pwisET" 1 ISOLATION OF NITROGEN BASES FROM CALIFORNIA PETROLEUM DISTILLATES Approved: Approved: Dean Graduate School. ISOLATION OF NITROGEN BASES FROM CALIFORNIA PETROLEUM DISTILLATES Ir MAZ THESIS Presented to the Faculty of the Graduate School of The University of Texas in Partial Fulfillment of the Requirements For the Degree of DOCTOR OF PHILOSOPHY By Willard Curtis Thompson,B.A.,M.3. (Austin, Texas) Austin, Texas June, 1930 Q I O f r- 6 ±'j Preface The experimental work described in this paper was carried out in connection with Project No. 20 of the American Petroleum Institute under the direction of Dr. J. R. Bailey. I wish to express to Dr. Bailey my sincere appreciation for the generous assistance and encouragement extended to me throughout the work. Willard c. Thompson Austin, Texas June, 1930. CONTENTS Page Introduction - 1 Results and Discussion 15 Experimental Part (1) Preliminary investigation 45 (2) Isolation of bases from California petroleum distillates 57 (3) Investigation of the bases, CnHygN and 61 Summary and Conclusion 87 Introduction ISOLATION OF NITROGEN BASES FROM CALIFORNIA PETROLEUM DISTILLATES An extensive program for fundamental research on petroleum was recently made possible by the contribution of the sum of $500,000 by John D. Rockfeller and the Universal Oil Products Company. The fund is administered by the American Petroleum Institute with the cooperation of the Central Petroleum Committee of the National Research Council. Project 20, entitled, Isolation and Investigation of the Nitrogen Compounds Present in Petroleum, was assigned to the Chemical Laboratory of the University of Texas with Dr. J. R. Sailey as Director. The investigation of the nitrogen compounds in petroleum was begun in 1926 and is now in its fourth year. The Union Oil Company of California has rendered indispensable assistance to the prosecution of th© research on nitrogen compounds, not only in contributing samples of crude petroleum and various refinery products for the preliminary investigations, but also by the extraction of nitrogen bases from large quantities of petroleum distillates. Without this assistance it would have been impossible to obtain a sufficient supply of nitrogen bases for an extended investigation. Acknowledgement is made to Mr. R. E. Haylett, Technical Assistant of the Union Oil Company, who has displayed great interest in the problem and has given generous cooperation in th® matter of obtaining the material for the investigation. This paper deals in general with the isolation and identification of nitrogen bases from California petroleum distillates and more particularly with the isolation and determination of structure of two bases, OHM and 11 X & C_ In order to give a connected account of the work carried out by the author in connection with Project 20 it has been necessary to include a brief discussion of the work of other investigators, to whom credit is given at the proper places. The presence of nitrogen compounds in petroleum was noted as early as 1817 by Saussure 1 who obtained a small amount of ammonia on exploding a mixture of the vapors of crude naptha from Amiano with an excess of oxygen in a eudiometer. The actual amount of nitrogen occurring in crude oils is small, varying from a trace to 2 per cent of 2 nitrogen as shown by a table compiled by Engler and Hofer from analyses of crude oils from representative oil fields 3 of the world. Mabery, using the Dumas method, found a mini- mum of 0.91 par cent, a maximum of 2.39 per cent, and an average of 1.66 per cent in 16 samples of California crudes analyzed. Mabery also stated that ’’the Kjeldahl method evolves only a part of the nitrogen as ammonia." Engler and Hofer*j giving Takano as their authority, state that the nitrogen content of Japanese petroleum was estimated by the Lumas method since the Kjeldahl method cannot be used with Japan-4 ese petroleum. Later, Mabery, analyzed 21 samples of crude petroleum from various parts of the United States by a modified Lunas method, obtaining values ranging from 0.048 per cent to 0.01 per cent. Those results chocked fairly well with determinations made by the Kjeldahl mothoi. No California crudes were included in this series of analyses. More recently, Poth, Armstrong, Cogburn, and Bailey 0 carried out an extensive investigation of the nitrogen content of crude petroleum and bitumens of the North American Continent, analyzing hundreds of samples from the more important fields. These investigators obtained concordant results with modified Dumas and Kjeldahl methods. No crude petroleum was found to contain mote than 0.815 per cent nitrogen, while the minimum for California crudes was found to be 0.102 per and the average 0.473 per cent. It should be pointed out that California pctroleums have an unusually high nitrogen content, greatly exceeding that of most crudes. The range in the amount of nitrogen occurring in the petroleum of the principal producing states is listed below: * Compiled from analyses of crude petroleum carried out at the University of Texas Laboratory Although it has been stated frequently in petroleum literature that pyridine and quinoline bases exist in petroleum, very little is known about the nature of the nitro gen compounds occurring in crude oils. Many investigators have obtained mixtures of nitrogen bases from crude petroleum or petroleum distillates by extraction with dilute acids, but it is doubtful that any individual compound has been isolated and its structure determined, The usual procedure consists of the agitation of crude petroleum or petroleum distillates with dilute acid, extraction of the acid solution with ether to remove hydrocarbons, and the neutralization of the acid solution with sodium hydroxide to free the bases. The alkaline residue is extracted with ether to insure complete removal of the bases, fractional distillation or precipitation of bases as salts is usually resorted to in the separation of individual substances. Weller extracted paraffin oils from Saxony with dilute sulfuric acid and obtained a mixture of bases, with a pyridine-like odor, which gave a yellow precipiatate 7 with picric acid. Bandrowski obtained a mixture of solid bases on the extraction of Galician petroleum with 10 per cent sulfuric acid. His bases gave characteristic tests Q for alkaloids and formed platinum chloride salts. Chlopin states that Caucasian petroleum contains 0.005 per cent of -pyridine bases of the general formula, C which he n 2n-15 obtained as a dark brown oil by acid extraction. Six dif- ferent platinum chloride salts were obtained with molecular weights varying from 104 to 308. The most complex fraction yielded numbers which agree with those of the formula, or base and (C22H29N) for the platinum chloride salt. By acid extraction of Rumanian petroleum Griffiths and Bluman 9 obtained a thick oily liquid boiling at 117° which had the composition of pyridine but was insoluble in water. On reduction the base yielded piperidine. As indicated above other workers investigated the nitrogen bases occurring in petroleum distillates. Zal- diluted the acids used in the refining of dis- tillates, added quicklime, and distilled the liberated bases with steam. The mixture of bases obtained exhibited many of the properties of pyridine. extracted distillates of crude oil from ? t omanisk with dilute sulfuric acid, and, after neutralization of the acid extract, ob- tained a brown'liquid having a characteristic pyridine odor. Those bases were insoluble in water and distilled undecomposed between 260° and 370 • On analysis the mixture was found to contain 6.60 per cent nitrogen, 85.72 per cent carbon, and 8.09 per cent hydrogen. The mixture yielded salts with mineral acids which were soluble in water and. alcohol. Picric acid, mercuric chloride, and dichromic acid yielded amorphous, insoluble salts. Crystalline salts were obtained with platinum chloride and lead chloride from acid solution. obtained amines from Baku petroleum. The acid sludge from kerosene agitators was diluted with water and steam distilled. The distillate was extracted with dilute sulfuric acid and the bases set free with caustic. Low boiling bases were obtained which fumed with hydrochloric acid and gave amorphous precipitates with platinum chloride. Pyhala expressed the opinion that these bases did not exist in petroleum but were produced, during the refining process, from more complex compounds. In 1899 began an investigation of nitrogen bases which were extracted with dilute sulfuric acid from fuel oil and lubricating oil distillates by Pecham and Solathd. The crude oil from which the distillates were obtained came from the Santa Paula District of California. The bases were distilled several times in a vacuum and 2 to 4° cuts were made. The fractions were analyzed and mole cular weight determinations made with results as shown below: Mabery was unable to prepare crystalline salts from these fractions. He found that the bases formed addition products with ethyl iodide, yielded nothing except a small amount of acetic acid on chromic acid oxidation, and when oxidized with permanganate in alkaline or neutral solution yielded their nitrogen as nitrogen or ammonia. All of the nitrogen was evolved as nitrogen on oxidizing with potassium permanganate in acid solution. No trace of pyridine or quinoline bases were found by the author who says: That these bases are totally unlike the quinoline derivatives is evident by their behavior towards exudation agents. Quinoline and its homologues readily give acids on oxidation but under no conditions could anything be obtained from these bases except, as shown above, a little acetic acid. There is no doubt of the existence of a tetrahydro quinoline ring as a part of the molecule, and the evidence of a benzol ring is strong. No doubt their hydro-condition accounts for their instability. Aside from the oxidation experiments the only evidence advanced that the bases are hydrides of quinoline consists of the following observation: "The proportion of carbon and hydrogen can only be explained by assuming the oils are tetrahydro bodies. The oil collected at 130°-140° corresponds in boiling point and composition with tetrahydro-Bz. 1-3, Py. 2 trimethyl quinoline." Seventeen years later the fractions described above 14 were re-investigated by Mabery and Wesson. At this time the material was described as follows: "these samples were brown, viscous to very viscous liquids, one or two of the highest fractions becoming amorphous resinous solids when cooled." The various fractions were oxidized with potassium permanganate in aqueous solution as shown below Three grams of a fraction was suspended in a 2 per cent solution of potassium permanganate and the solution boiled until the permanganate became decolorized. A portion of solid permanganate was then added and the heating continued until the solution became colorless. This process was repeated until the permanganate ceased to be decolorized. The reaction mixture was filtered and the filtrate concentrated by distillation after acidifying with nitric acid. The distillate was investigated for volatile acids without success. The concentrated residue was evaporated to dryness and the solid residue extracted with hot alcohol, which dissolved a part of the material. The alcohol insoluble residue was dissolved in water and gave a white, voluminous precipitate on the addition of silver nitrate. The silver salts obtained in this way from various fractions were analyzed and in each case the results obtained were in fair agreement with the calculated values for the tetra-silver salt of pyridine pentacarboxylic acid. The authors failed to isolate the free acid despite the fact that this substance had been prepar-1 5 ed by Hantzsch in crystalline form. The barium salt of the acid was also prepared and on distillation yielded a base which was converted to a picrate. The picrate had the melting point of the picrate of pyridine, and a mixed melting point with the known picrate of pyridine exhibited no lowering. The alcoholic extract also gave a white voluminous precipitate with silver nitrate. All of the silver salts obtained in this manner on analysis gave values which were in agreement with the calculated values for the di-silver salt of methyl pyridine tetracarboxylic acid. Here, again, the free acid was not prepared, although described in the . , . 16 literature. The calcium salt of a third acid was obtained from a chromic acid oxidation of one of the fractions. When distilled the calcium salt yielded a base whose picrate melted at 187-188? From this data Mabery concluded that he had obtained the picrate of-methyl quinoline. It was observed that the fractions did not give the Liebermann test for secondary amines. Several fractions were reduced with sodium and alcohol and the resulting products gave a positive Liebermann test. The reduced bases were not identified. In this paper Mabery reversed his earlier views on the nature of the bases under investigatL on. He now concluded that the organic bases of California petroleum consist mainly of an indefinite fixture of alkylated quinolines or iso-quinolines, which, as regards the pyridine ring, are completely alkylated. In view of his failure to obtain volatile acids during the permanganate oxidation Mabery assumed that the alkyl side chains consist of small groups only. The view that the nitrogen compounds in crud© petroleum are mainly bases and may ba easily extracted there- 17 from seems to be prevalent in petroleum literature. Beilby stated that the nitrogen compounds are present in petroleum , . . , _ .. i • .1 , 15,14,13 as bases which may be removed with sulfuric acid. Mabery on several occasions expressed toe belief that California crude petroleum may contain as much as 10 to 25 oer cent of nitrogen bases and that those bases may be extracted from the crude oil with dilute acids. Hamor and O X. o 1 say, "Nitrogen is present in crude petroleum generally in the form of derivatives of pyridine or quinoline." 20 However, Day ' states that none of the investigators has accounted for tho total nitrogen in various oils by the amount of pyridine bases extracted, ft is of interest 21 to note that Mabery attempted to extract the nitrogen from 21 samples of crude oils from Ohio, Pennsylvania, Texas, Kansas, and West Virginia with dilute acid but was able to remove only a part of the nitrogen as bases. He describes the following experiment: As a qualitative test one or two liters of the crude oil was vigorously agitated during several hours with dilute hydrochloric acid, the aqueous solution neutralized with sodium hydroxide and extracted with ether. Every specimen examined left on evaporation, an oily residue with the characteristic odor of the similar bases extracted from California and Russian petroleum in much larger amounts. The smallest yields were observed from the light Berea Grit and Pennsylvania crudes, but even these slight residues dissolved in acid and again precipitated with alkali. Since acids do not remove completely the basic constituents, quantitative results were not possible by this method. 9 Pyhala expressed the belief that the bases which he obtained from kerosene distillates did not exist as such in petroleum. 1. de Saussur®, Theodore: "Recherches sur la composition du naptha d'Amiano, dans les stats de Parma.” Annales chim. phys. 4, 314 (1817). 2. Bngler,C ~ and H. Ho’fer, "Das Erdol", Hirzel, Leipzig, 1913, Vol.l, p. 479. 3. Mabery, C.F.; "On the Nitrogen Compounds in California Petroleum." JS. Soc. Chern. Ind. 19, 502 (1900). 4. Mabery, Chas.F.: "The Genesis of Petroleum as Revealed by Its Nitrogen Constituents." J. A. C. 3. 41, 1691 (ISIS). 5. Poth,2.J., W.D.Armstrong, C.C.Cogburn, and J.R.Bailey: ’’The Estimation of Nitrogen in Petroleum and Bitumens.” Ind. Eng. Chern. 20, 83 (1928). 6. Weller, A.: "Vorkommen alkaloidartigen Basen in ParaffinSl." Ber. 20, 2098 (1887). 7. Bandrowski, Franz X.: "Über das Vorkommen alkaloidartiger Basen im galizischen Roherdole." Monats. 8, 224 (1887). 8. Chiopin, G.W.: "Die organ!schen Basen des russischen Srddlas. " Ber. 33, 2837 (1900). 9. Griffiths, A.B. and N.J.Bluman; "Les bases azotees dans le petrole roumain." Bull. Soc. Chim. (iii) 25, 725 (1901). 10. Oaloziecki, Homan: "Über pyridinartige 3asen im Srdßl." Monats. 13, 498 (1892). 11. Schestakow, P.J.: "Stickstoffverbindungen des kaukasischen Brdols." Chern. Ztg. 23, 41 (1899). 12. Pyh'dld, Ewald: "Einige neue t stickstoffhaltige Kohlenv/asserstoff e aus dem Bakuer Brdol.” Chern. Ztg. 46, 953 (1922). 13.Mabery, C.F. : ”0n the Nitrogen Compounds in California Petroleum.” Jr. Soo. Chern. Ind. 19, 505 14. Mabery, C.P. and L.G.Wesson: "The Constitution of the Organic Nitrogen Bases of California Petroleum."J.A.C.S. 42, 1014 (1920). 15. Hantzsch, Arthur: "Ueber die Synthese pyridinartiger Yerbindungen aus acetossigather und Aldehydammoniak." Ann. 215, 62 (1882). 16. Ibid. p. 57. 17.Beilby, George: ’’The Nitrogen of Crude Petroleum and Paraffin Oils." Jr. Soc. Chern. Ind. 10, 120 (1891). 18. Mahery, C.F. : "Composition of American Petroleum." J. A. C. S. 28, 426 (1906). 19. Hamor, W. A. , and Padgett: "The Examination of Petroleum," McGraw-Hill Book Co., New York, 1920, Ist ed., p. 9. 20. Day, D.T.: " A Handbook of th© Petroleum Industry," John Wiley and Sons, New York, 1922, Vol.l, p. 529. 21. Mabery, C.F.: "The Genesis of Petroleum as Revealed by its Nitrogen Constituents." J. A. G. S. 41, 1693 (1919). State % nitrogen Arkansas 0.051 - 0.088 California 0.101 - 0.815 Colorado 0.000 - 0.101 Indiana - 0.026 Kansas 0.040 - 0.098 Kentucky 0.054 - 0.062 Table I Range in the nitrogen content in crude petroleum of the principal producing states* Table I (continued) State % nitrogen Louisiana 0.006 - 0.090 Montana - 0.042 Oklahoma 0.036 - 0.165 Ohio 0.000 - 0.000 Pennsylvania ■ 0.000 - 0.000 Texas 0.014 - 0.135 Utah 0.000 - 0.000 ’Yost Virginia 0.000 - 0.000 Wyoming 0.017 - 0.317 Fraction Formula 130°-140° 197°-199° C 13 H 18 H 215°-217° 223°-225° C 15 H 19 H 243°-245° C 16 H 19 H 270°-275° C17H21N Table II RESULTS AND DISCUSSION 13 17 18 Contrary to the claims of Mabery ’ * preliminary 22 experiments " at the University of Texas laboratory indi- cated that only very small amounts of nitrogen bases may be extracted from California crude petroleum with dilute mineral acids. Experiments with McKittrick, Coalinga, and Ventura crude oils yielded minute amounts of basic oils, not more than 1 to 2oc. per kilogram of petroleum extracted. If the nitrogen of California petroleum is in the form of bases, assuming that the average nitrogen content of the bases Is as high as 10 per cent, McKittrick crude oil which contains 0.63 per cent nitrogen should yield 63 grams of bases per kilogram of oil extracted. That there is nothing p p in crude petroleum to prevent the extraction of bases if they are present is demonstrated by the fact that quino- line which has been added to crude oil may be quantitatively extracted with dilute mineral acids. These facts are in agreement with the observations made by that "none of the investigators has accounted for the total nitrogen in various oils by the amount of pyridine bases extracted." It is noteworthy that Mabery 2l failed to quantitatively extract the nitrogen from 21 samples of crude oil oxamined. However, no California petroleum was included in this experiment. When Ventura crude oil is treated with concentrated sulfuric acid in the cold practically all of the nitrogen is removed, most of it going into the dark tarry sludge formed, and, a small part, into the acid. This observation is in accord with the result obtained in refining petroleum distillates on the large scale. Concentrated sulfuric acid, therefore, in contrast to dilute sulfuric acid withdraws nitrogen, not so much through salt formation as through solvent action, sulfonation, or in other ways as yet undetermined. The distillates from crude petroleum were found to yield appreciable amounts of bases on extraction with dilute acids. However, the non-basic nitrogen compounds predominate in most fractions. Analyses of McKittrick crude oil and successive fractions obtained from it by distillation show that the first fractions are poor in nitrogen and that the amount of nitrogen increases with the distillation temperatures, reaching a maximum in the 24 residue." A similar phenomenon was observed by Beilby on the distillation of oil bearing shales, Furthermore, it was found that the ratio of basic nitrogen to non-basic nitrogen is higher in the lower boiling fractions than in the higher boiling fractions. For instance, dilute acid extraction of a light lubricating oil distillate reduces the nitrogen content from 0.11 to 0.07 per cent, while the nitrogen content of an extra heavy lubricating oil distillate is reduced from 0.639 to 0.542 per cent by similar treatment. In the first case the ratio of basic to non-basic nitrogen is 1: 1.75, and, in the second, 1: 5.59. Among the petroleum products, which wore investigated as a possible source of nitrogen bases, was a sample of extra heavy lubricating oil distillate obtained from the Oleum refinery of the Union Oil Company of California. After several hours agitation with an equal volume of 16 per cent sulfuric acid, the acid extract was neutralized with caustic soda solution and yielded a dark brown viscous basic oil having an odor similar to that of quinoline. 1.5 g. of bases was obtained for each liter of oil extracted. After the distillation of 1 liter of the extracted oil at atmospheric pressure, 6.6 g. of additional bases was obtained on acid extraction. Another 1 liter of the crude distillate was distilled at atmospheric pressure prior to acid extraction and yielded 8.7 g. of bases. Distillation of the distillate under a reduced pressure of 7-10 mm. increases the yield of bases only slightly, from 1.5 g. to 1.9 g. During the distillation at atmospheric pressure the odor of ammonia was noticeable, which, together with the increase in the amount of acid soluble nitrogen compounds, indicates that some decomposition of the non-basic nitrogen constituents of the oil was taking place. As in the case of the crude oil the lower boiling fractions from the distillation of the lubricating oil distillate are poor in nitrogen, the amount of which increases as the distillation temperature rises, reaching a maximum in the residue. Petroleum lubricating oil distillates are refined by agitating with concentrated sulfuric acid, the nitrogen compounds passing into the sludge and ’’black acid". An inpa vestigation of the sludge and acid washings from the refining of lubricating oil showed that considerable amounts of bases may; be obtained from these sources. If steam is passed through the heated sludge 2g. of bases may be extracted from the oil layer of the distillate for each kilogram of sludge treated. The yield of bases may be increased to 6g. of bases per kilogram of sludge by fusing the sludge with caustic soda or lime before distilling with steam. Neutralization of the ’’black acid” with caustic soda yields Bg. of bases per kilogram of acid. Another product furnished the Texas Laboratory by tho Union Oil Company was a sample of the residue from the refining of kerosene distillate by the Kdeleanu process. Briefly, the Bdeleanu process consists in agitating the kerosene distillate with 75 volume per cent of liquid sulfur dioxide. The sulfur dioxide is separated from the treat ed oil and evaporated for re-use leaving a residue, known as kerosene extract, which represents 30 per cent of the stock treated. The sulfur dioxide exerts a solvent action removing olefinic hydrocarbons, aromatic hydrocarbons, sulfur and sulfur compounds, and nitrogen compounds from tho kerosene distillate. The kerosene extract is not a waste product. It is used in the preparation of special solvents, in insecticides, and is added to motor fuels because of its detonation suppressing qualities. Considerable amounts of the kerosene extract are produced since four California refineries using this process have a total capacity of 10,000 barrels per day. The fundamental difference between the Edeleanji process and sulfuric acid treatment lies in the fact that, in the latter case, chemical action between the acid and certain constituents of the crude distillate occurs, while in the Edeleanu process the action is merely physical. The sludge produced in sulfuric acid refining represents a groat economic loss, whereas the residue from the Edeleanu process is a valuable by-product. In the case of distillates high in sulfur or rich in aromatic compounds the Edeleanu process effects the desired results where sulfuric acid treatment is too expensive or insufficient. The kerosene extract from the Oleum refinery of the Union Oil Company contains 0.055 per cent nitrogen which is reduced to 0.025 per cent on extraction with an equal volume of 16 per cent sulfuric acid. Neutralization of the acid extract yields a dark brown mixture of bases containing 6.75 per cent nitrogen. The yield is 2.75 g. per liter of kerosene extract. Distillation of the kerosene extract previous to the acid extraction seems to increase the amount of acid soluble compounds since the nitrogen content is reduced from 0.055 per cent to 0.014 per cent by such treatment. Soon after the inception of the investigation of the nitrogen compounds of petroleum it was decided to investigate first the bases present in petroleum distillates rather than the nitrogenous substances occurring in crude oil. It was hoped that such an investigation would throw some light on the nature of the inert nitrogen compounds in crude petroleum and might yield information of value to the refining industry. The separation of the bases from distillates is a relatively simple matter and their investigation Was expected to yield more immediate results. The nature of the inert nitrogen compounds in petroleum is unknown and methods for their separation have yet to be developed. It is possible that fractional extraction of crude petroleum with liquid sulfur dioxide may be used in obtaining these substances from petroleum or concentrate them to the point where their isolation becomes feasible. It is obvious that a very large amount of material must be treated to secure a sufficient amount of crude bases for an extended investigation. 50 kilograms is about the smallest quantity of crude bases that may be investigated with the probability of obtaining sufficient amounts of individual bases for the determination of structure. It was subsequently found that the base present in largest amount in 50 liters of crude bases investigated at this laboratory does not represent more than 1 per cent of the totaj bases. To obtain such a quantity of bases by simple acid extraction of extra heavy lubricating oil distillate, which yields 1.5 g. of bases per liter, would require the treatment of approximately 30 tons of crude distillate. To obtain the same amount of bases from the "black acid” from the refining of lubricating oil distillates would necessitate the neutralization of 6 tons of 50-60 per cent sulfuric acid. It would be necessary to extract approximately 18 tons of the kerosene extract to obtain 50kg. of bases. That it is impractical, if not impossible, to work up such large quantities of material in a university laboratory is readily apparent. Therefore the cooperation of some refinery utilizing crude petroleum of a high nitrogen content is imperative. Furthermore the procedure employed has to be as simple as possible and to involve little interference with routine factory operation. While a refinery is able to carry out an acid extraction of kerosene extract or light lubricating oil distillate without an inordinate use of time and equipment, one could not expect a plant to carry out special distillation of crude distillates prior to acid extraction or to prepare and steam distill caustic fusions of sludge. The decision to extract a considerable quantity of kerosene extract and lubricating oil distillate with dilute sulfuric acid was governed largely by the above considerations. In addition, the bases so obtained are in the same condition Ln which they exist in the original distillate since they have not been subjected to chemical treatment. However, any bases obtained from the "black acid" or sludge from the sulfuric acid refining of distillates have been changed chemically in all probability and may not be the same acid soluble nitrogen compounds originally present in the distillates. The Union. Oil Company through Mr- R. E. Haylott, Technical Assistant, offered, to carry out the extraction of any amount of kerosene extract and to permit a chemist from the Texas Laboratory to assist in the work. Accordingly, Hr. W. M. Slagle, of this laboratory, was sent to the Oleum refinery of the Union Oil Company, where, under his supervision, 150 barrels of kerosene extract and 150 barrels of lubricating oil distillate were extracted with 16 per cent sulfuric acid. The crude bases, which were obtained at a cost of approximately $lOOO, probably represent the largest quantity of nitrogen compounds from petroleum over assembled at one time for investigation. It was decided, to investigate first the 50 liters of bases obtained from the kerosene extract. An attempt was made to separate individual bases by fractional distillation under reduced pressure. A 5 liter still of the Washburn type was constructed by Dr. 3. -J. Poth for the fraction ation of the bases. All the material was carried through six distillations, and, in the temperature ranges where the largest fractions predominated, eight fractionations were made. The boiling points, at atmospheric pressure, of the various fractions ranged from 185° to 335° indicating that hundreds of bases may be present. At the end of the eighth distillation it was apparent that none of the fractions consisted of a single substance. The failure to separate individual compounds was not due to the formation of constant boiling mixtures of substances differing widely in boiling points but resulted from the fact that in any restricted temperature zone there occurred a number of compounds boiling very closely together. Por instance, three bases were isolated from the fraction boiling at 276° at atmospheric pressure and were found to have boiling points of 280.6°, 278.9°, and 278.2° C. at a pressure of 746 mm. The residual bases, after the removal of the compounds mentioned above, distilled at a constant temperature. The failure of fractional distillation to separate individual bases made it necessary to develop other methods for their isolation* In the course of an experiment on the fraction boiling at 276° a white crystalline precipitate was obtained on dissolving the bases in dilute sulfuric acid. The white solid melted at 275° on recrystallization from alcohol and was later found to be the acid sulfate of 2,3,8 trimethyl quinoline. Since the sulfates of the other bases of this fraction are highly soluble in water or dilute sulfuric acid, 2,3,8 trimethyl quinoline may be precipitated almost quantitatively in the form of nearly pure sulfate on dissolving the 276° fraction in 50 per cent sulfuric acid. 1000 g. of this fraction yielded an amount of sulfate equivalent to nearly 20 per cent of the original fraction. In this way a 20 per cent yield of 2,3,8 trimethyl quinoline was separated from the 276° fraction and its absence outside the 273-279° range established. It was also discovered, that 2,3,8 trimethyl quinoline and a second base may be fractionally precipiated as picrates from a 22 per cent solution of the 276° fraction in 50 per cent acetic acid. A hot concentrated solution of picric acid in 50 per cent acetic acid was added in successive portions to the hot acetic acid solution of the bases, between additions of picric acid, the solution was allowed to cool and the precipitate filtered off. The first fractions consisted of the picrate of 2,3,8 trimethyl quinoline and the last fractions of a picrate of a base, which was found subsequently to have the empirical formula, The picrate of the first base molts at 242° and that of the second at 150.9°. After the second base had been removed no further precipitation in crystalline form was obtained with picric acid in acetic acid solution. The residual bases were recovered from the acetic acid solution and it was found that a third base could be precipitated as a picrate, melting at 196.5°, by adding a hot alcoholic solution of picric acid to a hot alcoholic solution of the bases. It later developed that the third base may be best isolated as sulfate. The residual bases were converted to sulfates by treatment with concentrated sulfuric acid in the cold. When the viscous mass of sulfates was stirred with acetone most of the material dissolved leaving a white crystalline salt. Analysis of the free base gave results corresponding to the formula CllWan indication that more than one substance is involved here. The fractions boiling at 277° and 275° were also treated in the manner described above and yielded additional amounts of the three bases. Since only about 40 per cent by weight of the 276° fraction is accounted for by the products so far isolated, a method for the separation of other bases that are undoubtedly present is being developed by Dr. E. J. Poth and Miss Ivy Parker. The method consists, essentially, of fractional acid extractions with buffered solutions and fractional precipitation of the bases from the acid extract. % The base obtained from the picrate melting at 242° was 24 investigated by Dr. J. R. Bailey and Mr. W. A. King who proved it to be 2,3,8 trimethyl quinoline, and in addition effected its synthesis. This is probably the first base isolated from petroleum distillates whose structure has been definitely determined and undoubtedly represents the most complex substance of any type ever isolated in pure form from petroleum. The acetone insoluble sulfate melting at 258.6° gave, on decomposition with sodium hydroxide, a white odorless solid melting at 42.5°. It adds methyl iodide at water bath temperature and yields a condensation product with 3 mols of formaldehyde. Further investigation of this product is being carried out by Miss Ivy Parker of this laboratory. The determination of the structure of this compound is hampered by the fact that it probably contains some admixture and further by the fact that only a small quantity is available with no expectation of more occurring in fractions not yet investigated. Both its hydrogen content and high refractive index is conclusive evidence that we are dealing here with a base of the aromatic type. The base CngHggN is undoubtedly a tertiary, hydroaromatic amine and it thus appears that the bases obtained from California kerosene distillates are of two types, aromatic and hydroaromatic. A few hydroaromatic bases may be present in coal tar in very small amounts, but the only 25 one observed so far is dihydroacridine . On the other hand the present investigation in the Texas Laboratory has proved the existence of hydroaromatic types in all fractions of the California kerosene bases. While compounds of the pyrrole type are not bases, the Pr-hydroindoles and hydrocarbazoles are bases and may be present in petroleum distillates. Analyses of a number of the fractions obtained during the fractional distillation of the crude bases give hydrogen and carbon ratios corresponding to hydroaromatic bases containing pyrrole or pyridine nuclei. These fractions are riot pure compounds but analyses as well as refractive indices indicate that in all fractions the hydroaromatic bases predominate. An outstanding contrast between coal tar and petroleum nitrogen compounds is presented more especially by (1) the occurrence in large amounts in petroleum distillates of inert nitrogen compounds probably of very complex types as yet entirely iinknown, and (2) the predominance of very interesting, hydroaromatic nitrogen bases, which would undoubtedly be classed as alkaloids, were they produced with plant life. Furthermore, the purely aromatic bases, such as the 2,3,8 trimethyl quinoline isolated from the crude Oleum kerosene distillate by King and Bailey, are of more complicated structure than the bases formed in coal tar distillation. It may be expected also, as the higher petroleum distillates, such as heavy lubricating stocks, are investigated, that even more complex nitrogen bases will be found than those occurring in the lighter distillates. As regards the nitrogen, in compounds inert to mineral acids, it should be observed that representatives of this series also occur in coal tar: as a matter of fact all the carbazole, a non-basic substance of the pyrrole type, used in the dye industry is derived from this source. The conditions are much more favorable than those in coal for the formation of varied and numerous, heterocyclic, nitrogen complexes through pyrolysis on distillation: coal consists in the main of amorphous, non-reactive carbon, whereas in petroleum there are present a host of hydrocarbons many of them unsaturated and, therefore, reactive. 24 Therefore the claim made by King and Bailey that, "petroleum offers the greatest vzealth of nitrogen compounds of any natural source", seems substantiated. There is a possibility jrhat a third typo of bases, containing nitrogen and sulfur in the molecule, exists in kerosone distillates. The kerosene distillate refined by the Bdeleanu process at the Oleum refinery of the Union Oil Company contains 0.30 to 0.35 per cent sulfur and. the kerosene extract contains 1.5 per cent of sulfur. An analysis of several fractions of the bases obtained from the kerosene extract shows that sulfur is present to the extent of 0.1 per cent. If bases of the thiazole type are present they occur in very small amounts. The determination of the structure of the base, CifiW' is comparable to the determination of the structure of an alkaloid such as strychnine or morphine; on the investigation of these substances many chemists have worked over long periods of time without having realized their structural proof, much less their synthesis. However, these substances are comparativoly cheap, are available in quantity, and any chemist in this field may obtain, at negligible cost, all the material needed, furthermore, he has at his command the experience gained by others who have made valuable contributions to the chemistry of these products. In regard to the quantity of material at the disposal of the author for investigation of the base, the situation is quite different. 400 g. of this base was originally available for research and this represents the yield of this particular substance from 150 barrels of kerosene extract obtained in the refining of crude kerosene distillate, This is all of the material available, and more of the product can be obtained only at great cost and after considerable delay. The properties of the C H N base which have been 16 25 determined in the present research can doubtless be used to advantage in perfecting a greatly simplified procedure for its isolation from the original Edeleanu kerosene residue. The mixture of bases from the kerosene residue, after extraction with dilute acid and liberation with caustic soda, could be roughly fractionated and the distillate having the boiling point range of 265 to 285° subjected to fractional picrate formation. The bases could be liberated from the various picrate fractions with ammonium hydroxide and again subjected to fractional distillation. In the final isolation of the C H N fraction use could 16 25 be made of the benzene solubility of its picrate, the ready solubility of its sulfate in water, its peculiar behavior toward nitric acid, the formation of an addition product with bromine from which the base is readily regained by treatment with sodium hydroxide, and finally, its unusual stability toward oxidizing agents. The base under investigation is a tertiary amine for it does not yield an acetyl or benzoyl derivative and fails to react with toluene sulfonyl chloride. That it is an aliphatic amine is highly improbable since the composition of the base requires that the aliphatic amihe be highly unsaturated and, therefore, easily oxidized. An outstanding characteristic of the base, is its stability lb D toward oxidizing agents. Since it is necessary to rely largely on the action of oxidizing agents in order to gain an insight into the structure of heterocyclic nitrogen compounds, and since, the products of oxidation are obtained in small quantities, the greatest refinement must be used in the work. In most cases only a few tenths of a gram are available and it is necessary to use microanalytical methods and micro-molecular weight determinations in order to arrive at molecular formulas. The number of structural possibilities is enormous as the few examples given on the following pages illustrate. After determining the type of substance, the determination of the size and distribution of the alkyl groups presents no small problem. In view of the experimental difficulties, the limited amount of material at hand, and the short time at the disposal of the author, it is not possible to do more than characterize the base, and to gain some little insight into its structure. As will be shown later it does not belong to any of the more common series of hydroaromatic bases which only adds to the complexity of the problem. It was first supposed, that the base might represent a highly alkylated derivative of Py-tetrahydroquinoline but the results of the investigation have rendered this 26 vievz untenable. According to Tafel Py-tetrahydroquino- p 7 line reduces silver nitrate in the cold. Bochner and Miller found that Py-tetrahydroquinoline gives blue-red colorations with ferric chloride and potassium ferricyanide and state that N-methyl-tetrahydroquinolines give green colorations on heating with benzotrichloride and zinc chloride. The base, C-.XJ, fails to give any of these reactions. ’ lb 2b * o v However, more trustworthy evidence that the product under investigation is not an N-methyl-Py-tetrahydroquinoline has been obtained. Since the base is a tertiary amine it must have an alkyl group attached to nitrogen, if a ?y-tetrahydroquinoline. In that case it should yield an alkyl iodide and a secondary amine on heating with conp Q centrated hydriodic acid. Herzig and Meyer have devel- oped a procedure for the detection and estimation of Nalkyl groups by means of this reaction. This method has been applied to a great number of compounds containing N-alkyl groups - such as nicotin, cocaine, morphine, caffein, and trimethylphenyl ammonium iodide. An experi- ment carried out with our product in accordance with the procedure of Herzig and Meyer gave no evidence of the presence of an N-alkyl group. Hydrotropidine breaks up into methyl chloride and norhydrotropidine when heated in a 29 stream of hydrogen chloride . ’fhe hydrochloride of C n X O was heated to 250° in a current of hydrogen chloride with 30 out docomposition. Heating the base with bonzoic acid also failed to remove an alkyl group from nitrogen. It, therefore, is unlikely that the base possesses an N-alkyl group and so, cannot be a derivative of Py-totrahydroquinoline. furthermore, Py-tetrahydroquinolines are reduced by heating with hydriodic acid and red phosphorous to dekahydroquinolines. When the product is subjected to this treatment in accordance with the procedure used by 31 Bamberger it is regained unchanged. Lastly, tho experimental molecular refractivity was found to be 73.92 while the calculated value for an N-alkyl tetrahydroquinoline of the formula, is 75.74. Since the calculated molecular refractivities of a number of methylated Py-tetrahydroquinolines agree very closely with those obtained from the experimental results 32 reported by Yamaguchi , it is probable that this evidence is reliable. That the base is probably not a Bz-tetrahydroquino- line is shown by its behavior toward oxidizing and reduc's'6 ing agents, von Braun has demonstrated that Bz-tetrahydro- quinolines are easily oxidized to pyridine carboxylic acids by potassium permanganate in alkaline solution. The base, sl^2s^» is quite stable toward potassium permanganate under the same conditions, since, after prolonged treatment with permanganate, most of it is recovered unchanged with the production of a small amount of oxalic acid. A general reaction of quinolines is that reduction with sodium and alcohol effects hydrogenation in the pyridine nucleus only, e.g. von Braun states that Bz~tetrahydroquionlines are easily reduced to decahydroquinolines in this way. Attempts to reduce the base under investigation with sodium and alcohol or sodium and amyl alcohol were unsuccessful. Seduction with tin and hydrochloric acid also hydrogenates 55 in the pyridine nucleus. This method of reduction fails with our base. It is also probable that the kerosene base ie neither an N-alkyl tetrahydroisoquinoline not a Bz-tetrahydroisoquinoline. Py-tetrahydroisoquinolines are prepared by the 56 reduction of isoquinolines with sodium and alcohol and 3z-tetrahydroisoquinolines are prepared by the reduction 52 of isoquinolines with hydriodic acid and phosphorous. It is likely, therefore, that Py-tetrahydroisoquinolines would be reduced to decahy.droisoquinolines by hydriodic acid and phosphorous and that Bz-tetrahydroisoquinolines would be reduced to decahydroisoquinolines by sodium and alcohol as in the case of the corresponding tetrahydroquinolines. It seems evident, also, that the pro- duct, if a Bz-tetrahydroquinoline, should be oxidized to a pyridine carboxylic acid by potassium permanganate. It is also doubtful that it has an B-alkyl group which a Byte trahydrisoquinoline must possess if it is a tertiary amine. An N-alkyl indoline derivative (V) represents another possibility for the formula, However, the ease with which indolines are oxidized makes it improbable that the base is a substance of this type# Indolines are readily oxidized by permanganate in acid or alkaline solution and read-37 ily reduce silver or copper nitrate in acid solution. Kann 58 and Tafel oxidized dihydroindoles to indoles by heating the bases with silver sulfate, whereas the kerosene base is not oxidized by silver sulfate when the method of Kann and Tafel is employed, nor does it reduce silver or copper nitrate in acid solution. 39 von Braun has prepared a number of sym-N-alkyl-octa- hydrocarbazoles. A Tetramethyloctahydrocarbazole (VII) has the empirical formula, C_ _H nc N. However, compounds of this type are unstable and turn brown on exposure to air, finally forming dark brown viscous products. The base, does not discolor after long standing. The octahydrocarbazoles are readily oxidized by permanganate or ferric chloride and may be reduced with tin and hydrochloric acid, properties v/hich our base does not possess. 40 Perkin and Plant have also prepared N-alkyl-octa- hydrocarbazoles which are isomeric with the compounds des-41 cribed by von Braun. According to von Braun these com- pounds are not derivatives of 1,2,3,4,5,6,7,8 octahydrocarbazole but are derivatives of 1,2,3,6,7,8,10,13 octa- hydrocarbazolesfVlll). These bases also turn brown on standing and are reduced by tin and hydrochloric acid to decahydrocarbazole s. 42 Borsche oxidized tetrahydrocarbazoles to carbazoles by heating with lead dioxide. The base, is not oxidized on heating with lead dioxide. Other structural possibilities are found in decahydroacridines (IX,X) and 2,3, trimcthylene-hexahydroquinolines (XI,XII). It is unlikely that an interpretation corresponding to any of these formulas can be accepted since an K* alkyl group is required in each case. In addition, compounds having the structure represented by IX and XI should be reduced with sodium and alcohol to perhydroderivatives 43 since von Braun reduces 1,2,3,4,6,7,8,9 octahydroacri- dino and the corresponding 2,3, triemthylene-quinoline to the perhydroderivatives with sodium and alcohol. It has also been shown that partially hydrogenated pyridine nuclei may be completely hydrogenated by sodium and alcohol or 44 by tin and hydrochloric acid. The base, is readily attacked on heating with 45 per cent nitric acid in a sealed tube at 170°. Prom the complex mixture of reaction products four substances have been isolated in very small quantities. Doubtless the identification of the products obtained in the nitric acid oxidation will contribute greatly to the final elucidation of the structure of the kerosene base. One of the four products is a solid, colorless base having the molecular formula, C16H23N304. This substance is probably a dinitro-derivative of the original base. It may be argued that since this product is colorless it is not a nitro-compound. However, a considerable number of colorless nitro-compounds are known. Furthermore, hydroaromatic substances are aliphatic in character and such compounds are nitrated by heating at high temperatures with dilute nitric acid, precisely the conditions which obtain in the oxidation of the kerosene base. Another compound which has been obtained from the nitric acid oxidation is a yellow substance having the .molecular formula, c i6s2js4%* formation of this compound may be explained by assuming that one methyl group has been oxidized to carboxyl and three nitro-groups have been substituted for hydrogen. That this compound is an acid is demonstrated by its solubility in dilute alkali. It is evident that the investigation of this substance and the nitrocompound, will throw little light on the structure of the base, therefore the investigation is being concentrated on the two acids which will be next discussed. A very interesting product, obtained in the nitric acid oxidation, is a dicarboxylic acid having the molecular formula, C_ n H_ This substance does not give colorations with ferric chloride or ferrous sulfate and therefore cannot be a phenolic acid or a pyridine carboxylic acid with a carboxyl group in position 2. In addition there is no way to account for the required number of hydrogen atoms in the molecule if the substance is assumed to be a pyridine dicarboxylic acid. The acid does not give a Liebermann test for nitrosamines and, therefore, the second nitrogen atom is probably present in an oxime grouping. This assumption also accounts for the fifth oxygen atom. At the present stage of the investigation it is not possible to suggest a definite formula for this product. The fourth product isolated from the nitric acid oxi- dation is a tricarboxylic acid having the molecular formula, CqHcNO,.. It forms an insoluble barium salt and on boiling o D 6 with animal charcoal yields a highly insoluble calcium salt, having a characteristic crystalline form. The acid corresponds in its molecular formula, its melting point, and its slight solubility in cold water, with berberonic acid. The acid also gives a deep red coloration with ferrous sulfate, another property of 2,4,5 pyridine tricarboxylic acid. Its crystalline form and solubility in alcohol, however, are at variance with the description of these properties as given in the literature. Whereas the acid isolated from the nitric acid oxidation of the kerosene base crystallizes from water in fine, hairlike needles and is quite aoluble in cold alcohol, the literature states that berberonic acid crystallizes in small prisms and is difficultly soluble in cold alcohol. Since the literature is conflicting in regard to some of the properties of berberonic acid, this substance is being synthesized for direct comparison with the nitric acid oxidation product of the base, In this connection it is worthy of notice that berberonic acid may be obtained by the nitric 46 acid oxidation of the alkaloid berberine. At present it is not possible to suggest a definite formula for the base, C IZ .H OU N, but it is likely that further investigation of the nitric acid oxidation products will make possible the proposal of a formula which will explain the properties of the base. At the time of this writing it can be stated, however, that the product is not a simple alkylated derivative of the more common types of hydroaromatic bases such as tetrahydroquinolin^, tetrahydreisoquinoline, dihydroindole, octahydrocarbazole, or decahydroacridine. 22. Poth, E. J. , W.D.Armstrong, C.C.Cogburn, and J.R.Bailey: ’’The Estimation of Nitrogen in Petroleum and Bitumens." Jr. Ind. Eng. Chem. 20, 88 (1928). 23. Floyd, W.W.: Dissertation, The University of Texas,l92B. 24. Beilby, George: "Preparation of Ammonia from Nitrogenous Minerals." Dingl. polyt. J. 254, 342 (1884). 24. Poth, E. J. , W.A.Schulze, W.A.King, W.C.Thompson, W.M. Slagle, W.W.Floyd, and J.R.Bailey: ”An Investigation of the Bases in the Kerosene Distillate of California Petroleum.” J. A. G. S. 52, 1239 (1930). 25. Decker, H. and Georges Dunant: "Notiz über das Verkommon des Hydroacridins in Steinkohlenteer." Ber. 42, 1178 (1909). 26. Tafel, Julius: ”Ueber die Oxydation hydrierter Pyridin und Chinolinbasen.” Her. 25, 1619 (1892). 27. Doebner, o.' and W. v. Miller: ”Über Chinaldenbasen. ” Ser. 16, 2464 (1883). 28. Her zig, J. and H. Meyer: "Ueber den Nachweiss und die Bestimmung des an Stickstoff gobundenen Alkyls.” Monats. 15, 613 (1894). 29. Cohen, J. B. , '’Organic Chemistry for Advanced Students.” Longmans, Green and Company, Nev; York, 1928, Part 111, p. 353. 30. von Braun, J. and Karl Weissbach: "Entalkylierung tertiarer Amine durch organised© Sauren." Ber. 63, 489 (1930). 31. Bamberger, E. and ?. Lengfeld: "Neue deductions produete des Chinolins." Ber. 23, 1142 (1890). 32. Yamaguchi, S.: "Methylhydroquinolines." 1,11.C.A. 21 2696 (1927), J.Pharm. Soc. Japan, no. 533, 556 (1926), no. 535, 749 (1926). 33. von Braun, J., Adolf Petzold and Jon Seemann; "Kataiytische Hydrierung unter Pruck bei Gegenwart von Nickelsalzen, IV. Chinolin-Basen." Bcr. 55, 3782 (1922). 34. Kaku,T.: "Atophan Derivatives II: Seduction of 2- Anisyl- and Phenylquinolines.” C.A. 23, 392 (1929). J. Pharm.Soc. Japan, 48, 693 (1928). 35. Bamberger, E. and P. Wulz: ”Ueber Homologe des Tetrahydrochinolins." Ber. 24, 2055 (1891). 56. Bamberger, E. and W. Dieckmann: "Ueber Tetrahydrur des Isochinolins." Ber. 26, 1242 (1895). 37. Brunner, Karl: "Ueber Indolinone." Monats. 18, 118 (1897). 38. Kann, Moses and Julius Tafel: "Ueber die Oxydation hydrirter Indol." Ber. 27, 826 (1894). 39. von Braun, Julius and H. Ritter: "Katalytische Hydrierung unter Druck bei Gegenwart von Nickelsalzen, V: Der Garbazol-Komplex." Ber. 55, 3792 (1922). 40. Perkin, Jun., W.H. and S.G.P.Plant: "1:2:3:4:5:6:718- octahydrocarbazole and its Derivatives." J. Chem. Soc. 125, 1503 (1924). 41. von Braun, J., and Ludwig Schoring: "Über hydrierter carbazols. " Ber. 58, 2156 (1925) • 42. Borsche, Von W.: "Heber Tetra- und Hexahydriocarbazolverbindungen und eine neue Carbazolsynthese.” Ann. 359, 49 (1908). 43. von Braun, Julius, Adolf Petzold, and Adam Schultheiss: "Ueber Bz-tetrahydrochinolins und ihre Berivate, III.; Tricyclishe Vorbindungen." Ber. 56, 1348 (1923). 44. Preund, Martin and Georg Bode; "Über Einwirkung von Grignard-Losungen auf Halogen Ammoniumverbindungen." Ber. 42, 1746 (1909). Treund, Martin and E.Kessler; "Einwirkung von Organomagnesiumverbindungen auf Chinolinjodmethylat, ein Beitrage zur Stereochemie stickstoffhaltiger Verbindungen." J. Prakt. Chemie (2) 98, 233 (1919). 45. Cohen, J. 3., "Organic Chemistry for Advanced Students", Longmans, Green and Company, New York, 1928, Part 111, p. 390. I. 11. 17. 111. V. VI. VIII. VII. IX. X. XI < XII. EXPERIMENTAL PART Preliminary Investigation of Possible Sources of Nitrogen Bases The abnormally high nitrogen content of California crude petroleum suggested that California crude petroleum or products of its refining would he the most profitable sources of nitrogen bases. Consequently, a preliminary study was made of California crude oils, a lubricating oil distillate, the residue from the refining of kerosene distillates by the Edeleanu process, and the sludge and "black acid" from the refining of lubricating oil distillates. The greater part of the material was obtained from the Union Oil Company of California through the courtesy of Mr. R. E. Haylett, Technical Assistant. That only a very small part of the nitrogen compounds occurring in California petroleum are nitrogen bases is indicated by the following experiments in which only minute amounts of nitrogen bases were obtained from the crude oils by dilute acid extraction. Extraction of Ventura Prude Petroleum with Hydrochloric Acid. One kilogram of Ventura crude petroleum and lOOcc. of a 10 per cent solution of hydrochloric acid were placed in a large separatory funnel and agitated intermittently for four hours. After standing over night the mixture was heated in a water bath for some time to break the emulsion. The aqueous layer was then drawn off and extracted with ether to remove traces of hydrocarbons. After neutralization of the acid the aqueous layer was extracted with ether which left l-2cc. of basic oil on evaporation. The dark brown oil was soluble in dilute acid and gave a test for nitrogen. The original oil contained 0.38 per cent nitrogen and the extracted oil 0.42 per cent nitrogen. Had all the nitrogen existed in basic form approximately 40g. of bases should have been obtained, assuming their average nitrogen content to be as high as 10 per cent. Extraction of Coalinga Petroleum with Dilute Sulfuric Acid* One kilogram of Coalinga crude petroleum was also extracted as described above, except that 500 cc. of 20 per cent sulfuric acid was used in the place of the hydrochloric acid. Only l-2cc. of basic oil was obtained. The oil was soluble in dilute acid and gave a test for nitrogen. The crude oil contained 0.44 per cent nitrogen before extraction and 0.45 per cent after extraction. Extraction of Ventura Crude Petroleum with Acetic Acid. One hundred and fifty grams of Ventura crude oil were agitated several hours with 50cc. of 20 per cent acetic acid and allowed to stand two days. The mixture was then heated in a water bath until the two layers separated, whereupon the acid layer was neutralized in the usual way and yielded a very small amount of dark brown oil which gave tests for nitrogen and sulfur. The crude oil contained 0.38 per cent nitrogen before extraction and 0.39 per cent after extraction. Action of Concentrated Sulfuric Acid, on Crude Petroleum. Treatment of crude oil with concentrated sulfuric acid removes nearly all the nitrogen from the crude oil as is shown by the following experiment: Pifty cc. of concentrated sulfuric acid was added slowly to lOOcc. of Ventura crude oil under cooling. The mixture was shaken vigorously and allowed to stand for several days. The odor of sulfur dioxide was very pronounced during the addition of the acid. A black tarry sludge, 43cc. in volume, was deposited. The oil layer, which was nearly decolorized, had decreased in volume to 47cc. On analysis it was fuund to contain a negligible amount of nitrogen, the sludge 0.46 per cent, and the acid 0.06 per cent. Investigation of Hxtra Heavy Lubricating Oil Distillate. Because of the High nitrogen content, extra heavy, crude, lubricating oil distillate was investigated and the amount of bases extractable with dilute acid was quantitatively determined. The effect of distillation on the quantities of bases which may be extracted and the distribution of nitrogen in the various fractions from distillation were also determined. The extra heavy lubricating oil stock was a dark viscous oil* having a specific gravity, , of 0.976 and 15.5 a viscosity of 67 Saybolt at 210°]?. Analysis of the oil by the Kjeldahl method gave 0.699 per cent for the nitrogen content. The distillate did not give a pine splinter tost for pyrrole derivatives. Jfour liters of the lubricating oil stock were agitated with an equal volume of 16 per cent sulfuric acid and the mixture allowed to stand for several days. The acid solution was then separated from the oil layer and extracted with ether. After neutralization of the acid the free bases were taken up in ether leaving a residue of 5.712 g. of dark brown oil on evaporation. The yield was 1.42 g. of bases per liter of distillate. After drying, the extracted distillate was found to haze a nitrogen content of 0.659 per cent. That distillation of the lubricating oil distillate results in an increase in the amount of acid soluble nitro- gen compounds is shown by the following experiment: One liter of the distillate was distilled from a pyrex flask until oil ceased to pass over and the distillation flask began to melt. During the latter part of the distillation the odor of ammonia was pronounced and a red litmus paper turned blue when held in the fumes. 300 cc. fractions were collected. After analysis for nitrogen each fraction was extracted with an equal volume of 16 per cent sulfuric acid in the usual manner. The bases obtained were weighed and the extracted oil analyzed for nitrogen. A IOOOcc. sample of the lubricating oil distillate from which 1.54 g. of bases had been extracted was treated in a similar manner. A second sample of unextracted distillate was distilled at a pressure of 7-10 mm. The odor of ammonia was not noticeable during the distillation under reduced pressure. The data obtained in these experiments follows in tabular form: The data shows that on distillation the nitrogen is concentrated in the higher boiling fractions reaching a maximum in the residue* This effect is less pronounced in the distillation under reduced pressure. The fractions from the distillation of the untreated stock at atmospheric pressure yielded B.Bog. of bases per liter, or nearly six times that obtained by acid extraction alone. In the case of the sample which had been extracted previous to distillation 6.60 g. of additional bases were obtained. The distillation produced four times the amount of bases originally present in the oil. 1.99 g. of bases was obtained by distillation under reduced pressure, an increase of only 0.45 g. over that obtained by simple acid extraction. Obviously the yield of bases from extra heavy lubricating oil stock may be increased five or six times by a preliminary distillation at atmospheric pressure prior to the extraction with dilute acid. Investigation of Kerosene Extract. The kerosene extract, as received at the laboratory, had a light red-brown color and smelled strongly of sulfur dioxide. Its specific gravity, was 0.8941. The product dissolved completely ~° in ether, alcohol, and glacial acetic acid, but was very slightly soluble in water, dilute hydrochloric acid, dilute sodium hydroxide solution, or concentrated sulfuric acid. Analyses by the Kjeldahl method showed the nitrogen content to be 0.055 per cent. The sulfur content of the kerosene extract was found to be 1.58 per cent by a modified Carius 46 method. Before the sulfur analysis the extract was freed of as much sulfur dioxide as possible by exhausting a flask containing the product for six hours with a v/ater pump. The kerosene extract gives a pine splinter test for pyrrole derivatives. In order to determine the amount of bases present in the kerosene extract four liters of the product were agitated for several hours with an equal volume of 16 per cent sulfuric acid, and the mixture allowed to stand over night. The acid layer was separated from the oil and extracted with ether. The acid was then neutralized with sodium hydroxide, and the free bases extracted from the alkaline solution with ether. The ether solution left a residue of 11.0 g. of dark brown oil on evaporation. The basic oil has a nitrogen content of 6.78 per cent and gives a test for sulfur. The residual oil, after drying over anhydrous sodium sulfate, contains 0.025 per cent nitrogen, a decrease of 0.030 per cent from the nitrogen content of the original oil. A lOOcc. sample of the kerosene extract was twice extracted with an equal volume of 16 per cent sulfuric acid. An analysis of the extracted oil gave a result of 0.023 per cent nitrogen which indicates that one extraction with dilute acid is sufficient to remove the bases from the kerosene extract. Smaller amounts of 16 per cent sulfuric acid were also employed with the results as shown in the following table: Evidently one fourth of the amount of acid used in previous experiments was sufficient to remove the bases from the kerosene extract. 47 Hatcher and Skirrow state, ”The crude phenols must be extracted from coal tar oils by means of sodium hydroxide before the pyridine bases may be economically ex- 48 tracted with sulfuric acid." Oshima and Ishibashi find that preliminary treatment of low temperature coal tar increases the amounts of nitrogen bases which may be removed with dilute acids. To determine the effect of a preliminary treatment with caustic soda on the yield of bases which may be obtained with dilute acid the following procedure was carried out: Seventy co. of kerosene extract was agitated for some time with an equal volume of 16 per cent sodium hydroxide solution. The oil layer was separated, washed with water, dried over anhydrous sodium sulfate, and analyzed for nitro gen. 50cc. of the oil which had been treated with sodium hydroxide was extracted in the usual way with 16 per cent sulfuric acid, and the extracted oil analyzed again for nitrogen. The results of this analysis are tabulated below and show that treatment with dilute alkali, as would be expected. removes no nitrogen from the kerosene extract and does not increase the quantity of bases which may be extracted with dilute acid. It also seemed desirable to determine the effect of distillation at atmospheric pressure on the distribution of the nitrogen in successive fractions and on the quantity of bases which may be extracted with dilute acid. lOOcc of kerosene extract was distilled and lOcc. fractions collected. Nearly all of the extract distilled over leaving a very small residue. The separate fractions were analyzed for nitrogen. The data is set forth in the following table: 75c0. of another lOOcc. sample of kerosene extract,;was dis- tilled and the distillate and residue analyzed for nitrogen. The distillate and residue were then extracted with 16 per cent sulfuric acid and analyzed again for nitrogen. The results are given in Table IX; data for the extraction of the original kerosene extract is also given. To determine the effect of distillation on the amount of nitrogen bases which may be extracted with dilute acid, another sample of kerosene extract was twice extracted with dilute acid, distilled, and the distilled oil extracted with acid. The above facts indicate that (1) distillation of the kerosene extract concentrates the nitrogen in the higher boiling fractions, and (2) increases the ratio of basic to non-basie nitrogen. An explanation of the increase of the yield of bases due to distillation may be that non-basic substances of the pyrrole type undergo molecular rearrangement to isomeric basic pyridine types. Such a speculation finds encouragement in the many molecular rearrangements that have been studied by the organic chemist. 46. Devine, John M. and F.W.Lane: "The Use of the Carius Method for the Determination of Sulfur in the Less Volatile Petroleum Oils." J.A.C.S. 50, 1707 (1928). 47. Hatcher, W.H. and P. 7. Skirrow: ’’The Compounds of Phenol and the Cresols with Pyridine.” J.A.C.S. 39, 1939 (1917). 48. Oshima, Y. and K. Ishibashi: ’’Basic Compounds in Low- Temperature Tar.” Jr. Soo. Chem. Ind. Japan, 29, 445 (1926). Fraction Volume % nitrogen nitrogen after extraction Weight of base s 1. 300cc. 0.241 0.171 2.24 2. 300cc. 0.373 0.304 3.45 3. 165cc. 0.858 0.581 3.11 Hesidue 163g. 2.015 Table 111 Distillation of extra heavy lubricating oil stock at atmospheric pressure; IOOOcc. sample, Jo nitrogen equals 0.699. fraction Volume > nitrogen % nitrogen after extraction Weight of bases 1. 300cc. 0.278 0.210 2.10 2. 300cc. 0.482 0.296 2.90 3. 150co. 0.896 0.643 1.60 Hesidue 125g. 1.868 Table IV Distillation of extra heavy lubricating oil stock previously extracted with 16> sulfuric acid; IOOOcc. sample, % nitrogen equals 0.659. Table V Distillation of heavy lubricating oil stock under 7-10 mm. pressure; IOOOcc. sample, % nitrogen equals 0.699. Fraction Volume nitrogen nitrogen after extraction Weight of bases 1. -300c c. 0.548 0.449 1.25 2 300cc. 0.621 0.522 0.60 3. 300cc. 0.774 0.678 0.14 Hesidue 50 g. 1.195 Ratio of volume of acid % nitrogen in to volume of kerosene extract extracted oil 1/2 0.028 - 0.025 1/4 0.020 - 0.026 1/10 0.030 - 0.027 Table VI Extraction of kerosene extract with various amounts of 16 per cent sulfuric acid % nitrogen in original oil 0.055 % nitrogen in oil after alkali treatment 0.055 % nitrogen in oil after acid extraction 0.023 Table VII Sffect of alkali treatment on the amount of bases which may be extracted from kerosene extract on Temperature range nitrogen 1. 180-200° 0.000 2. 200-210° 0.013 3. 210-215° 0.000 4. 215-220° 0.023 5. 220-225° 0.025 6. 225-233° 0.039 7. 233-241° 0.063 8. 241-247° 0.071 9. 247-260° 0.062 10. 260-290° 0.175 Table VIII Effect of distillation on the distribution of nitrogen in the various fractions Material % nitrogen before extraction /o nitrogen after extraction Ratio of basic to non-basic nitrogen Kerosene extract 0.055 0.025 1.2 : 1 Distillate 0.041 0.014 1.92 : 1 Residue 0.131 0.027 3.9 : 1 Table IX Effect of distillation of kerosene extract on ratio of basic to non-basic nitrogen nitrogen in original kerosene extract 0.055 nitrogen after double extraction 0.026 % ni trogen after distillation 0.025 % nitrogen in distilled oil after extraction 0.014 Table X Effect of distillation on the amount of nitrogen which may be extracted from kerosene extract Separation of Bases From California Kerosene Distillates Through the courtesy of the Union Oil Company of California 150 barrels of kerosene extract and 150 barrels of a light lubricating oil distillate were extracted with 16 per cent sulfuric acid at the Oleum refinery under the direction of W. M. Slagle 49 of this laboratory* The investi- gation of the bases obtained from the kerosene extract was undertaken first. The three barrels of material received at the laboratory consisted of a solution of the bases in kerosene extract and a mucky brown precipitate which held a considerable portion of the bases and water. The separation of the bases from the mixture proved to be a tedious operation and consisted of the following steps: 1. Filtration of oil from suspended material 2. Extraction of filter cake from (1) with kerosene 3. Separation of bases from the oil and kerosene by extraction with dilute sulfuric acid and subsequent liberation of bases from the acid with caustic soda. The 50 liters of bases were fractionally distilled from a five liter Claissen flask with an eight inch fractionating column under a pressure of 30mm. By the use of a Bogert type receiver and of a second vacuum pump to evacuate the receiver after it had been emptied it was possible to make the distillation continuous. This procedure proved unsatisfactory and for exhaustive fractionation Dr. E. J. 24.50 Poth of this laboratory constructed a still of five liters capacity which was patterned after the apparatus design ed by E. W. Washburn and J. H. Brunner for use at the Bureau of Standards in connection with Project Do. 6 of the American Petroleum Institute. The complete distillation of all the material was repeated five times, and, in the temperature zones where the largest fractions were obtained, eight fractionations were made. Sach of the two largest fractions, boiling at 276° and 277° at atmospheric pressure, contained approximately 1400 g. of bases. ?rom these fractions three bases have been separated as picrates or sulfates. In order that the method of separation may be made clear the procedure followed with 400 g. of the fraction boiling at 276° will be explained in some detail. Sighty grams of 80 per cent picric acid, dissolved in 200 cc. of boiling 50 per cent acetic acid, were added to a hot solution of 400 g. of bases in 1400 cc. of 50 per cent acetic acid. On cooling the yellow, crystalline picrate of 2,3,8 trimethyl quinoline separated and was filtered from the solution. Three precipitations, carried out as described above, completely removed 2,3,8 trimethyl quinoline, as picrate, from the mixture of bases. This product was admixed with a small amount of a second picrate, melting at 150.9°, which was removed by extraction with not benzene. About 20 per cent of the 276° fraction was precipitated as the picrate of 2,3,8 trimethyl quinoline. The base, Ci6s2ss» was completely precipitated as the picrate melting at 150.9° by two additional precipitations. In each fractionation the picrate came down first as a viscous product. The supernatant liquid was transferred to another flask and yielded a crop of crystals after standing several days in the ice box. The viscous picrate was dissolved in hot alcohol from which it separated in a crystalline condition on cooling. The benzene extract of the picrate of 2,3,8 trimethyl quinoline yielded 35g. of the picrate of the base, C HN. A total of 167 g. of this 16 25 picrate was obtained. The acetic acid residue from the picric acid precipitation was concentrated by distilling off the acetic acid in a vacuum and the bases set free with ammonium hydroxide. The 190 g. of residual bases was treated in the cold with 50cc. of concentrated sulfuric acid, the viscous mixture of sulfates stirred with 150 cc. of acetone, and the acetone insoluble sulfate filtered off. The crude sulfate was washed with acetone and recrystallized from acetic acid. 12g. of purified sulfate, melting at 258.6°, was obtained. This product later proved to be the sulfate of a base having the empirical formula, and therefore could sot be a homogenous substance. From 1400 g. of the bases comprising the fraction boiling at 276° approximately 250 g. of 2,3,8 trimethyl quinoline was obtained as picrate. From the same fraction a total of 567 g. of the crude picrate of the C_ compound and 45g. of the sulfate of the base, CHS, were obtained. XX X 1400 g. of the fraction boiling at 277° yielded 130 g. of the picrate of 2,3,8 trimethyl quinoline, 610 g. of the picrate of the base, C,and 65g. of the sulfate of the base, From 470 g. of the fraction boiling at 274- 275° were obtained 70g. of the picrate of 2,3,8 trimethyl quinoline, 40g. of the picrate of the base, and 24g. of the sulfate of the base, C_-H_ B. From these fractions 11 X & there were obtained a total of 300 g. of 2,3,8 trimethyl quinoline, 1200 g. of the crude picrate of the base, and 134 g. of the sulfate of the base, 49. Slagle, W.M. : Dissertation, The university of Texas, 1928. 50. Poth, E.J.: "A Receiver for Vacuum Distillation.” Analytical Edition, Jr. Ind. Eng. Cheip. I, 111 (1929). Investigation of the Bases, C11H 12N and C16H25N The Base, The acetone insoluble sulfate melting at 258.6° yields an odorless, colorless, solid base on decomposition with sodium hydroxide. The base was prepared for analysis by recrystallization from 70 per cent alcohol and fractional distillation in a vacuum. It molts at 42.5°, boils at 278.9° under a pressure of 746 mm., and has a refractive index of 1.5828 (n 6^ • compound is insoluble in v/ater and is readily soluble in organic solvents. Analysis shows that the base has the empirical pormula, therefore, contains some admixture. Analysis. Calcd. for C n .H, o li: C, 83.54; H, 7.59; N, 8.86. Found: C, 83.43; H, 7.60; N, 8.76. On heating with formaldehyde for 7 hours at 100° the base forms a condensation product which yields a picrate melting at 160.8°. A crystalline product melting at 163° (unc.) is formed by heating the base with methyl iodide in a sealed tube at 100°. The investigation of this base is being continued by Miss Ivy Parker of this laboratory. The Base, C, H The picrate of the second base was l—ttr-25- purified by several recrystallizations from 95 per cent alcohol. On heating the picrate with ammonium hydroxide solution the base separated as an oil which was washed with water to remove traces of ammonium picrate and then dried at a temperature of 100° in a current of dry air. For further purification it was fractionated under a pressure of 11mm. and a heart cut redistilled for analysis. The base is a colorless, odorless oil having a density, (d~.°) , of 0*9391 and a refractive index, , of 1*5129. The compound boils at 278.2° under a pressure of 746 mm. and does not solidify when cooled in a freezing mixture. It is very slightly soluble in water but readily soluble in the common organic solvents. From analyses the molecular formula, was derived. 16 do Analysis. Calcd. for C I6 H 26 N: C, 83.12; H, 10.82; N, 6.06. Found: C, 83.25, 83.30; H, 10.78, 10.81; N, 5.95, 6.12. Molecular weight. Calcd. for C 1 : 231. Found: 235. Picrate, C-i , z OH. The picrate may be prepared by adding a hot alcoholic solution of picric acid to a hot alcoholic solution of the base. On cooling the bright yellow salt separates as small, pointed prisms which melt at 150.9°. The picrate is readily soluble in hot alcohol, glacial acetic acid, benzene, acetone, and ethyl acetate. It is slightly soluble in water, ether, and petroleum ether. Analysis. Cacld. for CHON. N, 12.17. Found; 22 2o 7 4 N. 12.63. Acid Sulfate, Ci6H2SN. This salt is best obtained by treating the base with concentrated sulfuric acid in the cold and can be purified by precipitation from glacial acetic acid with ether as a finely divided, white solid. The sulfate was obtained as square, colorless plates which melt at 195.7° on recrystallization from acetone. The salt is readily 'soluble in water, alcohol, and glacial acetic acid and insoluble in ether, ethyl acetate, and petroleum ether. Analysis. Calcd. for S, 9.72. Found; 5,9.69. Me thiodi de, C„ N The base can be quantitatively converted to the quaternary ammonium salt when heated with methyl iodide at 100° in a sealed tube for 24 hours. This salt may be purified by recrystallization from alcohol and crystallizes from hot water in the form of small, slender, colorless prisms. It begins to sinter on heating to 250° and finally decomposes without melting. The quaternary salt is somewhat soluble in cold water and cold alcohol and is readily soluble in hot water, hot alcohol, and glacial acetic acid. Analysis. Calcd. for C n „H oo NI: N, 3.75. Found; N,3.99. Sine Chloride Salt, Cq ZnClg. 3inc chloride solution precipitates a colorless oil on addition to a solution of the base in hydrochloric acid which solidifies on standing. It crystallizes from hot water in rosettes of short, pointed prisms which melt at 170.8°. The compound is slightly soluble in cold water and glacial acetic acid, but is readily soluble in alcohol, hot water, and hot acetic acid. Analysis. Calcd. for N, 3.81. Found; H, 4.13. Mercuric Chloride Salt. HgCl • On adding a hot solution of to a solution of the hydrochloride of the base an oil separates which on stirring solidifies. The salt crystallizes from alcohol in clusters of slender rods which melt at 157.5°. It is insoluble in water, cold alcohol, or glacial acetic acid and is quite soluble in hot alcohol or hot acetic acid. Analysis. Calcd. for G^H^N.HgCl^; N, 2.78. Found; N, 2.86. Hydrochloride, HCl* G one entrat edhydrochl or i c acid is added to the base, the solution evaporated to dryness, the residue taken up in warm acetone, and the hydro- chloride precipitated with ether. The salt crystallizes in short prisms, melting at 250.8°. It is readily soluble in water and alcohol and insoluble in ether. Analysis. Cacld. for C H iI.HCI: N, 5.23. Pound; 16 25 N, 5.14, 5.38. Nitrate, On dissolving the base in 45 per cent nitric acid the nitrate separates as a colorless oil which solidifies with cooling and vigorous stirring. It may be recrystallized from water in the form of small, pointed prisms melting at 78.8°. This product loses its mol of water of crystallization at 110°. The anhydrous salt, which melts at 141.2°, crystallizes from water in octahedra of bipyrimidal form. The nitrate is difficultly soluble in water, soluble in alcohol, glacial acetic acid, and acetone, and insoluble in ether. Analysis. Calcd. for CigHggKgO,!: N, 8.97. FoundjN, 8.82. Calcd. for C^gHg^N^O^(anhydrous form): N,9.52. Found: N,9.57. Chlorplatinate , ( c i 6 r H s ptG1 G ’ This salt is obtained as a light, orange precipitate on the addition of chlorplatinic acid to a solution of the base in dilute hydrochloric acid. It may be crystallized from dilute hydrochloric acid in the form of small, bluntly pointed rods melting at 240.3°. The salt is slightly soluble in water, cold alcohol, or cold dilute hydrochloric acid, and soluble in hot alcohol and hot dilute hydrochloric acid. Analysis. Caicd. for C x JL..,B o CI.Pt: Pt, 22. 35. Found: * 32 02 2 6 Pt,22.29. Proof of the Absence of an H-alkyl Croup in the c l6^aa H Base. On heating N-alkyl amines with concentrated hydriodic acid the alkyl group is removed as an alkyl iodide. Herzig and !£eyer were able to quantitatively estimate the alkyl groups in the case of a large number of compounds such as morphine, cocaine, nicotine, caffein, and trimethyl-phenylammonium iodide by employing a modification of the procedure and apparatus used by Sei sei in the determination of methoxyl groups. The bases are heated with concentrated hydriodic acid in a stream of carbon dioxide. The gas stream carries the alkyl iodide into an alcoholic solution of silver nitrate which precipitates the iodine as silver iodide. The authors state that the method may be applied to the detection of the higher alkyl groups and do not record a failure of this reaction in any of the experiments conducted. With the hope that the base, would yield an alkyl iodide under this treatment an experiment was carried out using the procedure and apparatus recommended by Herzig and Meyer. One gram of base, sg. of ammonium iodide,and 12cc. of hydriodic acid (5pg.1.70) were heated slowly to 300° in the course of several hours while a current of carbon dioxide was passed through the apparatus. The temperature was maintained for two hours longer at 300°. No trace of silver iodide appeared in the flask containing the silver nitrate solution. As a proof that the base remained unaltered the contents of the reaction flask were dissolved in hot water and, on the addition of sodium hydroxide solution, an oil separated having the same refractive index as the original product and not giving a nitroso compound with nitrous acid. A part of the base was converted to the picrate which melted at 146° (unc.). 0.95 g. of base was recovered. The experiment proves that the base, C_ .H.. N, does 1 o not possess an H-alkyl group. Since obtained nothydrotropidine and methyl chloride on heating hydrotropidine in a stream of hydrogen chloride, an attempt was made to remove an alkyl group from nitrogen by heating the hydrochloride of the base in a current of hydrogen chloride. One and. one half grains of were placed in a small tube connected to a small ice-cooled receiver, the system was evacuated by means of a water pump and dry hydrogen chloride gas was allowed to leak into the apparatus through a capillary tube dipping into the molten hydrochloride. The tube containing the salt was heated during this experiment to 190-200°. Ho decomposition was evident and a considerable portion of the hydrochloride sublimed into the colder part of the tube. The sublimate, in alcoholic solution, gave a picrate melting at 148 (unc.), the melting point of the picrate of the original base. A mix- ture of this picrate and that of the original base also melted at 148°. The unsublimed portion of the hydrochloride also gave a picrate melting at 148°. Identical results were obtained on heating the hydrochloride of the base in a stream of dry hydrogen chloride over a temperature range of 220-250° at atmospheric pressure. A third attempt was made to detect the presence of an N-alkyl group by heating the base with benzoic acid. Ac-50 cording to von Braun this treatment yields methyl alcohol and a benzoyl derivative of a secondary amine when applied to tertiary amines such as N-methyl tetrahydroisoquinoline or tropane. Two grams of the base, sX6s2ss* an s an e l u i va l en t ; amount of benzoic acid were heated together at 250-280° for twelve hours, kn ether solution of the discolored mixture was treated with sodium hydroxide to remove unchanged benzoic acid, and with hydrochloric acid to remove unchanged base. On evaporation of the ether 0.2 g. of a black gum was obtained. This was heated with concentrated hydrochloric acid for o twelve hours at 120 to saponify any benzoyl derivative that might have been formed. The greater part of the gum still remained after this treatment. The acid was neutralized and yielded only a trace of oily material. Since the dark gum did not yield a base with the hydrochloric acid treatment it was not a benzoyl derivative of a secondary amine and therefore the original compound probably does not contain an N-alkyl group. This product was not investigated further. Distillation with Zinc Dust.2,4-dimethy 1 tetrahydro- quinoline hydrochloride yields 2,4-dimethyl quinoline on 51 distillation with zinc dust. It was to be expected that the base, if a tetrahydroquinoline, would yield a quinoline by the loss of four hydrogen atoms on similar treatment. One half gram of C, was mixed with zinc dust and distilled. There was a copious evolution of white fumes and about o.scc. of yellow oil distilled from the reaction tube. A small part of the distillate formed a picrate which melted at 148° (unc.), the melting point of the picrate of the original base. The base, does not yield a con- densation product with formaldehyde while aromatic quinolines methylated in positions 2 or 4 of the pyridine nucleus yield such products. Accordingly the remainder of the distillate was sealed in a small tube with lee. of formalin and heated at 100° for 24 hours. The oily layer was still present after the heating and the formalin loft no residue on evaporation. Evidently the sole result of the zinc dust distillation was the dissociation of the hydrochloride into the base and hydrogen chloride, and it may be concluded that the compound is not a heptamethyl tetrahydroquino- line. Any other alkylated tetrahydroquinoline is also excluded. As is well knovm, many alkaloids yield decomposition products on distillation with zinc dust. l?or instance, morphine yields ammonia, trimethylamine, pyrrole, pyridine, and other products on distillation with zinc dust. With the object of effecting a decomposition of the kerosene base further experiments with zinc dust were instituted. One gram of the base, and sg. of zinc dust were placed in a test tube to which an air condenser was sealed. The mixture was heated to the boiling point of the base for eight hours. No reaction was apparent and 0.95 g. of base was recovered. The refractive index of the oil was not changed by the treatment and a small amount was converted to a picrate which melted at 149°, the melting point of the picrate of the base, C I6 H 2S H ‘ With the idea that higher temperatures might be necessary for the decomposition Ig. of the base and sg. of zinc dust were heated in a sealed tube for six hours at 350°. After cooling the tube was opened and the oil extracted with ether. Here again only a negative result was obtained. Action of Alcoholic Potash on the Base. One half gram of base, lee. of 50 per cent potassium hydroxide solution, and lee. of alcohol were heated in a sealed tube for three hours at 160°. Bo pressure was evident on opening the tube. Five grams of colorless oil were recovered and identified, in the usual way, as unchanged base. Reduction of the Base with Sodium and Alcohol. <X« von Braun states that Bz-tetrahydroquinolines are easily transformed to decahydroquinolines on reduction with sodium Ra and alcohol. The procedure used by Ladenburg in the reduction of pyridine to piperidine was applied as follows: One gram of the base in lOcc. of absolute alcohol was heated on the water bath under reflux and 2g. of sodium was added in the course of 20-30 minutes. From time to time a little more absolute alcohol was added as the reaction slowed down or sodium ethylate began to separate. After all the sodium had reacted the solution was diluted with water and steam distilled. The distillate yielded neither a nitroso compound nor a benzoyl derivative with benzoyl chloride. Here again a practically quantitative yield of unchanged base was recovered in the form of picrate. Reduction of the Base, C N, with Sodium and Amyl Alcohol. The above reduction experiment was next carried out in amyl alcohol solution. At the end of the reduction the alcoholic solution was acidified with hydrochloric acid and evaporated to dryness in a vacuum. The base liberated from the hydrochloride with caustic soda was rectified and analyzed. Calcd. for H, 10.82. Found: H, 10.97. Reduction to a product of the formula, is because here the calculated hydrogen per cent equals 15.08. Further confirmation that the desired reduction had not been effected was furnished through identification of the unchanged base through its picrate. Reduction of the Base, with Tin and Hydro- chloric Acid. The attempt to effect reduction with tin and hydrochloric acid was carried out in accordance with the 54 method used by Wikander in the reduction of 5,6,8 tri- methyl quinoline to 5,6,8 trimethyl-py-tetrahydroquinoline. sg. of the base, Ig. of granulated tin, and 3cc. of concentrated hydrochloric acid were heated together for 8-10 hours. The white stannous chloride salt which separated was filtered off and decomposed with strong sodium hydroxide solution. The solution was extracted with ether which yielded a colorless oil with the same refractive index (1.5080 at room temperature) as the original base. The unchanged base was further identified through the picrate as usual. It is evident that reduction was not effected by this procedure. Reduction of the C„l6?WL Base with Hydriodic Acid and Phosphorous. Here the procedure employed by Bamberger and 5 6 Williamson in the synthesis of decahydroquinoline from Py-tetrahydroquinoline was followed. One gram of base, 0.7 g. of red phosphorous and 3.5 cc. of concentrated hydriodic acid (spg. 1.70) were heated in a sealed tube slowly to 230° over a period of seven hours. Although considerable pressure was developed in the tube, only unchanged base was regained and there was no evidence that the desired reduction had been effected. Oxidation with Potassium Permanganate• Bz -1 et r ahydroquinolines behave like alkyl pyridines and are readily oxidized by a dilute = solution of potassium permanganate to pyridine carboxylic acids as has been shown by von Braun. Since the empirical formula, C_.H._N,among the many struct-16 do ural possibilities agrees with a number of 3z-tetrahydroquinolines, the oxidation of the base with permanganate was undertaken. One gram of base was suspended in lOOcc. of a 2 per cent aqueous solution of potassium permanganate and the mixture heated to boiling. After two hours 1 heating no apparent decolorization of the permanganate had taken place. A considerable excess of solid permanganate was then added and the heating continued for two weeks. After the reaction mixture had stood for four months the unchanged base was extracted with ether and the solution of permanganate decolorized with sulfur dioxide. The manganese dioxide sludge was filtered from the solution, which was then concentrated to a small bulk. A slight excess of nitric acid was added to the solution, but no precipitate of potassium salts resulted as should have been the case had pyridine pentacarboxylic acid or pyridine tetracarboxylic acid been present. Silver nitrate produced a voluminous precipitate. The silver salt, suspended in water, was decomposed with hydrogen sulfide, and after removal of the silver sulfide the acid was reprecipitated with silver nitrate. The acid, set free as before by treatment with hydrogen sulfide, was isolated in the form of a white crystalline solid. It melted at 101.4°, gave an insoluble calcium salt, and, in dilute sulfuric acid solution, decolorized potassium permanganate. The above properties revealed the oxidation product as oxalic acid. 26 Oxidation with Mercuric Acetate. Tafel oxidize d piperidine and Py-tetrahydroquinoline derivatives to the corresponding pyridine and quinoline derivatives by action of mercuric acetate solution at 130-150°. If the C, 16 25 compound is a Py-tetrahydroquinoline it must have an alkyl attached to nitrogen because it behaves in every respect like a tertiary amine. The bases oxidized by Tafel were secondary amines with an unsubstituted imino group. It was, therefore, evident that the action of the mercuric acetate could not go further than to oxidize away two hydrogen atoms. One half gram of base, 2.5 g. of mercuric acetate, and 2.5 cc. of 50 per cent acetic acid were heated in a sealed tube at 150° for 12 hours. Since there was no separation of mercury it was evident that the substance had not undergone oxidation and this was confirmed by recovery of the original base in the usual way. Oxidation with Silver Sulfate. According to Tafel*^ 8 indoline is energetically oxidized to indole on heating with silver sulfate and kieselguhr. Since a highly alkylated indoline represents one of the many structural possibilities for the base, the oxidation of the base with silver sillfate was attempted. One gram of base was added to a mixture of 0.75g.0f silver sulfate with several times its weight of kieselguhr. The small distilling flask containing the mixture was heated gently with a free flame. No vigorous reaction followed; a distillate of approximately lee. was collected and identified as the original base. Oxidation of the 3ase, with Lead Dioxide. 42 Borsche oxidized 1,2,3,4 tetrahydrocarbazole to carbazole by heating with lead dioxide and it is probable that more highly hydrogenated carbazoles can be oxidized to carbazoles under the same conditions. Since the formula, corresponds to a tetramethyl-octahydrocarbuzole the following experiment was carried out: A pyrex tube, 35cm. in length and l-2cm. in diameter, was drawn out at one end and a delivery tube was sealed on. A 15 cm. layer of pumice stone, impregnated with lead dioxide, was placed in the end of the combustion tube next to the delivery tube. This was followed by lead dioxide impregnated with Ig. of base and a second 15cm. layer of pumice mixed with lead dioxide. While a slow current of air was being aspirated through the combustion tube, the layer of lead dioxide was heated gently for some time and then the whole length of the tube as well. With the temperature increase 0.95 g. of oil distilled into the receiver, however, proved to be unchanged base despite the favorable conditions in this experiment for the formation of an oxidation product. Action of Bromine on the Base. With an excess - of bromine in chloroform Py-tetrahydroquinoline is both readily oxidized and brominated with the formation of tri-56 bromoquinoline. An attempt was made to oxidize the base in the same manner without success. 2o 0.5 g. of base dissolved in chloroform was added an excess of bromine. The mixture was warmed and allowed to stand over night. The chloroform was evaporated and the red oily residue taken up in a small amount of hot alcohol which gave no precipitate on cooling, in contrast to Py-tetrahydroquinoline under the same conditions. The alcoholic solution was treated with 0.5 g. of picric acid and gave a picrate melting at 148° (unc.), the melting point of the picrate of the original base. In a second experiment one half gram of base and o.2cc. of bromine were sealed in a tube. The mixture became slightly warm, the bromine disappeared, and a viscous red-brown gum, insoluble in water and dilute acids and soluble in glacial acetic acid, was formed. On treatment with caustic soda the product was reconverted to the original base. Since the true aromatic base, 2,3,8 trimethyl quinoline,accompanies the hydroaromatic base, in the 276° fraction of kerosene bases it was of interest to study its behavior toward bromine. The addition of bromine to a chloroform solution of 2,3,8 trimethyl quinoline is followed by the immediate precipi tation of a crystalline product which dissolves in hot alcohol and separates on cooling. The difference in behavior of the aromatic and hydroaromatic kerosene base toward bromine may possibly be utilized in developing a method for their separation. Nitrie Acid Oxidation, of the Base, T wenty grams of the base were oxidized by heating with 1-1 nitric acid in Ig. lots according to the following procedure: One gram of base and lOcc. of 1-1 nitric acid were sealed in a Carius tube and heated slowly to 170° in the course of two hours and the heating then continued for three hours at this temperature. Considerable pressure was evident on opening the tube. After the removal of the excess of nitric acid by distillation in a vacuum, there remained a thick, light yellow gum. This material was dissolved in hot water and the solution concentrated to a gummy consistency by distillation under reduced pressure. From the residue dissolved in scc. of hot water there separated on cooling a small amount of a viscous, yellow oil, to be referred to as product A. The nitric acid distillates yielded with ether extraction only a minute amount of white solid which was not further investigated. purification product A was dissolved in a small volume of hot alcohol from which 0.15 g. of yellow crystals separated on cooling. It crystallizes from alcohol in the form of small rhombic plates which melt at 237°. 'She substance is insoluble in water and difficultly soluble in dilute caustic soda solution. From analyses and a molecular weight determination the formula, » was derived. Analyses. Calcd. for C_ : C, 48.48; H, 5.05; 10 CU O T: K, 14.14. Found: C, 47.90; H, 4.92; B, 14.47. Molecular weight. Calcd. for C n :396. Found:3Bl. After removal of product A the remaining oxidation products consisting of acids and bases were investigated as follows: First the solution was made alkaline with caustic soda whereupon 5.1 g. of bases separated and was extracted with ether. After driving off the ether the residue was taken up in alcohol and precipitated with picric acid. In order to effect crystallization the viscous picrate was treated with a small amount of hot alcohol and then recrystallized from alcohol. In this way, I.lg. of a picrate was obtained in the form of diamond shaped plates, melting at 230.8°, apparently unchanged. The base, freed from the picrate in the usual way with ammonium hydroxide solution, was extracted with ether, and traces of ammonium picrate washed out with water. On evaporation of the ether there remained 0.5 g. of a new base to be referred to as product B. This was dissolved in 4cc. of hot alcohol from which 0.23 g. of base separated on cooling in the form of slender rods melting at 116.5°. It is insoluble in water and readily soluble in organic solvents. From analyses and a molecular weight determination the formula, was derived. Analyses. Calcd. for C n C, 59.81; H, 7.16; 16 23 4 3* * * N, 13.09. Found: C, 59.89; H, 7.32; N, 12.80. Molecular weight. Calcd. for CHON: 321. Found;32o. 16 23 4 3 The alcoholic solution of residual picrates was evaporated to dryness. The viscous picrate mixture defied all attempts at recrystallization and has not been investigated further. The alkaline residue from which the crude bases were obtained was concentrated to a volume of sOcc. and on the addition of 15cc. of 1-1 nitric acid a flocculent precipitate formed. After filtration from the solution this material was extracted with boiling alcohol and a product finally obtained in the form of a white granular solid with a wide solubility range in cold and hot water. Since on ignition it left an alkaline residue, the substance was dissolved in warm water and precipitated with silver nitrate. The colloidal silver salt was filtered off, thoroughly washed, suspended in water, and the silver precipitated as sulfide. From the filtered solution, concentrated to a very small bulk, the new acid, to be referred to as product 0, crystallized out as a mass of hair-like needles. Product C is soluble in cold water, more soluble in cold alcohol, and readily soluble in hot water or hot alcohol. The aqueous solution of the compound is acid towards methyl orange and gives a deep red coloration with ferrous sulfate. The compound may be purified by recrystallization from water from which it separates as hair-like needles, melting at 240-241° with decomposition. From analyses and a molecular weight determination the formula, C3HSOSN, was established. Analyses. Calcd. for 45.50; H, 2.37; N, 6.63. Found: C, 45.29; H, 2.56; N, 6.83. Equivalent weight. Calcd. for C 8 H 5 0 6 N: 70.3. Found: 71.3. Molecular weight. Calcd. for 211. Found: 216. On attempting to decolorize the acid with animal charcoal a product was obtained which, unlike the acid, is difficultly soluble in hot water and hot alcohol. On ignition the product chars and leaves an alkaline residue. The substance gives a test for calcium but none for phosphates and, evidently, is a calcium salt of A search of the literature of organic chemistry failed to disclose another instance of the preparation of a calcium salt through the use of animal charcoal. The salt crystallizes from a large volume of hot water in the form of characteristic thick rhombic prisms. The formation of an insoluble calcium salt on boiling with animal charcoal was subsequently used as a test for the presence of the acid. The acid solution, after the isolation of product C, was evaporated to dryness and the residue extracted with 30cc. of boiling alcohol. About 4g. of a brown gum remained on evaporation of the alcoholic extract. Attempts to separate individual compounds from this product were unsuccessful. Twenty-four grams of the base, were oxidized as described above. In this experiment the nitric acid was distilled off in a vacuum, water added to the viscous residue and the solution again concentrated under reduced pressure. This process was repeated several times and the free nitric acid almost completely removed. The gummy residue dissolved in 59cc. of hot water and on dilution a small amount of product A separated. On extraction with ether in a continuous extractor for several days, the extract yielded approximately 3g. of a light yellow, viscous product which has not been investigated. The residual solution from the ether extraction was treated with 278 cc. of N/3 barium hydroxide solution and the precipitated bases extracted with ether. In this way, 2.5 g. of bases was obtained which, worked up as described above, yielded a small amount of the picrate of product B. Th© alkaline solution was next treated with the calculated amount of N/l sulfuric acid to precipitate the barium as sulfate. After filtering from barium sulfate the solution was evaporated on the water bath. A little of the viscous residue was dissolved in a small amount of hot alcohol from which a white solid separated on cooling. A similar result was obtained when acetone was used as the solvent. All of the gum was dissolved in a small volume of hot alcohol and the solution allowed to stand in the ice box over night. Very little crystalline product separated. Since this procedure proved unsatisfactory, the material in the alcoholic solution was precipitated with ether and then dissolved in a small volume of hot acetone. After long standing in the ice box 0.84 g. of a white crystalline substance separated. This will be referred to as product 1). Product D recrystallized, from alcohol in the form of o small, colorless prisms which melt at 189-190 with decomposition. It is readily soluble in cold water and hot alcohol, but only slightly soluble in ether and cold alcohol. The compound is readily oxidized by dilute solution of potassium permanganate. Unlike product C it fails to give colorations with ferric chloride or ferrous sulfate, and, in contrast to product C does not form an insoluble calcium salt on boiling with animal charcoal. From analyses and a molecular weight determination the formula, c u h 14 o 6 b 2 , was derived. Titration showed this substance to be a dibasic acid and therefore the presence of two carboxyl groups is established. This account for four of the five oxygen atoms in the molecule. The fifth oxygen, since the substance is colorless, is probably in oxime form. Analyses. Calcd. for C, 51.97; H, 5.51; H, 11.02. Found: C, 52.48; H, 5.43; N, 11.03. Equivalent weight. Calcd. for CiiH l4 0 5 E 2: 27 ’ Found: 124. Molecular weight. Calcd. for 254. Found: 235. After removal of product D, the residual acetone solution was evaporated to dryness yielding 3.7 g. of viscous residue. This was dissolved in water and barium hydroxide added to faint alkalinity, whereupon a small amount of a barium salt separated. The investigation of this salt and the residual solution has not been completed. In a third experiment, v/here 24g. of base was oxidized and the excess of nitric acid removed as in the second experiment, the viscous oxidation products were dissolved in 150 cc. of water and filtered from a small amount of product A. M/3 barium hydroxide was then added until the solution reacted alkaline to methyl orange. Approximately Ig. of the barium salt of product C separated in a nicely crystalline form from which the acid was liberated by treat ment with the calculated amount of dilute sulfuric acid and vzorked up as previously described. The filtrate from the above barium salt was made alkaline to phenolphthalein and yielded 3.5 g. of bases on other extraction. The slightly alkaline solution was finally evaporated to dryness on the water bath and the solid residue treated with glacial acetic acid which dissolved the organic material leaving undissolved the barium nitrate. After driving off the acetic acid, the residue was taken up in alcohol and and a crystalline product precipitated on the addition of ether. The ether-alcohol solution was evaporated and yielded a dark brown viscous liquid. This material contained no barium and was only slightly soluble in water but dissolved readily in dilute acid or alkali. This has not been investigated. The crystalline precipitate from the ether-alcohol solution, dissolved in water, gave, with an excess of silver nitrate solution, a voluminous precipitate which was filtered off, suspended in water, and decomposed with hydrogen sulfide in the usual manner. After removal of the silver sulfide the acid was again precipitated with silver nitrate. The silver was again removed as sulfide and the solution evaporated to dryness. For purification of the organic acid obtained here the residue, weighing 2g., was dissolved in hot acetone from which a cream colored product separated on cooling. This was soluble in water and only slightly soluble in alcohol. Its aqueous solution gave a yellow coloration with ferrous sulfate and, unlike product C, failed to give an insoluble calcium salt on boiling with animal charcoal. It may, therefore, be concluded that this represents a third acid isolated from the nitric acid oxidation products of the base. The filtrate from the silver nitrate precipitation was treated with hydrogen sulfide to remove the excess silver nitrate. After filtering off the silver sulfide the solution was evaporated to dryness and yielded 7g. of crystalline residue, consisting chiefly of barium nitrate. It is doubtful whether any oxidation products other than the ones enumerated can be isolated. The following procedure was finally adopted for the isolation of the four oxidation products previously described and investigated: Twenty-four grams of base was oxidized with nitric acid in the usual way and the excess nitric acid removed as before. The residue was treated with 150 cc. of water and the solution decanted from product A which did not dissolve. The solution was now made alkaline to methyl orange with N/3 barium hydroxide and filtered from the barium salt of product C. On further addition of barium hydroxide to the phenolphthalein end point 4.2 g. of bases separated and was extracted with ether. After removal of the bases the alkaline solution was treated with the calculated amount of N/l sulfuric acid, the barium sulfate filtered off, and the solution evaporated to dryness. The viscous residue dissolved in a small volume of alcohol to which ether was added as long as it produced precipitation. The material obtained in this way was dissolved in a small volume of hot acetone and allowed to stand several days in the ice box; Ig. of product D in crystalline form was obtained. It is apparent that large amounts of base would be required to prepare the oxidation products described. The substances here that may be expected to be of greatest value in proof of structure of the CqgHg-N base are the acids referred to as products C and D. 51. Ferratini, Adolfo: ”Ueber die Verwandlung der Indole in Chinoline.” Ber. 26, 1812 (1893). 52. Cohen, J. 3., ’’Organic Chemistry for Advanced Students”, Longmans, Green and Company, New York, 1928, Part 111, p. 403. 53. Ladenburg, A: "Ueber Pyridin- und Piperidinbasen." Ann. 247, 43 (1888). 54. Wikander, Hjalmar: ft Ueber einige neue Derivative des o-p-ana-Trimethylchinolins." Ber. 33, 646 55. Bamber/rer, ling, and Sidney Williamson: ’’Uber das Lecahydrochlinolin. ’’ Ber. 27, 1495 (1894). 56.Hoffmann, Leo and W. Konigs: "Über Tetrahydrochinolin." 3er. 16, 727 (1883). SUMMARY AND CONCLUSION (1) Additional evidence has been obtained that only very small amounts of nitrogen bases occur in crude petroleum and that appreciable amounts of nitrogen bases first appear in petroleum distillates as a result of pyrolysis at distillation temperatures. (2) It is shown that the quantity of extradtable bases in petroleum distillates may be increased by redistillation. (3) It is shown that both aromatic and hydroaromatic bases exist in petroleum distillates. (4) The isolation of two bases, C U H I2 N and C l6 H 26 fi, from the residues obtained in refining California kerosene distillates by the Sdeleanu process, is described. (5) The base, 11513 been characterized by the preparation of several salts and experimental work carried out in an attempt to determine its structure. (6) It is shown that the base, is not a simple alkylated derivative of tetrahydroquinoline, tetrahydroisoquinoline , dihydroindole, octahydrocarbazole, or decahydroacridine. The structural problem involved is without doubt as complex as those presented in alkaloidal chemistry. BIBLIOGRAPHY Bamberger, X. and W.Dieckmann: "Ueber Tetrahydrur des Isochinolins." Ber. 26, 1242 (1893). Bamberger, B. and'F.Lengfeld: "Neue Seductions products des Chinolins." Bcr. 23, 1142 (1890). Bamberger, 5. and 3.Williamson, "Über das Lecahydrochinolin." Bor. 27, 1495 (1894). Bamberger, E. and P.Wulz: "Ueber Homologe des Tetrahydrochinolins." Ber. 24, 2055 (1891). 14 3andro T .vski, Franz X. : "Über das Vorkominen Alkaloidartiger Basen im galizischen Soherdßle." Monats. 8, 224 (1887). Beilby, George: "The Nitrogen of Crude Petroleum and Paraffin Oils." Jr. Soc. Chem. Ind. 10, 120 (1891). "Preparation of Ammonia from Nitrogeneous Minerals." Dingl. polyt. J. 254, 342 (1884). Borsche, Von W. : "Ueber Petra- und Hexahydriocarbazolver bindungen und eine neue Carbazolsynthese. " Ann. 359, 49 (1908). von Braun, J., A.Petzold, and A.Schultheiss: "Ueber Bz-tctrahydrochinoline und ihre Berivate % III: Tricyclishe Verbin dungen. " Ber. 56, 1348 (1923). von Braun, J., A.Petzold, and J.Seemann: "Katalytische Hy- dri erung unter Bruck bei Gegenwart von Nickelsalzen, IV. Chinolin-Basen." Ber. 55, 3782 (1922). von Braun, J. and H.Ritter: "Katalytische Hydrierung untcr Bruck bei Gegenwart von Nickelsalzen, V; Ber Carbazol- Komplex." Ber. 55, 3792 (1922). von Braun, J. and K.Weissbach: "Bntalkylierung tertiarer Amine durch organische Sauren." Ber. 63, 489 (1930). von Braun, J. and L.Schering: "Über hydrierter Carbazole." Ber. 58, 2156 (1925). Brunner, K. : "Ueber Indolinone." Monats. 18, 118, (1897). Chlopin, G.W.: "Die organischen Basen des russischen Hrdoles." Ber. 33, 2837 (1900). Cohen, J.B. , "Organic Chemistry for Advanced Students", Longmans, Green and Company, New York, 1928, Part 111. Day, D.T., "A Handbook of the Petroleum Industry", John Wiley and Sons, New York, 1922, Vol. I. Decker, H. and G.Dunant: "Notiz ueber das Vorkommen des Hydroacridine in Steinkohlenteer." Ber. 42, 1178 (1909). Devine, J. M. and F.W.Lane: "The Use of the Carius Method for the Determination of Sulfur in the Less Volatile Petroleum Oils." J.A.C.S. 50, 1707 (1928). Doebner, O. ; and W*v. Miler: "Übor Chinaldenbasen." Her. 16, 2464 (1883). Hngler, C. and H.Hofer, "Das Erdol", Hirzel, Leipzig, 1913, Vol.l. Ferratini, A.: "Ueber die Verwandlung der Indole in Chino- line." Ber. 26, 1812 (1893). Floyd, W.W., Dissertation, The University of Texas, 1928. Freund, M. and G.Bode: "Über Einwirkung von Grignard-Losungen auf Halogen Ammoniumverbindungen." Ber. 42, 1746 (1909). Freund, M. and E.Kessler: "Sinwirkung von Organomagnesiumverbindungen auf Chinolinjodmethylat, ein Beitrage zur Stereochemie stickstoffhaltiger Verbindungen." J. Prakt. Chemie (2) 98,233 (19191. Griffiths, A.B. and N. J. Bluman: "Les bases azotees dans le petrole roumain." Bull. Soc. Chim. (iii) 25, 725 (19011,. Hamor, W.A. , and F.W.Padgett, "The Examination of Petroleum", McGraw-Hill Book Company, New York, 1920, Ist ed. Hantzsch, A.: "Ueber die Synthese pyri dinartiger Verbindungen aus acetessigather und Aldehydammoniak." Ann. 215, 62 (1882). Hatcher, W.H. and P.W.Skirrow: "The Compounds of Phenol and the Cresols with Pyridine." J.A.C.S. 39, 1939 (1917). Herzig, J. and H.Meyer: TtT Jeber den Nachweiss und die Bestimmung des an Stickstoff gebundenen Alkyls." Llonats. 15, 613 (1894). Kaku, 7: "Atophan Derivatives II: Reduction of 2-Anisyland 2-Phenylquinolines." C.A. 23, 392 (1929). J. Pharm. Sec. Japan, 48, 693 (1928). Kann, M. and J. Tafel: "Ueber die Oxydation hydrirter Indol." Ber. 27, 826 (1894). Ladenburg, A.: "Ueber Pyridin- und Piperidinbasen." Ann. 247, 43 (1888). Mabory, C.F.: "On the Nitrogen Compounds in California Petroleum." Jr. Soc. Chern. Ind. 19, 502 (1900). "The Genesis of Petroleum as Revealed by Its Nitrogen Constitutents." J.A.C.S. 41,1696 (1919). "Composition of American Petroleum." J.A.C.S. 28, 426 (1906). Mabery, C.F. and L.G.Wesoon: "The Constitution of the Organic Nitrogen Bases of California Petroleum." J.A.C.S. 42, 1014 (1920). Oshima, V. and K.lshibashi: "Basic Compounds in Low-Temperature Coal." Jr. Soc. Chern. Ind. Japan, 29, 445 (1926). Perkin, and S.G.P.Plant: "1:2:3:4:5 :6|7 :8-octa- hydrocarbazole and its Derivatives." J. Chern. Soc. 125, 1503 (1924). Poth, 3.J. : "A Receiver for Vacuum Distillation." Analytical Edition, Jr. Ind. Eng. Chern. 1, 111 (1929). Poth, E. J. , W.D.Armstrong, C.C.Cogburn, and J.H.Baoley: "The Estimation of Nitrogen in. Petroleum and Bitumens." Ind. Eng. Chern. 20, 83 (1928). Poth, E. J., W. A. Schulze, W.A.King, W.C.Thompson, W.M. Slagle, W.W.Floyd, and J.R.Bailey: "An Investigation of the Bases in the Kerosene Distillate of California Petroleum." J.A.C.S. 52, 1239 (1930). ?yhal&, SU : "Einige neue stickstoffhaltige Kohlonwasserstoffe aus dem Bakuer Srdol." Chem Otg. 46, 953 (1922). de Saussure, T.: "Recherches sur la composition du naptha d’Amiano, dans les etats de Parma." Annales chim. phys. 4, 314 (1817). Schestakow, P.J.: "Stickstoffverbindungen des kaukasischen Brdols." Chem. 3tg. 23, 41 (1899). Slagle, W.M. , Dissertation, The University of Texas, 1928. Tafel, J.: "Usher die Oxydation hydriertor Pyridin- und Chinolinbasen." Ber. 25, 1619 (1902). Weller, A.: "Vorkommen alkaloidartigen Basen in Paraffinol." Ber. 20, 2098 (1887). Wikander, H.: "Ueber einige neue Derivative des o-p-ana- Trimsthychinolins." Bor. 33, 646 (1900). Yamaguchi,3.; "Methylhydroquinolines. " I, 11. C.A. 21, 296 (1927). J. Pharm. Soc. Japan, No. 533, 556 (1926). No. 535, 749 (1926). Saloziocki, R. ; "Über pyridinartige Basen im Brdol." Monats. 13, 498 (1892).