By Charles W. Trigg

Industrial Fellow of the Mellon Institute of Industrial Research, Pittsburgh, 1916–1920

When the vast extent of the coffee business is considered, together with the intimate connection which coffee has with the daily life of the average human, the relatively small amount of accurate knowledge which we possess regarding the chemical constituents and the physiological action of coffee is productive of amazement.

True, a painstaking compilation of all the scientific and semi-scientific work done upon coffee furnishes quite a compendium of data, the value of which is not commensurate with its quantity, because of the spasmodic nature of the investigations and the non-conclusive character of the results so far obtained. The following general survey of the field argues in favor of the promulgation of well-ordered and systematic research, of the type now in progress at several places in the United States, into the chemical behavior of coffee throughout the various processes to which it is subjected in the course of its preparation for human consumption.

Green Coffee

One of the few chemical investigations of the growing tree is the examination by Graf of flowers from 20-year-old coffee trees, in which he found 0.9 percent caffein, a reducing sugar, caffetannic acid, and phytosterol. Power and Chestnut[102] found 0.82 percent caffein in air-dried coffee leaves, but only 0.087 percent of the alkaloid in the stems of the plant separated from the leaves. In the course of a study[103] instituted for the purpose of determining the best fertilizers for coffee trees, it developed that the cherries in different stages of growth show a preponderance of potash throughout, while the proportion of P₂O₅ attains a maximum in the fourth month and then steadily declines.

Experiments are still in progress to ascertain the precise mineral requirements of the crop as well as the most suitable stage at which to apply them. During the first five months the moisture content undergoes a steady decrease, from 87.13 percent to 65.77 percent, but during the final ripening stage in the last month there is a rise of nearly 1 percent. This may explain the premature falling and failure to ripen of the crop on certain soils, especially in years of low rainfall. Malnutrition of the trees may result also in the production of oily beans.[104]

The coffee berry comprises about 68 percent pulp, 6 percent parchment, and 26 percent clean coffee beans. The pulp is easily removed by mechanical means; but in order to separate the soft, glutinous, saccharine parchment, it is necessary to resort to fermentation, which loosens the skin so that it may be removed easily, after which the coffee is properly dried and aged. There is first a yeast fermentation producing alcohol; and then a bacterial action giving mainly inactive lactic acid, which is the main factor in loosening the parchment. For the production of the best coffee, acetic acid fermentation (which changes the color of the bean) and temperature above 60° should be avoided, as these inhibit subsequent enzymatic action.[105]

Various schemes have been proposed for utilizing the large amount of pulp so obtained in preparing coffee for market. Most of these depend upon using the pulp as fertilizer, since fresh pulp contains 2.61 percent nitrogen, 0.81 percent P₂O₅, 2.38 percent potassium, and 0.57 percent calcium. One procedure[106] in particular is to mix pulp with sawdust, urine, and a little lime, and then to leave this mixture covered in a pit for a year before using. In addition to these mineral matters, the pulp also contains about 0.88 percent of caffein and 18 to 37 percent sugars. Accordingly, it has been proposed[107] to extract the caffein with chloroform, and the sugars with acidulated water. The aqueous solution so obtained is then fermented to alcohol. The insoluble portion left after extraction can be used as fuel, and the resulting ash as fertilizer.

The pulp has been dried and roasted for use in place of the berry, and has been imported to England for this purpose. It is stated that the Arabs in the vicinity of Jiddah discard the kernel of the coffee berries and make an infusion of the husk.[108]

Quality of green coffee is largely dependent upon the methods used and the care taken in curing it, and upon the conditions obtaining in shipment and storage. True, the soil and climatic conditions play a determinative rôle in the creation of the characteristics of coffee, but these do not offer any greater opportunity for constructive research and remunerative improvement than does the development of methods and control in the processes employed in the preparation of green coffee for the market.

That green coffee improves with age, is a tenet generally accepted by the trade. Shipments long in transit, subjected to the effects of tropical heat under closely battened hatches in poorly ventilated holds, have developed into much-prized yellow matured coffee. Were it not for the large capital required and the attendant prohibitive carrying charges, many roasters would permit their coffees to age more thoroughly before roasting. In fact, some roasters do indulge this desire in regard to a portion of their stock. But were it feasible to treat and hold coffees long enough to develop their attributes to a maximum, still the exact conditions which would favor such development are not definitely known. What are the optimum temperature and the correct humidity to maintain, and should the green coffee be well ventilated or not while in storage? How long should coffee be stored under the most favorable conditions best to develop it? Aging for too long a period will develop flavor at the expense of body; and the general cup efficiency of some coffees will suffer if they be kept too long.

One of the first efforts toward artificial maturing was that of Ashcroft[109], who argued from the improved nature of coffee which had experienced a delayed voyage. His method consisted of inclosing the coffee in sweat-boxes having perforated bottoms and subjecting it to the sweating action of steam, the boxes being enclosed in an oven or room maintained at the temperature of steam.

Timby[110] claimed to remove dusts, foreign odors, and impurities, while attaining in a few hours or days a ripening effect normally secured only in several seasons. In this process, the bagged coffee is placed in autoclaves and subjected to the action of air at a pressure of 2 to 3 atmospheres and a temperature of 40° to 100° F. The temperature should seldom be allowed to rise above 150° F. The pressure is then allowed to escape and a partial vacuum created in the apparatus. This alteration of pressure and vacuum is continued until the desired maturation is obtained. Desvignes[111] employs a similar procedure, although he accomplishes seasoning by treating the coffee also with oxygen or ozone.[112] First the coffee is rendered porous by storage in a hot chamber, which is then exhausted prior to admission of the oxygen. The oxygen can be ozonized in the closed vessel while in contact with the coffee. Complete aging in a few days is claimed.

Weitzmann[113] adopts a novel operation, by exposing bags of raw coffee to the action of a powerful magnetic field, obtained with two adjustable electro-magnets. The claim that a maturation naturally produced in several years is thus obtained in 1⁄2 to 2 hours is open to considerable doubt. A process that is probably attended with more commercial success is that of Gram[114] in which the coffee is treated with gaseous nitrogen dioxid.

By far the most notable progress in this field, both scientifically and commercially, has been made by Robison[115] with his "culturing" method. Here the green coffee is washed with water, and then inoculated with selected strains of micro-organisms, such as Ochraeceus or Aspergillus Wintii. Incubation is then conducted for 6 to 7 days at 90° F. and 85 percent relative humidity. Subsequent to this incubation, the coffee is stored in bins for about ten days; after which it is tumbled and scoured. With this process it is possible to improve the cupping qualities of a coffee to a surprising degree.

Renovating Damaged Coffees

Sophistication has often been resorted to in order ostensibly to improve damaged or cheap coffee. Glazing, coloring, and polishing of the green beans was openly and covertly practised until restricted by law. The steps employed did not actually improve the coffee by any means, but merely put it into condition for more ready sale. An apparently sincere endeavor to renovate damaged coffee was made by Evans[116] when he treated it with an aqueous solution of sulphuric acid having a density of 10.5° Baumé. After agitation in this solution, the beans were washed free from acid and dried. In this manner discolorations and impurities were removed and the beans given a fuller appearance.

The addition of glucose, sucrose, lactose, or dextrin to green coffees is practised by von Niessen[117] and by Winter[118], with the object of giving a mild taste and strong aroma to "hard" coffees. The addition is accomplished by impregnating, with or without the aid of vacuum, the beans with a moderately concentrated solution of the sugar, the liquid being of insufficient quantity to effect extraction. When the solution has completely disseminated through the kernels, they are removed and dried. Upon subsequent roasting, a decided amelioration of flavor is secured.

Another method developed by von Niessen[119] comprises the softening of the outer layers of the beans by steam, cold or warm water, or brine, and then surrounding them with an absorbent paste or powder, such as china clay, to which a neutralizing agent such as magnesium oxid may be added. After drying, the clay can be removed by brushing or by causing the beans to travel between oppositely reciprocated wet cloths. In the development of this process, von Niessen evidently argued that the so-called "caffetannic acid" is the "harmful" substance in coffee, and that it is concentrated in the outer layers of the coffee beans. If these be his precepts, the question of their correctness and of the efficiency of his process becomes a moot one.

A procedure which aims at cleaning and refining raw coffee, and which has been the subject of much polemical discussion, is that of Thum[120]. It entails the placing of the green beans in a perforated drum; just covering them with water, or a solution of sodium chloride or sodium carbonate, at 65° to 70° C.; and subjecting them to a vigorous brushing for from 1 to 5 minutes, according to the grade of coffee being treated. The value of this method is somewhat doubtful, as it would not seem to accomplish any more than simple washing. In fact, if anything, the process is undesirable; as some of the extractive matters present in the coffee, and particularly caffein, will be lost. Both Freund[121] and Harnack[122] hold briefs for the product produced by this method, and the latter endeavors analytically to prove its merits; but as his experimental data are questionable, his conclusions do not carry much weight.

The Acids of Coffee

The study of the acids of coffee has been productive of much controversy and many contradictory results, few of which possess any value. The acid of coffee is generally spoken of as "caffetannic acid." Quite a few attempts have been made to determine the composition and structure of this compound and to assign it a formula. Among them may be noted those of Allen,[123] who gives it the empirical formula C₁₄H₁₆O₇; Hlasiwetz,[124] who represents it as C₁₅H₁₈O₈; Richter, as C₃₀H₁₈O₁₆; Griebel,[125] as C₁₈H₂₄O₁₀, and Cazeneuve and Haddon,[126] as C₂₁H₂₈O₁₄. It is variously supposed to exist in coffee as the potassium, calcium, or magnesium salt. In regard to the physical appearance of the isolated substance there is also some doubt, Thorpe[127] describing it as an amorphous powder, and Howard[128] as a brownish, syrup-like mass, having a slight acid and astringent taste.

The chemical reactions of "caffetannic acid" are generally agreed upon. A dark green coloration is given with ferric chloride; and upon boiling it with alkalies or dilute acids, caffeic acid and glucose are formed. Fusion with alkali produces protocatechuic acid.

K. Gorter[129] has made an extensive and accurate investigation into the matter, and in reporting upon the same has made some very pertinent observations. His claim is that the name "caffetannic acid" is a misnomer and should be abandoned. The so-called "caffetannic acid" is really a mixture which has among its constituents chlorogenic acid (C₃₂H₃₈O₁₉), which is not a tannic acid, and coffalic acid. Tatlock and Thompson[130] have expressed the opinion that roasted coffee contains no tannin, and that the lead precipitate contains mostly coloring matter. They found only 4.5 percent of tannin (precipitable by gelatin or alkaloids) in raw coffee.

Hanausek[131] demonstrated the presence of oxalic acid in unripe beans, and citric acid has been isolated from Liberian coffee. It also has been claimed that viridic acid, C₁₄H₂₀O₁₁, is present in coffee. In addition to these, the fat of coffee contains a certain percentage of free fatty acids.

It is thus apparent that even in green coffee there is no definite compound "caffetannic acid," and there is even less likelihood of its being present in roasted coffee. The conditions, high heat and oxidation, to which coffee is subjected in roasting would suffice to decompose this hypothetical acid if it were present.

In the method of analysis for caffetannic acid (No. 24) given at the end of this chapter, there are many chances of error, although this procedure is the best yet devised. Lead acetate forms three different compounds with "caffetannic acid," so that this reagent must be added with extreme care in order to precipitate the compound desired. The precipitate, upon forming, mechanically carries down with it any fats which may be present, and which are removed from it only with difficulty. The majority of the mineral salts in the solution will come down simultaneously. All of the above-mentioned organic acids form insoluble salts with lead acetate, and there will also be a tendency toward precipitation of certain of the components of caramel, the acidic polymerization products of acrolein, glycerol, etc., and of the proteins and their decomposition products.

In view of this condition of uncertainty in composition, necessity for great care in manipulation, and ever-present danger of contamination, the significance of "caffetannic acid analysis" fades. It is highly desirable that the nomenclature relevant to this analytical procedure be changed to one, such as "lead number," which will be more truly indicative of its significance.

The Alkaloids of Coffee

In addition to caffein, the main alkaloid of coffee, trigonellin—the methylbetaine of nicotinic acid—sometimes known as caffearine, has been isolated from coffee.[132] This alkaloid, having the formula C₁₄H₁₆O₄N₂, is also found in fenugreek, Trigonella fœnum-græcum, in various leguminous plants, and in the seeds of strophanthus. When pure it forms colorless needles melting at 140° C., and, as with all alkaloids, gives a weak basic reaction. It is very soluble in water, slightly soluble in alcohol, and only very slightly soluble in ether, chloroform or benzol, so that it does not contaminate the caffein in the determination of the latter. Its effects on the body have not been studied, but they are probably not very great, as Polstorff obtained only 0.23 percent from the coffee which he examined.

Caffein, thein, trimethylxanthin, or C₅H(CH₃)₃N₄O₂, in addition to being in the coffee bean is also found in guarana leaves, the kola nut, maté, or Paraguay tea, and, in small quantities, in cocoa. It is also found in other parts of these plants besides those commonly used for food purposes.

A neat test for detecting the presence of caffein is that of A. Viehoever,[133] in which the caffein is sublimed directly from the plant tissue in a special apparatus. The presence of caffein in the sublimate is verified by observing its melting point, determined on a special heating stage used in connection with a microscope.

The chief commercial source of this alkaloid is waste and damaged tea, from which it is prepared by extraction with boiling water, the tannin precipitated from the solution with litharge, and the solution then concentrated to crystallize out the caffein. It is further purified by sublimation or recrystallization from water. Coffee chaff and roaster-flue dust have been proposed as sources for medicinal caffein, but the extraction of the alkaloid from the former has not proven to be a commercial success. Several manufacturers of pharmaceuticals are now extracting caffein from roaster-flue dust, probably by an adaptation of the Faunce[134] process. The recovery of caffein from roaster-flue gases may be facilitated and increased by the use of a condenser such as proposed Ewé.[135]

Pure caffein forms long, white, silky, flexible needles, which readily felt together to form light, fleecy masses. It melts at 235–7° C. and sublimes completely at 178° C., though the sublimation starts at 120°. Salts of an unstable nature are formed with caffein by most acids. The solubility of caffein as determined by Seidell[136] is given in Table I.

The similarity between caffein and theobromin (the chief alkaloid of cocoa), xanthin (one of the constituents of meat), and uric acid, is shown by the accompanying structural formulæ.

These formulæ show merely the relative position occupied by caffein in the purin group, and do not in any wise indicate, because of its similarity of structure to the other compounds, that it has the same physiological action. The presence and position of the methyl groups (CH₃) in caffein is probably the controlling factor which makes its action differ from the behavior of other members of the series. The structure of these compounds was established, and their syntheses accomplished, in the course of various classic researches by Emil Fischer.[137]

As may be seen in Table II,[138] the caffein content of coffee varies with the different kinds, a fair average of the caffein content being about 1.5 percent for C. arabica, to which class most of our coffees belong. However, aside from these may be mentioned C. canephora, which yields 1.97 percent caffein; C. mauritiana, which contains 0.07 percent of the alkaloid (less than the average "caffein-free coffee"); and C. humboltiana, which contains no caffein, but a bitter principle, cafemarin. Neither do the berries of C. Gallienii, C. Bonnieri, or C. Mogeneti contain any caffein; and there has also been reported[139] a "Congo coffee" which contained no crystallizable alkaloid whatever.

Apparently the variation in caffein content is largely due to the genus of the tree from which the berry comes, but it is also quite probable that the nature of the soil and climatic conditions play an important part. In the light of what has been accomplished in the field of agricultural research, it does not seem improbable that a man of Burbank's ability and foresight could successfully develop a series of coffees possessed of all the cup qualities inherent in those now used, but totally devoid of caffein. Whether this is desirable or not is a question to be considered in an entirely different light from the possibility of its accomplishment.

The variation in the caffein content of coffee at different intensities of roasting, as shown in Table III[140] is, of course, primarily dependent upon the original content of the green. A considerable portion of the caffein is sublimed off during roasting, thus decreasing the amount in the bean. The higher the roast is carried, the greater the shrinkage; but, as the analyses in the above table show, the loss of caffein proceeds out of proportion to the shrinkage, for the percentage of caffein constantly decreases with the increase in color. If the roast be carried almost to the point of carbonization, as in the case of the "Italian roast," the caffein content will be almost nil. This is not a suitable coffee for one desiring an almost caffein-free drink, for the empyreumatic products produced by this excessive roasting will be more toxic by far than the caffein itself would have been.

Caffein-free Coffee

The demand for a caffein-free coffee may be attributed to two causes, namely: the objectionable effect which caffein has upon neurasthenics; and the questionable advertising of the "coffee-substitute" dealers, who have by this means persuaded many normal persons into believing that they are decidedly sub-normal. As a result of this demand, a variety of decaffeinated coffees have been placed on the market. Just why the coffee men have not taken advantage of naturally caffein-free coffees, or of the possibility of obtaining coffees low in caffein content by chemical selection from the lines now used, is a difficult question to answer.

In the endeavor to develop a commercial decaffeinated coffee the first method of procedure was to extract the caffein from roasted coffee. This method had its advantages and its disadvantages, of which the latter predominated. The caffein in the roasted coffee is not as tightly bound chemically as in the green coffee, and is, therefore, more easily extracted. Also, the structure of the roasted bean renders it more readily penetrable by solvents than does that of the green bean. However, the great objection to this method arises from the fact that at the same time as the caffein is extracted, the volatile aromatic and flavoring constituents of the coffee are removed also. These substances, which are essential for the maintenance of quality by the coffee, though readily separated from the caffein, can not be returned to the roasted bean with any degree of certainty. This virtually insurmountable obstacle forced the abandonment of this mode of attack.

In order to avoid this action, the attention of investigators was directed to extraction of the alkaloid in question from the green bean. Because of the difficulty of causing the solvent to penetrate the bean, recourse to grinding resulted. This greatly facilitated the desired extraction, but a difficulty was encountered when the subsequent roasting was attempted. The irregular and broken character of the ground green beans resisted all attempts to produce practically a uniformly roasted, highly aromatic product from the ground material.

Avoidance of this lack of uniformity in the product, and the great desirability to duplicate the normal bean as far as possible, necessitated the development of a method of extraction of the caffein from the whole raw bean without a permanent alteration of the shape thereof. The close structure of the green bean, and its consequent resistance to penetration by solvents, and the existence of the caffein in the bean as an acid salt, which is not easily soluble, offered resistance to successful extraction.

As a means of overcoming the difficulty of structure, the beans were allowed to stand in water in order to swell, or the cells were expanded by treatment with steam, or the beans were subjected to the action of some "cellulose-softening acids," such as acetic acid or sulphur dioxid. As a method of facilitating the mechanical side of extraction without deleterious effects, the treatment of the coffee with steam under pressure, as utilized in the patented process of Myer, Roselius, and Wimmer,[141] is probably the safest.

Many ingenious methods have been devised for the ready removal of the caffein from this point on. Several processes employ an alkali, such as ammonium hydroxid, to free the caffein from the acid; or an acid, such as acetic, hydrochloric, or sulphurous, is used to form a more soluble salt of caffein. Other procedures effect the dissociation of the caffein-acid salt by dampening or immersion in a liquid and subjecting the mass to the action of an electric current.

The caffein is usually extracted from the beans by benzol or chloroform, but a variety of solvents may be employed, such as petrolic ether, water, alcohol, carbon tetrachloride, ethylene chloride, acetone, ethyl ether, or mixtures or emulsions of these. After extraction, the beans may be steam distilled to remove and to recover any residual traces of solvent, and then dried and roasted. It is said[142] that by heating the beans before bringing them into contact with steam, not only is an economy of steam effected, but the quality of the resultant product is improved.

One clever but expensive method[143] of preparing caffein-free coffee consists in heating the beans under pressure, with some substance, such as sodium salicylate, with the resultant formation of a more soluble and more easily steam-distillable compound of caffein. The beans are then steam distilled to remove the caffein, dried, and roasted.

Another process of peculiar interest is that of Hubner,[144] in which the coffee beans are well washed and then spread in layers and kept covered with water at 15° C. until limited germination has taken place, whereupon the beans are removed and the caffein extracted with water at 50° C. It is claimed by the inventor that sprouting serves to remove some of the caffein, but it is quite probable that the process does nothing more than accomplish simple aqueous extraction.

In the majority of these processes the flavor of the resultant product should be very similar to natural roasted coffee. However, in the cases where aqueous extraction is employed, other substances besides caffein are removed that are replaced in the bean only with difficulty. The resultant product accordingly is very likely to have a flavor not entirely natural. On the other hand, beans from which the caffein is extracted with volatile solvents, if the operation be conducted carefully, should give a natural-tasting roast. Any residual traces of the solvent left in the bean are volatilized upon roasting.

Some of the caffein-free coffees on the market show upon analysis almost as much caffein as the natural bean. Those manufactured by reliable concerns, however, are virtually caffein-free, their content of the alkaloid varying from 0.3 to 0.07 percent as opposed to 1.5 percent in the untreated coffee. Thus, although actually only caffein-poor, in order to get the reaction of one cup of ordinary coffee one would have to drink an unusual amount of the brew made from these coffees.

The Aromatic Principles of Coffee

To ascertain just what substance or substances give the pleasing and characteristic aroma to coffee has long been the great desire of both practical and scientific men interested in the coffee business. This elusive material has been variously called caffeol, caffeone, "the essential oil of coffee," etc., the terms having acquired an ambiguous and incorrect significance. It is now generally agreed that the aromatic constituent of coffee is not an essential oil, but a complex of compounds which usage has caused to be collectively called "caffeol."

These substances are not present in the green bean, but are produced during the process of roasting. Attempts at identification and location of origin have been numerous; and although not conclusive, still have not proven entirely futile. One of the first observations along this line was that of Benjamin Thompson in 1812. "This fragrance of coffee is certainly owing to the escape of a volatile aromatic substance which did not originally exist as such in the grain, but which is formed in the process of roasting it." Later, Graham, Stenhouse, and Campbell started on the way to the identification of this aroma by noting that "in common with all the valuable constituents of coffee, caffeone is found to come from the soluble portion of the roasted seed."[145]

Comparison of the aroma given off by coffee during the roasting process with that of fresh-ground roasted coffee shows that the two aromas, although somewhat different, may be attributed to the same substances present in different proportions in the two cases. Recovery and identification of the aromatic principles escaping from the roaster would go far toward answering the question regarding the nature of the aroma. Bernheimer[146] reported water, caffein, caffeol, acetic acid, quinol, methylamin, acetone, fatty acids and pyrrol in the distillate coming from roasting coffee. The caffeol obtained by Bernheimer in this work was believed by him to be a methyl derivative of saligenin. Jaeckle[147] examined a similar product and found considerable quantities of caffein, furfurol, and acetic acid, together with small amounts of acetone, ammonia, trimethylamin, and formic acid. The caffeol of Bernheimer could not be detected. Another substance was separated also, but in too small a quantity to permit complete identification. This substance consisted of colorless crystals, which readily sublimed, melted at 115° to 117° C., and contained sulphur. The crystals were insoluble in water, almost insoluble in alcohol, but readily soluble in ether.

By distilling roasted coffee with superheated steam, Erdmann[148] obtained an oil consisting of an indifferent portion of 58 percent and an acid portion of 42 percent, consisting mainly of a valeric acid, probably alphamethylbutyric acid. The indifferent portion was found to contain about 50 percent furfuryl alcohol, together with a number of phenols. The fraction containing the characteristic odorous constituent of coffee boiled at 93° C. under 13 mm. pressure. The yield of this latter principle was extremely small, only about 0.89 gram being procured from 65 kilos of coffee.

Pyridin was also shown to be present in coffee by Betrand and Weisweiller[149] and by Sayre.[150] As high as 200 to 500 milligrams of this toxic compound have been obtained from 1 kilogram of freshly roasted coffee.

As stated above, the empyreumatic volatile aromatic constituents of the coffee are without question formed during and by the roasting process. According to Thorpe,[151] the most favorable temperature for development of coffee odor and flavor is about 200° C. Erdmann claimed to have produced caffeol by gently heating together caffetannic acid, caffein, and cane sugar. Other investigators have been unable to duplicate this work. Another authority,[152] giving it the empirical formula C₈H₁₀O₂, states that it is produced during roasting, probably at the expense of a portion of the caffein. These conceptions are in the main incomplete and inaccurate.

By means of careful work, Grafe[153] came closer to ascertaining the origin of the fugacious aromatic materials. His work with normal, caffein-free coffee and with Thum's purified coffee led him to state that a part of these substances was derived from the crude fiber, probably from the hemi-cellulose of the thick endosperm cells. Sayre[154] makes the most plausible proposal regarding the origin of caffeol. He considers the roasting of coffee as a destructive distillation process, summarizing the results, briefly, as the production of furfuraldehyde from the carbohydrates, acrolein from the fats, catechol and pyrogallol from the tannins, and ammonia, amins, and pyrrols from the proteins. The products of roasting inter-react to produce many compounds of varying degrees of complexity and toxicity.

The great difficulty which arises in the attempt to identify the aromatic constituents of coffee is that the caffeols of no two coffees may be said to be the same. The reason for this is apparent; for the green coffees themselves vary in composition, and those of the same constitution are not roasted under identical conditions. Therefore, it is not to be expected that the decomposition products formed by the action of the different greens would be the same. Also, these volatile products occur in the roasted coffee in such a small amount that the ascertaining of their percentage relationship and the recognition of all that are present are not possible with the methods of analysis at present at our disposal. Until better analytical procedures have been developed we can not hope to establish a chemical basis for the grading of coffees from this standpoint.

Coffee Oil and Fat

It is well to distinguish between the "coffee oils," as they are termed by the trade, and true coffee oil. In speaking of the qualities of coffee, connoisseurs frequently use erroneous terms, particularly when they designate certain of the flavoring and aromatic constituents of coffee as "oils" or "essential oils." Coffee does not contain any essential oils, the aromatic constituent corresponding to essential oil in coffee being caffeol, a complex which is water-soluble, a property not possessed by any true oil. True, the oil when isolated from roasted coffee does possess, before purification, considerable of the aromatic and flavoring constituents of coffee. They are, however, no part of the coffee fat, but are held in it no doubt by an enfleurage action in much the same way that perfumes of roses, etc., are absorbed and retained by fats and oils in the commercial preparation of pomades and perfumes. This affinity of the coffee oil for caffeol assists in the retention of aromatic substances by the whole roasted bean. However, upon extraction of ground roasted coffee with water, the caffeol shows a preferential solubility in water, and is dissolved out from the oil, going into the brew.

The true oil of coffee has been investigated to a fair degree and has been found to be inodorous when purified. Analysis of green and roasted coffees shows them to possess between 12 percent and 20 percent fat. Warnier[155] extracted ground unroasted coffee with petroleum ether, washed the extract with water, and distilled off the solvent, obtaining a yellow-brownish oil possessing a sharp taste. From his examination of this oil he reported these constants: d_24–5, 0.942; refraction at 25°, 81.5; solidifying point, 6° to 5°; melting point, 8° to 9°; saponification number, 177.5; esterification number, 166.7; acid number, 6.2; acetyl number, 0; iodin number, 84.5 to 86.3. Meyer and Eckert[156] carefully purified coffee oil and saponified it with Li₂O in alcohol. In the saponifiable portion, glycerol was the only alcohol present, the acids being carnaubic, 10 percent; daturinic acid, 1 to 1.5 percent; palmitic acid, 25 to 28 percent; capric acid, 0.5 percent; oleic acid, 2 percent, and linoleic acid, 50 percent. The unsaponifiable wax amounted to 21.2 percent, was nitrogen-free, gave a phytostearin reaction, and saponification and oxidation indicated that it was probably a tannol carnaubate. Von-Bitto[157] examined the fat extracted from the inner husk of the coffee berry and found it to be faint yellow in color, and to solidify only gradually after melting. Upon analysis, it showed: saponification value, 141.2; palmitic acid, 37.84 percent, and glycerids as tripalmitin, 28.03 percent.

Carbohydrates of the Coffee Berry

There has been considerable diversity of opinion regarding the sugar of coffee. Bell believed the sugar to be of a peculiar species allied to melezitose, but Ewell,[158] G.L. Spencer, and others definitely proved the presence of sucrose in coffee. In fat-free coffee 6 percent of sucrose was found extractable by 70 percent alcohol. Baker[159] claimed that manno-arabinose, or manno-xylose, formed one of the most important constituents of the coffee-berry substance and yielded mannose on hydrolysis. Schultze and Maxwell state that raw coffee contains galactan, mannan, and pentosans, the latter present to the extent of 5 percent in raw and 3 percent in roasted coffee. By distilling coffee with hydrochloric acid Ewell obtained furfurol equivalent to 9 percent pentose. He also obtained a gummy substance which, on hydrolysis, gave rise to a reducing sugar; and as it gave mucic acid and furfurol on oxidation, he concluded that it was a compound of pentose and galactose. In undressed Mysore coffee Commaille[160] found 2.6 percent of glucose and no dextrin. This claim of the presence of glucose in coffee was substantiated by the work of Hlasiwetz,[161] who resolved a caffetannic acid, which he had isolated, into glucose and a peculiar crystallizable acid, C₈H₈O₄, which he named caffeic acid.

The starch content of coffee is very low. Cereals may readily be detected and identified in coffee mixtures by the presence and characteristics of their starch, in view of the fact that coffee (chicory, too) is practically free from starch. On this score it is inadvisable for diabetics to use any of the many cereal substitutes for coffee. It is pertinent to note in this connection that persons suffering from diabetes may sweeten their coffee with saccharin (1⁄2 to 1 grain per cup) or glycerol, thus obtaining perfect satisfaction without endangering their health.

The cellulose in coffee is of a very hard and horny character in the green bean, but it is made softer and more brittle during the process of roasting. It is rather difficult to define under the microscope, particularly after roasting, even though the chief characteristics of the cellular tissue are more or less retained. Coffee cellulose gives a blue color with sulphuric acid and iodin, and is dissolved by an ammoniacal solution of copper oxid. Even after roasting, remnants of the silver skin are always present, the structure of which, a thin membrane with adherent, thick-walled, spindle-shaped, hollow cells, is peculiar to coffee.

The Chemistry of Roasting

The effect of the heat in the roasting of coffee is largely evidenced as a destructive distillation and also as a partial dehydration. At the same time, oxidizing and reducing reactions probably occur within the bean, as well as some polymerization and inter-reactions.

A loss of water is to be expected as the natural outcome of the application of heat; and analyses show that the moisture content of raw coffee varies from 8 to 14 percent, while after roasting it rarely exceeds 3 percent, and frequently falls as low as 0.5 percent. The loss of the original water content of the green bean is not the only moisture loss; for many of the constituents of coffee, notably the carbohydrates, are decomposed upon heating to give off water, so that analysis before and after roasting is no direct indication of the exact amount of water driven off in the process. If it be desired to ascertain this quantity accurately, catching of the products which are driven off and determination of their water content becomes necessary.

The carbohydrates both dehydrate and decompose. The result of the dehydration is the formation of caramel and related products, which comprise the principal coloring matters in coffee infusion. That portion of the carbohydrates known as pentosans gives rise to furfuraldehyde, one of the important components of caffeol.

The effect of roasting upon the fat content of the beans is to reduce its actual weight, but not to change appreciably the percentage present, since the decrease in quantity keeps pace fairly well with the shrinkage. Some of the more volatile fatty acids are driven off, and the fats break down to give a larger percentage of free fatty acids, some light esters, acrolein, and formic acid. If the roast be a very heavy one, or is brought up too rapidly, the fat will come to the surface, through breaking of the fat cells, with a decided alteration in the chemical nature of the fat and with pronounced expansion and cracking.

Decomposition of the caffein acid-salt and considerable sublimation of the caffein also occur. The majority of the caffein undergoes this volatilization unchanged, but a portion of it is probably oxidized with the formation of ammonia, methylamin, di-methylparabanic acid, and carbon dioxid. This reaction partly explains why the amount of caffein recovered from the roaster flues is not commensurate with the amount lost from the roasting coffee; although incomplete condensation is also an important factor. Microscopic examination of the roasted beans will show occasional small crystals of caffein in the indentations on the surface, where they have been deposited during the cooling process.

The compound, or compounds, known as "caffetannic acid" are probably the source of catechol, as the proteins are of ammonia, amins, and pyrrols. The crude fiber and other unnamed constituents of the raw beans react analogously to similar compounds in the destructive distillation of wood, giving rise to acetone, various fatty acids, carbon dioxid and other uncondensable gases, and many compounds of unknown identity.

During the course of roasting and subsequent cooling these decomposition products probably interact and polymerize to form aromatic tar-like materials and other complexes which play an important rôle among the delicate flavors of coffee. In fact, it is not unlikely that these reactions continue throughout the storage time after roasting, and that upon them the deterioration of roasted coffee is largely dependent. Speculation upon what complex compounds are thus formed offers much attraction. A notable one by Sayre[162] postulates the reaction between acrolein and ammonia to give methyl pyridin, which in turn with furfurol forms furfurol vinyl pyridin. This upon reduction would produce the alkaloid, conin, traces of which have been found in coffee.

Although furfuraldehyde is the natural decomposition product of pentosans, furfuryl alcohol is the main furane body of coffee aroma. This would indicate that active reducing conditions prevail within the bean during roasting; and the further fact that carbon monoxid is given off during roasting makes this seem quite probable. If one admits that caffetannic acid exists in the green bean; that upon oxidation it gives viridic acid; and that it is concentrated in the outer layers of the bean, as certain investigators have claimed, then there is chemical proof of the existence of oxidizing conditions about the exterior of the bean. In any event, however, the fact that oxidizing conditions predominate on the external portion of the bean is obvious. Accordingly, our meager knowledge of the chemistry of roasting indicates that while the external layers of the roasting beans are subjected to oxidizing conditions, reducing ones exist in the interior. Future experimentation will, no doubt, prove this to be the case.

Attempts have been made to retain in the beans the volatile products, which normally escape, both by coating previous to roasting[163] and by conducting the process under pressure.[164] However, the results so obtained were not practical, since the cup values were decreased in the majority of cases, and the physiological effects produced were undesirable. In cases where the quality was improved, the gain was not sufficient to recompense the roaster for the additional expense and difficulty of operation.

Various persons have essayed to control the roasting process automatically; but the extreme variance in composition of different coffees, the effect of changing atmospheric conditions, and the lack of constancy in the calorific power of fuels have conspired to defeat the automatic roasting machine.[165] It is even doubtful whether De Mattia's[166] process for roasting until the vapors evolved produce a violet color when passed into a solution of fuchsin decolorized with sulphur dioxid is commercially reliable.

Many patents have been granted for the treatment of coffees immediately prior to or during roasting with the object of thus improving the product. The majority of these depend upon adding solutions of sugar,[167] calcium saccharate,[168] or other carbohydrates,[169] and in the case of Eckhardt,[170] of small percentages of tannic acid and fat. In direct opposition to this latter practise, Jurgens and Westphal[171] apply alkali, ostensibly to lessen the "tannic acid" content. Brougier[172] sprays a solution containing caffein upon the roasting berries; and Potter[173] roasts the coffee together with chicory, effecting a separation at the end.