THE DISTINCTION OF TRUE EXTRACT OF VANILLA FROM

LIQUID PREPARATIONS OF VANILLIN.

BY WM. H. HESS,

Assistant State Analyst.

An extract of vanilla cannot be made artificially that will duplicate one made from the bean itself, either in respect to chemical behavior, or with regard to usefulness as a flavoring agent. During the sweating process in preparing vanilla beans for the market, chemical changes of unknown nature take place, and as a result of these changes, the aroma is much improved. The importance of this process is well known, and those conditions which produce the best flavor are carefully studied. To provide the various ingredients of the vanilla bean artificially and so treat them that the delicate aroma, incorporated in an alcoholic tincture, equals that acquired from the bean, is a hopeless task, and it is quite as difficult to deceive the analyst by such a product. The vanillin that is synthesized and put upon the market is identical with the vanillin extracted from the bean itself. There are other constituents, as tannin, gum and organic acids, which might be duplicated and whose origin could not be easily proved. There is one important constituent, however, which will be extremely difficult to replace, and that is the resin. This resin behaves so differently from other resins that its separation and identification is a most valuable means of proving the origin of an "extract of vanilla."

Vanilla beans contain from four to eleven per cent of this resin. It is of a dark red to brown color. The coloring matter furnished by this resin constitutes about one-half of the color of the tincture of vanilla. Since manufacturers desire a deeply colored product, they are careful to get all the resin in solution. This resin is soluble in 50% alcohol, so that in an “extract" of high grade, where plenty of alcohol is used, all the resin is kept in solution. In cheap extracts, where as little as 20% of alcohol by volume is sometimes used, an alkali-usually potassium bicarbonate—is added to aid in getting resin, gum, etc., in solution, and also for the purpose of preventing subsequent turbidity. The color is thus deepened very materially. As vanillin is acidic in character and decomposes carbonates to form salts, the delicate flavor is impaired to a considerable extent by the addition of an alkaline carbonate, and a foreign odor is thereby introduced that is quite disagreeable and not difficult to recognize.

In testing preparations of vanilla it is well to make a pharmacopoeial tincture to be used in parallel tests, along with the sample of unknown purity. The following simple tests have been found useful:

I. DILUTION TEST.

Place a few cubic centimeters of the alcoholic tincture in a test tube and add water slowly to three times the original volume and shape. If no alkali has been used in making the extract a somewhat flocculent reddish brown precipitate will separate out. There should be but a slight turbidity. A milky solution would indicate foreign resin. Now add hydrochloric acid a drop at a time to a portion of the diluted sample. Turbidity results at once, due to the setting free of extracted matter held in solution by the plant bases as well as by alkali purposely added. The turbidity produced on acidifying should be slight (absence of foreign resin). Some practice with this simple test will enable an inspector to judge somewhat of the character of a vanilla "extract." If alkali has been used, the turbidity is increased considerably and the color will fade perceptibly when acid is added.

II. SEPARATION AND IDENTIFICATION OF THE RESIN.

Place 25 c. c. of the "extract" to be examined in an evaporating dish and drive off the alcohol on a water bath. When the alcohol is all gone, make up to about the original volume with water. If alkali has not been used in the manufacture of the "extract," the resin will appear as in amorphous, flocculent, red to brown residue. Acidify with acetic acid to free the resin from bases, and the whole of the resin separates out, leaving a partly decolorized, and clear supernatent liquid after standing a short time. If it is desirous to weigh the resin for a quantitative determination, several hours are necessary for its complete separation. Collect the resin on a filter, wash with water, and reserve the filtrate for further tests.

Place a piece of the filter with the resin attached in a few cubic centimeters of dilute caustic potash. The resin dissolves to a deep red solution. On acidifying, the resin is reprecipitated.

Make a solution of the resin in alcohol. To one portion of this alcoholic tincture add a few drops of ferric chloride. No striking coloration is produced. To another portion of the alcoholic tincture add some hydrochloric acid, and again there is little change in color. Most resins Make a solution of the resin in alcohol. To one portion of this alcoholic solution.

To a portion of the filtrate obtained above, add a few drops of basic lead acetate. The precipitate formed is so bulky as to almost make a solid mass, due to the excessive amount of organic acids, gums and other extractive matter. The filtrate from this lead precipitate is nearly but not entirely colorless. To another portion of the filtrate from the resin, add a few drops of ammonia water, and a deep red coloration results. If some zine dust be added to this ammoniacal solution and allowed to stand a short time protected from the air, the color will fade.

Test a portion of the filtrate from the resin for tannin with a solution of gelatin. Tannin is present in varying but small quantities. It should not be present in great excess.

Vanilla “extract" is frequently adulterated with "extract of tonka." This sophistication is best proved by ascertaining the presence of coumarin, the odoriferous constituent of the tonka bean.

COUMARIN AND VANILLIN, THEIR SEPARATION, ESTIMATION AND IDENTIFICATION IN COMMERCIAL

FLAVORING EXTRACTS.

BY WILLIAM H. HESS AND ALBERT B.

PRESCOTT.

It is a common practice in preparing an extract of vanilla from vanilla beans, to introduce also, as a lower-priced material, a portion of tonka beans. This makes a cheaper article and one that is generally popular with consumers. It is understood, however, to be an inferior article when compared with the genuine, and it can no longer be correctly labeled "Vanilla Extract." Under this name the presence of tonka extract constitutes an adulteration and hence the separation, identification, and estimation of coumarin, the odoriferous principle of the tonka bean, has become a practical problem for the analytical chemist.

THEORY FOR THE SEPARATION OF COUMARIN FROM VANILLIN.

Vanillin is hydroxymethoxybenzoic aldehyde, while coumarin is the anhydride of coumaric (orthohydroxycinnamic) acid. The separation depends on the aldehydic and phenolic character of vanillin. If vanillin be dissolved in pure dry ether and dry ammonia gas passed through the solution, the aldehyde ammonia compound of vanillin will be precipitated in almost quantitative proportions. It is accordingly not very soluble in either; but like all aldehyde ammonias, or phenolates it is very soluble in water. Ammonia water should, then, extract all the vanillin from an ether solution. This theoretic principle is fully borne out in practice, as has been proved by repeated experiment. Coumarin, on the other hand, remains wholly in the ether solution.

METHOD OF ANALYSIS.

A sample of 25 to 100 grams of the commercial extract to be examined is placed in an evaporating dish, and the alcohol is allowed to evaporate at about 80° C. Water is added from time to time to keep the amount of liquid about equal to the original volume. When the alcohol is all gone, normal lead acetate solution is added drop by drop until no more precipitate forms. Stirring with a glass rod aids in the flocculation and settling of the precipitate. The solution is now filtered, best through an asbestos funnel under pressure, and the precipitate washed with a few cubic centimeters of hot water. The cooled filtrate is now extracted with ether by shaking out in a separatory funnel. About fifteen to twenty c. c. of ether are used each time, and the process of shaking out is repeated until a few drops of the ether evaporated on a watch glass leaves no residue. If the shaking is done thoroughly, three to four times' shaking out will usually be sufficient to extract all the vanillin and coumarin. The combined ether extracts are introduced into a second separatory funnel and shaken out with dilute ammonia (one part strong ammonia to two parts of distilled water) using about two c. c. of the alkaline water solution at a time, and repeating the process until

the ammoniacal solution is no longer colored yellow. The ether solution is washed with two c. c. of water and this wash-water added to the alkaline extract. The ammoniacal solution is set aside for the estimation of vanillin. The ether solution is now poured out into an evaporating dish, and the ether allowed to evaporate spontaneously, or better, the dish containing the ether is placed in a vacuum desiccator over sulphuric acid, when the ether may be removed quickly. To prevent bumping of the ether and consequent loss during distillation in vacuo, a capillary tube sealed at one end is so placed in the liquid that the open end projects into the centre of the solution, while the sealed end extends well out of the liquid, and rests on the side of the dish. It is necessary that the tube be long enough to prevent its resting horizontally on the bottom of the dish.

For Coumarin.-The residue having been left in vacuo over sulphuric acid long enough to remove all water, or else dried in an air-bath at not above 45° C., is treated with a few cubic centimeters of cold ligroin, the latter being poured off into a clean dry weighed dish. The extraction with ligroin is repeated until a drop evaporated on a watch-glass leaves no residue. The ligroin is allowed to evaporate (best in vacuo in a desiccator containing paraffin), is dried at not above 45° C. and weighed. This gives pure crystalline coumarin, melting at 67 C. After weighing, a melting-point determination should be taken, and this, with the odor which is characteristic at this point in the analysis when compared with that of a known sample, is sufficient for its identification.

For Vanillin. The ammoniacal solution obtained above is rendered slightly acid with ten per cent. hydrochloric acid. It is then shaken out with ether in the manner stated before. The ether is removed and the residue is dried in vacuo over sulphuric acid or else in an air-bath at not above 55° C. This residue is washed repeatedly with boiling ligroin, using a few cubic centimeters at a time to remove the vanillin and free it from adhering substances. The ligroin solution is decanted into a weighed dish, the ligroin is removed, the vanillin is dried at not above 55 C. and weighed. This gives a pure crystalline product, melting ar 80-81° C. After the weighing, a melting-point determination should be made, and then color tests may be applied. Some samples of vanilla extracts have been found to contain phenol-like bodies resembling creosol in odor and chemical behavior. These give color compounds with salts of iron; hence, any volumetric estimation, or even qualitative test, applied to the unpurified extract may be fallacious when based on color reactions of iron salts.

Results were obtained by this method under control as follows: From Mexican vanilla beans, a liquid product was made by following the directions of the United States Pharmacopœia. Of this, tincture twenty-five grams were taken in duplicate portions; to each was added five-tenths gram of pure crystallized coumarin, and the analysis was made for estimation of coumarin and of vanilin. The coumarin recovered from the duplicate portions was, respectively, 0.491 gram or 98.2 per cent. of that taken, and 0.482 gram or 96.4 per cent. of that taken. The vanillin recovered was, respectively, 0.03081 gram and 0.02997 gram, from each of the portions of twenty-five grams of the tincture.

This method for vanillin is much easier and quicker to work than the