Ethane Pyrolysis in the Presence of Steam D. S. CRYDER AND D. J. PORTER' The Pennsylvania State College, State College, Pa.
T
HE increasing importance of the pyrolyVarious ethane-steam mixtures were led through a silica sis Of certain petroleum fractions in the tube at different temperatures. A t each temperature level, presence Of hydrogen' data were secured for the decomposition of the gaseous pared by hydrocarbon-steam reactions, sugmixture with the tube empty, with a silica-gel catalyst, gested the possibility of simultaneous interacand with silica gel impregnated with nickel. The results tion of these petroleum fractions with steam and with the hydrogen thus produced. Such indicate that the decomposition of ethane proceeds accorda Process, if capable of realization, should be ing to the simple first-o;der reaction : C2H,If 2CH,. For of considerable interest from both practical steam the mechanism is more complicated; the data inand theoretical viewpoints. dicate the probability of a second-order reaction. Although the svrolysis of hydrocarbons alone nia; not have adirect bearing on the problem of their decomposition in the presence of steam, the underlying theories evolved as a result in the pyrolysis of propane a t temperatures below 700" C., but of the extensive work on this subject serve as a basis for the that small amounts of carbon monoxide, carbon dioxide, and development of the mechanism of hydrocarbon decomposialdehydes were formed at higher temperatures, indicating the tion in the presence of steam. Ellis (9) presented a comprepresence of some reaction between steam and propane or its hensive review of the literature on hydrocarbon pyrolysis; decomposition products. other reviewers (8, 7, 11, 15) reported additional points of The work to be described in this paper was undertaken in interest. Rice and Herzfeld (28) developed a detailed mechaan effort to determine the mechanism of ethane decomposinism for the thermal decomposition of ethane. tion in the presence of steam and the course of any interaction Fewer experimental data are available concerning the between the gases as a first step in the study of the effect of pyrolysis of hydrocarbon-steam mixtures. The pyrolysis of steam on the thermal decomposition of higher hydrocarbons. methane-steam mixtures was considered by a number of investigators and is the subject of numerous patents (8). NeuExperimental Procedure mann and Jacob (2.2) indicated that a t about 660" C. reacThe apparatus employed is illustrated in Figure 1. tions giving carbon monoxide and carbon dioxide tended to occur siniultaneously; a t higher temperatures the reaction Ethane gas, contained in the calibrated cylinder, A , was disproducing carbon monoxide predominated. An increase in placed at a constant rate by water flowing from the constantsteam concentration mas found to result in the formation of head reservoir, B , through capillary C. The ethane, dried in increased amounts of carbon dioxide together with decreased the calcium chloride tower, D,passed t o the water vaporizer, E. conversion (14). In this vaporizer, distilled water was admitted at a constant slow rate through orifice, F , which was in contact with an asbestos Steam was mentioned as a diluent in the pyrolysis of parafwick, G, internally heated by a small coil of resistance wire. fins a t 550" C. (6),in the production of olefins ( 4 ) , and in the Thus the water vaporized at a constant rate was mixed with the pyrolysis of hydrocarbons ( 8 ) . Steam was found to stabilize stream of ethane, giving a known definite ratio of steam to and to prevent further oxidation of the formaldehyde proethane in the gas mixture entering the reaction tube. The rate of water vaporization in this device was controlled and measured duced in the partial oxidation of ethylene ( 2 ) . Wright and by the rate of displacement of mercury from reservoir H by Gauger (50)stated that ethane and steam react similarly to means of a piston n-ith a calibrated screw, I . The rate of admethane and steam, but no evidence has been advanced in vance of this piston was indicated by a speedometer, J , and the support of this statement. I n the manufacture of oil gas odometer registered the integrated flow. The speedometer was used both as an indicator and reducing gear, and was driven using steam, the formation of oxides of carbon was ascribed by a small variable-speed motor. The odometer was graduated to the interaction of steam with deposited carbon rather than to tenths of a revolution, and hundredths were estimated. All to its direct interaction with hydrocarbons (3); Morgan (91) tubing from the vaporizer to the condenser was maintained at considered that both reactions occurred. Ode11 (23) reported or above 100" C. to prevent condensation. The silica reaction tube, K , was 2.3 cm. i. d. and 58 cm. long, complete decomposition of the ethane content of a natural and had a 38-cm. length inside the furnace, L. The furnace gas during a steam-reforming operation. The treatment of was made of Nichrome ribbon wound on asbestos paper around cracking-still gases with steam a t about 870" C., followed by a section of steel pipe. Gases were introduced into the reaction conversion of the resulting carbon monoxide to carbon dioxide tube through a silica tube, M , 8 mm. 0.d. and 4 mm. i. d., flared at the t o p t o form a catalyst support. The reaction tube was by further reaction with steam a t 455" C., was described as a sealed at top and bottom with charred corks. A platinum, commercial process for the production of hydrogen (10, 25). platinum-rhodium thermocouple enclosed in a sheath, N , of Lang and Morgan (19) found that steam is an inert diluent 5-mm. silica tubing was embedded in the center of the catalyst mass, 0. Thermocouple readings were taken by means of a 1 Present address, 1539 State Avenue, Coraopolis, Pa. 667
668
INDUSTRIAL AND ENGINEERIKG CHEMISTRY
VOL. 29, NO. 6
Brown portable potentiometer reading to 0.01 millivolt. The Materials temperature distribution in the section of the tube containing the catalyst was fairly uniform, with variation of less than 10" C. Three types of runs were made in this study: with no at the highest temperatures, and the catalyst was located in the catalyst, with a silica-gel catalyst, and with a nickel catalyst section of maximum temperature* Heat input to furnace was supported on silica gel, They are designated by the letters indicated by ammeter a. B, G, and N, respectively, in the first letter of the composite Gases leaving the reaction tube were conducted through the run numbers in Table I. condenser. P. Drovided with a thermometer, &, in intimate contact with theexit conThe silica gel was a densate and gases. Any uniformly sized (10- to condensate was collected 20-mesh), u n t r e a t e d , in the graduated tube, R, for me as u r e m e n t and prepared sand.* Prepreservation. The parliminary experiments tially dehumidified gases s h o w e d that the gel were collected over salt could be maintailled a t solution saturated with the gas mixture in bottle 800" C. without sinterS , and determination of ing or showing visible the product volume was of deterioration. signs made by measurement of In the runs atr900" and the amount of confining solution removed, T. 1000" C . the g e l sinA constant pressure tered, often t o a smooth was maintained i n t h e globular form. entire system by means The nickel c a t a l y s t of the mercury contact, was prepared by moisU , actuated by fluctuating gas pressure, which tening silica gel with a controlled valve V ; the solution containing latter, in turn, regulated 74.33 grams of nickel the efflux of salt water. n i t r a t e hexahydrate By adjusting stopcock W to approximate the [Ni(N0&6H20]in 100 rate of outflow, the conof s o l u t i o n . Ten cc. t r o l device worked cubic c e n t i m e t e r s of smoothly. FIGURE 1. DIAGRAM OF APPARATUS this solution were used At the conclusion of a run the gas was diswith 15 grams of silica placed by illowing saturated salt solution to flow from X into gel (0.1 gram of nickel per gram of silica). After thorough S. Gas samples were collected through sampling arm, Y , in mixing, the impregnated gel was air-dried on a steam bath 125-cc. bottles by displacement of saturated salt solution, and with constant stirring. Small portions of this air-dried mixthe bottles, closed with corks, were inverted in a mercury-filled pneumatic trough. Samples thus kept remained unchanged over ture were heated with conqtant shaking over the flame of a long periods. Bunsen burner in a small porcelain casserole until the nitrate Of the variables generally considered in a problem of this was entirely decomposed. About 6 to 7 minutes were required type-namely, contact time, temperature, surface-volume ratio, for calcination of the product. Portions of 100 to 375 cc. of catalytic materials, reactant ratios, and pressure-all but pressure and contact time were subjected to controlled variation. The the calcined product were then placed in a Pyrex glass tube data secured were obtained by passing steam-ethane mixtures mound with Nichrome ribbon. After flushing the air from the through the reaction tube at the rate of 125 cc. per minute, tube with a stream of hydrogen, the contents were heated t o measured at 100' C. and 760 mm. Steam-ethane ratios apabout 220" C.; at that temperature the nickel oxide was proximating 2 to 1, 4 to 1, and 8 t o 1 were used. At the start of a run the charred cork connections at the top readily reduced to a n apple-green oxide. The temperature and bottom of the reaction tube were made gas-tight, the furnace was increased to 330" C., but reduction was not complete was brought to the desired temperature, and the water-flow until the material was further heated in hydrogen a t 360" to rate through capillary C was adjusted to give the desired flow 370" C. for several hours. The reduced catalyst was cooled rate of ethane. When all conditions were adjusted, the motor driving the water-feed mechanism was brought up to the correct in hydrogen and stored in a gasketed tin container. speed for the ratio desired, as indicated by the speedometer, The percentage of free space in the catalyst mass was found and the mixed gases were allowed to flow through the system by determining the amount of liquid that could be added to a and were vented. After allowing sufficient time for a thorough definite volume of the catalyst to fill up the available openings. flushing of the apparatus, the thermocouple was removed and 25 cc. of catalyst were added, the thermocouple tube was reSeveral determinations using water and carbon tetrachloride placed, and conditions were maintained constant for 5 or 10 indicated 68 per cent free space. minutes more. The ethane was obtained from the Carbide and Carbon The vent line was then closed, and stopcock W was opened sufficiently to allow the pressure regulator to function. Time Chemicals Corporation with a stated analysis of ethane 96, was taken soon afterward, and the following data were recorded methane 3, and propane 1 per cent. at intervals of 5 or 10 minutes: time, ethane volume in the cylinder, ethane temperature, water-feed revolutions, volume of csndensate, condenser temperature, exit gas volume and temAnalytical Methods perature, thermocouple electromotive force, and barometric pressure. In order to secure enough gas for analysis and to Gas analyses were performed with a standard Burrell type obtain a satisfactory degree of accuracy for the material balof apparatus equipped with a new type of slow-combustion ances, it was necessary to introduce 300 cc. or more ethane in pipet designed in this laboratory ($4). Special absorbing each run. At the close of a run, stopcock W was closed and the pressure solutions were neceseary for some of the constituents deterin the system was allowed to build up to a few centimeters of mined. Mercuric cyanide in sodium hydroxide was used for water above atmospheric; then X and A were closed. The the estimation of acetylene (29), and 8 per cent oleum satuthermocouple tube and condenser assembly were removed, and rated with silver and mercuric sulfates for ethylene (16). the reaction tube was withdrawn from the furnace for removal and inspection of the catalyst. The outfit was then reassembled The absorbing solutions were protected from the atmosphere in preparation for the next run while the gas samples were taken. by a 1-cm. layer of medicinal oil on the exposed surfaces. A Any carbon deposits in the reaction tube were removed by com9 Obtained from the Silica Gel Corporation, Baltimore, Md. bustion in a stream of air or oxygen.
JUNE,1937
4 f
INDUSTRIAL AND ENGINEERING CHEMISTRY
.3"g$;rioz :?hddm000000
--la0
.
00000
.. . .. .. ..
, , . . . . . . 0 . .. . . . . . . . .. .. .. .. .. .. .. . . . .. .. . i
,
669
preliminary colorimetric qualitative test for acetylene was made on the gas from each run (W8). Consistent errors were unavoidable in most absorptions because of the lack of specific solvent action of the individual reagents. These errors were largely corrected by repeated passage of the gas sample through each absorbent until either a constant reading or a constant difference was obtained, with adjustment of the apparent percentage by proper consideration of the constant difference. The collected condensates were qualitatively analyzed for simple organic molecules by standard methods. The essential data for the steam-ethane runs appear in Table I. The items t a b u l a t e d were calculated from the experimental data. All gas volumes are expressed in cubic c e 11t i m e t e r s a t 100' C. and 760 mm. pressure. For the sake of brevity the data a t 400", 600", and 800' C. are omitted from Table I. They appear, however, in the various graphs.
,,
3
Sample Calculations From the volume and analysis of the product gases a material balance on each element involved (carbon, hydrogen, oxygen) was made, using cc.-atoms as units. The method is as follows for run 16B: Carbon input: (281) (2) = 562 cc.-atoms Hydrogen input: (281) (6) (281)(7.95)(2) = 6156
+
cc.-atoms Oxygen input: (281)(7.95) = 2235 Carbon recovery: 524 cc.-atoms = 93.24y0 Apparent carbon deposition = 6.76y0 Ethane decomposition: 562 - (0.3626)(394) (100) = 49.2% 562
Free volume of reaction tube: Empty = 25 cc. With catalyst = 17 CC. Barometer: 738.4 mm. Temperature: 702' C. K
'
1= -1 1 n 3 = - I n 6
n2
4.31
(60)(29) = 4.31 sec. 562 = 0.157 286
Discussion J
Ir:
I n the first few experiments, in which appreciable quantities of carbon monoxide and carbon dioxide appeared in the exit gases, it was proved that steam was not a n inert diluent in the pyrolysis of ethane, since these oxygenated products could have been formed only by some reaction of steam, as C2Hs CzHe
+ 2H20 +2CO + 5Hz
+ 4Hz0 +2C02 + 7H2
Further evidence of the activity of steam when present in the pyrolysis of ethane is shown by the differences in the results caused by varying the steam-ethane ratios a t a constant temperature. The data obtained indicate that interaction of steam and ethane in the presence of nickel begins a t 430" C.; with the empty silica tube or in the presence of silica gel alone this temperature is 600 O C. I n the blank and gel runs a n increase in the proportion of carbon monoxide and carbon dioxide accompanies an increase in the steam-ethane ratio.
.
INDUSTRIAL AND ENGINEERING CHEMISTRY
670
.
The catalytic activity of nickel results in greater formation of the oxides of carbon a t a given temperature; and because of the increased effect of the nickel-catalyzed water-gas reaction a t temperatures below 830" C. (at which temperature the equilibrium leads to the formation of carbon dioxide and hydrogen), a n increase in the concentration of steam results in the conversion of carbon monoxide to carbon dioxide. Consequently, in the runs with nickel catalyst, the decrease in carbon dioxide and the increase in carbon monoxide percentages with increasing steam-ethane ratios agrees with the predicted trend.
TEMPERATURE
- 'C.
FIGURE 2. K "./ K., VALUESFOR WATER-GASREACTION Steam-Ethane
Ratio
Blank
Nickel
8-1 4- 1 2-1
Gel
0 0
A
+X
0
To confirm the importance of the water-gas reaction, the partial pressures of carbon monoxide, water, carbon dioxide, and hydrogen in the gas leaving the catalyst were calculated. From these data the experimental values of the equilibrium constant for the reaction CO
+ HzO +COz + Hz
were calculated. The product-quotients K"
= PCOZ PCO
x PH? x PHzO
divided by the values of K p a t each temperature, gave ratios K"/K,, from which the approach towards an equilibrium with respect to this particular reaction could be detected (Figure 2). I n the blank and gel runs this ratio exceeded 0.14 only a t 1000" C., a t which temperature a value of 0.24 was attained. I n the presence of nickel, however, all the values fell between 0.5 and 1.5 (except a t 800" and 900" C., a t which temperature the actual driving force was small). These figures show that nickel is active in catalyzing the water-gas reaction above 500" C. Substantial support is thereby given to the postulation that, of the two molecules carbon monoxide and carbon dioxide, the former is a primary and the latter a secondary product in the interaction of steam and ethane. Ethylene, formed by the dehydrogenation of ethane and an important constituent of the reaction products in the absence of nickel, was present in greatest amount a t 700" C. (28 per cent). At higher temperatures either the ethylene formed was decomposed to yield methane, or a shift in the relative importance of the initial reactions led to the increased formation of methane from ethane a t the expense of reactions yielding ethylene. Although the steam-ethane ratio has little effect on the amount of ethylene formed, the methane content was diminished by increasing the steam concentration. I n the presence of nickel, ethylene was found only a t 800' 6. and then in relatively small amounts which decreased with increasing steam concentrations. The formation of methane
VOL. 29, NO. 6
was greatest a t 500" C.,a t which temperature some ethane may possibly have decomposed to yield methane and hydrogen direct, as observed by Sabatier (27). Hydrogen, a product of all the decomposition reactions of ethane alone or with steam, was present in amounts increasing with the temperature and the steam-ethane ratio, either in the presence or absence of catalytic material. Carbon deposition, as measured by material balances based on oxygen and hydrogen, was greatest in the blank and gel. runs, and was decreased by the addition of steam. These various observations are apparent from Table I. For further aid in the interpretation of the data, curves were prepared showing the various gas analyses and other quantities plotted against the steam-ethane ratio, with temperatures as parameters. The three or four points representing runs with one catalyst a t the same temperature were connected by straight lines, and from the intersections of these lines with the 2-1 and 8-1 steam-ethane ratio lines, values were obtained from which curves showing the variations of several quantities as functions of the temperature were constructed. Representative temperature curves are presented in Figures 3 and 4. In the curves of carbon dioxide for the various conditions (Figure 3A) no consistent differences between the blank a n d gel runs appear, although a t higher temperatures the absence of gel catalyst results in increased amounts of carbon dioxide. These amounts of carbon dioxide, though relatively small, increase steadily with rising temperature. I n the presence of nickel an opposite behavior is noted, with an indicated maximum formation near 450" C. and a steady decrease with rising temperatures. T i e curves of carbon monoxide percentage (Figure 3C) increase uniformly in the presence or absence of nickel from zero to maxima a t the highest temperatures. This similarity in behavior, considered t40getherwith the difference in the carbon dioxide curves, can most simply and logically be explained by assuming that carbon monoxide is the initial product of the reaction of steam with ethane and is subsequently converted to carbon dioxide through the agency of the water-gas reaction. This is a reaction which proceeds strongly towards the formation of carbon dioxide a t low temperatures when catalyzed by such a material as nickel, but yields carbon monoxide a t temperatures above 830" C. The partial pressures of the components correspond more nearly to equilibrium values in the nickel runs than in the blank and gel runs. The formation of ethylene (Figure 3 B ) occurs in the absence of the gel, and a t some temperatures it forms a considerable proportion of the product. The appearance of the curve indicates that reactions yielding ethylene cominenced between 500" and 600" C. and increased rapidly with temperature to about 750". At this point the decomposition of ethylene, which is more stable than ethane by a temperature differential of approximately 200" C., becomes a factor of importance. With further temperature increase, less ethylene appears in the products of reaction until a t 1000" C. less than 3 per cent remains. In the presence of nickel, ethane decomposes without the formation of ethylene, or the ethylene itself decomposes very rapidly a t all temperatures except 800" C., a t which point a small amount of ethylene is found. The formation of ethylene a t this temperature is accompanied by a decrease in the decomposition of ethane and steam, a decrease in the K " / K , ratio (Figure 2) for the water-gas reaction, and a n increase in the carbon deposition. This indicates a loss of the specific catalytic action of the nickel, apparently because of the formation of a layer of carbon which destroys the activity of the catalyst. Figures 3F and 4A show that in the presence of nickel, ethane decomposition begins at 160" lower and is substantially complete a t 300" C. lower than with the blank tube.
JUNE, 1937
INDUSTRIAL AND ENGINEERING CHEMISTRY
671
I
I
I
I
rSo
I
I
I
TEMPERITURE
I
I
I
'C.
FIGURE 4. REPRESENTATIVE TEMPERATURE CURVES
Reaction Mechanism A mechanism for the reaction occurring during the pyrolysis of ethane in the presence of steam may be formulated by utilizing the free-radical theory developed by Rice (96)and co-workers, who showed that methyl, ethyl, and hydrogen radicals are derivable from ethane, the former by direct scission, the latter by reaction chains. In the presence of large proportions of steam, such reactions as CH3 CHI CHa CzHj CzHj CzHj
+ Hz0 _I, CHa + OH + HzO e CH30H + H + HzO CHaO + Hz + H2O e CzHs + OH + HzO +CzH50H + H
+ HzO e Cz&O + H2
(1) (2) (3)
(4) (5) (6)
constitute typical additional possibilities which must be considered. Since the probability of collisions of methyl radicals with water is greater in the presence of excess steam than is the probability of collisions with ethane, reactions 1 to 6 are of considerable interest. Reactions 3 and 6 involve the severance of two oxygen-hydrogen bonds and will occur much less frequently than the others of this series in which only one oxygen-hydrogen linkage is destroyed. The activation energy, E, for the reaction TEMPERATURE
' C
FIGURE3. REPRESENTATIVE TEMPERATURE CURVES
CHa
+ CzHe +CH, + CzHz
(7)
INDUSTRIAL AND ENGINEERING CHEMISTRY
672
is 20 Calories (16). Reaction 2 involves the rupture of an oxygen-hydrogen bond and the formation of a carbon-oxygen linkage. The rupture of a n oxygen-hydrogen bond requires about 103 Calories (I8),and the formation of a carbon-oxygen linkage liberates energy in an amount less than the activation energy (SO Calories) of the decomposition (255) CHsOCH,
-
CH3
+ CHaO
It is apparent therefore that for reaction 2, E is greater than
(103 - 80 =>23 Calories. The difference between the activation energie. for reactions 2 and 7 is therefore equal to or greater than 3 Calories. Hence a t 700" C. the ratio of probabilities of occurrence of these two reactions (P,/P, = e(ao'JO/RT) = 4.7) indicates that in gaseous mixtures containing excess amounts of steam, as in this study, reaction 2 will probably occur a t the same rate as 7. By a similar argument it may be shown that reaction 5 will also occur, and that reactions 1, 3, 4, and 6 are of minor importance. The methanol and ethanol molecules may also decompose according to the following chain reactions with the formation of condensable oxygenated products: R-
+ CHsOH +RH + -CHzOH
or/CH30-
%H&+
~
R
H-
+H-CH~CH~OH-+C ~ +H -OH ~
Carbon monoxide, hydrogen, and traces of methane were reported in the decomposition products of formaldehyde at 800" C.; a t 475" carbon dioxide, hydrogen, carbon monoxide] and saturated hydrocarbons were found (16). Ethanol decomposes to yield methane and formaldehyde a t 800" C. in the presence of carbon. Acetaldehyde forms methane and carbon dioxide a t 400" C., and a t higher temperatures hydrogen, ethylene] and carbon (16). The collected condensates from all the runs were analyzed to detect any organic constituents. Qualitative tests indicated the presence of acetic and formic acids. The aldehyde grouping was detected, but formaldehyde was not indicated; consequently the presence of acetaldehyde was inferred. Apparently any formaldehyde formed during the reaction was instantaneously decomposed, and not enough remained unchanged to respond to the resorcinol test. The formation of acetic and formic acids may be represented as follows: CHsCO HCO
CHIGOOH + H + HzO + H20 e HCOOH + H
The decomposition of formic acid can yield carbon dioxide, HCOOH
COz
+ Hz
but the minute traces of this acid in the condensate indicate that the chief source of carbon dioxide is the carbon monoxide, formed by the decomposition of formaldehyde or possibly by direct action of steam on carbon. The formation of the observed carbon deposits can reasonably be explained by assuming that acetylene, formed by the dehydrogenation of ethylene, is instantaneously decomposed into its elements (I?'). The small traces of acetylene remaining in the gases in the cases of large carbon deposition support this viewpoint. Rice (36)presented convincing evidence that the reaction C2Hs
represents the initial step in the decomposition of ethane a t low pressures and showed it to be a first-order reaction with an activation energy of approximately 80 Calories; therefore i t is evident that, if the decomposition of ethane in the presence of steam can be shown to be of the first order, the above mechanism is given further support. From the number of moles of ethane entering and leaving the reaction zone, the average temperature, the free reaction tube volume, and the length of a run, reaction velocity constants were calculated for the decomposition of ethane, assuming various orders of the reaction. The values of K , for a first-order reaction showed the best agreement. These values are presented in Table I and Figure 5. Since the units
I O -STEAM-ETHANE
2CHa
A
,x -5
0
0
-+
41 2'1
BLANK
_ .
8 1
-6
___ ____ I
NiCKEL
RAT10
p-
\
VOL. 29, NO. 6
O
,o -4
.
*
-3
I
I
I
I
0
-e
-I
0
I +I
REACTION RATECONSTANTS FOR FIGURE 5. FIRST-ORDER ETHANE DECOMPOSITIOK misleading results arising from the use of partial pressures in constant-pressure experiments (1). From the slope of a straight line drawn through the points representing blank and gel runs in Figure 5, the activation energy is found to be 47 Calories. Although this does not check with Rice's value of SO Calories, it is in agreement with the value of 49 which Frey ( 1 1 ) calculated from the work of Hague and Wheeler (13)and Marek and McCluer (90). I n the present experiments the extent of decomposition was too great to give accurate values of rate constants; hence the above agreement is considered sufficiently accurate to establish the mechanism presented. Any catalytic action tends to reduce the experimental value of E calculated for a homogeneous reaction: this behavior is well shown by the location of the points representing the nickel runs. The decomposition of steam was found by similar treatment to be no simple first-order reaction, but probably of the second order and involving the methyl radical. For example, the values of the first-order decomposition velocity constants of steam in the nickel runs a t 600" C. were 0.17,0.22, and 0.46; and a t 700" C., 0.057, 0.30, and 0.91, a t 9-1, 4 1 , and 2-1 steam-ethane ratios, respectively. The trial of other orders of reaction involving steam produced similar results. Proof of this fact will rest upon the determination of the concentration of methyl radical in the mixture a t any time. Though not subject to direct evaluation, it was shown that this quantity may be estimated by experiments conducted under especially selected conditions (IS).
Acknowledgment The authors wish to acknowledge the assistance rendered by A. E. Clarke, A. L. M. Bixler, A. R. Krotzer, and F. S. Hanson in securing data for this investigation.
Literature Cited (1) Benton, A. F., S.Am. Chem. Soc., 53, 2984 (1931). (2) Blair, E. W., and Wheeler, T. S., J. SOC.Chem. Ind.,41, 303-10 (1922). (3) Cowles, Henderson, and Yard, in Morgan's "American Gas Practice," p. 588, New York, D. Van Nostrand Co., 1924.
JUNE, 1937
INDUSTRIAL AND ENGINEERING CHEMISTRY
Dreyfus, H., French Patent 763,942 (May 9, 1934). Dunstan and Wheeler. British Patent 309.455 (Oct. 8. 1927). Egloff, Sohaad, and Lowry, J. Phgs. Chem., 34, 1617’(1930j. Ibid., 35, 3489 (1931). Ellis, Carleton, “Chemistry of Petroleum Derivatives,” pp. 135, 146, 181, 183, 214-17, New York, Chemical Catalog Co., 1934. Ibid., p. 149. Ibid., p. 285. Frey, F. E., IND. ENQ. CHEM.,26, 198-203 (1934). P Hague, E. N., and Wheeler, R. V., J. Chem. SOC.,1929, 378-93. Hanson, F. S., undergraduate thesis in chem. eng., Pa. State Coll., 1935. Hawk, Golden, Storch, and Fieldner, IRD. ENQ.CHEM.,24, 23-7 (1932). Hurd, C. D., “Pyrolysis of Carbon Compounds,” pp. 11, 148, 236, 330, New York, Chemical Catalog Co., 1929. Hurd, C. D., and Meinert, R. N., J.Am. Chem. SOC.,52,4978-90 (1930). Kassel, L. S., J. Am. Chem. SOC.,54, 3949-61 (1932). Kassel, L. S., “Kinetics of Homogeneous Gas Reactions,” p. 317, New York, Chemical Catalog Co., 1932. Lang, J. W., and Morgan, J. J., IND.ENQ.CHEM.,27, 937 (1935).
673
(20) Marek, L. F., and McCluer, W. B., Ibid.,23, 878-81 (1931). (21) Morgan, J. J., “A Textbook of American Gas Practice,” 2nd ed., pp. 583, 610, Lancaster, Pa., Lancaster Press, Inc., 1931. (22) Neumann and Jacob, 2. Elektrochem., 30, 557-76 (1924). (23) Odell, W. W., U. S. Bur. Mines, Rept. Investigations 2973 (Dec., 1929). (24) Porter, D. J., and Cryder, D. S., IND.ENQ.CHEM.,Anal. Ed., 7, 191 (1935). (25) Reich, G., Refiner Natural Gasoline Mfr., 11, 448 (1932). (26) Rice, F. O., and Herzfeld, K. F., J. Am. Chem. Soc., 56, 284-9 (1933). (27) Sabatier and Senderens, Compt. rend., 124, 1358 (1897). (28) Schulze, A., 2. angew. Chem., 29, 341 (1916). (29) Treadwell, W. D., and Tauber, F. A . , Helu. Chim. Acta, 2, 601-7 (1919). (30) Wright, C. C., and Gauger, A. W., IND.ENQ.CHEM.,26, 164-9 (1934). RECEIVEDMay 25, 1936. Abstracted from a thesis mbmitted by D. J. Porter t o the faculty of The Pennsylvania State College in partial fulfillment of the requirements for the degree of doctor of philosophy in chemical engineering.
CORN PROTEINS J. F. WALSH American Maize-Products Company, New York, N. Y.
U
SUALLY we think of corn proteins only as components
of animal feeds, and as such they have found extensive and ready markets. Then, too, in the wet milling industry they are classed as a process waste, and no intensive study of their possible increased value as industrial raw materials has been made. However, considerable work has been done on their isolation, identification, and characteristics since the early part of the nineteenth century; Gorham, Bizzio, Ritthausen, Dill, Chittenden, Clapp, and T. B. Osborne particularly have made available much of the necessary knowledge which has led to the development of the commercial production of a very complete range of industrial corn-component proteins. In composition and reactivity the corn proteins are similar in many respects to the natural and synthetic resin products, such as casein, shellac, cellulose derivatives, thermoplastic synthetic resins, and the protein components of soybeans. The corn proteins are not readily identified through the usual chemical reactions, but like other proteins they are separated and classified by their selective solubility in various solutions. Broadly, they are comprised of albumins which are soluble in water,, globulins which are soluble in dilute salt solutions, glutelins which are soluble in dilute alkali, and prolamines which occur only in grains and are soluble in aqueous alcohol solutions. The proteins constitute about 10 per cent of the weight of the corn substance. The major portion of these is isolated in gluten, the process waste left from the purification of starch. These gluten proteins are all substantially insoluble in water. The water-soluble albumins, and some of the more soluble globulins, have been previously removed from the corn in the steeping process from which they are isolated with the soluble salts by concentration; in this form they are used as a component of animal feeds, as noncoagulable soluble proteins, and as yeast and bacterial nutrients.
Types of Gluten Proteins The gluten is composed of approximately 50 per cent protein, 35 per cent starch, 5 per cent oil, and a small amount of fiber and mineral matter. Because of this high protein
content, the various industrial proteins are isolated from the gluten. At present these proteins include three general types-namely, the carbohydrate-free protein, the carbohydrate protein free of the aqueous-alcohol-soluble portion, and the aqueous-alcohol-soluble protein called “zein”; each is available in various special forms. The carbohydrate-free protein is produced in the form of carbohydrate oil-free, carbohydrate-free bleached, and carbohydrate oil-free bleached; all differ slightly in their characteristics and uses. As a group they are utilized principally a8 a plastic base, filler, and reactive component in cellulose derivatives and in natural and synthetic resins, and as a raw material for the preparation of amino acids, particularly glutamic acid and leucine; glutamic acid is the essential component of monosodium glutamate, the synthetic beef flavoring. An approximate analysis of the substantially carbohydrateoil-free bleached protein is: Moisture Protein Ash Oil
.
11.0%
76.5
1.2 1.1
80Yo a1coh ol-soluble Fiber Starch
35.8% 4.0
None
This group of whole proteins is, in general, thermoplastic, relatively light colored, and reactive to dilute alkali, formaldehyde, phenol, and aqueous alcohol solutions; they thus offer a base for a relatively wide range of plastic products and fillers. The carbohydrate proteins, free of alcohol-soluble portion, differ from the former type in that they are almost completely soluble in dilute alkali and substantially insoluble in aqueous alcohol solutions. Comprised principally of the glutelins and a small amount of globulins, they find use as fillers, as components of phenolic resins, and as bases in alkaline coating solutions.
Zein The third and most important group constitutes the aqueous-alcohol-soluble prolamines, or zein. First isolated in 1821 from corn or Zea mays by Gorham, it has attracted more attention than any of the other proteins because of its