Sept., 19501 OF RAFFINOSE I N RAW SUGAKS 457 Determination of Theobromine in Cocoa Products BY K. E. HOLMES SYNoPsIs-The existing methods for determining theobromine in cocoa and its products are described and a new and improved procedure is proposed. In this, the alkaloid is completely extracted by means of boiling water and magnesia, clarified with lead acetate, filtered, concentrated and extracted with chloroform. The alkaloidal residue from the chloroform extract is then assayed by Boie’s method. By this procedure the results are equal to, and sometimes higher than those by Wadsworth’s method. A rapid continuous extractor, for use with chloroform, is described. THE determination of theobromine in cocoa products has always been difficult, and it has received considerable attention in recent years.The methods recently devised can be grouped under two headings. (a) Those using tetrachloroethane or chloroform after triturating with magnesia and a limited amount of water. ( b ) Those using aqueous solvent with added acid or alkali. The methods in group (a), which are the best known, were initiated by Wadsworth,l whose method was modified by Macdonald,2 Humphrie~,~ Kay and Haywood4 and Lowe.5 All these methods suffer from the defect that the amount of water to be left in the mixture of cocoa material and magnesia before the solvent is added is very critical. Insufficient water gives low results, too much yields an impure product, and there is always a doubt as to whether the extraction is complete or not. Jalade6 steeped cocoa in dilute caustic soda, and Pritzker and Jungkunz7 heated it under a reflux condenser with water and magnesia.Both took an aliquot part of the clean extract and extracted it with chloroform in a continuous extractor. Martin and Clergue8 refluxed cocoa with 10 per cent. acid and continuously extracted the mixture with chloroform. Moir and Hinksg extracted the alkaloid with aqueous ethanol and magnesia, evaporated the filtered extract to remove ethanol, extracted the clarified and filtered extract with successive amounts of chloroform, and assayed the final product by Kjeldahl’s method. Parkes and ParkeslO used a mixture of chloroform and phenol instead of pure chloroform. Moores and Campbell11 eluted cocoa material with boiling water, adsorbed the theobromine from an aliquot part of the clarified extract on a chromatographic column, eluted with.alkali and estimated the alkaloid by electrometric titration. The following is a rksurnk of recent work in the second group.458 HOLMES: DETERMINATION OF i-VOl. 75 Of these methods, those of Jalade and of Moir and Hinks are rather lengthy. In the author's experience, that of Pritzker and Jungkunz gives low results, and the same may be said of the method of Martin and Clergue. The Moores and Campbell method is laborious; it uses rather small quantities of sample and the extraction of theobromine is incomplete. EXPERIMENTAL In the present research a method was devised which obviated some of the faults of those mentioned above. Complete extraction of theobromine was effected, with reduced manipula- tion, in a working day of 8 to 10 hours.In this work the Wadsworthl method was used for the sake of comparison, and the Roie12 method for assaying the extracted theobromine. Despite the high solubility of theobromine in tetrachloroethane, this solvent was avoided and water was used instead. I t was shown firstly that all the alkaloid could be extracted from cocoa material of known theobromine content by refluxing 10 g., together with 5 g. Glass Wool- 1 I --!- 1 ; 11'5 1 1 I 1 I I25 I - - f - - I - - ; 50 I I - -I-- \ ( 80 Fig. 1. Rapid liquid - liquid extractor. Dimensions in millimetres. I I I I I I I I I . I I I I 1 350 1 I I I I I I I of magnesium oxide, three times with 250-ml. portioiis of water, followed by filtration and evaporation to dryness.The dried residue, when extracted with more magnesia and tetra- chloroethane by the Wadsworth method, gave a quantitative yield. It was found that if the mixture of cocoa, magnesia and water was clarified, better filtration resulted and that the impurities in the extract were reduced. For this purpose, basic lead acetate solution was found suitable and no loss was observed as a result of its use. An aliquot part of the filtrate was taken after only one extraction, but this led consistently to low results. The loss wasSept., 19501 THEOBROMINE I N COCOA PRODUCTS 459 traced to the adsorption of alkaloid by magnesia, for when pure theobromine done was carried through the aliquot part process, not more than 95 per cent. could be recovered. Complete washing, on the other hand, gave a quantitative yield.Having got all the theobromine into aqueous solution, the next problem was how best to isolate it. This could be done by evaporating the solution to dryness and applying the Wadsworth method to the residue. Moir and Hinksg extracted theobromine from aqueous solution by shaking with successive quantities of chloroform, and Parkes and ParkeslO added 5 per cent. of phenol to the chloroform. The method of continuous extraction with a liquid - liquid extractor seemed to be more promising, and the apparatus shown in Fig. 1 was devised for this purpose. It is a Soxhlet extraction apparatus with the syphon tube replaced by a continuous overflow pipe. With the dimensions shown, the volume of the aqueous layer is about 130ml.The special condenser adaptor has six nozzles with fine jets which ensure maximum contact of the condensed chloroform with the aqueous layer. If a sufficiently rapid reflux rate is maintained (about 50 ml. per minute), 0.3 g. of theobromine can be completely recovered from aqueous solution in less than 3& hours, although 5 hours were taken in practice. No advantage was gained by using a larger extractor; one of 400ml. capacity required 8 hours to complete the extraction of 0.3 g. theobromine, and one of 800 ml. capacity needed 16 hours to extract a similar amount. After extraction, the chloroform was distilled off leaving theobromine that was usually 90 to 92 per cent. pure. It was only slightly discoloured and contained caffeine and a trace of fat. Only the caffeine and fat could be removed by leaching with light petroleum; some of the other impurities could also be removed if tetrachloroethane and ether were used as at the end of the Wadsworth process, but no more than 98 per cent.purity could be attained by this means. All attempts to produce pure theobromine failed. The purity of the crude theobromine as extracted above was therefore determined by the Boie12 method; good sharp end-points have been consistently obtained with this quick and reliable method. It has been in continual use in these laboratories for two years. The following procedure is recommended. PREPARATION OF SAMPLE- petroleum. The defatted material is dried and powdered before use. defatted, dried and powdered. but cocoa powder requires no preparation.first be neutralised with 3 N hydrochloric acid. Cocoa beans or nib are ground to a mass and defatted with low-boiling (80" C.) light Mass or chocolate is Press cake, extrusion cake or shell are ground before use, Cocoa residues containing added lime should PROCEDURE Well mix 10 g. of material with 5 g. of magnesia that has been heated to 900" C. and kept for use in a stoppered bottle. Transfer the mixture to a flask of 500 to 1000 ml. capacity, add 30 ml. of boiling water and boil under a reflux condenser for 10 minutes. Take care at the commencement of the boiling to control the frothing, which may become excessive. Clarify the hot liquid with 10 to 15 ml. of basic lead acetate (the strong solution of lead sub- acetate of the B.P.) and then filter the liquid by suction through a Buchner funnel with a 9 cm.Whatman No. 1 filter-paper. Wash the precipitate, which up to this point must not be allowed to go completely dry, with 50 ml. of hot water, and then suck dry. Transfer the residue and filter-paper to the flask, add 200 ml. of hot water and boil under reflux for 10 minutes. Filter the liquid and wash the precipitate as before. Repeat the extracton procedure once more. Concentrate the combined filtrates to about 100 g. in a tared flask, or in a flask marked a t 100 ml. During the concentration of the filtrate, prepare the clean and dry extractor by the addition of a one inch layer of dry glass wool on the bottom of the extractor tube. Weigh a 250-ml. ground joint flask containing three small fragments of porcelain.Add 100 to 120 ml. of chloroform, connect the extractor flask and clamp the apparatus vertically. Add chloroform to the extractor to a depth of three inches and then, through a funnel, add the 100 ml. of concentrated solution at a temperature of 50" to 60" C. The liquid displaces some of the chloroform into the flask. Wash the flask used for concentrating the solution with a small volume of hot water and add the washings to the extractor tube, taking care that the layer of chloroform above the glass wool is not less than an inch deep. Fit the dispersal460 HOLMES DETERMINATION O F [Vol. 75 adaptor, the jets of which must be just below the surface of the aqueous phase. Add water, if necessary, until this condition is fulfilled. Connect the apparatus to an efficient double surface condenser with ground joint and support the flask on an asbestos sheet having a circular hole large enough to take the bottom quarter- to half-inch of the flask.Use an Amal burner practically full-on to give the necessary rapid rate of reflux. Make sure that, in addition to the jets of the adaptor being full, there is about a half-inch layer of chloroform in the body of the adaptor. If it is intended to extract overnight, use an electrically heated sand-bath. At the end of the extraction time, dismantle the apparatus and distil off the chloroform until a quarter-inch layer remains. Place the flask on a bath of boiling water and remove the remaining chloroform with a hand-blower, until there is no odour of chloroform. Dry the flask and weigh when cold.Determine the theobromine in the crude extract by the Boiel2 method as follows- Dissolve the crude extract in 100 ml. of water containing 1.5 ml. of N sulphuric acid by heating almost to boiling and maintaining at this temperature for 5 minutes. Cool the solution to 40" C. and add 1.5 ml. of phenol red indicator solution. Add N sodium hydroxide (approximately 1.5 ml.) until the solution turns red. Bring the colour back just to yellow by the careful addition of 0.10 N sulphuric acid. Add 0.10 N silver nitrate (which must be neutral to phenol red), taking care that the amount used in ml. is about 80 times the weight of crude theobromine in grams, e.g., for 0.2 g. use 16 ml. of 0.10 N silver nitrate. Titrate the solution with 0.10 N sodium hydroxide until the pink colour returns, The end-point is indicated by a marked change in colour with one drop of the alkali solution.Express the theobromine as a percentage of the original material; for beans, nib, mass or chocolate, the fat percentage is also required for the purpose of calculation. The results have been compared with those obtained by the Wadsworth method on various cocoa products. Some typical figures, expressed on moisture and fat-free material, are given in Table I. Do not use a gauze. For complete extraction, run the extraction for a t least 5 hours. 1 ml. of 0.10 N sodium hydroxide = 18.01 mg. of theobromine. TABLE I Material used Cocoa residue after solvent extraction of the fat .. .. .. .. .. (2) . . .. .. .. .. . . (1) * . (3) * - (4) - .( 5 ) * . (6) - - (7) * - (8) * * (9) * * (10) . . .. .. * . .. .. .. .. .. .. .. .. .. .. .. . . . . . . .. . * .. .. .. .. .. . . * . .. .. .. .. .. .. .. .. .. . . .. .. .. .. Cocoa shell- Unroasted . . .. .. .. .. Roasted . . .. .. .. .. Roasted . . * . .. .. .. Cocoa residues from theobromine extraction . . Cocoa expeller cake . . .. .. .. Milk chocolate . , .. .. .. .. Plain chocolate . . .. .. .. .. Cocoa powder . . .. .. .. .. Wadsworth method r 1 Theobromine, yo Crude Pure 3.14 3.05 3.12 3.30 3-24 3-25 3-23 3.28 2.65 2-s1 2-98 2.93 3.03 3.14 3.06 3-14 3.11 3.19 2-58 2.75 1.58 1.53 1-85 1.58 1-12 1-04 0.65 0-53 0-40 0.38 0.74 0.7 1 3.36 3.24 3.31 3.19 2-71 2.62 0.26 0.23 1-28 1-18 3.46 3.30 Proposed method Crude Pure 3.33 3.3 1 3-37 3.45 3-37 3.53 7 3.41 3.48 2.81 3.03 3.05 3.00 3-05 3.16 3.11 3-17 3.12 3.21 2.59 2.76 1:71 1-54 1-75 1-59 1-19 1-06 0.59 0.54 0-45 0.4 1 0.78 0.7 1 3.60 3.25 3.52 3.19 2.84 2.64 0.26 0.24 1.29 1.1s 3.65 3.32Sept., 19501 THEOBROMINE IN COCOA PRODUCTS 461 These figures show that the proposed method gives results that are as good as, and sometimes higher than, those obtained by the Wadsworth method.The validity of the method has been checked as follows- (1) Theobromine of known purity was carried through the process with and without cocoa material. Some of the results obtained are given in Table 11. TABLE I1 Pure theobromine added, g- 0.247 1 0.3014 0.0592 0.1026 0.1040 0.1327 Theobromine in cocoa material used, €5 nil nil 0.282 0.224 0.041 0.04 1 Theobromine found, g. Crude Pure h f $ 0.250 0.305 0.368 0.356 0.158 0.192 0-247 0.302 0.34 16 0.327 0.145 0-174 These figures show that no theobromine is lost in the course of the process.(2) Reproducible results have been obtained, as shown in Table 111. TABLE I11 Theobromine found in 10 g. Material used Cocoa residue from solvent plant-Sample X . . Ditto-Sample B . . .. .. .. Cocoa residue from theobromine extraction . . Crude, €5 0.312 0.311 0.311 0.313 0.326 0.325 0.324 0.045 0.04 7 Pure, g. 0.285 0.284 0.286 0.285 0.294 0.292 0-293 0.041 0.042 (3) Blank determinatiom-When blank determinations have been carried out, no weighable extract has been obtained, and less than 0.02 ml. of 0.10 N sodium hydroxide has been needed in the Boie assay. SUMMARY It is based upon the extraction of cocoa material with boiling water and magnesia, clarification, filtration and continuous extraction of the concentrated filtrate with chloroform for 5 hours.A design is given for a rapid liquid-liquid extractor for this purpose. The chloroform extract is assayed by the Boie method. The proposed method has been compared with the Wadsworth process. The author wishes to thank the Directors of Messrs. Cadbury Bros., Ltd. for permission to publish this paper, and he is indebted to Mr. S. B. Pliillips and to Dr. H. C. Lockwood for much helpful advice and guidance. 14 method is proposed for the determination of theobromine in cocoa products. REFERENCES 1. Wadsworth, R. V., ,4nalyst, 1921, 46, 32. 2. 3. 4. 5. Lowe, E. H., Ibid., 1948, 73, 679. 6. 7. 8. 9. 10. 11. 12. Boie, H., Pharm. Ztg., 1930, 75, 968. Macdonald, J. A., “6th Ann. Rep. Cocoa Res.,” Imp. Coll. Trop. -\griculture, 1936, p. 34. Humphries, E. C., “8th ,4nn. Rep. Cocoa Res.,” Imp. Coll. Trop. .4griculture, 1938, pp. 36-38. Kay, J., and Haywood, P. J. C., Analyst, 1946, 71, 162. Jalade, Ann. Falsif., 1929, 22, 396. Pritzker, J., and Jungkunz, R., Mitt. Lebensm. Hyg., 1943, 34, 185. Martin, F., and Clergue, H., Ann. Chinz. Analyt., 1942, 24, 202. Moir, 1). D., and Hinks, E., Analyst, 1935, 60, 439. Parkes, A. E., and Parkes, H. A., Ibid., 1937, 62, 791. Moores, R. G., and Campbell, H. A., Anal. Chew., 1948, 20, 40. CADBURY BROTHERS LIMITED BOURNVILLE March, 1950