix the syrup obtained by the concentration (40-50° Baumé) with mineral or organic materials, such as natural phosphates, sawdust, peat, etc., allowing direct use as fertilizer.
The evaporation of the vinasses is rarely carried out in the rum distilleries. However, for establishments producing concentrated vinasse, it represents a process of rational and economic use. The high viscosity of the product and the tendency to foul the devices are however a pitfall to avoid. It is important to use, to achieve concentration, boxes with separate partitions, vertical or inclined (Prache and Bouillon, Barbet), claiming to be better than of the sugar house that are rapid acting. Kestner evaporators are also very often used in Europe.
Furthermore, the mixture of concentrated vinasse syrup with inert materials (mineral fertilizers, sawdust, peat, fine bagasse, press cakes, etc.) has the disadvantage of being hygroscopic due to the presence of caramel and especially glycerine. Several methods have been recommended to remedy this. That of Metzl, which gives good results, consists in fixing glycerin by superphosphates: the syrup at 45° Baumé is heated in the presence of super at a certain temperature, then peat is added. This gives a fertilizer that stays dry. [not sure what super refers to.]
It is generally accepted that diluted vinasse such as in beet distilleries (density 1.003-1.010) or vesou (density 1.010-1.015), contain a proportion of water too high to consider their concentration. Various authors (Barbet, Kestner, Lefrançois, etc.) have shown that the operation was economically feasible in beet distilleries. A fortiori, would it be so in the agricultural rhummeries processing vesou, which have a quantity of bagasse superabundant as fuel. Unfortunately, the cost of the necessary installations is very high and out of proportion to the importance of these establishments.
The composition of the evaporated vinasse presents rather large variations. The fertilizer content of some samples from cane molasses distilleries (% dry matter) is given below:
(1) Vinasse previously neutralized with lime. [This note goes with the Brazilian sample.]
Boname (2) indicates the composition of salts obtained from vinasses 3 and 4 in the above (% of raw ash):
(2) Ann. Sc. Agron. 2, II, 265, 1895.
Spencer found that the ashes of Cuba’s cane juice (and consequently vesu vinasse) contained 25-45% K2O. N. Deer found 41.8% K2O in the ashes of molasses vinasse.
Where available land is suitable, land application on cropland is undoubtedly the most practical and economical method of using vinasse. It is widely practiced in beet distilleries in northern France: in 1905, out of 321 distilleries, 223 used their wastewater for irrigation.
The amount of vinasse that can be spread per hectare varies with the nature of the soil and the terrain: from 3 to 4,000 hl in clay soils, it can reach 10,000 and even 20,000 hl if the land is well permeable. There is interest in not returning the vinasses on the same field every 3 years. As a spreading method, the discharge into the ridges is used in preference to uniform flooding. Irrigation must be methodical and the plots irrigated one after the other, to facilitate the drying of the canals.
Pure vinasses can not be used without inconvenience. Their strong acidity hinders the action of microbes, especially that of nitrification bacteria.
On the other hand, suspended solids clog the distribution channels, preventing infiltration and therefore promoting stagnation and putrefaction of water, which emits bad odors. It is therefore necessary to neutralize, at least partially, the acidity and to let the molasses stay in settling basins for 3 or 4 days, before sending them to the fields.
When the acidity is not too strong, one can, in the absence of lime, be content to dilute the vinasse with 3 or 4 times their volume of water. An ammoniacal fermentation is soon established, which renders the reaction alkaline.
In terms of fertilizing value, a contribution of 4,000 hl. vinasse at 1.035 density corresponds to about 40 tons of good farmyard manure, obtained by using as straw litter.
Chemical precipitation has been proposed on a number of occasions to purify the vinasse and to recover, in the form of sludge, some of the fertile material.
N. Deerr recommends the use of lime. By treating cane molasses with commercial slaked lime, at a rate of 275 gr. per hl. of liquid, he was able to obtain a precipitation of about 60% of the nitrogen. The deposit, which represented 15% of the volume of vinasse treated, contained after drying 3.82% of N. However, lime alone is generally considered insufficient to produce good clarification; it is important that it be in a slight excess, which necessitates the use of relatively large quantities.
P. Gaillet recommended iron percholide (1 kg per cubic meter) and lime (3 kg per cubic meter), which would allow, in the case of beet vinasse, to have a 70% elimination of the total organic matter and nitrogen. Other authors have used lime and alumina, lime and ferric sulphate, etc. But it has been generally found that the liquid separated from the sludge still contained large quantities of organic matter and remained putrescible. Chemical precipitation can therefore only be considered as a preliminary stage of biological purification.
Nathan-Lévy (1) reports that he has successfully used vinasse in Peru to improve the physical state of ash from bagasse kilns. These are crushed finely and sent to a brick lined tank, where the vinasse is poured as it leaves the distillation apparatus, as well as the scums from sugar production. The mass, initially very fluid, dries quickly and can absorb large quantities of vinasse. After about ten months, it became pulverulent and very friable. An analysis gave the following percentage composition:
(1) Bull. Ass. Chim. XLI. 208. 1923.
Potash, which in the ashes was in the state of insoluble silicates, had become largely soluble.
Numerous tests carried out in England, France, etc., have shown that it is perfectly possible to apply to distillery vinasse biological purification processes by bacterial beds, used for the treatment of sewage. However, it is essential, for obtaining results, to neutralize beforehand vinasse whose high acidity hinders the action of bacteria; dilute them with a certain volume of ordinary water; finally to separate organic matter in suspension.
E. Rolants (1) advises to operate as follows for beet vinasse. The vinasses, diluted with 4 or 5 times their volume of water, are poured into a large settling basin, the capacity of which must be such that it can contain the dilutions of 5 days of work. Initiate the ammoniacal fermentation by neutralizing the liquid with carbonate of lime and introducing soil or a little manure. Fermentation then continues on its own.
(1) Les eaux résiduaires. Paris, 1925.
[Borrowing biomes becomes very key.]
Decanted and fermented waters leave the basin through a spillway and fall into a channel, which leads them into the distribution channel to reservoirs feeding the bacterial beds. These tanks, equipped with automatic flushing siphons, have a capacity of 800 liters each and are in number corresponding to the volume of the liquid to be treated (each tank discharges per day about 750 hl., at the rate of a course per quarter of an hour ). The bacterial bed must have a maximum width of 10 meters on each side of the line of the flushing cisterns and a length corresponding to the quantity of liquid. It requires 2 quare meters of surface per cubic meter of liquid. The height of the bed is about 2 m., the lower 30 cm being constituted by large slag and the rest by slag all coming off as dust. The bed is placed in slopes, to avoid the construction of retaining walls.
This purification method is applied after Cosculluela (2) at the Arechabala distillery (Cuba) for the treatment of cane molasses vinasse. The vinasses are diluted with 4 times their volume of water, are received in a sedimentation tank and seeded with a culture of ammonizing bacteria (pig manure or cow). After 4 days, a vigorous ammoniacal fermentation takes place: the pH, from 4 at the beginning, passes soon to 7 or 8. The optimal temperature is that of 37° C. It is not advisable to push the decomposition of suspended solids too far, since the fertilizing value of the sludge deposits would be reduced. After leaving the settling basin, the vinasse is filtered on beds of slag or coal: it requires 1 cubic meter of filtering surface per cubic meter of liquid to be treated.
(2) Rev. Cubana de Azue. Ale. 1, 60, 1935.
Hoover and Burr (3) compared stick beds, sand beds and coke beds. They found that first, once active, provided good purification: a molasses vinasse clarified and diluted so that the biochemical oxygen demand is reduced to 2 gr per liter, is freed of about 90% of the oxidizable materials with a flow rate of 1 cubic meter per square meter per day. The sand filters were easily clogged and had a low flow rate. As for powdered coke filters (particle diameter 1mm) of 6 m. At depth, they made it possible to reduce by 98% the biochemical oxygen demand, with a flow rate of 200 l. per square meter per day.
(3) Ind Eng. Chem. XXVIII, 38, 1836.
Artificial biochemical purification of vinasses has the disadvantage of requiring expensive facilities. So it is practiced only if it is not possible to do otherwise. On the other hand, fertilizers are lost almost entirely. The sludge from the settling pits, however, retains a certain amount of these. The deposits obtained at the Arechabala distillery, in Cuba, have, according to Cosculluela, the following percentages composition after drying:
Finally, if the waste water can be discharged into the sea without any prior purification, it is of interest to recover some of the nutrients contained in the bottom of the tank, by passing them to the filter press after neutralization with lime. Cakes consisting mainly of yeasts are thus obtained. According to N. Deerr, these would contain, for a moisture content of 60%, 6.6 to 7% of nitrogen. One cubic meter of molasses must yields an average of 5 kg of yeast cakes. The nitrogen thus recovered represents 5 to 20% of that originally contained in the mash (Peck and Deerr).
Yeast cakes and vinasse deposits can be used directly as fertilizer. But it is preferable to use them for the preparation of compost and artificial manure. Especially because of their high nitrogen content, they allow rapid decomposition of cane straw and even the toughness of bagasse, as Hinchy has shown (1). This author obtained composts with the following percentage composition, treating cane straw with residues of vinasse:
(1) Int. Sug. J. XXXVII, 296, 1937.
In the small and medium sized distilleries of the French West Indies, the vinasse is often sent into pits filled with bagasse mixed with cane straw. The installation is unfortunately most often misunderstood, the emptying of the beds takes place at intervals too distant, etc. so that acidic fermentations, giving off bad odors, and imperfect humification of vegetable matter are produced. As well, the effluent is still putrescible. However, perfected, the process could give interesting results. In particular, it should be possible to establish several successive pits which can be easily drained and to finish with a filtration on slags or on charcoal. The bagasse and the cane straw removed from the pits would be arranged in small heaps as in the Indore method for the manufacture of artificial manures (2), so as to allow an aerobic fermentation, without the release of unpleasant odors.
(2) Indian Med. Gaz, LXIX. 93, 1934.
Recovery of yeasts from the bottom of the tank
The bottoms are mainly yeasts (3-6 kgs of dry yeast per hl of alcohol at 100°), which have a high nutritional value and it is therefore of interest to recover for the feeding of livestock (fodder yeast) or even for human consumption. [marmite!]
According to Reich (3), the dry yeast from the distillation of cane molasses has the following percentage composition:
(3) Sugar XL. N° 3, 26, 1945.
This composition is comparable to that of brewery dry yeast. Protein and vitamin B levels, however, are sometimes lower.
In fact, when musts have not been the subject of a preliminary purification, the bottoms contain a high proportion of impurities (fine bagasse, gums, minerals, etc.), which make their desiccation very difficult and significantly lower their nutritional value. Here is the composition of bottoms of vats obtained in rhummeries of Martinique:
Preparation of dry yeast.
If it is desired to recover the yeasts of the vat room, it is therefore important to subject the wort, before fermentation, to a purification, by the Arroyo process, for example for molasses musts (Cf, chapter V), by defecation and filtration for cane juice.
The bottoms can be collected in a channel, where water flows into a reservoir. The yeast is then sieved to remove foreign matter. This operation is done by passing through two screens, the first in copper cloth (No. 12), the second in fine silk. Screening is facilitated by sifting water over the fine sieve.
The collected yeast is sent to settling tanks, where it is mixed with very cold water (9-12°). After 2 or 3 hours, it settles, forming 2 layers: at the bottom the pure yeast, above a grayish layer consisting of immature cells and foreign ferments (“gray yeast”) This layer is decanted with care and the yeast remaining in the tank is subjected to a second and even a third wash, followed by decantation, before being sent to the filter press.
However, it is much more practical, especially in hot countries, where it is difficult to achieve a low enough temperature in the settling tanks, to use for the separation of yeasts De Laval type centrifuges, turning at 4,500 t/m, in which the fermented must is passed before distillation.
The yeast milk supplied by the first separator is not concentrated enough to be subjected as such to desiccation. It is passed in a second centrifuge. Although the product leaving the latter has reached a sufficient concentration, it is generally advantageous to eliminate impurities (gums in particular, which hinder drying) to mix with water and to make a 3rd transition to the centrifuge. If it is desired to obtain a yeast intended for human consumption, several washes may be necessary.
The yeast dough is generally dried by passing over slow-spinning hollow cylinders (5-9 mt) heated with steam.
It is also possible to prepare, from the yeast, yeast extracts, the chemical composition of which is very similar to that of meat extract.
Various methods have been advocated. In the Wahl and Hénius process, for example, a simple yeast decoction is made by boiling with water and the resulting liquid is concentrated. The Peeters process consists of extracting the protoplasmic contents of the cells at a temperature of 60° with liquefying substances such as hydrochloric acid. Other authors (O’Sullivant, Nolf, etc.) use yeast autophagy at 40-60°, to cause the solubilization of its contents: the protein materials are hydrolysed and dissolved in the form of amino acids, amides and purine derivatives.
The latter method can be applied as follows. The yeast, well washed, is diluted with a little water and 10% of its weight of sea salt is added. It is then kept in a closed container at a temperature of 45-50°, for enough time sufficient to obtain liquefaction.
Once this is done, the product is spread in a thin layer, in contact with the air, and the temperature is raised to 70-75° for 1 or 2 hours. The mass is filtered through the filter press, left standing for 24 hours, decanted and concentrated by heating to a paste consistency. The product obtained has a flavor similar to that of the meat extract.
Finally, Jallowetz proposed the production of a sweet extract, by digesting yeast at 53-55° in a sugar syrup, and then concentrating to a syrupy consistency. We have been able to prepare, by this process, from the bottoms of the rhummerie, a product pleasant to the taste, flavor reminiscent of that of battery syrup, likely to be consumed as jams.