Nov., 1918
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
esting to note in this cqnnection t h a t Wood’ has used a system of cold water percolation in order t o concentrate acid in the solution from niter cake, and t h a t this treatment has been recommended by Prideaux.2 Unfortunately, a t the present time the d a t a are not available for temperatures between the two given. I t may be possible t h a t it is unnecessary t o use a temperature as low as o o in order t o obtain a satisfactory separation. Solubility determinations for this system at 12’ are being carried out in this laboratory by Professor Foote, who suggested this paper t o the writer. I n the preceding paper reference has been made t o the work of Matignon and Meyer on the solubility relations in the system N a 2 S 0 4 - ( N H ~ ) 2 S O ~ - H ~ 0T.h e writer proposes t o treat this system as he has treated the system discussed in the present paper. SUMMARY
General equations have been developed for the system NazSO4-KzSO4-HtO, a t 2j0, by means of which we can calculate how much of a n y one solid phase will separate from a solution if we know the composition of the original solute and the acid concentration of the solution after crystallization. General equations have also been developed for this system a t 2 5 ’ by means of which we may calculate the weight of water in the solution after crystallization, or the weight of water t o be added t o the solid niter cake in order to leave a calculated weight of one of the solid phases. A very simple type of calculation has been applied t o niter cake of several compositions, by which the maximum amount of each solid phase which can be removed from solution a t 2 5 ’ and a t o o has been calculated. Leaching or crystallizing processes have been suggested by which sulfuric acid may be concentrated in the solution a n d sodium sulfate in the solid, a t the two temperatures mentioned. It was found t h a t this separation can be done much more efficiently a t the lower temperature. SHEFFIELD CHEMICAL LABORATORY YALEUNIVERSITY A’EW HAVEN,CONNECTICUT
THE FORMATION OF AROMATIC HYDROCARBONS FROM NATURAL GAS CONDENSATE3
matic hydrocarbons by the thermal decomposition of straight-chain hydrocarbons of low molecular weight. Previous t o this, Bone and Coward’ had passed ethane, ethylene, and acetylene through porcelain tubes a t various temperatures from 500’ t o I O O O O C. and had noted t h a t the decomposition of ethylene gave a black, viscous tar. The quantity of tar was too minute t o admit of analysis b u t they mentioned the fact that a few crystals of naphthalene were noticed also. They hold aromatic formation t o be produced by the breaking down of ethylene t o acetylene from which the aromatic hydrocarbons are produced by polymerization. Pring and Fairlie2 found t h a t acetylene a t high temperatures and in the presence of hydrogen produces methane for the most part, although some ethane was formed also. When ethylene and hydrogen were heated together no acetylene was produced even a t very high temperatures. Methane, however, was produced in large quantities. Jones3 studied the formation of aromatic compounds in coal t a r and is of the op’inion t h a t acetylene plays an unimportant part in the reaction, inclining more t o the belief t h a t the ring bodies are formed directly from olefines with the splitting out of hydrogen. Previous work in this laboratory pointed t o conclusions which were similar t o Jones’, and in a n effort to get a further insight into the reaction the following work was undertaken: It was decided t o divide the work into several parts and investigate each as fully as time allowed, for i t was quite evident from the beginning t h a t any one of the separate fields was capable of large expansion with possible loss of the original aim. The divisions of the work are as follows: ( I ) The effect of catalyzers on the decomposition of straightchain hydrocarbons of low molecular weight. (2) The influence of temperature and of pressure on the production of aromatic hydrocarbons. (3) The formulation of the reaction, Straight-chain hydrocarbons + Aromatic hydrocarbons
By J. G . DAVIDSON
Received May 23, 1918 INTRODUCTION
I n several papers which have appeared recently Zanetti4 has shown t h a t it is possible to produce aro1 J . Sac. Chem. I n d . , 36 (1917), 1216A. a Ibid., 36 (1917), 1216B. 8 This paper is condensed from a dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Faculty of Pure Science of Columbia University. The work was begun under the direction of Dr. J. E. Zenetti and is, in part, a continuation of his work. After the summer of 1917, when Dr. Zanetti entered the Chemical Warfare Service, the work was carried on more or less independently although I am glad t o thank Dr. Nelson, Dr. Freas, and Dr. Fisher for their many invaluable suggestions, and without whose help the work could not have been finished. 4 “The Thermal Decomposition of the Propane-Butane Fraction from Natural Cas Condensate,” THISJOURNAL,8 (1916), 674; “The Thermal Decomposition of the Ethane-Propane Fraction from Natural Gas Condensate,” Ibid., 8 (1916), 777; “Aromatic Hydrocarbons from the Thermal Decomposition of Natural Gas Condensate,” Ibid., 9 (1917), 474.
901
EXPERIMENTAL
material used was the e t hane-propane fraction of natural gas condensate, supplied in steel tanks under high pressure. The tanks are built on the siphon system, a pipe reaching almost t o the bottom, so the composition of the delivered gas remains almost constant. Analysis of the gas showed it to be composed almost entirely of the two hydrocarbons, although some butane, and possibly some pentane, was also present. No other gases were present in the original material, although tests were made for oxygen, carbon dioxide, olefines, and hydrogen. MATERIAL-The
1
“Thermal Decomposition of Hydrocarbons.” J . Chem. SOC., 9s
11908). 1197.. I ,
“Synthesis of Hydrocarbons at High Temperatures,” Ibid., 99 (1911), 1796. a “Aromatic Formation,” J . Sac. Chem. Ind., 36 (1917), 3 .
902
T H E J O U R N A L O F I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y Vol.
li
IO,
No.
IT
Trans f o rmer
AspiraTor
Gas
Tar
5 a rnpier
6eporaTo Y
InducTion
co
I
I
HeaTing
Appa
roTus
I
Drytn 4
ReTer
Towers
I
FIG. 1
APPARATus---The arrangement of the apparatus will be understood from Fig. I. The gas was led through the referee meter, capable of being read t o one-thousandth part of a cubic foot, then through the calcium chloride towers t o the cracking chamber. The cracking chamber was a silica tube one inch in diameter about 2 f t . long and could be heated by the resistance furnace shown. The temperature could be controlled within a few degrees by the adjustable rheostat, while its actual value was read by means of the pyrometer. After the gas had been cracked i t was quickly cooled in the metal condenser and passed into the precipitator. (At first a small copper plate and a fine iron wire were kept charged a t opposite sides of the bottle. This did not work satisfactorily and the inlet tube was then surrounded with wire gauze as shown in the figure. This worked well for a time, as the gas had t o pass through the charged wire meshes t o escape, and deposition therefore was easy. After some time the meshes of the gauze became stopped up and required frequent renewal. The form of precipitator was then changed t o t h a t shown in Fig. 2, which was very satisfactory.) From the tar precipitator the gas was allowed to escape, as shown by the light arrow, or by-passed in the direction of the heavy arrow when a sample was being collected, This arrangement was necessary t o prevent a change in the rate of flow through the cracking chamber, which was caused when trying t o take a sample direct from the tar precipitator. After a sample of sufficient volume had collected in the gas sampler the gas was allowed t o pass in the same direction for a half hour longer in order t h a t all parts of the apparatus might be in equilibrium. Failure t o do this produced results t h a t could not be checked. After a half hour had elapsed the stopcocks were turned so the gas followed the light arrow
again and the collected sample of gas WBS forced out into a gas collecting bottle by way of the dotted arrow. The apparatus for the analysis of the cracked gas was a modification of Burrell’s gas apparatus.’ Nothing new is claimed for this apparatus except its greater accessibility and ease of manipulation. I t is shown in Fig. 3. Babb pipettes with an extra stopcock blown in the bend, as shown, were substituted for the Ostwald pipette in Burrell’s apparatus. This extra stopcock facilitates refilling and cleaning the pipettes without the necessity of disconnecting from the main part of the apparatus. Beyond this the form of t h e Babb pipette lends itself admirably t o rapid and complete absorption. The slow combustion pipette (B) was made of transparent quartz rather than glass in order t o reduce breakage. When using t h e ordinary glass pipette for slow combustions the oxygen would sometimes catch fire and burn a t the point where the capillary opens out into the pipette. This would always result in a fracture at t h a t point; furthermore, it was necessary after a combustion t o wait almost 5 min. before the glass was cool enough t o allow the mercury, which always had some drops of water on t h e surface, t o be raised. After considerable trouble from both of these causes i t was decided t o have t h e pipette made of silica. The pipette as described gives the best of satisfaction. Copper oxide was used t o determine the hydrogen. The copper oxide was also enclosed in a tube of transparent silica, in preference t o glass, which will break if drops of condensed water are drawn into the hot part of the tube. Use of a silica tube was suggested for this purpose b y Suydamj2 but i t was found best 1
“New Forms of Gas Analysis Apparatus,” THISJOURNAL, 4 (1912).
296. 2 “A New Model of the Burrell and Oberfell Apparatus for the Analysis of Illuminating Gas,” THIS JOURNAL, 9 (1917), 972.
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY t o make t h e tube horizontal rather than vertical as he advises. This position prevents stoppages, which happen after a time with t h e vertical form. P R O C E D U R E - The gas was led through t h e cracking chamber at the rate of 0 . .j cu. ft.per hr., the rate of flow being frequently checked with a stopwatch. The temperature of the silica tube was rapidly raised t o the desired point and kept there b y adjusting the rheostat. When the temperature had become constant and the gas had passed through the tube and tar precipitator for sufficient time t o sweep out all traces of gas from a preceding run, a sample was tal-\en as described before, and L A the temperature of the FIG 2 tube raised for the next determination. While equilibrium was being reached again the first sample of gas was analyzed. When catalyzers were used in the Eorm of foil or gauze, pieces of uniform size were cut, rolled up to fit t h e tube snugly, and pushed in so the entire heated zone was filled with catalyzer. TVher small pieces of material had t o be used for catalyzer the cracking tube was packed with loose material, which was held in place with a plug of copper gauze, preliminary work having shown t h a t copper has no decided effect as catalyzer. G A S ANALYSIS-All capillary errors and the larger error due t o gas left in the copper oxide tube have been carefully determined and allowed for in t h e reports of analysis. As preliminary work showed no other gases t o be present, unsaturated hydrocarbons, hydrogen, and saturated hydrocarbons were the only ones determined. Pipettes I and 2 (Fig. 3) contained 30 per cent potassium hydroxide. Pipette 3 contained saturated bromine water. The unsaturated hydrocarbons were determined by absorption in Pipette 3 , one passage of the gas being sufficient if the olefine content of the gas was below 17 per cent. When the unsaturated hydrocarbons existed in greater amounts i t was necessary t o pass t h e gas through this pipette twice. Complete absorption is definitely shown by the presence of bromine vapor above the aqueous layer. Before measuring
I I
, i
903
the amount of absorption i t was necessary t o pass the gas through Pipettes I and 2 t o remove all traces of bromine vapor. Otherwise, the bromine causes a heavy sludge of mercuric bromide t o form in the measuring pipette. This sludge clings t o t h e sides of the tube and makes further work impossible. I t was necessary t o use bromine water for the absorption of olefines, because sulfuric acid, either fuming or doncentrated, was found t o absorb some of t h e saturated hydrocarbons left in the gas which had been cracked a t low temperatures. Above 750' C. t h e residual saturated hydrocarbons consisted solely of methane, so either sulfuric acid or bromine water could be used for gas cracked above this temperature. After the contraction due t o the absorption of t h e unsaturated hydrocarbons was measured, t h e gas was slowly passed through the copper oxide tube t o t h e slow combustion pipette and back, until no further contraction occurred. The shrinkage was calculated as hydrogen. The temperature of the copper oxide tube was kept a t 310' C. by means of a nichrome resistance heater controlled by a rheostat. This temperature was found to give rapid absorption of hydrogen, without noticeably attacking t h e saturated hydrocarbons still present. The residue was then passed into the silica slow-combustion pipette, the wire brought t o low whiteness, and the oxygen passed in. I n samples taken above 600° C. the residue consisted solely of methane and ethane, and above 7 j o " C. only methane survived.
-
I
FIG.3
T H E J O U R N A L O F I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y Vol.
904
tar. 8 7 8 S
coke. 10 11 ia
No.
II
Remarks for nickel apply herefa). Heavy fog above 750° C. Considerable tar deposited. No free carbon. The chromium was seemingly unaffected. Yield of tar small. Some free carbon in cracking tube. The lumps of manganese which were made by the Goldschmidt process crumbled into pieces. N o visible fog at any time. No t a r deposited. At the end of run t h e pieces of calcium carbide were found cemented together b y pieces of hard Heavy fog above 750' C. Good tar deposit. A few hard lumps of coke were found adhering t o the wire after two or three runs. Heavy fog but not much tar. Large amount of carbon in the form of soft lampblack in the cracking tube. ~~~~y fog and good t a r yield. Platinum was tarnished a t end of run but was easily cleaned, No free carbon deposited
This was checked up so many times t h a t finally this last determination was not carried out above 750' C. The composition of the olefines will be taken up in a later part of this paper. I n no case where tar was formed during a run did i t deposit in noticeable amounts before a temperature of 700' to 750' C. was reached. Slight bluish fogs were sometimes observed earlier, but in no case was tar recovered from them. The tabulated results are shown above. See also Figs. 4 t o 7 inclusive. DISCUSSION O F R E S U L T S
I n accordance with former work i t is observed t h a t a tar is produced by heating straight-chain hydrocarbons of low molecular weight to a temperature of 700' C. and above. This tar has been shown1 t o consist of a mixture of simple aromatic compounds, such as benzene, with more complex ones, such as phenanthrene. With the exception of nickel, iron, and cobalt, metals do not seem t o have any great catalytic effect upon the reaction, nor does variation in the surface exposed seem t o influence the reaction. It is t o be noted in this connection t h a t Bone and Coward2 found t h a t the decomposition of methane was a surface effect, but the decomposition of ethane and of ethylene was not. The metals nickel, iron, and cobalt act as anticatalysts, so far as the production of tar is concerned. Their presence causes the main reaction t o be of the order Hydrogen Straight-chain hydrocarbons ---t Carbon
+
The above reaction always takes place t o a certain extent in the thermal decomposition of the hydr3carbons, but in most of the cases examined i t is much less important than the reaction 1
IO,
Zanetti, LOC.cit.
a LOG. C i f .
Straight-chain hydrocarbons Aromatic hydrocarbons
--f
+ Hydrogen + Methane
It may also be pointed out here t h a t many of t h e decompositions observed when a substance was "passed through a red-hot iron tube" may have been due t o the specific catalytic action of the iron pipe, particularly in those cases where large amounts of free carbon were produced. The carbon deposited in these experiments varied from hard, coke-like, and closely adherent material to soft,velvety lampblack (this latter when the three anticatalysts were used) t h a t did not adhere t o t h e tube. Bone and Coward' hold t h a t the decomposition of methane gives the hard variety, while ethylene gives the soft material. Our work seems t o confirm this, for in those cases where soft carbon was deposited the amount of methane in the gas was small. T H E I N B L U E N C E O F T E M P E R A T U R E AND O F P R E S S U R E O N T H E PRODUCTION
OF AROMATIC
HYDRO-
CARBONS
In this part of the experimental work it was first. attempted t o ascertain a t what temperature t h e maximum yield of t a r was obtained, and definitely establish whether any of the catalysts previously used promoted the formation of tar. The same apparatus was used, with the exception t h a t a n improved form of precipitator was used. This is shown in Fig. 2 and is simply a 2 in. tube drawn down on one end t o a point which is fitted with a rubber tube and pinchcock, A cylinder of gauze which fits the tube tight was one electrode, the other was simply a fine iron wire insulated by a glass tube. This insulation was necessary t o prevent sparking from the center electrode t o the gauze. The same z in. spark coil was used as a source of current and t h e whole precipitator was immersed in a freezing mix1 LOC.
cit.
Nov., 1918
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY
90.5
Tempera f ure PIG.
4
ture. This precipitator worked so well t h a t i t was possible t o pass the gas through the cracking tube a t a rate of 3 t o 4 cu. f t . per hr. and still have complete precipitation of the tar. PROCEDURE-The furnace was brought up to the desired temperature and the gas admitted a t the rate of approximately I cu. f t . per hr. The precipitator was started and the apparatus left t o itself. At intervals, the temperature and the rate of flow were checked up, and while it was found possible t o regulate the temperature within a few degrees, i t was not always possible t o regulate the rate of flow closer than 5 or I O per cent of the rate desired. This was chiefly due to the gradual choking of the delivery pipe with naphthalene. At the end of a run the time was noted, the t a r run out of the precipitator, its volume measured, and its specific gravity taken a t zoo C. No analysis of the t a r was attempted. The precipitator and tube were cleaned and a new run made the next day. A few of the catalyzers which seemed most promising were introduced into the cracking tube, as before, and t h e yields determined after the most favorable temperature for the formation of tar had been worked out. No records were made below 750' C., as the yields of t a r were negligible. The results are shown in both tabular and graphic form (Fig. 8). VOl. of VOl. of Gas Length Tar Rate Yield Used of Run Obtained Ou. Ft. per Cu. Ft. Hours Cc. Per Hr. Cu. Ft. Sp. Or. 8.2 7.7 16.3 1.06 2.24 0.9819 7.6 8.1 25.5 0.94 3.33 1.0040 7.9 8.2 42.0 0.96 5.30 1.0600 900(a) 7.3 7.2 1.2 1.01 0.16 Copper as Catalyzer 850 8.6 7.2 53.4 1.18 6.2 0,999 Chromium as Catalyzer 850 5.4 41.0 1.32 7.2 1.002 5.7 Silicon as Catalyzer 850 6.4 6.1 49.0 1.04 1.016 7.7 Tungsten'as Catalyzer 8.50 7.0 45.4 1.23 8.6 1.038 5.3 (a) In the run at 900' C. a large amount of naphthalene was produced. It deposited on the sides of the precipitator and on the electrodes in such a way that its volume could not be measured. Temp. Deg. C. 750 800 850
80
-Unsofuroted 70
1
I --
Hudrooan
r
-T
b
.,
..
It is evident from the above t h a t a temperature of 850' C., or thereabouts, is the best for t a r formation. The increasing specific gravity of the t a r with higher temperatures shows a decreasing content of the lighter aromatic hvdrocarbons such as benzene, and also in-
Tern p e r a t u r e
FIG.5
The construction of the furnace is simple, but a few words of explanation may not be out of place. The tube itself was extra heavy copper pipe, I in. in diameter and 3 f t . long. I t was found impossible to
FIG.6
FIG. 7
cut good threads on the ends, so 6-in. sections of heavy brass pipe were brazed on and silver soldered t o make a tight joint. These brass ends were threaded and fitted with hydraulic iron fittings and valves. All joints were made tight with litharge-glycerin cement. Twenty-inch strips of l / 4 in. asbestos board were first wired around the pipe, and the nichrome wire wound in two sections on the strips. The ends of the wires were brought out t o in. asbestos board heads. The furnace would heat up t o 900' C. in 2 0 min. Beyond t h a t temperature we did not think i t advisable t o venture. PROCEDURE-The procedure was about the same as before. The furnace was brought up to temperature, the rate of flow adjusted, and the pressure regulated. Vacuum was obtained by a motor-driven Nelson pump, and arrangement was made for altering the strength of vacuum by introducing a valve t h a t could be opened t o the atmosphere.
I
I
I 750
1
1 Showing yle/d of Tar with \ 800
Temperature FIG. 8
850
900
Pressure was obtained by connecting the furnace directly t o the tank of compressed gas through a pressure reduction valve t h a t could be adjusted within a pound or two. An escape valve in the far end of the furnace provided another adjustment. The same runs were made as before and the cracked gas analyzed as described in Part I. The work of high pressure was taken up first. After the furnace had heen brought u p t o temperature, gas was run through it a t atmospheric pressure, a sample taken, and the pressure raised t o z j lbs. While equilibrium was being reached in the furnace the first sample was analyzed. I n the same way the pressure was raised t o 50, 7 j , and roo lbs., the temperature being held constant meanwhile. This usually constituted a day's work. The next day the work was repeated a t the next higher temperature until the whole range of temperatures had been recovered. The results may be observed in the following graphs (Fig. 9 ) . From the results of the work i t seems t h a t t w o entirely different reactions take place in the cracking, and they may be divided into those t h a t take place below 700' C. and those t h a t take place above 7 0 0 ° C. Up t o 700' C. increase of pressure causes increase of unsaturated hydrocarbons and hydrogen in t h e cracked gas. At 700' C . , however, a sharp change is noted, and from that point on increase of pressure d e creases the amounts o€ unsaturated hydrocarbons a n d
Nov., 1918
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
Temperafure FIG.
10
hydrogen in the cracked gas. There is no doubt that many complex reactions are taking place in the heated tube, b u t below 700' C. decrease of volume is unquestionably taking place. Above t h a t temperature we have supposed t h a t condensation t o aromatic hydrocarbons takes place. This point is well sustained by other writers, but, a t the same time, large quantities of hydrogen are split off and much methane is liberated, so the sum total of the reaction is a n increase of volume. Pressure therefore should inhibit t h e reaction. This view is confirmed by the fact t h a t in all the runs above 700' C. at atmospheric pressure the vapor in the tar precipitator is brown with tar, and tar is deposited as usual. As soon as pressure is applied, however, the vapors become colorless and no more t a r is deposited. The first increment of 2 j lbs. pressure was sufficient t o prevent t h e formation of tar, in all cases but the last (850' C.), where a slight fog persisted until a pressure of 5 0 lbs. was reached. The curves (Figs. I O and 1 1 ) in which percentage composition of t h e gas is plotted against temperature, the pressure being constant, show great similarity among themselves and with the simple curves representing runs a t atmospheric pressure. I n the diminished pressure work, the procedure was the same as above except t h a t the samples had t o be taken in a slightly different way, t h a t need not be described here. The temperature was held constant while readings a t atmospheric pressure, at 61, 46, 31, and at 16 cm. of mercury were taken. The exact degree of diminished pressure was rather difficult t o maintain, b u t by careful adjustment of a stopcock which permitted access to the atmosphere at one end of a T-tube, the other end being connected to the t a r precipitator, this was finally accomplished. The results (Fig. 1 2 ) bear out in detail those obtained with increased pressure. As the pressure becomes less, the content of unsaturated hydrocarbons and of hydrogen in the cracked gas becomes less a t all temperatures up t o 700' C. This time 7 5 0 ' seems to be the transition point, the percentage of hydrogen falling slightly while the percentage of olefines increases somewhat. Thereafter the percentage of hydrogen decreases rapidly with diminished pressure
907
Pmperafure FIG. 11
while the percentage of unsaturated hydrocarbons increases correspondingly. Tar made its appearance a t 750' C., b u t diminished with the pressure. Very little tar was deposited at any temperature under diminished pressure, which may be due t o two things. Under conditions of diminished pressure the gas is not exposed to the effect of heat as long as it is under atmospheric pressure. On the other hand, the formation of t a r is a condensation, and as such would be impeded by diminished pressure. The point brought out is t h a t the formation of tar takes place in two stages. The first one involves splitting of the saturated bodies into unsaturated bodies, with the splitting out of hydrogen and consequent increase of volume. Diminished pressure accelerates this stage and in one case the percentage of unsaturated hydrocarbons rises as high as 39 per cent, which is considerably more than attained at any other time. The next step, however, requires the condensation of these unsaturated bodies into aromatic bodies, and pressures should favor this. I t is true t h a t hydrogen may split out also a t this point, but, even so, condensation is difficult t o effect under diminished pressure. The curves in which percentage composition is plotted against temperature, while the pressure is held constant, show a constantly increasing maximum for the unsaturated bodies and also show that this maximum is reached at a higher temperature as the pressure becomes less. I n short, the pressure experiments definitely show t h a t the main reaction concerned in the formation of aromatic hydrocarbons from straight-chain hydrocarbons of low molecular weight begins a t a temperature around 700' C., and is a reaction t h a t proceeds with increase of volume despite the fact t h a t i t is a condensation. They show, furthermore, t h a t the reaction proceeds in two steps, the first of which is impeded by pressure, the second of which is impeded by diminished pressure. Diminished pressure, however, largely increases the yield of unsaturated hydrocarbons, and as these will later be shown to be valuable substances, a new way is opened for the production of these bodies in large amounts.
T H E J O U R N A L O F I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y Vol.
908
Showtna Ef:ect
o f P r e s s u r e on Comoostfion of Croched 60.5
Pressure
Fro. 12 THE REACTION SIMPLE STRAIGHT-CHAIN HYDROCARBONS
----)
AROMATIC HYDROCARBONS
There are two views held concerning the formation of aromatic bodies from the straight-chain series. The theory claiming the largest number of adherents is the decomposition of the saturated hydrocarbons step by step t o acetylene, which then polymerizes t o benzene.' The other view holds t h a t aromatic bodies can be formed from olefine bodies directly without going through the acetylene stagee2 It must be admitted t h a t little experimental evidence has been produced to substantiate this latter view, the chief reliance being placed in the fact t h a t i t has been impossible t o detect acetylene a t any stage of the reaction. It was decided t o investigate the reaction concerned in this work, with the hope t h a t we could fit i t t o one or the other of these views. The evidence of the polymerization of acetylene to benzene is incontrovertible, so this work was not repeated. The work was therefore begun by mixing known amounts of acetylene with the gas used in our previous work. This was accomplished by passing a definite amount of acetylene through the referee meter into a n empty gas holder, following this b y the addition of sufficient natural gas condensate t o make u p a mixed gas of desired acetylene content. The mixture was allowed t o stand a day before being used, t h a t thorough mixture might take place. I t was then run through the cracking tube in the original apparatus and the cracked gas allowed t o bubble through ammoniacal silver nitrate, followed by ammoniacal cuprous chloride. The following mixtures were experimented with: Per cent
.......... .......... .......... .......... ........... ..........
Acetylene.. Acetylene.. Acetylene.. Acetylene.. Acetylene. Acetylene..
0.1
0.5 1 .O 2.5 5.0 10.0
Natural Natural Natural Natural Natural Natural
1st drop at 29' C.. 29'-50 5Oo-6O0....... 60'-64'. Residue..
..... ..... ......
Berthelot, Ann. chs'm., [4] 9, 469. D. T. Jones, "The Thermal Decomposition of Hydrogenated Hydrocarbons," J . Chem. Soc., 107 (1915), 1582, and "The Thermal Decomposition of Low Temperature Coal Tar," J . SOC.Chem. Ind., 36 (1917), 3. 1
2
.................... ............................... ........................ .............................. ............................. LOSS ..................................
Per cent 99.0 99.5 99 .O 97.5 95 .O 90.0
As the mixed gas was led through the cracking tilbe, the temperature was raised from a n initial tempera-
No.
II
ture of 550' C. t o one of 950' C., by steps of 50'. In every case the test for acetylene was positive. I n the 0.I per cent and 0 . 5 per cent mixtures the test was less pronounced at higher temperatures than a t lower ones, b u t this was t o be expected, as acetylene in the presence of large amounts of hydrogen passes p8rtly to ethylene or ethane, and partly bTeaks u p into carbon and hydrogen.' The point to be noted here is t h a t sufficient acetylene remains even in a 0.1 per cent mixture, which has been passed through a tube slowly (0.5 cu. f t . per hr.) and cracked a t a temperature up t o 950' C., t o give a decided test for a triple bonded component. We must conclude, therefore, t h a t acetylene, if formed in appreciable amounts in the thermal decomposition of natural gas condensate, would not entirely decompose again. I n any event, enough would remain t o give a reaction with silver nitrate or cuprous chloride. Although the absence of acetylene has been reported in this connection, we repeated the tests, using clean gas, and could find no trace of acetylene. Examination of the bromides formed b y passing the cracked gas through bromine under water and cooled with ice and salt showed no tetrabromacetylene. The absence of acetylene or other triple bonded hydrocarbons seems t o establish the fact t h a t aromatic formation is not in this case dependent upon their formation. I n order t o study the unsaturated bodies more thoroughly, about 500 g. of mixed bromides were prepared as described above. It has been pointed out before that methane was the only hydrocarbon remaining when the gas was cracked a t a temperature of 750' C., or above, so simple bromides of ethane or propane need not be looked for. I n order t o make certain t h a t substitution was not taking place in the methane, however, a preliminary experiment was run in which the bromine was used dissolved in carbon tetrachloride. Hydrobromic acid is given off from such a solution when substitution takes place as opposed t o addition. The gases, after passing through the bromine solution, were cooled t o -zoo C., and then passed through glass wool a t the same temperature t o catch any volatilized bromine vapor. They were then allowed t o bubble through standard potassium hydroxide. Back titration showed t h a t almost no acid had come over, and therefore t h a t substitution was not taking place to any appreciable extent. The 5 0 0 g. of mixed bromides prepared above were distilled under 4 cm. vacuum with the following results:
...... ...... ......
Gas Condensate, Gas Condensate. Gas Condensate. Gas Condensate., Gas Condensate.. Gas Condensate.
IO,
Volume Per cent
...
6.5
35.0 45.0
10.0
3.5
Sp. Gr.
...
1.46 1.89
2.13
...
100.0
N o trouble was experienced in the distillation, as decomposition of the bromides did not take place. The liquid distillates were then mixed and redistilled at atmospheric pressure with the following results: 1
Bone and Coward, LOG.cit.
Nov., 1918
T H E JOURNAL OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY Volume Per cent
909
Tube 3. Hydrogen mixed with a large amount of the vapor of methane will pass on uncondensed. 1 :442 Some solid did appear in Tube I, showirg t h a t the 2.108 precipitator does not remove all the benzene; Tubes 2.070 2 and 3 contained liquid condensates. These con... densates were very mobile, almost colorless liquids 100.0 with pronounced odors. I n Tube 2 we are concerned with the presence of The residue from t h e vacuum distillation was almost butadien, so after 2 0 t o 30 cc. of liquid had collected, entirely tetrabrombutane, while a large amount of the tube was disconnected and allowed t o warm up this same material was extracted from the thick liquid slowly to - 2 0 ' C. The propylene all boiled off in t h a t remained in the distillation flask above. It is the process and t h e volume shrunk to less than oneevident t h a t the bromides are, as reported b y Zanetti,' half. Keeping the tube in the cooling mixture, a mixture of ethylene bromide, a smaller amount of pure bromine was dropped into the liquid which still propylene bromide, and from I O per cent t o 20 per remained. A reaction of almost explosive violence cent tetrabrombutane (butadien tetrabromide). Small took place and t h e liquid began t o boil rapidly. amounts of other bromides are also present, b u t no Addition of bromine was continued until the liquid acetylene tetrabromide (b. p. 137' C., sp. gr. 2 . 9 ) remained slightly red, then a little liquid from Tube 3 was found. The bromine content of the ethylene was added t o combine with the excess bromine. About and propylene bromides was checked up by analysis half the liquid remaining, after t h e propylene had and found t o be correct within 0 . 5 per cent. The boiled off, was lost in the bromination process which melting point of the tetrabrombutane was found t o had generated enough heat to keep the bromides be 117' C. (correct value = 118' C.) and its bromine formed liquid for some time, even though the bath content corresponded t o t h e formula CbHOBr4. C. Shortly afterwas still at a temperature of -20' The source of the tetrabrombutane needs inquiry. ward, however, the whole residue in the test tube But very little butane exists in t h e original gas, which, solidified. The tube was withdrawn, a portion of t h e as remarked before, consists of almost nothing b u t crystals recrystallized from alcohol, and the melting ethane and propane. The butane content is not large, point determined. M. p. = 117' C. No butane a n d accounts in no way for the large yield of tetra- dibromide, which would result from the presence of brombutane. A building-up process must therefore butene, was observed. This experiment was repeated, b u t this time after be responsible €or the greatest part of t h e yield. It is a known fact t h a t butadien yields tetrabrombutane 30 t o 40 cc. of liquid had collected in Tube 2 , i t was on bromination and, since the presence of this com- disconnected and closed with a stopper and delivery pound was suspected, experiments were carried o u t tube, whose end was beneath bromine covered with water and cooled with a mixture of ice and salt. The t o isolate it. C. I n order t o prove this point, the gases, after cracking gases evolved as Tube z warmed up t o - z o o were allowed t o bubble through the bromine until t h e a t 850' C., and after all t h e tar was removed by the -zoo C. was reached. Another tube temperature of precipitator, were cooled down in three stages. The cleaned gas was slowly led through three wide test containing bromine was then substituted for the first tubes placed in thermos bottles containing cooling one and Tube 2 , containing the condensed gas, was liquids. I n Tube I, t h e gas was cooled t o -3' C. removed from the thermos bottle and allowed t o come Any benzene still present was frozen out in this tube, t o room temperature, the evolved gases passing through Tube 2 was kept at a temperature of -90' C. by the bromine as before. All the liquid in the tube C. was means of a mixture of alcohol and liquid air, while had vanished before a temperature of -2' Tube 3 was placed in liquid air direct. The tem- reached. By distillation, propylene bromide was readily obperature of t h e gas in this last tube was about tainable from the first tube of bromine as a heavy, -170' C. almost colorless liquid, with a pronounced and rather The boiling points of some of the substances we sweet odor, b. p. = 139.8' C., sp. gr. = 1.942a t 1 7 ' C. are dealing with are given as follows: The liquid in the second tube solidified and was identiMethane........ -164' Propylene ........ - Soo fied by its melting point as tetrabrombutane. E t h a n e . . ....... - 8 9 O Butylene.. ....... - S o Tube 3 was now disconnected from the apparatus Butadien ......... Propane. ........ - 39O Butane ......... + 0.6O Hydrogen. ........ -256O and transferred to a thermos bottle contahing a mixEthylene ........ -103O Benzene, f. p.. .... + ture of liquid air and alcohol a t a temperature of Ethane, propane, and butane can be ruled out a t -115' C. Violent ebullition took place as i t warmed once, as methane is the only saturated hydrocarbon up t o this temperature, and methane (b. p. = -164' C.) remaining in the gas a t this temperature. Assuming boiled off. A clear liquid remained which was pracall t h e rest of t h e gases t o be present, it is evident tically all ethylene. I t was identified by brominating t h a t only benzene will be deposited in Tube I . Pro- a small portion of it and taking the melting point pylene, butylene, and butadien will condense in Tube (m. p. = t-9' C.). 2 , while ethylene and some methane will condense in The experiment was repeated once more, b u t this 1 LOG. cit. time, after the propylene had boiled off Tube 2 a n d Temperature 1st drop "d; 89' C . .
........... 890-125 ................. 10.3 125D-1330.. ............... 44.7 Ethylene Bromide Fraction 133'137O.. .............. 12.8 Propylene Bromide Fraction Residue. .................. 2 9 . 4 Loss. ..................... 2.8
Sp. Gr.
5 O
5 O
.
T H E J O U R N A L O F I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y Voi.
9=0
the methane off Tube 3, the remaining material as it volatilized was led in each case to a small gasometer. The gas so obtained represented a mixture of approximately 30 per cent butadien and 7 0 per cent ethylene. This mixed gas was passed through the original apparatus and cracked a t a temperature of 850’ C. A small amount of tar similar in all respects t o the original t a r was obtained. Its specific gravity was I . 010. Further analysis was not attempted, owing t o the small amount formed. The cracked gas was then analyzed and shown t o be 50 per cent unsaturated, 2 0 per cent methane, and 30 per cent hydrogen. The cracked gas, when cooled down as before, gave a large amount of ethylene in Tube 3 along with some ethane. Tube 2 had only a trace of condensate, showing t h a t all the butadien had combined.’ CONCLUSIOKS
The presence of butadien in large quantities in the cracked gas can be explained only by the combination of two molecules of ethylene, with the splitting off of hydrogen, thus HZC = CHz
+ HzC
= CHz + HzC = C H - H C = CHz
+ Hz
If now we presume t h a t another molecule of ethylene can unite with the butadien, we have all the necessary steps for the formation of benzene. D. T. Jones2 finds t h a t cyclohexane on heating t o 500’ C. passes t o cyclohexene which then decomposes in two ways, yielding benzene on one hand and butadien with ethylene on the other.
CH The present work merely requires the union of the ethylene and butadien of this equation t o form cyclohexane which most likely has no separate existence, breaking down a t once into benzene and hydrogen. This reaction involves a n increase of volume and would therefore be inhibited by increase of pressure, a fact which has already been proved true above. I n support of these views we have the work of Jones,2 who inclines t o the belief t h a t olefines condense t o aromatic bodies. He states, “It is highly probable t h a t a necessary transient stage is the formation and condensation of the stable conjugated double linking, -CH = C H - CI-I = CH-.” The presence of this For references on the preparation and properties of butadien see J . Chem. Soc., 27 (1874), 406; J . Chem. SOC.,48 (1886), 80; and Ann., SO8 1
cit.
No.
II
linkage in its simplest form, butadien, has been demonstrated above. Staudinger,l starting with isoprene, showed t h a t 45 t o 5 5 per cent of this material was converted into a t a r by passing it through a tube a t 7 50’ C. This t a r contained aromatic hydrocarbons similar t o those Zanetti found in his tar. Staudinger also cracked butadien alone and obtained from it a t a r t h a t contained about 2 5 per cent benzene. It seems t o be established, thereiore, t h a t diolefines, on cracking, pass in large part t o closed chain bodies. It has been demonstrated in this work t h a t simple olefines and diolefines are produced by the cracking of the ethane-propane fraction of natural gas condensate. The aromatic bodies found in the tar comes from the condensation of the diolefines and this therefore gives LIS all the necessary steps in the formation of aromatic bodies from straight-chain hydrocarbons of less t h a n four carbon atoms, and without the necessity of passing through the stage of acetylene. SUMMARY
I-It has been shown t h a t most metals are without action on the reaction Paraffin hydrocarbons Aromatic hydrocarbons. The metals nickel, iron, and cobalt are anticatalysts for the above reaction, but promote t o a marked degree the reaction Paraffin hydrocarbons Carbon Hydrogen. 11-The effect of temperature and pressure on the production of aromatic hydrocarbons has been studied. I t has been pointed out t h a t a temperature of 850’ C. is most favorable for the formation of liquid t a r and t h a t the formation of complex aromatic bodies increases with the temperature. 111-Increase of pressure inhibits the formation of tar while diminished pressure increases the yield of unsaturated bodies but also decreases the actual yield of tar. IV-Butadien has been isolated in fairly large amounts from the unsaturated bodies produced in the thermal decomposition of natural gas condensate. V-Acetylene has been shown t o be without action in the formation of the aromatic compounds. VI-Tar containing aromatic bodies has been produced from the cracking of a mixture of butadien and ethylene. VII-The most probable reaction for the formation of aromatic bodies from natural gas condensate is
*
+
Saturated straightchain hydrocarbons (Ethane)
f
Simple o1efin es
DEPARTMENT OF CHEMISTRY COLUMBIA UNIVERSITY N E W YORKC I T Y 1
Ber., 46 (1913), 2466
-
(Ethylene)
Higher olefines with conjugated bonds (Butadien)
(1899), 333. 9 LOC.
IO,
--t
Aromatic hydrocarbons (Benzene)
