Chemical Aspects of Distilling Wines Into Brandy

7/29/2019 Chemical Aspects of Distilling Wines Into Brandy

1/22

11

Chemical Aspects of Distilling Wines intoBrandyJAMES F . GUYM ONUniversity of Cali fornia , Davis, Cal i f . 95616

Wines for brandy distillation should be made from whitegrapes by fermentation of juice separated from pomacewithout adding SO2 (to minimize aldehyde accumulation)at as low temperature as is practicable (less than 75F , tominimize fusel oil formation) and should be distilled as soonas fermentation is complete. In California, wines are dis-tilled into brandy in continuous plate columns arranged forseparating a low-boiling heads fraction to eliminate alde-hydes. Higher alcohols (fusel oil) are the most abundantgroup of congeners, and ethyl esters of C8 to C12 fattyacids are next most abundant provided yeast cells are pres-ent in the wine distilled. Fatty acid esters and fusel alcoholsaccumulate in column tray liquids at 100 to 135 proof anddistill most rapidly in the earliest fractions of simple potdistillation.

Brandy is a distillate of wine so its i n i t i a l chemical composition dependso n the compounds present in wine being sufficiently volati le to d i s t i l l .A l t h o u g h e t h y l alcohol and water are the two major components of any

d i s t i l l e d spirit, aroma and flavor character depend on a multitude ofm i n o r compounds usually referred to as congeners or congenerics. Someof the congeners are derived directly from the grapes or other fruits thatmay be used for brandy. The most abundant congeners of brandy arem i n o r products of alcoholic fermentation derived p r i m a r i l y from sugarsbut in part from components of the f r u i t . D i s t i l l a t i o n conditions affectthe relative quantities of m i n o r compounds recovered in the distillate, andsome arise from chemical reactions during d i s t i l l a t i o n such as heat-induced degradative changes.

Grape brandies and some f ru i t brandies are tradit ionally stored inoak casks for varying periods of time for aging or maturation. D u r i n g

232

DownloadedbyMITonJanuary5,

2013|http://pubs.acs.org

PublicationDate:November1,

1974

|doi:10.1

021/ba-1974-0137.c

h011

In Chemistry of Winemaking; Webb, A.;Advances in Chemistry; American Chemical Society: Washington, DC, 1974.


2/22

11. G U Y M O N Distilling Wines into Brandy 233aging in oak many substances are extracted from the wood, some ofw h i c h react w i t h compounds i n i t i a l l y present, especially e t h y l alcohol.Thus, the aging process produces a complexity of minor constituentsw h i c h generally improves the palatability and enhances the aroma andflavor.

The overall chemical composition of brandy is derived from four general sources: the fruit , alcoholic fermentation, disti llation, and aging inwood. The scope of this discussion is l imited to chemical aspects of thecomponents of wine and their behavior during disti llat ion.

The term brandy as used here refers to distillates of grape winebefore aging. Federa l regulators require that distillates derived fromother fruits must show the name of the fruit, for example, pear brandy.Wines for Disti llation

The production of grape wines destined for distillation into brandyis essentially the same as that for dry white table wines. However thereare some desired characteristics of d i s t i l l i n g wines w h i c h are relativelymore significant than those of table wines produced for consumption.W h i t e grape varieties are preferred to red or black varieties even thoughw h i t e wines can be prepared from red grapes by separating the juicep r i o r to fermentation. Red varieties tend to impart a measure of coarseness to the distillates and, compared w i t h white varieties, generally leadto greater formation of higher alcohols during fermentation. This seemsto occur whether the fruit is processed as white wines (fermentation ofseparated juice) or as red wines (fermentation of the entire must cont a i n i n g juice, skins, seeds, and pu lp) . Thus, the preferred procedure forproducing d i s t i l l i n g wines is to use non-pigmented grapes and to separatethe juice from the marc or pomace prior to fermentation.

Sulfur Dioxide and Aldehydes. Sulfur dioxide is commonly addedboth before and after fermentation in preparing white table wines. It isan effective antioxidant as w e l l as a selective i n h i b i t o r of unwantedmicroorganisms. However, sulfur dioxide, as the bisulfite ion in solution,combines w i t h aldehydes, especially acetaldehyde, during fermentationg i v i n g an accumulation of aldehydes in the bound form of aldehyde-sulfurous acid.

C H 3 C H O + H S 0 3 - ^ C H 3 C H O H S O 3 -The accumulation of aldehydes as w e l l as the presence of sulfurdioxide itself is undesirable in d i s t i l l i n g wines for reasons given later.

W e have experienced no difficulty in producing sound wines (that isw i t h o u t bacterial contamination or significant oxidation) if good qualityf r u i t is used. However, fruit that is significantly damaged by rain or




1974

|doi:10.1

021/ba-1974-0137.c

h011



3/22

234 C H E M I S T R Y OF W I N E M A K I N Gother conditions whereby rot or molds have developed and increasedp o l y p h e n o l oxidase y i e l d wines w h i c h oxidize and brown badly and arevulnerable to Acetobacter and lactic acid bacteria. In this case it isadvisable to add 50-75 ppm of sulfur dioxide prior to fermentation.A m e r i n e and Ough (I ) showed that moderate levels of sulfur dioxide(50 pp m) increased aldehyde accumulation only slight ly over winesfermented without sulfur dioxide.

Acetaldehyde content reaches a maximum at the time of fermentat i o n that shows the most rapid rate of carbohydrate dissimulation, it fallsto a low level at the end of fermentation, and then slowly increases (2,3).A l s o aldehydes are formed by oxidation as evidenced by browning ofw i n e when exposed to air. Fo r these and other reasons, distillation shouldbe carried out as soon as possible after fermentation is complete.Higher Alcohols. The most abundant, volatile minor products ofalcoholic fermentation are the higher alcohols or fusel alcohols. The mostimportant are isoamyl (3-methyl-l-butanol), d-active amyl (2-methyl-l-butanol) , isobutyl (2-methyl-l-propanol) , and -propyl (1-propanol)alcohols. It is now recognized (4, 5, 6,7) that these higher alcohols areformed by decarboxylation of particular -keto acids to y i e l d the corresponding aldehydes, and these in turn are reduced to the alcohols in amanner analogous to the formation of e t h y l alcohol from pyruvic acid.

- C 0 2 + 2 HR C O C O O H R C H O > R C H 2 O HH i g h e r alcohol formation by yeasts has long been attributed to the

transamination of exogeneous amino acids to y i e l d the analogous ketoacid. For example:

(transaminase)leucine + -ketoglutaric acid > glutaric acid + (1)-ketoisocaproic acid(decarboxylase)

-ketoisocaproic acid isovaleraldehyde + C O 2 (2)(alcohol dehydrogenase)isovaleraldehyde + N A D H + H+ (3)isoamyl alcohol + N A D +

However, now it is w e l l established that the appropriate a-ketoacids giving rise to a particular higher alcohol arise mostly from carbohydrate sources through the synthetic pathways by w h i c h yeast synthesizes its amino acid requirements.




1974

|doi:10.1

021/ba-1974-0137.c

h011



4/22

11. G U Y M O N Distilling Wines into Brandy 235Peynaud and Guimberteau (8) estimated that no more than one-sixth

of the leucine and valine in grape musts assimilated during fermentationgave rise to isoamyl and isobutyl alcohols, and since these amino acidsare low in grape musts, the formation of higher alcohols by amino aciddegradation is negligible. Instead, nearly al l of the higher alcohols arederived from carbohydrate. However, Reazin et al. (9), in 1 4 C tracerstudies during fermentation of grain mashes, concluded that the majorproportions of higher alcohols were derived from amino acids, and lesserproportions were derived from carbohydrates.

Since the higher alcohols, especially the two amyl alcohols, constitute the most abundant minor components or congeners of brandy andother distil led spirits, they have a very significant effect on the sensoryand taste character. Considerable variation in the concentrations of fuselo i l alcohols in wines (10) and brandies (11) occurs in practice. Factorsi n f l ue n c i n g the levels of higher alcohols produced during fermentationinclude composition of raw material, yeast strain (12), temperature offermentation, suspended matter, and aeration (13, 14). In practice, it isdesirable to minimize the formation of higher alcohols, especially theamyls. Although precise control is not f u l l y feasible, highest levels areformed at medium fermentation temperatures of about 75 F (7, 12, 15,16). The presence of suspended matter and aeration stimulate fusel oi lformation (14). So the practice of draining the relatively clear free-runjuice several hours after crushing is preferred to using the juice immediately after pressing, w h i c h yields more turbid l i q u i d musts.

Several workers have reported the amounts of higher alcohols foundi n wines (8, 12, 17). W h i l e differences as much as fivefold occur considering diverse genera and strains of yeast, the variation is generally lessfor strains of Saccharomyces cerevisiae. Guymon and Heit z (JO) founda mean fusel oil content of 25 grams/100 liters (250 ppm) for 120 samplesof Cal ifornia white table wines and 29 grams/100 liters for 130 samplesof red table wines. In our pilot plant scale production of experimentald i s t i l l i n g wines, we obtain fusel o il levels ranging from about 16-30grams/100 liters for white varieties w i t h an average of 20. W i t h redvarieties, the levels may attain 40 or more grams/100 liters.

The appropriate fusel oi l content of brandy distillate, though comm e r c i a l producers are not in f u l l agreement, ranges from 65 to 100grams/100 liters at 100 proof (50 vol % ) . If 100 grams/100 liters isdesired as the upper l i m i t , a d i s t i l l i n g wine containing 12 vol % e t h y la l c o h o l should thus contain no more than 24 grams of higher alcohols per100 liters, assuming al l of the higher alcohols are recovered i n the d i s t i l l a t i on process.

Fatty Acids and Their Ethyl Esters. The normal aliphatic acids andt h e i r e t h y l esters are often the second most abundant group of congeners




1974

|doi:10.1

021/ba-1974-0137.c

h011



5/22

236 C H E M I S T R Y OF W I N E M A K I N Gfound in distil led spirits. After e t h y l acetate, generally caprylate ( C 8 ) ,caprate ( C i 0 ) , and laurate ( C i 2 ) esters predominate (18), although thosew i t h both fewer and more carbon atoms are found. The amount of suspended yeast cells in the wine at the time of distillation significantlyaffects the amount of fatty acids and esters in the distillate. The highermolecular weight acids and esters are closely bound to the cells of yeastwhereas those w i t h lower molecular weight tend to be secreted intosolution (19).

Thus, the concentrations of fatty acids and esters found in brandydistillates are greatly affected by the nature of the wine at the time ofd i s t i l l a t i o n , particularly the time interval between fermentation andd i s t i l l a t i o n since most of the yeast cells settle out fairly quickly afterfermentation has ceased. Obviously the degree of resuspension of thesettled lees into the wine when distilled affects the amount of fatty acidsand esters recovered in the distillates. The method and techniques ofd i s t i l l a t i o n are also very important since this class of congeners, havingrelatively high boiling points and weak solubility in water, exhibit wideranges of volatility as affected by the alcohol content of the l i q u i dv o l a t i l i z e d .

The desirable properties of d i s t i l l i n g wines include wines: a) madef r o m white varieties by fermentation of separated juice clarified as muchas practicable, b) fermented without addit ion of sulfur dioxide, c) fermented w i t h strains of yeast w h i c h form comparatively low amounts offusel alcohols at temperatures below 75 F , and d) distilled as soon aspossible after fermentation.Disti llation

M a n y processes are used throughout the w o r l d for d i s t i l l i n g winein to brandy. Continuous distillation in plate columns, discontinuous orbatch distillation using pot stills (with and without adjunct rectifyingcolumns), and even distillation apparatuses i n v o l v i n g both continuousand discontinuous aspects are used.

I n France, the famous cognac brandy is distilled in simple direct-firedpot stills. The Mthode Charentaise (20) includes two successive distillations. In the-first, wine is distilled into low wines ( brouillis ) containing26-30% alcohol or approximately a threefold concentration. The lowwines are redistilled, and three fractions are collected, i.e., a small headscut amounting to about 1 vo l % of the charge; the brandy coeur as thea l c o h o l content decreases from about 80 to 60 vol % average 70 vol % ;and the tails (seconde) to recover the remaining alcohol in the charge.The heads and tails fractions are recycled.

I n the Department of Gers, the Armagnac brandies are disti lled inan apparatus known as the Verdier systme. W i n e , preheated by con-




1974

|doi:10.1

021/ba-1974-0137.c

h011



6/22

11. G U Y M O N Distilling Wines into Brandy 237densing vapor, flows continuously to the top of a short plate columnand is partly stripped of alcohol as it transverses the column and flowsin to the top section of a two-tiered, direct-fired boiler. The vapor generated in the lower section by the furnace is caused to bubble through thel i q u i d layer in the top section an d on up the plate column. Vapor fromthe top (feed) tray is condensed by the preheater and condenser intobrandy containing 50-54 vol % alcohol. Intermittently most of the spentw i n e in the lower level of the boiler is drawn off to waste, and then thesame volume of l i q u i d from the upper level is drawn into the lower part.Tims the process is both continuous and discontinuous.

I n Californi a, distillation is usually carried out in a continuous singleor split column unit as shown in Figure 1. The brandy distillate, at amaximum of 170 proof as required by federal regulations for standardsof identity, is drawn as a l i q u i d sidestream from a tray in the middle orupper region of the concentrating section. A small portion of the overhead distillate from the vent condenser is withdrawn as a heads cut at arate of about 5-15% of the brandy rate to remove most of the low-boilingaldehydes, sulfites, esters, etc. Some producers also separate anothersmall fraction (ca. 10-20% ) of the product rate as a low oils or high-b o i l i n g stream from a level one or two trays below the brandy tray to

OverheadProduct

OverheadProduct

Stillage Steam

SINGLE COLUMNStillage

DOUBLE OR SPLIT COLUMN

Figure I . Single and split columns for brandy distillation




1974

|doi:10.1

021/ba-1974-0137.c

h011



7/22

238 C H E M I S T R Y OF W I N E M A K I N Greduce the level of fusel oi l in the product. The concentration of higheralcohols in the tray liquids of a concentrating section producing brandyat 168-170 proof is usually at its maximum at two trays below theproduct level (21). This maximum concentration of higher alcohols isabout 2 to 2Vz times that of the product and occurs where the tray l i q u i dindicates approximately 12 0 -1 35 proof. Thus the separation of a lowoils fraction at a 20% rate should reduce the fusel o il content of thebrandy by about 30%. In practice it is usually less. The disadvantageof this technique is that the amount of brandy, produced from primequality wine, is reduced by the amount of low oils fraction removed w h i c his usually redistilled to recover the alcohol as neutral spirit.

Most distillation units in Cal ifornia have been installed w i t h thecapability of producing high proof wine spirit (189-19 2 proof ). Ac c o r d i n g l y the concentrating sections may contain 20-40 bubble cap trayswhereas only three or four ideal trays are required to concentrate alcoholf r o m 10 to 85 vol % (the strength of beverage brandy distillate). Inbeverage brandy production practice, perhaps two-thirds of the trays liebetween the feed and product trays, and the remaining one-third of thetrays at the top are used for concentrating the low-boil ing impuritiesin to the heads cut. Recent installations generally provide for at least12- inch tray spacings.

The stripping or wine section typically contains 18-22 sieve orperforated trays w i t h 15-18 in . tray spacing. Direct or open steam is theusual heat source although one major brandy producer uses reboilers.C o l u m n diameters are commonly 5-6 feet though some larger and smallerare in use. Feed rates to the stripping section are typically in the range3000-8000 gallons/hr.

Some recent developments in design of d i s t i l l i n g units include theuse of valve and sieve trays in l i e u of bubble caps in the concentratingsections, air-cooled instead of water-cooled condensers, stainless steel inl i e u of copper construction, energy economizers w h i c h incorporate athermocompressor to transfer heat energy from the bottoms or stillageto the entering steam, and location of d i s t i l l i n g units entirely outside ofbuildings. However, most of these developments pertain to the moreelaborate recovery of neutral wine spirit from residue materials.

Some three- and a few four-column units are used. The t h i r d columnof multicolumn units is generally an aldehyde concentrating column inw h i c h the heads cut from the brandy concentrating column is concentrated (Figure 2) as much as 20-fold; the alcohol, stripped of aldehydesand other low-boiling components, is recycled to the main unit. Anotherthree-column arrangement provides for the brandy to be distilled as a topproduct without a heads cut; the entire brandy is then fed into an aldehyde stripping and concentrating column. The low-boiling substances




1974

|doi:10.1

021/ba-1974-0137.c

h011



8/22

11. G U Y M O N Distilling Wines into Brandy 239

co\c.HEADSDSPLITCOLUMNPLUSANALDEHYDECONCENTRATNGCOLUMN1 Adehydeseparatonfromtheprincpepoduct sefectedinthemnconcentrangsecton2Areatveylargeheadscu fromthemncoumnstaken 10to20%3 Opensteamcanbeusedintheadehydecoumn4 Themaopart otheacoho ntheheadsfedtotheadehydecoumnsconanedinthebotomandisreurnedtothaleve otheprmaryunt a whchtheacohoconensareapproxmaeyequaOnya potono theoveral acoho

in thedstllingmatera (10to20%isredstlledFigure 2. Split column with a heads concentrating column

are thus removed as the top product of the aldehyde column and thebrandy is removed as the bottom stream. A reboiler, to avoid d i l u t i o n , istherefore required to supply vapor ( Figure 3 ).

Some ingenious designs have been developed for producing beverage brandy w i t h asubstantial removal of the amyl alcohols withoutexceeding 170 proof in the primary distillate. In one such case alowoils cut, about 20% of the brandy rate, is fed into a fusel oi l concentratingc o l u m n to concentrate the alcohol in the low oils from approximately130 to 190 proof. A fusel oil cut from this column is thus sufficientlyconcentrated to separate into oil and aqueous layers in aconventionalfusel oil decanter (22). The 190 proof alcohol (thus very low in higheralcohols) is recycled to the main column. Another arrangement effectsa partial separation of higher alcohols, especially the amyls, by concentrating the brandy w e l l above 170 proof in the primary concentratingc o l u m n w h i c h permits some separation of fusel oil in its decanter; thebrandy stream is subsequently reduced to 170 proof or less by blendingi t w i t h alow proof stream drawn from an appropriate plate low in theheads concentrating column w h i c h , of necessity, isheated w i t h opensteam.

O n l y asmall amount of pot-still brandy is produced in California.It is used primarily for blending, in small percentages, w i t h continuouss t i l l brandy after aging. Equipment used for this purpose consists of pot-rectifiers. In at least one case, the fractionating column is so arrangedthat it can serve either as the concentrating section for the pot or as apart of a continuous s t i l l unit.




1974

|doi:10.1

021/ba-1974-0137.c

h011



9/22

240 CHEMISTRY OF WINEMAKING

2l d4J ^

Pied-

ConcentratingSection

FeedAldehydeI Section

StrippingSection

OpenSteamJ St i l log

Condeneate LProductHeat Inierchanger Product

Figure 3. Single column with aldehyde separating column heated byreboerbrandy is bottom productAldehydes and Other Low-Boiling Components. As mentioned, a

l o w - b o i l i n g fraction, called heads, is normally taken from the vent condenser during the distillation of wine into brandy. The p r i n c i p a l i m p u r i ties removed are acetaldehyde, d i e t h y l acetal, e t h y l acetate, and acetalde-hyde-sulfurous acid.The heads fraction may be h i g h l y acidic and corrosive if sulfurdioxide or bisulfite is used in the production or preservation of thew i n e . The reversible reactions involved are:

H 2 0 + S 0 2 H S O 3 - + H+H S O 3 - + C H 3 C H O C H 3 C H O H S O 3 -

The acetaldehydesulfurous acid compound has the properties of asulfonic acid w i t h CS linkage rather than an ester structure as onceassumed. It is properly 1-hydroxyethane sulfonic acid (23, 24) and ish i g h l y acidic (25). Samples of concentrated heads from commercialaldehyde columns having aldehyde contents of 5-13% gave p H valuesof 0.7-0.9 and contained high levels of copper (25).




1974

|doi:10.1

021/ba-1974-0137.c

h011



10/22

11. G U Y M O N Distilling Wines into Brandy 241W h i t m o r e (26) and Deibner and Bnard (27) claim that sulfonic

acids are capable of d i s t i l l i n g intact. However, it would seem more l i k e l ythat at the temperature of the d i s t i l l i n g process, 1-hydroxyethane sulfonicacid in the wine would dissociate into the very volatile free acetaldehydeand therefore rapidly d i s t i l l along w i t h free sulfur dioxide and recombineat the top level of the fractionating column.

A n o t h e r reaction of interest is the production of acetal by the reactionof 2 moles of ethanol and 1 mole of acetaldehyde. This reaction isstrongly catalyzed by hydrogen ion and during distillation depends

[H+]C H 8 C H O + 2 C 2 H 6 O H ^ C H 3 C H ( O C 2 H 5 ) 2 + H 2 0

on the presence of bisulfites or the 1-hydroxyethane sulfonic acid toestablish the necessary acidity. E q u i l i b r i u m is rapidly established at lowpHs and the degree of acetalization obviously depends on the concentrations of the two major components, ethanol and water. However,attempts to determine an equilibrium constant are unsuccessful presumably because the reaction proceeds in two steps and the rate-determ i n i n g step is the formation of hemiacetal from one mole each of ethanoland acetaldehyde (24).

E t h y l acetate is a product of yeasts and a normal component ofw i n e . Its level can be increased by Acetobacter contamination, althoughmost wines showing excess volatile ( acetic ) acid do not necessarily cont a in excess e t h y l ester i n i t i a l l y . It is quite possible to obtain brandy ofn o r m a l composition and quality by continuous disti llation of newly fermented wine containing excess acetic acid, e.g., 0.1%. On the otherhand, e t h y l acetate can be formed in continuous columns, particularly ifthe distillation conditions provide for a relatively high ethanol concent rat ion on the feed tray or immediately below. Since acetic acid is weaklyvolatile in all mixtures of ethanol and water, it does not appreciably d i s t i l lupward. Therefore there is no opportunity for acetic acid to combinew t i h ethanol in tray liquids normally of high ethanol concentration.

E t h y l acetate is the major low-boiling impuri ty of heads fractionsf r o m continuous columns if bisulfites are absent in the d i s t i l l i n g material.A heads fraction from a typical brandy column usually contains less than1% of these volatile impurities, although the concentrated heads froman aldehyde-concentrating column may contain as much as 10-15%aldehydes. In either case, e t h y l alcohol is the major component of theheads cut, and its recovery in usable form has been a troublesome processing problem.

A simple method, based on the strongly reducing conditions duringalcoholic fermentation, was developed in our laboratory (25, 29, 30).The heads l i q u i d is recycled into an actively fermenting dilute grape




1974

|doi:10.1

021/ba-1974-0137.c

h011



11/22

242 C H E M I S T R Y OF W I N E M A K I N Gjuice or sweet pomace wash, i n i t i a l l y containing sugar equivalent to about10 B r i x . The titratable aldehydes, including acetal, are very effectivelyreduced to alcohols, and apparently about half of the e t h y l acetate presentis metabolized. Guymon and Nakag iri (3) found that acetaldehyde orits equivalent concentration of acetal added before fermentation in excessof about 0.1% caused important delays in fermentation. But up to 0.35%acetaldehyde or 0.9% acetal could be added to a vigorous fermentationw i t h o u t significant effect on fermentation rates.

E t h y l acetate, added prior to fermentation in amounts up to 3%,caused delays in the onset of fermentation proportional to the amountadded, but once begun, the fermentation rate was unaffected. W h e n thesame levels of e t h y l acetate were added to an active fermentation, thefermentation period was sluggish.

This simple method of heads disposal by refermentation eliminatedthe earlier deterrent to taking an adequate cut to control volatile aldehydes to the desired level, now considered to be approximately 10 ppmor less in new brandy distillates.

O t h e r aliphatic aldehydes have been identified and quantitativelymeasured in various brandies (31, 32). These include small amounts offormaldehyde, propionaldehyde, isobutyraldehyde, isovaleraldehyde, furf u r a l , etc. Furfural is a normal component of pot-still distillates, but itspresence in continuous s t i l l brandies is negligible before aging in wood.

Higher Alcohols and Fatty Acid Esters. Because of the remarkableu t i l i t y of gas chromatography for analysis of minor volatile components,an extensive literature on the composition, mechanism of formation, andvariables affecting concentration of higher alcohols for various kinds ofd i s t i l l e d spirits has appeared during the past 15 years. Reviews includeStevens (33), Suomalainen and Ronkainen (34), Lawrence (35), Webband Ingraham (5), Guymon (6), Suomalainen et al. (32), and Ayrp(7) . Data generally limited to the influence of distillation w i l l be reported here.

G u y m o n (21) reported the composition of tray liquids for brandyd i s t i l l e d in continuous column, respectively, at 130 , 170, and 181proof. The maximum level of fusel oi l occurred on the tray nearest inproof to about 130. This is the second tray below the product tray forthe customary 170 proof of distil lation of the product.

Later we measured the concentrations of -propyl, isobutyl, and thecombined isoamyl and active amyl alcohols (3-methyl-l-butyl alcoholand 2-methyl-l-butyl alcohol) in distillation tray liquids using a gaschromatographic method w i t h -butyl alcohol as the internal standard.The distribution of the higher alcohols in the 14-tray concentratingsection of a 12-inch pilot column during a run in w h i c h the product fromtray 7 was 169 proof is shown in the upper port ion of Figure 4. The




1974

|doi:10.1

021/ba-1974-0137.c

h011



12/22

G U Y M O N D i s t i l l i n g Wines i n t o B r a n d y 243

F i g u r e 4. D i s t r i b u t i o n of h igher a l c o h o h (above) an d eth y l ester so f th r ee fa t t y a c i d s ( bel ow ) i n t r ay l i q u i d s of c o n c e n t r a t i n g sect i on

p r o d u c i n g b ra nd y a t 169proof f r om t r ay No . 7




1974

|doi:10.1

021/ba-1974-0137.c

h011



13/22

244 C H E M I S T R Y OF W I N E M A K I N Ghighest level of -propyl and isobutyl alcohols was found on tray number6 ( 152 proof) or one tray below the product tray. The more abundantamyls (plotted on a scale one-tenth the size) were most concentrated ontray number 5 (at 135 proof), two trays below the product tray.G u y m o n and C r o w e l l (18) developed a quantitative procedure forextracting ( w i t h methylene chloride ) the e t h y l esters of the straight chainfatty acids from brandy or distillates of w i n e , and concentrating and measu r i n g by programmed temperature gas chromatography the three mostabundant e t h y l esters: caprylic, capric, and lauric acids. E t h y l pelargo-nate was used as the internal standard.

The lower portion of Figure 4 shows the di str ibution of three e t h y lesters in the same set of tray l i q u i d samples. These high b o i l i n g esterstend to concentrate at slightly lower proof and tray levels than the higheralcohols, but they all overlap. Consequently, a fusel oi l or low oils layerdrawn from a column w i l l include both higher alcohols and these fattyacid esters.

The higher alcohols and e t h y l esters for a set of plate samples froma run during w h i c h the product at 173 proof was w ithdraw n as the topproduct, i.e., from the condenser, are given in Tab le I. In this case, thehigher alcohols and e t h y l esters were most concentrated on the twouppermost trays.

Table I. Distillation of Beverage Brandy Drawn from CondenserHigher Alcoholsa Ethyl Esters*

Tray - I so- Capry-No. Proof Propyl butyl Amyls late Caprate Laurate

Product 173.1 26.4 30.214 147.5 32.2 39.613 107.5 24.8 27.512 74.2 13.6 13.111 44.2 6.5 5.910 28.8 Trace Trace

Grams per 100 liters at existing proof.

134 1.6 2.3 1.1319 4.4 25.6 10.4305 2.8 14.2 13.3150 4.3 7.7 5.546.6 3.6 8.9 3.511.4 1.4 3.2 1.8

The analyses for higher alcohols in a set of tray samples in w h i c hthe pr imary product was drawn as a sidestream at 186 proof, threetrays below the top, are given in Table II. Und er the operating conditions for this run, all of the higher alcohols were most concentrated inthe lower four trays of the section w i t h an increasing percentage of amylalcohols on the lower trays.

A set of sequential samples taken during the simple batch or potd i s t i l l a t i o n of a charge of low wines into brandy were analyzed for higheralcohols and fatty acid esters as shown in Table III . These represent the




1974

|doi:10.1

021/ba-1974-0137.c

h011



14/22

11. G U Y M O N Distilling Wines into Brandy 245Table II. Distillation of Wine Spirits as Sidestream Productfrom Tray 11

Higher Alcohols,grams/100 liters'1

Tray n - Iso-No. "Proof Propyl butylHeads 189.8 2.314 189.2 7.113 188.6 11.8 11.812 187.5 19.9 16.4Product 186.0 25.4 25.411 186.6 25.7 22.210 184.0 42.5 42.59 181.9 62.3 65.68 179.2 65.0 85.07 176.8 88.0 110.06 171.9 107.5 150.55 166.4 120.5 1774 160.2 135 2003 148.5 116 1902 128.5 128 2081 94.4 53 71 Grams per 100 liters at existing proof.

n-Pro- Iso- AmylsAmyls pyi, % butyl, % % 100 2.4 74.7 25.35.9 40.0 40.0 20.014.0 39.6 32.6 27.825.4 33.3 33.3 33.323.4 36.0 31.1 32.964.0 28.5 28.5 43.0158 21.8 23.0 55.3207.5 18.2 23.8 58.0354 15.9 19.9 64.2688 11.4 15.9 72.71000 9.3 13.7 77.01500 7.4 10.9 81.72150 4.7 7.7 87.63210 3.6 5.9 90.51178 4.1 5.4 90.5

Table III. Pot Distillation of Beverage BrandyCompositionof Sequential Samples

Higher Alcoholsa Ethyl Estersan - Iso- Cap- Cap- Lau-

Sample No. "Proof Propyl butyl Amyls rylate rate rateHeads 158.6 24.1 49.7 116.5 6.1 11.1 8.2Brandy 1 156.4 25.1 49.4 117.2 3.5 7.8 .3fractions 2 153.2 23.7 44.5 115.5 1.2 2.9 2.53 151.4 23.1 41.4 110.8 0.3 0.6 0.64 149.0 -23.0 38.2 108.9 0.1 0.1 0.15 144.6 19.8 34.4 102.36 139.8 19.8 28.2 91.37 121.8 16.7 19.6 57.38 112.0 14.4 10.4 25.2Tail s 9 84.8 9.0 5.0 10.5fractions 10 55.2 3.9 2.0 3.6L o w wines(charge) 57.0 10.7 9.1 22.1 0.3 0.5 0.4Aggregate

product 142.4 23.8 32.8 100.8 0.7 1.7 1.5 Grams per 100 liters at existing proof.




1974

|doi:10.1

021/ba-1974-0137.c

h011



15/22

246 CHEMISTRY OF WINEMAKINGconditions during the second distillation step of the Charente methodused for cognac brandies and for the all-malt highland whiskies of Scotl a n d . The unit used was a 350-gallon pot of copper metal heated by steamcondensing in a heating c o i l . Both the higher alcohols and fatty acidesters are di st il led over at the highest rate at the start of the run. Thee t h y l esters of the C 8 , C i 0 , and C12 fatty acids, al l having b o i l i n g pointsgreater than the higher alcohols of fusel o il , dis til led over almost completely in the early part of the run so none were detectable in the laterproduct fractions nor in the tails. -Propyl alcohol dis til led over moreu n i f o r m l y than any other component measured.

Though not reported in Table III, a peak for phenethyl alcoholemerged in the gas chromatograms of the ester concentrates of the laterbrandy fractions and increased i n size into the tails fractions. Phenethyla l c o h o l is bare ly detectable, if at all , in corresponding concentrates ofcontinuous s t i l l brandies.

The behavior of higher alcohols and fatty acid esters during d i s t i l l a t i o n is clarified by Figures 5 and 6, prepared by Unger (36) from vapor-

0.80.6

METHYL

\ X V ETHYL\ V \ ISOPROPYL

S ^ -PROPYL\\ v ISOBUTYL

\\ ISOAMYL

0 20 40 60 80 100ETHYL ALCOHOL (VOL PERCENT)

Figure 5. Volatility of aliphatic alcohols at variousethyl alcohol-^water concentrations




1974

|doi:10.1

021/ba-1974-0137.c

h011



16/22

11. G U Y M O N D i s t i l l i n g Win es in t o B r andy 247

ETHYL ACETATE

ETHYL ISOBUTYRATE

ETHYL VALERATE

20 40 60 80

ETHYL ALCOHOL (VOL PERCENT)

F i g u r e 6. V o l a t i l i t y o f th r ee eth y l este r s comp ar edw i t h ethy l a l coho l a t va r i o u s ethy l a lcoho l -boater

c o n c e n t r a t i o n s

l i q u i d equilibrium data obtained by Will iams (37) . The volatility, orconcentration of a component in the vapor phase, divided by that in thel i q u i d phase at equi librium for the several alcohols or esters added insmall amounts (less than 1% ) to various strengths of ethyl alcohol-watersolutions are shown as a function of the proof strengths of the equilibriuml i q u i d . A higher molecular weight and high boiling alcohol such asisoamyl alcohol is markedly less volatile than ethyl alcohol at high proofstrengths since isoamyl alcohol is soluble in ethyl alcohol in all proportions. Normal binary solution behavior for isoamyl-ethyl alcohol mixturesis exhibited. But w i t h water, in w h i c h isoamyl alcohol is only slightlysoluble, the volati lity of the isoamyl alcohol is greatly enhanced, characteristic of a partially miscible binary system w h i c h exhibits a minimumb o i l i n g point azeotrope.




1974

|doi:10.1

021/ba-1974-0137.c

h011



17/22

248 C H E M I S T R Y OF W I N E M A K I N GThe proof strength at w h i c h the volatili ty curve of a particular com

ponent intersects that of e t h y l alcohol indicates that proof at w h i c h them i n o r constituent w i l l be concentrated in a fractionating column at thel i m i t i n g condition of total reflux. Fo r practical conditions, the maximumconcentration of a particular congener or minor component occurs at aproof in the column at w h i c h its' volati lity is approximately equal to thei n t e r n a l reflux ratio (L/V where L is the molal l i q u i d or overflow rateand V the molal vapor rate). Thi s can be established by the technique

HETDmON TDK,MNUT E S

F i g u r e 7. Ethy l est er s a n d f ree f a t t y ac i d s i n meth y lenec h l o r i d e ex t r ac t o f b ran dy . F F A? colu mn , 6 f t X Vs in . ,p r o g r a m m e d l i n e a r l y fr om 100-215 a t 7.5/mi n .; 3 ;

i n t e r n a l s t a n d a r d , e thy l pe la rgona te .

of w r i t i n g material balance or operating line equations w i t h respect tothe particular congener, as in the well -known McCabe-Thiele method inw h i c h molal l i q u i d and vapor rates are assumed constant. The internalreflux ratio of the column when the data for Table II were taken wasapproximately 0.70. Reading from the curves of Figure 5 at 0.7 v o l a t i l i t y ,the maximum concentrations should have been obtained as follows: amylalcohols, 128 proof; isobutyl alcohol, 158 proof; -propyl alcohol, 170proof. The values agree fairly w e l l w i t h Table II.




1974

|doi:10.1

021/ba-1974-0137.c

h011



18/22

11. G U Y M O N Distilling Wines into Brandy 249

Figure 8. Ethyl esters extracted from brandy afterremoval of free fatty acids. Same conditions as for

Figure 7.The e t h y l esters of caprylic, capric, and lauric acids are not included

i n Figure 6. However, they should al l f a l l to the left of the curve for e t h y lvalerate and exhibit a somewhat steeper volatility gradient.

Nordstrom (19) demonstrated that esters are formed primarily by adirect biosynthetic process during fermentation in w h i c h acyl-CoA compounds containing the particular fatty acid moiety combine w i t h alcoholsof the medium, w h i c h explains the predominance of e t h y l esters. Esterformation during fermentation does not appear to be direct esterificationbetween alcohols and free fattty acids. However, some direct esterificat i on may occur on the plates of a d i s t i l l i n g column where acids andalcohols are most concentrated.

Free Fatty Acids. In our gas chromatographic analyses of brandyextracts prepared for fatty acid ester measurements, free fatty acids alsoemerged as broad tailing peaks at the end of the programmed temperature run. Free fatty acids have been identified as components of various




1974

|doi:10.1

021/ba-1974-0137.c

h011



19/22

250 C H E M I S T R Y OF W I N E M A K I N Gfermented and distilled beverages by several workers. Nyknen et al.(38) gave quantitative levels of acetic acid and relative proportions of21 other volatile fatty acids in a variety of brandy, rum, and whiskeysamples. They found that acetic, caproic, caprylic, capric, and lauricacids were generally most abundant. They measured acids by gas chromatography after conversion to methyl esters.

C r o w e l l and Guymon (39) developed a gas chromatographic procedure for quantitative measurement of the predominant free fatty acidsby successive extractions of brandy w i t h methylene chloride and thenw i t h dilute a l k a l i to separate the free acids from esters. The acids wereconverted to their -butyl esters for programmed temperature gas chromatography. Figure 7 shows the chromatogram of e t h y l esters extractedf r o m a brandy without removing the free acids. F igur e 8 shows the e t h y lester chromatogram w i t h the free fatty acids removed and converted to-butyl esters and chromatographed as shown i n Figure 9. E t h y l pelar-

r I If i g 44--J . . .fcI 44-- J . . .fcI

! 4\~- J . . .fc 1 -)-+- -- J . . .fcUTYL ESTERS OF FATTY ACIDSEXTRACTED FROMAGED BEVERAGE BRANDY

01 -)-+- -- J . . .fcUTYL ESTERS OF FATTY ACIDSEXTRACTED FROM

AGED BEVERAGE BRANDY- 1

si I

-1

BUTYL ESTERS OF FATTY ACIDSEXTRACTED FROM

AGED BEVERAGE BRANDY5 .

-


AGED BEVERAGE BRANDY 8 .-


AGED BEVERAGE BRANDY < fi- .w m " 7

- :: 7

; 7- 2 i t - 7W 3 : ! 7

: :

Documents

Chemical Aspects of Distilling Wines Into Brandy