jThe Decomposition of Hydrocarbons 
 
 and 
 
 The Influence of Hydrogen in 
 ! Carbureted Water Gas Manufacture 
 
 DISSERTATION 
 
 SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS 
 
 FOR THE DEGREE OF DOCTOR OF PHILOSOPHY 
 
 IN THE FACULTY OF PURE SCIENCE 
 
 COLUMBIA UNIVERSITY IN THE 
 
 CITY OF NEW YORK 
 
 BY 
 
 Eugene Hendricks Leslie, B.S. 
 
 NEW YORK CITY 
 1916 
 
The Decomposition of Hydrocarbons 
 
 and 
 
 The Influence of Hydrogen in 
 Carbureted Water Gas Manufacture 
 
 DISSERTATION 
 
 SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS 
 
 FOR THE DEGREE OF DOCTOR OF PHILOSOPHY 
 
 IN THE FACULTY OF PURE SCIENCE 
 
 COLUMBIA UNIVERSITY IN THE 
 
 CITY OF NEW YORK 
 
 BY 
 Eugene Hendricks Leslie, B.S, 
 
 NEW YORK CITY 
 1916 
 
rt 
 
 to 
 
 V 
 

 ACKNOWLEDGMENTS 
 
 The author wishes to express to Professor Milton C. 
 Whitaker his sincerest gratitude for his many sugges- 
 tions and helpful criticisms, and for his encouragement 
 and cooperation during the progress of this work. 
 
 To Professor Floyd J. Metzger, Professor Samuel A. 
 Tucker and Dr. Clive M. Alexander the author also 
 wishes to extend thanks for their help and suggestions. 
 
 EUGENE H. LESLIE 
 
 DEPARTMENT OP CHEMICAL ENGINEERING 
 COLUMBIA UNIVERSITY 
 
 342654 
 
THE DECOMPOSITION OF HYDROCARBONS AND THE 
 
 INFLUENCE OF HYDROGEN IN CARBURETED 
 
 WATER GAS MANUFACTURE 
 
 I HISTORICAL REVIEW 
 
 The task of reviewing and assembling the large 
 amount of material which has been written on the re- 
 actions of the hydrocarbons under the influence of 
 heat is in no sense an easy one. On account of the 
 variety of the materials which have been worked 
 upon, the extreme complexity of the changes which 
 take place in any case, the differences in the types of 
 apparatus used, and the apparent inclination of many 
 writers to allow the reader to do the greater part of 
 the interpretation of the results, which in many cases 
 is well nigh impossible, the presentation of this ma- 
 terial in condensed form is accompanied by readily 
 appreciated difficulties. Such a summary must neces- 
 sarily confine itself to the work of those who seem 
 to have added most to the knowledge in this field. 
 
 In the discussion which follows the topics taken 
 up will be: 
 
 I The work relative to the primary decomposi- 
 tion of high molecular weight paraffin and naphthene 
 hydrocarbons. 
 
 II The various ideas in regard to the mode of re- 
 action of the products of the primary decomposition. 
 
 Ill The pyrogenic reactions of the simpler com- 
 pounds such as methane, ethane, ethylene, and sim- 
 ilar hydrocarbons. 
 
 IV Aromatic hydrocarbons. 
 
 V The influence of hydrogen on these reactions. 
 
 VI The transfer of heat in gas machines. 
 
 VII Our own experiments. 
 
 I PRIMARY DECOMPOSITION 
 
 It is evident that when a high molecular weight 
 paraffin splits up into two simpler molecules both 
 of these cannot be saturated hydrocarbons. An 
 olefin and a paraffin result. 
 
 It might be thought that two olefins and hydrogen 
 
 d) 
 
>e- formed. If this were the case the decomposi- 
 tion of pehtane would be represented: 
 
 5X112 ^ C2A14 ~T Carle ~f" H2 
 
 If this were the manner of the decomposition the 
 gases arising should contain at least 33 per cent of 
 hydrogen whereas we know that low temperature 
 gases contain very little hydrogen. 
 
 The other possibility that two paraffins might re- 
 sult with simultaneous separation of carbon is also 
 not the case as can be seen from the analysis of low 
 temperature oil gases shown graphically elsewhere 
 in this paper, and also from the fact that low tempera- 
 ture oil cracking produces, only very small amounts 
 of carbon. That such a change is possible at low tem- 
 peratures, however, in a hydrocarbon system of this 
 sort can be seen from the fact that McAfee 1 in the 
 aluminum chloride catalytic process obtained low 
 boiling* saturated hydrocarbons with none of the un- 
 saturated hydrocarbons. The carbon separated out 
 in the form of a granular coky mass. Even when 
 the starting material was unsaturated the final prod- 
 ucts were saturated hydrocarbons and carbon. That 
 .the course of the reaction in the presence of aluminum 
 chloride is different from that of the thermal de- 
 composition at low temperatures is obvious. 
 
 That the primary decomposition is actually a break- 
 ing down to paraffin and olefin has been well estab- 
 lished by the work of investigators which will be cited 
 below, and is borne out by our own experiments which 
 will be discussed later. 
 
 The next question which naturally arises is where 
 does the long paraffin chain split, in the middle or 
 near the -end? And if the rupture is near the end 
 which is the product of low molecular weight, the olefin 
 or the paraffin? In the discussion of these questions 
 the work of several investigators will be cited. 
 
 Vohl 2 believed that the higher paraffins decomposed 
 primarily into a paraffin of high molecular weight 
 and an olefin with few carbon atoms. His views, 
 however, have not been accepted and are not in ac- 
 cordance with the greater portion of the experimental 
 evidence recorded in the literature. 
 , Thorpe and Young 3 heated solid paraffins in sealed 
 tubes, and found that under the combined influence 
 
 , J. Ind. Eng. Chem., 1 (1915), 737-741; Met. Chem. Eng., 13 (1915). 
 592-597. 
 
 2 Dingler's polytech. J., 177, 69. 
 
 s Proc. Roy. Soc., 11, 184-201. 
 
 (2) 
 
of heat and pressure a mixture of olefins and paraffins 
 of lower boiling point was obtained. They believed 
 this mode of decomposition to be general, but were 
 unable to bring forward experimental evidence as to 
 whether the change gave rise to an olefin of low molec- 
 ular weight with a larger paraffin residue, or vice versa. 
 
 Prunier 1 found that butylene, propylene, ethylene, 
 and some crotonylene were formed when light petro- 
 leum vapors were passed through a glowing tube. 
 
 H. E. Armstrong 2 reported the finding of consid- 
 erable quantities of amylene and hexylene along with 
 aromatic hydrocarbons in the compression liquids 
 from Pintsch gas manufacture. Paraffins were not 
 present. A gaseous compound was dissolved in 
 these liquids which gave a bromide of the formula 
 C 4 H 6 Br4, and which, according to Armstrong, was 
 probably methylallene, CH 3 CH = C = CH 2 . 
 
 Brochet 3 identified normal butylene, normal amylene, 
 normal hexylene, and piperylene in the compression 
 liquids from Pintsch gas manufacture. 
 
 Lewes 4 found hexylene and heptylene in the tar 
 obtained by the cracking of Russian oil. 
 
 Brooks 5 examined a compression liquid from the 
 Pintsch gas process and found it to contain 48.0 per 
 cent of unsaturated products removable by cold con- 
 centrated sulfuric acid. The rest was chiefly ben- 
 zene. Experiments carried out by him showed that 
 solar oil when treated at temperatures between 430 
 C. and 600 C. yielded 3 to 20 per cent of gasoline 
 boiling below 150 C. This product was highly un- 
 saturated. Treatment with cold concentrated sul- 
 furic acid caused a 20 per cent loss. When nickel was 
 used along with the oil more gas was produced and 
 the product contained more unsaturated hydrocar- 
 bons. 
 
 Armstrong and Miller 6 examined the liquid products 
 which resulted when oil gas was compressed, and 
 found that it contained considerable quantities of 
 amylene and its next two higher homologs, but not the 
 corresponding paraffins. This apparently is evidence 
 in favor of the view that the initial decomposition 
 of the oil is such as to give rise to a high molecular 
 weight olefin and a low molecular with paraffin; 
 
 Ber., 6 (1873), 72. 
 
 /. Soc. Ghent. Ind., 3 (1884), 462-468. 
 
 Compt. rend., 114 (1892). 60. 
 
 J. 'Soc. Chem. Ind., 11 (1892"), 584. 
 
 J. Frank. Inst., 180 (1915), 653-673. 
 
 J. Chem. Soc., 49 (1886), 74-93. 
 
 (3) 
 
for if paraffin and olefin of nearly the same molecular 
 weight had been formed we should expect that a pre- 
 ponderance of paraffins would have been found in 
 the liquid resulting from the compression. It is well 
 known that the paraffins with carbon chains of three 
 atoms or over are more stable with respect to the in- 
 fluence of heat than the corresponding olefins. 
 
 Norton and Andrews 1 passed vapors of hexane and 
 pentane through glass or porcelain tubes 15 mm. 
 in diameter and 70 cm. long heated to a bright red 
 heat. In their experiments on hexane a liquid col- 
 lected in the receiver which was approximately 10 
 per cent by weight of the hexane used, and which 
 contained unchanged hexane, hexylene, amylene, 
 and a little benzene. The hexylene and amylene 
 were not present in large amount, but this would 
 not be expected when the temperature of their experi- 
 ments is taken into consideration. They found large 
 amounts of propylene and ethylene which they col- 
 lected by passing the gases through bromine, and these 
 hydrocarbons no doubt arose from the secondary 
 decomposition of amylene. 
 
 J. F. Tocher 2 destructively distilled octane and dec- 
 ane. He says in the discussion of his results that they 
 "bear out that at low temperatures octane and dec- 
 ane are decomposed into ethylene and higher olefins, 
 methane, and hydrogen, while at higher tempera- 
 tures no higher olefins are formed, the gaseous prod- 
 ucts being simply ethylene, methane, and hydrogen." 
 Thus again emphasis is laid on the presence of higher 
 olefins without the corresponding paraffins. 
 
 Haber 3 studied the decomposition of normal hexane 
 believing that its mode of decomposition was similar 
 to that of the still higher paraffins. He vaporized the 
 hydrocarbons and passed the vapors through a tube 
 1 /s in. in diameter and 25 in. long heated to tem- 
 peratures at intervals between 448 and 800 C. 
 The low temperature was used to avoid the secondary 
 decomposition which plays so important a part at higher 
 temperatures. The hexane was not appreciably de- 
 composed at 518 C. At 600 to 700 C. the gaseous 
 products contained slightly over 50 per cent olefins, 
 34 to 37 per cent saturated hydrocarbons (of which 
 approximately 70 per cent was methane) and 10 to 
 
 1 Am. Chem. J., 8 (1886), 1-9. 
 
 2 J. Soc. Chem. Ind.. 13, 231-237. 
 
 3 J. Gasbel, 39 (1896), 377-382, 395-399, 435-439, 452-455, 799-805, 
 813-818, 830-834; Ber., 29 (1896), 2691-2700. 
 
 (4) 
 
13 per cent hydrogen. From a number of such ex- 
 periments Haber concluded that the primary decom- 
 position took place largely in accordance with the 
 equation C 6 Hi 4 < > C 5 Hi -f- CH 4 , and from the re- 
 lations between the amounts of gaseous and liquid 
 olefins, and the mean molecular weight of the olefins 
 in the gaseous state he was led to believe that a por- 
 tion of the amylene decomposed into propylene and 
 ethylene thus: C 5 Hi "7"** C 3 H 6 + C 2 H 4 . In any case 
 Haber was convinced that the primary decomposi- 
 tion of the higher fatty hydrocarbons involved the 
 splitting off of hydrocarbons of less than three carbon 
 atoms with the formation of practically no hydrogen. 
 The higher the temperature the greater the tendency 
 for the splitting off to take place near the end of the 
 chain. 
 
 Kramer and Spillker 1 distilled a heavy hydrocar- 
 bon oil at 450 C. and under 20 atmospheres pressure. 
 Considerable gas was formed which contained 20 
 per cent olefins and 80 per cent methane. The liquid 
 distillates contained large amounts of olefins. These 
 observations are in accordance with the view that 
 the chief initial reaction is a splitting into high molec- 
 ular weight olefins and low molecular weight paraffins. 
 
 Engler 2 believes that the decomposition of high 
 molecular weight paraffins begins at 200 to 250 C. 
 at ordinary pressure. Still residues which he has 
 examined were composed largely of olefins and their 
 polymerization products. In his study of the crack- 
 ing of a heavy Galician oil residue he found that the 
 amount of olefins in the cracked distillate was high- 
 est in the high boiling fractions, and that the higher 
 the cracking temperature the higher the per cent 
 of olefins. It should be remembered that the tempera- 
 tures used were never high, for the liquid phase was 
 always present. In "Das Erdol," page 567, Engler 
 expresses the belief that at ordinary pressure the effect 
 of heating is to split off low molecular weight hydro- 
 carbons from the end of the chain, while with heating 
 under pressure the tendency is for the long chain 
 to split in the middle. That there might be some 
 such difference can be seen from a consideration of 
 the application of Le Chatelier's principle to this 
 case. 
 
 Worstall and Burwell 3 in a study of the hydrocar- 
 
 i Ber., 33 (1910), 2205. 
 5 "Das Erdol," p. 566. 
 Am. Chem. J., 19 (1897), 815-845. 
 
 (5) 
 
bons of the tars from the decomposition of heptane 
 and octane in Pintsch gas retorts found that though 
 the tars contained unchanged heptane and octane, 
 they contained none of the lower paraffins. Unsatura- 
 ted compounds composed a third of the tars. These 
 observations appear as an excellent corroboration 
 of the idea that the primary decomposition is in large 
 part a splitting off of a low molecular weight paraffin 
 such as methane. Otherwise propane, butane, pen- 
 tahe, and hexane should have been found in these 
 tars, for these hydrocarbons are more stable under 
 the influence of heat than the corresponding olefins 
 or the original heptane and octane. 
 
 Hempel 1 passed oil vapors through an iron tube 
 at temperatures below 700 to 800 C. and concluded 
 from a consideration of his gas analyses that the groups 
 splitting off were chiefly those of one or two carbon 
 atoms. The decomposition into methane and higher 
 molecular weight olefins must predominate because 
 of the high content of methane in the gases obtained. 
 Hempel's views are essentially in accord with Haber's. 
 
 Hall 2 brought out well the dependence of the mode 
 of primary decomposition of a long chain paraffin 
 hydrocarbon on the temperature and pressure. He 
 found that increase of pressure greatly diminished the 
 yield of fixed gases, *. e., methane, etc. Tempera- 
 ture was of great importance. At 560 C. a change 
 of 20 made a difference of 50 per cent in the gas pro- 
 duction, and a difference in the gravity of the liquid 
 boiling below a certain point, the greatest difference 
 being in the unsaturated portion. A 40 C. tempera- 
 ture change made as much as 40 per cent difference in 
 the unsaturated portion. Distillates containing from 
 30 per cent to 90 per cent unsaturated hydrocarbons 
 can be produced by changing no other condition than 
 .the temperature. These facts mean that the hydro- 
 carbon chain can break at any point. The reactions 
 which actually take place depend on the v conditions. 
 It is natural to suppose that increase of pressure would 
 tend to give precedence to those reactions whose prod- 
 ucts occupy the least volume, i. e., a splitting near 
 the middle of the chain giving liquid products. The 
 large increase in the percentage of olefins in the liquid 
 products with rise of temperature indicates that the 
 higher temperatures favor decomposition near the end 
 of the chain, for otherwise the paraffins which are more 
 
 1 J. Gasbel. 53 (1910), 53-58, 77-83, 101-105, 137-141, 155-165. 
 
 2 Gas World, 62 (1915), 187, el seq. 
 
 (6) 
 
stable than the corresponding olefins would predom- 
 inate in the liquid products. 
 
 Ipatiew and Dowgelewitsch 1 found that hexane 
 was decomposed at 710 C. yielding 40 to 50 per cent 
 hydrocarbons of the C M H 2n series, 8 to 14 per cent 
 hydrogen, and 40 per cent paraffins. Hexamethylene 
 heated for some time at 500 to 510 C. and under 70 
 atmospheres pressure gave a quantity of gas which 
 was largely methane and hydrogen. The liquid 'prod- 
 ucts contained considerable amounts of olefins, and 
 also methyl cyclopentane. 
 
 It seems that the mode of the pyrogenic decomposi- 
 tion of the naphthenes is not so different from that 
 of the paraffins, with a few added complications such 
 as rearrangement to the more stable five-carbon ring 
 and dehydrogenation to aromatics. 
 
 Burgess and Wheeler 2 found a considerable evolu- 
 tion of higher olefins from coal up to 350 C. They 
 believed that these hydrocarbons arose from the de- 
 composition of a paraffin or other similar long chain 
 hydrocarbon. 
 
 Brooks, Bacon, Padgett, and Humphrey, 3 in dis- 
 cussing the cracking of higher hydrocarbons, pointed 
 out that the fatal difficulty with all processes at- 
 tempting to produce gasoline from the higher hydro- 
 carbons was the large percentage of olefins contained 
 in the product. According to these authors this fig- 
 ure will vary from 20-50 per cent, depending on the 
 temperature at which the oil is cracked. An inter- 
 esting observation which bears on the mechanism 
 of the primary decomposition was that the gases evolved 
 during the distillation of paraffin under 100 Ibs. pressure 
 at temperatures between 350 and 450 C. consisted 
 of 7562 per cent saturated hydrocarbons, 25 to 37 
 per cent olefins, with o.o to 3 per cent hydrogen. 
 The saturated hydrocarbons were not examined, but 
 as only the lower paraffins are gaseous, and as it is 
 common experience that, where the gaseous paraffins 
 are formed at all, methane always predominates it is 
 not unfair to assume that a large proportion of the 
 gases obtained here was methane. 
 
 Pressure would be unfavorable to such a reaction 
 as 
 
 C n H 2n + 2 > C tt _,H 2M _ 2 + CH 4 . 
 
 Nevertheless such reactions constitute the most im- 
 
 1 Ber., 44 (1911), 2987-2992. 
 
 2 J. Chem. Soc., 105 (1914), 131-140. 
 
 J. Ind. Eng. Chem., 7 (1915), 180-185. 
 
 (7) 
 
portant chemical change taking place when these 
 higher hydrocarbons are heated even at 100 Ibs. 
 pressure and only 450 C. The conclusion then must 
 be that the change in concentration on account of 
 increased pressure has a more important effect than 
 the shifting of the equilibrium point of the reaction. 
 
 The most exact work on the separation of the indi- 
 vidual hydrocarbons resulting from the decomposi- 
 tion of the hydrocarbons of high molecular weight is 
 that of Burrell, Seibert, and Robertson. 1 By a series 
 of careful fractionations at low temperatures these 
 investigators have determined the percentages of the 
 various components in Pittsburgh coal gas and water 
 gas. The following table shows their results: 
 
 CARBURETED 
 
 COMPONENT COAL GAS WATER GAS 
 
 Carbon dioxide 1.4 4.8 
 
 Oxygen 0.7 (a) 
 
 Carbon monoxide 7.9 29 .8 
 
 Hydrogen 50.6 32.0 
 
 Methane 31.1 13.1 
 
 Ethane 0.9 2.9 
 
 Ethylene 2.1 9.8 
 
 Propylene 0.3 2.8 
 
 Butylene 0.1 1.7 
 
 Butane 
 
 Propane 0.3 
 
 Benzene(fc) 1.4 15 
 
 Nitrogen 3.5 1.3 
 
 Total 100.0 per cent 100. per cent 
 
 (a) Results calculated to air-free basis. 
 
 (b) Benzene or vapor having an inappreciable pressure at 78 C. 
 
 The notable feature of these results, so far as their 
 bearing on the primary decomposition of hydrocar- 
 bons is concerned, lies in the fact that the percentages 
 of olefins were always higher than the percentages 
 of paraffins of the same number of carbon atoms. 
 Thus in water gas there was 2 . 8 per cent propylene, 
 but only o . 3 per cent propane and i . 7 per cent butyl- 
 ene, but only traces of butane. When the relative 
 stability of these hydrocarbons is kept in mind, and 
 also the temperatures at which these gases were made, 
 these results appear as good evidence that tinder water 
 gas carbureter and superheater conditions the splitting 
 off of low molecular weight paraffins with formation 
 of high molecular weight olefins is the chief primary 
 reaction. 
 
 Egloff and Twomey 2 have cracked a gas oil de- 
 rived from Pennsylvania crude petroleum at atmos- 
 pheric pressure, and at temperature intervals of 50 
 C. between 450 C. and 800 C. In the cracked oils 
 it was found that the per cent of unsaturated com- 
 
 1 U. S. Bur. of Mines, Tech. Paper, 104. 
 
 2 Met. Chem. Eng., 14 (1916), 247-250. 
 
 (8) 
 
pounds increased with the temperature. In the dis- 
 tillation cut to 150 C. the per cent of unsaturateds 
 reached a maximum of 44 per cent at 550 C. and de- 
 creased to practically o per cent at 800 C. Thefrac- 
 tionation of the cracked oils was made in six cuts: 
 95to 120 C., 120 to 150, i7oto 230, 230 to 270, 
 and 270 to tar. In the first of these cuts it was 
 found that the percentage of unsaturateds reached 
 a maximum at 550 to 600 C. while in the latter 
 three cuts there was a steady increase with tempera- 
 ture in the proportion of unsaturateds. The authors 
 believed that the decrease in olefin formation might 
 be accounted for by polymerization of olefins to 
 naphthenes which in turn go to aromatics, and secondly 
 direct polymerization of olefins to aromatics. An- 
 other factor, no doubt, contributes in large degree to 
 the observed diminution in the per cent of olefins in 
 the low boiling cuts with increase in temperature, 
 namely decomposition of the olefins to gaseous prod- 
 ucts of lower molecular weight. This is discussed 
 more at length elsewhere in this paper. 
 
 The results of Egloff and Twomey are just those 
 which would be expected if it is considered that at 
 low temperatures the tendency of the higher molec- 
 ular weight paraffins is to split into paraffin and olefin 
 of nearly equal chain, while at higher temperatures 
 the primary decomposition is into high molecular 
 weight olefin and low molecular weight paraffin. 
 
 II SECONDARY DECOMPOSITION OF HYDROCARBONS 
 
 Two classes of secondary reactions must be taken 
 into account, first those which result in the breaking 
 down of the hydrocarbon molecules to hydrocarbons 
 of lower molecular weight, and second those synthetic 
 reactions which give rise to more complex but more 
 stable hydrocarbons. 
 
 In reviewing the work bearing on the primary de- 
 composition of paraffin hydrocarbons of high molecu- 
 lar weight it has been shown that the greater portion 
 of the experimental evidence points to a splitting of 
 the carbon chain with formation of olefin and paraffin, 
 and that conditions determine where the rupture 
 takes place low temperature and high pressure 
 tending to favor the splitting near the middle of the 
 chain while at lower pressures and higher tempera- 
 tures the breaking off of low molecular weight hydro- 
 carbons, such as methane, ethane, and ethylene, 
 but particularly methane and ethane, becomes the 
 important reaction. 
 
 (9) 
 
The members of the paraffin series down to butane 
 in all probability follow some such mode of reaction 
 as this. The first problem to be considered in the 
 study of the secondary reactions, then, is the fate 
 of the high molecular weight olefins which arise. 
 
 If, for the sake of simplicity, pentylene, CH 2 CH 2 - 
 CH 2 CH = CH 2 , be regarded as typical of the higher 
 olefines we may list a few of the possible reactions as 
 follows: 
 
 (1) CH 3 CH 2 CH 2 CH=CH 2 ^ 
 
 CH 3 CH=CH 2 + CH 2 =CH 2 
 
 (2) CH 3 CH 2 CH 2 CH=CH 2 ^1 .. 
 
 CH 4 .+ CH 3 C^C CH 3 
 /CH 2 CH 2 
 
 (3) CH 3 CH 2 CH 2 CH=CH 2 ^ CH/ | 
 
 \CH? CH 2 
 
 (4) CH 3 CH 2 CH 2 CH=CH 2 + H 2 ^ 
 
 CH 3 CH 2 CH 2 CH 2 CH 3 
 
 (5) Condensation 
 
 It is not to be imagined that any one of these reac- 
 tions is the sole reaction, nor is this to be regarded as 
 in any sense a complete list of the possible concurrent 
 reactions. The extent to which any or all of these 
 changes take p'ace is dependent on the temperature, 
 pressure and concentrations. It is in an effort to dis- 
 cuss the extent to which each comes into play, and 
 the effect of varying conditions that the work is re- 
 viewed bearing on this phase of the subject and that 
 later our own experiments are considered. 
 
 Engler and Routala 1 heated 350 g. of amylene in 
 sealed glass tubes at 325 C. for 32 days. About ten 
 liters of gas were produced of which over 90 per cent 
 was paraffin, 7 . 5 per cent hydrogen, and the rest 
 olefins. The liquid product contained a great variety 
 of hydrocarbons of the paraffin and polymethylene 
 series. Also the higher boiling portions contained 
 compounds of lower hydrogen content than the cyclo- 
 paraffins. 
 
 The presence of the gaseous paraffin compounds 
 and the high boiling compounds with more than one 
 double bond can be explained if it is assumed that 
 the amylene splits into methane and divinyl, CH 2 = 
 CH CH=CH 2 , or into methane and crotonylene, 
 reactions analogous to (2) above. 
 
 CH 3 v 
 
 V=CH CH 3 ^ CH 4 + CHs C=C CH 3 
 CH/ Crotonylene 
 
 Ber., 42 (1909), 4620-31. 
 
 (10) 
 
The divinyl, or crotonylene, under the influence of 
 heat and pressure, might easily polymerize to high 
 molecular weight compounds of low hydrogen content. 
 
 The naphthenic compounds in all probability re- 
 sulted from the polymerization of the decomposition 
 products of amylene, such as ethylene and propylene 
 or the substituted ethylenes and propylenes; or by a 
 rearrangement of the amylene molecule as in reac- 
 tion (3). 
 
 Engler 1 stated that long chain olefins split easily 
 into ethylene. The importance of reactions result- 
 ing in higher molecular weight hydrocarbons of lower 
 hydrogen content was emphasized. Naphthenes were 
 found in the product from the heating of still residues 
 at 300 C., and since these compounds were not pres- 
 ent in the still residues originally, Engler concluded 
 that they were formed by polymerization of ethylenes. 
 He was of the opinion that direct closing of the olefin 
 chain did not take place, but this was not proved by 
 experimental evidence. 
 
 Worstall and Bur well 2 have worked on oil gas tars 
 in an effort to determine whether or not naphthenes 
 were present. They mentioned that the absence of 
 naphthenes in residues from nitration or sulfonation 
 is not evidence that there were no naphthenes in 
 the original tar, for the naphthenes nitrate to nitro- 
 hydrobenzenes. Such compounds have been ob- 
 tained by the use of dilute nitric acid. 3 From distilla- 
 tion methods and specific gravity tests the authors 
 concluded that their oil gas tars contained no naph- 
 thenes. 
 
 It is to be kept in mind here that oil gas in the Pintsch 
 gas process is made at temperatures of 850 to 900 C.,. 
 and it is probable that this accounts for the differ- 
 ence between the results of these authors and those of 
 Engler whose work was carried out at lower tempera- 
 tures; i. e., between 300 and 400 C. 
 
 Haber, 4 on the other hand, was convinced by a study 
 of the combustion data of oil gases made from hexane 
 at temperatures of 600 C. and above that naphthenes 
 were present. He also believed that crotonylene was 
 formed in small amounts by the splitting off of methane 
 from amylene. Emphasis was laid on the synthetic 
 reactions such as the building up of high molecular 
 
 1 Ber., SO (1897), 2908. 
 
 2 Am. Chem. J ., 19 (1897), 815-845. 
 
 3 Ber., 25 (1892), 107-108; 28 (1895), 577-578. 
 
 < J. Gasbel, 39 (1896), 377-382, 395-399, 435-439, 452-455, 799-805. 
 813-818, 830-834. 
 
 (II) 
 
weight compounds by the combination of "free un- 
 saturated residues." 
 
 H. E. Armstrong 1 found a gas dissolved in the com- 
 pression liquids from oil gas manufacture which gave 
 a bromide of the composition C 4 H6Br 4 , and which he 
 believed to be methylallene, CH 3 CH=C=CH 2 . 
 
 Armstrong and Miller, 2 in the examination 
 of similar liquids, found olefins of the type C M H 2M _ 2- 
 Vinyl ethylene, CH 2 =CH CH=CH 2 , was identified. 
 
 A. Harzer 3 commented on the fact that there was 
 a considerable proportion of diolefins among the 
 hydrocarbons of coal and coal tar. 
 
 Jones and Wheeler 4 made note of their observation 
 that one-fifth by weight of the tar formed in their 
 vacuum distillations of coal at low temperatures 
 was composed of olefins, the greater part of which 
 was richer in carbon than the mono-olefins. This 
 appears an excellent evidence that the decomposi- 
 tion of the higher olefins into methane and diolefins 
 or substituted acetylenes is really one of the reactions 
 in the course of the breaking down of these long chain 
 olefins; for the conditions used here, i. e., low tempera- 
 ture and low pressure, are those which tend to pre- 
 serve intermediate products, and not those which 
 would be likely to give rise to diolefins through syn- 
 thetic reactions. 
 
 W. A. Noyes, W. M. Blinks, and A. V. H. Morey 5 
 made oil gases at comparatively low temperatures 
 in a checkerbrick filled machine. The gases were 
 analyzed for olefins by passing them through bromine 
 water, and subsequently fractionating the bromides 
 formed. In this manner they found that of the 28. i 
 per cent olefins present 16.2 per cent was ethylene 
 and 11.9 per cent propylene. 
 
 Norton and Andrews, 6 in their work on the pyro- 
 genic decomposition of hexanes and pentanes, passed 
 the gases through bromine dissolved in carbon disul- 
 fide. The liquid bromides so formed were found to 
 be ethylene and propylene dibromides in nearly equal 
 proportion. This observation and that of Noyes, 
 Blinks, and Morey, is entirely in accord with the 
 belief that reactions such as the first one mentioned 
 in the list above take place, 
 
 1 J. Soc. Chem. Ind., 3 (1884), 462-468. 
 
 2 /. Chem. Soc., 49 (1886), 74-93. 
 
 3 Gas World, 59 (1913). 405. 
 
 * /. Chem. Soc., 105 (1914), 140-151, 2562-2565. 
 
 J. Am. Chem. Soc., 16 (1894), 688-697. 
 s Am. Chem. J., 8 (1886), 1-9. 
 
 (12) 
 
CsHio ^ CsHe -f- CoH4 
 
 i. e., a splitting of higher olefins into lower olefins. 
 Norton and Andrews also found that there was formed 
 a solid bromide of the composition C 4 H 6 Br 4 , which was 
 evidently the bromide of divinyl or crotonylene rather 
 than ethyl acetylene, since the hydrocarbons passed 
 through a solution of ammoniacal cuprous chloride 
 before being led into the bromine solution. Ethyl 
 acetylene would have formed a copper, compound, 
 and thus have been removed. 
 
 That the higher olefins are intermediate between 
 the paraffins and the low molecular weight gaseous 
 compounds can be seen by a study of some further 
 work of Norton and Andrews. When hexane was 
 passed through a 15 mm. diameter glass tube at 700 
 C., and the products collected it was found that a 
 large portion of the hexane passed through unchanged. 
 But the liquid which was collected also , contained a 
 mixture of butylene, amylene, and hexylene. The 
 bromides obtained by passing the gases through 
 bromine in carbon disulfide were propylene dibromide, 
 crotonylene dibromide, but no ethylene dibromide. 
 At higher temperatures the decomposition of isohexane 
 gave, beside the liquid products, gaseous hydrocar- 
 bons which proved to be largely propylene and croton- 
 ylene with smaller amounts of ethylene. Similar 
 results were obtained when pentane was pyrogenically 
 decomposed^. 
 
 The statement as to the absence of ethylene in the 
 700 C. gases from hexane should, as judged from 
 general experience, be accepted with reservations. 
 However, it must be kept in mind that 700 C. as 
 measured and applied to a certain kind and size of 
 apparatus may give results similar to those of very 
 different temperatures in a different apparatus used 
 by another observer. 
 
 Lewes 1 decomposed Russian oil in a short retort 
 at 500 C. in order to study the early stages of the de- 
 composition of hydrocarbons. He found the paraffins 
 and olefins present in nearly equal proportions in the 
 gases. Increase of temperature diminished the per 
 cent of olefins. As the temperature went up the 
 paraffins, too, reacted secondarily, but this was with 
 formation of methane so that the actual percentage 
 of paraffin in the gas did not decrease, but increased. 
 
 Hempel 2 was evidently of the opinion that the 
 
 1 J. Soc. Chem. Ind., 11 (1893), 584-590. 
 
 J. Gasbel, 53 (1910), 53-58, 77-83, 101-105, 137-141, 155-165. 
 
 (13) 
 
higher olefins changed rather easily into ring com- 
 pounds, and pointed to the considerable quantity of 
 these latter compounds which were contained in oil 
 gas tars. His experimental work showed that when 
 oils were cracked in atmospheres of hydrogen con- 
 siderable hydrogen was actually absorbed, and that 
 the yields of methane, ethane and ethylene per unit 
 weight of oil were larger. His observations are dis- 
 cussed more at length elsewhere in this paper, but it 
 should be noted here that hydrogenation of these 
 higher olefins is not an improbable reaction. 
 
 W. Ipatiew 1 studied the polymerization of ethylene 
 and some of its homologs. At 380 to 400 C. and 
 70 atmospheres pressure the polymerization of ethylene 
 is very rapid. The lower boiling fractions of the 
 products contained paraffins, and the higher boiling 
 fractions olefins and polymethylenes. No benzene 
 hydrocarbons were obtained. The higher fractions 
 also contained some hydrocarbons of higher carbon 
 percentages and lower hydrogen content than the 
 C 2 H 2w series. Isobutylene yielded products analogous 
 to those obtained from ethylene except that the hydro- 
 carbons richer in carbon than the mono-olefins were 
 not present. Ipatiew believed that condensation to 
 hexamethylene was the first step in all these reac- 
 tions. 
 
 The work of Burrell, Seibert and Robertson 2 has 
 already been referred to in our discussion of the primary 
 decomposition of hydrocarbons. Further valuable 
 conclusions can be drawn from it, however, which 
 concern the changes undergone by the high molecular 
 weight olefins. The authors found that Pittsburgh 
 carbureted water gas contained 15.9 per cent illumi- 
 nants, and Pittsburgh coal gas 3 . 9 per cent illuminants. 
 The following table shows the proportions of the 
 various components which go to make up these totals: 
 
 COMPONENT 
 Ethylene 
 
 COAL GAS 
 53.8 
 
 CARBURETED 
 WATER GAS 
 60.9 
 
 Propylene 
 Butylene 
 
 7.7 
 2.6 
 
 17.4 
 10.6 
 
 Propane 
 Butane 
 
 
 1.8 
 
 Benzene (a) 
 
 35.9 
 
 9.3 
 
 (a) Vapors having an inappreciable pressure at 78 C. 
 
 The proportions of ethylene, propylene and butylene 
 are notable. These results may be interpreted as. 
 indicating that reactions such as C 5 Hi ~^~ > C 3 H 6 + 
 
 1 Ber., 44 (1911), 2978-2987. 
 
 2 ;U. S. Bureau of Mines, Tech. Paper, 104. 
 
 (14) 
 
C 2 H 4 actually play an important part in the breaking 
 down of the higher olefins. 
 
 No mention was made by Burrell, Seibert and Rob- 
 ertson of the naphthenes. The hexamethylene com- 
 pounds boil from 81 C. up, and might pass largely 
 into the water gas tar. A portion of the component 
 marked benzene in the above table may easily have 
 been hexamethylene. Pentamethylene, however, 
 boils at 49 C., and if it were formed in significant 
 amounts it should be found in the gases. Apparently, 
 therefore, compounds of this class are not formed in 
 large amount when the hydrocarbons undergo a de- 
 composition under the conditions of carbureted water 
 gas manufacture. 
 
 SUMMARY It seems, then, that the chief reaction 
 undergone by these high molecular weight olefins 
 is a splitting into lower molecular weight olefins. 
 A decomposition into methane and compounds with 
 two double bonds or one triple bond also takes place. 
 The intramolecular change of olefins into cycloparaffins 
 is possible, but from the evidence available it is diffi- 
 cult to state what proportion of the naphthene forma- 
 tion must be ascribed to this reaction. Hydrogena- 
 tion of olefins takes place to some extent. Poly- 
 merization of olefins to naphthenes occurs, also poly- 
 merization of the high molecular weight unsaturated 
 compounds to tarry compounds. 
 
 Ill TERTIARY DECOMPOSITION 
 
 The problem now becomes one of studying the re- 
 actions of the hydrocarbons, which are formed in the 
 primary and secondary decompositions of the paraffins. 
 These products are ethylene, propylene, diolefins, 
 acetylenes, naphthenes, methane, ethane, propane, 
 and the high molecular weight tarry compounds. 
 The discussion of the changes which these hydrocar- 
 bons undergo separately may appear fundamentally 
 wrong in view of the fact that in the hydrocarbon 
 system undergoing change these reactions are all 
 interlocked and related in a most complex manner. 
 However, it is just on account of this complexity 
 that the manner of presentation of the material which 
 follows has been chosen that is a consideration of 
 the reactions of the separate hydrocarbons. Before 
 turning to such a discussion, however, it is desired 
 to point out a few things in connection with such hy- 
 drocarbon systems as a whole. 
 
 The earliest attempt to elucidate in anything like 
 
 d5) 
 
a comprehensive manner the changes taking place 
 in a hydrocarbon system was made by Berthelot. 1 
 He heated hydrocarbons in a retort, and also by pass- 
 ing them through a tube. From rather meager data 
 on a few of the lower molecular weight hydrocarbons 
 a theory of thermal decomposition was worked out. 
 Berthelot found that ethylene and hydrogen when 
 heated in a retort gave some ethane, and that ethane 
 formed ethylene in small quantities. Ethylene was 
 formed from acetylene and hydrogen. Ethylene alone 
 gave small amounts of acetylene and ethane; and 
 liquid products of the formula of crotonylene re- 
 sulted when acetylene and ethylene were heated to- 
 gether. When the gases were passed through the 
 tube methane gave acetylene, olefins, ethane, naph- 
 thalene, and tar. From these experiments Berthelot 
 concluded that two types of reactions were at work 
 reactions of decompositions, and synthetic reactions. 
 The breaking down and building up changes com- 
 peted with each other and a complex equilibrium 
 was supposedly established. 
 
 Berthelot's work has been criticized by Haber on 
 the ground that conclusions could not be drawn re- 
 garding hydrocarbon reactions in general from a study 
 of a few of these simple hydrocarbons because they 
 constitute a special case. For instance, ethylene 
 cannot split open a carbon bond without completely 
 disrupting the molecule, whereas a long chain olefin can 
 easily break open in this fashion. Furthermore, Ber- 
 thelot heated these substances in a closed retort. These 
 conditions are obviously not those which should be 
 chosen for a study of the mechanism of the reactions,, 
 for such heating would give only those products which 
 stand long heating, and not those which might have 
 been formed transitorily. Also Berthelot's equi- 
 librium idea is criticized by Haber for the reason 
 that it would necessitate the presence of almost every 
 conceivable hydrocarbon, whereas, though the sys- 
 tem is complex enough, it does not exhibit such a 
 heterogeneous composition as this. 
 
 That Haber's criticisms of Berthelot's experimental 
 methods and of his general conclusions are well placed 
 can be seen from a consideration of recent work 
 and also from the results of the experimental work, 
 recorded in this paper. 
 
 Whitaker and Alexander 2 have shown that when 
 
 1 Ann. chim. phys., 67, iii (1863), 53; 9, iv (1866), 413, 455; 12, iv 
 (1867), 5, 122; 16, iv (1869), 143, 148, 153, 162. 
 
 2 J. Ind. Eng. Chem., 7 (1915), 484-495. 
 
 (16) 
 
paraffin hydrocarbons are cracked at temperatures 
 of 1400 to 1600 C. the olefins no longer exist, though 
 the gases still contain approximately 2% of methane. 
 If the oil rate was still further decreased the methane 
 disappeared and carbon and hydrogen were the only 
 products. 
 
 It is generally recognized that the most stable gaseous 
 hydrocarbons are methane, ethane and ethylene. 
 Ethane is formed in much smaller amount than either 
 of the other two, hence even though it may be more 
 stable than ethylene the latter persists longer in de- 
 tectable amounts because it was originally present 
 in larger proportion. Any hydrocarbon system heated 
 above 700 C. will tend toward these products, and 
 if given time these gases, along with carbon and 
 hydrogen, will be the chief products. With the lapse 
 of more time the ethane and ethylene decompose 
 with formation of methane, carbon and hydrogen, 
 which are the ultimate products. The methane 
 equilibrium is discussed elsewhere in this paper, and 
 it is seen that methane exists in quantity only at those 
 temperatures below 750 to 800 C. 
 
 A gaseous hydrocarbon system is thus not really 
 in equilibrium. Rather the condition of such a 
 system may be compared to that of a mixture of 
 hydrogen and oxygen gases at temperatures of 200 
 to 300 C. The equilibrium condition of such a sys- 
 tem is practically a complete union forming water, 
 but the speed of the reaction 2H 2 + O 2 > 2H 2 O 
 is so small at these temperatures that the system 
 might easily be thought to be in equilibrium, i. e., 
 in that condition which is independent of the further 
 passage of time. 
 
 Even though it has been established that a true 
 condition of equilibrium does not exist there is still 
 an interdependence of reactions in a hydrocarbon 
 system. This idea has been elaborated by Whitaker 
 and Rittman. 1 
 
 The reactions which form the lower molecular 
 weight hydrocarbons are in general more rapid in 
 their progress than the reactions of decomposition of 
 these lower hydrocarbons. Methane, in particular, 
 is stable under the action of heat at those tempera- 
 tures which are used in the various apparatus used 
 in the manufacture of gas. Thus those reactions 
 which result in the formation of methane and ethylene 
 reach a condition nearly corresponding to equilibrium 
 
 1 J. Ind. Eng. Chem., 6 (1914). 383-392, 472-479. 
 (17) 
 
proportions on account of the slow decomposition of 
 ethylene and methane, but the system as a whole 
 cannot be regarded as in equilibrium. 
 
 In considering the discussion of the reactions of the 
 individual hydrocarbons the effect of the presence 
 of ttie end-products of a particular reaction must 
 always be kept in mind. Also the changing concen- 
 tration conditions as the gas volume increases with 
 the progress of the changes involved must not be 
 forgotten. 
 
 REACTIONS OF METHANE 
 
 The study of the influence of heat at various tem- 
 peratures on the hydrocarbon methane has usually 
 been made with the idea in mind of finding the equi- 
 librium proportions of methane and hydrogen in the 
 system carbon-hydrogen-methane. 
 
 The equilibrium proportions are never even approx- 
 imated in experiments made after the manner of 
 those discussed in the latter part of this paper, nor 
 under the conditions maintained in the technical 
 production of coal, oil, and water gas. The preserva- 
 tion of methane is desired in all these cases, for its 
 decomposition into carbon and hydrogen means loss 
 of valuable carbon from the gas and the production 
 of gases high in hydrogen which are unsuited for dis- 
 tribution. Hence the studies of the methane equi- 
 librium are of interest only in so far as they indicate 
 the tendency of methane to decompose under certain 
 temperature conditions. 
 
 Mayer and Altmayer 1 studied the methane equi- 
 librium between 475 and 625 C., and found that' 
 below 625 C. appreciable quantities of methane 
 were stable. They deduced a mathematical expression 
 by means of which they extrapolated their experi- 
 mental results to 250 C. on one hand and 850 C. 
 on the other. Finding that at 850 C. the calculated 
 percentage of methane in equilibrium with carbon 
 and hydrogen would be only 1.59, they stated that 
 at temperatures as high as 1200 C. no methane 
 could be formed synthetically. This conclusion was 
 severely criticized by Bone and Coward 2 on the ground 
 that the low temperature experimental range would 
 not justify such wholesale extrapolation. This ap- 
 pears as a most just criticism. Bone and Coward 
 have, in fact, converted 73 per cent of a quantity of 
 
 Ber., 40 (1907), 2134-2144. 
 
 2 J. Chem. Soc., 93 (1908), 1975-1993, and 97 (1910), 1219-1225. 
 
 (18) 
 
pure carbon into methane at noo-i2oo by passing 
 hydrogen over it. The equilibrium proportion of 
 methane at these temperatures is exceedingly minute, 
 however. 
 
 Earlier work done by Bone and Jerdan 1 on the 
 synthesis of methane at 1200 C. and above, has been 
 criticized by Berthelot, 2 who did not believe that 
 methane could be synthesized at such high tem- 
 perature, if the reaction materials were pure. Bone 
 and Coward's results, however, appear to show that 
 this is not true. 
 
 Pring and Hutton 3 obtained o.io to 0.25 per cent 
 methane by electrical heating of purified carbon 
 rods in atmospheres of hydrogen at 1250 to 1350 C. 
 They also found 0.5 to 3.6 per cent acetylene formed 
 at temperatures between 1800 and 2500 C. The 
 amount of acetylene increased fairly regularly with 
 the temperature. At these same temperatures 
 amounts of methane ranging from o. 5 to i.o per cent 
 were found in the gases, being formed no doubt largely 
 by the decomposition of the acetylene, and then 
 moving out of the heated zone and escaping destruc- 
 tion. 
 
 Pring and Fairlie 4 have studied the methane equi- 
 librium. They noted that at temperatures below 
 1000 C. the union of carbon and hydrogen was so 
 slow, even with application of pressure, that the 
 equilibrium proportions could not be reached in any 
 reasonable length of time. Ethylene began to be 
 formed at 1300 C., and the quantity in equilibrium 
 at 1400 was 0.005 per cent. Acetylene was first 
 formed at 1650 C., and the amount increased with 
 the temperature at those temperatures. Where ethyl- 
 ene and acetylene are capable of formation the exact 
 determination of the methane equilibrium is impossi- 
 ble because of the decomposition of these gases into 
 methane. The amounts of methane in equilibrium 
 are: at 1200 C., 0.20 per cent; and at 1500 C., 0.07 
 per cent. 
 
 T. Holgate 6 pointed out that the curve expressing 
 the rate of decomposition of methane with rise of 
 temperature has a maximum at about 600 C. He 
 believed that this was due to the accumulation of 
 hydrogen which acted as a brake to the decomposi- 
 
 ' J. Chem. Soc., 71 (1897), 41-61. 
 5 Ann. chim. phys., 6 viii (1905), 183. 
 J. Chem. Soc., 89 (1906), 1591-1601. 
 4 Reports 8th Internal. Congress, 21, 65. 
 * J. Gas Lighting, 106 (1909). 25-28, 84-86. 
 
 (19) 
 
tion. Holgate evidently believed that the reaction 
 C + 2H 2 > CH 4 had an appreciable speed, for he 
 said that above 607 C. "the retarding action of the 
 square of the hydrogen pressure is greater than the 
 effect of heat in driving the reaction, and as a result 
 there is at 800 C. a less increase in the rate of de- 
 composition of the methane per degree rise in tem- 
 perature than there is at 400 C." This explana- 
 tion is not in accordance with the various investiga- 
 tions of the methane equilibrium which have shown 
 that the speed of the reaction C + 2H 2 > CH 4 is 
 inappreciable at temperatures less than 1000 C. 
 
 Ipatiew 1 has shown that even in the presence of 
 the catalysts NiO or reduced nickel, which have proved 
 most effective in promoting gaseous reactions, the 
 union of carbon and hydrogen could not be detected 
 at 650 C. 
 
 On the other hand, general experience shows that 
 at temperatures above 600 C. the production of gases 
 of all kinds is greatly increased, and the time of heat- 
 ing of the gaseous mixture is diminished in propor- 
 tion to this increase in volume. This accounts for 
 the fact that the hydrocarbons are not destroyed be- 
 fore leaving the coal gas retorts of the water gas 
 superheater. In thinking of the hydrocarbon systems 
 involved here it must be kept in mind that the reverse 
 reactions of these breaking down reactions have small 
 velocities, largely on account of the removal of the 
 end-products of the decompositions by other changes. 
 
 It can be seen from the researches cited that methane 
 is actually tending to decompose into carbon and 
 hydrogen at the temperatures normally operative in 
 water gas or oil gas machines or coal gas retorts. 
 That this decomposition can actually be brought about 
 on a large scale is well shown by the work of Ostromiss- 
 linski and Burshanadse, 2 who made a gas containing 
 75 to 80 per cent hydrogen and suitable for filling bal- 
 loons by passing hydrocarbon gases over nickel, nickel 
 oxide, or oxides of iron. 
 
 Brownlee and Uhlinger 3 made a gas containing 
 90 to 95 per cent hydrogen by passing natural gas 
 or the vapors of liquid or solid hydrocarbons over 
 highly heated refractory material. 
 
 More interesting from a practical or working point 
 of view than the investigations of the methane equi- 
 
 * J. prakt. Chem., 87 (1913), 479-487. 
 
 2 J. Russ. Phys.-Chem. Soc., 42 (1910), 195-207. 
 
 U. S. Pat. 1,168,931, Jan. 18, 1916. 
 
 (20) 
 
librium, are the results of those experiments wherein 
 the working conditions approximate those of actual 
 practice. 
 
 Simmersbach 1 analyzed coke oven gases which were 
 heated over refractory materials for 10 to 14 seconds. 
 The proportion of methane in these gases at 800 C. 
 was 27.6 per cent, at 1000 C. 20. 2, at 1100 C. 15.6, 
 and at 1200 C. 5. 5 per cent. All other hydrocarbons 
 disappeared at 1000 C. The stability of methane 
 was well illustrated here. It was found that the 
 decomposition of methane was practically complete 
 in 90 seconds at 1000 C. 
 
 H. Rollings and J. W. Cobb 2 conducted valuable 
 experiments on the stability of methane. Mixtures 
 of methane and hydrogen were passed through an 
 electrically heated tube. At 800 C. with a mixture 
 of equal parts of methane and hydrogen it was found 
 that only 2 per cent of the methane was decom- 
 posed on passing the gases when the time of heating 
 was one minute. At 1100 C., 65 per cent of the 
 methane in a similar mixture was decomposed when 
 the duration of the heating was 47 seconds. Thus 
 methane is not greatly affected at 800 C., but is 
 rapidly decomposed at 1100 C. 
 
 The work of Whitaker and Alexander 3 showed 
 that methane was the last gas to be decomposed at 
 temperatures from 1400 to 1600 C. With oil feed 
 rates of 10 cc. per minute the illuminants disappeared 
 at 1400 C., whereas 3. 5 per cent of methane escaped 
 destruction. 
 
 Lewes 4 passed methane through a platinum tube, 
 six inches of which was heated to 1000 C. Un- 
 saturated hydrocarbons, to the extent of 2. 7 per cent, 
 and i . 8 per cent acetylene formed. He concluded 
 that the acetylene was formed in accordance with 
 the reaction 2CH 4 > C 2 H 2 + 3H 2 . In another 
 paper Lewes 4 stated that methane when heated to 
 900 C. underwent practically no change. 
 
 Bone and Coward 5 found that at 700 C. the de- 
 composition of methane was inappreciable. In an 
 experiment at 785 C., with an open tube, they found 
 that 91.6 per cent of the methane used was unchanged 
 at the end of one hour, the rest of the gas being hydro- 
 
 > Stahl u. Risen, 33 (1913), 239-45. 
 *Gos World. 60 (1914), 879-884. 
 3 J. Ind. Eng. Chem., 1 (1915), 484-495. 
 
 J. Chem. Soc., 61 (1892), 322-338; Proc. Roy. Soc., 55 (1894), 90; 57 
 (1895), 394. 
 
 1 J. Chem. Soc., 93 (1908), 1197-1225. 
 
 (21) 
 
gen. Fjo one of the other hydrocarbons exhibited any 
 such stability at these temperatures. Experiments 
 were conducted at 1000 and at 1150 C., from the 
 results of which the authors concluded that the nor- 
 mal decomposition of methane was into carbon and 
 hydrogen. This they believed to be a surface effect 
 almost exclusively. Only very small amounts of 
 acetylene and olefins were found in the first few min- 
 utes of the heating. That acetylene or olefins were 
 not the primary products of the decompositions which 
 then undergo decomposition into carbon and hydro- 
 gen, Bone and Coward believe was proved by the 
 nature of the carbon deposit. When methane decom- 
 posed, the carbon was of a peculiarly hard and lustrous 
 variety, while acetylene and ethylene gave a soft 
 dull variety. 
 
 In summing up the work on methane it can be said 
 that the chief reaction is the decomposition into 
 carbon and hydrogen, CH 4 > C -f- 2H 2 , and that 
 the reaction suggested by Lewes and Berthelot, 
 2CH 4 > C 2 H 2 + 3H 2 , takes place to a small ex- 
 tent only. 
 
 Under the conditions of operation of a carbureted 
 water gas set very little methane is decomposed. 
 Though the temperatures of the retort walls in coal 
 gas manufacture are much higher than those in the 
 interior of the carbureter and superheater of a water 
 gas set it is obvious that the gases do not reach the 
 temperature of the refractory surfaces. So again 
 methane is not decomposed to a large extent. 
 
 REACTIONS OF ETHANE 
 
 H. Rollings and J. W. Cobb 1 passed a gas of the 
 composition 4.2 per cent ethane, 47.5 methane, 
 and 48.3 hydrogen through an electrically heated 
 porcelain tube. At 800 C., the exit gases had the 
 composition 0.9 per cent ethane, 48.9 methane, 
 48.9 hydrogen, 1.3 ethylene, and a trace of acetylene. 
 Thus the decomposition of ethane took place to an 
 extent of 79 per cent under these conditions with a 
 time of heating of 47 seconds. At 1100 C. the de- 
 composition was 88 per cent. It is evident that 
 ethane decomposes to ethylene, methane, and hydro- 
 gen. No mention was made of the separation of car- 
 bon, but on the other hand the gases passed through 
 a tube filled with coke and the separation of carbon 
 might not have been noticed. 
 
 i Gas World. 60, 879-S84. 
 
 (22) 
 
Lewes 1 passed ethane through a platinum tube, 
 and obtained in the resulting gases 19.5 per cent 
 unsaturated hydrocarbons and 8. 2 per cent acetylenes. 
 When ethane diluted with 80 per cent hydrogen 
 was heated with 1 5 per cent air, Lewes found 7 . 7 
 per cent unsaturated hydrocarbons, and 3.9 per cent 
 acetylenes. He believed that ethane first broke 
 down to ethylene and hydrogen, and that the ethylene 
 gave rise to the acetylene. 
 
 Bone and Coward 2 showed that the primary reac- 
 tion which ethane underwent when heated was a 
 dissociation into ethylene and hydrogen: C 2 H 6 < > 
 C 2 H 4 -f H 2 . In the gases from the decomposition 
 of ethane they found the ratio of the methane to the 
 hydrogen to be about 2:1, whereas if it is considered 
 that the ethylene formed as shown in the equation 
 above decomposes thus C 2 H 4 > CH 4 -f C the 
 ratio of methane to hydrogen in the gaseous products 
 from the decomposition of ethane should be i : i. 
 This can be explained only on the basis of the assump- 
 tion that some of the hydrogen enters into combina- 
 tion, forming methane or products which by their 
 subsequent decomposition give methane. Bone and 
 Coward suggested that it was due to the hydrogena- 
 tion of such residues as CH 3 , =CH 2 , =CH, 
 which can be thought of as having fugitive existence. 
 Their results in this connection are further discussed 
 under our caption " Atmospheres of Hydrogen." 
 
 However, "substances" such as CH 3 , =CH2, 
 and =CH are in a class with ''nascent" hydrogen. The 
 high methane content of these resultant gases can 
 be explained if we give credence to the possibility of 
 reactions such as C2H6 + H 2 < > 2CH 4 . This seems 
 the more probable when we call to mind that small 
 amounts of acetylene and ethylene are found in the 
 gases which result from the heating of methane as 
 was shown by Lewes. The acetylene and ethylene 
 would come from the ethane formed as in the equa- 
 tion above, v'" 
 
 Bone and Coward found only small amounts of 
 acetylene in the gases from the decomposition of 
 ethane, nor did the formation of aromatic hydrocar- 
 bons take place to a large extent. At 800 C., with 
 one minute heating,, only 17.9 per cent ethane sur- 
 vived. The decomposition of ethane was not be- 
 lieved to be a surface effect. 
 
 1 J. Chem. Soc., 61 (1892), 322-338. 
 * Ibid., 93 (1908). 1197-1225. 
 
 (23) 
 
REACTIONS OF ETHYLENE 
 
 Day, 1 in summing up the work of Fourcroy, De- 
 Wilde, Buff, Hoffman, Berthelot, Marchand, Grove, 
 and Magnus, says: "It seems clear that at the high- 
 est temperatures ethylene separates directly into its 
 elements. Below this point marsh gas and carbon 
 are obtained (C 2 H 4 > CH 4 -f- C), then marsh gas 
 and several liquid products, among them benzene, 
 styrene, etc., under certain conditions." Day circu- 
 lated ethylene through a glass tube heated to various 
 temperatures for different lengths of time. No change 
 in volume was observed with 14 hours' heating at 
 300 C. At 344 C., a contraction of one-twentieth 
 of the original volume took place in 24 hours, and Day 
 concluded that condensation had taken place, for 
 no methane or hydrogen was found in the gas. At 
 400 C., 132 cc. of ethylene contracted to 63 cc. in 
 171 hours. No hydrogen was present in the final 
 gas, which was a mixture of 22.4 cc. methane, 24.8 
 cc. ethane, and 15.6 cc. ethylene or other olefins. 
 
 Norton and Noyes 2 passed ethylene slowly through 
 a glass tube heated to a low red heat for 60 cm. of its 
 length. The products passed out through a series 
 of U-tubes immersed in a freezing mixture, then through 
 ammoniacal cuprous chloride, through bromine, and 
 finally to a gasometer. Carbon was deposited in 
 the tube. From the liquid condensing in the freezing 
 mixture benzene and naphthalene were isolated. Only 
 traces of acetylene were in the gases, and the authors 
 concluded that if acetylene were formed it must have 
 been decomposed in the tube. The liquid bromides 
 collected consisted chiefly of ethylene dibromide, 
 but also contained some methylene dibromide, propyl- 
 ene dibromide, and butylene dibromide. The solid 
 bromides had the composition C 4 H 6 Br4, and Norton 
 and Noyes believed that it could be explained by 
 the reaction 2C 2 H 4 > C 4 H 6 + H 2 . The gases col- 
 lected consisted of methane and ethane. 
 
 Sabatier and Sendefens 3 passed pure dry ethylene 
 over freshly reduced nickel at 300 C. and upwards. 
 The products were carbon, hydrogen, methane and 
 ethane in varying proportions. These investigators 
 found that the hydrogen increased with increase in 
 temperature, and believed that it resulted from a 
 secondary reaction, i. e., decomposition of methane 
 
 i Am. Chem. J., 8 (1886), 153-167. 
 
 2 /Wd., 8 (1886), 362-364. 
 
 Compt. rend., 124 (1897), 616-618. 
 
 (24) 
 
into carbon and hydrogen. In a later paper 1 similar 
 decompositions were carried out with cobalt as the 
 catalyst. The analysis of a typical gas which resulted 
 was: ethylene, 67.4; ethane, 13.4; methane, 4.4; 
 and hydrogen 14.8 per cent. As would be expected, 
 if the methane equilibrium is kept in mind, the decom- 
 position of the methane in the presence of the catalyst 
 was quite extensive, and a fairly large proportion of 
 the hydrogen so formed reacted with the ethylene 
 to form ethane. 
 
 Bone and Wheeler, 2 in their study of the combus- 
 tion of ethylene, found that when insufficient oxygen 
 was present . to burn the ethylene to formaldehyde 
 a thermal decomposition took place. The products 
 were carbon, hydrogen, methane, and traces of acet- 
 ylene. Ipatiew's study of the condensation reactions 
 of the ethylene hydrocarbons has been discussed un- 
 der the secondary decomposition, but should be called 
 to mind again here. Engler also has laid emphasis 
 on the importance of such condensation reactions. 
 
 Pring and Fairlie 3 found that small amounts of 
 ethylene were formed from carbon and hydrogen; 
 and that the ethylene so formed reacted further to 
 give methane. 
 
 H. Rollings and J. W. Cobb 4 passed a gas mixture 
 containing 10.6 per cent ethylene, 47.7 methane, 
 and 41.7 hydrogen through a tube heated to 800 C. 
 The exit gases, with a heating time of 45 seconds, 
 contained 4.5 per cent ethylene, 54.4 methane, 40.3 
 hydrogen and 0.8 acetylene. No benzene or ethane 
 were found. Methane and acetylene appeared to 
 be the chief products. Apparently the reactions 
 taking place were: C 2 H 4 > CH 4 + C and C 2 H 4 > 
 C 2 H 2 + H 2 . The first of these was much the most 
 important reaction. A contraction in volume took 
 place which the authors believed was due to the forma- 
 tion of liquid products. At 1100 C. the ethylene 
 was decomposed completely in 35 seconds' heating. 
 
 Lewes 5 found that ethylene yielded considerable 
 proportions of acetylene when it was decomposed 
 by heat. From a 'study of the gases made from Rus- 
 sian oil he concluded that nothing happened till a 
 
 1 Compt. rend., 131, iv (1900), 267-270. 
 
 2 J. Chem. Soc., 85 (1904), 1637-1663. 
 
 Repts. 8th Int. Congress Applied Chem., 11, 65, et seq. 
 
 *Gas World, 60, 879-884. 
 
 * J. Chem. Soc., 61 (1892). 322-338; J. Soc. Chem. Ind., 11 (1892), 
 584-590; Proc. Roy. Soc., 55 (1894), 90; Trans. Inc. Inst. of Gas Eng., 10 
 (1900), 111-133. 
 
 (25) 
 
temperature of 800 C. was reached, but that then 
 the" ethylene began to decompose in accordance with 
 the equation 3C 2 H 4 > 2C 2 H 2 + 2CH 4 . 
 
 The experiments of Bone and Coward, however, 
 completely disproved any such mode of decomposi- 
 tion for ethylene. From their work 1 it appeared 
 that ethylene decomposed in different ways at differ- 
 ent temperatures. In the study of the thermal 
 decomposition of ethane at 675 C., it was shown 
 that the ethane dissociated first into ethylene and 
 hydrogen, and that the further change consisted largely 
 in the breaking down of the ethylene into carbon 
 and hydrogen. When the reactions undergone by 
 pure ethylene at 575 C. were studied it was found 
 that very complex gas mixtures were obtained. Acet- 
 ylene, ethane, methane, hydrogen and aromatic hydro- 
 carbons were all present in notable quantities while 
 the separation of carbon was small. The experiments 
 continued' over a time of 150 minutes and there was a 
 continuous decrease in the volume of the gases which 
 indicated a formation of aromatic hydrocarbons. 
 
 Bone and Coward regarded acetylene as one of the 
 primary decomposition products of ethylene at the 
 temperature of 575 C. The large percentage of 
 hydrogen without a corresponding separation of car- 
 bon precluded Lewes' contention that the acetylene 
 arose largely from the reaction 3C 2 H 4 > 2C 2 H 2 + 
 2CH 4 , for, were this the main reaction, hydrogen would 
 result from the secondary decomposition of the acet- 
 ylene which is of necessity accompanied by separation 
 of carbon. Experiments were conducted at 700, 800 
 and 950 C. The higher the temperature the greater 
 the separation of carbon, and the smaller the forma- 
 tion of aromatic hydrocarbons. Acetylene and ethane 
 were produced in smaller amount, but methane in 
 larger amount. 
 
 Bone and Coward have some difficulties in recon- 
 ciling their results on the study of ethylene alone, 
 with those of the ethylene resulting from the dissocia- 
 tion of ethane. The formation of acetylene is much 
 greater and the decomposition into carbon and methane 
 less with pure ethylene than with the ethylene result- 
 ing from the decomposition of ethane. It seems that 
 the presence of the ethane and the hydrogen in the 
 latter case are all-important. 
 
 SUMMARY It appears that at temperatures up to 
 700 C. the dissociation of ethylene into acetylene 
 
 i J. Chem. Soc.. 93 (1908), 1197-1225. 
 (26) 
 
and hydrogen is the most important reaction. Con- 
 densation also takes place to some extent. At higher 
 temperatures the rate of the decomposition into car- 
 bon and methane is much greater, so this reaction 
 plays an important part. The high proportion of 
 methane found by Bone and Coward at temperatures 
 of about 800 C., and explained by them as due to 
 hydrogenation of unsaturated residues can be under- 
 stood if it is assumed that reactions such as 
 C 2 H 4 + 2H, ^1 2 CH 4 take place. 
 
 REACTIONS OF ACETYLENE 
 
 It has been seen that acetylene is formed by the 
 dissociation of ethylene. Many authors have noted 
 its formation in the gases made from oil. 
 
 Lewes 1 has shown that acetylene was present in the 
 gases made from Russian petroleum at 700 to 1000 
 C., and that the amount increased with the tempera- 
 ture. The percentages were 0.084 at 700, 0.38 at 
 800 and 0.46 at 900 C. 
 
 Haber and Oechelhauser 2 claimed to have found 
 1. 10 per cent acetylene in gases made from hexane 
 at temperatures between 900 and 1000 C. 
 
 Worstali and Burwell 3 reported 1 1 . 8 per cent in 
 the gases from the decomposition of heptane and oc- 
 tane under the conditions of Pintsch oil gas manufac- 
 ture. The presence of thte amount of acetylene, 
 however, in the light of the work of all other investiga- 
 tors, is open to confirmation; e. g., ammoniacal cuprous 
 chloride was used to absorb the acetylene while it is 
 known that this reagent will absorb ethylene as well. 
 
 W. A. Noyes, W. M. Blinks and A. V. H. Morey* 
 found acetylene in the oil gases made at 900 to 1000 
 C. as indicated by the precipitate obtained when the 
 gases were passed through ammoniacal cuprous 
 chloride. The observation of these investigators, and 
 many more which might be mentioned, rather over- 
 balance the claim of Norton and Andrews 5 that they 
 could not obtain acetylene in the gases made by 
 passing the vapors of pentane and hexane through a 
 porcelain tube heated to bright red heat. It is ap- 
 parent, therefore,- that the reactions of acetylene 
 are of some moment to the gas manufacturer. 
 
 1 J. Soc. Chem. Ind.. 11 (1892), 584-590. 
 
 2 J. Gasbel, 39 (1896), 377-382. 395-399, 435-439. 452-455. 799-805. 
 813-818, 830-834. 
 
 3 Am. Chem. J., 19 (1897), 815-845. 
 
 4 J. Am. Chem. Soc., 16 (1894), 688-697, 
 4 Am. Chem. J., 8 (1886), 1-9. 
 
R. Meyer 1 made a study of the thermal reactions 
 of acetylene. If the gas alone was passed through an 
 electrically heated tube at 650 to 800 C. the acet- 
 ylene was decomposed into carbon and hydrogen. If 
 hydrogen was mixed with the acetylene this decom- 
 position was avoided. The gases leaving the ap- 
 paratus contained some methane. The tar yields 
 were high, in one case 60 per cent of the weight of 
 the acetylene used, showing the decided tendency 
 of acetylene to condense to higher molecular weight 
 compounds. Twenty per cent of the tar, by weight, 
 was composed of benzene. Naphthalene, anthracene, 
 indene, toluene, diphenyl, fluorene, pyrene and chry- 
 sene were also detected. Meyer's conclusion was 
 that the condensation of acetylene was an essential, 
 though not the sole, factor in the formation of the 
 aromatic components of coal tar. 
 
 Not all investigators have agreed that benzene is 
 formed from acetylene. 
 
 Norton and Noyes, 2 in their experiments on eth- 
 ylene, found only small amounts of acetylene in the 
 g'ases, and yet the liquid products contained fair 
 amounts of benzene. These authors said that if ben- 
 zene was formed from acetylene the conditions of 
 their experiments, i. e., low red heat, must have been 
 just those favorable for this reaction of condensation. 
 
 Lewes 3 was of the opinion that acetylene when heated, 
 as in the coal gas retorts, condensed at once to ben- 
 zene and other aromatic hydrocarbons. At higher 
 temperatures it was decomposed into its elements, 
 and this according to Lewes, was the only reaction 
 resulting in the separation of carbon. 
 
 .Berthelot 4 had previously announced similar views. 
 He passed acetylene through glass tubes at low tem- 
 peratures and obtained solid and liquid hydrocarbons 
 among which were benzene, naphthalene, and styrene. 
 
 Jacobson 5 believed that acetylene condensed to 
 benzene, and that the substituted acetylenes gave 
 rise to toluene, xylenes and mesitylene. 
 
 Anschiitz 6 pointed out that if it was considered 
 that a hydrogen molecule split out the formation of 
 naphthalene could be accounted for: 5C 2 H 2 > 
 C 10 H 8 + H 2 . 
 
 1 Ber., 45 (1912), 1609-1633. 
 
 2 Am. Chem. J ., 8 (1886). 362-364. 
 
 Trans. Inc. Inst. of Gas Engineers, 10 (1900). 111-113. 
 4 Ann. chim. phys., 9, iv, 445, 469. 
 * Ber., 10 (1877). 855. 
 Ibid., 11 (1878), 1215. 
 
 (28) 
 
Haber found that benzene and some naphthalene 
 were formed by passing acetylene through a tube 
 at 700 C. 
 
 Sabatier and Senderens 1 studied the action of the 
 metals copper, nickel, cobalt, platinum, and iron on 
 acetylene. At temperatures of 150 to 250 C. a 
 large part of the liquid products formed were aromatic 
 hydrocarbons. Decomposition into carbon and hydro- 
 gen also took place, and nearly all the hydrogen so 
 liberated combined with the acetylene to form ethylene 
 and ethane. 
 
 Bone and Coward 2 showed that at 500 C. the 
 principal change undergone by acetylene was condensa- 
 tion, with decomposition into its elements second 
 in importance, while hydrogenation to ethylene, 
 ethane, and methane was least important. At 600 C. 
 the relationship was found to be much the same. At 
 800 C. the gas "flashed," and the temperature locally 
 was much higher than 800 C. The most important 
 feature of the results at this temperature was the 
 higher proportion of methane in the resultant gases. 
 This could be accounted for, according to Bone and 
 Coward, only by assuming the hydrogenation of =CH 
 residues formed momentarily by the breaking across 
 a triple bond. They strengthened this argument 
 by heating acetylene in nitrogen and hydrogen, in 
 the ratio of C 2 H 2 + 3N 2 and C 2 H 2 + 3H 2 , when it 
 was found that five times as much methane was formed 
 when the gas was cracked in hydrogen as when the 
 change took place in nitrogen. Condensation oc- 
 curred at 800 C., but far less extensively than at 
 lower temperatures. At 1100 C. condensation 
 played a still less important part, decomposition into 
 the elements being the chief change. The most favor- 
 able temperatures for the condensation of acetylene 
 lay between 600 and 700 C. 
 
 SUMMARY It would appear, then, that at tem- 
 peratures up to 700 C. acetylene undergoes a fairly 
 rapid condensation to benzene and its homologs. 
 Decomposition into carbon and hydrogen is second 
 in importance at these temperatures while hydrogena- 
 tion to ethylene, ethane and methane is least important. 
 At slightly higher temperatures the importance of 
 condensation diminishes, while the decomposition 
 into the elements and hydrogenation both increase. 
 At still higher temperatures, such as 1100 C., the 
 
 1 Compt. rend., 131, iv (1900), 267-270. 
 
 2 7. Chem. Soc., 93 (1908), 1197-1225. 
 
 (29) 
 
decomposition into carbon and hydrogen is the im- 
 portant reaction. 
 
 REACTIONS OF PROPYLENE, DIOLEFINS, SUBSTITUTED 
 ACETYLENES, PROPANE, AND HIGH MOLECU- 
 LAR WEIGHT TARRY COMPOUNDS 
 
 No discussion of the pyrogenic reactions of pyro- 
 pylene, diolefins, substituted acetylenes, propane, and 
 high molecular weight tarry compounds was found. 
 A few conjectures as to their probable reactions may 
 at least be of interest. 
 
 PROPYLENE can condense to fo'rm substituted naph- 
 thenes. It has been established that such a reaction 
 does take place in the case of ethylene, and it is en- 
 tirely probable that propylene undergoes condensa- 
 tion as readily or more readily than ethylene. A 
 rupture of the propylene molecule with formation 
 of acetylene and methane is also not an improbable 
 reaction,: 
 
 CH 3 CH=CH 2 > CH 4 + CH=CH 
 
 The DIOLEFIN which has most frequently been 
 isolated and identified from the products of the ther- 
 mal decomposition of hydrocarbons is divinyl or 
 vinyl ethylene, CH 2 =CH CH=CH 2 . Whether it 
 is formed in small amount only, or whether it is 
 formed in large amount as an intermediate product 
 between the higher olefins and more complex condensa- 
 tion products cannot be definitely stated. Isoprene, 
 a diolefin of the supposed formula CH 2 =C(CH 3 ) 
 CH=CH 2 , is known to condense spontaneously 
 to dipentene, and ultimately into caoutchouc. If 
 judgment from analogy is allowable it would seem 
 that divinyl would also condense readily. This rapid 
 condensation, rather than a small formation, may 
 be responsible for the fact that divinyl is found in 
 small amount only in the products of hydrocarbon 
 decomposition. 
 
 It may be regarded as well established that acetylene 
 will condense to form benzene if passed through a tube 
 heated to 600 to 750 C. So it would be probable 
 that the substituted acetylenes would undergo similar 
 reactions with formation of substituted benzenes. 
 That such reactions actually take place is rendered 
 the more plausible by the observation of Rittman and 
 his collaborators, referred to under the discussion of 
 aromatic hydrocarbons in this paper, that the sub- 
 stituted benzenes, such as cymene, xylene, and toluene, 
 
 (30) 
 
are the first to be formed when paraffin hydrocarbons 
 are decomposed pyrogenically. 
 
 PROPANE may easily decompose into methane and 
 ethylene, thus, CH 3 CH 2 CH 3 > CH 4 + CH 2 = CH 2 . 
 This reaction is analogous to the primary reaction of 
 the higher paraffin hydrocarbons. 
 
 The HIGH MOLECULAR WEIGHT TARRY COMPOUNDS 
 
 have been built up by synthetic reactions. For this 
 reason they are no doubt fairly stable compounds as 
 regards the influence of heat, and in general pass on 
 through the apparatus and into the tar. Haber has 
 suggested that the chief reaction of the compounds 
 of this class is a splitting off of hydrogen. 
 
 IV AROMATIC HYDROCARBONS THEIR FORMATION AND 
 
 REACTIONS 
 
 Though the formation of aromatic hydrocarbons 
 takes place only to a limited extent in water gas ma- 
 chines, and to but a slightly greater extent in coal gas 
 retorts, the tars from these processes are important 
 sources of the commercial aromatic hydrocarbons. 
 
 The aromatic hydrocarbons contribute greatly to 
 the illuminating value of a gas as determined in an 
 open flame burner. However, should the calorific 
 standard for gas service in the course of time become 
 general, the presence of the aromatic compounds in 
 the gas would be of far less importance to the gas manu- 
 facturer than it is to-day. These compounds con- 
 tribute less to the heating value of the gas than the 
 compounds from which they are formed. Their chief 
 importance would then lie in the better prices obtain- 
 able for the tars containing them. 
 
 Notwithstanding these facts it is deemed well worth 
 while to discuss to some extent the literature relevant 
 to the formation of the aromatic hydrocarbons, and 
 the changes undergone by them. 
 
 Lewes, 1 in commenting on Armstrong's statement 2 
 that benzene was not formed from acetylene, said 
 that he believed this reaction to be responsible for a 
 part, only, of the benzene formed and that the poly- 
 methylene hydrocarbons, by their dissociation into 
 benzene and hydrogen, contributed largely. The 
 high percentage of benzene (20 per cent) in Russian 
 oil tars was cited as proof. 
 
 It was suggested by Armstrong and Miller 3 that 
 hexane might have a specific tendency to condense 
 
 1 J. Soc. Chem. Ind., 11 (1892), 584-590. 
 
 2 Ibid., 3 (1884), 462-468. 
 
 1 J. Chem. Soc., 49 (1886), 74-93. 
 
 (31) 
 
to hydro-aromatic hydrocarbons which in turn would 
 give rise to benzene. This has been disproved by 
 Haber; 1 approximately equal amounts of benzene 
 were found in the reaction products of hexane and 
 trimethylechylene. The latter compound could not 
 form benzene by a linking together of the ends of its 
 carbon chain. 
 
 Considerable recent work has been done with the 
 commercial production of aromatic hydrocarbons 
 as its object. Though these investigations have not 
 been made from the manufacturer's gas viewpoint, 
 valuable conclusions as to the mechanism of the reac- 
 tions of the aromatic hydrocarbons can be drawn 
 from them. 
 
 Brooks and Humphreys 2 discussed the presence 
 of aromatics in high boiling petroleum distillates. 
 Though they recognized that in oil gas making ben- 
 zene no doubt results from the polymerization of 
 acetylene they did not believe that such condensa- 
 tions were responsible for the benzene in their distillates 
 since the temperature of distillation did not exceed 
 420 C. at which temperatures they assumed acet- 
 ylene formation negligible. However improbable such 
 formation of acetylene may be, it has not been proved 
 that it does not take place. These authors also re- 
 ferred to the work of Ogloblin, 3 in which it was shown 
 that at temperatures of 525 to 550 C., Russian 
 naphtha (a hydrocarbon largely composed of naph- 
 thenes) broke down readily to give good yields of ben- 
 zene, toluene, xylene, and smaller amounts of naph- 
 thalene and anthracene. Brooks and Humphreys 
 referred to this work in the sense that the tempera- 
 tures of 525 to 550 C. were the lower limiting tem- 
 peratures at which these reactions would take place. 
 Ogloblin, however, intended to convey no such im- 
 pression, and stated that at these temperatures the 
 reactions took place readily. This fact convinces 
 us that Brooks and Humphreys were in no sense 
 justified in their conclusion that the benzene in their 
 high boiling petroleum distillates arose from hydro- 
 carbons containing a phenyl radical. 
 
 The work of Engler and Lehmann 4 can be mentioned 
 in this connection. These investigators found naph- 
 thenes and aromatic hydrocarbons in oily distillates 
 
 1 J. Gasbel, 39 (1886), 377-382, 395-399, 435-439, 452-455, 799-805, 
 813-818, 8-30-834. 
 
 2 J. Am. Chem. Soc., 38 (1916), 393-340. 
 
 3 Z. angew. Chem., 18 (1905), 540. 
 
 4 Ber., 30 (1897), 2365-2368. 
 
 (32) 
 
made by heating fats under pressure at temperatures 
 below those used by Brooks and Humphreys. 
 
 It must be kept in mind also that Rittman and his 
 co-workers have obtained maximum yields of ben- 
 zene at temperatures of 600 C. and slightly above 
 and 4 atmospheres pressure. They are of the opinion 
 that condensation of acetylenes, dehydrogenation of 
 naphthenes, and splitting of polycyclic or asphaltic 
 hydrocarbons all contribute. Egloff and Twomey 1 
 obtained considerable yields of benzene at 450 C. 
 and atmospheric pressure. 
 
 An excellent investigation of the relationships 
 between the most important aromatic hydrocarbons 
 has been made by Rittman, Byron, and Egloff 2 and 
 by Rittman and Egloff. 3 They used the pure hydro- 
 carbons cymene, xylene, toluene, benzene, naphthalene, 
 and anthracene, which were treated at various tem- 
 peratures and pressure in the furnace formerly used 
 by Whitaker and Rittman. 4 The authors came to 
 the conclusion that from cymene it was possible to 
 produce all the other hydrocarbons mentioned. 
 Xylene gave toluene, benzene, naphthalene, and 
 anthracene, but no cymene. Toluene gave benzene, 
 anthracene, and naphthalene, but no cymene or xylene. 
 Only naphthalene and anthracene may be produced 
 from benzene. Naphthalene yielded anthracene but 
 none of the others. Anthracene yielded only carbon 
 and gas. The authors noted that other compounds 
 such as diphenyl, methylnaphthalenes, and anthra- 
 cenes, and phenanthrene were formed to some extent. 
 It can be seen that the general course of the reac- 
 tions with monocyclic aromatic compounds was from 
 higher to lower molecular weight compounds, and 
 that the tendency was for monocyclic to go to poly- 
 cyclic compounds. 
 
 The influence of various conditions on the formation 
 of aromatic hydrocarbons has recently been reviewed 
 by Egloff and Twomey. 5 This paper also records 
 the results of an investigation of the effect of tempera- 
 ture on the individual aromatic hydrocarbon forma- 
 tion at atmospheric pressure. Temperatures of 459 
 to 600 C. produced more toluene and xylene than 
 benzene in the cracked oil, and more toluene than 
 xylene. More toluene than benzene was formed at 
 
 1 J. Phys. Chem., 20 (1916), 121, et seq. 
 
 2 J. Ind. Eng. Chem., 7 (1915), 1019, et seq. 
 
 3 Met. Chem. Eng., 14 (1916), 15-18. 
 
 4 J. Ind. Eng. Chem., 6 (1914), 383, 472. 
 * J. Phys. Chem., 20 (1916), 121-150. 
 
 (33) 
 
650 C., and more benzene than xylene. No naph- 
 thalene or anthracene was found in the recovered 
 oils at 650 C. In the temperature range of 700 to 
 875 C. .more benzene was formed than toluene or 
 xylene, but more toluene than xylene. At 750 C. 
 naphthalene began to be formed, and at 800 C. 
 anthracene formation commenced. The amounts of 
 both of these hydrocarbons increased with tempera- 
 ture. The authors seemed justified in their conclu- 
 sion that the individual aromatic formation with change 
 of temperature was as follows: 
 
 HIGHER HOMOLOGs] 
 
 , TOLUENE 
 
 such as mesity- I 
 
 , ^ f > AND 
 
 lene pseudocu- 
 
 XYLENE 
 
 mene, etc. 
 
 BENZENE 
 
 This is not to be interpreted, for instance, to mean 
 that toluene and xylene cannot form anthracene or 
 naphthalene directly. 
 
 THE REACTIONS or BENZENE have been the subject 
 of a number of investigations, and it is deemed worth 
 while to present them here in review. Schultz 1 
 found that diphenyl was formed by passing benzene 
 vapors through tubes heated to redness. Meyer, 2 
 in his paper on the condensations of acetylene, noted 
 the presence of diphenyl in the high boiling fractions 
 of the liquid obtained by passing acetylene through 
 heated tubes. Ipatiew 3 found that diphenyl was not 
 formed below 600 C. Smith and Lewcock 4 passed 
 benzene vapors through a heated iron tube at the rate 
 of 5 cc. of liquid benzene per minute. The yield of 
 diphenyl was not increased by the use of tempera- 
 tures above 720 C. The time factor was found to 
 be important. By the variance of conditions yields 
 of diphenyl ranging from o.o to 59 per cent could be 
 obtained. H. Rollings and J. W. Cobb 5 found that 
 the decomposition of benzene at 800 C. was negligi- 
 ble when a mixture of practically equal amounts of 
 methane and hydrogen saturated with benzene vapor 
 was passed through a heated tube. At 1100 C., no 
 trace of the benzene could be found in the products. 
 According 6 to Haber, benzene was not decomposed by 
 heat at temperatures below 900 C. Above 1000 C., 
 
 1 Ber., 9 (1876), 547. 
 
 2 Ibid., 45 (1912), 1609-1633. 
 
 3 J. Russ. Phys. Chem. Soc., 39 (1907), 681. 
 J. Chem. Soc., 101 (1912), 1453-1458. 
 
 s Gas World, 60 (1914), 879-884. 
 
 J. Gasbel, 39 (1896), 377-352, 395-399, 435-439, 452-455, 799-805. 
 813-818, 830-834. ' 
 
 (34) 
 
however, it was affected; diphenyl and crystalline sub- 
 stances not further examined were found; naphthalene 
 was not observed. 
 
 The great variation in the results of these authors 
 shows very well the great influence of conditions on 
 the reactions of benzene. Temperature, pressure, 
 and time of heating are extremely important. 
 
 SUMMARY The reactions responsible for the aro- 
 matic hydrocarbons formed in the pyrogenic decom- 
 position of hydrocarbon oils are 
 
 i Condensation of acetylenes 
 2 Dehydrogenation of naphthenes 
 3 Decomposition of complex compounds already 
 containing the phenyl radical 
 
 Toluene, xylene, ethyl benzene, and similar com- 
 pounds are formed first and most easily. These by 
 further change give rise to benzene, naphthalene, 
 and anthracene. With high molecular weight mono- 
 cyclic compounds the general course of the reaction 
 is toward compounds of lower molecular weight. In 
 general, monocyclic compounds tend to go to poly- 
 cyclic compounds. 
 
 V INFLUENCE OF ATMOSPHERES OF HYDROGEN AND 
 
 INERT GASES ON THE PYROGENIC DECOMPOSI- 
 TION OF HYDROCARBONS 
 
 The idea that an oil cracked in hydrogen or in 
 inert gases gave more valuable products has been 
 prevalent for many years. 
 
 Lewes 1 states that the better results are obtained by 
 cracking oil in gases such as hydrogen, carbon mon- 
 oxide, or water gas. According to Lewes these gases 
 exerted their helpful influence by ' 'separating and 
 partly protecting the molecules during the decomposi- 
 tion by heat." Thus the temperature ranges over 
 which good results could be obtained was increased, 
 and excessive loss and breaking down of the hydrocar- 
 bons prevented. The use of blue water gas in coal 
 gas retorts was also advocated. 
 
 Lewes 2 found a considerable increase in the illumi- 
 nants formed per given weight of Russian oil when the 
 oil was decomposed in atmospheres of inert gases. 
 He also commented on the analogous results obtained 
 by the use of the Lowe system, and was of the opinion 
 that the blue water gas prevented the decomposition 
 going too far, and also aided the breaking up of the 
 
 1 Trans. Inc. Inst. of Gas Engineers, 2 (1892), 77-85. 
 . 2 J. Soc. Chem. Ind., 11 (1892), 584-590. 
 
 (35) 
 
benzene. Lewes no doubt had in mind the fact that 
 dilution, by diminishing the partial pressure of the 
 acetylene, would tend to favor the acetylene side of 
 the equilibrium 3HC=CH ^~ C 6 H 6 . 
 
 Numerous patents on the various methods of using 
 hydrogen or inert gases as an aid to the decomposi- 
 tion of oils have been issued. The patents of H. 
 Blau 1 are illustrative: these involve a manufacture of 
 oil gas by "expansion and rarefaction with permanent 
 gases," whereby the separation of soot, and the conse- 
 quent irregularities of working were avoided. 
 
 W. Benthrup 2 patented a process for the introduc- 
 tion of water gas into coal gas retorts during the early 
 stages of the gasification. The gas yield was supposedly 
 increased, and a better quality of gas obtained. 
 
 H. Croissant 3 discussed at some length the introduc- 
 tion of water gas into coal gas retorts. The advan- 
 tages according to him are the increase in the gas 
 yield from the coal, and a gain in heating value of 7 
 to 12 per cent. The tar was more fluid, and stoppages 
 in the ascension pipes were less frequent. Both of 
 these latter points show a less destructive cracking 
 of the hydrocarbons distilled from the coal. The 
 water gas should be introduced in the proportion of 
 */2 to i volume for every 2 volumes of coal gas during 
 the first 2 hours of the period of carbonization. 
 
 The basis of the Del Monte process of carboniza- 
 tion 4 developed in England and was the recirculation 
 of the permanent gases (such as hydrogen and methane) 
 through the coal gas retorts. Excessive cracking 
 of the high hydrocarbons distilled out at moderate 
 temperatures was thus prevented. 
 
 Meyer's work, in which it was found that admix- 
 ture of hydrogen with acetylene prevented the de- 
 composition of the latter into carbon and hydrogen, 
 has already been mentioned. 
 
 Several other authors have mentioned the effects 
 of decomposing oils in water gas. Thus C. E. Munroe 
 stated that the hydrogen in the water gas had the 
 effect of increasing the olefins and diminishing the 
 amount of tar. A. E. Forstall 6 was of the belief that 
 the presence of water gas arrested the decomposition 
 
 1 French Pat. 332,115, 1903, and Ger. Pat. 258,610, 1912. 
 
 2 Ger. Pat. 157,470, 1903. 
 
 J. Gasbel, 47 (1904), 219-222. 
 
 * Gas World, 58 (1913), 321, 328, 430 and 477; J. Gas Lighting., 120 
 (1913), 370. 
 
 s J. Frank. Inst., 174 (1912), 1-33. 
 Ibid.. 174 (1912), 279-302. 
 
 (36) 
 
of the hydrocarbons before carbon was deposited, 
 and thus allowed the use of a larger surface of contact 
 for heating. Thomas Holgate 1 pointed out that the 
 curve for the rate of decomposition of methane with 
 increase of temperature had a maximum, and ascribed 
 this to the retarding effect of the increasing amounts 
 of hydrogen formed by the decomposition. 
 
 Redding 2 cited several average analyses of water 
 gas. These are shown in abbreviated form in the 
 following table, but calculated to an illuminants- 
 carbon- monoxide- methane-hydrogen basis. 
 
 CARBON 
 
 No. ILLUMINANTS METHANE MONOXIDE HYDROGEN 
 
 1 10.4 19.9 33.6 36.1 
 
 2 11.9 21.3 33.2 33.6 
 
 3 14.5 14.0 35.1 36.4 
 
 4 14.5 17.8 32.2 35.5 
 
 Keeping in mind that the water gas introduced has a 
 little more hydrogen than carbon monoxide it can be 
 seen that the proportion of hydrogen to carbon mon- 
 oxide has not been much increased. The average 
 analysis of Pintsch gas for hydrogen is 14 per cent, 
 but evidently the oil when cracked in the water gas 
 has not given rise to this amount of hydrogen. This 
 appears as good evidence, since the other reactions 
 which might tend to distort the carbon monoxide- 
 hydrogen ratio, namely, 2C + 2 V** 2CO; CO -f 
 HOH ^ CO 2 + H 2 ; and CO 2 + C ^ 2 CO, tend 
 to a balance, and in any case do not have a large 
 effect. 
 
 A study of the gasification of oils in hydrogen and 
 other gases was made by Hempel. In his experiments 
 the oil was gasified and led into a i6-in. tube heated 
 by a row of Bunsen burners. The rate of gasifica- 
 tion was 1.6 to 2.0 grams of oil per minute. The 
 gases contained from 30 to 36 per cent unsaturated 
 hydrocarbons, 47.5 to 51 per cent saturated hydro- 
 carbons, and 14 to 17 per cent hydrogen. The prod- 
 ucts of decomposition of the oil and the hydrogen 
 reacted as was evidenced by the 15 per cent diminu- 
 tion in total volume when the hydrogen was introduced 
 in the proportion i hydrogen : 2 oil gas. The resultant 
 gases were practically identical with those produced by 
 cracking oil alone, but the yields of methane, ethylene, 
 and ethane, according to Hempel, were increased per 
 unit weight of oil. The ethane particularly was in- 
 creased. The gases showed a 15 per cent gain in heat 
 of combustion over the simple oil gases. The tars 
 
 1 J. Gas Lighting, 106 (1909), 25-28, 84-86. 
 
 2 Progressive Age, 24 (1906), 485. 
 
 (37) 
 
formed were more fluid, and in smaller amount by 
 about 10 per cent when the oils were cracked in hydro- 
 gen. The separation of carbon was also less by about 
 1.5 to 3.0 per cent. The experiments indicated 
 that when the hydrogen reached a proportion of 50 
 to 60 per cent of the final gas the formation of gaseous 
 hydrocarbons was confined to a limited quantity. 
 When hydrogen was used in the proportion of i hydro- 
 gen : i oil gas, there was a falling off in the produc- 
 tion of ethane and much of the hydrogen was lost 
 in the tars, which increased in amount. The most 
 favorable amount of hydrogen is about 18 cu. ft. 
 per pound of oil. The opinion was expressed that 
 the 15 per cent hydrogen normally present in oil gas 
 exerts a considerable influence. From a very limited 
 investigation Hempel concluded that the cracking 
 of oils in atmospheres of nitrogen or carbon monoxide 
 caused no change in the volume of gas produced from 
 the oil, nor in the candle power. The secondary de- 
 composition was diminished and as a result the per- 
 centage of methane falls a little and the olefin per- 
 centage increases. 
 
 E. C. Jones 1 and Z. B. Jones 2 have carried out 
 experiments on an all-oil water gas set to determine 
 the effect of the cracking of oils in inert gases, and 
 also the extent to which carbon monoxide and hydro- 
 gen react under working conditions to form methane. 
 They established that the inert gases had the effect 
 of increasing the proportion of carbon in the gas per 
 unit weight of the oil, and in their technical opera- 
 tions cracked the oil in an atmosphere of gas from 
 their holder, with beneficial effect. Their experi- 
 ments showed, too, that carbon monoxide and hydro- 
 gen united to some extent to form methane and water. 
 
 Whitaker and Rittman 3 noted that the quantity 
 and quality of the gas formed from the oil increased 
 when the oil was cracked in atmospheres of hydrogen, 
 and that the tar and deposited carbon decreased. 
 They found that more hydrogen entered into combina- 
 tion with the decomposition products of the oil at 
 atmospheric pressure than at a pressure of 0.75 
 pound absolute. 
 
 The results obtained by the cracking of oil in atmos- 
 spheres of hydrogen may be ascribed to at least three 
 factors : 
 
 * Am. Gas Lighting, J ., 92 (1910), 437-445. 
 
 2 Progressive Age, 28 (1910). 373. i 
 
 J. Ind. Eng. Chem., 6 (1914), 472-479. 
 
 (38) 
 
i- The dilution effect, which is the same as diminish- 
 ing the partial pressures of all the reacting gaseous 
 hydrocarbons. 
 
 2 The increase in volume, whereby the time of 
 heating of the hydrocarbon molecules is less than in 
 the making of straight oil gas. 
 
 3 Hydrogenation of hydrocarbons. 
 
 The first of these factors indicates that cracking 
 oils in hydrogen or other gases should favor those re- 
 actions which result in greater, numbers of molecules: 
 e. g., the tendency of acetylene to condense to ben- 
 zene would be less in hydrogen than if the acetylene 
 were present alone. The diminution of the time of 
 heating would result in a less extensive decomposi- 
 tion of the various hydrocarbons. The products ob- 
 tained should be those of the earlier stages of hydro- 
 carbon decomposition rather than the large amounts 
 of methane, hydrogen, and aromatics which are 
 obtained by very extensive changes. The extent to 
 which hydrogenation is important can be better judged 
 by considering the experimental work which bears 
 directly on this factor. Some mention of this work 
 has been made under our discussion of the secondary 
 decomposition of hydrocarbons. 
 
 P. Sabatier and J. B. Senderens 1 passed acetylene 
 and hydrogen over various catalysts at low tempera- 
 tures. With cobalt as catalyst there was formed 
 26 per cent ethane, 70 hydrogen and 4 of a mixture 
 of benzene vapor and unchanged acetylene. Nickel 
 as catalyst gave even more ethane than cobalt, while 
 iron was far less active, giving chiefly ethylene and its 
 homologs. 
 
 C. Paal 2 found that acetylene was easily reduced 
 to ethylene and ethane by hydrogen in the presence 
 of colloidal palladium. If equal volumes of acetylene 
 and hydrogen are taken an 80 per cent yield of ethylene 
 will result. 
 
 Ipatiew's results 3 on the hydrogenation of terpene 
 hydrocarbons and aromatic hydrocarbons with un- 
 saturated side chain should also be mentioned. Various 
 catalytic substances were used, and the hydrogen was 
 introduced under 100 to no atmospheres pressure. 
 The hydrogenation of the side chain took place at 
 temperatures of 95 C. in the presence of NiO catalyzer^ 
 
 1 Compt. rend.. 124 (1897), 832. 
 
 2 Chem. Zlg.. 36 (1912), 60, et. seq. 
 
 3 J. Russ. Phys. Chem. Soc.. 43 (1911), 1754-1760; 45 (1913), 944- 
 945; 45 (1913), 1829-1834. 
 
 (39) 
 
while the aromatic nucleus was hydrogenated com- 
 pletely only at temperatures of 185 to 190 C. Though 
 these results are interesting and valuable it must 
 not be concluded from them that similar reactions 
 would take place when these substances and hydrogen 
 are heated together even though the temperatures 
 are considerably higher. 
 
 Meyer 1 in his study of the pyrogenic reactions of 
 acetylene passed mixtures of acetylene and hydrogen 
 through a vertical tutje furnace at temperatures of 
 650 to 800 C.: no appreciable formation of ethylene 
 or ethane took place. Haber found i . 3 per cent 
 ethylene in the gases resulting from passing a mix- 
 ture of acetylene and hydrogen through a tube heated 
 to 630 to 645 C. Lewes found that the hydrogena- 
 tion of acetylene to olefins was an important reac- 
 tion at 1000 C. 
 
 The noteworthy results of Bone and Coward 2 have 
 already been mentioned under the discussion of the 
 decomposition of ethane. To further validate their 
 explanation of the high ratios of methane to hydrogen 
 these authors decomposed mixtures of ethane and 
 nitrogen in the proportion of i ethane : 3 nitrogen, 
 and ethane and hydrogen in the proportion i ethane : 
 3 hydrogen under identical conditions. The ratio of 
 methane to hydrogen in the gases from the methane 
 nitrogen mixture had the approximate value of unity, 
 while from the ethane-hydrogen mixtures it was much 
 higher, showing a considerable hydrogenation of the 
 decomposition products of the ethane; 2.27 times as 
 much methane was formed in the hydrogen atmos- 
 phere as in the nitrogen atmosphere. Similar ex- 
 periments were carried out with ethylene and nitro- 
 gen, and ethylene and hydrogen. The results ob- 
 tained were entirely analogous to those of the acet- 
 ylene experiments. The ratio of the methane forma- 
 tion in the hydrogen atmosphere to that in the nitro- 
 gen atmosphere was 3.15 : i. 
 
 Bone and Coward have shown that both ethylene 
 and acetylene have a certain tendency to combine 
 with hydrogen at low temperatures, but that the im- 
 portance of these reactions is never very great, and 
 that they are insignificant above 1000 C. 
 
 In so far as the work of these several investigators 
 allows us to draw conclusions it seems that hydrogena- 
 tion of hydrocarbons such as ethylene and acetylene 
 
 i Ber., 45 (1912), 1609-1633. 
 
 * J. Chem. Soc., 93 (1908). 1197-1225. 
 
 (40) 
 
does take place, but that it is never a reaction of great 
 importance. The fact, however, that hydrogen is 
 actually absorbed when oils are cracked in atmospheres 
 of the gas shows that hydrogenation of some sort takes 
 place. Possibly the higher unsaturated hydrocar- 
 bons are hydrogenated more readily. The evidence 
 from our experiments in this connection is presented 
 elsewhere in this paper. 
 
 VI TRANSFER OF HEAT IN GAS MACHINES 
 
 The most striking feature of the literature which 
 records the result of investigations of hydrocarbon 
 decompositions is the great difference between the re- 
 sults of various investigators who report that they have 
 worked at the same temperature, and, in general, 
 under the same conditions. Irregularities too great 
 to be attributed to the personal equation are of fre- 
 quent occurrence, and these are in need of explana- 
 tion. 
 
 It is often suggsted that these differences are caused 
 by the catalytic effects of the materials of construc- 
 tion, but without doubt the catalytic effect of contact 
 surfaces has been overestimated. It is well known 
 from general experience that these surfaces are always 
 covered with a layer of hard carbon as a result of the 
 decomposition of hydrocarbons. Hence the gases 
 do not come into contact with an active material, 
 but rather into contact with a dense layer of carbon 
 deposit which is inactive as a catalyzer. 
 
 Unquestionably many of the effects ascribed to 
 catalysis are in reality due to the effectiveness of the 
 heating by conduction and convection close to the sur- 
 faces of the refractory materials. The importance of 
 radiant energy in causing hydrocarbon reactions has 
 also been overemphasized most consistently in the 
 literature of the gas industry. Too little attention 
 has been paid to the methods of making temperature 
 measurements, and also to the interpretation of tem- 
 perature measurements made in certain ways. Thus 
 it is obvious that a metal pyrod with a metal casing 
 or a protective sheath of solid material such as quartz 
 will absorb many times as much radiant energy as 
 the gases in the heated space, and therefore indicate 
 a temperature considerably above the true tempera- 
 ture of the gases. Also the conduction of heat along the 
 metal pyrod casing is more effective than the conduc- 
 tion through a gas. 
 
 The processes of conduction and convection are 
 
 (41) 
 
FIG. 1 Oil, GAS APPARATUS 
 (42) 
 
.'Electrode Holder 
 
 - Cooling Spray 
 
 'Asbestos 
 
 A sbestos Disks 
 
 "Carbon Resistor 
 
 Window 
 
 ---Carbon side Tube 
 ..' Carbon Jacket Tube 
 
 . . Petroleum Coke 
 
 '"- Pyrome 
 
 Tube 
 
 B 
 FIG. 2 RESISTANCE FORNACB 
 
 (43) 
 
dependent on the extent of the heated surfaces ex- 
 posed to the gas, and the manner in which the gases 
 are made to pass over these surfaces. The heat 
 transfer from the hot walls of the retort or checker- 
 brick is primarily a matter of heat conduction through 
 a thin layer of gas next to the hot surfaces, and second- 
 
 To 0/li 
 Tan) 
 
 To Pressure 
 Gauge 
 
 To Oil supply Tank 
 
 FIG. 3 STATIC HEAD AND FEED REGULATOR 
 
 arily a process of convection. Conceived of in terms 
 of the kinetic molecular hypothesis, the molecules 
 of the gas very close to the hot surfaces can be thought 
 of as striking the surfaces, and in so doing having 
 their kinetic molecular energy increased; then they 
 dive out again into the surface layer of the gas where 
 
 (44) 
 
by numerous impacts the newly acquired energy is 
 given up to numbers of other molecules. The mole- 
 cules from the main gas stream are constantly enter- 
 ing the surface layer and those in the surface layer 
 are at all times entering the gas stream. 
 
 In addition to the increase in the kinetic energy 
 of the molecules on account of their contact with the 
 heated bricks, a chemical effect must no doubt be 
 taken into account. Langmuir 1 has explained the 
 abnormal^ heat losses from lamp filaments glowing 
 in hydrogen on the assumption that the molecules 
 
 Asbesfos 
 
 Nichrome 
 Wire Heating 
 Elements 
 
 Asbestos 
 packing around 
 Pipe 
 
 m 
 
 - -Iron Wire Filling 
 FIG. 4 PREVAPORIZER 
 
 of the gas^dissociate into atoms with absorption of 
 heat near the filament, and that the atoms then diffuse 
 away from the wire and recombine with liberation 
 of the heat at first absorbed. 
 
 Magnanini and Malagnini 2 have shown that the 
 dissociation of a gas causes an abnormally high heat 
 conductivity. They found that partially dissociated 
 nitrogen tetroxide had three times as great a heat 
 
 1 Proc. Am. Electrochem. Soc., 20 (1911), 225-242. 
 
 2 Nuovo dm., 6 (1892), 352. 
 
 (45) 
 
conductivity as the completely dissociated gas. This: 
 idea can readily be applied to the transfer of heat 
 in gas machines, for in them there exists a complex 
 gaseous hydrocarbon system in which numerous 
 endothermic reactions are possible in the highly heated 
 layer close to the checkerbrick or retort surfaces. 
 The products of these dissociations then diffuse into 
 the outer layers, and recombine with liberation of 
 the heat previously absorbed, thus effecting the 
 heat transfer. 
 
 Lewes 1 cites experiments by Euchene showing that 
 the temperatures of the gas above the coal in a re- 
 tort varied from 700 C. */2 hour after charging to 
 950 C., 4 hours after charging; 9 in. of space were 
 left above the coal in the retort and the flue tempera- 
 ture was approximately 1100 C. On p. 144 Lewes 
 states that in a horizontal retort, with 1100 C. flue 
 temperature, the gases left the mouthpiece at 400 C. 
 during the first few hours of the period of carboniza- 
 tion and then dropped 100 C. during the last 2 hours. 
 These statements appear as conflicting. It is hard 
 to see why the temperature in the mouthpiece of 
 the retort should not vary somewhat with the tem- 
 perature of the gases in the retort. Further, it is 
 difficult to believe that gases at 700 to 950 C. could 
 be cooled to 400 C. in the mouthpiece. The tempera- 
 ture as observed in the mouthpiece is probably more 
 nearly the true temperature of the gases in that part 
 of the apparatus than the temperatures noted in 
 the gas space above the coal in the retort are for 
 the gases filling that space. 
 
 In the measurement of gas temperatures with a. 
 metal pyrod it is impossible to avoid the absorption 
 of radiations from the solid bodies in the immediate 
 vicinity. This would no doubt play an important 
 part in the case at hand. The temperatures recorded 
 by Euchene are in all probability too high on account 
 of the absorption of radiant energy by the pyrod, 
 whereas this energy is very incompletely absorbed 
 by the gases. This difficulty would not be so marked 
 at 400 C. because the emission of radiations of this 
 type is proportional to the fourth power of the abso- 
 lute temperature. Hence it seems not improbable 
 .that the temperatures for the gas above the coal in 
 the retort as recorded are too high. However, if 
 the gases were efficiently absorbing the radiations. 
 
 * "Carbonization of Coal," pp. 132-133. 
 (46) 
 
emitted by the hot walls of the retort, the tempera- 
 tures of these gases would be at least as high as those 
 recorded by him. Our inference is that this radiant 
 energy is not absorbed to a large extent by the gases, 
 and that conduction and convection play far more 
 important parts in the heat transfer than radiation. 
 
 An excellent paper by Langmuir 1 discusses from 
 a theoretical and practical point of view the heat 
 transfer by means of radiation, conduction and con- 
 vection from surfaces of various shapes into air. It 
 is interesting to note that the air near a hot metal 
 disk was heated to a point hot enough to char paper 
 for a distance of only l / 4 in. from the disk even when 
 the latter was heated to redness. 
 
 Thus in the superheater and carbureter the size 
 of the voids between the checkerbricks must be very 
 important in connection with the results to be ob- 
 tained by the use of a certain temperature. 
 
 Tyndall and Magnus have both worked on the dia- 
 thermacy of gases and vapors. 2 The following tables 
 show some of the results of these investigators. 
 
 THE WORK OF TYNDALL In these experiments 
 the radiation from the heat source passed through a 
 column of gas 33 in. long, and then upon a thermopile 
 used to measure the radiant energy passing through 
 the gas. The source of the radiant energy was a 
 copper plate heated by a small impinging flame from 
 a Bunsen burner. 
 
 EXPERIMENTS AT ATMOSPHERIC PRESSURE 
 
 SUBSTANCE 
 
 % Absorption 
 of Radiations SUBSTANCE 
 08 Hydrogen sulfide 
 
 % Absorption 
 of Radiations 
 31 
 
 Nitrogen 
 
 . .. 0.08 Methane 
 
 32.0 
 
 Hydrogen 
 
 08 Amylene vapor 
 
 60 3 
 
 Carbon monoxide . 
 
 . . . 7.1 Ethylene 
 
 71.0 
 
 Carbon dioxide. . 
 
 7 . 1 Ammonia 
 
 94.5 
 
 Benzol vapor 
 
 ... 21.2 
 
 
 The effect of pressure is shown by the following 
 
 data: 
 
 PER CENT ABSORPTION 
 SUBSTANCE . 1 Atm. . 5 Atm. 1 . Atm. 
 
 Benzene 5.25 14.5 21.2 
 
 Amylene 13.2 39.2 60.3 
 
 Thus the absorption increases with the pressure, 
 but not in direct proportion to it, nor in accordance 
 with any simple law. 
 
 The effect of temperature of the source of the radiant 
 energy can be seen in the following data: The figures 
 are not per cents of absorption but simply numbers 
 
 1 Proc. Am. Electrochem. Soc., 23 (1913). 299-332. 
 
 2 Preston's "Theory of Heat," p. 551-568. 
 
 (47) 
 
representing the relative absorptions as found in this 
 particular set of experiments. 
 
 BARELY BRIGHT WHITE 
 
 SUBSTANCE VISIBLE RED RED HEAT 
 
 Benzene... ... 26.3 20.6 16.5 
 
 Amylene 35.8 27.5 22.7 
 
 Thus the higher the temperature the less the ab- 
 sorption of radiant energy by these gases. 
 
 MAGNUS' EXPERIMENTS The source of the radiant 
 energy in Magnus' experiments was at the tempera- 
 ture of boiling water, and the distance through the 
 gas traversed by the radiations was 15 in. 
 
 SUBSTANCE % Absorption SUBSTANCE % Absorption 
 
 Vacuum 0.00 Methane 27.79 
 
 Oxygen 11.12 Cyanogen 27.79 
 
 Hydrogen 14.21 Ethylene 53.71 
 
 Carbon dioxide 19.77 Ammonia 61.12 
 
 Carbon monoxide 20 . 99 
 
 These results are considerably lower than those ob- 
 tained by Tyndall in the case of the hydrocarbon 
 gases and ammonia, but higher in the cases of oxygen r 
 hydrogen, carbon dioxide, and carbon monoxide. 
 The high results with these last-mentioned gases seem 
 to be due to the influence of convection and conduc- 
 tion in Magnus' experiments. 
 
 Langmuir has shown that out of a total heat loss 
 of 0.670 watt per sq. cm. for a vertical oxidized silver 
 plate 0.134 watt was radiated, and 0.536 watt 
 were conduction and convection losses. The measure- 
 ments were made in still air. 
 
 Although the radiation loss would be greater for a 
 body under black body conditions such as exist ap- 
 proximately in retorts and gas machines the conduc- 
 tion and convection losses would be also greatly in 
 creased by the rapid passage of the gases. Although 
 no data are available on this question it seems a fair 
 assumption that 25 per cent of the heat loss from 
 any particular checkerbrick is radiated energy, while 
 the rest of the heat is transferred by conduction and 
 convection. As can be seen from the data of Tyndall 
 and Magnus cited above, ethylene absorbs more 
 radiant energy than any of the other gases listed 
 except ammonia the absorption lying between 55 
 and 77 per cent of all the available radiations. But 
 this is for ethylene at i atmosphere pressure, while 
 the partial pressure of ethylene in a water gas machine 
 at no time exceeds o. 2 atmosphere. Also the measure- 
 ments of Magnus and Tyndall were made at compara- 
 tively low temperatures whereas the temperature 
 in the gas machine is in the neighborhood of 700 C. 
 
 (48) 
 
The absorption of radiations by gases falls off with 
 diminishing pressure, and also with increasing tem- 
 perature. Taking these factors into consideration 
 it is probable that not more than one-quarter of the 
 available radiation is absorbed by the ethylene. On 
 this basis only one-sixteenth of the total heat trans- 
 fer from the checkerbricks to the ethylene is due to ab- 
 sorption of radiant energy. In the cases of the other 
 gases it is less than this. For hydrogen it is almost 
 negligible, and for carbon monoxide very small. Al- 
 though molecular decomposition would not be expected 
 in the cases of these gases at these temperatures, the 
 fact that they are not effectively atherrm'c is of im- 
 portance; for did they absorb radiations to a large 
 extent they would then transfer the heat so gained by 
 molecular collision to the hydrocarbon molecules 
 which are undergoing important molecular changes, 
 and thus influence the course of the reactions. 
 
 In general then, it would seem that radiation 
 does not play as important a part in transferring 
 heat as has been taken for granted. Conduction and 
 convection are largely responsible for this transfer, 
 and they are helped to some extent by the dissocia- 
 tion reactions of the hydrocarbons. Catalysis has 
 been greatly overrated. The influence on reaction 
 rates supposedly brought about in this fashion can 
 be better understood if the true mechanism of heat 
 transfer is kept in mind. The different results of 
 experimental investigations are also often easily un- 
 derstood if the shape and size of the apparatus is con- 
 sidered in its relation to heat transfer. 
 
 (49) 
 
THE DECOMPOSITION OF HYDROCARBONS AND THE 
 
 INFLUENCE OF HYDROGEN IN CARBURETED 
 
 WATER GAS MANUFACTURE 
 
 H EXPERIMENTAL 
 
 Although numerous studies of hydrocarbon de- 
 composition have been made, no one, nor all combined, 
 comprise a complete investigation. 
 
 The recent work 1 done in connection with the com- 
 mercial production of gasoline and aromatic hydrocar- 
 bons has been the most exhaustive ever attempted. 
 However, these investigations have not been carried 
 out from the standpoint of gas production. 
 
 It is scarcely necessary to discuss the importance 
 of a thorough understanding of the possibilities of 
 controlling the decomposition of hydrocarbons for 
 the obtaining of the particular products desired. 
 
 The most profitable utilization of an oil is of extreme 
 importance to the water gas manufacturer. The con- 
 trol of the cracking of an oil is the thing of first im- 
 portance to the manufacturer of oil gas in any of the 
 various processes. The effect of the presence of 
 hydrogen on the products derived from an oil is of 
 great importance to the water gas manufacturer and 
 the all-oil-water-gas maker. 
 
 Also it is generally recognized that the carboniza- 
 tion of coal and the combustion of coal and oil are 
 allied problems, and that the results of a study of hydro- 
 carbon decomposition are of direct application in these 
 connections. The effect of the introduction of hy- 
 drogen into the carbonizing retorts can also be seen 
 by a study of the data given in this paper. 
 
 Brief mention of the experimental work of a few 
 investigators will serve to show the importance of 
 the subject of hydrocarbon decomposition to the users 
 of coal for gas making purposes or as a fuel. 
 
 Jones and Wheeler 2 have extracted solid paraffins 
 from several British coals by means of pyridine and 
 chloroform. Pictet and Ramseyer 3 have isolated 
 hexahydrofluorene from that portion of a gas coal 
 which was soluble in benzene. The same hydrocarbon 
 
 1 Whitaker and associates at Columbia University, and Rittman and 
 associates, U. S. Bureau of Mines. 
 
 * J. Chem. Soc., 103 (1913), 1704. 
 
 Ber., 44 (1911), 2486; Gas World, 56 (1911), 131. 
 
 (51) 
 
has been identified by them in the tar obtained by the 
 low temperature distillation of coal in a vacuum. 
 Burgess and Wheeler 1 found that paraffin hydrocar- 
 bons were predominant among the primary decomposi- 
 tion products of coal. The same authors 2 found a 
 considerable evolution of higher olefins when coal 
 was distilled at low .temperatures. E. Bornstein 3 
 discussed the decomposition of coal at temperatures 
 up to 450 C. He found that the gaseous products 
 consisted of 5 to 14 per cent heavy hydrocarbons, 55 to 
 76 per cent paraffins, and 5 to 16 per cent hydrogen. 
 Jones and Wheeler 4 distilled coals at temperatures 
 up to 450 C. in a vacuum of 5 to 40 mm. of mercury, 
 and obtained 6.5 per cent by weight of a tar which con- 
 sisted of 25 per cent olefin hydrocarbons and an equal 
 proportion of naphthenes and paraffins. Pictet and 
 Bouvier 5 have conducted experiments similar to those 
 of Jones and Wheeler, and found a large proportion 
 of hydroaromatic or naphthene hydrocarbons in the 
 tars. Porter and Taylor 6 found that the primary de- 
 composition products of coal were complex easily 
 liquefiable paraffins, hydrocarbons with smaller amounts 
 of water, carbon dioxide and hydrogen. 
 
 One of the points in the propaganda of the various 
 recently developed low temperature carbonization 
 schemes has been the high percentage of light hydro- 
 carbon oils which might be recovered from the tars 
 and used for motor spirit. The gases obtained in 
 these processes are also rich in higher hydrocarbons. 
 Lewes 7 gives the analysis of a gas which contained 10.1 
 per ce*nt of the members of the paraffin series higher 
 than methane. 
 
 White, Park, and Dunkerley 8 found the ethane 
 content of the gases of low temperature carbonization 
 processes to run from n per cent to 47 per cent. 
 Parr and Olin 9 have found that approximately 10 
 per cent of a light hydrocarbon oil was obtained from 
 tars made in low temperature carbonization experi- 
 ments between 400 and 500 C. 
 
 1 J. Chem. Soc., 99 (1911), 649-667. 
 *lbid., 105 (1914), 131-140. 
 
 Z. angew. Chem., 17 (1904), 1520. 
 
 J. Chem. Soc., 105 (1914), 140-151, 2562-2565. 
 
 Compt. rend., 157 (1913). 779. 
 
 Proc. Am. Gas Inst., 1914, 234-288. 
 
 "Carbonization of Coal," p. 164. 
 
 Proc. Michigan Gas Assoc., 17 (1908), 83. 
 
 "Coking of Coal at Low Temperatures," Bull. 79, Univ. of Illinois 
 Experiment Station. 1915. 
 
 (52) 
 
PURPOSE OF THE PRESENT INVESTIGATION 
 
 It is the purpose of this investigation in general to 
 show what results may be expected in the decomposi- 
 tion of an oil if temperature, rate of oil feed, and con- 
 centration of hydrogen are taken into account and 
 carefully controlled. More specifically it is proposed 
 to show: 
 
 i The variation of the composition of the gases 
 made from oil at constant temperature and pressure 
 with changing rate of oil feed. 
 
 2 The effect o'f changing the temperature on the 
 composition of gases made from oil alone at constant 
 oil feed and constant pressure. 
 
 3 The variation in the volume of the various gases 
 obtained per cc. of oil fed at constant temperature and 
 pressure but with changing rate of oil feed. 
 
 4 The effect of changing temperature on the volume 
 of various gases obtained per cc. of oil at constant 
 rate of oil feed and constant pressure. 
 
 5 The extent to which hydrogen is absorbed at 
 any particular concentration, and the effect of changing 
 the concentration. 
 
 6 The influence of hydrogen of certain concentra- 
 tion on the number of cc. of the various gaseous com- 
 ponents obtained per cc. of oil, and the relations be- 
 tween this and change of concentration of hydrogen, 
 change of temperature, and change of oil rate. 
 
 7 The results of a study of the mean molecular 
 weight of the olefins in the gases at certain tempera- 
 ture, and the influence of the presence of hydrogen 
 and change of the rate of oil feed in this connection. 
 
 8 The proportion of aromatic hydrocarbons present 
 in the gases and the influence of the presence of hydro- 
 gen and changing oil rate in this connection. 
 
 9 The percentages of tar formed at different tem- 
 peratures and rates of oil feed, and the influence of 
 hydrogen on tar formation. 
 
 PLAN AND SCOPE OF THE EXPERIMENTAL WORK 
 
 The plan of the present work was to study the de- 
 composition of paraffin hydrocarbons under atmos- 
 pheric pressure, and at a number of temperatures and 
 varying oil rate; and under identical conditions, to 
 investigate the effect of the presence of hydrogen of 
 different concentrations on the decompositions of the 
 paraffin hydrocarbons. 
 
 The working temperatures were 621 C., 723 C. 
 and 825 C. The temperatures used in the commercial 
 
 (53) 
 
manufacture of water gas lie between .700 and 775 
 C., and are thus well within the temperature range of 
 these experiments. At 927 C. two runs were made, 
 but the separation of carbon in the furnace tube was 
 so rapid it was impossible to keep the tube open while 
 the adjustments for the hydrogen-oil gas runs were 
 made. The hydrogen concentrations are discussed 
 under the caption "The Hydrogen Concentration." 
 
 The method used was to adjust the furnace to the 
 proper conditions and to run the oil in at the desired 
 rate. The oil gas so made was collected. Without 
 stopping the flow of oil, hydrogen was then admitted 
 in proper concentration and the gas produced by crack- 
 ing the oil in hydrogen collected in a second gasometer. 
 The two gases were thus made under identical furnace 
 conditions. The volumes of the tars formed were 
 measured. 
 
 The straight oil gas runs made in connection with 
 this research were a repetition of those previously 
 made by Dr. C. M. Alexander (private communication), 
 the results of whose work have not as yet been published. 
 The experimental data of the work recorded here are 
 in excellent agreement with those of Dr. Alexander. 
 
 APPARATUS AND PROCEDURE 
 
 THE FURNACE used in these experiments was designed 
 and built by Whitaker and Alexander, and used by 
 them in a study of the time factor in the making of 
 oil gas. 1 For a detailed description of the furnace con- 
 struction reference must be had to the original article. 
 
 The heating was effected by the passage of an al- 
 ternating current from a single phase, 60 cycle, 50 kilo- 
 watt generator. The current passed through the 
 carbon resistor tube of the furnace was readily con- 
 trolled by means of a field rheostat. Thus a very 
 accurate regulation of the temperature was obtained. 
 Fluctuation limits of i or 2 C. were attained by care- 
 ful operation. 
 
 A constant feed of oil was readily maintained by 
 means of the static head and feed regulator. The oil 
 was vaporized in the prevaporizer. The hydrogen 
 gas was introduced into the sight feed just below the 
 oil feed valve where it mixed with the oil vapors com- 
 ing from the prevaporizer, and passed on into the heated 
 tube of the furnace. The gas velocity in the tube 
 could be calculated from its dimensions (i in. I. D. X 
 38.5 in. long), and the total gas rate. 
 
 i J. Ind. Eng. Chem.. 7 (1915). 484-495. 
 
 (54) 
 
Certain features of the design of the furnace, othe r 
 than the fact that it was susceptible to exact control, 
 which made it particularly suitable for the study of 
 the reactions of hydrocarbon decomposition both from 
 a theoretical standpoint and from an operating stand- 
 point, must be pointed out. It is evident that the gas 
 in passing through the carbon tube is subjected to a 
 set of conditions similar to those existent in the interior 
 of the water gas carbureter and superheater; i. e., 
 heated by carbon-coated passageway walls. Though 
 the size of the furnace tube is less than the voids in 
 the checkerbricking in the gas machine, the general 
 conditions are the same. 
 
 Both from a theoretical standpoint and practical 
 standpoint the study of the kinetics of these various 
 hydrocarbon reactions is greatly to be desired as has 
 been pointed out elsewhere in this paper. The fur- 
 nace was designed in such manner as to avoid catalytic 
 effects as completely as possible, and is therefore suit- 
 able for a study of the kinetics of such reactions. 
 
 MEASUREMENT OF GAS VOLUMES The gases from 
 the runs were collected in 5-cu. ft. holders. The 
 dimensions of these tanks were carefully taken and the 
 volumes computed. A stationary millimeter scale 
 was attached to the tank standard, and a rigid pointer 
 to the movable bell so that readings could be taken at 
 given time intervals, and the gas rates and total 
 volumes calculated. 
 
 A wet meter was used for the measurement of the 
 volume and rate of the hydrogen flowing into the ma- 
 chine. This meter was filled with kerosene to avoid 
 the aspiration of water vapor into the furnace. The 
 oil level in the meter was adjusted carefully at all times, 
 and the meter kept perfectly level. The meter was 
 calibrated against the tank which was used for the 
 collection of the mixed oil-hydrogen gases. The hy- 
 drogen was allowed to flow into the meter, through the 
 furnace heated to 800 C. and into the receiving tank, 
 readings being taken, at short intervals, of the meter 
 rate and tank rate. This calibration was checked from 
 time to time and found not to vary appreciably. 
 
 THE PYROD AND ITS CALIBRATION The temperature 
 measurements were made with a base metal thermo- 
 couple attached to a direct reading Wilson- Maeulen 
 instrument. The thermocouple was calibrated by 
 checking it against the boiling point of sulfur (444.6 
 C.) and against the melting point of sodium chloride 
 (800 C.). The readings of the thermocouple at these 
 
 (55) 
 
points were 425 and 775 C., respectively. Using 
 these data a curve was plotted from which the true 
 temperature could be read corresponding to tempera- 
 tures as read from the instrument. 
 
 The -temperatures at which runs were made were 
 those observed when the pyrod projected through a 
 suitable stuffing box into the lowest of the sight tubes. 
 A hole the size of the end of the pyrod was made in 
 the carbon resistor tube, and the pyrod allowed to 
 project about l /* in. into the interior of the resistor 
 tube. The upper portion of the resistor tube was 
 about 25-30 C. colder than that part of the tube 
 where the pyrod was inserted, due, no doubt, to the 
 cooling effect of the incoming gases and to the endo- 
 thermic reactions taking place in that part of the tube. 
 
 THE HYDROGEN used in these experiments was a very 
 high-grade electrolytic gas. Analyses showed it to 
 contain 99.9 to 100.0 per cent hydrogen. 
 
 THE OIL used was a water-white oil (0.8000 sp. 
 gr.), which boiled between 150 and 265 C. 
 
 METHOD OF OPERATION The furnace was first 
 heated up well and the jacket cooling water adjusted 
 properly. When a constant temperature 10 to 30 
 above the temperature at which the run was to be 
 made had been established, the oil valve was opened 
 and adjusted to the proper rate of flow. The tem- 
 perature of the furnace was then regulated till it re- 
 mained constant at the desired point. The gases 
 formed during this preliminary operation were run to 
 a waste tank, and the tar discarded. 
 
 As soon as all the proper operating conditions 
 had been established, the gases were run into a gas 
 holder and readings of the oil rate, gas rate and tem- 
 perature were taken at suitable short intervals. The 
 pressure was always atmospheric. The temperature 
 was noted at frequent intervals, and any necessary 
 regulations of the field rheostat were made. 
 
 When the proper amount of gas had been collected, 
 the gas was again sent to the waste gas holder, and the 
 tar removed from the tar drip. Hydrogen was then 
 admitted, flowing from the compressed hydrogen tank 
 through the reducing valve and meter and into the 
 admixer at the top of the furnace tube. When the 
 rate of the hydrogen flow and all the other conditions 
 had been adjusted, the gas passing was run into a sec- 
 ond gas holder. Readings of the oil rate, hydrogen 
 rate, total gas rate, and temperature were again taken 
 at suitable intervals. 
 
 (56) 
 
Thus an oil gas and a hydrogen-oil gas were made 
 tinder exactly the same operating conditions. These 
 gases were analyzed 18 to 20 hrs. after making. 
 Analyses were made on low temperature, fast oil rate, 
 gases at periods of one to two hours after making to 
 see if the standing for longer periods caused any differ- 
 ence in the analytical results. No appreciable differ- 
 ence was found. 
 
 RELATION BETWEEN THE EXPERIMENTAL APPARATUS 
 AND THE COMMERCIAL APPLIANCE 
 
 Frequently it is a difficult matter to reproduce ex- 
 perimental results in technical operation. It is with 
 the object of calling attention to salient points which 
 must be kept in mind in order that the results recorded 
 here may be reproduced in a commercial appliance that 
 the paragraphs which follow are written. 
 
 Experimental results show that the time factor is all 
 important. But it must be remembered that the fur- 
 nace tube used in these experiments was only 30 in. 
 long, whereas the column of checkerbrick in the car- 
 bureter and superheater of the water gas set is many 
 times that length. Hence even though a pyrometer 
 may record the same temperature in this furnace and 
 in the checkerbrick of a commercial machine it would 
 not be expected that the gases would have the same 
 composition. 
 
 The diameter of the furnace tube used in these ex- 
 periments was i in. This, however, was soon car- 
 bonized so that it was more nearly 3 / 4 in. The inter- 
 stices in the checkering of commercial machines are 
 of greater sectional area than a circle of 3 /4 in. diameter. 
 Hence the opportunity for heat transfer in the checker- 
 brick is not so good as in this experimental furnace and 
 it would be expected that a longer column of checker- 
 brick than was necessary in this furnace would be 
 necessary to produce a given result, all conditions being 
 the same. 
 
 The actual time of contact of the gases with heated 
 surfaces is all important, but it is a very complicated 
 function of the rate of oil feed, the amount of blue 
 gas introduced, the volume of the checkerbrick voids, 
 and the temperature. 
 
 Commercial operation demands that an apparatus 
 have a reasonable gas making capacity, and for this 
 reason the higher oil rates are most desirable. On the 
 other hand, with high oil rates more tar is always 
 produced than at low oil rates, which makes the use of 
 
 (57) 
 
the oil uneconomical. It is not improbable that the 
 best results could be obtained by designing one ap- 
 paratus with proper control suitable to the production 
 of certain gaseous products, and a second apparatus 
 which would use the tar from this first machine as the 
 carbureting oil. An examination of the tars formed 
 in the experimental apparatus at medium to high rates 
 of oil feed have led to the belief that this would not be 
 impracticable. 
 
 Methods of temperature measurement must not be 
 neglected. Above all it must not be assumed that 
 a particular temperature as measured will produce 
 the same results in two different machines. 
 
 The experiments recorded in this paper merely 
 show the possibilities in the decomposition of a hydro- 
 carbon oil. To obtain particular results on a commer- 
 cial scale would necessitate a great deal of thought as 
 to the proper design for the machine to be used. 
 
 THE HYDROGEN CONCENTRATION 
 
 The introduction of hydrogen in certain concentra- 
 tion is a question which must be regarded from at 
 least two standpoints; i. e., the experimental and 
 the technical or operating standpoints. 
 
 In the manufacture of carbureted water gas the 
 oil is cracked in an atmosphere of carbon monoxide 
 and hydrogen. The final gas is roughly 1 / 3 hydro- 
 carbons, 1 / 3 hydrogen and 1 / 3 carbon monoxide. The 
 hydrogen and hydrocarbons are thus present in the 
 approximate ratio of i volume to i volume in the final 
 gas. In addition there is present i volume of carbon 
 monoxide, the influence of which has never been ex- 
 actly determined by a comprehensive study. 
 
 In these experiments it was thought desirable to 
 study two concentrations of hydrogen: 
 
 (i) i Volume Hydrogen : 2 Volumes (oil gas + tar gas) 
 (2) 2 Volumes Hydrogen : i Volume (oil gas + tar gas) 
 
 The experimental difficulties were such, however, 
 that the introduction of hydrogen in exactly these pro- 
 portions was practically impossible, for, with change 
 in temperature, the amount of gas produced from a 
 given quantity of oil at a particular rate changes, 
 while with change in oil rate at constant temperature 
 the amount of gas produced from a given quantity 
 of oil changes. Also the gas formed from a certain 
 amount of oil is different when the oil is decomposed 
 alone or in an atmosphere of hydrogen. Furthermore, 
 the concentration of hydrogen in the upper part of the 
 
 (58) 
 
furnace tube is much greater at any time than it is in 
 the lower part of the tube, for as the oil vapors pass 
 through the tube a progressive decomposition takes 
 place with formation of a greater volume of hydro- 
 carbon gases. 
 
 It would therefore be necessary to make several trial 
 runs at each oil rate at each temperature to determine 
 the proper hydrogen rate. When it is considered that 
 each run of that sort would consume several hours' 
 time, the impracticability of this method of procedure 
 is apparent. Furthermore, in commercial operation 
 the two factors which would be susceptible to control 
 would be the rate of introduction of blue gas and the 
 rate of introduction of the- oil. So it was decided 
 to base the hydrogen concentration arbitrarily on the 
 oil rate. After some trials at 723 C., the following 
 equations relating to the hydrogen rate and the oil rate 
 were decided upon: 
 (/) To approximate i Vol. Hydrogen : 2 Vols. (oil gas + tar gas) 
 
 Oil rate in cc. per minute _ $ Hydrogen rate in liters 
 
 5.28 I per minute. 
 
 (2) To approximate 2 Vols. Hydrogen : i Vol. (oil gas + tar gas) 
 
 Oil rate in cc. per minute _ \ Hydrogen rate in liters 
 i~;6 ~ ) per minute. 
 
 One set of runs only was made at 621 C. and this 
 at the supposed ratio / Vol. Hydrogen : 2 Vols. Gas. 
 The lower curve in Fi'g. 5 is plotted from the 
 equation given under (i) above. The points repre- 
 sent the actual hydrogen rates and show how closely 
 it was possible to adjust the hydrogen rate to that de- 
 sired as calculated from the oil rate. 
 
 The upper curve in Fig. 5 shows the actual ratio of 
 hydrogen to (oil gas + tar gas). In the furnace the 
 hydrocarbons which compose the tar are of course 
 gaseous and the volume of this gas is calculated on the 
 assumption that the specific gravities of the liquid tars 
 are 0.80 (water = i) and that the mean molecular 
 weight of the hydrocarbons contained in the tar is 
 142; i. e., that the average molecule contains 10 carbon 
 atoms. 
 
 The straight line 'at ordinate 0.5 is the theoretical 
 curve for the value of the ratio of hydrogen to (oil 
 gas -f tar gas). It can be seen that the value of this 
 ratio more nearly approximated unity than it did 0.5 
 in these runs. 
 
 A second series of runs with the value 2 for the ratio 
 of hydrogen to (oil gas + tar gas), was not made for 
 
 (59) 
 
f 30 
 
 .30 
 
 -20 
 
 10 
 
 Point 
 "Y 
 
 G. 
 
 Ro. 
 
 J 
 
 Oil Rate - cc. per minute 
 
 FIG. 
 
 Oil Rate - cc. per minute. 
 
 Point 
 
 Hydroqen 
 
 Actua 
 
 JO 3S 40 
 
 Oil Rate - cc. per minute - - 
 
 EFFECT OF DIFFERENT HYDROGEN CONCENTRATIONS WITH VARYING Om 
 
 RATES AT THREE TEMPERATURES 
 Data for Curves Calculated on Two Bases 
 
 (60) 
 
the reason that from the results at higher temperatures 
 it was judged that the chief effect of more hydrogen at 
 621 C. would be to blow the vapors through the heated 
 tube faster without producing extensive chemical 
 change. Since there was approximately i volume 
 of hydrogen to i volume of (oil gas + tar gas), in these 
 runs at 621 C., that is, practically the relations 
 existent in the carbureter and superheater of the water 
 gas set, it was thought that this series of runs sufficed 
 to show the possibilities of a temperature in the neigh- 
 borhood of 621 C. for purposes of gas manufacture. 
 
 Curves A in Figs. 6 and 7 are the plots of the equa- 
 tion under (2) given above at 723 and 825 C., 
 respectively, and curves B in Figs. 6 and 7 the plots 
 of equation (i) above at those temperatures. The 
 points show how closely the actual hydrogen rates 
 approximated those calculated by these equations. 
 
 The lower curve in the upper half of Fig. 6 shows that 
 the actual value of the ratio of hydrogen to (oil gas + 
 tar gas) approximated very closely to the value 0.5 
 drawn horizontally on the ordinate 0.5. The upper 
 curve in the upper half of Fig. 6 shows that at low oil 
 rates too little hydrogen was introduced to make the 
 value of the ratio hydrogen to (oil gas + tar gas) 
 equal to 2.0, and at high oil rates that too much hy- 
 drogen was introduced. 
 
 The actual values of the ratio hydrogen to (oil gas 
 4- tar gas) in the second series of runs at 825 C. are 
 shown in the curves in the upper half of Fig. 7. In 
 the first series of runs the value of the ratio lies be- 
 tween i.o and 2.0, whereas it was intended that it 
 should be 2.0. In the second series the desired value 
 0.5 is very closely approximated at all oil rates. 
 
 ANALYTICAL PROCEDURE FOR GASES 
 
 During the early part of the work recorded in this 
 paper the method of gas analysis was the ordinary 
 one employing the Hempel burette with single and 
 double pipettes for the absorbing reagents. Thus 
 the carbon dioxide was removed by a solution of i 
 part of KOH in 2 parts of water, the unsaturated 
 and aromatic hydrocarbons by fuming sulfuric acid 
 with 20 per cent free SO 3 , the oxygen by alkaline 
 pyrogallol made up in accordance with HempePs 
 directions, 1 and the carbon monoxide absorption in 
 ammoniacal cuprous chloride prepared according to 
 Winkler. 2 A portion of the residual gas was then mixed 
 
 1 Dennis' "Gas Analysis," p. 160. 
 
 2 C. Winkler, "Handbook Tech. Gas Anal.," translated by Lunge, p. 73. 
 
 (60 
 
with oxygen, and slowly passed back and forth over 
 palladium black in a glass tube immersed in water at 
 a temperature of about 85-90 C. After thus re- 
 moving the hydrogen the gas mixture was exploded 
 over mercury, and the resulting carbon dioxide ab- 
 sorbed in potassium hydroxide. 
 
 "CARBON MONOXIDE" As the work progressed it 
 became evident that this method could be improved 
 upon. It was difficult to understand where the carbon 
 monoxide, varying from 0.2 per cent to 1.9 per cent 
 as shown by the cuprous chloride absorption, could 
 have come from. The furnace was tight, and in any 
 case there would be a slight positive pressure out- 
 wards, so that the ingress of air in more than small 
 amounts was out of the ques.tion. The only other 
 possible source of oxygen was the water in the oil 
 used. However, had the carbon monoxide arisen 
 thus from the reaction of steam on the carbon it would 
 have been present in larger amount at 800 C. than 
 at 600 C. Also there should have been some re- 
 lation between the percentages of carbon monoxide 
 and carbon dioxide. But carbon monoxide, as indi- 
 cated by the absorption in ammoniacal cuprous 
 chloride, was present in largest amount at 600 C., in 
 smaller amounts at 700 C., and least of all at 800 
 C. Also it was noticed that at any particular tem- 
 perature the carbon monoxide tended to be highest 
 when the oil rate was highest. However, there was 
 no regular variation of this sort as there was with the 
 other components of the gaseous mixtures formed. 
 Hence it appeared that the "carbon monoxide" forma- 
 tion as shown by. the absorption in the ammoniacal 
 cuprous chloride was not solely a function of the furnace 
 conditions. 
 
 F. C. Phillips 1 stated that cuprous chloride solu- 
 tion dissolved the higher members of the paraffin 
 series to some extent. G. A. Burrell and F. M. Seibert 2 
 found that cuprous chloride solution caused a contrac- 
 tion of 0.5 to 0.6 per cent in Pittsburgh natural gas. 
 They have also shown that a two-minute contact of 
 cuprous chloride solution with pure ethane caused a 
 loss in volume of 0.6 per cent and that in five minutes 
 the contraction was 1.4 per cent. 
 
 Our experimental work showed that those gases 
 which would be expected to have the largest proportion 
 of high molecular weight hydrocarbons, i. e., those 
 
 i "Oil and Gas Levels." W. Va. Geol. Survey. 1A (1904). 552. 
 1 "The Sampling and Analysis of Mine and Natural Gases," Bur. 
 of Mines. Bull. 42, 46-77. 
 
 (62) 
 
gases made at low temperatures and fast rates of oil 
 feed, were also those which showed the highest per- 
 centages of "carbon monoxide." It appeared certain, 
 therefore, that the contractions found on passing the 
 gases into the ammoniacal cuprous chloride solution 
 were in reality .largely due to absorption of paraffin 
 hydrocarbons such as ethane, propane, and butane 
 rather than to carbon monoxide. 
 
 The use of ammoniacal cuprous chloride was there- 
 fore abandoned, and the carbon monoxide and hydrogen 
 determined by Jaeger's fractional combustion method 
 somewhat as described by H. C. Porter and G. B. 
 Taylor. 1 In place of the vertical Nichrome resistance 
 heater with the inverted U-tube to hold the copper 
 oxide, a horizontal heater with a straight copper oxide 
 tube was used as shown in the accompanying drawing. 
 
 tor factional Combustion o1~CO tnt H z ovtr CvO 
 
 The difference in temperature between the bottom 
 and top of the vertical heater was great enough so 
 that when the oxide was at the proper temperature 
 in one part of the containing tube, it was either too 
 hot or too cold in other parts of the tube. With the 
 horizontal heater and the copper oxide tube running 
 concentrically through it no such difficulty was ex- 
 perienced, nor did the water formed during the com- 
 bustion cause any trouble. 
 
 The fractional combustion of gaseous mixtures over 
 copper oxide consumes a little more time with gases 
 which are low in hydrogen and high in paraffins than 
 does the method in which the cuprous chloride pipettes 
 and the palladium black are used. However, as the 
 average time for carefully made analyses is not over 
 45 mins., this cannot be considered as a serious dis- 
 advantage. 
 
 It was- found that a temperature of 275-280 C. 
 burned the carbon monoxide and hydrogen to carbon 
 dioxide and water without affecting the methane 
 
 1 Proc.Am.GasInst.,9 (1914), 255; J. Ind'.Eng. Ghent.. 6 (1914). 845-8. 
 
 (63) 
 
and other paraffin hydrocarbons present. After pass- 
 ing the gases Over the oxide till no further contraction 
 took place, they were allowed to cool to room tempera- 
 ture, and the volume read. The contraction at this 
 point equals the per cent of hydrogen in the gas. The 
 carbon dioxide was then absorbed in potassium hy- 
 droxide. This contraction is equal to the per cent of 
 carbon monoxide in the gas. The importance of 
 allowing the gases to reach room temperature can 
 readily be seen. Otherwise the contact with the po- 
 tassium hydroxide will cause a contraction due to 
 the lowering of the temperature of the gas. 
 
 The method of burning the carbon monoxide and 
 hydrogen over copper oxide has recently been dis- 
 cussed by G. A. Burrell and G. G. Oberfell. 1 Their 
 experiences with this method were apparently similar 
 to ours. 
 
 In the case of many of the gases obtained in the 
 experimental work, the gas residue after the fractional 
 combustion was small enough so that the whole of it 
 could be mixed with oxygen and exploded over mercury. 
 Here again the method of burning over copper oxide 
 presents a distinct advantage over the methods first 
 mentioned above, for in the combustion with 
 palladium black it is necessary to mix the gas with 
 oxygen previous to the fractional combustion of the 
 hydrogen. Thus a much smaller proportion of the 
 gas could be put through the partial combustion and 
 explosion analysis, in consequence of which accuracy 
 was sacrificed. Accuracy at this point was particu- 
 larly desired in order to calculate the mean molecular 
 weight of the paraffins. 
 
 AROMATIC HYDROCARBONS At the Outset of this 
 
 work the desirability of determining the extent of the 
 formation of hydrocarbons of the benzene series was 
 evident. A careful survey of the literature, however, 
 showed that none of the methods proposed had given 
 satisfaction in the hands of all who had worked with 
 them. It was not until the work recorded here was 
 nearing completion that the method proposed by 
 Hulett and developed by the Bureau of Mines 2 came 
 to our attention. This method was used for the de- 
 termination of the aromatics in the gases from one 
 series of runs at 825 C. 
 
 1 J. Ind. Eng. Chem., 8 (1916), 228-231. 
 
 2 G. A. Burrell, F. M. Seibert. and I. W. Robertson, "Analysis of 
 Natural Gas and Illuminating Gas by Fractional Distillation at Low Tem- 
 peratures and Pressures," Technical Paper, 104, 26-27. 
 
 (64) 
 
The procedure in brief is to evacuate the apparatus 
 with a good pump, after which the gas is admitted 
 and the temperature and barometer readings noted. 
 The gas is allowed to stand for some time (in these 
 determinations 2 to 3 hrs.), in order that the phos- 
 phorus pentoxide may remove the water vapor 
 completely, and then immersed for 10 to 15 mins. 
 in a mush of carbon dioxide snow in acetone contained 
 in a Dewar flask. At the end of this time the gases 
 are sucked off by means of the pump. The bulbs 
 are then removed from the cold bath and allowed to 
 come to room temperature. The temperature is 
 noted and the partial pressure of the volatilized aro- 
 matics is read on the short arm manometer. From 
 this data the volumetric percentage of these com- 
 ponents in the gas can be calculated. A rotary oil 
 pump which gave a vacuum of less than i mm. of 
 mercury, and that in less than a minute, was used in 
 these experiments. 
 
 MEAN MOLECULAR WEIGHT OF THE OLEFINS 
 
 When the analysis and the specific gravity of a gas 
 are known the mean molecular weight of the heavy 
 hydrocarbons of the gas can be calculated. When, 
 in addition to the percentages of carbon dioxide, 
 heavy hydrocarbons, oxygen, carbon monoxide, hy- 
 drogen, methane, ethane, and nitrogen, the percentage 
 of aromatics is known, the mean molecular weight of 
 the olefins and acetylenes can be calculated. 
 
 In determining the specific gravity the temperature 
 of the gas and the barometer reading should be taken, and 
 the gas should be thoroughly shaken with water in order 
 to saturate it with water vapor at the temperature of 
 the room before introducing it into the specific gravity 
 apparatus. Time should be allowed for the subsidence 
 of any mist formed during the agitation with water. 
 
 The method used for the determination of the 
 specific gravity was the so-called effusion method and 
 the apparatus was similar to that described by Pan- 
 nertz. 1 The orifice was made by pricking a piece 
 of thin platinum foil with a needle, and then beating 
 the foil with a small leather mallet till the tiny hole 
 was visible only when held up to a strong light. This 
 foil was mounted on the end of a short brass tube 
 fitting with a carefully turned brass screw cap luted in 
 with litharge and glycerol cement. 
 
 Certain errors are inherent in the experimental 
 methods used. In the first place it is well known that 
 
 J. fUr Gasbel, 48 (1905), 901. 
 
 (65) 
 
fuming sulfuric acid absorbs higher paraffins during 
 the determination of the heavy hydrocarbons. 1 It 
 was thought that the use of bromine water might 
 obviate this difficulty, but analyses on 621 C. gases 
 made by both methods checked to within o.i per cent 
 showing that the two reagents were having similar 
 effects. F. C. Phillips' 2 says that bromine water ab- 
 sorbs the higher paraffins. 
 
 Disagreement between analyses made by the use of 
 bromine water and by fuming sulfuric acid may be 
 occasioned by the fact that bromine water does not 
 brominate benzene. In the analyses made in these 
 experiments there is little doubt but that the benzene 
 would have been completely scrubbed down on ac- 
 count of the time and shaking necessary to obtain 
 complete reaction between the olefins and the bromine 
 solution. Also the proportion of benzene present in 
 a 621 C. gas made at atmospheric pressure is not large. 
 
 The absorption of higher paraffins by the fuming 
 sulfuric acid can be avoided in large part by the use of 
 small portions of the reagent in an apparatus of the 
 type described by G. B. Taylor. 3 It was thought, 
 however, that the time required for analyses made in 
 this way would be too great. 
 
 Another error inherent in the method of determining 
 the mean molecular weight of the olefins lies in the fact 
 that the method for aromatics does not differentiate 
 between benzene, toluene, or xylene. An average 
 molecular weight must be assumed here which evidently 
 is not absolutely correct. In the calculations all 
 aromatics have been regarded as if they were benzene, 
 since it was believed that this hydrocarbon comprised 
 the greatest portion of the aromatics present in the- gas. 
 
 A further error lies in the fact that the paraffins 
 were all assumed to be methane and ethane. Higher 
 homologs of this series are without doubt present as has 
 been shown by Burrell, Seibert, and Robertson in their 
 analyses of carbureted water gas and coal gas by the 
 method of fractional distillation at low temperatures. 4 
 
 A certain proportion of naphthenes or polymethylenes 
 are also no doubt present among the products of the 
 pyrogenic decomposition of the hydrocarbons of 
 
 R. P. Anderson and J. C. Engelder, J. Ind. Eng. Chem., 6 (1914), 989- 
 92; R. A. Worstall, J. A. C. S. t 21 (1899), 245; Orndorff and Young, Am. 
 Chem. J.. 15 (1893), 249; Burrell and Seibert, "Sampling and Analysis of 
 Mine and Natural Gas," Bur. of Mines, Bull 42 (1913), 45-47. 
 
 * "Oil and Gas Levels," W. Va. Geol. Survey, 1A (1904), 522. 
 
 a J. Ind. Eng. Chem., 6 (1914). 845. 
 
 4 Bureau of Mines, Tech. Paper, 104 (1915); J. Ind. Eng. Chem., 7 
 (1915), 17-21. 
 
 (66) 
 
kerosene. To what extent these are affected by the 
 fuming sulfuric acid, and to what extent they are carried 
 through the analysis and credited to the paraffins 
 cannot be stated. Data on the exact behavior of the 
 cyclobutanes, cyclopentanes, and cyclohexanes when 
 treated with fuming sulfuric acid have not been found, 
 if, in fact, such information is at all available. 
 
 In addition to the errors already mentioned is the ex- 
 perimental error in the determination of the specific 
 gravity by the effusion method. Check determinations 
 were always made on the time of flow, and it was found 
 possible to get agreements to within less than o. 5 per cent. 
 
 The following analyses are cited to compare the 
 results obtained by the usual procedure with Hempel 
 burette and pipettes with those obtained by the use 
 of the various modifications discussed above. These 
 will be designated' for the sake of brevity as the "Stand- 
 ard" method and the "CuO" method, respectively. 
 
 GAS No. 24 GAS No. 24 GAS No. 25 
 
 METHOD: Standard CuO Standard CuO Standard CuO 
 
 Carbon dioxide... 0.3 0.0 0.1 0.1 0.1 0.0 0.0 
 
 Heavy hydrocarbons. .. 48.1 48.2 47.9 49.8 49.3 50.8 507 
 
 Oxygen 0.5 0.5 0.5 0.5 0.5 0.4 0.4 
 
 Carbon monoxide 1.2 0.2 0.2 0.9 0.1 1.1 01 
 
 Hydrogen 11.9 11.9 12.1 9.3 9.2 10.4 10.3 
 
 Paraffins 37.3 37.6 37.7 38.1 38.4 37.0 37.2 
 
 Total 99.3 98.4 98.5 98.7 97.6 99.7 98.7 
 
 THE GAS RATE 
 
 The gas rates for three temperatures are plotted 
 against the oil rates in Fig. 8. At 621 C. the gas rate 
 
 as 
 
 per 
 
 VARIATION OP GAS RATE WITH VARIOUS OIL RATBS AT THREE 
 TEMPERATURES 
 
 increases with the oil rate up to an oil rate of 20 cc. per 
 minute, but the introduction of more oil per minute 
 causes no further increase in the gas rate. Apparently 
 
 (67) 
 
with a heated tube of the dimensions used here only a 
 certain quantity of oil can be affected in a given time 
 by heat at a temperature of 621 C. A similar limita- 
 tion would be expected in a commercial machine 
 operating at this temperature. At 723 C. the gas 
 rate is greater with greater oil rate at all oil rates 
 studied. Judging from the shape of the curve and 
 from the analogy of the 621 C. curve it is apparent 
 that an increase in oil rate above 45 to 50 cc. per minute 
 would produce no increase in the gas rate. At 825 C. 
 the gas rate increases with increase in oil rate as can be 
 seen, and at the oil rates studied there is no apparent 
 tendency for a rate to be reached beyond which there 
 is no further increase. Obviously, however, such a 
 point would be reached. 
 
 In comparing the gas rate curves it is interesting to 
 note that at low oil rates the change in temperature 
 from 62 1 to 723 C. has a much greater effect than 
 the change from 723 to 825 C. At low oil rates 
 a temperature of 723 C. is sufficient to gasify per- 
 manently the greatest portion of the oil. The slightly 
 greater production of gas at 825 C. is largely due to 
 the decomposition of methane into carbon and hydrogen. 
 
 As the oil rate increases, a temperature of 723 C. 
 becomes less and less effective for the purpose of per- 
 manently gasifying the oil; 825 C. is much more 
 effective, as can be seen from the increasing divergence 
 between the curves. 
 
 THE DECOMPOSITION OF PARAFFIN HYDROCARBONS 
 
 Figs. 9 to 14 indicate the proportions of the various 
 components obtained when the oil was decomposed 
 alone. Fig. 9 shows that at 621 C. the gases are 
 composed of about 10 per cent hydrogen at all oil 
 rates above 5 cc. per minute. The reactions which 
 contribute to the formation of this hydrogen, named 
 in order of their probable importance, are: 
 
 (/) C 2 H 6 ^ C 2 H 4 + H 2 
 
 00 C 2 H 4 :^ C 2 H 2 + H 2 
 
 (3) CH 4 ^ C + 2 H 2 
 
 (4) C 2 H 2 -> 2 C + H 2 
 
 The conditions are those which are known to be 
 favorable to the condensation of acetylene. The 
 methods for the determination of acetylene are so 
 unsatisfactory that no effort was made to determine 
 the proportion of acetylene in the gases formed. 
 Other investigators have found the amount small in 
 similar experiments. 
 
 (68) 
 
IV 
 
 60 
 SO 
 40 
 
 : 30 
 
 & 
 
 g20 
 
 *J 
 ^ 
 
 fc/o 
 
 ^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 r/G 
 
 . 9 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 s 
 
 ^ 
 
 
 
 82^ 
 
 
 
 
 
 
 _\ 
 
 *- . 
 -~i 
 
 9 
 ~ 
 
 3 
 
 >=^ 
 
 723' 
 ~65? 
 
 M= 
 
 
 
 
 & 
 
 Oil 
 
 10 /5" 20 25 30 35 40 45 J 
 
 Rate - cc. p<rr minute - 
 
 10 
 
 \v 
 
 t. 
 
 
 
 
 
 
 
 
 r/c. 
 
 10 
 
 
 1 
 
 -Q 50 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 k ^^ 
 
 1. 
 
 ^L 
 
 *JO 
 
 XJ 
 
 ^^n 
 
 > 
 
 X. 
 
 ^^ 
 
 
 
 
 
 
 
 >7?Vr 
 
 
 
 '*~ - 
 
 . ~^^ 
 
 -* 
 
 *=Z 
 
 
 
 - 
 
 <-$ 
 
 '5*C 
 
 
 
 
 
 
 
 
 
 
 
 | 
 
 ^ /0 
 ^? 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 J /O /5 20 25 JO J5 4(5 45" J- 
 
 0/7 /?flf<r - c.c. per minute - 
 
 75 20 25 30 35 
 
 Oif Rate - cc. per minute 
 
 (69) 
 
The rapid increase in the proportion of hydrogen 
 with decreasing oil rates below 5 cc. per minute is 
 thought to be due to a marked increase in the extent 
 to which reactions (j) and (4) take place. At 723 C. 
 only a slightly larger proportion of hydrogen is formed 
 than at 621 C., indicating that reactions (j) and (4) 
 do not become rapid up to this temperature. A very 
 marked increase in the hydrogen production takes 
 place when the temperature is elevated to 825 C., 
 chiefly due to an increase in reactions (3) and (4), as is 
 evidenced by the very large amount of free carbon 
 which was liberated and which tended to stop the 
 furnace tube. No trouble was occasioned by this 
 carbon deposition at 621 C. or 723 C. 
 
 Fig. ii shows the proportion of illuminants, or com- 
 ponents removed by 20 per cent fuming sulfuric acid, 
 in the oil gases made at the three temperatures. The 
 percentage of illuminants is highest in the gases made 
 at 62iC. at all oil rates, and 'remains practically 
 constant at 52 per cent at all oil rates above 10 cc. per 
 minute. The proportion of illuminants formed at 
 723 C. is higher than at 825 C. except possibly at 
 high Oil rates. It is thought, however, that the pro- 
 portion of illuminants at high oil rates would not be 
 greatly different at any of these temperatures. 
 
 The fact that the proportion of illuminants is lower, 
 the higher the temperature at moderate to low oil 
 rates is due to the secondary reactions of these hydro- 
 carbons. Ethylene is decomposed into carbon and 
 methane to some extent. Condensation to naphthenes 
 takes place. At these low oil rates the proportion of 
 hydrogen present is considerable, and the higher the 
 temperature the higher this percentage. Hence it 
 would be expected that hydrogenation reactions such 
 as C 2 H 4 + H 2 ^ C 2 H 6 and C 2 H 4 + 2H 2 ^ 2CH* 
 would take place. 
 
 The relation between the proportions of saturated 
 hydrocarbons formed at different temperatures is 
 shown in Fig. 10. It can be seen that the proportion 
 of saturateds in the gas is higher at 723 C. than at 
 621 C. at all oil rates. This no doubt is due to the 
 faster speeds of such reactions as C 2 H 4 > C + CHi 
 and C 2 H 6 + H 2 ^1 2CH 4 . 
 
 It might reasonably be expected that the percentage 
 of saturateds in the gases would be greater at 825' C. 
 than at 723 C., but it can be seen from Fig. 10 that 
 this is not the case. At low oil rates the proportion 
 of saturateds is lower at 825 C. than at 723 C. 
 
 (70) 
 
JO 
 
 60 
 50 
 An 
 
 
 
 
 
 62 
 
 /C 
 
 
 r/G. 
 
 12 
 
 
 
 
 
 
 
 
 
 ///j. 
 
 
 
 V 
 7 
 
 X 
 v^ 
 
 
 ir 
 
 
 
 
 
 
 
 / componem 
 
 N> C4 J 
 O O O O < 
 
 / 
 
 
 
 
 
 r 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 />i 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Oil 
 
 r 10 15 20 25 30 35 4O 4-5 J< 
 
 y9afe - cc. per minute - 
 
 / u 
 
 60 
 50 
 40 
 30 
 20 
 10 
 O 
 
 
 
 
 
 723 
 
 'C 
 
 
 ri&. 
 
 / J 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 V 
 
 
 
 
 
 
 nit. 
 
 
 
 <r 
 
 
 
 
 
 Sat. 
 
 
 
 f 
 
 /" 
 
 
 
 
 
 
 
 
 
 1 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 ~- . 
 
 > 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Oil 
 
 /O IS 20 25" J( 
 
 Rate - cc. per mi 
 
 rwte 
 
 /O /5 20 25 JO JJ ^0 45 SO 
 
 Oil Rate - cc. per minute 
 
 (71) 
 
This is chiefly due to the fact that the reaction 
 CH 4 > C -f 2H 2 has an appreciable velocity at this 
 temperature. As the oil rate increases it would be 
 reasonable to expect that the effect of this reaction 
 would be less and less. However, the divergence be- 
 tween the 825 C. curve, and the 723 C. curve be- 
 comes greater the faster the oil rate. This state of 
 affairs is apparently due to the superimposed effect of 
 another reaction. When the high paraffin hydro- 
 carbons first break down under these conditions the 
 chief products are low molecular weight paraffin 
 hydrocarbons and high molecular weight olefins. At 
 621 C. and 723 C. a considerable proportion of the 
 high molecular weight olefins formed pass on through 
 the tube into the tar. The tars were larger in amount 
 the lower the temperature, and when the tar curves 
 are studied it will be seen that there is not a marked 
 difference between the tar formation at 621 C. and 
 723 C. but that a considerable difference is found at 
 825 C. These low temperature tars were treated 
 with concentrated sulfuric acid in the cold, and it was 
 found that 2025 P er cent by volume was removable in 
 this fashion, the greater portion of which was doubt- 
 less olefins of high molecular weight. At 825 C. the 
 effect of the heat is sufficient to break down these 
 higher olefins almost completely. In the process 
 ethylene and propylene are formed in large quantity 
 with the result that the percentage of saturated 
 hydrocarbons drops. The proportion of illuminants 
 and saturated hydrocarbons present in the gases at 
 825 C. is greatly lowered on account of the high 
 percentage of hydrogen present in these gases. 
 
 Figs. 12, 13 and 14 show the relations between the 
 component illuminants, saturateds, and hydrogen at 
 the temperatures 621 C., 723 C. and 825 C., 
 respectively. Of all the proportions of these com- 
 ponents those at the high oil rates at 621 C. most 
 nearly represent the products of the primary decom- 
 position of the oil. The hydrogen is largely the re- 
 sult of secondary reactions. If the illuminants and 
 saturateds are calculated to a 100 per cent basis the 
 proportion is 58 per cent illuminants and 42 per cent 
 saturateds. Such a ratio as this would be expected in 
 the reaction if the primary decomposition of the 
 paraffin was: 
 
 C w H 2M + 2 > C M _iH 2 -2 + CH 4 
 Paraffin Olefin 
 
 (72) 
 
and, if then, the high molecular weight olefins in part 
 broke down to lower olefins. 
 
 The mechanism of the reactions at work can be 
 judged somewhat from a consideration of the relations 
 between the curves for the percentages of illuminants, 
 saturateds, and hydrogen. Thus, in Fig. 12, if it is 
 assumed that the illuminants are chiefly ethylene and 
 the saturateds largely methane it can be seen that the 
 normal formation of ethylene is 52, that of methane 
 39, and that of hydrogen 10 per cent. 
 
 Consider the proportions in the gas at an oil rate 
 of 2.5 cc. per minute. They are ethylene 34, methane 
 50, and hydrogen 16 per cent. The decrease in 
 ethylene has been 18 per cent on the basis of the total 
 gas. If this had been due to the reaction C 2 H 4 > 
 C + CH 4 the methane should have increased 18 per 
 cent on the basis of the total gas. In fact it increases 
 only ii per cent. 
 
 The hydrogen increases 6per cent on thetotal gas basis. 
 If this were due to the reaction CH 4 > C 4- 2H 2 
 a 3 per cent decrease in the methane should have taken 
 place. But as has been seen, other reactions also give 
 rise to hydrogen. Assuming, however, that the above 
 reaction was the sole change of this sort, only 14 per 
 cent out of the 18 per cent increase in methane which 
 should have been found, had the ethylene reacted 
 entirely with formation of carbon and methane, would 
 be accounted for. Hence the ethylene must be re- 
 moved in other ways, for example by the condensation 
 to naphthenes. 
 
 It is not probable that the 6 per cent increase in the 
 hydrogen is entirely due to the decomposition of 
 methane. Dehydrogenation of naphthenes, dissocia- 
 tion of ethylene and ethane, etc., may give rise to 
 hydrogen. Then, too, not all the methane formed 
 comes from the decomposition of ethylene. Decom- 
 position of higher olefins in such a manner as 
 
 CH 4 + CH 2 =CH CH=CH 2 
 
 and hydrogenation of olefins, C 2 H 4 + 2H 2 < > 2CH 4 , 
 may contribute. 
 
 A similar argument may be worked out for the rela- 
 tion between the illuminants, saturateds, and hydrogen 
 at 723 C. and 825 C. 
 
 The marked decrease in illuminants with decreasing 
 oil rates at 825 C. is notable. No corresponding 
 increase in saturateds takes place. It is apparent 
 
 (73) 
 
(U 
 
 o 60 
 10 50 
 
 S. 
 
 
 
 
 
 
 
 r/ G 
 
 /5 
 
 
 
 
 
 
 
 * 
 
 
 Illj. 
 
 
 
 /^ 
 
 
 6 
 
 
 
 
 
 
 
 ^ 30 
 
 
 ^ 
 
 
 6 
 
 
 
 
 T f 
 
 
 
 
 
 
 
 
 
 
 
 
 
 | 20 
 | 10 
 
 
 
 
 
 
 Oil 
 
 yd 5 
 705 ^ 
 
 /* : 
 
 
 
 
 
 6 
 
 
 
 62/'< 
 
 : // 
 
 
 ; 
 
 
 
 
 
 
 
 I 10 
 
 
 
 
 
 
 
 
 
 
 
 Oil 
 
 10 IS 20 25 JO J5 40 45 50 
 
 Atafe - cc. per minute 
 
 \ 
 
 
 
 
 
 
 
 r/c 
 
 . /6 
 
 
 
 
 ^ 
 
 
 
 
 
 ///^. 
 
 
 
 
 ff^^ 
 
 
 
 
 
 
 
 
 
 ' 4f\ 
 
 
 W 
 
 *5*^ 
 
 
 
 
 s 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 
 
 
 
 
 
 
 
 
 % Component 
 
 r N* < 
 O C 
 
 
 
 
 
 
 Oil 
 
 Oil ( 
 
 ;as y 
 
 ^: 
 
 
 
 
 
 6 
 
 
 
 7; 
 
 :rc 
 
 /"; 
 
 :26 
 
 ers 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Oil 
 
 10 /x 
 
 Rate - c 
 
 r 20 25 JO J5 4O 45 JO 
 
 c. per minute 
 
 t /u 
 
 to 
 
 '^ 60 
 ^J 
 
 CO 
 ^ 50 
 
 
 i* 
 
 ^- 
 30 
 
 v 
 
 | 2 
 
 h 
 
 Vj 
 ^ 
 
 x\ 
 \\ 
 
 
 
 
 
 
 
 no 
 
 ,/7 
 
 
 
 V 
 
 cTS* 
 
 
 
 
 juL 
 
 4 
 
 
 
 <s 
 
 60*" 
 
 
 
 
 
 2 
 
 ^| 
 
 6 
 
 6<* 
 
 
 Sot 
 
 
 
 
 
 
 
 
 
 --J 
 
 
 y/ 
 / 
 
 
 
 
 
 
 
 
 
 
 7 
 
 
 
 
 
 
 0/7 G 
 Oil G 
 
 as 
 
 fl-S V- 
 
 HZ - 
 
 o 
 d 
 
 
 
 723- 
 
 C 2 
 
 H 2 :l 
 
 yds 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 S 10 15 20 25 30 3f 40 *5" / 
 
 Oil Rate - cc. per minute - * 
 
 (74) 
 
from the high percentage of hydrogen that methane 
 is decomposing extensively into carbon and hydrogen. 
 
 PROPORTION OF ILLUMINANTS AND SATURATED HYDRO- 
 CARBONS 
 
 In order that the effect of hydrogen on the com- 
 position of the gases as regards illuminants and satu- 
 rated hydrocarbons may be seen, Figs. 15, 16, 17, 18 
 and 19 are shown. These curves were drawn by 
 calculating the illuminants and saturated hydro- 
 carbons to a basis of 100 per cent. In this manner the 
 proportions of the two classes of compounds can be 
 seen. 
 
 It might be expected that hydrogenation reactions 
 would play an important part and that the proportion 
 of saturated hydrocarbons would be higher in the oil- 
 gas-hydrogen runs than in the straight oil-gas runs. 
 It will be shown later in this paper that hydrogenation 
 does take place to a considerable extent. 
 
 It should be noticed that at low temperatures the 
 presence of the hydrogen has no influence on the 
 relative amounts of illuminants and saturated hydro- 
 carbons at high oil rates; i. e., its presence has little 
 effect on the mechanism of the primary decomposition 
 and the early stages of the secondary decomposition. 
 But at low oil rates where the gases are exposed to the 
 effect of heat for a longer time, and where extensive 
 secondary and tertiary changes take place the hydrogen 
 has a considerable influence at temperatures of 723 C. 
 and 825 C. At 621 C. the influence of the hydrogen 
 is not marked. 
 
 It will be remembered that the first reaction under- 
 gone by a paraffin hydrocarbon when it is thermally 
 decomposed is that which gives rise to a high molecular 
 weight olefin and a low molecular weight paraffin. 
 It is probable that the higher the molecular weight of 
 an olefin the more readily it is hydrogenated. 
 
 If the high molecular weight olefins are hydro- 
 genated paraffins would be formed. These would 
 again decompose into long chain olefins and low 
 molecular weight paraffins. This sequence of re- 
 actions may be represented as follows: 
 
 C M H 2 + 2 > C M -iH 2 -2 + CH 4 
 Paraffin Olefin 
 
 C M 1 H 2 n 2 + H2 > Cn 1 H2n 
 
 Olefin Paraffin 
 
 C n 1 H2n *" C M 2H2n 4 + CH 4 
 
 Paraffin Olefin 
 
 (75) 
 
The net result of such a sequence of reactions would 
 be an increase in the proportion of paraffins in the 
 gas. But as can be seen in the figures, the proportion 
 of paraffins is less in the cases of the hydrogen-oil 
 gases. However, at high oil 'rates where secondary 
 reactions are not so important, and in the series of 
 runs at 825 C. where the concentration ratio was 
 
 IO 15 20 25 JO 35 40 45 5( 
 
 Oil Rate - cc. per minute 
 
 40 45 
 
 ^0 
 
 5 10 15 20 25 30 35 
 
 Oil Rate - cc. per minute - - 
 
 2H2 : i gas (Figs. 17 and 19), the proportion of 
 saturated hydrocarbons is greater in the hydrogen-oil- 
 gas runs. It is probable, therefore, that the above 
 reactions take place in all cases but that the effect of 
 the hydrogen at low oil rates on the extensive secondary 
 
 (76) 
 
changes is so great that the result of the hydrogenation 
 is masked. 
 
 Apparently the low molecular weight olefins are not 
 hydrogenated to a large extent, for were this 'the case 
 the proportion of saturated hydrocarbons present 
 would be greatly increased. This is not the case. 
 
 The increase in the proportion of olefins in the 
 hydrogen-oil-gas runs at low rates of oil feed may be 
 accounted for in two ways: 
 
 (i) The effect of the hydrogen may be to increase 
 those reactions which give rise to olefins. 
 
 (2) The effect of the hydrogen may be to retard 
 those reactions which tend to remove or destroy the 
 olefins. 
 
 The largest proportion of the ethylene and propylene 
 present comes from the direct splitting up of high 
 molecular weight olefins. If an olefin of fairly high 
 molecular weight may be used to illustrate, this re- 
 action may be represented: 
 CH 3 CH 2 CH 2 CH=CH 2 ^CH 3 CH=CH 2 + CH 2 =CH 2 
 
 It will be noticed that the result is an increase in 
 volume. The equilibrium point of the reaction would 
 therefore be shifted by a diminution of pressure in 
 such direction as to favor the production of ethylene 
 and propylene. The introduction of hydrogen has the 
 same effect as a reduction in pressure, and would 
 therefore have a similar effect on the equilibrium. 
 
 Chief among the reactions which remove olefins 
 such as ethylene and propylene are condensation 
 and decomposition. These reactions may be rep- 
 resented as 3C 2 H 4 > C 6 Hi 2 and C 2 H 4 > C + CH 4 , 
 
 The presence of the hydrogen would displace the 
 equilibrium point of the first of these reactions in 
 favor of the ethylene. The latter reaction, insofar 
 as the influence of volume relationships on the equilib- 
 rium point is concerned, would not be affected. 
 
 Whether the displacement of the equilibrium points 
 in any of these reactions is sufficient to be worthy of 
 mention can not be said. In no case is the equilibrium 
 condition attained, but the speeds of the various re- 
 actions would depend on the difference between the 
 actual condition of the system and the equilibrium 
 condition; hence any displacement of the equilibrium 
 point would be important. 
 
 The increase in the proportion of olefins may be 
 looked upon from another angle. When hydrogen is 
 introduced along with the gas the time of contact of 
 the gas with the heated tube is diminished, due to the 
 
 (77) 
 
increase in the total volume passing in unit time. If 
 it is the case that the reactions which give rise to the 
 olefins ethylene and propylene are fairly rapid, while 
 those which destroy them are slower, the summational 
 effect of an increased gas rate would be an increased 
 proportion- of olefins. That the speed of the re- 
 actions which produce olefins is fairly great can be 
 seen by reference to Fig. 22,' which shows the mean 
 molecular weight of the olefins formed at 825 C. 
 The mean molecular weight lies between 30 and 34. 
 The molecular weight of ethylene is 28, while that of 
 propylene is 42. The proportion of olefins higher 
 than propylene cannot be great, therefore, and it 
 would seem that they break down largely to ethylene 
 and propylene. That the reactions which cause a 
 removal or destruction of ethylene are only moderate 
 in speed has been seen under the discussion of the 
 reactions of ethylene in the first part of this paper. 
 
 This latter explanation appears more probable than 
 the one concerning the displacement of the equilib- 
 rium points of reactions; however, both of these 
 effects may be concerned in the production of the re- 
 sults observed. 
 
 It will be noticed that the two different hydrogen 
 concentrations produce similar results at 723 C. but 
 that at 825 C. the proportion of olefins is much 
 higher when the concentration ratio is 2 HI : i Gas. 
 The rather large difference in this last case indicates, 
 that the chief effect of the hydrogen is due to its 
 cutting down the time of heating, for at 825 C. the 
 decomposition of the higher olefins to ethylene is 
 doubtless very rapid, and takes place extensively in 
 spite of the reduced time of heating in the hydrogen- 
 oil gas runs. 
 
 THE ABSORPTION OF HYDROGEN 
 
 Calculations from the analytical data show that a. 
 considerable absorption of hydrogen takes place when 
 the oil is cracked in an atmosphere of this gas. Haber 
 was of the opinion that the oil produced no hydrogen 
 by its own decomposition when it was cracked in an 
 atmosphere of hydrogen. On the other hand, it might 
 be assumed that there would be as much hydrogen 
 produced under these conditions as when the oil was 
 cracked alone. These two assumptions offer two bases 
 on which the absorption gf hydrogen may be calcu- 
 lated. 
 
 (i) If no hydrogen is produced by the cracking 
 of the oil, the hydrogen absorption per cc. of oil would 
 
 (78) 
 
be equal to the difference between the hydrogen added 
 and that present in the final gas divided by the total 
 number of cc. of oil. 
 
 (2) If the oil produces as much hydrogen as when 
 cracked alone, the difference between the hydrogen 
 added plus that normally produced from the oil at the 
 particular oil rate and the hydrogen in the final gas 
 represents the absorption. This divided by the total 
 number of cc. of oil gives the absorption in cc. per 
 cc. of oil. 
 
 Fig. 21 shows the absorption per cc. of oil calcu- 
 lated on basis (i) for the several temperatures and 
 concentrations. Fig. 22 shows these absorptions calcu- 
 lated on basis (2). 
 
 It will be seen that the curves of Fig. 22 are much 
 smoother and more regular than those of Fig. 21, not 
 because they are drawn more smoothly, but because 
 the points fall on smoother curves. The general form 
 of the curves, too, in Fig. 22 is that which would be 
 expected from a consideration of the curves for the 
 hydrocarbon, components formed per cc. of oil. The 
 curves of Fig. 21 show no general similarity to each 
 other, while those of Fig. 22 show similar general 
 characteristics. The curves representing the forma- 
 tion of all the other components of the gases show 
 regular variations, and it would be expected that this 
 regularity would extend to the curves for hydrogen 
 absorption. 
 
 In general, therefore, it seems that the basis on 
 which the curves of Fig. 22 are calculated is more 
 nearly correct than the basis which assumes that no 
 hydrogen is produced from the oil when it is decom- 
 posed in hydrogen. This view is strengthened by 
 the fact that in the case of the 825 C. gases with 
 hydrogen concentrates iH 2 : 2 Gas, the absorption 
 curve in Fig. 21 falls below the o.o line, i. e., hydrogen 
 must have been formed from the oil since there was 
 more hydrogen in the final gas than was added through 
 the meter. 
 
 It is probable that the true value for the hydrogen 
 absorptions for any set of conditions falls between the 
 two values as calculated from the two limiting as- 
 sumptions. It is thought that the true values are 
 slightly less than the values of the absorptions as they 
 would be read from the curves in Fig. 22. 
 
 The curves of Fig. 22 show the interesting fact that 
 at any particular temperature the hydrogen absorption 
 per cc. of oil decreases with increasing oil rate. The 
 
 (79) 
 
200 
 
 i 
 
 I ' 
 
 ^ 
 
 8? 5V 
 
 8; 
 
 r/s. 
 
 1 6as 
 
 "ZTC 
 
 3 10 IS ZQ 25 JO JJ 40 45 JO 
 
 Ci Oil Rate - cc. per minute - 
 
 < -ten, . 
 
 in 
 
 I 
 
 
 
 8 
 
 g 
 
 ISO 
 
 - cc. per minute 
 
 t: 
 
 fcr 
 
 30 
 
 
 
 
 825 
 
 t 
 
 IH t : 
 
 2G( 
 
 15 
 
 
 no. 
 
 25 
 
 
 
 
 
 
 
 
 
 
 
 ^^f*^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 -^ I0 
 
 5.0 
 
 v*> 
 
 
 
 
 
 
 
 out 
 
 Oil 6 
 
 as 
 
 flS + 
 
 
 
 //, - 
 
 
 
 
 
 
 
 
 
 
 
 
 
 J /O /J- 2(? 2S 30 35 40 45 S 
 
 Oil Rate - cc. per minute. - 
 
 
 
 
 BZ5' 
 
 '", 
 
 1* i 
 
 ^6a 
 
 
 PIG. 
 
 4 
 
 *~~ 30 
 
 
 
 
 
 
 
 
 
 
 
 Q 20 
 
 
 
 
 
 
 
 Oil (. 
 Oil ( 
 
 *l , 
 
 
 
 */y,- 
 
 -- 
 
 D 
 
 
 
 
 
 
 -o- 
 
 
 
 
 
 ^ 1.0 
 
 o "^^ 
 
 
 
 
 
 
 
 
 
 
 -> /O /i 20 2i JO J5 40 45 J 
 
 Oil Rate - cc. per minute 
 (80) 
 
great importance of the time factor is well brought 
 out here. At constant oil rate, and approximately 
 the same hydrogen concentration the absorption per 
 , cc. of oil is greater the higher the temperature. There 
 would, however, be an upper limit to this on account 
 of the excessive decomposition of all hydrocarbons at 
 elevated temperatures. 
 
 The effect of increasing the concentration of hy- 
 drogen is clearly shown in Fig. 22, for the curve for the 
 2 Hz : i Gas runs in above the curve for the iH z : 2 
 Gas runs at both 723 C. and 825 C. The spee.d of 
 hydrogenation reactions is greater the higher the 
 concentration of hydrogen. 
 
 It is interesting to note that the curve for the 
 i H 2 : 2 Gas runs at 825 C. falls below the curve for 
 the 2 Hz : i Gas runs at 723 C. This shows that the 
 increasing temperature is tending to cause dehydro- 
 genation reactions or hydrocarbon dissociations to a 
 marked degree at 825 C. The effect of hydrogen 
 in greater concentration in reversing these dissocia- 
 tions is clearly brought out when the position of the 
 2 H 2 : i Gas curve for 825 C. is considered in its 
 relation to the iH 2 : 2 Gas curve at this same tempera- 
 ture. 
 
 MEAN MOLECULAR WEIGHT OF THE OLEFINS 
 
 In Fig. 23 the mean molecular weight of the olefin 
 hydrocarbons in gases made at 825 C. in oil-gas 
 runs and in hydrogen-oil-gas runs with the concentra- 
 tion ratio i H z : 2 Gas can be seen. 
 
 It should be kept in mind that the molecular weight 
 of ethylene is 28 and that of propylene is 42. From 
 the position of the curves it can be seen that ap- 
 proximately one-third of the olefins is propylene. 
 The curves lie very close together, and it is impossible 
 to say just what the influence of the hydrogen is on 
 the formation of the olefins. 
 
 If the method of calculation of the molecular weight 
 of the olefins, as explained under the discussion of the 
 analytical methods, is considered, it is apparent that 
 all the analytical errors pile up and are brought out 
 in this calculation. This no doubt accounts for the 
 irregularity in the curve, and also for the fact that 
 there is no consistent difference in the position for the 
 oil-gas run and the hydrogen-oil-gas runs. 
 
 It was thought that certain differences might be 
 brought to light by the curves for the mean molecular 
 weights of the olefins. If the higher olefins were more 
 easily hydrogenated than ethylene the curve for the 
 
 (81) 
 
mean molecular weight of the olefins in the hydrogen- 
 oil-gas runs would fall below that of the oil runs. 
 If, on the other hand, the presence of the hydrogen, on 
 account of its causing a more rapid passage of the gas 
 through the tube, resulted in a less extensive decom- 
 position of the higher olefins, the curve for the hy- 
 drogen-oil-gas runs would lie above that for the oil- 
 gas runs. 
 
 It may be thought that these two effects are balanc- 
 ing each other with the result that the curves are 
 practically the same. It would have been desirable 
 to have carried out a similar series of runs with a high 
 concentration of hydrogen, but the calculation of the 
 mean molecular weight of the olefins can be made only 
 when the per cent of benzene in the gas is known, and, 
 as has been noted, the method for the determination 
 of benzene was found only as this experimental work 
 was drawing to a close. 
 
 THE FORMATION OF AROMATIC HYDROCARBONS 
 
 Fig. 24 shows the percentage of aromatic hydro- 
 carbons present in the gases made at 825 C. when 
 oil is cracked alone or in hydrogen when the con- 
 centration ratio is iH 2 : 2 (Oil Gas + Tar Gas). The 
 method of determining these percentages "has been 
 described under the analytical methods. 
 
 The percentage of aromatics appears to increase 
 slightly with increase in oil rate. Whether this is 
 actually the case or not cannot be definitely stated. 
 The exact opposite would be expected. It is thought 
 that the apparent increase may be due to the freezing 
 out of high molecular weight hydrocarbons of other 
 types than the aromatic compounds. High molecular 
 weight paraffins and olefins are present in greater 
 proportion in the gases made at high oil rates than in 
 those made at low oil rates. 
 
 The smaller proportion of aromatic hydrocarbons 
 present in the gases made at low oil rates may possibly 
 be due to the removal of benzene to form compounds 
 such as diphenyl, naphthalene and anthracene, which 
 pass largely into the tars. 
 
 The hydrogen apparently has little effect on the 
 formation of aromatics at low oil rates, but decreases 
 the aromatic formation somewhat at higher oil rates. 
 This is possibly due to the retarding effect which the 
 presence of hydrogen would have on the formation of 
 aromatics or hydroaromatics by condensation re- 
 actions. Less gas is formed from the oil at high oil 
 
 (82) 
 
10 
 
 
 r/6. 
 
 25 
 
 
 
 
 
 
 . 
 
 o 
 
 60 
 
 
 
 
 
 / 
 
 p 
 
 . ' 
 
 
 
 o 
 
 
 
 
 
 _^c* 
 
 x' 
 X 
 
 "'" 
 
 
 
 
 
 JO 
 
 
 y 
 
 p rn 
 
 ^' 
 
 
 
 
 
 
 
 
 
 
 
 
 
 0/7 C 
 
 JflS 
 
 
 
 
 
 
 
 
 
 
 
 0/7 C 
 
 a5 * 
 
 / - 
 
 
 
 >c U 
 
 A 
 
 , 
 
 
 62/ 
 
 'C 
 
 ///,: 
 
 26C 
 
 5 
 
 
 
 
 
 r 
 
 
 
 
 
 
 
 
 
 
 ; 
 
 F /O /5 20 25 JO 35 40 45 JC 
 
 0/7 Pate - cc. per minute. - 
 
 10 
 
 60 
 
 
 r/6. 
 
 26 
 
 
 
 
 ^ 
 
 ^ 
 
 ^r 
 
 
 
 
 
 
 
 ^ X 
 
 X* 
 
 
 
 
 JU 
 
 
 
 
 
 / 
 / 
 / 
 
 X 
 
 
 
 
 
 40 
 30 
 
 
 
 
 y 
 
 '/ 
 
 72 
 
 re 
 
 IH 2 
 
 :26 
 
 (LS 
 
 
 
 /' 
 
 77 
 
 J 
 
 
 
 
 
 
 
 -c C* 
 /O 
 
 
 
 / 
 
 >^ 
 
 f 
 
 
 
 (9/7 < 
 0/7 t 
 
 ?<25 
 
 ids # 
 
 HI- 
 
 __ _ 
 
 
 
 
 
 
 
 
 
 
 
 
 (?/7 
 
 r /0 /J" 20 2J 30 JJ -/5 JC 
 
 /^arc ' cc. per minute. - 
 
 IU 
 
 60 
 
 
 r/6 
 
 21 
 
 
 
 
 f s 
 
 r^^" 
 
 r^^ 
 
 . 
 
 
 -f O 
 
 
 
 
 P^ / 
 
 P 
 
 
 
 
 
 
 30 
 20 
 
 /O 
 
 r 
 
 
 
 
 
 
 
 
 723' 
 
 / j 
 
 1*4 
 
 /6as 
 
 
 
 
 y 
 
 y 
 
 
 
 
 
 
 
 x 
 
 ^/ 
 
 7* 
 
 
 
 OH 
 Oil 
 
 0(2-5 
 
 ^ 
 
 
 
 
 " 
 
 
 
 
 
 
 
 
 
 
 o// 
 
 ' /O 15 2Q 2S 3O 3, 
 
 Rate - cc. p^r minutt 
 
 J" 40 45 J"< 
 
 9 ....-_ - - - - 
 
 (83) 
 
rates than at low oil rates, and as a consequence the 
 concentration of hydrogen is greater at high oil rates 
 than at low oil rates. 
 
 TARS 
 
 The tars were collected from the tar drip and the 
 volume measured. This volume divided by the 
 volume of the total oil used and multiplied by 100 
 gives the percentages of tar formed. Figs. 25 to 29 
 show these tar percentages for both oil-gas and 
 hydrogen-oil-gas runs plotted against the oil rate at the 
 temperatures indicated. It should be mentioned that 
 at low oil rates these percentages are not accurate. 
 The low-oil-rate tars are heavy and viscous, and as a 
 result do not run down through the condenser as 
 easily as the lighter high-oil-rate tars. 
 
 As far as can be judged from the curves in Figs. 
 25, 26, and 27, for temperatures of 621 C. and 723 C. 
 there is no marked regular difference between the tar 
 formation in the oil-gas runs and the hydrogen-oil- 
 gas runs. 
 
 At 825 C. the percentage of tar in the oil-hydrogen 
 runs is consistently less than in the oil runs except at 
 low oil rates where, as has been mentioned, the tar 
 percentages as shown mean little. This difference is 
 more marked in the curves of Fig. 26 where the con- 
 centration ratio was 2 Hydrogen: i (Oil Gas + Tar 
 Gas) than in the curves of Fig. 28. 
 
 In general two classes of compounds are contained in 
 the tar: first,' Unchanged or partially changed oil; 
 second, synthetic hydrocarbons which are the products 
 of extensive change. It may be thought that hy- 
 drogen, on account of its decreasing the time of contact 
 of the hydrocarbon vapors with the heated tube, 
 would tend to increase the proportion of tar since the 
 decrease in the time of heating would cause a less 
 extensive decomposition of the oil vapors. On the 
 other hand, this decrease in the time of heating would 
 also diminish the extent to which synthetic reactions 
 resulting in the formation of tarry products would 
 take place. Also the percentages of the hydrogen 
 would retard these reactions, since they are all re- 
 actions which result in decrease of volume. Ap- 
 parently these effects are balanced at temperatures 
 of 723 C. or below. At 825 C., however, the per- 
 centage of tar is less. This leads to the belief that 
 synthetic reactions are responsible for a considerable 
 proportion of the tars at temperatures in the neighbor- 
 hood of 825 C. 
 
 (84) 
 
/ u 
 
 r n 
 
 
 r/6 
 
 2Q 
 
 
 
 
 
 
 
 
 50 
 
 40 
 
 1 (-\ 
 
 
 
 
 825 
 
 't 
 
 !H 2 : 
 
 2Gc 
 
 5 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Oi 
 
 Oil 
 
 Ga. 
 
 6as 
 
 / ^ 
 
 _ 
 
 _ _ _ 
 
 20 
 
 i n 
 
 
 
 
 
 
 
 
 
 
 *-n 
 
 > 
 
 V 
 
 ^o- 
 
 _ 
 
 
 Tl 
 
 - 1 
 
 , 
 
 . 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 J 
 
 0/7 
 
 IO 15 20 25 JO 35 4J 45 5(. 
 
 Rate - cc. per minute. - 
 
 fO 
 
 60 
 SO 
 AO 
 30 
 
 
 
 
 riG 
 
 29 
 
 
 
 
 
 
 
 
 
 
 
 82 
 
 5'C 
 
 21- 
 
 {I 
 
 S.<25 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 on 
 on 
 
 Gas 
 Gas 
 
 */ 
 
 
 
 
 /o 
 
 
 ^ 
 
 1 
 
 i . 
 
 ' 
 
 , 
 
 1 
 
 
 
 
 7 
 
 V- 
 
 
 *o 
 
 
 
 <> 
 
 
 
 
 
 5 JO 15 20 25 JO 35 4O 
 
 Oil Rate - cc. per minute 
 
 j>0 
 
 is funs 
 
 10 15 20 25 3O 35 4O 4 
 
 Oil Rate - cc. per minute - 
 
 (85) 
 
Fig. 30 shows clearly the effect of temperature on 
 tar formation, and also the effect of increase of oil 
 rate at constant temperature. The proportion of tar 
 increases with increasing oil rate and most markedly 
 so, at temperatures of 621 C. and 723 C. The 
 largest proportion of these tars at moderate to high 
 oil rates is undecomposed oil, as shown by distillation 
 and treatment with concentrated sulfuric acid. This 
 is also indicated by the fact that a temperature change 
 from 621 C. to 723 C. produces no great difference 
 in tar formation, and also by the fact of the very rapid 
 increase in the percentage of tar with increasing oil 
 rate. 
 
 At 825 C. the proportion of tar does not increase 
 greatly with increasing oil rate, indicating that these 
 tars are largely composed of synthetic products, 
 which is further substantiated by other physical 
 characteristics, such as distinct aromatic odor and 
 their reactions with concentrated sulfuric acid. 
 
 All the tars were strongly fluorescent. 
 
 FORMATION OF ILLUMINANTS, METHANE, AND ETHANE, 
 AND THE OBTAINING OF PARTICULAR END-PROD- 
 UCTS FROM A PARAFFIN HYDROCARBON OIL 
 
 In Figs. 31 to 35 the cc. of illuminants, methane and 
 ethane formed from i cc. of oil are shown plotted 
 against the oil rate at the temperatures and hydrogen- 
 gas concentration ratios indicated. It can be seen 
 from Fig. 31 that at 621 C., with the exception of the 
 illuminants in the case of the straight oil-gas runs, the 
 number of cc. of all these hydrocarbons formed from 
 i cc. of oil increases with decreasing oil rate. There 
 would be a limit to this, however, for, were the oil rate 
 made low enough, a very, extensive decomposition of 
 the hydrocarbons would set in. 
 
 Fig. 31 shows that the illuminants are the most 
 easily decomposed of the gaseous hydrocarbons. The 
 curve for the illuminants in the straight oil-gas runs 
 has a maximum due to the fact that though the 
 longer time of contact of the oil vapors with the 
 heated tube at low oil rates causes a more extensive 
 formation of ethylene and other illuminants, an oil 
 rate is reached where extensive decomposition of these 
 hydrocarbons takes place, which more than overcomes 
 the more rapid illuminants formation at low oil rates. 
 Condensation and hydrogenation, which are more 
 extensive at low oil rates, are important in this con- 
 nection also. 
 
 (86) 
 
__ JOO 
 
 6 
 
 o 250 
 
 VJ 
 
 V, 
 
 200 
 
 Q./OO 
 
 50 
 
 O// 
 
 -as 
 
 Gas 
 
 350 
 
 Oil Pate - cc. per minute 
 
 0/7 Rate - cc. per minute 
 
 4(J(J 
 T *\n 
 
 f 
 
 \ 
 
 
 
 12? 
 
 c. ^ 
 
 !/4s 
 
 16^ 
 
 
 
 
 250 
 200 
 
 ^0 
 
 
 \ 
 
 \ 
 
 
 nc 
 
 .JJ 
 
 
 
 
 
 
 
 V 
 
 f- 
 
 
 
 
 
 
 
 
 
 \ 
 
 V 
 
 
 
 
 Oil ( 
 
 Oil ( 
 
 ~as 1 
 
 
 
 
 ^*- 
 
 
 
 - 6 
 
 V 
 
 A 
 
 ~^\ 6 
 
 
 
 V^ 
 
 
 
 
 
 
 
 
 
 6 
 
 ^ 
 
 ST" 
 
 
 
 
 
 
 , 
 
 
 
 
 ^: 
 
 
 aa -**^ 
 
 
 
 L_ 
 
 
 A 
 
 "6*. 
 
 "^A 
 
 ^ 
 
 
 r,H, 
 
 
 "^ 
 
 
 
 
 
 
 
 
 
 
 
 
 Oil 
 
 /O /5 20 25 30 JJ 40 45 5C 
 
 Rate - cc. per minute - 
 
 (87) 
 
No maximum is observed in the illuminants curve 
 for the hydrogen runs, doubtless because the time of 
 contact of the oil vapors with the heated tube is less 
 at any particular oil rate than in the case of the oil- 
 gas runs. The curve for the illuminants in the hy- 
 drogen-oil-gas runs, for this reason also, is always be- 
 low that for the illuminants in the oil-gas runs, except 
 at low oil rates. 
 
 Less methane is formed per cc. of oil in the hy- 
 drogen-oil-gas runs than in the straight oil-gas runs 
 except at low oil rates. That the time of contact here 
 is sufficient so that extensive hydrogenation takes 
 place is clearly brought out from a consideration of 
 Fig. 22 in connection with Fig. 31. At all oil rates 
 the decreased time of contact of the gases with the 
 heated furnace tube, on account of the absorption of 
 hydrogen, results in a lower formation of methane. 
 
 The relationships for ethane are much the same as 
 for methane, and for the same reasons. 
 
 Figs. 32 and 33 show the relationships between the 
 hydrocarbons at 723 C. at hydrogen concentration 
 ratios of i/7 2 : 2 Gas and 2# 2 : i Gas, respectively. 
 More of each of the components is formed per cc. of 
 oil at 723 C. than at 621 C. The maximum in the 
 illuminants curve falls at a higher oil rate than at 
 621 C. as would be expected, since the higher tem- 
 perature would cause a more rapid decomposition, 
 condensation, and hydrogenation of the illuminants to 
 take place. At 723 C. the maximum on the illu- 
 minants curve for the hydrogen runs can be seen 
 clearly. It is interesting to note that this falls to the 
 left of the maximum on the curve for the oil-gas runs. 
 The decreased time of heating on account of the 
 hydrogenation admixture is responsible for this. 
 
 The effect on the illuminants of increasing the con- 
 centration of hydrogen is clearly brought out in Figs. 
 32 and 33. The maximum on the curve for the 
 illuminants in the hydrogen-oil-gas runs at the higher 
 hydrogen concentration falls at a slightly lower oil 
 rate than the maximum on the curve for the lower 
 hydrogen concentration. The divergence between the 
 illuminants curves for the oil-gas runs and the hy- 
 drogen-oil-gas runs is greater both at low and high oil 
 rates at the higher hydrogen concentration than at the 
 lower hydrogen concentration on the time of contact 
 of the gases with the heated tube surfaces. 
 
 The relationships in the case of the methane and 
 ethane are exceedingly interesting. The formation 
 
 (88) 
 
60D 
 
 550 
 
 TO 15 20 <eJ JO JO 40 ^7 TO 
 
 // P^?(? - tc. per minute - 
 
 JPV 
 
 J 450 
 
 ."^ 400 
 
 
 <o 35"0 
 *0 
 
 V.. JOO 
 
 sc 
 
 2JO 
 
 t-K 
 
 s: 
 
 <0 200 
 
 c 
 
 O 
 
 ^ /5 
 
 t: 
 o 
 
 ^ 700 
 
 VJ 
 VO SO 
 
 V 
 
 
 
 823 
 
 c 
 
 ZH 2 >. 
 
 / COL 
 
 * 
 
 
 
 1 
 
 \ 
 \ 
 
 
 rk 
 
 ;. jj 
 
 
 
 
 
 
 
 \ 
 
 v d 
 
 ^ ,- 
 
 
 t 
 
 
 
 
 
 
 
 N 
 
 A 
 
 
 
 
 
 Oi 
 
 '/ 6a 
 
 f Ga 
 
 5 
 5 +1- 
 
 6 - 
 
 o 
 
 6 
 
 
 A 
 
 <\. 
 
 ^ 
 
 
 
 
 
 
 
 / 
 / 
 
 / 
 
 \\v* 
 
 \ 
 
 r^I 
 ^x 
 
 ^ 
 
 
 
 
 
 
 2 
 
 2 
 
 
 > 
 
 ^ 
 
 ^^ 
 
 ^. 
 
 ^ 
 
 
 
 
 / 
 
 
 
 
 
 ^- 
 
 ^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 4?,! 
 
 ^5:" 
 
 -oo-l 
 
 
 
 ~t> 
 
 C&l. 
 
 
 
 
 
 
 Oil Rate - cc. per minute 
 
 (89) 
 
of methane is less in the hydrogen-oil-gas runs than 
 in the oil-gas runs at moderate to high oil rates, due 
 to the decreased time of contact of the gases witty the 
 heated surfaces, and this effect is more pronounced 
 at the higher hydrogen concentrations, as would be 
 expected. As the oil rate decreases the hydrogenation 
 effect becomes important and the proportion of 
 methane formed from i cc. of oil is greatest in the 
 case of the hydrogen-oil-gas runs. The curve for 
 methane in the hydrogen-oil-gas runs crosses the 
 methane curve for the oil-gas runs. This crossing 
 is at a higher oil rate with the higher hydrogen con- 
 centration, showing clearly the effect of the increase 
 of concentration of hydrogen on the hydrogenation 
 reactions. The formation of methane is slightly 
 greater in the hydrogen-oil-gas runs than in the 
 straight oil-gas runs when the hydrogen concentration 
 ratio is iHz : 2 Gas. This difference is slightly 
 greater at low oil rates than at high oil rates. 
 
 When the hydrogen concentration ratio is 2# 2 ' i 
 Gas the ethane formation per cc. of oil is less in the 
 hydrogen-oil-gas runs than in the oil-gas runs at 
 high oil rates. This is the effect of the decreased time 
 of contact due to the admixture of a larger volume of 
 hydrogen. But at low oil rates the formation of 
 ethane is much greater in the hydrogen-oil-gas runs 
 as can be seen in Fig. 33. The ethane curve has a 
 maximum, too, which is interesting because it shows 
 that at low oil rates the reactions of the hydrocarbon 
 ethane itself have an important part to play. . 
 
 Figs. 34 and 35 show the relationships between these 
 hydrocarbons at 825 C. A much less pronounced 
 decrease in the formation of illuminants with in- 
 creasing oil rate is due to the fact that the temperature 
 of 825 C. is sufficient to promote actively the forma- 
 tion of illuminants. The maxima on the illuminants 
 curves for the oil-gas runs fall at the higher oil rates, 
 as would be expected when the higher temperature is 
 taken into consideration. The illuminants curves for 
 the hydrogen-oil-gas runs are entirely above the 
 illuminants curves for the oil-g'as runs. This is again 
 a result of the decreased time of heating when hy- 
 drogen is admixed. The effect is most marked when 
 the higher concentration of hydrogen is used. 
 
 The curves for methane in the hydrogen-oil-gas runs 
 fall above the curves for methane in the oil-gas runs 
 at all oil rates at 825 C. This is due to two effects: 
 first, the less extensive decomposition of the methane 
 
 (90) 
 
into carbon and hydrogen due to the decreased time of 
 heating in the hydrogen-oil-gas runs; second,, the 
 increased rate of hydrogenation reactions such as 
 C 2 H 4 + 2H 2 "*"V 2CH 4 . The effect of hydrogen in 
 
 JOO 
 450 
 400 
 * JJO 
 
 zoo 
 
 zso 
 
 200 
 
 ^ 
 o 
 Q. ISO 
 
 o 
 
 Vj /OO 
 
 so 
 
 600 
 
 TOO 
 
 600 
 
 nc. J6 
 
 OH (>45 
 
 Runs 
 
 Oil Rate - cc. per minute 
 
 ^soo 
 
 + 400 
 
 O JOO 
 
 200 
 
 /OO 
 
 \J 
 
 Vo 
 
 Oil Rate - cc per minute. - - 
 reversing the reaction CH 4 ~*~^ C + 2H 2 is probably 
 not important, as has been brought out in the dis- 
 cussion of the methane equilibrium and the reactions 
 of methane in the first part of this paper. 
 
 (91) 
 
The effect of the greater concentration of hydrogen 
 on methane production can be seen clearly by com- 
 paring Figs. 34 and 35. The divergence between the 
 hydrogen-oil-gas and the oil-gas methane curves is 
 greatest when the hydrogen concentration ratio is 
 2 # 2 : i Gas, and this is practically true at low oil 
 rates where hydrogenation reactions are most im- 
 portant. 
 
 More ethane is formed when hydrogen is mixed with 
 the vapors of the oil than when it is not added. This 
 is doubtless due to the combined influence of the 
 hydrogen in diminishing the decomposition of the 
 ethane and to its effect in hydrogenating the olefins. 
 These effects are particularly marked when the con- 
 centration ratio is 2H 2 : i Gas. 
 
 It may have been noticed that the curves for the 
 oil-gas runs made at the same temperature do not 
 coincide exactly, since the carbon tube used car- 
 bonizes somewhat and becomes of smaller internal 
 diameter, thus decreasing the time of contact of the 
 gas with the tube and consequently altering the com- 
 position somewhat. 
 
 The effect of temperature on the hydrocarbon 
 products of decomposition of an oil can be seen very 
 clearly from Fig. 36. Within the temperature range 
 studied, the cc. of illuminants per cc. of oil increases 
 with temperature, with one exception. At low oil 
 rates there are more illuminants formed at 723 C. 
 than at 825 C. The effect of the higher temperature 
 in increasing the speed of the reactions which de- 
 compose ethylene more than overcomes the effect of 
 the higher temperatures in promoting the decom- 
 position of the long chain olefins to ethylene and 
 propylene. 
 
 It is interesting to note the position of the maxima 
 of the curves for the illuminants at the several tem- 
 peratures. These maxima indicate where the balance 
 between the reactions of formation and the reactions 
 of decomposition falls. 
 
 Within the temperature range studied, the forma- 
 tion of methane is greater the higher the temperature. 
 At low oil rates the difference in the methane pro- 
 duced by a ioo-temperature rise is greater in the 
 range from 621 to 723 C. than from 723 to 825 C., 
 since at low oil rates 723 C. is a sufficiently "high 
 temperature to break down the original oil extensively. 
 The methane increase between 723 and 825 C. is 
 largely due to the decomposition and hydrogenation of 
 
 (92) 
 
olefins C 2 H 4 > C + CH 4 and C 2 H 4 + 2 H 2 J 
 2CH 4 , as can be seen from a consideration of the 
 illuminants curve for 825.C. 
 
 At high oil rates a temperature of 825 C. is neces- 
 sary to form methane largely, as can be seen from the 
 position of the curve for methane at 621, 723 and 
 825 C. 
 
 The formation of ethane per cc. of oil is- not large at 
 any temperature studied, as shown in Fig. 36. The 
 primary decomposition of the oil therefore involves 
 chiefly a splitting off of methane rather than ethane 
 or higher paraffin. The decomposition and dissocia- 
 tion of ethane are clearly shown by the falling off of 
 the ethane curve as the oil rate decreases at a tem- 
 perature of 825 C. On the other hand, a tem- 
 perature of 825 C. is necessary to cause an extensive 
 formation of methane and ethane per cc. of oil at high 
 ^oil rates. 
 
 TOTAL HYDROCARBONS OBTAINABLE FROM THE OIL 
 
 The greater the proportion of the carbon of the oil 
 which can be obtained in gaseous form the better the 
 utilization of the oil for gas-making purposes. 
 
 Fig. 37 shows the Total cc. of Illuminants + methane 
 + ethane obtainable from i cc. of oil under the varying 
 conditions. At 62iC. more hydrocarbons are ob- 
 tained per cc. of oil in the oil-gas runs than in the 
 hydrogen-oil-gas runs, except at low oil rates where 
 hydrogenation reactions become important. This is 
 due to the lower time of contact of the oil vapors with 
 the furnace tube in the case of the hydrogen-oil-gas 
 runs. The same relations hold at 723 C. except 
 that the hydrogen-oil-gas curves cross the oil-gas 
 curve at a higher oil rate because of the greater effect 
 of the higher temperature in hastening the hydrogena- 
 tion reactions. The effect of the higher concentration 
 of hydrogen is clearly shown. At 825 C. the hy- 
 drocarbons formed per cc. of oil are of greater volume 
 in the hydrogen-oil-gas runs at all oil rates studied. 
 The higher temperature promotes hydrogenation re- 
 actions at all oil rates. The effect of the greater 
 concentration of hydrogen can be seen. 
 
 It is interesting to note that the slope of the curve 
 for the hydrogen-oil-gas runs at 723 C. is much 
 steeper than the slope of these curves at 825 C. at 
 low oil rates, doubtless because at 825 C., and low 
 oil rates, dehydrogenation reactions and reactions of 
 decomposition of the hydrocarbons become of im- 
 portance. 
 
 (93) 
 
SUMMARY 
 
 i A critical review of the most important work on 
 hydrocarbon decomposition and the influence of 
 hydrogen on the reactions involved has been given. 
 This has concerned itself with: first, the hydrocarbons 
 of high molecular weight; second, the products of the 
 primary decomposition; and third, the reactions of the 
 simpler hydrocarbons. Summaries have been in- 
 cluded which state concisely the probable course of the 
 reactions of dissociation, decomposition, and condensa- 
 tion involved. 
 
 2 The subject of the mechanism of heat transfer 
 in gas machines has been discussed. 
 
 3 Difficulties in the measurement of the true tem- 
 perature of a gas have been pointed out. 
 
 4 In the experimental work a paraffin oil was 
 thermally decomposed alone and in hydrogen at 
 temperatures of 621, 723 and 825 C. Concentra- 
 tions of hydrogen approximating i// 2 : 2 Oil Gas and 
 2 H 2 : i Oil Gas were those studied. It has been 
 shown what results may be expected in the decom- 
 position of a hydrocarbon oil when temperature, rate 
 of oil feed, and concentration of admixed hydrogen 
 are carefully controlled. 
 
 5 The relationship between the rate of oil feed and 
 the rate of gas generation has been brought out. 
 
 6 The proportions of illuminants, saturated hy- 
 drocarbons, and hydrogen resulting at varying rates 
 of oil feed, and at temperatures of 621, 723 and 
 825 C. have been shown graphically and discussed. 
 
 7 The effects of hydrogen on the reactions which 
 give rise to saturated hydrocarbons and illuminants 
 have been shown graphically and discussed at some 
 length. Besides its effect in hydrogenating olefins 
 and other hydrocarbons, the hydrogen, since its addi- 
 tion causes an increase in the total volume of the gas 
 passing through the heated zone of the furnace in a 
 given time, decreases the time of contact of the gases 
 with the heated walls of the resistor tube. The 
 effects of this are discussed in connection with the 
 curves showing the relationships between the com- 
 ponents of the gas when the oil is cracked in hydrogen. 
 
 8 The mean molecular weight of the olefins in a 
 series of gases made at 825 C. has been determined, 
 and also the proportion of aromatic hydrocarbons 
 in these gases. 
 
 (94) 
 
9 The formation of tar was studied at the various 
 oil rates, temperatures and concentrations of hy- 
 drogen. 
 
 10 Curves showing the number of cc. of illuminants, 
 ethane, and methane obtainable from i cc. of oil have 
 been shown. 
 
 ii In general the manner of decomposition of a 
 paraffin hydrocarbon oil has been mapped out over a 
 considerable range of temperature, rate of oil feed, 
 and concentration of hydrogen. 
 
 12 The results recorded in this paper may serve 
 as a guide to the obtaining of more desirable results in 
 commercial operations involving the decomposition of 
 oil for gas-making purposes. 
 
 CONCLUSIONS 
 
 In addition to showing the proportions of products 
 which are obtainable under a variety of conditions, 
 which relationships have been fully set forth in the 
 figures shown and which it is impossible to briefly 
 summarize, it has been concluded as a result of this in- 
 vestigation: 
 
 I That the importance of radiation insofar as it is 
 concerned in the furnishing of the energy for the pro- 
 duction of hydrocarbon reactions has been over- 
 estimated. 
 
 II That effects often ascribed to catalysis are in 
 reality due to effective heat transfer by conduction 
 arid convection from the large heated surfaces exposed 
 to the gases. 
 
 Ill That the equilibrium condition is not attained 
 in a hydrocarbon system when an oil is decomposed 
 by heat under conditions analogous to those of car- 
 bureted water-gas manufacture. 
 
 IV That the course of the changes involved in the 
 breaking down of a hydrocarbon oil may be roughly 
 traced. 
 
 V That hydrogen is produced from an oil even when 
 the cracking takes place in hydrogen. 
 
 VI That considerable absorptions of hydrogen take 
 place when an oil is cracked in an atmosphere of 
 hydrogen, and this absorption is greater the higher 
 the concentration of hydrogen, the higher the tem- 
 perature (within the range studied), and the lower the 
 oil rate. 
 
 VII That propylene and higher olefins constitute 
 approximately one-third by volume of the illuminants 
 of the gas. 
 
 (95) 
 
VIII That the proportion of tar increases with de- 
 crease in temperature, and with increasing oil rate, 
 particularly at the lower temperatures. 
 
 IX That no marked and consistent difference in the 
 amount of tar formed when an oil is decomposed alone 
 or in hydrogen at temperatures of 723 C. or below is 
 noticeable. At 825 C. less tar is formed when the 
 oil is cracked in hydrogen. The tars formed below 
 723 C. are in large part unchanged or partly changed 
 oil, whereas those tars formed above 800 C. are 
 essentially composed of synthetic products. 
 
 X That the reactions which result in decreasing 
 the proportion of illuminants are the most rapid. 
 
 XI That the presence of hydrogen during the de- 
 composition of an oil has the effect of increasing 
 largely the proportion of the carbon of the oil ap- 
 pearing as hydrocarbons in the gas. 
 
 XII That within the temperature range studied 
 the volume of illuminants produced per volume of oil 
 increases with the temperature with one slight ex- 
 ception. The formation of methane is greater the 
 higher the temperature. The formation of ethane is 
 not large at any temperature and therefore the primary 
 decomposition of an oil involves chiefly a splitting off 
 of methane rather than ethane or higher homologs. 
 
 XIII That a temperature of 823 C. is desirable 
 in decomposing an oil provided that too great op- 
 portunity for extensive secondary and tertiary change 
 is not given. 
 
 XIV That with correct design of apparatus, and 
 proper adjustment of temperature, rate of oil feed, 
 and concentration of hydrogen it is possible to obtain 
 gases of widely varying compositions. 
 
 DEPARTMENT OF CHEMICAL ENGINEERING 
 COLUMBIA UNIVERSITY, NEW YORK CITY 
 
 (96) 
 
VITA 
 
 Eugene Hendricks Leslie was born at Ottawa, 
 Illinois, August 28, 1892. 
 
 He was graduated from the Ottawa High School in 
 June, 1909, and in the Fall of that year he entered the 
 University of Illinois. From this institution he was 
 graduated in June of the year 191 J with the degree of 
 Bachelor of Science in Chemical Engineering. 
 
 In September, 1913, he entered Columbia University 
 where he has been in continuous residence to this date. 
 During this time he has been engaged in graduate 
 study and research work and has held an assistantship 
 in the Department of Chemistry. He has also been an 
 officer of instruction in the intervening summer sessions 
 at Columbia, and in Extension Teaching. 
 
UNIVEESITY OF CALIFOENIA LIBEAEY 
 BEEKELEY 
 
 THIS BOOK IS DUE ON THE LAST DATE 
 STAMPED BELOW 
 
 Books not returned on time are subject to a fine of 
 50c per volume after the third day overdue, increasing 
 to $1.00 per volume after the sixth day. Books not in 
 demand may be renewed if application is made before 
 expiration of loan period. 
 
 ^26 t910 
 APK 8 ls24 
 
 50m-7,'16 
 
YC '3370 
 
 UNIVERSITY OF CALIFORNIA LIBRARY