THE PREPARATION PROPERTIES AND USES OF SILICODUODECITUNGSTIC ACID BY EDWARD OSCAR NORTH B.S. Beloit College, 1018 M.S. University of Illinois, 1922 THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN CHEMISTRY IN THE GRADUATE SCHOOL OF THE UNIVERSITY OF ILLINOIS, 1924 “ore PIDBABY AE THE tie Lignan! UF AEC roR OA 109% TURBANAZILEINOIS “wad = ea 2 2agZare INIVERSITY GF ILLINO UNIVERSITY Ur ILLINGLS Reprinted from the JOURNAL OF THE AMERICAN PHARMACEUTICAL SOCIETY, Vol. XIII, Nos. 10 and 11, October and November, 1924 DOM ENT = > LATS i wre c a “3 oa , : Cd . ae NT 4 7 - rs df 4 ee > y Le i Z = me oe eek ; 42 . oe f A aa < es) ' ‘a 4 f . y ft - x - s BT : a : i <2 : ~ a) - Bs ) r 5 e 7 4 ed -_ > J « eon , ‘ ~~ . “ k » . ¥. ‘ “ . Trak, : CY ‘ * re ee. er THE PREPARATION, PROPERTIES, ‘AND USES OF SILICO- DUODECITUNGSTIC ACID.* I. THE PREPARATION OF THE eACI DOAN hp oe sn a oa HISTORICAL, Silicoduodecitungstic? acid was discovered by Marignac (1) in 1863 while preparing ammonium tungstate. The acid was formed during extraction of tungs- ten from the ore, which contained silica, and, being extremely soluble, was found in the mother liquors. He then prepared the acid by boiling a mixture of tungstic acid with gelatinous silicic acid, obtaining the solid acid by evaporation of the filtered liquid. ‘The sodium, potassium, and ammonium salts were described, and the empirical formula SMO.20WO;.2SiO; suggested for the salts. In a second paper (2) in 1864 two other acids were described: tungstosilicic acid, an isomer of the first, to which the formula 12WO;.Si0:.4HO + Aq was given, and silico- decitungstic acid of the formula 9102.10WO;.4HO.3Aq. These acids were separated as mercurous salts, which were quite insoluble but easily decomposed by hydrochloric acid. Wyrouboff (3) in 1896 followed essentially the procedure of Marignac in pre- paring the acid, substituting nitric acid for hydrochloric in generating the tungstic acid. Copaux (4) in 1908 made one of the most valuable contributions to the knowledge of these compounds. ‘There were then three acids known, which, in modern nomenclature, had the formulas A. SiO».12W0O;.2H.0.31Aq B. > $i0..12WO3.2H20.22Aq Ca SiO2.10WOs3.2H20.3Aq isomers Copaux prepared ‘‘A’”’ in three ways. 1. By reaction in a boiling solution of sodium tungstate, Na,WO,, and sodium silicate, Na.SiO; (water glass), the solution being kept neutral with nitric acid. The Pompeu was ex- tracted by ether in the presence of sulfuric acid. 2. By reaction between sodium tungstate and sodium orthosilicate, Na,SiOu, in a sealed tube at 150° C. The compound was isolated by means of ether and sulfuric acid. 3. By reaction between boiling solutions of sodium tungstate and water glass made acid with acetic acid. ‘The compound was extracted as before. Acid ““B’’ was made by Procedure 3 under Acid “A,” with the substitution of sulphuric acid for the acetic acid. Acid ‘‘C’’ was formed by boiling ‘‘A”’ with a large excess of ammonia. Silica at first precipitated but almost completely redissolved and the ammonium salt of “C” crystallized from the filtrate. * Contribution from the Department of Chemistry of the University of Illinois. 1 Read at the 72nd Meeting of the American Pharmaceutical Association, Buffalo, New York, August 25-29, 1924. 2 Throughout this paper the term silicotungstic acid will be understood to refer to silicoduodecitungstic acid. 6 Goddefroy (5), 1876, found ‘that the alkaloidal salts of this acid were very insoluble, and also called attention. to the low solubility of the rubidium and caesium salts. Bertrand (6, 7), 1899) published the first detailed study of the alka- loidal salts of the acid.. He maintained that the paper of Goddefroy had not been available and was unknown to him, and that his study was entirely independent. He determined the sensitiveness of the reaction with eighteen or more alkaloids and, with Javillier (8, 9), used the acid in the quantitative determination of nicotine as a means of separating the pure alkaloid. Javillier (10) continued the work, study- ing the compounds of antipyrine (11), atropine, conine, pyramidone and spar- teine. | Guillemard (12) in 1903 described the use of the acid in determining the animal bases in the urine, which he referred to as alkaloidal nitrogen. Chapin (13), in 1911, published a direct method for the determination of nicotine in tobacco extracts, precipitating nicotine silicotungstate which was filtered and ignited, the determination being based on the weight of silicotungstic anhydride ob- tained. _.Ferenz and David (14) used the acid for the separation of alkaloids from im- pure solutions; recovering the alkaloids by decomposition of their salts with sodium hydroxide and ether. Rasmusson (15), 1917, outlined a procedure for the analysis of belladonna, using a correction factor because of the solubility of atropine silico- tungstate. Taigner (16), 1919, in studying the salts of atropine, strychnine and cocaine, found that when dried at 120° they all had the same type formula, vz., 12WO3.Si02.2H.O.4 alkaloid. This permitted a gravimetric determination of these alkaloids without ignition. Heiduschka and Wolff (17) have studied the ratio of alkaloid to silicotungstic acid in a number of salts, determining the condi- tions under which these salts are formed. Sindlinger and Mach (18), 1924, have described the determination of pyridine and nicotine. By making allowance for the solubility of the pyridine salt in dilute hydrochloric acid it has been possible to determine this base quantitatively, and by the use of selective solvents for the two salts to determine the percentage of each of the two bases in a mixture. THEORETICAL. t Of the three methods suggested by Copaux (loc. cit.) for the preparation of “A”’ the first yields the best results. The sealed tube method is cumbersome and not applicable to large amounts, and the acetic acid used in the third method at — times causes some reduction of the silicotungstic acid. In the first method, there are at least several sources of difficulties. The sodium tungstate solution must be continuously stirred during the neutralization with nitric acid, since a local excess produces a precipitate of tungstic anhydride which dissolves with difficulty. The water glass, if not neutralized cautiously, yields a troublesome gel. During the heating the reaction of the solution must be frequently determined and nitric acid cautiously added to maintain neutrality. | Sulphuric acid is added in this method to finally liberate the acid, which is extracted by shaking the solution with ether. The ether solution is apparently similar to a hydrate, and is a heavy oily liquid, miscible neither with water nor ether. The sulphuric acid has a tendency to cause a precipitation of tungstic an- hydride which is carried down with the ether and promotes the formation of emul- fi sions. Sulphuric acid is carried down by the ether complex, cannot be volatilized with the ether, and tends to initiate a spontaneous decomposition on standing. In the procedure which we have adopted, hydrochloric acid is substituted for the nitric, sulphuric and acetic acids made use of by Copaux. ‘The formation of the acid from sodium tungstate and sodium silicate may be represented by the following reaction: 12Na,WO, + NassiO; + 17H,O = 4H20.Si0..12WO; + 26NaOH If the reaction is to proceed from left to right it is only necessary to heat the acid to favor hydrolysis of the salts and to neutralize the alkali as it is set free. A slight theoretical excess of sodium silicate over sodium tungstate is used, and upon the slow addition of hydrochloric acid no tungstic anhydride is precipitated. ‘The only turbidity results from the liberation of the excess of silicic acid after the combina- tion has taken place, and this gel is readily removed by filtration. Addition of a large excess of hydrochloric acid does not produce a precipitate. After the reaction is complete, the silicotungstic acid is readily extracted by ether without emulsifica- tion forming a distinct bottom layer, and the hydrochloric acid carried down with the complex volatilizes with the ether. Previous investigators have been well agreed upon all of the properties of this acid except its basicity. Silicotungstic acid is quite stable. It will decompose chlorides and nitrates and is very resistant towards boiling sulphuric and perchloric acids. It is readily reduced by free metals such as copper, iron, aluminum and zinc, blue compounds being formed. Silicotungstic acid forms salts with mono-, di- and trivalent metals, but not with metals forming the oxide RO», according to Wyrouboff (loc. cit.). The salts may contain four or eight equivalents of the base. ‘Those which are of the nature of acid salts are very soluble in ether, alcohol and water, the neutral salts much less readily so. The most interesting property of the acid is that of forming sparingly soluble compounds with many organic bases. ‘These include the secondary and tertiary aliphatic and aromatic amines and quaternary ammonium derivatives. Insoluble salts are not formed by the primary amines. Salts of the secondary bases are usually soluble in hot water, those of the tertiary bases being least soluble. An ex- ception to the last statement is found in the effect of adjacent groups, such as the hydroxy! or carboxyl, which tend to neutralize the basicity. of the compound or to render its salts more soluble. Marignac (2) believed that the acid was octabasic. He found that a solution of a salt containing four equivalents of a base would decompose carbonates and turn blue litmus red. ‘The addition of more base turned the litmus blue, but on standing the red color returned. He also found that the acid was slowly decomposed by an excess of alkali with the formation of normal tungstate and silicate. When silicotungstic acid is titrated with standard alkali, using methyl orange as indicator, an end-point is obtained when the equivalent of four acid hydrogens have been neutralized. When the acid is titrated in hot solution with standard alkali and phenolphthalein indicator, the equivalent of twenty-six hydrogens are neutralized. Copaux (19) claimed that a sharp, though not permanent, end-point, was obtained with phenolphthalein at a point corresponding to the neutralization 8 of four hydrogens. ‘This is entirely contrary to results obtained in this laboratory. While the acid has been found to contain eight replaceable hydrogens, as shown by the formation of normal potassium and ammonium salts, the salts which are formed with tertiary organic bases represent the neutralization of but four eqtiva- lents of hydrogen. ‘The empirical formulas obtained upon analysis of these salts, however, do not support the theory of the octabasicity of the acid. In the usual manner of formation of such salts, there is a direct addition of the ions of the acid to the nitrogen of the base. Such an addition would mean that the number of molecules of water which must be eliminated to form the anhydride of the acid, would be the same, regardless of whether the acid or the alkaloidal salt was heated. Actual experiments show that only one-half as much water can be eliminated from the alkaloidal salt as from the free acid, forming the same anhydride in each ease. The neutral metallic salts were formed by the addition of an actual excess of alkali to the acid, the alkaloidal salts were precipitated in solutions acid with a strong mineral acid. It would thus appear that there is a condition of equilibrium existing in the acid solution, in which system there are both tetra- and octabasic acids present as members. The presence of a strong acid may repress the ioniza- tion of four hydrogens, permitting, at the same time, a dehydration with the forma- tion of a partial anhydride. ‘The condition is thus exactly the same as the state of equilibrium between chromic and dichromic acids, or between a dichromate and a true acid chromate in solution. The octapotassium salt, prepared in this laboratory, is acidic towards phenol- phthalein, neutral towards litmus, and alkaline towards methyl orange and methyl red. . The following is suggested as a structural formula for silicoduodecitungstic acid, showing it as a derivative of orthosilicic acid: W oF Ov On No HO 0 RS O OH \w LON ee Uo >w—0-si-0—w ORO i ee cag HO” No H;N (m. wt. 3396.7); SiO2.12WOs, 83.75%. CoH:N + H20, 16.25%. Found: Si02.12WOs;, 83.65%, 83.68%. CoH;N + HO, 16 Boe l0-32 Ye. Nicotine Silicotungstate—Freshly distilled water-white nicotine was quickly dissolved in dilute hydrochloric acid and precipitated with an excess of silicotungstic acid. The liquid con- taining the amorphous precipitate was heated to boiling and allowed to ‘cool, obtaining a white crystalline product, not salmon colored as described by Bertrand (7). The product was dried at 60°. Analysis. Calculated for 2H20.Si0O2.12WO3.2CioHisN2.5H2O (m. wt. 3294): SiO:.12WOs, 86.34%; CioHuN2 + H2O, 13.66%. Found: SiOs.12WOs, 86.22%, 86.24%; CioHuN: + H20, 13.78%, 13:10 %.- Another specimen was prepared and dried at 100°. The nicotine used was not water-white and the salt had a yellow tint. Analysis. Calculated for 2H»:0.5i02.12WO3.2CioHiuN2 (m. wt. 3204): SiO..12WOs, 88.75%; CioHuN: + HoO, 11.25%. Found: SiO2.12WOs, 88.25%, 88.05%; CioHuNe + H20, 11.75%, 11.95%. Quinine Silicotungstate—This was a white pulverulent compound extremely insoluble in water and dilute acids. Analysis. Calculated for 2H2,0.SiO2.12WO3.2C2.HasN2O2 (m. wt. 3528.9): SiOe.12WOs, 80.60%; CooH2aN2O2 + H,0, 19.40%. Found: Si02.12WOs, 80.47%, S012, 5 CsoHosNoO, + H.0, 19.53%, 19.49%. Cinchonine Silicotungstate—This was a white microcrystalline product extremely insoluble in water and dilute acids. ‘The last traces of water of crystallization are removed with difficulty. This compound was heated for six hours at 120° without appreciable change. Analysis. Calculated for 2H20.SiO2.12WO3.2CisH2N20.2H:0 (m. wt. 3504.9): SiOs.- 12WOs, SLD... Ci 9H22N20 + HO, 18.85%. Found: SiO». 12WOs, SLID, 81.35%. Cio- H»N2,O + H2O, 18.85%, 18.65%. Morphine Silicotungstate.—A bufi-colored compound which is moderately soluble in alcohol: and dilute acids. Analysis. Calculated for 2H2O.SiO2.12WO3.4Ci7HisNO3 (m. wt. 4021.3): Si02.12WOs,, 70:73:%: Cy7HigNO3 + H:0, ZO N, 37%. Found: SiO2.12WOs, 70.39%, 70.56%; Ci H,;NO; + H:O, 29.61%, 29.44%; N, 1.44%. Codeine Silicotungstate.—A buff-colored precipitate slightly soluble in alcohol and dilute acids. Analysis. Calculated for 2H20.SiO2.12WO3.4CisHaNO; (m. wt. 4077.3): SiO2.12WOs, 69.74%: CisHaiNO; + HO, 30.26%. Found: SiO.12WO3;, 69.51%, 69.65%; CisH»NOs; + HO, 30.49%, 30.35%. Apomorphine Silicotungstate—This compound resembled in color and solubility the salts of morphine and codeine. 12 Analysis. Calculated for 2H:0.Si0O2.12WO3.4Ci7HizNO2 (m. wt. 3949.2): SiOs.12WOs, 12.02%; Cyi7Hy7NO2 4+ H.0, 27.98%. Found: Si02.12WOs, CZAR 71.94%; Ci7Hi7NOsz, 27.85%, 28.06%. Narceine Silicotungstate—This compound was buff colored when first precipitated. It dissolved in hot dilute hydrochloric acid and on cooling was deposited in reddish crystals. Analysis. Calculated for 2H:O.SiO2.12WO3.4C2;H2z;NOs (m. wt. 4661.8): SiO..12WOs, 61.01%; CosHez7NOg + HO, 38.99%; N, 1.20%. Found: S$i0s.12WOs;, 61.14%, 60.97%; Co3Hoz,NOs + H20, 38.86%, 39.03%; N, 1.22%. Narcotine Silicotungstate—This was a yellowish pulverulent compound, very insoluble in water and dilute acid. Analysis. Calculated for 2H20.Si0O2.12WO3.4Cs2H»NO;7 (m. wt. 4533.5): SiO2.12WOs, 62.74%; CoHeaNO7 + HO, 37.26%; N, 1.24%. Found: SiOs.12WOs:, 62.55%, 62.50%; CooHo3NO7 a H20O, 31.40%; 37.50%; N, LST. 4 Strychnine Silicotungstate—This salt was previously prepared by Taigner (loc. cit.). The product obtained in this laboratory agrees with the one which he has described. The compound was extremely insoluble and required long washing with dilute hydrochloric acid to remove the excess of silicotungstic acid. The color was faintly yellow. Analysis. Calculated for 2H2O.SiO2.12WO3.4C2:Ha»N2O2 (m. wt. 4217.5): SiO2e.12WOs, 67.45%; CoHeN2Oe + HO, 32.55%; N, 2.64%. Found: SiO..12WOs, 67.68%, 67.46%; C;Ho2N2O2 + H2O, 32.32%, 32.54%; N, 2.46%. Brucine Silicotungstate—It had a slight bluish tint and was very insoluble in water and dilute acids. Like the strychnine salt long washing was required to remove the excess of silico- tungstic acid. Analysis. Calculated for 2H:O.SiO2.12WO3.4Co3H2sN204 (m. wt. 4457.9): SiO2.12WOs, 63.82%; Co3Hop N20. + HO, 36.18%. Found: SiO2.12WOs, 63.56 Zoy.0.04 Yon CosHoeN2Ou > H.0, 36.44%, 36.33%. Cocaine Silicotungstate—A white microcrystalline salt, already mentioned by Taigner (loc. ctt.). Analysis. Calculated for 2H2O.Si0O2.12W0O;3.4Ci7H2NO,. (m. wt. 4093.5). SiOQ2.12WOs, 69.60%; C,7HuNOs, + H.O, 30.40%. Found: Si02.12WOs, 69.80%, 69.70%; Ci7He NO, + H.O, 30.20%, 30.30%. Caffeine Silicotungstate—This first precipitated as a white amorphous compound but changed to pale yellow crystals on standing. Analysis. Calculated for 2H2O.SiO:.12WO3.4CsHi9N.O2 (m. wt. 3657): SiO».12WOs, LABS: CsHi9N4O2 + HO, Pa LES Found: SiO2.12WOs, 77.60%, 11.00%; CsgHi9N102 + H.O, 22.40%, 22.50%. Theobromine Silicotungstate—This formed as a faint yellow powder. Analysis. Calculated for 2H20.Si0O..12WO;.4C;HsN.O2 (m. wt. 3600.8): SiOc.12WOs, 78.99%; CyHsN,0O. + HO, 21.01%; N, 6.28%. Found: SiOc.12WO;, 78.77%, 78.84%; Cr- HsN.O2 + HO, 21.28%, 21.16%; N, 6.10%. . Veratrine Silicotungstate—The compound had a slight purple color, which was deepened by heating in dilute sulphuric acid. A discrepancy existed between the composition as deter- mined by theory and analysis. ‘The salt was prepared from several different lots of alkaloid with invariably the same composition. Heiduschka and Wolff (17) state that the 1:4 compound could only be obtained by precipitation in alcoholic solution. Analysis. Calculated for 2H2O.SiO2.12WO3.4C32HsgNO g (m. wt. 5246.5): SiO2.12WOs, 54.22%; CzeHagNOg + HO, 45.78%. Found: $i02.12WOs, DD.DOUe ODD Loe. Cz3eHsgNOg + H20, 44.45%, 44.49%. Colchicine Silicotungstate.—This was a bright yellow amorphous salt. Analysis. Calculated for 2H20.Si0O2.12W O3.4Co2HosN Oe (m. wt. 44773): SiO». 12WO3, 63.58%; CoHosNOs + HO, 36.47%; N, 1.25%. Found: SiOv12WO;, 63.55%, 63.45%; CooHosNOs + H20, 36.45%, 36.55%; N, 1.21%. 13 Jensen (23) claims to have prepared a salt in which the ratio of acid to base was 1:5. Atropine Silicotungstate.—This salt was also prepared by Taigner (loc. cit.). It is a white crystalline compound, slightly soluble in hot dilute hydrochloric acid. The solubility, however, is slight enough to permit of quantitative precipitation from a concentrated solution. Analysis. Calculated for 2H20.Si02.12W0O3.4C;7H2NO3 (m. wt. 4037.4): SiOs.12WOs, 70.45%; CizHesNOs + H:O, 29.55%; N, 1.89%. Found: $102.12WOs, 70.05%, 70.23%; Ci7Ha3NO3; + H:20, 29.95%, 29.77%; N, 1.46%. Hyoscyamine Silicotungstate.—Like the atropine salt, this compound was white, crystal- line and moderately soluble. Analysis. Calculated for 2H20.Si02.12WO3.4Ci7HasNO3 (m. wt. 4037.4): SiO2.12WOs, 70.45%; CizHesNO3; + HO, 29.55%; N, 1.89%. Found: SiOe.12WOs, 70.84%, 70.47%; Ci7He3NO3 + H,0O, 29.66%, 20.93%: N, L330 %, Berberine Silicotungstate—This was a saffron-colored compound. Pictet (24) gives CooHy7NO, as the composition of berberine; Schmidt (25), CooHisNOs. Schmidt’s formula was used in the following calculations. Analysis. Calculated for 2H2O.Si02.12WO3;.4C2oHisNOs (m. wt. 4293): SiO2.12WOs, 66.25%; CooHisNOs + H20, 33.75%; N, 1.31%. Found: SiO..12WO3,-66.00%, 66.06%; CooHisNOs + H20, 34.00%, 33.94%; N, 1.27%. Hydrastine Silicotungstate-—This was a pale yellow amorphous compound. Analysis. Calculated for 2H20.Si02.12WO3.4Cs:;H2NOs (m. wt. 4413.4): SiOs.12WOs, 64.45%; CoHauNOs + H20, 35.55%; N, 1.27%. Found: SiOe.12WOs, 64.27%, 64.20%; CoHaNO,. + H20, 35.78%, 35.80%; N, 1.25%. SUMMARY. 1. Silicoduodecitungstic acid has been prepared in a much more stable form than previously known by a simplification of a method devised by Copaux. 2. ‘The acid was found to have the empirical formula 4H2O.Si02.12W0O3.5H20. The acid formed normal salts with potassium and ammonium in which eight atoms of hydrogen were replaced by a base. 3. A difference in the degree of ionization of the acid hydrogens was observed. in that titrations of the acid by means of a base with methyl orange or methyl] red as indicator show a sharp endpoint when four of these hydrogens have been re- placed. 4. ‘The acid forms stable salts with tertiary organic bases. ‘The salts of a large number of alkaloids have been prepared, the majority of which are so in- soluble as to indicate the possibility of employing them in a quantitative procedure. 5. The salts of the tertiary bases, precipitated from acid solution, show a partial dehydration of the same order as that existing in a dichromate. 6. All of the phases of this investigation are being continued. BIBLIOGRAPHY. 1. Marignac, Ann. chim. phys., III, 69, 6. Bertrand, Bull. soc. chim., III, 21, 434, 5, 1863. 1899. 2. Ibid., IV, 3, 5 (1864). 7. Bertrand, Compt. rend., 128, 742, 1899. 3. Wyrouboff, Bull. soc. franc. mineral., 8. Bertrand and Javillier, Ann. chim. 19, 219, 1896. anal. et chim. appl., 14, 165, 1909; Bull, 4. Copaux, Bull. soc. chim., IV, 3, 101, — Soc. chim., IV, 5, 241, 1909. 1908. 9. Bertrand and Javillier, Anal. chim. 5. Goddefroy, Tagebl. d. 49 Naturf. anal. et chim. appl., 16, 251, 1911. Vers., (1876). Beilage, 83, through Chem. 10. Javillier, Bull. sci. pharmacol., 17, centr., III, 7, 809, 1876. 315, 629, 1910. 14 11. Jbid., 19, 70, 1912. 18. Sindlinger and Mach, Zezt. f. angew. 12. Guillemard, Compt. rend., 132, 1488, Chem., 37, 89, 1924. 1901. 19. Copaux, Bull. soc. chim., IV, 13, 324, 1913. 13. Chapin, U. S. Bureau of Animal Ind., 20. Rosenheim and Jaenicke, Zeit. f. Bull. 183 (1911). anorg. Chem., 101, 241, 1917. ‘ 14. Ferenz and David, Pharm. Puost., 47, 21. Perillon, Bull. soc. ind. min. (1884); 559, 1914. Treadwell-Hall, ‘Analytical Chemistry,” John 15. Rasmusson, Ber, d. deutsch. pharm. Wiley and Sons, 6th ed., H, 277 (1924). Ges., 27, 198, 1917. 22. Jensen, Pharm. Jour., 90, 658, 1918. 16. Taigner, Zezt. f. anal. Chem., 38, 346, 23. Pictet-Biddle, ‘Vegetable Alkaloids,”’ 1919. John Wiley and Sons, p. 321 (1904). 17. Heiduschka and Wolff, Schweiz. A poth. 24. Schmidt, Pharmaceutische Chemie, Fried. Zeil.. OS) ala eLoae. Vieweg and Sohn, Vol. II, Part 2, 1750, 1923. Il. THE USE OF THE ACID AS A VOLUMETRIC REAGENT FOR ALKALOIDS. HISTORICAL. In a preceding paper (1) the authors have described the method of prepara- tion of silicoduodecitungstic acid by a simple and convenient method which yields. a product of very definite and stable character. This acid, as prepared by the authors, was found to have the composition 4H»2O.SiO2.12WO3.5H2,O. ‘The com- position of the normal potassium and ammonium salts was determined, as well as the composition of the salts formed by this acid with alkaloids in hydrochloric acid solution. For the sake of convenience, the acid will be referred to in this article as silicotungstic acid. Goddefroy (2) was the first to call attention to the insolubility of alkaloidal silicotungstates. Bertrand later (3) published the results of a series of investiga- tions, claiming that the work of Goddefroy was unknown to him at the time the work was begun, and alone (4) and with Javillier (5) he studied in detail the proper- ties of salts of this type applying the acid in the quantitative precipitation of nico- tine. Javillier (6) continued the study on the salts of coniine, atropine, sparteine, antipyrine, and pyramidone. The first direct gravimetric method was described by Chapin (7), in which nicotine was precipitated from tobacco extracts and the precipitate ignited, the anhydride of the acid being weighed. Ferenz and David (8) used the acid for the qualitative detection of alkaloids; Rasmusson (9) determined atropine gravi- metrically, applying a correction for the solubility of the atropine salt. Taigner (10), working with atropine, cocaine, and strychnine, found that all of their salts had the same general type formula, which has been confirmed by the authors for all of the monacid tertiary bases studied. Heiduschka and Wolff (11) have studied the proper conditions for the formation of the alkaloidal salts and Sind- linger and Mach have described a method for the determination of pyridine and nicotine in mixtures. The authors, in their previous paper (loc. cit.), have shown that for the mon- acid tertiary bases, the salts have the type formula 4Alkaloid, 2H2O.SiO2.12WOs, and for the diacid tertiary bases, the type formula 2Alkaloid, 2H,O.SiO2.12W0Os3. They have also shown that towards methyl orange and methyl red the acid, in its neutralizing power towards inorganic bases, is tetrabasic. 15 THEORETICAL AND EXPERIMENTAL. The usual methods for the quantitative determination of alkaloids in vegetable drugs or their preparations involve the separation of the alkaloid by the so-called “‘shaking-out process.” In this procedure, advantage is taken of the solubility of the alkaloidal salts in water and of the free bases in the organic solvents which are immiscible with water. Following the purification of the alkaloid, it is de- termined quantitatively by evaporation of the immiscible solvent solution, when the residue is weighed, titrated alkalimetrically, or titrated with a standard solu- tion of one of the alkaloidal precipitants such as potassium-mercuric iodide or io- dine-potassium iodide. Beal and Lewis (13) and Beal and Hamilton (14) have reviewed the literature of the “‘shaking-out”’ process, calling attention to the sources of error therein and have established optimum conditions for these extractions. In this connection attention was called to the fact that the number of extractions required for com- plete removal of alkaloid from either the aqueous or organic solvent phase of a system could not be calculated directly from the simple coefficient of distribution which would assume the absence of acid salts or hydrates of a base. They have also cited the experience of other investigators, together with their own, in which the difficulty of avoiding decomposition of the alkaloidal residue during the final: removal of the solvent, especially chloroform, is shown. In the alkalimetric determination of alkaloids, difficulty is experienced in the selection of a suitable indicator, since the alkaloids are not only weak bases but vary widely in their degree of basicity. McGill (15) and others have recently met with some success in the application of the hydrogen electrode and other methods of electrometric titration to these determinations. Gordin (16) has precipitated various alkaloids from a solution containing a known volume of standard hydrochloric acid by means of a neutral reagent such as iodine-potassium iodide or potassium-mercuric iodide, finding that an equivalent quantity of acid is carried down by the precipitate. He, therefore, titrated acidi- metrically the excess of the standard acid in the filtrate from this precipitate. This method was further studied by Kippenberger (17). Heikel (18) precipitated various alkaloids with a standard solution of potassium-mercuric iodide, determining the ex- cess of reagent in the filtrate by titration with potassium cyanide and silver nitrate. There are at least two objections to these last methods. ‘The methods have, in the first place, the usual disadvantages attendant upon quantitative filtration and back titration. It is also an established fact that the greater number of alka- loidal precipitates of this character begin to decompose after a short time, with the accompanying liberation of iodine. The contribution of the authors of this paper to the gravimetric determination of alkaloids has been in the way of standardizing the method of preparing silico- tungstic acid and determining in some additional salts the constancy of their composition. ‘They have attempted to devise a rapid volumetric method for gen- eral application in the determination of alkaloids which would as far as possible be independent of the degree of ionization of the base and upon which moderate variations in hydrogen-ion concentration would be without effect.. The manner of formation of the alkaloidal silicotungstates suggested the possibility of a volu- metric precipitation method. 16 Heiduschka and Wolff (loc. cit.) added an excess of a standard solution of silicotungstic acid to the alcoholic solution of the alkaloid, filtered off the pre- cipitate so formed and determined the excess of the acid by titration with alkali. The majority of the alkaloidal silicotungstates are best precipitated from a rather strongly acid solution. ‘The authors have therefore made a search for an indicator which will show the presence of an excess of silicotungstic acid in a solution. Because of the fact that derivatives of tungstic acid may be reduced with the formation of lower oxides of tungsten having a blue color, attempts were at first made to use as an indicator a strong reducing agent such as titanous chloride or colorless ammonium sulphide. The solution of titanous chloride showed some prom- ise. It cannot be used as an inside indicator since it will reduce even the pre- cipitated silicotungstate. There is also some difficulty in the preservation of the reagent. As an outside indicator the results are fairly satisfactory, the end-point, a blue color, being rather faint. ‘The ammonium sulphide could not be used with the acid solution in which precipitation took place. The leuco bases of some blue or green dyestuffs, such as reduced malachite green and methylene blue, were next tried, with the idea that the color of the re- developed dye would reinforce the color of the tungsten blue. Again only moderate success was met with, the endpoint being rather faint and the solutions of the leuco bases being very unstable in the presence of atmospheric oxygen. When malachite green is dissolved in hydrochloric acid of moderate concen- tration, the solution instead of being bluish green in color has a tone which is a shade of reddish or brownish orange. ‘The characteristic color of the dye is restored upon dilution with a large amount of water, but one drop of 0.01 molar silico- tungstic acid in 200 cc. of 0.6 normal hydrochloric acid will at once restore the color | of the indicator when one drop of each solution is placed on a porcelain test plate and the drops mixed. ‘The best range of acidity for the proper behavior of the indicator lies between 0.25 and 1.5 normal hydrochloric acid. Preparation of the Indicator—Two grams of malachite green was dissolved in 300 cc. of 6 N hydrochloric acid. Standardization of Silicotungstic Acid.—The molecular weight of the acid is 3006.—An approximately 0.01 molar solution was prepared by dissolving 30 grams of the acid in enough water to make 1000 cc. of solution. A known amount of pure cinchonine was dissolved in 200 cc. of 0.6 normal hydrochloric acid and the solution titrated with silicotungstic acid. After allowing the precipitate to settle, a drop of the clear supernatant liquid was removed to a spot plate by means of a glass tube having an internal diameter of 3 mm. A drop of the indicator solution was then added to this by means of a platinum wire loop of 1.5 to 2 mm. diameter, and the two drops mixed. While the alkaloid was present in solution the color of the mixed drops was deep yellow. With silicotungstic acid in excess the color was bluish green. An endpoint with methyl red is obtained when four hydrogens of the acid have been neutralized. The acid solution was also standardized by titration with standard sodium hydrox- ide, using methyl red as the indicator. The solution of the acid may also be standardized by evaporating a definite volume to dry- ness, igniting the residue and weighing the silicotungstic anhydride formed. Volumetric Assay of Cinchona Bark.—The method used for the extraction of the mixed alkaloids was that proposed by Scoville (19) for the Tenth Revision of the United States Pharma- copeeia. Five grams of powdered cinchona bark were placed in a 250-cc. flask with 15 cc. of water and 5 cc. of hydrochloric acid. After mixing, the contents were digested for 2 hours on the steam- bath, then cooled and 200 cc. of a mixture of 3 volumes of ether and 1 volume of chloroform added, and after thorough shaking 10 cc. of ammonia water. ‘The mixture was then shaken frequently 7 during a period of 12 hours. After the drug had settled, 160 oc. of clear, supernatant liquid were decanted into a separatory funnel and repeatedly shaken/out with 2 normal sulphuric acid. ‘The combined acid extracts were made alkaline with ammonia ant shalveti/gut with chloroform to complete extraction. The chloroform extract was drawn off thréugh a small filter into a beaker and the chloroform evaporated on the steam-bath, expelling the last’ traces by means of alcohol. Instead of weighing the alkaloidal residue it was softened with alcohol and dissolved in a measured volume of 0.1 normal sulphuric acid, and the excess of acid determined by titration with 0.02 normal sodium hydroxide solution, using methyl red as indicator. Following the alkalimetric titration the solution was diluted to a volume of 200 cc. and 20 ec. of 6 normal hydrochloric acid added. The solution was then titrated with approximately 0.01 molar silicotungstic acid, using the same procedure as in standardizing against cinchonine, with malachite green hydrochloride as the indicator. In both the alkalimetric and silicotungstic acid titrations the value 309 was used as the average molecular weight of the mixed anhydrous alkaloids. , The principal alkaloids of cinchona contain two atoms of tertiary nitrogen, but the basicity of one of these is so slight that in the ordinary alkalimetric titrations the alkaloids are regarded as monobasic. Silicotungstic acid, however, reacts with both nitrogen atoms of these alkaloids. TABLE TI. RESULTS OF THE ASSAY OF CINCHONA BARK. Titration with H2S0s titration, 4H20.Si0O2.12WOs, No. Per cent. Per cent. i aa) 5.85 7 in aes ‘es 3 592 5.60 4 Di OW, 5 Ay. 5 Der 610) 6 Derb 5.68 Ave. Dee 5.64 Volumetric Determination of Cinchona Alkaloids without Purification—Ob- viously the original ether-chloroform extract of cinchona bark is unsuitable for the direct gravimetric determination of total alkaloids because of the presence of non- alkaloidal extractives and ammonia. ‘The presence of this ammonia also makes it impossible to extract the mixed alkaloids by shaking out with measured volumes. of standard acid, afterwards determining the excess of acid by titration. However, ammonium salts do not precipitate our reagent in acid solution and the ether- chloroform extraction followed by a shaking-out with sulphuric acid should give a. protein-free solution. It, therefore, seemed possible to save much time and ob- viate a source of error by making a direct titration with silicotungstic acid in the sulfuric acid solution obtained as a result of the first series of shakings. Accordingly in a new series of assays two 80-cc. portions of the ether-chloroform extract were withdrawn. One portion was treated according to the full U. S. P. method, the mixed alka- loids being weighed, titrated with sulphuric acid and silicotungstic acid. The other portion was extracted with 2 normal sulphuric acid, the combined extracts diluted with water to 200 cc. and 10 cc. of 12 normal hydrochloric acid added. ‘This solution was then titrated with silicotungstic acid in the 11sual manner. The volume of silicotungstic acid solution used in these titrations averages. about 19cc. ‘The volume of standard sulphuric acid actually consumed is only about 4cc. In spite of the fact that the excess of sulphuric acid is determined by titration 18 i a TABLE II. Aa 7" Titration with SPE Aiea Titration with 4H20.Si02.12WOs No. it | Gravimetric. _, H2SO: titration. 4H20.SiOx.12WO3. without purification. 1 6.690 Lh 5.85 5.90 59079 2 6.19 5.87 5.85 5.90 3 6.37 6 .0O 5.95 5.99 4 5.83 Sn ft 5.68 5.84 5) 2.91 3.95 5.84 5.89 6 5.99 5.70 5. .68 5.92 Ave. 6 .00 5.84 5.82 o.91 with a much weaker alkali, there is a potential source of error in the measurement of the sulfuric acid which is greater than the possible error in the endpoint with the silicotungstic acid. Various chemists have pointed out a possible source of error in both gravimetric and alkalimetric determinations of alkaloids due to the presence of ammonium salts in water solution dissolved in the immiscible solvent. On evaporation to dryness there is a possibility that the alkaloid, as a non-volatile base, will expel ammonia, forming an alkaloidal salt. White spots have frequently been observed in such residues which upon examination have proved to be the alkaloidal salt, as chloride or sulfate, entirely free from the ammonium radical. This will naturally give high results in the gravimetric assay, and, being neutral, will give low results in the alkalimetric titration. The reaction with silicotungstic acid is independent of the presence of an acid forming a soluble salt, and such a form of contamination will not interfere in the proposed method of determination. Assay of Fluidextract of Cinchona.*—These preparations were assayed according to the Scoville modification previously mentioned. Five cc. of the fluidextract were mixed with 5 grams of purified sawdust in a 250-cc. flask and carefully dried on the steam-bath. This residue was then digested with hydrochloric acid and extracted in the same manner as cinchona bark. The chloroform residue was weighed, and titrated with sulphuric acid and sodium hydroxide, and the sulphuric acid extract titrated with silicotungstic acid as above. TABLE III. ASSAY OF FLUIDEXTRACT OF CINCHONA. Titration with Gravimetric H2S0s, titration 4H20.Si02.12WOs Gms. per 100 cc. Gms. per 100 cc. Gms. per. 100 cc. Red cinchona 3.41 3.26 3.36 3.29 3.32 3.39 Cinchona!/ Usa.P: 3.34 3.21 3.36 3.39 3.35 3.33 Assay of Nux Vomica.—This was assayed by the method of the U. S. P. IX, weighing the chloroform residue before the alkalimetric titration and titrating these solutions with silico- tungstic acid after adding hydrochloric acid to the neutral solution. The sulphuric acid extract was also titrated as in the assay of cinchona. * Samples of percolates of cinchona and of red cinchona were furnished by Parke, Davis and Company through the courtesy of Mr. F. O. Taylor. 19 TaBie IV. i] ASSAY OF NUX VOMIGA! } | 4 isp gy : I } Titration with Pane H2SO4 Titration with , 4F1O.Si02.12WOs No. Gravimetric. titration. 4H20.Si02.12WQ;/ without purification. 1 3.08 2.49 3.06 >” 2 .64 2.09 2.0 3 2.54 1.85 2.41 4 2202 2:38 2.88 Re ee 5 3.56 6 Boe 7 3.34 8 3.40 Since the titrations with silicotungstic acid gave values which were very much higher than those by the alkalimetric method, the silicotungstic acid which had been standardized against cinchonine was titrated against pure strychnine and brucine. TABLE V. Amt. by Amt. by H25O4 4H20.SiO2.12WO3 Per cent recovery by Wt. alkaloid. titration. titration. 4H20.5i102.12W0Os. eaiGd AG. cinchonine “ot ate 0.1018 100.0 0.2109 Gm. strychnine 0.2080 0.2492 118.1 0.2066 Gm. strychnine 0.2042 0.2430 he Pes Ones oD) oy a blest eC: Re hie. a i 2 Nr Aa 0.2766 133 .4 OPDet Grim bricines fees chives 0.2715 132 .2 0.2065 Gm. brucine 0.2070 a a te Mee ght) RL NE oie os aeons 0.2090 Gm. brucine GR TOON Ma tach ans Sa). tS ol Ol te aR hig The apparent result in the brucine titration was 32.8% too high and for strychnine 17.8% too high. The average for both was 25.8%. With this correction the average value in the assay of the drug became 2.74%, a very fair agreement with the gravimetric assay. Assay of Belladonna Leaves —Belladonna leaves in No. 60 powder were extracted according to the method of the U. S. P. IX, and the final chloroformic residue weighed and titrated with sulphuric acid and sodium hydroxide, using methyl red as indicator. Other portions of the drug were analyzed by titrating the sulphuric acid solution with silicotungstic acid before the final purification as in the assay of cinchona bark. Because of the solubility of atropine silicotungstate in dilute acids (9), the sulphuric acid extract was titrated without dilution with water, and a cor- rection of 0.0048 Gm. of atropine per 100 cc. of solution was made. TABLE VI. ASSAY OF BELLADONNA Root. Titration with 4H20.Si0O2.12WO3 No. Gravimetric. H2SO, titration. without purification. i! 0.410 A re 90 FO La a Te Boe 2 0.415 Bee Aon wet od ee ORR ela oe are ae Tl a) OS ne eis eer 0.420 et Nae ON eal Ne, Gi yO a rh ss , 0.440 Oe en A ee EON Pd on hs a Kia 0.445 a MME at CY RS EA? iro ON ae Yee Se oe 0.445 Assay of Stramonium Leaves.—Stramonium leaves were assayed in the same manner as the belladonna leaves above, determining the total alkaloids volumetrically by the U. 5S. P. process and also by titration with silicotungstic acid without the final purification. The same correction was applied as in the assay of belladonna leaves. 20 uA TaBLE VII. . Assay \o#\StRAMONTUM LEAVES. 4 , Titration with 4H20.Si0O2.12WOs \A \ syn Re © 4\ \) HeSOs titration, without final purification. No. | Per cent. Per cent. ss LSP ee A ae eee 0.456 ; De nh SR ASC a i ees See 0.466 3 OSELOS ah cn'| he OG Sea ee cee fe 4 FAC ea se ee nc tar ee Assay of Hydrastis Root.—Hydrastis root was assayed gravimetrically according to the method of the U. S. P. IX, and by silicotungstic acid without the final purification. TaBLE VIII. ASSAY OF HypRASTIS Root. Titration with 4H2O.SiO2e.12WOs U. S. P. Gravimetric, without final purification. No. Merecent. Per cent. 1 3.44 2 3.52 a eo 3 3.09 4 3.15 3) 3,32 6 3.29 Realizing the possibility of error due to the solubility of the alkaloidal silicotungstates, Nos. 5 and 6 were titrated by adding the silicotungstic acid in excess, allowing the precipitate to stand for thirty minutes, then filtering and washing with 1% hydrochloric acid. The excess of silicotungstic acid in the filtrate was titrated with a standard solution of cinchonine hydrochloride. A Source of Error in the Method of Assay by Aliquot Parts.—In practically all of the alkaloidal assays of the U. S. P. IX, the alkaloids are extracted in the free TABLE IX. Loss BY EVAPORATION OF PROLLIUS SOLUTION IN CINCHONA ASSAYS. Weight of flask Weight of flask and solution, and solution, No. before. after. Loss 1 324.5 324 .0 0.5 2 331.5 dol .0 0.5 3 319.0 318.5 0.5 4 305.1 304.7 0.4 3) 320.8 320.8 0.0 6 312.7 312.4 0.3 (j 325.4 323 .8 1.6 8 305.1 304.7 0.4 9 325 .0 324.5 0.5 10 310.0 310.0 0.0 11 326.5 324.8 ier Ds 317 .0 317 .0 0.0 13 311.0 311.0 0.0 14 316.0 315.0 12.0 15 305.5 304.5 LEQ) 16 318.0 315.0 3.0 Ave. 0.7 Gm. state by macerating with some modification of Prollius mixture. F ollowing this maceration, the drug is agglutinated, if necessary, by the addition of a small 21 amount of water and a definite volume of the non-aqueous solution / decanted. The final calculation is based upon the supposed relationship! jof| this poems of solution to the volume of liquid with which the drug was originally, macerated. Unless perfectly stoppered vessels are used for the maceration, there is ‘inevitably some loss of solvent during maceration, an effect which is oftentimes intensified by shaking. For this reason we have made it our practice to weigh the extraction flask at the beginning of the maceration and just before the removal of the aliquot, making up to the original weight if necessary by adding more of the non-aqueous solvent. We are of the opinion that the Pharmacopceia should contain a caution- ary statement regarding this loss. The foregoing table contains the weights ob- tained in a series of extractions of cinchona, the macerations being made in Erlen- meyer flasks having well-fitting ground glass stoppers. SUMMARY. 1. Aqueous solutions of silicotungstic acid restore the green color to an orange- colored solution of malachite green in hydrochloric acid. 2. Using the hydrochloric acid solution of malachite green as an outside indicator, it has been found possible to titrate alkaloidal salts with standard solu- tions of silicotungstic acid in the presence of free hydrochloric or sulphuric acid. 3. The results so obtained by slightly modifying the U.S. P. assay processes compare favorably with those obtained by the official method. 4. It has been found possible to materially shorten the official methods by titrating the first sulphuric acid extract with silicotungstic acid. 5. Attention is called to the possibility of losses of solvent in the initial ex- traction of the drug in the official assay processes. 6. A further study is being made of the possibilities of this reagent in the volumetric determination of alkaloids. BIBLIOGRAPHY. 1. North and Beal, Jour A. Pu. A., pre- 10. Taigner, Zeit. anal. Chem., 58, 346, vious article. 1919. 2. Goddefroy, Tagebl. d. 49 naturf. wissen., 11. _Heiduschka and Wolff, Schweiz. A poth. p. 83 (1876). Leite, -58,- 2191920. 3. Bertrand, Bull. soc. chim., (III) 21, 12. Sindlinger and Mach, Zeit. angew. 434, 1899. 89, 1924. 4. Bertrand, Compt. rend., 128, 742, 1899. 5. Bertrand and Javillier, Ann. chim. Gui 145 16550909 16;4251, 1911; Bull. -soc. chim., (IV), 5, 241, 1909. 6. Ujavillier, ~ Bull, set. 315, 629, 1910; 19, 70, 1912. 7. Chapin, U. S. Bur. Animal Ind., Bull. 133 (1911). 8. Ferenz 47, 559, 1914. 9. Rasmusson, Ber. Ges., 27, 193, 1917. URBANA, ILLINOIS, AvucustT, 1924. pharmacol., 17, and David, Pharm. d. deutsch. pharm. Chem., 37, Jost 13. ° Beal and Lewis, Jour. A. PH: A., 5, 812, 1916. 14. Beal and Hamilton, Jour. A. Pu. A., 9, 9, 1920. 15. McGill, PMS oat AS Vopr 16. Gordin, Ber., 32, 2871, 1899. Jour. Am. Chem. Soc., 44, 17. Kippenberger, Zeit. anal. Chem., 42, 101, 1908. 18. Heikel, Chem. Zeit., 32, 1149, 1908. 19. Scoville, Jour. A. Pu. A., 9, 867, 1920. VITA Edward Oscar North was born October 29, 1895, in Rockford, Illinois. His early training was received in the Rockford schools. In 1914 he entered Beloit College where he was graduated in 1918 with the degree of Bachelor of Science. In 1922 he received the degree of Master of Science from the University of Llinois. He has held the following positions: Assistant in Chemistry, Beloit College, 1918-1919. Instructor in Chemistry, Beloit College, 1919-1920. Assistant in Chemistry, University of Illinois, 1920-1924. He is a member of Sigma Xi and Phi Lambda Upsilon. ee 3 0112 061414881 ”