284 THE ANALYST, ON THE COMPOSITION AND VALUATION OF OILS USED FOR GAS-MAKING PURPOSES. BY RAYMOND Ross F.I.C. AND J. P. LEATHER. (Read at the Meeting June 14 1906.) A LARGE number of investigations have been made by numerous workers into the composition of the various petroleum oils found in certain parts of the world. These investigations have been largely confined to the fractions of low boiling-point and their principal aim has been the separation and isolation of definite compounds. The aim of our work has been primarily directed to the valuation of those oils which are used for gas-making purposes. Such oils consist generally of those portions of the crude oil which boil between 200' C. and 400" C. In certain states of the oil market even the refined petroleum burning oils are also used.From the results obtained by previous investigators-to which we shall refer later-we know that the various classes of hydrocarbons are found in varying pro-portions according to the origin of the oil. This being the case it is evident that to obtain a satisfactory knowledge of the gas-making properties of any oil it is necessary to ascertain the proportion of the different classes of hydrocarbons present in the oil and the gas-making properties of each class. The presence of the following classes of compounds in given petroleum oils has been proved by various workers-notably Markownikoff Zelinsky Aschan, Schorlemmer Engler Schutzenberger Maybery and others : Paraffins. Olefines. Aromatic hydrocarbons. Naphthenes. In this last term are included saturated and unsaturated hydrogenated deriva-tives of both benzene and naphthalene and their homologues.Some preliniinarg work which we undertook-the result of which was communicated to the Society of Chemical Industry (May 1902)-1ed US to think that the determination of certain chemical and physical properties of an oil might enable a correct judgment to b THE ANALYST. 285 formed of the nature and composition of any gas-oil. For this purpose it was necessary that the physical properties and the value for gas-making properties of each class of hydrocarbons should be separately determined. The present communication sets forth the progress that we have so far made in this direction and also the results obtained from the examination of a large number of gas oils.Although owing to the difficulties inherent to such a research our progress is more limited than we had hoped we trust that our results will be found to be both of theoretical interest and practical utility. Before proceeding further it will perhaps be best to recapitulate the methods which we use for the valuation of an oil and which are substantially those men-tioned in our previous paper. For the purpose of gasifying the oil we make use of a retort about 9 inches long by 52 inches wide and 42- inches in height. The retort carries an electrical pyrometer and two tubes-one for introducing the oil, going only a short distance into the retort and the other for the exit of the gas, being carried the whole length of the retort. The oil is fed in from a burette and the gas after passing through a tar-bottle, is collected in a holder measured and analysed.After the retort has been brought So the required temperature the gas-supply is turned off and 15 C.C. of oil run in during a period of about three minutes. Experience has shown us that if these directions are followed there is practically no variation in the temperature of the retort during the period of gasification. The size of the retort as well as the rate of flow of the oil are important factors as if the time of contact is altered the quantity and quality of the resultant gas are affected. The resultant gas is measured and then the amount of hydrocarbons absorbed by fuming sulphuric acid is ascertained in a Hempel burette and this figure is given in all our results as (' hydrocarbons." The number of C.C.of gas at N.P.T. per C.C. of oil (at 15" C.) multiplied by the percentage of hydrocarbons ascertained as above gives a number which we have called the '' valuation figure." The temperatures which we have selected for cracking the oils are those at or between which we have after many experiments, invariably obtained the highest valuation figure. In addition to the foregoing cracking test we make a distillation test on the following lines 250 C.C. of the oil are distilled and the fractions are separated at intervals of 20" C. in such a way as to have an even number for the second digit. The quantity of each fraction is measured and its specific gravity and refractive index determined. The specific refraction is then calculated.Before proceeding further it will perhaps be best to give a rksumh of the results which we have obtained by the examination of a number of different oils from various parts of the world. This is heated in a mufie furnace 240" C. t o 260" C. ____ .___ 280" C. to 300" C. 300" C. to 320" C. 1 200" c. t o 220" c. -~ N - I I ND. yj--I ~ 1.4570 0-556 0 00141 0.001 1.46551 0.555 i I 1.4711 0'552 I I 1 ND. I -1- I I 1'4449 0'558 - 1 -I 1.4697 0.555 I ' 1.4659 0'552 1.4679 0-549 1.4775 0'548 0'0028 0'004 1'4709 0.553 1'4796 0'656 0'0064 0,002 - i -SP. Or. N - 1 ND. 7 -1.4734 0.555 0.0007 0*001 L=18161 0'555 - 1 - I L'4S29 0'549 1'5042 0.551 1.0025 0.002 1.5040 0 555 I 1'4959 0'557 1.0099 0.002 1'4S95 0559 i 1'4947 0'5i4 1'0015 0'003 1 I - I -'4651 0561 1.00231 0 001 I SP.Gr. __ 0.797 -0.829 0'844 0'852 0.873 0'014 0'861 0.861 0'011 -i l PENNSYLVANIAN : KANSAS : 0'804 I1'4475J 0.556 RUSSIAN REFINED : 0'825 11.46471 0.551 RUSSIAN : - 1 - 1 -TEXAS : 0'844 1'4607 0'546 . - ~0*0005j 0'001 CALIFORNIAN : 0'822 I 1.4554 I 0'554 ROUMANIAN : 0'829 1*4616/ 0556 . 0'015 10.0080 0.004 GALICIAN : - 1 - 1 -GROSNY : , BORNEO : ' O'S61 I1'4827l 0.560 SCOTCH : - _ _ r . - l - / -0.841 ~ 1.4681' 0-55E - I 0.00031 0*001 0'860 ,I-4780 0.556 , I I I - - I - ' I 0371 11.47941 0'550 0.851 0 001 0967 -0379 0'9t.5 0'003 0'908 O'S90 0'018 O'S76 0.894 0-007 I-O'S64 3 '003 0'821 0.002 0.838 0.853 0.833 1'4635 0'55f - 0'0004 0'001 j I 0 S50 11.47231 0'55 ' 1 ! - ! - I -0 862 ! 1.4749 0'551 I 1 - I - ! _ 0'855 1-4705 0551 1 ; 0-895 11.49181 0'529 0'0022 0,003 0'906 11'4974 0'549 0'895 1.4989 0'557 l 1 0'883 1'4920 0'557 0'584 1'4849 0.348 0.003 0'0019 0'004 0.003 0'880 0.882 3'012 3'856 3.876 1-007 3'936 1.844 1'004 0.836 I1.4628' 0.553 I 1 1'4873 0'554 1'4914 0.557 0.0057 0.009 1.4722 0.557 1'4S47 0'553 0'0013 0'003 1.5242 0.560 1.4732 0'560 0'00231 0'001 I ( 1 1 ' 0'847 1.4709' 0'556 0,009 0'0061 0'002 0 017 0.865 0.886 0'007 -0.553 0.004 -0 0097 0'002 1'4S30 0'558 I 1-4886 0'552 0'0028 0'001 i - I -I 1*47iB 0'560 0'0027 0-001 ~ ""-I - ""-- I -1 -0'860 0'018 0.920 I I 0'857 1.4ilCh 0'550 0*006 0'0014 0.002 1~ 0'900 1'5043 0'560 1 -I -- 1 -I 0'833 1.4673 0.330 0.007 0.0035; 0'001 N - 1 Nn=Refmctive index for Sodiiini light.=specific refractioa Lincs i~nrlred 1. give tlic cxtrem Name of Oil _ ~ _ _ _ -Pennsylvanian Kansas . Russian . Russian refined Texas . Californian Roumanian Galician . Grosny . Borneo . Scotch . .-.{ .{ . . *{ .{ .I .{ 1 *-I . .{ .I .{ THE ANALYST. AVERAGES OF GASIFICATION RESULTS. Tnipcrature of Cracking. 1260" F. 1400' F. 1510' F. 1260' F. 1400' F. 1260' F. 1510" F. 1260" F. 1510' F. 1700" F. 1130" F. 1260' F. 1400' F. 1450" F. 1510' F.1130" F. 1260' F. 1400" F. 1130' F. 1260" F. 1400" F. 1450" F. 1260" F. 1130' F. 1260" F. 1400" F. 1260" F. 1510' F. 1640" F. 1130' F. 1260" F, 1450" F. - __ C.C. of Gas N.T.P. per c. of Oil. 44.5 529-9 563 438.9 483.6 465.7 556 429 518 550 325 388.3 461.8 547*7 508.8 3'70.8 5299 573.9 301.3 388.8 459.7 557.3 452.8 341.8 421-9 501.6 301 472 . 495 286-5 426.4 491.5 Hydrocarbons per Cent. -35.8 30.1 26.6 33.6 28.4 34.2 22.8 31.8 28.4 21.5 30.1 29.8 25.3 18.6 21.1 29.8 26.6 25-4 40.5 33.3 28.6 20.2 35.5 34.9 34.6 25.8 26-8 17 15 40.3 34 -2 26.1 287 Valuation Figure. 15,944 15,950 14,967 14,747 13,706 15,927 12,677 13,642 14,711 11,825 9,769 11,490 11,697 10,187 10,741 11,050 14,095 14,577 12,302 12,895 13,147 11,257 16,074 11,925 14,489 12,860 8,067 8,024 7,425 11,548 14,620 12,790 As the worst results which we have obtained were got from Borneo oil we determined to make a special examination of this oil.We commenced work on the lower boiling fractions with an attempt to obtain some definite compound by means of repeated fractional distillations but without success. We therefore had recourse to chemical means and to this end treated that fraction of the oil which boiled in the neighbourhood of 200' C. with an excess of fuming nitric acid for a period of some days with constant mechanical stirring, adding the nitric acid gradually and warming when the violence of the reaction had abated.At the end of the time the mixture separated into three layers-viz. unacted on oil a wax-like substance and nitric acid. The residual oil was treated with sulphuric acid and then heated with fuming sulphuric acid (50 per cent. SO,) sever 288 THE ANALYST. times. I t has been shown that fuming nitric acid acts in the following manner (Francis and James Chzem. Xoc. Jown. 1898)-viz. it readily attacks aromatic compounds methyl derivatives of polymethylene compounds and the iso-paraffins. All these compounds are characterized by the presence of a CH group. Consequently in the fraction experimented upon only normal paraffins and decahydronaphthalene could be present in any quantity after the double treatment.An analysis of the remaining oil gave the following figures : Per Cent. ::::; 1 Specific gravity of oil 0-821 15"/15" This analysis shows less hydrogen than corresponds to CnHen and it therefore approximates more nearly to decahydronaphthalene than to a paraffin, The oil was next boiled with excess of bromine in carbon tetrachloride with a reflux condenser for many hours in the hope that some substitution might take place in the paraffins. Hydrobromic acid was given off and after distilling off the carbon tetrachloride the residue was boiled with caustic soda to decompose the bromine compounds and thus leave unsaturated products. After washing the oil was again treated with fuming sulphuric acid and warmed. After separation washing and drying the oil gave on analysis : Theory for CI,Hl8.C 86.77 . . 86.96 H 13.24 . . 13.04 Its specific gravity was 0,843 at 15" C. The N 1.4507. Boiling-point 169.5" C. Molecular weight determination by Beckmann's boiling - point method 1.33.4 (theory 138). According to Beilstein the gravity is 0.847 at 15" C. and the boiling-point 173" to 180" C. The substance obtained which was probably fairly pure deca-hydronaphthalene was then cracked in a similar manner to the gas oils at 1260" F., and gave results as below : 434.1 c,c. of gas at N.T.P. per'c.c. of oil containing 26.2 per cent. of hydrocarbons soluble in fuming sulphuric acid giving a valuation figure of 11,373. The yellow wax-like substance obtained by nitrification was washed separated, and dissolved in absolute alcohol.After many attempts two batches of crystals were obtained in which the carbon was estimated by moist combustion with chromic acid. Several ordinary combustions were spoilt by explosions even when air was used and great care exercised. The gas from the moist combustion mas led through a combustion-tube containing copper oxide and reduced copper. The first batch of crystals gave as a result of three closely agreeing analyses carbon 54.28 per cent., and the nitrogen as determined by the modified Kjeldahl method was 12.7 per cent. These figures point to a dinitrotetrahydronaphthalene for which the theory is C=54.04 per cent. and N=12-61 per cent. To verify this suggestion we treated about 70 grams of the nitro compound with 300 grams of tin and 1 litre of hydro-chloric acid heated under a reflux condenser until the tin was dissolved; we then added an excess of strong caustic soda.This was washed with water and finally distilled i7z wacuo. A thick pitchy mass rose to the top THE ANALYST+ 289 To the distillate hydrochloric acid was added and after standing crystals were obtained which on being recrystallized from alcohol were nearly colourless and gave on analysis-Theory for C,,H,, Per Cent. (NH,HCI),. C 50-7 . . . . C 51.1 H 7.4 . . . . H 6-8 N 11.8 . . . N 11.9 C1 29.9 (calculated from nitrogen) C1 30.2 99.8 100*0 . -The agreement with theory is sufficiently close to identify this compound as diamidotetrahydronaphthalene hydrochloride. From some uncrystallized nitro com-pound we obtained by the above methods an amido compound which went rapidly black and was probably an isomer of the first but was not analysed.The foregoing investigations having dtrnonstrated the presence of decahydronaphthalene and tetrahydronaphthalene in the fractions experimented upon we next attempted to prepare these substances synthetically. Tetrahydronaphthalene was prepared by the action of sodium on a solution of naphthalene in amyl alcohol (according to the method of Bamberger and Kitschelt), the sodium being added gradually to the boiling solution. The oil obtained was separated dried and distilled. After many repetitions we finally obtained about 200 C.C. of distillate boiling between 203" C. and 206" C. This was cooled to freeze out any naphthalene and redistilled.The distillate between 203" C. and 206°C. gave the following constants Specific gravity 0.977 at 15" ND= 1.5712 I-__ = 0.584. 205" C. with the following results : ND-1 d Bamberger and Kitschelt give the gravity as 0.978 and the boiling-point at The tetrahydronaphthalene obtained as above was next gasified at 1,260" F. C.C. of gas per C.C. of oil _ . . . 134.5 Hydrocarbons . . . . . 13.6 Valuation figure . . 1,829 . . . The significance of these results will be referred to later. From the tetrahydro-naphthalene we attempted to prepare decahydronaphthalene by passing its vapour together with hydrogen over reduced nickel but without success although by this method we had previously obtained a good yield of hexamethylene from benzene without difficulty.We finally prepared some decahydronaphthalene by treating tetrahydronaphthalene with hydriodic acid and phosphorus in sealed tubes. Eleven tubes each containing 10 C.C. of tetrahydronaphthalene 0.5 gram of phosphorus, and 25 C.C. of hydriodic acid saturated at 0" C. were sealed off and heated in a bomb furnace. As they gave a very poor yield the remaining seven tubes were heated only to 210° C. and the time was extended to four days. From these tubes we obtained 14 C.C. of a liquid boiling between 170" C. and 178' C. This was redistilled the greater portion coming over between 170' C. and 173' C. This fraction had the following constants The first four were heated to 260" C. for thirty-six hours 290 THE ANALYST. Specific gravity = 0.8426 at 15" C. . N . . = 1.4486 at 20" C.C . 86.91 H . 12-90 99.81 The gravity and boiling-point are thus = 0.532 Theory-for C,,H,, 86-96 13.04 100~00 very slightly lower than those given by Wreden and are in close agreement with the decahydronaphthalene obtained by US from Borneo oil. According to Engler when petroleum oils are dissolved in a mixture of alcohol and ether (equal parts) and cooled the paraffins solidify out and may thus be separated and estimated. Five representative oils were fractionated so as to obtain those portions which boiled between 300" C. and 350" C. with the exception of Borneo of which the portion boiling between 280" C. and 340" C. was retained. Each of these fractions was then mixed with about four times its bulk of alcohol and ether and cooled to - 18" C.by means of liquid ammonia. I n the case of the American and Roumanian oils solids were obtained which were separated off and pressed between blotting-paper. The Russian Texas and Borneo oils separated into two portions and in each case the portion insoluble in the alcohol-ether mixture was separated off. It was naturally thought that these portions would principally consist of paraffins and olefines and they were cracked in the usual manner. The solid from the Roumanian oil had a gravity of 0.839 at 42" C. and an iodine absorption (Hubl) at '7.09 per cent. As this evidently contained some olefine it was treated with Nordhausen and after treatment the gravity was 0.857 at 15" C. and the iodine absorption was nil. The refractive index at 27' C.was 1.4709 therefore = 0.549. These figures corrected to gravity at 15" C. and N at 20" C. give N - 1 d N - 1 d =0-553. The American solid portion had a gravity of 0.849 at 15' C. and N at 20" C. was therefore 1.4734 = 0.557. d The iodine absorption was 17.2 per cent. The following table gives the results obtained on gasification of the above : C.C. of Gas Per Cent. Valuation per C.C. of Oil. Hydrocarbons. Figure. Roumanian . 465.4 35.4 16,475 American . 431.2 40 -0 17,248 Texas . . 430.2 27.2 11,701 Borneo . . 342.7 26.4 9,047 Russian . 425.3 38-9 16,544 The insoluble portion of the Borneo oil had a gravity of 0.9286 at 15" C THE ANALYST. 291 According to Kraft the gravity of the higher paraffins at their melting-points approximates to 0.777.We prepared undecane by Krafft's method from oil of rue. The oil was obtained from the London Essence Company and our best thanks are due to Mr. Burgess, their chief chemist for the trouble he took to procure us oil rich in ketones and for information as to the best method to prepare them. His method with very slight variations was used and was carried out as follows 240 C.C. of the oil of rue was mixed with 1 litre of a 40 per cent. solution of sodium metabisulphite. The mixture was kept at 70" C. on a water-bath for five hours with vigorous mechanical stirring all the time. Eight lots of 4 grams each of sodium carbonate were added at intervals of fifteen minutes. At the end of the five hours the mass was allowed to stand overnight and the liquid portion poured off.The crystalline residue was washed three times with ether cooled to 0" C. and then drained and pressed between blotting-paper. To the residue an equal weight of sodium carbonate and enough water to cover the mass were added. The mixture was heated on a water-bath under a reflux condenser for some hours. The liberated ketone was separated off and the mass extracted with ether and the ethereal extract added to the ketone. The ether was distilled off and the ketone fractionated. The portion boiling between 223" C. and 225" C. had a specific gravity of 0.8266 at 20" C. and N 1.4271. Theory for ND=1*4231 and according to Beilstein the gravity is 0.8268 at 20-5" C. Several lots of methylnonylketone were prepared in the above manner and the yield was in some cases as high as 50 per cent.of the oil of rue taken. Some of the ketone thus obtained was mixed with an equal bulk of chloro-form and one-third more than the calculated weight of phosphorus pentachloride added gradually to the liquid through a reflux condenser. When all the phosphorus pentachloride had been added the contents of the flask were heated on a water-bath for a short time and then cooled and poured into water. The chloride and chloroform were separated off and the chloroform dissipated on the water-bath; 12 C.C. of the undecylene chloride were then treated with 25 C.C. of hydriodic acid (2.0 specific gravity) and a little phosphorus in sealed tubes at 210" C. for forty-eight hours. The contents of the tubes were poured into sodium thiosulphate solution and the oil separated and distilled, It was redistilled and the portion boiling between 193" C.and 195O C. (except 754 mm.) gave the following constants : The greater part came over between 190" C. and 200" C. Specific gravity - 0-7466 at 15'/15'-N 1.41817 N 1.41620 According to Krafft the density is 0.7448 15'/4" and the boiling-point 194.5" at The calculated refractive index according to Bruhl's formula is Na - 1-41254. 760 mm 292 THE ANALYST. The above substance is therefore fairly pure undecane. The undecane was next cracked in the usual manner with the following results : At 1260" I?. At 1400" F. C.C. of gas at N.T.P. per C.C. of oil . 492.4 544.4 Hydrocarbons . . . . 36.4% 33 ~80/~ Valuation figure . . . 17,923 18,400 As we were unable to separate olefines from the petroleum oil we give the preparation and the results obtained from an olefine prepared from oil of rue.Methylnonylketone prepared in the manner already described was converted into the carbinol by mixing it with amyl-alcohol and adding a considerable excess of sodium to the boiling mixture using a reflux condenser. This operation occupied about three days. When the reaction was completed water was added and the methylnonylcarbinol and amyl-alcohol separated. The amyl-alcohol was distilled off and the carbinol remaining was then very gradually run on to an excess of phosphorus pentoxide in a flask provided with a reflux condenser and heated on a water-bath. The resultant mass was next extracted several times with ether and after distilling off the ether the oil.obtained was fractionated in a four-bulb Young's dephlegmator ; the fraction boiling 191' C. to 193" C. gave the following constants : . Specific gravity 0.7735 at 15'/15" N D 1 -43 3 25 __ 0.560 N a 1.43148 8 0.558 N - 1 d NP 1 -44198 7 9 0.571 N -1 d The -?-.- calculated according to Bruhl's figures is 0,558. According to different authorities the boiling-point of undecylene varies from 192" C. to 196' C. and the gravities from 0-790 to 0.8398 at 0" C. The density calculated from the formula proposed by Isidor Traube (Bey. 1895, xxviii. 2924) is 0-765 k 0.02 which is in close agreement with our results. A Hiibl iodine absorption of the olefine gave 164.4 per cent iodine absorbed the theory for undecylene being 164.2 per cent.This oil was therefore fairly pure undecylene. The undecylene obtained was next gasified in the usual manner with the following results, 15 C.C. being used for each experiment : At 1260" F. A t 1400" 3'. C.C. of gas at N.T.P. per C.C. of oil . 493.6 550.4 Hydrocarbons . . . . 30*80/6 29.0% Valuation figure . 15,202 15,961 Benzene vapours . . , . 2.4 Heavy hydrocarbons . . . . 28.4 Methane . . . . \ 45.2 Hydrogen . . . . . 23-9 . . The analysis of the gas obtained at 1260" F. was as under: Per Cent. 99. THE ANALYST. 293 I n our first attempt to prepare undecylene we used amyl-alcohol and sodium amalgam for the preparation of the carbinol passing in a stream of carbon dioxide all the time. The resultant oil was treated with phosphorus pentoxide in the manner before described but the temperature was kept much higher (about 200' Cl.) by means of an oil-bath.Very little undecylene was obtained but a series of fractions of the following composition : B. P. Carbon. Hydrogen. 240" C. to 280" C. . . 86.0 13.6 280" C. to 350" C. . . 86.85 13.03 350" C. to 370" C. . . 86.55 13.17 370" C. to 375" C. . . 83.05 13.57 Although these fractions were further examined no very definite conclusions could be drawn as to their constitution. The last compound is possibly a ketone of approximately the formula C32H640 or C32H,,0. In order to institute a comparison between our method of valuation by means of cracking on the small scale and the results obtainable in practical working on the large scale we append a series of tests carried out on the large scale with certain oils.These oils were used in a carburetted water-gas plant erected by the Economical Gas Apparatus Construction Company. I n every case all possible precautions were taken to obtain the best results, which were as follows : Pennsylvanian Oil. Laboratory valuation figure 16,000. Quantity of gas made 391,000 cubic feet. Amount of oil used 969 gallons = 247 gallons per 1,000 cubic feet. Average candle-power 22.12 = 8.9 candles per gallon per 1,000 cubic feet. Quantity of coke used in generator 13,260 pounds=33-9 pounds per 1,000 cubic feet. Russian. Oil. Laboratory valuation figure 15,927. Quantity of gas made 1,186,000 cubic feet. Quantity of oil used 3,078 gallons = 2.59 gallons per 1,000 cubic feet.Average candle-power 22.61 = 8-73 candles per gallon per 1,000 cubic feet. Quantity of coke used in generator 38,556 pounds = 32-5 pounds per 1,000 cubic feet. Texas Oil. Laboratory valuation figure 11,697. Quantity of gas made 349,000 cubic feet. Amount of oil used 1,144 gdlons=3*277 gallons per 1,000 cubic feet. Average candle-power 20.51 = 6.2 candles per gallon per 1,000 cubic feet. Quantity of coke used in generator 12,240 pounds = 35.0'7 pounds per 1,000 cubic feet 294 THE ANALYST. Borneo Oil. Laboratory valuation figure 8,024. Quantity of gas made 5,410,000 cubic feet. Amount of oil used 19,645 gallons = 3.63 gallons per 1,000 cubic feet. Average candle-power 17-03 = 4.69 candles per gallon per 1,000 cubic feet. Amount of coke used in generator 178,500 pounds = 33 pounds per 1,000 cubic Taking American oil as a standard the following comparisons may be instituted : feet.Laboratory Test. Works Test. . 100.0 American oil . 100.0 Russian oil . . . 99.5 98.1 Texas oil . . . 73.1 70.3 Borneo oil . . . 50.1 52- 7 From the above it is evident that our method of valuation yields approximately correct results. Having now completed the experimental portion of our paper we will proceed to discuss the conclusions which we think may fairly be drawn from the foregoing result a. A consideration of the figures in the distillation test tables makes it evident that, comparing fractions of the same boiling-point the lower the gravity and refractive index, the better the oil for gas-making purposes. The following table gives a summary of the constants obtained for the different ' Boiling- Name.i point, I Undecane . . . ' 194" Undecylene . . 193' . Decahydronaphthalene . i 172" Hexahydrocymene (accord- , ~ Tetrahydronaphthalene . 1 205" to Markownikoff) ._. 1 161" Specific Gravity. 0.746 0-773 0.843 0.977 0.783 ND 1.4182 1.4332 1.4507 1.5712 1.4323 0.560 0-5 60 0,534 0-584 0.552 __ __--Valuation Pigure. 18,400 15,961 11,373 1,829 -From this table it will be seen that although the boiling-points are not far apart the paraffin has the lowest gravity the olefine decahydronaphthalene and tetrahydronaphthalene increase in gravity in the order named. The same also holds good for the refractive index while on the contrary the value for gas-making pur-poses decreases in the same order.A further class of bodies which we have not yet succeeded in preparing synthetically consist of such compounds as hexahydro-cymene. Such cycloparaffins partake of the nature of both open chain and ring compounds and their value for gas-making purposes will no doubt depend on the relative number of carbon atoms contained in the chain and ring respectively. The presence of such compounds is indicated by the high gravity of the soli THE ANALYST* 295 paraffins separated from American and Roumanian oils whilst the specific refraction is only slightly lower than that of normal paraffins; Generally we think we are justified in concluding from our results that-1. The open chain compounds have the best value for gas-making purposes.2. The presence of double bonds in the chain slightly decreases the value. 3. The presence of one or more rings diminishes the value considerably. 4. The more fully hydrogenated the ring the better its value as compared with 5. Benzene rings have practically no value for cracking purposes. Next considering the relation of the gravity to the refractive index as expressed by the specific refraction there are some important and interesting deductions to be other ring compounds. made. The open chain compounds both paraffin and olefine have a specific refraction of about 0.555 to 0.560. The presence of a hexamethylene ring lowers the specific refraction and a saturated double ring reduces it still further. On the other hand, the specific refraction is raised considerably by the presence of an unsaturated ring, as in the case of tetrahydronaphthalene.Having deduced these conclusions from an examination of the pure hydrocarbons, it will be well to consider their bearing on the constants given by some of the various gas oils examined. For this purpose it will be well to consider the oils from the different regions separately and in most cases for simplicity to compare the fractions boiling between 280' C. and 300" C. as these are contained in all the oils examined. I n Pennsylvanian oil the gravity and refractive index of this fraction are lower than those of any of the other oils examined. The specific refraction (0.556) approaches that of a paraffin. This would indicate that the fraction contained a large proportion of paraffins and olefines.This opinion is modifid by the fact that its gravity (0*833) and that of the solid portion obtained on freezing out from ether-alcohol solution (0.849) are too high in relation to the melting-point and boiling-point to consist entirely of these substances. They would however quite agree with a mixture of paraffins olefines and what we may term paraffinoid bodies-Le. com-pounds containing one or more long open chains together with a hexamethylene ring. Applying the same process of reasoning to the case of Russian oil we should conclude that this oil consists principally of paraffinoid substances as though the gravity (0.862) is somewhat higher than in the case of American the specific refrac-tion (0.551) is rather lower.This conclusion is supported by the fact that it is difficult to freeze out any solid substance from the ether-alcohol solution and that the figures obtained could not agree with any other conclusion or with the high valuation figure obtained. I n the case of Texas oil the gravity is 0.895 and is higher than that of either Pennsylvanian or Russian and the specific refraction is still further lowered to 0-549. This oil therefore probably contains considerable quantities of complex ring com-pounds fully hydrogenated but not tending to be paraffinoid in their nature. This is distinctly confirmed by the valuation figure. The Roumanian oils vary considerably but considering the average gravity (Om882) and the specific refraction (0-557) of the fraction previously indicated we think tha 296 THE ANALYST.they must contain a considerable quantity of unsaturated ring compounds as the gravity precludes any great amount of compounds with long open chains and the high specific refraction is inconsistent with a preponderance of fully saturated ring compounds. Some Roumanian oils however have a composition differing from the average given and fall between Pennsylvanian and Russian oils in their properties. The Borneo oil examined by us differs from any of the preceding oils in that it contains a larger amount of unsaturated and partially hydrogenated ring compounds. This is shown by its very high gravity (0.936) in the specified fraction together with ts high specific refraction (0.560). This conclusion is also entirely in accordance with the valuation figure.Of the other oils examined the Kansas with a gravity of 0.850 and a specific refraction of 0.555 is somewhat inferior to Pennsylvanian for gas-making purposes, and in composition it is between Pennsylvanian and Texas but more nearly approaches the former. Californian probably contains a fair proportion of paraffins and paraffinoid bodies, but in addition to these its constants indicate a certain amount of incompletely saturated compounds particularly in the higher fractions. The only Galician oil we have examined was very similar in gravity to Russian oil but had a higher specific refraction thus indicating a somewhat higher proportion of open chain compounds but less than in the case of Pennsylvanian. The Grosny oils approach the Russian in character but are somewhat inferior, probably due to the presence of more paraffinoid bodies and fewer true paraffins.The Scotch gas oils are of course not true petroleum oils. From their low gravity together with the specific refraction of 0.560 one would almost expect a better valuation figure than that obtained. By the method of formation destructive distillation one would naturally expect a considerable proportion of olefines and also some substances containing benzene rings. These latter are probably paraffinoid in their nature as otherwise the gravity would be much higher. We have not however, had time to devote the same attention to Scotch oils M to the others and hope at some future time to investigate them further. We recognise that none of our deductions as to the composition of the various oils can be taken too dogmatically as the number of possible hydrocarbons is so great while the constants from which the deductions have been drawn are few in number.Considering the present state of our knowledge of the higher hydro-carbons we think that the method of reasoning adopted will be found to give a useful index as to the relative value of oils used for gas-making purposes. DISCUSSION. Mr. L. nIYDDELToN NASEI asked whether in addition to the valuation of the oils from the volumes of gas and the proportions of hydrocarbons yielded any photo-metric tests of the oil-gas had been made. He had at one time had a good deal to do with the manufacture of oils of this kind and had always regarded Borneo oil as the lowest in value for gas-making purposes his experience thus agreeing with that of the authors in this respect THE ANALYST.297 Mr. Ross said that the carburetted water-gas in the works tests had been examined photometrically but not the oil-gas made in the laboratory. Some other kinds of Borneo oil and the higher fractions of certain Borneo wax oils gave con-siderably higher valuation results. Some of them contained solid paraffins and so were of low specific gravity. Mr. W. J. ATKINSON BUTTERFIELD thought that the laboratory figures were not quite on all-fours with the works results because the former were arrived at without any knowledge of the nature of the hydrocarbons in the gas. Very often the gas containing 20 per cent. of hydrocarbons from one oil would have a very much higher illuminating value than the gas containing 20 per cent.of hydrocarbons estimated in the same way from another oil owing to the different nature of the hydrocarbons in the two cases. The method proposed by the authors therefore could not be expected to afford an absolute criterion of the gas-making value of an oil. Never-theless it was a very useful auxiliary test and he had himself used it for some time in a very similar €orm to that now described. He thought it would be an improve-ment to make the carbonizing apparatus rather larger and to use a somewhat larger gas-holder so that a portion of the gas might be available for photometric and even for calorific tests. These tests should preferably be made on a mixture of the oil-gas with coal-gas because all kinds of oil-gas had not the same power of enriching coal-gas owing to the differences which occurred in the hydrocarbons according to the temperature at which the gas was made and the kind of oil used.The authors had remarked that benzene rings were of no value and perhaps in one sense that was the case; but as a matter of fact carburetted water-gas owed its illuminating power to hydrocarbons which were roughly speaking those contained in 90 per cent. commercial benzol. It was true that benzol and hydrocarbons containing benzene rings were useless for cracking but it was not altogether useless to make benzene and use the coal-gas as a carrier of it for in that way a greater illuminating duty was obtained from the oil than by converting it wholly into a permanent gas, while the mixture of coal-gas and benzene vapour was sufficiently permanent for transmission to the consumer so far as all ordinary degrees of enrichment of coal-gas were concerned.Mr. LEATHER said that theoretically Mr. Butterfield was no doubt quite right in saying that the hydrocarbons absorbed by fuming sulphuric acid had somewhat different illuminating values but from the practical standpoint of the working of the oils in carburetted water-gas it was found that the proportion of hydrocarbons absorbed corresponded very fairly with the photometrically determined illuminating power of the gas,'while the working tests showed that-in the case at any rate of the four oils they had tried-the laboratory results corresponded closely enough with practical values for the ordinary valuation of an oil.When they said that the benzene ring was of no value they did not refer to benzene itself there being no actual benzene in the oils they used. What they had found was that in the case of a hydrocarbon like tetrahydronaphthalene containing one saturated or practically saturated benzene ring a low valuation figure was obtained. Mr; W; Irwin on cracking some portions of ordinary coal-tar oil (which was practically speaking, creosote oil consisting chiefly of hydrocarbons containing benzene rings) had foun 298 THE ANALYST. these to show such poor values as to be really useless for gas-making. As to the illuminating value of benzene they agreed but if there were more than a certain quantity of benzene in the oil the greater part of it would go into the tar instead of into the gas.As long as the gas had to come into contact with tar the quantity of benzene it would carry was strictly limited; With regard to Mr. Butterfield's sug-gestion that they should use a larger laboratory apparatus and so obtain enough gas for a photometric test they would need for this 10 cubic feet of gas as an tbsolute minimum and it was questionable whether on such a scale the cracking temperature could be exactly maintained. I n the apparatus they used at present the cracking temperature was absolute-Le. the whole of the oil was exposed and cracked at an exact temperature within 10 degrees. The retort was heated in the muffle up to the temperature at which cracking was commenced (say about 1260" F.) the gaL a was turned off and the oil run in.The experiment took from three to three and a half minutes and during that time the temperature never varied more than 10 degrees, because the fireclay casing of the retort being slightly hotter than the retort itself, gave up heat to the latter at about the same rate at which the heat was taken up in cracking. This was a rather important point and anyone repeating their experiments would have to use a retort of the same size and apparatus of a similar description, and to see that the time taken for running the oil in was the same or at any rate, to see that the relationship between the various factors was such that the cracking temperature was constant. On a larger scale it would not be possible to insure this.Mr. W. J. ATKINSON BUTTERFIELD agreed that it was advantageous to work at 8 uniform temperature but thought that photometric tests might be made as he had sug-gested by mixing the oil-gas with coal-gas. Tests made in that way too would more nearly represent practical conditions. As to the comparison between the laboratory and works tests the works values for two of the petroleum oils seemed to him extremely low. He had generally found it to be quite as high as or even slightly higher than that of Russian or Pennsyl-vanian oil. He quite agreed with Mr. Leather's remarks with reference to tar having only a low value for gas-making. Mr. J. H. B. JENKINS said that a good deal of oil-gas was made by railway companies and used for illuminating purposes after compression. The value of an oil wbs usually judged by multiplying the candle-power into the number of cubic feet of gas made from a given quantity of oil and if such a figure could either now or at some future time be included in the authors' results it would be very useful inas-much as it would enable a direct comparison to be made between their laboratory valuations and the methods of valuation ordinarily adopted in practice. Mr. Ross said that they could not do exactly as Mr. Jenkins suggested but the factor he had mentioned was given in their works tests for the carburetted water-gas. He might mention that the exact composition of the oil-gas was not at present known but that they had reason to believe that certain bodies occurred in it the presence of which perhaps had not been anticipated. Amongst these was un-doubtedly a considerable quantity of cyclo-pentadiene. This they had succeeded in separating. The value given for the Scotch oil aleo seemed low