1042 BONE AND JERDAN: THE DIRECT UNION OF CIX.-The Direct Union of Carbon and Hydrogen. Part 11. By WILLIAM A. BONE and DAVID S. JERDAN. IN a previous paper on the subject (Trans., 71, 1897, 41) we showed that carbon. and hydrogen combine a t 1200' forming a saturated hydro- carbon; and further, that when the electric arc is passed between carbon terminals in an atmosphere of hydrogen, a saturated hydro- carbon is produced in addition to acetylene, the formation of each con- tinuing until a definite equilibrium bet ween the hydrocarbons, hydro- gen, and carbon vapour is established. It mas also shown that the same condition of equilibrium is attained when the arc is passed for some time in an atmosphere of either acetylene or methane under similar conditions. We had not, however, separated the saturated hydrocarbon, orCARBON AND HYDROGEN. PART 11.1043 hydrocarbons, formed a t 1200°, or in the arc, from the large excess of hydrogen always present in the gases obtained in the various experi- ments, consequently we were unable to produce any experimental evidence as to either the number or character of the saturated hydro- carbons formed in either case. I n absence of any such evidence, we provisionally assumed that the gases under examination contained no saturated hydrocarbon other than methane, and interpreted our analy- tical results accordingly. We felt a t the time that, so long as this question remained unsettled, the investigation of the subject could not be considered complete ; circumstances, however, €or a long time prevented our continuing the research, but during the past three years we have been continuously working on this and other points arising out of the earlier experiments. The present communication contains an account of a series of experi- ments which complete that part of our inquiry which refers to the character of the saturated hydrocarbons produced by the direct com- bination of their elements at l2OO0, and in the arc, and the successful solution of this problem has proved a very tedious and difficult matter.I n order to avoid undue repetition of experimental details, we must refer the reader to the account given in our previous paper of the methods employed for the puri6cation of the carbon and hydrogen required for the research, and also to the diagrams and explanations contained therein of the apparatus used both for the tube experiments a t 1200O and for the arc experiments.Practically the same arrange- ments, with certain minor improvements which further experience suggested, were adopted for the experiments recorded in this paper. The gas obtained by the interaction of carbon and hydrogen a t 1200°, or in the arc, contained, in addition to small amounts of hydro- carbon, a relatively very large volume of hydrogen, as well as a small percentage of nitrogen. The latter was unavoidably introduced with t,he original hydrogen employed, and to some extent also during the course of the experiment. Obviously, therefore, the only feasible method of identifying the saturated hydrocarbon, or hydrocarbons, in such a mixture, was to remove first of all any unsaturated hydro- carbon present (for example, acetylene in the case of the arc gases), and afterwards the free hydrogen by means of palladium. There would then remain a mixture of the saturated hydrocarbon, or hydro- carbons, and nitrogen ; an analysis of this mixture by explosion with excess of oxygen mould reveal whether it contained one hydrocarbon only, or a mixture of hydrocarbons.If the evidence so obtained pointed to the former alternative, as it was found to do in the case of the gas from the experiments at 1200°, the analytical results would alsoindicate the nature of the hydrocarbon. If, on the other hand,1044 BONE AND JERDAN: THE DIRECT UNION OF as in the case of the arc gases, the latter alternative were indicated, further investigation would be necessary.After submitting the gases under investigation to the series of operations indicated in the preceding paragraph, we were able to show that methane is the only hydrocarbon formed from its elements at 1200°, and that the saturated hydrocarbons produced in the arc con- sist of a mixture of methane with another (or possibly others) of its homologues. The methods employed for the identification of this other saturated hydrocarbon mill be fully explained later, a t present it will be suficient if we state that it ultimately proved to be ethane, A. Examination of the Gas obtained by the action of Hydrogen on Solid Carbon in tlie Tube Expriments at 1200'. Analyses published in the previous paper (pp. 49-51) had shown that the gases from the tube experiments contained no unsaturated hydrocarbon, It was therefore only necessary in this case to remove the diluting hydrogen in order to obtain the saturated hydrocarbon mixed with nitrogen only, This was accomplished by means of palladium black.Three different experiments were performed in which hydrogen, free from hydrocarbons, was passed over pure carbon heated to 1200' in the jacketed porcelain tube described in the previous paper (pp. 46-51) ; a current of pure hydrogen was maintained through the jaeket throughout each experiment. About 3 litres of the issuing gas were collected in each case in a glass holder over strong sulphuric acid. The holder, A , was attached to the palladium absorption apparatus as shown in Fig. 1 (p. 1045).The central portion of the apparatus consisted of a series of three glass bulbs, B, each of about 50 C.C. capacity, con- taining some 10 grams of freshly reduced palladium black, previously heated to redness in a vacuum. The tube leading from the bulbs was at one end twice bent at right angles, and sealed to a glass capillary tap, c6, on the other side of which was a T joint, 6, the latter serving to connect the bulbs through its vertical branch with the gas holder, A , and through its horizontal branch with the laboratory vessel, C', stand- ing over mercury in a wooden trough, D. The tube leading from the other end of the bulbs after being bent a t right angles was sealed to the glass capillary tap, c, through which a connection was made by means of a glass capillary tube with a manometer, E, standing over mercury in the same glass reservoir as the barometer, F, and also with the Sprengel pump, H, by means of which any gas in B could be diawn off and collected in tubes over mercury in the trough, K.The absorp- tion bulbs, B, were immersed in a rectangular tin trough, I;, containing water, so that they could be maintained at any temperature betweenCARBON AND HYDROGEN. PART 11. 1045 that of the room and looo, as circumstances required, There were only two rubber joints in the whole apparatus, namely, those connecting the holder, A, and the laboratory vessel, C, with the branches of the T piece leading to the tap, A, of the absorption apparatus. These two joints were made of stout pump tubing firmly wired at either end to the glass tubes which they served to connsct. A t the outset of an experiment, the taps, a and c, of the absorption bulbs and e, of the laboratory vessel, C, were opened and the whole apparatus, with the exception of the holder, A, was thoroughly ex FIG.1. The palladium absorption apparatus. hausted by means of the Sprengel pump, H. The tap e was then closed, the pump stopped, and by opening tap d of the gas holder, A, a quantity of the gas under examination was drawn over into B, until the level of the mercury in the manometer, E, had fallen within a few mm. of that in the vessel, G. The taps d and a were then closed, and the progress of the absorption of the hydrogen by the palladium in B could be followed by observing from time to time the mercury level in E.I n about 20 minutes, the absorption had ceased, and the residual gas was drawn off by the Sprengel pump into tubes standing over mercury in the trough, K. VOL. LXXIX. 4 B1046 BONE AND JERDAN: THE DIRECT UNION OF The water in the bath, L, was now heated to the boiling point, when much of the gas absorbed by the palladium during the previous opera- tion was liberated and drawn off through the pump, Finally, the bulba were exhausted at 100’ and cooled to the ordinary temperature t o he ready for the next absorption. These processes were repeated until the whole of the gas in A, usually about 3 litres, had been paased through the absorption bulbs and reduced to about 400-500 C.C. (first residue). This firat residue was then introduced in three or four p r - tions into the laboratory vessel, C, and subjected to a second concentra- tion in the same manner as before, when its volume was usually reduced to about 120-150 C.C.(second residue). This gas still contained hydrogen, and, indeed, i t is impossible to eliminate all the hydrogen from such a mixture of gases by a process of absorption with palladium, as the hydrogen in palladium hydride has an appreciable vapour pres- sure, even at the ordinary temperature. The last traces must be oxidised and removed as water by contact with ‘‘ oxidised ” palladium. After this second concentration, therefore, the absorption bulbs were detached from the apparatus, the palladium black taken out and heated to dull redness in a hard glass tube attached to a Sprengel pump until as much as possible of the occluded hydrogen had been removed.The pump was then detached, and the palladium maintained at a red heat in a gentle current of sir until i t had acquired the characteristic dark blue oxidation tint. The I ‘ oxidised ” palladium, after being cooled, was transferred again to the absorption bulbs, which were then sealed on to the apparatus, and the whole exhausted as before. The residual gas from the second concentration was intro- duced into C, and drawn over into the bulbs by momentarily opening taps e and a. The water in the bath, I;, was heated to the boiling point, and the gas allowed to remain in the bulbs at this temperature for 20 minutes or half an hour, after which the bulbs were allowed to cool, and the residual gas was drawn off through the pump into tubes standing over mercury and each containing a small piece of solid caustic potash.The residual gas so obtained from each of the three different experiments was analysed by exploding a measured volume of it with excess of oxygen, in some cases diluted with air free from carbon dioxide, and determining (1) the ‘ contraction ’ (C) on explosion ; (2) the ‘ absorption ’ (A) which occurred when the products of explosion were treated with potassium hydroxide, and (3) the oxygen (X) used up in the explosion.* * All gas analyses involved in this research were carried out by means of an im- proved form of McLeod apparatus designed by one of us, and described at a former meeting of the Society (Proc., 1898, 14, 154) ; we have found that this apparatus is especially adapted to the accurate analysis of hydrocarbon mixtures.The ‘ volumes ’ quoted in the text are deduced from the tensions of the moist gases in mm. ofCARBON AND HYDROGEN. PART 11. 1047 The results are tabulated as follows : Experiment. Volume of gas taken ..................... ,, oxygen added ............... ,, air added ..................... ,, treatment with KOH.. , , treatment with alkaline Contraction (C) ........................... Absorption (A) ........................... Oxygen (X) .............................. Volume aftor explosion ............... pyrogallol .............................. Ratio, CIA ........................... Ratio, C/X ........................... A. 65 '3 154.3 nil 152'0 119'3 67.3 32'7 - - 2.07 - (i) 40.3 66 '4 150.8 133-1 102.2 35.6 17 -7 35.5 - 2.01 1.0 B.(ii) 136'0 389.0 nil 406.8 346.0 178.0 118.2 60.8 121-0 1-94 0 *98 C. (i> 100.0 249'75 196.65 388.2 355.4 - 60.2 30.8 - 1-96 - - (ii) 121.5 175.2 232.1 431.5 383.6 97.3 47.9 97.05 2 *03 1 *o - - There can be no doubt that the hydrocarbon in these mixtures was in each case methane and methane only, for Methane requires C/A 2.0 and C/X 1.00. Propane ,, 1-00 ,) 0-60. Ethane ,, 1.25 ), o m . B. Examination of the Gae obtained in the Arc Expscl.imcnte. Our earlier experiments had shown that, in addition to acetylene, saturated hydrocarbons are formed by the direct union of carbon and hydrogen at the temperature of the electric arc. It wag necessary to ensure the complete removal of this acetylene and any other unsatur- ated hydrocarbons, as well as of the hydrogen, before attempting to identify the residual saturated hydrocarbon.As a fairly large volume of this residual gas was ultimately required, and as it was desirable to determine whether the character and compo- sition of the saturated hydrocarbon in any way depended upon the time during which the arc was maintained, the gas obtained in five different experiments was investigated. Each experiment was carried out in the arc apparatus as described in our previous paper (pp. 62-57), except that no samples of gas were collected during the time the arc was being passed, which varied from 10 minutes to an hour in the different experiments, As soon as the arc apparatus, 8, had cooled, the products were drawn off by means of the Sprengel pump, B (see Fig.2, p. 1048), and sent mercury when brought to a certain constant volume, at a constant temperature, in the meaauring vessel of the instrument, the vacuum ' in the barometer being always kept saturated with water vapour. 4 ~ 21048 BONE AND JERDAN: THE DIRECT UNION OF forward by means of a special device under a pressure of 2-3 mm. of mercury into the glass gas holder, C, previously filled with an ammoniacal solution of cuprous chloride. The solution so displaced was allowed to run away through the glass tap, D, into a large Winchester quart bottle. The device for drawing off the gas from A through the pump and sending it forward under pressure to C, consisted in immersing the delivery tube of the Sprengel in a large test-tube, E, three-quarters full of mercury, to which a side tube, F, had been fused.Over this delivery tube was fixed the funnel-shaped end of a vertical glass tube, G, the other end of which was attached to the tube leading into the holder, C. The mercury flowing over from the pump ran off through the side tube, P, into the bottle, El. The capacity of the gm holder, C, FIU, 2. was about halE as large again as that of the globe of the arc apparatus, so that after the products had been all collected in C, it still remained about one-third full of the ammoniacal cuprous chloride solution. The gas was allowed to remain in C for a t least two days, to ensure the complete absorption of acetylene and any trace of carbon monoxide, and was then transferred to another similar gas holder, previously filled with strong sulphuric acid, over vhich it was allowed to stand for other three or four days.The gas could now only consist of hydrogen, saturated hydrocarbons, and any nitrogen originally contained in the hydrogen. The gas holder was therefore attached to the palladium absorption apparatus for the removal of hydrogen, and treated as was the gas from the tube experiment (see p. 1044).CARBON AND HYDROGEN. PART 11. 1049 The residual mixture of hydrocarbons and nitrogen was anslysed by mixing a measured volume of it with excess of oxygen, diluted in some cases with air free from carbon dioxide and measuring (1) the contraction, C, on explosion ; (2) the absorption, A, which occurred when the products were treated with caustic potash.In certain cases, the oxygen used in the combustion of the hydrocarbons was also determined. (1) Blank Expriment.-In order to test the purity of the hydrogen used in these experiments and the validity of the method employed, a blank experiment was performed in which hydrogen prepared and purified as described in our previous paper (p. 45) was passed into the arc apparatus, and without the arc having been formed, was then transferred by means of the Sprengel pump, B (Fig. 2), to the holder containing the ammoniacal cuprous chloride solution. From thence it mas passed into a holder containing strong sulphuric acid, and after- wards subjected to the action of palladium in the absorption bulbs. The only differences between this blank and a real arc experiment were as follows.(I) A small quantity of nitrogen waa allowed to mix with the hydrogen be€ore it passed into the arc apparatus, in order that there might be a measurable residue for analysis after hydrogen had been removed in the final stages. (2) The arc was not formed, and (3) the final treatment of the gas with “oxidised” palladium at 100’ was omitted. As nearly as possible 2 litres of hydrogen were passed into the globe of the arc apparatus; after two “concentrations” in the palladium absorption apparatus it was reduced to about 80 C.C. This residue was exploded in the gas analysis apparatus with measured volumes of pure hydrogen prepared by the electroIysis of water, and subsequently purified by means of palladium, and of air free from carbon dioxide.The products of explosion were afterwards treated with caustic potash, when a very slight absorption occurred. The details of the analysis are as follows : Volume of residual gas taken ........................ 84.9 ,, hydrogen added ........................... 57.1 ,, air ,, ........................... 205.1 ,, after explosion .............................. 262.6 ,, after treatment with KOH ............... 262.3 Contraction, C = 86.7. Absorption, A = 0.3. Assuming that this small absorption was due to carbon dioxide formed by the combustion of methane originally present in the 2 litres of hydrogen used, it follows that this impurity did not exceed 0.015 per cent. by volume. We may therefore assume that the hydrogen used for the experiments recorded in this paper -was practically free from1050 BONE AND JERDAN: THE DIRECT UNION OF Time during which arc was main- tained .................................... Volume of residual gas after removal of acetylene and hydrogen .........Volunie of residual gas after de- ducting nitrogen ....................... hydrocarbon impurity. Further, since the observed contraction is very little more than the sum of the contraction required by the hydrogen added (85.7) and the methane (0.6) present, this blank experiment indicates that practically the whole of the original hydrogen used had been removed by the palladium black. It was, however, deemed advisable not to omit the final treatment with 6' oxidised " palladium in the actual experiments. (2) Actual Expcriments.-The arc apparatus was in each case about two-thirds filled with hydrogen (about 2 litres), and the arc (voltage = 160) maintained for a period varying from 10 minutes to an hour in the different experiments, After removal of acetylene and hydrogen in the manner already described, the residual gas was analysed by explosion with excesq of oxygen in the usual manner.The details of each experiment, and the analytical results are given in the following table : 10 mins. 30 mins 40 C.C. - 35 c. c. - Experiment. 1 I. 1 11. 1 hour I hour 1-1- 50 C.C. Volume of gas taken. .................... ,, oxygen added ............... Volume after explosion ................. ,, treatment with KOH.. Contraction (C) ........................... Absorption (A) ........................... 46 C.C. 60'9 510.9 456-2 382.9 115-6 73'3 62.7 528'4 476'6 405.2 114'5 71'4 36 '1 330'0 315.1 284-2 61.0 30'9 74-2 482'2 462.2 4029 94'2 59.5 Ratio, C/A ............................. ....I 1-58 1 1-65 111.30 mina, 60 C.C. 54 C.C. 55.7 4433 388.0 318-0 111'0 70'0 1.58 i- 1-62 I 1'58 I * The large amount of nitrogen in this reeidual gas was due to an accidental in- leakage of air during the final absorptiou of hydrogen by means of oxidised palladium, These experiments prove beyond all possible doubt that hydrocarbons of the methane series are produced when the electric arc is made between carbon terminals in an atmosphere of hydrogen, The actual quantity of these saturated hydrocarbons so formed is to some extent dependent upon the time during which the arc is maintained, attaining a maximum, as our earlier experiments showed, after about half an hour, and after- wards remaining fairly constant (see also previous paper, pp.57-68}. Further, it is a remarkable fact that the ratio C/A found on exploding the residual saturated hydrocarbons with excess of oxygen is practically constant (1.6) and independent of the time during which the arc isCARBON AND HYDROGEN. PART XI. 1051 maintained. It may therefore be concluded that whatever may be the number and composition of the saturated hydrocarbons obtained in a given experiment they are formed simultaneously, and at equal rates throughout, The experiments indicate, moreover, that methane is one of the saturated hydrocarbons as the ratio C/A= 1.6 could only be given by a mixture of saturated hydrocarbons containing methane (p.1047). The analytical numbers, however, afford no clue as to either the character or the number of the other saturated hydrocarbons. For example, the results obtained in the case of the residual gas from ex- periment I would be given by any one of the following mixtures : CH, ...... 62.5 CH, ... 69.5 CH, ... 75.1 per cent. C,H, ... 34.0 C,H8 ... 16.9 C,H,, ... 11.3 ,, *N, ...... 13.5 N, ...... 13.6 N, ...... 13.6 ,, (1). (2). (3). - - 100*0 100*0 100.0 or, indeed, by an unlimited number of other mixtures of nitrogen methane, and one or more other saturated hydrocarbons. Further, since the densities of all such possible mixtures are identical, it was not possible to distinguish between the various interpretations of the chemical analysis by means of a density determination, With the view, however, simply of checking the chemical analysis, the density of the residual gas obtained in experiment 11, referred to hydrogen, was determined ;, it was found to be 11 *7 inatead of 11 *4 as calculated from the analysis.0. Difukon Expavirncnts with the Residual Gas fiom the Arc Exprimcnts. It waa now necessary to obtain some evidence as to the character of the saturated hydrocarbons other than methane formed in the arc ex- periments. It seemed probable that one or other, if not both, of two methods would enable us to settle the question. The first method consisted in subjecting some of the residual gas from the arc experi- ments to a slow process of diffusion through porous clay tubes many times repeated, and comparing its behaviour with that of artificial mixtures of methane and nitrogen with otrher saturated hydrocarbons (ethane, propane, $c.) giving the same analytical results. The other method, which the kindness of Professor Ramsay enabled us to carry out, consisted in liquefying the hydrocarbons in the residual gas from * Results for nitrogen given in this paper are all calculated ' by difference.1052 BONE AND JERDAN: THE DIRECT UNION OF the arc experiments, and subsequently fractionally distilling the liquid.For the purpose of the diffusion experiments, we mixed together 30 C.C. of residual gas from experiment I. 45 $ 9 ,, 111. 45 9 , 9 9 IV. -- Total 120 C.C. The percentage composition of this mixture, according to analysis, is shown below, supposing that there are only two saturated hydrocarbons present, and that the second is (1) ethane, (2) propane, and (3) butane.CH, ...... 54.3 CH, ... 70.7 CH, ... 76.2 per cent. C,H, ... 32% C,H8 ... 16.4 C,HIo ... 10.9 .. N, ...... 32.9 N, ...... 12.9 N, ...... 12.9 .. (1). (2). (3). The relative ratio at which the constituents of such mixtures would diffuse through a porous plug can, OF course, be calculated from their densities and partial pressures; taking the rate for the methane in each case as unit, the rates for the other constituents are as follows : (1 )* (2). (3). Ethane - - ............ 0.441 - 0.139 - Butane ............ - - 0.074 Nitrogen ............ 0.18 0.133 0-128 Propane ............ Di$usion Apparatus (see Fig. 3, p. 1053).-This consisted of four glass gas wash-bottles, A , B, C, D, arranged in series, as shown in the dia- gram.The central wide glass tube, which fitted into the neck of each bottle by a ground glass joint, was drawn out in the blow-pipe flame and then cut off about an inch below the joint. To it was attached a piece of clay pipe stem about 3-4 inches long, closed at the bottom in the oxyhydrogen flame. The four bottles were connected by fused glass joints, and between each pair was inserted a glass T joint ter- minating in a glass stopcock (b, c, d). The bottle, A, was connected by means of fused glass joints through a similar T piece with the laboratory vessel, E, standing in a wooden trough over mercury. The fourth bottle, D, was connected also by means of fused glass joints with a T piece leading through its vertical branch to the manometer, M, and through its horizontal branch to the Sprengel pump, G.H is a barometer standing in the same mercury reservoir, K, and attached t o the same millimetre scale as the manometer, M. The vertical branches of the T pieces between the bottles, and be-CARBON AND HYDROGEN. PART 11. 1053 tween A and E, were continued downwards beyond the stopcocks, a, 6, c, d, through joints made of stout india-rubber pump tubing" to a horizontal glass tube, PP, which led to a second Sprengel pump (not shown in the diagram), with its barometer and manometer. This second Sprengel pump served to exhaust the apparatus at the outset of a diffusion experiment,, and at the finish to collect the last fraction. All other fractions mere drawn off through the first Sprengel, G, and collected in tubes, L, standing in the trough, N.At the outset of an experiment, the whole apparatus was exhausted by opening all the stop-cocks, a, 6, c, d, e, and leading to the two pumps. As soon as the laboratory vessel, B, was full of mercury, the stopcock, e, was shut. Finally, when the exhaustion was complete, The difwion apparatus. stopcocks, a, b, c, and d, leading to the second pump, were closed. The apparatus was now allowed to stand for 24 hours, to see whether the ground glass joints at the top of the four diffusion bottles, A, B, C, and D, were quite air-tight. It was found possible to make them so by pouring melted paraffin wax over the outer surface of each joint when the apparatus was exhausted.The gas to be diffused was introduced into the laboratory vessel, 3, and then by opening the stopcock, e, it was drawn over into the first diffusion bottle, A. The diffusion process a t once began, and the various it With the exception of these four joints, which were shut off from the bottles throughout the diffusion operations, and the four ground glass joints of the bottles, all other joints in the apparatus were of fused glass.1054 BONE AND JERDAN: THE DIRECT UKION OF fractions, excepting only the final fraction, were collected through the first Sprengel pump, G. Finally, the stopcocks a, 6, c, d were opened, and the residual gas, that is, the last fraction, was drawn off through the second Sprengel pump. It will therefore be clear that in each diffusion operation the gas passed through four porous plugs.The amount of gas taken for each ‘‘ fractionation ” varied within very wide limits, Usually, it amounted to between 25 and 120 c.c., and the time required for each operation varied from 9-18 hours, being shorter the larger the amount of gas involved, Difwion of a Mixture of equal Volumes of Methane and Oxygen in th Apparatus.-To obtain some idea of tho efficiency of the apparatus, 100 C.C. of a mixture of equal volumes of methane and oxygen (oxygen has nearly the same density as ethane, which is, next to methane, the lightest saturated hydrocarbon) were slowly fractionated. Four nearly equal fraction8 were collected, and the oxygen in each deter- mined. Oxygen in original gas ..................60.0 per cent. ,, fraction I ..................... 42.5 ,, ), fraction 11.. .................. 52.0 ,, ,, fraction I11 .................. 5 8 5 ,, ,, fraction I V .................. 69.0 ,, Twenty C.C. of fraction I mere rediffused, and the first 6 C.C. col- lected and analysed. Scheme of Dzfusion Experiments with Arc Gases.-One hundred and twenty C.C. of the arc gases were introduced into the diffusion apparatus as already described, allowed to diffuse, and collected in four fractions, as follows : It contained 30 per cent. of oxygen. Fraction A = 30 C.C. Fraction C = 40 C.C. ,, B = 30 ?, D = 20 Prccction C was then rediffused, tho first 5 C.C. collected were added to A, the next 13 C.C. to B, and the residual gas collected separately. (Fraction C’.) The first 15 c.c, were added to A, the next 14 C.C.to C’, and the residue to D. Fraction B was then rediffused. At this stage, therefore, we had Fraction A = 50 C.C. Fraction C’ = 36 C.C. Fraction D = 34 C.C. Praction A was diffused, and collected in two fractions. Fraction A’ = 25 C.C. Fraction A“ = 25 C.C. Fraction A’ was finally diffused and collected in three approxi- mately equal fractions, a, 6, and c.CARBON AND HYDROGEN. PART If. 1055 Fraction a. (1 1 (2) CH, = 72'0 CH, = 80.6 C,H, = 16.8 C,H, = 8.4 N, = 11.2 N, = 11.0 Fruction D was diffused, and collected in two fractions. Fraction D' = 10 C.C. Fraction D" = 24 C.C. Fraction x. OH, = 36.6 CH, = 59.3 C,H, = 45.3 C,H8 = 22.6 N, = 18.1 N, = 18.1 (1) (2) Methane - Methane - - - 4.29. - - 9.6. Ethane Propane Methane - Methane = 2.624. - - 0.1108.Ethane Propane The ratios of the two hydrocarbons are important in view of the re- sult of the two following experiments, in which artificial mixtures were diffused. Dafwion of a Mixtuoa of Hethane, &?ham, and Nitrogm-A mix- ture of methane, ethane, and nitrogen was prepared as nearly as possible of the same composition as that of the residual gas from the arc experiments, assuming that the second hydrocarbon was ethane. The methane" employed was prepared by the action of a mercury- aluminium couple on a mixture of methyl iodide and methyl alcohol, and in order to remove any traces of hydrogen it might contain, the gas was passed over oxidised palladium sponge heated a t looo. The ethane" used was prepared by decomposing zinc ethyl with * The purity of both the methane and ethane (as well as of the propane used in the next experiment) was in each case proved by a careful analysis (explosion method), the details of which it is not necessary to record.The ratio C/A for the methane was found to be 2.0, and for the ethane 1-254.1056 BONE AND JERDAN: THE DIRECT UNION OF water, and the nitrogen by passing a slow current of air over red-hot copper turnings until the whole of the oxygen had been removed. The gases were then mixed in a small graduat.ed holder over mer- cury. On analysis, the mixture was found t o have the following per- centage composition : Methane = 51-77 ; ethane = 33-15 ; nitrogen (by difference) = 15.18, with ratio C/A = 19575. One hundred and twenty C.C. of this mixture were subjected to a precisely similar series of fractional diffusions as have been described in the case of the residual gas from the arc experiments, and finally the fractions a and x so obtained were analysed, with the following results : Fraction a.Fraction z. Volume of gas taken .......................... 49.2 45.0 ,, oxygen added ..................... 483.95 440.0 Volume after explosion ........................ 445.4 40'7.2 ?) treatment with KOK ......... 395.9 354-3 Contraction (C) ................................. 87-75 77.8 Absorption (A) ................................ 49.5 52.9 Ratio C/A ....................................... 1-77 1-47 The behaviour of the mixture of methane, ethane, and nitrogen when subjected to the process of fractional diffusion was similar to that of the residual gas from the arc experiment.The percentage composition of the two fractions a and z were therefore : a. 2. Methane .............................. 70.1 34.7 Ethane ................................. 15.2 415 Nitrogen (by difference) ............ 14.7 25.8 Ratio ____ Ethane ..................... 4.62 0.835 Bi$.usim of a Mixture of Hethane, Propcine, and Nitrogen.-The propane used in this experiment was prepared by the action of sodium amalgam on a solution of isopropyl iodide in ethyl alcohol; analysis showed it to be pure (ratio C/A= 0.997). The three gases were mixed in a graduated holder over mercury; on analysis, the mixture was found to have the following percentage composition : Methane = 65.5 ; propane= 17.65 ; nitrogen (by difference) = 16.85, with ratio C/A = 1.55, which is nearly that of the residual arc gases, assuming for the moment that the second hydrocarbon was propane.The mixture was subjected to a process of diffusion precisely similar to that carried out with the residual arc gases, and with the mixture of methane, ethane, and nitrogen. This mixture seemed to behave differently during the diffusion MethaneCARBON ARD HYDROGEN. PART IJ. 1057 Fraction. a 19-75 2.10 operations from either the arc gases or the mixture of methane, ethane, and nitrogen. The ' heavier ' fractions passed through the apparatus much more slowly, and the final fractions were only pumped out with great difficulty; last traces of propane obstinately clung to the plugs, and were probably never removed.The fractions ct and x were finalIy analysed, with the following results : (1.) If ethane (2) If propane is present. is present. Fraction. Fraction. z a z n z 4.29 0.808 1 9-6 2'624 Fraction a. 42.1 Volume of gas taken ........................... ,, oxygen added ..................... 388.9 Volume after explosion ........................ 357-85 9 9 treatment with KOH ......... 318*70 Contraction (C) ................................. 73.15 Absorption (A) ................................. 39-15 Ratio C/A ...................................... 1-86 Fraction z. 55.3 494.3 466.7 408.0 8 2-9 58.7 1 *41 The percentage composition of these fractions would therefore be : Methane ........................... 80.80 43.8 Propane .............................. 4-09 20.8 Nitrogen ...........................15.11 35.4 a. x. are 19.75 and 2-10 respectively. methane and the ratios ~ propane The results of these diffusion experiments are most easily under- stood if the ratio for the lightest and heaviest fractions obtained by the diffusion of the two mixtures of known composition be compared with the same ratio for the corresponding fractions obtained when the arc gases are subjected to tlie same process. other hydrocarbon Fraction. a z Methane Other hydrocarbon = 4'62 0'836 - 4 '62 --= 5-53 R, R, 0.836 - Arc gases. Mixture, -- 19-75 - 9.4 -I 4-29 - 5.3 9*6 - 3.6 2'10 1 0.808 12.624- The similarity between the results obtained with the artificial1058 BONE ANT) JERDAN: THE DIRECT UNION OF mixture containing ethane and the residual gas from the arc experiment is so great that very little doubt can remain as to the presence of ethane in the arc gases.The similarity is especially brought out by comparison of the ratio R,/RZ. I n the two cases quoted, the ratios are almost identical, namely, 5-53 and 5.3, whereas if we assume the second hydrocarbon in the arc gases to be propane, we find Ra/R, = 3.6, whilst for the artificial mixture containing propane the value 9.4 is obtained. The evidence afforded by the diffusion experiments is therefore much more in favour of the supposition that the second hydrocarbon in the arc gases was ethane than that i t was propane. On the other hand, it does not exclude t h e possibility of the gases having con- tained some small amount of propane in addition to methane and ethane.The idea of separating the hydrocarbons contained in the arc gases by a process of liquefaction and subsequent fractional distillation occurred to us a t :an early stage of our work, indeed before the dif- fusion experiments were seriously contemplated. We, however, had no means of carrying out the idea until Professor Ramsay, hearing of our difficulty, kindly offered to help us, and we desire to express our best thanks to him for his goodness in enabling us to bring our investi- gation to a satiBfactory conclusion. The following table of boiling points, expressed in degrees absolute, shows that there is a greater difference between the boiling points of methane and ethane than between those of benzene and the xylenes, and as a difference of so many degrees at so low a temperature means relatively much more than a similar difference at higher temperatures, the prospect of almost completely separating methane and ethane seemed good : Boiling points.Boiling points. Diff. Diff. ....... Benzene., .......... 353.4O>29 *cJ Nitrogen.. 90>3 40 Propane ......... 228 >48 - Toluene 383.3 >31,7 Methane ......... 113 ............ Ethane ......... 180 >67 Xylenes ... 411 to 415 Butane ......... 274 61 *6 About 100 C.C. of various fractions of the arc gases left over from the diffusion experiments were thoroughly mixed in a small holder over mercury; a portion of the mixture was analysed with the follow- ing results :CARBON AND HYDROGEN. PART 11. 1059 Volume of gas taken ........................... 38.5 ,, oxygen added ........................344.3 Volume after explosion ........................ 312.25 ,, treatment with KOH ......... 268.25 Contraction (C) = 70.55. Absorption (A) = 44.00. Ratio C/A = 1.60. The gas would therefore have the following percentage composition, assuming (1) that the second hydrocarbon is ethane, (2) that i t is propane : (1) (2) Methane ............ 54.0 Methane ............ 69.0 Ethane ............ 30.0 Propane ............ 15.0 Nitrogen ............ 16.0 Nitrogen ............ 16.0 It will further be noticed that this was practically the composition of the original arc gases before diffusion. About 70 C.C. of this mixture were sealed up in an exhausted bulb, and forwarded to Professor Ramsay, who carried out the fractionation for u s in the laboratory of the University College, London, as follows.The gas was passed into a small gas bulb immersed in boiling liquid air ; a large portion of i t solidified as white, snow-like crystals on the inner surface of the bulb, whilst another portion either liquefied or formed a vitreous, glassy solid. A third constituent, namely the nitrogen, which was only liquefied under pressure, was allowed to pass off uncondensed. The contents of the bulb were slowly volatilised and collected over mercury in three as nearly as possible equal frac- tions, A, B, and C, of about 20 C.C. each. The crystalline solid (the methane) disappeared during the early stages of the process, and the liquid or vitreous solid was entirely converted into gas at a very low temperature, certainly much below the boiling point of butane, and probably that of propane also, After the third fraction had been collected, there remained no residuum of gas in the apparatus.From the behaviour of the gas during the process, Professor Ramsay con- cluded that it certainiy contained no butane, and that the second hydrocarbon was more probably ethane than propane. F-action A was found, on analysis, to consist of methane with some 15 per cent, of nitrogen, as the following figures indicate : Volume of gas taken ........................... 49.80 ,, oxygen added ........................ 124.90 ,, air added (oxygen = 76.37) ...... 364.35 Volume after explosion ........................ 455.0 9 ) treatment with KOH 412.9 ......... 9 , treatment with alkaline pyro-1 296,1 gallol........................... Oxygen used (124.9 + 76-37 - 116.8) ......... 84.471060 BONE AND JERDAN: THE DIRECT UNION OF Contraction (C) = 84.05. Absorption (A) = 42.10. Ratio CIA = 2. The gas therefore had the percentage composition Methane = 84.5. Nitrogen = 15-5. Fraction B on analysis gave the following numbers : Volume of gas taken ............................. 38.0 ,, oxygen added ........................ 99.8 ,, air added (oxygen = 77.7) ......... 370.5 Volume after explosion ........................... 432.5 treatment with KOH ............ 390.9 99 9 9 treatment with alkaline pyro-] 295.5 gallol ........................... Oxygen used (99.8 x 0*988+77.7-95a4)...... 80.9 Contraction (C) = 75% Absorption (A) = 41.6.Ratio CIA = 1.82. The gas evidently therefore contained a large quantity of methane with some ethane or propane and also nitrogen. Its percentage corn- position would be as follows, assuming (1) that the second hydrocarbon was ethane, (2) that it was propane : (1) (2) Methane ......... 83.53 Methane ......... 90-0 Ethane ........... 12-97 Propane .......... 6.5 Nitrogen.. ....... 3.50 Nitrogen ......... 3.5 The amount of oxygen actually used in the explosion agrees very closely with the 80.73 vols. required by 38 vols of either of the fore- going mixtures, if we allow for the fact that the oxygen added was found on analysis to contain 1.2 per cent. of nitrogen. Fruction C, on analysis, yielded numbers agreeing very closely with those required on the supposition that it consisted of nearly pure ethane : Volume of gas taken .............................32.7 ?, oxygen added ........................ 483.0 Volume after explosion ......................... 435.0 ,? treatment with KOH .......... 372.0 Contraction (C) = 80 7. Absorption (A) = 63-0. Ratio C/A = 1.28. It is evident therefore that this fraction either consisted of ethane with a small quantity of methane, or that it contained nearly equal volumes of methane and Now pure ethane requires C/A = 1-25.CARBON AND HYDROGEN. PART 11. 1061 propane. figures indicate : Methane = 2.6 ('7.92 per cent.) Methane = 17.7 (54.0 per cent.) Ethane =30*2(92.07 ,, ) Propane =15*1 (46.0 ,, ) There was evidently no nitrogen present, as the following -- -- 32.8 32.8 The percentage composition of the tiwee fractions mould be as follows, (1 ) on the supposition that the second hydrocarbon was ethane, (2) that it was propane : A.B. C. (1) Methane ............ 84.5 83.53 7.92 Ethane ............... nil 12.97 92.07 Nitrogen ............ 15.5 3.50 nil (2) Methane ............ 84.5 90.0 54.0 Propane ............... nil 6.5 46.0 Nitrogen ............ 15.5 3.5 nil The composition of the middle fraction, B, almost excludes the supposition that the second hydrocarbon was propane, for if it were so, the second series of figures would indicate that whereas we had completely separated nitrogen and methane whose boiling points differ by only 34O, we had not effected a separation of two hydrocarbons whose boiling points differ by as much as 115'. To further test the matter, however, 10 C.C.of the fraction C were subjected to a process of diffusion in the apparatus we have described, which in one operation effected a considerable separation in a mixture of equal volumes of methane and oxygen (see page 1054). I f fraction C consisted of nearly equal volumes of methane and propane, we should expect to get a very considerable separation on difhsion; on the other hand, if the gas was nearly pure ethane, on analysis very little difference should be found between the ratio C/A for the heaviest portion and 1.28, the ratio for the original gas diffused. The 10 C.C. of gas were accordingly very slowly diffused through the apparatus, the operation extending altogether over 12 hours, The first 6.5 C.C. collected were rejected, and the last 3.5 C.C.wereanalysed, with the following results : Volume of gas taken. .......................... 27-95 ,, oxygen added ..................... 488.35 Volume after explosion.. ...................... 450.9 ,, treatment with KOH ...... 398.5 Absorption (A) = 52.4. Contraction (0) = 65.4. Ratio CIA= 1.25. VOL. LXXIX. 4 c1062 THE DIRECT UNION OF CARBON AND HYDROGEN. PART IT. The result is, we are inclined to think, quite decisive. The evidence is overwhelmingly in favour of the view that the second saturated hydrocarbon formed when the electric aro is maintained between carbon terminals in an atmosphere of hydrogen is ethane, Discussim of Results. Our experiments prove that methane, the simplest of all hydrocar- bons, is the first to be formed by the direct combination of its elements, for it alone is produced at the lower temperature of 1200O when the velocity of combination is just measurable. At Some temperature, a t present unknown, between 1200O and 3600O the temperature of the arc, acetylene, and ethane begin to be formed; in the arc the formation of all these hydrocarbons continues until a certain equi- librium between them and carbon vapour and hydrogen is established.Calculating from the results of our earlier experiments, interpreted in the light of our later work, we find that the proportions between the three hydrocarbons and hydrogen when this equilibrium is attained are somewhat as follows : Hydrogen ........................... 90-91 per cent. Acetylene ........................... 7- 8 ,, Methane............................. 1-25 ,, Ethane .............................. 0.75 ,, It is, of course, open to discussion whether acetylene and ethane are formed in the arc directly from their elements, or indirectly by the decomposition of methane. At first sight it may seem probable that methane is the first product of the combination of carbon vapour with hydrogen in the arc, a13 it undoubtedly is of the action of hydrogen upon solid carbon a t 1200O. The fact, however, that methane and ethane are produced simultaneously, and a t rates which bear a con- stant ratio to each other during the whole time the arc is maintained, indicates, we think, that ethane is formed directly from its elements, and not indirectly by the decomposition of either methane or acetylene. With regard to the acetylene, a study of the results recorded in our previous paper, particularly those of the arc experiments I1 and 111, shows that its rate of formation bears a nearly constant ratio to the rates of formation of the other two hydrocarbons, and, further, that the quantity of it present at any given moment throughout an experi- ment is always far in excess of the proportions of the other two hydro- carbons, even when these are considerably below the 'equilibrium limite.' These considerations, we think, point to the conclusion that acetylene also is formed directly from its elements. It is, however, difficult to draw any hard and fast conclusiom as to the character and sequence of the chemical changes which occur in the arc experiments,COLLIE : DECOMPOSITION OF CARBON DIOXIDE. 1063 Our results aIso open up the question of the stability of hydro- carbons at high temperatures. The recent publication of preliminary notices of work on the decomposition of varioue organic compounds, including aome hydrocarbons, by W. Ipatieff (Bur., 1901, 34, 596) and also by W. Lob (Ber., 1901, 34, 915) make it desirable that we should now indicate the lines upon which we have been working on this subject for some months, in order that unnecessary overlapping may be prevented. If, as we have shown, methane is the only hydro- carbon to be formed a t 1200' from its element, it is probable that it alone can permanently exist at this temperature ; we are, therefore, investigating the decomposition of methane, ethane, ethylene, acetylene, and certain aromatic hydrocarbons at 1200' or t.hereabouts, and hope shortly to communicate some results to the Society. We have much pleasure in expressing our indebtedness to Professor Dixon for valuable criticisms at various times during the course of the research, to Professor Ramsay for so kindly helping us in the separa- tion of the saturated hydrocarbons obtained in the arc experiments, to Messrs. Johnson and Matthey for the loan of palladium required for the hydrogen absorptions, and finally to the Government Grant Committee of the Royal Society for repeated grants which have enabled ug to purchase the special apparatus required for the research, including that used for the gas analyses involved. THE OWENS COLLEGE, MANCHEGTER.