EXCHANGE Lib. JUL 31 The Fractionation of California Petroleum by Diffusion through Fuller's Earth DISSERTATION SUBMITTED TO THE BOARD OF UNIVERSITY STUDIES OI ; THE JOHNS HOPKINS UNIVERSITY IN CONFORMITY WITH THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY BY PHILIP SCHNEEBERGER JUNE, 1913 EASTON, PA. ESCHENBACH PRINTING Co. 1913 The Fractionation of California Petroleum by Diffusion through Fuller's Earth DISSERTATION SUBMITTED TO THE BOARD OF UNIVERSITY STUDIES OF THE JOHNS HOPKINS UNIVERSITY IN CONFORMITY WITH THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY BY PHILIP SCHNEEBERGER JUNE, 1913 EASTON, PA. ESCHENBACH PRINTING Co. 1913 TABLE OF CONTENTS. Page Acknowledgment 4 Introduction 5 Description of Petroleums 6 Investigation of I California Oil 7 Fractionation 10 Determination of Sulphur 1 1 Investigation of II California Oil 14 Determination of Nitrogen 17 Investigation of Benzene and Paraffin Oil 19 Investigation of III California Oil 30 Preliminary Fractionation 32 Preliminary Analyses 33 Fractionation 35 Sulphuric Acid Absorption 38 Bromine Absorption 39 Proportion of Different Hydrocarbons Present 42 Summary 45 Biography 47 ACKNOWLEDGMENT. The author wishes to express grateful acknowledgment to Professors Remsen, Morse, Jones, Acree, Gilpin and Lovelace for instruction in the lecture room and laboratory. He also takes this opportunity to thank Dr. Gilpin, under whose care- ful direction this investigation was carried on, for valuable sug- gestions and practical assistance in the research; Dr. Frazer for aiding him in arriving at theoretical deductions; and Dr. Day, of Washington, for helpful suggestions and for furnishing material for investigation. Fractionation of California Petroleum by Diffusion through Fuller's Earth. 1 For several years investigations have been in progress in this laboratory upon the effect produced by diffusion of pe- troleum oils through fuller's earth. These investigations were pursued in order to obtain some idea of the changes produced in the process of diffusion to which the oils have been sub- jected in their passage from the place of formation to their present location; also, to gain some idea of the substances present in the natural oil by a separation of the constituents by a process not involving the use of heat, and thereby pro- ducing changes by cracking and otherwise. The results ob- tained when a light oil from Pennsylvania and a heavier oil from Illinois were thus fractionated have already been pub- lished. 2 In the present investigation, this method of fractiona- tion was extended to a very heavy petroleum from California. As will be later described, the different fractions obtained by such diffusion were studied with regard to their content of 1 This research was aided by a grant received from the C. M. Warren Committee of the American Academy of Arts and Sciences. 2 Gilpin and Cram: "The Fractionation of Crude Petroleum by Capillary Diffusion," Am. Chem. /., 40, No. 6, December, 1908. Gilpin and Bransky: "The Diffusion of Crude Petroleum through Fuller's Earth," Ibid., 44, No. 3, Sep- tember, 1910. paraffin, benzene, and olefin hydrocarbons, and to the amount of sulphur and nitrogen compounds found in them and in the- earth through which they passed. For the sake of compari- son, the behavior of mixtures of known amounts of benzene and pure paraffin oil, when allowed to diffuse through fuller's earth, were also studied. DESCRIPTION OF OILS USED California Oil L Viscous, brownish black in color; of a syrupy consistency, and failed to flow from a small vessel when cooled to 10; possessed a disagreeable odor, sugges- tive of organic sulphur compounds; specific gravity at 20, 0.912; when distilled, the first fraction came over at 90, colorless and agreeable in odor; the last fraction came at 380, brown, disagreeable in odor, resembling garlic, but supposed to be due to hydrides of the hydrocarbons, formed during the cracking of the oil; analysis showed appreciable amounts of sulphur compounds present. This oil came from Kern County, California. California Oil II. Less viscous than the first oil, and of less disagreeable odor; specific gravity, taken with a delicate Westphal balance at 20, 0.8890; when distilled, fractions were obtained from 100 to 350; contained a small proportion of benzene hydrocarbons and 0.760 per cent, of nitrogen compounds; no trace of sulphur compounds was found. The oil came from Well No. i, Section 30-30-24, Elk Hills, Kern County. California Oil III. Fairly viscous, brownish black oil, of somewhat disagreeable, smoky odor; specific gravity, 0.9118 at 20; when distilled, fractions were obtained from 105 to 340, attempts to obtain fractions at higher temperatures resulting in the cracking of the oil, giving fractions ranging around 270; rich in benzene and olefin hydrocarbons, but entirely free from nitrogen and sulphur compounds; its prop- erties resembled closely those of the first California oil. The petroleum came from Well No. i, Section 30, Elk Hills, Kern County. Pennsylvania Oil. A light, thin, dark brown oil from Ve- nango County, Pennsylvania; possessed an agreeable odor; specific gravity at 20, 0.8470. This was the same oil that had been investigated by Gilpin and Cram, and Gilpin and Bransky. Between the time when this oil was first studied and its investigation in 1912 and 1913, its specific gravity had increased from 0.810 to 0.8470 by evaporation through the barrel staves. INVESTIGATION OF CALIFORNIA OIL I The oil first studied was the heavy petroleum from Kern County, California. A description of this oil, named the Cali- fornia Oil I, is found on page 6. The method of handling this oil was practically the same as that introduced by Gilpin and Cram and improved by Gilpin and Bransky. The apparatus in which the diffusion was car- ried out was similar to that employed by Gilpin and Bransky. Two minor additions were made to the apparatus. The first of these was a manometer which recorded pressures from 730 mm. to o mm. when attached to the exhaust system. The other was a device which it was found necessary to put in series with the exhaust system, owing to the fact that the exhaust was obtained by use of a large Chapman water pump. Fluc- tuations in the water pressure were accompanied by fluctua- tions in the amount of exhaust. The device by which the suc- tion could be maintained uniform for any length of time consisted of a sliding tube with perforations at its lower end, that could be adjusted by raising or lowering in a reservoir of mercury according as lower or higher pressures were de- sired. It was found at the start that an oil as heavy as this one could not be made to diffuse of itself by capillarity at room tem- peratures (19 to 23). To produce the necessary diffusion, reduced pressures were brought to bear, pressures as low as 12 mm. of mercury being maintained for days at a time. In the preliminary experiments, sixteen tubes were filled with fuller's earth. 1 The tubes were filled by dropping into them an amount of earth that would form a column about a 1 This earth, known as "XXF Clay," and the fuller's earth used in later work were obtained by courtesy of the Atlantic Refining Co., Philadelphia, Pa. 8 foot in height. Since all ranges of compactness of the earth were desired, the earth in some tubes was not compressed in any way. In others it was rammed slightly by a rod tipped with a rubber stopper, In a third set, it was rammed fairly hard, and in a fourth as hard as possible. The tubes were then allowed to stand undisturbed for a short while, so as to permit the cushions of air held between layers of the earth to escape. A second column of earth a foot in height was then added, the same degree of packing observed, and the process repeated until the tubes were filled. They were then placed with their lower ends in separate reservoirs of oil, and a re- duced pressure of 600 mm. was then applied to the upper ends. This failed to produce any diffusion, as could be deter- mined by the level of the oil in the reservoirs, so the pressure was gradually reduced until 20 millimeters of mercury were registered on the manometer. There was, however, no sign of actual fractionation of the oil in any of the tubes, but, in- stead, at the low pressure that was employed the oil was drawn up through the earth unchanged in any of its properties. An explanation of the failure of the oil to fractionate was found in the high viscosity of the petroleum. The oil, instead of diffusing through each minute particle of earth, was sucked around the particles and emerged unaltered at the top of the tube. The cause of the high viscosity of the oil was doubtless the very large quantity of bitumen and the complex hydrocar- bons present, whose boiling points were as high as 380. Ac- cordingly, efforts were directed toward reducing the viscosity of the oil by coagulating the bitumen. It was shaken with a solid electrolyte and with a solution of the same, but repeated experiments failed to show any perceptible reduction in specific gravity or viscosity, nor was there any noticeable coagulation of the bituminous material held in suspension in the oil. It was observed, however, that any rise in temperature of the oil was accompanied by a marked decrease in viscosity. Accordingly, the effect of this reduction in viscosity upon the fractionation of the crude petroleum was next studied. In order to secure uniform conditions, it was deemed advisable to have the reservoirs for the oil and also the lower part of the tubes immersed in a large water bath, the temperature of which could be regulated and maintained uniform by a gas thermostat. An appliance that could withstand the action of the heated water and oil was, therefore, devised. Glass bot- tles, fitted with water-tight stoppers and safety tubes, could not be employed, for these broke when temperatures above 60 were reached. For the fractionation at high tempera- tures, each tube was constructed of brass, and was made to fit snugly into a closed reservoir of a liter in volume. Contact between the tube and the neck of the reservoir was made per- fect by having a lead washer which was held rigidly between the two by a nut that screwed down upon the neck of the reservoir. To determine the most desirable temperature at which the experiments should be carried out, the viscosity of the oil was measured at different temperatures. The apparatus was the same as used by a previous experimenter. 1 With the viscosimeter it was found that a measured quantity of the oil used by Bransky diffused at the following rates: 4.5 cc. of Pennsylvania Oil (sp. gr. 0.8470) : Minutes At 20 10.25 At 40 6.00 When a similar quantity of the California Oil I was run through the viscosimeter, it was found that the time required was vastly greater. The following results were obtained at different temperatures : Table I 4.5 cc. California Oil I (sp. gr. 0.912) Minutes At 20 84.0 At 40 41.0 At 60 17.5 At 75 ii. 3 At 90 8.1 1 Am. Chem. /., 44, No. 3. 10 From the above results, it is apparent that a temperature of at least 75 is necessary for diffusion, since the viscosity of the oil at lower temperatures was so great as to prevent its fractionation by diffusion. The temperature of 90 was ruled out because at that temperature there was a tendency for the lightest ingredients of the oil to distil off. This was deter- mined by keeping a measured quantity of the oil at 90 for a week. A quantity amounting to about two per cent, was found to distil off. For this reason it was deemed advisable to carry on the diffusion of the oil at 75. Accordingly, eight tubes packed with varying degrees of hardness were maintained at 75, and at the end of nine days these yielded the first fractions of the heavy California oil (specific gravity, 0.912). Nearly 90 liters of the oil had to be used to obtain the four fractions in the amounts shown be- low: Table II First Fractionation of California Oil I Quantity Fraction Sp. gr. Liters I-A 0.8695 1 1.2 i-B 0.882 1.4 i-C 0.9025 1.8 i-D 0.904 1.75 Total 6.15 There was thus secured a small proportion of available frac- tions. This was obtained, however, only after very great difficulties had been overcome and with a loss of nearly 93 per cent, of the original oil. The results demonstrate, never- theless, the possibility of fractionating a heavy, viscous petro leum. Owing to the difficulties encountered in working with large quantities of the oil at elevated temperatures, it was de- cided to discontinue work on this heavy petroleum until more adequate means of handling it under the necessary conditions were at the disposal of the experimenter. The four fractions from the California Oil I, designated as i-A, i B, i-C and i-D, were then analyzed for their sulphur 1 All specific-gravity measurements were taken at exactly 20" with a delicate Westpbal balance. II content. Similar determinations were made for the oil re- tained by the earth. Determination of Sulphur Compounds in the California Oil I The odor of the high-boiling fractions of this oil led to the suspicion that sulphur compounds were present. Accordingly, qualitative tests were made to detect the presence of this element. These tests were as follows : About 10 cc. of the sample to be tested were heated to boil- ing in a flask provided with a reflux condenser. About half a gram of metallic sodium was introduced, and the liquid heated to the boiling point for about thirty minutes. After cooling to room temperature, water was gradually introduced through the condenser and the flask shaken until the sodium had gone into solution. The solution of the hydroxide containing the sodium sulphide was separated, and the addition of sodium nitroprusside produced a purplish blue or deep purple colora- tion. A quicker method for the qualitative determination of sulphur was one that is employed in the petroleum industry. A solution of litharge in concentrated potassium hydroxide solution was prepared. The undissolved litharge was allowed to settle, and the clear solution decanted. Several cc. of the oil to be tested were shaken with a small quantity of the potas- sium plumbite solution, and the oil allowed to rise. The plumb- ite solution was colored from pale brown to black, according to the amount of sulphur present. Quantitative determination of sulphur by the usual Carius method failed, the bombs exploding in every one of the fifteen analyses that were attempted. Carius determinations made in hard glass tubes gave no better results. A modified Carius method was then employed. This was essentially as follows: A weighed sample of the oil was introduced into a Carius tube and 5 cc. of fuming nitric acid were poured upon it. The open tube was then heated for two hours in a water bath at 100. Five more cc. of the acid were then added, and the heating re peated for two more hours. Then five cc. of the acid were again introduced, the tube was drawn out and sealed, and then heated for two hours in a bomb furnace at 280. The charge 12 | of the tube, when cool, was emptied into several hundred cc. of water, and the sulphur determined as barium sulphate by the usual method. This modified Carius method gave results that were somewhat lower than the results obtained by the methods finally employed. This was probably due to the volatilization of the sulphur compounds when the open tube was heated. The methods found to be most satisfactory, however, were the Sauer combustion method, described in detail in Morse's Exer- cises in Quantitative Chemistry (pp. 258-60), and a method by which the oil was oxidized in a flask by concentrated nitric acid and potassium chlorate. The two methods gave con- cordant results. In every instance duplicate analyses were made, and the mean result given in the table below. The frac- tions analyzed were those described on p. jo- Table 111 Analysis for Sulphur Fraction Sp. gr. Per cent, sulphur i-A o . 8695 o . 1 1 i-B 0.882 0.144 i-C 0.9025 0.174 i-D 0.904 0.29 Crude oil 0.912 o*54i In order to see if the proportion of sulphur could be further reduced by filtration through fuller's earth, Fractions A and B were refractionated and the results of the second fractiona- tion were analyzed with respect to their sulphur content. The fractions from A are designated as i-A i, i-A 2 and i-A3; those from i-B as i-B i, i-B 2 and i-B 3. Table IV Second Fractionation: Analysis for Sulphur Fraction Sp. gr. Per cent, sulphur -A i 0.^57 0.06 -A 2 0.8604 0.07 -A 3 0.869 - IO 4 -B i 0.8625 0.072 -B 2 0.8771 0.09 -B 3 0.8803 0.141 The results tabulated above show that there is a gradual absorption of sulphur compounds by the fuller's earth. Those fractions that collect at the top of the tubes have the smallest proportion of the sulphur compounds in them, and the amount increases as fractions lower down are taken. The explanation of the comparatively small amount of sulphur in the upper- most fraction is probably as follows: the original petroleum penetrates a grain of the porous fuller's earth, emerges on the other side and has a portion of its sulphur removed and re- tained by the earth. The possible mechanism of this selective absorption by the earth is discussed later. The oil then enters another grain of earth, the absorptive action of the earth is repeated, and more of the sulphur is removed. The oil that thus penetrates by diffusion up to the top of the tube has passed through the greatest number of grains of earth and thus contains less sulphur than the oil that has not penetrated so far. Moreover, the oil that follows in the track of the first particle of oil which penetrates the earth finds particles of earth that have already taken up the greatest quantity of sulphur that is possible for them to absorb. Hence the second particle of oil passes through earth that is already saturated with re- spect to its power to absorb sulphur compounds, and it may pass through unchanged, or nearly so. This accounts for the fact that the fractions that are nearest to the original petro- leum have chemical and physical properties that closely re- semble those of the crude petroleum. Judging from the results obtained from the second frac- tionation, it is probable that if enough of the lighter fractions were available for carrying on several more fractionations, the sulphur could be entirely removed. In order to show to what extent the earth retained the sul- phur compounds which were originally in the oil, the earth from which the oil had been displaced by the addition of water was extracted with ether. By this process, oils were ob- tained which were heavier than those oils expelled by the ad- dition of water. These were analyzed for sulphur, with the following results: 14 Table V Analysis of Oil Retained by Earth Section of earth Sp. gr. Per cent, sulphur Oil extracted from A o . 8955 o . 1 1 Oil extracted from B 0.9038 0.237 Oil extracted from C 0.9105 0.42 Oil extracted from D 0.927 0.675 These results show that it is undoubtedly the earth through which the oils pass that retains the sulphur compounds. They show, moreover, that the earth in the lowest parts of the tube absorbs so much of the sulphur compounds from the oil that passes through them that the proportion of these compounds in the lower end of the tube indicates a concentration that is greater than that in the original petroleum. INVESTIGATION OF CALIFORNIA OIL II In working with the heavy California petroleum from which the fractions described above were obtained by diffusion at elevated temperatures, it was pointed out that such fractiona- tion was exceedingly difficult to effect, and at an enormous loss of petroleum. Owing to these results a lighter oil from the same locality was secured. This lighter oil was the one described as the California Oil II (see p. 8). Its properties were somewhat different from the oil first studied. The prin- cipal physical differences noted were viscosity and specific gravity, the latter being o. 889 in contrast to 0.912, that of the California Oil I. The viscosity was taken at various tempera- tures with the following results: Table VI Viscosity Measurements of California Oil II (4.5 cc. Run Through) Time Temperature of oil Minutes 2O 24.0 3 I7.I 40 12.2 5 9-1 It was decided to work with this oil at room temperature, owing to the difficulties arising from the fractionation of large quantities of oil at elevated temperatures. Attention was 15 then turned toward the problem of finding the length of tube most desirable for fractionating an oil of specific gravity 0.889 at 20. It was necessary, in addition, to ascertain the proper amount of pressure that should be brought to bear to bring about diffu- sion. For this purpose glass tubes of 3 . 2 cm. in diameter and varying in length from 30 to 150 cm. were employed. The object in using glass tubes was to enable the observer to see to what extent fractionation was taking place as the amount of pressure on the tubes was changed. It was soon found that a fractionation of the petroleum could be produced work- ing at room temperatures. While the results obtained were not as satisfactory as those obtained with light Pennsylvania oil of specific gravity 0.8470, they demonstrated the possi- bility of fractionating an oil of specific gravity 0.889 by dif- fusion through fuller's earth at ordinary temperatures. The yield of available fractions, however, amounted to only twenty per cent, of the oil used, while in the case of the lighter Penn- sylvania oil, it was 45 to 50 per cent, of the petroleum used. The length of tube that gave the most satisfactory results was found to be about 90 cm. The degree of compactness of the earth that gave best results was obtained by tapping the sides of the tube, and refraining from ramming, as all previous packing had been done. A column of earth a foot long was in- troduced into the tube, and it was tapped lightly on its side until the top of the column presented a firm surface. The method employed in extracting the oil from the earth into which it diffused consisted in taking measured sections from the earth 1 after it had been carefully emptied into a cyl- inder that was split longitudinally, and adding water to each section. The water formed a thick emulsion with the earth and expelled a part of the oil, retaining, however, an oil that was heavier than that expelled. It was noticed that the lighter oils were displaced in much greater proportion than were the heavier oils, for the earth showed a tendency to re- main in mechanical combination with the oil, holding almost 1 The section of earth from the uppermost part of the tube was called A, and the oil extracted therefrom called 2-A. The next lower section was called B, and its oil 2-B, etc. i6 100 per cent, if the specific gravity approached 0.9. When an examination of the oil retained by the earth was desired, the oil was extracted from the dried earth by ether. By the ether extraction an oil of different physical and chemical properties was secured. This proved that the fractions ob- tained by displacement of the oil by water did not represent the actual fractions formed during diffusion through the earth, but only a small part of such fractions. The method of securing fractions of oil by displacement by water was then dispensed with and another method intro- duced. This consisted in allowing the oil to diffuse to the top of the tube and then letting it overflow into small tubes of about 70 cc. capacity. The oil that first overflowed and col- lected in the upper reservoir constituted the first fraction. The first fraction was usually very light in color and specific gravity. As the diminished pressure continued to be applied, a heavier oil began to flow into the upper reservoir. When a change in color was noticeable, the reservoir was emptied or exchanged for another, and the pressure was temporarily cut off from the tube by means of pinchcocks while the change was being made. The second fraction was then collected until a change in color was noticed, and so on with a third, until an oil that was colored brown by the bitumen of the petroleum drained into the upper reservoir. The reduced pressure was then cut off, the tubes cleaned and refilled with fresh fuller's earth, and the process repeated. By repeated experiment, the amount of pressure that gave the best results was determined and regulated as follows: the tubes were allowed to stand in the reservoirs of petroleum for abour 24 hours, without any reduced pressure being ap- plied. Then, for a period of about three days, the pressure was gradually reduced until the manometer registered 650 mm. of mercury. After being maintained at this height for several days, the pressure was then lowered to 500 mm. and kept there until the first fractions overflowed into the upper reservoirs. When this pressure failed to draw up a slow, steady stream of oil into the reservoirs, it was still further re- 17 duced, usually being made as low as 200 mm., by the time that the last available fraction had been collected. By this method of operating, the California Oil II was found to be capable of fractionation into five distinct fractions. The lightest, termed Fraction 2-A, was of specific gravity 0.8264; the heaviest, 2-E, 0.8737, showing a wide range in density (see p. 18)- The basis upon which the various fractions were differ- entiated was their color, as it was observed that the depth of color was an approximate measure of the specific gravity of the oil. The fractions 2-A to 2-E were examined for nitrogen, and all of them showed signs of its presence. Careful anal- yses of these were then made, and similar analyses of the earth in various parts of the tube were made. Fractions 2-A and 2-B were then subjected to a second fractionation, and the results also studied with regard to their nitrogen content. Determination of Nitrogen Compounds in California Oil II Nitrogen compounds were found to exist in the California Petroleum II to the extent of nearly 0.8 per cent. Accord- ingly, this oil was subjected to fractionation by the improved method that was adopted by the investigator (see p. 16). This was necessary since the oil was too heavy to be worked by the method of Cram and Bransky unless elevated tempera- tures were resorted to in order to reduce the viscosity. Be- sides, it was feared that high temperatures would bring about the loss of the nitrogen compounds by volatilization. The method for determining the nitrogen was that known as the Gunning-Arnold-Dyer modification of the Kjeldahl method. It is described in detail in Sherman's Organic Analy- sis (pp. 291-4). It was found necessary to digest the light oils as long as 3 or 4 hours and the heavy ones as long as 8 hours before the contents of the digestion flask became color- less. Another necessary precaution had to be observed in applying heat very gradually to the flask at the beginning of the digestion, otherwise the nitrogenous material distilled off, i8 giving results that were far below those obtained when greater precaution was exercised. Table VII Nitrogen Determinations Fraction Sp. gr. Per cent, nitrogen 2-A t>.8264 0.08 2-B 0.8421 0.116 2-C 0.852 0.289 2-D 0.8614 0-315 2-B 0.8737 0.332 Crude oil 0.889 0.761 These results show that the proportion of nitrogen com- pounds is reduced in the lightest fraction to nearly one per cent, of its total amount. This proportion of nitrogen was still further reduced by the second fractionation of the two lightest fractions. The fractions obtained from 2-A are desig- nated by 2-A i, 2-A 2, etc. Those from 2-B are designated by 2-B i, 2-B 2, etc. Table VIII Second Fractionation of California Oil II Fraction Sp. gr. Per cent, nitrogen 2-A i 0.8117 Trace 2-A 2 0.8 1 86 Trace 2-A 3 0.8193 0.03 2-A 4 o . 8240 o . 06 2-B i 0.8205 Trace 2-B 2 0.8386 0.045 2-B 3 0.8414 0.09 2-B 4 0.8421 o. 109 From the above results it is apparent that the fuller's earth is particularly potent in selectively absorbing nitrogen com- pounds from the oil. In view of the fact that but two filtra- tions through the earth succeeded in reducing the amount of nitrogen present to such a small proportion as compared to that in the petroleum, it is probable that a third filtration would have reduced it to zero. A third fractionation would have been made, but lack of workable material rendered this impossible. The analysis of the oil retained by the earth taken from the 19 upper, middle and lower ends of the tubes showed that the part in the lowest sections of the tube had absorbed the great- est amount of the nitrogen compounds from the petroleum. The oil was extracted from sections of earth, taken at differ- ent levels from the tubes, by ether. The ether was evaporated off in an electrical drying oven at 50. Table IX Analysis of Oil Retained by Fuller's Earth Section of tube Sp. gr. Per cent, nitrogen Upper end o . 84 1 3 o . 205 Middle section 0.8655 -43 Lower end 0.9172 0.94 The concentration of the nitrogen compounds in the lower end of the tubes is apparent. It is noticeable that the per- centage of nitrogen here is slightly greater than it was in the original petroleum. This was to be expected, inasmuch as the earth in the lower end of the tube came into contact with the greatest amount of oil and could thus selectively absorb more of the nitrogen compounds than fcarth in the upper end. Moreover, earth in the upper end came into contact with oil that had already had a portion of its nitrogen compounds re- moved and could not, for that reason, extract as much nitro- genous material from it as it could from oil that was much richer in nitrogen compounds. The study of the benzene and olefin hydrocarbons in Cali- fornia petroleum was next contemplated. Before taking this up, it was considered advisable to study with some degree of accuracy the behavior of known mixtures of benzene and pure paraffin oilwhere such mixtures were allowed to diffuse through fuller's earth. This problem had been investigated before by earlier workers in this field, and certain conclusions arrived at, but a further study of the same, working under somewhat different conditions, was now resumed. FRACTIONATION OF MIXTURES OF BENZENE AND PARAFFIN OIL In previous work by Gilpin, Cram and Bransky on the frac- tionation of crude petroleum, it was noted that there was a tendency for the pure paraffin hydrocarbons to collect in 20 the upper section of the tube through which the petroleum was allowed to diffuse. To investigate this more closely, Gilpin and Bransky studied the behavior of mixtures of ben- zene and paraffin oil, such mixtures being of definitely known composition, and plotted their results in curves that showed the relative amounts of benzene and paraffin oil that collected in all parts of the tube. The curves show, moreover, that the proportion of benzene decreases gradually as one passes from Grade F (the oil from the lowest section of the tubes) to Grade C (the oil from the third section of the tubes), Grade A being considered that fraction from the uppermost part of the tubes. Above Grade C, there is a sharp decrease in the proportion of benzene, the same dropping far below the propor- tion in the original mixture that was put into the reservoir. It was also noted that the curves representing the specific gravities of the various sections were about parallel with those representing the percentages of benzene. This was to be ex- pected, since the benzene was of considerably higher specific gravity than the paraffin oil. The tubes employed measured five feet six inches in length. With the view to determine more accurately the exact loca- tion of the break in the curve, and to ascertain if it was a func- tion of the length of the tube, investigation along this line was begun by the authors. In order to study the latter prob- lem, it was decided to use shorter tubes, those of two feet nine inches in length being chosen. Correspondingly smaller sections of earth were taken, and the oil was displaced from them by the addition of water. The benzene used was of specific gravity o . 879. The method of analysis used to determine the proportion of benzene and paraffin oil in each fraction consisted in shaking 10 cc. of the oil with three or four times its volume of concentrated sul- phuric acid, until all of the benzene had been sulphonated. Then the shaken material was poured into a burette and al- lowed to stand until all the paraffin oil mechanically held in combination with the acid had separated out. To determine the length of time necessary for shaking in order to remove all the benzene, 10 cc. of benzene were mixed 21 with an equal quantity of pure paraffin oil and the mixture shaken with four times its volume of concentrated sulphuric acid in a machine that agitated the mixture about 450 times per minute. The amounts of benzene that were absorbed after definite periods of time were as follows: Table X A ction of Sulphuric A cid on Benzene and Paraffin Oil Time of shaking . Per cent, benzene Minutes absorbed 15 52.1 30 79.0 45 9i-4 60 100. o All samples for analysis were, therefore, shaken for more than an hour, until further shaking failed to reduce the volume of the oil. The paraffin oil used was a light, pale yellow oil of very disagreeable odor. Its specific gravity was 0.7895 at 20. After purification by the method described below, the specific gravity became 0.7775. The oil was purified as follows: A quantity was agitated with one-third its volume of concen- trated sulphuric acid for 6 hours in two 2 -liter bottles that re- volved slowly about an axis placed between the two. The acid became dark brown, and the evolution of sulphur dioxide indicated chemical action. The acid was then drawn off in a large separatory funnel and a fresh supply added to the oil. This was again agitated for six hours and separated off as be- fore. By this process the volume of oil decreased 9 . 3 per cent., and lost its disagreeable odor and became colorless. It was then shaken with a small quantity of dilute alkali until neutral, washed with water, and finally shaken for several hours with calcium chloride, and filtered. By repeated tests it was shown that the oil, after this treatment, did not de- crease in volume when shaken with sulphuric acid. Mixtures of benzene and pure paraffin oil were then allowed to diffuse through fuller's earth. The earth in the tubes had been as tightly packed as it was possible to secure by ramming the earth with a rod tipped with a rubber stopper. The pro- 22 portions of benzene and paraffin oil in the mixtures were as follows : Name applied Per cent. Per cent. to mixture benzene paraffin oil Series No. 5 2O 80 Series No. 6 33 6 7 Series No. 7 50 50 Series No. 8 75 25 Series No. 9 20 80 Series No. 10 33 6 7 Series No. n 50 50 Series No. 12 20 80 Series No. 13 75 25 Series No. 14 20 80 Sections of varying lengths, as shown in the tables accom- panying the curves (pp. 23 to 29), were taken, and the oil was displaced by addition of water (in Series No. 5 to No. 8, inclu- sive), by extraction with ether (Series No. 9), or by letting the oil overflow into upper reservoirs by the new method described on page 16 (Series No. 14). The sections of earth from which the oil was extracted were made considerably smaller than similar sections taken by Bransky in his work. The uppermost section, called A, varied in length from 25 to 12 cm. The fraction of oil that it yielded was called 5~A, 6-A, according to the series to which it belonged. The next lower section of earth, usually a little shorter than A, was called B. Its oil was termed 5-6, 6-B, or 7~B, according to the series to which it belonged. The short tube and short sections of earth from which the oil was ex- tracted were chosen so as to locate more accurately the point at which a sharp decrease in the amount of benzene occurred. This point of sudden change was found to be invariably loca- ted from 20 to 40 cm. from the top. The specific gravity of each fraction of the oil was taken for the first five tubes. Since this in every case was a function of the proportion of benzene and paraffin oil present, it was discontinued after the first five series were run. Series No. 5 consisted of the following: A tube was filled with earth and the reservoir below filled with 500 cc. of a mix- ture of 20 per cent, of benzene of specific gravity 0.879 and 80 per cent, of paraffin oil of specific gravity 0.7775. When diffusion had taken place, the fractions were analyzed as before described, and it was found that the break in the curve oc- curred at about 35 cm. from the top. The specific gravity of that fraction that contained least benzene was the lowest. It was noticeable that the proportion of benzene to paraffin oil was nearly a constant, until the point B was reached (see Fig. I). Table XI Series No. 5 Benzene, 20 per cent. Paraffin Oil, 80 per cent. Rose to a height of 88 cm. Per cent. Per cent. Fraction Cm. Cc. Sp. gr. benzene paraffin 5-A 20 46 0.787 10.4 89.6 5-B 15 44 0.790 19.1 80.9 5-C 15 43 o-79i 21.3 78.7 5-D 15 36 0.792 21.0 79.0 5-B 12 30 0.792 21.3 78.7 5-F ii 39 o. 795 23.3 76.7 A - A \ \ B < \ \ B - fc 1 O 3 C " \ I D , \ D \ * V / E- E~ F- F - i JO 20 3O 4O $0 6O JO 20 3O 4O 50 60 Per cent. Benzene Per cent. Benzene Fig. I. Series No. 5 Fig. II. Series No. 6 In Series No. 6 there was not noticed as marked uniformity in specific gravity or in the proportion of benzene to paraffin oil as in Series No. 5. The break in the curve occurred about 25 cm. from the top (see Fig. II) . Table XII Series No. 6 Benzene, 33 per cent. Paraffin Oil, 67 per cent. Rose to a height of 95 cm. Fractions 6-A 6-B 6-C 6-D 6-B 6-F Series No. 7, with equal quantities of benzene and paraffin oil, gave a curve that broke sharply at Section B, 16 cm. from the top. Up to this point the amounts of the two oils remained nearly constant (see Fig. III). A '" Cm. Cc. Sp. gr. Per cent, benzene Per cent, paraffin 25 31 0.725 18.4 8l.6 15 38 0.798 25-7 74-3 15 42 0.798 25-3 74-7 12 38 0.799 25-9 74.1 12 35 0-7995 27-3 72.7 12 45 0.799 27.0 73-o c r E F 20 30 40 50 Per cent. Benzene Fig. III. Series No. 7 60 IO 20 3O 40 SO Per cent. Benzene Fig. IV. Series No. 8 60 70 25 Table Xlll Series No. 7 Benzene, 50 per cent. Paraffin Oil, 50 per cent. Rose to a height of 90 cm* Per cent. Per cent. Fractions Cm. Cc. Sp. gr. benzene paraffin 7-A 16 21 0.8023 30-5 69.5 7-B 16 34 0.8155 47.0 53.0 7-C 16 44 0.8160 45.8 54.2 7-D 1 6 50 0.8165 45.7 54.3 7-B 12 48 0.8165 45-7 54-3 7-F 12 53 0.817 47.2 52.8 In Series No. 8 the break was not such a sharp one, and oc- curred about 30 cm. from the top (see Fig. IV). Table XIV Series No. 8 Benzene, 75 per cent. Paraffin Oil, 25 per cent. Rose to a height of 84 cm. Per cent. Per cent. Fractions Cm. Cc. Sp. gr. benzene paraffin 8-A 15 25 0.832 58.1 31.9 8-B 15 37 0.833 64.5 35.5 8-C 12 38 0.8385 69.0 31.0 8-D 12 44 0.839 69.8 30.2 8-E 12 43 0.839 69.7 30.3 8-F 12 51 0.842 70.1 29.9 In order to determine whether the proportion of benzene to paraffin oil in the fractions was affected by the displacement of the oil by water, the fractions secured from Series No. 9 were extracted with ether. The results plotted in the curve on page 26 show that the water plays no part whatever in the action. The break in the curve, showing a sudden sharp de- crease in the proportion of benzene to paraffin oil, occurred in about the same locality. Table XV Series No. 9 Benzene, 20 per cent. Paraffin Oil, 80 per cent. Rose to a height of 70 cm. Per cent. Per cent. Fractions Cm. Cc. benzene paraffin 9- A 12 44 8.9 91 . i 9-B 12 34 18.4 81.6 9~C 12 38 20.0 80.0 9-D 10 37 21.7 79.3 9-E 10 39 21.3 79.3 9-F 10 44 22.1 78.9 26 In Series No. 10 and No. n (Figs. VI and VII) the indi- vidual fractions were analyzed immediately after the displace- ment of the oil in the earth by water. This was done so as to avoid any possible loss of either oil by evaporation. The same general characteristics are apparent in the curves that express the results of the fractionation. r E F 10 20 30 40 Per cent. Benzene Fig. V. Series No. 9 60 IO 20 30 4O $O 60 Per cent. Benzene Fig. VI. Series No. 10 Table XVI Series No. 10 Benzene, 33 per cent. Paraffin Oil, 67 per cent. Rose to a height of 81 cm. Fractions Cm. Cc. Per cent, benzene Per cent, paraffin IO-A 15 36 20.1 79-9 IO-B 15 53 26.2 73-68 IO-C 12 45 28.0 72.0 io-D 12 52 27.7 72.3 lo-E 12 42 28.1 71.9 io-F 10 37 30-4 69.6 Table XVII Series No. n Benzene, 50 per cent. Paraffin Oil, 50 per cent. Rose to a height of 74 cm. Per cent. Per cent. Fractions Cm. Cc. benzene paraffin ii-A 15 35 32.1 67.9 n-B 12 42 47.1 52.9 n-C 12 37 47.1 n-D ip 41 46.9 n-E 10 36 50.4 ii-F 10 35 50.9 B - C - E F \ IO 20 3O 40 SO 6O Per cent. Benzene Fig. VII. Series No. 11 10 52.9 55-1 49.6 49.1 20 3O 40 Per cent. Benzene Fig. VIII. Series No. 12 50 60 In all tubes up to Series No. 12 the oil was drawn up to a height under two feet nine inches. Although tubes five feet nine inches long were used, the quantity of oil placed in the reservoir was such as allowed only the lower half of the earth in the tube to become impregnated. In order to avoid the possibility of the more volatile oil evaporating into the dry earth above it, tubes of 2 feet 9 inches in length were used for Series No. 12 and No. 13. The curves representing the results from 28 these series (pp. 27 and 29) show that this precaution failed to produce any noticeable difference in the proportion of ben- zene and paraffin oil or in the characteristic behavior of the fractions that were obtained by diffusion through fuller's earth. Table XV III Series No. 12 Benzene, 20 per cent. Paraffin Oil, 80 per cent. Rose to a height of 80 cm. Per cent. Per cent. Fractions Cm. Cc. benzene paraffin I2-A 15 31 9.5 90.5 I2-B 15 40 16.8 83.3 I2-C 12 54 18.6 81.4 I2-D 12 52 18.8 81.2 I2-E 12 56 18.0 82.0 I2-F 12 53 21. i 79.9 Table -XIX Series No. 13 Benzene, 75 per cent. Paraffin Oil, 25 per cent. Rose to a height of 80 . 5 cm. Per cent. Per cent. Fractions Cm. Cc. benzene paraffin I3-A 15 30 60.0 40.0 I3-B 15 41 69.1 31.9 I3-C 12 37 72.8 27.2 1 3-D 12 41 71.3 28.7 I3~B 12 47 74.0 26.0 I3-F 12 48 74.9 25.1 Series No. 14 was set up to test out the improved method of fractionating by means of the earth. Eight hundred cc. of a mixture of 20 per cent, benzene and 80 per cent, paraffin oil were drawn up through tightly packed fuller's earth, and six fractions of 50 cc. each were collected in the upper reser- voir. The first of these was designated i4~A, and the last i4~F. The six fractions were separately analyzed and the results plotted in a curve (see p. 29). 2 9 2nd } A B > I jc t, * D 1 \ 5th i5 E ' , F 6th \ IO 20 JO 40 50 60 70 10 20 30 40 50 60 Per cent. Benzene Per cent. Benzene Fig. IX. Series No. 13 Fig. X. Series No. 14 Table XX Series No. 14 Benzene, 20 per cent. Paraffin Oil, 80 per cent. Per cent. Per cent. Fractions Cc. benzene paraffin I4-A 50 10.9 t 8 9 .I I4-B 50 17.1 82.9 I4-C 50 19.4 80.6 I4-D 50 21.6 78.4 I4-E 50 23.0 77-0 I4-F 50 20.0 80.0 Extract of Earth Left in Tube Per cent. Per cent. Fractions Cm. Cc. benzene paraffin Upper half 70 215 15.9 84.1 Lower half 70 225 13.8 86.2 It was found that the proportion of benzene in the first fraction was the lowest, and that it gradually increased in the subsequent fractions. The curve above shows that the increase in benzene in the successive fractions is more 30 gradual than when the oil was obtained by being expelled by water. There was, however, a marked increase in the amount of the benzene after the first 50 cc. had been drawn off. This corresponds to the increase indicated by the curves in every series that was run, and thus it appears that the new method of working gives the same results as the other methods of manipulation. These results show that the degree of fractionation is not a question of the absolute height of the earth through which the oil passes, but of the relative height. The advantage, however, of using long tubes is that more material can be obtained and a greater number of fractions between the two extremes are possible. FRACTIONATION OF CALIFORNIA OIL III With a view to studying in some detail the fractionation of a heavy California petroleum, a tank of this material was se- cured from Kern County. It had a specific gravity of 0.9118 at 20, and resembled in physical properties the California Oil I. A description of this oil, termed California Oil III, is found on page 6. In order to find the best conditions for se- curing large workable fractions of this oil, preliminary work was carried on by means of glass tubes so as to enable the in- vestigator to observe the progress of the fractionation. The glass tubes measured 1.25 inches in internal diameter, and varied in length from two to five feet. The object of this was to find the length of tube which would give a maximum yield of workable fractions. The tubes were filled with a fine-grained fuller's earth known as XXF clay. The method of packing these by ramming with a rod tipped with a rubber stopper was abandoned since this gave varying degrees of hardness, while strict uniformity was desired. It also failed to remove the cushions of air that persisted in remaining between layers of the earth, and these,, it had been found in earlier work, were a grave source of trouble. Instead, the earth was run into the tubes until they were full. Then the tubes were tapped on their sides through- out their length until the earth failed to subside. More earth was added to fill in the space left vacant by the earth that had settled, and they were tapped again until further subsidence ceased. With tubes packed as indicated above, the fractionation of the oil by suction was commenced. It was soon apparent that the amount of fractionation by using the fine-grained earth would be exceedingly small, and that it would possibly require four or five weeks' suction to bring the oil to the top of the longer tubes, so the effect of using a coarse-grained earth was suggested. Accordingly, fuller's earth of the size 30 to 60 mesh was secured. The tubes were packed uniformly by the method above described, and placed in the reservoirs. The pressure was reduced to 650 mm. of mercury. Even at this slight reduction in pressure, the oil was drawn up through the tubes, the products showing, however, no signs of frac- tionation. Accordingly, after repeated experiments, the tubes were allowed to stand 48 hours with no suction applied, and then the pressure on the top of the tubes was reduced to 730 mm. With this, the oil started to rise slowly and steadily through the earth, the uppermost part showing signs of marked fractionation. When the oil had risen to the height of about 1 8 inches, the pressure was reduced to 700 mm. ; at a height of 3 feet, it was reduced to 650 mm. ; and above 4 . 5 feet it was maintained at 600 mm. until all of the available oil was drawn over into the upper reservoirs. By use of the coarse-grained fuller's earth the time required for fractionation of a series of tubes was found to be from ten to twelve days. The frac- tions of oil that were drawn up and collected were classified on the basis of color. With this as a criterion, six distinctly different fractions were obtained, and these showed a wide range in color and specific gravity. The total amount of these six fractions was, however, only 15 per cent, of the oil put into the lower reservoir, there being a loss of 85 per cent, due to the speedy darkening of the earth by the bitumen present in the petroleum. Results, with a brief description of the fractions obtained, are tabulated be- low: 32 Table XXI Preliminary Fractionation of California Oil 111 Fraction Sp. gr. Description 1 0.8364 Nearly colorless; pale green fluor- escence 2 0.8449 Pale yellow; pale green fluor- escence 3 0.8609 Yellow; quite fluorescent 4 0.8701 Light brown; strong fluorescence 5 0.8770 Brown; strongly fluorescent 6 0.8866 Dark brown; deep green fluor- escence These were the first results that were ever obtained with an oil of as high specific gravity and viscosity as this one, for all investigations carried on at room temperature with an oil of this high density had failed thus far to produce any results. When these six fractions were next examined as to their chem- ical properties, it was found that the diffusion through the earth had not only lowered the viscosity of the oil, removed the bitumen, and thereby greatly decreased the specific grav- ity of the oil, but it had also absorbed from the petroleum a large proportion of benzene and olefin hydrocarbons. The amounts of the two last-named ingredients were ascer- tained by treatment of the oil with concentrated sulphuric acid. This did not determine the benzene and olefin hydro- carbons separately. It is fair to assume that the concentrated acid acted upon the other materials in the oil, but that it removed all the benzene and olefin hydrocarbons was conclu- sively proved by a method described later on. The method by which the benzenes plus the olefins were de- termined was as follows: Ten cc. of each fraction were care- fully measured from a burette. The light oils were shaken with three times their volume of concentrated sulphuric acid until no further diminution in the volume of the oil occurred. The shaking was accomplished in a machine that vibrated 400 times per minute. The bottles containing the oil thus treated were emptied into burettes, rinsed with a few cc. of acid and allowed to drain overnight. The oil not acted upon rose in this time above the acid, and could be read off and directly translated into percentages of paraffin hydrocarbons present. 33 This last statement is based upon the fact that paraffin hydro- carbons are not acted upon by cold, concentrated sulphuric acid, while the benzene and olefin hydrocarbons react with the acid to form sulphonic and alkylsulphuric acids, respect- ively. That the benzene and olefin hydrocarbons were en- tirely removed by two hours' agitation with a large excess of sulphuric acid was proved by the absence of these com- pounds in the oil after it had been acted upon. The test for benzene was the action of nitric acid to give nitro derivatives and the subsequent reduction to give amino compounds. The tests for olefin were the direct addition of bromine and the action of alkaline permanganate solution. The treatment of the heavier oils with sulphuric acid was somewhat different from that of the light oils, in that in the case of the former the action of the acid produced a mixture so dark and viscous that the line of demarcation between the acid and the unabsorbed oil was invisible. Therefore, the sam- ple of heavy oil was treated in one of the following ways: I. It was mixed with twice its volume of pure paraffin oil that had been treated previously with concentrated acid until none of it was absorbed by further action of the acid upon it (see p. 21). The diluted oil was then shaken for five hours or more with three times its volume of acid until further diminution in volume of the oil ceased; 1 or II. The sample was shaken with thirty cc. of sulphuric acid for two hours, and the mixture was thinned out by diluting with twenty cc. of the paraffin oil, shaken for a few minutes, drained into the burette, and the amount of benzenes and olefins present determined by difference between thirty cc. and the amount of oil unabsorbed by the acid. The methods of determining the sulphuric acid absorption gave results that were concordant to within 0.5 per cent. The accuracy of the analysis by this method was within one per cent., as was proved by analysis of known mixtures. The percentages of benzene and olefin hydrocarbons that 1 For diluting the 10 cc. sample for analysis and shaking with 90 cc. of sulphuric acid, it was necessary to have a burette of over 120 cc. capacity. This was secured by blowing a bulb of 80 cc. at the lower end of a 50 cc. burette, leaving a volume of about 40 cc. above by which to read the amount of oil that was not acted upon. 34 were found in the various fractions of the oil (see p. 32) were found to vary from 5. 15 per cent, to 27.7 per cent., as shown in the following table : Table XXII Preliminary Determination of Benzenes and Ole- fins in California Oil III Per cent, benzenes Fractions and olefins 1 5-15 2 10-4 3 15-2 4 16.8 5 20.6 6 27.7 It is thus seen that the earth through which the oil passed exerted an absorptive effect upon the benzene and olefin hydrocarbons. This effect has been termed selective absorp- tion or adsorption, by which is meant that the earth exerts an action upon the complex oil by which it retains an apprecia- ble quantity of certain of its ingredients. It is not a filtration effect, for when the petroleum is drawn quickly through coarse or fine fuller's earth by means of low pressure, it filters through, depositing any solid matter that may be suspended, but being otherwise unaltered. If the action of the earth is explained as a phenomenon of adsorption, the statement might be made that the separa- tion of the bituminous material from the petroleum was by the coagulation and adhesion of the bitumen to the very ex- tensive internal surface that the grains of fuller's earth pos- sess. For it is a well-known fact that porous media like char- coal, dried clays, colloidal and finely divided metals, platinum sponge, etc., possess an enormous amount of surface energy, due to the forces that are active at their extensive surface, and that such substances show the phenomenon of adsorption to a marked degree. The separation of the bitumen, carry- ing with it the benzene hydrocarbons, the olefins, the sulphur and nitrogen compounds, may thus be regarded as a special case of adsorption. 35 If the bitumen is considered as existing in the colloidal con- dition, the effect of the internal surface of the fuller's earth could be explained as bringing about the coagulation of the colloidal bitumen into discrete particles which would carry with them all materials in the oil, save the paraffin hydrocar- bons. That all the above-mentioned materials are held be- hind by the fuller's earth has been conclusively established by direct and indirect proof. In retaining the bitumen with benzenes, olefins, nitrogen and sulphur compounds and per- mitting the paraffin oils to diffuse through, the fuller's earth acts as a dialyzer, proving more or less impervious to the substances held in solution in the paraffin oils, but not so to the solvent itself. Fractionation of California Oil III For the more accurate study of the physical and chemical properties of the fractions obtained from this oil, large quan- tities of these fractions were necessary. Accordingly, a slight alteration was made in the apparatus employed, so as to be able to handle larger quantities. The tubes originally used measured 1.25 inches by 5.5 feet. It was now decided to test out tubes of greater diameter and length, and the size finally adopted was 2.75 inches in diameter by six feet in length. In order to be able to observe the behavior of the oil, a glass tube of the same size was joined in parallel with the tin tubes. With tubes of this diameter it was found that the best results were obtained by using very little suc- tion, and by extending the suction over a period of about two weeks, as the yield of available fractions was found to increase through this method of working. This was ascer- tained by measuring the amounts of each of the fractions that were obtained from a single tube placed in a measured amount of petroleum. From this tube 14 distinct fractions were secured, the basis of distinction being the color of each. The amounts of each fraction and a brief description of the same are as follows: 36 Table XXIII Preliminary Fractionation of California OH 111 Amount Fraction Cc. Description 1 35 Colorless, pale blue fluorescence 2 38 Colorless, pale green fluorescence 3 40 Pale yellow, pale green fluorescence 4 40 Yellow, pale green fluorescence 5 48 6 45 Yellow to deep orange-brown, increasing 7 47 green fluorescence as fractions increased 8 60 9 65 10 63 11 75 12 85 13 88 14 no in sp. gr. and viscosity Light to dark brown in transmitted light; fluorescence less pronounced, but still very noticeable The total amount of available fractions was 839 cc. The amount of petroleum used was 3500 cc. The yield was 23.7 per cent. It is seen from the above figures that the yield of heavier fractions is increasingly greater than that of the lighter oils. It is also to be noted that this slower method of working in- creased the total yield of available fractions from 15 per cent* to nearly 24 per cent. Fractions that were colored darker than the fourteenth were discarded as being contaminated with too much bitumen. The fractionation of considerable quantities of California Oil III was then undertaken, and differentiation was made be- tween the various fractions on the basis of specific gravity, a hydrometer small enough to fit into the upper reservoirs being used 1 to indicate the specific gravity approximately. The boiling points of the various fractions were taken at atmospheric pressure. Every fraction was found to be a mix- ture of oils, for not any of the boiling points remained con- stant, but, instead, rose through a range of n to 35 degrees. In determining the boiling points, 5 cc. of each fraction were taken. In order to get the boiling point of the greatest part 1 The upper reservoirs were glass tubes 2.5 cm. X 16.5 cm. closed at one end. Two-hole rubber stoppers 6tted in the tops which were flanged slightly to give air- tight connections. 37 of each, i. e., of that part which represented the average of the constituents, one cc. was distilled off, and the boiling point noted during the distillation of the next three cc. The boiling point of the last cc. was not taken. The distillates ranged from a colorless, thin oil with a smoky smell to a thick, dark brown oil of a very disagreeable odor resembling garlic. The fractions from the California Oil III were termed 3~A, 3 B, etc. Those of a second frac donation were termed 3 A i, 3- A 2, etc. The range of the fractions obtained was as fol- lows: Table XXIVFractionation of California Oil III Fraction Sp. gr. Boiling point 1 3-A 0.8325 i6o-i95 2 3-B 0.8347 I72-20I 3 3-C 0.8372 i86-2i 9 4 3-D 0.8462 2io-23i 5 3-E 0.8524 2 3 5-26o 6 3-F 0.8551 2 4 7-269 7 3-G 0.8680 256-28o 8 3-H 0.8781 268-289 9 3-1 0.8840 275-3io 10 3-J 0.8885 28 4 -3i7 11 3-K 0.8903 299-326 12 3-L 0.895 3ii-328 13 3-M 0.8972 3i7-334 14 3-N 0.8984 32 9 -34o Crude petroleum 0.9118 105 ~34O These^results show the very wide range in specific gravity between the first and last fractions of the petroleum under in- vestigation. For comparison, the range of fractions obtained from a lighter Pennsylvania petroleum by Gilpin and Bransky are given in the following table: Fractionation of Pennsylvania Petroleum Fraction Sp. gr. Fraction Sp. gr. A-i 0.8250 D-2 0.8495 A-2 0.8287 D-3 0.8515 B-i 0.8367 D-4 0.8555 B-2 0.8392 E-i 0.8527 C-i 0.8413 E-2 0.8540 C-2 0.8460 E-3 0.8570 C-3 0.8488 D-i o . 8470 OQ 5000000 Specific Gravity Fig. XI. California Oil III JO 20 30 40 50 Per cent. Benzenes + Ole fines Fig. XII. Sulphuric Acid Absorption of Cali- fornia Oil III The range obtained by Bransky's method of working shows specific gravities varying from 0.8250 to 0.8570, as contrasted with the range of 0.8325 to 0.8984 obtained by the present investigator upon the heavy California oil. Sulphuric Acid Absorption of the California Oil 111 The sulphuric acid absorption was determined for the four- teen fractions. The method used was that described on pages 32 to 33. The results are tabulated with respect to the percentage of paraffin hydrocarbons in contrast to that of the benzene and olefin hydrocarbons taken together. These figures show the great extent to which the diffusion through fuller's earth removes the benzene and olefin hydro- carbons. The first fraction consisted of nearly pure paraffin oils, reaching a degree of purity of 96 per cent., while, by contrast, the crude petroleum contained about 50 per cent, of paraffins. 39 '. Table XXV Sulphuric Acid Absorption Per cent, benzenes Per cent. Fraction and olefins paraffins 3-A 3-7 96.3 3-B 4.14 95.86 3-C 5-i 94-9 3-D 7.44 92.56 3-E 10.13 89.87 3-F 13.06 86.94 3-G 15.2 84.8 3-H 15.8 84.2 3-! 19-89 80. i i 3-J 20.6 79.4 3-K 23.47 76.53 3-L 27.9 72.1 3~M 31-45 68.55 3-N 32.72 67.28 Crude petroleum 49 . 7 50 . 3 Bromine Absorption of California Oil HI In order to determine the amount of unsaturated or olefin hydrocarbons in the fourteen fractions of this oil, the quan- tity of bromine absorbed at room temperature (i9-23) in the dark by a weighed amount of the oil was determined. The method employed for the determination of the olefin content of the fractions was as follows: A weighed sample of the oil to be analyzed, about 0.6 gram for each determina- tion, was dissolved in fifteen cc. of redistilled carbon tetra- chloride. The vessel into which the solution was introduced was a 250 cc. Erlenmeyer flask with a ground glass stopper that fitted accurately and was sunk some distance into the neck of the flask so as to leave a gutter between the neck and the stopper. When the absorption of bromine was taking place, the gutter was filled with a couple of cc. of potassium iodide. It effectually prevented the escape of bromine vapor. The bromine was introduced in the form of a solution in pure carbon tetrachloride. The solution was made practically decinormal by dissolving 3.3 cc. of redistilled bromine in a liter of solvent. Its exact strength was determined by titrating against a known volume of a standard sodium thiosulphate 4 o solution. The bromine solution kept best in the dark, but frequent tests of its strength were necessary. In determining the bromine absorption, a known quantity of the bromine solution was added to the solution of the oil in carbon tetrachloride. The amount added was more than twice that necessary to combine with the total amount of olefins present (as determined by a previous analysis). The flask was closed and shaken, the gutter filled with 2 cc. of a ten per cent, solution of potassium iodide, and the flask al- lowed to stand in the dark with occasional agitating for thirty minutes. Longer contact of the oil with the bromine gave substitution products as well as addition products, the former being indicated by the presence of hydrobromic acid. After thirty minutes, the flask was brought out of the dark, ten cc. of the potassium iodide solution were added, the flask closed and violently shaken, and the amount of iodine liberated by the excess of bromine present determined Jby titrating against the thiosulphate solution. Toward the end of the titration, a few cc. of a very dilute starch solution were added to indi- cate sharply the end point. Repeated shaking of the con- tents of the flask was necessary during the last part of the titra- tion in order to free the iodine from its solution in the carbon tetrachloride. A blank determination was made parallel with each analysis in order to ascertain how much of the thio- sulphate solution was exactly equivalent to the amount of bromine solution that was added. The results were trans- lated directly into the percentage of olefins present in the various fractions and in the crude petroleum. By the action of concentrated sulphuric acid on the oils, the percentage of benzenes and olefins together was ascer- tained. The action of bromine gave the percentage of ole- fins. In order to determine whether the difference between these results gave a value that represented the percentages of benzene alone, an investigation was carried out on the ist, 5th, Qth and i4th fractions of the California Oil III. After being shaken for several hours with sulphuric acid, they were tested for the presence of benzenes and olefins, and blank 41 results were obtained. Then fresh samples of these fractions were treated with an excess of bromine so as to brominate the olefins present, washed with water, dried and shaken with concentrated sulphuric acid for several hours. By this, an amount of the oil was absorbed which corresponded to the benzenes present. This amount, added to that secured by the action of bromine, gave the true percentages of benzenes and olefins together. The results, however, were in each case a little high, showing that the sulphuric acid must have acted upon the brominated oils, but the difference was not greater than i . 3 per cent. Since the sulphuric acid absorption method was accurate to within one per cent., it was concluded that the difference between the percentages of the oil absorbed by sulphuric acid and that acted upon by bromine could be taken to represent the amount of benzene hydrocarbons present. The results that justified this conclusion are here given: Table XXVI Per cent, benzenes and olefins by Per cent. Per cent. Sum of sulphuric acid ab- Fraction benzenes olefins columns 2 and 3 sorption 3-A 3-E 3-1 3-N In the table below, the percentages of the various hydro- carbons present in the oils investigated are given. In the first column are given the designations of the various fractions; in the second are given the mean values of the percentages of the olefins found by determining the bromine absorption (duplicate determinations made in every analysis) ; in the third are given the percentages of benzenes and olefins taken to- gether, and determined by. the sulphuric acid absorption method; in the fourth are given the percentages of benzenes de- termined by difference between the second and third columns. These results demonstrate the selective absorption of the fuller's earth in its action upon the olefin hydrocarbons. As is indicated above, the proportion of olefins in the crude oil is 1.09 2.79 3-88 3-7 5-35 5-05 10.4 10.13 8.72 12.46 21. 18 19.89 14.03 19-35 33-38 32.72 Table XXV 11 Proportion of Various Hydrocarbons in Frac- lions of California Oil 111 Per cent. Per cent, ben- Per cent. Fraction olefins zenes and olefins benzenes 3-A 2\79 3-7 0.8l 3-B 3-25 4.14 0.76 3-C 3-62 5-i 1.48 3-D 4.06 *-44 3-38 3-B 5-05 10.13 5.08 3-F 5.84 13.06 7.22 3-G 7-44 15.2 7.76 3-H 8-43 18.8 7-37 3-1 12.46 19.89 7-43 3-J 13-44 20.6 7.16 3-K 14.66 23-47 8.81 3-L 14.81 27.9 I3-I9 3-M 18.34 31-45 13.11 3-N 19-35 32.72 J 3-37 Crude petroleum 28 . 24 49-7 21.47 3 6 Q 12 15 18 21 Per cent. Olefines Fig. XIII. California Oil III 24 27 5 JO 15 Per cent. Benzenes Fig. XIV. California Oil III 43 28.24 per cent. In a single fractionation this is reduced to 2 . 79 per cent, in the first fraction obtained. The gradual in- crease in the proportion of olefins as the specific gravity of the fractions increases is parallel to the gradual increase, in successive fractions, of all the compounds in the petroleum thus far studied, i. e., of the sulphur compounds, nitrogen compounds, benzene hydrocarbons and olefin hydrocarbons. The results, when plotted in the form of curves, show that there is a noticeable parallelism in selective action of the fuller's earth upon the compounds above mentioned. The curves are found on pages 38 and 42. In order to determine to what extent the olefins could be removed by further filtration through fuller's earth, a liter of Fraction 3-! was ref ractionated by means of a tube of smaller diameter. The nine fractions that were obtained were anal- yzed for their olefin content. The results were as follows: Table XXV 1 11 Second Fractionation of California Oil HI (Refractionation of 3-!) Fraction Sp. gr. Per cent, olefins 3-1 i 0.8661 8.55 3-1 2 0.8685 8.93 3-! 3 0.8740 10. 06 3-14 0.8751 10.82 3-15 0.8759 io-77 3-16 0.8773 11.40 3-! 7 0.8782 11.67 3-18 0.8801 11,79 3-19 0.8807 11-89 3-! 10 0.884 12.06 These results demonstrate that a further removal of bitu- minous material is possible by ref rac donating a heavy frac- tion, for there was a slight loss of color occasioned by the diffusion through the earth. Only a part of the olefins was removed, however, by this second fractionation, and the re- duction in viscosity was also very slight. Efforts were made to determine the effect of shaking frac- tions of oil with large amounts of fuller's earth. Accordingly several portions of Fraction 3~E were shaken with varied quan- 44 titles of fuller's earth for different periods of time. First, a quantity was shaken with three times its weight of earth for 30 hours and separated from the earth by suction. It was analyzed, with the following results: Table XXIX Fraction 3~E Shaken with Three Times Its Weight of Earth Time in hours Sp. gr. Per cent, olefins Color 20 0.8305 4.01 Yellow o 0.8524 5.05 Pale orange When shaken with one-half its weight of earth for periods ranging from 10 to 50 hours, the following results were ob- tained : Table XXX Fraction 3~E Shaken with One-half Its Weight of Earth Time in hours Sp. gr. Per cent, olefins Color o 0.8524 5.05 Pale orange 10 0.8524 5.08 Pale orange 20 o . 852 1 5 . oo Pale orange 30 o . 852 5 . oo Pale orange 50 0.8513 4-90 Same color, but slightly less fluor- escence These results show that the time that an oil is in contact with fuller's earth is of little or no importance as a factor in determining to what extent the earth absorbs certain in- gredients from it. The important factor is the amount of earth with which it comes in contact. These experiments substantiate, then, the deductions that are given on page 88 et seq.j i. e., that it is the amount of surface of the earth to which the oil is exposed that determines the extent of its ab- sorptive action. This is in line with the action of those substances whose surface energy is capable of affecting a colloid. That the bitumen in the petroleum investigated exists in this form was proved by the following: i. There was effected an actual separation of the petroleum into two distinct layers when an electromotive force of no 45 volts was impressed upon parts of the oil separated by an un- glazed porcelain septum. In order to make the oil conduct the current, one portion was shaken for several hours with one-fourth its volume of a 20 per cent, solution of hydro- chloric acid, and another with an equal amount of a 20 per cent, solution of potassium hydroxide. The emulsions formed in this way proved poor conductors, but were sufficiently good to enable the bitumen partially to precipitate out as a brown layer of a very viscous liquid, containing no mineral residue. Its form indicated that the bitumen was held in the original petroleum as a colloid that type that has been named by Oswald an emulsoid. 2. The high temperature coefficient of viscosity of Cali- fornia petroleum, as shown by earlier experiments (p. 9, Table I), is one of the most marked characteristics of emulsoids, i. e., that type of colloid solution in which the colloid is a liquid in a state of minute subdivision in a liquid medium. 3. The oil and bituminous material that were held back by the fuller's earth could not be removed by mechanical means. For, after the earth had been extracted with ether and carbon tetrachloride until no more could be extracted, it yielded a small quantity of oil upon distillation. This showed that the bituminous material that was in the oil had undergone a change of condition when it was adsorbed by the earth. SUMMARY 1. The diffusion of petroleum through fine-grained fuller's earth failed to effect the fractionation of the petroleum when the latter was of specific gravity as high as 0.912 at 20. Raising the temperature of such an oil to 75 made fractiona- tion possible. 2. The effect of such a diffusion of a petroleum containing compounds of sulphur is to separate out the light fractions of the oil containing smaller proportions of sulphur than are found in the original petroleum. 3. The effect of fractionating by means of diffusion through fuller's earth of a petroleum containing nitrogen compounds is to remove the nitrogen compounds from the oil that diffuses 4 6 upward through the earth, and to cause them to concentrate in the earth through which the oil has passed. 4. Mixtures of benzene and paraffin oil, when fractionated by capillary diffusion through fuller's earth, give fractions that have marked general characteristics, both chemical and physical, based on the proportions of benzene and paraffin oil in each. 5. The f ractionation of a petroleum rich in benzene and olefin hydrocarbons by the diffusion through fuller's earth gives fractions in which the proportions of benzene and olefin hydro- carbons increase regularly with the increase in specific gravity of the successive fractions. 6. An explanation of the above phenomena was found in the conception of the petroleum as an emulsoid, and in the action of the fuller's earth as a dialyzing septum, permitting the free passage of the paraffin oils, and causing by its ex- tensive surface the adsorption and coagulation of the bituminous material, carrying with it the sulphur and nitrogen com- pounds and the benzene and olefin hydrocarbons. BIOGRAPHY Philip Schneeberger was born in Baltimore, Maryland, on November 22, 1887. His primary education was obtained in the public schools of that city and at the Baltimore City Col- lege, from which he graduated in 1906. His Collegiate educa- tion was obtained at the Johns Hopkins University, which he entered in 1906, and from which he received his A.B. in 1909. Thereafter he pursued graduate courses in chemistry at the Johns Hopkins University, and was laboratory assistant there in general inorganic and organic chemistry for the years 1910-1 1 and 1912-13. During his graduate work in chemistry, since October, 1909, his subordinate subjects have been physical chemistry and geology. THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW AN INITIAL FINE OF 25 CENTS WILL BE ASSESSED FOR FAILURE TO RETURN THIS BOOK ON THE DATE DUE. THE PENALTY WILL INCREASE TO 5O CENTS ON THE FOURTH DAY AND TO $1.OO ON THE SEVENTH DAY OVERDUE. AUG 13 193"* ten? 2B 1997 LD c UNIVERSITY OF CALIFORNIA LIBRARY