26.4
Distillation

26.4.1 Introduction

Distillation is an industrially important process for partially or completely separating (fractionating) volatile compounds in a liquid from each other or from non‐volatile components. At a basic level, distillation involves boiling a liquid and condensing the vapors to produce distillate fractions, with separation of components arising from differences in vapor composition relative to that of the boiling liquid. Crude forms of distillation have been around for thousands of years, and the use of distillation in Asia may date to as early 800 BC. While there is conjecture that alcoholic beverages were used to produce distillates in these very earlier stills, others date the first distillation of alcohol to China in the first century [1]. Distillation was also refined by the Islamic alchemists of the eighth to fifteenth centuries, and Arabic has given us several distillation‐related terms, including alcohol (al kohl/kohol) and alembic (the historical term for a still).

In Europe, the concept of distillation (and associated equipment) was first documented by the ancient Greeks around the first century AD, initially in relation to distilling sea water to produce potable water [2, 3]. Alcohol distillation in Europe was first reported in a recipe for obtaining “aqua ardens” (burning water) from wine in 1200 AD [3]. Contemporaneous accounts of “aqua vitae” (water of life) were also reported, and although the alcohol was used primarily for medicinal purposes, without doubt such distillates were also consumed as a beverage. These early distillations involved some degree of sophistication – fractions were collected and redistillation was undertaken, and at times herbs and flowers were steeped in the alcohol prior to a final distillation. Consequently, the technology of distillation progressed and spread across Europe and other parts of the world, and fermented beverages produced from sources other than grapes, such as those from grains and other fruits, were also utilized as feedstocks. Thus began the evolution of the spectrum of contemporary distilled beverages such as whisky, gin, rum, vodka, and of course grape‐derived spirits such as brandy and variations of it, as well as high strength (i.e., highly rectified) grape alcohol used for fortifying wines [4]. This chapter describes some of the principles of distillation [e.g., 5–7] as they relate to water–ethanol solutions [e.g., 8–10] (i.e., the main components of wine), and focuses on the production of alcohol from grapes and grape byproducts [e.g., 11–14].

26.4.2 Vapor–liquid equilibria

Separation by distillation is an interphase (liquid–vapor) mass transfer process that exploits differences in volatility of components within a liquid mixture. Volatility is the ratio of mole fractions of a single substance between vapor and liquid phases (i.e., equilibrium ratio K, Table 26.4.1)¹ and signifies the ease by which a compound is vaporized (evaporated). This relates to vapor–liquid equilibrium (VLE) – the coexistence of liquid and vapor phases – and the equilibrium vapor pressure of a pure compound, which is the pressure caused by molecules in the saturated vapor phase in equilibrium with the saturated liquid phase. The relative volatilities of different components in a mixture determine the ability to separate those components by distillation. The concept of volatility was presented in Chapter 1 in terms of Henry’s Law (Table 26.4.1), which is appropriate for low‐concentration solutes where volatility is primarily dependent on solvent–solute interactions. For components at high concentrations – like ethanol and water – volatility will depend more on self‐interactions and is better described by Raoult’s Law (Table 26.4.1).

Table 26.4.1 Useful equations and explanations of their relationships as they apply to distillation

Name	Equation	Explanation
Henry's Law		Pa is the partial vapor pressure of component a in the mixture, Xa is the mole fraction of component a, and Ha is the Henry's law constant of component a for a given temperature and solvent.
Raoult's Law		Pa is the partial vapor pressure of component a in the mixture, Xa is the mole fraction of component a, and P o a is the vapor pressure of pure component a at a given temperature. For a non‐ideal solution the activity coefficient is included as a multiplier (modified Raoult's Law).
Activity coefficient (γ)		Used for non‐ideal solutions to account for molecular interactions (and partial vapor pressures) that vary with the composition of the liquid. The activity coefficient for each component in a mixture may be greater than 1 (positive deviation from ideality) or less than 1 (negative deviation from ideality).
Dalton's Law		P is the total vapor pressure, Pa is the partial vapor pressure of component a, Pb is the partial vapor pressure of component b, etc, at a given temperature.
		This expresses Dalton's Law in terms of Raoult's Law by substituting in XaP o a in place of Pa, etc.
		The Dalton's Law relationship between Pa and P can also be expressed in terms of vapor mole fraction, Ya.
		This equation can be derived by substituting in XaP o a in place of Pa in the equation above, which allows liquid and vapor compositions to be related with pressure.
Volatility (K)		Ka is the volatility of component a, and Ya and Xa are the mole fractions of component a in the vapor and liquid, respectively. From this relationship it can be seen that vapor enrichment rises as K increases above 1.
		Ka can also be expressed in terms of Raoult's and Dalton's Law expressions given above. This form of the equation shows that the volatility of component a is the ratio between the vapor pressure of pure a and the total pressure. For non‐ideal solutions the activity coefficient of the component is included as a multiplier for the vapor pressure of the pure component.
Relative volatility (α)		Ka and Kb are the volatilities of components a and b, respectively. It is necessary to specify which components are being compared when assessing relative volatility.
		αab can also be expressed using Ka and Kb written in terms of Raoult's and Dalton's Laws as shown above. This reveals that the relative volatility of an ideal solution is given by the ratio of vapor pressures of the pure components (note that temperature dependence is ignored). For non‐ideal solutions the activity coefficients of each component are included as multipliers for the respective vapor pressures of the pure components.

26.4.2.1 Raoult’s and Dalton’s Laws

Vapor pressures differ based on the type of liquid and strength of attractive forces holding the molecules together. The vapor pressure of pure liquids is a direct function of temperature, and as the temperature increases so too does vapor pressure.³ For liquid mixtures, each component will exert its own vapor pressure (i.e., partial vapor pressure) as a function of its concentration in a liquid of given composition. This is the basis of Raoult’s Law, which states that the ratio of partial vapor pressure of a component to its vapor pressure as a pure liquid is equal to the mole fraction of that component in the liquid mixture. Put another way – a typical wine is around 95% water on a molar basis, so we expect the vapor pressure of water in a wine headspace to be about 95% that of the vapor pressure of pure water, assuming ideal Raoult’s Law behavior. Summing partial vapor pressures gives the total vapor pressure exerted by the liquid (i.e., Dalton’s Law, the pressure exerted by a mixture of gases is the sum of their partial pressures, Table 26.4.1). Raoult’s, Dalton’s and Henry’s Laws are represented diagrammatically in Figure 26.4.1 (a) and (d).

Figure 26.4.1 VLE diagrams showing (a) isothermal (arbitrary temperature of 70 °C) pressure–composition (P–X) plot (showing partial and total vapor pressures, and Dalton’s and Raoult’s Laws), (b) isobaric (101.325 kPa) temperature–composition (T–X) plot (highlighting regions for liquid, vapor, and vapor + liquid), (c) X–Y diagram (X = Y diagonal shown for reference) for an ideal mixture of methanol–ethanol,Figure 26.4.1 (continued) (d) P–X plot (dashed lines show the ideal behavior scenario, dashed red lines show where Henry’s Law would apply), (e) T–X plot (dashed lines indicate how distillation separates components based on an arbitrary starting temperature and liquid composition, a₁), and (f) X–Y diagram (with some relative volatilities indicated), for a non‐ideal mixture of ethanol–water. Data obtained from www.vle‐calc.com/phase_diagram.html or calculated⁴ in some instances

26.4.2.2 Ideal and non‐ideal behavior

For a binary solution containing similar chemical components (e.g., ethanol and methanol or benzene and toluene), the intensity of the molecular interactions between these different entities is comparable to the interactions of either pure liquid alone; this system behaves as an ideal solution and obeys Raoult’s Law throughout the composition range (Figure 26.4.1a–c). The situation is not the same in the case of water and ethanol mixtures that feature in wine distillation, as these two components are chemically quite different (Chapter 1),⁵ This leads to non‐ideal behavior and a positive deviation from Raoult’s Law (Figure 26.4.1d–f) because the attractive forces between water and ethanol are weaker than those of either pure component.⁶ Compared to an ideal solution, ethanol–water mixtures demonstrate higher partial vapor pressures for each component of the mixture, meaning a higher total pressure (Figure 26.4.1d). In a certain ethanol–water composition, these mixtures will also form a solution that has a lower boiling point than either pure component (Figure 26.4.1e).⁷ At the azeotropic point, liquid and vapor compositions are identical and further enrichment is not possible with ordinary fractional distillation. Hence, under normal conditions ethanol can only be distilled to 96% alcohol by volume (abv).

26.4.2.3 Separation and enrichment

A plot of temperature versus composition (T‐X plot) for ethanol–water (Figure 26.4.1e) has some features added to show how separation by distillation is possible, using a solution with an ethanol mole fraction of 0.1 (roughly 20% abv) as an example:

1. Beginning at point a₁ (or any temperature below the bubble point for this particular composition) the mixture is heated towards boiling until the temperature rises to its bubble point (when the first bubble of vapor forms) of about 87 °C (point a₂).
2. Drawing a horizontal “tie line” to meet the dew point curve (when the first drop of vapor condenses) shows the vapor in equilibrium with this liquid (an equilibrium “stage”) is richer in the more volatile component with composition a_2' (ethanol mole fraction of about 0.44).
3. The condensed liquid shown at point a₃, now at a lower bubble point of about 80 °C (approaching that of pure ethanol), also has the same composition as the vapor at a_2', showing that ethanol is enriched compared to the initial liquid composition of 0.1 mole fraction.
4. Repeating this cycle of vaporizing and condensing produces a new VLE stage⁸ and vapor composition with an ethanol mole fraction of around 0.64 (point a_3'), which condenses to a liquid with a lower bubble point again (around 79 °C, point a₄); ultimately the azeotrope for ethanol–water can be reached in this way using enough stages.

The above example assumes that each distillation stage, or “plate,” is 100% efficient (i.e., equilibrium is reached at each theoretical stage). In practice, most distillation stages are not so efficient, and production of highly purified alcohol from wine requires stills with a large number of physical (actual) stages (i.e., columns containing plates or packing material, see Section 26.4.3 below) that maximize mass transfer rates and more closely approach equilibrium through enhanced contact time and mixing of phases. Nonetheless, the example serves to highlight how the process of distillation separates the components in a liquid mixture because of differences in composition between liquid and vapor phases in equilibrium.

26.4.2.4 Relative volatility and ease of separations

X–Y diagrams are particularly useful when considering the separation of components by distillation. The distance of the curve from the X = Y diagonal (drawn for reference in Figure 26.4.1c and f) indicates the possible enrichment of the vapor phase with the more volatile component as compared to the liquid phase. Greater enrichment leads to better separation of the components in solution. More specifically, the relative volatilities for the components in a mixture (i.e., the ratio of volatility of one component with that of the other, designated α, Table 26.4.1)⁹ can be used to determine the point in distillation where the greatest enrichment occurs. For ethanol–water mixtures, the greatest enrichment occurs for dilute ethanol solutions (high α values, Figure 26.4.1f), whereas it is marginal at higher ethanol concentrations (the curve approaches the diagonal as the composition nears the azeotropic value where α = 1).

26.4.2.5 VLE of a multicomponent mixture

While considering the distillation of wine as a binary system of ethanol and water is a useful exercise, many other volatile compounds exist in wines and distillates. Termed congeners, these are the aroma and flavor components of wine that can be carried over to provide the characteristic sensory attributes of the distillate. They are minor components, so it is assumed that they do not affect the volatility of the ethanol–water matrix or that of each other. Because of their low concentrations, their volatility is often better described by Henry’s Law than Raoult’s Law, and is usually a function of the ethanol strength of the liquid in addition to temperature (Chapter 1). When ethanol strength is high, many congeners have lower volatility than ethanol itself, and, conversely, in dilute ethanol solutions some congeners have higher volatility than ethanol (Figure 26.4.2). For example:

• Alcohols such as 2‐methyl‐1‐propanol, 2‐methyl‐1‐butanol, and 3‐methyl‐1‐butanol (i.e., higher alcohols, Chapter 6), commonly called fusel alcohols/oils by distillers, are more volatile than ethanol below about 20% abv.
• Methanol is more volatile than ethanol only above around 80% abv (Figure 26.4.2a).
• For ethyl esters (Chapter 7), ethyl acetate is more volatile than ethanol across the composition range (as is diethyl acetal, Chapter 9), ethyl hexanoate is more volatile than ethanol below about 65% abv, and ethyl lactate is always less volatile than ethanol at any ethanol concentration (Figure 26.4.2b).

Figure 26.4.2 Volatility data as a function of ethanol concentration in wine distillates for (a) representative alcohols and (b) representative esters and acetal. Calculated from data in Reference [15]

This knowledge is crucial to understanding the separation of components by distillation of wine and determines the point at which certain congeners are best separated from ethanol depending on the distillation system, as elaborated below.

26.4.3 Batch and continuous distillation

Production of grape spirit involves the separation and enrichment of alcohol from the predominantly aqueous wine matrix. Distillation of wine (approx. 9–12% abv) is used to produce a product with up to 83% abv in the case of brandy and variants (i.e., flavored spirit containing congeners), or up to 96% abv (the azeotropic point) in the case of neutral grape spirit (spiritus vini rectificatus, SVR) used for fortifying wines such as port and sherry. The two methods used to produce these spirits are batch distillation using a pot still and continuous distillation using one or more fractionating columns [16–19].

26.4.3.1 Reflux and reflux ratio

Condensing and returning of vapors to the system – known as reflux – can improve the efficiency of a separation and reduce the number of stages required to achieve a certain ethanol concentration. The two types of reflux are:

• Internal reflux, where vapors have cooled and condensed before leaving the column.
• External reflux, where vapors that have exited the column are condensed by a partial condenser (dephlegmator) and returned to the still.

Reflux improves the effectiveness of the separation by allowing greater vapor–liquid contact time, which allows the system to more closely approach equilibrium. For external reflux, the volume of distillate returned to the column divided by the volume being taken off as product gives the reflux ratio, which is an important operational factor used to modify the separation of components. A high reflux ratio, which returns a greater amount of externally condensed distillate, gives greater separation of components (especially important for those with similar boiling points) but also increases the energy requirements and distillation time to produce a given amount of product.

26.4.3.2 Batch distillation with a pot still

Batch distillation for producing grape spirit predates the invention of continuous stills. As the name implies, distillation is performed batch‐wise, in a pot still almost always constructed from copper (Figure 26.4.3).¹⁰ A pot still consists of:

• Container (i.e., pot) for the distillation substrate (e.g., wine)
• Heat source
• Column and still head
• Vapor pipe
• Condenser.

Figure 26.4.3 Images of pot stills used for distillation showing (a) a copper pot still with open column, (b) the top portion of the open column, still head, vapor pipe, and condenser (on the next floor above in the still house), and (c) a Cognac charentais pot still consisting of open flame (gas or wood fire, hidden), copper pot (only the top is visible), still head, vapor pipe which passes through a wine pre‐heater (to pre‐heat the following batch of wine), and condenser.

Source: Reproduced with permission of Michael Hage

Double distillation is performed with a pot still to obtain the required alcohol strength and separation of components – that is, a portion of the initial distillate is returned to the pot and redistilled [11, 12, 20, 21]. This method is used for commercial production of wine spirits (e.g., eau de vie, brandy, Cognac, Armagnac, pisco) and grape marc spirits (e.g., grappa and rakia). During both distillation steps, fractions (“cuts”) are taken by switching the flow of distillate from one receiving container to another. Typically, these cuts are classified as “heads” (beginning), “hearts” (middle), and “tails” (end). The hearts cut of the first distillation produces “low wine” (brouillis) – approximately 30% of initial pot volume – which contains 20–50% abv; heads (60% abv) from the first distillation run are less than 1% of pot volume whereas tails (3% abv) are approximately 6% of pot volume (Figure 26.4.4). These cuts contain various congeners relative to their volatilities (see Figure 26.4.2). The heads cut contains a relatively high concentration of ethanol, but is enriched in highly volatile, malodorous compounds like acetaldehyde and ethyl acetate.¹¹ Conversely, the tails cut is low in alcohol and contains low‐volatility congeners such as acetic acid, ethyl lactate, and 2‐phenylethanol. The alcohol in the heads and tails cuts can be recovered by recycling these cuts into the next batch of wine destined for distillation (or simply by diluting with water and redistilling). The first distillation is also referred to as a “stripping run” since it will strip out the majority of the ethanol from water and non‐volatiles; this generally requires recovery of about 35% of the initial volume in the pot.

Figure 26.4.4 Schematic of cuts made during the double distillation process for transforming wine into low wine and then brandy spirit, showing alcohol strength and cut volume (as a % of initial pot content), and how the heads and tails can be recycled (or diluted with water and redistilled). Late hearts cuts made during the second distillation may be redistilled with another batch of low wine if necessary

The batch process requires four first distillations (or a larger pot still, as is often the case in Cognac) to produce enough low wine to fill the pot for the second distillation, where again heads, hearts (sometimes several cuts), and tails are collected. The hearts cut contains the spirit used for brandy – up to about 40% of initial pot volume and about 70–80% abv – whereas the heads (around 1% of pot volume and 75% abv) and tails (about 6% of pot volume and 3% abv) can be recycled, as mentioned above. Late hearts cuts may be redistilled with another batch of low wine. The behavior of congeners during the second distillation is quite different from that of the first, due to the higher ethanol strength of both the low wine and subsequent distillate, and provides greater ability to separate esters and aldehydes (heads cut) and fusel alcohols (tails cut) from the spirit (see Figure 26.4.2). Depending on the pot still design, cool water can also be run through a brandy ball (Figure 26.4.2b) during the second distillation to provide external reflux, further improving the separation of components.

Instead of an open column and brandy ball, a pot still may have a rectifier column containing 20–30 distillation trays (actual stages) that improve internal reflux, as well as a reflux condenser to enhance separation. This configuration is a batch process but overcomes the need for double distillation, and can provide higher strength spirit (around 92% abv) with reduced volume of the cuts taken. Brandy strength spirit can be produced by lowering the reflux ratio with less cooling to the reflux condenser. An Armagnac still¹² is similar to the pot rectifier, having a tray‐filled rectifying column but operating as a continuous still, as described in the next section.

26.4.3.3 Continuous distillation with fractionating columns

In contrast to batch distillation in a pot still, continuous distillation involves a constant feed of wine (or low wine) and input of energy balanced with a continual production of distillates and stillage (waste), that is, a dynamic equilibrium. Continuous distillation can have lower operating costs than batch distillation due to higher rates of throughput and ability to automate, but is also more expensive to install and complex to operate. Continuous stills are usually made of stainless steel components, and a system may utilize one column, as is often the case for brandy production (Figure 26.4.5), or can consist of multiple columns (Figure 26.4.6) to produce highly rectified, neutral SVR [22]. There are many column designs and operational details that are beyond the scope of this chapter and can be found elsewhere [8, 23]. Unlike the open column typical of a pot still, a continuous still column can contain horizontal trays of different designs (e.g., sieve, bubble cap, fixed, or floating valves) or packings (random such as Raschig/Pall rings and saddles, or structured such as gauze and mesh bundles) [5, 24]. These devices within a fractionating column provide very large surface areas for liquid and vapor to interact, increasing internal reflux and improving the vapor–liquid mass transfer efficiency. The distillate may be removed at multiple locations along a column, based on the volatility of congeners as a function of ethanol strength. Highly volatile congeners are taken off overhead, much like with a pot still. Less‐volatile congeners (with higher boiling points) such as fusel alcohols can concentrate on trays lower in the column, below the take‐off point of the desired spirit.

Figure 26.4.5 Single‐column continuous still (analyzer) containing bubble cap trays and smaller column containing supports and packing (brandy column, concentrates heads) used to produce brandy spirit at up to 83% abv. When lower quality wine or lees are used the still can be operated to produce low wine at around 40% abv destined for multicolumn distillation.

Source: Reproduced with permission of Tarac Technologies

Figure 26.4.6 Three‐column continuous still consisting of purifier, rectifier, and methanol columns used to produce 96% abv neutral spirit from approximately 40% abv low wine. Note the alcohol strengths entering and leaving the columns and that live steam is used to heat the purifier (where hydroselection water is also added) but not in the rectifier or methanol column, which feature external reboilers for heating.

Source: Reproduced with permission of Tarac Technologies

A continuous single‐column analyzer with bubble cap trays (Figure 26.4.5) can be used for brandy production. This may be part of a larger multicolumn still arrangement or can stand alone with some variations in design (e.g., use of sieve or valve trays, and ability to produce high‐strength spirit or draw off fusel oils due to tray configuration) [11]. Wine is pre‐heated and enters a few trays below the top of the column, where it is heated by live steam. The bottom section of the column (stripping section) contains more bubble cap trays than the top, where rectification takes place (concentrating section). The more volatile heads (e.g., acetaldehyde) are cooled and separated, and some vapors are returned as external reflux. The remaining vapors containing ethanol and congeners are cooled and sent to a smaller column to further remove heads components, with brandy spirit exiting from the bottom of the column. Depending on the operating conditions (e.g., external reflux, feed rate, heat input) the still can be run flexibly enough to produce different distillate strengths, so if the quality of the feed is high enough¹³ then brandy spirit with an alcohol content >80% abv can be targeted (maximum % abv depends on local regulations). On the other hand, if lees or low‐quality wine are distilled, then the still is run so as to produce a low wine at about 40% abv, which is then further distilled in a multicolumn system such as the one described below.

The three‐column still example in Figure 26.4.6 consists of a purifier column to remove heads, a rectifier column to concentrate the alcohol, and a methanol column to remove methanol. The feed enters near the top few bubble cap trays of a purifier column heated by live steam, and hot hydroselection water is added to the trays above to enable extractive distillation to occur. Thus, the purpose of the hydroselection water (and live steam) is to raise the boiling point of the mixture by diluting the ethanol to about 20% abv or less. This increases the relative volatility of the heads components (refer to Figure 26.4.2) and reduces their miscibility in the diluted solution, allowing them a greater chance to boil off. Heads vapors pass through a small packed column and are returned as external reflux or cooled and separated. The weaker ethanol solution leaves from the bottom of the purifier and enters the rectifier column at a bubble cap tray with about the same ethanol content (feed tray), where it is heated by a reboiler to rectify the ethanol to 96% abv. The trays below the feed point strip ethanol from the dilute solution (stripping section, fewer trays) whereas the trays above concentrate the ethanol and remove other congeners (concentrating section, contains twice the number of trays to enable rectification to 96% abv; compare this with the analyzer column in Figure 26.4.5).

The concentrating section enables separation of so‐called “oils of wine” (i.e., C₃ and C₄ alcohols concentrating at around 75% abv) and “fusel oils” (i.e., C₅ alcohols concentrating at around 40% abv) due to their lower volatility than ethanol at 96% abv (refer to Figure 26.4.2). These congeners collect at a higher temperature (i.e., closer to the reboiler, at around 80–90 °C) on trays lower down in the column and can be drawn off (oils of wine above the feed plate and fusel oils below).¹⁴ Heads are condensed overhead and returned to the rectifier as external reflux or sent to the purifier to be redistilled, and the spent waste exits from the bottom, pre‐heating the low wine entering the purifier.

The ethanol fraction that emerges from the spirit trays near the top of the rectifier can contain a relatively high concentration of methanol (see Chapter 6 for more on methanol).¹⁵ Methanol is more volatile than ethanol at >70% abv, but the difference in volatilities is not that great (refer to Figure 26.4.2), and a very efficient packed column¹⁶ called a demethylizer (or methanol column) is used to obtain the separation. The highly rectified ethanol is gently reboiled as it flows down to exit from the bottom of the methanol column, whereas methanol and other remaining congeners are condensed overhead. Again, heads from the column can be returned as external reflux or recycled back to the purifier.

Variations are possible in the column designs of multicolumn stills, with different numbers of trays and combinations of columns, such as a packed column purifier or a methanol column with trays in the lower region and packing in higher regions. In any case, the principles are the same, and differences in volatility relative to the ethanol strength of the liquid in the column are exploited to remove congeners and concentrate the ethanol. Wine, lees, and other lower quality feedstocks with higher concentrations of unwanted congeners, for example grape marc with higher amounts of aldehydes and methanol, require greater effort to separate the ethanol and produce neutral SVR. In these cases additional columns need to be used, such as a five‐column still with packed and bubble cap tray columns (wash or analyzer columns similar to Figure 26.4.5) to produce low wine and remove volatiles at low ethanol strength, followed by a rectifier with bubble cap trays and two packed methanol columns to produce SVR with a very low congener content.

26.4.3.4 Alcohol reduction in wine

Beyond its role in spirits production, different forms of distillation can feature in the production of reduced alcohol wines [25, 26]. Specialized distillation equipment in the form of an evaporator or spinning cone column use lower temperature distillation under vacuum for alcohol removal, producing an ethanol‐rich distillate that may be fractionally distilled in the conventional way to rectify and recover the ethanol. Permeate obtained from wine that has undergone alcohol reduction by reverse osmosis or other membrane techniques (Chapter 26.3) can also be distilled in the conventional way to remove the ethanol, where the water‐rich remainder is returned to the original wine. This approach has the advantage that only the low molecular weight fraction of the wine will be subjected to distillation, and should result in fewer changes to the wine flavor. Whatever the distillation technique, the principles remain the same and involve mass transfer as a result of liquid and vapor interacting.

26.4.4 Spirit composition and cask maturation

Whereas processing decisions and equipment selection can have many effects on distillate composition, the degree of rectification (ethanol content) is of particular importance because higher rectification indicates lower congener concentrations. Brandy production involves lower rectification, since congeners contribute to the desirable sensory properties of the product, while SVR is used when only the alcohol is required and not any of the flavors, such as when fortifying some ports and sherries. Table 26.4.2 provides an idea of the concentrations of congeners found in brandies (including Cognac and Armagnac) and SVR arising from pot or continuous stills. In particular, there are lower levels of heads components (i.e., acetaldehyde, diethyl acetal, ethyl acetate) and furfural in continuous still spirits, and substantially elevated levels of methanol and heads components in pot still marc spirit. Furthermore, some difference in the concentration of a number of congeners (especially fusel oils and ethyl lactate) between brandies and Armagnacs obtained from continuous stills is evident due to variations in still design. Note that the maximum limits of congeners permitted in the European Union for SVR (96% abv) are markedly lower than typical values observed in other spirits as a result of the multicolumn stills used to produce highly rectified neutral alcohol.

Table 26.4.2 Comparison between generic wine data and the composition of major volatiles in a selection of grape spirits produced by batch or continuous distillation. Data from Section A and references [12] and [27] to [31]

Congener (mg/L of absolute ethanol) a	Pot Still					Continuous Still
	Wine	Brandy	Cognac	Armagnac	Marc	Brandy	Armagnac	SVR b
Acetaldehyde	110	103−202	140−276	191−247	517−767	57−101	59−69	50
Diethyl acetal	445	58−102	58−101	58−81	277−432	21−100	33−54	–
Methanol	370	372−542	102−510	389−408	5769−7917	138−490	392−433	300
1‐Propanol	280	408−495	294−372	277−285	439−697	285−490	269−287	50
2‐Methyl‐1‐propanol	600	385−640	1092−1335	907−1030	626−878	248−352	884−920
2‐Methyl‐1‐butanol	465	1742−2022 c	2797−2910 c	642−729	427−579	985−1342 c	659−669
3‐Methyl‐1‐butanol	2020			2660−2860	1185−2007		2420−2530
Ethyl acetate	370	390−535	440−512	254−331	738−1669	148−225	228−246	13
Ethyl hexanoate	4	6−10	6−9	3	5−14	2−7	3
Ethyl decanoate	3	31−50	15−44	4−6	16−69	12−24	4
Ethyl lactate	3030	55−69	184	340−347	171−257	5−72	390−505
Furfural	90	33−34	24−29	10−14	–	2−4	2	ND d

^a Congeners are often expressed in terms of liters of pure alcohol, so for a spirit that is 40% abv, the listed mg/L value would be multiplied by 0.40 to get the actual mg/L of congener in that spirit. The data for wine are based on a value of 11% abv.

^b Maximum values according to European Union regulations. Total aldehydes expressed as acetaldehyde, total higher alcohols as sum of individuals, and total esters expressed as ethyl acetate.

^c Isomers of 2‐methyl‐1‐butanol and 3‐methyl‐1‐butanol reported together.

^d Not detectable.

Brandy spirit is usually aged in oak casks where some oak components, including aroma compounds are extracted [14, 17]. Reactions that occur during wine aging (Chapter 25) can also potentially contribute to aging of spirits, with some important differences:

• Brandy has a higher pH than wines (4.2–4.4), and thus acid‐catalyzed reactions like ester hydrolysis and formation will occur proportionally more slowly. Interestingly, these reactions appear to occur more rapidly in barrel‐aged brandies than those aged in glass [32], possibly because organic acids that lower the pH can be extracted from oak wood.
• Brandy will have a lower transition metal concentration, a lower phenolic content, and negligible SO₂ content as compared to most wines, and thus will consume oxygen much more gradually. While aldehyde concentrations are typically higher in brandy than in wine, this is because brandy is often stored semi‐aerobically. When both wine and brandy are exposed to large amounts of air, aldehydes will accumulate in wine more rapidly.

The combination of aging in new and old oak of different origins for various periods of time (decades in some cases) and blending of different spirits produces the characteristics of the commercial spirit. Additionally, most spirits will undergo reduction with distilled water to dilute it from cask to bottle strength (often step‐wise over a period of time) [20, 21]. Very old brandies cellared in oak for many years will naturally lose alcohol over time due to evaporation, and may already be at the appropriate strength at the time of blending and bottling so no reduction is required.

Beyond ethanol, the aroma of Cognac is reported to originate largely from fermentation derived congeners – particularly branched‐chain ethyl esters, and fusel alcohols and their corresponding aldehydes – and wood‐derived odorants. The only high‐odor activity odorant of likely grape origin reported in Cognac was β‐damascenone [33], and concentrations over 100 μg/L have been reported in some brandies [34], which is an order of magnitude higher than most wines (Chapter 8). Such elevated concentrations of β‐damascenone likely arise due to its liberation from glycosylated precursors as a result of heating of the wine (Chapter 23.1).

References

1. 1. Huang, H.T. (2000) Evolution of distilled wines in China, in Part V: Fermentations and food science , Cambridge University Press, Cambridge, UK, pp. 203–231.
2. 2. Sherwood, T.F. (1945) The evolution of the still. Annals of Science , 5 (3), 185–202.
3. 3. Liebmann, A.J. (1956) History of distillation. Journal of Chemical Education , 33 (4), 166.
4. 4. Lea, A.G.H. and Piggott, J.R. (eds) (2003) Fermented beverage production , 2nd edn, Kluwer Academic/Plenum Publishers, New York.
5. 5. Bennett, B.L. and Kovak, K.W. (2000) Optimize distillation columns. Chemical Engineering Progress , 96, 19–34.
6. 6. Luyben, W.L. (2006) Fundamentals of vapor–liquid phase equilibrium (VLE), in Distillation design and control ssing aspen simulation , 2nd edn (ed. Luyben, W.L.), John Wiley & Sons, Inc., Hoboken, New Jersey, pp. 1–26.
7. 7. Kiss, A.A. (2013) Basic concepts in distillation, in Advanced distillation technologies , John Wiley & Sons, Ltd, Chichester, UK, pp. 1–35.
8. 8. Jacques, K.A., Lyons, T.P., Kelsall, D.R. (eds) (2003) The alcohol textbook , 4th edn, Nottingham University Press, Nottingham.
9. 9. Meirelles, A.J.A., Batista, A.C., Scanavini, H.F.A., et al. (2009) Distillation applied to the processing of spirits and aromas, in Extracting bioactive compounds for food products (ed. Meireles, M.A.A.), CRC Press, Boca Raton, FL, pp. 75–135.
10. 10. Buglass, A.J. (2011) Handbook of alcoholic beverages: technical, analytical and nutritional aspects , John Wiley & Sons, Ltd, Chichester, West Sussex, UK.
11. 11. Guymon, J.F. (1974) Chemical aspects of distilling wines into brandy, in Chemistry of winemaking , American Chemical Society, Washington, DC, pp. 232–253.
12. 12. Léauté, R. (1990) Distillation in alambic. American Journal of Enology and Viticulture , 41 (1), 90–103.
13. 13. Silva, M.L., Macedo, A.C., Malcata, F.X. (2000) Review: steam distilled spirits from fermented grape pomace. Revision: Bebidas destiladas obtenidas de la fermentación del orujo de uva. Food Science and Technology International , 6 (4), 285–300.
14. 14. Tsakiris, A., Kallithraka, S., Kourkoutas, Y. (2014) Grape brandy production, composition and sensory evaluation. Journal of the Science of Food and Agriculture , 94 (3), 404–414.
15. 15. Martin, A., Carrillo, F., Trillo, L.M., Rosello, A. (2009) A quick method for obtaining partition factor of congeners in spirits. European Food Research and Technology , 229 (4), 697–703.
16. 16. Piggot, R. (2003) From pot stills to continuous stills: flavor modification by distillation, in The alcohol textbook , 4th edn (eds Jacques, K.A., Lyons, T.P., Kelsall, D.R.), Nottingham University Press, Nottingham, pp. 255–266.
17. 17. Buglass, A.J., McKay, M., Lee, C.G. (2011) Brandy, in Handbook of alcoholic beverages: technical, analytical and nutritional aspects (ed. Buglass, A.J.), John Wiley & Sons, Ltd, Chichester, West Sussex, UK, pp. 574–594.
18. 18. Buglass, A.J., McKay, M., Lee, C.G. (2011) Grape and other pomace spirits, in Handbook of alcoholic beverages: technical, analytical and nutritional aspects (ed. Buglass, A.J.), John Wiley & Sons, Ltd, Chichester, West Sussex, UK, pp. 595–601.
19. 19. Bertrand, A. (2003) Armagnac, brandy, and Cognac and their manufacture, in Encyclopedia of food sciences and nutrition , 2nd edn (ed. Caballero, B.), Academic Press, Oxford, pp. 584–601.
20. 20. Cantagrel, R. and Galy, B. (2003) From vine to Cognac, in Fermented beverage production , 2nd edn (eds Lea, A.G.H. and Piggott, J.R.), Kluwer Academic/Plenum Publishers, New York, pp. 195–212.
21. 21. Bertrand, A. (2003) Armagnac and wine‐spirits, in Fermented beverage production , 2nd edn (eds Lea, A.G.H. and Piggott, J.R.), Kluwer Academic/Plenum Publishers, New York, pp. 213–238.
22. 22. Technical Symposium on Australian grape spirit and brandy production (1978) Waite Agricultural Research Station, Urrbrae SA, Adelaide, SA, 12–13 June 1978, The Australian Wine Resarch Institute.
23. 23. Kiss, A.A. (2013) Design, control and economics of distillation, in Advanced distillation technologies , John Wiley & Sons, Ltd, Chichester, UK, pp. 37–65.
24. 24. Pilling, M. and Holden, B.S. (2009) Choosing trays and packings for distillation. Chemical Engineering Progress , 105, 44–50.
25. 25. Pickering, G.J. (2000) Low‐ and reduced‐alcohol wine: a review. Journal of Wine Research , 11 (2), 129–144.
26. 26. Schmidtke, L.M., Blackman, J.W., Agboola, S.O. (2012) Production technologies for reduced alcoholic wines. Journal of Food Science , 77 (1), R25–R41.
27. 27. Bertrand, A. (1989) Role of continuous distillation process on the quality of Armagnac, in Distilled beverage flavour: recent develoments (eds Piggott, J.R. and Paterson, A.), Ellis Horwood Ltd., Chichester, UK, pp. 97–115.
28. 28. Panosyan, A.G., Mamikonyan, G.V., Torosyan, M., et al. (2001) Determination of the composition of volatiles in cognac (brandy) by headspace gas chromatography–mass spectrometry. Journal of Analytical Chemistry , 56 (10), 945–952.
29. 29. López‐Vázquez, C., Herminia Bollaín, M., Berstsch, K., Orriols, I. (2010) Fast determination of principal volatile compounds in distilled spirits. Food Control , 21 (11), 1436–1441.
30. 30. EC Regulation No. 110/2008 (2008).
31. 31. Etiévant, P.X. (1991) Wine, in Volatile compounds in foods and beverages , 1st edn (ed. Maarse, H.), Marcel Dekker, New York, pp. 483–546.
32. 32. Onishi, M., Guymon, J.F., Crowell, E.A. (1977) Changes in some volatile constituents of brandy during aging. American Journal of Enology and Viticulture , 28 (3), 152–158.
33. 33. Uselmann, V. and Schieberle, P. (2015) Decoding the combinatorial aroma code of a commercial Cognac by application of the sensomics concept and first insights into differences from a German brandy. Journal of Agricultural and Food Chemistry , 63 (7), 1948–1956.
34. 34. Sefton, M.A., Skouroumounis, G.K., Elsey, G.M., Taylor, D.K. (2011) Occurrence, sensory impact, formation, and fate of damascenone in grapes, wines, and other foods and beverages. Journal of Agricultural and Food Chemistry , 59 (18), 9717–9746.

Notes

1 Mole fractions for liquid and vapour are denoted by X and Y, respectively. The expression for volatility is explained in Table 26.4.1.

2 The kinetic energy of molecules in the liquid phase results in some molecules leaving the surface of the liquid to form a vapor. Molecules in the vapor phase also condense and return to the liquid phase. At equilibrium, the rate of molecules evaporating is balanced by the rate of molecules condensing, giving a certain vapor pressure. Input of energy (heat) increases both the temperature of the liquid and the kinetic energy of the molecules, meaning they can evaporate more easily, which increases the vapor pressure above the liquid and translates into higher volatility. At a given temperature a compound with a lower boiling point has a higher vapor pressure and is more volatile than one with a higher boiling point. Exploiting these differences enables separation by distillation.

3 If enough heat is added a liquid will reach boiling point, which occurs when the vapor pressure has reached atmospheric pressure. That is why water will boil at a lower temperature atop a mountain, for example, compared to sea level where the atmospheric pressure is higher, and why pressure cookers can achieve higher temperatures and faster cooking times than conventional boiling. Conversely, boiling points are lowered by applying a vacuum because the “system” pressure is lower than atmospheric pressure.

4  The calculations are beyond the scope of this chapter but involved the use of readily available constants, Antoine and van Laar equations, and Raoult’s and Dalton’s Laws.

5 Note the relative symmetry of the methanol-ethanol VLE diagram as compared to the ethanol–water diagram (Figure 26.4.1).

6 In this situation a modified version of Raoult’s Law can be used which introduces an activity coefficient, the ratio between actual and ideal partial vapor pressures, for each component to account for the altered behavior.

7 Conversely, non-ideal behavior of liquid mixtures showing a negative deviation from Raoult’s Law can lead to a maximum boiling azeotrope, where the vapor pressures are lower than expected if the solution was ideal, and the boiling point of the mixture at a specific composition is higher than that of either pure component.

8 This can be thought of as if the vapors were condensed once the still had reached equilibrium and the resulting liquid was distilled again, this time boiling at a lower temperature and producing more enriched vapor. In reality, this occurs within the distillation system with separate stages (mass transfer zones) where rising enriched vapors interact with cooler liquid higher within the still, enabling more ethanol to transfer to the vapor phase while more water condenses into the liquid phase at each stage. A greater number of stages results in more enriched vapor exiting the still.

9 Note the relative volatility of a non-ideal ethanol–water solution strongly depends on the liquid composition because the activity coefficients change with composition. As such, the relative volatility of ethanol with respect to water varies markedly across the composition range.

10 Copper provides several advantages for use in stills: it is highly malleable during manufacturing, it has excellent thermal conductivity, and it can strip out malodorous thiols from the vapors.

11 More acetaldehyde in a base wine would thus result in a larger heads fraction and a less efficient distillation process. Brandy producers will typically not use SO₂ additions before fermentation to avoid high levels of bound acetaldehyde accumulation (Chapter 22.1).

12 Strict regulations govern the production of Armagnac and Cognac. Batch distillation in a pot still is mandatory for Cognac whereas Armagnac can be produced using a pot still, much like those used for Cognac, or more commonly with a continuous Armagnac still (armagnacais).

13 Production of quality brandy typically requires wine that is made especially for distillation, as is the case for Cognac and Armagnac. Wines of neutral grape varieties from a clean fermentation (one where minimal fusel oils, aldehydes, and sulfur compounds are produced), having moderate alcohol levels, high acid (microbial stability, better aroma quality), and no added SO₂, are distilled soon after fermentation. Some lees may be included during distillation for additional flavor (e.g., from fatty acids and their ethyl esters).

14 Components such as heads and fusel oils that are separated and drawn off during distillation also contain ethanol. Collectively termed feints, these components are sent to a feints still to recover the ethanol prior to disposal of the remaining waste.

15 Although higher trays also contain high-strength alcohol, the spirit is not drawn from those trays because it contains a small amount of the more volatile heads components not removed by the purifier, and these are quite soluble at high ethanol strength.

16 Instead of actual trays, the number of which determines the ability to separate components, packed columns are described in terms of their height, which determines the number of equilibrium stages. The height equivalent of a theoretical plate (HETP) is used to express efficiency of a packed column; the smaller the HETP, the higher the theoretical plate count and the better the separation for a given column length.